gel development using cellulose nanocrystals

University of Calgary

PRISM: University of Calgary's Digital Repository

Graduate Studies The Vault: Electronic Theses and Dissertations

2020-06-18 0:00

Gel Development Using Cellulose Nanocrystals

Abbasi Moud, Aref

http://hdl.handle.net/1880/112911

doctoral thesis

University of Calgary graduate students retain copyright ownership and moral rights for their

thesis. You may use this material in any way that is permitted by the Copyright Act or through

licensing that has been assigned to the document. For uses that are not allowable under

copyright legislation or licensing, you are required to seek permission.

Downloaded from PRISM: https://prism.ucalgary.ca

UNIVERSITY OF CALGARY

Gel Development Using Cellulose Nanocrystals

by

Aref Abbasi Moud

A THESIS

SUBMITTED TO THE FACULTY OF GRADUATE STUDIES

IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE

DEGREE OF DOCTOR OF PHILOSOPHY

GRADUATE PROGRAM IN CHEMICAL AND PETROLEUM ENGINEERING

CALGARY, ALBERTA

JUNE, 2020

© Aref Abbasi Moud 2020

I

Abstract

Cellulose Nano Crystal (CNC) are naturally derived nanoparticles with a slender shape that own a

remarkably high aspect ratio. Due to its mechanical properties with Young modulus in the range of

120-160 GPa and proven biocompatibility, CNC is an attractive nanoparticle for many applications.

In the acid-based method of CNC production, the particles develop naturally negative charges,

inducing an electrostatic repulsion between CNC particles and, consequently, when suspended in the

water, a stable suspension. Upon introducing a coagulant, such as NaCl above a threshold value in

the CNC suspension, phase separation happens where the system evolves toward gelation. The

individual CNC particles, through aggregation, contribute to building a 3D gel fractal structure.

Where a porous self-similar 3D structure spans in space, the design and synthesis of CNC-based gels

are tunable and flexible. Mechanical properties of the hydrogels can also be tuned when CNC is

coupled with polymers such as Polyvinyl alcohol (PVA).

Herein, the CNC-based gelation process is monitored, and the formed gels are characterized. The

zeta potential and Dynamic Light Scattering are employed to measure the hydrodynamic radii and

the surface charges of particles in different CNC-coagulant loadings. The gel morphology and CNC

cluster fractal dimensions are recorded using Scanning Election Microscopy (SEM) and

Transmission Electron Microscopy (TEM). The CNC-based gel behavior under large amplitude

strains is also characterized by non-linear rheology. The gel collapse behavior and the self-healing

dynamics of CNC-based gels are also quantified using fluorescence recovery after photobleaching

(FRAP) analysis.

It is shown that CNCs can coagulate upon increasing the ionic strength of the medium, where the

mechanical stability of the CNC-based gels (i.e., storage modulus), is an increasing function of NaCl

and CNC concentration. Non-linearity of the gel was shown to be more influenced by NaCl

concentration. The addition of PVA makes the CNC hydrogels mechanically robust, where two

jumps in values of storage modulus as a function of frequency is observed. The jumps are attributed

to the network formation between CNCs and CNC-polymer. Finally, the FRAP analysis using

Confocal Laser Scanning Microscopy reveals that the CNC mobility in gel media is influenced by

both CNC and NaCl concentration. The result of this study can be used in controlling CNC hydrogels

properties, such as the gel self-healing and mechanical properties, and in assembling a 3-D hydrogel

structure with CNC. In practice, the developed CNC-based hydrogels can be used in tissue

engineering.

II

To:

My parents, friends, and Siblings

for their unconditional supports

III

They did not know it was impossible, so

they did it!

Mark Twain

IV

Acknowledgments

I would like to express my gratitude to Dr. Hossein Hejazi and Dr. Sanatinezhad for their supervision,

direction, and guidance from early days of PhD journey. In various ways, they provided me with

encouragement and support in my academic endeavours. Their scientific support made a great source

of ideas and passions in pursuing science, which remarkably helped me to grow as a student and a

researcher. I am grateful to them and their support more they can imagine.

In a nutshell, I treasure my supervisors; whose expertise, knowledge and support made my Ph.D.

academic career an enjoyable and productive journey.

I would like to extend my appreciation to my supervisory committee members and external

examiners, Dr. Mahinpey, Dr. Wadood, Dr. Federico, and Dr. Wong. I would also like to express my

gratitude to my group members at the University of Calgary, whose contribution made the whole

process of pursuit of science much easier. I would also like to thank Dr. Sundararaj and Milad

Kamkar for their help with the acquisition of rheological data of this thesis.

The financial support from the Natural Science and Engineering Research Council (NSERC) of

Canada and Mitacs accelerate is highly appreciated. The deepest gratitude goes to my family because

of their unconditional support and encouragement.

V

Table of Contents

Table of Contents ............................................................................................................................... V

Table of Figures ................................................................................................................................ XI

Table of Tables .............................................................................................................................. XIX

CHAPTER 1: Introduction ........................................................................................................... 1

1-1 Cellulose .............................................................................................................................. 2

1-1-1 Cellulose at the nanoscale ............................................................................................. 3

1-1-2 Cellulose nanocrystals .................................................................................................. 4

1-2 CNC in colloidal suspensions and gels ................................................................................ 5

1-2-1 Parameters affecting aggregation and gelation of CNCs .............................................. 7

1-2-2 Gelation mechanisms of CNC ...................................................................................... 9

1-2-3 CNC-composite hydrogels: review ............................................................................ 11

1-3 Methodology ...................................................................................................................... 19

1-3-1 CNCs confocal laser scanning microscopy ................................................................ 19

1-3-2 Gel healing monitoring using rheometry and CLSM ................................................. 22

1-3-3 Fiber orientation models ............................................................................................. 24

1-3-4 Monitoring the mechanical behavior of gels using rheometry ................................... 27

1-3-5 Large amplitude oscillatory shear test (LAOS) .......................................................... 32

1-4 Problem statement .............................................................................................................. 35

1-5 Dissertation outline ............................................................................................................ 36

CHAPTER 2: Colloidal Behavior of Cellulose Nanocrystals in the Presence of Sodium Chloride

39

2-1 Introduction ............................................................................................................................ 40

2-2 Results and discussion ............................................................................................................ 41

2-2-1 Transmission electron microscopy of CNC suspensions ........................................... 41

VI

2-2-2 Scanning electron microscopy of freeze-dried gels .................................................... 44

2-2-3 Confocal laser microscopy ......................................................................................... 45

2-2-4 Zeta potential and hydrodynamic radius .................................................................... 47

2-2-5 Dynamicity of the gel ................................................................................................. 49

2-2-6 Gravity drivel gel collapse of CNC ............................................................................ 50

2-2-7 Re-dispersion of already formed gel .......................................................................... 53

2-3 Conclusion .............................................................................................................................. 54

2-4 Supporting information (CHAPTER 2) .................................................................................. 55

2-4-1 Materials ..................................................................................................................... 55

2-4-2 Materials preparation .................................................................................................. 55

2-4-3 Scanning and transmission electron microscopies ..................................................... 55

2-4-4 Confocal laser scanning microscopy .......................................................................... 56

2-4-5 Zeta potential and particle size measurements ........................................................... 57

CHAPTER 3: Cellulose Nano Crystals structure in the presence of salt .................................. 59

3-1 Introduction ............................................................................................................................ 60

3-2 Experimental section .............................................................................................................. 61

3-2-1 Materials ..................................................................................................................... 61


3-2-3 Materials characterization ........................................................................................... 62

3-3 Result and discussion .............................................................................................................. 64

3-3-1 Transmission electron microscopy of CNC suspensions ................................................ 64

3-3-2 Confocal laser scanning microscopy ............................................................................... 67

3-3-3 Zeta potential and hydrodynamic radius ......................................................................... 70

3-3-4 Dynamics of CNC gel and its relevancy to eventual gel collapse ................................... 71

3-3-5 Gravity driven collapse of CNC gel ................................................................................ 72

VII

3-3-6 Molecular dynamic simulation ........................................................................................ 75

3-4 Conclusion .............................................................................................................................. 78

CHAPTER 4: Nonlinear Viscoelastic Characterization of Charged Cellulose Nanocrystal

Network Structure in the Presence of Salt in Aqueous Media ......................................................... 79

4-1 Introduction ............................................................................................................................ 80

4-2 Experimental section .............................................................................................................. 82

4-2-1 Materials ..................................................................................................................... 82


4-2-3 Materials characterization ........................................................................................... 82

4-2-4 Background ................................................................................................................. 83

4-3 Result and discussion .............................................................................................................. 86

4-3-1 Morphological characterization of freeze-dried hydrogels under SEM ..................... 86

4-3-2 Zeta potential measurement ........................................................................................ 89

4-3-3 Linear viscoelastic behavior of CNC suspensions ..................................................... 90

4-3-4 Inter-cycle nonlinear viscoelastic behavior of CNC suspensions .............................. 90

4-3-5 Intra-cycle nonlinear viscoelastic parameters ............................................................. 93

4-3-6 Frequency dependence of intra-cycle viscoelastic parameters ................................... 97

4-4 Conclusion .............................................................................................................................. 98

4-5 Supporting information (Chapter 4) ..................................................................................... 100

4-5-1 Confocal laser scanning microscopy ........................................................................ 100

4-5-2 Linear viscoelastic behavior of CNC suspensions ................................................... 102

4-5-3 The effect of Cellulose Nano Crystals (CNC) concentration of inter-cycle viscoelastic

behavior of CNC-salt suspensions .......................................................................................... 104

4-5-4 Intra-cycle nonlinear viscoelastic parameters for 30g/L CNC suspension ............... 105

4-5-5 Effect of frequency on inter-cycle response of 20 g/L CNC suspension containing 85.5

mM salt 106

VIII

4-5-6 Lissajous-Bowditch plots of CNC/salt suspensions at 20 g/L CNC suspensions at

different salt concentrations .................................................................................................... 106

4-5-7 Lissajous-Bowditch plots of CNC/salt suspensions at 30 g/L CNC suspensions at

different salt concentrations .................................................................................................... 109

4-5-8 Nonlinear parameters ................................................................................................ 110

CHAPTER 5: Viscoelastic properties of poly (vinyl alcohol) hydrogels with cellulose

nanocrystals fabricated through NaCl addition ............................................................................. 112

5-1 Introduction .......................................................................................................................... 113

5-2 Experimental section ............................................................................................................ 115

5-2-1 Materials and Materials preparation .............................................................................. 115

5-2-2 Materials characterization .............................................................................................. 116

5-2-2-1 Scanning electron microscopy ................................................................................... 116

5-2-2-2 Transmission electron microscopy ............................................................................. 116

5-2-2-3 Rheology .................................................................................................................... 116

5-2-2-4 Compression tests ....................................................................................................... 117

5-3 Result and discussion ............................................................................................................ 117

5-3-1 Morphological characterization of freeze-dried hydrogels under SEM ........................ 117

5-3-2 Morphological characterization of CNC-PVA colloids ................................................ 118

5-3-3 Rheological characterization of CNC-PVA samples ..................................................... 119

5-3-4 Storage modulus-recovery relationship ......................................................................... 122

5-3-5 Lissajous-Bowditch plots of CNC/salt suspensions at different CNC and salt

concentrations .......................................................................................................................... 123

5-3-6 The sequence of physical processes .............................................................................. 126

5-3-7 Overshoot during a start-up experiment ........................................................................ 129

5-3-8 Modeling of CNC orientation based on Folgar-Tucker based models .......................... 130

5-3-9 Mechanical properties of CNC-PVA hydrogels ............................................................ 134

IX

5-4 Conclusion ............................................................................................................................ 135


5-5-1 Mechanical properties of CNC-PVA hydrogels ............................................................ 137

CHAPTER 6: Self-healing and Collapse in CNC-based Gels and Suspensions ...................... 139

6-1 Introduction .......................................................................................................................... 140

6-2 Experimental section ............................................................................................................ 143

6-2-1 Materials ................................................................................................................... 143

6-2-2 Materials preparation ................................................................................................ 143

6-2-3 Materials characterization ......................................................................................... 144

6-3 Results and discussion .......................................................................................................... 147

6-3-1 Confocal imaging accuracy verification ................................................................... 147

6-3-2 Relationship between the CLSM signal and the CNC concentration ....................... 148

6-3-3 Quantitative analysis of CLSM images of CNC gels ............................................... 151

6-3-4 Dynamicity of CNC gel and eventual gel collapse ................................................... 152

6-3-5 Dynamics characterization of CNC clusters in gel using FRAP .............................. 156

6-4 Conclusion ............................................................................................................................ 162


6-5-1 Theory of FRAP ....................................................................................................... 164

6-5-3 Dye binding to CNCs ............................................................................................... 165

6-5-4 Additional considerations ......................................................................................... 165

CHAPTER 7: Summary and conclusion .................................................................................. 167

7-1 Future works ......................................................................................................................... 170

Copyrights ...................................................................................................................................... 174

References ...................................................................................................................................... 184

XI

Table of Figures

Figure 1.1 Molecular structure of Cellulose. The picture was reprinted from Wikipedia ................. 2

Figure 1.2 Cellulose structures in trees from logs size to molecules scale. Figure reproduced from

reference[10] ....................................................................................................................................... 3

Figure 1.3 Fractal dimension as a function of sticking probability in 2-D and 3-D. .......................... 6

Figure 1.4 Depiction of simulated fractal structures with a sticking probability of 1 and 0.1 is

shown. ................................................................................................................................................. 7

Figure 1.5 variation in zeta potential values as a function of pH of the system. The line has been

drawn just as a guide to the eye. ......................................................................................................... 9

Figure 1.6 Aggregation of 10 g/L CNC concentration under 120 C° annealing condition and during

the period of 10 hrs. .......................................................................................................................... 11

Figure 1.7 A) attachment of adamantane and β‐cyclodextrin to hyaluronic acid. B) illustration of

the extrusion process for ink (red) into the support gel (designated as green) C). The capability of

CLSM in showing the results (reprinted from reference [90]) ......................................................... 21

Figure 1.8 Utility of confocal scanning laser microscopy (CSLM) images in capturing gelation of

full-fat milk containing Nile blue. Numbers signify minutes after dye addition. Scale bar = 25 µm

(reprinted from reference [95]) ......................................................................................................... 22

Figure 1.9 Capturing phase separation of glucono-δ-lactone (GDL)-induced gelation of skim milk

using CLSM. Bright areas are protein. Numbers signify minutes after the introduction of the rennet

(phase separation trigger). Scale bar = 25 µm. (reprinted from reference [95]) .............................. 22

Figure 1.10 visual self-healing experiment (a) cut segmented parts, (b) segmented parts were just

brought into contact, (c) crack partially healed after passage of 24 h and (d) thoroughly healed

crack after passage of 48 h (reprinted from reference [99]) ............................................................. 23

Figure 1.11 Storage modulus reported from various references reflecting the effect of CNC on the

reinforcement of different matrices, numbers associated with reference numbers [43, 46-49, 59, 72,

83, 116-123] sketched on double logarithmic axes. ......................................................................... 30

Figure 1.12 Strain sweep test, strain changing g from small to large values [124] .......................... 31

Figure 2.1 Low-mag and high-mag transmission electron microscopy (TEM) images of cellulose

nanocrystals (CNC) aggregation at different CNC and sodium chloride (NaCl) concentrations. (a)

XII

and (d) 5 g/L CNC at 1 mM NaCl; (b) and (e) 15 g/L CNC and 5 mM NaCl; (c) and (f) 15 g/L

CNC and 10 mM NaCl. .................................................................................................................... 43

Figure 2.2 Scanning electron microscopy (SEM) images of the CNC network at different

magnifications for CNC concentrations of (a-c) 7.5 g/L and (d-f) 15 g/L. The concentration of

NaCl for all images is 10 mM .......................................................................................................... 45

Figure 2.3 Growth of CNC network at 15 g/L CNC at different concentrations of NaCl: a (0 mM),

b (0.33 mM), c (0.45 mM), and d (1 mM). The dimensions of the visualization cube are

100×636×636 µm3. The 3-D confocal laser scanning microscopy (CLSM) images were twisted to

obtain a better view of the gel network ............................................................................................ 46

Figure 2.4 Variation of σ as a function of NaCl concentration for CNC gels with 15 g/L

concentration. ................................................................................................................................... 47

Figure 2.5 Rearrangement and slow coarsening of the gel network (black parts) of the aqueous

suspension of 10 g/L CNC with 2 mM NaCl concentration over a period of 30 min. The

dimensions of the two-dimensional visualization box are 636×636 µm2 ......................................... 50

Figure 2.6 Gradual collapse of the gel network of the aqueous suspension of 5 g/L CNC with 10

mM NaCl concentration as a function of time. The dimension of the visualization box is

100×636×636 µm3. The dispersion had a height of 3 mm, and the images were recorded at the

height of ~ 1 mm above the base of the cell. .................................................................................... 52

Figure 2.7 Depiction of a robust gel network of the aqueous suspension of 15 g/L of CNC with 10

mM NaCl concentration as a function of time. The dimension of the visualization box is


height of ~ 1 mm above the base of the cell. .................................................................................... 53

Figure 2.8 Gel network of the aqueous suspension of 7.5 g/L CNC with 10 mM NaCl diluted with

10 ml deionized water: (a) before sonication, and (b) after sonication for 1 min. The dimension of

the visualization box is 100×636×636 µm3. The dispersion had a height of 3 mm, and the images

were recorded at the height of ~ 1 mm above the base of the cell ................................................... 54

Figure 2.9 Gradual collapse of the gel network of the aqueous suspension of 7.5 g/L of CNC with

10 mM NaCl concentration as a function of time. The dimension of the visualization cube is


height of ~ 1 mm above the base of the cell ..................................................................................... 58

XIII

Figure 3.1 High magnification transmission electron microscopy (TEM) images of cellulose

nanocrystals (CNC) with 10 g/L concentration aggregation at different Magnesium chloride

(MgCl2) concentration. (A) 17 mM, (B) 21 mM, (C) 32 mM, (D) 42 mM. The ratio of salt/CNC

varies from 0.17 to 0.42. ................................................................................................................... 65

