http://chesterrep.openrepository.com
by
Paul Wassell
May 2013
Declaration of Originality
I hereby declare that work contained here is original and is entirely my own
work (unless indicated otherwise). It has not been previously submitted in
support of a Degree, qualification or other course.
I wish to acknowledge the Swedish Institute for Food and Biotechnology
(SIK). Göteborg, Sweden, who contributed to the research leading to Chapter
3, by supporting experiments through analysis and interpretation of data for the
Ultrasonic Velocity Profiling with Pressure Difference studies (Wassell et al.,
2010b; Young et al., 2008).
I wish to acknowledge the Laboratory of Food Biophysics, Hiroshima
University, Higashi-Hiroshima, Japan, who contributed to the research leading
to Chapter 6, by making the requested measurements and data analysis for the
Synchrotron Radiation Macrobeam and Microbeam X-ray Diffraction studies
of Interfacial Crystallisation of Fats in Water-in-Oil Emulsions (Wassell et al.,
2012).
Signed
Date
Acknowledgements
I would like to thank the following friends and colleagues who have inspired,
helped and supported me, before and during this study:
Allan Bech, Thue Christensen, Fernanda Davoli, Mark Farmer, Brad Forrest,
Torben Isak, Henrik Kragh, Cecilie Kristensen, Neils Krog, Christian Kjølby,
Jorgen Madsen, John Neddersen, Mogens Nielsen, Birgitte Pedersen, Lisbeth
Dahl Pedersen, John Podmore, Kirsten Aaby Sorensen, Jorn Borch Søe, Ralph
Timms, Stuart Warner, Niall Young and many others too numerous to mention.
I wish to acknowledge the contribution made by Danisco / DuPont and
particularly mention the technical support and the excellent library team in
Brabrand.
Thanks to my supervisors (Professors) at the University of Chester especially
to Graham Bonwick and Niall Young for being available in the good and
testing times.
I need to thank Dr. Johan Wiklund and all those involved in the joint
collaboration between Danisco / DuPont and the Swedish Institute for Food
and Biotechnology (SIK). Göteborg, Sweden.
Especially, I wish to thank very much Airi Okamura and Professors Kiyo Sato
and Satoru Ueno involved in the joint collaboration between Danisco / DuPont
and the Laboratory of Food Biophysics, Hiroshima University, Higashi-
Hiroshima, Japan, for their time and patience.
Finally, of course, Julia, for all your incredible support. Thank you.
List of Publications, Patents and Presentations
Papers:
Paper 1 Wassell, P. & Young, N.W.G (2007). Food applications of trans fatty
acid substitutes. International Journal of Food Science and
Technology, 42, 503–517.
Paper 4 Wassell, P., Wiklund, J,. Stading, M., Bonwick, G., Smith, C. J.,
Almiron-Roig, E. & Young, N. W. G. (2010) Ultrasound Doppler
Based In-Line Viscosity and Solid Fat Profile Measurement of Fat
Blends. International Journal of Food Science and Technology, 45,
877‐883.
Paper 5 Wassell, P., Okamura, A., Young, N.W.G., Bonwick, G., Smith, C.,
Sato, K., and Ueno, S., (2012) Synchrotron Radiation Macrobeam and
Microbeam X-ray Diffraction Studies of Interfacial Crystallization of
Fats in Water-in-Oil Emulsions. Langmuir 28 (13), 5539-5547.
Book Chapters
Young, N.W.G. and Wassell, P. (2008). Margarines and Spreads. In: G.L.
Hasenhuettl and R.W. Hartel (eds.), Food Emulsifiers and Their Applications,
2nd ed. Springer, New York.
Wassell, P. (2014). Bakery fats. In: Fats in Food Technology (edited by K.K.
Rajah), 2nd edn. Oxford, UK: Wiley-Blackwell (for publication).
Patents:
Patent 1 Wassell, P., Farmer, M., Warner, S.A., Bech, A.T., Young,
N.W.G., Bonwick, G., Smith, C., DuPont Nutrition Biosciences
APS, (2012). Food or Feed Including Moringa Oil. World
Intellectual Property Organisation No. WO2012168722.
Patent 2 Wassell, P., Farmer, M., Warner, S.A., Bech, A.T., Young, N.W.G.,
Bonwick, G., Smith, C., DuPont Nutrition Biosciences APS,
(2012). Triglyceride Fat Crystallisation. World Intellectual
Property Organisation No. WO2012168727.
Patent 3 Wassell, P., Farmer, M., Warner, S.A., Bech, A.T., Young,
N.W.G., Bonwick, G., Smith, C., DuPont Nutrition Biosciences
APS, (2012). Low Fat Spread. World Intellectual Property
Organisation No. WO2012168726.
Patent 4 Wassell, P., Farmer, M., Warner, S.A., Bech, A.T., Young,
N.W.G., Bonwick, G., Smith, C., Forrest, B.A., DuPont Nutrition
Biosciences APS, (2012). Dispersion of Triglycerides. World
Intellectual Property Organisation No. WO2012168723.
Patent 5 Wassell, P., Farmer, M., Warner, S.A., Bech, A.T., Young,
N.W.G., Bonwick, G., Smith, C., DuPont Nutrition Biosciences
APS, (2012). Spread. World Intellectual Property Organisation
No. WO2012168724.
Patent 6 Bech, A.T., Farmer, M., Forrest, B.A., Wassell, P., Young,
N.W.G., DuPont Nutrition Biosciences APS, (2013).
Composition. World Intellectual Property Organisation No.
WO2013050944.
Presentations (Oral)
Presentations (Poster)
Abstract
This study (Wassell & Young 2007; Wassell et al., 2010a) shows that behenic
(C22:0) fatty acid rich Monoacylglycerol (MAG), or its significant inclusion,
has a pronounced effect on crystallisation (Wassell et al., 2010b; 2012; Young
et al., 2008) and interfacial kinetics (3.0; 4.0). New interfacial measurements
demonstrate an unusual surface-interactive relationship of long chain MAG
compositions, with and without Polyglycerol Polyricinoleate (PGPR). A novel
MAG synthesised from Moringa oleifera Triacylglycerol (TAG) influenced
textural behaviour of water-in-oil (W/O) emulsions and anhydrous TAG
systems (4.0: 5.0; 6.0).
Emulsifier mixtures of PGPR and MAG rich in C18:1 / 18:2 and C16:0 / C18:0
do not decrease interfacial tension compared with PGPR alone. Only those
containing MAG with significant proportion of C22:0 impacted interfacial
behaviour. A mixture of C22:0 based MAG and PGPR results with decreasing
tension from ~20°C and is initially dominated by PGPR, then through
rearrangement, the surface is rapidly dominated by C22:0 fatty acids.
i
Wassell, P. A Multidisciplinary Approach to Structuring in Reduced Triacylglycerol Based Systems
The application of a Moringa oleifera based MAG in low TAG (35% - 41%),
W/O emulsions, results in high emulsion stability without a co-surfactant
(PGPR). The bi-functional behaviour of Moringa oleifera based MAG is
probably attributed to miscibility (Ueno et al., 1994) of its fatty acids, ranging
~30% of saturated fatty acids (SAFA), with ~70% of C18:1 (5.0). It is
concluded that the surface-interactive behaviour of Moringa oleifera based
MAG, is attributed to approximately 10% of its SAFA commencing from
C20:0.
When examined separately and compared, results showed that physical effect
of a Moringa oleifera based MAG was not dissimilar to PGPR, influencing the
crystallisation kinetics of the particular anhydrous TAG system. When either
was combined with a C22:0 rich MAG, enhanced gelation onset and strong
propensity to form dendrite structure occurred (5.0).
ii
Table of Contents
Title Page
Declaration of Originality
Acknowledgements
List of Publications, Patents and Presentations
Abstract i
Table of Contents iii
List of Figures xi
List of Tables xxv
Abbreviations xxxii
iii
1.5 A Pilot Study – Preliminary Investigations for Enhancing 31
Crystallisation of Anhydrous TAG Dispersions and W/O
Emulsions
1.5.1 Introduction and Background 31
1.5.1.1 Materials and Methods 32
1.5.1.2 Results - Crystallisation of RBD Palm Oil and Commercial 33
Hard and Soft Anhydrous Fat Blends
1.5.1.3 Interim Conclusions of Pilot Study on Crystallisation 39
1.5.2 Interface and Diversity: Effect of Fatty Acids on W/O 40
Emulsions
1.5.2.1 Introduction and Background 40
1.5.2.2 Reduced TAG W/O Emulsion 41
1.5.2.3 Materials and Methods 42
1.5.2.4 Results and Evaluation Test 44
1.5.2.5 Conclusion 46
1.5.3 Very Low W/O TAG (12%) Based Emulsion - With and 47
Without MAG/TAG Additive
1.5.3.1 Introduction and Background 47
1.5.3.2 Materials and Methods 47
1.5.3.3 Results 49
1.5.3.4 Conclusion 52
1.5.4 General Conclusion of Pilot Studies 52
1.6 A Critical Review of Current Issues: Alternative 53
Mechanisms for Aiding Structure in TAG Based Systems
1.6.1 Introduction and Background 53
1.6.2 New Applications and Novel Materials 53
1.6.3 Techniques for Analysis 55
1.6.4 Structure – Effects of Minor Components 57
1.6.5 Conclusion 59
1.7 Aims and Objectives 63
1.7.1 Background 63
1.7.2 Outstanding Problems 64
1.7.3 Empirical Quantification 65
1.7.4 Experimental Approach 66
1.7.5 Analytical Approach 69
1.7.5.1 UVP-PD 69
1.7.5.2 Tensiometry 69
1.7.5.3 Rheology 69
1.7.5.4 A Study Series of Application Tests: Observations on the 70
Behaviour of Novel MAG, Behenic rich MAG and PGPR in
Real Systems
1.7.5.5 Synchrotron Radiation X-ray Diffraction (SR-XRD) 70
Macrobeam and Microbeam small angle X-ray Diffraction
(SR-μ-SAXD) Analysis
iv
2.0 General Materials and Methods 71 - 118
2.1 Emulsifiers 71
2.2 Fatty Acid and Monoacylglycerol Compositions (%) 71
2.3 Additional Distillations of Natural Moringa oleifera 77
2.4 Synthetic Monoglycerides 78
2.5 High and Low Temperature Distillation of Moringa oleifera 80
TAG
2.6 Triacylglycerol (TAG) for Water-in-Oil (W/O) Emulsions 81
and Anhydrous Dispersions
2.7 Other and Minor Ingredients for W/O Emulsions 82
2.7.1 Hydrocolloid 82
2.7.2 Flavours 82
2.7.3 Antioxidant 82
2.7.4 Skimmed Milk Powder 82
2.7.5 Salt 82
2.7.6 Antimicrobial 82
2.7.7 β-carotene and Preparation 82
2.7.8 EDTA 83
2.7.9 Tap Water Specification (Data on Water Hardness, Aarhus, 83
Denmark)
2.8 Preparation of W/O Emulsions and Bulk Anhydrous – 83
TAG Systems
2.8.1 Equipment and Processing 84
2.8.1.1 Water Phase 84
2.8.1.2 Fat Phase 85
2.8.1.3 Lab Pilot Scale Process Conditions 85
2.9 Confocal laser scanning microscopy (CLSM) 87
2.9.1 Method 87
2.9.1.1 Reagents 88
2.9.1.2 Equipment Beam Path Settings 88
2.9.1.3 Calculations 88
2.10 Differential Scanning Calorimetry (DSC) 89
2.11 Gas Chromatography (GC) 89
2.11.1 GC Analysis 89
2.11.2 Monoglyceride Analysis Method: Fatty Acid Methyl Ester 90
(FAME) and Glycerols (Tri, Di, Mono)
2.11.3 Preparation of Methyl Esters for GC Determination 91
2.12 Polarized Light Microscopy (PLM) 92
2.12.1 Method 92
2.12.2 Induction Heat /Cool / PLM Micrographs 92
2.13 Rheology 93
2.13.1 Interfacial Rheological System (IRS) Measurements 93
2.13.2 Haake Controlled Stress Rotational Rheometer 96
2.13.3 Rheometrics Controlled Stress Rotational Rheometer 96
2.14 SFC Determination 97
v
2.15 Isothermal Rate of Crystallisation (RoC) Test 97
2.16 Surface Tension - Interfacial Tensiometry 97
2.16.1 Tensiometry Materials and Methods 99
2.16.1.1 Solvent 99
2.16.1.2 Preparation of Samples 100
2.16.1.3 Interfacial Tension Method 100
2.17 Texture Analysis (TA-XT2i) 101
2.18 Droplet Size Distribution (DSD) in Low-Fat Spread 102
2.18.1 Method 103
2.18.2 Analytical Principle 103
2.19 Ultrasonic Velocity Profiling with Pressure Difference 105
(UVP-PD) Materials and Methods
2.19.1 Pilot Plant (Gerstenberg-Schröder A/S) 105
2.19.2 UVP-PD System 106
2.19.3 UVP-PD Method and Experimental Parameters 108
2.19.4 Solid Fat Content (SFC) Measurement – p-NMR 110
2.19.5 Solid Fat Content (SFC) Measurement – Ultrasound 110
2.20 Synchrotron Radiation X-ray Diffraction (SR-XRD) using 112
Microbeam small angle X-ray Diffraction, (SR-μ-SAXD)
2.20.1 Method 112
2.20.2 Macrobeam X-ray Diffraction (XRD) Measurements 115
2.20.3 Microbeam Small-angle X-ray Diffraction Measurements 115
2.20.4 DSC Measurements 118
2.20.5 Polarised Optical Microscopic (POM) Observation 118
vi
4.3 A Rheological and Interfacial Examination of Single, 148
Mixed and Novel MAG Behaviour: Effect of Saturation
and Chain Length in Anhydrous Bulk and Water-Oil
Systems
4.3.1 Introduction 148
4.3.2 Interfacial Rheological System (IRS) Measurements 148
4.3.3 “True” Dynamic Complex 149
4.3.4 Results and Discussion 150
4.3.5 Emulsifier Similarity 156
4.3.6 Tension Reduction 157
4.3.7 Novel MAG 157
4.3.8 Conclusions and Recommendations 164
vii
5.5 Microscopy Examination of a Forced Cooled Model TAG 229
Based System Containing a Distilled Moringa MAG,
Behenic Based MAG and PGPR
5.5.1 Introduction 229
5.5.2 Materials and Methods 229
5.5.3 Results and Discussion 230
5.5.4 Conclusion 237
5.6 The Rheological Behaviour of Model TAG Systems 238
Containing a Behenic based MAG, PGPR and Moringa
MAG, During Forced Cooling Velocities
5.6.1 Introduction 238
5.6.2 Materials and Methods 239
5.6.3 Results and Discussion 240
5.6.3.1 Cooling Velocity - 1°C/min 240
5.6.3.2 Cooling Velocity - 10°C/min 245
5.6.3.3 Cooling Velocity - 30°C/min 247
5.6.4 Conclusion 253
5.7 Impact of Fatty Acid Profile: Distinctions Between 254
Moringa MAG and Behenic Based MAG in Low TAG W/O
Emulsions
5.7.1 Introduction 254
5.7.2 Materials and Methods 255
5.7.3 Results and Discussion 258
5.7.3.1 CLSM 262
5.7.3.2 Texture Analysis 265
5.7.4 Conclusion 266
5.8 Effect of PGPR Concentration in Low TAG W/O
Emulsions 268
5.8.1 Introduction 268
5.8.2 Materials and Methods 269
5.8.3 Results and Discussion 270
5.8.3.1 Droplet Size Distribution (DSD) 270
5.8.3.2 Texture Analysis 272
5.8.3.3 CLSM 273
5.8.3.4 Cardboard Test 277
5.8.4 Conclusion 279
5.9 The Performance of Varying Distillations of Moringa 280
MAG in Low TAG W/O (40%) Emulsions
5.9.1 Introduction 280
5.9.2 Materials and Methods 280
5.9.3 Results and Discussion 284
5.9.3.1 Droplet Size Distribution (DSD) 285
5.9.3.2 CLSM 286
5.9.3.3 Texture Analysis 290
5.9.4 Conclusion 291
viii
5.10 Blended MAG Compositions to Equal Moringa Based 292
MAG in Low TAG W/O Emulsions
5.10.1 Introduction 292
5.10.2 Materials and Methods 292
5.10.3 Results and Discussion 296
5.10.3.1 Spread Test 300
5.10.3.2 CLSM 304
5.10.3.3 Texture Analysis 307
5.10.4 Conclusion 308
5.11 Effect of High and Low Temperature Distillation on the 309
Functionality of MAG based on Moringa oleifera TAG in
Low TAG W/O Emulsions
5.11.1 Introduction 309
5.11.2 Materials and Methods 310
5.11.3 Results and Discussion 314
5.11.3.1 Droplet Size Distribution (DSD) 314
5.11.3.2 Sensory 315
5.11.3.3 Effect of Distillation at 0.3% Moringa MAG Concentration 319
5.11.3.4 Effect of Distillation at 0.6% Moringa MAG Concentration 319
5.11.4 Conclusion 320
5.12 Inventions: a Moringa MAG and Modified Crystalliser 321
Composition
5.12.1 Summary 321
5.12.2 Moringa oleifera based Monoglycerides 321
5.12.3 Interim Conclusion 323
5.12.4 Modified Crystalliser Composition 324
5.12.5 Interim Conclusion 325
5.12.6 Significance for Pickering Stabilisation 325
5.12.7 Metastable Region 327
5.12.8 Dendrite Structure 328
ix
7.0 General Discussion, Final Conclusions and 340 - 351
Recommendations
7.1 Multidisciplinary Approach to Structuring in Reduced 340
Triacylglycerol (TAG) Based Systems
7.1.1 Implications for Pickering Stabilisation 344
7.1.2 Interfacial Stabilisation in a Metastable Region 344
7.1.3 Dendrite Behaviour 345
7.2 Key Findings 346
7.3 Novelty and Future Context 347
7.4 Recommended Research 349
7.4.1 Additional Measurements 349
7.4.2 High Internal Aqueous Phase Emulsions 349
7.4.3 Interactions at the Interfacial Region 350
7.5 Concluding Remarks 350
Appendices
x
List of Figures
Figure 1.3.2 Fatty acid units showing the short spacing between the 17
individual fat units (source: Young & Wassell, 2008)
xi
Figure 1.5.5 Relative increase in SFC% of RBD Palm Oil, Hard and 38
Soft blends
Figure 1.5.6 The effect of temperature and fatty acid diversity at the 42
interface: a preparation of water-soybean oil (1:1) with or without 4%
emulsifier (E471) in the oil phase is observed at three temperatures
5°C, 20°C & 40°C (Image supplied courtesy of Danisco A/S., 2003)
Figure 1.5.8 Textural firmness at 5°C of W/O 12% very low fat 49
spread emulsion
Figure 2.3 (a) Anton Paar Physica MCR Interfacial Rheology System 94
(IRS) and (b / c) Bicone measuring system
xii
Figure 2.7 Schematic image of pilot scale scraped surface heat 107
exchanger unit and pin rotor units with flow-cell and by-pass loop
measuring apparatus with differential pressure gauge and ultrasound
transducers.
Figure 2.8 Data analyses of SR-μ-SAXD patterns. (a) 2D SR-μ- 116
SAXD pattern. (b) Lamellar direction in a fat crystal. (c) χ extension
pattern
Figure 2.9 Schematic showing cold stage temperature control for 117
emulsion preparation (Ueno et al. 2008)
Figure 3.2 Schematic image of pilot scale scraped surface heat 122
exchanger unit and pin rotor units with flow-cell and by-pass loop
measuring apparatus with differential pressure gauge and ultrasound
transducers
Figure 3.4 Spectral plot together with an arithmetic average of some 126
30 measured profiles (green) together with the resulting power-law fit
(red) for the control system, i.e. 25% Akomic / 75% rapeseed oil at a
flow rate of 70 kg/h
xiii
Figure 3.6 Comparison of the Akomic / Rapeseed (RSO) TAG 128
blends; without additive (tetrahedrons) and with 1% CRY110 (circles)
Figure 3.7 Measured arithmetic average over 28 velocity profiles and 130
the resulting power-law fit for 25% Akomic / 75% rapeseed TAG
without additive at a flow rate of 70 kg h-1 ((The profiles were
measured both opposite to- (Transducer 1, TDX1) and in the direction
of the flow (Transducer 2, TDX2))
Figure 3.8 Measured arithmetic average over 39 velocity profiles and 131
the resulting power-law fit for 25% Akomic / 74% rapeseed TAG with
1% CRY110 at a flow rate of 70 kg h-1 ((The profiles were measured
both opposite to- (Transducer 1, TDX1) and in the direction of the
flow (Transducer 2, TDX2))
Figure 3.9 Solid fat content (SFC) expressed as percentage values 133
versus temperature for 30% palm stearin / 70% rapeseed TAG
measured by standard p-NMR technique (triangles) and in-line
dynamic conditions from UVP-PD (circles)
xiv
Figure 4.5 Interfacial tension behaviour of emulsifiers used in 143
preliminary application tests for 12% WO emulsion (1.5)
Figure 4.7 Interfacial Rheology System (IRS) temperature sweep and 150
resulting complex modulus Gi’ and Gi’’ of CRY110 (rich in C22:0)
alone compared to PGPR and other emulsifier mixtures
Figure 4.8 Bulk temperature sweep and resulting G’ and G’’ of 151
CRY110 (rich in C22:0) alone compared to PGPR and other
emulsifier mixtures.
Figure 4.11 The influence of a MAG rich in C22:0, where CRY110 154
affects both G’’ and G’
Figure 4.15 Bulk behaviour of Moringa MAG v CRY110 with PGPR 160
xv
Figure 4.17 Bulk behaviour of Moringa and CRY110 compared to 161
PGPR and CRY110
Figure 4.18 The complex IRS for Moringa MAG / CRY110 or PGPR 162
Figure 4.19 Complex IRS: CRY110 & Moringa v CRY110 & PGPR 163
Figure 4.20 The effect on G’’ after interfacial temperature sweep for 163
Moringa MAG v PGPR
Figure 5.2.1 Effect of Moringa MAG and PGPR in Rapeseed TAG 172
Figure 5.2.2 Effect of Moringa MAG and PGPR in Peanut TAG 173
Figure 5.4.1 Water droplet size distribution (DSD) of the 60% TAG 195
blends
Figure 5.4.2 CLSM image of 60% TAG spread with DIMODAN® 196
RT at 0.3% concentration. (Top images 375 x 375 µm. Bottom images
188 x 188µm)
xvi
Figure 5.4.3 CLSM image of 60% TAG spread with DIMODAN® 197
RT at 0.6% concentration (Top images 375 x 375 µm. Bottom images
188 x 188µm)
Figure 5.4.4 CLSM image of 60% TAG spread with Moringa at 0.3% 198
concentration. (Top images 375 x 375 µm. Bottom images 188 x
188µm)
Figure 5.4.5 CLSM image of 60% TAG spread with Moringa at 0.6% 199
concentration. (Top images 375 x 375 µm. Bottom images 188 x
188µm)
Figure 5.4.6 CLSM image of 60% TAG spread with Lesquerella at 200
0.3% concentration. (Top images 375 x 375 µm. Bottom images 188 x
188µm)
Figure 5.4.7 CLSM image of 60% TAG spread with Lesquerella at 201
0.6% concentration. (Top images 375 x 375 µm. Bottom images 188 x
188µm)
Figure 5.4.8 Water droplet size distribution (DSD) of W/O 40% TAG 204
blends
Figure 5.4.9 CLSM image of 40% TAG spread with DIMODAN® UJ 206
at 0.3% concentration. (Top images 375 x 375 µm. Bottom images
188 x 188µm)
Figure 5.4.10 CLSM image of 40% TAG spread with Moringa at 207
0.3% concentration. (Top images 375 x 375 µm. Bottom images 188 x
188µm)
Figure 5.4.11 CLSM image of 40% TAG spread with Lesquerella at 208
0.3% concentration. (Top images 375 x 375 µm. Bottom images 188 x
188µm)
xvii
Figure 5.4.12 CLSM image of 40% TAG spread with DIMODAN® 209
UJ at 0.6% concentration. (Top images 375 x 375 µm. Bottom images
188 x 188µm)
Figure 5.4.13 CLSM image of 40% TAG spread with Moringa at 210
0.6% concentration. (Top images 375 x 375 µm. Bottom images 188 x
188µm)
Figure 5.4.14 CLSM image of 40% TAG spread with Lequerella at 211
0.6% concentration. (Top images 375 x 375 µm. Bottom images 188 x
188µm)
Figure 5.4.15 CLSM image of 40% TAG spread with DIMODAN® 212
UJ 0.3% / GRINDSTED® PGPR 90 at 0.2% concentration. (Top
images 375 x 375 µm. Bottom images 188 x 188µm)
Figure 5.4.16 CLSM image of 40% TAG spread with Moringa 0.3% / 213
GRINDSTED® PGPR 90 at 0.2% concentration. (Top images 375 x
375 µm. Bottom images 188 x 188µm)
Figure 5.4.17 CLSM image of 40% TAG spread with Lesquerella 214
0.3% / GRINDSTED® PGPR 90 at 0.2% concentration. (Top images
375 x 375 µm. Bottom images 188 x 188µm)
Figure 5.4.18 The hardness of the high TAG (60%) samples – (1-6), 215
and the low TAG (40%) samples – (11-19)
Figure 5.4.19 The stickiness of the high TAG (60%) samples- (1-6), 215
and the low TAG (40%) TAG samples – (11-19)
Figure 5.4.20 Water droplet size distributions for the 40% low TAG 219
spreads of the validation tests
Figure 5.4.21 CLSM image of 40% TAG spread with DIMODAN® 220
UJ - 0.3% concentration. (Top images 375 x 375 µm. Bottom images
188 x 188µm)
xviii
Figure 5.4.22 CLSM image of 40% TAG spread with Moringa - 0.3% 221
concentration. (Top images 375 x 375 µm. Bottom images 188 x
188µm)
Figure 5.4.23 CLSM image of 40% TAG spread with DIMODAN® 222
UJ - 0.6% concentration. (Top images 375 x 375 µm. Bottom images
188 x 188µm)
Figure 5.4.24 CLSM image of 40% TAG spread with Moringa - 0.6% 223
concentration. (Top images 375 x 375 µm. Bottom images 188 x
188µm)
Figure 5.4.25 CLSM image of 40% TAG spread with DIMODAN® 224
UJ - 0.3% / GRINDSTED® PGPR 90 – 0.2% concentration. (Top
images 375 x 375 µm. Bottom images 188 x 188µm)
Figure 5.4.26 CLSM image of 40% TAG spread with Moringa - 0.3% 225
/ GRINDSTED® PGPR 90 – 0.2% concentration. (Top images 375 x
375 µm. Bottom images 188 x 188µm)
Figure 5.4.27 Textural hardness for the 40% low TAG spread 226
validation tests, where the order from left to right is as follows;
DIMODAN® UJ, 0.3%; Moringa 0.3%; DIMODAN® UJ, 0.6%,
Moringa, 0.6%; DIMODAN® UJ, 0.3% and GRINDSTED® PGPR
90, 0.2%; Moringa 0.3% and GRINDSTED® PGPR 90, 0.2%
xix
Figure 5.6.1 Schematic diagram of the energy barrier to crystallisation 239
together with the relative amount of structure formed for a given
cooling rate
Figure 5.6.3 c / d Viscosity cooling curves at 1°C per minute of base 244
fat blend (70% palm olein / 30% palm stearin) [green]; 1% CRY110
[red]; 1% CRY110 / 0.5% PGPR 90 [light blue]; 0.5% PGPR 90
[pink]; 1% CRY110 / 1% Moringa MAG [dark blue]; and 1%
Moringa MAG [yellow]
Figure 5.6.5 Viscosity cooling curves at 30°C/min of base TAG blend 250
(70% palm stearin/ 30% palm olein) [Blue]; 1% CRY110 [green]; 1%
CRY110 / 0.5% PGPR 90 [red]; 0.5% PGPR 90 [light blue]; 1%
CRY110 / 1% Moringa MAG [pink]; and 1% Moringa MAG [yellow].
Figure 5.6.6 Viscosity cooling curves at 30°C per minute of base fat 251
blend (70% palm olein / 30% palm stearin) [green]; 1% CRY110
[red]; 1% CRY110 / 0.5% PGPR 90 [light blue]; 0.5% PGPR 90
[yellow]; 1% CRY110 / 1% Moringa MAG [pink]; and 1% Moringa
MAG [dark blue]
xx
Figure 5.7.1, (1a) Spread test for samples 11 and 12; (1b) spread test 261
for samples 13 and 14; (1c) spread test for samples 15 and 16; (1d)
samples in storage tub for samples 21, 22, 23, 24, 25 and 26; (1e)
spread test for sample 26
Figure 5.7.2 (2a – 2f) CLSM images of low TAG (40%) spread 263
samples 11 to 16, with stabilised water phase
Figure 5.7.3 (3a – 3f) CLSM images of low TAG (35%) spread 264
samples 21 – 26 with empty water phase
Figure 5.7.4 Hardness results: texture analysis for samples 11 to 16, 265
40% TAG, full water phase
Figure 5.7.5 Hardness result: texture analysis for sample 26, 35% 265
TAG, empty water phase.
Figure 5.8.1 DSD normal distribution shows data from Table 5.8.3 272
Figure 5.8.2 Texture analysis on 61 – 64; average of 10 measurements 272
shows decreasing hardness with increasing PGPR concentration: 0.15,
0.3, 0.6 & 1.2%
Figure 5.8.3 CLSM image of 40% TAG emulsion (Sample 61, dose 273
0.15% PGPR)
Figure 5.8.4 CLSM image of 40% TAG emulsion (Sample 62, dose 274
0.3% PGPR)
Figure 5.8.5 CLSM image of 40% TAG emulsion (Sample 63, dose 275
0.6% PGPR)
Figure 5.8.6 CLSM image of 40% TAG emulsion (Sample 64, dose 276
1.2% PGPR)
Figure 5.8.7 Photographic evidence of spread tests: PGPR at 0.15% & 277
0.3% (top) and 0.6% & 1.2% (bottom)
Figure 5.9.1 (1a) Spread tests: samples 41 & 42 (Moringa MAG 102); 287
(1b) 43 & 44 (Moringa MAG 102); (1c) 45 & 46 (Moringa MAG 105)
and (1d) 47 & 48 (Moringa MAG 105)
xxi
Figure 5.9.2 CLSM images, Moringa MAG 102: sample 41 (0.15%); 288
sample 42 (0.3%); sample 43 (0.6%), and sample 44 (1.2%) (scaled to
375 x 375µm)
Figure 5.9.3 CLSM images, Moringa MAG 105: sample 45 (0.15%); 289
sample 46 (0.3%); sample 47 (0.6%), and sample 48 (1.2%) (scaled to
375 x 375µm)
Figure 5.9.4 Texture analysis: hardness profile for Moringa MAG 102 290
emulsions (41-44) and Moringa MAG 105 emulsions (45-48), at 5°C
Figure 5.10.3 CLSM images of synthetic MAG (SM90) mono content 304
~96%, samples 31, 32, 33, & 34, of concentration 0.15, 0.3, 0.6 &
1.2% respectively
Figure 5.10.5 CLSM images of synthetic MAG (SM80) mono content 306
~82%, samples 49, 51, 52, & 53, of concentration 0.15, 0.3, 0.6 &
1.2% respectively
Figure 5.10.6 Hardness results from texture analyser for synthetic 307
SM90 (31 to 34) and synthetic SM60 (35 to 38)
Figure 5.10.7 Hardness results from texture analyser for synthetic 308
SM80 (49, 51, 52 & 53)
xxii
Figure 5.11.1 40% emulsion samples: (1a) no.71, 0.3% dose, HIGH 317
T°C distillation; (1b) no.72, 0.3% dose, low T°C distillation; (1c)
no.73, 0.6% dose, HIGH T°C distillation; (1d) no.74, 0.6% dose, low
T°C distillation
Figure 5.11.2 35% emulsion samples: (1a) no.75, 0.3% dose, HIGH 318
T°C distillation; (1b) no.76, 0.3% dose, low T°C distillation; (1c)
no.77, 0.6% dose, HIGH T°C distillation; (1d) no.78, 0.6% dose low
T°C distillation
Figure 5.11.3 Texture analysis: hardness for 40% TAG emulsions 318
(samples 71-74), and 35% TAG emulsions (samples 75-78)
Figure 6.2 DSC cooling thermopeaks taken for (a) PGPR emulsion 332
and (b) PGPR + MB emulsion
xxiii
7.0 General Discussion, Final Conclusions and Recommendations
xxiv
List of Tables
Table 1.2.4 Typical emulsifier selection for W/O spreads with and 14
without proteins (Information received from Danisco / DuPont Oils &
Fats, 2009)
Table 1.5.1 Fatty acid profiles of two palm based fat blends. 35
Table 1.5.4 Formula of 50% reduced fat W/O emulsions and details of 44
the major fatty acids (SMP = skimmed milk powder)
xxv
Table 1.5.6 12% very low fat spread formula, assembled on 48
Gerstenberg scraped surface pilot plant. (interesterified blend = PK4
INES, a palm stearin / lauric. Cargill GmbH)
Table 1.6.1 Approved novel foods related to fats and oils (source: 54
Wang et al., 2012)
Table 2.10 RBD Ricebran oil (Thai Edible Oil Co., LTD (Bangkok, 76
Thailand)
xxvi
Table 2.11 Moringa oleifera (Earth Oil Plantations Ltd. Lichfield, 76
Staffs., UK)
Table 2.20 Droplet size distribution results presented as volume and 104
number size distribution
Table 2.21 Droplet size distribution calculated from log-scale values 104
of standardised normal distribution
xxvii
Table 3.2 Exothermic crystallisation behaviour recorded after pin- 132
worker, prior to UVP-PD measurement
Table 5.2.1 Fatty acid composition of Moringa oleifera TAG and 169
MAG
Table 5.3.1 Natural Moringa MAG with the breakdown of mono, di- 177
and tri- glycerides.
Table 5.3.4 Fatty acid composition of the natural Moringa MAG 179
Table 5.4.1 Recipe for 60% TAG spreads trials 1-6 189
xxviii
Table 5.4.2 Pilot plant processing conditions for the 60% TAG spread 190
samples
Table 5.4.3 Recipe for 40% TAG spreads trials 11-19 191
Table 5.4.4 Pilot plant processing conditions for the 40% TAG spread 192
samples
Table 5.4.5 Recipe for 40% TAG spreads (validation test) 193
Table 5.4.6 Plant process conditions for the 40% TAG spreads 193
(validation test)
Table 5.4.8 Water droplet size distribution for 40% TAG spread 204
samples
Table 5.4.9 Validated water droplet size distribution (DSD) results 218
(plain text) of 40% low TAG spreads. Figures in bold are the results
from the original samples (Table 5.4.8)
Table 5.7.3 Pilot plant processing conditions for the formula’s in 256
Tables 5.7.1 and 5.7.2
Table 5.7.4 Fatty acid profiles for MM 191 and CRY110 (Report 5.3 - 257
tables 5.3.4 & 5.3.6)
Table 5.7.5 Water droplet size distribution (DSD) for 40% TAG 258
spreads (samples 21-26), and 35% TAG spreads (sample 11-16)
Table 5.8.1 Formulations for low TAG W/O (40%) emulsions 269
Table 5.8.2 Processing parameters used for formula in Table 5.8.1 270
xxix
Table 5.8.3 Low TAG W/O (40%) emulsions samples 61 – 64 to test 271
PGPR water droplet size distribution at concentrations: 0.15, 0.3, 0.6
& 1.2%
Table 5.8.4 Results of average DSD volumes for low TAG W/O 271
(35%) emulsions (no protein / stabiliser) 0.6% Moringa 191; 0.4%
PGPR 90 and 0.3% PGPR + 0.15% CRY110))
Table 5.9.2 Moringa (Moringa oleifera) based MAG: mono, di- and 281
tri- glycerides (#2559/104 = 105 with antioxidant)
Table 5.9.3 Formulations: low TAG W/O (40%) emulsions with 282
Moringa oleifera based MAG
Table 5.9.4 Processing conditions for the low TAG W/O emulsions 283
Table 5.9.5 DSD data: samples 41-44 (concentration 0.15, 0.3, 0.6 284
and 1.2%) correspond to Moringa MAG 102, monoglyceride content
~53%; samples 45-48 (concentration 0.15, 0.3, 0.6 and 1.2%)
correspond to Moringa MAG 105, monoglyceride content ~83%
Table 5.9.6 DSD: low TAG W/O (40%) spreads containing Moringa 284
MAG 191 (monoglyceride content of ~91%)
Table 5.10.4 Formulations for 40% low TAG emulsions with 294
synthetic MAG (SM90 SM60)
Table 5.10.5 40% low TAG emulsions with synthetic MAG (SM80) 295
xxx
Table 5.10.6 Pilot plant scraped surface processing conditions for all 295
samples
Table 5.10.7 DSD data for SM 90 (samples 31-34), SM60 (samples 296
35-38), and SM80 (samples 49-53)
Table 5.10.8 DSD data for 40% TAG spread samples containing 298
natural Moringa MAG 191 monoglyceride ~91% (from Table 5.4.8 in:
5.4; and Table 5.7.5 in: 5.7)
Table 5.10.9 DSD data from 5.9 shows: samples 41-44 (conc. 0.15, 298
0.3, 0.6, 1.2%) correspond to Moringa MAG 102 (mono 53%); and
samples 45-48 (conc. 0.15, 0.3, 0.6, 1.2%) correspond to Moringa
MAG 105 (mono 83%)
Table 5.11.1 Formulations for 35% and 40% TAG emulsions using 311
two Moringa MAG distillations: sample 2559/132 (87%
Monoglyceride - high temperature distillation 210°C) and sample
2559/134 (97% Monoglyceride – low temperature distillation 185°C)
Table 5.11.2 Processing conditions for 35% and 40% TAG W/O 312
emulsions
Table 5.11.3 Fatty acid compositions of high and low temperature 313
MAG distillations from Moringa oleifera TAG
Table 5.11.4 DSD analysis results for W/O emulsions (samples 71 – 314
78)
xxxi
Abbreviations
G* complex modulus
IU international unit
IV iodine value
ND no date
O/W oil-in-water
Pas viscosity
xxxii
Pst palm stearin
W/O water-in-oil
Symbols
α alpha
β beta
β’ beta prime
θ contact angle
T°C temperature
xxxiii
Tγ critical interfacial temperature
µm micrometer
χ azimuthal angle
xxxiv
1.0 General Introduction
Herein lies the problem, removal of these ‘building’ blocks, i.e. saturates and their
trans isomer from fat systems, make it difficult to achieve food products which are
functionally acceptable to the consumer. The mechanics and successful removal
of saturated and trans fatty acids in foods (Wassell & Young, 2007) is a complex
problem, as is the successful structuring of low or reduced fat based foods, that
require additional reduction of total saturates (Wassell et al., 2010a; Young &
Wassell 2008). Accomplishing any net reduction of saturated fatty acid materials
for dietary requirements (Winwood, 2011) without negative impact on structure
and functional properties of a particular food (Wassell & Young, 2007; Wassell et
al., 2010a) is more problematic for foods which are already regarded as low
saturated. Therefore, the challenge is intensified for food manufacturers to find
innovative solutions to structuring issues whilst meeting consumer expectations.
1
This study examines this current situation through a literature review. It then
reports on a small pilot study which investigated the effect of several emulsifiers
on crystallisation behaviour in dispersion and emulsion systems, and consequently
led to a critical review. As a result of this preliminary research, extensive work
explored the use of traditional and novel materials, methods of measurement and
analysis. This is subsequently described.
To set the scene, the following sections present a brief overview and history of oils
and fats and their use in water – oil emulsions.
2
1.1 Oils and Fats – Triacylglycerol Based Structurants
Both oils and fats are composed of triacylglycerols (TAG), generally known as
triglycerides and are liquid and solid at ambient temperatures, respectively. A
triglyceride consists of glycerol esterified with three fatty acids, which can either
be three similar ones, called a simple triglyceride, or two or three different ones, in
which case it is a mixed triglyceride. A schematic example is given in Figure
1.1.1 (Madsen, 2003).
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
H O H H H H H H H H H H H H H H H H H
H C O C C C C C C C C C C C C C C C C C C H
H H H H H H H H H H H H H
Linoleic acid
O H H H H H H H H H H H H H H H H H
H C O C C C C C C C C C C C C C C C C C C H
H H H H H H H H H H H H H H H
Oleic acid
O H H H H H H H H H H H H H H H H H
H C O C C C C C C C C C C C C C C C C C C H
H H H H H H H H H H H H H H H H H H
Glycerol Stearic acid
As can be seen in Figure 1.1.1, double bonds are present in some of these fatty
acids. Modification of the fatty acids, usually by means of hydrogenation, is where
the unsaturated fatty acids are transformed into saturated fatty acids. Here, an
example could be C18:1 (oleic acid) going to C18:0 (stearic acid). Such
modifications offer the oils and fats manufacturer a greater flexibility and the
chance to dramatically alter the melting point of the fat. The fatty acid composition
of some natural fats along with other important information is summarised in
Table 1.1.1. Given current trends, the down side to hydrogenation is that during
the addition of hydrogen, trans fatty acids are formed, which are schematically
3
shown in Figure 1.1.2. Selective hydrogenation involves the saturation of the most
polyunsaturated fats first, such that the trans fatty acid concentration increases up
to a point until they themselves are hydrogenated. If the reaction runs to
completion, then the trans isomer is absent. The trans fatty acids can have
substantially higher melting points than the corresponding cis fatty acid, where the
difference can be in excess of 30°C.
Cis
C C C
Trans
H
C C C C
H H H C
Figure 1.1.2 Schematic diagrams of cis and trans configurations (Young and
Wassell 2008).
Table 1.1.1 gives the melting points of the individual oils that are used to make up
the fat blends, but it is also important to know and recognise the melting points of
the fatty acids themselves. These are outlined together with the number of double
bonds they contain in Table 1.1.2.
4
Table 1.1.1 Composition of fatty acids in vegetable oils and fats (Madsen, 2003).
a b
T = trace. refers to the typical value, and refers to the range due to natural
variability – valid throughout the entire table.
acid
0.5a 0.5
Caproic, C6:0
0.2-0.8b T-1.6
8 4
Caprylic C8:0
6-9 3-10
7 5
Capric C10:0
6-10 3-14
48 50
Lauric C12:0
44-51 37-52
17 15 2 1 0.5
Myristic C14:0 T T
13-18 7-17 0.5-5 0.5-2 0-1.5
2 9 7 42 10 21 8 6 5 2
Palmitic C16:0
1-2 8-10 2-9 32-47 7-12 20-27 7-10 4-8 4-7 1-4
1 2 2 5 3 2 4 4 4 1
Stearic C18:0
1-2 1-3 1-3 2-8 2-6 1-3 3-6 2-5 2-5 0.5-2
T 3 0.5 0.5 0.5 0.5
Arachidic C20:0 T
T-0.6 2-4 0.2-1 0-2 0-1 0-1
2 0.5 1
Behenic C22:0
T-3 0-1 0.5-1.5
2 1
Lignoceric C24:0 T T
1-3 1-2
0.2 0.5 0.5 0.5
Palmitoleic C16:1 T T T
T-0.4 T-0.6 0-2 T-1
7 7 15 41 50 29 28 28 22 15
Oleic C18:1
T-8.5 5.5-7.5 11-23 40-52 35-70 22-35 20-35 20-40 12-34 11-24
7
Gadoleic C20:1
5-12
50
Eurcic C22:1
40-55
87
Ricinoleic C18:1
86-92
3 1.3 1 10 30 45 53 61 17 15
Linoleic C18:2
3-6 T-2.5 1-3 5-11 20-25 42-54 40-57 45-68 14-20 11-29
1 6 52 7
Linolenic C18:3 T
T-2 5-14 35-65 6-13
Pure Glycerine 8.8-9.8 13.2-13.5 12.2-12.8 5.5-10.0 8.7-9.9 10.6 10.2 10-12 0.4-10.5 9-9.7
I2 value (Wijs) 81-91 7.5-10.5 14-23 44-54 84-100 99-113 120-141 126-136 55-205 97-108
Saponification
177-187 250-264 245-255 195-205 188-195 189-198 189-195 186-194 188-196 170-180
Value
Melting point °C -10 b. -12 23-26 24-26 27-50 -2 -2b. +2 -20 b. -23 -10 -20 -9
Titer °C 20-24 20-24 40-47 26-32 30-37 20-21 16-20 19-21 11-15
5
Table 1.1.2 Common fatty acids showing their melting points and the number of
double bonds they naturally contain (Source: Chemical Abstracts Service,
www.cas.org)
Fatty Acid No. of Double Bonds Melting point °C
Lauric C12:0 - 44 – 46°C
Myristic C14:0 - 52 – 54°C
Myristoleic C14:1 1 Liquid at 5°C
Palmitic C16:0 - 61 – 62.5°C
Palmitoleic C16:1 1 Liquid at 5°C
Stearic C18:0 - 67 – 72°C
Oleic C18:1 1 13 – 14°C
Linoleic C18:2 2 Liquid at 5°C
Linolenic C18:3 3 Liquid at 5°C
Arachidic C20:0 - 74 – 76°C
Probably all oils and fats will undergo some form of modification before use in a
food product. The most well used modifications are hydrogenation,
interesterification and fractionation. Depending on the naturally occurring starting
point of saturation or unsaturation of the said oil, the degree of modification might
result in a more saturated or unsaturated fat having a higher or lower melting point
than the starting material. The importance of this cannot be overemphasised,
because it provides structure, texture and hardness to the food product, lubrication,
and in terms of emulsion stability, can provide specific crystallisation behaviour,
necessary for continuous fat phase dispersions ⁄ emulsions. The extent and degree
of satisfying these criteria are dependent on the amount of solid fat, size, shape and
interactions of the crystals (Wassell & Young, 2007). Figure 1.1.3, shows
6
schematically, the degree of modification liquid oil required to obtain the desired
solid fat structure, which could be used, for example, in margarine or shortening
production. Detailed reviews about the subject of structuring edible fat based
systems are discussed by Wassell & Young (2007) and Wassell et al., (2010a).
100
Palm Kernel
Palm
Area of Liquid Oils
desired fat
crystal
properties
Solids (%)
Modification of oils
0
10 45
Temperature (°C)
Figure 1.1.3 Desired solid phase lines for hardstocks (dark area) compared to
naturally occurring oils and fats (adapted from van Duijn et al., 2006).
7
1.2 Edible Water-in-Oil Emulsions – Background
1.2.1 History
The term “margarine and spreads”, refers to a group of products, which are likened
to butter, but have different fat contents. Margarine was invented in response to a
request from the French Government of Napoleon III for a less expensive, longer
life replacement for butter, available in higher quantities and at a lower price. The
big difference between butter and margarine is the source of the raw materials.
Butter is made with the cream of cow’s milk and margarine with vegetable oils or
in the beginning with animal fat (Young & Wassell, 2008; Verstraete, 2011).
The definition of margarine has been established. Anything below 80% fat
content, by definition, is not margarine, and must be referred to as a spread;
reduced fat, low fat or otherwise. Legislation is complex and terms such as three-
quarter fat, half-fat and fat content X% are routinely seen. Three-quarter fat refers
to a fat percentage of 60-62%, half-fat to 39-41% specifically. Reduced fat falls
8
within the range 41-62%, and low fat or light products under 39%. Overall, there
is a general consensus that a fat spread product, butter, margarine or other must
have a fat content between 10 and 90%. These options are summarised in Table
1.2.1
Table 1.2.1 Product type versus their fat content in percent Source a Article 5 of
the Council Regulation (EC) No. 2991 / 94, laying down standards for spreadable
fats, b Other industry classifications
9
Table 1.2.2 The average composition of margarine and table spreads in Europe
1.2.4 Hardness
10
debated. One study found the hardness of the margarine has no linear correlation
with the SFC (Moziar, De Man, & De Man, 1989). However, hardness is also
linked to the total lipid composition (including emulsifiers), polymorphism,
crystallisation behaviour and the microstructure. (Vereecken, Meeussen, Lesaffer,
& Dewettinck, 2010; Wassell et al., 2010a).
There are strong correlations between the SFC% and the plasticity of margarine.
This also has implications for the overall melting characteristics of margarines and
spreads at various temperatures. This is particularly important for low fat spreads.
At low temperatures, i.e. 4–10°C the value of the SFC% will lead to an indication
of spreadability when the product is removed straight from the fridge. The SFC
values should not exceed 32% at 10°C (Lida & Ali, 1998). Higher ambient
temperatures – of 20–22°C – determine the product stability and resistance to oil
exudation in domestic conditions and therefore the SFC should be at least 10%.
SFC’s at approx. 35–37°C provides data on the thickness and mouth feel
characteristics as well as flavour release, and should be around 1–3% (Wassell &
Young, 2007).
Ideally, the SFC, melting profile, and crystallisation process is such that texturally
the fat blend gives the optimal structure and sensory properties characteristic of
table margarine or spread (Chrysan, 2005).
11
As discussed in 1.2.5, the SFC% must be determined so as to allow the emulsion
to attain spreadibility. Selected fat blends tend to be softer where there is much
less fat content and therefore, demands on the emulsifier are greater to maintain
the stabilisation of the water phase. The preconditions for a stable low fat spread
are small water droplets and a stable emulsion. Other components in the system,
such as milk proteins, act to give a more open emulsion resulting in improved
flavour release; but they also make controlling the water dispersion more difficult,
with the consequence of shorter shelf life. The selection of emulsifier therefore
depends not only on the fat content of the spread product, but also on the aqueous
phase and its composition. This involves consideration of things such as local
water hardness, proteins, and hydrocolloids.
1.2.7 Emulsifiers
On the interface, the hydrophilic groups are in contact with water and the
hydrophobic groups are in contact with the oil. The interfacial tension is then
reduced by the emulsifier, resulting in an easy mixing of oil and water.
Typically in margarine spread type W/O emulsions, the most common emulsifiers
(surfactants) used are monoacylglycerols (know as Mono and diglycerides or
monoglycerides or E471) or MAGs (Juriaanse & Heertje, 1988), which have
amphiphilic properties (Krog, 1977). Table 1.2.3 shows distilled monoglycerides
with varying degrees of saturation (a higher iodine value = more unsaturation), and
its relationship to emulsion stability.
12
1.2.8 PGPR
In low-fat spreads with high water and milk protein content, polyglycerol
polyricinoleate (PGPR), known for its exceptional water-binding properties
secures the necessary emulsion stability and water dispersion. According to the EC
Directive 95/2/EC, PGPR (E-476) is allowed in low-fat spread with 41% fat and
less in a maximum dosage of 0.4% (Bastida-Rodríguez, 2013).
For very low fat spreads the use of a combination of monoglyceride and
polyglycerol ester emulsifiers is claimed to result in wider processing latitude and
enhanced finished product stability (Chrysan, 2005; Stewart & Hughes, 1972)
Table 1.2.3 Stability of a 40% W/O emulsion over time (Information received
from Danisco / DuPont Oils & Fats, 2009)
WATER SEPARATION
Iodine value 55 8% 27 %
Iodine value 80 3% 16 %
Iodine value 80 0% 0%
+ PGPR
In the preparation of the margarine / spread and depending on the ratio of water to
oil in the said emulsion, it is usually necessary to select the correct type of
emulsifiers to aid product integrity. Table 1.2.4 shows typical emulsifier selection
according to fat content and presence of protein.
13
Table 1.2.4 Typical emulsifier selection for W/O spreads with and without
proteins (Information received from Danisco / DuPont Oils & Fats, 2009).
There are several common tests for emulsion stability (Chrysan, 2005), these being
electrical resistance (conductivity), droplet size distribution (DSD), cardboard test
(simple backward and forwards spreading action), ambient and cooled storage to
assess oiling out and texture analysis. The principles and methods of these and
other analyses used within the scope of this thesis are discussed in the General
Materials and Methods (2.0). The physical properties of the emulsion can have
microbiological implications (Juriaanse & Heertje, 1988; Charteris, 2007); this
does not form part of this study.
