Food emulsions - principles, practices, and techniques.pdf

THIRD EDITION Food Emulsions PRINCIPLES, PRACTICES, AND TECHNIQUES THIRD EDITION Food Emulsions PRINCIPLES, PRACTICES, AND TECHNIQUES David Julian McClements Boca Raton London New York CRC Press is an imprint of the Taylor & Francis Group, an informa business CRC Press Taylor & Francis Group 6000 Broken Sound Parkway NW, Suite 300 Boca Raton, FL 33487-2742 © 2016 by Taylor & Francis Group, LLC CRC Press is an imprint of Taylor & Francis Group, an Informa business No claim to original U.S. Government works Version Date: 20150713 International Standard Book Number-13: 978-1-4987-2669-6 (eBook - PDF) This book contains information obtained from authentic and highly regarded sources. Reasonable efforts have been made to publish reliable data and information, but the author and publisher cannot assume responsibility for the validity of all materials or the consequences of their use. The authors and publishers have attempted to trace the copyright holders of all material reproduced in this publication and apologize to copyright holders if permission to publish in this form has not been obtained. If any copyright material has not been acknowledged please write and let us know so we may rectify in any future reprint. Except as permitted under U.S. Copyright Law, no part of this book may be reprinted, reproduced, transmitted, or utilized in any form by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying, microfilming, and recording, or in any information storage or retrieval system, without written permission from the publishers. For permission to photocopy or use material electronically from this work, please access www.copyright.com (http:// www.copyright.com/) or contact the Copyright Clearance Center, Inc. (CCC), 222 Rosewood Drive, Danvers, MA 01923, 978-750-8400. CCC is a not-for-profit organization that provides licenses and registration for a variety of users. For organizations that have been granted a photocopy license by the CCC, a separate system of payment has been arranged. Trademark Notice: Product or corporate names may be trademarks or registered trademarks, and are used only for identification and explanation without intent to infringe. Visit the Taylor & Francis Web site at http://www.taylorandfrancis.com and the CRC Press Web site at http://www.crcpress.com This book is dedicated to my wife, Jayne, and daughter, Isobelle. Contents Preface..................................................................................................................................................... xxi Author....................................................................................................................................................xxiii 1. Context and Background.................................................................................................................. 1 1.1 Emulsion Science and Technology in the Food Industry........................................................ 1 1.2 General Characteristics of Food Emulsions............................................................................ 2 1.2.1 Definitions.................................................................................................................. 2 1.2.2 Mechanisms of Emulsion Instability......................................................................... 6 1.2.3 Ingredient Partitioning in Emulsions......................................................................... 6 1.2.4 Dynamic Nature of Emulsions................................................................................... 7 1.2.5 Complexity of Food Emulsions................................................................................. 7 1.3 Emulsion Properties................................................................................................................. 8 1.3.1 Dispersed-Phase Volume Fraction............................................................................. 8 1.3.2 Particle Size............................................................................................................... 9 1.3.2.1 Collecting Particle Size Data................................................................... 10 1.3.2.2 Presenting Particle Size Data.................................................................. 10 1.3.2.3 Mean and Standard Deviation................................................................. 13 1.3.2.4 Mathematical Models.............................................................................. 15 1.3.3 Interfacial Properties................................................................................................18 1.3.4 Droplet Charge..........................................................................................................18 1.3.5 Droplet Crystallinity................................................................................................ 19 1.3.6 Droplet Interactions................................................................................................. 20 1.4 Hierarchy of Emulsion Properties......................................................................................... 20 1.5 Understanding Food Emulsion Properties............................................................................. 21 1.5.1 Factors Influencing Topics and Directions of Research.......................................... 22 1.5.2 General Approaches Used to Study Food Emulsions.............................................. 23 1.6 Overview and Philosophy...................................................................................................... 25 References......................................................................................................................................... 26 2. Molecular Characteristics.............................................................................................................. 29 Introduction............................................................................................................................ 29 2.1 2.2 Forces of Nature..................................................................................................................... 30 2.3 Origin and Nature of Molecular Interactions........................................................................ 30 2.3.1 Covalent Interactions............................................................................................... 30 2.3.2 Electrostatic Interactions..........................................................................................31 2.3.3 Van der Waals Interactions...................................................................................... 34 2.3.4 Steric Overlap Interactions...................................................................................... 36 2.4 Overall Intermolecular Pair Potential.................................................................................... 38 2.4.1 Lennard–Jones Potential: Understanding Bond Strengths and Lengths................. 38 2.4.2 Thermal Energy: Judging Bond Strengths.............................................................. 38 2.4.3 Converting Potential Energies into Forces.............................................................. 39 vii viii Contents Molecular Structure and Organization Is Determined by a Balance of Interaction Energies and Entropy Effects.......................................................................... 39 2.5.1 Forms of Entropy..................................................................................................... 40 2.5.2 Physicochemical Basis of Molecular Transitions.....................................................41 2.6 Thermodynamics of Mixing.................................................................................................. 43 2.6.1 Potential Energy Change on Mixing....................................................................... 43 2.6.2 Entropy Change on Mixing..................................................................................... 44 2.6.3 Overall Free Energy Change on Mixing................................................................. 44 2.6.4 Complications.......................................................................................................... 46 2.7 Molecular Conformation....................................................................................................... 46 2.8 Compound Interactions.......................................................................................................... 48 2.8.1 Hydrogen Bonds....................................................................................................... 48 2.8.2 Hydrophobic Interactions......................................................................................... 49 2.9 Computer Modeling of Liquid Properties............................................................................. 49 2.9.1 Monte Carlo Techniques.......................................................................................... 50 2.9.2 Molecular Dynamics Techniques............................................................................ 50 2.10 Measurement of Molecular Characteristics............................................................................51 References......................................................................................................................................... 52 2.5 3. Colloidal Interactions..................................................................................................................... 55 Introduction............................................................................................................................ 55 3.1 3.2 Colloidal Interactions and Droplet Aggregation................................................................... 55 3.3 Van der Waals Interactions.................................................................................................... 58 3.3.1 Origin of van der Waals Interactions....................................................................... 58 3.3.2 Modeling van der Waals Interactions...................................................................... 58 3.3.2.1 Interdroplet Pair Potential......................................................................... 58 3.3.2.2 Hamaker Function..................................................................................... 59 3.3.3 General Features of van der Waals Interactions...................................................... 64 Electrostatic Interactions....................................................................................................... 64 3.4 3.4.1 Origins of Electrostatic Interactions........................................................................ 64 3.4.2 Modeling Electrostatic Interactions......................................................................... 64 3.4.2.1 Interdroplet Pair Potential......................................................................... 64 3.4.2.2 Factors Influencing Electrical Characteristics of Surfaces....................... 66 3.4.2.3 Influence of Ionic Strength on the Magnitude and Range of Interactions...... 68 3.4.2.4 Influence of Ion Bridging on Electrostatic Interactions............................ 69 3.4.3 General Characteristics of Electrostatic Interactions.............................................. 69 3.5 Steric Interactions.................................................................................................................. 70 3.5.1 Origin of Steric Interactions.................................................................................... 70 3.5.2 Modeling Steric Interactions.................................................................................... 71 3.5.2.1 Interdroplet Pair Potential......................................................................... 71 3.5.2.2 Mixing Contribution.................................................................................. 71 3.5.2.3 Elastic Contribution................................................................................... 73 3.5.2.4 Distance Dependence of Steric Interactions............................................. 74 3.5.2.5 Optimum Characteristics of Steric Stabilizers.......................................... 74 3.5.3 General Characteristics of Steric Interactions......................................................... 75 3.6 Depletion Interactions............................................................................................................ 76 3.6.1 Origin of Depletion Interactions.............................................................................. 76 3.6.2 Modeling of Depletion Interactions......................................................................... 77 3.6.3 General Characteristics of Depletion Interactions................................................... 80 3.7 Hydrophobic Interactions....................................................................................................... 80 3.7.1 Origin of Hydrophobic Interactions......................................................................... 80 3.7.2 Modeling Hydrophobic Interactions........................................................................ 81 3.7.3 General Characteristics of Hydrophobic Interactions............................................. 81 Contents ix Hydration Interactions........................................................................................................... 82 3.8.1 Origin of Hydration Interactions............................................................................. 82 3.8.2 Modeling Hydration Interactions............................................................................. 83 3.8.3 General Characteristics of Hydration Interactions.................................................. 84 3.9 Thermal Fluctuation Interactions.......................................................................................... 85 3.9.1 Origin of Thermal Fluctuation Interactions............................................................ 85 3.9.2 Modeling Thermal Fluctuation Interactions............................................................ 85 3.9.3 General Characteristics of Fluctuation Interactions................................................ 86 3.10 Nonequilibrium Effects......................................................................................................... 86 3.10.1 Molecular Rearrangements at the Interface............................................................. 86 3.10.2 Hydrodynamic Flow of Continuous Phase.............................................................. 86 3.10.3 Gibbs–Marangoni Effect......................................................................................... 87 3.11 Total Interaction Potential..................................................................................................... 88 3.11.1 Van der Waals and Steric Interactions..................................................................... 88 3.11.2 Van der Waals, Steric, and Electrostatic Interactions.............................................. 90 3.11.3 Van der Waals, Steric, Electrostatic, and Hydrophobic Interactions....................... 92 3.11.4 Van der Waals, Steric, Electrostatic, and Depletion Interactions............................ 93 3.12 Measurement of Colloidal Interactions................................................................................. 94 3.13 Prediction of Colloidal Interactions in Food Emulsions....................................................... 94 References......................................................................................................................................... 95 3.8 4. Emulsion Ingredients...................................................................................................................... 99 4.1 Introduction............................................................................................................................ 99 4.2 Fats and Oils........................................................................................................................ 100 4.2.1 Molecular Structure and Organization...................................................................101 4.2.2 Bulk Physicochemical Properties.......................................................................... 102 4.2.3 Fat Crystallization.................................................................................................. 104 4.2.3.1 Supercooling.......................................................................................... 105 4.2.3.2 Nucleation.............................................................................................. 106 4.2.3.3 Crystal Growth...................................................................................... 109 4.2.3.4 Crystal Morphology...............................................................................110 4.2.3.5 Polymorphism.........................................................................................110 4.2.3.6 Crystallization of Edible Fats and Oils..................................................111 4.2.3.7 Fat Crystallization in Emulsions............................................................112 4.2.4 Chemical Changes..................................................................................................113 4.2.5 Selection of an Appropriate Lipid...........................................................................114 4.2.5.1 Nutritional Profile...................................................................................114 4.2.5.2 Flavor Profile..........................................................................................114 4.2.5.3 Crystallization Behavior.........................................................................114 4.2.5.4 Oxidative Stability..................................................................................115 4.2.5.5 Bulk Physicochemical Properties...........................................................115 4.2.5.6 Oil Quality..............................................................................................115 4.3 Water.....................................................................................................................................115 4.3.1 Molecular Structure and Organization...................................................................116 4.3.2 Bulk Physicochemical Properties...........................................................................117 4.3.3 Influence of Solutes on the Organization of Water Molecules...............................117 4.3.3.1 Interaction of Water with Ionic Solutes..................................................118 4.3.3.2 Interaction of Water with Polar Solutes..................................................121 4.3.3.3 Interaction of Water with Nonpolar Solutes: The Hydrophobic Effect............................................................................... 122 4.3.4 Influence of Solutes on the Physicochemical Properties of Solutions................... 124 4.3.5 Selection of an Appropriate Aqueous Phase.......................................................... 124 x Contents Emulsifiers........................................................................................................................... 125 4.4.1 Surfactants............................................................................................................. 125 4.4.1.1 Molecular Characteristics...................................................................... 125 4.4.1.2 Physicochemical Properties................................................................... 127 4.4.1.3 Surfactant Classification Schemes..........................................................133 4.4.1.4 Common Food-Grade Surfactants........................................................ 140 4.4.2 Amphiphilic Biopolymers.......................................................................................142 4.4.2.1 Molecular Characteristics.......................................................................142 4.4.2.2 Interfacial Activity and Emulsion Stabilization.....................................143 4.4.2.3 Biopolymer-Based Food Emulsifiers......................................................145 4.4.2.4 Protein–Polysaccharide Complexes.......................................................149 4.4.3 Selection of an Appropriate Emulsifier..................................................................149 4.5 Texture Modifiers..................................................................................................................151 4.5.1 Thickening Agents..................................................................................................151 4.5.1.1 Effective Volume of Biopolymers in Aqueous Solutions.......................151 4.5.1.2 Relationship between Biopolymer Molecular Structure and Effective Volume in Solution..................................................................152 4.5.1.3 Viscosity Enhancement by Biopolymers in Solution.............................153 4.5.1.4 Shear-Thinning in Biopolymer Solutions.............................................. 154 4.5.2 Gelling Agents....................................................................................................... 156 4.5.3 Commonly Used Texture Modifiers...................................................................... 160 4.5.3.1 Polysaccharides......................................................................................161 4.5.3.2 Proteins.................................................................................................. 166 4.5.3.3 Biopolymer Blends.................................................................................167 4.5.4 Selection of an Appropriate Texture Modifier........................................................169 4.6 Other Food Additives............................................................................................................170 4.6.1 pH Control...............................................................................................................170 4.6.2 Minerals..................................................................................................................170 4.6.3 Chelating Agents.....................................................................................................171 4.6.4 Antioxidants............................................................................................................171 4.6.5 Antimicrobial Agents..............................................................................................172 4.6.6 Flavors.....................................................................................................................173 4.6.7 Colorants.................................................................................................................173 4.6.8 Weighting Agents....................................................................................................174 4.6.9 Fat Replacers...........................................................................................................174 4.7 Factors Influencing Ingredient Selection..............................................................................175 References........................................................................................................................................176 4.4 5. Interfacial Properties and Their Characterization....................................................................185 5.1 Introduction...........................................................................................................................185 5.2 General Characteristics of Interfaces...................................................................................186 5.2.1 Interfaces Separating Two Pure Liquids.................................................................186 5.2.2 Interfaces in the Presence of Solutes......................................................................188 5.3 Adsorption of Solutes to Interfaces..................................................................................... 192 5.3.1 Definition of Surface Excess Concentration.......................................................... 192 5.3.1.1 Gas–Liquid Interface in the Absence of Solutes................................... 192 5.3.1.2 Gas–Liquid Interface in the Presence of Solutes.................................. 193 5.3.1.3 Liquid–Liquid Interfaces....................................................................... 193 5.3.2 Relationship between Adsorbed and Bulk Solute Concentrations........................ 194 5.3.3 Stipulating Interfacial Properties of Surface-Active Solutes................................. 196 5.3.4 Adsorption Kinetics............................................................................................... 197 5.3.4.1 Movement of Molecules to the Vicinity of an Interface....................... 197 5.3.4.2 Attachment of Emulsifier Molecules to Interface................................. 199 Contents xi Electrical Characteristics of Interfaces............................................................................... 200 5.4.1 Origin of Interfacial Charge.................................................................................. 200 5.4.2 Ion Distribution near a Charged Interface............................................................. 202 5.4.2.1 Inner Region.......................................................................................... 205 5.4.2.2 Outer Region.......................................................................................... 207 5.4.3 Factors Influencing Interfacial Electrical Properties of Emulsions...................... 207 5.4.4 Characterization of Interfacial Electrical Properties............................................. 208 5.5 Interfacial Composition and Its Characterization............................................................... 208 5.5.1 Factors Influencing Interfacial Composition......................................................... 208 5.5.2 Characterization of Interfacial Composition in Emulsions....................................211 5.6 Interfacial Structure..............................................................................................................212 5.6.1 Factors Influencing Interfacial Structure................................................................212 5.6.2 Characterization of Interfacial Structure in Emulsions..........................................216 5.6.2.1 Microscopy Techniques..........................................................................216 5.6.2.2 Spectroscopy Techniques.......................................................................217 5.6.2.3 Interference Reflection Techniques........................................................218 5.6.2.4 Scattering Techniques............................................................................219 5.6.2.5 Langmuir Trough Measurements...........................................................219 5.6.2.6 Surface Force Measurements................................................................ 221 5.6.2.7 Calorimetry Techniques........................................................................ 221 5.6.2.8 Biochemical Techniques........................................................................ 221 5.7 Interfacial Tension and Its Measurement............................................................................. 222 5.7.1 Factors Influencing Interfacial Tension................................................................. 222 5.7.2 Characterization of Interfacial Tension................................................................. 222 5.7.2.1 Du Nouy Ring Method.......................................................................... 222 5.7.2.2 Wilhelmy Plate Method......................................................................... 224 5.7.2.3 Sessile and Pendant Drop Methods....................................................... 225 5.7.2.4 Drop Volume Method............................................................................ 226 5.7.2.5 Spinning Drop Method.......................................................................... 227 5.8 Interfacial Rheology and Its Measurement......................................................................... 229 5.8.1 Factors Influencing Interfacial Rheology.............................................................. 229 5.8.2 Characterization of Interfacial Rheology.............................................................. 230 5.8.2.1 Measurement of Interfacial Shear Rheology........................................ 230 5.8.2.2 Measurement of Interfacial Dilational Rheology...................................231 5.9 Chemical and Biochemical Properties of Interfaces........................................................... 232 5.10 Practical Implications of Interfacial Phenomena................................................................ 233 5.10.1 Properties of Curved Interfaces............................................................................. 233 5.10.2 Contact Angles and Wetting.................................................................................. 234 5.10.3 Capillary Rise and Meniscus Formation............................................................... 236 5.10.4 Interfacial Phenomenon in Food Emulsions.......................................................... 238 References....................................................................................................................................... 238 5.4 6. Emulsion Formation..................................................................................................................... 245 6.1 Introduction.......................................................................................................................... 245 Overview of Emulsion Formation....................................................................................... 245 6.2 6.3 Flow Profiles in Homogenizers........................................................................................... 248 6.4 Physical Principles of Emulsion Formation..........................................................................251 6.4.1 Droplet Disruption..................................................................................................251 6.4.1.1 Interfacial Forces................................................................................... 252 6.4.1.2 Disruptive Forces................................................................................... 252 6.4.1.3 Role of the Emulsifier in Droplet Disruption........................................ 257 6.4.1.4 Role of Nonideal Fluid Behavior on Droplet Disruption...................... 257 xii Contents 6.4.2 Droplet Coalescence.............................................................................................. 258 6.4.3 Role of the Emulsifier............................................................................................ 259 6.5 Homogenization Devices..................................................................................................... 260 6.5.1 High Shear Mixers................................................................................................. 260 6.5.2 Colloid Mills...........................................................................................................261 6.5.3 High-Pressure Valve Homogenizers...................................................................... 262 6.5.4 Microfluidization................................................................................................... 265 6.5.5 Ultrasonic Homogenizers...................................................................................... 266 6.5.6 Membrane and Microchannel Homogenizers........................................................ 267 6.5.7 Homogenization Efficiency.................................................................................... 269 6.5.8 Comparison of Homogenizers............................................................................... 270 6.6 Factors Influencing Droplet Size......................................................................................... 272 6.6.1 Emulsifier Type and Concentration....................................................................... 272 6.6.2 Energy Input............................................................................................................274 6.6.3 Properties of Component Phases........................................................................... 276 6.6.4 Temperature........................................................................................................... 277 6.6.5 Predicting Droplet Sizes Produced by Homogenization....................................... 278 6.7 Low-Energy Homogenization Methods............................................................................... 278 6.7.1 Spontaneous Emulsification................................................................................... 279 6.7.2 Emulsion Inversion Point Methods........................................................................ 279 6.7.3 Phase Inversion Temperature Methods.................................................................. 279 6.7.4 Comparison with High-Energy Methods............................................................... 280 6.8 Demulsification.................................................................................................................... 280 6.8.1 Nonionic Surfactants............................................................................................. 281 6.8.2 Ionic Surfactants.................................................................................................... 282 6.8.3 Biopolymer Emulsifiers......................................................................................... 282 6.8.4 General Methods of Demulsification..................................................................... 283 6.8.5 Selection of the Most Appropriate Demulsification Technique............................. 283 6.9 Future Developments........................................................................................................... 283 References....................................................................................................................................... 284 7. Emulsion Stability......................................................................................................................... 289 7.1 Introduction.......................................................................................................................... 289 7.2 Thermodynamic and Kinetic Stability of Emulsions.......................................................... 290 7.2.1 Thermodynamic Stability...................................................................................... 290 7.2.2 Kinetic Stability..................................................................................................... 292 7.3 Gravitational Separation...................................................................................................... 293 7.3.1 Physical Basis of Gravitational Separation............................................................ 294 7.3.1.1 Stokes’ Law........................................................................................... 294 7.3.1.2 Deviations from Stokes’ Law................................................................ 295 7.3.2 Methods of Controlling Gravitational Separation................................................. 304 7.3.2.1 Minimizing Density Difference............................................................ 304 7.3.2.2 Reducing Droplet Size........................................................................... 306 7.3.2.3 Modifying Continuous Phase Rheology............................................... 306 7.3.2.4 Increasing Droplet Concentration......................................................... 306 7.3.2.5 Altering the Degree of Droplet Flocculation........................................ 306 7.3.3 Experimental Characterization of Gravitational Separation................................. 307 7.4 Droplet Aggregation: General Features................................................................................310 7.4.1 Droplet–Droplet Encounters...................................................................................310 7.4.2 Film Thinning.........................................................................................................310 7.4.3 Thin Film Formation..............................................................................................311 7.4.4 Film Rupture...........................................................................................................312 Contents 7.5 7.6 7.7 7.8 xiii Flocculation..........................................................................................................................312 7.5.1 Physical Basis of Flocculation................................................................................312 7.5.1.1 Collision Frequency................................................................................313 7.5.1.2 Collision Efficiency................................................................................317 7.5.1.3 Overall Particle Growth Rate.................................................................317 7.5.2 Methods of Controlling Flocculation.....................................................................318 7.5.2.1 Collision Frequency................................................................................318 7.5.2.2 Collision Efficiency................................................................................319 7.5.3 Structure and Properties of Flocculated Emulsions.............................................. 328 7.5.3.1 Influence of Colloidal Interactions on Floc Structure........................... 328 7.5.3.2 Use of Fractal Geometry to Describe Floc Structure........................... 329 7.5.3.3 Influence of Floc Structure on Emulsion Properties............................. 330 7.5.4 Experimental Measurement of Flocculation..........................................................331 7.5.4.1 Microscopy Methods..............................................................................331 7.5.4.2 Particle Sizing Methods..........................................................................332 7.5.4.3 Bulk Physicochemical Properties...........................................................333 Coalescence......................................................................................................................... 334 7.6.1 Physical Basis of Coalescence............................................................................... 334 7.6.1.1 Physical and Molecular Processes Associated with Coalescence..........335 7.6.1.2 Mechanisms of Film Rupture................................................................ 336 7.6.1.3 Hole Formation...................................................................................... 338 7.6.1.4 Rate-Limiting Step for Coalescence..................................................... 339 7.6.1.5 Modeling Droplet Growth due to Coalescence......................................341 7.6.2 Methods of Controlling Coalescence..................................................................... 342 7.6.2.1 Prevention of Droplet Contact............................................................... 342 7.6.2.2 Prevention of Rupture of Interfacial Layers.......................................... 343 7.6.3 Factors Affecting Coalescence.............................................................................. 343 7.6.3.1 Emulsifier Type...................................................................................... 343 7.6.3.2 Influence of Environmental Conditions................................................ 344 7.6.3.3 Influence of Impurities and Surfaces.................................................... 346 7.6.4 Measurement of Droplet Coalescence................................................................... 346 7.6.4.1 Microscopy Methods............................................................................. 346 7.6.4.2 Particle Sizing Methods......................................................................... 348 7.6.4.3 Oiling Off Tests..................................................................................... 348 7.6.4.4 Accelerated Test Methods..................................................................... 349 Partial Coalescence...............................................................................................................351 7.7.1 Physical Basis of Partial Coalescence....................................................................352 7.7.2 Methods of Controlling Partial Coalescence..........................................................355 7.7.2.1 Prevention of Close Contact...................................................................355 7.7.2.2 Prevention of Interfacial Layer Disruption............................................ 356 7.7.2.3 Control of Crystal Concentration, Structure, and Location.................. 356 7.7.3 Experimental Characterization of Partial Coalescence......................................... 356 7.7.3.1 Fat Crystal Properties............................................................................ 357 7.7.3.2 Emulsion Microstructure....................................................................... 357 7.7.3.3 Macroscopic Properties......................................................................... 358 Ostwald Ripening................................................................................................................ 358 7.8.1 Physical Basis of Ostwald Ripening...................................................................... 358 7.8.2 Methods of Controlling Ostwald Ripening............................................................361 7.8.2.1 Droplet Size Distribution........................................................................361 7.8.2.2 Solubility................................................................................................361 7.8.2.3 Interfacial Layer.................................................................................... 362 7.8.2.4 Droplet Composition............................................................................. 362 7.8.3 Experimental Characterization of Ostwald Ripening........................................... 365 xiv Contents Phase Inversion.................................................................................................................... 365 7.9.1 Physical Basis of Phase Inversion.......................................................................... 365 7.9.1.1 Surfactant-Induced Phase Inversion...................................................... 366 7.9.1.2 Fat Crystallization–Induced Phase Inversion........................................ 367 7.9.2 Methods of Controlling Phase Inversion............................................................... 368 7.9.2.1 Disperse Phase Volume Fraction........................................................... 368 7.9.2.2 Emulsifier Type and Concentration....................................................... 368 7.9.2.3 Mechanical Agitation............................................................................ 369 7.9.2.4 Temperature........................................................................................... 369 7.9.3 Characterization of Phase Inversion...................................................................... 369 7.9.3.1 Electrical Conductivity.......................................................................... 370 7.9.3.2 Rheology................................................................................................ 370 7.9.3.3 Optical Properties.................................................................................. 370 7.9.3.4 Microscopy.............................................................................................371 7.9.3.5 Droplet Size Analysis.............................................................................371 7.9.3.6 Interfacial Tension..................................................................................371 7.9.3.7 Coalescence Stability.............................................................................371 7.9.3.8 Emulsion Miscibility..............................................................................371 7.10 Chemical and Biochemical Stability................................................................................... 372 7.10.1 Lipid Oxidation...................................................................................................... 372 7.10.2 Enzyme Hydrolysis................................................................................................ 372 7.10.3 Flavor and Color Degradation............................................................................... 372 References....................................................................................................................................... 373 7.9 8. Emulsion Rheology....................................................................................................................... 383 8.1 Introduction.......................................................................................................................... 383 8.2 Rheological Properties of Materials.................................................................................... 384 8.2.1 Solids...................................................................................................................... 384 8.2.1.1 Ideal Elastic Solids................................................................................ 384 8.2.1.2 Nonideal Elastic Solids.......................................................................... 386 8.2.2 Liquids................................................................................................................... 386 8.2.2.1 Ideal Liquids.......................................................................................... 386 8.2.2.2 Nonideal Liquids................................................................................... 388 8.2.3 Plastics................................................................................................................... 392 8.2.3.1 Ideal Plastics.......................................................................................... 393 8.2.3.2 Nonideal Plastics................................................................................... 393 8.2.4 Viscoelastic Materials............................................................................................ 394 8.2.4.1 Transient Tests....................................................................................... 394 8.2.4.2 Dynamic Tests....................................................................................... 396 8.3 Measurement of Rheological Properties............................................................................. 397 8.3.1 Simple Compression and Elongation..................................................................... 398 8.3.2 Shear Measurements.............................................................................................. 400 8.3.2.1 Capillary Viscometers........................................................................... 400 8.3.2.2 Mechanical Viscometers and Dynamic Rheometers............................ 401 8.3.2.3 Possible Sources of Experimental Error............................................... 403 8.3.3 Advanced Rheological Methods............................................................................ 404 8.3.3.1 Rheometers Combined with Other Analytical Methods....................... 404 8.3.3.2 Rheometers Utilizing Complex Deformation Profiles.......................... 405 8.3.3.3 Thin Film Rheology (Tribology)........................................................... 405 8.3.3.4 Microrheology Methods........................................................................ 407 8.3.3.5 Interfacial Rheology Methods............................................................... 408 8.3.4 Empirical Techniques............................................................................................ 408 Contents xv Rheological Properties of Emulsions.................................................................................. 408 8.4.1 Dilute Suspensions of Rigid Spherical Particles................................................... 409 8.4.2 Dilute Suspensions of Fluid Spherical Particles.....................................................411 8.4.3 Dilute Suspensions of Rigid Nonspherical Particles..............................................411 8.4.4 Dilute Suspensions of Flocculated Particles...........................................................413 8.4.5 Concentrated Suspensions in the Absence of Long-Range Colloidal Interactions......415 8.4.6 Concentrated Suspensions with Repulsive Interactions.........................................417 8.4.7 Concentrated Suspensions with Attractive Interactions: Flocculated Systems...... 420 8.4.8 Emulsions with Semisolid Continuous Phases...................................................... 425 8.5 Computer Simulation of Emulsion Rheology...................................................................... 425 8.6 Major Factors Influencing Emulsion Rheology................................................................... 427 8.6.1 Disperse-Phase Volume Fraction........................................................................... 427 8.6.2 Rheology of Component Phases............................................................................ 428 8.6.3 Particle Size and Polydispersity............................................................................. 428 8.6.4 Colloidal Interactions............................................................................................. 429 8.6.5 Droplet Charge....................................................................................................... 430 8.7 Concluding Remarks and Future Directions........................................................................431 References........................................................................................................................................432 8.4 9. Emulsion Flavor............................................................................................................................ 437 Introduction.......................................................................................................................... 437 9.1 9.2 Flavor Partitioning............................................................................................................... 438 9.2.1 Partitioning between a Homogeneous Liquid and a Vapor................................... 439 9.2.2 Influence of Flavor Ionization................................................................................ 441 9.2.3 Influence of Flavor Binding on Partitioning.......................................................... 442 9.2.4 Influence of Surfactant Micelles on Partitioning................................................... 444 9.2.5 Partitioning in Emulsions in the Absence of an Interfacial Layer........................ 445 9.2.6 Partitioning in Emulsions in the Presence of an Interfacial Layer........................ 447 Flavor Release...................................................................................................................... 449 9.3 9.3.1 Overview of the Physicochemical Process of Flavor Release............................... 449 9.3.2 Release of Nonvolatile Compounds (Taste)........................................................... 449 9.3.2.1 Maximum Amount of Flavor Released................................................. 450 9.3.2.2 Kinetics of Flavor Release..................................................................... 450 9.3.3 Release of Volatile Compounds (Aroma).............................................................. 454 9.3.3.1 Flavor Release from Homogeneous Liquids......................................... 454 9.3.3.2 Influence of Ingredient Interactions...................................................... 458 9.3.3.3 Flavor Release from Emulsions............................................................. 459 9.4 Emulsion Mouthfeel and Oral Processing........................................................................... 462 9.4.1 Colloidal Aspects................................................................................................... 462 9.4.2 Rheological Aspects.............................................................................................. 463 9.4.3 Lubrication Aspects............................................................................................... 463 9.4.4 Coating Aspects..................................................................................................... 463 9.4.5 Thermal Aspects.................................................................................................... 464 9.5 Measurement of Emulsion Flavor........................................................................................ 464 9.5.1 Analysis of Volatile Flavor Compounds................................................................ 464 9.5.2 Analysis of Nonvolatile Flavor Compounds.......................................................... 466 9.5.3 Analysis of Oral Processing................................................................................... 467 9.5.3.1 Large-Deformation Rheology............................................................... 468 9.5.3.2 Small-Deformation Rheology............................................................... 468 9.5.3.3 Tribology............................................................................................... 468 9.5.3.4 Extensional/Elongational Flow Rheology............................................. 469 9.5.3.5 Miscellaneous Tests............................................................................... 469 9.5.4 Sensory Analysis.................................................................................................... 470 xvi Contents Overview of the Factors Influencing Emulsion Flavor........................................................ 473 9.6.1 Disperse-Phase Volume Fraction........................................................................... 473 9.6.2 Droplet Size.............................................................................................................475 9.6.3 Interfacial Characteristics...................................................................................... 477 9.6.4 Oil Phase Characteristics........................................................................................478 9.6.5 Aqueous Phase Characteristics...............................................................................478 9.7 Concluding Remarks and Future Directions....................................................................... 479 References....................................................................................................................................... 480 9.6 10. Appearance.................................................................................................................................... 489 10.1 Introduction.......................................................................................................................... 489 10.2 General Aspects of Optical Properties of Materials........................................................... 490 10.2.1 Interaction of Light with Matter............................................................................ 490 10.2.2 Human Vision........................................................................................................ 495 10.2.3 Quantitative Description of Appearance............................................................... 495 10.3 Mathematical Modeling of Emulsion Color........................................................................ 496 10.3.1 Calculation of Scattering Characteristics of Emulsion Droplets........................... 498 10.3.2 Calculation of Spectral Transmittance or Reflectance of Emulsions.................... 500 10.3.3 Relationship of Tristimulus Coordinates to Spectral Reflectance and Transmittance.................................................................................................. 503 10.3.4 Influence of Polydispersity..................................................................................... 504 10.3.5 Numerical Calculations of Emulsion Color........................................................... 505 10.3.6 Influence of Measurement Cell.............................................................................. 509 10.4 Measurement of Emulsion Color..........................................................................................510 10.4.1 Spectrophotometric Colorimeters...........................................................................511 10.4.2 Trichromatic Colorimeters......................................................................................513 10.4.3 Light Scattering.......................................................................................................514 10.4.4 Sensory Analysis.....................................................................................................514 10.5 Major Factors Influencing Emulsion Color..........................................................................514 10.5.1 Droplet Concentration and Size..............................................................................514 10.5.2 Relative Refractive Index of Droplets.....................................................................517 10.5.3 Colorant Type and Concentration...........................................................................519 10.5.4 Factors Affecting Color of Real Food Emulsions.................................................. 520 10.6 Concluding Remarks and Future Directions....................................................................... 520 References........................................................................................................................................521 11. Gastrointestinal Fate of Emulsions............................................................................................. 523 11.1 Introduction.......................................................................................................................... 523 11.2 Overview of Emulsion Passage through the GIT................................................................ 523 11.2.1 Mouth..................................................................................................................... 523 11.2.2 Stomach.................................................................................................................. 524 11.2.3 Small Intestine....................................................................................................... 526 11.2.4 Colon...................................................................................................................... 527 11.2.5 Hormonal and Neurological Responses................................................................. 528 11.3 Potential Changes in Emulsion Characteristics................................................................... 529 11.3.1 Droplet Composition.............................................................................................. 529 11.3.2 Particle Size........................................................................................................... 530 11.3.3 Interfacial Properties..............................................................................................531 11.3.4 Physical State..........................................................................................................531 11.4 Reasons for Controlling Gastrointestinal Fate of Emulsions...............................................532 11.4.1 Development of Reduced Calorie Products............................................................532 11.4.2 Control of Hormonal Responses.............................................................................532 11.4.3 Delivery of Bioactive Components.........................................................................533 Contents xvii Characterization of Gastrointestinal Fate of Emulsions...................................................... 534 11.5.1 In Vitro Approaches............................................................................................... 534 11.5.1.1 Passage through GIT..............................................................................535 11.5.1.2 Absorption............................................................................................. 536 11.5.2 In Vivo Approaches.................................................................................................537 11.5.3 In Vitro versus In Vivo Correlations.......................................................................537 11.5.4 Measurement of Changes in Emulsion Properties................................................. 538 11.6 Conclusions and Future Directions...................................................................................... 540 References....................................................................................................................................... 540 11.5 12. Food Emulsions in Practice.......................................................................................................... 547 12.1 Introduction.......................................................................................................................... 547 12.2 Milk and Cream................................................................................................................... 547 12.2.1 Composition........................................................................................................... 547 12.2.1.1 Dispersed Phase..................................................................................... 547 12.2.1.2 Interfacial Layer.................................................................................... 548 12.2.1.3 Continuous Phase.................................................................................. 550 12.2.2 Microstructure....................................................................................................... 550 12.2.3 Production.............................................................................................................. 550 12.2.4 Physicochemical Properties....................................................................................551 12.2.4.1 Stability...................................................................................................551 12.2.4.2 Rheology.................................................................................................552 12.2.4.3 Appearance.............................................................................................553 12.2.4.4 Flavor......................................................................................................553 12.2.5 Dairy Products........................................................................................................553 12.2.5.1 Whipped Cream.....................................................................................553 12.2.5.2 Butter..................................................................................................... 554 12.2.5.3 Ice Cream............................................................................................... 554 12.2.5.4 Yogurt.................................................................................................... 556 12.2.5.5 Cheese.................................................................................................... 556 12.3 Beverage Emulsions..............................................................................................................557 12.3.1 Composition............................................................................................................557 12.3.1.1 Dispersed Phase......................................................................................557 12.3.1.2 Interfacial Layer.....................................................................................558 12.3.1.3 Continuous Phase.................................................................................. 560 12.3.2 Microstructure....................................................................................................... 560 12.3.3 Production...............................................................................................................561 12.3.3.1 Beverage Emulsion Concentrate.............................................................561 12.3.3.2 Finished Product.....................................................................................561 12.3.4 Physicochemical Properties....................................................................................561 12.3.4.1 Stability...................................................................................................561 12.3.4.2 Texture................................................................................................... 563 12.3.4.3 Flavor..................................................................................................... 563 12.3.4.4 Appearance............................................................................................ 563 12.4 Dressings.............................................................................................................................. 564 12.4.1 Composition........................................................................................................... 565 12.4.1.1 Dispersed Phase..................................................................................... 565 12.4.1.2 Continuous Phase.................................................................................. 566 12.4.1.3 Interfacial Layer.................................................................................... 567 12.4.2 Microstructure....................................................................................................... 567 12.4.3 Production.............................................................................................................. 568 12.4.4 Physicochemical Properties................................................................................... 569 12.4.4.1 Stability.................................................................................................. 569 xviii Contents 12.4.4.2 Rheology.................................................................................................571 12.4.4.3 Appearance............................................................................................ 572 12.4.4.4 Flavor..................................................................................................... 572 References....................................................................................................................................... 572 13. Emulsion-Based Delivery Systems............................................................................................... 577 13.1 Introduction.......................................................................................................................... 577 13.1.1 Active Ingredients and Their Need for Encapsulation.......................................... 577 13.1.2 Challenges to Incorporating Active Ingredients in Foods..................................... 578 13.1.3 Desirable Characteristics of Delivery Systems...................................................... 580 13.1.4 Delivery System Design......................................................................................... 582 13.2 Emulsions and Nanoemulsions............................................................................................ 582 13.2.1 Composition and Structure.................................................................................... 582 13.2.2 Formation............................................................................................................... 583 13.2.3 Properties............................................................................................................... 585 13.2.4 Applications........................................................................................................... 586 13.3 Multiple Emulsions.............................................................................................................. 587 13.3.1 Composition and Structure.................................................................................... 587 13.3.2 Formation............................................................................................................... 588 13.3.3 Properties............................................................................................................... 588 13.3.4 Applications............................................................................................................591 13.4 Multilayer Emulsions........................................................................................................... 592 13.4.1 Composition and Structure.................................................................................... 592 13.4.2 Formation............................................................................................................... 592 13.4.3 Properties............................................................................................................... 595 13.4.4 Applications........................................................................................................... 596 13.5 Solid Lipid Particles............................................................................................................. 598 13.5.1 Composition and Structure.................................................................................... 598 13.5.2 Formation............................................................................................................... 598 13.5.3 Properties............................................................................................................... 600 13.5.4 Applications........................................................................................................... 602 13.6 Filled Hydrogel Particles..................................................................................................... 602 13.6.1 Composition and Structure.................................................................................... 602 13.6.2 Formation............................................................................................................... 602 13.6.3 Properties............................................................................................................... 604 13.6.4 Applications........................................................................................................... 606 13.7 Microclusters....................................................................................................................... 607 13.7.1 Composition and Structure.................................................................................... 607 13.7.2 Formation............................................................................................................... 607 13.7.3 Properties............................................................................................................... 608 13.7.4 Applications........................................................................................................... 609 13.8 Miscellaneous Systems.........................................................................................................610 13.8.1 Particle-Stabilized Emulsions.................................................................................610 13.8.2 Emulsified Microemulsions and Cubosomes..........................................................610 13.8.3 Nanocrystal Suspensions........................................................................................611 13.9 Summary...............................................................................................................................611 References........................................................................................................................................611 14. Characterization of Emulsion Properties................................................................................... 623 14.1 Introduction.......................................................................................................................... 623 14.2 Testing Emulsifier Effectiveness.......................................................................................... 623 14.2.1 Emulsifying Capacity............................................................................................ 624 14.2.2 Emulsion Stability Index........................................................................................ 625 Contents xix 14.3 Microstructure and Droplet Size Distribution..................................................................... 627 14.3.1 Microscopy............................................................................................................ 627 14.3.1.1 Optical Microscopy............................................................................... 627 14.3.1.2 Laser Scanning Confocal Microscopy.................................................. 630 14.3.1.3 Electron Microscopy............................................................................. 630 14.3.1.4 Atomic Force Microscopy..................................................................... 634 14.3.2 Static Light Scattering........................................................................................... 636 14.3.2.1 Principles............................................................................................... 636 14.3.2.2 Measurement Techniques...................................................................... 638 14.3.2.3 Applications........................................................................................... 640 14.3.3 Dynamic Light Scattering and Diffusing Wave Spectroscopy.............................. 642 14.3.3.1 Principles............................................................................................... 642 14.3.3.2 Measurement Techniques...................................................................... 642 14.3.3.3 Applications........................................................................................... 646 14.3.4 Electrical Pulse Counting...................................................................................... 647 14.3.5 Sedimentation Techniques..................................................................................... 648 14.3.5.1 Principles............................................................................................... 648 14.3.5.2 Measurement Techniques...................................................................... 648 14.3.5.3 Applications........................................................................................... 649 14.3.6 Ultrasonic Spectrometry........................................................................................ 649 14.3.6.1 Principles............................................................................................... 649 14.3.6.2 Measurement Techniques.......................................................................651 14.3.6.3 Applications........................................................................................... 652 14.3.7 Nuclear Magnetic Resonance................................................................................ 652 14.3.8 Neutron Scattering................................................................................................. 653 14.3.9 Alternative Methods.............................................................................................. 654 Disperse Phase Volume Fraction......................................................................................... 655 14.4.1 Proximate Analysis................................................................................................ 655 14.4.2 Density Measurements........................................................................................... 655 14.4.2.1 Principles............................................................................................... 655 14.4.2.2 Measurement Techniques...................................................................... 656 14.4.2.3 Applications........................................................................................... 657 14.4.3 Electrical Conductivity.......................................................................................... 657 14.4.3.1 Principles............................................................................................... 657 14.4.3.2 Measurement Techniques...................................................................... 657 14.4.3.3 Applications........................................................................................... 658 14.4.4 Alternative Methods.............................................................................................. 658 Droplet Crystallinity............................................................................................................ 658 14.5.1 Dilatometry............................................................................................................ 658 14.5.1.1 Principles............................................................................................... 658 14.5.1.2 Measurement Techniques...................................................................... 659 14.5.1.3 Applications........................................................................................... 659 14.5.2 Nuclear Magnetic Resonance................................................................................ 659 14.5.2.1 Principles............................................................................................... 659 14.5.2.2 Measurement Techniques.......................................................................661 14.5.2.3 Applications........................................................................................... 662 14.5.3 Thermal Analysis................................................................................................... 662 14.5.3.1 Principles............................................................................................... 662 14.5.3.2 Measurement Techniques...................................................................... 663 14.5.3.3 Applications........................................................................................... 664 14.5.4 Ultrasonics............................................................................................................. 665 14.5.4.1 Principles............................................................................................... 665 14.4 14.5 xx Contents 14.5.4.2 Measurement Techniques...................................................................... 665 14.5.4.3 Applications........................................................................................... 666 14.6 Droplet Charge..................................................................................................................... 666 14.6.1 Particle Electrophoresis......................................................................................... 666 14.6.2 Electro-Acoustics................................................................................................... 668 14.7 Droplet Interactions............................................................................................................. 670 14.8 Summary...............................................................................................................................671 References....................................................................................................................................... 672 Index������������������������������������������������������������������������������������������������������������������������������������������������������ 677 Preface It has been over a decade since the second edition of Food Emulsions: Principles, Practice, and Techniques was published. During this time, there have been important advances in a number of traditional areas within this subject, as well as the emergence of some new areas that were not covered previously. The purpose of this new edition of the book is to update those parts of the subject that have seen these recent advances and to give an overview of the new areas that have emerged. For this reason, all of the previous chapters have been revised and updated, and the figures have been redrawn or supplemented where necessary. In addition, two new chapters have been added to this edition of the book to reflect two important areas that have been the focus of intense recent research efforts. First, a chapter on the gastrointestinal fate of food emulsions has been included to give an overview of the current understanding of this important topic (Chapter 11). Traditionally, most of the research on food emulsions focused on their behavior within the actual product, for example, their optical, rheological, and stability characteristics. However, it is now recognized that an understanding of the fate of emulsions within the human gastrointestinal tract can aid in the design of functional foods that may promote health and wellness. For example, by controlling the location where an emulsion is broken down within the gastrointestinal tract, it is possible to control the flavor release profile, the satiety response, and the bioavailability of nutraceuticals. This research area has been driven by the focus of the modern food industry on designing and marketing foods based on their potential health benefits. Second, a chapter on emulsion-based delivery systems is included to highlight the important advances that have been made in this area (Chapter 13). In the previous editions of this book, the main focus was on the conventional oil-in-water emulsions that form the basis of many common food products, such as sauces, dressings, beverages, desserts, dairy products, and dips. Recent work has focused on creating a range of structured emulsions that are specifically designed to have novel functional attributes, such as encapsulation, protection, controlled release, textural properties, or fat mimicking. Structured emulsions are colloidal materials that utilize emulsion droplets as building blocks and include nanoemulsions, multiple emulsions, multilayer emulsions, solid lipid nanoparticles, filled hydrogel particles, colloidosomes, and microclusters. The possible advantages and limitations of these structured emulsions for particular applications are highlighted, as well as methods of designing and fabricating them. As in previous editions, the main focus of this book is on presenting the fundamental principles that underlie all types of emulsion-based food products rather than on specific food products themselves. An improved understanding of these basic principles will aid in the design of new products, the improvement of existing products, and the rapid solution of processing problems. I thank all of the students, postdoctoral researchers, and visiting scientists who have worked with me throughout my career. Their hard work, creativity, and dedication have played a major role in the development of my understanding of the subject of food emulsions. I also thank all of my colleagues at the University of Massachusetts and other academic and industrial institutions. My collaborations with these individuals have allowed me to apply the basic principles of emulsion science to a wide range of problems that it would have been difficult for me to tackle on my own, such as improving the chemical stability of nutrients, developing effective antimicrobial delivery systems, designing and testing delivery systems for bioactive molecules, and creating reduced-calorie foods with desirable sensory attributes. I also thank all of my teachers and advisors at the University of Leeds, the University of California, and the University College Cork for giving me a strong foundation in food colloids and biopolymers. In particular, I thank Professors Eric Dickinson and Malcolm Povey, who were such strong role models at the beginning of my career. I also recognize all the support of the excellent staff at CRC Press for bringing this book into existence. Finally, I acknowledge the continuous support, encouragement, and understanding of my family in England, and my wife, Jayne, and daughter, Isobelle. xxi Author David Julian McClements is a professor at the Department of Food Science at the University of Massachusetts, Amherst, Massachusetts. He specializes in the areas of food biopolymers and colloids, and in particular on the development of food-based structured delivery systems for bioactive components. Dr. McClements received his PhD in food science (1989) from the University of Leeds (United Kingdom). He then did postdoctoral research at the University of Leeds, University of California, Davis (California), and University College Cork (Ireland). Dr. McClements is the sole author of Nanoparticle- and Microparticle-based Delivery Systems: Encapsulation, Protection and Release of Active Components, three editions of Food Emulsions: Principles, Practice, and Techniques, coauthor of Advances in Food Colloids with Prof. Eric Dickinson, and coeditor of Developments in Acoustics and Ultrasonics, Understanding and Controlling the Microstructure of Complex Foods, Designing Functional Foods, Oxidation in Foods and Beverages (Volumes 1 and 2), and Encapsulation and Delivery Systems for Food Ingredients and Nutraceuticals. In addition, he has published more than 550 scientific articles in peer-reviewed journals (with an H-index of over 72). Dr. McClements has previously received awards from the American Chemical Society, American Oil Chemists Society, Society of Chemical Industry (United Kingdom), Institute of Food Technologists, and University of Massachusetts in recognition of his scientific achievements. He is also a fellow of the Royal Society of Chemistry (United Kingdom), the American Chemical Society, and the Institute of Food Technologists. His research has been funded by grants from the U.S. Department of Agriculture, National Science Foundation, NASA, U.S. Department of Commerce, Dairy Management Incorporated, and the food industry. He is a member of the editorial boards of a number of journals and has organized workshops, symposia, and conferences in the field of food colloids, food emulsions, and delivery systems. xxiii 1 Context and Background 1.1 Emulsion Science and Technology in the Food Industry Knowledge of the science and technology of emulsions is important for those working in the food and related industries for a number of reasons. First, many natural and processed foods consist either partly or wholly as emulsions, or have been in an emulsified state sometime during their production, including milk, cream, beverages, infant formula, soups, cake batters, salad dressings, mayonnaise, cream liqueurs, sauces, deserts, dips, salad cream, ice cream, coffee whitener, spreads, butter, and margarine. Second, emulsions are increasingly being utilized as delivery systems for functional food ingredients, such as colors, flavors, preservatives, vitamins, and nutraceuticals (Velikov and Pelan 2008, McClements and Li 2010). Emulsion-based delivery systems are usually designed to encapsulate, protect, and release these functional ingredients so as to improve their handling, stability, or efficacy. Emulsion-based food products and delivery systems exhibit a wide range of physicochemical, sensory, and biological characteristics depending on the kinds of ingredients and processing conditions used to create them. Despite this diversity, there are a number of underlying features that are common to this group of products that makes them amenable to study by the scientific discipline known as emulsion science, which combines aspects of physics, chemistry, biology, and engineering. Traditionally, the fundamental principles of emulsion science were largely derived from the disciplines of colloid science, interfacial chemistry, polymer science, and fluid mechanics. Nevertheless, as emulsion science has evolved within the food industry, it has incorporated a range of other scientific disciplines, such as sensory science and human physiology, as researchers attempt to correlate organoleptic qualities (such as taste, odor, mouthfeel, and appearance) and biological responses (such as digestion, absorption, and hormone release) to emulsion composition, structure, and physicochemical properties. A particularly notable aspect of modern emulsion research in the food industry is the integration of knowledge from disparate scientific disciplines. The manufacture of an emulsion-based product with specific functional attributes depends on the selection of the most suitable types and concentrations of raw materials (e.g., water, oil, emulsifiers, thickening agents, minerals, acids, bases, vitamins, flavors, colorants, and preservatives) and the most appropriate processing, storage, transport, and usage conditions (e.g., mixing, homogenization, pasteurization, sterilization, chilling, freezing, and cooking). Traditionally, the food industry largely relied on craft and tradition for the formulation of food products and the establishment of optimum processing conditions. This approach is unsuitable for the modern food industry, which must rapidly respond to changes in consumer preferences for a greater variety of cheaper, higher-quality, healthier, more exotic, and more convenient foods. In addition, the modern food industry relies increasingly on large-scale production operations to produce large quantities of foods at relatively low cost. The design of new foods, the improvement of existing foods, and the efficient operation of food manufacturing processes require a rigorous scientific understanding of food properties. Two major factors that have contributed to the more rational design and fabrication of emulsion-based products with improved or novel properties are highlighted as follows: 1. Development of a more rigorous scientific approach to understanding food emulsion properties: There has been an increasing tendency within the food industry toward relating the bulk physicochemical, organoleptic, and nutritional properties of food emulsions to the type, 1 2 Food Emulsions: Principles, Practices, and Techniques concentration, structure, and interactions of their constituent components. Research in this area is carried out at many different hierarchical levels, ranging from the study of the structure and interactions of molecules and colloidal particles to the study of the rheology, stability, and optical properties of emulsions, to the study of the taste, smell, mouthfeel, and appearance of final products, to the study of the behavior of emulsions within the human body after ingestion. In particular, there is a growing emphasis on integrating information determined at different hierarchical levels, so as to obtain a more holistic understanding of the properties of the whole system. The improved understanding of the physicochemical basis of food emulsion properties that has resulted from this approach has enabled manufacturers to create low-cost high-quality food products in a more systematic and reliable fashion. 2. Development of new analytical techniques to characterize food properties: The boundaries of our understanding of the physicochemical basis of food emulsion properties are often determined by the availability of analytical techniques that are capable of investigating the appropriate characteristics of the system. As analytical instrumentation progresses, we are able to study things that were not possible earlier, which often results in a deeper and broader understanding of the subject. In recent years, many new and improved analytical techniques or experimental protocols for probing the molecular, interfacial, colloidal, physicochemical, and biological properties of emulsions have become available. The application of these techniques has led to considerable advances in basic research, product development, and quality control within the food industry. These analytical techniques and protocols are used in research laboratories to enhance the fundamental understanding of the physicochemical basis of emulsion properties. They are also used in factories to monitor the properties of foods during processing so as to ensure that they meet the required quality specifications and to provide information that can be used to optimize the processing conditions required to produce consistently high-quality products. As new analytical instrumentation continues to become available, there will certainly be further developments in the abilities of food scientists to understand, predict, and control the properties of emulsion-based food products. In addition, the study of food emulsions can provide an excellent paradigm for the study of more structurally complex food materials, since many of the concepts, theories, and techniques developed to model or probe emulsion properties can be applied (with some modification) to understanding these systems. Ultimately, the aim of the emulsion scientist working in the food industry is to utilize the basic principles and techniques of emulsion science to enhance the quality of the food supply and the efficiency of food production. This book presents the conceptual and theoretical framework required by food scientists to understand and control the properties of emulsion-based food products. 1.2 General Characteristics of Food Emulsions 1.2.1 Definitions An emulsion consists of two immiscible liquids (usually oil and water), with one of the liquids being dispersed as small spherical droplets in the other (Figure 1.1). In most foods, the diameters of the droplets usually lie somewhere between 100 nm and 100 μm, but there has been growing interest in the utilization of emulsions with smaller diameters (d <100 nm) recently due to their novel physicochemical properties. Emulsions can be conveniently classified according to the relative spatial distribution of the oil and aqueous phases. A system that consists of oil droplets dispersed in an aqueous phase is called an oil-in-water (O/W) emulsion, for example, milk, cream, dressings, mayonnaise, beverages, soups, and sauces. A system that consists of water droplets dispersed in an oil phase is called a water-in-oil (W/O) emulsion, for example, margarine and butter. The substance that makes up the droplets in an emulsion is referred to as the dispersed, discontinuous, or internal phase, whereas the substance that makes up the surrounding liquid is called the continuous or external phase. To be consistent, I will refer to the 3 Context and Background 50 µm 10 µm FIGURE 1.1 An example of a food O/W emulsion (salad dressing) consisting of oil droplets dispersed in an aqueous medium evaluated using differential interference contrast (DIC), a general contrast enhancement optical method that highlights differences in refractive indices in heterogeneous samples. (Courtesy of Kraft Foods, Chicago, IL.) droplets as the dispersed phase and the surrounding liquid as the continuous phase throughout this book. The concentration of droplets in an emulsion is usually described in terms of the dispersed-phase volume fraction, ϕ (Section 1.3.1). In addition to the conventional O/W or W/O emulsions described earlier, it is also possible to prepare various types of multiple emulsions, for example, oil-in-water-in-oil (O/W/O) or water-in-oil-in-water (W/O/W) emulsions. For example, a W/O/W emulsion consists of water droplets dispersed within larger oil droplets, which are themselves dispersed in an aqueous continuous phase. Recently, research has been carried out to create stable multiple emulsions that can be used to control the release of certain ingredients, reduce the total fat content of emulsion-based food products, or isolate one ingredient from another ingredient that it might normally interact with. Multiple emulsions are part of a growing group of structurally designed emulsions that are likely to find increasing utilization within the food industry because of their potential advantages over conventional emulsions (McClements and Li 2010, McClements 2012). As well as multiple emulsions, this group also includes nanoemulsions, multilayer emulsions, colloidosomes, filled hydrogel particles, and microclusters (Figure 1.2). Researchers are currently trying to develop structured emulsions that can be economically produced using food-grade ingredients and that have desirable quality attributes, functional performances, and shelf lives. A discussion of the fabrication, properties, and applications of these novel forms of emulsions will be given in a later chapter. The process of converting two separate immiscible liquids into an emulsion, or of reducing the size of the droplets in a preexisting emulsion, is known as homogenization. In the food industry, this process is usually carried out using high-energy methods that utilize mechanical devices (homogenizers) to form small droplets by subjecting the phases to disruptive forces, for example, high-speed blenders, highpressure valve homogenizers, and colloid mills (Chapter 6). However, under certain circumstances, it is also possible to form emulsions using low-energy methods that rely on the spontaneous formation of small droplets when two phases are mixed together (Chapter 6). It is possible to form an emulsion by homogenizing pure oil and pure water together, but the two phases usually rapidly separate into a system that consists of a layer of oil (lower density) on top of a 4 Food Emulsions: Principles, Practices, and Techniques (a) (e) (b) (f) (c) (g) (d) (h) FIGURE 1.2 Examples of structurally designed emulsion-based systems that can be used in foods (not drawn to scale): (a) emulsions, (b) multilayer emulsions, (c) filled solid particles, (d) microclusters, (e) multiple emulsions, (f) colloidosomes, (g) filled hydrogel particles, and (h) nanoemulsions. layer of water (higher density). This is because droplets tend to merge with their neighbors when they collide with them, which eventually leads to complete phase separation. The driving force for this process is the fact that the contact between oil and water molecules is thermodynamically unfavorable, so that emulsions are thermodynamically unstable systems (Chapter 7). It is possible to form emulsions that are kinetically stable (metastable) for a reasonable period of time (a few days, weeks, months, or years), by including substances known as stabilizers (Chapter 4). A stabilizer is any ingredient that can be used to enhance the kinetic stability of an emulsion and may be classified as an emulsifier, a texture modifier, a weighting agent, or a ripening inhibitor depending on its mode of action. Emulsifiers are surface-active molecules that adsorb to the surface of freshly formed droplets during homogenization, forming a protective layer that prevents the droplets from coming close enough together to aggregate (Chapters 6 and 7). Most emulsifiers are amphiphilic molecules, that is, they have polar and nonpolar regions on the same molecule. The most common emulsifiers used in the food industry are smallmolecule surfactants, phospholipids, proteins, and polysaccharides (Section 4.4). Texture modifiers can be divided into two categories depending on their mode of operation and the rheological characteristics of their solutions: thickening agents and gelling agents (Section 4.5). Thickening agents are ingredients that are used to increase the viscosity of the continuous phase of emulsions, whereas gelling agents are ingredients that are used to form a gel in the continuous phase of emulsions. Texture modifiers therefore enhance emulsion stability by retarding the movement of the droplets. In the food industry, the most commonly used thickening and gelling agents are usually polysaccharides or proteins in O/W emulsions and fat crystals in W/O emulsions (Section 4.5). A weighting agent is a substance that is added to the dispersed phase (droplets) so as to decrease the density contrast between the droplets and surrounding liquid, thereby retarding gravitational separation. Weighting agents are commonly used in the beverage industry to increase the density of flavor oil droplets and therefore improve their stability to creaming. Ripening inhibitors are highly nonpolar substances (very low water solubility) that are added to the oil phase of O/W emulsions to inhibit Ostwald ripening through an entropy of mixing effect (Chapter 7). When developing an emulsion-based product, it is extremely important to identity the most appropriate stabilizer or combination of stabilizers to use based on knowledge of the instability mechanisms in the system. An appreciation of the difference between the thermodynamic stability of a system and its kinetic stability is crucial for an understanding of the properties of food emulsions. Consider a system that consists of a large number of molecules that can occupy two different free energy states: Glow and Ghigh (Figure 1.3). 5 Context and Background Emulsion Thermodynamically unstable Gf Kinetically stable Energy barrier ΔG* Kinetically unstable Thermodynamically stable ΔG Gi Seperated phases FIGURE 1.3 Schematic demonstration of the difference between thermodynamic and kinetic stability. A system can remain in a thermodynamically unstable (metastable) state for some time if there is a sufficiently large free energy barrier preventing it from reaching the state with the lowest free energy. The state with the lowest free energy is the one that is thermodynamically favorable and therefore the one that the molecules are most likely to occupy. At thermodynamic equilibrium, the two states are populated according to the Boltzmann distribution (Walstra 2003): æ ( Ghigh - Glow ) ö nhigh = exp ç ÷÷ ç nlow kT è ø (1.1) where nlow and nhigh are the number of molecules that occupy the energy levels Glow and Ghigh k is Boltzmann’s constant (k = 1.38 × 10 −23 J K−1) T is the absolute temperature The larger the difference between the two free energy levels compared to the thermal energy of the system (kT), the greater the fraction of molecules in the lower free energy state. In practice, a system may not be able to reach equilibrium during the timescale of an observation because of the presence of a free energy barrier, ΔG*, between the two states (Figure 1.3). A system in the high free energy state must increase its free energy by more than ΔG* before it can move into the low-energy state. The rate at which a transformation from a high to a low free energy state occurs therefore decreases as the height of the free energy barrier increases. When the free energy barrier is sufficiently large, the system may remain in a thermodynamically unstable state for a considerable length of time, in which case it is said to be kinetically stable or metastable. In food emulsions, there are actually a large number of intermediate metastable states between the initial emulsion and the completely separated phases, and there are free energy barriers associated with transitions between each of these states. Nevertheless, it is often possible to identify a single free energy barrier, associated with a particular physicochemical process, which is the most important factor determining the overall kinetic stability of an emulsion (Chapter 7). 6 Food Emulsions: Principles, Practices, and Techniques 1.2.2 Mechanisms of Emulsion Instability The term “emulsion stability” is broadly used to describe the ability of an emulsion to resist changes in its properties with time (Chapter 7). An emulsion may break down due to physical changes (changes in the relative location of its components) or chemical changes (changes in the chemistry of its components). There are a variety of physicochemical mechanisms that may be responsible for alterations in emulsion properties, and it is usually necessary to establish which of these mechanisms are important in the particular system under consideration before effective strategies can be developed to improve the stability. A number of the most common physical mechanisms that are responsible for the instability of food emulsions are shown schematically in Figure 1.4. Creaming and sedimentation are both forms of gravitational separation. Creaming describes the upward movement of droplets due to the fact that they have a lower density than the surrounding liquid, whereas sedimentation describes the downward movement of droplets due to the fact that they have a higher density than the surrounding liquid. Flocculation and coalescence are both types of droplet aggregation. Flocculation occurs when two or more droplets come together to form an aggregate in which the droplets retain their individual integrity, whereas coalescence is the process whereby two or more droplets merge together to form a single larger droplet. Extensive droplet coalescence can eventually lead to the formation of a separate layer of oil on top of a sample, which is known as “oiling-off.” Phase inversion is the process whereby an O/W emulsion is converted into a W/O emulsion, or vice versa. The physicochemical origin of these and the other major forms of emulsion instability are discussed in Chapter 7, along with factors that influence them, strategies for controlling them, and analytical techniques for monitoring them. In addition to the physical processes mentioned earlier, it should be noted that there are also various chemical, biochemical, and microbiological processes that occur in food emulsions that can also adversely affect their shelf-life and quality, for example, lipid oxidation, enzyme hydrolysis, and bacterial growth. 1.2.3 Ingredient Partitioning in Emulsions Emulsions are heterogeneous materials whose composition and properties vary from region to region when examined at length scales of the order of nanometers or micrometers. To a first approximation, many conventional food emulsions can be considered to consist of three distinct regions with different physicochemical properties: the interior of the droplets, the continuous phase, and the interface (Figure 1.5). The molecules in an emulsion distribute themselves between these three regions according to their Kinetically stable emulsion Gravitational separation Coalescence Phase separation Flocculation FIGURE 1.4 Food emulsions may become unstable through a variety of physical mechanisms, including creaming, sedimentation, flocculation, coalescence, and phase inversion. 7 Context and Background Particle Characteristics: • Size • Composition • Physical properties Food-grade Emulsifiers Phospholipids -- - - - Surfactants - - -- Interfacial Characteristics: • Composition • Charge • Thickness and structure • Chemistry • Physical properties Polysaccharides and proteins FIGURE 1.5 The ingredients in an emulsion partition themselves between the oil, water, and interfacial regions according to their concentration and interactions with the local environment. concentration and polarity (Chapter 9). Nonpolar molecules tend to be located primarily in the oil phase, polar molecules in the aqueous phase, and amphiphilic molecules at the interface. Even at equilibrium, there is a continuous exchange of molecules between the different regions, which occurs at a rate that depends on the mass transport of the molecules through the various phases of the system. Molecules may also move from one region to another when there is some alteration in the environmental conditions of an emulsion, for example, a temperature change or dilution. The location and mass transport of the molecules within an emulsion has a significant influence on the stability (Chapter 7), flavor (Chapter 9), and release characteristics (Chapter 13) of emulsions. 1.2.4 Dynamic Nature of Emulsions Many of the properties of emulsions can only be understood with reference to their dynamic nature. The formation of emulsions by homogenization is a highly dynamic process that involves the violent disruption of droplets and the rapid movement of surface-active molecules from the bulk liquids to the interfacial region (Chapter 6). Even after their formation, the droplets in an emulsion are in continual motion and frequently collide with one another because of their Brownian motion, gravity, or applied mechanical forces (Chapter 7). The continual movement and interactions of droplets cause the properties of emulsions to evolve over time due to the various destabilization mechanisms mentioned earlier (Section 1.2.2). Biopolymers adsorbed to the surface of emulsion droplets may undergo relatively slow conformational changes over time, which result in alterations in the stability and physicochemical properties of the overall system. Surfaceactive molecules in the continuous phase may exchange with those adsorbed to the droplet surfaces, thus changing the composition and properties of the droplet interfaces. The properties of the system may also change over time due to chemical reactions that occur in the droplet interior, interfacial region, or continuous phase, for example, oxidation of lipids or proteins, or hydrolysis of proteins or polysaccharides. An appreciation of the dynamic processes that occur in food emulsions is therefore extremely important for a thorough understanding of their physicochemical, organoleptic, and physiological properties. 1.2.5 Complexity of Food Emulsions Most food emulsions are much more complex than the simple three-component (oil, water, and emulsifier) systems described earlier (Section 1.2.1). The aqueous phase may contain a variety of 8 Food Emulsions: Principles, Practices, and Techniques water-soluble ingredients, including sugars, salts, acids, bases, buffers, alcohols, surfactants, proteins, polysaccharides, and preservatives. The oil phase may also contain a complex mixture of lipid-soluble components, such as triacylglycerols, diacylglycerols, monoacylglycerols, free fatty acids, sterols, vitamins, fat replacers, weighting agents, colors, flavors, and preservatives. The interfacial region may contain a mixture of different surface-active components, including proteins, polysaccharides, phospholipids, surfactants, alcohols, and molecular complexes. In addition, these components may form various types of structural entities in the oil, water, or interfacial regions (such as fat crystals, ice crystals, biopolymer aggregates, air bubbles, liquid crystals, and surfactant micelles), which in turn may associate to form larger structures (such as biopolymer or particulate networks). A further complicating factor is that foods are subjected to variations in their temperature, pressure, and mechanical agitation during their production, storage, and handling, which can cause significant alterations in their overall properties. It is clear from the earlier discussion that food emulsions are compositionally, structurally, and dynamically complex materials and that many factors contribute to their overall properties. One of the major objectives of this book is to present the conceptual framework needed by food scientists to understand these complex systems in a more systematic and rigorous fashion. Much of our knowledge about these complex systems has come from studies of simple model systems (Section 1.5). Nevertheless, there is an increasing awareness of the need to elucidate the factors that determine the properties of actual emulsion-based food products. For this reason, many researchers are now systematically focusing on understanding at a fundamental level the factors that determine the properties of real food emulsions, such as ingredient interactions (e.g., biopolymer–biopolymer, biopolymer–surfactant, and biopolymer– water) and processing conditions (e.g., homogenization, freezing, chilling, cooking, sterilization, pasteurization, mechanical agitation, pressurization, and drying). 1.3 Emulsion Properties 1.3.1 Dispersed-Phase Volume Fraction The concentration of droplets in an emulsion plays an important role in determining its cost, appearance, texture, flavor, stability, and nutritional attributes (Chapters 7 through 11). It is therefore important to be able to clearly specify and reliably report the droplet concentration of emulsions. The droplet concentration is usually described in terms of the dispersed-phase volume fraction (ϕ), which is equal to the volume of emulsion droplets (VD) divided by the total volume of the emulsion (VE): ϕ = VD/VE. In some situations, it is more convenient to express the droplet concentration in terms of the dispersed-phase mass fraction (ϕm), which is equal to the mass of emulsion droplets (mD) divided by the total mass of the emulsion (mE): ϕm = mD / mE. Frequently, the droplet concentration is expressed as a volume or mass percentage, rather than as a volume or mass fraction. The mass fraction and volume fraction are related to each other through the following equations: fr2 r2f + (1 - f)r1 (1.2) fmr1 r1fm + (1 - fm )r2 (1.3) fm = f= where ρ1 and ρ2 are the densities of the continuous and dispersed phases, respectively. When the densities of the two phases are equal, the mass fraction is equivalent to the volume fraction. However, if the densities of the two phases are appreciably different, then there are significant differences between these two ways of reporting the particle concentration. 9 Context and Background In reality, emulsion droplets are not homogeneous spheres, but actually consist of a core of dispersed phase surrounded by a shell of emulsifier molecules. If the thickness of the interfacial layer (δ) is appreciable compared to the droplet radius (r), then the effective volume fraction of the coated droplets will be larger than that of the volume fraction of the dispersed phase: ϕeff = ϕ(1 + δ/r)3. This increase has important consequences for the stability and rheology of nanoemulsions and some conventional emulsions (Chapters 7 and 8). The dispersed-phase volume fraction of an emulsion is often known because the concentration of the ingredients used to prepare it is carefully controlled during the manufacturing process. Nevertheless, local variations in dispersed-phase volume fraction occur within emulsions when the droplets accumulate at either the top or bottom of an emulsion due to creaming or sedimentation. In addition, the dispersed-phase volume fraction of an emulsion may vary during a food processing operation, for example, if a mixer or valve is not operating efficiently. Consequently, it is often important to have analytical techniques to measure dispersed-phase volume fraction (Chapter 14). 1.3.2 Particle Size Many of the most important properties of emulsion-based food products are determined by the size of the droplets that they contain, for example, shelf-life, appearance, texture, release characteristics, flavor profile, and biological fate (Chapters 7 through 11). Consequently, it is important for food scientists to be able to reliably control, predict, measure and report the size of the droplets in emulsions. In this section, the most widely used methods for reporting droplet size data are discussed. Methods of controlling, predicting, and measuring droplet size are covered in later chapters (Chapters 6, 7, and 14). If all the droplets within an emulsion have exactly the same dimensions, it is referred to as a monodisperse emulsion, but if there is a range of droplet sizes present, it is referred to as a polydisperse emulsion (Figure 1.6). The droplet size (x) of a monodisperse emulsion can be completely characterized by a single number, such as the droplet diameter (d) or radius (r). Monodisperse emulsions are sometimes prepared and used for fundamental studies because the interpretation of experimental measurements is much simpler than for polydisperse emulsions. Nevertheless, real food emulsions always contain a distribution of droplet sizes, and so the specification of their droplet size is more complicated than for monodisperse systems. In some situations it is important to have information about the full particle size distribution of an emulsion (i.e., the fraction of droplets in different size classes), whereas in other situations knowledge of the average droplet size is sufficient. A common error that occurs when particle size data are presented in scientific publications and presentations is that the investigator neglects to say whether the size is reported as a radius or a diameter. Obviously, this practice should be avoided because it can cause considerable confusion and lead to misleading interpretations of reported data. (a) (b) FIGURE 1.6 Schematic representation of (a) monodisperse and (b) polydisperse emulsions. In a monodisperse emulsion, all the droplets have the same size, but in a polydisperse emulsion, they have a range of different sizes. 10 Food Emulsions: Principles, Practices, and Techniques 1.3.2.1 Collecting Particle Size Data Information about the size of the particles within an emulsion may be obtained using various analytical methods, including microscopy, light scattering, particle counting, and sedimentation methods (Chapter 14). The nature of the particle size data collected depends on the method used, for example, it may consist of an average particle size or a particle size distribution. For some analytical methods, information about particle size is obtained by direct observation of the sample (such as microscopy). In other methods, the particle size is inferred from measurements of some particle-size-dependent physical property of the system by fitting the data with an appropriate mathematical model, for example, using light scattering theory to interpret the diffraction pattern of an emulsion. In this latter case, it is important that the mathematical model used is appropriate for the system being tested; otherwise, erroneous results will be obtained (Chapter 14). 1.3.2.2 Presenting Particle Size Data The number of droplets in most emulsions is extremely large (Table 1.1), and so their size can be considered to vary continuously from some minimum value to some maximum value (Walstra 2003). When presenting particle size data, it is usually convenient to divide the full particle size range into a number of discrete particle size classes and then stipulate the concentration of droplets that fall into each class (Hunter 1986). The resulting data can then be presented in tabular form (Table 1.2) or as a histogram (Figure 1.7a). Histograms are usually plotted so that the height of each bar represents the amount of particles in the stipulated size class, and the central position (xi) of each bar on the x-axis represents the average size of the particles within the size class, for example, xi = (x low + x high)/2, where x low and x high are the lower and upper boundaries of the size class. Ideally, the width of the bars on the histogram should be proportional to the width of the size classes (which may be the same or different for each size class). In practice, this is often not done because the graphic programs used to plot data are not sufficiently flexible. In addition, the data are often presented as a continuous curve rather than as a histogram. It is usually more convenient to represent the amount of particles in each size class as a fraction rather than as an absolute value because it is then possible to directly compare particle size distributions of emulsions with different total droplet concentrations. The fraction of particles in a size class can be defined in a number of different ways, for example, the number ratio, fi = ni /N, where N is the total number of droplets, or the volume ratio, ϕi = vi /V, where vi is the volume of the droplets in the ith size class and V is the total volume of all the droplets in the emulsion. The shape of a particle size distribution changes appreciably depending on whether it is presented as a number or volume ratio (Table 1.2; Figure 1.8). The volume of a droplet is proportional to x3, and so a volume distribution is skewed more toward the larger droplets, whereas a number distribution is skewed more toward the smaller droplets. For convenience, the x-axis of particle size distributions is often plotted on a logarithmic scale because the size of the particles within a distribution may vary by a few orders of magnitude. TABLE 1.1 Effect of Particle Dimensions on the Physical Characteristics of 1 g of Oil Dispersed within Water in the Form of Spherical Droplets Droplet Radius (μm) 100 10 1 0.1 0.01 Number of Droplets per Gram Oil (g−1) Droplet Surface Area per Gram Oil (m2 g−1) Percentage of Oil Molecules at Droplet Surface (vol%) 2.6 × 105 2.6 × 108 2.6 × 1011 2.6 × 1014 2.6 × 1017 0.03 0.3 3 30 300 0.02 0.2 1.8 18 100 Note: Values were calculated assuming the oil had a density of 920 kg m−3 and the end-to-end length of an individual oil molecule was 6 nm. 11 Context and Background TABLE 1.2 Particle Size Distribution of an Emulsion Can Be Conveniently Represented in Tabular Form dmin (μm) 0.041 0.054 0.071 0.094 0.123 0.161 0.211 0.277 0.364 0.477 0.626 0.821 1.077 1.414 1.855 2.813 3.191 3.620 dmax (μm) 0.054 0.071 0.094 0.123 0.161 0.211 0.277 0.364 0.477 0.626 0.821 1.077 1.414 1.855 2.433 3.192 3.620 4.107 di (μm) ni f i (%) ϕ i (%) 0.048 0.063 0.082 0.108 0.142 0.186 0.244 0.320 0.421 0.551 0.724 0.949 1.245 1.634 2.144 3.002 3.406 3.864 0 2 20 38 89 166 243 360 420 361 256 145 78 23 6 1 0 0 0.0 0.1 0.9 1.7 4.0 7.5 11.0 16.3 19.0 16.3 11.6 6.6 3.5 1.0 0.3 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.2 0.5 1.8 4.7 9.1 14.5 18.6 22.6 15.1 8.9 4.1 0.0 0.0 F(di) (μm−1) 0.0 117.6 888.9 1333.3 2342.1 3320.0 3681.8 4161.8 3716.8 2431.0 1312.8 566.4 231.8 52.2 10.4 2.6 0.0 0.0 C(di) (%) 0.0 0.1 1.0 2.7 6.7 14.3 25.3 41.6 60.6 76.9 88.5 95.1 98.6 99.7 100.0 100.0 100.0 100.0 Selected mean particle diameters and relative standard deviations 0.49 µm c0 0.63 d10 d20 0.57 µm c1 0.59 d30 0.67 µm c2 0.55 0.92 µm d32 d43 1.20 µm Note: The volume frequency (ϕi) is much more sensitive to larger droplets than the number frequency ( fi). A particle size distribution can also be represented as a smooth curve such as the frequency distribution F(xi) or cumulative distribution C(xi) shown in Figures 1.7b and c, respectively. For a continuous particle size distribution, the number of particles in a given size class can be calculated from the frequency distribution by the following equation (Walstra 2003): xi + 12 Dxi ni = ò F ( xi ) d ( x ) (1.4) xi - 12 Dxi where Δxi is the width of the size class xi is the value of the particle size at the midpoint of the size class For a discrete particle size distribution, the number of particles in a size class is given by ni ≈ FiΔxi, where Fi and Δxi are the number frequency distribution and width of the ith size class. The number frequency distribution is therefore constructed so that the area under the curve between two droplet sizes is approximately equal to the number of droplets ni in that size range (Hunter 1986). This relationship can be used to convert a histogram to a frequency distribution curve, or vice versa. The number ratio in a particular size class can be related to the number frequency distribution by the following equation: fi = FiΔxi /N. 12 Food Emulsions: Principles, Practices, and Techniques 20 fi (%) 15 10 0 Polydisperse emulsion (a) (b) 0.048 0.063 0.082 0.108 0.142 0.186 0.244 0.320 0.421 0.551 0.724 0.949 1.245 1.634 2.144 3.002 3.406 3.864 5 di (µm) 5000 100 75 3000 C(di) (%) F(di) (µm–1) 4000 2000 25 1000 0 (c) 50 0 1 di (µm) 2 0 3 dm 0 (d) 1 2 3 di (µm) FIGURE 1.7 The (a) particle size distribution of an emulsion can be represented in a number of different ways: (b) a histogram, (c) a frequency distribution F(d), and (d) a cumulative distribution C(d). These different ways of representing the particle size distribution are described in the text. Droplets in size-class (%) 25 20 Volume Number 15 10 5 0 0.01 0.1 1 10 Mid-point diameter (μm) FIGURE 1.8 Comparison of the particle size distribution of an emulsion plotted as either number or volume ratio versus diameter (see Table 1.2). The data are plotted as curves, rather than as a histogram, to facilitate comparison. 13 Context and Background A similar approach can be used to represent a volume frequency distribution, but using volume rather than number as a measure of the particle concentration in each size class. The cumulative distribution represents the percentage of droplets that are smaller than xi (Figure 1.7c). The resulting curve usually has an S shape that varies from 0% to 100% as the particle size increases from the smallest to the largest value. The particle size at which half the droplets are smaller and the other half are larger is known as the median particle size, xm. The size class that contains the largest amount of particles is called the mode or modal particle size (xmodal). The numerical values of the median and the modal droplet sizes depend on the way that the concentration of particles in each size class is expressed, for example, number or volume. Hence, there are number and volume median sizes, and number and volume modal sizes of a distribution. 1.3.2.3 Mean and Standard Deviation It is often convenient to represent the size of the droplets in a polydisperse emulsion by one or two numbers, rather than stipulating the full particle size distribution (Hunter 1986). The most useful numbers are the mean particle size, x, which is a measure of the central tendency of the distribution, and the standard deviation, σ, which is a measure of the width of the distribution. The mean and standard deviation of a particle size distribution can be calculated using the following simple equations: ån x i i x= i =1 å n (x - x ) i s= (1.5) N i 2 i =1 (1.6) N Here, the summations are carried out over the total number of size classes used to represent the particle size distribution. Nevertheless, this is only one way of expressing the mean and standard deviation, and there are numerous other ways that emphasize different physical characteristics of the particle size distribution that are often more appropriate (Hunter 1986, Walstra 2003). To understand these different ways of representing the mean and standard deviation of particle size distributions, it has proved useful to define the concept of the “moment of a distribution”: ¥ ò Sn = x n F ( x ) dx » ån x n i i (1.7) i =1 0 where Sn is the nth moment of the distribution.* The mean particle size and relative standard deviation can then be defined as 1/( a - b ) æS ö xab = ç a ÷ è Sb ø (1.8) 1/ 2 æS S ö cn = ç n 2 n + 2 - 1 ÷ è Sn +1 ø (1.9) where a and b are integers (usually between 0 and 6) cn is the (dimensionless) relative standard deviation weighted with the nth power of x * Note: one must be careful here to distinguish between the “n” that represents the number of droplets in a size-class (ni) and the “n” that represents the nth moment of the distribution. 14 Food Emulsions: Principles, Practices, and Techniques The moment of the distribution has no physical meaning itself, but it can often be simply related to important physical characteristics of the distribution. For example, if the particle size is expressed as the diameter (di) of the particles in each size class, and ni is expressed as the number of particles in each size class per unit volume of emulsion, S0 » ån » N (1.10) i i =1 S1 » ån d = L N (1.11) AN p (1.12) 6VN 6f = p p (1.13) i i i =1 S2 » ån d 2 i i = i =1 S3 » ån d 3 i i i =1 = where L N, AN, and VN are the total length, surface area, and volume of the droplets per unit emulsion volume ϕ is the dispersed-phase volume fraction Some of the most important ways of expressing the mean droplet size of particle size distributions are highlighted in Table 1.3. Each of these mean sizes has dimensions of length (meters), but stresses a different physical aspect of the distribution. In general, mean particle sizes can be calculated by multiplying a weighting factor (such as the number, surface area, or volume of particles in a particular size class) by some parameters related to the size of the particles within that size class (such as the droplet diameter) after a suitable conversion has been made to obtain the appropriate units of length (meters). The three most commonly used mean particle size values are the number-weighted mean diameter (d10 = Σnidi/Σni), the surface-weighted mean diameter (d32 = Σnidi3/Σnidi2), and the volume-weighted mean diameter (d43 = Σnidi4/Σnidi3). All of these values are equal for a monodisperse emulsion, but they may be considerably different for a polydisperse emulsion. Typically, the larger the difference between different mean size values, the greater the polydispersity of the emulsion. Generally, the volume-weighted mean diameter is more sensitive to the presence of large particles within a polydisperse system than the numberweighted mean diameter (Figure 1.8). Consequently, d43 is often more useful than d10 or d32 for detecting small amounts of flocculation or coalescence in an emulsion. Because the mean particle size depends on the approach used to calculate it, it is always important when interpreting or reporting particle size data to identify which mean particle size (and corresponding standard deviation) is being used. In addition, mean particle size values should be treated with caution when used to represent highly polydisperse TABLE 1.3 Different Ways of Expressing the Mean Particle Diameter of Polydisperse Emulsions Name of Mean Number–length mean diameter Number–area mean diameter Number–volume mean diameter Area–volume mean diameter Volume–length mean diameter Symbol Definition Quantity Averaged Weighting Factor dNL or d10 dNA or d20 dNV or d30 dAV or d32 dVL or d43 d10 = S1/S0 d20 = (S2/S0)1/2 d30 = (S3/S0)1/3 d32 = S3/S2 d43 = S4/S3 Diameter (∝ L) Diameter squared (∝ A) Diameter cubed (∝ V) Diameter cubed (∝ V) Diameter (∝ L) Number in class Number in class Number in class Area in class Volume in class Note: Here L, A, and V represent length, surface area, and volume, respectively. 15 Context and Background emulsion (e.g., aggregated systems). In these cases, it is usually better to utilize and report the full particle size distribution to obtain a more accurate and reliable representation of the system. The surface-weighted mean diameter (d32) is related to the average surface area of droplets exposed to the continuous phase per unit volume of emulsion, AN: AN = 6f d32 (1.14) This relationship is particularly useful for calculating the total surface area of droplets in an emulsion from knowledge of the mean droplet diameter (d32) and the dispersed-phase volume fraction (ϕ). This information is important for determining the amount of emulsifier required to cover the droplet surfaces in an emulsion, or to understand the role of droplet surface area on interfacial chemical reactions, such as lipid oxidation or digestion. In general, the higher the order of the mean (a + b, Equation 1.8) used to describe the particle size distribution, the higher its numerical value. For narrow particle size distributions, the different mean values are fairly similar, but for wide particle size distributions, they may differ appreciably (Table 1.2). For narrow particle size distributions, it is often sufficient to report only the mean particle size, but for wide particle size distributions, it is often important to provide some measure of the width of the distribution also. The width of a particle size distribution can be conveniently represented by the relative standard deviation (Equation 1.9). For example, for n = 2, the relative standard deviation (c2) is the standard deviation of the size distribution weighted by the particle surface area divided by d32 (Walstra 2003). In general, c2 ranges from around 0.1 for a very narrow distribution to 1.3 for a very wide distribution, but in most food emulsions, it usually ranges from around 0.5 to 1. It should be noted that the absolute standard deviation (Equation 1.6) is not a good representation of the width of a distribution, since it is highly sensitive to the mean particle size. For example, consider two particle size distributions with an absolute standard deviation of 0.1 μm, but with mean particle diameters of 0.2 and 20 μm. The droplets in the emulsion with the smallest mean diameter span a relatively wide range compared to the mean (∼0.1–0.3 μm, c2 = 0.5), whereas those in the emulsion with the largest mean diameter span a fairly narrow range (∼19.9–20.1 μm, c2 = 0.05). Another important reason for being aware of the various ways of calculating and reporting the mean particle size is that different particle characterization techniques determine different mean values (Hunter 1986, Orr 1988, Hiemenz and Rajagopalan 1997). For example, analysis of polydisperse emulsions using microscopy measurements of droplet length gives d10, imaging processing of droplet area gives d20, electrical pulse counting gives d30, and low-angle laser light scattering gives d43 (Rawle 2004). Consequently, it is always important to be clear about which mean size has been determined in an experiment when using or quoting particle size data. In addition, one must be particularly careful in using data determined by one type of particle sizing technique to calculate a mean particle size that is different from the one that the technique is most sensitive to (Rawle 2004). For example, static light scattering is much more sensitive to the volume of particles in each size class than to the number of particles. Hence, there could be a large number of small particles (but with a small total volume) that would not be accurately detected by this kind of light scattering, so that a calculation of d10 from the data would not be particularly accurate. Finally, it should be stressed that one must also be careful when choosing either a mode, median, or mean diameter to represent a full particle size distribution in a polydisperse emulsion (Rawle 2004). For a normal distribution (which is symmetrical), the mode, median, and mean have similar values, but for a nonsymmetrical or multimodal distribution, they have very different values. One must therefore select one or more of these parameters to represent the full particle size distribution based on the physical property that is most pertinent to the person who is going to use the information. 1.3.2.4 Mathematical Models The particle size distribution of an emulsion can often be modeled using a mathematical theory, which is convenient because it means that the full data set can be described by a small number of parameters (Hunter 1986). In addition, many analytical instruments designed to measure particle size distributions 16 Food Emulsions: Principles, Practices, and Techniques assume that the distribution has a certain mathematical form so as to facilitate the conversion of the measured physical parameters (e.g., light intensity versus scattering angle) into a particle size distribution (Hunter 1986). If a plot of frequency distribution versus droplet size is symmetrical about the mean droplet size, the curve can often be described by a normal frequency distribution function (Figure 1.9a): F ( x) = é -( x - x )2 ù 1 exp ê ú 2 s 2p ë 2s û (1.15) This function has a maximum value when x = x . Most (∼68%) of the droplets fall within one standard deviation of the mean ( x ± s), while the vast majority (∼99.7%) fall within three standard deviations (x ± 3s). Only two parameters are needed to describe the full particle size distribution of an emulsion that can be approximated by a normal distribution: the mean and the standard deviation. The number ratio of droplets within a particular size range (x low to x high) can be calculated from the earlier equation: xhigh fi = ò F( x)dx /N (1.16) xlow The particle size distribution of most food emulsions is not symmetrical about the mean but tends to extend much further at the high droplet size end than at the low droplet size end (Figure 1.9b). This type of distribution can often be described by a log-normal frequency distribution function (Hiemenz and Rajagopalan 1997): F (ln x ) = 1 ln sg é -(ln x - ln xg )2 ù exp ê ú 2 ln 2 sg 2p ë û 12 12 10 8 6 σ = 0.2 μm 4 2 (a) Number percentage, fi (%) Number percentage, fi (%) ln(σg) = 1.1 σ = 0.1 μm 10 0 (1.17) 0.5 1 1.5 Droplet diameter, di (μm) 2 6 4 ln(σg) = 1.2 2 σ = 0.3 μm 0 8 0 2.5 (b) ln(σg) = 1.4 0 0.5 1 1.5 2 2.5 Droplet diameter, di (μm) FIGURE 1.9 Particle size distributions, represented as droplet number percentage ( f i) versus droplet diameter (di) calculated assuming different standard deviations for normal and log-normal distributions. For both emulsions, the mean droplet diameter was assumed to be 1 μm. (a) Normal distribution, f i = F(di)d(di); (b) log-normal distribution, f i = F(ln di)d(ln di). 17 Context and Background where xg and σg are the geometric mean and the standard deviation of the geometric mean, which are given by the following expressions: å n ln x i ln xg = (1.18) N å n ( ln x - ln x ) i ln sg = i i =1 i 2 g i =1 (1.19) N If the log-normal curve shown in Figure 1.9b was plotted as fi versus ln di rather than as fi versus di, it would be symmetrical about ln dg. It should be stressed that the particle size distribution of many food emulsions cannot be adequately described by the simple mathematical models given earlier. Bimodal distributions that are characterized by two peaks (Figure 1.10) are often encountered in food emulsions, for example, due to droplet flocculation or coalescence (Chapters 6 and 7). For these systems, it is often better to present the data as the full particle size distribution; otherwise, considerable errors may occur if an inappropriate mathematical model is used. This kind of problem can occur when one is using an analytical instrument that assumes a particular mathematical model when calculating the particle size distribution, for example, a light scattering or ultrasonic spectrometry instrument. If the mathematical model is inappropriate, then the instrument may still report a particle size distribution, but this distribution will be incorrect. The user of the instrument should therefore be aware of this potential problem and, if necessary, ensure that the mathematical model is correct by using some independent technique to verify the particle size distribution (e.g., microscopy). For example, a static light scattering instrument calculates the particle size distribution from the measured laser diffraction assuming that the scattering particles are spherical and 20 Volume percentage, Фi (%) 15 10 5 0 0.1 1 10 Droplet diameter, di (μm) FIGURE 1.10 Example of a bimodal distribution resulting from the heat-induced flocculation of droplets in a hexadecane O/W emulsions stabilized by β-lactoglobulin. (Adapted from Kim, H.J. et al., Langmuir, 18(20), 7577, 2002.) 18 Food Emulsions: Principles, Practices, and Techniques homogeneous. For a flocculated emulsion, such as that shown in Figure 1.10 (Kim et al. 2002), some of the particles are nonspherical and nonhomogeneous, and therefore the reported data should only be treated as an indication of the actual particle size distribution and should usually be confirmed using another method, such as microscopy. 1.3.3 Interfacial Properties The droplet interface consists of a narrow region (usually a few nanometers thick) that surrounds each emulsion droplet and contains a mixture of oil, water, and other surface-active molecules (Figure 1.5). The interfacial region only makes up a significant fraction of the total volume of an emulsion when the droplet radius is less than about 1 μm (Table 1.1). Even so, it plays a major role in determining many of the most important bulk physicochemical and organoleptic properties of food emulsions. For this reason, food scientists are particularly interested in elucidating the factors that determine the composition, structure, thickness, rheology, and charge of the interfacial region and in elucidating how these interfacial characteristics are related to the bulk physicochemical, sensory, and nutritional properties of emulsions. The composition and structure of the interfacial region are determined by the type and concentration of surface-active species present prior to emulsion formation, as well as by the events that occur during and after emulsion formation, for example, competitive adsorption and displacement (Chapters 5 and 6). The thickness and rheology of the interfacial region may influence the stability of emulsions to gravitational separation, coalescence, and flocculation (Chapters 3 and 7), the rheology of emulsions (Chapter 8), the mass transport rate of molecules in or out of droplets (e.g., Ostwald ripening, compositional ripening, and flavor release) (Chapters 7 and 9), or the gastrointestinal fate of emulsions (Chapter 11). The interface acts as a region where surface-active components accumulate, which may lead to acceleration of certain types of chemical reactions (e.g., lipid oxidation), either by increasing the local concentration of molecules or by bringing together different reactive species (McClements and Decker 2000). The major factors that determine the characteristics of the interfacial region are discussed in Chapter 5, along with experimental techniques to characterize its properties. The electrical characteristics of the interface are discussed in the following section. 1.3.4 Droplet Charge The bulk physicochemical, organoleptic, and nutritional properties of many food emulsions are governed by the magnitude and sign of the electrical charge on the droplets (Chapters 7 through 11). The origin of this charge is normally the adsorption of emulsifier molecules that are ionized or ionizable* (Section 4.4). Surfactants have hydrophilic head groups that may be neutral, positively charged, or negatively charged. Proteins may also be neutral, positively charged, or negatively charged depending on the pH of the solution compared to their isoelectric point. Surface-active polysaccharides may also have an electrical charge depending on the type of functional groups along their backbone. Electrically charged hydrophilic substances, such as mineral ions or polyelectrolytes, may also adsorb onto the surfaces of emulsifier-coated oil droplets and thereby alter their charge. Consequently, emulsion droplets may have an electrical charge that depends on the types of ionizable molecules present and the pH of the aqueous phase. The electrical charge on a droplet can be characterized in a number of different ways (Hunter 1986), that is, the surface charge density (σ), the electrical surface potential (ψ0), and the zetapotential (ζ). The surface charge density is the amount of electrical charge per unit surface area, whereas the surface potential is the free energy required to increase the surface charge density from zero to σ. The ζ-potential is the effective surface potential of a particle suspended in a medium, which takes into account that charged species in the surrounding medium may adsorb to the surface of the droplet and alter its net charge. The ζ-potential can be conveniently measured in the laboratory using commercially available analytical instrumentation (Chapter 14). The charge on an emulsion droplet is important because it determines the nature of its interactions with other charged species (Chapters 2 through 4) or its behavior in the presence of an electrical field * There is experimental evidence that even oil droplets stabilized by nonionic surfactants have an electrical charge because the oil preferentially adsorbs either OH− or H3O+ ions from water or contains ionic impurities (Pashley 2003). Context and Background 19 (Chapter 14). Two species that have charges of opposite sign are attracted toward each other, whereas two species that have charges of similar sign are repelled (Chapters 2 and 3). All of the droplets in an emulsion are usually coated with the same type of emulsifier, and so they have the same electrical charge (if the emulsifier is ionized). When this charge is sufficiently large, the droplets are prevented from aggregating because of the electrostatic repulsion between them (Chapter 3). The properties of emulsions stabilized by ionized emulsifiers are particularly sensitive to the pH and ionic strength of the aqueous phase. If the pH of the aqueous phase is adjusted so that the emulsifier loses its charge, or if salt is added to screen the electrostatic interactions between the droplets, the repulsive forces may no longer be strong enough to prevent the droplets from aggregating (Chapters 3 and 7). Droplet aggregation often leads to a large increase in emulsion viscosity (Chapter 8) and may cause the droplets to cream more rapidly (Chapter 7). Electrostatic interactions also influence the interactions between emulsion droplets and other charged species, such as biopolymers, surfactants, vitamins, antioxidants, flavors, and minerals (Chapters 3, 4, 7, and 9). These interactions often have significant implications for the overall quality of an emulsion product. For example, the volatility of a flavor is reduced when it is electrostatically attracted to the surface of an emulsion droplet, which alters the flavor profile of a food (Landy et al. 1996), or the susceptibility of oil droplets to lipid oxidation depends on whether the catalyst is electrostatically attracted to the droplet surface (Mei et al. 1998, 1999). The accumulation of charged species at a droplet surface and the rate at which this accumulation takes place depend on the sign of their charge relative to that of the surface, the strength of the electrostatic interaction (which depends on ionic strength), their concentration, and the presence of any other charged species that might compete for the surface. The earlier discussion has highlighted the importance of droplet charge in determining both the physical and chemical properties of food emulsions. It is therefore important for food scientists to be able to predict, control, and measure droplet charge. For most food emulsions, it is difficult to accurately predict droplet charge because of the complexity of their composition and the lack of suitable theories. Nevertheless, there is a fairly good understanding of the major factors that influence droplet charge (Chapters 3 through 5) and of the effect of droplet charge on the stability and rheology of emulsions (Chapters 7 and 8). In addition, a variety of experimental techniques have been developed to measure the magnitude and sign of the charge on emulsion droplets (Chapter 14). 1.3.5 Droplet Crystallinity The physical state of the droplets in an emulsion can influence a number of its most important bulk physicochemical, organoleptic, and biochemical properties, including appearance, rheology, flavor, stability, and gastrointestinal fate (Chapters 7 through 11). The production of margarine, butter, whipped cream, and ice cream depends on a controlled destabilization of an O/W emulsion containing partly crystalline droplets (Chapter 13). The stability of cream to shear and temperature cycling depends on the crystallization of the milk fat droplets. The rate that milk fat droplets cream depends on their density, which is determined by the fraction of each droplet that is solidified. The cooling sensation that occurs when fat crystals melt in the mouth contributes to the characteristic mouthfeel of many food products (Walstra 2003). Knowledge of the factors that determine the crystallization and melting of emulsified substances, and of the effect that droplet phase transitions have on the properties of emulsions, is therefore particularly important to food scientists.* In O/W emulsions, we are concerned with phase transitions of emulsified fat, whereas in W/O emulsions, we are concerned with phase transitions of emulsified water. In the food industry, we are primarily concerned with the crystallization and melting of emulsified fats, because these transitions occur at temperatures that are commonly encountered during the production, storage, or handling of O/W emulsions, and because they usually have a pronounced influence on the bulk properties of food emulsions. In contrast, phase transitions of emulsified water are less likely to occur in foods because of the high degree of supercooling required to initiate crystallization (Clausse 2010, Clausse and Dalmazzone 2014). * It should be noted that the continuous phase of an emulsion is also capable of melting or crystallizing that can have a profound influence on the overall properties. For example, the characteristic texture of ice cream is partly due to the presence of ice crystals in the aqueous continuous phase, whereas the rheology of butter and margarine is determined by the existence of a network of aggregated fat crystals in the oil continuous phase. 20 Food Emulsions: Principles, Practices, and Techniques The percentage of the total fat in a sample that is solidified at a particular temperature is known as the solid fat content (SFC). The SFC varies from 100% at low temperatures where the fat is completely solid to 0% at high temperatures where the fat is completely liquid (Walstra 2003). The precise nature of the SFC–temperature curve is an important consideration when selecting a fat for a particular food product. The shape of this curve depends on the composition of the fat, the thermal and shear history of the sample, whether the sample is heated or cooled, the heating or cooling rate, the size of the emulsion droplets, and the type of emulsifier. The melting and crystallization behavior of emulsified substances can be quite different from that of the same substance in bulk. The crystallization of bulk fats is considered in Chapter 4, while the additional factors that influence the crystallization of emulsified fats are considered in Chapter 7. Experimental techniques that are used to provide information about the crystallization and melting of emulsion droplets are described in Chapter 14. 1.3.6 Droplet Interactions Many of the bulk physicochemical and sensory properties of food emulsions are strongly affected by the attractive and repulsive interactions acting between the droplets (Chapters 7 through 11). There are many different kinds of colloidal interactions that may operate in food emulsions, including van der Waals, electrostatic, steric, depletion, and hydrophobic interactions (Chapter 3). These interactions vary in their sign (attractive or repulsive), magnitude (strong to weak), and range (long to short). The overall characteristics of the droplet–droplet interactions in a particular food emulsion are determined by the relative contribution of the different kinds of colloidal interactions operating in that specific system, which depends on emulsion composition, microstructure, and environment. When the attractive forces dominate, the droplets tend to associate with each other, but when the repulsive forces dominate, the droplets tend to remain as individual entities (Chapter 3). The interactions between emulsion droplets can lead to large changes in the stability, rheology, appearance, flavor, and gastrointestinal fate of food emulsions, and so it is crucial to understand their physicochemical origin and characteristics. A brief overview of some of the analytical techniques that have recently been developed to provide information about droplet–droplet interactions is given in Chapter 14. 1.4 Hierarchy of Emulsion Properties Scientists are becoming increasingly aware of the hierarchical nature of food emulsion properties (Mezzenga et al. 2005, Ubbink et al. 2008) (Figure 1.11). Ultimately, a food emulsion consists of an extremely complex mixture of many different kinds of molecular species, for example, water, lipids, proteins, carbohydrates, surfactants, salts, and flavors. These molecular species vary in their chemical structure, polarity, reactivity, molar mass, conformation, flexibility, and dynamics. The different types of molecules present in an emulsion interact with each other to form the oil, water, and interfacial phases, as well as any other structural entities distributed within these phases, such as surfactant micelles, molecular aggregates, particle or polymer networks, air bubbles, fat crystals, or ice crystals. The physicochemical properties of the overall emulsion (e.g., optical properties, rheology, stability, and molecular partitioning) depend on the physicochemical properties of the individual oil, water, and interfacial phases, as well as the interactions that occur between these phases. The sensory properties of a food emulsion (e.g., appearance, texture, aroma, taste) depend on the direct or indirect interaction of the food emulsion and its components with the sensors in the human body, for example, light waves reflected from the emulsion reaching the eye, sound waves generated by the emulsion reaching the ear, flavor molecules released from the emulsion reaching receptors in the mouth and nose, and forces and heat generated by the emulsion interacting with tactile and temperature sensors in the hands and mouth. The way that a person responds to these sensory inputs depends on the physiology of the human sensory system; the way that the sensory information is processed, stored, and retrieved by the brain; and the way that this information is represented to consciousness. In addition, an individual’s perception of a food 21 Context and Background Human sciences (Physiology, neuroscience, and psychology) Sensory science (Appearance, aroma, taste, and texture) Physicochemical properties of emulsion (Optical properties, rheology, molecular partitioning, and stability) Physicochemical properties of individual phases Colloidal properties (Optical properties, rheology, density, and polarity) (Concentration, size, and interactions) Molecular properties (Size, conformation, flexibility, polarity, and reactivity) FIGURE 1.11 The properties of emulsion-based food products can be understood at different hierarchical levels, ranging from molecular characteristics to structural organization of molecules, to bulk physicochemical properties, to sensory properties, and ultimately to the interaction of the emulsion and its components with the human body. product is strongly dependent on their background and experiences (e.g., culture, age, sex, ethnicity, and social class). Hence, the quality or desirability of the same food product may be perceived differently by two different people or by the same person at different times. This brief discussion has highlighted the many different hierarchical levels involved in the study of food emulsions. A more complete understanding of the factors that determine the properties of emulsions depends on establishing the most important processes that operate at each hierarchical level and then linking the processes that occur at different levels to one another. It should be noted that specialized analytical techniques and theoretical concepts are often required to study each hierarchical level, and so scientists with particular areas of expertise often focus their research programs on studying a particular level. For this reason, an integrated understanding of the physicochemical basis of the properties of food emulsions usually requires collaboration of scientists with different specializations, for example, physicists, physical chemists, analytical chemists, biochemists, polymer scientists, chemical engineers, sensory scientists, physiologists, psychologists, food scientists, and nutritionists. The integration of knowledge from different hierarchical levels of organization is an extremely ambitious and complicated task that requires many years of painstaking research. Nevertheless, the knowledge gained from such an endeavor will enable food manufacturers to design and produce higher-quality foods in a more cost-effective and systematic fashion. For this reason, the connection between molecular, colloidal, bulk physicochemical, biochemical, and sensory properties of food emulsions will be stressed throughout this book. 1.5 Understanding Food Emulsion Properties Food emulsions are compositionally and structurally complex materials (Section 1.2.5). It is therefore particularly challenging to understand their properties at a fundamental scientific level. Nevertheless, 22 Food Emulsions: Principles, Practices, and Techniques appreciable progress has been made over the past few decades due to the coordinated efforts of scientists working in industry, academia, and government laboratories. The purpose of this section is to give a general overview of the factors that influence the topics and directions of research on food emulsions and to highlight the general approaches that are used to provide information about food emulsion properties. 1.5.1 Factors Influencing Topics and Directions of Research As in other areas of science, progress in understanding the properties of food emulsions has not been uniform. Instead, certain aspects of the field have been the focus of intense study during a particular period, whereas other aspects have been largely ignored. A number of the most important factors that influence the choice of topics and directions of research on food emulsions are reviewed in this section. Relevance to industry and society: Certain research topics are perceived as being of higher commercial priority to the food industry or of greater relevance to government organizations at a particular time, and therefore attract greater financial and institutional support. The food industry usually supports research that improves manufacturing efficiency, reduces product costs, increases product stability, or leads to the development of new or improved products that will give it a market edge. Government agencies often support research that will improve the health or quality of life of its citizens, or that will increase the general efficiency or competitiveness of the food industry. Research scientists may find support for their research programs by working on topics that are already recognized as being of considerable commercial or societal value. Alternatively, they may identify a research topic that has previously been neglected, but which they believe is of importance to industry or society. They then have to convince other scientists and funding agencies that the research topic has scientific merit and is of sufficient importance to warrant funding. Availability of scientific personnel with the appropriate expertise: Certain topics are the focus of investigation because the research scientists working in the field have previously been trained in that particular area of expertise (e.g., they may have done graduate studies or postdoctoral research in a laboratory that specializes in this area). Conversely, other important topics are not studied because the current generation of scientists does not have the required expertise or conceptual framework to address them. This is particularly true in the study of food emulsions, which has grown rapidly to incorporate various aspects of many diverse scientific disciplines, including mathematics, physics, chemistry, biology, engineering, sensory science, psychology, and physiology. This has meant that it is often difficult for individual scientists to make appreciable advances in knowledge when working in isolation.* Instead, progress has become increasingly dependent on research being carried out by multidisciplinary teams consisting of scientists with different expertise, methodologies, and instrumentation. Indeed, the integration of knowledge from different disciplines is one of the dominant characteristics of many current research programs on food emulsions. Development of our understanding of food emulsion properties therefore depends on bringing together individuals with the diverse range of skills that can effectively work together as a team. Availability of appropriate theory and instrumentation: Certain topics are amenable to study because analytical instrumentation is already available that can be used to probe the characteristics of the system that are of interest and/or because appropriate physical theories are available to facilitate the design of experiments and the interpretation of results. On the other hand, other topics are not studied because they are too complicated to understand with the available analytical instrumentation or theoretical models. These topics may be of great industrial or societal importance, but knowledge of them cannot progress significantly until appropriate expertise, instrumentation, and theory are developed or adopted from another field. There are many examples of the rapid progress in the field of food emulsions resulting from the introduction of new theories or techniques. For example, the commercial availability of analytical instruments to rapidly and accurately measure the particle size distribution of emulsions meant that many experiments could be carried out that were not previously possible. An example of this phenomenon is the rapid development in the understanding of emulsion flavor and sensory perception that has * In reality, scientists never work in isolation since they always rely on the published work of earlier scientists. Context and Background 23 occurred during the past couple of decades (Chapter 9). A number of new analytical techniques became available that enabled researchers to measure the release of volatile components from foods on timescales relevant to mastication. This allowed researchers to design and perform experiments to systematically investigate the factors that influence flavor release in food emulsions, which in turn stimulated the development of physical theories to describe and predict flavor release. These theories made predictions that could be tested against experimental measurements and provided valuable new insights into the physicochemical and physiological basis of flavor release that were not possible earlier. More recently, there have been major advances in understanding the fate of food emulsions after ingestion due to the availability of simulated gastrointestinal tracts and analytical instrumentation to study this important phenomenon (Chapter 11). Access to previous knowledge: Another issue that has a strong influence on the topics and directions of research on food emulsions is accessibility to the knowledge accumulated by previous scientists working in similar or related areas. Developments in a particular field of scientific study are largely built upon the experimental and theoretical work that has been done previously. A great deal of research has been carried out on food emulsions, and much of this work has been published in scientific journals, conference proceedings, books, and online. It is therefore extremely important that researchers carry out extensive literature reviews of their field of interest, plus related fields, since this helps identify gaps in the current knowledge where research is needed, helps identify inconsistencies in previous studies, helps identify new techniques or theories that can be used, and helps avoid repetition of previous work. The number of scientific publications has increased enormously during the past few decades, which has made it increasingly difficult for scientists to be sure that they are thoroughly familiar with all of the previous research carried out in their field of interest. Nevertheless, advances in online information retrieval systems especially devoted to the scientific literature have certainly facilitated access to published research. Finally, it should be noted that a considerable amount of important research on food emulsions carried out by basic scientists working within food companies has not been published. 1.5.2 General Approaches Used to Study Food Emulsions Our understanding of the factors that determine the properties of food emulsions normally progresses through a synthesis of observation, experimentation, and theory development. Some of the most important general approaches that have been employed by scientists to increase the knowledge of emulsion properties are presented later. An appreciation of the advantages and limitations of each of these approaches may help investigators design and interpret experiments. It should be stressed that there is no single unified approach to making scientific advances. Instead, scientists have to use their experience, creativity, and imagination to select the most appropriate approach for the particular system studied. In addition, the success of a particular research program largely depends on the motivation, persistence, and commitment of the individuals involved. Trial-and-error approach: The trial-and-error approach is widely used in industry for solving manufacturing problems, as well as for developing new and improved products and processes. In this approach, the investigator usually prepares a sample for study that has certain characteristics (e.g., composition or microstructure) and then subjects it to some form of processing treatment (e.g., storage, heating, chilling, freezing, stirring, and pressure treatment). The properties of the sample are then measured, and the investigator establishes whether the sample characteristics or the processing treatment lead to a final product with desirable characteristics. If the final product meets these requirements, then the initial sample characteristics and/or processing treatments are selected, but if it does not meet these requirements, then the sample characteristics and/or processing treatments are changed and the procedure is repeated until a final product with suitable properties is obtained. The major advantage of the trial-and-error method is that it is sometimes possible to rapidly solve a problem or develop a product using minimal resources. For example, an investigator with some prior knowledge of a system may be able to rapidly select the optimum initial sample characteristics and processing treatments required to produce a desirable final product. The major disadvantages of this method are that it may not be possible 24 Food Emulsions: Principles, Practices, and Techniques to solve the problem in a reasonable time (if the wrong input values are selected), it may not produce a solution that is robust, or it may not produce the optimum solution to the problem. If a more rigorous study was carried out, it might have been possible to identify a solution that was more efficient, more robust, or less expensive. In addition, the trial-and-error method is largely dependent on the accumulated expertise of the investigator and provides little insight into the basic physicochemical processes that govern the properties of food emulsions. Black-box approach: The black-box approach is a more systematic means of obtaining information about the system being studied, but it is also not concerned with providing detailed information about the fundamental physicochemical processes occurring within the system. Instead, the system (the black box) is subjected to one or more treatments (inputs), and the change in one or more properties of the system in response to these treatments (outputs) is measured. The investigator then reports the measured system properties for each of the various treatments and/or uses a statistical model to correlate the system properties to the treatments. To provide a concrete example of this approach, the system could be a protein-stabilized O/W emulsion, the treatment could be a change in temperature, and the measured property could be the emulsion viscosity. The investigator would prepare an emulsion, subject samples of it to different temperature treatments, measure the viscosities of each of the samples, and then report the change in emulsion viscosity with temperature. This approach is particularly useful for identifying the major factors that influence the properties of food emulsions and their components, and in assessing their magnitude and relative importance. It is widely used as an initial screening procedure by investigators who are studying a system that has not been extensively studied before, since it enables them to rapidly develop an understanding of the dominant factors that influence its properties. In some situations, the black-box approach may be the only option available since the theoretical concepts or analytical techniques required to probe the internal operation of the system are not available. Nevertheless, the black-box approach has limited value when used improperly or when taken to extremes. For example, there are many examples of experiments on food emulsions that have been designed and analyzed primarily on the basis of sophisticated statistical models (such as surface response methodology), which have produced confusing and misleading results. Rather than using existing knowledge of the fundamental physicochemical properties of the system to design experiments or interpret data, investigators select (an often inappropriate) range of input and output variables based on some statistical model, and then find statistical correlations between the input and output variables. These statistical models can be useful in establishing the relative importance of different factors, but the author believes that they should always be used extremely carefully and in combination with the physicochemical approach described later whenever possible. Indeed, if an investigator truly had no knowledge of the behavior of the system being studied, it would be difficult to select the type of input parameters to vary, the range of input parameters to select, and the material properties to measure. In practice, investigators usually have a fairly good a priori expectation of the factors that are likely to be important, which facilitates the selection of the most appropriate input and output variables to use. Physicochemical approach: The main disadvantages of the trial-and-error and black-box approaches is that they do not provide any direct understanding of the fundamental physicochemical processes that occur within a system. Knowledge of these processes is usually desirable since it provides a deeper and more quantitative insight into the factors that determine the overall properties of the system, which greatly facilitates the identification of effective strategies for controlling the properties of the system and can allow one to make predictions about how the system (or related systems) will behave under different conditions. For this reason, many researchers utilize a more fundamental physicochemical approach to investigating emulsion properties. This approach attempts to understand at a fundamental physicochemical level why a system behaves in precisely the way it does when it is subjected to a particular treatment. This approach can be used in two broad ways depending on the starting point. First, an investigator could start with some empirical observation or experimental result and then attempt to establish its physicochemical origin using a range of analytical techniques, physical concepts, and mathematical models. Second, an investigator could start with a conceptual or theoretical model and then use it to make predictions about how a real system should behave. These predictions could then Context and Background 25 be compared with experimental measurements made on a real system, and the investigator could determine how well the model describes the properties of the real system. If there are deviations between the theoretical predictions and experimental measurements, then the mathematical model could either be discarded or it could be modified to take them into account. By comparing how closely theoretical predictions and experimental measurements agree, it is often possible to obtain quantitative insights into the physicochemical basis of food emulsion properties. In reality, these two different ways of using the physicochemical approach are closely related to each other, and investigators often use a combination of them both. The level of understanding that is achievable using the physicochemical approach is largely determined by the complexity of the system studied and the sophistication of the analytical instrumentation and theoretical models available. Reductionism–integrationist approach: The reductionism–integrationist approach has proved to be an extremely powerful means of advancing our understanding of food emulsion properties. Food emulsions are extremely complex systems, and many factors operate in concert to determine their overall properties. For this reason, experiments are usually carried out using simplified well-defined model systems that retain the essential features of the real system, but which ignore many of the secondary effects. For example, the emulsifying properties of proteins are often investigated by using an isolated individual protein, pure oil, and pure water. In reality, a protein ingredient used in the food industry consists of a mixture of different proteins, sugars, salts, fats, and minerals, and the oil and aqueous phases may contain a variety of different chemical constituents (Section 1.2.5). Nevertheless, by using a well-defined model system, it is possible to elucidate the primary factors that influence the properties of proteins in emulsions in a more quantitative fashion. Once these primary factors have been established, it is possible to increase the complexity of the model by introducing additional variables and systematically examining their influence on the overall properties. This incremental approach eventually leads to a thorough understanding of the factors that determine the properties of actual food emulsions and to the development of theories that can be used to describe and predict their behavior. Open-minded approach: Finally, another valuable means of advancing our understanding of food emulsions is to be receptive to the potential importance of unexpected results (Beveridge 1950). If an experiment is designed that consistently produces an unexpected result, it is important to be aware that the result may be due to poor experimental design or that it may be due to some interesting new phenomenon. Many of the most interesting discoveries in science are made by researchers who try to explain some result that did not correspond to their initial expectations. In these situations, it is usually important to carefully design further experiments that will provide evidence about the factors that influence the observed effect and about its physicochemical origin. 1.6 Overview and Philosophy It is impossible to cover every aspect of food emulsions in a book of this size. By necessity, one must be selective about the material presented and the style in which it is presented. Rather than reviewing the practical knowledge associated with each particular type of emulsion-based food product, I will focus primarily on the fundamental principles of emulsion science as applied to food systems because these principles are generally applicable to all types of food emulsion. Even so, I will use real food emulsions as examples where possible in order to emphasize the practical importance of the fundamental approach (particularly in Chapter 14). As mentioned earlier, I will pay particular attention to the relationship between molecular, colloidal, and bulk physicochemical properties of food emulsions because I believe that this approach leads to the most complete understanding of their behavior. Throughout this book, it will be necessary to introduce a number of theories that have been developed to describe the properties of emulsions. Rather than concentrating on the mathematical derivation of these theories, I will highlight their physical significance and focus on their relevance to food scientists. A feeling for the major factors that determine the properties of food emulsions can often be gained by programming these theories onto a personal computer and systematically examining the role that each physical parameter in the equation plays. 26 Food Emulsions: Principles, Practices, and Techniques Before ending this introductory chapter, I will give a brief overview of the subject matter of the remaining chapters in the book. In Chapters 2 and 3, I reviewed the various types of attractive and repulsive forces that can act on molecules and colloidal particles and discuss how these interactions influence the organization of the molecules or particles within a system. In Chapter 4, I reviewed the major functional ingredients that are utilized in food emulsions (e.g., oil, water, emulsifiers, weighting agents, thickening agents, gelling agents), with particular emphasis on understanding the physicochemical basis of their functional properties. In Chapter 5, I discussed the structure and characteristics of the interfacial region that separates bulk phases (e.g., oil and water) since this narrow region plays a crucial role in determining the overall properties of emulsions. In Chapter 6, I discussed the physicochemical principles underlying emulsion formation and review the various mechanical devices available for preparing emulsions for research and industrial applications. In Chapters 7 through 10, I discussed the molecular-colloidal basis of the bulk physicochemical properties of emulsions, that is, stability, rheology, appearance, and flavor. In Chapter 11, I focused on the gastrointestinal fate of food emulsions, which has become a major research area in the past decade due to the desire to create functional foods specifically designed to improve human health and wellness. In Chapter 12, I demonstrated the practical application of the principles of emulsion science and technology in the food industry by reviewing the formulation, formation, and physicochemical properties of three common types of food emulsion (i.e., dairy emulsions, beverages, and dressings). In Chapter 13, I discussed the structural design of emulsion-based delivery systems, as this topic has emerged as an important area for the delivery of nutraceuticals and other active agents in foods. Finally, in Chapter 14, I discussed the most important analytical techniques used to provide information about the composition, microstructure, and physicochemical properties of food emulsions. REFERENCES Beveridge, W. I. B. (1950). The Art of Scientific Investigation. New York: Vintage Books. Clausse, D. (2010). Differential thermal analysis, differential scanning calorimetry, and emulsions. Journal of Thermal Analysis and Calorimetry 101(3): 1071–1077. Clausse, D. and C. Dalmazzone (2014). Freezing within emulsions: Theoretical aspects and engineering applications. Oil & Gas Science and Technology-Revue D Ifp Energies Nouvelles 69(3): 415–434. Hiemenz, P. C. and R. Rajagopalan (1997). Principles of Colloid and Surface Chemistry. New York: Marcel Dekker. Hunter, R. J. (1986). Foundations of Colloid Science: Volume 1. Oxford, U.K.: Oxford Science Publications. Kim, H. J., E. A. Decker, and D. J. McClements (2002). Role of postadsorption conformation changes of betalactoglobulin on its ability to stabilize oil droplets against flocculation during heating at neutral pH. Langmuir 18(20): 7577–7583. Landy, P., J. L. Courthaudon, C. Dubois, and A. Voilley (1996). Effect of interface in model food emulsions on the volatility of aroma compounds. Journal of Agricultural and Food Chemistry 44(2): 526–530. McClements, D. J. (2012). Advances in fabrication of emulsions with enhanced functionality using structural design principles. Current Opinion in Colloid & Interface Science 17(5): 235–245. McClements, D. J. and E. A. Decker (2000). Lipid oxidation in oil-in-water emulsions: Impact of molecular environment on chemical reactions in heterogeneous food systems. Journal of Food Science 65(8): 1270–1282. McClements, D. J. and Y. Li (2010). Structured emulsion-based delivery systems: Controlling the digestion and release of lipophilic food components. Advances in Colloid and Interface Science 159(2): 213–228. Mei, L. Y., E. A. Decker, and D. J. McClements (1998). Evidence of iron association with emulsion droplets and its impact on lipid oxidation. Journal of Agricultural and Food Chemistry 46(12): 5072–5077. Mei, L. Y., D. J. McClements and E. A. Decker (1999). Lipid oxidation in emulsions as affected by charge status of antioxidants and emulsion droplets. Journal of Agricultural and Food Chemistry 47(6): 2267–2273. Mezzenga, R., P. Schurtenberger, A. Burbidge and M. Michel (2005). Understanding foods as soft materials. Nature Materials 4(10): 729–740. Orr, C. (1988). Determination of particle size. In Encyclopedia of Emulsion Technology, P. Becher, ed., vol. 3, pp. 137–169. New York: Marcker Dekker. Context and Background 27 Pashley, R. M. (2003). Effect of degassing on the formation and stability of surfactant-free emulsions and fine teflon dispersions. Journal of Physical Chemistry B 107(7): 1714–1720. Rawle, A. (2004). Basic Principles of Particle Size Analysis. Technical Paper, Malvern, Worcs, U.K.: Malvern Instruments. Ubbink, J., A. Burbidge and R. Mezzenga (2008). Food structure and functionality: A soft matter perspective. Soft Matter 4(8): 1569–1581. Velikov, K. P. and E. Pelan (2008). Colloidal delivery systems for micronutrients and nutraceuticals. Soft Matter 4(10): 1964–1980. Walstra, P. (2003). Physical Chemistry of Foods. New York: Marcel Decker. 