A thesis submitted in fulfillment of the requirements for the degree of Doctor of Philosophy: Use of Z-stack imaging to quantify the phase behaviour of biomaterial composites in relation to theoretical predictions of blending laws from rheological measurements

Use of Z-stack imaging to quantify the phase behaviour of biomaterial composites in relation to theoretical predictions of blending laws from rheological measurements

A thesis submitted in fulfillment of the requirements for the degree of Doctor of Philosophy

Pranita Mhaske

B. Tech. (Food Science and Technology), MIT College of Food Technology, Pune, India

School of Science

Science, Technology, Engineering and Mathematics College

RMIT University

March 2021

Declaration

I certify that except where due acknowledgement has been made, the work is that of the

author alone, and the work has not been submitted previously, in whole or in part, to qualify

for any other academic award. The content of the thesis is the result of work which has been

carried out since the official commencement date of the approved research program at RMIT

University; and any editorial work, paid or unpaid, carried out by a third party has been

acknowledged; and ethics procedures and guidelines have been followed.

Pranita Mhaske

23rd March 2021

Copyright

Pranita Mhaske

2020

Notice 1

The 1968 Copyright Act, mandates that this thesis be used only under regular conditions of

scholarly fair dealing. In particular, no results or conclusion should be extracted from it, nor

should it be copied or closely paraphrased in whole or in part without the written consent of

the author. Proper written acknowledgement should be made for any assistance obtained from

this thesis.

Notice 2

I certify that I have made all reasonable efforts to secure copyright permissions for third-party

content included in this thesis and have not knowingly added copyright content to my work

without the owner’s permission.

ACKNOWLEDGEMENTS

I owe my sincere gratitude to my senior supervisor, Prof. Stefan Kasapis for his support

and supervision throughout my PhD candidature. His guidance, motivation and scientific

insight has enabled me to develop a thorough understanding of the subject and meet his high

publication standards.

I would also like to express my gratitude to my co-supervisor, A/Prof. Asgar Farahnaky,

for his patient and kind guidance, and critical outlook that helped me be on track during my

candidature and for providing scientific exposure and opportunities that were beyond the scope

of the candidature. Working with him has impressed and inspired me on both the professional

and personal front.

A word of gratitude also goes to Lita Katopo, for patiently guiding, supporting, training

and motivating me during the early days of my candidature, without which I probably wouldn’t

have been able to see this through.

My research would not have reached its successful completion without the support of

the technical staff at RMIT University. I am thankful to Sanaz Salehi, Yan Chen, Spiros

Tsaroumis, Hadi Ranjiburachaloo, Jake Torrisi, Badyn Cook, James Lauer, Chaitali Dekiwadia

and especially Mina Dokouhaki for providing more than just technical assistance throughout

the years.

A big shoutout to my colleagues, associates and friends at the Food Science Laboratory

– Manisha Singh, Billy Lo, Lloyd Condict, Shahla Teimouri, Felicity Whitehead, Courtney

Morrish, Diah Ikasari, Kourosh Abdollahi, Cameron Ince, Carine Semasaka, Jasmeet Singh,

Geethu Kurup, Pham Loc, Arissara Phosanam, Lili Mao, Mayumi Silva, Nelum Pematilleke

and Mithila Jayasundera for their support and company which made the time spent working in

the labs fun and memorable.

Last but certainly not the least, I am grateful to my family for their continual love and

iii

support, to which I owe everything.

Page i ii iii iv ix xiv xv xvii xx 1 Summary

Title Declaration Copyright Acknowledgements Table of contents List of Figures List of Tables List of abbreviations List of units and symbols Explanatory notes Chapter 1 Background and Literature Review

1.1. Introduction 1.2. Phase behaviour in mixed biopolymer systems 1.3. Phase diagrams of protein-polysaccharide-water systems 1.4. Techniques used to probe phase behaviour

1.4.1. Visual observation and phase volume ratio method 1.4.2. Refractometry 1.4.3. Chemical analysis for concentration estimation 1.4.4. Viscometry 1.4.5. Spectroscopy 1.4.5.1. UV-Vis spectroscopy

1.4.5.2. FTIR spectroscopy

1.4.5.3. Raman spectroscopy 1.4.5.4. NMR spectroscopy 1.4.5.5. Ultrasonic spectroscopy 1.4.5.6. Diffusing wave spectroscopy 1.4.5.7. Electron spin resonance spectroscopy 1.4.6. Diffraction and scattering methods 1.4.6.1. Light scattering 1.4.6.2. X-Ray diffraction 1.4.6.3. SAXS and SANS 1.4.7. Calorimetry 1.4.8. Rheology 1.4.8.1. Mechanical properties of the composite 1.4.8.2. Interfacial tension between phases 1.4.9. Microscopy 1.4.9.1. Optical microscopy 1.4.9.2. Phase contrast microscopy 5 8 10 13 13 18 20 23 26 26 27 30 31 32 34 34 35 36 37 38 40 43 43 47 50 51 52

1.4.9.3. Atomic force microscopy 1.4.9.4. Fluorescent microscopy 1.4.9.5. Electron microscopy

1.5. Significance of the Research 1.6. Research questions 1.7. Research objectives

References

Chapter 2 Materials and Methods

Abstract 2.1. Materials

laser scanning 56 57 58 59 61 61 62 83 84 84 84 85 85 86 86 87 88 88 88 89 89 89 2.1.1. Biomaterials 2.1.1.1. Agarose 2.1.1.2. Gelatin 2.1.1.3. Lipids 2.1.1.4. Microcrystalline cellulose 2.1.2. Chemicals and Dyes 2.1.3. Instruments 2.1.4. Sample preparation

2.1.4.1. Agarose-canola oil composites 2.1.4.2. Agarose-ghee composites 2.1.4.3. Agarose-MCC composites 2.1.4.4. Agarose-gelatin composites 2.1.4.5. Staining for confocal microscopy

2.2. Methods

2.2.2.1. Micro Differential Scanning Calorimetry 2.2.2.2. Modulated Differential Scanning Calorimetry

2.2.1. Fourier Transform Infrared (FTIR) Spectroscopy 2.2.2. Differential Scanning Calorimetry 2.2.3. Rheology 2.2.4. Scanning Electron Microscopy 2.2.5. Confocal Laser Scanning Microscopy 2.2.6. Image Analysis References

90 90 94 97 98 99 107 110 115 116

in comparison to blending

123 123 Chapter 3 Quantitative analysis of the phase volume of agarose- canola oil gels law predictions using 3D imaging based on confocal laser scanning microscopy Abstract 31. Introduction

3.2. Materials and methods

3.2.1. Materials 3.2.2. Methods 3.2.2.1. Sample preparation 3.2.2.2. SEM analysis 3.2.2.3. Fourier transform infrared (FTIR) spectroscopy 3.2.2.4. Differential scanning calorimetry 3.2.2.5. Rheological measurements 3.2.2.6. Confocal laser scanning microscopy (CLSM) 3.2.2.6.1. Sample preparation and image acquisition 3.2.2.6.2. Image analysis 3.2.2.7. Statistical analysis

3.3. Results and discussion

the structural 124 124 124 124 124 124 124 124 124 124 125 125 125 125 3.3.1. Experimental observations on

characteristics of agarose-canola oil mixtures

127 3.3.2. Theoretical modelling of the phase behaviour in

agarose-canola oil composites

128 3.3.3. Development of a CLSM protocol and image

analysis for agarose-canola oil composites

3.4. Conclusions References

131 131

Chapter 4 Phase volume quantification of agarose-ghee gels using 3D confocal laser scanning microscopy and blending law analysis: A comparison Abstract 4.1. Introduction 4.2. Materials and methods

4.2.3.1. Scanning electron microscopy

4.2.3.2. Fourier transform infrared spectroscopy (FTIR)

4.2.3.3. Differential scanning calorimetry 4.2.3.4. Rheological measurements 4.2.3.5. Confocal laser scanning microscopy (CLSM) 4.2.3.6. Image analysis 4.2.3.7. Statistical analysis 4.2.1. Materials 4.2.2. Methods 4.2.3. Experimental analysis

4.3. Results and discussion

134 134 135 135 135 135 135 135 135 135 135 135 136 136 136 4.3.1. Investigation of the composite microstructure and its

structural characteristics 4.3.1.1. Composite microstructure 4.3.1.2. Chemical properties 136 136

structure 136 137

138

139

4.3.1.3. Thermal characterization 4.3.1.4. Rheological observations of formation 4.3.2. Theoretical modelling of mechanical functions in support of phase topology 4.3.3. Z-stack CLSM imaging and image analysis of agarose-ghee blends

4.4. Conclusions References

141 141 Chapter 5 3D Confocal Laser Scanning Microscopy

for Quantification of the Phase Behaviour in Agarose- MCC co-gels in Comparison to the Rheological Blending-law Analysis Abstract 5.1. Introduction 5.2. Materials and methods

5.2.2.4. Rheological measurements

144 144 145 145 145 145 145 145 145 146 5.2.1. Materials and Sample Preparation 5.2.2. Experimental analysis 5.2.2.1. SEM Analysis 5.2.2.2. Fourier transform infrared spectroscopy (FTIR) 5.2.2.3. Differential scanning calorimetry

5.2.2.5. Confocal laser scanning microscopy (CLSM) and Image Analysis

5.2.2.6. Statistical analysis

5.3. Results and discussion

146 146 146 5.3.1. Structural Characterisation of the Agarose/MCC

Mixtures

148 5.3.2. Rheological Modelling of the Phase Behaviour in

Agarose/MCC Composite Gels

149 5.3.3. 3D CLSM imaging and image analysis for Phase

Volume Estimations in Agarose/MCC Mixture

5.4. Conclusions References

151 151

scanning microscopy confocal laser

vii

Chapter 6 Comparison of rheological blending-law analysis and for 3D quantification of the phase behaviour in agarose- gelatin co-gels Abstract 153

6.1. Introduction 6.2. Materials and methods

6.2.2.1. Sample preparation 6.2.2.2. Fourier transform infrared spectroscopy (FTIR) 6.2.2.3. Differential scanning calorimetry 6.2.2.4. Rheological measurements

6.2.1. Materials 6.2.2. Methods 153 155 155 155 155 156 156 156 157

6.2.2.5. Confocal laser scanning microscopy (CLSM) and Image Analysis

6.2.2.6.Statistical analysis

6.3. Results and discussion

157 158 158 6.3.1. Structural Characterisation of agarose-gelatin

164

171

mixtures 6.3.2. Theoretical modelling of the Phase Behaviour in Agarose-gelatin Co-Gels using blending law analysis 6.3.3. 3D CLSM imaging and image analysis of agarose- gelatin co-gels

6.4. Conclusions References

175 176 Chapter 7 Conclusions and Future Research

181 7.1. Conceptual basis of the present study

viii

182 185 190 193 7.2. Conclusions 7.3. Future Research References Appendix

List of Figures

Chapter 1

8 Figure 1.1 Scale of measurement of the various techniques used to probe phase

behaviour

11 16 Figure 1.2 Phase diagram in an aqueous biopolymer mixture Figure 1.3 Phase diagram of 3% gelatin/0,9% iota-carrageenan as a function of

sodium chloride concentration and pH

Figure 1.4 DSC exotherms for a) 2% agarose, b) 15% gelatin, c) 2% agarose plus 15% gelatin, d) Dalda vanaspati lipid, e) soyabean oil, and f0 2% agarose plus 15% gelatin plus 2.5% Dalda vanaspasti lipid. Scan rate 1°C/min

Figure 1.5 A drop of caseinate in alginate rich phase at rest (1), start-up transient of deformation under steady shear (2,3), steady state shape (4), and relaxation after shear is stopped (5,6). Graph represents the deformation parameter as a function of shear rate in comparison with theory

Figure 1.6 Kappa carrageenan (A, B) and Iota carrageenan (C, D) in mixture with meat proteins at pH 5.6 and 7.1, respectively. Black arrows indicate carrageenan molecules with attached meat proteins.

Chapter 2

84 85 91 93 94 95 Figure 2.1 Chemical structure of an agarose polymer Figure 2.2 Chemical structure of gelatin Figure 2.3 Schematic Diagram of FTIR Figure 2.4 FTIR- ATR Spectrum 100 at RMIT University Figure 2.5 Schematic diagram of a heat flux type DSC Figure 2.6 Typical material transition as a function of temperature change in a

DSC

97 98 99 102

Figure 2.7 The heat flux micro DSC with a cylinder type measuring system Figure 2.8 Micro DSC Setaram VII at RMIT University Figure 2.9 Modulated DSC Q2000 from TA instruments at RMIT University Figure 2.10 Basic shear rate versus shear stress behaviour of Newtonian, shear thinning, shear-thickening, and Bingham and Herschel-Bulkley materials

104

Figure 2.11 The principle of oscillation viscometry. Applied strain versus time (a) and resultant stress versus time that is measured in an elastic solid (b), Newtonian liquid (c) and viscoelastic liquid (d).

106 108 110 Figure 2.12 Discovery Hybrid Rheometer by TA Instruments at RMIT University Figure 2.13 Schematic diagram of a SEM Figure 2.14 FEI Quanta 200 ESEM from FEI Corporate at RMIT University

112 Figure 2.15

Illustration of the confocal imaging principle (solid lines = in-focus light; dashed lines = out-of-focus light)

Chapter 3

126

Figure 3.1 SEM images of (a) 1% (w/w) agarose at 50x magnification, (b) 1% (w/w) agarose at 400x magnification, (c) 1% (w/w) agarose with 3% (w/w) canola oil at 400x magnification, and (d) 1% (w/w) agarose with 11% (w/w) canola oil at 400x magnification.

126

Figure 3.2 FTIR spectra of 1% (w/w) agarose with 0, 1, 3, 5, 7, 9 and 11% (w/w) canola oil arranged from top to bottom (solid lines), and canola oil (dashed line)

127

Figure 3.3 Cooling profiles of 1% (w/w) agarose with 0, 1, 3, 5, 7, 9 and 11% (w/w) canola oil arranged from top to bottom (solid lines), and canola oil (dotted line) obtained using DSC at a scan rate of 1°C/min.

127

Figure 3.4 Cooling profiles of G' for 1% (w/w) agarose with 0 (○), 1% (■), 3% (x), 5% (●), 7% (Δ), 9% (-) and 11% (♦) (w/w) canola oil from 50°C to 5°C at a scan rate of 2°C/min. The inset corresponds to calibration curve of G' at 5°C as a function of agarose concentration.

128

129 Figure 3.6

Figure 3.5 Modeling the phase topology of 1% (w/w) agarose plus 3% (w/w) canola oil using the isostrain blending law. Predicted storage modulus of agarose (G'ag) and composite (G'C(BL)) are illustrated by dashed and solid lines, respectively, and the experimental composite modulus (G'exp) is shown by a dotted line. Sag is the solvent content of the agarose phase. The inset shows experimental G' values (■) and calculated phase volume of canola oil (▲) for 1% (w/w) agarose with increasing oil concentration of (solid line fit on the inset and the oil phase volume are derived from the isostrain blending law). (a) 2D, (b) Z-stack, and (c) 3D CLSM images of 1% (w/w) agarose plus 5% (w/w) canola oil. The oil droplets are either in purple (a-b) or grey (c), and the agarose is in black (a-c).

130

Figure 3.7 Phase volume of canola oil in mixture with 1% (w/w) agarose determined via a Z-stack method coupled with image analysis software (FIJI and Imaris).

Chapter 4

137

Figure 4.1 SEM micrographs at 50x magnification of (a) 1% (w/w) agarose, and at 400x magnification of 1% (w/w) agarose containing (b) 0% ghee, (c) 3% (w/w) ghee and (d) 15% (w/w) ghee.

137

Figure 4.2 FTIR spectra of ghee (dotted line), and 1% (w/w) agarose with 0, 1, 3, 6, 9, 12 and 15% (w/w) ghee arranged from top to bottom (solid lines).

138

Figure 4.3 Exothermic peaks from the cooling profiles of (a) ghee (solid line) and agarose (dotted line) by themselves and (b) 1% (w/w) agarose with 1, 3, 6, 9, 12 and 15% (w/w) ghee arranged from bottom to top.

139

Figure 4.4 Cooling profile of storage modulus for 1% (w/w) agarose with 0 (□), 1% (җ), 3% (-), 6% (x), 9% (●), 12% (Δ) and 15% (♦) (w/w) ghee and ghee by itself (■) during cooling from 50 to 5°C at a scan rate of 2°C/min. The inset denotes the agarose calibration curve of G' at 5°C as a function of its concentration

140

140 Figure 4.6

Figure 4.5 Computerized modeling for 1% (w/w) agarose plus 6% (w/w) ghee, using the Lewis and Nielsen blending law, is illustrated depicting the predicted storage modulus of agarose (G'ag; dashed line), composite (G'C(LN); solid line) and the experimental composite modulus (G'exp; dotted line) as a function of the solvent content of the agarose phase (Sag). The experimental Gʹ values (●) and phase volume estimates of ghee (♦) for 1% (w/w) agarose with increasing ghee concentrations are depicted in the inset (blending law is used to derive the solid line fit of composite modulus and the ghee phase volume in the inset). (a) 2D, (b) Z-stack, and (c) 3D CLSM images of 1% (w/w) agarose plus 6% (w/w) ghee. The green droplets depict the ghee whereas the black background corresponds to agarose.

141 Figure 4.7 Confocal images of 1% (w/w) agarose and 20% (w/w) ghee captured

from (a) the surface and (b) a depth of 80 µm from the surface.

Chapter 5

147

Figure 5.1 SEM images of (a) 1% (w/w) agarose network at 50x and, walls of 1% (w/w) agarose network containing (b) 0, (c) 0.6 and (d) 1.2 % (w/w) MCC under 600x.

148

Figure 5.2 FTIR spectra between 400-4000 cm-1 for agarose with 0, 0.1, 0.3, 0.6, 0.9, 1.2 and 1.5% (w/w) MCC and MCC by itself stacked upwards from the the bottom.

148

Figure 5.3 DSC thermograms of (a) cooling and heating profiles of hydrated MCC paste and (b) cooling profiles of agarose with 0, 0.1, 0.3, 0.6, 0.9, 1.2 and 1.5% (w/w) MCC arranged successively upwards

149

Figure 5.4 Profiles of Gʹ for 1% (w/w) agarose with 0 (●), 0.1% (x), 0.3% (▲), 0.6% (□), 0.9% (♦), 1.2% (-) and 1.5% (○) (w/w) MCC during cooling from 50 to 5°C with a controlled strain of 0.1% and a scan rate of 2°C/min. The agarose calibration curve of Gʹ at 5°C as a function of its concentration is illustrated in the inset.

149

Figure 5.5 Values of composite Gʹ obtained via rheological experiments (x) and theoretical blending laws predictions (solid line) along with the estimated phase volumes of MCC (♦) for samples containing 1% (w/w) agarose with increasing MCC concentrations.

150 Figure 5.6

(a) 2D, (b) Z-stack, and (c) 3D CLSM images of 1% (w/w) agarose containing 0.6% (w/w) MCC. The blue background denotes the stained agarose phase whereas the black cavities correspond to the unstained MCC particles.

150

Figure 5.7 Phase volume estimates of MCC in mixture with 1% (w/w) agarose determined by analysing 3D CLSM images using image analysis software – FIJI and Imaris.

Chapter 6

159

Figure 6.1 FTIR spectra for (a) single systems of 3% agarose and 3% gelatin and (b) 3% gelatin with 0.1, 0.4, 0.8, 1.2, 1.6, 2, 2.5 and 3% agarose arranged upwards.

160

Figure 6.2 Cooling profiles of single systems of 3% gelatin (dotted line), 3% agarose (dashed line), and 3% gelatin with 0.1, 0.4 0.8, 1.2, 1.6, 2, 2.5 and 3% agarose (solid lines) arranged from bottom to top, obtained at a scan rate of 1°C/min using DSC.

162

164

168 Figure 6.5

172 Figure 6.6

Figure 6.3 Variation in storage modulus (a) as a function of temperature and time of observation for single systems of 3% gelatin and 3% agarose, and (b) on heating 3% gelatin plus 0.1, 0.4, 0.8, 1.2, 1.6, 2, 2.5 and 3% agarose, arranged upwards. Scan rate: 2°C/min, frequency: 1 rad/s, strain: 0.1%. Figure 6.4 2D CLSM images of 3% gelatin with (a) 0.1, (b) 0.8, (c) 1.6 and (d) 2.5% agarose, at 10x. Agarose phase is denoted in black and gelatin phase in purple. Scalebar denotes 100 µm. (a) Calibration curves of Gʹ at 5°C as a function of agarose (▲) and gelatin (●) concentrations. Modeling phase topology using – the isostrain (GʹU) and isostress (GʹL) blending laws for 3% gelatin and (b) 0.1, (c) 0.8 and (d) 2.5% agarose, and blending law for bicontinuous binary gels for (e) 3% gelatin plus 1.2,1.6, 2.0, 2.5 and 3% agarose (solid lines) arranged upwards along with the experimental values of the same concentrations (dotted lines) arranged upwards; and (f) predictions of the p factor for all the experimental concentrations based on the isostrain (■), isotress (●) and bicontinuous (♦) blending laws with the inset representing the estimated phase volume of the gelatin phase with increasing agarose concentrations. (a, d) 2D, (b,e) Z-stack and (c, f) 3D CLSM images of 3% gelatin and 0.4 and 2.5% agarose respectively. The gelatin phase is depicted in purple whereas agarose phase is in black.

174

xii

Figure 6.7 Gelatin phase volume estimates plotted against increasing agarose concentrations added to 3% gelatin determined by quantifying 3D CLSM images using image analysis software – FIJI and Imaris, and the rheology based blending laws.

Chapter 7

187 Figure 7.1

188 Figure 7.2

xiii

(a) 2D, (b) Z-stack, and (c) 3D CLSM images of 3% (w/w) agarose plus 2% (w/w) gelatin and 2% (w/w) canola oil. The agarose, gelatin and oil phase is denoted in green, black and red respectively. The composite 3D CLSM image is further split to denote 3D images of the constituent (d) agarose and gelatin, and (e) oil phase separately (Images captured at 10x, scale bar denotes 200μm). (a) 2D, (b) Z-stack, and (c) 3D CLSM images of 10% (w/w) gelatin plus 5% (w/w) WPI and 5% (w/w) canola oil. The gelatin, WPI and oil phase is denoted in red, black, and green, respectively. The composite 3D CLSM image is further split to denote 3D images of the constituent (d) gelatin and WPI, and (e) oil phase separately. (Images captured at 10x, scale bar denotes 200μm)

List of Tables

Chapter 1 Table 1.1

28 53 Summary of analytical techniques used to measure biopolymer concentration in separated phases in an aqueous mixture Spectroscopic techniques used to probe phase behaviour Table 1.2 Table 1.3 Microscopic techniques used for structural characterisation of phase

behaviour

87 87 92 Chapter 2 Table 2.1 List of chemicals and dyes used Table 2.2 List of instruments used Table 2.3 Absorption frequencies of various organic functional groups in the mid-

100

130

IR region Table 2.4 Rheological parameters Chapter 3 Table 3.1 Comparison of canola oil phase volume in mixture with agarose calculated using different FIJI thresholding methods (n=27)

141 Chapter 4 Table 4.1 Experimental ghee concentrations in the composite and estimated phase

volumes with FIJI and Imaris image analysis

xiv

189 Chapter 7 Table 7.1 Experimental constituent concentrations in the composites and estimated phase volumes with Imaris image analysis

List of abbreviations

BSE Back scattered electron

CLSM Confocal Laser Scanning Microscopy

Degree of polymerisation DP

Differential Scanning Calorimeter DSC

DMSO Dimethyl sulfoxide

DMA Dynamic Mechanical Analyser

DMTA Dynamic Mechanical Thermal Analyser

Dichlorotriazinyl Aminofluorescein DTAF

Fourier Transform Infrared Spectrometer FTIR

High performance liquid chromatography HPLC

Infrared IR

Linear Viscoelastic Region LVR

Microcrystalline cellulose MCC

MDSC Modulated Differential Scanning Calorimeter

Nuclear Magnetic Resonance NMR

Potential of Hydrogen pH

RMIT Royal Melbourne Institute of Technology

Scanning Electron Microscope SEM

Secondary Electron SE

Transmission Electron Microscope TEM

UV/Vis Ultraviolet/visible

versus vs

X-ray diffraction XRD

2D Two dimensional

3D Three dimensional

ANOVA Analysis of variance

e.g. Example

RMMF RMIT Microscopy and Microanalysis Facility

Milk protein concentrate MPC

Skim milk powder SMP

Diffusing wave spectroscopy DWS

Kilo Dalton kDa

Molar M

milligram mg

Pascal Pa

radian rad

Electronic spin resonance ESR

Small angle neutron scattering SANS

Small angle light scattering SALS

Sodium hydroxide NaOH

nanometer nm

Atomic force microscopy AFM

Weight by weight w/w

Weight by volume w/v

ESEM Environmental scanning electron microscopy

alpha α

xvi

Beta β

List of symbols and units

absorbance AU

Change in enthalpy ΔH

Change in temperature ΔT

Dalton Da

Degree °

Degree Celsius °C

°C/min Degree Celsius per minute

Storage/elastic modulus Gʹ

Loss modulus G′′

Glass transition temperature Tg

g/mol Gram per mole

Heat capacity Cp

Hertz Hz

Kilo volt kV

Litre L

Micrometer µm

Milliampere mA

Milligram mg

mg/kg Milligram per kilogram

Millimeter mm

Millilitre mL

minute min

xvii

molar M

Newton N

Pascal Pa

Per centimetre cm-1

Phase angle δ

Rotation per min rmp

second s

Water activity aw

Weight by weight w/w

Percent %

centimetre cm

nanometer nm

MPa Megapascal

gram g

density ρ

Angular frequency ω

Phase volume ɸ

wavelength λ

Rayleigh ratio Rθ

Solvent avidity parameter p

Complex modulus G*

magnification X

xviii

hour h

Explanatory notes

The following notes briefly delineate the points that were taken into consideration during the

writing of this thesis.

i) Attempts have been made to use British spellings in the text except in the published

journal articles where the journal guidelines have been followed.

ii) Symbols or abbreviations, used in place of a lengthy name or expression, have been

defined or explained in appropriate places as far as practicable.

iii) Wherever possible, SI units have generally been used in expressing results

throughout this thesis.

iv) The current guidelines to authors for Food Hydrocolloids (published by Elsevier)

has been followed in this thesis except in the published articles where journal

xix

guidelines have been followed.

Summary

The importance of estimating phase separation in protein and polysaccharide gels in

order to obtain desired structural properties and textural profile in food product formulations

remains paramount. Due to thermodynamic imbalance, biopolymer mixtures segregate into

distinct phases, with one phase acting as a continuous matrix in which the second phase remains

dispersed as a discontinuous ‘filler’. Studies up till now show that theoretical models relate the

elastic modulus of the biopolymer phases to the topology of their mixtures and can be

successfully employed for an indirect estimation of the solvent partition between the

biopolymers. Theoretical modelling, though robust, is a time consuming and tedious method

of estimating phase volume of phase separated polymers. Acknowledging this, a need for

developing a rapid, efficient, and accurate method of determining solvent partition between the

biopolymers is identified.

Therefore, this PhD study aims at exploiting the accelerated advancement in image

processing technology and developing a novel Confocal Laser Scanning Microscopy (CLSM)-

based approach using Z-stack imaging and image analysis to quantify phase behaviour of

biopolymer composites. This is achieved by employing numerous analytical and

physicochemical techniques such as scanning electron microscopy, Fourier transform infrared

(FTIR) spectroscopy, differential scanning calorimetry, rheology based theoretical modelling

and confocal laser scanning microscopy.

The first experimental chapter focused mainly on developing a protocol for obtaining

3D images of a simple model system of agarose and canola oil and quantifying the lipid phase

volume using two image analysis software, FIJI and Imaris in parallel. For comparison,

rheology based- theoretical blending laws were employed. The phase behaviour of composite

gels made of agarose and various concentrations of canola oil were examined using a variety

of techniques including SEM, FTIR, DSC, dynamic oscillation, polymer blending laws and

CLSM-based Z-stack imaging coupled with image analysis. Microscopic, spectroscopic, and

thermomechanical observations recorded continuous agarose networks supporting

discontinuous canola oil inclusions with increasing levels of canola oil reinforcing the rigidity

of the continuous phase. The outcome of the microscopic protocol is in close agreement with

the oil phase volume predictions from the isostrain blending law indicating the suitability of

the developed protocol in quantifying the phase behaviour of composite gels.

The second experimental chapter probed the phase behaviour of a model system

comprising agarose and a varying concentration of a hard lipid, ghee. Results obtained from

SEM, microDSC, FTIR and dynamic oscillation in-shear revealed discontinuous and hard

inclusions of ghee reinforcing the continuous, weaker agarose matrix with increasing

concentrations of the former. Phase behaviour of the system was quantified in parallel with a

novel method combining 3D CLSM imaging and image analysis software - FIJI and Imaris -

in an effort to substantiate the efficacy of the microscopic protocol in quantifying phase

behaviour. Phase volumes recorded with the microscopic protocol were in close agreement to

those modelled with the isostress blending law using small-deformation dynamic oscillation.

However, results indicated that the inner filtering effect or ‘self-shadowing’ observed

commonly in CLSM images due to the diffraction of laser as it passes through components

with different densities and refractive indices may pose a limitation to the application of this

technique, necessitating further development before it can be applied to more complex,

industrially relevant systems.

This merited further consideration of the suitability of the developed microscopic

protocol in estimating solvent partition in an industrially relevant hydrogel. In doing so, the

efficacy of confocal laser scanning microscopy (CLSM) paired with image analysis software

– FIJI and Imaris - in quantifying phase volume in a model system of agarose with varying

concentrations of microcrystalline cellulose (MCC) in comparison to the rheological blending

laws was probed. Structural studies performed using SEM, FTIR, differential scanning

calorimetry and dynamic oscillation in-shear unveiled a continuous, weak agarose network

supporting the hard, rod-shaped MCC inclusions where the composite gel strength increased

with higher ‘filler’ concentration. The phase volumes of MCC, estimated with the microscopic

protocol, matched the predictions obtained from computerized modelling using the Lewis-

Nielsen blending laws. Results highlighted the suitability of the microscopic protocol in

estimating the water partition and effective phase volumes in the agarose-MCC composite gel.

These were encouraging outcomes which lead us to estimate phase behaviour in a

mixed gelling system of a polysaccharide (agarose) and a protein (gelatin) in comparison with

the rheology based blending laws. Structural properties of the composites were probed using

FTIR spectroscopy, microDSC and small-deformation dynamic oscillation in shear.

Throughout the experimental range of concentrations, gelatin formed a continuous network

whereas the agarose phase remained dispersed as either a soft or a hard filler at low agarose

concentrations (0.1-0.8%) and formed a continuous network alongside that of gelatin at higher

concentrations (1.2-3%). The phase volumes of gelatin, recorded using Imaris, were a close

match with those obtained from the blending law predictions, whereas FIJI yielded statistically

different estimates. Results suggest that the microscopic protocol using Imaris shows promise

and can potentially be used as an alternative to the theoretical blending laws in estimating phase

behaviour in biomaterial composites.

Work outlined in this thesis thus lays the groundwork for future research, where it can be

used to determine phase behaviour accurately and rapidly in complex binary and tertiary gelling

biomaterial composites and perhaps to materials in the nanoscale.

Chapter 1 Background and Literature Review

1.1. Introduction

Industrial processing generally involves multiple polymers interacting intricately to

develop products whose physiochemical properties are distinct from those of their constituents.

Mixtures of biopolymers thereby broaden the spectrum of structural and textural variations

more than deemed possible and offer a better control over processability, nutritional value of

the product, kinetics of active compound release and texture and shelf stability of the final

product (Prameela, Mohan, & Ramakrishna, 2018). Hence, it is no surprise that biopolymers

today have a marked predominance in the food, pharmaceutical, cosmetic, and packaging

industry to name but a few.

Biopolymer composites, however, are stable and co-soluble only in very dilute systems

or in cases where the mixing entropy dominates. When the concentration of the biopolymers is

increased beyond the critical concentration, the system becomes unstable and phase

separation/demixing occurs (Walter, Johansson, & Brooks, 1991). Phase separation can either

be segregative, where the biopolymers repel each other leading to an enrichment of each

constituent in a distinct phase, or associative, where due to weak attractive forces and

nonspecific interactions (hydrogen bonding, electrostatic, hydrophobic and Van der Waals

interactions) both the biopolymers concentrate in one phase forming complex coacervates or

insoluble compounds (Tolstoguzov, 2007).

Since the pioneering works of Beijerinck (1896), outlining the formation of tiny

droplets in a mixed solution of starch and gelatin, there is no scarcity of publications

investigating and reviewing phase separation in biopolymer mixtures, the mechanism behind

it, the factors influencing it and ways in which it can be exploited by the processing industries.

Despite the plethora of literature available, phase behaviour of biopolymer mixtures remains

of interest and an active research area among researchers and industrialists.

Generally, while investigating mixed biopolymer systems in the aqueous state, a phase

diagram is created of the constituents at desired processing parameters to establish the

boundaries of their compatibility. Depending on the accuracy required, several methods can be

employed to construct a biopolymer phase diagram. Rudimentary methods such as visual

observation (Hemar, Tamehana, Munro, & Singh, 2001) and the phase volume ratio method

(Antonov, Gonçalves, 2018; Schorsch, Clark, Jones, & Norton, 1999), frequently accompanied

by centrifugation (Thaiudom & Goff, 2003) to speed up the process and induce bulk phase

separation, have also been successfully utilised for simple binary mixtures. The construction

of a more detailed and precise phase diagram entails determination of biopolymer

concentration in the separated phases using analytical techniques. Techniques like

refractometry (Semenova & Savilova, 1998), flow injection analysis (Kontogiorgos, Tosh, &

Wood, 2009), phenol-sulphuric method (Masuko et al., 2005), protein analysis method (C

Schorsch, Jones, & Norton, 1999), turbidimetry (Kontogiorgos et al., 2009), viscometry and

quantitative HPLC have been successfully utilized in the past to determine

protein/polysaccharide concentration in the individual phases.

Industrially relevant biopolymer systems, however, often comprise of highly viscous

and gel forming polymers, making it difficult to accurately study their phase behaviour using

the aforementioned techniques. A biopolymer gel, which is essentially a three-dimensional

network of polymer chains, entraps a large volume of solvent within its network and in the

presence of a gel forming polymer; therefore, gelation and phase separation act as competing

processes (Anderson & Jones, 2001). The onset of gelation arrests the polymer demixing in a

state of non-equilibrium, yielding a complex, unique structure for the composite. Although this

opens up endless possibilities of manipulating the properties of the resultant gel by altering the

effects of the two processes, pinning the exact mechanism and extent of phase separation

becomes a formidable problem. As such, there is no time-temperature superposition principle

that can sustain a composite gel’s state of thermodynamic equilibrium as the phase diagram

does in solutions (Shrinivas, Kasapis, & Tongdang, 2009).

In composite gels, the concentration of the phases for one, can vary considerably within

the system subject to the degree of phase separation attained when gelation occurred. The

demixed phases may not be pure solutions and can exhibit the presence of multiple inclusions

of each phase within each other. The internal and external conditions of the system may further

lead to changes in the micro and macroscopic structure of the composite gel. With their

complex chemical composition and intricate structural hierarchy, the systems often need to be

‘adjusted’ to overcome interferences from other components. Systems may need to be

physically or chemically treated to make them suitable to be analysed, introducing the risk of

physical contamination and chemical alteration, making the results additionally dubious and

the whole endeavour laborious (Agbenorhevi & Kontogiorgos, 2010; Tolstoguzov, 1999).

Given that biopolymer mixtures encompass a broad range of products with a broader

spectrum of textural and structural properties, accurately determining their phase behaviour

with a single technique is a major challenge as it results in an incomplete, tunnel visioned

understanding of the composite. In the past, different types of gravimetric (Mousia, Farhat,

Blachot, & Mitchell, 2000), calorimetric (Conti, Yoshida, Pezzin, & Coelho), spectroscopic

(Icoz & Kokini, 2007), microscopic (van de Velde et al., 2003), rheological analysis

(Semasaka, Mhaske, Buckow, & Kasapis, 2018), and scattering measurements have been used

to characterize phase behaviour of multiple biopolymer composites from the molecular level

up to their macroscopic behaviour and properties (Fig.1.1). Each of these techniques have their

own advantages and limitations while probing specific parameters of phase-separated

composites under certain experimental conditions and a thorough understanding of the

characterisation techniques can facilitate a better, interdisciplinary approach of following phase

behaviour.

Fig.1.1. Scale of measurement of the various techniques used to probe phase behaviour.

This review therefore, aims at summarising the basic principles of the techniques

inherent in the estimation of phase behaviour of mixed biopolymer systems in the aqueous and

gel state, along with their key advantages and limitations with relevant examples without

attempting to provide an exhaustive list of applications.

1.2. Phase behaviour in mixed biopolymer systems

Biopolymers in a mixture seldom behave in the same way as they would in the absence

of the other. In mixture of similar biopolymers and polyelectrolytes, a biopolymer can change

the ionic environment of the solution inducing gelation, complex formation or other associative

interactions of the other (Sarika, Pavithran, & James, 2015), or compete with it for ions from

added salts in the mixture (Donato, Garnier, Novales, & Doublier, 2005). On the other hand, a

mixture of dissimilar biopolymers, proteins and polysaccharides for example, can have one of

three interactions – co-solubility, complexing or segregation due to incompatibility. As

mentioned earlier, a co-soluble biopolymer mixture is a rare possibility, occurring in cases

where the total polymer concentration does not exceed 2-4% for aqueous mixtures of linear

polysaccharides or gelatin and 12% for mixtures containing globular proteins, making phase

separation in biopolymer composites a rule rather than an exception (Tolstoguzov, 2003;

Zasypkin, Braudo, & Tolstoguzov, 1997).

Depending on the polymeric nature of the biopolymers, their molecular weights and the

different functional groups present in their macromolecules, the constituent biopolymers in a

mixture are either attracted to or repulsed from each other. This forms the basis for their

complexing or thermodynamic incompatibility (de Kruif & Tuinier, 2001). Electrostatic

attraction between polyelectrolytes such as the anionic polysaccharides and cationic proteins

below their isoelectric points leads to the formation and precipitation of the two polymers as a

complex coacervate. Formation of a coupled network as a result of association between two

gelling biopolymers is also observed in certain, but rare cases (Tanaka, 2002). The most

common outcome of mixing two biopolymers is a segregative interaction or thermodynamic

incompatibility which occurs due to the fact that macromolecules tend to be surrounded by

others of the same type and polarity, resulting in a reduced concentration of one polymer in

close proximity of the other (Grinberg & Tolstoguzov, 1997; Tolstoguzov, 1993). When the

concentration is high enough, the biopolymers separate completely into two phases each rich

in one biopolymer and depleted in another (Frith, 2010).

Gibbs free energy of mixing which considers the entropy and enthalpy of mixing is a

popular method of deducing whether the constituents of a mixture will phase separate or not.

Non-covalent interactions between the solutes including polar and non-polar attractions, ion

pairing, van der Waals forces and hydrogen bonding are exothermic reactions whereas

endothermic reactions include repulsion between molecules or disruption of molecular

bonding. Entropy denotes the energy changes corresponding to the alterations in the molecular

arrangement. A positive entropy corresponds to the increase in the degree of freedom of

molecular motions, while a negative value denotes a decrease in it (Schmitt, Sanchez, Desobry-

Banon, & Hardy, 1998). Gibbs free energy of mixing (ΔGmix), is given by the following

equation (Goh, Teo, Sarkar, & Singh, 2020):

ΔGmix = ΔHmix - TΔSmix (1)

Where ΔH is the enthalpy of mixing and ΔS the entropy of mixing at temperature T in Kelvin.

Considering systems in equilibrium, according to the second law of thermodynamics,

co-existing phases must have the lowest ΔGmix at temperature T. In most cases, ΔHmix is

positive, favouring demixing of the solutes. Solutes with a low molecular weight have an

adequately large ΔSmix value that results in the co-solubility of solutes. In biopolymers,

however, ΔSmix is much smaller and ΔHmix dominates, leading to phase separation. Depending

on the charge on the two biopolymers, the system then undergoes either an associative or a

segregative phase separation. In segregative interactions, there is no thermodynamic drive for

heterotypic association and the enthalpic drive for segregation dominates (Tolstoguzov, 1991).

Grinberg & Tolstoguzov (1997), in their work report the incompatibility conditions for over a

hundred different protein-polysaccharide composites. The thermodynamics and kinetics of

phase separation have been discussed in detail elsewhere (Bergfeldt, Piculell, & Linse, 1996;

Flory, 1953; Fredrickson, Liu, & Bates, 1994; Johansson, Karlström, Tjerneld, & Haynes,

1998; Robeson, 2007).

1.3. Phase diagrams of protein-polysaccharide-water systems

A phase diagram depicts the phase behaviour of two thermodynamically distinct

biopolymers (Fig.1.2) (Goh et al., 2020; Tolstoguzov, 1991, 2003, 2006). The concentrations

of biopolymers are generally plotted on the axes in weight percent, the rest of the phase being

the solvent. Each point in the diagram denotes a system composition, with the binodal curve

separating the miscible composite concentrations below it from the two-phase system

concentrations above. The closer the binodal is to the axes, the higher the incompatibility

between the biopolymers. Lowering the temperature of the system, increasing the ionic strength

or the molecular weights of the constituents decreases the entropy of mixing and displaces the

binodal towards the origin (Morris, 2009). A composite with a total biopolymer concentration

C, constituting a% of biopolymer 1 and b% of biopolymer 2, phase separates into two bulk

biopolymer-rich phases. The biopolymer 1 enriched phase will have d% biopolymer 1, while

biopolymer 2 enriched phase would have a composition of e%. The biopolymer 1 enriched

phase contains negligible traces of biopolymer 2 and vice-versa. Line d-C-e denotes the tie

line. The ratio of the line segments dC/Ce corresponds to the phase volume ratio of the two

separated phases. Therefore, any composite composition that lies on the tie line would result

in the same effective concentration in the separated phases, but their phase volumes will vary.

A rectilinear diameter can be obtained by joining the midpoints of two or more tie-lines. Any

composition on the rectilinear diameter would phase separate into two phases with the same

phase volume and biopolymer concentration.

Fig.1.2. Phase diagram in an aqueous biopolymer mixture.

When the overall composite concentration shifts closer to the origin, the tie lines

become shorter till the two points converge on a single point on the binodal. This point is

known as the critical point, or separation threshold, F in Fig.1.2 which denotes the minimum

critical concentration at which the biopolymer composite will phase separate into two phases.

A lower critical point denotes lower compatibility between the constituent biopolymers. In an

ideal biopolymer composite, when the composite concentration is varied, the corresponding

tie-lines do not cross over. However, this is the case in an ideal biopolymer composite. In reality

most composites exhibit a much more complex behaviour.

An important aspect to consider is the spinodal, the dotted line in Fig.1.2. Lying within

the binodal, the spinodal separates composites based on the nature of their phase separation. If

a composite lies between the binodal and the spinodal, it is likely to phase separate via

nucleation, taking more time than the timescale of the experiment and making it difficult to

study. Composites within the spinodal region undergo spontaneous phase separation owing to

the increasing thermal concentration fluctuations. Gelation of one or more of the biopolymers

of the composite is another factor that needs to be considered as it may arrest phase separation

at different stages depending on various factors or prevent it completely.

As the phase diagram provides critical information of biopolymer compatibility and can

be used to predict the behaviour of two biopolymers in a composite, it is a requisite for the

study of phase behaviour and the fabrication of novel textures and products. Innumerable

publications outline the works of researchers and industrialists studying phase behaviour and /

or constructing phase diagrams of biopolymer composites. Use and development of multiple

techniques is undertaken in an effort to not overlook any of the plenitude of factors that affect

phase behaviour of a composite. The rest of this review will focus on the various techniques

used to probe phase behaviour.

1.4. Techniques used to probe phase behaviour

The various analytical techniques used to probe phase behaviour in aqueous mixtures will be

discussed in detail in the following sections and briefly summarised in Table 1.1.

1.4.1. Visual observation and phase volume ratio method

The earliest technique used to study phase behaviour was visual observation. Two

polysaccharides in varying concentrations were mixed to form composites with the same total

polysaccharide concentration. The mixtures would either be left undisturbed over a longer

period of time or be centrifuged to speed up phase separation. The binodal is plotted by

following the different concentration of both biopolymers at which phase separation occurs.

Depending on the degree of incompatibility between the constituents, phase separation within

the composite ranged between microscopic and macroscopic. Michon, Cuvelier, Launay,

Parker, & Takerkart (1995) constructed a phase diagram of acid gelatin and iota carrageenan

at different pH and added salt and further detailed it by classifying the mixtures as ‘clear’ when

it was possible to see through them, ‘cloudy’ when objects appeared blurred when viewed

through the composite, ‘opaque’ when nothing could be seen through them (Fig.1.3).

Thaiudom & Goff (2003) followed the compatibility of skim milk powder with locust bean

gum, guar gum and xanthan gum by centrifuging the mixtures and observing the composite

while Hemar, Tamehana, Munro, & Singh (2001) successfully constructed the phase diagram

of milk protein concentrate (MPC) or skim milk powder (SMP) when mixed with xanthan gum

by visually observing the components phase separate into two distinct layers with a clear phase

boundary.

To qualitatively determine the binodal, Grinberg & Tolstoguzov (1972) added a

solution of one constituent biopolymer into the second biopolymer solution till turbidity

Table 1.1: Summary of analytical techniques used to measure biopolymer concentration in separated phases in an aqueous mixture.

Technique

Principle

Advantages

Disadvantages

References

 Suitable for a wide

 Suitable only for bulk separated

Application (examples)  Locust bean

(Antonov & Gonçalves, 2018)

Phase volume ratio

range of composites

phases

gum + gelatin

 Doesn’t need sophisticated instruments for analysis

 Determination of cloud point concentration results in a coexistence curve which may involve solutions of higher concentrations than initially used

Volumes of two bulk phase separated phases is measured in a calibrated capillary tube; graphical calculation using geometrical analysis is used to calculate weight fraction of the two polymers in individual phases.

 Assumes complete depletion of one polymer in the separated phase of the other which isn’t always the case

Refractometry Refractive index of a

 Sucrose in agar

 Sample needs to have low optical

gels

density

 Micellar casein +

 Significant differences between the refractive indices of the two biopolymers necessary

(Wang et al., 2014; Yang et al., 2015; Bourriot et al., 1999; Semenova & Savilova, 1998)

 More suitable for non-interacting systems that have undergone bulk phase separation

 Easy to use  Quick  Small amounts of sample needed  Doesn’t need a sophisticated laboratory set up for estimation

gaur gum  11S globulin of faba beans + fibrinogen/ dextran/ carboxymethylce llulose/ sodium alginate



Phenol sulphuric method

Interferences from protein component significantly impact absorption readings



(Zhang & Foegeding, 2003); (Li et al., 2016); (Perrechil et al., 2009)

 β-lactoglobulin + dextran sulfates / kappa- carrageenan soluble rice protein + dextran

 Addition of sulphuric acid to proteins produces varying amounts of ammonium sulphate,

 Quick  Easy  Reliable  Reagents are easily available and affordable

medium is linearly proportional to its concentration within a range of polymer concentration at a specific temperature. Thus, measuring the refractive index can be an indirect method of estimating the polymer concentration. Phenol and sulphuric acid are added to polysaccharide solution. Dehydrated polysaccharides react with phenol to produce furfural which forms a

CO2 and sulphur that alter the native system

 As the protein concentration

/ kappa- carrageenan  Sodium caseinate + locust bean gum

coloured solution whose absorbance is measured. The intensity of the colour produced is directly proportional to the polysaccharide concentration.

varies, different blank solutions are needed for blank correction  Certain systems necessitate the removal of protein from system before estimating the polysaccharide concentration

 BSA +

Bradford Assay

 Results in varied response between different proteins

Ribonuclease A

(González- González, 2011); (Svensson, 1999)

 Assay is linear over a short range

 Dextran +

 High sensitivity  Easy operation  Universally accepted

of concentrations

technique

albumin, pluronic acid

 Dilution is often necessary

Biuret method



 Affected by the association state

Independent of protein composition

(Farouk et al., 2011); (Zhuang et al., 2020)

 Meat proteins + carrageenan  Myofibrillar



and purity of proteins Isn’t as sensitive as other protein assays

protein + dietary fibre/ modified starch/ konjac glucomannan

 Canola protein

Kjeldahl method

(Neirynck et al., 2007); (Klassen et al., 2011)

 Universally accepted  High precision  High reproducibility

isolate + alginate + carrageenan  Sodium caseinate

 Involves high set up cost  Other components containing nitrogen can result in errors in protein estimation

guar gum

 Different proteins need different



correction factors Is an indirect method of protein estimation

The anionic part of the Coomassie blue dye, attaches to argenine or lysine residues. The absorbance of the coloured complex is measured. In alkaline conditions, copper sulphate is added to a protein solution. Cupric ions bind with peptide bonds developing a purple colour, the absorbance of which is measured using a colorimeter. Proteins are digested using an acid to release ammonia. The released ammonia/nitrogen is calculated by titration as an indirect estimation or the protein content.

occurred and then added the second solution till the composite formed a clear solution. This

was repeated for a series of individual biopolymer solutions with varying concentrations.