14
1.3 Crystallisation and Crystal Growth
1.3.1 Introduction
An overview of some key areas in the physical properties of fats and oils
(triacylglycerols / TAG) are presented in brief, and where appropriate additional
explanation is provided.
Within edible oils and fats, the subject of crystallisation is usually discussed in the context
of two kinds of nucleation:
Once stable nuclei have formed, these grow into crystals by incorporating
molecules in the interface from the melted oil. So that as crystallisation continues,
the degree of super-saturation in the system decreases (Timms, 1985).
Once a fat has crystallised, the individual crystals may aggregate to form a three
dimensional network which traps liquid oil through capillary forces. With time,
crystals transform into more stable state, depending on the chemical composition
of the fat and environmental conditions. The rate of crystal growth is proportional
to the degree of supercooling and inversely proportional to the viscosity of the
melt (Hartel, 2001; Timms, 2007).
During crystal growth, heat is released. Rapid cooling of the melted oil leads to the
simultaneous formation of many nuclei which grow into small crystals. Slow
cooling leads to a smaller amount of nuclei which have time to grow into larger
crystals (Himawan et al., 2006).
15
1.3.2 Polymorphism and its Identification
Figure 1.3.1 A schematic of triacylglycerol arrangement in the β'-2, and β-3 form.
LS is the long spacing, t is the angle of tilt of the triglyceride configuration
(adapted from Timms, 1984).
16
The angle of tilt, t, will depend on the LS value such that larger LS values result in
smaller angles of tilt and vice versa.
α, β’ and β crystals (Hoerr, 1960) can be formed directly from the melt, or α to β’
to β, but this is not reversible. Although any polymorph can be returned to the
liquid phase by raising the temperature above its melting point, polymorphic
transformations are irreversible, where β cannot transform back to β’ and β’ cannot
revert back to α (Sato, 1989).
SS
Figure 1.3.2 Fatty acid units showing the short spacing between the individual fat
units (source: Young & Wassell, 2008)
Measurement of short spacing between the fatty acids enables quantification of the
type and ratio of the fat crystal forms in a given blend. A number of analytical
techniques have established that margarines and spreads are preferred when the
crystal polymorph is in the β’crystal form, because this form imparts desirable
textural qualities (Wassell & Young, 2007; Wassell et al., 2010a).
17
The influence of processing however can have dramatic impact on crystallisation
kinetics. Palm oil is probably the most widely used of vegetable oils. It is naturally
β’ tending, largely because of its diverse fatty acid profile and particularly high
content of palmitic acid (Berger & Idris, 2005). However, if processed incorrectly,
these benefits are lost, and because palm oil also contains unusually high content
of diglycerides ~ 6 to 7 %, the diglycerides have anti-crystallisation properties that
can influence crystal kinetics (Siew, 2002).
The different crystal types, α, β’ and β each have their own configurations, shown
in Figure 1.3.3., (Hernqvist, 1988), and it is well known that fat crystals with
similar chain lengths, e.g. hydrogenated sunflower oils, transform more quickly
from the β’ to β form (Yap, deMan, & deMan, 1989).
α (H)
β’ (Ο⊥) b
β (Τ ) b
a
c b c
a a b
Figure 1.3.3 The three projections of α, β’, and β crystal forms (Source:
Hernqvist, 1988).
18
As the crystal form changes, the texture likewise changes. This typically takes
place under storage, with the usual transition being from β’ to β. During this
transition crystal size increases dramatically, from ~3-5 microns to ~100 microns
respectively, and likewise melting point. The result is that the margarine now
infers a sandy / gritty mouth feel (Kristott, 1999).
19
1.3.3 Supercooling
Crystallisation can take place once a liquid phase is cooled below its melting point.
This is due to activation energy (ΔG*), which must be overcome before the
transition from liquid to solid (Figure 1.3.4). Crystallisation of a single species of
TAG molecules requires lowering the temperature of the melted fat below its
melting point. The degree of sub-cooling, known as ΔT, is simply the difference
between the melting point and the actual temperature. The lower the temperature,
the higher the driving force for crystallisation.
Pure oils can be supercooled by more than 10°C, before any crystallisation is
observed.
∆G*
Liquid
∆G
Solid
20
Table 1.3.1 The influence of supercooling on the solubility and radius of
triglyceride crystals.
RADIUS OF CRYSTAL
SUPERCOOLING INCREASE IN
(°C) SOLUBILITY
(µ) (Å)
Data for Table 1.3.1 is supplied from Dr. Ralph Timms (Danisco AS, lecture, 2003
& 2004).
Table 1.3.1 shows that small crystals have an enormously increased solubility and
require a lot of supercooling to cause expanded crystallisation. It can be seen that
at a radius <10Å the degree of supercooling required would be very large. Hence
spontaneous or homogeneous nucleation rarely, if ever, occurs.
21
1.3.4 Post-Crystallisation
Fat hardness increases during post-crystallisation and Figure 1.3.5 shows a process
described as “sintering”, according to the following sequence:
A. Nucleation
B. Crystal growth
C. Solid bridge formation
D. Flocculation of small nuclei between fat crystals
(Johansson & Bergenstahl, 1995).
Figure 1.3.5 Schematic stages of sintering (Source: Johansson & Bergenstahl, 1995)
22
1.3.5 Technical Impact of Process Conditions on Post-Crystallisation
The texture of reduced fat spread products can be critical, especially where
reduction of fat in very low fat spread products leads to loss of texture due to
“uncontrolled” post-crystal growth (Sato, 2001). Again, cooling and agitation
(Bell, Gordon, Jirasubkunakorn, & Smith, 2007) in the production process may
influence product quality as shown in Figure 1.3.6. This effect could be
experienced by the end-user noticing textural issues such as graininess.
Rapid
Cooling
with
agitation
Slow
Cooling
without
agitation
Figure 1.3.6 Effect of cooling and agitation on crystal growth. One scale unit =
100 microns. (Image supplied courtesy of Danisco / DuPont in their Technical
Paper: Fat Crystallography – a Review. TP1504-2e)
23
1.3.6 Palm Oil
The effect of switching to trans free fat blends, has usually meant working with
palm oil products. Palm oil is not the subject of this thesis, but it is considered
necessary to acknowledge its use within the edible oils and fats industry and
thereby provide some general back ground to it (Basiron, 2005). Palm oil is
inherently β’ tending and has a natural semisolid consistency, useful to margarine
applications.
In tackling the issue of reducing trans saturates, palm oil offers a natural solution.
It does not necessarily require a hydrogenation step, which can lead to formation
of (dependant on degree of hydrogenation) trans isomers (Wassell & Young,
2007). The whole palm oil naturally contains significant amounts of saturates
(Wassell et al., 2010a).
Palm is probably the most important oil in terms of fractionation because of its
unique fat profile (Sahasranamam, 2005) which can be broken down into
individual fractions and will naturally and obviously, become more or less
saturated (Timms, 2007). The degree of unsaturation of the palm oil (fractions),
can present crystallisation issues due to their inherent diacylglyceride (DAG)
content (Siew, 2002; Wassell & Young, 2007; Wassell et al., 2010a). The presence
of DAG slows the crystallisation process and changes the melting behaviour (Siew
& Ng, 1990; Siew, 2002). Slowing of the crystallisation process leads to post-
hardening during storage of palm-based margarines (Wassell & Young, 2007). By
gaining control over the DAG content of palm oil, Kristensen and Wassell (2006)
showed that not only control, but also improvement over the crystallisation rate
was possible (Kristensen, Wassell, Mikkelsen, & Søe, 2005).
Control of the crystallisation rate can be altered by either varying the degree of
cooling, type of additive and its concentration, or by manipulating the DAG
content (Wassell & Young, 2007). The selection and concentration of emulsifier to
enhance crystallisation kinetics has been reviewed extensively (Smith, Bhaggan,
Talbot, & van Malssen, 2011; Wassell & Young, 2007).
24
1.4 Literature Review
The issue of trans fatty acid (TFA) replacement has been the subject of much
discussion and debate within the food industry (Wassell & Young, 2007) as well
as in the clinical field for health reasons (FSA, 2008). Now, there is even stronger
focus on reducing total saturated fatty acid contents (Beaglehole et al., 2011;
Bradley, 2012; Gortmaker et al., 2011; Micha & Mozaffarian, 2010; Mozaffarian
et al., 2010; NICE, 2010; Siri-Tarino et al., 2010; Yamagishi et al., 2010;
Swinburn et al., 2011; van Camp, Hooker, & Lin, 2012; Vesper, Kuiper, Mirel,
Johnson, & Pirkle, 2012; Wang, Y.C. et al., 2011; WHO, 2011; Winwood, 2011).
Currently, some governments are considering its taxation to control its presence in
foods (Lucas & Rappeport, 2012). However, reformulating is often expensive and
not simple, because consumers expect particular textural properties in a given food
product.
1.4.1.1 Texture
The aroma, texture and mouth sensations experienced when consuming foods are
strongly dependent on their fat content and can eventually shape dietary choices in
the long term (Drewnowski, Shrager, Lipsky, Stellar, & Greenwood, 1989;
Drewnowski & Almiron-Roig, 2009). Thus, the idea of reformulating familiar
foods which historically contained significant amounts of fats and oils (and usually
contain significant levels of saturated fat) into lower fat versions, is challenging
not only technologically (Pothiraj, Zuñiga, Simonin, Chevallier, & Le-Bail, 2012),
but also from a consumer acceptance point of view. Research into consumer
perceptions of a new food product, is often compared to memorized experience
from a similar consumed food. Consumers may tend to conclude their decisions on
the basis of their memory of a food, and not on the basis of their actual perceptions
(Mojet & Koster, 2005; Wassell et al., 2010a).
25
1.4.1.2 Health
1.4.1.3 Legislation
Legislation, and literature (WHO, 2011; Wassell et al., 2010a) state that trans fatty
acids should be removed from food, but current research (Stender, Astrup, &
Dyerberg, 2009) suggests that this is insufficient in terms of creating healthy
foods. Removal of, or at least drastic reduction of, not only trans containing fats,
but all saturated fats is also required (Beaglehole et al., 2011; Bradley, 2012;
Gortmaker et al., 2011; Micha & Mozaffarian, 2010; Mozaffarian et al., 2010;
NICE 2010; Siri-Tarino et al., 2010; Yamagishi et al., 2010; Swinburn et al., 2011;
van Camp et al., 2012; Vesper et al., 2012; WHO, 2011; Wassell & Young, 2007;
Wassell et al., 2010a).
26
removal from bulk fat based systems has been the main focus of research in this
field (Wassell & Young, 2007), and in some countries remains so (Wassell et al.,
2010a), e.g. Brazil and USA. However, the predominant global attention is on
reducing total saturated fats, which of course includes trans fats (Wassell &
Young, 2007; Wassell et al., 2010a). That said, while the problem is global; for
some cases in emerging economies, the degree of importance placed on providing
food products with lower trans and saturated fats has taken lower priority (Wassell
et al., 2010a).
Whilst there is general agreement that total fat levels should be reduced, there is
disagreement as to the best route (Wassell et al., 2010a). Talbot, Favre, and
ThÖrig (2007), advocate the use of saturated fats preferably from natural sources
(Wassell et al., 2010a). Marangoni (2007) suggests that the use of saturated fat in
the absence of trans fats is not a solution.
The dramatic switch to palm oil products and away from high trans containing oils
created challenges for food producers to satisfy nutritional demands, and still
deliver acceptable functional performance (Wassell & Young, 2007; Wassell et al.,
2010a).
A review article entitled: Food applications of trans fatty acid substitutes (Wassell
& Young, 2007), assessed the increasing pressure to remove trans fatty acids from
food products. Further, it reviewed how practical applications of trans-free
solutions could be successfully implemented and suggested this might be achieved
via multiple routes (Wassell & Young, 2007).
When pulled together using a multidisciplinary approach several areas are thought
to demand further study. As a consequence, following on from Wassell & Young
(2007) a review to study the current state of the art was continued and identified
important areas for consideration: Raw materials, reducing total saturates;
crystallisation kinetics and interfacial behaviour (Wassell et al., 2010a). These are
discussed next.
27
1.4.2 Raw Materials
The initiatives to either phase out or to reduce trans fatty acids, have not entirely
addressed the issue of total saturates; nor have they entirely addressed the debate
over inclusion of fully hydrogenated oil and fats (van Duijn, Dumelin, &
Trautwein, 2006), which are trans-free (Wassell & Young, 2007; Wassell et al.,
2010a).
28
texture were reviewed and all were TAG based solutions (Wassell & Young,
2007). Currently, the literature acknowledges the necessity to couple traditional
TAG based solutions with novel materials, because it is unlikely that one novel
structurant could satisfy all requirements. An example might be to use a novel
lipid based mixture in a TAG-based dispersion or W/O emulsion, a more realistic
approach (Johansson, Bergenståhl, & Lundgren, 1995; Rousseau & Hodge, 2005;
Wassell et al., 2010a). Whichever route is adopted, this would have implications
on either crystallisation behavior or emulsification, or both.
The effect of switching to alternative raw materials for TAG based systems (with
or without aqueous phase), and reducing total saturates, will likely influence
kinetic behaviour and eventually textural properties. This could be more prevalent
in those retail products to which consumers are accustomed. It is likely that
synergies obtained through several disciplines, coupled with an understanding of
the thermodynamic mechanical properties and critical design of real systems, play
a vital influence (Wassell et al., 2010a).
29
For low TAG W/O, it is usual to soften the TAG continuous phase, which
automatically results in less saturation. However, this results in less solid TAG
content, hence reduced steric stabilisation of the internal (aqueous) phase (Krog,
1990), from a three dimensional network of crystal aggregates. As a consequence,
there is potential compromise of the particle attachment to the water-in-oil
interface which is essential to Pickering stabilisation (Johansson et al., 1995;
Ghosh, Tran, & Rousseau, 2011; Pickering 1907). To compensate, aid from a
surface active material that could possibly fulfil both crystallisation and
emulsification tasks is likely needed. This situation may be strongly linked to the
rate of fat crystallisation in the TAG phase as well as on the amount and form of
fat crystals formed. Variations in temperature and/or shear rate may also be
important (Johansson et al., 1995).
1.4.5 Conclusion
A review of TAG based materials, crystal kinetics and interfacial behaviour shows
that whilst low fat emulsion technology is not new, the use of one or several novel
applications of long-chain behenic based emulsifiers to low or reduced TAG based
W/O emulsions is not found in the literature (Wassell & Young, 2007). Stabilising
W/O emulsion structures at lower TAG levels, using these alternate strategies,
required more investigation.
To support this, pilot studies were conducted. First, the use of MAG / TAG
mixtures for enhancing crystallisation rates in several anhydrous dispersions.
Second, an investigation of the effect of diversified fatty acid mixtures of MAG in
the presence of an aqueous interface (shown in 1.5).
30
1.5 A Pilot Study – Preliminary Investigations for Enhancing Crystallisation
of Anhydrous TAG Dispersions and W/O Emulsions
Several aspects of crystallisation have been considered previously (1.3 and 1.4).
There is no intention to repeat the theory of crystallisation and nucleation in lipid
based systems as this has been presented and discussed by others (Garti & Sato,
2001; Himawan et al., 2006; Povey, 2001; Timms, 1984; 1985; Wassell & Young,
2007).
For practical reasons, it was thought necessary to provide supporting data from
preliminary investigations to show the effects of MAG / TAG additives on the
behaviour of anhydrous continuous model systems. These considerations provide
additional comprehensive background to support this thesis, but which is not
specifically detailed in publicly available literature.
The literature review (1.4) identified that long-chain emulsifiers are thought to be
potentially important. A pilot study was conducted on the use of a MAG based on
C22:0.
The following are pilot study results and data from investigations into
crystallisation behaviour of anhydrous dispersions together with available MAG
and TAG materials.
31
1.5.1.1 Materials and Methods
Arbitrary concentrations were used simply to observe effect and achieve proof of
concept. It was not to focus on prohibitive commercial or legal issues on the use of
a behenic based emulsifier. Samples were assembled and tested as follows:
32
1.5.1.2 Results - Crystallisation of RBD Palm Oil and Commercial Hard and
Soft Anhydrous Fat Blends
Based on Figure 1.5.1 it is possible to see that adding DIMODAN HP (HP) and
DIMODAN MB-90 (MB-90) impacts on solid fat content of palm oil. The solid
fat content was increased. Examination of Figure 1.5.2 shows this more clearly. A
plateau occurs at 3-5 minutes, as emulsifier concentration was increased, this
resulted in a direct relative influence on SFC%.
1% MB-90 showed increased performance compared to 1% HP. Nucleation onset
was most significant from time zero.
As crystallisation continued, a concentration above 1% of either HP or MB-90
showed similar performance compared to 1% MB-90. This suggested from this
test, no clear advantage with addition of MB-90 beyond 1% in RBD palm oil.
20
15 palm oil
1% Dim HP
10 2% Dim HP
SFC %
3% Dim HP
5 1% Dim MB90
2% Dim MB90
3% Dim MB90
0
0 5 10 15 20
-5
Time (min.)
33
20
Plateau
15
Crystallisation palm oil
onset. 1% Dim HP
10 2% Dim HP
SFC %
3% Dim HP
5 1% Dim MB90
2% Dim MB90
3% Dim MB90
0
0 2 3 4 5 6 8 10
-5
Time (min.)
Further trials using two commercial fat blends, a hard (H) and soft (S) fat blend
were conducted. Their fatty acid compositions are shown in Table 1.5.1. The rate
of crystallisation results are shown in Figures 1.5.3 and 1.5.4. DSC results are
shown in Table 1.5.2, where data is based on the crystallisation onset starting
point.
34
Table 1.5.1 Fatty acid profiles of two palm based fat blends.
Fatty Acid (%) Profile Hard blend (H) Soft blend (S)
Methyl-ester:
C8:0 0.2 <0.1
C10:0 0.2 <0.1
C12:0 2.6 0.3
C14:0 1.7 0.7
C15:0 0.1 0.1
C16:0 40.9 28.7
C16:1 0.1 0.3
C17:0 0.2 0.2
C18:0 4.6 3.5
C18:1 33.4 44.3
C18:2 10.4 13.8
C18:3 3.4 5.2
C20:0 0.4 0.5
C20:1 0.5 0.8
C20: unclassified 0.2 0.3
C22:0 0.2 0.3
C24:0 0.1 0.1
Total 99.32 99.12
35
Hard blend
30 1% Dimodan HP (H)
2% Dimodan HP (H)
25
3% Dimodan HP (H)
15
Soft blend
1% Dimodan HP (S)
10
2% Dimodan HP (S)
3% Dimodan HP (S)
5
1% Dimodan MB90 (S)
-5
Time (min. at 20°C)
35
35
Hard Blend
30
1% Dimodan HP (H)
2% Dimodan HP (H)
25
3% Dimodan HP (H)
15
Soft Blend
1% Dimodan HP (S)
10
2% Dimodan HP (S)
3% Dimodan HP (S)
5
1% Dimodan MB90 (S)
-5
Time (min. at 20°C)
Rate of Crystallisation:
Hard Blend (H): From Figure 1.5.3 it is difficult to say which and/or if the
emulsifiers promoted any effect. However, in Figure 1.5.4, it can be seen that
between 4 and 8 minute, MB-90 curves are above the pure H, whereas HP curves
are below. So, it seems that MB-90 presented a greater and more pronounced
effect than HP.
Soft Blend (S): It is much easier to see the emulsifier effect on S than H. Both
emulsifiers (all concentrations) presented higher SFC% than S. Figure 1.5.4 shows
that at 4 to 8 minutes, all emulsifier curves are above S and MB-90 is above HP.
36
DSC:
Hard Blend (H): The tabulated DSC data shown in Table 1.5.2, shows that H
starts crystallisation onset at ~24.6°C. With the addition of 1% HP, there was little
emulsifier effect noted. This suggests that DSC of pure H and H+1% HP are
similar. Increasing to 2% HP, it was possible to see some effect, where the
crystallisation onset started at a higher temperature (~29.2°C). Whereas, increasing
to 3% HP significantly enhanced the crystallisation onset, which then started at
~34.5°C.
In comparison, Table 1.5.2 shows that adding 1% MB-90 to blend H, had a similar
effect to that of adding 3% HP. The addition of 2% MB-90 significantly increased
the crystallisation onset to ~45.0°C. Increasing the MB-90 to 3%, raised onset to
~50.7°C.
Soft Blend (S): The DSC data (Table 1.5.2) revealed that S started crystallisation
onset at ~18.7°C. The inclusion of 1% HP had a small effect, increasing the
crystallisation onset to 22.9°C. The addition of 2% HP and 3% HP increased the
effect further, so that crystallisation onset began at ~32.7°C and ~51.1°C
respectively.
37
In comparison, the addition of 1% MB-90 to S, increased onset to ~38.2°C. This is
significantly more functional, than the addition of 2% HP. At 2% MB-90,
crystallisation onset started at ~47.6°C, almost 10°C higher than with 1% MB-90.
Increasing to 3% MB-90, led to crystallisation onset at ~54.6°C, which is only
slightly greater than the effect of 3% HP.
These results showed both HP and MB-90 had the most impact on the
crystallisation onset temperature in the soft (S) fat blend.
In the soft fat blend (S), the comparative effect of MB-90 to HP was greater at
lower concentrations (1-2%), but when the concentration was increased to 3%,
MB-90 showed only a small advantage.
Comparing the two emulsifiers, it was shown that MB-90 had a greater effect than
HP in both fat blends.
To show the relative increase of the hard (H) and soft (S) blends (Table 1.5.1)
compared to RBD Palm oil, the rate of crystallisation of the three pure fats are
plotted together in Figure 1.5.5 which shows their relative SFC% increase over
time.
Figure 1.5.5 Relative increase in SFC% of RBD Palm Oil, Hard and Soft blends
38
1.5.1.3 Interim Conclusions of Pilot Study on Crystallisation
The results showed that MB-90 tends to increase both SFC and crystallisation
onset compared to HP. The influence of both these distilled monoglycerides
seemed dependent on both the concentration and the chain length of the fatty
acids.
Overall, there was a positive effect on the selected fat blends, when a MAG rich in
C22:0 was included. This promoted a higher rate of crystallisation and earlier
onset than other tested MAG / TAG mixtures. In this work it was generally
concluded that the harder the fat blend, the greater the influence of the
monoglyceride performance when compared to triglyceride. Whereas, for the
softer TAG based blend, when C22:0 rich MAG (MB-90) and HLEAR (fully
hardened low erucic acid rapeseed oil) were combined 1:1, this resulted in equal
affect compared to a single C22:0 based MAG.
The measurement of the nucleation and crystal growth rates of MAG in TAGs is
not precise (Himawan et al., 2006). However, this is partially solved by the
determination of overall crystallisation kinetics and the use of models such as the
Avrami equation, which is not considered central to this investigation, because in a
number of industrial applications fat crystallisation occurs in dispersions
(emulsions). MAGs have been found to accelerate not only the onset of
crystallisation of various fats under agitation, but also in the presence of water
(Chrysan, 2005; Himawan et al., 2006). Since retail fat blends tend to be a softer
39
(less saturation) continuous phase together with a secure internal aqueous phase,
this has implications not only for crystallisation but also emulsification, especially
where C22:0 rich MAGs are present. This was explored next.
Real food systems are normally emulsions and inherently have interfaces which
are most often stabilised by the assistance of emulsifiers (Krog, 1975; 2001). As
already described (1.2.7), emulsifiers concentrate at the surface, either creating
interfacial film or strongly influencing other material at or close to the interface.
They thereby act as a mechanical barrier between two immiscible liquids,
preventing flocculation or coalescence (Krog, 1977).
The stability and rheological properties of emulsions are largely determined by the
size of the dispersed droplets and how they interact with each other. The degree of
particle-particle interactions is mainly dependent on the structure and composition
of the interfacial film which may consist of adsorbed layer of proteins and/or
emulsifiers as in the case of a W/O interface (Johansson et al., 1995; Krog, 1990;
Krog & Larson, 1992; Krog & Sparsø, 2003).
Usually distilled monoglycerides (E471) are used for margarine and spread
production. A distilled monglyceride composition is related to the fat source from
which it is made (1.4). Depending on the fatty acid residue of the distilled
monoglyceride, the interfacial film can have different rigidity or visco-elastic
behaviour (Krog & Sparsø, 2003; Rousseau, 2000). The type of film affects not
only stability of the emulsion during formation but also stability of the final
product during shelf life and consumption (spreading, eating properties).
40
strong van der Waals attraction forces between the hydrocarbon chains (Krog,
1997). Emulsifiers in adsorbed monolayers can act as templates for the surface
crystallisation of triglycerides. Further, emulsifiers which contain saturated
hydrocarbon chains are known to be effective initiators of fat crystallisation,
whereas those with unsaturated chains are less effective (Euston, 2008; Wassell &
Young, 2007).
It should be noted that the standard hydrophilic - lipophilic balance (HLB) was not
used in this pilot study because the mixing complex properties of several
emulsifiers is removed from individual emulsifiers and is therefore an unreliable
guide for emulsion stability (Boyd, Parkinson, & Sherman, 1972). Further,
Bergenståhl (2008), says the HLB values do not include the important
crystallisation properties of monoglycerides; fatty acid composition is more
important (Keogh, 2006).
In the context of reduced or low TAG W/O emulsions, the degree of saturation /
unsaturation of the fatty acids have a direct bearing on the technical influence of
the emulsifier (Chrysan, 2005). Compared to lower TAG contents, it is known
(Garti & Remon, 1984) that a distilled highly saturated fatty acid based MAG e.g.
(IV 2), is more suitable for full-fat margarine. Whereas for low TAG spreads, a
distilled TAG with more unsaturation (>IV 40) has a theoretically more flexible
film (interface) because of the loosely packed monolayers, stabilising the emulsion
and final product when spread. However, this is not always correct and can only be
used as a rough guideline.
41
The effect of emulsifiers on interfacial tension between oil and water is highly
influenced by temperature (Krog, 1990). It is also known that the actual
composition of the emulsion droplet surface is the key denominator of most
surface interactions (Bergenståhl, 2008). There appears to be no information or
practical examples reported in the literature about a wide distribution of MAG
fatty acid chain compositions to aid stability. The effect at the interface of
temperature and fatty acid diversity is shown in Figure 1.5.6, where four
preparations are made and held at three temperature regimes. It is easy to observe
how, depending on the selected temperature and degree of saturation, there is a
tendency for mono-crystal formations to occur towards the bulk oil or towards the
aqueous phases. Subsequently, as part of this pilot study, this was explored next.
1 2 3 4 1 2 3 4 1 2 3 4
Figure 1.5.6 The effect of temperature and fatty acid diversity at the interface: a
preparation of water-soybean oil (1:1) with or without 4% emulsifier (E471) in the
oil phase is observed at three temperatures 5°C, 20°C & 40°C (Image supplied
courtesy of Danisco A/S., 2003)
42
soybean oil, palm oil, sunflower oil and fully hardened palm oil. The last column
shows a fatty acid composition of two emulsifiers: Fully hardened Palm
(DIMODAN® HP) and Sunflower (DIMODAN® U/J) in a ratio of 1:1 This
mixture or ratio, results in a more diverse fatty acid composition, which
hypothetically may lead to additional emulsifying properties (Bergenståhl 2008).
Table 1.5.3 Commercial distilled monoglycerides and their prevalent fatty acids
(*approx. 50% trans fatty acids)
Application trials were performed on reduced fat 50% spreads with milk proteins,
using an abnormal (hard) fat blend under typical scraped surface pilot plant
processing conditions (refer to General Materials and Methods 2.0). The
formulation used, is described in Table 1.5.4 to test the following: a) Milk proteins
against monoglycerides, destabilising emulsion during production; b) high water
43
content giving risk of emulsion inversion during production; c) hard fat blend,
crystals of which can cause water separation (coalescence) when spread (worked).
Samples were assessed using a simple cardboard test; an abuse test, by spreading
back and forth across the cardboard surface and then subjectively gauging the
degree of separation of the emulsion.
Table 1.5.4 Formula of 50% reduced fat W/O emulsions and details of the major
fatty acids (SMP = skimmed milk powder).
Trial no. 1 2 3 4 5
WATER PHASE:
Water 48 48 48 48 48
Salt 0.5 0.5 0.5 0.5 0.5
SMP 1 1 1 1 1
pH 5.5 5.5 5.5 5.5 5.5
FAT PHASE:
Coconut oil (fully hardened) 15 15 15 15 15
RBD liquid Rapeseed oil 15 15 15 15 15
Soya 41°C 50 50 50 50 50
RBD Palm oil 20 20 20 20 20
FAT total 49.8 50 50 50 50
DIMODAN® HP (C16:0+C18:0) 0.4 0.2
DIMODAN® S-T (C18:1+C16:0) 0.4
DIMODAN® UP (C18:2+C18:1+C16:0) 0.4
DIMODAN® U/J (C18:2+C18:1) 0.4 0.2
Lecithin 0.2
PPM ß-carotene 6 6 6 6 6
Butter Flavouring 0.01 0.01 0.01 0.01 0.01
A simple evaluation test was carried out by “abuse” test on cardboard with
stainless steel knife moving the prepared emulsion back and forth over the surface
of the cardboard as shown in Figure 1.5.7
44
Figure 1.5.7 Image of cardboard evaluation test by spreading (Image supplied
courtesy of Danisco A/S., 2009)
A separation test does not give a complete, objective picture of sample stability; it
is a subjective evaluation, but nevertheless, it is very informative. Regarding the
five samples prepared in Table 1.5.4, the following was observed:
3. Same as 2.
4. As 1.
45
Table 1.5.5 Results of preparations from Table 1.5.4; the average water droplet
size distribution
1.5.2.5 Conclusion
These results showed that capability to secure a narrow droplet size distribution
was likely the result of diversified fatty acid residue in the lypophilic part of the
monoglyceride mixtures, thereby aiding interfacial and functional properties. In
this case it was an improvement on using individual monoglycerides that may not
have had sufficient fatty acid diversity.
As total TAG levels decrease below 50% of the total emulsion, and therefore by
default naturally follow a decrease in total saturates, the emulsifier interactions
may become more critical especially where proteins are still present. (Chrysan,
2005; Madsen, 1987).
There are still unanswered questions as to the effects on W/O emulsion, when
utilising a range of diversified fatty acid mixtures, incorporating the use of hydro-
carbons beyond C18:0 chain length. As reported earlier, there appear to be some
benefits to using much longer saturated chain lengths (1.4).
46
1.5.3 Very Low W/O TAG (12%) Based Emulsion - With and Without
MAG/TAG Additive
In the case of those products already classified as low fat (<41%), reducing total
saturates by re-engineering to lower fat contents, or improving the general melting
/ firmness of W/O spreads, presents a problem for the product developer. This is
because TAG’s are the building blocks of the continuous phase for W/O
emulsions. Therefore, if reducing the structuring material, a “scaffold” of sorts is
still necessary.
A second obstacle is the increasing use of palm oil products, which may
potentially (depending on degree of DAG content) effect crystallisation. Thirdly,
PGPR is often utilised in this product category to secure emulsion stability (Garti
& Remon, 1984; Goubran & Garti, 1988; Rousseau, 2000). It is well documented
that blending emulsifiers will always produce a more stable emulsion (ICI
Americas Inc. 1980) as opposed to using a single emulsifier, by the fact that
combining a water soluble hydrophilic one with an oil soluble lipophilic type will
produce a denser interfacial film due to the hydrophilic and lipophilic portions of
the molecules sitting on different sides of the interface (Garti & Remon, 1984).
Based on the preliminary findings in 1.5.2 for the 50% reduced fat W/O emulsion
(Table 1.5.4), a series of application trials were made to produce a very low W/O
TAG (12%) emulsion (Table 1.5.6); a significant proportion of the aqueous phase
contained a polydextrose (Litesse® Ultra – DuPont, Denmark).
47
Table 1.5.6 12% very low fat spread formula, assembled on Gerstenberg scraped
surface pilot plant. (interesterified blend = PK4 INES, a palm stearin / lauric.
Cargill GmbH)
Ingredient Name 1 2 3 4
Water phase
Fat phase
The emulsifier system was based on highly unsaturated (DIMODAN® U/J – based
on sunflower C18:1, C18:2) and highly saturated (Grindsted® Crystallizer 100 –
containing significant levels of C22:0, abbreviated CRY100) distilled
monoglycerides (a combination of DIMODAN® MB-90 + HLEAR (fully
hardened low erucic acid rapeseed oil)). It is highly unusual to consider the use of
a MAG based on C22:0 for aqueous binding properties in a W/O emulsion where
the total fat content is either low and or very soft (Garti & Remon, 1984). One
specific reason for introducing C22:0 as previously shown, is that behenic fatty
acid seems to have a generally positive influence on crystallisation and additional
influence on emulsification (1.4; 1.5.1.3; Wassell & Young, 2007).
48
With regards to the stabilisation of the W/O low TAG emulsions and the choice of
saturated MAG, why were MAG’s based on C16:0 or C18:0 not considered? The
answer is because of partial hydrogenation (presence of trans isomer) and/ or the
MAG being palm based, with additional commercial implications (1.4).
Irrespective of these factors, palm triglycerides were used for the fat continuous
systems in this study. To aid and or secure emulsion stability, a co-surfactant,
PGPR was still part of the emulsifier system (Garti & Remon, 1984; Goubran &
Garti, 1988; Rousseau, 2000).
Samples were prepared using a G&A pilot plant, and then assessed by DSD and
Texture Analysis (refer to General Materials & Methods 2.0)
1.5.3.3 Results
Results from water droplet size distribution were not available. However, results
from texture test using Stable Micro Systems TX2 Texture Analyser revealed
interesting data; both samples 3 and 4 exhibited additional resilience (Figure
1.5.8). Both contained the MAG rich in C22:0.
Figure 1.5.8 Textural firmness at 5°C of W/O 12% very low fat spread emulsion
49
To determine if the resilience was because of the addition of the emulsifiers
increasing the total solid fat content (SFC), the SFC% using Bruker p-NMR, was
measured on three samples (Table 1.5.7), two of which were the same emulsifier
combination as used in the 12% very low fat spread formula (Table 1.5.6). A third
sample was also added to include Grindsted Crystallizer 110, abbreviated CRY110
(refer to Materials & Methods, 2.0).
Table 1.5.7 Anhydrous bulk TAG blend with emulsifier mixtures: three samples
assembled for SFC% determinations
Samples were triple determined and results presented in Table 1.5.8, then
graphically in Figure 1.5.9. The SFC% for all samples appears to be very similar.
Arguably, those samples with either CRY100 or CRY110 should, at 5°C, be
differentiated because of additional saturation. However, an emulsifier
contribution of 0.2% in either case on the total TAG / emulsifier load is thought to
be negligible (Smith et al., 2011) because a significant increase in the total SFC%
is not observed.
50
Table 1.5.8 Triple SFC% determinations for emulsifier mixtures in Table 1.5.7
Figure 1.5.9 Graph to show SFC% of MAG/TAG and PGPR mixtures in 15%
PKINES / 85% RBD RP (re: Table 1.5.7 / 1.5.8)
51
1.5.3.4 Conclusion
These interim results seem to suggest that the presence of a fully hydrogenated
TAG/MAG containing significant presence of behenic (C22:0) monoglyceride,
seems to have a technical influence on interfacial properties, causing apparent
thickening of the final emulsion. The results from SFC% showed no appreciable
differences to suggest firmness is attributed to increase in overall total SFC% in
the samples containing CRY100 or CRY110. This leads to the question - what is
the cause of the apparent increase in emulsion resilience? If the textural response
is not attributed to increased SFC%, then it can only be attributed to the emulsifier
relationship of PGPR, CRY110 with or without UJ. This needs to be investigated
further.
Exploratory pilot studies were conducted and indicated that MAG / TAG mixtures
give an enhancing effect on the crystallisation rates in several anhydrous
dispersions (1.5.1; Appendix A,B,C,D). Second, the effect of diversified fatty acid
mixtures of MAG in the presence of an aqueous interface (shown in 1.5.2) was
shown to have a positive influence, increasing the W/O emulsion stability. Third,
pilot studies showed it was possible to stabilise exceptionally low TAG W/O
emulsion structures (1.5.3) using an application of long-chain behenic based
emulsifier. This approach is not known to be reported in the literature.
Based on these pilot studies, a critical review of current issues behind finding
alternative structuring mechanisms for water-oil emulsions and for aiding structure
in TAG based systems, needed to be undertaken.
52
1.6 A Critical Review of Current Issues: Alternative Mechanisms for
Aiding Structure in TAG Based Systems
Exploratory pilot studies have shown that saturated MAG / TAG mixtures give an
enhancing effect on the crystallisation rates in TAG based anhydrous dispersions
(1.5.1; Appendix A,B,C, D). Secondly, diversified fatty acid mixtures of MAGs in
the presence of an aqueous interface (1.5.2) have a positive influence for W/O
emulsion stability. Thirdly, interim pilot studies suggested it is possible to stabilise
exceptionally low TAG W/O emulsion structures (1.5.3) using the application of a
long-chain behenic based emulsifier. This approach is not known to be reported in
the literature.
Since conducting a literature review (1.4) and series of pilot studies (1.5), a more
critical review is now necessary. This is because to progress this learning into
finding new and novel approaches to structuring, a consideration for current issues
is required. Realistically, this must therefore also encompass issues surrounding
the sourcing of alternative, natural lipid materials, their sustainability and a
consideration of the techniques used to observe and quantify structural and
interfacial events.
An edible oil and its commoditised fractions which has grown widely in use
(Wassell & Young, 2007) and status since the 1990’s, is palm oil. While palm oil
is not necessarily regarded as a low saturated fat, it does have many fractions,
some of which are more unsaturated, eg. olein fractions. However, despite being
53
inherently natural, there is controversy surrounding agricultural ethics and
sustainability of palm oil (Carmichael, 2011; Young, 2011).
In sourcing alternative natural lipid materials, China (the largest vegetable oil-
consuming country in the world) is, as part of its continued growth strategy,
developing several uncommon oil resources (Table 1.6.1). This is to minimise its
land dependence and achieve maximum yield. It is also regulatory friendly to
“Novel foods” or food components which are not traditionally used in China
(Wang, Xu, & Jiang, 2012).
Table 1.6.1 Approved novel foods related to fats and oils (source: Wang et al.,
2012)
54
Other researchers have explored different strategies including, lipid crystal
engineering (Schaink, van Malssen, Morgado-Alves, Kalnin, & van der Linden,
2007); phytosterol systems (Bot et al., 2009); organogel-based emulsion systems
(Duffy, Blonk, Beindorff, Cazade, Bot, & Duchateau, 2009) and wax technology
to form ‘lipo-colloid’ structures (Shigemi, 2006). A comprehensive discussion of
possibilities is put forward by Pernetti, van Malssen, Flöter, & Bot, (2007), who
state succinctly that finding alternatives to traditional triacylglyceride (TAG)
structuring whilst maintaining healthy properties, versatility and performance, is a
tremendous challenge. More recent and current reviews confirm the same (Rogers
2009) and have considered the influence of minor components (Smith et al., 2011)
e.g. emulsifiers, functional ingredients and their characterisation at micro and
nano-scale (Marangoni et al., 2012; Sato & Ueno, 2011; Silva, Cerqueira, and
Vicente, 2011).
Additionally, it is clear that population increase and the issues of sustainability are
indirect catalysts to finding alternate lipid materials for the function of structuring /
stabilising food products (Carmichael, 2011; Wang et al., 2012; Wassell & Young,
2007; Young, 2011). Whatever forms of approach are adopted, removal of, or at
least drastic reduction of all saturated fats is also required (Beaglehole et al., 2011;
Bradley, 2012; Gortmaker et al., 2011; Micha & Mozaffarian, 2010; Mozaffarian
et al., 2010; NICE, 2010; Siri-Tarino et al., 2010; Yamagishi et al., 2010;
Swinburn et al., 2011; van Camp et al., 2012; Vesper et al., 2012; WHO, 2011;
Merchant et al., 2008; Wassell & Young, 2007; Wassell et al., 2010a).
55
Schaink et al., 2007; Arima, Ueno, Ogawa, & Sato, 2009; Rogers, 2009; Tanaka,
Tanaka, Yamato, Ueno, & Sato, 2009; Young et al., 2008).
New techniques are necessary to observe and quantify structural properties and
events in both static and under dynamic conditions. (Arima et al., 2009; Boodhoo,
Humphrey, & Narine, 2009; Tang & Marangoni, 2006). This should help provide
stronger information that could be transposed to, for example, “real” emulsion
systems. One example is ultrasonic velocity profiling with pressure difference
(UVP-PD) instrumentation, allowing non-invasive in-line rheological
measurement. Another is synchrotron radiation microbeam small-angle XRD (SR-
μ-SAXD) technique which more precisely observes microstructures of fat crystals
(Awad, Moharram, Shaltout, Asker, & Youssef, 2012; De Graef et al., 2006;
Maleky et al., 2007; Mazzanti, Guthrie, Sirota, Marangoni, & Idziak, 2003; 2005;
McClements & Rao, 2011; Okamura, Wassell, Young, Bonwick, Smith, Almiron-
Ruig, Sato, Ueno, 2011; Pernetti et al., 2007; Prakash & Ramana, 2003; Silva et
al., 2011; Tanaka et al., 2009; Young et al., 2008).
56
1.6.4 Structure – Effects of Minor Components
Pernetti et al., (2007) suggest crystal structure modifiers are potentially interesting
not only within traditional systems, but also within novel structures and mixtures,
where a combination of these materials might provide several functions. The
effects of hydrophobic adsorption properties of additives, template behaviour and
acceleration of nucleation have been studied (Awad & Sato, 2001 & 2002;
Cerdeira, Martini, Hartel, & Herrera, 2003; Cerdeira, Pastore, Vera, Martini,
Candal, & Herrera, 2005; Cerdeira, Martini, Candal, & Herrera 2006; Sakamoto et
al., 2003; Sakamoto et al., 2004). More recently, the advanced analysis of the
behaviour of unsaturated and saturated surfactants, their polymorphism and time
temperature effects (Arima et al., 2009; Tanaka et al., 2009), has provided
information on the interactions of minor components (Vereecken, Foubert, Smith,
& Dewettinck, 2009a) and crystallisation with and without a water phase
(Vereecken, Foubert, Meeussen, Lesaffer, & Dewettinck, 2009b; Vereecken,
Meeussen, Foubert, Lesaffer, Wouters, & Dewettinck, 2009c; Vereecken et al.,
2010). However, it is important to gain a clearer understanding on the effects of
mixed surfactant systems, and how these interactions influence TAG mixtures in
W/O emulsions (Pernetti et al 2007; Rogers, 2009).
57
conditions e.g. static vs. dynamic crystallisation (Young et al., 2008). Rousseau et
al. (2005) suggest liquid emulsifiers such as PGPR compared to emulsifiers such
as saturated MAG are not as likely to influence crystallisation unless PGPR is
combined with agitation conditions, where polymorphic transformation was
retarded. They found that static crystallisation in the presence of PGPR showed no
“apparent” differences. This relationship ought to be more clearly understood,
especially within the area of high internal-phase-ratio emulsions (HIPREs), where
interfacial film strength and mechanical flexibility will probably impact on final
textural properties (Charteris, 2007; Povey et al., 2007; Pothiraj et al., 2012), and
primarily where both MAGs and PGPR are combined (Garti & Remon, 1984;
Garti, Binyamin, & Aserin, 1998; Goubran & Garti, 1988).
58
Therefore, attempts to find novel strategies for making healthier (reduced fat)
spreadable butters or to improve total saturated fatty acid contributions in low fat
spreads (W/O), may be limited by the diversity of accepted (FAO/WHO, 2011)
raw materials. As a result, to bring forth any subsequent or new improvements,
there is requirement to better understand current application of traditional
ingredients (e.g. emulsifiers), and/or to explore new possibilities with novel
ingredients to perform similar and or improved function.
1.6.5 Conclusion
The literature shows first, the necessity to move away from the addition of hard
fatty material and substitute with alternative structure forming strategies
(Dassanayake, Kodali, & Ueno, 2011; Wassell & Young, 2007). Second (related to
first theme) is the need to cause beneficial interactions through inclusion of a small
amount of additive such as an emulsifier (Smith et al., 2011).
Given the aforementioned approaches, there are two areas for consideration; not
just structure and assembly within the bulk oil, but the mechanism of structure at
interfaces e.g. W/O emulsions, and how the interface might be modified through
emulsifier initiated processes (Krog, 1975; 2001) to provide adequate structure,
strength, and flexibility via a range of mechanisms (Povey et al., 2007). One
mechanism is to employ the use of “traditional” TAG based structurants (Lupi,
Gabriele, de Cindio, Sánchez, Gallegos, 2011b; Smith at al., 2011), coupled with
novel structuring material (Duffy et al., 2009; Pernetti et al., 2007). The addition
of a co-surfactant could aid nucleation through standard thermodynamic events
(Pernetti et al., 2007; Rogers, 2009). Co-surfactants such as MAGs derived from
TAGs, which are commonly used for W/O emulsions (Garti & Remon 1984),
could fit this description, as they can enhance textural properties which consumers
are familiar with (Foster et al., 2007; Pothiraj et al., 2012).
59
Among the TAG components discussed by Rogers (2009), fatty acids, particularly
long chain fatty acids, are ignored due to the negative health implications
associated with traditional TAG based networks. However, there are still
unanswered questions regarding the use of emulsifiers synthesised from them
(Vereecken et al., 2009c & 2010). Pernetti et al. (2007) suggested that longer chain
fatty acids tended to have more structuring potential per gram of material than
shorter chains. Further investigation of how these subtle variations in the amounts
and chain length of structurants influence the structural properties of TAG based
systems is of interest (Krog & Larsson 1992). This is especially the case when
combining surfactants of differing molecular structure, e.g. monodiglyceride and
polyglycerol polyricinoleate (PGPR) – a classic combination, for low fat W/O
emulsions (Garti & Remon 1984; Garti et al., 1998; Annon, 2005).
60
In the case of a given interfacial emulsifier concentration, temperature, and shear
rate (Lupi, Gabriele, de Cindio, 2011a), the surface of water droplets will be
optimally or sub-optimally covered, especially in the context of HIPREs.
However, explaining these mechanisms could be problematic, in that laboratory
scale techniques using standard models / methods could be limited because of their
inability to provide data whilst under true dynamic conditions (Young et al.,
2008).
61
rheological mechanics of W/O emulsions (Arima et al., 2009; Awad et al., 2012;
Ishizuka et al., 2009; Ojijo et al., 2004; Pernetti et al., 2007; Povey et al., 2007;
Rogers, 2009; Schaink et al., 2007; Young et al., 2008). Pothiraj et al., (2012)
explain that low TAG W/O emulsions cannot easily be formulated to mimic butter
or similar, because of the challenges to developing melting behaviour, which is
largely governed by emulsion characteristics such as volume of the aqueous phase
and size of water droplets.
62
1.7 Aims and Objectives
1.7.1 Background
There is a need to find alternative structuring materials, since many food products,
especially fat based food dispersions and emulsions are currently still highly
dependent on the presence of saturated fats and traditional emulsifiers. However,
the problem, of removing these ‘building’ blocks, i.e. saturates and their trans
isomer from fat systems, make it difficult to produce food products which are
functionally acceptable to the consumer. The mechanics and successful removal of
saturated and trans fatty acids in foods (Wassell & Young, 2007) is a complex
problem, as is the successful structuring of low or reduced fat based foods, that
require additional reduction of total saturates (Wassell et al., 2010a; Young and
Wassell 2008). Achieving further reduction of saturated fatty acid materials for
dietary requirements (Winwood, 2011) without negative impact on structure and
functional properties, presents challenges for those food systems already regarded
as low saturated. The challenge is for food manufacturers to find innovative
solutions to structuring whilst meeting consumer expectations (Wassell & Young,
2007; Wassell et al., 2010a).
63
time process rheology (Young, Wassell, Wiklund, & Stading, 2008; Wassell,
Wiklund, Stading, Bonwick, Smith, Almiron-Roig, & Young, 2010b) and analytical X-
ray measurement (Wassell, Okamura, Young, Bonwick, Smith, Sato, & Ueno, 2012).
A pilot study (1.5) looked at the effect of MAG on crystallisation time and
diversifying their fatty acids. This included mixing C22:0 with other MAG chain
lengths and TAG mixtures. Some positive influence was found when tested in
anhydrous dispersions (Basso et al., 2010). The degree of benefit was influenced
also by the degree of saturation of the solvent (TAG mixtures). An effect was also
found on the visco-elastic properties of a very low TAG W/O emulsion by minor
addition of a C22:0 based MAG (1.5.1; 1.5.2; 1.5.3; 1.5.4).
Currently, many low cost emulsifiers tend to be based on palm TAG products,
characterised by its major fatty acids C16:0 / C18:1 and its intrinsically natural,
unmodified arrangement. Alternative TAG materials which are natural,
unmodified and provide a whole range of fatty acids may currently be difficult to
source in sufficient quantities. Hence, other naturally occurring novel resources
must be considered (Dierig & Thompson, 1993; Dierig, Thompson, & Nakayama,
1993). Moringa TAG, has a significant range of saturated fatty acids to unsaturated
fatty acids, including naturally occurring saturated fatty acids >C16:0 – C24:0
(Abdulkarim, Long, Lai, Muhammad, Ghazali, 2005; Lalas & Tsaknis, 2002).
64
In low or reduced fat systems it is known that as the oil phase is decreased, and
water phase increased, the emulsifiers usually become softer, more liquid in form
(Table 1.2.4). Furthermore, as the fat content is reduced, there is also tendency to
soften the SFC, so that both the continuous TAG phase and the emulsifier are
similar in rigidity (Garti & Remon 1984). So that, whilst utilisation of a MAG
based on e.g. liquid sunflower or similar (IV = ~105) is able to facilitate assembly
of a W/O emulsion, in W/O of approximately <41% this may not be entirely
secure. Hence, where TAG content is <41% it is common to use both PGPR and a
second emulsifier (a co-emulsifier) e.g E476 (EU max permitted use – 0.4%).
For commercial and product quality reasons, several manufacturers have requested
replacement of PGPR because of its negative effect on the sensory melting
behaviour (too stable) and to achieve clean label by removing E-numbers. Another
reason is because of handling re-melt (a given volume of product will re-circulate
through the production process), where the WO emulsion must be re-melted and
sometimes separated. PGPR is very robust at binding water.
In the context of real, low TAG based W/O emulsion systems, there is still a need
to report the affects of the interfacial behaviour of MAG mixtures with other
common emulsifiers e.g. PGPR (Garti & Remon 1984; Garti et al., 1998). While
texture analysis (1.5.3.3) revealed “apparent” increased firmness in the finished
product, it is more difficult to quantify affects in a dynamic state. This leads to
another necessity; to find a method to measure rheological change in real-time so
that rheological interactions can be reported both off-line and in-line.
65
Empirical quantification to determine how much impact a minor ingredient e.g. a
MAG based on significant quantities of saturated longer chain >C18:0, may have
on a “real”, low or reduced TAG based W/O emulsion system has not been
reported in the literature. Only the behaviour of behenic (C22:0) based emulsifiers
in O/W systems is reported (Arima et al., 2007; Awad & Sato, 2001; 2002;
Fujiwara et al., 2000; Sakamoto et al., 2003; Wassell & Young, 2007) and not in
W/O. Therefore, as part of a multidisciplinary investigation, the rheological and
interfacial properties after the addition of a >C18:0 based MAG within the bulk
and W/O emulsion needs to be characterised. Then to quantify its impact in a
dynamic state, requires both direct and indirect methods of analysis.
The intention was to investigate how structuring and texturing low saturated fat
based systems can be achieved through a multidisciplinary route (Wassell &
Young, 2007). The innovative nature of this approach intended to provide answers
to these outstanding questions:
66
3) What is revealed through investigating real emulsion systems, as opposed
exclusively to single phase model systems?