2 Molecular Characteristics 2.1 Introduction The rational design and fabrication of an emulsion-based food product with specific physicochemical and sensory properties relies on understanding the characteristics of the different kinds of molecules* within the product, as well as the nature of the interactions between them (Bishop et al. 2009, Israelachvili 2011). The various molecules within an emulsion may form a variety of different structures depending on their interactions with each other (Figure 2.1). Molecules may be part of a bulk phase (which may be solid or liquid) where they are surrounded by similar types of molecule, for example, oil molecules in an oil phase. Molecules may be part of a regular solution where they are surrounded by a mixture of similar and dissimilar molecules that are randomly organized, for example, sugar molecules dissolved in water. Molecules may be part of an ordered solution where they are preferentially surrounded by dissimilar molecules that tend to have a specific organization, such as mineral ions dissolved in water that are surrounded by a shell of highly organized water molecules. Molecules may accumulate at an interface between two phases, such as surfactants at an oil–water interface. Molecules may associate with similar or dissimilar molecules and form molecular clusters dispersed within a bulk phase, such as surfactant micelles in water. Molecules may be incorporated into a 3D network that extends throughout the system and gives it some solid-like characteristics, such as gelatin molecules within a hydrogel. Finally, molecules may be part of complex biological structures, such as phospholipid molecules within cell membranes. The overall physicochemical properties of food emulsions ultimately depend on the nature, properties, and interactions of the structures formed by the molecules they contain, for example, separate phases, interfaces, clusters, and networks. The structural organization of a particular set of molecules is determined by the forces that act between them, as well as the environmental conditions, such as temperature and pressure. From a thermodynamic viewpoint, a certain arrangement of molecules may have the lowest free energy since it is the best comprise that maximizes the number of favorable interactions, minimizes the number of unfavorable interactions, and maximizes the various entropy contributions of the system. Nevertheless, foods are rarely in their most thermodynamically stable state and, therefore, the structural organization of the molecules is often governed by kinetic factors that prevent them from reaching the arrangement with the lowest free energy (Section 1.2.1). For this reason, the structural organization of the molecules in foods is often determined by their previous history, that is, the temperatures, pressures, gravitational forces, and applied mechanical forces that they experienced during their lifetime. To understand, predict, and control the behavior of food emulsions, it is important to be aware of the origin and nature of the forces responsible for holding the molecules together, the factors that impact these forces, and how these forces lead to the various types of structures formed in food emulsions. Only then will it be possible to rationally design foods that have internal structures that are known to be advantageous to food quality. The purpose of this chapter is to give a brief overview of the major types of molecular forces and entropy effects important in materials, and to show how these factors influence the conformation and structural organization of molecules. * The term “molecule” is used broadly to refer to molecular species, such as atoms, molecules, and ions. 29 30 Food Emulsions: Principles, Practices, and Techniques (a) (b) (c) (d) (e) (f) (g) (h) FIGURE 2.1 The molecules in food emulsions may adopt a variety of different structural arrangements depending on the nature of their interactions with their neighbors. (a) Bulk phase (solid), (b) regular solution, (c) ordered solution, (d) immiscible phases, (e) bulk phase (liquid), (f) molecular aggregate, (g) adsorption to interface, and (h) molecular network. 2.2 Forces of Nature There are four distinct types of force in nature: strong nuclear interactions, weak nuclear interactions, electromagnetic interactions, and gravity (Israelachvili 2011, Atkins and de Paula 2014). The strong and weak nuclear forces act over extremely short distances and are chiefly responsible for holding together subatomic particles in the nucleus. As nuclear rearrangements are not usually important for understanding the properties of food emulsions, these forces will not be considered further. Gravitational forces are relatively weak and act over large distances compared to other types of forces. Their strength is proportional to the product of the masses of the objects involved, and consequently they are insignificant at the molecular level because molecular masses are extremely small. Nevertheless, they do affect the behavior of food emulsions at the macroscopic level, for example, sedimentation or creaming of droplets, the shape adopted by large droplets, meniscus formation, and capillary rise (Israelachvili 2011). For this reason, gravitational forces will only be considered later in the book when we turn to the behavior of larger objects. The dominant forces that act at the molecular level are all electromagnetic in origin and can conveniently be divided into four types: covalent, electrostatic, van der Waals, and steric overlap (Israelachvili 2011, Atkins and de Paula 2014). Despite acting over extremely short distances, often on the order of a few angstroms or less, intermolecular forces are ultimately responsible for the structural organization of food ingredients, and therefore the bulk physicochemical and sensory properties of emulsions and other foods. 2.3 Origin and Nature of Molecular Interactions 2.3.1 Covalent Interactions Covalent bonds involve the sharing of outer shell electrons between two or more atoms, so that the individual atoms lose their discrete nature (Israelachvili 2011, Atkins and de Paula 2014). The number of electrons in the outer shell of an atom governs its valency, which determines the optimum number of covalent bonds that it forms with other atoms. Covalent bonds may be saturated or unsaturated depending on the number of electrons involved. Unsaturated bonds tend to be shorter, stronger, and more rigid than saturated ones. The distribution of electrons within a covalent bond determines its Molecular Characteristics 31 polarity. When the electrons are shared equally between the atoms, the bond has a nonpolar character, but when the electrons are shared unequally, the bond has a polar character. The polarity of a molecule depends on the symmetry of the various covalent bonds it contains (see Section 2.3.2). Covalent bonds are also characterized by their directionality, that is, their tendency to be located at clearly defined angles relative to each other. The valency, saturation, polarity, strength, and directionality of covalent bonds determine the 3D structure, flexibility, chemical reactivity, and physical interactions of molecules. Chemical reactions involve the breaking and formation of covalent bonds (Atkins and de Paula 2014). The bulk physicochemical and organoleptic properties of food emulsions are altered by various types of chemical and biochemical reactions that occur during their production, storage, and consumption, such as oxidation, reduction, hydrolysis, and polymerization (Belitz et al. 2009, Brady 2013). Some of these reactions are beneficial to food quality, while others are detrimental. It is therefore important for food scientists to be aware of the various types of chemical reaction that occur in food emulsions, and to establish their influence on the overall properties of the system. The chemical reactions that occur in food emulsions are similar to those that occur in any other multicomponent heterogeneous food material, for example, oxidation of lipids and proteins, hydrolysis of proteins or polysaccharides, cross-linking of proteins, and Maillard reactions between reducing sugars and free amino groups (Belitz et al. 2009). Nevertheless, the rates and pathways of these reactions are often influenced by the physical environment of the molecules involved, for example, whether they are located in the oil, water, or interfacial region (McClements and Decker 2000). Until fairly recently, emulsion scientists were principally concerned with understanding the physical changes that occur in food emulsions, rather than the chemical changes. Nevertheless, there is growing interest in establishing the relationship between emulsion properties and the mechanisms of various chemical reactions that occur within them (McClements and Decker 2000, Boon et al. 2010, Choi et al. 2010, Waraho et al. 2011). Despite the importance of chemical reactions on emulsion quality, it should be stressed that many of the most important changes in emulsion properties are a result of alterations in the spatial distribution of the molecules, rather than of alterations in their chemical structure, for example, creaming, flocculation, coalescence, Ostwald ripening, and phase inversion (Chapter 7). The spatial organization of molecules is governed principally by their noncovalent (physical) interactions with their neighbors, for example, electrostatic, van der Waals, and steric overlap. It is therefore particularly important to have a good understanding of the origin and nature of these interactions. 2.3.2 Electrostatic Interactions Electrostatic interactions occur between molecular species that possess a permanent electrical charge, such as ions and polar molecules (Israelachvili 2011). An ion is an atom or molecule that has either lost or gained one or more outer shell electrons so that it obtains a permanent positive or negative charge (Atkins and de Paula 2014). A polar molecule has no net charge (i.e., as a whole the molecule is neutral), but it does have an electrical dipole because of an uneven distribution of the charges within it. Certain atoms are able to pull the electrons in the covalent bonds toward them more strongly than other atoms (Atkins and de Paula 2014). As a consequence, they acquire a partial negative charge (δ−), and the other atom acquires a partial positive charge (δ+). If the partial charges within a molecule are distributed symmetrically, they cancel each other and the molecule has no dipole (e.g., CCl4), but if they are distributed asymmetrically the molecule will have a dipole (Israelachvili 2011). For example, the chlorine atom in HCl pulls the electrons in the covalent bond more strongly than the hydrogen atom, and so a dipole is formed: Hδ+Clδ−. The strength of a dipole is characterized by the dipole moment μ = ql, where l is the distance between two charges q+ and q−. The greater the magnitude of the partial charges, or the further they are apart, the greater the dipole moment of a molecule. The interaction between two molecular species is characterized by an intermolecular pair potential, w(s), which is the energy required to bring two molecules from an infinite distance apart to a center-to-center separation s (Israelachvili 2011). There are a number of different types of electrostatic interaction that can occur between permanently charged molecular species (ion–ion, ion–dipole, and 32 Food Emulsions: Principles, Practices, and Techniques dipole–dipole) (Figure 2.2), but they can all be described by a similar general equation (Hiemenz and Rajagopalan 1997): wE (s) = Q1Q2 4pe0e R s n (2.1) where Q1 and Q2 are the effective charges on the two species ε0 is the dielectric constant of a vacuum (8.85 × 10 −12 C2 J−1 m−1) εR is the relative dielectric constant of the intervening medium s is the center-to-center distance between the charges n is an integer that depends on the nature of the interaction For ions, the value of Q is determined by their valency z and electrical charge e (1.602 × 10 −19 C), whereas for dipoles, it is determined by their dipole moment μ and orientation (Table 2.1). Numerical calculations of the intermolecular pair potential for representative ion–ion, ion–dipole, and dipole–dipole interactions are illustrated in Figure 2.3a. These interactions are representative of molecular interactions between sodium and chloride ions (Na+–Cl−), sodium ions and water molecules (Na+–H2O), and two water molecules (H2O–H2O), respectively. The fact that the magnitude and range of the ion–dipole interactions is much stronger than the dipole–dipole interactions accounts for the fact that addition of salt ions to aqueous phases causes large changes in the structural organization of the water molecules around the ions (Chapter 4). + – + δ– δ+ Ion–ion δ– δ+ Ion–dipole δ– δ+ Dipole–dipole FIGURE 2.2 Schematic representation of the most important types of intermolecular electrostatic interaction that arise between molecules. TABLE 2.1 Parameters Needed to Calculate the Interaction Pair Potential for Ion–Ion, Ion–Dipole, and Dipole–Dipole Electrostatic Interactions Using Equation 2.1 (See Also Figure 2.3a) Interaction Type Example Q1 Q 2 n Ion–Ion Ion–Dipole Dipole–Dipole Na Cl Na+ H2O H2O H2O (z1e) (z2e) (z1e) μ2 cosϕ μ1μ2  f(ϕ) 1 2 3 + − Source: Data from Hiemenz, P.C. and Rajagopalan, R., Principles of Colloid and Surface Chemistry, Marcel Dekker, New York, 1997. Note: Here z is the valence, μ is the dipole moment, e is the electronic charge, and ϕ is the angle between the dipole charges. 33 Molecular Characteristics 0 Dipole-dipole δ– δ+ 0 δ– δ+ CH4–CH4 –50 –2 Ion-ion + – w(s)/kT w(s)/kT Ion-dipole + δ– δ+ H2O–CH4 –4 H2O–H2O –6 –100 s –8 –150 (a) 0 1 2 3 4 5 s (nm) 50 45 Soft shell –10 (b) 0.2 0.3 0.4 0.5 0.6 kT Hard shell 40 35 w(s)/kT 30 25 20 15 10 5 0 (c) 0.2 0.3 0.4 s (nm) 0.5 FIGURE 2.3 Dependence of the intermolecular pair potential on intermolecular separation for (a) electrostatic, (b) van der Waals, and (c) steric overlap interactions. Equation 2.1 is based on the assumption that the medium separating the two charged species is isotropic and homogeneous. In reality, the charged species in foods are usually separated by one or more type of solvent molecules that have discrete sizes, shapes, and properties. When the separation between the charged species is relatively large compared to the size of the solvent molecules the assumption that the intervening medium is a continuum is a reasonably good one. On the other hand, when the separation between the charged species is relatively small, this assumption breaks down and a more sophisticated analysis of the electrostatic interactions between the charged species is required (Israelachvili 2011). This analysis must take into account the size, shape, orientation, and interactions of all the molecules involved. Examination of Equation 2.1 and Figure 2.3a provides a number of valuable insights into the nature of intermolecular electrostatic interactions, and the factors that influence them: 1. The sign of the interaction may be either positive (repulsive) or negative (attractive) depending on the signs of charged molecules involved. If the charges have similar signs, w E(s) is positive and the interaction is repulsive, but if they have opposite signs, w E(s) is negative and the interaction is attractive. 34 Food Emulsions: Principles, Practices, and Techniques 2. The strength of the interaction depends on the magnitudes of the charges involved, that is, Q1 and Q2. Thus, ion–ion interactions are stronger than ion–dipole interactions, which are in turn stronger than dipole–dipole interactions (Figure 2.3a). In addition, the strength of interactions involving ions increases as their valency increases, whereas the strength of interactions involving polar species increases as their dipole moment increases. Thus, multivalent ions (such as Ca2+) will tend to interact more strongly with other charged species than similarly sized monovalent ions (such as K+). 3. The strength of the interaction increases as the center-to-center separation of the charged species decreases (Figure 2.3a). Thus, interactions between small ions or molecules (which can get closer together) are stronger than those between large ions or molecules with the same charge. 4. The strength of the interaction depends on the nature of the material separating the charges: the higher the relative dielectric constant (εR), the weaker the electrostatic interaction (Equation 2.1). Electrostatic interactions between two charged species in water (εR = 80) are therefore much weaker than those between the same two charged species in oil (εR = 2). This phenomenon accounts for the much higher solubility of salts in water than in nonpolar solvents and is also important for predicting the strength of the electrostatic interactions that occur within the hydrophobic core of globular proteins (Israelachvili 2011). 5. The strength of the interaction depends on the orientation of any dipoles involved, being strongest when partial charges of opposite sign are brought close together. When the electrostatic interaction between a dipole and another charged species is much stronger than the thermal energy (Section 2.5), the dipole becomes permanently aligned so as to maximize the strength of the attraction. This alignment of dipoles is responsible for the high degree of structural organization of molecules in bulk water and the ordering of water molecules around ions in aqueous solutions (Chapter 4). 6. The range of the interaction depends on the nature of the charged molecular species involved, with ion–ion (1/s) interactions being longer range than ion–dipole interactions (1/s2), which are in turn longer range than dipole–dipole interactions (1/s3). 2.3.3 Van der Waals Interactions Van der Waals forces act between all types of molecular species, whether they are ionic, polar, or nonpolar (Hiemenz and Rajagopalan 1997, Israelachvili 2011). They are conveniently divided into three separate contributions, which all rely on the polarization of molecules (Figure 2.4): δ– δ– δ+ δ– δ+ δ+ Dispersion δ+ δ– δ– δ+ δ– δ+ Induction Orientation FIGURE 2.4 Schematic representation of van der Waals intermolecular interactions that involve either the electronic or orientational polarization of molecules. 35 Molecular Characteristics Dispersion forces: These forces arise from the interaction between an instantaneous dipole and a dipole induced in a neighboring molecule due to the presence of the instantaneous dipole. The electrons in a molecule are continually moving around the nucleus. At any instant in time there is an uneven distribution of the negatively charged electrons around the positively charged nucleus, and so an instantaneous dipole is formed. This instantaneous dipole generates an electrical field that induces a dipole in a neighboring molecule. Consequently, there is an instantaneous attractive force between the two dipoles. The attraction between the molecules is therefore finite, even though the net charge on the individual molecules involved in the interaction is zero when averaged over time. Induction forces: These forces arise from the interaction between a permanent dipole and a dipole induced in a neighboring molecule by the presence of the permanent dipole.* A permanent dipole causes an alteration in the distribution of electrons of a neighboring molecule that leads to the formation of an induced dipole. The interaction between the permanent dipole and the induced dipole leads to an attractive force between the molecules. Orientation forces: These forces arise from the interaction between two permanent dipoles that are continuously rotating. On average, each individual rotating dipole has no net charge, but there is still a weak attractive force between different dipoles because the movement of one dipole induces some correlation in the movement of a neighboring dipole. When the interaction between the two dipoles is strong enough to cause them to be permanently aligned, this contribution is replaced by the electrostatic dipole–dipole interaction described in the previous section. As will be seen in Chapter 3 on the colloidal interactions between emulsion droplets, an understanding of the origin of these three contributions to the van der Waals interaction has important consequences for predicting the stability of emulsions to aggregation (Section 3.3). The overall intermolecular pair potential due to van der Waals interactions is given by wVDW (s) = - ( Cdisp + Cind + Corient ) ( 4pe0eR ) 2 s6 (2.2) where Cdisp, Cind, and Corient are positive constants that depend on the dispersion, induction, and orientation contributions, respectively (Hiemenz and Rajagopalan 1997). Their magnitude depends on the dipole moment (for permanent dipoles) and the polarizability (for induced dipoles) of the molecules involved in the interaction (Table 2.2). The polarizability is a measure of the strength of the dipole induced in a molecule when it is in the presence of an electrical field: the larger the polarizability the easier it is to induce a dipole in a molecule. For most biological molecules, the dominant contribution to the van der Waals interaction is the dispersion force, with the important exception of water where the major contribution is from the orientation force (Israelachvili 2011). Examination of Equation 2.2 and Figure 2.3b provides some useful physical insights into the factors that influence the van der Waals interactions between two molecules: 1. The sign of the interaction is always negative (attractive) because the values of Cdisp, Cind, and Corient are always positive. 2. The strength of the interaction increases as the polarizability and dipole moment of the molecules involved increases. 3. The strength of the attraction decreases as the dielectric constant of the intervening medium increases, which highlights the electromagnetic origin of van der Waals interactions. 4. The range of the interaction is relatively short, decreasing rapidly with increasing intermolecular separation (1/s6). * Ions may also induce dipoles in neighboring molecules, but this ion polarization interaction has a 1/s4 dependence on intermolecular separation and is therefore not usually considered as a van der Waals interaction. 36 Food Emulsions: Principles, Practices, and Techniques TABLE 2.2 Compilation of Molecular Properties of Some Common Liquids and Solutes Needed to Calculate Intermolecular Interactions Taken from the Literature Static relative dielectric constants εR Water 78.5 Chloroform Ethylene glycol 40.7 Edible oils Methanol 32.6 Carbon tetrachloride Ethanol 24.3 Liquid paraffin Acetone 20.7 Dodecane Propanol 20.2 Hexane Acetic acid 6.2 Air Molecular diameters, polarizabilities, and dipole moments Molecule Type H2O CH4 HCl CH3Cl CCl4 NH3 Methanol Ethanol Acetone Benzene 4.8 2.5 2.2 2.2 2.0 1.9 1.0 σ (nm) α/4πε 0 (×10 −30 m3) μ (Da) 0.28 0.40 0.36 0.43 0.55 0.36 0.42 1.48 2.60 2.63 4.56 10.5 2.26 3.2 5.2 6.4 10.4 1.85 0 1.08 1.87 0 1.47 1.69 1.69 2.85 0 b b 0.53 Source: Data from Buffler, C.R., Advances in dielectric characterization of foods, in Characterization of Food: Emerging Methods, Gaonkar, A.G., ed., Elsevier, Amsterdam, the Netherlands, Chapter 10, 1995; Israelachvili, J., Intermolecular and Surface Forces, 3rd edn., Academic Press, London, U.K. a D = 3.336 × 10−30 C m. b Cannot be treated as spheres. Although van der Waals interactions act between all types of molecular species, they are considerably weaker than electrostatic interactions (Figure 2.3 and Table 2.3). For this reason, they are most important in determining interactions between nonpolar molecules, where electrostatic interactions do not make a significant contribution. Indeed, the structure and physicochemical properties of organic liquids is largely governed by the van der Waals interactions between the molecules (Israelachvili 2011). 2.3.4 Steric Overlap Interactions When two atoms or molecules come so close together that their electron clouds overlap, there is an extremely large repulsive force generated between them (Figure 2.3c). This steric overlap force is very short range and increases rapidly when the separation between the two molecules becomes less than the sum of their radii (σ = r1 + r 2). A number of empirical equations have been derived to describe the dependence of the steric overlap intermolecular pair potential, wsteric(s) on molecular separation (Israelachvili 2011). The hard-shell model assumes that the repulsive interaction is zero when the separation is greater than σ, but infinitely large when it is less than σ: æsö wsteric (s) = ç ÷ èsø ¥ (2.3) 37 Molecular Characteristics TABLE 2.3 Approximate Bond Strengths for Some of the Most Important Types of Molecular Interaction That Occur in Foods at Room Temperature In Vacuum Interaction Type Covalent Bonds C−O C−C C−H O−H C=C C≡N w(s*) (kJ mol ) −1 In Water w(s*) (RT) w(s*) (kJ mol−1) w(s*) RT 340 360 430 460 600 870 140 140 170 180 240 350 Electrostatic Ion–Ion Na+ Cl− Mg2+ Cl− Al3+ Cl− 500 1100 1800 200 460 730 6.3 14.1 22.5 2.5 5.7 9.1 Ion–Dipole Na+ H2O Mg2+ H2O Al3+ H2O 97 255 445 39 103 180 1.2 3.2 5.6 0.5 1.3 2.3 Dipole–Dipole H2O H2O 38 15 0.5 0.2 Ion Polarization Na+ CH4 24 10 Van der Waals CH4 CH4 C6H14 C6H14 C12H26 C12H26 C18H38 C18H38 CH4 H2O H2O H2O 1.5 7.4 14.3 21.2 2.6 17.3 0.60 3.0 5.7 6.1 0.7 6.9 Dipole interactions were calculated assuming that the molecules were aligned to get maximum attraction. Van der Waals forces were calculated from Israelachvili (2011) assuming that w(s*) was approximately equal to the cohesive energy divided by 6. In reality, molecules are slightly compressible and so the increase in the steric overlap repulsion is not as dramatic as indicated by Equation 2.3. The slight compressibility of molecules is accounted for by a soft-shell model, such as the power-law model: 12 æsö wsteric (s) = ç ÷ èsø (2.4) At separations greater than σ, the steric overlap repulsion is negligible, but at separations less than this value, there is a steep increase in the interaction pair potential, which means that the molecules strongly repel one another. The strong repulsion that arises from steric overlap determines the effective size and shape of atoms and molecules and determines how closely they can come together. It therefore has a strong influence on the packing of molecules in liquids and solids. 38 Food Emulsions: Principles, Practices, and Techniques 2.4 Overall Intermolecular Pair Potential We are now in a position to calculate the overall interaction between a pair of molecules. Assuming that no chemical reactions occur between the molecules, the overall intermolecular pair potential is the sum of the various physical interactions mentioned earlier: w(s) = w E(s) + wVDV(s) + wsteric(s) (2.5) The magnitude of each of the individual contributions to the overall interaction potential is strongest at close separations and decreases as the molecules move apart. Nevertheless, the overall intermolecular pair potential has a more complex dependence on separation, which may be attractive at some separations and repulsive at others, because it is the sum of a number of interactions that each have different magnitudes, ranges, and signs. 2.4.1 Lennard–Jones Potential: Understanding Bond Strengths and Lengths To highlight some of the most important features of intermolecular interactions, it is useful to consider the interaction of a pair of spherical nonpolar molecules (i.e., no electrostatic interactions). The overall intermolecular pair potential for this type of system is given by an expression known as the Lennard– Jones potential (Norde 2011): w(r ) = -A B + s 6 s12 (2.6) where the A-term represents the contribution from the van der Waals interactions (Equation 2.2) and the B-term represents the contribution from the steric overlap interaction (Equation 2.4). The dependence of the intermolecular pair potential on center-to-center separation is illustrated in Figure 2.5. The van der Waals interactions are attractive at all separations, whereas the steric overlap interactions are repulsive. At large separations, w(s) is so small that there is no effective interaction between the molecules. As the molecules are brought closer together, the pair potential becomes increasingly attractive (negative) because the van der Waals interactions dominate. Eventually, the molecules get so close together that their electron clouds overlap and the pair potential becomes strongly repulsive (positive) because steric overlap interactions dominate. Consequently, there is a minimum in the overall intermolecular pair potential at some intermediate separation, s*. Two molecules will tend to remain associated in this potential energy minimum in the absence of any disruptive influences (such as thermal energy or applied external forces), with a bond-length of s* and a bond strength of w(s*). 2.4.2 Thermal Energy: Judging Bond Strengths The molecules in a substance are in continual motion (translational, rotational, and vibrational) because of their thermal energy, kT (Israelachvili 2011, Atkins and de Paula 2014). Here k is Boltzmann’s constant and T is the absolute temperature. The thermally induced movement of molecules has a disorganizing influence, which opposes the formation of intermolecular bonds. For this reason, the strength of intermolecular interactions is usually measured relative to the thermal energy: kT ≈ 4.1 × 10 −24 kJ per bond or RT ≈ 2.5 kJ mol−1. If the bond strength is sufficiently greater than kT, the molecules will remain together, but if the bond strength is sufficiently smaller than kT, the molecules will tend to move apart because of the disorganizing influence of the thermal energy. At intermediate bond strengths, the molecules spend part of their time together and part of their time apart, that is, bonds are rapidly breaking and reforming. The bond strengths of a number of important types of intermolecular interaction are summarized in Table 2.3. In a vacuum, the strength of these bonds decreases in the following order: ion–ion, covalent > ion–dipole > dipole–dipole > van der Waals (Israelachvili 2011). With the exception of methane (a small nonpolar molecule), the bonds between the molecules shown in Table 2.3 are sufficiently 39 Molecular Characteristics 3 Repulsion 2.5 2 1.5 B/s12 w(s)/kT 1 Total 0.5 0 –0.5 s* 0.2 0.3 w(s*) 0.4 0.5 0.6 0.7 –1 –1.5 –2 –A/s6 Attraction s (nm) FIGURE 2.5 Intermolecular pair potential for a pair of spherical nonpolar molecules. The curves were calculated assuming typical values for the constants: A = 10 −77 J m6 and B = 10 −134 J m12. strong (compared to the thermal energy) to hold them together in a liquid or solid at room temperature. The strength of electrostatic and van der Waals interactions decreases appreciably when the molecules are surrounded by a solvent rather than a vacuum, especially when the solvent has a high dielectric constant, for example, water (Israelachvili 2011). When solute molecules are relatively large and sufficiently far apart (compared to the size of the solvent molecules), then the solvent can often be treated as a continuum with well-defined physicochemical properties, for example, relative dielectric constant. On the other hand, when solute molecules are relatively small and in close proximity, it is usually necessary to take into account the dimensions, locations, and interactions of both the solute and solvent molecules on the overall interaction potential (see Section 2.6.4). 2.4.3 Converting Potential Energies into Forces It is often more convenient to describe the interaction between a pair of molecules in terms of forces rather than potential energies (Israelachvili 2011). The force acting between two molecules can simply be calculated from the intermolecular pair potential using the following relationship: F(s) = −dw(s)/ds. The minimum in the potential energy curve therefore occurs at a separation where the net force acting between the molecules is zero, that is, the attractive and repulsive forces exactly balance. If the molecules move closer together, they experience a repulsive force, and if they move further apart, they experience an attractive force. 2.5 Molecular Structure and Organization Is Determined by a Balance of Interaction Energies and Entropy Effects In bulk materials, such as food emulsions, we are concerned with huge numbers of molecules, rather than with a pair of isolated molecules in a vacuum. The overall structure and organization of molecules within a molecular ensemble depends on the interactions of each molecule with all of its neighbors (which 40 Food Emulsions: Principles, Practices, and Techniques may be similar or dissimilar) and with various entropy effects (Ninhan and Nostro 2010, Israelachvili 2011). One of the most powerful means of understanding the relationship between molecular structure, interactions, and organization in molecular ensembles is to use statistical thermodynamic techniques in combination with computational methods (Palma et al. 2012). A molecular ensemble tends to organize itself so that the molecules are in an arrangement that minimizes the free energy of the system. The overall Gibbs free energy of a molecular ensemble is governed by both enthalpy and entropy contributions. The enthalpy contributions are determined by the molecular interaction energies discussed earlier in this chapter, while the entropy contributions are determined by the tendency of a system to adopt its most disordered state. 2.5.1 Forms of Entropy A variety of different molecular phenomenon may contribute to the overall entropy of a molecular ensemble (Walstra 2003): Translational entropy: The translational entropy is determined by the freedom of the molecules to move from one location to another. If the molecules are completely free to move throughout the system, then they have a high translational entropy, but if their movement is restricted in some way, then they have a lower translational entropy, for example, due to phase separation, adsorption to a surface, formation of molecular clusters, or adopting a crystalline arrangement (Figure 2.6). Rotational entropy: The rotational entropy is related to the number of different angular positions that an anisometric molecule can adopt. If the molecules are free to rotate at any angle, then they have high rotational entropy, but if their rotation is restricted in one or more directions, they have lower rotational entropy, for example, due to adsorption to an interface. Conformational entropy: The conformational entropy is determined by the number of different conformations that a molecule can adopt. If a molecule can adopt many different conformations, then it has high entropy (e.g., a flexible random coil molecule), but if the number of conformations it can adopt is restricted, then it has a low entropy (e.g., a globular or rodlike conformation). Mixing entropy: The mixing entropy is determined by the number of different ways that two or more different kinds of molecules can adopt in a given volume. When the different kinds of molecules are (a) (b) (c) (d) (e) (f) FIGURE 2.6 Examples of physicochemical phenomenon that involve changes in the molecular organization of the system. (a) Disordered (High S), (b) self-assembly (Low S),(c) ordered (Low S), (d) network formation (Low S), (e) adsorption (Low S), and (f) phase-separation (Low S). 41 Molecular Characteristics randomly distributed throughout the volume, the system has the highest entropy, but when one type of molecule is confined to one region and the other type of molecule is confined to another region, then the system has a lower entropy. There are many physicochemical processes that occur in food emulsions that involve one or more of the entropy changes mentioned earlier, for example, mixing, self-association, binding, adsorption, solvent structuring, helix–coil transitions, and protein denaturation. A number of these physicochemical processes will be encountered later in this book when we discuss the functional properties of various kinds of food ingredients, for example, water, lipids, proteins, polysaccharides, surfactants, and minerals. 2.5.2 Physicochemical Basis of Molecular Transitions An understanding of the molecular basis for the organization of molecules within a particular system is usually obtained by comparing the strength of the molecular interactions and entropy contributions in that system to those in an appropriate reference system. Some examples of transitions between different spatial arrangements of molecules that are important in food emulsions are listed later (Figures 2.6 and 2.7): • Mixing: Will a given collection of molecules form an intimate mixture of randomly distributed molecules or will it separate into two or more phases? • Self-association: Will solute molecules dispersed in a solvent exist as individual molecules or as molecular clusters? • Ordering: Will the molecules in a given system arrange themselves into an ordered structure (e.g., a crystalline solid) or will they be randomly distributed throughout the system (e.g., a simple liquid)? • Binding: Will solute molecules dispersed in a biopolymer solution exist as unbound molecules or will they bind to the biopolymer molecules? • Adsorption: Will solute molecules exist as individual molecules dispersed throughout the solvent or will they adsorb to a surface? • Conformation: Will a biopolymer molecule dispersed in a solvent adopt a random coil or helix conformation? In general, these different kinds of physiochemical process can be represented in terms of an equilibrium between two states with different molecular characteristics: State (1) ↔ State (2) (2.7) The transition from one state to another is accompanied by a change in the free energy of the system: ΔG tr = ΔHtr − TΔStr (2.8) Mixing Phase separation Immiscible liquids Regular solution FIGURE 2.7 System in which two types of molecules may be completely miscible or form a regular solution depending on the strength of the interactions between them and the entropy of mixing. 42 Food Emulsions: Principles, Practices, and Techniques where ΔG tr, ΔHtr, and ΔStr are the free energy, enthalpy, and entropy changes associated with the transition, respectively. When considering molecular interactions, it is often convenient to replace the enthalpy term (and some of the entropy term) with a molecular interaction energy (ΔEtr), which contains both enthalpy and entropy contributions: ΔG tr = ΔEtr − TΔStr (2.9) This is because many important types of molecular interactions are not purely enthalpic in origin, such as hydrogen bonds and hydrophobic interactions (see Section 2.8). If ΔG tr is negative, the transition is thermodynamically favorable; if ΔG tr is positive, the transition is thermodynamically unfavorable; and if ΔG tr ≈ 0, the transition is thermodynamically neutral. The free energy change associated with a transition can often be related to the molecular characteristics of the system using an appropriate physical model to calculate the change in interaction energies and entropy contributions that occur on going from one state to the other. The relative sign and magnitude of these contributions depends on the nature of the transition and on the type of molecules involved. In certain cases, it is possible to measure the enthalpy and entropy changes associated with specific transitions, for example, using calorimetry methods. In general, the interaction energy (E) of a particular arrangement of molecules can be determined by calculating the sum of all of the different types of interactions involved: E= ån w i (2.10) i i where ni is the number of interactions of strength wi. The change in interaction energies associated with a transition from state 1 to state 2 is then given by DE = ån w - ån 2i 2,i 1,i w1,i (2.11) i i The entropy (S) of a particular system can be calculated by the following expression (Atkins and de Paula 2014): S = k blnΩ (2.12) where k b is the Boltzmann constant Ω is the number of ways that the system can be arranged The change in entropy associated with a transition can therefore be calculated from knowledge of the number of ways that the system can arrange itself in each different state: ΔStr = k b(lnΩ2 − lnΩ1) (2.13) The aforementioned equations can be used to relate changes in the organization of molecular ensembles to changes in the molecular interactions and entropy contributions in the system. Nevertheless, it is often difficult to develop appropriate physical models that can be used to calculate changes in molecular interaction energies and entropy contributions for real systems because of the lack of information about molecular interactions and structural organization. Despite this limitation, it is often useful to think of physicochemical processes in terms of the change in interaction energies and entropy contributions that occur due to a transition in their molecular structure or organization. In the following sections, we examine two physicochemical processes (mixing and conformational changes) in more detail to highlight the advantages of taking a molecular approach to understanding physicochemical phenomenon. These two phenomena were considered because they often play an important role in the formation, stabilization, and physicochemical properties of food emulsions. 43 Molecular Characteristics 2.6 Thermodynamics of Mixing The utility of molecular models for understanding the relationship between molecular organization, interaction energies, and entropy contributions is demonstrated by considering the thermodynamics of mixing of a simple system. Nevertheless, understanding the physicochemical basis of this relatively simple phenomenon greatly facilitates the understanding of many more complex phenomenon that occur in food emulsions. Consider a hypothetical system that consists of a collection of two different types of equally sized spherical molecules, A and B (Figure 2.7). The free energy change that occurs when these molecules are mixed is given by ΔGmix = ΔEmix − TΔSmix (2.14) where ΔE mix and ΔS mix are the differences in the molecular interaction energy and entropy of the mixed and unmixed states, respectively. Practically, we may be interested in whether the resulting system consists of two immiscible liquids or as a simple mixture where the molecules are more or less intermingled (Figure 2.7). To a first approximation, thermodynamics tells us that if ΔG mix is positive, mixing is unfavorable and the molecules tend to exist as two separate phases (i.e., they are immiscible); if ΔG mix is negative, mixing is favorable and the molecules tend to be intermingled with each other (i.e., they are miscible); and if ΔG mix ≈ 0, the molecules are partly miscible and partly immiscible. In practice, more complicated situations can occur depending on the relationship between ΔG mix and the composition of the system. For simplicity, we assume that if the two types of molecules do intermingle with each other, they form a regular solution, that is, a completely random arrangement of the molecules (Figure 2.7), rather than an ordered solution, in which the type A molecules are preferentially surrounded by type B molecules, or vice versa. In practice, this means that the attractive forces between the two different types of molecules are not much stronger than the thermal energy of the system. This argument is therefore only applicable to mixtures that contain nonpolar or slightly polar molecules, where strong ion–ion or ion–dipole interactions do not occur. Despite the simplicity of this model system, we can still gain considerable insight into the behavior of more complex systems that are relevant to food emulsions. In the following sections, we separately consider the contributions of the interaction energy and the entropy to the overall free energy change that occurs on mixing. 2.6.1 Potential Energy Change on Mixing An expression for ΔEmix can be derived by calculating the total interaction energy of the molecules before and after mixing (Israelachvili 2011, Norde 2011). For both the mixed and the unmixed system, the total interaction energy is determined by summing the contribution of each of the different types of bond: E = nAA × wAA + nBB × wBB + nAB × wAB (2.15) where nAA, nBB, and nAB are the total number of bonds wAA, w BB, and wAB are the intermolecular pair potentials at equilibrium separation, which correspond to interactions between A–A, B–B, and A–B molecules, respectively. The total number of each type of bond formed is calculated from the number of molecules present in the system, the coordination number of the individual molecules (i.e., the number of molecules in direct contact with them), and their spatial arrangement. For example, many of the A–A and B–B interactions that occur in the unmixed system are replaced by A–B interactions in the mixed system. The difference in the total interaction energy between the mixed and unmixed states is then calculated: ΔEmix = Emix − Eunmixed. This type of analysis leads to the following equation (Evans and Wennerstrom 1999). 44 Food Emulsions: Principles, Practices, and Techniques ΔEmix = nXA XBw (2.16) where n is the total number of moles w is the effective interaction parameter XA and XB (=1 − XA) are the mole fractions of molecules of type A and B, respectively The effective interaction parameter is a measure of the compatibility of the molecules in a mixture, and is related to the intermolecular pair potential between isolated molecules by the expression (Norde 2011): ( w = z wAB - 12 éë wAA + wBB ùû ) (2.17) where z is the coordination number of a molecule (i.e., the number of contacting neighbors). The effective interaction parameter determines whether the mixing of dissimilar molecules is energetically favorable (w negative), unfavorable (w positive), or indifferent (w = 0). It should be stressed that even though there may be attractive forces between all the molecules involved (i.e., wAA, wBB, and wAB may all be negative), the overall interaction potential can be either negative (favorable to mixing) or positive (unfavorable to mixing) depending on the relative magnitude of the interactions. If the strength of the interaction between two different types of molecule (wAB) is greater (more negative) than the average strength between similar molecules (wAB < [wAA + wBB]/2), then w is negative, which favors the intermingling of the different types of molecules. On the other hand, if the strength of the interaction between two different types of molecule is weaker (less negative) than the average strength between similar molecules (wAB > [wAA + wBB]/2), then w is positive, which favors phase separation. If the strength of the interaction between different types of molecule is the same as the average strength between similar molecules (wAB = [wAA + wBB]/2), then the system has no preference for any particular arrangement of the molecules within the system (an ideal mixture). In summary, the change in the overall interaction energy may either favor or oppose mixing, depending on the relative magnitudes of the intermolecular pair potentials. 2.6.2 Entropy Change on Mixing An expression for ΔSmix is obtained from simple statistical considerations (Israelachvili 2011, Norde 2011). The mixing entropy of a system depends on the number of different ways that the molecules can be arranged in a given volume. For an immiscible system, there is only one possible arrangement of the two different types of molecule (i.e., zero entropy), but for a regular solution, there are a huge number of different possible arrangements (i.e., high entropy). A statistical analysis of this situation leads to the derivation of the following equation for the entropy of mixing (Atkins and de Paula 2014): ΔSmix = −nR (XAlnXA + XBlnXB) (2.18) ΔSmix is always positive because the mole fractions (XA and XB) are always between zero and one (so that the natural logarithm terms are always negative), which reflects the fact that there is always an increase in entropy after mixing. For regular solutions, the entropy contribution (−TΔSmix) always decreases the free energy of mixing, that is, favors the intermingling of the molecules. It should be stressed that for more complex systems, there may be additional contributions to the entropy due to the presence of some order within the mixed state, for example, organization of solvent molecules around a solute molecule (Section 4.3). In these cases, a more sophisticated analysis of the entropy contributions would be required. 2.6.3 Overall Free Energy Change on Mixing For a regular solution, the free energy change on mixing depends on the combined contributions of the interaction energies and the entropy: ΔGmix = n [XA XBw + RT(XAlnXA + XBlnXB)] (2.19) 45 Molecular Characteristics We are now in a position to investigate the relationship between the strength of the interactions between molecules and their structural organization within a bulk liquid. The dependence of the free energy of mixing on the effective interaction parameter and composition of a system consisting of two different types of molecule is illustrated in Figure 2.8. The two liquids are completely miscible when the interactions between the dissimilar molecules are not too energetically unfavorable (i.e., w < 2 RT) because the entropy of mixing contribution dominates. This accounts for the miscibility of liquids in which the interactions between the different types of molecules are fairly similar, for example, two nonpolar oils or ethanol with water. Two liquids are almost completely immiscible when the interactions between the dissimilar molecules are highly energetically unfavorable (i.e., w > 4 RT). This accounts for the immiscibility of oil and water, where the water molecules can form strong hydrogen bonds with each other but not with oil molecules. Two liquids are partially miscible when the interactions between the dissimilar molecules are moderately unfavorable (i.e., 2 RT < w < 4 RT). At these intermediate interaction strengths, there are two minimum values in the ΔG mix versus X A curve: X AL and X AH, which represent the lower and higher values, respectively (Figure 2.8). Under these circumstances, if the initial composition (X AT) of the overall system falls between these two minimum values, then the system separates into two phases, one with composition X A = X AL and the other with composition X A = X AH. The relative proportion of these two phases depends on the initial composition: ϕAL = (X AT − X AH)/(X AH − X AL), where ϕAH = (1 − ϕAL). The positions of X AL and X AH on the composition axis depend on the magnitude of the effective interaction parameter: the higher w, the smaller X AL , and the larger X AH (Norde 2011). Knowledge of the effective interaction parameter and composition of a system enables one to use the aforementioned equation for ΔG mix to construct a phase diagram that describes the conditions under which the molecular ensemble exists as an intimate mixture or as a phase separated system (Figure 2.9). The aforementioned approach therefore enables one to use thermodynamic considerations to relate bulk physicochemical properties of liquids (such as immiscibility) to molecular properties (such as the effective interaction parameter and the coordination number). Phase separation 0.4 w/RT = +4RT 0.2 w/RT = +3RT ΔGmix/RT 0 w/RT = +2RT –0.2 w/RT = +1RT –0.4 Mixing –0.6 –0.8 w/RT = +0RT 0 0.2 0.4 XA 0.6 0.8 1 FIGURE 2.8 Dependence of the free energy of mixing (calculated using Equation 2.19) on the composition and effective interaction parameter of a binary liquid. When ΔG mix is much less than –RT, the system tends to be mixed; otherwise it will be partly or wholly immiscible. 46 Food Emulsions: Principles, Practices, and Techniques Strong repulsion 7 Two-phase 6 ω/RT 5 Phase separation 4 3 Weak repulsion 2 One-phase 1 0 0 0.2 0.4 XA 0.6 0.8 1 Mixing FIGURE 2.9 Phase diagram describing the compositions and effective interaction parameters where the system exists as a one-phase (miscible) or two-phase (immiscible) system. This diagram was calculated using Equation 2.19 to find the minimum values in the ΔG mix versus X A graph. 2.6.4 Complications The derivation of the equation for ΔG mix given earlier depends on making a number of simplifying assumptions about the properties of the system that are not normally valid in practice, for example, that the molecules are spherical, that they all have the same size and coordination number, and that there is no ordering of the molecules within the mixture (Israelachvili 2011). It is possible to incorporate some of these features into the aforementioned theory, but a much more elaborate mathematical analysis is required. Food molecules come in all sorts of different sizes, shapes, and flexibilities; they may be nonpolar, polar, or amphiphilic; they may have one or more specific binding sites; or they may have to be in a certain orientation before they can interact with their neighbors. In addition, a considerable degree of structural organization of the molecules within a solvent often occurs when a solute is introduced if the solute–solvent interactions are sufficiently strong, for example, when mineral ions are added to water (Chapter 4). The variety of molecular characteristics exhibited by food molecules accounts for the great diversity of structures that are formed in food emulsions, such as bulk liquids, bulk solids, regular solutions, ordered solutions, molecular clusters, molecular networks, and immiscible liquids (Figure 2.1). Another problem with the thermodynamic approach is that food systems are rarely at thermodynamic equilibrium because of the presence of various kinetic energy barriers that prevent the system from reaching its lowest free energy state. This approach cannot therefore tell us whether two liquids will exist as an emulsion or not, because an emulsion is a thermodynamically unstable system. Nevertheless, it can tell us whether two liquids are capable of forming an emulsion, that is, whether they are immiscible or miscible. Despite the obvious limitations of the simple thermodynamic approach, it does highlight some of the most important features of molecular organization, especially the importance of considering both interaction energies and entropy effects. 2.7 Molecular Conformation So far we have only considered the way that molecular interactions influence the spatial distribution of molecules in a system. Molecular interactions can also determine the 3D conformation and flexibility of individual molecules (Walstra 2003, Norde 2011). Small molecules, such as H2O and CH4, normally exist in a single conformation that is determined by the relatively strong covalent bonds that hold the atoms together (Atkins and de Paula 2014). On the other hand, many larger molecules can exist in a number of different conformations because of the possibility of rotation around saturated covalent bonds, for example, proteins and polysaccharides (Tolstoguzov 2002, Dickinson and Semenova 2010). A macromolecule 47 Molecular Characteristics will tend to adopt the conformation that has the lowest free energy under the prevailing environmental conditions (Walstra 2003). The conformational free energy of a molecule is determined by the various interaction energies and entropy contributions within the particular system. The molecular interactions may be between different parts of the same molecule (intramolecular) or between the molecule and its neighbors (intermolecular). Similarly, the entropy is determined by the number of conformations that the molecule can adopt, as well as by any changes in the entropy caused by interactions with its neighbors, for example, restriction of their translational or rotational motion. To highlight the importance of molecular interactions and entropy in determining the conformation of molecules in solution, it is useful to examine a specific example. Consider a dilute aqueous solution containing hydrophilic biopolymer molecules that can exist in either a helical or a random coil conformation depending on the environmental conditions (Figure 2.10). Many types of food biopolymer used in emulsions are capable of undergoing this type of transformation, including proteins (such as gelatin) and polysaccharides (such as xanthan, carrageenan, and alginate). The free energy associated with the helix-to-coil transition between these two different conformations is given by ΔGh → c = ΔEh → c − TΔSh → c (2.20) where ΔGh → c, ΔEh → c, and ΔSh → c are the free energy, interaction energy, and entropy changes associated with the helix-to-coil transition. If ΔGh → c is negative, the transition is thermodynamically favorable and so the random coil conformation is most favored; if ΔGh → c is positive, the transition is unfavorable and the helix conformation is favored; and if ΔGh → r ≈ 0, the molecule spends part of its time in each of the conformations. In principle, a molecular interpretation of the free energy change associated with the helix-to-coil transition can be obtained by calculating the sum of the various intermolecular and intramolecular interaction energies in both the helix and coil states and by calculating the number of different ways that all the molecules in the system (polymer and solvent) could arrange themselves in both the helix and coil states. In practice, this is extremely difficult to carry out because of the lack of information about the strength and number of all of the different kinds of interactions that occur in the system, and about the magnitude of the entropy effects. Nevertheless, a broad understanding of the relative importance of interaction energy and entropy contribution effects can be obtained. A helical conformation allows a molecule to maximize the number of energetically favorable intermolecular and intramolecular interactions, while minimizing the number of energetically unfavorable ones (Bergethon 2010). Nevertheless, it has a much lower entropy than the random coil state because the molecule can only exist in a single conformation, whereas in the random coil state the molecule can exist in a large number of different conformations. At low temperatures, the interaction energy term dominates the entropy term and so the molecule tends to exist as a helix, but as the temperature is raised the entropy term (−TΔSh → c) becomes increasingly important until eventually it dominates and the molecule unfolds. The temperature at which the helix-to-coil transition takes place is referred to as the transition temperature, Th → c, which occurs when ΔG h → r = 0. Similar arguments can be used to account for the unfolding of globular proteins when they are heated above a particular temperature, although the relative Unfolding Folding Helix Random coil FIGURE 2.10 The conformation of a molecule in solution is governed by a balance of interaction energies and entropic effects. A helical molecule unfolds when it is heated above a certain temperature because the random coil conformation is entropically more favorable than the helical conformation. 48 Food Emulsions: Principles, Practices, and Techniques contribution of the various types of interaction energies is different (Dickinson and McClements 1996). It must be stressed that many food molecules are unable to adopt their thermodynamically most stable conformation because of the presence of various kinetic energy barriers (Section 1.2.1). When an energy barrier is much greater than the thermal energy of the system, a molecule may be trapped in a metastable state indefinitely. The flexibility of molecules in solution is also governed by both thermodynamic and kinetic factors. Thermodynamically, a flexible molecule has a large number of conformations that have fairly similar (±kT) low free energies. Kinetically, the energy barriers that separate these energy states must be small compared to the thermal energy of the system. When both of these criteria are met, a molecule will rapidly move between a number of different configurations and therefore be highly flexible. If the free energy difference between the conformations is large compared to the thermal energy, the molecule will tend to exist predominantly in the minimum free energy state (unless it is locked into a metastable state by the presence of a large kinetic energy barrier). Knowledge of the conformation and flexibility of a macromolecule under a particular set of environmental conditions is particularly important in understanding and predicting the behavior of many ingredients in food emulsions. The conformation and flexibility of a molecule determine its chemical reactivity, catalytic activity, intermolecular interactions, and functional properties, for example, solubility, dispersability, water holding capacity, gelation, foaming, and emulsification (Chapter 4). 2.8 Compound Interactions When one consults the literature dealing with molecular interactions in foods and other biological systems, one often comes across the terms “hydrogen bonding” and “hydrophobic interactions” (Belitz et al. 2009, Norde 2011, Brady 2013). In reality, these terms are a shorthand way of describing certain combinations of more fundamental interactions that occur between specific chemical groups commonly found in food molecules. Both of these compound interactions consist of contributions from various types of interaction energy (i.e., van der Waals, electrostatic, and steric overlap), as well as various entropy effects (e.g., translational or rotational). It is useful to highlight the general features of hydrogen bonds and hydrophobic interactions in this section before discussing their importance in determining the properties of individual food components later (Chapter 4). As mentioned earlier, if one considers molecular organization in terms of compound interactions (rather than fundamental interactions), then it is inconvenient to divide the overall free energy change (ΔG) with a transition into separate enthalpy (ΔH) and entropy (ΔS) changes (Equation 2.8), since the compound interactions contain both enthalpy and entropy contributions. Instead, it is more convenient to group the enthalpy and some of the entropy contributions together into specific compound interactions (e.g., hydrophobic interactions), and consider other entropy contributions separately (e.g., translational, mixing, or conformational entropy). For example, the hydrophobic interaction contains enthalpy changes associated with alterations in molecular interactions, as well as entropy changes associated with alterations in molecular organization (Chapter 4). The general approach used to relate the free energy of mixing to the molecular interactions and entropy contributions can still be used (Section 2.5), but the effective interaction parameter is expressed in terms of interaction free energies (e.g., gAA, g BB, and gAB) rather than just interaction energies (e.g., wAA, w BB, and wAB) and Equation 2.9 is used rather than Equation 2.8 (Norde 2011). 2.8.1 Hydrogen Bonds Hydrogen bonds play a crucial role in determining the functional properties of many of the most important molecules present in food emulsions, including water, proteins, lipids, carbohydrates, surfactants, and minerals (Chapter 4). They are formed between a lone pair of electrons on an electronegative atom (such as oxygen) and a hydrogen atom on a neighboring group, that is, O–Hδ+⋯Oδ− (Bergethon 2010, Norde 2011). The major contribution to hydrogen bonds is electrostatic (dipole–dipole), but van der Waals forces and steric repulsion also make a significant contribution. Typically, they have bond strengths Molecular Characteristics 49 between about 10 and 40 kJ mol−1 and lengths of about 0.18 nm (Israelachvili 2011). The actual strength of a particular hydrogen bond depends on the electronegativity and orientation of the donor and acceptor groups (Stone 2013). Hydrogen bonds are stronger than most other examples of dipole–dipole interaction because hydrogen atoms have a strong tendency to become positively polarized and because they have a small radius. In fact, hydrogen bonds are so strong that they cause appreciable alignment of the molecules involved. The strength and directional character of hydrogen bonds are responsible for many of the unique properties of water (Chapter 4). 2.8.2 Hydrophobic Interactions Hydrophobic interactions also play a major role in determining the behavior of many important ingredients in food emulsions, particularly lipids, surfactants, and proteins (Brady 2013). They manifest themselves as a strong attractive force that acts between nonpolar groups separated by water (Israelachvili 2011, Norde 2011, Stone 2013). Nevertheless, the actual origin of hydrophobic interactions is the ability of water molecules to form relatively strong hydrogen bonds with their nearest neighbors, whereas nonpolar molecules can only form relatively weak van der Waals bonds. When a nonpolar molecule is introduced into liquid water, it causes the water molecules in its immediate vicinity to rearrange themselves, which changes both the interaction energy and entropy of the system (Chapter 4). Overall, these changes are thermodynamically unfavorable and so the system attempts to minimize contact between water and nonpolar groups, which appears as an attractive force between the nonpolar groups. It is this effect that is largely responsible for the immiscibility of oil and water, the adsorption of emulsifiers to interfaces, the aggregation of protein molecules, and the formation of surfactant micelles, and it is therefore particularly important for food scientists to have a good understanding of its origin and the factors that influence it. 2.9 Computer Modeling of Liquid Properties Our understanding of the way that molecules organize themselves in a liquid can be greatly enhanced by the use of computer modeling techniques (Norde 2011, Palma et al. 2012, Stone 2013). Computer simulations of the relationship between molecular properties and structural organization have provided a number of valuable insights that are relevant to a better understanding of the behavior of food emulsions, including the miscibility/immiscibility of liquids, the formation of surfactant micelles, the adsorption and displacement of emulsifiers at interfaces, the transport of nonpolar molecules through aqueous phases, the conformation and flexibility of biopolymers in solution, polymer interactions, and the formation of gels (Euston 2004, Pugnaloni et al. 2004, Zahn 2009, Dickinson 2013, Euston 2013, Jusufi 2013). An example of the power of computer simulation techniques for understanding molecular processes relevant to food emulsions is shown in Figure 2.11. The displacement of one type of molecule from an interface by another more surface-active type of molecule is simulated. This type of simulation can be used to provide insights into the molecular factors that influence the displacement of proteins from droplet surfaces by other surface-active substances in the surrounding aqueous phase. The stability and physicochemical properties of emulsions are strongly influenced by interfacial properties, and so it is important to have a fundamental understanding of the factors that influence the type, concentration, interactions, and arrangement of surface-active molecules at interfaces. The first step in a molecular simulation is to define the characteristics of the molecules involved (e.g., size, shape, flexibility, and polarity) and the nature of the intermolecular pair potentials that act between them (Frenkel and Smit 2001).* A collection of these molecules is arbitrarily distributed within a box that represents a certain region of space, and the change in the conformation and/or organization of the molecules is then monitored as they are allowed to interact with each other. Depending on the simulation technique used, one can obtain information about the evolution of the structure with time and/or * It should be noted that it is usually necessary to make a number of simplifying assumptions about the properties and interactions of the molecules in order to create computer programs that can be solved in a reasonable time period. 50 Food Emulsions: Principles, Practices, and Techniques Initial Final Top view Side view (a) (b) FIGURE 2.11 Brownian dynamics simulation of the displacement of a gel-like adsorbed layer of bond-forming white spheres by more surface-active non–bond forming small black spheres. In (a), the displacer black spheres have been just introduced to the system beneath the interface. In (b), the black spheres have partially displaced the gel-forming white spheres. The thickening of the displaced layer can be appreciated on the side-on profiles shown. (Courtesy of Professor Eric Dickinson and Dr. Luis Pugnaloni, University of Leeds, Leeds, U.K.) about the equilibrium structure of the molecular ensemble. The two most commonly used computer simulation techniques are the Monte Carlo approach and the Molecular Dynamics approach (Cramer 2004, Euston 2013). 2.9.1 Monte Carlo Techniques This technique is named after Monte Carlo, a town in the principality of Monaco (near southern France), which is famous for its gambling and casinos. The reason for this name is the fact that the movement of the molecules in the box is largely determined by a random selection process, just as the winner in a Roulette game is selected. Initially, one starts with an arbitrary arrangement of the molecules in the box. The overall interaction energy is then calculated from knowledge of the positions of all the molecules and their intermolecular pair potentials. One of the molecules is then randomly selected and moved to a new location and the overall interaction energy is recalculated. If the energy decreases, then the move is definitely allowed. However, if the energy increases, the probability of the move being allowed depends on the magnitude of the energy change compared to the thermal energy (RT). When the increase in energy is much greater than RT, the move is highly unlikely and will probably be rejected, but if it is on the same order as RT, it is much more likely to be accepted. This procedure is continued until there is no further change in the average interaction energy, which is taken to be the minimum potential energy state of the system. The free energy and entropy of the system are calculated by monitoring the fluctuations in the overall interaction energy after successive molecules have been moved once the system has reached equilibrium. The Monte Carlo technique therefore provides information about the equilibrium properties of a system, rather than about the evolution in its properties with time. 2.9.2 Molecular Dynamics Techniques This technique is named after the fact that it relies on monitoring the movement of molecules with time. Initially, one starts with an arbitrary arrangement of the molecules in the box. The computer then calculates the force that acts on each of the molecules as a result of its interactions with the surrounding molecules. Newton’s equations of motion are then used to determine the direction and speed that each of the molecules moves within a time interval that is short compared to the average time between molecular collisions (typically 10 −15 to 10 −14 s). By carrying out the computation over a large number of successive time intervals, it is possible to monitor the evolution of the system with time. The main limitation of this Molecular Characteristics 51 technique is that a huge number of computations have to be carried out to model events on timescales relevant to pertinent physicochemical phenomena using a sufficiently high number of molecules to accurately represent these phenomena. Consequently, powerful computers and relatively long computation times are required to model even relatively simple molecular processes. Even so, as computer technology advances these computation times are decreasing, which enables researchers to model increasingly complex systems. A molecular dynamics simulation should lead to the same final state as a Monte Carlo simulation if it is allowed to proceed long enough to reach equilibrium. The free energy of the system is determined by calculating the fraction of molecules in different energy states once the system has reached equilibrium. In practice, molecular dynamics simulations are more difficult to set up and take much longer to reach equilibrium than Monte Carlo simulations. For this reason, Monte Carlo simulations are more practical if a researcher is only interested in equilibrium properties, whereas molecular dynamic simulations are used when information about both the kinetics and thermodynamics of a system is required. Molecular dynamics techniques are particularly suitable for studying nonequilibrium processes, such as mass transport, fluid flow, adsorption kinetics, and solubilization processes (Dickinson and McClements 1996, Euston 2013). It is clear from the aforementioned discussion that each technique has its own advantages and disadvantages and that both techniques can be used to provide useful insights into the molecular basis for some of the most important bulk physiochemical properties of food emulsions. For example, researchers can systematically examine the influence of specific molecular properties on their behavior, such as molecular size, shape, polarity, charge, or hydrophobicity. 2.10 Measurement of Molecular Characteristics Prediction of the organization of molecules using statistical thermodynamics or computer simulation techniques requires information about the total number and strength of the different kinds of bonds in the system, as well as of the spatial distribution of the molecules within the system. Direct measurement of these molecular properties is usually extremely difficult because of their extremely small size, large number, and rapid movement. For this reason, most of the information that we have about molecular characteristics has to be inferred from more indirect measurements of the physicochemical properties of materials (Israelachvili 2011, Stone 2013): • Measurements of the thermodynamic properties on gases, liquids, and solids (e.g., boiling points, heats of vaporization, lattice energies, and pressure–volume–temperature data) provide information on short-range attraction between molecules. • Measurements of the thermodynamic properties of solutions (e.g., solubility, miscibility, partitioning, phase diagrams, and osmotic pressure) provide information about short-range solute–solute and solute–solvent interactions. • Measurements of the bulk physicochemical properties of gases, liquids, and solids (e.g., density, compressibility, viscosity, diffusion, scattering, and spectroscopy) provide information about short-range attractive and/or repulsive interactions between molecules. • Measurements of adhesion forces holding solid surfaces together can provide information about short-range attractive interactions. • Measurements of surface or interfacial properties (tensions or contact angles) can provide information about short-range liquid–liquid and solid–liquid interactions. In addition, specialized analytical methods have been developed to directly measure the forces acting between molecules or surfaces as a function of their separation, for example, surface force apparatus, total internal reflection microscopy, atomic force microscopy, and single molecule force microscopy (Neuman and Nagy 2008, Israelachvili 2011, Noy and Friddle 2013). These techniques often provide detailed information about force–distance profiles for different kinds of molecular interactions (Noy and Friddle 2013). 52 Food Emulsions: Principles, Practices, and Techniques REFERENCES Atkins, P. and J. de Paula (2014). Physical Chemistry: Thermodynamics, Structure, and Change. Oxford, U.K.: Oxford University Press. Belitz, H. D., W. Grosch, and P. Schieberle (2009). Food Chemistry. Berlin, Germany: Springer. Bergethon, P. R. (2010). The Physical Basis of Biochemistry: The Foundations of Molecular Biophysics. New York: Springer. Bishop, K. J. M., C. E. Wilmer, S. Soh, and B. A. Grzybowski (2009). Nanoscale forces and their uses in selfassembly. Small 5(14): 1600–1630. Boon, C. S., D. J. McClements, J. Weiss, and E. A. Decker (2010). Factors influencing the chemical stability of carotenoids in foods. Critical Reviews in Food Science and Nutrition 50(6): 515–532. Brady, J. W. (2013). Introductory Food Chemistry. Ithaca, NY: Cornell University Press. Buffler, C.R. (1995). Advances in dielectric characterization of foods. In Characterization of Food: Emerging Methods, A.G. Gaonkar, ed., Chapter 10. Amsterdam, the Netherlands: Elsevier. Choi, S. J., E. A. Decker, L. Henson, L. M. Popplewell, and D. J. McClements (2010). Inhibition of citral degradation in model beverage emulsions using micelles and reverse micelles. Food Chemistry 122(1): 111–116. Cramer, C. J. (2004). Essentials of Computational Chemistry: Theories and Models. New York: Wiley. Dickinson, E. (2013). Structure and rheology of colloidal particle gels: Insight from computer simulation. Advances in Colloid and Interface Science 199: 114–127. Dickinson, E. and D. J. McClements (1996). Advances in Food Colloids. Glasgow, U.K.: Blackie Academic and Professional. Dickinson, E. and M. G. Semenova (2010). Biopolymers in Food Colloids: Thermodynamics and Molecular Interactions. Boca Raton, FL: CRC Press. Euston, S. R. (2004). Computer simulation of proteins: Adsorption, gelation and self-association. Current Opinion in Colloid & Interface Science 9(5): 321–327. Euston, S. R. (2013). Modelling and computer simulation of food structures. In Food Microstructures: Microscopy, Measurement and Modelling, V. J. Morris and K. Groves, eds., pp. 336–385. Cambridge, U.K.: Woodhead Publishing. Evans, E. D. and W. Wennerstrom (1999). The Colloidal Domain: Where Physics, Chemistry and Biology Meet. New York: Wiley-VCH. Frenkel, D. and B. Smit (2001). Understanding Molecular Simulation: From Algorithms to Applications, 2nd edn. Waltham, MA, Academic Press. Hiemenz, P. C. and R. Rajagopalan (1997). Principles of Colloid and Surface Chemistry. New York: Marcel Dekker. Israelachvili, J. (2011). Intermolecular and Surface Forces, 3rd edn. London, U.K.: Academic Press. Jusufi, A. (2013). Molecular simulations of self-assembly processes of amphiphiles in dilute solutions: The challenge for quantitative modelling. Molecular Physics 111(21): 3182–3192. McClements, D. J. and E. A. Decker (2000). Lipid oxidation in oil-in-water emulsions: Impact of molecular environment on chemical reactions in heterogeneous food systems. Journal of Food Science 65(8): 1270–1282. Neuman, K. C. and A. Nagy (2008). Single-molecule force spectroscopy: Optical tweezers, magnetic tweezers and atomic force microscopy. Nature Methods 5(6): 491–505. Ninhan, B. W. and P. L. Nostro (2010). Molecular Forces and Self Assembly: In Colloid, Nano Sciences and Biology. Cambridge, U.K.: Cambridge University Press. Norde, W. (2011). Colloids and Interfaces in Life Sciences and Bionanotechnology. Boca Raton, FL: CRC Press. Noy, A. and R. W. Friddle (2013). Practical single molecule force spectroscopy: How to determine fundamental thermodynamic parameters of intermolecular bonds with an atomic force microscope. Methods 60(2): 142–150. Palma, C. A., M. Cecchini, and P. Samori (2012). Predicting self-assembly: From empirism to determinism. Chemical Society Reviews 41(10): 3713–3730. Pugnaloni, L. A., E. Dickinson, R. Ettelaie, A. R. Mackie, and P. J. Wilde (2004). Competitive adsorption of proteins and low-molecular-weight surfactants: Computer simulation and microscopic imaging. Advances in Colloid and Interface Science 107(1): 27–49. Molecular Characteristics 53 Stone, A. J. (2013). The Theory of Intermolecular Forces. Oxford, U.K.: Oxford University Press. Tolstoguzov, V. (2002). Thermodynamic aspects of biopolymer functionality in biological systems, foods, and beverages. Critical Reviews in Biotechnology 22(2): 89–174. Walstra, P. (2003). Physical Chemistry of Foods. New York: Marcel Decker. Waraho, T., D. J. McClements, and E. A. Decker (2011). Mechanisms of lipid oxidation in food dispersions. Trends in Food Science & Technology 22(1): 3–13. Zahn, D. (2009). On the role of the solvent in biosystems: Atomistic insights from computer simulations. Frontiers in Bioscience 14: 3586–3593. 3 Colloidal Interactions 3.1 Introduction Food emulsions contain a variety of structural entities that have at least one dimension that falls within the colloidal size range (from a few nanometers to a few micrometers), for example, biopolymer aggregates, micelles, emulsion droplets, fat crystals, ice crystals, and air cells. The characteristics of these colloidal particles, and their interactions with each other, are responsible for many of the most important physicochemical, physiological, and sensory properties of emulsion-based foods. The ability of food scientists to understand, predict, and control the properties of food emulsions therefore depends on knowledge of the interactions that arise between colloidal particles. In this chapter, we examine the origin and nature of the most important types of colloidal interaction, while in later chapters, we consider the relationship between these interactions and the stability, rheology, and appearance of food emulsions (Chapters 7 through 9). The interaction between a pair of colloidal particles is the result of interactions between all of the molecules within them, as well as those within the intervening medium (Bergethon 2010, Israelachvili 2011). For this reason, many of the interactions between colloidal particles appear at first sight to be similar to those between molecules, for example, van der Waals, electrostatic, and steric (Chapter 2). Nevertheless, the characteristics of colloidal interactions are often different from their molecular counterparts, because of additional features that arise due to the relatively large size of colloidal particles compared to molecules, and due to the relatively large number of different kinds of molecules involved (Bishop et al. 2009). The major emphasis of this chapter will be on interactions between fat droplets dispersed within water, although the same principles can be applied to the various other types of colloidal particles that are commonly found in foods. 3.2 Colloidal Interactions and Droplet Aggregation Colloidal interactions govern whether the droplets within an emulsion remain as separate entities or whether they associate with each other, as well as determining the characteristics of any aggregates formed, for example, their size, shape, porosity, and deformability (Dickinson 2006, 2010, 2013). Many of the bulk physicochemical, physiological, and sensory properties of food emulsions are determined by the degree of droplet aggregation and the nature of the aggregates formed, such as their stability, texture, appearance, mouthfeel, and digestion (Chapters 7 through 11). It is therefore extremely important for food scientists to understand the relationship between colloidal interactions, droplet aggregation, and emulsion properties. In the previous chapter, the interaction between two isolated molecules was described in terms of an intermolecular pair potential (Chapter 2). In a similar fashion, the interactions between two emulsion droplets can be described in terms of an interdroplet pair potential. The interdroplet pair potential, w(h), is the energy required to bring two emulsion droplets from an infinite distance apart to a surface-tosurface separation of h (Figure 3.1). Before examining specific types of interactions between emulsion droplets, it is useful to examine the features of colloidal interactions in a more general fashion. Consider a system that consists of two similar droplets of radius r at a surface-to-surface separation of h (Figure 3.1). For convenience, we assume that only two types of interaction occur between the emulsion droplets, one attractive and one repulsive: 55 56 Food Emulsions: Principles, Practices, and Techniques Medium 1 Medium 2 h r FIGURE 3.1 Dispersed phase emulsion droplets (medium 2) of radius r separated by a surface-to-surface separation h though a liquid continuous phase (medium 1). w(h) = wattractive(h) + wrepulsive(h) (3.1) The overall interaction between the droplets depends on the relative magnitude and range of these attractive and repulsive interactions. A number of different types of behavior can be distinguished depending on the nature of the interactions involved (Figure 3.2): Attractive interactions dominate at all separations: If the attractive interactions are greater than the repulsive interactions at all droplet separations, then the overall interaction is always attractive (Figure 3.2a), which means that the droplets will tend to aggregate (provided the strength of the interaction is greater than the disorganizing influence of the thermal energy). Repulsive interactions dominate at all separations: If the repulsive interactions are stronger than the attractive interactions at all separations, then the overall interaction is always repulsive (Figure 3.2b), which means that the droplets tend to remain as individual entities. Attractive interactions dominate at large separations, but repulsive interactions dominate at short separations: At very large droplet separations, there is no effective interaction between the droplets. As the droplets move closer together, the attractive interaction initially dominates, but at closer separations, the repulsive interaction dominates (Figure 3.2c). At some intermediate surface-to-surface separation, there is a minimum in the interdroplet interaction potential at a separation hmin. The depth of this minimum, w(hmin), is a measure of the strength of the interaction between the droplets, while the position of the minimum, hmin, corresponds to the most likely separation of the droplets. Droplets tend to aggregate when the strength of the interaction is large compared to the thermal energy w ( hmin )  kT , remain as separate entities when the strength of the interaction is much smaller than the thermal energy w ( hmin ) » 0 , and spend some time together and some time apart at intermediate interaction strengths. When droplets fall into a deep potential energy minimum, they are said to be strongly flocculated or coagulated because a large amount of energy is required to pull them apart. When they fall into a shallow minimum, they are said to be weakly flocculated because they are fairly easy to pull apart. The fact that there is an extremely large repulsion between the droplets at close separations prevents them from coming close enough together to coalesce. Repulsive interactions dominate at large separations, but attractive interactions dominate at short separations: At very large droplet separations, there is no effective interaction between the droplets. As the droplets move closer together, the repulsive interaction initially dominates, but at closer separations, the attractive interaction dominates (Figure 3.2d). At some intermediate surface-to-surface separation (hmax), there is an energy barrier that the droplets must overcome before they can move any closer together (Friberg 1997). If the height of this energy barrier is large compared to the thermal energy of the system w ( hmax )  20 kT , the droplets are effectively prevented from coming close together and will therefore remain as separate entities. If the height of the energy barrier is small compared to the thermal energy w ( hmax ) < 5 kT , the droplets easily have enough thermal energy to jump over it, and they rapidly fall into the deep minimum that exists at close separations. At intermediate values, the droplets still tend to aggregate, but this process occurs slowly because only a fraction of droplet–droplet collisions has ( ( ) ( ( ) ) ) 57 Colloidal Interactions 100 100 R R 50 0 0 5 T 10 w(h)/kT w(h)/kT 50 0 5 A –100 h (nm) (b) A –100 h (nm) 100 100 R R 0 T 50 0 5 10 w(h)/kT w(h)/kT 50 (c) 0 T 0 5 10 –50 –50 –100 10 –50 –50 (a) T 0 A A h (nm) (d) –100 h (nm) FIGURE 3.2 The overall interaction of a pair of emulsion droplets depends on the relative magnitude and range of any attractive and repulsive interactions. The overall interaction may be attractive at some separations and repulsive at others. (a) Attractive interactions dominate at all separations, (b) repulsive interactions dominate at all separations, (c) attractive interactions dominate at large separations, but repulsive interactions dominate at short separations, and (d) repulsive interactions dominate at large separations, but attractive interactions dominate at short separations. See text for discussion of differences in figures. sufficient energy to jump over the energy barrier. The fact that there is an extremely strong attraction between the droplets at close separations is likely to cause them to coalesce, that is, merge together. Despite the simplicity of the aforementioned model (Equation 3.1), we have already gained a number of valuable insights into the role that colloidal interactions play in determining whether emulsion droplets are likely to be nonaggregated, flocculated, or coalesced. In particular, the importance of the sign, magnitude, and range of the colloidal interactions has become apparent. As would be expected, the colloidal interactions that arise between the droplets in real food emulsions are much more complex than those considered earlier. First, there are a number of different types of repulsive and attractive interactions that contribute to the overall interaction potential, each with a different sign, magnitude, and range. Second, food emulsions contain a huge number of droplets and other colloidal particles that have different sizes, shapes, and properties that may all interact with each other. Third, the liquid that surrounds the droplets may be compositionally and structurally complex, containing various types of ions and molecules. Droplet–droplet interactions in real food emulsions are therefore influenced by the presence of the neighboring droplets, as well as by the precise nature of the surrounding liquid phase. For these reasons, it is difficult to accurately account for colloidal interactions in real food emulsions because of the mathematical complexity of describing interactions between huge numbers of molecules, ions, and particles in complex systems. Nevertheless, considerable insight into the major factors that 58 Food Emulsions: Principles, Practices, and Techniques influence the properties of food emulsions can be obtained by examining the interaction between a pair of droplets in a simple liquid. In addition, our progress toward understanding complex food systems depends on us first understanding the properties of simpler model systems. These model systems can then be incrementally increased in complexity and accuracy as advances are made in our knowledge and computational power. In the following sections, the origin and nature of the major types of colloidal interaction that arise between emulsion droplets are reviewed. In Section 3.11, we then consider ways in which these individual interactions combine with each other to determine the overall interdroplet pair potential and thus the stability of emulsion droplets to aggregation. Knowledge of the contribution that each of the individual colloidal interactions makes to the overall interaction is often of great practical importance since it enables one to identify the most effective means of controlling the stability of a given system to droplet aggregation. For example, the influence of pH or salt on the aggregation of electrostatically stabilized emulsions can be investigated in a systematic fashion, or the influence of interfacial thickness on the stability of sterically stabilized emulsions. 3.3 Van der Waals Interactions 3.3.1 Origin of van der Waals Interactions Intermolecular van der Waals interactions operate between all of the different kinds of molecular species in the dispersed and continuous phases of emulsions, which results in a net colloidal van der Waals interaction between emulsion droplets (Hiemenz and Rajagopalan 1997, Israelachvili 2011). Ultimately, the colloidal van der Waals interaction is a result of the orientation, induced, and dispersion contributions to the intermolecular van der Waals interaction discussed in Section 2.3.3 at Chapter 2. An appreciation of these different contributions is important because it enables one to understand some of the physicochemical factors that influence the strength of colloidal van der Waals interactions between emulsion droplets, for example, retardation and electrostatic screening (see Section 3.3.2.1). 3.3.2 Modeling van der Waals Interactions The van der Waals interactions between macroscopic bodies can be calculated using two different mathematical approaches (Hunter 1986, Bishop et al. 2009, Israelachvili 2011). In the microscopic approach, the van der Waals interaction between a pair of droplets is calculated by carrying out a pair-wise summation of the interaction energies of all the molecules in one of the droplets with all of the molecules in the other droplet. Calculations made using this approach rely on knowledge of the properties of the individual molecules, such as polarizabilities, dipole moments, and electronic energy levels. In the macroscopic approach, the droplets and surrounding medium are treated as continuous liquids that interact with each other because of the fluctuating electromagnetic fields generated by the movement of the electrons within them. Calculations made using this approach rely on knowledge of the bulk physicochemical properties of the liquids, such as dielectric constants, refractive indices, and absorption frequencies. Under certain circumstances, both theoretical approaches give similar predictions of the van der Waals interaction between emulsion droplets. In general, however, the macroscopic approach is usually the most suitable for describing interactions between emulsion droplets because it automatically takes into account the effects of retardation and of the liquid surrounding the droplets (Hunter 1986). 3.3.2.1 Interdroplet Pair Potential The van der Waals interdroplet pair potential, w VDW(h), of two emulsion droplets of equal radius, r, separated by a surface-to-surface distance, h (Figure 3.1), is given by the following expression: wVDW (h) = - A212 6 éæ 2r 2 ö æ ö æ h2 + 4rh 2r 2 + ln êç 2 ÷ ç 2 ÷+ç 2 2 2 êëè h + 4rh ø è h + 4rh + 4r ø è h + 4rh + 4r öù ÷ú ø úû (3.2) 59 Colloidal Interactions Here, A212 is the Hamaker function for emulsion droplets (medium 2) separated by a liquid (medium 1). The value of the Hamaker function can be calculated using either the microscopic or macroscopic approaches mentioned earlier (Mahanty and Ninham 1976). At close separations h ≪ r, the aforementioned equation can be simplified: wVDW (h) = - A212r 12h (3.3) This equation indicates that van der Waals interactions between a pair of colloidal particles (w ∝ 1/h) are of much longer range than those between a pair of molecules (w ∝ 1/s6), which has important consequences for determining the stability of food emulsions. Equations for calculating the van der Waals interaction between spheres of unequal radius are given elsewhere (Bishop et al. 2009). 3.3.2.2 Hamaker Function In general, an accurate calculation of the Hamaker function of a pair of emulsion droplets is a complicated task (Bowen and Jenner 1995, Israelachvili 2011). Knowledge of the optical properties (dielectric permittivities) of the oil, water, and interfacial phases is required over a wide range of frequencies, and this information is not readily available for most substances. In addition, the full theory must be solved numerically using a computer (Pailthorpe and Russel 1982). Nevertheless, approximate analytical expressions for the Hamaker function have been derived, which can be calculated using data that can easily be found in the literature (Israelachvili 2011): A212 = Av = 0 + Av > 0 (3.4) where Av= 0 = Av > 0 3 kT 4 ¥ å s =1 ( ( 1 æ e1 - e2 ö s 3 çè e1 + e2 ÷ø ) ) 2s 2 2 2 3hve n1 - n2 = 3/ 2 16 2 n12 + n22 ε is the static relative dielectric constant n is the refractive index ve is the major electronic absorption frequency in the ultraviolet region of the electromagnetic spectrum (which is assumed to be equal for both phases) h is Planck’s constant, and the subscripts 1 and 2 refer to the continuous phases and droplets, respectively Equation 3.4 indicates that the Hamaker function of two similar droplets is always positive, which means that wVDW(h) is always negative, so that the van der Waals interaction is always attractive. However, the interactions between colloidal particles with different compositions may be either attractive or repulsive, depending on the relative physicochemical properties of the particles and the intervening medium (Israelachvili 2011). In Equation 3.5, the Hamaker function is divided into two contributions: a zero-frequency component (Av = 0) and a frequency-dependent component (Av > 0). The overall interdroplet pair potential is therefore given by wVDW(h) = wv = 0(h) + wv > 0(h) (3.5) where wv = 0(h) and wv > 0(h) are determined by inserting the expressions for Av = 0 and Av > 0 into Equation 3.2 or 3.3. The zero-frequency component is due to orientation and induction contributions to the van der Waals interaction, whereas the frequency-dependent component is due to the dispersion contribution 60 Food Emulsions: Principles, Practices, and Techniques 0 0 2 4 6 8 10 –10 –20 –30 w(h)/kT –40 h –50 –60 –70 w(v = 0) –80 w(v > 0) –90 w(Total) –100 h (nm) FIGURE 3.3 Predicted dependence of the interaction potential on droplet separation for van der Waals interactions: total (w VDW), zero-frequency contribution (wv = 0), and frequency-dependent contribution (wv > 0). See Table 3.1 for the physicochemical properties of the oil and water phases used in the calculations (r = 1 μm). (Section 2.3.3). The separation of the Hamaker function into these two components is particularly useful for understanding the influence of electrostatic screening and retardation on van der Waals interactions. The variation of wVDW(h), wv = 0(h) and wv > 0(h) with droplet separation for two oil droplets dispersed in water is shown in Figure 3.3. Calculations show that for food emulsions, the Hamaker function is typically about 0.56 × 10 −20 J (1.37 kT), with about 43% of this coming from the zero-frequency contribution and 57% from the frequency-dependent contribution. The physicochemical properties needed to calculate Hamaker functions for ingredients typically found in food emulsions are summarized in Table 3.1. In practice, the magnitude of the Hamaker function depends on droplet separation and is considerably overestimated by Equation 3.4 because of the effects of electrostatic screening, retardation, and interfacial layers. Electrostatic screening effects: The zero-frequency component of the Hamaker function (Av = 0) is electrostatic in origin because it depends on interactions that involve permanent dipoles, that is, orientation and induction forces (Section 2.3.3). Consequently, this part of the van der Waals interaction is screened (reduced) when droplets are suspended in an electrolyte solution because of the TABLE 3.1 Physicochemical Properties Needed to Calculate the Nonretarded Hamaker Function (Equation 3.4) for Some Materials Commonly Found in Food Emulsions Medium Water Oil Pure protein x% protein in water Pure Tween 20 Static Relative Dielectric Constant Εr 80 2 5 5x + 80 (1 − x) Note: Data compiled from various sources. Refractive Index N Absorption Frequency ve/1015 s−1 1.333 1.433 1.56 1.56x + 1.333(1 − x) 1.468 3.0 2.9 2.9 2.9 2.9 61 Colloidal Interactions accumulation of counterions around the droplets (Section 3.4). Electrostatic screening causes the zerofrequency component to decrease with increasing droplet separation, and with increasing electrolyte concentration (Mahanty and Ninham 1976, Israelachvili 2011). At high electrolyte concentrations, the zero-frequency contribution decays rapidly with distance, and makes a negligible contribution to the overall interaction energy at distances greater than a few κ−1 (Figure 3.4a), where κ−1 is the Debye 100 Screening of VDW Relative Interaction Strength (%) 90 80 1 mM 70 60 10 mM 50 40 100 mM 1000 mM 0 2 1 3 (a) 5 4 6 h (nm) 100 Retardation of VDW Relative Interaction Strength (%) 95 90 85 80 75 70 65 60 (b) 0 10 20 30 40 50 h (nm) FIGURE 3.4 Influence of (a) electrostatic screening and (b) retardation on the normalized van der Waals attraction between two oil droplets suspended in water. The normalized interaction potentials are reported as w(h) in the presence of the stipulated effect relative to w(h) in the absence of the effect expressed as a percentage. See Table 3.1 for the physicochemical properties of the phases used in the calculations. 62 Food Emulsions: Principles, Practices, and Techniques screening length (see Section 3.4.2.1 and Equation 3.7). On the other hand, the frequency-dependent component (Av > 0) is unaffected by electrostatic screening because the ions in the electrolyte solution are so large that they do not have time to move in response to the rapidly fluctuating dipoles. Consequently, the van der Waals interaction may decrease by as much as 42% in oil-in-water emulsions at high ionic strength solutions because the zero-frequency component is completely screened. Equations for calculating the influence of electrostatic screening on van der Waals interactions have been developed (Israelachvili 2011). To a first approximation, the influence of electrostatic screening on the zero-frequency contribution to the van der Waals interaction can be accounted for by replacing Av = 0 with Av = 0 × e −2κh in the earlier equations. Retardation: The strength of the van der Waals interaction between emulsion droplets is reduced because of a phenomenon known as retardation (Israelachvili 2011). The origin of retardation is the finite time taken for an electromagnetic field to travel from one droplet to another and back. The frequency-dependent contribution to the van der Waals interaction (wv > 0) is the result of a transient dipole in one droplet inducing a dipole in another droplet, which then interacts with the first dipole (Section 2.3.3). The strength of the resulting attractive force is reduced if the time taken for the electromagnetic field to travel between the droplets is comparable to the lifetime of a transient dipole, because then the orientation of the first dipole will have changed by the time the field from the second dipole arrives back. This effect becomes appreciable at dipole separations greater than a few nanometers, and results in a decrease in the frequency-dependent (Av > 0) contribution to the Hamaker function with droplet separation. The zero-frequency contribution (Av = 0) is unaffected by retardation because it is electrostatic in origin. Consequently, the contribution of the Av > 0 term becomes increasingly small as the separation between the droplets increases, which leads to a decrease in the overall interaction potential (Figure 3.4b). Any accurate prediction of the van der Waals interaction between droplets should therefore include retardation effects. To a first approximation, the influence of retardation on the frequency-dependent contribution to the van der Waals interaction can be accounted for by replacing Av > 0 with Av > 0 × (1 + 0.11 h) −1 in the previous equations (Gregory 1981). Thus, the retarded value of wv > 0(h) between two emulsion droplets at a separation of 20 nm is only about 30% of the nonretarded value. Influence of interfacial membranes: So far we have assumed that the van der Waals interaction occurs between two homogeneous spheres separated by an intervening medium (Figure 3.1). In reality, emulsion droplets are normally surrounded by a thin layer of emulsifier molecules, and this interfacial layer has different physicochemical properties (ε R, n and ve) than either the oil or water phases (Figure 3.5). The molecules nearest the surface of a particle make the greatest contribution to the overall van der Waals interaction, and so the presence of an interfacial layer can have a large effect on the interactions between emulsion droplets, especially at close separations (Vold 1961, Israelachvili 2011). An approximate mathematical expression for the influence of an adsorbed layer on the van der Waals interactions between emulsion droplets has been developed (Vold 1961): Medium 1 Medium 2 h r δ Medium 3 FIGURE 3.5 The droplets in food emulsions are normally surrounded by an interfacial layer (medium 3), which modifies their van der Waals interactions. 63 Colloidal Interactions wVDW (h) = - æ h ö 1 é æ h + 2¶ ö æ h + ¶ r + ¶ öù , ,1 ÷ + A232 H ç ,1 ÷ + 2 A132 H ç ê A131H ç ú r ÷ø û 12 ë è 2r è 2r ø è 2(r + ¶ ) ø (3.6) where the subscripts 1, 2, and 3 refer to the continuous phase, disperse phase, and interfacial layer, respectively h is the surface-to-surface separation between the outer regions of the interfacial layers δ is the thickness of the interfacial layer H(x,y) is a function given by H ( x, y ) = æ x 2 + xy + x ö y y + 2 + 2 ln ç 2 ÷ x + xy + x x + xy + x + y è x + xy + x + y ø 2 The dependence of the (nonretarded and nonscreened) van der Waals interaction between two emulsion droplets on the thickness and composition of an interfacial layer consisting of a mixture of protein and water was calculated using Equation 3.6 and the physical properties listed in Table 3.1 (Figure 3.6). In the absence of the interfacial layer, the attraction between the droplets was about −110 kT at a separation of 1 nm. Figure 3.6 clearly shows that the presence of an interfacial layer causes a significant alteration in the strength of the interactions between the droplets, leading to either an increase or decrease in the strength of the van der Waals attraction depending on the concentration of protein in the interfacial layer. At high protein concentrations (>60%), the attraction is greater than that between two bare emulsion droplets, whereas at low protein concentrations (<60%), it is smaller. Thus, if emulsion droplets are coated by a thick interfacial layer that is highly hydrated, then the van der Waals attraction between them may be greatly weakened. 0 –50 0 20 40 60 δ = 10 nm 80 100 h = 1 nm δ = 1 nm w(h)/kT –100 –150 –200 –250 –300 %Protein in interfacial layer FIGURE 3.6 Influence of the composition of an interfacial layer, consisting of different amounts of water and protein, on the van der Waals interactions between emulsion droplets. The interdroplet pair potential is reported at an outer surfaceto-surface separation of 1 nm for 1 μm droplets. The physical properties of the oil, water, and interfacial layer used in the calculations are reported in Table 3.1. 64 Food Emulsions: Principles, Practices, and Techniques 3.3.3 General Features of van der Waals Interactions 1. The interaction between two oil droplets in water (or between two water droplets in oil) is always attractive. 2. The strength of the interaction decreases with droplet separation, and the interaction is fairly long range (w ∝ 1/h). 3. The interaction becomes stronger as the droplet size increases. 4. The strength of the interaction depends on the physical properties of the droplets and surrounding liquid (through the Hamaker function). 5. The strength of the interaction depends on the thickness and composition of the interfacial layer. 6. The strength of the interaction decreases as the concentration of electrolyte in an oil-in-water emulsion increases because of electrostatic screening. Van der Waals interactions act between all types of colloidal particle, and therefore, they must always be considered when calculating the overall interaction potential between emulsion droplets. Nevertheless, it must be stressed that an accurate calculation of their magnitude and range is extremely difficult, because of the lack of physicochemical data required to perform the calculations, and because of the need to simultaneously account for the effects of screening, retardation, and interfacial layers. The fact that van der Waals interactions are relatively strong and long range, and that they are always attractive, suggests that emulsion droplets will always tend to associate with each other in the absence of any repulsive interactions. In practice, many food emulsions are stable to droplet aggregation, which suggests the existence of repulsive interactions that are strong enough to overcome the van der Waals attraction. In the following sections, we discuss some of the most important types of these repulsive interactions, including electrostatic, steric, hydration, and thermal fluctuation interactions. 3.4 Electrostatic Interactions 3.4.1 Origins of Electrostatic Interactions The droplets in most food emulsions have an appreciable electrical charge (Section 5.4), and therefore, electrostatic interactions often play an important role in determining their overall stability and physicochemical properties. The magnitude and sign of this charge mainly depends on the type of emulsifier used to stabilize the emulsion, the concentration of the emulsifier at the interface, and environmental conditions (e.g., pH, temperature, and ionic strength). All the droplets in an emulsion are usually stabilized by the same type of emulsifier and therefore have the same electrical charge. The electrostatic interaction between similarly charged droplets is repulsive, and so electrostatic interactions play a major role in preventing droplets from coming close enough together to aggregate. The electrostatic interactions between emulsion droplets depend on the electrical characteristics of the droplet surfaces and the ionic composition of the surrounding aqueous phase (Hunter 1986, Hiemenz and Rajagopalan 1997, Evans and Wennerstrom 1999, Norde 2011). The electrical properties of a surface are usually characterized by the surface charge density (σ) and the electrical surface potential (Ψ0). The surface charge density is the amount of electrical charge per unit surface area, whereas the surface potential is the free energy required to increase the surface charge density from zero to σ (Chapter 5). These values depend on the type and concentration of emulsifier present at the droplet surfaces, as well as the properties of the surrounding liquid, for example, pH, ionic composition, dielectric constant, and temperature. 3.4.2 Modeling Electrostatic Interactions 3.4.2.1 Interdroplet Pair Potential In foods, we are primarily interested in the strength of the electrostatic interactions between oil droplets dispersed within an aqueous continuous phase (i.e., oil-in-water emulsions), and so these systems will 65 Colloidal Interactions + – + + – + + – – – + + – + + + – – – + + – – + + – – + – + – + – + + – + – – + – – – – + + – – + – + + – + + + – + + – – – + + – – – + – – – + + – + + + – – + + – – – + – – + + + – – + + – + – FIGURE 3.7 Emulsion droplets can be considered to be surrounded by clouds of counterions, whose thickness is determined by the distance that has to be moved into the electrolyte solution before the charge of the counterions completely neutralizes the charge on the droplet surface. The thickness of this layer is therefore related to the Debye screening length. When the clouds overlap, there is a strong electrostatic repulsion. be the main focus of this section. An isolated electrically charged droplet is surrounded by a cloud of counterions and co-ions, with the oppositely charged counterions having a higher concentration in the immediate vicinity of the droplet (Figure 3.7). When two similarly charged droplets approach each other, their counterion clouds overlap and this gives rise to a repulsive interaction. The range of the electrostatic interaction is primarily determined by the Debye screening length (κ−1), which is a measure of how far the electrical properties of an interface are sensed within the surrounding solution (Chapter 5): k -1 = e0e R kT e 2 ån z 2 0i i (3.7) where n 0i is the number concentration of ionic species of type i in the bulk electrolyte solution (in molecules per cubic meter) zi is their valency e is the elementary electrical charge (1.602 × 10 −19 C) ε0 is the dielectric constant of a vacuum εR is the relative dielectric constant of the solution For aqueous solutions at room temperatures, κ−1 ≈ 0.304/√I nm, where I is the ionic strength expressed in moles per liter (Israelachvili 2011). 66 Food Emulsions: Principles, Practices, and Techniques The electrostatic repulsion between two similarly charged droplets in close proximity can be divided into two contributions: (1) an enthalpy contribution associated with the change in the strength of the attractive and repulsive electrostatic interactions between the various charged species involved and (2) an entropy contribution associated with the confinement of the counterions between the droplets to a smaller volume. The entropy contribution is strongly repulsive, whereas the enthalpy contribution is weakly attractive, and therefore, the overall interaction is repulsive (Evans and Wennerstrom 1999). Mathematical models have been developed to relate the electrostatic interdroplet pair potential to the physical characteristics of the emulsion droplets and the intervening electrolyte solution (Hiemenz and Rajagopalan 1997, Bishop et al. 2009, Israelachvili 2011). Analytical equations based on these models can be derived by making some simplifying assumptions. For example, if it is assumed that there is a relatively low surface potential (|Ψ| < 25 mV) and that the Debye screening length and surface-to-surface separation are much less than the droplet size (i.e., κ−1 < r/10 and h < r/10), then fairly simple expressions for the electrostatic interdroplet pair potential between two similar droplets can be derived (Hunter 1986). The equation that is applicable for a particular system depends on whether the interface can be treated as having constant potential or constant charge density when the droplets approach each other (see Section 3.4.2.2): At constant surface potential, y welectrostatic (h) = 2pe0 e Rry d2 ln(1 + exp(- kh)) (3.8) s welectrostatic (h) = -2pe0 e Rry 2d ln(1 - exp(- kh)) (3.9) At constant surface charge, The smallest droplets in most food emulsions are about 100 nm in radius, which means that these equations are likely to be applicable at droplet separations less than about 10 nm, and at electrolyte concentrations greater than about 1 mM. Whether the electrostatic interaction between two droplets takes place under conditions of constant surface potential or constant surface charge depends on the ability of the surface groups to regulate their charge (Israelachvili 2011). 3.4.2.2 Factors Influencing Electrical Characteristics of Surfaces Charge regulation: As two similarly charged emulsion droplets move closer together, the interaction between them becomes increasingly repulsive (Figure 3.8). Certain systems are capable of reducing the magnitude of this increase by undergoing some form of reorganization of the molecular species present, which is referred to as charge regulation. For example, the surface charge may be regulated by adsorption–desorption of ionic emulsifiers or by association–dissociation of charged groups. Depending on the physical characteristics of a system, it is possible to discern three different situations that may occur when two droplets approach each other (Israelachvili 2011): 1. Constant surface charge: As the droplets move closer together, the number of charges per unit surface area remains constant, that is, no adsorption–desorption or association–dissociation of ions occurs. In this case, the electrostatic repulsion between the surfaces is at the maximum possible value because the surfaces are fully charged. 2. Constant surface potential: As the droplets move closer together, the number of charges per unit surface area decreases so as to keep the surface potential constant, for example, by an ion adsorption–desorption or association–dissociation mechanism. In this case, the electrostatic repulsion between the surfaces is at the minimum possible value because the surface charge density is reduced as the droplets move closer together. 3. Intermediate situation: In reality, the electrostatic repulsion usually falls somewhere between the two extremes mentioned earlier. The number of charges per unit surface area depends on the characteristics of the adsorption–desorption or association–dissociation mechanisms, 67 Colloidal Interactions 400 10 mM Constant charge 350 300 w(h)/kT 250 200 150 Constant potential 100 50 0 0 5 10 15 20 25 h (nm) FIGURE 3.8 Comparison of electrostatic interaction between a pair of emulsion droplets under conditions of constant surface charge and constant surface potential. for example, the surface activity of an ionic emulsifier or the pKa value of an ionizable surface group. These processes take a finite time to occur, and therefore, the surface charge density may also depend on the speed at which the droplets come together. The variation of the interdroplet pair potential with separation is shown for two similarly charged droplets under constant charge and constant potential conditions in Figure 3.8. In general, there is a strong repulsive interaction between the droplets at close separations, which decreases as the droplets move further apart. This repulsive interaction is often sufficiently strong and long range to prevent droplets from aggregating. At relatively large droplet separations, Equations 3.8 and 3.9 give approximately the same predictions for the electrostatic interaction, but at closer separations, the assumption of constant surface charge predicts a significantly higher repulsion than the assumption of constant surface potential. In practice, the interdroplet pair potential always lies somewhere between these two extremes and depends on the precise nature of the system. Ion binding: The surface charge density of emulsion droplets is often influenced by adsorption of ionic substances present in the continuous or dispersed phases to the droplet surfaces, for example, ionic surfactants (e.g., phospholipids, fatty acids, or small-molecule surfactants), multivalent mineral ions (e.g., Ca2+, Cu2+, and Fe3+), charged biopolymers (e.g., proteins or polysaccharides), or charged nanoparticles (e.g., silica or titanium dioxide). The primary driving force for the adsorption of these surface-active ions to charged surfaces is usually either hydrophobic or electrostatic attraction depending on ion type. For example, ions with an appreciable number of nonpolar groups (such as ionic surfactants, phospholipids, or proteins) are mainly driven to adsorb through hydrophobic interactions, whereas ions that are primarily hydrophilic (such as mineral ions and many polysaccharides) are mainly driven by electrostatic interactions. The contribution of adsorbed ions to surface charge is governed mainly by the type and concentration of surface-active ions present in the system, and their relative affinities for the droplet surface. Depending on the nature of the ionic substances involved, ion adsorption can either decrease or increase the magnitude of the electrical charge, and under some circumstances, it may even lead to charge reversal (Section 5.4). 68 Food Emulsions: Principles, Practices, and Techniques Ionic strength: When the electrostatic interaction between a charged surface and the counterions is relatively weak, the surface charge density is simply related to the surface potential: σ = εRε0κΨδ (Evans and Wennerstrom 1999). This equation indicates that the electrical properties of a surface are altered by the presence of electrolyte in the aqueous phase, and has important consequences for the calculation of the electrostatic interdroplet pair potential. If it is assumed that the surface charge density remains constant when salt is added to the aqueous phase, then the surface potential decreases (because less energy is needed to bring a charge from infinity to the droplet surface through an electrolyte solution due to screening effects). Conversely, if the electrical potential remains constant as the salt concentration is increased, this means that the surface charge density must increase. In general, both σ and Ψδ tend to change simultaneously. In food emulsions, one can usually assume that the surface charge density is independent of ionic strength at low to moderate electrolyte concentrations for monovalent counterions, and so one must take into account the variation of Ψδ with ionic strength when calculating the electrostatic repulsion between emulsion droplets (Kulmyrzaev et al. 2000a,b). Heterogeneous charge distribution: So far it has been assumed that the charge on the droplets is evenly spread out over the whole of the surface. In practice, droplets may have surfaces that have some regions that are negatively charged, some regions that are positively charged, and some regions that are neutral. The heterogeneous distribution of the charges on a droplet influences their electrostatic interactions (Miklavic et al. 1994). Thus, two droplets or molecules that have no net charge may still be electrostatically attracted to each other if they have patches of positive and negative charges on their surfaces (Kayitmazer et al. 2013). Similarly, an electrically charged polymer may adsorb onto a droplet surface with the same net charge if the droplet surface has a heterogeneous distribution of positive and negative charges (Dickinson 2003). 3.4.2.3 Influence of Ionic Strength on the Magnitude and Range of Interactions The magnitude and range of the electrostatic repulsion between two droplets decrease as the ionic strength of the solution separating them increases because of electrostatic screening, that is, the accumulation of counterions around the surfaces (Figure 3.9). Electrostatic screening effects become more pronounced as the concentration and valency of the counterions in the solution surrounding the emulsion droplets increase (Israelachvili 2011). The Debye screening length (κ−1) provides a good estimate of the 300 – – + + + – – + – w(h)/kT – + + + – + + – – + – 250 – + 200 + – + + – + – – + – 150 100 1 mM 10 mM 50 1000 mM 0 100 mM 0 5 10 15 20 h (nm) 25 30 35 FIGURE 3.9 The presence of electrolytes (such as salts) reduces the magnitude and range of the electrostatic repulsion between emulsion droplets due to electrostatic screening effects. 69 Colloidal Interactions + + + + + + + + + + + – – – + – – – + + + + FIGURE 3.10 Polyvalent ions are capable of forming ion bridges between emulsion droplets, thereby leading to bridging flocculation. influence of electrostatic screening effects on the range of electrostatic interactions since it corresponds to the distance where the electrical potential falls to 1/e of its value at droplet contact (h = 0). Multivalent ions are much more effective at screening electrostatic interactions than monovalent ions (Equation 3.7). Consequently, much smaller concentrations of multivalent ions are required to promote emulsion instability (Hunter 1986, Keowmaneechai and McClements 2002). This has important consequences for the texture, stability, appearance, and gastrointestinal fate of many food emulsions and explains the susceptibility of electrostatically stabilized emulsions to flocculation when the electrolyte concentration is increased above a critical level (Chapter 7). 3.4.2.4 Influence of Ion Bridging on Electrostatic Interactions Ion bridging is another type of colloidal interaction that involves electrostatic interactions (Dickinson 2003, Guzey and McClements 2006). It occurs when a polyvalent ion simultaneously binds to the surface of two or more emulsion droplets that have an opposite charge to the ion (Figure 3.10). These polyvalent ions may be low molecular weight species (e.g., Ca2+, Mg2+, or Al3+) or high molecular weight species (e.g., polysaccharides or proteins). The tendency for ion bridging to occur depends on the strength of the electrostatic interaction linking the polyvalent ion to the droplets compared to the strength of the electrostatic repulsion between the similarly charged droplets. For this reason, large polyvalent species, such as ionic polysaccharides, are often particularly effective at forming ion bridges because they are able to act as a bridge between the droplets without allowing them to get too close together. The ability of polyvalent ions to form ion bridges is superimposed on their ability to modulate the electrostatic repulsion between droplets through ion binding and charge screening effects mentioned earlier. 3.4.3 General Characteristics of Electrostatic Interactions 1. Electrostatic interactions may be either attractive or repulsive depending on the sign of the charges on the droplets. The interaction is repulsive when droplets have similar charges (which is usually the case), but is attractive when they have opposite charges. 2. The strength of the interaction decreases with droplet separation, and may be either long or short range depending on the ionic strength and dielectric constant of the electrolyte solution surrounding the droplets (Figure 3.9). The interaction becomes increasingly short range as the ionic strength increases because of electrostatic screening. 3. The strength of the interaction is proportional to the size of the emulsion droplets. 4. The strength of the interaction depends on the electrical characteristics of the droplet surfaces, that is, the number of emulsifier molecules adsorbed per unit surface area and the number of ionized groups per emulsifier molecule. The fraction of ionized groups present depends on their pKa values relative to the pH of the surrounding solution. 70 Food Emulsions: Principles, Practices, and Techniques 5. The interaction becomes more difficult to predict when charge regulation occurs (e.g., due to association–dissociation of ionizable groups or adsorption–desorption of ionic emulsifiers), especially at close droplet separations. 6. Ion binding and bridging effects have to be taken into account when polyvalent ions are involved. Ion binding may alter the surface charge density and surface potential of a droplet, whereas ion bridging may also cause droplets to be linked together. In this section we have seen that under certain conditions, repulsive electrostatic interactions may be relatively strong and long range compared to attractive van der Waals interactions (compare Figures 3.3 and 3.8). This suggests that they may be strong enough to prevent droplets from aggregating in certain systems. Indeed, it is widely recognized that electrostatic stabilization plays an important role in determining the aggregation of droplets in many food emulsions, and particularly those stabilized by proteins (McClements 2004). Electrostatic interactions also influence numerous other physicochemical phenomena in food emulsions that may alter their overall properties, such as the partitioning of ingredients, the rates of chemical reactions, and the interactions of emulsion droplets with surfaces (Chapters 7 through 10). The partitioning of ionized volatile flavor compounds between the head space and bulk of emulsions is influenced by electrostatic interactions between flavor molecules and electrically charged interfaces (Guyot et al. 1996). The oxidation of lipids in food emulsions is often catalyzed by polyvalent ions, such as Fe3+, that are normally present in the aqueous phase (McClements and Decker 2000). The rate of ironcatalyzed lipid oxidation in oil-in-water emulsions has been shown to increase when the droplets have a negative charge because the Fe3+ catalyst and oil molecules are brought into close contact (Mei et al. 1998, 1999). The interaction of emulsion droplets with bacterial surfaces also depends on their charges (Ziani et al. 2011). Knowledge of the factors that determine the magnitude and range of electrostatic interactions is therefore extremely important to food scientists for a variety of reasons. 3.5 Steric Interactions 3.5.1 Origin of Steric Interactions The droplets in most food emulsions are coated by an interfacial layer comprising of emulsifier molecules (Figure 3.11), such as surfactants, phospholipids, proteins, polysaccharides, or their complexes (Chapter 4). In addition, the thickness and properties of the interfacial layers can be increased after they have been formed by coating them with other substances, for example, proteins or polysaccharides (Guzey and McClements 2006, Dickinson 2008). When two droplets approach each other sufficiently closely, then their interfacial layers start to overlap and interact with each other (Figure 3.12). Steric interactions are a result of the intermingling and/or compression of the interfacial layers. At close droplet separations, steric interactions are strongly repulsive and may therefore prevent emulsion droplets from aggregating. Nevertheless, the overall magnitude, sign, and range of steric interactions are strongly dependent on the characteristics of the interfacial layers (e.g., thickness, packing, rheology, and molecular interactions) and therefore depend on emulsifier type. An understanding of the relationship between Water Oil (a) Oil (b) Oil (c) Oil (d) FIGURE 3.11 Structural organization of some emulsifiers and biopolymers at an oil–water interface: (a) small-molecule surfactants, (b) flexible biopolymers, (c) globular biopolymers, and (d) mixed layers. 71 Colloidal Interactions (a) (b) FIGURE 3.12 Steric interactions between emulsion droplets can be divided into (a) a mixing contribution that involves interpenetration of the polymer chains and (b) an elastic contribution that involves compression of the polymer layers. the properties of interfacial layers and the stability of emulsions is important for food scientists because it facilitates the rational selection of emulsifiers with optimum performance for each application. 3.5.2 Modeling Steric Interactions 3.5.2.1 Interdroplet Pair Potential As mentioned earlier, steric interactions arise when emulsion droplets get so close together that the interfacial layers overlap (Figure 3.12). This type of interaction can be conveniently divided into two contributions (Hunter 1986, Hiemenz and Rajagopalan 1997): wsteric(h) = welastic(h) + wmix(h) (3.10) The elastic contribution is due to the compression of the interfacial layers, whereas the mixing contribution is due to the intermingling of the emulsifier molecules within the interfacial layers when they overlap (Figure 3.12). 3.5.2.2 Mixing Contribution If it is assumed that the emulsifier molecules within the interfacial layers interpenetrate without the layers being compressed (Figure 3.12a), then the interaction is entirely due to mixing of the emulsifier molecules. The theories describing steric interactions are much less well developed than those describing electrostatic or van der Waals interactions, primarily because they are particularly sensitive to the precise structure, orientation, packing, and interactions of the emulsifier molecules within 72 Food Emulsions: Principles, Practices, and Techniques the interfacial layers. These parameters vary from system to system and are difficult to account for theoretically or to measure experimentally. Mathematical theories have been developed for a number of simple well-defined systems, and it is informative to examine these because they provide some useful insights into more complex systems (Hunter 1986). Consider a system that consists of polymeric emulsifier molecules that are permanently attached to the droplet surfaces, with a constant number of polymer chains per unit surface area. A mathematical analysis of the interactions that occur between the polymer chains when the interfacial layers approach each other leads to the following equation for the mixing contribution: wmix (h) = 4prkTm 2 N A v P2 1 æ 1 hö ( - c) ç 1 ÷ VS 2 è 2 dø 2 (3.11) where m is the mass of polymer chains per unit area δ is the thickness of the adsorbed layer NA is Avogadro’s number χ is the Flory–Huggins parameter vP is the partial specific volume of the polymer chains VS is the molar volume of the solvent The Flory–Huggins parameter depends on the relative magnitude of the solvent–solvent, solvent– segment, and segment–segment interactions and is a measure of the quality of a solvent. It is related to the effective interaction parameter (w), which was introduced in Chapter 2 to characterize the compatibility of molecules in mixtures: χ = w/RT. In a good solvent (χ < 1/2), the polymer molecules prefer to be surrounded by solvent molecules. In a poor solvent (χ > 1/2), the polymer molecules prefer to be surrounded by each other. In an indifferent (theta) solvent (χ = 1/2), the polymer molecules have 20,000 Repulsion 15,000 w(h)/kT 10,000 Good solvent 5,000 Neutral solvent 0 Poor solvent –5,000 –10,000 Attraction 0 2 4 6 8 10 12 h (nm) FIGURE 3.13 Interdroplet pair potential due to steric polymeric interactions. At intermediate separations, the steric polymeric interaction can be either attractive or repulsive depending on the quality of the solvent because of the mixing contribution, but at short separations, it is strongly repulsive because of the elastic contribution. 73 Colloidal Interactions no preference for either solvent or polymer molecules. In the original Flory–Huggins theory, it was assumed that χ was entirely due to enthalpy contributions associated with the molecular interactions. In practice, it is more convenient to assume that χ also contains entropy contributions, since then interactions involving changes in the structural organization of the solvent can be accounted for, for example, hydrophobic interactions (Evans and Wennerstrom 1999, Norde 2011). Whether the mixing contribution is attractive or repulsive depends on the quality of the solvent (Figure 3.13). In a good solvent, the increase in concentration of polymer molecules in the interpenetration zone is thermodynamically unfavorable (wmix positive) because it reduces the number of polymer–solvent contacts and therefore leads to a repulsive interaction between the droplets. Conversely, in a poor solvent, it is thermodynamically favorable (wmix negative) because it increases the number of polymer–polymer contacts and therefore leads to an attractive interaction between the droplets. In an indifferent solvent, the polymer molecules have no preference as to whether they are surrounded by solvent or by other polymer molecules, and therefore, the mixing contribution is zero. Thus, by altering solvent quality, it is possible to change the mixing contribution from attractive to repulsive, or vice versa. In food emulsions, this could be done by changing the solvent quality for the polymer chains, for example, by altering the temperature or adding alcohol to the aqueous phase. 3.5.2.3 Elastic Contribution If it is assumed that the interfacial layers surrounding the emulsion droplets are compressed without any interpenetration of the emulsifier molecules (Figure 3.12b), then the interaction is entirely elastic. When the interfacial layers are compressed, a smaller volume is available for the emulsifier molecules to occupy, and therefore, their configurational entropy is reduced, which is thermodynamically unfavorable, and so this type of interaction is always repulsive (welastic positive). For certain simple systems, the magnitude of the elastic contribution can be calculated from a statistical analysis of the number of configurations that the emulsifier molecules can adopt before and after the layers are compressed (Hiemenz and Rajagopalan 1997). Nevertheless, it is usually not possible to carry out such an analysis for most real systems because of the structural complexity of the interfacial layers. For these systems, it is often better to use semiempirical models to account for the elastic interaction. For example, a simple exponential model has been proposed to describe the steric repulsion between emulsion droplets stabilized by polymeric emulsifiers: welastic(h)/kT = A Eexp(−πh/L), where L is the polymer chain length and A E is a parameter that depends on polymer chain length, chain packing, and droplet size (Quemada and Berli 2002). An alternative semiempirical expression was developed by considering the force required to compress interfacial layers with specified rheological characteristics (Jackel 1964): 1 ö æ1 welastic (h) = 0.77E ç d - h ÷ 2 ø è2 welastic (h) = 0 5/ 2 (r + d) (h < d) (3.12) (h  d) where E is the elastic modulus of the interfacial layer. This equation indicates that there is a negligible interaction between the droplets when the separation is greater than the thickness of the interfacial layer (δ), but that there is a steep increase in the interaction energy when the droplets approach closer than this distance (Figure 3.13). As a first approximation, it may be possible to use measurements of the elastic modulus of macroscopic solutions and gels formed using emulsifier concentrations similar to those found in the interfacial region in Equation 3.12. Alternatively, the elastic modulus of an interfacial layer could be measured directly using various types of surface force apparatus (Claesson et al. 1996, Butt et al. 2005). To a first approximation, the steric repulsion between a pair of emulsion droplets can simply be described using a similar approach to that used for the steric repulsion between a pair of molecules (Chapter 2): 74 Food Emulsions: Principles, Practices, and Techniques 12 æ 2d ö welastic (h) = ç ÷ kT è h ø (h < d) (3.13) welastic (h) = 0 (h  d) where δ is the thickness of the interfacial layers kT is the thermal energy This equation is particularly useful for roughly estimating the colloidal interactions between emulsion droplets because it only requires information about interfacial thickness. However, it does not account for any molecular interactions between the overlapping interfacial layers in the interpenetration regime (see Section 3.5.2.4). 3.5.2.4 Distance Dependence of Steric Interactions Steric interactions between emulsion droplets can be conveniently divided into three regimes, according to the separation of the surface of the bare droplets h relative to the thickness of the interfacial layers δ (Figure 3.13) 1. Zero interaction regime (h ≥ 2δ). At sufficiently large droplet separations, the interfacial layers do not overlap with each other and the steric interaction between the droplets is zero. 2. Interpenetration regime (δ ≤ h < 2δ). When the droplets are sufficiently close together for their interfacial layers to overlap with each other, so that the emulsifier molecules on different droplets can intermingle, but the interfacial layers are not significantly compressed, then the major contribution to the steric interaction is the mixing contribution (wmix). This contribution may be either repulsive (positive) or attractive (negative) depending on the solvent quality. 3. Interpenetration and compression regime (h < δ). When the droplets get so close together that the interfacial layers start to compress each other, the overall steric interaction is a combination of elastic and mixing contributions, although the strongly repulsive elastic component usually dominates, and so the overall interaction is repulsive. It should be stressed that the length of the interpenetration and elastic regions actually depends on the precise nature of the emulsifier molecules within the interfacial layers. Flexible biopolymer molecules will have relatively large interpenetration regions, whereas compact globular proteins will have relatively small ones. Consequently, the choice of δ as the distance where the elastic contribution first contributes to the interaction is rather arbitrary, and different values will be more appropriate for some systems. As mentioned earlier, the only way these values can accurately be established is by measuring the force between two well-defined surfaces (with interfacial layers similar to those around emulsion droplets) as they are brought closer together (Claesson et al. 1996, Butt et al. 2005, Israelachvili 2011). 3.5.2.5 Optimum Characteristics of Steric Stabilizers To be effective at providing steric stabilization, an emulsifier must have certain physicochemical characteristics (Dickinson 1992). Firstly, it must have some segments that bind strongly to the droplet surfaces (to anchor the emulsifier molecules to the surface) and other segments that protrude a significant distance into the surrounding liquid (to prevent the droplets from coming close together). This means that the emulsifier must be amphiphilic, having some hydrophobic segments that protrude into the oil phase, and some hydrophilic segments that protrude into the aqueous phase (Figure 3.11). The binding to the interface must be strong enough to prevent the emulsifier from desorbing from the droplet surface as the droplets approach one another. Theoretical calculations suggest that a good polymeric stabilizer should 75 Colloidal Interactions have 10%–20% of the molecule that adsorbs to the droplet surfaces, and 80%–90% of the molecule that protrudes into the continuous phase (Claesson et al. 2004). Secondly, the continuous phase surrounding the droplets must be a sufficiently good solvent for the segments of the emulsifier molecules that protrude into it, so that the mixing contribution to the overall interaction energy is repulsive (wmix positive). Thirdly, the steric repulsive interaction must act over a distance that is comparable to the range of the attractive van der Waals interactions. Thus, emulsifiers that form thick interfacial layers (such as modified starch or gum arabic) are much more effective at stabilizing emulsions against flocculation through steric repulsion than emulsifiers that form thin layers (such as globular proteins at their isoelectric point*). Finally, the surface must be covered by a sufficiently high concentration of emulsifier. If too little of a polymeric emulsifier is present, a single emulsifier molecule may adsorb onto the surface of more than one emulsion droplet, forming a bridge that causes the droplets to flocculate. In addition, some of the nonpolar regions will be exposed to the aqueous phase, which leads to a hydrophobic attraction between the droplets (see Section 3.7). Many emulsifiers are charged, and therefore, they stabilize emulsion droplets against aggregation through a combination of electrostatic and steric repulsion (Dickinson 2003, Claesson et al. 2004). 3.5.3 General Characteristics of Steric Interactions 1. Steric interactions are always strongly repulsive at short separations (h < δ), but may be either attractive or repulsive at intermediate separations (δ < h < 2δ), depending on the quality of the solvent (Figure 3.13). 2. The range of steric interactions increases with the thickness of the adsorbed layer. 3. The strength of steric interactions increases with droplet size. 4. The strength of steric interactions depends on the precise molecular characteristics of the interfacial layer, for example, packing, flexibility, rheology, and molecular interactions. Consequently, it varies considerably from system to system and is difficult to predict from first principles. Steric interactions are one of the most common and important stabilizing mechanisms in food emulsions. Unlike electrostatic interactions, they occur in almost every type of food emulsion because most droplets are stabilized by a layer of adsorbed emulsifier molecules. Some food emulsions are stabilized almost entirely by steric stabilization, whereas others are mainly stabilized by a combination of steric and electrostatic stabilization. Food scientists must often decide the most appropriate emulsifier for a particular application, and so it is useful to compare the differences between steric and electrostatic stabilization (Table 3.2). The principal difference is their sensitivity to pH and ionic strength. TABLE 3.2 Comparison of the Advantages and Disadvantages of Polymeric and Electrostatic Stabilization Mechanisms in Food Emulsions Polymeric Steric Stabilization Insensitive to pH Insensitive to electrolyte Large amounts of emulsifier needed to cover droplet surface Weak flocculation (easily reversible) Good freeze–thaw stability Electrostatic Stabilization pH dependent: aggregation tends to occur when emulsifier loses charge. Aggregation tends to occur at high electrolyte concentrations (>CFC). Small amounts of emulsifier needed to cover droplet surface. Strong flocculation (often irreversible) Poor freeze–thaw stability. Source: Hunter, R.L., Foundations of Colloid Science, Oxford University Press, Oxford, U.K., 1986. * Thick biopolymer layers may also help stabilize droplets against aggregation by reducing the magnitude of the attractive van der Waals interaction (Section 3.3.6). 76 Food Emulsions: Principles, Practices, and Techniques The electrostatic repulsion between emulsion droplets is dramatically decreased when the electrical charge on the droplet surfaces is reduced (e.g., by altering the pH) or screened (e.g., by increasing the concentration of electrolyte in the aqueous phase). In contrast, steric repulsion is fairly insensitive to both electrolyte concentration and pH.* Another major difference is the fact that the electrostatic repulsion is usually weaker than the van der Waals attraction at short distances, whereas the steric stabilization is stronger. This means that emulsions stabilized entirely by electrostatic repulsion are prone to coalescence when the droplets approach sufficiently closely, whereas emulsions stabilized by steric interactions may flocculate, but they are unlikely to coalesce because of the extremely strong short-range repulsion. From a practical standpoint, another important difference is the fact that considerably more emulsifier is usually required to provide steric stabilization (because a thick interfacial layer is required) than to provide electrostatic stabilization. Thus, an emulsifier-to-oil ratio of >1:4 is required when gum arabic is used as an emulsifier, whereas <1:10 is required when whey protein is used (Chanamai et al. 2002, Charoen et al. 2011). The amount of emulsifier required to prepare an emulsion is often an important financial consideration when formulating a food product. 3.6 Depletion Interactions 3.6.1 Origin of Depletion Interactions Many food emulsions contain small colloidal particles that are dispersed in the continuous phase that surrounds the droplets (Figure 3.14). These nonadsorbed colloidal particles may be surfactant micelles, polymer molecules, molecular clusters, or solid particles. The presence of these colloidal particles causes an attractive interaction between the droplets that is often large enough to promote emulsion instability (Dickinson 2003, Guzey and McClements 2006). The origin of this interaction is the exclusion of colloidal particles from a narrow region surrounding each droplet (Figure 3.14). This region Droplet Examples: Exclusion zone Depletion flocculation Colloidal particle Solid particles Molecular clusters Hydrated polymers Surfactant micelles FIGURE 3.14 An attractive depletion interaction arises between emulsion droplets when they are surrounded by small nonadsorbing colloidal particles, such as surfactant micelles, biopolymers, molecular clusters, or nanoparticles. * It should be stressed that the polymeric steric interaction may be effected by pH and ionic strength if the polymer molecules are charged, because this will alter the thickness of the interfacial layer and the interaction of the polymer chains. 77 Colloidal Interactions extends a distance approximately equal to the radius of a colloidal particle away from the droplet surface. The concentration of colloidal particles in this exclusion zone is effectively zero, while it is finite in the surrounding continuous phase. As a consequence, an osmotic pressure is generated that favors the movement of solvent molecules from the exclusion zone into the bulk liquid, so as to dilute the colloidal particles and thus reduce the concentration gradient. The only way this process can be achieved is by two droplets coming into close proximity and thereby reducing the volume of the exclusion zone, which manifests itself as an attractive force between the droplets (Figure 3.14). Thus, there is an osmotic driving force that favors droplet aggregation, and which increases as the concentration of colloidal particles in the aqueous phase increases. 3.6.2 Modeling of Depletion Interactions When the separation between two droplets is small compared to their size (h ≪ rd), the interdroplet pair potential due to exclusion of the nonadsorbing colloidal particles from the exclusion zone is given by the following expression (Sperry 1982): 2 3 æ æ r ö3 æ h ö 2 h öö æ r ö æ wdepletion (h) = - pr 3 P OSM ç 2 ç 1 + c ÷ + ç 1 + ÷ - 3 ç 1 + c ÷ ç 1 + ÷ ÷ ç è 3 r ø è 2r ø r ø è 2r ø ÷ø è è (3.14) where POSM is the osmotic pressure arising from the exclusion of the colloidal particles r is the radius of the emulsion droplets rc is the radius of the colloidal particles This equation is applicable for h < 2rc, for higher droplet separations wdepletion(h) = 0. To a first approximation, the osmotic pressure difference can be described by one of the following equivalent expressions (Hiemenz and Rajagopalan 1997): P OSM = kTni (1 + 2ni v) (3.15) P OSM = cRT æ 2N cv ö 1+ A ÷ M çè M ø (3.16) P OSM = cRT æ 2cRV ö ç 1+ ÷ M è rc ø (3.17) where c, M, v, ni, and ρc are the concentration (in kg m−3), molecular weight (in kg mol−1), effective volume (in m3), number density (in m−3), and mass density (in kg m−3) of the colloidal particles NA is Avogadro’s number Equations 3.15 and 3.16 are applicable to nonadsorbing particles of any type. Equation 3.17 is suitable for nonadsorbing polymers, where ρc is the mass density of the polymer backbone, rather than of the overall polymer and trapped liquid. The parameter, Rv, is called the volume ratio and is equal to the effective volume of a nonadsorbing colloidal particle divided by the actual volume of the constituent atoms added to the system (McClements 2000). For compact spherical colloidal particles, such as solid spheres, surfactant micelles, or globular proteins, Rv ≈ 1. However, for asymmetrical solid particles or biopolymers that entrain large quantities of solvent as they rotate in solution, Rv ≫ 1 (Chapter 4). This phenomenon has important consequences for the ability of nonadsorbing colloidal particles to promote depletion flocculation in emulsions (see the following text). 78 Food Emulsions: Principles, Practices, and Techniques 0 1 0 2 3 4 1% –5 2% w(h)/kT –10 3% –15 4% –20 –25 –30 h (nm) (a) 0 0 1 1 kg 2 3 4 5 6 mol−1 100 kg mol−1 –5 w(h)/kT –10 –15 10 kg mol−1 –20 –25 –30 (b) h (nm) FIGURE 3.15 (a) Influence of the concentration of the nonadsorbing colloidal particles (shown in wt.% in the annotation) on the attractive depletion interaction between emulsion droplets: rd = 1 μm, rc = 1.36 nm (Rv = 1). (b) Influence of the molecular weight of the nonadsorbing colloidal particles (shown in kg mol−1 in the annotation) on the attractive depletion interaction between emulsion droplets: c = 5 wt.%, rd = 1 μm, rc = 0.63–2.9 nm. It was assumed that the colloidal particles were dense spheres, so that rc ∝ M1/3, and Rv = 1.  (Continued ) 79 Colloidal Interactions 0 –50 0 10 20 30 40 2 kg mol−1 5 kg mol−1 10 kg mol−1 –100 w(h)/kT –150 –200 –250 –300 –350 –400 (c) h (nm) FIGURE 3.15 (Continued) (c) Influence of the molecular weight of the nonadsorbing colloidal particles (shown in kg mol−1 in the annotation) on the attractive depletion interaction between emulsion droplets: c = 5 wt.%, rd = 1 μm, rc = 2.8 – 14 nm (Rv = 44–1100). It was assumed that the colloidal particles were linear rigid rods, so that rc ∝ M, and Rv ∝ M2. In this case, the effective volume fraction of the nonadsorbed colloidal particles is much greater than the actual volume fraction of the polymer chain. The range of the depletion interactions is approximately equal to twice the radius of the nonadsorbed colloidal particles: 2rc. An estimate of the maximum strength of the attractive depletion interaction between two droplets can be obtained by calculating the interdroplet pair potential when the droplets are in contact (i.e., h = 0): é 2 ù wdepletion (h = 0) = -2prc2 POSM êr + rc ú ë 3 û (3.18) Theoretical predictions of the attractive depletion interactions between a pair of emulsion droplets are shown in Figure 3.15. The interaction potential is zero at droplet separations greater than the diameter (2rc) of the nonadsorbing colloidal particles and decreases to a finite negative value (given by Equation 3.18) when the droplets come into close contact. For nonadsorbing colloidal particles of constant molecular weight (or radius), the strength of the interaction increases with increasing particle concentration (Figure 3.15a). If it is assumed that an emulsion contains a constant mass concentration (wt.%) of nonadsorbing colloidal particles that are compact spheres (Rv = 1), then the range of the depletion attraction increases with increasing molecular weight of the particles, but the maximum strength of the interaction decreases with increasing M (Figure 3.15b). More complex behavior can be observed when the effective volume of the nonadsorbing colloidal particles is much greater than the actual volume of the added material (i.e., Rv ≫ 1). For example, the attractive depletion interaction for an emulsion containing a constant mass concentration (wt.%) of nonadsorbing colloidal particles that are assumed to be linear rigid rods is shown in Figure 3.15c. In this case, both the range and maximum strength of the depletion attraction increase with increasing molecular weight, since rc ∝ M and Rv ∝ M2 (hence, POSM can increase with increasing M, Equation 3.17). 80 Food Emulsions: Principles, Practices, and Techniques Depletion interactions rely on the colloidal particles not interacting strongly with the surface of the emulsion droplets. Otherwise, the bound particles would have to be displaced as the droplets moved closer together, which would require the input of energy and therefore be repulsive. 3.6.3 General Characteristics of Depletion Interactions 1. The maximum strength of the depletion interaction increases as the size of the emulsion droplets increases. 2. The maximum strength of the depletion interaction increases as the concentration (ni or c) of nonadsorbing colloidal particles in the continuous phase increases (at constant M or rc). 3. The maximum strength of the interaction may either decrease or increase with increasing molecular weight of the nonadsorbing colloidal particles (at constant ni or c) depending on their volume ratio, Rv. 4. The range of the depletion interaction (2rc) increases as the radius of the colloidal particles increases. Equation 3.14 suggests that the strength of the depletion interaction is independent of pH and ionic strength. Nevertheless, these parameters may indirectly influence the depletion interaction by altering the effective size of the colloidal particles and the depletion zone. For example, changing the number of charges on a biopolymer molecule by altering the pH can either increase or decrease its effective size. Increasing the number of similarly charged groups usually causes a biopolymer to become more extended because of electrostatic repulsion between the charged groups. On the other hand, decreasing the number of similarly charged groups or having a mixture of positively and negatively charged groups usually causes a biopolymer to reduce its effective size. Altering the ionic strength of an aqueous solution also causes changes in the effective size of biopolymer molecules, for example, adding salt to a highly charged biopolymer molecule screens the electrostatic repulsion between charged groups and therefore causes a decrease in biopolymer size. Thus, if the colloidal particles are ionic, one would expect the strength of the depletion interaction to depend on pH and ionic strength, but if they are nonionic, one would expect them to be fairly insensitive to these parameters. It should be noted that at high concentrations of free colloidal particles in the continuous phase, one can actually have depletion stabilization (Hiemenz and Rajagopalan 1997). 3.7 Hydrophobic Interactions 3.7.1 Origin of Hydrophobic Interactions Hydrophobic interactions are believed to play an important role in determining the stability and physicochemical properties of a number of food emulsions. Hydrophobic interactions are important when the surfaces of the droplets have some nonpolar character, either because they are not completely covered by emulsifier (e.g., during homogenization or at low emulsifier concentrations) or because the emulsifier has some hydrophobic regions exposed to the aqueous phase (e.g., denatured globular proteins). For example, fat droplets coated with globular proteins may become flocculated after surface or thermal denaturation of the adsorbed proteins because when the globular proteins unfold, the droplet surfaces become more hydrophobic (McClements 2004, Dickinson 2013). There has been considerable debate about the physicochemical origin of the hydrophobic interactions that act between nonpolar surfaces separated by water (Meyer et al. 2006, Tabor et al. 2014). Experimental measurements of the forces between different kinds of hydrophobic surfaces have shown that the force–distance curves can be highly irreproducible and that they vary greatly from system to system. It appears that the various types of behaviors observed in these experiments can be classified into two groups: (1) a long-range strong irreproducible attractive force and (2) a short-range weaker reproducible attractive force (Christenson and Claesson 2001, Attard 2003). The origin of the long-range interaction appears to be highly system dependent and has been attributed to various physicochemical 81 Colloidal Interactions mechanisms, such as the presence of undissolved gas (nanobubbles) that adhere to nonpolar surfaces and the generation of electrostatic interactions due to molecular arrangements at the surfaces (Meyer et al. 2006). The short-range interaction is believed to be the authentic hydrophobic interaction associated with the structuring of water molecules around nonpolar groups. The short-range hydrophobic attraction is considerably stronger than the van der Waals attraction (Attard 2003), and therefore plays an important role in determining the stability of emulsion droplets to aggregation. The molecular origin of the short-range hydrophobic attraction between droplets can be attributed to the ability of water molecules to form relatively strong hydrogen bonds with each other but not with nonpolar molecules (Chapter 4). Consequently, the interaction between nonpolar substances and water is thermodynamically unfavorable, which means that a system will attempt to minimize the contact area between these substances by causing them to associate (Israelachvili 2011, Tabor et al. 2013). This process manifests itself as a relatively strong attractive force between hydrophobic substances dispersed in water and is responsible for many important phenomenon that occur in food emulsions, such as protein conformation and aggregation, micelle formation, adsorption of surfactants to interfaces, and the low water solubility of nonpolar compounds (Chapters 4 and 5). In this section, we only consider the short-range hydrophobic attraction since this is considered to be the true hydrophobic interaction (Meyer et al. 2006). 3.7.2 Modeling Hydrophobic Interactions The interdroplet pair potential between two emulsion droplets with hydrophobic surfaces separated by water can be represented as an exponential decay (Israelachvili 2011): whydrophobic (h) = -2prg i fH l 0e - h / l0 (3.19) where γi is the interfacial tension between the nonpolar groups and water (typically around 50 mJ m−2) λ 0 is the decay length of the interaction (typically between 1 and 2 nm) ϕH is a measure of the surface hydrophobicity of the droplets The surface hydrophobicity varies from 0 (for a fully polar surface) to 1 (for a fully nonpolar surface). Thus, the magnitude of the hydrophobic interaction increases as the surface hydrophobicity increases, that is, ϕH tends toward unity (Figure 3.16). It should be noted that the aforementioned equation is actually an oversimplification and the actual short-range hydrophobic attraction will depend on the precise nature of the system involved. For example, experiments have shown that the hydrophobic interaction is not directly proportional to the number of nonpolar groups at a surface, because the alteration in water structure imposed by nonpolar groups is disrupted by the presence of any neighboring polar groups (Israelachvili 2011). Thus, it is not possible to assume that ϕ is simply equal to the fraction of nonpolar sites at a surface. As a consequence, it is difficult to accurately predict their magnitude from first principles. Measurements of the force versus distance profile of nonpolar surfaces have shown that the hydrophobic attraction is stronger than the van der Waals attraction up to relatively high surface separations (Claesson et al. 2004, Israelachvili 2011). When hydrophobic surfaces are covered by amphiphilic molecules, such as small-molecule surfactants, phospholipids, or biopolymers, the hydrophobic interaction between them is effectively screened and the overall attraction is mainly due to van der Waals interactions (Israelachvili 2011). Nevertheless, hydrophobic interactions are likely to be significant when the surface has some hydrophobic character, for example, if the surface is not completely saturated with emulsifier, if it is bent to expose the underlying oil molecules, or if the emulsifier molecules have some hydrophobic regions exposed to the aqueous phase, for example, denatured globular proteins. 3.7.3 General Characteristics of Hydrophobic Interactions Few studies have been carried out to systematically characterize the influence of environmental and solution conditions on the strength and range of hydrophobic interactions between emulsion droplets. 82 Food Emulsions: Principles, Practices, and Techniques 0 0 2 4 –50 6 8 10 12 φH = 25% –100 w(h)/kT –150 –200 50% –250 75% –300 –350 100% –400 h (nm) FIGURE 3.16 An attractive hydrophobic interaction arises between emulsion droplets when their surfaces have some hydrophobic character, where ϕH is approximately equal to the fraction or percentage of the droplet surface that is nonpolar. Nevertheless, it is possible to gain some insight into the factors that would be expected to influence interdroplet hydrophobic interactions by examining the factors that influence the strength of intermolecular hydrophobic interactions. Intermolecular hydrophobic interactions become increasingly strong as the temperature is raised (Israelachvili 2011). Thus, hydrophobic interactions between emulsion droplets should become stronger at higher temperatures. Because the strength of hydrophobic interactions depends on the magnitude of the interfacial tension, then any change in the properties of the solvent that increases the interfacial tension will increase the hydrophobic attraction, or vice versa. The addition of small amounts of alcohol to the aqueous phase of an emulsion lowers γi and therefore would be expected to reduce the hydrophobic attraction between nonpolar groups. Electrolytes that alter the structural arrangement of water molecules also influence the magnitude of the hydrophobic effect when they are present at sufficiently high concentrations. Structure breakers tend to enhance hydrophobic interactions, whereas structure promoters tend to reduce them (Chapter 4). Variations in pH have little direct effect on the strength of hydrophobic interactions, unless there are accompanying alterations in the structure of the water or the interfacial tension (Israelachvili 2011). 3.8 Hydration Interactions 3.8.1 Origin of Hydration Interactions Hydration interactions arise from the structuring of water molecules around polar and ionic groups (Figure 3.17), which is in contrast to hydrophobic interactions that arise from structuring of water around nonpolar groups. Most food emulsifiers naturally have polar or ionic groups that are hydrated (e.g., –OH, –COO −, and –NH3+), and some of these are also capable of binding hydrated ions (e.g., –COO − + Na+ → –COO − Na+). As two emulsion droplets approach each other, the bonds between the polar groups and the water molecules in their immediate vicinity must be disrupted, which results in a repulsive interaction (Israelachvili 2011). The magnitude and range of the hydration interaction therefore depends on 83 Colloidal Interactions No confinement Entropic confinement (a) (b) (c) FIGURE 3.17 There are a variety of repulsive short-range interactions that may arise when interfaces come into close proximity that are associated with entropic confinement effects: (a) hydration, (b) undulation, and (c) protrusion interactions. the number and strength of the bonds formed between the polar or ionic groups and the water molecules: the greater the degree of hydration, the more repulsive and longer range the interaction. 3.8.2 Modeling Hydration Interactions Just as with hydrophobic interactions, it is difficult to develop precise theoretical models to describe this type of interaction from first principles because of the complex nature of its origin and its dependence on the specific type of ions and polar groups involved. Nevertheless, experimental measurements of the forces between two liquid surfaces have shown that hydration interactions are fairly short-range repulsive forces that decay exponentially with surface-to-surface separation (Israelachvili 2011): whydration (h) = Arl 0e - h / l0 (3.20) where A is a constant that depends on the degree of hydration of the surface (typically between 3 and 30 mJ m−2) λ 0 is the characteristic decay length of the interaction (typically between 0.6 and 1.1 nm) The greater the degree of hydration of a surface group, the larger the values of A and λ 0. The hydration interaction is negligible at large droplet separations but becomes strongly repulsive when the droplets get closer than a certain separation (Figure 3.18). In practice, it is often difficult to isolate the contribution of the hydration forces from other short-range interactions that are associated with mobile interfacial layers at small separations (such as steric and thermal fluctuation interactions), and so there is still much controversy about their origin and nature. Nevertheless, it is widely accepted that they make an important contribution to the overall interaction energy in many systems (Claesson et al. 2006). Experimental measurements of the forces between extremely smooth solid surfaces separated by water reveal an oscillating force versus distance profile, rather than the smooth one predicted by the aforementioned equation (Israelachvili 2011). The spacing between the peaks in this oscillating force curve is equal to the radius of water molecules, which suggests that energy needs to be supplied to expel 84 Food Emulsions: Principles, Practices, and Techniques 300 Hydration Protrusion Undulation 250 w(h)/kT 200 150 100 50 0 0 2 4 6 8 10 h (nm) FIGURE 3.18 Short-range repulsive interactions arise between emulsion droplets when they come into close contact due to hydration, protrusion, and undulation of interfacial layers. each layer of water molecules. Nevertheless, these oscillations are not observed when the surfaces are relatively fluid or rough because the effects are averaged out, which would likely be the case for the surfaces of emulsifier-coated emulsion droplets. 3.8.3 General Characteristics of Hydration Interactions At high electrolyte concentrations, it is possible for ionic surface groups to specifically bind hydrated ions to their surfaces (Hunter 1986, Miklavic and Ninham 1990). Some of these ions have large amounts of water associated with them (Kiriukhin and Collins 2002), and can therefore provide strong repulsive hydration interactions (Israelachvili 2011). Specific binding depends on the radius and valency of the ion involved, because these parameters determine the degree of ion hydration. Ions that have small radii and high valencies tend to bind less strongly because they are surrounded by a relatively thick layer of tightly bound water molecules and some of these must be removed before the ion can be adsorbed (Israelachvili 2011). As a general rule, the adsorbability of ions from water can be described by the lyotropic series: I− > Br− > Cl− > F− for monovalent ions, and K+ > Na+ > Li+ for monovalent cations (in order of decreasing adsorbability). On the other hand, once an ion is bound to a surface, the strength of the repulsive hydration interaction between the emulsion droplets increases with degree of ion hydration because more energy is needed to dehydrate the ion as the two droplets approach each other. Therefore, the ions that tend to adsorb the least strongly are also those the ones that provide the greatest hydration repulsion if they do adsorb. Thus, it is possible to control the interaction between droplets by altering the type and concentration of ions present in the aqueous phase. Hydration interactions are often strong enough to prevent droplets from aggregating (Israelachvili 2011). Thus, oil-in-water emulsions that should contain enough electrolyte to cause droplet flocculation through electrostatic screening have been found to be stable because of specific binding of ions. This effect is dependent on the pH of the aqueous phase because the electrolyte ions have to compete with 85 Colloidal Interactions the H+ or OH− ions in the water (Miklavic and Ninham 1990). For example, at relatively high pH and electrolyte concentrations (>10 mM), it has been observed that Na+ ions can adsorb to negatively charged surface groups and prevent droplets from aggregating through hydration repulsion, but when the pH of the solution is decreased, the droplets aggregate because the high concentration of H+ ions displaces the Na+ ions from the droplet surface (Israelachvili 2011). Nonionic emulsifiers are less sensitive to pH and ionic strength, and they do not usually bind highly hydrated ions. The magnitude of the hydration interaction decreases with increasing temperature because polar groups become progressively dehydrated as the temperature is raised (Israelachvili 2011). In summary, the importance of hydration interactions in a particular system depends on the nature of the hydrophilic groups on the droplet surfaces, as well as on the type and concentration of any ions present in the aqueous phase. 3.9 Thermal Fluctuation Interactions 3.9.1 Origin of Thermal Fluctuation Interactions At the molecular levels, the interfacial region that separates the oil and aqueous phases of an emulsion is often highly dynamic (Israelachvili 2011). In particular, interfaces that are comprised of small-molecule surfactants tend to exhibit undulations because their bending energy is relatively small compared to the thermal energy of the system (Figure 3.17). In addition, the surfactant molecules may be continually twisting and turning, as well as moving in and out of the interfacial region. When two dynamic interfaces move close to each other, they experience a number of short range repulsive thermal fluctuation interactions that are entropic in origin. In emulsions, the two most important of these are protrusion and undulation interactions. 3.9.2 Modeling Thermal Fluctuation Interactions Protrusion interactions are short-range repulsive interactions that arise when two surfaces are brought so close together that the movement of the surfactant molecules in and out of the interface of one droplet is restricted by the presence of another droplet, which is entropically unfavorable (Figure 3.17). The magnitude of this repulsive interaction depends on the distance that the surfactant molecules are able to protrude from the interface, which is governed by their molecular structure. The interdroplet pair potential due to protrusion interactions is given by the following expression (Israelachvili 2011): wprotrusion (h) » 3pGrkTl 0e - h / l0 (3.21) where Γ is the number of surfactant molecules (or head groups) per unit surface area λ 0 is the characteristic decay length of the interaction (typically between 0.07 and 0.6 nm), which depends on the distance the surfactant molecules can protrude from the interface Undulation interactions are short-range repulsive interactions that arise when the wavelike undulations of the interfacial region surrounding one emulsion droplet are restricted by the presence of another emulsion droplet, which is entropically unfavorable (Figure 3.17). The magnitude and range of this repulsive interaction increases as the amplitude of the oscillations increases. The interdroplet pair potential due to undulation interactions is given by the following expression (Israelachvili 2011): wundulation (h) » pr ( kT ) 4 kb h 2 (3.22) 86 Food Emulsions: Principles, Practices, and Techniques here, k b is the bending modulus of the interfacial layer, which typically has values of between about 0.2 and 20 × 10 −20 J depending on the surfactant type. The magnitude of the bending modulus is related to the molecular geometry of the surfactant molecules (Chapter 4) and tends to be higher for surfactants that have two nonpolar chains, than those that have only one. Predictions of the protrusion and undulation interactions made using the aforementioned equations are shown in Figure 3.18. 3.9.3 General Characteristics of Fluctuation Interactions Thermal fluctuation interactions are much more important for small-molecule surfactants that form dynamic and flexible interfacial layers, than for biopolymers that form more rigid and inflexible interfacial layers. Both types of interaction tend to increase with temperature because the interfaces become more mobile. Nevertheless, this effect may be counteracted by increasing dehydration of any polar groups with increasing temperature. These interactions may play a significant role in stabilizing droplets against aggregation in emulsions stabilized by small-molecule surfactants, particularly when they act in conjunction with other types of short-range repulsive interactions, such as steric or hydration interactions. The strength of this interaction is mainly governed by the structure and dynamics of the interfacial layer and therefore varies considerably from system to system. 3.10 Nonequilibrium Effects So far it has been assumed that the interactions between emulsion droplets occur under equilibrium conditions. In practice, the various molecular species and particles in emulsions are in continual motion, which influences the colloidal interactions in a number of ways (Evans and Wennerstrom 1999, Israelachvili 2011). 3.10.1 Molecular Rearrangements at the Interface The system may not have time to reach equilibrium when two droplets rapidly approach each other because molecular rearrangements take a finite time to occur, for example, adsorption–desorption of emulsifiers, ionization–deionization of charged groups, and conformational changes of biopolymers (Israelachvili 2011). As a consequence, the colloidal interactions between droplets may be significantly different from those observed under equilibrium conditions. These nonequilibrium effects depend on the precise nature of the system and are difficult to account for theoretically. 3.10.2 Hydrodynamic Flow of Continuous Phase The movement of a droplet causes an alteration in the flow profile of the intervening continuous phase, which can be felt by another droplet (Walstra 2003). As two droplets move closer together, the continuous phase must be squeezed out from the narrow gap separating them against the friction of the droplet surfaces (Figure 3.19). This effect manifests itself as a decrease in the effective diffusion coefficient of the emulsion droplets, D(h) = D 0 G(h), where D 0 is the diffusion coefficient of a single droplet and G(h) is a correction factor that depends on surface-to-surface separation between the droplets (Hunter 1986). Mathematical expressions for G(h) have been derived from a consideration of the forces that act on particles as they approach each other in a viscous liquid (Zhang and Davis 1991). For rigid spherical particles, the hydrodynamic correction factor can be approximated by the following expression (Hunter 1986): 2h G (h ) = r 3h 2r 3h 3h2 + 1+ 2r r 2 1+ (3.23) 87 Colloidal Interactions The value of G(h) varies from 0, when the particles are in close contact (h = 0), to 1, when they are far apart and therefore have no influence on each other (h → ∞). Thus, as particles approach each other, their speed gets progressively slower, and therefore, they would not aggregate unless there was a sufficiently strong attractive colloidal interaction to overcome this repulsive hydrodynamic interaction. Equation 3.23 must be modified for emulsions to take into account the fact that there is less resistance to the movement of the continuous phase out of the gap between the droplets when their surfaces have some fluid-like characteristics (Zhang and Davis 1991, Blawzdziewicz et al. 1999). Thus, the hydrodynamic resistance to the approach of fluid droplets is less than that for solid droplets. Hydrodynamic interactions are particularly important for determining the stability of droplets to flocculation and coalescence in emulsion systems (Chapter 7). 