Figure 3.2 Growth of CNC network at 15 g/L CNC at different contents of MgCl2: (A) 0 g, (B) 4.2

mM, (C) 8.5 mM, (D) 17 mM, (E) 21 mM and (F) 42 mM. The dimensions of the visualization

cube are 100×1272×1272 µm3. The 3-D confocal laser scanning microscopy (CLSM) images are

rotated to obtain a better view of the gel hybrid system. Resolution: 500 nm. Images were taken

once the salt was added into the mixture. The ratio of salt/CNC varies from 0 to 0.27 ................... 68

Figure 3.3 Variation in σ as a function of MgCl2 concentration for CNC gels with 15 g/L CNC

concentration .................................................................................................................................... 69

Figure 3.4 Semi-logarithmic variation of volume percentage as a function of the hydrodynamic

radius of CNC. Inset depicts gradual changes in zeta potential as the MgCl2/CNC ratio changes

from 0 to 0.25. .................................................................................................................................. 70


mM MgCl2 content in the time span of 30 min with intervals of 5 min. The dimension of the

visualization box is 200×1272×1272 µm3, and the resolution is 500 nm. The dispersion has a height

of 3 mm, and the images are recorded at the height of approximately 1 mm above the base of the

cell .................................................................................................................................................... 74


mM MgCl2 content in a span of 30 min with intervals of 5 min. The dimension of the visualization

box is 200×1272×1272µm3, and the resolution is 500 nm. The dispersion has a height of 3 mm, and

the images are recorded at the height of approximately 1 mm above the base of the cell ............... 75

Figure 3.7 Molecular dynamic simulation. (A) Snapshot of two CNC rods in the sodium chloride

solution. The rods are fixed on the x and z-axis, and the rods are parallel (distance between two

rods set at 3.5nm). The transparent material is an aqueous solution in which the Na+ is yellow, and

the Cl- is blue. In the CNC rod system, the carbon atom is cyan, the oxygen atom is red, and the

hydrogen atom is white. (B) Potential mean force (PMF) of two CNC rods (T=298 K, P=0.1 MPa)

in salts solutions of MgCl2. ............................................................................................................... 76

Figure 4.1 Scanning electron microscopy (SEM) images of (a-b) 20 g/L CNC with 42.7 mM salt at

250x and 1000x magnifications, respectively, (c-d) 20 g/L CNC with 85.5 mM salt, (e-f) 20 g/L

XIV

with 172 mM salt, (g-h) 30 g/L with 42.7 mM salt, (i-j) 30 g/L CNC with 85.5 mM salt, (k-l) 30

g/L with 172 mM salt at 250x and 1000x magnifications, respectively. All samples were freeze-

dried out of 10 mL gelled suspension. .............................................................................................. 88

Figure 4.2 Changes in zeta potentials of CNC suspensions at a fixed concentration of 0.5 g/L CNC

and as a function of NaCl concentrations. ........................................................................................ 90

Figure 4.3 Oscillatory amplitude sweep response of CNC 20 g/L suspensions containing different

amount of salt ((a) 1.72, (b) 17.2, (c) 34.4, (d) 85.5, (e) 172 mM) for strain amplitudes of γ0=0.1-

1000% at an angular frequency of ω =1 rad/s using a cone-plate geometry (at a truncation of 101

μm and cone tip angle of 1°) at 25˚C. (f) Critical strain amplitude c (linear to nonlinear transition)

and crossover strain amplitude T (solid to liquid transition). .......................................................... 91

Figure 4.4 Nonlinear viscoelastic measures of CNC 20 g/L-salt as a function of strain amplitude at

an angular frequency of ω =1 rad/s obtained using a cone-plate geometry (at a truncation of 101

μm and cone tip angle of 1°) at 25˚C. (a) Dynamic viscosities (ηM' and ηL') and (b) local

viscoelastic moduli (GM' and GL') for CNC 20 g/L- 85.5 mM salt. (c) Dynamic viscosities (ηM'

and ηL') and (d) local viscoelastic moduli (GM' and GL') for CNC 20 g/L- 172 mM salt. .............. 94

Figure 4.5 Elastic (S) and viscous (T) intra-cycle nonlinearity indices as a function of strain

amplitude for CNC 20 g/L suspensions at 85.5 mM and 172 mM salt contents, measured using a

cone-plate geometry (at a truncation of 101 μm and cone tip angle of 1°) at 25˚C and angular

frequency of ω = 1 rad/s. Schematics show the state of the systems in each regime. ...................... 96


amplitude for CNC 20 g/L suspensions at 85.5 mM salt, measured using a cone-plate geometry (at

a truncation of 101 μm and cone tip angle of 1°) at 25˚C and angular frequency of (a) 0.5, (b) 1, (c)

5, and (d) 10 rad/s. ............................................................................................................................ 98

Figure 4.7 The growth of cellulose nanocrystal (CNC) network at 20 g/L CNC and at different

contents of sodium chloride (NaCl): (a) 0 mM, (b) 42.7 mM, (c) 85.5 mM, (d) 172 mM in

deionized water (DI). The dimensions of the visualization cube are 100×1000×1000 µm3. The

three-dimensional (3D) confocal laser scanning microscopy (CLSM) images are rotated to obtain a

better view of the gel hybrid system. Images were taken immediately after adding the salt into the

mixture. Resolution: 500 nm .......................................................................................................... 101

Figure 4.8 The growth of CNC network at 30 g/L CNC at different contents of NaCl salt: (a) 0

mM, (b) 42.7 mM, (c) 85.5 mM, (d) 172 mM salt. The dimensions of the visualization cube are

XV

100×1272×1272 µm3. The 3D CLSM images are rotated to obtain a better view of the gel hybrid

system. Images were taken immediately after adding the salt into the CNC mixture. Resolution:

500 nm ............................................................................................................................................ 102

Figure 4.9 Linear viscoelastic characterization of CNC-salt solutions/gels for CNC concentration

and at different salt concentrations. (a) Storage modulus (G'), (b) Loss modulus (G''), and (c)

Complex viscosity (|𝜼*|) of CNC/salt solution at different salt concentrations for strain amplitudes

of 1% using a cone-plate geometry (at a truncation of 101μm and cone tip angle of 1°) at 25˚C. (d)

The storage modulus (G') versus salt concentration at an angular frequency of 1 rad/s ................ 102

Figure 4.10 Oscillatory frequency sweep response of CNC solutions containing different amount of

CNC (20 g/L closed symbols and 30g/L open systems) and different salt concentrations at strain

amplitudes of γ0= 1% using a cone-plate geometry (with a truncation of 101μm and a cone tip

angle of 1°) at 25˚C. ....................................................................................................................... 104

Figure 4.11 Oscillatory amplitude sweep response of CNC solutions containing different amount

of CNC (20 g/L closed symbols and 30 g/L open systems) and different salt concentrations for

strain amplitudes of γ0= 0.1-1000% at an angular frequency of ω = 1rad/s using a cone-plate

geometry (with a truncation of 101μm and a cone tip angle of 1°) at 25˚C. .................................. 105

Figure 4.12 Nonlinear viscoelastic measures () of CNC/salt suspensions with 30 g/L CNCs and two

different concentrations of a,c) 85.5 mM and b,d) 172 mM as a function of strain amplitude at an

angular frequency of ω = 1rad/s using a cone-plate geometry (with a truncation of 101μm and a

cone tip angle of 1°) at 25˚C. .......................................................................................................... 105


amplitude for 30g/L CNC solutions and at a) 85.5 and b) 172 mM salt using a cone-plate geometry

(with a truncation of 101μm and a cone tip angle of 1°) at 25˚C and angular frequency of ω=1rad/s.

........................................................................................................................................................ 106

Figure 4.14 Oscillatory amplitude sweep response of CNC 20 g/L solutions containing at salt 85.5

mM for strain amplitudes of γ0=0.1-1000% at different angular frequencies using a cone-plate

geometry (with a truncation of 101μm and a cone tip angle of 1°) at room temperature. ............. 106

Figure 4.15 Dimensionless Lissajous-Bowditch loops for CNC 20 g/L solutions containing (a, b)

17.2, (c, d) 85.5, and (e, f) 172mM salt, measured using cone-plate geometry (with a truncation of

101μm and a cone tip angle of 1°) at 25˚C. Projections on the elastic (τ - γ) and viscous (τ - dγdt)

XVI

planes are presented at strain amplitudes of γ0 = 1, 40, 100, and 250% and at an angular frequency

of ω = 1rad/s .................................................................................................................................. 108

Figure 4.16 Dimensionless Lissajous-Bowditch loops for CNC 30 g/SL suspension containing a,

b) 85.5, c, d) 172 mM salt using cone-plate geometry (with a truncation of 101μm and a cone tip

angle of 1°) at 25˚C. Projections on the elastic (τ - γ) and viscous (τ - dγdt) planes are presented at

strain amplitudes of γ0 = 1, 40, 100, and 250% and an angular frequency of ω =1 rad/s. .......... 110

Figure 5.1 Scanning Electron Microscopy (SEM) micrographs of cellulose nanocrystals (CNC)-

poly (vinyl alcohol) hydrogels (PVA) freeze dried samples: a) (CNC 10 g/L), b) (CNC 15 g/L), c)

(CNC 25 g/L), d) (CNC 30 g/L) at magnification of 100x. e) The average pore size of samples as a

function of CNC concentration ....................................................................................................... 118

Figure 5.2 Distribution of CNC particles embedded in PVA at different concentrations and at 0.5

µm resolutions respectively: a-d (CNC 10-15-25-30 g/L) ............................................................. 119

Figure 5.3 a) Storage, b) loss modulus and c) complex viscosity as a function of frequency at strain

amplitude of γ0 = 1%. d) Storage (solid symbols) and loss moduli (open symbols) as a function of

strain amplitude at angular frequency of ω = 1rad/s. ..................................................................... 120

Figure 5.4 Storage modulus as a function CNC concentration. Values are extracted from Figure

5.3. .................................................................................................................................................. 120

Figure 5.5 a) Flow curve (γ = 1s-1) and b) reconstruction of PVA-CNC hybrid hydrogel network

after breakage as a function of time (γ0 = 1% and ω = 1rad/s). ...................................................... 123

Figure 5.6 Lissajous Bowditch plots: a) stress versus strain b) stress versus strain rate for CNC-

PVA/salt hydrogels at different CNC contents and strain amplitudes of γ0 = 1, 7, 10, 40% and

angular frequency of ω=1rad/s. ...................................................................................................... 125

Figure 5.7 Lissajous-Bowditch plots of a) CNC(10 g/L)/PVA/salt at strain amplitudes of γ0 = 10,

14, and 19%, b) CNC(30 g/L)/PVA/salt at strain amplitudes of γ0 = 3.5, 5, 10, 14, and 19%. c) and

d) open gray circles represent power-law flow response of CNC(30g/L)/PVA/salt (solid line, in

corresponding to a strain amplitude of c) 10% and d) 100% and angular frequency of ω=1 rad/s,

raw waveform as a black solid line). The direction of traversal is indicated by the dashed arrows.

........................................................................................................................................................ 129

Figure 5.8 Transient shear stress of the CNC-PVA hydrogel measured at a shear rate of 1/s and 25

C°. ................................................................................................................................................... 130

XVII

Figure 5.9 Fitting FT equation on experimental data a) CNC 30 g/L using μ2 = 4300 CI = 0.006

and quadratic approximation. b) CNC 10 g/L using μ2 = 2700 CI = 0.002 and quadratic

approximation at shear rate of 1/s. c-d) Fitting SRF equations on experimental data for CNC 30 g/L

using μ2 = 4300 CI = 0.006 and quadratic approximation under k values of 0.7, 0.8 and 0.85 at

two magnifications. ........................................................................................................................ 133

Figure 5.10 Stress vs. Strain relationship for CNC-PVA freeze-dried samples under compression

test ................................................................................................................................................... 134

Figure 5.11 a) recovery of 10 g/L sample at strain of 1% and 30% b) recovery of 30 g/L sample at

strain amplitude of 1% and 30%. .................................................................................................... 137

Figure 5.12 a) Storage modulus recovery of 10 g/L sample at strain amplitude of 1% after 3 cyclic

strain- storage modulus recovery b) storage modulus recovery of 30 g/L sample at strain amplitude

of 1% after 3 cyclic strain- storage modulus recovery ................................................................... 138

Figure 6.1 Gaussian bleached area (circular) immediately after bleaching in the sample of CNC

with 45 g/L concentration and 20 mM NaCl .................................................................................. 146

Figure 6.2 Distribution of Polystyrene microparticle sizes and their respective CLSM micrographs

presented at 5 µm scale bar. ........................................................................................................... 148

Figure 6.3 A linear relationship between the mean signal value and the concentration of CNC. The

concentration ratio CNC/FB28 for all samples is set at 0.01. The measurement is done at 1 mm

above the base of the petri-dish. ..................................................................................................... 150

Figure 6.4 CNC concentration distribution for original suspension of CNC 5g/L with the addition

of NaCl at 10, 15, 20, 25, and 30 mM ............................................................................................ 152

Figure 6.5 Mean signal intensity decay for the CNC concentrations of (a) 6 g/L and (b) 30 g/L at 0-

43.1 mM NaCl loadings. Results are captured at 1 mm location above the base of the petri dish (the

initial gel height is 5mm) with a 10x apo lens (NA=0.45) and at the timing of 5 min after gelation.

........................................................................................................................................................ 153

Figure 6.6 Mean signal intensity decay for 30 g/L CNC concentration and 43.1 mM NaCl

concentration at different gel initial heights. Results were captured at a 1 mm location above the

base of the petri dish with a 10x apo lens (NA=0.45) 5 min after the introduction of NaCl. ........ 154

Figure 6.7 Average loss of CNC out of the control box over 400 s period of the experiment for

CNC 30 g/L (top) CNC 10 g/L (middle) and CNC 6 g/L (bottom). ............................................... 155

XVIII

Figure 6.8 Depiction of changes in cluster size and mobility of CNCs in the CNC-DI water

suspension system on semi-logarithmtic scale. The minimum in the z-factor can be due to the

retraction of double layer. ............................................................................................................... 156

Figure 6.9 (a-c) Frap recovery curves for CNC with a concentration of 6g/L at 0, 20 mM, and 86.2

mM NaCl loadings. (d-e) Temporal-spatial CLSM 3-D images of the sample with 6g/L CNC

captured at 0, 10, and 15 seconds after bleaching. Visualization box size: 148.347 µm×148.347

µm, resolution 500 nm. ................................................................................................................... 157

Figure 6.10 (a-f) FRAP recovery curves of samples with the CNC concentrations of 45 g/L and 30

g/L at various concentrations of NaCl (0, 17.2, 34.4, 51.7,70 mM) ............................................... 158

Figure 6.11 (a-c) Variation of diffusion coefficient as a function of NaCl loadings (0, 8.62, 20, 86.2

mM) depicted on left-hand side and immobile particle percentage as a function of NaCl loadings

(0, 8.62, 20, 86.2 mM) depicted on right-hand side of each figure ................................................ 158

Figure 6.12 Time series of FRAP done on CNC 45 g/L sample with PL APO 10x optic (NA=0.45).

The sample immobile fraction is 0% after 40 seconds. Scale bar;10 µm resolution 500 nm; and ROI

size 10 µm. ...................................................................................................................................... 159

Figure 6.13 The FRAP recovery of the sample with 45 g/L CNC and 20 mM NaCl that over a span

of 8.4 min has not healed. The height of the 3-D image shows changes in time that were

continuously captured. The visualization box size is 140 µm ×140µm and resolution is 500 nm. 160

Figure 6.14 (a) Measured diffusion coefficients normalized by the Dinfinite, as a function

of(a/b)2φ. Note that (a/b)2φ is proportional to the number of rods in the volume L3 with L the

length of the rods. The results of Brownian dynamics and Edwards-Evans equation are also given

as a function of concentration. (b) universality graph that connects data of zeta potential, storage

modulus, and immobile fraction obtained through FRAP analysis. The lines in the above graphs of

FRAP data, zeta potential values, and storage modulus are drawn as a guide to the eye. ............. 161

Figure 6.15 Filtered CNC-FB28 in condition (a) Before and (b) after exposure to UV light ........ 165

XIX

Table of Tables

Table 1.1 Highlights of studies on CNC-PVA hydrogels with the reference number and main

contents. ............................................................................................................................................ 11

Table 1.2 Highlights of studies on CNC only and CNC composites hydrogels with the reference

number and main contents. ............................................................................................................... 14

Table 2.1 Changes in CNC suspensions zeta potential at a fixed concentration of 0.5 g/L CNC as a

function of NaCl concentration. ....................................................................................................... 48

Table 3.1 Force field parameters employed in the system. .............................................................. 77

Table 4.1 Characterizing the elastic nonlinearity in response to imposed large amplitude oscillatory

shear (LAOS) deformation. ............................................................................................................ 110

Table 4.2 Characterizing viscous nonlinearity in response to imposed LAOS deformation.......... 111

XX

List of Symbols and Abbreviations

Cellulose Nano Crystal (CNC)

Calcium chloride (CaCl2)

Scanning Electron Microscopy (SEM)

Sodium chloride (NaCl)

Transmission Electron Microscopy (TEM)

Magnesium chloride (MgCl2)

Dynamic Light Scattering (DLS)

Carbon dioxide (CO2)

Thermogravimetric analysis (TGA)

Polyvinyl alcohol (PVA)

Confocal Laser Scanning Microscopy (CLSM)

Hyaluronic acid (HA)

Diffusion limited aggregation (DLA)

Hydroxyethylcellulose (HEC)

Reaction limited aggregation (RLA)

Sodium carbonate (Na2CO3)

Hydroxyapatite (Hap)

Carboxymethylcellulose (CMC)

Extracellular matrix (ECM)

Poly (ethylene glycol) (PEG)

Poly (e-caprolactone) (PCL)

Polyurethane (PU)

XXI

Polymethyl methacrylate (PMMA)

Polycarbonate (PC)

Polyethylene terephthalate (PET)

Equivalent Oxygen Percentage (EOP)

Brunauer and Emmett and Teller (BET)

Fluorescent brightener 28 (FB 28)

PDI (molecular weight distribution)

Relative humidity (RH)

DMSO (dimethyl sulfoxide)

Drug delivery system (DDS)

Linear viscoelastic regime (LVR)

Small amplitude oscillatory shear (SAOS)

Medium amplitude oscillatory shear (MAOS)

Large amplitude oscillatory shear (LAOS)

Small angle neutron scattering (SANS)

Derjaguin-Landau-Verwey-Overbeek (DLVO) theory

Alberta Innovates Technology Futures (AITF)

Critical aggregation concentration (CAC)

Potential of mean force (PMF)

Particle mesh Ewald (PME)

Steered molecular simulation (SMD)

Natural Sciences and Engineering Research Council of Canada (NSERC)

Three-dimensional (3-D)

XXII

Two-dimensional (2-D)

XXIII

Greek letters

Storage modulus (G')

Loss modulus (G'')

Complex viscosity (|𝜼*|))

Critical strain amplitude (c)

Cross-over strain amplitude (T)

Fractal dimension (df)

Shear stress (σ)

Temporal phase shift (δ1).