Specifically, the thesis reviews current approaches to structuring fat based foods /
emulsions and points to potential new mechanisms to modify fat crystallisation
and emulsion behaviour (Wassell & Young, 2007; Wassell et al., 2010a) and
considers new ways with which to measure these phenomena (Young et al., 2008;
Wassell et al., 2010b; 2012).
This thesis aims to show the potential of reducing total saturated fats by using
existing and novel emulsifiers, demonstrating their impact on functionality
(Wassell et al., 2012a; 2012b; 2012c; 2012d; 2012e; Bech et al., 2013).
A decrease in total fats tends to lead to increase in the water phase (in respect to
water – oil emulsions) of the emulsion system. This being the case, the thesis takes
into consideration fat mimetic properties (Wassell & Young, 2007; Wassell et al.,
2010a). The objective therefore is to investigate behaviour and influence of
emulsifiers both at the interface and in the bulk to demonstrate how their role is
critical to the ultimate mechanics of structuring fat based dispersions and
emulsions (Wassell et al., 2012a; 2012b; 2012c; 2012d; 2012e; Bech et al., 2013).
The uses of novel structuring agents are investigated within this thesis to observe
their influence on final fat texture and functionality. Within the scope of this work,
67
extensive focus is given to a novel emulsifier developed from an innovative source
of edible oil (Wassell et al., 2012a; 2012b; 2012c; 2012d; 2012e; Bech et al.,
2013). The study looks at emulsifiers that do not necessarily conform to the broad
international legislative demands, but which nonetheless, provide deeper insight
into structuring and texturing of low saturated fat based systems.
It is intended to show that innovation and novelty can be provided not only
through combining the use of emulsifiers (nucleation inducers) in real systems, but
also analysis through static and dynamic conditions. Techniques to analyse the
functionality of the given systems under static and dynamic conditions, include
controlled stress rheology, interfacial rheology, tensiometry, X-ray diffraction,
CLSM, DSC, DSD, UVP-PD.
The key aim of this multidisciplinary approach was to develop a more thorough
and complex understanding of the mechanics and development of structure within
commercially relevant, novel, low saturated fat based systems. Successful
achievement of the key aim would therefore enable development of new routes to
structuring fat based systems, either by enhancing traditional options, or switching
to new alternatives and away from traditional methods which are currently under
legislative scrutiny and lobbying pressure to be abandoned (Wassell et al., 2010a;
Smith et al., 2011; Wang et al., 2012).
68
1.7.5 Analytical Approach
1.7.5.1 UVP-PD
1.7.5.2 Tensiometry
1.7.5.3 Rheology
69
any previously observed effects (interfacial behaviour) of saturation and chain
length in anhydrous bulk and water-oil systems. The intention of this study is to
observe the rheological behaviour of selected emulsifiers through temperature,
varying fatty acid chain length, degree of saturation with / without PGPR and
novel MAG material. Any undiscovered improvement to viscoelastic behaviour
would have important commercial value for water – oil emulsions and other food
systems. This knowledge is currently unreported in the literature.
70
2.0 General Materials and Methods
2.1 Emulsifiers
Moringa oleifera oil (Code: 126089, Batch No: DE05040243, Ref: 804903823/1)
was obtained from Earth Oil Plantations Ltd. Lichfield, Staffordshire, United
Kingdom. RBD ricebran oil was obtained from Thai Edible Oil Co., Ltd.
(Bangkok, Thailand). RBD Lesquerella fendleri oil was obtained through Danisco
A/S (known as DuPont – Bioscience ApS, Denmark).
Emulsifiers and their source are shown (Tables 2.1, 2.2, 2.3, 2.4, 2.5, 2.6, 2.7, 2.9,
2.10, 2.11) and characterised (Table 2.8) by one or two specific fatty acids:
DIMODAN® HP is characterised by palmitic acid (C16:0); DIMODAN® P by
palmitic acid (C16:0) and oleic acid (C18:1c); Crystallizer 110 by behenic acid (C
22:0); DIMODAN® UJ by linoleic acid (C18:2c), and RT by oleic acid (C18:1c)
and linoleic acid (C18:2c). Table 2.1 also shows degree of saturation by iodine
value (IV).
71
Table 2.1 Emulsifier fatty acid carbon chain lengths
Table 2.2 Distilled Ricebran monoglyceride from RBD ricebran oil (Table 2.10)
72
Table 2.3 Fatty acid composition of Moringa oleifera TAG and Moringa
monoglyceride
Analysis (%) Moringa TAG / mono Analysis (%) Moringa TAG / mono
C14 0.1 C19 0.1
C15 <0.1 C20 3.4
C16 5.8 C20:1 2.2
C16:1 1.8 C22 5.8
C17 0.2 C22:1 0.1
C18 5.4 C24 1.0
C18:1 73.0 C26 -
C18:2 0.7 Unknown 0.2
C18:3 0.2
73
Table 2.6 Analyses of Lesquerella monoglyceride
Table 2.7 Table of Fatty acid profile and saturation (IV) of Lesquerella fendleri
TAG
74
Table 2.8 Emulsifiers are characterised by source from which they were produced
PGPR 90# C16 C17 C18 C18:1 C18:2 C20 C18:1 C18:1 unknowns
Analysis % Ricinoleic oleic
acid acid
1.3 0.0 1.3 3.6 5.1 0.0 86.1 0.0 2.4
#
PGPR Palsgaard 4150 also known as PGPR 90 Plus
75
Table 2.10 RBD Ricebran oil (Thai Edible Oil Co., LTD (Bangkok, Thailand)
Table 2.11 Moringa oleifera (Earth Oil Plantations Ltd. Lichfield, Staffs., UK)
Testsa Specification
Physical State MOBILE LIQUID -MAY SOLIDIFY
Colour PALE YELLOW -YELLOW/ORANGE
Peroxide Value (Meq) Maximum 6
Free Fatty Acid content (%) Maximum 3
Fatty Acid profile PASS
Palmitic Acid (16:0) 2.0 -10.0
Palmitoleic Acid (16:1) 1.0 -5.0
Stearic Acid (18:0) 2.0 -7.5
Oleic Acid (18:1) 65.0 -85.0
Linoleic Acid (18:2) Maximum 1
Arachidic Acid (20:0) 2.0 -5.0
Gadoleic Acid (20:1) Maximum 4
a = Test parameters not specified within material specification
76
2.3 Additional Distillations of Natural Moringa oleifera
Table 2.12 Other partial and full distillations of natural Moringa monoglyceride
with mono, di- and triglyceride specifications
77
2.4 Synthetic Monoglycerides
Monoglycerides described in Table 2.1 and Table 2.8 are mixed to provide a
similar fatty acid profile to natural distilled Moringa monoglycerides (Table 2.3),
and similar GC (Table 2.5) are shown in Table 2.13. Resultant compositions are
shown in Table 2.14.
Analysis (%) SM 90 SM 60 SM 80
GL 0.16 0.24 0.20
Digl 0.14 0.1 0.18
FFA 0.30 0.40 0.40
Mono 96.50 64.56 82.87
Di 2.64 29.02 15.28
Tri 0.22 2.59 1.10
In summary the natural samples, align approximately with synthetic samples and
are as follows:-
Natural Synthetic
2472/191 = SM 90
2559/102 = SM 60
2559/104 = SM 80
78
The fatty acid compositions of synthetic Moringa monoglycerides are shown in
Table 2.15.
79
2.5 High and Low Temperature Distillation of Moringa oleifera TAG
High temperature distillation of the Moringa oleifera TAG (Table 2.11) enables
preservation of the whole fatty acid distribution being carried over into the
resultant monoglyceride. The effects of producing fully distilled Moringa MAG at
a low (185°C) and high (210°C) temperature are shown in Table 2.16.
Table 2.16 Effect of high / low temperature distillation on Moringa oleifera TAG
80
2.6 Triacylglycerol (TAG) for Water-in-Oil (W/O) Emulsions and Anhydrous
Dispersions
The multidisciplinary approach used throughout this study, led to the use of
several base TAGs. These are listed and described in Table 2.17
Table 2.17 Major TAGs for W/O emulsions and anhydrous dispersions
TAG
ChocoFill PK4 Palmotex Palmotex Palmovit Kokowar
°C l BR60a INESb Bc 98d 200e 31f COLZAOg,k
10 72 - 80 82 10
20 45 - 48 49 - 55 64 1 60
25 31 29
30 10 - 14 23-28 42 9
35 1-2 27 4
40 3–4
Slip melth °C 33 37 - 42 36 53 - 57 31
Iodine Valuei 22 - 32 35 min 63 110 - 121
SAT %j 74 74 39 7
MONO %j 22 31 47 62
POLY %j 4 5 14 30
trans %j <1 max 2 <1 max 1 <4 max 1
a = Interesterified palm stearin / lauric (previous names include: Akomic or Akocrem M), Aarhus Karlshamn
(AAK), Denmark.
b = Interesterified palm stearin / lauric. Cargill GmbH., Hamburg, Germany.
c = RBD Palm oil. AarhusKarlshamn (AAK), Denmark.
d = Palm Stearin. AarhusKarlshamn (AAK), Denmark.
e = RBD Palm Olein. ADM Hamburg AG, Hamburg.
f = hydrogenated and lauric oil based on coconut oil. Aarhus Karlshamn (AAK), Denmark.
g = RBD Rapeseed oil. AarhusKarlshamn (AAK), Denmark.
h = AOCS Cc 3-25
i = IUPAC 2.205
j = IUPAC 2.304
k = Cloudpoint (°C) -16, ASTM D97 SS-EN (23015)m
l = Solid Fat Content (%), IUPAC 2.150a
RBD Peanut oil (Lot no: J0112KA) supplied by Columbus Foods Company, inc.
Des Plaines, IL., USA), with following specifications:-
Colour (Lovibond): 0.5 (Cc13b-45); Flavour: Bland; Free Fatty Acid (%): 0.025 (Ca5a-40);
Peroxide Value meq/kg: 0.25 (Cd8b); Iodine Value: 90.3 (Cd1-25); Cold Test: 5.5 Clear &
Brilliant (Cc11-53); Additives: none.
81
2.7 Other and Minor Ingredients for W/O Emulsions
2.7.1 Hydrocolloid
2.7.2 Flavours
Butter flavour (water phase) 050001 T03007, and Butter flavour (oil phase)
050001 T04184, (Firmenich. Denmark).
2.7.3 Antioxidant
2.7.5 Salt:
2.7.6 Antimicrobial
82
2.7.8 EDTA:
Besides the chemical properties of the fat blend, the process conditions have a
major impact on the finished product. Emulsions are subjected to applied cooling
temperatures and shear, whilst passing through a series of scraped surface units,
often referred as A-units, and pin-workers, often referred as B-units. Depending on
the pump speed and back-pressure, the residence time of the product can be
controlled. These process parameters will strongly determine the properties of the
margarine / spread and are very important to obtain margarine with a good quality
and storage stability.
83
2.8.1 Equipment and Processing
35%, 40% and 60% water-in-oil (W/O) emulsions and continuous bulk oil
compositions were processed (Table 2.19) on a cooled scraped surface heat
exchanger pilot plant (Figure 2.1), known as a Gerstenberg & Schröder
(previously named Gerstenberg & Agger) : 3-tube pilot perfector 3x57 (Serial No:
11803). Design pressure 80 bar – 120bar. Refrigerant: Ammonia (NH3), from SPX
Flow Technology Copenhagen A/S, Vibeholmsvej 22, Brøndby, Denmark.
Specifications are as follows: Scraped surface heat exchanger (SSHE):
Diameter/length 57mm/430mm. Annular space 6.5 mm. Volume 0.42ltr. Cooling
surface area /tube: 0.0567m2. RPM each tube: 200-1000. Cooling (max) -25°C.
High pressure 3-piston pump (Model: P-3-15/35-Labo). Capacity: 20-210 l/h. Max
pressure: 160 bar. Centrifugal Pump type: CS 25 (serial no: 48258) with a 115mm
inbuilt propeller, 0.55KW Bi Polar motor 0.75 hp motor (2900rpm). AISI316
stainless steel heated pump housing (Model T82)
The Processed 35%, 40% and 60% W/O emulsions and or bulk oils were
assembled as per the following general procedure:
All dry ingredients were weighed and mixed and slowly added to the water phase,
while stirring intensively for 4 minutes. The water phase was then allowed to cool
to 40°C or adjusted accordingly (see Table 2.19). The pH of the water phase was
adjusted with standard citric acid (Citric Acid Anhydrous, (Food Grade) CAS No.
77-92-9)).
84
2.8.1.2 Fat Phase
The emulsifiers, beta carotene (2% solution) and the selected oils and fats were
weighed into the same container and then heated to 80°C. Periodically, the fat
phase was stirred to ensure homogeneous melting.
The fat phase was then cooled to 40°C or adjusted accordingly (see Table 2.19),
until required for processing. In the case of emulsions, the water phase was
carefully added to the fat phase while stirring intensively.
Table 2.19 Processing parameter using Gerstenberg & Schröder 3-tube lab scale
scraped surface heat exchanger units
85
The pilot scale scraped surface heat exchanger units are shown in Figure 2.1
Figure 2.1 A Gerstenberg & Schröder pilot scale: 3-tube Pilot Perfector 3x57
86
2.9 Confocal Laser Scanning Microscopy (CLSM)
In the confocal microscope, all structures not in focus are suppressed at image
formation. This is obtained by a pair of pinholes which limit the specimen’s focal
plane to a confined volume. Relatively thick specimens can be imaged in
successive volumes by acquiring a series of sections along the optical axis of the
microscope. This results in image stacks that can be reconstructed into images
showing the 3D structure of the sample. Structural information can thus be
obtained with minimum risk of integrity to the sample under examination.
The CLSM system in this study enabled simultaneous detection of signals from
four photo multipliers, including one for transmitted light which allowed use of
ordinary light microscopy. There were 3 detection channels for simultaneous
staining with fluorescent probes and 6 laser lines available for inducing maximum
available fluorescence. The protein was dyed using FITC, which turns green. The
fat was dyed using Nile Red, which turns red. Both of these are fluorescent colours,
which are dissolvable in acetone.
2.9.1 Method
The water-in-oil emulsions were placed and pushed lightly upon a 1.5 x 1 x 0.5 cm
dyed cover glass (using 15µL FITC, Flourescein-5-Isothiocyanate, dissolved in
acetone, and 15µL Nile Red, dissolved in acetone), then placed in refrigerator at
87
5°C for dying for a minimum of 30 minutes. Final analysis occurred at ambient
approximately 21°C.
2.9.1.1 Reagents
FITC (Flourescein-5-Isothiocyanate)
Acridin Orange.
2.9.1.3 Calculations
Four images at two magnifications were taken using a 40X and 100X object lenses
respectively. CLSM images were then run through a MATLAB program R2009b
(MathWorks, Matrix House, Cambridge Business Park, Cambridge, UK) to
remove the blue reflection channel, and to ‘colour combine’ the images. The
images are then reproduced (scaled) to 375 x 375 µm and 188 x 188µm.
88
2.10 Differential Scanning Calorimetry (DSC)
2.11.1 GC Analysis
2.11.2 Monoglyceride Analysis Method: Fatty Acid Methyl Ester (FAME) and
Glycerols (Tri, Di, Mono)
A Perkin Elmer model Autosystem XL gas chromatograph was used fitted with a
Perkin Elmer AS-8300 automatic injector, programmable temperature vaporiser
(PTV), flame ionisation detector and TotalChrom software for data analysis
(PerkinElmer Instruments, Massachusetts, USA). The chromatographic column
used was a WCOT fused silica 12.5m x 0.25mm ID x 0.1µ film thickness, with
coating: 5% phenyl methyl silicone (figures are length of column x diameter x
thickness of the coating).
This following method was used for determination of fatty acid composition in all
monoglycerides and fat containing samples.
Carrier gas: Helium, 10-12 psi inlet pressure. Automatic injector programming:
injection time 1 second, rinse cycle 10-20 seconds. 0.1-0.2 µl of sample injected if
on-column injection is used, and 1-2 µl of sample injected if split injection is used.
Injector: 0.01 minutes after injection, the temperature of the PTV is increased to
400°C. The Oven was programmed from 60-80°C to 200°C at 15-20°C/min., then
to 360°C at 10-TC/min and held at 360°C for 7-12 minutes. The Detector
temperature was 385°C. Detector gases: H2 35 ml/min., and air 430 ml/min.
Data processing: peak data was collected and computed using a Perkin Elmer
mini-computer based integrating system.
90
2.11.3 Preparation of Methyl Esters for GC Determination
91
2.12 Polarised Light Microscopy (PLM)
2.12.1 Method
Several analyses of W/O emulsions and continuous bulk oil phase systems were
observed using an Olympus BX60 optical microscope (Serial no: 6M02546), fitted
with polarised filter (Olympus Optical Co. GmbH. Hamburg, Germany). The
desired amount of sample (~40 mg) is placed on a carrier glass slide which has
been pre-cooled or preheated to ~5°C. A cover slip was then placed parallel to the
plane of the carrier slide and centred on the drop of sample to ensure uniformity
and desirability of sample thickness. The micrograph of the crystal was taken at
40x and 200x magnification unless otherwise indicated. A number of images were
acquired each representing a typical field. Scale bars were added as appropriate.
92
2.13 Rheology
The IRS allows measurements of interfaces covered with surface active molecules.
The raw data is produced by well known rheological standard tests and the
interfacial properties can then be calculated by analysing the raw data.
Interfacial Rheology and Bulk Rheology, was measured using Physica MCR 301
(Figure 2.3a) and data driven from RHEOPLUS/32 V3.21 software (Anton Paar
GmbH, Germany). Oscillatory interfacial method was used to measure interfacial
properties of interfaces covered with surface active molecules. The IRS bicone
system is shown in Figure 2.3b & c. Geometries used for measurements were as
follows: Interfacial Rheology: Bicone - BIC68-5 (Figure 2.4a), and Bulk rheology:
Bob/cup - CC27 (Figure 2.4b).
Oil phase: Emulsifiers were weighed for rheology measurements at 0.2% w/w
(unless otherwise indicated) and the RBD sunflower oil balanced to 100%. The
preparation is heated to 10°C above melting point of emulsifier, and held for 1
hour, then cooled to ambient temperature and deaerated (~12hrs).
A very low strain was chosen to make sure that at every stage in the crystallisation
process, measurements were performed within the linear viscoelastic region so as
to avoid disturbing the crystallisation process. The low frequency was preferred in
order to minimise influence of crystallisation as little as possible. Rheograms were
obtained by plotting dynamic change in the elastic modulus (G’) and viscous
93
modulus (G’’) of the crystallising sample as a function of the crystallisation
temperature (non-isothermal).
Figure 2.3 (a) Anton Paar Physica MCR Interfacial Rheology System (IRS) and (b
/ c) Bicone measuring system
94
Figure 2.4 Geometries used for measurements: (a) Interfacial Rheology: Bicone -
BIC68-5 and (b) Bulk rheology: Bob/cup - CC27
Bulk rheology measurements were used to provide real constants for upper fluid
density and viscosity values for interfacial analyses (Appendix I).
95
2.13.2 Haake Controlled Stress Rotational Rheometer
A controlled stress rheometer (Haake RS 150, from Fisher Scientific Ltd, Bishop
Meadow Road, Loughborough, Leicestershire, UK), was set up to run from 0.001 -
60Pa of stress range, and it was programmed to measure over a period 300s, in a
logarithmic distribution. Geometries were used as follows: Bob Diameter =
20.000mm. Measuring gap = 0.850mm. Cup Diameter = 21.700mm. Sample
volume = 8.20cm3. The Haake internally calculates the stress range and time range.
Software: RheoWin version 3.610004 (Thermo Fischer group, UK) was used for
computation of readings at the stress corresponding to an even log distribution
over stress and time range requested.
Investigation of bulk oil blends subjected to the effects of controlled cooling rate
while under shear were analysed using a shear stress controlled rotational
rheometer, Rheometrics SR 5 (proRheo, Germany), operating in simulated rate
control mode. Target shear rate of 10 s-1. Crystal history was removed by melting
and holding at 90°C for 15 minutes before loading onto the rheometer. A parallel
plate geometry (40mm diameter top plate. Gap = 1mm) with thermoelectric
cooling plate using Peltier cooling was used. The temperature ramp was 70°C to
25°C at either 1°C/min, 10°C/min, 30°C/min, was used. A 2 minute delay without
shear at 70°C prior to thermo-cooling was also used.
The fat blend tested in all cases comprised of a base of 70% palm stearine (35 IV)
and 30% palm olein (56 IV), to which the emulsifiers GRINDSTED® Crystallizer
110, GRINDSTED® PGPR 90, and Monoglycerides of Moringa were added at
1%, 0.5% and 1% respectively (2.1, 2.2, 2.3, 2.6)
96
2.14 SFC Determination
Following the AOCS Official Method Cd 16b-93 (Firestone, 2004), the SFC of the
samples was determined on a PC120 pulsed nuclear magnetic resonance (pNMR)
spectrometer (Bruker AXS GmbH. Karlsruhe, Germany). The sample was placed
in the NMR tube and successively melted at 80°C for 30 min, tempered at 0°C for
90 min, and then kept at the desired temperature (10°C, 20°C, 25°C, 30°C, 35°C
40°C) for 30 min before measurement was recorded. Triplicate measurements
were obtained.
The SFC of the separated fat was measured at isothermal crystallisation conditions
using the following procedures: samples were successively filled into the glass
NMR tube (each~3 g), heated to 80°C and then maintained for 60 minutes, and
placed into a refrigerated bath circulator set at ambient 20°C. SFC readings were
obtained at 1 minute intervals for 20 minutes for each RoC test. The crystallisation
curves (SFC vs. time) were produced under isothermal conditions. Triplicate
measurements were obtained (Appendix B & C).
97
and is also known as the Wilhelmy plate method. This method is ideal for
measuring solutions containing surfactants and observing change of surface
tension over time without disturbing the liquid surface.
Figure 2.5 Principle action of the Wilhelmy plate method (Source: Danisco
Physical Food Science, Brabrand, Denmark).
98
The Wilhelmy plate consists of a thin plate usually on the order of a few square
millimetres in area. The plate is often made from glass or platinum which may be
roughened to ensure complete wetting. The plate is cleaned thoroughly and
attached to a scale or balance via a thin metal wire. The force on the plate due to
wetting is measured via a tensiometer or microbalance and used to calculate the
surface tension (γ) using the Wilhelmy equation:
Eq. (2.1)
contact angle between the liquid phase and the plate. In practice the contact angle
is rarely measured, instead either literature values are used, or complete wetting
(θ=0) is assumed (Holmberg 2002).
2.16.1.1 Solvent
Refined, bleached and deodorised (RBD) sunflower oil, iodine value 127, was
obtained from AAK (Aarhus, Denmark). Purification was then carried out using
the following procedure: Mix 30g of Fluorisil PR60/100 mesh (Sigma-Aldrich
Denmark A/S ) with 500g Sunflower Oil in a vessel. The mixture was stirred for
60 min at 80°C, and protected from UV light. After cooling over 12hrs, the
sunflower oil was passed slowly at room temperature through a glass column,
containing filter paper (glass fibre GA55, 47 mm) into 800ml UV light protected
beaker. This procedure results in the sunflower oil having an interfacial tension at
20°C of 28-30mN/m (oil – water)
99
2.16.1.2 Preparation of Samples
Oil phase: Emulsifiers were weighed for tensiometer measurements at 0.02% w/w
(unless otherwise indicated) and the RBD sunflower oil balanced to 100%. The
preparation was heated to 10°C above the melting point of emulsifier, and held for
1 hour, then cooled to ambient temperature and deaerated (~12hrs).
Measurements were started at 50°C after preheating the oil phase and the water
phase to 50°C separately. Prior to commencing with a temperature sweep, the
interfacial tension was maintained at 50°C for 5 minutes to achieve equilibrium
between the oil and water phases. Then the temperature was decreased to 5°C at
0.3°C/min and kept at 5°C for 5 minutes.
100
Figure 2.6 Digital-Tensiometer K10ST from Krüss, Germany
Texture analysis was measured using a Stable Micro Systems Texture analyser:
TA-XT2i and Software Texture expert Exceed version 2.64, from Stable Micro
Systems Ltd., Godalming, Surrey, United Kingdom. In all measurements the probe
was: SMS P/0.5 (½ diameter cylinder: Ebonite) and samples were prepared
directly from the scrapped surface cooling plant (Gerstenberg & Schroder pilot
plant), described in the methods (2.8). All samples were measured at 5°C and force
was according to standard compression test at 1mm/s to depth of 15mm. The
maximum penetration force was recorded. Hardness is reported as the maximum
penetration force (g) based on triplicate measurements. Test settings are shown in
Appendix J.
101
2.18 Droplet Size Distribution (DSD) in Low-Fat Spread
The smaller the water droplet size, the less attractive the environment for the
micro-organisms since less nutrients are available to them. A small water droplet
size (average droplet range from 5 to 50 microns) increases the shelf life of the
product, and this is aided by the side function of emulsifiers which are able to
influence texture and stability.
When the droplet size is less than 10 microns, this environment will be more
restrictive to microbiological growth (Charteris, 2007; Delamarre & Batt, 1999).
In reality, good manufacturing practice (GMP) will come into play, because the
102
margarine and spreads (water-in-oil) industry is generally regarded as low risk,
sometimes larger size droplets are found because of the acceptable trade-off to
attain a required flavour release (Young & Wassell, 2008).
2.18.1 Method
Pulsed NMR analysis using a pulsed gradient unit Bruker Minispec mq 20,
20MHz low field pulsed pNMR Analyzer, Magnet unit ND2172, equipped with a
Pulsed Gradient Unit 1059. High / low temperature probe head assembly mq-
PA231 (-120°C - +200°C). Software: SSL, system status logging. Pulsed gradient
system for 10mm tubes (10 x 180 x 0.6mm = diameter x length x thickness). Mq-
SOFT EDMs Oil droplets / Water droplets and Diffusio. Bruker gas tempering unit
for high and low temperature analysis: mq-BVT3000c (for minispec probe PA231).
Measurements are performed at 20°C and field gradients of 2.0 T/m (Telsa per
meter) or higher.
A Hahn spin echo experiment with field gradient pulses involves calculating the
reduction in spin echo amplitude compared with the Hahn spin echo amplitude
without field gradient pulses (R).
103
performed at 5°C and with 8 R values. Log-normal particle size distribution is
typically seen in w/o emulsions and is used in the mathematical calculation of
droplet size distribution. Results are given as volume and number size distribution
(Table 2.20) and derived from a log-scale using values of standardised normal
distribution (Table 2.21).
Table 2.20 Droplet size distribution results presented as volume and number size
distribution
104
2.19 Ultrasonic Velocity Profiling with Pressure Difference (UVP-PD)
Materials and Methods
The fats described earlier (Table 2.17) were blended forming the following:
25% Akomic (non trans fat filling based on non-hydrogenated partly lauric
vegetable oils and fats; Aarhus Karlshamn, Karlshamn, Sweden) and 75%
rapeseed oil – the control.
25% Akomic, 74% rapeseed oil and 1% GRINDSTED® Crystallizer 110 (distilled
monoglyceride based on behenic acid; Danisco / DuPont, Grindsted, Denmark).
Fats used for off-line (Haake) rheological measurements (2.13) shown in Figure
3.3 (3.0), were Kokowar and Akocrem (see footnote in Table 2.17)
Fat blends used for the SFC measurements (2.19.4; 2.19.5) consisted of 30% of
palm stearin, added to 70% rapeseed oil before being processed (2.19.1).
At time of writing the Akomic (Table 2.17) has been renamed to Chocofill BR 60
(Aarhus Karlshamn, Karlshamn, Sweden).
105
pilot plant via quick valve expansion couplings such that a smooth transition to the
UVP-PD tube diameter of 22.5 mm was achieved. The sample then flowed
through the UVP-PD equipment and exited directly into 150ml plastic cups or
collected and re-melted before sending the fat blend on a repeat process. The
measurement apparatus was attached directly at and immediately after the pin
worker, the area where the greatest heat of crystallisation occurs.
The fat blends were independently introduced to the pilot plant at 80°C, and
thermodynamic drive constants: ammonia -30°C, Flow rate at high-pressure pump
= 70 kg/h. Pressure on transducers <4 bars. Shaft rotation speed perfector 1 = 1000
r.p.m. Pin rotor =302 r.p.m. These conditions achieved an approximate exit
temperature in the order of 18°C after ~2 min flow time.
The UVP-PD system and method for in-line rheometry used here are well
described in the literature (Wiklund, 2007; Wiklund, Shahram, & Stading, 2007;
Wiklund & Stading, 2008). However, the upgraded system including new
hardware and new transducers used in this work has not been described in detail
before (Young et al., 2008; Wassell et al., 2010b).
The UVP-PD testing section was designed and manufactured by SIK – The
Swedish Institute for Food and Biotechnology, Goteborg, Sweden SIK, Sweden
and AB WI-KA Mekaniska Verkstad, Sweden and connected to the Gertenberg &
Schröder pilot plant (refer to 2.8.1) using a series of quick valve expansion
couplings, facilitating rapid movement of UVP-PD test sections between different
measuring points. Figure 2.12 shows the pilot plant and testing section including
the by-pass loop and the flow adapter cell for housing the ultrasound transducers,
located towards the exit point, prior to sampling.
106
Figure 2.7 Schematic image of pilot scale scraped surface heat exchanger unit and
pin rotor units with flow-cell and by-pass loop measuring apparatus with
differential pressure gauge and ultrasound transducers.
The testing section comprised a flow adapter cell for housing a pair of custom
made 4 MHz high-temperature ultrasound transducers (TR0405LH-X., Signal-
Processing SA., Savigny, Switzerland). The high temperature transducers, at the
time of writing are a new development and part of confidential contract work,
allowing measurements directly from the transducer front and thus to record more
or less zero velocity at the wall. As a result, higher precision in determining the
wall positions were achieved compared to previous work (Young et al 2008).
The transducers were mounted opposite each other at a fixed angle of inclination
(20°), which allowed simultaneous non-invasive measurement of the flow velocity
profiles and monitoring of the acoustic properties. The active diameter of the piezo
107
was 5 mm and the transducers were in contact with the liquid but were pulled back
6.9 mm to avoid measurements within the near field region where the acoustic
pressure field is irregular (Young et al 2008). The testing section also contained a
differential pressure sensor, (ABB 265DS; ETP80, ABB Automation Technology
Products AB, Sollentuna, Sweden), and the pressure drop was measured over a
distance of 0.68m (shown in Figure 2.7).
The pulser/receiver and the differential pressure sensor were connected to a master
PC via Ethernet and a high-speed DAQ card (National Instruments Sweden AB).
UVP data acquisition was implemented with an ActiveX library (Met-Flow SA,
Switzerland). All data acquisition, data processing and analysis were made using
novel software, RheoFlowTM, developed at SIK, Sweden for in-line UVP-PD-based
rheological measurements.
The UVP-PD in-line rheometer system used in this work only required initial
calibration of the transducer angle. No other calibration was needed since only
physical parameters i.e. velocity, pressure and velocity of sound are measured
108
directly. The emitting frequency of the ultrasound was set to 4 MHz, 150V with 2-
4 cycles per pulse and 512 repetitions resulting therefore in a sampling rate of
300ms per demodulated echo amplitude (DMEA) data set. Short bursts of
ultrasonic pulses were emitted and then echoed back towards the transducer by
reflective surfaces such as particles moving with the flowing suspension. The local
velocities are obtained by detecting the Doppler shift frequency of the reflected
ultrasound and the local position of the particles by simultaneous measurements of
the time delay between emission and echo reception. The power density spectra of
one set of profiles, typically 25-40 were determined using Fast Fourier Transform
(FFT) algorithm and averaged. The Doppler shift frequency (flow velocity) was
then determined from the geometrical mean of the spectra above an offset. The
total measurement time per recorded and processed velocity profile was 450ms.
Thus, radial velocity profiles can be continuously obtained in real-time both in the
direction of the flow and counter to the direction of the flow, and the volumetric
flow rate were obtained by integration of the velocity profiles.
The shear rate distribution could be determined directly from the velocity gradient
of the acquired velocity profiles. Shear viscosities and rheological model
parameters presented here were determined from a non-linear fit to the integrated
form of the power-law model.
R∆P
1
n
R r (1+ 1n )
Integrated form: v(r ) = ⋅ ⋅ 1 − Eq. (2.3)
2 LK (1+ 1
n
) R
Where τ and γ are the shear stress and shear rate, K is the consistency index and
n is the flow exponent. In the integrated form of the model r and R are the inner
and outer radius, ∆P is the pressure drop and L is the length of pipe section used
for ∆P . Details on the UVP-PD system and method for in-line rheometry used in
this work have been published in the literature (Young et al., 2008; Wassell et al.,
2010b).
109
2.19.4 Solid Fat Content (SFC) Measurement – p-NMR
The solid fat content was determined off line using the established p-NMR
technique as described by IUPAC (1987) and the calculation principle outlined by
van Putte and van den Enden (1974). Also refer to 2.14
A method for determining the solid fat content in-line using an ultrasonic velocity
technique has been improved. The method is based on the pulse transmission
technique described e.g. by McClements and Povey (1987; 1988) where an
equation derived by Urick (1947) is used to convert the ultrasonic velocity in a
sample into SFCs for many oil/fat mixtures. The velocity of sound in a two-
component suspension of particles or emulsion of droplets depends on the mean
density and mean compressibility and is given by the Urick (1947) equation:
1
v= Eq. (2.4)
κρ
Here φ is the volume fraction of the dispersed phase, ρ is the density, κ is the
adiabatic compressibility, ν is the velocity of sound and the subscripts 1 and 2
represent the continuous and dispersed phases respectively. Equation (2.5) can be
rewritten as a cubic equation in terms of φ . For oil- and fat based systems an
analytical solution exists:
(
φ = − B − (B 2 − 4 AC )
1/ 2
)/ 2 A Eq.(2.7)
Where,
110
A = ν 12 (1 − ρ1 / ρ 2 ) + ν 22 (1 − ρ 2 / ρ1 ) ,
B = ν 22 (ρ 2 / ρ1 − 2) + ν 12 ρ1 / ρ 2 ,
(
C = ν 22 1 −ν 12 /ν total
2
)
The SFC is then calculated by converting the volume fraction φ , into a mass
percentage
The velocities in, and densities in the solid and liquid phases are required to
determine SFC from Eq. (2.8). The densities were measured using a Densito-30PX
instrument (Mettler-Toledo AB, Stockholm, Sweden). The sound velocity in palm
stearin at 10°C was determined to 1548 ±5 m/s and its density 899 ±1 kg/m3. The
density in rapeseed oil was determined to 916 ±1 kg/m3. The sound velocity in
rapeseed oil was determined as function of temperature using a pulsed method to c
= 1536.4-3.487xT m/s with a correlation coefficient of 0.997. This procedure was
needed to allow for temperature compensation of the measured sound velocities. A
significant limitation of the ultrasonic technique is that vacuole formation may
occur in the sample during cooling, which practically restricts SFCs to less than
about 40% according to McClements and Povey (1988).
111
2.20 Synchrotron Radiation X-ray Diffraction (SR-XRD) using Microbeam
small angle X-ray Diffraction, (SR-μ-SAXD)
This study encompassed multiple analyses using DSC, PLM, macrobeam and
microbeam synchrotron radiation X-ray diffraction (SR-XRD), experiments on fat
crystals in the continuous phase of W/O emulsions (35 wt. % fat).
2.20.1 Method
112
Table 2.22 Sample recipes for 35% W/O emulsions
113
Table 2.23 Properties of fats used for W/O emulsions
e = AOCS Cc 3-25
f = IUPAC 2.205
g = IUPAC 2.304
114
2.20.2 Macrobeam X-ray Diffraction (XRD) Measurements
The basic principle of SR-µ-SAXD has been reported elsewhere (Ueno, Nishida,
Sato, 2008; Shinohara et al 2008; Arima et al 2009; Tanaka et al 2009; Wassell et
al., 2012). By scanning the X-ray microbeam on a thin section of the sample two
dimensions with steps on the order of the beam size, and by collecting each two-
dimensional (2D) XRD pattern with a 2D X-ray sensitive area detector, the
polymorphic structure can be assessed by measuring the long spacing value, which
is calculated by the diffraction angle (2θ) extension (Fig. 2.10a). In addition, the
lamellar plane direction (Fig. 2.10b) of the fat crystal can be assessed by
measuring the azimuthal angle (χ) extension pattern at a fixed 2θ position (Fig.
2.10c).
115
Figure 2.8 Data analyses of SR-μ-SAXD patterns. (a) 2D SR-μ-SAXD pattern. (b)
Lamellar direction in a fat crystal. (c) χ extension pattern
When all the fat crystals are highly oriented, two sharp 2D diffraction peaks (arc
peaks) should appear because of the preferred orientation of crystals. The average
direction of the lamellar planes of the fat crystals is directed normal to the
direction connecting the two arc peaks. For example, in Figure 2.8(c), sharp arc
peaks appear at χ =110° and χ = 290°, which are superimposed by a 180° rotation.
This set of two peaks (twin peaks), correspond to the symmetric diffraction peaks
from the crystals whose lamellar planes are aligned along the same direction at χ
= 200° (arrow in Figure 2.8(c)).
The degree of orientation of the lamellar planes of TAG crystals can be evaluated
by calculating the half width of χ value (∆χ). A smaller ∆χ yields a higher degree
of orientation of the lamellar planes (Arima et al., 2009).
116
The μ-SAXD measurement was performed at BL-4A of the Photon Factory, the
synchrotron radiation facility of the High-Energy Accelerator Research
Organization (KEK) in Tsukuba, Japan. Details of the μ-SAXD method have been
reported previously (Riekel, Burghammer, & Muller, 2000).
The X-ray microbeam wavelength was 0.11nm, and the beam area was 5 x 5μm2.
The emulsion sample was sealed in a 50μm-thick cell made of mica covered with
polyethylene terephthalate (PET) film, and set on a temperature controlled stage
(Figure 2.9). The sample was thermally treated using a Linkam (Linkam, UK) as
follows. First, the sample was kept at 60°C for 5min. The temperature was then
reduced to 5°C at a rate of 2°C/min and kept at 5°C during the μ-SAXD
measurement. An image of the Linkam temperature control stage unit is shown in
Figure 2.10.
Figure 2.9 Schematic showing cold stage temperature control for the W/O
emulsion preparation (Ueno et al. 2008)
It was necessary to decrease the distance between the sample and the 2D detector
to 30cm so that both the small- and wide-angle diffraction patterns could be
imaged on a 2D detector with a 6in x 6in (15.24 x 15.24cm) area. Keeping the
microbeam position fixed, the measured sample was moved by an x-y-z stepping
motor for observation by an optical microscope (magnification ×200). The sample
was moved automatically within a 2D plane in 5μm steps.
117
Figure 2.10 Image of Linkam temperature control stage unit
Water droplet distribution and fat crystal morphology were observed using a
CX31-P POM (Olympus Co., Tokyo, Japan) with a DP 12 digital camera
(Olympus). The POM was set under the crossed Nicols condition. Samples were
set in a Linkam (TU-600PM, Cambridge, UK) furnace set to a temperature of 5°C
and placed on the sample stage of the POM.
118
3.0 A Study on the Crystallisation of TAG Based Systems using UVP-PD
The physical properties of Triacylglycerol (TAG) based systems have mainly been
studied and performed under isothermal conditions where observations of crystal
behaviour can be quite different from those observed whilst in or from a dynamic,
non-isothermal environment (Wassell et al., 2010a). Ultrasonic velocity profiling
with pressure difference (UVP-PD), offers the chance to probe the mechanics of
TAG blend physics under real, dynamic conditions.
There are many potential factors influencing crystal nucleation and many have
been reviewed (Povey et al., 2007). There is a need for methods to analyse and
understand the crystallisation processes occurring under dynamic conditions. A
range of different analytical techniques have been reported to study TAG
crystallisation processes, including rheology, DSC, pNMR, ultrasonic
spectroscopy (De Graef et al., 2006; Janssen & MacGibbon, 2007; Maleky et al.,
2007; Martini et al., 2006; Wassell et al., 2010a), but so far all these techniques
have been constrained to static, and semi-static conditions. None of these analyses
measure while the material is potentially in a dynamic – supercooled condition,
where crystallisation occurs very quickly and in bulk (Garbolino et al., 2005).
Ultrasound is regarded as the most sensitive (Povey & Challis 2006).
119
3.2 Effect of Cooling
Energy Structure
0°C
15°C ∆G*
20°C
Liquid
∆G
Solid
Cooling Temperature
Figure 3.1 Schematic diagrams of the energy barrier to crystallisation and the
relative amount of structure formed as a function of cooling temperature.
The effect cooling on the structure forming ability of a system is shown in Figure
3.1, where a schematic energy diagram depicts the amount of activation energy
(McClements, 1999) required to ‘kick start’ nucleation; energy on the y-axis and
cooling temperature on the x-axis, and structure on the second y-axis. Reducing
the cooling rate to 15°C causes a small change. Reducing to 0°C and cooler,
120
suggests an optimum has been reached. The energy barrier to crystallisation is
surpassed. The benefit of moving to lower cooling rates, together with mechanical
contact (shear), encourages the process of crystallisation (refer 1.3).
Until now, a relatively new technique (Young et al., 2008; Wassell et al., 2010b)
using UVP-PD, draws attention to a new method of observing rheological
behaviour of a crystallising TAG blend under real fluid / dynamic process
conditions, neatly demonstrating fundamental differences. Importantly, this non-
invasive technique is also able to observe, the effect on additions of minor
additives, e.g. addition of a relatively small inclusion of a longer chain MAG such
as monobehenate (see 1.5; 1.6). These additives are known to influence
crystallisation mechanics (seeding / promoting crystal networks). This aspect is
well documented (Sakamoto et al., 2003 & 2004). The effect has been less easily
observed in real dynamic processing conditions, until now.
121
A review of findings (Wassell et al., 2010a) clearly shows the necessity for a
multiple approach to structuring. Part of this multiple discipline highlights
investigation into the use of crystal modifiers; specifically, this being the use of
MAG. The influence of the longer chain behenic MAG has been discussed in
review (Wassell & Young 2007), and offers some influence on nucleation (For
comprehensive supporting detail, refer 1.6). Indirectly, the UVP-PD technique was
used to quantify the effect of crystallisation on a typical TAG blend; the TAG
blend being anhydrous. This investigation was to characterise a true in-line
measurement, under real processing conditions.
Figure 3.2 shows schematics of the test section (Appendix E shows previous and
current photo images of UVP test sections). The study characterised the influence
of an additive, based on behenic based MAG (GRINDSTED® Crystallizer 110) on
the selected TAG blend (Young et al., 2008).
Figure 3.2 Schematic image of pilot scale scraped surface heat exchanger unit and
pin rotor units with flow-cell and by-pass loop measuring apparatus with
differential pressure gauge and ultrasound transducers.
122
3.3.1 Off-line Haake Measurements
For comparison, several MAG / TAG mixtures were first measured off-line on a
controlled stress rheometer from Haake (Thermo Electron Corp., Karlsruhe,
Germany) (Haake RS 150), using a cup-bob geometry where the bob is of 20 mm
diameter (refer to 2.13.2). The temperature was cooled from 70°C to 40°C at 1°C
min-1 and the shear rate was held constant at 10 s-1. Results for viscosity as a
function of temperature are given in Figure 3.3.
The samples represent filling fats and were chosen for their resemblance to the
model systems to be tested with the UVP equipment. The filling fats (refer to
materials and methods Table 2.17) are: Kokowar, (hydrogenated, deodorised lauric
oil based on coconut) from Aarhus Karlshamn, Akocrem (the same as Akomic),
GRINDSTED® PS 209 (a blend of MAG and TAG, fully hydrogenated rapeseed
and palm-based oil) from Danisco A/S, GRINDSTED® Crystallizer 110
(frequently abbreviated throughout the subsequent text as CRY110 and described
previously in 2.0) and Dimodan® HP (fully distilled MAG, from fully
hydrogenated palm TAG, described previously in 2.0).
The data in Figure 3.3 depicts that, at a given temperature, in this case c.48°C,
there is the onset of a dramatic rise in viscosity, not seen in the samples where the
additive (CRY110) is absent. This increase in viscosity continues for the duration
of the experiment, whereby the actual viscosity increase is approaching a factor of
2 for the sample with the lowest viscosity and still significantly higher than the
previous highest result.
The results show a distinct trend for the viscosity increase as a result of CRY110
addition under the semi-static conditions of the rheometer, where shear rates and
cooling rates are predominantly lower than those experienced in real process
conditions. Hence, the need to be able to characterise the same rheological
parameters in-line and under real process conditions, which gives not only real
commercial value, but adds to the understanding of TAG crystallisation discipline
generally.
123
Figure 3.3 Viscosity as a function of temperature for TAG blends measured off-
line
124
heterogeneous nucleation induced by the CRY110, thereby promoting a more solid
TAG-crystal network structure. Isothermal change is partially explained by
increased energy consumption (ampere) drawn when supercooling the samples
(Amps first tube). Additional energy consumption (c. 1.8%) is drawn on the
sample with CRY110, while still maintaining steady-state thermodynamic-
processing conditions. Ideally, control of in situ isothermal heat of crystallisation
can help minimise post-structural changes, post-process. Most importantly, these
tests provided sufficient continual structural changes to be ‘captured’ downstream
at the point of UVP measurement.
125
Figure 3.4 Spectral plot together with an arithmetic average of some 30 measured
profiles (green) together with the resulting power-law fit (red) for the control
system, i.e. 25% Akomic / 75% rapeseed TAG at a flow rate of 70 kg/h
As seen for the sample without additive (CRY110) in Figure 3.4, there is a strong
shear-thinning behaviour, the power-law index is 0.1, and consistency index K is
8.53 Pas. The fit coefficient is satisfactory at R2 = 0.98. However, the consistency
index K between the two TAG blends is almost double for the fats blend with the
CRY110 additive (Figure 3.5) at 14.4 Pas. The fit coefficient is satisfactory at R2 =
0.98.
126
Figure 3.5 Corresponding spectral plot and ultrasound profile as an arithmetic
average of 30 measured profiles (green) and resulting power-law fit (red) for the
25% Akomic / 74% rapeseed TAG and 1% CRY110 (Distilled monoglyceride) at
70 kg/h.
To appreciate the significance of these in-line results (Figure 3.4 & 3.5), the flow
curves constructed from the velocity profiles for the samples with and without
added CRY110 are given in Figure 3.6.
127
Figure 3.6 Comparison of the Akomic / Rapeseed (RSO) TAG blends; without
additive (tetrahedrons) and with 1% CRY110 (circles)
The results in Figure 3.6 indicate that the sample containing the CRY110 had
increased the viscosity of the control, and these in-line results were in agreement
with previous off-line results, using a Haake Rheometer (3.3.1). It is clearly shown
that the UVP-PD method could rheologically differentiate the TAG blend, with
CRY110 (GRINDSTED® Crystallizer 110) and without additive. The viscosity of
the TAG blend increased by almost a factor of 2 after the addition of CRY110 and
thereby showed good agreement with the Haake off-line measurements.
The UVP-PD method directly allows measurements that were previously not
possible with common rheometers; namely the inline determination of rheological
properties and velocity profiles in real time. This therefore offers advantages over
commercially available process rheometers and off-line measurements.
The systems tested were limited to model systems because of current transducer
limitations and flow cell diameters, but nevertheless, a significant difference,
essentially a doubling of the viscosity, was measured and quantified (Young et al.,
2008).
128
3.4 Comparison and Validation
A further exploration was the next step, using upgraded UVP-PD apparatus with
new upgraded transducers, measuring to wider temperature and pressure range,
combined with rapid super-cooling in real-time. At this point the SFC values were
determined under dynamic conditions and these values were compared from the
UVP-PD measurement from those obtained from p-NMR data. This allowed
comparison and validation between the two techniques.
The data presented in Wassell et al. (2010b), clearly shows the upgraded UVP-PD
apparatus is comparable to the data reported previously (Young et al., 2008).
Specifically, when considering the velocity profiles found within Figures 3.7 and
3.8, there is obvious effect in the power law consistency index values e.g. 6.6 –
6.7Pas, compared to 14.2 – 14.3Pas, the latter, being with the CRY110. These
results (Wassell et al., 2010b) compare favourably with previous results (Young et
al., 2008).
Clearly, the sample with added CRY110 has a shear viscosity of twice that found
in the TAG blend without this additive. This compares well with previous off-line
data (3.3.1) measurements (Young et al., 2008).
129
Figure 3.7 Measured arithmetic average over 28 velocity profiles and the resulting
power-law fit for 25% Akomic / 75% rapeseed TAG without additive at a flow
rate of 70 kg h-1 (The profiles were measured both opposite to- (Transducer 1,
TDX1) and in the direction of the flow (Transducer 2, TDX2))
130
Figure 3.8 Measured arithmetic average over 39 velocity profiles and the resulting
power-law fit for 25% Akomic / 74% rapeseed TAG with 1% CRY110 at a flow
rate of 70 kg h-1 (The profiles were measured both opposite to- (Transducer 1,
TDX1) and in the direction of the flow (Transducer 2, TDX2))
131
Table 3.2 (Wassell et al., 2010b), shows response behaviour when subjected to
process modification, where exothermic behaviour is observed at the pin-worker
prior to UVP measurement.
UVP-PD measurement
This in-situ effect is potentially helpful to minimising any negative effects on the
TAG blend during post-process storage. Therefore, the net effect is possibly more
pronounced in the TAG blend containing CRY110, because this may minimise
observable structural changes within the transducer flow-cell, in respect to the
location (down-stream) of the UVP-PD analysis (Figure 3.2).
132
3.5 In-line Solid Fat Content
In line solid fat content (SFC) measurement was also explored and compared to a
well established method using Bruker p-NMR (2.14). Other reviewed authors
(Wassell et al., 2010a; Young et al., 2008), have examined the prospects of
following TAG crystallisation by means of ultrasonic, in-line analysis. They have
shown in-line detection of SFC is possible. However, SFC results gained under
true dynamic processing conditions had not yet been reported until now (Wassell
et al., 2010b). Results are presented in Figure 3.9, showing the SFC values as a
function of temperature for a 30% palm stearin / 70% rapeseed TAG blend (Table
2.17), comparing the two methods, p-NMR and UVP-PD. While there is good
correlation, the largest deviation is observed at the high temperature (40°C) and
low temperature (10°C).
Figure 3.9 Solid fat content (SFC) expressed as percentage values versus
temperature for 30% palm stearin / 70% rapeseed TAG measured by standard p-
NMR technique (triangles) and in-line dynamic conditions from UVP-PD (circles)
133
A correlation of the data and minimal variation, confirms from the initial
hypothesis (Young et al., 2008), that the UVP-PD can monitor the in-line rheology
of anhydrous TAG blends, with a dynamic process condition. Specifically, the
SFC measured differences are shown to result in a variation of not more than +/-
2% over the entire range, and within accepted experimental variation.
Many studies (Flöter & van Duijn 2006; Frederick et al., 2008; Garbolino et al.,
2005; Mazzanti et al., 2005) have been performed using isothermal conditions
where observations of crystal behaviour can be quite different from those observed
whilst in a dynamic, non-isothermal environment (Wassell et al., 2010a).
134
ultrasound can be used in an entirely different way, to stimulate nucleation by
applying high-intensity ultrasound (Sato & Ueno 2011). This approach was not
within the scope of the UVP-PD investigation, where the intention was to
characterise bulk crystallisation, not stimulate it.
Future work is still necessary to observe behaviour and robustness of water – oil
emulsion systems. The next apparent step is to introduce a water phase to observe
effects of emulsifiers on physical behaviour. Testing a water-in-oil (W/O)
emulsion proved problematic, due to the effects of pressure across the test
apparatus (to be maintained below 4 bar max), specifically on the flow cells, due
to potential stress and weakening to the prototype transducer housings. Despite
this current limitation, it will be necessary to run these experiments with updated
transducer housings (at time of writing, the transducer casing was upgraded. Refer
to Figure E4: Appendix E) in order for instrumentation to cope with additional
back-pressures. If several formulas are measured with a range of process
parameters, information ought to become available about the optimal addition of a
structurant e.g. single, mixed emulsifiers or other additive in reduced TAG
dynamic environments (Wassell et al., 2010a; 2010b; Young et al., 2008). While
this is important, because normally water droplet size is a usual parameter and
indicator for emulsion integrity, it is not the only parameter. Attention was
therefore turned towards interfacial tensiometry (4.0).