Construction of the phase diagram using this cloud point system is convenient when the

constituent biopolymer solutions have a low optical density, are in the aqueous state and the

difference in their refractive coefficient is significant (Tolstoguzov, 2006). However, visual

observation is a qualitative method, where the concentration of biopolymers in the separated

phases and the tie lines of compatibility cannot be determined in the phase diagram.

Fig.1.3. Phase diagram of 3% gelatin/0.9% iota-carrageenan as a function of sodium chloride

concentration and pH (Michon, Cuvelier, Launay, Parker, & Takerkart, 1995).

Polyakov, Grinberg, & Tolstoguzov (1980) developed a more refined technique called

the phase volume ratio method that could be used to deduce the concentration of the polymers

as well as the tie lines in the phase diagram. Similar to visual observation method, mixtures of

two biopolymers A and B with individual weight concentrations, Ai and Bi were mixed together

in varying concentration and allowed to phase separate. This was carried out in graduated

centrifuge tubes or calibrated capillary tubes in order to determine the volumes (V) of the

separated phases. The phase volume ratio (r) of the phases could be calculated using the

following formula (Schorsch et al., 1999):

r↑ = V↑/ (V↑+ V↓) (2)

Where ↑ denotes the separated upper phase and ↓ is the lower phase.

The apparent weight fraction of the individual separated phases, y is given by:

(3) y = Ac/Ai

(4) (1-y) = Bc/Bi

Where Ac and Bc are the weight fractions of polymer A and B in the mixture.

The ratio of the volume is then plotted against the apparent weight fraction of the

respective component. Extrapolating the curve of the phase volume ratio as a function of the

apparent weight fraction, with the values of r ranging between 0 and 1, give the values of r on

the boundary of a stable homogeneous mixture and two separated phases. r↑ = 0 denotes one

phase similar to the lower phase while r↑ = 1 corresponds to one phase similar to the upper

phase. This procedure is repeated with solutions of varying polymer concentrations Ai and Bi

with different polymer A/polymer B ratio. The binodal is drawn through the points representing

the composition of the phases after it has been determined. The points denoting composition

of two co-soluble phases are joined to get the tie lines on the phase diagram. The system

composition where r = 0.5 corresponds to the point on the rectilinear diameter (Tolstoguzov,

1999). Antonov & Gonçalves (2018) used the phase volume ratio to construct a phase diagram

of locust bean gum and gelatin. Though this technique overcomes the drawbacks of visual

observation in terms of being able to estimate the polymer concentration in the separated phases

and enables determination on tie-lines, the method only deals with systems that bulk phase

separate and the determination of cloud point concentrations gives a coexistence curve which

isn’t always the same as the initial concentration of phases in the composite left to phase

separate. The method has the added disadvantage that it considers complete depletion of one

polymer from the other leading to anomalies in concentration estimation. To overcome these

draw backs, refractometry, turbimetric methods and methods to chemically estimate polymer

concentration in the separated phases are used which have been explained in detail in the

following sections.

1.4.2. Refractometry

The simplest and fastest method of estimating biopolymer concentration in a solution

is refractometry. The speed of light changes in different materials based on the material’s

density. This difference in the speed changes the direction of light as it passes through different

materials. The ratio of the speed of light in vacuum to the speed of light in a specific material

is a constant physical characteristic of the material known as index of refraction (n). The

behaviour of light as it passes between two materials is determined using the Snell’s law of

refraction given by (George, 2001):

(5) 𝑛 = 𝑠𝑖𝑛𝛩𝑖 𝑠𝑖𝑛𝛩𝑟

Where Θi and Θr are angles made by the incident ray and the refracted ray, respectively, at the

intersection of the two materials. The value of n varies according to the wavelength of incident

light and in an aqueous solution is influenced by the temperature of the solution, the solute

added and its concentration. Every solute, or biopolymer, has a specific value of n per

concentration in g/dL, termed as specific refractivity. The refractive index of a biopolymer

solution is linearly proportional to the biopolymer concentration over a wide range of

concentrations at a specific temperature. With an increase in temperature, the value of n

decreases due to the increase in solvent volume. By determining the n of a standard solution at

a specific temperature, it is possible to determine the biopolymer concentration at that

temperature.

According to the generally accepted Stockmayer theory (Stockmayer, 1950), in a three

component system, containing biopolymer 1, biopolymer 2 and water, when the concentrations

are expressed in molal units, the interaction between the two biopolymers is measured directly

by the difference between the increments of the refractive index. The refractive indices are

measured under conditions at which the chemical potential of the second biopolymer and its

molality are maintained at the same level as in the solution and reference solvent. The change

in the refractive index, dn is given by (Pittz, Lee, Bablouzian, Townend, & Timasheff, 1973):

𝑇,𝑝,𝑚2

𝑇,𝑝,𝑚1

(6) ) ) 𝑑𝑚2 𝑑𝑛 = ( 𝜕𝑛 𝜕𝑚1 𝑑𝑚1 + ( 𝜕𝑛 𝜕𝑚2

Where n is the refractive index, m1 and m2 is the molal concentration of biopolymer 1 and 2

respectively, p is the pressure (Pa) and T is the temperature (°C).

The value of n can be converted to biopolymer concentration using a refractometer

reticule, or an established calibration curve of the standard. However, the angle of refraction

produced by a biopolymer solution, Θr is a result of the combined concentration of the

constituents also called, “total solids” in a solution. In a biopolymer composite, therefore, if

the constituent biopolymers have similar specific refractivities, the presence of the second

biopolymer will interfere with accurate concentration measurement. To overcome this, Ding et

al. (2002) combined refractometry with UV-Vis spectroscopy to measure the concentration on

dextran and gelatin in separated phases. UV-Vis spectroscopy was first used to determine

concentration of gelatin in both the phases and the increments in refractive index of both the

biopolymers were predetermined. With this information at hand, the concentration of dextran

was calculated using the refractive index of the separated phase. Other examples of applications

of refractometry in concentration estimation include, but aren’t limited to sucrose in agar gels

(Wang, Yang, Brenner, Kikuzaki, & Nishinari, 2014; Yang et al., 2015), micellar casein- gaur

gum (Bourriot, Garnier, & Doublier, 1999) and fibrinogen, dextran, carboxymethylcellulose,

sodium alginate and 11S globulin of faba beans (Semenova & Savilova, 1998).

Use of refractometry, however, necessitates that both the biopolymers have a low

optical density with a significant difference in their refractive indices, which is not always the

case with biopolymers. In cases like Ding et al., (2002), it is assumed that the refractive index

of the composite is the sum of the refractive indexes of the individual composites, an

assumption that holds true just for non-interacting systems. The interference of the second

biopolymer makes this inapplicable to systems that haven’t undergone bulk phase separation

and its high temperature sensitivity necessitates that temperature is monitored throughout the

analysis (Loret, Schumm, Pudney, Frith, & Fryer, 2005).

1.4.3. Chemical analysis for concentration estimation

A popular technique of estimating biopolymer concentration in the separated phases is

by chemical analysis. Due to its simplicity and high sensitivity, concentration measurement by

UV absorbance is widely used. However, alternative methods such as optical activity, HPLC,

electrophoresis, and other radiometric and chromatographic methods have also been

successfully used and are discussed in the following sections.

Perhaps the most popular method for estimating the concentration of pure sugar

solution, oligosaccharide of polysaccharides is the phenol sulphuric method (Dubois, Gilles,

Hamilton, Rebers, & Smith, 1956). In this method, trace quantities of phenol are added to an

aliquot of sample along with sulphuric acid. Polysaccharides, when dehydrated due to the

reaction with concentrated sulphuric acid produce furfural derivatives. These furfural

derivatives further react with phenol to form a coloured solution. Absorption, which is

proportional to the intensity of the colour developed, is measured at 490nm (Jain, Karibasappa,

Dodamani, & Mali, 2017). This method has been used for quantification of polysaccharides in

a number of studies including the works of Zhang & Foegeding (2003), Li et al. (2016) and

Perrechil, Braga, & Cunha (2009). Masuko et al. (2005) further modified the technique to make

it suitable for microplate analysis. The popularity of the method can be attributed to the fact

that it is quick, easy and reliable involving affordable and easily available reagents. However,

the method is established for pure solutions and the presence of even minor traces of proteins

in the solution have a significant impact on the absorption measurements. Sulphuric acid reacts

with protein to form ammonium sulphate, carbon dioxide and sulphur which alter the overall

chemistry of the reaction. Moreover, multiple blanks are needed for blank correction to

attenuate signals from non-polysaccharide sources (Agbenorhevi & Kontogiorgos, 2010).

A more accurate method of protein estimation is binding it with the anionic dye,

Coomassie blue. This technique is also known as the Bradford assay (Bradford, 1976). The

anionic form of the dye adsorbs to the non-polar amino acid (ideally arginine and lysine)

residues, and the cationic region of proteins. The absorption of the protein-Coomassie blue

complex is measured at 595 nm (González-González, Mayolo-Deloisa, Rito-Palomares, &

Winkler, 2011; Svensson, Berggren, Veide, & Tjerneld, 1999). Since the dye exhibits a

preferential binding to arginine and lysine, it subsequently results in a varied response between

different proteins. The assay is linear typically just within the short range of concentrations

within 0 µg/mL to 2000 µg/mL making dilution necessary before analysis and error in one

dilution is compounded throughout affecting the accuracy of the assay (Ernst & Zor, 2010). In

spite of these complications, the high sensitivity and ease of the assay make it a popular

technique in protein estimation.

Another popular method for the determination of protein concentration in a solution is

the use of biuret reaction, the approaches of which have been around since the works of Gornall

et al. in 1949. In this method, copper sulphate is mixed with a protein in a high alkaline solution.

Cupric ions from copper sulphate form a complex with the proteins by binding at the peptide

bonds, producing a violet-purple colour (Farouk, Frost, Krsinic, & Wu, 2011; Zhuang, Wang,

Jiang, Chen, & Zhou, 2020). This reaction, unlike the Bradford assay, is independent of the

protein composition but is affected by association state and purity of proteins. This method,

however, isn’t as sensitive as the other methods of protein estimation and attempts have been

made to modify the technique to increase its sensitivity (Matsushita, Irino, Komoda, &

Sakagishi, 1993). The interference caused by the presence of other components in the solution

such as lactose, fat and turbidity have been successfully eliminated by adding hydrogen

peroxide to the solution (Reichardt & Eckert, 1991). The biuret reaction was amplified by

including a second reaction with Folin phenol reagent in the Lowry method (Waterborg, 2009).

However, in the Lowry method, specific amino acids, namely tyrosine and tryptophan affect

the development of colour resulting in substantial variation in accordance to protein

composition. With the development of simplistic and sensitive protein assays, these methods

however have witnessed a dip in popularity.

The classic Kjeldahl method is an alternative to the colorimetric analysis where the

proteins are ‘digested’ or wet oxidised in the presence of a catalyst in sulphuric acid. The acidic

ammonium sulphate formed is diluted in water and alkalinised. This is followed by titration to

estimate the content of ammonia/nitrogen representing the amount of protein in the sample

(Bradstreet, 1954). Kjeldahl method is popular globally due to its high reproducibility and

precision. However, it involves a high set up cost and is an indirect method of protein

estimation that may result in errors as samples may have other sources nitrogen apart from

protein. Furthermore, different proteins need different correction factors due to the varying

amino acid sequence (Mæhre, Dalheim, Edvinsen, Elvevoll, & Jensen, 2018).

All the chemical assays mentioned here, along with others that aren’t so popular, for

example, use of methylene blue, ninhydrin, bicinchoninic acid, Dumas method, BCA assay,

‘blue value’ assay, have been explained in detail elsewhere in literature (Copeland, 1994;

Gilbert & Spragg, 1964; Jiang, Tsao, Li, & Miao, 2014; Marshall, 2005; Sapan, Lundblad, &

Price, 1999; Soedjak, 1994). A common drawback for all chemical assays of concentration

estimation is that it is only feasible for a polymer if there is an established test for it.

1.4.4. Viscometry

Along with its wide application in determining degree of polymerisation, molecular

weight and molecular weight distribution (Masuelli & Illanes, 2014), there have been various

attempts of correlating the viscosity of ternary polymer solutions with its miscibility. When a

biopolymer is added to a solvent, its configuration and hydrodynamic volume is influenced by

the solvent used, affecting the solution’s viscosity. In a biopolymer composite, it is assumed

that beside solvent-biopolymer interactions, mutual interactions of the macromolecules greatly

influence composite viscosity. The attractive or repulsive forces between the unlike polymers

can lead to the contraction or expansion of the polymer strands. According to the classic

Huggins theory, the dimensions of the strands can be correlated to the intrinsic viscosity of a

composite. Viscosity of biopolymers can be measured using viscometers and rheometers that

have been explained in detail elsewhere (Gupta, Wang, & Vanapalli, 2016).

Characterisation of the thermodynamic and hydrodynamic interaction between the

biopolymer strands can help determine the phase behaviour of the composite. According to the

framework established by Cragg-Bigelow (1955) and Krigbaum-Wall (1950), the adaptation

(𝜂𝑠𝑝)

𝑖

of the Huggins equation to a polymer-solvent system is denoted as

𝑐𝑖

(7) = [𝜂]𝑖 + 𝑏𝑖𝑖𝑐𝑖

Where (ηsp)i is the specific viscosity and ci is the polymer concentration. [η]i is the intrinsic

viscosity given by [η]i = limCm→0 (ηsp)i/Ci. bii is the viscometric interaction parameter related

i (Vis et al.2015).

to the Huggins coefficient as bii = kH[η]2

(𝜂𝑠𝑝)

𝑚

For a mixture of polymers in a solvent, Eq. 7 can be extended as:

𝑐𝑚

(8) = [𝜂]𝑚 + 𝑏𝑚𝑐𝑚

Where ‘m’ stands for mixture. Since reduced viscosity is an additive property, Eq. 8 can be

(𝜂𝑠𝑝)

𝑚

𝑖

rewritten in the weight-average form as follows:

𝑐𝑚

𝑐𝑖

(9) = ∑ 𝑖 𝑤𝑖

Where wi is the weight fraction of the biopolymer i (i =2,3) given by wi = ci/cm. Combining

(𝜂𝑠𝑝)

𝑚

Eqs. 1 and 3, leads to :

𝑖

𝑐𝑚

1/2

+ = ∑ ([𝜂]𝑖 + 𝑏𝑖𝑖𝑐𝑖) 𝑤𝑖 = ∑ [𝜂]𝑖𝑤𝑖 𝑖 + 𝑐𝑚 ∑ 𝑏𝑖𝑖𝑤2 𝑖 = ∑ [𝜂]𝑖𝑤𝑖 𝑖

𝑖

2 𝑤𝑖)

(10) 𝑐𝑚(∑ 𝑏𝑖𝑖

Comparing Eqs. 2 and 4, with i=2,3 it can be deduced that:

𝑖

(11) [𝜂]𝑚 = ∑ [𝜂]𝑖𝑤𝑖 = [𝜂]2𝑤2 + [𝜂]3𝑤3

1 2 𝑤𝑖)

1 2𝑤2 + 𝑏33

1 2𝑤3)

And 𝑏𝑚 = (∑ 𝑏𝑖𝑖 = (𝑏22

2 + 𝑏33𝑤3

2 + 2𝑏22

1/2𝑏33

1/2𝑤2𝑤3

id between the strands

(12) = 𝑏22𝑤2

The ‘ideal’ value of the universal viscometric interaction parameter, bm

of the constituent biopolymers in the mixture is defined using Eq. (6). This is done with an

id, the geometric

assumption that for the ideal value of the specific interaction parameter b23

1/2

mean of those parameters corresponds to interaction between like-stranded polymers as

id = 𝑏22

1/2𝑏33

(13) b23

By parallelism, the experimental value of the same parameter can be defined as:

𝑒𝑥𝑝 = 𝑏22𝑤2

2 + 𝑏33𝑤3

2 + 2𝑏23

𝑒𝑥𝑝𝑤2𝑤3

𝑒𝑥𝑝 is estimated from the slope of Eq. (2), and b22 and b33 are determined using Eq 1

(14) 𝑏𝑚

Where 𝑏𝑚

from the composite formed by polymers 2 and 3, respectively.

𝑒𝑥𝑝

The compatibility between the biopolymers in a mixture is based on the comparison

between the theoretical and experimental b23 values (Krigbaum & Wall, 1950). When 𝑏23

id or Δbm > 0 the biopolymers are compatible or have attractive forces between them and

𝑒𝑥𝑝 < b23

id or Δbm < 0, the polymers are incompatible with repulsive

> b23

𝑒𝑥𝑝 = b23

id or Δbm = 0, the polymers have

conversely when 𝑏23

molecular interactions. In hypothetical cases where 𝑏23

neither attractive nor repulsive forces acting between them.

Viscometry has been used to determine the phase behaviour of a number of systems

involving both organic solvents (Chee, 1990; Danait & Deshpande, 1995) and water (Inamura

& Jinbo, 1991; Parets, Garcia, Soria, & Campos, 1990; Wanchoo & Sharma, 2003) as solvent.

The method is simple and straightforward and does not require sophisticated equipment.

id is based on the geometric

However, there seems to be ambiguity in the theoretical framework that has been a subject of

debate among researchers. As pointed out in Eq. 7, calculation of b23

mean value, while there have been arguments that support the use of arithmetic mean value to

id (Garcı́a, Melad, Gómez, Figueruelo, & Campos, 1999). Garcı́a et al. (1999) also

define b23

raise the concern of mathematical errors in the framework that has been explained in their work.

1.4.5. Spectroscopy

Spectroscopy deals with the generation, measurement and interpretation of the spectra

produced during the interaction of electromagnetic radiation with matter. Multiple

spectroscopic techniques are available, differing in terms of the region of the electromagnetic

spectrum used for analysis (e.g. UV, visible light, IR), the type of radiation-matter interaction

being observed (e.g. diffraction, absorption, emission) and the material being analysed (atomic

or molecular spectroscopy). The common techniques used to analyse phase behaviour have

been discussed below and summarised in Table 1.2.

1.4.5.1. UV-Vis spectroscopy

This type of spectroscopy analyses compounds using electromagnetic radiations

between 10 nm to 700 nm and has the most applications in the realms of food biopolymer

science. With relatively low energies, the UV-Vis spectrum can only excite the chromophores,

molecular sub-structures that interact with electromagnetic radiation, to the first couple of

levels of the electronic excited state. Chromophores are able absorb energy from the incident

UV-Vis range of the electromagnetic spectrum, and this absorption gives the analyte its

apparent colour which can be measured to estimate the concentration of samples.

Estimation of sample concentration is done based on the principles of the Beer-Lambert

given as follows:

(15) = 𝜀 × 𝑐 × 𝑑 = 𝐴 lg 𝐼𝑜 𝐼 = lg 1 𝑇

where Io and I are the initial and the drop in the intensity of light as it passes the sample. T, ε,

c, d and A, are transmission, the molar absorption coefficient, molar concentration, path length

and the samples absorbance that is displayed on the spectrophotometer.

However, absorbance is not always linearly proportional to the concentration of the

chromophores based on the limitations of the spectrophotometer and at high concentration

chromophores might dimerise resulting in different spectroscopic parameters. In samples

exhibiting substantial turbidity, light may not necessarily be absorbed as much as it is scattered

by the sample. The spectrophotometer in these conditions measures the apparent absorbance

(or attenuance) relating to the optical density of the sample and not its concentration. The most

common application of the UV-Vis spectroscopy while probing the phase behaviour of

biopolymers is estimation of changes in polymer concentration based on the assays discussed

in Section 4.3.

1.4.5.2. FTIR spectroscopy

Fourier transform infrared (FTIR) spectroscopy is type of infrared spectroscopy that

uses the mathematical process termed ‘Fourier transform’ to convert raw data into an actual

spectrum. Infrared (IR) analyses compounds using the infrared spectrum that is further divided

into near IR, mid-IR and far IR. Near IR, having the greatest energy penetrates the sample

much deeper than mid or far IR but is therefore the least sensitive. As the energies involved are

relatively low, they do not excite the electrons. However, when samples are subjected to

infrared radiation, they undergo a change in dipole moment leading to vibrational excitation of

groups and atoms that are covalently bonded (Gaca-Zając et al., 2018). Distinct molecules each

have their own unique vibrational energy spectra characterized by their frequency, band shape

and intensity, that serves as the fingerprint for that particular molecule (Gutiérrez Sama et al.,

2018).

In a composite, where the components are in distinct phases, it can be assumed that the

spectral fingerprint of one component isn’t affected by the presence of the other. This would

Table 1.2. Spectroscopic techniques used to probe phase behaviour.

Principle

Advantages

Technique Ultrasonic  Measures the intermolecular

Disadvantages  Not suitable for dilute

 Highly sensitive to

suspensions less than 1% concentration

intermolecular interactions  Can investigate a wide range of

References (Chanamai & McClements, 2001, Chappellaz, Alexander, & Corredig, 2010; Corredig, Sharafbafi, & Kristo, 2011; Clark, 1996)

repulsions caused in aqueous samples due to compression in ultrasonic waves

molecular processes  Non-destructive and non-

invasive

 Can be applied to highly

concentrated, optically opaque systems

Raman

 Can be used for optically unclear

 Due to the inherent weak

(Normand et al., 2000, Pudney et al., 2003, Haug et al., 2003)



samples  In situ analysis possible  Sample size as less as 1 µL or

 Measures wavelength and intensity of inelastic light scattering from molecules along with changes in the polarizability of molecules

1 mg is sufficient

 Enhanced signal-to-noise and



simplification of the resonance Raman spectrum enable time- resolved studies

scattering, not applicable for dilute systems Intense lasers can induce sample lability and photo/thermal decomposition Inherent fluorescence of constituents in sample may obscure the Raman scattering signal

 The variety of functional

groups and their microenvironments present result in overlapping broad spectral bands

FTIR

 Measures the changes in



Identifies functional groups present in the molecule

(Shamsuri, Abdullah, & Daik, 2012; Rafe & Razavi, 2015)

 Sampling chamber has a relatively small size  Some materials absorb

intrinsic dipole moment and bond vibration frequencies in molecules

infrared radiations making the results unreliable

 Mounted pieces can obstruct

the IR beam

Advantages  Fast, non-destructive and

Disadvantages  Needs additional

Table 1.2. Spectroscopic techniques used to probe phase behaviour (continued). Technique Principle Diffusing Wave

reproducible

 Doesn’t necessitate sample

References (Weinbreck, 2004; Blijdenstein, 2003; Gancz, 2005; Liu, 2007)

experimentation to predetermine various parameters

dilution

 Quantifies in-situ dynamics of a colloidal turbid system by measuring the intensity fluctuations of scattered occurring due to Brownian motion of colloidal particles

UV-Vis

 Measures sample

 Necessitates tedious sample

preparation

 Quick and accurate  Easy to use

 Due to high accuracy,

(Jain, Karibasappa, Dodamani, & Mali, 2017; Agbenorhevi & Kontogiorgos, 2010)

concentration based on the absorption of UV-Vis radiation by the sample during electron excitation

readings can be affected by minute contaminants, light or vibration form equipment nearby

 Generates a weaker signal

 Non-destructive technique  Allows examination of materials

Nuclear Magnetic Resonance

compared to other techniques

in the solid state

(Voron’ko, Derkach, Vovk, & Tolstoy, 2016; Dai & Matsukawa, 2013)

 Analyses the environment of hydrogen or carbon nuclei in a compound in an externally applied magnetic field

 Paramagnetic species have a

Electron spin resonance

(Schorsch, Jones, & Norton, 2000; Williams, Clegg, Langdon, Nishinari, & Piculell, 1993)

 Measures the microwave energy absorbed by unpaired electrons in a magnetic field

low chemical lifetime  Some species require low temperatures for detection

 Employs non-ionizing radiation  Preparation steps or extraction not needed prior to analysis  Greater sensitivity than NMR  Offer higher sensitivity  Require low sample volumes  Can analyse dilute samples

result in a composite spectrum that depicts the peaks of both the components in intensities

corresponding to their concentrations. When composites are partially or completely miscible,

the composite spectrum is likely to show formation of new bands, disappearance of some

component bands or shifts in bands indicating shifts in component specific bonds to bonds

between components (Icoz & Kokini, 2007).

In their work, Shamsuri, Abdullah, & Daik (2012) used FTIR to study the phase

behaviour in agar/biopolymer (cellulose, rice starch or zein protein) blend aerogel fabricated

in ionic liquid and co-solvent mixture. The blends of the composites replicated all the chemical

moieties seen in the component spectral scan, indicating that the components did not undergo

any chemical changes and remained in their distinct phases in the composite. In another study

conducted by Rafe & Razavi (2015) on β-lactoglobulin and basil seed gum composite it was

reported that the components were thermodynamically incompatible as the composite spectra

reproduced the individual peaks of the components with a significant increase in intensity with

relation to concentration.

1.4.5.3. Raman spectroscopy

Raman spectroscopy is a branch of vibrational spectroscopy, complementary to that of

IR spectroscopy. Though both IR absorption and Raman scattering induce transitions within

vibrational levels, they produce distinct spectra and cannot be used as alternative techniques.

While IR absorption depends on a change in the intrinsic dipole moment with molecular

vibration, Raman scattering relies on changes in the polarizability of functional groups due to

vibration of the atoms (Li-Chan, 1996). Therefore, non-polar groups such as C=C, S-S and C-

C have dominant Raman bands compared to polar groups (O-H, N-H and C=O) which exhibit

strong IR stretching vibrations.

Water being a polar compound exhibits a strong IR absorption that interferes with the

sample spectra as most food systems are aqueous in nature. Therefore, Raman spectroscopy is

better suited for such systems as it produces less interference. The phase behaviour of

gelatin/maltodextrin composite has been studied using Raman spectroscopy in conjunction

with multivariate curve resolution which enabled determination of individual polymer

concentration in their respective domains (Normand, Pudney, Aymard, & Norton, 2000;

Pudney, Hancewicz, Cunningham, & Gray, 2003). Haug, Williams, Lundin, Smidsrød, &

Draget (2003) probed the phase behaviour of fish gelatin and κ-carrageenan composites using

Raman spectroscopy.

Being a scattering technique, samples of any shape, size or thickness can be analysed,

down to the microscopic range of 10 microns. Samples can be measured in-situ in the solid,

liquid or gaseous state without any sample preparation. The many advantages make Raman

spectroscopy a promising technique in spite of its limited application to materials that

autofluoresce. Sample heating due to radiations can destroy the sample or mask the Raman

spectrum.

1.4.5.4. NMR spectroscopy

Nuclear magnetic resonance (NMR) utilizes nuclear spin states and resonance

spectroscopy for spectroscopic analysis. It is widely used for the structure determination of

biopolymers and their composites. The method utilizes spectral lines of different atomic nuclei

which are excited in a strong magnetic field when a radiofrequency transmitter is applied.

Changes in the 3D structure of molecules can be determined as neighbouring atoms influence

the signals from individual nuclei.

It is a non-destructive technique that allows visualization of atoms and molecules in a

solid as well as in various media in solution. NMR spectroscopy uses the lowest irradiation

energy for excitation within the spectroscopic techniques and hence, relaxation and sensitivity

of NMR spectroscopy are significantly different, enabling observation of molecular dynamics.

In the works of Voron’ko, Derkach, Vovk, & Tolstoy (2016), NMR spectroscopy was used to

study the various bonds involved in the formation of κ-carrageenan-gelatin complexes. The

results elucidated that while electrostatic interactions and hydrogen bonds were at play within

the temperature range of the coil to helix transition, the complexes were stabilised by

hydrophobic interactions at higher temperatures. They further expanded their work to study the

kinetics of gel formation using NMR in aqueous mixtures of κ-carrageenan and gelatin and

documented a decrease in the gelation rate as the κ-carrageenan/gelatin ratio increased. NMR

has also been successfully used to study the gelation mechanism of curdlan (Zhang, Nishinari,

Williams, Foster, & Norton, 2002), konjac mannan (Williams et al., 2000) and agarose (Dai &

Matsukawa, 2013) at the molecular level.

1.4.5.5. Ultrasonic spectroscopy

Ultrasonic spectroscopy studies ultrasonic waves resolved into their Fourier frequency

components. As an analytical technique, it is based on precise measurements of structural (size,

shape, orientation, etc) or material (attenuation, velocity, etc) properties. Ultrasound, by

definition, has a frequency greater than 20 kHz and in an ultrasonic wave, the oscillating

pressure (which is stress in general terms) causes oscillation of compressions (mechanical

deformation), thus acting like a rheological wave. Therefore, simplistically ultrasonic

parameters deal with elasticity and viscosity. However, ultrasound involves high frequency

deformations as opposed to the low frequency deformations (below 1 kHz) involved in classical

rheology.

Ultrasonic spectroscopy deals with elastic properties of materials in contrast to other

spectroscopic techniques that probe the electric and magnetic properties. Compression in

ultrasonic wave alters the intermolecular distances in a sample causing intermolecular

repulsions as a response. Due to the elasticity of water, sound waves travel up to five times

faster in water than in air making it highly sensitive to intermolecular interactions. With the

high sensitivity of the ultrasonic parameters to intermolecular interaction, ultrasonic analysis

of a wide range of molecular processes in aqueous media are made possible.

The measurable parameters of interest are the speed of sound v and attenuation α which, when

assuming a Newtonian medium with minimal thermo-acoustic losses, are directly related to the

complex longitudinal elastic modulus:

(16) M = Mʹ +iMʹʹ

Where Mʹ = ρv2 and Mʹʹ = (2ραv3)/ω and ω is the angular frequency of the wave and ρ is the

material density. The complex longitudinal elastic modulus is related to the bulk K and shear

moduli G by:

(17) M = K + 4G/3

Ultrasound has been successfully used to study biopolymer gels and sol-gel transitions in the

works of Audebrand, Doublier, Durand, & Emery, (1995) and Toubal, Nongaillard,

Radziszewski, Boulenguer, & Langendorff, (2003). It is also gaining popularity for the

investigation of phase behaviour in mixed biopolymer systems such as gum arabic and

modified starch (Chanamai & McClements, 2001), cellulose and amylose gels (Clark, 1996)

and dairy proteins (Chappellaz, Alexander, & Corredig, 2010; Corredig, Sharafbafi, & Kristo,

2011).

1.4.5.6. Diffusing wave spectroscopy

Recently, diffusing wave spectroscopy (DWS) has gained substantial attention as a light

scattering technique used to measure the structural, viscoelastic, and dynamic properties of

turbid colloidal samples. DWS utilises the multiple scattering events that occur when light

passes through a colloidal system to quantify its in-situ dynamics. The diffusion equation can

be used to approximate the propagation of light in such a system. Similar to any light scattering

technique, DWS measures the intensity fluctuations of scattered light occurring due to the

Brownian motion of colloidal particles. By treating the propagation of photons through turbid

samples as a process of diffusion, it makes it possible to extract the dynamics of the scattering

particles from measured correlation functions. DWS is a fast, reproducible, non-destructive

technique that doesn’t necessitate sample dilution unlike the conventional dynamic light

scattering techniques. DWS has been successfully used to probe the phase behaviour of whey

protein and gum arabic (Weinbreck, 2004), β-lactoglobulin and dextran (Blijdenstein, 2003)

and various emulsions (Gancz, 2005; Liu, 2007)

1.4.5.7. Electron spin resonance spectroscopy

Electron spin resonance (ESR) spectroscopy is a technique used for probing materials

with unpaired electrons. Electron spin resonance spectroscopy is similar to nuclear magnetic

resonance with the key difference that in electron spin resonance, the electron spins are excited

due to the absorption of radiation in the microwave radiation instead of spins of atomic nuclei

as in nuclear magnetic resonance. Microwave radiations are absorbed in the presence of an

external magnetic field. Given that electrons have a greater magnetic moment than nuclei, ESR

spectroscopy has a greater sensitivity than NMR spectroscopy. It is also more specific, which

can also be a disadvantage as most biopolymers are not paramagnetic (having unpaired

electrons). Sample volumes as low as 300 µL and concentrations as low as 1 µM can be

analysed. Phase behaviour of micellar casein and κ-carrageenan systems and κ-carrageenan

and konjac mannan mixtures have been investigated using ESR spectroscopy in the past

(Schorsch, Jones, & Norton, 2000; Williams, Clegg, Langdon, Nishinari, & Piculell, 1993)

1.4.6. Diffraction and scattering methods

A classic example of the interaction between radiation and matter is the scattering of

light as it strikes particles. This phenomenon initiated a number of theoretical approaches, out

of which the Rayleigh theory, that pursues the scattering of light by particles substantially

smaller than incident wavelength, has been adapted and most widely applied to the realm of

biopolymer science as the Rayleigh-Gans-Debye (RGD) theory. Most biopolymer mixtures

behave as ‘water-in-water’ emulsions and can be modelled as particles in a solvent. This is

made possible as the biopolymer forming the continuous phase acts like the solvent, and the

dispersed phase, which often forms spherical droplets to reduce interfacial energy acts as the

dispersed particles. As both the phases are water rich, with biopolymers being inherently weak

scatterers, they manifest a relatively small difference in the refractive index (Aymard,

Williams, Clark, & Norton, 2000).

A number of scattering techniques such as small-angle neutron scattering (SANS),

small-angle light scattering (SALS), and X-ray diffraction have been used to characterise phase

behaviour in biopolymer mixtures. The underlying principle of all these techniques is very

similar, with the main difference being the wavelengths of the radiation impacting the

investigation of local structure. As a general thumb rule, short wavelengths of X rays (λ =

0.151 nm) and neutrons (λ = 1.0-2.5 nm) are better than the relatively longer wavelengths of

visible light (λ = 350-800 nm). A combination of the three scattering techniques can provide

an exhaustive analysis of the morphology, but such comprehensive studies are rare in

biopolymer science.

SANS exhibits a strong angular dependence for the scattered intensity, denoting sharp

interfaces between the continuous and dispersed phase (Tuinier, Ten Grotenhuis, Holt,

Timmins, & De Kruif, 1999). Time-resolved SALS can help distinguish between the spinodal

and nucleation growth mechanisms (Tromp, Rennie, & Jones, 1995) while turbidity

measurements too can help detect the onset of phase separation.

1.4.6.1 Light scattering

According to Debye (1959) and Smoluchowski (1912), in a solution, scattered light

intensity arises from fluctuations in concentration and density. For biopolymer solutions,

fluctuations in the density are much smaller than the fluctuations in concentration even in dilute

solutions and can be approximated by the density fluctuations in pure solvents. Therefore, in

solutions where the radius of gyration, Rg < 1/20 λ, the Rayleigh ratio or Rθ (normalized

scattering intensity) is given by:

(18) Rθ = K (Δc2)

Where Rθ = (iθ /Io) r2 and, Io is the initial beam intensity, iθ is the corresponding intensity of the

scattered beam at angle θ, r is the distance between the detector and the centre of the scattering

volume, and K is the constant of optical contrast. Thus, at low concentrations, the intensity of

the scattered light is directly proportional to concentration c, but decreases eventually due to

the repulsive interactions, which lead to a decrease in osmotic compressibility with an increase

in concentration. This enable differentiation between molecular structures at higher

concentrations. Scattering intensity in such systems is given by:

(19) Rθ = KcRT(dc/dπ)

Where RT(dc/dπ) is the osmotic compressibility and π the osmotic pressure (Burchard, 1994).

This principle of light scattering has been used to estimate parameters like interfacial tension

and turbidity, the changes in which can be linked to phase behaviour as discussed in the latter

sections.

Puyvelde et al. (2002) used SALS combined with shear rheometry to measure the

interfacial tension in dextran-gelatin composites. Under shear, the composite yields an

anisotropic SALS pattern, where the degree of anisotropy relates to droplet deformation. The

interfacial tension can be determined from the deformation of the droplet under shear, and from

droplet relaxation after cessation of shear. This technique is also applicable to the theory of

fibril break up set forth by Tomotika (1935), which follows the break-up of fibrils formed under

shear, when shear is halted. Antonov, Puyvelde, & Moldenaers (2004) in their work used SALS

to study the interfacial tension and subsequent phase behaviour of water-sodium caseinate-

sodium alginate systems located at different distances from the binodal. They revealed that

close to the critical point, interfacial tension is much lower than for concentrations farther off

from it.

1.4.6.2 X-ray diffraction

Compared to the huge contribution of X-ray diffraction to conventional structural

chemistry, its application to biopolymers is somewhat confined to molecular biology. The less

direct link between macroscopic behaviour, characteristic of food materials, and detailed

molecular features limit the application of X-rays to biopolymer science.

When a beam of X-ray passes from a material, a small fraction of the radiation is

scattered due to the inhomogeneities in electron density, and the variations in the scattered

intensity and the direction of scatter are used to draw conclusions on the material’s microscopic

structure. A simplifying feature, or a limiting factor of the X-ray technique is that fact that

only electrons present scatter the radiation as opposed to both electrons and neutrons for other

radiations. Even out of the small fraction of X-ray diffraction studies that focus on biopolymer

investigations, a smaller fraction deals with gels and solutions, with amorphous, crystalline

powders and fibre being a popular choice.

At more hydrated conditions, in solutions and gels, systems are less ordered in structure

and hence produce a diffused scattering pattern. Miles et al. (1985) in their work successfully

utilized X-ray diffraction and turbidity analysis to describe the gelation of amylose solutions.

The changes in the scattering intensity was studied as gelation commenced. The rate of

crystallinity development was much slower than the time of experimentation for turbidimetric

and rheometric analysis (1000 min vs 100 and 300 min, respectively). It was concluded that

concentration of amylose into dense light scattering domains, leading to network formation

occurred in the early stages accompanied by an increase in shear modulus. On a much slower

time scale the onset of crystallinity was observed, with insignificant contribution to modulus

increase. Other examples of materials investigated by X-ray diffraction include wheat and corn

starch gels (Xu et al., 2013), amylose-amylopectin mixtures (Leloup, Colonna, & Buleon,

1991), agarose (Foord & Atkins, 1989), and carrageenan/ konjac glucomannan compound gel

(Yuan, Xu, Cui, & Wang, 2019).

1.4.6.3. SAXS and SANS

Investigating properties and characteristics of a food product are best done while

maintaining their native surrounding environment. In the given sense, small angle scattering

possesses attractive attributes as a non-invasive technique, enabling analysis in the solution or

gel state. Due to the inverse relationship between the scattering angle and scale lengths

investigated, small angle scattering makes it possible to investigate proteins, emulsions,

porosity and phase separation, to name a few. At these larger scales, the technique is sensitive

to the arrangement of assemblies of atoms rather than atomic separation in materials enabling

the determination of the molecular structures and spatial skills. Consequently, small angle

scattering (SAS) techniques, either with X-rays (SAXS) or with neutrons (SANS) have been

widely used in biopolymer systems. The theory behind SAS has been covered in detail in

literature (Hammouda, 2010; King, 1999).

A SAS experiment typically measures the intensity of scattered radiation versus the

scattering vector q, which is given by:

(20) q = (4π/λ) sinθ

where λ is the wavelength of the incident radiation, and θ is the scattering angle.

The scattering length in SAXS is determined by the electronic structure of the target

atom. As the number of electrons in an atom is equal to the number of protons, the intensity of

X-ray scattering increases linearly with atomic number. Naturally then, the heavier elements

in the material dominate the X-ray scattering signal. In contrast, in SANS, neutrons are

scattered by the atomic nucleus, which means that scattering can vary significantly between

isotopes.

In their work, Horkay & Hammouda (2008), studied the effect of temperature, nature

of solvent and charge in solutions of poly-lysine and poly-proline. In semi dilute solutions of

polyaspartic acid, chondroitin sulfate hyaluronic acid and DNA, addition of salts result in

structural changes around the polymer chains. Xu et al. (2018) used SAXS and SANS to

characterise the microstructure of the β-lactoglobulin-pectin complex coacervates.

1.4.7. Calorimetry

Even though considerable work was done on the thermodynamic properties of polymer

solutions and blends, the fundamental approach to the thermodynamics of polymeric solutions

appears to have been neglected till the 1990s, with the works of Tseretely and Smirnova (1992)

investigating the melting enthalpy of gelatin being the first of its kind. Flory-Huggins lattice

model (Flory, 1953), along with its generalisation (Doi, Edwards, & Edwards, 1988) had been

the broadly accepted thermodynamic theory used for analysis of experimental data. However,

in case of composite systems, studies seldom split excess Gibbs free energy (parameter χ) to

the constituent enthalpy (χS) and entropy (χH) contributions (Cesàro, Cuppo, Fabri, & Sussich,

1999). This was probably due to the late adoption of calorimeters in the field of biochemistry

making it difficult to accurately determine the enthalpy of biopolymer phase transitions (Wadsö

& Chemistry, 1980). Detailed reviews on the development of calorimeters, their working

principle and their application in biopolymer science has been documented in literature

(Leharne, 2017; Chiu, Prenner, & Sciences, 2011).

Differential scanning calorimetry (DSC) is a physical characterization technique used

to study the purity, thermal behaviour and stability of polymers and polymer blends. First order

transitions like gelation and melting, along with second order transitions like glass transition

are characteristic properties of biopolymers, impacted by the presence of and interaction with

other biopolymers and additives, along with different processing techniques employed

(Gregorova & Microcalorimetry, 2013). Simplistically, a calorimeter is used to measure the

specific heat capacity of thermally induced events in biopolymer systems as a function of

temperature. The apparent specific heat (cp2) of a biopolymer solution is calculated by:

(21) cp2 = cp1 + 1/w2(cp-cp1)

where cp is the solution’s specific heat, cp1 is the specific heat of the solvent, and w2 is the

weight fraction of the solute. Principally, in a calorimeter, the reference and the sample are

maintained at the same temperature, while the changes in the energy during a thermal transition

are recorded by the DSC as a rate dQ/dt against temperature. This sudden change in the value

𝑑𝑄

of specific heat Cp corresponds with the transition temperature as:

𝑑𝑇

(22) = 𝑚𝐶𝑝

Where m is the mass of the sample. First order transitions can be described using the following

𝑑𝑄

formula:

𝑑𝑡

(23) = 𝜅𝛥𝑇 = 𝜅Ṫ(𝑡 − 𝑡0) + 𝑑𝑄 |𝑡0

Where κ is the thermal conductance between the sample holder and the sample, Ṫ is the rate of

increase in temperature and t0 the start of transition.

For the analysis of calorimetric results, the most widely accepted model is the van’t Hoff

equation:

(24) 𝛥𝐻° = 𝑅𝑇2 𝑑 ln 𝐾 𝑑𝑇

Where K is the equilibrium constant, ΔH° is the enthalpy change, R is the ideal gas constant

and T is temperature. This equation is applicable to two state processes which do not have

significant intermediate populations during transition, protein denaturation for example.

Change in entropy, ΔS° and change in standard free energy ΔG° are obtained using the

following equations:

(25) 𝛥𝐺° = −𝑅𝑇 ln 𝐾

(26) 𝛥𝐺° = 𝛥𝐻° − 𝑇𝛥𝑆°

Calorimetry can provide not only the calorimetric enthalpy ΔH (from the area of the

heat absorption), but also the process’ effective enthalpy using the van’t Hoff equation and the

sharpness of the transition. It can hence be used to observe the impact of a second polymer on

the thermal transitions of the first biopolymer and vice versa both qualitatively and

quantitatively.

Fig.1.4. DSC exotherms for a) 2% agarose, b) 15% gelatin, c) 2% agarose plus 15% gelatin, d)

Dalda vanaspati lipid, e) soyabean oil, and f) 2% agarose plus 15% gelatin plus 2.5% Dalda

vanaspasti lipid. Scan rate 1°C/min.

Shrinivas et al. (2009) in their work used the DSC to study the phase behaviour of

binary mixtures of gelatin and agarose, and the ternary mixtures of agarose, gelatin, and a lipid.

Peaks in Figures 1.4 a and 1.4 b depict the coil to helix transition of single agarose and gelatin

solutions, respectively, Fig.1.4 e denotes the crystalline aggregation and lattice formation of

dalda vanaspati, a fat solid at room temperature, as opposed to a featureless baseline heat flow

of soybean oil reflecting that it remains in the liquid state throughout the experimental

temperature range (Fig. 1.4 e). Binary mixtures of agarose and gelatin (Fig. 1.4 c) and ternary

mixtures of agarose/gelatin/lipid (Fig. 1.4 f) maintain the overall peak form of each of the

constituent transition revealing that they remain phase separated without interacting with each

other in a mixture.

1.4.8. Rheology

1.4.8.1. Mechanical properties of the composite

When either one or both of the biopolymers in a biopolymer mixture can form a gel,

the process of phase separation is arrested by network formation. If just one biopolymer can

form a gel, it forms a continuous network in which the second biopolymer remains dispersed.

In a scenario where both the biopolymers are gelling agents, the first component will form a

continuous network, with the solution of the second polymer permeating it. The gelation of the

second polymer may follow in one of three ways - the first component remains as a continuous

network within which the second biopolymer forms a dispersed gel network; in the case the

second biopolymer is a stronger gelling agent, it disrupts the primary network and forms a

continuous network supporting the broken network as a dispersed phase; the third possible

outcome is that the primary network remains intact and the second component forms another

continuous network through the interstices of the primary network leading to a bi-continuous

network (Morris, 2009). Various factors can influence the onset of gelation of one or both of

the polymers and the level of phase separation attained before the structure is trapped by

network formation. This has been discussed in detail in the works of Kasapis (2008) and Norton

& Frith (2001). Biopolymer gelation along with evolution of phase behaviour can thus be used

to create a plethora of phase structures, allowing the manipulation of the composite’s final

properties (Firoozmand & Rousseau, 2015). The phase structure of a given composite acutely

determines its bulk mechanical properties.

The rheological approach of determining phase behaviour correlates the mechanical

properties of a simple binary mixture to the corresponding properties of its constituents by a

set of theoretical blending laws (Frith, 2010). The different parameters that account for the

gel’s mechanical properties and the techniques used for characterising them are outlined in the

works of Picout & Ross-Murphy (2003) and Clark & Ross-Murphy (2009).

Adapted from the realm of polymer science, the Takayanagi blending laws state that in

a biopolymer composite where the individual components X and Y have a low miscibility, the

composite modulus (Gc) can be estimated by considering the extreme cases of strain and stress

distribution within the composite as follows (Takayanagi, 1963):

(27) 𝐺𝑐 = ɸ𝑋𝐺𝑋 + ɸ𝑌𝐺𝑌

and

(28) 𝐺𝑐 = (ɸ𝑋 𝐺𝑋⁄ + ɸ𝑌 𝐺𝑌⁄ )−1

Where GX and GY are the storage modulus and ɸX and ɸY (ɸX + ɸY = 1) are the phase volumes

of the polymers X and Y. Eq. 27 refers to the upper bound or isostrain condition where the

stronger gel forms continuous matrix supporting the weaker dispersed phase and the entire

composite has a uniform strain. Eq. 28 on the other hand applies to isostress conditions where

the weaker matrix supports a more rigid filler and the composite is subjected to the same stress.

Davies in his works (Davies, 1971a, 1971b) developed the following equation that

could be applied to phase separated biopolymer composites where the constituents formed a

bi-continuous gel network:

5⁄

5⁄ = ɸ𝑋(𝐺𝑋)1

5⁄ + ɸ𝑌(𝐺𝑌)1

(29) (𝐺𝑐)1

Since the majority of the biopolymer composites include a large amount of solvent

(water) with a relatively smaller concentration of the biopolymers, it introduces another

complication in accurately determining the phase behaviour of a three component system

(water-polymer X-polymer Y) - variable phase volume (Morris, 1992). Due to the varying

avidity of the two active gelling biopolymers to attract water, the total amount of water in the

system is split between the two in unequal quantities. This influences the values of ɸx and ɸy

needed in Eqs. 27, 28 and 29. The modulus-concentration relationship is also affected as the

effective biopolymer concentration after phase separation is much higher than the initial

concentration. This was overcome by introducing the ‘solvent avidity parameter’, p as follows

(Clark, Richardson, Ross-Murphy, & Stubbs, 1983):

𝑤𝑋 𝑤𝑌 (30) 𝑝 = ( ⁄ 𝑋⁄ ) ( 𝑌⁄ )

Eq. 30 can be used in systems where the weights of the two biopolymers (X and Y) and

the water content of both (w = wX + wY) is known, in order to determine p, the phase volumes

of the respective phases and subsequently their effective concentration in the separated phases.

Numerous studies have documented the successful application of the blending laws in phase

behaviour estimation in biopolymers (Dekkers, Emin, Boom, & van der Goot, 2018; Hassan

Firoozmand & Rousseau, 2013; Mhaske, Condict, Dokouhaki, Katopo, & Kasapis, 2019).

The original blending laws are applicable for composites where one biopolymer formed

a continuous network and the other remained dispersed as spherical inclusions. This is true in

most scenarios as the dispersed phase tends to form isotropic spheroidal fillers due to the drive

to reduce the interfacial tension between the two phases (Ding et al., 2002). However, when

the spherical inclusions aggregate to form fibre or oriented particles, or when the system

comprises of biopolymer fillers whose inherent structures aren’t spherical, the application of

the original blending laws cannot accurately probe the structural properties of these systems.