3.10.3 Gibbs–Marangoni Effect There may be an additional nonequilibrium contribution to the colloidal interactions between emulsion droplets due to the Gibbs–Marangoni effect (Walstra 1993, Wilde et al. 2004, Skartlien et al. 2012). As two droplets approach each other, the liquid in the continuous phase is forced out of the narrow gap that separates them (Figure 3.19). As the liquid is squeezed out, it drags some of the emulsifier molecules along the droplet surface, which leads to the formation of a region where the emulsifier concentration on the surfaces of the two emulsion droplets is lowered. This causes a surface tension gradient at the interface, which is thermodynamically unfavorable. The emulsifier molecules therefore have a tendency to flow toward the region of low emulsifier concentration and high interfacial tension, dragging some of the liquid in the surrounding continuous phase along with them. This motion of the continuous phase is in the opposite direction to the outward flow that occurs when it is squeezed from between the droplets, and therefore, it opposes the movement of the droplets toward each other and therefore increases their stability to close approach and coalescence. This effect is most important for emulsifiers that are relatively mobile at the oil–water interface, such as small-molecule surfactants rather than surface-active biopolymers. Droplets approach Low γ Surfactant High γ Low γ Surfactant Water Water Surfactant Surfactant FIGURE 3.19 Schematic representation of the Gibbs–Marangoni effect. As two emulsion droplets approach each other (top), some of the intervening fluid separating their surfaces must flow out (bottom). The resulting viscous drag on the interfacial layers may lead to an emulsifier concentration gradient at the surfaces, which opposes fluid flow from between the droplets. 88 Food Emulsions: Principles, Practices, and Techniques 3.11 Total Interaction Potential The overall interdroplet pair potential is the sum of the various attractive and repulsive contributions*: wtotal (h) = wVDW (h) + welectrostatic (h) + wsteric (h) + wdepletion (h) + whydrophobic (h) +  (3.24) Not all of the interactions play an important role in every type of food emulsion, and it is often possible to identify two or three interactions that dominate the overall interaction. For this reason, it is informative to examine the characteristics of certain combinations of colloidal interaction that are particularly important in food emulsions. A summary of the characteristics of the various types of interaction is given in Table 3.3. In this section, the utility of predicting the overall interdroplet pair potential as a function of droplet separation for understanding the behavior of food emulsions is demonstrated. We begin by considering a simple system where only van der Waals attraction and steric repulsion operates, and then build up the complexity of the system by incorporating the effects of other important types of attractive and repulsive interactions. The physicochemical parameters used in the theoretical calculations are shown in the captions to the figures. 3.11.1 Van der Waals and Steric Interactions The most basic model for describing the colloidal interactions between emulsion droplets is to consider that only van der Waals and steric interactions are important. Van der Waals interactions always act between emulsion droplets and must therefore always be taken into account. Similarly, the droplets in emulsions are nearly always stabilized by an interfacial layer of adsorbed emulsifier molecules and so steric interactions must also be taken into account. This type of model would be appropriate for describing the behavior of emulsion droplets stabilized by nonionic surfactants or neutral polymers. It would also be appropriate for describing the behavior of emulsion droplets stabilized by charged surfactants or biopolymers at high salt concentrations where the electrostatic interactions are effectively screened. The overall interdroplet pair potential for this simple model system is given by w(h) = wVDW(h) + wsteric(h) (3.25) TABLE 3.3 Summary of the Characteristics of the Various Types of Colloidal Interactions between Emulsion Droplets Interaction Type Van der Waals Electrostatic Strength Range Major Factors Affecting A R Sign S W→S LR SR → LR ε, n, I Ψδ, σ, pH, I R A or R A A R R S W→S W→S S S S SR SR SR LR SR→ MR SR→MR δ, E δ, w ϕc, rc ϕH, T T T Steric Elastic Mixing Depletion Hydrophobic Hydration Thermal fluctuation Note: The interactions are classified according to the following symbols: A, attractive; R, repulsive; S, strong; W, weak; SR, short range (<10 nm); MR, medium range (10–20 nm); LR, long range (>20 nm). The major factors affecting the interactions are dielectric constant (ε), refractive index (n), ionic strength (I), surface potential (Ψδ), surface charge density (σ), thickness of interfacial layer (δ), elastic modulus of interfacial layer (E), effective interaction parameter for emulsifier–solvent interactions (w), and temperature (T). * In reality, it is not always appropriate to simply sum the contribution from all of the separate interactions because some of them are coupled (Ninham and Yaminsky 1997). Nevertheless, this approach gives a good first approximation. 89 Colloidal Interactions 40 w(h)/kT 0 –40 –80 δ (nm) 1 2 5 10 –120 –160 0 5 10 h (nm) 15 20 FIGURE 3.20 Predicted interdroplet pair potentials for emulsions where only van der Waals and steric interactions are important. Predictions are carried out for oil-in-water emulsions with different thicknesses of adsorbed layers as stated in the annotation (r = 1 μm, T = 25°C). The dependence of the overall interdroplet pair potential on droplet separation for emulsions with interfacial layers of different thickness is shown in Figure 3.20. It is assumed that the continuous phase is an indifferent quality solvent for the polymer, so that the mixing contribution to the polymeric steric interaction is zero (Section 3.5). At wide separations, the overall interaction between the droplets is negligible. As the droplets move closer together, the attractive van der Waals interaction begins to dominate and so there is a net attraction between the droplets. However, once the droplets get so close together that their interfacial layers overlap, then the repulsive steric interaction dominates and there is a net repulsion between the droplets. At a particular separation, there is a minimum in the interdroplet pair potential, and this is the location where the droplets tend to reside. If the depth of this minimum is large compared to the thermal energy of the system, the droplets remain aggregated; otherwise, they move apart. The depth and position of the minimum depends on the thickness and properties of the interfacial layer surrounding the droplets. As the thickness of the adsorbed layer increases, the strong steric repulsion interaction between droplets becomes more significant at larger separations, and consequently, the depth of the minima decreases. If the interfacial layer is sufficiently thick, it may prevent the droplets from aggregating altogether, because the depth of the minimum is relatively small compared to the thermal energy. This phenomenon accounts for the effectiveness of emulsifiers that form relatively thick interfacial layers (e.g., polysaccharides) at preventing droplet flocculation and coalescence in emulsions containing high salt concentrations, whereas emulsifiers that form relatively thin interfacial layers (e.g., globular proteins) can prevent coalescence but not flocculation (Chanamai et al. 2002, Charoen et al. 2011). As mentioned in Section 3.3, the composition and thickness of the adsorbed layer may have an additional influence on the overall interaction potential due to its modification of the van der Waals interactions (which was not taken into account in the calculations carried out here). It should also be mentioned that emulsion droplets coated by emulsifiers that should in principle have no net charge (such as nonionic surfactants) may actually have a net charge because of ionic impurities in the system (such as free fatty acids), and therefore, electrostatic interactions would have to be included in the calculations for these systems. 90 Food Emulsions: Principles, Practices, and Techniques 3.11.2 Van der Waals, Steric, and Electrostatic Interactions The droplets in many food emulsions have an electric charge because of adsorption of surface-active ions or emulsifiers (Section 3.4). A more realistic model of the colloidal interactions between emulsion droplets is therefore obtained by considering van der Waals, steric, and electrostatic interactions: w(h) = w VDW(h) + wsteric(h) + welectrostatic(h) (3.26) The dependence of the overall interdroplet pair potential on droplet separation for electrically charged droplets is shown in Figure 3.21. For simplicity, a more schematic representation of the general form of the interaction potential for this type of system is shown in Figure 3.22. When the two droplets are separated by a large distance, there is no effective interaction between them. As they move closer together, the van der Waals attraction dominates initially and there is a shallow minimum in the profile, which is referred to as the secondary minimum, w(h2o min ). When the depth of this minimum is large compared to the thermal energy ( w(h2o min )  kT ), the droplets tend to be flocculated, but if it is small compared to the thermal energy, then the droplets tend to remain nonaggregated. At closer separations the repulsive electrostatic interaction dominates and there is an energy barrier, w(hmax), that must be overcome before the droplets can come any closer. At still closer separations, the attractive van der Waals interaction dominates the repulsive electrostatic interaction and there is a relatively deep primary minimum, w(h1o min ). If the energy barrier is sufficiently large compared to the thermal energy (|w(hmax)| > 20 kT), then it will effectively prevent the droplets from falling into the primary minimum. On the other hand, if it is relatively small compared to the thermal energy, then the droplets tend to fall into the primary minimum, which would lead to strong droplet aggregation. When the droplets get so close together that their interfacial interfaces overlap, there is an extremely strong steric repulsion that dominates the other interactions. This short-range repulsive interaction should prevent the droplets from getting close enough together to coalesce. 140 I (mM) 10 20 50 100 200 100 w(h)/kT 60 20 –20 + – –60 + –100 0 5 + – – – + + – + + – + + – + – – – + 10 h (nm) + – – – + – + – – – – + + – + + – + + + – + + – – – + 15 20 FIGURE 3.21 Predicted interdroplet pair potentials for emulsions where only van der Waals, steric, and electrostatic interactions are important. Predictions are carried out for oil-in-water emulsions with different ionic strengths as stated in the annotation (r = 1 μm, δ = 1 nm, T = 25°C, Ψ0 = 20 mV). The height of the energy barrier decreases as the ionic strength of the intervening medium increases because of electrostatic screening. 91 Colloidal Interactions w(h) Energy barrier h 2° Min 1° Min FIGURE 3.22 Schematic representation of the overall interaction potential between a pair of electrically charged droplets covered by an interfacial membrane, assuming only van der Waals, steric, and electrostatic interactions are important. The depths of the primary and secondary minima and the height of the energy barrier determine the stability of the system to droplet aggregation (see text for details). There are a number of complicating factors that need to be taken into account when implementing the aforementioned approach. First, one must decide the location of the electrical charge in the system (e.g., at the oil droplet surfaces or at the outer edge of the interfacial layers), since this will have a pronounced influence on the strength and range of the electrostatic interactions. Second, one may have to take into account the influence of electrostatic screening, retardation, and the interfacial layer on the strength of the van der Waals interactions (Section 3.3). Third, one may have to take into account changes in the thickness or characteristics of the interfacial layer if it consists of charged emulsifier molecules, since this will influence the strength and range of the steric interactions (Section 3.5). Despite these complicating factors, the aforementioned approach provides valuable insights into the factors that influence the stability of electrically charged emulsion droplets. Electrostatically stabilized emulsions are particularly sensitive to the ionic strength and pH of the aqueous phase (Figure 3.21). At low electrolyte concentrations, there may be a sufficiently high energy barrier to prevent the droplets from coming close enough together to aggregate into the primary minimum. As the ion concentration is increased, the screening of the electrostatic interaction becomes more effective (Section 3.4), which reduces the height of the energy barrier. Above a certain electrolyte concentration, often referred to as the critical aggregation concentration (CAC), the energy barrier is no longer high enough to prevent the droplets from falling into the deep primary minimum, and so the droplets tend to aggregate. This accounts for the susceptibility of many electrostatically stabilized food emulsions to droplet aggregation when salt is added to the aqueous phase (McClements 2004, Dickinson 2008). The electrical charge of many food emulsifiers is sensitive to the pH of the aqueous phase (Chapter 4). For example, the charge on the droplets in protein-stabilized emulsions goes from positive below the 92 Food Emulsions: Principles, Practices, and Techniques isoelectric point to negative above it (McClements 2004). The magnitude of the droplet charge therefore decreases as the isoelectric point is approached, which reduces the electrostatic repulsion acting between the droplets. This accounts for the tendency of protein-stabilized emulsions to become flocculated when their pH is adjusted to the isoelectric point of the adsorbed proteins. Nevertheless, the droplets are often stable to coalescence because of the presence of a short-range steric repulsion associated with the adsorbed protein layers. The classical approach to describing colloidal interactions between electrically charged particles is the DLVO theory, named after the four scientists who first proposed it: Derjaguin, Landau, Verwey, and Overbeek (Hiemenz and Rajagopalan 1997). This theory accounts for the van der Waals attraction and electrostatic repulsion, but does not include the short-range steric repulsion that acts between droplets at close separations, and it is therefore not a particularly realistic model for describing the behavior of emulsion droplets coated by interfacial layers. The DLVO theory predicts that emulsion droplets should coalesce once they fall into the primary minimum because there is no short-range repulsive force stopping them from getting closer together. 3.11.3 Van der Waals, Steric, Electrostatic, and Hydrophobic Interactions The droplet surfaces in many food emulsions acquire some hydrophobic character during their manufacture, storage, or consumption. A typical example is a whey protein–stabilized emulsion that is subjected to heat treatment (Kim et al. 2002, 2005, Dickinson and Parkinson 2004). Heating the emulsion above 65°C causes the protein molecules adsorbed to the oil–water interface to partially unfold and thus expose some of the nonpolar amino acids to the aqueous phase. An increase in surface hydrophobicity may also occur at ambient temperatures due to surface denaturation of globular proteins after adsorption to droplet surfaces (Kim et al. 2002, 2005, Dickinson and Parkinson 2004). The overall interdroplet pair potential for this type of system is given by 150 φH (%) 0 0.1 100 0.25 0.5 w(h)/kT 50 1 0 –50 –100 –150 0 2 4 6 8 10 12 h (nm) FIGURE 3.23 Predicted interdroplet pair potentials for emulsions where van der Waals, steric, electrostatic, and hydrophobic interactions are important. Predictions are carried out for oil-in-water emulsions with different surface hydrophobicities as stated in the annotation box (r = 1 μm, δ = 1 nm, T = 25°C, Ψ0 = 20 mV, and I = 20 mM). The height of the energy barrier decreases as the surface hydrophobicity increases because of the increase in the hydrophobic attraction. 93 Colloidal Interactions w(h) = wVDW(h) + wsteric(h) + welectrostatic(h) + whydrophobic(h) (3.27) The dependence of the overall interdroplet pair potential on droplet separation for droplets with different degrees of surface hydrophobicity is shown in Figure 3.23. As the hydrophobicity of the droplet surface increases, the hydrophobic attraction increases, which causes a decrease in the height of the energy barrier. When the surface hydrophobicity is sufficiently large, the energy barrier becomes so small that the droplets can aggregate into the primary minimum. This accounts for the experimental observation that whey protein–stabilized emulsions become more susceptible to aggregation when they are heated above a temperature where the protein molecules unfold (Dickinson and Parkinson 2004, McClements 2004). 3.11.4 Van der Waals, Steric, Electrostatic, and Depletion Interactions Depletion interactions are important when the continuous phase of an emulsion contains a significant concentration of small colloidal particles, such as surfactant micelles or nonadsorbing biopolymers (Jenkins and Snowden 1996, Dickinson 2003, 2010). The interdroplet pair potential for a system in which depletion interactions are important is given by w(h) = wVDW (h) + wsteric(h) + welectrostatic(h) + wdepletion(h) (3.28) The variation of the interdroplet pair potential with droplet separation for this type of system is shown in Figure 3.24. At low concentrations of colloidal particles, the energy barrier is sufficiently large to prevent the droplets falling into the primary minimum. As the concentration of colloidal particles is increased, the attraction between the droplets increases. A number of workers have shown that depletion interactions promote droplet flocculation in emulsions when the concentration of nonadsorbed colloidal particles exceeds some critical concentration, for example, nonionic surfactants (McClements 1994), 30 φc (%) 2 4 6 8 10 20 w(h)/kT 10 0 –10 –20 –30 0 2 4 6 h (nm) 8 10 12 14 FIGURE 3.24 Predicted interdroplet pair potentials for emulsions where van der Waals, steric, electrostatic, and depletion interactions are important. Predictions are carried out for oil-in-water emulsions containing different volume fractions of nonadsorbed colloidal particles dispersed in the continuous phase as stated in the annotation box (r = 1 μm, δ = 1 nm, rc = 5 nm, T = 25 °C, Ψ0 = 20 mV, and I = 50 mM). The height of the energy barrier decreases as the concentration of colloidal particles increases because of the increase in the depletion attraction. 94 Food Emulsions: Principles, Practices, and Techniques ionic surfactants (Bibette et al. 1992), proteins (Dickinson et al. 1997, Radford and Dickinson 2004), and polysaccharides (Reiffers-Magnani et al. 2000, Gu et al. 2005, Moschakis et al. 2005). It is therefore particularly important to take depletion interactions into account in any system that has an appreciable amount of nonadsorbed colloidal particles in the continuous phase. 3.12 Measurement of Colloidal Interactions One of the major technological advances in understanding and characterizing colloidal interactions was the development of analytical instruments that could accurately measure the forces between surfaces down to separations of a fraction of a nanometer (Claesson et al. 1996, Israelachvili 2011). A variety of different instruments have been developed for this purpose, including surface force apparatus, atomic force microscopy, and light-lever instruments (Claesson et al. 1996, Christenson and Claesson 2001, Butt et al. 2005, Liang et al. 2007, Israelachvili 2011). All of the instruments use some means of measuring the separation distance and force acting between two macroscopic surfaces as one of the surfaces is moved through the intervening liquid in a controlled fashion. The surfaces can be chemically or physically modified to control their hydrophobicity, hydrophilicity, electrical charge, roughness, and thickness. In addition, it is possible to adsorb different types of emulsifier onto the surfaces and to vary the composition and properties of the liquid separating the surfaces (e.g., solvent type, ionic strength, pH, and concentration of nonadsorbing colloidal particles). One drawback of these techniques for understanding the characteristics of particular types of interactions is that they only measure the overall interaction potential, and so it is necessary to design ways of disentangling the contributions from the various individual interactions. Nevertheless, the application of these techniques has led to considerable advances in our knowledge of the origin, sign, magnitude, and range of colloidal interactions, as well as providing a better understanding of the factors that influence them, such as pH, ionic environment, solvent composition, surface properties, and temperature. Ultimately, the knowledge gained from application of these techniques will help food scientists to gain a better understanding of the factors that determine the stability of food emulsions. 3.13 Prediction of Colloidal Interactions in Food Emulsions In this chapter, we have examined the origin, magnitude, and range of the most important types of attractive and repulsive interactions that can arise between emulsion droplets. In principle, it is possible to predict the likelihood that the droplets in an emulsion will be in an aggregated or a nonaggregated state using the theories given earlier. In practice, it is extremely difficult to make quantitative predictions about the aggregation stability of food emulsions for a number of reasons. First, food emulsions contain a huge number of different emulsion droplets (rather than just two) that interact with each other and with other components within the system, and it is difficult to quantify the overall nature of these interactions. Second, there is often a lack of information about the relevant physical parameters needed to carry out the calculations. Third, certain simplifying expressions often have to be made in the theories in order to derive tractable expressions for the interaction energies, and these are not always justified. Fourth, food systems are not usually at thermodynamic equilibrium and so many of the aforementioned equations do not strictly apply. Fifth, covalent interactions are important in some systems, and these are not taken into account in the aforementioned analysis. Finally, food emulsions may be subjected to external forces that affect the interactions between the droplets, for example, gravity, centrifugation, or mechanical agitation. Despite these limitations, an understanding of the various types of colloidal interactions that act between emulsion droplets gives food scientists a powerful tool for understanding and predicting the effects of ingredient formulations and processing conditions on the properties of many food products. It is often possible to predict the major factors that determine the stability of emulsions (albeit in a fairly qualitative fashion). Alternatively, in some systems it may be possible to experimentally measure the forces between surfaces using a force measurement technique (see Section 3.12) for a system that closely Colloidal Interactions 95 mimics the food system of interest. For example, it may be possible to coat the solid surfaces in the force-measuring device with the same type of emulsifier as used in the food product and to use an aqueous solution with the same composition as found in the food product (e.g., pH, ionic composition). The force–distance curves could then be measured and the factors that influence them determined. REFERENCES Attard, P. (2003). Nanobubbles and the hydrophobic attraction. Advances in Colloid and Interface Science 104: 75–91. Bergethon, P. R. (2010). The Physical Basis of Biochemistry: The Foundations of Molecular Biophysics. New York: Springer. Bibette, J., D. Roux, and B. Pouligny (1992). Creaming of emulsions—The role of depletion forces induced by surfactant. Journal De Physique Ii 2(3): 401–424. Bishop, K. J. M., C. E. Wilmer, S. Soh, and B. A. Grzybowski (2009). Nanoscale forces and their uses in selfassembly. Small 5(14): 1600–1630. Blawzdziewicz, J., E. Wajnryb, and M. Loewenberg (1999). Hydrodynamic interactions and collision efficiencies of spherical drops covered with an incompressible surfactant film. Journal of Fluid Mechanics 395: 29–59. Bowen, W. R. and F. Jenner (1995). The calculation of dispersion forces for engineering applications. Advances in Colloid and Interface Science 56: 201–243. Butt, H. J., B. Cappella, and M. Kappl (2005). Force measurements with the atomic force microscope: Technique, interpretation and applications. Surface Science Reports 59(1–6): 1–152. Chanamai, R., G. Horn, and D. J. McClements (2002). Influence of oil polarity on droplet growth in oilin-water emulsions stabilized by a weakly adsorbing biopolymer or a nonionic surfactant. Journal of Colloid and Interface Science 247(1): 167–176. Charoen, R., A. Jangchud, K. Jangchud, T. Harnsilawat, O. Naivikul, and D. J. McClements (2011). Influence of biopolymer emulsifier type on formation and stability of rice bran oil-in-water emulsions: Whey protein, gum arabic, and modified starch. Journal of Food Science 76(1): E165–E172. Christenson, H. K. and P. M. Claesson (2001). Direct measurements of the force between hydrophobic surfaces in water. Advances in Colloid and Interface Science 91(3): 391–436. Claesson, P. M., E. Blomberg, and E. Poptoshev (2004). Surface forces and emulsion stability. In Food Emulsions, S. Friberg, K. Larsson, and J. Sjoblom, eds. New York, Marcel Dekker. Claesson, P. M., T. Ederth, V. Bergeron, and M. W. Rutland (1996). Techniques for measuring surface forces. Advances in Colloid and Interface Science 67: 119–183. Claesson, P. M., M. Kjellin, O. J. Rojas, and C. Stubenrauch (2006). Short-range interactions between nonionic surfactant layers. Physical Chemistry Chemical Physics 8(47): 5501–5514. Dickinson, E. (1992). Introduction to Food Colloids. Cambridge, U.K., Royal Society of Chemistry. Dickinson, E. (2003). Hydrocolloids at interfaces and the influence on the properties of dispersed systems. Food Hydrocolloids 17(1): 25–39. Dickinson, E. (2006). Colloid science of mixed ingredients. Soft Matter 2(8): 642–652. Dickinson, E. (2008). Interfacial structure and stability of food emulsions as affected by protein–polysaccharide interactions. Soft Matter 4(5): 932–942. Dickinson, E. (2010). Flocculation of protein-stabilized oil-in-water emulsions. Colloids and Surfaces B-Biointerfaces 81(1): 130–140. Dickinson, E. (2013). Structure and rheology of colloidal particle gels: Insight from computer simulation. Advances in Colloid and Interface Science 199: 114–127. Dickinson, E., M. Golding, and M. J. W. Povey (1997). Creaming and flocculation of oil-in-water emulsions containing sodium caseinate. Journal of Colloid and Interface Science 185(2): 515–529. Dickinson, E. and E. L. Parkinson (2004). Heat-induced aggregation of milk protein-stabilized emulsions: Sensitivity to processing and composition. International Dairy Journal 14(7): 635–645. Evans, E. D. and W. Wennerstrom (1999). The Colloidal Domain: Where Physics, Chemistry and Biology Meet. New York, Wiley-VCH. Friberg, S. (1997). Emulsion stability. In Food Emulsions, S. Friberg and K. Larsson, eds., pp. 1–56. New York, Marcel Dekker, Inc. 96 Food Emulsions: Principles, Practices, and Techniques Gregory, J. (1981). Approximate expressions for retarded van der Waals interaction. Journal of Colloid and Interface Science 83(1): 138–145. Gu, Y. S., E. A. Decker, and D. J. McClements (2005). Influence of pH and carrageenan type on properties of beta-lactoglobulin stabilized oil-in-water emulsions. Food Hydrocolloids 19(1): 83–91. Guyot, C., C. Bonnafont, I. Lesschaeve, S. Issanchou, A. Voilley, and H. E. Spinnler (1996). Effect of fat content oil odor intensity of three aroma compounds in model emulsions: Delta-decalactone, diacetyl, and butyric acid. Journal of Agricultural and Food Chemistry 44(8): 2341–2348. Guzey, D. and D. J. McClements (2006). Formation, stability and properties of multilayer emulsions for application in the food industry. Advances in Colloid and Interface Science 128: 227–248. Hiemenz, P. C. and R. Rajagopalan (1997). Principles of Colloid and Surface Chemistry. New York: Marcel Dekker. Hunter, R. J. (1986). Foundations of Colloid Science. Oxford, U.K.: Oxford University Press. Israelachvili, J. (2011). Intermolecular and Surface Forces, 3rd edn. London, U.K.: Academic Press. Jackel, V. K. (1964). Uber die funcktionen des schuzkolloids. Kolloid-Zeitschrift und Zeitshrift fur Polymere 197: 143–154. Jenkins, P. and M. Snowden (1996). Depletion flocculation in colloidal dispersions. Advances in Colloid and Interface Science 68: 57–96. Kayitmazer, A. B., D. Seeman, B. B. Minsky, P. L. Dubin, and Y. Xu (2013). Protein–polyelectrolyte interactions. Soft Matter 9(9): 2553–2583. Keowmaneechai, E. and D. J. McClements (2002). Effect of CaCl2 and KCl on physiochemical properties of model nutritional beverages based on whey protein stabilized oil-in-water emulsions. Journal of Food Science 67(2): 665–671. Kim, H. J., E. A. Decker, and D. J. McClements (2002). Role of postadsorption conformation changes of betalactoglobulin on its ability to stabilize oil droplets against flocculation during heating at neutral pH. Langmuir 18(20): 7577–7583. Kim, H. J., E. A. Decker, and D. J. McClements (2005). Influence of protein concentration and order of addition on thermal stability of beta-lactoglobulin stabilized n-hexadecane oil-in-water emulsions at neutral pH. Langmuir 21(1): 134–139. Kiriukhin, M. Y. and K. D. Collins (2002). Dynamic hydration numbers for biologically important ions. Biophysical Chemistry 99(2): 155–168. Kulmyrzaev, A., R. Chanamai, and D. J. McClements (2000a). Influence of pH and CaCl2 on the stability of dilute whey protein stabilized emulsions. Food Research International 33(1): 15–20. Kulmyrzaev, A., M. P. C. Sivestre, and D. J. McClements (2000b). Rheology and stability of whey protein stabilized emulsions with high CaCl2 concentrations. Food Research International 33(1): 21–25. Liang, Y., N. Hilal, P. Langston, and V. Starov (2007). Interaction forces between colloidal particles in liquid: Theory and experiment. Advances in Colloid and Interface Science 134–135: 151–166. Mahanty, J. and B. W. Ninham (1976). Dispersion Forces. New York: Academic Press. McClements, D. J. (1994). Ultrasonic determination of depletion flocculation in oil-in-water emulsions containing a nonionic surfactant. Colloids and Surfaces A: Physicochemical and Engineering Aspects 90(1): 25–35. McClements, D. J. (2000). Comments on viscosity enhancement and depletion flocculation by polysaccharides. Food Hydrocolloids 14(2): 173–177. McClements, D. J. (2004). Protein-stabilized emulsions. Current Opinion in Colloid & Interface Science 9(5): 305–313. McClements, D. J. and E. A. Decker (2000). Lipid oxidation in oil-in-water emulsions: Impact of molecular environment on chemical reactions in heterogeneous food systems. Journal of Food Science 65(8): 1270–1282. Mei, L. Y., E. A. Decker, and D. J. McClements (1998). Evidence of iron association with emulsion droplets and its impact on lipid oxidation. Journal of Agricultural and Food Chemistry 46(12): 5072–5077. Mei, L. Y., D. J. McClements, and E. A. Decker (1999). Lipid oxidation in emulsions as affected by charge status of antioxidants and emulsion droplets. Journal of Agricultural and Food Chemistry 47(6): 2267–2273. Meyer, E. E., K. J. Rosenberg, and J. Israelachvili (2006). Recent progress in understanding hydrophobic interactions. Proceedings of the National Academy of Sciences of the United States of America 103(43): 15739–15746. Colloidal Interactions 97 Miklavic, S. J., D. Y. C. Chan, L. R. White, and T. W. Healy (1994). Double-layer forces between heterogeneous charged surfaces. Journal of Physical Chemistry 98(36): 9022–9032. Miklavic, S. J. and B. W. Ninham (1990). Competition for adsorption sites by hydrated ions. Journal of Colloid and Interface Science 134(2): 305–311. Moschakis, T., B. S. Murray, and E. Dickinson (2005). Microstructural evolution of viscoelastic emulsions stabilised by sodium caseinate and xanthan gum. Journal of Colloid and Interface Science 284(2): 714–728. Ninham, B. W. and V. Yaminsky (1997). Ion binding and ion specificity: The Hofmeister effect and Onsager and Lifshitz theories. Langmuir 13(7): 2097–2108. Norde, W. (2011). Colloids and Interfaces in Life Sciences and Bionanotechnology. Boca Raton, FL: CRC Press. Pailthorpe, B. A. and W. B. Russel (1982). The retarded van der Waals interaction between spheres. Journal of Colloid and Interface Science 89(2): 563–566. Quemada, D. and C. Berli (2002). Energy of interaction in colloids and its implications in rheological modeling. Advances in Colloid and Interface Science 98(1): 51–85. Radford, S. J. and E. Dickinson (2004). Depletion flocculation of caseinate-stabilised emulsions: What is the optimum size of the non-adsorbed protein nano-particles? Colloids and Surfaces A: Physicochemical and Engineering Aspects 238(1–3): 71–81. Reiffers-Magnani, C. K., J. L. Cuq, and H. J. Watzke (2000). Depletion flocculation and thermodynamic incompatibility in whey protein stabilised O/W emulsions. Food Hydrocolloids 14(6): 521–530. Skartlien, R., B. Grimes, P. Meakin, J. Sjoblom, and E. Sollum (2012). Coalescence kinetics in surfactant stabilized emulsions: Evolution equations from direct numerical simulations. Journal of Chemical Physics 137(21): 214701. Sperry, P. R. (1982). A simple quantitative model for the volume restriction flocculation of latex by watersoluble polymers. Journal of Colloid and Interface Science 87(2): 375–384. Tabor, R. F., F. Grieser, R. R. Dagastine, and D. Y. C. Chan (2014). The hydrophobic force: Measurements and methods. Physical Chemistry Chemical Physics 16(34): 18065–18075. Tabor, R. F., C. Wu, F. Grieser, R. R. Dagastine, and D. Y. C. Chan (2013). Measurement of the hydrophobic force in a soft matter system. Journal of Physical Chemistry Letters 4(22): 3872–3877. Vold, M. J. (1961). Effect of adsorption on van der waals interaction of spherical colloidal particles. Journal of Colloid Science 16(1): 1–10. Walstra, P. (1993). Principles of emulsion formation. Chemical Engineering Science 48(2): 333–349. Walstra, P. (2003). Physical Chemistry of Foods. New York: Marcel Dekker. Wilde, P., A. Mackie, F. Husband, P. Gunning, and V. Morris (2004). Proteins and emulsifiers at liquid interfaces. Advances in Colloid and Interface Science 108: 63–71. Zhang, X. G. and R. H. Davis (1991). The rate of collisions due to brownian or gravitational motion of small drops. Journal of Fluid Mechanics 230: 479–504. Ziani, K., Y. H. Chang, L. McLandsborough, and D. J. McClements (2011). Influence of surfactant charge on antimicrobial efficacy of surfactant-stabilized thyme oil nanoemulsions. Journal of Agricultural and Food Chemistry 59(11): 6247–6255. 4 Emulsion Ingredients 4.1 Introduction Commercial food emulsions typically contain a wide variety of different ingredients, including oils, emulsifiers, thickening agents, gelling agents, buffering systems, preservatives, antioxidants, chelating agents, sweeteners, salts, colorants, flavors, etc. Each of these ingredients has its own unique molecular, physical, and functional properties. Ultimately, the physicochemical, sensory, and nutritional properties of an emulsion-based food product depend on the type of ingredients present, their physical location, and their interactions with each other. The efficient production of high-quality food emulsions therefore depends on knowledge of the contribution that each individual ingredient makes to the overall properties and how this contribution is influenced by the presence of the other ingredients. One of the most important decisions that a food manufacturer must make during the design, formulation, and production of an emulsion-based food product is the selection of the most appropriate ingredients for that particular application. Each ingredient must exhibit its desired functional properties within the food, while also being economically viable, convenient to use, of reliably high quality, compatible with other ingredients, readily available, and possibly “label friendly.” It is possible to define the composition of an emulsion in a number of different ways: concentrations of specific atoms (e.g., H, C, O, N, Na, Mg, Cl, and P); concentrations of specific molecules (e.g., water, sucrose, amylose, and β-lactoglobulin); concentrations of general classes of molecules (e.g., proteins, lipids, carbohydrates, and minerals); concentrations of specific functional ingredients (e.g., emulsifiers, texture modifiers, chelating agents, buffers, and preservatives); or concentrations of composite ingredients (e.g., flour, milk, salt, and egg). Composite ingredients contain a number of different constituents that may have different functions in the final product; for example, eggs contain emulsifiers, gelling agents, and chelating agents. Food manufacturers are usually concerned with the concentration of composite or specific functional ingredients, because food components are normally purchased and utilized in this form. On the other hand, research scientists may be more interested in the concentrations of specific atoms, molecules, or molecular classes, depending on the purpose of their investigations. In this chapter, we mainly categorize ingredients according to their functional roles within emulsions, since this seems to be the most logical and convenient means of discussing them. The formulation of food products has traditionally been more of a craft than a science. Many of the foods that are familiar to us today are the result of a long and complex history of development. Consequently, there has often been a rather poor understanding of the role (or multiple roles) that each chemical constituent plays in determining their overall quality. The twentieth century saw the development of largescale industrial manufacturing operations where foods were mass produced. Mass production has led to the availability of a wide variety of inexpensive foods that are quick and easy to prepare, and are therefore appealing to consumers. Nevertheless, increasing reliance on mass production has meant that food manufacturers have had to develop a more thorough understanding of the behavior of food ingredients before, during, and after processing. This knowledge is required for a number of reasons: 1. The properties of the ingredients entering a food factory often vary from batch to batch. Food manufacturers utilize their knowledge of the behavior of food ingredients under different conditions to adjust the food processing operations so that the final product has consistent properties. 99 100 Food Emulsions: Principles, Practices, and Techniques 2. Food manufacturers are often looking for cheaper alternatives to existing ingredients, for ingredients with improved functional properties, or for ingredients that are more “label friendly.” An understanding of the role(s) that the original ingredient plays in a food facilitates the rational selection of an alternative ingredient. 3. Food companies are often trying to improve the quality, variety, and convenience of the processed foods they sell to improve their competitiveness. Knowledge of ingredient properties enables food scientists to develop these foods in a more systematic and informed manner. 4. The food industry is trying to develop high-quality products with reduced amounts of food constituents associated with human health concerns (e.g., saturated fat, trans fatty acids, cholesterol, sugar, and salt), or that are fortified with food constituents that have been associated with maintaining or improving human health (e.g., ω-3 fatty acids, dietary fiber, specific minerals, and nutraceuticals). The removal of certain ingredients or the addition of new ingredients may cause significant changes in the taste, texture, or appearance of foods that consumers find undesirable. For example, many no-fat or low-fat products do not exhibit the desirable taste or textural characteristics of the full-fat products that they are designed to replace. Consequently, it is important to understand the role that each ingredient plays in determining the overall physicochemical and organoleptic properties of foods, so that this role can be mimicked by a healthier alternative ingredient. This chapter provides an overview of the molecular, physicochemical, and functional characteristics of the major categories of functional ingredients present in food emulsions. Special emphasis is given to those ingredients that are particular to food emulsions, that is, emulsifiers, texture modifiers, and stabilizers. 4.2 Fats and Oils Fats and oils are part of a group of compounds known as lipids (Damodaran et al. 2007, Gunstone et al. 2007, Akoh and Min 2008, Belitz et al. 2009, Brady 2013). By definition a lipid is a compound that is soluble in organic solvents, but insoluble or only sparingly soluble in water. This group of compounds contains a large number of different types of molecules, including triacylglycerols, diacylglycerols, monoacylglycerols, free fatty acids, sterols, and phospholipids. Triacylglycerols are by far the most common lipid in foods, and it is this type of molecule that is usually referred to as a fat or oil. Edible fats and oils come from a variety of different sources, including plants, seeds, nuts, animals, and fish (Gunstone 2008). By convention a fat is solid-like at room temperature, whereas an oil is liquid-like, although these terms are often used interchangeably. Because of their high natural abundance and their major importance in food emulsions, we are mainly concerned with the properties of triacylglycerols in this section. Nevertheless, it should be mentioned that other types of lipids are more important in certain types of emulsion-based foods and beverages, for example, the major lipid source in many beverage emulsions are flavor oils (Chapter 12). Fats and oils influence the nutritional, sensory, and physicochemical properties of food emulsions in a variety of ways. Lipids are a major source of energy and essential nutrients in the human diet (Leray 2014); however, overconsumption of certain types of lipid (cholesterol, saturated fat, and trans fatty acids) have been linked to human health concerns, such as obesity, cardiovascular disease, diabetes, and cancer (Bray et al. 2004, Muoio and Newgard 2006, Akoh and Min 2008, Leray 2014). Consequently, there has been a trend in the food industry to reduce the overall fat content of many traditional foods, as well as reducing the proportion of undesirable lipids within the fat phase, for example, saturated fats and cholesterol (Ma and Boye 2013). The challenge to the food scientist is to create a product that has the same desirable quality attributes as the original, but with a reduced fat content, which is often extremely difficult. On the other hand, underconsumption of certain types of polyunsaturated fats (e.g., ω-3 fatty acids) has also been linked to human health problems, such as heart disease, diabetes, cancer, and psychological disorders (Ruxton et al. 2004, Freeman et al. 2006, Gebauer et al. 2006). Consequently, many food manufacturers are attempting to find effective strategies of incorporating these polyunsaturated Emulsion Ingredients 101 lipids into foods, which is often problematic because of their poor oxidative stability (McClements and Decker 2000, Waraho et al. 2011, Berton-Carabin et al. 2014). The perceived flavor of a food emulsion is strongly influenced by the type and concentration of lipids present (Chapter 9). Lipids undergo a variety of chemical changes during the processing, storage, and handling of foods that generate products that can be either desirable or deleterious to their flavor profile. Controlling these reactions requires knowledge of both lipid chemistry and emulsion science (Waraho et al. 2010, 2011). The flavor of food emulsions is also indirectly influenced by the presence of the lipid phase because flavor compounds can partition between the oil, water, and gaseous phases according to their polarities (Chapter 9). For this reason, the perceived aroma and taste of food emulsions are often strongly influenced by the type and concentration of lipids present. The lipid phase may also act as a solvent for various other important food components, including oil-soluble vitamins, antioxidants, preservatives, and essential oils. Reducing the lipid content of an emulsion can therefore have a profound influence on its flavor profile, stability, and nutritional content. The characteristic appearance and rheology of food emulsions is largely a result of the immiscibility of oil and water, since this leads to a system where the droplets of one phase are dispersed in the other phase. Food emulsions usually appear turbid, cloudy, or opaque because the light passing through them is scattered by these droplets (Chapter 10). The intensity of the scattering depends on the concentration of droplets present, so that both the color and opacity of food emulsions are strongly influenced by their fat content. The rheology of food emulsions also depends on fat content, since emulsion viscosity increases with increasing droplet concentration. The textural characteristics of emulsions are important in numerous commercial products, including creams, desserts, dressings, and mayonnaise (Chapter 8). The characteristic textural attributes of some food emulsions is due to the ability of the oil phase to crystallize (Fredrick et al. 2010). For example, the “spreadability” of water-in-oil emulsions, such as margarines and butters, is determined by the formation of a three-dimensional network of aggregated fat crystals in the continuous phase that gives the product some mechanical rigidity (Walstra 2003, Marangoni et al. 2012, Sato et al. 2013). On the other hand, the creation of products such as ice cream and whipped cream depends on the controlled destabilization of partially crystalline oil droplets in oil-in-water emulsions (Fredrick et al. 2010, Mendez-Velasco and Goff 2012). The tendency of a cream to thicken or “clot” when it is cooled below a certain temperature is due to the formation of fat crystals in the oil droplets, which causes them to aggregate (Boode et al. 1991). The melting of fat crystals in the mouth causes a cooling sensation that is an important sensory attribute of many fatty foods (Walstra 2003). The ability of food scientists to improve the quality of food emulsions therefore depends on an improved understanding of the multiple roles that fats and oils play in determining their properties. 4.2.1 Molecular Structure and Organization Chemically, triacylglycerols are esters of a glycerol molecule and three fatty acid molecules (Figure 4.1). Each of the fatty acids may contain different numbers of carbon atoms and may have different degrees of unsaturation and branching (Akoh and Min 2008, Belitz et al. 2009). Nevertheless, most naturally occurring fatty acids have an even number of carbon atoms (usually less than 24) and are nonbranched. The fact that there are many different types of fatty acid molecule, and that these fatty acids can be located at different positions on the glycerol molecule, means that there are a huge number of possible triacylglycerol molecules present in foods. Indeed, edible fats and oils always contain a mixture of many different types of triacylglycerol molecules, with the precise type and concentration depending on their biological origin. Triacylglycerol molecules have a “tuning-fork” structure, with the two fatty acids at the ends of the glycerol molecule pointing in one direction, and the fatty acid in the middle pointing in the opposite direction (Figure 4.1). Triacylglycerols are predominantly nonpolar molecules and so the most important types of molecular interaction with their neighbors are the van der Waals attraction and steric overlap repulsion (Chapter 2). At a certain molecular separation, there is a minimum in the intermolecular pair potential whose depth is a measure of the strength of the attractive interactions that hold the molecules together in the solid and liquid states (Section 2.4). Whether a triacylglycerol exists as a liquid or solid at 102 Food Emulsions: Principles, Practices, and Techniques FIGURE 4.1 Chemical structure of a triacylglycerol molecule (tristearin), which is assembled from three fatty acids and a glycerol molecule. a particular temperature depends on a balance between these attractive interactions and the disorganizing influence of the thermal energy (Section 4.2.3). 4.2.2 Bulk Physicochemical Properties The bulk physicochemical properties of edible fats and oils depend on the molecular structure and interactions of the triacylglycerol molecules that they contain (Marangoni and Wesdorp 2012). The strength of the attractive interactions between molecules and the effectiveness of their packing in a condensed phase determines their melting point, density, and rheology (Israelachvili 2011). Triacylglycerols that contain branched or unsaturated fatty acids are not able to pack as closely together as those that contain linear saturated fatty acids, and so they have lower densities and higher compressibilities than saturated triacylglycerols (Coupland and McClements 1997, Walstra 2003). The temperature at which a triacylglycerol melts also depends on the packing of the molecules: the more effective the packing the higher the melting point (Walstra 2003, Israelachvili 2011). Thus, the melting points of triacylglycerols increase with increasing chain length; are higher for saturated than for unsaturated fatty acids; are higher for straight chained than branched fatty acids; and are higher for triacylglycerols with a more symmetrical distribution of fatty acids on the glycerol molecule (Table 4.1). Triacylglycerol molecules have a relatively low dielectric constant because of their low polarity (Table 4.2). Knowledge of the dielectric constant of oils is important because it influences the range and magnitude of the colloidal interactions between droplets in emulsions, especially the van der Waals and electrostatic interactions (Chapter 3). Many of the bulk physicochemical properties of edible fats and oils have an important influence on the formation and stability of food emulsions. The creaming stability of emulsions depends on the density contrast between the oil and aqueous phases, and hence changes in the density of the oil phase may cause changes in the long-term stability of an emulsion (Chapter 7). The minimum size of droplets that can be produced by some homogenizers depends on the ratio of the viscosity of the dispersed phase to that of the continuous phase (Chapter 6). The viscosity of edible lipids decreases appreciably with temperature, and the nature of the viscosity–temperature profile depends on lipid type and composition. 103 Emulsion Ingredients TABLE 4.1 Melting Points and Heats of Fusion of the Most Stable Polymorphic Forms of Selected Triacylglycerol Molecules Triglyceride LLL MMM PPP SSS OOO LiLiLi LnLnLn SOS SOO Melting Point (°C) ΔHf /J g−1 46 58 66 73 5 −13 −24 43 23 186 197 205 212 113 85 — 194 — Note: L, lauric acid (C12:0); M, myristic acid (C14:0); P, palmitic acid (C16:0); S, stearic acid (C16:0); O, oleic acid (C18:1); Li, Linoleic (C18:2); Ln, Linolenic (C18:3). TABLE 4.2 Comparison of Some Bulk Physicochemical Properties of Liquid Oil (Triolein) and Water at 20°C Molecular weight Melting point (°C) Density (kg m−3) Compressibility Viscosity (mPa s) Thermal conductivity (W m−1 K−1) Specific heat capacity (J kg−1 K−1) Thermal expansion coefficient (°C−1) Dielectric constant Surface tension (mN m−1) Refractive index Oil Water 885 5 910 5.03 × 10−10 ≈50 0.170 1980 7.1 × 10−4 3 ≈35 1.46 18 0 998 4.55 × 10−10 1.002 0.598 4182 2.1 × 10−4 80.2 72.8 1.333 Hence, the ability to produce an emulsion containing small droplets may depend on the nature of the oil used, as well as on the homogenization conditions utilized in the emulsion preparation, for example, pressure and temperature. The interfacial tension of an oil–water interface may also influence the size of the droplets produced during homogenization, since droplet disruption usually becomes easier as the interfacial tension decreases (Chapter 6). The interfacial tension may also affect the long-term stability of emulsions by influencing the composition and properties of the interface formed. The interfacial tension of an oil depends on the polarity of the major lipid molecules present (e.g., triacylglycerols or terpenes), as well as on the presence of any minor surface-active components (e.g., free fatty acids, monoacylglycerols, diacylglycerols, or phospholipids). There can be significant variations in the interfacial tensions produced by oils depending on their origin and purity. Oil polarity may also influence the partitioning of functional constituents (such as flavors, antioxidants, preservatives, or colors) between the oil and aqueous phases, which may alter the physicochemical or sensory properties of the system. The strength and range of the colloidal interactions between the droplets in emulsions are determined by the dielectric constant and refractive index of the component phases (Chapter 3). The appearance of an emulsion depends on the scattering of light by the droplets and the absorption of light by any chromophores present (Chapter 10); hence, utilization of oils of different refractive index or color may lead to differences in emulsion appearance. Some oils are digestible by gastrointestinal enzymes, whereas others are indigestible. In summary, differences in the bulk physicochemical properties of oils can cause appreciable changes in the stability and properties of food emulsions. 104 Food Emulsions: Principles, Practices, and Techniques The bulk physicochemical properties of many liquid triacylglycerol oils are fairly similar (Coupland and McClements 1997), and so the choice of oil type may not have a major influence on the overall properties of many emulsion-based food products. Nevertheless, certain oil types do have significantly different properties, which may influence their functional performance in emulsions. This may be particularly important when trying to replace one type of oil with another chemically different type, for example, oil rich in monosaturated lipids, with oil rich in polyunsaturated lipids, or conventional oil with a fat substitute. It should be noted that many studies aimed at establishing the colloidal basis of emulsion properties have used simple model systems containing highly purified oils with known chemical structures, for example, hydrocarbons. These model oils may facilitate the interpretation of experimental data, but one should be careful to ensure that conclusions drawn from these model systems do actually apply to real food emulsions. 4.2.3 Fat Crystallization One of the most important characteristics of fats and oils is their ability to undergo solid–liquid phase transitions at temperatures that occur during the processing, storage, and handling of food emulsions (Walstra 2003, Marangoni and Tang 2008, Marangoni et al. 2012). The texture, mouthfeel, stability, and appearance of many food emulsions depend on the physical state of the lipid phase. The conversion of milk into butter relies on the controlled destabilization of an oil-in-water emulsion (milk) into a waterin-oil emulsion (butter), which is initiated by the formation of crystals in the milk fat globules (Walstra 2003). The spreadability of the butter produced by this process is governed by the amount and type of fat crystals formed (Wright et al. 2001). If the amount of fat crystals is too high the product is firm and difficult to spread, and if it is too low the product is too soft and tends to collapse under its own weight. The creation of food emulsions with desirable properties therefore depends on an understanding of the major factors that influence the crystallization and melting of lipids in foods (Birker and Padley 1987). The arrangement of triacylglycerol molecules in the solid and liquid state is shown schematically in Figure 4.2. The physical state of a triacylglycerol at a particular temperature depends on its free energy, which is made up of contributions from enthalpy and entropy terms: ΔG S→L = ΔHS→L − TΔS S→L (Atkins and dePaula 2014). The enthalpy term (ΔHS→L) represents the change in the overall strength of the molecular interactions between the triacylglycerols when they are converted from a solid to a liquid, whereas the entropy term (ΔS S→L) represents the change in the organization of the molecules that is brought about by the melting process (Walstra 2003). The strength of the bonds between the molecules is greater in the solid state than in the liquid state because the molecules are able to pack more efficiently, and so ΔHS→L is positive, which favors the solid state. On the other hand, the entropy of the molecules in the liquid state is greater than that in the solid state, and therefore ΔS S→L is positive, which favors the liquid state. At low temperatures, the enthalpy term dominates the entropy term (ΔHS→L > TΔS S→L), and therefore the solid state has the lowest free energy. As the temperature increases, the Liquid oil Solid fat Melt Crystallize Lower entropy stronger interactions Higher entropy weaker interactions FIGURE 4.2 The arrangement of triacylglycerols in the solid and liquid states depends on a balance between the organizing influence of the attractive interactions between the molecules and the disorganizing influence of the thermal energy. 105 Emulsion Ingredients entropy contribution becomes increasingly important. Above a certain temperature, known as the melting point, the entropy term dominates the enthalpy term (TΔS S→L > ΔHS→L) and so the liquid state has the lowest free energy. A material therefore changes from a solid to a liquid when its temperature is raised above the melting point. A solid-to-liquid transition (melting) is endothermic because energy must be added to the system to pull the molecules further apart. Conversely, a liquid-to-solid transition (crystallization) is exothermic because energy is released as the molecules come closer together. The endothermic nature of fat crystal melting is the reason for the cooling sensation that is perceived when fatty foods melt in the mouth (Walstra 2003). The temperature dependence of the free energies of the solid and liquid states shows that below the melting point the solid state has the lowest free energy, but above it the liquid state has the lowest (Figure 4.3). Thermodynamics informs us whether or not a phase transition can occur, but it tells us nothing about the rate at which this process occurs or about the physical mechanism by which it is accomplished (Atkins and dePaula 2014). As seen below, an understanding of lipid phase transitions requires knowledge of both the thermodynamics and kinetics of the process. The crystallization of fats can be conveniently divided into three stages: supercooling, nucleation, and crystal formation (Marangoni et al. 2012). 4.2.3.1 Supercooling Crystallization can only take place after a liquid phase is cooled below its thermodynamic melting point (Hartel 2001, Kashchiev and van Rosmalen 2003, Walstra 2003). Even so, a material can persist as a liquid below its melting point for a considerable time before any crystallization is observed. This is because of an activation energy that must be overcome before the liquid–solid phase transition can occur (Figure 4.4). If the magnitude of this activation energy is sufficiently high compared to the thermal energy (kT) of the system, then crystallization will not occur even though the transition is thermodynamically favorable. The supercooled liquid is then said to exist in a metastable state. The height of the activation energy depends on the ability of crystal nuclei to be spontaneously formed in the liquid oil that are stable enough to grow into crystals (see Section 4.2.3.2). The degree of supercooling of a liquid is defined as ΔT  =  Tmp −  T, where T is the temperature and Tmp is the melting point. The value of ΔT at which crystallization is first observed depends on the chemical structure of the oil, the presence of any contaminating materials, the cooling rate, microstructure (e.g., bulk versus emulsified oil), and the application of external forces (Kashchiev and van Rosmalen 2003, Lindfors et al. 2008). Pure oils containing no impurities can often be supercooled by more than 10°C before any crystallization is observed (Vanapalli et al. 2002, Degner et al. 2014). Free energy Melting point GS Solid favorable Liquid favorable GL Temperature FIGURE 4.3 Temperature dependence of the free energies of the solid and liquid states. At low temperatures the solid state is thermodynamically favorable, but above the melting point the liquid state is more favorable. 106 Food Emulsions: Principles, Practices, and Techniques Nuclei formation Liquid Energy barrier ΔG* Disordered ΔG Solid T < Tm Ordered FIGURE 4.4 When there is a sufficiently high activation energy between the solid and liquid states, a liquid oil can persist in a metastable state below the melting point of a fat. 4.2.3.2 Nucleation Crystal growth can only occur after stable nuclei have been formed in a liquid (Hartel 2001, Himawan et al. 2006, Hartel 2013). These nuclei are believed to consist of clusters of oil molecules that transiently exist as small ordered regions within the liquid and are formed when a number of oil molecules collide and become spontaneously associated with each other. There is a free energy change associated with the formation of one of these nuclei. Below the melting point, the bulk crystalline state is thermodynamically favorable, and so there is a decrease in free energy when some of the oil molecules in the liquid cluster together to form a nucleus. This negative (favorable) free energy (ΔGV) change is proportional to the volume of the nucleus formed. On the other hand, the formation of a nucleus leads to the creation of a new interface between the solid and liquid phases which requires an input of free energy to overcome the interfacial tension (Chapter 5). This positive (unfavorable) free energy (ΔG S) change is proportional to the surface area of the nucleus formed. The total free energy change associated with the formation of a nucleus is therefore a combination of a volume and a surface term (Hartel 2001, Walstra 2003): DG = DGV + DGS = 4 3 DH fus DT pr + 4pr 2 g i Tmp 3 (4.1) where r is the radius of the nuclei ΔHfus is the enthalpy change of fusion per unit volume associated with the liquid–solid transition (which is negative) γi is the solid–liquid interfacial tension The volume contribution becomes increasingly negative as the size of the nucleus increases, whereas the surface contribution becomes increasingly positive (Figure 4.5). The surface contribution dominates for small nuclei, whereas the volume term dominates for large nuclei. The overall free energy has a maximum value at a certain critical nucleus radius (r*): 107 Emulsion Ingredients ΔG Unfavorable ΔG associated with creation of new surface ΔGs Spontaneous nuclei formation Unstable nuclei Stable nuclei r r* ΔG ΔGv Favorable ΔG associated with formation of fat crystal volume FIGURE 4.5 The critical size of a nucleus required for crystal growth depends on a balance between the volume and surface contributions to the free energy of nuclei formation. dDG DH fus DT = 4pr 2 + 8prg i = 0 dr Tmp (4.2) This equation can be rearranged to give an expression for the critical radius of the nucleus that must be achieved for crystallization to occur: r* = 2g iTmp DH fus DT (4.3) If a nucleus is formed that has a radius below this critical size (r < r*), it will tend to dissociate so as to reduce the free energy of the system. On the other hand, if a nucleus is formed that has a radius above this critical value, it will tend to grow into a crystal. This equation indicates that the critical size of nuclei required for crystal growth decreases as the degree of supercooling increases, which accounts for the increase in nucleation rate with decreasing temperature. The rate at which nucleation occurs can be related to the activation energy ΔG* that must be overcome before a stable nuclei is formed (Boistelle 1988): æ -DG * ö J = A exp ç ÷ è kT ø (4.4) where J is the nucleation rate, which is equal to the number of stable nuclei formed per second per unit volume of material A is a pre-exponential factor k is Boltzmann’s constant T is the absolute temperature 108 Food Emulsions: Principles, Practices, and Techniques J The value of ΔG* is calculated by replacing r in Equation 4.1 with the critical radius given in Equation 4.3. The variation of the nucleation rate predicted by Equation 4.4 with the degree of supercooling (ΔT) is shown in Figure 4.6. The formation of stable nuclei is negligibly slow at temperatures just below the melting point, but increases dramatically when the liquid is cooled below a certain temperature, T*. In reality, the nucleation rate increases with cooling up to a certain temperature, but then decreases on further cooling. This is because the increase in viscosity of the oil that occurs as the temperature is decreased slows down the diffusion of oil molecules toward the liquid–nucleus interface (Boistelle 1988, Hartel 2001). Consequently, there is a maximum in the nucleation rate at a particular temperature (Figure 4.6). The type of nucleation described above occurs when there are no impurities present in the oil and is usually referred to as homogeneous nucleation (Himawan et al. 2006, Hartel 2013). If the liquid oil is in contact with foreign surfaces, such as the surfaces of dust particles, fat crystals, oil droplets, air bubbles, reverse micelles, or the vessel containing the oil, then nucleation can be induced at a higher temperature than expected for a pure system (Hartel 2001). Nucleation due to the presence of these foreign surfaces is referred to as heterogeneous nucleation and can be divided into two types: primary and secondary (Smith et al. 2011). Primary heterogeneous nucleation occurs when the foreign surfaces have a different chemical structure to that of the oil, whereas secondary heterogeneous nucleation occurs when the foreign surfaces are crystals with the same chemical structure as the liquid oil. Heterogeneous nucleation occurs when the impurities provide a surface where the formation of stable nuclei is more thermodynamically favorable than in the pure oil. As a result the degree of supercooling required to initiate fat crystallization is reduced. On the other hand, certain types of impurities are capable of decreasing the nucleation rate of oils because they are incorporated into the surface of the growing nuclei and prevent any further oil molecules being incorporated (Smith et al. 2011). Whether an impurity acts as a catalyst or an inhibitor of nucleation depends on its molecular structure and interactions with the nuclei. It should be noted that there is still considerable debate about the mathematical modeling of nucleation, since existing theories often give predictions of nucleation rates that are greatly different from experimental measurements (Walstra 2003, Vekilov 2010, Hartel 2013). Nevertheless, the general form of the dependence of nucleation rates on temperature are predicted fairly well by existing theories. ∆T* Supercooling, ∆T FIGURE 4.6 Theoretically, the rate of the formation of stable nuclei increases with supercooling (solid line), but in practice, the nucleation rate decreases below a particular temperature because the diffusion of oil molecules is retarded by the increase in oil viscosity (broken line). 109 Emulsion Ingredients 4.2.3.3 Crystal Growth Once stable nuclei have been formed they grow into crystals by incorporating molecules from the liquid oil at the solid–liquid interface (Hartel 2001, Walstra 2003, Himawan et al. 2006). Crystals have different faces, and each face may grow at a different rate, which partially accounts for the wide variety of different crystal shapes that can be formed by fats, for example, needles and spherulites. The overall crystal growth rate depends on a number of factors, including mass transfer of the liquid molecules to the solid–liquid interface, mass transfer of noncrystallizing species away from the interface, incorporation of the liquid molecules into the crystal lattice, or removal of the heat generated by the crystallization process from the interface (Hartel 2001). Any of these processes can be rate limiting depending on the molecular characteristics of the system and the prevailing environmental conditions, for example, temperature profile and mechanical agitation. Consequently, a general theoretical model of crystal growth is difficult to construct. In crystallizing lipid systems, the incorporation of a molecule at the crystal surface is often rate limiting at high temperatures, whereas the diffusion of a molecule to the solid–liquid interface is often rate limiting at low temperatures. This is because the viscosity of the liquid oil increases as the temperature is lowered and so the diffusion of a molecule is retarded. The crystal growth rate therefore increases initially with supercooling, has a maximum rate at a certain temperature, and then decreases on further supercooling (Hartel 2001). The dependence of the growth rate on temperature therefore shows a similar general trend to the nucleation rate, but the shape of the two curves is different (Figure 4.7). Experimentally, it has been observed that the rate of crystal growth is proportional to the degree of supercooling, and inversely proportional to the viscosity of the melt (Timms 1991). A variety of mathematical theories have been developed to model the rate of crystal growth in crystallizing fats (Hartel 2001, Himawan et al. 2006). The most appropriate model for a specific situation depends on the rate limiting step for that particular system under the prevailing environmental conditions, for example, mass transfer of the liquid molecules to the solid–liquid interface, mass transfer of Nucleation or growth rate Nucleation Growth Supercooling (Tmp – T) FIGURE 4.7 The size of crystals produced when a melted fat is cooled below its melting temperature depends on the relative rates of nucleation and crystal growth. 110 Food Emulsions: Principles, Practices, and Techniques noncrystallizing species away from the interface, incorporation of the liquid molecules into the crystal lattice, or removal of the heat generated by the crystallization process from the interface. Once crystallization is complete, it is still possible for there to be changes in crystal size and morphology during storage due to postcrystallization processes such as crystal aggregation and Ostwald ripening (Hartel 2001, 2013, Walstra 2003, Marangoni et al. 2012). Crystal aggregation occurs when two or more crystals come together and form a larger crystal, whereas Ostwald ripening occurs when oil molecules migrate from smaller crystals to larger ones through the intervening medium. Aggregation and Ostwald ripening therefore both lead to an increase in the average size of the crystals present within a fat. Crystal growth during storage is often undesirable since it adversely affects the physicochemical and sensory properties of the final product (Walstra 2003). 4.2.3.4 Crystal Morphology The morphology of the crystals formed depends on a number of internal factors (e.g., molecular structure, composition, packing, and interactions) and external factors (e.g., temperature–time profile, mechanical agitation, and impurities). In general, when a liquid oil is cooled rapidly to a temperature well below its melting point a large number of small crystals are formed, but when it is cooled slowly to a temperature just below its melting point a smaller number of larger crystals are formed (Hartel 2001, Walstra 2003, Degner et al. 2014). This is because the nucleation rate increases more rapidly with decreasing temperature than the crystallization rate (Figure 4.7). Thus, rapid cooling produces many nuclei simultaneously that subsequently grow into small crystals, whereas slow cooling produces a smaller number of nuclei that have time to grow into larger crystals before further nuclei are formed. Crystal size has important implications for the rheology and organoleptic properties of many types of food emulsion. When crystals are too large they are perceived as being “grainy” or “sandy” in the mouth (Walstra 2003). The efficiency of molecular packing in crystals also depends on the cooling rate. If a fat is cooled slowly, or the degree of supercooling is small, then the molecules have sufficient time to be efficiently incorporated into a crystal. At faster cooling rates, or higher degrees of supercooling, the molecules do not have sufficient time to pack efficiently before another molecule is incorporated. Thus, rapid cooling tends to produce crystals that contain more dislocations, and in which the molecules are less densely packed (Timms 1991). The cooling rate therefore has an important impact on the morphology and functional properties of crystalline lipids in foods. 4.2.3.5 Polymorphism Triacylglycerols exhibit a phenomenon known as polymorphism, which is the ability of a material to exist in a number of crystalline structures with different molecular packing (Hartel 2001, Himawan et al. 2006, Lee et al. 2011). The three most commonly occurring types of packing in triacylglycerols are hexagonal, orthorhombic, and triclinic, which are usually designated as α, β′, and β polymorphic forms, respectively. The thermodynamic stability of the three forms decreases in the order: β > β′ > α. Even though the β form is the most thermodynamically stable, triacylglycerols often crystallize in one of the metastable states because they have a lower activation energy for nuclei formation (Figure 4.8). With time the crystals transform to the most stable state at a rate that depends on environmental conditions, such as temperature, pressure, and the presence of impurities (Timms 1991, Smith et al. 2011). Polymorphic transitions often occur at a different rate in emulsified fats than in bulk fats, for example, the α- to β-transition of tripalmitin is much faster in emulsions (Helgason et al. 2008). In addition, the morphology and spatial arrangement of the crystals formed in emulsified fats is often different from those formed in bulk fats, which has been attributed to differences in heat transfer rates when crystallizing fats are surrounded by water rather than by oil and because of the physical limitations imposed by the droplet surfaces (Walstra 2003, McClements 2012). Knowledge of the polymorphic form of the crystals in an emulsified fat is often important because it can impact the physicochemical and sensory properties of food emulsions. 111 Emulsion Ingredients ΔG* Melt α G β΄ β FIGURE 4.8 The polymorphic state that is initially formed when an oil crystallizes depends on the relative magnitude of the activation energies associated with nuclei formation. 4.2.3.6 Crystallization of Edible Fats and Oils The melting point of a triacylglycerol depends on the chain length and degree of unsaturation of its constituent fatty acids, as well as their relative positions along the glycerol molecule (Table 4.1). Edible fats and oils contain a complex mixture of many different types of triacylglycerol molecules, each with a different melting point, and so they usually melt over a wide range of temperatures, rather than at a distinct temperature as would be the case for a pure triacylglycerol (Figure 4.9). The melting profile of a fat is not simply the weighted sum of the melting profiles of its constituent triacylglycerols, because high melting point triacylglycerols are soluble in lower melting point ones (Walstra 2003, McClements 2012). For example, in a 50:50 mixture of tristearin and triolein it is possible to 100 Pure triglyceride SFC (%) 80 60 40 20 0 Fatty food 0 20 40 60 80 100 Temperature (°C) FIGURE 4.9 Comparison of the melting profile of a pure triacylglycerol and a typical edible fat. The edible fat melts over a much wider range of temperatures because it consists of a mixture of many different pure triacylglycerol molecules each with different melting points. 112 Food Emulsions: Principles, Practices, and Techniques dissolve 10% of solid tristearin in liquid triolein at 60°C. The solubility of a solid component in a liquid component can be predicted assuming they have widely differing melting points (>20°C): é 1 D 1ù ln x = H fus ê - ú R êë T mp T úû (4.5) where x is the solubility, expressed as a mole fraction, of the higher melting point component in the lower melting point component ΔHfus is the molar heat of fusion (Walstra 1987) The structure and physical properties of crystals produced by cooling a complex mixture of triacylglycerols is strongly influenced by the cooling rate and temperature (Walstra 2003). If an oil is cooled rapidly all the triacylglycerols crystallize at approximately the same time and a solid solution is formed, which consists of homogeneous crystals in which the triacylglycerols are intimately mixed with each other. On the other hand, if the oil is cooled slowly the higher melting point triacylglycerols crystallize first, whereas the low melting point triacylglycerols crystallize later, and so mixed crystals are formed. These crystals are heterogeneous and consist of some regions that are rich in high melting point triacylglycerols and other regions that are depleted in these triacylglycerols. Whether a crystalline fat forms mixed crystals or a solid solution influences many of its physicochemical properties, such as density, compressibility, and melting profile (Walstra 2003), which could have an important influence on the properties of a food emulsion. Once a fat has crystallized the individual crystals may aggregate to form a three-dimensional network that traps liquid oil through capillary forces (Marangoni et al. 2012). The interactions responsible for crystal aggregation in pure fats are primarily van der Waals interactions between the solid fat crystals (Marangoni et al. 2012). Once aggregation has occurred, the fat crystals may partially fuse together which strengthens the crystal network (Johansson and Bergenstahl 1995, Walstra 2003). The system may also change over time due to the growth of larger crystals at the expense of smaller ones, that is, Ostwald ripening (Chapter 7). 4.2.3.7 Fat Crystallization in Emulsions The influence of fat crystallization on the bulk physicochemical properties of food emulsions depends on whether the fat forms the continuous phase or the dispersed phase. The characteristic stability and rheological properties of water-in-oil emulsions, such as butter and margarine, is determined by the presence of a network of aggregated fat crystals within the continuous (oil) phase (Marangoni et al. 2012). The fat crystal network is responsible for preventing the water droplets from sedimenting under the influence of gravity, as well as determining the spreadability of the product. If there are too many fat crystals present the product is too firm and difficult to spread, but when there are too few crystals present the product is soft and collapses under its own weight. Selection of a fat with the appropriate melting characteristics is therefore one of the most important aspects of margarine and spread production (Gunstone 2008). The melting profile of natural fats can be optimized for specific applications by various physical or chemical methods, including blending, interesterification, fractionation, and hydrogenation (Gunstone 2008, Belitz et al. 2009, Brady 2013). Fat crystallization also has a pronounced influence on the physicochemical properties of many oilin-water emulsions, such as milk, cream, or salad dressings (Fredrick et al. 2010, McClements 2012). When the fat droplets are partially crystalline, a crystal from one droplet can penetrate into another droplet during a collision which causes the two droplets to stick together (Fredrick et al. 2010). This phenomenon is known as partial coalescence and leads to a dramatic increase in the viscosity of an emulsion, as well as a decrease in the stability to creaming (Chapter 7). Extensive partial coalescence can eventually lead to phase inversion, that is, conversion of an oil-in-water emulsion to a water-in-oil emulsion (Thivilliers-Arvis et al. 2010, Buldo et al. 2013). This process is one of the most important Emulsion Ingredients 113 steps in the production of butters, margarines, and spreads (Walstra 1987). Partial coalescence is also important in the production of ice cream and whipped creams, where an oil-in-water emulsion is cooled to a temperature where the emulsified fat partially crystallizes, and is then mechanically agitated to promote droplet collisions and aggregation (Goff 2002, Goff and Vega 2007, Bazmi and Relkin 2009). The aggregated droplets form a shell around the air bubbles and a three-dimensional network in the continuous phase that contribute to the stability and texture of the product (Goff 1997, Goff and Hartel 2013). 4.2.4 Chemical Changes The type and concentration of molecules within the lipid phase can change with time due to chemical reactions. The two most important chemical changes that occur in edible fats and oils are lipolysis and oxidation (Belitz et al. 2009, Brady 2013). Lipolysis is the process where ester bonds of fats and oils are hydrolyzed by certain enzymes, or by a combination of heat and moisture. The result of lipolysis is the liberation of free fatty acids, which can be either detrimental or desirable to food quality. Lipolysis has deleterious effects on the quality of some food products because it leads to the generation of rancid offflavors and off-odors, which is known as “hydrolytic rancidity.” In addition, free fatty acids are more surface active than triacylglycerols and therefore accumulate preferentially at an oil–water or air–water interface, which increases their susceptibility to oxidation and may increase the tendency for emulsion droplets to coalesce (Coupland and McClements 1996, McClements and Decker 2000, Waraho et al. 2011). On the other hand, a limited amount of lipolysis is beneficial to the quality of some foods because it leads to the formation of desirable flavors and aromas, for example, cheese and yogurt (McSweeney 2004, Cheng 2010). Many food emulsions contain polyunsaturated lipids that are highly susceptible to lipid oxidation (Jacobsen et al. 2008, Waraho et al. 2010). Indeed, lipid oxidation is one of the most serious causes of quality deterioration in many foods because it leads to the generation of undesirable off-flavors and off-odors (“oxidative rancidity”), as well as potentially toxic reaction products. In other foods, a limited amount of lipid oxidation is beneficial because it leads to the generation of a desirable flavor profile, for example, cheese. The term lipid oxidation describes an extremely complex series of chemical reactions that involves unsaturated lipids and oxygen (Akoh and Min 2008, Belitz et al. 2009). It has proved convenient to divide these reactions into three different types: initiation, propagation, and termination. Initiation occurs when a hydrogen atom is extracted from the methylene group (–CH=CH–) of a polyunsaturated fatty acid, leading to the formation of a free radical (–CH = C.–). This process can be started by a variety of different initiators that are present in foods, including naturally occurring lipid peroxides, transition metal ions, UV light, and enzymes. It is worthwhile noting that many of these initiators are predominantly water soluble, which has important implications for the oxidation of emulsified oils, because the initiator must either travel through or interact across the interfacial membrane in order to come into contact with the oil (Coupland and McClements 1996, McClements and Decker 2000, Waraho et al. 2011). Once a free radical has formed, it reacts with oxygen to form a peroxy radical (–CH–COO.–). These radicals are highly reactive and can extract hydrogen atoms from other unsaturated lipids and therefore propagate the oxidation reaction. Termination occurs when two radicals interact with each other to form a nonradical, and thus end their role as propagators of the reaction. During lipid oxidation a number of decomposition reactions occur simultaneously, which leads to the formation of a complex mixture of reaction products, including aldehydes, ketones, alcohols, and hydrocarbons. Many of these products are volatile and therefore contribute to the characteristic odor associated with lipid oxidation. Some of the products are surface active and would therefore accumulate at oil–water interfaces in emulsions, whereas others are water soluble and would therefore leach into the aqueous phase of emulsions (McClements and Decker 2000). The growing trend of incorporating polyunsaturated lipids into food products in order to improve their nutritional profiles has meant that there has been a considerable research effort to elucidate the relationship between emulsion properties and lipid oxidation, and to develop effective strategies to inhibit oxidation in foods (Jacobsen et al. 2008, Waraho et al. 2011, Berton-Carabin et al. 2014). Some of the work carried out in this area is discussed in the chapter on emulsion stability (Chapter 7). 114 Food Emulsions: Principles, Practices, and Techniques 4.2.5 Selection of an Appropriate Lipid A variety of edible fats and oils are available for utilization in food emulsions, and the choice of the most appropriate type for a particular application depends on the nutritional, physicochemical, and sensory characteristics desired for that specific product (Gunstone 2008). Some of the most important characteristics to consider when selecting a lipid source are briefly highlighted below. 4.2.5.1 Nutritional Profile A major trend in the food industry is to improve the healthiness of foods, either by decreasing the level of those lipids associated with health problems (such as cholesterol, saturated fats, and trans fatty acids) or by increasing the level of those lipids associated with improved human health and wellness (such as polyunsaturated fats, ω-3 fatty acids, oil-soluble vitamins, and nutraceuticals). Many food manufacturers are therefore reformulating their products to replace existing oil sources with lipids with more healthful nutritional profiles. These more healthful lipids could be oils from different natural sources (e.g., fish oils, flaxseed oils, or algae oils), modified oils (e.g., chemically, physically, enzymatically, or genetically modified oils), or fat substitutes with low calorific values (e.g., OlestraTM). Nevertheless, changing the nutritional profile of an oil may also cause appreciable changes in its physicochemical and sensory characteristics (e.g., flavor profile, crystallization characteristics, and viscosity), which may adversely influence its functional attributes within a specific product (e.g., emulsion formation, stability, or properties). For this reason, research is currently being carried out to produce emulsions containing oils with improved nutritional profiles, but which also maintain their desirable functional attributes. Advances in the development of emulsion-based delivery systems for oil-soluble vitamins and nutraceuticals are discussed in Chapter 13. 4.2.5.2 Flavor Profile Triacylglycerols are relatively large molecules that have a low volatility and hence little inherent flavor. Nevertheless, different natural sources of edible fats and oils do have distinctive flavor profiles because of the characteristic volatile breakdown products and impurities that they contain, for example, compare the aromas of corn oil, olive oil, and fish oil. Oil from a specific natural source may therefore be selected for utilization in a particular food product because it contributes to the overall flavor profile. The oil phase may also indirectly influence the flavor profile because of its ability to act as a solvent for volatile nonpolar molecules. The partitioning of flavor molecules between oil, water, and gaseous regions and their release rate during mastication depends on factors such as the polarity, viscosity, and crystallinity of the lipid phase, which may vary from one source of oil to another (Chapter 9). 4.2.5.3 Crystallization Behavior The suitability of edible fats and oils for many applications within food emulsions depends on their melting and crystallization temperatures, SFC–temperature profile, crystal morphology, and polymorphic tendency. In some emulsions, it is important that the fat does not crystallize during the lifetime of the product since this would lead to instability through partial coalescence. For example, it is important that the oils used to produce salad dressings do not crystallize (“cloud”) when exposed to refrigerator temperatures. This can be achieved either by using oil sources that naturally have low melting points, by removing high melting fractions by selective crystallization (“winterization”), or by adding components that retard crystal formation, such as oil-soluble surfactants. In other food emulsions, the crystallization of the lipid phase is an integral part of their production and determines their desirable physicochemical and sensory attributes, for example, margarine, butter, whipped cream, and ice cream. In these products, it is usually important to select an oil that has a particular SFC vs. temperature profile, and that forms crystals of the appropriate size, shape, and polymorphic form. A variety of analytical techniques are available to characterize the crystallization behavior of oils (Chapter 14). The desired crystallization characteristics can be obtained by selection of a natural oil with an appropriate triacylglycerol composition, or the Emulsion Ingredients 115 triacylglycerol composition of the oil phase can be obtained by blending, fractionation, interesterification, or hydrogenation of oils (Gunstone 2008). 4.2.5.4 Oxidative Stability Many edible fats and oils naturally contain significant quantities of polyunsaturated lipids, which are highly susceptible to lipid oxidation. Lipid oxidation leads to a reduction in the concentration of these health-promoting polyunsaturated lipids, as well as to the generation of volatile compounds that may cause an undesirable rancid flavor. Flavor oils also contain components that are susceptible to oxidative degradation reactions that lead to loss of desirable flavors and/or production of undesirable off-flavors. When selecting an oil for use in an emulsion-based food product, it is often important to ensure that it has not undergone a significant amount of lipid oxidation prior to use, and that it will have good oxidative stability throughout the lifetime of the product. Analytical tests are available to assess the extent of lipid oxidation that has already occurred in an oil and to predict the susceptibility of oils to oxidation (Barriuso et al. 2013). The oxidative stability of an emulsion can be improved by using an oil source naturally low in polyunsaturated fats or by reducing the polyunsaturated fat content of a natural oil, for example, by partial hydrogenation.* Nevertheless, many food manufacturers want to increase the concentration of polyunsaturated fats in food products because of their potential health benefits. For these products, it is important to develop effective strategies for preventing or retarding lipid oxidation during the shelf-life of the product (Waraho et al. 2011). 4.2.5.5 Bulk Physicochemical Properties The type and concentration of molecules within an oil phase determine its bulk physicochemical properties, for example, viscosity, density, refractive index, dielectric constant, polarity, and interfacial tension. These properties may have an appreciable influence on the formation, stability, and quality of a food emulsion (Section 4.2.2). Hence, oils from different natural sources or that have been processed differently may behave differently when used in an emulsion. These differences may have to be taken into account when reformulating an emulsion to change the type of lipid used to make up the oil phase. 4.2.5.6 Oil Quality In addition to the impurities mentioned above, the oils used to prepare emulsions may contain a variety of other impurities that adversely affect their suitability for particular applications, including off-flavors, pigments, phospholipids, and free fatty acids. For this reason, components that have a negative impact on emulsion quality are usually removed from oils prior to their utilization in food products, for example, by deodorization, neutralization, degumming, and bleaching (Akoh and Min 2008). A variety of analytical procedures are routinely used by food scientists to test the quality of an oil so as to ensure that it is suitable for utilization in a product, such as classical, enzyme, spectroscopy, spectrometry, and chromatography methods (Gunstone 2008). 4.3 Water Water plays an extremely important role in determining the bulk physicochemical and organoleptic properties of food emulsions. Its unique molecular and structural properties largely determine the solubility, conformation, and interactions of the other components present in aqueous solutions (Bergethon 2010). It is therefore crucial for food scientists to understand the contribution that water makes to the overall properties of food emulsions. * It should be noted that hydrogenation leads to the production of trans fatty acids, which have been linked to human health problems. Consequently, many food manufacturers are attempting to find means of reducing the trans fatty acid content of foods. 116 Food Emulsions: Principles, Practices, and Techniques 4.3.1 Molecular Structure and Organization A water molecule is comprised of two hydrogen atoms covalently bonded to an oxygen atom (Figure 4.10). The oxygen atom is highly electronegative and pulls the electrons associated with the hydrogen atoms toward it (Norde 2011). This leaves a partial positive charge (δ+) on each of the hydrogen atoms, and a partial negative charge (δ−) on each of the lone pairs of electrons on the oxygen atom. The tetrahedral arrangement of the partial charges on an individual water molecule means that it can form hydrogen bonds with four of its nearest neighbors (Figure 4.10). A hydrogen bond is formed between a lone pair of electrons on the oxygen atom of one water molecule and a hydrogen atom on a neighboring water molecule, that is, O−Hδ + ⋯Oδ−. A hydrogen bond is actually a composite of more fundamental interactions, that is, dipole–dipole, van der Waals, steric, and partial charge transfer. The magnitude of the hydrogen bonds in water is typically between 13 and 25 kJ mol−1 (5–10 kT), which is sufficiently strong to cause the water molecules to overcome the disorganizing influence of the thermal energy and become highly aligned with each other (Israelachvili 2011). In order to maximize the number of hydrogen bonds formed, water molecules organize themselves into a three-dimensional tetrahedral structure because this allows each water molecule to form hydrogen bonds with four of its nearest neighbors. In the solid state, the number of hydrogen bonds formed per molecule is four. In the liquid state, the disorganizing influence of the thermal energy means that the number of hydrogen bonds per molecule is between about 3 and 3.5 at room temperature, and decreases with increasing temperature. The three-dimensional tetrahedral structure of water in the liquid state is highly dynamic, with hydrogen bonds continually being broken and reformed as the water molecules move about. Water molecules that dissociate to form ions, such as H3O+ and OH−, do not fit into the normal tetrahedral structure of water; nevertheless, they have little effect on the overall structure and properties of water because there concentration is so low (Fennema 2008). As well as forming hydrogen bonds with each other, water molecules are also capable of forming them with other polar molecules, such as organic acids, bases, proteins, and carbohydrates. The strength of these interactions varies from about 2 to 40 kJ mol−1 (1–16 kT) depending on the electronegativity and orientation of the donor or acceptor groups (Baker and Hubbard 1984). Many ions form relatively strong ion–dipole interactions with water molecules, which has a pronounced influence on the structure and physicochemical properties of water (Fennema 2008, Bergethon 2010, Israelachvili 2011). It is the ability of water molecules to form relatively strong bonds with each other and with other types of polar or ionic molecules that determines many of the characteristic properties of food emulsions. Oxygen has strongly positive nucleus (pulls electrons) Tetrahedral structure δ– δ+ H O δ– H δ+ FIGURE 4.10 Molecular structure and tetrahedral organization of water molecules. Emulsion Ingredients 117 4.3.2 Bulk Physicochemical Properties The bulk physicochemical properties of pure water are determined by the mass, dimensions, bond angles, charge distribution, and interactions of the water molecule (Fennema 2008, Israelachvili 2011, Norde 2011). Water has a high dielectric constant because the uneven distribution of partial charges on the molecule means that it is easily polarized by an electric field. It has a relatively high melting point, boiling point, enthalpy of vaporization, and surface tension, compared to other molecules of a similar size that also contain hydrogen (e.g., CH4, NH3, HF, and H2S), because a greater amount of energy must be supplied to disrupt the strong hydrogen bonds holding the water molecules together in the condensed state. The relatively low density of ice and of liquid water is because the water molecules adopt a structure in which they are in direct contact with only four of their nearest neighbors rather than forming a more close packed structure. The relatively low viscosity of water is because of the highly dynamic nature of hydrogen bonds compared to the time scale of a rheology experiment. Even though energy is required to break the hydrogen bonds between water molecules as they move past each other, most of this energy is regained when they form new hydrogen bonds with their new neighbors. The crystallization of water has a pronounced effect on the bulk physiochemical properties of food emulsions. The presence of ice crystals in the aqueous phase of an oil-in-water emulsion, such as ice cream, contributes to the characteristic mouthfeel and texture of the product (Hartel 1996, Goff 1997). When these ice crystals grow too large a product is perceived as being “grainy” or “sandy,” which is commonly experienced when ice cream is melted and then refrozen. Many emulsion-based foods are designed to be freeze–thaw stable, that is, their quality should not be adversely affected once the product is frozen and then thawed (Ghosh and Coupland 2008, Degner et al. 2014). Considerable care must be taken in the choice of ingredients and freezing/thawing conditions to create a food emulsion that is freeze–thaw stable. The basic principles of ice crystallization are similar to those described for fats and oils (Section 4.2.3). Nevertheless, water does exhibit some anomalous behavior because of its unique molecular properties, for example, it expands when it crystallizes, whereas most other substances contract. This is because the increased mobility of the water molecules in the liquid state means that they can get closer together, and so the density of the liquid state is actually greater than that of the solid state. Some of the most important bulk physicochemical properties of liquid water are compared with those of a liquid oil in Table 4.2. A more detailed discussion of the molecular basis of the physicochemical properties of water in relation to food quality is given by Fennema (2008). 4.3.3 Influence of Solutes on the Organization of Water Molecules The aqueous phase of most food emulsions contains a variety of water-soluble constituents, including minerals, acids, bases, flavors, preservatives, vitamins, sugars, surfactants, proteins, and polysaccharides. The solubility, partitioning, conformation, interactions, and chemical reactivity of many of these food ingredients are determined by their interactions with water. It is therefore important for food scientists to understand the nature of solute–water interactions and their influence on the bulk physicochemical and organoleptic properties of food emulsions. When a solute molecule is introduced into pure water, the normal structural organization and interactions of the water molecules are altered. This results in changes in the physicochemical properties of the water molecules that are affected by the presence of the solute, such as density, compressibility, melting point, boiling point, and mobility (Fennema 2008, Israelachvili 2011, Norde 2011, Brady 2013). The extent of these changes depends on the molecular characteristics of the solute, that is, its size, shape, and polarity. The water molecules in the immediate vicinity of the solute experience the largest modification of their properties and are often referred to as being “bound” to the solute. In reality, these water molecules are not permanently bound to the solute, but rapidly exchange with the bulk water molecules, albeit with a reduced mobility. The mobility of “bound” water increases as the strength of the attractive interactions between it and the solute decreases, that is, nonpolar–water > dipole–water > ion–water. The amount of water “bound” to a solute can be defined as the number of water molecules whose properties are significantly altered by its presence. In practice, it is difficult to unambiguously define or stipulate the amount of “bound” water. First, the water molecules “bound” to a solute do not all have the same 118 Food Emulsions: Principles, Practices, and Techniques properties: the water molecules closest to the solute are more strongly influenced by its presence than those furthest away. Second, the physicochemical properties that are measured in order to determine the amount of “bound” water are each influenced to a different extent (e.g., density, compressibility, mobility, and melting point). As a consequence, different analytical techniques often measure different amounts of “bound” water, depending on the physical principles on which they operate. 4.3.3.1 Interaction of Water with Ionic Solutes Many of the solutes present in food emulsions are either ionic are capable of being ionized, including salts, acids, bases, proteins, and polysaccharides. The degree of ionization of many of these solutes is governed by the pH of the surrounding aqueous solution, and so their interactions are particularly sensitive to pH. The ion–dipole interactions that occur between an ionic solute and a water molecule are usually stronger than the dipole–dipole interactions that occur between a pair of water molecules (Table 4.3). As a consequence, the water molecules in the immediate vicinity of an ion tend to orientate themselves so that their oppositely charged dipole faces the ion. Thus, a positively charged ion causes the water molecules to align themselves so that a δ− group faces the ion, whereas the opposite is true for a negatively charged ion (Figure 4.11). The relatively strong nature of ion–dipole interactions means that the mobility of the water molecules near the surface of an ion is significantly less than that of bulk water. The residence time of a water molecule in the vicinity of an ionic group is ≈10 −8 s, whereas it is ≈10 −11 s in bulk water. The influence of an ion on the mobility and alignment of the water molecules is greatest at its surface because the electric field is strongest there. As one moves away from the ion surface, the strength of the electric field diminishes, so that the ion–water interactions become progressively weaker. Thus, the water molecules become more mobile and are less likely to be aligned toward the ion. At a sufficiently large distance from the ion surface the water molecules are uninfluenced by its presence and have properties similar to those of bulk water. Alterations in the structural organization and interactions TABLE 4.3 Typical Water–Solute Interactions Found in Food Emulsions Interaction Type Water–ion Water–dipole Water–nonpolar Examples Strength Compared to Water–Water Interactions Free ions (Na+, Cl−) Ionic groups (–CO2−, –NH3+) –C = 0, –NH, –OH Alkyl group Greater Similar Much smaller 2δ– δ+ δ+ Na+ Cl– FIGURE 4.11 Organization of water molecules around ions in aqueous solutions. 119 Emulsion Ingredients of water molecules in the vicinity of an ion cause significant changes in the physicochemical properties of water. The water that is “bound” to an ionic solute is denser, less mobile, less compressible, and has a lower freezing point, and a higher boiling point than bulk water. Most ionic solutes have a high water solubility because the formation of many ion–dipole bonds in an aqueous solution helps to compensate for the loss of the strong ion–ion bonds in the crystals, which is coupled with the favorable entropy of mixing contribution (Chapter 2). Advances in the development of experimental and computational methods for providing detailed information about the structural organization of ions and water in aqueous solutions have been reviewed (Bowron and Moreno 2014). The number of water molecules whose mobility and structural organization is altered by the presence of an ion increases as the strength of its electric field increases (Israelachvili 2011, Norde 2011). The strength of the electric field generated by an ion is determined by its charge divided by its radius. Thus, ions that are small and/or multivalent generate strong electric fields that influence the properties of the water molecules up to relatively large distances from their surface, for example, Li+, Na+, H3O+, Ca2+, Ba2+, Mg2+, Al3+, and OH−. On the other hand, ions that are large and/or monovalent generate relatively weak electrical fields, and therefore their influence extends a much shorter distance into the surrounding water, for example, K+, Rb+, Cs+, NH4+, Cl−, Br−, and I−. The number of water molecules “bound” to an ion is usually referred to as the hydration number. Thus, the hydration number of small multivalent ions is usually larger than that of large monovalent ions. When an ionic solute is added to pure water, it disrupts the existing tetrahedral arrangement of the water molecules, but imposes a new order on the water molecules in its immediate vicinity (Marcus 2009, Norde 2011). The overall structural organization of the water molecules in an aqueous solution can therefore either increase or decrease after a solute is added, depending on the amount of structure imposed on the water by the ion compared to that lost by disruption of the tetrahedral structure of bulk water. If the structure imposed by the ion is greater than that lost by the bulk water, the overall structural organization of the water molecules is increased, and the solute is referred to as a structure maker (Figure 4.12). Ionic solutes that generate strong electric fields are structure makers, and the magnitude of their effect increases as the size of the ions decreases and/or their valance increases. If the structure Ion Structure breaker Ion-ordered region (a) Intermediate region Water-ordered region Structure maker (b) Na+ (c) FIGURE 4.12 Schematic representation of organization of water molecules around ionic solutes that act as either structure breakers or structure makers. The water molecules surrounding an ionic solute can be conceptually divided into three regions (a) water molecules in the immediate vicinity of the solute that are highly organized, (b) water molecules in the intermediate region between the solute-organized region and the bulk water region, and (c) water molecules having the normal tetrahedral organization of bulk water. 120 Food Emulsions: Principles, Practices, and Techniques imposed by an ion is not sufficiently large to compensate for that lost by disruption of the tetrahedral structure of bulk water, then the overall structural organization of the water molecules in the solution is decreased, and the solute is referred to as a structure breaker (Figure 4.12). Ionic solutes that generate weak electric fields are structure breakers, and the magnitude of their effect increases as their size increases or their valance decreases. The influence of ionic solutes on the overall properties of water depends on their concentration. At low solute concentrations, the majority of water is not influenced by the presence of the ions and therefore has properties similar to that of bulk water. At intermediate solute concentrations, some of the water molecules have properties similar to those of bulk water, whereas the rest have properties that are dominated by the presence of the ions. At high solute concentrations, all the water molecules are influenced by the presence of the solute molecules and therefore have properties that are appreciably different from those of bulk water. At relatively high salt concentrations, the solubility of biopolymer molecules in aqueous solutions generally decreases when the concentration of ionic solutes increases above a certain level, which is known as “salting-out,” because the solutes compete with the biopolymers for the limited amount of water that is available to hydrate them (Li et al. 2014). Ionic solutes may also influence the molecular conformation and association of biopolymers, and therefore their functional properties, by screening electrostatic interactions, by binding to oppositely charged groups, or by acting as salt bridges. Consequently, at relatively low salt concentrations, biopolymer solubility may either increase or decrease with increasing ionic strength depending on the precise nature of the interactions involved. It is also useful to outline the various ways that ionic solutes can influence droplet–droplet interactions in oil-in-water emulsions since this has a major impact on the stability and properties of these emulsions, especially those stabilized by ionic emulsifiers: • At relatively low concentrations (<10 mM), multivalent ions may bind to the surface of oppositely charged emulsion droplets, thereby decreasing the magnitude of their electrical charge (ζ-potential) and reducing the strength of the electrostatic repulsion between them (Section 3.4). In addition, these ions may form salt bridges between charged oil droplets in water, thereby promoting bridging flocculation (Section 3.4). • At intermediate concentrations (<250 mM), ionic solutes screen electrostatic interactions (Section 3.4) and screen the zero-frequency contribution to the van der Waals interaction (Section 3.3), thus altering the magnitude and range of the repulsive and attractive interactions between the droplets. • At relatively high concentrations (>500 mM), ionic (and other) solutes increase the attraction between emulsion droplets (and other types of colloidal particles) because of a steric exclusion effect. Hydrated ions are significantly bigger than water molecules. Hence, there is a region surrounding the surface of an emulsion droplet where the water molecules can enter, but the hydrated ions are excluded (McClements 2002). This generates a solute concentration gradient between the solute-depleted region and the bulk aqueous solution, which is thermodynamically unfavorable and generates an attractive force between the emulsion droplets. This mechanism is similar to depletion flocculation, but has a much shorter range. • At relatively high concentrations (>500 mM), ionic solutes alter the structural organization of water (Israelachvili 2011), which influences the strength of hydrophobic interactions (Section 3.7). Structure-breakers increase the hydrophobic attraction, whereas structure makers decrease the hydrophobic attraction. • Ionic solutes may cause changes in the conformation of biopolymer molecules adsorbed to the surface of emulsion droplets or dispersed in the continuous phase, which will alter the strength of the steric and depletion interactions between droplets (Sections 3.5 and 3.6). • The binding of hydrated ions to the surface of emulsion droplets may increase the hydration repulsion between the droplets (Section 3.8). The fact that ions influence the interactions between emulsion droplets in so many different ways means that it is often difficult to accurately predict or quantify their effect on emulsion properties. 121 Emulsion Ingredients 4.3.3.2 Interaction of Water with Polar Solutes Many food constituents are noncharged molecules that are entirely polar or contain polar regions, including alcohols, sugars, polyols, proteins, polysaccharides, and surfactants. Water is capable of participating in dipole–dipole interactions with the polar groups on these solutes (Fennema 2008, Israelachvili 2011, Brady 2013). By far the most important type of dipole–dipole interaction is between water and those solutes that have hydrogen bond donors (e.g., –O–Hδ+) or acceptors (e.g., δ− O–). The strength of hydrogen bonds between water molecules and this type of polar solute is similar to that between two water molecules (Table 4.3). The addition of a polar solute to water therefore has much less influence on the mobility and organization of the water molecules in its immediate vicinity than does a similarly sized ionic solute. The influence of polar solutes on the properties of water is largely governed by the ease at which they can be accommodated into the existing tetrahedral structure of the water molecules (Figure 4.13). When a polar solute is of an appropriate size and shape, and has hydrogen bond acceptors and donors at positions where they can easily form bonds with the neighboring water molecules, it can fit into the tetrahedral structure. For this type of solute, there need be little change in the number of hydrogen bonds formed per water molecule or the overall structural organization of the water molecules. This type of solute therefore tends to be highly water soluble because of the entropy of mixing (Chapter 2). If the solute molecule is not of an appropriate size and shape, or if its hydrogen bond donors and acceptors are incapable of aligning with those of neighboring water molecules, then it cannot easily fit into the tetrahedral structure of water. This causes a dislocation of the normal water structure surrounding the solute molecules, which is thermodynamically unfavorable. In addition, there may be a significant alteration in the physicochemical properties of the water molecules in the vicinity of the solute. For this reason, polar solutes that are less compatible with the tetrahedral structure of water tend to be less soluble than those that are compatible. Just as with ionic solutes, the effect of polar solutes depends on their concentration. At low solute concentrations, most of the water has the same properties as bulk water, but at high concentrations a significant proportion of the water has properties that are altered by the presence of the solute. Nevertheless, it takes a greater concentration of a polar solute to cause the same effect as an ionic solute because of the Sugar molecules vary in their shape, dimensions and bond orientations Water Sugar Cavity in water tetrahedral structure Correct shape and charge distribution δ+ δ+ δ– δ– δ– δ– δ+ δ+ High solubility Correct shape and wrong Wrong shape and correct charge distribution charge distribution δ+ δ– δ– δ– δ+ δ– δ+ δ+ Low solubility Low solubility FIGURE 4.13 Schematic representation of the ability of a polar solute (such as a sugar molecule) to fit into the tetrahedral structure of water. 122 Food Emulsions: Principles, Practices, and Techniques greater strength of ion–water interactions compared to dipole–water interactions. At high solute concentrations, there may also be a steric exclusion effect as mentioned in the previous section. Interactions between polar groups and water determine a number of important properties of food components in emulsions. The hydration of the polar head groups of surfactant molecules is believed to be partly responsible for their stability to aggregation. When surfactants are heated, the head groups become progressively dehydrated, which eventually causes the molecules to aggregate (Section 4.5). These hydration forces also play an important role in preventing the aggregation of emulsion droplets stabilized by nonionic surfactants (Section 3.8). The three dimensional conformation and interactions of proteins and polysaccharides is influenced by their ability to form intramolecular and intermolecular hydrogen bonds (Section 4.5). The solubility, partitioning, and volatility of polar solutes depend on their molecular compatibility with the surrounding solvent: the stronger the molecular interactions between a solute and its neighbors in a liquid, the greater its solubility and lower its volatility. 4.3.3.3 Interaction of Water with Nonpolar Solutes: The Hydrophobic Effect The attraction between a water molecule and a nonpolar solute is much weaker than that between two water molecules, because nonpolar molecules are incapable of forming hydrogen bonds (Israelachvili 2011, Norde 2011, Brady 2013). For this reason, when a nonpolar molecule is introduced into pure liquid water, the water molecules surrounding it change their orientation so that they can maximize the number of hydrogen bonds formed with neighboring water molecules (Figure 4.14). The structural rearrangement and alteration in the physicochemical properties of water molecules in the immediate vicinity of a nonpolar solute is known as hydrophobic hydration. At relatively low temperatures, it is believed that a “cage-like” structure of water molecules exists around a nonpolar solute, in which the water molecules involved have a coordination number of four, which is greater than that of the water molecules in the bulk phase (3–3.5). Despite gaining some order, the water molecules in these cage-like structures are still highly dynamic. The alteration in the organization and interactions of water molecules surrounding a nonpolar solute has important implications for the solubility and interactions of nonpolar groups in water. The behavior of nonpolar solutes in water can be understood by considering the transfer of a nonpolar molecule from an environment where it is surrounded by similar molecules to one where it is surrounded by water molecules (Tanford 1980). When a nonpolar solute is transferred from a nonpolar solvent into water, there are changes in both the enthalpy (ΔHtransfer) and entropy (TΔStransfer) of the system. The enthalpy change is related to the alteration in the overall strength of the molecular interactions, whereas the entropy change is related to the alteration in the structural organization of the solute and solvent molecules. The overall free energy change (ΔGtransfer) depends on the relative magnitude of these two contributions: ΔG transfer = ΔHtransfer − TΔStransfer (4.6) The relative contribution of the enthalpy and entropy contributions to the free energy depends on temperature (Figure 4.15). An understanding of the temperature dependence of the free energy of transfer δ– δ– δ+ δ+ Nonpolar solute Water molecules organized in tetrahedral structure FIGURE 4.14 Schematic representation of the reorganization of water molecules near a nonpolar solute. 123 Emulsion Ingredients 40 ΔGtransfer 30 Free energy (kJ/mol) 20 ΔHtransfer 10 –25 0 0 25 50 75 100 125 150 –10 –20 TΔStransfer –30 –40 Temperature (°C) FIGURE 4.15 Temperature dependence of the typical thermodynamics associated with the transfer of a nonpolar solute from a nonpolar liquid into water. is important for food scientists because it governs the behavior of many food components during food processing, storage, and handling. At relatively low temperatures (<25°C), the number of hydrogen bonds formed by the water molecules in the cage-like structure surrounding the nonpolar solute is slightly higher than in bulk water and so ΔHtransfer is negative (i.e., favors transfer). On the other hand, the water molecules in direct contact with the nonpolar solute are more ordered than those in bulk water and so the entropy term is positive (i.e., opposes transfer). Overall, the entropy term dominates and so the transfer of a nonpolar molecule into water is thermodynamically unfavorable. As the temperature is raised the water molecules become more thermally agitated and so their organization within the cage-like structure is progressively lost, which has consequences for both the enthalpy and entropy contributions. First, some of the partial charges on the water molecules face toward the nonpolar group and are therefore unable to form hydrogen bonds with the surrounding water molecules. Thus, the number of hydrogen bonds formed by the water molecules in the cage-like structure decreases with increasing temperature. At a certain temperature, the number of hydrogen bonds formed by the water molecules in the cage-like structure becomes less than that of bulk water. Below this temperature the enthalpy associated with transferring a nonpolar molecule into water is negative (exothermic) and so is favorable to transfer, but above this temperature it is positive (endothermic) and so is unfavorable to transfer. The enthalpy term therefore makes an increasing contribution to opposing the transfer of nonpolar molecules into water as the temperature raises. Second, the increasing disorganization of the water molecules surrounding a nonpolar molecule as the temperature is raised means that the entropy difference between the water molecules in the cage-like structure and those in the bulk water is lessened. Thus, as the temperature is increased, the contribution of the entropy term becomes progressively less important. In summary, at low temperatures the major contribution to the unfavorable free energy change associated with transfer of a nonpolar molecule into water is the entropy term, but at higher temperatures it is the enthalpy term. Overall, the transfer of a nonpolar molecule from an organic solvent into water becomes increasingly thermodynamically unfavorable as the temperature is raised up to a certain temperature (around 100°C–150°C). The free energy associated with transferring a nonpolar molecule from an environment where it is surrounded by similar molecules to one in which it is surrounded by water molecules has been shown to be a product of its surface area and the interfacial tension between the bulk nonpolar liquid and water, that is, ΔG = γΔA (Israelachvili 2011). An aqueous solution containing a nonpolar solute can decrease its free energy by reducing the unfavorable contact area between the nonpolar groups and water, which is known 124 Food Emulsions: Principles, Practices, and Techniques as the hydrophobic effect. The strong tendency for nonpolar molecules to associate with each other in aqueous solutions is a result of the attempt of the system to reduce the contact area between water and nonpolar regions and is known as the hydrophobic interaction (Section 3.7). The hydrophobic effect is responsible for many of the characteristic properties of food emulsions, including the aggregation of proteins, the formation of surfactant micelles, the adsorption of emulsifiers at oil–water interfaces, the aggregation of hydrophobic particles, and the immiscibility of oil and water. The strength of the hydrophobic interaction between nonpolar substances in water is affected by the presence of ions in the aqueous phase separating them. Ions can either increase or decrease the structural organization of water molecules in an aqueous solution, depending on whether they are structure makers or structure breakers (previous section). As one of the major driving forces for hydrophobic interactions is the difference in structural organization (entropy) between the water molecules in the immediate vicinity of the nonpolar solute and those in the bulk water, then changes in the organization of the water molecules in bulk water alters its strength. Structure makers decrease the magnitude of the hydrophobic interaction and therefore increase the water solubility of nonpolar solutes because the difference in structural organization of water molecules in the bulk solution and in the immediate vicinity of a nonpolar solute is reduced, whereas structure breakers have the opposite effect. The strength of the hydrophobic interaction also depends on temperature, increasing as the temperature is raised up to a temperature somewhere above 100°C (Figure 4.15). The temperature dependence of the hydrophobic interaction has important implications for the functionality of many food constituents, since there may be appreciable changes in product temperature during the manufacture, storage, and utilization of food emulsions. 4.3.4 Influence of Solutes on the Physicochemical Properties of Solutions Emulsion formation, stability, and properties are strongly influenced by the bulk physicochemical properties of the aqueous phase, for example, density, viscosity, refractive index, specific heat capacity, thermal conductivity, and dielectric constant. For example, the ability to produce small emulsion droplets during homogenization depends on the viscosity and interfacial tension of the aqueous phase (Chapter 6). The creaming stability of oil-in-water emulsions depends on the density and viscosity of the aqueous phase (Chapter 7). The flocculation and coalescence stability of emulsions depends on the strength of the attractive and repulsive interactions between the droplets, which is influenced by the ionic strength, dielectric constant, and refractive index of the aqueous phase (Chapter 3). The appearance of an emulsion is influenced by the contrast in refractive index between the oil and aqueous phases, since this determines the fraction of light scattered by the droplets (Chapter 9). The flavor of emulsions depends on the partitioning and release rate of flavor molecules in a food, which are influenced by the polarity and viscosity of the aqueous phase (Chapter 10). Knowledge of the bulk physicochemical properties of the aqueous phase is therefore important for predicting, understanding, and controlling the behavior of food emulsions. The bulk physicochemical properties of an aqueous phase depend on the type, concentration, and interactions of the various solutes present. Solutes may influence the bulk properties of water directly through their molecular interactions with the water molecules, or indirectly due to thermodynamic “colligative” effects (Atkins and dePaula 2014). For example, the presence of solutes in an aqueous solution causes a decrease in melting point, an increase in boiling point, and a decrease in water activity of aqueous solutions. Each solute type influences the bulk physicochemical properties of water in a different manner because of differences in its intrinsic molecular characteristics (e.g., molecular weight, size, shape, and charge) and in its ability to alter the properties of the water molecules in its vicinity. The physicochemical properties of aqueous solutions containing different concentrations of many common solutes have been tabulated in the literature (e.g., CRC Handbook of Chemistry and Physics). 4.3.5 Selection of an Appropriate Aqueous Phase Water is usually considered to be a chemically distinct substance with the chemical formula H2O. In practice, the “water” used in the commercial preparation of food emulsions typically contains significant amounts of organic and inorganic contaminants that influence its physicochemical and sensory Emulsion Ingredients 125 properties, for example, acids, bases, minerals, microorganisms, and off-flavors. Many of these contaminants can have an adverse effect on the quality of emulsion-based food products, and, therefore, they are often removed by treating the water prior to utilization, for example, using precipitation, centrifugation, filtration, or evaporation methods. A variety of different solutes are often incorporated into the aqueous phase of food emulsions to control their physicochemical and sensory properties, including acids, bases, salts, buffers, sugars, emulsifiers, and biopolymers. The molecular and functional properties of the most important of these constituents are discussed elsewhere in this chapter. Each constituent plays one or more specific roles in determining the overall properties of a food emulsion, and these roles are often influenced by the presence of the other constituents. It is therefore important for food scientists to identify the role or multiple roles that each functional ingredient plays within a particular food emulsion, and to understand how these roles are influenced by the presence of other constituents. This knowledge will facilitate the rational selection of the optimum combination of ingredients required to produce the desired physicochemical and sensory properties of the final product. 4.4 Emulsifiers The term “emulsifier” is used in this book to describe any surface-active substance that is capable of adsorbing to an oil–water interface and protecting emulsion droplets from aggregation (flocculation and/ or coalescence). The most commonly used emulsifiers in the food industry are small-molecule surfactants, phospholipids, amphiphilic biopolymers, and certain types of particulate matter. These emulsifiers vary widely in their ability to form and stabilize emulsions depending on their molecular and physicochemical characteristics. Ideally, an emulsifier should rapidly adsorb to the oil–water interface during homogenization, reduce the interfacial tension by an appreciable amount, and prevent droplet coalescence from occurring during homogenization (Chapter 6). In addition, it is usually important that the emulsifier forms an interfacial layer that prevents droplet aggregation under the environmental conditions that the product experiences during manufacture, transport, storage, and utilization (Chapter 7). In this section, we review the major types of emulsifiers used in food products, and discuss some of the factors that should be considered in selecting an emulsifier for a particular application. 4.4.1 Surfactants 4.4.1.1 Molecular Characteristics The term “surfactant” is used to refer to those relatively small surface-active molecules that consist of a hydrophilic “head” group that has a high affinity for water, attached to a lipophilic “tail” group that has a high affinity for oil (Stauffer 1999, Walstra 2003, Friberg et al. 2004, Hasenhuettl 2008a,b, Kralova and Sjoblom 2009). The principal role of surfactants in food emulsions is to improve emulsion formation and stability. Nevertheless, they may also alter emulsion properties in a number of other ways, including forming surfactant micelles, interacting with biopolymers, or modifying the formation, growth, and structure of fat crystals. A wide variety of surfactants are available for utilization in food products, and a number of the most commonly used are listed in Table 4.4. These surfactants can be represented by the formula RX, where X represents the hydrophilic head and R the lipophilic tail. The characteristics of a particular surfactant depend on the nature of its head and tail groups. The head group may be anionic, cationic, zwitterionic, or nonionic, although most surfactants used in the food industry are nonionic (e.g., monoglycerides, Tweens, Spans, ACETEM, and LACTEM), anionic (e.g., fatty acid salts, stearoyl lactylate salts, DATEM, and CITREM), or zwitterionic* (e.g., lecithin). Lauric arginate is one of the very few examples of cationic surfactants that are allowed in certain food applications. The tail group of surfactants usually consists of one or more hydrocarbon chains, having between 10 and 20 carbon atoms per chain. Surfactant tails may be saturated or unsaturated, linear or branched, and aliphatic and/ * A zwitterionic surfactant is one which has both positively and negatively charged groups on the same molecule. 126 Food Emulsions: Principles, Practices, and Techniques TABLE 4.4 Classes of Small Molecules Surfactants Commonly Used in Food Emulsion Chemical Name Abbreviation EU Number U.S. FDA ADI (mg kg−1) Solubility E 322 E 470 E 481 E 482 E 472c E 472e 184.1400 172.863 172.846 172.844 172.832 184.1101 NL NL 0–20 0–20 NL 0–50 Oil/water Oil/water Water Oil Water Water E 471 E 472a E 472b — E 475 E 477 E 473 E 491 E 492 E 435 E 436 E 433 184.1505 172.828 172.852 172.830 172.854 172.856 172.859 172.842 — 172.836 172.838 172.840 NL NL NL — 0–25 0–25 0–10 0–25 0–15 0–25 0–25 0–25 Oil Oil Oil Ionic Lecithin Fatty acid salts Sodium stearoyl lactylate Calcium stearoyl lactylate Citric acid esters of MG Diacetyl tartatric acid esters of MG FA SSL CSL CITREM DATEM Nonionic Monoglycerides Acetic Acid esters of MG Lactic acid esters of MG Succinic acid esters of MG Polyglycerol esters of FA Propylene glycol esters of FA Sucrose esters of FA Sorbitan monostearate Sorbitan tristearate Polyoxyethylene (20) sorbitan monostearate Polyoxyethylene (20) sorbitan tristearate Polyoxyethylene (20) sorbitan monooleate MG ACETEM LACTEM SMG PGE PGMS SMS STS Polysorbate 60 Polysorbate 65 Polysorbate 80 Water Oil Oil/watera Water Oil Water Water Water The solubility of some classes of surfactants depends on the relative lengths of their hydrophilic and hydrophobic parts. Note: The table includes their chemical name, abbreviation, acceptable daily intake (ADI), and solubility. NL = not limited. a or aromatic, but most food surfactants have either one or two linear aliphatic chains, which may be saturated or unsaturated. Each type of surfactant has functional properties that are determined by its molecular characteristics and the environment it operates in. It is important to note that commercial surfactants vary considerably in their cost, usage levels, legal status, ingredient compatibility, and ease of utilization. Hence, there is no single surfactant that is suitable for every food application, and it is necessary to select the most appropriate surfactant for each product. Food-grade surfactants are produced industrially by chemical processes utilizing a variety of different raw materials, such as fats, oils, glycerol, organic acids, sugars, and polyols (Hasenhuettl 2008a,b). Despite normally being called by a specific chemical name (Table 4.4), most commercial surfactants are actually highly complex mixtures of a number of different chemical species. This compositional heterogeneity can have a large impact on their functional properties in both laboratory studies of emulsion properties and in their utilization as ingredients in food products. Consequently, it is often important to have analytical methods to determine the type and concentration of different chemical species present in a commercial surfactant. Surfactants are often used in combinations with other types of surfactants, rather than as individual components, since improved functional properties can often be obtained. Surfactant ingredients for utilization in the food industry come in a variety of different forms, including liquids, pastes, solids, powders, and beads. The surfactant is usually suspended in the phase that it is most soluble in prior to homogenization, so that water-soluble surfactants are dispersed in the aqueous phase and oil-soluble surfactants are dispersed in the lipid phase. Even so, there are also examples in the food industry where the surfactant is simply blended directly with the oil and aqueous phases. A variety of processing treatments may be required to ensure that the surfactant is adequately dispersed (e.g., shearing and heating), and these are usually stipulated by the ingredient supplier. 127 Emulsion Ingredients 4.4.1.2 Physicochemical Properties This section provides a brief overview of the most important functional properties of surfactants that are relevant to their application in food emulsions. 4.4.1.2.1 Molecular Organization of Surfactants in Solution At sufficiently low concentrations, surfactants exist as monomers in solution because the entropy of mixing overweighs the attractive forces operating between the surfactant molecules. Nevertheless, as their concentration is increased they can spontaneously aggregate into a variety of thermodynamically stable structures known as association colloids, for example, micelles, bilayers, vesicles, and reverse micelles (Figure 4.16). The primary driving force for the formation of these structures is the hydrophobic effect, which causes the system to adopt a molecular organization that minimizes the unfavorable contact area between the nonpolar tails of the surfactant molecules and water. At still higher concentrations, surfactants may organize themselves into a variety of liquid crystalline structures, such as hexagonal, lamellar, and reversed hexagonal phases. In addition, the surfactant solution may separate into a number of phases, with different compositions and molecular organizations. The molecular organization of surfactants in solutions depends mainly on the geometry and interactions of the surfactant molecules, the nature of the solvent, the solution composition, and the temperature. The influence of solution and environmental conditions on the molecular organization of surfactants can be conveniently described by phase diagrams. Phase diagrams are usually determined empirically for a given surfactant system and allow one to determine the type, number, and composition of the phases formed under a given set of experimental conditions. The surfactant concentrations normally used in food emulsions are insufficient to lead to the formation of liquid crystalline structures, although they are often high enough to lead to the formation of association colloids (Figure 4.16). In the remainder of this section, the focus is on the properties of surfactant micelles, since these are the most common type of association colloid formed in food emulsions. 4.4.1.2.2 Critical Micelle Concentration A surfactant forms micelles in an aqueous solution when the amount added exceeds a certain level known as the critical micelle concentration or CMC (Hiemenz and Rajagopalan 1997, Evans and Wennerstrom 1999, Israelachvili 2011). Below the CMC, the surfactant molecules are dispersed predominantly as monomers, but once the CMC is exceeded, any additional surfactant molecules form micelles, and the monomer concentration remains fairly constant. Micelles have highly dynamic Micelle Nonspherical micelle Vesicle Reverse micelle Bilayer FIGURE 4.16 Some typical structures formed due to the self-association of surfactant molecules at relatively low surfactant concentrations. At higher surfactant concentrations many different kinds of liquid crystalline phases may form. 128 Food Emulsions: Principles, Practices, and Techniques structures because they are only held together by physical interactions that are relatively weak compared to the thermal energy of the system. Despite the highly dynamic nature of their structure, surfactant micelles have a fairly well-defined average size and shape under a given set of environmental conditions. Thus, when surfactant is added to a solution above the CMC the number of micelles tends to increase, rather than the size or shape of the individual micelles (although this may not be true at high surfactant concentrations). There is an abrupt change in the physicochemical properties of a surfactant solution when the CMC is exceeded, for example, surface tension, electrical conductivity, turbidity, and osmotic pressure. This is because the properties of surfactant molecules dispersed as monomers are different from those in micelles. For example, surfactant monomers are amphiphilic and have a high surface activity, whereas micelles have little surface activity because their surface is covered with hydrophilic head groups. Consequently, the surface tension of a solution decreases with increasing surfactant concentration below the CMC, but remains fairly constant above it (Chapter 5). The CMC of a surfactant solution depends on the chemical structure of the surfactant molecules, as well as on the solution composition and prevailing environmental conditions. The CMC tends to decrease as the hydrophobicity of surfactant molecules increases (e.g., by increasing the hydrocarbon tail length) or their hydrophilicity decreases (e.g., by decreasing the length of a nonionic head group or by exchanging an ionic head group for a nonionic one). For ionic surfactants, the CMC decreases appreciably with increasing ionic strength, since counter-ions screen the electrostatic repulsion between charged head groups, reducing the magnitude of this unfavorable contribution to micelle formation. The CMC is not usually strongly temperature dependent over the temperature ranges normally found in foods (e.g., 0°C–100°C). For many commercial food-grade surfactants, the CMC does not occur at a well-defined concentration, but occurs over a range of concentrations, because the surfactant ingredient contains a mixture of components with different chain lengths, degrees of unsaturation and head group size. 4.4.1.2.3 Krafft Point It is usually necessary for a surfactant to be adequately dispersed in a solvent before it can exhibit its desired functional properties. For this reason, it is often necessary to heat surfactants that have a high melting point (usually ionic surfactants) above a critical temperature, known as the Krafft point, before they become soluble enough to function properly (Holmberg et al. 2002). The Krafft point occurs at the temperature where the solubility of the monomers equals the CMC of the surfactant (Figure 4.17). Below the Krafft point the solubility of the surfactant is low, but once the Krafft point is exceeded the surfactant solubility increases dramatically because micelles are much more soluble than monomers. Surfactant concentration Solubility Micelles + Water Crystals + Water CMC T* Monomers + Water Temperature FIGURE 4.17 Simplified phase diagram for a typical surfactant at relatively low surfactant concentrations. At higher surfactant concentrations many different kinds of liquid crystalline phases may form. T* is the Krafft point temperature. 129 Emulsion Ingredients 4.4.1.2.4 Cloud Point When a surfactant solution is heated above a certain temperature, known as the cloud point, it becomes turbid (Holmberg et al. 2002). This occurs because the hydrophilic head groups of the surfactant molecules become progressively dehydrated as the temperature is raised, which alters their molecular geometry and decreases the hydration repulsion between them (Israelachvili 2011). Above a certain temperature, known as the cloud point, the micelles form aggregates that are large enough to scatter light and therefore make the solution appear turbid. As the temperature is increased further the aggregates may grow so large that they sediment under the influence of gravity and form a separate phase that can be observed visually. The cloud point of nonionic surfactants tends to increase as the size of their hydrophilic head group increases and depends on the type and concentration of electrolytes present in aqueous solutions. Knowledge of the cloud point may be an important factor to consider when selecting a surfactant for a particular food emulsion application. The interfacial tension tends to decrease appreciably as the temperature is increased toward the cloud point of a surfactant, which means that droplets are easier to disrupt, but are also more prone to coalescence (Figure 4.18). Consequently, if an emulsion is going to receive some kind of thermal processing treatment, it may be necessary to ensure that the cloud point of the surfactant used to stabilize the system is considerably higher than the maximum temperature experienced by the product. On the other hand, it may be advantageous to homogenize an emulsion at a temperature close to the cloud point of the surfactant since droplet disruption is facilitated and smaller droplet sizes can be achieved using the same input energy. The emulsion can then be rapidly cooled to a temperature well below the cloud point to ensure that the droplets are stable to coalescence. Emulsions and nanoemulsions produced by low-energy homogenization methods are particularly susceptible to destabilization when they are heated close to their cloud points (Figure 4.18) (Chapter 6). 4.4.1.2.5 Solubilization Nonpolar molecules, which are normally insoluble or only sparingly soluble in water, can be solubilized in an aqueous surfactant solution by incorporating them into micelles or other types of association 4.0 Coalescence 3.5 Turbidity increase (cm–1) 3.0 2.5 2.0 1.5 1.0 0.5 0.0 Cloud point 40 50 60 Temperature (°C) 70 80 FIGURE 4.18 Effect of temperature on the turbidity of oil-in-water nanoemulsions containing 10% vitamin E, 10% Tween 80, and 20% ethanol. Rapid coalescence occurs as the cloud point is approached. 130 Food Emulsions: Principles, Practices, and Techniques colloids (Dickinson and McClements 1996, Moulik 1996, Holmberg et al. 2002). The resulting system is thermodynamically stable; however, it may take an appreciable time to reach equilibrium because of the time taken for molecules to diffuse through the system and because of the activation energy associated with transferring a nonpolar molecule from a bulk phase into a micelle. Micelles containing solubilized materials are referred to as swollen micelles or microemulsions, whereas the material solubilized within the micelle is referred to as the solubilizate. The solubilization of lemon oil by nonionic surfactant micelles is shown in Figure 4.19. The ability of micelle solutions to solubilize nonpolar molecules has a number of potentially important applications in the food industry, including selective extraction of nonpolar molecules from oils, controlled ingredient release, incorporation of nonpolar substances into aqueous solutions, transport of nonpolar molecules across aqueous membranes, and modification of chemical reactions. There are three important factors that determine the functional properties of swollen micelle solutions (1) the location of the solubilizate within the micelles, (2) the maximum amount of material that can be solubilized per unit mass of surfactant, and (3) the solubilization rate. The concentration of surfactants in oil-in-water emulsions is often high enough to form micelles in the aqueous phase. These micelles may be capable of solubilizing various kinds of nonpolar and amphiphilic molecules, including flavors, antioxidants, prooxidants, and preservatives, thereby altering their location and functional characteristics. Food manufacturers may therefore have to take into account the possibility that micelle solubilization may influence the bulk physicochemical and sensory properties of food emulsions. 4.4.1.2.6 Surface Activity and Droplet Stabilization Surfactant molecules adsorb to oil–water interfaces because they can adopt an orientation in which the hydrophilic part of the molecule is located in the water, whereas the hydrophobic part is located in the oil. This minimizes the thermodynamically unfavorable contact between hydrophilic and hydrophobic regions, and therefore reduces the interfacial tension (Chapter 5). This reduction in interfacial tension is important during homogenization because it facilitates the further disruption of emulsion droplets, that is, less energy is needed to break-up a droplet when the interfacial tension is lowered (Chapter 6). Once adsorbed to the surface of a droplet, the surfactant must provide a repulsive force that is strong enough 1.2 1 Turbidity (cm–1) 0.8 0.6 0.4 0.2 0 0 1 2 Lemon oil concentration (wt%) 3 FIGURE 4.19 Dependence of turbidity on lemon oil concentration when lemon oil-in-water nanoemulsions are added to a surfactant solution (1% sucrose monopalmitate). Initially, the lemon oil droplets dissolve until the surfactant micelles become saturated, and then the turbidity increases. (Courtesy of JiaJia Rao, University of Massachusetts, Amherst, MA.) Emulsion Ingredients 131 to prevent the droplet from aggregating with its neighbors (Chapters 3 and 7). Ionic surfactants primarily provide stability by causing all of the emulsion droplets to have the same electric charge and therefore electrostatically repel each other. Nonionic surfactants primarily provide stability by generating short range repulsive forces that prevent the droplets from coming too close together, such as steric, hydration, and thermal fluctuation interactions (Chapter 3). It should be noted that even oil droplets stabilized by nonionic surfactants often have an electrical charge (due to ionic impurities) and therefore electrostatic repulsion may also contribute to their stability. Some surfactants form multilayers (rather than monolayers) at the surface of an emulsion droplet, which has been reported to greatly enhance the stability of the droplets against aggregation (Friberg et al. 2004). In summary, surfactants must have three characteristics to be effective at enhancing the formation and stability of emulsions (Chapter 6). First, they must rapidly adsorb to the surface of the freshly formed emulsion droplets during homogenization. Second, they must reduce the interfacial tension by a significant amount. Third, they must form an interfacial layer that prevents the droplets from aggregating under the solution and environmental conditions pertaining to the emulsion. The interfacial layers formed by some surfactants (especially those containing saturated hydrocarbon chains) are capable of undergoing liquid–solid phase transitions upon changes in temperature (Walstra 2003). Above a critical temperature (Tc), the hydrocarbon chains have a relatively high molecular mobility and can be considered to be “fluid-like,” but below Tc the chains lose their molecular mobility, pack closely together and can be considered to be more “solid-like.” The transition of the chain packing from fluid-like to solid-like usually causes an appreciable decrease in the interfacial tension, and may have important consequences for the functional properties of some emulsions. It should be noted that the ability of surfactants to form micelles in the continuous phase of an emulsion can have a negative impact on emulsion stability, because they can induce depletion flocculation or facilitate the transport of oil molecules between droplets when present at a sufficiently high concentration (Dickinson and McClements 1996). The ability of surfactants to regulate the depletion interactions between droplets can also have a pronounced influence on the rheological properties of emulsions (Chapter 8). 4.4.1.2.7 Interaction with Biopolymers Under certain circumstances, surfactant molecules bind to proteins and polysaccharides and the resulting surfactant–biopolymer complexes may have very different functional characteristics than either of the individual components. These surfactant–biopolymer interactions can occur through a variety of different mechanisms, with the two most important usually being electrostatic and hydrophobic interactions. The number of surfactant molecules bound to a biopolymer, and whether the surfactant molecules bind as individual monomers or as micelle-like clusters, depends on the origin and nature of the interaction. The binding of surfactants to biopolymers can lead to large changes in the conformation, stability, and interactions of biopolymer molecules. These changes can have a large influence on the bulk physicochemical properties of biopolymer solutions, such as appearance, rheology, and phase behavior. In addition, these interactions can lead to the formation of structures that may have novel functional properties, for example, for encapsulation and release. When surfactant molecules are mixed with a solution of polymer molecules, they may exist in either a free or a bound state (Figure 4.20). In either of these states, the surfactant may exist as individual monomers or molecular clusters (e.g., micelles). The partitioning of surfactant molecules between these different molecular forms depends on the concentration and molecular characteristics of the biopolymer and surfactant (e.g., molecular weight, hydrophobicity, electrical charge, and flexibility), as well as the prevailing solution and environmental conditions (e.g., temperature, pressure, pH, ionic strength, and external forces). A variety of different physicochemical mechanisms may either favor or oppose binding, for example, hydrophobic interactions, electrostatic interactions, hydrogen bonding, and entropy effects. In a system of a fixed biopolymer concentration and increasing amounts of surfactant, it is possible to define two critical surfactant concentrations: “C1” and “C2.” C1 is usually referred to as the critical aggregation concentration (CAC) and represents the onset concentration at which the interaction between the surfactant and the biopolymer first occurs. Above this concentration, the surfactant molecules may either bind as monomers or as micelle-like clusters. C2 is the surfactant 132 Food Emulsions: Principles, Practices, and Techniques Increasing surfactant concentration C1 C2 CMC* Binding begins Saturation occurs Micelle formation begins FIGURE 4.20 Schematic representation of surfactant binding to biopolymers (assuming that the surfactants do not form micelle-like structures upon binding, which is often the case): C1 is the surfactant concentration where binding begins; C2 is the surfactant concentration where the biopolymer becomes saturated with surfactant; and, CMC* is the effective CMC of the surfactant in the presence of the biopolymer. concentration at which the polymer becomes saturated with surfactant. Above this concentration, additional surfactant goes into the aqueous phase and forms monomers or micelles depending on whether or not the free surfactant concentration is below or above the CMC, respectively. C1 is generally well below the CMC of the surfactant and is only weakly dependent on the amount of polymer in solution. On the other hand, C2, which represents the surfactant concentration at saturation of the polymer, is usually proportional to the polymer concentration. Interactions between surfactants and polysaccharides are utilized in many types of food process to improve food properties. For example, surfactants (e.g., monoglycerides and stearoyl lactylates) are often incorporated into starch-based products, such as breakfast cereals, pasta, and potato products, to improve their quality (Stauffer 1999). The surfactants form inclusion complexes with starch by inserting their hydrocarbon tails into helical coils formed by amylose or linear regions of amylopectin. These lipid–starch complexes are believed to improve the quality of starch-based products such as bread by increasing loaf volume, reducing crumb firmness, and delaying staling, mainly through their ability to retard the retrogradation of starch. The ability of surfactants to bind to starch depends on the molecular characteristics of the starch (e.g., chain length), as well as of the surfactants (e.g., head group polarity, tail group length, and degree of unsaturation). Starch tends to bind more ionic than nonionic surfactant and binds more saturated than unsaturated surfactants. Surfactants may also interact with a wide variety of other types of polysaccharides (e.g., cellulose, pectin, carrageenan chitosan, etc.), thereby altering their conformation, association, and/or stability, which in turn leads to alterations in their functional properties, such as rheology, appearance, stability, and phase separation. Judicious utilization of these interactions can be used to create food emulsions with novel properties or to develop encapsulation or delivery systems. Interactions between surfactants and proteins are also commonly used to improve processing operations or product properties (Friberg et al. 2004). These interactions may be either direct or indirect. Direct interactions involve binding of surfactants to proteins and can cause substantial changes in the conformation, stability, or interactions of protein molecules. Depending on the nature of the interaction, these changes may have either a beneficial or detrimental influence on the functional properties of 133 Emulsion Ingredients proteins, for example, surface activity, foaming capacity, gelation, and solubility. Surfactants may also interact indirectly with proteins by either competing with them or displacing them from interfaces. For example, small molecule surfactants are added to some emulsified food products to displace proteins from the surface of oil droplets, thereby facilitating the coalescence of the droplets during subsequent chilling and shearing operations, for example, ice cream and whipped cream. Modification of fat crystallization: Certain types of surfactants have been shown to be capable of modifying the nucleation and crystallization of lipids, which is utilized to control crystal formation in some food products (Friberg et al. 2004, Smith et al. 2011). Surfactants have been shown to be capable of preventing clouding in salad oils by retarding the growth of fat crystals. The surfactants are believed to adsorb to the surface of any nuclei or small fat crystals formed in the oil, thereby inhibiting their further growth by preventing adsorption of additional lipid molecules. Surfactants have also been shown to inhibit undesirable polymorphic transitions of lipid crystals in chocolates, shortenings, and margarines. 4.4.1.3 Surfactant Classification Schemes A number of classification schemes have been proposed to facilitate the rational selection of surfactants for particular applications (Dickinson and McClements 1996, Salager et al. 2005, Israelachvili 2011, McClements 2014). Classification schemes have been developed that are based on a surfactants solubility in oil and/or water (Bancrofts rule), its ratio of hydrophilic to lipophilic groups (HLB number), its relative affinity for oil and water phases (HLD number), and its molecular geometry (Packing parameter). Ultimately, all of these properties depend on the chemical structure of the surfactant, and so the different classification schemes are often closely related to each other. 4.4.1.3.1 Bancroft’s Rule One of the first empirical rules developed to describe the type of emulsion that could be stabilized by a given surfactant was proposed by Bancroft (Davis 1994). Bancroft’s rule states that the phase in which the surfactant is most soluble will form the continuous phase of an emulsion (Figure 4.21). Hence, a water-soluble surfactant should stabilize oil-in-water emulsions, whereas an oil-soluble surfactant should stabilize water-in-oil emulsions. It should be stressed that the solubility used should be the total surfactant concentration (monomers + micelles) in a phase, not just the monomers. This rule works well for a wide range of surfactants, although there are numerous exceptions. For example, some amphiphilic molecules are highly soluble in either one phase or the other, but they do not form stable emulsions because they are not particularly surface active or they do not protect droplets against aggregation. W/O O/W Oil Water Water-soluble surfactant Intermediate surfactant Oil-soluble surfactant FIGURE 4.21 Bancroft’s rule states that a surfactant that is more soluble in water will form an O/W emulsion, whereas one that is more soluble in oil will form a W/O emulsion. (Adapted from McClements, D.J., Nanoparticle- and Microparticle-Based Delivery Systems: Encapsulation, Protection and Release of Active Components, CRC Press, Boca Raton, FL, 2014.) 134 Food Emulsions: Principles, Practices, and Techniques In summary, Bancroft’s rule is a useful empirical method of determining the type of emulsion a surfactant will potentially stabilize (O/W or W/O); however, it provides little insight into the relationship between the molecular structure of a surfactant and the long-term stability of the emulsions formed. 4.4.1.3.2 Hydrophile–Lipophile Balance (HLB) The HLB concept is a semiempirical method that is widely used for classifying surfactants (Pasquali et al. 2008). The hydrophile–lipophile balance is described by a number that depends on the molecular properties of the surfactant and gives an indication of its relative affinity for the oil and aqueous phases. Each surfactant is therefore assigned an HLB number according to its chemical structure, that is, head group and tail group structure. A molecule with a high HLB number has a high ratio of hydrophilic groups to lipophilic groups, and vice versa. The HLB number of a surfactant can be calculated from knowledge of the number and type of hydrophilic and lipophilic groups it contains, or it can be estimated from experimental measurements of its cloud point (Shinoda and Friberg 1986). The HLB numbers of many surfactants have been tabulated in the literature (Becher 1983, 1985). A widely used semiempirical method of calculating the HLB number of a surfactant is as follows (Davis 1994): HLB = 7 + ∑(hydrophilic group numbers) − ∑(lipophilic group numbers) (4.7) Group numbers have been assigned to many different types of hydrophilic and lipophilic groups (Table 4.5). The sums of the group numbers of all the lipophilic groups and of all the hydrophilic groups are substituted into the above equation and the HLB number is calculated. The HLB numbers of many foodgrade surfactants have been calculated or determined experimentally (Table 4.6). Despite originally being developed as a semiempirical equation, Equation 4.7 has been shown to have a thermodynamic basis, with the sums corresponding to the free energy changes in the hydrophilic and lipophilic parts of the molecule when micelles are formed (Becher 1985). The HLB number of a surfactant gives a useful indication of its solubility in either the oil and/or water phases and can be used to predict the type of emulsion that will be formed by a surfactant (Table 4.7; Figure 4.21). A surfactant with a low HLB number (3–6) is predominantly hydrophobic, dissolves preferentially in oil, stabilizes water-in-oil emulsions, and forms reverse micelles in oil. A surfactant with a high HLB number (10–18) is predominantly hydrophilic, dissolves preferentially in water, stabilizes oilin-water emulsions, and forms micelles in water. A surfactant with an intermediate HLB number (7–9) has no particular preference for either oil or water and is considered a good “wetting agent.” Molecules with HLB numbers below 3 (very hydrophobic) and above 18 (very hydrophilic) are often not particularly surface active since they tend to accumulate preferentially in bulk oil or bulk water, rather than at an oil–water interface. Emulsion droplets are particularly prone to coalescence when they are stabilized by surfactants that have extreme or intermediate HLB numbers. At very high or low HLB numbers, a surfactant may have such a low surface activity that it does not accumulate appreciably at the droplet TABLE 4.5 Selected HLB Group Numbers Hydrophilic Group –SO4 Na −COO−H+ Tertiary amine Sorbitan ester Glycerol ester –COOH –OH –O– –(CH2–CH2–O)– − + Group Number Lipophilic Group Group Number 38.7 21.2 9.4 6.8 5.25 2.1 1.9 1.3 0.33 –CH– –CH2− –CH3 –CH= 0.475 0.475 0.475 0.475 Source: Adapted from various sources. 135 Emulsion Ingredients TABLE 4.6 Summary of the Properties of Small Molecule Surfactants That Can Be Used to Formulate Food Emulsions Chemical Name Ionic Lecithin Lysolecithin Fatty acid salts Sodium stearoyl lactylate Calcium stearoyl lactylate Citric acid esters of MG Diacetyl tartaric acid esters of MG Lauroyl arginate Nonionic Mono- and diglycerides Acetyl esters of MG Lactyl esters of MG Succinic acid esters of MG Polyglycerol esters of FA Propylene glycol esters of FA Sucrose monooleate Sucrose monostearate Sucrose monopalmitate Sucrose monolaurate Sucrose distearate Sorbitan monooleate Sorbitan monostearate Sorbitan monopalmitate Sorbitan monolaurate Sorbitan tristearate POE sorbitan monooleate POE sorbitan monostearate POE sorbitan monopalmitate POE sorbitan monolaurate POE sorbitan tristearate Polyglycerol polyricinoleate Abbreviation Solubility — — FA SSL CSL CITREM DATEM LAE O&W W O&W W O W W W MDG ACETEM LACTEM SMG PGE PGMS SMO SMS SMP SML SDS Span 80 Span 60 Span 40 Span 20 Span 65 Tween 80 Tween 60 Tween 40 Tween 20 Tween 65 PGPR O O O O O&W O W W W W O O O O O O W W W W W O HLB Number 2–8 8–11 1–3 11 7–9 9.2 Parameter Charge Zwitterionic Zwitterionic Negative Negative Negative Negative Negative Positive 2–5 2.5–3.5 3–4 5.3 2–13 1–3 15 16 15 6 4.3 4.7 6 8.6 2.2 15 14.9 15.6 16.7 10.5 1.5 Cloud point — — — — — 65 80 73 76 — Source: Data from McClements, D.J., Nanoparticle- and Microparticle-Based Delivery Systems: Encapsulation, Protection and Release of Active Components, CRC Press, Boca Raton, FL, 2014. Note: MG, monoglycerides; POE, polyoxyethylene. a The solubility of some classes of surfactants depends on the relative lengths of their hydrophilic and hydrophobic parts. surface and therefore does not provide protection against coalescence. At intermediate HLB numbers (7–9), emulsions are unstable to coalescence because the interfacial tension is so low that very little free energy is required to disrupt oil–water interfaces. Empirical observations suggest that maximum emulsion stability is obtained for oil-in-water emulsions using surfactants with an HLB number around 10–12, and for water-in-oil emulsions around 3–5. This is because the surfactants are surface active, but do not lower the interfacial tension so much that the droplets are easily disrupted. Under certain circumstances, it is possible to adjust the “effective” HLB number by using a combination of two or more surfactants with different HLB numbers (Stauffer 1999). Surfactant blends are often used in the food industry to improve the overall functionality of surfactant systems in commercial products. In addition, most commercial surfactants actually contain a blend of different types of surfactant molecule, which may vary from batch to batch. 136 Food Emulsions: Principles, Practices, and Techniques TABLE 4.7 Comparison of Functional Attributes of Different General Classes of Emulsifiers Chemical Name Solubility Emulsion Type Usage Level (g/goil) Oil Water Water W/O O/W O/W ∼0.05 ∼0.05 ∼0.05 Water O/W Water O/W pH Stability Salt Stability Temperature Stability Good Good Good Good Good Poor at I > CFC — Poor at T ∼ PIT Poor at T ∼ PIT ∼0.05 Poor at IEP Poor at I > CFC Poor at T > Tm ∼1–1.5 Good Good Good Surfactants Nonionic (low HLB) Nonionic (high HLB) Ionic Proteins Polysaccharides Note: It should be stressed that the behavior of a specific emulsifier may be different from these general characteristics, and the reader is referred to the text for additional information about the behavior of the different emulsifiers. The symbols in the table are PIT, phase inversion temperature; Tm, thermal denaturation temperature; IEP, isoelectric point; I, ionic strength; and CFC, critical flocculation concentration. One of the major drawbacks of the HLB concept is that it does not take into account the fact that the functional properties of a surfactant molecule are altered significantly by changes in temperature or solution conditions. Thus, a surfactant may be capable of stabilizing oil-in-water emulsions at one temperature, but water-in-oil emulsions at another temperature, even though it has exactly the same chemical structure. The HLB concept could be extended to include temperature effects by determining the group numbers as a function of temperature, although this would be a rather tedious and time-consuming task. Another limitation is that the optimum HLB number required for a surfactant to create a stable emulsion depends on oil type. Hence, the optimum “required” HLB number has to be empirically established for different kinds of oil (Pasquali et al. 2008). 4.4.1.3.3 Molecular Geometry and the Phase Inversion Temperature (PIT) The molecular geometry of a surfactant molecule can be described by a packing parameter, p (Evans and Wennerstrom 1999, Israelachvili 2011): p= v la0 (4.8) where v and l are the volume and length of the hydrophobic tail a 0 is the cross-sectional area of the hydrophilic head group (Figure 4.22) When surfactant molecules associate with each other, they tend to form monolayers that have a curvature that allows the most efficient packing of the molecules. At this optimum curvature, the monolayer has its lowest free energy, and any deviation from this curvature requires the expenditure of free energy. The optimum curvature (H0) of a monolayer depends on the packing parameter of the surfactant: for p = 1, monolayers with zero curvature (H0 = 0) are preferred; for p < 1, the optimum curvature is convex (H0 < 0); and, for p > 1 the optimum curvature is concave (H0 > 0) (Figure 4.22). Simple geometrical considerations indicate that spherical micelles are formed when p is less than 1/3, nonspherical micelles when p is between 1/3 and 1/2, and bilayers when p is between 1/2 and 1. Above a certain concentration, bilayers join-up to form vesicles because energetically unfavorable end-effects are eliminated. At values of p greater than 1 reverse micelles are formed, in which the hydrophilic head groups are located in the interior (away from the oil), and the hydrophobic tail groups are located at the exterior (in contact with the oil) (Figure 4.22). The packing parameter therefore gives a useful indication of the type of association colloid that a surfactant molecule forms in solution. 137 Emulsion Ingredients Packing parameter: p = aT/aH Optimum curvature p<1 p=1 p>1 FIGURE 4.22 The physicochemical properties of surfactants can be related to their molecular geometry. The packing parameter is also useful because it accounts for the temperature dependence of the physicochemical properties of surfactant solutions and of emulsions (Kabalnov and Wennerstrom 1996, Kabalnov 1998). The temperature at which a surfactant solution converts from a micelle to a reversemicelle system or that an oil-in-water emulsion changes to a water-in-oil emulsion is known as the phase inversion temperature or PIT (Shinoda and Friberg 1986). Consider what happens when an emulsion that is stabilized by a surfactant is heated (Figure 4.23). At temperatures well below the PIT (≈20°C), the packing parameter is significantly less than unity, and so a system that consists of an oil-in-water emulsion in equilibrium with a swollen micelle solution is favored. As the temperature is raised, the hydrophilic head groups of the surfactant molecules become progressively dehydrated, which causes p to increase toward unity. Thus, the emulsion droplets become more prone to coalescence and the swollen micelles grow in size. At the phase inversion temperature, p = 1, and the emulsion breaks down because the droplets have an ultralow interfacial tension and therefore readily coalesce with each other. The resulting system consists of excess oil and excess water (containing some surfactant), separated by a third phase that contains surfactant molecules organized into bilayer structures. At temperatures sufficiently greater than the PIT (≈20°C), the packing parameter is much larger than unity, and the formation of a system that consists of a water-in-oil emulsion in equilibrium with swollen reverse-micelles is favored. A further increase in temperature leads to a decrease in the size of the reverse micelles and in the amount of water solubilized within them. The method of categorizing surfactant molecules according to their molecular geometry is now widely used for determining the type of emulsions they are best at stabilizing. 4.4.1.3.4 Hydrophile–Lipophile Deviation (HLD) Another approach has recently been developed to rationalize surfactant performance specifically takes into account the environment in which they are actually used (Salager et al. 2004, Queste et al. 2007). The behavior of a surfactant–oil–water (SOW) system is described by a formulation–composition map (Figure 4.24). The composition-axis represents changes in the water-to-oil ratio (WOR), whereas the formulation-axis represents changes in the relative affinity of the surfactant for the oil and water phases, which is expressed as the hydrophilic–lipophilic deviation (HLD). Formulation–composition maps 138 Food Emulsions: Principles, Practices, and Techniques P.I.T. Surface tension Coalescence instability p<1 p=1 O/W Unstable p>1 W/O FIGURE 4.23 The phase inversion temperature occurs when the optimum curvature of a surfactant monolayer is zero. provide a convenient means of specifying the types of emulsions or microemulsions that are stable for a given SOW system when either surfactant properties are altered (e.g., by changing pH, ionic strength, solvent type, or temperature) or when the overall system composition is altered (e.g., by changing the relative amounts of oil and water). The HLD value characterizes the behavior of a surfactant within a particular environment, and is therefore more comprehensive than the HLB number that mainly focuses on the properties of the surfactant itself (Leal-Calderon et al. 2007). The HLD value includes the influence of oil phase properties (such as oil type), aqueous phase properties (such as salt, alcohol, or other cosolvents content), and environmental factors (such as temperature) on the relative affinity of a surfactant for the oil and water phases. The HLD number of certain types of surfactant can be calculated using simple empirical equations that depend on surfactant type, oil type, solvent composition, and temperature (Queste et al. 2007). The relationship between the HLD number of a surfactant and its ability to stabilize emulsions or microemulsions is highlighted below (McClements 2014): • HLD < 0: The surfactant (1) has a higher affinity for water than oil; (2) tends to form normal micelles or microemulsions in water; and (3) tends to stabilize O/W emulsions. The more negative the HLD number, the greater the affinity for the water phase. • HLD = 0: The surfactant (1) has an equal affinity for water and oil; (2) tends to form bicontinuous microemulsion or liquid crystalline phases; and (3) tends to stabilize neither O/W nor W/O emulsions. • HLD > 0: The surfactant (1) has a higher affinity for oil than water; (2) tends to form reverse micelles or microemulsions in oil; and (3) tends to stabilize W/O emulsions. The more positive the HLD number, the greater the affinity for the oil phase. 139 Emulsion Ingredients 0 C+ WOR W/O B– A– C– O/W O/W W/O Unstable W /O /W W/O O/ W /O HLD formulation A+ B+ + O/W HLD Unstable A,B,C + or – – 0 Oil 1 Composition WOR Regime Composition ∞ Water Formulation Favored system C+ WOR > 1 (O/W) HLD > 0 (W/O) Unstable C– WOR > 1 (O/W) HLD < 0 (O/W) O/W A+ WOR ≈ 1 (W/O or O/W) HLD > 0 (W/O) W/O A– WOR ≈ 1 (W/O or O/W) HLD < 0 (O/W) O/W B+ WOR < 1 (W/O) HLD > 0 (W/O) W/O B– WOR < 1 (W/O) HLD < 0 (O/W) Unstable FIGURE 4.24 A formulation–composition map can be used to describe the behavior of a surfactant–oil–water system for different formulation variables (HLD) and compositions (WOR). Knowledge of the HLD number, water-to-oil ratio (WOR), and formulation–composition map for a particular SOW system can be used to rationalize its behavior (Figure 4.24). The formulation–composition map can be conveniently divided into a number of different regimes, which are designated by a letter and a sign (Leal-Calderon et al. 2007). The sign determines the influence of formulation (HLD number) on the type of emulsions that will remain stable under a given set of conditions. In regions where the HLD number is negative (A−, B−, and C−), the surfactant favors the formation of O/W emulsions and microemulsions, whereas in regions where the HLD number is positive (A+, B+, and C+), the surfactant favors the formation of W/O emulsions and microemulsions (Salager et al. 2005, Witthayapanyanon et al. 2008). The letter determines the influence of system composition (WOR) on the type of emulsion formed: A refers to a system where the oil and water phases have fairly similar amounts (WOR ≈ 1) and so formation of both O/W and W/O emulsions is favorable; B refers to a system where the oil phase is in excess and so the formation of W/O emulsions is favored; and C refers to a system where the water phase is in excess and so the formation of O/W emulsions is favored. If both the formulation variable (HLD number and sign) and composition variable (WOR and letter) favor the formation of a particular emulsion type (e.g., O/W), then this emulsion is said to be “normal” and will tend to be stable (Mira et al. 2003, RondónGonzaléz et al. 2006). Conversely, if the formulation variable favors one emulsion type (e.g., W/O) while the composition variable favors the other type (e.g., O/W), then this system is said to be “abnormal” and will tend to be unstable. In the formulation–composition map shown in Figure 4.24 there are two regimes where O/W emulsions should be stable (A− and C−), two regimes where W/O emulsions should be stable (A+ and B+), and two regimes where no emulsions are stable (B− and C+). The emulsions formed in the abnormal regimes are usually highly unstable to droplet coalescence and phase separation, but multiple emulsions may be formed near certain phase boundaries. For example, O/W/O emulsions may be formed near the B− to A− boundary, whereas W/O/W emulsions may be formed near the C+ to A+ boundary (Figure 4.24). The above formulation–composition maps can also be useful for characterizing phase transitions from one emulsion type to another, for example, O/W to W/O or vice versa. 140 Food Emulsions: Principles, Practices, and Techniques 4.4.1.3.5 Other Factors The classification schemes mentioned above provide information about the type of emulsion that a surfactant tends to stabilize (i.e., O/W or W/O), but they do not provide much insight into the size of the droplets that are formed during homogenization, the amount of surfactant required to form a stable emulsion, or the stability of the emulsion droplets once formed. In choosing a surfactant for a particular application, these factors must also be considered. The speed at which a surfactant adsorbs to the surface of the emulsion droplets produced during homogenization determines the minimum droplet size that can be produced: the faster the adsorption rate, the smaller the size (Chapters 5 and 6). The amount of surfactant required to stabilize an emulsion depends on the total surface area of the droplets, the surface area covered per unit mass of surfactant, and the binding affinity for the interface (Chapters 5 and 6). The magnitude and range of the repulsive interactions generated by an interfacial surfactant layer, as well as its viscoelasticity, determine the stability of emulsion droplets to aggregation (Chapters 3 and 7). 