135
4.0 Rheology and Interfacial Surface-Interactive Behaviour of Mixed
Surfactant Systems
4.1 Summary
This work follows a natural progression from the pilot study examinations of
crystallisation speed and fatty acid diversity (1.5; 3.0).
New learning from these measurements helps to provide important data, which
supports several new patent applications (Wassell et al., 2012a; 2012b; 2012c;
2012d; 2012e; Bech et al., 2013).
136
4.2 Interfacial Tension - An Interfacial Examination of Single and Mixed
MAG Behaviour: Effect of Saturation and Chain Length
4.2.1 Introduction
Although many food products are essentially emulsions, interest in the structuring
of water-continuous emulsions is not necessarily straightforward, especially where
more novel structuring materials are used either wholly or partially to replace
traditional structurants (Wassell et al., 2010a).
Theory, suggests some emulsifiers are able to act as templating agents with other
surfactants, thereby resulting in increased strength of the interfacial membrane
between the oil and water phases. At a given interfacial emulsifier concentration,
temperature, and time, the surface of the water droplets will be partially or totally
137
covered (Johansson et al., 1995; Garti et al., 1998) with fatty acid chains likened to
a fat crystal surface (Krog & Larsson 1992).
In the case where PGPR and a MAG are used, the MAG is used to form template
or monolayer for heterogeneous crystallisation, while the PGPR may be used for
emulsion stability. However, at a certain critical micelle concentration (CMC) and
or change in temperature, the interaction to maintain, or form further template
formations may be either disturbed or strengthened. The formation for a template
for further crystallisation (Arima et al., 2009) is described in Figure 4.1.
Although Figure 4.1 describes an O/W emulsion and the type of emulsifier could
be different from case to case, it nevertheless shows the basic principle of
interaction and can be applied to Figure 4.2 which describes the same process with
a W/O emulsion.
138
Figure 4.2 Schematic of proposed template mechanism for heterogeneous
crystallisation of W/O emulsion with additives (PGPR, Monooleate,
Monobehenate) (Wassell et al., 2010a)
139
Figure 4.3 Proposed schematic showing interfacial crystallization at the water-in-
oil interface (Source: Wassell et al., 2012)
The intention of this study is to show the effect of temperature induced changes
through a series of complementary non-isothermal measurements as a function of
temperature rather than time (De Graef et al., 2006); then to investigate the surface
tension and rheology of several MAG of differing fatty acid profiles (different
saturation and chain length), which are examined as single MAG and mixed MAG.
These are also combined with novel Moringa MAG and PGPR.
140
4.2.2 Materials and Methods
141
Interfacial tension (Water / oil)
Cooling 0.3°C/min. Emulsifier 0.02%
35
30
25
0.02% Dimodan UJ
20
mN/m
0.06% Dimodan UJ
15
0.02%PGPR90 PLUS
10
0.02%PGPR90 PLUS+0.02%Dim. UJ
5
0
0 10 20 30 40 50 60
Temperature °C
Figure 4.4 Tensiometry values for PGPR 90 compared with unsaturated MAG
based on sunflower oil (DIMODAN® UJ)
Assuming that PGPR inhibits the effect of CRY110, then the following question
must be addressed: What is the effect of CRY110 and does it have any interaction
with PGPR or Dim UJ?
Figure 4.5 shows that a fully saturated MAG, rich in C22:0 (CRY110) both
individually and together with UJ (C18:1 / C18:2). Upon reaching towards 20°C, a
critical temperature is attained and a considerable rate of reduction in surface
tension occurs.
Where PGPR is present there is initial rapid decrease of ~5mN/m, from the
starting temperature at 50°C. All PGPR containing variants continued to gradually
decrease in tension from ~22mN/m to ~12mN/m, following a linear slope. When
reaching a critical interfacial tension temperature (Tγ) 15°C the PGPR blends fall
dramatically to 10°C. However, the PGPR alone continues on a linear slope to
approximately 5°C where measurement continues for 5 minutes. No further
reduction in tension occurs.
142
Interfacial tension (water / oil)
Cooling 0.3°C/min. Emulsifier addition 0.02%
35
30
Pure sunf lower oil
0.02%PGPR90 PLUS
mN/m
20
0.02%Cryst.110+0.02% Dim UJ
0
0 10 20 30 40 50 60
Temperature °C
Where PGPR is present, the situation in Figure 4.5 occurs in all cases irrespective
of emulsifier concentration, ratio, relative to PGPR (Dim UJ at 0.06%).
Given that PGPR and CRY110 are entirely different emulsifiers, it would seem
reasonable to assume, that through exchange, competitive adsorption occurs (Krog
1990). Both CRY110 and PGPR are lipophilic. PGPR is highly polar towards
water, having a larger hydrophilic head group (polyglycerol) compared with
CRY110 (MAG) and also has a large hydrophobic polyricinoleic acid chain.
The situation in Figure 4.5, strongly suggests PGPR is highly surface active
(Wassell et al., 2010a), at lower concentrations, dominating the interface
(Dedinaite & Campbell, 2000; Rousseau, 2000). The likely reason for this is due to
required physical properties of MAG, which are governed by temperature (Krog &
Larsson, 1992). The development of surface-active formations from the CRY110
does not occur until the C22:0 rich MAG reaches Tγ.
143
Table 4.1 Emulsifier specifications supplied courtesy of Danisco A/S, Denmark
acids (%)
C14:0 0.1 0.4 - 1 -
C16:0 7 20.5 44 55 -
C18:0 5 2.1 5 43 5
C18:1 28 41.3 39 0.1 -
C18:2 58 32.5 10 0.2 -
C18:3 0.1 1.3 0.3 - -
C18:0 total 91.1 75.9 54.3 43.3 5
C20:0 > 1 2.2 0.2 0.7 95
C22:0 total - 0.2 - - 85
Saturated total 12 24 49 100 100
IV (app.) 105 103 45 2 2
PGPR combined with a second or third emulsifier is perhaps more surface active
until ~15°C, whereby Tγ is observed. Standalone or mixtures of CRY110 with
PGPR follow a near linear shift from 20°C to 10°C, and converge with the same
inflection where PGPR is present as a single emulsifier at ~15°C. This suggests a
rearrangement / displacement (Marze, 2009) of the PGPR at the interface; and is
effectively “forced” or exchanged into a new geometry in the presence of
CRY110. PGPR continues to dominate (Gülseren & Corredig 2012) the surface
arrangement until broken at 15°C (Shimada & Ohashi, 2003). Therefore, the effect
of PGPR suppression is now dominated by CRY110 at a Tγ ~15°C onwards. The
behenic rich (C22:0) MAG crystals are more surface active and an exchange
through competitive adsorption occurs (Krog 1990).
144
Speculation exists as to whether a similar effect might be observed when
interchanging the C22:0 based MAG for one predominately rich in C16:0 or
C18:0, where not only the degree of saturation but also the chain length might
have an interaction (Table 4.1). The next step asked whether this happens
specifically because the MAG is C22 chain length.
A mixture of PGPR and MAG rich in C18:1 / 18:2 (DIMODAN® UJ) & C16:0 /
C18:0 (DIMODAN® HP) did not improve on PGPR alone. Only a mixture of
PGPR and MAG rich in C18:1 / 18:2 (DIMODAN® UJ) & C22:0 (CRY110)
resulted on tension impact at Tγ ~15°C. This effect must be entirely influenced by
C22:0 rich MAG (CRY110).
145
Interfacial Tension (water / oil)
Cooling 0.3°C/min.
35
30
0.02% Dimodan P/PEL
0.02% Dimdan HP
25
20
0.02%PGPR
15 0.02%PGPR90 PLUS+0.02%Cryst.110
0.02%PGPR90
PLUS+0.02%Cryst.110+0.02% Dim. UJ
10
0.02%PGPR90 PLUS+0.02%Dim.
UJ+0.2% Dim. HP
0
0 10 20 30 40 50 60
Temperature °C
After an initial steady fall in tension from the upper temperature towards lower
temperature, results from Krog (1990) showed dramatic fall of interfacial tension
once a critical temperature (Tγ) had been reached. This behaviour could possibly
have occurred more accentuated if the emulsifier concentration had been increased
to individual concentrations of between 0.03% - 0.3%. However, a risk of total
emulsifier overload could possibly mask small differentiations.
In this examination only liquid sunflower oil was used as the solvent. Where there
were obvious signs of crystallisation, the test samples containing C22:0 changed
the translucence of the oil, turning milky white. In the context and study of
interfacial behaviour, crystallisation at the surface or interface is suggested to be
faster than nucleation and crystal growth in the bulk oil, particularly when the
146
emulsification process also involves cooling. As the emulsifier crystals orientate,
they position methyl end group towards the oil and the polar heads towards water.
These are referred to as surface-active crystals (Krog & Larsson 1992).
4.2.4 Conclusion
This study not only confirms findings from the pilot study preliminary
investigations (1.5), where it was shown that behenic fatty acid based MAG or
dilutions have a pronounced effect on crystallisation; These new interfacial
measurements clearly demonstrate an unusual surface-interactive relationship.
They may also partially explain textural behaviour previously observed in
application trials of 12% WO emulsions (1.5).
A mixture of PGPR and MAG rich in C18:1 / 18:2 (DIMODAN® UJ) & C16:0 /
C18:0 (DIMODAN® HP) did not improve on PGPR alone. Only a mixture of
PGPR and MAG rich in C18:1 / 18:2 (DIMODAN® UJ) & CRY110 resulted on
tension impact at Tγ ~15°C. This effect must be entirely dominated by C22:0 rich
MAG (CRY110).
Whether similar behaviour could be observed in other solvents e.g liquid / solid
TAG mixtures remains to be discovered. Irrespective of this, it may be beneficial
to re-run these samples in controlled rheological measurements to observe if there
is any functional effect on viscoelastic behaviour, the outcome of which may have
implications for real bulk TAG and W/O (TAG) emulsions.
147
4.3 A Rheological and Interfacial Examination of Single, Mixed and Novel
MAG Behaviour: Effect of Saturation and Chain Length in Anhydrous Bulk
and Water-Oil Systems.
4.3.1 Introduction
Based on the conclusion in 4.2.4, attention was turned to interfacial and bulk
behaviour where single, mixed and novel MAG behaviour was examined to
observe the effects of saturation and chain length in anhydrous bulk and water-oil
systems. This is important, because in real emulsion systems, interest lies in the
properties of the interfacial film and its viscoelastic behaviour, where surface
elasticity is thought to be the determining factor in minimising film rupture (Boyd
et al., 1972) and hence coalescence. The strength of the interfacial film formed by
the emulsifier may be more important than its effect on interfacial tension
(Scherze, Knotha, Muschiolika, 2006).
Rheograms were obtained by plotting dynamic change in the elastic modulus (G’)
and viscous modulus (G’’) of the crystallising sample as a function of the
crystallisation temperature (non-isothermal).
Interfacial rheology and bulk rheology, were investigated using Physica MCR 301,
and data driven from RHEOPLUS/32 V3.21 software (Anton Paar, Germany
GmbH). Oscillatory interfacial method was used to measure interfacial properties
of interfaces covered with surface active molecules. Emulsifiers were weighed for
rheology measurements at 0.2% w/w (unless otherwise indicated) and the solvent
(RBD sunflower oil) balanced to 100% (Refer to general Materials & Methods
2.0).
148
Bulk rheology measurements were used to provide real constants for upper fluid
density and viscosity values for interfacial analyses. Complex viscosity is based on
constants of upper fluid density and viscosity parameter for Sunflower oil at 50°C
and the lower fluid density and viscosity parameter for water was 50°C. These
constants were used as a function of all 151 data points. The analysis method for
calculation of complex interfacial viscosity (η*i) by frequency sweep
measurements, with a bi-cone measuring system, are computed by known
rheological and mathematical regression formula (Oh & Slattery, 1978; Ray, Lee,
Jiang, & Jiang, 1987; Lee, Jiang, Jiang, & Avramdis, 1991; Nagarajan, Chung, &
Wasan, 1998). The interfacial properties are then calculated by analysing the raw
data using RHEOPLUS/32 V3.21 software. Ideally, the calculation of complex
interfacial viscosity eta-i, is completed for each of the 151 data points.
149
4.3.4 Results and Discussion
It is important to note that where measured and studied, the viscous (G’’) and
elastic (G’) moduli of several types of commercial MAG, using interfacial
rheology system (IRS), shown in Figure 4.7 were observed only with CRY110 and
mixtures of other MAG that resulted in development of both G’’ / G’ in samples.
Pure CRY110 is highly surface active at <20°C as shown from previous
tensiometer measurements (4.2).
Figure 4.7 Interfacial Rheology System (IRS) temperature sweep and resulting
complex modulus Gi’ and Gi’’ of CRY110 (rich in C22:0) alone compared to
PGPR and other emulsifier mixtures.
The temperature sweep shows that CRY110 is quite different from PGPR. The
presence of highly unsaturated MAG – Dim U/J with CRY110, impacts on both
Gi’ and Gi’’. The bulk oil phase reveals a similar picture (Figure 4.8), in that
CRY110 shows an elastic modulus (G’). The presence of an aqueous interface
150
(Figure 4.7) affects the onset of both Gi’ and Gi’’ at 30°C compared with the
anhydrous (Figure 4.8) bulk phase.
Figure 4.8 Bulk temperature sweep and resulting G’ and G’’ of CRY110 (rich in
C22:0) alone compared to PGPR and other emulsifier mixtures.
The results shown in Figure 4.7 and 4.8 seem to correspond to the tensiometer
measurement in Fig. 4.5 (see 4.1). It is suggested that the behaviour is due to a
development of stronger monolayer formation and corresponding affinity to the oil
phase is observed because of its C22 fatty acid configuration. This argument is
now supported by results from interfacial rheology (Figure 4.7), whereby with
continued onset of cooling to ~29°C, both G’ and G’’ begin to develop. Upon
reaching 18 – 20°C, the CRY110, with and without Dim U/J attains a G’’ plateau.
At a similar gradient, a continued trend is observed in development of the elastic
modulus G’.
151
Figure 4.9 Bulk measurement of the single emulsifiers compared to CRY110
In Figure 4.9 and 4.10, the bulk and interfacial temperature sweeps show the same
single emulsifiers measured using tensiometry (Figure 4.4 / Figure 4.5). Only the
DIMODAN® HP (C16:0 / C18:0) exhibited a small increase in G’’ compared to
CRY110.
152
Figure 4.10 Comparison of interfacial behaviour of single emulsifiers compared to
CRY110 (C22:0). Dim HP shows small change in G’’
153
When combined with PGPR and UJ, the CRY110 has slightly greater G’’ values
compared with DIMODAN® HP. The tensiometry readings in Figure 4.6 clearly
showed the effect of combining PGPR with UJ, and CRY110 or HP. A MAG rich
in C22:0 (CRY110) has stronger interaction with PGPR / UJ compared to HP. A
similar evidence is seen in Figure 4.11 and Figure 4.12 which show a small
deviation observed from ~15°C – 18°C where there is a small build-up of G’’ in
bulk and interfacial behaviour. A slight build-up of G’ is also found in the bulk.
Figure 4.11 The influence of a MAG rich in C22:0, where CRY110 affects both
G’’ and G’
154
Figure 4.12 The effect of CRY110 compared with HP in combination with PGPR
and UJ
The results from IRS and bulk rheology seem to concur with tensiometry data;
when PGPR is not present, higher tension values are observed. The interfacial
rheology data show higher viscosity without PGPR. At 18 – 20°C in both
interfacial rheology and tensiometry, there is clear increase in G’ and sharp
decrease in interfacial tension respectively (Figure 4.5 & 4.7). This particular
effect is likely observed because of the factors previously mentioned (temperature
/ time dependant changes occur). A similar outcome occurs whether in the bulk
phase, or with additional presence of an interface (water), so that while PGPR does
seem to suppress CRY110, this does not occur across the whole measured
temperature range.
155
A form of synergy seems to occur, where one emulsifier (PGPR) lowers interfacial
tension – early in process and the second emulsifier (CRY110) dominates the
surface tension faster as the material approaches a dramatic fall of interfacial
tension once a critical temperature (Tγ) had been reached (Krog 1990). A
comparison of IRS and bulk rheology suggests this behaviour is onset earlier in the
presence of a water phase. This will be examined in more detail through
application measurements (5.0).
In the Bulk oil phase (Figure 4.8) PGPR or Dim U/J are causing suppression to G’
development, whereas interfacial rheology (Figure 4.7) demonstrates that Dim U/J
is not disrupting assemblies or causing displacement because there is definite build
up of G’ in the presence of CRY110. Similar is observed in Figure 4.5, where Dim
U/J does not have strong interaction with CRY110 to cause earlier tension drop
and follows the same pattern as standalone CRY110. Possibly this occurs due to
similar affinity at the interface and or similar dielectric compatibility to each other
so that despite these structures having radically different fatty acid configurations,
these binary MAG mixtures are apparently still miscible (Ueno, Suetake, Yano,
Suzuki, & Sato, 1994). However, a broader examination of binary ratio’s of MAG,
coupled with a new temperature constant for each change in density as well as the
viscosity of the bulk and upper phase change, if this was possible (refer to 4.3.3),
may reveal undiscovered transition temperatures (Ueno et al., 1994).
156
4.3.6 Tension Reduction
The interactive behaviour of a MAG combined with PGPR observed within the
scope of these measurements has significance for real W/O emulsions, specifically
e.g. low fat spreads. It is found to be the case that tension is very low at typical
product packing temperatures, between 10°C – 20°C. The interfacial tension of
C22:0 fatty acids (in CRY110) and interactive behaviour of surface active
‘crystals’ may partially explain the presence of G’. Further, when CRY110 is
combined with PGPR, it is considered potentially important as a partial
explanation to the cause of why apparent thickening is observed (1.5).
If it was possible to find a novel emulsifier from an edible oil source, being non-
hydrogentated (unmodified), this would have potential commercial advantages
(Wassell & Young 2007; Wassell et al., 2010a; Wassell et al., 2012a; 2012b;
2012c; 2012d; 2012e; Bech et al., 2013).
It is reasonably established from the tensiometry and IRS temperature sweeps that
CRY110 functions because of it having a relatively rich source of C22:0, having a
longer fatty acid moiety than other common MAG with shorter fatty acid chains.
However, the problem to date is finding and developing surfactants which contain
a significant source of C22:0 and or other long chain moieties, and yet have not
undergone hydrogenation or other modification steps.
157
Figure 4.13 shows several MAG which, irrespective of their fatty acid
compositions, all give high tension measurements across a wide temperature
spectrum (50°C to 5°C). The ricebran MAG (predominant in C18:1/C18:2 and
~20% C16:0) gave a tension value similar to those grouped with MAG previously
show in Figure 4.4 and 4.6. This result is not surprising given that it has a similar
IV value (103) to sunflower based DIMODAN® UJ (105).
Cooling 0.3°C/min
Unless otherwise indicated, all dosed at 0.02% w/w Pure sunflower oil
35
0.02% Dim. UJ
0.02% Ricebran
20 0.06% Dimodan UJ
mN/m
0.02% Dimdan HP
15 0.02%PGPR90K
0.02%PGPR Super
10
0
0 10 20 30 40 50 60
Temperature °C
Figure 4.13 Interfacial tension measurements for several MAG, including novel
ricebran and Moringa, compared to several PGPR’s (PGPR Palsgaard 4150,
known as PGPR 90 Plus)
158
Moringa, as a standalone emulsifier, gives a similar γ to PGPR Plus, and is
actually similar to PGPR 90 (Figure 4.13). Possibly the Moringa behaviour is
dominating at the surface, causing the early onset of Tγ, when combining CRY110
and Moringa (Figure 4.14). From a commercial prospective, this gives certain
labelling advantages because both emulsifiers are declared E471.
0.02%PGPR90
PLUS+0.02%Cryst.110/water
20
mN/m
0.02%PGPR90 PLUS/water
15
0.02% Moringa 191
10
0.02%Moringa 191+ 0.02%Cryst. 110
0
0 10 20 30 40 50 60
Temperature °C
159
Figure 4.15 Bulk behaviour of Moringa MAG v CRY110 with PGPR
160
Figure 4.16 Bulk behaviour of Moringa MAG v PGPR
Figure 4.17 Bulk behaviour of Moringa and CRY110 compared to PGPR and
CRY110
161
Turning attention to the interfacial behaviour, it is apparent from Figure 4.18 that
Moringa / CRY110 combination is active in the G’ / G’’. The response curve is
similar to CRY110 / UJ shown in Figure 4.7.
When comparing Figure 4.5 and Figure 4.14, the tension value is slightly lower
overall with the CRY110 / Moringa. It is more surface active. But crucially, it
seems Moringa MAG is able to enhance and or build structure similar to CRY110
and yet compared with PGPR 90 (Figure 4.13), have remarkably similar γ
behaviour.
Figure 4.18 The complex IRS for Moringa MAG / CRY110 or PGPR
Another possible reason for the large differences in response behaviour shown in
Figure 4.19 and Figure 4.20 is because the PGPR 90 Plus is more surface active
compared to PGPR 90 and Moringa. This is seen in Figure 4.13 and helps to
explain the reason for Moringa having more G’’ compared with PGPR 90 Plus as
shown in Figure 4.20
162
Figure 4.19 Complex IRS: CRY110 & Moringa v CRY110 & PGPR
Figure 4.20 The effect on G’’ after interfacial temperature sweep for Moringa
MAG v PGPR
163
4.3.8 Conclusions and Recommendations
If it was possible for the Anton Paar RheoPlus software to handle complex
regression analysis for absolute values and each density / viscosity for each of the
151 data plots, the IRS and bulk rheology information would surely be more
informative and accurate (4.3.3). However, this was not possible because only one
T°C constant could be used. If a low T°C constant was used, then high T°C plots
would be inaccurate or lost. Conversely, in this procedure 50°C was the T°C
constant, which likely resulted in the low T°C plots being less reliable. However,
this does not minimise the information about viscoelastic crystallisation behaviour,
which could not be provided from more classic techniques e.g. pNMR, DSC (De
Graef et al., 2006, Wassell et al., 2010a)
164
The TAG solvent used for both tensiometry and IRS rheology was pure liquid
sunflower oil. In this situation, it not known if interstitial (Shiota, Iwasawa,
Kotera, Konno, Isogai, & Tanaka, 2011) liquid crystalline TAG structures may
have formed during thermal treatment in the presence of PGPR and or MAG
containing C22:0. As with interfacial tensiometry (4.2.4) there were obvious
crystallisation events where MAG containing C22:0 changed the translucence of
the oil at and near the interface, turning milky white. These must be surface-active
crystals (Krog & Larsson 1992), which may account for the resulting interfacial
behaviour, which in turn may have important implications for droplet stabilisation
in W/O low TAG emulsions specifically (Rousseau, Zilnik, Khan, & Hodge,
2003). Surface active PGPR was clearly dominant at the upper temperature region,
but as thermal treatment continued towards lower temperatures, insoluble surface
active crystalline MAG then competes at the interface. Rousseau et al (2003),
suggest a highly viscous and rigid interfacial film would result from sufficient
monolayer coverage, in effect – Pickering stabilisation (Pickering, 1907). In the
case of PGPR / CRY110 mixture, competitive adsorption (Bergenståhl, 2008),
together with previous factors discussed in this conclusion may explain these
rheological and interfacial results and therefore help to explain the textural
“thickening” observed in very low TAG (12%) W/O emulsions, from earlier
preliminary studies (1.5.3).
A dynamic interfacial tension measurement (Krüss DSA 100) would enable the
observation of thermal conditions where surface-active ingredients are still mobile
and molecular organisation is incomplete. Drop shape analysis video can measure
165
on time scales ranging from 5 minutes to 30 seconds. Comparing this information
with a fixed isothermal condition using a Krüss K10ST analysis (2.0) e.g. at ex-
packing (~10°C – 15°C) and domestic refrigeration conditions (~5°C), might show
new data, which if found, could provide additional insight into affects of chain
length transitions (Larsson et al., 1969). New information could potentially be
extended to other food and non-food applications (Carmichael 2011; Marangoni &
Garti 2011; Menaa et al., 2013; Piller 2011; Wang et al., 2012)..
Results obtained from tensiometry and interfacial rheology now require testing
(Wassell & Young 2007; Wassell et al., 2010a; Wassell et al., 2012) within
conditions that replicate those of the actual food products in which MAG
combinations, with and without PGPR, will be used (McClements, 2007), e.g.
W/O emulsions. These application trials and additional measurements will reveal
their emulsification capabilities and crystallisation behaviour under real dynamic
conditions.
166
5.0 A Study Series on the Behaviour of a Moringa MAG, Behenic Based
MAG and PGPR
5.1 Introduction
167
In attempting to discover the practical capabilities and application of the
emulsifying properties of novel Moringa MAG and other long chain fatty acid
mixtures, a series of application measurements are made. While reviews (Wassell
& Young 2007; Wassell et al., 2010a), techniques (Young et al., 2008; Wassell et
al., 2010b; 2012) and methodologies (2.0) for characterisation of emulsions or
dispersions have been made to test emulsifier efficiency (McClements 2007), a
consideration to the characteristics of the food in which the emulsifier is present is
key. Consequently, it is preferable to examine the effectiveness of an emulsifier
within conditions relevant to those used in the actual food product in which it will
be used (McClements 2007).
Fat based food systems, including dispersions (Wassell et al., WIPO No.
WO2012168722, 2012a; WO2012168723, 2012d), crystallisation (Wassell et al.,
WO2012168727 2012b; Bech et al., WO2013050944, 2013), water-in-oil low fat
spreads (Wassell et al., WO2012168726, 2012c), reduced fat water-in-oil spreads
(Wassell et al., WO2012168724, 2012e).
168
5.2 A Rheological Evaluation on a Moringa Monoglyceride (MAG) and
PGPR in Triacylglyceride (TAG)
5.2.1 Introduction
The work described herein investigated the functionality of Moringa MAG, based
on Moringa oleifera TAG and PGPR (GRINDSTED® PGPR 90) specifically in
peanut TAG – for peanut butter dispersions, and more generally in rapeseed TAG.
The rationale of the study was to investigate the effect of adding either Moringa
MAG or PGPR over a range of concentrations to peanut TAG or rapeseed TAG
respectively by measuring the rheological properties during cooling. This work
supports inventive claims for a texturant for bulk food, TAG blend and emulsion
(Wassell et al., 2012a; 2012b; 2012c; 2012d; 2012e; Bech et al., 2013).
RBD Peanut TAG and RBD rapeseed TAG (specification in 2.0) were used as the
base TAGs, to which a Moringa monoglyceride (MAG) and a PGPR
(GRINDSTED® PGPR 90) were added such that the concentrations were 0.1, 0.2,
0.4, 1, and 3%. Fatty acid content and various analyses of Moringa oleifera TAG
and monoglyceride are given in Table 5.2.1, 5.2.2 and 5.2.3. Refer to General
Materials & Methods (2.0) for additional detail.
Table 5.2.1 Fatty acid composition of Moringa oleifera TAG and MAG
Analysis (%) Moringa TAG / MAG Analysis (%) Moringa TAG / MAG
C14 0.1 C19 0.1
C15 <0.1 C20 3.4
C16 5.8 C20:1 2.2
C16:1 1.8 C22 5.8
C17 0.2 C22:1 0.1
C18 5.4 C24 1.0
C18:1 73.0 C26 -
C18:2 0.7 Unknown 0.2
C18:3 0.2
169
Table 5.2.2 Analyses of Moringa oleifera TAG
Each sample was then pre-heated in a microwave oven for 2 minutes at maximum
power before testing. This was to ensure the temperature exceeded 90°C, thereby
removing crystal history before the sample was then loaded and cooled on the
rheometer. In each case the measurements were carried out using a controlled
stress Haake RS 150 rheometer (2.13.2) fitted with a serrated parallel plate of
35mm in diameter, both top and bottom. Cooling took place from 85°C to 25°C at
the rate of 1°C per minute. The strain used was 0.004, 120 data points were
collected, the frequency was fixed at 0.5Hz, and the gap was reduced to 0.5mm.
Polarised Light Microscopy (PLM) was carried out at x20 magnification using a
polarised microscope (Olympus SC300).
170
5.2.3 Results and Discussion
Figures 5.2.1 and 5.2.2 show the results for the effect of PGPR and Moringa MAG
on rapeseed and peanut TAG respectively.
In Figure 5.2.1, the full concentration range of 0.1%, 0.2%, 0.4%, 1% and 3%
PGPR is tested, along with only Moringa MAG at 1% and 3% in rapeseed TAG.
The assumption was that with 1% Moringa displaying the same behaviour as the
raw rapeseed TAG sample, the results for the lower Moringa concentrations would
have followed a similar trend. For concentrations 0.1%, 0.2%, 0.4% and 1%
together with the pure rapeseed TAG the dynamic viscosity measured was
essentially constant between 0.03 and 0.04 Pas, whereas for 3% PGPR and 3%
Moringa MAG, differences were seen. For PGPR there is an observed shift in the
curve to a higher constant viscosity centred at 0.06 Pas, but for Moringa MAG,
dramatic changes in the profile of the curve occurred.
The Initial viscosity for Moringa MAG at the lower concentrations was 0.03 –
0.04Pas, but at a temperature corresponding to 32°C the onset of a dramatic
increase in viscosity is seen. Viscosity value increased to ~20 Pas, approaching 3
orders of magnitude greater than its start point, and occurred over a temperature
range of 6 – 7°C (25°C – 32°C).
171
100.000 04845
|ƒ*| = f (T) Rapeseed oil, pure
04839
|ƒ*| = f (T) PGPR 90, 3%
04840
|ƒ*| = f (T) PGPR 90, 1%
10.000 04841
|ƒ*| = f (T) PGPR 90, 0.4%
04844
|ƒ*| = f (T) PGPR 90, 0.2%
04843
|ƒ*| = f (T) PGPR 90, 0.1%
04847
1.000 |ƒ*| = f (T) Moringa
MoringaMAG 3%
oil, 3%
04848
|ƒ*| = f (T) Moringa
MoringaMAG 1%
oil, 1%
|ƒ*| in Pas
0.100
0.010
0.001
20 25 30 35 40 45 50 55 60 65 70
T in °C
HAAKE RheoWin 3.61.0005
In Figure 5.2.2, where the results for Peanut TAG are presented, essentially the
same picture is seen. For concentrations of PGPR below 3% and for Moringa
MAG at 1% the measured viscosity starts at around 0.03 Pas, and rises upon
cooling in a steady state to a value of 0.05 – 0.06 Pas, i.e. there is a greater rise in
viscosity of peanut TAG compared to rapeseed TAG. This however, can be
explained by the fact that the peanut TAG itself follows this curve, and could
likely be attributed to either its intrinsic wax content, or natural content of C22:0
of ~1.5 – 4.5%, and the generally more saturated nature, IV 86 – 107, of the
peanut TAG (FAO/WHO 2011). PGPR at 3% does show greater viscosity than the
samples so far mentioned, but not to the same extent as it did in Figure 5.1. Here
the viscosity begins at 70°C around 0.04 Pas and ends around 0.06 Pas at 26°C.
Moringa MAG at 3% mirrors the tendency displayed from Figure 5.2.1, i.e. in this
case it follows the same gently increasing viscosity of the other samples until a
172
temperature of 30°C, a similarly dramatic increase in viscosity is recorded. The
increase is less than in Figure 5.2.1, with a final measured value plateau round 7
Pas, but still enough to record a 2 order of magnitude increase.
10.00 04846
|ƒ*| = f (T) Peanut oil, pure
04854
|ƒ*| = f (T) PGPR 90, 3%
04856
|ƒ*| = f (T) PGPR 90, 1%
04857
|ƒ*| = f (T) PGPR 90, 0.4%
04855
|ƒ*| Moringa
= f (T) Moringa oil, 3%
MAG
04858
1.00 |ƒ*| = f (T) Moringa oil, 1%
Moringa MAG 1%
|ƒ*| in Pas
0.10
0.01
20 25 30 35 40 45 50 55 60 65 70
T in °C
The results from Figures 5.2.1 and 5.2.2 show that both the tested TAGs (rapeseed
and peanut) are affected at around 30°C to 32°C, which does not occur in samples
below the Moringa MAG concentration of 3%. The reason why Moringa MAG
shows this dramatic increase in viscosity, can likely be attributed to its own fatty
acid composition, which is retained when synthesised from the original TAG (refer
to 5.11). Moringa oleifera TAG fatty acid composition is C16:0 – 6.0%, C16:1 –
1.4%, C18:0 – 4.0%, C18:1 – 75.0%, C20:0 – 2.5%, C20:1 – 2.5%, C22:0 – 5.8%,
173
and >C24:0 - 1.0% (Abdulkarim et al., 2005; Lalas & Tsaknis, 2002) which is in
good agreement with the analysis in table 5.2.1. There is a significant proportion
of Moringa MAG with C22:0 fatty acids (behenic acid), similar to CRY110 (refer
to 1.5; 2.0; 3.0; 4.0). Therefore, it is reasonable to suggest that the increase in
viscosity is likely due to the presence of long chain saturated fatty acids, which
also include C20:0 and >C24:0.
Supporting evidence as to why the viscosity rise occurs is shown via microscopy.
The micrographs for the 3% samples of both PGPR and Moringa MAG in the
rapeseed and peanut TAGs are shown in Figure 5.2.3.
Figures 5.2.3a and 5.2.3c show 3% Moringa MAG in rapeseed and peanut TAG
respectively show a higher level of crystalline structure than Figures 5.2.3b or
5.2.3d with 3% PGPR in the same respective TAGs. Figures 5.2.3a and 5.2.3c,
shows the presence of crystalline entities, which may be attributed to the presence
of the Moringa MAG. In contrast Figures 5.2.3b and 5.2.3d detailing 3% PGPR in
the same rapeseed or peanut TAG show much less in the way of crystal formation,
and thereby explain why the viscosity profile remains at the low levels see in
Figures 5.2.1 and 5.2.2.
174
Figure 5.2.3 Micrographs of 3% Moringa MAG in rapeseed and peanut TAG (3a,
3c), and 3% PGPR in rapeseed and peanut TAG (3b, 3d) at approximately 25°C –
26°C
5.2.4 Conclusion
The rheological behaviour of Moringa MAG and PGPR 90 in two different liquid
TAGs are essentially the same once below a concentration of 3%. At 3% there is a
significant increase in dynamic viscosity from a temperature of 32°C for rapeseed
TAG and 30°C for peanut TAG. The increase continues reaching levels up to 3
orders of magnitude greater than the start point. This increase is attributed to the
greater fat crystal density seen in the Moringa samples, as shown by the
micrographs; in the peanut TAG specifically, there may also likely be another
factor where the TAG solvent i.e. degree of its total saturation, may also have a
contributing factor. However, a Moringa based MAG with a fatty acid
specification similar to those described in tables 5.2.1, 5.2.2 and 5.2.3 is thought to
be the reason for the proliferation of crystal growth (Wassell & Young 2007;
Wassell et al., 2012a; 2012b; 2012c; 2012d; 2012e; Bech et al., 2013). This aspect
requires further investigation.
175
5.3 Crystallisation Effects of Natural Moringa MAG, a Synthetic Behenic
Based MAG Composition and PGPR by Cold Stage Polarised Light
Microscopy
5.3.1 Introduction
A Moringa (Moringa oleifera TAG) based MAG possibly has some unusual
similarities to PGPR in certain applications (refer to 5.2). This investigation
attempts to provide clear visual evidence to confirm these similarities. Through an
examination of samples cooled by controlled thermal induction, and then using a
polarised light microscope, it is possible to gain optical evidence, resulting in a
snapshot of the crystal structure at a given temperature. Therefore, an induction
heat /cool regime can provide evidence that is being sought.
176
5.3.2 Materials and Methods
The samples were used from the melt after pre-heating to remove crystal history,
starting at 85°C and then cooling to 20°C at a temperature cooling rate of 1°C per
minute. Photographic images were taken approximately once every 2°C and
recorded. Using an Olympus BX 60 microscope, all images were taken under
polarised light at a magnification of x200.
The TAG blend used in all cases was 70% Palm Stearin, with an iodine value
(IV35) and 30% Palm Olein (IV 56). 1% of either Moringa MAG or
GRINDSTED® Crystallizer 110 (CRY110) or 0.5% GRINDSTED® PGPR 90
(PGPR) was added (2.0)
Moringa MAG samples, were either classified as natural (sourced from: Moringa
oleifera TAG) or synthetic, i.e. the natural samples are true Moringa sourced
materials (2.0), whereas the synthetic samples have been assembled by attempting
to blend a similar fatty acid and glyceride content similar to that of the natural
samples as given in Table 5.3.1.
To try and match the natural Moringa MAG in Table 5.3.1, a series of synthetic
samples were blended (Table 5.3.2) and dosed at the same 1% inclusion rate.
Table 5.3.1 Natural Moringa MAG with the breakdown of mono, di- and tri-
glycerides
In Tables 5.3.1 and 5.3.3, the abbreviations stand for; GL – Glycerol, Digl – Diglycerol,
FFA – Free Fatty Acids, Mono – Monoglycerides, Di – Diglycerides, and Tri –
Triglycerides.
177
Synthetic MAG: Blends of commercial MAG (Table 5.3.2) were assembled in
such a way as to provide a similar fatty acid profile to natural distilled Moringa
monoglycerides (Table 5.3.4). Resultant compositions are shown in Table 5.3.3
and 5.3.5. The predominant fatty acid base for all three compositions is based on
Table 5.3.6.
Table 5.3.3 Glyceride distribution of synthetic Moringa MAG (mono, di-, tri-
glycerides)
Analysis (%) SM 90 SM 60 SM 80
GL 0.16 0.24 0.20
Digl 0.14 0.1 0.18
FFA 0.30 0.40 0.40
Mono 96.50 64.56 82.87
Di 2.64 29.02 15.28
Tri 0.22 2.59 1.10
178
The natural MAG samples, align approximately with synthetic MAG samples as follows:-
Natural Synthetic
2472/191 = SM 90
2559/102 = SM 60
2559/104 = SM 80
The iodine values (IV) of these samples was calculated according to the principles
The fatty acid compositions of the natural Moringa MAG and synthetic MAG are
shown in Tables 5.3.4 and 5.3.5.
179
Table 5.3.5 Fatty acid composition of synthetic Moringa MAG: sample 1 (SM
90), sample 2 (SM 60) and sample 3 (SM 80)
180
5.3.3 Results and Discussion
The micrographs shown in Figure 5.3.1a to 5.3.1f depict the crystal structure as
seen via microscope at 20°C after the samples were cooled from 85 – 20°C at a
cooling rate of 1°C per minute. The samples represented from Figure 5.3.1a to
5.3.1f respectively are:
1f 1% Moringa (191)
Moringa (191) refers to the Moringa MAG being fully distilled (see Table 5.3.1 –
sample 2472/191)
The blank sample, Figure 5.3.1a, shows clear evidence of TAG crystallisation,
with pentagonal and hexagonal snowflake like patterned crystal structures. Figure
5.3.1b, shows 1% of CRY110, with many small discrete TAG crystals. A similar
micrograph image has recently been reported (Basso et al., 2010) when adding a
behenic based monoglyceride. The dramatic difference between Figures 5.3.1a and
5.3.1b is possibly attributed to the crystallisation onset, which may occur at higher
temperatures where the CRY110 is present. The solubility and degree of total
saturation of the TAG solvent will also have a bearing (Refer to 5.2, A rheological
evaluation on a Moringa Monoglyceride (MAG) and PGPR in TAG and 1.5, A
Pilot Study – preliminary investigations for enhancing crystallisation of anhydrous
TAG dispersions and W/O Emulsions) .
181
The next four micrographs from Figure 5.3.1 are examined in pairs, starting next
with the single samples of PGPR and Moringa MAG (191). In Figures 5.3.1d and
5.3.1f respectively, there are clear dendrite structures (Mullin, 1993) with the
presence of PGPR. Though not as comparatively large, Figure 5.3.1f (Moringa
MAG 191) is densely populated with irregular crystal structures of which also
dendrite structures are present. There is also evidence of several large cluster
formations, which were seen in Figure 5.3.1a, but without doubt it is the dendrite
like structure which dominates. This is completely different from the crystal
structure of Figure 5.3.1b, and given that all observed images are in reality an
intertwined 3-D network of interpenetrating crystal flocculation’s, then the
physical dimensions of the dendrite structures by their surface area, suggests these
could have high function in real food systems.
The similarities between Figure 5.3.1d and 5.3.1f leads to the suggestion that the
behaviour of Moringa MAG could be similar to PGPR, but given that the dendrite
structures in 5.3.1f are not so well “developed”, then performance will need to be
determined in application trials.
The potentially similar functionality of Moringa MAG and PGPR can be seen
when comparing the last two micrographs of Figure 5.3.1, i.e. Figure 5.3.1c and
Figure 5.3.1e. These refer to a combination of two samples, where Figure 5.3.1c is
a combination of 1% CRY110 and 0.5% PGPR and Figure 5.3.1e is 1% Moringa
MAG (191) and 1% CRY110. In both cases the micrograph show an entirely
different picture compared to these components in isolation i.e. Figure 5.3.1b, 1d,
1f. This suggests that whatever the effect, the interaction of adding PGPR or
Moringa MAG (191) to the CRY110 has a similar outcome.
Three important observations come from Figure 5.3.1, first; Moringa MAG (191)
which naturally contains ~6% behenic acid (C22:0) behaves differently to
CRY110, which contains a high source of C22:0. Second, Moringa MAG (191)
appears to have some similar crystal structures to PGPR, suggesting the possibility
of similar behaviour.
182
Figure 5.3.1 Micrographs showing the crystal structures of samples 1a – 1f at
20°C in 70% Palm Stearin (IV35) and 30% Palm Olein (IV 56) (scale bars =
20µm)
183
Third, Moringa MAG (191) and PGPR appear to demonstrate similar interaction
with CRY110. These conclusions strongly point to the potential of Moringa MAG
(191) possessing similar properties to PGPR.
Moringa TAG (Moringa oleifera), being a natural source of saturated fatty acids
(Abdulkarim et al., 2005; Anwar and Rashid 2007; Fahey 2005; Lalas and
Tsaknis, 2002; Mohammed et al., 2003) gives this material potentially attractive
nutrition labelling. Currently, while TAG from Moringa oleifera is a known
source of food energy and has multipurpose use (Marikkara & Ghazali 2011;
Pandey et al., 2011), it is not currently recognised as a commodity TAG at present.
However, once synthesised to a MAG it could (subject to Codex & Novel foods
regulation) be identified as E471. This is advantageous for low TAG W/O
emulsions < 41%, where it is not unusual to combine both a MAG (E471) and
PGPR (E476); it may be possible to use only E471. The use of mono- and di-
glycerides prepared from Moringa oleifera TAG has not previously been taught
for such applications (Wassell et al., 2012a; 2012b; 2012c; 2012d; 2012e; Bech et
al., 2013).
Given the results observed in the application of a natural Moringa MAG (sample
2472/191 in Table 5.3.1), could a “synthetic” mix of fatty acids and Mono, Di, Tri
composition offer similar properties as the natural Moringa MAG. Three synthetic
versions of specific fatty acid composition were assembled (Table 5.3.2) and
tested.
184
Figure 5.3.2 Micrographs 2a – 2f show crystal structures of natural and synthetic
Moringa MAG (dose 1%) at 20°C in 70% Palm Stearin (IV35) and 30% Palm
Olein (IV 56) (scale bars = 10µm)
185
The results in Figure 5.3.2 show Figures 5.3.2a to 5.3.2f, where 2a, 2b, and 2c
represent the natural Moringa MAG, and 2d, 2e, and 2f represent the synthetic
blends made to match the natural Moringa MAG components (fatty acid profile).
In general the micrographs in Figure 5.3.2 seem to show that synthetic Moringa
MAG samples have similar fractural patterns, compared to the natural Moringa
MAG. More specifically, Figures 5.3.2a, and 5.3.2d which represent the natural
and synthetic versions at 91.14%, 96.5% mono content respectively, both show the
scattered TAG crystallisation in the bulk, but also the evidence of single larger
branched clusters together with the dendrite structures similar to those seen in
Figure 5.3.1d and 5.3.1f. Both are clearly distinct from CRY110 (Figure 5.3.1b).
Figures 5.3.2b and 5.3.2e, natural and synthetic versions of 53.16% and 64.56%
mono content respectively; Figure 5.3.2e shows larger clusters as opposed to the
discrete nuclei of the CRY110 in Figure 5.3.1b. Less evident in both cases is the
dendrite structures associated with PGPR (Figure 5.3.1d) and Moringa 191 (Figure
5.3.1f) at higher monoglyceride contents. However, Figure 5.3.2e has structures
found in both Figure 5.3.2a and 5.3.2b, i.e. structures that equate to as yet
underdeveloped fern-like structures, which may be signs of a delayed crystal
development due to the higher di-glyceride (DAG) content within the TAG olein.
Figures 5.3.2c and 5.3.2f, natural and synthetic versions at 82.55%, 82.87% mono
content respectively, show similar smaller crystals coupled with several large
clusters. This is not dissimilar to 5.3.2b.
Summarising Figure 5.3.2 it is concluded that firstly the ability to make synthetic
samples of Moringa by blending MAG to achieve a fatty acid composition is
possible. Secondly, it seems as though either the mono or di-glyceride content is
important, but this matter requires further investigation before explanations can be
attributed.
186
5.3.4 Conclusion
This study has shown that based on micrograph evidence, the influence on TAG
crystallisation of Moringa MAG (191) is different from CRY110, despite Moringa
MAG (191) containing some 5.8% - 6% of C22:0 (behenic acid) and yet exhibiting
behaviour not dissimilar to PGPR 90.
Figure 5.3.1f shows the TAG blend with added Moringa MAG to be essentially
like viewing a composite image of Figures 5.3.1a and 5.3.1d together, where bulk
TAG crystallisation and obvious evidence of fern-like dendrite structures are
present. Given that both aspects are now present in a single system, this may
possibly provide evidence as to why the interfacial tension measurements (Figure
4.14 in 4.2) showed PGPR and a Moringa MAG behaving with not dissimilar
tensiometry values.
Combining Moringa MAG (191) with CRY110, and PGPR with CRY110 resulted
in similar micrograph observations.
187
5.4 High and Low TAG W/O Emulsion Application Trials: A Functional
Evaluation of MAG based on Moringa, Lesquerella, Rapeseed, Sunflower
TAG and a PGPR
5.4.1 Introduction
It has been shown (4.0) that a MAG based on Moringa oleifera TAG, can have
similar interfacial and crystallisation (5.2 & 5.3) characteristics to PGPR
(GRINDSTED® PGPR 90). Previous examinations have centred on model
systems. In real emulsion systems, interest lies in the properties of the interfacial
film and its viscoelastic behaviour, where surface elasticity is thought to be the
determining factor in minimising film rupture (Boyd et al., 1972) and hence
coalescence. The strength of the interfacial film formed by the emulsifier may be
more important than its effect on interfacial tension (Scherze, Knotha,
Muschiolika, 2006). An emulsifier which is very effective in stabilising an
emulsion may be much less effective in aiding the construction of an initial small
water droplet and narrow DSD, i.e. it may help form visco-elastic films at the
TAG / water interface, but may result in emulsions of relatively large DSD. This
may be irrespective of its hydrophobic influence on reducing interfacial tension to
low values (Boyd et al., 1972). To obtain confirmation about the functionality of
Moringa MAG, proof of concept by way of application trials were required. A
series of W/O emulsion spreads are prepared at 60% and 40% TAG concentration
respectively. Results from water Droplet Size Distribution (DSD) analysis,
Confocal Laser Scanning Microscopy (CLSM) imaging, texture analysis and
sensory tasting are used as a basis for evaluation.
High and low TAG WO emulsion application trials were prepared for evaluation
of monoglycerides based on Moringa oleifera TAG and Lesquerella fendleri TAG
(Appendix F, G); commercial rapeseed based (DIMODAN® RT); sunflower based
(DIMODAN® UJ) and a PGPR (GRINDSTED® PGPR 90).
188
The W/O emulsion samples were made according to formula and process
conditions given below for 60% W/O TAG spreads in Tables 5.4.1 and 5.4.2, and
for 40% W/O TAG spreads in Tables 5.4.3 and 5.4.4.
For more details of emulsifier and general process conditions and analysis, refer to
Materials and Methods (2.0)
Ingredients in %
Ingredient Name 1 2 3 4 5 6
Water phase
Water (Tap) 38.400 38.400 38.400 38.400 38.400 38.400
Salt 1.000 1.000 1.000 1.000 1.000 1.000
Skimmed milk powder (MILEX 240) 0.500 0.500 0.500 0.500 0.500 0.500
Potassium Sorbate 0.100 0.100 0.100 0.100 0.100 0.100
Water phase total 40.000 40.000 40.000 40.000 40.000 40.000
pH 5.5 5.5 5.5 5.5 5.5 5.5
Fat phase
PK4 - INES 25.000 25.000 25.000 25.000 25.000 25.000
Rapeseed oil 75.000 75.000 75.000 75.000 75.000 75.000
Fat blend total 100.000 100.000 100.000 100.000 100.000 100.000
Other fat ingredients
Distilled Lesquerella Monoglyceride 0.300 0.600
DIMODAN R-T PEL/B - K 0.300 0.600
Moringa MAG (191) 0.300 0.600
TOCO 50 0.010 0.010 0.010 0.010 0.010 0.010
2% sol. beta-carotene 0.020 0.020 0.020 0.020 0.020 0.020
Other fat ingredients total 0.330 0.630 0.330 0.630 0.330 0.630
Fat phase total 60.000 60.000 60.000 60.000 60.000 60.000
RECIPE total (calc. batchsize) 100.000 100.000 100.000 100.000 100.000 100.000
189
The emulsifiers tested in 60% W/O TAG spreads samples according to Table 5.4.1
are as follows:
1. DIMODAN® RT 0.3%
2. DIMODAN® RT 0.6%
3. Moringa MAG (191) 0.3%
4. Moringa MAG (191) 0.6%
5. Lesquerella 0.3% (Fatty acid details in 2.0)
6. Lesquerella 0.6%
Table 5.4.2 Pilot plant processing conditions for the 60% TAG spread samples
Pilot Plant
Processing (3-tube lab perfector): 1 2 3 4 5 6
Oil phase temperature 50 50 50 50 50 50
Water phase temperature 20 20 20 20 20 20
Emulsion temperature 50 50 50 50 50 50
Centrifugal pump Auto Auto Auto Auto Auto Auto
Capacity high pressure pump 40 40 40 40 40 40
Cooling (NH3) tube 1: -10 -10 -10 -10 -10 -10
Cooling (NH3) tube 2: -10 -10 -10 -10 -10 -10
Cooling (NH3) tube 3: -10 -10 -10 -10 -10 -10
Rpm tube 1: 1000 1000 1000 1000 1000 1000
Rpm tube 2: 1000 1000 1000 1000 1000 1000
Rpm tube 3: 1000 1000 1000 1000 1000 1000
190
Table 5.4.3 Recipe for 40% TAG spreads trials 11-19
Ingredients in %
Ingredient Name 11 12 13 14 15 16 17 18 19
Water phase
Water (Tap) 57.300 57.300 57.300 57.300 57.300 57.300 57.300 57.300 57.300
Salt 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000
Skimmed milk powder (MILEX
0.100 0.100 0.100 0.100 0.100 0.100 0.100 0.100 0.100
240)
GRINDSTED® LFS 560 1.500 1.500 1.500 1.500 1.500 1.500 1.500 1.500 1.500
Potassium Sorbate 0.100 0.100 0.100 0.100 0.100 0.100 0.100 0.100 0.100
Water phase total 60.000 60.000 60.000 60.000 60.000 60.000 60.000 60.000 60.000
pH 5.5 5.5 5.5 5.5 5.5 5.5 5.5 5.5 5.5
Fat phase
Fat blend
PK4 – INES 25.000 25.000 25.000 25.000 25.000 25.000 25.000 25.000 25.000
Rapeseed oil 75.000 75.000 75.000 75.000 75.000 75.000 75.000 75.000 75.000
Fat blend total 100.000 100.000 100.000 100.000 100.000 100.000 100.000 100.000 100.000
Other fat ingredients
Distilled Lesquerella
0.300 0.600 0.300
Monoglyceride
DIMODAN U/J – K 0.300 0.600 0.300
Moringa MAG (191) 0.300 0.600 0.300
PGPR 90 – K 0.200 0.200 0.200
TOCO 50 0.010 0.010 0.010 0.010 0.010 0.010 0.010 0.010 0.010
2% sol. beta-carotene 0.020 0.020 0.020 0.020 0.020 0.020 0.020 0.020 0.020
Other fat ingredients total 0.330 0.330 0.330 0.630 0.630 0.630 0.530 0.530 0.530
Fat phase total 40.000 40.000 40.000 40.000 40.000 40.000 40.000 40.000 40.000
RECIPE total (calc. batchsize) 100.000 100.000 100.000 100.000 100.000 100.000 100.000 100.000 100.000
The emulsifiers tested in 40% W/O TAG spreads samples according to Table 5.4.3
are as follows:
191
Table 5.4.4 Pilot plant processing conditions for the 40% TAG spread samples
Pilot Plant
Processing (3-tube lab 11 12 13 14 15 16 17 18 19
perfector):
Oil phase temperature 50 50 50 50 50 50 50 50 50
Water phase temperature 50 50 50 50 50 50 50 50 50
Emulsion temperature 50 50 50 50 50 50 50 50 50
Centrifugal pump Auto Auto Auto Auto Auto Auto Auto Auto Auto
Capacity high pressure pump 40 40 40 40 40 40 40 40 40
Cooling (NH3) tube 1: -10 -10 -10 -10 -10 -10 -10 -10 -10
Cooling (NH3) tube 2: -10 -10 -10 -10 -10 -10 -10 -10 -10
Cooling (NH3) tube 3: -10 -10 -10 -10 -10 -10 -10 -10 -10
Rpm tube 1: 1000 1000 1000 1000 1000 1000 1000 1000 1000
Rpm tube 2: 1000 1000 1000 1000 1000 1000 1000 1000 1000
Rpm tube 3: 1000 1000 1000 1000 1000 1000 1000 1000 1000
Texture analysis was carried out with a Stable Micro Systems: TA-XT2i texture
analyser and run under standard methodology for measuring hardness (2.0).