Large-strain modulus (GL′ )

Minimum-strain modulus (GM′ )

Minimum-strain rate viscosity (η′M)

Large-strain rate viscosity (η′L))

Angular frequency ( ω )

Electrophoretic mobility (µ)

Electric field (E)

Zeta potential (ζ)

Dielectric constant (εr)

The permittivity of the free space (ε0)

Dynamic viscosity (η)

Volume fraction (φ)

1

CHAPTER 1: Introduction

Nanomaterials have great utility in many applications, including medicine tablet production, boards

in electronics, scaffolds in biomaterials, and aerogels in energy storage/production. A key constituent

of nanomaterials with the option of being renewable is cellulose, produced in the shape of cellulose

nanocrystals (CNCs), nano-fibrillated cellulose, and cellulose with bacteria origin [1-3]. These

nanoparticles can be obtained from various natural sourced cellulose; cellulose is most generated

plant material in nature. These nanoparticles are advantageous over inorganic materials, such as

biocompatibility, biodegradability, and good mechanical properties. Nanocellulose in either of these

three shapes has potential usage in many areas, including but not limited to: scaffolds, surface

coatings, hydrogels, polymer composites, and as an emulsifier.

Programs running worldwide are focusing on the extensive development of nanocellulose products

such as Suomen Nanoselluloosakeskus [4]; Technical Association of the Pulp and Paper Industry

(TAPPI) [5] and joint companies of FPInnovations and CelluForce [5]. From these organizations,

CNCs are commercially available at large quantities with uniform properties and high purity.

The main goal for research on CNCs is to fully exploit the incredible physical and chemical

properties of CNCs in various applications. Distinct properties of CNCs, has led them to be used

vastly in different polymers as reinforcing agents. There are, however, challenges faced, such as

tuning the interactions between CNCs and polymer matrices and achieving uniform distribution and

dispersion of CNCs in matrices. Moreover, CNC on its own can be assembled into films and gels.

The very shape of CNCs enables them to behave like liquid crystals or fibers. The rod shape of CNCs

can enable them to generate different morphologies as a nanostructure, this potential caused them to

have various final properties and functions.

Addressing multiple health-related problems using engineered and biocompatible materials is a

common practice in the industrial world. Development of scaffolds can be put forth as an example

for tissue engineering application. Tissue engineering as a growing interdisciplinary realm of science

involves usage and development of bio-related materials, cell biology, and cell-matter exchanges.

Mainly regeneration tissue and promotion of tissue functions are targeted in this field. It also targets

to replace defective or damaged organs inside the body or tissues that have been heavily damaged.

Many requirements are envisioned for scaffolds such as mechanical properties, biocompatibility,

2

degradability, cell adhesion, to name a few. Mechanical properties are among key parameters for

fabrication of scaffolds. Mechanical strength is characterized by the impact resistance of products

with the aim of maintaining the scaffold mechanically during implantation [6, 7] and after

implantation. The most common mechanical tests to evaluate scaffold include tensile and

compressive tests. This thesis focuses on the development of engineered hydrogels out of cellulose

nanocrystals (CNC) suitable for performing as a scaffold.

1-1 Cellulose

Cellulose, with the molecular structure depicted in Figure 1.1, is abundant in nature.

Figure 1.1 Molecular structure of Cellulose. The picture was reprinted from Wikipedia

According to the depicted structure of cellulose (Figure 1.1), three hydroxyl groups on the side of

chains could promote hydrogen bonding between chains, which yield a highly concentrated crystal

system [8]. Theoretically, the crystalline part of CNC could get to one hundred percent. However, if

the external extraction of amorphous regions is not sufficient, the normal attainable crystallinity

range is between 55 to 90% based on different cellulosic sources and reaction conditions [9].

Generally, this water-insoluble biomacromolecule is an important part of plant cells where it grants

the plant high tensile strength. Figure 1.2 depicts the wood structure at different length scales. Meters

(the whole tree), centimeters (the cross-section), millimeters (size of growth rings), tens of

micrometers (cellular level), micrometers (a layered structure within plant cell walls), tens of

nanometers (cellulose fibrils) and nanometer (hemicellulose and lignin)[10]. Cellulose is not only

found in plants, as it also has sources in several animals such as tunicates, and to a lesser degree in

microorganisms such as bacteria (A. xylinum, A. hansenii), algae (Chaetomorpha and Cladophora),

fungi (mycelium or yeast cell), invertebrates, and amoebae (protozoa)[3].

3

Figure 1.2 Cellulose structures in trees from logs size to molecules scale. Figure reproduced from reference[10]

Due to cellulose being rooted in nature and hydrophilicity, it has been widely researched and studied

for more than 150 years. Currently, applications for cellulose range from construction, the food

industry, paper industry, biomaterials, and pharmaceuticals [11]. Moreover, properties of CNCs such

as biodegradability, renewability, eco-friendliness is also a major driving force behind further

developments of these products.

1-1-1 Cellulose at the nanoscale

Nanocellulose, in the form of CNCs, is a promising nanomaterial that can be produced at a low cost.

It also has special properties such as high strength, lightweight, liquid crystalline behavior,

biodegradability, and general biocompatibility [2, 3]. CNCs also have been named as cellulose

nanowhiskers or nanocrystalline cellulose (NCC) and nanofibrillated cellulose, also as is micro

fibrillated cellulose (MFC) or cellulose nanofibrils (CNF) [11-13].

Cellulose nanocrystals and nano fibrillated cellulose are the most widely manufactured fibers. With

more new manufacturing facilities and technologies, it is expected that the cost will be controlled in

4

an acceptable range (below $10/kg) in the future to further promote the use of nanocellulose in a

variety of applications [14, 15]. Cellulose being a economical and environmentally friendly material

is estimated to vastly contribute to technology in the 21st century [5].

1-1-2 Cellulose nanocrystals

Cellulose nanocrystals are rod-shaped particles with a few nanometer diameters, and lengths that

spans ranges from hundreds of nanometers to microns. CNCs, in comparison to cellulosic fibers,

have advantages such as higher length, higher surface area, and additional mechanical properties.

Studies have shown that CNCs display low cytotoxicity with a range of animal and human cell types

[16-20]. Moreover, it has been corroborated that CNCs pose low ecotoxicological risk and toxicity,

whether oral or dermal, is minimal [16, 20]. Therefore, in 2013, CNCs have been validated as the

very first safe nanomaterial on Environment Canada’s domestic substance list[21, 22].

Sulfuric acid is commonly used for the manufacturing of cellulose nanocrystals through eliminating

amorphous regions inside the native cellulose and then fabricating the quasi-stable suspensions with

negative charges on the surface. As a result of treating with the acid (i.e., esterification reaction with

hydroxyl groups), the -SO42- groups are generated and installed on the surface of nanocrystalline

cellulose. Through reaction time, one can tune the extent of conversion of hydroxyl groups on the

crystalline regions into sulfate groups [23]. As will be covered in subsequent sections for biomedical

applications, cellulose fibrils are able to provide mechanical support to the cells that they host [24].

Since cellulose is of a natural origin, biocompatibility [25] and bio-degradability are other traits that

make it an attractive biocompatible material. The reported average length of CNCs changes between

200-600 nm, while the width varies substantially between 3 nm to 50 nm. In rare cases, the cross-

sectional area will get to higher values due to the tiny aggregations of nanoparticles [26].

The ordered crystalline portions of nanocrystalline cellulose could also highly influence the

penetration of organic solvents and take in of water. Consequently, the ordered crystalline portions

decrease the number of bounded waters and fabricate a very difficult path for transmittance of gas

and water [27]. Thus, CNCs with qualities such as high stiffness and good barrier ability properties

can be a good reinforcement agent in biopolymers.

5

CNC owns a high inherent elastic modulus, which is the most important mechanical parameters for

fabrication of nanocomposites materials [28]. Some reports predict that Young’s modulus of CNCs

is approximately a few folds higher than steel or magnesium metal mixtures [29]. Theoretically, the

elastic modulus of CNCs could reach 145 GPa [30], and Young’s modulus has a wide range from

100 GPa to 200 GPa depending on the cellulose sources [31]. In the following section, suspension,

and gels of CNCs will be discussed.

1-2 CNC in colloidal suspensions and gels

To employ colloid as a term for a suspension, the mixture constituent must not settle at all, or it

should, at minimum, take a long time to settle. Moreover, the size of the particles suspended inside

the continuous phase should vary, between approximately 1 nm to 1000 nm. CNCs, due to their

diameters being in nanometric size, fulfill the requirement to be considered colloid in water.

Colloidal stability is governed by V, the interaction potential between particle and specifically by the

amount of the energy barrier ∆V. If the energy barrier well surpasses the thermal energy, then

aggregation is prevented. On the contrary, if the energy barrier drops below kT (k is the Boltzmann

constant and T represents temperature), then aggregation gets initiated, if the other neighboring

particles get in proximity. In the presence of enough salt, the condition of having energy barrier

dipping below kT occurs. Under this condition, CNC particles adhere, and depending on CNC

concentration, collapse, or a 3-D gel can be readily obtained [32]. For instance, in a study carried out

by Cherhal et al. [33], due to the slender shape of their CNCs upon the gradual addition of salt, the

system changed into a space-filling gel at low concentration of CNC.

A stabilized colloid consists of particles that are disallowed to aggregate due to repulsive forces.

However, when the existing repulsive interactive forces become weaker or get screened through the

incorporation of a coagulant, particles start to aggregate. Accordingly, when the interaction potential

acting between particles becomes positive, the aggregation process is limited by solely Brownian

diffusion of the particles, a regime also known as diffusion-limited aggregation (DLA). Upon

reaching to intermediate values, the aggregation gets slower since a higher number of collisions are

needed for successful aggregation, a regime also known as limited reaction aggregation (RLA).

Placement in either of these zones occurs depending on the salt concentration in a CNC-salt system.

6

Diffusion-limited aggregation (DLA) is a regime where, because of Brownian motions, nanoparticles

go through random walk fashion movement and eventually assemble to form aggregates. The clusters

fabricated in the aforesaid process are called Brownian trees. Herein, we have briefly performed

monte Carlo simulations in a lattice for 2-D and 3-D diffusion and found fractal dimensions of

clusters. The DLA attachment models of colloids were first introduced by Witten and Sander [34].

In their models, a nucleus particle was fixed at a certain point, and then particles were set loose into

the system one after another to migrate onto the clusters. Each particle migrates to a neighboring

particle in the cluster until it becomes part of it. It is accepted that a cluster formed this way has a

self-similarity property in all considered length scales. The fractal dimension associated with the

formed clusters 1.68±0.05 [35] in 2D and 2.5 ±0.05 [34] in 3D. In DLA simulations, when a particle

joins an existing cluster, it becomes stuck. However, if one defines a probability of sticking then the

particles would stick to the cluster at a probability between 0 and 1 which in the present work, we

refer to it as sticking probability and found the fractal dimension as a function of sticking probability

for 2-D and 3-D (see Figure 1.3 and Figure 1.4). The simulations have been conducted with kill

zone boundary conditions, and the initial particles number was set at 500000. If the cluster size were

to exceed the simulation box, the simulation would become terminated. Fractal dimensions were

estimated using image J and its box counting module. For each point on the graph, simulations have

been run 20 times, and results show the average values with the standard deviation.

Figure 1.3 Fractal dimension as a function of sticking probability in 2-D and 3-D.

7

Figure 1.4 Depiction of simulated fractal structures with a sticking probability of 1 and 0.1 is shown.

1-2-1 Parameters affecting aggregation and gelation of CNCs

One of the goals of rod suspension preparation is its applicability in a networked state. Rod

suspension is preferable to be in the state of jammed or gelled to produce required mechanical

properties. Jamming is a reference to arrested state of the particle kinetically, while, gelled state is a

reference to the structure formed with particles at the gelled state. Therefore, knowing mechanism

that leads to either of these two states is important. The forces between colloidal particles normally

involves, electrostatic, depletion forces or the nature of the force is frictional. These forces cause the

particles to form networks of rods without interaction or fractal clusters.

Forces acting between particles dictate whether the system is at the jammed or gelled state. Forces

acting between rod particles are either attractive Van der Waals forces and hydrogen bonding or

repulsive such as electrostatic forces. Van der Waals forces and hydrogen bonding interactions

between CNC happens naturally when two CNC particles get in proximity to one another. For the

case of CNCs, due to the acid involved in the method of production, there is an inherently negative

charge on CNCs. Electrostatic forces are due to the negative charge on the CNC surface that keeps

two CNC particles separated, and they act at long distances.

Now that the nature of forces has been reviewed, questions about the magnitude of such forces and

their impact must be considered. For the case of CNC suspension, quantification of the strength of

forces can happen using quantity Ucontact/kBT (non-dimensional), where Ucontact is potential pair

8

interactions of two cylindrical particles at the time of contact and kBT is the thermal energy attributed

to Brownian forces. Ucontact/kBT <<1 is for non-interacting regime while Ucontact/kBT >>1 is related to

interacting regime in which particles have high affinity towards each other. In such cases collision

between particles leads to irreversible aggregation. Values situated in between is related to slow

aggregation or phase separation that happens slowly.

In combination to the parameters mentioned here number density will complement the set of

parameters responsible for controlling the microstructure. ρ is proportional to the volume fraction φ;

φ=ρVp where Vp is the volume of the particle. If we assume the system is non-interacting (particles

have no affinity towards one another), there zone of interest can be defined. In the following, consider

L, length of particles, b their diameter, and r their aspect ratio. When ρ << (1/L3) or equivalently (φ

<< (1/r2)), the suspension is dilute, and rods do not make any contact both structurally and

dynamically. When (1/L3) <<ρ<<(1/bL2) or equivalently (1/r2<<φ<<(1/r)), rods make few structural

contacts with the particles in their vicinity. Finally, when the concentration is high, i.e. ρ >> (1/bL2)

or equivalently (φ>>(1/r)), rod rotation is hindered by neighboring particles, both spatial and

dynamically. Putting a suspension into this concentration regime also yields the order-disorder

transition of liquid crystals[36-39].

Electrostatic forces between CNCs are under the influence of ionic strength of medium and external

parameters such as the pH of the media. For instance, at various pH values (2 < pH < 14) CNC

colloidal suspension stability is different. Generally, increasing pH values in CNC suspension will

drive zeta potential toward higher negative values. Experimentally, the effect of pH on gelation can

be studied via zeta potential. For instance, in Figure 1.5, the change in the zeta potential of CNC

with 0.5 g/L and ten mM as a function of pH has been displayed. For this system, zeta potential

values surge to higher absolute values with an increase in pH.

9

Figure 1.5 variation in zeta potential values as a function of pH of the system. The line has been drawn just as a guide

to the eye.

The anisotropicity of the excluded volume of cylindrical particles also shapes their packing geometry

and changes their stability. Both Brownian motions, i.e. translational and rotation are under influence

of rods length [40]. The aspect ratio (length over diameter) of particles can influences the points the

system morphs into a liquid crystalline phases [36, 37], as also can polydispersity [41].

1-2-2 Gelation mechanisms of CNC

Due to the addition of a coagulant CNCs in this thesis, coagulate in the form of gelled structure;

however, there are other methods of coagulation in the literature. Freeze-thaw cycles [42], annealing

at high temperature [43], flocculation with polymers [44], and coagulation through depletion

mechanism [45] are other methods currently used for gelation. However, this thesis primarily focuses

on using a coagulant, such as NaCl or MgCl2, on promoting gelation. Using this method of gelation,

for instance, in the case of CNCs, drawing from examples in the literature, depicts that CNCs above

10 wt % will transition into an aggregated gel-like phase [46]. Alternatively, gelation can also

happen with usage of chemical bonds or physical bonds using other multivalent ions [47, 48].

Gelation in the presence of coagulant is also an adjustable process. Pragmatically, the gel point

signifies a threshold upon which CNC as a function of concentration reaches a critical state due to

solution conditions, modifications of surface of CNCs, and/or addition of adsorbing or nonadsorbing

10

polymers into the mix [47, 49-51]. Studies showed, in the report of Chau et al. [48] that upon

increasing ionic strength of medium around suspended CNCs via salt addition, electrostatic repulsion

will be screened, and attractive forces will become more dominant (e.g., van der Waals and hydrogen

bonding). Specifically, reports show increasing the charge number, and the radii of the ions both led

to increases in gel stiffness. Regardless of the cation type added (Na+, Mg2+, Al3+, Ca2+, Sr2+), the

sol−gel transition was determined to occur around 1.5 wt %, roughly one order of magnitude lower

than the threshold of aggregation for pure CNC suspensions investigated by Uren a-Benavides et al.

[46].

Alternatively, physical stimuli can be used to drive gelation. Way et al. [47] used the

functionalization of CNCs using placing carboxyl or amine groups on top of CNCs to make them

pH-responsive. Considering amine functional groups, at high pH, the amine functional groups had a

neutral charge that allowed the attractive forces to become dominant; however, at low pH,

protonation of amine groups caused electrostatic repulsions. For carboxylic acid, the opposite trend

was observed as the storage modulus of a suspension of 2.7wt% CNCs surged by three orders of

magnitude when pH decreased from 11 to 1.41. Way et al. [47]. also showed CNC films made with

the PVA matrix could become mechanically adaptive through changes in pH.