4.4.1.4 Common Food-Grade Surfactants The properties of a number of food-grade surfactants commonly used in the food industry are briefly discussed below and summarized in Tables 4.6 and 4.7. Water-soluble surfactants with relatively high HLB numbers (10–18) are normally used to stabilize oil-in-water emulsions, such as beverages, dressings, deserts, and coffee whiteners. Nevertheless, they are also used to displace proteins from the surfaces of protein-stabilized fat droplets during the production of ice creams, whipped creams, and toppings. Watersoluble surfactants may also bind to proteins or polysaccharides and modify their functional properties, for example, surfactant binding to starch molecules can inhibit bread staling. Oil-soluble surfactants with relatively low HLB numbers (3–6) are often used to stabilize water-in-oil emulsions, such as margarines and spreads. They are also used to inhibit fat crystallization in some oil-in-water emulsions, since this improves the stability of the food product to refrigeration conditions, for example, dressings. Oil-soluble and water-soluble surfactants can also be used in combination to extend their range of functional properties and applications. Surfactants with intermediate HLB numbers (6–9) have a poor solubility in both oil and water phases and are not particularly good emulsifiers when used in isolation. Nevertheless, their emulsification properties can be improved by using them in combination with other surfactants. As mentioned earlier, most surfactants do not consist of an individual molecular species, but consist of a complex mixture of different types of molecular species. Some of the impurities in surfactant mixtures may adversely affect the physical or chemical stability of emulsions, for example, peroxides in nonionic surfactants can affect lipid oxidation. Hence, it may be necessary to ensure that a surfactant is of a reliable high purity and quality before it is used to prepare a product. Detailed descriptions of various surfactants have been given elsewhere (Friberg et al. 2004, Whitehurst 2004, McClements 2005, Kralova and Sjoblom 2009, Hasenhuettl and Hartel 2008, McClements 2014), and so only a brief outline is given here. Many ingredient manufacturers provided detailed information about the composition, properties, and functional performances of the surfactants they supply. 4.4.1.4.1 Mono- and Diglycerides The term mono- and diglycerides is commonly used to refer to a series of surfactants produced by interesterification of fats or oils with glycerol (Moonen and Bas 2004). This manufacturing operation produces a complex mixture of monoglycerides, diglycerides, triglycerides, glycerol, and free fatty acids. The monoglyceride fraction can be isolated (>90% purity) from the other fractions by molecular distillation to produce a more pure ingredient, referred to as distilled monoglycerides. Distilled monoglycerides are available with hydrocarbon chains of differing lengths and degrees of unsaturation. Generally, monoglycerides are nonionic oil-soluble surfactants with relatively low HLB numbers (∼2–5) and are therefore most suitable for forming reverse micelles, W/O microemulsions, and W/O emulsions. 4.4.1.4.2 Organic Acid Esters of Mono- and Diglycerides Mono- and diglycerides can be esterified with a variety of different organic acids (e.g., acetic, citric, diacetyl tartaric, and lactic acids) to form small molecule surfactants with different functional properties Emulsion Ingredients 141 (Gaupp and Adams 2004). The most common examples of this type of surfactant are acetylated monoglycerides (ACETEM), lactylated monoglycerides (LACTEM), diacetyl tartaric acid monoglycerides (DATEM), and citric acid esters of monoglycerides (CITREM). Each of these surfactants is available with hydrocarbon chains of differing lengths and degrees of unsaturation. ACETEM and LACTEM are nonionic oil-soluble surfactants with low HLB numbers, whereas DATEM and CITREM are anionic water-dispersible surfactants with intermediate or high HLB numbers. 4.4.1.4.3 Polyol Esters of Fatty Acids Small molecule surfactants can also be produced by esterification of polyols with fatty acids (Cottrell and vanPeij 2004, Nelen and Cooper 2004, Norn 2004, Sparso and Krog 2004). The polyols form the hydrophilic head group, whereas the fatty acid chains form the hydrophobic tail group. The physicochemical properties and functional attributes of a particular surfactant produced using this approach depends on the type of polyols and fatty acids used. The most commonly used polyols are polyglycerol, propylene glycol, sorbitan, polyoxyethylene sorbitan, and sucrose. The fatty acids used to prepare these kinds of surfactant may vary in their chain length and degree of unsaturation. The solubility and functional properties of polyol esters of fatty acids depend on the relative sizes of the hydrophilic and lipophilic parts of the molecules. Surfactants with relatively large polyol head groups tend to be water dispersible and have high HLB numbers (e.g., sucrose, polyglycerol, and polyoxyethylene sorbitan esters), whereas those with small polyol head groups tend to be oil soluble and have low HLB numbers (e.g., propylene glycol esters). The ratio of hydrophilic to lipophilic groups can also be varied by changing the number of fatty acids attached to the polyol group, which leads to both oil-soluble and water-dispersible surfactants in the same class, for example, sucrose, sorbitan, or polyoxyethylene sorbitan esters. Sorbitan esters of fatty acids are one of the most commonly used oil-soluble nonionic surfactants, which are often sold under the trade name “Span™.” On the other hand, polyoxyethylene sorbitan esters are one of the most commonly used water-soluble nonionic surfactants, which are often sold under the trade names of “polysorbate™” or “Tween™.” Oil-soluble and water-soluble surfactants in this class are often used in combination to facilitate the formation and stability of colloidal delivery systems (Myers 2006). 4.4.1.4.4 Stearoyl Lactylate Salts Surfactants can be produced by esterification of lactic acid with fatty acids in the presence of either sodium or calcium hydroxide (Boutte and Skogerson 2004). Sodium stearoyl lactylate (SSL) is an anionic water-dispersible surfactant with an intermediate HLB number, whereas calcium stearoyl lactylate (CSL) is an anionic oil-soluble surfactant with a low HLB number. 4.4.1.4.5 Lecithins Lecithins are naturally occurring surface-active molecules that can be extracted from a variety of sources, including soybeans, milk, rapeseed, and egg (Bueschelberger 2004). In nature, they are present in the cell and organelle walls of plants, animals, and microorganisms where they form a natural barrier with important functions in protection, separation, and transport of components. Lecithins isolated from natural sources contain a complex mixture of different types of phospholipids and other lipids, although they can be fractionated to form more pure ingredients that are enriched with specific fractions. The most common phospholipids in lecithin are phosphatidylcholine (PC), phosphatidylethanolamine (PE), and phosphatidylinositol (PI). The hydrophilic head groups of these molecules are either anionic (PI) or zwitterionic (PC and PE), whereas the lipophilic tail groups consist of two fatty acids. Natural lecithin ingredients tend to have low-to-intermediate HLB numbers (2–8) and are therefore most suitable for stabilizing water-in-oil systems (low HLB), or forming bilayers or liposomes in aqueous solutions (intermediate HLB). However, lecithin can also be used in combination with other types of surfactants to improve stability and to form different structures. In addition, lecithin can be chemically or enzymatically hydrolyzed to break off one of the hydrocarbon tails to produce more hydrophilic surfactants called lysolecithins that are capable of forming micelles, microemulsions, or O/W emulsions. One of the advantages of using lecithins is that they are perceived as being natural by consumers. 142 Food Emulsions: Principles, Practices, and Techniques 4.4.1.4.6 Saponins Saponins offer another potential source of naturally derived small molecule surfactants. A natural extract from the bark of the Quillaja saponaria tree has been found to contain surface active components capable of forming surfactant micelles and of stabilizing oil-in-water emulsions (Waller and Yamasaki 1996a,b, Mitra and Dungan 1997). The major surface-active components within this extract have been identified as saponins (van Setten et al. 1995, 1998), which are high molecular weight glycosides consisting of a sugar moiety attached to a triterpene or steroid aglycone (Hostettmann and Marston 1995). The saponins are surface active because they contain both hydrophilic regions (such as sugar groups) and hydrophobic regions (such as phenolic groups) on the same molecule (Sidhu and Oakenfull 1986, Mitra and Dungan 1997). A food ingredient based on the quillaja saponin extract has recently been marketed under the trade name Q-Naturale® (Ingredion, Bridgewater, NJ) This ingredient has been shown to be capable of forming oil-in-water emulsions containing small droplets that are stable over a wide range of pH values, ionic strengths, and temperatures (Yang et al. 2013, Yang and McClements 2013). 4.4.2 Amphiphilic Biopolymers 4.4.2.1 Molecular Characteristics Proteins and polysaccharides are naturally occurring polymers that can often be used as emulsifiers in foods (Damodaran et al. 2007, Belitz et al. 2009, Brady 2013). Proteins are polymers of amino acids, whereas polysaccharides are polymers of monosaccharides. The functional properties of food biopolymers (e.g., solubility, surface activity, thickening, and gelation) are ultimately determined by their molecular characteristics (e.g., molecular weight, conformation, flexibility, polarity, hydrophobicity, and interactions) (Figure 4.25). These molecular characteristics are determined by the type, number, and sequence of the monomers that make up the polymer chain. Monomers vary according to their polarity (ionic, polar, nonpolar, or amphiphilic), dimensions, interactions, and chemically reactive groups. If a biopolymer contains only one type of monomer it is referred to as a homopolymer (e.g., amylose or cellulose), but if it contains different types of monomer it is referred to as a heteropolymer (e.g., gum arabic, pectin, and all natural proteins). Random coil Globular Rigid rod (a) – High Low (b) – – – – Negative + + + + + Positive (c) + + Unbranched (d) – + Branched Low + + + + + + High (e) FIGURE 4.25 Biopolymers, such as proteins and polysaccharides, may have a variety of different molecular characteristics depending on their biological origin and processing conditions used to isolate, purify, and modify them. (a) Conformation, (b) molecular weight, (c) charge sign, (d) branching, and (e) charge density. Emulsion Ingredients 143 Both proteins and polysaccharides have covalent linkages between the monomers around which the polymer chain can rotate at certain well-defined angles. The fact that biopolymers contain relatively large numbers of monomers (typically between 20 and 20,000) and that rotation around the links in the chain is possible, means that they can potentially take up a huge number of different configurations in solution. In practice, many biopolymers adopt fairly well-defined conformations in an attempt to minimize the free energy of the system under the prevailing solution and environmental conditions. This conformation is determined by a delicate balance of physicochemical phenomena, including hydrophobic interactions, electrostatic interactions, hydrogen bonding, van der Waals forces, and configurational entropy (Chapter 2). It should be stressed that most foods are actually nonequilibrium systems, and so a biopolymer may be trapped in a metastable state, because there is a large activation energy preventing it from reaching the most thermodynamically stable state. The configurations that biopolymer chains tend to adopt in aqueous solutions can be conveniently divided into three general categories: globular, random coil, or rod-like (Figure 4.25). Globular biopolymers have fairly rigid compact structures, rodlike biopolymers have fairly rigid extended structures (often helical), and random-coil biopolymers have highly dynamic and flexible structures. Biopolymers can also be classified according to the degree of branching of the chain. Most proteins have linear chains, whereas polysaccharides can have either linear (e.g., amylose) or branched (e.g., amylopectin) chains. In these systems, the location, length, and composition of the branches play an important role in biopolymer functionality. In practice, many biopolymers do not have exclusively one type of conformation, but have some regions that are random coil, some that are rod-like, and some that are globular. Biopolymers in solution may be present as individual molecules or they may be present as clusters where they are associated with one or more molecules of the same or different kind. Finally, it should be mentioned that biopolymers may undergo transitions from one conformation to another, or from one aggregation state to another, if their environment is altered, for example, pH, ionic strength, solvent composition, or temperature. The conformation and aggregation state of biopolymers play a major role in determining their functional attributes, and so it is usually important that food scientists are aware of the molecular characteristics of the biopolymers present in each particular food emulsion. 4.4.2.2 Interfacial Activity and Emulsion Stabilization Usually, amphiphilic biopolymers must be fully dispersed and dissolved in an aqueous solution before they are capable of exhibiting their desirable emulsifying properties. Solvation of biopolymer ingredients prior to homogenization is therefore an important step in the formation of many food emulsions. This process usually involves a number of stages, including dispersion, wetting, swelling, and dissolution. The rate and extent of dissolution depends on many factors, including the nature of the ingredient (e.g., liquid, powder, or granules), biopolymer type and conformation, solution conditions (pH, ionic strength, and temperature), and the application of shearing forces. Generally, factors that favor biopolymer– biopolymer interactions tend to oppose good dissolution, whereas factors that favor biopolymer–solvent interactions tend to promote good dissolution. These factors are primarily governed by the nature of the molecular interactions that dominate in the particular system, which depends strongly on biopolymer type and solvent composition. Guidelines about the most appropriate conditions required to disperse and dissolve specific food-grade biopolymers are usually given by ingredient suppliers. After a biopolymer ingredient has been adequately dissolved in the aqueous phase, it is important to ensure that the solution and environmental conditions (e.g., pH, ionic strength, temperature, and solvent composition) will not promote droplet aggregation during homogenization or after the emulsion is formed. For example, it is difficult to produce protein-stabilized emulsions at pH values close to the isoelectric point of the proteins or at high salt concentrations because the electrostatic repulsion between the droplets is insufficient to prevent droplet aggregation once the emulsions are formed. The interfacial activity of many biopolymers is due to the fact that they have both hydrophilic and lipophilic regions distributed along their backbones (Dickinson 2003). For example, most proteins have significant numbers of exposed nonpolar amino acid side groups, whereas some polysaccharides have nonpolar side chains attached to their polar backbones. The major driving force for adsorption of these amphiphilic biopolymers to oil–water interfaces is therefore the hydrophobic effect. When the 144 Food Emulsions: Principles, Practices, and Techniques biopolymer is dispersed in an aqueous phase, some of the nonpolar groups are in contact with water, which is thermodynamically unfavorable because of hydrophobic interactions (Section 4.3). When a biopolymer adsorbs to an interface, it can adopt a conformation where the nonpolar groups are located in the oil phase (away from the water) and the hydrophilic groups are located in the aqueous phase (in contact with the water) (Figure 4.26). Adsorption also reduces the contact area between the oil and water molecules at the oil–water interface, which lowers the interfacial tension (Chapter 5). Both of these factors favor the adsorption of amphiphilic biopolymers to oil–water interfaces. The conformation that a biopolymer adopts at an interface, and the physicochemical properties of the membrane formed, depends on its molecular structure and interactions. Flexible random-coil biopolymers adopt an arrangement where the predominantly nonpolar segments protrude into the oil phase, the predominantly polar segments protrude into the aqueous phase, and the neutral regions lie flat against the interface (Figure 4.26). The interfaces formed by these types of molecules tend to be relatively open, thick, and of low viscoelasticity. Globular biopolymers (usually proteins) adsorb to an interface so that the predominantly nonpolar regions on the surface of the molecule face the oil phase, whereas the predominantly polar regions face the aqueous phase, and so they tend to have a particular orientation at an interface (Figure 4.26). Once they have adsorbed to an interface, biopolymers often undergo structural rearrangements so that they can maximize the number of contacts between nonpolar groups and oil (Norde 2011). Random-coil biopolymers are relatively flexible molecules and can therefore rearrange their structures fairly rapidly, whereas globular biopolymers are more rigid molecules and therefore rearrange more slowly. The unfolding of a globular protein at an interface often exposes amino acids that were originally located in the hydrophobic interior of the molecule, which can lead to enhanced interactions with neighboring protein molecules through hydrophobic attraction or disulfide bond formation (Tcholakova et al. 2006). Consequently, globular proteins tend to form relatively thin, compact interfaces with high viscoelasticity. To be effective emulsifiers, biopolymers must rapidly adsorb to the surfaces of the emulsion droplets created during homogenization, and then form an interfacial coating that prevents the droplets from aggregating with one another (Chapter 6). The interfacial coatings formed by biopolymers can stabilize emulsion droplets against aggregation by a variety of different mechanisms, for example, steric, electrostatic, and hydration repulsion (Chapters 3 and 7). The stabilizing mechanism that dominates in a particular system is largely determined by the characteristics of the interfacial coating formed, for example, thickness, electrical charge, internal packing, hydrophobicity, and exposed chemically reactive groups. The dominant stabilizing mechanism operating in a particular emulsion determines the sensitivity of the system to droplet aggregation under different solution and environmental conditions, for example, pH, ionic strength, temperature, and solvent quality. In the following sections, the interfacial properties and emulsion stabilizing abilities of a number of food proteins and polysaccharides commonly used as emulsifiers is described. Flexible biopolymer Polar segments Globular biopolymer Water Oil Nonpolar segments FIGURE 4.26 biopolymers. The properties of an interfacial layer depend on the molecular structure and interactions of the adsorbed Emulsion Ingredients 145 4.4.2.3 Biopolymer-Based Food Emulsifiers Many food emulsions are stabilized by surface-active biopolymers that adsorb to droplet surfaces and form protective coatings. Some of these functional biopolymers are integral components of more complex food ingredients used in food products (e.g., milk, eggs, meat, fish, and flour), whereas others have been isolated from their natural environments and possibly modified before being sold as specialty ingredients (e.g., protein concentrates, protein isolates, and amphiphilic hydrocolloids). In this section, we focus primarily on those surface-active biopolymers that are sold as functional ingredients specifically designed for use as emulsifiers in foods. In addition, we focus on the ability of biopolymers to create stable oil-in-water emulsions, rather than on their interfacial activity, since the former is more relevant to their application as emulsifiers in the food industry. This point can be clearly illustrated by considering the interfacial characteristics of globular proteins near their isoelectric point. Globular proteins are capable of rapidly adsorbing to oil–water interfaces and forming thick viscoelastic layers near their isoelectric points, but they will not form stable emulsions because the electrostatic repulsion between the droplets is insufficient to prevent droplet aggregation. 4.4.2.3.1 Proteins The interfacial coatings formed by proteins are usually relatively thin and electrically charged; hence, the major mechanism preventing droplet flocculation in protein-stabilized emulsions is electrostatic repulsion (Dickinson 2003, McClements 2004, Lam and Nickerson 2013). Consequently, protein-stabilized emulsions are particularly sensitive to pH and ionic strength effects, and will tend to flocculate at pH values close to the isoelectric point of the adsorbed proteins and when the ionic strength exceeds a certain level. Emulsions stabilized by globular proteins are also particularly sensitive to thermal treatments, because these proteins unfold when the temperature exceeds a critical value exposing reactive nonpolar and sulfhydryl groups. These reactive groups increase the attractive interactions between droplets, which may lead to droplet flocculation. A number of strategies have been developed to improve the emulsifying properties of protein ingredients, including limited hydrolysis to form peptides, modification of protein structure by chemical, physical, enzymatic or genetic means, and blending of the proteins with other ingredients, although not all of these processes are currently legally allowed. 4.4.2.3.1.1 Milk Proteins Protein ingredients isolated from bovine milk are used as emulsifiers in numerous emulsion-based food products, including beverages, frozen desserts, ice creams, sports supplements, infant formula, and salad dressings. Milk proteins can be conveniently divided into two major categories: caseins (∼80 wt%) and whey proteins (∼20 wt%) (Swaisgood 2008). Casein and whey protein fractions can be separated from each other by causing the casein to precipitate from solution (the “curd”) and leaving the whey proteins in solution (the “whey”). Casein precipitation can be achieved by adjusting the pH close to the isoelectric point (∼4.6) of the caseins or by adding an enzyme called rennet that cleaves the hydrophilic fraction of casein that is normally responsible for stabilizing casein micelles. If isoelectric precipitation is used the separated fractions are called “acid casein” and “acid whey,” whereas if enzyme precipitation is used the separated fractions are called “rennet casein” and “sweet whey.” The fractions separated using these two processes have different compositions, and therefore ingredients produced from them may have different functional properties. Curd formation is a critical step in the creation of cheese, and there are large quantities of whey remaining from this process that can be used to make functional whey protein ingredients. A variety of milk protein ingredients are available for utilization as emulsifiers in foods, including whole milk, whey proteins, and caseins. These ingredients are usually sold in a powdered form, which is light cream to white in appearance and has a bland flavor. These powders are normally available in the form of protein concentrates (25%–80% protein) or protein isolates (>90% protein). It should be noted that there are numerous different kinds of proteins in both casein and whey ingredients (see the following text), and that it is possible to fractionate these ingredients into individual purified fractions (e.g., β-lactoglobulin or β-casein). Purified fractions are normally too expensive to be used as emulsifying ingredients in the food industry, but they are frequently used in research studies because they facilitate the development of a more fundamental understanding of protein functionality in emulsions. 146 Food Emulsions: Principles, Practices, and Techniques There are four main protein fractions in casein: αS1 (∼44%), αS2 (∼11%), β (∼32%), and κ (∼11%). In general, these molecules have relatively random and flexible structures in solution, although they do have a limited amount of secondary and tertiary structure. The caseins also have some regions that are highly nonpolar and others that are highly charged, which plays a major role in determining their molecular and functional properties in foods. There is considerable interest in the use of plant proteins because of their potential advantages over other ingredients in terms of sustainability, clean labels, and specialist diets (vegan, vegetarian, Kosher, etc.). In their natural state, the caseins tend to exist as complex molecular clusters called “micelles” that are typically between 50 and 250 nm in diameter and are partly held together by mineral ions (such as calcium phosphate). In commercial ingredients, caseins may also be present in a number of other sorts of molecular cluster depending on the way that the proteins were isolated, for example, sodium caseinate, calcium caseinate, acid casein, and rennet casein. Caseinate-stabilized emulsions have been shown to be unstable to droplet flocculation at pH values (3.5–5.3) close to the protein’s isoelectric point and at relatively high ionic strengths (Hunt and Dalgleish 1995, Agboola and Dalgleish 1996). Caseinate-stabilized emulsions tend to be more stable to heating than whey protein-stabilized emulsions, presumably because the relatively flexible casein molecules do not undergo appreciable heat-induced conformational changes like globular proteins do (Hunt and Dalgleish 1995). It should be noted that sufficiently high concentrations of nonadsorbed caseinate can promote emulsion instability through a depletion flocculation mechanism (Dickinson and Golding 1997). The caseinate concentration where depletion flocculation occurs depends on the size of the nonadsorbed casein micelles, which is governed by factors such as solution composition and environmental conditions. Whey protein is also a complex mixture of different individual proteins, with the most common being β-lactoglobulin (∼55%), α-lactalbumin (∼24%), serum albumin (∼5%), and immunoglobulins (∼15%) (Swaisgood 2008). Normally, β-lactoglobulin dominates the functional characteristics of whey proteins because of its relatively high concentration and specific physicochemical properties. Whey protein-stabilized emulsions tend to flocculate at pH values (∼4–5.5) close to their isoelectric point (pI ∼ 5), at high salt concentrations, and upon heating above the thermal denaturation temperature of the adsorbed proteins in the presence of salt (McClements 2004). Users of whey protein emulsifiers in the food industry have reported that large variations in their functional properties can occur from batch to batch, which has been attributed to the presence of mineral impurities and partial denaturation of the proteins during their isolation. Preferential adsorption and competitive displacement of milk proteins with each other and with other types of emulsifiers have been widely studied because this process alters interfacial composition, and therefore emulsion stability and performance (Pugnaloni et al. 2004, Dickinson 2011). 4.4.2.3.1.2 Meat and Fish Proteins Meat and fish contain a number of proteins that are surface active and capable of stabilizing emulsions, such as gelatin, myosin, actin, and actomyosin, with gelatin being the most commonly used (Karim and Bhat 2009, Gomez-Guillen et al. 2011, Taherian et al. 2011). Many of these proteins play an important role in stabilizing meat emulsions, that is, products formed by blending or homogenizing fat, meat, and other ingredients together. Emulsion stabilization is partly due to their ability to adsorb to the oil–water interface and partly due to their ability to increase the aqueous phase viscosity or to form a gel in the aqueous phase. Gelatin is one of the few proteins that have been isolated from meat and fish and sold commercially as a functional emulsifier ingredient. Gelatin is a relatively high molecular weight protein derived from animal or fish collagen, for example, pig, cow, or fish. Gelatin is prepared by hydrolyzing collagen by boiling in the presence of acid (Type A gelatin) or alkaline (Type B gelatin). The isoelectric point (pI) of Type A gelatin (∼7 to 9) tends be higher than that of Type B gelatin (∼5). Gelatin exists as a random coil molecule at relatively high temperatures, but undergoes a helix-to-coil transition upon cooling below a critical temperature, which is about 10°C–25°C for pig and cow gelatin and about 0°C–5°C for fish gelatin (Leuenberger 1991). Gelatin has been shown to be surface active and capable of acting as an emulsifier in oil-in-water emulsions (Surh et al. 2006). Nevertheless, when used on its own gelatin often produces relatively large droplet sizes during homogenization that are not highly stable (Dickinson and Lopez 2001), so that it has to be hydrophobically modified by attachment of nonpolar side groups (Toledano and Magdassi 1998) or used in conjunction with anionic surfactants to improve its effectiveness as an emulsifier (Olijve et al. 2001). Research has Emulsion Ingredients 147 been carried out to establish the ability of various other protein fractions of fish and meat muscle to act as emulsifiers, for example, actomyosin (Mignino et al. 2011) and myosin (Petursson et al. 2004). The ultimate objective of this work is to be able to convert waste products from fish and meat production into value-added functional ingredients for use as emulsifiers in foods. Nevertheless, there are currently few examples of functional ingredients derived from fish or meat products (other than gelatin) designed, especially as emulsifiers. 4.4.2.3.1.3 Egg Proteins Both egg yolk and egg white contain a mixture of protein and nonprotein components that are surface active (Mine 2002, Anton 2013). Egg ingredients can be purchased in a variety of different forms for utilization in food emulsions, including fresh egg yolks, frozen egg yolks, dried egg yolks, fresh whole eggs, frozen whole eggs, and dried whole eggs. Different egg ingredients are usually prepared using different processing treatments, which often influences their effectiveness at stabilizing emulsions. In the food industry, egg white is more commonly used for stabilizing foams, whereas egg yolk is more commonly used for stabilizing emulsions. Nevertheless, a number of studies have shown that egg white proteins can be used to stabilize oil-in-water emulsions. Egg yolk is widely used as an emulsifier in the production of mayonnaise, salad dressings, sauces, and cake batters. The effectiveness of whole egg yolk and its individual constituents (plasma and granules) at forming oil-inwater emulsions have been compared. Studies of the ability of whole egg yolk, plasma, and granules to stabilize oil-in-water emulsions prepared using a high-pressure valve homogenizer have also been carried out (Le Denmat et al. 1999, Le Denmat et al. 2000). These researchers found that the main contributors to egg yolk functionality as an emulsion stabilizer were the plasma constituents rather than the granules. Emulsions stabilized by egg yolk were found to be stable to droplet flocculation at pH 3 at relatively low salt concentrations (150 mM NaCl), but unstable to flocculation at pH 3 at high salt concentrations (550 mM NaCl) and at pH 7 (150 and 550 mM NaCl) (Anton et al. 2002). The instability of these emulsions was attributed to depletion, bridging, and electrostatic screening effects. It therefore seems that egg yolk is better at forming emulsions at high pH, but stabilizing emulsions at low pH. Understanding the influence of pH and salt concentration on the stability of egg yolk stabilized emulsions is often complicated because these factors influence the solubility and structural organization of the protein molecules, as well as the interactions between the emulsion droplets. Like other globular proteins, the proteins in eggs will unfold and aggregate upon heating above their thermal denaturation temperature, which influences the stability and rheological properties of emulsions. Emulsions stabilized by egg yolk also have poor stability to freeze–thaw cycling. Preferential adsorption and competitive displacement of egg yolk proteins with each other and with other types of emulsifiers have been reviewed (Mine 2002). 4.4.2.3.1.4 Plant Proteins Surface active proteins can be extracted from a variety of plant sources, including legumes and cereals (Aoki et al. 1980, Moure et al. 2006). A considerable amount of research has been carried out to establish the ability of these proteins to stabilize emulsions, and whether they could be made into commercially viable value-added ingredients for utilization as emulsifiers in foods. One of the most widely studied proteins extracted from a plant source is soy protein, which is commercially available as a protein concentrate or isolate (Nishinari et al. 2014). Soy protein ingredients are a complex mixture of many individual protein fractions with different molecular and functional characteristics, for example, 2S, 7S, 11S, and 15S fractions. In addition, each of these fractions contains a mixture of different protein subunits that also have different molecular and functional characteristics. Soy proteins are surface active molecules that can adsorb to oil–water interfaces during homogenization and form a protective coating around droplets that provides stability during storage. Nevertheless, the ability to form and stabilize emulsions depends strongly on quality of the soy protein ingredient used. Soy proteins can vary widely in their composition and in their denaturation state depending on how they are isolated and purified, which can impact their functionality. As with other globular proteins, soy protein-stabilized emulsions are highly susceptible to aggregation when environmental conditions such as pH, ionic strength, and temperature are changed. 148 Food Emulsions: Principles, Practices, and Techniques 4.4.2.3.2 Polysaccharides 4.4.2.3.2.1 Gum Arabic Gum arabic is widely used as an emulsifier in the beverage industry to stabilize cloud and flavor emulsions (Islam et al. 1997, Williams and Phillips 2009, Piorkowski and McClements 2014). It is derived from the natural exudate of Acacia senegal and consists of at least three high molecular weight biopolymer fractions. The surface-active fraction is believed to consist of branched arabinogalactan blocks attached to a polypeptide backbone (Dickinson 2003). The hydrophobic polypeptide chain is believed to anchor the molecules to the droplet surface, whereas the hydrophilic arabinogalactan blocks extend into solution. The thick hydrophilic interfacial coating formed by gum arabic provides stability against droplet aggregation primarily through steric repulsion, but with some contribution from electrostatic repulsion also. The influence of a variety of processing conditions on gum arabic functionality has been examined. For example, it has been shown that gum arabic stabilized emulsions remain stable to droplet flocculation when exposed to a wide range of conditions, for example, pH (3–9), ionic strength (0–25 mM CaCl2), and thermal treatment (30°C–90°C) (Chanamai and McClements 2002, Qian et al. 2011). Nevertheless, gum arabic has a relatively low affinity for oil–water interfaces compared to most other surface-active biopolymers, which means that it has to be used at relatively high concentrations to form stable emulsions. For example, a ≈ 1:1 mass ratio of gum arabic-to-oil phase is often needed to form stable oil-in-water emulsions, although this amount has been lowered for some of the newer gum arabic ingredients. For this reason, its application as an emulsifier is restricted to products that have relatively low droplet concentrations, for example, beverage emulsions. In addition, there have been frequent problems associated with obtaining a reliable source of consistently high-quality gum arabic that has led many food scientists to investigate alternative sources of biopolymer emulsifiers for use in beverages. Gum arabic has a high water solubility and a relatively low solution viscosity compared to other gums, which facilitates its application as an emulsifier. 4.4.2.3.2.2 Modified Starches Natural starches are hydrophilic molecules that have poor surface activity. Nevertheless, they can be made into effective emulsifiers by chemically attaching hydrophobic moieties along their backbones (Sweedman et al. 2013). These modified starches are widely used as emulsifiers in the beverage industry (Piorkowski and McClements 2014). One of the most commonly used modified starches is an octenyl succinate derivative of waxy maize. It consists primarily of amylopectin that has been chemically modified to contain a side group that is anionic and nonpolar. These side groups anchor the molecule to the oil droplet surface, whereas the hydrophilic starch chains protrude into the aqueous phase and protect droplets against aggregation through steric repulsion. Because the dominant stabilizing mechanism is steric repulsion, emulsions stabilized by modified starch are resistant to changes in pH (3–9), ionic strength (0–25 mM CaCl2), and temperature (30°C–90°C) (Chanamai and McClements 2002, Qian et al. 2011). Like gum arabic, modified starch has a relatively low interfacial activity (compared to proteins or surfactants), and so a large excess must be added to ensure that all the droplet surfaces are adequately coated. For example, about a 1:1 mass ratio of modified starch-to-oil phase is required to produce stable oil-in-water emulsions using traditional OSA starch ingredients, but this value has been reduced somewhat for more recently developed modified starch ingredients (Charoen et al. 2011). Modified starches usually come in powdered or granular forms that are easily dispersible in cold water. 4.4.2.3.2.3 Modified Celluloses In its natural state, cellulose is not usually suitable for utilization as an emulsifier because it forms strong intermolecular hydrogen bonds, which makes it insoluble in water. Nevertheless, it can be isolated and modified in a number of ways to produce food-grade ingredients that have interfacial activity and can be used as emulsifiers (Huang et al. 2001). The most commonly used surface-active cellulose derivatives are methyl cellulose (MC), hydroxypropyl cellulose (HPC), and methyl hydroxypropyl cellulose (MHPC). These ingredients are all nonionic polymers that are soluble in cold water, but tend to become insoluble when the solution is heated above a critical temperature (around 50°C–90°C). They have good stability to pH (2–11), salt and freeze–thaw cycling, which may be beneficial in a number of food emulsion applications. 4.4.2.3.2.4 Other Polysaccharides A number of studies have shown that various other types of polysaccharide are capable of reducing oil–water interfacial tensions and forming stable emulsions, for Emulsion Ingredients 149 example, galactomannans, pectin, and chitosan (Garti 1999, Garti and Leser 2001, Huang et al. 2001, Dickinson 2003, Leroux et al. 2003). There is some debate about the molecular origin of the surface activity of these hydrocolloids (e.g., nonpolar regions on the polysaccharide molecule itself, protein contaminants, or proteins covalently bound to the polysaccharide backbone), and about whether their ability to form stable emulsions is primarily due to their surface activity, coadsorption with other molecules, or ability to thicken the aqueous phase (Dickinson 2003, 2011). 4.4.2.4 Protein–Polysaccharide Complexes Proteins tend to be better at producing small emulsion droplets when used at low concentrations than polysaccharides, whereas polysaccharides tend to be better at producing emulsions that are stable to a wider range of environmental conditions than proteins, for example, pH, ionic strength, temperature, and freeze–thaw cycling. It may therefore be advantageous to combine the beneficial attributes of these two kinds of biopolymer to produce small emulsion droplets with good environmental stability. A number of researchers have shown that protein–polysaccharide complexes may have better emulsifying properties than either of the biopolymers used in isolation (Dickinson 2003, 2011, Guzey and McClements 2006). In particular, substantial improvements have been shown in the stability of oil-in-water emulsions to environmental stresses such as pH alterations, high salt contents, thermal processing, freezing–thawing, and dehydration. Protein–polysaccharide complexes may be held together by either physical or covalent bonds and may be formed either before, during, or after homogenization. Ingredients based on protein– polysaccharide interactions will have to be legally acceptable, economically viable, and show benefits over existing ingredients before they find widespread utilization in the food industry. It should be noted that gum arabic is a naturally occurring protein–polysaccharide complex that is already widely used in the food industry as an emulsifier. 4.4.3 Selection of an Appropriate Emulsifier In this section, we discuss some schemes for classifying and comparing the effectiveness of different types of food emulsifier, as well as some of the factors that should be considered when selecting an emulsifier for a particular application. As has been mentioned earlier an effective emulsifier should have the following general characteristics (1) it should be capable of rapidly adsorbing to the surface of freshly formed droplets during homogenization; (2) it should be capable of reducing the interfacial tension by a significant amount; and (3) it should be capable of forming an interfacial coating that is either resistant to rupture and/or provides a sufficiently strong repulsive interaction between the droplets. A number of food-grade ingredients exhibit these general characteristics and can be used as emulsifiers, but they vary considerably in their ability to form and stabilize emulsions, as well as in their sensitivity to environmental conditions, for example, pH, ionic strength, temperature, solvent composition, shearing, and dehydration (Table 4.7). It would therefore be useful to have a standardized means of assessing the relative efficiency of different types of emulsifiers for specific applications. Unfortunately, there has been little attempt to systematically compare the advantages and disadvantages of different emulsifiers under standardized conditions, so that it is currently difficult for food manufacturers to rationally select the most suitable ingredient for particular products. One of the purposes of this section is to highlight some criteria that could form the basis for such a comparison. Food manufacturers usually measure and compare the functional properties of emulsifiers in terms of parameters that depend on the processing procedure and formulation of their actual food product (McClements 2007), for example: • The minimum droplet size (dmin) that can be produced by a certain amount of emulsifier for a specified emulsion system using specified homogenization conditions. • The minimum amount of emulsifier (cmin) required to produce a desired droplet size for a specified emulsion system using specified homogenization conditions. • The long-term stability (e.g., to creaming, flocculation, or coalescence) of a specified emulsion system produced by an emulsifier using specified storage conditions. 150 Food Emulsions: Principles, Practices, and Techniques The characteristics of the specified emulsion system (e.g., oil type, oil concentration, and aqueous phase composition) used to establish the efficiency of an emulsifier depends on the food being produced and will vary considerably from product to product. In addition, the specified homogenization conditions will also vary according to the type of homogenizer used (e.g., high-speed blender, high-pressure valve homogenizer, microfluidizer, or colloid mill) and the precise operating conditions (e.g., energy input, flow rate, and temperature). The above approach is particularly suited for food manufacturers trying to determine the best emulsifier for utilization in their specific product, but it is not particularly suited for development of a general classification scheme because of the wide variation in the composition and processing of different foods. This approach could be used to develop a more general classification scheme by stipulating standardized emulsion systems and homogenization conditions. The analytical methods developed to measure emulsifier capacity and emulsion stability index (Chapter 14) are attempts at developing emulsifier classification schemes based on this principle. Colloid and interfacial scientists often characterize emulsifier properties in terms of quantitative physical parameters that can be measured using fundamental analytical instruments under well-defined environmental conditions (McClements 2007): • Surface load, Γsat: The surface load at saturation is the mass of emulsifier adsorbed per unit surface area of interface when the interface is saturated with emulsifier, and is usually expressed as mg m−2 (Chapters 5 and 14). The surface load provides a measure of the minimum amount of emulsifier required to produce an emulsion with a given surface area (or droplet size): the higher Γ, the greater the amount of emulsifier required to completely cover the same surface area. • Maximum surface pressure, πmax: The maximum surface pressure is the interfacial tension of an oil–water interface in the absence of emulsifier minus the interfacial tension of the same interface when it is saturated with emulsifier (Chapter 5). It provides a measure of the ability of an emulsifier to decrease the oil–water interfacial tension, and thereby facilitate droplet disruption: the higher πmax, the lower the Laplace pressure and the smaller the droplets that can be produced in a homogenizer at a fixed energy input, provided there is sufficient emulsifier present and that it adsorbs rapidly to the droplet surfaces (Chapter 6). • Binding affinity, c1/2: The binding affinity is a measure of how strongly an emulsifier adsorbs to an oil–water interface (Chapter 5). It can be expressed as the emulsifier concentration at which the surface pressure is half the maximum surface pressure. The stronger the binding affinity (the lower c1/2), the lower the concentration of emulsifier required to reach interfacial saturation. • Adsorption kinetics, τads: Adsorption kinetics can be defined in terms of the average time required for an interface to become saturated with emulsifier (Chapter 5). It is important that this time be measured under conditions that adequately represent the highly dynamic conditions that occur in most homogenizers. In practice, it is difficult to establish an accurate measure of the adsorption kinetics of different emulsifiers under realistically dynamic conditions. • Droplet aggregation stability: The aggregation stability is a measure of the tendency for droplets to become aggregated (flocculated or coalesced) under a specified set of environmental conditions, for example, pH, ionic strength, temperature, and shearing rate (Chapter 7). It can be expressed in a number of different ways, for example, the percentage of droplets that are flocculated or coalesced, the percentage of droplets larger than a specified size, or the percentage increase in the mean size of the particles in an emulsion due to droplet aggregation. One of the major challenges of food scientists is to relate these more fundamental parameters to the more practical parameters mentioned above that are of interest to food manufacturers. An attempt has been made to compare the relative efficiencies of different types of emulsifiers at stabilizing food emulsions (Table 4.7). This comparison shows that nonionic surfactants can be used at Emulsion Ingredients 151 low levels, and provide good stability to droplet aggregation over a range of environmental conditions. Proteins can also be used at relatively low levels, but their ability to stabilize emulsions against droplet aggregation is strongly influenced by pH, ionic strength, and temperature. Emulsions stabilized by polysaccharides have much better stability to environmental stresses than proteins due to the fact that the predominant stabilizing mechanism is steric rather than electrostatic, but they usually have to be used at much higher levels. The discussion above has highlighted the wide variety of emulsifiers available for use in food products. A food manufacturer must decide which of these emulsifiers is the most suitable for utilization in each particular product. In addition to the physicochemical characteristics considered above, a food manufacturer must also consider a number of economic, legal, and marketing factors when selecting a suitable emulsifier. The most important of these are discussed at the end of this chapter (Section 4.7). 4.5 Texture Modifiers A number of ingredients commonly used in food emulsions are added because of their ability to modify the texture of the continuous phase (usually the aqueous phase of oil-in-water emulsions). These texture modifying ingredients can be conveniently divided into “thickening agents” and “gelling agents” depending on the molecular origin of their functional characteristics. Thickening agents are ingredients whose functional characteristics are due to their highly extended molecular conformation in solution, whereas gelling agents are those ingredients whose functional characteristics are due to their ability to associate with each other through intermolecular cross-links (see below). Nevertheless, in practice there is often no clear distinction between these two different categories of texture modifiers, since thickening agents can form gels when used at sufficiently high concentrations and gelling agents can increase the viscosity of aqueous solutions (without forming gels) when used at sufficiently low concentrations. In addition, a particular type of biopolymer may act as a thickening agent under some conditions, but a gelling agent under other conditions, for example, at a different temperature, pH, or ionic strength. The major roles of texture modifiers in food emulsions are to provide the product with desirable textural and mouthfeel characteristics, and to improve emulsion stability by reducing the rate at which particulate matter moves, such as oil droplets, herbs, spices, and air bubbles. 4.5.1 Thickening Agents The primary function of thickening agents in food emulsions is to increase the viscosity of the aqueous phase of oil-in-water emulsions. This viscosity enhancement modifies the texture and mouthfeel of food products (“thickening”), as well as reducing the rate at which particles sediment or cream (“stabilization”). Thickening agent ingredients are usually supplied as powders or granules consisting of one or more types of biopolymer, and possibly other components (such as sugars or minerals). The biopolymers found in thickening agents usually exist as highly hydrated and extended molecules (or molecular assemblies) in aqueous solutions. Their ability to increase the viscosity of a solution depends principally on their molecular weight, degree of branching, conformation, and flexibility. In this section, we consider the relationship between the molecular characteristics of biopolymers and their ability to act as thickening agents. Specific types of thickening agents commonly used in the food industry are outlined in Section 4.5.3. 4.5.1.1 Effective Volume of Biopolymers in Aqueous Solutions The effectiveness of a biopolymer at enhancing the viscosity of an aqueous solution is largely determined by its molecular structure. The effective volume of a biopolymer in solution may be considerably greater than the volume occupied by the actual biopolymer chain alone because it entraps a large volume of solvent (Figure 4.27). It is convenient to characterize this phenomenon in terms of a volume ratio, RV: 152 Food Emulsions: Principles, Practices, and Techniques 10 Rv 1 10 100 1,000 10,000 Relative viscosity 8 6 4 2 0 0.01 0.1 1 10 100 Biopolymer concentration (kg m–3) FIGURE 4.27 Prediction of change in relative viscosity of aqueous biopolymer solutions with biopolymer concentration for different effective volume ratios, RV (shown in box). The viscosity increases dramatically when the biopolymer molecules start to overlap with one another, which occurs at lower biopolymer concentrations for higher RV. RV = VE 4prH3rN A » VA 3M (4.9) where VE is the “effective” volume of the biopolymer molecule in solution VA is the actual volume occupied by the biopolymer chain r H is the radius of hydration of the molecule ρ is the density of the biopolymer chain NA is Avagadro’s number M is the molecular weight of the biopolymer 4.5.1.2 Relationship between Biopolymer Molecular Structure and Effective Volume in Solution The effective volume of a biopolymer depends on its three-dimensional structure in solution (Figure 4.27). For molecules that form compact globular structures (such as many globular proteins), the actual volume of the molecule is close to its effective volume and therefore R V ≈ 1. The average end-to-end length (L) of random coil molecules is given by L ≈ l√n, whereas for rigid rod-like molecules it is given by L ≈ ln, where l is the length of the monomer unit and n is the number of monomers per molecule (Grosberg and Khokhlov 2010). If we assume that the radius of hydration of a biopolymer 153 Emulsion Ingredients molecule is half the end-to-end length, then we can obtain expressions for the effective volume of different types of molecule: Globular biopolymers: RV ≈ 1 (4.10) RV » pn3 / 2l 3rN A 6M (4.11) RV » pn3l 3rN A 6M (4.12) Random coil biopolymers: Rigid rod-like biopolymers: where M0 is the molecular weight of a monomer segment n = M/M0 In practice, real biopolymers often have some regions that are compact, some that are rod-like and some that are flexible and therefore they fall somewhere between these extremes. Nevertheless, these equations give us some indication of the expected volume ratios of real biopolymers. For example, the molecular weight of polysaccharide segments is typically about 168 Da, and the length of a segment is typically about 0.47 nm (Voet and Voet 2010). The molecular weights of polysaccharides used as thickening agents typically vary between about 5 and 2000 kDa (Cui 2005, Belitz et al. 2009). We would therefore expect volume ratios ranging from around unity to thousands of millions depending on the structure and molecular weight of the polysaccharide. The above discussion indicates that biopolymers that have highly extended structures in solution have larger volume ratios than those that have compact structures. Thus, RV tends to be higher for linear than for branched biopolymers with the same molecular weight and tends to increase as the electrostatic repulsion between different segments on charged biopolymer molecules increase because this causes the molecule to become more extended (Walstra 2003). 4.5.1.3 Viscosity Enhancement by Biopolymers in Solution The apparent viscosity (η) of a colloidal dispersion containing spherical rigid particles suspended in an ideal liquid can be described over a wide range of particle concentrations using the following semiempirical equation (Liu and Masliyah 1996): h æ fö = 1h1 çè P ÷ø -[ h] P (4.13) where η1 is the viscosity of the continuous phase æ h / h1 - 1 ö [η] is the intrinsic viscosity = lim ç ÷ f® 0 f ø è ϕ is the volume fraction of the particles P is a packing parameter The packing parameter is related to the volume fraction at which the particles become close packed, which depends on the applied shear stress and the polydispersity of the particles (Hunter 1986). 154 Food Emulsions: Principles, Practices, and Techniques For rigid monodisperse spherical particles, the following parameters have been determined experimentally: P = 0.57 at low shear stresses, P = 0.68 at high shear stresses, and [η] = 2.67. To a first approximation, the viscosity of a suspension of hydrated biopolymer molecules rotating in solution can be treated in a similar manner (McClements 2000): f ö h æ » 1 - eff ÷ P ø h1 çè -[ h] P æ R cö » ç1 - V ÷ Pr ø è -[ h] P (4.14) where ϕeff is the effective volume fraction of the biopolymer molecules in solution (=ϕ × RV) ϕ is the actual volume fraction occupied by the biopolymer chains (≈c/ρ) c is the polysaccharide concentration (in kg m−3 of emulsion) ρ is the density of the biopolymer chains (in kg m−3), which is approximately 1600 kg m−3 (Rahman 2009) Theoretical predictions of viscosity versus biopolymer concentration for molecules with different volume ratios are shown in Figure 4.27. For convenience, it was assumed that the shear stresses applied to the emulsions were in the low shear regime so that P = 0.57. The viscosity increases dramatically when the biopolymer concentration exceeds a critical concentration, whose value decreases as the volume ratio (RV) increases. In practice, Equation 4.14 only gives a very rough approximation of the viscosity of aqueous biopolymer solutions because the flexible biopolymer molecules cannot be treated as rigid spherical particles. The biopolymer molecules may become aligned with the shear field, interact with each other, or become entangled, thus changing their effective volume with shear stress. Nevertheless, the above equation does provide some useful insights into the relationship between the viscosity of polysaccharide solutions and the molecular structure of polysaccharide molecules. The dependence of the rheology of an aqueous solution on biopolymer concentration can be divided into a number of different regions depending on the interaction between the molecules (Dickinson 1992). In the “dilute region,” the biopolymer concentration is so low that the molecules (or molecular aggregates) do not interact with each other and can be treated as separate entities. As the concentration of biopolymer increases above some critical value, c* (≈P/RV), the viscosity of the solution increases rapidly because the spheres swept out by the biopolymers begin to interact with each another (Figure 4.28). Between the dilute region and this critical concentration, the solution is known as a semidilute solution, because even though the molecules are interacting with one another, each individual biopolymer is still largely surrounded by solvent molecules. At polymer concentrations above this critical concentration, the molecules pack so close together that they become entangled with each other and the system has more gel-like characteristics. Biopolymers that are used to thicken the aqueous phase of emulsions are often used in the semidilute concentration range. A more detailed discussion of the influence of particle concentration on the rheology of colloidal dispersions is given in Chapter 8. 4.5.1.4 Shear-Thinning in Biopolymer Solutions Solutions containing extended biopolymers often exhibit strong shear-thinning behavior (pseudoplasticity), that is, their apparent viscosity decreases with increasing shear stress (Williams and Phillips 2003). The molecular origin of this pseudoplasticity has been attributed to a number of molecular events that occur when shear is applied to a biopolymer solution, such as alignment and stretching of biopolymers along the shear field, disentanglement of biopolymers, or disruption of weak physical interactions holding biopolymers together. Each of these molecular events has a characteristic relaxation time associated with it, which is the time taken for the system to adjust to the applied mechanical stresses. At relatively low shear rates (long deformation times), the system has sufficient time to relax on the experimental time scale, and so the viscosity remains high (e.g., since biopolymers remain randomly orientated, associated, or entangled). As the shear rate is increased, these molecular events occur on a time scale similar to the experimental time scale, and so the viscosity begins to decrease. At sufficiently high shear rates (short deformation times), the system does not have time to relax within the experimental time scale and so a 155 Emulsion Ingredients Biopolymer Hydrodynamically entrained water Relative viscosity Conc. Semi-dilute Dilute Biopolymer concentration Dilute solution Semidilute solution Concentrated solution FIGURE 4.28 The impact of biopolymers on the viscosity of aqueous solutions can be divided into three concentration regimes: dilute, semidilute, and concentrated. low constant viscosity is reached (e.g., the biopolymers remain aligned, dissociated, or disentangled). The viscosity of many biopolymer solutions therefore changes from a relatively constant high value at low shear rates, decreases at intermediate shear rates, and reaches a relatively constant low value at high shear rates (Figure 4.29). Some biopolymer solutions may even have a yield stress due to the formation of a three-dimensional network of interacting molecules that gives some solid-like characteristics to the system. When a biopolymer solution experiences an applied stress below the yield stress it acts like an elastic solid, but when it experiences an applied stress that exceeds the yield stress it acts like a liquid (Chapter 8). The characteristic rheological behavior of biopolymer solutions plays an important role in determining their functional properties in food emulsions. For example, a salad dressing must be able to flow when it is poured from a container, but must maintain its shape under its own weight after it has been poured onto a salad. The amount and type of biopolymer used must therefore be carefully selected so that it provides a low viscosity when the salad dressing is poured (high applied stress), but a high viscosity when the salad dressing is allowed to sit under its own weight (low applied stress). The viscosity of biopolymer solutions is also related to the mouthfeel of a food product during oral processing of foods. Liquids that do not exhibit extensive shear thinning behavior at the shear stresses experienced within the mouth are perceived as being “slimy.” On the other hand, a certain amount of viscosity is needed to contribute to the “creaminess” of a product. The shear thinning behavior of biopolymer solutions is also important for determining the stability of food emulsions to creaming. As an oil droplet moves through an aqueous phase it only exerts a very small shear stress on the surrounding liquid. As a result of the shear-thinning behavior of the solution, it experiences a very high viscosity which greatly slows down the rate at which it creams.* Many biopolymer solutions also exhibit a shear-thinning behavior known as thixotropy, that is, their apparent viscosity decreases with time when they are sheared at a constant rate. The molecular origin of thixotropy can also be attributed to be the fact that applied shear stresses can cause biopolymer alignment, biopolymer disentanglement, or disruption of weak physical forces between biopolymer molecules. * It should be noted that biopolymers can actually promote creaming at certain concentrations because they cause depletion flocculation (Section 3.6). 156 Food Emulsions: Principles, Practices, and Techniques Elongation Dissociation Disentanglement Apparent viscosity (Pa s) η0 η∞ Shear stress (Pa) FIGURE 4.29 Schematic representation of the dependence of apparent shear viscosity on applied shear stress for a biopolymer thickening agent that exhibits shear thinning. Shear thinning may occur due to changes in biopolymer conformation, alignment, association, or entanglement when the shear stress is increased. Once the shearing stress is removed, the biopolymer molecules may be able to undergo molecular rearrangements that enable the biopolymers to become nonaligned, entangled, or associated with their neighbors again, and so the system regains its original structure and rheological properties. This type of system is said to be reversible, and the speed at which the structure is regained may be important for the practical application of a biopolymer in a food. If the molecular rearrangements are unable to take place once the stress is removed, or if they are only able to partially take place, then the system is said to be irreversible or partially reversible, respectively. A food manufacturer must therefore select an appropriate biopolymer or combination of biopolymers to produce a final product that has a desirable mouthfeel, stability, and texture. Both proteins and polysaccharides can be used as thickening agents, but polysaccharides are usually preferred because they tend to have higher molecular weights and be more extended so that they can be used at much lower concentrations (higher RV values). 4.5.2 Gelling Agents Biopolymers are used as functional ingredients in many food emulsions because of their ability to cause the aqueous phase to gel, for example, yogurts, cheeses, deserts, egg, and meat products (Rossmurphy 1995, Williams and Phillips 2003, Foegeding 2006, Foegeding and Davis 2011, Rao 2013). Gel formation often imparts desirable textural and sensory attributes, as well as preventing the droplets and other particles from creaming or sedimenting. A biopolymer gel consists of a three dimensional network of aggregated or entangled biopolymers that entraps a large volume of water, giving the whole structure some “solid-like” characteristics. 157 Emulsion Ingredients The properties of biopolymer gels depend on the type, structure, and interactions of the molecules they contain. Gels may be transparent or opaque, hard or soft, brittle or rubbery, and homogeneous or heterogeneous; exhibit syneresis; or have good water holding capacity. Gelation may be induced by a variety of different methods, including altering the temperature, pH, ionic strength, or solvent quality, or by adding enzymes, denaturants, or other cross-linking agents. Biopolymers may be cross-linked to one another either by covalent and/or noncovalent bonds. The type of cross-links formed depends on the nature of the molecules involved, as well as the prevailing environmental conditions. Some common types of molecular interactions responsible for holding the molecules together in biopolymer gels are schematically illustrated in Figure 4.30. It is often convenient to categorize food gels according to the nature of their basic structure as either particulate or filamentous (Foegeding 2006, Foegeding and Davis 2011). Particulate gels consist of a three-dimensional network of relatively large compact particles, which themselves are usually formed from numerous aggregated biopolymer molecules. This type of gel tends to be formed when the individual biopolymer molecules are able to interact with their neighbors at any point on their surface. Particulate gels are optically opaque because the particles are large enough to strongly scatter light, and are prone to syneresis because the relatively large pore sizes between the particles means that the water is not held tightly within the gel network by capillary forces. Common examples of particulate gels are those formed by heating aqueous solutions of globular proteins (e.g., whey, egg, or soy proteins) at pH values close to their isoelectric point or at high salt concentrations. Under these conditions, individual protein molecules aggregate with each other to form relatively large particles, and then these particles aggregate with each other to form the final gel network. In contrast, filamentous gels consist of thin filaments of individual or aggregated biopolymer molecules (Figure 4.31). Filamentous gels tend to be optically transparent because the filaments are so thin that they do not scatter light strongly. They also tend to have good water holding capacity because the small pore size of the gel network means that the water molecules are held tightly by capillary forces. Examples of filamentous gels are those formed by many hydrocolloids (e.g., gelatin, pectin, gellan, agar, and alginates) and those formed by heating globular proteins at low ionic strengths and pH values sufficiently far from the protein’s isoelectric point. In hydrocolloid gels the filaments are individual molecules, but in globular protein gels the filaments are linear chains containing many protein molecules linked together (Foegeding 2006, Foegeding and Davis 2011). There may be considerable variations in the gelation characteristics of biopolymers depending on their nature (Williams and Phillips 2003, Brady 2013). Some biopolymers form gels upon heating (heat-setting gels), whereas others form them upon cooling (cold-setting gels). Gels may also be either thermo-reversible or thermo-irreversible, depending on whether the sol–gel process is reversible or not. Gelatin is an example of a cold-setting thermo-reversible gel: when a solution of gelatin molecules S (a) (c) COO– Ca2+ –OOC S (b) (d) (e) FIGURE 4.30 Schematic representation of some common junction zones found in biopolymer gels. (a) Covalent bond, (b) salt bridge, (c) hydrogen bonding, (d) hydrophobic attraction, and (e) VDW attraction. 158 Food Emulsions: Principles, Practices, and Techniques (a) (b) FIGURE 4.31 Many food gels can be conveniently categorized as being either (a) particulate or (b) filamentous, depending on the structural organization of the molecules. is cooled below a certain temperature a gel is formed, but when it is reheated the gel melts. Egg white is an example of a heat-setting thermo-irreversible gel: when egg white is heated above a certain temperature a characteristic white gel is formed, but when it is cooled back to room temperature it remains as a white gel, rather than reverting back into the relatively clear liquid from which it was formed. Whether a gel is reversible or irreversible depends on the type of bonds holding the biopolymer molecules together, as well as any changes in the molecular structure and organization of the molecules during gelation. Biopolymer gels that are stabilized by noncovalent interactions, and which do not involve permanent changes in the structure of the individual molecules during the gelation processes, tend to be reversible. On the other hand, gels that are held together by covalent bonds, or which involve permanent changes in the structure of the individual molecules prior to gelation, tend to form irreversible gels. The type of interactions holding the molecules together in gels varies from biopolymer to biopolymer (Figure 4.30), and plays a large role in determining the response of a gel to changes in its environment. Some proteins and polysaccharides form helical junction zones through extensive hydrogen bond formation (Table 4.8). This type of junction zone tends to form when a biopolymer solution is cooled and be disrupted when it is heated, and is thus responsible for the formation of cold-setting reversible gels. Below the gelation temperature, hydrogen bonding favors junction zone formation between helices on different biopolymers, but above this temperature the configurational entropy favors a random-coil type structure and the junction zones are disrupted. Biopolymers with extensive nonpolar groups tend to associate via hydrophobic interactions, for example, caseins or denatured whey proteins. Many biopolymers have electrical charges that vary with pH and mineral environment Figure 4.32. For these biopolymers, electrostatic interactions often play an important role in determining their gelation behavior, and so gelation is particularly sensitive to the pH and ionic strength. For example, at pH values sufficiently far away from their isoelectric point, proteins may be prevented from gelling because of the strong electrostatic repulsion between the molecules; however, if the pH is adjusted near to the isoelectric point, or if salt is added, the proteins tend to aggregate and form a gel. The addition of cationic multivalent ions (such as Ca2+) can promote gelation of anionic biopolymers (e.g., alginate, pectin, or carrageenan) by forming salt bridges between anionic groups or helical regions on different molecules. Similarly, anionic multivalent ions (such as tripolyphosphate) can promote gelation of cationic biopolymer molecules (e.g., chitosan). Proteins with thiol groups are capable of forming covalent linkages through thiol–disulfide interchanges, which help to strengthen and enhance the stability of gels. The tendency for a biopolymer to form a gel under certain conditions, and the physical properties of the gel formed, depend on a delicate balance of various kinds of biopolymer–biopolymer, biopolymer–solvent, and solvent–solvent interactions. The properties of food emulsions that have a gelled aqueous phase are dependent on the nature of the interactions between the emulsifier adsorbed to the surface of the droplets and the biopolymer molecules in the gel network (Dickinson 2012). If there is a strong attractive interaction between the droplet surfaces and the gel network, then the network is reinforced and a strong gel is formed. On the other hand, 159 Emulsion Ingredients TABLE 4.8 Summary of Molecular and Functional Properties of Thickening and Gelling Agents Commonly Used in Food Emulsion Name Structure Solubility Function Aggregation Mechanism Notes Carrageenan κ, ι, λ Linear Anionic 200–400 kDa Hot water Cold water Thickening Gelling Helix association Cold-set Thermoreversible Not acid stable Linear Nonionic* 80–140 kDa Hot water Thickening Gelling Helix association Cold-set Thermoreversible* Linear Anionic 32–200 kDa Hot water Low Ca2+ Thickening Gelling Ca2+ Cold-set Thermoreversible Partly acid stable Multivalent ions should be added slowly Linear Anionic 5–150 kDa Linear Anionic 5–150 kDa Hot water Cold water (Low Ca2+) Hot water Cold water (Low Ca2+) Thickening Gelling Ca2+ Cold-set Thermoreversible Acid + sugar Cold-set Thermoirreversible Acid stable, degrade on heating at pH > 5 Linear Nonionic Linear Nonionic Cold water Hot water Hot water Thickening Linear Anionic ∼2500 kDa Agar Alginate Pectin LM HM Thickening Gelling Acid stable, degrade on heating at pH > 5 Seed gums Guar gum LBG Poor acid stab. Thickening Gelling Helix association Freeze-set Thermoirreversible Poor acid stab. Cold water Hot water Thickening Gelling Helix association Cold-set Thermoreversible Acid, alkali, heat, and freeze–thaw stable Linear Anionic Hot water Cold water (Low divalent) Thickening Gelling Helix association + salt Cold setting Thermoreversible* Poor acid stability Transparent gels *Gels formed in presence of multivalent ions may be irreversible Native Granules Nonionic Hot water Thickening Gelling Opaque Modified Linear/ branched Nonionic Cold water Hold water Thickening Gelling Granule swelling Heat-set Irreversible Helix association Cold-set Reversible Xanthan Gellan gum Starch A variety of modified starches are available for different applications (Continued) 160 Food Emulsions: Principles, Practices, and Techniques TABLE 4.8 (Continued) Summary of Molecular and Functional Properties of Thickening and Gelling Agents Commonly Used in Food Emulsion Name Structure Solubility Function Aggregation Mechanism Notes Cellulose derivatives MC MHPC Linear Nonionic Cold water Thickening Gelling HPC Linear Nonionic Cold water Thickening CMC Linear Anionic Microcrystals MCC Insoluble Dehydration Heat-set Reversible Tgel ∼50°C–90°C Precipitates Tppt ∼40°C–45°C Acid and base Heating Freeze-thaw Acid and base Heating Freeze-thaw Thickening Gelling Thickening Gelling Salt bridges Particle gel Acid and base Heating Freeze-thaw Gelatin Linear Amphoteric Amphiphilic Cold water Thickening Gelling Helix formation Cold-set Thermoreversible Transparent gels Linear Amphoteric Amphiphilic Cold water Warm water Thickening Gelling Rennet IEP precipitation Ca2+ Alcohol Opaque gels Linear Amphoteric Amphiphilic Cold water Warm water Thickening Gelling Hydrophobic Heat-set Thermoirreversible Transparent or opaque gels depending on pH and salt Casein Globular proteins Note: Biopolymers whose functional properties are influenced by Ca2+ ions may also be influenced by the presence of other types of multivalent cations. It should be noted that many of the biopolymers mentioned below come in different forms that may have appreciably different functional properties than those mentioned here. * Depends on agarose to agaropectin ratio. if the droplet surface does not interact favorably with the gel network then the droplets may disrupt the network and weaken the gel strength. The magnitude of this effect depends on the size of the emulsion droplets. The larger the droplets compared to the pore size of the gel network, the greater the disruptive effect. Some of the components in food emulsions may also influence the formation and properties of biopolymer gels. Certain types of emulsifiers interact with gelling biopolymers and alter their thermal transition temperatures and gel strengths. 4.5.3 Commonly Used Texture Modifiers A variety of substances have the molecular characteristics required to make them suitable as thickening or gelling agents for use in food emulsions (Table 4.8). The most commonly used texture modifiers are biopolymers (polysaccharides and proteins) that are added to the aqueous phase of oil-in-water emulsions.* A brief overview of some of the biopolymers most commonly used as * In water-in-oil emulsions, such as margarine and butter, fat crystals play the role of texture modifiers by forming a threedimensional network of aggregated crystals. 161 Emulsion Ingredients 60 40 Lactoferrin Chitosan ζ-potential (mV) 20 0 3 4 5 6 8 7 9 –20 Pectin –40 –60 β-Lg Alginate –80 pH FIGURE 4.32 Biopolymers, such as proteins and ionic polysaccharides, exhibit a range of different electrical characteristics, which can be characterized by their ζ-potential versus pH profiles. texture modifiers in food emulsions is given in this section, and more detail is given elsewhere (Nussinovitch 1997, Walstra 2003, Williams and Phillips 2003, Cui 2005, Stephen et al. 2006, Kasapis et al. 2009, Brady 2013). 4.5.3.1 Polysaccharides 4.5.3.1.1 Carrageenans Carrageenans are natural hydrocolloids extracted from certain species of red seaweed. They are linear sulfated polysaccharides consisting of alternating β(1–3)- and α(1–4)-linked galactose residues. There are three major types of carrageenan, which primarily differ in the number and position of sulfate ester groups on the galactose residues: kappa (κ), iota (ι), and lambda (λ). These differences in primary structure have a large influence on the functional characteristics of the different carrageenans, for example, solubility, thickening, gelation, environmental sensitivity, and ingredient compatibility. λ-carrageenan is commonly used as a thickening agent, whereas κ- and ι-carrageenans are usually used as cold-setting reversible gelling agents. Carrageenan ingredients come in a variety of different forms with different functional attributes, for example, molecular weights, salts, and blends. Typically, they are sold as salts (Na, K, and Ca) and have number average molecular weights between 200 and 400 kDa. Carrageenans usually have a random coil conformation at relatively high temperatures, but undergo a helical-to-coil transition when they are cooled below a transition temperature (∼30°C to 70°C). The transition temperature depends on carrageenan structure, salt type and concentration, and the presence of sugars. In the presence of sufficiently high quantities of salt, helical regions of gelling carrageenans (κ and ι) can associate with each other to form hydrogen bonded junction zones that promote gel formation. Knowledge of the transition temperature is important when utilizing carrageenans in foods since it determines the temperature above which they must be heated to adequately disperse and solubilize them in water, and the temperature they must be cooled below to form gels. Carrageenan is widely used in food emulsions such as milk shakes, coffee creamers, ice cream, and desserts. However, in many of these applications it is used as a stabilizer (by forming a protective coating around oil droplets), rather than a gelling agent. 162 Food Emulsions: Principles, Practices, and Techniques Carrageenan is often used in blends with other polysaccharides (e.g., LBG, konjac, or starch) to improve functional characteristics such as water-holding capacity, thickening, and gelation. Negatively charged carrageenan molecules may also interact with positively charged groups on proteins under certain circumstances, for example, pH, ionic strength, and temperature. These interactions have been utilized to improve the stabilizing, thickening, gelling, and water-holding properties of various food products. 4.5.3.1.2 Agars Agars are a group of natural hydrocolloids extracted from certain species of red seaweed. They are linear polysaccharides consisting primarily of alternating β(1–3)- and α(1–4)-linked galactose units. Different agars vary in the number and type of substituents (e.g., sulfate, pyruvate, urinate, or methoxyl) on the hydroxyl groups of the sugar residues and in the fraction of the α(1–4)-linked galactose units that are present in the 3–6 anhydride form. Agar can be roughly divided into two fractions: agarose, a nonionic polysaccharide that gels; and agaropectin, a slightly negatively charged polysaccharide that does not gel. The negatively charged fraction contains anionic substituents (usually sulfates) along its backbone. Commercial agars vary in the relative proportions of the nonionic and ionic fractions present. Typically, the mean weight average MW of agars is between 80 and 140 kDa, but they are usually highly polydisperse. Agars usually require heating in aqueous solutions to dissolve them. When the system is cooled it forms a viscous solution, which gels over time without the need for specific additives (e.g., multivalent ions or sugars). Agars are unusual in that their gelation temperatures upon cooling (30°C–40°C) are usually considerably lower than their melting temperatures upon heating (85°C–95°C). The gelation mechanism has been attributed to the transition of an appreciable part of the agar molecules from a random coil to a helical structure upon cooling, and subsequent association of the helical regions to form junction zones that are separated by fairly irregular flexible chain regions. Agars form thermo-reversible cold-set gels. 4.5.3.1.3 Alginates Alginates are natural hydrocolloids usually extracted from certain species of brown seaweed. Alginates are linear copolymers of D-mannuronic acid (M) and l-guluronic acids (G), which can be distributed as blocks of M, blocks of G, or blocks of alternating M and G residues. The M-blocks tend to have a flexible conformation, the G-blocks tend to have a relatively inflexible conformation, and the MG blocks tend to have an intermediate flexibility between these two extremes. Alginates vary in their molecular weights (typically between 32 and 200 kDa) and in the proportions and distributions of the M and G groups along the chain, which leads to appreciable differences in their functional characteristics. The alginic acid extracted from brown seaweed is usually reacted with bases to produce sodium, potassium, calcium, or ammonium alginate salts. Alternatively, it can be reacted with propylene oxide to produce propylene glycol alginate (PGA), in which partial esterification of the carboxylic acid groups on the uronic acid residues occurs. The monovalent salts of alginate tend to have good water solubility, whereas alginic acid and multivalent salts of alginate tend to have fairly poor water solubility and form paste-like materials. Often special care is needed to adequately disperse and dissolve alginates when preparing them for use in food products. In the absence of multivalent ions, alginate tends to form viscous solutions since there is little intermolecular cross-linking. Conversely, in the presence of multivalent cations alginates tend to form cold-set thermo-irreversible gels because the positively charged ions form electrostatic bridges between negatively charged polysaccharides. The junction zones are believed to be between relatively stiff G-block regions on different alginate molecules. The gelation characteristics of a particular alginate are therefore strongly dependent on the number and length of the G-blocks. Alginates have been used as thickening agents, gelling agents, and stabilizers in a variety of food emulsions. For example, they have been used as thickening agents in ice cream, soups, sauces, dressings, mayonnaise, and beverages and as gelling agents in desserts and whipped cream. There functional attributes are primarily due to their texture modifying characteristics, but there may also be additional contributions arising from their interactions with other components, for example, other polysaccharides, Emulsion Ingredients 163 proteins, and fat droplets. PGA is widely used as a stabilizer and thickening agent in food emulsions, such as dressings and fruit beverages. 4.5.3.1.4 Pectins Pectins are natural hydrocolloids found in the cell walls and intercellular regions of high plants. Most commercial pectins used in the food industry are extracted from citrus or apple pomace and sold as powders. The term “pectin” actually refers to a broad range of different molecular species. In general, pectin molecules tend to be comprised of “smooth” linear regions consisting of α(1–4) linked d-galacturonic acids separated by “hairy” branched regions consisting of various sugars. The galacturonic acid groups may be partly esterified by methyl groups and partly neutralized by bases. The fraction of esterified galacturonic groups is one of the main factors influencing the functional characteristics of commercial pectins. Pectins are usually classified as either high methoxyl (HM) or low methoxyl (LM) pectins depending on whether their degree of methylation (DM) is greater or less than 50%, respectively. HM pectins form gels under acidic conditions at high sugar contents, which is attributed to the reduction of electrostatic repulsion between the chains at low pH and the increased osmotic attraction at high sugar contents. Gels formed by HM pectins are thermo-irreversible cold-setting gels. The junction zones are believed to be hydrogen bonds and hydrophobic attraction between helical regions formed in the linear smooth regions of the molecules. LM pectins form gels in the presence of calcium, which is attributed to the ability of the positively charged calcium ions to form electrostatic bridges between the linear smooth regions of the negatively charged pectin molecules. Gels formed by LM pectins are thermo-reversible cold-setting gels. The precise gelation characteristics of a particular pectin depend on its molecular structure (e.g., DE, amidation, molecular weight, and branching) and the prevailing environmental conditions (e.g., pH, ionic strength, and sugar content). Pectins are water soluble, but usually have to be dispersed in warm water prior to use to ensure proper dissolution. Pectins are relatively stable to heating at low pH (3–5), but may degrade due to hydrolysis at higher or lower pH values, with the effects being more pronounced the higher the DM. Typically, the average molecular weight of pectins is between 50 and 150 kDa. The viscosity of pectin solutions depends on the concentration and type of pectin used, as well as solution conditions such as pH, ionic strength, and temperature. Typically, the pKa value of the acid groups on pectin is around 3.5, so that it starts to lose its negative charge as the pH is lowered around and below this value. Pectins are used as stabilizers, thickening agents, and gelling agents in a variety of different food emulsions, for example, drinkable yogurts, dressings, mayonnaise, beverages, and ice cream. 4.5.3.1.5 Seed Gums (Galactomannons) A number of polysaccharide texture modifiers are extracted from the seeds of various bushes, trees, and plants, for example, locust bean gum (LBG), guar gum, and tara gum. These polysaccharides are primarily linear nonionic polysaccharides known as galactomannans (∼103 kDa), which consist of β(1–4)linked d-mannose residues with single α-d-galactose residues linked to the main chain. One of the main differences between galactomannans from different sources is the degree of galactose substitution, with galactose-to-mannose ratios of 1:4.5 for LBG, 1:3 for tara gum, and 1:2 for guar gum. The galactose side chains tend to inhibit molecular associations and hence these differences in galactose content lead to differences in the functional properties of different galactomannans, for example, solubility, thickening, and gelation. For example, guar gum can be dissolved in cold water, whereas LBG and tara gum require hot water for dissolution. At ambient temperatures, galactomannans tend to exist as individual molecules in aqueous solutions because close intermolecular associations are inhibited by the presence of the galactose substituents. For this reason, seed gums are primarily used as thickening agents, rather than as gelling agents. Nevertheless, LGB has been shown to form irreversible gels upon freezing, which has been attributed to self-association of nonsubstituted regions along the mannose backbone. Galactomannan solutions tend to be highly viscous, pseudoplastic, and thixotropic, and their rheological characteristics are not strongly influenced by pH or ionic strength because they are nonionic biopolymers. Galactomannans are sensitive to thermal degradation in acidic solutions (pH < 4.5), which limits 164 Food Emulsions: Principles, Practices, and Techniques their application in some foods. Guar gum and LBG are widely used as thickening agents in food emulsions, such as dressings, mayonnaise, sauces, and deserts. Seed gum functional properties are often improved by using them in combination with other kinds of polysaccharide, for example, xanthan or carrageenan. 4.5.3.1.6 Tree Gum Exudates A variety of polysaccharides can be extracted from the exudates of certain trees, for example, gum arabic, gum tragacanth, and gum karaya. Gum arabic is the most widely used tree gum exudates in the food industry, but it is mainly used as an emulsifier in beverage emulsions (Section 4.4.2.3). Gum tragacanth is an exudate collected from the shrubs of the Astragalus species. It is a complex heterogeneous polysaccharide with a high molecular weight that has protein moieties attached. It contains a variety of different types of sugars and is acidic. It is used in foods to provide high viscosity and pseudoplastic properties. It has also been reported to be surface active and capable of stabilizing emulsions. Gum tragacanth has good stability in acidic conditions, which makes it suitable for application in salad dressings and other low pH products. 4.5.3.1.7 Xanthan Gum Xanthan gum is the trivial name given to extracellular polysaccharides secreted by bacteria of the genus Xanthanomonas. Generally, xanthan gum ingredients used in the food industry are relatively high molecular weight molecular polysaccharides that are produced commercially from Xanthomonas campestris. The primary structure of xanthan gum consists of a β-(1–4)-d-glucose backbone that is substituted with trisaccharide side chains at the C-3 positions of alternate glucose residues. The trisaccharide chains usually consist of mannose—glucuronic acid—mannose, with a relatively high proportion of the terminal mannose units containing either pyruvate or acetate residues. Consequently, the side chains of the xanthan molecules tend to have an appreciable negative charge. In aqueous solutions at relatively low temperatures, xanthan gum is believed to exist as stiff extended molecules with a largely helical structure, but at higher temperatures it exists as more random coil molecules. The helix-coil transition temperature is highly sensitive to ionic strength, and may range from around 40°C to > 90°C. Under appropriate solution conditions, helical regions on different xanthan molecules may associate with each other, which leads to the formation of a weak gel. Xanthan gum ingredients are readily soluble in both hot and cold water and are stable over a wide range of solution and environmental conditions, for example, pH, ionic strength, heating, freeze–thaw cycling, and mixing. Xanthan gum ingredients come in a range of molecular weights, typically around 1000 kDa. Xanthan gum forms highly viscous solutions at relatively low concentrations because it is a fairly stiff molecule that is highly extended in aqueous solutions. In addition, xanthan gum solutions exhibit pronounced reversible shear-thinning behavior, for example, the viscosity of a 0.5% solution has been shown to decrease by over three orders of magnitude from low to high applied shear rates. At high salt concentrations, the rheology of xanthan gum solutions is relatively insensitive to temperature. The unique rheological characteristics of xanthan gum solutions are widely utilized in the formulation of food emulsions such as dressings, sauces, beverages, deserts, and cake batters. Xanthan can interact synergistically with a variety of other polysaccharides, leading to improved viscosity or gelation characteristics. In particular, xanthan gum is often used in food emulsions in conjunction with galactomannans, such as guar gum and LBG. The xanthan gum–galactomannan combination can be used to provide a rheological profile (viscosity versus shear stress) that gives better emulsion stability, texture, and mouthfeel than xanthan gum alone. Xanthan gum also has a synergistic interaction with galactomannans, leading to the formation of thermo-reversible gels. 4.5.3.1.8 Gellan Gum Gellan gum is an extracellular polysaccharide produced commercially as a fermentation product of the bacterium Pseudomonas elodea. It is a linear anionic hetero-polysaccharide with a molecular weight of approximately 500 kDa. The linear chain consists of a repeating unit of four saccharides, glucose, glucuronic acid, glucose, and rhamnose. In nature, there are approximately 1.5 substituents Emulsion Ingredients 165 per repeating unit, comprising mainly of glycerate or acetate. These substituents hinder intermolecular association and therefore influence the gelling characteristics of gellan gums. Two forms of gellan gum are commonly produced commercially that have different functional properties: a low-acylated form that produces strong nonelastic brittle gels and a high-acylated form that produces soft elastic nonbrittle gels. Gellan gums can be dissolved at ambient temperatures provided significant amounts of divalent ions are not present, otherwise they have to be heated. They give solutions that are highly viscous and pseudoplastic. The solution viscosity decreases steeply with increasing temperature due to a reversible helixto-coil transition that occurs upon heating (around 25°C–50°C). Gellan gums have good heat stability at neutral pH, but are susceptible to thermal degradation under acidic conditions. They form gels when cooled from high temperatures due to the formation of helical regions that can associate with each other and form junction zones. Since they are electrically charged their thickening and gelling properties are highly sensitive to salt type and concentration. Divalent ions usually promote gelation by forming salt bridges between negatively charged helical regions. Gels formed in the presence of monovalent ions are usually thermo-reversible, whereas those formed in the presence of multivalent ions may be thermoirreversible. A variety of gel characteristics can be achieved by altering the degree of esterification of the gellan gum and the mineral composition. Gellan gums can be used in food emulsions as thickening or gelling agents. 4.5.3.1.9 Starch and Its Derivatives Starch is one of the most abundant naturally occurring polysaccharides, being found in the roots, stems, seeds, and fruits of all green leaf plants (Damodaran et al. 2007). Starch is extracted from a wide variety of sources with the most common being corn, potato, wheat, tapioca, and rice. There are two main fractions in starch: amylose and amylopectin. Amylose is primarily a linear chain (MW ∼ 106) of α-d-(1–4)-linked glucose units, although there may be a limited number (<0.5%) of α-d-(1–6)-linked branches. Amylopectin is a very large (MW = 107 − 5 × 108) highly branched molecule also consisting primarily of α-d-(1–4)-linked glucose units, but with a much higher fraction of (∼5%) of α-d-(1–6)linked branches. Natural sources of starch vary appreciably in the ratio of amylose to amylopectin, which partially accounts for differences in their functional characteristics. In nature, amylose and amylopectin are organized into complex biological structures within starch granules that consist of crystalline regions separated by amorphous regions. When aqueous solutions of starch granules are heated above a critical temperature, they incorporate water and the crystalline regions are disrupted. The resultant swelling of the starch granules leads to an appreciable increase in their effective volume fraction, thereby leading to an increase in solution viscosity (“gelatinization”). Upon further heating, a fraction of the starch leaches out of the granules and there is a subsequent decrease in viscosity. When the solution is cooled, linear regions of starch molecules associate with each other though hydrogel bonding (“retrogradation”) and there may be an increase in viscosity or gelation. The rheological characteristics of a particular native starch depend on the structural organization of the molecules within the starch granule, the ratio of amylose to amylopectin, the precise molecular characteristics of each of these fractions, the solution composition (e.g., pH, ionic strength, and sugar content), and environmental factors (e.g., shearing, temperature, and pressure). The gels formed by native starch often have limited application in the food industry, because they do not have the desired solubility, textural, or stability characteristics. For this reason, starches are often physically, chemically, or enzymatically modified to improve their functional properties, for example, pregelatinization, limited hydrolysis, addition of side groups (polar, ionic, or hydrophobic) or crosslinking. Starch ingredients are currently available that are soluble in cold or hot water, that thicken or gel with or without heating, that exhibit a wide range of gel characteristics (e.g., opacity, strength, and water holding capacity), and that have different stabilities to environmental conditions (e.g., heating, freezing, pH, ionic strength, and shearing). These starches are used in a wide variety of different food emulsions as thickening agents, gelling agents, and stabilizers. For example, they are used in dressings, sauces, desserts, and beverages to provide desirable textural characteristics and to prevent gravitational separation of suspended matter. 166 Food Emulsions: Principles, Practices, and Techniques 4.5.3.1.10 Cellulose and Its Derivatives Cellulose is the most abundant natural polysaccharide, being the major structural component of land plants. Cellulose is a linear polymer with a relatively high molecular weight consisting of d-glucose units joined together by d-β(1–4) linkages. In its natural state, cellulose is not usually suitable for utilization as a texture modifier in processed foods because it forms strong intermolecular hydrogen bonds that make it insoluble in water. Nevertheless, it can be isolated and chemically modified in a number of ways to produce products that are useful as food ingredients. The most common cellulose derivatives used in foods are methyl cellulose (MC), carboxymethyl cellulose (CMC), hydroxypropyl cellulose (HPC), and methyl hydroxypropyl cellulose (MHPC). These ingredients consist of cellulose molecules that have been chemically modified by adding substituents (M, CM, HP, or MHP) to the cellulose backbone. These substituents provide a steric hindrance that helps prevent strong intermolecular associations between cellulose backbones. MC, MHPC, and HPC are all soluble in cold water, but tend to become insoluble when the solution is heated above a critical temperature (around 50°C–90°C). MC and MHPC both form reversible gels or highly viscous solutions upon heating, whereas HPC just precipitates out of solution. The driving force for the aggregation of these cellulose derivatives at high temperatures has been attributed to the increase in hydrophobic attraction between the molecules, favoring cellulose–cellulose interactions. MC, MHPC, and HPC are all nonionic polymers and therefore have good stability to pH and salt, as well as good compatibility with other ingredients. These products have been used as texture modifiers in a variety of food products, including dressings, sauces, creams, and deserts. CMC, also known as cellulose gum, is an anionic linear polymer, which is manufactured by chemically attaching carboxymethyl groups to the backbone of native cellulose. It is normally sold in the form of either sodium or calcium salts and is available in different molecular weights and degrees of substitution (DS). At a sufficiently high DS (‰0.4), CMC is readily soluble in water and forms viscous solutions. Because CMC is ionic, the viscosity of these solutions is sensitive to pH and ionic strength, as well as to the presence of other types of electrically charged molecules. CMC can form gels in the presence of multivalent ions due to electrostatic screening and bridging effects. CMC is an odorless and tasteless ingredient that is commonly used in foods and beverages to prevent gravitational separation of suspended particles and to create desirable textural attributes and mouthfeel, for example, deserts, dressings, sauces, bakery products, and beverages. Another commonly used cellulose-based product in the food industry is microcrystalline cellulose (MCC). This product is manufactured by treating native cellulose with hydrochloric acid to dissolve the amorphous regions leaving crystalline regions as colloidal sized particles. MCC is water-insoluble and so exists as small colloidal particles that are predominately dispersed in the aqueous phase. In aqueous solutions, MCC can form three-dimensional matrices of aggregated particles that form viscous solutions or gels depending on the concentration used. These solutions are pseudoplastic and thixotropic because the particle network breaks down upon application of shear forces, but the viscosity or gel strength is regained once the shearing stress is removed. MCC functions over a wide range of temperatures, providing freeze/thaw and heat stability to many food products. This product is dispersible in water at relatively high pH (>3.8), but may need addition of protective hydrocolloids to disperse it at lower pH values. MCC may also be advantageous in the formulation of low-fat products because it provides a creamy mouthfeel and opacity due to light scattering. MCC is used in a variety of food emulsions to improve emulsion stability and provide desirable textural attributes, including soups, sauces, meat products, dressings, and beverages. 4.5.3.2 Proteins 4.5.3.2.1 Gelatin Gelatin is a relatively high molecular weight protein derived from animal collagen, for example, pig, cow, or fish. Gelatin is prepared by hydrolyzing collagen by boiling in the presence of acid (Type A gelatin) or alkaline (Type B gelatin). The isoelectric point of Type A gelatin (∼7 to 9) tends be higher than that of type B gelatin (∼5). Type A gelatin is therefore quite unusual because it is positively charged Emulsion Ingredients 167 over the entire pH range typically found in foods. Gelatin exists as a random coil molecule at relatively high temperatures, but undergoes a helix-to-coil transition upon cooling, which is at about 10°C–30°C for mammalian gelatin and at about 0°C–5°C for fish gelatin. Gelatin forms a thermo-reversible cold-set gel upon cooling below the coil–helix transition temperature due to formation of helical junction zones between segments of two or three gelatin molecules. Gelatins are used in a number of emulsion-based food products as thickening agents and gelling agents, including deserts, beverages, soups, sauces, and dairy emulsions. 4.5.3.2.2 Caseins As mentioned earlier, casein is a complex mixture of different proteins usually derived from bovine milk by acid or enzyme precipitation (Section 4.4.2.3). The ability of casein to act as a texture modifier is mainly determined by the ability of the casein molecules to associate with each other under suitable conditions. Caseins have significant fractions of nonpolar regions along their polypeptide chains, which favor self-association through hydrophobic interactions. They also have a relatively high amount of negatively charged phosphoseryl residues, which favors self-association through electrostatic bridge formation by multivalent cations, such as Ca2+. More generally, the self-association of casein is strongly influenced by electrostatic interactions between the molecules and is therefore sensitive to pH and ionic strength. Casein molecules can be made to aggregate in a variety of ways to form viscous solutions or gels, for example, addition of ethanol, addition of rennet, or pH adjustment to the isoelectric point. Casein ingredients are available in a variety of different powdered forms for utilization in food products, for example, whole casein or sodium, potassium, or calcium caseinate. Caseins are used in a wide variety of food emulsions as thickening and gelling agents, with the most important being yogurt and cheese. 4.5.3.2.3 Globular Proteins A number of texture modifiers used in food emulsions are based on the utilization of globular proteins extracted from a variety of sources, for example, whey, eggs, and soy. These proteins tend to be fairly water soluble at ambient temperatures, providing the pH is sufficiently far from their isoelectric point. Nevertheless, they can thicken solutions or form gel when they are heated above a temperature where the globular proteins unfold (typically 60°C–80°C). Protein unfolding exposes reactive amino acid side groups that are normally buried in the globular proteins hydrophobic interior, such as nonpolar or sulfhydryl groups. Exposure of these groups promotes intermolecular interactions through hydrophobic attraction and disulfide bond formation. Gelation is particularly sensitive to the magnitude and range of the electrostatic interactions between protein molecules, so that gel characteristics are strongly dependent on pH and ionic strength. A range of different gel types can be produced by varying pH, ionic strength, and heating conditions, for example, brittle vs. rubbery, strong vs. weak, transparent vs. opaque, and good vs. bad water holding capacity. The heat-set gels formed by globular proteins tend to be irreversible, that is, when the gels are cooled they do not melt. 4.5.3.3 Biopolymer Blends Biopolymers are often used in combination with other biopolymers, rather than in isolation, to form systems with novel structures and rheological properties (Tolstoguzov 2002, 2003, Turgeon et al. 2007, Dickinson 2011, Schmitt and Turgeon 2011). When two different biopolymers are mixed together, they may either form a one-phase or a two-phase system depending on the nature of the biopolymers involved, the solution composition and the prevailing environmental conditions (Figure 4.33). In a one-phase system, the two biopolymers can exist either as individual molecules or as soluble molecular complexes that are evenly distributed throughout the system, so that the solution composition is the same at every location. In a two-phase system, the solution separates into two distinct phases that have different biopolymer compositions. Phase separation can occur through two different physicochemical mechanisms: complex coacervation or thermodynamic incompatibility. 168 Food Emulsions: Principles, Practices, and Techniques = Protein (a) (b) = Polysaccharide = Complex (c) (d) FIGURE 4.33 Schematic representation of organization of biopolymer molecules in a mixed biopolymer system. The biopolymer solution may form one or two phases, containing aggregated or nonaggregated biopolymer molecules. (a) One-phase individual molecules, (b) one-phase molecular complexes, (c) two-phase incompatibility, and (d) two-phase coacervation. Complex coacervation: Phase separation occurs due to complex coacervation when the two biopolymers associate with each other through relatively strong attractive interactions, for example, when they have opposite electrical charges. The resulting two-phase system consists of an insoluble phase that is rich in both biopolymers, and an aqueous phase that is depleted in both biopolymers (Figure 4.33). Thermodynamic incompatibility: Phase separation occurs due to thermodynamic incompatibility when the free energy of mixing of the biopolymers is positive, which is common when biopolymers have different molecular conformations, dimensions, rigidities, or solvent affinities. This type of phase separation often occurs when one or both of the biopolymers are uncharged, or when both biopolymers have similar electrical charges. At sufficiently low biopolymer concentrations, the two biopolymers are intimately mixed and form a one-phase solution, but once the biopolymer concentration exceeds a certain level phase separation occurs and a two-phase solution is formed with one of the phases being rich in one type of biopolymer and depleted in the other type, and vice versa (Figure 4.33). The behavior of biopolymer blends under different solution and environmental conditions can be conveniently characterized in terms of phase diagrams (Tolstoguzov 2002). For example, a typical phase diagram for a mixed biopolymer system that undergoes phase separation due to thermodynamic incompatibility is shown in Figure 4.34. These phase diagrams can often be used to optimize the biopolymer composition required to produce a solution with a particular microstructure and physicochemical properties. Once a particular microstructure has been formed by phase separation of a mixed biopolymer solution, it is often possible to trap the system in a kinetically stable state and thus create novel food microstructures and rheological properties (Norton and Frith 2001). For example, kinetic trapping can be achieved by changing solution or environmental conditions so that one or both of the phases thickens or gels, for example, by changing temperature, pH, ionic composition, or solvent quality. If this process is carried out in the presence of shear forces, it is possible to produce a wide variety of different microstructures, for example, spheres, tear-drops, and fibers. 169 Emulsion Ingredients W1/W2 W1 Biopolymer 2 Bicontinuous Two phase W2/W1 One phase W2 Biopolymer 1 FIGURE 4.34 The structural organization of a mixed biopolymer system that undergoes phase separation due to thermodynamic incompatibility can be described by a phase diagram. Different types of gel microstructure can be created using biopolymer blends by varying the nature of the biopolymers involved, the solution composition, and the prevailing environmental conditions, for example, interpenetrating networks comprised of different biopolymers, a single network that incorporates both types of biopolymer, or a “filled gel” consisting of regions rich in one biopolymer dispersed in regions rich in the other biopolymer. Each of these microstructures will have unique rheological and physicochemical properties, for example, gel strength, gelation rate, gelation temperature, water holding capacity, and opacity. Many food scientists are currently attempting to understand the fundamental processes involved in the formation of structured biopolymer blends and in utilizing these systems to create foods with novel or improved physicochemical, sensory, or delivery properties. In particular, mixed biopolymer systems appear to be an effective means of creating low-fat products with similar properties to high-fat products, for example, deserts, yogurts, dressings, and spreads. 4.5.4 Selection of an Appropriate Texture Modifier There are a large number of different types of food ingredient that can be used by food manufacturers to modify the texture of their products. The choice of a particular type of ingredient or combination of ingredients depends on a number of physicochemical, legal, economic, and marketing factors (see Section 4.7). In this section, we focus on the rheological and other physicochemical aspects influencing the selection of texture modifiers for use in food emulsions. Initially, a food manufacturer should stipulate the physicochemical and sensory properties that are desired for the particular product of interest. Some of the factors that might be considered are listed below: • Should the product be capable of passing through a homogenizer, flowing through a pipe, being stirred or being packaged into a container during the manufacturing process? • Should the product be capable of pouring easily from a container during its utilization by a consumer? • Are there special textural requirements that are desirable in the final product, for example, cling, spreadability, and stirability? • Should the final product be a low viscosity liquid, a highly viscous liquid, a paste, a gel, or a solid? 170 Food Emulsions: Principles, Practices, and Techniques • What kind of mouthfeel is desirable in the final product, for example, “watery,” “creamy,” “smooth,” and “thick”? • Is the texture modifier going to be used primarily to modify the texture of the product or to prevent gravitational separation of droplets or other particulate matter? • Should the texture modifier produce a transparent, translucent, or optically opaque solution? • Is it necessary for the texture modifier to have good freeze–thaw, thermal, or acid stability? • Should the desirable textural properties of the system only manifest themselves after the food has been processed in a certain way, for example, chilling and cooking? • Is it important to control the gastrointestinal fate of the food after ingestion? After considering these factors, the manufacturer should establish certain measurable parameters that can be used to define the rheological (and other physicochemical) characteristics of the product, such as a viscosity versus shear stress profile, a yield stress, a modulus, a breaking stress or strain, and a texture versus temperature profile (Chapter 8). The manufacturer should then specify the optimum rheological characteristics desired for an acceptable product, which often involves correlating the results of rheological tests made on the product with sensory measurements made on the same product. Once the optimum rheological characteristics of the product have been specified, a food manufacturer can then experiment with different types and concentrations of texture modifiers within the food to determine the ingredient(s) that provides the desired functional characteristics. 4.6 Other Food Additives Food emulsions also contain a variety of other ingredients that contribute to their stability, taste, texture, and appearance, such as acidulants, preservatives, flavorings, colorings, vitamins, minerals, and antioxidants (Damodaran et al. 2007, Igoe 2011, Smith and Hong-Shum 2011). In this section, a brief overview of the most important of these food additives will be presented. 4.6.1 pH Control The pH of the aqueous phase plays an extremely important role in determining the physicochemical, microbiological, and organoleptic properties of food emulsions. The pH of the majority of food emulsions lies within the range 2.5 (e.g., beverage emulsions) to 7.5 (e.g., infant formulations). The pH of the aqueous phase can be adjusted by adding organic or inorganic acids or bases. The pH can be lowered by adding organic or inorganic acids, such as acetic, lactic, citric, malic, fumaric, succinic, or phosphoric acids. It can also be lowered by adding bacteria (streptococci lactobacilli) or enzymes (δ-gluconolactone) to a food to promote biochemical reactions that lead to acid production. The pH can be increased by adding various types of organic and inorganic salts, such as phosphate, citrate, carbonate, bicarbonate, oxide, and hydroxide salts. The pH of an aqueous solution can be stabilized at a particular value using an appropriate buffering system. There may be some functional ingredients present within a food emulsion that were originally added for a different purpose, but which also have a significant buffering capacity, for example, proteins. Alternatively, specific ingredients can be added to emulsions as buffering agents, for example, weak organic or inorganic acids in combination with salts. The type of buffering system used depends on the pH of the food. For example, the effective buffering ranges of some commonly used buffering systems are: pH 2.1–4.7 for citric acid–sodium citrate; pH 3.6–5.6 for acetic acid–sodium acetate; pH 2.0–3.0, pH 5.5–7.5, and pH 10–12 for the three ortho- and pyrophosphate anions, respectively. 4.6.2 Minerals Many minerals are essential for the maintenance of human health, as well as making an important contribution to the physicochemical and sensory properties of foods. The minerals in foods may exist Emulsion Ingredients 171 in a variety of different forms, including free ions, complexes, and compounds, depending on their type and the environmental conditions, for example, pH, ionic strength, temperature, and solution composition. The solubility of the minerals in the aqueous and oil phases can vary considerably depending on the form they exist in, which has important consequences for their functional properties in foods. For example, a chelated form of a mineral may act very differently than the nonchelated form. It is therefore often important for food manufacturers to control the form that the minerals are present within a food. There are currently deficiencies in the consumption of certain minerals that are essential for the maintenance of good health, for example, calcium, iron, selenium, and zinc. Consequently, many food manufacturers are fortifying their foods with these minerals. On the other hand, overconsumption of other minerals (e.g., Na+) has been linked to adverse health effects, such as hypertension. For this reason, food manufacturers are developing effective strategies to reduce the levels or completely remove these types of mineral from foods. It should be noted that changing the mineral composition of food emulsions to improve their nutritional aspects may cause undesirable changes in their physicochemical and sensory properties. High concentrations of minerals can have an adverse affect on the aggregation stability of oil-inwater emulsions containing electrostatically stabilized droplets due to electrostatic screening and ion binding effects (Chapters 3 and 7). These effects can occur at relatively low mineral concentrations (<5 mM) when multivalent counter-ions are present, for example, Ca2+ in an emulsion containing negatively charged droplets. Certain mineral ions may also promote undesirable chemical reactions that lead to product deterioration, for example, iron and copper ions can promote lipid oxidation. In these systems, it is usually necessary to add chelating agents to sequester the mineral ions and prevent them from causing chemical instability. Certain types of minerals also influence the functional properties of other food ingredients. For example, the ability of many biopolymers to thicken or gel a solution is depends on the type and concentration of mineral ions present. Careful selection and control of the mineral ions present in food emulsions is therefore important when formulating a successful product. 4.6.3 Chelating Agents Chelating agents are often added to foods to sequester multivalent mineral ions. Chelation of mineral ions can have a number of beneficial functions in food emulsions, including improving the solubility of mineral ions, inhibiting lipid oxidation, retarding color or flavor loss, and preventing aggregation of charged droplets. Many of the most effective chelating agents currently used in food emulsions are synthetic, for example, EDTA, phosphoric acid, and polyphosphates. Nevertheless, there is some concern about the use of synthetic chelating agents because they are believed to bind minerals so strongly that they may not be bioavailable and because consumers do not regard them as “label friendly.” Natural chelating agents, such as citric acid, can be used to sequester minerals, but they tend to be less effective and have limited use in many foods because of their flavor, solubility, and/or requirement for acid environments. Research is therefore being carried out to identify alternative natural chelating agents that can be used in a wider range of food applications. A variety of proteins, protein hydrolysates and polysaccharides have been shown to be effective at chelating transition metals. It is important to ensure that the chelating system chosen is effective under the solution conditions in the product (e.g., pH, ionic composition, and temperature), and that it does not adversely affect the functional properties of other food ingredients. 4.6.4 Antioxidants The oxidation of lipids is one of the most important chemical reactions that occur in food emulsions that causes deterioration in product quality (McClements and Decker 2000, Waraho et al. 2011). Lipid oxidation can lead to the production of an off-flavor, loss of beneficial polyunsaturated lipids, and formation of potentially toxic reaction products. “Lipid oxidation” is a general term that is used to describe a complex sequence of chemical changes that result from the interaction of lipids with oxygen active species. The precise mechanism of lipid oxidation in a particular food depends on the nature of the reactive species present and their physicochemical environment. Lipid oxidation can be conveniently divided into three 172 Food Emulsions: Principles, Practices, and Techniques distinct stages: initiation, propagation, and termination. One of the most effective means of retarding lipid oxidation in fatty foods is to incorporate antioxidants. Antioxidants work by a variety of different methods, including control of oxidation substrates (e.g., oxygen and lipids), control of prooxidants (e.g., reactive oxygen species and prooxidant metals), and inactivation of free radicals. Antioxidants can be broadly divided into two categories depending on the mechanism by which they operate: primary antioxidants and secondary antioxidants. Primary antioxidants retard lipid oxidation because they are capable of accepting free radicals, thereby retarding the initiation step or interrupting the propagation step. The effectiveness of these antioxidants depends on their chemical structure, solution conditions (pH, ionic strength, and temperature), and physicochemical environment (oil, water, or interfacial region). Antioxidants that are effective at retarding lipid oxidation in bulk oils may not be as effective in emulsions due to differences in their location relative to lipid substrates or prooxidants. Many synthetic food additives, such as BHA, BHT, and TBHQ, are common chain-breaking antioxidants used in food systems. These synthetic antioxidants are often highly effective at controlling lipid oxidation; however, consumer demand for all natural foods has prompted the food industry to look for more “label friendly” alternatives. For this reason, a number of studies have been carried out to assess the effectiveness of natural chain-breaking antioxidants in bulk oils and emulsions, including tocopherols, fruit extracts, and plant extracts. Secondary antioxidants can retard lipid oxidation through a variety of mechanisms, including chelation of transition metals, replenishing of hydrogen to primary antioxidants, oxygen scavenging, and deactivation of reactive species. It should be noted that none of these mechanisms involves conversion of free radical species to more stable products. From the standpoint of oil-in-water emulsions, the most important type of secondary antioxidants are those that chelate transition metal ions. The presence of transition metals, such as iron or copper, in the aqueous phase of oil-in-water emulsions has been shown to be a major factor in the promotion of lipid oxidation. The effectiveness of transition metals at promoting lipid oxidation increases dramatically when they are located near droplet surfaces because they are then in closer proximity to the lipid substrate. Consequently, any aqueous phase component that chelates transition metals and removes them from the vicinity of the droplet surface would be expected to retard lipid oxidation. A variety of synthetic and natural chelating agents are available as additives to prevent lipid oxidation in foods, for example, EDTA, phosphoric acid, polyphosphates, citric acid, other organic acids, proteins, and polysaccharides. The choice of a particular chelating agent depends on the specific food type. There are a number of means of retarding oxidation of emulsified oils, which are not available for retarding oxidation of bulk oils. For example, lipid oxidation can be retarded in oil-in-water emulsions by coating the oil droplets with a relatively thick interfacial membrane that is positively charged so that it prevents transition metal ions coming into close contact with the lipids inside the droplets. In practice, the most effective means of controlling lipid oxidation in emulsions is often to use a combination of different antioxidant strategies. 4.6.5 Antimicrobial Agents Chemical preservatives that have antimicrobial properties are added to many types of food emulsion to prevent spoilage during storage and to ensure their safety for human consumption. The type of antimicrobial agent used in a particular food emulsion depends on the pH and thermal processing of the product, as well as its compatibility with the other ingredients present. Some common chemical antimicrobial agents and their effective pH ranges are: acetic acid (pH 3.0–5.0), benzoic acid (pH 2.5–4.0), sorbic acid (3.0–6.5), propionic acid (2.5–5.0), sulfites (2.5–5.0), and nitrites (4.0–5.5). Due to growing consumer demands for “natural” food products, food manufacturers are increasingly trying to replace chemical preservatives with more “label friendly” antimicrobials extracted from natural sources, for example, herbs, spices, and plants. It should be stressed that in addition to the utilization of antimicrobial additives, microbial growth in food emulsions is also inhibited using various other methods, for example, pH control, moisture control, thermal processing, nonthermal processing, chilling, and freezing. Emulsion Ingredients 173 4.6.6 Flavors The flavor of a food emulsion is one of the most important factors determining its overall quality (Chapter 9). A food manufacturer must therefore design each food product so that it has the desired characteristic flavor profile expected by consumers for that kind of product. A desirable flavor profile can be achieved by incorporating known concentrations of particular types of flavor molecules into a food (e.g., NaCl, sucrose, d-limonene, and citric acid) or by using multicomponent ingredients that contain flavor molecules (e.g., lemon juice, herbs, spices, flavor oils, and milk fat). Alternatively, the flavor profile might be generated or modified by ingredients that undergo chemical or biochemical reactions during food production, storage, or preparation, for example, lipid oxidation, browning reactions, or enzymatic reactions. A food manufacturer must decide the type and amount of flavoring components that must be incorporated into a food emulsion during the manufacturing process to produce a desirable flavor profile in the final product. This is by no means a simple task, since the perceived flavor of a food emulsion is governed not only by the type and concentration of flavors present, but also by their partitioning and release rate, which depends on the composition and microstructure of the emulsion (Chapter 9). The most important factors that should be considered when selecting flavors for use in food emulsions are their partitioning between the oil, aqueous, interfacial, and headspace regions, and their rate of mass transport to the taste and odor receptors during consumption (Chapter 9). Due to the difficulty in predicting the flavor of a food emulsion from first principles, the creation of a final product with a desirable flavor profile often requires extensive formulation and reformulation of products after testing of their flavor profiles using analytical instruments and sensory tests. Currently, there are trends within the food industry toward the utilization of more natural flavors, and to reduce the concentration of certain sugars and salts. For example, sugar-based sweeteners (such as sucrose, high-fructose corn syrup, fruit juices, and honey) that have widely been used to produce desirable flavor profiles in beverage emulsions are being replaced with reduced calorie sweeteners based on polyhydric alcohols (mannitol, xylitol, and sorbitol) or nonnutritive sweeteners (e.g., aspartame, acesulfame K, saccharin, and sucralose). The food manufacturer should be aware that changing the level of sweeteners or salts in an emulsion may alter its overall physicochemical characteristics, so that the system has to be reformulated to produce a desirable product. 4.6.7 Colorants Appearance plays a major role in determining whether or not a consumer will purchase a particular product, as well as their perception of the quality of the product once it is consumed. The desired appearance for a particular product depends mainly on the product type and its description on the label. The overall appearance of an emulsion is determined by the amount of light it absorbs and scatters across the visible region of the electromagnetic spectrum (Chapter 10). For food emulsions, the most important elements of appearance are usually the color and the opacity. The opacity is mainly determined by particulate matter that scatters light, such as emulsion droplets, air bubbles, biopolymer aggregates, fat crystals, and ice crystals. The color is mainly determined by chromophoric materials that selectively absorb light in the wavelength range from 380 to 780 nm. Substances that can absorb electromagnetic energy in this region are usually referred to as dyes or pigments. A pigment has been defined as a colored substance that is soluble in the medium in which it is dispersed, whereas a pigment is insoluble. The more general term “colorant” can be used to encompass both dyes and pigments. Substances that contribute to the color of an emulsion-based product may be naturally present in the other major ingredients used to formulate the product (e.g., oils and egg yolk) or they may be added as specific colorants. A wide variety of natural and synthetic colorants are available to provide characteristic appearances to food emulsions, including fruit, vegetable, or plant extracts and FD&C colorants. FD&C colorants are those synthetic food color additives approved by the United States Food and Drug Administration (FDA) for usage in foods, drugs, and cosmetics, for example, FD&C Red No. 40, FD&C Yellow No. 5, and FD&C Blue No. 1. Colorants may be either oil soluble or water soluble, which will determine the phase that they must be dispersed in during the production of emulsions. Many colorants undergo chemical degradation reactions that lead to a change or fading in color over time. Consequently, a manufacturer 174 Food Emulsions: Principles, Practices, and Techniques may have to develop effective strategies to prevent these undesirable changes, for example, by controlling light levels, oxygen content, pH, and storage temperatures, or by adding preservatives. The overall appearance or an emulsion may also be controlled by adding particulate material that scatters light and therefore makes the product look cloudy or opaque, for example, titanium dioxide or particulated biopolymer aggregates. The type and concentration of colorants used in a product depends on the desired final appearance, as well as the composition and microstructure of the matrix. For example, a higher concentration of a colorant may be required to produce a certain color-intensity in a high-fat oil-in-water emulsion than in a low-fat version of the same product due to the greater amount of light scattering by the droplets (McClements 2002). 4.6.8 Weighting Agents Weighting agents are often used in beverage emulsions to increase their creaming stability (Piorkowski and McClements 2014). The purpose of weighting agents is to reduce the density contrast between the oil droplets and the surrounding aqueous phase, thereby reducing the driving force for creaming. A variety of natural and synthetic weighting agents are available for utilization in beverage emulsions. The most common are brominated vegetable oil (BVO), sucrose acetate isobutyrate (SAIB), dammar gum, and ester gum. Brominated vegetable oil is produced by addition of bromine molecules to unsaturated bonds on the fatty acid chains of the triacylglycerols in food oils, for example, corn oil, soybean oil, cotton seed oil, or olive oil. Ester gum is made by esterification of wood rosin with glycerol. Damar gum is a natural exudate obtained from the shrubs of the Caesalpiniaceae and Dipterocarpaceae families. SAIB is made by the esterification of sucrose with acetic and isobutyric anhydrides. Weighting agents are usually incorporated into the oil phase prior to homogenization. The density of the weighting agent determines how much of it is required to match the oil and aqueous phase densities: BVO = 1240 – 1330 kg m−3, SAIB = 1150 kg m−3; ester gum = 1080 kg m−3; and damar gum = 1060 kg m−3. Nevertheless, the type and amount of weighting agents that can be used in beverage emulsions is restricted by governmental and international regulations in many countries. For example, in the United States only 15 ppm BVO and 100 ppm ester gum can be present in the finished product. These relatively low levels mean that it is only possible to use weighting agents to improve the stability of oil-in-water emulsions with very low droplet concentrations, typically <0.1 wt%, which practically limits their use to beverage emulsions. 4.6.9 Fat Replacers Over consumption of fatty foods is a major cause of obesity, which has been linked to major human health problems such as heart disease, diabetes, and cancer. There has therefore been a considerable trend in the food industry toward the development of reduced fat, low fat, or fat-free versions of traditional products. Fats play a variety of different roles in determining the overall appearance, texture, flavor, stability, and nutrition of food emulsions (Chapters 7 through 11). Consequently, it is usually necessary to use a combination of fat replacement ingredients with different functional roles to replace the quality attributes lost when fat droplets are removed. Some different possible ingredients that can be used as fat replacers are shown in Figure 4.35. Particles that scatter light strongly (such as titanium dioxide or protein nanoparticles) can be added to emulsions to mimic the optical properties normally provided by fat droplets. Biopolymers that can thicken solutions, such as proteins or polysaccharides, can be added to replace some of the desirable textural attributes that are lost when fat droplets are removed. These biopolymers may be added at individual molecules, molecular clusters, or colloidal particles. For example, microparticulated whey proteins (Simplesse™) have been developed that have similar dimensions as fat droplets and can mimic many of their textural and sensory properties. One of the hardest quality attributes to imitate when the fat is removed is the flavor profile, because the fat phase acts as a solvent for many characteristic flavors and controls their release rate during consumption. When the fat content is reduced the partitioning and release rate of flavor compounds is changed, which changes the overall flavor profile. In addition, the flavor profile may be changed due 175 Emulsion Ingredients (a) (d) (b) (e) (c) (f) FIGURE 4.35 Some of the desirable food attributes provided by fat droplets (opacity, texture, and mouthfeel) can be provided by polymeric or particulate substances, such as (a) thickening agents, (b) hydrogel particles, (c) protein particles, (d) indigestible fat droplets, (e) starch granules, and (f) titanium dioxide. to interactions of the flavor molecules with proteins, polysaccharides, or surfactant micelles. For these reasons, it is often necessary to supplement biopolymer fat replacers with other types of ingredient to obtain the desired flavor profile, such as surfactants or flavorings. An alternative method of reducing the fat content of emulsions is to replace a fraction or all of the conventional oil with nondigestible fat-like molecules (such as Olestra™, a sucrose fatty acid ester) or specially designed triacylglycerols with reduced caloric levels (such as Salatrim™ and Caprenin™). An ideal fat replacer should provide all of the quality attributes provided by conventional fat, while being safe to consume and significantly reducing the overall fat and calorie content. One of the advantages of this method is that the overall droplet concentration remains the same so that the physicochemical and sensory attributes of the product are fairly similar to those of a conventional product. Nevertheless, there are some problems associated with having appreciable quantities of nondigested fats within the lower gastrointestinal tract, as well as concerns about poor absorption of oil-soluble vitamins and other hydrophobic nutrients. 4.7 Factors Influencing Ingredient Selection As this chapter has shown there are a large number of different functional ingredients that can be used in food emulsions. Each of these ingredients has its own unique functional properties that contribute to the overall physicochemical and sensory properties of the final product in a distinctive way. We have considered some of the physiochemical factors that food manufacturers should consider when selecting particular types of functional ingredients in previous sections. In this section, an overview of some of the other important factors that must be considered in choosing an ingredient for a particular application is given. • There are differences in the sensitivity of ingredients to solution composition and environmental conditions, for example, pH, ionic strength, and temperature. The ingredients chosen must be capable of exhibiting their desired functional properties under the conditions that the product experiences during its production, storage, transport, and utilization. 176 Food Emulsions: Principles, Practices, and Techniques • Certain ingredients can interact with other types of ingredient in ways that dramatically alter their functional attributes. These interactions may either be beneficial or detrimental to the overall properties of the system. It is therefore important to select a combination of ingredients that are compatible with one another. • There are often differences in supplier reliability, consistency from batch to batch, ease of handling, ease of utilization, and shelf-life for each functional ingredient. • There are differences in the total amount and cost of ingredients required to provide the desired functional attributes to food emulsions. • There are legal limits, which vary from country to country, that specify the types and amounts of ingredients that can be used in particular foods. • There may be labeling and marketing requirements on the type of ingredients used. For example, consumers are becoming more interested in purchasing “all natural” products and therefore many food companies are examining the possibility of replacing synthetic ingredients with natural ones. In addition, it may be important to utilize an ingredient that is suitable for consumption by particular ethnic, religious, or social groups, for example, kosher, vegetarian, or vegan. The food manufacturer must consider all of the above factors when selecting a combination of ingredients that is suitable for application in a particular product. Usually, it will not be possible to identify a series of ingredients that satisfies all of the desired characteristics, and it will be necessary to come to some compromise between functionality, cost, and labeling requirements. REFERENCES Agboola, S. O. and D. G. Dalgleish (1996). Effects of pH and ethanol on the kinetics of destabilisation of oil-in-water emulsions containing milk proteins. Journal of the Science of Food and Agriculture 72(4): 448–454. Akoh, C. C. and D. B. Min (2008). Food Lipids: Chemistry, Nutrition, and Biotechnology. Boca Raton, FL: CRC Press. Anton, M. (2013). Egg yolk: Structures, functionalities and processes. Journal of the Science of Food and Agriculture 93(12): 2871–2880. Anton, M., V. Beaumal, C. Brossard, and G. Gandemer (2002). Droplet flocculation and physical stability of oil-in-water emulsions prepared with hen egg yolk. In Food Emulsions and Dispersions, M. Anton, ed., pp. 15–28. Trivandrum, India: Research Signpost. Aoki, H., O. Taneyama, and M. Inami (1980). Emulsifying properties of soy protein—Characteristics of 7s and 11s proteins. Journal of Food Science 45(3): 534. Atkins, P. and J. de Paula (2014). Physical Chemistry: Thermodynamics, Structure, and Change. Oxford, U.K.: Oxford University Press. Baker, E. N. and R. E. Hubbard (1984). Hydrogen-bonding in globular-proteins. Progress in Biophysics & Molecular Biology 44(2): 97–179. Barriuso, B., I. Astiasaran, and D. Ansorena (2013). A review of analytical methods measuring lipid oxidation status in foods: A challenging task. European Food Research and Technology 236(1): 1–15. Bazmi, A. and P. Relkin (2009). Effects of processing conditions on structural and functional parameters of whipped dairy emulsions containing various fatty acid compositions. Journal of Dairy Science 92(8): 3566–3574. Becher, P. (1983). Encyclopedia of Emulsion Technology. New York: Marcel Dekker. Becher, P. (1985). Encyclopedia of Emulsion Technology. New York: Marcel Dekker. Belitz, H. D., W. Grosch, and P. Schieberle (2009). Food Chemistry. Berlin, Germany: Springer. Bergethon, P. R. (2010). The Physical Basis of Biochemistry: The Foundations of Molecular Biophysics. New York: Springer. Berton-Carabin, C. C., M. H. Ropers, and C. Genot (2014). Lipid oxidation in oil-in-water emulsions: Involvement of the interfacial layer. Comprehensive Reviews in Food Science and Food Safety 13(5): 945–977. Emulsion Ingredients 177 Birker, P.J.M.W.L. and F.B. Padley (1987). Physical properties of fats and oils. In: Recent Advances in Chemistry and Technology of Fats and Oils, R.J. Hamilton and A. Bhati, eds., Chapter 1. London: Elsevier Applied Science. Boistelle, R. (1988). Fundamentals of nucleation and crystal growth. In Crystallization and Polymorphism of Fats and Fatty Acids, N. Garti and K. Sato, eds., pp. 189–226. New York: Marcel Dekker. Boode, K., C. Bisperink, and P. Walstra (1991). Destabilization of o/w emulsions containing fat crystals by temperature cycling. Colloids and Surfaces 61: 55–74. Boutte, T. and L. Skogerson (2004). Stearoyl-2-lactylates and oleoyl lactylates. In Emulsifiers in Food Technology, R. J. Whitehurst, ed., pp. 206–225. Oxford, U.K.: Blackwell Publishing. Bowron, D. T. and S. D. Moreno (2014). Using synchrotron X-ray and neutron methods to investigate structural aspects of metal ion solvation and solution structure: An approach using empirical potential structure refinement. Coordination Chemistry Reviews 277: 2–14. Brady, J. W. (2013). Introductory Food Chemistry. Ithaca, NY: Cornell University Press. Bray, G. A., S. Paeratakul, and B. M. Popkin (2004). Dietary fat and obesity: A review of animal, clinical and epidemiological studies. Physiology & Behavior 83(4): 549–555. Bueschelberger, H. G. (2004). Lecithins. In Emulsifiers in Food Technology, R. J. Whitehurst, ed., pp. 1–39. Oxford, U.K.: Blackwell Publishing. Buldo, P., J. J. K. Kirkensgaard, and L. Wiking (2013). Crystallization mechanisms in cream during ripening and initial butter churning. Journal of Dairy Science 96(11): 6782–6791. Chanamai, R. and D. J. McClements (2002). Comparison of gum arabic, modified starch, and whey protein isolate as emulsifiers: Influence of pH, CaCl(2) and temperature. Journal of Food Science 67(1): 120–125. Charoen, R., A. Jangchud, K. Jangchud, T. Harnsilawat, O. Naivikul, and D. J. McClements (2011). Influence of biopolymer emulsifier type on formation and Sstability of rice bran oil-in-water emulsions: Whey protein, gum arabic, and modified starch. Journal of Food Science 76(1): E165–E172. Cheng, H. F. (2010). Volatile flavor compounds in yogurt: A review. Critical Reviews in Food Science and Nutrition 50(10): 938–950. Cottrell, T. and J. van Peij (2004). Sorbitan esters and polysorbates. In Emulsifiers in Food Technology, R. J. Whitehurst, ed., pp. 162–165. Oxford, U.K.: Blackwell Publishing. Coupland, J. N. and D. J. McClements (1996). Lipid oxidation in food emulsions. Trends in Food Science & Technology 7(3): 83–91. Coupland, J. N. and D. J. McClements (1997). Physical properties of liquid edible oils. Journal of the American Oil Chemists Society 74(12): 1559–1564. Cui, S. W. (2005). Food Carbohydrates: Chemistry, Physical Properties and Applications. Boca Raton, FL: Taylor & Francis. Damodaran, S., K. L. Parkin, and O. R. Fennema (2007). Fennema’s Food Chemistry. Boca Raton, FL: CRC Press. Davis, H. T. (1994). Factors determining emulsion-type - hydrophile-lipophile balance and beyond. Colloids and Surfaces A: Physicochemical and Engineering Aspects 91: 9–24. Degner, B. M., C. Chung, V. Schlegel, R. Hutkins, and D. J. McClements (2014). Factors influencing the freeze-thaw stability of emulsion-based foods. Comprehensive Reviews in Food Science and Food Safety 13(2): 98–113. Dickinson, E. (1992). Introduction to Food Colloids. Cambridge, U.K.: Royal Society of Chemistry. Dickinson, E. (2003). Hydrocolloids at interfaces and the influence on the properties of dispersed systems. Food Hydrocolloids 17(1): 25–39. Dickinson, E. (2011). Mixed biopolymers at interfaces: Competitive adsorption and multilayer structures. Food Hydrocolloids 25(8): 1966–1983. Dickinson, E. (2012). Emulsion gels: The structuring of soft solids with protein-stabilized oil droplets. Food Hydrocolloids 28(1): 224–241. Dickinson, E. and M. Golding (1997). Depletion flocculation of emulsions containing unadsorbed sodium caseinate. Food Hydrocolloids 11(1): 13–18. Dickinson, E. and G. Lopez (2001). Comparison of the emulsifying properties of fish gelatin and commercial milk proteins. Journal of Food Science 66(1): 118–123. Dickinson, E. and D. J. McClements (1996). Advances in Food Colloids. Glasgow, U.K.: Blackie Academic and Professional. 178 Food Emulsions: Principles, Practices, and Techniques Evans, E. D. and W. Wennerstrom (1999). The Colloidal Domain: Where Physics, Chemistry and Biology Meet. New York: Wiley-VCH. Fennema, O. R. (2008). Water and Ice. In Food Chemistry, S. Damodaran, K. L. Parkin, and O.R. Fennema, eds., 4th edn. Boca Raton, FL: CRC Press. Foegeding, E. A. (2006). Food biophysics of protein gels: A challenge of nano and macroscopic proportions. Food Biophysics 1(1): 41–50. Foegeding, E. A. and J. P. Davis (2011). Food protein functionality: A comprehensive approach. Food Hydrocolloids 25(8): 1853–1864. Fredrick, E., P. Walstra, and K. Dewettinck (2010). Factors governing partial coalescence in oil-in-water emulsions. Advances in Colloid and Interface Science 153(1–2): 30–42. Freeman, M. P., J. R. Hibbeln, K. L. Wisner, J. M. Davis, D. Mischoulon, M. Peet, P. E. Keck et al. (2006). Omega-3 fatty acids: Evidence basis for treatment and future research in psychiatry. Journal of Clinical Psychiatry 67(12): 1954–1967. Friberg, S., K. Larsson, and J. Sjoblom (2004). Food Emulsions. New York: Marcel Dekker. Garti, N. (1999). Hydrocolloids as emulsifying agents for oil-in-water emulsions. Journal of Dispersion Science and Technology 20(1–2): 327–355. Garti, N. and M. E. Leser (2001). Emulsification properties of hydrocolloids. Polymers for Advanced Technologies 12(1–2): 123–135. Gaupp, R. and W. Adams (2004). Acid esters of mono- and diglycerides. In Emulsifiers in Food Technology, R. J. Whitehurst, ed., pp. 59–85. Oxford, U.K.: Blackwell Publishing. Gebauer, S. K., T. L. Psota, W. S. Harris, and P. M. Kris-Etherton (2006). n-3 fatty acid dietary recommendations and food sources to achieve essentiality and cardiovascular benefits. American Journal of Clinical Nutrition 83(6): 1526S–1535S. Ghosh, S. and J. N. Coupland (2008). Factors affecting the freeze-thaw stability of emulsions. Food Hydrocolloids 22(1): 105–111. Goff, H. D. (1997). Colloidal aspects of ice cream—A review. International Dairy Journal 7(6–7): 363–373. Goff, H. D. (2002). Formation and stabilisation of structure in ice-cream and related products. Current Opinion in Colloid & Interface Science 7(5–6): 432–437. Goff, H. D. and R. W. Hartel (2013). Ice Cream. New York: Springer. Goff, H. D. and C. Vega (2007). Structure-engineering of ice-cream and foam-based foods. In Understanding and Controlling the Microstructure of Complex Foods, D.J. McClements, ed., (149): 557–574. Cambridge, U.K.: Woodhead Publishing. Gomez-Guillen, M. C., B. Gimenez, M. E. Lopez-Caballero, and M. P. Montero (2011). Functional and bioactive properties of collagen and gelatin from alternative sources: A review. Food Hydrocolloids 25(8): 1813–1827. Grosberg, A. Y. and A. R. Khokhlov (2010). Giant Molecules: Here, There, and Everywhere. Hackensack, NH: World Scientific Publishing. Gunstone, F. D. (2008). Oils and Fats in the Food Industry. Chichester, U.K.: Blackwell Publishing. Gunstone, F. D., J. L. Harwood, and A. J. Dijkstra (2007). The Lipid Handbook. Boca Raton, FL: CRC Press. Guzey, D. and D. J. McClements (2006). Formation, stability and properties of multilayer emulsions for application in the food industry. Advances in Colloid and Interface Science 128: 227–248. Hartel, R. W. (1996). Ice crystallization during the manufacture of ice cream. Trends in Food Science & Technology 7(10): 315–321. Hartel, R. W. (2001). Crystallization in Foods. Gaithersburg, MD: Aspen Publishers. Hartel, R. W. (2013). Advances in food crystallization. Annual Review of Food Science and Technology 4: 277–292. Hasenhuettl, G.L (2008a). Overview of food emulsifiers. In Food Emulsifiers and their Applications, G.L. Hasenhuettl and R.W. Hartel, eds., pp 1–10. New York: Springer Science. Hasenhuettl, G.L. (2008b). Synthesis and commercial preparation of food emulsifiers. In Food Emulsifiers and their Applications, G.L. Hasenhuettl and R.W. Hartel, eds., pp 11–38. New York: Springer Science. Hasenhuettl, G. L. and R. W. Hartel (2008). Food Emulsifiers and Their Applications. New York: Springer Science. Emulsion Ingredients 179 Helgason, T., T. S. Awad, K. Kristbergsson, D. J. McClements, and J. Weiss (2008). Influence of polymorphic transformations on gelation of tripalmitin solid lipid nanoparticle suspensions. Journal of the American Oil Chemists Society 85(6): 501–511. Hiemenz, P. C. and R. Rajagopalan (1997). Principles of Colloid and Surface Chemistry. New York: Marcel Dekker. Himawan, C., V. M. Starov, and A. G. F. Stapley (2006). Thermodynamic and kinetic aspects of fat crystallization. Advances in Colloid and Interface Science 122(1–3): 3–33. Holmberg, K., B. Jonsson, B. Kronberg, and B. Lindman (2002). Surfactants and Polymers in Aqueous Solution. New York: Wiley. Hostettmann, K. and A. Marston (1995). Saponins. Cambridge, U.K.: Cambridge University Press. Huang, X., Y. Kakuda, and W. Cui (2001). Hydrocolloids in emulsions: Particle size distribution and interfacial activity. Food Hydrocolloids 15(4–6): 533–542. Hunt, J. A. and D. G. Dalgleish (1995). Heat-stability of oil-in-water emulsions containing milk-proteins— Effect of ionic-strength and PH. Journal of Food Science 60(5): 1120. Hunter, R. J. (1986). Foundations of Colloid Science. Oxford, U.K.: Oxford University Press. Igoe, R. S. (2011). Dictionary of Food Ingredients. New York: Springer Scientific. Islam, A. M., G. O. Phillips, A. Sljivo, M. J. Snowden, and P. A. Williams (1997). A review of recent developments on the regulatory, structural and functional aspects of gum arabic. Food Hydrocolloids 11(4): 493–505. Israelachvili, J. (2011). Intermolecular and Surface Forces, 3rd edn. London, U.K.: Academic Press. Jacobsen, C., M. B. Let, N. S. Nielsen, and A. S. Meyer (2008). Antioxidant strategies for preventing oxidative flavour deterioration of foods enriched with n-3 polyunsaturated lipids: A comparative evaluation. Trends in Food Science & Technology 19(2): 76–93. Johansson, D. and B. Bergenstahl (1995). Sintering of fat crystal networks in oil during post-crystallization processes. Journal of the American Oil Chemists Society 72(8): 911–920. Kabalnov, A. (1998). Thermodynamic and theoretical aspects of emulsions and their stability. Current Opinion in Colloid & Interface Science 3(3): 270–275. Kabalnov, A. and H. Wennerstrom (1996). Macroemulsion stability: The oriented wedge theory revisited. Langmuir 12(2): 276–292. Karim, A. A. and R. Bhat (2009). Fish gelatin: Properties, challenges, and prospects as an alternative to mammalian gelatins. Food Hydrocolloids 23(3): 563–576. Kasapis, S., I. T. Norton, and J. B. Ubbink (2009). Modern Biopolymer Science: Bridging the Divide between Fundamental Treatise and Industrial Application. New York: Academic Press. Kashchiev, D. and G. M. van Rosmalen (2003). Review: Nucleation in solutions revisited. Crystal Research and Technology 38(7–8): 555–574. Kralova, I. and J. Sjoblom (2009). Surfactants used in food industry: A review. Journal of Dispersion Science and Technology 30(9): 1363–1383. Lam, R. S. H. and M. T. Nickerson (2013). Food proteins: A review on their emulsifying properties using a structure-function approach. Food Chemistry 141(2): 975–984. Le Denmat, M., M. Anton, and V. Beaumal (2000). Characterisation of emulsion properties and of interface composition in O/W emulsions prepared with hen egg yolk, plasma and granules. Food Hydrocolloids 14(6): 539–549. Le Denmat, M., M. Anton, and G. Gandemer (1999). Protein denaturation and emulsifying properties of plasma and granules of egg yolk as related to heat treatment. Journal of Food Science 64(2): 194–197. Leal-Calderon, F., V. Schmitt, and J. Bibette (2007). Emulsion Science: Basic Principles. Springer Verlag, New York. Lee, A. Y., D. Erdemir, and A. S. Myerson (2011). Crystal polymorphism in chemical process development. Annual Review of Chemical and Biomolecular Engineering 2: 259–280. Leray, C. (2014). Lipids: Nutrition and Health. Boca Raton, FL: CRC Press. Leroux, J., V. Langendorff, G. Schick, V. Vaishnav, and J. Mazoyer (2003). Emulsion stabilizing properties of pectin. Food Hydrocolloids 17(4): 455–462. Leuenberger, B. H. (1991). Investigation of viscosity and gelation properties of different mammalian and fish gelatins. Food Hydrocolloids 5(4): 353–361. 180 Food Emulsions: Principles, Practices, and Techniques Li, X. P., K. Huang, J. Y. Lin, Y. Z. Xu, and H. Z. Liu (2014). Hofmeister ion series and its mechanism of action on affecting the behavior of macromolecular solutes in aqueous solution. Progress in Chemistry 26(8): 1285–1291. Lindfors, L., S. Forssen, J. Westergren, and U. Olsson (2008). Nucleation and crystal growth in supersaturated solutions of a model drug. Journal of Colloid and Interface Science 325(2): 404–413. Liu, S. J. and J. H. Masliyah (1996). Rheology of suspensions. Suspensions: Fundamentals and Applications in the Petroleum Industry 251: 107–176. Ma, Z. and J. I. Boye (2013). Advances in the design and production of reduced-fat and reduced-cholesterol salad dressing and mayonnaise: A review. Food and Bioprocess Technology 6(3): 648–670. Marangoni, A. G., N. Acevedo, F. Maleky, E. Co, F. Peyronel, G. Mazzanti, B. Quinn, and D. Pink (2012). Structure and functionality of edible fats. Soft Matter 8(5): 1275–1300. Marangoni, A. G. and D. Tang (2008). Modeling the rheological properties of fats: A perspective and recent advances. Food Biophysics 3(2): 113–119. Marangoni, A. G. and L. H. Wesdorp (2012). Structure and Properties of Fat Crystal Networks. Boca Raton, FL: CRC Press. Marcus, Y. (2009). Effect of ions on the structure of water: Structure making and breaking. Chemical Reviews 109(3): 1346–1370. McClements, D. J. (2000). Comments on viscosity enhancement and depletion flocculation by polysaccharides. Food Hydrocolloids 14(2): 173–177. McClements, D. J. (2002). Modulation of globular protein functionality by weakly interacting cosolvents. Critical Reviews in Food Science and Nutrition 42(5): 417–471. McClements, D. J. (2004). Protein-stabilized emulsions. Current Opinion in Colloid & Interface Science 9(5): 305–313. McClements, D. J. (2005). Food Emulsions: Principles, Practice, and Techniques. Boca Raton, FL: CRC Press. McClements, D. J. (2007). Critical review of techniques and methodologies for characterization of emulsion stability. Critical Reviews in Food Science and Nutrition 47(7): 611–649. McClements, D. J. (2012). Crystals and crystallization in oil-in-water emulsions: Implications for emulsionbased delivery systems. Advances in Colloid and Interface Science 174: 1–30. McClements, D. J. (2014). Nanoparticle- and Microparticle-Based Delivery Systems: Encapsulation, Protection and Release of Active Components. Boca Raton, FL: CRC Press. McClements, D. J. and E. A. Decker (2000). Lipid oxidation in oil-in-water emulsions: Impact of molecular environment on chemical reactions in heterogeneous food systems. Journal of Food Science 65(8): 1270–1282. McSweeney, P. L. H. (2004). Biochemistry of cheese ripening. International Journal of Dairy Technology 57(2–3): 127–144. Mendez-Velasco, C. and H. D. Goff (2012). Fat structure in ice cream: A study on the types of fat interactions. Food Hydrocolloids 29(1): 152–159. Mignino, L. A., M. C. Tomas, and M. E. Paredi (2011). Effect of frozen storage on emulsifying properties of actomyosin from mantle and fins of squid (Illex argentinus). European Food Research and Technology 233(3): 437–445. Mine, Y. (2002). Recent advances in egg protein functionality in the food system. Worlds Poultry Science Journal 58(1): 31–39. Mira, I., N. Zambrano, E. Tyrode, L. Márquez, A. Peña, A. Pizzino, and J. Salager (2003). Emulsion catastrophic inversion from abnormal to normal morphology. 2. Effect of the stirring intensity on the dynamic inversion frontier. Industrial and Engineering Chemistry Research 42(1): 57–61. Mitra, S. and S. R. Dungan (1997). Micellar properties of quillaja saponin.1. Effects of temperature, salt, and pH on solution properties. Journal of Agricultural and Food Chemistry 45(5): 1587–1595. Moonen, H. and H. Bas (2004). Mono- and di-glycerides. In Emulsifiers in Food Technology, R. J. Whitehurst, ed., pp. 40–58. Oxford, U.K.: Blackwell Publishing. Moulik, S. P. (1996). Micelles: Self-organized surfactant assemblies. Current Science 71(5): 368–376. Moure, A., J. Sineiro, H. Dominguez, and J. C. Parajo (2006). Functionality of oilseed protein products: A review. Food Research International 39(9): 945–963. Muoio, D. M. and C. B. Newgard (2006). Obesity-related derangements in metabolic regulation. Annual Review of Biochemistry 75: 367–401. Myers, D. (2006). Surfactant Science and Technology. Hoboken, NJ: John Wiley & Sons. Emulsion Ingredients 181 Nelen, B. A. P. and J. M. Cooper (2004). Sucrose esters. In Emulsifiers in Food Technology, R. J. Whitehurst, ed., pp. 131–161. Oxford, U.K.: Blackwell Publishing. Nishinari, K., Y. Fang, S. Guo, and G. O. Phillips (2014). Soy proteins: A review on composition, aggregation and emulsification. Food Hydrocolloids 39: 301–318. Norde, W. (2011). Colloids and Interfaces in Life Sciences and Bionanotechnology. Boca Raton, FL: CRC Press. Norn, V. (2004). Polyglycerol esters. In Emulsifiers in Food Technology, R. J. Whitehurst, ed., pp. 110–130. Oxford, U.K.: Blackwell Publishing. Norton, I. T. and W. J. Frith (2001). Microstructure design in mixed biopolymer composites. Food Hydrocolloids 15(4–6): 543–553. Nussinovitch, A. (1997). Hydrocolloid Applications: Gum Technology in the Food and Other Industries. New York: Springer. Olijve, J., F. Mori, and Y. Toda (2001). Influence of the molecular-weight distribution of gelatin on emulsion stability. Journal of Colloid and Interface Science 243(2): 476–482. Pasquali, R. C., M. P. Taurozzi, and C. Bregni (2008). Some considerations about the hydrophilic-lipophilic balance system. International Journal of Pharmaceutics 356(1–2): 44–51. Petursson, S., E. A. Decker, and D. J. McClements (2004). Stabilization of oil-in-water emulsions by cod protein extracts. Journal of Agricultural and Food Chemistry 52(12): 3996–4001. Piorkowski, D. T. and D. J. McClements (2014). Beverage emulsions: Recent developments in formulation, production, and applications. Food Hydrocolloids 42: 5–41. Pugnaloni, L. A., E. Dickinson, R. Ettelaie, A. R. Mackie, and P. J. Wilde (2004). Competitive adsorption of proteins and low-molecular-weight surfactants: computer simulation and microscopic imaging. Advances in Colloid and Interface Science 107(1): 27–49. Qian, C., E. A. Decker, H. Xiao, and D. J. McClements (2011). Comparison of biopolymer emulsifier performance in formation and stabilization of orange oil-in-water emulsions. Journal of the American Oil Chemists Society 88(1): 47–55. Queste, S., J. L. Salager, R. Strey, and J. M. Aubry (2007). The EACN scale for oil classification revisited thanks to fish diagrams. Journal of Colloid and Interface Science 312(1): 98–107. Rahman, M. S. (2009). Food Properties Handbook. Boca Raton, FL: CRC Press. Rao, M. A. (2013). Rheology of Fluid, Semisolid, and Solid Foods: Principles and Applications. New York: Springer Science. Rondón-Gonzaléz, M., V. Sadtler, L. Choplin, and J. Salager (2006). Emulsion catastrophic inversion from abnormal to normal morphology. 5. Effect of the water-to-oil ratio and surfactant concentration on the inversion produced by continuous stirring. Industrial and Engineering Chemistry Research 45(9): 3074–3080. Rossmurphy, S. B. (1995). Structure-property relationships in food biopolymer gels and solutions. Journal of Rheology 39(6): 1451–1463. Ruxton, C. H. S., S. C. Reed, M. J. A. Simpson, and K. J. Millington (2004). The health benefits of omega-3 polyunsaturated fatty acids: A review of the evidence. Journal of Human Nutrition and Dietetics 17(5): 449–459. Salager, J. L., R. Antón, D. Sabatini, J. Harwell, E. Acosta, and L. Tolosa (2005). Enhancing solubilization in microemulsions—State of the art and current trends. Journal of Surfactants and Detergents 8(1): 3–21. Salager, J. L., A. Forgiarini, L. Marquez, A. Pena, A. Pizzino, M. P. Rodriguez, and M. Rondo-Gonzalez (2004). Using emulsion inversion in industrial processes. Advances in Colloid and Interface Science 108: 259–272. Sato, K., L. Bayes-Garcia, T. Calvet, M. A. Cuevas-Diarte, and S. Ueno (2013). External factors affecting polymorphic crystallization of lipids. European Journal of Lipid Science and Technology 115(11): 1224–1238. Schmitt, C. and S. L. Turgeon (2011). Protein/polysaccharide complexes and coacervates in food systems. Advances in Colloid and Interface Science 167(1–2): 63–70. Shinoda, K. and S. Friberg (1986). Emulsions and Solubilization. New York: Wiley-Interscience Sidhu, G. and D. Oakenfull (1986). A mechanism for the hypocholesterolaemic activity of saponins. British Journal of Nutrition 55(3): 643–649. Smith, J. and L. Hong-Shum (2011). Food Additives Data Book. Chichester, U.K.: Wiley-Blackwell. 182 Food Emulsions: Principles, Practices, and Techniques Smith, K. W., K. Bhaggan, G. Talbot, and K. F. van Malssen (2011). Crystallization of fats: Influence of minor components and additives. Journal of the American Oil Chemists Society 88(8): 1085–1101. Sparso, F. V. and N. Krog (2004). Propylene glycol fatty acid esters. In Emulsifiers in Food Technology, R. J. Whitehurst, ed., pp. 186–205. Oxford, U.K.: Blackwell Publishing. Stauffer, S. E. (1999). Emulsifiers. St Paul, MN: Eagen Press. Stephen, A. J., G. O. Phillips, and P. A. Williams (2006). Food Polysaccharides and Their Applications. Boca Raton, FL: CRC Press. Surh, J., E. A. Decker, and D. J. McClements (2006). Properties and stability of oil-in-water emulsions stabilized by fish gelatin. Food Hydrocolloids 20(5): 596–606. Swaisgood, H. E. (2008). Characteristics of Milk. In Food Chemistry, S. Damodaran, K. L. Parkin, and O. R. Fennema, eds., pp. 886–921. Boca Raton, FL: CRC Press. Sweedman, M. C., M. J. Tizzotti, C. Schafer, and R. G. Gilbert (2013). Structure and physicochemical properties of octenyl succinic anhydride modified starches: A review. Carbohydrate Polymers 92(1): 905–920. Taherian, A. R., M. Britten, H. Sabik, and P. Fustier (2011). Ability of whey protein isolate and/or fish gelatin to inhibit physical separation and lipid oxidation in fish oil-in-water beverage emulsion. Food Hydrocolloids 25(5): 868–878. Tanford, C. (1980). The Hydrophobic Effect. New York: John Wiley & Sons. Tcholakova, S., N. D. Denkov, I. B. Ivanov, and B. Campbell (2006). Coalescence stability of emulsions containing globular milk proteins. Advances in Colloid and Interface Science 123: 259–293. Thivilliers-Arvis, F., E. Laurichesse, V. Schmitt, and F. Leal-Calderon (2010). Shear-induced instabilities in oil-in-water emulsions comprising partially crystallized droplets. Langmuir 26(22): 16782–16790. Timms, R. E. (1991). Crystallization of fats. Chemistry and Industry 20: 342–345. Toledano, O. and S. Magdassi (1998). Emulsification and foaming properties of hydrophobically modified gelatin. Journal of Colloid and Interface Science 200(2): 235–240. Tolstoguzov, V. (2002). Thermodynamic aspects of biopolymer functionality in biological systems, foods, and beverages. Critical Reviews in Biotechnology 22(2): 89–174. Tolstoguzov, V. (2003). Some thermodynamic considerations in food formulation. Food Hydrocolloids 17(1): 1–23. Turgeon, S. L., C. Schmitt, and C. Sanchez (2007). Protein-polysaccharide complexes and coacervates. Current Opinion in Colloid & Interface Science 12(4–5): 166–178. van Setten, D. C., G. J. Ten Hove, E. J. H. J. Wiertz, J. P. Kamerling, and G. van de Werken (1998). Multiplestage tandem mass spectrometry for structural characterization of saponins. Analytical Chemistry 70(20): 4401–4409. van Setten, D. C., G. van de Werken, G. Zomer, and G. F. A. Kersten (1995). Glycosyl compositions and structural characteristics of the potential immuno-adjuvant active saponins in the Quillaja saponaria Molina extract Quil A. Rapid Communications in Mass Spectrometry 9(8): 660–666. Vanapalli, S. A., J. Palanuwech, and J. N. Coupland (2002). Stability of emulsions to dispersed phase crystallization: Effect of oil type, dispersed phase volume fraction, and cooling rate. Colloids and Surfaces A: Physicochemical and Engineering Aspects 204(1–3): 227–237. Vekilov, P. G. (2010). Nucleation. Crystal Growth & Design 10(12): 5007–5019. Voet, D. and J. G. Voet (2010). Biochemistry. New York: Wiley Scientific. Waller, G. R. and K. Yamasaki (1996a). Saponins Used in Food and Agriculture. New York: Plenum Press. Waller, G. R. and K. Yamasaki (1996b). Saponins Used in Traditional and Modern Medicine. New York: Plenum Press. Walstra, P. (1987). Fat crystallization. In Food Structure and Behavior, J. M. V. Blandshard and P. Lillford, eds. London, U.K.: Academic Press. Walstra, P. (2003). Physical Chemistry of Foods. New York: Marcel Decker. Waraho, T., V. Cardenia, E. A. Decker, and D. J. McClements (2010). Lipid oxidation in emulsified food products. Oxidation in Foods and Beverages and Antioxidant Applications, Management in Different Industry Sectors 2(200): 306–343. Waraho, T., D. J. McClements, and E. A. Decker (2011). Mechanisms of lipid oxidation in food dispersions. Trends in Food Science & Technology 22(1): 3–13. Whitehurst, R. J. (2004). Emulsifiers in Food Technology. Oxford, U.K.: Blackwell Publishing. Williams, P. A. and G. O. Phillips (2003). The use of hydrocolloids to improve food texture. In Texture in Foods, vol. 1: Semi-Solid Foods, B. M. McKenna, ed., pp. 251–274. Boca Raton, FL: CRC Press. Emulsion Ingredients 183 Williams, P. A. and G. O. Phillips (2009). Gum arabic. In Handbook of Hydrocolloids, 2nd edn., Cambridge, U.K.: Woodhead Publishing. vol. 173, pp. 252–273. Witthayapanyanon, A., J. Harwell, and D. Sabatini (2008). Hydrophilic-lipophilic deviation (HLD) method for characterizing conventional and extended surfactants. Journal of Colloid and Interface Science 325(1): 259–266. Wright, A. J., M. G. Scanlon, R. W. Hartel, and A. G. Marangoni (2001). Rheological properties of milkfat and butter. Journal of Food Science 66(8): 1056–1071. Yang, Y., M. E. Leser, A. A. Sher, and D. J. McClements (2013). Formation and stability of emulsions using a natural small molecule surfactant: Quillaja saponin (Q-Naturale (R)). Food Hydrocolloids 30(2): 589–596. Yang, Y. and D. J. McClements (2013). Encapsulation of vitamin E in edible emulsions fabricated using a natural surfactant. Food Hydrocolloids 30(2): 712–720. 5 Interfacial Properties and Their Characterization 5.1 Introduction The interfacial region that separates the oil from the aqueous phase constitutes only a small fraction of the total volume of conventional emulsions (Table 1.1). Nevertheless, it has a major influence on the bulk physicochemical and sensory properties of food emulsions, including their formation, stability, rheology, and flavor (Chapters 6 through 9). Food scientists would like to know how interfacial characteristics (such as composition, structure, thickness, and rheology) impact emulsion properties, and how these interfacial characteristics depend on the type, concentration, and properties of the surface-active components in the system, so that they can rationally create emulsion-based foods with improved quality. An interface is a narrow region that separates two phases, which could be a gas and a liquid, a gas and a solid, two liquids, a liquid and a solid, or two solids. The two phases may consist of different kinds of molecules (e.g., oil and water) or different physical states of the same kind of molecule (e.g., liquid oil and solid fat). By convention, the region separating two condensed phases (solids or liquids) is referred to as an interface, while the region separating a condensed phase and a gas is called a surface. Nevertheless, the terms interface and surface are frequently used interchangeably, and in this chapter, the term interface will often be used to cover both terms. A number of different types of surfaces and interfaces commonly occur in food emulsions, including oil–water (e.g., oil droplets in water), air–water (e.g., gas bubbles in water), solid water–liquid water (e.g., ice crystals in water), and solid fat–lipid oil (e.g., fat crystals in oil). In this chapter, we will mainly focus on the oil–water interface, because it is present in all food emulsions. Nevertheless, various other types of surface and interface will also be considered where appropriate, and it should be recognized that much of the discussion about oil–water interfaces is also applicable to other systems. Initially, it is useful to provide a brief overview of the interfacial characteristics that are most important in determining the overall properties of food emulsions: • Interfacial composition: The type and concentration of surface-active substances present at an interface strongly influence its free energy, structure, dimensions, electrical characteristics, and rheology, and therefore plays a major role in determining emulsion properties. • Interfacial structure: The thickness and internal structure of the interfacial layer play an important role in determining the magnitude and range of the colloidal forces acting between emulsion droplets (Chapter 3). For example, the range of the steric repulsion between droplets increases as the thickness of the interfacial layer increases (Section 3.5). • Interfacial electrical properties: The electrical characteristics of the interface (e.g., surface charge density and surface potential) play an important role in determining the magnitude and range of the electrostatic interactions between emulsion droplets (Section 3.4), as well as influencing the adsorption of ions to emulsion droplet surfaces. • Interfacial energy: The free energy stored in the interface, which is described by the surface or interfacial tension, determines the ease at which the interfacial area can be changed. This interfacial energy is important in emulsion formation, since it influences the amount of mechanical energy that must be input during homogenization to deform and break up droplets 185 186 Food Emulsions: Principles, Practices, and Techniques (Section 6.4.1). It is also important in determining the stability of some surfactant-stabilized emulsions to coalescence (Sections 4.4.1 and 7.6). Finally, measurements of the interfacial energy can be used to provide valuable information about interfacial composition, emulsifier adsorption kinetics, and interfacial rheology. • Interfacial rheology: The rheological characteristics of an interface, such as viscosity and elasticity, influence the formation, stability, and functional performance of many food emulsions. In addition, measurements of the rheological characteristics of interfaces can be used to provide valuable information about the interactions of surface-active molecules within the interfacial layer. • Interfacial responsiveness: The changes in interfacial properties in response to changes in environmental conditions (such as pH, ionic strength, temperature, solvent type, or enzyme activity) are important in determining the stability and performance of food emulsions. For example, some interfaces do not change their properties appreciably when the pH is changed, whereas others may undergo appreciable alterations in packing, thickness, and charge. In the remainder of this chapter, we consider the molecular characteristics of the interfacial region, thermodynamic relationships for describing its properties, the role that it plays in determining the bulk physicochemical properties of emulsions, and experimental techniques available for characterizing its properties. 5.2 General Characteristics of Interfaces 5.2.1 Interfaces Separating Two Pure Liquids The interface that separates the oil and water phases is often assumed to be a planar surface of infinitesimal thickness (Figure 5.1a). This assumption is convenient for many purposes, but it ignores the highly dynamic nature of the interfacial region, as well as the structure and organization of the various types of molecules involved (Figure 5.1b). On the molecular level, the oil and water molecules intermingle with each other over distances of the order of a few molecular diameters (Evans and Wennerstrom 1999, Jonsson et al. 2003, Israelachvili 2011). The composition of the system therefore varies gradually across the interfacial region (Figure 5.1b), rather than changing abruptly (Figure 5.1a). The thickness and dynamics of the interfacial region depend on the relative magnitude of the interactions between the Oil Water (a) (b) FIGURE 5.1 Interfaces are often assumed to be planar surfaces of infinitesimally small thickness, but in reality, they are highly dynamic and have a thickness that depends on the dimensions and interactions of the molecules (a) continuum theory and (b) molecular theory. 187 Interfacial Properties and Their Characterization molecules involved (oil–oil, water–water, and oil–water): the more unfavorable the oil–water interactions, the thinner and more inflexible is the interfacial region. Consider the various types of molecular interactions that operate in a two-phase system consisting of oil and water (Figure 5.2). The water molecules are capable of forming relatively strong hydrogen bonds with their neighbors in the bulk water phase, whereas the oil molecules are only capable of forming relatively weak van der Waals bonds with each other in the bulk oil phase. At the oil–water interface, oil molecules can only form relatively weak van der Waals bonds with water molecules, because they do not have polar groups that would enable them to form hydrogen bonds. Consequently, increasing the number of interactions between oil and water molecules by increasing the interfacial area is unfavorable, because it involves replacing relatively strong water–water bonds with relatively weak water–oil bonds. In reality, one must take into account the structural organization (entropy) of the molecules at the interface, as well as their interaction energies (enthalpy), particularly because of the dominant role of the hydrophobic effect (Norde 2011). In summary, the interaction of oil and water molecules at the interface is strongly thermodynamically unfavorable because of the hydrophobic effect (Section 4.3.3). It is therefore necessary to supply free energy to the system to increase the contact area between oil and water molecules. The amount of free energy that must be supplied is proportional to the increase in contact area between the oil and water molecules (Hiemenz and Rajagopalan 1997): ΔG = γiΔA (5.1) where ΔG is the free energy required to increase the contact area between the two immiscible liquids by ΔA (at constant temperature and pressure) γi is a constant of proportionality called the interfacial tension If one of the phases is a gas, the interfacial tension is replaced by the surface tension, γs. Ultimately, the interfacial tension is determined by the magnitude of the imbalance of molecular interactions across an interface: the greater the imbalance of interactions, the greater is the interfacial tension. Conceptually, the interfacial tension can be thought of as a contractile force that manifests itself as a tendency for the system to minimize the contact area between the two phases. The interfacial tension can be expressed Weak interactions Oil Weak interactions Strong interactions Water FIGURE 5.2 The molecular origin of interfacial tension at a liquid–liquid interface is the imbalance of the attractive forces acting on the molecules at the interface. The length of the arrows is related to the strength of the attraction between molecules. Hence, water–water interactions are considerably stronger than oil–oil or oil–water interactions. 188 Food Emulsions: Principles, Practices, and Techniques in units of energy per unit interfacial area (J m−2) or force per unit length of interface (N m−1). Selected values for the surface and interfacial tensions of some substances relevant to food scientists are presented in Table 5.1. For some simple systems, equations have been developed to relate the interfacial and surface tensions of materials to the number and magnitude of the various types of molecular interactions involved (Israelachvili 2011, Norde 2011). Many important properties of food emulsions are governed by the imbalance of molecular forces at interfaces, including the tendency for droplets to be spherical, the surface activity of emulsifiers, the nucleation and growth of ice and fat crystals, meniscus formation, and the rise of liquids in a capillary tube (Section 5.9). 