The recipes and plant conditions for the validation trails are given below in Tables
5.4.5 and 5.4.6. Moringa MAG 191 is the same Moringa MAG identified in 5.2
and 5.3. Tests with Lesquerella MAG were not repeated, but abandoned from this
point, because of obvious emulsion separation (refer to results).
192
Table 5.4.5 Recipe for 40% TAG spreads (validation test)
Ingredient Name 1 2 3 4 5 6
Water (Tap) 57.300 57.300 57.300 57.300 57.300 57.300
Salt 1.000 1.000 1.000 1.000 1.000 1.000
Skimmed milk powder (MILEX 240) 0.100 0.100 0.100 0.100 0.100 0.100
GRINDSTED® LFS 560 1.500 1.500 1.500 1.500 1.500 1.500
Potassium Sorbate 0.100 0.100 0.100 0.100 0.100 0.100
Butter Flavouring 050001 T03007 0.010 0.010 0.010 0.010 0.010 0.010
Water phase total 60.010 60.010 60.010 60.010 60.010 60.010
pH 5.5 5.5 5.5 5.5 5.5 5.5
PK4 - INES 25.000 25.000 25.000 25.000 25.000 25.000
Rapeseed oil 75.000 75.000 75.000 75.000 75.000 75.000
Fat blend total 100.000 100.000 100.000 100.000 100.000 100.000
DIMODAN U/J 0.300 0.600 0.300
PGPR 90 - K 0.200 0.200
Moringa MAG (191) 0.300 0.600 0.300
TOCO 50
Butter Flavouring 050001 T04184 0.020 0.020 0.020 0.020 0.020 0.020
Other fat ingredients total 0.320 0.320 0.620 0.620 0.520 0.520
Fat phase total 39.990 39.990 39.990 39.990 39.990 39.990
RECIPE total (calc. batchsize) 100.000 100.000 100.000 100.000 100.000 100.000
Table 5.4.6 Plant process conditions for the 40% TAG spreads (Validation test)
Processing (3-tube lab perfector): Default 1 2 3 4 5 6
Oil phase temperature 50 50 50 50 50 50 50
Water phase temperature 50 50 50 50 50 50 50
Emulsion temperature 50 50 50 50 50 50 50
Centrifugal pump Auto Auto Auto Auto Auto Auto Auto
Capacity high pressure pump 40 40 40 40 40 40 40
Cooling (NH3) tube 1: -10 -10 -10 -10 -10 -10 -10
Cooling (NH3) tube 2: -10 -10 -10 -10 -10 -10 -10
Cooling (NH3) tube 3: -10 -10 -10 -10 -10 -10 -10
Rpm tube 1: 1000 1000 1000 1000 1000 1000 1000
Rpm tube 2: 1000 1000 1000 1000 1000 1000 1000
Rpm tube 3: 1000 1000 1000 1000 1000 1000 1000
193
5.4.4 Results and Discussion
Table 5.4.7 shows the DSD for the samples from 60% TAG spreads based on
those trials shown in Table 5.4.1
The DSD reveals that water droplet size volume at 97.5% <µm, are grouped, with
the exception of samples containing Lesquerella, which considered highly
unstable. At 60% TAG concentration the Moringa and DIMODAN® RT results
suggest stable emulsions and thereby smaller water droplet sizes. The DSD data
suggests the presence of “large lakes” of water. This assumption is further
supported by the data of the distribution as expressed graphically in Figure 5.4.1.
Clearly visible in Figure 5.4.1 is the distribution of the water droplet sizes for the
60% TAG spreads, showing the extremely large size of water droplets
corresponding to Lesquerella. Conversely, Moringa sample 4 at the concentration
of 0.6% is showing a water droplet size broadly akin to that of sample 1 which
corresponds to DIMODAN® RT at 0.3%. This suggests that the stability of the
0.6% Moringa is potentially similar DIMODAN® RT 0.3% sample.
194
Figure 5.4.1 Water droplet size distribution (DSD) of the 60% TAG blends.
The data presented in Figures 5.4.2 to 5.4.7 are the CLSM images of application
trials 1 – 6, as described in Table 5.4.1. Four images at two magnifications (40X
and 100X) were taken respectively. The images are then reproduced (scaled) to
375 x 375 µm and 188 x 188µm. All samples are stained with FITC, stains protein
green, and Nile Red, stains TAG red.
195
Figure 5.4.2 CLSM image of 60% TAG spread with DIMODAN® RT at 0.3%
concentration. (Top images 375 x 375 µm. Bottom images 188 x 188µm)
196
Figure 5.4.3 CLSM image of 60% TAG spread with DIMODAN® RT at 0.6%
concentration (Top images 375 x 375 µm. Bottom images 188 x 188µm)
197
Figure 5.4.4 CLSM image of 60% TAG spread with Moringa at 0.3%
concentration. (Top images 375 x 375 µm. Bottom images 188 x 188µm)
198
Figure 5.4.5 CLSM image of 60% TAG spread with Moringa at 0.6%
concentration. (Top images 375 x 375 µm. Bottom images 188 x 188µm)
199
Figure 5.4.6 CLSM image of 60% TAG spread with Lesquerella at 0.3%
concentration. (Top images 375 x 375 µm. Bottom images 188 x 188µm)
200
Figure 5.4.7 CLSM image of 60% TAG spread with Lesquerella at 0.6%
concentration. (Top images 375 x 375 µm. Bottom images 188 x 188µm)
201
The slides above in Figures 5.4.2 to 5.4.7 present excellent visual images to
support the data in Table 5.4.7 and Figure 5.4.1. DIMODAN® RT 0.3% and 0.6%
samples have a small water droplet size and a narrow water droplet distribution.
The water droplets are seen via the green or dark colouring, since the protein sits at
or in the water phase, these appear together. Apart from Lesquerella (Figures 5.4.6
& 5.4.7) these systems are inherently stable. To help determine stability, sensory
properties are discussed later.
Figures 5.4.4 and 5.4.5 refer to the samples with Moringa at 0.3% and 0.6%
respectively, and it can be seen that the water droplet size is considerably larger in
Figure 5.4.4 at 0.3% as opposed to 0.6% - Figure 5.4.5. These images do not show
the very small water droplet structures as in Figure 5.4.3 (DIMODAN® RT 0.6%),
however, in Figure 5.4.5, where Moringa is at 0.6% concentration the CLSM
images are similar to those for DIMODAN® RT at 0.3%. This is also indicated in
the graphical data shown in Figure 5.4.1, and indicates that Moringa here is
behaving in a manner that is capable of forming stable spread products and this
high TAG concentration is in parity with DIMODAN® RT – albeit at a higher
concentration.
Figures 5.4.6 and 5.4.7, corresponding to Lesquerella at 0.3 and 0.6% respectively
show the presence of large areas of coalesced water droplets (lakes of water). This
was enough to confirm that this product was not stable and basically would fail
any storage, sensory or spreading test post production. In 60% W/O TAG spread
emulsions Lesquerella MAG is not recommended.
202
5.4.4.2 40% TAG based W/O Spreads
Use of Moringa MAG (E471) in low TAG (<41%) applications could represent a
cleaning up of the food product ingredient label since there would be potential
opportunity to omit a co-emulsifier, PGPR (E476). With this objective all MAG
based emulsifier components, plus PGPR were then tested in 40% TAG spreads
with the aim of assessing the effect of stabilising W/O low TAG content
emulsions. The water droplet size results (DSD) are shown in Table 5.4.8.
All samples relating to Lesquerella in 40% TAG spreads show the similar
tendency as for the results above for 60% TAG spreads i.e. the water droplet size
was so large that the emulsion samples have failed. When combined with a co-
emulsifier PGPR, the recovery and stabilisation of the water droplet size was
insufficient, leading to emulsion failure.
Moringa showed a trend that is also similar to the 60% TAG spreads, where in
40% TAG spreads, the droplet size where Moringa is present is smaller than all
samples with DIMODAN® UJ either alone, or in combination with PGPR. This
suggests from DSD analysis that the spreads made with Moringa were stable and
exhibit adequate function similar to standard palm or hydrogenated MAG products
for reduced or low TAG spreads.
203
Table 5.4.8 Water droplet size distribution for 40% TAG spread samples.
Figure 5.4.8 Water droplet size distribution (DSD) of W/O 40% TAG blends.
204
Results in Figure 5.4.8 show the samples having the tightest water droplets -
samples 12, 15, and 18; those with Moringa at 0.3%, Moringa at 0.6% and
Moringa 0.3% / PGPR 0.2% respectively. These results suggest the Moringa
containing spreads are likely to be stable.
The CLSM images of the 40% TAG spreads are given in Figures 5.4.9 – 5.4.17.
All the CLSM images of samples containing Lesquerella showed the presence of
very large lakes of water. Visually these are beyond droplet classification and can
be deemed as failed.
Sample 18, i.e. Moringa at 0.3% with co-emulsifier PGPR at 0.2%, had the
smallest droplet size from Figure 5.4.8. But of particular significance was the
CLSM images of Moringa at 0.6% – sample 15, which as a standalone, gave
comparatively, the smallest water droplet. Moringa at 0.3% had slightly larger
water droplets shown in Figure 5.4.8 and in fact showed the product itself was
stable and held up to both storage and spreadability tests. The same argument was
confirmed in the basic sensory results (5.4.4.4).
205
Figure 5.4.9 CLSM image of 40% TAG spread with DIMODAN® UJ at 0.3%
concentration. (Top images 375 x 375 µm. Bottom images 188 x 188µm)
206
Figure 5.4.10 CLSM image of 40% TAG spread with Moringa at 0.3%
concentration. (Top images 375 x 375 µm. Bottom images 188 x 188µm)
207
Figure 5.4.11 CLSM image of 40% TAG spread with Lesquerella at 0.3%
concentration. (Top images 375 x 375 µm. Bottom images 188 x 188µm)
208
Figure 5.4.12 CLSM image of 40% TAG spread with DIMODAN® UJ at 0.6%
concentration. (Top images 375 x 375 µm. Bottom images 188 x 188µm)
209
Figure 5.4.13 CLSM image of 40% TAG spread with Moringa at 0.6%
concentration. (Top images 375 x 375 µm. Bottom images 188 x 188µm)
210
Figure 5.4.14 CLSM image of 40% TAG spread with Lequerella at 0.6%
concentration. (Top images 375 x 375 µm. Bottom images 188 x 188µm)
211
Figure 5.4.15 CLSM image of 40% TAG spread with DIMODAN® UJ 0.3% /
GRINDSTED® PGPR 90 at 0.2% concentration. (Top images 375 x 375 µm.
Bottom images 188 x 188µm)
212
Figure 5.4.16 CLSM image of 40% TAG spread with Moringa 0.3% /
GRINDSTED® PGPR 90 at 0.2% concentration. (Top images 375 x 375 µm.
Bottom images 188 x 188µm)
213
Figure 5.4.17 CLSM image of 40% TAG spread with Lesquerella 0.3% /
GRINDSTED® PGPR 90 at 0.2% concentration. (Top images 375 x 375 µm.
Bottom images 188 x 188µm)
Using a Stable Micro-Systems TX2 analyser, the texture of the spread samples
was measured and treated together. The grouping on both low TAG and high TAG
spreads shown in Figures 5.4.18 (hardness) and Figure 5.4.19 (stickiness).
214
Hardness of high TAG 60% & low TAG 40% Jr. No -1
W/O emulsion spreads at 5°C Jr. No -2
Jr. No -3
500
450 Jr. No -4
400 Jr. No -5
Hardness (force /g)
350 Jr. No -6
300 Jr. No -11
250 Jr. No -12
200
Jr. No -13
150
Jr. No -14
100
50 Jr. No -15
0 Jr. No -16
Average Jr. No -17
Jr. No -18
Jr. No -19
Figure 5.4.18 The hardness of the high TAG (60%) samples – (1-6), and the low
TAG (40%) samples – (11-19)
Jr. No -1
Stickiness at 5°C
Jr. No -2
Average
Jr. No -3
0
Jr. No -4
-500 Jr. No -5
Stickiness (area)
Jr. No -6
-1000 Jr. No -11
Jr. No -12
-1500 Jr. No -13
Jr. No -14
-2000 Jr. No -15
Jr. No -16
-2500 Jr. No -17
Jr. No -18
Jr. No -19
Figure 5.4.19 The stickiness of the high TAG (60%) samples- (1-6), and the low TAG
(40%) TAG samples – (11-19)
215
Considering the 60% TAG samples first, it was clear that the DIMODAN® RT
spreads resulted in the greatest degree of hardness, but with the exception of
sample 2 where the DIMODAN® RT concentration is 0.6%, the hardness level of
the Moringa samples was broadly comparable. The hardness of the Lesquerella
samples was significantly less for all concentrations and was difficult to measure
because of total product failure.
For the 40% TAG spreads, samples 11-19 in Figures 5.4.18 and 5.4.19, the
hardness results showed that DIMODANT® UJ at 0.3% and Moringa at 0.3%
gave the highest results, higher than the 60% spreads. The remaining samples
showed hardness values similar to the 60% spreads, although the presence of
PGPR seemed to result in a decrease in hardness for both DIMODAN® UJ and
Lesquerella. Moringa and PGPR resulted in a much firmer product, and could
indicate a synergy, a co-emulsifier relationship taking place. In terms of stickiness
the same general trends were observed.
Sensory tasting of the 60% TAG spreads (1-6) by an expert panel resulted in the
following comments for the high TAG samples, 24hrs post production:
Sample 1. (DIMODAN® RT - 0.3%) Soft and smooth with pleasant overall mouth
feel.
Sample 2. (DIMODAN® RT - 0.6%) Firmer than sample 1 and generally still soft
and smooth. This resulted in a firmer mouth feel, but one that was still acceptable.
Sample 3. (Moringa - 0.3%) Soft and smooth with a pleasant overall mouth feel
Sample 4. (Moringa - 0.6%) Firmer than sample 3, softer than sample 2, and
smooth to taste but firmer over all mouth feel than sample 3.
216
Sample 6. (Lesquerella - 0.6%) Very soft, yellow in appearance, appeared
unstable.
The conclusions which were drawn for 60% TAG spreads, was that Moringa at
0.3% or 0.6% concentration can form acceptable stable spreads. The degree of
firmness achieved was comparable to the control (DIMODAN® RT).
Sensory tasting of the 40% TAG spreads (11- 19) by an expert panel resulted in
the following comments for the high TAG samples, 24hrs post production:
Sample 11. (DIMODAN® UJ - 0.3%) Very pleasant and smooth, good overall
mouth feel.
Sample 12. (Moringa – 0.3%) Very pleasant and smooth, good overall (thick)
mouth feel.
Sample 13. (Lesquerella - 0.3%) Appeared unstable.
Sample 14. (DIMODAN® UJ - 0.6%) Pleasant, smooth and good overall mouth
feel.
Sample 15. (Moringa – 0.6%) Very pleasant and smooth, good overall (thick)
mouth feel
Smaple 16. (Lesquerella – 0.6%) Unstable.
Sample 17. (DIMODAN® UJ 0.3% / GRINDSTED® PGPR 90 – 0.2%) Very
pleasant and smooth, good overall mouth feel.
Sample 18. (Moringa – 0.3% / GRINDSTED® PGPR 90 – 0.2%) Very pleasant
and smooth, good overall (thick) mouth feel.
Sample 19. (Lesquerella 0.3% / GRINDSTED® PGPR 90 – 0.2%) Unstable.
These results confirmed the results from earlier, relating to the water droplet size
and the CLSM images. Lesquerella failed in the 40% spread emulsions, as it did
for 60% spreads. However, Moringa performed well and provided both firmness
and good mouth feel characteristics to the emulsion in question. There was also
evidence of a potential synergistic interaction with PGPR.
217
5.4.4.5 Validation Testing
Only 40% TAG spreads were examined, and the samples of Lesquerella were
omitted due to obvious product failure (including 60%). The 40% TAG samples
were produced according to repeat formula / conditions as previously described in
Tables 5.4.3 and 5.4.4. Validation samples and conditions are described in Tables
5.4.5 and 5.4.6
The results tabulated in Table 5.4.9 show the validated water droplet size
distributions for samples 1 – 6 (Table 5.4.5) compared to earlier trial samples 11,
12, 14, 15, 17, 18 in Table 5.4.3.
Table 5.4.9 Validated water droplet size distribution (DSD) results (plain text) of
40% low TAG spreads. Figures in bold are the results from the original samples
(Table 5.4.8)
The DSD validation results for a 40% TAG spread in Table 5.4.9, compared well
with the previous results (shown bold within same table), showing that Moringa
performs equally as well compared to the initial trials in producing stable spreads
with small DSD. In each case the Moringa samples showed smaller DSD than the
218
initial trials. Moringa MAG used in the validation is the same identical Moringa
monoglyceride 191 as used in previous work.
Figure 5.4.20 shows graphically the data presented in Table 5.4.9, but without data
from Figure 5.4.8. However, a comparison clearly reveals that generally, all peaks
in this validation are in a narrow range. This was possibly attributed to absence of
TOCO 50 (antioxidant) a minor component (0.01% w/w) containing 50%
tocopherol, a wetting agent.
Figure 5.4.20 Water droplet size distributions for the 40% low TAG spreads of the
validation tests.
CLSM images of the validation studies are presented in Figures 5.4.21 to 5.4.26.
219
Figure 5.4.21 CLSM image of 40% TAG spread with DIMODAN® UJ - 0.3%
concentration. (Top images 375 x 375 µm. Bottom images 188 x 188µm)
220
Figure 5.4.22 CLSM image of 40% TAG spread with Moringa - 0.3%
concentration. (Top images 375 x 375 µm. Bottom images 188 x 188µm)
221
Figure 5.4.23 CLSM image of 40% TAG spread with DIMODAN® UJ - 0.6%
concentration. (Top images 375 x 375 µm. Bottom images 188 x 188µm)
222
Figure 5.4.24 CLSM image of 40% TAG spread with Moringa - 0.6%
concentration. (Top images 375 x 375 µm. Bottom images 188 x 188µm)
223
Figure 5.4.25 CLSM image of 40% TAG spread with DIMODAN® UJ - 0.3% /
GRINDSTED® PGPR 90 – 0.2% concentration. (Top images 375 x 375 µm.
Bottom images 188 x 188µm)
224
Figure 5.4.26 CLSM image of 40% TAG spread with Moringa - 0.3% /
GRINDSTED® PGPR 90 – 0.2% concentration. (Top images 375 x 375 µm.
Bottom images 188 x 188µm)
In comparing these CLSM images above (Figures 5.4.21 – 5.4.26) with the
original trials for the corresponding samples in Figures 5.4.9 to 5.4.17 there are
similarities, especially with the samples that contain Moringa. Taking Figures
5.4.10 and 5.4.22, both of which represent Moringa at 0.3% concentration, when
viewed together as a whole, there is similarity. In Figure 5.4.13 and 5.4.24,
225
Moringa at 0.6% concentration, it appears that Figure 5.4.24 has the smallest water
droplets, as indeed indicated by the water droplet size results (DSD). For Figures
5.4.16 and 5.4.26, where Moringa was present together with PGPR, again broad
similarities in the image were found. Similar trends were seen for the samples
which contain DIMODAN® UJ when compared against the original trials, with
the exception of DIMODAN® UJ in the presence of PGPR, where large lakes of
water are seen in Figure 5.4.15, which are not present in Figure 5.4.25. The
DIMODAN® UJ used in the validation trials came from a new commercial sample
than that used in the initial trials (Its material specification being consistent with
the first commercial sample). In Figure 5.4.25 the droplets are larger, but remain
as discrete droplets. This is possibly explained by the number of variations of
selected surfactants used in each trial, and which are run continuously, without any
separate product processing interface between each trial. Validation trials were
fewer in number, and as previously indicated excluded Lesquerella MAG.
The texture analysis results for the validation testing showing the hardness of the
validation samples at 5°C is given in Figure 5.4.27.
1
2
3
4
5
6
Figure 5.4.27 Textural hardness for the 40% low TAG spread validation tests,
where the order from left to right is as follows; DIMODAN® UJ, 0.3%; Moringa
0.3%; DIMODAN® UJ, 0.6%, Moringa, 0.6%; DIMODAN® UJ, 0.3% and
GRINDSTED® PGPR 90, 0.2%; Moringa 0.3% and GRINDSTED® PGPR 90,
0.2%.
226
The validation test confirmed the suggestion that Moringa MAG seems to be
outperforming DIMODAN® UJ and also in combination with PGPR, there is
small increased firmness. This could coincide with reduced DSD. But, only at the
lower concentration of 0.3% is the difference in textural firmness significant where
Moringa MAG is present.
5.4.5 Discussion
The water binding properties of PGPR are one of the reasons that it is essentially
the stabilising emulsifier of choice for many water – oil based system food
systems (refer to 1.2; 1.5). However, the level of stability that the PGPR can
confer is often such that any re-work of the system is made difficult, and this can
result in production down time. This has commercial implications because of the
negative effects of PGPR on emulsion breakdown characteristics and
227
consequently, this also has a powerful sensory impact on melting and flavour
release.
All samples were made with natural Moringa MAG. Proof of concept to ascertain
the performance of MAG compositions to mimic Moringa MAG is required.
Therefore, this would be systems based on blended fatty acid compositions to
match the fatty acid profile of naturally based (Moringa oleifera) Moringa MAG
to achieve similar functionality. This is plausible because it has already been
shown in bulk TAG model systems that such compositions were performing
similar to natural versions (refer to 5.3).
5.4.6 Conclusion
The results presented, show that Moringa MAG, when incorporated into a high
TAG spread (60%) or a low TAG spread (40%), is highly functional, resulting in a
commercially acceptable product which is both stable to processing and storage.
The formed emulsions had both acceptable commercial structure and flavour
release. Compared to other MAGs in this work, weight for weight, a Moringa
MAG conferred additional textural resilience and spreadability.
Further application trials would support the model system results (4.0); because it
was shown previously (5.3), that synthetic (blended composition) MAG samples
seemed to behave similarly to the naturally based (Moringa oleifera) Moringa
MAG samples, as used in this study.
228
5.5 Microscopy Examination of a Forced Cooled Model TAG Based System
Containing a Distilled Moringa MAG, Behenic Based MAG and PGPR
5.5.1 Introduction
Processing conditions usually run at faster cooling rates than 1°C per minute, but
typical examinations are often run at this much slower cooling rate. This is to
avoid excessive temperature gradients and therefore nullify any potential anomaly
that may occur. However, this approach may alter the perspective of the original
measurement when the cooling rate is changed closer to industrially relevant
cooling rates, as used in 5.4.
In a previous study (5.3), cooling rates of 1°C/min from the melt to 20°C resulted
with interesting dendrite crystal structure. Therefore the aim of this study is to
investigate the effect of forced cooling on the crystal structure of TAG based
systems where the cooling rate moves over two orders of magnitude, i.e. from
1°C/min to 100°C/min.
The control TAG blend was 70% Palm Stearin (IV35), and 30% Palm Olein
(IV56). Emulsifiers were added to the control sample at 1% of either Moringa
MAG, or GRINDSTED® Crystallizer 110 or 0.5% GRINDSTED® PGPR 90 as
per the previous study (5.3); the same method is also used. The microscopy was
run using an Olympus BX 60 with all images taken using PLM at x200
magnification. The samples were heated to 85°C to remove any previous crystal
229
history and then treated to the following cooling rates: 1°C/min, 10°C/min,
50°C/min, and 100°C/min to a final ambient temperature of 20°C. Refer to
General Materials & Methods (2.0) for additional detail.
The results in Figures 5.5.1 – 5.5.4 show the micrographs of the control sample,
CRY110, PGPR and Moringa MAG. Each Figure is split further into four
individual micrographs that correspond to the four different cooled temperature
rates.
In Figure 5.5.1, the control shows a slow cooling rate of 1°C/min the formation of
what can be described as clear evidence of TAG crystallisation; with pentagonal
and hexagonal snowflake (dendrite) like patterned crystal structures are very
apparent.
When the cooling rate is increased to 10°/min, the formation of dendrites become
smaller and less distinct; therefore, this is taken as clear indication that the forced
cooling at a faster rate as a significant effect on the kinetics of crystal formation.
At 50°C/min, the polarised image was poor and it was difficult to draw
conclusions, but at 100°C/min no evidence of dendrite structure is visible and the
crystals are smaller relative to those at 1°C/min.
Figure 5.5.2, shows the results over the same cooling rates for 1% of CRY110.
The polarised image shows homogenous distribution of many small crystals. A
similar micrograph image has been observed by Basso et al. (2010) when they
used a behenic based monoglyceride. However, the effect is dramatic in their
observation probably due to the degree of total saturation of the TAG solvent used
(Basso et al., 2010). Moving to faster cooling rates simply reduces the size of the
crystal structures progressively, at no point with CRY110 is there any indication of
dendrite or any other structural connotation.
230
Figure 5.5.3 shows results from 0.5% PGPR at 1°C/min cooling rate. The
polarised image is dominated by dramatic dendrite structures. There is no evidence
of small discrete crystals. In this respect it appears as though PGPR has completely
altered the crystallisation kinetics of the bulk system, crystal nuclei have
developed into fern-like dendrite structures of notable size in comparison to Figure
5.5.1.
Accelerating the cooling rate to 10°C/min, the large dendrite crystal structures
have reduced in number and size. The dark background, non-crystallised region
(liquid olein) as seen at 1°C/min is now dominated by homogeneous scattering of
mixed crystal shape and size. At 50°C/ min are not obvious and the crystal
distribution is now more uniform in size and shape; increasing the cooling rate still
further to 100°C/min, results in a structure similar to the polarised image as shown
in Figure 5.5.1 and 5.5.2 at the same cooling rate (100°C).
Lastly, Figure 5.5.4 shows the results for Moringa MAG. At 1°C/min there
appears to be a composite distribution across the polarised micrograph. Several
regions within image appear similar the micrographs at 1°C in Figure 5.5.2, and
Figure 5.5.3. This could suggest that Moringa MAG influences crystal behaviour
similar to both CRY110 and PGPR. There are areas where fern-like structures are
visible as well as areas where discrete crystal clusters are to be found. Increasing
the cooling rate to 10°C/min reduces the size of the crystals, but importantly the
fern-like structures and clusters are still present; they are more numerous
compared to the PGPR in Figure 5.5.3 at the same cooling rate. At a cooling rate
of 50°C/min the dendrite structures are arguably still visible, but the overall crystal
numeration is possibly more compared to either PGPR or CRY110. Increasing the
cooling rate to 100°C/min, results in a vastly reduced crystal structure and no
apparent indication of fern-like dendrite structures.
The relevance of these results is in the context of the aim, which is to approach
cooling gradients associated with typical production processes. However, this is
highly dependent on flow rates and the nature of the product itself, not least the
fact that alterations in the TAG solvent will immediately alter the thermal response
231
and hence solubility potential of forming crystal structures (Metin & Hartel 2005).
Conservative estimates suggest that the cooling rates on closed scrapped surface
heat-exchanger plants can be anything between 35°C/min., to 45°C/min. Unlike
environments associated with real process conditions, these observations were
performed under static conditions, as opposed to the dynamic cooling gradients
and shear conditions in real processes. Therefore, these results can be taken as
indication as to what is potentially happening in the dispersion or bulk TAG phase
of a W/O or O/W emulsion system at accelerated cooling rates. Until now, this
information was not found within the literature.
There are clear physical differences in the influence of crystal forming kinetics of
Moringa MAG, compared with a behenic rich MAG (CRY110), and consequently,
these may exert on performance and functionality. As discussed previously (5.2 &
5.3), Moringa MAG has portions of its fatty acid profile composed of C22:0 which
is the main component of CRY110 (~89%). Moringa MAG contains
approximately 5 - 8% of C22:0 which could be significant.
The polarised micrograph evidence also suggests parallels with the crystal forming
nature of PGPR. These parallel crystal formations, coupled with the parallel
functionality in high TAG (60%) and low TAG (40%) spreads, is supportive
evidence to commercial applications in a number of food products, containing a
TAG crystal network structure (Wassell et al., 2012a; 2012b; 2012c; 2012d; 2012e;
Bech et al., 2013).
Clearly, and most importantly, Moringa MAG shown in Figure 5.5.4 reveals
similar physical properties to PGPR shown in Figure 5.5.3. Theoretically, it is
possible that similar behaviour could be manifest at cooling rates experienced in
production process conditions, where these crystal structures may influence and be
transposed into the final products, so that if present, would influence the physical
properties of the bulk oil (refer to 5.4).
232
Figure 5.5.1 Micrographs of blank control exposed to forced cooling from 1 –
100°C/min (scale bars = 20µm).
233
Figure 5.5.2 Micrographs of CRY110 exposed to forced cooling from 1 –
100°C/min (scale bars = 20µm).
234
Figure 5.5.3 Micrographs of PGPR 90 exposed to forced cooling from 1 –
100°C/min (scale bars = 20µm).
235
Figure 5.5.4 Micrographs of Moringa MAG exposed to forced cooling from 1 –
100°C/min (scale bars = 20µm).
236
5.5.4 Conclusion
237
5.6 The Rheological Behaviour of Model TAG Systems Containing a
Behenic based MAG, PGPR and Moringa MAG, During Forced Cooling
Velocities
5.6.1 Introduction
The aim of this study is to investigate a model TAG blend under large scale
deformatory rheology and examine if there are similarities in behaviour between
PGPR 90 and Moringa MAG.
As in the previous studies (5.3; 5.5), the cooling gradient of the rheological
experiments ranged between 1°C/min to 30°C/min. With both cooling and shear,
an attempt to reach super-cooled conditions is sought, because of its effects on the
level of structure that occurs for a given rate of cooling (Ghosh & Rousseau 2009).
Schematically, it can be viewed as in Figure 5.6.1 which shows the energy barrier
to crystallisation and the relative amount of structure formed as a function of
cooling temperature (Section 1.3).
238
Figure 5.6.1 Schematic diagram of the energy barrier to crystallisation together
with the relative amount of structure formed for a given cooling rate.
Two TAGs blends with differing degrees of saturation were used as the solvent.
First a blend consisting of 70% palm stearin (35 IV) and 30% palm olein (56 IV),
then secondly a more unsaturated blend consisting of 70% palm olein (56 IV) and
30% palm stearin (35 IV), to which in both cases respectively, the emulsifiers
GRINDSTED® Crystallizer 110 (CRY110), GRINDSTED® PGPR 90 (PGPR),
and Moringa MAG (191) were added at 1%, 0.5% and 1%.
239
commencement of shear each sample was subject to a 2 minute waiting time with
the temperature held at 70°C. The TAG blend and selected emulsifiers were pre-
heated to 90°C to remove all crystallisation history prior to rheological
measurements. Refer to General Materials & Methods (2.0) for additional detail.
The results in Figure 5.6.2(a) and 5.6.3(c) show the entire curve for all the samples
cooled at the rate of 1°C/min, where the control sample represents the TAG blend
of 70% palm stearin / 30% palm olein and 70% palm olein / 30% palm stearin
respectively. The remaining samples are combinations of the TAG blend and the
emulsifiers under investigation: CRY110, PGPR 90, and Moringa MAG. Figures
5.6.2(a) and 5.6.2(c) show a large portion of the graph from 70°C to about 45°C
where there is only a gradual increase in viscosity. It is viewed that in this portion
of the graph essentially nothing is happening to the TAG based system due to the
fact that it is still in the melt, and therefore expected to behave in true Newtonian
fashion. Below 45°C however, dramatic increase in viscosity is observed
occurring at several temperatures depending on the nature of the sample. The
dramatic increase in viscosity is possibly the onset of crystallisation, and indeed
the values which can be taken from Figure 5.6.1 correspond well with previous
results for similar systems containing CRY110 (Young et al., 2008), where onset
temperatures for CRY110 in TAG based samples also occurred around 42°C –
45°C.
For clarity, the area of viscosity in Figures 5.6.2(b) and 5.6.3(c) show the
expanded section of Figures 5.6.2(a) and 5.6.3(c), where the data is expressed
from 50°C and cooler.
Figures 5.6.2(b) and 5.6.3(d) show that CRY110 alone begins the onset of
viscosity increase at 40°C and 38°C respectively, whereas the samples with
240
CRY110 and PGPR 90 or Moringa MAG show an onset at 42.5°C and 41°C
respectively. In isolation, Figure 5.6.2(b) shows PGPR 90 and Moringa MAG first
show an onset of viscosity rise at around 31-32°C, and the control sample
increases from 34°C. Figure 5.6.3(d) shows PGPR 90 and Moringa MAG with
onset of viscosity at approximately 30°C-34°C, and the control sample increases
from approximately 27°C.
The results from Figure 5.6.2(b) and 5.6.3(d) clearly suggest that when PGPR 90
and Moringa MAG are used in isolation, they both behave similarly, showing the
onset of viscosity increase to occur at essentially the same temperature. This
similarity in behaviour was suggested earlier (5.3; 5.5), where strong similarities in
crystal structure, (i.e. the presence of fern-like dendrite structures) were clearly
observed.
When CRY110 is mixed with a co-emulsifier; PGPR 90 and Moringa MAG, there
are obvious distinctions. The combination of CRY110 with PGPR 90 shows the
earliest onset of viscosity, occurring at 42.5°C. This is likely attributed to several
aspects discussed previously (see 4.0), namely, crystal structure and surface
activity respectively. The micrographs reported previously (5.3 & 5.5) showed
PGPR 90 has an inherent influence on crystal kinetics of TAG toward dendrite
structures - (which is likely governed by the “inherent” thermal response
behaviour of a given solvent); so that at a cooling rate of 1°C/min, a slow cooling
gradient, there is adequate time for PGPR molecules to influence crystal kinetics
(Marze 2009), such that PGPR molecules could act as a co-emulsifier (template),
to explain the early onset of the CRY110 / PGPR 90 system viscosity increase.
Observations by others confirm that PGPR is highly surface active (Rousseau,
2000) as is a behenic based MAG (Krog & Larsson, 1992). This interactive
relationship of CRY110 / PGPR 90 was observed earlier (see 4.0) and revealed
through thermal treatment that PGPR lowers interfacial tension – early and the
second emulsifier (CRY110) dominates the surface tension faster as the emulsifier
mixture approaches (Tγ). Then a dramatic fall of interfacial tension occurs once
the critical temperature is reached.
241
In trying to explain the mechanism in the anhydrous bulk, it is known that PGPR
adsorbs at polar sites from anhydrous solution. In one example Dedinaite &
Campbell (2000) showed that mixture of phospholipids (PE) and PGPR in a
anhydrous liquid vegetable oil (triolein), facilitated the adsorption of non-polar PE
crystals on a polar surface. In this study two TAG based solvents were used having
differing degrees of saturation. Dedinaite & Campbell (2000) explain how PGPR
is able to induce crystal kinetics in bulk TAG blends containing a smaller
molecular weight surfactant (e.g. behenic MAG). The onset of viscosity occurs at
higher temperatures with the CRY110 / PGPR 90 combination than with CRY110
alone, or in combination with Moringa MAG.
Figure 5.6.2(b) shows CRY110 and Moringa MAG have onset of viscosity
increases which begins at a slightly cooler temperature of 41°C. Approaching
38°C, viscosity builds almost one order of magnitude more compared with
CRY110 alone. The reason for this is possibly explained from the evidence in the
work reported previous (5.5), which showed that both Moringa MAG and PGPR
strongly influenced TAG crystals towards dendrite structures (5.5 also used the
same TAG solvent as in this study). Both Moringa MAG and PGPR appear to
have a similar physical response pattern. This similarity may possibly explain why
Moringa MAG and PGPR both behave according to a similar mechanism when
combined with CRY110 (Wassell et al., 2012a; 2012b; 2012c; 2012d; 2012e; Bech et
al., 2013).
242
Figure 5.6.2 a / b Viscosity cooling curves at 1°C/min of base TAG blend (70%
palm stearin/ 30% palm olein) [dark blue]; 1% CRY110 [green]; 1% CRY110 /
0.5% PGPR 90 [red]; 0.5% PGPR 90 [light blue]; 1% CRY110 / 1% Moringa
MAG [pink]; and 1% Moringa MAG [yellow]
243
All treatments compared: 1°C/min
All treatments compared: 1°C/min (Zoom)
10 5
15.0
Eta 1
10 Eta
Jr16904 #31 1°C/min
4 Jr16904 #31 1°C/min
10 Jr16904 #32 1°C/min
Jr16904 #33 1°C/min Jr16904 #32 1°C/min
Jr16904 #34 1°C/min Jr16904 #33 1°C/min
Jr16904 #35 1°C/min Jr16904 #34 1°C/min
3 Jr16904 #36 1°C/min Jr16904 #35 1°C/min
10 Jr16904 #36 1°C/min
)
0
)
10 2 10
[Pa-s ]
[Pa-s ]
Eta (
10 1
Eta (
10 0
10 -1
244
10 -1
Figure 5.6.4(a) and 5.6.4(b) show the same rheological cooling profiles for the
same samples, measured at a faster cooling rate of 10°C/min. The response
behaviour when compared to the slower cooling rate of 1°C/min in Figures
5.6.2(a) and 5.6.2(b), shows on closer examination, a delay in onset temperature to
lower values in Figure 5.6.4(a) and 5.6.4(b).
Figure 5.6.4(b) reveals Moringa MAG and PGPR 90 behave similarly, until a
temperature of 28°C whereupon differences in their profiles are seen. This
manifests itself as a large ‘kink’ in the profile shape, where there is first a lowering
of viscosity followed by another rise in viscosity. Given that the rheometer
geometry is rotating throughout the measurement, such a profile may be attributed
to an initial build up of viscosity (i.e. structure) which reached a critical point and
is then broken by the continual applied shear (Corke, 2007). Quite why this should
be observed for the Moringa MAG and PGPR 90 and not the control sample of the
base TAG blend alone is not altogether clear. It is possible that the presence of
PGPR may influence molecular kinetics in the bulk oil phase, it being highly
surface active (Rousseau, 2000; Claesson et al, 1997). If this in turn caused
different dielectric parts of the mixed TAG systems to have irregular crystal
shapes – dendrites (Ghosh & Rousseau 2010), this together with the presence of a
nano-crystalline formation adhered by non-polar polyricinoleate part of the PGPR
may cause adhesive properties (Dedinaite & Campbell, 2000).
Shearing through was also not observed in the slower cooling rate of 1°C, and this
might be attributed to the slower cooling rate allowing the structure more time to
arrange such that breakage or adhesive properties are stronger (Marze 2009).
This would then tend to suggest that the faster cooling rates lead to weaker
structures. However, closer examination of Figures 5.6.2(a) and 5.6.4(a) show that
the final viscosity value of the sample cooled at 10°C/min in fact gives the highest
viscosity. Similar cooling rate structural observations, i.e. greater structure
development after gelation onset has been reported in the literature (Ojijo et al,
245
2004). This suggests that although faster cooling rates may influence differences in
either crystallisation onset or shearing through, this does not equate to lower levels
of final structure in the final application (Toro-Vazquez et al, 2001).
Clearly, the influence of cooling and its consequential effects to the bulk reaching
equilibrium conditions cannot be underestimated. Marze, (2009) demonstrated the
effects of PGPR concentrations in sunflower oil, and observed behaviour is
controlled by long relaxation times, whereas shorter time relaxation becomes
manifest, and takes over when approaching and exceeding the saturation of
interfacial concentration. This was reported and understood as a shift from a
diffusion-dominated (long process) regime to a rearrangement-dominated (faster
process) regime. The significance of this is important and enhances appreciation
for, other environmental conditions affecting the bulk, e.g. temperature ramp,
solvent type, degree of water activity (Wells, 1998).
Figure 5.6.4(b) with CRY110, shows that the onset of the viscosity rise –
essentially gelation – has reduced to between 36°C and 37.5°C at this faster
cooling rate of 10°C/min, compared to the onset temperatures of 41°C to 42.5°C in
Figure 5.6.2(b). The same trend observed in Figure 5.6.2(b) is seen that the
crystalliser sample with PGPR 90 still has the onset at the highest temperature,
here being 37.5°C compared to the crystalliser sample with Moringa MAG (36°C).
The profile for the CRY110 and Moringa MAG/ CRY110 were essentially the
same throughout the cooling regime of 10°C/min, contrary to the results obtained
for the slower cooling rate in Figure 5.6.2(b).
The observation that the CRY110 based samples are showing delayed onset of
gelation – or viscosity increase – at the higher cooling rate of 10°C/min is
understood as super-cooling. Specifically for monoglycerides, any increase in
cooling rate, results in crystallisation temperature below normal, due to ‘super-
cooling’ effect (Krog, 2001). There is confirmation of this seen in Figures 5.6.5
and 5.6.6, which show the rheological curves at the accelerated cooling rate of
30°C/min.
246
5.6.3.3 Cooling Velocity - 30°C/min
In the case of higher saturation and faster cooling rates (Figure 5.6.5), the Moringa
MAG and PGPR 90 still show behavioural patterns that are the same, (as seen at
lower temperature gradient), and basically follow the control sample, but the
viscosity increase has dropped to lower temperatures. When used in combination
with CRY110 the onset of viscosity increase – gelation, has also moved to lower
temperatures, but still maintaining a distinct gap from the control and single
samples.
The new onset temperature for the combined Moringa MAG with CRY110 and
PGPR 90 with CRY110 is now the same at 32°C. There appears to be no
difference between the sample containing PGPR 90 or Moringa MAG.
Figure 5.6.6 (less saturation faster cooling rate) shows similar response behaviour.
However, it is clear that examination of the Moringa MAG and PGPR 90 as single
additives cause quicker induction rates compared to the control (without additive).
Benefits are also seen again when combining Moringa MAG with CRY110 and
PGPR 90 with CRY110.
The increased cooling rate has exerted influence on the onset temperature of the
control sample. The data in Figures 5.6.5 and 5.6.6 show, the combined samples
and the control and individual samples converge at the lower temperatures and
merge. Therefore, this would seem to indicate a physical limit is being reached. At
higher cooling rates the non-isothermal crystallisation temperature is reached at
shorter periods, but in this case, the time it takes to reach the critical lamellar size
for nucleation is prolonged. (Toro-Vazquez et al, 2001).
The results in Figures 5.6.2, 5.6.3, 5.6.4, 5.6.5 and 5.6.6 are very specific for the
given model TAG blends and emulsifier concentrations, as well as the selected
process conditions. Both Toro-Vazquez et al (2001) and Fredrick et al (2008),
suggest conditions ultimately dictate the induction time for crystallisation, and
therefore will change for other TAG blends. Ojijo et al (2004) goes on to report
that at slow cooling rates sufficient time is available for lamellar structures to
247
form, these being the precursors to crystal nuclei, and they can form to such an
extent as to reach a given critical size for nucleation. This critical size is then
reached before the isothermal temperature is reached, whereas at faster rates of
cooling much of the lamellar organisation and formation takes place at the
isothermal temperature – and therefore more time is required to reach the critical
lamellar size. In this case, longer time results in a lower temperature being
reached before any onset of viscosity increase is seen, and hence the apparent
delay in onset temperatures.
The cooling rate and the solvent have considerable impact and bearing on the
mechanical and physical properties, including aspects like stability of the final
network structure. In this study two palm oil fractions palm stearin (IV35) and
palm olein (IV56), were used at two ratios’ to achieve differing unsaturation.
Other ratio may have given a different outcome. Studies of palm oil and its
numerous fractions show that crystallisation behaviour changes depending on the
degree of supercooling (Fredrick et al, 2008).
When comparing the effect of increased cooling rates with their consequences on
the structures to the results from the work in 5.5, there is agreement, in that during
higher cooling rates the TAG crystals formed were smaller at the same given
temperatures than was the case for lower cooling rates.
248
All samples 10°C per min All samples 10°C per min
10 4 10 0
Eta Eta
All samples 10°C per min All samples 10°C per min
Bipartite_13_10_min Bipartite_13_10_min
103
Bipartite_14_10_min Bipartite_14_10_min
Bipartite_15_10_min Bipartite_15_10_min
Bipartite_16_10_min Bipartite_16_10_min
Bipartite_11_10_min Bipartite_11_10_min
102
)
)
[Pa-s ]
[Pa-s ]
-1
Eta (
Eta (
10
0
10
249
-1
10
-2
10 0.02
20.0 30.0 40.0 50.0 60.0 70.0 80.0 20.0 25.0 30.0 35.0 40.0 45.0
Temp [°C] Temp [°C]
palm stearin/ 30% palm olein) [yellow]; 1% CRY110 [green]; 1% CRY110 / 0.5%
Figure 5.6.4 a / b Viscosity cooling curves at 10°C/min of base TAG blend (70%
250
Figure 5.6.6 Viscosity cooling curves at 30°C per minute of base fat blend (70%
palm olein / 30% palm stearin) [green]; 1% CRY110 [red]; 1% CRY110 / 0.5%
PGPR 90 [light blue]; 0.5% PGPR 90 [yellow]; 1% CRY110 / 1% Moringa MAG
[pink]; and 1% Moringa MAG [dark blue]
It is known that addition of hard material, i.e. CRY110 based on C22:0 can, under
specific usage and process conditions, induce template nucleation (Wassell et al
2010). The crystalliser material can ‘anchor’ the surface, but at high dosages, act
as a destabilising influence, especially in low saturated TAG systems (Low fat
emulsions). However, if a less saturated material were added which still possessed
sufficient ratio of saturation : unsaturation to interact with a PGPR dominated
surface or have characteristics similar to PGPR, then additional surface
strengthening might occur. Based on model systems and application trials in
sections 5.3; 5.4 and 5.5, Moringa MAG is such a material.
251
Coinciding with this study, it has previously been demonstrated (5.7; 5.8) that
Moringa MAG is capable of producing stable W/O low fat emulsions where
addition of CRY110 or even PGPR (5.8) alone does not. This could suggest that
Moringa MAG may fit into the schematic diagram given in Figure 5.6.7 in
potentially two distinct manners.
Water phase
Interface
heterogeneous nucleation
fat
crystals
Polyglycerol Template of
Additive
Polyricinoleate additive
Figure 5.6.7 Additive to induce crystal nucleation, whilst retarding crystal growth.
From Wassell, Bonwick, Smith, Almiron-Roig, and Young (2010)
In Figure 5.6.7 the blue PGPR molecules are dominating the surface and only the
presence of the co-emulsifier (CRY110) at a Tγ, influences surface kinetics and
TAG crystals. Moringa MAG could fulfil combined roles outlined in this
schematic model.
252
5.6.4 Conclusion
Previous studies (3.0; 4.0; 5.2; 5.3; 5.4; 5.5) along with this work should be used
as part of a composite conclusion, which suggests that Moringa oleifera based
MAG is capable of performing a bi-functional role - template nucleation as well as
surface stabilising.
These results have established a clear link between cooling rate and onset of
gelation (viscosity increase) of both saturated and less saturated TAG systems. At
a cooling rate of 1°C/min the onset for PGPR 90 / CRY110 was 42.5°C, whereas
at a cooling rate of 10°C/min was 37.5°C. The corresponding onset temperatures
for Moringa oleifera based MAG / CRY110 were 41°C and 36°C respectively. At
a cooling rate of 30°C/min the onset temperature has fallen further to 32°C.
This study showed that a Moringa oleifera based MAG and PGPR 90 behave very
similarly when used in the given TAG blend alone, but in combination with
CRY110 differences do occur within cooling rates of 1°C/min to 30°C/min.
Acceleration of gelation was more evident with CRY110 combined with PGPR in
the less saturated TAG blend (70% palm olein / 30% palm stearin). These
differences are explained via difference in surface activity because of either
diffusion and or rearrangement of molecular geometry of the tested surfactants
(Bergenståhl 2008). Although a strict relation between surface activity and cooling
rate is sought, it is realistic to recognise the complexity of many environmental
conditions. At the highest cooling rate studied, 30°C/min, no difference was found
between the Moringa oleifera based MAG or PGPR 90 when used in combination
with CRY110 in the highly saturated TAG blend. The convergence of the curves at
the cooler temperature, the final diffusion behaviour and or rearranged interfacial
structures resulting from addition of surfactants cannot be quantified within the
scope of this measurement.
A study at the same cooling rate (30°C), in a less saturated TAG blend, revealed
the same gelation patterns are as distinct to those observed for 1°C/min, although
at lower onset temperature.
253
5.7 Impact of Fatty Acid Profile: Distinctions Between Moringa MAG and
Behenic Based MAG in Low TAG W/O Emulsions
5.7.1 Introduction
A previous study (5.4) of W/O low TAG spread emulsions with Moringa MAG
concentrated solely on the functionality of the Moringa MAG as being a
convincing functional ingredient for commercial applications (Wassell et al., 2012a;
2012b; 2012c; 2012d; 2012e; Bech et al., 2013). The effectiveness of Moringa MAG
may in part be due to its natural fatty acid profile consisting of C20:0, C22:0, and
>C24:0.
Both CRY110 and PGPR 90 were used as comparisons in previous studies (5.2;
5.3; 5.4; 5.5; 5.6) for several reasons. First, the behenic (C22:0) based MAG
(CRY110) is characterised by its high C22:0 content. Moringa MAG is 6 – 8%
C22:0 (total >C20:0 = ~10%) which may be unusually high for a natural MAG
(Wassell et al., 2012a; 2012b; 2012c; 2012d; 2012e; Bech et al., 2013). Second, the
PGPR is normally used in conjunction with a second emulsifier (1.6; 1.7) for low
TAG emulsions (<41%). Studies in this thesis have revealed Moringa MAG to
have unusual rheological (5.2; 5.6), crystallisation (5.3; 5.5) and interfacial (4.0)
similarities to PGPR.