Adsorption of polymeric chains onto the surface of CNC can also induce gel formation, although the

yielding gels are not CNC purely made gels. Hu et al. specially reported that 3wt% of CNCs could

be pushed to make a nematic gel via adsorption of polymers, with 0.2 wt% nonionic polysaccharides

(hydroxyethylcellulose polymer, locust gum, or hydroxypropyl guar polymer). Mechanism

adsorption onto CNC can make the particles artificially bigger, and this will shift the liquid

crystalline phase diagram to lower concentrations [49].

It is worth mentioning, as stated earlier, that gelation could also be thermally initiated although the

thermal treatment at elevated temperatures. This mode of gelation was related primarily to

desulfation at high temperatures. Lewis et al. [43] showed that heating suspension of CNCs above

80 °C resulted in gel formation. Dorris and Gray [52] also showed in a similar manner that CNC

desulfation could induce gel formation in dilute CNC suspension floating in glycerol/water matrix.

To verify whether CNC can gel through annealing at high temperatures, we performed a simple,

short experiment. To promote gelation, CNCs in an autoclave chamber were annealed at high

11

temperatures. Figure 1.6 shows the aggregation of 10 g/L CNC suspension happened in autoclave

format via thermal annealing at 120 C° during the period of 10 hours. This method of gelation is

equally important, as there is no additive involved. However, the focus of this thesis is not gelation

of CNCs, which means other gelation mechanisms will not be followed further. In short, CNC can

gel using methods mentioned earlier, and depending on the final application, one or multiple

strategies should be followed. Moreover, it should be noted that throughout the thesis same CNC has

been used therefore surface chemistry between different sections of this thesis does not change.

Figure 1.6 Aggregation of 10 g/L CNC concentration under 120 C° annealing condition and during the period of 10

hrs.

1-2-3 CNC-composite hydrogels: review

The ability of CNC to form hydrogels, alone or in combination with polymers, has been extensively

investigated. Polymers inclusion can significantly enhance the mechanical properties of CNC-based

hydrogels. Herein, a review of the CNC-based hydrogels with and without polymers is presented

(see Table 1.1 for the highlights of existing work). As we target biocompatible and biodegradable

hydrogels, the emphasis is on the CNC-based hydrogels coupled with polymers such as polyvinyl

alcohol (PVA) that meets the bio-related requirements (see Table 1.2). This is a brief review

summarizing the reports on improving the properties of CNC-composite hydrogels. One may refer

to the thesis chapters for a detailed review.

Table 1.1 Highlights of studies on CNC-PVA hydrogels with the reference number and main contents.

Reference-Hydrogel composite developed Highlights of the study

12

Chen et al. [53]- CNC-PVA-TPS An increase of ~20% and 33% in the composite of starch

and ~40% and ~50% increase in PVA composite tensile

modulus was reported for CNC nanocomposites with 1

and 2% CNC loadings.

Abitbol et al. [54]- CNC-PVA The water sorption capacity of the gels was augmented with

an increased amount of CNC due to CNC being hydrophilic

and reduction in PVA crystallinity. In the compression and

tensile test, the elastic modulus of PVA-CNC showed

improvement compared to PVA pure hydrogel.

Tanpichai and Oksman [55]- CNC-PVA The compressive strength of hydrogels at 60% strain for

the hydrogels with only 1 wt% CNCs surged from 17.5

kPa to 53 kPa. Creep elasticity decreased in presence of

CNCs as molecular chain mobilities were restricted. The

strain recovery of about 97% was observed for samples

containing CNCs, while it was 92% for the PVA-

crosslinked system.

Song et al. [56]. PVA-CNC With samples having 1.5 wt% CNCs, the compressive

strength of PVA foams was augmented from 7 to 58 kPa

for a period of 10s initial reaction time. These values

changed to 65 to 115 kPa if the initial reaction time were

adjusted to 120s.

Song et al. [57]. PVA-CNC Due to hydrogel bonding and intermolecular

interactions, the interactions between PVA and CNCs

caused the hydrogels to gain better mechanical

properties.

Butylina et al. [58]. PVA-CNC The loading of PVA had the largest influence on the

morphology of hydrogels. In comparison to the hydrogel

made with PVA only, the presence of CNCs decreased

the crystallinity of PVA/CNC. In the case of 5% PVA

hydrogel, addition of CNC was reported to increase the

degree of swelling and water content.

13

Zhou et al. [59]. PVA-CNC Due to orientation in the direction of shear at low

concentrations, the system showed shear thinning

behavior. Upon changes in concentrations, the

possibility of collision in the nanoparticle’s population

disallowed the decrease in viscosity, which

subsequently lead to stability after the shear-thinning

region.

Gonzalez [60]. PVA-CNC It was found that the addition of cellulose nanoparticles

to the gel allowed authors to control the pore

morphology of the samples. It was also found that the

presence of CNCs maintained the hydrogel composite

transparency, while thermal stability and mechanical

properties were increased.

Li et al. [61]. PVA/carboxylated CNCs At swelling condition, it was found that hydrogel films

could get stretched 3 to approximately 3.4 times their

initial length. The tensile strength was also found to be

in the range of 7.9 to approximately 11.6 MPa.

Mihranyan [62], Chemical crosslinked CNC-PVA The viscoelastic characteristic of the fabricated

hydrogels was improved by cross-linking, which pushed

the values of G′ and G″ on the order of 10 kPa, which is

noticeable for biomedical applications.

Ben shalom et al. [63]. CNC-PVA The presence of CNC improved tensile strain at break

and toughness to 570% and 202 MJ m-3 values,

respectively. Using the crosslinker helped with

improvement in tensile strength, toughness and modulus

as compared to CNC purely made sheets. Upon an

increase in density, cross-linking increased while

transparency improved, while water absorption level of

cross-linked CNC and CNC-PVA sheets decreased.

14

Anirudhan and Rejeena [64], Poly(acrylic acid-co-

acrylamide-co22-acrylamide-2-methyl-1-propane

sulfonic acid)-grafted nanocellulose/poly(vinyl alcohol)

The hydrogel showed good swelling behavior, and it

was found the drug delivery curve is promising.

Table 1.2 Highlights of studies on CNC only and CNC composites hydrogels with the reference number and main

contents.

Reference-Hydrogel composite developed Highlights of the study

Ooi et al. [65]. Gelatin-CNC The overall crystallinity level and properties of gelatin-

CNC hydrogel showed increase. In the case of addition

of 25% CNC, the impact of crystallinity on storage

modulus caused the modulus to modify from 122 Pa to

468 Pa. pH sensitivity was shown through inspection of

swelling tests of CNC-gelatin hydrogels.

You et al. [59]. Polysaccharide-CNC Authors showed that fabrication of an injectable

polysaccharide hydrogels is beneficial to biomedical

applications. However, they found that poor mechanical

properties inhibit such hydrogels from being used

efficiently.

Ghavimi et al. [66]. Gelatin-CNC Desirability of hydrogel in terms of mechanical

properties was dependent on ionic and covalent ratios.

The significance of osteoinductivity (bone forming

ability) of these hydrogels showed their capability to be

employed as an injectable ensemble for spinal fracture

cases.

Bertch et al. [67]. CNC only hydrogel Authors verified injectability of hydrogels were using

combination of shear and oscillatory rheology. This

aspect showed that in capillaries, flow is mostly plug

flow and the structure of the hydrogel will remain intact

after injection.

15

Nigmatullin et al. [68]. CNC only hydrogels

(functionalized CNC)

Authors showed the possibility of making of cellulose

nanocrystals that glue to each other due to their

associative hydrophobic interactions by modifying

sulfated CNCs (sCNCs) with octyl-CNCs. It was found

that functionalized CNCs gellify at a significantly lower

concentration than un-modified CNCs, and they can

make strong hydrogels.

Liu et al. [69]. Cellulose acetoacetate (CAA),

hydroxypropyl chitosan (HPCS), and amino-modified

cellulose nanocrystals (CNC-NH2)

Authors reported the impact of amine functionalized

CNC loading on mechanical properties, the interior

morphology, and gelation time. Elastic modulus for the

loading of CNC equal to 0.80 wt% showed a maximum.

The resulting hydrogel depicted pH-responsive

properties and excellent overall stability under

conditions like the human body. The hydrogel also

showed extremely well self-repairing behavior under

acidic conditions.

Lenfant et al. [32]. CNC only hydrogels In the case of the addition of only 4 wt% CNC, for

duration of 2hrs, the storage modulus rose to 1390 Pa in

the case of CaCl2, but for NaCl, this value only increased

to 443 Pa. At 6 wt%, these values were 3156 Pa for

CaCl2, in comparison to 1453 Pa for NaCl.

Khabilulin et al. [70]. CNC hydrogel decorated with

graphene quantum dots

Authors showed that the formation of physically cross-

linked gel that can have varying levels of mechanical

properties. As samples showed shear thinning, they

were a good candidate for 3-D printing as well.

Tang et al. [71]. Cellulose nanocrystal (CNC) and

sodium alginate (SA)

Authors witnessed that CNCs through performing as

macro-cross linkers, improved structural robustness and

of the matrix SA. In their work, the hydrogels displayed

uniform chemical macroscopic structures and could

efficiently self-heal at room temperature within the span

of several hours.

16

Wu et al. [72]. CNC only hydrogels Authors chose two ensembles of CNCs with different

aspect ratios were investigated. CNC, with a higher

aspect ratio, morphed into a biphasic (containing two

phases) state and formed a hydrogel at lower CNC

concentration compared to shorter CNCs. Complex

viscosity trend did not superimpose on shear viscosity

values of both CNCs, which the resulting observation

can reveal the formation of liquid crystal domains.

Zhou et al. [73]. CNC- PAM (polyacrylamide) hydrogel Authors observed a good dispersion of CNCs that

causes an improvement in the storage modulus,

compression strength, and elastic modulus of

nanocomposite. Among the CNC loadings employed, a

loading of 6.7 wt% led to maximum mechanical

properties for hydrogels.

Mckee et al. [74]. CNC only hydrogels Variation of CNC concentration cause changes in

viscoelastic storage modulus to vary between 1.0 to 75

Pa when CNC concentration varied between 0 to 3.5

wt% in vicinity of 1.0 wt% MC (methylcellulose). At

higher temperatures (60 °C), a gel was fabricated that

had higher storage modulus values (i.e., 110 vs 900 Pa)

with identical loadings of CNC and MC. For the current

set of constituents, cross links were the reason behind

higher mechanical properties of the gels.

Lewis et al. [42] CNC hydrogels Reports showed CNC in water and other polar solvents

can gel due to repeated freeze-thaw cycles.

Characterization wad done with Rheological

measurements to show gelation process. Gel strength

was looked at using rheology as a function of freeze-

thaw cycles.

Oechsle et al. [75]. CNC hydrogels It was conjectured that CO2-responsive CNC hydrogels

can potentially be used in applications that requires

change in suspension properties as a response to

17

external factors. CO2 acted as an ingredient that

switched its charges with modifications in pH.

Talantikite et al. [76]. CNC-xyloglucan (XG) The influence of molar mass of XG on mechanical

properties of solutions having 10 g/L CNC was

inspected. At lower molar mass of XG, the solution still

behaved viscoelastically. At higher molar mass, system

experienced an increase in magnitude of both

viscoelastic storage modulus and loss modulus. System

in these cases showed ability to reach gel point.

Zhou et al. [73]. PAM (polyacrylamide) -CNC The result showed that CNCs that own lower aspect

ratios can facilitate the formation of PAM-CNC

nanocomposite hydrogel. The onset of gelation of CNC-

free and CNC-embedded system calculated was 4.1±0.4

and 2.5±0.3 min, respectively. The authors observed

that sol-gel transition due to presence of CNCs was

expedited.

Hou et al. [77]. CNC/poly (ethylene glycol) diacrylate

hydrogel

The initial gel rheology assessment that gel showed

shear-thinning and viscoelastic behavior. The

mechanical properties showed a surging trend in the

vicinity of CNCs.

Shafiei-Sabet et al. [78]. CNC only hydrogels When the concentration was diminished to 0.5 wt%, the

G’ and G’’ curve shifted to higher values as the angular

frequency increased, showing that the behavior started

to morph from solid-like to gel-like. In the case of 0.25

wt% S-CNC suspension, a behavior similar to gel

materials was observed, where G’ and G’’ were strongly

reliant on frequency of oscillation. For all CNC

suspensions, both storage modulus and loss modulus

depicted a strong reliance on the frequency of tests, and

G’ was slightly higher than G’’.

Wang et al. [79]. surface-modified CNCs-bis(acyl)

phosphane oxide derivative

Hydrogels printed using a 3-D printer showed good

swelling behavior and improved mechanical properties.

18

Huang et al. [80]. Carboxymethyl chitosan (CMC) and

dialdehyde cellulose nanocrystals (DACNC) network.

It was also shown that hydrogel could repeatedly be

extended longitudinally to 4 times its original length and

had a tensile strength of 244 kPa. The composite also

had complete healing when compressed by 90% and

showed compressive strength up to 8 MPa. Moreover,

the stretched hydrogel could recover 81.3 % of the

dissipated energy without any external changes.

Rao et al. [81]. CNC-Xanthan-Chitosan Using hydrogen bond interactions and electrostatic

forces hydrogels were made. A surge in mechanical

properties was observed when CNC loading changed

from 2 to 10 wt%.

Ghorbani et al. [82]. CNC-collagen hydrogel

Concentration ratio of CNC to collagen and its effect on

pore shape, swelling index and mechanical properties

was inspected.

Han et al. [83]. CNC-PVA The compression and dynamic oscillation measurement

showed that the incorporation of CNCs significantly

improved the mechanical properties. The compression

stress of CNP-PB-PVA hydrogel was 21 times higher

than pure PVA hydrogel. CNC acted as crosslinkers that

could weave interiors of the 3-D network of hydrogels

both physically and chemically.

Hu et al. [49]. Hydroxyethylcellulose (HEC)-

hydroxypropyl guar (HPG)-locust bean gum (LBG)-

CNC

By adding 3 wt% CNC and 0.2 wt% polysaccharide in

their experiment, storage modulus at the plateau level

changed from 32 to 100 Pa. In the same manner for

CNC-COOH, range of storage modulus variations were

the same.

García-Astrain et al. [84]. Gelatin-CNC Swelling index and rheological characterization were

employed to depict effect of functionalization and

formation of the network in the hydrogel.

19

The above review indicates that CNC-polymer hydrogels have tunable properties where swelling,

rheological properties, enhancement in mechanical properties, functionalization, injectability have

been extensively analyzed. The existing studies demonstrate the improvement in the mechanical

properties of the CNC-composite hydrogels. However, there are other characteristics, such as

stability under gravity, and the self-healing ability of these hydrogels that have received less

attention. Analyzing such parameters, which are critical in many applications, including 3-D printing

[85], forms the basis of the present thesis.

1-3 Methodology

In this thesis, confocal laser scanning microscopy (CLSM), which has been successfully employed

in medical, geological, and biological areas, would be put forth as an alternative investigative tool

for CNC hydrogel structure analysis [86]. Moreover, we use rheometry to monitor the mechanical

behavior of gels.

1-3-1 CNCs confocal laser scanning microscopy

CLSM has good capability towards the visualization of hydrogels due to its non-invasive

characteristics in monitoring the evolving structure of gels. As hydrogels have water, a major

advantage of CLSM is monitoring the bulk structure of the hydrogel without interfering with it prior

to any investigation. Dehydration that usually comes into play when observing the structure using

SEM unavoidably alters the morphology and structure of the hydrogel. Pictographs of the hydrogel

bulk structure, using CLSM, can be taken at pre-defined time steps, without altering the structure

such as freeze-fracturing.

For instance, in work by Fergg et al. [87] on PVA hydrogels, CLSM imaging of PVA bulk hydrogel

structure showed continuous, 3-D intertwining structure that stemmed from prior phase separation

during the freezing period. In the sample, no porosity gradient or any preferred orientation was

observed, so the sample was isotropic. Considering volume fractions of 18.33, 15.04, and 9.20 %,

the hydrogels displayed a uniform pore size of 2-3, 4-5, and 6-7 µm, respectively. Moreover, it was

observed that raising the PVA concentration caused to an augmentation in hydrogel tortuosity and

finally inhibited the mobility of the tracer material. After inspection of morphology of the surface of

20

the PVA structure, it was found that PVA macromolecules consisted of randomly distributed

domains.

In a separate study by Savina et al. [88], CLSM was employed to gain knowledge of the whereabouts

of polymer grafting inside the hydrogel. In this case, CLSM had an advantage over SEM, as it could

observe the structure internally, as opposed to SEM. The CLSM pictographs showed an uneven

distribution of grafted polymer inside the porous gel as the initiator of the chemical reaction between

grafting polymer, and the gel was insoluble. The insolubility of the initiator pushed the chemical to

mainly deposit locally on the surface.

In the study by Koyano et al. [89], using CLSM, the surface of PVA/chitosan was examined, and the

result showed that chitosan had an island type structure, while being blended on top of the hydrogel.

Recently, CLSM has also found applicability in nano-ink technology and 3-D printing of hydrogels

as well. For instance, in a study by Highly et al. [90], the hydrogel injectability and the state of the

hydrogel after processing was monitored by CLSM. For instance, in Figure 1.7, injected hydrogel

does not mix with the supporting material. In Figure 1.7C, ink filament has been labeled Rhodamine-

B, while the supporting material has been labeled by Fluorescein.

21

Figure 1.7 A) attachment of adamantane and β‐cyclodextrin to hyaluronic acid. B) illustration of the extrusion process

for ink (red) into the support gel (designated as green) C). The capability of CLSM in showing the results (reprinted from

reference [90])

In Figure 1.8 and Figure 1.9, the capability of CLSM towards monitoring proteins and fat droplets

phase separation. The system shows phase separation towards the adjustment of the pH of the system.