5.2.2 Interfaces in the Presence of Solutes So far only the molecular characteristics of an interface that separates two pure liquids have been considered. In practice, food emulsions contain various types of surface-active substances that can accumulate at interfaces and alter their properties, for example, surfactants, phospholipids, proteins, polysaccharides, alcohols, and particulate matter. In this section, the molecular origin of adsorption of surfaceactive substances to interfaces is considered. Consider a system that consists of an amphiphilic solute, an oil phase, and a water phase (Figure 5.3). The solute tends to accumulate at the interface, separating the oil and water phases when the free energy of the adsorbed state is lower than that of the nonadsorbed state (Hiemenz and Rajagopalan 1997, Norde 2011). The difference in free energy between the adsorbed and nonadsorbed states, ΔGads, is determined by changes in the interaction energies of the various molecules involved in the process, as well as by various entropy effects. The change in the interaction energies that occurs as a result of adsorption of a solute comes from two sources: one associated with the interface and the other with the solute itself. The hydrophobic effect makes a major contribution to both of these sources.* Firstly, when a solute adsorbs TABLE 5.1 Approximate Values at Ambient Temperature for Surface and Interfacial Tensions (mJ m−2) of Selected Materials Relevant to Food Scientists Substance Aqueous phases Water SDS solutiona Tween 20 solutiona Protein solutiona Gum Arabic solutiona 40% Sucrose solution 20% NaCl solution Organic phases n-octane n-dodecane n-hexadecane Triacylglycerol (liquid) Triacylglycerol (solid) Ethanol γS (Against Air) γI (Against Water) γI (Against Oil) 72 37 35 50 47 74 82 — — — — — — — 27.5 10 26 20–25 43 22 25 27 35 22 54 30 31 0 — — — — 4 Source: Data from various sources. Note: Interfacial tension values actually depend strongly on oil type used (e.g., hydrocarbon, triglyceride, or flavor oil). a Interfacial or surface tension at saturation coverage. * It is assumed that the interaction energy includes both enthalpy and entropy contributions, so that hydrophobic interactions can be conveniently treated. 189 Interfacial Properties and Their Characterization Oil Water (a) (b) FIGURE 5.3 Surface-active molecules accumulate in the interfacial region, because this minimizes the free energy of the system. (a) Continuum theory and (b) molecular theory. to an oil–water interface, the number of unfavorable contacts between water molecules and oil molecules is reduced. The direct contact between oil and water molecules is replaced by contacts between the nonpolar segments of the solute and oil molecules, and between the polar segments of the solute and water molecules. Thermodynamically, these interactions are more favorable than the direct interactions between oil and water molecules. Secondly, surface-active solutes usually have both polar and nonpolar segments, and when they are dispersed in bulk water, some of the nonpolar segments come into contact with water, which is thermodynamically unfavorable because of the hydrophobic effect. By adsorbing to an interface, they are able to maximize the number of thermodynamically favorable interactions between the polar segments and water, while minimizing the number of unfavorable interactions between the nonpolar segments and water (Figure 5.3b). In addition to the hydrophobic effect, various other types of interaction energy may also contribute to the propensity of an amphiphilic solute to adsorb to an interface, including van der Waals, hydration, electrostatic, and steric interactions. The relative magnitude of these interactions, and whether they favor or oppose adsorption, depends on the type of solute involved, but these contributions are often appreciably smaller than the hydrophobic effect. There are also a number of entropic contributions that influence the tendency of solute molecules to adsorb to an interface (Norde 2011): 1. Configuration entropy: When a molecule adsorbs to an interface, it is confined to a region that is considerably smaller than the region it could potentially occupy in the bulk liquid. This reduction in the number of possible configurations that the molecule can adopt within the system leads to a decrease in entropy, which opposes adsorption. 2. Orientation entropy: When a molecule adsorbs to an interface, its ability to adopt different three-dimensional orientations due to rotation may be limited compared to when it is dispersed in the bulk liquid. This reduction in the number of possible orientations that the molecule can have leads to a decrease in entropy, which opposes adsorption. 3. Conformation entropy: When a polymeric molecule adsorbs to an interface, the number of different conformations that it can adopt may either increase or decrease, depending on the nature of the molecule. For example, the number of conformations adopted by flexible random coil type biopolymers usually decreases after adsorption, whereas the number of conformations adopted by compact globular biopolymers usually increases. This contribution may therefore either decrease or increase the entropy, and so either oppose or favor adsorption, depending on whether the number of molecular conformations decreases or increases after adsorption. 190 Food Emulsions: Principles, Practices, and Techniques 4. Interaction entropy: The interaction entropy is primarily associated with changes in the organization of molecules resulting from adsorption. The major contribution to the interaction entropy is the hydrophobic effect, which has already been included in the molecular interaction contribution discussed earlier. The number of highly ordered water molecules surrounding nonpolar groups decreases when an amphiphilic molecule adsorbs to an interface, which increases the entropy of the solvent molecules and favors adsorption. Nevertheless, there may also be other entropy contributions associated with changes in the organization of solvent molecules due to adsorption, for example, redistribution of counterions around charged groups. The adsorption of a molecule to an interface will therefore only occur when the various favorable interaction energy and entropy contributions outweigh the unfavorable ones. If the adsorption free energy is highly negative (i.e., ΔGads/RT ≪ 0), then a molecule has a strong affinity for the surface and has a high surface activity. If the adsorption free energy is relatively small compared to the thermal energy (i.e., ΔGads/RT ≈ 0), then a molecule tends to be located mainly in the bulk liquid and has a low surface activity. If the adsorption free energy is highly positive (i.e., ΔGads/RT ≫ 0), then there is a deficit of solute in the interfacial region, which is referred to as negative adsorption. The change in free energy of a system that occurs when a surface-active solute is present manifests itself as a change in the interfacial (or surface) tension, that is, in the amount of free energy required to increase the interfacial (or surface) area between the water and oil (or air) phases by a unit amount. The interfacial tension is reduced in the presence of a surface-active solute, because the thermodynamically unfavorable contacts between the oil and water phases are reduced: the higher the solute concentration at the interface, the greater is the reduction in interfacial tension. The reduction of the interfacial tension by the presence of a surface-active solute is referred to as the surface pressure: π = γo/w − γ (5.2) where γo/w is the interfacial tension of a pure oil–water interface γ is the interfacial tension in the presence of the surface-active solute (Hiemenz and Rajagopalan 1997) Typical plots of the dependence of the interfacial tension and surface pressure on the concentration of a surface-active solute in the bulk solution are shown in Figure 5.4. As the solute concentration is increased, the interfacial tension continues to fall from its value in the absence of solute (γo/w), until it reaches a relatively constant level at high solute concentrations where the interface has become saturated with solute. In contrast, the surface pressure increases from 0 in the absence of solute to a constant value (π∞) at high solute concentrations where the interface has become saturated with solute. The value of π∞ is a measure of how effectively the adsorbed solute molecules are able to minimize the thermodynamically unfavorable interactions between the oil and water phases at saturation. The higher the value of π∞ for a particular interface, the better the solute is at minimizing thermodynamically unfavorable interactions at that interface. The variation of surface tension with solute concentration for different types of molecules is shown in Figure 5.5. For a solute that has a high surface activity (i.e., ΔGads/RT ≪ 0), like SDS, the surface tension decreases rapidly with increasing solute concentration, and then reaches a relatively constant value when the surface is saturated. For a solute with a moderate surface activity, like methanol, the surface tension decreases more slowly with increasing solute concentration. For a solute with little or no surface activity (i.e., ΔGads/RT ≈ 0), like sucrose, there is a little change in surface tension with increasing solute concentration. For a solute with a negative surface activity (i.e., ΔGads/RT > 0), like NaCl, there is actually an increase in surface tension with increasing solute concentration. Thermodynamic equations for relating the interfacial concentration of a solute to its bulk concentration and surface activity are discussed in Section 5.3.2. 191 Interfacial Properties and Their Characterization 80 40 Slope = –ΓRT 30 60 π (mN m–1) Γ (mN m–1) 70 35 50 40 20 π 15 10 30 20 25 5 –2 –1 0 1 0 0 0.25 ln (c) 0.5 0.75 1 c c FIGURE 5.4 The interfacial properties of an emulsifier can be conveniently characterized by plotting the interfacial ­tension or surface pressure as a function of emulsifier concentration in the bulk solution. As the bulk emulsifier concentration increases, the interfacial emulsifier concentration increases, which leads to a decrease in γ and an increase in π until the interface is saturated with emulsifier. 85 NaCl 80 Surface tension (mN m−1) 75 Sucrose 70 65 60 55 50 Methanol 45 40 35 SDS 0 5 10 15 20 25 Concentration (wt%) FIGURE 5.5 Solutes have different surface tension versus concentration profiles because of differences in their affinity for the surface. 192 Food Emulsions: Principles, Practices, and Techniques 5.3 Adsorption of Solutes to Interfaces In the previous section, the molecular origin of the ability of solutes to adsorb to interfaces was highlighted. In this section, we introduce mathematical quantities and thermodynamic relationships that can be used to describe the adsorption of solutes to interfaces. As a whole, emulsions are thermodynamically unstable systems because of the unfavorable contact between oil and water molecules (Section 7.2.1). Nevertheless, their interfacial properties can often be described by thermodynamics, because the ­adsorption–desorption of surface-active solutes occurs at a rate that is much faster than the time-scale of the kinetic destabilization of the overall emulsion (Hunter 1986). 5.3.1 Definition of Surface Excess Concentration To define the concentration of solute that accumulates at an interface from a thermodynamic perspective, it is convenient to assume that the interface is a smooth infinitesimally thin plane that separates two homogeneous liquids (Figure 5.1a). Initially, one has to decide precisely where this imaginary plane should be located in the system, as this location influences the value of the interfacial solute concentration (Hiemenz and Rajagopalan 1997). In the following sections, the standard convention for assigning the location of the imaginary plane (the Gibbs dividing surface) is introduced, and then the definition of the interfacial solute concentration (surface excess concentration) is given. 5.3.1.1 Gas–Liquid Interface in the Absence of Solutes For simplicity, consider a system that consists of liquid water in equilibrium with its vapor in the absence of any solutes (Figure 5.6). The volume fraction of water molecules in the liquid water is approximately unity, and decreases to approximately zero as one moves up through the interfacial region and into the vapor phase. The imaginary plane interface could be located anywhere in the interfacial region indicated in Figure 5.6. In practice, it is convenient to assume that the interface is located at a position where the excess concentration of the substance on one side of the interface is equal to the deficit concentration of the substance on the other side of the interface: cexcess = cdeficit. In this example, the excess concentration corresponds to the amount of water above the interface that exceeds that which would have been Gibbs dividing surface Height Water vapor Excess concentration Deficit concentration Liquid water Concentration FIGURE 5.6 From a thermodynamic standpoint, it is convenient to locate the interface (Gibbs dividing surface) separating a liquid and its vapor where cexcess = cdeficit. 193 Interfacial Properties and Their Characterization present if the concentration of water was the same as that in the bulk vapor phase right up to the interface. Similarly, the deficit concentration corresponds to the amount of water that is below the interface that is less than that which would have been present if the concentration of the water was the same as that in the bulk liquid phase right up to the interface. This location of the interface is known as the Gibbs dividing surface, after the scientist who first proposed this convention. 5.3.1.2 Gas–Liquid Interface in the Presence of Solutes The concept of the Gibbs dividing surface is particularly useful for defining the amount of a surfaceactive solute that accumulates at an interface (Hunter 1986). Consider a system that consists of a surfactant solution in contact with its vapor (Figure 5.7). The surface-active solute is distributed between the bulk aqueous phase, the vapor phase, and the interfacial region. The excess solute concentration at the surface (ni) corresponds to the total amount of solute present minus that which would have been present if the solute were not surface-active, and equals the shaded area shown in Figure 5.7. The accumulation of solute molecules at an interface is characterized by a surface excess concentration, Γ, which is equal to the excess solute concentration divided by the surface area: Γ = ni/A. Food emulsifiers typically have Γ values of a few mg m−2. It is important to note that the solute molecules are not actually concentrated at the Gibbs dividing surface (which is infinitely thin), because of their finite size and the possibility of multilayer formation. Nevertheless, this approach is extremely convenient for developing thermodynamic descriptions of the properties of surfaces and interfaces. The surface excess concentration is often identified with an experimentally measurable parameter called the surface load, which is the amount of emulsifier adsorbed per unit surface area (Section 4.4). 5.3.1.3 Liquid–Liquid Interfaces For an interface between pure oil and pure water, the Gibbs dividing surface could either be positioned at the point where the excess and deficit concentrations of the oil or of the water were equal on either sides of the interface, which will in general be different. For convenience, it is usually assumed that the phase in which the surface-active solute is most soluble is the one used to decide the position of the Gibbs dividing surface. The surface excess concentration of a solute is then equal to that which is present in the system minus that which would be present if there were no accumulation at the interface. Surfactant molecules Water molecules Water vapor Liquid water Height Excess concentration Gibbs dividing surface Surface excess concentration Deficit concentration Concentration FIGURE 5.7 When an emulsifier is present, the Gibbs dividing surface is conveniently located at the position where cexcess = cdeficit for the liquid in which the emulsifier is most soluble, and the surface excess concentration is equal to the shaded region. 194 Food Emulsions: Principles, Practices, and Techniques 5.3.2 Relationship between Adsorbed and Bulk Solute Concentrations Consider a system that consists of a solution containing a surface-active solute in contact with a surface (Figure 5.3). There will be an equilibrium between the solute molecules adsorbed to the surface and those present in the bulk solution. As the solute concentration in the bulk liquid is increased, so does its concentration at the surface. The presence of solute molecules at the surface reduces the thermodynamically unfavorable contacts between air and water molecules, thereby reducing the surface tension (Norde 2011). At a certain solute concentration, the surface tension reaches a constant value, because the surface becomes saturated with solute molecules. It is useful to be able to mathematically describe the relationship between the adsorbed and free solute concentrations, and to quantify how this relationship depends on the surface activity of the solute molecules. In this section, two different thermodynamic approaches that have been developed to describe this relationship are presented: the Langmuir adsorption isotherm and the Gibbs adsorption isotherm (Hunter 1986, Hiemenz and Rajagopalan 1997, Norde 2011). These approaches are based on a thermodynamic analysis of the adsorption process, assuming that the adsorption–desorption of solutes at the surface is reversible, and that solute–solute interactions do not occur in the bulk solution or at the surface. The Langmuir adsorption isotherm is useful for relating the amount of solute present at a surface to the concentration and surface activity of the solute in the bulk solution: q= c /c1/ 2 G = G ¥ 1 + c /c1/ 2 (5.3) where θ is the fraction of adsorption sites that are occupied Γ∞ is the surface excess concentration when the surface is completely saturated with solute c1/2 is the solute concentration in the bulk solution where θ = 1/2 The equilibrium constant for adsorption (K = 1/c1/2) provides a good measure for the surface activity or binding affinity of an emulsifier: the greater the 1/c1/2, the higher is the binding affinity. The surface activity of a molecule is related to the free energy of adsorption by the following equation: K= 1 æ DGads ö = exp ç ÷ c1/ 2 è RT ø (5.4) where ΔG ads corresponds to the free energy change associated with exchanging a solvent molecule with a solute molecule at the surface. Thus, the more negative is the free energy change associated with solute adsorption (ΔG ads/RT), the higher is the affinity of the solute for the surface. As mentioned earlier, the adsorption free energy has both enthalpy and entropy contributions. The enthalpy contributions are due to changes in the strength of the fundamental interaction energies associated with adsorption, for example, van der Waals, electrostatic, and steric (Chapter 2). The entropy contributions are due to differences in the number of ways the molecules can be arranged in the system in the nonadsorbed and adsorbed states, for example, configuration, orientation, conformation, and interaction entropies (Section 5.2). Attempts have been made to calculate the relative contributions of the various enthalpy and entropy contributions to ΔG ads for different kinds of emulsifiers (Norde 2011). The magnitude of the enthalpy and entropy contributions to the overall free energy of adsorption can often be determined by measuring the temperature-dependence of the surface activity of the molecules. Predictions made using the Langmuir adsorption equation for emulsifiers with low and high binding affinities are shown in Figure 5.8. The emulsifier concentration at the surface (Γ) increases approximately linearly with increasing emulsifier concentration in the bulk solution (c) at low emulsifier concentrations. At high emulsifier concentrations, Γ/Γ∞ tends toward unity, indicating saturation of the Interfacial Properties and Their Characterization 195 1 High affinity 0.8 Γ/Γ∞ 0.6 Low affinity 0.4 0.2 0 c FIGURE 5.8 Theoretical calculations of the dependence of the normalized surface excess concentration (Γ/Γ∞) on emulsifier concentration in a bulk solution for two emulsifiers with different surface activities. The surface activity of a solute can be determined from a plot of surface excess concentration (Γ) versus surfactant concentration in the bulk solution (c). The surface activity is defined as the concentration where Γ = ½Γ∞, where Γ∞ is the surface excess concentration when the interface is saturated with solute. interface. In practice, there are often appreciable deviations between predictions made by the Langmuir equation and experimental measurements, because the assumptions used in the derivation of the theory are not met for many real systems (Walstra 2003, Norde 2011). Some important sources of deviation are: (1) surfactants may form micelles in solution so that their activity coefficient is very different from their concentration; (2) emulsifiers may interact appreciably with each other at the surface, so the binding sites cannot be treated as being independent; (3) biopolymers unfold and interact at the interface so that the adsorption process cannot be considered to be reversible; and (4) solvent and emulsifier molecules have different sizes. For these reasons, a number of researchers have developed more sophisticated theories that take into account some of these effects (Fainerman et al. 1998, Karakashev et al. 2008, Rabe et al. 2011). The Gibbs adsorption isotherm is useful for relating the amount of solute present at a surface (Γ) to the surface tension (or surface pressure) and solute concentration in the bulk solution (which are both experimentally measurable): G=- 1 æ dg ö 1 æ dp ö ç ÷= ç ÷ pRT è d ln(c) ø pRT è d ln(c) ø (5.5) where c is the concentration of solute in the aqueous phase R is the gas constant T is the absolute temperature p is a parameter that depends on solute type and solution conditions For nonionic solutes, p = 1. For monovalent ionic solutes, p = 2 at low ionic strengths, p = 1 at high ionic strengths, and p = c/(c + cE) at intermediate ionic strengths, where cE is the concentration of monovalent electrolyte present in the aqueous solution (Norde 2011). The parameter p takes into account the fact that there may be counterions closely associated with ionic solutes and that these ions can also accumulate at the interface. The Gibbs adsorption isotherm can be used to determine the surface excess concentration of a solute from experimental measurements of the surface tension as a function of solute concentration, 196 Food Emulsions: Principles, Practices, and Techniques since Γ is related to the slope of a plot of γ (or π) versus ln(c) (Figure 5.4). The Gibbs adsorption isotherm can also be presented in the following form: c ò p = pRT G(c)d ln c (5.6) 0 This expression enables one to calculate the surface pressure (or tension) from knowledge of the relationship between the surface and bulk solute concentrations. Equations for Γ(c) have been derived for various systems (Norde 2011), for example, the Langmuir adsorption isotherm described earlier (Equation 5.3). Insertion of the Langmuir adsorption isotherm into the aforementioned equation and expressing the surface excess concentration in units of mass per unit area rather than moles per unit area gives æ c ö æ pRT ö p=ç ÷ G ¥ ln ç 1 + c ÷ M è ø 1/ 2 ø è (5.7) This equation is only strictly applicable at low solute concentrations, since it does not take into account solute–solute interactions in the bulk solution or at the interface. Nevertheless, it provides some valuable insights into the factors that influence the surface pressure. The surface pressure should increase as the emulsifier concentration increases (c), the surface activity increases (1/c1/2), or the surface excess concentration (Γ∞) at saturation increases. Knowledge of Γ∞ is important for formulating food emulsions, because it determines the minimum amount of emulsifier that can be used to create an emulsion with a given size distribution (Section 6.6.1). The smaller the value of Γ∞, the greater is the area of oil–water interface that can be covered per gram of emulsifier, and therefore the smaller the size of droplets that can be effectively stabilized by the same amount of emulsifier. Plots of surface tension versus emulsifier concentration are also useful, because they indicate the maximum surface pressure πmax that can be achieved when the surface is saturated by an emulsifier, which has important consequences for the formation and stability of food emulsions (Chapters 6 and 7). 5.3.3 Stipulating Interfacial Properties of Surface-Active Solutes Overall, the interfacial characteristics of a surface-active solute can be described by plotting the surface excess concentration and surface pressure (or tension) versus surfactant concentration (Figure 5.4). The interfacial characteristics of a solute can then be conveniently described in terms of three thermodynamic parameters that can be determined from these curves: 1. Surface activity: The surface activity (1/c1/2) of a solute is determined by the free energy change associated with adsorption of the solute from the bulk solution to the interface. It provides a quantitative measure of the affinity of a solute molecule for the interface: the higher is 1/c1/2, the greater is the surface activity. 2. Saturation surface pressure: The surface pressure of the interface when it is saturated with solute molecules (Π∞) is determined by how efficient the solute molecules are at minimizing the thermodynamically unfavorable contacts at the interface. It therefore depends on the packing of the solute molecules at the interface, as well as their interactions with the other molecules present there, for example, oil and water. 3. Saturation surface excess concentration: The surface excess concentration at the interface when it is saturated with solute (Γ∞) is determined by the mass of the individual solute molecules as well as how efficiently they can pack at the interface. The aforementioned parameters are derived assuming that the adsorption–desorption process is reversible and that there are no solute–solute interactions in the bulk solution or at the interface. In practice, 197 Interfacial Properties and Their Characterization a thermodynamic interpretation of these parameters may therefore be invalid for many real systems, because these assumptions are not met. Nevertheless, these parameters still provide a useful means of characterizing and comparing the interfacial properties of surface-active solutes in terms of experimentally measurable quantities. The interfacial properties of a typical small molecule surfactant and a typical protein are compared in Figure 5.9. The protein has a much higher surface activity than the surfactant (lower c1/2), but the surfactant can lower the interfacial tension appreciably more at saturation than the protein. This means that proteins adsorb to interfaces at much lower bulk emulsifier concentrations than do surfactants, but that surfactants are effective at displacing proteins from interfaces at sufficiently high concentrations. 5.3.4 Adsorption Kinetics The rate at which an emulsifier adsorbs to an interface is one of the most important factors determining its efficacy as a food ingredient, since it influences the size of the droplets produced during homogenization (Section 6.4). The adsorption rate depends on the molecular characteristics of the emulsifier (e.g., size, flexibility, conformation, and interactions), the nature of the bulk liquids it is dispersed in (e.g., viscosity, polarity), and the prevailing environmental conditions (e.g., temperature and fluid flow profile). It is often convenient to divide the adsorption process into different stages, such as movement of the emulsifier molecules from the bulk liquid to the vicinity of the interface, followed by attachment of the emulsifier molecules to the interface (Figure 5.10). In practice, emulsifier molecules are often in a dynamic equilibrium between the adsorbed and nonadsorbed states, and so the rate at which emulsifier molecules leave the interface must also be considered when calculating the overall adsorption rate (Norde 2011). 5.3.4.1 Movement of Molecules to the Vicinity of an Interface In this section, it is assumed that an emulsifier molecule is adsorbed to an interface as soon as it encounters it, that is, there are no energy barriers that retard adsorption. In an isothermal quiescent liquid, 30 Surfactant 25 π (mN m–1) 20 Protein 15 10 5 0 10–6 10–4 10–2 cbulk (mol m–3) 100 102 FIGURE 5.9 Comparison of the affinity of amphiphilic biopolymers and small molecule surfactants for an oil–water interface. Biopolymers tend to have a higher surface activity and therefore saturate the interface at lower concentrations, but surfactants are usually more effective at minimizing unfavorable interactions at the interface and therefore have higher π∞. 198 Food Emulsions: Principles, Practices, and Techniques (a) (b) FIGURE 5.10 The adsorption of a surface-active solute at an interface can be divided into a number of steps: (a) movement to the vicinity of the interface, (b) attachment to the interface; postadsorption conformational changes or interactions. emulsifier molecules move from a bulk liquid to an interface by molecular diffusion with an initial adsorption rate given by (Norde 2011) dG D =c dt pt (5.8) where D is the translational diffusion coefficient of the emulsifier Γ is the surface excess concentration t is the time c is the concentration of emulsifier initially present in the bulk liquid The variation of the surface excess concentration with time is obtained by integrating this equation with respect to time: G(t ) = 2c Dt p (5.9) Thus, a plot of the surface excess concentration versus √t should be a straight line that passes through the origin. This equation indicates that the accumulation of an emulsifier at an interface occurs more rapidly as the concentration of emulsifier in the bulk liquid increases or as the diffusion coefficient of the emulsifier through the bulk liquid increases. The diffusion coefficient increases as the size of molecules decreases, and one would therefore expect smaller molecules to adsorb more rapidly than larger ones. This equation typically gives a good description of the early stages of adsorption to clean interfaces under diffusion-controlled conditions, but is unsuitable for describing the later stages of adsorption, because the interface becomes saturated with emulsifier molecules, and therefore, there are less sites available for the emulsifier to adsorb to (Figure 5.11). Nevertheless, mathematical models have been developed to describe the change in interfacial concentration with time at the later stages of adsorption (Norde 2011, Rabe et al. 2011). Models have been developed that take into account solute–solute interactions at the interface, the orientation of the solute molecules, and postadsorption conformational changes of solutes. In practice, the initial rate is often faster than that given by Equation 5.9 because of convection currents caused by temperature gradients within a liquid. Consequently, considerable care must be taken to ensure that the temperature within a sample is uniform when measuring diffusion-controlled adsorption processes. 199 Interfacial Properties and Their Characterization 1.8 1.6 1.4 1.2 Γ (mg m–2) 1 0.8 0.6 0.4 0.2 0 0 50 100 t1/2 150 200 (s1/2) FIGURE 5.11 Typical example of adsorption kinetics for a diffusion-controlled system. The surface excess concentration increases with time as emulsifier molecules accumulate at the interface. The aforementioned equations do not apply during the homogenization of emulsions, because homogenization is a highly dynamic process and mass transport is governed mainly by convection rather than diffusion (Walstra 2003). Under isotropic turbulent conditions, the initial increase of the surface excess concentration with time is given by (Dukhin et al. 1995) 3 æ r ö G(t ) = Crd c ç 1 + e ÷ t r d ø è (5.10) where C is a constant that depends on the experimental conditions rd and re are the radii of the droplet and emulsifier, respectively This equation predicts that the adsorption rate increases as the concentration of emulsifier increases, the size of the emulsion droplets increases, or the size of the emulsifier molecules increases relative to the size of the droplets. This equation implies that when an emulsion is homogenized, the emulsifier molecules initially adsorb preferentially to the larger droplets, and that larger emulsifier molecules tend to adsorb more rapidly than smaller ones (which is the opposite to diffusion-controlled adsorption). This explains why large casein micelles adsorb faster than individual casein molecules during the homogenization of milk (Mulder and Walstra 1974). 5.3.4.2 Attachment of Emulsifier Molecules to Interface So far it has been assumed that as soon as an emulsifier molecule reaches an interface, it is immediately adsorbed. In practice, there may be one or more energy barriers that must be overcome before a 200 Food Emulsions: Principles, Practices, and Techniques molecule actually adsorbs, and so only a fraction of the encounters between an emulsifier molecule and an interface leads to adsorption (Norde 2011, Rabe et al. 2011). In these systems, adsorption kinetics may be governed by the height of the energy barrier, rather than by the rate at which the molecules reach the interface. There are a number of reasons that an energy barrier to adsorption may exist: 1. As an emulsifier molecule approaches an interface, there may be various types of repulsive interaction between it and the emulsifier molecules already adsorbed to the interface, for example, electrostatic, steric, hydration, or thermal fluctuation (Chapter 3). 2. Some surface-active molecules will only be adsorbed if they are in a specific orientation when they encounter the interface. For example, globular proteins with hydrophobic patches on their surface may have to have these patches facing toward the oil phase during an encounter. 3. The ability of surfactant molecules to form micelles plays a major role in determining their adsorption kinetics (Dukhin et al. 1995). Surfactant monomers are surface active, because they have a polar head group and a nonpolar tail, but micelles are not surface active, because their exterior is surrounded by hydrophilic head groups. The adsorption kinetics therefore depends on the concentration and mass transfer coefficients of monomers and micelles present, as well as the dynamics of micelle assembly–disassembly. The adsorption of emulsifier molecules at a surface or interface can be measured by using a variety of experimental methods (Couper 1993, Dukhin et al. 1995, Milling 1999, Norde 2011). The most commonly used is to measure the variation in surface or interfacial tension with time using a tensiometer (Section 5.7). A number of workers have also used radio-labeled emulsifier molecules to measure adsorption kinetics. A radioactivity detector is placed immediately above a water–air surface. The radio-labeled emulsifier is injected into the water and the increase in the radioactivity at the surface is recorded over time by the detector. The radioactivity is highly attenuated by the water, and so only those molecules that are close to the air–water surface are detected. A variety of other experimental methods that can detect changes in interfacial properties due to the adsorption of emulsifier molecules have also been used to monitor adsorption kinetics, including spectroscopy, scattering and reflection of various forms of radiation, mechanical methods, and interfacial rheology (Sections 5.6 through 5.8). 5.4 Electrical Characteristics of Interfaces 5.4.1 Origin of Interfacial Charge The droplets in most food emulsions have an appreciable electrical charge, and therefore, electrostatic interactions may play an important role in determining their overall stability and physicochemical properties. Oil droplets in emulsifier-free oil-in-water emulsions have been shown to have an electrical charge that depends on both pH and ionic strength. For example, the ζ-potential of sunflower oil droplets dispersed in aqueous solutions goes from positive at low pH (<5) to increasingly negative at higher pH (Hsu and Nacu 2003). Similar trends have been observed for hydrocarbon droplets dispersed in water (Pashley 2003). The origin of this effect has been attributed to preferential adsorption of either H3O+ (low pH) or OH− (high pH) species from the water onto the droplet surfaces. In addition, there may be surface-active ionic impurities present in many commercial oils, such as free fatty acids. The droplets in food emulsions are normally stabilized by emulsifiers, and so their electrical characteristics are largely determined by the characteristics of the interfacial layer of adsorbed emulsifier molecules. The emulsifiers used in foods may be anionic, cationic, or nonionic, depending on their molecular characteristics and the prevailing environmental conditions (e.g., pH, ionic strength, and temperature). Contrary to expectations, emulsion droplets stabilized by nonionic surfactants may have an appreciable electrical charge. For example, sunflower oil droplets stabilized by Tweens have been shown to have positive charges at low pH (<4) and negative charges at higher pH (Hsu and Nacu 2003). Hence, they follow similar trends to bare oil droplets, but they tend to have more negative charges at the same pH, which causes their isoelectric points to shift to lower pH values. This phenomenon has been attributed to the Interfacial Properties and Their Characterization 201 preferential adsorption of OH− ions to the hydrophilic head groups of the surfactants, but it may also be due to surface-active ionic impurities in the oil or surfactant ingredients, such as free fatty acids. The electrical properties of oil droplets stabilized by nonionic surfactants therefore seem to be dominated by the electrical characteristics of the bare droplets, but are modified somewhat by the presence of the interfacial layer of adsorbed surfactant molecules. Many types of commonly used food emulsifiers are either ionic or capable of being ionized, for example, proteins, polysaccharides, and surfactants (Section 4.4). All food proteins have acidic (–COOH → COO − + H+) and basic (NH2 + H+ → NH3+) groups whose degree of ionization depends on the pH and ionic strength of the surrounding aqueous phase. Some surface-active polysaccharides, such as modified starch and gum arabic, also have acidic groups that may be ionized. Ionic surfactants may be either positively or negatively charged, depending on the nature of their hydrophilic head group (Section 4.4). The magnitude and sign of the electrical charge on an emulsion droplet therefore depends on the type of emulsifier used to stabilize it, the concentration of the emulsifier at the interface, and the prevailing environmental conditions (e.g., pH, temperature, and ionic strength). All the droplets in an emulsion are usually stabilized by the same type of emulsifier and therefore have the same electrical charge. The electrostatic interaction between similarly charged droplets is repulsive, and so electrostatic interactions play a major role in preventing droplets from coming close enough together to aggregate (Section 3.4). The sign and magnitude of the electrical charge on emulsion droplets plays an important role in determining the stability and physicochemical properties of food emulsions. For example, the electrical characteristics of the interface influence the magnitude and range of the colloidal interactions between emulsion droplets, as well as the tendency for various types of electrically charged species to accumulate at the droplet surfaces, for example, biopolymers, antioxidants, pro-oxidants, and flavors. The electrical properties of an interface are usually characterized by the surface charge density (σ) and the electrical surface potential (Ψ0) (Hunter 1986, Israelachvili 2011). The surface charge density is the amount of electrical charge per unit interfacial area, whereas the surface potential is the amount of free energy required to increase the surface charge density from 0 to σ. The surface charge density is governed by the type and concentration of ionizable surface-active molecules adsorbed to the interface, as well as by the characteristics of the solution surrounding the interface (e.g., ion concentration, ion type, and dielectric constant) and the prevailing environmental conditions (e.g., temperature). The various kinds of ionic species present in food emulsions that influence the electrical properties of interfaces can be conveniently divided into three categories (Hunter 1986): 1. Potential determining ions: This type of ion is responsible for the association–dissociation of charged groups, for example, –COOH → COO − + H+. In food emulsions, the most important potential determining ions (PDIs) are H+ and OH−, because they govern the degree of ionization of acidic and basic groups on many proteins, polysaccharides, and phospholipids. The influence of PDIs on surface charge is therefore determined principally by the pH of the aqueous phase relative to the pKa values of the ionizable surface groups. 2. Indifferent electrolyte ions: This type of ion accumulates around charged groups because of attractive electrostatic interactions: for example, Na+ ions may accumulate around a negatively charged –COO − group. These ions reduce the strength of the electrical field around a charged group principally due to electrostatic screening, rather than by altering the surface charge density. Nevertheless, at sufficiently high ionic strengths, some indifferent electrolyte ions may alter the degree of ionization of charged groups, either by altering their dissociation constants (i.e., their pKa values) or by competing with H+ or OH− ions (e.g., –COO − + Na+ → –COO −Na+). The influence of indifferent electrolyte ions on surface charge is therefore determined principally by their effect on the ionic strength of the surrounding solution, and is usually characterized by the Debye screening length (see Section 5.4.2). 3. Adsorbed ions: The electrical characteristics of the interface can also be altered by adsorption of ions to the interface, which changes the surface charge density. In food emulsions, the most important types of adsorbed ions are ionic emulsifiers (e.g., many surfactants, proteins and polysaccharides) and polyvalent ions (e.g., multivalent mineral ions and polyelectrolytes) 202 Food Emulsions: Principles, Practices, and Techniques (Chapter 4). The primary driving force for the adsorption of ionic emulsifiers to an interface is often the hydrophobic effect, whereas the major driving force for the adsorption of polyvalent ions is usually electrostatic attraction. The contribution of adsorbed ions to surface charge is governed mainly by their type and concentration in the overall system, and their relative affinities for the interface. Food scientists are particularly interested in understanding the role that each of these different types of ions plays in determining the overall properties and stability of food emulsions. 5.4.2 Ion Distribution near a Charged Interface The development of mathematical models to describe the electrical properties of interfaces relies on an appreciation of the way that ions are organized close to charged surfaces (Evans and Wennerstrom 1999, Norde 2011). Consider a charged surface that is in contact with an electrolyte solution (Figure 5.12). Ions of opposite charge to the surface (counterions) are attracted toward it, whereas ions of similar charge (co-ions) are repelled from it. Nevertheless, the tendency for ions to be organized in the vicinity of a charged surface is opposed by the disorganizing influence of the thermal energy. Consequently, the concentration of counterions is greatest at the charged surface, and decreases as one moves away from the surface until it reaches the bulk counterion concentration, whereas the concentration of co-ions is smallest at the charged surface, and increases as one moves away from the surface until it reaches the bulk coion concentration (Figure 5.12). The concentration of counterions near a charged surface is always greater than the concentration of co-ions, and so a charged surface can be considered to be surrounded by a cloud of counter-ions. Nevertheless, the overall system must be electrically neutral, and Electrolyte solution Charged surface + – + – + + – + + + – + – + + – – + Counter-ions – + – – + + + – + – – – + – + – – + + – – + Co-ions FIGURE 5.12 The organization of ions near a charged surface is governed by two opposing tendencies: (1) electrostatic interactions favor accumulation of counterions near an oppositely charged surface and (2) thermal energy favors a random distribution of ions. 203 Interfacial Properties and Their Characterization so the charge on the surface must be completely balanced by the excess charge of the counterions in the surrounding electrolyte solution. The distribution of ions close to a charged surface is referred to as the electrical double layer, because it is convenient to assume that the system consists of two oppositely charged layers. The first layer is the charged surface itself, and the second layer is the neutralizing layer of counterions in the liquid in contact with the surface. The effective thickness of the second layer is determined primarily by the ionic composition of the liquid, decreasing with increasing counterion valency and concentration (see Section 5.4.2.2). It is possible to develop mathematical equations to predict the distribution of ions in the immediate vicinity of a charged surface from knowledge of the electrical characteristics of the surfaces and the properties of the solution in contact with it (Hunter 1986, Evans and Wennerstrom 1999, Israelachvili 2011). The electrical properties of the surface are usually characterized in terms of the surface charge density (σ) and electrical surface potential (Ψ0), whereas the properties of the solution are usually characterized in terms of ion concentration, ion type, and dielectric constant. A mathematical relationship, known as the Poisson–Boltzmann equation, has been derived to relate the electrical potential in the vicinity of a charged surface to the concentration and type of ions present in the adjacent electrolyte solution: d 2 y( x ) e =2 dx e0e R åz n i 0i i æ - z ey( x ) ö exp ç i ÷ è kT ø (5.11) where n 0i is the concentration of ionic species of type i in the bulk electrolyte solution (in molecules per cubic meter) zi is their valency e is the elementary charge (1.602 × 10 −19 C) ε0 is the dielectric constant of a vacuum εR is the relative dielectric constant of the solution Ψ(x) is the electrical potential at a distance x from the charged surface This equation is of central importance to emulsion science, because it is the basis for the calculation of electrostatic interactions between emulsion droplets (Chapter 3). Nevertheless, its widespread application has been limited, because it does not have an explicit analytical solution. When accurate calculations are required, it is necessary to solve the Poisson–Boltzmann numerically using a digital computer. For certain systems, it is possible to derive much simpler analytical equations that can be used to calculate the electrical potential near a surface by making certain simplifying assumptions, for example, that the electrolyte ions are symmetrical (valency 1:1, 2:2, etc.) or that the surface charge is not too high. If it is assumed that the electrostatic attraction between the charged surface and the counterions is relatively weak compared to the thermal energy, that is, zieΨ0 < kT (which means that |Ψ0| must be less than about 25 mV at room temperature in water), then a simple expression, known as the Debye– Huckel approximation, can be used to calculate the dependence of the electrical potential on distance from the surface: Ψ(x) = Ψ0 exp(−κx) (5.12) This equation indicates that the electrical potential decreases exponentially with distance from the surface at a decay rate that is determined by the parameter, κ−1, that is known as the Debye screening length. The Debye screening length is a measure of the “thickness” of the electrical double layer, and it is related to the properties of the electrolyte solution by the following equation: k -1 = e0e r kT e 2 å n0 i zi2 (5.13) 204 Food Emulsions: Principles, Practices, and Techniques For aqueous solutions at room temperatures, κ−1 ≈ 0.304/√I nm, where I is the ionic strength expressed in moles per liter (Israelachvili 2011). For example, the Debye screening lengths for NaCl solutions with different ionic strengths are: 0.3 nm for a 1 M solution, 0.96 nm for a 100 mM solution, 3 nm for a 10 mM solution, 9.6 nm for a 1 mM solution, and 30.4 for a 0.1 mM solution. The Debye screening length is an extremely important characteristic of an electrolyte solution, because it determines how rapidly the electrical potential decreases with distance from the surface. Physically, κ−1 corresponds to the distance from the charged surface where the electrical potential has fallen to 1/e of its value at the surface. This distance is particularly sensitive to the concentration and valency of the ions in an electrolyte solution. As the ion concentration or valency increases, κ−1 becomes smaller, and therefore, the electrical potential decreases more rapidly with distance (Figure 5.13). The physical explanation for this phenomenon is that the neutralization of the surface charge occurs at shorter distances when the concentration of opposite charge in the surrounding solution increases. As will be seen in later chapters, the screening of electrostatic interactions by electrolytes has important consequences for the stability and rheology of many food emulsions, because it can promote droplet flocculation (Chapters 7 and 8). The Poisson–Boltzmann theory (and hence the Debye–Huckel approximation) assumes that an electrolyte solution is a continuum that contains ions that are infinitesimally small. It therefore allows ions to accumulate at unrealistically large concentrations near to a charged surface (Evans and Wennerstrom 1999). In reality, ions have a finite size and shape, and this limits the number of them that can be present in the first layer of molecules that are in direct contact with the surface (Israelachvili 2011). This assumption is not particularly limiting for systems in which there is a weak interaction between the ions and the charged surface. Nevertheless, it becomes increasingly unrealistic as the strength of the electrostatic interactions between a charged surface and the surrounding ions increases relative to the thermal energy. In these cases, the Poisson–Boltzmann theory must be modified to take into 25 Debye screening length ψ0 (mV) 20 Particle 15 0.1 mM 10 5 1 mM 1000 mM 0 100 mM 0 10 mM 10 Distance, x (nm) 20 30 FIGURE 5.13 Influence of ionic strength on the electric field near a charged surface. The electrical double layer shrinks as the electrolyte concentration increases or the ion valency increases. 205 Interfacial Properties and Their Characterization account the finite size of the ions in the electrolyte solution. Conceptually, it is convenient to divide the distribution of counterions near a highly charged surface into two regimes: an inner region and an outer region (Figure 5.14). 5.4.2.1 Inner Region In the inner region, the attraction between the counterions and charged surface is relatively strong and therefore, the counterions are relatively immobile, whereas in the outer region, the attraction is much weaker and therefore, the counterions are more mobile. The thickness of the inner region, δ, is approximately equal to the radius of the hydrated counterions, rather than their diameter, because the effective charge of an ion is located at its center. The inner region is sometimes referred to as the Stern layer, while the boundary between the inner and outer regions is referred to as the Stern plane (Figure 5.14). The electrical potential at the Stern plane (Ψδ) is different from that at the surface (Ψ0) because of the presence of the counterions in the Stern layer. For monovalent indifferent electrolyte counterions, Ψδ is less than Ψ0, because the surface charge is partly neutralized by the charge on the counterions (Figure 5.15a). The extent of this decrease depends on the number and packing of the counterions within the Stern plane. The same behavior is observed for polyvalent counterions at low concentrations (such as multivalent mineral ions and polyelectrolytes), but at higher concentrations, the surface may adsorb such a large number of oppositely charged multivalent counterions that its charge is actually reversed, so that Ψδ has an opposite sign to Ψ0 (Figure 5.15b). If a charged surface adsorbs surface-active co-ions (e.g., ionic emulsifiers), it is even possible for Ψδ to be larger than Ψ0 (Figure 5.15c). An increase in surface charge may occur when the hydrophobic attraction between the nonpolar tail of a surfactant and a surface is greater than any electrostatic repulsive interactions. In real food emulsions, there is usually a mixture of different types of ions present in the aqueous phase in contact with the charged surfaces, and these ions may all compete with each other at the surface. The magnitude of the electrical potential at the Stern plane therefore depends on the precise type and concentration of ions present in the system. Ψ0 Ψδ ζ – + + – + – + + – – + – – – + + + – – Stern plane Shear plane FIGURE 5.14 When the electrostatic attraction between a charged surface and the surrounding counterions is relatively strong compared to the thermal energy, it is convenient to divide the electrolyte solution into an inner and an outer region. 206 Food Emulsions: Principles, Practices, and Techniques + + – + + + + – – – + + – – 2– + + + – + 2– 2– + + + 2– x + + + 2– + + + + + – – + – – Ψ(x) Ψ(x) Ψ(x) (a) + 2– x x (b) (c) FIGURE 5.15 The electrical potential at the Stern plane may be (a) lower, (b) a different sign, or (c) higher than that at the bare surface depending on the strength of the interaction and the type of ions adsorbed. A number of theories have been developed to take into account the effect of the finite size and limited packing of ions in the Stern layer on the relationship between Ψδ and Ψ0 (Hiemenz and Rajagopalan 1997). One of the most widely used is the Langmuir adsorption isotherm mentioned earlier (Section 5.3), which assumes that there are only a finite number of binding sites at the surface, and that once these are filled, the surface becomes saturated and cannot adsorb any more ions. In this case, the ion adsorption free energy is given by ΔGads = −(zeΨδ + ϕ). The zeΨδ term is due to the electrostatic attraction between the ion and the surface, while the ϕ term accounts for any specific binding effects. These specific binding effects could be due to hydrophobic interactions (e.g., when an emulsifier adsorbs) or due to chemical interactions (e.g., –COO − + Na+ → –COO −Na+). The fraction of surface sites that are occupied increases as the free energy of adsorption of an ion increases. The electrical potential at the Stern layer can be related to the electrical potential at the charged surface using the following equation: yd = y0 - dqs * ede0 (5.14) where σ* is the surface charge density when the surface is completely saturated with ions δ is the thickness of the Stern layer θ is the fraction of ion binding sites that are occupied εδ is the relative dielectric constant of the Stern layer This equation indicates that the difference between the potential at the surface and that at the Stern plane depends on the fraction of surface sites that are occupied. In principle, this equation can be used to calculate the change in the electrical potential of a surface due to ion adsorption. In practice, this equation is difficult to use because of a lack of knowledge about the values of δ, ϕ, and εδ in the Stern layer. These parameters are unique for every ion–surface combination and are difficult to measure experimentally. For this reason, it is usually more convenient to experimentally measure the electrical potential at the Stern plane, rather than attempting to predict it theoretically (Hunter 1986). Interfacial Properties and Their Characterization 207 Experiments have shown that Ψδ is closely related to the electrical potential at the shear plane (Hunter 1986). When a liquid flows past a charged surface, it pulls those counterions that are only weakly attached to the surface along with it, but leaves those ions that are strongly attached in place, that is, those ions in the Stern plane. The shear plane is defined as the distance from the charged surface below which the counterions remain strongly attached and is approximately equal to the diameter of the hydrated ions (Figure 5.14). The electrical potential at the shear plane is referred to as the zeta-potential (ζ), and can be measured using various types of electrokinetic techniques (Chapter 14). 5.4.2.2 Outer Region In the outer region, the electrostatic interaction between the surface and the counterions is fairly weak (because the ions in the Stern layer partially screen the surface charge), and so the variation in electrical potential with distance can be described by replacing Ψ0 with Ψδ in Equation 5.12: Ψ(x) = Ψδ exp(−κx) (5.15) where x is now taken to be the distance from the shear plane, rather than from the charged surface. The dependence of the electrical potential on distance from the shear plane can then be calculated once Ψδ is known. An appreciation of the factors that influence the distribution of ions near an electrically charged interface is important for understanding the interactions between charged emulsion droplets and the accumulation of ions near droplet surfaces. 5.4.3 Factors Influencing Interfacial Electrical Properties of Emulsions The most important factor influencing the sign and magnitude of the electrical charge on emulsion droplets is the type of emulsifier used to stabilize them (Section 5.4.1). Many types of commonly used food emulsifiers are either ionic or capable of being ionized, for example, proteins, polysaccharides, phospholipids, and ionic surfactants (Chapter 4). The electrical characteristics of the interfaces formed by these emulsifiers depend on the number and type of ionizable groups present on the molecules. In particular, the pKa values of the charged groups are particularly important for determining their extent of ionization at a particular pH. All food proteins have acidic (e.g., –COOH → COO − + H+, pKa ≈ 4-5; −SH → S− + H+, pKa ≈ 8.5; –OH → O − + H+, pKa ≈ 10) and basic (e.g., –C2N2H2 + H+ → –C2N2H3+, pKa ≈ 8.5; NH2 + H+ → NH3+, pKa ≈ 10–12) groups whose degree of ionization depends on the pH and ionic strength of the surrounding aqueous phase (Damodaran et al. 2007, Belitz et al. 2009, Brady 2013). Proteins are positively charged at pH values below their isoelectric point (pI), have no net charge at their pI, and are negatively charged above their pI. This has important consequences for the pH stability of protein-stabilized emulsions, since the electrostatic repulsion between droplets is much smaller near the pI than at higher or lower pH values (Chapter 3). Consequently, protein-stabilized emulsions usually exhibit a high degree of droplet flocculation at pH values near their isoelectric points (Chapter 7). Some surface-active polysaccharides, such as modified starch and gum arabic, also have acidic groups that may be ionized (Damodaran et al. 2007, Belitz et al. 2009, Brady 2013). Ionic surfactants may be either positively or negatively charged, depending on the nature of their hydrophilic head group, although most ionic surfactants used in foods are negatively charged (Kralova and Sjoblom 2009). As mentioned earlier, even droplets stabilized by nonionic surfactants may have an appreciable electrical charge (Section 5.4.1). In addition to the adsorption of charged emulsifiers, the surface charge density of emulsion droplets is also determined by adsorption of other types of ionic substances present in either the continuous and/or dispersed phases to the droplet surfaces, such as multivalent mineral ions (e.g., Ca2+, Cu2+, Fe3+) or polyelectrolytes (e.g., proteins or polysaccharides). Adsorption of these ionic substances can alter the magnitude of the electrical charge on the emulsion droplets, change the sign of the charge on the droplets, or act as bridges between charged droplets. The electrical potential (or ζ-potential) of emulsion droplets is also reduced by the presence of indifferent ions that do not specifically bind to the droplet surface but increase the ionic strength of the surrounding liquid, for example, salts. 208 Food Emulsions: Principles, Practices, and Techniques 5.4.4 Characterization of Interfacial Electrical Properties A variety of analytical techniques are available for characterizing the electrical properties of interfaces. The electrical charge at an interface can often be determined by titrating PDIs into the system and calculating the concentration of these ions that adsorb to the interface: PDIadsorbed = PDItotal − PDI free (Norde 2011). For example, the PDIs for many food emulsifiers are H+ and OH−; thus, the concentration of adsorbed PDIs can be monitored by measuring the change in pH as the concentration of acid or base added to the system is changed. To determine the absolute interfacial charge using this technique, it is necessary to define a reference point of known charge, which is usually taken to be the point of zero charge. The sign and magnitude of the electrical charge on an interface can also be determined using various electrokinetic techniques, for example, electro-osmosis, electrophoresis, streaming potential, and streaming current (Norde 2011). The particle electrophoresis technique is the most commonly used technique for characterizing the electrical properties of the droplet interfaces in emulsions. This technique measures the direction and velocity of particle movement in a well-defined electric field and then uses a mathematical model to determine the ζ-potential of the particles (Chapter 14). 5.5 Interfacial Composition and Its Characterization 5.5.1 Factors Influencing Interfacial Composition The physicochemical properties of emulsions, such as their ease of formation, stability, and texture, are governed by the nature of the interface, and therefore, it is important for food scientists to understand the factors that determine the composition of the interfacial region. If an emulsion is prepared using a single type of emulsifier, then the interfacial layer will be comprised of this emulsifier (Figure 5.16). Even so, the amount of emulsifier that adsorbs to the droplet surfaces (i.e., the surface load) may depend on a number of factors, including the initial emulsifier concentration, temperature, pH, ionic strength, and homogenization conditions. For example, the surface load increases as the initial emulsifier concentration in the aqueous phase increases for some globular protein–stabilized emulsions (Dalgleish 1997). This phenomenon has been attributed to the fact that at high protein concentrations, adsorption is relatively rapid, so that adsorbed protein molecules are unable to undergo extensive conformational changes and can therefore pack more efficiently (Norde 2011). The surface load may also increase at relatively high protein concentrations due to the formation of multiple layers of adsorbed proteins, rather than a single layer. The influence of temperature on the surface load depends on the emulsifier type and solution (a) (d) (b) (e) (c) (f) FIGURE 5.16 The orientation and conformation of surface-active solutes at an interface is determined by their tendency to reduce the free energy of the system. The interfacial composition and structure depends on the type, concentration, and surface activity of the different surface-active molecules present in the system. (a) Surfactant, (b) globular biopolymer, (c) mixed layer, (d) particulate matter, (e) flexible biopolymer, and (f) multiple layers. Interfacial Properties and Their Characterization 209 conditions (e.g., pH and ionic strength). The surface load of some globular proteins has been found to increase with temperature, presumably because the proteins partially unfold, which increases their surface hydrophobicity (Dickinson and Hong 1994). The surface load of some protein-stabilized emulsions increased when the pH is adjusted towards the protein’s isoelectric point or when the salt concentration is increased, presumably because the electrostatic repulsion between adsorbed and nonadsorbed proteins was reduced (Dalgleish 1997). Rather than containing a single type of surface-active substance, many food emulsions contain mixtures of different surface-active components, and so the interfacial layers surrounding the droplets may be compositionally complex (Figure 5.16). In this case, the interfacial composition is determined by the concentrations of the various kinds of surface-active substances present, their relative affinity for the interface, the method used to prepare the emulsion, the solution conditions (e.g., temperature, pH, and ionic strength), and the history of the emulsion (e.g., the order in which the emulsifiers were added). Some of the most important factors influencing the interfacial composition of food emulsions containing a mixture of different emulsifiers are discussed in the following. In a mixed emulsifier system, the interfacial composition depends on the relative adsorption rates of the different types of emulsifier when the bulk solution is brought into contact with the interface, as well as on any subsequent changes that occur during storage (Dickinson 1992a,b, Wilde et al. 2004). The relative adsorption rate of the emulsifiers depends on their molecular characteristics (e.g., size, shape, and polarity) and the flow profile within the bulk solution when it is brought into contact with the interface (e.g., static, laminar, or turbulent) (Section 5.3). For the case of food emulsions prepared by homogenization, the droplets will tend to be initially coated by those surface-active molecules that absorb to the interface most rapidly under turbulent conditions. Nevertheless, the interfacial composition may change during storage, because some of the surface-active molecules that were initially present at the droplet surface are displaced by molecules in the bulk liquid that have a greater affinity for the surface. Alternatively, additional surface-active components may be added to the continuous phase of an emulsion after homogenization, and these may displace some of the original emulsifier molecules from the droplet surface. For example, small molecule surfactants are often added to ice-cream premixes so as to displace the proteins from the surface of the milk fat globules prior to cooling and shearing (Goff and Hartel 2013). This causes the droplets to become more susceptible to partial coalescence, which leads to the formation of a network of aggregated droplets that stabilizes the air bubbles and gives the final product its characteristic texture and shelf-life. It is possible to relate the composition of an interface to the concentrations and surface activities of the emulsifier molecules present in a mixed emulsifier system by modifying the Langmuir adsorption isotherm (Razumovsky and Damodaran 2001): G1 c1 /c1,1/ 2 = G ¥,1 1 + c1 /c1,1/ 2 + c2 /c2,1/ 2 G2 c2 /c2,1/ 2 = G ¥,2 1 + c1 /c1,1/ 2 + c2 /c2,1/ 2 (5.16) (5.17) where c1 and c2 are the concentrations of the two types of emulsifiers in the bulk solution Γ1 and Γ2 are the surface excess concentrations of the emulsifiers at the interface Γ1,∞ and Γ2,∞ are the surface excess concentrations at saturation c1,1/2 and c2,1/2 are the emulsifier concentrations where Γ1/Γ1,∞ and Γ2/Γ2,∞ = ½ This equation assumes that the solvent molecules and the two types of solute molecules all have the same dimensions. These equations can be rearranged to give 210 Food Emulsions: Principles, Practices, and Techniques c1 /c1,1/ 2 G1 = G TOT c1 /c1,1/ 2 + c2 /c2,1/ 2 (5.18) c2 /c2,1/ 2 G2 = G TOT c1 /c1,1/ 2 + c2 /c2,1/ 2 (5.19) Hence, the composition of the interface depends on the concentrations of the two different types of emulsifiers, as well as their relative affinities for the interface (i.e., their surface activities, 1/c1/2). The change in interfacial composition when a more surface-active emulsifier (1/c1/2 = 1) is added to a solution containing a fixed concentration of a less surface-active emulsifier (1/c1/2 = ½) is shown in Figure 5.17. Initially, the interface is comprised entirely of Type 1 emulsifier, but as the concentration of Type 2 emulsifier is increased in the bulk solution, then some of the Type 1 emulsifier is displaced. In practice, the earlier approach usually has to be modified somewhat, because the assumptions used in its derivation are not appropriate. For example, there may be interactions between the emulsifiers (either in the bulk solution and/or at the interfacial region) or the adsorption–desorption process may be irreversible due to conformational changes of one or both of the emulsifiers after adsorption. It should be noted that the relative affinities of emulsifiers for an interface are influenced by solution and environmental conditions, such as temperature, pH, and ionic strength. In addition, the phase in which a surfactant is most soluble also determines its effectiveness at displacing proteins from an interface. For example, water-soluble surfactants have been shown to be more effective at displacing proteins from the surface of oil droplets than oil-soluble surfactants (Dickinson et al. 1993). The interfacial composition of an emulsion containing different types of emulsifier may also depend on droplet size, since this determines the total amount of interfacial area available for the emulsifier molecules to adsorb to. If one of the emulsifiers is present at a relatively low concentration, then it may make up a substantial 1 Γ1/ΓTOT Γ/ΓTOT 0.8 0.6 0.4 0.2 Γ2/ΓTOT 0 0 5 10 c2 15 20 FIGURE 5.17 Theoretical calculations of the dependence of interfacial composition on the composition of a bulk solution containing two different types of surface-active solute (Type 1 and 2). Initially, 2 wt% of Type 1 molecule (1/c1/2 = 0.5) is present in the bulk solution, then increasing amounts of Type 2 molecule (1/c1/2 = 1) are added to the bulk solution. The Type 1 molecules are progressively displaced from the interface. Interfacial Properties and Their Characterization 211 fraction of the interfacial layer when the droplets are relatively large (small total surface area), but only a minor fraction when the droplets are relatively small (large total surface area) (Walstra 2003). The displacement of emulsifier molecules from an interface may be retarded if they are capable of undergoing some form of conformational change that enables them to bind strongly to their neighbors. Some globular proteins become surface-denatured after adsorption to an interface because of the change in their molecular environment (Norde 2011). When these proteins unfold, they expose amino acids that are capable of forming disulfide bonds with their neighbors, and thus form an interfacial layer that is partly stabilized by covalent bonds. This accounts for the experimental observation that the ease at which β-lactoglobulin can be displaced from the surface of oil droplets decreases as the emulsion ages (Dalgleish 1996). This effect is even more pronounced in globular protein-stabilized emulsions that are aged or heated above their thermal denaturation temperature, since the degree of interfacial covalent (disulfide) bond formation becomes more extensive (Dickinson and Matsumura 1991, Monahan et al. 1996). It may then be extremely difficult to displace the aggregated protein molecules from the interface using small molecule surfactants, unless a chemical is added first to disrupt the covalent cross-links, for example, mercaptoethanol for disulfide bonds. The earlier discussion has highlighted the wide variety of factors that can influence interfacial composition, such as emulsifier concentration, emulsifier type, solution conditions, temperature, and time. For this reason, a great deal of research is being carried out to establish the relative importance of each of these factors, and on establishing the relationship between interfacial composition and the bulk physicochemical properties of food emulsions. 5.5.2 Characterization of Interfacial Composition in Emulsions A variety of analytical techniques are available for determining the composition of interfacial layers; however, many of these are only suitable for making measurements at planar gas–liquid, liquid–liquid, or solid–liquid interfaces (Sections 5.6 through 5.8). In this section, the primary focus will be on those techniques that can be used to determine the interfacial composition of emulsions, but a few of the more commonly used techniques for analyzing planar interfaces will also be mentioned, because they can provide valuable insights into the fundamental factors that influence the composition and structure of droplet coatings. The most commonly used method of obtaining information about interfacial composition in emulsions is to use the “depletion” technique (Dickinson 1992a). An emulsion is prepared using a known concentration of emulsifier (c T), and then the concentration of nonadsorbed emulsifier remaining in the continuous phase (cNA) after homogenization is determined using an appropriate analytical method, for example, a chemical, spectroscopic, electrophoretic, or chromatographic method. The concentration of adsorbed emulsifier is then deduced from the difference between the total emulsifier present and that which remains in the continuous phase: cA = (c T − cNA). The depletion method assumes that all of the “adsorbed” emulsifier is present at the oil–water interface. Some emulsifiers may have a significant solubility in the dispersed phase so that an appreciable fraction is present in the interior of the droplets rather than at the interface. In many food emulsions, there may be more than one type of surface-active material present at the oil–water interface, so that it is important to determine the interfacial concentrations of the different species. This can also be achieved using the depletion method, but in this case, it is necessary to use one or more analytical techniques to measure the concentration of the different kinds of surface-active species present in the total emulsion and in the continuous phase after homogenization. Chromatography and electrophoresis techniques are often used for this purpose, for example, SDS-PAGE for proteins. The depletion method is particularly suitable for determining the interfacial composition of emulsions prepared by the investigator, since then the initial concentrations of the various surface-active species present in the system are known. If the interfacial composition of a pre-existing emulsion needs to be determined and this emulsion was not prepared by the investigator, then a different method must be used, because the initial concentration of surface-active species present in the system may not be known. In these cases, it is often possible to separate the droplets in an emulsion from the continuous phase by centrifugation or filtration. The droplets can then be collected and washed to remove any residual 212 Food Emulsions: Principles, Practices, and Techniques continuous phase. Washing can be achieved by dispersing the droplets in an appropriate buffer solution and then centrifuging or filtering them repeatedly. Once the droplets have been successfully separated from the continuous phase, a highly surface-active component can be added to the system to displace the original emulsifiers from the oil–water interfaces. In some cases, it may also be necessary to add an additional chemical component to break down any covalent bonds formed between adsorbed emulsifier molecules: for example, mercaptoethanol can be added to breakdown disulfide bonds between globular proteins and facilitate emulsifier displacement (Monahan et al. 1996). Once all the emulsifier molecules have been displaced from the droplet surfaces, the continuous phase containing the emulsifier could be separated from the droplets, and their concentration and identity could be established using appropriate analytical methods to analyze the continuous phase, for example, chromatography, electrophoresis, NMR, or mass spectrometry. Usually, it will be necessary to have a fairly good idea about the type of emulsifier molecules present in a system in order to select an appropriate analytical method to measure its concentration and identity. Information about the nature of the emulsifier molecules present at the surface of an emulsion droplet can sometimes be obtained indirectly by measuring the change in droplet characteristics with solution or environmental conditions. For instance, the change in the electrical charge on the droplet surfaces could be determined as a function of pH, or the stability of droplets to flocculation could be measured as a function of pH, ionic strength, or temperature. Droplets stabilized by different types of emulsifier behave differently when solution or environmental conditions are changed; hence, it may be possible to obtain some insight into emulsifier type using this approach. For example, the ζ-potential of the droplets in a protein-stabilized emulsion will go from positive to negative as the pH is raised from below to above their isoelectric point, whereas the electrical charge on the droplets in an ionic surfactant-stabilized emulsion may be relatively high and fairly insensitive to pH. For some emulsions, it is possible to obtain information about the type and concentration of emulsifiers present at the droplet surfaces using instrumental methods. For example, it is often possible to infer the interfacial composition in emulsions from microscopy or spectroscopy techniques that are sensitive to the molecular environment of emulsifier molecules, such as infrared, fluorescence, Raman, circular dichroism, and UV–visible spectroscopy (see Section 5.6). 5.6 Interfacial Structure 5.6.1 Factors Influencing Interfacial Structure The structural organization of emulsifier molecules within the interfacial layers surrounding emulsion droplets strongly influences the physicochemical and sensory properties of many types of food emulsion. Interfaces are often treated as thin homogeneous layers of well-defined thickness, but in reality, they are usually heterogeneous and structurally complex regions (Figure 5.16). Interfacial composition may vary both horizontally and vertically across the interface. For example, there is often a greater concentration of polymeric emulsifiers in the immediate vicinity of an interface where the polymers are densely packed, than at the outer edge of the interface where the polymers have a more flexible and open structure (Dickinson 2001). In addition, there may be aggregation or phase separation of emulsifier molecules at an interface so that interfacial composition, thickness, and properties vary across the two-dimensional plane of the interface (Norde 2011). A dramatic example of this kind of behavior has been observed at interfaces containing adsorbed proteins when small molecule surfactants are added to the bulk solution (Mackie et al. 2000, Wilde 2000, Pugnaloni et al. 2004). At relatively low surfactant concentrations, surfactant molecules adsorb to the interface and form small islands of surfactant located within the protein network (Figure 5.18). As the surfactant concentration is increased, the size of the surfactant-rich regions expands, restricting the protein network to a smaller surface area. At relatively high surfactant concentrations, the protein region increases appreciably in thickness and eventually the protein molecules are completely displaced from the interface. The two-dimensional phase separation of the interface into a protein-rich and a surfactant-rich region can clearly be observed using various microscopy techniques (Figures 5.19 and 5.20). The structural evolution of the interfacial film depends on the type of 213 Interfacial Properties and Their Characterization FIGURE 5.18 Schematic representation of the changes in the structure of protein interfaces when increasing amounts of surfactant are added. A two-dimensional phase separation occurs, leading to protein-rich and surfactant-rich areas. (a) (b) FIGURE 5.19 (a) BAM and (b) AFM images of a planar interface consisting of a protein film (β-lactoglobulin) to which a nonionic surfactant (Tween 20) was added. The BAM image is 300 × 260 μm, while the AFM image is 6 × 6 μm. (Courtesy of Dr. Alan Mackie, Institute of Food Research, Norwich, U.K.) protein that was initially present at the interface (Figure 5.20). For example, a nonionic surfactant forms fairly circular domains in β-casein films, because there are only weak cross-links between the protein molecules and so the surfactants can easily push them aside. On the other hand, a nonionic surfactant forms irregular shaped domains in β-lactoglobulin films, because strong bond formation between protein molecules means that the film has to fracture before the surfactant domains can expand. This type of two-dimensional phase separation also depends on the type of surfactant used to displace the adsorbed proteins (Gunning et al. 2004). In addition, to the structural characteristics mentioned earlier, there may also be appreciable changes in the conformation and interactions of emulsifier molecules after they are adsorbed to an interface, for example, surface- or thermal-denaturation of adsorbed globular proteins (Kim et al. 2002a,b, Norde 2011). It is therefore important for food scientists to understand the factors that influence the structural organization of the surface-active molecules within interfacial layers. The conformation and orientation of surface-active molecules adsorbed to an interface is governed by their attempt to reduce the free energy of the system (Evans and Wennerstrom 1999, Norde 2011). Amphiphilic molecules usually arrange themselves so that the maximum number of nonpolar groups is in contact with the oil phase, while the maximum number of polar groups is in contact with the aqueous phase, so as to reduce unfavorable hydrophobic interactions (Figure 5.16). For this reason, small molecule surfactants tend to have their 214 Food Emulsions: Principles, Practices, and Techniques (a) (b) FIGURE 5.20 AFM images of planar interfaces comprising of protein films ((a) β-casein and (b) β-lactoglobulin) to which a nonionic surfactant (Tween 20) has been added to displace the proteins. Both images are 4 μm across. The dark regions are surfactant, and the lighter regions are protein. The surfactant forms fairly circular domains in the β-casein film, because there are only weak interactions between the protein molecules. On the other hand, the surfactant forms irregular shaped domains in the β-lactoglobulin films, because there are strong interactions between the protein molecules, so the film has to fracture before the domains can expand. (Courtesy of P. Gunning, Institute of Food Research, Norwich, U.K.) polar head groups protruding into the aqueous phase, and their hydrocarbon tails protruding into the oil phase. Similarly, biopolymer molecules adsorb so that predominantly nonpolar segments are located within the oil phase, whereas predominantly polar segments are located within the water phase. In addition, many biopolymer molecules undergo structural rearrangements after adsorption to an interface so as to further optimize the number of favorable interactions (Norde 2011, Rabe et al. 2011). In aqueous solutions, globular proteins adopt a three-dimensional conformation in which the nonpolar amino acids are predominantly located in the hydrophobic interior of the protein molecule so that they can avoid unfavorable hydrophobic interactions with the surrounding water molecules. When globular proteins adsorb to an oil–water interface, they are no longer completely surrounded by water, and so the protein can reduce its free energy by altering its conformation so that many of the nonpolar amino acids are located in the oil phase, and many of the polar amino acids are located in the water phase. The rate at which the conformation of a biopolymer changes at an oil–water interface depends on its molecular structure. Flexible random coil molecules can rapidly alter their conformation, whereas rigid globular molecules change more slowly because of various kinetic constraints. Immediately after adsorption to an interface, a globular protein has a conformation that is similar to that in the bulk aqueous phase. With time, it alters its conformation so that it can optimize the number of favorable interactions between the nonpolar amino acids and the oil molecules. An intermediate stage in this unfolding process is the exposure of some of the nonpolar amino acids to water, which is thermodynamically unfavorable because of the hydrophobic effect, and so there is an energy barrier that must be overcome before unfolding can occur. In this case, the rate of any conformational changes will depend on the height of the energy barriers compared to the thermal energy. The conformation of biopolymer emulsifiers at an interface may also depend on when they arrive (Fleer et al. 1993, Norde 2011). Emulsifier molecules that arrive at the beginning of the adsorption process have a large surface area over which to spread, whereas those that arrive latter have less room because of the presence of the other adsorbed emulsifier molecules. The final structure of the biopolymer molecules may also depend on the adsorption rate compared to the rate of any conformational changes (Figure 5.21). If adsorption to the surface is faster than unfolding, then the surface excess concentration will be higher and the interface will tend to be thicker. On the other hand, if adsorption is slower than unfolding, then the surface excess concentration will be lower and the interface will be 215 Interfacial Properties and Their Characterization Adsorption rate >> unfolding rate high Γ Adsorption rate << unfolding rate low Γ FIGURE 5.21 For polymers that are able to undergo conformational changes at an interface, the surface excess concentration depends on the rate of the conformational changes compared to the adsorption rate. Relatively rapid adsorption gives less time for conformational changes and therefore leads to a more densely packed interface. thinner, because significant spreading of the biopolymer molecules can occur. Finally, the conformation of relatively flexible biopolymer emulsifiers may also depend on the strength of their attractive interactions with the interface, which can be described in terms of the solvent quality. If the attraction between the monomers and the surface is extremely strong, then the biopolymer will tend to form a thicker interfacial layer with tails and loops that extend further into the continuous phase. On the other hand, if the attraction between the monomers and surface is relatively weak, then the biopolymer will tend to form a thinner interfacial layer with a higher proportion of trains and loops and tails that do not extend as far into the continuous phase. This phenomenon occurs, because when the attraction between the biopolymer and interface is relatively strong, the monomers will tend to stick where they first encounter the interface and remain there, but when the attraction is relatively weak, the biopolymer can undergo conformational changes after adsorption to maximize the number of favorable interactions with the interface. The thickness and structural organization of biopolymer molecules within an interfacial layer may be strongly dependent on pH, ionic strength, and temperature (Claesson et al. 1995, Dickinson 2001). Electrostatic interactions involving biopolymer molecules depend on the sign, type, number, and distribution of the charged groups along the biopolymer chain. Many food biopolymers contain groups that change their degree of ionization in response to changes in pH, for example, carboxyl, amino, phosphate, sulphate, and imidazole groups. An understanding of the influence of pH on interfacial structure depends on an appreciation of the change in the intramolecular and intermolecular electrostatic interactions involving adsorbed and nonadsorbed molecules. At pH values where biopolymers have no net charge, the intramolecular electrostatic repulsion between similarly charged groups is usually reduced, and there may even be attraction between oppositely charged groups, and so the molecules tend to adopt a more compact interfacial structure (Sanchez et al. 2003, Nino et al. 2005). On the other hand, at pH values where the molecules have a high net charge, there tends to be a strong electrostatic repulsion between different segments of the same biopolymer chain, which causes the molecules to adopt a more open structure. Similarly, intermolecular interactions between charged groups on different molecules also tend to favor a denser interfacial layer when the biopolymers have a low net charge (since there is less electrostatic repulsion between them), and a more open interfacial structure when the biopolymers have 216 Food Emulsions: Principles, Practices, and Techniques a high net charge (since there is a strong electrostatic repulsion between them). Finally, it should be noted that the thickness of the interfacial layer may increase if the pH is adjusted to a value where adsorption of biopolymer molecules to the interface is promoted, for example, near the point of zero net charge. Electrostatic interactions involving charged biopolymer molecules are also strongly influenced by ionic strength, and so the mineral content can have a pronounced influence on the structure of interfacial layers. The range and magnitude of electrostatic interactions increases as the ionic strength decreases. Hence, the presence of salt will tend to decrease the thickness of interfacial layers comprised of biopolymers with high charges by screening the electrostatic repulsion between different segments of the same molecule. On the other hand, the addition of salt may actually increase the thickness of interfacial layers with no net charge if they contain a balance of positive and negative charges, because the salt screens the electrostatic attraction between the oppositely charged groups. The thickness and structural organization of emulsifier molecules within an interfacial layer may also depend on temperature, since temperature may influence the solvent quality or the conformation and interactions of adsorbed molecules. The thickness and structural organization of the emulsifier molecules within an interfacial layer can have an important influence on the bulk physicochemical properties of food emulsions (Dickinson 2001, McClements 2004). The flocculation and coalescence stability of many protein-stabilized emulsions depends on the unfolding and interaction of the protein molecules at the droplet surface. When globular proteins unfold, they expose reactive amino acids that are capable of forming hydrophobic and disulfide bonds with their neighbors, thus generating a highly viscoelastic interface that is resistant to droplet coalescence under quiescent conditions (Dickinson 1992). Conversely, if hydrophobic attraction and disulfide bond formation occurs between proteins adsorbed onto different droplets, then extensive droplet flocculation can occur (Kimet al. 2002a,b). The steric repulsion between emulsion droplets is strongly dependent on the thickness of the interfacial layer (Chapter 3); hence, any factor that changes the thickness of this layer may have an important impact on the stability of droplets to aggregation. The susceptibility of certain proteins to enzymatic hydrolysis may depend on their orientation at the droplet surface (Dalgleish 1996). The susceptibility of surfactants with unsaturated hydrocarbon tails to lipid oxidation depends on whether their tails are orientated perpendicular or parallel to the droplet surface, the latter being more prone to oxidation by free radicals generated in the aqueous phase (Coupland and McClements 1996). 5.6.2 Characterization of Interfacial Structure in Emulsions The impact of interfacial structure on emulsion properties means that it is important to have analytical techniques that can be used to measure interfacial thickness and the structural organization of emulsifier molecules within interfacial layers. A number of the most important experimental techniques that are available to provide information about interfacial structure are listed in Table 5.2. Each of these techniques works on a different physical principle and is sensitive to a different aspect of the properties of an interface. The majority of these techniques can only be used in fundamental studies of emulsifiers adsorbed to planar interfaces, although some of them can be also used to directly provide information about the properties of the interfaces in emulsions. 5.6.2.1 Microscopy Techniques One of the most direct methods of obtaining information about the structural organization of molecules at an interface is to use microscopy, often in combination with suitable staining methods. There have been rapid developments in many areas of microscopy during the past few decades that have led to the availability of a variety of powerful analytical instruments for probing interfacial structure. Relatively large (>200 nm) changes in the structural organization of molecules at surfaces, such as the two-dimensional phase separation of protein–surfactant mixtures, can be observed by optical microscopy, for example, scanning near field optical microscopy, laser scanning confocal fluorescence, or Brewster angle microscopy (BAM) (Blonk and Vanaalst 1993, Patino et al. 1999, Mackie et al. 2001). Information about finer structural details can be obtained using electron microscopy or atomic force microscopy (AFM) (Stokes 2003, 217 Interfacial Properties and Their Characterization TABLE 5.2 Experimental Techniques to Characterize the Structural Properties of Interfaces Analytical Techniques Depletion techniques Scattering techniques Light scattering Small angle x-ray scattering Neutron scattering Reflection techniques Ellipsometry Surface plasmon resonance Neutron reflection Absorption techniques Infrared Ultraviolet–visible Fluorescence Nuclear magnetic resonance Circular dichroism Biochemical techniques Enzyme hydrolysis Immunoassays Microscopic techniques Fluorescent microscopy Raman microscopy AFM Electron microscopy BAM Mechanical methods Interfacial rheometers Quartz crystal microbalance Applications Interfacial composition Thickness Thickness, concentration profile, packing Thickness, concentration profile Thickness, refractive index Thickness, refractive index Thickness, concentration profile Composition, conformational changes Composition, conformational changes Composition, conformational changes Composition Secondary structure Orientation, conformation Composition, orientation, conformation Interfacial location and composition Interfacial location and composition Interfacial structure and topology Interfacial structure and topology 2-D organization of molecules Interfacial rheology and interactions Interfacial load and rheology Keerati-u-rai and Corredig 2009, Dudkiewicz et al. 2011, Liu and Wang 2011). AFM constructs a topographical image of a sample surface by scanning a small tip across it (Morris et al. 1999, Sitterberg et al. 2010, Raigoza et al. 2013). The major advantage of AFM is that little sample preparation is often needed, so that the structure of the surface is not adversely altered. A more detailed discussion of microscopy techniques is given in the chapter on emulsion characterization (Chapter 14). The utility of microscopy techniques for characterizing structural changes at interfaces is demonstrated in Figures 5.19 and 5.20, which show BAM and AFM images of the two-dimensional phase separation of protein–surfactant mixtures at a planar interface. Other microscopy methods that can provide information about the chemical composition of interfaces at specific locations may also be useful, such as Raman or Fourier transform infrared (FTIR) microscopy methods (Mudunkotuwa et al. 2014, Zheng and He 2014). 5.6.2.2 Spectroscopy Techniques A number of spectroscopic techniques have been developed to provide information about interfacial structure based on absorption of electromagnetic radiation by specific molecules at the interface, for example, infrared, ultraviolet, visible, fluorescence, and x-ray techniques (Kallay et al. 1993, Bajpai and Rajpoot 1999, Stefaniu and Brezesinski 2014). These techniques can either be used in a transmission mode, where the electromagnetic wave is propagated through the sample, or in a reflection mode, where the electromagnetic wave is reflected from the interface. In both cases, the absorption of energy from the electromagnetic wave is measured and correlated to some property of the molecules at the 218 Food Emulsions: Principles, Practices, and Techniques interface. The different techniques are sensitive to different characteristics of the molecules. Infrared and microwave techniques rely on characteristic molecular vibrations or rotations of different groups within a molecule, whereas ultraviolet–visible and fluorescence techniques rely on characteristic electronic transitions involving outer shell electrons (Nielsen 2010). Spectroscopy techniques have been used to determine the type and concentration of emulsifier molecules at interfaces, to study adsorption kinetics, and to provide information about changes in molecular conformation after adsorption (Kallay et al. 1993, Lee et al. 2007). When a molecule undergoes a conformational change, the adsorption spectra is altered because of the change in the environment of the adsorbing groups. A number of these techniques have been adapted so that they can be used to analyze the structures of interfaces in emulsions (Zhai et al. 2013). For example, FTIR and front-face fluorescence spectroscopy (FFFS) have been used to provide information about changes in the conformation of globular proteins adsorbed to droplet surfaces in emulsions (Fang and Dalgleish 1998, Rampon et al. 2003, Lee et al. 2007). An alternative approach has been developed to monitor the properties of adsorbed emulsifiers in emulsions using conventional spectroscopy methods. Normally, emulsions cannot be analyzed by spectroscopy methods that rely on the transmission of electromagnetic waves through the sample, because the droplets scatter the waves so strongly. However, this problem can be overcome by making an emulsion transparent by adding solutes to the aqueous phase so that its refractive index becomes the same as that of the oil phase (Husband et al. 2001). Because the resulting “refractive index–matched emulsion” (RIME) is optically transparent, it can be analyzed using standard spectroscopy techniques, such as fluorescence, FTIR, circular dichroism, or UV–visible spectrometry. It is then possible to use this approach to provide information about interfacial structure in concentrated emulsions. Nevertheless, there are concerns about the influence of the solutes added to the emulsions to increase the refractive index of the aqueous phase (e.g., glycerol, sorbitol, polyethylene glycol, or sucrose) on the structural properties of the proteins, since many solutes change protein conformation and aggregation at the high concentrations needed to create RIMEs (McClements 2002). 5.6.2.3 Interference Reflection Techniques A number of techniques have been developed that rely on the interference pattern produced by the reflection of a beam of light or subatomic particles from a thin interface. Ellipsometry relies on the reflection of a polarized light beam from a highly reflective planar surface (Elwing 1998, Arwin 2014). Initially, the phase and amplitude of the light beam reflected from the clean surface are measured. Emulsifier is then allowed to adsorb to the surface and the experiment is repeated. A mathematical analysis of the reflected wave provides information about the thickness and refractive index profile of the adsorbed layer. The refractive index profile can then be converted into a concentration profile. Imaging ellipsometry techniques have been developed that enable one to scan a beam of light across a surface so that a two-dimensional image of the thickness and optical properties of an interface can be obtained (Niu and Jin 2013). Information about interfacial composition can often be inferred from the optical properties. BAM is another optical interference reflectance technique that has been used to provide information about the structural organization of surface-active molecules at planar interfaces (Murray et al. 2009, Stefaniu et al. 2014). This technique has been used primarily to observe the two-dimensional organization of emulsifiers at planar surfaces and interfaces (Patino et al. 1999, Mackie et al. 2001, van Kempen et al. 2013). This technique is based on the fact that there is no reflection of polarized light from a clean surface when the angle of incidence of the light is set at (or below) the Brewster angle. However, light is reflected when a relatively thin emulsifier layer (∼2 nm) is present at the surface, provided that the refractive index of the layer is significantly different from the refractive index of the liquid below it. The reflected light can be used to form an image of the two-dimensional morphology of the layer (Figure 5.19). The BAM technique has been used to provide valuable information about the changes in interfacial structure that occur when small molecule surfactants displace proteins from surfaces or interfaces (Mackie et al. 2001, Murray et al. 2009). The two-dimensional phase separation of the interface into a protein-rich and a surfactant-rich region can clearly be observed using a BAM (Figure 5.19). Interfacial Properties and Their Characterization 219 Neutron reflection techniques work on a similar principle to optical reflection techniques, but they use a beam of neutrons, rather than a beam of light, to probe the interface (Atkinson et al. 1995, 1996, Lu et al. 2007, Heinrich and Loesche 2014). The beam of neutrons is directed at an air–water or oil–water interface at an angle. The beam of neutrons reflected from the surface is analyzed and provides information about the thickness of the interface and the neutron refractive index profile, which can be related to the emulsifier concentration profile across the interface. These techniques have been widely used to study the characteristics of protein layers and the influence of pH and ionic strength on their properties. 5.6.2.4 Scattering Techniques A number of analytical instruments based on scattering of radiation or subatomic particles are suitable for providing information about the structural organization of molecules adsorbed to interfaces, for example, optical, x-ray, or neutron scattering techniques (Cristofolini 2014, Stefaniu and Brezesinski 2014). Analytical techniques based on the scattering of radiation or subatomic particles are most commonly used to provide information about interfacial thickness, but some of them can also provide detailed information about molecular organization and interactions. They are often used to compare the thickness of interfaces formed by different emulsifiers, or to study changes in interfacial thickness or structural organization in response to alterations in environmental conditions, such as pH, ionic strength, or temperature. The application of these techniques has provided valuable insights into the structures formed by proteins at oil–water and air–water interfaces, which is useful for predicting the stability of emulsion droplets to flocculation or coalescence. An example of this approach is the use of light scattering (or some other particle sizing technique) to indirectly determine the thickness of the emulsifier layer surrounding colloidal particles. It is usually advantageous to use colloidal particles that are fairly monodisperse and that are stable to aggregation in the absence of emulsifier. The particle radius is then measured before and after addition of the emulsifier, and the difference in radius is taken to be equal to the thickness of the interfacial layer. 5.6.2.5 Langmuir Trough Measurements Useful information about the structural organization of molecules at gas–liquid and liquid–liquid interfaces can be obtained using a Langmuir trough apparatus (Norde 2011, Stefaniu et al. 2014). A Langmuir trough consists of a container that holds the liquid(s) to be analyzed, a moveable barrier that is capable of changing the area of the liquid–air interface, and a device for measuring the surface tension at the liquid–air interface, usually a Wilhelmy plate (Figure 5.22). The surface-active solute to be analyzed is either dissolved in the liquid or spread across the surface of the liquid depending on its solubility. The interfacial area is then decreased in a controlled fashion using a motor to drive the moveable barrier, and the surface pressure (π) is measured as a function of interfacial area (A). The resulting π-A plot depends on the interfacial characteristics of the individual solute molecules, as well as on the sign, magnitude, and range of solute–solute interactions at the interface. A highly schematic π-A plot is shown in Figure 5.23 for a model surface-active solute. It is convenient to divide the π-A plot into different regimes depending on the strength of the solute–solute interactions. At high interfacial areas, the solute molecules are far apart and do not interact with each other, and therefore, the surface pressure is mainly determined by the interfacial characteristics of the individual molecules. By analogy to three-dimensional systems, this regime is usually called the “gas” phase, because each solute molecule acts independently of its neighbors. As the interfacial area is decreased, the interfacial concentration of the solute molecules increases, and solute–solute interactions begin to occur, which leads to an increase in the surface pressure. Even so, the solute molecules are still sufficiently far apart to be able to move freely past each other and so this regime is referred to as the “liquid” phase. When the interfacial area is decreased further, the solute molecules become so closely packed together that there is a strong repulsive force between them, which leads to a steep increase in the surface pressure. This regime is usually referred to as the “solid” phase, because the solute molecules are densely packed and have low mobility. The interfacial area where the surface pressure increases dramatically can provide valuable information about the packing of the surface-active 220 Food Emulsions: Principles, Practices, and Techniques Wilhelmy plate Movable barriers Trough Liquid FIGURE 5.22 Schematic diagram of Langmuir trough apparatus used to measure packing of solutes at interfaces, and interfacial dilational rheology, by measuring changes in surface pressure as a function of interfacial area. Solid Π Gas Liquid Solid Liquid Gas Area FIGURE 5.23 Schematic representation of the change in surface pressure with interfacial area when the area is compressed. The π-A profile can often be divided into a gas, liquid, and solid regime, depending on the strength of the solute– solute interactions. Interfacial Properties and Their Characterization 221 molecules within the interfacial layer at saturation, for example, the area per solute molecule. In practice, π-A plots are often more complex than that shown in Figure 5.23 because of changes in the orientation and interactions of solute molecules or because of phase separation processes. The Langmuir trough can also be used in combination with various microscopy, reflection, and scattering techniques to provide information about changes in the structural organization and interactions of solute molecules as the interfacial area is changed, for example, BAM, fluorescence microscopy, neutron reflection, or x-ray scattering (Murray et al. 2009, Stefaniu et al. 2014). 5.6.2.6 Surface Force Measurements The thickness of a layer of emulsifier molecules adsorbed to an interface can be determined by measuring the forces between two interfaces as they are brought into close contact (Claesson et al. 1995, Meyer et al. 2006). There is usually a steep rise in the repulsive forces at close separations between the interfaces because of electrostatic or steric repulsion. When the electrostatic forces are relatively weak (e.g., at high ionic strength or at pH values where surface charge is low), the distance at which the steep rise in repulsion occurs can be assumed to be the outer edge of the emulsifier layer. By compressing the interfacial layer further, some information about the packing of the adsorbed molecules within the interface can be obtained. Interesting information about the properties of molecules at interfaces (such as proteins) can be obtained using single-molecule force spectroscopy, where the force required to pull a molecule from the interface is measured (Touhami et al. 2011). 5.6.2.7 Calorimetry Techniques Information about the conformation of proteins adsorbed to the surfaces of emulsion droplets has also been obtained using sensitive differential scanning calorimetry (DSC) techniques (Corredig and Dalgleish 1995). The emulsion being analyzed is heated at a controlled rate, and the amount of heat absorbed or released by the proteins when they undergo conformational changes is recorded by the DSC instrument. The temperature at which the conformation change occurs (Ttransition) and the amount of heat adsorbed or released by the protein during the transition (ΔHtransition) are determined. The values for the adsorbed protein are then compared with those for a solution containing a similar concentration of nonadsorbed proteins. Experiments with oil-in-water emulsions containing adsorbed globular proteins, such as α-lactalbumin, lysozyme, and β-lactoglobulin, have shown that there is a decrease in ΔHtransition after adsorption, which suggests that they have undergone some degree of surface denaturation (Corredig and Dalgleish 1995). After the proteins had been desorbed from the droplet surfaces, it was observed that the denaturation of β-lactoglobulin was irreversible, but that of α-lactalbumin and lysozyme was at least partially reversible. Flexible biopolymers that have little structure, such as casein, cannot be studied using this technique, because they do not undergo any structural transitions that absorb or release significant amounts of heat. It should be noted that the data from DSC experiments must be interpreted carefully because a reduction in ΔHtransition may be because the protein has no secondary or tertiary structure or because the structure is so stable that it does not unfold at the same temperature as in water. In addition, the adsorption of a protein at an interface may also reduce ΔHtransition, because its thermodynamic environment has altered: in solution the protein is completely surrounded by water, but at an interface, part of the molecule is in contact with a hydrophobic surface. Despite these limitations, the DSC technique is a powerful tool for studying structural changes caused by adsorption, especially when used in conjunction with other techniques. 5.6.2.8 Biochemical Techniques A number of biochemical techniques have also been developed to provide information about the structure and organization of biopolymer molecules at the surface of emulsion droplets (Dalgleish 1996). The susceptibility of proteins to hydrolysis by specific enzymes can be used to identify the location of particular peptide bonds. When a protein is dissolved in an aqueous solution, the whole of its surface is usually accessible to enzyme hydrolysis, but when it is adsorbed to an interface, it adopts a conformation 222 Food Emulsions: Principles, Practices, and Techniques where some of the amino acids are located close to the interface (and are therefore inaccessible to proteolysis), whereas others are exposed to the aqueous phase (and are therefore accessible to proteolysis). By comparing the bonds that are susceptible to enzyme hydrolysis and the rate at which hydrolysis of the adsorbed and nonadsorbed proteins occurs, it is possible to obtain some idea about the orientation and conformation of the adsorbed proteins at the interface. This technique is more suitable for the study of flexible random-coil proteins, than for globular proteins, because the latter have compact structures that are not particularly accessible to enzyme hydrolysis in either the adsorbed or nonadsorbed states. Immunological techniques can also be used to provide similar information. These techniques utilize the binding of antibodies to specific sites on a protein to provide information about its orientation at an interface or about any conformational changes that take place after adsorption. 5.7 Interfacial Tension and Its Measurement 5.7.1 Factors Influencing Interfacial Tension Interfacial tension is a measure of the free energy required to increase the area of an interface by a unit amount, and is usually specified in units of J m−2 (or N m−1) (Israelachvili 2011). The origin of the interfacial tension is the imbalance of molecular interactions between molecules located at the interface (Section 5.2). When a surface-active solute is present, the interfacial tension is reduced, because the solute can help to minimize the thermodynamically unfavorable contacts between the different kinds of molecules at the interface. From a practical standpoint, the interfacial tension is important, because it influences the droplet size produced during homogenization (Chapter 6), the stability of droplets to coalescence and Ostwald ripening (Chapter 7), and the packing of large droplets in concentrated emulsions, such as mayonnaise and dressings (Chapter 12). In addition, the measurement of interfacial tension can provide valuable information about emulsifier and interfacial characteristics, such as surface excess concentration, surface activity, critical micelle concentration, adsorption rates, and interfacial rheology. Finally, the interfacial tension also influences many of the macroscopic properties of food materials, such as capillary rise, spreading, and wetting (Section 5.9). 5.7.2 Characterization of Interfacial Tension The purpose of this section is to give an overview of some of the most important analytical instruments that have been developed to measure interfacial and surface tensions (Couper 1993, Miller et al. 1994, Hiemenz and Rajagopalan 1997, Hartland 2004, Javadi et al. 2013). By definition, surface tension is measured at a gas–fluid interface (e.g., air–water or air–oil), while interfacial tension is measured at a fluid–fluid interface (e.g., oil–water). These quantities are measured using analytical instruments called surface or interfacial tensiometers, respectively. A wide variety of different types of tension meters are available for providing information about the properties of surfaces and interfaces (Table 5.3). These instruments differ according to the physical principles on which they operate, their mechanical design, whether measurements are static or dynamic, and whether they are capable of measuring surface tension, interfacial tension, or both. Static measurements are carried out on surfaces or interfaces that are considered to be at equilibrium, whereas “dynamic” measurements are carried out under nonequilibrium conditions. 5.7.2.1 Du Nouy Ring Method The Du Nouy ring method is commonly used to measure static surface and interfacial tensions of liquids. The apparatus required to carry out this type of measurement consists of a vessel containing the liquid(s) to be analyzed and a ring that is attached to a sensitive force measuring device (Figure 5.24). The vessel is capable of being moved upward and downward in a controlled manner, while the position of the ring is kept constant. Initially, the vessel is positioned so that the ring is submerged just below the surface of the liquid being analyzed. The vessel is then slowly lowered and the force exerted on the ring is recorded. 223 Interfacial Properties and Their Characterization TABLE 5.3 Summary of the Instruments Used for Measuring Surface and Interfacial Tensions Surface Tension Interfacial Tension Equilibrium Du Nouy ring Wilhelmy plate Pendant drop Sessile drop Spinning drop Capillary rise Dynamic Maximum bubble pressure Oscillating jet Drop volume Surface waves Du Nouy ring Wilhelmy plate Pendant drop Sessile drop Spinning drop Drop volume Pendant drop Spinning drop Force measuring device Force measuring device Ring Air Ring Sample holder Liquid held by surface forces Liquid Volume of liquid being weighed FIGURE 5.24 Du Nouy ring method of determining the interfacial and surface tension of liquids. The force on the ring is measured as the vessel holding the liquid is lowered. As the surface of the liquid moves downward, some of the liquid “clings” to the ring because of its surface tension (Figure 5.24). The weight of the liquid that clings to the ring is recorded by the forcemeasuring device, and used to calculate the surface tension. The Du Nouy ring method is often used in a “detachment” mode. The vessel is lowered until the liquid clinging to the ring ruptures and the ring becomes detached from the liquid. The force exerted on the ring at detachment is approximately equal to the surface tension multiplied by the length of the ring perimeter: F = 4πRγ, where R is the radius of the ring. In practice, this force has to be corrected, because the surface tension does not completely act in the vertical direction and because some of the liquid remains clinging to the ring after it has become detached: F = 4πRγβ (5.20) where β is a correction factor, which depends on the dimensions of the ring and the density of the liquid(s) involved. Values of β have been tabulated in the literature or can be calculated using semiempirical equations (Couper 1993). One of the major problems associated with the Du Nouy ring method, as well as any other detachment method, is that serious errors may occur when it is used for measuring the surface tension of emulsifier solutions, rather than pure liquids. When a ring detaches from a liquid, 224 Food Emulsions: Principles, Practices, and Techniques it leaves behind some fresh surface that is initially devoid of emulsifier. The measured surface tension therefore depends on the speed at which the emulsifier molecules move from the bulk liquid to the fresh surface during the detachment process. If an emulsifier adsorbs rapidly compared to the detachment process, the surface tension measured by the detachment method will be similar to the equilibrium value, but if the emulsifier adsorbs relatively slowly, the surface tension will be greater than the equilibrium value, because the surface has a lower emulsifier concentration than expected. The Du Nouy ring method can also be used to determine the interfacial tension between two liquids. In this case, the ring is initially placed below the surface of the most dense liquid (usually water), and then the less dense liquid is poured on top (usually oil). The force acting on the ring is then measured as it is pulled up through the interface and into the oil phase. A similar equation can be used to determine the interfacial tension from the force exerted on the ring, but one has to take into account the densities of the oil and water phases and use a different correction factor. For two liquids that are partially immiscible with each other, the interfacial tension can take an appreciable time to reach equilibrium because of the diffusion of water molecules into the oil phase and vice versa. The Du Nouy ring method can also be used to determine surface or interfacial tensions by continuously monitoring the force acting on the ring as the vessel containing the liquid is slowly lowered, rather than just measuring the detachment force. As the liquid is lowered, the force initially increases, but at a certain position, it reaches a maximum (when the surface tension acts vertically), before decreasing slightly prior to detachment. In this case, the maximum force, rather than the detachment force, is used in the equations to calculate the surface or interfacial tension (Couper 1993). This method has the advantage that it does not involve the rupture of the liquid, and therefore, there are less problems associated with the kinetics of emulsifier adsorption during the detachment process. For accurate measurements, it is important that the bottom edge of the ring is kept parallel to the surface of the fluid, and that the contact angle between the liquid and the ring is close to zero. Rings are usually manufactured from platinum or platinum–iridium, because these give contact angles that are approximately equal to zero. The Du Nouy ring method can be used to determine surface tensions to an accuracy of about 0.1 mN m−1. 5.7.2.2 Wilhelmy Plate Method The Wilhelmy plate method is normally used to determine the static surface or interfacial tensions of liquids. Nevertheless, it can also be used to monitor adsorption kinetics provided that the accumulation of the emulsifier at the surface is slow compared to the measurement time, for example, for dilute biopolymer solutions. The apparatus consists of a vessel that contains the liquid to be analyzed and a plate, which is attached to a sensitive force-measuring device (Figure 5.25). The vessel is capable of being moved upwards and downwards, while the plate remains stationary. The vessel is positioned so that the liquid just comes into contact with the plate, that is, the bottom of the plate is parallel to the surface of the bulk liquid. Some of the liquid climbs up the edges of the plate, because this reduces the thermodynamically unfavorable contact area between the plate and the air. The amount of liquid that moves up the plate depends on its surface tension and density. If the force measuring device is zeroed prior to bringing the plate into contact with the liquid (to take into account the mass of the plate), then the force recorded by the device is equal to the weight of liquid clinging to the plate. This weight is balanced by the vertical component of the surface tension multiplied by the length of the plate perimeter: F = 2 (l + L) γcosθ (5.21) where l and L are the length and thickness of the plate θ is the contact angle Thus, the surface tension of a liquid can be determined by measuring the force exerted on the plate. Plates are often constructed of materials that give a contact angle close to 0, such as platinum or platinum– iridium as this facilitates the analysis (cosθ = 1). The Wilhelmy plate can also be used to determine the 225 Interfacial Properties and Their Characterization Force, F Plate Liquid climbs up plate due to γs Liquid FIGURE 5.25 Wilhelmy plate method of determining the interfacial and surface tension of liquids. The force on the plate is measured as the vessel holding the liquid is lowered. interfacial tension between two liquids. In this case, the plate is positioned so that its bottom edge is parallel to the interface between the two bulk liquids, and the force-measuring device is zeroed when the plate is located in the less dense liquid (usually oil). A major advantage of the Wilhelmy plate method over the Du Nouy ring method is that it does not rely on the disruption of the liquid surface, and therefore is less prone to errors associated with the adsorption kinetics of emulsifiers (see Section 5.3.4). The Wilhelmy plate method is widely used to study the adsorption kinetics of biopolymers. It can be used for this purpose, because their adsorption rate is much slower than the time required to carry out a surface tension measurement. The plate is positioned at the surface of the liquid at the beginning of the experiment, and then the force required to keep it at this position is recorded as a function of time. The Wilhelmy plate method is usually unsuitable for studying the adsorption kinetics of small molecule surfactants, because they adsorb too rapidly to be followed using this technique. For accurate measurements, it is important that the bottom of the plate is kept parallel to the liquid surface, and that the contact angle is either known or close to zero. Accuracies of about 0.05 mN m−1 have been reported using this technique. 5.7.2.3 Sessile and Pendant Drop Methods The sessile and pendant drop methods can be used to determine the static surface and interfacial tensions of liquids. The shape of a liquid droplet depends on a balance between the gravitational and surface forces. Surface forces favor a spherical droplet, because this shape minimizes the contact area between the liquid and its surroundings. On the other hand, gravitational forces tend to cause droplets to become flattened (if they are resting on a solid surface) or elongated (if they are hanging on a solid surface or from a capillary tip). A flattened drop is usually referred to as a “sessile” drop, whereas a hanging one is referred to as a “pendent” drop (Figure 5.26). The equilibrium shape that is adopted by a droplet is determined by its volume, density, and surface or interfacial tension. The surface or interfacial tension is determined by measuring the shape of a drop using an optical microscope, often in conjunction with an image analysis program, and mathematical equations that 226 Food Emulsions: Principles, Practices, and Techniques (a) (b) FIGURE 5.26 (a) Pendant drop and (b) sessile drop methods of determining the interfacial and surface tension of liquids. The shape of the drops is recorded and analyzed mathematically. describe the equilibrium shape. This technique can be used to determine the surface or interfacial tensions as low as 10 −4 mN m−1, has an accuracy of about 0.01%, and can be used to simultaneously determine the contact angle. Analytical instruments based on this principle have been developed that combine interfacial tension measurements with interfacial rheology measurements, which provides a powerful tool for studying the properties of many interfaces (see Section 5.8.2). 5.7.2.4 Drop Volume Method The drop volume method is used to measure surface and interfacial tensions of liquids by a detachment method. The liquid to be analyzed is pumped through the tip of a vertical capillary tube whose tip protrudes into air or an immiscible liquid (Figure 5.27). A droplet detaches from the tip of the capillary tube when the separation force (due to gravity) is balanced by the adhesion force (due to surface Oil Droplet Water Oil FIGURE 5.27 Drop volume method of determining the interfacial and surface tension of liquids. Interfacial Properties and Their Characterization 227 or interfacial tension). To measure surface tension, the tip should point downward into air, whereas to measure interfacial tension, the tip may point either upward or downward, depending on the relative densities of the two liquids being analyzed. If the liquid in the tip has a higher density than the surrounding medium, the opening of the tip faces down and the drop moves downward once it becomes detached. On the other hand, if the liquid in the tip has a lower density, then the tip faces up and the drop moves upward when it becomes detached. The surface or interfacial tension can be related to the volume of the detached droplet by analyzing the forces that act on it just prior to detachment (Couper 1993): g= VDROP Drg pd (5.22) where d is the diameter of the tip Δρ is the density difference between the liquid being analyzed and the surrounding liquid VDROP is the volume of the droplet g is the gravitational constant The volume of the droplets can be determined using a graduated syringe or by weighing them using a sensitive balance (once the liquid density is known). As with other detachment methods, the surface area of the droplet expands rapidly during the detachment process, and therefore, the method is unsuitable for analyzing liquids that contain emulsifiers whose adsorption time is comparable to the detachment time. To obtain reliable results, it is a normal practice to bring the droplet slowly to the detachment process. The accuracy of this technique has been reported to be better than 0.1 mN m−1. 5.7.2.5 Spinning Drop Method The spinning drop method is used to measure static interfacial tensions of liquids. It was designed to characterize liquids that have particularly low interfacial tensions, that is, between about 10 −6 to 5 mN m−1. Most food emulsions have interfacial tensions that are well above this range, and so this technique is not widely used in the food industry for studying oil–water interfaces, though it is being used to study the very low interfacial tensions between phase-separated biopolymer solutions (Scholten et al. 2002, 2006). This technique may also be useful for some fundamental studies, and therefore, its operating principles are briefly outlined in the following. A drop of oil (or water) is injected into a glass capillary tube that is filled with a more dense liquid (usually water). When the tube is made to spin at a particular angular frequency, ω, the droplet becomes elongated along its axis of rotation so as to reduce its rotational kinetic energy (Figure 5.28). The elongation of the droplet causes an increase in its surface area, and is therefore opposed by the interfacial tension. Consequently, the equilibrium shape of the droplet at a particular angular frequency depends on the interfacial tension. The shape of the elongated droplet is measured using optical microscopy, and the interfacial tension is then calculated using the following equation: g = kr 3w2 Dr (5.23) where k is an instrument constant that can be determined using liquids of known interfacial tension r is the radius of the elongated droplet Δρ is the difference in density between the oil (or water) and the surrounding liquid This equation is applicable when the droplet becomes so elongated that it adopts an almost cylindrical shape. More sophisticated equations are needed when the shape of the droplet does not change so dramatically (Couper 1993). 228 Food Emulsions: Principles, Practices, and Techniques Spinning droplet Aqueous phase ω Oil droplet Spinning capillary tube FIGURE 5.28 Spinning drop method of determining the interfacial tension of liquids. The change in shape of a liquid droplet suspended in another immiscible liquid is measured as the tube containing them is rotated at a known speed. 5.7.2.5.1 Maximum Bubble Pressure Method The maximum bubble pressure method is used to measure the static or dynamic surface tensions of liquids. The apparatus required to carry out this type of measurement consists of a vertical capillary tube whose tip is immersed below the surface of the liquid being analyzed (Figure 5.29). Gas is pumped into the tube that causes an increase in the pressure and results in the formation of a bubble at the end of the tip. The build-up of pressure in the tube is monitored using a pressure-sensor inside the instrument. As the bubble grows, the pressure increases until it reaches a maximum when the bubble becomes hemispherical and the surface tension acts in a completely vertical direction. Any further bubble growth beyond this point causes a decrease in the pressure. Eventually, the bubble formed breaks away from the tip and moves to the surface of the liquid, and then another bubble begins to form and the whole process is repeated. The maximum bubble pressure can be related to the surface tension using suitable mathematical equations. The maximum bubble pressure method can be used to measure the static surface tension of pure liquids at any bubble frequency. It can also be used to monitor the dynamic surface tension of emulsifier Air Gas Water Gas bubble moves to surface Pressure Time FIGURE 5.29 Maximum bubble pressure method of determining the surface tension of liquids. A sensor measures the change in pressure with time as a gas bubble is forced into a liquid, which provides information about surface tension. 229 Interfacial Properties and Their Characterization solutions by varying the bubble frequency. The age of the gas–liquid interface is approximately half the time interval between detachment of successive bubbles, and can be varied by changing the flow rate of the gas. It is therefore possible to monitor adsorption kinetics of emulsifiers by monitoring the surface tension as a function of bubble frequency. The dynamic surface tension decreases as the bubble frequency increases, because there is less time for the emulsifier molecules to move to the surface of the droplets. The variation in the dynamic surface tension with bubble frequency therefore gives an indication of the speed at which emulsifier molecules are adsorbed to the surface. This information is important to food manufacturers, because the adsorption rate of emulsifiers determines the size of the droplets produced during homogenization: the faster the rate, the smaller is the droplet size (Chapter 6). Nevertheless, it should be pointed out that the maximum bubble pressure method can only be used to measure dynamic surface tensions down to about 50 ms, whereas many surfactants have faster adsorption rates than this. More rapid techniques are therefore needed to study these systems, such as the oscillating jet, capillary wave, punctured membrane, or microfluidic methods described elsewhere (Couper 1993, Stang et al. 1994, Steegmans et al. 2009). One of the major advantages of the maximum bubble method is that it can be used to analyze optically opaque liquids, because it is not necessary to visually observe the bubbles during the measurements. In addition, it is not necessary to know the contact angle of the liquid, because the maximum pressure occurs when the surface tension acts in a completely vertical direction. 5.8 Interfacial Rheology and Its Measurement 5.8.1 Factors Influencing Interfacial Rheology The rheological properties of the interfacial layer surrounding emulsion droplets influence the formation, stability, and texture of many food emulsions, and therefore, it is important to establish the factors that influence interfacial rheology (Murray and Dickinson 1996, Bos and van Vliet 2001, Javadi et al. 2013, Sagis and Scholten 2014). Interfacial rheology has been defined as the “study of the mechanical and flow properties of adsorbed layers at fluid interfaces” (Murray and Dickinson 1996). When an emulsion is subjected to mechanical agitation, the surfaces of the droplets experience a number of different types of deformation as a result of the stresses acting upon them (Walstra 2003). Stresses may cause different regions of the interface to move relative to each other, without altering the overall surface area, which is known as interfacial shear deformation. Alternatively, they may cause the surface area to expand or contract (like the surface of a balloon when it is inflated or deflated), which is known as interfacial dilational deformation. An interface may have solid-like characteristics that are described by an interfacial elastic constant, or fluid-like characteristics that are described by an interfacial viscosity. In practice, most interfaces have partly “solid-like” and partly “fluid-like” characteristics and therefore exhibit viscoelastic behavior. It is convenient to assume that an interface is infinitesimally thin so that it can be treated as a two-dimensional plane, because its rheological characteristics can then be described using the twodimensional analogs of the relationships used to characterize bulk materials (Chapter 8). A number of the most important rheological characteristics of interfaces are summarized in Table 5.4, where τi is the interfacial shear stress, σi is the interfacial shear strain, ηi is the interfacial shear viscosity, Gi is the interfacial shear modulus, εi is the interfacial dilational elasticity, κ i is the interfacial dilational viscosity, A is TABLE 5.4 Rheological Characteristics of Interfaces Viscosity Shear deformation ti = hi s i Dilatational deformation dg i = e dA A Elasticity τi = Giσi dg i = k A dg dA 230 Food Emulsions: Principles, Practices, and Techniques ( ) ( ) the surface area, and γ is the surface or interfacial tension. In general, Gi =Gi¢ + iGi² and e i =e¢i + ie²i are frequency-dependent complex numbers, whose real part is related to the storage modulus (elastic part) and whose imaginary part is related to the loss modulus (liquid part). For a predominantly elastic material, Gi¢ /Gi² or e¢i / e²i > 1, whereas for a predominantly fluid material, Gi¢ /Gi² or e¢i /e²i < 1. The frequency dependence of the shear and elastic modulus depend on the time taken for any molecular rearrangements to occur in the interface relative to the time that the deforming stress is applied. Generally, those emulsifiers that tend to undergo extensive entanglement or cross-linking at an interface will tend to form a layer that has a high interfacial viscosity or elastic modulus, such as globular proteins and some polysaccharides. On the other hand, emulsifiers that do not strongly entangle or crosslink tend to form interfacial layers with relatively low viscosities or elastic moduli, for example, small molecule surfactants and flexible biopolymers (such as casein). The interfacial rheology is therefore strongly governed by factors that alter the nature and strength of the interactions between the molecules adsorbed at the interface, for example, emulsifier concentration, temperature, pH, and ionic strength. For example, the interfacial rheology of globular proteins tends to increase over time due to conformational changes that lead to interfacial aggregation. 5.8.2 Characterization of Interfacial Rheology 5.8.2.1 Measurement of Interfacial Shear Rheology A variety of experimental methods have been developed to measure the shear rheology of surfaces and interfaces (Murray and Dickinson 1996, Derkach et al. 2009, Sagis and Scholten 2014). One of the most commonly used methods is analogous to the concentric cylinder technique used to measure the shear properties of bulk materials (Chapter 8). The sample to be analyzed is placed in a temperature-controlled vessel, and a thin disk is placed in the plane of the interface that separates the two phases, for example, water and air or oil and water (Figure 5.30). The vessel is then rotated, and the torque on the disk is measured. The sample can be analyzed in a number of ways depending on whether it is solid-like, liquid-like, or viscoelastic. For liquid-like interfaces, the shear viscosity is determined by measuring the torque on the disk as the vessel is rotated continuously. For solid-like interfaces, the shear modulus is determined by measuring the torque on the disk after the vessel has been moved to a fixed angle. For viscoelastic interfaces, the complex shear modulus is usually determined by measuring the torque continuously as the vessel is made to oscillate backward and forward at a certain frequency and angle. The interfacial shear viscosity or elasticity of surfactant layers is usually several orders of magnitude less than that of biopolymer layers, because biopolymer molecules often become entangled or interact Torque on wire Disk Liquid Rotating dish FIGURE 5.30 Experimental technique for measuring interfacial shear rheology. A disk is placed at the air–water or oil– water interface, and the shear stress versus rate of strain is measured as the disk is rotated or oscillated. 231 Interfacial Properties and Their Characterization with each other through various covalent or strong physical forces. The rheology of emulsions depends on the concentration, size, and interactions of the droplets that they contain (Chapter 8). Similarly, the rheology of interfaces depends on the concentration, size, and interactions of the adsorbed emulsifier molecules that they contain. Interfacial shear rheology measurements are particularly useful for providing information about adsorption kinetics, competitive adsorption, and the structure and interactions of molecules at an interface, especially when they are used in conjunction with experimental techniques that provide information about the concentration of the emulsifier molecules at the interface, for example, interfacial tension or radioactive labeling techniques. The concentration of emulsifier molecules at an interface often reaches a constant value after a particular time, while the shear modulus or viscosity continues to increase because of interactions between the adsorbed molecules (Dickinson 1992, Norde 2011). 5.8.2.2 Measurement of Interfacial Dilational Rheology One of the most convenient methods of characterizing both the interfacial tension and dilational rheology of liquid–gas and liquid–liquid interfaces is to use the oscillating drop method (Benjamins and Lucassen-Reynder 2009, Derkach et al. 2009, Lucassen-Reynders et al. 2010, Ravera et al. 2010) Sophisticated analytical instruments based on this principle are commercially available and are being widely used to provide valuable insights into the factors that influence interfacial rheology. One of the fluids is placed into a syringe that is attached to a capillary tube, while the other fluid is poured into a temperature-controlled cuvette (Figure 5.31). A light source and digital camera are used to record the shape of the droplet formed at the tip of the capillary tube when it is dipped into the cuvette. The interfacial tension is determined by analyzing the shape of the drop using a suitable mathematical model. The interfacial dilational rheology is determined by measuring the change in interfacial tension (γ) and interfacial area (A) of the droplet when its volume is increased or decreased in a controlled manner by applying a pressure to the liquid in the capillary tube via a piston: ε = dγ/dlnA. The change in droplet volume can be carried out in a step-wise fashion or periodically (usually sinusoidally). If a sinusoidal wave is used, then the complex viscoelastic modulus of the interfacial layer can be determined at a particular frequency: ε* = ε′ + iε″, where ε′ is the storage modulus and ε″ is the loss modulus. The frequency of the sinusoidal wave applied to the fluid in the capillary tube can be varied, which can provide useful information about the time-scale of molecular rearrangements occurring at the interface. This type of instrument can be used to monitor changes in interfacial rheology with solution composition, time, or temperature. The oscillating drop method is being widely used to study the characteristics of Syringe Applied oscillating pressure Digital camera Light source Oscillating drop Cuvette FIGURE 5.31 Oscillating drop technique for measuring the tension and dilational rheology of surfaces and interfaces. The change in the shape of a droplet suspended in air or another immiscible liquid is measured as an oscillating pressure wave is applied to the liquid. 232 Food Emulsions: Principles, Practices, and Techniques interfacial layers formed by emulsifiers relevant to food systems, and to establish the factors that affect these characteristics. A variety of other analytical techniques are also available for measuring interfacial dilational rheology, for example, trough and overflowing cylinder methods. Trough methods measure the surface or interfacial tension of a liquid using a Wilhelmy plate as the interfacial area is varied by changing the distance between two solid barriers that confine the liquid (Figure 5.22). In some instruments, it is possible to vary the distance between the barriers in a sinusoidal fashion so that the complex dilational modulus can be determined. The overflowing cylinder method can also be used to measure the dynamic dilational rheology of surfaces or interfaces. A Wilhelmy plate is used to measure the surface or interfacial tension of a liquid as it is continuously pumped into a cylinder so that it overflows at the edges. The dilational viscosity is measured as a function of the surface age by altering the rate at which the liquid is pumped into the cylinder. The interfacial dilational rheology of liquids can also be determined by capillary wave methods (Maestro et al. 2012). In these methods, a laser beam reflected from the surface of a liquid is used to determine the amplitude and wavelength of the surface waves, which can then be related to the dilational modulus or viscosity of the surface using an appropriate theory. These surface waves are believed to play an important role in the coalescence of certain types of emulsion droplets (Chapter 7), and therefore, this technique may provide information that has direct practical importance to food scientists. It should be noted that the dilational rheology of an interface is often influenced by the adsorption kinetics of emulsifiers (Murray and Dickinson 1996). When a surface undergoes a dilational expansion, the concentration of emulsifier per unit area decreases and therefore, its surface tension increases, which is energetically unfavorable. The dilational elasticity or viscosity is a measure of this resistance of the surface to dilation (Table 5.4). When there are emulsifier molecules present in the surrounding liquid, they may be adsorbed to the surface during dilation and thereby reduce the surface tension. The dilational rheology therefore depends on the rate at which emulsifier molecules are adsorbed to a surface relative to the rate at which the interfacial area is changed, which is determined by emulsifier concentration, molecular structure, and the prevailing environmental conditions (Section 5.3). The faster the molecules adsorb to the freshly formed interface, the lower is the resistance to dilation, and therefore, the lower is the dilational modulus or viscosity. When an interface undergoes dilational compression, some of the emulsifier molecules may leave the interface to reduce the resulting strain, and therefore, the desorption rate may also influence the rheological characteristics of an interface. 5.9 Chemical and Biochemical Properties of Interfaces There has been an increasing interest in understanding the chemical and biochemical properties of the interfacial layers surrounding emulsion droplets due to the recent focus on the design of food products with improved nutritional attributes. A number of bioactive components that are incorporated into oil droplets, such as polyunsaturated lipids, are highly susceptible to chemical degradation due to lipid oxidation. The chemical stability of these bioactive agents can often be improved by careful design of the composition and structure of the interfacial layer coating the oil droplets (McClements and Decker 2000, Waraho et al. 2010, 2011). For example, lipid oxidation can be suppressed by including antioxidant substances in the interfacial layer (such as proteins), by having cationic interfaces that repel transition metals that normally catalyze oxidation (such as Fe2+), or by designing interfaces that limit contact between pro-oxidants and lipids (such as thick and densely packed interfaces). There has also been interest in controlling the behavior of oil droplets within the human gastrointestinal tract (GIT) by altering interfacial properties. In particular, food scientists are trying to produce specific nutritional or biochemical effects by altering droplet characteristics: increased satiety, improved bioavailability, and targeted release (Chapter 11). A number of strategies have been developed to control the gastrointestinal fate of emulsified lipids by altering their interfacial properties. Droplet interfaces with different properties (e.g., charge, thickness, polarity, digestibility) can be fabricated by using different kinds of emulsifiers, for example, specific surfactants, phospholipids, proteins, polysaccharides, or solid particles. The stability of oil droplets to flocculation, coalescence, creaming, and digestion in different regions of the GIT 233 Interfacial Properties and Their Characterization depends on the nature of the emulsifier coating them (Hur 2009, Torcello-Gomez et al. 2012, Singh and Ye 2013). Alternatively, oil droplets can be coated with one or more layers of dietary fiber using electrostatic deposition methods (McClements 2010). These dietary fiber coatings can be designed so that their permeability or integrity changes in different regions of the GIT, thereby controlling oil droplet stability to lipid digestion and bioactive release. In addition, interfacial cross-linking methods can be used to alter the gastrointestinal fate of ingested oil droplets, such as physical, chemical, or enzyme methods. This area is likely to continue to grow in importance in the near future. 5.10 Practical Implications of Interfacial Phenomena In this section, we consider a number of the important practical implications of interfacial properties for bulk liquids and emulsions. 5.10.1 Properties of Curved Interfaces The majority of surfaces or interfaces found in food emulsions are curved, rather than planar. The curvature of an interface alters its characteristics in a number of ways (Hiemenz and Rajagopalan 1997, Jonsson et al. 2003, Israelachvili 2011). The interfacial tension tends to cause an emulsion droplet to shrink in volume so as to reduce the unfavorable contact area between the oil and water phases. As the droplet shrinks, there is an increase in its internal pressure because of the compression of the water molecules. Eventually, an equilibrium is reached where the inward stress due to the interfacial tension is balanced by the outward stress associated with compressing the bonds between the liquid molecules inside the droplet.* At equilibrium, the pressure within the droplet is larger than that outside, and can be related to the interfacial tension and radius of the droplets using the Young–Laplace equation: Dp = 2g r (5.24) This equation indicates that the pressure difference across the interface of an emulsion droplet increases as the interfacial tension increases or the size of the droplet decreases. The properties of a material depend on the pressure exerted on it, and so the properties of a material within a droplet are different from those of the same material in bulk. This effect is usually negligible for liquids and solids that are contained within particles that have radii greater than a few micrometers, but it does become significant for smaller particles. For example, the solubility of a substance in a liquid surrounding it increases as the radius of the spherical particle containing the substance decreases (Sun et al. 2012): S æ 2gv ö = exp ç ÷ S* è rRT ø (5.25) where S is the solubility of the substance when it is contained within a spherical particle of radius r S* is the solubility of the same substance when it is contained within a spherical particle of infinite radius (i.e., the solubility of the bulk substance) v is the molar volume of the substance For a typical food oil contained within a droplet surrounded by water (v = 10 −3 m3 mol−1, γ = 10 mJ m−2), the values of S/S* are 2.24, 1.08, 1.01, and 1.00 for oil droplets with radii of 0.01, 0.1, 1, and 10 μm, * The shrinkage of a droplet due to the interfacial tension is usually negligibly small, because liquids have a very low compressibility. 234 Food Emulsions: Principles, Practices, and Techniques respectively. The dependence of the solubility of substances on the size of the particles that they are contained within has important implications for the stability of emulsion droplets, fat crystals, and ice crystals in many foods because of Ostwald ripening, that is, the growth of larger particles at the expense of smaller ones due to diffusion of the substance contained within the particles through the intervening medium (Chapter 7). So far it has been assumed that the interfacial tension of a droplet is independent of its radius. Experimental work has indicated that this assumption is valid for oil droplets, even down to sizes where they only contain a few molecules, but that it is invalid for water droplets below a few nanometers because of the disruption of long range hydrogen bonds (Israelachvili 2011). It should also be noted that the droplets in emulsions are usually covered with a layer of emulsifier molecules, which will alter the interfacial tension and the mass transfer rate of molecules across the interface. 5.10.2 Contact Angles and Wetting In food systems, we are often interested in the ability of a liquid to spread over or “wet” the surface of another material. In some situations, it is desirable for a liquid to spread over a surface (e.g., when coating a food with an edible film), while in other situations, it is important that a liquid does not spread (e.g., when designing water-proof packaging). When a drop of liquid is placed on the surface of a material, it may behave in a number of ways, depending on the nature of the interactions between the various types of molecules present. The two extremes of behavior that are observed experimentally are outlined in the following (Figure 5.32): 1. Poor wetting: The liquid gathers up into a lens, rather than spreading across the surface of a material. 2. Good wetting: The liquid spreads over the surface of the material to form a thin film. The situation that occurs in practice depends on the relative magnitude of the interactions between the various types of molecules involved, that is, solid–liquid, solid–gas, and liquid–gas. A system tends to organize itself so that it can maximize the number of thermodynamically favorable interactions and minimize the number of thermodynamically unfavorable interactions between the molecules. Consider what may happen when a drop of liquid is placed on a solid surface (Figure 5.33). If the liquid remained as a lens, there would be three different interfaces: solid–liquid, solid–gas, and liquid–gas, each with its own interfacial or surface tension. If the liquid spreads over the surface, there would be a decrease in the area of the solid–gas interface, but an increase in the areas of both the liquid–gas and solid–liquid interfaces. The tendency for a liquid to spread therefore depends on the magnitude of the solid–gas interactions (γSG) compared to the magnitude of the solid–liquid and liquid–gas interactions that replace it (γSL + γLG). This situation is conveniently described by a spreading coefficient, which is defined as (Hunter 1986, Norde 2011) S = γSG − (γSL + γLG) (5.26) If the interfacial tension of the solid–gas interface is greater than the sum of the interfacial tensions associated with the solid–liquid and liquid–gas interfaces (γSG > γSL + γLG), then S is positive and the liquid tends to spread over the surface to reduce the thermodynamically unfavorable contact area between the solid and the gas. On the other hand, if the interfacial tension associated with the solid–gas interface is less than the sum of the interfacial tensions associated with the solid–liquid and liquid–gas interfaces (γSG < γSL + γLG), then S is negative and the liquid tends to form a lens. The shape of a droplet on a solid surface can be predicted by carrying out an equilibrium force-balance at the point on the surface where the solid, liquid, and gas meet (Figure 5.33) using the Young–Dupré Equation (Norde 2011): g SG = g SL + g LG cos q (5.27) 235 Interfacial Properties and Their Characterization Gas Liquid Poor wetting Solid Gas γ LG γ Liquid γ SG SL Solid Gas Liquid Good wetting Solid FIGURE 5.32 The wetting of a surface by a liquid depends on a delicate balance of molecular interactions between solid, liquid, and gas phases. Gas γLG Liquid θ γSG γSL Solid FIGURE 5.33 Force balance of a droplet at a solid–gas interface. The shape of the droplet formed depends on the relative magnitude of the interfacial tensions at the liquid–gas, solid–gas, and solid–liquid interfaces. 236 Food Emulsions: Principles, Practices, and Techniques so that cosq = g SG - g SL g LG (5.28) Here θ is known as the contact angle, which is the angle of a tangent drawn at the point where the liquid contacts the surface (Figure 5.33). The shape of a droplet on a surface can therefore be predicted from knowledge of the contact angle: the smaller θ, the greater is the tendency for the liquid to spread over the surface. The Young–Dupré Equation contains two parameters that cannot be determined independently (γSG and γSL). This equation is therefore only useful when it is used in combination with another equation that allows one to estimate one of the unknown terms. So far we have only considered the situation where a liquid spreads over a solid surface, but similar equations can be used to consider other three component systems, such as a liquid spreading over the surface of another liquid (e.g., oil, water, and air), or of a solid particle at an interface between two other liquids (e.g., a fat crystal at an oil–water interface). The latter case is important when considering the nucleation and location of fat crystals in oil droplets, and has a pronounced influence on the stability and rheology of many important food emulsions, including milk, cream, butter, and whipped cream (Hartel 2001, Walstra 2003, Goff and Hartel 2013). The earlier equations assume that the materials involved are completely insoluble in each other, so that the values of γSG, γSL, and γLG (or the equivalent terms for other three component systems) are the same as those for pure systems. If the materials are partially miscible, then the interfacial tensions will change over time until the equilibrium is reached (Hunter 1994). The solubility of one component in another generally leads to a decrease in the interfacial tension. This means that the shape that a droplet adopts on a surface may change with time, for example, a spread liquid may gather into a lens, or vice versa, depending on the magnitude of the changes in the various surface or interfacial tensions. When surface-active solutes are present in the system, there will be changes in the relative magnitudes of the various surface and interfacial tensions, which may drastically alter the ability of a liquid to spread over or wet a surface. The contact angle of a liquid can conveniently be measured using a digital camera, which is often attached to a computer with image analysis software. A droplet of the liquid to be analyzed is placed on a surface, and its shape is recorded via the digital camera. The contact angle is determined by analyzing the shape of the droplet using an appropriate theoretical model (Hiemenz and Rajagopalan 1997). There are a number of important practical considerations that must be taken into account in order to perform accurate contact angle measurements, including the effects of surface roughness, surface heterogeneity, and adsorption of vapor or surfactants to the solid surface (Chau et al. 2009, Norde 2011). The advantages and disadvantages of a variety of other techniques available for measuring contact angles have been considered elsewhere (Hunter 1986, Chau 2009). The concepts of a contact angle and a spreading coefficient are useful for explaining a number of important phenomena that occur in food emulsions. The contact angle determines the distance that a fat crystal protrudes from the surface of a droplet into the surrounding water, and whether nucleation occurs within the interior of a droplet or at the oil–water interface (Rousseau 2000, Fredrick et al. 2010, McClements 2012, Douaire et al. 2014). It also determines the amount of liquid that is drawn into a capillary tube and the shape of the meniscus at the top of the liquid (Hunter 1986, Hiemenz and Rajagopalan 1997). The contact angle also determines the effectiveness of particulate matter at stabilizing emulsion droplets against aggregation, since it determines how far these particles protrude out of the droplets, that is, Pickering stabilization (Aveyard et al. 2003, Rousseau 2013). Knowledge of the contact angle is also often required in order to make an accurate measurement of the surface or interfacial tension of a liquid (Couper 1993). Finally, measurement of the contact angle of a water droplet placed on a surface can be used to quantify the surface hydrophobicity of that surface (Norde 2011). 5.10.3 Capillary Rise and Meniscus Formation The surface tension of a liquid governs the rise of liquids in capillary tubes and the formation of menisci (curved surfaces) at the top of liquids (Hiemenz and Rajagopalan 1997, Evans and Wennerstrom 1999). 237 Interfacial Properties and Their Characterization When a glass capillary tube is dipped into a beaker of water, the liquid climbs up the tube and forms a curved surface (Figure 5.34). The origin of this phenomenon is the imbalance of intermolecular forces at the various surfaces and interfaces in the system. When water climbs up the capillary tube, some of the air–glass contact area is replaced by water–glass contact area, while the air–water contact area remains fairly constant. This occurs because the imbalance of molecular interactions between glass and air is much greater than that between glass and water. Consequently, the system attempts to maximize the glass–water contacts and minimize the glass–air contacts, by having the liquid climb up the inner surface of the capillary tube. This process is opposed by the downward gravitational pull of the liquid. When the liquid has climbed to a certain height, the surface energy it gains by optimizing the number of favorable water–glass interactions is exactly balanced by the potential energy that must be expended to raise the mass of water up the tube. A mathematical analysis of this equilibrium leads to the derivation of the following equation: g= Drghr 2cos q (5.29) where g is the gravitational constant h is the height that the meniscus rises above the level of the water r is the radius of the capillary tube Δρ is the difference in density between the water and the air θ is the contact angle (Figure 5.34) When a capillary tube is made from a solid material that is completely wetted by the fluid into which it is dipped (e.g., θ = 0°, cos θ = 1), then the fluid moves upward, for example, glass and water. On the other hand, when a capillary tube is made from a solid material that is poorly wetted by the fluid into which it is dipped (e.g., θ = 180°, cos θ = −1), then the fluid will move downward, for example, glass and mercury. The aforementioned equation indicates that the surface tension can be estimated from a measurement of the height that a liquid rises up a capillary tube and the contact angle: the greater the surface tension, the higher the liquid rises up the tube. This is one of the oldest and simplest methods of determining the surface tension of pure liquids, but it has a number of problems that limit its application to emulsifier solutions (Couper 1993). For these systems, it is better to use the surface and interfacial tension–­ measuring devices described earlier in this chapter. r θ Capillary tube h Liquid FIGURE 5.34 The rise of a liquid up a capillary tube is a result of its surface tension. The greater the surface tension, the higher will the liquid rise. 238 Food Emulsions: Principles, Practices, and Techniques Capillary forces are responsible for the entrapment of water in biopolymer networks (such as hydrogels) and oil in fat crystal networks (such as in margarine and butter). When the pores in the network are relatively small, the capillary force is strong enough to hold large volumes of liquid, but when the pores exceed a certain size, the capillary forces are no longer strong enough and syneresis or “oiling-off” occurs. Knowledge of the origin of capillary forces is therefore important for understanding the relationship between the microstructure of foods and many of their quality attributes. 5.10.4 Interfacial Phenomenon in Food Emulsions We conclude this chapter by briefly outlining some of the most important practical implications of interfacial phenomena for food emulsions. One of the most striking features of food emulsions when observed under a microscope is the sphericity of the droplets. Droplets tend to be spherical, because this shape minimizes the thermodynamically unfavorable contact area between oil and water molecules, which is described by the Laplace equation (see Equation 5.23). Droplets become nonspherical when they experience an external force that is large enough to overcome the Laplace pressure, for example, gravity, centrifugal forces, or mechanical agitation. The magnitude of the force needed to deform a droplet decreases as their interfacial tension decreases or their radius increases. This accounts for the ease at which the relatively large droplets in highly concentrated emulsions, such as mayonnaise, are deformed into polygons. The thermodynamic driving force for coalescence and “oiling-off” is the interfacial tension between the oil and water phases caused by the imbalance of molecular forces across an oil–water interface (Chapter 7). On the other hand, the ability of the interfacial layer to resist rupture or to prevent droplets from coming close together is responsible for their kinetic stability (Chapter 3). The reason that surfaceactive molecules adsorb to an air–fluid or oil–water interface is because of their ability to reduce the surface or interfacial tension (Section 5.2.2). The tendency of a liquid to spread over the surface of another material or to remain as a lens depends on the relative magnitude of the interfacial and surface tensions between the different types of substance involved (Section 5.9.2). The formation of stable nuclei in a liquid is governed by the interfacial tension between the crystal and the melt (Chapter 4). The larger the interfacial tension between the solid and liquid phases, the greater is the degree of supercooling required to produce stable nuclei. The interfacial tension also determines whether an impurity is capable of promoting heterogeneous nucleation. The solubility of a substance increases as the size of the particle containing it decreases (Section 5.9.1). If a suspension contains particles (emulsion droplets, fat crystals, ice crystals, or air bubbles) that have a range of different sizes, then there is a greater concentration of the substance dissolved in the region immediately surrounding the smaller particles than surrounding the larger particles. Consequently, there is a concentration gradient that causes material to move from the smaller particles to the large particles (Walstra 2003). With time, this process manifests itself as a growth of the large particles at the expense of the smaller ones, which is referred to as Ostwald ripening or disproportionation (Chapter 7). This process may be responsible for the growth of emulsion droplets, fat crystals, ice crystals, and air bubbles in certain types of food emulsions, which often has a detrimental effect on food quality. For example, the growth of ice crystals in ice-cream causes the product to be perceived as “gritty.” It should be clear from this chapter, that even though the interfacial region only comprises a small fraction of the total volume of an emulsion, it plays an extremely important role in determining their bulk physicochemical properties. REFERENCES Arwin, H. (2014). Adsorption of proteins at solid surfaces. In Ellipsometry of Functional Organic Surfaces and Films, K. Hinrichs and K. J. Eichhorn, eds., vol. 52, pp. 29–46. Berlin, Germany: Springer-Verlag. Atkinson, P. J., E. Dickinson, D. S. Horne, F. A. M. Leermakers, and R. M. Richardson (1996). Theoretical and experimental investigations of adsorbed protein structure at a fluid interface. Berichte Der BunsenGesellschaft-Physical Chemistry Chemical Physics 100(6): 994–998. Interfacial Properties and Their Characterization 239 Atkinson, P.J., E. Dickinson, D.S. Horne, R.M. Richardson (1995). Neutron reflectivity of adsorbed β-casein and β-lactoglobulin at the air/water interface. Journal of the Chemical Society - Faraday Transactions 91: 2847. Aveyard, R., B. P. Binks, and J. H. Clint (2003). Emulsions stabilised solely by colloidal particles. Advances in Colloid and Interface Science 100: 503–546. Bajpai, A. K. and M. Rajpoot (1999). Adsorption techniques—A review. Journal of Scientific and Industrial Research 58(11): 844–860. Belitz, H., W. Grosch, and P. Schieberle (2009). Food Chemistry. Berlin, Germany: Springer. Benjamins, J. and E. H. Lucassen-Reynder (2009). Interfacial rheology of adsorbed protein layers. In Interfacial Rheology, R. Miller and L. Liggieri, eds., vol. 1, pp. 253–302. Boca Raton, FL: CRC Press. Blonk, J. C. G. and H. Vanaalst (1993). Confocal scanning light-microscopy in food research. Food Research International 26(4): 297–311. Bos, M. A. and T. van Vliet (2001). Interfacial rheological properties of adsorbed protein layers and surfactants: A review. Advances in Colloid and Interface Science 91(3): 437–471. Brady, J. W. (2013). Introductory Food Chemistry. Ithaca, NY: Cornell University Press. Chau, T. T. (2009). A review of techniques for measurement of contact angles and their applicability on mineral surfaces. Minerals Engineering 22(3): 213–219. Chau, T. T., W. J. Bruckard, P. T. L. Koh, and A. V. Nguyen (2009). A review of factors that affect contact angle and implications for flotation practice. Advances in Colloid and Interface Science 150(2): 106–115. Claesson, P. M., E. Blomberg, J. C. Froberg, T. Nylander, and T. Arnebrant (1995). Protein interactions at solid-surfaces. Advances in Colloid and Interface Science 57: 161–227. Corredig, M. and D. G. Dalgleish (1995). A differential microcalorimetric study of whey proteins and their behavior in oil-in-water emulsions. Colloids and Surfaces B-Biointerfaces 4(6): 411–422. Couper, A. (1993). Surface tension and its measurement. In Physical Methods of Chemistry, B. W. Rossiter and R. C. Baetzold, eds. New York: John Wiley & Sons. Coupland, J. N. and D. J. McClements (1996). Lipid oxidation in food emulsions. Trends in Food Science & Technology 7(3): 83–91. Cristofolini, L. (2014). Synchrotron x-ray techniques for the investigation of structures and dynamics in interfacial systems. Current Opinion in Colloid & Interface Science 19(3): 228–241. Dalgleish, D. G. (1996). Conformations and structures of milk proteins adsorbed to oil-water interfaces. Food Research International 29(5–6): 541–547. Dalgleish, D. G. (1997). Food emulsions stabilized by proteins. Current Opinion in Colloid & Interface Science 2(6): 573–577. Damodaran, S., K. L. Parkin, and O. R. Fennema (2007). Fennema’s Food Chemistry. Boca Raton, FL: CRC Press. Derkach, S. R., J. Kragel, and R. Miller (2009). Methods of measuring rheological properties of interfacial layers (Experimental methods of 2D rheology). Colloid Journal 71(1): 1–17. Dickinson, E. (1992a). Introduction to Food Colloids. Oxford, U.K.: Oxford University Press. Dickinson, E. (1992b). Structure and composition of adsorbed protein layers and the relationship to emulsion stability. Journal of the Chemical Society-Faraday Transactions 88(20): 2973–2983. Dickinson, E. (2001). Milk protein interfacial layers and the relationship to emulsion stability and rheology. Colloids and Surfaces B-Biointerfaces 20(3): 197–210. Dickinson, E. and S. T. Hong (1994). Surface coverage of beta-lactoglobulin at the oil-water interface— Influence of protein heat-treatment and various emulsifiers. Journal of Agricultural and Food Chemistry 42(8): 1602–1606. Dickinson, E., G. Iveson, and S. Tanai (1993). Competitive adsorption in protein stabilized emulsions containing oil-soluble and water-soluble surfactants. In Food Colloids and Polymers: Stability and Mechanical Properties, E. Dickinson and P. Walstra, eds., pp. 312–322. Cambridge, U.K.: Royal Society of Chemistry. Dickinson, E. and Y. Matsumura (1991). Time-dependent polymerization of beta-lactoglobulin through disulfide bonds at the oil-water interface in emulsions. International Journal of Biological Macromolecules 13(1): 26–30. Douaire, M., V. di Bari, J. E. Norton, A. Sullo, P. Lillford, and I. T. Norton (2014). Fat crystallisation at oilwater interfaces. Advances in Colloid and Interface Science 203: 1–10. 240 Food Emulsions: Principles, Practices, and Techniques Dudkiewicz, A., K. Tiede, K. Loeschner, L. H. S. Jensen, E. Jensen, R. Wierzbicki, A. B. A. Boxall, and K. Molhave (2011). Characterization of nanomaterials in food by electron microscopy. TrAC Trends in Analytical Chemistry 30(1): 28–43. Dukhin, S. S., G. Kretzschmar, and R. Miller (1995). Dynamics of Adsorption at Liquid Interfaces. New York: Elsevier Science. Elwing, H. (1998). Protein absorption and ellipsometry in biomaterial research. Biomaterials 19(4–5): 397–406. Evans, E. D. and W. Wennerstrom (1999). The Colloidal Domain: Where Physics, Chemistry and Biology Meet. New York: Wiley-VCH. Fainerman, V. B., E. Lucassen-Reynders, and R. Miller (1998). Adsorption of surfactants and proteins at fluid interfaces. Colloids and Surfaces A-Physicochemical and Engineering Aspects 143(2–3): 141–165. Fang, Y. and D. G. Dalgleish (1998). The conformation of alpha-lactalbumin as a function of pH, heat treatment and adsorption at hydrophobic surfaces studied by FTIR. Food Hydrocolloids 12(2): 121–126. Fleer, G., M. A. Cohen Stuart, J. M. H. M. Scheutjens, T. Cosgrove, and B. Vincent (1993). Polymers at Interfaces. New York: Springer Scientific. Fredrick, E., P. Walstra, and K. Dewettinck (2010). Factors governing partial coalescence in oil-in-water emulsions. Advances in Colloid and Interface Science 153(1–2): 30–42. Goff, H. D. and R. W. Hartel (2013). Ice Cream. New York: Springer. Gunning, P. A., A. R. Mackie, A. P. Gunning, N. C. Woodward, P. J. Wilde, and V. J. Morris (2004). Effect of surfactant type on surfactant-protein interactions at the air-water interface. Biomacromolecules 5(3): 984–991. Hartel, R. W. (2001). Crystallization in Foods. Gaithersburg, MD: Aspen Publishers. Hartland, S. (2004). Surface and Interfacial Tension: Measurement, Theory, and Applications. Boca Raton, FL: CRC Press. Heinrich, F. and M. Loesche (2014). Zooming in on disordered systems: Neutron reflection studies of proteins associated with fluid membranes. Biochimica et Biophysica Acta-Biomembranes 1838(9): 2341–2349. Hiemenz, P. C. and R. Rajagopalan (1997). Principles of Colloid and Surface Chemistry. New York: Marcel Dekker. Hsu, J. P. and A. Nacu (2003). Behavior of soybean oil-in-water emulsion stabilized by nonionic surfactant. Journal of Colloid and Interface Science 259(2): 374–381. Hunter, R. J. (1986). Foundations of Colloid Science. Oxford, U.K.: Oxford University Press. Hunter, R. J. (1994). Introduction to Modern Colloid Science. Oxford, U.K.: Oxford University Press. Hur, S. J., E. A. Decker, and D. J. McClements (2009). Influence of initial emulsifier type on microstructural changes occurring in emulsified lipids during in vitro digestion. Food Chemistry 114(1): 253–262. Husband, F. A., M. J. Garrood, A. R. Mackie, G. R. Burnett, and P. J. Wilde (2001). Adsorbed protein secondary and tertiary structures by circular dichroism and infrared spectroscopy with refractive index matched emulsions. Journal of Agricultural and Food Chemistry 49(2): 859–866. Israelachvili, J. (2011). Intermolecular and Surface Forces, 3rd edn. London, U.K.: Academic Press. Javadi, A., N. Mucic, M. Karbaschi, J. Y. Won, M. Lotfi, A. Dan, V. Ulaganathan et al. (2013). Characterization methods for liquid interfacial layers. European Physical Journal-Special Topics 222(1): 7–29. Jonsson, K., B. Kronberg, B. Lindman, and B. Holmberg (2003). Surfactants and Polymers in Aqueous Solution. Chichester, U.K.: John Wiley & Sons. Kallay, N., V. Hlady, J. Jednacak-Biscan, and S. Milonjic (1993). Techniques for the study of adsorption from solution. In Physical Methods of Chemistry, Investigations of Surfaces and Interfaces, B. W. Rossiter and R. C. Baetzold, eds., vol. IXA. New York: John Wiley & Sons. Karakashev, S. I., A. V. Nguyen, and J. D. Miller (2008). Equilibrium adsorption of surfactants at the gasliquid interface. In Interfacial Processes and Molecular Aggregation of Surfactants, R. Narayanan, ed., vol. 218, pp. 25–55. Berlin, Germany: Springer-Verlag. Keerati-u-rai, M. and M. Corredig (2009). Heat-induced changes in oil-in-water emulsions stabilized with soy protein isolate. Food Hydrocolloids 23(8): 2141–2148. Kim, H. J., E. A. Decker, and D. J. McClements (2002a). Impact of protein surface denaturation on droplet flocculation in hexadecane oil-in-water emulsions stabilized by beta-lactoglobulin. Journal of Agricultural and Food Chemistry 50(24): 7131–7137. Interfacial Properties and Their Characterization 241 Kim, H. J., E. A. Decker, and D. J. McClements (2002b). Role of postadsorption conformation changes of beta-lactoglobulin on its ability to stabilize oil droplets against flocculation during heating at neutral pH. Langmuir 18(20): 7577–7583. Kralova, I. and J. Sjoblom (2009). Surfactants used in food industry: A review. Journal of Dispersion Science and Technology 30(9): 1363–1383. Lee, S.-H., T. Lefevre, M. Subirade, and P. Paquin (2007). Changes and roles of secondary structures of whey protein for the formation of protein membrane at soy oil/water interface under high-pressure homogenization. Journal of Agricultural and Food Chemistry 55(26): 10924–10931. Liu, S. Y. and Y. F. Wang (2011). A review of the application of atomic force microscopy (AFM) in food science and technology. Advances in Food and Nutrition Research 62: 201–240. Lu, J. R., X. Zhao, and M. Yaseen (2007). Protein adsorption studied by neutron reflection. Current Opinion in Colloid & Interface Science 12(1): 9–16. Lucassen-Reynders, E. H., J. Benjamins, and V. B. Fainerman (2010). Dilational rheology of protein films adsorbed at fluid interfaces. Current Opinion in Colloid & Interface Science 15(4): 264–270. Mackie, A. R., A. P. Gunning, M. J. Ridout, P. J. Wilde, and J. R. Patino (2001). In situ measurement of the displacement of protein films from the air/water interface by surfactant. Biomacromolecules 2(3): 1001–1006. Mackie, A. R., A. P. Gunning, P. J. Wilde, and V. J. Morris (2000). Orogenic displacement of protein from the oil/water interface. Langmuir 16(5): 2242–2247. Maestro, A., C. Kotsmar, A. Javadi, R. Miller, F. Ortega, and R. G. Rubio (2012). Adsorption of beta-caseinsurfactant mixed layers at the air-water interface evaluated by interfacial rheology. Journal of Physical Chemistry B 116(16): 4898–4907. McClements, D. J. (2002). Modulation of globular protein functionality by weakly interacting cosolvents. Critical Reviews in Food Science and Nutrition 42(5): 417–471. McClements, D. J. (2004). Protein-stabilized emulsions. Current Opinion in Colloid & Interface Science 9(5): 305–313. McClements, D. J. (2010). Design of nano-laminated coatings to control bioavailability of lipophilic food components. Journal of Food Science 75(1): R30–R42. McClements, D. J. (2012). Crystals and crystallization in oil-in-water emulsions: Implications for emulsionbased delivery systems. Advances in Colloid and Interface Science 174: 1–30. McClements, D. J. and E. A. Decker (2000). Lipid oxidation in oil-in-water emulsions: Impact of molecular environment on chemical reactions in heterogeneous food systems. Journal of Food Science 65(8): 1270–1282. Meyer, E. E., K. J. Rosenberg, and J. Israelachvili (2006). Recent progress in understanding hydrophobic interactions. Proceedings of the National Academy of Sciences of the United States of America 103(43): 15739–15746. Miller, R., P. Joos, and V. B. Fainerman (1994). Dynamic surface and interfacial-tensions of surfactant and polymer-solutions. Advances in Colloid and Interface Science 49: 249–302. Milling, A. J. (1999). Surface Characterization Methods: Principles, Techniques, and Applications. Boca Raton, FL: CRC Press. Monahan, F. J., D. J. McClements, and J. B. German (1996). Disulfide-mediated polymerization reactions and physical properties of heated WPI-stabilized emulsions. Journal of Food Science 61(3): 504–509. Morris, V. J., A. P. Gunning, and A. R. Kirby (1999). Atomic Force Microscopy for Biologists. London, U.K.: Imperial College Press. Mudunkotuwa, I. A., A. Al Minshid, and V. H. Grassian (2014). ATR-FTIR spectroscopy as a tool to probe surface adsorption on nanoparticles at the liquid-solid interface in environmentally and biologically relevant media. Analyst 139(5): 870–881. Mulder, H. and P. Walstra (1974). Milk Fat Globule: Emulsion Science as Applied to Milk Products and Comparable Foods. Wageningen, the Netherlands: CABI Publishing. Murray, B. S. and E. Dickinson (1996). Interfacial rheology and the dynamic properties of adsorbed films of proteins and surfactants. Food Science and Technology International 2: 131–145. Murray, B. S., R. Xu, and E. Dickinson (2009). Brewster angle microscopy of adsorbed protein films at airwater and oil-water interfaces after compression, expansion and heat processing. Food Hydrocolloids 23(4): 1190–1197. 