It is known that to stabilise a low TAG W/O emulsion (<50%) using fully
saturated MAG is not conducive to stability; especially where a fully saturated
and/or longer chain saturated fatty acid e.g. C22:0 (behenic acid) is used.
However, evidence in this thesis (1.5) has shown the presence of a C22:0 rich
MAG (CRY110) can and does aid structuring performance and in low TAG
systems, provided dosage, process conditions and ingredient balance is correct –
primarily this mechanism is linked to the presence of the C22:0 fatty acids
(Sakamoto et al, 2003; Wassell & Young 2007; Wassell et al., 2010a).
Therefore, assuming that Moringa MAG seems to have a bi-functional effect (5.6),
it is necessary to distinguish by application tests, a comparison of only a C22:0
based MAG (CRY110) and Moringa MAG in low TAG emulsion systems.
254
5.7.2 Materials and Methods
Two TAG concentrations were studied, 35% and 40%/, in W/O emulsions
formula’s given in Tables 5.7.1 and 5.7.2. In the 35% TAG samples (Table 5.7.1)
the water phase is empty, i.e. does not contain hydrocolloid thickeners and
represent a spread that is “stressed”. In the 40% TAG spreads (Table 5.7.2) the
water phase contains GRINDSTED® LFS 560 Stabiliser System (Pectin / Alginate
blend). The plant process conditions are subsequently given for the 35% and 40%
TAG samples in Table 5.7.3, and were the same in each case. The Moringa MAG
191 is the same specification as described previously (5.3 - Tables 5.3.4 & 5.3.6),
see Table 5.7.4. The emulsifier concentrations were 0.15, 0.3, 0.6 and 1.2%. The
reliability of Moringa MAG at 0.3% and 0.6% is already confirmed positive (5.4),
but the extreme low 0.15% and high 1.2% concentration had not been tested until
now. Refer to General Materials & Methods (2.0) for emulsion assembly
procedure.
Ingredients in %
Ingredient Name 21 22 23 24 25 26
Water phase
Water (Tap) 64.000 64.000 64.000 64.000 64.000 64.000
Salt (Sodium Chloride) 1.000 1.000 1.000 1.000 1.000 1.000
Butter Flavouring 050001 T03007 0.010 0.010 0.010 0.010 0.010 0.010
Water phase total 65.010 65.010 65.010 65.010 65.010 65.010
pH 5.5 5.5 5.5 5.5 5.5 5.5
Fat phase
Fat blend
PK4 - INES 25.000 25.000 25.000 25.000 25.000 25.000
COLZAO 75.000 75.000 75.000 75.000 75.000 75.000
Fat blend total 100.000 100.000 100.000 100.000 100.000 100.000
Other fat ingredients
Moringa. monoglyceride 191 0.150 1.200
GRINDSTED® CRYSTALLIZER 110
0.150 0.300 0.600 1.200
Distilled Monoglyceride
2% sol. beta-carotene 0.020 0.020 0.020 0.020 0.020 0.020
Butter Flavouring 050001 T04184 0.020 0.020 0.020 0.020 0.020 0.020
Other fat ingredients total 0.190 0.190 0.340 0.640 1.240 1.240
Fat phase total 34.990 34.990 34.990 34.990 34.990 34.990
RECIPE total (calc. batchsize) 100.000 100.000 100.000 100.000 100.000 100.000
255
Table 5.7.2 Formula of spread samples at 40% TAG content.
Ingredients in %
Ingredient Name 11 12 13 14 15 16
Water phase
Water (Tap) 57,300 57,300 57,300 57,300 57,300 57,300
Salt (Sodium Chloride) 1,000 1,000 1,000 1,000 1,000 1,000
Skimmed milk powder (MILEX 240) 0,100 0,100 0,100 0,100 0,100 0,100
GRINDSTED® LFS 560 Stabiliser System 1,500 1,500 1,500 1,500 1,500 1,500
Potassium Sorbate 0,100 0,100 0,100 0,100 0,100 0,100
Butter Flavouring 050001 T03007 0,010 0,010 0,010 0,010 0,010 0,010
Water phase total 60,010 60,010 60,010 60,010 60,010 60,010
Ph 5,5 5,5 5,5 5,5 5,5 5,5
Fat phase
Fat blend
PK4 – INES 25,000 25,000 25,000 25,000 25,000 25,000
COLZAO 75,000 75,000 75,000 75,000 75,000 75,000
Fat blend total 100,000 100,000 100,000 100,000 100,000 100,000
Other fat ingredients
Moringa, monoglyceride 191 0,150 1,200
GRINDSTED® CRYSTALLIZER 110
0,150 0,300 0,600 1,200
Distilled Monoglyceride
2% sol. beta-carotene 0,020 0,020 0,020 0,020 0,020 0,020
Butter Flavouring 050001 T04184 0,020 0,020 0,020 0,020 0,020 0,020
Other fat ingredients total 0,190 0,190 0,340 0,640 1,240 1,240
Fat phase total 39,990 39,990 39,990 39,990 39,990 39,990
RECIPE total (calc. batchsize) 100,000 100,000 100,000 100,000 100,000 100,000
Table 5.7.3 Pilot plant processing conditions for the formula’s in Tables 5.7.1 and
5.7.2
Refer to General Materials & Methods (2.0) for emulsion assembly procedure.
256
Table 5.7.4 Fatty acid profiles for Moringa MAG 191 and CRY110 (5.3 - Tables
5.3.4 & 5.3.6)
The methods of analysis for water droplet size distribution (DSD) and confocal
laser scanning microscopy (CLSM), scaled to 375 x 375µm are outlined in 2.0. All
samples are stained with FITC, stains protein green, and Nile Red, stains fat red.
Texture analysis was outlined previously (5.3). Photographic images were
recorded using a Canon G12.
257
5.7.3 Results and Discussion
Table 5.7.5 Water droplet size distribution (DSD) for 40% TAG spreads (samples
21-26), and 35% TAG spreads (sample 11-16).
Average/
Sample ID 2.5% <μm 50% <μm 97.5% <μm
St.dev
11 Average 1.08 5.38 26.80
St.dev 0.02 0.07 0.54
12 Average 1.10 5.62 28.80
St.dev 0.05 0.03 1.14
13 Average 0.84 6.50 50.41
St.dev 0.04 0.13 2.75
14 Average 0.60 10.14 171.08
St.dev 0.04 0.20 17.17
16 Average 2.01 3.64 6.58
St.dev 0.09 0.02 0.36
21 Average 0.23 3.46 51.73
St.dev 0.01 0.07 5.26
22 Average 0.58 3.81 24.82
St.dev 0.03 0.06 1.16
23 Average 0.91 10.20 115.23
St.dev 0.05 0.75 21.16
24 Average 1.01 21.66 481.56
St.dev 0.05 3.66 196.53
25 Average 0.85 23.01 665.20
St.dev 0.13 4.21 346.97
26 Average 3.48 3.48 3.49
St.dev 0.01 0.01 0.01
The results presented in Table 5.7.5 show the water droplet size distribution for the
35% TAG spreads (samples 21-26) and the 40% TAG spreads (samples 11-16).
Sample 15 could not be measured due weak to signal. Samples 21, 22, 23 and 24
covering the 35% TAG spreads were phase separated (except sample 26);
abundance of liquid oil was in the bottom of the 150g sample container. This
observation indicates instability. An indicator of instability is revealed by large
water droplet size distribution (DSD).
The clear conclusion that is drawn from the results given in Table 5.7.5 is that the
size of the water droplets for all samples containing CRY110 are large and
258
therefore the spread samples are prone to instability, and hence separation. This
was found irrespective of TAG content, either 35% or 40%, although the samples
at 40% were distinctly more stabilised, likely because of 5% more TAG content
and aided by stabiliser in the water phase.
A different situation was apparent for the samples containing Moringa MAG.
These samples generally showed a stable performance, with the exception of
sample 22, which contained Moringa MAG at 0.15% dosage in the 35% TAG
spreads. This was the most stressed of the Moringa MAG containing samples since
the water phase of this spread was empty, i.e. no hydrocolloid thickener. Sample
22 showed clear phase separation and therefore instability. In a 35% W/O
emulsion, and empty water phase, it is possible that 0.15% Moringa MAG is the
lower inclusion limit.
At the higher Moringa MAG inclusion level sample 26, dosed at 1.2% for the
same 35% TAG spread with empty water phase, resulted in a dramatic reduction
of the water droplet size, and no phase separation, adding positively to previous
results (5.4).
In the 40% TAG spreads, with the water phase stabilised with GRINDSTED®
LFS 560 Stabiliser System, Moringa MAG dosed at 0.15% (sample 11) showed
water droplet sizes of 26.8, which was enough to provide a stable emulsion,
whereas when the dose is increased to 1.2% (sample 16) the water droplet size
dropped to 6.58, and the level of stability increased.
In Figures 5.7.1a to c the cardboard test is seen for the samples at 40% TAG
content with a stabilised water phase. Sample 11, containing Moringa MAG at
0.15% produced a thick and creamy emulsion that was stable, and acceptable to
259
spread testing. There was no adverse sign of emulsion breakdown or leakage of
water.
The results suggest that the optimum Moringa MAG dose for low TAG W/O
emulsions lies between 0.15% and 1.2%. As demonstrated previously (5.4) where
a co-emulsifier (PGPR) might be added, it is now shown that PGPR is not
required.
260
Figure 5.7.1, (1a) Spread test for samples 11 and 12; (1b) spread test for samples
13 and 14; (1c) spread test for samples 15 and 16; (1d) samples in storage tub for
samples 21, 22, 23, 24, 25 and 26; (1e) spread test for sample 26
261
5.7.3.1 CLSM
The data presented in Figures 5.7.2 a to f (samples 11 to 16), and Figures 5.7.3 a to
f (samples 21 to 26) show the confocal laser scanning microscopy (CLSM) images
for the samples of full water phase and a TAG content of 40%, and empty water
phase and a TAG content of 35% respectively. Images are reproduced from x40
magnification and scaled to 375 x 375 µm.
For Figure 5.7.2 a and f, which contain Moringa MAG at 0.15% and 1.2%
respectively, the CLSM images show a compact-like water droplet size
distribution. This is indicative of a stable emulsion.
In terms of a maximum Moringa MAG limit (1.2%), the CLSM image in Figure
5.7.3 f (sample 26), appears unstable. It is suggested this is possibly an artefact of
an O/W/O emulsion, where Moringa MAG has gone beyond critical micelle
concentration, and formed a separate laminar phase. Figure 5.7.1 e (Sample 26)
shows a highly stable white emulsion, demonstrated by the tight emulsion (97.5%
DSD = 3.49).
262
Figure 5.7.2 (2a – 2f) CLSM images of low TAG (40%) spread samples 11 to 16,
with stabilised water phase
263
Figure 5.7.3 (3a – 3f) CLSM images of low TAG (35%) spread samples 21 – 26
with empty water phase
264
5.7.3.2 Texture Analysis
11
12
13
14
15
16
Figure 5.7.4 Hardness results: texture analysis for samples 11 to 16, 40% TAG,
full water phase
26
Figure 5.7.5 Hardness result: texture analysis for sample 26, 35% TAG, empty
water phase.
265
Figure 5.7.4 shows (samples 11 – 16) a gradual increase in hardness, i.e. moving
from Moringa MAG at 0.15% and CRY110 from 0.15%, 0.3%, 0.6% and 1.2%
respectively. The emulsion firmness at the higher crystalliser concentrations is
harder and this could be attributed to the continued water leakage from the
emulsion making the solid part of the sample appear harder than would otherwise
have been. At 1.2% sample 16 (Moringa MAG) is much softer. It is worth noting
that despite this leakage of water being attributed for the increase in hardness, the
level of water leakage has not led to the catastrophic failure of the systems
represented in samples 21 to 25. It can be seen that Figure 5.7.5 only has data for
one sample, the Moringa MAG containing Sample 26 at 1.2% dosage. All other
samples in this range failed and were not possible to measure. Interestingly, the
effect of Moringa MAG at 1.2% in either the 35% empty water phase or 40%
hydrocolloid-protein enriched water phase W/O emulsion gives basically the same
textural force response.
Given the water leakage apparent for the samples in Figure 5.7.4, it is difficult to
assign these textural results to a particular trend, as they are perhaps best described
as average apparent values.
5.7.4 Conclusion
The results in this study have differentiated the distinction between a natural non-
hydrogenated monoglyceride based on Moringa oleifera TAG and a fully saturated
long chain monoglyceride based mainly on C22:0, (GRINDSTED® Crystallizer
110), in W/O low TAG emulsions. The results show that low TAG spreads cannot
be adequately stabilised by behenic based (89% C22:0) MAG (CRY110) alone in
either normal full, or empty water phase environments at 40% or 35% TAG
content. All results show there is water leakage resulting in breakdown of the
emulsion or indeed complete emulsion failure.
In contrast to this, Moringa MAG were shown to be able to stabilise the emulsions
in the full water phase 40% TAG content systems at dosages between 0.15% and
266
1.2%, with the optimal being in between this range (5.4). These systems did not
exhibit water leakage, were stable and spreadable. At a high concentration of
1.2%, a tendency towards over stabilisation resulted, leading to the inability of the
emulsion to give good flavour release. When stressing the systems further by
reducing the TAG content and keeping the water phase empty, the Moringa MAG
at 0.15% dosage could not stabilise the spread. This is possibly an artefact of
dosage, where a dose of 0.15% is below the lowest limit of securing stability,
while a 1.2% dosage leads to over stabilisation.
There is clear distinction between two distilled MAG, one both natural, non-
hydrogenated (Moringa MAG), and one fully hydrogenated (CRY110). Moringa
MAG is able to stabilise low TAG spread emulsions, which might be attributed in
part to the presence of its approximately 30% range of saturated fatty acids
including C16:0, C18:0, C20:0, C22:0, C24:0 and C26:0, but clearly not
exclusively because of this, but also because of a significant proportion being
>C20:0. Possibly there are also other aspects of a Moringa oleifera based MAG
structure and combination of fatty acids (~70% C18:1) that allow it to stabilise low
TAG spread emulsions, where as a single emulsifier, CRY110 (essentially ~85% -
90% C22:0) does not. In the product applications and at the concentrations tested
in this investigation, the addition CRY110 as a single emulsifier in low TAG W/O
emulsions led to instability.
267
5.8 Effect of PGPR Concentration in Low TAG W/O Emulsions
5.8.1 Introduction
For comparison, studies in this thesis have revealed that a MAG based on Moringa
oleifera TAG has unusual rheological (5.2; 5.6), crystallisation (5.3; 5.5; 5.6) and
interfacial properties (4.0) similar to PGPR. The evidence suggests that the
optimum Moringa MAG dose for low TAG W/O emulsions is positioned between
0.15% and 1.2%. As demonstrated previously (5.4 & 5.7) where a co-emulsifier
(PGPR) might be added, it was shown that PGPR is not required. However to
support the argument that Moringa MAG seems to have a bi-functional effect
(5.6), it was necessary to distinguish by application tests, a comparison solely of a
behenic based MAG (CRY110) and Moringa MAG as single emulsifiers in low
TAG emulsion systems (5.7). Goubran and Garti (1988) reported the benefit of
using high molecular weight polyglycerols (PGPR) and recognised that while
much had been reported on their behaviour in O/W emulsions, there appeared to
be little reported regarding their behaviour in W/O emulsions.
The effect of PGPR as a single emulsifier is now tested in the context of the
maximum permitted dose of PGPR in food emulsions is 0.4 wt%, and specifically
it is limited in use to emulsions of less than 41% fat/oil (permitted use of E476 -
EC Directive 95/2/EC). The aim in this study is to distinguish the effect PGPR 90
as a single emulsifier system at varying concentration in W/O 40% low TAG
emulsions in comparison with a Moringa MAG tested previously (5.6; 5.7).
(Wassell et al., 2012a; 2012b; 2012c; 2012d; 2012e; Bech et al., 2013).
268
5.8.2 Materials and Methods
Ingredients in %
Ingredient Name 61 62 63 64
Water phase
Water (Tap) 57.300 57.300 57.300 57.300
Salt 1.000 1.000 1.000 1.000
Skimmed milk powder (MILEX 240) 0.100 0.100 0.100 0.100
GRINDSTED® LFS 560 1.500 1.500 1.500 1.500
Potassium Sorbate 0.100 0.100 0.100 0.100
Butter Flavouring 507104 A 0.010 0.010 0.010 0.010
Water phase total 60.010 60.010 60.010 60.010
pH 5.5 5.5 5.5 5.5
Fat phase
Fat blend
PK4 - INES 25.000 25.000 25.000 25.000
Rapeseed oil 75.000 75.000 75.000 75.000
Fat blend total 100.000 100.000 100.000 100.000
Other fat ingredients
PGPR 90 0.150 0.300 0.600 1.200
2% sol. beta-carotene 0.020 0.020 0.020 0.020
Butter Flavouring 050001 T04184 0.020 0.020 0.020 0.020
Other fat ingredients total 0.190 0.340 0.640 1.240
Fat phase total 39.990 39.990 39.990 39.990
RECIPE total (calc. batchsize) 100.000 100.000 100.000 100.000
269
Table 5.8.2 Processing parameters used for formula in Table 5.8.1
Pilot Plant
Processing (3-tube lab perfector): 61 62 63 64
Oil phase temperature 50 50 50 50
Water phase temperature 50 50 50 50
Emulsion temperature 50 50 50 50
Centrifugal pump Auto Auto Auto Auto
Capacity high pressure pump 40 40 40 40
Cooling (NH3) tube 1: -10 -10 -10 -10
Cooling (NH3) tube 2: -10 -10 -10 -10
Cooling (NH3) tube 3: -10 -10 -10 -10
Rpm tube 1: 1000 1000 1000 1000
Rpm tube 2: 1000 1000 1000 1000
Rpm tube 3: 1000 1000 1000 1000
The results reveal an apparent inconsistency. PGPR at 0.6% has larger DSD
(average 47.56μm) than PGPR at 0.4% (average 6.10μm). One would imagine the
35% TAG emulsion to be relatively less stable and have larger DSD, but the
270
results in Table 5.8.4 show that PGPR at 0.4% results in a smaller DSD. The
reason for the discrepancy is possibly explained, because 0.6% PGPR is within a
40% TAG emulsion and contains skimmed milk powder (SMP), which can
destabilise the emulsion, which seems contradictory, but protein within the SMP is
often used to “open” the emulsion and enhance flavour release and melting
behaviour. Therefore, despite 5% less TAG content and hence increased water
phase for the 35% TAG, it is possibly easier to process than the 40% TAG.
Table 5.8.3 Low TAG W/O (40%) emulsions samples 61 – 64 to test PGPR water
droplet size distribution at concentrations: 0.15, 0.3, 0.6 & 1.2%
Average/
Sample ID 2.5% <μm 50% <μm 97.5% <μm Dose wt%
St.dev
61 Average 1.46 12.54 107.93 0.15
St.dev 0.05 0.46 10.39
62 Average 1.02 12.89 164.41 0.30
St.dev 0.04 0.67 22.91
63 Average 0.93 6.59 47.56 0.60
St.dev 0.09 0.37 9.61
64 Average 1.82 4.23 9.84 1.20
St.dev 0.03 0.05 0.27
Table 5.8.4 Results of average DSD volumes for low TAG W/O (35%) emulsions
(no protein / stabiliser) 0.6% Moringa 191; 0.4% PGPR 90 and 0.3% PGPR +
0.15% CRY110))
Average/
Sample ID 2.5% <μm 50% <μm 97.5% <μm Dose wt%
St.dev
7 Average 3.52 3.54 3.56 0.6% Moringa 191
St.dev 0.05 0.03 0.02
8 Average 1.92 3.42 6.10 0.4% PGPR
St.dev 0.10 0.02 0.28
18 Average 3.58 3.59 3.60 0.3% PGPR + 0.15% CRY110
St.dev 0.01 0.01 0.02
271
Figure 5.8.1 DSD normal distribution shows data from Table 5.8.3
Hardness at 5°C
350
293 61
Jr. No
300 17124-1-61
250 226
194 62
Jr. No
Force / g
200 17124-1-62
164
150
63
Jr. No
17124-1-63
100
50 64
Jr. No
17124-1-64
0
Average of 10 measurings
272
5.8.3.3 CLSM
Confocal laser scanning microscopy (CLSM) images show trials 61 – 64, in
Figures 5.8.3 - 5.8.6. Each figure shows four images at two magnifications (40X
top and 100X bottom) and then scaled to 375 x 375µm and 188 x 188µm. All
samples were stained with FITC, stains protein green, and Nile Red, stains fat red.
Comparing Figure 5.8.3 (sample 61, 0.15% PGPR) and Figure 5.8.4 (sample 62,
0.3% PGPR) there is apparent similarity in size and shape of water droplets despite
significant difference in PGPR dose. The DSD data in Table 5.8.3 shows that as
concentration is increased from 0.15% to 0.3% an increase in droplet size occurs,
this is an unexplained anomaly.
Figure 5.8.3 CLSM image of 40% TAG emulsion (Sample 61, dose 0.15% PGPR)
273
As emulsifier load is increased to 0.6% PGPR (sample 63, Figure 5.8.5) and 1.2%
PGPR (sample 64, Figure 5.8.6) the results show decreased droplet size. This is
consistent with the DSD and textural analysis.
Figure 5.8.4 CLSM image of 40% TAG emulsion (Sample 62, dose 0.3% PGPR)
274
Figure 5.8.5 CLSM image of 40% TAG emulsion (Sample 63, dose 0.6% PGPR)
275
Figure 5.8.6 CLSM image of 40% TAG emulsion (Sample 64, dose 1.2% PGPR)
276
5.8.3.4 Cardboard Test
In Figure 5.8.7 simple cardboard test (abuse test) results are shown for PGPR
samples at 0.15, 0.3, 0.6 and 1.2% dosages. The sensory comments were: At
0.15%, the emulsion was falling apart, was generally weak, a watery mouth feel.
At 0.3%, signs of water separation, non-creamy and watery mouth feel. At 0.6%
acceptable, mouth feel thicker than at 0.3% (sample 62), chewy in character. At
1.2%, good stability, but mouth feel was very chewy.
Sample 62
(0.3%)
Sample 61
(0.15%)
Sample 64
(1.2%)
Sample 63
(0.6%)
Figure 5.8.7 Photographic evidence of spread tests: PGPR at 0.15% & 0.3% (top)
and 0.6% & 1.2% (bottom)
277
The sensory and simple cardboard test shows the use of PGPR as a single
emulsifier in low TAG W/O emulsions is to be severely inadequate at stabilising
the emulsion at 0.15 to 0.3%. At low emulsifier concentrations, which are possibly
below the surface excess concentration, larger droplets are formed because of
insufficient surface coverage, in turn leading to coalescence. Smaller droplets with
increasing concentrations are formed because of sufficient surface coverage and
less coalescence (Taylor, 2011). However, other considerations must also be
weighed, such as sensory, and spreadability. Likely between 0.3 and 0.6% there is
sufficient PGPR concentration to stabilise the emulsion; according to sensory and
abuse test, the emulsion at 0.6% is stabilised (Figure 5.8.7).
The water binding properties of PGPR 90 are one of the reasons that it is
essentially the stabilising emulsifier of choice for many water – oil based food
systems (1.2.8; 1.6; 1.7). Hence, PGPR is normally used in conjunction with a
second emulsifier for low TAG emulsions <41% (Garti & Remon 1984).
However, the maximum permitted dose of 0.4% (EC Directive 95/2/EC) is
potentially insufficient to achieve stability. This examination has shown that
PGPR is not adequately able to stabilise low TAG W/O emulsions to 0.3%. In
contrast, Moringa MAG (191) was shown to be able to stabilise the emulsions in a
full water phase 40% TAG emulsion at between 0.15% and 1.2%, with the optimal
being in between this range (5.4). These systems did not exhibit water leakage,
were stable and spreadable.
In typical TAG blends e.g. low fat W/O emulsions, there is usually a proportion
which is higher melting. On cooling, the higher melt will eventually crystallise out
of solution and in doing so establish additional surfaces for the emulsifier to
adsorb to. A higher melting emulsifier e.g. behenic rich MAG, when crystallising
out of solution, may establish a template structure of micro or nano crystalline
material (Taylor, 2011) for new born crystal growth within the bulk TAG solvent.
278
The use of a highly surface-active emulsifier such as PGPR may result in a
strengthened interface; so that there is mutual dependence with a co-emulsifier. It
is suggested that Moringa MAG has a bi-functional role, first as co-emulsifier
acting as a template for nucleation (5.2; 5.3; 5.5; 5.6) and secondly high surface
activity (4.0).
5.8.4 Conclusions
279
5.9 The Performance of Varying Distillations of Moringa MAG in Low
TAG W/O (40%) Emulsions
5.9.1 Introduction
Previously, the use of fully distilled Moringa MAG 191, from Moringa oleifera
TAG has been studied for high (60%) and low (40%) TAG W/O emulsion systems
(5.4). The purpose of this study was to examine the behaviour of previously
untested partially distilled Moringa MAG 102 / 105 samples in 40% low TAG
W/O emulsions (5.3) and compare the results obtained from 5.4.
The Moringa MAG used in this study were the same as outlined in previous work
(5.3). Briefly, the fatty acid profiles of the samples of the monoglyceride from
Moringa oleifera TAG are given in Table 5.9.1, and Table 5.9.2 shows their
mono-, di-, and tri-glycerides. It should be noted here that Moringa MAG 191
(~91% mono) was not used in this particular study, because it has previously been
tested, validated and documented earlier (5.4).
280
Table 5.9.1 Fatty acid composition of MAG based on Moringa oleifera TAG
Table 5.9.2 Moringa (Moringa oleifera) based MAG: mono, di- & tri-glycerides
(#2559/104 = 105 with antioxidant)
The sample formulation used in this work can be seen in Table 5.9.3, together with
the processing parameters for the same in, Table 5.9.4.
281
The sample number, concentration and identification are given as follows:
Table 5.9.3 Formulations: low TAG W/O (40%) emulsions with Moringa oleifera
based MAG
Ingredients in %
Ingredient Name 41 42 43 44 45 46 47 48
Water phase
Water (Tap) 57.300 57.300 57.300 57.300 57.300 57.300 57.300 57.300
Salt 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000
Skimmed milk powder (MILEX
0.100 0.100 0.100 0.100 0.100 0.100 0.100 0.100
240)
GRINDSTED® LFS 560 1.500 1.500 1.500 1.500 1.500 1.500 1.500 1.500
Potassium Sorbate 0.100 0.100 0.100 0.100 0.100 0.100 0.100 0.100
Butter Flavouring 050001 T03007 0.010 0.010 0.010 0.010 0.010 0.010 0.010 0.010
Water phase total 60.010 60.010 60.010 60.010 60.010 60.010 60.010 60.010
pH 5.5 5.5 5.5 5.5 5.5 5.5 5.5 5.5
Fat phase
Fat blend
PK4 - INES 25.000 25.000 25.000 25.000 25.000 25.000 25.000 25.000
Rapeseed oil 75.000 75.000 75.000 75.000 75.000 75.000 75.000 75.000
Fat blend total 100.000 100.000 100.000 100.000 100.000 100.000 100.000 100.000
Other fat ingredients
Moringa MAG102 0.150 0.300 0.600 1.200
Moringa MAG105 0.150 0.300 0.600 1.200
2% sol. beta-carotene 0.020 0.020 0.020 0.020 0.020 0.020 0.020 0.020
Butter Flavouring 050001 T04184 0.020 0.020 0.020 0.020 0.020 0.020 0.020 0.020
Other fat ingredients total 0.190 0.340 0.640 1.240 0.190 0.340 0.640 1.240
Fat phase total 39.990 39.990 39.990 39.990 39.990 39.990 39.990 39.990
RECIPE total (calc. batchsize) 100.000 100.000 100.000 100.000 100.000 100.000 100.000 100.000
282
Table 5.9.4 Processing conditions for low TAG W/O emulsions
Pilot Plant
Processing (3-tube lab perfector): 41 42 43 44 45 46 47 48
Oil phase temperature 50 50 50 50 50 50 50 50
Water phase temperature 50 50 50 50 50 50 50 50
Emulsion temperature 50 50 50 50 50 50 50 50
Centrifugal pump Auto Auto Auto Auto Auto Auto Auto Auto
Capacity high pressure pump 40 40 40 40 40 40 40 40
Cooling (NH3) tube 1: -10 -10 -10 -10 -10 -10 -10 -10
Cooling (NH3) tube 2: -10 -10 -10 -10 -10 -10 -10 -10
Cooling (NH3) tube 3: -10 -10 -10 -10 -10 -10 -10 -10
Rpm tube 1: 1000 1000 1000 1000 1000 1000 1000 1000
Rpm tube 2: 1000 1000 1000 1000 1000 1000 1000 1000
Rpm tube 3: 1000 1000 1000 1000 1000 1000 1000 1000
The procedure for water droplet size distribution (DSD) analysis, confocal laser
scanning microscopy (CLSM), and texture analysis were reported earlier (5.4;
5.7). Four images at two magnifications (40X and 100X) were taken respectively.
The images are then reproduced (scaled) to 375 x 375 µm and 188 x 188µm. All
samples are stained with FITC, stains protein green, and Nile Red, stains fat red.
283
5.9.3 Results and Discussion
The DSD data is given in Table 5.9.5 for samples 102 and 105, and Table 5.9.6 for
comparison shows sample 191 (5.4).
Table 5.9.5 DSD data: samples 41-44 (concentration 0.15, 0.3, 0.6 and 1.2%)
correspond to Moringa MAG 102, monoglyceride content ~53%; samples 45-48
(concentration 0.15, 0.3, 0.6 and 1.2%) correspond to Moringa MAG 105,
monoglyceride content ~83%
Average/
Sample ID 2.5% <μm 50% <μm 97.5% <μm Dose wt%
St.dev
41 Average 2.12 9.17 39.71 0.15
St.dev 0.03 0.20 2.08
42 Average 2.02 7.80 30.12 0.30
St.dev 0.03 0.21 1.32
43 Average 1.65 6.33 24.30 0.60
St.dev 0.02 0.02 0.48
44 Average 1.29 4.84 18.20 1.20
St.dev 0.07 0.08 1.54
45 Average 2.06 11.61 66.25 0.15
St.dev 0.01 0.80 11.62
46 Average 1.88 8.95 43.37 0.30
St.dev 0.05 0.27 3.55
47 Average 1.45 6.51 29.32 0.60
St.dev 0.01 0.21 1.92
48 Average 1.46 4.13 11.71 1.20
St.dev 0.04 0.09 0.80
Table 5.9.6 DSD: low TAG W/O (40%) spreads containing Moringa MAG 191
(monoglyceride content of ~91%)
284
5.9.3.1 Droplet Size Distribution (DSD)
Table 5.9.5 shows the DSD for Moringa 102 at 0.15, 0.30, 0.60 and 1.2%
concentrations, corresponding to 97.5 % of droplet volume of 39.71, 30.12, 24.30,
and 18.20µm. This shows a clear decreased DSD with increasing concentration.
Similarly for Moringa 105 over the same concentrations, 97.5 % of droplet volume
was 66.25, 43.37, 29.32 and 11.71µm, showing an increasing trend towards
decreased DSD. These results are in good agreement with the trends shown earlier
(5.3). For comparison Table 5.9.6, shows the DSD for Moringa 191 which had
good stability and mouth feel properties. Noteworthy, is that the DSD for Moringa
191 (91% mono) at 0.3% and 0.6% concentrations are closer to those from Table
5.9.5 for Moringa 105 (~83% mono) than Moringa 102 (~53% mono). The DSD
for Moringa 102 is smallest overall. This leads to the assumption that stability is
not inherently linked only to monoglyceride content or DSD exclusively.
Figure 5.9.1 shows photographic images of the samples after preparing an abuse
test of spreading arbitrary portion of the sample onto cardboard, then manually
spreading backward / forward. As the concentration of the Moringa MAG
increased the samples became more resilient (thicker).
Sensory comments were made for the samples which started for Moringa 102 at
sample 44, working to more dilute systems. The emulsion was stable, thick and
creamy, and then with subsequent dilutions proceeded to become less thick, and
less creamy in mouth feel. The regression in texture continued until the lowest
concentration was reached whereby the emulsion was described as uneven.
For Moringa 105, again starting at the highest concentration (sample 48), a
regression was found, from a stable and thick emulsion to one that is showing
clear signs of water separation, and not as thick or creamy in terms of mouth feel
(sample 45). Samples 46 / 47 sat placed on a sliding scale between these (sample
45 / 48) two extremes.
285
5.9.3.2 CLSM
The data presented in Figures 5.9.2 and 5.9.3 show the confocal laser scanning
microscopy (CLSM) images relating to Moringa MAG samples 102, and 105
respectively. Four images at magnification 40X were taken respectively. The
images are then reproduced (scaled) to 375 x 375µm
In both Figures 5.9.2 and 5.9.3 it can be seen that the images corresponding to the
lower dosage of the given Moringa MAG (top left) showed a much looser
structure compared to the remaining images where concentration increases. This is
manifested in larger DSD.
286
Figure 5.9.1 (1a) Spread tests: samples 41 & 42 (Moringa MAG 102); (1b) 43 &
44 (Moringa MAG 102); (1c) 45 & 46 (Moringa MAG 105) and (1d) 47 & 48
(Moringa MAG 105)
287
Figure 5.9.2 CLSM images, Moringa MAG 102: sample 41 (0.15%); sample 42
(0.3%); sample 43 (0.6%), and sample 44 (1.2%) (scaled to 375 x 375µm)
288
Figure 5.9.3 CLSM images, Moringa MAG 105: sample 45 (0.15%); sample 46
(0.3%); sample 47 (0.6%), and sample 48 (1.2%) (scaled to 375 x 375µm)
289
5.9.3.3 Texture Analysis
Texture analysis results on hardness are presented in Figure 5.9.4, and show a
general reduction in hardness as the concentration of either Moringa MAG 102 or
105 increases.
Figure 5.9.4 Texture analysis: hardness profile for Moringa MAG 102 emulsions
(41-44) and Moringa MAG 105 emulsions (45-48), at 5°C
The outcome of this study was to test the integrity of W/O low TAG emulsion
spreads; which was shown to be positive when using natural Moringa MAG with
lower mono % contents. Generally, the results have confirmed acceptable stability.
290
Next would be to ascertain if an earlier study (5.3) where synthetic Moringa MAG
were tested, would have any correspondence to the degree of final mono % shown
in resulting samples produced with Moringa (Moringa oleifera TAG) based MAG
distillations in this work.
5.9.4 Conclusion
In comparison to a fully distilled Moringa oleifera based MAG (191) tested earlier
(5.4), this work concludes that low TAG W/O emulsion spreads made with a
Moringa oleifera based MAG with mono contents of 53.16% and 82.55%, are
equally capable of producing commercially viable low TAG W/O emulsions.
291
5.10 Blended MAG compositions to Equal Moringa Based MAG in Low
TAG W/O Emulsions
5.10.1 Introduction
Previous studies have shown the use of Moringa MAG (based on Moringa oleifera
TAG) in low TAG W/O emulsion systems with natural distillations (5.4; 5.9) and
comparisons have been made to a specific MAG identified as being characterised
rich in C22:0 (5.7). Natural Moringa MAG contains in total ~10% of the three
saturated fatty acids: C20:0 + C22:0 + C24:0. The aim of this study was to assess
the effect blended commercial MAG compositions to reproduce a synthetic
version of natural Moringa MAG. The natural Moringa MAG samples were made
from base Moringa oleifera TAG as a raw material to give MAG as outlined
previously (5.3).
292
Table 5.10.1 Blending ratios for synthetic MAG
SM 90* SM 60** SM 80***
GRINDSTED® CRYSTALLISER 110 10 10 10
DIMODAN RT 90 50
GRINDSTED® Mono-di PR40 90 40
Total % 100 100 100
Mono (approximate %) * 96 ** 64 *** 82
293
Table 5.10.3 Total distribution of natural Moringa MAG saturated, unsaturated
fatty acids in relation to synthetic MAG and actual total saturated chain length
from C20:0
Table 5.10.4 Formulations for 40% low TAG emulsions with synthetic MAG
(SM90 SM60)
Ingredients in %
Ingredient Name 31 32 33 34 35 36 37 38
Water phase
Water (Tap) 57.300 57.300 57.300 57.300 57.300 57.300 57.300 57.300
Salt (Sodium Chloride) 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000
Skimmed milk powder (MILEX
0.100 0.100 0.100 0.100 0.100 0.100 0.100 0.100
240)
GRINDSTED® LFS 560
1.500 1.500 1.500 1.500 1.500 1.500 1.500 1.500
Stabiliser System
Potassium Sorbate 0.100 0.100 0.100 0.100 0.100 0.100 0.100 0.100
Butter Flavouring 507104 A 0.010 0.010 0.010 0.010 0.010 0.010 0.010 0.010
Water phase total 60.010 60.010 60.010 60.010 60.010 60.010 60.010 60.010
pH 5.5 5.5 5.5 5.5 5.5 5.5 5.5 5.5
Fat phase
Fat blend
PK4 - INES 25.000 25.000 25.000 25.000 25.000 25.000 25.000 25.000
COLZAO (Rapeseed Oil) 75.000 75.000 75.000 75.000 75.000 75.000 75.000 75.000
Fat blend total 100.000 100.000 100.000 100.000 100.000 100.000 100.000 100.000
Other fat ingredients
Synthetic Moringa – 1 (SM90) 0.150 0.300 0.600 1.200
Synthetic Moringa – 2 (SM60) 0.150 0.300 0.600 1.200
2% sol. beta-carotene 0.020 0.020 0.020 0.020 0.020 0.020 0.020 0.020
Butter Flavouring 050001 T04184 0.020 0.020 0.020 0.020 0.020 0.020 0.020 0.020
Other fat ingredients total 0.190 0.340 0.640 1.240 0.190 0.340 0.640 1.240
Fat phase total 39.990 39.990 39.990 39.990 39.990 39.990 39.990 39.990
RECIPE total (calc. batchsize) 100.000 100.000 100.000 100.000 100.000 100.000 100.000 100.000
294
Table 5.10.5 40% low TAG emulsions with synthetic MAG (SM80)
Ingredients in %
Ingredient Name 49 51 52 53
Water phase
Water (Tap) 57.300 57.300 57.300 57.300
Salt (Sodium Chloride) 1.000 1.000 1.000 1.000
Skimmed milk powder (MILEX 240) 0.100 0.100 0.100 0.100
GRINDSTED® LFS 560 Stabiliser System 1.500 1.500 1.500 1.500
Potassium Sorbate 0.100 0.100 0.100 0.100
Butter Flavouring 050001 T03007 0.010 0.010 0.010 0.010
Water phase total 60.010 60.010 60.010 60.010
pH 5.5 5.5 5.5 5.5
Fat phase
Fat blend
PK4 - INES 25.000 25.000 25.000 25.000
COLZAO (Rapeseed Oil) 75.000 75.000 75.000 75.000
Fat blend total 100.000 100.000 100.000 100.000
Other fat ingredients
Synthetic Moringa – 3 (SM80) 0.150 0.300 0.600 1.200
2% sol. beta-carotene 0.020 0.020 0.020 0.020
Butter Flavouring 050001 T04184 0.020 0.020 0.020 0.020
Other fat ingredients total 0.190 0.340 0.640 1.240
Fat phase total 39.990 39.990 39.990 39.990
RECIPE total (calc. batchsize) 100.000 100.000 100.000 100.000
Refer to General Materials & Methods (2.0) for emulsion assembly procedure. The
processing conditions on the plant for all samples are given in Table 5.10.6.
Table 5.10.6 Pilot plant scraped surface processing conditions for all samples
Analytical measurements, DSD, CLSM and texture analysis were carried out as
described earlier (2.0; 5.4).
295
5.10.3 Results and Discussion
Table 5.10.7 DSD data for SM 90 (samples 31-34), SM60 (samples 35-38), and
SM80 (samples 49-53)
Average/
Sample ID 2.5% <μm 50% <μm 97.5% <μm Dose wt%
St.dev
31 Average 1.33 5.19 20.22 0.15
St.dev 0.03 0.07 0.94
32 Average 1.31 4.39 14.70 0.30
St.dev 0.05 0.02 0.60
33 Average 1.38 3.95 11.34 0.60
St.dev 0.08 0.04 0.81
34 Average 1.33 3.73 10.56 1.20
St.dev 0.15 0.03 1.22
35 Average 1.61 6.32 24.75 0.15
St.dev 0.02 0.04 0.45
36 Average 1.48 5.52 20.51 0.30
St.dev 0.04 0.05 0.72
37 Average 1.36 4.34 13.90 0.60
St.dev 0.01 0.03 0.20
38 Average 1.90 3.52 6.53 1.20
St.dev 0.11 0.04 0.47
49 Average 1.72 6.50 24.58 0.15
St.dev 0.05 0.04 1.06
51 Average 1.47 5.33 19.37 0.30
St.dev 0.06 0.14 1.00
52 Average 1.39 4.27 13.12 0.60
St.dev 0.04 0.01 0.42
53 Average 1.72 3.42 6.81 1.20
St.dev 0.07 0.03 0.17
The data from Table 5.10.7 shows a trend as seen previously, where at lower
synthetic MAG concentrations DSD is larger, and decreases as synthetic MAG
concentration increases. The smaller the DSD led a tendency towards more stable
emulsion. The mean DSD during the emulsification process was heavily
dependent on emulsifier concentration. At low emulsifier concentrations, which
296
are possibly below the surface excess concentration, larger droplets were formed
because of insufficient surface coverage, in turn leading to coalescence. Smaller
droplets with increasing concentrations were formed because of sufficient surface
coverage (Taylor 2011). However, other considerations must also be weighed,
such as sensory, and spreadability (5.8).
While there has been attempt to control all variables, there was always “natural”
process variation (with a normal distribution) from batch to batch processing
conditions. From a pilot scale perspective, the natural process variation made it
extremely difficult to statistically validate because the pilot plant setup for the
studies in this work and others (5.4; 5.7; 5.9) was exclusively setup for this
product specification. However, where there were anomalies, these have been
mentioned.
The DSD data showed all synthetic (SM 90 / 60 / 80) MAG concentrations were
considerably lower than the corresponding natural Moringa MAG reported earlier
at the 0.3% and 0.6% emulsifier load. This immediately suggests that the synthetic
MAG at 0.3 / 0.6% gave tighter structures (5.4; 5.7). For ease of comparison the
DSD from previous tests of Moringa MAG at 0.15 0.3, 0.6, 1.2%, are shown in
Tables 5.10.8 and 5.10.9. As previously mentioned, the anomaly, may possibly
have been attributed to batch to batch variation.
297
Table 5.10.8 DSD data for 40% TAG spread samples containing natural Moringa
MAG 191 monoglyceride ~91% (from Table 5.4.8 in: 5.4; and Table 5.7.5 in: 5.7)
Table 5.10.9 DSD data from 5.9 shows: samples 41-44 (conc. 0.15, 0.3, 0.6, 1.2%)
correspond to Moringa MAG 102 (mono 53%); and samples 45-48 (conc. 0.15,
0.3, 0.6, 1.2%) correspond to Moringa MAG 105 (mono 83%)
Average/
Sample ID 2.5% <μm 50% <μm 97.5% <μm Dose wt%
St.dev
41 Average 2.12 9.17 39.71 0.15
St.dev 0.03 0.20 2.08
42 Average 2.02 7.80 30.12 0.30
St.dev 0.03 0.21 1.32
43 Average 1.65 6.33 24.30 0.60
St.dev 0.02 0.02 0.48
44 Average 1.29 4.84 18.20 1.20
St.dev 0.07 0.08 1.54
45 Average 2.06 11.61 66.25 0.15
St.dev 0.01 0.80 11.62
46 Average 1.88 8.95 43.37 0.30
St.dev 0.05 0.27 3.55
47 Average 1.45 6.51 29.32 0.60
St.dev 0.01 0.21 1.92
48 Average 1.46 4.13 11.71 1.20
St.dev 0.04 0.09 0.80
298
It was observed that SM90 (31-34) where the monoglyceride content was highest
(96.50%) was not the sample resulting with the smallest water droplet size at 1.2%
concentration, as may have been expected. The high concentrations of SM60 and
SM80 with monoglyceride contents of ~64%, and ~82% respectively were lower
and similar to each other. This seems to suggest that the monoglyceride /
diglyceride content of the synthetic blended MAG may play a more specific role in
adjusting water droplet size than was previously thought.
Also to consider is that DSD is not the only determining factor to confirm a
stabilised emulsion; the interfacial film thickness and strength are likely also
important considerations. This aspect has shown the importance of interfacial
tension and visco-elastic behaviour especially, in conjunction with combined
emulsifiers (4.0) e.g GRINDSTED® Crystallizer 110 (CRY110) and
GRINDSTED® PGPR90 (PGPR). It was shown previously (5.6) that Moringa
MAG (191), has similar viscous profiles to PGPR and yet these two emulsifiers
could not be more molecularly dissimilar, given that monoglycerides are low
molecular weight, compared to the high molecular weight PGPR.
Application tests have characterised and made distinction between a behenic based
MAG (CRY110) and Moringa MAG and have shown a behenic rich C22:0 ~89%)
MAG is not conducive to stabilised W/O low TAG emulsions (5.7).
The strong water binding properties of PGPR are also well known (Claesson et al,
1997; Garti & Remon 1984; Marze, 2009; Dedinaite & Campbell, 2000; Rousseau,
2000). Application tests have shown that PGPR used alone in low TAG W/O
emulsions, resulted in relatively increased volume of water droplet size, when
compared to the natural Moringa MAG (5.8)
How does Moringa MAG work at the interface? Fatty acid compositions showed
(Tables 5.10.2 and 5.10.3) that the synthetic MAG have roughly 8.5 – 12.3% of
their saturation starting from C20:0. Importantly, for the natural Moringa MAG,
diversity exists from the same fatty acid chain length (C20:0) and there is a
significant proportion of C22:0 and C24:0 as seen in Table 5.10.2 (also detected
299
C26:0, but not shown). Whereas, within the same spectrum, CRY110 (5.7) and
synthetic MAG are mainly centred at the C22:0 chain length. Another distinction
is that natural Moringa MAG also contain lower chain (C16:0 and C18:0) and a
high degree of unsaturated fatty acids (~72%). Given the complexity of the
natural Moringa MAG fatty acid profile it was not possible within the scope of this
examination to precisely mimic this fatty acid range exactly for the synthetic
MAG.
Factors such as resilience / abuse (cardboard test) and compression (texture test)
must also be considered.
The next test looked at integrity of the samples by way of a physical spreading
(abuse) test. The results of which are shown in Figure 5.10.1. The samples were
spread out manually onto a piece of cardboard using a regular kitchen knife. The
use of cardboard is designed to mimic a surface offering resistance e.g bread /
toast and the tongue to a certain degree, i.e. the cardboard although smooth is a
roughened surface (specification not known). The spreading action gives a better
representation than would be the case for spreading on plastic, glass or steel.
Together with the spread tests, sensory evaluation of the sample was also
completed. These results can be summarised as follows:
SM90 with monoglyceride ~96% (samples 31-33) had acceptable structure, a thick
and creamy mouth feel and acceptable in the mouth melt profile, indicating good
flavour release. SM 90 sample 34 gave an acceptable emulsion, but the mouth feel
was not as smooth or creamy, and the melt profile was slower. This is at high
concentration – 1.2%.
SM60 with monoglyceride 64.56% (sample 35) gave an acceptable emulsion, but
was duller in appearance, though creamy as the preceding samples (SM90 31-33),
but the in mouth melt profile was poorer. SM60 sample 36 gave a better emulsion
300
and was creamy and thicker than SM60 sample 35. SM60 sample 37 gave a good
thick emulsion and was creamy and thick in the mouth feel. SM60 sample 38 gave
a thick emulsion was creamy and thick to taste, but showed a slow melting profile.
Likely this was again attributed to high concentration 1.2%.
SM80 with monoglyceride ~82% (sample 49) gave a good emulsion and was thick
and creamy to taste. SM80 sample 51 gave a good shiny emulsion, was thick, but
not as creamy as SM80 sample 49. SM80 sample 52 gave an acceptable thick
emulsion with a creamy taste. SM80 sample 53 gave a very thick emulsion, with
an equally thick mouth feel, and poor flavour release, once again attributed to high
concentration 1.2%.
301
Figure 5.10.1 Photographic evidence of spread testing of the synthetic MAG 31-
34 (SM90); 35-38 (SM60); and 49, 51-53 (SM80)
302
As is seen from the sensory results and images in Figure 5.10.1, an increase in
concentration of the synthetic Moringa MAG generally leads to greater emulsion
stability, but eventually at a concentration between 0.6% and 1.2%, mouth feel and
the flavour release became poor. It was also interesting to compare back to the
natural Moringa MAG spreads described earlier (5.9) to note that as the
concentration of the natural Moringa MAG samples increased, there was not the
same consistent comment regarding poor flavour release despite the concomitant
increase in emulsion stability. This suggests the synthetic MAG have different
interfacial film behaviour and are interfacially active according to a different
mechanism when compared directly to the natural Moringa MAG. This aspect has
been shown through tensiometry measurements (method as before: 2.0; 4.0),
whereby natural Moringa MAG (191) was similar to, or more surface active
(lower tension), than synthetic equivalents as shown in Figure 5.10.2.
It is important to note that where measured and studied, the viscosity (G’’) and
elastic (G’) modulus of several fatty acid variations of commercial MAG, showed
that only behenic rich (CRY110) and mixtures of the same blended together with
303
other MAG, resulted in development of both G’’ / G’ in samples (4.0). All three
synthetic MAG (SM90, SM80, SM60) contained a minimum of ~7.2% behenic
contribution from CRY110 (see Table 5.10.2). Interfacial tension measurements
shown throughout this work have also shown CRY110 is highly surface active at
~<20°C, within the final exit temperature (packing T°C) of the W/O emulsions
(Bech et al., 2013).
5.10.3.2 CLSM
Figure 5.10.3 CLSM images of synthetic MAG (SM90) mono content ~96%,
samples 31, 32, 33, & 34, of concentration 0.15, 0.3, 0.6 & 1.2% respectively
304
Figure 5.10.4 CLSM images of synthetic MAG (SM60), mono content ~64%,
samples 35, 36, 37, & 38, of concentration 0.15, 0.3, 0.6 & 1.2% respectively
305
Figure 5.10.5 CLSM images of synthetic MAG (SM80) mono content ~82%,
samples 49, 51, 52, & 53, of concentration 0.15, 0.3, 0.6 & 1.2% respectively
306
5.10.3.3 Texture Analysis
31
32
33
34
35
36
37
38
Figure 5.10.6 Hardness results from texture analyser for synthetic SM90 (31 to
34) and synthetic SM60 (35 to 38)
Figure 5.10.6 shows textural hardness data for synthetic MAG (SM90 / SM60)
with mono contents of ~96% and ~64 respectively, which broadly agree with
findings in previous work (5.9). The hardness values decreased as the
concentration of the Moringa MAG increased, and the same was observed with
synthetic MAG. The hardness values themselves were greater for the synthetic
MAG than was the case for the natural Moringa MAG described earlier by
between 50 to 100g (5.9).
Figure 5.10.7 shows synthetic MAG (SM80) with a mono content of ~82%, which
did not correlate with decreasing hardness with increasing emulsifier
concentration. The hardness values were still more than the values for the
corresponding natural Moringa MAG. This is likely explained by the ratio of
saturated fatty acids to unsaturated. The natural Moringa MAG have a higher
degree of unsaturation, hence softer, which is known to have proven advantages
for manufacture of low TAG W/O emulsion spreads (1.2.6; 1.5; 1.6).