Kerner (1956a, 1956b), followed by Halpin (1976), were the first to introduce a method that

took into account the shape of the filler as follows:

(31) Gʹ / GʹX = (1 + ABɸY) / (1 - BɸY)

(32) B = [(GʹY/GʹX) – 1]/[(GʹY/GʹX) + A]

Where A is related to Einstein co-efficient kE (A = kE -1), sensitive to phase topology of the

composite and takes into account the shape of the dispersed inclusions and Poisson’s ratio of

the matrix (Kasapis, 2009).

Lewis and Nielsen further refined the blending laws by taking into consideration the

concept of maximum packing fraction of the dispersed phase, ɸm (Lewis & Nielsen, 1970;

Nielsen, 1970):

(33) Gʹc/ GʹX = (1 + ABɸY)/(1-BψɸY)

2]ɸY

(34) Ψ = 1 + [(1 - ɸm)/ɸm

ɸm relates to the shape of the filler particles and its degree of polydispersity (Mhaske, Condict,

Dokouhaki, Farahnaky, & Kasapis, 2020).

When shear is applied to the composite gels, it often permanently orients the filler

particles fibrous filler particles and enhances the network strength. Using the single parameter

exponential equation, the filler orientation distribution can be defined as (Kacir, Narkis, &

Ishai, 1975):

(35) γ = 1 – e-λα

Where λ is the empirical parameter, correlated to material architecture or operational

parameters, α is the orientation angle between 0 and 90 (π/2) degrees and γ is the percent of

filler particles aligned between 0 and ± α degrees (Chen & Cheng, 1996). Using a simplified

version of Eq. 35 Koh & Kasapis (2011) further modified Eq. 33 to take into account the

angular orientation of aligned filler particles in the composite as follows:

(36) Gʹc/ GʹX = [(1 + ABɸY)/(1-BψɸY)]eλ

When λ = 0, Eq. (11) will default to the original Lewis and Nielsen model.

The biopolymer structure-rheology relationship in co-relation with phase behaviour has

been studied in detail in multiple other publications (Chronakis & Kasapis, 1995; Clark &

Ross-Murphy, 1987; Stefan Kasapis, 2008, 2009; S. Kasapis & Bannikova, 2017; Kasapis,

Morris, Norton, & Clark, 1993; E. Morris, 2000; Richardson & Kasapis, 1998). Though the

application of the blending laws is well established, it is as such, an indirect approach of

probing phase behaviour that is time consuming, involving tedious experimentation and

substantial modelling expertise. Furthermore, the blending laws are based on certain

assumptions that may not always hold true for the system for example, a) polymer confinement

to individual phases, b) the modulus is proportional to the square of the polymer concentration

and the volume of polymer network is negligible when compared to volume of the solvent and

c) there is massive increase in effective molecular weight at the onset of gelation.

1.4.8.2. Interfacial tension between phases

For phase separation to take place in aqueous polymer mixtures, the driving force of

phase separation must be greater than the increase in interfacial free energy which is the product

of the total interfacial area and interfacial tension between the two phases (Antonov et al.,

2004). This makes interfacial tension between the coexisting phases of a phase separated

system a property of interest.

The onset of phase separation is marked by the formation of droplet structures in the

spatial distribution of the two polymers. In the absence of external dynamic effects such as

shear or applied flow field, the phase with the greater volume fraction forms the continuous

phase. Since the microstructure formation is spherical, it is evident that interfacial tension is at

play. It is worth noting however that the interfacial tension is extremely low, typically in the

range of 0.01-10 µN/m (Vis et al., 2015). In addition, the liquid-liquid interface is permeable

to solvent and small ions and the interplay between phase separation of the charged polymers

leads to an electrical potential difference between the two phases which further reduces the

interfacial tension. The effect of biopolymer concentration and molecular weight of the

biopolymers on the interfacial tension has been investigated by a number of researchers (Ding

et al., 2002; Guido, Simeone, & Alfani, 2002; Spyropoulos et al., 2008; Puyvelde et al., 2002).

Bamberger, Seaman, Sharp, & Brooks (1984) reported that interfacial tension could be

correlated either with the length of the tie line (TLL) in the phase diagram:

(37) 𝜎 = 𝐶1 . (𝑇𝐿𝐿)𝐶2

Or with the difference in the polymer concentrations in the separated phases:

𝐶2

(38) 𝜎 = 𝐶1 . (𝛥𝑋𝑝𝑜𝑙𝑦𝑚𝑒𝑟)

Where C1, C2 are experimental constants related to the biopolymer, temperature, concentration,

and molecular weight.

A number of techniques can be used to measure the interfacial tension and are classified

into three broad groups (Ding et al., 2002): measurement of interfacial tension from the balance

between surface forces and gravitational forces; measuring the equilibrium size and shape of

the droplets; dynamic measurements of deformation and breakage of thin threads of one phase

(Scholten, Tuinier, Tromp, & Lekkerkerker, 2002), steady-state shape of droplets in a

controlled flow field (Sigillo, Di Santo, Guido, & Grizzuti, 1997) or retraction of deformed

drop after cessation of flow (Guido & Villone, 1999).

Selection of the measurement technique depends on the physical properties of both the

phases, mainly their viscosity and density. Since the density of the phases is quite similar and

at least one or both the phases are highly viscous using the techniques of the first two categories

is either impossible or highly unreliable. Therefore, techniques of dynamic deformation are

most widely used to estimate interfacial tension (Stokes, Wolf, & Frith, 2001). The interfacial

tension can be directly correlated to the complex modulus of the composite using Palierne’s

model as follows (Palierne, 1990, 1991):

∗ (1+3∑𝑖ɸ𝑖𝐻𝑖 1−2∑𝑖ɸ𝑖𝐻𝑖

(39) ) 𝐺 ∗ = 𝐺𝐶

Where Hi = fn(α/Ri, Gc*, Gd*). Here Hi is a function of frequency (ω), given for each particle

size Ri. Gc* and Gd* are the complex modulus of the continuous and the filler phase,

respectively. ɸi is the phase volume for a monodispersed emulsion and α is the interfacial

tension between the two phases.

Alternatively, steady state deformation of droplets as result of applied flow field can be

easily investigated in a mixture at very low shear rates. The relationship between the steady

5(19𝑝+16)

state deformation of the droplet in a shear flow is given by (Cox, 1969):

4(1+𝑝)√(19𝑝)2+ (20/𝐶𝑎)2

(40) 𝐷 =

Where p is the ratio of the viscosity of the internal phase to that of the external phase and Ca

is the capillary number given by (Tucker III & Moldenaers, 2002):

(41) 𝐶𝑎 = 𝜂𝑐𝑅𝛾 𝜎

Where ηc is the viscosity of the external phase, Rγ is the radius of the droplet under shear and σ

is the interfacial tension.

Fig.1.5. A drop of caseinate in alginate rich phase at rest (1), start-up transient of deformation under

steady shear (2,3), steady state shape (4), and relaxation after shear is stopped (5,6). Graph represents

the deformation parameter as a function of shear rate in comparison with theory (Guido et al. 2002)

Fig.1.5 shows the results of such an experiment extracted from the works of Guido et

al. (2002) and demonstrates that low shear rates are capable of producing significant

deformations in a caseinate-alginate mixture. This method is combined with microscopy or

small angle light scattering (SALS) to visualise/characterise droplet deformation and

reformation and have been explained in detail elsewhere in the review. The role of interfacial

tension in the phase behaviour of mixed biopolymer systems has been explained in detail in

the works of Frith, (2010).

1.4.9. Microscopy

The different microscopy techniques used to probe behaviour in biopolymer composites

have been summarised in Table 1.3 and discussed in detail below:

1.4.9.1. Optical microscopy

The late nineteenth century saw a rapid advancement in the field of microscopy

enabling researchers to probe specimens from a multitude of disciplines in the micron and

submicron level. Optical microscopy being the simplest form of microscopy was the most

accessible and hence most popular. Early microscopists were limited by poor lens design,

blurred images and optical aberration which have been successfully overcome with

technological advancements. In its simplest form, an optical microscope consists of an

objective lens placed at an adjustable distance from the object, known as the focal length. Light

rays reflecting from the object, pass through the objective to form an inverted aerial

intermediate image at a fixed distance within the optical tube. This is further magnified by the

microscope eyepiece that produces an erect image. The total magnification of the microscope

is equal to the magnification of the objective times that of the eyepiece (Mertz, 2019).

While imaging a mixed biopolymer system, light from the microscope lamp reaches

the sample via a condenser. Some of this light passes through the sample undeviated, whilst

some of it gets diffracted due to the differences in density and refractive indices of the

constituent biopolymers within the sample. The diffracted light is rendered 180 degrees or half

a wavelength out of phase with the undeviated light and causes destructive interference with

the direct light at the intermediate image plane. The eye piece lens further magnifies the image

which is finally projected onto the camera film, light-sensitive computer chip or the retina. The

objective projects the undeviated light, spreading it evenly throughout the image plane at the

eyepiece’s diaphragm. The diffracted light is brought to focus at localized places on the same

image plane where it causes a destructive interference reducing intensity resulting in lighter or

darker areas. These lighter or darker patterns are what form the image of the sample (Davidson,

Abramowitz, 2002). Optical microscopy is a quick, easy technique that does not require much

sample preparation. However, it is suitable only for dark imaging of samples that have a

significant density difference.

Freitas, Albano, & Telis (2017) in their work used the optical microscope to study the

complexing of soy protein isolate (SPI) and high methoxyl pectin (HM) at varying pH and at

different proportions. Capron, Costeux, & Djabourov (2001) studied the phase separated

alginate and sodium caseinate water in water emulsion using the optical microscope.

1.4.9.2. Phase contrast microscopy

Phase contrast microscopy is based on the advances in optical principles by Frits

Zernike introduced before the World War II, but the technique did not become universally

recognised till the 1950s. Frits Zernike’s study revealed that amplitude and phase differences

between deviated light and zeroth order could be altered to enhance contrast and interference

(Zernike, 1942a, 1942b). As mentioned in the previous section, refracted waves that are ½

wavelength or 180 degrees out of phase exhibit a good contrast in an optical image, but samples

with a lower difference in refractive indices cause a lag of just ¼ wavelength, which isn’t

enough to decrease its intensity. This creates an image without contrast, making the details

obscured. Zernike proposed a method to speed up the undeviated light by ¼ wavelength to

induce a difference of ½ wavelength between the undeviated and deviated light. This would

then subsequently lead to the formation of a dark image against a light background as described

earlier (Pluta & Maksymilian, 1988). Another less popular option is to slow down the direct

light by ¼ wavelength, which results in both the diffracted and direct light reaching the

eyepiece in step causing a constructive interference. This creates a bright image on a dark

background (Abramowitz, 1987).

Table 1.3: Microscopic techniques used for structural characterisation of phase behaviour.

Principle

Advantages

Disadvantages

Type of Microscope

Radiation type

Photons

Optical

 Simplest of the microscopes to

 Have a lower

use

magnification compared to other microscopes

 Light reflected/ transmitted through the sample passes through a number of glass lenses that magnify the object

 Uses visible light  Affordable



 Lower resolution compared to other microscopes Internal structures can’t be viewed unless samples are fixed

 Cannot operate in

darkness

Electrons

Electron

 Has a high magnification and

resolution

 Samples need to be under vacuum in most cases  Most samples need to be

 Quick  Capable of generating three-

coated with a non- conducting material

 The sample is scanned using a beam of electron. The energy lost by the electrons as it interacts with the specimen is used as a signal to form an image.

dimensional and topographical images

 Excess charge can be carried over during imaging and affect the image quality  Cannot image live

samples

 Only generates black and

white images

Photons

Phase contrast

 Can observe samples in-situ in

 Not suitable for thick

natural state

 Differences in refractive indices used to provide contrast in constituents



samples Images usually have a halo effect

 No sample preparation required  High contrast, high resolution

images obtained

Laser

Confocal

 Allows identification of multiple

components as long as the dyes used to stain them have different excitation and emission wavelengths

 Depth of view limited to a few hundred microns  Smaller field of view compared to other microscopic techniques

 High resolution images can be

 Staining required

 Laser is used to fluoresce only one plane of sample; emitted light is captured through a pinhole to block out-of-focus photons; stepping feature allows to change focal plane along the Z axis

obtained rapidly

 Non-invasive 3D sectioning made

possible

Laser

Fluorescence



 Capable of generating high

 Samples need to be

Instead of transmitted/ reflected light, natural or induced fluorescence is captured

stained in order to be observed

contrast, high resolution images  Capable of using and visualising

multiple fluorophores simultaneously

 Not best suited for thick samples and cross- sections

Atomic force

N/A (Physical cantilever used)

 Being a new technique, it needs to be standardised for widespread application

 A cantilever is scanned across the sample surface, changes in the position of the cantilever induced by surface morphology is recorded

 Possibility of damaging

 High scan speed  High resolution  Minimal or no sample preparation needed  Can be paired with other

soft samples

techniques to give a holistic picture

 Can be used at extreme

temperatures

 Can be operated in air or in a

liquid

To speed up the undeviated zeroth order of light, a phase plate with a ring-shaped phase

shifter is installed to objective’s rear focal plane. The phase plate is optically much thicker than

the ring’s narrow area. Consequently, while traversing the objective, direct light passing

through the phase ring travels a shorter distance than the diffracted light. This causes the

deviated and direct light to be ½ wavelength out of phase when they reach the image plane,

where they interfere destructively and creating a dark image on a light background. To slow

down the speed of direct light, the phase shifter area of the plate is made optically thicker than

the plate. In this scenarios, direct light reaches the image plane with the diffracted light

inducing constructive interference, creating a bright image on a darker background (Delly,

1988).

Fig.1.6: Kappa carrageenan (A, B) and Iota carrageenan (C, D) in mixture with meat proteins

at pH 5.6 and 7.1, respectively. Black arrows indicate carrageenan molecules with attached

meat proteins.

Farouk, Frost, Krsinic & Wu (2011) used phase contrast microscopy to study the phase

behaviour of kappa and iota carrageenan in mixture with meat protein (Fig.1.6). Phase contract

microscopy has also been successfully used in the works of Bourriot, Garnier & Doublier

(1999) to study the phase behaviour of κ-carrageenan- micellar casein mixtures using different

ionic forms of carrageenan (Na+, K+).

1.4.9.3. Atomic force microscopy

Also known as scanning-force microscopy, atomic force microscopy (AFM) is a type

of scanning probe microscope technique. Originally designed to characterise surfaces at the

atomic scale, AFM allows imaging of solutions and soft composites at molecular scale with a

resolution of <1 nm.

Disregarding the conventional notions of microscopy, AFM collects data for images by

‘feeling’ the samples, making it a physical method of imaging (Morris, Kirby, & Gunning,

1999). A cantilever probe is run over the sample surface and the changes in the magnitude of

interaction between the sample surface and the probe (mainly van der Waal’s force) is

measured. A laser beam is focused directly over the tip of the cantilever and reflected on to a

mirror. The mirror further reflects the laser onto a position positive photodiode detector.

As the sample is scanned, the cantilever tip moves in accordance with the surface

topography. Consequently, the angle of the laser reflected from the cantilever changes

according to Hooke’s law, resulting in changes in the position of the laser spot falling on the

photodiode. As a result, the intensity in each quadrant of the photodiode changes producing an

electric signal that quantifies the normal motion of the tip. The cantilever deflections are thus

measured to produce a map of the surface topography on the computer.

Kim et al. (2004) successfully employed AFM to probe the morphology of protein-

DNA oligonucleotide complexes. Whereas Ikeda et al. (2004b) studied the concentration

dependence of gel formation in mixed gellan-xyloglucan systems using AFM revealing that at

0.05% gellan and 0.7% xyloglucan, the constituents exhibited a synergistic interaction forming

a single-phase gel. Whilst single systems of gellan and xyloglucan did not gel at concentrations

lower than 0.5 and 0.75%, respectively. A detailed account of the applications of AFM in food

and polymer science can be found elsewhere in literature (Nguyen-Tri, Ghassemin, Carriere,

Assadi, & Nguyen, 2020; H. Yang et al., 2007).

Given the high scan speeds and resolution of AFM, it is helping unveil information that

was thus far inaccessible to researchers. Samples probed using AFM can be studied in their

native state without any processing. AFM can also be modified to study aqueous systems in-

situ. Since it is a relatively new technique, it needs to be standardised to find widespread

application. In addition, as the probe comes in direct contact with the samples, there is a

possibility of it damaging the soft samples and gels. Lastly, since AFM is a physical technique,

it doesn’t help in the chemical characterisation of the samples but recent developments in the

technique like infrared-AFM are looking at overcoming these drawbacks (Magonov &

Instrumentation, 2006).

1.4.9.4. Fluorescent microscopy

Epifluorescence or fluorescence microscopy exploits the capability of fluorochromes

to emit light when excited by light of a specific wavelength and is one of the fastest growing

areas of investigation in microscopy (Winter & Shroff, 2014). Some materials auto-fluoresce

or fluorochromes can be externally added to samples. The fluorescent probe molecules spread

over the sample according to its affinity with the different constituents in the sample and local

accessibility and attach to a biopolymer non-covalently or covalently (Patonay, Salon, Sowell,

& Strekowski, 2004). At a particular wavelength of light, fluorochromes, being rich in

chromophores absorb photons and emit them at a longer wavelength. Multiple dyes can be used

to stain different components as long as the emission wavelength of the dyes is different.

Illumination in normal optical microscopy is attained by widefield illumination where

the entire sample, and not just the focal plane is simultaneously and uniformly illuminated.

This necessitates that samples are thin and relatively transparent and often causes an out-of-

focus blur that decreases contrast and resolution of the image (Auty, 2013). To overcome these

drawbacks, the confocal microscope was developed. The essential feature of confocal

microscopy is that both illumination and detection systems are focused on a single point in the

sample. This is brought about by focusing the illuminating light or laser on a specific point in

the sample and the emitted light is detected by means of a small pinhole (aperture) in front of

the detector. The pinhole helps in reducing the out of focus blur giving a sharp image. The

detector amplifies the signal and converts the photon electron intensity into pixels (Davidson

et al., 2002). Confocal microscopy has been extensively used to study food microstructure at

the meso scale and the applications can be found summarised in the works of Cardona, Iriart,

& Herrera (2013) and Auty (2013). Confocal microscopy has major advantages over other

forms of microscopy including 3D imaging of bulk samples is possible, minimal sample

preparation required as samples can be imaged in situ, has a higher resolution and more

sensitive detection as compared to conventional microscopy and it is possible to image dynamic

processes. Disadvantages of confocal microscopy are limited to the fact that staining generally

involves solvents that may affect the sample and/or introduce artefacts to the sample. As of

now diffraction limited lateral resolution is around 200nm and prolonged exposure to

fluorescent light may induce photobleaching and loss of fluorescence (Jonkman & Brown,

2015).

1.4.9.5. Electron microscopy

Electron microscopes comprise an electron emitter in a vacuum chamber. Electron

beams have a substantially shorter wavelength compared to light radiation and hence give a

much higher resolution. Electromagnets are used to focus the electron beam on the sample and

an image is produced by striking the surfacing of a sample and capturing the emitted electrons

as in scanning electron microscopy (SEM) or by passing the electron beam through a thin

section of sample as in transmission electron microscopy (Ikeda et al. 2004). SEM is more

common in the study of the morphology of biopolymer gels.

For TEM thin sections of the sample need to be prepared either as a metallic replica or

as negatively stained dispersions through which a narrow beam of electrons is passed. Sample

preparation is tedious involving chemical fixation, solvent dehydration and embedding in a

suitable resin. Glass or diamond knives are used to section ultrathin samples of ~90-150 nm

thickness using an ultramicrotome. Sections are further stained to increase contrast and carbon

coated to prevent charging.

On the other hand, SEM provides topographic details with a high depth of field.

Conventional SEM is suitable for studying surface morphology of low moisture samples like

freeze dried gels, dairy powder or starches. Cryo-SEM and environmental SEM (ESEM) are

advancements in SEM which allow a sample to be rapidly frozen and fractured under vacuum

and directly imaged under the electron beam without any sample preparation for successful

visualisation of internal structures, whilst ESEM allows samples to be imaged under low

vacuum mode.

A major challenge to EM techniques is minimising damage to non-conductive

specimens caused by the electron beam. Samples usually need to be coated with a conductive

layer of carbon or heavy metal and may mask minute morphological details.

1.5. Significance of the Research

Phase separation in protein and polysaccharide gels remains one of the basic tools of

obtaining the required structural properties and textural profile in food product formulations.

In phase separated biopolymer gels, the overall structural properties and techno-functionality

of the mixture depend on the volume and composition of each phase, hence it is critical to

determine the water (solvent) partition between two hydrocolloid phases. A number of

techniques have been used in the past with varying popularity to probe the driving kinetics,

mechanism and extent of phase separation. These, however, have numerous drawbacks as listed

above and fail to provide a holistic understanding of phase behaviour in biopolymer

composites. There is a need for a simplistic and direct approach to probe phase behaviour in

biopolymer composites. Therefore, the significance of this PhD study is summarised as

follows:

i. It is established that phase behaviour in biopolymer gels affects the overall structural

and textural properties of the composite. Therefore, it is critical to accurately determine

the water partition between the two hydrocolloid phases, along with the volume and

composition of each phase.

ii. Although numerous techniques have attempted to probe phase behaviour in biomaterial

composites, these provide an incomplete, tunnel visioned understanding of phase

behaviour.

iii. Of the many techniques, application of rheology based theoretical blending laws is most

popular, and though robust, these are an indirect method of estimating phase behaviour

which based on numerous assumptions and require tedious experimentation and

modelling expertise.

iv. Confocal scanning laser microscopy is used for qualitative analysis in biomedical

science, biosciences and recently in food science but not for quantitative analysis.

v. This research, therefore, aims at developing a new approach for the estimation of the

phase volume in biopolymer mixtures using a Z-stack function in confocal microscopy.

1.6. Research Questions

i. Can rheological data be obtained on the small deformation properties of the model mixed

biopolymer gels?

ii. Can blending law analysis be applied to the rheological data of the model composite

systems?

iii. Can the Z-stack generate three-dimensional images of the biopolymer composites?

iv. Can the three-dimensional images be quantified using the image processing software?

v. Can theoretical predictions of phase behaviour be correlated to tangible evidence from

microscopic observations/analysis?

1.7. Research Objectives

To address the research questions outlined in the previous section, the following research

objectives are developed:

i. Designing model biopolymer composites that are suitable for both rheological as well as

microscopic analysis.

ii. Obtaining rheological data on the small deformation properties of model biopolymer

composites and applying blending law analysis to it.

iii. Developing a method to capture three-dimensional images of the composite gels using

the Z-stack feature of CLSM.

iv. Developing a novel technique of quantifying three-dimensional images to estimate phase

volume using two image analysis software- FIJI and Imaris.

v. Testing the efficacy of the established protocol in probing biopolymer composites of

varying complexity.

vi. Comparing the two image analysis software for accuracy and customizability to ensure

wide-spread applicability of the protocol.

References

Abramowitz, M. (1987). Contrast Methods in Microscopy: Transmitted Light: Olympus

Corporation, Precision Instrument Division.

Affdl, J. H., & Kardos, J. (1976). The Halpin‐Tsai equations: a review. Polymer Engineering

& Science, 16(5), 344-352.

Agbenorhevi, J. K., & Kontogiorgos, V. (2010). Polysaccharide determination in

protein/polysaccharide mixtures for phase-diagram construction. Carbohydrate

polymers, 81(4), 849-854.

Anderson, V., & Jones, R. J. P. (2001). The influence of gelation on the mechanism of phase

separation of a biopolymer mixture. 42(23), 9601-9610.

Antonov, Y. A., & Gonçalves, M. (2018). Phase equilibria and the mechanical properties of

gel-like water-gelatin-locust bean gum systems. Journal of Characterization and

Development of Novel Materials, 10(2), 77-87.

Antonov, Y. A., Van Puyvelde, P., & Moldenaers, P. (2004). Interfacial tension of aqueous

biopolymer mixtures close to the critical point. International Journal of Biological

Macromolecules, 34(1), 29-35. doi:https://doi.org/10.1016/j.ijbiomac.2004.01.001

Audebrand, M., Doublier, J., Durand, D., & Emery, J. J. F. h. (1995). Investigation of gelation

phenomena of some polysaccharides by ultrasonic spectroscopy. 9(3), 195-203.

Auty, M. (2013). Confocal microscopy: principles and applications to food microstructures. In

Food microstructures (96-P98): Elsevier.

Aymard, P., Williams, M., Clark, A., & Norton, I. J. L. (2000). A turbidimetric study of phase

separating biopolymer mixtures during thermal ramping. 16(19), 7383-7391.

Bamberger, S., Seaman, G. V., Sharp, K., & Brooks, D. E. (1984). The effects of salts on the

interfacial tension of aqueous dextran poly (ethylene glycol) phase systems. Journal of

Colloid and Interface Science, 99(1), 194-200.

Beijerinck, M. J. Z. f. B., Parasiten und Infektionskrankheiten. (1896). Über eine

Eigentümlichkeit der löslichen Stärke. 2, 679-699.

Bergfeldt, K., Piculell, L., & Linse, P. (1996). Segregation and association in mixed polymer

solutions from Flory− Huggins model calculations. The Journal of Physical Chemistry,

100(9), 3680-3687.

Bourriot, S., Garnier, C., & Doublier, J. (1999). Phase separation, rheology and microstructure

of micellar casein–guar gum mixtures. Food Hydrocolloids, 13(1), 43-49.

Bourriot, S., Garnier, C., & Doublier, J. L. (1999). Micellar-casein–κ-carrageenan mixtures. I.

Phase separation and ultrastructure. Carbohydrate polymers, 40(2), 145-157.

Bradford, M. M. (1976). A rapid and sensitive method for the quantitation of microgram

quantities of protein utilizing the principle of protein-dye binding. Analytical

biochemistry, 72(1), 248-254. doi:https://doi.org/10.1016/0003-2697(76)90527-3

Bradstreet, R. B. (1954). Kjeldahl method for organic nitrogen. Analytical chemistry, 26(1),

185-187.

Burchard, W. (1994). Light scattering techniques. In Physical techniques for the study of food

biopolymers (151-213): Springer.

Capron, I., Costeux, S., & Djabourov, M. (2001). Water in water emulsions: phase separation

and rheology of biopolymer solutions. Rheologica Acta, 40(5), 441-456.

Cardona, J. A. R., Iriart, C. H., & Herrera, M. L. (2013). Applications of confocal laser

scanning microscopy (CLSM) in foods. In Confocal Laser Microsc.-Princ. Appl. Med.

Biol. Food Sci. (203-234).

Cesàro, A., Cuppo, F., Fabri, D., & Sussich, F. (1999). Thermodynamic behavior of mixed

biopolymers in solution and in gel phase. Thermochimica Acta, 328(1), 143-153.

Chanamai, R., & McClements, D. J. (2001). Depletion flocculation of beverage emulsions by

gum arabic and modified starch. Journal of food science, 66(3), 457-463.

Chappellaz, A., Alexander, M., & Corredig, M. (2010). Phase separation behavior of caseins

in milk containing flaxseed gum and κ-carrageenan: a light-scattering and ultrasonic

spectroscopy study. Food Biophysics, 5(2), 138-147.

Chee, K. K. (1990). Determination of polymer-polymer miscibility by viscometry. European

Polymer Journal, 26(4), 423-426.

Chen, C.-H., & Cheng, C.-H. (1996). Effective elastic moduli of misoriented short-fiber

composites. International Journal of Solids and Structures, 33(17), 2519-2539.

Chiu, M. H., & Prenner, E. J. (2011). Differential scanning calorimetry: An invaluable tool for

a detailed thermodynamic characterization of macromolecules and their

interactions. Journal of Pharmacy and Bioallied Sciences, 3(1), 39.

Chronakis, I. S., & Kasapis, S. (1995). Food applications of biopolymer—theory and practice.

In G. Charalambous (Ed.), Developments in Food Science 37(75-109): Elsevier.

Clark, A., Richardson, R., Ross-Murphy, S., & Stubbs, J. (1983). Structural and mechanical

properties of agar/gelatin co-gels. Small-deformation studies. Macromolecules, 16(8),

1367-1374.

Clark, A. H. (1996). Biopolymer gels. Current Opinion in Colloid & Interface Science, 1(6),

712-717.

Clark, A. H., & Ross-Murphy, S. B. (1987). Structural and mechanical properties of

biopolymer gels. In Biopolymers (57-192): Springer.

Clark, A. H., & Ross-Murphy, S. B. (2009). Biopolymer network assembly: measurement and

theory. In Modern biopolymer science (1-27): Elsevier.

Conti, D. S., Yoshida, M. I., Pezzin, S. H., & Coelho, L. A (n.d.). On the miscibility of

biopolymers and their interactions. Paper presented at the 2nd International

Symposium on Calorimetry and Chemical Thermodynamics P.

Copeland, R. A. (1994). Methods for protein analysis: Springer.

Corredig, M., Sharafbafi, N., & Kristo, E. J. F. H. (2011). Polysaccharide–protein interactions

in dairy matrices, control and design of structures. 25(8), 1833-1841.

Cox, R. (1969). The deformation of a drop in a general time-dependent fluid flow. Journal of

fluid mechanics, 37(3), 601-623.

Cragg, L., & Bigelow, C. (1955). The viscosity slope constant k′—ternary systems: Polymer–

polymer–solvent. Journal of Polymer Science, 16(82), 177-191.

Dai, B., & Matsukawa, S. J. C. r. (2013). Elucidation of gelation mechanism and molecular

interactions of agarose in solution by 1H NMR. 365, 38-45.

Danait, A., & Deshpande, D. D. (1995). A novel method for determination of polymer-polymer

miscibility by viscometry. European Polymer Journal, 31(12), 1221-1225.

Davidson, M. W., & Abramowitz, M. (2002). Optical microscopy. Encyclopedia of imaging

science and technology.

Davies, W. (1971a). The elastic constants of a two-phase composite material. Journal of

Physics D: Applied Physics, 4(8), 1176.

Davies, W. (1971b). The theory of composite dielectrics. Journal of Physics D: Applied

Physics, 4(2), 318.

de Kruif, C. G., & Tuinier, R. (2001). Polysaccharide protein interactions. Food Hydrocolloids,

15(4), 555-563.

Debye, P. (1959). Angular dissymmetry of the critical opalescence in liquid mixtures. The

Journal of Chemical Physics, 31(3), 680-687.

Dekkers, B. L., Emin, M. A., Boom, R. M., & van der Goot, A. J. (2018). The phase properties

of soy protein and wheat gluten in a blend for fibrous structure formation. Food

Hydrocolloids, 79, 273-281.

Delly, J. G. (1988). Photography Through The Microscope. 9e. Eastman Kodak Co.,

Rochester, New York.

Ding, P., Wolf, B., Frith, W., Clark, A., Norton, I., & Pacek, A. (2002). Interfacial tension in

phase-separated gelatin/dextran aqueous mixtures. Journal of Colloid and Interface

Science, 253(2), 367-376.

Ding, P., Wolf, B., Frith, W. J., Clark, A. H., Norton, I. T., & Pacek, A. W. (2002). Interfacial

Tension in Phase-Separated Gelatin/Dextran Aqueous Mixtures. Journal of Colloid and

Interface Science, 253(2), 367-376.

Doi, M., Edwards, S. F., & Edwards, S. F. (1988). The theory of polymer dynamics (Vol. 73):

Oxford University Press.

Donato, L., Garnier, C., Novales, B., & Doublier, J. L. (2005). Gelation of globular protein in

presence of low methoxyl pectin: effect of Na+ and/or Ca2+ ions on rheology and

microstructure of the systems. Food Hydrocolloids, 19(3), 549-556.

Dubois, M., Gilles, K. A., Hamilton, J. K., Rebers, P. t., & Smith, F. (1956). Colorimetric

method for determination of sugars and related substances. Analytical chemistry, 28(3),

350-356.

Ernst, O., & Zor, T. (2010). Linearization of the Bradford protein assay. Journal of Visualized

Experiments, (38), e1918.

Farouk, M., Frost, D., Krsinic, G., & Wu, G. (2011). Phase behaviour, rheology and

microstructure of mixture of meat proteins and kappa and iota carrageenans. Food

Hydrocolloids, 25(6), 1627-1636.

Farouk, M. M., Frost, D. A., Krsinic, G., & Wu, G. (2011). Phase behaviour, rheology and

microstructure of mixture of meat proteins and kappa and iota carrageenans. Food

Hydrocolloids, 25(6), 1627-1636.

Firoozmand, H., & Rousseau, D. (2013). Microstructure and elastic modulus of phase-

separated gelatin–starch hydrogels containing dispersed oil droplets. Food

Hydrocolloids, 30(1), 333-342.

Firoozmand, H., & Rousseau, D. (2015). Controlled phase separation for texture modification.

In Modifying Food Texture (157-181): Elsevier.

Flory, P. J. (1953). Principles of polymer chemistry: Cornell University Press.

Foord, S., & Atkins, E. J. B. O. R. o. B. (1989). New x‐ray diffraction results from agarose:

Extended single helix structures and implications for gelation mechanism. 28(8), 1345-

1365.

Foreman, J., Sauerbrunn, S., & Marcozzi, C. (2013). ‘Thermal Analysis & Rheology. Kyoritsu

Chem. & Co., Ltd, Kisarazu, Japan.

Fredrickson, G. H., Liu, A. J., & Bates, F. S. (1994). Entropic corrections to the Flory-Huggins

theory of polymer blends: Architectural and conformational effects. Macromolecules,

27(9), 2503-2511.

Freitas, M. L. F., Albano, K. M., & Telis, V. R. N. (2017). Characterization of biopolymers

and soy protein isolate-high-methoxyl pectin complex. Polímeros, 27(1), 62-67.

Frith, W. J. (2010). Mixed biopolymer aqueous solutions–phase behaviour and rheology.

Advances in Colloid and Interface Science, 161(1-2), 48-60.

Gaca-Zając, K. Z., Smith, B. R., Nordon, A., Fletcher, A. J., Johnston, K., & Sefcik, J. (2018).

Investigation of IR and Raman spectra of species present in formaldehyde-water-

methanol systems. Vibrational Spectroscopy, 97, 44-54.

Garcı́a, R., Melad, O., Gómez, C. M., Figueruelo, J. E., & Campos, A. (1999). Viscometric

study on the compatibility of polymer–polymer mixtures in solution. European

Polymer Journal, 35(1), 47-55. doi:https://doi.org/10.1016/S0014-3057(98)00106-2

George, J. W. (2001). The usefulness and limitations of hand‐held refractometers in veterinary

laboratory medicine: an historical and technical review. Veterinary Clinical Pathology,

30(4), 201-210.

Gilbert, G., & Spragg, S. (1964). Iodine sorption:" blue value. Whistler, RL et al. Methods in

carbohydrate chemistry, 4, 168-169.

Goh, K. K., Teo, A., Sarkar, A., & Singh, H. (2020). Milk protein-polysaccharide interactions.

In Milk proteins (499-535): Elsevier.

González-González, M., Mayolo-Deloisa, K., Rito-Palomares, M., & Winkler, R. (2011).

Colorimetric protein quantification in aqueous two-phase systems. Process

Biochemistry, 46(1), 413-417.

Gornall, A. G., Bardawill, C. J., & David, M. M. (1949). Determination of serum proteins by

means of the biuret reaction. Journal of biological chemistry, 177(2), 751-766.

Gregorova, A. (2013). Application of differential scanning calorimetry to the characterization

of biopolymers. Applications of Calorimetry in a Wide Context-Differential Scanning

Calorimetry, Isothermal Titration Calorimetry and Microcalorimetry, 3-20.

Grinberg, V. Y., & Tolstoguzov, V. (1972). Thermodynamic compatibility of gelatin with some

D-glucans in aqueous media. Carbohydrate Research, 25(2), 313-321.

Grinberg, V. Y., & Tolstoguzov, V. (1997). Thermodynamic incompatibility of proteins and

polysaccharides in solutions. Food Hydrocolloids, 11(2), 145-158.

Guido, S., Simeone, M., & Alfani, A. (2002). Interfacial tension of aqueous mixtures of Na-

caseinate and Na-alginate by drop deformation in shear flow. Carbohydrate polymers,

48(2), 143-152.

Guido, S., & Villone, M. (1999). Measurement of interfacial tension by drop retraction

analysis. Journal of Colloid and Interface Science, 209(1), 247-250.

Gupta, S., Wang, W. S., & Vanapalli, S. A. (2016). Microfluidic viscometers for shear rheology

of complex fluids and biofluids. Biomicrofluidics, 10(4), 043402.

Gutiérrez Sama, S., Farenc, M., Barrère-Mangote, C., Lobinski, R., Afonso, C., Bouyssière,

B., & Giusti, P. (2018). Molecular Fingerprints and Speciation of Crude Oils and Heavy

Fractions Revealed by Molecular and Elemental Mass Spectrometry: Keystone

between Petroleomics, Metallopetroleomics, and Petrointeractomics. Energy & Fuels,

32(4), 4593-4605.

Hammouda, B. (2010). SANS from polymers—review of the recent literature. Journal of

Macromolecular Science®, Part C: Polymer Reviews, 50(1), 14-39.

Haug, I., Williams, M. A. K., Lundin, L., Smidsrød, O., & Draget, K. I. (2003). Molecular

interactions in, and rheological properties of, a mixed biopolymer system undergoing

order/disorder transitions. Food Hydrocolloids, 17(4), 439-444.

Hemar, Y., Tamehana, M., Munro, P., & Singh, H. J. F. H. (2001). Viscosity, microstructure

and phase behavior of aqueous mixtures of commercial milk protein products and

xanthan gum. 15(4-6), 565-574.

Hemar, Y., Tamehana, M., Munro, P. A., & Singh, H. (2001). Viscosity, microstructure and

phase behavior of aqueous mixtures of commercial milk protein products and xanthan

gum. Food Hydrocolloids, 15(4), 565-574. doi:https://doi.org/10.1016/S0268-

005X(01)00077-7

Horkay, F., Basser, P. J., Hecht, A.-M., & Geissler, E. J. T. J. o. c. p. (2018). Ionic effects in

semi-dilute biopolymer solutions: A small angle scattering study. 149(16), 163312.

Horkay, F., Hammouda, B. J. C., & Science, P. (2008). Small-angle neutron scattering from

typical synthetic and biopolymer solutions. 286(6-7), 611-620.

Icoz, D. Z., & Kokini, J. L. J. C. p. (2007). Probing the boundaries of miscibility in model

carbohydrates consisting of chemically derivatized dextrans using DSC and FTIR

spectroscopy. 68(1), 68-76.

Ikeda, S., Nitta, Y., Kim, B. S., Temsiripong, T., Pongsawatmanit, R., & Nishinari, K. J. F. H.

(2004). Single-phase mixed gels of xyloglucan and gellan. 18(4), 669-675.

Inamura, I., & Jinbo, Y. (1991). Interaction between poly (vinyl alcohol) and poly (ethylene

glycol) in water studied by viscosity and density. Polymer journal, 23(9), 1143-1147.

Jain, V. M., Karibasappa, G. N., Dodamani, A. S., & Mali, G. V. (2017). Estimating the

carbohydrate content of various forms of tobacco by phenol-sulfuric acid method.

Journal of education and health promotion, 6.

Jiang, B., Tsao, R., Li, Y., & Miao, M. (2014). Food Safety: Food Analysis

Technologies/Techniques. In N. K. Van Alfen (Ed.), Encyclopedia of Agriculture and

Food Systems (273-288). Oxford: Academic Press.

Johansson, H.-O., Karlström, G., Tjerneld, F., & Haynes, C. A. (1998). Driving forces for phase

separation and partitioning in aqueous two-phase systems. Journal of Chromatography

B: Biomedical Sciences and Applications, 711(1), 3-17.

Jonkman, J., & Brown, C. M. J. J. o. b. t. J. (2015). Any way you slice it—a comparison of

confocal microscopy techniques. 26(2), 54.

Kacir, L., Narkis, M., & Ishai, O. (1975). Oriented short glass‐fiber composites. I. Preparation

and statistical analysis of aligned fiber mats. Polymer Engineering & Science, 15(7),

525-531.

Kasapis, S. (2008). Phase Separation in Biopolymer Gels: A Low- to High-Solid Exploration

of Structural Morphology and Functionality. Critical reviews in food science and

nutrition, 48(4), 341-359.

Kasapis, S. (2009). CHAPTER 7 - Unified Application of the Materials-Science Approach to

the Structural Properties of Biopolymer Co-Gels throughout the Industrially Relevant

Level of Solids. In S. Kasapis, I. T. Norton, & J. B. Ubbink (Eds.), Modern Biopolymer

Science (225-260). San Diego: Academic Press.

Kasapis, S., & Bannikova, A. (2017). Chapter 2 - Rheology and Food Microstructure. In J.

Ahmed, P. Ptaszek, & S. Basu (Eds.), Advances in Food Rheology and Its Applications

(7-46): Woodhead Publishing.

Kasapis, S., Morris, E. R., Norton, I. T., & Clark, A. H. (1993). Phase equilibria and gelation

in gelatin/maltodextrin systems — Part IV: composition-dependence of mixed-gel

moduli. Carbohydrate polymers, 21(4), 269-276. doi:https://doi.org/10.1016/0144-

8617(93)90058-C

Kerner, E. (1956a). The elastic and thermo-elastic properties of composite media. Proceedings

of the physical society. Section B, 69(8), 808.

Kerner, E. (1956b). The electrical conductivity of composite media. Proceedings of the

physical society. Section B, 69(8), 802.

Kim, J. M., Jung, H. S., Park, J. W., Lee, H. Y., & Kawai, T. J. A. c. a. (2004). AFM phase lag

mapping for protein–DNA oligonucleotide complexes. 525(2), 151-157.

King, S. M. (1999). Small-angle neutron scattering. Modern Techniques for Polymer

Characterisation, 12.

Klassen, D. R., Elmer, C. M., & Nickerson, M. T. (2011). Associative phase separation

involving canola protein isolate with both sulphated and carboxylated polysaccharides.

Food Chemistry, 126(3), 1094-1101.

Koh, L. W., & Kasapis, S. (2011). Orientation of short microcrystalline cellulose fibre in a

gelatin matrix. Food Hydrocolloids, 25(5), 1402-1405.

Kontogiorgos, V., Tosh, S. M., & Wood, P. J. J. F. B. (2009). Kinetics of phase separation of

oat β-glucan/whey protein isolate binary mixtures. 4(3), 240.

Krigbaum, W. R., & Wall, F. T. (1950). Viscosities of binary polymeric mixtures. Journal of

Polymer Science, 5(4), 505-514.

Leharne, S. J. C. (2017). The physical chemistry of high-sensitivity differential scanning

calorimetry of biopolymers. 3(1), 1.

Leloup, V., Colonna, P., & Buleon, A. J. J. o. C. S. (1991). Influence of amylose-amylopectin

ratio on gel properties. 13(1), 1-13.

Lewis, T. B., & Nielsen, L. E. (1970). Dynamic mechanical properties of particulate‐filled

composites. Journal of applied polymer science, 14(6), 1449-1471.

Li-Chan, E. C. (1996). The applications of Raman spectroscopy in food science. Trends in

Food Science & Technology, 11(7), 361-370.

Li, X., Liu, Y., Li, N., Xie, D., Yu, J., Wang, F., & Wang, J. (2016). Studies of phase separation

in soluble rice protein/different polysaccharides mixed systems. LWT-Food Science

and Technology, 65, 676-682.

Loret, C., Schumm, S., Pudney, P. D., Frith, W. J., & Fryer, P. J. (2005). Phase separation and

molecular weight fractionation behaviour of maltodextrin/agarose mixtures. Food

Hydrocolloids, 19(3), 557-565.

Mæhre, H. K., Dalheim, L., Edvinsen, G. K., Elvevoll, E. O., & Jensen, I.-J. (2018). Protein

determination—method matters. Foods, 7(1), 5.

Magonov, S. N. (2006). Atomic force microscopy in analysis of polymers. Encyclopedia of

Analytical Chemistry: Applications, Theory and Instrumentation.

Marshall, R. J. (2005). FOOD AND NUTRITIONAL ANALYSIS | Dairy Products. In P.

Worsfold, A. Townshend, & C. Poole (Eds.), Encyclopedia of Analytical Science

(Second Edition) (312-319). Oxford: Elsevier.

Masuelli, M. A., & Illanes, C. O. (2014). Review of the characterization of sodium alginate by

intrinsic viscosity measurements: Comparative analysis between conventional and

single point methods.

Masuko, T., Minami, A., Iwasaki, N., Majima, T., Nishimura, S.-I., & Lee, Y. C. (2005).

Carbohydrate analysis by a phenol–sulfuric acid method in microplate format.

Analytical biochemistry, 339(1), 69-72.

Matsushita, M., Irino, T., Komoda, T., & Sakagishi, Y. (1993). Determination of proteins by a

reverse biuret method combined with the copper-bathocuproine chelate reaction.

Clinica chimica acta, 216(1-2), 103-111.

Melish, K. A Literature Review on Viscometers that Measure the Viscosity at Extreme

Conditions.

Mertz, J. (2019). Introduction to optical microscopy: Cambridge University Press.

Mhaske, P., Condict, L., Dokouhaki, M., Farahnaky, A., & Kasapis, S. (2020). Phase volume

quantification of agarose-ghee gels using 3D confocal laser scanning microscopy and

blending law analysis: A comparison. LWT, 109567.

Mhaske, P., Condict, L., Dokouhaki, M., Katopo, L., & Kasapis, S. (2019). Quantitative

analysis of the phase volume of agarose-canola oil gels in comparison to blending law

predictions using 3D imaging based on confocal laser scanning microscopy. Food

Research International, 125, 108529.

Michon, C., Cuvelier, G., Launay, B., Parker, A., & Takerkart, G. (1995). Study of the

compatibility/incompatibility of gelatin/iota-carrageenan/water mixtures.

Carbohydrate polymers, 28(4), 333-336.

Miles, M. J., Morris, V. J., Orford, P. D., & Ring, S. G. J. C. r. (1985). The roles of amylose

and amylopectin in the gelation and retrogradation of starch. 135(2), 271-281.

Morris, E. (2000). Rheology of biopolymer co-gels. In Hydrocolloids (135-146): Elsevier.

Morris, E. R. (1992). The effect of solvent partition on the mechanical properties of biphasic

biopolymer gels: an approximate theoretical treatment. Carbohydrate polymers, 17(1),

65-70.

Morris, E. R. (2009). CHAPTER 5 - Functional Interactions in Gelling Biopolymer Mixtures.

In S. Kasapis, I. T. Norton, & J. B. Ubbink (Eds.), Modern Biopolymer Science (167-

198). San Diego: Academic Press.

Morris, V. J., Kirby, A. R., & Gunning, A. P. (1999). Atomic force microscopy for biologists

(Vol. 57): World Scientific.

Mousia, Z., Farhat, I., Blachot, J., & Mitchell, J. J. P. (2000). Effect of water partitioning on

the glass-transition behaviour of phase separated amylopectin–gelatin mixtures. 41(5),

1841-1848.

Neirynck, N., Dewettinck, K., & Van der Meeren, P. (2007). Influence of pH and biopolymer

ratio on sodium caseinate—guar gum interactions in aqueous solutions and in O/W

emulsions. Food Hydrocolloids, 21(5-6), 862-869.

Nguyen-Tri, P., Ghassemin, P., Carriere, P., Assadi, A. A., & Nguyen, D. D. (2020). Recent

Progressive Use of Advanced Atomic Force Microscopy in Polymer Science: A

Review.

Nielsen, L. E. (1970). Generalized equation for the elastic moduli of composite materials.

Journal of Applied Physics, 41(11), 4626-4627.

Normand, V., Pudney, P. D., Aymard, P., & Norton, I. T. J. J. o. A. P. S. (2000). Weighted‐

average isostrain and isostress model to describe the kinetic evolution of the mechanical

properties of a composite gel: Application to the system gelatin: Maltodextrin. 77(7),

1465-1477.

Norton, I. T., & Frith, W. J. (2001). Microstructure design in mixed biopolymer composites.

Food Hydrocolloids, 15(4), 543-553.

Palierne, J. (1990). Linear rheology of viscoelastic emulsions with interfacial tension.

Rheologica acta, 29(3), 204-214.

Palierne, J. (1991). Rheologica Acta Rheol Acta 30: 497 (1991). Rheologica acta, 30(497).

Parets, M. J., Garcia, R., Soria, V., & Campos, A. (1990). Solution properties of

polyelectrolytes—V. Viscometric study of mixed polyanions in pure water. European

Polymer Journal, 26(7), 767-773. doi:https://doi.org/10.1016/0014-3057(90)90127-P

Patonay, G., Salon, J., Sowell, J., & Strekowski, L. J. M. (2004). Noncovalent labeling of

biomolecules with red and near-infrared dyes. 9(3), 40-49.