There are many more examples in the literature that demonstrate the capability of CLSM for

monitoring hydrogels in their native state [91-94], but here, due to the scope of the research, they

will not be mentioned.

22

Figure 1.8 Utility of confocal scanning laser microscopy (CSLM) images in capturing gelation of full-fat milk

containing Nile blue. Numbers signify minutes after dye addition. Scale bar = 25 µm (reprinted from reference [95])

Figure 1.9 Capturing phase separation of glucono-δ-lactone (GDL)-induced gelation of skim milk using CLSM. Bright

areas are protein. Numbers signify minutes after the introduction of the rennet (phase separation trigger). Scale bar = 25

µm. (reprinted from reference [95])

1-3-2 Gel healing monitoring using rheometry and CLSM

23

The transient trait of entanglements makes the physical hydrogels in a position to tolerate the

externally imposed stress, through the mechanism of displacement, rearrangements, and re-

orientation of the junctions. This specific trait of physical gels cannot be found in chemical gels [96].

After physical or chemical damage, self-healing goes through mechanisms that get triggered, either

autonomously or by other factors such as light, sound, and pH. Self-healing specifically becomes

important, if maintaining a certain strength estimated before processing for the hydrogel is to be met

after the processing is complete. For instance, hydrogels can go through several cycles of shearing

before reaching the final stage, and returning to original strength is important. Also, hydrogels might

go through a few cycles of extensional, compressional or shear cycles, maintaining certain

mechanical properties during or between each cycle becomes important. Moreover, knowledge of

the recovery rate at the molecular and macro-level is also equally important for designing purposes.

In the report by Yu et al. [97], a ring structure with pore sizes ranging from nanoscale to microscale

was obtained first through an organogel system. The gelation and breakage could take place by a

simple shaking and resting intervals. Therefore, the system under study had a self-healing property

that could be triggered through shear or any other deformation. In another study by Yuan et al. [98],

Metallo-supramolecular gels were fabricated using transition metal ions and/or lanthanide ions. Due

to unique structure of the hydrogel, the gel depicted a self-healing ability in a facile manner. In

another study by Herbst et al. [99], rheological characterization for unearthing healing ability of

supramolecular poly(isobutylene)s (PIBs) materials and fluorescence recovery after photobleaching

(FRAP) to examine the samples at rest were employed.

Figure 1.10 visual self-healing experiment (a) cut segmented parts, (b) segmented parts were just brought into contact,

(c) crack partially healed after passage of 24 h and (d) thoroughly healed crack after passage of 48 h (reprinted from

reference [99])

Both technique displayed that the materials shows self-healing properties (see Figure 1.10).

However, contrary to rheology experiments, FRAP could probe the samples at rest, without altering

the structure via deformations. The results of this study showed the ability of both techniques that

24

probe microstructure of hydrogel, i.e. rheology and FRAP, in comprehending the healing of

hydrogels and mechanisms. Readers are encouraged to read reference [85] for through information

on gel self healing and mechanism that can improve it. Aligned with the mentioned studies, in this

thesis, both rheology and FRAP will be employed to study the dynamics of CNC suspension and

gels.

1-3-3 Fiber orientation models

Distribution of filler orientation dictates the final properties of the finished product. One of the

processing routes for employing CNC or CNC composite hydrogels is through injection or going

through a 3-D printing apparatus. Therefore, it seems imperative to assess the level of CNC

orientation during such flow, due to the impact it might wield on the finished products’ overall

mechanical properties. In order to predict the state of fiber distribution inside a matrix (usually a

polymer), model equations are developed to use theories restricted to dilute and semi-dilute regimes

[100] [101]. The models used are based on the protocol that originally developed and is known as

Jeffery’s equation [102] for an isolated freely rotating fiber in a Newtonian fluid. The models are

usually geared towards taking into consideration hydrodynamic interactions amongst fibers. The

succession of these interactions causes very tiny alterations in fiber orientation states. The tiny

changes to an orientation state of one fiber caused by its neighboring particle through hydrodynamic

interactions being accounted for with a rotary diffusion process in the developed models [100, 103-

105]. This rotary diffusion can neither dislocate a fiber horizontally, nor it can affect its length. The

most commonly used rotational diffusion model is the standard Folgar-Tucker model [100]. The

model builds on Jeffery’s rotation rate of fibers, with the anisotropic diffusive process that is

dependent on the state of orientation. When there is isotropic diffusion occurring in the system, the

orientation state of fibers evolves over time, and the possibility of alterations is equal in all directions,

and it assumes a scalar value. However, it has been shown that this diffusivity can decrease fiber

alignment unjustly, through compensating for factors that are also responsible for fiber dispersions.

Another drawback of isotropic diffusivity is that it cannot predict and tune all components of second-

order orientation moments at the same time [106, 107]. Therefore, many researchers [107, 108] have

justly proposed the idea of diffusivity parameter in the Folgar-Tucker equation that is also dependent

on orientation state and is not isotropic.

25

Based on theory of Jeffery, rheology of suspension of fibers can be modelled [102]. Considering the

unit vector p, the main axis involved in the evolution of the state of the fibers can be expressed as

follow:

�� = 𝑊𝑝 + 𝜆(𝐸𝑝 − 𝐸: 𝑝𝑝𝑝)

1-1

Where, 𝜆 = (𝑟𝑝2 − 1)/ (𝑟𝑝

2 + 1) is the shape factor of the fibers that is dependent on its aspect ratio,

�� is the material derivative of p, and 𝑟𝑝 = 𝐿𝑑⁄ is aspect ratio assigned to the fiber. For very thin and

long fibers, 𝜆 goes to 1, 𝑊 is the vorticity tensor, and 𝐸 is the strain rate tensor.

As discussed earlier, the rotation motion of rod particles can be well described by p. p is a vector

aligned in the direction of the main axis of particles. Second-order and fourth-order tensors

introduced by Advani and Tucker[109] can be described as follow:

𝑎2 ⟺ 𝑎𝑖𝑗 = ∫ 𝑝𝑖 𝑝𝑗𝜓(𝑝)𝑑𝑝

1-2

𝑎4 ⟺ 𝑎𝑖𝑗𝑘𝑙 = ∫ 𝑝𝑖 𝑝𝑗𝑝𝑘𝑝𝑙𝜓(𝑝)𝑑𝑝

1-3

Where 𝑎2 is a second-order tensor with trace equal to 1. For the 4th order tensor one needs to use an

approximation.

Derivative of 𝑎2 can be obtained through rotation tensor introduced by Jeffery [102]. For dilute state

of suspension of particles embedded in a Newtonian fluid under shear and in low Reynold number

following equation can be used:

��2 =𝐷𝑎2

𝐷𝑡=

1

2(𝑊𝑎2 − 𝑎2𝑊) +

1

2𝜆(𝐸𝑎2 + 𝑎2𝐸 − 2𝐸: 𝑎4)

1-4

In this equation, Ω is related to the aspect ratio as follows. Other parameters that need to be defined

are:

26

𝐸 = (𝐾 + 𝐾𝑡)

1-5

W = (𝐾𝑡 − 𝐾)

1-6

In this equation 𝐾𝑡 is the velocity gradient tensor. It is assumed that particle have no interaction with

the particles in their vicinity, this model can only give accurate results for dilute suspension of

axisymmetric particles.

For suspensions that are not dilute, FT model (Folgar-Tucker) considers the interaction between rod

particles using the following equation:

��2 =𝐷𝑎2

𝐷𝑡=

1

2(𝑊𝑎2 − 𝑎2𝑊)

+1

2𝜆(𝐸𝑎2 + 𝑎2𝐸 − 2𝐸: 𝑎4)

+2𝐶𝐼 ��(𝐼 − 3𝑎2)

1-7

In this equation �� effective shear rate and 𝐶𝐼 is a semi-empirical constant and effective shear rate

is �� = √1

2𝐸: 𝐸 .Bay et al. [110] suggested that the relationship between concentration and aspect

ratio be considered for estimation of 𝐶𝐼:

𝐶𝐼 = 0.0184 exp (−0.7148 𝜙 𝑟)

1-8

In this equation, 𝜙 and 𝑟 are volume fraction and aspect ratio of particles, respectively. There are

many approximations for 4th order tensor, which the simplest one is a second-order estimation as

follow:

𝑎4𝑞

⟺ 𝑎𝑖𝑗𝑘𝑙𝑞

= 𝑎𝑖𝑗𝑎𝑘𝑙

1-9

For obtaining the transient stress after application of shear, Jeffery hand and Lipscomb suggested

the following equation:

σ = −𝑃𝐼 + 𝜂𝑚�� + 𝜂𝑚𝜙{𝜇1�� + 𝜇2𝛾: 𝑎4}

1-10

27

Where, 𝜂𝑚 is the viscosity of the matrix and 𝜇1 and 𝜇2 are rheological constants that act as fitting

parameters in the model. Usually 𝜇1 is assumed to be equal of 2.

1-3-4 Monitoring the mechanical behavior of gels using rheometry

In in-vivo tissue engineering applications, matching of mechanical properties of cells plus injected

scaffold with the natural tissue at the location of implantation are important. Therefore, due to this

complexity, studying the mechanical behavior of gels through rheometry deemed necessary.

Previously, few works have studied the effect of salt on CNC suspension rheology and has been

studied in the linear viscoelastic zone. For instance, Lenfant et al. [32] studied the linear viscoelastic

response of CNC and electrically stabilized CNC in the presence of sodium and calcium ions. It was

shown that at moderate ionic strength values, CNC suspension forms agglomerations. Among the

tested salts calcium chloride (CaCl2) showed a greater effect than sodium chloride (NaCl) on both

properties of flow-related shear (and viscoelasticity), due to the fabrication of a stronger network.

Due to higher repulsion between individual particles, electrically stabilized CNC could tolerate

increased amounts of salt before the aggregation occurrence. In another study, Shafiei-Sabet et al.

[111] showed that for isotropic CNC suspensions, augmentation of ionic strength via the introduction

of salt till 5mM of NaCl decreases the viscosity of the system, due to weakening of the electro-

viscous effects.

Innately, mechanical traits of the hydrogel are crucial for providing mechanical support to the

surrounding tissue, when load-bearing applications are involved such as bone and to provide an

adapted environment for the cohabiting cells. It has been conjectured that the stiffness of the hydrogel

and stresses originated from the cell surrounding environment impacts the cell’s fate, in particular

for the aim of the differentiation of stem cells [112, 113]. In a recent study [114], it was shown that

in addition to the mechanical properties such as stiffness, relaxation and retardation time scale of

hydrogel also influences stem cells fate. Most tissue due to having cells, Extracellular matrices and

high percentage of water easily operate under nonlinear regime. Hence, for tissue engineering

studying viscoelastic behavior is key for successful hydrogel implantation as a scaffold inside the

body. Even for rigid tissues, like bone [115], viscoelastic traits are impactful, especially strain rates

that are low and in bound of frequency ranges suitable to a normal body. Therefore, to know how

28

flawlessly the beneath scaffold material mimics the targeted tissue viscoelastic nature such as their

time and frequency behavior nonlinear and linear rheology should be assessed in details.

Due to their rigid nature (spindle in shape), CNC acts as an effective filler in hydrogel

nanocomposites. However, their inability to create physical anchors to one another (contrary to CNTs

or other flexible nanofillers) causes CNC to produce gels with low mechanical properties. Few

published papers that show the extent of CNC pure hydrogels mechanical properties limitation has

been reflected in Figure 1.11. CNCs above 10 wt% (Corroborated by rheological measurements)

can enter the gelled phase automatically. Alternatively, CNC in the presence of external electrolytes,

changing pH, or getting cross-linked with multivalent ions or chemical bonds can enter the gelation

phase as well. These methods of gelation by destabilization of the suspension cause the gelation to

happen at much lower CNC concentrations. Reports in the literature shows that gelation points

change if one changes solution condition, add polymers or alter the surface chemistry of the particles

(the driving force behind gelation in the case of non-adsorbing polymers is depletion mechanism).

Chau et al. [48] have shown that the method of gelation in the presence of external electrolytes is

decreasing in double-layer thickness that is expected by DLVO theory. The addition of salts limits

the electrostatic repulsion interactions and causes the van der Waals interaction to become more

dominant, and this causes gelation. Through rheological measurement, it has been shown that gel

strength can go through alteration depending on the ions added into the mix (i.e., changing charge

values of cation and radii of physical ion bonds established between CNC can get stronger). In work

by Ureña-Benavides et al. [46], a rheological study was reported on pure CNC gel. In their report,

the independency of storage modulus towards CNC concentration happened around 14.5 and 17.3

vol% of CNCs. The transition to a pure gel happened between 12.6 and 14.5 vol% of CNCs. Large

amplitude shear tests were done on the 14.5 vol% sample at 20% strain, to shear the structure of

suspension above their linear regime. The storage modulus measured during LAOS was almost 50%

smaller than the one obtained under 1.5% strain. After resting for 40 mins, the measured modulus

was only 5 percent lower and after 80 min the recovery was found to be complete.

Alteration of pH similarly causes the gelation point to change. Way et al. [47]., though

functionalization of CNC with either carboxyl or amine groups, showed that gelation could become

pH-responsive. For instance, decorating the CNC surface with amine groups causes the gel to attain

positive, neutral, or negative charge depending on the pH of the medium. Therefore, driving CNC

29

towards each other or repelling from one another is possible through this route. For the case of

functionalization using carboxylic acid, storage modulus as a function of pH was reported to change

three order magnitude depending on the pH.

Lewis et al. [43] showed that through temperature annealing of the CNC suspension at high

temperature, one could drive CNC suspension state into gelation. They attributed the finding to CNC

desulfation that happens because of keeping the suspension in an autoclave under heated conditions.

Desulfation, in their case, means less electrostatic repulsion and, therefore, the higher affinity of

CNC towards one another. Dorris and Gray [52] similarly found that CNC desulfation was the only

way to drive gelation in a dilute suspension of CNC in the presence of glycerol and water.

Polymer adsorption to the surface of CNCs is another route to gelify CNC suspension gel. Hu et

al.[49] showed that in suspension, CNC with a 3 wt% adsorption of nonionic polysaccharides causes

gel formation. In their case, gel formation happens through polymeric chains adsorption, which

causes the effective volume fraction to increase and shift the gelation point to lower CNC

concentrations.

Figure 1.11 depicts storage modulus magnitudes of CNC hydrogel and CNC composite hydrogels

on a double logarithmic scale for various references mentioned in the graph. Polymers so far

mentioned in the literature as a matrix for CNC composite hydrogels are poly(vinyl alcohol) (PVA),

polyacrylamide, poly(meth)acrylates, poly(ethylene glycol), polysaccharides, and nature sourced

polymers such as alginate and gelatin. In these hydrogels, the CNC amount varied between 0.1-19.6

wt% based on total gel weight. Gel storage modulus in these cases showed sharp increases with the

addition of CNCs, with the maximum mechanical reinforcement enhancement reported in this thesis

152 kPa.

30

Figure 1.11 Storage modulus reported from various references reflecting the effect of CNC on the reinforcement of

different matrices, numbers associated with reference numbers [43, 46-49, 59, 72, 83, 116-123] sketched on double

logarithmic axes.

Figure 1.11 depicts power of rheology to gain information about the structure and strength of gels.

In fact, an extensive range of materials, can be studied with rheology, and particularly rheology can

be used to observe and control their viscoelastic properties. Small amplitude oscillatory shear tests

(SAOS) are a useful tool to probe the linear viscoelastic properties; however, large amplitude

oscillatory shear tests (LAOS) are as important to assess efficiently nonlinear viscoelastic properties

[124]. When deformations are small, SAOS’s tests are generally applicable, and this way, the

material rheological response stays in the linear regime. On the other hand, upon increasing the

deformation, LAOS’s tests can be a helpful tool to characterize the nonlinear response. Figure 1.12

shows two zones in which the strain is has an ascending trend.

31

Figure 1.12 Strain sweep test, strain changing g from small to large values [124]

The following set of equation can be used to define applied deformation and shear rate in the dynamic

oscillatory shear test [124]:

𝛾 = 𝛾0 𝑠𝑖𝑛(𝜔𝑡)

1-11

�� = ��0𝜔 𝑐𝑜𝑠 (𝜔𝑡) 1-12

Shear stress response can be:

𝜏𝑦𝑥 = 𝐺′(𝜔)𝛾0 𝑠𝑖𝑛(𝜔𝑡) + 𝐺′′(𝜔)𝛾0 𝑐𝑜𝑠(𝜔𝑡)

1-13

To have a better understanding of the meaning of these linear viscoelastic properties, it is useful to

repeat that for a material with no viscous component 𝐺′ is in value equal to the constant shear

modulus 𝐺 and loss modulus is zero. In other words, storage modulus of the material in this scenario

gives information about the elastic character of the fluid. Quite similarly, for a completely Newtonian

fluid, 𝜂′ is identical to the viscosity 𝜇 and 𝜂′′ is zero. 𝐺′′ (= 𝜂′𝜔) is known as loss modulus, and it

depicts the viscous behavior of the fluid or the amount wasted energy in each cycle of deformation.

In 1958, W. P. Cox and E. H. Merz [125] suggested the following relation, which empirically they

found for a range of solutions and melts of many unlinked and unfilled polymers (𝜂(��) = |𝜂∗(𝜔)|.

This relation applies if the values of both �� and 𝜔 are equal in size. It is also noteworthy to mention

that complex viscosity can also be obtained using |𝜂∗| = √(𝜂′)2 + (𝜂′′)2.

32

1-3-5 Large amplitude oscillatory shear test (LAOS)

LAOS is a technique with expanding popularity among researchers [124, 126]. This approach is

destined to mark the onset of nonlinearities in complex materials. The LAOS tests involve oscillation

cycles at multiple strain amplitudes.