307
49
51
52
53
Figure 5.10.7 Hardness results from texture analyser for synthetic SM80 (49, 51,
52 & 53)
Where the synthetic MAG have been blended to mimic the natural, non-
hydrogenated Moringa MAG samples, the same diversity of long chain fatty acids
(>C22:0) is not viable because of availability and hence price prohibitive.
Therefore, synthetic samples are essentially based on the fully hydrogenated
CRY110. From a commercial perspective there is implication; synthetic MAG are
able to produce firmer spreads, which have a smaller average water droplet size
than natural Moringa MAG. Whereas the natural non-hydrogenated MAG samples
may result in a softer, creamier mouth feel, whilst maintaining interfacial film
strength and which can fulfil re-work / re-melt requirements.
5.10.4 Conclusion
308
5.11 Effect of High and Low Temperature Distillation on the Functionality
of MAG based on Moringa oleifera TAG in Low TAG W/O Emulsions
5.11.1 Introduction
Another distinction is that natural Moringa MAG also contain other fully saturated
fatty acids of shorter chain lengths (C16:0 and C18:0) and a high degree of
unsaturated fatty acids (~72%). Given the complexity of the natural Moringa
MAG fatty acid profile it is was not possible within the scope of previous
examination (5.10), to precisely mimic its fatty acid range exactly (5.6 – 5.10).
The distillation process used to synthesise Moringa MAG from the raw TAG
material was carried out at an unusually higher temperature than normal, because
distillation at high temperature preserves a wider spectrum of fatty acids of
different chain lengths, as high as plus C24:0, thereby resembling the natural fatty
acid profile of the original TAG composition. However, could the same functional
properties be achieved from a Moringa MAG when distilled at a lower, normal
distillation temperature?
The aim of this study was to investigate two samples of Moringa MAG, distilled at
two temperatures, 210°C and 185°C respectively, and to test if functional
309
differences are apparent. The resultant Moringa MAG were tested in W/O low
TAG emulsions (40% and 35% TAG), at similar process conditions described
earlier (5.4, 5.7 – 5.10).
Water DSD was carried out using a Bruker Minispec NMS (20MHz) according to
the standard method (outlined in Materials & Methods 2.0). Texture analysis was
carried out using a TA-XT2i texture analyser from Stable Micro Systems. Visual
evaluation of the emulsions was carried out after 16 weeks storage at 5°C, by
spreading with a typical kitchen table knife onto cardboard (abuse test), and the
results photographed. The W/O emulsions were evaluated by an expert sensory
panel, for texture / mouth feel and flavour release parameters. Table 5.11.3 gives
the fatty acid profile of the high (210°C) and low (185°C) temperature distillations
for two Moringa MAG distillations from Moringa oleifera TAG (see 2.0):
310
Table 5.11.1 Formulations for 35% and 40% TAG emulsions using two Moringa
MAG distillations: sample 2559/132 (87% Monoglyceride - high temperature
distillation 210°C) and sample 2559/134 (97% Monoglyceride – low temperature
distillation 185°C)
Ingredients in %
Ingredient Name 71 72 73 74 75 76 77 78
Water phase
Water (Tap) 57.300 57.300 57.300 57.300 64.000 64.000 64.000 64.000
Salt 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000
Skimmed milk powder (MILEX 240) 0.100 0.100 0.100 0.100
GRINDSTED® LFS 560 1.500 1.500 1.500 1.500
Potassium Sorbate 0.100 0.100 0.100 0.100
Butter Flavouring 507104 A 0.010 0.010 0.010 0.010 0.010 0.010 0.010 0.010
Water phase total 60.010 60.010 60.010 60.010 65.010 65.010 65.010 65.010
pH 5.5 5.5 5.5 5.5 5.5 5.5 5.5 5.5
Fat phase
Fat blend
PK4 - INES 25.000 25.000 25.000 25.000 25.000 25.000 25.000 25.000
Rapeseed oil 75.000 75.000 75.000 75.000 75.000 75.000 75.000 75.000
Fat blend total 100.000 100.000 100.000 100.000 100.000 100.000 100.000 100.000
Other fat ingredients
87% Monoglyceride (Moringa oil), 0.300 0.600 0.300 0.600
2559/132
97% Monoglyceride (Moringa oil), 0.300 0.600 0.300 0.600
2559/134
Butter Flavouring 050001 T04184 0.020 0.020 0.020 0.020 0.020 0.020 0.020 0.020
Other fat ingredients total 0.320 0.320 0.620 0.620 0.320 0.320 0.620 0.620
Fat phase total 39.990 39.990 39.990 39.990 34.990 34.990 34.990 34.990
RECIPE total (calc. batchsize) 100.000 100.000 100.000 100.000 100.000 100.000 100.000 100.000
311
Table 5.11.2 Processing conditions for 35% and 40% TAG W/O emulsions
312
Table 5.11.3 Fatty acid compositions of high and low temperature MAG
distillations from Moringa oleifera TAG
313
5.11.3 Results and Discussion
The results from the water droplet size distribution (DSD) are given in Table
5.11.4 where emphasis is placed towards 97.5% <µm.
Table 5.11.4 DSD analysis results for W/O emulsions (samples 71 – 78)
The 40% TAG (water phase with hydrocolloid / protein) emulsion (samples 71 –
74), DSD (97.5% <µm), indicates all emulsions are likely to be stable with
minimal phase separation, but need to be concluded in relationship to sensory
tests. These results are different from those previously found (Table 5.4.8 in 5.4; &
Table 5.7.5 in 5.7), where Moringa MAG from high temperature distillation (Lot
2559/ 191) only were tested.
Previous DSD results (5.4 & 5.7) for Moringa MAG at 0.3% concentration and
Moringa MAG at 0.6% were; 42.41, and 38.56 respectively, which compares to
18.19, and 15.44 here. This suggests that the stability of these new samples is
greater than the stability of the original samples tested previously (5.4 & 5.7). A
314
possible reason is the difference in diglyceride content which, although not
discussed here, is known to aid emulsification (Shimada and Ohashi 2003).
Factors such as insufficient emulsification or inconsistencies in the shear scraped
surface heat exchanger process may also have a bearing.
The low temperature distillation samples (72 & 74) gave larger respective DSD
sizes compared to the high temperature distillations; where the low temperature
distillation samples showed here are within the DSD range shown earlier (5.4) and
correspond to samples that were fundamentally stable in appearance and abuse
testing (spreading).
The 35% TAG (empty water phase) emulsions, samples 75 – 78, gave DSDs to
values consistently under 10 microns. This tends to indicate an extremely stable
emulsion, indeed if not over-stable, which would negatively impact the flavour
release properties (see 5.11.4.2). No material difference could be seen from high
temperature distillation to the low temperature distillation samples.
5.11.3.2 Sensory
It was often difficult to make final conclusions from the DSD alone, and therefore
a visual (photographic) evaluation combined with sensory evaluation, where each
sample was spread onto cardboard and then tasted and characterised
organoleptically was carried out. Comments from sensory after spreading and
tasting are given as follows:
71. Very stable, seemed as if break down would occur but did not. Improved with
more working. Mouth feel; initially thick/creamy, breaks down easily, good
flavour release.
72. As stable as 71, no separation. Mouth feel; not as thick as 71, but persists in
the mouth for longer not breaking down as easily.
315
73. Very thick, bit of separation to begin with, then stabilised. Mouth feel; thicker
than 71/72, but still acceptable flavour release. Retains structure longer, does not
breakdown so easily, not as sticky/waxy as 71/72.
74. Very thick, some oiling out, but emulsion ok, slightly more unstable than 73.
Mouth feel; not as thick as 73, detection of lumps possible, like scrambled eggs.
75. Very very stable emulsion, thick, very spreadable. Mouth feel; thick, not too
tight, firm, breaks down, melting ok, bit lumpy.
76. Very very stable, thick spreadable. Mouth feel; thick, but more salt release
than 75.
77. Extremely stable emulsion, thick, spreadable. Mouth feel; smooth, thick, slow
flavour release.
78. Extremely stable, spreadable, thick, slightly more sticky than 74. Mouth feel;
smooth, thick, no breakdown, possibly stabilised?
40% W/O emulsions: On assessment of samples 71-74, the 40% TAG emulsions,
the samples with the high temperature distillation, (71 and 73) seemed to form
stable emulsions, but are breaking down in the mouth easier than the low
temperature counterparts, giving good flavour release. It would seem that the
Moringa MAG produced via the high temperature distillation, i.e. with the
broadest range of fatty acids (see Table 5.11.3), seem to infer a beneficial function
in terms of in-mouth sensorial properties and flavour release.
35% W/O emulsions: At 35% TAG level, a similar trend is observed; the high
temperature distillation samples resulted with products which had improved mouth
feel perception and flavour release. However, all the 35% TAG emulsions as a
whole were extremely stable and as such could not be classified (in the opinion of
expert sensory panel) as being representative of a typical table emulsion. These
could be summarised as all being too stable and too thick for acceptable product
use. Probably for a more realistic consistency, the dosage of the Moringa MAG
should be reduced to perhaps 0.15%. Again, although the stability and taste levels
316
of the high temperature distillation samples were improved, the low temperature
samples were not unacceptable.
1a 1b
no. 71 no. 72
0.3% 0.3%
1c 1d
no. 73 no. 74
0.6% 0.6%
Figure 5.11.1 40% emulsion samples: (1a) no.71, 0.3% dose, HIGH T°C
distillation; (1b) no.72, 0.3% dose, low T°C distillation; (1c) no.73, 0.6% dose,
HIGH T°C distillation; (1d) no.74, 0.6% dose, low T°C distillation
317
1a 1b
no. 75 no. 76
0.3% 0.3%
1c 1d
no. 77 no. 78
0.6% 0.6%
Figure 5.11.2 35% emulsion samples: (1a) no.75, 0.3% dose, HIGH T°C
distillation; (1b) no.76, 0.3% dose, low T°C distillation; (1c) no.77, 0.6% dose,
HIGH T°C distillation; (1d) no.78, 0.6% dose low T°C distillation
Texture analysis
The results from texture analysis in terms of hardness are given in Figure 5.11.3
after measurement at 5°C.
71No 17124-1-71
Jr.
Hardness at 5°C 72No 17124-1-72
Jr.
400 73No 17124-1-73
Jr.
350 332 323
300 74No 17124-1-74
Jr.
282 283
300 75No 17124-1-75
Jr.
229 236 242
250 76No 17124-1-76
Jr.
Force /g
Figure 5.11.3 Texture analysis: hardness for 40% TAG emulsions (samples 71-
74), and 35% TAG emulsions (samples 75-78)
318
5.11.3.3 Effect of Distillation at 0.3% Moringa MAG Concentration
When comparing samples 73 / 74, though small, there appeared similar agreement
when compared to water-droplet sizes. The same pattern did not follow for
samples 77 / 78. This was probably attributed because at the emulsifier
concentrations and shear on the emulsion, a mean DSD of approximately 3 - 4µm,
had already approached the maximum decreased droplet size.
The result of high shear during emulsification and cooling will always produce
droplets, however the Laplace pressure within the droplets increases as droplet size
decreases, consequentially resulting in a resistance to further deformation. The
addition of an emulsifier decreases the Laplace pressure allowing the production
of smaller droplets. However, 0.6% Moringa MAG in this examination probably
exceeds the concentration necessary to produce stable emulsions.
In the case of 40% emulsions, the initial trend in textural resilience is thought to be
reliable. The “elastic” resilience experienced in sensory observations is likely the
319
effect of strengthened interfacial film properties (4.0) at the water-oil surface
(observed in the pilot studies in 1.5). This may have some connection with the
miscibility of the Moringa MAG, where the proportion of saturated fatty acids
>C20:0 are likely to be more surface-active (Krog 1992), forming an insoluble
microcrystalline arrangement, where an interstitial stabilised interface may exist
(Johansson, Bergenståhl & Lundgren, 1995).
Data for 35% emulsions did not show any significant trend in the case of DSD –
evident from droplets at <5µm. The expert sensory findings follow a similar
pattern to the 40% emulsions. It is assumed therefore, that such resilience was the
effect of maintaining the whole preserved fatty acid profile in the high temperature
distillation, 87% mono (2559/132), and this is despite the higher mono content
(97%) in the low temperature distillation (2559/134).
Finally for both 40% and 35% TAG emulsion – in all cases - both high and low
temperature distillations were an improvement on PGPR as shown previously
(Tables 5.11.3; 5.11.4 in 5.11 & 5.8).
5.11.4 Conclusion
The results show a direct effect of securing a diversified fatty acid spectrum. It is
concluded that Moringa MAG synthesised at a higher distillation temperature has
generally led to direct increase on the influence of initial firmness and is manifest
through sensory resilience, compared to normal distillations. These physical
properties have a direct functional effect, leading to “elastic” resilience because of
strengthened interfacial film properties (4.0) at the water-oil surface (observed in
the pilot studies in 1.5). This may have some connection with the miscibility
(Ueno et al., 1994) of the Moringa oleifera based MAG, where the proportion of
saturated fatty acids >C20:0 are likely to be more surface-active (Krog 1992),
forming Pickering and insoluble microcrystalline arrangement; in turn, a steric,
interstitial stabilised interface may exist (Johansson, Bergenståhl & Lundgren,
1995).
320
5.12 Inventions: a Moringa MAG and Modified Crystalliser Composition
5.12.1 Summary
Recently, there has been a need to find ways to promote crystallisation in oil / fat
blends (triacylglycerols) which require an increase in their crystal kinetics
(Wassell & Young 2007; Wassell et al., 2010b; Young et al., 2008). One issue is
the removal of trans isomers which are derived from partial hydrogention of liquid
triacylglycerol (TAG) and semisolid triacylglycerols (TAGs). It is known that
partial trans isomers which have a degree of saturation (low iodine value) are able
to induce crystallisation (Flöter & van Duijn 2006).
There is still a problem to find materials which can overcome crystallisation and
emulsification issues when formulating TAG based oil blends which meet the
following criteria:- First, componants are trans free. Second, they should not come
from fully hydrogenated hardstock, and third, they should provide adequate
function within a low saturated TAG environment (Wassell et al., 2010a). These
may be especially interesting for applications where water-in-oil or oil-in-water
emulsions are used.
321
Currently, commercial reasons dictate a need to find a solution for substituting
fully hydrogenated fats (Wassell et al., 2010a). While these do not inherently
contain trans isomers of any significance, the perception in the media and more
importantly from medical sources, is that these should also be removed because
hydrogenation is linked to formation of trans and human coronory risks. Therefore
there is now requirement even with minor ingredients to make these fit under the
category of non-hydrogenated material (Wassell et al., 2010a), therefore attaining
a more natural and/or ethical status (Bech et al., 2008; Carmichael 2011; Wang et
al., 2012).
Emulsifiers designed from Moringa oleifera TAG (4.0, 5.0), are thought to be able
to provide unique functions because of the following reasons:-
322
A bi-functional Moringa MAG could be a “one-stop” solution for both low and
high TAG systems (5.4), where emulsification and crystallisation are deemed
critical.
Moringa oleifera TAG based monoglycerides may also provide unique new esters
of monoglycerides. Therefore, this could lead to construction of a monoglyceride
designed with a diverse fatty acid range, which may or may not contain a
significant degree of diglyceride content (possibly similar to Moringa MAG) (5.9,
5,10, 5.11).
Finally, application trials (5.0), to test performance of Moringa MAG have
successfully investigated the suitability of a Moringa oleifera based MAG in 60%
and 40% TAG W/O emulsion spreads. In the case of 40% spreads, these were all
tested with and without PGPR (5.4; 5.7; 5.8; 5.9; 5.11).
323
5.12.4 Modified Crystalliser Composition
It has been previously shown in discussion (5.12.2 & 5.12.3), that a minor
ingredient can be strongly effective at influencing crystallisation and interfacial
behaviour (3.0; 4.0; 5.0). An essential feature of CRY110 is that it provides a
relatively rich source of behenic (C22:0), longer chain fatty acid moiety than
commonly used fatty acids (Krog & Larsson 1992). This feature is considered to
offer some important advantages, both from a crystallisation and emulsification
perspective in low and high TAG spreads (1.5). However, it has been
demonstrated that a MAG characterised as behenic based is not suitable as a
standalone solution in low TAG (<41%) based water-in-oil emulsion spreads, and
either causes destabilisation or fails quickly (days) during shelf life (5.7).
A second emulsifier may enhance the first emulsifier in such systems. This is
particularly the case in demanding applications such as low (<41%) TAG based
spreads (Bech et al., 2013). A MAG having a bi-functional effect would be
desirable to provide a composition which can function as both an emulsifier and a
crystallisation improver. An emulsifier able to satisfy the following would provide
unique functions:-
A rich source of saturated fatty acids is interesting for use in W/O emulsions,
particularly low TAG W/O (<41% oil) where it is common industrial practice to
include the use of Polyglycerol Polyricinoleate (PGPR) (1.5). There was strong
evidence to suggest that low TAG emulsions (W/O) could be successfully
assembled without the use of PGPR (5.4; 5.8; 5.9; 5.11). Depending on its
inclusion level, PGPR has tendency to “over stabilise” the emulsion, leading to
adverse sensory (5.8) resilience. Therefore, where there is a preference to include
PGPR, the dosage can be reduced (5.8) thereby offering potential cost-in-use
savings.
It was found that MAG with fatty acid composition (FAC) containing in the region
of 8.5 – 12.3% saturation starting from C20:0 seemed to enhance emulsification
and crystallisation properties (5.0) In natural Moringa MAG, diversity in the FAC
324
starts from the same chain length (C20:0) and there is a significant proportion of
C22:0 and C24:0 (Table 5.10.2). Natural Moringa MAG also contain lower chain
(C16:0 and C18:0) and a high degree of unsaturated fatty acids (~72%). A similar
designed fatty acid composition, using alternatively sourced MAG to mimic a
natural Moringa MAG is considered novel. Wassell et al (2012a; 2012b; 2012c;
2012d; 2012e), describe the use of a MAG based on natural Moringa TAG), where
it could provide a “one-stop” solution for both low and high TAG systems (5.3;
5.10). Optionally, when combined with e.g. PGPR (a co-emulsifier) this may also
provide additional structuring advantages (Bech et al., 2013).
A series of application trials have revealed the possibility to make a synthetic like
Moringa monoglyceride by mimicking the fatty acid composition (5.3; 5.10) of
Moringa oleifera based monoglycerides, thereby overcoming potential restrictions
of supply and dependence on sources of Moringa oleifera TAG (Wassell et al.,
2012a; 2012b; 2012c; 2012d; 2012e; Bech et al., 2013).
Most studies have looked at the effect of high TAG Pickering stabilisation
compared to low TAG Pickering stabilisation. Garti et al., (1998) found sub-
microcrystalline surface-active fully hydrogenated palm stearin, could in the
325
presence of PGPR, stabilise water droplets (25% aqueous phase in 75% TAG
consisting of liquid soybean oil). Ghosh and Rousseau (2011) investigated surface
active fully hydrogenated canola TAG in the presence of GMO (a Glycerol
monooleate based MAG), testing in emulsions consisting 80% liquid canola oil /
20% aqueous phase. Both these studies and others (Johansson et al., 1995) have
shown effects of Pickering stabilisation to reduce coalescence and smaller droplet
size for a given TAG crystal concentration and speed of thermal treatment (Hodge
& Rousseau 2005). However, these were found to be very system specific, and
subsequently this has technical consequences for low TAG W/O emulsions.
Since conducting this thesis, Ghosh and Rousseau (2011) have shown Pickering
stabilisation to be more efficient with TAG (HCO) and GMS (a Glycerol
monostearate based MAG), than with TAG (HCO) in the presence of GMO (a
Glycerol monooleate based MAG). Stabilisation appeared superior with the more
saturated MAG. This could be due to hydrogen bonding with the TAG solvent
(Chen, van Damme, & Terentjev, 2009). It is therefore suggested that when TAG
particles are covered by an adsorbed layer of an emulsifier, the interaction may
strongly influence the macroscopic properties of the TAG continuous phase
(Bergenståhl 2008). Further, if the adsorbed emulsifier is of a liquidus nature e.g.
PGPR, this would cause the Pickering TAG crystals to be more surface-active
(Ghosh & Rousseau 2011).
Rousseau et al. (2003), say in the presence of fewer TAG crystals, stabilisation
must be dependent on the interfacial properties, so that any visco-elastic behaviour
could take place. It would therefore seem to be the case that an interstitial region
326
and its interaction with Pickering surface-active TAG and MAG crystals are
therefore critical. Rousseau et al. (2003), also alluded to this when examining low
(38%) TAG margarine, indicating this aspect had not been examined.
The low TAG W/O emulsions measured in this research had considerably
decreased bulk TAG network. Consequently, network stabilisation would have had
even less influence on interfacial stabilisation. In this situation, Pickering surface-
active TAG / MAG crystals are more important, so that in the presence of a co-
emulsifier such as PGPR or additional MAG (4.0; 5.4; 5.11), this effect would be
enhanced. Additionally, should crystal structures exist as smaller, liquidus and or
nano-arrays, then this might also account for interfacial visco-elastic behaviour,
stabilising the W/O emulsion (4.0; Johansson et al., 1995; Rousseau, 2000).
327
aqueous surface and ester groups of TAG occur (Claesson et al., 1997; Petrov et
al., 1995). The nature of the solvent plays a vital role on surface interactions
(Dedinaite & Campbell 2000). Though not adequately explained here or
understood (5.6), PGPR may induce additional adhesive properties when mixed
with another emulsifier in the presence of water. Additionally, the polar region of
the PGPR may influence the water droplet properties (Ghosh & Rousseau 2009).
Dendrite structures were observed (2-D) microscopically (5.3; 5.5) in the bulk
phase. Logically, this situation 3-D, combined with their proliferation would likely
influence the total G* due to their relative surface area. The phenomenon of
dendrite structures (Ben-Jacob & Garik 1990) observed in this thesis (5.0) is
possibly linked to both saturation, the rate of thermal treatment of the TAG blends
(5.3) and irregularity of the structure magnifying existing wetting properties
(Rousseau 2002). This is possibly linked to the effect of the adsorbed emulsifier
enhancing crystallisation (Marikkara & Ghazali 2011), causing dendrite forms. In
the presence of water and oil, this could be additional explanation for comparative
thickening observed in the W/O emulsions (5.4; 5.7, 5.11), where adhesion is
caused by increased polarity (Rousseau 2002). The additional sites of contact on
wetted TAG crystal due to the presence of adsorbed emulsifier (forming surface
active crystals) e.g. Moringa MAG or PGPR combined with CRY110 (5.6), could
also explain observed rheological and interfacial behaviour (4.0).
328
6.0 Synchrotron Radiation Macrobeam and Microbeam X-ray Diffraction
Studies of Interfacial Crystallisation of Fats in Water-in-Oil Emulsions
6.1 Introduction
6.2 Background
329
of the interface. There would seem to be contradiction, because studies have
shown (Arima et al., 2009; Tanaka et al., 2009) long-chain saturated fatty acid
moieties (behenic rich) to be effective emulsifier additives at the interface of TAG
and water emulsion droplets (O/W).
It has been shown (Wassell et al., 2010a) that some emulsifiers can template with
other emulsifiers resulting in increased strength of the interfacial membrane
between the water and TAG phases so that the surfaces of the water droplets are
either partially or wholly covered, forming monolayers for heterogeneous
crystallisation. This feature has been analysed in TAG oil-in-water emulsions
(Arima et al., 2009), but has not been investigated in water-in-TAG oil emulsions
(Figure 6.1).
330
Figure 6.1Proposed schematic showing interfacial crystallisation at the water-in-
oil interface.
As described earlier (1.5; 4.0; 5.0), it is common practice to use two emulsifiers
(e.g., polyglycerol polyricinoleic acid (PGPR) and monoacylglycerol (MAG)) in
the production of W/O emulsions, in which MAG may act as a templating agent,
whereas PGPR assures emulsion stability.
The effect of PGPR and MAG has been studied in high TAG (>75%) W/O
emulsions (Garti et al., 1998; Ghosh & Rousseau, 2009), but until now it is not
specifically known about the interactive behaviour of a long-chain MAG and
PGPR in low TAG (<41%) W/O emulsions (Wassell et al., 2012).
PGPR is highly surface active and has a large complex structure (molecular weight
(MW) range >4000 g/mol)), compared to MAG which has a smaller molecular
structure (MW 580 g/mol) than PGPR. Interactions of the two emulsifiers have an
interesting effect on interfacial and bulk crystallisation behaviour. According to
available information, the influence of PGPR and MAG containing behenic acid
moiety (monobehenate (MB) previously abbreviated as CRY110 or MB90 (2.0
General Materials & Methods)) has not previously been analysed to determine the
effects of interfacial heterogeneous crystallisation of low TAG (35% oil phase)
W/O emulsions (See Table 2.22 in Materials and Methods (2.0)).
331
Synchrotron radiation X-ray diffraction (SR-XRD) macrobeam and microbeam
small angle X-Ray diffraction (SR-μ-SAXD) analysis, DSC, polarised optical
Microscopy (POM) were used to enable observation of TAG crystals near the W/O
interface because this method can provide microscopic information about
crystallised materials on the order of micrometer to submicrometer dimensions and
provide learning about the interactions between TAGs and emulsifiers in the
continuous TAG phase and at the W/O interface.
Figure 6.2 depicts the DSC cooling thermopeaks of two emulsions using PGPR
alone and PGPR + MB. Although not shown here, the DSC heating thermograms
did not indicate any differences between the two emulsion samples, whereas the
cooling thermopeaks were quite different.
Figure 6.2 DSC cooling thermopeaks taken for (a) PGPR emulsion and (b) PGPR
+ MB emulsion
332
The emulsifiers influenced kinetic crystallisation processes at 16.1°C in the PGPR
emulsion, whereas the crystallisation temperature increased to 17.9 °C in the
PGPR + MB emulsion. Exothermic behaviour resulted in a single peak (PGPR
emulsion) to double peak (18°C and 10 °C in the PGPR + MB emulsion). This
may have been caused by the promotion of the crystallisation of a high-melting
fraction of solid TAG by the addition of MB.
Figure 6.3 Macrobeam SR-XRD patterns of PGPR emulsion taken during heating
and cooling. Unit: nm. Experiment noise is denoted by arrows
333
When the emulsion was heated, both SAXD and WAXD patterns disappeared at
40 °C, due to the melting of the β form. During cooling, crystallisation was
detected with the occurrence of the SAXD peak of 4.65 nm at 14 °C,
corresponding to the exothermic peak of crystallisation observed in the DSC
cooling pattern (Figure 6.2a). The long spacing of 4.65 nm means that the
polymorphic form of these crystals is α form, as the chain axis in the TAG crystals
is normal to the lamellar interface (Sato & Ueno 2011). In addition, the α form
tends to crystallise faster than the other more stable forms during the ambient rate
of cooling. Despite the strong SAXD peak, no strong peaks were detectable in the
WAXD area during cooling (Figure 6.3b). This may be due to the weak diffraction
of the hexagonal subcell packing of α form in the W/O emulsion. Figure 6.4
depicts the melting and crystallisation of the PGPR + MB emulsion examined by
macrobeam SR-XRD. Similar to the PGPR emulsion, the SAXD patterns
confirmed the presence of the β form at 5 °C, as evidenced by a long spacing peak
of 4.07 nm (Figure 6.4a), and the WAXD patterns of strong peaks of 0.462 nm and
weak peaks of 0.386 nm (Figure 6.4b).
334
No strong peak was detectable in the WAXD area when the strong SAXD peak of
4.65 nm appeared. However, a weak WAXD peak of 0.383 nm appeared when the
SAXD peak of 4.15 nm appeared below 10 °C. It is assumed that this form is a
low temperature sub-α form (Yano et al., 1999).
PGPR: The lamellar directions of the fat crystals near the W/O interface are found
parallel to the W/O interfacial plane using PGPR alone as the emulsifier agent.
The azimuthal angle (χ) extension patterns did not reveal this tendency for all
positions in the bulk TAG phase, but it may be assumed that the PGPR membrane
at the W/O interface may induce the interfacial heterogeneous crystallisation
illustrated in Figure 6.1.
PGPR + MB Emulsion: When compared to PGPR only, the two emulsions had
basically common results in terms of lamellar directions of the TAG crystals at the
positions near the W/O interface. These were confirmed by analysis of the χ
extension patterns at various positions surrounding the water droplets. However,
the intensity of the diffracted X-ray beams was less in the PGPR + MB emulsion
than in the PGPR emulsion, possibly due to the formation of tiny TAG crystals
(see Polarised Optical Microscope (POM) micrographs Figure 6.5)) caused
through addition of MB. This effect made analysis of the results of μ-SAXD data
from the PGPR + MB emulsions difficult for calculating the degree of orientation
of the lamellar planes of TAG crystals; these are evaluated by calculating the half
width of χ value (Δχ). A smaller Δχ yields a higher degree of orientation of the
lamellar planes. The lamellar planes of the TAG crystals in the PGPR + MB
335
emulsion were more highly ordered than those in the PGPR emulsion. However,
such a comparison of Δχ values could not be made for all positions of the PGPR +
MB emulsion, hence making it difficult to obtain a clear conclusion about the
preferred orientation of the lamellar planes of the TAG crystals, near the W/O
interface in the PGPR + MB emulsion and in comparison with those in the PGPR
emulsion.
The crystal morphology and dimensions of the fat crystals were observed in the
PGPR emulsion and PGPR + MB emulsion polarised optical microscope (Figure
6.5). Clear differences were detectable between the crystal morphologies of the
two emulsions, in that no large crystal aggregate is observed in the TAG phase in
the PGPR + MB emulsion. Instead, thin layers of bright images were attached to
the water phases (arrows in Figure 6.3). This morphological change can be
explained by the addition of MB increasing the crystallisation temperature, as
confirmed by the cooling experiments of SR-XRD and by the reduction in crystal
size due to increased rates of crystal nucleation.
Figure 6.5 Polarised optical micrographs of W/O emulsions. (a) PGPR emulsion
and (b) PGPR + MB emulsion. Scale bar: 25μm
336
It should be noted that 35% W/O emulsions assembled for these measurements,
originally had a smaller water droplet size (see Table 5.8.4) to that inferred in
Figure 6.5. Due to confidentiality (Wassell et al., 2012a; 2012b; 2012c; 2012d; 2012e)
several samples (containing Moringa oleifera based MAG) were not measured.
The POM images of Figure 6.5 (originally Figure 13 in Wassell et al., 2012) were
the result of heating the sample to 50°C, then cooling to 5°C at a rate of 2°C/min
and keeping at this temperature for 10 minutes. By comparison, this may explain
the difference in droplet sizes observed in this multiple study (5.0) which shows
other examples of low fat (<41%), W/O emulsions (Wassell et al., 2012c).
This study (Wassell et al., 2012) revealed the following results. First, the addition
of MB promoted TAG crystallisation as observed by DSC and macrobeam SR-
XRD analysis. Second, the microbeam SR-XRD indicated that the lamellar planes
of TAG crystals near the water and oil interfaces were arranged almost parallel to
the interface planes in both the PGPR emulsion and the PGPR + MB emulsion. No
appreciable differences between them were confirmed. Third, the POM
observations indicated that adding MB eliminated the formation of large crystal
aggregates, resulting in the formation of tiny TAG crystals.
Although PGPR contains large amounts of liquidus fatty acid moieties at the
temperatures examined in the present μ-SAXD experiments, it can be assumed that
certain molecular interactions between PGPR and TAG molecules may be
operating at the W/O interface to cause interfacial crystallisation to some extent,
and the addition of MB may strengthen this interfacial crystallisation. Such
interactions may include not only hydrophobic chain-chain interactions between
the fatty acid moieties of the emulsifiers and TAG molecules, but also hydrophilic
interactions between the polar groups of the emulsifiers and glycerol groups of
TAG molecules (Ghosh & Rousseau, 2011; Ghosh et al., 2011).
337
Regarding the effects of adding MB on intensifying the interfacial crystallisation
illustrated in Figure 6.1, which was considered plausible to observe at the
beginning of this study, no significant evidence was observed, except for the
promotion of TAG crystallisation in the continuous TAG phase and close to the
water phase perimeter. A possible reason may be low adsorption of MB at the
W/O interface in cooperation with PGPR membranes, whose adsorbability exceeds
that of MB.
The polymorph behaviour of Pickering TAG and MAG (Malkin, 1954) crystalline
material has been examined but requires more research. Possibly a relationship
exists between their form, quantity and quality. Hodge and Rousseau (2005) found
in their study of W/O emulsions that TAG based on hydrogenated canola was
more resistant to destabilisation in the β-form (Basso et al., 2010). When
comparing the relative β intensity (R β) within W/O emulsions, Shiota et al.
(2011), using TAG blends prepared with MAG based on palmitic, stearic, and
behenic acid, found that the influence of the behenic based MAG resulted in the
lowest (R β) within W/O emulsions.
It is known that distilled MAG crystallise into metastable α-crystal, then transform
into higher melting β-form. Metastable β’-form does not occur in commercial
MAG. On cooling, a solid state transition occurs to sub-α (Krog, N., Danisco /
DuPont TP18-1e). The behaviour of MAG / TAG in hydrophobic solutions is
shown not to be the same or similar as in water, because hydrogen bonding of
glycerol groups would be different (Chen et al., 2009). The lamellar phase
338
structure appears more ordered at a low temperature, where a sub-α form result in
rigidity of lamellar bilayers (Chen et al., 2009). It is not known if assumed low
temperature sub-α form seen in macrobeam SR-XRD patterns of PGPR + MB
(CRY110) emulsion (Wassell et al., 2012) could also explain interfacial (4.0) and
macroscopic (5.0) visco-elastic properties (Rousseau & Hodge 2005; Bergenståhl
2008). These areas require more research.
It should be noted that a Moringa oleifera based MAG in a 35% TAG W/O
emulsion (5.4) was assembled for μ-SAXD measurement, but not analysed (6.3.3)
due to confidentiality (WIPO applications: Wassell et al., 2010a; 2012; 2012a;
2012b; 2012c; 2012d; 2012e). Analysis of this sample would still be ideal in order to
compare with those measured previously (Wassell et al., 2012).
339
7.0 General Discussion, Final Conclusions and Recommendations
Since the commencement of this thesis, both a literature and critical review
(Wassell & Young 2007; Wassell et al., 2010a) have clearly shown the need for a
multiple focus approach to solving some structuring problems of TAG based water
– oil systems, specifically those already characterised as reduced and low fat.
In contrast, this study investigated the effects of a MAG that contains a significant
content of C22:0 in anhydrous bulk fats (3.0; 4.0; 5.0). It then investigated the
same effect in emulsions with aqueous phase contents initially at 82% (1.5), then
between 35% and 65% (5.0; 6.0). These emulsions systems have inherently
reduced total TAG (containing SAFA) and are more complex to stabilise. Further
TAG reduction whilst avoiding impact on structure and functional properties
(Wassell & Young 2007; Wassell et al., 2010a) is more problematic where a food
is already regarded as low saturated (1.4; 1.6; 1.7).
This study has investigated the effects of emulsifier on the physical behaviour,
both to the TAG, and at the TAG/water interface and shows new information
340
about interfacial properties to be of considerable importance (Wassell et al.,
2010a; 2012; 2012a; 2012b; 2012c; 2012d; 2012e; Bech et al., 2013).
In contrast to other studies (1.6; 1.7), this study (Wassell & Young 2007; Wassell
et al., 2010a) has shown that C22:0 rich MAG, or its significant inclusion, has a
pronounced effect on crystallisation (Wassell et al., 2010b; 2012; Young et al.,
2008) and interfacial kinetics (4.0). New interfacial measurements clearly
demonstrated an unusual surface-interactive relationship of longer chain MAG
compositions, both with and without PGPR in W/O emulsions and anhydrous
TAG systems. This is thought to partially explain textural behaviour observed in
multiple application trials (1.5; 5.0; 6.0; Wassell et al., 2010b; 2012; Young et al.,
2008).
Results show that the influence on interfacial tension by a single MAG emulsifier
is more pronounced with long-chain MAG than with medium-chain MAG (4.0).
The relative decrease in interfacial tension upon decreased temperature, was
greater the longer the chain length. While this confirms studies by others (Krog &
Larsson 1992), this thesis moreover, now provides new additional insight, because
bulk and interfacial rheology showed that the presence of C22:0 fatty acids have a
pronounced effect on both G’ and G’’. This effect was more manifest in the
presence of an aqueous phase.
341
by C22:0 fatty acids. This resulting interactive behaviour may partially explain
increased textural resilience (thickening) due to viscoelasticity (4.0; 5.0).
Application studies of Moringa oleifera based MAG in low TAG (35% - 41%)
W/O emulsions, resulted with high emulsion stability without a co-surfactant
(PGPR) 4.0; 5.0)). These results provided positive verification of the interfacial
and rheological analyses (4.0, 5.6). This may offer technological advantage, where
under certain circumstances, a co-emulsifier (PGPR) can be avoided (1.7).
When tested separately, both a Moringa oleifera based MAG and PGPR seemed to
have similar influence on the crystallisation kinetics of the particular anhydrous
TAG systems (5.3; 5.5; 5.6). A combination of either of these two emulsifiers with
a behenic rich (CRY110) MAG, resulted in enhanced gelation onset behaviour,
which was dependent on the rate of thermal treatment and selected solvent (5.5;
5.6).
342
Additional application studies confirmed the potential to mimic the fatty acid
composition of Moringa oleifera TAG (5.3; 5.10), meaning restriction of supply
and dependence on a particular TAG source is now reduced (Bech et al., 2013).
It is thought that PGPR, which has liquidus fatty acid moieties at the temperatures
examined in the μ-SAXD measurements, may have certain molecular interactions
with TAG molecules at the W/O interface to cause a degree of interfacial
crystallisation. The addition of a MAG containing long-chain C22:0 may
strengthen the interfacial crystallisation (4.3.8) of liquidus interstitial crystals near
the aqueous droplets (Shiota et al, 2011).
The following sections now examine the implications of several of these findings,
summarise the key results and then discuss implementation of innovative
outcomes from this study and future use.
343
7.1.1 Implications for Pickering Stabilisation
Studies and outcomes in this thesis (1.4; 1.5; 4.0; 5.0; 6.0; Wassell et al., 2012b),
collectively now provide new insight concerning the effects of Pickering surface-
active MAG (with significant content of >C20:0) in reduced and low TAG W/O
emulsions. The low TAG W/O emulsions measured in this research had
considerably decreased bulk TAG network. Network stabilisation would have had
less influence on interfacial stabilisation. In this situation, it is likely that Pickering
surface-active TAG / MAG crystals become strongly mobilised to the interface
due to dielectric interaction.
344
causing a steric, interstitial stabilised interface to be formed (Johansson,
Bergenståhl & Lundgren, 1995).
345
7.2 Key Findings:
A Moringa oleifera based MAG can successfully stabilise low TAG (35%
- 41%) WO emulsions without PGPR (5.4; 5.7; 5.8; 5.11). Moreover, a
Moringa oleifera based MAG is distinctly more functional and contrasted
in performance to a behenic rich (CRY110) MAG used as a single
emulsifier (5.7; 5.8). Maintaining the whole fatty acid composition, close
to the original (Moringa oleifera) TAG profile is shown to be more
functional (5.10., 5.11; 5.12).
346
7.3 Novelty and Future Context
The literature (1.4; 1.6) shows health and sustainability have become important
reasons for finding alternative lipid materials for the function of structuring /
stabilising food products (Beaglehole et al., 2011; Bradley 2012; Carmichael 2011;
Gortmaker et al., 2011; Menaa et al., 2013; Wang et al., 2012; Wassell & Young
2007; Wassell et al., 2010a;). The reformulation of reduced fat systems is not
simple (Pothiraj et al., 2012; Wassell et al., 2010a; Windwood, 2011).
While low fat emulsion technology is not new, the use of one or several novel
applications of long-chain e.g. behenic (C22:0) based emulsifiers to low or
reduced TAG W/O emulsions was not found in the literature (1.4; 1.6). This novel
research aimed to observe rheological and interfacial crystallisation behaviour in
W/O emulsions and anhydrous bulk TAG (Wassell et al., 2010b; Young et al.,
2008; 4.0; 5.0; 6.0). It has discovered potential advantages for the inclusion of
certain concentrations of saturated long chain fatty acids for stabilising and
emulsifying low TAG (<41%), W/O emulsions and dispersions. A study of
Moringa oleifera as a new TAG material has resulted in the development of a
MAG, which is a novel approach to stabilising TAG based W/O emulsions or
dispersions (Wassell et al., 2012a; 2012b; 2012c; 2012d; 2012e).
347
From this multidisciplinary study, there are resulting inventions (Patents) that
contribute towards consumer requirements for foods that satisfy the current and
future demand for natural, sustainable sources of raw materials. These are briefly
described:-
Practical application of these inventions may also potentially extend to areas such
as: pharmaceutical products, cosmetics and bio-fuel.
348
7.4 Recommended Research
Low TAG (<41%) W/O emulsions contain a large dispersed aqueous phase, whose
composition cannot be overlooked (Keogh 2006; Rousseau et al., 2003), because
ionic strength may influence hydrophilic stabilisation (Larsson et al., 1969). This
aspect was not within the scope of this thesis. Scherze et al (2006), suggest
electrolytes may have a bearing on the rigidity of the interfacial film, because in
TAG (70%) based W/O emulsions they found PGPR in the presence of NaCl
349
seemed to stiffen the interface. Further research may demonstrate this is caused by
influence to the polar region of the PGPR molecule.
Research from studies within other disciplines may bring insight into high internal
phase emulsions (Peng et al., 2009) where the influence of the droplet curvature
must be considered (Shinohara et al., 2008). In turn dielectric behaviour may
possibly influence mobility of surface-active crystals to the interface stronger than
van der Waals forces where there is decreasing bulk TAG network, as in low
(<41%) TAG emulsions (Goubran & Garti 1988). These factors could also have
implications in more complex emulsions e.g. O/W/O or W/O/W.
350
With reference to Figures 6.1 and 6.5 (6.0), subsequent studies regarding emulsion
stabilisation should give attention to the interstitial region. Figure 7.1 shows some
important elements to consider at the W/O interface. In addition, there are
questions regarding the interrelationship of aspects such as super-cooling / shear,
time / adsorption development, crystal structure / morphology and aqueous phase
behaviour at the interfacial region, which are areas for further research.
351
352
References
Abdulkarim, S.M., Long, K., Lai, O.M., Muhammad, S.K.S., Ghazali, H.M.,
(2005). Some physico-chemical properties of Moringa oleifera seed oil
extracted using solvent and aqueous enzymatic methods. Food Chemistry. 93,
253–263.
Arima, S., Ueji, T., Ueno, S., Ogawa, A. and Sato, K. (2007). Retardation of
crystallisation-induced destabilization of PMF-in-water emulsion with
emulsifier additives. Colloids and Surfaces B: Biointerfaces, 55, 98–106.
Arima, S., Ueno, S., Ogawa, A. and Sato, K. (2009). Scanning microbeam
small-angle X-ray diffraction study of interfacial heterogeneous crystallisation
of fat crystals in oil-in-water emulsion droplets. Langmuir, 25, 9777–9784.
353
Awad, T.S., Moharram, H.A., Shaltout, O.E., Asker, D. and Youssef, M.M.
(2012). Applications of Ultrasound in Analysis, Processing and Quality
Control of Food: A Review, Food Research International. 48 (2), 410-427.
Basheva, E.S., Gurkov, T.D., Ivanov, I.B., Bantchev, G.B., Campbell, B., and
Borwankar, R.P., (1999). Size Dependence of the Stability of Emulsion Drops
Pressed against a Large Interface. Langmuir. 15: 6764-6769.
Basso, R.C., Ribeiro, A.P.B., Masuchi, M.H., Gioielli, L.A., Gonçalves, LA.
G., Santos, A.O., Cardoso, L.P. and Grimaldi, R. (2010). Tripalmitin and
monoacylglycerols as modifiers in the crystallisation of palm oil, Food
Chemistry, 122, 1185–1192.
Basiron, Y. (2005). Palm oil in: Bailey’s, Industrial Oil and Fat Products, 6th
ed. (edited by F. Shahidi). New York: John Wiley & Sons.
Bech, A.T., Farmer, M., Forrest, B.A., Wassell, P., Young, N.W.G., DuPont
Nutrition Biosciences APS, (2013). Composition. World Intellectual Property
Organisation No. WO2013050944.
Bell, A., Gordon, M.H., Jirasubkunakorn, W. and Smith, K.W. (2007). Effects
of composition on fat rheology and crystallistation. Food Chemistry, 101, 799–
805.
354
Bergenståhl, B. (2008). Physicochemical Aspects of an Emulsifier
Functionality. In: G.L. Hasenhuettl and R.W. Hartel (eds.), Food Emulsifiers
and Their Applications. 2nd ed. Springer, New York.
Berkowitz, M.L. and Vácha, R., (2012). Aqueous Solutions at the Interface
with Phospholipid Bilayers. Accounts of Chemical Research, 45, 74-82.
Birkhofer, B., Jeelani, S.A.K., Windhab, J., et al. (2008). Monitoring of fat
crystallisation process using UVP-PD technique. Flow Measurement and
Instrumentation, 19, 163–169.
Bockisch, M. (1998). Fats and Oils Handbook. Champaign, AOCS Press, 838.
Boodhoo, M.V., Humphrey, K.L. and Narine, S.S. (2009). Relative hardness of
fat crystal networks using force displacement curves. International Journal of
Food Properties, 12, 1129–1144.
Bot, A., den Adel, Regkos, C., Sawalha, H., Venema, P., and Flöter, E. (2011).
Structuring in β-sitosterol + γ-oryzanol-based emulsion gels during various
stages of a temperature cycle. Food Hydrocolloids 25, 639-646.
Bot, A., Veldhuizen, Y.S.J., del Adel, R. and Roijers, E.C. (2009). Non-TAG
structuring of edible oils and emulsions. Food Hydrocolloids, 23, 1184–1189.
Boyd, J.V., Krog, N., Sherman, P., (1976). Comparison of Rheological Studies
on Adsorbed Emulsifier Films with X-Ray Studies of the Bulk Solutions, in:
A.L. Smith, edit., Proceedings of a Symposium by Society of Chemical
Industry, Brunel University, September 16-18, 1974., Theory and Practice of
Emulsion Technology., Academic Press. 123-133.
Bradley, P. (2012). Diet composition and obesity. The Lancet, 379, (9821),
1100.
355
van Camp, D., Hooker, N.H., and Lin, C.J., (2012). Changes in fat contents of
US snack foods in response to mandatory trans fat labelling. Public Health
Nutrition. 15, (6), 1130-1137.
Cerdeira, M., Martini, S., Hartel, R.W. and Herrera, M.L. (2003). Effect of
sucrose ester addition on nucleation and growth behaviour of milk fat-
sunflower oil blends. Journal of Agricultural Food and Chemistry, 51, 6550–
6557.
Cerdeira, M., Pastore, V., Vera, L.V., Martini, S., Candal, R.J. and Herrera,
M.L. (2005). Nucleation behaviour of blended high-melting fractions of milk
fat as affected by emulsifiers. European Journal of Lipid Science and
Technology, 107, 877–885.
Cerdeira, M., Martini, S., Candal, R.J. and Herrera, M.L. (2006).
Polymorphism and growth behaviour of low-trans fat blends formulated with
and without emulsifiers. JOACS, 83, 489–496.
Chen, C.H., Van Damme, I., Terentjev, E.M., (2009). Phase Behavior of C18
monoglyceride in hydrophobic solutions, Soft Matter, 5, 432-439.
Chrysan, M.M. (2005). Margarines and Spreads in: Bailey’s, Industrial Oil and
Fat Products, 6th edn (edited by F. Shahidi). New York: John Wiley & Sons.
356
Corke, H. (2007). “Speciality Cereal and Noncereal Starches.” In: The RVA
Handbook , Eds. Crosbie, G.B., and Ross, A.S. AACC International, USA: 49-
62.
Delamarre, S. and Batt C.A. (1999). Review – The microbiology and historical
safety of margarine. Food Microbiology, 16, 327─333.
Dierig, D.A. and Thompson, A.E., (1993). Vernonia and Lesquerella potential
for commercialization. In: J. Janick and J.E. Simon, Editors, New Crops.
Proceedings of the Second National Symposium. New Crops: Exploration,
Research, Commercialization Indianapolis, Ind., October 6–9, 1991, John
Wiley & Sons, Inc., New York, 362–367.
Drewnowski, A., Shrager, E.E., Lipsky, C., Stellar, E. and Greenwood, M.R.C.
(1989). Sugar and fat: sensory and hedonic evaluations of liquid and solid
foods. Physiology and Behavior, 45, 177–183.
Duffy, N., Blonk, H.C.G., Beindorff, C.M., Cazade, M., Bot, A. and
Duchateau, G.S. (2009). Organogel-based emulsion systems, microstructural
features and impact on in vitro digestion. JAOCS, 86, 733–741.
357
van Duijn, G., Dumelin, E.E. and Trautwein, E.A. (2006). Virtually trans free
oils and modified fats. In: Improving the Fat Content of Foods (edited by C.
Williams & J. Buttriss). p 490–507. Cambridge: Woodhead Publishing.
Euston, S.R. (2008). Emulsifiers in Dairy Products and Dairy Substitutes. In:
G.L. Hasenhuettl and R.W. Hartel (eds.), Chp 7., Food Emulsifiers and Their
Applications (2ed). Springer, New York.
Fahey, J.W., (2005). Moringa oleifera: A Review of the Medical Evidence for
Its Nutritional, Therapeutic, and Prophylactic Properties. Part 1. Trees for Life
Journal (1:5)., http://www.TFLJournal.org/article.php/20051201124931586
Flöter, E. and van Duijn, G. (2006). Trans-free fats for use in food. In:
Modifying Lipids for Use in Food (ed. F.D. Gunstone). 429–443. Cambridge:
Woodhead Publishing.
Ford, R. (2012). Weight-Watchers gets back into dairy spreads with Half Fat,
The Grocer, William Reed Business Media Ltd., UK.
Foster, T., Russell, A., Farrer, D. et al. (2007). Instant emulsions. In: Food
Colloids – Self Assembly and Material Science (edited by E. Dickinson & M.
Leser). 413–422. London: RSC Publishing.
Food Standards Agency (FSA) (2008). Saturated fat and energy intake
programme. http://www.food.gov.uk/multimedia/pdfs/satfatprog.pdf (last
accessed 15th April 2013).
358
Fujiwara, K., Nagahisa, S., Yano, J., Ueno, S. and Sato, K. (2000). Kinetics of
heterogeneous nucleation, of n-alcohol crystals from solution assisted by
template thin films of monoacylglycerols and fatty acids. Journal of Physical
Chemistry, 104, 8116–8123.
Gortmaker, S.L., Swinburn, B.A., Levy, D., Carter, R., Mabry, P.L., Finegood,
D.T., Huang, T., Marsh, T. and Moodie, M.L. (2011). Changing the future of
obesity: science, policy, and action. The Lancet. 378, (9793), 838-847.
359
Goubran, R.. and Garti, N. (1988). Stability of W/O Emulsions Using High
Molecular Weight Emulsifiers, J. Dispersion Sci. Technol. 9 (2), 131–148.
Gülseren, I., and Corredig, M., (2012). Interactions at the interface between
hydrophobic and hydrophilic emulsifiers: polyglycerol polyricinoleate (PGPR)
and milk proteins, studied by drop shape tensiometry, Food Hydrocolloids. 29,
(1) 193–198.
Himawan, C., Starov, V.M. and Stapley, A.G.F. (2006). Thermodynamic &
kinetic aspects of fat crystallisation. Advances in Colloid and Interface
Science, 122, 3–33.
Hodge, S.M. and Rousseau, D., (2005). Continuous-phase fat crystals strongly
influence water-in-oil emulsion stability, J. Am. Oil Chem. Soc., 82, (3), 159–
164.
Hughes, N., Marangoni, A.G., Wright, A.J., Rogers, M.A. and Rush, J.W.E.
(2009). Potential food applications of edible oil organogels. Trends in Food
Science & Technology, 20, 470–480.