Perrechil, F., Braga, A., & Cunha, R. (2009). Interactions between sodium caseinate and LBG

in acidified systems: Rheology and phase behavior. Food Hydrocolloids, 23(8), 2085-

2093.

Picout, D. R., & Ross-Murphy, S. B. (2003). Rheology of biopolymer solutions and gels.

TheScientificWorldJOURNAL, 3.

Pittz, E. P., Lee, J. C., Bablouzian, B., Townend, R., & Timasheff, S. N. (1973). [10] Light

scattering and differential refractometry. In Methods in enzymology 27(209-256):

Elsevier.

Pluta, M., & Maksymilian, P. (1988). Advanced light microscopy (Vol. 1): Elsevier

Amsterdam.

Polyakov, V. I., Grinberg, V. Y., & Tolstoguzov, V. B. (1980). Application of phase-volume-

ratio method for determining the phase diagram of water-casein-soybean globulins

system. Polymer Bulletin, 2(11), 757-760. doi:10.1007/BF00255893

Prameela, K., Mohan, C. M., & Ramakrishna, C. (2018). Biopolymers for food design:

Consumer-friendly natural ingredients. In Biopolymers for food design (pp. 1-32):

Elsevier.

Pudney, P. D. A., Hancewicz, T. M., Cunningham, D. G., & Gray, C. (2003). A novel method

for measuring concentrations of phase separated biopolymers: the use of confocal

Raman spectroscopy with self-modelling curve resolution. Food Hydrocolloids, 17(3),

345-353.

Rafe, A., & Razavi, S. M. (2015). Effect of thermal treatment on chemical structure of β-

lactoglobulin and basil seed gum mixture at different states by ATR-FTIR

spectroscopy. International Journal of Food Properties, 18(12), 2652-2664.

Rahman, M. S., Al-Marhubi, I. M., & Al-Mahrouqi, A. (2007). Measurement of glass transition

temperature by mechanical (DMTA), thermal (DSC and MDSC), water diffusion and

density methods: A comparison study. Chemical Physics Letters, 440(4), 372-377.

Reichardt, W., & Eckert, W. (1991). The determination of protein content of milk, cheese, and

meat with the use of the biuret reaction. Die Nahrung, 35(7), 731.

Richardson, R. K., & Kasapis, S. (1998). Rheological methods in the characterisation of food

biopolymers. In Developments in Food Science, 39(1-48): Elsevier.

Robeson, L. M. (2007). Polymer blends. A Comprehensive Review.

Sapan, C. V., Lundblad, R. L., & Price, N. C. (1999). Colorimetric protein assay techniques.

Biotechnology and applied Biochemistry, 29(2), 99-108.

Sarika, P. R., Pavithran, A., & James, N. R. (2015). Cationized gelatin/gum arabic

polyelectrolyte complex: Study of electrostatic interactions. Food Hydrocolloids, 49,

176-182.

Schmitt, C., Sanchez, C., Desobry-Banon, S., & Hardy, J. (1998). Structure and

technofunctional properties of protein-polysaccharide complexes: a review. Critical

reviews in food science and nutrition, 38(8), 689-753.

Scholten, E., Tuinier, R., Tromp, R. H., & Lekkerkerker, H. (2002). Interfacial tension of a

decomposed biopolymer mixture. Langmuir, 18(6), 2234-2238.

Schorsch, C., Clark, A., Jones, M., & Norton, I. (1999). Behaviour of milk

protein/polysaccharide systems in high sucrose. Colloids and surfaces B: Biointerfaces,

12(3-6), 317-329.

Schorsch, C., Jones, M., & Norton, I. J. F. H. (1999). Thermodynamic incompatibility and

microstructure of milk protein/locust bean gum/sucrose systems. 13(2), 89-99.

Schorsch, C., Jones, M., & Norton, I. J. F. H. (2000). Phase behaviour of pure micellar casein/κ-

carrageenan systems in milk salt ultrafiltrate. 14(4), 347-358.

Semasaka, C., Mhaske, P., Buckow, R., & Kasapis, S. J. F. c. (2018). Modelling water partition

in composite gels of BSA with gelatin following high pressure treatment. 265, 32-38.

Semenova, M., & Savilova, L. (1998). The role of biopolymer structure in interactions between

unlike biopolymers in aqueous medium. Food Hydrocolloids, 12(1), 65-75.

Shamsuri, A., Abdullah, D. K., Daik, R. J. C. C., & Technology. (2012). FABRICATION OF

AGAR/BIOPOLYMER BLEND AEROGELS IN IONIC LIQUID AND CO-

SOLVENT MIXTURE. 46, 45-52.

Shrinivas, P., Kasapis, S., & Tongdang, T. J. L. (2009). Morphology and mechanical properties

of bicontinuous gels of agarose and gelatin and the effect of added lipid phase. 25(15),

8763-8773.

Sigillo, I., Di Santo, L., Guido, S., & Grizzuti, N. (1997). Comparative measurements of

interfacial tension in a model polymer blend. Polymer Engineering & Science, 37(9),

1540-1549.

Smoluchowski, M. J. T. L., Edinburgh,, Magazine, D. P., & Science, J. o. (1912). XIII. On

opalescence of gases in the critical state. 23(133), 165-173.

Soedjak, H. S. (1994). Colorimetric determination of carrageenans and other anionic

hydrocolloids with methylene blue. Analytical chemistry, 66(24), 4514-4518.

Spyropoulos, F., Ding, P., Frith, W., Norton, I., Wolf, B., & Pacek, A. (2008). Interfacial

tension in aqueous biopolymer–surfactant mixtures. Journal of Colloid and Interface

Science, 317(2), 604-610.

Stockmayer, W. (1950). Light scattering in multi‐component systems. The Journal of Chemical

Physics, 18(1), 58-61.

Stokes, J., Wolf, B., & Frith, W. (2001). Phase-separated biopolymer mixture rheology:

Prediction using a viscoelastic emulsion model. Journal of Rheology, 45(5), 1173-1191.

Svensson, M., Berggren, K., Veide, A., & Tjerneld, F. (1999). Aqueous two-phase systems

containing self-associating block copolymers: Partitioning of hydrophilic and

hydrophobic biomolecules. Journal of Chromatography A, 839(1), 71-83.

Takayanagi, M. (1963). Viscoelastic behavior of polymer blends and its comparison with

model experiments. J. Soc. Mater. Sci. Jpn., 12, 389-394.

Tanaka, F. (2002). Theoretical study of molecular association and thermoreversible gelation in

polymers. Polymer journal, 34(7), 479-509.

Thaiudom, S., & Goff, H. D. (2003). Effect of κ-carrageenan on milk protein polysaccharide

mixtures. International Dairy Journal, 13(9), 763-771.

Thaiudom, S., & Goff, H. D. (2003). Effect of κ-carrageenan on milk protein polysaccharide

mixtures. International Dairy Journal, 13(9), 763-771.

Tolstoguzov, V. (1991). Functional properties of food proteins and role of protein-

polysaccharide interaction. Food Hydrocolloids, 4(6), 429-468.

Tolstoguzov, V. (1999). Compositions and phase diagrams for aqueous systems based on

proteins and polysaccharides. In International review of cytology 192(3-31): Elsevier.

Tolstoguzov, V. (2003). Some thermodynamic considerations in food formulation. Food

Hydrocolloids, 17(1), 1-23. doi:https://doi.org/10.1016/S0268-005X(01)00111-4

Tolstoguzov, V. (2006). 17 Phase Behavior in Mixed Polysaccharide Systems. Food

polysaccharides and their applications, 589.

Tolstoguzov, V. B. (1993). Thermodynamic Incompatibility of Food. Food colloids and

polymers: Stability and mechanical properties, 113, 94.

Tolstoguzov, V. B. (2007). 7 - Ingredient interactions in complex foods: aggregation and phase

separation. In D. J. McClements (Ed.), Understanding and Controlling the

Microstructure of Complex Foods (185-206): Woodhead Publishing.

Tomotika, S. J. P. o. t. R. S. o. L. S. A. (1935). On the Instability of a Cylindrical Thread of a

Toubal, M., Nongaillard, B., Radziszewski, E., Boulenguer, P., & Langendorff, V. (2003). Ultrasonic

monitoring of sol–gel transition of natural hydrocolloids. Journal of Food Engineering, 58(1),

1-4.

Viscous Liquid Surrounded by Another Viscous Fluid. 150, 322.

Tromp, R., Rennie, A., & Jones, R. J. M. (1995). Kinetics of the simultaneous phase separation

and gelation in solutions of dextran and gelatin. 28(12), 4129-4138.

Tseretely, G. I., & Smirnova, O. I. (1992). DSC study of melting and glass transition in

gelatins. Journal of Thermal Analysis and Calorimetry, 38(5), 1189-1201.

Tucker III, C. L., & Moldenaers, P. (2002). Microstructural evolution in polymer blends.

Annual Review of Fluid Mechanics, 34(1), 177-210.

Tuinier, R., Ten Grotenhuis, E., Holt, C., Timmins, P. A., & De Kruif, C. G. (1999). Depletion

interaction of casein micelles and an exocellular polysaccharide. Physical Review

E, 60(1), 848.

van de Velde, F., Weinbreck, F., Edelman, M. W., van der Linden, E., & Tromp, R. H. (2003).

Visualisation of biopolymer mixtures using confocal scanning laser microscopy

(CSLM) and covalent labelling techniques. Colloids and Surfaces B:

Biointerfaces, 31(1-4), 159-168.

Van Puyvelde, P., Antonov, Y. A., & Moldenaers, P. (2002). Rheo-optical measurement of the

interfacial tension of aqueous biopolymer mixtures. Food Hydrocolloids, 16(5), 395-

402.

Vis, M., Peters, V. F., Blokhuis, E. M., Lekkerkerker, H. N., Erné, B. H., & Tromp, R. H.

(2015). Decreased interfacial tension of demixed aqueous polymer solutions due to

charge. Physical review letters, 115(7), 078303.

Voron’ko, N. G., Derkach, S. R., Vovk, M. A., & Tolstoy, P. M. (2017). Complexation of κ-

carrageenan with gelatin in the aqueous phase analysed by 1H NMR kinetics and

relaxation. Carbohydrate polymers, 169, 117-126.

Voron’ko, N. G., Derkach, S. R., Vovk, M. A., & Tolstoy, P. M. (2016). Formation of κ-

carrageenan–gelatin polyelectrolyte complexes studied by 1H NMR, UV spectroscopy

and kinematic viscosity measurements. Carbohydrate polymers, 151, 1152-1161.

Wadsö, I. (1980). Calorimetric instrumentation for studies of biopolymer model

compounds. Pure and Applied Chemistry, 52(2), 465-477.

Walter, H., Johansson, G., & Brooks, D. E. (1991). Partitioning in aqueous two-phase systems:

recent results. Analytical biochemistry, 197(1), 1-18.

Wanchoo, R. K., & Sharma, P. K. (2003). Viscometric study on the compatibility of some

water-soluble polymer–polymer mixtures. European Polymer Journal, 39(7), 1481-

1490.

Wang, Z., Yang, K., Brenner, T., Kikuzaki, H., & Nishinari, K. (2014). The influence of agar

gel texture on sucrose release. Food Hydrocolloids, 36, 196-203.

Waterborg, J. H. (2009). The Lowry method for protein quantitation. In The protein protocols

handbook (7-10): Springer.

Williams, M. A., Foster, T. J., Martin, D. R., Norton, I. T., Yoshimura, M., & Nishinari, K. J.

B. (2000). A molecular description of the gelation mechanism of konjac mannan. 1(3),

440-450.

Williams, P., Clegg, S., Langdon, M., Nishinari, K., & Piculell, L. J. M. (1993). Investigation

of the gelation mechanism in. kappa.-carrageenan/konjac mannan mixtures using

differential scanning calorimetry and electron spin resonance spectroscopy. 26(20),

5441-5446.

Winter, P. W., & Shroff, H. (2014). Faster fluorescence microscopy: advances in high speed

biological imaging. Current opinion in chemical biology, 20, 46-53.

Xu, A. Y., Melton, L. D., Ryan, T. M., Mata, J. P., Rekas, A., Williams, M. A. K., &

McGillivray, D. J. (2018). Effects of polysaccharide charge pattern on the

microstructures of β-lactoglobulin-pectin complex coacervates, studied by SAXS and

SANS. Food Hydrocolloids, 77, 952-963.

Xu, J., Fan, X., Ning, Y., Wang, P., Jin, Z., Lv, H., ... & Xu, X. (2013). Effect of spring dextrin

on retrogradation of wheat and corn starch gels. Food Hydrocolloids, 33(2), 361-367.

Yang, H., Wang, Y., Lai, S., An, H., Li, Y., & Chen, F. (2007). Application of atomic force

microscopy as a nanotechnology tool in food science. Journal of food science, 72(4),

65-75.

Yang, K., Wang, Z., Brenner, T., Kikuzaki, H., Fang, Y., & Nishinari, K. (2015). Sucrose

release from agar gels: Effects of dissolution order and the network inhomogeneity.

Food Hydrocolloids, 43, 100-106.

Yuan, C., Xu, D., Cui, B., & Wang, Y. (2019). Gelation of κ-carrageenan/Konjac

glucommanan compound gel: Effect of cyclodextrins. Food Hydrocolloids 87, 158-

164.

Zasypkin, D. V., Braudo, E. E., & Tolstoguzov, V. B. (1997). Multicomponent biopolymer

gels. Food Hydrocolloids, 11(2), 159-170.

Zernike, F. J. P. (1942a). Phase contrast, a new method for the microscopic observation of

transparent objects. 9(7), 686-698.

Zernike, F. J. P. (1942b). Phase contrast, a new method for the microscopic observation of

transparent objects part II. 9(10), 974-986.

Zhang, G., & Foegeding, E. A. (2003). Heat-induced phase behavior of β-

lactoglobulin/polysaccharide mixtures. Food Hydrocolloids, 17(6), 785-792.

Zhang, H., Nishinari, K., Williams, M. A., Foster, T. J., & Norton, I. T. J. I. J. o. B. M. (2002).

A molecular description of the gelation mechanism of curdlan. 30(1), 7-16.

Zhuang, X., Wang, L., Jiang, X., Chen, Y., & Zhou, G. (2020). The effects of three

polysaccharides on the gelation properties of myofibrillar protein: Phase behaviour and

moisture stability. Meat Science, 108228.

Chapter 2 - Materials and Methods

Abstract

This chapter summarises the specifications of the materials used in this work along with

the equipment, instrumentation and the methods employed in this research. Specifications of

the biopolymers, chemicals and reagents used in this research are provided along with the

experimental methods of analysis such as scanning electron microscopy (SEM), Fourier

transform infrared spectroscopy (FTIR), differential scanning calorimetry (DSC), rheology,

and confocal laser scanning microscopy. The working principle and reasoning behind the

analytical instruments are provided in detail. Further explanation specific to each experimental

chapter can be found in Chapters 3 to 6.

2.1. Materials

2.1.1. Biomaterials

2.1.1.1. Agarose

Agarose is a natural polysaccharide and the gelling component of Agar, obtained from the

Rhodophyceae family of red seaweeds. Agarose has negligible traces of sulphur content as

opposed to the charged fraction of agar, agaropectin and exhibits an unbranched linear structure

(Figure 2.1) . In its solid states, it is in the powder form, existing as a threefold, left-handed

double helix. The hydroxyl groups line the central cavity along the helix axis and are

responsible for the hydrogen bonding with water. When dissolved in hot water (~70°C),

agarose remains dispersed as fluctuating disordered coils. Upon cooling, these disordered coils

form an ordered double helix, forming a gel network. In this study, agarose is used as the main

gelling biopolymer within which other biomaterials are entrapped. Type-1 agarose used for the

experiments was supplied by Sigma-Aldrich Co., Sydney, Australia. The powder has an off-

white colour, with the moisture content being less than 10 g H2O/g dry matter and less than

0.15% sulphate according to the supplier. Agarose forms a strong gel with a storage modulus

of 7208 Pa at 1% w/w (at 5°C, angular frequency of 1 rad/sec and 0.01% strain).

Figure 2.1 Chemical structure of an agarose polymer (Lee & Bahaman, 2012)

2.1.1.2. Gelatin

As a common animal-derived protein, gelatin is extracted via partial acid or alkaline hydrolysis

and proteolysis of collagen present in the skin, cartilage, bone and connective tissue of fish,

cows or pigs. The physical and chemical properties of gelatin, such as the gel strength,

isoelectric point and molecular weight, vary with the source of gelatin, as well as the type and

extent of extraction. Figure 2.2 illustrates the chemical structure of gelatin. Type-B bovine

gelatin generally results in higher viscosity, obtained via chemically intensive alkaline

hydrolysis for extended time and was used in this work to form a bicontinuous network along

with agarose. It was purchased from Sigma-Aldrich Co., Sydney, Australia. The powder was

light yellow in colour with a bloom strength of 225 with a molecular weight of 173kDa. Gelatin

formed a soft gel matrix with a storage modulus of 81 Pa at 1% w/w (at 5°C, angular frequency

of 1 rad/sec and 0.01% strain).

Figure 2.2. Chemical structure of gelatin (Deshmukh et al. 2017)

2.1.1.3. Lipids

Canola oil is produced from the variety of rapeseeds that have a low erucic acid content. It

mainly constitutes of a mixture of α-linolenic acid and oleic acid. In the present study it was

used as a discontinuous filler that remained dispersed in a continuous gel network exhibiting

distinct phase separation. Food grade canola oil was purchased from Manildra Group, NSW,

Australia. 0.06% free fatty acid as oleic acid was present in the oil with a peroxide value of 0.1.

It remains in the liquid state at room temperature with a negligible storage modulus.

Pure cow ghee, a form of clarified butter manufactured by Amul (Gujrat, India) was

used in the study. The density of ghee was measured to be 0.911g/ml. Ghee is a complex

mixture of triacylglycerols (TAGs), which constitutes 98% of its composition. In this research,

ghee is seen to crystalize between 10-15°C.

2.1.1.4. Microcrystalline cellulose

Microcrystalline cellulose (MCC) is widely used in the food and pharmaceutical industries as

a texture modifier/fat replacer and tableting excipient, respectively. MCC is obtained from the

acid hydrolysis of naturally occurring cellulose, such as wood pulp. The extracted colloidal

MCC when dispersed in water, can form a three-dimensional network as the hydrocolloid layer

that surrounds the cellulose particles absorbs water and swells. The high-water holding capacity

of the MCC particles make it ideal as a fat replacer. As it cannot be digested, MCC is also used

in food products to increase their fibre content. MCC was obtained from Sigma Aldrich Co.,

Sydney, Australia. The product was in the form of white fibres with a bulk density of 0.6g/cm3.

The cellulose crystals, being hygroscopic, absorbed water about 4.1 times their weight from

the system without disintegrating.

2.1.2 Chemicals and Dyes

Details of all the chemicals including dyes used for staining in this study are stated in Table

2.1. Throughout the study, Milli-Q water is used for preparing chemical solutions.

Table 2.1 List of chemicals and dyes used.

Chemical CAS Number Supplier

Agarose 9004-34-6 Sigma Aldrich

DMSO 67-68-5 Sigma Aldrich

Nile Red 7385-67-3 Sigma Aldrich

DTAF 51306-35-5 AAT Bioquest

FITC 27072-45-3 Sigma Aldrich

Rhodamine B 81-88-9 Sigma Aldrich

Microcrystalline Cellulose 9004-34-6 Sigma Aldrich

Gelatin 9000-70-8 Sigma Aldrich

2.1.3 Instruments

The instruments used in this study along with their manufacturers are listed in Table 2.2

Table 2.2 List of Instruments used

Instrument Type Manufacturer

MDSC Q2000 Calorimeter TA Instruments

(New Castle, DE, USA)

Quanta™ 200 Microscope FEI

(Hillsboro, OR, USA)

FriZone Triade freeze dryer Labconco (Kansas City,

Missouri, USA)

Spectrum Two FTIR spectrometer Perkin Elmer (Norway, CT,

USA)

Setaram VII Calorimeter Setaram (Seturau, Caluire,

France)

Discovery Hybrid Rheometer TA Instruments (New Castle,

DE, USA)

Eclipse Ti-E Microscope Nikon (Minato, Tokyo,

Japan)

2.1.4 Sample preparation

2.1.4.1. Agarose-Canola oil composites

The concentration of agarose was kept constant at 1% (w/w) and the concentration of canola

oil was varied (1, 3, 5, 7, 9 and 11%, w/w). Agarose powder was dissolved in milliQ water at

80°C and stirred for 5 min before cooling to 45°C to prepare agarose solution. Respective

quantity of canola oil was then added to the agarose solution and homogenised at 3000 rpm for

30 s using an IKA T25 Ultra Turrax homogeniser (Wilmington, North Carolina, USA).

2.1.4.2. Agarose-ghee composites

Binary composites were prepared by maintaining the concentration of agarose at 1% (w/w) and

varying the ghee concentrations (0, 1, 3, 6, 9, 12 and 15%) with the remainder being milliQ

water to make the total volume of sample, 100 mL. I g of agarose powder was dispersed in

calculated amount of milliQ water and heated to 75°C. It was stirred for 10 minutes and cooled

to 45°C before the addition of ghee. The mixture was then homogenised using the IKA

homogeniser-T25 Turrax at 5000 rpm for 30s.

2.1.4.3. Agarose-MCC composites

99 ml of MCC solutions were prepared by adding varying concentrations of MCC (0.1, 0.3,

0.6, 0.9, 1.2 and 1.5 g) to calculated quantities of milliQ water. These solutions were stood

overnight at room temperature to ensure complete hydration of the MCC particles. These

suspensions were then heated to 80°C with constant magnetic stirring before adding 1g agarose

to the mixture. The agarose/MCC composites were stirred for 15 min for the complete

dissolution of agarose and cooled to 50°C for further analysis.

2.1.4.4. Agarose-gelatin composites

The concentration of gelatin in the composite was maintained at 3% and the concentration of

agarose was varied to yield 0.1, 0.4, 0.8, 1.2, 1.6, 2.0, 2.5 and 3% of agarose in the final

composite. Required amounts of agarose were dissolved in milliQ water at 75°C and allowed

to dissolve completely before cooling the mixture down to 60°C and adding gelatin. Samples

were stirred for 60 minutes for homogenous dissolution prior to analysis.

2.1.4.5. Staining for confocal laser scanning microscopy

The agarose/lipid composites in Chapters 3 and 4, were stained using 5mg/L Nile red, which

is extensively used to stain lipids. The stained sample solution was carefully pipetted into

precooled Coverwell imaging chamber for quick and homogenous gel formation.

In Chapter 5, the procedure outlined in the works of Russ et al. (2013) was followed to

covalently label agarose using 5-DTAF (5-(4, 6-Dichlorotriazinyl) Aminofluorescein). 1g of

agarose powder was dispersed in 50 ml milliQ water at heated to 80°C with constant stirring.

After cooling the solution to 60°C, 8mg 5-DTAF was added followed by the slow addition of

15 ml of 1%(w/v) Na2SO4. After the addition of 3-4 drops of 10% (w/v) NaOH, the mixture

was stirred for 2 hours before adding 120 ml absolute ethanol to get a yellow precipitate. The

solution was refrigerated for 20 min at -20°C before filtering it and washing the residue with

ethanol. The yellow residue was dried to obtain a fine agarose powder.

In chapter 6, the protein in the agarose-gelatin mixture was stained using 0.001% w/w FITC.

The dye was added to the composite and stirred for 60 min before allowing gel formation for

imaging.

2.2.2 Methods

2.2.2.1 Fourier Transform Infrared (FTIR) Spectroscopy

Spectroscopy focuses on, measures and interprets the spectra obtained due to the interaction of

electromagnetic radiation with materials (Penner, 2010). Spectroscopy can be further classified

into different types based on the region of electromagnetic spectrum (UV, visible light or IR),

type of interaction (absorption, transmission or diffraction) and type of matter (atomic or

molecular spectroscopy). FTIR specifically focuses on the infrared region of the

electromagnetic spectrum which can be further classified into near – IR (14000-4000 cm-1),

mid IR (4000-400 cm-1) and far-IR (400-10 cm-1) based on their wavenumbers. The mid-IR

spectra is the most commonly used range for FTIR analysis (Türker-Kaya & Huck, 2017).

When the material interacts with infrared radiation, it excites the molecules changing the

charge distribution of the molecule and subsequently, its electric dipole moment. The most

essential vibration modes include stretching and bending (rocking, twisting, torsioning,

scissoring and wagging) (Gaca-Zając et al., 2018). Distinct molecules have a unique vibrational

energy spectra characterized by its frequency, band shape and intensity, that serves as the

fingerprint for that particular molecule (Gutiérrez Sama et al., 2018).

Figure 2.3 Schematic Diagram of FTIR (Dole, Patel, Sawant, & Shedpure, 2011)

As depicted in Figure 2.3, a typical FTIR comprises of an infrared light source, a detector, a

fixed and a sliding mirror and a beam splitter. Infrared radiation from the source hits the beam

splitter and is ‘split’ into two and each half is diverted to the sliding and fixed mirror,

respectively. The beams are reflected back to the splitter at a point, where they combine to

induce either a destructive or constructive interference. The combined beam passes through the

sample to a detector which compiles an interferogram. The interferogram is transferred into the

source spectra with and without the absorption spectra of the sample. The proportion of the

source spectra to the absorption spectra is considered to be the sample’s FTIR transmission

spectrum (Komal Kumar & Devi Prasad, 2012).

Table 2.3 Absorption frequencies of various organic functional groups in the Mid- IR region

adapted from the works of Stuart, 2005

Groups Frequency (cm-1)

1470-1420 and 1380-1340 -CH2 and -CH3 bending Alkanes -CH stretching and bending 3000-2800

Olefinic –CH stretching 3100-3000 Alkenes

Acetylenic –CH stretching 3300 Alkynes

-CH stretching 3100-3000

-C=C- stretching 1600 Aromatics -OH bending 1500-1300

C-O stretching 1220-1000

C-O asymmetric stretching 1220-1000 Ethers

-NH stretching 3500-3300 Amines

-C=O stretching 1670-1640 Amides

-NH stretching 3500-3100

-NH bending 1640-1550

-C=O stretching 1735-1700 Aldehydes and Ketones

-C=O stretching 1740-1720 Carboxylic Acids

The IR radiation of a molecule can be determined by measuring its frequency as follows:

E = hv (1)

Where E represents the system’s energy, h denotes Planck’s constant (6.63 x 10-27 J.s) and

the frequency (in hertz) is given by v. Frequency is generally denoted in wavenumbers, ṽ

(cm-1) as follows:

ṽ = 1/ λ (2)

where λ is light’s wavelength in cm.

In the food industry, FTIR analysis is widely used as a technique in quality control. FTIR can

rapidly produce an IR spectra of the sample with great accuracy and with extreme convenience.

This can be co-related with the reference spectra to detect any anomalies that can affect the

quality of the product. The interpretation of the spectra and the peaks needs a fundamental

grasp of the composition and properties of food in terms of the micro and macro components

present and their molecular structure. Table 2.3 gives a broad classification of the absorption

frequencies of organic functional groups in the mid – IR region.

A Spectrum 100 with MIRacleTM ZnSe single reflection ATRplate from Perkin Elmer

(Norwalk, Connecticut) given in Figure 2.4 was employed for analysis.

Figure 2.4 FTIR- ATR Spectrum 100 at RMIT University

2.2.2.2 Differential Scanning Calorimetry

Calorimetric approaches have been widely used in food research for the quantitative and

qualitative analysis of the thermodynamic properties of the food systems. It studies and

measures the temperatures, the heat flow and enthalpy during the kinetic reactions and thermal

transitions of a material like heating, cooling, melting, crystallisation, denaturation, etc. All of

the reactions are classified into either endothermic (absorption of heat by the system) or

exothermic (release of heat) reactions.

Figure 2.5: Schematic diagram of a heat flux type DSC (Barron et al., 2012)

There are two types of Differential Scanning Calorimetry (Markus B. Dürrenberger, Stephan

Handschin, Béatrice Conde-Petit, & Felix Escher) – heat flux DSC and power compensation

DSC. Both types of DSC essentially study the heatflow in a system secluded from

environmental influences (Figure 2.5). In power compensation DSC, the sample is kept in a

hermetically sealed pan and an empty reference pan are places in two different chambers. Both

the pans are designed to be heated or cooled at the same rate (dT/dt). However, since one pan

is empty and the other contains the sample, the rate of supply of the thermal energy or the heat

flow (dH/dt) is different. In the heat flow DSC, the hermetically sealed sample and the empty

reference pan are kept in one sample chamber and heated. The heat flow (dH/dt) is the same

for both the pans but the rate of heating (dT/dt) achieved by both of them is different. A

computer is set up with the instrument to monitor and record, the initial temperature, the rate

of heating and the heat flow of the system (Guigo & Sbirrazzuoli, 2018).

When the sample is heated, the molecules or atoms gain energy to break free from the ordered

rigid state in solids, to a more unordered state in liquids. Figure 2.6 shows a typical example

of the effect of change in temperature on the material. Both exothermic (crystallisation) and

endothermic (melting) reactions or transitions can occur in a sample as a result of heating.

Figure 2.6: Typical material transition as a function of temperature change in a DSC adapted

from Verdonck, Schaap, & Thomas, 1999

A transition that is result of energy absorbed by the system, is termed as endothermic transition,

eg. melting. The energy absorbed by the system per unit mass to bring about the transition, is

known as latent heat (Emadi, Whiting, DJURDJEVIC, Kierkus, & Sokolowski, 2018). Glass

transition in amorphous materials has no latent heat associated with it and is referred to as a

second order transition.

At a fixed heating rate (dT/dt), the instrument can measure the system’s heat flow (dH/dt). The

heat capacity (Cp) of the material can be further calculated by dividing the heat flow (dH/dt)

by the rate of heating (dT/dt).

⁄ ] (3) 𝐶𝑝 = [(𝑑𝐻 𝑑𝑡⁄ ) (𝑑𝑇 𝑑𝑡⁄ )

When the mass of the sample (m), is known, the specific heat capacity (Cp/m) of the sample

𝑑𝐻

can be obtained as a function of temperature as follows (Thomas, 2005):

𝑑𝑡

(4) + 𝑓(𝑇, 𝑡) = 𝐶𝑝 𝑑𝑇 𝑑𝑡

Where dH/dt refers to the rate of total heat flow (mW or mJ/s)

Cp refers to the heat capacity of the sample in (J/°C)

dT/dt refers to the rate of heating (°C/min)

f(T,t) refers to heat flow as a function of temperature and time (mW)

A user defined heating/cooling rate and regime is used to measure the gain or loss of energy of

the system during it. The result is displayed as a thermogram as displayed in Figure 2.6. The

area of a peak in the thermogram relates to the energy lost or gained during the thermal

transition (Liu, Yao, Fan, Wang, & Hu, 2018). Two different types of DSC – micro and

modulated DSC, were employed in this study.

2.2.2.2.1 Micro Differential Scanning Calorimetry

The micro DSC is fabricated to work on bulk samples and dilute solutions with great versatility,

precision, and sensitivity. The characteristic feature of the micro DSC is its three-dimensional

sensor facilitated to achieve greater contact area with the sample yielding a higher degree of

sensitivity and accuracy. Sensitive Peltier elements are used to ensure contact with the

calorimetric block as shown in Figure 2.7 (Kodre, Attarde, Yendhe, Patil, & Barge, 2014).

Thermal conductors are used to ensure that the temperature of the vessel is same as that of the

calorimetric block. Thermistors, thermocouples, thermophiles, and resistance wires are used as

detectors to detect the heat flux of the sample in the vessel. Transducers are fit on both the

sample and reference vessel to eliminate changes common in both.

Figure 2.7: The heat flux micro DSC with a cylinder type measuring system (Kodre, Attarde,

Yendhe, Patil, & Barge, 2014).

Setaram VII from Setaram was employed in this research (Figure 2.8).

Figure 2.8: Micro DSC Setaram VII at RMIT University

2.2.2.2.2 Modulated Differential Scanning Calorimetry

Most food systems are complex composites made up of two or more components. Hence,

during a thermal regime, the heat flux of a given sample may comprise of multiple weak, broad

or overlapping transition signals arising from the different components, material impurities or

the presence of water. The modulated DSC overcomes this issue by delineating complex

transitions into individual contributing transitions. This is achieved by heating the samples with

two heating rates simultaneously, resulting in complementary information that facilitates the

deconvolution of the total heat flow into reversing and non reversing heat flow (Jana & Das,

2018). In addition, a sinusoidal modulation is applied in the modulated DSC on the linear

heating ramp, supplying uneven heatflow to the sample in a periodic fashion (Verdonck et al.,

1999). This causes to average sample temperature to continuously rise but in a non linear

fashion. The rate of the heating ramp, the period of modulation and the temperature amplitude

of modulation are critical parameters that need to be considered in the modulated DSC.

Figure 2.9: Modulated DSC Q2000 from TA instruments at RMIT University

2.2.2.3 Rheology

Rheology is the study of the flow and deformation of matter. First developed in synthetic

biopolymers owing to their extensive application in daily utilities, rheology has found

widespread application in various industries like food, pharmaceutical, ceramics, metals, to

name but a few. Food rheology has a direct impact on the final product and its acceptability as

it influences the product’s texture, flavour and appearance (Cheremisinoff, 2018). For example

the ease with which ketchup can be poured from its bottle, the disintegration of food in the

mouth affecting the release of flavour compounds, etc. Hence, rheology is one of the most

important components in product development and quality control. The main objectives of

performing rheological studies are:

i) Obtaining quantitative data of the materials’ mechanical properties

ii) Evaluating food texture and correlating it with sensory analysis

iii) Acquiring information pertaining to the material’s molecular composition and

arrangement

iv) Measuring and correlating quality of the medial and final products

v) Evaluating shelf life of products

In rheological measurements, the most important parameter used for quantification is viscosity,

which is defined as a material’s resistance to flow. The knowledge of a product’s viscosity

helps determine the processing and packaging parameters of a product. Viscosity

measurements are generally involve the use of a Brookefield viscometer and/or a rheometer.

Besides viscosity, other key rheological parameters used in rheological analysis are given in

Table 2.4.

Table 2.4 Rheological parameters

Parameter Definition Symbol Unit (SI)

Stress Force per unit area σ Pa

Strain Relative deformation in shear γ -

Shear rate Change of strain per unit time γ̇ s-1

Viscosity Resistance to flow η Pa.s

Storage Modulus Measure of solid like properties of Gʹ Pa

material

Loss Modulus Measure of fluid like properties of a Gʹʹ Pa

100

material

Phase angle Degree of viscoelasticity Tan δ -

Complex viscosity Resistance to flow of the sample in a η* Pa.s

structured state

Dynamic viscosity Internal friction of a liquid ηʹ Pa.s

Rheological behaviour of fluids and semi-solids can be broadly categorized as Newtonian or

Non-Newtonian behaviour. In Newtonian behaviour shear rate and shear stress exhibit a linear

relationship, with a plot beginning at the origin. Water, clarified juices, milk, etc show a

Newtonian behaviour.

Non-Newtonian materials like yoghurt, salad dressings ketchup, etc show a rheological

behaviour depended on time as a result of structural changes in the material. Non-Newtonian

behaviour includes shear thinning, shear thickening and time dependent shear shinning or shear

thickening. The shear stress-shear strain plot of Non-Newtonian fluids is either non-linear

and/or doesn’t begin at the origin. Shear thinning fluids, also known as pseudoplastic materials

exhibit an increase in shear stress with an increase in shear rate at a rate that is less than

proportional. Salad dressings, concentrated fruit juices are examples of shear thinning

materials. Materials like gelatinised starch dispersions show a shear thickening behaviour

where an increase in shear rate gives a more than proportional increase in the shear stress.

Many materials do not commence flow till the yield stress σ0 (the threshold of stress) is

exceeded, once they start flowing they may further exhibit shear thinning (thixotropic) or shear

thickening (antithixotropic) behaviour (Rao, 2010). The shear rate vs shear stress behaviour of

101

Newtonian and Non-Newtonian fluids is given in Figure 2.10.

Figure 2.10: Basic shear rate versus shear stress behaviour of Newtonian, shear thinning,

shear-thickening and Bingham and Herschel-Bulkley materials

Rheological parameters can be measured in a broad spectrum of materials ranging from solids

to liquids. Data available to the present day researchers has been based on the rheological

study of two extremes – Hookean solids and Newtonian liquids. The Hookean solid and the

Newtonian liquids are idealised examples of material behaviour as both establish a linear

relationship between stress and strain (or shear rate) based on direct proportionality,

irrespective of the applied force and amount of stress (Douglas, 2018). Modulus (Gʹ) is the

proportionality constant for solids,while viscosity (η) is the proportionality constant of liquids,

where modulus = σ / γ and viscosity = σ/(dγ/dt).

Majority of the food materials however, exhibit the ideal conditions of a Hookean solid or

Newtonian liquid only within a limited range of time and stress applied. More commonly, both

viscous (liquid like) and elastic (solid like) behaviour, termed as viscoelastic behaviour is

observed. Figure 2.11 shows the difference in behaviour between an ideal elastic solid, ideal

102

viscous liquid and a viscoelastic material (Shaw & MacKnight, 2018). When the time of

experimentation is short, all substances show a solid like behaviour. Whereas in an experiment

conducted over a longer period of time, the material is more likely to exhibit a viscous

behaviour. Dr. Marcus Reiner states that when given enough time, everything will flow

(Reiner, 1964). He proposed the following equation to describe this phenomenon

(5) 𝐷 = 𝜏 𝑇

Where D is the Deborah number, τ is the characteristic time of the material and T is the time

over which the experiment is run. When D < 1, liquid like behaviour is observed whilst, at D>1

materials exhibit a solid like behaviour (Barborik & Zatloukal, 2018).

Viscoelasticity of a material can be measured by using a rheometer to run small scale

deformation experiments. A bi-directional (oscillating) shear is used for viscoelastic

measurements as contrasted with the unidirectional oscillations used in viscosity

measurements. A sinusoidal stress or strain is applied on the material within what is known as

its ‘linear viscoelastic region (LVR)’. The pre-edtablished LVR is the domain where the

magnitudes of stress and strain are linearly related and the behaviour of the material is

independent of the stress or the strain applied. This implies that both the applied stress (or

strain) and the consequent response follow a sine wave and ‘excite’ the sample but does not

destroy its elastic structure. This is achieved by using a deformation that is relatively low or

slow, so as to maintain the sample’s structure close to its equilibrium. In large scale

deformation measurements used for measuring viscosity, permanent deformation of the sample

103

structure occurs.

Figure 2.11: The principle of oscillation viscometry. Applied strain versus time (a) and

resultant stress versus time that is measured in an elastic solid (b), Newtonian liquid (c) and

viscoelastic liquid (d).

Interpreting the data from rheological studies necessitates the knowledge of the terms: Elastic

modulus (Gʹ), loss modulus (Gʹʹ), tan delta (δ) and complex viscosity (η*). The elastic

properties of a sample, the strength of its network and its subsequent ability to withstand

deformation under applied stress or strain are all denoted by Gʹ, the storage modulus of the

sample. It takes into consideration the number and strength of the molecular interactions within

a sample. Whilst, Gʹʹ or the viscous modulus denotes the viscous properties of a sample and

deals with just the number of molecular interactions within the sample. If Gʹ > Gʹʹ, then the

sample is more like a solid than like a liquid. Conversely, when Gʹ < Gʹʹ the sample exhibits

104

liquid like properties(Dealy, Read, & Larson, 2018).

The complex modulus G*, which combines the storage and the loss moduli with equation 6

indicates the sample’s overall resistance to oscillatory shear.

G* = (Gʹ2 + Gʹʹ2)1/2 (6)

Another method for assessing the comparative importance of Gʹ and Gʹʹ is calculating the tan

δ or phase angle of the substance that is given by Equation 7. Tan δ is regarded as a measure

of viscosity.

Tan δ = Gʹʹ/Gʹ (7)

Materials with tan δ above 1 are extremely liquid-like whilst the value of tan δ is less than 1

for highly elastic substances.

In this study, TA Instruments’ Discovery Hybrid Rheometer (New Castle, DE, USA) was used

to perform small scale deformation oscillation experiments to obtain insight on the rheology of

materials (Figure 2.10). Keeping all variables constant while varying the following (one at a

time): time, temperature, frequency and strain provides important data about the materials

being examined (Puig-Rigall et al., 2018).

Strain Sweep

This is generally the first experiment conducted in rheological studies. It helps in determining

the LVR of the sample. The amplitude of the stress/strain is varied over a range and the

response of the sample is measured. The LVR is the region where the stress/strain and the Gʹ

are linearly related. The point where Gʹ shows a 10% deviation from the constant value, is

105

taken to be the end of the LVR region.

Figure 2.12: Discovery Hybrid Rheometer by TA Instruments at RMIT University

Time sweep

During time sweep, the sample is observed over a specific period while the other variables

(frequency, temperature, strain) are held constant. It enables the user to determine if there are

any changes in the properties of the sample over time. It indicates clearly the time taken by a

sample to reach an equilibrium state and also provides information regarding the storage

stability of the sample.

Temperature sweep

During the temperature sweep test, the sample is subjected to a predefine temperature range

while the other parameters are held constant. This allows the user to observe changes in the

sample’s properties as a function of temperature eg. gelation, gelatinisation, denaturation. It is

106

important to note that the scan rate employed for the test has a significant impact on the results.

Frequency sweep

Frequency, expressed as the inverse of time is the time needed to complete one oscillation. A

high frequency corresponds to a shorter time span and vice versa. The resultant spectrum of

the frequency sweep helps classify the sample that’s being examined into a weak gel, a strong

gel, a dilute solution or an entangled solution (Richardson & Kasapis, 1998).

2.2.2.4 Scanning Electron Microscopy

Scanning Electron Microscopy(SEM) was used to study the microstructure and phase topology

of the individual components in the biopolymer composites. It was employed to provide

tangible evidence of the phase separation and the presence of an architecture suitable for

rheological modeling.

SEM is capable of revealing surface information of a wide range of materials at the microspace

and nanospace which are inaccessible by light microscopy (magnification range 10 – 100,000x)

(Dudkiewicz et al., 2011). Conventional optical microscopy, on the other hand is limited to

examining particles 0.2 µm or greater in size (Cardott & Curtis, 2018). Another advantage of

SEM is its large imaging chamber which enables analysis of large samples. Furthermore, the

capacity to rotate, tilt and move the sample in three dimensions in the imaging chamber

increases its efficiency in probing the structure of the sample. SEM however, can’t capture

coloured images and has a lower resolution compared to transmission electron microscopy (Z.

Zhang, Law, & Lian, 2010). Typically, SEM consists of an electron gun that emits electrons, a

lens system that focuses the electrons, an electron detector that collects electrons from the

beam-specimen interaction, cathode ray tubes (CRT) that enable visual and photo recording

107

and miscellaneous electronics as shown in Figure 2.13.

Figure 2.13 Schematic diagram of a SEM (AMMRF, 2010)

The cathodic electron gun most commonly made of tungsten, emits electrons which are

accelerated to the anode through a voltage difference ranging between 0.1 to 50 keV. Pairs of

electromagnetic coils act as the focusing (condenser) lenses, the power of which depends on

the current passing through the lenses. The lenses focus the electron beam on to the specimen,

adjusting the spot size as much as necessary to improve the sharpness of the image. The

deflection coils in the objective lens are used to direct the beam across the sample in a raster

fashion scan. When the electron beam hits the specimen surface, and excites the atoms on the

surface creating signals. The signal created is specific to the structural features, chemical

proportion and crystallography characteristics of the specimen. The most common signals

generated by the SEM are secondary electrons (SE) and back scattered electrons (BSE) (Wang

& Petrova, 2012). Low energy secondary electrons (<50eV) provide the best resolution and

are best suited for generating the information about the sample topology. Whereas the

backscattered electrons have high energy (>50 eV), and reduce the resolution of the images.

108

BSE is most commonly used to probe into the sample composition (Fowler et al., 2018) .

Depending on the conditions of imaging, SEM is further classified into four main categories as

follows (Wandrol, Novák, Wu, Vesseur, & Vystavěl, 2018 ; Carter & Shieh, 2015):

Conventional (High Vacuum) SEM

This is the most common type of SEM where the sample is coated with a conducting

metal by a method called sputtering. A beam of electrons is used to excite the conducting

surface and create secondary electron emission. Imaging is done by recording either the SE or

BSE.

Low Vacuum SEM

When imaging non conducting samples like biological, archaeological or geological samples

among others which cannot be coated, SEM in a low vacuum mode is used. This works along

similar principle as the High vacuum SEM, other than the fact that the pressure in the chamber

can be adjusted (lowered) to prevent charging of the non conducting samples. Imaging is done

using BSE and the image quality is generally lower than that achieved in the high vacuum

mode due unpredictable electron discharge causing jumps or black spots in the image.

Cryo SEM

Cryo- SEM is specially used for imaging hydrated samples in their frozen state without altering

the sample properties. A cryo set up is added to the conventional high vacuum SEM that allows

the flash frozen samples to be viewed at sub-zero temperatures. The frozen samples can be

fractured or cracked to reveal internal structures. This is most suited for samples like hydrogels,

biofilms, foams, etc.

Environmental SEM (ESEM)

Imaging dynamic processes like melting, crystallisation or growth of biological specimens can

109

be made possible with use of an ESEM. The equipment set up allows the sample to be imaged

in its natural state without the need of any pre-processing steps. Sample temperature and the

relative humidity of the specimen chamber can be controlled allowing the user to wet, dry, heat

and cool the sample.

Figure 2.14: FEI Quanta 200 ESEM from FEI Corporate (Oregon, USA) at RMIT University

2.2.2.5 Confocal Laser Scanning Microscopy

The Confocal Microscope is based on the principle underlying Paul Nipkow’s Nipkow disk

which was a rotating disk with spirally arranged holes, which made it possible to scan and

110

transmit images via a telegraphic wire. The first application of confocal microscopy in food

science was published in 1987 (Heertje, 1987) and has witnessed an exponential rise in

applications since then.

Conventional optical microscopy, has been widely employed to get an insight of the food

structure at the meso scale and relate it to the texture and sensory attributes, along with the

physico-chemical properties of the final product since decades (Auty, 2013). However, light

microscopy requires pre-processing of the sample, such as freeze drying, sectioning, fixation,

or embedding, which may alter its microstructure (Plucknett et al., 2001). Confocal microscopy

is an adapted form of optical microscopy that has much in common with the epiflourescence

microscope. The development of confocal microscopy has helped overcome the problems in

light microscopy and has supplemented its advantages with the ability to capture images in

situ, image bulk specimens without sectioning them and obtain 3 dimensional images of the

specimen (N. Zhang et al., 2018).

In confocal microscopy, a diffraction limited laser beam from an optical fibre (generally from

an Argon source) is focused on a small subsurface spot on the sample via an objective lens as

shown in Figure 2.15. This enables imaging of a small section of the specimen as opposed to

the widefield illumination achieved by a typical optical microscope that illuminates the entire

sample. The incident laser excites the naturally present or labelled flourophores in the samples

at the given spot. The flourophores absorb the incident light and emit it at a longer wavelength.

The fluorescence emission is deflected via a dichoric mirror onto the condenser lens that

focuses it onto a photon detector via a pinhole. The photon detector, such as a photomultiplier

tube or CCD measures the intensity of the emitted fluorescence for this one spot. The signal is

then amplified and converted to picture element or pixel that makes up a digital image. The

focused spot is scanned across the sample, in the given focal plane in a raster fashion,

generating a 2D image pixel by pixel (Bayguinov et al., 2018). The pinhole is the distinguishing

111

feature of a confocal microscope as it controls the amount of fluorescence that reaches the

detector. The fluorescence from the out-of-focus planes is blocked by the pinhole stopping it

from reaching the photon detector. This makes it possible for the confocal microscope to focus

on any single focal plane along the z axis allowing emitted fluorescence from that specific

plane to reach the photon detector barring emissions from the out of focus planes. By adjusting

the pinhole, the z-depth or thickness can be adjusted. When the pinhole is completely open, the

confocal microscope is equivalent to a conventional scanning microscope (Blouin et al., 2018).

This 3D imaging technique, also known as Z-stack imaging has widely been used for

qualitative analysis, but the its use in quantitative analysis is scanty (Markus B. Dürrenberger

et al., 2001).

Figure 2.15: Illustration of the confocal imaging principle (solid lines = in-focus light; dashed

lines = out-of-focus light) (Franck, Hong, Maskarinec, Tirrell, & Ravichandran, 2007)

The laser is scanned across the sample in a raster fashion to form an image of various sizes,

112

most common being 512 x 512 or 1024 x 1024 pixels for a 2D image. A stepping motor fitted

to the microscope can be used to change the focal plane and a number of 2D images (x-y planes)

can be captured at consecutive focal planes through the z-axis. These multiple 2D images form

a Z – Stack that can be rendered into a 3D image. The presence of the pinhole facilitates the

CLSM to not only capture images with excellent resolution within the focal plane (≥ 0.25µm

in the x and y direction) but also obtains a good resolution between consecutive focal planes (≥

0.25µm in the z direction).