242 Food Emulsions: Principles, Practices, and Techniques Nielsen, S. S. (2010). Food Analysis. New York: Springer. Nino, M. R. R., C. C. Sanchez, V. P. Ruiz-Henestrosa, and J. M. R. Patino (2005). Milk and soy protein films at the air-water interface. Food Hydrocolloids 19(3): 417–428. Niu, Y. and G. Jin (2013). Surface modification methods to improve behavior of biosensor based on imaging ellipsometry. Applied Surface Science 281: 84–88. Norde, W. (2011). Colloids and Interfaces in Life Sciences and Bionanotechnology. Boca Raton, FL: CRC Press. Pashley, R. M. (2003). Effect of degassing on the formation and stability of surfactant-free emulsions and fine teflon dispersions. Journal of Physical Chemistry B 107(7): 1714–1720. Patino, J. M. R., C. C. Sanchez, and M. R. R. Nino (1999). Morphological and structural characteristics of monoglyceride monolayers at the air-water interface observed by Brewster angle microscopy. Langmuir 15(7): 2484–2492. Pugnaloni, L. A., E. Dickinson, R. Ettelaie, A. R. Mackie, and P. J. Wilde (2004). Competitive adsorption of proteins and low-molecular-weight surfactants: Computer simulation and microscopic imaging. Advances in Colloid and Interface Science 107(1): 27–49. Rabe, M., D. Verdes, and S. Seeger (2011). Understanding protein adsorption phenomena at solid surfaces. Advances in Colloid and Interface Science 162(1–2): 87–106. Raigoza, A. F., J. W. Dugger, and L. J. Webb (2013). Review: Recent advances and current challenges in scanning probe microscopy of biomolecular surfaces and interfaces. ACS Applied Materials & Interfaces 5(19): 9249–9261. Rampon, V., A. Riaublanc, M. Anton, C. Genot, and D. J. McClements (2003). Evidence that homogenization of BSA-stabilized hexadecane-in-water emulsions induces structure modification of the nonadsorbed protein. Journal of Agricultural and Food Chemistry 51(20): 5900–5905. Ravera, F., G. Loglio, and V. I. Kovalchuk (2010). Interfacial dilational rheology by oscillating bubble/drop methods. Current Opinion in Colloid & Interface Science 15(4): 217–228. Razumovsky, L. and S. Damodaran (2001). Incompatibility of mixing of proteins in adsorbed binary protein films at the air-water interface. Journal of Agricultural and Food Chemistry 49(6): 3080–3086. Rousseau, D. (2000). Fat crystals and emulsion stability—A review. Food Research International 33(1): 3–14. Rousseau, D. (2013). Trends in structuring edible emulsions with Pickering fat crystals. Current Opinion in Colloid & Interface Science 18(4): 283–291. Sagis, L. M. C. and E. Scholten (2014). Complex interfaces in food: Structure and mechanical properties. Trends in Food Science & Technology 37(1): 59–71. Sanchez, C. C., S. E. M. Ortiz, M. R. R. Nino, M. C. Anon, and J. M. R. Patino (2003). Effect of pH on structural, topographical, and dynamic characteristics of soy globulin films at the air-water interface. Langmuir 19(18): 7478–7487. Scholten, E., L. M. C. Sagis, and E. van der Linden (2006). Effect of bending rigidity and interfacial permeability on the dynamical behavior of water-in-water emulsions. Journal of Physical Chemistry B 110(7): 3250–3256. Scholten, E., R. Tuinier, R. H. Tromp, and H. N. W. Lekkerkerker (2002). Interfacial tension of a decomposed biopolymer mixture. Langmuir 18(6): 2234–2238. Singh, H. and A. Ye (2013). Structural and biochemical factors affecting the digestion of protein-stabilized emulsions. Current Opinion in Colloid & Interface Science 18(4): 360–370. Sitterberg, J., A. Ozcetin, C. Ehrhardt, and U. Bakowsky (2010). Utilising atomic force microscopy for the characterisation of nanoscale drug delivery systems. European Journal of Pharmaceutics and Biopharmaceutics 74(1): 2–13. Stang, M., H. Karbstein, and H. Schubert (1994). Adsorption-kinetics of emulsifiers at oil-water interfaces and their effect on mechanical emulsification. Chemical Engineering and Processing 33(5): 307–311. Steegmans, M. L. J., A. Warmerdam, K. Schroen, and R. M. Boom (2009). Dynamic interfacial tension measurements with microfluidic Y-junctions. Langmuir 25(17): 9751–9758. Stefaniu, C. and G. Brezesinski (2014). X-ray investigation of monolayers formed at the soft air/water interface. Current Opinion in Colloid & Interface Science 19(3): 216–227. Stefaniu, C., G. Brezesinski, and H. Moehwald (2014). Langmuir monolayers as models to study processes at membrane surfaces. Advances in Colloid and Interface Science 208: 197–213. Interfacial Properties and Their Characterization 243 Stokes, D. J. (2003). Recent advances in electron imaging, image interpretation and applications: Environmental scanning electron microscopy. Philosophical Transactions of the Royal Society of London Series A-Mathematical Physical and Engineering Sciences 361(1813): 2771–2787. Sun, J., F. Wang, Y. Sui, Z. She, W. Zhai, C. Wang, and Y. Deng (2012). Effect of particle size on solubility, dissolution rate, and oral bioavailability: Evaluation using coenzyme Q(10) as naked nanocrystals. International Journal of Nanomedicine 7: 5733–5744. Torcello-Gomez, A., A. B. Jodar-Reyes, J. Maldonado-Valderrama, and A. Martin-Rodriguez (2012). Effect of emulsifier type against the action of bile salts at oil-water interfaces. Food Research International 48(1): 140–147. Touhami, A., M. Alexander, M. Kurylowicz, C. Gram, M. Corredig, and J. R. Dutcher (2011). Probing protein conformations at the oil droplet-water interface using single-molecule force spectroscopy. Soft Matter 7(21): 10274–10284. van Kempen, S., K. Maas, H. A. Schols, E. van der Linden, and L. M. C. Sagis (2013). Interfacial properties of air/water interfaces stabilized by oligofructose palmitic acid esters in the presence of whey protein isolate. Food Hydrocolloids 32(1): 162–171. Walstra, P. (2003). Physical Chemistry of Foods. New York: Marcel Decker. Waraho, T., V. Cardenia, E. A. Decker, and D. J. McClements (2010). Lipid oxidation in emulsified food products. In Oxidation in Foods and Beverages and Antioxidant Applications: Management in Different Industry Sectors, E.A. Decker,R.J. Elias, and D.J. McClemetns, eds., vol. 2, issue 200, pp. 306–343. Cambridge, U.K.: Woodhead Publishing. Waraho, T., D. J. McClements, and E. A. Decker (2011). Mechanisms of lipid oxidation in food dispersions. Trends in Food Science & Technology 22(1): 3–13. Wilde, P., A. Mackie, F. Husband, P. Gunning, and V. Morris (2004). Proteins and emulsifiers at liquid interfaces. Advances in Colloid and Interface Science 108: 63–71. Wilde, P. J. (2000). Interfaces: their role in foam and emulsion behaviour. Current Opinion in Colloid & Interface Science 5(3–4): 176–181. Zhai, J. L., L. Day, M. I. Aguilar, and T. J. Wooster (2013). Protein folding at emulsion oil/water interfaces. Current Opinion in Colloid & Interface Science 18(4): 257–271. Zheng, J. K. and L. L. He (2014). Surface-enhanced raman spectroscopy for the chemical analysis of food. Comprehensive Reviews in Food Science and Food Safety 13(3): 317–328. 6 Emulsion Formation 6.1 Introduction Fresh milk is an example of a naturally occurring emulsion that can be consumed directly by human beings. In practice, however, most milk is subjected to a number of processing operations prior to consumption to ensure its safety, to extend its shelf-life, and to create new products. Processing operations, such as homogenization, pasteurization, whipping, churning, enzyme treatment, and aging, are responsible for the wide range of properties exhibited by dairy products, for example, homogenized milk, cream, ice-cream, butter, and cheese (Chapter 12). Unlike dairy products, most other food emulsions are manufactured by combining raw materials that are not normally found together in nature. For example, a salad dressing may be prepared using water, proteins from milk, oil from soybeans, vinegar from apples, and polysaccharides from seaweed. The physicochemical and sensory properties of a particular food emulsion depend on the type and concentration of ingredients it contains, as well as the production method used to fabricate it. To improve the quality of existing products, develop new products, and reduce production costs, it is important for food manufacturers to have a thorough understanding of the physical processes that take place during emulsion formation. This chapter discusses the physical principles of emulsion formation, the various techniques available for creating emulsions, and the factors that affect the efficiency of emulsion formation. In general, the homogenization methods used to form emulsions can be divided into two categories: high-energy and low-energy methods. High-energy methods typically use specialized equipment (homogenizers) capable of generating intense disruptive forces to intermingle and disrupt oil and water phases. On the other hand, low-energy methods rely on the spontaneous formation of small droplets (usually in surfactant, oil, and water systems) when conditions such as temperature or composition are changed in a specific manner. Most food emulsions are currently produced by highenergy methods, and so they will be the main focus of this chapter. However, low-energy methods are utilized for certain specialist applications within the food industry (e.g., production of some beverage emulsions), and so they will also be considered. 6.2 Overview of Emulsion Formation The formation of an emulsion may involve a single step or a number of consecutive steps, depending on the nature of the starting material and the method used to create it (Figure 6.1). Prior to converting separate oil and aqueous phases into an emulsion, it is often necessary to disperse the various functional ingredients into the phase in which they are most soluble. Oil-soluble ingredients, such as lipophilic vitamins, colors, antioxidants, and surfactants, are mixed with the oil phase, while water-soluble ingredients, such as hydrophilic proteins, polysaccharides, sugars, salts, buffers, vitamins, colors, antioxidants, and surfactants, are mixed with the water phase. Having said this, in industry, sometimes predominantly watersoluble ingredients are initially dispersed in the oil phase (such as certain surfactants, phospholipids, or biopolymers), since practical experience has demonstrated that more stable systems can be formed or the processing is more efficient. In some situations, it is more convenient to incorporate powdered functional ingredients directly into an oil–water mixture, regardless of the phase in which the ingredients are most soluble, since this helps to prevent clumping and facilitates dispersion during subsequent mixing and homogenization processes. Certain types of powdered ingredients are often mixed together in the powder 245 246 Food Emulsions: Principles, Practices, and Techniques Pre-homogenization Mixing Dispersing Dissolving Temperature control Homogenization Method used Operating conditions System composition 2 3 4 1 Off Post-homogenization Thermal processing Chilling or freezing Dehydration Blending Storage and transport FIGURE 6.1 The formation of a food emulsion may involve a number of different processes that may be carried out before, during, or after homogenization. form before adding to the mixing vessel to facilitate dispersion and dissolution. High-speed mixing is often required to prevent clumping of ingredients and to stop them sticking to the sides of the vessel. Certain functional ingredients used to formulate emulsions require a heat treatment after incorporation into the system, since this promotes some kind of conformational change that facilitates their dispersion and dissolution, for example, heating above a helix-to-coil transition temperature of a polysaccharide or protein. On the other hand, it is important not to overheat other ingredients, because this adversely affects their functionality: for example, heating a globular protein above its thermal denaturation temperature. It may therefore be important to be aware of any critical temperatures associated with specific ingredients used to formulate an emulsion. The intensity and duration of the mixing process depends on the time required to solvate and uniformly distribute the ingredients. Adequate solvation is important for the functionality of a number of food components: for example, the emulsifying properties of proteins are often improved by allowing them to fully hydrate prior to homogenization. If the lipid phase contains any crystalline material, it is necessary to warm it to a temperature where all the fat melts prior to homogenization, otherwise it is difficult to create a stable emulsion (and homogenizers may become clogged). On the other hand, excessive heating of thermally labile lipids may adversely affect product quality: for example, oxidation of polyunsaturated lipids such as ω-3 fatty acids or carotenoids. Most ingredient suppliers provide detailed instructions on the optimum conditions required to disperse ingredients during emulsion formation, for example, mixing, solvent, and temperature requirements. The process of converting two immiscible liquids into an emulsion is known as homogenization, and a mechanical device designed to carry out this process is called a homogenizer. High-energy homogenization can be divided into two categories, depending on the nature of the starting material used. 247 Emulsion Formation The creation of an emulsion directly from two separate liquids is known as primary homogenization, whereas the reduction in size of the droplets in an already existing emulsion is known as secondary homogenization (Figure 6.2). The creation of a particular type of food emulsion may involve the use of either of these types of homogenization, or a combination of them both. For example, the preparation of a salad dressing at home in the kitchen is usually carried out by direct homogenization of the aqueous and oil phases using a fork, whisk, or blender, and is therefore an example of primary homogenization. Conversely, homogenized milk is manufactured by reducing the size of pre-existing fat globules in raw milk, and so is an example of secondary homogenization. In many food-processing operations and research laboratories, it is more efficient to prepare an emulsion using two sequential steps. The separate oil and water phases are first converted into a coarse emulsion that contains fairly large droplets using one type of homogenizer (e.g., a high shear mixer), and then the size of the droplets is reduced using another type of homogenizer (e.g., a high-pressure valve homogenizer). Many of the same physical processes occur during primary and secondary homogenization (e.g., mixing, droplet disruption, and droplet coalescence), and so there is no clear distinction between them. Emulsions that have undergone secondary homogenization usually contain smaller droplets than those that have undergone primary homogenization, although this is not always the case. Some homogenizers are capable of producing emulsions with small droplet sizes directly from the separate oil and water phases: for example, high-intensity ultrasound, microfluidizers, or membrane homogenizers (see Section 6.5). The physical processes that occur during homogenization can be highlighted by considering the formation of an emulsion from pure oil and pure water. When the two liquids are placed in a container, they tend to adopt their thermodynamically most stable state, which consists of a layer of oil on top of a layer of water (Figure 6.2). This arrangement is adopted, because it minimizes the contact area between the two immiscible liquids, and because oil has a lower density than water (Section 7.2). To create an emulsion using high-energy methods, it is necessary to supply mechanical energy to the system to disrupt and intermingle the oil and water phases. The type of emulsion formed in the absence of an emulsifier depends primarily on the initial concentration of the two liquids: at high oil concentrations, a water-in-oil emulsion tends to be formed, but at low oil concentrations, an oil-in-water emulsion tends to be formed.* In this example, we assume that the oil concentration is so low that an oil-in-water emulsion is formed. High-energy homogenization can be achieved using a variety of different approaches (Section 6.5), with the simplest being to vigorously shake the oil and water together in a sealed container. Immediately after shaking an emulsion is formed that appears optically opaque, because the emulsion droplets scatter light (Chapter 10). The oil droplets formed during the application of the mechanical agitation are Primary homogenization Secondary homogenization Oil Premix Water 23 1 4 Off FIGURE 6.2 Homogenization can be conveniently divided into two categories: primary and secondary homogenization. Primary homogenization is the conversion of two bulk liquids into an emulsion, whereas secondary homogenization is the reduction in size of the droplets in an existing emulsion. * In the presence of an emulsifier, the type of emulsion formed is governed mainly by the properties of the emulsifier, that is, the HLB number and optimum curvature (Chapter 4). 248 Food Emulsions: Principles, Practices, and Techniques Stabilization Formation Rapidly adsorb Lower interfacial tension Facilitate breakup (a) (b) Generate repulsive forces Form resistant membrane Prevent aggregation FIGURE 6.3 An emulsifier plays two important roles in determining the overall properties of food emulsions: (a) emulsion formation and (b) emulsion stability. constantly moving around and frequently collide and coalesce with neighboring droplets. As this process continues, the large droplets formed move to the top of the container due to gravity and merge together to form a separate layer. As a consequence, the system reverts back to its initial state—a layer of oil sitting on top of a layer of water (Figure 6.2). The thermodynamic driving forces responsible for this phenomenon are the hydrophobic effect, which favors the minimization of the contact area between the oil and water phases, and gravity, which favors the upward movement of the oil (Section 7.2). To form an emulsion that is (kinetically) stable for a reasonable period of time, one must prevent the droplets from merging together after they have been formed. This is normally achieved by having a sufficiently high concentration of emulsifier present during the homogenization process. As will be discussed later, the emulsifier plays two distinctly different but important roles in determining the overall properties of food emulsions: (1) emulsion formation; (2) emulsion stability (Figure 6.3). During emulsion formation, the emulsifier must rapidly adsorb to the surfaces of the newly formed droplets, reduce the interfacial tension, and form a protective coating that prevents them from coalescing inside the homogenizer. After emulsion formation, the emulsifier must continue to protect the droplets from aggregating when they are exposed to different environmental conditions founds in foods (such as pH, ionic strength, temperature, and ingredient interactions). Knowledge of the impact of emulsifier type on emulsion formation and stability is essential for creating food products with the required quality attributes (Section 6.4). Many of the physicochemical, sensory, and nutritional properties of food emulsions depend on the size of the droplets they contain, including their stability, texture, flavor, appearance, and gastrointestinal fate (Chapters 7 through 11). One of the major objectives of homogenization is therefore to create an emulsion with a particle size distribution known to give desirable quality attributes for the specific product being manufactured. It is therefore important for food scientists to appreciate the major factors that determine the size of the droplets produced during homogenization. This brief introduction to homogenization has highlighted some of the most important aspects of emulsion formation by high-energy methods, including the necessity to mechanically process the system, the competing processes of droplet formation and droplet coalescence, and the role of the emulsifier. These topics will be considered in more detail in the rest of the chapter. 6.3 Flow Profiles in Homogenizers The rates of droplet disruption, droplet coalescence, and emulsifier adsorption within a particular homogenizer depend on the flow profile that the fluids experience (Walstra and Smulder 1988, Tesch et al. 2003, 249 Emulsion Formation Walstra 2003, Schubert and Engel 2004, Hakansson, Tragardh et al. 2009). For this reason, we begin by providing a brief outline of the major types of flow profile that emulsions can experience within homogenizers used to prepare food emulsions (Figure 6.4): • Laminar flow: At relatively low flow rates, fluid flow tends to be regular, smooth, and well defined. Laminar flow may involve simple shear, rotational, or extension flow. • Turbulent flow: At relatively high flow rates, fluid flow tends to be irregular, chaotic, and ill-defined due to the formation of eddies within the fluid. • Cavitational flow: In the presence of highly fluctuating pressure variations within a fluid, the flow profile is extremely complex because of the formation of small cavities that violently implode and generate shock waves. A more detailed explanation of these flow profiles can be found elsewhere (Walstra 1983, Walstra and Smulder 1988, Walstra 2003). In practice, the flow regime within a homogenizer is often a combination of two or more of these different flow types. The tendency for laminar or turbulent flow to occur depends on the balance of viscous (frictional) and inertial forces acting on the fluid, which is normally characterized by the Reynolds number: Re = Inertial forces LvrC = Viscous forces hC (6.1) where L is some characteristic length of the system (e.g., the diameter of a pipe or a droplet) v is the average fluid flow velocity ρ C is the density of the fluid ηC is the viscosity of the fluid When the viscous forces generated within a fluid dominate the inertial forces (low Re), the flow profile is laminar. However, when the Reynolds number in the fluid exceeds some critical value (ReCr[fluid]), the flow goes from laminar to turbulent and inertial forces dominate. For the flow of a Newtonian fluid (a) (b) (c) (d) FIGURE 6.4 Examples of some types of flow profile that may occur in fluids within mechanical homogenizers under different conditions. At relatively low Reynolds numbers: (a) rotational, (b) simple shear, and (c) extensional (or elongational) flow. At relatively high Reynolds numbers: (d) turbulent flow. 250 Food Emulsions: Principles, Practices, and Techniques through a cylindrical pipe (L = D, the diameter of the pipe), ReCr[fluid] = 2300. For the flow of a Newtonian fluid around a spherical droplet (L = d, the diameter of the droplet, v = v, the velocity of the droplet relative to the fluid), ReCr[drop] = 1. Based on the aforementioned definitions of the Reynolds numbers for fluids and droplets, it is useful to distinguish different flow regimes responsible for droplet disruption in homogenizers that utilize laminar or turbulent flow profiles, depending on the type of fluid flow they experience and the type of forces mainly responsible for droplet disruption (Walstra and Smulder 1988): • Laminar-viscous (LV) regime: The dominant flow profile in the fluid is laminar (Re[fluid] < ReCr[fluid]), and viscous forces are predominantly responsible for droplet disruption (Re[drop] < ReCr[drop]). • Turbulent-viscous (TV) regime: The dominant flow profile in the fluid is turbulent (Re[fluid] > ReCr[fluid]), and viscous forces are predominantly responsible for droplet disruption (Re[drop] < ReCr[drop]). • Turbulent-inertial (TI) regime: The dominant flow profile in the fluid is turbulent (Re[fluid] > ReCr[fluid]), and inertial forces are predominantly responsible for droplet disruption (Re[drop] > ReCr[drop]). The viscous forces acting on the droplets are due to the flow of fluid parallel to the surface of the droplets, whereas the inertial forces are due to local pressure fluctuations in the fluid and tend to act perpendicular to the surface of the droplets. The flow regime responsible for droplet disruption depends on the type of homogenizer used to create the emulsion (Table 6.1), as well as the physicochemical characteristics of the fluid (e.g., density and viscosity) (Table 6.2). TABLE 6.1 Comparison of the Attributes of Different Types of Homogenizer Used to Prepare Food Emulsions Homogenizer Type High speed mixer Colloid mill Throughput Batch or continuous Continuous Dominant Flow Regime Energy Density (J m−3) Relative Energy Efficiency Minimum Droplet Size Sample Viscosity TI, TV, LV Low–high Low 2 μm Low to medium LV (TV) Low–high 103–108 Medium–high 106–108 Medium–high 106–108 Medium–high 106–108 Medium–high 106–2 × 108 Low–medium <103–108 Intermediate 1 μm Medium to high High 0.1 μm Low to medium Low 0.1 μm Low to medium High 1 μm Low to medium High <0.1 μm Low to medium Very high 0.3 μm Low to medium High pressure homogenizer Continuous TI, TV, (CI) LVa Ultrasonic probe Batch or continuous CI Ultrasonic jet homogenizer Continuous CI Micro-fluidization Continuous TI, TV Membrane processing Batch or continuous Injection Sources: Walstra, P., Formation of emulsions, in Encyclopedia of Emulsion Technology, Becher, P., ed., Marcel Dekker, New York, 1983, vol. 4, pp. 57–128; Schubert, H. Advances in the mechanical production of food emulsions, in Engineering and Food, Jowitt, R., ed., Sheffield Academic Press, Sheffield, U.K., 1997, pp. 82–102, Walstra, P. and Smulder, P.E.A., Emulsion formation, in Modern Aspects of Emulsion Science, Binks, B.P., ed., The Royal Society of Chemistry, Cambridge, U.K., 1998, pp. 56–99. Symbols: TI, turbulent-inertial; TV, turbulent-viscous; LV, laminar-viscous; CI, cavitational. a In a high pressure homogenizer the dominant droplet disruption mechanism is highly dependent on the nature of the homogenization nozzle, and may be either turbulent or laminar elongational. 251 Emulsion Formation TABLE 6.2 Equations for Calculating the Stresses, Mean Particle Diameters, Adsorption Times, Deformation Times, and Collision Times for Droplets in Emulsions under Laminar and Turbulent Flow Conditions Regime Laminar-Viscous (LV) Re, Flow < ∼1000 > ∼2000 > ∼2000 Re, Droplet Stress acting on droplets Mean diameter (d) <1 ηCG 2gWe Cr hCG <1 (εηC)1/2 g >1 (ε2d2ρC)1/3 Collision time (τCOL) p 8Gf Adsorption time (τABS) Turbulent-Viscous (TV) (ehC ) 1 2 Turbulent-Inertial (TI) æ g3 ö çç 2 ÷÷ è e rC ø 1 5 — 1 æ d 2 rC ç 15f çè e 6pG dmCG 6pGh1C/ 2 dmC e1/ 2 Deformation time (τDEF) hD hCG hD (hC e)1/ 2 G æ rC ö mC çè de ÷ø hD (e2 d 2r)1/3 Duration of disruptive forces (τDIS) 1 G h1C/ 2 e1/ 2 1/ 3 ö ÷÷ ø 1/ 3 1 æ rg 2 ç 2 çè e3 1/ 5 ö ÷÷ ø Sources: Adapted from Walstra, P., Formation of emulsions, in Encyclopedia of Emulsion Technology, Becher, P., ed., Marcel Dekker, New York, 1983, vol. 4, pp. 57–128; Walstra, P., Chem. Eng. Sci., 48(2), 333, 1993a; Walstra, P. and Smulder, P.E.A., Emulsion formation, in Modern Aspects of Emulsion Science, Binks, B.P., ed., The Royal Society of Chemistry, Cambridge, U.K., 1998, pp. 56–99. Notes: Γ, excess surface concentration; G, shear rate, ε, power density; ρC, continuous phase density; ηC and ηD, viscosities of continuous and dispersed phases; d, droplet diameter; ϕ, dispersed phase volume fraction; γ, interfacial tension; mC, emulsifier concentration (mol m−3). 6.4 Physical Principles of Emulsion Formation The size of the droplets produced by a homogenizer depends on a balance between two opposing physical processes: droplet disruption and droplet coalescence (Figure 6.5). A better understanding of the factors that influence these two processes would help food manufacturers to select the most appropriate ingredients and homogenization conditions required to produce a particular food product. An overview of droplet disruption, droplet coalescence, and the role of the emulsifier in these processes is given in this section. The reader is referred elsewhere for more thorough discussions of the physicochemical basis of emulsion formation (Walstra and Smulder 1988, Walstra 2003, Schubert and Engel 2004, Hakansson et al. 2009). 6.4.1 Droplet Disruption The precise nature of the physical processes that occur during emulsion formation depends on the type of homogenizer used and how it is operated, since this determines the type of flow profile that the droplets experience (Table 6.2). Nevertheless, there are some common aspects of droplet disruption that apply to most types of homogenizer. The initial stages of primary homogenization involve the breakup and intermingling of the bulk oil and aqueous phases so that fairly large droplets of one of the liquids become dispersed throughout the other liquid. The later stages of primary homogenization, and the entire process of secondary homogenization, involve the disruption of larger droplets into smaller ones (Figure 6.5). It is therefore particularly important to understand the nature of the forces that are responsible for the disruption of droplets during homogenization. Whether or not a droplet breaks up is determined by a 252 Food Emulsions: Principles, Practices, and Techniques Rapid adsorption: Stabilization Droplet disruption Emulsifier properties: • Adsorption rate • Interfacial tension • Colloidal interactions Slow adsorption: Coalescence FIGURE 6.5 The size of the droplets produced during homogenization depends on the balance between the time for an emulsifier to adsorb to the surface of a droplet (τADS) and the time between droplet–droplet collisions (τCOL). balance between interfacial forces that tend to hold the droplets together and disruptive forces generated within the homogenizer that tend to pull them apart. 6.4.1.1 Interfacial Forces An emulsion droplet tends to be spherical, because this shape minimizes the thermodynamically unfavorable contact area between the oil and aqueous phases (Chapter 5). Changing the shape of a droplet, or breaking it up into a number of smaller droplets, increases this contact area and therefore requires an input of free energy. The interfacial force responsible for keeping a droplet in a spherical shape is characterized by the Laplace pressure (ΔP L), which acts across the oil–water interface toward the center of the droplet so that there is a larger pressure inside the droplet than outside of it: DPL = 4g d (6.2) where γ is the interfacial tension between oil and water d is the droplet diameter To deform and disrupt a droplet during homogenization, it is necessary to apply an external force that is significantly larger than the interfacial force. Equation 6.2 indicates that the pressure required to disrupt a droplet increases as the interfacial tension increases or as the droplet size decreases. It also indicates that intense pressure gradients must be generated within a homogenizer in order to overcome the interfacial forces holding the emulsion droplets together. For example, the Laplace pressure of a 1 μm diameter droplet with an interfacial tension of 0.01 N m−1 is about 40 kPa, which corresponds to a pressure gradient of ΔP L/d ≈ 40 × 109 Pa m−1 across the droplet. These large pressure gradients cannot be achieved using simple mixers or blenders, instead specially designed homogenizers are normally required to generate the intense mechanical stresses needed to create emulsions containing small droplets (Section 6.5). 6.4.1.2 Disruptive Forces The nature of the disruptive forces that act on a droplet during homogenization depends on the flow conditions it experiences (that is, laminar, turbulent, or cavitational), and therefore on the type of 253 Emulsion Formation homogenizer used. The dominant flow regimes prevailing in some commonly used homogenizers are summarized in Table 6.2. For a droplet to be broken up during homogenization, the disruptive forces must exceed the interfacial forces and their duration must be longer than the time required for droplet deformation. The relative magnitude of disruptive and interfacial forces is conveniently characterized by the Weber number (We = disruptive forces/interfacial forces) (Walstra 1983, Seekkuarachchi et al. 2006). Droplets tend to be disrupted when the Weber number exceeds some critical value (around unity), which depends on the physical characteristics of the oil and aqueous phases. The relative duration of disruptive forces (τDIS) and droplet deformation (τDEF) is conveniently characterized by the ratio τDIS/τDEF. Droplets tend to be disrupted when the duration of the applied disruptive forces is longer than the time required for the droplet to become deformed, that is, τDIS/τDEF > 1. Approximate expressions for the duration of the disruptive forces and the droplet deformation time under different flow regimes are summarized in Table 6.1. These expressions can be used to predict whether droplets will be disrupted in a particular flow regime. The flow profile of an emulsion within a homogenizer is usually extremely complex and is difficult to model mathematically, although recently, advances have been made using computational fluid dynamics (CFD) simulation methods (Dubbelboer et al. 2014). Without these simulation methods, it is not easy to accurately calculate the disruptive forces that a droplet experiences during homogenization. Nevertheless, it is possible to gain some useful physical insights into the major factors that affect droplet disruption by considering droplet breakup under simpler flow conditions that approximate those occurring in actual homogenizers, that is, laminar, turbulent, or cavitational flow conditions (Walstra and Smulder 1988, Walstra 2003, Sajjadi et al. 2013). Laminar flow conditions: This type of flow profile is predominant at low flow rates (i.e., low Reynolds numbers), where the fluid moves in a regular and well-defined pattern. Different types of laminar flow profile are possible, depending on the direction and velocity at which different regions within the fluid move relative to one another, for example, simple shear, rotational, and elongational flow (Figure 6.4). For convenience, we will mainly consider droplet disruption under simple shear flow conditions in order to highlight some of the most important factors that influence droplet disruption. Nevertheless, it should be stressed that simple shear flow is rarely the dominant droplet disruption mechanism in commercial homogenizers; instead, it is usually elongational or turbulent flow. In the presence of a simple shear field, a droplet experiences a combination of normal and tangential stresses. These stresses cause the droplet to rotate and become elongated, as well as causing the liquid within the droplet to circulate (Figure 6.6). At sufficiently high shear rates, the droplet becomes so elongated that it is broken up into a number of smaller droplets (Stone 1994, Windhab et al. 2005). The manner in which the droplets break up depends on the ratio of the viscosities of the droplet and continuous phase (ηD/ηC). Experiments in which droplets were photographed under different flow conditions have shown that at low values of ηD/ηC the droplets break up at their edges, at intermediate values they break up near their middle, and at high values, they may not break up at all, because there is insufficient time for the droplets to deform during the application of the disruptive forces (Williams et al. 1997, Fischer and Erni 2007). The disruptive forces that a droplet experiences during simple shear flow are determined by the shear stress (GηC) that acts upon the droplet, and so the Weber number is given by (Walstra 1993)* We = Shear forces GhC d = Interfacial forces 2g (6.3) where G is the shear rate ηC is the viscosity of the continuous phase * The reason that a factor of 4 appears in Equation 6.3, while a factor of 2 appears in Equation 6.2, is because only half of the applied shear force goes to deforming the droplet, the remainder causes the droplet to rotate and is therefore not responsible for the droplet disruption. 254 Food Emulsions: Principles, Practices, and Techniques Rotation of whole droplet Simple shear flow Circulation of fluid within droplet Droplet elongation Droplet budding Droplet disruption FIGURE 6.6 In the presence of a simple shear flow, droplets may rotate and become elongated. In addition, the fluid inside of them may circulate around the center of the droplet. For a given system, it is possible to define a critical Weber number, WeCr, which is the value of We where the droplets are just stable to disruption. If the Weber number is above this critical value (i.e., high shear rates or large droplets), then the droplets will be broken up, otherwise they will remain intact (Figure 6.7). The critical Weber number of an emulsion in the absence of emulsifier depends principally on the ratio of the viscosities of the dispersed and continuous phases (Karbstein and Schubert 1995a,b, Walstra 2003, Seekkuarachchi et al. 2006). WeCr has a minimum value when ηD/ηC is between about 0.05 and 0.5, and increases significantly as the viscosity ratio decreases below about 0.05 or increases above about 5. The behavior of droplets during the disruption process has been widely studied (Williams et al. 1997). Droplets are resistant to break up at low viscosity ratios (<0.05), because they are able to become extremely elongated before any disruption occurs. They are resistant to breakup at high viscosity ratios (>5), because they do not have sufficient time to become deformed before the flow field causes them to rotate to a new orientation and therefore alter the distribution of disruptive stresses acting on them. At intermediate viscosity ratios, the droplets tend to form a dumb-bell shape just prior to breaking up. In practice, droplet disruption under laminar flow conditions often occurs by a combination of both viscous and elongational contributions (Walstra and Smulder 1988). Under these conditions, there is reduced droplet rotation, reduced circulation of fluid within the droplet, and an increase in the effective viscosity of the fluid. Consequently, elongational flow exerts a higher stress on droplets than simple shear flow and is therefore more effective at breaking up droplets. In addition, the dependence of the critical Weber number on ηD/ηC is much less when there is a significant elongational flow component, there being a gradual decrease in WeCr from around 1 to 0.1 as ηD/ηC goes from 10 −4 to 102 (Walstra and Smulder 1988). 255 Emulsion Formation 10 Droplet disruption We (critical) 8 6 4 Droplet stability 2 0 0.001 0.01 0.1 ηd/ηc 1 10 FIGURE 6.7 Dependence of the critical Weber number on the viscosity of the dispersed and continuous phases under simple shear flow conditions. Optimum droplet disruption occurs when the viscosities of the continuous and dispersed phases are fairly similar. In the presence of emulsifiers, such as small molecule surfactants or proteins, the behavior of droplets in flow fields is different from that in the absence of emulsifiers, which has been primarily attributed to their influence on the rheology of the interfacial layer (Lucassen-Reynders and Kuijpers 1992, Williams et al. 1997, Fischer and Erni 2007). Droplets are more difficult to disrupt than would be expected from their equilibrium interfacial tension, because the emulsifier imparts rheological properties to the droplet interface that increase its resistance to tangential stresses. These rheological properties may be due to the formation of interfacial tension gradients caused by movement of surfactant molecules across the interface, that is, the Gibbs–Marangoni effect (Walstra and Smulder 1988). Alternatively, they may be due to the intrinsic dilational and shear rheology of the interfacial membrane: for example, many proteins and polysaccharides form highly viscoelastic layers (Bos and van Vliet 2001, Fischer and Erni 2007). A food manufacturer usually wants to produce an emulsion that contains a majority of droplets below some specified size, and so it is important to establish the factors determining the size of the droplets produced during homogenization. For simple shear flow, the following relationship gives a good description of the maximum size of droplets that can persist in an emulsion during homogenization under steadystate conditions (Walstra and Smulder 1988): dmax = 2gWe Cr GhC (6.4) Any droplets larger than dmax will be disrupted, whereas any smaller droplets will remain intact. This equation indicates that the size of the droplets produced during homogenization decreases as the interfacial tension γ decreases, as the shear rate increases, or as the viscosity of the continuous phase increases. It also indicates that a higher shear rate is required to decrease the droplet size when the viscosity of the continuous phase is low. It is for this reason that homogenizers that rely principally on simple shear flow conditions, such as colloid mills with smooth disks, are not suitable for generating emulsions with small droplet sizes when the continuous phase has a low viscosity (Walstra 1983). For this type of system, it is better to use homogenizers that utilize elongational, turbulent, or cavitational flow to break up the droplets: for example, high-pressure valve homogenizers, microfluidizers, or sonicators. Turbulent flow conditions: Turbulence occurs when the flow rate of a fluid exceeds some critical value, which is determined by the Reynolds number (Walstra 1993, 2003). Turbulence is characterized by rapid and chaotic fluctuations in the velocity of the fluid with time and location. The disruption of droplets under turbulent flow conditions is caused by the extremely large shear and pressure gradients associated with eddies generated in the fluid. An eddy is a region within a fluid where there is a close correlation 256 Food Emulsions: Principles, Practices, and Techniques between the fluid velocities of different elements. There is normally a range of different sized eddies formed within a liquid during turbulence. The shear and pressure gradients associated with these eddies increase as their size decreases. As a consequence, large eddies are believed to be relatively ineffective at disrupting emulsion droplets. Very small eddies are also believed to be ineffective at breaking up droplets, because they generate such high shear stresses that most of their energy is dissipated through viscous losses, rather than through droplet disruption. For these reasons, intermediate-sized eddies are thought to be mainly responsible for droplet disruption under turbulent flow conditions. When a droplet is in the vicinity of one of these intermediate-sized eddies, it is deformed and disrupted because of the large shear gradient acting across it. As mentioned earlier, the turbulent flow conditions can be divided into two categories depending on whether the droplets are primarily disrupted by viscous or inertial forces, that is, the turbulent-viscous (TV) and turbulent-inertial (TI) regimes. For turbulent conditions, the Weber number is given by (Karbstein and Schubert 1995) We = Turbulent forces Interfacial forces (6.5) Mathematical expressions have been derived for the disruptive forces acting on droplets under the TV and TI regimes; hence, it is possible to calculate the Weber number under turbulent conditions (Walstra and Smulder 1988). These equations have been used to predict the maximum size of droplets that can persist during homogenization once a steady-state has been reached: dmax = dmax g 1 (ehC ) 2 æ g3 ö =ç 2 ÷ è e rC ø (For TV) (6.6) (For TI) (6.7) 1 5 For emulsions created using a homogenizer that operates under turbulent conditions, the droplet size should decrease as the power density increases and the interfacial tension decreases, which is supported by experimental observations. In addition, the droplet size should decrease with increasing fluid viscosity under TV conditions, or decreasing fluid density under TI conditions. A number of experimental studies have shown that the viscosities of the dispersed and continuous phases influence the maximum droplet size that can persist during homogenization under turbulent (TI) conditions, there being a minimum in dmax when ηD/ηC is between about 0.05 and 5 (Seekkuarachchi et al. 2006). The dependence of the droplet size produced during homogenization on the viscosity ratio under turbulent flow conditions is therefore similar in form to that produced under laminar flow conditions (Figure 6.7). It is therefore possible to reduce the size of the droplets produced during homogenization by ensuring that the viscosity ratio falls within the optimum range for droplet breakup (0.05 < ηD/ηC < 5), which could be achieved by varying the temperature, changing the oil type, or adding solutes to the aqueous phase (such as sugars or polyols). The viscosity may also influence the droplet size if it is large enough to suppress turbulence. For example, an increase in viscosity due to the presence of thickening agents or high concentrations of droplets may be sufficient to prevent turbulent flow conditions and therefore lead to inefficient homogenization (Walstra 1993). An emulsion does not normally remain in a homogenizer long enough for steady-state conditions to be attained, and so the aforementioned equations are not strictly applicable. In practice, the size of the droplets in an emulsion decreases as the length of time they spend in the disruption zone of a homogenizer increases, until eventually a constant value is reached (Karbstein and Schubert 1995). This is because droplets take a finite time to be deformed, and so the turbulent forces must act over a period that is sufficiently longer than this time if all the droplets are to be effectively disrupted. The deformation time is proportional to the viscosity of a droplet, and therefore, the more viscous the disperse phase Emulsion Formation 257 is, the less likely is the droplet breakup within a specified time (Table 6.2). Emulsions produced under turbulent flow conditions are always polydisperse because of the distribution of eddy sizes in the fluid. In fact, the statistical theories used to derive the aforementioned equations indicate that droplets formed under turbulent conditions should follow a log-normal distribution, which is often observed in practice (Walstra and Smulder 1988). Cavitational flow conditions: Cavitation occurs in fluids that are subjected to rapid changes in pressure, and is particularly important in ultrasonic and high-pressure valve homogenizers (Freudig et al. 2003, Floury et al. 2004, Patist and Bates 2008, Hakansson et al. 2010). A fluid contracts when the pressure acting on it increases, and expands when the pressure decreases. When the instantaneous pressure that a fluid experiences falls below some critical value, a cavity is formed. As the fluid continues to expand, the cavity grows and some of the surrounding liquid evaporates and moves into it. During a subsequent compression, the cavity catastrophically collapses, generating an intense shock wave that propagates into the surrounding fluid and causes any droplets in its immediate vicinity to be deformed and disrupted. Extremely high temperatures and pressures are associated with these shock waves, but they are of very short duration and highly localized, so that limited damage is usually caused to the vessel containing the fluid. Nevertheless, over time cavitational effects cause significant damage to the surfaces of high pressure valve homogenizers and ultrasonic transducers, which become pitted. Cavitation only occurs in fluids when the intensity of the fluctuating pressure field exceeds a critical value, known as the cavitational threshold. This threshold is high in pure liquids, but is reduced when cavitational nuclei, such as gas bubbles or impurities, are present. The cavitational threshold also depends on the frequency of the pressure fluctuations, decreasing with decreasing frequency (Gopal 1968). 6.4.1.3 Role of the Emulsifier in Droplet Disruption The ease at which a droplet is disrupted during homogenization increases as the interfacial tension decreases (Equation 6.2). Thus, it should be possible to produce droplets with smaller sizes by homogenizing in the presence of an emulsifier that reduces the interfacial tension (Walstra 1983, 2003). For example, adding an emulsifier that decreases the interfacial tension from 50 to 5 mN m−1 should decrease the size of the droplets produced under laminar flow conditions 10-fold. Nevertheless, there are a number of other factors that also determine the effectiveness of emulsifiers at reducing the droplet size. Firstly, the rate at which an emulsifier adsorbs to the surface of the droplets during homogenization must be considered. Immediately after their formation, droplets have a low concentration of emulsifier adsorbed to their surface, and are therefore more difficult to disrupt because of the relatively high interfacial tension. With time a greater amount of emulsifier accumulates at the surface, which decreases the interfacial tension and therefore facilitates droplet disruption. Thus, the quicker the emulsifier adsorbs to the surface of the droplets during homogenization, the smaller are the droplets produced. Secondly, the ability of emulsifiers to enhance the interfacial rheology of emulsion droplets hampers the breakup of droplets, which leads to larger droplets sizes than those expected from the equilibrium interfacial tension (Lucassen-Reynders and Kuijpers 1992, Williams et al. 1997, Fischer and Erni 2007). These two effects partly account for the poor correlation between droplet size and equilibrium interfacial tension often reported in the literature. 6.4.1.4 Role of Nonideal Fluid Behavior on Droplet Disruption The equations given earlier are only strictly applicable to homogenization of ideal (Newtonian) liquids (Chapter 8). In practice, many liquids used in the food industry exhibit nonideal behavior, which can have a pronounced influence on the efficiency of droplet disruption and, therefore, on the size of the droplets produced during homogenization (Walstra 1993, 2003). Many biopolymers used to thicken or stabilize emulsions exhibit pronounced shear thinning behavior (Chapters 4 and 8). As a consequence, the viscosity used in the aforementioned equations should be that which the droplet experiences at the shear rates that occur during homogenization, rather than that which is measured in a viscometer at low shear rates. In addition, biopolymers may be capable of suppressing the formation of eddies, which may reduce the efficiency of homogenization carried out under turbulent flow conditions. 258 Food Emulsions: Principles, Practices, and Techniques 6.4.2 Droplet Coalescence Emulsions are highly dynamic systems in which the droplets continuously move around and frequently collide with each other, for example, due to Brownian motion, gravity, or applied mechanical forces (Chapter 7). Droplet–droplet collisions are particularly rapid during homogenization because of the intense mechanical agitation of the emulsion. If droplets are not protected by a sufficiently strong interfacial coating, they tend to coalesce with one another during a collision. Immediately after the disruption of an emulsion droplet, there is insufficient emulsifier present to completely cover the newly formed surface, and therefore, the new droplets are more likely to coalesce with their neighbors during a collision (Walstra 1993, Jafari et al. 2008, Lee et al. 2013). To prevent coalescence, it is necessary to form a sufficiently concentrated emulsifier coating around the droplets before they have time to collide with their neighbors (Figure 6.3). The size of the droplets produced during homogenization therefore depends on the time taken for the emulsifier to be adsorbed to the surface of the droplets (τADS) relative to the time between droplet–droplet collisions (τCOL). These times depend on the flow profile that the droplets experience, as well as the nature of the emulsifier used. Estimates of the adsorption and collision times have been established for laminar and turbulent flow conditions (Table 6.2). The equations given in Table 6.2 are only strictly applicable to dilute emulsions, even so, they do give some useful insights into the factors that influence the droplet size produced by homogenization. Ideally, a food manufacturer wants to minimize droplet coalescence during homogenization by ensuring that the emulsifier adsorption time is much shorter than the droplet collision time (τADS/τCOL << 1). The following expression has been shown to be applicable for both laminar and turbulent flow regimes (Walstra 2003): tADS 6pGf » dmc tCOL (6.8) Thus, coalescence within a homogenizer should decrease as the disperse phase volume fraction (φ) decreases, the diameter of the droplets (d) increases, the excess surface concentration (Γ) decreases, and the concentration of emulsifier (mc) increases. The effect of emulsifier adsorption kinetics on the size of the droplets produced during homogenization has been demonstrated experimentally (Schubert 1997, Baret, Kleinschmidt et al. 2009). Under fixed homogenization conditions, it was shown that emulsifiers that adsorb more rapidly produce smaller droplet sizes than those that adsorb more slowly. Most food emulsifiers do not adsorb quickly enough to completely prevent droplet coalescence, and so the droplet size achieved during homogenization is greater than that which is theoretically possible (Stang et al. 1994, Jafari et al. 2008). In addition to the factors already mentioned, the tendency for droplets to coalesce during (or shortly after) homogenization depends on the effectiveness of the interfacial coating to resist coalescence during a droplet–droplet encounter. The resistance of an interfacial layer to coalescence depends on the type and concentration of emulsifier molecules present, as well as the structural and physicochemical properties of the interface, for example, thickness, electrical charge, packing, and interactions (Chapters 4 and 7). Information about the rate of coalescence within a homogenizer can be obtained using a variety of methods. A number of methods that have been proved particularly useful are briefly discussed in the following. In the first method, an emulsion is recirculated through a homogenizer at a fixed operating pressure until a constant droplet size distribution is obtained where droplet coalescence is balanced by droplet disruption (Narsimhan and Goel 2001). The operating pressure is then reduced to a new value and the change in droplet size distribution with homogenization time (or number of passes) is monitored. The droplet size increases with homogenization time due to droplet coalescence until a new (larger) steady-state value is reached where droplet coalescence is again balanced by droplet disruption. The faster the coalescence rate within the homogenizer, the faster is the increase in droplet size with homogenization time. In the second method, an oil-in-water emulsion is prepared that contains a hydrophobic fluorescent probe in a fraction of the droplets, but no probe in the remainder of the droplets (Lobo et al. 2002). This “mixed emulsion” is prepared by combining an emulsion that contains no probe with an emulsion that contains some probe, with all other aspects of the emulsions being similar, for example, droplet size Emulsion Formation 259 distribution, oil type, and emulsifier type. The fluorescent probe is chosen so that the fluorescent emission spectrum depends on the local probe concentration in the oil phase, for example, due to formation of dimers that have a different spectrum than monomers. Coalescence during homogenization causes the dispersed phases of different droplets to be mixed, thereby reducing the local probe concentration and changing the fluorescence spectrum. Measurement of the change in fluorescence spectrum with time provides information about the coalescence rate. In the third method, a “mixed emulsion” is prepared by combining two oil-in-water emulsions that contain different colored oil-soluble dyes within the droplets (e.g., blue and yellow), but are similar in other aspects (Schubert et al. 2003). The mixed emulsion is then homogenized by recirculating it through a homogenizer at a fixed operating pressure and the change in droplet composition with homogenization time (or number of passes) is monitored by measuring the change in the color of the droplets using optical microscopy. As homogenization proceeds, the droplets coalesce with each other and their contents are mixed, which results in a decrease in the fraction of droplets with the initial dye colors (yellow and blue) and an increase in the fraction of droplets of mixed color (green). In the fourth method, a “mixed emulsion” is prepared by combining two oil-in-water emulsions that contain different oil types (e.g., hexadecane and octadecane), but are otherwise similar (Hakansson et al. 2012). The mixed emulsion is then homogenized by recirculating it through a homogenizer at a fixed operating pressure and the change in droplet composition due to coalescence is monitored with homogenization time (or number of passes). The change in droplet composition with time is determined by measuring some physical property of the system that depends on oil composition, for example, refractive index (Taisne et al. 1996) or melting point (Elwell et al. 2004). In the fifth method, no probes or mixed oils are required to determine the coalescence (and fragmentation) rates of droplets during homogenization (Hakansson and Hounslow 2013). In this case, the coalescence rate is inferred by using a population balance model to interpret changes in the particle size distribution under different homogenization conditions (pressure). The main advantage of this method is that it does not require the use of external probes or particular mixtures of oils. However, the main disadvantage is that it requires sophisticated mathematical analysis and computation skills that may not be available to many research groups. The aforementioned methods can be used to quantify the factors that influence droplet coalescence within a homogenizer. For protein-stabilized emulsions in a high-pressure valve homogenizer, the coalescence rate was found to be higher for higher homogenization pressures, larger drop sizes, higher dispersed phase volume fractions, pH values closer to the isoelectric point, and high ionic strengths (Mohan and Narsimhan 1997). For emulsions stabilized by ionic surfactants prepared using a high-pressure valve homogenizer, the coalescence rate was found to be higher for higher homogenization pressures, for lower surfactant concentrations, and to depend on dispersed phase fraction and ionic strength (Taisne et al. 1996, Narsimhan and Goel 2001, Lobo and Svereika 2003). Coalescence rates of oil-in-water emulsions stabilized by different kinds of emulsifier have also been compared: for example in high-pressure valve homogenizers, the coalescence rate was higher for whey protein than for sodium caseinate–stabilized droplets (Mohan and Narsimhan 1997), was higher for egg yolk than for Tween-stabilized droplets (Schubert et al. 2003), and was higher for Tween 20– than for caseinate-stabilized droplets (Elwell et al. 2004). 6.4.3 Role of the Emulsifier The discussion so far has highlighted two of the most important functions of emulsifiers during the homogenization process: 1. They decrease the interfacial tension between the oil and water phases, thereby reducing the amount of free energy required to deform and disrupt the droplets. 2. They form a protective coating around the droplets that prevents them from coalescing with each other. The size of the droplets produced during homogenization therefore depends on a number of different characteristics of an emulsifier: (1) the ratio of emulsifier to dispersed phase—there must be sufficient 260 Food Emulsions: Principles, Practices, and Techniques emulsifier present to completely cover the surfaces of all the droplets formed, (2) the time required for the emulsifier to move from the bulk phase to the droplet surface—the faster the adsorption time, the smaller is the droplet size, (3) the probability that an emulsifier molecule will be adsorbed to the surface of a droplet during an encounter between it and the droplet—the greater the adsorption efficiency, the smaller the droplet size, (4) the amount that the emulsifier reduces the interfacial tension—the greater the amount, the smaller is the droplet size, (5) the extent to which the emulsifier alters the dynamic interfacial rheology—the greater the resistance to deformation, the more difficult is droplet disruption (larger size) and the less likely is droplet coalescence (smaller size), and (6) the effectiveness of the emulsifier layer at protecting the droplets against coalescence—the better the protection, the smaller is the droplet size. A more detailed discussion of the influence of emulsifier characteristics on the size of the droplets produced in homogenizers is given in Section 6.6.1. 6.5 Homogenization Devices Numerous types of homogenization device have been developed to produce food emulsions using highenergy methods. Each of these devices has its own particular advantages and disadvantages, and range of materials where it is most suitably applied. The choice of a particular homogenizer depends on whether the emulsion is being prepared in a factory or in a laboratory, the equipment available, the volume of material to be homogenized, the desired throughput, the nature of the starting materials, the desired droplet size distribution, the required physicochemical properties of the final product, and the cost of purchasing and running the equipment. The most important types of homogenizer used in the food industry or by food scientists working in research laboratories are discussed in the following text. The general characteristics of these different homogenizers are compared in Table 6.1. 6.5.1 High Shear Mixers High shear mixers (which are a type of rotor-stator device) are the most commonly used instruments for directly homogenizing oil and aqueous phases in the food industry (Urban et al. 2006, Hall et al. 2011, Singh and Heldman 2013). In a batch process, the oil, water, and other ingredients to be homogenized are placed in a suitable vessel (Figure 6.8), which may contain as small as a few cm3 (for laboratory use) or as large as several m3 (for industrial use) of liquid. The components are then agitated by a mixing head that rotates at high speeds (typically up to 3600 rev min−1). The various ingredients may all be added at the beginning of the process, or they may be added sequentially to improve dispersion and/or reduce homogenization times. The rapid rotation of the mixing head generates a combination of longitudinal, rotational, and radial velocity gradients in the fluids, which disrupts the interfaces between the oil and water phases, causes the liquids to become intermingled, and breaks the larger droplets into smaller ones. Efficient homogenization is achieved when the horizontal and vertical flow profiles distribute the liquids evenly throughout the vessel, which can be facilitated by having baffles fixed to the inside walls of the vessel. The design of the mixing head also determines the efficiency of the homogenization process, and a number of different types are available for different situations, for example, blades, propellers, and turbines. Specially designed mixing heads are often used to generate more intense and evenly distributed disruptive forces, so as to create smaller droplets, reduce homogenization times, and/or ensure more uniform mixing. The homogenization efficiency can also be increased by using countercurrent mixing devices, where the container rotates in one direction and the mixing head rotates in another. For industrial purposes, many high shear mixing devices are available that are capable of in-line (rather than batch) operation so that products can be produced continuously. Typical throughputs for in-line devices range from a few liters to a few hundred liters per hour. Blending generally leads to an increase in the temperature of an emulsion, because some of the mechanical energy is converted into heat due to viscous dissipation. If any of the ingredients in the emulsion are sensitive to heat, it may be necessary to control the temperature of the vessel during homogenization. High shear mixers are particularly useful for preparing emulsions with low or intermediate viscosities (Table 6.1). The droplet size usually decreases as the homogenization time or 261 Emulsion Formation 2 3 4 1 Off High-shear mixer Rotating shaft Typical flow pattern FIGURE 6.8 High-speed mixers are often used in the food industry to directly homogenize oil and aqueous phases. the rotation speed of the mixing head is increased, until a lower limit is achieved, which depends on the nature and concentration of the ingredients used and the power density of the mixer. Typically, the droplets produced by a high shear mixer range between about 1 and 10 μm in diameter. Industrial mixers are often designed to avoid excessive incorporation of air bubbles during homogenization, because this can have an adverse effect on the subsequent processing and physicochemical properties of many emulsion-based food products. As well as being used to create coarse food emulsions, high shear mixers are commonly used to ensure effective dispersion and dissolution of ingredients, particularly powdered ingredients. 6.5.2 Colloid Mills Colloid mills are widely used in the food industry to homogenize medium and high viscosity liquids (Walstra 1983, Schubert 1997, Urban et al. 2006). A variety of different designs of colloid mill are commercially available, but they all operate on fairly similar physical principles (Figure 6.9). A colloid mill usually contains two disks: a rotor (a rotating disk) and a stator (a static disk). The liquids to be homogenized are usually fed into the center of the colloid mill in the form of a coarse emulsion, rather than as separate oil and aqueous phases, because the device is much more efficient at reducing the size of the droplets in a pre-existing emulsion (secondary homogenization), than at homogenizing two separate phases (primary homogenization). The coarse emulsion is usually prepared directly from the oil, water, and other ingredients using a high shear mixer. The rapid rotation of the rotor within a colloid mill generates a shear stress in the gap that causes the larger droplets to be broken down into smaller ones, and generates a centrifugal force that causes the fluid to move from the center to the periphery of the disks where it is either collected or passed through a pipe to another unit-operation. Colloid mills are available as batch versions or as in-line versions that allow continuous emulsion production. The intensity of the shear stresses (and therefore droplet disruption forces) can be altered by varying the thickness of the gap between the rotor and stator (typically from about 50 to 1,000 μm), varying the rotation speed (typically from about 1,000 to 20,000 rev min−1), or by using disks that have roughened surfaces or interlocking 262 Food Emulsions: Principles, Practices, and Techniques Inlet Rotor Stator Outlet FIGURE 6.9 Colloid mills are mainly used in the food industry to homogenize intermediate and high viscosity materials. teeth. When the surfaces of the rotor and stator are smooth, the dominant droplet disruption mechanism is laminar shear flow, but when the surfaces are roughened or toothed, it is turbulence. Droplet disruption can also be enhanced by increasing the length of time the emulsion spends in the colloid mill, either by decreasing the flow rate or by passing the emulsion through the device a number of times. Typically, the flow rate can be varied between about 4 and 20,000 L h−1, depending on the mechanical device used and the operating conditions. It should be noted that many of the factors that increase the effectiveness of droplet disruption also increase the manufacturing costs (by increasing energy costs or reducing product throughput). Food manufacturers must therefore select the rotation speed, gap thickness, rotor/stator type, and throughput that give the best compromise between droplet size and manufacturing costs. It is usually necessary to have some form of cooling device as part of a colloid mill to offset the increase in temperature caused by viscous dissipation losses. Colloid mills are more suitable for homogenizing intermediate and high viscosity fluids (such as peanut butter, fish, or meat pastes) than high-pressure valve or ultrasonic homogenizers (Table 6.2). Typically, they can be used to produce emulsions with droplet diameters around about 1 and 5 μm. 6.5.3 High-Pressure Valve Homogenizers High-pressure valve homogenizers are probably the most common instrument used in producing fine emulsions in the food industry (Schultz et al. 2004, Santana et al. 2013). Like colloid mills, they are more effective at reducing the size of the droplets in a pre-existing emulsion, than at creating an emulsion directly from two separate liquids. A coarse emulsion is usually produced using a high shear mixer and then fed directly into the input of the high-pressure valve homogenizer (Figure 6.10). The homogenizer has a pump that pulls the coarse emulsion into a chamber on its backstroke and then forces it through a narrow valve at the end of the chamber on its forward stroke. As the coarse emulsion passes through the valve, it experiences a combination of intense disruptive forces that cause the larger droplets to be broken down to smaller ones. The actual flow regime that is responsible for disrupting the droplets in a particular high pressure valve homogenizer depends on the characteristics of the material being homogenized (e.g., viscosity), the size of the homogenizer (e.g., bench-top or 263 Emulsion Formation Coarse emulsion Inlet Impact ring Fine emulsion Outlet Piston Adjustable value Value seat FIGURE 6.10 High-pressure valve homogenizers are used to produce emulsions with fine droplet sizes. A coarse emulsion is introduced into the inlet, and then forced through a narrow gap under pressure using a piston, thereby causing droplet breakup. The homogenization pressure can be varied by adjusting the gap size using an adjustable valve. production), and the design of the homogenization nozzle (Walstra and Smulder 1988, Stang et al. 2001). Different types of nozzle have been designed to increase the efficiency of droplet disruption for different kinds of applications (Schubert et al. 2003, Schultz et al. 2004, Donsi et al. 2012). The dominant droplet disruption mechanism tends to be inertial forces in turbulent flow for standard nozzle and microfluidizer nozzles, and shear forces in laminar elongational flow for jet disperser and orifice valves. Acoustic and optical imaging methods have recently been used to provide information about the different flow profiles inside model homogenizer valves so as to establish the relative importance of cavitation and turbulence in droplet breakup (Hakansson et al. 2010, 2011). In addition, mathematical modeling approaches, such as population balance and CFD models, have provided valuable insights into the different flow profiles in homogenizer valves (Hakansson et al. 2012, 2013, Dubbelboer et al. 2014). Most commercial homogenizers use spring-loaded valves so that the gap through which the emulsion passes can be varied (typically between about 15 and 300 μm). Decreasing the gap size increases the pressure drop across the valve, which causes a greater degree of droplet disruption and smaller droplets to be produced. On the other hand, decreasing the gap size increases the energy input required to form an emulsion, thereby increasing manufacturing costs. The throughputs of industrial homogenizers typically vary between about 100 and 20,000 L h−1, while homogenization pressures vary between about 3 and 20 MPa. Experiments have shown that there is an approximately linear relationship between the logarithm of the homogenization pressure (P) and the logarithm of the droplet diameter (d), that is, log d ∝ log P (Walstra and Smulder 1988, Stang et al. 2001). The constant of proportionality depends on the dominant flow regime inside the homogenizer, which in turn depends on the dimensions of the homogenizer, the dominant droplet disruption mechanism, and the fluid viscosity. For a large homogenizer and a low fluid viscosity, the flow regime is predominantly turbulent-inertial (TI) and d ∝ P−0.6. For a large homogenizer and a high fluid viscosity, the flow regime is predominantly turbulent-viscous (TV) and d ∝ P−0.75. For a small homogenizer, like those used in many laboratory studies, the flow regime may even be laminarviscous (LV) and d ∝ P−1.0. In addition, the efficiency of droplet disruption depends on the dimensions of the homogenizer because of differences in the time that the emulsion spends in the region where droplet 264 Food Emulsions: Principles, Practices, and Techniques disruption actually occurs. In large homogenizers, the droplets may experience many more disruption events than in a small homogenizer operating at the same pressure, and therefore, droplet disruption is more efficient. These differences in the dominant flow profiles present within homogenizers of different dimensions have to be taken into account when one wants to scale up from laboratory experiments to actual production (Walstra and Smulder 1988). Some commercial devices use a “two-stage” homogenization process, in which the emulsion is forced through two consecutive valves. In the dairy industry, the first valve is often set at high pressure and is responsible for breaking up the droplets, while the second valve is set at a lower pressure and is mainly responsible for disrupting any flocs that are formed during the first stage. High-pressure valve homogenizers can be used to produce a wide variety of different food products, although they are most suitable for low and intermediate viscosity materials, particularly when a small droplet size is required. If the oil and aqueous phases have been blended prior to homogenization, it is often possible to create an emulsion with submicron particles using a single pass through the homogenizer, although microfluidizers are usually more efficient for this purpose (Lee and Norton 2013). If very fine emulsion droplets are required using a high-pressure valve homogenizer, then it is usually necessary to pass the emulsion through the device a number of times (Figure 6.11). Emulsion droplets with diameters as low as 0.1 μm can be produced using this method, provided there is sufficient emulsifier present to completely cover the oil–water interface formed, the emulsifier adsorbs rapidly enough to prevent droplet coalescence inside the homogenization, and the viscosity ratio and interfacial tension are within an appropriate range. The temperature rise in a high-pressure valve homogenizer is often fairly small, but it can become appreciable if the emulsion is recirculated a number of times or if high homogenization pressures are used. In these cases, it may be necessary to keep the emulsion cool by using a water-jacketed homogenization chamber. Conversely, it may be necessary to keep the homogenizer warm during the homogenization process for certain applications, for example, to prevent crystallization of a lipid phase (Schubert et al. 2003). It is always important to ensure that a high-pressure valve homogenizer is clean and operating correctly prior to utilization to avoid damaging it. 3 Microfluidizer HPVH Mean diameter (μm) 2.5 2 1.5 6000 psi 1 0.5 0 0 1 2 Number of passes 3 4 FIGURE 6.11 Comparison of the influence of number of passes on mean particle size produced in whey-protein-stabilized oil-in-water emulsions using a laboratory-scale high-pressure valve homogenizer and microfluidizer operating at a similar pressure (6000 psi). 265 Emulsion Formation 6.5.4 Microfluidization Microfluidization is a highly efficient method of creating emulsions with very fine droplets (Jafari et al. 2007, Qian and McClements 2011). This type of homogenizer usually consists of a fluid inlet, some kind of pumping device, and an interaction chamber containing two channels through which the fluids are made to flow and interact with each other (Figure 6.12). Fluids are introduced into the homogenizer, accelerated to a high velocity within the channels using a pumping device, and then made to simultaneously impinge with each other on a solid surface. Intense disruptive forces are generated when the two fluid streams collide with each other, which cause the fluids to intermingle and any larger droplets to be disrupted. Microfluidizer devices are available with either single or double inlets. Single-inlet microfluidizers are used to reduce the size of the droplets in a pre-existing emulsion by causing different portions of a coarse emulsion to flow through the two different channels. Double-inlet microfluidizers can create emulsions directly from the individual oil and aqueous phases by having different inlets for the two phases to be homogenized (Panagiotou and Fisher 2008, Panagiotou et al. 2008, 2009). The size of the droplets produced by a microfluidizer can be decreased by increasing homogenization pressure, number of passes, and emulsifier concentration (Wooster et al. 2008, Qian and McClements 2011). However, microfluidizers typically produce very small droplets after a single pass, and therefore, passing emulsions through multiple times is not usually energy efficient (Figure 6.11). The droplet size also depends on the oil–water interfacial tension, the viscosity ratio of the oil and water phases, the adsorption kinetics of the emulsifier, and the ability of the emulsifier to prevent coalescence within the homogenizer (Jafari et al. 2008, Iqbal et al. 2013). A variety of different channel types have been designed to increase the efficiency of droplet disruption within the homogenization zone, including straight or zigzag channels. Microfluidizers are available in bench top, pilot plant, and production versions. The minimum volumes that can be produced using specially designed laboratory-scale microfluidizers can be as small as a few milliliters, which is convenient for preparation of research-grade emulsions using ingredients that are costly or scarce. Microfluidizers are considered to be the most efficient high-energy method for producing emulsions containing very small droplets (e.g., d < 100 nm). Typically, microfluidizers are most suitable for homogenizing low and intermediate viscosity fluids. Input Single inlet Output Water input Double inlet Output Oil input FIGURE 6.12 In a microfluidizer, the fluids are brought together at a high velocity that causes intermingling and droplet disruption. Single-inlet microfluidizers can be used to reduce the droplet size in an existing emulsion, whereas double-inlet versions can produce an emulsion from separate oil and water phases. 266 Food Emulsions: Principles, Practices, and Techniques Operating pressures up to 275 MPa and throughputs up to 12,000 L h−1 can be achieved using industrial microfluidizers. Similar maximum operating pressures can be achieved using laboratory models, which accounts for their ability to produce very small droplets, that is, <0.1 μm. 6.5.5 Ultrasonic Homogenizers Ultrasonic homogenizers are widely used to produce emulsions and nanoemulsions, particularly in research laboratories (Canselier et al. 2002, Kentish et al. 2008, Leong et al. 2009, Abbas et al. 2013). This type of homogenizer utilizes high-intensity ultrasonic waves that generate intense shear and pressure gradients within a material that disrupt droplets mainly through cavitation and turbulent effects (Leong et al. 2011). A number of methods are available for generating high-intensity ultrasonic waves, but only two are commonly used commercially: piezoelectric transducers and liquid jet generators (Canselier et al. 2002). The principles behind ultrasonic homogenization methods and the main factors that affect the particle size have been reviewed in detail elsewhere (Leong et al. 2009, Delmas et al. 2011, Abbas et al. 2013). Piezoelectric transducers are used in the bench-top ultrasonic probe homogenizers that are found in many research laboratories (Figure 6.13). These homogenizers are ideally suitable for rapidly preparing small volumes of emulsion (a few cm3 to a few 100 cm3), which is often an important consideration in research laboratories, because the reagents may be limited or expensive. An ultrasonic probe (a) Input Output (b) FIGURE 6.13 (a) Ultrasonic probe and (b) ultrasonic jet homogenizers may be used for the production of emulsions. These devices work principally by generating intensity disruptive forces through a cavitation mechanism, and come in batch (probe homogenizer) and continuous (jet homogenizer) versions. Emulsion Formation 267 homogenizer contains a piezoelectric crystal mounted within a protective metal casing, which is usually tapered at the end to increase the intensity of the waves generated. A high-intensity electrical wave is applied to the probe, which causes the piezoelectric crystal inside it to rapidly oscillate and generate an ultrasonic wave. The ultrasonic wave is directed toward the tip of the transducer where it radiates into the surrounding liquids and generates intense pressure and shear gradients (mainly due to cavitational effects) that cause the liquids to be broken up into smaller fragments and intermingled with one another. Ultrasonic energy is focused on a small volume of the sample near the tip of the ultrasonic transducer: it is important to have a good agitation in the sample container. In small vessels, this is achieved by the fluid flow induced by the ultrasonic field itself, but in large vessels, it is often necessary to have additional agitation to ensure effective mixing and homogenization. To create a stable emulsion, it is usually necessary to irradiate a sample with ultrasound for periods ranging from a few seconds to a few minutes. Continuous application of ultrasound to a sample can cause appreciable heating, and so it is often advantageous to apply the ultrasound in a number of short bursts. In addition, a cooling coil may be used to reduce excessive heating of the sample. Traditionally, ultrasonic probe homogenizers were used for batch preparation of emulsions, but in-line flow-through versions have also been developed (Canselier et al. 2002, Kentish et al. 2008). Prolonged exposure of certain food components to high-intensity ultrasound may promote their physical or chemical degradation, for example, oxidation of lipids, depolymerization of polysaccharides, or denaturation of proteins (Pingret et al. 2013). It is therefore important to carefully consider the potential effects of ultrasonic treatment on these sensitive components. Ultrasonic jet homogenizers are mainly used for the industrial preparation of food emulsions (Figure 6.13). A stream of fluid is made to impinge on a sharp edged blade that causes the blade to rapidly vibrate, thus generating an intense ultrasonic field that breaks up any droplets in its immediate vicinity due to a combination of cavitation, shear, and turbulence (Canselier et al. 2002). The major advantages of this device are that it can be used for the continuous production of emulsions, it can generate very small droplets, and it is usually more energy-efficient than high-pressure valve homogenizers, that is, less energy is required to produce droplets of the same size (Schubert et al. 2003). Even so, the vibrating blade is prone to erosion because of the high-intensity ultrasonic field: it has to be replaced frequently. Fluid flow rates between 1 and 500,000 L h−1 are possible using this technique. The principle factors determining the efficiency of ultrasonic homogenizers are the intensity, duration, and frequency of the ultrasonic waves (Abbas et al. 2013, Tang et al. 2013). Below a frequency of about 18 kHz, ultrasonic waves become audible and are therefore objectionable to users. In principle, emulsions can be formed using ultrasonic waves with frequencies up to about 5 MHz, but the efficiency of homogenization decreases with increasing ultrasonic frequency. For these reasons, most commercial devices use ultrasonic waves with frequencies between about 20 and 50 kHz. The size of the droplets produced during homogenization can be decreased by increasing the intensity or duration of the ultrasonic radiation. At a fixed intensity, the size of the droplets produced usually decreases with increasing sonication time until a fairly constant value is reached. At a fixed sonication time, the droplet size tends to decrease with increasing intensity until a constant value is reached. Some of the other factors that influence the size of the droplets produced are: the design of the device used (e.g., batch versus continuous); dissolved gas concentration; hydrostatic pressure; temperature; oil–water ratio; surfactant type and concentration; and oil and aqueous phase viscosities (Abbas et al. 2013). All of these parameters must be carefully optimized to produce emulsions with the desired particle size distributions. 6.5.6 Membrane and Microchannel Homogenizers At present membrane and microchannel homogenizers are mainly utilized within research laboratories to produce particles with well-defined dimensions and structures (Vladisavljevic and Williams 2005, Nisisako 2008, Nazir et al. 2010). They are not commonly used for industrial purposes due to their relatively low volume throughput. Membrane homogenizers can be used in two main ways to process emulsions. Direct homogenization involves forming an emulsion directly from the separate oil and water phases in the presence of a 268 Food Emulsions: Principles, Practices, and Techniques suitable emulsifier. Premix homogenization involves reducing the size of the droplets present within an existing coarse emulsion. In direct homogenization, an emulsion is formed when one immiscible liquid (the disperse phase) is forced into another immiscible liquid (the continuous phase) through a solid membrane that contains small pores (Figure 6.14). The continuous phase usually contains an emulsifier that adsorbs to the droplet surfaces as they are formed, thereby stabilizing them against aggregation. The droplet size attained depends on the membrane pore size, the oil–water interfacial tension, the applied pressure, the flow profile of the continuous phase, and the type and amount of emulsifier used. Membranes can be manufactured with different pore sizes so that emulsions with different droplet sizes can be produced. Some membranes contain pores with well-defined sizes and shapes, whereas others have less well-defined ones, which has an impact on the polydispersity of the emulsion droplets produced. The membrane polarity should also be carefully selected as it determines the type of emulsions that can be created: hydrophobic membranes are needed to produce water-in-oil emulsions, whereas hydrophilic membranes are needed to produce oil-in-water emulsions. Membrane homogenization can be carried out as either a batch or continuous process. In the batch process, droplets are formed by forcing the dispersed phase through a membrane in contact with the continuous phase (Figure 6.14). In the continuous process, the dispersed phase is made to flow through a cylindrical inner membrane, while the continuous phase is made to flow through an external outer cylinder. The dispersed phase is pressurized so that it is forced through the inner membrane, where it forms small droplets in the continuous phase flowing through the outer membrane. Membrane homogenizers can also be used to reduce the size of droplets in pre-existing coarse emulsions. In this approach, a coarse emulsion is formed first (e.g., by simple stirring) and is then forced through the membrane to reduce the droplet size. The major advantages of this approach are that higher fluxes can be achieved and emulsions with higher droplet concentrations can be produced more easily. Pressure Oil Membrane Water + surfactant Batch membrane homogenizer Flow Water + surfactant Pressure Oil Continuous membrane homogenizer FIGURE 6.14 Schematic representation of batch and continuous versions of membrane homogenizers for producing oil-in-water emulsions. The disperse phase is forced into the continuous phase through small pores in a solid membrane. 269 Emulsion Formation Microchannel homogenizers work on a somewhat similar principle to membrane homogenizers, except that the disperse phase is forced through microchannels with well-defined geometries to form droplets (Neethirajan et al. 2011, Vladisavljevic et al. 2012). This technique is particularly useful for fabricating emulsion droplets with very narrow particle size distributions, and with well-defined structures, such as core-shell or dispersion structures. For example, they have been used to produce fairly monodisperse oil-in-water and water-in-oil emulsions, and to produce multiple emulsions (O/W/O or W/O/W) with controllable numbers of internal droplets. These more elaborate droplet structures are usually produced using coaxial (concentric) microchannels, and by controlling the relative flow rates of the different fluids involved. These kinds of structured droplets are suitable for highly specialized commercial applications or for research purposes, but are typically unsuitable for large-scale commercial applications in the food industry, because they are too expensive and time-consuming. Emulsions with mean droplet diameters as low as 300 nm can be produced using these types of homogenizers, but typically, they are more suitable for producing larger droplets (1–100 μm). Membrane homogenizers can be purchased commercially from a number of companies, which has increased their utilization as a research tool. On the other hand, microchannel homogenizers still have to be prepared using specialized fabrication methods. 6.5.7 Homogenization Efficiency The energy efficiency of a homogenizer (εH) can be calculated by comparing the minimum amount of energy theoretically required to form an emulsion (Emin), with the actual amount of energy that is expended during homogenization (EV) (Santana et al. 2013): eH = Emin ´100 EV (6.9) The minimum amount of energy required to form an emulsion is equal to the free energy needed to increase the interfacial area between the oil and water phases: Emin = ΔAγ, where ΔA is the increase in interfacial area (per unit volume of emulsion) and γ is the interfacial tension. For a typical oil-in-water emulsion, Emin has a value of about 3 kJ m−3, assuming that φ = 0.1, r = 1 μm, and γ = 10 mN m−1 (Walstra 1983). The actual amount of energy required to form an emulsion depends on the type of homogenizer used and the operating conditions. For a high-pressure valve homogenizer, EV is typically about 10,000 kJ m−3, and so the homogenization efficiency is less than 0.1%. The reason that homogenization is such an inefficient process is because the disruption of small droplets requires the generation of extremely high-pressure gradients. These pressure gradients must be large enough to overcome the Laplace pressure gradient (≈2γ/r 2), which is about 2 × 1010 Pa m−1 for a 1 μm emulsion droplet. The pressure gradient due to shear forces acting across a droplet is given by Gηc/r, which indicates that the shear rate must exceed about 2 × 107 s−1 in order to disrupt the droplets. The movement of liquids at such high shear rates leads to large amounts of energy dissipation because of frictional losses. The conversion of mechanical work into heat also accounts for the increase in temperature observed during homogenization. The total amount of energy supplied to an emulsion during the homogenization process is often referred to as the energy density, which has been defined as the energy input per unit volume of emulsion, or the power input per unit volume flow rate of emulsion (Schubert et al. 2003). In general, the energy density is given by the following relation (Walstra and Smulder 1988): ò EV = PV (t ) dt (6.10) where PV is the net power density t is the duration of the homogenization procedure Droplet disruption only occurs when P V exceeds some critical value, which depends on the Laplace pressure of the droplets. If the net power density is below this in any region within the homogenizer, 270 Food Emulsions: Principles, Practices, and Techniques then droplet disruption will not occur and the energy is wasted. The values of EV or P V used in the aforementioned equation depend on the type of homogenizer used to create the emulsion (Walstra and Smulder 1988, Canselier et al. 2002, Schubert et al. 2003). For most of the major types of homogenizer used in the food industry, theoretical or semiempirical relationships are available to calculate the energy density. For example, in a high-pressure valve homogenizer, the energy density is equal to the operating pressure: EV = P. Alternatively, an estimation of the net energy consumption of a homogenizer can be determined by measuring the increase in temperature of the emulsion during homogenization (since >99.9% of the energy is lost as heat) or by measuring the electrical power requirements of the homogenizer. A number of approaches can be used to improve the energy efficiency of homogenizers (Walstra and Smulder 1988): • Homogenization efficiency is normally improved by increasing the net power intensity and decreasing the duration of the emulsification process. This can be achieved by increasing the region within the homogenizer where efficient droplet disruption occurs relative to that where droplet disruption is inefficient. • Homogenization efficiency can often be improved by increasing the disperse phase volume fraction of the emulsion, then diluting it to the desired droplet concentration afterward. This approach is effective, because most of the energy lost during homogenization is due to friction, which is proportional to the overall emulsion volume. Experiments have shown that the same droplet diameter can often be achieved by a homogenizer if the emulsifier-to-disperse phase ratio and homogenization conditions are kept constant. Hence, increasing the disperse phase volume fraction reduces the overall frictional losses required to produce the same droplet size. • Homogenization efficiency can be improved by increasing the emulsifier concentration, since this leads to a greater decrease in interfacial tension and faster emulsifier adsorption, thereby facilitating droplet disruption and retarding droplet coalescence. • Homogenization efficiency can often be improved by using a combination of different homogenizers, or by using the same homogenizer at increasing intensities. This approach is based on the principle that different kinds of flow regime are more or less efficient at disrupting different sized droplets. For example, preparing a coarse emulsion using a high-speed blender and then reducing the droplet size using a high-pressure valve homogenizer is more energy-efficient than using either technique alone. 6.5.8 Comparison of Homogenizers The choice of a homogenizer for a particular application depends on a number of factors, including the desired droplet size distribution, the amount of material to be homogenized, the desired throughput, the energy consumption, the physicochemical properties of the component phases and the final product, the equipment available, the initial costs, and the running costs (Karbstein and Schubert 1995, Schubert 1997). After choosing the most suitable type of homogenizer, one must select the optimum operating conditions for that particular device. For example, the pressure, number of passes, and valve design may be optimized for a high-pressure valve homogenizer or microfluidizer, whereas the rotor-stator design, rotation speed, and duration may be varied for a high shear mixer or colloid mill. Some of the most important differences between homogenizers are briefly discussed in the following text: Energy densities and efficiency: Homogenizers vary considerably in the range of energy densities that they are capable of generating, and in the efficiency of these energy levels at disrupting emulsion droplets. A plot of the mean droplet size produced by a homogenizer as a function of the energy density is one of the most convenient means of establishing its effectiveness and efficiency at creating emulsions (Figure 6.15). High shear mixers and colloid mills are only suitable for preparing emulsions with relatively large droplet sizes (r > 1 μm), whereas the other major types of homogenizer can be used to prepare submicron droplets. Microfluidizers are particularly suitable for producing fine emulsions using relatively low energy inputs. 271 Emulsion Formation 100 Colloid mill High shear mixer 10 d32 (μm) Membrane 1 0.1 0.01 HPVH Sonicator Microfluidizer 0.1 1 Ev (MJ m–3) 10 100 FIGURE 6.15 Comparison of the energy efficiency of different types of mechanical homogenizers—variation of mean droplet diameter with energy input: HPVH = high-pressure valve homogenizer. In practice, the precise relationship for a given device depends on the specific characteristics of the emulsion and homogenizer. (Adapted from Walstra, P., Formation of emulsions, in Encyclopedia of Emulsion Technology, Becher, P., ed., Marcel Dekker, New York, 1983, pp. 57–128; Schubert, H. et al., Trends Food Sci. Technol., 14(1–2), 9, 2003.) Primary versus secondary homogenization: Some homogenizers can be used to convert separate oil and aqueous phases directly into an emulsion (primary homogenization), whereas others can only be used to reduce the size of the droplets in a pre-existing emulsion (secondary homogenization). High shear mixers, membrane homogenizers, ultrasonic homogenizers, and some forms of microfluidizer can be used for primary homogenization, whereas high-pressure valve homogenizers and colloid mills are most suitable for secondary homogenization. Product rheology: Another important difference between homogenizers is the rheological characteristics of the materials that they can handle during the homogenization process. High shear mixers and colloid mills can be used to homogenize highly viscous fluids, whereas most other types of homogenizer are only suitable for low and/or intermediate viscosity fluids. Product volume or throughput: For industrial applications, one usually wants to produce the largest volume of material in the shortest amount of time. Most of the homogenizers mentioned in this chapter are capable of high volume throughputs, either in a batch or continuous operation mode, although there are significant differences in their maximum capacities. In particular, membrane and microchannel homogenizers usually have appreciably lower maximum flow rates than the other major types of homogenizer. In research and development applications, one often needs to prepare small volumes of emulsion, because the ingredients are relatively expensive. In these situations, it is often possible to use scaled-down versions of industrial equipment or to use small-scale instruments specifically designed for laboratory utilization. Ultrasonic transducers are widely available in many laboratories to produce small volumes of emulsions, but one has to be careful that the high-intensity levels used do not promote degradation reactions, such as protein denaturation, lipid oxidation, or polysaccharide depolymerization. Microfluidizers are commercially available that can produce emulsions using very small volumes of the component phases. Droplet distribution characteristics: There are appreciable differences in the distribution of droplet sizes that can be produced by different kinds of homogenizer, that is, the polydispersity. Some homogenizers are capable of generating narrow droplet size distributions (e.g., membrane and microchannel homogenizers), whereas others are only capable of generating rather broad distributions (e.g., high shear 272 Food Emulsions: Principles, Practices, and Techniques mixers, high-pressure valve homogenizers, colloid mills, microfluidizers, and ultrasonic homogenizers). The use of membrane or microchannel homogenizers may be particularly useful in situations where narrow droplet size distributions are important, for example, in fundamental studies of the relationship between droplet characteristics and emulsion properties. Selecting and purchasing homogenizers: The selection of an appropriate homogenizer for a particular application usually involves close cooperation between the food processor and the manufacturer of the equipment. The food processor must first specify the desired throughput, pressure, temperature, particle size, hygiene requirements, and properties of the sample. The equipment manufacturer will then be able to recommend a piece of equipment that is most suitable for the specific product, for example, size, valvedesign, flow rates, and construction materials. It is a good practice for food processors to test a number of homogenizers from different manufacturers under conditions that approximate those that will be used in the factory prior to making a purchase. 6.6 Factors Influencing Droplet Size The size of the droplets produced during homogenization is important, because it determines the stability, appearance, texture, taste, and gastrointestinal fate of the final product (Chapters 7 through 11). To create a product with specific properties, it is therefore necessary to ensure that the majority of droplets fall within some pre-established optimum size range. For this reason, it is important for food scientists to be aware of the major factors that influence the size of the droplets produced by homogenizers. 6.6.1 Emulsifier Type and Concentration For a fixed concentration of oil, water, and emulsifier, there is a maximum interfacial area that can be completely covered by the emulsifier. As homogenization proceeds, the size of the droplets decreases and therefore, the interfacial area increases (Figure 6.16). Once the droplets fall below a certain size, 1 0.9 0.8 d32 (µm) 0.7 0.6 Emulsifier-poor regime 0.5 Emulsifier-rich regime 0.4 0.3 0.2 0.1 0 Size limited by emulsifier 0 0.5 Size limited by homogenizer 1 c (wt.%) 1.5 2 FIGURE 6.16 The change in mean droplet size with increasing emulsifier concentration can be divided into two regions: an emulsifier-poor regime, where the droplet size is limited by the amount of emulsifier present; an emulsifier-rich regime, where the droplet size is limited by the maximum disruptive energy that the homogenizer can generate. 273 Emulsion Formation there may be insufficient emulsifier present to completely cover their surface and so they will tend to coalesce with their neighbors. The minimum size of stable droplets that can be theoretically produced during homogenization (assuming monodisperse droplets*) is therefore governed by the type and concentration of emulsifier present: rmin = 3 × Gsat × f 3 × Gsat × f = cS cS¢ (1 - f) (6.11) where Γsat is the excess surface concentration (or surface load) of the emulsifier at saturation (in kg m−2) ϕ is the disperse phase volume fraction cS is the concentration of emulsifier in the emulsion (in kg m−3) cS¢ is the concentration of emulsifier in the continuous phase (in kg m−3) This equation indicates that the minimum droplet size can be decreased by increasing the emulsifier concentration, decreasing the droplet concentration, or using an emulsifier with a lower Γsat. For a 10 vol.% (ϕ = 0.1) oil-in-water emulsion containing 1% (∼10 kg m−3) of emulsifier, the minimum droplet radius is about 60 nm (assuming Γsat = 2 × 10 −6 kg m−2). In practice, there are a number of factors that mean that the droplet size produced during homogenization is greater than this theoretical minimum. In order to reach the theoretical minimum, a homogenizer must be capable of generating a pressure gradient that is large enough to disrupt any droplets greater than rmin, that is >2γ/rmin2 (Section 6.4). Some types of homogenizer are incapable of generating such high-pressure gradients and are therefore unsuitable for producing emulsions with small droplet sizes, even though there may be sufficient emulsifier present. The emulsion must also spend sufficient time within the homogenization zone for all of the droplets to be completely disrupted. If an emulsion passes through a homogenizer too rapidly or if there is an uneven distribution of disruptive energy within the homogenization zone, then some of the droplets may not be disrupted (Walstra and Smulder 1988). Even if a homogenizer is capable of producing small droplets, the emulsifier molecules must adsorb rapidly enough to form a protective interfacial layer around the droplets that prevents them from coalescing with their neighbors (Baret et al. 2009). The emulsifier also influences the droplet size by reducing the interfacial tension between the oil and aqueous phases, thereby facilitating droplet disruption. Consequently, the more rapidly an emulsifier adsorbs, and the greater the reduction of the interfacial tension, the smaller the droplets that can be produced at a fixed homogenizer energy input. Many different types of emulsifier can be used in the food industry, and each of these exhibits different characteristics during homogenization, for example, the speed at which they adsorb, the maximum reduction in interfacial tension, and the effectiveness of the interfacial coating at preventing droplet coalescence. Factors that influence the adsorption kinetics of emulsifiers, and their effectiveness at reducing interfacial tension, were discussed in Chapter 5, while factors that affect the stability of droplets against coalescence will be covered in Chapter 7. Some of the different kinds of behavior that could be observed during homogenization are illustrated schematically in Figure 6.16, which shows the change in droplet size with emulsifier concentration. The influence of emulsifier concentration on droplet size can be divided into two regions (Tcholakova et al. 2006): 1. Emulsifier-poor regime: When the emulsifier concentration is limiting (i.e., there is insufficient emulsifier present to cover all of the droplet surface area created by the homogenizer), then the droplet size is governed primarily by the emulsifier concentration, rather than the energy input of the homogenizer. The surface load of the emulsifier remains relatively constant in this regime and is close to the value of the excess surface concentration of the emulsifier at saturation (ΓSAT). Under these conditions, the mean droplet size produced by homogenization is given by Equation 6.11. * For a polydisperse emulsion, the radius used in Equation 6.11 should be the volume-surface mean radius (Chapter 1). 274 Food Emulsions: Principles, Practices, and Techniques 2. Emulsifier-rich regime. When the emulsifier concentration is in excess (i.e., there is more emulsifier present than is required to completely cover all of the droplet surface area created by the homogenizer), then the droplet size is relatively independent of emulsifier concentration and depends primarily on the energy input of the homogenizer. Under these circumstances, the mean droplet diameter that can be produced depends on the flow conditions prevalent in the homogenizer (see Section 6.6.2). If the emulsifier does not form multiple layers at the interface, then the surface load of the emulsifier will remain relatively constant in this regime (Γ ∼ ΓSAT). Alternatively, if the emulsifier is capable of forming multiple layers at the interface, then the surface load may increase as the overall concentration of emulsifier in the system is increased (Γ > ΓSAT). This latter affect has been observed for globular proteins, such as whey proteins, where the surface load increases up to a certain level as the overall emulsifier concentration is increased. Typical mean droplet size versus emulsifier concentration curves for food-grade surface active biopolymers (whey protein, gum arabic, and modified starch) are shown in Figure 6.17. This experimental data highlights the fact that different emulsifiers are more or less effective at producing small droplets under similar homogenization conditions. In this example, the whey protein (β-lactoglobulin) is much more effective at producing small droplets at low concentrations than the gum arabic or modified starch. However, the polysaccharides usually produce emulsion droplets that are more stable to changes in environmental conditions than the proteins (Chanamai and McClements 2002, Charoen et al. 2011). The differences in the efficiencies of different types of food-grade emulsifier at reducing the droplet size during homogenization can be attributed to various factors, including their surface loads, adsorption rates, ability to decrease the interfacial tension, and ability to protect droplets from coalescence. 6.6.2 Energy Input The size of the droplets in an emulsion can be reduced by increasing the intensity or duration of disruptive energy supplied during homogenization (as long as there is sufficient emulsifier to cover all the 5 Gum arabic Modified starch BLG Mean particle diameter (μm) 4 3 2 1 0 0 1 2 3 4 5 Emulsifier concentration (wt.%) FIGURE 6.17 Experimental measurements of the change in mean droplet diameter with increasing emulsifier concentration for three food-grade biopolymers: β-lactoglobulin (BLG); gum arabic; and modified starch. The emulsions were 5 wt.% orange oil-in-water emulsions produced at pH 3 using a microfluidizer (9000 psi, 3 passes). 275 Emulsion Formation surfaces of the droplets formed). The range of energy inputs that can be achieved by a given homogenization device, and the effectiveness of this energy at disrupting the droplets, depends on the type of homogenizer used (Figure 6.15). The energy input can be increased in a number of different ways, depending on the nature of the homogenizer. In a high shear mixer, the energy input can be enhanced by increasing the rotation speed or the length of time that the material is blended. In a high-pressure valve homogenizer, it can be enhanced by increasing the homogenization pressure or recirculating the emulsion through the device a number of times. In a colloid mill, it can be increased by using a narrower gap between the stator and rotor, increasing the rotation speed, by using disks with roughened surfaces, or by passing the emulsion through the device a number of times. In an ultrasonic homogenizer, the energy input can be increased by increasing the intensity of the ultrasonic wave or by sonicating for a longer time. In a microfluidizer, the energy input can be increased by increasing the homogenization pressure or by recirculating the emulsion through the device a number of times. In a membrane homogenizer, the energy input can be increased by increasing the pressure at which the liquid is forced through the membrane or the flow rate of the continuous phase across the membrane. Under a given set of homogenization conditions (energy intensity, emulsion composition, and temperature), there is a certain size below which the emulsion droplets cannot be reduced any further with repeated homogenization, and therefore homogenizing the system any longer would be inefficient. A schematic illustration of the influence of homogenization pressure on mean droplet size that demonstrates this effect is shown in Figure 6.18. The droplet size continues to decrease with increasing homogenization pressure if there is excess emulsifier present, but it may reach a constant value if there is insufficient emulsifier present. Increasing the energy input usually leads to an increase in production costs, and therefore, a food manufacturer must establish the optimum compromise between droplet size and manufacturing costs. As mentioned earlier, the energy input required to produce an emulsion containing droplets of a given size depends on the energy efficiency of the homogenizer used. Consequently, a manufacturer should take this into account when selecting a homogenizer to produce droplets with a given particle size distribution. 1 0.9 0.8 d32 (µm) 0.7 0.6 0.5 0.4 0.3 Limited emulsifier 0.2 0.1 0 Excess emulsifier 0 5 10 Pressure (kpsi) 15 FIGURE 6.18 The change in mean droplet size with increasing homogenization pressure depends on the amount of emulsifier present. If there is sufficient emulsifier, then the droplet size usually decreases with increasing homogenization pressure. However, if there is insufficient emulsifier, the droplet size will reach a limiting value determined by emulsifier concentration. 276 Food Emulsions: Principles, Practices, and Techniques 1 Flocculation or coalescence 0.9 0.8 0.7 d32 (μm) 0.6 0.5 0.4 0.3 Over processing 0.2 0.1 0 0 5 10 Pressure (kpsi) 15 FIGURE 6.19 For certain types of emulsifier, the droplet size may actually increase with increasing homogenization pressure once a critical pressure is exceeded due to flocculation or coalescence, which is known as overprocessing. Under most circumstances, there is a decrease in droplet size as the energy input is increased. Nevertheless, there may be occasions when increasing the energy input actually leads to an increase in droplet size, which is referred to as overprocessing (Figure 6.19). A number of possible physicochemical mechanisms have been proposed for this kind of overprocessing, including a relatively slow rate of emulsifier adsorption compared to the increased coalescence frequency, a relatively short residence time of droplets within the disruptive zone, depletion of emulsifier, and loss of emulsifier functionality (Jafari et al. 2008, Santana et al. 2013). For example, protein-stabilized emulsions may be destabilized at high homogenization pressures due to surface or thermal denaturation of the adsorbed globular proteins (Rampon et al. 2003). On the other hand, nonionic surfactants may lose their ability to stabilize emulsions if the temperature within the homogenizer becomes too close to the phase inversion temperature (PIT) due to rapid droplet coalescence (Rao and McClements 2011). 6.6.3 Properties of Component Phases The composition and physicochemical properties of both the oil and aqueous phases also influence the size of the droplets produced during homogenization (Walstra and Smulder 1988, Wooster et al. 2008, Qian and McClements 2011, Lee and Norton 2013). Variations in the type of oil or aqueous phase will alter the viscosity ratio, ηD/ηC, which influences the minimum droplet size that can be produced under steady-state conditions (Section 6.4). This effect is shown in Figure 6.20 for oil-in-water emulsions produced using a microfluidizer: very small oil droplets can be produced by reducing the viscosity ratio toward unity. Different oils also have different interfacial tensions when placed in contact with water, because they have different molecular structures or because they contain different amounts of surfaceactive impurities, such as free fatty acids, monoacylglycerols, or diacylglycerols. These surface-active impurities will tend to accumulate at the oil–water interface and lower the interfacial tension, thereby lowering the amount of energy required to disrupt a droplet. The aqueous phase of an emulsion may contain a wide variety of components, including minerals, acids, bases, biopolymers, sugars, alcohols, and gas bubbles. Many of these components may alter the 277 Emulsion Formation 150 Mean droplet diameter (nm) 130 110 90 70 50 0 10 20 30 40 50 log (ηD/ηC) FIGURE 6.20 Example of the influence of viscosity ratio on the mean droplet diameter of oil-in-water emulsions produced using a microfluidizer. The viscosity ratio was varied by using different ratios of corn oil and octadecane in the oil phase, and different ratios of glycerol and water in the aqueous phase. size of the droplets produced during homogenization because of their influence on rheology, interfacial tension, coalescence stability, emulsifier available, or emulsifier adsorption kinetics. For example, the presence of low concentrations of short chain alcohols in the aqueous phase of an emulsion reduces the size of the droplets produced during homogenization because of the reduction in interfacial tension (Zeeb et al. 2014). The presence of biopolymers in an aqueous phase has been shown to increase the droplet size produced during homogenization due to their ability to suppress the formation of small eddies during turbulence (Walstra 1983). Protein-stabilized emulsions cannot be produced close to the protein isoelectric point or at high ionic strengths, because the proteins are susceptible to aggregation. A knowledge of the composition of both the oil and aqueous phases of an emulsion and the role that each component plays during homogenization is therefore important when optimizing the size of the droplets produced by a homogenizer. Some studies have shown that the smallest droplet size that can be achieved using a high-pressure valve homogenizer increases as the disperse phase volume fraction increases (Phipps 1985). There are a number of possible reasons for this: (1) increasing the viscosity of an emulsion may suppress the formation of eddies responsible for breaking up droplets; (2) if the emulsifier concentration is kept constant, there may be insufficient present to completely cover the droplets; and (3) the rate of droplet coalescence is increased. On the other hand, other studies have shown that there is little change in the mean droplet diameter with increasing disperse phase volume fraction, providing that the ratio of emulsifier to disperse phase is kept constant and there is sufficient emulsifier present to cover all of the droplets formed (Schubert et al. 2003). 6.6.4 Temperature Temperature influences the size of the droplets produced during homogenization in a number of ways (Santana et al. 2013). The viscosity of both the oil and aqueous phases is temperature-dependent, and therefore, the minimum droplet size that can be produced may be altered because of a variation in the 278 Food Emulsions: Principles, Practices, and Techniques viscosity ratio, ηD/ηC (Section 6.4). Usually the viscosity of oils decreases more rapidly with increasing temperature than the viscosity of water; hence, ηD/ηC for an oil-in-water emulsion would tend to decrease, thereby facilitating droplet disruption at higher temperatures. Heating an emulsion usually causes a slight reduction in the interfacial tension between the oil and water phases that would be expected to facilitate the production of small droplets (Section 6.4). Certain types of emulsifiers lose their ability to stabilize emulsion droplets against flocculation and aggregation when they are heated above a certain temperature. For example, when nonionic surfactants are heated close to their PIT, they are no longer effective at preventing droplet coalescence,* or when globular proteins are heated above a critical temperature, they unfold and aggregate (Chapter 4). Alterations in temperature also influence the competitive adsorption of surface-active components, thereby altering interfacial composition, which may in turn alter the physicochemical properties of emulsions. The temperature is also important, because it determines the physical state of the lipid phase (Chapter 4). It is difficult to homogenize a fat that is solid, because it will not flow through a homogenizer or because a high amount of energy is required to break up the fat crystals into small particles. There are also problems associated with the homogenization of oils that contain even small amounts of fat crystals because of partial coalescence (Chapter 7). The crystals from one droplet penetrate into another droplet, leading to the formation of clumps. Extensive clump formation leads to the generation of large particles and to a dramatic increase in viscosity that would cause a homogenizer to become blocked. For this reason, it is usually necessary to warm a sample prior to homogenization to ensure that the lipid phase is completely liquid. For example, milk fat is usually heated to about 40°C to melt all the fat crystals prior to homogenization (Phipps 1985). 6.6.5 Predicting Droplet Sizes Produced by Homogenization There have been considerable advances in the development of mathematical models to predict the major factors influencing the particle size distribution produced during homogenization (Hakansson et al. 2009, Maindarkar et al. 2012, Hakansson et al. 2013, Becker et al. 2014). Population balance models have proved particularly effective as a tool for predicting particle size distributions based on homogenizer operating conditions (such as valve design, homogenization pressure, and number of passes) and system properties (such as disperse phase volume fraction, viscosities, and interfacial tensions) (Hakansson et al. 2009, 2013, Maindarkar et al. 2012, 2013). Population balance models typically involve solving a series of equations that take into account droplet disruption, droplet coalescence, and emulsifier adsorption within the homogenizer. These models are often coupled with CFD simulations to input information about the flow profiles in different regions of the homogenizer (Dubbelboer et al. 2014). Computational models that can predict the particle size distribution of emulsions during homogenization may be particularly useful for optimizing product formulation and operating conditions of commercial food products. 6.7 Low-Energy Homogenization Methods There has been considerable interest in the development of low-energy homogenization methods to produce emulsions due to their potential advantages for certain applications, for example, low equipment costs, simplicity of operation, and ability to produce very small droplet sizes (Santana et al. 2013). These methods rely on the spontaneous formation of oil droplets in certain types of surfactant–oil–water (SOW) mixtures when either their composition or environment is altered in a specific manner (Anton and Vandamme 2009, McClements and Rao 2011, Solans and Sole 2012). A number of these low-energy methods can be utilized to fabricate food-grade emulsions (Figure 6.21). * It should be noted that droplet disruption is highly efficient near the surfactant PIT, so that it is often possible to efficiently homogenize an emulsion at this temperature to produce small droplets, and then rapidly cool it to a lower temperature to reduce coalescence. 279 Emulsion Formation W/O emulsion T > PIT Cool Microemulsion (a) T = PIT Heat O/W emulsion W/O O/W/O O/W (b) T < PIT (c) FIGURE 6.21 Schematic representation of various low-energy methods that can be used to produce food-grade emulsions: (a) spontaneous emulsification, (b) emulsion inversion point, and (c) phase inversion temperature. 6.7.1 Spontaneous Emulsification Spontaneous emulsification methods simply involve titrating a mixture of oil and water-soluble surfactant into a water phase with continuous stirring (Saberi et al. 2013a–c). Small oil droplets are spontaneously formed at the oil–water boundary as the surfactant molecules move from the oil phase to the water phase. 6.7.2 Emulsion Inversion Point Methods Emulsion inversion point methods involve titrating water into a mixture of oil and water-soluble surfactant with continuous stirring. As increasing amounts of water are added, a W/O emulsion is initially formed, then an O/W/O emulsion, and then an O/W emulsion. It has been proposed that the small internal oil droplets within the O/W/O emulsion (that later become the oil droplets in the O/W emulsion) are formed by spontaneous emulsification at the oil–water boundary (Ostertag et al. 2012). Consequently, the size of the final droplets formed using both methods are closely related (Komaiko and McClements 2015). 6.7.3 Phase Inversion Temperature Methods PIT methods rely on heating an SOW mixture around or slightly above its PIT and then quench cooling with continuous stirring (Anton and Vandamme 2009, Roger et al. 2010). This method relies on changes in the optimum curvature and solubility of nonionic surfactant molecules when they are heated (Figure 6.21c). At temperatures above the PIT, the surfactant is more soluble in the oil phase and has a curvature that favors W/O emulsions. When the emulsion passes through the PIT, the optimum curvature tends toward unity, thereby leading to an ultralow interfacial tension and a highly dynamic interface. In addition, the surfactant molecules become more hydrophilic and water-soluble as the head groups become hydrated at lower temperatures. It has been proposed that ultrafine oil droplets are formed when an SOW mixture is rapidly cooled from around the PIT due to the movement of surfactant molecules from the oil to the aqueous phases (in a process similar to spontaneous 280 Food Emulsions: Principles, Practices, and Techniques Particle diameter (μm) 50 Low energy High energy 5 0.5 0.05 0.0 1.0 SOR 2.0 FIGURE 6.22 Comparison of influence of surfactant-to-oil ratio on the size of the droplets in oil-in-water emulsions produced by high-energy (microfluidizer) and low-energy (EPI) methods using a nonionic surfactant. The oil phase consisted of MCT and vitamin E acetate (2:8), and the surfactant was Tween 80. emulsification) (Anton and Vandamme 2009). Alternatively, it has been proposed that small oil droplets are produced upon rapid cooling due to trapping of the small particles present in the microemulsion phase that forms around the PIT (Roger et al. 2010). 6.7.4 Comparison with High-Energy Methods The major advantages of low-energy homogenization methods are that they are very simple to implement, they do not require any expensive equipment, and they can produce very fine emulsion droplets (often so small that they produce transparent systems). The major disadvantages are that they typically require relatively high surfactant-to-oil ratios, and can currently only be utilized with synthetic surfactants and certain types of oil phase (Komaiko and McClements 2015). A comparison of the influence of surfactant-to-oil ratio on the droplet size produced by high-energy (microfluidizer) and low-energy (emulsion phase inversion) methods is shown in Figure 6.22. Both methods are capable of producing very fine droplets at high surfactant concentrations, but only the high-energy method is able to produce very small droplets at low surfactant levels. 6.8 Demulsification Demulsification is the process whereby an emulsion is converted into the separate oil and aqueous phases from which it was comprised, and is therefore the opposite process to homogenization (Menon and Wasan 1985, Al-Sabagh et al. 2011). There are a number of technological processes in the food industry where demulsification is important, for example, oil recovery from plant or animal tissue or the separation of lipid and aqueous phases. Demulsification is also important in research and development, because it is often necessary to divide an emulsion into the separate oil and aqueous phases so that their composition or properties can be characterized. For example, the oil phase may be extracted from an emulsion to determine the extent of lipid oxidation, to measure the oil–water partition coefficient of a food additive, or to determine the amount of a lipophilic bioactive component solubilized within it. Demulsification is Emulsion Formation 281 achieved by causing the droplets to come into close contact with each other and then coalesce. As this process continues, it eventually leads to the complete separation of the oil and aqueous phases. Knowledge of the physical principles of demulsification requires an understanding of the factors that determine the stability of emulsions to flocculation, coalescence, and gravitational separation (Chapter 7). A variety of different types of emulsifier are used in the food industry to stabilize droplets against flocculation and coalescence (Chapter 4). Each type of emulsifier relies on different physicochemical mechanisms to prevent droplet aggregation, with the most important being electrostatic and steric repulsion (Chapter 3). The selection of the most appropriate demulsification technique for a given emulsion therefore depends on knowledge of the type of emulsifier used to stabilize the system and of the mechanisms by which it provides stability. In the following sections, we begin by considering demulsification methods appropriate for specific types of emulsifier, and then we consider more general methods suitable for most types of emulsifier. 6.8.1 Nonionic Surfactants Nonionic surfactants usually stabilize emulsion droplets against aggregation through a combination of steric, hydration, and thermal fluctuation interactions (Chapter 3), which are often grouped together and referred to simply as steric interactions. In some cases, droplets stabilized by nonionic surfactants may have an electrical charge due to adsorption of ionic impurities to the droplet surfaces, and so electrostatic interactions may also be important. Emulsion droplets coated by nonionic surfactants may therefore be destabilized by reducing the steric repulsion between them (e.g., by heating to reduce the hydration of the polar head groups) or by reducing the electrostatic repulsion (e.g., by altering the pH or ionic strength). Once the droplets come into close contact, the interfacial layers coating them may rupture, thereby leading to coalescence and phase separation. Demulsification can therefore be achieved by altering the properties of an emulsion so that the droplets come into close contact for prolonged periods. In the remainder of this section, some possible mechanisms of promoting demulsification in nonionic stabilized systems are briefly discussed: Heating: The head groups of nonionic surfactants become progressively dehydrated when they are heated, which reduces the hydration repulsion between the droplets and allows them to come closer together. In addition, the optimum curvature of the surfactant monolayer tends toward zero as the size of the head group decreases, which increases the likelihood of droplet coalescence and phase separation (Chapter 4). The efficacy of this method depends on the PIT of the nonionic surfactant used, which depends on head-group and tail-group properties. The PIT must be within an experimentally attainable range in order to promote rapid droplet coalescence and phase separation. The PIT can sometimes be reduced by adding cosolvents or cosurfactants to the system, such as polyols, salts, or alcohols (Saberi et al. 2013a,b, 2014). This demulsification technique cannot therefore be used for emulsions stabilized by nonionic surfactants with PIT greater than about 100°C, or for systems where heating causes degradation or evaporation of one of the components being analyzed. In these cases, it is necessary to induce demulsification using alternative methods. Freezing: When emulsions containing droplets coated by nonionic surfactants are frozen and then thawed, they often undergo extensive coalescence and phase separation. Consequently, one or more freeze–thaw cycles can be utilized to promote demulsification of these systems. An example of the change in microstructure of an oil-in-water emulsion after freezing and thawing is shown in Figure 6.23. It should be noted that this method may not work in emulsions containing relatively high solute concentrations (such as sugars or polyols) because of freeze-concentration effects, that is, the formation of nonfrozen regions that prevent the droplets from being forced into close proximity (Degner et al. 2014). Alcohols: The addition of medium chain alcohols has also been found to be effective at promoting demulsification of some emulsions (Menon and Wasan 1985). There are two possible explanations for this behavior: (1) The alcohol displaces some of the surfactant molecules from the interface and forms an interfacial layer that provides little protection against droplet aggregation, and (2) the alcohol molecules are able to get between the tails of the surfactant molecules at the interface, thereby causing the optimum curvature of the interface to tend towards zero and increasing the likelihood of droplet coalescence (Chapter 4). 282 Food Emulsions: Principles, Practices, and Techniques (a) (b) FIGURE 6.23 Oil-in-water emulsions stabilized may be destabilized by subjecting them to a freeze–thaw cycle. These optical microscopy images show the microstructure of an emulsion (a) before and (b) after freezing and thawing. Acids, bases, and enzymes: In some emulsions, it is possible to promote droplet coalescence by adding a strong acid or base that cleaves the head groups of the surfactants from their tails so that the polar head group moves into the aqueous phase and the nonpolar tail moves into the droplet, thereby providing little protection against droplet coalescence. Some enzymes are also able to cleave the bond between the head group and tail group of surfactants, e.g., ester bonds. 6.8.2 Ionic Surfactants Ionic surfactants stabilize droplets against coalescence principally by electrostatic repulsion (Chapter 3). Like nonionic surfactants, the interfacial layers formed by ionic surfactants are not particularly resistant to rupture once the droplets are brought into close contact (Evans and Wennerstrom 1999). The most effective method of inducing droplet coalescence in these systems is therefore to reduce the magnitude of the electrostatic repulsion between the droplets. This can be achieved by adding electrolyte to the aqueous phase of the emulsion so as to screen the electrostatic interactions. Sufficient electrolyte must be added so that the energy barrier between the droplets decreases below a critical level, ∼20 kT (Chapter 3). This process can most easily be achieved using multivalent ions, because they are more effective at screening electrostatic interactions at low concentrations than monovalent ions. Moreover, the addition of electrolyte may cause the optimum curvature of ionic surfactants to tend toward zero, thereby promoting droplet coalescence (Chapter 4). Alternatively, the pH may be altered so that the surfactant loses its charge, which depends on the dissociation constant (pKa) of the ionizable groups (Choi et al. 2011). Electromechanical methods can also be used to promote demulsification (Ichikawa et al. 2004). An electric field is applied across an emulsion that causes the charged droplets to move toward the oppositely charged electrode. A semipermeable membrane is placed across the path of the droplets that captures the droplets but allows the continuous phase to pass through. The droplets are therefore forced against the membrane until their interfacial layers are ruptured and the droplets coalesce. 6.8.3 Biopolymer Emulsifiers Biopolymers principally stabilize droplets against coalescence through a combination of electrostatic and steric interactions (Dickinson 2003, McClements 2004). In addition, they tend to form thick viscoelastic layers that are highly resistant to rupture. There are two different strategies that can be used to induce droplet coalescence in this type of system: 1. The biopolymer can be digested by strong acids, bases, or enzymes so that it is broken into small fragments that are either not surface-active or do not form a sufficiently strong interfacial layer (Rosenthal et al. 1996). 2. The biopolymers are displaced from the interface by small molecule surfactants and then the surfactant-coated droplets are destabilized using one of the methods described in the previous Emulsion Formation 283 sections. Some proteins are capable of forming an interfacial layer in which the molecules are covalently bound to each other through disulfide bonds. In order to displace these proteins, it may be necessary to cleave the disulfide bonds prior to displacing the proteins, for example, by adding mercaptoethanol. Gerber and Babcock methods of determining the total fat content of milk are examples of the first of these strategies, while the detergent method is an example of the second (O’Keefe and Pike 2010). 6.8.4 General Methods of Demulsification A variety of physical techniques are available that can be used to promote demulsification in most types of emulsions (Menon and Wasan 1985). In all of the demulsification processes mentioned earlier, the separation of the oil phase from the aqueous phase can be facilitated by centrifuging the emulsion after the coalescence process has been initiated. In some emulsions, it is also possible to separate the phases directly by centrifugation at high speeds, without the need for any pretreatment. Centrifugation forces the droplets to one end of the container, which causes their interfacial layers to become ruptured and therefore leads to phase separation. Demulsification can also be achieved using various types of filtration device. The emulsion is passed through a filter that adsorbs emulsion droplets. When a number of these adsorbed droplets come into close contact, they merge together to form a single large droplet that is released back into the aqueous phase. As the emulsion passes through the filter, this process continues until eventually the oil and water phases are completely separated from each another. Finally, freeze–thaw cycling is a particularly efficient method of promoting droplet coalescence and oiling-off in many types of emulsion (Figure 6.23). The emulsion is cooled to a temperature where the water freezes, and is then cooled back to room temperature. This process can be repeated a number of times to improve its effectiveness. 6.8.5 Selection of the Most Appropriate Demulsification Technique As well as depending on the type of emulsifier present, the choice of an appropriate demulsification technique also depends on the sensitivity of the other components in the system to the separation process. For example, if one is monitoring lipid oxidation or trying to determine the concentration of an oil-soluble volatile component, it is inadvisable to use a demulsification technique that requires excessive heating. On the other hand, if the sample contains a lipid phase that is crystalline, it is usually necessary to warm the sample to a temperature where all the fat melts prior to carrying out the demulsification procedure. 6.9 Future Developments Homogenization is an extremely important step in the production of emulsion-based food products. The efficiency of this process has a large impact on the bulk physicochemical and sensory properties of the final product. This chapter has reviewed the progress that has already been made in identifying the factors that influence homogenization using both high-energy and low-energy methods. Nevertheless, a great deal of research is still required before we can fully understand these methods because of the inherent complexity of the physicochemical processes involved, and because foods are often complex multicomponent and multiphase systems. For high-energy methods, rapid progress has been made in understanding and predicting the major factors that influence the particle size distribution of emulsions produced by homogenizers using advanced computational models. However, these theories need to be extended to give a more realistic description of emulsion formation in complex food matrices using commercial homogenization devices. In addition, systematic experiments using well-characterized emulsions and homogenization devices still need to be carried out to better understand the factors that govern the droplet size distribution produced. The development of novel methods of fabricating emulsions (such as membrane and 284 Food Emulsions: Principles, Practices, and Techniques Spray drop Feed fluid Atomizer Solvent Drying chamber Powder collection Wall material Encapsulated oil droplet Powder particle Finishing: Agglomeration Coating FIGURE 6.24 Spray driers can be used to produce powdered forms of emulsions. Many emulsion-based products involved posthomogenization steps that impact their functional properties. microchannel homogenizers) is likely to continue, which may lead to the introduction of more efficient methods of creating droplets with tailored dimensions and internal structures. However, much work is still required before these devices can be scaled up so that they can be utilized in a factory environment for large-scale production of food emulsions. Low-energy methods have potential advantages for certain applications in the food industry due to their low cost, simplicity, and ability to form very small droplets. Nevertheless, research is still needed to determine whether the amount of surfactant required to produce these emulsions can be reduced to commercially acceptable levels, and to establish the range of oil and surfactant types that are suitable for use in this method. In particular, it would be highly advantageous if these emulsions could be produced from natural emulsifiers (such as phospholipids, proteins, or polysaccharides), rather than synthetic surfactants (such as Tweens). Before finishing this chapter, it is important to mention that homogenization is only one step in the formation of a food emulsion. A number of other processing operations usually come before or after homogenization, including chilling, freezing, pasteurization, drying, mixing, churning, and whipping. For example, certain types of food emulsion are converted into powders using spray drying (Figure 6.24) or other dehydration methods. The quality of the final product is determined by the effect that each of these processing operations has on the properties of the food. Homogenization efficiency may be influenced by the effectiveness of any of the preceding processing operations, and it may alter the effectiveness of any of the following processing operations. Thus, it is important to establish the interrelationship between the various food processing operations on the final properties of a product. REFERENCES Abbas, S., K. Hayat, E. Karangwa, M. Bashari, and X. M. Zhang (2013). An overview of ultrasound-assisted food-grade nanoemulsions. Food Engineering Reviews 5(3): 139–157. Al-Sabagh, A. M., N. G. Kandile, and M. R. N. El-Din (2011). Functions of demulsifiers in the petroleum industry. Separation Science and Technology 46(7): 1144–1163. Anton, N. and T. F. Vandamme (2009). The universality of low-energy nano-emulsification. International Journal of Pharmaceutics 377(1–2): 142–147. Baret, J. C., F. Kleinschmidt, A. El Harrak, and A. D. Griffiths (2009). Kinetic aspects of emulsion stabilization by surfactants: A Microfluidic Analysis. Langmuir 25(11): 6088–6093. Emulsion Formation 285 Becker, P. J., F. Puel, A. Dubbelboer, J. Janssen, and N. Sheibat-Othman (2014). Coupled population balance-CFD simulation of droplet breakup in a high pressure homogenizer. Computers & Chemical Engineering 68: 140–150. Bos, M. A. and T. van Vliet (2001). Interfacial rheological properties of adsorbed protein layers and surfactants: A review. Advances in Colloid and Interface Science 91(3): 437–471. Canselier, J. R., H. Delmas, A. M. Wilhelm, and B. Abismail (2002). Ultrasound emulsification: An overview. Journal of Dispersion Science and Technology 23(1–3): 333–349. Chanamai, R. and D. J. McClements (2002). Comparison of gum arabic, modified starch, and whey protein isolate as emulsifiers: Influence of pH, CaCl(2) and temperature. Journal of Food Science 67(1): 120–125. Charoen, R., A. Jangchud, K. Jangchud, T. Harnsilawat, O. Naivikul, and D. J. McClements (2011). Influence of biopolymer emulsifier type on formation and stability of rice bran oil-in-water emulsions: Whey protein, gum arabic, and modified starch. Journal of Food Science 76(1): E165–E172. Choi, S. J., E. A. Decker, L. Henson, L. M. Popplewell, H. Xiao, and D. J. McClements (2011). Formulation and properties of model beverage emulsions stabilized by sucrose monopalmitate: Influence of pH and lyso-lecithin addition. Food Research International 44(9): 3006–3012. Degner, B. M., C. Chung, V. Schlegel, R. Hutkins, and D. J. McClements (2014). Factors influencing the freeze-thaw stability of emulsion-based foods. Comprehensive Reviews in Food Science and Food Safety 13(2): 98–113. Delmas, T., H. Piraux, A. C. Couffin, I. Texier, F. Vinet, P. Poulin, M. E. Cates, and J. Bibette (2011). How to prepare and stabilize very small nanoemulsions. Langmuir 27(5): 1683–1692. Dickinson, E. (2003). Hydrocolloids at interfaces and the influence on the properties of dispersed systems. Food Hydrocolloids 17(1): 25–39. Donsi, F., M. Sessa, and G. Ferrari (2012). Effect of emulsifier type and disruption chamber geometry on the fabrication of food nanoemulsions by high pressure homogenization. Industrial & Engineering Chemistry Research 51(22): 7606–7618. Dubbelboer, A., J. Janssen, H. Hoogland, A. Mudaliar, S. Maindarkar, E. Zondervan, and J. Meuldijk (2014). Population balances combined with computational fluid dynamics: A modeling approach for dispersive mixing in a high pressure homogenizer. Chemical Engineering Science 117: 376–388. Elwell, M. W., R. F. Roberts, and J. N. Coupland (2004). Effect of homogenization and surfactant type on the exchange of oil between emulsion droplets. Food Hydrocolloids 18(3): 413–418. Evans, E. D. and W. Wennerstrom (1999). The Colloidal Domain: Where Physics, Chemistry and Biology Meet. New York: Wiley-VCH. Fischer, P. and P. Erni (2007). Emulsion drops in external flow fields—The role of liquid interfaces. Current Opinion in Colloid & Interface Science 12(4–5): 196–205. Floury, J., J. Legrand, and A. Desrumaux (2004). Analysis of a new type of high pressure homogeniser. Part B. Study of droplet break-up and recoalescence phenomena. Chemical Engineering Science 59(6): 1285–1294. Freudig, B., S. Tesch, and H. Schubert (2003). Production of emulsions in high-pressure Hhmogenizers— Part II: Influence of cavitation on droplet breakup. Engineering in Life Sciences 3(6): 266–270. Gopal, E. S. R. (1968). Principles of emulsion formation. In Emulsion Science, P. Sherman, ed., pp. 1–47. London, U.K.: Academic Press. Hakansson, A., L. Fuchs, F. Innings, J. Revstedt, B. Bergenstahl, and C. Tragardh (2010). Visual observations and acoustic measurements of cavitation in an experimental model of a high-pressure homogenizer. Journal of Food Engineering 100(3): 504–513. Hakansson, A., L. Fuchs, F. Innings, J. Revstedt, C. Tragardh, and B. Bergenstahl (2011). On flow-fields in a high pressure homogenizer and its implication on drop fragmentation. 11th International Congress on Engineering and Food (Icef11) 1: 1353–1358. Hakansson, A. and M. J. Hounslow (2013). Simultaneous determination of fragmentation and coalescence rates during pilot-scale high-pressure homogenization. Journal of Food Engineering 116(1): 7–13. Hakansson, A., F. Innings, J. Revstedt, C. Tragardh, and B. Bergenstahl (2012). Estimation of turbulent fragmenting forces in a high-pressure homogenizer from computational fluid dynamics. Chemical Engineering Science 75: 309–317. Hakansson, A., F. Innings, C. Tragardh, and B. Bergenstahl (2013). A high-pressure homogenization emulsification model-Improved emulsifier transport and hydrodynamic coupling. Chemical Engineering Science 91: 44–53. 286 Food Emulsions: Principles, Practices, and Techniques Hakansson, A., C. Tragardh, and B. Bergenstahl (2009). Dynamic simulation of emulsion formation in a high pressure homogenizer. Chemical Engineering Science 64(12): 2915–2925. Hakansson, A., C. Tragardh and B. Bergenstahl (2012). A method for estimating effective coalescence rates during emulsification from oil transfer experiments. Journal of Colloid and Interface Science 374: 25–33. Hall, S., M. Cooke, A. W. Pacek, A. J. Kowalski, and D. Rothman (2011). Scaling up of silverson rotor-stator mixers. Canadian Journal of Chemical Engineering 89(5): 1040–1050. Ichikawa, T., K. Itoh, S. Yamamoto, and M. Sumita (2004). Rapid demulsification of dense oil-in-water emulsion by low external electric field—I. Experimental evidence. Colloids and Surfaces A: Physicochemical and Engineering Aspects 242(1–3): 21–26. Iqbal, S., M. K. Baloch, G. Hameed, and D. J. McClements (2013). Controlling W/O/W multiple emulsion microstructure by osmotic swelling and internal protein gelation. Food Research International 54(2): 1613–1620. Jafari, S. M., E. Assadpoor, Y. H. He, and B. Bhandari (2008). Re-coalescence of emulsion droplets during high-energy emulsification. Food Hydrocolloids 22(7): 1191–1202. Jafari, S. M., Y. He, and B. Bhandari (2007). Production of sub-micron emulsions by ultrasound and microfluidization techniques. Journal of Food Engineering 82(4): 478–488. Karbstein, H. and H. Schubert (1995a). Developments in the continuous mechanical production of oil-in-water macro-emulsions. Chemical Engineering and Processing 34(3): 205–211. Karbstein, H. and H. Schubert (1995b). Parameters influencing the selection of a machine for producing finely dispersed oil-in-water emulsions. Chemie Ingenieur Technik 67(5): 616–619. Kentish, S., T. J. Wooster, A. Ashokkumar, S. Balachandran, R. Mawson, and L. Simons (2008). The use of ultrasonics for nanoemulsion preparation. Innovative Food Science & Emerging Technologies 9(2): 170–175. Komaiko, J. and D. J. McClements (2015). Low-energy formation of edible nanoemulsions by spontaneous emulsification: Factors influencing particle size. Journal of Food Engineering 146: 122–128. Lee, L. and I. T. Norton (2013). Comparing droplet breakup for a high-pressure valve homogeniser and a microfluidizer for the potential production of food-grade nanoemulsions. Journal of Food Engineering 114(2): 158–163. Lee, L. L., N. Niknafs, R. D. Hancocks and I. T. Norton (2013). Emulsification: Mechanistic understanding. Trends in Food Science & Technology 31(1): 72–78. Leong, T., M. Ashokkumar, and S. Kentish (2011). The fundamentals of power ultrasound—A review. Acoustics Australia 39(2): 54–63. Leong, T. S. H., T. J. Wooster, S. E. Kentish, and M. Ashokkumar (2009). Minimising oil droplet size using ultrasonic emulsification. Ultrasonics Sonochemistry 16(6): 721–727. Lobo, L. and A. Svereika (2003). Coalescence during emulsification 2. Role of small molecule surfactants. Journal of Colloid and Interface Science 261(2): 498–507. Lobo, L., A. Svereika, and M. Nair (2002). Coalescence during emulsification—1. Method development. Journal of Colloid and Interface Science 253(2): 409–418. Lucassen-Reynders, E. H. and K. A. Kuijpers (1992). The role of interfacial properties in emulsification. Colloids and Surfaces 65(2–3): 175–184. Maindarkar, S. N., P. Bongers, and M. A. Henson (2013). Predicting the effects of surfactant coverage on drop size distributions of homogenized emulsions. Chemical Engineering Science 89: 102–114. Maindarkar, S. N., N. B. Raikar, P. Bongers, and M. A. Henson (2012). Incorporating emulsion drop coalescence into population balance equation models of high pressure homogenization. Colloids and Surfaces a-Physicochemical and Engineering Aspects 396: 63–73. McClements, D. J. (2004). Protein-stabilized emulsions. Current Opinion in Colloid & Interface Science 9(5): 305–313. McClements, D. J. and J. Rao (2011). Food-grade nanoemulsions: Formulation, fabrication, properties, performance, biological fate, and potential toxicity. Critical Reviews in Food Science and Nutrition 51(4): 285–330. Menon, W. B. and D. T. Wasan (1985). Demulsifcation. Encyclopedia of Emulsion Technology, P. Becher, ed., vol. 2, pp. 1–76. New York: Marcel Dekker. Mohan, S. and G. Narsimhan (1997). Coalescence of protein-stabilized emulsions in a high-pressure homogenizer. Journal of Colloid and Interface Science 192(1): 1–15. Emulsion Formation 287 Narsimhan, G. and P. Goel (2001). Drop coalescence during emulsion formation in a high-pressure homogenizer for tetradecane-in-water emulsion stabilized by sodium dodecyl sulfate. Journal of Colloid and Interface Science 238(2): 420–432. Nazir, A., K. Schroen, and R. Boom (2010). Premix emulsification: A review. Journal of Membrane Science 362(1–2): 1–11. Neethirajan, S., I. Kobayashi, M. Nakajima, D. Wu, S. Nandagopal, and F. Lin (2011). Microfluidics for food, agriculture and biosystems industries. Lab on a Chip 11(9): 1574–1586. Nisisako, T. (2008). Microstructured devices for preparing controlled multiple emulsions. Chemical Engineering & Technology 31(8): 1091–1098. O’Keefe, S. F. and O. A. Pike (2010). Fat characterization. In Food Analysis, S. S. Nielsen, ed., pp. 239–260. New York: Springer. Ostertag, F., J. Weiss, and D. J. McClements (2012). Low-energy formation of edible nanoemulsions: Factors influencing droplet size produced by emulsion phase inversion. Journal of Colloid and Interface Science 388: 95–102. Panagiotou, T. and R. J. Fisher (2008). Form nanoparticles via controlled crystallization. Chemical Engineering Progress 104(10): 33–39. Panagiotou, T., S. V. Mesite, J. M. Bernard, K. J. Chomistek, and R. J. Fisher (2008). Production of Polymer Nanosuspensions Using Microfluidizer (R) Processor Based Technologies. Newton, MA : Microfluidics Corporation. Panagiotou, T., S. V. Mesite, and R. J. Fisher (2009). Production of norfloxacin nanosuspensions using microfluidics rteaction technology through solvent/antisolvent crystallization. Industrial & Engineering Chemistry Research 48(4): 1761–1771. Patist, A. and D. Bates (2008). Ultrasonic innovations in the food industry: From the laboratory to commercial production. Innovative Food Science & Emerging Technologies 9(2): 147–154. Phipps, L. W. (1985). The High Pressure Dairy Homogenizer. Reading, U.K., The National Institute for Research in Dairying. Pingret, D., A. S. Fabiano-Tixier, and F. Chemat (2013). Degradation during application of ultrasound in food processing: A review. Food Control 31(2): 593–606. Qian, C. and D. J. McClements (2011). Formation of nanoemulsions stabilized by model food-grade emulsifiers using high-pressure homogenization: Factors affecting particle size. Food Hydrocolloids 25(5): 1000–1008. Rampon, V., A. Riaublanc, M. Anton, C. Genot, and D. J. Mc Clements (2003). Evidence that homogenization of BSA-stabilized hexadecane-in-water emulsions induces structure modification of the nonadsorbed protein. Journal of Agricultural and Food Chemistry 51(20): 5900–5905. Rao, J. J. and D. J. McClements (2011). Food-grade microemulsions, nanoemulsions and emulsions: Fabrication from sucrose monopalmitate & lemon oil. Food Hydrocolloids 25(6): 1413–1423. Roger, K., B. Cabane, and U. Olsson (2010). Formation of 10–100 nm size-controlled emulsions through a sub-PIT cycle. Langmuir 26(6): 3860–3867. Rosenthal, A., D. L. Pyle, and K. Niranjan (1996). Aqueous and enzymatic processes for edible oil extraction. Enzyme and Microbial Technology 19(6): 402–420. Saberi, A. H., Y. Fang, and D. J. McClements (2013a). Effect of glycerol on formation, stability, and properties of vitamin-E enriched nanoemulsions produced using spontaneous emulsification. Journal of Colloid and Interface Science 411: 105–113. Saberi, A. H., Y. Fang, and D. J. McClements (2013b). Fabrication of vitamin E-enriched nanoemulsions by spontaneous emulsification: Effect of propylene glycol and ethanol on formation, stability, and properties. Food Research International 54(1): 812–820. Saberi, A. H., Y. Fang, and D. J. McClements (2013c). Fabrication of vitamin E-enriched nanoemulsions: Factors affecting particle size using spontaneous emulsification. Journal of Colloid and Interface Science 391: 95–102. Saberi, A. H., Y. Fang and D. J. McClements (2014). Stabilization of vitamin E-enriched mini-emulsions: Influence of organic and aqueous phase compositions. Colloids and Surfaces a-Physicochemical and Engineering Aspects 449: 65–73. Sajjadi, B., A. A. A. Raman, R. Shah, and S. Ibrahim (2013). Review on applicable breakup/coalescence models in turbulent liquid-liquid flows. Reviews in Chemical Engineering 29(3): 131–158. 288 Food Emulsions: Principles, Practices, and Techniques Santana, R. C., F. A. Perrechil, and R. L. Cunha (2013). High- and low-energy emulsifications for food applications: A focus on process parameters. Food Engineering Reviews 5(2): 107–122. Schubert, H. (1997). Advances in the mechanical production of food emulsions. In Engineering and Food, R. Jowitt, ed., pp. 82–102, Sheffield, U.K.: Sheffield Academic Press. Schubert, H., K. Ax, and O. Behrend (2003). Product engineering of dispersed systems. Trends in Food Science & Technology 14(1–2): 9–16. Schubert, H. and R. Engel (2004). Product and formulation engineering of emulsions. Chemical Engineering Research & Design 82(A9): 1137–1143. Schultz, S., G. Wagner, K. Urban, and J. Ulrich (2004). High-pressure homogenization as a process for emulsion formation. Chemical Engineering & Technology 27(4): 361–368. Seekkuarachchi, I. N., K. Tanaka, and H. Kumazawa (2006). Formation and charaterization of submicrometer oil-in-water (O/W) emulsions, using high-energy emulsification. Industrial & Engineering Chemistry Research 45(1): 372–390. Singh, R. P. and D. R. Heldman (2013). Introduction to Food Engineering. London, U.K.: Academic Press. Solans, C. and I. Sole (2012). Nano-emulsions: Formation by low-energy methods. Current Opinion in Colloid & Interface Science 17(5): 246–254. Stang, M., H. Karbstein, and H. Schubert (1994). Adsorption-kinetics of emulsifiers at oil-water interfaces and their effect on mechanical emulsification. Chemical Engineering and Processing 33(5): 307–311. Stang, M., H. Karbstein, and H. Schubert (2001). Emulsification in high pressure homogenizers. Engineering in Life Sciences 1: 151–162. Stone, H. A. (1994). Dynamics of drop deformation and breakup in viscous fluids. Annual Review of Fluid Mechanics 26: 65–102. Taisne, L., P. Walstra, and B. Cabane (1996). Transfer of oil between emulsion droplets. Journal of Colloid and Interface Science 184(2): 378–390. Tang, S. Y., P. Shridharan, and M. Sivakumar (2013). Impact of process parameters in the generation of novel aspirin nanoemulsions: Comparative studies between ultrasound cavitation and microfluidizer. Ultrasonics Sonochemistry 20(1): 485–497. Tcholakova, S., N. D. Denkov, I. B. Ivanov, and B. Campbell (2006). Coalescence stability of emulsions containing globular milk proteins. Advances in Colloid and Interface Science 123: 259–293. Tesch, S., B. Freudig, and H. Schubert (2003). Production of emulsions in high-pressure homogenizers—Part I: Disruption and stabilization of droplets. Chemical Engineering & Technology 26(5): 569–573. Urban, K., G. Wagner, D. Schaffner, D. Roglin, and J. Ulrich (2006). Rotor-stator and disc systems for emulsification processes. Chemical Engineering & Technology 29(1): 24–31. Vladisavljevic, G. T., I. Kobayashi, and M. Nakajima (2012). Production of uniform droplets using membrane, microchannel and microfluidic emulsification devices. Microfluidics and Nanofluidics 13(1): 151–178. Vladisavljevic, G. T. and R. A. Williams (2005). Recent developments in manufacturing emulsions and particulate products using membranes. Advances in Colloid and Interface Science 113(1): 1–20. Walstra, P. (1983). Formation of emulsions. In Encyclopedia of Emulsion Technology, P. Becher, ed., vol. 4, 57–128. New York: Marcel Dekker. Walstra, P. (1993). Principles of emulsion formation. Chemical Engineering Science 48(2): 333–349. Walstra, P. (2003). Physical Chemistry of Foods. New York, Marcel Decker. Walstra, P. and P. E. A. Smulder (1988). Emulsion formation. In Modern Aspects of Emulsion Science, B. P. Binks, ed., pp. 56–99. Cambridge, U.K.: The Royal Society of Chemistry. Williams, A., J. J. M. Janssen, and A. Prins (1997). Behaviour of droplets in simple shear flow in the presence of a protein emulsifier. Colloids and Surfaces a-Physicochemical and Engineering Aspects 125(2–3): 189–200. Windhab, E. J., M. Dressler, K. Feigl, P. Fischer, and D. Megias-Alguacil (2005). Emulsion processing—From single-drop deformation to design of complex processes and products. Chemical Engineering Science 60(8–9): 2101–2113. Wooster, T., M. Golding, and P. Sanguansri (2008). Impact of oil type on nanoemulsion formation and Ostwald ripening stability. Langmuir 24(22): 12758–12765. Zeeb, B., E. Herz, D. J. McClements, and J. Weiss (2014). Impact of alcohols on the formation and stability of protein-stabilized nanoemulsions. Journal of Colloid and Interface Science 433: 196–203. 7 Emulsion Stability 7.1 Introduction The term “emulsion stability” refers to the ability of an emulsion to resist changes in its properties over time: the more stable the emulsion, the slower its properties change. An emulsion may become unstable due to a number of different types of physical and chemical processes.* Physical instability results in an alteration in the spatial distribution or structural organization of the molecules, whereas chemical instability results in an alteration in the kind of molecules present. Creaming, flocculation, coalescence, partial coalescence, phase inversion, and Ostwald ripening are all examples of physical instability, whereas oxidation and hydrolysis are common examples of chemical instability in emulsions. The development of an effective strategy to prevent undesirable changes in the properties of a particular food emulsion depends on the dominant physicochemical mechanism(s) responsible for the changes. In practice, two or more of these mechanisms may operate in concert. For example, the oil droplets in an oil-in-water emulsion may flocculate first, which then accelerates the rate of phase separation due to creaming. It is therefore important for food scientists to identify the relative importance of each mechanism, the relationship between them, and the factors that influence them, so that effective means of controlling the stability and physicochemical properties of emulsions can be established. The length of time that an emulsion must remain stable depends on the nature of the food product. Some emulsions are formed as intermediate steps during a food manufacturing process, and therefore only need to remain stable for a few seconds, minutes, or hours, for example, cake batter, ice cream mix, and margarine premix. Conversely, other food emulsions must remain stable for days, months, or even years prior to consumption, for example, dressings, sauces, dips, beverages, and cream liqueurs. It should be noted that, the production of certain foods actually involves a controlled destabilization of an emulsion during the manufacturing process, for example, margarine, butter, whipped cream, and ice cream (Chapter 12). One of the major objectives of emulsion scientists working in the food industry is to establish the specific factors that determine the stability of each particular type of food emulsion, as well as to elucidate general principles that can be used to predict the behavior of food products or processes. In practice, it is very difficult to quantitatively predict the stability of food emulsions from first principles because of their compositional and structural complexity. Nevertheless, an appreciation of the origin and nature of the various destabilization mechanisms is still an invaluable tool for controlling and improving emulsion stability. Because of the difficulties in theoretically predicting emulsion stability, food scientists often rely on the use of analytical techniques to experimentally monitor changes in emulsion properties over time. By using a combination of theoretical understanding and experimental measurements food manufacturers are able to predict the influence of different ingredients, processing operations, and storage conditions on the stability and properties of food emulsions. The rate at which an emulsion breaks down, and the mechanism by which this process occurs, depends on its composition and microstructure, as well as on the environmental conditions it experiences, for * It should be noted that the properties of emulsions may also change with time due to microbiological changes, for example, the growth of specific types of bacteria or mold. 289 290 Food Emulsions: Principles, Practices, and Techniques example, temperature variations, mechanical agitation, and storage conditions. In this chapter, we examine the physicochemical basis of each of the major destabilization mechanisms responsible for changes in food emulsion properties, as well as discussing the major factors that influence them, methods of controlling them, and experimental techniques for monitoring them. This information is particularly useful for food manufacturers who need to formulate emulsions with enhanced shelf-life or to promote emulsion instability in a controlled fashion. 7.2 Thermodynamic and Kinetic Stability of Emulsions When we consider the “stability” of an emulsion, it is extremely important to distinguish between its thermodynamic stability and its kinetic stability (Dickinson 1992). Thermodynamics tells us whether a given process will occur or not, whereas kinetics tells us the rate at which it will proceed if it does occur (Atkins and de Paula 2014). All food emulsions are thermodynamically unstable systems that will eventually break down if they are left long enough. For this reason, it is differences in kinetic stability that are largely responsible for the diverse range of physicochemical and sensory properties exhibited by different food emulsions. 7.2.1 Thermodynamic Stability The thermodynamic instability of emulsions can be easily demonstrated practically by agitating a sealed vessel containing pure oil and pure water, and then observing the change in the appearance of the system over time. The optically opaque emulsion that is initially formed by agitation breaks down over time until a layer of oil is observed on top of a layer of water (Figure 6.2). The origin of this thermodynamic instability can be illustrated by comparing the free energy of a system consisting of oil and water before and after emulsification (Hunter 1989). To simplify this analysis, we initially assume that the oil and water have similar densities so that no creaming or sedimentation occurs. As a consequence, the final state consists of a single large droplet suspended in the continuous phase (Figure 7.1), rather than a layer of oil on top of a layer of water (Figure 6.2). In its initial state, prior to emulsification, the free energy is given by: i i G i = GOi + GW + GIi - TSconfig (7.1) Formation Increase in interfacial area Destabilization Separated phases Emulsion FIGURE 7.1 The formation of an emulsion is thermodynamically unfavorable because of the increase in surface area between the oil and water phases. In this diagram, it is assumed that the oil forms a single droplet in the phase separated state, but in reality the oil will usually form a layer at the top due to gravity. 291 Emulsion Stability and in its final state, after emulsification, it is given by: f f G f = GOf + GW + GIf - TSconfig (7.2) where G O, GW, and GI are the free energies of the oil phase, water phase, and the oil–water interface, respectively T is the absolute temperature S is the configurational entropy of the droplets in the system The superscripts i and f refer to the initial and final states of the system The free energies of the bulk oil and water phases remain constant before and after homogenization: i f and so the difference in free energy between the initial and final states is given by: GOi = GOf and GW = GW ( ) f i DGformation = G f - G i = GIf - GIi - TSconfig - TSconfig = DGI - TDSconfig (7.3) By definition the difference in interfacial free energy between the initial and final states (ΔG I) is equal to the increase in contact area between the oil and water phases (ΔA) multiplied by the interfacial tension (γ): ΔG I = γΔA. Hence, DGformation = gDA - TDSconfig (7.4) The change in interfacial free energy (γΔA) is always positive, because the contact area increases after homogenization, and therefore it opposes emulsion formation. On the other hand, the configurational entropy term (−TΔSconfig) is always negative, because the number of arrangements accessible to the droplets in the emulsified state is much greater than in the nonemulsified state, and therefore it favors emulsion formation. An expression for the configurational entropy can be derived from a statistical analysis of the number of configurations emulsion droplets can adopt in the initial and final states: DSconfig = - nk (f ln f + (1 - f)ln(1 - f)) f (7.5) where k is Boltzmann’s constant n is the number of droplets in the emulsion ϕ is the disperse phase volume fraction In most food emulsions, the configurational entropy is much smaller than the interfacial free energy and can be ignored. As an example, consider a 10 vol% oil-in-water emulsion containing droplets with a radius of 1 μm and interfacial tension (γ) of 0.01 N m−1. The interfacial free energy term (γΔA) is about 3 kJ m−3 of emulsion, whereas the configurational entropy (TΔS) term is about 3 × 10 −7 kJ m−3. Even for the very small droplets found in some nanoemulsions (r = 25 nm), the configurational entropy term (0.02 kJ m−3) is still much smaller than the interfacial free energy term (120 kJ m−3). The overall free energy change associated with the creation of a food emulsion can therefore be represented by the following simple expression: ΔG formation = γΔA (7.6) Thus, the formation of a food emulsion is always thermodynamically unfavorable, because of the increase in interfacial area after emulsification. It should be noted that the configurational entropy term may dominate the interfacial free energy term in emulsions in which the interfacial tension is extremely 292 Food Emulsions: Principles, Practices, and Techniques small, and that these systems are therefore thermodynamically stable (Hunter 1989). This type of thermodynamically stable system is usually referred to as a microemulsion, to distinguish it from thermodynamically unstable (macro)emulsions. In practice, the oil and water phases normally have different densities, and so it is necessary to include a free energy term that accounts for gravitational effects, that is, the tendency for the liquid with the lowest density to move to the top of the emulsion. This term contributes to the thermodynamic instability of emulsions and accounts for the observed creaming or sedimentation of droplets (Section 7.3). 7.2.2 Kinetic Stability The free energy change associated with emulsion formation determines whether an emulsion is thermodynamically stable or not, but it does not give any indication of the rate at which the properties of an emulsion change over time, the type of changes that occur, or the physical mechanism(s) responsible for these changes. Information about the time dependence of emulsion stability is particularly important to food scientists who need to create food products that retain their desirable properties for a sufficiently long time under a variety of environmental conditions. For this reason, food scientists are usually more interested in the kinetic stability of emulsions than in their thermodynamic stability. The importance of kinetic effects can be highlighted by comparing the long-term stability of emulsions with the same composition but with different droplet sizes. An emulsion that contains small droplets usually has a longer shelf-life (greater kinetic stability) than one that contains large droplets, even though it is more thermodynamically unstable (because it has a larger interfacial area, ΔA). Despite the fact that food emulsions exist in a thermodynamically unstable state, many of them remain kinetically stable (metastable) for months or even years. What is the origin of this kinetic stability? Conceptually, the kinetic stability of an emulsion can be attributed to an activation energy (ΔG*) that must be overcome before the emulsion can reach its most thermodynamically favorable state (Figure 7.2). Kinetically stable ΔG* Gf Emulsion Thermodynamically unfavorable Kinetically unstable ΔG Gi Separated phases Thermodynamically favorable FIGURE 7.2 Emulsions are thermodynamically unstable systems, but may exist in a metastable state, and therefore be kinetically stable. The kinetic stability depends on the height of the activation energy (ΔG*) between the emulsion and the separated states. 293 Emulsion Stability An emulsion that is kinetically stable has to have an activation energy that is significantly greater than the thermal energy of the system (kT). For most emulsions, an activation energy of about 20 kT is sufficient to provide long-term stability. In reality, emulsions have a number of different metastable states, and each of these has its own activation energy. Thus, an emulsion may move from one metastable state to another before finally reaching the most thermodynamically stable state. A change from one of these metastable states to another may be sufficient to have a deleterious effect on food quality. The kinetic stability of emulsions can only be understood with reference to their dynamic nature. The droplets in an emulsion are in a continual state of motion and frequently collide into one another due to their Brownian motion, gravity, or applied external forces. Whether droplets move apart, remain loosely associated with each other, or fuse together after a collision depends on the nature of the interactions between them. The kinetic stability of emulsions is therefore largely determined by the dynamics and interactions of the droplets they contain. Consequently, a great deal of this chapter is concerned with the nature of the interactions between droplets and the factors that determine droplet movement in emulsions. Earlier it was mentioned that if pure oil and pure water are agitated together a temporary emulsion is formed that rapidly reverts back to its individual components. This is because there is a very low activation energy between the emulsified and separated states in the absence of a suitable stabilizer. To create an emulsion that is kinetically stable for a reasonably long period of time, it is necessary to have either an emulsifier or a texture modifier present that produces an activation energy that is sufficiently large to prevent instability. An emulsifier adsorbs to the surface of freshly formed droplets and forms a protective coating that prevents them from merging together, while a texture modifier increases the viscosity of the continuous phase or forms a gel so that droplets collide less frequently with one another (Chapter 4). The role of emulsifiers and texture modifiers on emulsion stability will therefore be another common theme of this chapter. 7.3 Gravitational Separation Gravitational separation is one of the most common forms of instability in food emulsions (Robins 2000, Robins et al. 2002). In general, the droplets in an emulsion have a different density to that of the liquid that surrounds them, and so a net gravitational force acts upon them. If the droplets have a lower density than the surrounding liquid, they have a tendency to move upward, which is referred to as creaming (Figure 7.3). Conversely, if they have a higher density than the surrounding liquid they tend to move Fg Ff Creaming FIGURE 7.3 Food emulsions are prone to creaming because of the density difference between the oil and water phases. Inset: the forces acting on an emulsion droplet. 294 Food Emulsions: Principles, Practices, and Techniques downward, which is referred to as sedimentation. The densities of most edible oils (in their liquid state) are lower than that of water, and so there is a tendency for oil to accumulate at the top of an emulsion and water at the bottom. Thus, droplets in an oil-in-water emulsion tend to cream, whereas those in a waterin-oil emulsion tend to sediment. Gravitational separation is usually regarded as having an adverse effect on the quality of food emulsions. A consumer expects to see a product that appears homogeneous and therefore the separation of an emulsion into an optically opaque droplet-rich layer and a less opaque droplet-depleted layer is undesirable. The textural attributes of a product are also adversely affected by gravitational separation, because the droplet-rich layer tends to be more viscous than expected, whereas the droplet-depleted layer tends to be less viscous. The taste and mouthfeel of a portion of food therefore depend on the location from which it was taken from the emulsion. A sample selected from the top of an oil-in-water emulsion that has undergone creaming will seem too “rich” because of the high fat content, whereas a sample selected from the bottom will seem too “watery” because of the low fat content. Gravitational separation is also a problem because it causes droplets to come into close contact for extended periods, which can lead to enhanced flocculation or coalescence, and eventually to oiling off—the formation of a layer of pure oil on top of the emulsion. When a food manufacturer is designing an emulsion-based product, it is therefore important to control the rate at which gravitational separation occurs. Each food product is unique, containing different types of ingredients and experiencing different environmental conditions during its processing, storage, and consumption. As a consequence, the optimum method of controlling gravitational separation varies from product to product. In this section, we consider the most important factors that influence gravitational separation, as well as strategies for controlling it. 7.3.1 Physical Basis of Gravitational Separation 7.3.1.1 Stokes’ Law The rate at which an isolated rigid spherical particle creams in an ideal liquid is determined by the balance of forces that acts upon it (Figure 7.3 inset). When a particle has a lower density than the surrounding liquid an upward gravitational force acts upon it (Hiemenz and Rajagopalan 1997): 4 Fg = - pr 3 (r2 - r1 )g 3 (7.7) where r is the radius of the particle g is the acceleration due to gravity ρ is the density The subscripts 1 and 2 refer to the continuous and dispersed phases, respectively As the particle moves upward through the surrounding liquid, it experiences a hydrodynamic frictional force that acts in the opposite direction and therefore retards its motion: Ff = 6ph1rv (7.8) Here v is the creaming velocity η is the shear viscosity The particle rapidly reaches a constant velocity, where the upward force due to gravity balances the downward force due to friction, that is, Fg = Ff. By combining Equations 7.7 and 7.8, we obtain Stokes’ law equation for the creaming rate of an isolated spherical particle in a liquid: 295 Emulsion Stability vStokes = - 2 gr 2 (r2 - r1 ) 9h1 (7.9) The sign of vStokes determines whether the droplet moves upward (+) due to creaming or downward (−) due to sedimentation. To a first approximation, the stability of a food emulsion to creaming can be estimated using Equation 7.9. For example, an oil droplet (ρ2 = 910 kg m−3) with a radius of 1 μm suspended in water (η1 = 1 mPa s, ρ1 = 1000 kg m−3) will cream at a rate of about 17 mm day−1. Thus, one would not expect an emulsion containing droplets of this size to have a particularly long shelf-life. As a useful rule of thumb, an emulsion in which the calculated creaming rate is less than about 0.5 mm day−1 can be considered to be stable toward creaming. This corresponds to droplets with radii less than about 0.15 μm for a typical triglyceride oil dispersed in water. In the rest of this section, we mainly consider creaming, rather than sedimentation, because it is more common in food systems. Nevertheless, the same physical principles occur in both cases, and the methods of controlling them are similar. In the initial stages of creaming, the droplets move upward and a droplet-depleted layer is observed at the bottom of the container (Figure 7.4). When the droplets reach the top of the emulsion they cannot move upward any further and so they pack together to form a “creamed layer.” The final thickness of the creamed layer depends on the initial droplet concentration in the emulsion and the effectiveness of the droplet packing. Droplets may pack tightly or loosely depending on their polydispersity and the nature of the interactions between them. Tightly packed droplets form relatively thin creamed layers, whereas loosely packed droplets form relatively thick creamed layers. Many of the factors that determine the packing of droplets in a creamed layer, also determine the structure of flocs (see Section 7.4). For example, a loosely packed layer occurs when the attractive interactions between the droplets in the cream layer are so strong that they do not allow them to rearrange themselves once they have come into contact with their neighbors. The droplets in a creamed emulsion can often be redispersed by mild agitation, providing that they are not too strongly attracted to each other or that coalescence has not occurred. 