360
ICI Americas Inc. (1980). The HLB System: A Time-saving Guide to
Emulsifier Selection. Wilmington, DE 19897, USA.
Ishizuka, C., Ahmed, T., Arima, S. and Aramaki, K. (2009). Viscosity boosting
effect of added ionic surfactant in nonionic wormlike micellar aqueous
solutions. Journal of Colloid and Interface Science, 339, 511–516.
Keogh, M.K., (2006). Chemistry and Technology of Butter and Milk Fat
Spreads in: Advanced Dairy Chemistry. Vol 2: Lipids, 3rd edition. Ed. P.F.
Fox and P.L.H. McSweeney, Springer, New York. 333-363.
361
Kristensen, C.J. and Wassell, P. (2006). Diglycerides in palm oil – influence on
crystallisation behaviour in production of puff pastry margarine. Poster and
abstract from 4th Euro Fed Lipid Congress, 1–4 October 2006, University of
Madrid (UCM), Madrid, Spain. 491.
Kristensen, C.J., Wassell, P., Mikkelsen, J.D. and Søe, J.B. (2005). Enzymatic
treatment of oils. Danisco
⁄ S, ADenmark, International patent:
WO⁄2005⁄066351.
Kristott, J.U. (1999). Physical Properties of Edible Vegetable Oils. In: LFRA
Oils and Fats Handbook Series, Vol. 1. Vegetable Oils and Fats. (B. Rossell,
ed.). Leatherhead Food RA, Leatherhead UK.
Krog, N. (1975). Interactions between water and surface active lipids in food
systems. In: DUCKWORTH, R. B. (ed.). Water relations of foods. Academic
Press: New York.
Krog, N.J. (1997). 'Food emulsifiers and their Chemical and physical
properties', in Friberg, S.E. and Larsson, K., (eds). Food Emulsions, Third
Edition, Revised and Expanded. New York, Marcel Dekker. 141-188.
Krog N. (2001). Danisco A/S Technical Paper- TP-17., New Insights in Food
Emulsions. Paper: South African Association for Food Science and
Technology 16th Int. Congress. Durban Elangeni, 10-12 September, 2001 by:
N. Krog, Danisco A/S, DK-8220 Brabrand, Denmark.
362
Krog, Niels and Vang Sparsø, Flemming (2003). Food Emulsifiers Their
Chemical And Physical Properties. In Food Emulsion: - Edited by: Stig E.
Friberg, Kåre Larsson and Johan Sjöblom.
Lee, H.O., Jiang, T.L., Jiang, TS. and Avramdis, K.S. (1991). Measurements of
Interfacial Shear Viscoelasticity with an Oscillatory Torsional Viscometer. J.
Colloid Interface Sci. 146 (1) 90-122.
Lin, C., He, G.H., Dong, C.X., Liu, H.J., Xiao, G.K. and Liu, Y.F. (2008).
Effect of Oil Phase Transition on Freeze/Thaw-Induced Demulsification of
Water-in-Oil Emulsions, Langmuir 24 (10) 5291–5298.
Lupi, F.R., Gabriele, D., de Cindio, B., (2011a). Effect of Shear Rate on
Crystallisation Phenomena in Olive Oil-Based Organogels. Food and
Bioprocess Technology. Food Biopr Tech., 4: 111-17.
Lupi, F.R., Gabriele, D., de Cindio, B., Sánchez, M.C., Gallegos, C., (2011b).
A rheological analysis of structured water-in-olive oil emulsions, Journal of
Food Engineering, 107, (3–4), 296-303.
363
Madsen, J. (1987). Emulsifiers used in margarine, low calorie spread,
shortening bakery compound and filling. Fett-Wiss. Technol., 89, 165-172.
Marangoni, A.G. (2005). Fat Crystal Networks. Marcel Dekker, New York.
Marangoni, A.G., Acevedo, N., Maleky, F., Co, E., Peyronel, F., Mazzanti, G.,
Quinn, B., and Pink, D. (2012). Structure and functionality of edible fats. Soft
Matter, 8, 1275-1300.
Marikkara, J.M.N. and Ghazali, H.M. (2011). Effect of Moringa Oleifera Oil
Blending on Fractional Crystallization Behavior of Palm Oil. International
Journal of Food Properties. 14 (5), 1049-1059.
Martini, S., Awad, T. and Marangoni, G., (2006). Structure and properties of
fat crystal networks. In: Modifying lipids for use in food, Ed by Gunstone, F.D.,
Woodhead publishing, England. 142 – 169.
364
Mazzanti, G., Guthrie, S.E., Sirota, E.B., Marangoni, A.G. and Idziak, S.H.J.
(2003). Orientation and phase transition of fat crystals under shear. Crystal
Growth & Design. 3, 721-725.
Mazzanti, G., Marangoni, A.G., Guthrie, S.E., Idziak, S.H.J., Sirota, E.B.,
(2005). Crystallisation of Bulk Fats Under Shear in: Soft Materials Structure
and Dynamics. Dutcher, J.R. and Marangoni, A.G. (Eds). Marcel Dekker, N.Y.
McClements, D.J. and Povey, M.J.W. (1987). Solid fat content determination
using ultrasonic velocity measurements. International Journal of Food Science
and Technology, 22, 491–499.
McClements, D.J. and Povey, M.J.W. (1988). Comparison of pulsed NMR and
ultrasonic velocity techniques for determining solid fat content. International
Journal of Food Science and Technology, 23, 159–170.
McNeill, G.P. (2009). Saturated fats and the risk of heart disease. inform,
AOCS, 20, 340–341.
365
Metin, S. and Hartel, R.W. (2005). Crystallization of Fats and Oils, in:
Bailey’s, Industrial Oil and Fat Products, 6th ed (edited by F. Shahidi). New
York: John Wiley & Sons.
Mohammed, A.S., Lai, O.M., Muhammad, S.K.S., Long, K., and Ghazali,
H.M. (2003). Moringa oleifera, Potentially a New Source of oleic acid-type oil
for Malaysia. In: Investing in Innovation 2003, Vol 3: Bioscience and
Biotechnology, ed. Mohd. Ali Hassan et al., pp 137-140. Universiti Putra
Malaysia Press, Serdang Press, Selangor, Malaysia
Mojet, J. and Koster, E.P. (2005). Sensory memory and food texture. Food
Quality and Preference, 16, 251–266.
Moziar, C., De Man, J.M. and De Man, L. (1989). Effect of Tempering on the
Physical Properties of Shortening, Can. Inst. Food Sci. Technol. J. 22: 238–
242.
Nagarajan, R., Chung, S.I. and Wasan, D.T. (1998). Biconical Bob Oscillatory
Interfacial Rheometer. J. Colloid Interface Sci. 204, 53-60.
Narine, S.S. and Marangoni, A.G. (1999). Relating structure of fat crystal
networks to mechanical properties: a review, Food Research International, 32
(4), 227–248.
366
Narine, S.S and Humphrey K.L., (2004). A comparison of lipid functionality as
a function of molecular ensemble and shear: Microstructure, polymorphism,
solid fat content and texture. Food Res Int, 37, 28-38.
Oh, S.G. and Slattery, J.C. (1978). Disk and Biconical Interfacial Viscometers.
Journal of Colloid and Interface Science. 67 (3) 516-525.
Okamura, A., Wassell, P., Young, N.W.G., Bonwick, G., Smith, C., Almiron-
Ruig, E., Sato, K., Ueno, S., (2011). Influence of Emulsifiers in W/O Low Fat
Spreads for Fat Crystallization. (PC-007) Physical chemistry poster session.
9th Euro Fed Lipid Congress, Rotterdam, 18-21 September 2011.
Pandey, A., Pradheep, K., Gupta, R., Roshini Nayar, E., Bhandari, D.C.,
(2011). ‘Drumstick tree’ (Moringa oleifera Lam.), a multipurpose potential
species in India, Genetic Resources and Crop Evolution. 58 (3), 453-460.
Pernetti, M., van Malssen, K.F., Flöter, E. and Bot, A. (2007). Structuring of
edible oils by alternatives to crystalline fat. Current Opinion in Colloid &
Interface Science, 12, 221–231.
Peng, J., Xia, H., Liu, K. et al. (2009). Water-in-oil gel emulsions from a
cholesterol derivative Structure and unusual properties. Journal of Colloid and
Interface Science, 336, 780–785.
367
Petrov, P., Miklavcic, S., Olsson, U., Wennerstroem, H. (1995). A Confined
Complex Liquid. Oscillatory Forces and Lamellae Formation from an L3
Phase. Langmuir, 11 (10), 3928–3936.
Piller, D., (2011). Innovations in the “post-trans” era of fats and oils. inform.
AOCS., 22(10), 647-648.
Podmore, J. (2008). Food applications of trans fatty acids. In: Trans Fatty
Acids (edited by A. Dijkstra, R.J. Hamilton & W. Hamm). 203–217. London,
UK: Blackwell Publishing.
Pothiraj, C., Zuñiga, R., Simonin, H., Chevallier S. and Le-Bail, A. (2012).
Methodology assessment on melting and texture properties of spread during
ageing and impact of sample size on the representativeness of the results.
Journal of Stored Products and Postharvest Research. 3 (10), 137 – 144.
Prakash, M.N.K. and Ramana, K.V.R. (2003). Ultrasound and its application in
the food industry. Journal of Food Science and Technology, 40, 563–570.
van Putte, K. and van den Enden, J.C. (1974). Fully automated determination
of solid fat content by pulsed NMR. JAOCS, 51, 316–320.
368
Qiao, B.F., Sega, M., Holm, C., (2012). Properties of water in the interfacial
region of a polyelectrolyte bilayer adsorbed onto a substrate studied by
computer simulations. Phys. Chem. Chem. Phys 14, 11425–11432.
Ray, Y.C., Lee, H.O., Jiang, T.L. and Jiang, TS. (1987). Oscillatory Torsional
Interfacial Viscometer. J. Colloid Interface Sci. 119 (1) 81-99.
Rousseau, D., Hodge, S.M., Nickerson, M.T. and Paulson, A.T. (2005).
Regulating the Beta prime to Beta polymorphic transition in food fats, JAOCS,
82 (1), 7-12.
Rousseau, D., Zilnik, L., Khan, R. and Hodge, S. (2003). Dispersed phase
destabilization in table spread. J Am Oil Chem Soc. 80, 957-961.
369
Sahasranamam, U.M. (2005). EU Patent EP1552751 A1 Trans free hard
structural fat for margarine blend and spread.
Sakamoto, M., Mauro, K., Kuiryama, J., Kouno, M., Ueno, S. and Sato, K.
(2003). Effects of adding polyglycerol behenic acid esters on the crystallisation
of palm oil. Journal of Oleo Science, 52, 639–645.
Sakamoto, M., Ohba, A., Kuiryama, J., Mauro, K., Ueno, S. and Sato, K.
(2004). Influences of fatty acid moiety and esterification of polyglycerol fatty
acid esters on the crystallisation of palm mid fraction in oil-in-water emulsion.
Colloids and Surfaces, B, Biointerfaces, 37, 27–33.
Sato, K. (1989). Crystallization of fats and fatty acids. In: Crystallisation and
polymorphism of fats and fatty acids, Ed. by Garti N. and Sato K., 227-260.
Schaink, H.M., van Malssen, K.F., Morgado-Alves, S., Kalnin, D. and van der
Linden, E. (2007). Crystal network for edible oil organogels: possibilities and
limitations of the fatty acid and fatty alcohol systems. Food Research
International, 40, 1185–1193.
Shigemi, U. (2006). Patent: Plastic Oil and Fat Composition and Method for
Producing the Same. Japan: Fuji Oil Co. Ltd. JP2006129819.
370
Shimada, A. and Ohashi, K. (2003). Interfacial and Emulsifying Properties of
Diacylglycerol. Food Sci. Technol. Res., 9 (2).142- 147.
Shinohara, Y., Takamizawa, T., Ueno, S., Sato, K., Kobayashi, I., Nakajima,
M. and Amemiya., Y. (2008). Microbeam X-ray diffraction analysis of
interfacial heterogeneous nucleation of n-hexadecane inside oil-in-water
emulsion droplets. Cryst. Growth Des. 8, 3123–3126.
Shiota, M., Iwasawa, A., Kotera, M., Konno, M., Isogai, T., Tanaka, L.,
(2011). Effect of Fatty Acid Composition of Monoglycerides and Shear on the
Polymorph Behavior in Water-in-Palm Oil-Based Blend. J Am Oil Chem Soc.
88 (8) 1103-1111.
Silva, H.D., Cerqueira M.A., and Vicente, A.A., (2011). Nanoemulsions for
Food Applications: Development and Characterization. Food and Bioprocess
Technology. 5, (3), 854-867.
Smith, K.W., Bhaggan, K., Talbot, G. and van Malssen, K.F. (2011).
Crystallization of Fats: Influence of Minor Components and Additives.
JAOCS, 88, (8), 1085-1101.
371
Stender, S., Astrup, A. and Dyerberg, J. (2009). What went in When trans went
out? New England Journal of .Medicine, 361, 314–316.
Stewart, M.F. and Hughes, E.J., (1972). Polyglycerol Esters as Food Additives,
Process Biochemistry, 7 (12), 27-28.
Swinburn, B.A., Sacks, G., Hall, K.D., McPherson, K., Finegood, D.T.,
Moodie, M.L. and Gortmaker, S.L. (2011). The global obesity pandemic:
shaped by global drivers and local environments. Lancet, 378, (9793), 804-814.
Talbot, G., Favre, L. and ThÖrig, L. (2007). Palm oil: The healthy alternative
to trans fats. Inform, 18, 198–200.
Tanaka, L., Tanaka, K., Yamato, S., Ueno, S. and Sato, K. (2009). Microbeam
X-ray diffraction study of granular crystals formed in water-in-oil emulsion.
Food Biophysics, 4, 331–339.
Timms, R.E. (1985). Physical properties of oils and mixtures of oils. Journal of
American Oil Chemists Society, 62, 241–249.
Timms, R.E. (1991). Crystallisation of Fats, Chemistry & Industry (SCI), May
20th, 342-344.
Timms, R.E. (2007). Fat Crystallisation: mechanism and methods for studying
OFI Middle East: SCI Technical Conference Edible Oils and Fats Lecture-
Trends in Raw Materials, Processing and Applications 20–21 March 2007,
Cairo
372
Toro-Vazquez, J.F., Dibildox-Alvarado, E., Herrera-Coronado, V. and Charo-
Alonso M.A. (2001). Triglyceride crystallization in vegetable oils: application
of models, measurements, and limitations, in Crystallization and Solidification
Properties of Lipids. Ed by Widlak N, Hartel R and Narine S. AOAC Press,
Champaign, IL, 53–78.
Ueno, S., Suetake, T., Yano, J., Suzuki, M., Sato, K., (1994). Structure and
polymorphic transformations in elaidic acid, (trans-ω9-octadecenoic acid).
Chemistry and Physics of Lipids. 72, 27-34.
Ueno, S., Nishida, T., Sato, K., (2008). Synchrotron Radiation Microbeam X-
ray Analysis of Microstructures and the Polymorphic Transformation of
Spherulite Crystals of Trilaurin. Cryst. Growth Des. 8 (3), 751–754.
Vereecken, J., Foubert, I., Smith, K.W., and Dewettinck, K., (2009a). Effect of
SatSatSat and SatOSat on crystallisation of model fat blends, Eur J, Lipid Sci,
Technology, 111, 243-258.
Vereecken, J., Foubert, I., Meeussen, W., Lesaffer, A., and Dewettinck, K.,
(2009b). Fat Structuring with partial acylglycerols, Effect on solid fat profiles,
Eur J, Lipid Sci, Technology, 111, 259-272.
Vereecken, J., Meeussen, W., Foubert, I., Lesaffer, A., Wouters, J., and
Dewettinck, K., (2009c). Comparing the crystallization and polymorphic
behaviour of saturated and unsaturated monoglycerides, Food Research
International, 42, 1415-1425.
Vereecken, J., Meeussen, W., Lesaffer, A., Dewettinck, K., (2010). Effect of
water and monoglyceride concentration on the behaviour of monoglyceride
containing fat systems. Food Research International, 43, 872–881.
373
Verstraete, E. (2011). MSc thesis: Methods for monitoring fat crystallization
under shear for margarine applications. Faculty of Bioscience Engineering.
University Gent.
Vesper, H.W., Kuiper, H.C., Mirel, L.B., Johnson, C.L. and Pirkle, J.L. (2012).
Levels of plasma trans-fatty acids in non-Hispanic white adults in the United
States in 2000 and 2009. JAMA. 8, 307 (6), 562-3.
Wang, F., Liu, Y., Jin, Q., Huang, J., Meng, Z., Wang, X., (2011). Kinetic
analysis of isothermal crystallization in hydrogenated palm kernel stearin with
emulsifier mixtures, Food Research International, 44, (9), 3021-3025.
Wang, Y.C., McPherson, K., Marsh, T., Gortmaker, S.L. and Brown M.
(2011). Health and economic burden of the projected obesity trends in the USA
and the UK. The Lancet., Vol 378, (9793), 815-825.
Wang, R.., Xu, X. and Jiang, Y. (2012). Vegetable oils and fats in China:
current and future trends. inform. AOCS. 23 (5), 329-334.
Wassell, P. and Young, N.W.G (2007). Food applications of trans fatty acid
substitutes. International Journal of Food Science and Technology, 42, 503–
517.
Wassell, P., Bonwick, G., Smith, C. J., Almiron-Roig, E. and Young, N.W.G.
(2010a). Towards a Multidisciplinary Approach to Structuring in Reduced
Saturated Fat- Based Systems – A Review. International Journal of Food
Science & Technology, 45, 642–655.
Wassell, P., Wiklund, J., Stading, M., Bonwick, G., Smith, C. J., Almiron-
Roig, E. and Young, N. W. G. (2010b). Ultrasound Doppler based in-line
viscosity and solid fat profile measurement of fat blends, International Journal
of Food Science and Technology, 45, 877‐883.
Wassell, P., Farmer, M., Warner, S.A., Bech, A.T., Young, N.W.G., Bonwick,
G., Smith, C., DuPont Nutrition Biosciences APS, (2012a). Food or Feed
Including Moringa Oil. World Intellectual Property Organisation No.
WO2012168722.
374
Wassell, P., Farmer, M., Warner, S.A., Bech, A.T., Young, N.W.G., Bonwick,
G., Smith, C., DuPont Nutrition Biosciences APS, (2012b). Triglyceride Fat
Crystallisation. World Intellectual Property Organisation No. WO2012168727.
Wassell, P., Farmer, M., Warner, S.A., Bech, A.T., Young, N.W.G., Bonwick,
G., Smith, C., DuPont Nutrition Biosciences APS, (2012c). Low Fat Spread.
World Intellectual Property Organisation No. WO2012168726.
Wassell, P., Farmer, M., Warner, S.A., Bech, A.T., Young, N.W.G., Bonwick,
G., Smith, C., Forrest, B.A., DuPont Nutrition Biosciences APS, (2012d).
Dispersion of Triglycerides. World Intellectual Property Organisation No.
WO2012168723.
Wassell, P., Farmer, M., Warner, S.A., Bech, A.T., Young, N.W.G., Bonwick,
G., Smith, C., DuPont Nutrition Biosciences APS, (2012e). Spread. World
Intellectual Property Organisation No. WO2012168724.
Wassell, P., Okamura, A., Young, N.W.G., Bonwick, G., Smith, C., Sato, K.
and Ueno, S., (2012). Synchrotron Radiation Macrobeam and Microbeam X-
ray Diffraction Studies of Interfacial Crystallization of Fats in Water-in-Oil
Emulsions. Langmuir 28 (13), 5539-5547.
375
Wiklund, J. and Stading, M. (2008). Application of in-line ultrasound Doppler
based UVP-PD rheometry method to concentrated model and industrial
suspensions. Flow Measurement and Instrumentation, 19, 171–179.
Winwood. R. (2011). So you think you know the effects of dietary lipids on
human health? – Fat Chance! J. Food Science and Technology, (25), 2, 26-28.
Wood, A.B. (1964). A textbook of Sound, 3rd edn. London: Bell and Sons.
Yamagishi, K., Iso, H., Yatsuya, H., Tanabe, N., Date, C., Kikuchi, S.,
Yamamoto, A., Inaba, Y. and Tamakoshi, A., (2010). “Dietary intake of
saturated fatty acids and mortality from cardiovascular disease in Japanese: the
Japan Collaborative Cohort Study for Evaluation of Cancer Risk Study”
American Journal of Clinical Nutrition.
Yano, J., Sato, K., Kaneko, F., Small, D.M., and Kodali, D.R., (1999).
Structural analyses of polymorphic transitions of sn-1,3-distearoyl-2-
oleoylglycerol (SOS) and sn-1,3-dioleoyl-2-stearoylglycerol (OSO):
assessment on steric hindrance of unsaturated and saturated acyl chain
interactions. J. Lipid Res. 40: 140–151.
Young, N.W.G. and Wassell, P. (2008). Margarines and Spreads. In: G.L.
Hasenhuettl and R.W. Hartel (eds.), Food Emulsifiers and Their Applications,
2nd ed. Springer, New York.
376
Young, N.W.G., Wassell, P, Wiklund, J. and Stading, M. (2008). Monitoring
structurants of fat blends with ultrasound based in-line rheometry (UVP-PD).
International Journal of Food Science and Technology, 43, 2083–2089.
Yusoff, M.S.A., Kifli, H., Lida, H.M.D.N. & Rozie, M.P. (1998). The
formulation of trans fatty acid-free margarines. In: Proceedings of the World
Conference on Oilseed and Edible Oils Processing (ed. S.S. Koseoglu). 156–
158. Champaign, IL: AOCS Press.
377
Appendices
Dr. Renato Grimald, Fats & Oils group, tested the crystallisation of palm oil
and palm kernel oil behaviour with additive. Independently, his team tested a
sample of DIMODAN MB-90 (CRY110) and found an increase in palm oil
crystallisation speed. Their study found it was possible to increase fat kinetics
in biscuit cream filling fats and firmness. From DSC evaluation of Dimodan
MB-90 (CRY110) using the following:
Figure A1.1
PO and PKO have different crystallisation behaviour, when decreasing
temperature from:
In order to accommodate a range of fats with varying solid fat content (SFC),
the emulsifier blends were applied into three fat blends, otherwise known as
hard, medium and soft. The emulsifier blends shown in Table B1.1, were
designed to enable to compare performance of monoglyceride and triglyceride
with similar fatty acid profile, such as palm based products like: DIMODAN
HP (Monoglyceride) and GRINDSTED PS 101 (Triglyceride). All emulsifier
blends were applied at 1% dosage, and then rate of crystallisation and DSC
were determined.
Table B1.1 – Description of emulsifier blends composition
Emulsifier Ingredient 1 2 3 4 5 6 7 8 9 10 11 12
DIMODAN MB-90 100 50 50 50 50 50 50 50 70 30 34 34
DIMODAN HP 50 30 70 33 33
DIMODAN HR 50 33
GRINDSTED PS101 50
HLEAR (low erucic) 50
PS105 ALT1 50
PS105 ALT2 50
HHEAR (high erucic) 50 33
Remarks:
89% A
100%
A B C D 11% m/90%
C or D
HPPaSt
Item No. 126269 042301 035800 004556 European PS 105 126324 061191
ALT 1 ALT2
Fatty acid HHEAR HLEAR HPAST HP PS 105 PS 209 PS 101
(52%A 48%B) (52%A 48%D)
C 12 - - 0,2 0,5 - 0,2 0,0
C 14 0,2 - 1,2 2,5 0,1 1,3 0,3
C 15 4,1 - 0,1 - 2,1 2,1 3,8
C 16 - 6,0 58,0 43,0 9,5 2,9 20,6 4,9
C 17 0,1 - 0,1 - 0,1 0,1 0,1
C 18 39,0 91,0 40,0 51,0 58,0 64,0 44,8 39,1
C 18-1 - - 0,3 0,3 - 0,1 0,0
C 20 8,9 2,0 0,7 1,5 5,6 5,3 8,2
C 22 46,0 0,6 - - 24,0 24,2 23,9 42,1
C 24 1,1 - - - 0,6 0,6 1,0
Sum 99,4 99,6 100,6 98,8 91,5 99,5 99,1 99,5
Fat blends:
A. Palm Oil
B. A Brazilian commercial low trans table margarine fat blend.
C. Puff pastry margarine fat blend, composed of:
60% palm stearin
25% palm oil
15% rapeseed oil
70
60
Fat Blend A
50 (medium SFC)
40
SFC (%)
Remarks:
16
12
8 Determination1
SFC %
Determination2
4 Determination3
Average
0
0 5 10 15
-4
t (min)
Results
Results are shown using average values from Rate of Crystallisation triplicate
determinations (Appendix C).
MEDIUM Fat Blend A - (Palm Oil)
Figure B1.4 shows all variants tested in the medium (palm oil) fat blend. The
following
Pure Fat
16 MB-90
14 HP/MB-90
HR/MB-90
12
PS101/MB-90
10 HLEAR/MB-90
SFC(%)
8 PS105-1/MB-90
6 PS105-2/MB-90
HHEAR/MB-90
4
70MB-90/30HP
2 70HP/30MB-90
0 HP/HR/MB-90
0 3 6 9 12 15
-2 HP/HHEAR/MB-
90
Time (min)
16
14
12 Pure Fat
10 MB-90
SFC (%)
8 HP/MB-90
6 70MB-90/30HP
4 70HP/30MB-90
2 PS101/MB-90
0
-2 0 3 6 9 12 15
Time (min)
16
14
Pure Fat
12
MB-90
10
PS101/MB-90
SFC (%)
8
HLEAR/MB-90
6
PS105-1/MB-90
4 PS105-2/MB-90
2 HHEAR/MB-90
0
-2 0 3 6 9 12 15
Time (min)
16
14
12
Pure Fat
10
MB-90
SFC (%)
8
HP/MB-90
6
HHEAR/MB-90
4 HP/HHEAR/MB-90
2
0
-2 0 3 6 9 12 15
Time (min)
DSC
Data is provided in Appendix D
45
Pure Fat
40 MB-90
35 HP/MB-90
30 HR/MB-90
PS101/MB-90
25
T (ºC)
HLEAR/MB-90
20 PS105-1/MB-90
15 PS105-2/MB-90
10 HHEAR/MB-90
70MB-90/30HP
5
70HP/30MB-90
0
HP/HR/MB-90
Crystallisation onset Crystallisation onset
HP/HHEAR/MB-90
of initiation of main peak
Comments:
Crystallisation onset
of initiation
Pure Fat
12
MB-90
10 HP/MB-90
HR/MB-90
8
PS101/MB-90
SFC (%)
6 HLEAR/MB-90
4 PS105-1/MB-90
PS105-2/MB-90
2
HHEAR/MB-90
0 70MB-90/30HP
0 3 6 9 12 15 70HP/30MB-90
-2
HP/HR/MB-90
Time (min)
HP/HHEAR/MB-90
Figure B1.10
Pure Fat
6
MB-90
5 HP/MB-90
HR/MB-90
4
PS101/MB-90
SFC (%)
3 HLEAR/MB-90
2 PS105-1/MB-90
PS105-2/MB-90
1
HHEAR/MB-90
0 70MB-90/30HP
0 1 2 3 70HP/30MB-90
-1
HP/HR/MB-90
Time (min)
HP/HHEAR/MB-90
Figure B1.11
12
10
8 Pure Fat
MB-90
SFC (%)
6
HP/MB-90
4 HR/MB-90
2 HP/HR/MB-90
0
0 3 6 9 12 15
-2
Time (min)
Figure B1.12
Comments: From the beginning (first 3 mins), both blends with HP and HR
improved the onset temperature comparing with pure fat blend. But, after 6 min
MB-90 was more efficient.
12
10
Pure Fat
8
MB-90
SFC (%)
6 HP/MB-90
4 70MB-90/30HP
70HP/30MB-90
2 PS101/MB-90
0
0 3 6 9 12 15
-2
Time (min)
Figure B1.13
10
8 Pure Fat
MB-90
SFC (%)
6
HR/MB-90
4 HLEAR/MB-90
2 HHEAR/MB-90
0
0 3 6 9 12 15
-2
Time (min)
Figure B1.14
Comments:
Blends containing triglycerides (HLEAR and HHEAR) performance much
better than the blend HR/MB-90 (monoglycerides), the onset temperature was
clearly enhanced.
12
10
Pure Fat
8 MB-90
PS101/MB-90
SFC (%)
6
HLEAR/MB-90
4 PS105-1/MB-90
2 PS105-2/MB-90
HHEAR/MB-90
0
0 3 6 9 12 15
-2
Time (min)
Figure B1.15
Comments:
All blends containing triglycerides perform more or less the same and better
than MB-90. Except for PS101/MB-90 blend.
12
10
8 Pure Fat
MB-90
SFC (%)
6
HP/MB-90
4 HHEAR/MB-90
2 HP/HHEAR/MB-90
0
0 3 6 9 12 15
-2
Time (min)
Figure B1.16
45
Pure Fat
40 MB-90
35 HP/MB-90
30 HR/MB-90
PS101/MB-90
25
T (ºC)
HLEAR/MB-90
20 PS105-1/MB-90
15 PS105-2/MB-90
10 HHEAR/MB-90
5 70MB-90/30HP
70HP/30MB-90
0
HP/HR/MB-90
Crystallisation onset Crystallisation onset
HP/HHEAR/MB-90
of initiation of main peak
Figure B1.17
HARD Fat Blend C (60% palm stearin / 25% palm oil / 15%
rapeseed oil)
Rate of crystallisation (at 20°C, 15min plotted), 1% emulsifier blend dosage.
Pure Fat
40
MB-90
35
HP/MB-90
30 HR/MB-90
25 PS101/MB-90
SFC (%)
20 HLEAR/MB-90
15 PS105-1/MB-90
10 PS105-2/MB-90
5 HHEAR/MB-90
70MB-90/30HP
0
70HP/30MB-90
-5 0 3 6 9 12 15
HP/HR/MB-90
Time (min)
HP/HHEAR/MB-90
Figure B1.18
Comments: Pure MB-90 is the only emulsifier that clearly made positive
difference on the onset temperature. It means that only MB-90 improved the
rate of crystallisation in this fat blend.
Pure Fat
20
MB-90
HP/MB-90
15
HR/MB-90
PS101/MB-90
10
SFC (%)
HLEAR/MB-90
PS105-1/MB-90
5
PS105-2/MB-90
HHEAR/MB-90
0
0 1 2 70MB-90/30HP
70HP/30MB-90
-5
HP/HR/MB-90
Time (min)
HP/HHEAR/MB-90
Figure B1.19
40
35
30
Pure Fat
25
MB-90
SFC (%)
20
HP/MB-90
15
HR/MB-90
10 HP/HR/MB-90
5
0
-5 0 3 6 9 12 15
Time (min)
Figure B1.20
40
35
30 Pure Fat
25 MB-90
SFC (%)
20 HP/MB-90
15 70MB-90/30HP
10 70HP/30MB-90
PS101/MB-90
5
0
-5 0 3 6 9 12 15
Time (min)
Figure B1.21
40
35
30 Pure Fat
25 MB-90
SFC (%)
20 HR/MB-90
15 HLEAR/MB-90
HHEAR/MB-90
10
5
0
0 3 6 9 12 15
Time (min)
Figure B1.22
40
35
Pure Fat
30
MB-90
25
PS101/MB-90
SFC (%)
20
HLEAR/MB-90
15
PS105-1/MB-90
10 PS105-2/MB-90
5 HHEAR/MB-90
0
-5 0 3 6 9 12 15
Time (min)
Figure B1.23
40
35
30
Pure Fat
25
MB-90
SFC (%)
20
HP/MB-90
15
HHEAR/MB-90
10 HP/HHEAR/MB-90
5
0
-5 0 3 6 9 12 15
Time (min)
Figure B1.24
50
Pure Fat
45 MB-90
40 HP/MB-90
35 HR/MB-90
30 PS101/MB-90
T (ºC)
25 HLEAR/MB-90
20 PS105-1/MB-90
15 PS105-2/MB-90
HHEAR/MB-90
10
70MB-90/30HP
5
70HP/30MB-90
0
HP/HR/MB-90
Crystallisation onset Crystallisation onset
HP/HHEAR/MB-90
of initiation of main peak
1a 2a 3a 4a
DIMODAN MB-90 50 30 100 (applied at 0.5% into fat)
TS-XYZ 100 50 70
Remarks:
50
45
40 Pure Fat
35 MB-90
30 Pure Fat
T (ºC)
25 TS-945
20 TS-945/MB-90
15 70TS-945/30MB-90
0,5%MB-90
10
5
0
Crystallisation onset Crystallisation onset
of initiation of main peak
50
45
40 Pure Fat
35 MB-90
30 Pure Fat
T (ºC)
25 TS-945
20 TS-945/MB-90
15 70TS-945/30MB-90
0,5%MB-90
10
5
0
Crystallisation onset Crystallisation onset
of initiation of main peak
50
45
40 Pure Fat
35 MB-90
30 Pure Fat
T (ºC)
25 TS-945
20 TS-945/MB-90
70TS-945/30MB-90
15
0,5%MB-90
10
5
0
Crystallisation onset Crystallisation onset
of initiation of main peak
Generally it can be concluded from this study that the harder the fat blend, the
better the monoglyceride performance when compared to triglyceride. The
following is found:
Pure MB-90 increased the rate of crystallisation and the onset temperature. The
degree of saturation of the harder fat blend has an intrinsic influence; so that the
DIMODAN MB-90 performance compared better with other alternatives.
When a softer fat blend is used, more options (alternatives) can be found which
could change positively the fat crystallisation behaviour. An example, such as
50:50 combination of DIMODAN MB-90 + HLEAR (fully hardened low erucic
acid rapeseed oil) showed a good alternative in performance.
Note:
Emulsifier 1 (1%)
Emulsifier 2 (1%)
Emulsifier 3 (1%)
Emulsifier 4 (1%)
Emulsifier 5 (1%)
Emulsifier 6 (1%)
Emulsifier 7 (1%)
Emulsifier 9 (1%)
Emulsifier 10 (1%)
Emulsifier 11 (1%)
Emulsifier 12 (1%)
Emulsifier 2 (1%)
Emulsifier 4 (1%)
Emulsifier 6 (1%)
Emulsifier 7 (1%)
t (min) Determ.1 Determ.2 Determ.3 Average
0 0,220 0,411 0,336 0,322
1 2,726 2,715 2,082 2,508
2 4,482 4,530 4,819 4,610
3 5,514 5,355 5,133 5,334
4 6,551 6,254 6,332 6,379
5 7,656 7,411 7,452 7,506
6 8,756 8,845 8,787 8,796
7 9,584 9,348 9,651 9,528
8 10,335 9,937 9,954 10,075
9 10,316 10,065 10,171 10,184
10 10,859 10,405 10,471 10,578
11 10,739 10,354 10,488 10,527
12 10,909 10,508 10,910 10,776
13 10,866 10,647 10,876 10,796
14 10,970 10,607 10,789 10,789
15 11,079 10,736 10,803 10,873
Emulsifier 8 (1%)
Emulsifier 9 (1%)
Emulsifier 10 (1%)
Emulsifier 12 (1%)
Emulsifier 1 (1%)
Emulsifier 2 (1%)
Emulsifier 3 (1%)
Emulsifier 4 (1%)
Emulsifier 5 (1%)
Emulsifier 6 (1%)
Emulsifier 7 (1%)
Emulsifier 8 (1%)
Emulsifier 9 (1%)
Emulsifier 11 (1%)
The details of the UVP-PD equipment are the same as before, and have been
adequately outlined previously (Wassell et al., 2010b = Paper 4; Young, Wassell,
Wiklund, & Stading, 2008 = Paper 3), with the exception of the transducer
sensors.
The experimental Oils and Fats pilot plant flow loop at Danisco A/S consists of a
closed stainless steel piping circulation system with the possibility to switch
between different heat exchanger combinations etc. The sample fluids can be re-
circulated through the 12 mm piping from a stainless steel tank with an agitator
using a centrifugal pump. The project started with the adaptation of the UVP-PD
in-line measuring section to the Oils and Fats pilot plant at Danisco A/S, Braband,
Denmark. The measuring section, featuring a custom made flow adapter cell, was
fitted with a pair of ultrasound transducers and a pair of pressure sensors.
Figures E1-2 show photos of the experimental Oils and Fats pilot plant flow loop
at Danisco A/S together with the UVP-PD in-line measuring section.
Figure E1: Pilot plant at Danisco A/S installation of UVP-PD in-line measuring
section
The UVP-PD in-line measuring section was connected to the pilot plant piping
through an expansion and had an inner diameter of 22.5 mm. The measuring
section was attached to the exit of the pin roller and the pilot plant loop could
easily be modified to closely mimic real industrial conditions
Figure E2 shows close up of UVP-PD in-line measuring section and flow adapter
cell fitted with two ultrasound transducers
Images of previous test section are shown in Figure E3 (not previously shown)
from the schematic Figure 3.2 in 3.0. The current apparatus is shown in Figure E4
Figure E3 UVP section as described in Figure 3.2
The current flow cell and pressure sensor apparatus shown in Figure E4, was used
to investigate DAG concentration in palm products (Wassell, Wiklund, Farmer,
Hogan Bonwick, Smith, Young, Dynamic In-Line and Static Off-Line Rheological
Profiling on Palm Oil and the Effect of Diglyceride Content on Structure
Formation (for publication)
Figure E4 Photographic image of current UVP-PD flow cell apparatus. The flow
cell is in the centre with two fixed delay line transducers in place. Far right and
left are the pressure sensors which in turn are connected to the pressure gauge (top
right).
APPENDIX F
The tree is considered one of the world’s most useful trees, as almost every
part of the Moringa tree can be used for food, or has some other beneficial
property. The tree's bark, roots, fruit, flowers, leaves, seeds, and gum are also
used medicinally.
The Moringa seeds yield 38–40% edible oil, called ben oil from the high
concentration of behenic acid contained in the oil.
While it grows best in dry sandy soil, it tolerates poor soil, including
coastal areas. It is a fast-growing (3m in 10 months from seed) drought-
resistant tree.
References:
EUROPE
ENGLISH: Horseradish tree, Radish tree, Drumstick tree, Mother's Best
Friend, West Indian ben.
FRENCH: Bèn ailé, Benzolive, Ben oléifère, Arbre radis du cheval.
GERMAN: Behenbaum, Behenussbaum, Flügelsaniger bennussbaum,
Pferderettichbaum
ITALIAN: Sàndalo ceruleo
PORTUGUESE: Acácia branca, Marungo, Muringa, Moringuiero; Cedro
(Brazil)
SPANISH: Árbol del ben, Ben, Morango, Moringa
AFRICA
BENIN:
(Adia): Kpashima
(Bariba): Yuru ara, Yorwata, Yoroguma
(Dendi): Windibundu
(Fon): Patima, Kpatima, Yovokpatin, Kpano, Yovotin
(Gun): Ékwè kpatin, Kpajima
(Mina): Yovo vigbe, Yovo kpati
(Natemba): Tekpinda
(Peul): Guildandeni, Latj iri, Legi-Lakili
(Saxwe): Kotba
(Waama): Yori kununfa
(Yoruba): Ewé ilé
(Yoruba-Nago): Ewé igablé, Ewé ilé, Ewé oyibo (White man’s tree), Agun,
Manyieninu, Ayere, Oyibo.
BURKINA FASO :
(Bwaba): La-Banyu.
(Dioula): Ardjeneyiri, Ardjian jirri.
(Gourmanche/Gourmantchema): Alj an-tiiga, Ki gambaga (region of Tapua),
Diegu kanlobuga (region of Gnagna) (meaning “one will never lack a child in
the courtyard”), Makkakomboanga.
(Fulfuldé): Aljannahi, Guilgandani, Gigandjah.
(Morée/Mossi): Argentiga, Arzan tiiga ("The tree of paradise").
CAMEROUN:
(Daggai): Paizlava
(Foulfouldé): Guiligandja, Giligandjahi
(Hausa): Zogalagandi
(Mafa): Gagawandalahai
(Mandara): Djhiré
(Moundang): Naa-toukoré
(Podoko): Chabané
(Toupouri): Naa-nko
CHAD:
(Sara): Kag n’dongue
(Shuwa Arabic): Alim, Halim
COMOROS ARCHIPELAGO:
Anambo, Mvungué.
GHANA:
(Ewe): Atiuwuse (the tree with tender/slim leaves), Babati, Babatsi, Kpotowuzie
(feeble tree; easily broken). Kpokpoti (the “illness tree”), Nukunaya (wonderful news),
Yevu-ti (white man’s tree), Yevutsi
(Dagari): Obnukuo, Ornyyukuo, Zangala.
KENYA:
(Swahili): Mlonge, Mronge, Mrongo, Mlongo, Mzunze, Mzungu., Mjungu
moto, Mboga chungu, Shingo.
(Sokoki - Indian spoken in Mombasa): Mborongi.
MADAGASCAR:
(Malagasy): Anamambo, Anamorongo, Feliimorongo, Felikambo,
Felikamoranga, Landihazo, Moringa, Moringy.
MALAWI:
(Chichewa): Cham'mwanba, Kangaluni.
(Lomwe): Sangoa, Shangoa
(Senna): Nsangoa.
(Yao): Kalokola.
ALSO: Maula tengo, Mpundi, Muula, Mbula, Mpempu, Chakate, Mpenba
MALI:
(Bambara): Gnougou Jirini, Kandjirini, Manjirini, Massa Jirini.
(Segou): Verdaye
MAURITIUS:
(Creole): Drède mouroungue.
(Indian Creole): Mouroungue
NIGER:
(Hausa): Zôgala gandi.
(Shuwa Arabic): Alim, Halim.
(Zarma): Windi-bundu.
NIGERIA:
(Fulani): Gawara, Gaware, Konamarade, Rini maka, Habiwal hausa.
(Hausa): Bagaruwar maka, Bagaruwar masar, Barambo, Koraukin zaila, Rimin
nacara, Rimin turawa, Samarin danga, Shipka hali, Shuka halinka, Zogall,
Zogalla-gandi,
(Ibo): Odudu oyibo,Okwe oyibo,Okwe olu,Uhe,Oku-ghara-ite, Okochi egbu
(“cannot be killed by the dry season”).
(Nupe): Chigban Wawa
(Yoruba): Adagba malero, Ewele, Ewé ilé, Ewe igbálé, Idagbo monoyé ("The
tree which grows crazily").
SENEGAL:
(Wolof): Neverday, Nébéday, Sap-Sap.
(Serer): Nébéday.
SEYCHELLES :
(Creole): Drède mouroungue.
SIERRA LEONE:
Boganja
SOMALIA:
(Indian): Mrongo
SUDAN:
(Arabic): Ruwag, Alim, Halim, Shagara al ruwag.
(Dinka): Anid.
(Kordofan Arabic): Shagara zaki al moya.
TANZANIA:
(Swahili): Mlonge, Mronge, Mrongo, Mlongo, Mzunze, Mzungu., Mjungu
moto, Mboga chungu, Shingo.
TOGO:
(Dagomba): Baganlua, Bagaelean.
(Ewe): Kpotima, Kpoti, Yevu-ti, Yovoviti.
(Hausa): Mágurua maser.
(Mina): Yovoviti.
(Moba): Gambaduk.
(Mouroungue): Jevoti, Jovoviti.
ALSO: Amedoti, Ekpoti, Molo-Kpoti.
ZAMBIA:
(Tonga): Zagalanda, Zakalanda
ZIMBABWE:
(Tonga): Mupulanga, Zakalanda.
ASIA
BURMA:
(Burmese): Dandalun, Daintha, Dandalun-bin, Dandalonbin.
CAMBODIA:
Ben ailé, Daem mrum.
INDIA :
(Bengalese): Munga ara, Sajna, Sojna, Sujana
(Gujarati): Midho-saragavo, Saragavo, Saragvo, Suragavo.
(Hindi): Munga ara, Shajmah, Shajna, Segra.
(Hindi/Orissa): Sanjna, Saijna, Shajna, Soandal
(Kanarese): Nugga egipa, Nugge, Noogay, Nuggi Mara.
(Kol): Mulgia, Munga ara, Mungna
(Kumao – Himalayan region): Sunara
(Konkani/Goa): Moosing, Mosing
(Malayalam): Sigru, Moringa, Muringa, Murinna, Morunna.
(Marathi): Sujna, Shevga, Shivga.
(Modesia/W. Bengal): Mangnai
(Monghye/Punjab): Sejana
(Oriya): Munigha, Sajina.
(Punjabese): Sanjina, Soanjana.
(Rajasthan): Lal Sahinjano
(Sanskrit): Danshamula, Shobhanjana, Sigru Shobhanjan, Sobhan jana.
(Sindhi): Swanjera
(Tamil): Morunga, Murungai, Murunkak-kai.
(Telegu): Sajana, Tella-Munaga.
(Teling): Morunga, Morungai
(Urdu): Sahajna
(Central provinces): Mulaka, Saihan
(Western region): Sundan
ALSO: Sweta Maricha
INDONESIA:
(Alor): Maroenga, Motong.
(Bali): Kelor, Tjelor.
(Flores): Moltong
(Java): Kelor
(Madura): Marongghi
(Moluccan islands): Oho Gaairi
(Roti): Kafok, Kai fok
(Sumatra): Kalor, Kerore
(Sumba): Kawona, Wona
(Ternate): Kelo, Oege Kelo
(Tidore): Kelo
(Timor): Baoe fo, Maroenga
ALSO: Remoenggai, Sajor Kelor
LAOS:
(Lao): B'Loum
MALAYSIA:
Kachang Kelur, Lemunggai, Meringgai, Semunggai, Smunggai, Semunggai,
Remunggai
NEPAL:
Sitachini
PAKISTAN:
Saijan, Sohanjna
PHILIPPINES:
(Tagálog): Kalungai, Kamalungua, Malongai, Malungai, Mulanggay,
Malunkai.
(Bikol): Kalungai
(Bisáya): Alúngai, Dool, Malungit.
(Kisaya): Kalungai
(Ibanág): Marongai, Marungai
(Ilóko): Marongai, Marungai, Komkompilan
(Pampángan): Dool, Kamalungua, Malúngit.
(Panay Bisáya): Kalamúngai, Kamalongan
(Pangasinán): Rúnggai
(Sambáli): Marongai, Marungai
(Simeulue): Aroenggai
THAILAND:
(Thai): Kaanaeng-doeng, Phak eehuem, Phak eehum, Phak-nuea-kai, Se-cho-
ya
(Central highlands): Ma rum
(North): Ma khonkom
VIETNAM:
Chum Ngay
YEMEN:
(Arabic?): Saisam
COLOMBIA:
(Spanish): Aceite, Aceitoso, Angela, Colirio, Goma, Jeringa, Marango,
Maranjo, Marangon, Sen de la tierra.
COSTA RICA:
(Spanish): Marango, Marangon.
CUBA:
(Spanish): Acacia, Ben, Calicita, Leno nefrítico, Palo blanco, Palo de jeringa,
Palo de Tambor, Paraíso francés.
DOMINICAN REPUBLIC:
(Spanish): Ben, La libertad, Libertad, Palo de abejas, Palo de aceiti.
DUTCH ANTILLES:
(Dutch): Ben boom
(Spanish): Brenolli, Morenga, Orselli.
EL SALVADOR:
(Spanish): Ceiba, Marengua, Narango, Paraíso extranjero, Teberinto.
GUADALOUPE:
(Spanish): Maloko, Moloko, Ben-ailé.
GUATAMALA:
(Spanish): Perla, Perlas, Paraíso blanco.
GUYANA:
Saijhan.
HAITI:
(French): Benzolive, Benzolivier, Ben oleifere, Bambou-bananier, Graines
benne, Olivier.
(English): Benzolive tree.
HONDURAS:
Maranga calalu.
MEXICO:
(Spanish): Arbol de las perlas, Chinto borrgo, Flor de Jacinto, Jacinto, Paraíso
blanco, Paraíso de Espana, Perla, Perlas, Perla de la India, Perlas del oriente,
San Jacinto.
NICARAGUA:
(Spanish): Marango, Maranjo, Marangon.
PANAMA:
(Spanish): Jacinto.
PUERTO RICO:
(Spanish): Angela, Ben, Colirio, Jasmin francés, Resada, Sen de la tierra.
SAN SALVADOR:
(Spanish): Marango, Marangon, Maranjo.
SURINAM:
(Dutch): Peperwortel boom.
(Indonesian): Kelor.
TRINIDAD:
(Hindi?): Saijan.
VENEZUELA:
(Spanish): Aceite de Ben, Azucarillo, Ben, Sen.
ALSO: Arbol do los aspáragos, Bamboubamamoer, Cedro, Cenauro, Chinto borrego, Chuva
de prata, Desengaño, Gailito, Guaireña, Hoja de sen, Macasar, Marenque, Moongay,
Moriengo, Noz de bem, Orenga, Palo de geringa, Palo jeringa, Paraíso, Pois quinique, Quiabo
de tres quinas, Sainto John, Salaster, Salibau, Sen, Seringa.
MORINGA DROUHARDII
MADAGASCAR
(Malagasy): Hazomalana, Maroserano, Moringy.
MORINGA LONGITUBA
ETHIOPIA
(Somali): Hane, Mane.
KENYA
(Somali): Borrant, Haduma.
(Boran): Saffara.
SOMALIA
(Majindi?): Fintir.
(Somali): Hawe, Mawe, Mawow, Wame, Wuame.
MORINGA OVALIFOLIA
ANGOLA Hungua, Mungua.
NAMIBIA
(Afrikansk): Meelsakboom.
(Herero): Omutindi.
MORINGA PEREGRINA
EGYPT, SYRIA, ARABIAN PENINSULA
(Arabic): Bãn, El bãn, Vassar, Ithi, Yassar
IRAN
Gazrokh, Gazroghan.
ORAN
(Arabic): Shuh
SOMALIA
(Somali): Dankap, Dongop, Dumok, Mereh, Moroh.
SUDAN
(Arabic): Shagara al rauwãq.
(Bisharin, Hadendowa): Mai
(Tigri): Khal erbal.
ETHIOPIA
(Tigri): Khal erbal.
MORINGA STENOPETALA
ETHIOPIA
(Konso): Shifara, Shalchada.
KENYA
(Turkana): Etebusoit.
(Njemps): Loresienjo.
(Samburu): Larsanjo, Lorsenjo, Lossantscho.
(Somali): Mau, Mawa.
(Boran): Saffara. (Konso): Shifara, Shalchada, Shalqueida.
APPENDIX G
As stated:-
Lesquerella has several novel properties that set it apart from castor and other
oilseeds. One property, is its oil functionality, including difunctional hydroxy
moities (in contrast to other seed oils such as castor that are trifunctional).
Lesquerella has been reported to contain natural estolides (secondary esters
derived from the addition of a fatty acid moiety to the hydroxyl functionality of
the hydroxy triglyceride) that improve performance of the oil. Estolides have
been encountered in only a few other seed oils, and are not naturally found in
castor oil (although they can be synthetically fabricated). Estolides have been
shown to improve vegetable oil performance in motor oils by improving pour
points (temperature at which the oil no longer pours), and estolides have also
been used as viscosity modifiers in lubricating oils. The second of lesquerella's
novel properties is its antioxidants, converted from the seed meal fraction that
contain unique glucosinolates and have superior oxidative stability properties.
The seed meal also has a binder application, unique for seed meals. The third
novel property is the seed coat of lesquerella, which contains a unique gum that
has rheological properties equivalent to guar or xanthan that can be useful in
coatings and food thickeners” (Dierig et al., 1993; Dierig and Thompson,
1993).
0.600
0.400
0.200
Interfacial moduli [Pam]
0.000
0 10 20 30 40 50 60
-0.200
G'i Pa
m
-0.600
-0.800
Temperature [°C]
APPENDIX I
Example data from RHEOPLUS/32 V3.21 shows upper fluid parameter: bulk
sunflower oil at 50°C and lower fluid parameter: water at 50°C
APPENDIX J