The resolution of the confocal microscope has the same limitations as those that affect light

microscopy. Theoretically, the diffraction limitation is given by the following equation (Zuo,

Liu, Cheng, Chen, & Tian, 2018):

∆x, ∆y = λ/2 n sinα (8)

Where α, n and λ are the aperture of the objective lens, refractive index of the sample and the

wavelength of the incident light respectively.

Along the optical axis, z-resolution, ∆z = λ/2 n sinα a minimum lateral resolution of 190nm

and an axial resolution of about 500nm is theoretically possible (Balashov et al., 2018). Under

the ideal conditions, in the absence of out of focus blur, close resolutions are attainable,

however, in practice, a z resolution of 750nm and an xy resolution of 250 nm is realistic (Auty,

2013).

Most commonly, optical sections are imaged in the z direction, in the xy planes perpendicular

to the laser beam. However, using the stepping principle, it is possible to capture images in the

xz or the yz plane. Further, scanning one image in time yields an x-t image. This makes it

possible to study any change in the fluorescent intensity in the given time frame. This feature

of the CLSM has been largely used in the biosciences to study chemical reactions in cells,

changes in ion concentrations, changes in pH and release of caged fluorescing probes due to

113

photo-activation. This technique is being explored in food studies and has been successful in

capturing in-situ images of gelation, aggregation and break up of β – lactoglobulin (Ko &

Gunasekaran, 2009; Olsson, Langton, & Hermansson, 2002) and baking of bread (Markus B

Dürrenberger, Stephan Handschin, Béatrice Conde-Petit, & Felix Escher, 2001) among others.

The advantages of CLSM over conventional light microscopy include (Pawley, 2006):

1) Ability to image bulk samples

2) Minimal sample preparation enables samples to be visualised in their native form

3) Optical sectioning makes it possible to capture a 3D Z Stack and subsequent 3D

reconstruction

4) Better resolution than conventional optical microscopy

5) In situ imaging enables capturing images of dynamic processes under controlled

environmental conditions using appropriate sample stages.

The disadvantages of CLSM include:

1) A flat surface is needed for imaging to ensure that the field of view has an even

illumination and emission signal

2) Flourescent labelling of the sample with a fluoroschrome may change intrinsic

properties and/or introduce non-native particles in the sample.

3) As the lateral resolution is limited to 250 nm by diffraction, colloidal food systems such

as casein micelles or interfacial films are difficult to properly resolve.

4) Common lasers have a limited number of excitation wavelengths, which occur over

narrow bands and risk overlapping and colocalisation

5) Since image is formed pixel by pixel as opposed to a snapshot in light microscopy,

imaging with CLSM is a much slower process. Imaging can be made faster by reducing

114

the sensitivity, resolution, etc.

2.2.2.6 Image Analysis

The Z-stack images captured by the CLSM were further analysed using two image analysis

software – FIJI with Java 6 and Imaris 9.0. Both the software employ distinct methods of

analysis and were hence used in the study to determine and compare their suitability in

accurately measuring the phase volume.

FIJI (FIJI Is Just ImageJ) is a free image processing software provided by ImageJ. It regards a

3D Z-stack as multiple 2D images and processes each image on the xy plane separately. The

result for each of the 2D images is then added and multiplied by the distance between two

consecutive xy planes to get volumetric measurements (Schneider, Rasband, & Eliceiri, 2012)

(Schindelin et al., 2012).

This is contrasted by Bitplane’s core scientific software module, Imaris. Imaris is made

exclusively for visualising and analysing 3D and 4D microscopy datasets. The Z stack images

captured using CLSM are rendered into 3D images by default and all measurements are done

on the single rendered 3D image (Li, Zhou, Voss, Quick, & LaBarbera, 2016).

The image analysis procedures followed and the results obtained are discussed and compared

115

in the experimental chapters.

References

AMMRF (2010). Electron Microscopy. Booklet.

Auty, M. (2013). Confocal microscopy: principles and applications to food microstructures

Food microstructures (pp. 96-P98): Elsevier.

Balashov, V., Efimov, A., Agapova, O., Pogorelov, A., Agapov, I., & Agladze, K. (2018). High

resolution 3D microscopy study of cardiomyocytes on polymer scaffold nanofibers

reveals formation of unusual sheathed structure. Acta Biomaterialia, 68, 214-222.

Barborik, T., & Zatloukal, M. (2018). Effect of die exit stress state, Deborah number, uniaxial

and planar extensional rheology on the neck-in phenomenon in polymeric flat film

production. Journal of Non-Newtonian Fluid Mechanics, 255, 39-56.

Barron, A. R., Allen, J., Bott, S., Jebb, M., Jiang, C., & Kienast, W. (2012). Physical methods

in chemistry and nano science. CONNEXIONS Rice University: Houston, TX, USA.

Bayguinov, P. O., Oakley, D. M., Shih, C. C., Geanon, D. J., Joens, M. S., & Fitzpatrick, J. A.

(2018). Modern Laser Scanning Confocal Microscopy. Current protocols in cytometry,

e39.

Blouin, S., Roschger, A., Varga, F., Misof, B., Spitzer, S., Roschger, P., & Klaushofer, K.

(2018). Confocal laser scanning microscopy—a powerful tool in bone research. Wiener

Medizinische Wochenschrift, 1-8.

Cardott, B. J., & Curtis, M. E. (2018). Identification and nanoporosity of macerals in coal by

scanning electron microscopy. International Journal of Coal Geology, 190, 205-217.

Carter, M., & Shieh, J. (2015). Chapter 5 – Microscopy. In M. Carter & J. Shieh (Eds.), Guide

to Research Techniques in Neuroscience (Second Edition) (pp. 117-144). San Diego:

Academic Press.

116

Cheremisinoff, N. P. (2018). Introduction to Polymer Rheology and Processing: 0: Crc Press.

Dealy, J. M., Read, D. J., & Larson, R. G. (2018). Structure and rheology of molten polymers:

from structure to flow behavior and back again: Carl Hanser Verlag GmbH Co KG.

Dole, M. N., Patel, P. A., Sawant, S. D., & Shedpure, P. S. (2011). Advance applications of

Fourier transform infrared spectroscopy. Int. J. Pharm. Sci. Rev. Res, 7(2), 159-166.

Douglas, J. (2018). Weak and Strong Gels and the Emergence of the Amorphous Solid State.

Gels, 4(1), 19.

Dudkiewicz, A., Tiede, K., Loeschner, K., Jensen, L. H. S., Jensen, E., Wierzbicki, R., . . .

Molhave, K. (2011). Characterization of nanomaterials in food by electron microscopy.

TrAC Trends in Analytical Chemistry, 30(1), 28-43.

Dürrenberger, M. B., Handschin, S., Conde-Petit, B., & Escher, F. (2001). Visualization of

Food Structure by Confocal Laser Scanning Microscopy (CLSM). LWT – Food Science

and Technology, 34(1), 11-17.

Dürrenberger, M. B., Handschin, S., Conde-Petit, B., & Escher, F. (2001). Visualization of

food structure by confocal laser scanning microscopy (CLSM). LWT-Food Science and

Technology, 34(1), 11-17.

Emadi, D., Whiting, L. V., DJURDJEVIC, M., Kierkus, W. T., & Sokolowski, J. (2018).

Comparison of newtonian and fourier thermal analysis techniques for calculation of

latent heat and solid fraction of aluminum alloys. Metallurgical and Materials

Engineering.

Fowler, C., Lynch, R., Shingler, D., Walsh, D., Carson, C., Neale, A., . . . Brown, A. (2018).

A Novel Electron-Microscopic Method for Measurement of Mineral Content in Enamel

Lesions. Archives of oral biology.

Franck, C., Hong, S., Maskarinec, S., Tirrell, D., & Ravichandran, G. (2007). Three-

117

dimensional full-field measurements of large deformations in soft materials using

confocal microscopy and digital volume correlation. Experimental Mechanics, 47(3),

427-438.

Gaca-Zając, K. Z., Smith, B. R., Nordon, A., Fletcher, A. J., Johnston, K., & Sefcik, J. (2018).

Investigation of IR and Raman spectra of species present in formaldehyde-water-

methanol systems. Vibrational Spectroscopy, 97, 44-54.

Guigo, N., & Sbirrazzuoli, N. (2018). Thermal Analysis of Biobased Polymers and

Composites. Handbook of Thermal Analysis and Calorimetry, 6, 399-429.

Gutiérrez Sama, S., Farenc, M., Barrère-Mangote, C., Lobinski, R., Afonso, C., Bouyssière,

B., & Giusti, P. (2018). Molecular Fingerprints and Speciation of Crude Oils and Heavy

Fractions Revealed by Molecular and Elemental Mass Spectrometry: Keystone

between Petroleomics, Metallopetroleomics, and Petrointeractomics. Energy & Fuels,

32(4), 4593-4605.

Jana, P. P., & Das, J. (2018). Precise estimation of glass transition and crystallization

temperatures of Zr55Cu30Ni5Al10 metallic glass using step-scan modulated

temperature differential scanning calorimeter. Thermochimica Acta, 660, 18-22.

Ko, S., & Gunasekaran, S. (2009). In situ microstructure evaluation during gelation of β-

lactoglobulin. Journal of food engineering, 90(2), 161-170.

Kodre, K., Attarde, S., Yendhe, P., Patil, R., & Barge, V. (2014). Differential scanning

calorimetry: A review. Research and Reviews: J. Pharmaceutical Analysis (RRJPA),

3(3), 11-22.

Komal Kumar, J., & Devi Prasad, A. (2012). Fourier transform infrared spectroscopy an

advanced technique for identification of biomolecules. Drug Invention Today, 4(12,

Co), 616-618.

Li, L., Zhou, Q., Voss, T. C., Quick, K. L., & LaBarbera, D. V. (2016). High-throughput

118

imaging: Focusing in on drug discovery in 3D. Methods, 96, 97-102.

Liu, Y., Yao, X., Fan, C., Wang, B., & Hu, J. (2018). Effect of a smectic liquid crystal polymer

as new β-nucleating agent on crystallization structure, melting, and rheological

behavior of isotactic polypropylene. Polymer Bulletin, 1-15.

Olsson, C., Langton, M., & Hermansson, A.-M. (2002). Dynamic measurements of β-

lactoglobulin structures during aggregation, gel formation and gel break-up in mixed

biopolymer systems. Food Hydrocolloids, 16(5), 477-488.

Pawley, J. B. (2006). Fundamental limits in confocal microscopy Handbook of biological

confocal microscopy (pp. 20-42): Springer.

Penner, M. H. (2010). Ultraviolet, visible, and fluorescence spectroscopy Food analysis (pp.

387-405): Springer.

Plucknett, K., Pomfret, S., Normand, V., Ferdinando, D., Veerman, C., Frith, W., & Norton, I.

(2001). Dynamic experimentation on the confocal laser scanning microscope:

application to soft‐solid, composite food materials. Journal of microscopy, 201(2), 279-

290.

Puig-Rigall, J., Obregon-Gomez, I., Monreal-Pérez, P., Radulescu, A., Blanco-Prieto, M. J.,

Dreiss, C. A., & González-Gaitano, G. (2018). Phase behaviour, micellar structure and

linear rheology of tetrablock copolymer Tetronic 908. Journal of colloid and interface

science, 524, 42-51.

Rao, M. A. (2010). Rheology of fluid and semisolid foods: principles and applications:

Springer Science & Business Media.

Reiner, M. (1964). The deborah number. Physics today, 17(1), 62.

Richardson, R. K., & Kasapis, S. (1998). Rheological methods in the characterisation of food

119

biopolymers Developments in Food Science (Vol. 39, pp. 1-48): Elsevier.

Schindelin, J., Arganda-Carreras, I., Frise, E., Kaynig, V., Longair, M., Pietzsch, T., . . .

Schmid, B. (2012). Fiji: an open-source platform for biological-image analysis. Nature

methods, 9(7), 676.

Schneider, C. A., Rasband, W. S., & Eliceiri, K. W. (2012). NIH Image to ImageJ: 25 years of

image analysis. Nature methods, 9(7), 671.

Shaw, M. T., & MacKnight, W. J. (2018). Introduction to polymer viscoelasticity: John Wiley

& Sons.

Stuart, B. (2005). Infrared spectroscopy. Kirk‐Othmer Encyclopedia of Chemical Technology.

Thomas, L. C. (2005). MDSC Paper# 2 Modulated DSC Basics; Calculation and Calibration

of MDSC Signals: New Castle: TA Instruments.

Türker-Kaya, S., & Huck, C. (2017). A review of mid-infrared and near-infrared imaging:

principles, concepts and applications in plant tissue analysis. Molecules, 22(1), 168.

Verdonck, E., Schaap, K., & Thomas, L. C. (1999). A discussion of the principles and

applications of Modulated Temperature DSC (MTDSC). International journal of

pharmaceutics, 192(1), 3-20.

Wandrol, P., Novák, L., Wu, M., Vesseur, E. J., & Vystavěl, T. (2018). In-situ Heating in SEM

and FIB/SEM Systems. Microscopy and Microanalysis, 24(S1), 336-337.

Doi:10.1017/S1431927618002179

Wang, Y., & Petrova, V. (2012). Scanning electron microscopy. Nanotechnology Research

Methods for Foods and Bioproducts, 103-126.

Zhang, N., Thompson, C. E., Townend, I. H., Rankin, K. E., Paterson, D. M., & Manning, A.

J. (2018). Non-destructive 3D imaging and quantification of hydrated biofilm-sediment

aggregates using X-ray micro-computed tomography. Environmental Science &

120

Technology.

Zhang, Z., Law, D., & Lian, G. (2010). Characterization methods of encapsulates

Encapsulation Technologies for Active Food Ingredients and Food Processing (pp.

101-125): Springer.

Zuo, R., Liu, W., Cheng, H., Chen, S., & Tian, J. (2018). Breaking the Diffraction Limit with

Radially Polarized Light Based on Dielectric Metalenses. Advanced Optical Materials,

121

1800795.

Chapter 3 Quantitative analysis of the phase volume of agarose- canola oil gels in comparison to blending law predictions using 3D imaging based on confocal laser scanning microscopy

Mhaske, P., Condict, L., Dokouhaki, M., Katopo, L., & Kasapis, S. (2019).

Quantitative analysis of the phase volume of agarose-canola oil gels in comparison to

blending law predictions using 3D imaging based on confocal laser scanning microscopy. Food

122

Research International, 125, 108529.

Food Research International 125 (2019) 108529

Contents lists available at ScienceDirect

Food Research International

journal homepage: www.elsevier.com/locate/foodres

Quantitative analysis of the phase volume of agarose-canola oil gels in comparison to blending law predictions using 3D imaging based on confocal laser scanning microscopy

⁎

Pranita Mhaske, Lloyd Condict, Mina Dokouhaki, Lita Katopo, Stefan Kasapis

School of Science, RMIT University, Bundoora West Campus, Plenty Road, Bundoora, VIC 3083, Australia

A R T I C L E I N F O

A B S T R A C T

Keywords: Blending law analysis Phase volume Confocal laser scanning microscopy Z-stack 3D imaging

Studying the phase behaviour of composite gels facilitates understanding of their structural and textural prop- erties at low and intermediate levels of solids. In this work, the phase behaviour of a model system of agarose including various concentrations of canola oil was studied. This was pursued using a variety of techniques including SEM, FTIR, microDSC and dynamic oscillation in-shear. The structural studies recorded strong, con- tinuous agarose networks supporting soft, discontinuous canola oil inclusions, with increasing levels of canola oil strengthening the composite system. A novel confocal laser scanning microscopy (CLSM) method for quantita- tive in situ examination of the oil phase volume was developed using three-dimensional (3D) imaging and image analysis software - FIJI and Imaris. Microscopic observations were assessed in relation to theoretical predictions from rheology-based blending-law analysis. Quantitative outcomes from the combined 3D imaging and image analysis are in close agreement with the volume predictions for the oil phase obtained from the isostrain blending law indicating the suitability of this approach in quantifying the phase behaviour of composite ma- terials. The results of this work indicate that the developed microscopic method shows promise and could be used in the determination of phase volume in more complex and industrially relevant systems.

1. Introduction

2018) and the tertiary system of agarose-gelatin-lipid (Shrinivas, Kasapis, & Tongdang, 2009).

synthetic polymer

Accurate determination of the phase volume for each component in the aqueous mixture is critical, since it determines the structural properties of the commercial formulation. Blending laws aim to achieve this but they reﬂect an indirect approach that relies on certain as- sumptions (e.g. the massive increase in eﬀective molecular weight of the biopolymer at the onset of gelation) and a time-consuming method that requires considerable rheological work. On the other hand, con- focal laser scanning microscopy (CLSM) may oﬀer a direct and rapid alternative to phase volume determination. One of CLSM's advantages is minimal sample manipulation, hence its capability to analyse the original state of a material without the pre-processing steps of freeze drying, sectioning or ﬁxation, which are typically employed in other techniques (e.g. electron or optical microscopy) and may alter micro- structure (Auty, 2013).

CLSM can capture a sequence of two-dimensional (2D) optical sec- tions through a soft matrix to produce a three-dimensional (3D) image. Multiple 2D images are obtained by moving a focal plane, i.e. an x-y plane, on which the laser is focused, along the z-axis in steps of

By and large, industrial processing results in complex systems con- taining two or more components with well-deﬁned physicochemical properties, with end products often possessing structural characteristics that are distinct from the individual constituents (Sikorski, Pokorny, & Damodaran, 2008). Examples include processed food, ceramics and plastic composites to name but a few. In the absence of a thermo- dynamic drive to heterotypic association, components tend to segregate leading to phase separation, which increases the “eﬀective” con- centration of each constituent in its phase (Morris, 2009). Drawing from the “sophisticated” research (Nielsen, 1974; Takayanagi, Harima, & Iwata, 1963), workers in the area of bioma- terials adapted the blending law analysis to predict the mechanical properties of composites from the shear modulus of the individual components in relation to their eﬀective concentration and phase vo- lume. Successful applications have been documented in various systems including agarose-whey protein (Katopo, Kasapis, & Hemar, 2012), bovine serum albumin-gelatin (Semasaka, Katopo, Buckow, & Kasapis, 2018), soy protein-wheat gluten (Dekkers, Emin, Boom, & van der Goot,

⁎

Corresponding author. E-mail address: stefan.kasapis@rmit.edu.au (S. Kasapis).

https://doi.org/10.1016/j.foodres.2019.108529 Received 18 November 2018; Received in revised form 18 May 2019; Accepted 26 June 2019

123

P. Mhaske, et al.

Food Research International 125 (2019) 108529

speciﬁed thickness through the depth of the sample. This technique, also known as Z-stack imaging, oﬀers unique insights into the material's true 3D structure. Numerous studies have demonstrated the applic- ability of CLSM in phase-separated mixtures including wheat protein- starch (Peighambardoust, Dadpour, & Dokouhaki, 2010), gelatin- starch-olive oil (Firoozmand & Rousseau, 2014) and whey protein iso- late-locust bean gum (Ur-Rehman et al., 2016), but its use has been focused on qualitative observations.

concentrations of canola oil were freeze dried using Labconco FriZone Triade freeze dryer (Kansas City, Missouri, USA) at −40 °C for 3 days. Then, they were placed on 12.6 mm carbon coated aluminium pins and gold coated. Surface micrographs were obtained using the FEI Quanta 200 ESEM (Hillsboro, Oregon, USA). An overview of the gel network was captured in the low vacuum mode with an accelerated electron beam of 25 kV and spot size of 5, with the working distance being about 10 mm at 50× magniﬁcation. Details of the agarose-canola oil topology were captured at a higher magniﬁcation of 400× in the high vacuum mode also using an accelerated electron beam of 25 kV and a spot size of 5.

2.2.3. Fourier transform infrared spectroscopy (FTIR)

A Perkin Elmer Spectrum Two FTIR spectrometer (Norway, CT, USA) ﬁtted with a GladiATR from Pike Technologies (Wisconsin, United States) was used to obtain absorbance spectra for the samples of this −1 investigation. Measurements were carried out from 400 to 4000 cm −1 and averaged using 64 scans. The ﬁnal spectra at a resolution of 4 cm of agarose-canola oil mixtures were collected by subtracting the water spectrum, and measurements were done in triplicate.

2.2.4. Diﬀerential scanning calorimetry (DSC)

Some examples of CLSM work, coupled with image analysis soft- ware, have been reported for the visualisation of water droplets in table spreads (Van Dalen, 2002), starch granules in guar gum-maize starch mixtures (Nagano, Tamaki, & Funami, 2008) and the fractal dimension of proteins in wheat dough (Li, Liu, Wu, & Zhang, 2017). Earlier at- tempts to develop CLSM-based techniques that can directly determine phase volume in mixtures were restricted to viscous aqueous solutions (Blonk, Van Eendenburg, Koning, Weisenborn, & Winkel, 1995) or single 2D images (Lorén, Langton, & Hermansson, 1999), hence they have not had widespread appeal that prevented further development. In this communication, a novel CLSM-based approach is developed using 3D imaging and quantitative image analysis. For the ﬁrst appli- cation of this approach, we have chosen a mixture of agarose hydrogel and canola oil that remains in the liquid state at the experimental temperatures of this investigation (5 to 35 °C). The aqueous agarose network is thermodynamically incompatible with canola oil ensuring conditions of phase separation in the binary mixture (McClements, 2015). The aim is to estimate the phase volume of canola oil globules dispersed homogeneously within the supporting matrix of an agarose gel.

Thermal analysis of agarose-oil mixtures was performed using a microDSC Setaram VII (Seturau, Caluire, France). Samples of about 700 mg were transferred into aluminium cells and sealed, with the re- ference cell containing water at an equal weight. Both cells were then deposited into the instrument chamber and equilibrated at 25 °C for 60 min prior to ramping the temperature to 80 °C, cooling to 5 °C and then heating again to 80 °C. A constant scan rate of 1 °C/min was used throughout and all experiments were carried out in duplicate returning eﬀectively overlapping thermograms.

Results from the microscopic experimentation will be compared with estimates from the corresponding rheological tests thus aﬀording an evaluation of the predictions of rheological modelling on the phase behaviour of this model system. Successful application of this protocol to the agarose-canola oil gel will allow expansion to other systems of commercial interest, including mixtures with a solid lipid phase, sus- pending reﬁned wood pulp (for example, microcrystalline cellulose), and matrices with two phase separated biopolymer networks where it is important but diﬃcult to record directly the volumes and composition of the two phases; these, however, determine the textural character- istics of formulations (for example, blends of agarose and gelatin).

Additional measurements on canola oil were carried out using a modulated DSC by TA Instruments, Q2000 (New Castle, DE, USA). The instrument is ﬁtted with a refrigerated cooling unit to reach tempera- tures well below 0 °C and a nitrogen purge cell with a ﬂow rate of 20 ml/min. About 11 mg of the oil was hermetically sealed in an alu- minium pan with an empty pan being used as the reference. It was equilibrated at 25 °C for 60 min followed by successive heating and cooling routines between −80 and 80 °C at a scan rate of 1 °C/min. A modulated amplitude of ± 0.53 °C for 40 s was used throughout. Work was performed in duplicate producing consistent results.

2. Materials and methods

2.1. Materials

2.2.5. Rheological measurements

Agarose (Type-1) was purchased from Sigma-Aldrich (Sydney, Australia). The powder has an oﬀ-white colour, with its moisture and sulphate content being < 10 and 0.15%, respectively, according to the speciﬁcation. Canola oil was purchased from Manildra supplier (Gladesville, Australia), and has a density of 0.918 g/ml.

2.2. Methods

2.2.1. Sample preparation

Small-deformation dynamic oscillation in-shear was used to mea- sure the elastic modulus (G′) and viscous modulus (G″) of both agarose and canola oil as well as their binary mixtures. Measurements were conducted with TA Instruments' Discovery Hybrid Rheometer (New Castle, DE, USA). A 40 mm diameter parallel plate geometry with 1 mm gap was employed. Samples were loaded on the Peltier plate at 45 °C and coated with polydimethylsiloxane to minimize evaporation. They were then cooled from 45 to 5 °C, held at 5 °C for 30 min followed by a frequency sweep at 5 °C from 0.1 to 100 rad/s and a heating run to 80 °C. A constant rate of 2 °C/min and a strain of 0.01%, which was within the linear viscoelastic region (LVR), were used throughout. Triplicate tests were carried out yielding congruent outcomes.

Binary mixtures of agarose and canola oil were prepared by adding a ﬁxed amount of the polysaccharide power that would have produced 1% (w/w) agarose in the aqueous preparation, to the remainder of water once various concentrations of the oil (0, 1, 3, 5, 7, 9 and 11%, w/w) were added to a total of a hundred. In doing so, the agarose so- lution was made ﬁrst by dissolving the powder in milliQ water at 80 °C with agitation for 10 min followed by cooling to 45 °C. Canola oil was then added to the agarose solution at this temperature followed by homogenisation using an IKA T25 Ultra Turrax homogeniser (Wilmington, North Carolina, USA) at 3000 rpm for 30 s.

2.2.2. SEM analysis

2.2.6. Confocal laser scanning microscopy (CLSM) 2.2.6.1. Sample preparation and image acquisition. Agarose-canola oil composites containing various concentrations of the oil were stained with Nile red (0.005 g/l), which is a sensitive stain for the detection of oil droplets, and then were placed into the well of a precooled Coverwell Press-Seal imaging chamber with 20 mm diameter and 2.8 mm depth (Sigma-Aldrich, NSW, Australia). Precooling of the chamber at 4 °C allowed a rapid and homogenous gel formation in situ following deposition of the warm solution. A Nikon A1R laser on an ECLIPSE Ti-E inverted microscope (Minato, Tokyo, Japan) was used to

Gel cubes of about 10 mm3 in size of agarose with varying

124

P. Mhaske, et al.

Food Research International 125 (2019) 108529

3,6-anhydro-galactose

−1), CeC bending (1396 cm (1068 cm bridges

generate confocal images. An excitation wavelength of 488 nm was used to induce ﬂuorescence and an emission ﬁlter of 500–530 nm was applied to observe the stained oil droplets and collect images (Auty, 2013).

−1), CeH a series of chemical moieties, i.e. OeH stretching (2980 cm −1), CeOeC vibration vibration (2900 cm −1), of esters sulfate −1), and CeH angular deformation of anomeric carbon (1246 cm −1) (Almrhag et al., 2013; Samiey & Ashoori, 2012; Wang, Guo, (877 cm & Yang, 2011; Zamora-Mora, Velasco, Hernández, Mijangos, & Kumacheva, 2014).

Each 3D image was composed of 15 consecutive z-series of 2D images collected using a 10× objective and a z-axis step of 5.72 μm. The 3D image was generated with the xyz dimensions of 1279 μm × 1279 μm × 80 μm, respectively. Stained oil droplets were observed at up to 80 μm from the surface due to the signiﬁcant loss of ﬂuorescent signal at greater depths. A Z-stack of 2D images from nine sections across each sample was captured to avoid the eﬀect of any inhomogeneous distribution of oil droplets.

Canola oil produces characteristic peaks representing symmetrical and asymmetrical CeH stretching vibration from CH2 groups of organic −1, respectively), C]O stretching of fatty acids (2856 and 2920 cm −1), CH3 aliphatic group bending vibration aliphatic esters (1746 cm −1), CH2 bending vibration or CeO stretching vibration of (1452 cm −1), and out of plane bending vibration of cis- aliphatic esters (1155 cm −1) (Rohman & Man, 2010; Zhang, Ma, Miao, HC]CHe (721 cm Tuchiya, & Chen, 2015). Mixtures of agarose and canola oil reproduce the characteristics of the individual components, with ﬁngerprints from agarose dominating the interferograms. Since the mixtures show marked similarity with the individual components both in band wa- venumber and intensity, the earlier assumption in relation to the ab- sence of speciﬁc molecular interactions between the two components stands.

2.2.6.2. Image analysis. Obtained images were analysed using two image analysis software: FIJI with Java 6 software (Madison, Wisconsin, USA) and Imaris 9.1.0 (Bitplane, Belfast, Northern Ireland) in parallel. The former utilised the iterative ‘Default’ automatic global thresholding method (Calvard & Ridler, 1978), with the ‘analyse particle’ function then being used to estimate pixels denoting the lipid phase prior to multiplication by the Z-depth to obtain the volume fraction. The latter processed the obtained images using the ‘surface’ feature, the 3D images were then thresholded using the default method and the phase volume was estimated for the thresholded images.

2.2.7. Statistical analysis

Volume measurements of the oil phase were carried out in duplicate and data are expressed in mean ± standard deviation. Results were evaluated by one-way analysis of variance (ANOVA) using Minitab 18 (Minitab Inc., Pennsylvania, USA). Signiﬁcance diﬀerences were de- ﬁned as p < .05 with the Tukey test.

3. Results and discussion

Complementary characterization of the phase behaviour in aqueous preparations of agarose-canola oil mixtures is performed using micro diﬀerential scanning calorimetry. Fig. 3 depicts thermograms of the polysaccharide and mixtures that were cooled from 80 to 5 °C at 1 °C/ min. The former exhibits a relatively sharp exothermic peak with a mid- point transition temperature of 26 °C signifying the coil-to-helix tran- sition traversing the solution-to-gel consistency. Lipid inclusion also reveals the presence of a single peak, which remains unaltered in terms of temperature band and cooperativity. Increasing levels of canola oil generate thermal events of comparable enthalpy change arguing that the gelation process of the polysaccharide remains unaﬀected. Results further support the statement of microscopic phase separation between agarose and canola oil in the blend.

3.1. Experimental observations on the structural characteristics of agarose- canola oil mixtures

Scanning electron microscopy has been used in the past to unveil complex composite morphology (Shrinivas et al., 2009). In the present study it is employed to provide tangible evidence of phase separation in agarose-canola oil mixtures. Fig. 1a exhibits the SEM image of freeze- dried agarose gels at 50× magniﬁcation, where it is evident that there is a considerable degree of porosity in a honeycomb pattern. At the same magniﬁcation, we observed a similar mesh-type assembly in the presence of oil arguing for a continuous agarose phase in the composite but the oil droplets were not visible (images not shown). This suggests that the size of oil droplets is much smaller than the agarose pores and as such it requires a higher level of magniﬁcation for observation.

The absence of a noticeable calorimetric event attributed to the lipid constituent of the mixture prompted us to examine it separately from the polysaccharide. Thermal analysis of canola oil using microDSC fails to record a visible peak during cooling from 80 to 5 °C. A plausible explanation to this outcome lies in the composition of triacylglycerols (TAGs) in canola oil, which contains a high proportion of unsaturated TAGs crystallizing at very low temperature (Marikkar, Ghazali, Che Man, & Lai, 2002; Tang & Marangoni, 2007). To examine this, we employed modulated diﬀerential scanning calorimetry that was able to look for potential thermodynamic transitions at subzero temperatures. We observed an intense exothermic peak, culminating at about −57 °C, during a cooling routine between 80 and - 80 °C, as recorded in Fig. 3. This should correspond to the aggregation of unsaturated TAGs into clusters and the subsequent formation of a crystalline network. The subzero crystallisation temperature of canola oil shown in this work strongly supports our statement that canola oil exists as liquid inclu- sions in the agarose gel.

The agarose structure at higher magniﬁcation (400×) reveals a smooth featureless surface in Fig. 1b. In contrast, oil addition generates spheroidal clumps that correspond to the oil phase being dispersed within the agarose network (Fig. 1c,d). There is a noticeable alteration in the polysaccharide architecture showing a decreased level of smoothness with increasing oil concentration from 3 to 11% (w/w). It is worth noting that the size of the lipid domains remains fairly consistent (i.e. < 100 (80 ± 20) μm in diameter) with varying concentration arguing for limited lipid coalescence. Overall, the agarose-canola oil blends demonstrate a continuous polysaccharide structure supporting homogeneously dispersed oil globules. This topology provides the basis for quantitative analysis of the phase behaviour of the agarose-canola oil gel via theoretical modelling discussed in the next section.

Supramolecular evidence on the structure of the materials of this investigation is aﬀorded by mechanical measurements, which en- compassed the individual components as a prelude to work on their mixtures. The general pattern of elastic modulus development for the single agarose preparation during controlled cooling at 2 °C/min from 50 to 5 °C is shown in Fig. 4. There is a sharp increase in the values of G′ at temperature below 30 °C indicating a gelation process in which the long and ﬂexible coils of the polysaccharide transform into rigid double helices via hydrogen bonds. Subsequent aggregation of the double he- lices yields a three-dimensional network (Katopo et al., 2012). In con- trast, canola oil shows a Newtonian behaviour of a very low viscous modulus (G″ ~ 0.02 Pa), with the complex dynamic viscosity (η*) re- maining constant at about 0.16 Pa.s at 5 °C.

Following the measurements on single components, a series of binary mixtures was prepared by keeping the agarose at 1% (w/w) and

Results obtained from SEM indicate phase separation between polysaccharide and lipid. To conﬁrm the absence of molecular inter- actions between the two constituents, we carried out Fourier transform infrared spectroscopy, and the interferograms for agarose, canola oil and their blends are illustrated in Fig. 2. The agarose spectrum exhibits

125

P. Mhaske, et al.

Food Research International 125 (2019) 108529

Fig. 1. SEM images of (a) 1% (w/w) agarose at 50× magniﬁcation, (b) 1% (w/w) agarose at 400× magniﬁcation, (c) 1% (w/w) agarose with 3% (w/w) canola oil at 400× magniﬁcation, and (d) 1% (w/w) agarose with 11% (w/w) canola oil at 400× magniﬁcation.

1.8

1068

2980

2900

1396

1246

1.6

877

1.4

1.2

1.0

0.8

e c n a b r o s b A

0.6

1746

2920

2856

1155

0.4

721

1452

illustrates the cooling proﬁles of blends recorded under the same ex- perimental routine as for the single systems. These follow similar pat- terns with the agarose network, an outcome that supports phase se- paration behaviour with the continuous agarose network suspending discontinuous inclusions of canola oil. Results from various techniques in Figs. 1 to 4 are consistent, and it is evident that the storage modulus of the composite gels increases with higher oil addition; G′ records values from 103.80 to 104.03 Pa at 1 to 11% (w/w) oil, respectively. Such enhancement in the overall rigidity of the mixture should be attributed to the rising agarose concentration in the polymer phase (eﬀective concentration), which is the result of volume exclusion between aqu- eous and lipid phase.

0.2

0.0

3500

3000

2500

1500

1000

500

2000 Wavenumber (cm-1)

Fig. 2. FTIR spectra of 1% (w/w) agarose with 0, 1, 3, 5, 7, 9 and 11% (w/w) canola oil arranged from top to bottom (solid lines), and canola oil (dashed line).

varying the canola oil from 1 to 11% (w/w) in concentration steps of increases the eﬀective concentration of the 2% which, of course, polysaccharide in the gradually diminishing aqueous phase. Fig. 4 also

This outcome is distinctly diﬀerent to the results reported by Shrinivas et al. (2009) who have shown that in tertiary systems of 1% agarose, 10% gelatin and variable concentrations of soybean oil at 25 °C, when gelatin dominates the phase topology, the strength of the network gradually decreases on increasing concentrations of oil. These diﬀerences should be attributed to the much steeper calibration curve of agarose when compared to gelatin; for example, the log G' value of gelatin at a concentration of 30% (w/w) is ~ 4.5 (Pa), i.e. considerably lower than that of agarose at concentrations of < 5% (w/w) with a log G' value of ~5 (Pa). This seems to indicate that the initial strength of the continuous matrix has a substantial eﬀect on how the addition of oil aﬀects the mechanics of the aqueous/lipid composite system.

126

P. Mhaske, et al.

Food Research International 125 (2019) 108529

2.25

900

850

1.75

)

800

)

W m

(

750

1.25

W m

700

s e r u t x i m

0.75

650

600

0.25

550

500

l i o / e s o r a g a d n a

-0.25

( l i o a l o n a c f o w o l f t a e H

450

e s o r a g a

-0.75

400

350

f o w o l f

-1.25

300

t a e H

-1.75

250

-80

-60

-40

-20

Temperature (°C)

Fig. 3. Cooling proﬁles of 1% (w/w) agarose with 0, 1, 3, 5, 7, 9 and 11% (w/w) canola oil arranged from top to bottom (solid lines), and canola oil (dotted line) obtained using DSC at a scan rate of 1 °C/min.

ϕ x

ϕ y

(2)

3.2. Theoretical modelling of the phase behaviour in agarose-canola oil composites

where, G'c, G'x and G'y are the shear elastic moduli of the composite, X- phase constituent and Y-phase constituent, respectively, with ϕx and ϕy being the phase volumes of the two constituents. Eq. (1) is applicable to a format of a strong matrix supporting a soft ﬁller (isostrain relationship in Morris, 2009), which is the topology of the current system; agarose network suspending liquid inclusions of canola oil.

To put the experimental observations of the preceding section on a quantitative footing, we constructed a calibration curve of elastic modulus for agarose networks at 5 °C. As shown in the inset of Fig. 4, increasing polysaccharide concentration from 1.0 to 2.5% (w/w) yields a double logarithmic plot with a good linear ﬁt (r2 = 0.995). This in- formation is vital for modelling work that relates the overall mechan- ical property of the composite gel to the modulus of the two component phases via the so-called “blending laws”. Their mathematical expres- sion, being appropriate for this work, is as follows (Semasaka et al., 2018):

′

′ = C

′ + x

ϕ G x

ϕ G y

(1)

Fig. 5 displays an example of the application of the isostrain blending law (designed to model hard matrices supporting soft ﬁller particles), for 1% (w/w) agarose with 3% (w/w) canola oil, but results are consistent for all combinations in the mixture. It shows a compu- terised calculation of shear modulus of the continuous agarose phase (G'ag) for all possible solvent (water) distributions in the agarose phase (Sag), i.e. from 0 to 1, in steps of 0.01. The values of the shear modulus

4.2

3.8

3.4

5.5

) a P ( '

3.0

5.0

) a P ( '

G g o L

4.5

2.6

y = 3.805x + 3.8521 R² = 0.995

G g o L

4.0

3.5

2.2

0.1

0.2

0.3

0.4

0.5

Log agarose concentration (%)

1.8

20 Temperature (°C)

Fig. 4. Cooling proﬁles of G' for 1% (w/w) agarose with 0 (○), 1% (■), 3% (x), 5% (●), 7% (Δ), 9% (-) and 11% (♦) (w/w) canola oil from 50 °C to 5 °C at a scan rate of 2 °C/min. The inset corresponds to a calibration curve of G' at 5 °C as a function of agarose concentration.

127

P. Mhaske, et al.

Food Research International 125 (2019) 108529

Fig. 5. Modelling the phase topology of 1% (w/w) agarose plus 3% (w/w) canola oil using the isostrain blending law. Predicted storage modulus of agarose (G'ag) and composite (G'C(BL)) are illustrated by dashed and solid lines, respec- tively, and the experimental composite modulus (G'exp) is shown by a dotted line. Sag is the solvent content of the agarose phase. The inset shows experimental G' values (■) and calculated phase volumes of canola oil (▲) for 1% (w/w) agarose with increasing oil concentration (solid line ﬁt on the inset and the oil phase volume are derived from the isostrain blending law).

analysis of phase behaviour in agarose-canola oil composites, which involves selective ﬂuorescent staining of oil droplets followed by con- focal imaging using Z-stack function and processing by image analysis software. A typical example of a single optical section (2D image) in the xy-plane captured from a section of 1% agarose with 5% canola oil gel is depicted in Fig. 6a. The oil droplets appear as purple spheres whereas the black regions correspond to the unstained agarose network. Fur- thermore, the droplets that are in the focal plane display bright purple colour whilst those that are outside the focal plane show a reduction in the colour intensity. Consecutive 2D images of the sample were ac- quired at a constant z-axis step of 5.72 μm (Fig. 6b) and the Z-stack image was rendered into a 3D image shown in Fig. 6c. To obtain esti- mations of the oil phase volume, the stack of 2D images (Fig. 6b) was processed using FIJI, whilst Imaris analysed the 3D image (Fig. 6c).

of the composite (G'C(BL)), which are calculated via the isostrain model of Eq. (1), and the experimental composite modulus (G'exp) are also depicted as a function of Sag. Gratifyingly, calculated and experimental values of the composite shear modulus intersect at the Sag value of 1.0, an outcome which indicates that the entirety of the solvent is dis- tributed within the agarose phase, hence demonstrating the capacity of the blending law to predict physically realistically results. This is also shown for the whole experimental range of increasing lipid con- centrations in the inset of Fig. 5, which illustrates a perfect ﬁt between experimental moduli, obtained following frequency sweeps of the composites at 5 °C, and predicted moduli from blending-law analysis. Knowledge of solvent partition is crucial in estimating the phase volume of the individual components in composite systems, since it also predicts the eﬀective (ﬁnal) polymer concentration in its phase; generic relationships that predict solvent distribution for phase separated sys- tems are given in the following framework (Kasapis, Morris, Norton, & Clark, 1993; Morris, 2009):

x = +

S (

)

(3)

y = +

((1

−

)

(4)

ϕ x

tw x +

(5)

FIJI is an open source analysis software that processes individual 2D images, which make up a Z-stack (Schindelin et al., 2012). The 8-bit 2D image provided by CLSM comprises numerous pixels with the value of each pixel ranging from 0 to 255 (Luo et al., 2018). Pixels carry the intensity value of images, hence being vital to their quantiﬁcation, an unstained sample results in a completely black image with the pixel intensity of all the image pixels being 0. This also acts as the negative control for imaging and the baseline calibration for image analysis. To correctly determine the number of pixels constituting an object (i.e. the oil droplet presently), an accurate description of the object edge is necessary. Thresholding was selected as the method for estimating the boundary of an object (Unnikrishnan, Pantofaru, & Hebert, 2007), since it can deﬁne a precise pixel intensity as the cut-oﬀ point between oil droplet and agarose background. Choosing a threshold that is too low leads to false positives, i.e. overestimating the phase volume, whereas too high a threshold causes false negatives that overlook oil droplets emitting a low-intensity signal. The threshold can be deﬁned either manually or by applying automated algorithms. The former relies on the operator's practical skills for visually detecting the threshold and can be arbitrary, with inter-operator subjectivity impacting the re- producibility of results (Bergouignan et al., 2009). In contrast, auto- matic thresholding, which uses an algorithm to set the threshold, eliminates subjectivity and as such it was employed in this work.

In Eqs. (3) and (4), w, x, y and Sx are the total weight of water in the system, weight of polymer X, weight of polymer Y and the solvent fraction in X-phase, respectively. In Eq. (5), twx and twy refer to the total weights of X- and Y-phase, respectively. The phase volumes (ϕx and ϕy), in Eqs. (5) and (2), are deﬁned by the total weight of the phases. The phase volume of canola oil in the mixtures, calculated from the isostrain blending law based on the Sag values of 1.0, is also plotted against the experimental oil concentration in the inset of Fig. 5. It was found that the predicted phase volume of the oil using Eqs. (3) to (5) tracks well its actual concentration in the mixture across the entire experimental range, hence its phase volume depends solely on its weight. Outcomes of theoretical modelling from rheological algorithms will be compared with those from 3D imaging and image analysis in the following sec- tion.

3.3. Development of a CLSM protocol and image analysis for agarose- canola oil composites

There are two categories of automatic thresholding that could be employed, i.e. global and local, with the former utilising a single threshold value for the entire image while the latter employs a diﬀerent threshold value for each pixel depending on the intensity of adjacent pixels. Global thresholding works well for high quality images where

We developed a CLSM microscopic protocol

for quantitative

128

P. Mhaske, et al.

Food Research International 125 (2019) 108529

Fig. 6. (a) 2D, (b) Z-stack, and (c) 3D CLSM images of 1% (w/w) agarose plus 5% (w/w) canola oil. The oil droplets are either in purple (a-b) or grey (c), and the agarose is in black (a-c). (For interpretation of the references to colour in this ﬁgure legend, the reader is referred to the web version of this article.)

estimated oil phase volume is closely matched to its experimental concentration in the composites, and Table 1 demonstrates that the best estimation is provided by the Default algorithm.

there is a good contrast between object and background pixels, while local thresholding is better suited for poorly illuminated images with either low contrast or multiple colour spectra (Sakila & Vijayarani, 2017). The CLSM images obtained in this study (Figs. 6a,b) reveal that the ﬂuorescently labelled oil droplets can be clearly distinguished from the surrounding agarose background that has negligible ﬂuorescence, indicating high contrast images, hence automatic global thresholding was utilised.

The ‘Default’ thresholding method employs an iterative algorithm, which randomly selects pixels from the image (oil droplets and agarose) and averages them to form an initial threshold value (Calvard & Ridler, 1978). This is used to achieve a separation, with the average of the pixels at or above the threshold value (i.e. oil droplets) and the average of the pixels below the threshold value (i.e. agarose) being computed. The average of the two means is then used as the optimal threshold limit for re-thresholding of the image. The process is repeated until the optimal threshold is greater than the composite average, as follows:

μ TH (

)

TH (

)

obj

THopt

(6)

= ⎛ ⎜ ⎝

⎞ ⎟ ⎠

where, THopt is the optimal threshold, and μbg and μobj are the average of the background and object pixel intensities separated by the initial threshold (TH).

FIJI includes sixteen distinct methods of global thresholding, each employing a diﬀerent algorithm (Landini, 2017). To ascertain which algorithm works best for oil phase volume calculations, individual 3D images, from one section of each agarose-canola oil mixture, were thresholded using the sixteen algorithms. The oil volume calculation was carried out by extracting the total area of all droplets from the thresholded 2D images, regardless of the droplet diameter or circu- larity, and then multiplying the total area by the z-step between two optical cross sections in the xy-plane (i.e. 5.72 μm in this case). Table 1 summarises estimations of the oil phase volume derived from all the thresholding algorithms for a stack of 2D images in our composites. Blending law modelling in the preceding section argues that the

In our samples, an optimal threshold value of 55 was obtained from

129

P. Mhaske, et al.

Food Research International 125 (2019) 108529

Table 1 Comparison of canola oil phase volume in mixture with agarose calculated using diﬀerent FIJI thresholding methods (n = 27).

Thresholding method

Phase volume of canola oil (%)

1% agarose plus 1% oil

1% agarose plus 3% oil

1% agarose plus 5% oil

1% agarose plus 9% oil

1% agarose plus 7% oil

1% agarose plus 11% oil

Default

Huang

Intermodes

IsoData

IJ IsoData

MaxEntropy

Mean

MinError

9.39 ± 0.03 17.80 ± 0.59 12.99 ± 0.04 11.74 ± 0.52 11.53 ± 0.28 16.94 ± 0.21 12.28 ± 1.22 24.23 ± 0.62 84.04 ± 0.82

Minimum

Moments

Otsu

Percentile

Reniyentropy

3.02 ± 0.01 8.09 ± 1.12 3.36 ± 0.06 3.59 ± 0.03 3.47 ± 0.02 6.55 ± 0.08 3.89 ± 0.01 13.94 ± 0.11 46.02 ± 0.57 2.20 ± 0.11 3.69 ± 0.09 3.59 ± 0.08 42.34 ± 0.75 3.85 ± 0.24 0.01 ± 0.00

5.32 ± 0.03 18.90 ± 0.22 2.59 ± 0.38 7.36 ± 0.33 6.78 ± 0.33 14.08 ± 0.72 6.16 ± 0.47 21.61 ± 0.48 48.50 ± 0.49 0.01 ± 0.00 6.89 ± 0.15 7.36 ± 0.07 45.44 ± 1.32 6.36 ± 0.54 0.0013 ± 0.02

Shanbagh

1.16 ± 0.03 36.11 ± 0.14 1.18 ± 0.04 1.21 ± 0.09 1.18 ± 0.03 2.25 ± 0.09 1.66 ± 0.17 17.58 ± 0.58 32.31 ± 0.23 0.26 ± 0.05 1.23 ± 0.05 1.24 ± 0.04 37.15 ± 0.17 1.66 ± 0.13 0.0006 ± 0.00

Triangle

Yen

10.99 ± 0.19 4.88 ± 0.23

41.98 ± 0.13 6.46 ± 0.50

7.33 ± 0.16 30.11 ± 1.27 3.55 ± 0.27 12.51 ± 0.43 11.28 ± 0.21 23.49 ± 0.28 5.10 ± 0.15 26.99 ± 0.23 56.19 ± 0.86 0.05 ± 0.01 7.65 ± 0.47 8.84 ± 0.30 46.90 ± 0.76 7.65 ± 0.20 0.004 ± 0.00 50.29 ± 0.02 9.60 ± 0.26

8.23 ± 0.49 11.95 ± 0.54 11.74 ± 0.25 46.96 ± 0.88 12.39 ± 0.09 1.57 ± 0.11 22.13 ± 1.70 13.67 ± 0.38

10.72 ± 0.15 15.32 ± 0.29 9.71 ± 0.18 9.38 ± 0.21 9.24 ± 0.08 14.09 ± 0.30 11.92 ± 0.11 17.77 ± 0.30 65.92 ± 0.91 4.36 ± 0.44 10.06 ± 1.05 10.60 ± 0.29 43.44 ± 0.40 11.60 ± 1.56 0.06 ± 0.04 21.51 ± 1.12 12.06 ± 1.08

29.59 ± 0.40 1.61 ± 0.08

Fiji

Imaris

)

( l i o a l o n a c f o e m u l o v e s a h P

11%

Concentration of canola oil in composite with agarose (%)

Fig. 7. Phase volume of canola oil in mixture with 1% (w/w) agarose determined via a Z-stack method coupled with image analysis software (FIJI and Imaris).

software was found to be utilising the ‘Default’ automatic global thresholding method and the ‘analyse particle’ function, prior to mul- tiplication by the Z-depth to obtain the volume fraction of the lipid phase.