In the LAOS region, the sinusoidal input strain waveform is translated to a non-sinusoidal stress

response. There are various approaches to analyze the non-sinusoidal stress response, such as Fourier

transform rheology (FT- rheology) [127] and stress decomposition [128] methods. The shear stress

(𝜎) can be inscribed as in-phase and out-of-phase components of a time-domain Fourier series of

odd harmonics[129] being in steady-state condition for an oscillatory input strain (𝛾(𝑡) =

𝛾0sin (𝜔𝑡)):

𝝈(𝒕) = 𝜸𝟎 ∑ [𝑮𝒏′ (𝝎, 𝜸𝟎) 𝐬𝐢𝐧 𝒏𝝎𝒕𝑵

𝒏=𝟏 + 𝑮′′𝒏(𝝎, 𝜸𝟎) 𝐜𝐨𝐬 𝒏𝝎𝒕] 1-14

In the above equation, 𝛾0 is strain amplitude, and 𝐺𝑛′ and 𝐺𝑛

′′ are amplitudes of n harmonics with

frequencies (nω). In the linear viscoelastic framework, the output stress waveform is the only

function of the first harmonic coefficients, 𝑛 = 1. The emergence of higher harmonics in the

resulting stress waveform depicts the appearance of nonlinear viscoelastic response, meaning that

the stress signal cannot be displayed by a simple sinusoidal waveform any longer. Furthermore,

𝐺′and 𝐺" lose their physical meaning in the nonlinear region, meaning another technique should be

implemented to explain the output stress signal.

FT- rheology is developed based on a sophisticated mathematical framework that is a powerful

technique to spot nonlinearities and higher-order harmonics in the stress waveform. However, it

cannot give a clear physical interpretation of higher-order harmonics and the resulting nonlinear

behaviors[130]. So, this method is insufficient to describe the material response. In 2008, Ewoldt et

al.[130] proposed novel measures based on the stress decomposition method introduced by Cho et

al.[128] in 2005 to give meaning to LAOS results.

Based on symmetric arguments proposed by Cho et al.,[128], the generic nonlinear stress response

(𝜎(𝑡)) can be decomposed into superposition of elastic and viscous stresses as below:

1) elastic stress component (𝜎′) as an odd function of normalized strain (𝑥(𝑡) = 𝛾(𝑡)

𝛾0 ),

33

2) viscous stress component (𝜎′′) as an odd function of the normalized strain rate (𝑦(𝑡) =

�� (𝑡)

��0). Thus, the total resulting stress can be described as the following:

𝜎(𝑡) = 𝜎′(𝑡) + 𝜎′′(𝑡). 1-15

Afterward, Ewoldt et al.[130] suggested a polynomial regression fit to the elastic (𝜎′) and viscous

(𝜎′′) lines. In their work, they argued the limitation of different polynomial basis functions, such as

Jacobi, Laguerre, Hermite, Chebyshev of the first and second kind, and Legendre. Considering the

mathematical and physical limitations, e.g., elastic(𝜎′) and viscous (𝜎′′) stresses are orthogonal over

a finite domain, they proposed that the set of Chebyshev polynomials of the first kind is the best

choice for fitting the output stress contributions. Then, the authors established a physical

interpretation of nonlinear viscoelasticity using Chebyshev coefficients.

Based on this method, a series of Chebyshev polynomials of the first kind in the orthogonal

space made up of the input strain and strain-rate can be used to represent the elastic 𝜎′ and viscous

𝜎′′stress components via the following equations:

𝜎′(𝑥: 𝜔, 𝛾0) = 𝛾0 ∑ 𝑒𝑛(𝜔, 𝛾0) 𝑇𝑛(𝑥), 1-16

𝜎′′(𝑥: 𝜔, 𝛾0) = 𝛾0 ∑ 𝑣𝑛(𝜔, 𝛾0) 𝑇𝑛(𝑦) 1-17

where �� =𝑥

𝛾0=

𝛾

𝛾0 and �� =

𝑦

𝛾0=

𝛾

��0

depicts the normalized version of strain and strain-rate, and 𝑇𝑛

symbolizes Chebyshev polynomials. ‘‘𝑒’’ and ‘‘𝑣’’ are elastic and viscous contributions and have units

of modulus (Pa) and viscosity (Pa.s-1), respectively.

The criteria for specification of the physical interpretation of the nonlinearity based on “𝑒”

and “𝑣” is defining the concavity of 𝜎′and 𝜎′′. As the magnitude of each Chebyshev coefficient

decays monotonically by increasing “𝑛”, the third-order Chebyshev coefficients (𝑒3 and 𝑣3)

determine the concavity of the elastic and viscous stress curves. According to these coefficients, the

following intra-cycle nonlinear behaviors can be observed: strain-stiffening (𝑒3 > 0), strain-

softening (𝑒3 < 0), shear-thickening (𝑣3 > 0) and shear-thinning (𝑣3 < 0)[130]. Moreover, the nth-

order Chebyshev coefficient and Fourier coefficients can be related to each other via the following

equations [130]:

𝑒𝑛 = 𝐺𝑛′ (−1)(𝑛−1)/2 1-18

34

𝑣1 =𝐺𝑛

′′

𝜔= 𝜂𝑛

′ 1-19

In the nonlinear regime, the measured dynamic moduli (𝐺1′and 𝐺1

′′) do not have a clear physical

meaning. Hence, using ‘’𝑒’’ and ‘’𝑣’’, Ewoldt et al.[130] defined local viscoelastic moduli and

viscosities to interpret the distorted stress signal. Hence, comparing the local viscoelastic moduli

(i.e., large-strain modulus (𝜎

𝛾|

𝛾=±𝛾0

≡ 𝐺𝐿′ ) and minimum-strain modulus (

𝑑𝜎

𝑑𝛾|

𝛾=0≡ 𝐺𝑀

′ )) can assist

to interpret intra-cycle elastic nonlinear behavior [130]. It is noted that both 𝐺𝑀′ and 𝐺𝐿

′ converge to

linear elastic modulus in the linear viscoelastic region, i.e., 𝐺𝑀′ =𝐺𝐿

′ =𝐺1′=𝐺′. These elastic measures

have been used by Ewoldt et al. [130, 131] to develop a dimensionless index for interpretation of

intra-cycle elastic nonlinearity defined as:

S≡𝐺𝐿

′ −𝐺𝑀′

𝐺𝐿′

1-20

S (strain stiffening ratio) value equal to 0 corresponds to linear viscoelastic response, a positive S

indicates intra-cycle strain-stiffening behavior, and a negative S is indicative of intra-cycle strain-

softening. Like the above-mentioned elastic measures, viscous parameters have been introduced to

characterize intra-cycle viscous nonlinearity. In this context, a set of local dynamic viscosities have

been defined as minimum-rate dynamic viscosity 𝑑𝜎

𝑑��|

��=0≡ 𝜂′𝑀 and large-rate dynamic viscosity

𝜎

��|

��=±��0

≡ 𝜂′𝐿[130, 131]. Similar to the elastic measures, in the linear regime, dynamic viscosities

converge to the linear real viscosity value 𝜂′ =𝐺"

𝜔, i.e., η'L=η'M=η'. The dimensionless index for

dissipative (viscous) intra-cycle nonlinearity has been proposed as:

T≡𝜂𝐿

′ −𝜂𝑀′

𝜂𝐿′

1-21

T=0 signifies linearity, T>0 implies intra-cycle shear-thickening, and T<0 corresponds to intra-cycle

shear-thinning behavior. It should be born in mind that there are other methods and approaches, such

as the sequence of physical processes [132] and intrinsic nonlinearity [133, 134], which researchers

used to interpret nonlinear data. Compared to the mentioned methods (e.g., FT- rheology), the

method that I used in this work provides us the physical interpretation of nonlinearity with the aid of

unambiguous material measures, which quantify nonlinear elastic and viscous behavior,

simultaneously. Thus, this method provides us with more substantial information regarding the

35

mechanism governing the microstructural changes under LAOS flow.In the process of analyzing 𝑆

and 𝑇, physical mechanisms should be considered to avoid any misinterpretation. These intra-cycle

nonlinearities can be defined by the set of following formulas that relate coefficients to 𝑆 and 𝑇

indirectly:

and:

𝜂𝑀′ ≡

𝑑𝜎

𝑑��≈

1

𝜔∑ 𝑛𝐺𝑛

′′(−1)𝑛−1

2

= 𝑣1 − 3𝑣3 + 5𝑣5 − 7𝑣7 + ⋯

1-24

𝜂𝐿′ ≡

𝜎

��≈

1

𝜔∑ 𝐺𝑛

′′ = 𝑣1 + 𝑣3 + 𝑣5 + 𝑣7 + ⋯

1-25

1-4 Problem statement

CNC-polymer hydrogels due to their adjustable properties can easily fit into design of new materials

protocol. Despite many existing studies, a few CNC-based hydrogel properties, including gel healing

rates, relations among gel mechanical properties and CNC orientations, and nonlinear rheological

properties of gels, have received attention. Traits mentioned above, are key design parameters in the

formulation of hydrogels. In summary, CNC has the following characteristics:

• CNC is the building block for CNC-based hydrogels

• CNC is biocompatible therefore its does not cause an immune reaction from the body

• CNC-based hydrogels have tunable mechanical properties

• CNC, through chemical modification, can be bio-resorbable.

𝐺𝑀′ ≡

𝑑𝜎

𝑑𝛾≈ ∑ 𝑛𝐺𝑛

′ = 𝑒1 − 3𝑒3 + 5𝑒5 − 7𝑒7 + ⋯

1-22

𝐺𝐿′ ≡

𝜎

𝛾≈ ∑ 𝐺𝑛

′ (−1)𝑛−1

2 = 𝑒1 + 𝑒3 + 𝑒5 + 𝑒7 + ⋯

1-23

36

This thesis provides information on the CNC-based gels from the nano level and the connection

between the macro properties to microstructures. The use of salts and polymers in tuning CNC-based

hydrogels are addressed. Polymer and salt addition in CNC-based hydrogel can strengthen the

mechanical properties of gel and also can avoid the erratic nature of gel formation. Equipment and

techniques that are used to control and characterize the CNC gelation include confocal laser scanning

microscopy (CLSM), scanning electron microscopy (SEM), linear and nonlinear rheology, dynamic

light scattering, zeta potential, compression tests, and computer simulations

This dissertation investigates; (i) the gel porosity as a function of PVA, CNC and salt using SEM

imaging, (ii) the gel mechanical properties using linear and non-linear rheology and compression

tests, and (iii) the self-healing ability of CNC gel through measuring the particle diffusions in the gel

media by CLSM monitoring. The goal here is to fabricate and characterize a hydrogel ready for

future use in important fields such as tissue engineering applications, including scaffold fabrication.

1-5 Dissertation outline

This dissertation has been prepared in the paper-based format and comprised of seven chapters. The

first one is the introduction chapter in which the problem is stated, and the goals of the dissertation

are discussed. The next five chapters, Chapters 2 to 6, comprise the main contents of the dissertation.

Each chapter is prepared as a modified version of a manuscript and is published or submitted for

publication in peer-reviewed journals.

Chapter 2 discusses the gelation of CNC monitored with CLSM. Moreover, methodology and

procedure for the determination of pore size and effect of CNC and salt loadings on porosity would

be outlined. Evaluation of microstructure shows cluster formation with NaCl addition into the

system. This chapter is the same version of the manuscript entitled “Colloidal behavior of cellulose

nanocrystals in the presence of sodium chloride” published in Chemistry Select1.

The highlight of Chapter 3 is a comparison of MgCl2 as a divalent ion compared to monovalent NaCl

on gelation. Discussion about gravity effects on gel stability is another subject of this chapter.

Molecular dynamic simulations are conducted to understand the magnitude of forces in play in CNC

1 Moud, A. A.; Arjmand, M.; Yan, N.; Nezhad, A. S.; Hejazi, S. H., Colloidal behavior of cellulose nanocrystals in

presence of sodium chloride. ChemistrySelect 2018, 3 (17), 4969-4978.

37

suspensions. This chapter is a modified version of the manuscript entitled “Cellulose nanocrystal

structure in the presence of salts” published in Cellulose2.

In Chapter 4, the effect of CNC and salt concentrations on the rheology of CNC hydrogels are

discussed. Both linear and non-linear rheological properties of hydrogels, as a function of CNC and

salt concentrations, are evaluated. In particular, the intra-cycle viscoelasticity is analyzed. Chapter 4

is a modified version of the manuscript entitled “Nonlinear Viscoelastic Characterization of Charged

Cellulose Nanocrystal Network Structure in the presence of Salt in Aqueous Media,” published in

Cellulose3.

Chapter 5 is the continuation of Chapter 4, targeting the mechanical and rheological properties of

CNC hydrogel. The increase in mechanical properties and versatility of CNC hydrogel by

incorporation of 5 wt% PVA is investigated. This chapter is a modified version of the manuscript

entitled “Viscoelastic properties of poly (vinyl alcohol) hydrogels with cellulose nanocrystals

fabricated through NaCl addition” ready for submission4.

Chapter 6 presents the diffusion rate of CNC particles and gravity effect on CNC hydrogel, through

the employment of CLSM (FRAP) and DLS. This chapter is a modified version of the manuscript

entitled “Probing Dynamics of CNCs in Gel and Suspension using FRAP and DLS” ready for

submission5.

Finally, the last chapter summarizes the main conclusions of this study and provides several

recommendations for further research.

2 Moud, A. A.; Arjmand, M.; Liu, J.; Yang, Y.; Sanati-Nezhad, A.; Hejazi, S. H., Cellulose nanocrystal structure in the

presence of salts. Cellulose 2019, 1-15.

3 Moud, A. A.; Kamkar, M.; Sanati-Nezhad, A.; Hejazi, S. H., Sundararaj. U.T, Nonlinear Viscoelastic

Characterization of Charged Cellulose Nanocrystal Network Structure in Presence of Salt in Aqueous Media, 1-15,

2020. Cellulose, In press.

4 Moud, A. A.; Kamkar, M.; Sanati-Nezhad, A.; Hejazi, S. H., Sundararaj. U.T, Viscoelastic properties of poly (vinyl

alcohol) hydrogels with cellulose nanocrystals fabricated through NaCl addition. To be submitted.

5 Moud, A. A.; Sanati-Nezhad, A.; Hejazi, S. H., Self-healing and collapse in CNC-based gels and suspensions. To be

submitted.

39

CHAPTER 2: Colloidal Behavior of Cellulose Nanocrystals in

the Presence of Sodium Chloride 6

Aggregation and gelation of cellulose nanocrystals (CNCs) induced by sodium chloride (NaCl) were

investigated as a function of NaCl and CNC concentrations. Incorporation of NaCl improved CNCs'

ability to form clusters via screening surface charges of CNCs. Transmission electron microscopy

(TEM) images revealed the formation of porous CNC clusters following NaCl addition. The confocal

laser scanning microscopy (CLSM) micrographs indicated the presence of regions with colloid-rich

and colloid-poor patterns in CNC clusters. Fluorescent brightener 28 was found to have a strong

hydrogen bonding to the cellulose surface and used as the staining agent in CLSM. The CLSM

images also indicated a dynamic structure for gels, continually rearranging over the course of time.

Zeta potential data, coupled with CLSM images, confirmed the impact of NaCl on the gel formation

of CNCs.

6 Moud, A. A.; Arjmand, M.; Yan, N.; Nezhad, A. S.; Hejazi, S. H., Colloidal behavior of cellulose nanocrystals in

presence of sodium chloride. ChemistrySelect 2018, 3 (17), 4969-4978.

40

Graphical abstract

2-1 Introduction

Cellulose nanocrystals (CNCs) can be sourced from multiple different natural sources, such as plant

cell walls (cotton, algae, wood particles) and bacteria [135]. Tiny crystals obtained from cotton have

a square cross-section with a dimension of approximately 6×6 nm2 and an average length ranging

from 100 to 200 nm. Sulfuric acid hydrolysis, put negative charges on CNCs, and increases their

polarity [136]. It is widely accepted that geometry of CNCs (aspect ratio) and chemistry of the surface

of CNC nanorods, along with ionic strength of the aqueous medium, governs the colloidal behavior

of CNC suspensions [137].

In the presence of coagulants, CNC enter a phase that are unstable colloidally and aggregate

after they come into vicinity of one another. Depending on CNC concentration, CNC clustering can

happen that leads to precipitation or a fractal gel [137]. In a study reported by Cherhal et al. [33], it

CNC CNC – N C

300 μm

41

was shown that the incorporation of NaCl causes the formation of a gel structure that can be formed

at low concentrations of CNCs due to the elongated nature of CNC nanorods. In fact, the suspension

stability of CNC particles is highly affected following the addition of NaCl.

CNC rod particles due to acid involved method of synthesis have roughly one negative charge

per 10 anhydroglucose units [136]. Generally, charged CNCs, because of electrostatic repulsive

interaction, do not aggregate in deionized water, but the addition of NaCl can induce random

aggregation [138]. Analysis of structure with small-angle neutron scattering of various CNC-NaCl

combination loadings of namely 2, 10, 50, and 200 mM has been carried out recently [33].

Aggregates with self-similar shaped structure. The process of aggregation was rapid after NaCl

passed a certain threshold. Larger clusters were formed when more NaCl was added (more than 10

mM). Self similarity of aggregates was proven with interpretation of a strong upturn at the lowest

scattering wave vectors. The network stability towards gravity was good at high CNC concentration

but the network failed when CNC loading was low. Similar system of CNC showed aggregation with

changes in ionic strength of the media [139]. Reviewing of reports in the literature shows that

underlying mechanisms regarding CNC gel and its colloidal behavior have not been fully

investigated.

Therefore, in this study, we try to present a systematic investigation of CNC aggregation and

network formation as a function of CNC and NaCl concentrations. By employing transmission

electron microscopy (TEM), scanning electron microscopy (SEM), and confocal laser scanning

microscopy (CLSM), the structure of the developed gels and its evolution were investigated. To the

best of our knowledge, this is the first study using CLSM to investigate the gel structures of CNCs.

Furthermore, employing CLSM, this study investigates the effect of ionic strength and CNC

concentration on the extent of aggregation and structure of the CNC network. Development of gel

structures with specific structural characteristics helps make efficient aerogels with desired porosity

and mechanical strength for applications such as air and water filters, and also provides a perfect

substrate for aerogel nanocomposites.