7.3.1.2 Deviations from Stokes’ Law Stokes’ law can only be strictly used to calculate the velocity of an isolated rigid spherical particle suspended in an ideal liquid of infinite extent. In practice, there are often large deviations between the creaming velocity predicted by Stokes’ law and experimental measurements of creaming in food emulsions, because many of the assumptions used in deriving Equation 7.9 are invalid. Some of the most important factors that alter the creaming rate in food emulsions are considered below: (a) (b) (c) (d) (e) (f ) (g) (h) FIGURE 7.4 Schematic representation of the time dependence of droplet creaming in oil-in-water emulsions. Droplets move upward until they cannot move any further and then form a “creamed” layer. Larger droplets tend to move upward faster than smaller ones. The diagram (a–g) shows a progressive change in an emulsion from a stable system and (h) a completely separated system due to creaming. 296 Food Emulsions: Principles, Practices, and Techniques 7.3.1.2.1 Droplet Fluidity Stokes’ equation assumes that there is no slip at the interface between the droplet and the surrounding fluid, which is only strictly true for solid particles. The liquid within a droplet can move when a force is applied to the droplet’s surface, thus the frictional force that opposes the movement of a droplet is reduced, which causes an increase in the creaming velocity (Dickinson and Stainsby 1982): v = vStokes 3(h2 + h1 ) (3h2 + 2h1 ) (7.10) This expression reduces to Stokes’ equation when the viscosity of the droplet is much greater than that of the continuous phase (η2 ≫ η1). Conversely, when the viscosity of the droplet is much less than that of the continuous phase (η2 ≪ η1), the creaming rate is 1.5 times faster than that predicted by Equation 7.9. In practice, the droplets in most food emulsions can be considered to act like rigid spheres because they are surrounded by a viscoelastic interfacial layer that prevents the fluid within them from moving (Walstra 2003). 7.3.1.2.2 Nondilute Systems The creaming velocity of droplets in concentrated emulsions is less than that in dilute emulsions because of hydrodynamic interactions between the droplets (Hunter 1989, Chanamai and McClements 2000). As an emulsion droplet moves upward due to gravity, an equal volume of continuous phase must move downward to compensate. Thus, there is a net flow of continuous phase downward, which opposes the upward movement of the droplets, and therefore decreases the creaming velocity. In fairly dilute emulsions (i.e., <2% droplets) the creaming velocity of spherical particles has been derived mathematically from a consideration of fluid hydrodynamics: v = vStokes (1 - 6.55f) (7.11) In more concentrated emulsions, a number of other types of hydrodynamic interaction also reduce the creaming velocity of the droplets. These hydrodynamic effects can be partly accounted for by using a value for the viscosity of a concentrated emulsion in Equation 7.9, rather than that of the continuous phase (see Chapter 8). A semiempirical equation that has been found to give relatively good predictions of the creaming behavior of concentrated emulsions has been developed: æ fö v = vStokes ç 1 - ÷ f c ø è kfc (7.12) Here, ϕ c and k are parameters that depend on the nature of the emulsion. Normally, ϕ c is taken to be the volume fraction at which the spherical particles become close packed. For monodisperse latex suspensions the following values have been determined experimentally: k = 5.4 and ϕ c = 0.585 (Hunter 1989). Nevertheless, different values have been found to be appropriate for quasi-monodisperse oilin-water emulsions stabilized by an anionic surfactant, that is, k = 8 and ϕ c = 0.585 (Chanamai and McClements 2000). The k and ϕ c parameters can therefore be thought of as semiempirical quantities, whose precise value for a particular system needs to be determined experimentally. These values will depend on the polydispersity, shape, and interactions of the particles in an emulsion. The above equation predicts that the creaming velocity decreases as the droplet concentration increases, until creaming is completely inhibited once a critical disperse phase volume fraction (ϕ c) has been exceeded (Figure 7.5). In general, the value of ϕc used in the above equation depends on the packing of the droplets within an emulsion, which is governed by their polydispersity, shape, and colloidal interactions. Polydisperse droplets are able to fill the available space more effectively than monodisperse droplets because the small 297 Emulsion Stability 1 ν/ν Stokes 0.8 Dilute 0.6 0.4 0.2 0 0 10 20 φ (%) 30 40 Concentrated FIGURE 7.5 Comparison of measured (points) and predicted (line) relative creaming rates as a function of disperse phase volume fraction for quasi-monodisperse oil-in-water emulsions. Creaming is effectively retarded when the emulsion droplets become close packed. droplets can fit into the gaps between the larger ones, hence ϕc tends to be higher for polydisperse than for monodisperse emulsions (Das and Ghosh 1990, Cheetangdee and Fukada 2012). When the droplets are strongly attracted to each other they can form a particle gel network at relatively low droplet concentrations, which severely restricts droplet movement at lower ϕ values (Figure 7.6). On the other hand, when the droplets are strongly repelled from each other their effective size increases, which also severely restricts droplet movement at lower ϕ values. In the case of strong droplet attraction or repulsion, the above equation must therefore be modified (see Sections 7.3.1.2.4 and 7.3.1.2.8). 7.3.1.2.3 Polydispersity Food emulsions normally contain a range of different droplet sizes and the larger droplets tend to cream more rapidly than the smaller droplets (Figure 7.7), so that there is a distribution of creaming rates within (a) (b) FIGURE 7.6 Creaming is prevented when the disperse phase volume fraction exceeds a critical value (ϕc) where the droplets become so closely packed that they cannot easily move past each other. This critical value occurs at lower disperse phase volume fractions when the droplets are flocculated. (a) Close packing (high ϕ c) and (b) open packing (low ϕc). 298 Food Emulsions: Principles, Practices, and Techniques 3000 ν (mm day–1) 2500 v 2000 r2 1500 1000 500 0 0 2 4 6 8 10 12 14 r (µm) FIGURE 7.7 The rate at which a droplet moves upward due to creaming increases as the droplet diameter increases. The creaming rate was calculated using Stokes’ law assuming a density contrast of 80 kg m−3 and a continuous phase viscosity of 1 mPa s. an emulsion. Larger droplets tend to cream faster than smaller ones, which leads to an accumulation of larger droplets at the top of an emulsion (Figure 7.4). Thus, there is both a droplet concentration profile and a droplet size profile in the vertical direction within an emulsion. As the larger droplets move upward more rapidly, they collide with smaller droplets that are moving more slowly (Dukhin and Sjoblom 1996). If the droplets aggregate after a collision, they will cream at a faster rate than either of the isolated droplets. Detailed information about the evolution of the droplet concentration and size profiles within an emulsion can be obtained by using computer simulations that take into account polydispersity (Davis and Gecol 1994, Pinfield et al. 1997, Burger et al. 2000). For many purposes, it is only necessary to have an average creaming velocity, which can usually be estimated using a mean droplet radius (r54) in the Stokes equation (Dickinson 1992, Walstra 2003). 7.3.1.2.4 Droplet Flocculation In many food emulsions, the droplets aggregate to form flocs (Section 7.4). The size and structure of the flocs within an emulsion has a large influence on the rate at which the droplets cream. At low or intermediate droplet concentrations, where flocs do not substantially interact with one another, flocculation tends to increase the creaming velocity because the flocs have a larger effective size than the individual droplets (which more than compensates for the fact that the density difference between the flocs and the surrounding liquid is reduced). In concentrated emulsions, flocculation retards creaming because a three-dimensional network of aggregated flocs is formed that prevents the individual droplets from moving (Figure 7.6). The droplet concentration at which creaming is prevented depends primarily on the internal structure of the flocs formed. A network can form at lower disperse phase volume fractions when the droplets in a floc are more openly packed, and therefore creaming is prevented at lower droplet concentrations. Openly packed flocs tend to form when there is a strong attraction between the droplets (Section 7.5.3). Mathematical models can be derived to predict the creaming rate of a flocculated emulsion if one assumes that the flocs are spherical and do not break down during creaming (Chanamai and McClements 2000). For example, Stokes’ law can be used to predict the creaming velocity of the flocs in a dilute emulsion (where floc–floc interactions are negligible) by replacing the characteristics of the individual droplets (r, ρ2,) by those of the flocs (rfloc, ρfloc,): 299 Emulsion Stability vfloc = - 2 2 grfloc (rfloc - r1 ) 9h1 (7.13) The radius (rfloc) and density (ρfloc) of flocs depend on the number of droplets per floc (n) and the internal packing (ϕi) of the droplets within the flocs (Figure 7.8). If it is assumed that the flocs are spherical, then their internal packing is given by the expression: ϕi = Vdroplets/Vfloc, where Vdroplets is the volume of the drop3 /3). The expressions for Vdroplets lets within the floc (=n4πr 3/3) and Vfloc is the volume of the floc (= 4prfloc and Vfloc can be inserted into the expression for ϕi to give the following equation: rfloc = r × 3 n fi (7.14) where, ϕi is independent of the floc size. For random close packing of hard spheres ϕi ≈ 0.63 (Quemada and Berli 2002). This equation indicates that the floc size should increase as the number of droplets within it increases or as the packing becomes more open. The density of the flocs can also be calculated from knowledge of the packing of the droplets within them: rfloc = fir2 + (1 - fi )r1 (7.15) Flocculation increases the effective size of the particles within the emulsion (which enhances creaming), while decreasing the density contrast between these particles and the surrounding fluid (which retards creaming). The overall influence of flocculation on the creaming velocity can be conveniently characterized by a creaming instability ratio: vfloc/vStokes. For dilute emulsions, this ratio can be calculated using Stokes’ law (Equation 7.9) and the density of the flocs (Equation 7.15): vfloc r2 f = floc2 i vStokes r (7.16) As expected this equation predicts that the creaming rate should increase as the size of the flocs increases or as the droplets become more densely packed within the flocs. rfloc φi = Vdroplets/Vfloc FIGURE 7.8 A floc can be considered to consist of a spherical particle with a particular radius (rfloc) and internal packing (ϕi). 300 Food Emulsions: Principles, Practices, and Techniques In a concentrated nonflocculated suspension, the creaming velocity is reduced because of hydrodynamic interactions between the particles and can be described by the same semiempirical equation as used for concentrated nonflocculated suspensions by replacing v with vfloc and ϕ with ϕfloc (=ϕ/ϕi): æ f ö vfloc,f = vfloc ç 1 ÷ è fc f i ø kfc = vStokes 2 rfloc fi æ f ö 1÷ 2 ç r è fc f i ø kfc (7.17) Here, the values of ϕc and k have the same meanings (and approximately the same numerical values) as in nonflocculated emulsions, that is, ϕc ∼ 0.585 and k ∼ 5.4. This equation can be used to predict the influence of floc size and internal packing on their creaming velocity (Figure 7.9). In dilute emulsions (ϕ → 0), the normalized creaming velocity (vfloc,ϕ/vStokes) increases with increasing floc radius (higher rfloc) and internal packing (higher ϕi). For concentrated emulsions, the normalized creaming velocity decreases more rapidly with increasing droplet concentration as the internal packing decreases. These predictions confirm experimental observations that droplet flocculation promotes creaming in dilute emulsions because of the increase in particle size, but may retard it in concentrated emulsions because of the formation of a network of aggregated droplets the extends throughout the emulsion volume (Chanamai and McClements 2000). Experimental measurements of the creaming velocity in nonflocculated and flocculated quasi-monodisperse oil-in-water emulsions largely confirm the predictions of the above theories (Figure 7.10). 7.3.1.2.5 Non-Newtonian Rheology of Continuous Phase The continuous phase of many food emulsions is non-Newtonian, that is, the viscosity depends on shear stress or has some elastic characteristics (Chapter 8). As a consequence, it is important to consider the most appropriate apparent viscosity to use in Stokes’ equation (Walstra 2003). Biopolymers, such as modified starches or gums, are often added to oil-in-water emulsions to increase the viscosity of the aqueous phase (Section 4.5). Many of these biopolymer solutions exhibit pronounced shear-thinning behavior, having a high viscosity at low shear rates that decreases dramatically as the shear rate is increased (Chapter 8). This property is important because it means that the droplets are prevented from creaming, but that the food emulsion still flows easily when poured from a container. Creaming usually occurs when an emulsion is at rest, and therefore it is important to know the apparent viscosity that a droplet experiences as it moves through the continuous phase under these conditions. The shear stress acting on a droplet undergoing gravitational separation is τ ≈ 2Δρgr, which is typically between 10 −4 and 10 −2 Pa for food emulsions (Walstra 2003). Solutions of thickening agents have extremely high apparent shear viscosities at these relatively low shear stresses, and hence the droplets will tend to cream extremely slowly. Some aqueous biopolymer solutions have a yield stress (τB), below which the solution acts like an elastic solid, and above which it acts like a viscous fluid (van Vliet and Walstra 1989). In these systems, droplet creaming is effectively eliminated when the yield stress of the solution is larger than the stress exerted by a droplet as it moves through the continuous phase, that is, τB ≥ |2Δρgr| (Walstra 2003). Typically, a value of about 10 mPa is required to prevent emulsion droplets of a few micrometers from creaming, which is often exceeded in practice. Similar behavior is observed in water-in-oil emulsions containing a network of aggregated fat crystals when it has a sufficiently high yield stress to prevent the water droplets from undergoing sedimentation. The above discussion highlights the importance of carefully defining the rheological properties of the continuous phase. For this reason, it is good practice to measure the viscosity of the continuous phase over the range of shear rates that an emulsion droplet is likely to experience during processing, storage, and handling, which may be as wide as 10 −7–103 s−1 (Dickinson 1992). 7.3.1.2.6 Electrical Charge Charged emulsion droplets tend to move more slowly in a gravitational field than uncharged droplets for a number of reasons (Dickinson and Stainsby 1982, Hiemenz and Rajagopalan 1997). First, repulsive electrostatic interactions between similarly charged droplets mean that they cannot get as close together 301 Emulsion Stability 100 rfloc = 5 µm ν/νStokes 10 0.2 1 0.4 0.6 0.1 NF 0.01 (a) 0 10 20 φ (%) 30 40 1000 100 ν/νStokes φi = 0.6 10 µm 5 µm 10 2 µm 1 0.1 0.01 (b) NF 0 10 20 φ (%) 30 40 FIGURE 7.9 Predicted influence of floc radius and internal packing on normalized creaming velocity of oil-in-water emulsions with different disperse phase volume fractions. The curves in (a) represent flocs with different internal packing (ϕi = 0.2, 0.4, 0.6) but the same radius (rfloc = 5 μm). The curves in (b) represent flocs with the same internal packing (ϕi = 0.6) but different radius (rfloc = 2, 5 or 10 μm). NF represents nonflocculated emulsions. as uncharged droplets. Thus, as a droplet moves upward there is a greater chance that its neighbors will be caught in the downward flow of the continuous phase. Second, the cloud of counter-ions surrounding a droplet moves less slowly than the droplet itself, which causes an imbalance in the electrical charge that opposes the movement of the droplet. Third, if the electrostatic repulsive interactions between droplets are sufficiently strong and long range then the droplets may be prevented from moving because of the substantial increase in the effective volume fraction of the droplets. 7.3.1.2.7 Fat Crystallization Many food emulsions contain a lipid phase that is either partly or wholly crystalline (Fredrick et al. 2010, McClements 2012). In oil-in-water emulsions, the crystallization of the lipid phase affects the overall 302 Food Emulsions: Principles, Practices, and Techniques 35 30 Nonflocculated ν (mm day−1) 25 20 15 10 5 0 0 10 20 φ (%) 30 40 40,000 35,000 Flocculated ν (mm day−1) 30,000 25,000 20,000 15,000 10,000 5,000 0 0 10 20 30 40 φ (%) FIGURE 7.10 Dependence of creaming velocities on droplet concentration for n-hexadecane oil-in-water emulsions containing monodisperse droplets (r = 0.86 μm) stabilized by SDS. The emulsions were either nonflocculated (7 mM SDS) or flocculated (80 mM SDS) depending on the concentration of free SDS micelles added to promote depletion flocculation. Note the difference in scales on the two graphs. creaming rate because solid fat (ρ ≈ 1200 kg m−3) has a higher density than liquid oil (ρ ≈ 910 kg m−3). The density of a droplet containing partially crystallized oil is given by: ρdroplet = ϕSFC ρsolid + (1 − ϕSFC) ρliquid, where ϕSFC is the volume fraction of solid fat (relative to the total fat). At a solid fat content of about 30% an oil droplet has a similar density as water and will therefore neither cream nor sediment. At lower solid fat contents the droplets cream, and at higher solid fat contents they sediment. This partially accounts for the more rapid creaming of milk fat globules at 40°C (where they are completely liquid) than at 20°C (where they are partially solid) (Mayhill and Newstead 1992). As mentioned above, crystallization of the fat in a water-in-oil emulsion may lead to the formation of a three-dimensional network of aggregated fat crystals that prevents the water droplets from sedimenting, for example, in butter and margarine. In these systems, there is a critical solid fat content necessary for the formation of a network, which depends on the morphology of the fat crystals 303 Emulsion Stability (Fredrick et al. 2010). The importance of network formation is illustrated by the effect of heating on the stability of margarine. When margarine is heated above a temperature where most of the fat crystals melt, the network breaks down and the water droplets sediment, leading to the separation of the oil and aqueous phases. 7.3.1.2.8 Adsorbed Layer The presence of a layer of adsorbed emulsifier molecules at the surface of an emulsion droplet affects the creaming rate in a number of ways (McClements 2011, McClements and Rao 2011). First, it increases the effective size of the emulsion droplet, and therefore the creaming rate is increased by a factor of (1 + δ/r)2, where δ is the thickness of the adsorbed layer. Typically, the thickness of an adsorbed layer is between about 2 and 10 nm, and therefore this effect is only significant for very small emulsion droplets (<0.1 μm). Second, the adsorbed layer may alter the effective density of the dispersed phase, ρ2. The effective density of the dispersed phase when the droplets are surrounded by an adsorbed layer can be calculated using the following relationship, assuming that the thickness of the adsorbed layer is much smaller than the radius of the droplets (δ ≪ r): r2 = rdroplet + 3(d /r )rlayer 1 + 3(d /r ) (7.18) The density of the emulsifier layer is usually greater than that of either the continuous or dispersed phases and, therefore, the adsorption of the emulsifier increases the effective density of the dispersed phase. The density of large droplets (r ≫ δ) is approximately the same as that of the bulk dispersed phase, but that of smaller droplets may be altered significantly. It is therefore possible to retard creaming by using a surface-active biopolymer that forms a relatively thick and dense interfacial layer. This mechanism may be important in slowing down creaming in beverage emulsions, since the droplet size is relatively small and the interfacial layers are relatively thick (Piorkowski and McClements 2014). 7.3.1.2.9 Brownian motion Another major limitation of Stokes’ equation is that it ignores the effects of Brownian motion on the creaming velocity of emulsion droplets (Hiemenz and Rajagopalan 1997, Walstra 2003). Gravity favors the accumulation of droplets at either the top (creaming) or bottom (sedimentation) of an emulsion. On the other hand, Brownian motion favors the random distribution of droplets throughout the whole of the emulsion because this maximizes the configurational entropy of the system. The equilibrium distribution of droplets in an emulsion that is susceptible to both creaming and Brownian motion is given by the following equation (ignoring the finite size of the droplets): æ -4pr 3Drgh ö f(h) = f0 exp ç ÷ 3kT è ø (7.19) where ϕ(h) is the concentration of the droplets at a distance h below the top of the emulsion ϕ 0 is the concentration of the droplets at the top of the emulsion If ϕ(h) = ϕ 0, then the droplets are evenly dispersed between the two locations (i.e., Brownian motion completely dominates), but if ϕ(h) ≪ ϕ 0, the droplets tend to accumulate at the top of the emulsion (i.e., creaming dominates). Predictions of the dependence of ϕ(h)/ϕ 0 on emulsion height (h) for typical food emulsions (Δρ = 100 kg m−3) are shown in Figure 7.11. These predictions suggest that Brownian motion may play a role on the creaming behavior of emulsions when the droplet radius is less than about 0.1 μm, and that the accumulation of droplets at the top of an emulsion due to creaming should be almost completely retarded when the droplet radii is less than about 10 nm. This effect is responsible for the fact that many microemulsions are stable to gravitational separation. 304 Food Emulsions: Principles, Practices, and Techniques 10 nm 1 20 nm φ/φ0 0.8 0.6 Small 50 nm 0.4 0.2 0 100 nm 200 nm 0 Large 20 40 60 Sample height (mm) 80 100 FIGURE 7.11 Predicted influence of Brownian motion on the distribution of oil droplets in model oil-in-water emulsions with different droplet sizes (Δρ = 100 kg m−3). 7.3.1.2.10 Complexity of Creaming The above discussion has highlighted the many factors that need to be considered when predicting the rate at which droplets cream in emulsions. In some situations, it is possible to combine two or more of the factors mentioned above into a simple analytical equation. However, it may be difficult to accurately model the creaming behavior of an emulsion if a large number of factors operate simultaneously. In these situations, the most comprehensive method of predicting gravitational separation in emulsions it is to use computer simulations (Pinfield et al. 1997, Gonzalez et al. 2004). 7.3.2 Methods of Controlling Gravitational Separation The discussion of the physical basis of gravitational separation in Section 7.3.1 has highlighted a number of ways of retarding its progress in food emulsions: 7.3.2.1 Minimizing Density Difference The driving force for gravitational separation is the density difference between the droplets and the surrounding liquid: Δρ = (ρ2 − ρ1). It is therefore possible to inhibit or prevent gravitational separation by “matching” the densities of the oil and aqueous phases (Piorkowski and McClements 2014). Most naturally occurring triacylglycerol-based edible oils have fairly similar densities (≈910 kg m−3) and therefore food manufacturers have limited flexibility in preventing creaming by changing the type of oil used in their product.* Nevertheless, a number of alternative strategies have been developed that enable food manufacturers to match the densities of the dispersed and continuous phases more closely. Density matching can be achieved by mixing oil-soluble or oil-dispersible “weighting agents” that have a higher density than water with the oil phase prior to homogenization, so that the overall density of the oil droplets becomes similar to that of the aqueous phase (Piorkowski and McClements 2014). * It should be noted that the flavor oils used in beverage emulsions usually have appreciably lower densities than triacylglycerol oils (Chapter 12). 305 Emulsion Stability Different types of weighting agents are available to food manufacturers, including brominated vegetable oil (BVO), ester gum (EG), damar gum (DG), and sucrose acetate isobutyrate (SAIB). BVO is made by bromination of vegetable oil (ρ ≈ 1330 kg m−3). Ester gum is made by esterification of wood rosin with glycerol (ρ ≈ 1080 kg m−3). Damar gum is a natural exudate obtained from the shrubs of the Caesalpiniaceae and Dipterocarpaceae families (ρ ≈ 1050–1080 kg m−3). SAIB is made by the esterification of sucrose with acetic and isobutyric anhydrides (ρ ≈1146 kg m−3). Because of the differences in their densities, different amounts of the weighting agents are required to completely match the density of the oil phase to that of the aqueous phase, with their effectiveness at density matching decreasing in the following order: BVO > SAIB > EG, DG (Chanamai and McClements 2000). Weighting agents also vary in their legal status and maximum permissible usage level in different countries, as well as in their cost and ease of use. The influence of adding a weighting agent (BVO) to the oil phase of an oil-in-water emulsion is demonstrated in Figure 7.12 (Chanamai and McClements 2000). At low BVO concentrations, the density of the droplets is less than that of the aqueous phase and creaming occurs. At high BVO concentrations, the density of the droplets is greater than that of the aqueous phase and sedimentation occurs. At a certain BVO concentration in the oil phase (∼25%), the density of the droplets equals that of the aqueous phase and gravitational separation is completely suppressed. If the droplets in an oil-in-water emulsion are sufficiently small, it may be possible to prevent gravitational separation by using an emulsifier that forms a relatively thick and dense interfacial layer, because this decreases the density contrast between the oil droplets and the surrounding liquid (Section 7.3.1). This mechanism is important in beverage oil-in-water emulsions because the droplet size is relatively low and the thickness of the interfacial coating is relatively thick (Piorkowski and McClements 2014). In some emulsions, it may be possible to control the degree of gravitational separation by varying the solid fat content of the lipid phase. As mentioned in Section 7.3.1, an oil droplet with a solid fat content of about 30% has a similar density to water and should therefore be stable to gravitational separation. The solid fat content of a droplet could be controlled by altering the composition of the lipid phase or 10 5 ν (mm day−1) 0 900 1000 1100 1200 1300 –5 –10 –15 –20 –25 Droplet density (kg m–3) FIGURE 7.12 Influence of droplet density on the creaming velocity of 1 wt% oil-in-water emulsions containing different ratios of soybean oil to brominated vegetable oil (0%–100%) at 25°C. When ρ2 < ρ1, creaming occurs, when ρ2 > ρ1 sedimentation occurs, and when ρ2 = ρ1 no droplet movement occurs. Densities of pure liquids: Aqueous phase = 1000.4 kg m−3; Soybean oil = 911.1 kg m−3; and BVO = 1329.5 kg m−3. 306 Food Emulsions: Principles, Practices, and Techniques by controlling the temperature (Akoh and Min 2008, McClements 2012). In practice, this procedure is unsuitable for many food emulsions because partially crystalline droplets are susceptible to partial coalescence, which severely reduces their stability (Section 7.7). 7.3.2.2 Reducing Droplet Size Stokes’ law indicates that the velocity at which a droplet moves is proportional to the square of its radius (Equation 7.9). The stability of an emulsion to gravitational separation can therefore be enhanced by reducing the size of the droplets it contains (Figure 7.7). Homogenization of raw milk is one of the most familiar examples of the retardation of creaming in a food emulsion by droplet size reduction. A food manufacturer typically aims to reduce the droplet size in an emulsion below some critical size known to be small enough to prevent creaming during the lifetime of the product. In practice, homogenization leads to the formation of emulsions that contain a range of different sizes, and the largest droplets are most susceptible to gravitational separation. For this reason, a food manufacturer usually specifies the minimum percentage of droplets that can be above the critical droplet size without leading to a significant decrease in perceived product quality. For example, cream liqueurs are usually designed so that less than 3% of the droplets have radii greater than 0.2 μm (Dickinson 1992). Even though a small fraction of the droplets are greater than this size, and therefore susceptible to creaming, this does not cause a major problem because the presence of the droplet-rich creamed layer at the top of the emulsion is obscured by the opacity produced by the smaller droplets remaining in the bulk of the emulsion that do not cream appreciably. The creaming stability can also be improved by preventing any changes in the system that lead to an increase in the droplet size, such as flocculation, coalescence, or Ostwald ripening (see Sections 7.5 through 7.8). 7.3.2.3 Modifying Continuous Phase Rheology Increasing the viscosity of the liquid surrounding a droplet, η1, decreases the velocity at which the droplet moves (Equation 7.9). Thus, the stability of an emulsion to gravitational separation can be enhanced by increasing the viscosity of the continuous phase, for example, by adding a thickening agent (Section 4.5). Gravitational separation may be completely retarded if the continuous phase contains a three-dimensional network of aggregated molecules or particles that traps the droplets and prevents them from moving. Thus, the droplets in oil-in-water emulsions can be completely stabilized against creaming by using biopolymers that form a gel in the aqueous phase (Section 4.5), while the droplets in water-in-oil emulsions can be completely stabilized against sedimentation by ensuring there is a network of aggregated fat crystals in the oil phase. 7.3.2.4 Increasing Droplet Concentration The rate of gravitational separation can be retarded by increasing the droplet concentration. At a sufficiently high disperse phase volume fraction, the droplets are prevented from moving because they are so closely packed together (Figure 7.5). It is for this reason that the droplets in mayonnaise, which has a high disperse phase volume fraction, are more stable to creaming than those in salad dressings, which have a lower disperse phase volume fraction. Nevertheless, it should be mentioned that it is often not practically feasible to alter the droplet concentration, and therefore one of the alternative methods of preventing creaming should be used. It may be possible to increase the effective volume fraction of the droplets in an emulsion, without increasing the overall fat content, by using water-inoil-in-water (W/O/W) emulsions, rather than conventional oil-in-water (O/W) emulsions (JimenezColmenero 2013). 7.3.2.5 Altering the Degree of Droplet Flocculation The rate of gravitational separation can be controlled by altering the degree of flocculation of the droplets in an emulsion. In dilute emulsions, flocculation causes enhanced gravitational separation 307 Emulsion Stability because it increases the effective size of the particles. To improve the stability of these systems, it is important to ensure that the droplets are prevented from flocculating (Section 7.5). In concentrated emulsions, flocculation reduces the rate of gravitational separation because the droplets are prevented from moving past one another (Figure 7.6). The critical disperse phase volume fraction at which separation is prevented depends on the structural organization of the droplets within the flocs (Section 7.5.3). The creaming stability of concentrated emulsions may therefore be enhanced by altering the nature of the colloidal interactions between the droplets and therefore the structure of the flocs formed. 7.3.3 Experimental Characterization of Gravitational Separation The long-term stability of an emulsion to gravitational separation can often be predicted using Stokes’ law and its modifications. To theoretically predict the rate at which gravitational separation occurs in an emulsion it is necessary to have information about the densities of the dispersed and continuous phases, the droplet size distribution, and the rheological properties of the continuous phase. The density of the liquids can be measured using a variety of techniques, including density bottles, hydrometers, and oscillating U-tube density meters. The droplet size distribution can be measured by microscopy, light scattering, electrical pulse counting, or ultrasonic methods (Chapter 14). The rheological properties of the continuous phase can be characterized using various types of viscometer and dynamic shear rheometer (Chapter 8). In principle, it is therefore possible to predict the creaming stability of a food emulsion from knowledge of its physicochemical properties and a suitable mathematical model. In practice, this approach has limited use because the mathematical models are not currently sophisticated enough to take into account the inherent complexity of most food emulsions. For this reason, it is often more appropriate to directly measure the gravitational separation of the droplets in an emulsion. The simplest method of monitoring gravitational separation is to place an emulsion in a transparent test tube, leave it for a certain length of time, and then measure the height of the interfaces between the different layers formed (Figure 7.13). For example, in oil-in-water emulsions it is often possible to visually discern a lower droplet-depleted “serum” layer (ϕ < ϕ initial), an intermediate “emulsion” Upper “Creamed” HU Middle “Emulsion” HM Lower “Serum” HE HL Creaming index: CI = 100 × HL/HE FIGURE 7.13 Different layers are often observed in an emulsion undergoing creaming (i) an upper “cream” layer (ϕ > ϕinitial); (ii) a middle layer (ϕ = ϕinitial); and (iii) a lower “serum” layer (ϕ < ϕinitial). 308 Food Emulsions: Principles, Practices, and Techniques CI One-layer system Three-layer system Two-layer system 50 CIfinal CI (%) 40 30 20 10 0 v = d(CI)/dt 0 20 40 60 80 Time (h) FIGURE 7.14 The creaming velocity can simply be determined by measuring the change in the height of the serum layer as a function of time. An emulsion usually forms three layers first (serum, emulsion, and cream) and then two layers (serum and cream) after long-term storage. layer (ϕ = ϕ initial), and a droplet-rich “creamed” upper layer (ϕ > ϕ initial). In general, an emulsion tends to separate into these three phases at the early stages of creaming, but then forms only two phases toward the latter stages, that is, a serum layer and a creamed layer (Figure 7.14). The creaming velocity can be estimated by manually measuring the change in the height in the serum layer over time. This procedure can often be accelerated by centrifuging an emulsion at a fixed speed for a certain length of time. Nevertheless, the use of accelerated creaming tests as a means of predicting the long-term stability of emulsions to gravitational separation should be treated with caution because the factors that determine droplet movement in a gravitational field may be different from those that are important in a centrifugal field. For example, the continuous phase may have a yield stress that is exceeded in a centrifuge, but which would never be exceeded under normal storage conditions. The two major problems associated with determining the extent of creaming visually are (1) it is only possible to obtain information about the location of the boundaries between the different layers, rather than about the full vertical concentration profile of the droplets and (2) in some systems it is difficult to clearly locate the boundaries between the different layers because the boundaries are diffuse or the layers are optically opaque. A more sophisticated method of monitoring gravitational separation is to use vertical-scanning light scattering methods (Mengual et al. 1999). An emulsion is placed in a vertical glass tube and a monochromatic beam of near infrared light is directed through it (Figure 7.15). The percentage of transmitted and/or scattered light is measured as a function of emulsion height using one or two detectors by scanning the light beam up and down the sample using a stepper motor. The variation of droplet concentration with emulsion height can sometimes be deduced from the percentage of transmitted and/or 309 Emulsion Stability Scattered light Incident light Height Scattered light Transmitted light Transmitted light 0% 50% Signal 100% FIGURE 7.15 Light scattering device for monitoring creaming or sedimentation of droplets in emulsions. The light source and detectors scan vertically up the emulsion and record the intensity of transmitted and scattered light. scattered light using a suitable theory or calibration curve. Nevertheless, it is often difficult to quantify the actual droplet concentration versus height profile within emulsions because the intensities of the scattered and transmitted light do not change appreciably with changes in ϕ at high droplet concentrations and are also dependent on droplet radius (Chapter 10). In principle, this technique could be used to measure both the size and concentration of the droplets in dilute emulsions at any height by measuring the angular dependence of the intensity of the scattered light. This technique is finding increasing use for the characterization of gravitational separation in food emulsions due to the fact that fully automated analytical instruments based on this principle have recently become commercially available. The major disadvantages of this technique are that it is unsuitable for monitoring gravitational separation in some concentrated emulsions, and it is difficult to accurately determine the full profile of droplet concentration versus emulsion height. Traditionally, the kinetics of gravitational separation was monitored in concentrated emulsions by physically removing sections of an emulsion from different heights and then analyzing the concentration of droplets in each section, for example, by measuring the density or by evaporating the water (Pal 1994). These techniques cause the destruction of the sample being analyzed and cannot therefore be used to monitor creaming in the same sample as a function of time. Instead, a large number of similar samples have to be prepared and each one analyzed at a different time. Recently, a number of nondestructive analytical methods have been developed to monitor gravitational separation in concentrated emulsions without disturbing the sample, for example, electrical conductivity, ultrasound, and NMR (Chapter 14). Information about gravitational separation can be obtained by inserting electrodes into an emulsion and measuring the change in electrical conductivity across them at different heights and times. Using a suitable theoretical model, the electrical conductivity at a particular emulsion height can be converted into a droplet concentration. The ultrasonic device is very similar to the light scattering technique described above, except that it is based on the propagation of ultrasonic waves through an emulsion, rather than light waves. An ultrasonic transducer is scanned vertically up and down an emulsion, which enables one to determine the droplet concentration (and sometimes droplet size) as a function of emulsion height. NMR imaging techniques, which are based on differences in the response of oil and water to the application of a radio frequency pulse, have also been used to monitor gravitational separation in emulsions. These techniques enable one to obtain a three-dimensional image of the 310 Food Emulsions: Principles, Practices, and Techniques droplet concentration (and sometimes droplet size) within a concentrated emulsion without the need for dilution, but they are expensive to purchase and require highly skilled operators, which has somewhat limited their application. 7.4 Droplet Aggregation: General Features The droplets in emulsions are in continual motion because of the effects of thermal energy, gravity, or applied mechanical forces, and as they move about they frequently collide with their neighbors. After a collision, emulsion droplets may either move apart or remain aggregated, depending on the relative magnitude of the attractive and repulsive interactions between them (Chapter 3). Droplets aggregate when there is a minimum in the interdroplet pair potential that is sufficiently deep and accessible to the droplets. The two major types of aggregation in food emulsions are flocculation and coalescence. Flocculation is the process whereby two or more droplets come together to form an aggregate in which the droplets retain their individual integrity, whereas coalescence is the process whereby two or more droplets merge together to form a single larger droplet. In this section, we consider some of the more general features of droplet aggregation, while in the following sections we discuss droplet flocculation and coalescence separately in order to highlight the most important factors that influence them in food emulsions. Consider a system that initially consists of a number of nonaggregated spherical particles dispersed in a liquid. Over time, the particles may either remain as individual entities or they may associate with their neighbors. Droplet association may take the form of flocculation or coalescence, where flocculation may be either reversible (weak flocculation) or irreversible (strong flocculation or coagulation). As an emulsion scientist one is interested in predicting the evolution of the particle size distribution of the system. In particular, one would like to know the change in the concentration of the different types of particles present in the system with time, that is, individual particles, particles present in weak or strong flocs (dimers, trimers, etc.), and particles that have become coalesced (dimers, trimers, etc.). Considerable progress has been made in developing mathematical models to describe the kinetics of particle aggregation in colloidal systems (Saether et al. 2004, Dickinson 2013). In general, the aggregation kinetics depends on the mechanism responsible for particle–particle encounters, the hydrodynamic and colloidal interactions acting between the particles, and the susceptibility of the thin film separating the particles to become ruptured. Some of the most important physiochemical mechanisms that influence the rate of particle aggregation in emulsions are identified in Figure 7.16. The relative importance of these processes on droplet aggregation is briefly discussed below assuming that the colloidal interactions in the system are similar to those shown in Figure 7.16, that is, a secondary minimum, an energy barrier, a deep primary minimum, and a strong short-range repulsion. This kind of profile is representative of many protein-stabilized emulsions at pH values away from the isoelectric point where they are stabilized by both electrostatic and steric stabilization. 7.4.1 Droplet–Droplet Encounters The first prerequisite for droplet aggregation to occur is that the droplets move toward each other and come into close proximity. The rate at which droplets encounter each other is largely determined by the dominant mechanism responsible for droplet movement in the emulsion, for example, Brownian motion, gravity, or applied shear. A droplet encounter time (τEnc) can be defined, which provides a measure of the average time between droplet collisions. 7.4.2 Film Thinning When the droplets come into close proximity, a relatively thin film of continuous phase is formed between them and this fluid must be squeezed out before the droplets can get any closer. This process generates a hydrodynamic resistance to droplet approach because of the friction associated with fluid flow out 311 Emulsion Stability w(h) Energy barrier h τCoag τFD Coalesced 1° min τFloc 2° min τFrag FIGURE 7.16 Droplet aggregation involves a number of physiochemical processes, including droplet approach, film thinning, thin film formation, and thin film rupture. These processes are strongly dependent on the colloidal and hydrodynamic interactions between the droplets. The overall aggregation rate and the type of aggregation that occurs depend on which of these processes are rate limiting. of the thin film (Ivanov et al. 1999, Krebs et al. 2013). In addition, there may be various attractive and repulsive colloidal interactions between the droplets with different signs, magnitudes, and ranges, which will also alter the rate at which droplets approach each other. A characteristic film thinning time (τFT) can be defined, whose magnitude depends on the nature of the colloidal and hydrodynamic interactions acting between the droplets. 7.4.3 Thin Film Formation The film of continuous phase separating the droplets continues to thin up to a certain value, after which a number of events may occur depending on the nature of the colloidal and hydrodynamic interactions in the system (Ivanov et al. 1999, Dukhin et al. 2001, 2003, Mishchuk 2005). The droplets may move apart (no aggregation), remain in a secondary minimum (weak flocculation), remain in a primary minimum (coagulation), or move closer together and coalesce (Figure 7.16). • No aggregation: If the secondary minimum is shallow, and there is a high energy barrier, then the droplets will tend to move apart immediately after a collision. • Weak flocculation: If the secondary minimum is fairly deep, and there is a high energy barrier, then the droplets will tend to weakly flocculate with a relatively thick film of continuous phase (but still only a few nanometers) separating the droplets. The fragmentation time (τFrag) is a measure of the average time that droplets spend in the secondary minimum before moving apart. This time increases as the depth of the secondary minimum increases. • Coagulation (strong flocculation): If the energy barrier is relatively low, but there is a strong short-range repulsion, then the droplets may fall into the primary minimum and be strongly flocculated with a relatively thin film of continuous phase between the droplets. Droplets may move directly into the primary minimum immediately following a droplet–droplet encounter, or (more usually) they may jump over the energy barrier after they have been trapped in a 312 Food Emulsions: Principles, Practices, and Techniques secondary minimum for some time. In the latter case, the coagulation time (τCoag) is a measure of the average time that droplets take to move from the secondary minimum into the primary minimum. The coagulation time increases as the height of the energy barrier increases. In all these examples, the height of the energy barrier and the depth of the primary and secondary minima are measured relative to the thermal energy (kT) of the system. 7.4.4 Film Rupture Droplet coalescence occurs if the thin film of fluid (the continuous phase) separating the droplets is ruptured and the fluids within the droplets (the dispersed phase) merge together (Kabalnov 1998, van Aken et al. 2003, Tcholakova et al. 2006a). If there is no strong short-range repulsion between the droplets, then they will tend to rapidly coalesce after falling into the primary minimum because there is nothing preventing them from getting close together. In this case, the rate of droplet coalescence is largely determined by the probability that the droplets obtain sufficient energy to jump over the primary energy barrier. In the presence of a high short-range repulsion, the droplets should be stable to coalescence. Nevertheless, droplet coalescence is often observed in real systems even though a shortrange repulsive force does exist, which is due to the disruption of the interfacial coatings surrounding the droplets. Interfacial disruption can occur through a variety of different mechanisms depending on the nature of any emulsifiers present at the droplet surfaces (see Section 7.6). The rate of droplet coalescence depends on the film disruption time (τ FD), which is the average time required for a rupture to appear in a film. The goal of theoreticians is to derive mathematical expressions for each of the characteristic times associated with these different physical events, since mathematical models can then be developed to predict the change in the number of the different types of particles (nonaggregated, flocculated, and coalesced droplets) in a system with time (Saether et al. 2004). The relative magnitude of these different characteristic times determines whether the system remains stable, undergoes flocculation, or undergoes coalescence. 7.5 Flocculation As mentioned earlier, flocculation is the process whereby two or more droplets associate with each other, but maintain their individual integrities. Droplet flocculation may be either advantageous or detrimental to emulsion quality depending on the nature of the food product. Flocculation accelerates the rate of gravitational separation in dilute emulsions, which is usually undesirable because it reduces their shelflife. It also causes a pronounced increase in emulsion viscosity, and may even lead to the formation of a gel at sufficiently high droplet concentrations. Some food products are expected to have a low viscosity and therefore flocculation is detrimental. In other products, a controlled amount of flocculation may be advantageous because it leads to the creation of a more desirable texture. Improvements in the quality of emulsion-based food products therefore depends on a better understanding of the factors that determine the degree of floc formation, the structure of the flocs formed, the strength of the bonds holding the droplets together within the flocs, and the rate at which flocculation proceeds. In addition, it is important to understand the effect that flocculation has on the bulk physicochemical, sensory and gastrointestinal properties of emulsions, for example, shelf-life, texture, taste, appearance, and digestion (Chapters 8 through 11). 7.5.1 Physical Basis of Flocculation In general, mathematical models can be derived to account for the change in the number of nonflocculated, flocculated, and coalesced particles in an emulsion with time (Saether et al. 2004). In this section, we present a relatively simple model to describe droplet flocculation in colloidal dispersions containing monodisperse spherical particles. As flocculation proceeds there is a decrease in the total number of 313 Emulsion Stability particles (monomers + aggregates) in an emulsion, which can be described by the following equation (Evans and Wennerstrom 1999): dnT 1 = - FE dt 2 (7.20) where dnT/dt is the flocculation rate nT is the total number of particles per unit volume t is the time F is the collision frequency E is the collision efficiency A factor of ½ appears in the equation because a collision between two particles leads to a reduction of one in the total number of particles present. Equation 7.20 indicates that the rate at which flocculation proceeds depends on two factors: the frequency of collisions between the droplets, and the fraction of collisions that leads to aggregation. 7.5.1.1 Collision Frequency The collision frequency is the total number of droplet encounters per unit time per unit volume of emulsion. Any factor that increases the collision frequency increases the flocculation rate (provided that it does not also decrease the collision efficiency). Collisions between droplets occur as a result of their movement, which may be induced by Brownian motion, gravitational separation, or applied mechanical forces depending on the system. 7.5.1.1.1 Collisions due to Brownian Motion In quiescent systems, collisions between droplets are mainly a result of Brownian motion. By considering the diffusion of particles in a dilute suspension, von Smoluchowski was able to derive the following expression for the collision frequency (Evans and Wennerstrom 1999): FB = 16pD0rn2 (7.21) Here FB is the collision frequency due to Brownian motion (m−3 s−1) D 0 is the diffusion coefficient of a single particle (m2 s−1) n is the number of particles per unit volume (m−3) r is the droplet radius (m) For rigid spherical particles, D 0 = kT/6πη1r, where η1 is the viscosity of the continuous phase, k is Boltzmann’s constant, and T is the absolute temperature. Hence: FB = kBn2 = 8kTn2 3kTf2 = 2h1p2r 6 3h1 (7.22) Here k B is a second-order rate constant (m3 s−1) ϕ is the disperse phase volume fraction For particles dispersed in water at room temperature, the collision frequency is ≈ 0.64 × 1018ϕ2/r6 (m−3 s−1), when the radius is expressed in micrometers. Equation 7.22 indicates that the frequency of collisions between droplets can be reduced by decreasing their volume fraction, increasing their size, or increasing 314 Food Emulsions: Principles, Practices, and Techniques the viscosity of the continuous phase. If it is assumed that every collision between two particles leads to aggregation, and that the rate constant is independent of aggregate size, then the flocculation rate is given by: dnT/dt = −(1/2)FB, which can be integrated to give the following expression for the change in the total number of particles with time: nT = n0 1 + (1/ 2)kBn0 t (7.23) Here n 0 is the initial number of particles per unit volume. The time taken to reduce the number of droplets in an emulsion by half can be calculated from the above equation: t1/ 2 = 2 3h1 æ ph ö r 3 = =ç 1 ÷ kBn0 4kTn0 è kT ø f (7.24) For a system where the particles are suspended in water at room temperature, τ1/2 ≈ r 3/ϕ 0 seconds when r is expressed in micrometers. Thus, an oil-in-water emulsion with ϕ = 0.1 and r = 1 μm would have a half-life of about 10 s, which is on the same order as the existence of an emulsion prepared by shaking oil and water together in the absence of a stabilizer. It is also possible to derive an equation to describe the change in the number of dimers, trimers, and other aggregates with time (Evans and Wennerstrom 1999): æ t ö nk = n0 ç ÷ è t1/ 2 ø k -1 æ t ö ç1 + ÷ t1/ 2 ø è - k -1 (7.25) Here, nk is the number of aggregates per unit volume containing k particles. The predicted variation in the total concentration of particles and of the concentration of monomers (k = 1), dimers (k = 2), and trimers (k = 3) with time is shown in Figure 7.17. As would be expected the total number of particles and the number of monomers decreases progressively with time as flocculation proceeds, while the number 1.0 Monomers (n1) 0.8 Dimers (n2) 0.6 n/n0 Trimers (n3) 0.4 ntotal n1 0.2 0.0 n2 n3 0 1 2 t/t1/2 3 4 FIGURE 7.17 Dependence of the concentration of the total number of particles (nT), monomers (n1), dimers (n2), and trimers (n3) on time t/τ1/2. The number of monomers decreases with time, whereas the number of aggregates initially increases and then decreases. 315 Emulsion Stability of dimers, trimers, and other aggregates initially increases with time and then decreases as they interact with other particles and form larger aggregates. The above equations are only applicable to dilute suspensions containing identical spherical particles suspended in an ideal liquid (Lattuada 2012). Many of the assumptions used in their derivation are not valid for actual food emulsions, which may be concentrated, polydisperse, and have nonideal continuous phases. In addition, the properties of the flocs cannot be assumed to be the same as those of the monomers, and therefore the above theory has to be modified to take into account the dimensions, structure, and hydrodynamic behavior of the flocs (Bremer et al. 1993). 7.5.1.1.2 Collisions due to Gravitational Separation In polydisperse emulsions, droplet–droplet encounters can occur because of the different creaming (or sedimentation) rates of the differently sized droplets. Large droplets move more quickly than smaller ones and therefore they collide with them as they move upward (or downward). The collision frequency for gravitationally induced flocculation is given by (Melik and Fogler 1988, Zhang and Davis 1991): FG = p(v2 - v1 )(r1 + r2 )2 n1n2 FG = kG n1n2 = ( ) 2 2 2 gDrf1f2 éê r2 - r1 (r1 + r2 ) ùú 8ph1 ê r13r23 ú ë û (7.26) (7.27) Here FG is the collision frequency due to gravitational separation vi is the Stokes creaming velocity of a particle with radius ri Δρ is the density difference between the droplets and the surrounding liquid This equation indicates that the collision frequency increases as the difference between the creaming velocities of the particles increases. The rate of gravitationally induced flocculation can therefore be retarded by ensuring that the droplet size distribution is not too wide, decreasing the density difference between the oil and aqueous phases, decreasing the droplet concentration, or increasing the viscosity of the continuous phase. Equation 7.27 would have to be modified before it could be applied to systems that do not obey Stokes’ law (Section 7.3). In addition, it does not take into account the fact that the droplets reach a position at the top or bottom of an emulsion where they cannot move any further and are therefore forced to encounter each other (Figure 7.14). 7.5.1.1.3 Collisions due to Applied Shear Forces Food emulsions are often subjected to various kinds of shear flow during their production, storage, and transport. Consequently, it is important to understand the effect that shearing has on their stability to flocculation. In a system subjected to Couette flow, the collision frequency is given by (Dickinson 1992): FS = kSn 2 = 16 3 2 æ 3G ö f2 Gr n = ç 2 ÷ 3 3 è p ør (7.28) Here, FS is the collision frequency due to shear. Thus, the frequency of shear-induced collisions can be retarded by decreasing the shear rate, increasing the droplet size, or decreasing the disperse phase volume fraction. It should be noted that the collision frequency is independent of the viscosity of the continuous phase. 7.5.1.1.4 Relative Importance of Different Collision Mechanisms In general, each of the above mechanisms may contribute to the droplet collision frequency in an emulsion. In practice, one or other of the mechanisms usually dominates, depending on the composition 316 Food Emulsions: Principles, Practices, and Techniques and microstructure of the product, as well as the prevailing environmental conditions. To effectively control the collision frequency, it is necessary to establish the mechanism that is the most important in the particular system being studied. It is convenient to use the collision frequency due to Brownian motion as a reference value, since this process occurs in most fluid emulsions. The ratio of the ­shear-to-Brownian motion collision frequencies (FS/F B) and the gravitational-to-Brownian motion collision frequencies (FG/F B) are plotted as a function of shear rate (G) and particle size ratio (=r 2/r 1), respectively, in Figure 7.18 for a typical oil-in-water emulsion. At low shear rates (G < 2 s−1), collisions due to Brownian motion dominate, but at high shear rates those due to mechanical agitation of the system dominate. Gravitationally induced collisions dominate those due to Brownian motion Shear vs Brownian 100 FS/FB 10 1 0 1 2 3 4 5 0.1 (a) 0.01 G (s–1) Gravity vs Brownian 100 FG/FB 10 1 0 2 4 6 8 0.1 (b) 0.01 r2/r1 FIGURE 7.18 Relative importance of the different collision mechanisms for typical food emulsions (Δρ = 90 kg m−3, r = 1 μm, ϕ = 0.1). (a) Shear induced collisions become increasingly important as the shear rate is increased. (b) Gravitationally induced collisions become increasingly important as the ratio of droplet sizes increases or the viscosity of the continuous phase decreases. 317 Emulsion Stability when the particle size ratio exceeds about 2, and thus it is likely to be most important in emulsions that have a broad particle size distribution. 7.5.1.2 Collision Efficiency If every encounter between two droplets led to aggregation, then emulsions would not remain stable long enough to be practically useful. To prevent droplets from flocculating during a collision, it is necessary to have a sufficiently high repulsive energy barrier to stop them from coming too close together (Chapter 3). The height of this energy barrier determines the likelihood that a collision leads to flocculation, that is, the collision efficiency. The collision efficiency, E, has a value between 0 (no flocculation) and 1 (every collision leads to flocculation)* and depends on the hydrodynamic and colloidal interactions between the droplets. The flocculation rate therefore depends on the precise nature of the interactions between the emulsion droplets (Ivanov et al. 1999). For collisions induced by Brownian motion (Dukhin and Sjoblom 1996): - dnB 4kTn 2 EB = dt 3h1 æ ¥ exp[ w(s) /kT ] ö EB = ç 2 ds ÷ ç ÷ s 2G ( s ) è 2 ø (7.29) -1 ò (7.30) Here s is the dimensionless center-to-center distance between the droplets (s = [2r + h]/r) r is the droplet radius h is the surface-to-surface separation Colloidal interactions are accounted for by the w(s) term, and hydrodynamic interactions by the G(s) term (Chapter 3). When there are no colloidal interactions between the droplets (w(s) = 0) and no hydrodynamic interactions (G(s) = 1), Equation 7.29 becomes equivalent to that derived by Smoluchowski (Equations 7.20 and 7.22). The stability of an emulsion to aggregation is governed primarily by the maximum height of the energy barrier, wmax(s), rather than by its width (Friberg and Larsson 1997). To enhance the stability of an emulsion against flocculation, it is necessary to have an energy barrier that is large enough to prevent the droplets from coming close together. The half-lives of emulsions with different height energy barriers have been estimated and are shown in Table 7.1. An energy barrier of about 20kT is usually sufficient to provide good long-term stability to emulsions. Expressions for the efficiency of shear and gravitational induced collisions also depend on colloidal and hydrodynamic interactions and have been derived for some simple systems (Liao and Lucas 2010). 7.5.1.3 Overall Particle Growth Rate To a first approximation, the increase in mean particle diameter with time in an emulsion due to flocculation can be calculated by assuming that at any given time the particles formed are monodisperse and have an “effective” particle diameter given by: d (t ) = 3 6f pnT (t ) (7.31) * In practice, E can have a value which is slightly higher than 1 because droplet collisions are accelerated when there is a strong attraction between the droplets. 318 Food Emulsions: Principles, Practices, and Techniques TABLE 7.1 Approximate Flocculation Half-Lives for Electrostatically Stabilized Emulsions with Different Energy Barriers w(hmax)/kT Half-Life 0 1 5 10 15 20 50 0.6 s 1s 30 s 1.2 h 23 h 3 years 3 × 1013 years Source: Friberg, S. and Larsson, K., Food Emulsions, Marcel Dekker, New York, 1997. The change in effective particle diameter with time can then be calculated by substituting this expression into the equation for the change in the total number of particles with time given above: d 3 = d03 + p3 fFB Et (7.32) Here, d 0 and d are the mean particle diameters at times zero and t, respectively. This equation indicates that there should be a linear increase in the cube of the mean particle diameter with time, and that the particle growth rate should increase with increasing droplet concentration, collision frequency, and collision efficiency. In practice, although the mean size of the particles in a flocculating emulsion may increase steadily with time (Figure 7.19a), the growth of the particles is not usually uniform throughout the size distribution (Figure 7.19b). Often, a fraction of the droplets become flocculated while the rest remain nonflocculated so that a bimodal particle size distribution is observed. More complex mathematical models are required to predict the change in the full particle size distribution with time. 7.5.2 Methods of Controlling Flocculation Knowledge of the physical basis of droplet flocculation facilitates the development of effective strategies of controlling it in food emulsions. These strategies can be conveniently divided into those that influence the collision frequency and those that influence the collision efficiency. 7.5.2.1 Collision Frequency The droplet flocculation rate can be controlled by manipulating the collision frequency of the droplets. The most effective means of achieving this depends on the dominant collision mechanism in the emulsion, that is, Brownian motion, gravity, or mechanical agitation. The rate at which droplets encounter each other in an unstirred emulsion can be reduced by increasing the viscosity of the continuous phase (Equation 7.22). Flocculation may be completely retarded if the continuous phase contains a threedimensional network of aggregated molecules or particles that prevents the droplets from moving, for example, a biopolymer gel or a fat crystalline network. The collision frequency increases when an emulsion is subjected to sufficiently high shear rates (Equation 7.28), and therefore it may be important to ensure that a product is protected from mechanical agitation during its storage and transport in order to avoid flocculation. The collision frequency increases as the droplet concentration increases or the droplet size decreases, with the precise nature of this dependence being determined by the type of collision mechanism that dominates. The rate of collisions due to gravitational separation depends on the relative velocities of the particles in an emulsion, and therefore decreases as the density difference between the droplet and surrounding liquid decreases or as the viscosity of the continuous phase increases (Equation 7.27). 319 Emulsion Stability 5 150 mM NaCl Mean diameter (µm) 4 3 2 1 0 mM NaCl 0 0 10 20 30 Storage time (h) (a) 40 50 Volume frequency 48 h 24 h 7h 0h 0.1 (b) 1 10 100 Particle diameter (µm) FIGURE 7.19 Evolution of mean particle diameter and particle size distribution of 5wt% n-hexadecane oil-in-water emulsions (1wt% β-lactoglobulin, pH 7.0, 0 or 150 mM NaCl) during storage at 30°C. (a) Mean particle diameter for 0 and 150 mM NaCl and (b) particle size distribution for 150 mM NaCl after 0 and 24 h storage. The droplet size increases because of flocculation induced by surface denaturation of adsorbed globular proteins. (Adapted from Kim, H.J. et al., J. Agric. Food Chem., 50(24), 7131, 2002a.) 7.5.2.2 Collision Efficiency The most effective means of controlling the rate and extent of flocculation in an emulsion is to regulate the colloidal interactions between the droplets. Flocculation can be prevented by designing an emulsion in which the repulsive interactions between the droplets are appreciably greater than the attractive interactions. Numerous different types of colloidal interaction can act between the droplets in an emulsion, for example, van der Waals, steric, electrostatic, hydrophobic, and depletion (Chapter 3). Which of these is important in a given system depends on the type of ingredients present, the microstructure of the 320 Food Emulsions: Principles, Practices, and Techniques emulsion, and the prevailing environmental conditions. To control flocculation in a particular system, it is necessary to identify the most important types of colloidal interaction. 7.5.2.2.1 Electrostatic Interactions Many oil-in-water emulsions used in the food industry are (at least partly) stabilized against flocculation using electrically charged emulsifiers that generate an electrostatic repulsion between the droplets, for example, ionic surfactants, proteins, or polysaccharides (Chapter 4). The flocculation stability of electrostatically stabilized oil-in-water emulsions depends mainly on the electrical properties of the emulsion droplets (ψ0 and σ), and the pH and ionic strength of the surrounding aqueous phase. The number, position, sign, and dissociation constants of the ionizable groups on adsorbed emulsifier molecules determine the electrical behavior of emulsion droplets under different environmental conditions. For each type of food product, it is therefore necessary to select an emulsifier with appropriate electrical characteristics. Hydrogen ions are potential determining ions for many food emulsifiers (e.g., –COOH → –COO − + H+ or –NH2 + H+ → –NH3+), and therefore the sign and magnitude of the electrical charge on emulsion droplets is determined principally by the pH of the surrounding solution (Chapter 3). In protein-stabilized emulsions, the electrical charge on the droplets goes from positive at low pH, to zero at the isoelectric point, to negative at high pH (Figure 7.20a). This change in droplet charge has a major impact on the stability of protein-stabilized emulsions to droplet flocculation (Figure 7.20b). At pH values sufficiently above or below the isoelectric point of the proteins, the droplet charge is large enough to prevent flocculation because of the relatively strong electrostatic repulsion between the droplets. At pH values near to the isoelectric point (pI ± 2), the net charge on the proteins is relatively low and the electrostatic repulsion between the droplets is no longer sufficiently strong to prevent flocculation. Droplet flocculation leads to a pronounced increase in the viscosity of an emulsion, as well as a decrease in creaming stability, and therefore has important implications for food quality. The ionic strength of an aqueous solution depends on the concentration and valency of the ions it contains (Israelachvili 2011). As the ionic strength is increased the electrostatic repulsion between droplets is progressively screened, until eventually it is no longer strong enough to prevent flocculation (Chapter 3). The minimum amount of electrolyte required to cause flocculation is known as the critical flocculation concentration or CFC. The CFC decreases as the surface potential of the emulsion droplets decreases and as the valency of the counter-ions increases. It has been shown that CFC ∝ Y 04 /z 2 (where Ψ0 is the surface potential and z is the counter-ion valency) for droplets with relatively low surface potentials, that is, Ψ0 < 25 mV (Hunter 1986). These low surface potentials are found in many food emulsions that are susceptible to flocculation. Under certain conditions, Ψ0 is inversely proportional to the valency of the counter-ions (z), so that CFC ∝ 1/z6, which is known as the Schultz–Hardy rule. This relationship indicates that a much lower concentration of a multivalent ion is required to cause droplet flocculation, than a monovalent ion, for example, 64-times less of a divalent counter-ion should be required than a monovalent counter-ion. This effect is clearly shown in Figure 7.21 where the flocculation stability of anionic protein-coated lipid droplets in compared in the presence of monovalent (K+) and divalent (Ca2+) counter-ions. The Schultz–Hardy rule can be derived from the DLVO theory by assuming that the CFC occurs when the potential energy barrier, which normally prevents droplets from aggregating, falls to a value of zero due to the addition of salt. Consequently, when two droplets collide with each other they immediately aggregate into the primary minimum. In practice, significant droplet flocculation occurs when the potential energy barrier is slightly higher than zero, and therefore the Schultz–Hardy rule is expected to slightly overestimate the CFC. The ability of ions to promote droplet flocculation in emulsions also depends on whether they are indifferent ions or specifically bound ions (Chapter 3). Monovalent counter-ions (e.g., K+, Na+, and Cl−) tend to be indifferent ions that screen electrostatic interactions, but do not alter the surface charge density or isoelectric point of electrically charged emulsion droplets by binding to the droplet surfaces (Kulmyrzaev and Schubert 2004). On the other hand, multivalent counter-ions (e.g., Ca2+, Cu2+, Fe2+, Fe3+, Al3+, and SO42−) may bind to the surface of emulsion droplets, thereby altering the surface charge density and isoelectric point of the droplets, as well as screening the electrostatic interactions (Mei et al. 1998, Kulmyrzaev et al. 2000, Ramkumar et al. 2000, Keowmaneechai and McClements 2002). 321 Emulsion Stability 80 60 ζ-Potential (mV) 40 20 pI 0 –20 –40 –60 –80 3 4 5 (a) 6 7 8 pH Particle diameter (µm) 100 Floc 10 1 Stable Stable 0.1 (b) 3 4 5 6 7 8 pH FIGURE 7.20 Influence of pH on the (a) ζ-potential and (b) flocculation stability of oil-in-water emulsions stabilized by whey protein. Extensive flocculation is observed near the protein isoelectric point (pI ∼ 5) because the electrostatic repulsion is no longer sufficiently strong to overcome the attractive interactions. Specifically adsorbed mineral ions usually decrease the surface charge density on droplets by an amount that depends on their valency and concentration, but they may also cause charge reversal if present at sufficiently high concentrations (Kippax et al. 1999). In addition, to their influence on surface charge specifically adsorbed ions are often highly hydrated and therefore increase the short-range hydration repulsion between droplets (Ivanov et al. 1999, Israelachvili 2011). In general, multivalent counter-ions tend to be much more effective at reducing the flocculation stability of electrostatically stabilized emulsions than monovalent counter-ions. The influence of monovalent and divalent counter-ions on the stability of protein-stabilized oil-inwater emulsions is illustrated in Figure 7.21. In this system, the proteins are negatively charged, so that the counter-ions are K+ and Ca2+. Appreciable droplet flocculation was observed in the emulsions when the counter-ion concentration exceeded about 250–300 mM for K+ ions or about 3–4 mM for 322 Food Emulsions: Principles, Practices, and Techniques 14 Mean diameter (µm) 12 10 8 6 4 2 0 0 100 200 (a) 300 400 500 KCl (mM) 14 Mean diameter (µm) 12 10 8 6 4 2 0 (b) 0 5 10 CaCl2 (mM) 15 20 FIGURE 7.21 Influence of mineral ion concentration on droplet flocculation in 20wt% corn oil-in-water emulsions stabilized by whey protein isolate (1wt% WPI, pH 7.0). At this pH, the protein-stabilized droplets have a negative charge, so that Ca2+ counter-ions (b) are more effective at promoting flocculation than K+ counter-ions (a). Ca2+ ions. These results are in accordance with the Schultz–Hardy rule and highlight the much greater effectiveness of multivalent ions at promoting droplet flocculation in emulsions stabilized by electrostatic repulsion. The flocculation stability of emulsions containing electrically charged droplets can be controlled in a variety of ways depending on the system. To prevent flocculation it is necessary to ensure that the droplets have a sufficiently high surface potential under the existing solution conditions (which often requires that the pH be controlled) and that the electrolyte concentration is below the CFC for the specific kinds of minerals present in the aqueous phase and the prevailing pH. In some foods it is necessary to have relatively high concentrations of multivalent mineral ions present for nutritional purposes, Emulsion Stability 323 for example, mineral-fortified emulsions used in infant, elderly, and athlete formulations. The potential negative impact of multivalent ions on emulsion stability can be reduced in a number of ways (1) using an emulsifier that provides stability through a nonelectrostatic mechanism, for example, steric repulsion; (2) ensuring that multivalent ions are excluded from the formulation, for example, by using purified water or other ingredients; and (3) by adding ingredients that sequester multivalent ions, for example, EDTA, citrate, or polyphosphates (Keowmaneechai and McClements 2002, 2006). 7.5.2.2.2 Steric Interactions Many food emulsifiers prevent droplet flocculation through steric repulsion, for example, some polysaccharides and nonionic surfactants (Chapters 3 and 4). This repulsion must be sufficiently strong and long range to overcome any attractive interactions (Chapter 3). Sterically stabilized emulsions are usually much less sensitive to variations in pH and ionic strength than electrostatically stabilized emulsions (Hunter 1986). Nevertheless, they can become unstable to flocculation under certain conditions. If the composition of the continuous phase or the temperature is altered so that polymer–polymer interactions become more favorable than solvent–solvent/solvent–polymer interactions, then the mixing contribution to the steric interaction becomes attractive and may promote droplet flocculation (Chapter 3). A sterically stabilized emulsion may also become unstable if the thickness of the interfacial layer is reduced (Chapter 3), which could occur if the polymeric segments on the emulsifier were chemically or biochemically cleaved (e.g., by acid or enzyme hydrolysis), if the continuous phase became a poor solvent for the polymer segments (e.g., by adding an antisolvent), or if the electrostatic repulsion between charged biopolymer molecules in the interfacial layer were reduced (e.g., by changing pH or increasing ionic strength). Short-range hydration forces make an important contribution to the flocculation stability of many sterically stabilized emulsions (Evans and Wennerstrom 1999, Israelachvili 2011). In these systems, droplet flocculation may occur when the emulsion is heated, because emulsifier head groups are progressively dehydrated with increasing temperature. This effect is common in oil-in-water emulsions stabilized by nonionic surfactants, where rapid particle growth is observed upon heating above a certain temperature (Figure 4.18). 7.5.2.2.3 Biopolymer Bridging Many types of biopolymer promote flocculation by forming bridges between two or more droplets (Dickinson 2003, Walstra 2003, Guzey and McClements 2006). In oil-in-water emulsions, biopolymers may adsorb onto bare oil droplets or onto emulsifier-coated oil droplets. To be able to bind to the droplets, there must be a sufficiently strong attractive interaction between segments of the biopolymer and the droplet surface. The most common types of interaction that operate in food emulsions are hydrophobic and electrostatic. When a biopolymer has a number of nonpolar residues along its back bone, some of them may associate with hydrophobic patches on one droplet, while others associate with hydrophobic patches on another droplet. This type of bridging flocculation tends to occur when there is insufficient biopolymer present to completely cover all the oil droplet surfaces formed during homogenization. Bridging may occur either during the homogenization process or after it is complete, for example, when a biopolymer is only weakly associated with a droplet then some of its segments can desorb and become strongly attached to a neighboring droplet. This type of bridging flocculation can usually be prevented by ensuring there is a sufficiently high concentration of biopolymer present in the continuous phase prior to homogenization. Bridging flocculation can also occur when a biopolymer in the continuous phase has an electrical charge that is opposite to that of the droplets (Guzey and McClements 2006, Dickinson 2011). In this case, bridging flocculation can be avoided in a number of ways (1) ensuring the droplets and biopolymer have similar charges; (2) ensuring that either the droplets or the biopolymer are uncharged; and (3) adding sufficient biopolymer to completely coat the droplets before flocculation can occur. An example of this type of bridging flocculation is shown in Figure 7.22, which shows the change in mean particle size with pectin concentration for protein-stabilized emulsions. In this case, the emulsion droplets are positively charged and the negatively charged pectin molecules act as bridges that hold two or more droplets together into flocs. At sufficiently high biopolymer concentrations the flocs may not form (or can easily 324 Food Emulsions: Principles, Practices, and Techniques 16 14 + Particle diameter (μm) 12 + + 10 8 – 4 + + 2 0 – + 6 0 0.05 0.1 – 0.15 0.2 Pectin concentration (wt%) FIGURE 7.22 Influence of pectin concentration on droplet flocculation in oil-in-water emulsions stabilized by β-lactoglobulin (pH 3.0). At this pH, the protein-stabilized droplets have a positive charge and the pectin has a negative charge, which leads to charge neutralization and bridging flocculation. The microscopy and optical images show differences in the microstructure and creaming stability of the emulsions in the absence and presence of pectin. be disrupted) because there is sufficient biopolymer present to completely cover all of the droplet surfaces, and so a single biopolymer does not link more than one droplet (Guzey and McClements 2006, Dickinson 2011). 7.5.2.2.4 Hydrophobic Interactions This type of interaction is important in emulsions that contain droplets that have nonpolar regions exposed to the aqueous phase. Hydrophobic interactions are believed to be responsible for the influence of surface and thermal denaturation of adsorbed globular proteins on the flocculation stability of oil-inwater emulsions (Kim et al. 2002a,b). At room temperature, β-lactoglobulin-stabilized emulsions (pH 7, 0 mM NaCl) are stable to flocculation because of the relatively strong electrostatic repulsion between the droplets. Nevertheless, they become unstable to flocculation when a sufficiently high level of salt is present in the continuous phase (Figure 7.23). At relatively low temperatures (<65°C), droplet flocculation has been attributed to surface denaturation of the proteins after adsorption. Surface denaturation occurs because of differences in the molecular environment of proteins in the nonadsorbed state (where they are surrounded by water) and the adsorbed state (where they are surrounded by water on one side and oil on the other). Consequently, the globular proteins undergo conformational changes after adsorption so as to maximize the number of favorable interactions and minimize the number of unfavorable interactions in their new environment (Zhai et al. 2013). It has been proposed that the conformational changes resulting from surface denaturation lead to an increased exposure of nonpolar and sulfhydryl containing amino acids to the aqueous phase, which promotes droplet aggregation through increased hydrophobic attraction and disulfide bond formation between proteins adsorbed to different droplets. When β-lactoglobulin-stabilized emulsions are heated above the thermal denaturation temperature of the adsorbed globular proteins (∼70°C) in the presence of salt (150 mM NaCl), protein unfolding becomes much more extensive, which leads to an increase in the extent of droplet flocculation (Figure 7.23). Nevertheless, it is interesting to note that very little droplet flocculation is observed when a β-lactoglobulin-stabilized emulsion is heated above the thermal denaturation temperature in the absence of salt, and then salt is added after the emulsion has been cooled to room temperature (Figure 7.23). These results suggest that interactions between proteins adsorbed onto different droplets are favored when the droplets are in close proximity during heating (i.e., high salt), but that interactions 325 Emulsion Stability 100 Particle diameter (µm) 150 mM NaCl (before) Thermal denaturation 10 Surface denaturation 150 mM NaCl (after) 1 0.1 No salt 30 40 50 60 70 Temperature (°C) 80 90 100 FIGURE 7.23 Influence of temperature and salt on the flocculation stability of n-hexadecane oil-in-water emulsions stabilized by β-lactoglobulin (pH 7, 0 or 150 mM NaCl). Flocculation may occur due to surface or thermal denaturation of adsorbed proteins depending on holding temperatures, salt content, and whether salt is added before or after heating. between proteins adsorbed onto the same droplets are favored when droplets are not in close proximity during heating (i.e., low salt) (Figure 7.24). This knowledge may provide a useful practical method of reducing the susceptibility of globular protein-stabilized emulsions to droplet flocculation during heat processing. Hydrophobic interactions are also likely to be important in emulsions in which there is insufficient emulsifier present to completely saturate the surfaces of the droplets (Dickinson 2003). This may occur when there is insufficient emulsifier present in an emulsion prior to homogenization or when an emulsion is diluted so much that some of the emulsifier desorbs from the droplet surfaces. Flocculation due to hydrophobic interactions can be avoided by ensuring that there is sufficient emulsifier present to completely cover the droplet surfaces or by selecting an emulsifier that does not undergo detrimental conformational changes that increase the surface hydrophobicity at the temperatures used during processing, storage, or handling. 7.5.2.2.5 Depletion Interactions The presence of nonadsorbing colloidal “particles” (such as some proteins, polysaccharides, and surfactant micelles) in the continuous phase of an emulsion causes an increase in the attractive force between the droplets due to an osmotic effect associated with the exclusion of the colloidal particles from a narrow region surrounding each droplet (Chapter 3). This attractive force increases as the concentration of colloidal particles increases, until eventually it may become large enough to overcome the repulsive interactions between the droplets and cause them to flocculate (Jenkins and Snowden 1996, Kleshchanok et al. 2008). This type of droplet aggregation is usually referred to as depletion flocculation. A wide variety of different food-grade colloidal particles have been shown to be capable of inducing depletion flocculation when added in sufficiently high concentrations, including surfactant micelles (Tween 20, SDS), polysaccharides (xanthan gum, gum arabic, modified starch, maltodextrin, pectin, and carrageenan), and protein aggregates (whey and caseinate). The lowest concentration required to cause depletion flocculation is referred to as the critical flocculation concentration by analogy to the CFC used to characterize the effect of salt on the stability of electrostatically stabilized emulsions. The CFC decreases as the size of the emulsion droplets increases and the effective volume fraction of the colloidal particles 326 Food Emulsions: Principles, Practices, and Techniques Flocculation –S–S– – –S –S H –S High salt –S–S– Heat –SH Inter-droplet interactions HS– HS– No flocculation –SH –S–S– Low salt –S–S– Conformational change Intra-droplet interactions FIGURE 7.24 Schematic diagram of the influence of interfacial cross-linking on the flocculation stability of emulsions coated by globular proteins. In the presence of high salt, the electrostatic repulsion between the droplets during heating is weak causing protein–protein interactions between different droplets. Conversely, in the presence of low salt, the electrostatic repulsion during heating is strong only allowing protein–protein interactions to occur within interfacial layers on similar droplets. increases (McClements 2000, Radford and Dickinson 2004). There is an optimum size of colloidal particles required to promote depletion flocculation which depends on a balance between (1) the number of particles per unit volume (which increases with decreasing size) and (2) the volume of the depletion zone (which decreases with decreasing size) (Radford and Dickinson 2004). The flocculation rate initially increases as the concentration of nonadsorbing colloidal particles is increased because of the enhanced attraction between the droplets, that is, a higher collision efficiency. However, once the concentration of colloidal particles exceeds a certain concentration, the flocculation rate often decreases because the viscosity of the continuous phase increases so much that the movement of the droplets is severely retarded, that is, a lower collision frequency. The influence of depletion flocculation on the creaming stability of oil-in-water emulsions containing different concentrations of a nonadsorbing uncharged biopolymer (maltodextrin) is illustrated in Figure 7.25. In the absence of biopolymer, the emulsions are stable to creaming over a 1 week period. However, once the biopolymer concentration in the continuous phase of the emulsions exceeds the CFC, then the net attraction between the droplets is sufficiently large to cause them to flocculate. The increase in mean particle size brought about by droplet flocculation initially leads to rapid creaming. However, at higher biopolymer concentrations the creaming rate is decreased, even though there is still a strong depletion attraction between the droplets, because the movement of the droplets is restricted due to the large increase in continuous phase viscosity. The CFC is strongly dependent on the molecular weight of the biopolymer molecules, decreasing as the molecular weight of maltodextrin increases (Figure 7.25). This can be attributed to the fact that the effective volume (Rv) of the maltodextrin molecules increases as their molecular weight increases; hence, they are more effective at promoting depletion flocculation (McClements 2000). For ionic colloidal particles, such as ionic surfactants or charged biopolymers, the CFC is expected to be dependent on electrolyte concentration because of its ability to screen electrostatic interactions, 327 Emulsion Stability 100 Creaming index (%) 80 60 1800 1200 720 40 500 20 0 0 10 20 Concentration (wt%) 30 40 FIGURE 7.25 Influence of maltodextrin concentration on the creaming stability of 5% corn oil-in-water emulsions. Depletion flocculation was observed above the critical flocculation concentration (CFC), which depended on the molecular weight of the maltodextrin (shown in Daltons beside the data lines). thereby reducing the effective volume of the colloidal particles. Studies of nonionic surfactant stabilized oil-in-water emulsions containing a nonadsorbed biopolymer (dextran sulfate) have shown that the CFC increases as the salt concentration increases because the effective volume of the ionic biopolymer is reduced by electrostatic screening effects (Demetriades and McClements 1998). In general, any change in solution conditions that alters the effective volume of the nonadsorbed colloidal particles will influence the CFC, for example, temperature, ionic strength, or solvent quality (Radford et al. 2004). It should be stressed that when an emulsion that exhibits depletion flocculation is diluted for particle sizing measurements (e.g., light scattering or electrical pulse counting), the flocs usually break down since then the concentration of nonadsorbed colloids particles falls below the CFC, that is, depletion flocculation is usually weak and reversible. This phenomenon can lead researchers to believe that an emulsion is stable, when actually it is flocculated in its normal state. 7.5.2.2.6 Hydrodynamic Interactions The efficiency of the collisions between droplets is also determined by the strength of the hydrodynamic interactions between them (Dukhin and Sjoblom 1996, Tanaka and Araki 2000). As two droplets approach each other, a repulsion arises because of the resistance associated with the flow of the continuous phase from the thin gap between them. The magnitude of this resistance decreases as the droplet surfaces become more mobile, leading to an increase in the collision efficiency (Chapter 3). On the other hand, the collision efficiency may be reduced when the droplet surfaces are stabilized by small molecule surfactants because of the Gibbs–Marangoni effect (Walstra 2003). 7.5.2.2.7 Covalent Interactions The droplets in flocs may also be held together by covalent interactions. For example, disulfide bond formation between protein molecules adsorbed onto different emulsion droplets has been proposed to account for the stability of flocs formed by surface and thermal denaturation of globular proteins (Kim et al. 2002a,b). Covalent bonds are short-range interactions and therefore droplets have to be in close proximity for this type of interaction to occur. This accounts for the fact that when β-lactoglobulinstabilized emulsions are heated above the thermal denaturation temperature of the adsorbed proteins in the presence of salt (small droplet separation) extensive interdroplet disulfide bond formation occurs, 328 Food Emulsions: Principles, Practices, and Techniques but when they are heated to the same temperature in the absence of salt (large droplet separation) only intradroplet disulfide bonds are formed (Figure 7.24). Consequently, extensive droplet flocculation is observed when β-lactoglobulin-stabilized emulsions are heated in the presence of salt (150 mM NaCl), but little droplet flocculation is observed when the same emulsions are heated in the absence of salt and then the same amount of salt (150 mM NaCl) is added after the emulsions are cooled to room temperature (Figure 7.23) (Kim et al. 2002a,b). 7.5.3 Structure and Properties of Flocculated Emulsions The appearance, taste, texture, and stability of emulsions is strongly influenced by the characteristics of any flocs formed, for example, their number, size, flexibility, and packing (van Vliet and Walstra 1989, Tadros 1994, Walstra 2003). The structure and properties of flocs are mainly determined by the nature of the colloidal and hydrodynamic interactions between the droplets, but they also depend on the mechanism responsible for the droplet collisions, that is, Brownian motion, gravity, or mechanical agitation (Bremer et al. 1993, Evans and Wennerstrom 1999). Valuable insights into the relationship between these parameters and floc characteristics have been obtained from a combination of computer simulations and experimental measurements. 7.5.3.1 Influence of Colloidal Interactions on Floc Structure When the attraction between the droplets is relatively strong compared to the thermal energy, the flocs formed tend to have open structures in which each droplet is only linked to two or three of its neighbors (Figure 7.26). This type of open structure is formed because the droplets “stick” firmly together at the point where they first come into contact and are unable to undergo any subsequent structural rearrangements. As a consequence, a droplet that encounters a floc cannot move very far into its interior before becoming attached to another droplet. This type of floc is characterized by a tenuous structure that traps large amounts of continuous phase within it. The volume fraction of particles within such a floc may be as low as 0.13, depending on its size and the strength of the interactions (Dickinson 1992). When the attraction between the droplets is relatively weak compared to the thermal energy, the droplets do not always stick together after a collision and they may be able to roll around each other after sticking together. Thus, a droplet that encounters a floc is able to penetrate closer to its center and flocs are able to undergo structural rearrangements, which means that the droplets can pack more closely together (Blijdenstein et al. 2004). This type of floc is characterized by a more compact structure that traps less of the continuous phase. The volume fraction of the particles in this type of floc may be as high as 0.63, which is close to the value for random packing of monodisperse particles. (a) (b) FIGURE 7.26 Schematic representation of different types of floc structure. The internal structure of flocs tends to be more open (lower ϕi) when the attractive forces between the droplets are stronger. (a) Floc with close packing and (b) floc with open packing. 329 Emulsion Stability 7.5.3.2 Use of Fractal Geometry to Describe Floc Structure Fractal geometry has proven to be a valuable tool for characterizing the structural organization of droplets in flocs (Walstra 2003, Dickinson 2013). Fractal geometry is applicable to systems that exhibit a phenomenon known as self-similarity, that is, have structures that appear self-similar when observed at different levels of magnification (Figure 7.27). There are no truly fractal objects in nature because all real objects have a definite upper and lower size. Nevertheless, many natural objects do show self-similarity over a number of levels of magnification and can therefore be described by fractal geometry (Peleg 1993). Despite their extremely complex structures, these fractal objects can be described by a single parameter, D, known as the fractal dimension. The concept of a fractal dimension is best illustrated by considering the model two-dimensional fractal flocs shown in Figure 7.27. The floc with the “string-of-beads” type structure has a fractal dimension of one (D = 1), because increasing the level of magnification by a factor of X changes the number of particles per floc by a factor of X1, that is, N ∝ X1. The floc that contains the array of closely packed particles has a fractal dimension of two (D = 2), because increasing the level of magnification by a factor of X changes the number of particles per floc by a factor of X2, that is, N ∝ X2. These are the two extreme values of the fractal dimension of a two-dimensional structure. There are other types of floc that have self-similar structures with intermediate fractal dimensions (Figure 7.27). The number of particles in these flocs is described by the relationship: N ∝ XD, where the fractal dimension D has a noninteger value between 1 and 2. The closer the value is to one, the more tenuous is the floc structure, and the closer it is to two, the more compact. The number of particles in a fractal floc is given by the relationship: N = (R/r)D. In nature, flocs are three-dimensional structures and so D ranges from 1 to 3: the higher the fractal dimension, the more compact the floc structure (Peleg 1993). The volume fraction of droplets in a threedimensional floc is given by (Walstra 2003): æRö fF = ç ÷ èrø D -3 (7.33) This equation indicates that the internal packing of a floc becomes more open as the floc size increases at the same fractal dimension, that is, ϕF decreases as R increases (because D < 3). One of the major factors influencing the fractal dimension is whether the reaction is diffusion-limited (every collision leads to aggregation, E → 1) or reaction-limited (only a fraction of collisions leads to aggregation, E < 1). D=1 1<D<2 D=2 FIGURE 7.27 Two-dimensional structures that illustrate the concept of self-similar (fractal) flocs. 330 Food Emulsions: Principles, Practices, and Techniques For diffusion-limited aggregation, the fractal dimension is usually around 1.7–1.75, but for reactionlimited aggregation it may be appreciably higher (up to 2.4). Thus, the flocs formed by diffusion-limited aggregation tend to have more open structures than those formed by reaction-limited aggregation. The fractal dimension of flocs can be determined using a variety of experimental techniques, including settling, rheology, scattering, and microscopy methods (Bushell et al. 2002). Thus, it is possible to quantify the influence of different parameters on the structure of the flocs formed, for example, pH, ionic strength, and temperature. It should be mentioned that fractal flocs are only kinetically stable structures (Evans and Wennerstrom 1999). When there is a strong attraction between the droplets the most thermodynamically stable arrangement would be one in which the droplets formed a close packed structure, as this would maximize the number of favorable attractive interactions. It is only because there is a large kinetic energy barrier associated with the rearrangement of droplets within a floc that it retains its fractal structure. If the external stresses acting on the flocs are sufficiently larger than the forces holding the droplets together, then the flocs may change from a fractal to a nonfractal structure, for example, Brownian motion, mechanical agitation, gravity, or centrifugal forces. Nevertheless, if the rearrangements only occur over a short distance, then the overall system may still have a fractal structure, but the primary units are then clusters of droplets rather than individual droplets (Walstra 2003). In this case, the radius of the clusters, rather than that of the individual droplets, should be used in Equation 7.33. It should also be stressed that not all types of aggregation mechanism lead to the formation of flocs with fractal structures. 7.5.3.3 Influence of Floc Structure on Emulsion Properties The structure and properties of flocs has a pronounced effect on the stability and rheological properties of emulsions. An emulsion containing flocculated droplets has a higher viscosity than one containing the same concentration of nonflocculated (Figure 7.13) droplets (Chapter 8). This is because the effective volume fraction of a floc is greater than the sum of the volume fractions of the individual droplets due to the presence of the continuous phase trapped within it (Tadros 1994, Quemada and Berli 2002). The apparent viscosity of emulsions usually increases as the floc structure becomes more tenuous for the same reason. Emulsions that contain flocculated droplets tend to exhibit pronounced shear thinning behavior, that is, the viscosity decreases as the shear rate or shear time increases (Chapter 8). Shear thinning occurs for two reasons (1) the flocs are deformed and become aligned with the shear field, which decreases their resistance to flow and (2) the flocs are disrupted by the shear forces, which decreases their effective volume fraction (Figure 7.28). The ease at which the flocs in an emulsion are deformed and disrupted decreases as the number and strength of the attractive interactions between the droplets increases (Quemada and Berli 2002). Thus, one would expect that a higher shear stress would be required to disrupt flocs in emulsions containing droplets that are strongly bound to each another, than one in which they are only weakly bound. Once the shearing forces are removed from an emulsion, the bonds between the droplets may reform. The rate at which this process occurs and the type of structures formed often has an important influence on the quality of food emulsions. In some cases, shearing an emulsion can actually promote droplet flocculation because the efficiency and frequency of collisions between the droplets is increased (Spicer and Pratsinis 1996a,b). Thus, an emulsion that is stable under quiescent conditions may become flocculated when it is subjected to shear forces. This type of emulsion initially shows shear thickening behavior because the formation of flocs leads to an increase in viscosity, but at higher shear rates these flocs may become deformed and disrupted, which leads to a decrease in viscosity (Pal 1996). Emulsions may therefore exhibit complex rheological behavior depending on the nature of the colloidal interactions between the droplets. At sufficiently high disperse phase volume fractions, flocculation leads to the formation of a threedimensional network of aggregated droplets that extends throughout the emulsion (Dickinson 2013). This type of system is referred to as a particle gel and can be characterized by a yield stress that must be exceeded before the emulsion will flow. The value of the yield stress increases as the disperse phase volume fraction increases and as the strength of the attractive forces between the droplets increases (Pal 1996). 331 Emulsion Stability 10 Elongation Viscosity (Pa s) Alignment Disruption 1 0.1 0.01 0.0001 0.01 1 100 10,000 Shear rate (s–1) FIGURE 7.28 Schematic diagram of events that occur during the shearing of a flocculated emulsion and their effect on emulsion viscosity. Flocs become increasingly deformed, aligned, and disrupted with increasing shear stress, which leads to a decrease in emulsion viscosity. The minimum disperse phase volume fraction required to form a particle gel decreases as the structure of the flocs become more tenuous (i.e., D → 1) because this type of structure is able to fill the available space more effectively (Figure 7.6). Thus, two emulsions may have exactly the same disperse phase volume fractions; yet, one could be a low viscosity liquid and the other a viscoelastic gel depending on the nature of the colloidal interactions between the droplets. Flocculation also affects the stability of emulsions to creaming. In dilute emulsions, flocculation increases the creaming rate because the effective size of the particles in the emulsion is increased, which more than compensates for the decrease in Δρ (Equation 7.13). In concentrated emulsions, the droplets are usually prevented from creaming because the formation of a three-dimensional network of aggregated droplets prevents them from moving (Figure 7.9). 7.5.4 Experimental Measurement of Flocculation 7.5.4.1 Microscopy Methods Microscopy methods are often the simplest and most direct means of providing information about droplet flocculation in food emulsions (McClements 2007). Various types of microscopy instruments are available to characterize flocculated emulsions, including optical, electron, or atomic force microscopy. The most suitable type of microscopy for a particular application depends on the dimensions of the objects to be resolved, and the sensitivity of emulsion microstructure to the preparation procedures used for each type of microscopy (Mikula 1992). Optical microscopy is the most widely used type of microscopy for characterizing droplet flocculation because it is relatively inexpensive, is readily available in many laboratories, and can provide valuable information about floc properties (Dickinson 2007, Loren et al. 2007). An aliquot of emulsion is placed on a microscope slide and then its overall microstructure is observed. Typically, the microscope is linked to a digital camera and computer to capture and store the images obtained. The digital images can then be manipulated using imaging processing software to obtain valuable qualitative and quantitative information about floc structure (Russ 2004). 332 Food Emulsions: Principles, Practices, and Techniques If the droplets in an emulsion are clearly separated and evenly distributed across the image, then one can assume that the original emulsion was nonflocculated (Figure 7.22). On the other hand, if the droplets are closely associated with one another and clump together in certain regions, then one can assume that the original emulsion was flocculated. It should be noted that in emulsions where the initial droplets are relatively small (d < 1 μm), it may be difficult to observe the individual droplets clearly, but it may still be possible to detect flocculation because the flocs themselves are relatively large (d > 1 μm). By observing the overall microstructure of a flocculated emulsion, it is often possible to obtain useful information about the flocs, for example, their size, morphology, and internal packing. This knowledge can often be used to infer information about the strength and nature of the attractive forces between the individual droplets. Flocs tend to form more compact regularly shaped structures when the attractive forces are relatively weak, whereas they form more open irregular shaped structures when the attractive forces are strong because there is less chance for droplet rearrangement after aggregation. Thus, it may be possible to distinguish between bridging flocculation (strong interactions) and depletion flocculation (weak interactions) by observing overall floc microstructure. Alternatively, it may be possible to gain information about the strength of the attractive interactions by monitoring the change in floc microstructure when a well-defined shear force is applied to an emulsion using specially designed rheo-optical cells (van der Linden et al. 2003). One must always be careful in interpreting microstructural images of flocculated emulsions since (Aguilera et al. 1999) (1) sample preparation may alter the characteristics of the flocs so that they are not representative of those present in the original sample; (2) a two-dimensional image of a threedimensional object may be misleading; and (3) in concentrated systems it is often difficult to determine whether droplets are just in close proximity or whether they are flocculated. Laser scanning confocal fluorescence microscopy is a particularly powerful method of providing information about floc structures because it can generate three-dimensional images by compiling a sequence of two-dimensional slices of the emulsion (Loren et al. 2007). Appropriate image analysis software tools can then be used to obtain quantitative information about floc characteristics, such as their size, morphology, and internal packing. Particle sizing techniques (see Section 7.5.4.2) are usually capable of providing more rapid information about particle aggregation than microscopy techniques. Nevertheless, they have the limitation that they cannot directly distinguish between different types of particle aggregation, for example, growth of individual droplets (coalescence or Ostwald ripening) versus association of a number of droplets (flocculation). Microscopy techniques are particularly good at distinguishing different aggregation mechanisms, since the particles (individual droplets or flocs) can be observed directly. Consequently, it is usually good practice to combine particle size analysis measurements with microscopy measurements. 7.5.4.2 Particle Sizing Methods Flocculation can be monitored indirectly by measuring the change in the particle size distribution with time using a particle sizing instrument (McClements 2007), for example, light scattering, ultrasonic spectrometry, NMR, or electrical pulse counting (Chapter 14). These instruments often provide a fairly qualitative indication of the extent of flocculation in an emulsion, because of a lack of suitable mathematical models to convert the experimental measurements into physical characteristics of the flocs, for example, internal packing, bond strength, and dimensions. For example, light scattering techniques usually assume that the scattering particles in a sample are isolated homogeneous spheres, whereas flocs may be interacting, inhomogeneous nonspherical aggregates, and therefore the theories normally used to interpret light scattering data are no longer valid. In many systems, it is important to determine whether the increase in droplet size is caused by flocculation, coalescence, or Ostwald ripening. The simplest method of achieving this is to alter the emulsion in a way that would be expected to break down any flocs that are present. If there are no flocs present, the particle size will not change after the alteration, but if there are flocs present there will be a decrease in the particle size. A variety of methods are available for breaking down flocs (1) altering solvent conditions, such as pH, ionic strength, polarity, or temperature; (2) applying mechanical agitation, such as stirring or sonication; and (3) adding a more surface active agent, such as a small molecule surfactant, which displaces the original emulsifier from the droplet interface but does not cause flocculation itself. Emulsion Stability 333 The choice of a suitable method depends on the nature of the emulsion, and in particular on the type of emulsifier used to stabilize the system. In emulsions where the interfacial layer consists of proteins that are held together by extensive intermolecular disulfide bonds, it may be necessary to use a combination of mercaptoethanol (to break the disulfide bonds) and a small molecule surfactant (to displace the proteins and stabilize the droplets). As with microscopy, the preparation of samples for analysis using particle sizing instruments often disturbs the structures of the flocs. For example, emulsions often have to be diluted and stirred before they can be analyzed, which can cause disruption of the flocs, particularly when there is only a weak attraction between the droplets. In addition, it is important to carry out dilution using a solvent that has similar properties to the continuous phase in which the droplets were originally dispersed (e.g., ionic strength, pH, and temperature), otherwise the floc structure may be altered. The dispersion of droplets by stirring may also cause some of the flocs to break down, particularly when they are only held together by weak attractive forces. Many of these problems can be overcome using modern particle sizing techniques based on ultrasonics, electroacoustics, NMR, or dielectric spectrometry, because these instruments can be used to analyze concentrated emulsions without the need for any sample preparation (Chapter 14). As mentioned previously, flocculation causes an increase in the viscosity of an emulsion and may eventually lead to the formation of a particle gel (Dickinson 2013). As a consequence, flocculation can often be monitored by measuring the change in viscosity or shear modulus of an emulsion. An indication of the strength of the attractive forces between flocculated droplets can be obtained from measurements of the viscosity or shear modulus versus shear stress (Pal 1996, Quemada and Berli 2002). As the shear stress is increased the flocs become deformed and disrupted so that the viscosity or shear modulus decreases. The stronger the attractive forces between the flocculated droplets, the higher the shear stress needed to disrupt them. Another indirect measurement of the degree of flocculation in an emulsion is to determine the rate at which droplets cream or sediment. As mentioned earlier, flocculation may either increase or decrease the creaming rate depending on the droplet concentration and the structure of the flocs formed. Experimental methods that can be used to monitor creaming and sedimentation were discussed in Section 7.3.3. 7.5.4.3 Bulk Physicochemical Properties Measurements of changes in the bulk physicochemical properties of emulsions can often be used to provide useful information about flocculation in food emulsions (McClements 2007). For example, rheology measurements can be used to follow floc formation in emulsions (Quemada and Berli 2002). The amount that the viscosity increases depends on the fraction of droplets involved in floc formation, the size of the flocs formed, the packing of the droplets within the flocs, and the strength of the attraction between the droplets. Measurements of the decrease in emulsion viscosity with increasing applied shear stress have been used to obtain information about the strength of the attractive forces between droplets in flocs (Quemada and Berli 2002). The applied shear stress at which the emulsion viscosity falls steeply increases as the strength of the attractive forces increases (Figure 7.28). Measurements of the creaming profile of the particles in an emulsion can also be used to monitor flocculation (Chanamai and McClements 2000). Analytical methods capable of measuring the vertical distribution of droplets in an emulsion can be used to monitor flocculation, for example, visual observation, light scattering, ultrasonic velocity measurements, or NMR imaging (Section 7.3). In dilute oil-in-water emulsions, the creaming velocity usually increases when flocculation occurs because of the resultant increase in mean particle size (Chanamai and McClements 2000). In concentrated oil-in-water emulsions, flocculation may actually retard or prevent creaming due to the formation of a three-dimensional network of aggregated droplets. Measurements of the droplet concentration in the creamed layer of an oil-in-water emulsion that undergoes flocculation can provide valuable information about the packing of the droplets in the flocs. Droplets that are held together by weak attractive forces tend to pack closely together so that the droplet concentration in the cream layer is relatively high, leading to the formation of a relatively thin cream layer. On the other hand, when the droplets are held together by strong attractive forces, they tend to pack less closely together so that the droplet concentration in the cream layer is relatively low, leading to the formation of a relatively thick cream layer. Thus, one would expect fairly 334 Food Emulsions: Principles, Practices, and Techniques thin cream layers to form for depletion flocculation (where the attractive forces are relatively weak) and fairly thick layers for bridging flocculation (where the attractive forces are relatively strong). 7.6 Coalescence Coalescence is the process whereby two or more liquid droplets merge together to form a single larger droplet (Figure 7.29). It is the principal mechanism by which an emulsion moves toward its most thermodynamically stable state because it involves a decrease in the contact area between the oil and water phases (Section 7.2). Coalescence causes emulsion droplets to cream or sediment more rapidly because of the increase in their size (Section 7.3). In oil-in-water emulsions, coalescence eventually leads to the formation of a layer of oil on top of the material, which is referred to as oiling off. In water-in-oil emulsions, it leads to the accumulation of water at the bottom of the material. An understanding of the factors that influence coalescence is therefore important to food manufacturers attempting to create products with extended shelf-lives. 7.6.1 Physical Basis of Coalescence Coalescence is the result of the liquid within two or more emulsion droplets coming into molecular contact (Kabalnov and Wennerstrom 1996, Evans and Wennerstrom 1999, Walstra 2003, Sanfeld and Steinchen 2008, Tcholakova et al. 2008). This process can only occur when droplets are in close proximity and the thin film of material separating them is ruptured.* The fact that the droplets must be in close contact means that coalescence is much more dependent on short-range forces and the precise molecular details of a system, than either gravitational separation or flocculation. The rate at which coalescence proceeds and the physical mechanism by which it occurs are therefore highly dependent on the nature of the emulsifier used to stabilize the system. For this reason, our knowledge of coalescence is much less well developed than that of the other major forms of emulsion instability. Even so, a combination of theoretical and experimental studies has led to a fairly good understanding of the major factors that influence coalescence in some important systems. In general, the susceptibility of droplets to coalescence is determined by physical mechanism responsible for droplet encounters (e.g., Brownian motion, simple shear, turbulence, and gravity), the nature of the interactions between droplets (i.e., colloidal and hydrodynamic forces), and the resistance of the thin film of material separating the droplets to rupture. FIGURE 7.29 Droplet coalescence leads to a growth in the mean droplet diameter and may eventually lead to complete separation of the oil and aqueous phases. * This material consists of the continuous phase, plus any interfacial layers surrounding the droplets. 335 Emulsion Stability 7.6.1.1 Physical and Molecular Processes Associated with Coalescence The most important physical and molecular processes that must occur in order for two droplets to coalesce were discussed earlier, that is, droplet encounters, film thinning, film formation, and film rupture (Figure 7.16, Section 7.4). In this section, we focus on some additional factors that are particularly important for determining the stability of droplets to coalescence. 7.6.1.1.1 Droplet Deformation In certain systems, emulsion droplets become deformed as they approach each other, which has a pronounced influence on their stability to coalescence (Ivanov et al. 1999, Walstra 2003, Sanfeld and Steinchen 2008). When the film separating the droplets has thinned to a certain level, the surfaces of the droplets become flattened (Figure 7.30). Droplet flattening occurs when the external forces acting upon the droplets exceed the internal forces responsible for holding the droplets into a spherical shape. The internal forces are a result of the Laplace pressure (as well as the resistance of the interfacial layers to deformation), whereas the external forces may have a variety of origins, including colloidal, hydrodynamic, mechanical, and gravitational forces. The tendency for droplets to become deformed can be described by a Weber number: We = s EXTr 2 2 gh (7.34) Here σEXT is the external force r is the droplet radius γ is the interfacial tension h is the surface-to-surface separation For We < 1 the droplets tend to remain spherical, but for We > 1 they tend to be deformed and a flat film is formed between them (Figure 7.30). Expressions for the Weber number for different kinds of external stresses have been given elsewhere (Walstra 2003). Calculations using these expressions show that relatively small emulsion droplets (r < 1 μm) with interfacial tensions similar to those found in foods tend to remain spherical (unless they are centrifuged at relatively high speeds), but relatively large droplets (r > 10 μm) or droplets with low interfacial tensions are often appreciably deformed by the colloidal, hydrodynamic, gravitational, or mechanical forces in the system. Theoretical and experimental studies have shown that the overall coalescence rate of an emulsion is strongly dependent on the propensity for droplets to become deformed (Sanfeld and Steinchen 2008). It is therefore important to take this factor into account when considering droplet coalescence in food emulsions containing relatively large droplets. 7.6.1.1.2 Film Rupture Theoretical modeling of the hydrodynamic and colloidal interactions between emulsion droplets predict that certain emulsions should be stable to coalescence. For example, if there is a sufficiently high energy barrier or short-range repulsion between the droplets, the emulsions should remain indefinitely stable We < 1 We > 1 FIGURE 7.30 Under certain circumstances, emulsion droplets become deformed, for example, large diameter, low interfacial tension, and high external stresses. The increase in contact area between flattened droplets tends to promote coalescence. 336 Food Emulsions: Principles, Practices, and Techniques to coalescence. Alternatively, if the droplet surfaces are immobile, then hydrodynamic theory suggests that the droplets should not coalescence because the velocity of film thinning is proportional to the droplet separation (V ∝ h). Hence, the velocity of film thinning approaches zero as h approaches zero, so that the droplets should never actually contact each other. Nevertheless, coalescence is often observed experimentally in systems with these characteristics, which suggests that some other mechanism must be involved in initiating droplet coalescence. The propensity for coalescence to occur depends on the tendency of the thin film of liquid separating the droplet surfaces to rupture. In general, the kinetics of film rupture can be described by the following expression (Walstra 2003): æ DGFR ö fFR = f0 exp ç ÷ è kBT ø (7.35) Here f FR is the frequency of film rupture per unit area (m−2 s−1) f0 is the natural frequency (m−2 s−1) ΔG FR is the free energy change associated with causing a rupture in the film (e.g., a hole) Typically, the natural frequency of film rupture is around 1030 m−2 s−1 (Kabalnov 1998). In general, the free energy penalty associated with film rupture depends on many factors including the interfacial tension, hole size, film thickness, colloidal interactions, and mechanical properties of the interfaces. An expression for ΔG FR is given later for a simple model system where film rupture is attributed to hole formation (see Section 7.6.1.3). If droplet coalescence occurs when emulsion droplets are in prolonged contact, then film disruption is likely to be the rate limiting step for droplet coalescence. 7.6.1.2 Mechanisms of Film Rupture Before coalescence can occur it is necessary for the thin film separating the droplets to be ruptured. A number of mechanisms have been proposed to account for the rupture of this thin film depending on the nature of the system (Kabalnov 1998, van Aken et al. 2003, Tcholakova et al. 2006b). The relative importance of these different mechanisms is largely determined by the characteristics of the continuous phase separating the droplets (e.g., thickness and viscosity), and the interfacial layers surrounding the droplets (e.g., thickness, dilational modulus, interfacial tension, shear modulus, and colloidal interactions). In the absence of emulsifier at the droplet surfaces, the following mechanisms could lead to film disruption and coalescence: • Capillary wave formation: Capillary waves form spontaneously in thin films because of the thermal motion of the system (Figure 7.31). If the amplitude of the thermal fluctuations exceeds approximately half the film thickness, a point will be created where the fluids in the different droplets come into molecular contact with each other, which leads to the formation of a hole through which the dispersed phases can flow. Capillary waves are damped by repulsive interactions between droplets, high interfacial tensions, and high interfacial mechanical rigidity (Kabalnov 1998, Walstra 2003). In practice, it is believed they are principally responsible for film disruption in systems where the interfacial tensions are extremely low or where there are no repulsive interactions between droplets. • Spontaneous hole formation: Small holes may form spontaneously in a thin film because of the thermal motion of the system (Figure 7.32). If the size of these holes is below some critical value they tend to shrink and collapse, but if it is above this value they tend to grow and film rupture occurs (Kabalnov 1998). In most food emulsions, the droplets are surrounded by a layer of emulsifier molecules and there may be mechanical stresses applied to the system. The above mechanisms may also promote film disruption in these systems, but additional factors should also be considered (van Aken et al. 2003, van Aken 2004). 337 Emulsion Stability (a) (b) FIGURE 7.31 As two droplets approach each other, a thin film of continuous phase is formed between them. Once droplets get closer than a critical distance, they may spontaneously merge because thermal fluctuations in the thin film generate capillary waves that promote hole formation. (a) Film thinning and (b) capillary wave formation. rH,o Dispersed phase Continuous phase rH,i Dispersed phase FIGURE 7.32 A hole may spontaneously form in the thin film separating emulsion droplets leading to coalescence. • Insufficient emulsifier: If there is insufficient emulsifier present in a system to completely cover the oil–water interfaces present, then there will be gaps in the interfacial layers surrounding the droplets. Coalescence could then occur if two gaps on different droplets came into close proximity, for example, due to spontaneous hole or capillary wave formation. This type of coalescence is likely to be most important during homogenization where new surfaces are continually being created by the intense forces generated within a homogenizer. • Film stretching: If a sufficiently large stress is applied parallel to an emulsifier-coated interface, then some of the emulsifier molecules may be dragged along the interface, leaving some regions where there is excess emulsifier and other regions where there is depleted emulsifier. Coalescence could then occur if two emulsifier-depleted regions on different droplets came into close proximity during a droplet–droplet encounter. This process is only likely to be important if the adsorption of the emulsifier is relatively slow compared to the duration of the applied stresses and droplet encounter frequency, otherwise emulsifier would have time to adsorb to the droplet surfaces and cover the gaps. Film stretching is likely to be important in emulsions that are subjected to intense mechanical stresses, especially if the droplets are in close proximity, for example, in flocculated or concentrated emulsions. • Film tearing: If a sufficiently large stress is applied parallel to an interface that is comprised of a highly cohesive layer of emulsifier molecules, then it may cause the interfacial layer to tear, leaving exposed emulsifier-depleted patches that promote coalescence. This mechanism is likely to be important in systems where the interfacial layers are highly cohesive (e.g., protein 338 Food Emulsions: Principles, Practices, and Techniques layers with extensive cross-linking), particularly under high applied mechanical stresses, for example, shearing, homogenization, or centrifugation. It is likely to be less important in dilute, nonflocculated, and quiescent emulsions because the forces generated in these systems are not strong enough to tear the interfacial layers. 7.6.1.3 Hole Formation In this section, we focus on the spontaneous formation of a hole in the thin layer of material separating two emulsion droplets. Hole formation involves overcoming a free energy penalty associated with creating a hole in the thin layer (Kabalnov 1998). A physical model, known as the de Vries theory, has been developed to calculate the magnitude of the free energy change associated with the formation of a hole in a thin film. Hole formation involves two contributions that cause changes in the contact area between the oil and water phases (1) there is an increase in contact area inside the film due to hole formation and (2) there is a decrease in contact area at the planar edges of the film due to hole formation (Figure 7.32). The first contribution dominates for small holes, whereas the second contribution dominates for large holes. Since increasing the contact area between the oil and water phases is thermodynamically unfavorable, then there is an increase in free energy associated with hole formation for small holes, but a decrease for large holes. A geometrical analysis of hole formation has led to the following approximate expression for the overall free energy change associated with hole formation (Kabalnov 1998): DGH = -2pg ëérH2 - (p - 2)rH h - (p - 3)h2 ùû (7.36) Here γ is the interfacial tension r H is the hole radius h is the film thickness Calculations of the free energy associated with hole formation versus the hole radius are shown in Figure 7.33. Initially, there is an increase in free energy with increasing hole size, until a maximum value Spontaneous growth ∆GH* 20 ∆G/kT 15 10 5 Formation-collapse rH,i* 0 0 0.25 0.5 0.75 rH,i (nm) 1 1.25 FIGURE 7.33 There is a free energy penalty (ΔG H) associated with hole formation that determines the frequency of film rupture. Holes only grow once they exceed a critical size (rH* ), otherwise they shrink and disappear. 339 Emulsion Stability (∆GH* ) is reached at a critical hole radius (rH* ), after which there is a decrease in free energy with further increase in hole size. The critical hole radius occurs at 57% of the film thickness. If a hole is spontaneously formed that is below the critical hole radius (rH < rH* ), then the hole will tend to collapse, but if a hole is spontaneously formed that is above the critical hole radius (rH > rH* ), then the hole will expand, the film will rupture and coalescence will occur. The propensity for hole formation to occur is strongly influenced by the presence of emulsifiers at the oil–water interfaces. First, emulsifiers decrease the interfacial tension, thereby reducing the free energy associated with altering the contact area between the oil and water phases when a hole is formed. Second, there are additional contributions to hole formation associated with the rheology of the interfacial layers, for example, the bending energy and dilational modulus (Kabalnov 1998, Ivanov et al. 1999). For example, the formation of a hole in the thin film that separates the droplets depends on the development of a highly curved edge (Figure 7.34). If the curvature of the edge is close to the optimum curvature of the emulsifier layer (H0 ≈ Hedge), then the formation of the hole is thermodynamically favorable. On the other hand, if the optimum curvature of the emulsifier is opposite to that of the edge, then the formation of a hole is thermodynamically unfavorable, and free energy will need to be expended to bend the interface to the correct curvature. The dependence of hole formation on the optimum curvature of an interfacial layer means that coalescence is related to the molecular geometry of emulsifier molecules (Chapter 4). The relationship between the molecular geometry and coalescence stability of oil-in-water and water-in-oil emulsions is highlighted in Figure 7.34. When two oil droplets are separated by a thin film of water (as in an oil-in-water emulsion), then hole formation will be much more likely for a surfactant with a high packing parameter (p > 1), than a surfactant with a low packing parameter (p < 1). In summary, the coalescence rate tends to increase as the interfacial tension and/or rigidity of the interfacial layers formed by the adsorbed emulsifier molecules decrease. 7.6.1.4 Rate-Limiting Step for Coalescence The above discussion of the physicochemical processes that may occur when droplets approach each other suggests that droplet coalescence can be divided into different categories depending on the rate limiting step. Molecular geometry Optimum curvature Hole formation Favorable Unfavorable FIGURE 7.34 Coalescence occurs more rapidly when the optimum curvature of a surfactant is similar to that of the edge of the hole formed in the thin film. Thus, coalescence is more favorable in oil-in-water emulsions when the surfactant tail group cross-sectional area is greater than that of the head group 340 Food Emulsions: Principles, Practices, and Techniques 7.6.1.4.1 Coalescence Immediately after a Collision In this situation, coalescence occurs immediately after two or more droplets encounter each other during a collision. This type of coalescence is likely to dominate in emulsions where the droplets are free to move around and collide with each other, and there is no deep secondary minimum, high energy barrier, or strong short-range repulsion. Droplets will tend to exist as individual droplets or as coalesced droplets, without substantial flocculation occurring. The rate at which this type of coalescence occurs is governed by many of the same factors as flocculation, that is, the collision frequency and efficiency (Section 7.5.1). The collision frequency is determined by the dominant mechanism responsible for droplet movement, for example, Brownian motion, gravity, or applied mechanical forces (Section 7.5.1.1). The collision efficiency is determined by the probability that the droplets can jump over the energy barrier. If the energy barrier is small, then the coalescence collision efficiency is high (EC → 1), which may occur when there is no emulsifier present and no electrostatic repulsion between the droplets. If the energy barrier is relatively large, the collision efficiency may be extremely low (EC → 0), which is likely to occur in the presence of an emulsifier or if the droplets have an appreciable electrical charge. This type of coalescence is only likely to be significant in systems in which there is insufficient emulsifier present to completely saturate the droplet surfaces or during homogenization where emulsifiers may not have sufficient time to cover the droplet surfaces before a droplet–droplet collision occurs (Chapter 6). It may also be important in emulsions that are subjected to high shear forces, because the impact forces that act on the droplets as the result of a collision may be sufficient to cause them to jump over the energy barrier and into the primary minimum. This type of coalescence depends on the collision frequency and therefore follows second-order kinetics (Walstra 2003). 7.6.1.4.2 Coalescence from the Secondary Minimum In this situation, coalescence occurs after the emulsion droplets have been trapped in the secondary minimum for a certain period, that is, the droplets are already in fairly close proximity. The rate of coalescence depends on how quickly the droplets can move from the secondary to the primary minimum over the energy barrier, which is governed primarily by the height of the energy barrier (Petsev 2000, Toro-Mendoza and Petsev 2010). This type of coalescence is likely to be important in emulsions in which there is a relatively deep secondary minimum, and no strong short-range repulsion to prevent the droplets from coalescing once they have moved into the primary minimum. The rate limiting step for this type of coalescence is the average time taken for the droplets to move from the secondary to primary minimum, τ2 → 1. This type of process has first-order kinetics (Walstra 2003). 7.6.1.4.3 Coalescence from the Primary Minimum In this situation, coalescence spontaneously occurs after the droplets have been in contact for a prolonged period in the primary minimum. This type of coalescence is likely to be important in systems in which there is a relatively strong short-range repulsion between the droplets, which prevents them from coalescing immediately after moving over the energy barrier and into the primary minimum. The rate limiting step for this type of coalescence is the average time taken for film rupture to occur, that is, τFR. This type of process has first-order kinetics (Walstra 2003). This mechanism is likely to be the most important in food systems, because they usually have an interfacial layer that generates very strong short-range repulsion between the droplets. Coalescence from the primary or secondary minimum occurs when the droplets are in close proximity for extended periods, and is therefore particularly important in emulsions that have high droplet concentrations, that contain flocculated droplets, or that contain droplets that have accumulated at the top (or bottom) of the sample due to creaming (or sedimentation). In these situations, it is no longer convenient to describe the rate at which coalescence occurs in terms of a collision frequency and efficiency, because the droplets are already in contact with each other. Instead, it is more useful to characterize these types of coalescence in terms of a coalescence time, that is, the length of time the droplets remain in contact before coalescence occurs. In practice, all of the droplets in an emulsion do not usually coalesce after the same time, and therefore it is more appropriate to define an average coalescence time or to stipulate a range of times over which the majority of the droplets coalesce. 341 Emulsion Stability 7.6.1.5 Modeling Droplet Growth due to Coalescence The most appropriate mathematical model to describe coalescence in a particular system depends on the physicochemical phenomenon that is the rate limiting step for coalescence in that particular system, for example, droplet encounters, movement from primary to secondary minimum, or film rupture. In most food emulsions, it is likely that film rupture is the rate limiting step because of the relatively strong shortrange repulsion generated by the emulsifiers. An approximate expression for the change in the mean droplet size with time in an emulsion due to coalescence has been derived assuming that the droplets are densely packed together and film rupture is a random process (Kabalnov 1998): 1 1 Z = - ft d 2 d02 3 (7.37) Here d is the droplet diameter after time t d 0 is the initial droplet diameter Z is the number of neighbors in contact with each droplet (∼6) f is the frequency of film rupture This equation can be used to estimate the time required for complete coalescence to occur (i.e., d → ∞): τ ≈ 1/(2d 02f), assuming Z = 6. Thus, the coalescence time decreases with increasing initial droplet size and rupture frequency. In general, the above equation has to be modified to take into account the fact that the droplets are not densely packed together, that is, the contact area between droplets is often only a relatively small fraction of their overall surface areas. If it is assumed that the droplets remain spherical during coalescence, and that the thickness of the thin film remains constant, then the following expression can be derived for the change in droplet diameter with time based on the approach described by Kabalnov (1998): 1 1 Z = fph*t d d0 12 (7.38) Here, h* is the film thickness, which is taken to be the surface-to-surface droplet separation at the primary minimum. This equation can also be used to predict the time for complete coalescence to occur (i.e., d → ∞): τ ≈ 1/(d0  fh*), assuming Z = 6. Coalescence should therefore occur more rapidly with increasing initial droplet size, rupture frequency, and film thickness (although it should be noted that the rupture frequency decreases with film thickness). The above equations suggest that the mean droplet size increases steadily with time during coalescence, but they give no indication of the expected change in the particle size distribution over time. Experimental studies of coalescence in concentrated oil-in-water emulsions stabilized by small molecule surfactants have shown that coalescence may proceed by either a homogeneous or a heterogeneous process depending on the system composition and environmental conditions (Demiere et al. 1998). In the homogeneous process all the droplets in the emulsion grow uniformly throughout the different regions of the system, but in the heterogeneous process a few of the droplets grow very rapidly, whereas the rest remain approximately the same size as in the original system (Figure 7.35). Homogeneous coalescence therefore tends to lead to the formation of monomodal droplet size distributions with the peak moving upward with time, whereas heterogeneous coalescence tends to lead to the formation of bimodal droplet size distributions with the fraction of droplets in the larger peak growing with time. Heterogeneous coalescence may lead to extensive “oiling off,” where an oil layer is formed on top of the emulsion. More sophisticated physical models than the ones discussed above are therefore required to model the change in the particle size distribution with time due to coalescence. Finally, it should be noted that if the rate limiting step for coalescence is droplet encounters, rather than film rupture, then the equations developed for droplet flocculation can be used to predict the change in mean particle size with time. 342 Food Emulsions: Principles, Practices, and Techniques Coalescence rate independent of droplet size Volume frequency Homogeneous coalescence Time Coalescence rate increases with droplet size T2 T1 1 10 0 100 Particle diameter Volume frequency Heterogeneous coalescence 0.1 T3 T3 T2 T1 0 0.1 1 10 100 Particle diameter FIGURE 7.35 Droplet coalescence may occur by a homogeneous mechanism when the coalescence rate is largely independent of droplet size (which leads to monomodal distributions), or a heterogeneous mechanism when the coalescence rate increases with droplet size (which leads to bimodal distributions). Here the designations 0 through T3 represent a progressive increase in storage time. 7.6.2 Methods of Controlling Coalescence The rate at which coalescence occurs is strongly dependent on the colloidal and hydrodynamic interactions between the droplets, as well as the physicochemical properties of the interfacial layers that coat the droplets. As a consequence, the most appropriate method of controlling coalescence is highly dependent on the type of emulsifier used to stabilize the system, as well as the prevailing environmental conditions (such as pH, ionic strength, solvent quality, and temperature). Even so, it is possible to give some general advice about the most effective methods of avoiding coalescence. These methods can be conveniently divided into two categories: those that prevent droplet contact, and those that prevent rupture of the interfacial layers surrounding the droplets. 7.6.2.1 Prevention of Droplet Contact The coalescence rate can be decreased by reducing the length of time that droplets are in close contact. The droplet contact time can be reduced in a number of ways, including: (1) decreasing their collision frequency, (2) ensuring that they do not flocculate, (3) preventing them from forming a concentrated layer at the top (or bottom) of an emulsion due to creaming (or sedimentation), and (4) ensuring that the droplet concentration in the emulsion is not so high that the droplets become close packed. Even when the droplets are in contact with each other for extended periods (e.g., in flocs, creamed layers, or concentrated emulsions), they can still be prevented from coalescing by ensuring that they do not get too close. The probability of coalescence occurring increases as the thickness of the layer of continuous phase separating them decreases, because thermal fluctuations may then become large enough to form a hole that extends from one droplet to another (Figure 7.32). The thickness of this layer is determined principally by the magnitude and range of the various attractive and repulsive forces that act between the droplets (Chapter 3). The coalescence stability can therefore be enhanced by ensuring there is a sufficiently strong repulsive interaction to prevent the droplets from coming into close contact. This can be achieved in a number of ways, including varying the emulsifier type, pH, ionic strength, or temperature. Emulsion Stability 343 7.6.2.2 Prevention of Rupture of Interfacial Layers The rupture of the thin film between the droplets depends on changes in its shape caused by thermal energy or applied mechanical forces (Demiere et al. 1998, Kabalnov 1998, Evans and Wennerstrom 1999, van Aken 2004). The magnitude of these changes is governed by the interfacial tension and rheology of the interfacial layer. The lower the interfacial tension or rheology, the more mobile is the interface and therefore the more likely a hole will form that leads to film rupture. Consequently, coalescence becomes less likely as the interfacial tension or the viscoelasticity of the interfacial layer increases (Dickinson 1992a). The thickness of the interfacial layers around the droplets is also likely to influence the coalescence stability of emulsions. Droplets with thicker layers would be expected to provide the greatest stability because they are less likely to be ruptured and they provide a greater steric repulsion. For small molecule surfactants, it is important to select an emulsifier that has an optimum curvature that does not favor coalescence under the conditions used (Section 7.6.1.3). Some nonionic surfactants stabilize droplets against coalescence at low temperatures, but coalesce when they are heated due to dehydration of the head groups changing the curvature (Saberi et al. 2013). The likelihood of a hole forming somewhere in a thin film increases as its overall area increases. Consequently, the larger the contact area between two droplets, the greater the coalescence rate. The coalescence rate therefore increases as the size of the droplets in an emulsion increases, or when the droplets become flattened against one another. Droplet flattening tends to occur in highly concentrated emulsions, in creamed layers, or when emulsions are subjected to external forces (Walstra 1983). Large droplets are more prone to flattening than smaller droplets because the interfacial forces that tend to keep them in a spherical shape are lower (Chapter 6). Large droplets are also more prone to collision-induced coalescence because the impact forces generated during a collision are greater, and the magnitude of the attractive forces between the droplets is larger. On the other hand, increasing the size of the droplets decreases the frequency of the encounters between droplets, which may be the dominant effect in emulsions where the rate limiting step is the collision frequency. Emulsion droplets stabilized by relatively thick cohesive interfacial layers (e.g., proteins and polysaccharides) are relatively resistant to coalescence under quiescent conditions, but may become unstable when mechanical forces are applied to the system (e.g., intense stirring, homogenization, or centrifugation), especially in concentrated systems (van Aken 2004). This instability has been attributed to stretching and tearing of the interfacial layers surrounding the droplets, which leads to the exposure of emulsifier-depleted regions on the droplet surfaces. 7.6.3 Factors Affecting Coalescence 7.6.3.1 Emulsifier Type Protein emulsifiers have been found to be extremely effective at providing protection against coalescence, especially under quiescent conditions (van Aken 2004). The main reason for this is that proteins are capable of producing emulsions with small droplet sizes, they provide strong repulsive forces between droplets (due to a combination of electrostatic and steric interactions), the interfacial tension is relatively high, and they form interfaces that are highly resistant to rupture. Emulsion droplets stabilized by polysaccharides are also highly stable to coalescence for the same reasons. Partially hydrolyzed proteins are less effective at preventing coalescence because they tend to form thinner, less viscoelastic interfacial layers that are easier to rupture (Euston et al. 2001, Gallier et al. 2013). Extensive coalescence has been observed in protein-stabilized emulsions when they are subjected to mechanical stresses, such as shear, elongational, or turbulent flow (van Aken et al. 2003, van Aken 2004). Mechanical stresses may cause the adsorbed proteins to undergo extensive clumping at the droplet surfaces. Clump formation may lead to exposure of oil patches in the interfacial layers that promote droplet coalescence. Alternatively, strong intermolecular interactions between proteins adsorbed onto different droplets may lead to “tearing” of the interfacial layers when the emulsions are exposed to mechanical stresses and the droplets are separated. Presumably, this tearing process would lead to the 344 Food Emulsions: Principles, Practices, and Techniques formation of oil patches in the interfacial layers that were temporarily not covered by protein, thereby making the droplets more susceptible to coalescence. Coalescence of emulsions stabilized by small molecule surfactants may be governed by their ability to keep droplets apart, as well as the resistance of the interfacial layers to rupture. Nonionic surfactants, such as the Tweens, have polymeric hydrophilic head groups that typically provide a large steric overlap and hydration repulsion at ambient temperatures, thereby inhibiting coalescence (Section 3.5). However, when they are heated the head groups become progressively dehydrated, which reduces the hydration repulsion and also changes the interfacial optimum curvature and flexibility. Above a certain temperature the interfacial tension becomes low and the interfacial flexibility becomes high, which promotes droplet coalescence. Ionic surfactants, such as SDS and fatty acids, generate strong electrostatic repulsive forces that can prevent droplets from coalescing (Section 3.4). Nevertheless, this electrostatic repulsion is only appreciable when the aqueous phase has a low ionic strength. At high ionic strengths, the electrostatic repulsion is screened by the counter-ions, so that the droplets come close together, and are prone to coalescence, because the interfacial tension becomes low and the interfacial flexibility becomes high, which promotes coalescence. 7.6.3.2 Influence of Environmental Conditions Food manufacturers often need to create emulsions with extended shelf-lives and so it is important for them to understand the influence of various types of processing and storage condition on droplet coalescence. As mentioned above, emulsions stabilized by milk proteins are fairly stable to coalescence under quiescent conditions, provided that the droplets are completely liquid (van Aken et al. 2003, van Aken 2004). However, droplet coalescence may occur when the droplets are subjected to shear forces or brought into close contact for extended periods, for example, in a creamed layer, a floc, or a concentrated emulsion. The presence of small amounts of low-molecular weight surfactants in the aqueous phase greatly enhances the tendency of emulsion droplets to coalesce during shearing (Chen et al. 1993, Dickinson et al. 1993). The stability of emulsions to shear forces therefore depends on the structure and properties of the adsorbed interfacial layers. Centrifugation of an emulsion may also lead to extensive coalescence because the droplets are forced together into a compact droplet-rich layer with sufficient force to flatten the droplets and disrupt the interfacial layers (Tcholakova et al. 2005, Cheetangdee et al. 2011). When an oil-in-water emulsion is frozen, only part of the water is initially crystallized and the oil droplets are forced into the remaining liquid region (Figure 7.36) (McClements 2012, Degner et al. 2014). The ionic strength of this region is increased significantly because of the concentration of salts and other components. The combination of forcing the droplets into a more confined space and of altering the solvent conditions is often sufficient to disrupt the droplet membranes and promote coalescence once an emulsion is thawed. In addition, the oil droplets may crystallize during the freezing process, which can lead to emulsion instability through partial coalescence (Section 7.7). Under certain circumstances, freezing can cause cold denaturation of proteins, which may lead to a reduction in their functionality (Walstra 2003). Finally, freezing of the water phase may lead to dehydration of the emulsifier molecules adsorbed to the surface of the droplets, which promotes droplet–droplet interactions. There are many factors that contribute to the instability of emulsions during freezing and thawing, and the development of freeze–thaw stable emulsions still remains a major challenge to food scientists (Degner et al. 2014). Coalescence may also be promoted when an emulsion is dried into a powder, because drying may disrupt the integrity of the interfacial layer surrounding the droplets, for example, during freeze or spray drying (Taneja et al. 2013), which leads to coalescence once the emulsion is reconstituted. The coalescence stability can often be improved by adding relatively high concentrations of proteins or carbohydrates to the system prior to drying. These molecules form a wall material around the oil droplets that keeps them from coming into close contact during the dehydration process (Figure 7.37). The coalescence of droplets in an emulsion may also be influenced by various chemical or biochemical changes that occur over time. Lipid oxidation leads to the development of surface active reaction products that may be capable of displacing emulsifier molecules from the droplet surface and thereby 345 Emulsion Stability Cool Cool Cool Cool Cool Heat Cool Cool Heat (a) (b) (c) FIGURE 7.36 The droplets in an oil-in-water emulsion may coalesce due to crystallization of the water and/or oil phases: (a) partial coalescence may occur when the fat phase crystallizes; (b) the fat phase may crystallize first, and then the water phase (forcing the droplets together); (c) the water phase may crystallize first (forcing the droplets together) and then the fat phase crystallizes. These emulsions may coalesce once the system is heated to melt the fat and water phases. 0wt% sucrose 20wt% sucrose Wall matrix FIGURE 7.37 Influence of dehydration on the stability of oil-in-water emulsions in the absence and presence of a wall material (sucrose). Micrographs show emulsion structure after freeze-drying and then rehydration. In the absence of a wall material, droplets are forced together leading to coalescence when they are rehydrated. 346 Food Emulsions: Principles, Practices, and Techniques promoting coalescence (Coupland and McClements 1996). Extensive enzymatic hydrolysis of proteins or polysaccharides could cause an interfacial layer to be disrupted (Singh 2011), again promoting droplet coalescence. An understanding of the various chemical and biochemical factors that determine coalescence is therefore essential in creating emulsions with extended shelf-lives or that can be broken down under specific conditions. 7.6.3.3 Influence of Impurities and Surfaces In many food emulsions, droplet coalescence is promoted by the presence of impurities and surfaces, for example, gas bubbles, solid particles, crystals, and surfaces (van Aken 2004). For example, coalescence can be promoted when the droplets are in close proximity to a fluid or solid surface, provided the dispersed phase is capable of wetting the surface (i.e., the contact angle is lower than 90°). This type of coalescence is believed to be important in aerated oil-in-water emulsions, where the oil droplets become coalesced when they spread around air bubbles, for example, in whipped cream (van Aken et al. 2003, Hotrum et al. 2005). Surface-induced droplet coalescence may also be promoted when emulsion droplets are confined between thin moving surfaces, for example, in a colloid mill during homogenization or in the mouth during mastication (van Aken 2004, Dresselhuis et al. 2008). Finally, droplet coalescence may be promoted by the presence of solid particles or crystals due to their ability to disrupt the thin film separating the droplets, especially during shearing, for example, fat, ice, sugar, or salt crystals. 7.6.4 Measurement of Droplet Coalescence Experimental characterization of coalescence can be carried out using a variety of analytical techniques, many of which are similar to those used to monitor flocculation (Section 7.5.4). 7.6.4.1 Microscopy Methods The most direct method for observing droplet coalescence in an emulsion is to use optical microscopy (Mikula 1992, McClements 2007). An aliquot of emulsion is placed on a microscope slide and the change in droplet size distribution is measured as a function of time or after some specific treatment by monitoring the number and size of the droplets. This can be done manually or most usually by using a computer with suitable image processing software. It is possible to observe individual coalescence events using a high-speed camera, but these events are usually so unlikely in food emulsions that they are difficult to catch (Dickinson 1992). In some cases, coalescence can be observed directly using a microscope when an emulsion is subjected to destabilizing stresses, such as shearing, freezing, or aeration. The change in droplet size distribution of an emulsion during storage can also be measured using other forms of microscopy, such as confocal laser scanning microscopy, electron microscopy, or atomic force microscopy. An alternative optical microscopy method involves the observation of the coalescence of individual emulsion droplets with a planar oil–water interface (Dickinson et al. 1988). An oil droplet is released from a capillary tube into an aqueous solution and moves upward to the oil–water interface due to gravity (Figure 7.38). The time (tC) taken for the oil droplet to merge with the planar interface after it has arrived there is determined by observing it using an optical microscope. Typically, one observes two stages in the droplet coalescence process (1) a lag phase (corresponding to film thinning) where the droplet remains at the interface but no coalescence occurs and (2) a coalescence phase where (corresponding to film disruption) where the droplet merges with the bulk liquid. Droplets exhibit a spectrum of coalescence times because film disruption is a random (stochastic) process. Consequently, there is an approximately exponential decrease in the number of noncoalesced droplets remaining at the interface with time after the lag phase. The major disadvantages of this technique are that only droplets above about 1 μm can be observed directly using an optical microscope, and that coalescence often occurs so slowly that it is impractical to monitor it continuously using a microscope. To detect coalescence over a reasonably short period, it is necessary to have relatively low concentrations of emulsifier at the surfaces of the droplets, 347 Emulsion Stability Microscope Syringe FIGURE 7.38 Microscopic technique for monitoring coalescence of single droplets at a planar oil–water interface. The time taken for a droplet to merge with the interface is determined microscopically. which is unrealistic because the droplets in food emulsions are nearly always saturated with emulsifier. Computer simulation techniques have been used to predict coalescence lifetimes measured using this method in terms of the colloidal and hydrodynamic interactions involved (Rojas et al. 2010). The film trapping technique (FTT) also involves measuring the coalescence of individual droplets at an oil–water interface using a microscope (Tcholakova et al. 2002). In this case, the oil droplets are trapped between a bulk oil phase held in a capillary tube and a bulk water phase contained within a vessel (Figure 7.39). The capillary tube is connected to a pressure control system that enables one to force the oil–water interface downward. The pressure in the capillary tube is gradually increased until coalescence of the oil droplets with the oil–water interface is observed using the microscope. The pressure where coalescence is first observed is called the critical capillary pressure ( PCCR ), which can be calculated from measurable experimental parameters (Tcholakova et al. 2002). More recently microfluidic devices have been developed to monitor the coalescence of oil droplets under well-defined flow conditions (Bremond and Bibette 2012, Krebs et al. 2013). These devices can Pressure port r Capillary tube h Water Air Oil droplets compressed against O–W interface Oil Microscope objective FIGURE 7.39 Observation of single oil droplets compressed against an oil–water interface can be used to quantify droplet coalescence. 348 Food Emulsions: Principles, Practices, and Techniques be used to explore the influence of different parameters on the coalescence stability of droplets, such as emulsifier type, dispersed and continuous phase viscosities, droplet size, solution pH and ionic strength, and flow rates. 7.6.4.2 Particle Sizing Methods Droplet coalescence can also be monitored by measuring the time dependence of the particle size distribution using particle sizing methods, such as light scattering, electrical pulse-counting, ultrasonic spectrometry, or NMR (McClements 2007) (Chapter 14). These instruments are usually simple to operate, highly automated, and provide measurements of the size of a large sample of droplets within a few minutes or less. Nevertheless, there are some important practical factors that need to be considered when analyzing coalesced systems. In oil-in-water emulsions that contain relatively large oil droplets (d > 10 μm) or a layer of free oil (“oiling off”), one must be careful to analyze a representative sample. Large particles cream faster than small particles and so there may be a vertical distribution of droplets within an emulsion that changes over time. This phenomenon will occur in the original emulsion from which the sample to be analyzed is collected. If the emulsion is not made homogeneous before sample collection (e.g., by stirring or agitation), then the measured particle size will depend on where the sample was collected from (e.g., the bottom, middle, or top). Rapid creaming of large droplets may also occur within the measurement chamber of a particle sizing instrument. Consequently, the large droplets move to the top of the measurement chamber and may not be detected by the instrument, for example, they move out of the laser beam in a light scattering device. On the other hand, if there is a layer of free oil on top of an emulsion due to oiling off, then it may be converted into large droplets due to mixing and agitation steps used in sample preparation. These large droplets may be detected by the instrument, but they were not actually present in the original emulsion from which the sample was collected. In summary, particle sizing techniques tend to work fairly well in the initial stages of coalescence when the droplets are still relatively small, but may not give reliable results in systems that have undergone extensive coalescence or oiling off. It is usually important to establish the origin of an observed increase in measured particle size in an emulsion, for example, coalescence, flocculation, or Ostwald ripening (McClements 2007). Coalescence can often be distinguished from flocculation using the following procedure (1) measure the particle size distribution of the emulsion of interest, (2) add a deflocculant to break down any flocs that are present (e.g., a surfactant), and (3) measure the particle size distribution of the deflocculated emulsion. If there were no flocs present in the original emulsion, then the mean particle size remains unchanged, otherwise it decreases. Coalescence is more difficult to distinguish from Ostwald ripening because they both involve an increase in the average size of the individual droplets with time (Urbina-Villalba et al. 2009). Nevertheless, it is sometimes possible to distinguish between coalescence and Ostwald ripening by measuring the full particle size distribution with time. Coalescence usually leads to a bimodal distribution (heterogeneous coalescence), whereas Ostwald ripening leads to a monomodal distribution with the cube of the mean particle diameter increasing linearly with time (Kabalnov 2001). However, this is not always the case, since Ostwald ripening can lead to a bimodal distribution when the disperse phase contains mixed oils. 7.6.4.3 Oiling Off Tests Extensive droplet coalescence often leads to the formation of a separate oil layer on top of an oil-in-water emulsion. The amount of “oiling off” that has occurred can be assessed by quantifying the amount of oil released (McClements 2007). In some emulsions, the amount of free oil released can be determined by simple visual observation. In this case, the height of the oil layer (HO) and the height of the total emulsion (HE) are measured: %Oiling off = 100 × HO/HE. In this case, the degree of oiling off is calculated relative to the total volume of the emulsion. This procedure may be carried out after the emulsion has been left to stand for a certain period or after a specific environmental stress. Sometimes mild centrifugation is used to enhance the separation of the oil phase from the remainder of the sample. This method is unsuitable for emulsions where the amount of free oil released is relatively small because the oil layer is too thin to accurately measure. Methods have been developed to determine the amount of free oil released from emulsions based on the principle that a lipid dye solution added to an oil-in-water emulsion will Emulsion Stability 349 be diluted by free oil but not by emulsified oil (Palanuwech et al. 2003). The amount of oil released can then simply be determined by measuring the change in absorbance of the lipid dye solution before and after mixing with the emulsion. Alternatively, the degree of coalescence or oiling off can be measured using differential scanning calorimetry (DSC) methods (Thanasukarn et al. 2004). In these methods, the degree of droplet coalescence is determined by measuring the fraction of oil that crystallizes at low temperatures (small droplets) compared to the fraction that crystallizes at high temperatures (large droplets or bulk oil). These latter two methods calculate the degree of oiling off relative to the total amount of oil in the emulsion. 7.6.4.4 Accelerated Test Methods 7.6.4.4.1 Mechanical Agitation Methods The rate of droplet coalescence often increases when an emulsion is exposed to mechanical agitation, which can be used as the basis for the development of accelerated coalescence tests designed to better predict the long-term coalescence stability of emulsions (McClements 2007). The influence of shearing on the coalescence stability of emulsions can be established by measuring the change in their particle size distribution when they are sheared under defined conditions, for example, flow profile, shear rate, and duration. The coalescence stability can be characterized by measuring (1) the shear stress (τC) at which coalescence is first observed when the shear stress is increased in a controlled fashion or (2) the length of time (tC) that an emulsion must be sheared at a constant shear stress before coalescence is observed (Dickinson et al. 1993, Dickinson and Williams 1994). The influence of whipping on the coalescence stability of emulsions can be determined in a similar way (Goff 1997, van Aken 2001, Hotrum et al. 2005). The particle size distribution, microstructure, or physicochemical properties of an emulsion are measured after whipping at a fixed speed for a specific time. By varying the whipping times while keeping the shear stress constant or varying the shear stress while keeping the whipping time constant, it is possible to define a critical shear stress (τC) or time (tC) for coalescence to occur in a particular emulsion. The influence of homogenization on the coalescence stability of emulsions can be established using a variety of different methods (Chapter 6). 1. An emulsion is recirculated through a homogenizer at a fixed operating pressure until a constant particle size distribution is obtained where droplet coalescence is balanced by droplet disruption. The operating pressure is then reduced to a new value and the change in droplet size distribution with homogenization time (or number of passes) is monitored. The droplet size increases with homogenization time due to droplet coalescence until a new (larger) steady state value is reached where droplet coalescence is again balanced by droplet disruption. The faster the coalescence rate within the homogenizer, the faster the increase in droplet size with homogenization time. 2. An oil-in-water emulsion is prepared that contains a highly hydrophobic fluorescent probe in a fraction of the droplets, but no probe in the remainder of the droplets. This “mixed emulsion” is prepared by combining an emulsion that contains no probe with an emulsion that contains some probe, with all other aspects of the emulsions being similar, for example, droplet size distribution, oil type, and emulsifier type. The fluorescent probe is chosen so that the fluorescent emission spectrum depends on the local probe concentration in the oil phase, for example, due to formation of dimers that have a different spectrum than monomers. Coalescence during homogenization causes the dispersed phases of different droplets to be mixed, thereby reducing the local probe concentration and changing the fluorescence spectrum. Measurement of the change in fluorescence spectrum with time provides information about the coalescence rate. 3. A “mixed emulsion” is prepared by combining two oil-in-water emulsions that contain different colored oil-soluble dyes within the droplets (e.g., blue and yellow), but are similar in other aspects. The mixed emulsion is then homogenized by recirculating it through a homogenizer at a fixed operating pressure and the change in droplet composition with homogenization time (or number of passes) is monitored by measuring the change in the color of the 350 Food Emulsions: Principles, Practices, and Techniques droplets using optical microscopy. As homogenization proceeds, the droplets coalesce with each other and their contents are mixed, which results in a decrease in the fraction of droplets with the initial dye colors (yellow and blue) and an increase in the fraction of droplets of mixed color (green). 4. A “mixed emulsion” is prepared by combining two oil-in-water emulsions that contain different oil types (e.g., hexadecane and octadecane), but are otherwise similar. The mixed emulsion is then homogenized by recirculating it through a homogenizer at a fixed operating pressure and the change in droplet composition due to coalescence is monitored with homogenization time (or number of passes). The change in droplet composition with time is determined by measuring some physical property of the system that depends on oil composition, for example, refractive index (Taisne et al. 1996, Taisne and Cabane 1998) or melting point (Elwell et al. 2004). 7.6.4.4.2 Centrifugation Methods The rate of droplet coalescence in an emulsion can also be accelerated by centrifugation since this process forces the droplets together. In an oil-in-water emulsion, the coalescence stability is determined by measuring the change in droplet size distribution or the extent of “oiling off” after the emulsion has been centrifuged at a specified speed and time (van Aken and van Vliet 2002). The coalescence stability can be characterized in terms of the minimum centrifugation force that the emulsion can tolerate or the length of time that it will endure at a particular speed before coalescence is observed (van Aken and van Vliet 2002). Centrifugation methods have been developed to provide quantitative information about the stability of emulsions to coalescence (Tcholakova et al. 2002, 2005, 2006a). An emulsion is placed in an optically transparent centrifugation tube, which is then placed within a centrifuge. The emulsion is subjected to a centrifugal acceleration for a specific length of time. The oil droplets tend to move toward the axis of rotation of the centrifuge because they have a lower density than the surrounding aqueous phase. Initially, the droplets form a cream layer where they are forced into close proximity to one another but they retain their original size. However, if the centrifugal force is sufficiently high, the droplets will be forced into even closer proximity and the interfacial layers surrounding them will be ruptured, thereby releasing a layer of oil on top of the emulsion (Figure 7.40). A critical osmotic pressure (POSMCR) can be Oil Cream HC HReleased PCR = OSM Serum ∆ρg (VTotal – VReleased) A FIGURE 7.40 Observation of changes in the thickness of oil layers or cream layers during centrifugation under controlled conditions can be used to quantify droplet coalescence. In particular, a critical pressure can be determined that is related to the susceptibility of droplet interfaces to breaking. 351 Emulsion Stability defined as the osmotic pressure of an emulsion at the top of the emulsion column, where the emulsion is in mechanical equilibrium with the continuous layer of released oil (Tcholakova et al. 2005). The critical osmotic pressure can be measured experimentally: CR POSM = Drg(VTotal - VReleased ) A Here Δρ (=ρ1 − ρ2) is the difference in the mass density between the continuous and dispersed phases g is the acceleration VTotal is the total volume of dispersed phase in the emulsion VReleased is the volume of dispersed phase released during centrifugation A is the cross-sectional area of the centrifuge tube This method has proved particularly useful for monitoring the coalescence stability of protein-stabilized emulsions (Tcholakova et al. 2002, 2003, 2005, 2006b). 7.6.4.4.3 Limitations of Accelerated Coalescence Tests The advantages of accelerated coalescence tests are that they can give you an indication of the longterm stability of an emulsion from measurements made on a relatively short time. Nevertheless, accelerated coalescence tests (such as centrifugation, shearing, whipping, or homogenization) should be treated with caution because they may not always give a good correlation with the long-term coalescence stability of an emulsion. For example, chemical or biochemical changes may occur in an emulsion that is stored for a long period that eventually lead to coalescence (e.g., lipid oxidation and protein hydrolysis), but they may not be detected in an accelerated coalescence test. Alternatively, there may a critical force that is required to cause interfacial rupture that is exceeded in a centrifuge or shearing device, but which would never be exceeded under normal storage conditions. As a consequence, one must use these accelerated coalescence tests with caution and always carry our preliminary experiments to ensure that there is a good correlation between the results of the accelerated test and the actual long-term stability. 