Microscopic work was complemented with another image proces- sing tool (Imaris) that employs an alternative way of phase volume estimation (Calvard & Ridler, 1978; Fishman et al., 2006; Fogarty, Hammond, Kanjhan, Bellingham, & Noakes, 2013). We used a feature of this licensed software called maximum intensity projection (MIP) to transform a stack of 2D confocal images of the agarose-canola oil gel

the Default algorithm using Eq. (6). In addition, we found that this is the only algorithm that can consistently compute the phase volumes as a function of increasing oil levels in the various composites (Table 1). Therefore, the Default thresholding method was applied to determine the phase volume of oil in all 3D images, captured from nine sections across each sample, and outcomes are given in Fig. 7. Accordingly, FIJI's image-analysis predictions of oil phase volumes are very close to the experimental concentrations in the composites, and these are also congruent with the phase volume estimates from blending law model- ling (inset of Fig. 5). Therefore, the best suited methodology for the FIJI

130

P. Mhaske, et al.

Food Research International 125 (2019) 108529

approach lays the groundwork for future research, which could utilise more complex biomaterial composites, e.g. two hydrocolloid aqueous phase supporting lipid inclusions, with the goal of accurately and ra- pidly determining phase behaviour in industrially relevant matrices.

into a 3D image. Then, we activated another feature in Imaris called “Surfaces” for identiﬁcation of the object of interest (i.e. oil droplets) and subsequent image quantiﬁcation. In doing so, a Gaussian ﬁlter was applied on a 3D image of the present composite to smoothen it followed by subtraction of the background intensity from the image to imple- ment a thresholding process.

Acknowledgement

The authors acknowledge the facilities, and the scientiﬁc and the RMIT University's Microscopy &

assistance of

technical Microanalysis Facility.

Funding

This research did not receive any speciﬁc grant from funding

agencies in the public, commercial, or not-for-proﬁt sectors.

Declarations of interest

None.

References

Once the above were accomplished, measurements could be carried out with a marching cube algorithm (Lorensen & Cline, 1987). The algorithm works by creating a polygonal mesh around the oil droplets and those contained within the mesh were analysed volumetrically. To ensure that all oil droplets are included in the polygonal mesh selection, we used the diameter of the largest oil droplet as the limiting factor. Measurement of the droplet diameter is performed manually in Imaris, using a feature called “Slice”, and it was found that the largest diameter of oil droplets in the images of all mixtures varied from 80 to 100 μm. It should be noted that in some cases boundaries are in contact with an- other and they may be considered as one larger droplet in Imaris. Therefore, measurements were carried out taking into consideration the possible presence of fused oil droplets yielding an imaging limiting factor of a large diameter. As such the methodology determined to be best suited for this application in the Imaris software was found to be processing of the obtained images using the ‘surface’ feature, followed by thresholding using the default method followed by phase volume estimation.

Almrhag, O., George, P., Bannikova, A., Katopo, L., Chaudhary, D., & Kasapis, S. (2013). Investigation on the phase behaviour of gelatin/agarose mixture in an environment of reduced solvent quality. Food Chemistry, 136, 835–842.

Auty, M. A. E. (2013). 4 - Confocal microscopy: Principles and applications to food mi- crostructures. In V. J. Morris, & K. Groves (Eds.). Food microstructures (pp. 96–98). Woodhead Publishing.

Bergouignan, L., Chupin, M., Czechowska, Y., Kinkingnéhun, S., Lemogne, C., Le Bastard,

G., & Fossati, P. (2009). Can voxel based morphometry, manual segmentation and automated segmentation equally detect hippocampal volume diﬀerences in acute depression? NeuroImage, 45, 29–37.

Blonk, J., Van Eendenburg, J., Koning, M., Weisenborn, P., & Winkel, C. (1995). A new CSLM-based method for determination of the phase behaviour of aqueous mixtures of biopolymers. Carbohydrate Polymers, 28, 287–295.

Calvard, T., & Ridler, S. (1978). Picture thresholding using an iterative selection method.

IEEE Transactions on Systems, Man, and Cybernetics, 8, 630–632.

Dekkers, B. L., Emin, M. A., Boom, R. M., & van der Goot, A. J. (2018). The phase

properties of soy protein and wheat gluten in a blend for ﬁbrous structure formation. Food Hydrocolloids, 79, 273–281.

Firoozmand, H., & Rousseau, D. (2014). Tailoring the morphology and rheology of phase-

separated biopolymer gels using microbial cells as structure modiﬁers. Food Hydrocolloids, 42, 204–214.

Fishman, E. K., Ney, D. R., Heath, D. G., Corl, F. M., Horton, K. M., & Johnson, P. T.

The above protocol allowed volume determination of oil droplets within the speciﬁed mesh. It was applied on 3D images from nine sections of each composite gel, and results are also reproduced in Fig. 7. They show that the Imaris-based phase volumes are closely matched to the experimental oil concentration in the mixtures and as such are in good agreement with those predicted using the rheology-based blending law. Statistical analysis conducted on the phase volumes ob- tained by Imaris and FIJI for each composite preparation show no signiﬁcant diﬀerence (p > .05) between the two imaging tools. It ap- pears that although two distinct processing techniques have been de- veloped, i.e. Imaris processes 3D images whereas FIJI does so on the 2D counterparts, they can both determine accurately the phase volume of canola oil in agarose gels. These inclusions appear to belong to the same statistical population, giving us conﬁdence in the application of the developed protocols for comparison with the theoretical modelling from rheology.

(2006). Volume rendering versus maximum intensity projection in CT angiography: What works best, when, and why. Radiographics, 26(3), 905–922.

4. Conclusions

Fogarty, M., Hammond, L., Kanjhan, R., Bellingham, M., & Noakes, P. (2013). A method for the three-dimensional reconstruction of Neurobiotin™-ﬁlled neurons and the lo- cation of their synaptic inputs. Frontiers in Neural Circuits, 7, 153.

Kasapis, S., Morris, E. R., Norton, I. T., & Clark, A. H. (1993). Phase equilibria and ge-

lation in gelatin/maltodextrin systems - Part IV: Composition-dependence of mixed- gel moduli. Carbohydrate Polymers, 21, 269–276.

Katopo, L., Kasapis, S., & Hemar, Y. (2012). Segregative phase separation in agarose/ whey protein systems induced by sequence-dependent trapping and change in pH. Carbohydrate Polymers, 87, 2100–2108.

Landini, G. (2017, 29 September). Auto threshold. Retrieved from http://imagej.net/

Auto_Threshold.

Li, Q., Liu, R., Wu, T., & Zhang, M. (2017). Interactions between soluble dietary ﬁbers and wheat gluten in dough studied by confocal laser scanning microscopy. Food Research International, 95, 19–27.

Lorén, N., Langton, M., & Hermansson, A.-M. (1999). Confocal laser scanning microscopy and image analysis of kinetically trapped phase-separated gelatin/maltodextrin gels. Food Hydrocolloids, 13, 185–198.

Lorensen, W. E., & Cline, H. E. (1987). Marching cubes: A high resolution 3D surface construction algorithm. Paper presented at the ACM siggraph computer graphics.

Luo, T. L., Eisenberg, M. C., Hayashi, M. A. L., Gonzalez-Cabezas, C., Foxman, B., Marrs, C. F., & Rickard, A. H. (2018). A sensitive thresholding method for confocal laser scanning microscope image stacks of microbial bioﬁlms. Scientiﬁc Reports, 8(1), 13013.

Through the years, rheology-based blending law analysis has been extensively utilised to predict phase equilibria and gelation properties in binary gels. Ultimately, blending laws provide estimates of the component phase volume that determines structural performance in composites. Recent advances in imaging technology (mainly for clinical analysis and medical intervention) have allowed us to develop a novel microscopic method using Z-stack CLSM, in combination with image processing tools, for application in a soft biocomposites. In particular, the model system of this study comprised a binary composite of a continuous agarose network supporting discontinuous inclusions of canola oil that show a distinct phase separation. The developed mi- croscopic method was successful in determining the phase volume of spheroidal oil inclusions in a range of mixed gel formulations. Results agree remarkably well with phase volume predictions from the isostrain model of blending laws also obtained presently for agarose-canola oil systems.

Marikkar, J. M. N., Ghazali, H. M., Che Man, Y. B., & Lai, O. M. (2002). The use of cooling and heating thermograms for monitoring of tallow, lard and chicken fat adulterations in canola oil. Food Research International, 35, 1007–1014.

McClements, D. J. (2015). Emulsion formation. Food emulsions: principles, practices, and

techniques(3rd ed.). Boca Raton: CRC Press.

Morris, E. R. (2009). Funtional interactions in gelling biopolymer mixtures. In S. Kasapis, I. T. Norton, & J. B. Ubbink (Eds.). Modern biopolymer science – Bridging the divide between fundamental treatise and industrial application. New York: Elsevier. Nagano, T., Tamaki, E., & Funami, T. (2008). Inﬂuence of guar gum on granule

morphologies and rheological properties of maize starch. Carbohydrate Polymers, 72,

Furthermore, this is the ﬁrst time that the validity of rheological predictions in determining phase volume in an aqueous/lipid phase has been tested, as opposed to the widely used application of rheological modelling in aqueous mixtures. Known concentrations of canola oil were used in this study to ensure known phase volumes, such that the validity of both the theoretical/rheological and microscopic meth- odologies for determining phase volume could be assessed. The working

131

P. Mhaske, et al.

Food Research International 125 (2019) 108529

95–101.

of components in food system. In S. Damodaran, K. L. Parkin, & O. R. Fennema (Eds.). Fennema's food chemistry(4th ed.). Boca Raton: CRC Press.

Nielsen, L. E. (1974). Morphology and the the elastic modulus of block polymers and

polyblends. Rheologica Acta, 13, 86–92.

Peighambardoust, S., Dadpour, M., & Dokouhaki, M. (2010). Application of epi-

Takayanagi, M., Harima, H., & Iwata, Y. (1963). Viscoelastic behavior of polymer blends and its comparison with model experiments. 23, Memoirs-Faculty of Engineering Kyushu University1–13.

ﬂuorescence light microscopy (EFLM) to study the microstructure of wheat dough: A comparison with confocal scanning laser microscopy (CSLM) technique. Journal of Cereal Science, 51, 21–27.

Rohman, A., & Man, Y. B. C. (2010). Fourier transform infrared (FTIR) spectroscopy for

Tang, D., & Marangoni, A. G. (2007). Modeling the rheological properties and structure of colloidal fat crystal networks. Trends in Food Science & Technology, 18, 474–483. Unnikrishnan, R., Pantofaru, C., & Hebert, M. (2007). Toward objective evaluation of image segmentation algorithms. IEEE Transactions on Pattern Analysis and Machine Intelligence, 29, 929–944.

analysis of extra virgin olive oil adulterated with palm oil. Food Research International, 43, 886–892.

Ur-Rehman, A., Khan, N. M., Ali, F., Khan, H., Khan, Z. U., Jan, A. K., & Ahmad, S. (2016). Kinetics study of biopolymers mixture with the help of confocal laser scanning mi- croscopy. Journal of Food Process Engineering, 39, 533–541.

Sakila, A., & Vijayarani, S. (2017). A hybrid approach for document image binarization. Paper presented at the 2017 International Conference on Inventive Computing and Informatics (ICICI).

Samiey, B., & Ashoori, F. (2012). Adsorptive removal of methylene blue by agar: Eﬀects of

NaCl and ethanol. Chemistry Central Journal, 6, 14.

Van Dalen, G. (2002). Determination of the water droplet size distribution of fat spreads using confocal scanning laser microscopy. Journal of Microscopy, 208, 116–133. Wang, W., Guo, X., & Yang, Y. (2011). Lithium iodide eﬀect on the electrochemical be-

Schindelin, J., Arganda-Carreras, I., Frise, E., Kaynig, V., Longair, M., Pietzsch, T., & Cardona, A. (2012). Fiji: An open-source platform for biological-image analysis. Nature Methods, 9, 676.

havior of agarose based polymer electrolyte for dye-sensitized solar cell. Electrochimica Acta, 56, 7347–7351.

Semasaka, C., Katopo, L., Buckow, R., & Kasapis, S. (2018). Modeling water partition in composite gels of BSA with gelatin following thermal treatment. Food Hydrocolloids, 76, 141–149.

Zamora-Mora, V., Velasco, D., Hernández, R., Mijangos, C., & Kumacheva, E. (2014). Chitosan/agarose hydrogels: Cooperative properties and microﬂuidic preparation. Carbohydrate Polymers, 111, 348–355.

Shrinivas, P., Kasapis, S., & Tongdang, T. (2009). Morphology and mechanical properties of bicontinuous gels of agarose and gelatin and the eﬀect of added lipid phase. Langmuir, 25, 8763–8773.

Zhang, H., Ma, J., Miao, Y., Tuchiya, T., & Chen, J. Y. (2015). Analysis of carbonyl value of frying oil by fourier transform infrared spectroscopy. Journal of Oleo Science, 64, 375–380.

Sikorski, Z. E., Pokorny, J., & Damodaran, S. (2008). Physical and chemical interactions

132

Chapter 4 Phase volume quantification of agarose-ghee gels using

3D confocal laser scanning microscopy and blending law

analysis: A comparison

Mhaske, P., Condict, L., Dokouhaki, M., Farahnaky, A., & Kasapis, S. (2020). Phase volume

quantification of agarose-ghee gels using 3D confocal laser scanning microscopy and

133

blending law analysis: A comparison. LWT, 129, 109567.

LWT - Food Science and Technology 129 (2020) 109567

Contents lists available at ScienceDirect

LWT - Food Science and Technology

journal homepage: www.elsevier.com/locate/lwt

Phase volume quantiﬁcation of agarose-ghee gels using 3D confocal laser scanning microscopy and blending law analysis: A comparison

∗

Pranita Mhaske, Lloyd Condict, Mina Dokouhaki, Asgar Farahnaky, Stefan Kasapis

School of Science, RMIT University, Bundoora West Campus, Plenty Road, Melbourne, VIC, 3083, Australia

A R T I C L E I N F O

A B S T R A C T

Keywords: Lewis-Nielsen blending law Phase behaviour Confocal laser scanning microscopy 3D imaging Image analysis

A thorough understanding of the phase behaviour of biomaterial composites is imperative for manipulating the structural and textural properties in novel food products. This study probed the phase behaviour of a model system comprising agarose and a varying concentration of ghee. Results obtained from scanning electron mi- croscopy (SEM), micro diﬀerential scanning calorimetry (DSC), Fourier transform infrared spectroscopy (FTIR) and dynamic oscillation in-shear revealed discontinuous and hard inclusions of ghee reinforcing the continuous, weaker agarose matrix with increasing concentrations of the former. Phase behaviour of the system was quantiﬁed in parallel with a novel method combining 3D confocal laser scanning microscopy (CLSM) imaging and image analysis software - FIJI and Imaris - in an eﬀort to substantiate the eﬃcacy of the microscopic protocol in quantifying phase behaviour. Phase volumes recorded with the microscopic protocol were in close agreement to those modelled with the Lewis-Nielsen blending law using small-deformation dynamic oscillation. However, results indicated that the inner ﬁltering eﬀect or ‘self-shadowing’ observed commonly in CLSM images may pose a limitation to the application of this technique, necessitating further development before it can be applied to more complex, industrially relevant systems.

1. Introduction

adapting the classic works of Nielsen (1974) and Takayanagi et al. (1963) in biopolymer research (Semasaka, Buckow, & Kasapis, 2018). This approach has been successfully applied to a plethora of studies in relation to the development speciﬁc textures, fat replacement, etc. (Dekkers, Emin, Boom, & van der Goot, 2018; Kasapis & Tay, 2009). However, the application of blending laws is a time-consuming exercise necessitating a signiﬁcant amount of rheological experimentation and modeling expertise. They are, as such an indirect application, and until recently were conﬁned to aqueous biopolymer mixtures (Mhaske, Condict, Dokouhaki, Katopo, & Kasapis, 2019).

Over the years, attempts have been made to develop a direct method using confocal laser scanning microscopy (CLSM) to study and quantify phase behaviour and the eﬀective phase volume of components in binary mixtures. Such studies aimed mainly to qualitatively probe various images of biopolymer mixtures, e.g. interpenetrating networks of soy protein and sugar beet pectin (Hou et al., 2015). They also at- tempted to quantitatively analyse aqueous systems and 2D images (Blonk, van Eendenburg, Koning, Weisenborn, & Winkel, 1995; Lorén, Langton, & Hermansson, 1999), but were not met with widespread appeal leading to further development.

In order to keep up with the increasing competition in the market, the food industry is continuously challenged to improve products in terms of value addition, cost eﬃciency and contribution to maintaining a healthy lifestyle without compromising on their quality, taste and texture (Kasapis, 2008). As a consequence, the industry has seen an increase in the use of biopolymers such as proteins and dietary ﬁbre not only to reinforce the nutritional value but also to act as thickening and stabilizing agents, creating formulations with intriguing and distinct textural properties (Çakır & Foegeding, 2011). Mixtures of biopolymers widen the sensory window of products by enabling the manipulation of microstructure and textural properties in semi-solid foods. This is made possible due to the thermodynamic imbalance that segregates the binary mixture into distinct phases, depending on the morphology of the microstructure, hence altering the product's textural properties. Segregative phase separation commonly leads to the enrichment of the two components in distinct phases with one phase acting as a con- tinuous matrix supporting the second discontinuous “ﬁller” (Gaaloul, Turgeon, & Corredig, 2010).

Recently, a novel technique of quantifying phase volume in agarose- canola oil composites using 3D imaging coupled with image analysis

By considering these phase-separated components as distinct in- dividual systems, their mechanical properties can be predicted by

∗

Corresponding author. E-mail address: stefan.kasapis@rmit.edu.au (S. Kasapis).

https://doi.org/10.1016/j.lwt.2020.109567 Received 19 September 2019; Received in revised form 5 May 2020; Accepted 6 May 2020

134

P. Mhaske, et al.

LWT - Food Science and Technology 129 (2020) 109567

2.3.2. Fourier transform infrared spectroscopy (FTIR)

Single and binary mixtures of agarose and ghee were analysed using a PerkinElmer Spectrum 2 FTIR spectrometer (Norwalk, Connecticut, United States) with a GladiATR (Pike Technologies, Maddison, Wisconsin, United States) diamond crystal attenuated total reﬂectance (ATR) device. The absorbance interferograms were obtained over 400- −1. The −1 and averaged over 32 scans at a resolution of 4 cm 4000 cm spectra of agarose and agarose-ghee composites were ﬁnalised by subtracting the spectrum of water in triplicate measurements.

was developed by Mhaske et al. (2019). Results obtained were in ex- cellent agreement with estimates from traditional theoretical modeling, supporting the eﬃcacy and accuracy of this work. CLSM is a quick, direct method that enables the user to study phase behaviour while keeping the composite microstructure intact by making tedious sample preparation like sectioning, coating and freeze drying redundant. This, coupled with the accuracy of image analysis software in quantifying phase volume shows potential for application over a wide range of materials saving time and eﬀort. The primary objective of this paper, therefore, is to test the eﬃcacy of the protocol in quantifying the phase volume of a more complex, hard lipid-biopolymer composite.

2.3.3. Diﬀerential scanning calorimetry (DSC)

Setaram Micro DSC VII (Setu-rau, Caluire, France) was used to carry out thermal analysis of the agarose-ghee mixtures. Samples (700 mg) were weighed into cylindrical vessels and sealed. A vessel containing equal weight of milliQ water (Millipore Corporation, Burlington, United States) served as a reference. Samples were equilibrated at 40 °C for 30 min before heating to 70 °C, cooling to 5 °C and heating to 70 °C again. Triplicate runs of averaged exotherms obtained at a scan rate of 1 °C/min are reported in this work.

The research also deals with the application of blending laws in quantifying the phase volumes in a hard lipid-biopolymer mixture in contrast to their widespread application to aqueous gels. Since lipids are an integral part of most food products (e.g. dressings, spreads), this ﬁnding holds potential in broadening the spectrum of scrutinizable products using blending laws. Attempts will be made to quantify the phase volume of a model system comprising of known quantities of hard lipid (i.e. ghee) in an agarose-ghee mixture using the blending laws and theoretical estimates will be compared with phase volumes obtained by the microscopic protocol. Ghee (anhydrous milk fat made by clarifying butter) remains in the solid form at ambient temperature (and within the temperature range of the present experimentation) and is expected to phase separate due to thermodynamic incompatibility with the aqueous network of agarose (Mortensen, 2016).

2. Materials and methods

2.1. Materials

Ghee by itself was analysed using a modulated DSC-Q2000, by TA Instruments (New Castle, DE, USA), ﬁtted with a cooling unit, capable of reaching temperatures of about −50 °C. The accumulation of con- densate was prevented by purging nitrogen at a ﬂow rate of 20 ml/min. About 10 mg of ghee was weighed and hermatically sealed in an alu- minium pan with an empty pan used as a reference. The sample was held at 40 °C for 30 min, heated to 70 °C and cooled down to −35 °C before heating it to 70 °C at a scan rate of 1 °C/min. A modulated amplitude of ± 0.53 °C for 40 s was used throughout. Averaged results of triplicates yielding eﬀectively overlapping exotherms are reported in this study.

2.3.4. Rheological measurements

Agarose: Agarose (Type-1) was obtained from Sigma-Aldrich (Sydney, Australia). According to the supplier, the moisture, ash and sulphate content was less than 10, 0.25 and 0.15 g/100 g, respectively. Ghee: Pure cow ghee was purchased from Amul (Gujrat, India), the

density of which was measured to be 0.911 g/ml.

2.2. Methods

Density measurement: The density of ghee was volumetrically mea- sured using a 10 ml measuring cylinder. The cylinder was ﬁlled up to the 10 ml mark with ghee to measure its weight, which was divided by the weight of equal volume of water.

Discovery Hybrid Rheometer by TA Instruments (New Castle, DE, USA) was used to carry out small-deformation dynamic oscillation in- shear tests to measure the storage modulus of single and binary systems of agarose and ghee. A 20 mm geometry was used for ghee by itself while a 40 mm geometry was used for agarose and the composites with a 1 mm gap. Samples were loaded on a Peltier plate pre-heated at 45 °C and covered with polydimethylsiloxane to minimize evaporation. These were then cooled to 5 °C (at a rate of 2 °C/min) with a controlled strain of 0.01% and a constant oscillatory frequency of 1 rad/s prior to a 30 min time sweep. At the conclusion of the time sweep, a frequency sweep was performed in the range of 0.01–100 rad/s, followed by a heating ramp to 80 °C. Tests were carried out in triplicate for each sample.

2.3.5. Confocal laser scanning microscopy (CLSM)

Sample preparation: Binary mixtures were prepared by maintaining agarose concentration at 1 g/100 g and varying concentrations of ghee (0, 1, 3, 6, 9, 12 and 15 g/100 g) while adjusting the amount of water added to make the total volume of sample, 100 ml. In doing so, 1 g agarose was dissolved in calculated amount of milliQ water at 75 °C, stirred for 10 min and cooled to 45 °C prior to the addition of ghee. This mixture was then homogenized at 5000 rpm for 30 s using an IKA homogeniser-T25 Turrax (Wilmington, North Carolina, USA).

2.3. Experimental analysis

2.3.1. Scanning electron microscopy (SEM)

Ghee was stained with a lipophilic stain, Nile red (0.005 g/L) and then used to prepare agarose-ghee composites containing 1 g/100 g agarose and a varying concentration of ghee as outlined in Section 2.2. The homogenized sample solutions were pipetted into the well of a CoverWell imaging chamber, having a diameter of 20 mm and a depth of 2.8 mm (Sigma-Aldrich, NSW, Australia). Pre-cooling the chamber to 4 °C was essential for rapid in-situ gelation where the ghee is kinetically trapped within the agarose network and not bulk phase separated.

Confocal images were captured using a Nikon ECLIPSE Ti-E inverted microscope (Minato, Tokyo, Japan) ﬁtted with an A1R laser. A 488 nm laser was used to excite Nile red and the emissions were captured be- tween 500 and 530 nm. A stack of 2D images was captured using a 10x objective. Fifteen consecutive images along the z-axis with a z-step of 5.72 μm were captured, imaging a depth of about 86 μm from the surface. Nine images were captured across each sample to avoid the eﬀect of any inhomogeneous distribution of the ghee droplets.

2.3.6. Image analysis

The Z-stack obtained from CLSM imaging was quantiﬁed using two

Agarose gel cubes (10 mm3) with varying concentrations of ghee were freeze dried for 72 h at −40 °C in a Labconco FriZone Triade freeze dryer (Kansas City, Missouri, USA). They were then glued on to 12.6 mm carbon coated aluminium pins prior to being subjected to gold coating. An overview of the surface microstructure was obtained using FEI Quanta 200 ESEM (Oregon, USA) in low vacuum mode at 50x magniﬁcation. The working distance was 12 mm with an accelerating voltage of 25 kV and a spot size of 5. A 400x magniﬁcation was used to zoom into the pores of the agarose network to capture details of the agarose-ghee topology using a spot size of 5 and an electron beam of 25 kV.

135

P. Mhaske, et al.

LWT - Food Science and Technology 129 (2020) 109567

3.1.2. Chemical properties

image analysis software – FIJI with Java 6 (Madison, Wisconsin, USA) and Imaris 9.1.0 software (Bitplane, Belfast, Northern Ireland) in par- allel to estimate the lipid phase volume using the protocol outlined by Mhaske et al. (2019) as follows:

bridges),

(sulphate

FIJI: The Z-stack was thresholded using the “Default” algorithm, after which, the total area of object pixels was calculated irrespective of the size and circularity of the droplet. The generated area was ﬁnally multiplied by the z-step to determine the volume of the fat droplets present in that Z-stack. Averaged results of the nine images captured across each sample are reported.

1246 −1 (C–H vibration) and 2980 cm

−1 (C–C bending), 2900 cm

The molecular nature of agarose, ghee and their composites was examined using Fourier transform infrared (FTIR) spectroscopy. This was done to identify the presence of potential molecular interactions between the two components using as reference their molecular ﬁn- gerprints. The interferograms obtained are illustrated in Fig. 2. Spectral −1 (anomeric scans of agarose exhibited absorption bands at 877 cm −1 (C–O–C vibration of carbon's angular C–H deformation), 1068 cm −1 3,6-anhydro-galactose cm esters), −1 1396 cm (O–H stretching). These bands are characteristic of the polysaccharide as reported earlier (Awadhiya, Tyeb, Rathore, & Verma, 2017; Wang, Guo, & Yang, 2011; Zamora-Mora, Velasco, Hernández, Mijangos, & Kumacheva, 2014).

Imaris: The 3D image was opened in the “Surfaces” feature and a Gaussian ﬁlter was applied to smoothen it, prior to subtracting the background intensity to reduce noise. The volume of ghee is measured in voxels by setting the diameter of the largest droplet as the limiting factor. This was repeated for the nine images and the averaged results are reported.

−1 (C–H rocking of CH2 and cis double bond), 1155 cm

2.3.7. Statistical analysis

Volume measurement of the lipid phase was done in triplicate and the data is expressed in mean ± standard deviation. Minitab 18 (Minitab Inc., Pennsylvania, USA) was used to carry out one-way ana- lysis of variance (ANOVA). Signiﬁcance diﬀerences were deﬁned as p < 0.05 with the Tukey test.

3. Results and discussion

The interferogram of ghee produced typical peaks, which reﬂect distinct characteristics of edible oils (Mansor, Man, & Rohman, 2011; Upadhyay, Jaiswal, & Jha, 2016). Absorption bands were observed at −1 721 cm −1 (bending of CH3 (stretching vibration of C–O in esters), 1452 cm −1 (C]O stretching of aliphatic esters) and at aliphatic group), 1746 cm −1 (symmetrical and asymmetrical C–H stretching of 2856 and 2920 cm CH2 groups in the organic fatty acids) (Antony, Sharma, Mehta, Ratnam, & Aparnathi, 2017; Upadhyay, Jaiswal, & Jha, 2018). The spectral scans of the blends replicate the characteristics of the two constituents with the spectra of the polymer dominating. Since its peaks deviated neither in band wavenumber nor intensity in the composites, and no new peaks were formed, it can be inferred that there is no molecular interaction between ghee and agarose.

3.1. Investigation of the composite microstructure and its structural characteristics

3.1.3. Thermal characterisation

3.1.1. Composite microstructure

In order to be able to apply the blending law analysis, it is vital to have a phase separated system with distinct, non-interacting compo- nent phases. To attest that the agarose-ghee composites meet this morphological requirement, and avoid bulk phase separation, freeze dried samples were analysed using SEM. The eﬃcacy of SEM in probing into the complex microstructures of composite gels has been well documented in the past (Pang, Deeth, Sharma, & Bansal, 2015).

Complementary characterisation to the preceding sections for the phase behaviour of our composite is aﬀorded by calorimetric tests (Du, Brenner, Xie, & Matsukawa, 2016). The typical exotherms of agarose and ghee during controlled cooling are shown in Fig. 3a. Changes ac- companying the exothermic proﬁle of agarose exhibited a mid-point transition temperature of about 26 °C, an outcome of the coil-to-helix transition of the polysaccharide leading to gelation at 1 g/100 g solids. The crystallization and lattice formation of the fatty acids, having a medium melting fraction (MMF; mainly stearin), lead to the formation of one minor peak at 15 °C, with the low melting fraction (LMF; mainly olein) resulting in a major peak at 9.5 °C. The bimodal behaviour re- ﬂected the thermodynamically distinct molecular domains present in ghee (Dimick, Reddy, & Ziegler, 1996; Verret et al., 2009).

An overview of the freeze-dried agarose gel (1 g/100 g) captured using 50x magniﬁcation is depicted in Fig. 1a. The image reveals a highly porous structure with the polysaccharide network forming a honeycomb pattern. Addition of ghee did not have any perceivable eﬀect on the continuous network of agarose, as all samples displayed a similar morphology (results not shown). Since the fat droplets weren't visible at this magniﬁcation, a higher exposure (400x) was used to zoom into the walls of the agarose network. This indicates that the size of the ghee inclusions is much smaller than the pores of the agarose matrix.

At 400x, in the absence of ghee, the walls of agarose network ex- hibited a smooth, uninterrupted surface as shown in Fig. 1b. This was contrasted in the micrographs of the binary mixture seen in Fig. 1c and d, which show the spherical inclusions of ghee embedded within the continuous polysaccharide mesh. The micrographs reveal that when the concentration of ghee was increased from 3 to 15 g/100 g, the number of lipid inclusions increased, while their size remained fairly constant (60 ± 15 μm).

Ghee is a complex mixture of triacylglycerols (TAGs), which con- stitutes 98% of its composition and comprise at least 60 diﬀerent fatty acids, with their characteristic chemical and thermal properties (Christelle Lopez, Bourgaux, Lesieur, Riaublanc, & Ollivon, 2006). The onset of crystallization and peak intensity depend largely on the rate of cooling (C. Lopez, Lesieur, Bourgaux, & Ollivon, 2005). Based on the above, Fig. 3b examines the cooling exotherms of the blends of 1 g/ 100 g agarose with increasing concentrations of ghee. Dominant peaks denoting the crystallization of fat are visible between −10 and 15 °C, with the peaks containing higher concentrations of ghee depicting two independent but partially overlapping phase transitions of the con- stituent melting fractions. The absence of the native agarose peaks in the blend can be attributed to the much larger intensity of the peaks of ghee masking the low intensity coil-to-helix events of the former.

Single ghee preparations or in a composite with agarose generated comparable thermal events, varying only in intensity as a function of concentration. This argues that the crystallization of fat is unaﬀected by the presence of the polysaccharide in their blends reinforcing the con- cept of phase separation in the mixture. Furthermore, agarose gels at temperatures higher than ghee crystallization and the isothermal test used in the following rheological experiments (5 °C for 30 min), hence the composite morphology should be that of a continuous biopolymer matrix supporting discontinuous fat crystals.

From the micrographs, it is clear that the blending of agarose and ghee lead to the formation of isotropic structures where the dis- continuous inclusions of ghee was supported by the continuous poly- meric matrix. The spheroidal lipid inclusions will serve as the basis for modeling the mechanical properties of the binary composite. Similar results were reported by Mhaske et al. (2019), who studied the mi- crostructure of agarose-canola oil composite and unveiled a continuous polysaccharide network supporting the oil droplets acting as a dis- continuous “ﬁller”.

136

P. Mhaske, et al.

LWT - Food Science and Technology 129 (2020) 109567

Fig. 1. SEM micrographs at 50x magniﬁcation of (a) 1% (w/w) agarose, and at 400x magniﬁcation of 1 g/100 g agarose containing (b) 0 g/100 g ghee, (c) 3 g/100 g ghee and (d) 15 g/100 g ghee.

Fig. 2. FTIR spectra of ghee (dotted line), and 1 g/100 g agarose with 0, 1, 3, 6, 9, 12 and 15 g/100 g ghee arranged from top to bottom (solid lines).

3.1.4. Rheological observations of structure formation

Finally, oscillatory tests were performed to investigate the me- chanical behaviour of the agarose network and crystalline ghee as a prerequisite to advancing a study on the eﬀect of increasing the con- centration of the latter in the composite. The proﬁle of storage modulus

(Gʹ) for 1 g/100 g agarose during controlled cooling at 2 °C/min is depicted in Fig. 4. On hydration and subsequent solubilisation, the disordered coils of agarose undergo a conformational transition at about 30 °C to form a strong gel at the end of the cooling routine (103.79 Pa at 5 °C). On the other hand, formation of the crystalline ghee

137

P. Mhaske, et al.

LWT - Food Science and Technology 129 (2020) 109567

Fig. 3. Exothermic peaks from the cooling proﬁles of (a) ghee (solid line) and agarose (dotted line) by themselves and (b) 1 g/100 g agarose with 1, 3, 6, 9, 12 and 15 g/100 g ghee arranged from bottom to top.

network lead to a sharp increase in the values of Gʹ at around 22 °C and a reinforced structure of 105.66 Pa at 5 °C.

aqueous phase and the hardening of ghee. Rheological observations suggest that ghee acts as the strong discontinuous ﬁller (its contribution is noticeable in the mixture) supported by a weaker agarose phase hence providing the basis for theoretical modeling using blending law.

3.2. Theoretical modeling of mechanical functions in support of phase topology

Turning our attention to the composites, these were made by adding 1 g of agarose to, for example, 9 g of ghee in a beaker. Subsequent amount of water was added to make up the total sample volume, re- sulting in an increase in the eﬀective concentration of agarose in the diminished aqueous phase as the concentration of fat increased in the composite (from 1 to 15 g/100 g). Addition of ghee to the agarose gels produced a systematic variation in viscoelasticity of the cooling pro- ﬁles, as depicted in Fig. 4. Blends showed a similar trend of gelation as that of the single agarose system up till 15 °C, with higher ghee con- centrations increasing the composite Gʹ value, as a result of volume exclusion between the aqueous and lipid phase, leading to eﬀective polysaccharide concentrations higher than the initial 1 g/100 g.

As shown in Fig. 4, a second wave of structure formation was wit- nessed at about 15 °C that reﬂected the onset of fat crystallization. Storage modulus values correspond to the concentration of added ghee reaching 103.84 and 104.32 Pa for 1 and 15 g/100 g ghee, respectively. Thus, the overall composite strength at the end of the cooling run (5 °C) is the outcome of higher eﬀective concentrations of the polymer in the

Having qualitatively established in the preceding sections that agarose and ghee remain in mixture in non-interacting phase-separated domains, we turned to an in-depth quantitative analysis using rheolo- gical blending law. In doing so, a double-logarithmic calibration curve was obtained (inset of Fig. 4 with a correlation coeﬃcient, R2 of 0.995) to provide information on the concentration dependence of storage modulus for agarose ranging from 1.0 to 2.5 g/100 g. This was indis- pensable for the interpolation between measured values of Gʹ at dif- ferent concentrations during theoretical modeling. The topology of a soft continuous polymer phase supporting a hard discontinuous ﬁller can follow progress in viscoelasticity of our results using the mathe- matical expression of Lewis and Nielsen (Shrinivas, Kasapis, &

138

P. Mhaske, et al.

LWT - Food Science and Technology 129 (2020) 109567

Fig. 4. Cooling proﬁle of storage modulus for 1% (w/w) agarose with 0 (□), 1 (җ), 3 (−), 6 (x), 9 (●), 12 (Δ) and 15 g/100 g (♦) (w/w) ghee and ghee by itself (■) during cooling from 50 to 5 °C at a scan rate of 2 °C/min. The inset denotes the agarose calibration curve of G′ at 5 °C as a function of its concentration.

Tongdang, 2009):

ABΦ y

−

BψΦ y

(1)

′ G ′ G x

−

fraction in the polymer-X phase and the total weights of the X and Y phases, respectively. Further, the relative weight of the two phases needed to be adjusted for density diﬀerences. This was carried out by dividing Wx and Wy by their respective densities (water = 1 g/ml; ghee = 0.911 g/ml) to obtain the values of Vx and Vy, i.e. the volumes of phase X and Y, respectively. Finally, the relative phase volume in the composite is given for Φx, for example, as follows:

(2)

′ G y ′ G x ′ G y ′ G x

Φ x

V x +

V x

V y

(6)

⎜

⎟

Φ y

Φ − 2 Φ m

(3)

⎞ ⎠

= + ⎛ 1 ⎝

=Φ y

+Φx

where, Gʹ, Gʹx, and Gʹy are the storage modulus of the composite, soft continuous polymer phase and rigid dispersed phase, respectively and ɸy is the relative phase volume of the ﬁller. The approach takes into consideration the shape of the ﬁller via the Einstein coeﬃcient (A), which is 1.5 for dispersed spheroids in an elastomeric matrix, and the maximum packing fraction (ɸm) of the ﬁller, which is 0.48 for loose packing of spheroidal aggregates reﬂecting the random arrangement of extremely hard ghee particles within the agarose matrix.

and these relative phase volumes are used in where, equations (1) and (3). Accordingly, the phase volumes of ghee were calculated for Sag = 1 and equations (4)–(6), which are also reproduced in the inset of Fig. 5 against the experimental lipid concentrations. Estimated phase volumes of ghee correlate well with the experimental concentrations in the composite, upholding that the phase volume de- pended entirely on its concentration. Phase volume predictions by theoretical modeling will be compared with those of 3D imaging cou- pled with image analysis to assess its eﬃcacy in estimating phase vo- lumes in demixed systems.

3.3. Z-stack CLSM imaging and image analysis of the agarose-ghee blends

Fig. 5 illustrates a typical outcome of calculations using equations (1)–(3) for 1 g/100 g agarose with 6 g/100 g ghee. A detailed section of the computerized output reproduced the storage modulus of the con- tinuous agarose phase (Gʹag) for all possible water distributions from 0.5 to 1.0, in steps of 0.01, in the agarose phase (Sag). The calculated hy- pothetical values of the composite (GʹC(LN)) and the experimental composite values (Gʹexp) as a function of Sag are also depicted. Calcu- lated and experimental values intersected at Sag of 1.0, positing that the total amount of water remained in the polymer phase. Modeling of the entire working range is depicted in the inset of Fig. 5 for Sag = 1, re- vealing a close agreement between the experimental and modelled composite moduli as a function of lipid concentration.

The above understanding has now been utilised to estimate the phase volume of the individual phases. This can be obtained by the generic equations developed by Kasapis, Morris, Norton, and Clark (1993):

x = +

S (

)

(4)

W x

= +

((1

−

)

W y y

(5)

3D images of all the agarose-ghee blends were captured and pro- cessed using image analysis software, FIJI and Imaris, following the protocol developed by Mhaske et al. (2019). In doing so, ﬂuorescent probes were employed to selectively stain the lipid. A single, 2D image captured from a section of a sample containing 1 g/100 g agarose and 6 g/100 g ghee in the xy plane is represented in Fig. 6a, as an example. The stained ghee droplets appeared as green spheres whilst the un- stained agarose formed the black background. The bright green spheres correspond to the droplets in the focal plane, which are contrasted to those outside the focal plane showing a reduced colour intensity. Each 2D depiction is an 8 bit image with 512 × 512 pixel in size (1 pixel = 2.49 μm on the X and Y-axis). Keeping a ﬁxed z-step of 5.72 μm, 15 consecutive 2D images along the Z-axis were captured as shown in Fig. 6b, collecting data up to a depth of 85 μm from the surface. This stack of 2D images was processed with FIJI, an open source analysis software for phase volume estimation, by considering separately each 2D plane. On the other hand, Imaris rendered a 3D image using the Z- stack (Fig. 6c) and performed volumetric measurements.

where, w, x, y, Sx, Wx and Wy denote the total weight of water in the composite, weight of polymer X, weight of polymer Y, the solvent

139

P. Mhaske, et al.

LWT - Food Science and Technology 129 (2020) 109567

Fig. 5. Computerized modeling for 1 g/100 g agarose plus 6 g/100 g ghee, using the Lewis and Nielsen blending law, is illustrated depicting the predicted storage modulus of agarose (G'ag; dashed line), composite (G'C(LN); solid line) and the experi- mental composite modulus (G'exp; dotted line) as a function of the solvent content of the agarose phase (Sag). The experimental Gʹ values (●) and phase vo- lume estimates of ghee (♦) for 1 g/100 g agarose with increasing ghee concentrations are depicted in the inset (blending law is used to derive the solid line ﬁt of composite modulus and the ghee phase volume in the inset).

Each pixel in the 2D images denotes a speciﬁc intensity value ran- ging from 0 to 255 (Xu, Li, Li, & Hua, 2016). A sample containing no ﬂorescent probe produces a completely black image with all the pixels having an intensity value of 0. This is exploited as the negative control that deﬁnes the baseline calibration in image analysis. While analysing the images, thresholding was the most crucial step to accurately de- lineate the boundary of fat droplets so that the object (fat) pixels are distinguished from the background (agarose) pixels. The phase volume of ghee estimated using FIJI is reported in Table 1 along with the weight-per-weight experimental concentration of the lipid added to the mixture and its conversion to the corresponding weight-per-volume concentration considering that its density is below 1.0. Thus, the phase volumes obtained using FIJI are very close to the experimental values (entries in the ﬁrst two columns), and the theoretical estimates in the inset of Fig. 5.

Turning our attention to the second image analysis software – Imaris, we once again followed the procedure outlined by Mhaske et al. (2019), with the phase volume estimates being also reported in Table 1. It is evident that the Imaris estimates are in good agreement with those obtained from FIJI, the experimental values, and the theoretical pre- dictions from rheological blending-law analysis seen in the inset of Fig. 5. Statistically, there is no signiﬁcant diﬀerence (p ˃ 0.05) between the measurements of phase volume obtained using FIJI and Imaris, indicating that both types of software systems are capable of generating reliable results.

A closer look at Table 1, revealed a discernible trend of decreasing proximity between the weight-per-volume ghee concentrations and the phase volumes estimated using the microscopic protocol. To scrutinize this further, a binary gel comprising 1 g/100 g agarose and 20 g/100 g ghee was prepared and analysed. Taking into account density con- siderations, this created a weight-per-volume ghee concentration of 21.55%, with FIJI and Imaris returning values of 20.12 ± 0.05 and 20.09 ± 0.11% (w/v), respectively. Such outcomes are rationalised in Fig. 7a that depicts the ﬁrst optical section captured along the Z-axis and Fig. 7b with the 15th optical section at a depth of approximately 85 μm from the surface. As before, green spheres denote the stained ghee droplets whereas the agarose is present as the black background. The clear spherical inclusions of ghee in Fig. 7a are contrasted with those in Fig. 7b, which appear to be eclipsed and akin to crescents.

Fig. 6. (a) 2D, (b) Z-stack, and (c) 3D CLSM images of 1 g/100 g agarose plus 6 g/100 g ghee. The green droplets depict the ghee whereas the black back- ground corresponds to agarose. Scale bar denotes 100 μm. (For interpretation of the references to colour in this ﬁgure legend, the reader is referred to the Web version of this article.)

This outcome is attributed to the limitation of CLSM in imaging deep sections, with the illuminating beam undergoing partial reﬂection

140

P. Mhaske, et al.

LWT - Food Science and Technology 129 (2020) 109567

Table 1 Experimental ghee concentrations in the composite and estimated phase volumes with FIJI and Imaris image analysis.

Weight per weight (g/100 g) ghee concentrations in the composite

Weight per volume (g/100 cm3) ghee concentrations in the composite

Lipid phase volumes estimated using FIJI (g/100 cm3)

Lipid phase volumes estimated using Imaris (g/100 cm3)

1 3 6 9 12 15

1.10 3.30 6.59 9.89 13.10 16.40

1.14 ± 0.02 3.42 ± 0.16 6.48 ± 0.11 9.68 ± 0.18 12.78 ± 0.13 15.60 ± 0.22

1.21 ± 0.07 3.41 ± 0.09 6.59 ± 0.14 9.54 ± 0.04 12.63 ± 0.21 15.79 ± 0.19

CRediT authorship contribution statement

Pranita Mhaske: Methodology, Formal analysis, Writing - original draft. Lloyd Condict: Writing - review & editing. Mina Dokouhaki: Writing - review & editing. Asgar Farahnaky: Writing - review & editing. Stefan Kasapis: Funding acquisition, Conceptualization, Supervision, Writing - review & editing.

Acknowledgement

The authors acknowledge the facilities, and the scientiﬁc and the RMIT University's Microscopy and

technical assistance of Microanalysis Facility.

References

Fig. 7. Confocal images of 1 g/100 g agarose and 20 g/100 g ghee captured from (a) the surface and (b) a depth of 80 μm from the surface. Scale bar de- notes 100 μm.

Antony, B., Sharma, S., Mehta, B. M., Ratnam, K., & Aparnathi, K. (2017). Study on FT- MIR spectra of ghee (anhydrous milk fat). British Food Journal, 119(1), 181–189. Awadhiya, A., Tyeb, S., Rathore, K., & Verma, V. (2017). Agarose bioplastic‐based drug delivery system for surgical and wound dressings. Engineering in Life Sciences, 17(2), 204–214.

Blonk, J. C. G., van Eendenburg, J., Koning, M. M. G., Weisenborn, P. C. M., & Winkel, C. (1995). A new CSLM-based method for determination of the phase behaviour of aqueous mixtures of biopolymers. Carbohydrate Polymers, 28(4), 287–295. Çakır, E., & Foegeding, E. A. (2011). Combining protein micro-phase separation and

protein–polysaccharide segregative phase separation to produce gel structures. Food Hydrocolloids, 25(6), 1538–1546.

Dekkers, B. L., Emin, M. A., Boom, R. M., & van der Goot, A. J. (2018). The phase

properties of soy protein and wheat gluten in a blend for ﬁbrous structure formation. Food Hydrocolloids, 79, 273–281.

Dimick, P., Reddy, S. Y., & Ziegler, G. R. (1996). Chemical and thermal characteristics of milk‐fat fractions isolated by a melt crystallization. Journal of the American Oil Chemists Society, 73(12), 1647–1652.

Du, L., Brenner, T., Xie, J., & Matsukawa, S. (2016). A study on phase separation behavior

and scattering that reduces the photon intensity striking the specimen being sampled (“inner ﬁlter” eﬀect). Loss of excitation reduces the ﬂuorescence yield from the focal plane, which is further attenuated on its return to the detector by the same factors (Pioch, Sorg, Stadler, Hagge, & Dörfer, 2004; Laurent et al., 1994). In lipid-biopolymer composites, component phases have signiﬁcantly diﬀerent refractive indices, and when the laser beam passes from a phase with a speciﬁc refractive index to another some of it is scattered back towards the source. Reduction of the laser intensity reaching deeper sections causes a “shadowing” eﬀect observed in Fig. 7b. It can be minimized by the use of an objective with a higher numerical aperture, increasing illumina- tion intensity, using the microscope in the transmission mode, etc. but cannot be completely overcome within the bounds of current tech- nology (Sandison, Williams, Wells, Strickler, & Webb, 1995, pp. 39–53).

in kappa/iota carrageenan mixtures by micro DSC, rheological measurements and simulating water and cations migration between phases. Food Hydrocolloids, 55, 81–88.

Gaaloul, S., Turgeon, S. L., & Corredig, M. (2010). Phase behavior of whey protein ag- gregates/κ-carrageenan mixtures: Experiment and theory. Food Biophysics, 5(2), 103–113.

4. Conclusions

Hou, J.-J., Guo, J., Wang, J.-M., He, X.-T., Yuan, Y., Yin, S.-W., et al. (2015). Edible

double-network gels based on soy protein and sugar beet pectin with hierarchical microstructure. Food Hydrocolloids, 50, 94–101.

Kasapis, S. (2008). Phase separation in biopolymer gels: A low-to high-solid exploration of structural morphology and functionality. Critical Reviews in Food Science and Nutrition, 48(4), 341–359.