2-2 Results and discussion

2-2-1 Transmission electron microscopy of CNC suspensions

42

Figure 2.1 depicts the low-magnificationand high-magnification TEM images of CNC aqueous

suspensions made at different CNC and NaCl concentrations. TEM images show the rod-like shape

of individual CNCs with the aggregation state for 5 g/L CNC, and one mM NaCl (Figure 2.1(a) and

(d) Figures 6(a)). Moreover, it was observed that increasing both CNC and NaCl contents led to

CNC aggregation, where CNC clusters containing a substantial number of individual CNCs

appeared. The addition of salt increased the number of junctions per CNC, leading to denser clusters

and elongated particles, and in some cases, increased the apparent aspect ratio through attachments

at extremities.

Generally, during the production of CNCs, the sulfuric acid hydrolysis step leads to the

formation of ester sulfate groups at CNC surfaces [136]. Negative charges, causing electrostatic

repulsions between CNCs, bring about CNCs stabilization in aqueous medium at low ionic strengths

through the prevention of aggregation caused by attractive van der Waals interactions. In colloidal

suspensions, attractions cause the gelation through the formation of particle-rich and particle-poor

regions. However, the transition to full separation might be stopped. Depending on ionic strength of

the medium, at certain inter-particle potentials and particle concentrations, the attractions that induce

separation of phases can also retard or stop gel growth, thereby rigidifying a non-equilibrium

configuration for the total volume of particles, and thus resulting in the formation of a gel structure

[140, 141].

The TEM images of charged CNCs in the present study are qualitatively like those reported

for other aggregating rod-like colloidal particles in the literature [142-144]. The kinetic of gelation,

gel point, and gel strength depends on the charge density and the shape of the CNC particles, which

are strong functions of the precursor materials and the processing protocol [137].

1 µm

(a)

1 µm

(b)

1 µm

(c)

43

Figure 2.1 Low-mag and high-mag transmission electron microscopy (TEM) images of cellulose nanocrystals

(CNC) aggregation at different CNC and sodium chloride (NaCl) concentrations. (a) and (d) 5 g/L CNC at 1 mM

NaCl; (b) and (e) 15 g/L CNC and 5 mM NaCl; (c) and (f) 15 g/L CNC and 10 mM NaCl.

Charge density screening of ions in the system, made gelation faster. In absence of salt, gelation of

CNCs with low electrostatic repulsive interactions was also reported to be rapid [33, 43].

Furthermore, annealing at higher temperature causes reduction of charge density and led gelation

[43].

The root of destabilization is in short-range interaction energy between two neighboring particles.

According to Derjaguin-Landau-Verwey-Overbeek theory, inter-particle pair potential and the

repulsive Yukawa potential take into account van der Waals attractions and electrostatic repulsions,

respectively [145, 146]. The repulsive Yukawa potential is mainly dictated by two independent

parameter (1) net surface charge of nanoparticles, and distribution of charge on their surface, and (2)

Debye length κ−1, which is associated with the ionic strength of the medium (being water here).

Generally, changing coagulant ionic strength ranging between 0 to 10 mM at CNC loading

range of 2 to 50 g/L has been the targeted by the majority of publications [138, 147, 148]. Debye

length is always found to be higher than 3 nm, which is large enough to limit local aggregations.

Some reports show destabilization is under influence of coagulant concentration [147, 148], although

ionic strength necessary for aggregation depends on surface charge and distribution of charges on

the CNCs. Surface charge and its distribution depends on the cellulosic source, method of processing

and chemical agent used for treatment of CNCs.

200 nm

(d)

44

The compact nature of CNC clusters observed in this study (Figure 2.1 (c) and (f)) is similar to

the results reported by Cherhal et al. [33], where the fractal dimensions for charged and uncharged

CNCs in the presence of NaCl were reported to be 2.1 and 2.3, respectively. The discrepancy in

fractal dimensions reveals a denser three-dimensional aggregation for the uncharged CNCs

compared to the charged CNCs. Enhancement in the possibility of obtaining denser aggregates for

the uncharged CNCs is due to the lack of neutralization of surface charges, decrease in repulsive

electrostatic interactions between CNC pairs, and impact of the effectively excluded volume.

Another striking feature of the colloidal suspension of nanorods is the innate ability to create

porous clusters. As shown in Figure 2.1, the density of the porous cluster increased as the

concentration of salt increased from 5 mM to 10 mM. Colloidal stability and aggregation criteria

have been studied in the past for system of ions with different valences [34], where density of

aggregates increases with increase in number of CNCs available in the system. The morphology of

the porous clusters observed here via TEM resembles the structures obtained via hydrothermal

gelation route [43]. The resemblance between the two structures further validates the role of surface

charges in gelation.

2-2-2 Scanning electron microscopy of freeze-dried gels

To visualize the structure of the generated gels, the gels made at 7.5 g/L and 15 g/L CNC

concentration and ten mM NaCl were freeze-dried and imaged with the SEM setup. Figure 2.2

illustrates the bundles of CNC fibrous networks around water droplets etched out of the system within

freeze-dryer. It is noted that the clusters observed in the SEM images mimic the morphology

perceived in the TEM images, as individual CNCs appear to have a random spatial orientation. The

gel network prepared in the presence of NaCl exhibited a random orientation of nanofibrils in the

length scale up to several micrometers. The SEM images also revealed that the variation in CNC

concentration at a constant salt content influenced the gel mesh size. An increase in the concentration

from 7.5 g/L to 15 g/L led to a decrease in mesh size from 1.4 µm to 1.2 µm, as measured by the

ImageJ software. The larger mesh-size at the lower CNC concentration was the result of a smaller

number of associating CNCs and a lower number of contact points per gel volume.

45

Figure 2.2 Scanning electron microscopy (SEM) images of the CNC network at different magnifications for CNC

concentrations of (a-c) 7.5 g/L and (d-f) 15 g/L. The concentration of NaCl for all images is 10 mM

2-2-3 Confocal laser microscopy

Confocal laser scanning microscopy (CLSM), as shown in Figure 2.3, demonstrates gradual changes

in the suspension morphology with the addition of NaCl. In the absence of the salt, no sign of

agglomeration is detected with CLSM, as verified by the homogeneous green color. In fact, CNC

suspensions without salt are homogeneous without any gel formation; thus, the fluorescence intensity

is almost the same for each pixel. Upon addition of a small amount of salt (0.33 mM), the first sign

of aggregation appeared in the composition. The addition of larger amounts of NaCl pushes the

system toward gelation, which spanned the entire visualization cube. The green regions signify the

presence of CNC gel structure holding FB 28 fluorescence dye, whereas the black parts indicate

CNC-free regions.

(a)

10 μm 2 μm

(b) (c)

1 μm

10 μm

(d) (e)

2 μm

(f)

1 μm

46

Figure 2.3 Growth of CNC network at 15 g/L CNC at different concentrations of NaCl: a (0 mM), b (0.33 mM),

c (0.45 mM), and d (1 mM). The dimensions of the visualization cube are 100×636×636 µm3. The 3-D confocal

laser scanning microscopy (CLSM) images were twisted to obtain a better view of the gel network

To quantify the CLSM images in Figure 2.3, the width of the distribution was characterized via

normalization of standard deviation:

𝜎 = ⟨𝐴⟩−1√𝑛−1 ∑(𝐴𝑖 − ⟨𝐴⟩)2

𝑛

𝑖=1

2-1

Where ⟨𝐴⟩ is the signal that has been spatially averaged, and Ai represents the value of pixel i. The

smallest value of 𝜎 is, thus, determined by changes in the intensity of the fluorescence dye across the

sample. In CLSM, the illumination volume is determined by the microscope, optics, and lenses used,

and does not depend on the pixel size. Illumination duration has also been adjusted with scanning

speed across the sample. For systems homogeneous in length scales larger than the resolution of

CLSM, 𝜎 has been reported to have an inverse relationship with volume and duration of illumination

(b)

(c) (d)

47

and fluorescence dye concentration [149]. Accordingly, one needs to be sure that the spatial

distributions of the fluorescence dye and CNCs, and the volume and duration of the illumination

remain the same for the repeated measurements.

Figure 2.4 Variation of σ as a function of NaCl concentration for CNC gels with 15 g/L concentration.

The values of 𝜎 gave a rough indication of heterogeneity in the gels and were plotted as a function

of NaCl concentration in Figure 2.4. For each point shown in Figure 2.4, the degree of heterogeneity

of 3 samples was measured. It can be observed that as the amount of salt added into the system

increased, the value of 𝜎, indicating the degree of heterogeneity, increased across the system.

2-2-4 Zeta potential and hydrodynamic radius

To have a better understanding of CLSM images, changes in zeta-potential and hydrodynamic radius

as a function of CNC concentration were further measured. For aqueous suspensions made of CNC,

the Smoluchowski equation was employed for the conversion of mobility values to zeta potential

[136, 139],

µ𝐸 = 𝜀𝑟 𝜀0 𝜁/𝜂, 2-2

Where ζ is zeta potential, εr is the dielectric constant of water, ε0 is the permittivity of the free space,

and η is the dynamic viscosity of the water (Pa.s). This equation is only valid for a thin double layer

48

comparatively smaller than the hydrodynamic radius of the particle [150]. The first layer contains

ion adsorbed onto the surface due to chemical interaction, which will render this layer either positive

or negative in terms of total charge. The second layer is ions migrated to the vicinity of the surface

due to Coulomb forces, which practically screens the first layer.

The degree of dilution (depending on concentration) and ionic strength of the medium impact

the calculated zeta potential [136, 151]. Therefore, reported zeta potential values have been used only

for the sake of comparison, and might not be considered to indicate the exact value of the CNC

surface charge. The results of the measurement of zeta potential values of CNC suspensions as a

function of nominal NaCl concentration of 0 to 1 mM are presented in Table 2.1. The CNC particles

without any electrolyte had a zeta potential of −64 mV, close to the results reported by Boluk et al.

[147] and Shafiei-Sabet et al. [136]. The reduction of the absolute value of zeta potential due to the

addition of NaCl is possibly due to migration and adsorption of Na+ ions on the negatively charged

CNC elongated surfaces, therefore retracting the double layer surrounding the nanocrystal particles.

In similar colloidal systems, zeta potential has been impacted by adjusting the ionic strength of the

medium, which is in agreement with the double-layer theory [152]. Please note that the zeta potential

of CNCs is considerably higher than the cellulosic fibers produced by copious extraction processes

(between −20 mV and −50 mV) [151, 153]. Variation in zeta potential values for different cellulosic

sources (e.g., bacteria) is due to the method of processing and chemicals involved in CNC production,

and it is very common [111, 154].

Table 2.1 Changes in CNC suspensions zeta potential at a fixed concentration of 0.5 g/L CNC as a function of NaCl

concentration.

CNC (g/L) NaCl Content (mM) Zeta Potential (mV)

0.5 0 -64±4

0.5 0.33 -60.2±3

0.5 0.50 -55.9±3

0.5 0.87 -42.2 ±2

0.5 1.00 -31.7 ±3

The impact of NaCl on the hydrodynamic size of CNC particles was also investigated with dynamic

light scattering (DLS). The results showed that the equivalent hydrodynamic size (z-average) of CNC

nanoparticles at 3 gr/L concentration increased from 45±2 nm to 56±6 nm and 75±8 nm with the

https://en.wikipedia.org/wiki/Electric-field_screening

49

addition of 0.33 and 0.50 mM NaCl, respectively. These results show that even a small amount of

salt can induce aggregation. Zhong et al. [151] also reported a similar trend for their CNC

suspensions in the presence of NaCl.

2-2-5 Dynamicity of the gel

A series of images were taken from the suspension of 10 gr/L CNC in the presence of 2 mM salt over

the course of 30 min at intervals of 5 min (Figure 2.5). The images have been binarized using the

ImageJ software. The confocal micrographs reveal that the structure of the gel continually rearranged

over the course of time, implying that it was a dynamic gel. Simulations have shown that this

continuous rearrangement compacts the clusters. Similarly, experiments show that colloidal gel

thickens over time and pores grow larger [141, 155, 156]. For instance, evidences show that

microstructure continuously changes the interior of the gel [156-158]. Theories developed for

describing coarsening, are Kramer’s escape-time theory [155, 159, 160] and transient network theory

[161], which both are consistent with our observations. In accordance with our study, Zia et al. [140]

reported that colloidal gel coarsen and this process changes rheology and dynamic of the gel.

According to some models, microstructural changes of the gel happen through the diffusion of

particles from cluster to cluster, movements of particles along the contour of the network, and

advective flow connected with condensed (liquid and solid) phase (see for example [141, 154, 156]).

Dynamic of Breakage or coalescence of network branches are the primary focus of other models

[157, 162-164]. For example, d’Arjuzon et al. [163], through simulation of hard spheres equipped

with short-range attractions, showed that migration of individual particles plays a minor role in

structural changes of the gel. In fact, visual inspection of their simulated system demonstrated that

only a few mobile particles exist in the suspension during microstructural changes. Although

individual particles are fast, due to their scarcity across the sample, they cannot contribute to the

coarsening of the gel. In another study, dynamical analysis of the gel revealed that coarsening does

not happen through the merging of large scale networks, and instead, changes are due to breakage of

an entire strand into smaller pieces and its displacement through solvent and subsequent merging

[140].

50

Figure 2.5 Rearrangement and slow coarsening of the gel network (black parts) of the aqueous suspension of 10 g/L

CNC with 2 mM NaCl concentration over a period of 30 min. The dimensions of the two-dimensional visualization

box are 636×636 µm2

2-2-6 Gravity drivel gel collapse of CNC

The origin of collapse due to the effect of gravity has been reported to be coarsening [141, 160];

however, the critical condition necessary for the collapse has remained unclear. Poon et al. [141] by

employing dark-field imaging, showed that the collapse is a complex process and involves the

formation of voids and channels. Another probable scenario is that non-stop coarsening brings the

system to a brink where the counter-flow of liquid due to the pressing of the porous gel causes some

pores to grow suddenly by erosion deep within the structure [165].

Regarding the timing of the collapse, two complex behaviors have been reported. The first

behavior is slow pace precipitation, in which pores grow smaller while the bulk of the gel gets thicker

[161]. The second behavior is the delayed reaction of the gel with respect to gravity. In this case,

after a critical aging time, the gel network collapses. Research shows critical aging time can be

impacted by the strength of cohesion between particles, size of primary particles, and so on [155,

166, 167]. Delayed collapse brings the gel to a next to the collapse state [155]. At its initial stages,

5 min 10 min 15 min

20 min 25 min 30 min

51

gelation proceeds through the percolation of particles. At longer times, particles can detach slowly

from the branches of the gel and get displaced through the solvent [158, 168]; these integral changes

induce reconfiguration of the gel network and unavoidable eventual collapse [158, 162, 169, 170].

In the present study, dynamics of the collapsing gel is recorded by taking 3D CLSM images,

spanning a cube of 100×636×636 µm3, over the course of time. As shown in Figure 2.6, it is evident

that the gel network is initially distributed throughout the whole cell prior to gravity-driven collapse.

Because of the mismatch in the density of gel and water (1.6 vs. 1.0 g/cm3), a gravity-driven flow is

expected. The microscopic changes in the morphology of the network, ending in gravity-driven

collapse, are evident in Figure 2.6. In fact, the collapse occurs when the CNC network at low

concentrations could not support its own weight, and precipitates [33]. Analogous results were also

observed for the CNC concentration of 7.5 g/L (in the supporting information). Comparing Figure

2.7 and Figure 2.9, indicates that the CNC networks were densified with increasing CNC

concentration.

10 min

15 min 20 min

52

Figure 2.6 Gradual collapse of the gel network of the aqueous suspension of 5 g/L CNC with 10 mM NaCl

concentration as a function of time. The dimension of the visualization box is 100×636×636 µm3. The dispersion

had a height of 3 mm, and the images were recorded at the height of ~ 1 mm above the base of the cell.

Figure 2.7 shows CLSM images of the suspension of 15 g/L CNC with 10 mM NaCl as a function

of time. Contrary to the results of Figures 6 and S1, the gel structure of CNC with 15 g/L

concentration did not collapse during the time frame of the experiment. Evidently, the strength of the

gel at 15 g/L was high enough to resist the gravity effects within the time frame of the experiment.

Indeed, at 15 g/L, no movement was observed in length scales probed by CLSM. Generally, the

higher the concentration of the CNC, the lower the collapse rate of the network. Similar observations

were reported by Teece et al. [156], who performed CLSM imaging of colloid-polymer mixture with

the long-range attraction to study gravity-driven gel collapse.

25 min 30 min

10 min

15 min 20 min

53

Figure 2.7 Depiction of a robust gel network of the aqueous suspension of 15 g/L of CNC with 10 mM NaCl

concentration as a function of time. The dimension of the visualization box is 100×636×636 µm3. The dispersion had a

height of 3 mm, and the images were recorded at the height of ~ 1 mm above the base of the cell.

2-2-7 Re-dispersion of already formed gel

It was noticed that the bonds established between the CNC individual particles in the gel were sturdy

enough to withstand dilution in deionized water. This implies that the thermal energy of CNC

particles was not enough to overcome the attractive forces, denoting an external source of energy

such as sonication for shattering the structure and dispersing CNC gels in water is necessary. Figure

2.8 depicts CLSM images of a CNC suspension (7.5 g/L CNC with 10 mM NaCl) diluted in deionized

water prior to and the following sonication. As evident, after sonication for 1 min, the CNC cluster

structure demolished. This observation is in accord with a study reported by Peddireddy et al. [138].

This functionality further highlights the versatility of CLSM for monitoring the colloidal behavior of

CNCs.

In fact, dilution with deionized water to a lower effective concentration along with sonication

induced a partial breakage of the network decreased the aggregate volume fraction below the

percolation threshold and led to precipitation. Moreover, the bonds formed between CNC clusters

appeared weak because they broke via swelling by the invasion of water molecules, leading to the

disentanglement of the 3D gel into its components. It should be noted that sonication was found to

be effective in demolishing the structure of CNC clusters at higher NaCl and CNC concentrations.