7.7 Partial Coalescence Originally, it was proposed that partial coalescence occurs when two or more partly crystalline oil droplets come into contact and form an irregularly shaped aggregate (Figure 7.41). The aggregate partly retains the shape of the original droplets from which it was formed because the mechanical strength of the fat crystal network within the droplets prevents them from completely merging together (Fredrick et al. 2010, Thivilliers-Arvis et al. 2010). The partially crystalline oil droplets may encounter each other during a collision within a bulk aqueous medium or they may encounter each other after adsorption to the surfaces of air bubbles formed during shearing. Partial coalescence is particularly important in dairy products, because milk fat globules are partly crystalline over a fairly wide range of temperatures (Goff 1997, Walstra 2003, Goff and Hartel 2013). The application of shear forces or temperature cycling to cream containing partly crystalline milk fat globules can cause partial coalescence, which leads to a marked increase in viscosity (“thickening”) and subsequent phase separation. Partial coalescence is an essential process in the production of ice cream, whipped toppings, butter, and margarine. Oil-in-water emulsions are cooled to a temperature where the droplets are partly crystalline and a shear force is applied, which leads to droplet aggregation via partial coalescence. In butter and margarine aggregation results in an O/W to W/O phase inversion, whereas in ice cream and whipped cream the aggregated fat droplets form a network that surrounds the air cells and extends throughout the aqueous phase, thus providing the necessary mechanical strength required to produce good stability and texture. 352 Food Emulsions: Principles, Practices, and Techniques Aggregation Fusion FIGURE 7.41 Partial coalescence occurs when a crystal from one partially crystalline lipid droplet penetrates into the liquid portion of another partially crystalline lipid droplet. 7.7.1 Physical Basis of Partial Coalescence Partial coalescence occurs when a solid fat crystal located in one droplet penetrates into a liquid oil region located in another droplet (Boode et al. 1991, 1993, Boode and Walstra 1993, Walstra 2003, McClements 2012). Normally, a protruding fat crystal would be surrounded by water, but when it penetrates into another fat droplet it is surrounded by liquid oil. This causes the droplets to remain aggregated because it is thermodynamically more favorable for a fat crystal to be surrounded by oil molecules than by water molecules due to the reduction in interfacial energy (hydrophobic effect). Over time the droplets tend to partially merge together because this reduces the unfavorable contact area of oil exposed to water (Figure 7.41). Hence, the junctions holding the partially crystalline droplets together often become stronger and more difficult to break after an emulsion has been aged, leading to a gel network in concentrated systems. A nice example of this effect is shown in electron microscopy images of hydrocarbon oil-in-water emulsions stored under conditions where the droplets were fully liquid or partly crystalline (Figure 7.42). More recently it has been proposed that partial coalescence may also occur in oil-in-water emulsions by other mechanisms. For example, emulsions that would be unstable to normal coalescence in the absence of fat crystals may undergo partial coalescence in the presence of fat crystals because the droplets are prevented from fully merging together due to the rigidity of the fat crystal network (Pawar et al. 2012). The solid fat content (ϕSFC) is the percentage of the fat phase that is crystalline at a particular temperature, which varies from 0% for a completely liquid oil to 100% for a completely solid fat (Chapter 4). Partial coalescence only occurs in emulsions that contain partially crystalline regions because a key Gelled sample Fluid sample (a) (b) FIGURE 7.42 Cryo-SEM images of 40wt% Tween 20 stabilized emulsions quench cooled from either (a) 40°C (liquid droplets) or (b) 0°C (partially crystalline droplets). Irregular fat aggregates are formed in the latter case due to partial coalescence. Scale bar is 1 μm. (SEM images by Professor John Coupland, Pennsylvania State University, University Park, PA.) 353 Emulsion Stability Partially crystalline 100 80 EC (%) 60 Completely liquid Completely solid 40 20 0 0 20 40 60 SFC (%) 80 100 FIGURE 7.43 Schematic diagram of the influence of droplet solid fat content on the efficiency of partial coalescence in oil-in-water emulsions. A maximum partial coalescence rate is usually observed at an intermediate solid fat content. requirement is the penetration of solid fat crystals into liquid oil phases. If the droplets were completely liquid they would undergo normal coalescence, whereas if they were completely solid they would undergo flocculation (rather than partial coalescence) because the rigid particles could not merge together. There is typically a range of solid fat contents at which the rate of partial coalescence is highest (Figure 7.43). Indeed, it has been found that increasing the solid fat content of the droplets causes an initial increase in the partial coalescence rate until a maximum value is reached, after which the partial coalescence rate decreases (Walstra 2003, Fredrick et al. 2010). The solid fat content at which this maximum rate occurs depends on the morphology and location of the fat crystals within the droplets, as well as the magnitude of the applied shear stresses. A practical example of this effect is demonstrated in Figure 7.44, which shows the extent of droplet coalescence in emulsions containing a hydrogenated palm oil after holding at different temperatures. Extensive droplet coalescence was observed after emulsions stored at temperatures ranging from 0°C to 8°C were warmed, which can be attributed to partial coalescence of the droplets since they were partially crystalline in this temperature range (Thanasukarn et al. 2004a,b). The morphology and location of fat crystals within an oil droplet also plays an important role in determining its susceptibility to partial coalescence (Darling 1982, Campbell 1989, Walstra 2003, Fredrick et al. 2010). The further a fat crystal protrudes from the surface of an oil droplet into the surrounding water, the more likely it is to penetrate into another oil droplet and therefore promote partial coalescence. The different types of partially crystalline fat globule commonly observed in milk, a representative oil-in-water emulsion, are shown in Figure 7.45. The small platelet-shaped crystals may be evenly distributed within the interior of the droplet (type N), located exclusively at the oil–water interface (Type L), or a combination of the two (type M). The type of crystals formed depends on the nucleation mechanism (homogeneous, surface heterogeneous, or volume heterogeneous), as well as the cooling conditions used to induce nucleation and crystallization (Walstra 2003, Fredrick et al. 2010). In other types of emulsified food products, there may be different crystal structures within the lipid droplets than those observed in milk. In general, the crystal structure depends on factors such as the chemical composition of the fat, the cooling rate, temperature cycling, the 354 Food Emulsions: Principles, Practices, and Techniques 100 Oiling off (%) 80 Fat crystallization Water freezing 60 Tween 20 40 + + + 20 + Casein 0 –40 –20 0 Temperature (°C) + 20 + + 40 FIGURE 7.44 Experimental measurements of the degree of oiling off in hydrogenated palm oil-in-water emulsions that had been held at different temperatures and then taken back to ambient temperature. Extensive droplet coalescence and oiling off occurs when the fat phase crystallizes (due to partial coalescence) or when the water phase crystallizes (because droplets are forced together). L M N FIGURE 7.45 Typical microstructures of fat crystals reported in milk fat globules. Note: L, crystals located at droplet surface; N, crystals distributed through droplet interior; M, mixed system. application of shear forces, droplet size distribution, the type of emulsifier used to stabilize the lipid droplets, and the presence of any impurities that can either poison or catalyze crystal growth (Walstra 2003, Fredrick et al. 2010). The rate of partial coalescence is also strongly influenced by the nature of the attractive and repulsive interactions acting between the lipid droplets, as well as the properties of the interfacial layer surrounding the lipid droplets. Emulsifiers that produce strong repulsive interactions between the oil droplets, or that form thick interfacial layers that are resistant to penetration and/or rupture by fat crystals, can reduce the susceptibility of emulsions to partial coalescence (Walstra 2003). For example, a number of studies have shown that oil-in-water emulsions containing lipid droplets coated by dairy proteins that form relatively thick layers are less susceptible to partial coalescence than those coated by small molecule surfactants that form relatively thin layers (Palanuwech and Coupland 2003, Thanasukarn et al. 2004a). This effect is clearly shown in Figure 7.44, which shows that a greater amount of instability due to partial coalescence occurs when oil droplets are coated by a thin layer of nonionic surfactant than when they are coated by a thick layer of protein. In emulsions with sufficiently high droplet concentrations, partial Emulsion Stability 355 coalescence leads to the formation of a network of aggregated particles that makes the emulsion highly viscous or gel-like (Vanapalli and Coupland 2001, Thivilliers et al. 2006). The formation of fat droplet networks due to partial coalescence can be seen in scanning electron microscopy images of oil-in-water emulsions containing partially crystalline droplets (Figure 7.42). It should be noted that fully crystalline fat droplets may undergo extensive aggregation if they undergo an appreciable shape change that increases their interfacial area. Then, the emulsifier present may be insufficient to completely cover the lipid particle surfaces, leading to a hydrophobic attraction between exposed nonpolar patches (Helgason et al. 2009). 7.7.2 Methods of Controlling Partial Coalescence The various factors that influence partial coalescence in oil-in-water emulsions have been reviewed in detail elsewhere (Walstra 2003, Fredrick et al. 2010). The major factors are the disperse phase volume fraction, mechanical agitation, crystal morphology, contact angle, colloidal interactions, and emulsifier type. In this section, we use this knowledge to highlight methods of controlling partial coalescence in oil-in-water emulsions. 7.7.2.1 Prevention of Close Contact Partial coalescence is more likely to occur in emulsions that contain droplets that remain in contact for extended periods, that is, flocculated emulsions, concentrated emulsions, and creamed layers. For example, partial coalescence occurs more rapidly when oil droplets are located in a creamed layer than when they are freely suspended in an aqueous phase. This phenomenon occurs because the droplet contact time is increased and the droplet separation is reduced in a creamed layer. In emulsions containing freely suspended droplets, the rate of partial coalescence is proportional to the frequency of encounters between emulsion droplets. Thus, anything that increases the collision frequency will increase the rate of partial coalescence (provided it does not also reduce the collision efficiency). The precise nature of these effects depends on whether the droplet collisions are induced by Brownian motion, shear, or gravity. In general, the collision frequency is proportional to the square of the disperse phase volume fraction (Section 7.5). Experiments with oil-in-water emulsions have shown that the partial coalescence rate is roughly proportional to ϕ2 up to oil concentrations around 20%, after which it increases more rapidly than expected. Oil-in-water emulsions that are stable to partial coalescence under quiescent conditions are often prone to it when they are subjected to mechanical agitation. A number of physicochemical mechanisms may contribute to the increased rate of partial coalescence during shearing of emulsions. First, shearing increases the collision frequency. Second, shearing may increase the collision efficiency because a crystal protruding from one droplet is more likely to penetrate into another droplet when the droplets “roll over” each other in a shear field. Final, a high shear stress may force droplets closer together, thus allowing fat crystals to penetrate through the interfacial layers surrounding the droplets more effectively. Partial coalescence is usually more rapid when emulsions are subjected to turbulent flow conditions than laminar flow conditions. Partial coalescence can only occur when droplets get so close together that a fat crystal from one droplet protrudes into another droplet. Thus, any droplet–droplet interaction that prevents droplets from coming into close contact should decrease the rate of partial coalescence, for example, electrostatic or steric repulsion. On the other hand, any droplet–droplet interaction that causes the droplets to come into close contact should increase the rate of partial coalescence, for example, van der Waals, hydrophobic, depletion, shear, centrifugal, and gravitational forces. The size of the droplets in an emulsion affects partial coalescence in a number of ways. The efficiency of partial coalescence can increase with increasing droplet size because the contact zone between the droplets increases, which increases the probability of their being a crystal present. On the other hand, increasing the droplet size decreases their collision frequency, which can lead to a decrease in partial 356 Food Emulsions: Principles, Practices, and Techniques coalescence with increasing size. The influence of droplet size is therefore quite complex and depends on the rate limiting step in the partial coalescence process. 7.7.2.2 Prevention of Interfacial Layer Disruption Emulsifiers that form thicker and more viscoelastic films at the oil–water interface are more resistant to penetration by fat crystals and so provide greater stability to partial coalescence (Walstra 1983). Consequently, the rate of partial coalescence is less for droplets stabilized by proteins (thick layers) than for those stabilized by small molecule surfactants (thin layers) (Palanuwech and Coupland 2003, Thanasukarn et al. 2004b), as highlighted in Figure 7.44. The nature of the emulsifier at the surfaces of emulsion droplets has an important impact on the stability of a number of foods. The presence of phospholipids in dairy emulsions has been observed to decrease their stability to thickening under the influence of shear forces, which has been attributed to the displacement of protein molecules from the oil–water interface by the more surface active phospholipids, leading to the formation of an emulsifier film that is more susceptible to penetration by fat crystals (Boode 1992). When an ice cream mix or dairy whipped topping is aged in the presence of small molecule surfactants prior to cooling and shearing the resulting product has improved texture and better stability (Goff 1997, Goff and Hartel 2013). This is because emulsifiers displace the milk proteins from the oil–water interface and so enhance the tendency for partial coalescence to occur during subsequent cooling and shearing. For these reasons, small molecule surfactants are often added to ice creams and dairy toppings to improve their physical characteristics. Only about one in a million encounters between droplets typically leads to partial coalescence (Vanboekel and Walstra 1981). This suggests that the rate limiting step of partial coalescence is the penetration of the emulsifier layer by a crystal and subsequent nucleation, rather than the collision frequency. 7.7.2.3 Control of Crystal Concentration, Structure, and Location The degree of partial coalescence in an emulsion can be regulated by controlling the concentration, structure, and location of the fat crystals within the droplets (Walstra 2003, Fredrick et al. 2010). The most effective method of preventing partial coalescence is to ensure that all the droplets are either completely liquid or completely solid (Figure 7.43). This can be achieved by carefully controlling the temperature of the emulsion so that it is above or below some critical range where the droplets are only partially crystalline. If it is not possible to alter the temperature, then the solid fat content could be controlled by selecting a fat with a different melting profile. The susceptibility of an emulsion to partial coalescence also depends on the morphology and the precise location of the fat crystals within the emulsion droplets. The greater the number of crystals that protrude from a droplet and the further that they protrude into the water phase, the more effective is partial coalescence. At low solid fat contents, the crystals often do not penetrate very far into the aqueous phase, but once a critical solid fat content is reached a network of aggregated crystals is formed, which tends to increase the likelihood that crystals will protrude out of the droplets. Thus, two emulsions with the same solid fat content may have very different stabilities because of differences in crystal properties. The number and morphology of crystals in droplets can be controlled by varying the cooling rate at which they are formed, by selecting an appropriate fat source, or by adding components that modify fat nucleation and crystal growth rates (Awad et al. 2001, Awad and Sato 2001, 2002). The presence of small molecule surfactants in the surrounding continuous phase may pull crystals fully or partly out of lipid droplets, which would alter the tendency for partial coalescence to occur (Ergun et al. 2015). 7.7.3 Experimental Characterization of Partial Coalescence Food scientists would like to have analytical protocols to provide information about: the susceptibility of a particular emulsion to partial coalescence; the extent of partial coalescence that has occurred in an emulsion; the influence of partial coalescence on the macroscopic properties of emulsions (stability, rheology, and appearance); and the factors that influence partial coalescence. Emulsion Stability 357 7.7.3.1 Fat Crystal Properties One of the most important factors that determine the susceptibility of an oil-in-water emulsion to partial coalescence is the concentration, morphology, and location of the fat crystals within the emulsified lipids, as well as how these properties change with temperature. Thus, analytical instruments are required to provide information about fat crystal properties in emulsions. Fat crystallization in emulsions can be monitored using experimental techniques that utilize differences in the physicochemical properties of the solid and liquid phases (e.g., density, compressibility, birefringence, molecular mobility, or packing) or changes associated with the solid–liquid phase transition (e.g., absorption or release of heat). The physicochemical properties that are of most interest to food scientists are (1) the final melting point of the fat, (2) the variation of the solid fat content (SFC) with temperature, (3) the morphology, interactions, and location of the crystals within the droplets, (4) the packing of the fat molecules within the crystals, and (5) the influence of droplet crystallization on the overall stability and physicochemical properties of the emulsion. The variation of the SFC with temperature can be measured using a number of analytical techniques, including density measurements, differential scanning calorimetry (DSC), nuclear magnetic resonance (NMR), ultrasonic velocity measurements, and electron spin resonance (ESR) (Sato 1999, Rousseau 2000, McClements 2007). The technique used in a particular experiment depends on the equipment available, the information required, and the nature of the sample being tested. The position of fat crystals relative to an oil–water interface depends on the relative magnitude of the oil–water, oil–crystal, and crystal–water interfacial tensions and is characterized by the contact angle (Rousseau 2000, Walstra 2003). The contact angle between liquid oil, solid fat, and water phases can be measured (Darling 1982). Two fats can have exactly the same SFC but very different physical characteristics because of differences in the crystal habit and spatial distribution of the crystals. The location of crystals within a system, and their crystal habit can be studied by optical, electron, or atomic force microscopy depending on the crystal size (Boode et al. 1991, Rousseau and Hodge 2005, Rousseau 2006, Tang and Marangoni 2006). Transmission electron microscopy has proved to be particularly powerful for providing information about the packing of molecules within crystalline fat droplets (Jores et al. 2004, Kuntsche et al. 2011). The packing of the molecules in the crystals can be determined by techniques that utilize the scattering or adsorption of radiation (Rousseau 2000, Hartel 2001, Rousseau and Hodge 2005). X-ray diffraction and small angle neutron scattering have been used to determine the long and short spacings of the molecules in fat crystals (Hartel 2001). Infrared and Raman spectroscopy have been used to obtain information about molecular packing via its effect on the vibration of certain chemical groups in fat molecules (Yano and Sato 1999, Dufour et al. 2000, Bresson et al. 2011). Each polymorphic form has a unique spectrum that can be used to identify it. The polymorphic form of fat crystals can also be identified by measuring the temperature at which phase transitions occur and the amount of heat absorbed/released using DSC (Avendano-Gomez et al. 2005, Clausse et al. 2005, Qian et al. 2012, Shukat et al. 2012). Techniques have been developed that combine different analytical techniques to simultaneously monitor changes in solid fat content and polymorphism in emulsions, for example, DSC, ultrasonics and X-ray diffraction (Awad et al. 2001, Awad and Sato 2001, 2002). 7.7.3.2 Emulsion Microstructure Partial coalescence may lead to the formation of an extensive network of aggregated fat droplets at temperatures where the lipid phase is partly crystalline. This type of droplet aggregation can conveniently be monitored by optical or electron microscopy (Boode et al. 1993, Relkin et al. 2003, Relkin and Sourdet 2005). The electron micrographs of oil-in-water emulsions before and after partial coalescence shown in Figure 7.42 clearly indicate the formation of a particle network. If an emulsion initially containing partially coalesced fat droplets is heated to a temperature where the fat crystals melt, then the oil droplets usually merge together and coalesce (Figure 7.36). This process may lead to an appreciable increase in mean droplet size or to extensive oiling off in the warmed emulsion. The change in particle size distribution in a warmed emulsion can be determined by the particle 358 Food Emulsions: Principles, Practices, and Techniques sizing methods discussed earlier, for example, light scattering or electrical pulse counting. The extent of oiling off can be determined by visually measuring the height of the free oil layer or by extraction and quantification of the free oil (Palanuwech et al. 2003). 7.7.3.3 Macroscopic Properties Ultimately, a food scientist is interested in the influence of droplet crystallization on the bulk physicochemical properties of a food emulsion, such as its appearance, stability, and texture. The large aggregates formed due to partial coalescence cause an increase in emulsion viscosity and may eventually lead to solidification, so that it can be monitored by measuring the increase in viscosity or shear modulus, either as a function of time or temperature (Vanapalli and Coupland 2001, Thivilliers et al. 2006). The stability of emulsion droplets to creaming or sedimentation is also influenced by partial coalescence, which can be followed by the techniques described in Section 7.3. 7.8 Ostwald Ripening Ostwald ripening is the process whereby large droplets grow at the expense of smaller ones in a polydisperse emulsion because of the mass transport of dispersed phase from one droplet to another through the intervening continuous phase (Kabalnov and Shchukin 1992, Taylor 1998, Kabalnov 2001) (Figure 7.46). Ostwald ripening is normally negligible in food emulsions containing long chain triacylglycerols as the oil phase (such as canola, corn, fish, olive, palm, peanut, or sunflower oil) because the mutual solubility’s of these triacylglycerols and water are so low that the mass transport rate is insignificant. However, it is important in oil-in-water emulsions that contain more water-soluble oil phases, for example, short chain triacylglycerols (Li et al. 2009), flavor oils (McClements et al. 2012, Rao and McClements 2012), and essential oils (Chang et al. 2012). Ostwald ripening may also promote instability in long chain triglyceride emulsions that contain alcohol in the aqueous phase since this increases oil solubility, for example, cream liqueurs (Dickinson and Golding 1998). Thus, there is a range of products within the food industry where Ostwald ripening may be important and needs to be controlled. 7.8.1 Physical Basis of Ostwald Ripening Ostwald ripening occurs because the solubility of the material (the “solute”) within a spherical particle increases as the size of the particle decreases (Weers 1998, Kabalnov 2001): Large droplets grow Small droplets shrink FIGURE 7.46 Ostwald ripening involves the growth of large droplets at the expense of smaller ones due to diffusion of dispersed phase through the continuous phase. The driving force for this process is the fact that the solubility of a substance within a droplet in the continuous phase surrounding it increases with decreasing droplet radius. 359 Emulsion Stability æ 2gVm S (r ) = S (¥)exp ç è RTr ö æaö ÷ = S (¥)exp ç r ÷ ø è ø (7.39) Here S(∞) is the solubility of the solute in the continuous phase for a particle with infinite curvature (a planar interface) S(r) is the solubility of the solute when contained within a spherical particle of radius r α (=2γVm /RT) is a characteristic length scale Vm is the molar volume of the solute γ is the interfacial tension The increase in solubility with decreasing particle size means that there is a higher concentration of solute around a small droplet than around a larger one in a polydisperse emulsion (Figure 7.46). The solute molecules therefore move from the smaller droplets to the larger droplets because of this concentration gradient. This process causes the smaller droplets to shrink and the larger droplets to grow, leading to an overall net increase in the mean droplet size with time. In general, Ostwald ripening can be separated into two phases (Weers 1998, Kabalnov 2001). In the initial nonsteady state period, the emulsion has a droplet size distribution that is mainly determined by the homogenization conditions used to product it. As Ostwald ripening proceeds the shape of the droplet size distribution evolves toward a particular shape, which is determined by the physicochemical processes associated with droplet shrinkage and growth. In the steady state period, the droplet size distribution maintains a time-independent form, and only shifts up the particle size axis during aging (Figure 7.47). 35 0h 24 h 48 h Volume (%) 30 25 20 15 10 5 0 0.4 4 Diameter (µm) 0.04 Monomodal PSD evolution d3 (µm3) d3 proportional to time 0.03 0.02 0.01 0 0 200 400 Time (h) 600 FIGURE 7.47 Under steady state conditions, Ostwald ripening of oil-in-water emulsions containing a single oil type tends to adopt a certain monomodal particle size distribution that moves up the time axis and has an OR rate that is proportional to the particle size cubed. 360 Food Emulsions: Principles, Practices, and Techniques Once steady state has been achieved, the mean droplet size should evolve according to the following equations derived from the Lifshitz–Slyozov–Wagner (LSW) theory (Weers 1998): d ár ñ 3 4 = aS (¥ ) D dt 9 r 3 - r03 = wt = 4 aS (¥) Dt 9 (7.40) (7.41) where D is the translation diffusion coefficient of the solute molecules through the continuous phase separating the droplets. This equation indicates that the cube of the mean particle size should increase linearly with time, and that the rate of this process increases as the equilibrium solubility of the solute molecules within the continuous phase increases. Experiments have shown that the droplet size distribution does tend to adopt a time-independent form during Ostwald ripening, but that the shape of the distribution is slightly different from that predicted from the LSW theory (Weers 1998). For emulsions containing oil phases that only contain a single molecular species, experiments have shown that the cube of the mean particle size increases linearly with time, and that the Ostwald ripening (OR) rate is proportional to the solubility of the disperse phase in the continuous phase. Nevertheless, experimentally determined values of the Ostwald ripening rate in oil-in-water emulsions (in the absence of micelle solubilization effects) have been found to be about two to three times higher than that predicted by the LSW theory, which was attributed to the Brownian motion of the oil droplets. In the presence of micelle solubilization effects, the measured Ostwald ripening rates may be much greater than those predicted by the LSW theory. The time dependence of the droplet size distribution and mean droplet diameter for an oil-in-water emulsion is shown in Figure 7.47. For emulsions containing two or more oils with different water solubilities, the change in the particle size distribution with time may not be monomodal and OR rate may not be proportional to the droplet size cubed (Kabalnov and Shchukin 1992, Wooster et al. 2008). The above equations also assume that the rate limiting step is the diffusion of the solute molecules through the continuous phase. Most food emulsions contain droplets that are surrounded by interfacial layers comprising of emulsifiers and other substances, and in some instances these layers may retard the diffusion of solute molecules in or out of the droplets. Under these circumstances, the above equation must be modified to take into account the diffusion of solute molecules across the droplet coatings (Kabalnov and Shchukin 1992): dñ r á 3 3 æ S i - Sc ö = ç ÷ dt 4p è Ri + Rc ø (7.42) where Si and Sc are the solubilities of the solute in the interfacial layer and the continuous phase Ri and Rc are the diffusion resistances of the interfacial layer and continuous phase Ri = 1 4prDi Rc = dCi,¥ 4pr 2 DcCc,¥ (7.43) Here δ is the thickness of the interfacial layer Ci,∞ and Cc,∞ are the solubilities of the solute in the specified phases The subscripts i and c refer to the properties of the interfacial and continuous phases, respectively When the diffusion of the solute molecules through the interfacial layer is limiting, the growth rate of the droplet size is proportional to r 2, rather than r 3 (Kabalnov and Shchukin 1992). 361 Emulsion Stability 7.8.2 Methods of Controlling Ostwald Ripening The various factors that influence Ostwald ripening in emulsions have been reviewed in detail previously, for example, droplet size, solute solubility, interfacial tension, interfacial diffusion, and droplet composition (Kabalnov and Shchukin 1992, Taylor 1998, Weers 1998). Knowledge of these factors can be used to control the rate of Ostwald ripening in emulsions. 7.8.2.1 Droplet Size Distribution Ostwald ripening proceeds more rapidly when the average size of the droplets in an emulsion decreases, because the solubility of the dispersed phase increases with decreasing droplet radius. Hence, the droplet size increases more rapidly in emulsions containing small droplets than large droplets. The initial rate also increases as the width of the particle size distribution increases. Ostwald ripening can therefore be retarded by ensuring that an emulsion has a narrow droplet size distribution and that the droplets are fairly large. Nevertheless, there may be other problems associated with having relatively large droplets in an emulsion, such as accelerated creaming, flocculation, or coalescence. 7.8.2.2 Solubility The greater the equilibrium solubility of the dispersed phase in the continuous phase, the faster the rate of Ostwald ripening (Equation 7.40). Ostwald ripening is therefore extremely slow in oil-in-water emulsions containing lipids that are sparingly soluble in water (e.g., long chain triacylglycerols), but may occur at an appreciable rate in emulsions containing lipids that are smaller and/or more polar (e.g., flavor oils, essential oils, or short chain triacylglycerols). Certain substances are capable of increasing the water solubility of lipids in water, and are therefore able to enhance the Ostwald ripening rate, for example, alcohols or surfactant micelles (Dickinson and Golding 1998, Pena and Miller 2006). Ostwald ripening could therefore be retarded by excluding these substances from the emulsion, or by using lipids with a low water solubility. An example of the influence of the water solubility of the oil phase on Ostwald ripening is given in Figure 7.48, which shows that droplet growth is considerably faster for decane than for hexadecane. This is because decane has a considerably higher water solubility than hexadecane. 4 Decane d3 (μm3) 3 2 Hexadecane 1 0 0 25 50 75 Time (h) 100 FIGURE 7.48 Influence of oil solubility on droplet growth in oil-in-water emulsions due to Ostwald ripening (OR). The OR rate is much less for hexadecane than for decane because hexadecane has a much lower water solubility due to its longer chain length. 362 Food Emulsions: Principles, Practices, and Techniques 7.8.2.3 Interfacial Layer The rate of Ostwald ripening increases as the interfacial tension increases (Equation 7.40). Consequently, it is possible to retard its progress by using an emulsifier that is highly effective at reducing the interfacial tension (Weers 1998). The mass transport of molecules from one droplet to another depends on the rate at which the molecules diffuse across the interfacial layer (Kabalnov and Shchukin 1992). It may therefore be possible to retard Ostwald ripening by decreasing the diffusion coefficient of the dispersed phase in the interfacial layer, or by increasing the thickness of the interfacial layer. Little work has been carried out in this area; however, it may prove to be an useful means of controlling the stability of some food emulsions. Finally, the resistance to deformation of the interfacial layers surrounding droplets may also be able to reduce the Ostwald ripening rate. The shrinkage or growth of droplets stabilized by biopolymers (proteins or polysaccharides) that form cohesive interfacial layers may be retarded because of the mechanical resistance of the interfaces to changes in their area (Weers 1998). For example, cross-linking the proteins within the interfacial layers surrounding oil droplets has been shown to inhibit Ostwald ripening in protein-stabilized oil-in-water emulsions (Dickinson et al. 1999). Coating oil droplets with multiple layers of biopolymers using electrostatic deposition has also been shown to inhibit Ostwald ripening by providing a mechanical resistance to droplet deformation (Mun and McClements 2006, Zeeb et al. 2012). 7.8.2.4 Droplet Composition The Ostwald ripening rate is particularly sensitive to the composition of emulsion droplets that contain a mixture of components with different solubilities in the continuous phase (Kabalnov and Shchukin 1992, Weers 1998). Consider an oil-in-water emulsion that contains droplets comprised of two different types of oil-soluble components: ML has a low water solubility, and MH has a high water solubility. The diffusion of MH molecules from the small to the large droplets occurs more rapidly than the ML molecules. Consequently, there is a greater percentage of MH in the larger droplets than in the smaller droplets. Differences in the composition of emulsion droplets are thermodynamically unfavorable because of entropy of mixing effects: it is entropically more favorable to have the two oils distributed evenly throughout all of the droplets, rather than concentrated in particular droplets. Consequently, there is a thermodynamic driving force that operates in opposition to the Ostwald ripening effect, which is often referred to as compositional ripening (Figure 7.49). The change in droplet size distribution with time then depends on the concentration and solubility of the two components within the oil droplets. OR Promotes droplet growth Size difference CR Opposes droplet growth Concentration gradient Water-soluble oil FIGURE 7.49 In mixed oil systems, the increase in droplet size due to Ostwald ripening (OR) is often opposed by compositional ripening (CR) due to changes in droplet composition. If OR occurs, the concentration of a low-water solubility component within the smaller droplets will increase, which is opposed due to entropy of mixing effects. 363 Emulsion Stability The dependence of droplet growth due to OR on droplet composition can be divided into three regimes for systems containing a water-soluble and a water-insoluble component (Kabalnov et al. 1987, Taylor 1998): 1. Unstable regime: X2 < 2a1 3d0 (7.44) 2. Kinetically stable regime: 2a1 2a < X2 < 1 d0 3d0 (7.45) 3. Thermodynamically stable regime: X2 > 2a1 d0 (7.46) Here X2 is the initial mole fraction of water-insoluble component (ML) in the disperse phase α1 (=2γVm /RT) is the characteristic length scale of the water-soluble component (MH) d 0 is the initial mean droplet diameter In the unstable regime, the OR effect dominates the entropy of mixing effect because the initial concentration of water-insoluble component in the droplets is too low to generate a sufficiently high concentration gradient. Hence, the droplets continue to grow at a rapid rate due to OR. Usually, a bimodal distribution is observed after a certain time with a population of large droplets enriched with the watersoluble component (Kabalnov et al. 1987, Taylor 1998). In the thermodynamically stable regime, the entropy of mixing effect dominates the OR effect and the droplets are stable to droplet growth. Under these conditions, the size and composition of the droplets remains constant over time (Kabalnov et al. 1987, Kabalnov and Shchukin 1992). In the kinetically stable regime, droplet growth is thermodynamically favorable, but there is a kinetic energy barrier that retards droplet growth. If the initial particle size distribution is fairly narrow, then the emulsion may remain relatively stable to OR for some time, but if the particle size distribution is broad then OR occurs but at a rate appreciably lower than that of the pure water-soluble component. To a first approximation, the rate of Ostwald ripening for an emulsion containing two oils with different water-solubilities is given by the following equation in the kinetically stable regime (Weers 1998): æf f ö wmix = ç L + H ÷ w w H ø è L -1 (7.47) where ϕ and ω are the volume fraction and Ostwald ripening rates of the pure substances. This equation predicts that the presence of the low-solubility component (ML) will slow down the overall Ostwald ripening rate. Nevertheless, the general characteristics of the ripening process are similar to those of pure oils, that is, d3 increases linearly with time, and the shape of the droplet size distribution is time-­ invariant. On the other hand, if the system is in the metastable regime (i.e., there is insufficient lowsolubility component within the droplets), then a bimodal droplet size distribution develops over time. It is therefore possible to control the rate of Ostwald ripening in oil-in-water food emulsions by using an oil phase that contains a mixture of oils with different water solubilities (McClements et al. 2012, Rao and McClements 2012). An example of this effect is shown in Figure 7.50, which shows the influence 364 Food Emulsions: Principles, Practices, and Techniques Mean droplet diameter (μm) 4 0% EG 4.8% EG 3 9.1% EG 16.7% EG 2 1 0 0 15 Days 4 Mean droplet diameter (μm) 15 5 10 Storage time (day) 3 2 1 0 0 5 10 15 20 Ester gum in oil phase (%) 25 FIGURE 7.50 Increase in mean particle diameter during storage in oil-in-water emulsions due to Ostwald ripening. The oil phase contained a mixture of orange oil (high water solubility) and ester gum (low water solubility). The ester gum (EG) acts as a ripening inhibitor, effectively retarding OR when its concentration in the oil phase exceeds 10%. of adding increasing amounts of ester gum (a low water solubility oil) to orange oil (a high water solubility oil). When the ester gum concentration exceeds a particular value, it completely inhibits droplet growth since the entropy of mixing effect opposes Ostwald ripening. Similar improvements in stability to Ostwald ripening can be obtained in water-in-oil emulsions (such as margarine or butter) by including water-soluble components that have a low solubility in the lipid continuous phase, for example, salts (Walstra 2003). Finally, it should be noted compositional ripening can occur within emulsions, even if they have similar droplet sizes. If an emulsion contains droplets that have different internal compositions, then there is a thermodynamic driving force (entropy of mixing) that favors the exchange of the disperse phase material between the droplets until they all have similar compositions. This process can be achieved by diffusion Emulsion Stability 365 of oil molecules through the continuous phase separating the droplets. Compositional ripening may be a useful practical means of introducing an oil-soluble component into the droplets of a preexisting oil-inwater emulsion. An emulsion of the oil-soluble component could be prepared and then mixed with the preexisting emulsion. Provided the oil-soluble component has sufficient solubility in the aqueous phase, then it will be incorporated into all of the droplets given sufficient time. 7.8.3 Experimental Characterization of Ostwald Ripening Most of the analytical methods used to monitor Ostwald ripening are similar to those used to monitor droplet coalescence, that is, techniques that measure changes in droplet size distribution over time (Section 7.6.4). If the droplets are sufficiently large (>1 μm) then optical microscopy can be used, otherwise particle sizing instruments can be used, such as light scattering or electrical pulse counting. Nevertheless, it is often difficult to directly distinguish Ostwald ripening and coalescence using particle sizing instruments because both instability mechanisms cause an increase in the mean size of the droplets over time. It is sometimes possible to distinguish between Ostwald ripening and coalescence by measuring the change in particle size distribution or mean particle size with time, and/or by examining the factors that influence the rate of droplet growth. For OR, the particle size distribution should attain a specific time-independent form that moves up the size axis with time (Weers 1998), whereas with (heterogeneous) coalescence a bimodal distribution is usually observed (Deminiere et al. 1998). In addition, for OR the mean particle diameter cubed usually increases linearly with time (d3 ∝ t), whereas for coalescence the reciprocal of the particle diameter squared should increase linearly with time (1/d2 ∝ t) (Kabalnov 1998). Having said this, coalescence may sometimes lead to a monomodal distribution (e.g., homogeneous coalescence), whereas OR may lead to a bimodal distribution (e.g., mixed oil systems). Consequently, it may be necessary to carry out additional experiments to distinguish these two mechanisms. OR should be fairly insensitive to the type of emulsifier used to coat the droplets (being mainly governed by disperse phase solubility), whereas coalescence should be highly dependent on emulsifier type. 7.9 Phase Inversion Phase inversion is the process whereby a system changes from an oil-in-water emulsion to a water-in-oil emulsion, or vice versa (Figure 7.51). Phase inversion is an essential step in the manufacture of a number of important food products, including butter and margarine. In other foods, phase inversion is undesirable because it has an adverse effect on their appearance, texture, stability, and taste. In these products, a food manufacturer wants to avoid the occurrence of phase inversion. 7.9.1 Physical Basis of Phase Inversion Phase inversion is usually triggered by some alteration in the composition or environment of an emulsion, for example, disperse phase volume fraction, emulsifier type, emulsifier concentration, solvent type, additives, temperature, or mechanical agitation (Brooks et al. 1998, Kabalnov 1998, McClements 2005). Only certain types of emulsion are capable of undergoing phase inversion, rather than separating into their individual phases. These emulsions are capable of existing in a kinetically stable state both before and after phase inversion takes place. Typically, it is necessary to agitate an emulsion throughout the phase inversion process, otherwise it will separate into its component phases rather than forming another emulsion. The physicochemical basis of phase inversion is extremely complex, involving aspects of droplet flocculation, coalescence, and disruption (Dickinson 1992a). At the point where phase inversion occurs, the system may contain regions of oil-in-water (O/W) emulsion, water-in-oil (W/O) emulsion, multiple emulsions (e.g., W/O/W or O/W/O), and bicontinuous phases (Brooks et al. 1998). In food emulsions, phase inversion can be conveniently divided into two types depending on its origin: surfactant-induced and fat crystallization-induced. Surfactant-induced phase inversion is 366 Food Emulsions: Principles, Practices, and Techniques Phase inversion Oil-in-water emulsion Water-in-oil emulsion FIGURE 7.51 Phase inversion involves the conversion of an oil-in-water emulsion to a water-in-oil emulsion, or vice versa. usually caused by a change in the molecular geometry (optimum curvature) of the surfactant monolayer used to stabilize the initial emulsion due to some alteration in emulsion composition or environment (e.g., temperature, salt, alcohol, or surfactant) (Brooks et al. 1998, Kabalnov 1998). Fat crystallization-induced phase inversion is caused by a change in an emulsion that leads to extensive partial coalescence, for example, temperature, shearing, and alteration of interfacial layer thickness (Walstra 2003). 7.9.1.1 Surfactant-Induced Phase Inversion Surfactant-induced phase inversion occurs in emulsions stabilized by small molecule surfactants and may be classified as “transitional” or “catastrophic” phase inversion depending on the underlying physicochemical mechanisms (Binks 1998). Transitional phase inversion is caused by changes in the molecular geometry of the surfactant molecules in response to alterations in solution or environmental conditions, for example, temperature, ionic strength, or effective HLB number (Shinoda and Friberg 1986, Salager 1988, Evans and Wennerstrom 1999). For example, an emulsion stabilized by a nonionic surfactant undergoes a transition from an oil-in-water emulsion, to a bicontinuous system, to a water-in-oil emulsion on heating because of progressive dehydration of the surfactant head groups (Salager et al. 2003, 2004). This process is characterized by a phase inversion temperature (PIT), which is governed by the molecular geometry of the surfactant molecules. An example of phase inversion induced by heating an oil-in-water nanoemulsion stabilized by a nonionic surfactant is shown in Figure 7.52. At low temperatures an O/W emulsion is stable, but upon heating this emulsion becomes unstable to coalescence and then inversion to a W/O emulsion due to dehydration of the head groups. An oil-in-water emulsion stabilized by an ionic surfactant exhibits a similar kind of behavior when the concentration of electrolyte in the aqueous phase is increased. Increasing the ionic strength causes the system to undergo a transition from an oil-in-water emulsion, to a bicontinuous system, to a waterin-oil emulsion, because the electrical charge on the surfactant head groups is progressively screened by the counter ions (Chapter 4). Phase inversions may also be induced by changing the effective HLB number of the surfactants by mixing surfactants together. A surfactant-stabilized emulsion may switch from a water-in-oil emulsion at a low HLB number, to a bicontinuous system at intermediate 367 Emulsion Stability 4 3.5 Turbidity (cm–1) 3 O/W 2.5 2 1.5 W/O 1 0.5 0 PIT 50 60 70 Temperature (°C) 80 90 FIGURE 7.52 Phase inversion occurs in some oil-in-water emulsions stabilized by nonionic surfactants due to progressive dehydration of the head groups during heating. This can be observed by measuring the change in turbidity of oil-inwater nanoemulsions upon heating. HLB, to an oil-in-water emulsion at high HLB number. In general, the tendency for this type of phase inversion to occur is determined by the packing parameter (p) of the surfactant (Chapter 4), with p < 1 favoring an oil-in-water emulsion, p ≈ 1 favoring a bicontinuous system, and p > 1 favoring a waterin-oil emulsion. Catastrophic phase inversion usually occurs when the disperse phase volume fraction is increased above a critical level (Brooks et al. 1998). As the droplet concentration is gradually increased, there is suddenly a dramatic change in the system characteristics. The point where phase inversion occurs also depends on the intensity of agitation and the rate at which the disperse phase is added to the emulsion. Transitional phase inversions are usually reversible, whereas catastrophic phase inversions are usually irreversible. For example, when a transitional phase inversion is induced in an oil-in-water emulsion by increasing its temperature above the PIT, the emulsion will usually revert back into an oil-in-water emulsion when it is cooled back below the PIT. Nevertheless, it is often necessary to continuously agitate the system during this process, or else it will separate into the individual oil and water phases. In addition, there is often an effect that is analogous to that of super-cooling, that is, the temperature at which the phase inversion occurs on heating is different from that on cooling. This is because there is an activation energy that must be overcome before a system can be transformed from one state to another. 7.9.1.2 Fat Crystallization–Induced Phase Inversion When an oil-in-water emulsion containing completely liquid droplets is cooled to a temperature where the droplets are partly crystalline and then sheared it may undergo a phase inversion to a water-inoil emulsion (Dickinson and Stainsby 1982, Walstra 2003). The principle cause of this type of phase 368 Food Emulsions: Principles, Practices, and Techniques inversion is partial coalescence of the droplets, which leads to the formation of a continuous fat crystal network that traps water droplets within it. This is one of the principal manufacturing steps in the production of margarine and butter. When the emulsion is heated to a temperature where the fat crystals melt, the emulsion breaks down because the water droplets are released and sediment to the bottom of the sample where they coalesce with other water droplets. This is clearly seen when one melts margarine or butter and then cools it back to the original temperature: the sample is very different before and after heating. This type of phase inversion depends mainly on the crystallization of the fat and on the resistance of the droplets to partial coalescence (see Section 7.7). 7.9.2 Methods of Controlling Phase Inversion The propensity for phase inversion to occur in an emulsion can be controlled in a number of ways. 7.9.2.1 Disperse Phase Volume Fraction If the dispersed phase volume fraction of an emulsion is increased, while all the other experimental variables are kept constant (e.g., emulsifier type, emulsifier concentration, temperature, and shearing rate), then a critical volume fraction (ϕcp) may be reached where the system either undergoes a catastrophic phase inversion or completely breaks down so that the excess dispersed phase forms a layer on top of the emulsion. There is usually a range of volume fractions over which an emulsion can exist as either a water-in-oil emulsion or as an oil-in-water emulsion (Dickinson 1992). Within this region the emulsion can be converted from one state to another by altering some external property, such as the temperature or shear rate. From geometrical packing considerations, it has been estimated that this range extends from 1 − ϕcp < ϕ < ϕcp, where ϕcp refers to the volume fraction when the droplets are packed closely together without being distorted. In practice, factors other than simple geometric considerations will influence this range, including the fact that the droplets can become deformed and the chemical structure of the emulsifier used. Catastrophic phase inversion can therefore be prevented by ensuring that the droplet concentration is kept below ϕcp (∼0.6). 7.9.2.2 Emulsifier Type and Concentration The most important factor determining the susceptibility of an emulsion to surfactant-induced transitional phase inversion is the molecular geometry of the surfactant used to stabilize the droplets (Evans and Wennerstrom 1999, Israelachvili 2011). Surfactant-stabilized emulsions undergo a phase inversion when some change in the environmental conditions causes the optimum curvature of the surfactant monolayer to tend toward 0 (or p → 1), for example, temperature, ionic strength, cosolvents, or cosurfactants. The surfactants that are susceptible to this kind of phase inversion usually have an intermediate HLB number (nonionic surfactants) or are electrically charged (ionic surfactants). Alternatively, mixtures of two different types of surfactants can be used, one to stabilize the oil-in-water emulsion and the other to stabilize the water-in-oil emulsion. The point at which phase inversion occurs is then sensitive to the ratio of the surfactants used (Dickinson 1992). Emulsions stabilized by proteins do not exhibit this type of phase inversion because proteins are incapable of stabilizing water-in-oil emulsions. The total concentration of emulsifier present in the system is also important because there must be sufficient present to cover all of the droplets formed in both the oil-in-water and water-in-oil states on either side of the phase inversion. The emulsifier type and concentration is also important in determining the stability of emulsions to fat crystallization-induced phase inversion. Emulsifiers that form thick viscoelastic interfacial layers are more likely to protect an emulsion from this type of phase inversion because they retard partial coalescence (Section 7.7). It is therefore important to select an emulsifier that exhibits the appropriate behavior over the experimental conditions that a food emulsion experiences during its lifetime. Emulsion Stability 369 7.9.2.3 Mechanical Agitation It is often necessary to subject an emulsion to a high shearing force to induce phase inversion in the region where it can possibly exist as either an oil-in-water or water-in-oil emulsion. The higher the applied shearing force, the more likely that phase inversion will occur. If the shearing force were not applied, then the system may just undergo separation into the individual oil and water phases, rather than being transformed into a phase-inverted emulsion. 7.9.2.4 Temperature Increasing or decreasing the temperature of emulsions is one of the most important means of inducing phase inversion. The mechanism by which this process occurs depends on whether the phase inversion is induced by surfactant changes or fat crystallization. Cooling an oil-in-water emulsion to a temperature where the oil partly crystallizes and shearing causes fat crystallization-induced phase inversion. On the other hand, heating an oil-in-water emulsion stabilized by a nonionic surfactant may cause surfactantinduced phase inversion above the PIT (Figure 7.52). Fat crystallization-induced phase inversion is the most important type in the food industry because it is an essential step in the manufacture of butter and margarine (Dickinson and Stainsby 1982). Surfactantinduced phase inversion may be important in emulsions stabilized by nonionic surfactants that must be heated to high temperatures, for example, for pasteurization, sterilization, or cooking. In these systems, it is important for the food manufacturer to ensure that the PIT of the emulsifier is above the highest temperature that the emulsion experiences during processing, storage, and handling. 7.9.3 Characterization of Phase Inversion When an emulsion changes from an O/W to a W/O system there are changes in the nature of the continuous phase, as well as possible changes in the concentration, size, and interactions of the disperse phase (droplets). Consequently, any analytical technique that is sensitive to one or more of these changes can be used to monitor phase inversion. In a typical experiment, one starts with an initial emulsion (either O/W or W/O) and well-defined solution and environmental conditions (e.g., pH, ionic strength, solution composition, and temperature). This emulsion is then stirred at a fixed rate using a high speed mixer, and either the system composition or temperature is changed in a controlled fashion. An appropriate analytical instrument is then used to measure some physicochemical property of the system that is sensitive to emulsion type (O/W versus W/O), for example, electrical conductivity, viscosity, droplet size, turbidity, etc. The critical composition or temperature where the emulsion undergoes a phase inversion is then determined. A phase inversion temperature (PIT) is determined for emulsions that undergo a ­temperature-induced inversion, whereas a phase inversion concentration (PIC) is determined for emulsions that undergo a composition-induced inversion. Another commonly used experimental protocol is to establish the emulsion inversion point (EIP) (Brooks et al. 1998). In this case, water (or oil) containing a known amount of emulsifier is placed in a vessel and stirred at a constant speed. Oil (or water) is then titrated into the stirred system and an O/W (or W/O) emulsion is initially formed. However, once the oil (or water) concentration exceeds a certain level, defined as the EIP, the emulsion undergoes a phase inversion from O/W to W/O (or vice versa). The PIT, PIC, and EIC may depend on a number of different parameters, such as oil type, emulsifier type, emulsifier concentration, solution composition, disperse phase volume fraction, mixing conditions (shearing rate, time, and type), rate of additive addition, and temperature. Consequently, experiments can be designed to examine one or more of these parameters, depending on which is most relevant for the system studied. Some of the analytical techniques that are commonly used to monitor phase inversion in emulsions are briefly discussed below: 370 Food Emulsions: Principles, Practices, and Techniques 7.9.3.1 Electrical Conductivity An O/W emulsion has an aqueous continuous phase with a high electrical conductivity, whereas a W/O emulsion has a lipid continuous phase with a low electrical conductivity. Thus, there is a dramatic reduction in electrical conductivity when an O/W emulsion inverts to a W/O emulsion, or vice versa (Gu et al. 2000, Allouche et al. 2004, Tyrode et al. 2005). Electrical conductivity meters can therefore be conveniently used to monitor emulsion phase inversion (Figure 7.53). 7.9.3.2 Rheology In general, the viscosity of an emulsion is proportional to the viscosity of the continuous phase, and increases as the effective droplet concentration (ϕ) increases (Chapter 8). When an emulsion undergoes a phase inversion, there is a change in the nature of the continuous phase (from water to oil, or vice versa), as well as a change in the effective droplet concentration; thus, there is usually an appreciable change in emulsion viscosity. Measurements of emulsion viscosity can therefore be used to follow phase inversion in emulsions (Allouche et al. 2004, Tyrode et al. 2005). 7.9.3.3 Optical Properties The overall appearance of an emulsion depends on the concentration and size of the droplets that it contains, consequently there should be changes in the optical properties (e.g., “lightness” or “chromaticity”) of an emulsion as a result of phase inversion (Keowmaneechai and McClements 2002). In optically opaque emulsions, these changes can be determined continuously and nondestructively using a colorimeter or a spectrophotometer in reflectance mode. In optically transparent emulsions, phase inversion can be measured using a spectrophotometer in transmission mode (Figure 7.52). 700 Conductivity (μs.cm–1) 600 500 400 300 W/O O/W 200 100 0 20 30 40 Temperature (°C) 50 60 FIGURE 7.53 Temperature dependence of the electrical conductivity of tetradecane oil-in-water nanoemulsions. The conductivity decreases above the PIT due to droplet coalescence and phase inversion. Emulsion Stability 371 7.9.3.4 Microscopy Measurements of the bulk physicochemical properties of emulsions (such as electrical conductivity, viscosity, or optical properties) provide little information about the nature and characteristics of the structural changes occurring during phase inversion. Information about these morphological changes can be obtained by examining samples of emulsion using an appropriate microscopy method (Binks and Lumsdon 2000). Microscopy is particularly powerful if a dye that preferentially partitions into either the oil or water phases is present in the system. Optical microscopy has been used to show that multiple emulsions (W/O/W or O/W/O) form around the phase inversion point in certain systems (Salager et al. 2004, Rondon-Gonzalez et al. 2006). 7.9.3.5 Droplet Size Analysis There is usually an appreciable change in the droplet size distribution of an emulsion at and around the phase inversion point. Consequently, measurements of the particle size distribution can be used to monitor phase inversion. However, it should be stressed that the measured particle size distribution is likely to depend on the shearing rate of the mixer used to agitate the emulsion during the phase inversion process (particularly for the new emulsion formed after phase inversion). In addition, it is often necessary to dilute emulsions prior to making particle size measurements and so it is important to use the appropriate continuous phase for dilution (e.g., water for a O/W emulsion and oil for a W/O emulsion). 7.9.3.6 Interfacial Tension Measurements of the interfacial tension between oil and aqueous phases can also be used to predict the likelihood that surfactant-induced phase inversion will occur in an emulsion (Brooks et al. 1998). The interfacial tension may be measured as the temperature is changed to determine the PIT, or as the additive concentration is changed to determine the PIC. Typically, the interfacial tension has a very low value around the phase inversion point, which can be measured using highly sensitive tensiometers (Brooks et al. 1998). 7.9.3.7 Coalescence Stability The large reduction in interfacial tension that occurs around the phase inversion point for surfactantinduced phase inversion leads to an appreciable increase in coalescence instability. Consequently, the phase inversion point can be established by measuring the droplet coalescence rate in emulsions, for example, by measuring the increase in droplet diameter with time (Kabalnov 1998). There is usually a large increase in the coalescence rate (decrease in emulsion lifetime) as a system approaches the phase inversion point (Figure 7.52). 7.9.3.8 Emulsion Miscibility Information about the type of emulsion (O/W or W/O) present under specific conditions can often be established by determining whether it is miscible with oil or water. This method can be applied in a number of different ways. Small aliquots of the emulsion to be analyzed are placed into one vessel containing oil and another vessel containing water (or buffer solution). If the emulsion is of the O/W type then it should rapidly disperse in water but not in oil, but if it is of the W/O type then it should rapidly disperse in oil but not in water. Alternatively, an aliquot of oil or of water can be placed on top of the emulsion to be analyzed. If the emulsion is of the O/W type then the water should rapidly disperse in it, whereas the oil should remain sitting on the top. Conversely, if the emulsion is of the W/O type then the oil should rapidly disperse in it, but the water should not. This process can sometimes be seen more clearly if a dye that is only soluble in one of the phases is present. 372 Food Emulsions: Principles, Practices, and Techniques 7.10 Chemical and Biochemical Stability The majority of this chapter has been concerned with the physical instability of food emulsions, rather than with their chemical instability. This is largely because emulsion scientists have historically focused mainly on the physical aspects of food emulsions. Nevertheless, there are many types of chemical or biochemical reaction that can have adverse effects on the quality of food emulsions, for example, color fading, flavor loss, lipid oxidation, and biopolymer hydrolysis. For this reason, there has been a growing interest in the influence of various chemical and biochemical reactions on the stability of food emulsions. 7.10.1 Lipid Oxidation One of the most common forms of instability in foods that contain fats is lipid oxidation (McClements and Decker 2000, Waraho et al. 2011, Berton-Carabin et al. 2014). Lipid oxidation leads to the development of undesirable “off-flavors” (rancidity) and potentially toxic reaction products. In addition, it may also promote the physical instability of some emulsions. For example, many of the reaction products generated during lipid oxidation are surface active, and may therefore be able to interact with the interfacial layer surrounding the droplets in such a way as to lead to droplet coalescence. The importance of lipid oxidation in food emulsions has led to a considerable amount of research being carried out in this area over the past few years. The main emphasis of this work has been to develop effective strategies for retarding lipid oxidation in emulsions by ensuring high-quality ingredients, incorporating antioxidants, removing or deactiviating prooxidants, controlling storage conditions, or engineering droplet interfacial properties. One of the most important factors that has been identified is the role of transition metals (such as iron or copper) that are potent prooxidants. Lipid oxidation in many food emulsions can be greatly improved by deactivating transition metal ions using effective chelating agents. It should also be noted that lipid oxidation may promote oxidation of adsorbed or nonadsorbed proteins in an emulsion, and that this may alter their nutritional and functional properties (Rampon et al. 2001). 7.10.2 Enzyme Hydrolysis There has been considerable interest in the influence of enzymes on the properties of food emulsions recently for a number of reasons. Enzymes can be used to cross-link certain types of interfacial layers surrounding lipid droplets, thereby modifying their functional performance (Zeeb et al. 2014). This method may be used to improve the physical or chemical stability of emulsions during storage, or to create novel structures or textures in emulsions. The response of food emulsions to enzyme hydrolysis has become a major focus of research efforts recently due to the desire to better understand the gastrointestinal fate of emulsified foods after ingestion (McClements et al. 2009, Singh 2011, McClements 2013). Proteases within the stomach and small intestine may hydrolyze adsorbed or nonadsorbed proteins in food emulsions, thereby altering the physicochemical stability of the emulsion in the GIT (Mackie and Macierzanka 2010). Similarly, gastric and pancreatic lipases in the stomach and small intestine may hydrolyze the lipid phase within food emulsions (Reis et al. 2009). Studies have been carried out to determine the influence of protein and lipid type, emulsion microstructure, and solution composition on the rate and extent of enzyme hydrolysis in the GIT (McClements and Li 2010). This information may be used to design food products that perform in the human gastrointestinal tract in a particular way, for example, release nutrients or bioactive components at a particular location (mouth, stomach, small intestine, or colon), to increase the bioavailability of nutraceuticals, or to induce a satiety response (van Aken 2010). A more detailed discussion of the GIT fate of food emulsions is given in Chapter 11. 7.10.3 Flavor and Color Degradation Food emulsions contain a variety of additives to improve their perceived sensory attributes, including flavors and colors (Chapter 4). Many of these additives are susceptible to chemical degradation in Emulsion Stability 373 the product during storage, which can lead to a loss of overall product quality and acceptability. For example, citral is one of the most important flavor constituents in the lemon, lime, and orange oils used as flavorants in foods and beverages (Tan 1997). Citral rapidly decomposes at acidic pH due to a series of cyclization and oxidation reactions (Tan 1997, Ueno et al. 2004). The acid-catalyzed cyclization of citral decreases the desirable fresh-like aroma of citral and generates undesirable off-flavors (Kimura et al. 1983, Peacock and Kuneman 1985, Schieberle and Grosch 1998). Consequently, there has been a considerable amount of research aimed at understanding the factors that influence citral degradation (Choi et al. 2010a,b, Yang et al. 2011). Many food colors are also susceptible to degradation in products leading to color fading and loss of desirable visual appearance. For example, carotenoids (such as β-carotene, zeaxanthin, astaxanthin, or lycopene) are natural oil-soluble pigments used to give red, pink, orange, or yellow colors to many food products (Brady 2013). However, carotenoids are conjugated polyunsaturated lipids that are highly prone to lipid oxidation (Boon et al. 2009). Numerous factors promote the oxidation of carotenoids, including highly acidic environments (Konovalov and Kispert 1999), light (Mortensen and Skibsted 1996), elevated temperatures (Mader 1964), singlet oxygen (Krinsky 1998), transition metals (Williams et al. 2001, Gao and Kispert 2003), and free radicals (Liebler and Mc Clure 1996, Woodall et al. 1997). Research has been carried out to identify effective strategies to inhibit carotenoid oxidation in emulsions, including adding antioxidants, using chelating agents, utilizing encapsulation techniques, and engineering interfacial layers (Boon et al. 2009). Natural water-soluble pigments, such as anthocyanins, are also highly susceptible to chemical degradation during storage, with the rate and extent of degradation depending on their chemical structure, concentration, pH, temperature, oxygen, light, polymeric form, and the presence of cofactors and/or ascorbic acid (Poei-Langston and Wrolstad 1981, Francis and Markakis 1989, Bridle and Timberlake 1997, Marín et al. 2002). Foods contain a variety of flavor and color molecules that degrade through different mechanisms, and much more research is required to identify how emulsion properties influence the rate, extent, and pathway of degradation. This information could then be used to develop effective strategies to inhibit these undesirable reactions. REFERENCES Aguilera, J. M., D. W. Stanley, and G. V. Barbosa-Cánovas (1999). Microstructural Principles of Food Processing Engineering. Gaithersburg, MD: Aspen Publishers. Akoh, C. C. and D. B. Min (2008). Food Lipids: Chemistry, Nutrition, and Biotechnology. Boca Raton, FL: CRC Press. Allouche, J., E. Tyrode, V. Sadtler, L. Choplin, and J. L. Salager (2004). Simultaneous conductivity and viscosity measurements as a technique to track emulsion inversion by the phase-inversion-temperature method. Langmuir 20(6): 2134–2140. Atkins, P. and J. de Paula (2014). Physical Chemistry: Thermodynamics, Structure, and Change. Oxford, U.K.: Oxford University Press. Avendano-Gomez, J. R., J. L. Grossiord, and D. Clausse (2005). Study of mass transfer in oil-water-oil multiple emulsions by differential scanning calorimetry. Journal of Colloid and Interface Science 290(2): 533–545. Awad, T., Y. Hamada, and K. Sato (2001). Effects of addition of diacylglycerols on fat crystallization in oil-inwater emulsion. European Journal of Lipid Science and Technology 103(11): 735–741. Awad, T. and K. Sato (2001). Effects of hydrophobic emulsifier additives on crystallization behavior of palm mid fraction in oil-in-water emulsion. Journal of the American Oil Chemists Society 78(8): 837–842. Awad, T. and K. Sato (2002). Acceleration of crystallisation of palm kernel oil in oil-in-water emulsion by hydrophobic emulsifier additives. Colloids and Surfaces B: Biointerfaces 25(1): 45–53. Berton-Carabin, C. C., M.-H. Ropers, and C. Genot (2014). Lipid oxidation in oil-in-water emulsions: Involvement of the interfacial layer. Comprehensive Reviews in Food Science and Food Safety 13(5): 945–977. Binks, B. P. (1998). Emulsions—Recent advances in understanding. In Modern Aspects of Emulsion Science, B. P. Binks, ed., pp. 1–55. Cambridge, U.K.: The Royal Society of Chemistry. 374 Food Emulsions: Principles, Practices, and Techniques Binks, B. P. and S. O. Lumsdon (2000). Transitional phase inversion of solid-stabilized emulsions using particle mixtures. Langmuir 16(8): 3748–3756. Blijdenstein, T. B. J., A. J. M. van Winden, T. van Vliet, E. van der Linden, and G. A. van Aken (2004). Serum separation and structure of depletion- and bridging-flocculated emulsions: A comparison. Colloids and Surfaces A: Physicochemical and Engineering Aspects 245(1–3): 41–48. Boode, K. (1992). Partial coalescence in oil-in-water emulsions. PhD thesis, Wageningen Agricultural University, Wageningen, the Netherlands. Boode, K., C. Bisperink, and P. Walstra (1991). Destabilization of O/W emulsions containing fat crystals by temperature cycling. Colloids and Surfaces 61: 55–74. Boode, K. and P. Walstra (1993). Partial coalescence in oil-in-water emulsions. 1. Nature of the aggregation. Colloids and Surfaces A: Physicochemical and Engineering Aspects 81: 121–137. Boode, K., P. Walstra, and A. E. A. Degrootmostert (1993). Partial coalescence in oil-in-water emulsions. 2. Influence of the properties of the fat. Colloids and Surfaces A: Physicochemical and Engineering Aspects 81: 139–151. Boon, C. S., D. J. McClements, J. Weiss, and E. A. Decker (2009). Role of iron and hydroperoxides in the degradation of lycopene in oil-in-water emulsions. Journal of Agricultural and Food Chemistry 57(7): 2993–2998. Brady, J. W. (2013). Introductory Food Chemistry. Ithaca, NY: Cornell University Press. Bremer, L. G. B., B. H. Bijsterbosch, P. Walstra, and T. van Vliet (1993). Formation, properties and fractal structure of particle gels. Advances in Colloid and Interface Science 46: 117–128. Bremond, N. and J. Bibette (2012). Exploring emulsion science with microfluidics. Soft Matter 8(41): 10549–10559. Bresson, S., D. Rousseau, S. Ghosh, M. El Marssi, and V. Faivre (2011). Raman spectroscopy of the polymorphic forms and liquid state of cocoa butter. European Journal of Lipid Science and Technology 113(8): 992–1004. Bridle, P. and C. F. Timberlake (1997). Anthocyanins as natural food colours—Selected aspects. Food Chemistry 58(1–2): 103–109. Brooks, B. W., H. N. Richmond, and M. Zerfa (1998). Phase inversion and drop formation in agitated liquidliquid dispersions in the presence of non-ionic surfactants. In Modern Aspects of Emulsion Science, B. P. Binks, ed., pp. 175–204. Cambridge, U.K.: The Royal Society of Chemistry. Burger, R., F. Concha, K. K. Fjelde, and K. H. Karlsen (2000). Numerical simulation of the settling of polydisperse suspensions of spheres. Powder Technology 113(1–2): 30–54. Bushell, G. C., Y. D. Yan, D. Woodfield, J. Raper, and R. Amal (2002). On techniques for the measurement of the mass fractal dimension of aggregates. Advances in Colloid and Interface Science 95(1): 1–50. Campbell, I. J. (1989). The role of fat crystals in emulsion stability. In Food Colloids, R. D. Bee, P. Richmond, and J. Mingins, eds., p. 272. Cambridge, U.K.: The Royal Society of Chemistry. Chanamai, R. and D. J. McClements (2000a). Creaming stability of flocculated monodisperse oil-in-water emulsions. Journal of Colloid and Interface Science 225(1): 214–218. Chanamai, R. and D. J. McClements (2000b). Dependence of creaming and rheology of monodisperse oilin-water emulsions on droplet size and concentration. Colloids and Surfaces A: Physicochemical and Engineering Aspects 172(1–3): 79–86. Chang, Y., L. McLandsborough, and D. J. McClements (2012). Physical properties and antimicrobial efficacy of thyme oil nanoemulsions: Influence of ripening inhibitors. Journal of Agricultural and Food Chemistry 60(48): 12056–12063. Cheetangdee, N. and K. Fukada (2012). Protein stabilized oil-in-water emulsions modified by uniformity of size by premix membrane extrusion and their colloidal stability. Colloids and Surfaces A: Physicochemical and Engineering Aspects 403: 54–61. Cheetangdee, N., M. Oki, and K. Fukada (2011). The coalescence stability of protein-stabilized emulsions estimated by analytical photo-centrifugation. Journal of Oleo Science 60(8): 419–427. Chen, J. S., E. Dickinson, and G. Iveson (1993). Interfacial interactions, competitive adsorption and emulsion stability. Food Structure 12(2): 135–146. Choi, S. J., E. A. Decker, L. Henson, L. M. Popplewell, and D. J. McClements (2010a). Influence of droplet charge on the chemical stability of citral in oil-in-water emulsions. Journal of Food Science 75(6): C536–C540. Emulsion Stability 375 Choi, S. J., E. A. Decker, L. Henson, L. M. Popplewell, and D. J. McClements (2010b). Inhibition of citral degradation in model beverage emulsions using micelles and reverse micelles. Food Chemistry 122(1): 111–116. Clausse, D., F. Gomez, I. Pezron, L. Komunjer, and C. Dalmazzone (2005). Morphology characterization of emulsions by differential scanning calorimetry. Advances in Colloid and Interface Science 117(1–3): 59–74. Coupland, J. N. and D. J. McClements (1996). Lipid oxidation in food emulsions. Trends in Food Science & Technology 7(3): 83–91. Darling, D. F. (1982). Recent advances in the destabilization of dairy emulsions. Journal of Dairy Research 49(4): 695–712. Das, A. K. and P. K. Ghosh (1990). Concentrated emulsions—Investigation of polydispersity and droplet distortion and their effect on volume fraction and interfacial area. Langmuir 6(11): 1668–1675. Davis, R. H. and H. Gecol (1994). Hindered settling function with no empirical parameters for polydisperse suspensions. AIChE Journal 40(3): 570–575. Degner, B. M., C. Chung, V. Schlegel, R. Hutkins, and D. J. McClements (2014). Factors influencing the freeze-thaw stability of emulsion-based foods. Comprehensive Reviews in Food Science and Food Safety 13(2): 98–113. Demetriades, K. and D. J. McClements (1998). Influence of dextran sulfate and NaCl on the flocculation of oil-in-water emulsions stabilized by a nonionic surfactant. Journal of Agricultural and Food Chemistry 46(10): 3929–3935. Demiere, B., A. Colin, G. Leal-Calderon, and J. Bibette (1998). Lifetime and destruction of concentrated emulsions undergoing coalescence. In Modern Aspects of Emulsion Science, B. P. Binks, ed. pp. 261–291. Cambridge, U.K.: The Royal Society of Chemistry. Deminiere, B., A. Colin, F. L. Calderon, and J. Bibette (1998). Lifetime and destruction of concentrated emulsions undergoing coalescence. In Modern Aspects of Emulsion Science, B. P. Binks, ed., pp. 261–291. Cambridge, U.K.: The Royal Society of Chemistry. Dickinson, E. (1992a). An Introduction to Food Colloids. Oxford, U.K.: Oxford Science Publishers. Dickinson, E. (1992b). Structure and composition of adsorbed protein layers and the relationship to emulsion stability. Journal of the Chemical Society-Faraday Transactions 88(20): 2973–2983. Dickinson, E. (2003). Hydrocolloids at interfaces and the influence on the properties of dispersed systems. Food Hydrocolloids 17(1): 25–39. Dickinson, E. (2007). Colloidal systems containing droplets and bubbles. In Understanding and Controlling the Microstructure of Complex Foods, D. J. McClements, ed. pp. 153–184. Cambridge, U.K.: Woodhead Publishing. Dickinson, E. (2011). Mixed biopolymers at interfaces: Competitive adsorption and multilayer structures. Food Hydrocolloids 25(8): 1966–1983. Dickinson, E. (2013). Structure and rheology of colloidal particle gels: Insight from computer simulation. Advances in Colloid and Interface Science 199: 114–127. Dickinson, E. and M. Golding (1998). Influence of alcohol on stability of oil-in-water emulsions containing sodium caseinate. Journal of Colloid and Interface Science 197(1): 133–141. Dickinson, E., B. S. Murray, and G. Stainsby (1988). Coalescence stability of emulsion-sized droplets at a planar oil-water interface and the relationship to protein film surface rheology. Journal of the Chemical Society-Faraday Transactions I 84: 871–883. Dickinson, E., R. K. Owusu, and A. Williams (1993). Orthokinetic destabilization of a protein-stabilized emulsion by a water-soluble surfactant. Journal of the Chemical Society-Faraday Transactions 89(5): 865–866. Dickinson, E., C. Ritzoulis, Y. Yamamoto, and H. Logan (1999). Ostwald ripening of protein-stabilized emulsions: Effect of transglutaminase crosslinking. Colloids and Surfaces B: Biointerfaces 12(3–6): 139–146. Dickinson, E. and G. Stainsby (1982). Colloids in Foods. London, U.K.: Elsevier. Dickinson, E. and A. Williams (1994). Orthokinetic coalescence of protein-stabilized emulsions. Colloids and Surfaces A: Physicochemical and Engineering Aspects 88(2–3): 317–326. Dresselhuis, D. M., M. A. C. Stuart, G. A. van Aken, R. G. Schipper, and E. H. A. de Hoog (2008). Fat retention at the tongue and the role of saliva: Adhesion and spreading of ‘protein-poor’ versus ‘protein-rich’ emulsions. Journal of Colloid and Interface Science 321(1): 21–29. Dufour, E., G. Mazerolles, M. F. Devaux, G. Duboz, M. H. Duployer, and M. N. Riou (2000). Phase transition of triglycerides during semi-hard cheese ripening. International Dairy Journal 10(1–2): 81–93. 376 Food Emulsions: Principles, Practices, and Techniques Dukhin, S. S., N. A. Mishchuk, G. Loglio, L. Liggieri, and R. Miller (2003). Coalescence coupling with flocculation in dilute emulsions within the primary and/or secondary minimum. Advances in Colloid and Interface Science 100: 47–81. Dukhin, S. S. and J. Sjoblom (1996). Kinetics of Brownian and gravitational coagulation in dilute emulsions. In Emulsions and Emulsion Stability, J. Sjoblom, ed. New York: Marcel Dekker. Dukhin, S. S., J. Sjoblom, D. T. Wasan, and O. Saether (2001). Coalescence coupled with either coagulation or flocculation in dilute emulsions. Colloids and Surfaces A: Physicochemical and Engineering Aspects 180(3): 223–234. Elwell, M. W., R. F. Roberts, and J. N. Coupland (2004). Effect of homogenization and surfactant type on the exchange of oil between emulsion droplets. Food Hydrocolloids 18(3): 413–418. Ergun, R., R. W. Hartel, and P. T. Spicer (2015). Kinetic effects on interfacial partitioning of fat crystals. Food Structure, 5: 1–9. Euston, S. R., S. R. Finnigan, and R. L. Hirst (2001). Heat-induced destabilization of oil-in-water emulsions formed from hydrolyzed whey protein. Journal of Agricultural and Food Chemistry 49(11): 5576–5583. Evans, E. D. and W. Wennerstrom (1999). The Colloidal Domain: Where Physics, Chemistry and Biology Meet. New York: Wiley-VCH. Francis, F. J. and P. C. Markakis (1989). Food colorants: Anthocyanins. Critical Reviews in Food Science and Nutrition 28(4): 273–314. Fredrick, E., P. Walstra, and K. Dewettinck (2010). Factors governing partial coalescence in oil-in-water emulsions. Advances in Colloid and Interface Science 153(1–2): 30–42. Friberg, S. and K. Larsson (1997). Food Emulsions. New York: Marcel Dekker. Gallier, S., H. Tate, and H. Singh (2013). In vitro gastric and intestinal digestion of a walnut oil body dispersion. Journal of Agricultural and Food Chemistry 61(2): 410–417. Gao, Y. L. and L. D. Kispert (2003). Reaction of carotenoids and ferric chloride: Equilibria, isomerization, and products. Journal of Physical Chemistry B 107(22): 5333–5338. Goff, H. D. (1997). Instability and partial coalescence in whippable dairy emulsions. Journal of Dairy Science 80(10): 2620–2630. Goff, H. D. and R. W. Hartel (2013). Ice Cream. New York: Springer. Gonzalez, A. E., G. Odriozola, and R. Leone (2004). Colloidal aggregation with sedimentation: Concentration effects. European Physical Journal E 13(2): 165–178. Gu, Y. X., Y. H. Huang, B. Liao, G. M. Cong, and M. Xu (2000). Studies on the characterization of phase inversion during emulsification process and the particle sizes of water-borne microemulsion of poly(phenylene oxide) ionomer. Journal of Applied Polymer Science 76(5): 690–694. Guzey, D. and D. J. McClements (2006). Formation, stability and properties of multilayer emulsions for application in the food industry. Advances in Colloid and Interface Science 128: 227–248. Hartel, R. W. (2001). Crystallization in Foods. Gaithersburg, MD: Aspen Publishers. Helgason, T., T. S. Awad, K. Kristbergsson, D. J. McClements, and J. Weiss (2009). Effect of surfactant surface coverage on formation of solid lipid nanoparticles (SLN). Journal of Colloid and Interface Science, 334: 75–81. Hiemenz, P. C. and R. Rajagopalan (1997). Principles of Colloid and Surface Chemistry. New York: Marcel Dekker. Hotrum, N. E., M. A. C. Stuart, T. van Vliet, S. F. Avino, and G. A. van Aken (2005). Elucidating the relationship between the spreading coefficient, surface-mediated partial coalescence and the whipping time of artificial cream. Colloids and Surfaces A: Physicochemical and Engineering Aspects 260(1–3): 71–78. Hunter, R. J. (1986). Foundations of Colloid Science. Oxford, U.K.: Oxford University Press. Hunter, R. J. (1989). Foundations of Colloid Science. Oxford, U.K.: Oxford University Press. Israelachvili, J. (2011). Intermolecular and Surface Forces, 3rd edn. London, U.K.: Academic Press. Ivanov, I. B., K. D. Danov, and P. A. Kralchevsky (1999). Flocculation and coalescence of micron-size emulsion droplets. Colloids and Surfaces A: Physicochemical and Engineering Aspects 152(1–2): 161–182. Jenkins, P. and M. Snowden (1996). Depletion flocculation in colloidal dispersions. Advances in Colloid and Interface Science 68: 57–96. Jimenez-Colmenero, F. (2013). Potential applications of multiple emulsions in the development of healthy and functional foods. Food Research International 52(1): 64–74. Emulsion Stability 377 Jores, K., W. Mehnert, M. Drechsler, H. Bunjes, C. Johann, and K. Mader (2004). Investigations on the structure of solid lipid nanoparticles (SLN) and oil-loaded solid lipid nanoparticles by photon correlation spectroscopy, field-flow fractionation and transmission electron microscopy. Journal of Controlled Release 95(2): 217–227. Kabalnov, A. (1998). Coalescence in emulsions. In Modern Aspects of Emulsion Science, B. P. Binks, ed., pp. 205–260. Cambridge, U.K.: The Royal Society of Chemistry. Kabalnov, A. (2001). Ostwald ripening and related phenomena. Journal of Dispersion Science and Technology 22(1): 1–12. Kabalnov, A. and H. Wennerstrom (1996). Macroemulsion stability: The oriented wedge theory revisited. Langmuir 12(2): 276–292. Kabalnov, A. S. (1998). Coalescence in emulsions. In Modern Aspects of Emulsion Science, B. P. Binks, ed. pp. 205–260. Cambridge, U.K.: The Royal Society of Chemistry. Kabalnov, A. S., A. V. Pertzov, and E. D. Shchukin (1987). Ostwald ripening in 2-component disperse phase systems—Application to emulsion stability. Colloids and Surfaces 24(1): 19–32. Kabalnov, A. S. and E. D. Shchukin (1992). Ostwald ripening theory—Applications to fluorocarbon emulsion stability. Advances in Colloid and Interface Science 38: 69–97. Keowmaneechai, E. and D. J. McClements (2002a). Effect of CaCl2 and KCl on physiochemical properties of model nutritional beverages based on whey protein stabilized oil-in-water emulsions. Journal of Food Science 67(2): 665–671. Keowmaneechai, E. and D. J. McClements (2002b). Influence of EDTA and citrate on physicochemical properties of whey protein-stabilized oil-in-water emulsions containing CaCl2. Journal of Agricultural and Food Chemistry 50(24): 7145–7153. Keowmaneechai, E. and D. J. McClements (2006). Influence of EDTA and citrate on thermal stability of whey protein stabilized oil-in-water emulsions containing calcium chloride. Food Research International 39(2): 230–239. Kim, H. J., E. A. Decker, and D. J. McClements (2002a). Impact of protein surface denaturation on droplet flocculation in hexadecane oil-in-water emulsions stabilized by beta-lactoglobulin. Journal of Agricultural and Food Chemistry 50(24): 7131–7137. Kim, H. J., E. A. Decker, and D. J. McClements (2002b). Role of postadsorption conformation changes of beta-lactoglobulin on its ability to stabilize oil droplets against flocculation during heating at neutral pH. Langmuir 18(20): 7577–7583. Kimura, K., H. Nishimura, I. Iwata, and J. Mizutani (1983). Deterioration mechanism of lemon flavor. 2. Formation mechanism of off-odor substances arising from citral. Journal of Agricultural and Food Chemistry 31: 801–804. Kippax, P., J. D. Sherwood, and D. J. McClements (1999). Ultrasonic spectroscopy study of globule aggregation in parenteral fat emulsions containing calcium chloride. Langmuir 15(5): 1673–1678. Kleshchanok, D., R. Tuinier, and P. R. Lang (2008). Direct measurements of polymer-induced forces. Journal of Physics-Condensed Matter 20(7): 1–25. Konovalov, V. V. and L. D. Kispert (1999). AM1, INDO/S and optical studies of carbocations of carotenoid molecules. Acid induced isomerization. Journal of the Chemical Society-Perkin Transactions 2(4): 901–909. Krebs, T., C. Schroen, and R. M. Boom (2013). Coalescence kinetics of oil-in-water emulsions studied with microfluidics. Fuel 106: 327–334. Krinsky, N. I. (1998). The antioxidant and biological properties of the carotenoids. Towards Prolongation of the Healthy Life Span 854: 443–447. Kulmyrzaev, A., R. Chanamai, and D. J. McClements (2000). Influence of pH and CaCl2 on the stability of dilute whey protein stabilized emulsions. Food Research International 33(1): 15–20. Kulmyrzaev, A. A. and H. Schubert (2004). Influence of KCl on the physicochemical properties of whey protein stabilized emulsions. Food Hydrocolloids 18(1): 13–19. Kuntsche, J., J. C. Horst, and H. Bunjes (2011). Cryogenic transmission electron microscopy (cryo-TEM) for studying the morphology of colloidal drug delivery systems. International Journal of Pharmaceutics 417(1–2): 120–137. Lattuada, M. (2012). Predictive model for diffusion-limited aggregation kinetics of nanocolloids under high concentration. Journal of Physical Chemistry B 116(1): 120–129. 378 Food Emulsions: Principles, Practices, and Techniques Li, Y., S. Le Maux, H. Xiao, and D. J. McClements (2009). Emulsion-based delivery systems for tributyrin, a potential colon cancer preventative agent. Journal of Agricultural and Food Chemistry 57(19): 9243–9249. Liao, Y. X. and D. Lucas (2010). A literature review on mechanisms and models for the coalescence process of fluid particles. Chemical Engineering Science 65(10): 2851–2864. Liebler, D. C. and T. D. Mc Clure (1996). Antioxidant reactions of beta-carotene: Identification of carotenoidradical adducts. Chemical Research in Toxicology 9(1): 8–11. Loren, N., M. Langton, and A. M. Hermansson (2007). Confocal fluorescence microscopy for structure characterization. In Understanding and Controlling the Microstructure of Complex Foods, D. J. McClements, ed., pp.232–260. Cambridge, U.K.: Woodhead Publishing. Mackie, A. and A. Macierzanka (2010). Colloidal aspects of protein digestion. Current Opinion in Colloid & Interface Science 15(1–2): 102–108. Mader, I. (1964). Beta-carotene—Thermal degradation. Science 144(361): 533. Marín, F. R., M. J. Frutos, J. A. Pérez-Alvarez, F. Martinez-Sánchez, and J. A. Del Río (2002). Flavonoids as nutraceuticals: Structural related antioxidant properties and their role on ascorbic acid preservation. In Studies in Natural Products Chemistry, R. Attaur, ed., vol. 26, Part G, pp. 741–778. New York: Elsevier. Mayhill, P. G. and D. F. Newstead (1992). The effect of milkfat fractions and emulsifier type on creaming in normal-solids uht recombined milk. Milchwissenschaft-Milk Science International 47(2): 75–79. McClements, D. J. (2000). Comments on viscosity enhancement and depletion flocculation by polysaccharides. Food Hydrocolloids 14(2): 173–177. McClements, D. J. (2005). Food Emulsions: Principles, Practice, and Techniques. Boca Raton, FL: CRC Press. McClements, D. J. (2007). Critical review of techniques and methodologies for characterization of emulsion stability. Critical Reviews in Food Science and Nutrition 47(7): 611–649. McClements, D. J. (2011). Edible nanoemulsions: Fabrication, properties, and functional performance. Soft Matter 7(6): 2297–2316. McClements, D. J. (2012). Crystals and crystallization in oil-in-water emulsions: Implications for emulsionbased delivery systems. Advances in Colloid and Interface Science 174: 1–30. McClements, D. J. (2013). Edible lipid nanoparticles: Digestion, absorption, and potential toxicity. Progress in Lipid Research 52(4): 409–423. McClements, D. J. and E. A. Decker (2000). Lipid oxidation in oil-in-water emulsions: Impact of molecular environment on chemical reactions in heterogeneous food systems. Journal of Food Science 65(8): 1270–1282. McClements, D. J., E. A. Decker, and Y. Park (2009). Controlling lipid bioavailability through physicochemical and structural approaches. Critical Reviews in Food Science and Nutrition 49(1): 48–67. McClements, D. J., L. Henson, L. M. Popplewell, E. A. Decker, and S. J. Choi (2012). Inhibition of ostwald ripening in model beverage emulsions by addition of poorly water soluble triglyceride oils. Journal of Food Science 77(1): C33–C38. McClements, D. J. and Y. Li (2010). Structured emulsion-based delivery systems: Controlling the digestion and release of lipophilic food components. Advances in Colloid and Interface Science 159(2): 213–228. McClements, D. J. and J. Rao (2011). Food-grade nanoemulsions: Formulation, fabrication, properties, performance, biological fate, and potential toxicity. Critical Reviews in Food Science and Nutrition 51(4): 285–330. Mei, L. Y., E. A. Decker, and D. J. McClements (1998). Evidence of iron association with emulsion droplets and its impact on lipid oxidation. Journal of Agricultural and Food Chemistry 46(12): 5072–5077. Melik, D. H. and H. S. Fogler (1988). Fundamentals of colloidal stability in quiescent media. In Encyclopedia of Emulsion Technology, P. Becher, ed., vol. 3., pp. 3–78. New York: Marcel Dekker. Mengual, O., G. Meunier, I. Cayre, K. Puech, and P. Snabre (1999). TURBISCAN MA 2000: Multiple light scattering measurement for concentrated emulsion and suspension instability analysis. Talanta 50(2): 445–456. Mikula, R. J. (1992). Emulsion characterization. In Emulsions: Fundamentals and Applications in the Petroleum Industry, L. L. Schramm, ed., Chapter 3. Washington, DC: American Chemical Society. Emulsion Stability 379 Mishchuk, N. A. (2005). Theoretical analysis of coagulation kinetics in Brownian disperse systems. Colloid Journal 67(3): 341–350. Mortensen, A. and L. H. Skibsted (1996). Kinetics of photobleaching of beta-carotene in chloroform and formation of transient carotenoid species absorbing in the near infrared. Free Radical Research 25(4): 355–368. Mun, S. H. and D. J. McClements (2006). Influence of interfacial characteristics on Ostwald ripening in hydrocarbon oil-in-water emulsions. Langmuir 22(4): 1551–1554. Pal, R. (1994). Techniques for measuring the composition (oil and water-content) of emulsions—A state-ofthe-art review. Colloids and Surfaces A: Physicochemical and Engineering Aspects 84(2–3): 141–193. Pal, R. (1996). Rheology of emulsions containing polymeric liquids. In Encyclopedia of Emulsion Technology, P. Becher, ed., vol. 4., pp.93–264. New York: Marcel Dekker. Palanuwech, J. and J. N. Coupland (2003). Effect of surfactant type on the stability of oil-in-water emulsions to dispersed phase crystallization. Colloids and Surfaces A: Physicochemical and Engineering Aspects 223(1–3): 251–262. Palanuwech, J., R. Potineni, R. F. Roberts, and J. N. Coupland (2003). A method to determine free fat in emulsions. Food Hydrocolloids 17(1): 55–62. Pawar, A. B., M. Caggioni, R. W. Hartel, and P. T. Spicer (2012). Arrested coalescence of viscoelastic droplets with internal microstructure. Faraday Discussions, 158: 341–350. Peacock, V. E. and D. W. Kuneman (1985). Inhibition of the formation of α-p-dimethylstyrene and p-cymen8-ol in a carbonated citral-containing beverage system. Journal of Agricultural and Food Chemistry 33: 330–335. Peleg, M. (1993). Fractals and foods. Critical Reviews in Food Science and Nutrition 33(2): 149–165. Pena, A. A. and C. A. Miller (2006). Solubilization rates of oils in surfactant solutions and their relationship to mass transport in emulsions. Advances in Colloid and Interface Science 123: 241–257. Petsev, D. N. (2000). Theoretical analysis of film thickness transition dynamics and coalescence of charged miniemulsion droplets. Langmuir 16(5): 2093–2100. Pinfield, V. J., E. Dickson, and M. J. W. Povey (1997). Modeling of combined creaming and flocculation in emulsions. Journal of Colloid and Interface Science 186(1): 80–89. Piorkowski, D. T. and D. J. McClements (2014). Beverage emulsions: Recent developments in formulation, production, and applications. Food Hydrocolloids 42: 5–41. Poei-Langston, M. S. and R. E. Wrolstad (1981). Color degradation in an ascorbic acid-anthocyanin-flavanol model system. Journal of Food Science 46(4): 1218–1236. Qian, C., E. A. Decker, H. Xiao, and D. J. McClements (2012). Solid lipid nanoparticles: Effect of carrier oil and emulsifier type on phase behavior and physical stability. Journal of the American Oil Chemists Society 89(1): 17–28. Quemada, D. and C. Berli (2002). Energy of interaction in colloids and its implications in rheological modeling. Advances in Colloid and Interface Science 98(1): 51–85. Radford, S. J. and E. Dickinson (2004). Depletion flocculation of caseinate-stabilised emulsions: What is the optimum size of the non-adsorbed protein nano-particles? Colloids and Surfaces A: Physicochemical and Engineering Aspects 238(1–3): 71–81. Radford, S. J., E. Dickinson, and M. Golding (2004). Stability and rheology of emulsions containing sodium caseinate: Combined effects of ionic calcium and alcohol. Journal of Colloid and Interface Science 274(2): 673–686. Ramkumar, C., H. Singh, P. A. Munro, and A. M. Singh (2000). Influence of calcium, magnesium, or potassium ions on the formation and stability of emulsions prepared using highly hydrolyzed whey proteins. Journal of Agricultural and Food Chemistry 48(5): 1598–1604. Rampon, V., L. Lethuaut, N. Mouhous-Riou, and C. Genot (2001). Interface characterization and aging of bovine serum albumin stabilized oil-in-water emulsions as revealed by front-surface fluorescence. Journal of Agricultural and Food Chemistry 49(8): 4046–4051. Rao, J. and D. J. McClements (2012). Impact of lemon oil composition on formation and stability of model food and beverage emulsions. Food Chemistry 134(2): 749–757. Reis, P., K. Holmberg, H. Watzke, M. E. Leser, and R. Miller (2009). Lipases at interfaces: A review. Advances in Colloid and Interface Science 147–48: 237–250. Relkin, P. and S. Sourdet (2005). Factors affecting fat droplet aggregation in whipped frozen protein-­stabilized emulsions. Food Hydrocolloids 19(3): 503–511. 380 Food Emulsions: Principles, Practices, and Techniques Relkin, P., S. Sourdet, and P. Y. Fosseux (2003). Fat crystallization in complex food emulsions—Effects of adsorbed milk proteins and of a whipping process. Journal of Thermal Analysis and Calorimetry 71(1): 187–195. Robins, M. M. (2000). Emulsions—Creaming phenomena. Current Opinion in Colloid & Interface Science 5(5–6): 265–272. Robins, M. M., A. D. Watson, and P. J. Wilde (2002). Emulsions—Creaming and rheology. Current Opinion in Colloid & Interface Science 7(5–6): 419–425. Rojas, C., G. Urbina-Villalba, and M. Garcia-Sucre (2010). Lifetime of micrometer-sized drops of oil pressed by buoyancy against a planar interface. Physical Review E 81(1), Article 016302. Rondon-Gonzalez, M., V. Sadtler, L. Choplin, and J. L. Salager (2006). Emulsion inversion from abnormal to normal morphology by continuous stirring without internal phase addition—Effect of surfactant mixture fractionation at extreme water-oil ratio. Colloids and Surfaces A: Physicochemical and Engineering Aspects 288(1–3): 151–157. Rousseau, D. (2000). Fat crystals and emulsion stability—A review. Food Research International 33(1): 3–14. Rousseau, D. (2006). On the porous mesostructure of milk chocolate viewed with atomic force microscopy. LWT—Food Science and Technology 39(8): 852–860. Rousseau, D. and S. M. Hodge (2005). Stabilization of water-in-oil emulsions with continuous phase crystals. Colloids and Surfaces A: Physicochemical and Engineering Aspects 260(1–3): 229–237. Russ, J. C. (2004). Image Analysis of Food Microstructure. Boca Raton, FL: CRC Press. Saberi, A. H., Y. Fang, and D. J. McClements (2013). Effect of glycerol on formation, stability, and properties of vitamin-E enriched nanoemulsions produced using spontaneous emulsification. Journal of Colloid and Interface Science 411: 105–113. Saether, O., S. S. Dukhin, and J. Sjoblom (2004). Droplet flocculation and coalescence in dilute oil-in-water emulsions. In Food Emulsions, S. Friberg, K. Larsson, and J. Sjoblom, eds., pp.175–200. New York: Marcel Dekker. Salager, J. L. (1988). Phase transformation and emulsion inversion on the basis of the catastrophe theory. In Encyclopedia of Emulsion Technology, P. Becher, ed., vol. 3, pp.79–136. New York: Marcel Dekker. Salager, J. L., A. Forgiarini, L. Marquez, A. Pena, A. Pizzino, M. P. Rodriguez, and M. Rondo-Gonzalez (2004). Using emulsion inversion in industrial processes. Advances in Colloid and Interface Science 108: 259–272. Salager, J. L., L. Marquez, I. Mira, A. Pena, E. Tyrode, and N. B. Zambrano (2003). Principles of emulsion formulation engineering. In Adsorption and Aggregation of Surfactants in Solution, K. L. Mittal and D. O. Shah, eds., vol. 109, pp. 501–523. Boca Raton, FL: CRC Press. Sanfeld, A. and A. Steinchen (2008). Emulsions stability, from dilute to dense emulsions—Role of drops deformation. Advances in Colloid and Interface Science 140(1): 1–65. Sato, K. (1999). Solidification and phase transformation behaviour of food fats—A review. Fett-Lipid 101(12): 467–474. Schieberle, P. and W. Grosch (1998). Identification of potent flavor compounds formed in an aqueous lemon oil/citral acid emulsion. Journal of Agricultural and Food Chemistry 36: 797–800. Shinoda, K. and S. Friberg (1986). Emulsions and Solubilization. New York: John Wiley & Sons. Shukat, R., C. Bourgaux, and P. Relkin (2012). Crystallisation behaviour of palm oil nanoemulsions carrying vitamin E. Journal of Thermal Analysis and Calorimetry 108(1): 153–161. Singh, H. (2011). Aspects of milk-protein-stabilised emulsions. Food Hydrocolloids 25(8): 1938–1944. Spicer, P. T. and S. E. Pratsinis (1996a). Coagulation and fragmentation: Universal steady-state particle-size distribution. AIChE Journal 42(6): 1612–1620. Spicer, P. T. and S. E. Pratsinis (1996b). Shear-induced flocculation: The evolution of floc structure and the shape of the size distribution at steady state. Water Research 30(5): 1049–1056. Tadros, T. F. (1994). Fundamental principles of emulsion rheology and their applications. Colloids and Surfaces A: Physicochemical and Engineering Aspects 91: 39–55. Taisne, L. and B. Cabane (1998). Emulsification and ripening following a temperature quench. Langmuir 14(17): 4744–4752. Taisne, L., P. Walstra, and B. Cabane (1996). Transfer of oil between emulsion droplets. Journal of Colloid and Interface Science 184(2): 378–390. Tan, C.-T. (1997). Beverage emulsions. In Food Emulsions, S. E. Friberg and K. Larsson, eds., pp. 491–524. New York: Marcel Dekker. Emulsion Stability 381 Tanaka, H. and T. Araki (2000). Simulation method of colloidal suspensions with hydrodynamic interactions: Fluid particle dynamics. Physical Review Letters 85(6): 1338–1341. Taneja, A., A. Ye, J. R. Jones, R. Archer, and H. Singh (2013). Behaviour of oil droplets during spray drying of milk-protein-stabilised oil-in-water emulsions. International Dairy Journal 28(1): 15–23. Tang, D. M. and A. G. Marangoni (2006). Microstructure and fractal analysis of fat crystal networks. Journal of the American Oil Chemists Society 83(5): 377–388. Taylor, P. (1998). Ostwald ripening in emulsions. Advances in Colloid and Interface Science 75(2): 107–163. Tcholakova, S., N. D. Denkov, I. B. Ivanov, and B. Campbell (2002). Coalescence in beta-lactoglobulinstabilized emulsions: Effects of protein adsorption and drop size. Langmuir 18(23): 8960–8971. Tcholakova, S., N. D. Denkov, I. B. Ivanov, and B. Campbell (2006a). Coalescence stability of emulsions containing globular milk proteins. Advances in Colloid and Interface Science 123: 259–293. Tcholakova, S., N. D. Denkov, and A. Lips (2008). Comparison of solid particles, globular proteins and surfactants as emulsifiers. Physical Chemistry Chemical Physics 10(12): 1608–1627. Tcholakova, S., N. D. Denkov, D. Sidzhakova, and B. Campbell (2006b). Effect of thermal treatment, ionic strength, and pH on the short-term and long-term coalescence stability of beta-lactoglobulin emulsions. Langmuir 22(14): 6042–6052. Tcholakova, S., N. D. Denkov, D. Sidzhakova, I. B. Ivanov, and B. Campbell (2003). Interrelation between drop size and protein adsorption at various emulsification conditions. Langmuir 19(14): 5640–5649. Tcholakova, S., N. D. Denkov, D. Sidzhakova, I. B. Ivanov, and B. Campbell (2005). Effects of electrolyte concentration and pH on the coalescence stability of beta-lactoglobulin emulsions: Experiment and interpretation. Langmuir 21(11): 4842–4855. Thanasukarn, P., R. Pongsawatmanit, and D. J. McClements (2004a). Impact of fat and water crystallization on the stability of hydrogenated palm oil-in-water emulsions stabilized by whey protein isolate. Colloids and Surfaces A: Physicochemical and Engineering Aspects 246(1–3): 49–59. Thanasukarn, P., R. Pongsawatmanit, and D. J. McClements (2004b). Influence of emulsifier type on freezethaw stability of hydrogenated palm oil-in-water emulsions. Food Hydrocolloids 18(6): 1033–1043. Thivilliers, F., N. Drelon, V. Schmitt, and F. Leal-Calderon (2006). Bicontinuous emulsion gels induced by partial coalescence: Kinetics and mechanism. Europhysics Letters 76(2): 332–338. Thivilliers-Arvis, F., E. Laurichesse, V. Schmitt, and F. Leal-Calderon (2010). Shear-induced instabilities in oil-in-water emulsions comprising partially crystallized droplets. Langmuir 26(22): 16782–16790. Toro-Mendoza, J. and D. N. Petsev (2010). Brownian dynamics of emulsion film formation and droplet coalescence. Physical Review E 81(5), Article 051404. Tyrode, E., J. Allouche, L. Choplin, and J. L. Salager (2005). Emulsion catastrophic inversion from abnormal to normal morphology. 4. Following the emulsion viscosity during three inversion protocols and extending the critical dispersed-phase concept. Industrial & Engineering Chemistry Research 44(1): 67–74. Ueno, T., H. Masuda, and C.-T. Ho (2004). Formation mechanism of p-methylacetophenone from citral via a tert-alkoxy radical intermediate. Journal of Agricultural and Food Chemistry 52: 5677–5684. Urbina-Villalba, G., A. Forgiarini, K. Rahn, and A. Lozsan (2009). Influence of flocculation and coalescence on the evolution of the average radius of an O/W emulsion. Is a linear slope of (R)over-bar(3) vs. t an unmistakable signature of Ostwald ripening? Physical Chemistry Chemical Physics 11(47): 11184–11195. van Aken, G. A. (2001). Aeration of emulsions by whipping. Colloids and Surfaces A: Physicochemical and Engineering Aspects 190(3): 333–354. van Aken, G. A. (2004). Coalescence mechanisms in protein-stabilized emulsions. In Food Emulsions, S. Friberg, K. Larsson, and J. Sjoblom, eds., pp.299–326. New York: Marcel Dekker. van Aken, G. A. (2010). Relating food emulsion structure and composition to the way it is processed in the gastrointestinal tract and physiological responses: What are the opportunities? Food Biophysics 5(4): 258–283. van Aken, G. A., T. B. J. Blijdenstein, and N. E. Hotrum (2003). Colloidal destabilisation mechanisms in protein-stabilised emulsions. Current Opinion in Colloid & Interface Science 8(4–5): 371–379. van Aken, G. A. and T. van Vliet (2002). Flow-induced coalescence in protein-stabilized highly concentrated emulsions: Role of shear-resisting connections between the droplets. Langmuir 18(20): 7364–7370. van der Linden, E., L. Sagis, and P. Venema (2003). Rheo-optics and food systems. Current Opinion in Colloid & Interface Science 8(4–5): 349–358. 382 Food Emulsions: Principles, Practices, and Techniques van Vliet, T. and P. Walstra (1989). Weak particle networks. In Food Colloids, R. D. Bee, P. Richmond, and J. Mingins, eds., pp. 206–220. Cambridge, U.K.: The Royal Society of Chemistry. Vanapalli, S. A. and J. N. Coupland (2001). Emulsions under shear—The formation and properties of partially coalesced lipid structures. Food Hydrocolloids 15(4–6): 507–512. Vanboekel, M. and P. Walstra (1981). Stability of oil-in-water emulsions with crystals in the disperse phase. Colloids and Surfaces 3(2): 109–118. Walstra, P. (1983). Formation of emulsions. In Encyclopedia of Emulsion Technology, P. Becher, ed., vol. 4, pp. 57–128. New York: Marcel Dekker. Walstra, P. (2003). Physical Chemistry of Foods. New York: Marcel Decker. Waraho, T., D. J. McClements, and E. A. Decker (2011). Mechanisms of lipid oxidation in food dispersions. Trends in Food Science & Technology 22(1): 3–13. Weers, J. G. (1998). Ostwald ripening in emulsions. In Modern Aspects of Emulsion Science, B. P. Binks, ed., pp. 292–327. Cambridge, U.K.: The Royal Society of Chemistry. Williams, M. A. K., D. Fabri, C. D. Hubbard, L. Lundin, T. J. Foster, A. H. Clark, I. T. Norton, N. Loren, and A. M. Hermansson (2001). Kinetics of droplet growth in gelatin/maltodextrin mixtures following thermal quenching. Langmuir 17(11): 3412–3418. Woodall, A. A., S. W. M. Lee, R. J. Weesie, M. J. Jackson, and G. Britton (1997). Oxidation of carotenoids by free radicals: Relationship between structure and reactivity. Biochimica et Biophysica Acta—General Subjects 1336(1): 33–42. Wooster, T. J., M. Golding, and P. Sanguansri (2008). Impact of oil type on nanoemulsion formation and ostwald ripening stability. Langmuir 24(22): 12758–12765. Yang, X., H. Tian, C.-T. Ho, and Q. Huang (2011). Inhibition of citral degradation by oil-in-water nanoemulsions combined with antioxidants. Journal of Agricultural and Food Chemistry 59(11): 6113–6119. Yano, J. and K. Sato (1999). FT-IR studies on polymorphism of fats: Molecular structures and interactions. Food Research International 32(4): 249–259. Zeeb, B., L. Fischer, and J. Weiss (2014). Stabilization of food dispersions by enzymes. Food & Function 5(2): 198–213. Zeeb, B., M. Gibis, L. Fischer, and J. Weiss (2012). Influence of interfacial properties on Ostwald ripening in crosslinked multilayered oil-in-water emulsions. Journal of Colloid and Interface Science 387: 65–73. Zhai, J. L., L. Day, M. I. Aguilar, and T. J. Wooster (2013). Protein folding at emulsion oil/water interfaces. Current Opinion in Colloid & Interface Science 18(4): 257–271. Zhang, X. G. and R. H. Davis (1991). The rate of collisions due to Brownian or gravitational motion of small drops. Journal of Fluid Mechanics 230: 479–504. 8 Emulsion Rheology 8.1 Introduction Rheology is the science concerned with the deformation and flow of matter (Tadros 2010, Rao 2013, van Vliet 2013). Most rheological tests involve applying a specific force to a material and measuring the resulting flow and/or deformation of the material. The rheological properties of a material are then established by analyzing the relationship between the applied force and the resultant flow or deformation. Knowledge of the rheological properties of food emulsions is important for a variety of reasons (Chen and Stokes 2012, Selway and Stokes 2014). The efficiency of droplet disruption in a homogenizer depends on the viscosity of the individual components, as well as on the overall rheology of the product (Chapter 6). The shelf life of many food emulsions depends on the rheological characteristics of the component phases, for example, the creaming of oil droplets in oil-in-water emulsions is strongly dependent on the viscosity of the aqueous phase (Chapter 7). Information about the rheology of food emulsions is used by food engineers to design processing operations that depend on the way that the product flows, for example, flow through a pipe, stirring in a mixer, passage through a heat exchanger, packaging into containers (Singh and Heldman 2013). Many of the sensory attributes of food emulsions are directly related to their rheological properties, for example, creaminess, thickness, smoothness, spreadability, pourability, flowability, brittleness, and hardness (Chapter 9). A food manufacturer must therefore be able to design and consistently produce a product that has the desirable rheological properties expected by the consumer. Finally, rheological measurements are frequently used by food scientists as an analytical tool to provide fundamental insights about the structural organization and interactions of the components within emulsions, for example, measurements of viscosity versus shear rate can be used to provide information about the strength of the colloidal interactions between droplets (Quemada and Berli 2002, Tadros 2010). The purpose of this chapter is to present the general principles of rheology, to discuss the relationship between the rheological characteristics of food emulsions and their colloidal properties, and to provide an overview of analytical instruments used to characterize the rheological properties of food emulsions. Food emulsions are compositionally and structurally complex materials that can exhibit a wide range of different rheological behaviors, ranging from low-viscosity fluids (e.g., milk and soft drinks), to viscoelastic gels (e.g., yogurt and desserts), to fairly hard solids (e.g., refrigerated margarine and butter). Food scientists aim to develop theories that can be used to describe and predict the rheological behavior of this diverse group of products, as well as experimental techniques that can be used to quantify their rheological properties. Despite the diversity of rheological behavior exhibited by food emulsions, it is often possible to characterize their rheology in terms of a few simple models: the ideal solid, the ideal liquid, and the ideal plastic (Tadros 2010). More complex systems can then be described by combining two or more of these simple models. In the following sections, the concepts of ideal solid, ideal liquid, and ideal plastic are introduced, as well as some of the deviations from these models that are commonly observed in food emulsions. 383 384 Food Emulsions: Principles, Practices, and Techniques 8.2 Rheological Properties of Materials 8.2.1 Solids In our everyday lives, we come across solid materials that exhibit quite different rheological characteristics, for example, the resistance of the material to an applied force (soft versus hard) or the amount of deformation or force required to cause the material to fracture (brittle versus tough). Despite this range of different behavior, it is possible to characterize the rheological properties of many solid foods in terms of a few simple concepts (Tadros 2010). 8.2.1.1 Ideal Elastic Solids An ideal elastic solid is often referred to as a Hookean solid after Robert Hooke, the scientist who first described this type of behavior. Hooke observed experimentally that there is a linear relationship between the deformation of a solid material and the magnitude of the force applied to it, provided the deformation is not too large (Figure 8.1). He also observed that when the force was removed from the material, it returned back to its original length. In general, the force per unit area (or stress) acting upon the material is proportional to the relative deformation (or strain) of the material. Hooke’s law can therefore be summarized by the following statement: Stress (τ) = Constant (E) × Strain (γ) (8.1) A stress can be applied to a material in a number of ways, including simple shear, simple compression (or elongation), and bulk compression (or expansion) (Figure 8.2). Equation 8.1 is applicable to each of these situations, but the nature of the stresses, strains, and constants used depends on the type of deformation (Table 8.1). In addition, the strain can also be defined in a number of ways depending on the way that the length of the material is expressed (Walstra 2003). For example, in a compression experiment, the Cauchy or engineering strain is defined as the change in material length (Δl = l − l 0) divided by the original material length (l 0): γ E = Δl/l 0. Alternatively, the Hencky Fracture strain Stress (τ) Fracture stress Strain (γ) FIGURE 8.1 At small deformations, there is a linear relationship between the applied stress and the resultant strain for an ideal elastic (Hookean) solid. At higher deformations, the stress is no longer linearly related to strain, and the material will eventually fracture. 385 Emulsion Rheology F F Δl A A Δl Simple shear l l Simple compression P Bulk compression FIGURE 8.2 An elastic solid can be deformed in a number of different ways depending on the nature of the applied stress. TABLE 8.1 Rheological Parameters for Different Types of Deformation of Elastic Solids Deformation Simple shear Simple compression Bulk compression Stress F t= A Strain Dl g= = tan j l Elastic Modulus G= t F = g A tan f t= F A g= Dl l Y= t Fl = g ADl t= F =P A g= DV V K= t PV = g DV Notes: G, shear modulus; Y, Young’s modulus; K, bulk modulus; P, pressure; and the other symbols are defined in Figure 8.2. or natural strain is defined as the change in material length divided by the length at the time of measurement: γ H = |ln(l/l 0)|. The Hencky strain is usually more appropriate than the Cauchy strain for characterizing large material deformations. Similarly, the stress can be defined as being equal to the applied force divided by the original cross-sectional area of a material (engineering stress), or as the applied force divided by the cross-sectional area of the material at the time of measurement (natural stress). These values are appreciably different from each other when the cross-sectional area of a material changes during deformation, and it is advisable to use the natural stress rather than the engineering stress, particularly for large material deformations. 386 Food Emulsions: Principles, Practices, and Techniques The equations given in Table 8.1 assume that the material is homogeneous and isotropic, that is, its properties are the same in all directions. To characterize the rheological constants of an ideal elastic solid, it is therefore necessary to measure the change in its dimensions when a force of known magnitude is applied. The elastic behavior of a solid is related to the intermolecular forces that hold the structural units (molecules or particles) together (Walstra 2003). When a stress is applied to a material, the bonds between the structural units are compressed or expanded and therefore they store energy. When the stress is removed, the bonds give up this energy, and the material returns to its original shape. The elastic modulus of an ideal elastic solid is therefore related to the strength of the interactions between the structural units within it and the number of interactions per unit area of material. In reality, the elastic modulus of a solid may also depend on the internal structure of a solid, for example, if there are any cracks or dislocations present. 8.2.1.2 Nonideal Elastic Solids Hooke’s law is strictly applicable only to elastic materials at relatively low strains, and therefore many fundamental rheological studies of solid foods are carried out using very small material deformations (<1%). Nevertheless, the rheological behavior of foods at large deformations is often more relevant to their actual use, for example, spreadability, slicing, or mastication (Walstra 2003, Rao 2013, van Vliet 2013). For this reason, it is also important to characterize the rheological behavior of solids at large deformations. At strains just above the region where Hooke’s law is obeyed, the stress is no longer proportional to the strain, and therefore an apparent modulus is defined, which is equal to the stress/ strain at a particular value of the strain. It is therefore important to stipulate the strain (or stress) that an apparent modulus of a material is measured when reporting rheological data on nonideal solids. In this range of deformations, the material still returns to its original shape once the force is removed, even though it does not obey Hooke’s law. Above a certain deformation, however, a solid may not return back to its original shape after the applied stress is removed, because it either fractures or flows. A material that breaks at low strains is referred to as being brittle, whereas a material that flows is referred to as being plastic or viscoelastic depending on the nature of the flow (see later). The stress at which a material fractures is referred to as the fracture stress (τFr), whereas the strain at which it fractures is referred to as the fracture strain (γFr). A material usually fractures or flows when the forces holding the structural units (e.g., atoms, molecules, or particles) together in the material are exceeded. This often begins at regions where the bonds holding the material together are relatively weak, for example, a crack or dislocation. Knowledge of the fracture stress or fracture strain of a material is often a useful indication of its ability to be broken down during mastication, cut with a knife, or damaged during storage and transport. 8.2.2 Liquids Liquid food emulsions also exhibit a wide range of rheological properties (Rao 2013, van Vliet 2013). Some have low viscosities and flow easily (e.g., milk or soft drinks), while others are highly viscous (e.g., double cream). Even so, it is often possible to characterize their rheological properties using a few simple concepts. 8.2.2.1 Ideal Liquids An ideal liquid is often referred to as a Newtonian liquid, after Isaac Newton, the scientist who first described its behavior. When a shear stress is applied to an ideal liquid, it continues to flow as long as the stress is applied (Figure 8.3). Once the applied stress is removed, then the liquid will continue to flow until the kinetic energy stored within it is dissipated as heat due to friction. In this case, there is no elastic recovery of the material once the applied stress is removed, that is, it does not return to its original shape. 387 Stress (τ) Viscosity (η) Emulsion Rheology Stress (τ) Rate of strain (γ) FIGURE 8.3 Stress is proportional to the rate of strain for an ideal (Newtonian) liquid, and the viscosity is independent of the applied rate of strain (or shear stress). The viscosity of a liquid is a measure of its resistance to flow: the higher the viscosity, the greater the resistance. The concept of viscosity can be understood by considering a liquid that is contained between two parallel plates (Figure 8.4). The bottom plate is at rest while the top plate moves in the x-direction with a constant velocity, v. It is assumed that the liquid between the plates consists of a series of infinitesimally thin layers. The liquid layers in direct contact with the bottom and top plates are assumed to “stick” to them, so that they have velocities of 0 and v, respectively. The intervening liquid layers slide over each other with velocities that range between 0 and v, the actual value being given by dy(dv/dy), where dy is the distance from the bottom plate and dv/dy is the velocity gradient between the plates. The shear stress applied to the fluid is equal to the shear force divided by the area over which it acts. (τ = F/A). The rate of strain is given by the change in displacement of the layers per unit time: dγ/dt (or g ) = dv/dy. For an ideal liquid, the shear stress is proportional to the rate of strain (Figure 8.3): t = hg (8.2) Here, the constant of proportionality, η, is called the viscosity. Conceptually, the viscosity can be considered to arise from the friction acting between the imaginary “layers” of liquid as they slide past one x y v v=0 FIGURE 8.4 The viscosity of a liquid can be envisaged as arising from the friction generated by thin layers of liquid as they slide across each other: the greater the friction, the higher the energy dissipation, and the greater the viscosity. 388 Food Emulsions: Principles, Practices, and Techniques another. The lower the viscosity of a liquid, the less resistance between the liquid layers, and therefore the smaller the force required to cause the top plate to move with a given velocity, or the faster the top plate moves when a given force is applied. The ideal viscous fluid differs from the ideal elastic solid because the shear stress is proportional to the rate of strain (Figure 8.3), rather than the strain (Figure 8.1). It should be noted that the value of a fluid viscosity actually depends on the type of flow profile that it experiences, for example, simple shear flow (ηss or η) or elongational flow (ηel) (Walstra 2003). The elongational viscosity is always higher than the shear viscosity: ηel = Tr × η, where the Trouton ratio (Tr) depends on the precise nature of the elongational flow. Most rheometers used to characterize food emulsions utilize shear flow conditions and therefore measure the shear viscosity, but some do use fully or partly elongational flow conditions. The units of shear stress (τ) are Ν m–2 (or Pa) and those of shear rate ( g ) are s–1; thus, the viscosity (η) has units of N s m–2 (or Pa s) in the SI system. Viscosity can also be expressed in the older c.g.s. units of Poisse, where 1 Pa s = 10 Poisse. Thus, the viscosity of water at room temperature is around 1 mPa s, 0.001 Pa s, 0.01 Poise, or 1 centipoise, depending on the units used. Ideally, a Newtonian liquid should be incompressible (the volume does not change when a force is applied to it), isotropic (the properties are the same in all directions), and structureless (homogeneous). Although many liquid foods do not strictly meet these criteria, their rheological behavior can still be described excellently by Equation 8.2, for example, milk, sugar solutions, brine, and honey. Nevertheless, there are many other food fluids that exhibit nonideal liquid behavior, and so their properties cannot be described by Equation 8.2, for example, flocculated emulsions, concentrated emulsions, or biopolymer solutions (see Section 8.2.2.2). The type of flow depicted in Figure 8.4 occurs at low applied shear rates and is known as laminar flow, because the liquid travels in a well-defined laminar pattern. At higher shear rates, eddies form in the liquid, and the flow pattern is much more complex (Walstra 2003, van Vliet 2013). This type of flow is referred to as turbulent, and it is much more difficult to mathematically relate the shear stress to the shear rate under these conditions. For this reason, instruments that measure the viscosity of liquids are designed to avoid non-laminar flow conditions. 8.2.2.2 Nonideal Liquids Nonideal rheological behavior may manifest itself in a variety of ways in liquids, for example, the viscosity may depend on the magnitude of the applied shear and/or the length of time it is applied, or the material may exhibit some elastic as well as viscous properties (Rao 2013, van Vliet 2013). Plastic and viscoelastic materials, which have some elastic characteristics, are considered in later sections. Shear-rate-dependent nonideal liquids. In an ideal liquid, the viscosity is independent of the magnitude or duration of the applied shear stress, that is, the ratio of the shear stress to the shear rate does not depend on shear rate or time (Figure 8.5). In practice, many food emulsions have viscosities that do depend on the shear rate and the length of time that the system is sheared. In this section, we examine emulsions in which the viscosity depends on shear rate but is independent of the shearing time. In the following section, we examine emulsions in which the viscosity depends on both the shear rate and the time. The viscosity of an emulsion may either increase or decrease as the shear rate is increased, rather than staying constant as for a Newtonian liquid (Figure 8.5). In these systems, the viscosity at a particular shear rate is referred to as the apparent viscosity. The dependence of the apparent viscosity on shear rate means that it is crucial to stipulate the shear rate used to carry out the measurements when reporting data. The choice of shear rate to use when measuring the apparent viscosity of a nonideal liquid is a particularly important consideration when carrying out rheological measurements that are supposed to mimic some process that occurs in a food naturally, for example, flow through a pipe, stirring, mixing, pouring from a bottle, creaming of an individual emulsion droplet, or mastication of a food (Table 8.2). A laboratory test should use a shear rate that is as close as possible to that which the food experiences in practice. 389 Emulsion Rheology Dilatant Dilatant Viscosity (η) Stress (τ) Ideal Ideal Pseudoplastic (a) Rate of strain (γ) Pseudoplastic (b) Stress (τ) FIGURE 8.5 Comparison of the viscosity of ideal and nonideal liquids: (a) shear stress versus rate of shear strain; (b) apparent viscosity versus applied shear stress. TABLE 8.2 Typical Rates of Shear Strain Observed in Some Common Processes Relevant to Food Emulsions Process Pumping Mixing and stirring Chewing and swallowing Pouring Droplet creaming Shear Rate (s –1) 100 to 103 101 to 103 101 to 102 10–2 to 102 10–6 to 10–3 The two most common types of shear-rate-dependent nonideal liquids are as follows: 1. Pseudoplastic fluids: Pseudoplastic flow is the most common type of nonideal behavior exhibited by food emulsions. It manifests itself as a decrease in the apparent viscosity of a fluid as the shear rate is increased and is therefore often referred to as shear thinning (Figure 8.5). Pseudoplasticity may occur for numerous reasons in food emulsions, for example, the spatial distribution of the particles may be altered by the shear field, nonspherical particles may become aligned with the flow field, solvent molecules bound to the particles may be removed, or flocs may be deformed and disrupted (see Sections 8.4.1, 8.4.3, and 8.4.4). 2. Dilatant fluids: Dilatant behavior is much less common than pseudoplastic behavior. It manifests itself as an increase in the apparent viscosity as the shear rate is increased and is therefore often referred to as shear thickening (Figure 8.5). Dilantency is often observed in concentrated emulsions or suspensions when the particles are packed tightly together (Hunter 1989). At intermediate shear rates, the particles form two-dimensional sheets that slide over each other relatively easily, but at higher shear rates, these sheets are disrupted and so the viscosity increases. Shear thickening may also occur when the applied shear causes the droplets in an emulsion to flocculate because of an increased collision frequency (Section 7.5); however, this process usually leads to time-dependent behavior and so will be considered in the following section. This kind of shear thickening is also observed in starch suspensions, which have a low viscosity when sheared slowly but a very high viscosity when sheared rapidly. 390 Food Emulsions: Principles, Practices, and Techniques Liquids that exhibit pseudoplastic behavior often have a viscosity versus shear stress profile similar to that shown in Figure 8.6. The viscosity decreases from a constant value at low shear stresses (η0) to another constant value at high shear stresses (η∞). A number of mathematical equations have been developed to describe the rheological behavior of shear-rate-dependent nonideal liquids (Hunter 1989). The major practical difference is the range of shear stresses over which they are applicable. The “Meter” model can be used to describe the dependence of the apparent viscosity on shear stress across the whole shear stress range: h = h¥ + h0 - h¥ 1 + ( t /t i ) n (8.3) where τi is the shear stress where the viscosity is midway between the low- and high-shear rate limits (=½[η0 + η∞]) n is the power index The rheological properties of this type of system can therefore be characterized by four parameters: η0, η∞, τi, and n. Equations similar to Equation 8.3 have been developed to relate the shear dependence of the apparent viscosity to the strength of the attractive interactions acting between the droplets in flocculated emulsions (see later). The dependence of the apparent viscosity on shear rate can be described using a similar equation by replacing τ/τi with g /g i. If measurements are carried out only at shear rates that are sufficiently lower than the high-shear-rate plateau, then the rheology can be described by the Ellis model (Hunter 1994): h= h0 1 + ( t /t i ) (8.4) n Apparent viscosity (Pa s) 10 1 0.1 0.01 0.01 0.1 1 10 100 Shear stress (Pa) FIGURE 8.6 Typical apparent viscosity versus shear stress profile for a shear-thinning (pseudoplastic) material. The apparent viscosity decreases from a constant value (η0) at low shear rates to another constant value (η∞) at high shear rates (predicted using Meter model). 391 Emulsion Rheology If measurements are carried out only at intermediate shear rates (i.e., above the low-shear plateau and below the high-shear plateau), then the viscosity can often be described by a simple power-law model (Hunter 1994): h = A ( g ) B -1 (8.5) The constants A and B are usually referred to as the consistency index and the power index, respectively. For an ideal liquid B = 1, for an emulsion that exhibits shear thinning B < 1, and for an emulsion that exhibits shear thickening B > 1. Equation 8.5 is easy to use since it contains only two unknown parameters that can simply be obtained from a plot of log(η) versus log(g ). Nevertheless, these equations should be used only after it has been proved experimentally that the relationship between log(τ) and log(dγ/dt) is linear over the shear stresses or shear rates used. In addition, the power-law model is applicable only over a relatively narrow shear stress range. Time-dependent nonideal liquids: The apparent viscosity of the fluids described in the previous section depended on the shear rate (or shear stress), but not on the length of time that the shear stress was applied. There are many food emulsions whose apparent viscosity either increases or decreases with time during the application of shear. In some cases, this change is reversible, and the fluid will recover its original rheological characteristics if it is allowed to stand for a sufficiently long period. In other cases, the change brought about by shearing the sample is irreversible, and the sample will not recover its original characteristics. An appreciation of the time dependency of the flow properties of food emulsions is of great practical importance in the food industry. The duration of pumping or mixing operations, for instance, must be carefully controlled so that the food sample has an apparent viscosity that is suitable for the next processing operation. If a food is mixed or pumped for too long, it may become too thick or too runny and thus lose its desirable rheological properties. The dependence of a liquid’s rheology on time is usually associated with some kind of relaxation process (Hunter 1994). When an external force is applied to a system that is initially at equilibrium, the material takes a certain length of time to reach the new equilibrium condition, which is characterized by a relaxation time, τR. When the measurement time is of the same order as the relaxation time, it is possible to observe changes in the system properties with time. Thus, the rheological properties of an emulsion depend on the timescale of the experiment. Time-dependent nonideal fluids are classified into two categories: 1. Thixotropic behavior: A thixotropic fluid is one in which the apparent viscosity decreases with time when the fluid is subjected to a constant shear rate (Figure 8.7). Emulsions that exhibit this type of behavior often contain particles (e.g., droplets, crystals, or biopolymers) that associate with each other through weak attractive forces. Shearing of the material causes the aggregated particles to be progressively deformed and disrupted, which decreases the resistance to flow and therefore causes a reduction in the apparent viscosity over time. If the relaxation time associated with the disruption of the flocs is considerably shorter than the measurement time, then the apparent viscosity will tend to a constant low final value. This value may correspond to the point where the rate of structure disruption is equal to the rate of structure reformation, or where there is no more structure to be broken down. In pseudoplastic liquids, the breakdown of the aggregated particles occurs so rapidly that the system almost immediately attains its new equilibrium position, and so it appears as though the apparent viscosity is independent of time. 2. Rheopectic behavior: A rheopectic fluid is one in which the apparent viscosity increases with time when the fluid is subjected to a constant shear rate (Figure 8.7). One of the most common reasons for this type of behavior is that shearing increases both the frequency and the efficiency of collisions between droplets, which leads to enhanced aggregation (Section 7.4.1) and consequently to an increase in the apparent viscosity over time. 392 Food Emulsions: Principles, Practices, and Techniques 20 Gelling 18 Apparent viscosity 16 Rheopectic 14 12 Ideal 10 8 6 Thixotropic 4 2 0 0 10 20 Shearing time 30 40 FIGURE 8.7 Comparison of the apparent viscosity versus time profiles of ideal and time-dependent nonideal liquids. The viscosity may increase or decrease to a constant value with time. Alternatively, it may increase steeply due to network formation after a particular time. The rheological properties of some liquids are irreversible, that is, once the shear stress is removed, the system does not fully regain its original rheology. Liquids that exhibit this type of permanent change in their properties are called “rheo-destructive.” This type of behavior might occur when flocs are disrupted by an intense shear stress and are unable to reform when the shear stress is removed. On the other hand, the rheological properties of other liquids are partially or fully reversible, that is, once the shear stress is removed, the system eventually regains some or all of its original rheology. In this case, the recovery time is an important characteristic of the liquid. The rheological properties of time-dependent nonideal liquids can be characterized by measuring the change in their apparent viscosity with time during the application of a constant shear stress. Alternatively, the apparent viscosity of the liquid can be measured when the shear rate is increased from zero to a certain value and then decreased back to zero again (Figure 8.8). When there is a significant structural relaxation in a system, the upward curve is different from the downward curve, and one obtains a hysteresis loop. The area within the loop depends on the degree of relaxation that occurs and the rate at which the shear rate is altered. The slower the shear rate is altered, the more time the system has to reach its equilibrium value and therefore the smaller the area within the hysteresis loop. By carrying out measurements as a function of the rate at which the shear rate is increased, it is possible to obtain information about the relaxation time. Information about the time required for a liquid to recover its rheological properties after an applied shear stress has been removed can be obtained by measuring the rheology after the liquid has been left to stand for a certain period in the absence of shear. By varying the time between the applied shear and the rheological measurements, it is possible to obtain some information about how quickly the system recovers its original rheological properties. 8.2.3 Plastics A number of food emulsions exhibit rheological behavior known as plasticity, for example, mayonnaise, margarine, butter, and certain spreads. A plastic material has solid-like elastic properties below a certain applied stress, known as the yield stress, but liquid-like viscous properties when this stress is exceeded (Rao 2013, van Vliet 2013). 393 Emulsion Rheology Apparent viscosity Increase Hysteresis Decrease Shear stress FIGURE 8.8 A typical apparent viscosity versus shear stress hysteresis curve for a liquid whose viscosity depends on the length of time that it is sheared for. 8.2.3.1 Ideal Plastics The ideal plastic material is referred to as a “Bingham plastic” after the scientist who first characterized this type of rheological behavior. Two equations are needed to describe the rheological behavior of a Bingham plastic, one below the yield stress and one above it: t = Gg for t < tY t - tY = hg for t ³ tY (8.6) (8.7) where G is the shear modulus η is the viscosity τY is the yield stress The rheological properties of an ideal plastic are shown in Figure 8.9. Foods that exhibit plastic behavior usually consist of a network of aggregated molecules or particles dispersed in a liquid matrix. For example, margarine and butter consist of a network of tiny fat crystals dispersed in a liquid oil phase. Below a certain applied stress, there is a small deformation of the sample, but the weak bonds between the crystals are not disrupted. When the applied stress exceeds the yield stress, the weak bonds are broken, and the crystals slide past one another leading to the flow of the sample. Once the force is removed, the flow stops. A similar type of behavior can sometimes be observed in oil-in-water emulsions containing three-dimensional networks of flocculated droplets. 8.2.3.2 Nonideal Plastics Above the yield stress, the fluid flow may exhibit non-Newtonian behavior similar to that described earlier for liquids, for example, pseudoplastic, dilatant, thixotropic, or rheopectic behavior. The material may also exhibit nonideal elastic behavior below the yield stress, for example, the yield point may not be sharply defined; instead, the stress may increase steeply, but not instantaneously, as the shear rate is increased (Figure 8.9). This would occur if the material did not all begin to flow at a particular stress, but there was a gradual breakdown of the network structure over a range of applied stresses. 394 Food Emulsions: Principles, Practices, and Techniques Shear stress Ideal plastic τY Nonideal plastic Shear rate FIGURE 8.9 Rheological behavior of an ideal (Bingham) plastic and a nonideal plastic. 8.2.4 Viscoelastic Materials Many food emulsions are not pure liquids or pure solids but have rheological properties that are partly viscous and partly elastic (Walstra 2003, Rao 2013, van Vliet 2013). Plastic materials exhibit elastic behavior below a certain value of the applied stress and viscous behavior above this value. In contrast, viscoelastic materials exhibit both viscous and elastic behavior simultaneously. In an ideal elastic solid, all the mechanical energy applied to the material is stored in the deformed bonds and is returned to mechanical energy once the force is removed, that is, there is no loss of mechanical energy. On the other hand, in an ideal liquid, all of the mechanical energy applied to the material is dissipated due to friction, that is, the mechanical energy is converted to heat. In a viscoelastic material, part of the energy is stored as mechanical energy within the material, and part of the energy is dissipated as heat. For this reason, when a force is applied to a viscoelastic material, it does not instantaneously adopt its new dimensions nor does it instantaneously return back to its original nondeformed state when the force is removed (as an ideal elastic material would). In addition, the material may even remain permanently deformed once the force is removed. The rheological properties of a viscoelastic material are characterized by a complex elastic modulus, E*, which is comprised of an elastic and a viscous contribution: E* = E′ + iE″ (8.8) where E′ is known as the storage modulus E″ is known as the loss modulus Two types of experimental tests are commonly used to characterize the rheological properties of viscoelastic materials: one based on transient measurements and the other based on dynamic measurements. Both types of test can be carried out by the application of simple shear, simple compression, or bulk compression to the material being analyzed. Simple shear tests are the most commonly used to analyze food emulsions, and so only these will be considered here. Nevertheless, the same basic principles are also relevant to other kinds of applied stresses. 8.2.4.1 Transient Tests In a transient experiment, either a constant stress is applied to a material and the resulting strain is measured as a function of time (creep), or a material is deformed to a constant strain, and the stress required to keep the material at this strain is measured as a function of time (stress relaxation): 395 Emulsion Rheology Creep: In a creep experiment, a constant stress is applied to a material, and then the change in its dimensions with time is monitored, which results in a strain versus time curve (Sherman 1968). The data are usually expressed in terms of a parameter called the compliance, J, which is equal to the ratio of the strain to the applied stress (and is therefore the reciprocal of the modulus). The compliance is proportional to the strain, but it is a better parameter to use to characterize the rheological properties of the material because it takes into account the magnitude of the applied stress. The time dependence of the compliance of a material can also be measured when the stress is removed, which is referred to as a creep recovery experiment. A typical compliance versus time curve for a viscoelastic material that is put under stress is shown in Figure 8.10. This curve can be divided into three regions: 1. A region of instantaneous elastic deformation in which the bonds between the particles are stretched elastically. In this region, the material acts like an elastic solid with a compliance (J0) given by the ratio of the strain to the applied stress. 2. A region of retarded elastic compliance in which some bonds are breaking and some are reforming. In this region, the material has viscoelastic properties, and its compliance is given by JR = JM(1 − exp(−t/τM)), where JM and τM are the mean compliance and retardation time. 3. A region of Newtonian compliance, JN, when the bonds are disrupted and do not reform so that the material only flows: JN = t/ηN. The total creep compliance of the system is therefore given by J(t) = J0 + JR(t) + JN(t) = J0 + JM(1 − exp(−t/τM)) + t/ηN (8.9) This type of material is usually referred to as a viscoelastic liquid, because it continues to flow for as long as the stress is applied. Some materials exhibit a different type of behavior and are referred to as viscoelastic solids. When a constant stress is applied to a viscoelastic solid, the creep compliance increases up to a finite equilibrium value (JE) at long times, rather than increasing continuously. When the force is removed, the compliance returns to zero, unlike a viscoelastic liquid, which does not return to its initial shape once the force is removed. Stress relaxation: Instead of applying a constant stress and measuring the change in the strain with time, it is also possible to apply a constant strain and measure the stress required to keep the material at this strain as a function of time. This type of experiment is referred to as “stress relaxation.” The same type of information can be obtained from creep and stress relaxation experiments, and the method used largely depends on the type of rheological instrument available. Recovery Compliance Creep Time FIGURE 8.10 A representative compliance versus time curve for a viscoelastic material, such as ice cream. 396 Food Emulsions: Principles, Practices, and Techniques 8.2.4.2 Dynamic Tests The utilization of dynamic tests to characterize the rheological properties of viscoelastic materials has become routine in many laboratories due to the commercial availability of sophisticated dynamic rheometers (Rao 2013, van Vliet 2013). In a dynamic experiment, a sinusoidal stress is applied to a material, and the resulting sinusoidal strain is measured, or vice versa. In this section, we will consider only the case where a sinusoidal strain is applied to the sample and the resultant stress is measured. The applied strain is characterized by its maximum amplitude (γ0) and its angular frequency (ω). The resulting stress has the same frequency as the applied strain, but its phase is different because of relaxation mechanisms associated with the material, which cause viscous dissipation of some of the applied mechanical energy. Information about the viscoelastic properties of the material can therefore be obtained by measuring the maximum amplitude (τ0) and phase shift (δ) of the stress (Figure 8.11). The amplitude of the applied strain used in this type of test is usually chosen to be sufficiently small that the material is in the linear viscoelastic region (LVR). In this region, the stress is proportional to the strain, and the properties of the material are unaffected by sample deformation. The magnitude of a sinusoidal applied strain varies with time according to the following expression: γ = γ0 cos(ωt) (8.10) This applied strain causes the magnitude of the resulting stress to vary over time according to the following equation: τ = τ0 cos(ωt + δ) (8.11) The dynamic shear modulus of the material can be derived from these equations by dividing the stress by the strain: G = G′ + iG″ (8.12) where G¢ = t0 cos ( d ) g0 (8.13) Strain (γ) γ0 δ Stress (τ) τ0 Time FIGURE 8.11 The rheological properties of a viscoelastic material can be determined by measuring the relationship between an applied sinusoidal stress and the resultant sinusoidal strain (or vice versa). 397 Emulsion Rheology G¢¢ = t0 sin ( d ) g0 (8.14) where G′ is known as the “storage modulus” G″ is known as the “loss modulus” The in-phase component (G′) is determined by the elastic properties of the material, whereas the 90° out-of-phase component (G″) is determined by the viscous properties. This is because the stress is proportional to the strain (τ ∝ γ) for elastic materials, whereas it is proportional to the rate of strain (τ ∝ dγ/dt) for viscous materials. The dynamic rheological properties of a material can therefore be characterized by measuring the frequency dependence of the applied strain and the resulting stress, and then plotting a graph of G′ and G″ versus frequency. Alternatively, the data are often presented in terms of the magnitude of the complex modulus (G*) and the phase angle (δ): G * = G¢2 + G¢¢2 (8.15) æ G¢¢ ö d = tan -1 ç ÷ è G¢ ø (8.16) The phase angle of a material provides a useful insight into its viscoelastic properties: δ = 0° for a perfectly elastic solid, δ = 90° for a perfectly viscous fluid, and 0 < δ < 90° for a viscoelastic material. The more elastic a material, the smaller the phase angle, and the lower the amount of energy dissipated per cycle. Gels are often defined as having phase angles less than 45°, but this value actually depends on the measurement frequency, and so it is also important to specify the frequency when reporting gelation times or gelation temperatures determined using this definition. In addition, measurements of the rheological properties of viscoelastic materials as a function of frequency can provide valuable information about relaxation times associated with structural changes within the material. Sophisticated analytical instruments are available to measure the dynamic rheological properties of viscoelastic materials, and utilization of these techniques is providing valuable insights into the factors that influence the rheology of food emulsions. Typically, the dynamic shear modulus of a food material is measured as a function of time, temperature, or frequency. The data are usually represented as either G′ and G″ versus the relevant parameter or as G* and δ depending on the experimenter’s preference. 8.3 Measurement of Rheological Properties Food emulsions can exhibit a wide range of different types of rheological behavior, including liquid, solid, plastic, and viscoelastic. Consequently, a variety of instrumental methods have been developed to characterize their rheological properties (Rao 2007, Rao 2013, Selway and Stokes 2014). Instruments vary according to the type of deformation they apply to the sample (e.g., shear, compression, elongation, or some combination), the rheological properties that they can measure (e.g., shear modulus, viscosity, viscoelasticity, or friction coefficient), the nature of the samples that they can test (e.g., liquids, solids, or gels), their cost, their sensitivity, the range of accessible shear stresses or strains, the ability to scan temperature or not, their ease of operation, data processing, and their ability to make off-line or in-line measurements. In many industrial applications, it is necessary to have instruments that make measurements that are rapid, low cost, simple to carry out, and reproducible, rather than giving absolute fundamental data. Thus, simple empirical measurement techniques are often used in quality assurance laboratories, rather than the more sophisticated and expensive instruments used in research and development laboratories. 398 Food Emulsions: Principles, Practices, and Techniques The information obtained from these simple empirical instruments is often difficult to relate to the fundamental rheological constants of a material because the applied stresses and strains are not easily measured or defined. Rather than having a simple elongation, shear, or compression, different types of forces may be applied simultaneously. For example, when a blade cuts through a margarine or butter, both shear and compression forces are applied simultaneously, and the sample is deformed beyond the limit where Hooke’s law is applicable. To compare data from different laboratories, it is necessary to carefully follow standardized test procedures. These procedures may define experimental parameters such as the design of the device used, the magnitude of the applied force or deformation, the speed of the probe, the length of time the force is applied for, the measurement temperature, the sample dimensions, and preparation procedure. For food scientists involved in research and development, it is usually necessary to use instruments that provide information about the fundamental rheological constants of the material being tested, for example, η, G′, G″. These instruments are designed to apply well-defined stresses and strains to a material in a controlled manner so that stress–strain relationships can be measured and interpreted using available mathematical models. Rheological properties determined using these techniques can be compared with measurements reported by other researchers in the literature or made by colleagues working in other laboratories with different instruments. In addition, measured fundamental rheological properties can be compared with theoretical predictions made by mathematical models developed to relate the structure and composition of materials to their fundamental rheological properties. It is convenient to categorize rheological instruments according to the nature of the forces they apply to the sample being tested, for example, compression/elongation, shear, or combined methods. 8.3.1 Simple Compression and Elongation Measured strain This type of test is most frequently carried out on solid or semisolid foods that are capable of supporting their own weight, for example, gels, butter, margarine, and ice cream (Bourne 2002, Rao 2013). Measurements are often carried out using instruments referred to generally as “Universal Testing Machines” or “Texture Analyzers.” The solid sample to be analyzed is placed between a fixed plate and a moving probe (Figure 8.12). The probe can have many different designs depending on the type of information required, including a flat plate, a blade, a cylindrical spike, or even a set of teeth. Movable probe Sample Pressure sensor Applied stress FIGURE 8.12 Universal Testing Machine that can be used to measure the rheological properties of materials using compression or elongation tests. Emulsion Rheology 399 The probe can be moved vertically, either upward or downward, at a controlled speed. Either the probe or the plate contains a pressure sensor that measures the force exerted on the sample when it is deformed by the probe. The instrument also records the distance that the probe moves through the sample. The stress and strain experienced by a material can therefore be calculated from the knowledge of its dimensions, and the force and deformation recorded by the instrument. Often, it is necessary to measure the change in the dimensions during the compression (or elongation) test so as to calculate the true stress. Some of the common tests carried out using Universal Testing Machines are as follows: 1. Stress–strain curve: The stress on a sample is measured as a function of strain as it is compressed at a fixed rate (Figure 8.1). The resulting stress–strain curve is used to characterize the rheological properties of the material being tested. The slope of stress versus strain at relatively small deformations is often a straight line, with a gradient equal to the elastic modulus (Table 8.1). At intermediate deformations, the stress may no longer be proportional to the strain, and some flow may occur, so that when the stress is removed, the sample does not return to its original shape. At larger deformations, the sample may rupture, and the fracture stress, strain, and modulus can be determined. The operator must decide the distance and speed at which the probe will move through the sample. For viscoelastic materials, the shape of the upward and downward curves may be different and depends on the speed at which the probe moves. This type of test is used commonly to test solid and semisolid samples, such as gels, margarine, butter, spreads, and desserts. 2. Repeated deformation: The sample to be analyzed is placed between the plate and the probe, and then the probe is lowered and raised a number of times at a fixed speed so that the sample experiences a number of compression cycles. An ideal elastic solid would show the same stress–strain curve for each cycle. However, the properties of many materials are altered by compression (e.g., due to fracture or flow), and therefore successive compression cycles give different stress–strain curves. Analysis of the stress–strain relationship over a number of cycles is often used to calculate a variety of parameters that can sometimes be related to the sensory texture of foods, such as hardness, fracturability, cohesiveness, springiness. This type of test is often used to give an indication of the processes that occur when a food is chewed in the mouth, that is, the breakdown of food structure due to oral processing. 3. Transient experiments: A material is placed between the plate and the probe, then compressed to a known deformation, and then the relaxation of the stress with time is measured (stress relaxation). Alternatively, a constant stress is applied to the sample, and the variation of the strain is measured over time (creep). This type of experiment is particularly useful for characterizing the rheological properties of viscoelastic food emulsions (see Section 8.2.4). By using different fixtures, the same type of instrument can be used to carry out elongation experiments. A sample is clamped at both ends, and then the upper clamp is moved upward at a controlled speed, and the force required to elongate the sample is measured by the pressure sensor as a function of sample deformation. Again the elastic modulus and fracture properties of the material can be determined by analyzing the resulting stress–strain relationship. Universal Testing Machines can also be adapted to perform various other types of experiments, for example, bending, slicing, or forcing a test material through an orifice. More sophisticated instruments based on dynamic rheological measurements have been developed to characterize the rheological properties of solids, plastics, and viscoelastic materials. As well as carrying out the standard compression measurements mentioned earlier, they can also be used to carry out dynamic compression measurements on viscoelastic materials. The sample to be analyzed is placed between a plate and a probe, and an oscillatory compression strain (or stress) of known amplitude and frequency is applied to it. The amplitude and the phase of the resulting strain (or stress) are measured and converted into a storage and loss modulus using suitable equations (Section 8.2.4.2). The amplitude of the applied strain (or stress) must be small enough to be in the LVR of the material. These instruments are relatively expensive to purchase and are therefore only expected to be used by research laboratories in large food companies, government institutions, and universities. Nevertheless, they are extremely powerful tools 400 Food Emulsions: Principles, Practices, and Techniques for carrying out fundamental studies on food emulsions. The rheological properties of a sample can be measured as a function of storage time or temperature, and thus processes such as gelation, aggregation, crystallization, melting, and glass transitions can be monitored. The measurement frequency can also be varied, which provides valuable information about relaxation processes occurring within a sample. Some complications can arise when carrying out simple compression experiments (Walstra 2003). There may be friction between the compressing plates and the sample, which can lead to the generation of shear as well as compression forces. For this reason, it is often necessary to lubricate the sample with oil to reduce the effects of friction. In addition, the cross-sectional area of the sample may change during the course of the experiment, which would have to be taken into account when converting the measured forces into stresses. Finally, for viscoelastic materials, some stress relaxation may occur during the compression or expansion so that the results depend on the rate of sample deformation. An interesting adaptation of compression testing, called “squeezing flow viscometry,” has been developed for the rheological testing of liquids and semisolid foods (Campanella and Peleg 2002). This technique is based upon compressing a sample between two parallel plates and measuring the resulting force–height relationship. A variety of measurement protocols are possible for analyzing different kinds of samples. The squeezing flow viscometry technique has potential advantages over many of the conventional methods used to measure the rheological properties of liquids because it can minimize problems associated with slip at the sample–measurement cell boundary and reduce structural disruption caused by insertion of the sample into a narrow measurement cell. 8.3.2 Shear Measurements Instruments that utilize shear measurements are used to characterize the rheological properties of liquids, viscoelastic materials, plastics, and solids (Bourne 2002, Rao 2013, van Vliet 2013). The type of instrument and test method used in a particular situation depends on the physicochemical characteristics of the sample being analyzed as well as on the kind of information required. Some instruments can be used to characterize the rheological properties of both solids and liquids, whereas others can be used for only either solids or liquids. Certain types of viscometer are capable of measuring the viscosity of fluids over a wide range of shear rates and can therefore be used to analyze both ideal and nonideal liquids, whereas in others, the shear rate cannot be controlled, and so they are suitable only for analyzing ideal liquids. A number of instruments can be used to characterize the rheological behavior of viscoelastic materials using both transient and dynamic tests, whereas others can use only either one or the other type of test. To make accurate and reliable measurements, it is important to select the most appropriate instrument and test method and to be aware of any possible sources of experimental error. 8.3.2.1 Capillary Viscometers The simplest and most commonly used capillary viscometer is called the “Ostwald viscometer.” This device usually consists of a glass U-tube into which the sample to be analyzed is poured. The whole arrangement is usually placed in a temperature-controlled water bath to reach the measurement temperature (Figure 8.13). The viscosity of the liquid is measured by sucking it into one arm of the tube using a slight vacuum and then measuring the time taken for a fixed volume of the material to flow back through a capillary of fixed radius and length. The time taken to travel through the capillary is related to the viscosity by the following equation: t =C h r (8.17) where ρ is the density of the fluid t is the measured flow time C is a constant that depends on the precise size and dimensions of the U-tube and can be determined by calibration of the device 401 Emulsion Rheology Start mark (i) Pull liquid upwards Stop mark (ii) Measure time for liquid to flow back Narrow capillary FIGURE 8.13 Capillary viscometer used to measure the viscosity of liquids. This method is most suitable for ideal liquids. The higher the viscosity of the fluid, the longer it takes to flow through the capillary. The simplest method for determining the viscosity of a liquid is to measure its flow time and compare it with that of a liquid of known viscosity, such as distilled water: æt r ö hS = ç S S ÷ h0 è t 0 r0 ø (8.18) where the subscripts S and 0 refer to the sample being analyzed and the reference fluid, respectively. This type of viscometer is used principally to measure the viscosity of Newtonian liquids. It is normally unsuitable for analyzing non-Newtonian liquids because the sample does not experience a uniform and controllable shear rate. U-tubes with different diameters are required to analyze liquids with different viscosities: the larger the diameter, the higher the viscosity of the sample that can be analyzed. In more modern U-tube instruments, the fluid is made to flow through the tube by applying an external pressure to it, rather than relying on its hydrostatic pressure. This external pressure can be applied using a piston or compressed gas. 8.3.2.2 Mechanical Viscometers and Dynamic Rheometers Numerous mechanical rheometers have been designed to measure the shear properties of liquids, viscoelastic materials, plastics, and solids. These instruments are usually computer controlled and can often carry out sophisticated rheological tests as a function of time, temperature, shear rate, or oscillation frequency. Basically, the sample to be analyzed is placed in a temperature-controlled measurement cell (Figure 8.14), where it is subjected to a controlled shear stress (or strain). The resulting strain (or stress) is measured by the instrument, and so the rheological properties of the sample can be determined from the stress–strain relationship. The type of rheological test carried out depends on whether the sample is liquid, solid, or viscoelastic. The instruments can be divided into two different types: constant stress instruments that apply a constant torque to the sample and measure the resultant strain or rate of strain, and constant strain instruments that apply a constant strain or rate of strain and measure the torque 402 Food Emulsions: Principles, Practices, and Techniques (a) (b) (c) (d) FIGURE 8.14 Different types of measurement cell commonly used with dynamic shear rheometers and viscometers. (a) Concentric cylinder, (b) cone and plate, (c) parallel plate, and (d) vane. generated in the sample. For convenience, we will discuss only constant stress instruments in the following text although both types of instrument are commonly used in the food industry. In addition, with many of the modern instruments, it is possible to make a constant stress instrument operate like a constant strain instrument and vice versa. A number of different types of measurement cell can be used to contain the sample during an experiment (Bourne 2002, Rao 2013): 1. Concentric cylinder: The sample is placed in the narrow gap between two concentric cylinders (Figure 8.14). Normally, the inner cylinder (the bob) is driven at a constant torque (angular force) and the resultant strain (angular deflection) or rate of strain (speed at which the cylinder rotates) is measured, depending on whether one is analyzing a predominantly solid or liquid sample. For a solid, the angular deflection of the inner cylinder from its rest position is an indication of its elasticity: the larger the deflection, the smaller the shear modulus. For a liquid, the speed at which the inner cylinder rotates is governed by the viscosity of the fluid between the plates: the faster it spins at a given torque, the lower the viscosity of the liquid being analyzed. The torque can be varied in a controlled manner so that the (apparent) elastic modulus or viscosity can be measured as a function of shear stress. This instrument can be used for measuring the viscosity of Newtonian liquids, the apparent viscosity of non-Newtonian liquids, the viscoelasticity of semisolids, and the plasticity and elasticity of solids. In some instruments, the outer cylinder rotates, and the inner cylinder remains fixed, but the principles of the measurements are the same. 2. Parallel plate: In this type of measurement cell, the sample is placed between two parallel plates (Figure 8.14). The lower plate is stationary, while the upper one can rotate. A constant torque is applied to the upper plate, and the resultant strain or rate of strain is measured, depending on whether one is analyzing a predominantly solid or liquid sample. The main challenge with this type of experimental arrangement is that the shear strain varies across the sample: the shear strain in the middle of the sample being less than that at the edges. The parallel-plate arrangement is therefore suitable for analyzing only samples that have rheological properties that are independent of shear rate, and it is therefore usually unsuitable for analyzing nonideal liquids or solids. 3. Cone and plate: This is essentially the same design as the parallel-plate measurement cell, except that the upper plate is replaced by a cone (Figure 8.14). The cone has a slight angle that is designed to ensure that a more uniform shear stress acts across the sample. The cone-and-plate arrangement can therefore be used to analyze nonideal materials. 4. Vane: A vane consists of a multibladed bob that is placed in a sample and then rotated around its axis (Figure 8.14). This method is finding increasing utilization for characterizing semisolid Emulsion Rheology 403 food emulsions because it overcomes many of the problems associated with conventional measurement geometries, such as disruption of sample structure during insertion into the device and wall slip. 5. Serrated edges: The effects of wall slip may also be overcome by using measurement cells that have serrated edges (rather than being smooth). Often, the rheological properties of samples are measured either as a function of storage time at a fixed temperature or as the temperature is varied in a controlled manner. 8.3.2.3 Possible Sources of Experimental Error This section highlights a number of possible sources of experimental error that should be avoided or taken into account when carrying out rheology measurements on food emulsions. Other possible sources of error that are common to all types of rheology measurement have been discussed in detail elsewhere (Malone et al. 2003). Rheometer gap effects: The gap between the cylinders or plates should be at least 10 times greater than the diameter of the droplets, so that the emulsion appears as a homogeneous material within the device (Barnes 2000). On the other hand, the gap must be narrow enough to ensure a fairly uniform shear stress across the whole of the sample. Wall slip effects: A phenomenon known as wall slip may occur within a viscometer or rheometer, which can cause serious errors in rheological measurements if not properly taken into account (Franco et al. 1998, Sanchez et al. 2001, Meeker et al. 2004). It is normally assumed that the liquid in direct contact with the surfaces of the measurement cell moves with them at the same velocity. This assumption is usually valid for simple liquids because the small molecules are caught within the surface irregularities on the walls and are therefore dragged along with them. For an emulsion, this assumption may not hold because the droplets or flocs are greater in size than the surface irregularities. Under these circumstances, phase separation occurs at the measurement cell surfaces, and a thin layer of continuous phase acts as a lubricant so that slip occurs. The instrument response is then determined mainly by the properties of this thin layer of liquid rather than by the bulk of the material being tested. Wall slip effects can be minimized by roughening the surfaces of measurement cells or by using a range of different gap widths. Alternatively, different measurement geometries or rheological techniques can be used to overcome this effect, such as vanes, serrated edge devices, or squeezing flow techniques. Sample history: The rheological properties of many food emulsions depend strongly on their thermal and shear history, and so sample handling must be carefully controlled to obtain reproducible measurements (Franco et al. 1998). For example, the viscosity of many flocculated food emulsions decreases substantially upon shearing due to disruption of particle flocs, and the recovery of the original viscosity takes a certain length of time to achieve after the shear stress is removed. For these systems, it is extremely important to establish a consistent thermal and shear sample history prior to starting any rheological measurements. For example, it may be necessary to place an emulsion in a temperaturecontrolled measurement cell, then apply a fixed shear stress for a constant time, then allow it to sit for a fixed time, and then begin the rheological test. The objectives of this process are to break down and reform the structure of the emulsion in a reproducible and consistent manner so that samples can be compared under similar conditions. Gravitational separation: Many emulsions are susceptible to creaming or sedimentation during the course of an experiment, which causes the vertical distribution of droplets in the emulsion to become inhomogeneous (Macosko 1994). For example, in emulsions where the density of the dispersed phase is less than the density of the continuous phase, creaming leads to the formation of a droplet-rich layer at the top of the emulsion and a droplet-depleted layer at the bottom (Chapter 7). The separation of an emulsion into a creamed and a serum layer should be avoided because the rheological characteristics of a separated emulsion may be appreciably different from those of a homogeneous emulsion. The importance of this effect depends on the geometry of the measurement cell. In a concentric cylinder measurement cell, the formation of a viscoelastic or plastic creamed layer may dominate the rheology of the whole emulsion because the shear stress is applied to the sides of the emulsion. On the other hand, 404 Food Emulsions: Principles, Practices, and Techniques in a cone-and-plate or parallel-plate rheometer, the formation of a viscoelastic or plastic creamed layer may have a different effect because the shear stress is applied to the top and bottom of the emulsions. An approximate criterion that has been proposed to ensure that gravitational separation effects do not greatly affect measurements is that the droplets should move less than 10% of the emulsion height in the rheometer during the course of a measurement (Macosko 1994): t10%h = 0.45hh1 gr 2 ( r2 - r1 ) (8.19) where h is the height of the emulsion in the rheometer ρ2 is the density of the droplets ρ1 is the density of the continuous phase η1 is the viscosity of the continuous phase g is the acceleration due to gravity r is the droplet radius For a typical oil-in-water emulsion without thickening or gelling agent in the aqueous phase inside a typical concentric cylinder measurement cell (η1 = 1 m Pa s; Δρ = 80 kg m–3, h = 50 mm), t10%h ∼ 470/r 2 min, when r is expressed in micrometers. Thus, for 1 μm droplets, the sample will be stable for almost 8 h, but for 10 μm droplets, the sample will be stable only for about 5 min. Hydrodynamic instabilities: Hydrodynamic instabilities occur in fluids at sufficiently high flow rates, which generate secondary flows that interfere with the rheological measurements, for example, inertial effects (Larsson 1998). These effects can be observed in concentric cylinder–type rheometers at high shear rates, particularly for low-viscosity fluids, where it appears that the shear viscosity increases with increasing shear rate. For a concentric cylinder (Couette) measurement cell, the critical rotational speed at which Taylor vortices are first observed can be calculated from the following expression (Triantafillopoulos 1988): ΠC = 0.394 η/(Rδ3)1/2, where ΠC is the critical rotational speed (in rpm), η is the viscosity of the fluid (in Pa s), R is the radius of the inner cylinder (in meters), and δ is the gap width (in meters). Thus, one should be sure to operate a rheometer at a rotational speed below this critical level so as to avoid potential errors caused by hydrodynamic instabilities. Instrument sensitivity: There may also be considerable errors introduced into measurements due to limitations in instrument sensitivity. Many viscometers and dynamic shear rheometers cannot make accurate measurements of the rheological properties of low-viscosity liquids (such as water) when operated at relatively low shear rates due to a lack of sensitivity. One should therefore be careful to carefully calibrate an instrument with an ideal liquid of known rheological properties (such as water or a mineral oil) to be sure that the results obtained are reliable. 8.3.3 Advanced Rheological Methods There have been a number of advances in the development of analytical tools to characterize the rheological properties of emulsions and other food products (Melito and Daubert 2011, Chen and Stokes 2012, Selway and Stokes 2014). The application of these advanced rheological methods is leading to important new insights into the behavior of food emulsions under different conditions. 8.3.3.1 Rheometers Combined with Other Analytical Methods An important advance has been to combine rheometers with oth