Kasapis, S., Morris, E. R., Norton, I. T., & Clark, A. H. (1993). Phase equilibria and ge- lation in gelatin/maltodextrin systems — Part IV: Composition-dependence of mixed- gel moduli. Carbohydrate polymers. 21(4), 269–276.

Kasapis, S., & Tay, S. L. (2009). Morphology of molecular soy protein fractions in binary

composite gels. Langmuir, 25(15), 8538–8547.

Laurent, M., Johannin, G., Gilbert, N., Lucas, L., Cassio, D., Petit, P. X., et al. (1994). Power and limits of laser scanning confocal microscopy. Biology of the Cell, 80(2), 229–240.

Lopez, C., Bourgaux, C., Lesieur, P., Riaublanc, A., & Ollivon, M. (2006). Milk fat and

primary fractions obtained by dry fractionation: 1. Chemical composition and crys- tallisation properties. Chemistry and Physics of Lipids, 144(1), 17–33.

Lopez, C., Lesieur, P., Bourgaux, C., & Ollivon, M. (2005). Thermal and structural be- havior of anhydrous milk fat. 3. Inﬂuence of cooling rate. Journal of Dairy Science, 88(2), 511–526.

Lorén, N., Langton, M., & Hermansson, A. M. (1999). Confocal laser scanning microscopy and image analysis of kinetically trapped phase-separated gelatin/maltodextrin gels. Food Hydrocolloids, 13(2), 185–198.

Rheology-based blending laws have been widely applied over the years to estimate the phase volumes and water partitioning between segregative biopolymers in the composite gel. These, however, are te- dious, demanding a substantial amount of experimental work and specialised expertise for successful modeling of phase behaviour. Advances in imaging technology and high precision image analysis software have made it possible to probe specimens non-invasively and capture in-situ 3D images that can be further quantiﬁed to estimate the phase volume in binary composites. Our microscopic protocol was tested on a model system of agarose-ghee that showed distinct phase separation in order to estimate the phase volume of the discontinuous ﬁller. Estimates compared well with those obtained from the rheolo- gical modeling, making the microscopic protocol a potential substitute for accurately determining phase behaviour in industrially relevant composite gels. Perceivable eﬀects of self-shadowing during imaging merit further research in order to achieve wide-spread applicability for this protocol.

Mansor, T. S. T., Man, Y. B. C., & Rohman, A. (2011). Application of fast gas chroma- tography and Fourier transform infrared spectroscopy for analysis of lard adultera- tion in virgin coconut oil. Food Analytical Methods, 4(3), 365–372.

141

P. Mhaske, et al.

LWT - Food Science and Technology 129 (2020) 109567

Mhaske, P., Condict, L., Dokouhaki, M., Katopo, L., & Kasapis, S. (2019). Quantitative

of bicontinuous gels of agarose and gelatin and the eﬀect of added lipid phase. Langmuir, 25(15), 8763–8773.

analysis of the phase volume of agarose-canola oil gels in comparison to blending law predictions using 3D imaging based on confocal laser scanning microscopy. Food Research International, 125, 108529.

Mortensen, B. K. (2016). Anhydrous milk fat/butter oil and ghee reference Module in food

science. Elsevier.

Takayanagi, M. (1963). Viscoelastic behavior of polymer blends and its comparison with model experiments. Journal of the Society of Materials Science Japan, 12, 389–394. Upadhyay, N., Jaiswal, P., & Jha, S. N. (2016). Detection of goat body fat adulteration in pure ghee using ATR-FTIR spectroscopy coupled with chemometric strategy. Journal of Food Science and Technology, 53(10), 3752–3760.

Nielsen, L. E. (1974). Morphology and the elastic modulus of block polymers and poly-

blends. Rheologica Acta, 13(1), 86–92.

Pang, Z., Deeth, H., Sharma, R., & Bansal, N. (2015). Eﬀect of addition of gelatin on the

Upadhyay, N., Jaiswal, P., & Jha, S. N. (2018). Application of attenuated total reﬂectance Fourier Transform Infrared spectroscopy (ATR–FTIR) in MIR range coupled with chemometrics for detection of pig body fat in pure ghee (heat clariﬁed milk fat). Journal of Molecular Structure, 1153, 275–281.

rheological and microstructural properties of acid milk protein gels. Food Hydrocolloids, 43, 340–351.

Verret, C., El Moueﬀak, A., Largeteau, A., Frimigacci, M., Demazeau, G., Leal-Calderon, F., et al. (2009). Eﬀects of high pressure on anhydrous milk fat crystallization in emulsion. High Pressure Research, 29(1), 57–60.

Wang, W., Guo, X., & Yang, Y. (2011). Lithium iodide eﬀect on the electrochemical be-

Pioch, T., Sorg, T., Stadler, R., Hagge, M., & Dörfer, C. E. (2004). Resin penetration through submicrometer hiatus structures: A SEM and CLSM study. Journal of Biomedical Materials Research Part B: Applied Biomaterials: An Oﬃcial Journal of The Society for Biomaterials, The Japanese Society for Biomaterials, and The Australian Society for Biomaterials and the Korean Society for Biomaterials, 71(2), 238–243. Sandison, D. R., Williams, R. M., Wells, K. S., Strickler, J., & Webb, W. W. (1995).

havior of agarose based polymer electrolyte for dye-sensitized solar cell. Electrochimica Acta, 56(21), 7347–7351.

Xu, L., Li, Z., Li, J., & Hua, W. (2016). A novel bit-level image encryption algorithm based

Quantitative ﬂuorescence confocal laser scanning microscopy (CLSM) Handbook of bio- logical confocal microscopy. Springer.

on chaotic maps. Optics and Lasers in Engineering, 78, 17–25.

Semasaka, C., Buckow, R., & Kasapis, S. (2018). Modeling counterion partition in com- posite gels of BSA with gelatin following thermal treatment. Food Hydrocolloids, 74, 97–103.

Shrinivas, P., Kasapis, S., & Tongdang, T. (2009). Morphology and mechanical properties

142

Chapter 5 3D Confocal laser scanning microscopy for quantification

of the phase behaviour in agarose-MCC co-gels in

comparison to the rheological blending-law analysis

Mhaske, P., Farahnaky, A., & Kasapis, S. (2020). 3D Confocal Laser Scanning Microscopy

for Quantification of the Phase Behaviour in Agarose-MCC co-gels in Comparison to the

143

Rheological Blending-law Analysis. Food Biophysics, 1-8.

Food Biophysics https://doi.org/10.1007/s11483-020-09656-6

ORIGINAL ARTICLE

3D Confocal Laser Scanning Microscopy for Quantification of the Phase Behaviour in Agarose-MCC co-gels in Comparison to the Rheological Blending-law Analysis

Pranita Mhaske 1 & Asgar Farahnaky 1 & Stefan Kasapis 1

Received: 3 September 2020 / Accepted: 5 October 2020 # Springer Science+Business Media, LLC, part of Springer Nature 2020

Abstract The need for a rapid and direct alternative to the rheology-based blending laws in quantifying phase behaviour in biopolymer composite gels is explored in this study. In doing so, the efficacy of confocal laser scanning microscopy (CLSM) paired with image analysis software – FIJI and Imaris - in quantifying phase volume was studied. That was carried out in a model system of agarose with varying concentrations of microcrystalline cellulose (MCC) in comparison to the rheological blending laws. Structural studies performed using SEM, FTIR, differential scanning calorimetry and dynamic oscillation in-shear unveiled a continuous, weak agarose network supporting the hard, rod-shaped MCC inclusions where the composite gel strength increased with higher ‘filler’ concentration. The phase volumes of MCC, estimated with the microscopic protocol, matched the predictions obtained from computerized modelling using the Lewis-Nielsen blending laws. Results highlight the suitability of the micro- scopic protocol in estimating the water partition and effective phase volumes in the agarose-MCC composite gel.

Keywords Phase behaviour . Blending law analysis . 3D imaging . Confocal laser scanning microscopy . Image analysis

Introduction

using the classical methods of centrifugal separation or deter- mination of osmotic pressure from composite solutions [3, 4]. Other attempts to probe phase behaviour of composite sys- tems employed turbidimetric [5], spectrometric [6] and rheo- logical protocols. The latter, in particular, paired with blend- ing law analysis estimated accurately the phase behaviour of synthetic and natural polymer mixtures [7, 8]. In the latter, blending laws correlate the overall mechanical properties of the composite to those of the individual constituents taking into consideration the solvent partition between the two poly- meric phases. Though accurate in estimating phase volumes, their application demands considerable experimentation and modelling expertise. Clearly, there is a need for a rapid and direct approach capable of predicting the phase behaviour in biopolymer mixtures.

The recent years have seen a drastic increase in the use of proteins and polysaccharides in food product formulations to improve/control processability, shelf life, texture/mouthfeel and the kinetics of bioactive compound release [1]. For the most part, such techno-functionality is governed by the sol- vent partitioning amongst the macromolecules in the phase- separated mixture [2]. Fundamental understanding of solvent distribution and molecular interactions between constituents depends on their concentration, water addition, ionic strength, pH and processing, for example, thermal treatment or applied shear. Despite the range and depth of research in the literature, a thorough understanding of phase behaviour and interactions in complex biomaterial systems remains of great interest. These systems generally comprise highly viscous and gelled phases, whose phase behaviour cannot be accurately predicted

* Stefan Kasapis

stefan.kasapis@rmit.edu.au

Over the years, confocal laser scanning microscopy (CLSM) has been used sporadically in this area and proved popular for qualitative analysis [9–11]. A few studies coupled CLSM imaging with image analysis but they were limited to the investigation of single 2D images [12] and viscous aque- ous solutions [13]. Exploiting advances in image processing technology, Mhaske et al. (2019) developed a technique of quantifying phase volume in a lipid-biopolymer system using 3D CLSM imaging paired with image analysis [14]. The

School of Science, RMIT University, Bundoora West Campus, Plenty Road, Melbourne, VIC 3083, Australia

144

Food Biophysics

protocol yielded values for the volume of the lipid phase that depend entirely on its concentration (w/w) added to the sys- tem. These were in close agreement with the estimates obtain- ed from the rheology based theoretical modelling, highlight- ing the potential of the combined approach.

days in a Labconco FriZOne Triade freeze drier (Kansas City, Missouri, USA). These were then placed on 12.6 mm carbon coated aluminium stubs and gold coated. FEI Quanta 200 ESEM (Hillsboro, Oregon, USA) was used to capture the surface micrographs of the freeze dried samples. Images were captured in the high vacuum mode, with a spot size of 5, working distance of 12 mm and an accelerating voltage of 30 kV. A 50x magnifica- tion was used to obtain an overview of the gel network, whilst details of the composite topology were captured at 600 x.

Fourier Transform Infrared Spectroscopy (FTIR) A Spectrum 2 FTIR spectrometer by Perkin Elmer (Norwalk, Connecticut, USA) with a diamond crystal attenuated total reflectance de- vice from GladiATR (Pike Technologies, Maddison, Wisconsin, USA) was used to analyse single and binary mix- tures of agarose and MCC. The absorbance spectra were ob- tained at a resolution of 4 cm− 1 between 400 and 4000 cm− 1 and averaged over 64 scans in triplicate measurements by subtracting the spectrum of water.

The present work aims to quantify the phase behaviour of a complex, industrially relevant hydrogel comprising a structur- ing polysaccharide (agarose) and a rigid dispersion in the form of microcrystalline cellulose (MCC). The latter is also capable of absorbing and retaining water, hence being used in the cosmetics, plastic, pharmaceutical and cement industries. Applications in processed food are based on its functionality as a dietary fibre and bulking agent to manipulate the nutri- tional profile and consistency of preparations [15]. Thus, MCC’s interactions with other biopolymers is a topic of cur- rent fundamental and applied interest. In the present system, the retention of water by agarose and MCC is unknown and yet it will go a long way in determining the structural proper- ties of this mixture. Phase volume estimations will be attempted with the combined protocol of microscopic exami- nation and blending law modelling in order to confirm the suitability of this approach in identifying the solvent partition between two hydrophilic polymeric phases.

Materials and Methods

Materials and Sample Preparation

Differential Scanning Calorimetry (DSC) Thermal analysis of agarose and agarose-MCC mixtures was carried out using the Setaram Micro DSC VII (Setu-rau, Caluire, France). About 700 mg sample was transferred into a cylindrical vessel and sealed, with an equal weight of milliQ water being added to the reference vessel. Both vessels were then deposited into the instrument chamber and held at 40 °C for 30 min before ramping the temperature to 70 °C, cooling to 5 °C and once again heating to 70 °C. Averaged exotherms from triplicate runs, obtained at a constant scan rate of 1 °C/min, are reported. MCC paste, obtained by hydrating MCC overnight in excess water and centrifuging the suspension, was also analysed using a modulated DSC Q2000 (TA Instruments, New Castle, DE, USA). The instrument was interfaced to a refrigeration unit to achieve temperatures well below 0 °C and a nitro- gen purge at a flow rate of 25 mL/min was used. About 12 mg of sample was weighed in an aluminium pan and hermetically sealed, with an empty pan being the reference. Samples were subjected to successive heating and cooling routines between 10 and 350 °C at a constant scan rate of 1 °C/min. Triplicate runs produced consistent results.

Agarose (Type-1) and microcrystalline cellulose (Avicel, PH- 101) were supplied by Sigma-Aldrich (Sydney, Australia). The sulphate, ash and moisture content of agarose was less than 0.15, 0.25 and 10%, respectively, according to the sup- plier. The particle size of cellulose crystals was ~ 50 µm, with a relative density of 0.6 g/cm3. 5-DTAF [5-(4, 6– Dichlorotriazinyl)aminofluorescein], a bright yellow coloured powder was purchased from AAT Bioquest (California, USA) and used to covalently label agarose. Varying quantities of MCC (0.1, 0.3, 0.6, 0.9, 1.2, 1.5 g) were added to specific amounts of milliQ water to make the total volume of the aque- ous solutions 99 ml. This was stood overnight at ambient temperature to ensure complete hydration of the cellulose crystals. MCC suspensions were then heated to 80 °C in a water bath with constant magnetic stirring before dispersing 1 g agarose powder. Mixtures (1% (w/w) agarose and 0.1, 0.3, 0.6, 0.9, 1.2 and 1.5% (w/w) MCC) were stirred continuously for 15 min to dissolve the polysaccharide and cooled to 50 °C for subsequent analysis.

Experimental

Rheological Measurements Small-deformation dynamic oscil- lation in-shear was carried out on a Discovery Hybrid Rheometer (TA Instruments, New Castle, DE, USA) to mea- sure the elastic modulus (Gʹ) of the hydrated MCC paste, aga- rose gel and agarose-MCC composites. Samples were loaded at 50 °C onto the preheated Peltier plate with a 40 mm parallel plate geometry and 1 mm gap. The exposed edges of the com- posite were coated with polydimethylsiloxane (PDMS) to

SEM Analysis Gel cubes of agarose (10 mm3) with varying concentrations of MCC were freeze dried at -40 °C for 3

145

Food Biophysics

Statistical Analysis Phase volume results obtained were eval- uated using Minitab 18 (Minitab Inc., Pennsylvania, USA) by one-way analysis of variance (ANOVA). Volume measure- ment is done in duplicate and expressed in mean ± standard deviation, with significant differences being p < 0.05 (Tukey test).

minimise moisture loss. Analysis involved a temperature sweep from 50 to 5 °C at 2 °C/min, a 30 min isothermal step at 5 °C with a frequency of 1 rad/s, followed by a frequency sweep from 0.1 to 100 rad/s prior to ramping the temperature to 70 °C with a controlled strain of 0.01% throughout, which is within the linear viscoelastic region of this matrix (LVR).

Results and Discussion

Structural Characterisation of the Agarose/MCC Mixtures

Confocal Laser Scanning Microscopy (CLSM) and Image Analysis Agarose was covalently labelled with 5-DTAF (5-(4, 6-Dichlorotriazinyl) Aminofluorescein) using the procedure outlined by Russ et al. (2013) [16]. In doing so, 1 g of agarose was dissolved in 50 ml milliQ water at 80 °C with constant stirring. The solution was cooled to 60 °C and 8 mg 5-DTAF was added before slowly adding 15 ml of 1% (w/v) Na2SO4. Then, 3–4 drops of 10% (w/v) NaOH were added and stirred for 2 hours prior to adding 120 ml absolute ethanol to obtain a yellow precipitate. The mixture was refrigerated at -20 °C for 15 min before filter- ing it and washing the residue with absolute ethanol. The residue was then dried to obtain a fine yellow agarose powder.

Scanning electron microscopy is a technique capable of pro- viding tangible evidence on the morphology of biomaterials [17], and it is used in the present work to visualise the structure of agarose/MCC composites. As shown in Fig. 1a, agarose exhibits a highly porous, honeycomb like assembly under a low magnification of 50x. This assembly remains unchanged upon addition of MCC (images not shown), indicating that the gelation of the polysaccharide is independent of the presence of the cellulose crystal. The presence of MCC particles could not be seen at 50x magnification, which prompted us to in- crease it to 600x. As depicted in Fig. 1b, the walls of the agarose network appear uninterrupted and smooth at the higher magnification. Addition of MCC results in the devel- opment of rough surfaces, with rod-like MCC particles being embedded within the featureless background (Fig. 1c, d). As the concentration of MCC is raised from 0.6 to 1.2% (w/w), the surface roughness corresponding to the volume of MCC particles entrapped in the polysaccharide mesh also increases. SEM micrographs suggest that the topology of the composite gel consists of agarose strands forming a continuous matrix that supports the dispersed rod-shaped particles of the MCC filler.

The 5-DTAF-labelled agarose was used to make samples as outlined earlier. These were transferred into the well of a CoverWell imaging chamber (diameter – 20 mm; depth – 2.8 mm) and precooled to 4 °C to ensure rapid gelation of agarose before the settling of cellulose crystals. Nikon A1R laser fitted to an inverted microscope, ECLIPSE Ti-E (Minato, Tokyo, Japan) was used to obtain confocal images. An argon laser at 488 nm was used to excite the 5-DTAF-labelled aga- rose and the emissions were recorded between 500 and 530 nm. Fifteen 8-bit images, 1279 µm x 1279 µm in size were captured along the Z-axis, at an interval of 5.72 µm. The distribution of MCC from 9 different regions across the samples was studied.

Two image analysis software, Imaris 9.1.0 (Bitplane, Belfast, Northern Ireland) and FIJI with Java 6 (Madison, Wisconsin, USA) were used to quantify the Z-stacks obtained in parallel. The MCC phase volume was determined using the steps outlined by Mhaske et al. (2019) [14]. In Imaris, the Z- stack was opened in the “Surfaces” feature. A Gaussian filter was applied before subtracting the background intensity to smoothen the image and reduce noise. The volume of hollow spaces within the agarose network, attributed to the presence of unstained MCC particles, was measured. In FIJI, the “Default” algorithm was used to threshold the images, and the images were inverted to make the unstained MCC particles the object and the agarose phase the background. The area of the object pixels was calculated for the Z-stack and then mul- tiplied by the Z-step to get the volume of the MCC particles. Results obtained for the nine images captured across each sample were averaged before reporting the phase volume of the MCC particles.

Infrared spectroscopy was employed next to record the chemical fingerprints of agarose, MCC and their composites (Fig. 2). The former featured characteristic absorbance bands at 3650 cm− 1 (O-H stretching), 2980 and 2900 cm− 1 (C-H vibration), 1396 cm− 1 (C-C bending), 1246 cm− 1 (sulphate esters), 1068 cm − 1 (C-O-C vibration of 3,6-anhydro-galac- tose bridges) and 877 cm− 1 (anomeric carbon’s angular C-H deformation) [18, 19]. MCC interferograms showed the gen- eral attributes of cellulose, reproducing typical absorbance bands at 1059 cm− 1 (ring vibration and C-OH bending), 1160 cm− 1 (C-O and C-O-C stretching), 1640 cm− 1 (-O- ten- sile vibration neighbouring H atoms) and a broad band be- tween 3225 and 3350 cm− 1 (stretching vibrations of O-H groups) [20–22]. Composites reproduced the absorbance bands of both individual components, with no variation in their wavenumber, hence suggesting that there is no chemical interaction between the constituents of this mixture. Furthermore, as the concentration of MCC increased in the

146

Food Biophysics

Fig. 1 SEM images of (a) 1% (w/w) agarose network at 50x and, walls of 1% (w/w) agarose network containing (b) 0, (c) 0.6 and (d) 1.2% (w/w) MCC under 600x

exhibit a single exothermic event that reproduces the charac- teristics of the agarose thermogram in terms of shape and temperature band. Once more, it is confirmed that the pres- ence of MCC does not affect the structural properties of the agarose gel.

composite, the absorbance bands of both agarose and MCC displayed a slight amplification in intensity. This is an expect- ed outcome, congruent with the increase in polymer concen- trations in their respective domains as explained in the follow- ing sections.

In addition to infrared spectroscopy, differential scanning calorimetric tests were carried out to obtain information on the micromolecular aspects of the structural behaviour of our mix- ture. Featureless thermograms are recorded from the cooling cycles of the MCC paste in Fig. 3a indicating the absence of a first-order thermodynamic transition. The major features in the heating run are two endothermic peaks. The first event with a peak at about 108 °C indicates the desorption of water from the cellulose molecule, whereas the second event at 320 °C traverses the thermal decomposition of the MCC crys- tal [22]. Figure 3b depicts the exothermic peak of agarose upon cooling with a midpoint at 23 °C, reflecting the coil-to- helix transition of the polysaccharide leading to subsequent gel formation. Corresponding profiles of the blends also

Finally, we studied the mechanical behaviour of single and binary mixtures of agarose and MCC. Figure 4 reproduces typical patterns of elastic modulus variation as a function of temperature when the polymer is cooled from 40 to 5 °C at a rate of 2 °C/min. A significant increase in Gʹ values is ob- served at around 35 °C (onset of gelation), approaching as- ymptotically an equilibrium at the end of the cooling cycle (G’ is about 103.8 Pa at 5 °C). In contrast, the storage modulus of the MCC paste produced a relatively flat pattern that remained unaffected by temperature, yielding a storage modulus value of about 105.4 Pa (results not shown). Qualitatively, the me- chanical spectra of the composite exhibit a similar trend of storage modulus development upon cooling from vestigial gelation to eventual levelling off. As the MCC concentration

147

Food Biophysics

1160

hydrated MCC paste

1640

1059

Fig. 2 FTIR spectra between 400–4000 cm-1 for agarose with 0, 0.1, 0.3, 0.6, 0.9, 1.2 and 1.5% (w/w) MCC and MCC by itself stacked upwards from the bottom

3225-3350

1% agarose + 1.5% MCC

1% agarose + 1.2% MCC

1% agarose + 0.9% MCC 1% agarose + 0.6% MCC

e c n a b r o s b A

1% agarose + 0.3% MCC

1% agarose + 0.1% MCC

1068

2980

2900

1396 1246

3650

1% agarose

877

4500

4000

3500

3000

1500

1000

500

2500

2000

Wavenumber (cm-1)

agarose phase and increasing its effective concentration, an outcome that leads to the formation of a stronger hydrogel.

Rheological Modelling of the Phase Behaviour in Agarose/MCC Composite Gels

increases, however, so does the mechanical strength of the composite, reaching 104.1 Pa with 1.5% (w/w) MCC addition. This outcome is partially attributed to the reinforcing effect of the hard MCC particles on the softer agarose matrix. It should also be noted that the cellulose particles absorb water [23], thereby decreasing the amount of solvent available to the

1000

Cooling

500

)

Heating

W m

(

-500

w o l f t a e H

-1000

-1500

Analysis thus far, established qualitatively the noninteractive na- ture of MCC crystal and agarose molecules leading to a filler composite with a relatively weak matrix supporting hard inclu- sions. We are now interested to relate quantitatively the mechan- ical strength of the constituent phases to the overall behaviour of the composite gel. That was initiated by obtaining a double- logarithmic calibration curve of storage modulus for the agarose network at concentrations between 1.0 to 2.5% (w/w); reported in the inset of Fig. 4. Mechanical properties were then followed by a set of equations adapted from the ‘sophisticated synthetic polymer research’ proposed by Lewis and Nielsen [24]:

-2000

100

300

350

400

150

250

ð1Þ

200 Temperature (°C)

G0 G

¼ 1 þ ABΦy 1 − BψΦy

0 x

1.5

− 1

1% (w/w) agarose + 1.5 % MCC

ð2Þ

B ¼

1% (w/w) agarose + 1.2 % MCC

y 0 G x 0 G

1% (w/w) agarose + 0.9 % MCC

)

0.5

y 0 x

1% (w/w) agarose + 0.6 % MCC

W m

(

1% (w/w) agarose + 0.3 % MCC

ð3Þ

Φy

1% (w/w) agarose + 0.1 % MCC

þ A ψ ¼ 1 þ 1−Φm Φ2 m

w o l f t a e H

-0.5

1% (w/w) agarose

-1

-1.5

100

150

-50

50 Temperature (°C)

Fig. 3 DSC thermograms of (a) cooling and heating profiles of hydrated MCC paste and (b) agarose cooling profiles with 0, 0.1, 0.3, 0.6, 0.9, 1.2 and 1.5% (w/w) MCC arranged successively upwards scanned at a rate of 1 °C/min

where, Gʹ, Gʹx, Gʹy are the elastic modulus of the composite, the continuous agarose network and the discontinuous MCC phase, respectively, and ɸy is the phase volume of the filler. The framework considers the concept of maximum packing fraction (ɸm) of the filler particles, which corresponds to 0.52 for random MCC rods. The shape of the filler is taken into account via the Einstein co-efficient, rendering A to be 2.08 for dispersed cylindrical, rod-shaped inclusions [25].

148

Food Biophysics

4.2

3.7

5.5

y = 3.7321x + 3.8717 R² = 0.9945

) a P (

′

3.2

Fig. 4 Profiles of Gʹ for 1% (w/w) agarose with 0 (●), 0.1% (x), 0.3% (▲), 0.6% (□), 0.9% (♦), 1.2% (-) and 1.5% (○) (w/w) MCC during cooling from 50 to 5 °C at a scan rate of 2 °C/min; the agarose calibration curve of Gʹ at 5 °C as a function of its concentration is illustrated in the inset

4.5

) a P (

′

G g o L

2.7

3.5

0.1

0.3

0.4

0.2 Log agarose concentration (%)

2.2

20 Temperature (°C)

3D CLSM Imaging and Image Analysis for Phase Volume Estimations in the Agarose/MCC Mixture

In an effort to determine the contribution of each com- ponent to the elastic modulus of the composite, prepara- tions were centrifuged at 40 °C and 10,000 rpm for 15 min. Supernatants of clear agarose solutions were decanted and their weights were recorded in order to determine the ef- fective agarose concentration, which ranged from 1.01 to 1.08% (w/w). Respective values of Gʹx for these agarose concentrations were calculated using the calibration curve from the inset of Fig. 4. Centrifuged MCC pellets were also weighed to work out the amount of absorbed water, which was ~ 4.1 times the weight of the cellulose powder in each composite. Weights of the decanted agarose supernatant and MCC pellet obtained after centrifugation were used to calculate the values of ɸx and ɸy (phase volumes of agarose and MCC, respectively) for each composite, with ɸx + ɸy = 1.

In order to capture and then process 3D images of our com- posites, agarose was covalently labelled with DTAF and the procedure outlined by Mhaske et al. (2019) [14] was followed. Figure 6a depicts a typical 2D image of the surface of the gel containing 1% (w/w) agarose and 0.6% (w/w) MCC. Stained agarose forms a continuous blue background in which the unstained MCC particles are dispersed as black inclusions. This is an 8 bit 2D image with a dimension of 512 × 512 pixels, corresponding to an area of 1279 × 1279 µm. Fifteen such images were obtained along the Z-axis at a fixed interval or Z-step of 5.72 µm, reproduced in Fig. 6b, which allowed imaging the sample up to a depth of ~ 85 µm. The Z-stack was then utilized to generate a 3D image, as seen in Fig. 6c. FIJI, which is an open source analysis software, analyses the Z- stack for phase volume estimation by quantifying each 2D

4.1

)

(

3.9

) a P

C C M

, x (

′

3.8

G g o L

3.7

f o e m u l o v e s a h P

3.6

3.5

0.6

0.3

1.2

1.5

1.8

0.9 MCC concentration (%, w/w) Fig. 5 Values of composite Gʹ obtained via rheological experiments (x) and blending law predictions for G’ (solid line) and phase volumes of MCC (♦) for samples containing 1% (w/w) agarose with increasing MCC concentrations

The storage modulus value of the MCC paste (about 251 kPa), obtained from rheological measurements in the preceding section, was used as a constant Gʹy for all com- posites, with the assumption that equally hydrated MCC particles in the various composites (0.1 to 1.5% w/w MCC) would yield the same elastic response. Blending law modelling was able to produce a Gʹ trace that exhib- ited a close fit with the corresponding experimental values of the composites depicted in Fig. 5, an outcome that showcases its ability to predict phase behaviour in binary composites. Corresponding volumes of the MCC phase in all composites, as predicted by blending law, are also depicted in Fig. 5 being congruent with the amount of experimentally added polysaccharide. Phase volume pre- dictions of MCC will be compared with those obtained from microscopy, paired with image analysis, in the fol- lowing section in order to assess the suitability of the microscopic protocol in quantifying phase behaviour in binary co-gels.

149

Food Biophysics

Fig. 6 (a) 2D, (b) Z-stack, and (c) 3D CLSM images of 1% (w/w) agarose containing 0.6% (w/w) MCC. The blue background denotes the stained agarose phase whereas the black cavities correspond to the unstained MCC particles

image separately. In contrast, Imaris renders a 3D image using the stack of 2D images before quantifying the 3D image volumetrically.

et al. (2019) [14] was followed to estimate the MCC phase volumes and results are also shown in Fig. 7. Similarly, the Imaris predictions are a close fit to those estimated using FIJI image analysis and rheological blending laws. Statistically, the difference between the phase volume values obtained from the two types of imaging software is not significant (p > 0.05) indicating that both can equally measure water absorption and its retention in the polymeric phases of the binary mixture. Nevertheless, the two types of software have distinct ap- proaches of image quantification and their “plug-ins” vary in

FIJI

Imaris

)

(

C C M

f o e m u l o v e s a h P

0.1

0.3

0.6

0.9

1.2

1.5

Concentration of MCC (%)

In an 8 bit image, pixels denote a specific intensity value of the wavelength emitted by the sample within the range of 0 and 255, where a pixel with the intensity of 0 appears black and refers to an unstained sample. This serves as the negative control and is considered to be the baseline calibration in image analysis [26]. Thresholding is the most critical step in image analysis in order to accurately segregate the object pixels (MCC particles) from the background pixels (agarose network) [27]. The phase volume of MCC particles in the composite was measured using the image analysis software, FIJI. Figure 7 illustrates the final phase volume of MCC in the composite in relation to the original amount of added cellulose powder. The phase volumes estimated by FIJI are comparable with those obtained from the rheological blending-law analy- sis in Fig. 5. Increase in the MCC phase volume in the com- posite also corresponds to ~ 4.1 times the initial added amounts documented in Fig. 7 microscopically and in Fig. 5 rheologically from the MCC concentration (%, w/w) vs.. phase volume of MCC (%) relationship. Phase volume esti- mations, of course, allow quantification of the amount of sol- vent held in a given polymeric phase.

Fig. 7 Phase volume estimates plotted against MCC concentrations in mixture with 1% (w/w) agarose determined by analysing 3D CLSM images using image analysis software – FIJI and Imaris

Next, we turned our attention to the second image analysis software, Imaris. Once more the protocol used by Mhaske

150

Food Biophysics

Rheological Properties of Materials· Rheo-Optics· Biorheology. (Springer, Berlin, 1975), p. 594–600

customisability, which may affect their capacity in estimating phase volume in more complex systems, for example, in ter- tiary preparations of protein, polysaccharide and a lipid phase.

8. M. Takayanagi, H. Harima, Y. Iwata, Viscoelastic behaviour of polymer blends and its comparison with model experiments. Mem. Fac. Eng. Kyusha Univ. 23, 1–13 (1963)

Conclusions

9. X. Li et al., Studies of phase separation in soluble rice protein/ different polysaccharides mixed systems. LWT Food Sci. Technol. 65, 676–682 (2016)

10. T. Moschakis et al., in Microrheology and microstructure of water- in-water emulsions containing sodium caseinate and locust bean gum. Food Funct. 9(5), 2840–2852 (2018)

13.

11. H. Pan et al., Effect of the extent and morphology of phase separa- tion on the thermal behavior of co-blending systems based on soy protein isolate/alginate. Food Hydrocoll. 52, 393–402 (2016) 12. N. Lorén, M. Langton, A.M. Hermansson, Confocal laser scanning microscopy and image analysis of kinetically trapped phase- separated gelatin/maltodextrin gels. Food Hydrocoll. 13(2), 185– 198 (1999) J.C.G. Blonk et al., A new CSLM-based method for determination of the phase behaviour of aqueous mixtures of biopolymers. Carbohyd. Polym. 28(4), 287–295 (1995)

15.

14. P. Mhaske et al., Quantitative analysis of the phase volume of agarose-canola oil gels in comparison to blending law predictions using 3D imaging based on confocal laser scanning microscopy. Food Res. Int. 125, 108529 (2019) J. Nsor-Atindana et al., Functionality and nutritional aspects of microcrystalline cellulose in food. Carbohyd. Polym. 172, 159– 174 (2017)

Rheology-based blending laws have been used extensively in the literature to predict the mechanical strength of synthetic polymer matrices, and this is augmented in biomaterial com- posite gels by the estimation of solvent partition between polymeric phases, since these are largely aqueous prepara- tions. Nevertheless, such attempts are built around painstaking rheological experimentation and they also require consider- able modelling expertise, hence making it difficult for labora- tories to implement or reproduce. Given this, the suitability of CLSM based 3D imaging, coupled with image analysis, in quantifying phase volumes in a binary mixture was investigat- ed as a rapid alternative approach. The microscopic protocol could successfully render phase volumes of hydrated MCC particles suspended in an agarose network, an outcome that agreed remarkably well with those obtained from rheological modelling. The protocol could cope with increasing concentrations/density of the filler phase in phase volume es- timation and merits further consideration in more complex preparations.

16. N. Russ et al., Influence of nongelling hydrocolloids on the gelation of agarose. Biomacromolecules 14(11), 4116–4124 (2013) 17. C. Kyomugasho et al., Evaluation of cation-facilitated pectin-gel properties: Cryo-SEM visualisation and rheological properties. Food Hydrocoll. 61, 172–182 (2016)

18. X. Qi et al., Facile formation of salecan/agarose hydrogels with tunable structural properties for cell culture. Carbohyd. Polym. 224, 115208 (2019)

Acknowledgements The authors acknowledge the facilities and technical assistance of the RMIT University’s Microscopy and Microanalysis Facility.

19. V. Zamora-Mora et al., Chitosan/agarose hydrogels: Cooperative properties and microfluidic preparation. Carbohydr. Polym. 111, 348–355 (2014)

References

20. C.P. Azubuike, A.O. Okhamafe, Physicochemical, spectroscopic and thermal properties of microcrystalline cellulose derived from corn cobs. Int. J. Recycl. Org. Waste Agric. 1(1), 9 (2012) 21. H. Toğrul, N. Arslan, Flow properties of sugar beet pulp cellulose and intrinsic viscosity–molecular weight relationship. Carbohyd. Polym. 54(1), 63–71 (2003)

1. K. Prameela, C.Murali Mohan, C. Ramakrishna, Chap. 1 - Biopolymers for Food Design: Consumer-Friendly Natural Ingredients, in Biopolymers for Food Design, ed. by A.M. Grumezescu, A.M. Holban (Academic Press, Cambridge, 2018), p. 1–32

22. M. Murigi et al., Comparison of physicochemical characteristics of microcrystalline cellulose from four abundant Kenyan biomasses. IOSR J. Polym. Text. Eng. 1(2), 53–63 (2014)

2. C. Stenger et al., Formation of concentrated biopolymer particles composed of oppositely charged WPI and pectin for food applica- tions. J. Dispersion Sci. Technol. 38(9), 1258–1265 (2017) 3. V. Tolstoguzov, Ionic strength and poly ion charges. Int. Food

23. A.J. Gravelle, S. Barbut, A.G. Marangoni, Food-grade filler parti- cles as an alternative method to modify the texture and stability of myofibrillar gels. Sci. Rep. 7(1), 11544 (2017)

Ingred. 2, 8–9 (1990)

24. C. Semasaka et al., Modeling water partition in composite gels of BSA with gelatin following high pressure treatment. Food Chem. 265, 32–38 (2018)

25. L.W. Koh, S. Kasapis, Orientation of short microcrystalline cellu- lose fibers in a gelatin matrix. Food Hydrocoll. 25(5), 1402–1405 (2011)

4. M.M.G. Koning, J. van Eendenburg, D.W. de Bruijne, Mixed Biopolymers in Food Systems: Determination of Osmotic Pressure, in Food Colloids and Polymers, ed. by E. Dickinson, P. Walstra (Woodhead Publishing. Cambridge, 2005), p. 103–112 5. T. Giancone et al., Protein–polysaccharide interactions: Phase be- haviour of pectin–soy flour mixture. Food Hydrocoll. 23(5), 1263– 1269 (2009)

26. A.A. Karamitrou, G.N. Tsokas, M. Petrou, A pixel-based semi‐ stochastic algorithm for the registration of geophysical images. Archaeol. Prospect. 24(4), 413–424 (2017)

27. T.Y. Goh et al., Performance analysis of image thresholding: Otsu

technique. Measurement 114, 298–307 (2018)

6. P.D. Pudney, T.M. Hancewicz, D.G. Cunningham, The use of con- focal Raman spectroscopy to characterise the microstructure of complex biomaterials: foods. J. Spectrosc. 16(3–4), 217–225 (2002)

Publisher’s Note Springer Nature remains neutral with regard to jurisdic- tional claims in published maps and institutional affiliations.

7. L.E. Nielsen, in Morphology and the elastic modulus of block poly- mers and polyblends, in Rheological Theories· Measuring Techniques in Rheology Test Methods in Rheology· Fractures

151

Chapter 6 Comparison of rheological blending-law analysis and 3D

confocal laser scanning microscopy for quantification of the

phase behaviour in agarose-gelatin co-gels

Abstract

The efficacy of image analysis software – Imaris and FIJI – in quantifying 3D images

captured using confocal laser scanning microscopy (CLSM) to estimate phase behaviour in a

mixed gelling system of a polysaccharide (agarose) and a protein (gelatin) in comparison with

the rheology based blending laws is explored. Structural properties of the composites were

probed using FTIR spectroscopy, micro DSC and small-deformation dynamic oscillation in

shear. Throughout the experimental range of concentrations, gelatin formed a continuous

network whereas the agarose phase remained dispersed as either a soft or a hard filler at low

agarose concentrations (0.1-0.8%) and formed a continuous network alongside that of gelatin

at higher concentrations (1.2-3%). The phase volumes of gelatin, recorded using Imaris, were

a close match with those obtained from the blending law predictions, whereas FIJI yielded

statistically different estimates. Results suggest that the microscopic protocol using Imaris

shows promise and can potentially be used as an alternative to the theoretical blending laws in

estimating phase behaviour in biomaterial composites.

Key words: Blending law analysis; phase behaviour; 3D imaging; image analysis; confocal

152

laser scanning microscopy;

6.1. Introduction

A food system typically consists of several components, from simple molecules to macro–

molecular aggregates, held together by intermolecular forces and arranged in several structural

levels, for instance, colloidal phase. Geometric features and interactions among food

components are responsible for the ‘food structure’ which has a direct impact on the texture of

the product. A thorough understanding of such interactions is therefore important for

developing food products with intriguing texture, different from that of the original components

(Sikorski, Pokorny, & Damodaran, 2008).

Texture of the food product can be manipulated by altering the processing conditions (e.g.

ionic strength, pH, temperature), polymer concentration used, or by using a combination of

polymers at varying concentrations. When used in combination, biopolymer can exhibit

attractive or repulsive interactions resulting in complexation or thermodynamic

incompatibility, respectively based on their nature and the solvent conditions (Çakır &

Foegeding, 2011). Out of these, thermodynamic incompatibility leading to phase separation in

a protein and polysaccharide mixture is widely observed due to a low mixing entropy (Grinberg

& Tolstoguzov, 1997). However, when one or more components of a composite are gel

forming, gelation and phase separation act as competing processes, with their kinetics affecting

the morphology of the final microstructure.

The interplay between gel formation and phase separation has been successfully studied

using techniques like light scattering (Norton & Frith, 2001), spectroscopy (Anvari, Pan, Yoon,

& Chung, 2015), microscopy (Geonzon, Bacabac, & Matsukawa, 2019), calorimetry (Wang,

Dumas, & Gharsallaoui, 2019) and rheology (Wurm, Pham, & Bechtold, 2019). Rheology

based blending law analysis in particular, have been extensively used to estimate phase

153

behaviour of synthetic and natural polymer composites. These laws estimate the mechanical

properties of the mixtures from the shear modulus of the constituents in relation to their phase

volume and effective concentration (Nielsen, 1975; Takayanagi, 1963). The blending laws

however are an indirect approach, based on assumptions (e.g. the onset of gelation brings about

a substantial increase in the molecular weight of the biopolymer), and are a time-consuming

approach necessitating considerable rheological work and modelling expertise.

In the recent past, confocal laser scanning microscopy has seen increasing popularity for

qualitatively probing phase behaviour in biopolymer mixtures (Lan, Ohm, Chen, & Rao, 2020;

Zhang et al., 2017). In their work, Mhaske et al. (2019) developed a method of estimating phase

volume in a binary model system (lipid-polysaccharide) using CLSM based 3D imaging

coupled with image analysis. The technique could also accurately quantify solvent partition

and phase volume in a composite where both components were hydrophilic (agarose and

microcrystalline cellulose) (Mhaske, Farahnaky, & Kasapis, 2020). Given that the technique is

quick, easy, requiring minimum sample preparation and capable of producing phase volume

estimates that are highly comparable to those obtained from theoretical blending laws, it is

worth probing its efficacy in measuring phase volume in a complex, industrially relevant

system comprising of two gel forming polymers.

Agarose and gelatin are both extensively used as texturisers, stabilisers, emulsifiers and

gelling agents in the food, pharma, and cosmetic industry. There is a plethora of research

probing their gelation mechanism, effect induced by changes in physical and chemical

environment and their subsequent applications. Therefore, in the present study, the phase

behaviour and topology of agarose-gelatin mixtures is revisited and followed using the classical

blending laws at increasing polysaccharide concentration. Due to their inherent thermodynamic

incompatibility, the two polymers are expected to phase separate, yielding varying phase

154

topologies depending on their concentrations. The phase volume estimates obtained using the

blending laws will be compared to those from the microscopic protocol coupled with image

analysis software.

Successful implementation of this protocol will allow application to other composites of

commercial interest, making it easier for laboratories to probe the phase behaviour of

composites without painstaking rheological experimentation and modelling expertise.

6.2. Materials and methods

6.2.1. Materials

Type -1 Agarose and Type- B bovine gelatin was procured from Sigma-Aldrich (Sydney,

Australia). Agarose had less than 0.15% sulphate, 0.25% ash and 10% moisture content, whilst

gelatin had a molecular weight of about 173kDa and a Bloom value of 225 according to the

supplier. Fluorescein isothiocyanate (FITC) was also purchased from Sigma-Aldrich (Sydney,

Australia) and used to stain the protein.

6.2.2. Methods

6.2.2.1. Sample preparation

Single solutions of agarose and gelatin were prepared by dispersing powders in milliQ

water at 75 and 60°C, respectively to produce 3% aqueous solutions of each. Binary solutions

had a constant gelatin concentration of 3% and varied agarose concentrations to yield 0.1, 0.4,

0.8, 1.2, 1.6, 2.0, 2.5 and 3% of agarose in the final composite. Solutions were prepared by

dissolving required quantities of agarose in milliQ water at 75°C and cooling the mixture down

to 60°C before adding gelatin. Samples were stirred at 60°C using a magnetic stirrer for 60

155

minutes for homogeneous dissolution prior to analysis.

6.2.2.2. Fourier Transform Infrared Spectroscopy (FTIR)

Single and binary solutions of agarose and gelatin were analysed using a Perkin Elmer

(Norwalk, Connecticut, USA) spectrometer, Spectrum 2 FTIR with an attenuated total

reflectance device by GladiATR (Pike Technologies, Maddison, Wisconsin, USA). The

absorbance spectra were recorded between 400 and 4000 cm-1 with a resolution of 4 cm-1 and

averaged over 32 scans before subtracting the spectrum of water.

6.2.2.3. Differential Scanning Calorimetry (DSC)

A Setaram Micro DSC VII (Setu-rau, Caluire, France) was employed for the thermal

analysis of agarose, gelatin, and their composites. About 700 mg sample was added to

aluminium vessels and sealed, while equal weight of water was used as reference. The vessels

were transferred into an insulated instrument chamber and equilibrated at 25°C for 30 min. The

temperature was then raised to 80°C, lowered to 5°C and ramped to 80°C again. Thermographs

reported are the averaged triplicate scans obtained at a scan rate of 1°C/min.

6.2.2.4. Rheological analysis

The elastic moduli of agarose, gelatin and their composites were measured using small-

deformation dynamic oscillation in-shear on a Discovery Hybrid Rheometer by TA instruments

(New Castle, DE, USA). Samples were loaded on a preheated Peltier plate at 45°C with a 40

mm parallel plate geometry and a fixed gap of 1mm. Exposed edges of the sample were coated

with silicon oil to prevent moisture loss. Samples were cooled to 5°C, followed by a 30 min

156

time sweep and a frequency sweep from 0.1 to 100 rad/s at 5°C with a final heating run to

80°C. For all the tests, a scan rate of 2°C/min and a strain of 0.1%, which is within the linear

viscoelastic region (LVR), were used.

6.2.2.5. Confocal Laser Scanning Microscopy (CLSM) and Image Analysis

0.001% w/w FITC was added to the agarose-gelatin composites prepared in section

2.2.1 and stirred for 60 min to stain the protein. Samples were then transferred into wells of

CoverWell imaging chambers (20mm wide and 2.8 mm deep) and allowed to gel. An inverted

microscope, ECLIPSE Ti-E (Minato, Tokyo, Japan) fitted with a Nikon A1R laser was used to

capture confocal images at 10x. A 488nm argon laser was used to excite the FITC-labelled

gelatin and the emissions were captured between 500 and 530 nm. Thirty 8-bit images, 1279

µm x 1279 µm in size were captured along the Z-axis at an interval of 10 µm. A Z-stack of 2D

images from ten sections across each sample was captured to study the distribution of agarose

and gelatin phases.

The images captured were analysed in parallel using two image analysis software: FIJI

with Java 6 software(Madison, Wisconsin, USA) and Imaris 9.1.0 (Bitplane, Belfast, Nothern

Ireland). The phase volume of gelatin was estimated using the procedure outlined by Mhaske

et al. (2019). In FIJI, the images were thresholding using the “Default” algorithm. The area of

the object (gelatin) pixels was calculated for each 2D image of the Z-stack. This, multiplied by

the Z-step generated the volume of the gelatin phase. Whereas in Imaris, the Z-stack was

opened using the ‘Surfaces’ feature and the background intensity was subtracted to reduce

noise and smoothen the image. The total volume of the gelatin phase was estimated using the

generated 3D image.

6.2.2.6. Statistical Analysis

The phase volume estimates were evaluated using one-way analysis of variance

157

(ANOVA) using Minitab 18 (Minitab Inc., Pennsylvania, USA). Volume measurement was

carried out in duplicate and expressed as mean ± standard deviation, with the significant

differences being p<0.05 (Tukey test).

6.3. Results and discussion

6.3.1. Structural characterisation of agarose-gelatin mixtures

Fourier transform infrared spectroscopy was used as a preliminary test to confirm the

absence of molecular interaction between agarose and gelatin by comparing the chemical

fingerprints of the constituents with those of the composites (Fig. 6.1). The interferogram of

agarose featured a series of characteristic chemical moieties, i.e. C-H angular deformation of

anomeric carbon (877 cm-1), sulphate esters (1246 cm-1), C-O-C vibration of 3,6-anhydro-

galactose bridges (1078 cm-1), C-C bending (1396 cm-1), C-H vibration (2900 cm-1) and O-H

stretching (2980 cm-1) (Awadhiya, Tyeb, Rathore, & Verma, 2017; Qi et al., 2019). Spectral

scan of gelatin yielded major peaks in four different amide regions – Amide A denoting the N-

H bending vibrations (3325-3157), Amide I arising due to C=O stretching vibration, NH

bending and C-N stretching (1700-1600 cm-1), Amide II absorption band containing

contributions from C-N stretching and N-H bending vibrations (1565-1520 cm-1) and Amide

III representing N-H in-bending vibrations paired with C-N stretching with contributions from

C=O in-plane bending and C-C stretching (1240-670 cm-1) (Fig. 6.1a) (Cebi, Durak, Toker,

Sagdic, & Arici, 2016; Mirzapour‐Kouhdasht, Sabzipour, Taghizadeh, & Moosavi‐Nasab,

2019).

The spectral scans of the composites reproduced the absorbance bands of both the

constituents without a variance in their wavenumber, thereby suggesting that there is no

chemical interaction between the individual components of the composite (Fig. 6.1b).

158

However, a slight increase in the intensity of all the absorbance bands is observed on increasing

the agarose concentration. This is an expected outcome attributed to the increase in the effective

polymer concentration in the respective phases with the increase in agarose concentration in

the composite.

Figure 6.1.: FTIR spectra for (a) single systems of 3% agarose and 3% gelatin and (b) 3%

159

gelatin with 0.1, 0.4, 0.8, 1.2, 1.6, 2, 2.5 and 3% agarose arranged upwards.

Calorimetric tests were carried out for the complementary characterisation of phase

behaviour in aqueous mixtures of agarose and gelatin. Exotherms of the protein,

polysaccharide, and their mixtures during a cooling run from 80 to 5°C are depicted in Fig. 6.2.

A single system of 3% agarose denoted by the dashed line exhibits an exothermic peak with a

midpoint at 25°C, reflecting the formation of a gel network due to the coil to helix transition

of the agarose strands. The exotherm of gelatin (dotted line) yielded a peak culminating at a

much lower temperature of 12°C attributed to its gelation.

Figure 6.2: Cooling profiles of single systems of 3% gelatin (dotted line), 3% agarose (dashed

line), and 3% gelatin with 0.1, 0.4 0.8, 1.2, 1.6, 2, 2.5 and 3% agarose (solid lines) arranged

from bottom to top, obtained at a scan rate of 1°C/min using DSC.

When mixed, agarose and gelatin maintained the temperature sequence and overall

peak form of each transition in the composite (Fig. 6.2). A dominant peak denoting the gelation

of gelatin was observed between 0 and 20°C in all composites. Whereas the peak attributed to

the gelation of agarose is virtually absent in samples with the lowest concentrations of agarose

160

and increases in intensity as the polysaccharide concentration is increased in the composite. An

obvious explanation of the observed behaviour is that gelatin forms a continuous matrix in all

the samples, whereas contributions from the formation of agarose network are only evident in

samples containing increased agarose concentrations. As the composites neither exhibit

distortion of the individual peaks of the constituents nor generate new thermal events it can be

argued that agarose and gelatin remain in two distinct, phase separated domains in the

composite without any heterotypic association. It is worth noting that as the agarose

concentration in the composites is increased, the temperature band of agarose in the composites

shifts to higher temperatures signifying an increased thermal stability of the concentrated

agarose phase in the composite. Similar results supporting phase separation are reported in the

works of Shrinivas, Kasapis, & Tongdang (2009), who investigated the phase behaviour of

binary and ternary mixtures of agarose, gelatin, and a lipid.

Finally, mechanical measurements were carried out to obtain supramolecular evidence

on structure formation of the constituents and their mixtures. The asymmetric variations in the

values of storage modulus of single agarose and gelatin systems during a cooling and a heating

run separated by a 30 min isothermal run at 5°C exhibit a similar sequence of structure

formation as for the calorimetric thermograms in Fig. 6.2, and are depicted in Fig. 6.3a. When

the disordered and fluctuating coils of 3% agarose in an aqueous solution are cooled to 5°C,

they form an ordered three-dimensional network resulting in a significant increase in the

storage modulus, commencing at about 33.5°C. Substantial thermal hysteresis is observed

during the heating run, with the network melting at 80°C. This is a result of the highly

aggregated intermolecular associations forming the structural knots of the enthalpic network

(Fernández et al., 2008). Gelatin on the other hand, exhibits a gelation and a melting

temperature of 20.7°C and 26.2°C, respectively. A constant thermal hysteresis of six to seven

degrees throughout the experimental range for gelatin argues for a gelation brought about by

161

helices as opposed to aggregation (Parker & Povey, 2012).

Figure 6.3: Variation in storage modulus (a) as a function of temperature and time of

observation for single systems of 3% gelatin and 3% agarose, and (b) on heating 3% gelatin

plus 0.1, 0.4, 0.8, 1.2, 1.6, 2, 2.5 and 3% agarose, arranged upwards. Scan rate: 2°C/min,

162

frequency: 1 rad/s, strain: 0.1%.

Figure 6.3b shows the melting profile of the composites containing 3% gelatin with

increasing concentrations of agarose during controlled heating at 2°C/min. At the lower

experimental concentrations of agarose (0.1, 0.4 and 0.8% in Fig. 6.3b), the composite network

melts completely at the temperature range of gelatin on heating. This disintegration of the

composite with gelatin melting suggests that at low concentrations of agarose, gelatin forms a

continuous network within which agarose is dispersed as a discontinuous filler. In contrast,

when the concentration of agarose is increased to 1.2% or more, upon heating, composites

show a decrease in composite moduli on the melting of the gelatin network, but the gel remains

intact within the experimental range of temperature. This postulates that at increased

concentrations of agarose, the polysaccharide too forms a continuous network that holds the

composite gel together after gelatin melts. Thus, the melting of gelatin weakens the composite

gel but does not destroy it. This is congruent with the calorimetry results which yield a single

peak of gelatin gelation at low agarose concentrations along with the gradual development of

the second peak illustrating the formation of agarose network in composites containing a higher

concentration of agarose.

Concluding evidence on the phase behaviour of agarose-gelatin co-gels is afforded by

confocal microscopy in Fig. 6.4. The gelatin phase which is stained using FITC appears purple

in the images while the unstained agarose appears black. Figs. 6.4a and 6.4b denote samples

containing 3% gelatin with 0.1 and 0.8% agarose, respectively. It is evident from the purple

background of the images that gelatin forms a continuous matrix in which agarose is dispersed

as a discontinuous filler. The filler particles increase in number as the concentration of agarose

163

increases (Fig. 6.4b vs Fig. 6.4a).

Figure 6.4: 2D CLSM images of 3% gelatin with (a) 0.1, (b) 0.8, (c) 1.6 and (d) 2.5% agarose,

at 10x. Agarose phase is denoted in black and gelatin phase in purple. Scalebar denotes 100

µm.

Fig. 4c representing the composite containing 3% gelatin and 1.6% agarose shows

larger aggregates of agarose branching out to form a continuous network. In Fig. 4d, sample

containing 3% gelatin and 2.5% agarose depicts a clear bicontinuous network meshed together.

This confirms the hypothesis made in the preceding sections that at lower concentrations of

agarose (0.1, 0.4 and 0.8%), the composite gel has a continuous gelatin phase with agarose as

the discontinuous filler, and at higher agarose concentrations (1.2, 1.6, 2.0, 2.5 and 3%),

agarose too forms a continuous matrix along with gelatin.

6.3.2. Theoretical modelling of phase behaviour in agarose-gelatin co-gels using blending

law analysis

In order to advance a quantitative footing to the qualitative observations on phase behaviour

164

made thus far, we obtained double logarithmic calibration curves of elastic moduli within the

experimental range of concentrations for agarose (between 1 and 3.25%) and gelatin (between

1 and 10%) at 5°C; reported in Fig. 6.5a. The calibration curves with a correlation coefficient,

R2 of 0.995, are indispensable for interpolating between the experimentally obtained Gʹ values

at varying concentrations during theoretical modelling. The isostrain and isostress blending

laws given by Eqs. 1 and 2, respectively are used to follow progress in viscoelasticity of the

composites (Morris, 2009):

−1

(1) 𝐺ʹ𝑐 = ɸ𝑥𝐺ʹ𝑥 + ɸ𝑦𝐺ʹ𝑦

ɸ𝑦 𝐺ʹ𝑦

(2) + ) 𝐺ʹ𝑐 = (ɸ𝑥 𝐺ʹ𝑥

Where Gʹc, Gʹx, and Gʹy are the elastic moduli of the composite and the constituent phases X

and Y, respectively, with ɸx and ɸy being the respective phase volumes of the constituents and

ɸx + ɸy = 1. According to this framework, the isostrain conditions or upper bound behaviour is

applicable when a strong continuous matrix supports a weaker discontinuous filler, and in the

converse situation where a stronger filler is dispersed in a weak supporting matrix, the isostress

conditions or lower bound predictions are suitable.

The Takayanagi (isostrain and isostress) blending laws predict the solvent distribution

between the constituent biopolymer phases by estimating the elastic modulus on shear for all

possible solvent (water) distributions in the gelatin phase (SGEL), via the calibration curves in

Fig. 6.5a, and finding the one estimated value that matches the experimental storage modulus

of the composite (Gʹexp). Examples of the application of blending law analysis to the agarose-

gelatin co-gels in the present experimental framework are represented in Figs. 6.5b, 6.5c and

6.5d. On plotting the experimental modulus for the composite containing 3% gelatin and 0.1%

agarose in Fig. 6.5b, it intersects both agarose continuous (GʹU(Ag)) and gelatin continuous

165

(GʹU(gel)) topologies of the upper bound. Results in the preceding sections and specifically the

microscopy image presented in Fig. 6.4a, eliminate the possibility of an agarose continuous

topology at low agarose concentrations. On the other hand, a strong gelatin continuous phase

supporting agarose as a weak filler with gelatin claiming about 3/4th of the water (SGEL =0.76)

is plausible.

In contrast, when the agarose concentration in the composite was increased to 0.8%,

the experimental Gʹ value intersects just the gelatin continuous topology of the lower bound

(GʹL(gel)) of the computerized modelling at SGEL = 0.52 as illustrated in Fig. 6.5c. Results for the

composite containing 0.4% agarose were consistent to these (with SGEL = 0.69) and are hence,

not represented to avoid repetition. However, application of the blending laws to agarose-

gelatin composites containing greater than 0.8% agarose, yielded inconclusive results as shown

in Fig. 6.5d, which represents the composite containing 2.5% agarose. The experimental

modulus intersects the agarose and gelatin continuous topologies of both the upper and the

lower bound at their convergence. Evidence presented in Figs. 6.2, 6.3b and 6.4 suggests that

at higher agarose concentrations, agarose too forms a continuous network. Therefore, the

possibility of the system undergoing phase inversion at higher agarose concentrations and

forming an agarose continuous matrix with gelatin as a filler, cannot be entirely dismissed. In

addition, the arrangement of either an agarose or gelatin continuous phase, with agarose

trapping a disproportionately large amount of water in its phase (SGEL < 0.07) is physically

improbable as it would render the gelatin phase effectively dehydrated.

Results from the theoretical modelling along with the supporting evidence in the

preceding sections help conclude that in the composites containing 3% gelatin with 0.1, 0.4

and 0.8% agarose, the system is gelatin continuous with the dispersed agarose phase at 0.1%

concentration being weaker than the gelatin matrix, and stronger than it at 0.4 and 0.8%

166

concentrations. The isostrain and isostress blending law analysis however, fails to rationalise

the co-relation between the composite moduli and the constituents at higher concentrations of

agarose.

Since the microscopy images (Figs. 6.4c and 6.d) provide a direct visualisation of the

distribution of agarose and gelatin in the composite and suggest that agarose forms a continuous

network alongside that of gelatin, we pursued this molecular architecture next. This was done

by using the theoretical framework set forth by Davies that relates the composite modulus to

phase volumes of the individual networks and their respective moduli in a bicontinuous gel

1 5⁄

network (Davies, 1971):

1 5⁄ = ɸ𝑥𝐺ʹ𝑥

1 5⁄ + ɸ𝑦𝐺ʹ𝑦

(3) 𝐺ʹ𝑐

Fig. 6.5e summarises the results of this analysis for gelatin gels in the presence of the higher

concentrations of agarose studied in the present work (1.2-3%). Gratifyingly, values of

experimental composite moduli when plotted on the theoretically estimated bounds yielded

physically realistic results and the intersected predictions shift to lower volumes of water in the

gelatin phase on increasing the agarose content in the mixture. This once again confirms the

hypothesis that at agarose concentrations of 1.2% or more, the composite comprises a

bicontinuous network of agarose and gelatin.

Since agarose and gelatin undergo a segragative phase separation in the composite, the

solvent fraction in their respective phases, the final phase volumes along with the concentration

of each biopolymer in its phase and its corresponding mechanical contribution all depend on

the avidity of the two biopolymers for the aqueous medium. Thus, we utilized the ‘solvent-

avidity’ factor, p to allow for a meaningful comparison of water partition in the composites in

167

the given study.

168

169

Figure 6.5: (a) Calibration curves of Gʹ at 5°C as a function of agarose (▲) and gelatin (●)

concentrations. Modeling phase topology using – the isostrain (GʹU) and isostress (GʹL)

blending laws for 3% gelatin and (b) 0.1, (c) 0.8 and (d) 2.5% agarose, and blending law for

arranged upwards along with the experimental values of the same concentrations (dotted lines)

arranged upwards; and (f) predictions of the p factor for all the experimental concentrations

based on the isostrain (■), isotress (●) and bicontinuous (♦) blending laws with the inset

representing the estimated phase volume of the gelatin phase with increasing agarose

concentrations.

constituent X divided by the corresponding ratio of constituent Y, given by (Clark, Richardson,

Ross-Murphy, & Stubbs, 1983; A. J. F. s. Clark, Blanshard, & Lillford, 1987):

(4) 𝑝 = 𝑆𝑋 𝑐𝑋⁄ 𝑆𝑌 𝑐𝑌⁄

With X and Y corresponding to gelatin and agarose, respectively, and SX + SY = S, the total

amount of water (solvent) in the sample. Intersection data in all composites (Figs. 6.5b, 6.5c

and 6.5e) were replotted as a function of the p factor in Fig. 5f. It was found that as phase

behaviour of the composites varied between isostrain, isostress and biocontinuous conditions

at different agarose concentrations, according to Eqs. 1, 2 and 3, respectively, it resulted in

three distinct clusters of p values. As seen in Fig. 6.5f, there is an increase in the values of p as

the composites move from an isostrain to isotress to biocontinuous makeup, with gelatin

displaying a greater relative affinity for water with the changes in phase behaviour. The p

values within a specific family however (isostress or bicontinuous in Fig. 6.5f), exhibit a

systematic decrease, indicating a reduced proportion of water associated with the gelatin as the

170

concentration of agarose is increased. The asystematic variation of p with polymer

concentration as the system moves from a gelatin continuous to a bicontinuous makeup

suggests fundamental differences in the interplay between gelation kinetics and microscopic

phase separation at different architectural setup.

Finally, we estimated the phase volumes of the gelatin phase for the SGEL values of

intersection across all experimental composites (represented in the inset of Fig. 6.5f in percent)

using the framework established by Kasapis, Morris, Norton, & Clark (1993):

(5) Wx = x + (Sx x w)

(6) Wy = y +((1 – Sx) x w)

𝑊𝑥+ 𝑊𝑦

(7) ɸ𝑥 = 𝑊𝑥

Where, x, y, w, Sx, Wx and Wy denote the weight of polymer X, weight of polymer Y, weight of

total water in the system, solvent fraction in polymer-X phase, and the total weights of

constituent X and Y phases, respectively. These phase volume predictions will be compared

with those of the microscopic protocol to assess the efficacy of the latter in quantifying phase

volumes in phase separated systems.

6.3.3. 3D CLSM imaging and image analysis of agarose-gelatin co-gels

A Z-stack of all the experimental composites was captured and analysed using two image

analysis software – FIJI and Imaris, in parallel, according to the procedure outlined by Mhaske

et al. (2019). A typical 2D image captured from the surface of the composite containing 3%

gelatin with 0.4 and 2.5% agarose is depicted in Fig. 6.6a and Fig. 6.6d, respectively. As in

Fig. 6.4, FITC stained gelatin appears purple, whereas the unstained agarose appears black. In-

171

line with the discussion thus far, agarose phase remains as dispersed inclusions in a continuous

gelatin matrix at a low concentration of 0.4% in Fig. 6.6a, and forms a continuous matrix

alongside

Figure 6.6: (a, d) 2D, (b,e) Z-stack and (c, f) 3D CLSM images of 3% gelatin and 0.4 and 2.5%

172

agarose respectively. The gelatin phase is depicted in purple whereas agarose phase is in black.

gelatin when the concentration increases to 2.5% in the composite (Fig. 6.6d). The 2D images

are 8-bit images, 512 x 512 pixels in dimension corresponding to an area of 1279 x 1279 µm.

30 such images were captured along the Z axis at a fixed distance or Z-step of 10 µm, imaging

a ~300 µm deep section from the surface, as seen in Figs. 6.6b and 6.6e (depicting samples

containing 0.4 and 2.5% agarose, respectively). These Z-stacks (Figs. 6.6b and 6.6e) are then

used to render the 3D images seen in Figs. 6.6c and 6.6f, respectively. The Z-stacks are

analysed using an open source analysis software, FIJI, which quantifies area fraction in each

2D image and multiples it with the Z-step to give the volume fraction of a phase in the imaged

cross-section. In contrast, Imaris generates a 3D image using the Z-stack and then performs

volumetric quantification.

In the 8-bit images captured, pixels denote the intensity of the wavelength emitted by the

sample between the range of 0 and 255. An unstained sample appears black with a pixel

intensity of 0, serving as a negative control and baseline calibration in image analysis (Lu,

Chen, Wang, & Chang, 2016). The phase volume of the gelatin phase measured using the two

image processing software – Imaris and FIJI are illustrated in Fig. 6.7 along with the phase

volume estimates obtained from theoretical modelling for comparison. As seen in the figure,

the volume estimates obtained using Imaris are comparable with the theoretical predictions

indicating its suitability as an alternative for the theoretical blending laws. Phase volumes

estimated using FIJI however vary between 40 and 60% throughout the experimental range and

173

are statistically, significantly different from the theoretical outcomes.

Figure 6.7: Gelatin phase volume estimates plotted against increasing agarose concentrations

added to 3% gelatin determined by quantifying 3D CLSM images using image analysis

software – FIJI and Imaris, and the rheology based blending laws.

Given that FIJI has sixteen distinct methods of thresholding (separating the object

pixels from the background pixels), with each method employing a different algorithm, it is

possible that the method used for thresholding in the present work isn’t the most appropriate

for the system investigated. During the initial exploration of the applicability of 3D imaging

paired with image analysis in estimating phase behaviour in biopolymer composites, Mhaske

et al. (2019) found that the ‘Default’ thresholding method that utilized the iterative algorithm

yielded values closest to the theoretical predictions. This iterative algorithm by Ridler &

Calvard (1978), has also been successfully used to estimate the phase volumes in systems

comprising agarose-lipid (Mhaske, Condict, Dokouhaki, Farahnaky, & Kasapis, 2020) and

agarose-microcrystalline cellulose (Mhaske et al., 2020) previously. It is probable that while

the algorithm could accurately quantify components with a distinct phase interface, it is not

best suited for a system comprising constituents that are co-soluble. Further exploration would

174

perhaps help find the suitable algorithm in FIJI. However, as the aim of this study was to

evaluate the efficacy of the microscopic protocols developed previously in conjunction with

image analysis software in estimating phase behaviour, this wasn’t pursued. A nonadaptive

thresholding, as is the case with FIJI, would also necessitate that theoretical modeling is carried

out to verify the suitability of the thresholding method for every system, negating the purpose

of its development as an alternative to theoretical modeling.

6.3.4. Conclusions

Exploiting the advances made in imaging technology, recent attempts have been made at

developing a microscopic protocol paired with image analysis for quantifying phase volumes

in simple biopolymer composites as an alternative to the theoretical blending laws. Successful

implementation of the quick, direct microscopic protocol could potentially render the tedious

experimental work and modelling expertise that the theoretical blending laws entail,

dispensable. Therefore, the suitability of 3D CLSM imaging, paired with image analysis using

two software – Imaris and FIJI in parallel, in estimating the phase volume in a gelling

biopolymer composite of agarose and gelatin was investigated. Imaris, estimated the phase

volume of the continuous gelatin phase in close comparison with the rheological modelling as

the composite phase topology shifted from the upper bound and lower bound to a bicontinuous

network with increasing agarose concentration. Whereas, the ‘Default’ thresholding method of

FIJI which had been successfully used in the past for simpler composites, failed to cope up and

yielded inconsistent results.

The microscopic protocol thus lays the groundwork for future research, where it can be

used to determine phase behaviour accurately and rapidly in complex binary and tertiary gelling

175

biomaterial composites.

References

Anvari, M., Pan, C. H., Yoon, W. B., & Chung, D. (2015). Characterization of fish gelatin–

gum arabic complex coacervates as influenced by phase separation

temperature. International Journal of Biological Macromolecules, 79, 894-902.

Awadhiya, A., Tyeb, S., Rathore, K., & Verma, V. (2017). Agarose bioplastic‐based drug

delivery system for surgical and wound dressings. Engineering in Life Sciences, 17(2),

204-214.

Çakır, E., & Foegeding, E. A. (2011). Combining protein micro-phase separation and protein–

polysaccharide segregative phase separation to produce gel structures. Food

Hydrocolloids, 25(6), 1538-1546.

Cebi, N., Durak, M. Z., Toker, O. S., Sagdic, O., & Arici, M. (2016). An evaluation of Fourier

transforms infrared spectroscopy method for the classification and discrimination of

bovine, porcine and fish gelatins. Food Chemistry, 190, 1109-1115.

Clark, A. H., Richardson, R. K., Ross-Murphy, S. B., & Stubbs, J. M. (1983). Structural and

mechanical properties of agar/gelatin co-gels. Small-deformation

studies. Macromolecules, 16(8), 1367-1374.

Clark, A. H. (1987). Application of network theory to food systems. Food structure and

behaviour;edited by JMV Blanshard and P. Lillford. 13-34.

Davies, W. E. A. (1971). The theory of elastic composite materials. Journal of Physics D:

Applied Physics, 4(9), 1325.

Fernández, E., López, D., Mijangos, C., Duskova‐Smrckova, M., Ilavsky, M., & Dusek, K.

(2008). Rheological and thermal properties of agarose aqueous solutions and

176

hydrogels. Journal of Polymer Science Part B: Polymer Physics, 46(3), 322-328.

Geonzon, L. C., Bacabac, R. G., & Matsukawa, S. (2019). Microscopic characterization of

phase separation in mixed carrageenan gels using particle tracking. Journal of The

Electrochemical Society, 166(9), B3228.

Grinberg, V. Y., & Tolstoguzov, V. B. (1997). Thermodynamic incompatibility of proteins and

polysaccharides in solutions. Food Hydrocolloids, 11(2), 145-158.

Kasapis, S. (2008). Phase separation in biopolymer gels: A low to high-solid exploration of

structural morphology and functionality. Critical Reviews in Food Science and

Nutrition, 48, 341-359.

Kasapis, S., Morris, E. R., Norton, I. T., & Clark, A. H. (1993). Phase equilibria and gelation

in gelatin/maltodextrin systems—Part IV: Composition-dependence of mixed-gel

moduli. Carbohydrate Polymers, 21(4), 269-276.

Lan, Y., Ohm, J. B., Chen, B., & Rao, J. (2020). Phase behavior, thermodynamic and

microstructure of concentrated pea protein isolate-pectin mixture: Effect of pH,

biopolymer ratio and pectin charge density. Food Hydrocolloids, 101, 105556.

Lu, C. T., Chen, Y. Y., Wang, L. L., & Chang, C. F. (2016). Removal of salt-and-pepper noise

in corrupted image using three-values-weighted approach with variable-size

window. Pattern Recognition Letters, 80, 188-199

Mhaske, P., Condict, L., Dokouhaki, M., Farahnaky, A., & Kasapis, S. (2020). Phase volume

quantification of agarose-ghee gels using 3D confocal laser scanning microscopy and

blending law analysis: A comparison. LWT, 109567.

Mhaske, P., Condict, L., Dokouhaki, M., Katopo, L., & Kasapis, S. (2019). Quantitative

analysis of the phase volume of agarose-canola oil gels in comparison to blending law

predictions using 3D imaging based on confocal laser scanning microscopy. Food

177

Research International, 125, 108529.

Mhaske, P., Farahnaky, A., & Kasapis, S. (2020). 3D Confocal Laser Scanning Microscopy for

Quantification of the Phase Behaviour in Agarose-MCC co-gels in Comparison to the

Rheological Blending-law Analysis. Food Biophysics, 25, 1-8.

Mirzapour‐Kouhdasht, A., Sabzipour, F., Taghizadeh, M. S., & Moosavi‐Nasab, M. (2019).

Physicochemical, rheological, and molecular characterization of colloidal gelatin

produced from Common carp by‐products using microwave and ultrasound‐assisted

extraction. Journal of texture studies, 50(5), 416-425.

Morris, E. R. (2009). Functional interactions in gelling biopolymer mixtures. In Modern

biopolymer science (pp. 167-198): Elsevier.

Nielsen, L. E. (1975). Morphology and the elastic modulus of block polymers and polyblends.

In Rheological Theories· (pp. 594-600): Springer.

Norton, I. T., & Frith, W. J. (2001). Microstructure design in mixed biopolymer

composites. Food Hydrocolloids, 15(4-6), 543-553.

Parker, N. G., & Povey, M. J. W. (2012). Ultrasonic study of the gelation of gelatin: Phase

diagram, hysteresis and kinetics. Food Hydrocolloids, 26(1), 99-107.

Qi, X., Su, T., Tong, X., Xiong, W., Zeng, Q., Qian, Y., ... & He, X. (2019). Facile formation

of salecan/agarose hydrogels with tunable structural properties for cell

culture. Carbohydrate polymers, 224, 115208.

Ridler, T. W., & Calvard, S. (1978). Picture thresholding using an iterative selection

method. IEEE trans syst Man Cybern, 8(8), 630-632.

Shrinivas, P., Kasapis, S., & Tongdang, T. (2009). Morphology and mechanical properties of

bicontinuous gels of agarose and gelatin and the effect of added lipid

phase. Langmuir, 25(15), 8763-8773.

Sikorski, Z. E., Pokorny, J., & Damodaran, S. (2008). 14Physical and Chemical Interactions of

178

Components in Food Systems. Fennema’s food chemistry, 849-884.

Takayanagi, M. (1963). Viscoelastic behavior of polymer blends and its comparison with

model experiments. Journal of the Society of Material Science Japan, 12, 389-394.

Wang, J., Dumas, E., & Gharsallaoui, A. (2019). Low Methoxyl pectin/sodium caseinate

complexing behavior studied by isothermal titration calorimetry. Food

Hydrocolloids, 88, 163-169.

Waterborg, J. H. (2009). The Lowry method for protein quantitation. In The protein protocols

handbook (pp. 7-10): Springer.

Wurm, F., Pham, T., & Bechtold, T. (2019). Modelling of phase separation of alginate-

carrageenan gels based on rheology. Food Hydrocolloids, 89, 765-772.

Zhang, T., Xu, X., Ji, L., Li, Z., Wang, Y., Xue, Y., & Xue, C. (2017). Phase behaviors involved

in surimi gel system: Effects of phase separation on gelation of myofibrillar protein and

179

kappa-carrageenan. Food Research International, 100, 361-368.

Chapter 7 Conclusions and Future Research

180

7.1 Conceptual basis of the present study

Mixed biopolymer systems are increasingly used in biomedical, pharmaceutical and

food industries due to their functional and structure forming properties. While broadening the

spectrum of textural variations, they offer a better control over the nutritional value of the

product, processibility, kinetics of active compound release and shelf stability of the final

product (Prameela, Mohan, & Ramakrishna, 2018). With the growing appreciation of

biopolymers in altering techno-functionality in a wide range of products, a thorough

understanding of the structure-function relationship of biopolymer composites is crucial for the

fabrication of materials with desired functional properties (Sikorski, Pokorny & Damodaran,

2008).

This has been successfully obtained via theoretical modelling using rheology based

blending laws which assume that constituents of a biopolymer composite undergo extensive

segragative phase separation. Characterisation of solvent partition between the constituents is

done with respect to the relative affinity for water of each component. Theories borrowed from

the realm of polymer science are utilized to follow the mechanical properties of the mixtures

as outlined by Takayangi, Lewis and Nielsen (Nielsen, 1974; Takayanagi et al., 1963):

(𝐺𝑐)𝑛 = ɸ𝑋(𝐺𝑋)𝑛 + ɸ𝑌(𝐺𝑌)𝑛 (1)

(2) Gʹc/ GʹX = (1 + ABɸY)/(1-BψɸY)

2]ɸY

(3) Ψ = 1 + [(1 - ɸm)/ɸm

(4) B = [(GʹY/GʹX) – 1]/[(GʹY/GʹX) + A]

Where n equals to 1, -1 and 0.2 for isostrain, isostress and bicontinuous conditions,

respectively. GʹC, GʹX and GʹY are the storage modulus of the composite, continuous and

dispersed phases, respectively while ɸY is the phase volume of the filler. A is the Einstein

coefficient which corresponds to the shape of the filler, and ɸm is the maximum packing factor

of the filler.

Though the blending laws are robust, they are an indirect method of estimating phase

181

volume and require considerable experimentation and modeling expertise. On the other hand,

confocal laser scanning microscopy (CLSM) which is widely used in the biosciences and

biomedical sciences for qualitative analysis shows potential in being utilised for direct

quantitative estimation of phase volume in biomaterial composites when paired with the

accelerating advancement in image processing. CLSM allows the sample to be analysed in its

native state without needing pre-processing steps that other microscopic techniques entail like

fixation, sectioning and/or freeze drying. Multiple two-dimensional (2D) images can be

captured along the Z-axis by moving the focal plane in steps of specified thickness through the

sample depth. This Z-stack feature of the confocal offers unique insight into a gel’s true 3D

microstructure (Auty, 2013).

Numerous studies have demonstrated the qualitative application of CLSM in phase-

separated biopolymer mixtures including wheat protein/starch (Peighambardoust, Dadpour &

Dokouhaki, 2010), gelatin/starch/olive oil (Firoozmand & Rousseau, 2014), and whey protein

isolate/locust bean gum (Ur-Rehman et al., 2016). Examples of application of Z-stack CLSM

imaging, paired with image analysis software (e.g. ImageJ, FIJI or Image-Pro) for quantitative

analysis do exist but are limited for example, measuring size of starch granules (Nagano,

Tamaki & Funami, 2008), analysis of water droplet size distribution in spreads (Van Dalen,

2002), and protein’s fractal dimension and area in wheat dough (Li, Liu, Wu & Zhang, 2017).

7.2. Conclusions

To address the gaps outlined above, this PhD study began by developing a novel

microscopic method using Z-stack CLSM in combination with image processing software for

application in soft biomaterials. Work started off with a simple agarose-canola oil system,

where agarose upon cooling formed a continuous network within which the discontinuous

fillers of canola oil remained dispersed exhibiting distinct phase separation. As there was no

182

water partitioning between the two components, water remained in the agarose phase in its

entirety and the volume of the lipid phase depended solely on its concentration added to the

sample. Using the lipid phase volume estimates obtained via rheology based theoretical

modelling as reference, different algorithms in image analysis, using two software, FIJI and

Imaris in parallel were tested for suitability by comparing the accuracy of the phase volume

estimates obtained. Gratifyingly, the microscopic protocol developed employing the iterative

algorithm for thresholding (separating the object pixels from the background pixels) (Ridler &

Calvard, 1978), successfully determined the phase volume of the oil droplets across the

experimental range in congruence with the isostrain blending laws. The successful

implementation of the microscopic protocol merited further consideration of the protocol’s

application to more complex mixtures.

In addition, this was the first time that the rheological blending laws were applied to an

aqueous/lipid mixture, as opposed to their popular application in aqueous composites. Using

known quantities of lipid allowed for the validation of both the rheology based theoretical

modeling and the microscopic procedures. The successful application of the blending laws

suggested that they could potentially be used to analyse samples containing lipids, thus

broadening the range of biomaterials whose phase behaviour could be probed.

Next, a binary system consisting of agarose and a hard lipid, ghee that showed distinct

phase separation was probed using the developed microscopic protocol in order to quantify the

phase volume of the discontinuous filler. Ghee, being a lipid, primarily helped verify the

applicability of the blending laws to non-aqueous composites. Exhibiting a greater storage

modulus than the agarose network, it changed the composite topology from isostrain to

isostress, verifying the applicability of the isostress blending law in probing non aqueous

mixtures, while validating the novel microscopic protocol. Values of the estimated phase

volumes compared well with the theoretical estimates making the microscopic procedure a

183

potential alternative to rheological modeling in soft biopolymer hydrogels. However, at higher

lipid concentrations, perceivable inner-filtering was observed during imaging due to diffraction

of the illuminating laser as it passed through constituents with different densities and refractive

indices (Pioch, Sorg, Stadler, Hagge, & Dörfer, 2004). This in turn impacted the obtained phase

volume estimates and highlighted the need for further research in order to achieve wide-spread

applicability of the microscopic procedure.

Thus far, the efficacy of the microscopic protocol had just been tested in estimating

phase volume in a lipid-polymer system where the lipid didn’t absorb any water/solvent and

there was no water partitioning between the two components. Therefore, in the third

experimental chapter, attempts were made at estimating the water absorbed by, and the

resultant phase volume of microcrystalline cellulose (MCC) suspended in an agarose network.

The MCC particles, being hygroscopic, were expected to absorb water from the system but not

disintegrate (Nsor-Atindana et al., 2017). The microscopic protocol rendered phase volume

estimates of hydrated cellulose crystals in close comparison with those obtained from

rheological modelling and could cope with the increasing density of the MCC fillers. This

proved that the developed protocol can be used to accurately estimate not just the phase volume

of a component in a composite but also the amount of water it absorbs from the system.

With such encouraging results, work was expanded to a mixed biopolymer system

comprising two gelling biopolymers – agarose and gelatin. Upon cooling, both biopolymers

were expected to form a gel network entrapping a fraction of water within its domain. The two

image processing software were used to estimate the phase volume of the gelatin phase with

increasing agarose concentration in the composite. One of the image processing software,

Imaris, estimated the phase volume of the continuous gelatin phase in close comparison with

the rheological modelling as the composite phase topology shifted from the isostrain to

184

isostress to a bicontinuous network with increasing agarose concentration. Whereas the

‘Default’ thresholding method of FIJI which had been successfully used in the past for simpler

composites, failed to cope and yielded inconsistent results.

7.3. Future research

The microscopic protocol thus lays the groundwork for future research, where it can be

used to determine phase behaviour accurately and rapidly in complex, industrially relevant

gelling biomaterial composites. To further advance research in this area, a crucial step would

be overcoming or minimising the effect of self-shadowing or inner filtering. Within the bounds

of current research, the effect of self-shadowing has been successfully minimized by a number

of measures including - increasing the illumination intensity, using an objective with a higher

numerical aperture (NA), using the microscope in the transmission mode, etc (Sandison,

Williams, Wells, Strickler, & Webb, 1995). However, it cannot be completely overcome. Non-

diffractive bessel beams and airy beams could probably be used for illumination instead of

lasers to reduce shadowing (Di Domenico, G., Ruocco, G., Colosi, C. et al., 2018).

It would be interesting to test the efficacy of the developed protocol in quantifying

phase volume of composites in the nanoscale. CLSM has been successfully used to image nano

particles in the field of biomedical sciences, with its application limited to qualitative analysis

(Zhang, L., & Monteiro-Riviere, N., 2012; Lee et al., 2014). As nanoparticles are increasingly

used in food, packaging and pharmaceutical industries to alter microstructure, rheology and

functional properties at the nanoscale (Ameta et al., 2020), the successful application of this

technique could help researchers probe the structure-function relationship of nanoparticles with

ease.

Another interesting endeavour would be to modify the developed protocol to include

185

the additional dimension of time. With the recent advancements in imaging technology, there

have been successful applications of continuous CLSM imaging of live cells to follow progress

of biological events over time (Grossi et al., 2016). If this technique is adapted in the realm of

polymer science, it would be possible to visualise and quantify the dynamic changes in phase

behaviour in composites during gelation and/or storage.

Finally, though the technique was developed as an alternative to the blending laws, it

has the potential to surpass them in terms of application. Exploiting the capability of CLSM to

excite multiple fluorophores simultaneously, it is possible to accurately differentiate between

multiple constituents of a composite, provided they are adequately labelled (Dürrenberger et

al., 2001). This implies that the microscopic protocol can be used to capture and further

quantify images of complex tertiary or quaternary composites, whilst the application of

blending laws mostly remains restricted to binary composites.

An example of this is illustrated in Figures 7.1 and 7.2. Figure 7.1a is a 2D image of

Sample 1, containing 3% (w/w) agarose plus 2% (w/w) gelatin and 2% (w/w) canola oil. The

stained agarose and lipid phases appear green and red, respectively and the unstained gelatin is

denoted in black. The agarose phase forms the continuous phase within which the gelatin and

lipid phase remains dispersed. Figure 7.2a which is a 2D image of another tertiary composite,

Sample 2, comprising 10% (w/w) gelatin plus 5% (w/w) WPI and 5% (w/w) canola oil denotes

the stained gelatin and oil phases in red and green respectively while the whey protein isolate

appears black. Here, the discontinuous WPI and oil fillers are dispersed in the continuous

gelatin matrix. For both samples, multiple 2D images were captured along the Z-axis to obtain

a Z-stack as shown in Figures 7.1b and 7.2b, respectively. These Z-stacks are then used to

generate 3D images using the image processing software, Imaris, as shown in Figures 7.1c and

186

7.2c, respectively.

Figure 7.1: (a) 2D, (b) Z-stack, and (c) 3D CLSM images of 3% (w/w) agarose plus 2% (w/w)

gelatin and 2% (w/w) canola oil. The agarose, gelatin and oil phase is denoted in green, black

and red respectively. The composite 3D CLSM image is further split to denote 3D images of

the constituent (d) agarose and gelatin, and (e) oil phase separately (Images captured at 10x,

scale bar denotes 200μm).

Image analysis allows the user to split the 3D images based on the wavelength used for

illumination, making it possible to view and capture images of the distribution of a specific

component in a composite. Figure 7.1c is split into Figures 7.1d and 7.1e which are 3D images

187

of the agarose and gelatin, and oil phase, respectively. Figure 7.2c is split into Figures 7.2d and

7.2e which denote the 3D images of the distribution of gelatin and WPI, and the oil phase

within the composite.

(a) 2D, (b) Z-stack, and (c) 3D CLSM images of 10% (w/w) gelatin plus 5% (w/w) WPI and

5% (w/w) canola oil. The gelatin, WPI and oil phase is denoted in red, black and green

respectively. The composite 3D CLSM image is further split to denote 3D images of the

constituent (d) gelatin and WPI, and (e) oil phase separately. (Images captured at 10x, scale

188

bar denotes 200μm).

These images can then be quantified using the developed microscopic protocol using Imaris as

described in previous chapters. Phase volume estimates obtained by processing Figures 7.1d,

7.1e, 7.2d and 7.2e are summarised in Table 7.1.

Table 7.1: Experimental constituent concentrations in the composites and estimated phase

volumes with Imaris image analysis

Phase volume estimated using Imaris (%) Concentration added in the composite (%, w/w) Sample 1

3 2 2 74.5 1.9 23.6 Agarose Oil Gelatin Sample 2

10 5 5 60.6 5.1 34.3 Gelatin Oil WPI

The phase volume of canola oil in the composite depends entirely on its concentration

added to the composite, as there is no water partitioning between the aqueous and the lipid

phases. Table 7.1 shows close agreement in the oil phase volume estimates generated by Imaris

and the actual concentration of canola oil added to both the samples. Though this gains

confidence regarding the accuracy and credibility of the results obtained, further

189

experimentation using complementary techniques is needed for confirmation.

References

Ameta S.K., Rai A.K., Hiran D., Ameta R., Ameta S.C. (2020) Use of Nanomaterials in Food

Science. In: Ghorbanpour M., Bhargava P., Varma A., Choudhary D. (eds) Biogenic

Nano-Particles and their Use in Agro-ecosystems. Springer, Singapore.

https://doi.org/10.1007/978-981-15-2985-6_24

Athimethphat, M. (2011). A review on global binarization algorithms for degraded document

images. AU JT, 14(3), 188-195.

Auty, M. A. E. (2013). 4 - Confocal microscopy: principles and applications to food

microstructures. In V. J. Morris & K. Groves (Eds.), Food Microstructures (pp. 96-

P98): Woodhead Publishing.

Calvard, T. & Ridler, S. (1978). Picture Thresholding Using an Iterative Selection Method.

IEEE Transactions on Systems, Man, and Cybernetics, 8, 630-632

Di Domenico, G., Ruocco, G., Colosi, C. et al. (2018).Cancellation of Bessel beam side lobes

for high-contrast light sheet microscopy. Sci Rep 8, 17178

Dürrenberger, M. B., Handschin, S., Conde-Petit, B., & Escher, F. (2001). Visualization of

food structure by confocal laser scanning microscopy (CLSM). LWT-Food Science and

Technology, 34(1), 11-17.

Firoozmand, H., & Rousseau, D. (2014). Tailoring the morphology and rheology of phase-

separated biopolymer gels using microbial cells as structure modifiers. Food

Hydrocolloids, 42, 204-214.

Grossi, M., Morgunova, M., Cheung, S., Scholz, D., Conroy, E., Terrile, M., ... & O’Shea, D.

F. (2016). Lysosome triggered near-infrared fluorescence imaging of cellular

trafficking processes in real time. Nature communications, 7(1), 1-13.

Lee, P. L., Chen, B. C., Gollavelli, G., Shen, S. Y., Yin, Y. S., Lei, S. L., ... & Ling, Y. C.

190

(2014). Development and validation of TOF-SIMS and CLSM imaging method for

cytotoxicity study of ZnO nanoparticles in HaCaT cells. Journal of hazardous

materials, 277, 3-12.

Li, Q., Liu, R., Wu, T. & Zhang, M. (2017). Interactions between soluble dietary fibers and

wheat gluten in dough studied by confocal laser scanning microscopy. Food Research

International, 95, 19-27.

Nagano, T., Tamaki, E., & Funami, T. (2008). Influence of guar gum on granule morphologies

and rheological properties of maize starch. Carbohydrate Polymers, 72, 95–101

Nielsen, L. E. (1974). Morphology and the the elastic modulus of block polymers and

polyblends. Rheologica Acta, 13, 86-92.

Nsor-Atindana, J., Chen, M., Goff, H. D., Zhong, F., Sharif, H. R., & Li, Y. (2017).

Functionality and nutritional aspects of microcrystalline cellulose in

food. Carbohydrate Polymers, 172, 159-174.

Peighambardoust, S., Dadpour, M., & Dokouhaki, M. (2010). Application of epifluorescence

light microscopy (EFLM) to study the microstructure of wheat dough: a comparison

with confocal scanning laser microscopy (CSLM) technique. Journal of Cereal

Science, 51, 21-27.

Pioch, T., Sorg, T., Stadler, R., Hagge, M., & Dörfer, C. E. (2004). Resin penetration through

submicrometer hiatus structures: a SEM and CLSM study. Journal of Biomedical

Materials Research Part B: Applied Biomaterials: An Official Journal of The Society

for Biomaterials, The Japanese Society for Biomaterials, and The Australian Society

for Biomaterials and the Korean Society for Biomaterials, 71(2), 238-243.

Prameela, K., Mohan, C. M., & Ramakrishna, C. (2018). Biopolymers for food design:

Consumer-friendly natural ingredients. In Biopolymers for food design (pp. 1-32):

191

Elsevier.

Sikorski, Z.E., Pokorny, J. & Damodaran, S. (2008). Physical and chemical interactions of

components in food system. In S. Damodaran, K.L. Parkin and O.R. Fennema (Eds),

Fennema’s Food Chemistry (4th ed). Boca Raton: CRC Press.

Takayanagi, M., Harima, H., & Iwata, Y. (1963). Viscoelastic behavior of polymer blends and

its comparison with model experiments. Memoirs-Faculty of Engineering Kyushu

University, 23, 1-13.

Ur‐Rehman, A., Khan, N. M., Ali, F., Khan, H., Khan, Z. U., Jan, A. K., Ahmad, S. (2016).

Kinetics Study of Biopolymers Mixture with the Help of Confocal Laser Scanning

Microscopy. Journal of food process engineering, 39, 533-541.

Van Dalen, G. (2002). Determination of the water droplet size distribution of fat spreads using

confocal scanning laser microscopy. Journal of Microscopy, 208, 116-133.

Zhang, L. W., & Monteiro-Riviere, N. A. (2012). Use of confocal microscopy for nanoparticle

192

drug delivery through skin. Journal of biomedical optics, 18(6), 061214.

APPENDIX

PUBLICATIONS AND PRESENTATIONS

I. Journal articles:

Mhaske, P., Condict, L., Dokouhaki, M., Katopo, L., & Kasapis, S. (2019). Quantitative

analysis of the phase volume of agarose-canola oil gels in comparison to blending law

predictions using 3D imaging based on confocal laser scanning microscopy. Food

Research International, 125, 108529.

(This published paper is presented as Chapter 3 in this Thesis)

Mhaske, P., Condict, L., Dokouhaki, M., Farahnaky, A., & Kasapis, S. (2020). Phase volume

quantification of agarose-ghee gels using 3D confocal laser scanning microscopy and

blending law analysis: A comparison. LWT, 109567.

(This published paper is presented as Chapter 4 in this Thesis)

Mhaske, P., Farahnaky, A., & Kasapis, S. (2020). 3D Confocal Laser Scanning Microscopy

for Quantification of the Phase Behaviour in Agarose-MCC co-gels in Comparison to

the Rheological Blending-law Analysis. Food Biophysics, 25, 1-8.

(This published paper is presented as Chapter 5 in this Thesis)

II. Poster presentation

Mhaske, P., Condict, L., Dokouhaki, M., Katopo, L. and Kasapis, S.: 3D imaging using

confocal laser scanning microscopy and image analysis to quantify phase volume in

composite mixtures – a novel, direct approach. In 15th International Hydrocolloids

conference (2-5th March), Melbourne, Australia.

III. Oral presentation

Mhaske, P., Condict, L., Dokouhaki, M., Farahnaky, A. and Kasapis, S.: Quantitative phase

volume estimation of agarose-ghee gels using 3D confocal laser scanning microscopy

and blending law analysis: A comparison. In 15th International Hydrocolloids

193

conference (2-5th March), Melbourne, Australia.

A thesis submitted in fulfillment of the requirements for the degree of Doctor of Philosophy: Use of Z-stack imaging to quantify the phase behaviour of biomaterial composites in relation to theoretical predictions of blending laws from rheological measurements

Use of Z-stack imaging to quantify the phase behaviour of biomaterial composites in relation to theoretical predictions of blending laws from rheological measurements

Pranita Mhaske

B. Tech. (Food Science and Technology), MIT College of Food Technology, Pune, India

Declaration

Copyright

Pranita Mhaske

ACKNOWLEDGEMENTS

TABLE OF CONTENTS

List of Figures

Chapter 1

Chapter 6

Chapter 7

List of Tables

Chapter 1 Table 1.1

List of abbreviations

List of symbols and units

Explanatory notes

Summary

Chapter 1

Background and Literature Review

Table 1.1: Summary of analytical techniques used to measure biopolymer concentration in separated phases in an aqueous mixture.

Technique

Principle

Advantages

Disadvantages

References

Application (examples)  Locust bean

Phase volume ratio

Refractometry Refractive index of a

Phenol sulphuric method

Bradford Assay

Biuret method

Kjeldahl method

Table 1.2. Spectroscopic techniques used to probe phase behaviour.

Principle

Advantages

Technique Ultrasonic  Measures the intermolecular

Disadvantages  Not suitable for dilute

References (Chanamai & McClements, 2001, Chappellaz, Alexander, & Corredig, 2010; Corredig, Sharafbafi, & Kristo, 2011; Clark, 1996)

Raman

FTIR

Advantages  Fast, non-destructive and

Disadvantages  Needs additional

Table 1.2. Spectroscopic techniques used to probe phase behaviour (continued). Technique Principle Diffusing Wave

References (Weinbreck, 2004; Blijdenstein, 2003; Gancz, 2005; Liu, 2007)

UV-Vis

Nuclear Magnetic Resonance

Electron spin resonance

Fig.1.5. A drop of caseinate in alginate rich phase at rest (1), start-up transient of deformation under

Table 1.3: Microscopic techniques used for structural characterisation of phase behaviour.

Principle

Advantages

Disadvantages

Type of Microscope

Radiation type

Optical

Electron

Phase contrast

Confocal

Fluorescence

Atomic force

Chapter 2 - Materials and Methods

Abstract

Chapter 3

Quantitative analysis of the phase volume of agarose- canola oil gels in comparison to blending law predictions using 3D imaging based on confocal laser scanning microscopy

Chapter 4

Phase volume quantification of agarose-ghee gels using

3D confocal laser scanning microscopy and blending law

analysis: A comparison

Chapter 5

3D Confocal laser scanning microscopy for quantification

of the phase behaviour in agarose-MCC co-gels in

comparison to the rheological blending-law analysis

3D Confocal Laser Scanning Microscopy for Quantification of the Phase Behaviour in Agarose-MCC co-gels in Comparison to the Rheological Blending-law Analysis

Chapter 6

Comparison of rheological blending-law analysis and 3D

confocal laser scanning microscopy for quantification of the

phase behaviour in agarose-gelatin co-gels

Chapter 7