25 min 30 min

54

Figure 2.8 Gel network of the aqueous suspension of 7.5 g/L CNC with 10 mM NaCl diluted with 10 ml deionized

water: (a) before sonication, and (b) after sonication for 1 min. The dimension of the visualization box is 100×636×636

µm3. The dispersion had a height of 3 mm, and the images were recorded at the height of ~ 1 mm above the base of the

cell

2-3 Conclusion

The addition of NaCl causes aggregation of negatively charged CNCs, which subsequently causes

the formation of self-similar clusters that grow until space-filling gel forms. It was shown that the

addition of salt increased the number of contact points per CNC particle, leading to denser clusters

and elongated particles. SEM images revealed that the gel’s mesh size had an inverse relationship

with the CNC content, ascribed to a smaller number of associating CNCs, and a lower number of

contact points per gel volume. In line with the TEM images, CLSM was found to be a versatile

technique to monitor the colloidal behavior and gel structure of the cellulose nanocrystals. The results

indicated that the zeta potential and hydrodynamic radius are important parameters to trace the

genesis of the evolution of the gel structure in cellulose nanocrystals. CLSM images also revealed

that the structure of the gels continually rearranged over the course of time, representing a dynamic

gel. Moreover, the gel was found to be resilient at high concentrations of CNC but collapsed at low

concentrations. The branches connecting clusters were found to be weak, breaking upon swelling

and leaving suspended isolated clusters behind. However, sonication almost completely shattered the

structure formed during gelation.

(b)

55

2-4 Supporting information (CHAPTER 2)

The experimental section has been included in the supporting information section. Supporting

information contains experimental details and schematic of the gradual collapse of the gel network

of the aqueous suspension of 7.5 g/L CNC with 10 mM NaCl concentration as a function of time.

2-4-1 Materials

CNC was purchased from Innotech Alberta with a reported length of 100-200 nm and a diameter of

5-15 nm. According to the supplier, the density and crystallinity index of CNCs is 1.6 g/cm3 and 80%,

respectively, with an average length of ca. 150 nm by TEM images. The purchased CNCs were

extracted via sulfuric acid hydrolysis, leading to the formation of ester sulfate groups at CNC

surfaces. Fluorescent brightener 28 (FB 28) (Sigma Aldrich) was used as the staining agent in CLSM.

2-4-2 Materials preparation

3 wt% stock of CNC suspension was prepared by sonication of batch of CNCs in DI water. The pH

of the suspension was measured using a Mettler Toledo Seven Compact pH-meter (Mettler-Toledo

135 International Inc., Columbus, OH, USA) and was set to be 6.8. The ionic strength of the

suspension was adjusted by addition of a concentrated 20 mM NaCl solution.

2-4-3 Scanning and transmission electron microscopies

TEM and SEM were utilized to study the gel structure of CNCs and the location of CNC clusters.

TEM images were generated using a Tecnai TF20 G2 FEG-TEM (FEI, Hillsboro, Oregon, USA) at

an acceleration voltage equal to 200 kV. A droplet (5 μL) of the generated suspension was dripped

on a carbon-coated electron microscopy grid for improved observation. The micro-morphology of

the developed gel was observed using a scanning electron microscope (XL30, Philips). Prior to SEM

imaging, the generated gels were freeze-dried using liquid nitrogen. A small piece of freeze-dried

hydrogel was mounted onto a silica wafer. A layer of gold was sprayed on the samples by a vacuum

sputter to form a conductive surface and avoid electrostatic discharging. The ImageJ software was

employed to find the distribution of pore size in the freeze-dried hydrogels.

56

2-4-4 Confocal laser scanning microscopy

Fluorescent FB 28 stain that binds efficiently to cellulose was used to monitor the location of cellulose

clusters in the suspension. The nitrogen, hydroxyl and sulfonic acid groups in FB 28 are responsible

for strong hydrogen bonding to the cellulose surface [171]. FB 28 binds to polysaccharides through ß-

1, 3 and ß-1, 4 linkages, such as chitin and cellulose, and is, therefore, able to stain starch-based

materials.

CNCs were labeled with FB 28 dye by adding 20 ppm of the dye to the CNC suspension, leading to

physical adsorption of the dye onto the CNC surface. The concentration of the dye needs to be selected

below the threshold to neither influence neither the behavior of CNCs nor the gel; however, the

concentration needs to be sufficient to provide enough fluorescence to CNCs [172]. Following

mixing and short sonication of the dye and CNC suspension, the samples were left in the dark

environment for 30 min for incubation of the fluorescence dye with CNC particles. CNC suspensions

were then sandwiched between a concave slide and a cover glass. CLSM monitoring was performed

for suspensions at different NaCl and CNC concentrations. CLSM measurements were carried out

with an inverted Nikon confocal microscope (Ti-A1R) equipped with apochromatic lens objectives

of 10X and 20X with resolutions of 500 and 300 nm, respectively. The microscope galavno’s scanner

enabled us to achieve high-resolution images up to 4096 x 4096 pixels.

57

2-4-5 Zeta potential and particle size measurements

Nano-Zetasizer (Malvern Instruments, Nano ZS, Malvern, UK) was used to probe zeta potential and

size of CNC particles dispersed and distributed in DI water. The device is also equipped with a zeta

potential analyzer that employs electrophoretic light scattering for studying particles, molecules, and

surfaces. Zetasizer using light scattering technique to find the mobility of the particles due to

Brownian motions [111].

58

Figure 2.9 Gradual collapse of the gel network of the aqueous suspension of 7.5 g/L of CNC with 10 mM NaCl

concentration as a function of time. The dimension of the visualization cube is 100×636×636 µm3. The dispersion

had a height of 3 mm, and the images were recorded at the height of ~ 1 mm above the base of the cell

5 min 10 min

15 min 20 min

25 min 30 min

59

CHAPTER 3: Cellulose Nano Crystals structure in the

presence of salt 7

Aggregation and gelation of cellulose nanocrystals (CNCs) induced by magnesium chloride

(MgCl2) are investigated as a function of CNC and MgCl2 concentrations. Transmission electron

microscopy (TEM) and confocal laser scanning microscopy (CLSM) is employed to study the

effect of ionic strength and CNC concentration on the extent of aggregation and structure of the

CNC network. The location of CNC particles is traced with a Fluorescent brightener 28 staining

agent. The results show that the addition of different amounts of MgCl2 causes a cluster formation

of CNCs with different fractal dimensions, confirmed by TEM. The fractal dimension of CNC

clusters is varied from approximately 1.56 ± 0.08 to 1.98 ± 0.01 as the MgCl2/CNC concentration

ratio is increased from 0.17 to 0.42. We use the MgCl2/CNC concentration ratio as a global

parameter to correlate the results of different measurements and imaging data, including TEM,

zeta potential, and CLSM. Furthermore, we conduct molecular dynamics simulations to

quantitatively examine different CNC behavior in MgCl2 salt-CNC suspension. The results on the

potential of mean force (PMF) indicate that the PMF of different ions concentration gravitates to

zero, where the distance between CNCs is increased from 3.1 nm to 3.5 nm. However, adding ions

to the system changes the energy of the system and leads to the different behavior of CNC

interactions.

7 Moud, A. A.; Arjmand, M.; Liu, J.; Yang, Y.; Sanati-Nezhad, A.; Hejazi, S. H., Cellulose nanocrystal structure in

the presence of salts. Cellulose 2019, 1-15.

ABM did the experimental design, data collection and interpretation, and manuscript preparation. JL contributed in

molecular dynamic simulations.

60

Graphical abstract

3-1 Introduction

Cellulose nanocrystals (CNCs) are rod or whisker shaped particles produced using a hydrolysis

reaction using acid out of wood particles or other resources. CNCs are whisker shape and can have

different crystallinity levels depending on the source of cellulose and method of processing. Sizes of

CNCs, van der Waals forces, surface charges and its distribution, hydrophobic and hydrogen

interactions dictate the gel formation and colloidal behavior[138]. Electrostatic repulsion among

individual CNCs can be adjusted through reducing charge density, for instance, desulfation or

annealing at high temperatures [43]. Short-range attraction forces between CNCs with addition of

coagulant can be empowered over repulsive forces [173]. Hence, CNC aggregates from a

suspension into a precipitated ensemble of clusters or a fractal gel [33, 138]. Gel formation will

also happen for pure CNCs if concentration reaches thresholds of 10 wt% [147]. Self-similar

aggregates can also form through addition of coagulant such as salt and adsorbing or nonadsorbing

polymers. There are other methods such as freeze thaw cycles gel formation mechanism as well.

61

A study reported by Cherhal et al. [33] shows gel formation after salt introduction. Chau et al. [48]

showed addition of salt causes empowerment of van der Waals forces over electrostatic repulsion

forces. Authors also claimed that stiffness of the gel increases with increase in ion size and charge

number on the ions. In another study, Uren a-Benavides et al. [174] reported solid to gel transition

point is approximately one order of magnitude lower in presence of coagulant in comparison to

pure CNC.

The solid to gel transition is abrupt and dubbed as the critical aggregation concentration

(CAC) via empirical relationship of Schulze–Hardy empirical as 𝐶𝐴𝐶 ∝1

𝑍𝑛 where Z is valence of

counter ion (n = 6 is a number assigned to the highly charged particles and n = 2 a number assigned

for weakly charged particles) [175]. Through experiments accurate prediction of this empirical

relationship has been proven over and over for nanotubes [176], nanofibers of peptide [177], and

CNC [178-181].

Despite many recent reports on the suspension stability of many types of nano cellulose [33,

147, 148, 151, 179, 182-185], to the best of our knowledge, no systematic study has been reported

on colloidal stability of CNC particles with confocal laser microscopy. Moreover, authors believe

more attention must be paid to the impact of these parameters on the aggregates. The structure of

these aggregates and the concentration and valence dependence of the aggregation onset are pivotal

to develop materials out of CNC gels.

In the present study, we seek to present a systematic investigation of CNC aggregation and

network formation as a function of CNC and MgCl2 concentrations. Employing transmission

electron microscopy (TEM) and CLSM, the structure of the developed gels and their evolution are

investigated. Furthermore, employing CLSM, this study investigates the effect of ionic strength

and CNC concentration on the extent of aggregation and structure of the CNC network.

Development of gel structures with controllable porosity and rate of gelation and collapse

facilitates the production of efficient gels with desired porosity, which is pivotal for applications

such as air and water filters and provides an ideal matrix for aerogel nanocomposites.

3-2 Experimental section

3-2-1 Materials

62

InnoTech Alberta provided CNC that had a reported length of 100-200 nm and a diameter of 5-15

nm. The reported density and crystallinity index of cellulose was 1.6 g/cm3 and 80%, respectively,

according to the manufacturer data, with an average length of ca. 150 nm quantified using TEM

images. Extraction of CNC happened through hydrolysis via H2SO4, with formation of half ester

sulfate functional groups on the CNCs. Fluorescent brightener 28 (FB 28) (Sigma Aldrich) was

used to tag CNCs.


A stock suspension with 3 wt% of CNCs was made by dispersing and distributing the spray-dried

CNC powder in deionized water. Suspensions were made with a pH of 6.8 as measured using a

Mettler Toledo Seven Compact pH-meter (Mettler-Toledo 135 International Inc., Columbus, OH,

USA). The ionic strength of the suspension was adjusted through the addition of 80 mM MgCl2

solution. To prepare the final suspension with desired concentrations, samples were diluted with

deionized water (DI) and sonicated for 20 min. Ultra-sonication (125 W Qsonica Sonicators Q125

Sonicator, Qsonica) was used for dispersion of CNC into DI water. Ultrasonic treatment was done

in an ice bath to disallow overheating as the surface charge of CNC particles is sensitive to

temperature [50].

3-2-3 Materials characterization

3-2-3-1 Scanning and transmission electron microscopies

TEM was employed to assess the gel structure of CNCs and the location of CNC clusters within

the evolving structure. TEM images were produced using Tecnai TF20 G2 FEG-TEM (FEI,

Hillsboro, Oregon, USA) at an acceleration voltage equal to 200 kV. A droplet (5 μL) of the

generated suspension (or gel) was dripped on a carbon-coated electron microscopy grid for

improved observation.

3-2-3-2 Confocal laser microscopy

Fluorescent FB 28 fluorescent dye that binds preferentially to cellulose [173] was employed to

assess the location of cellulose clusters in the suspension. Carbohydrate-Aromatic (CA)

interactions from van der Waals forces (CH–p interactions) and the hydrophobic effect have been

63

reported to explain the adsorption of aromatic molecules to carbohydrates such as cellulose in

aqueous environments [186, 187]. These CA interactions, as opposed to electrostatic interactions,

is possibly responsible for dye binding in our report [186]. It is believed that in our samples, CA

interaction is the main force that attaches FB28 to CNC.

CNCs were labeled with FB 28 dye by adding 20 ppm of the dye to the CNC suspension, leading

to physical adsorption of the dye onto the CNC surface. The concentration of the dye must be

selected below a threshold that does not influence the colloidal behavior of CNCs. Zeta potential

values in our system did not change for sample of 10 g/L CNCs even if the FB28 concentration

increased to 500 ppm. However, 20 ppm of dye was enough for tagging CNCs [172]. After

physical mixing of CNC with dye, the samples were left in the dark for duration of half an hour

for the dye tagging to happen on the CNCs. CLSM monitoring was performed for suspensions at

different MgCl2 and CNC concentrations. CLSM measurements were carried out with an inverted

Nikon confocal microscope (Ti-A1R) equipped with the apochromatic lens objectives of 10X and

20X providing the resolutions of 500 and 300 nm, respectively. The microscope’s galvanometer-

based scanner enables achieving high-resolution images up to 4096 x 4096 pixels.

3-2-3-3 Zeta potential and size measurements

Nano-Zetasizer (Malvern Instruments, Nano ZS, Malvern, UK) was used to characterize the zeta

potential and the size of CNC particles suspended in DI water. The Zetasizer Nano ZS equipped

with two analyzers was used for the detection of aggregates and the measurement of small to

relatively large CNCs (0.3 nm to 10 µm in diameter) in dilute samples. The device was also

equipped with a zeta potential analyzer that employs electrophoretic light scattering for studying

particles, molecules, and surfaces. Zetasizer uses the dynamic light scattering technique to trace

the movement and size of particles while they are in the Brownian motion regime [111]. Given the

rod geometry of CNC particles and considering the original development of the Stokes-Einstein

equation for particles with spherical geometry, the size of CNC particles measured by the Zetasizer

does not represent the real particle size and shape. Nevertheless, it represents an equivalent

hydrodynamic size of particles. Therefore, the result of the size measurement presented in this

study has only been used to compare sizes among different CNC samples. To make the

measurements report reliable, the size measurements were performed for 10 replicates.

64

3-2-3-4 The calculation method of MD simulation

The molecular dynamic simulations were performed using LAMMPS package [188]

Subsequently, VMD package was used for visualization [189]. The potential of mean force (PMF)

[190] is computed by the colvars tool package [191] in LAMMPS. Van der Waals's interactions

were calculated between different components by Lorentz-Berthelot combining rules [192]. The

cut off distance is set at 1.2 nm. The system implies the particle mesh Ewald (PME) method to

compute the electrostatic interactions. Initially, the energy of the system was minimized, and 3 ns

simulation with the NVE ensemble and the NPT ensemble were used. Then, steered molecular

simulation (SMD) and umbrella sampling methods were used to calculate PMF for 45 ns where

the temperature is maintained at 298 K by Nosé-Hoover thermostats. Following the calculation of

PMF for two CNC rods, the first CNC was fixed in place and the second CNC was pulled to the

first one along the Y direction to compute the free energy.

3-3 Result and discussion

3-3-1 Transmission electron microscopy of CNC suspensions

Figure 3.1 depicts TEM images of CNC aqueous suspensions made at different CNC and MgCl2

concentrations. TEM images show the rod-shape of individual CNCs with the aggregation state

for 10 g/L CNC and different concentrations of MgCl2. It is observed that increasing the MgCl2

content leads to CNC aggregation, where CNC clusters containing a substantial number of

individual CNCs are imaged in Figure 3.1 A-D. The addition of salt increases the number of

junctions per CNC, leading to denser clusters and elongated particles. In some cases, the apparent

aspect ratio is increased through attachments at extremities. The TEM images of charged CNCs in

the present study are qualitatively similar to those reported for other aggregating rod-like colloidal

particles in the literature [142-144].

65

Figure 3.1 High magnification transmission electron microscopy (TEM) images of cellulose nanocrystals (CNC) with

10 g/L concentration aggregation at different Magnesium chloride (MgCl2) concentration. (A) 17 mM, (B) 21 mM,

(C) 32 mM, (D) 42 mM. The ratio of salt/CNC varies from 0.17 to 0.42.

It is noteworthy that similar behavior was observed in the previous report for CNC-NaCl systems

[173]. Also, the compactness of CNC clusters at high salt concentrations of MgCl2 as observed in

this study (Figure 3.1D), is similar to those of the CNC-NaCl coagulated system [33]. A higher

concentration of salt means a higher screen level of surface charges. For instance, Cherhal et al.

[33] used two different CNCs, with and without charges, and reported that the fractal dimensions

for charged and uncharged CNCs after the addition of NaCl are 2.1 and 2.3, respectively.

Attainment of denser aggregates for the uncharged CNCs can be translated into a decrease in the

electrostatic repulsion. This result corroborates why obtaining denser structures is more probable

at high concentrations of salt.

1µm

A

1µm

B

1µm

C

1µm

D

https://pubchem.ncbi.nlm.nih.gov/compound/magnesium_chloride

66

A method for finding the fractal dimension of CNC clustered aggregates is the box-counting

method [193]. A series of boxes with size L2 pixels are used on the binarized 2-D image of the

aggregates. The estimation of the fractal dimension technique is finding the slope of the graph that

correlate the number of boxes with filed pixels labelled as (N(L)) to size of the primary box sizes

sketched in a double logarithmic diagram. For evenly dispersed and distributed systems, 𝑁(𝐿) ∝

𝐿−2, and for isotropic self-similar aggregates such as CNC here, 𝑁(𝐿) ∝ 𝐿1−𝑑𝑓, with df is the

fractal dimension. Applying this method to the TEM images depicts gradual changes in the fractal

dimension of CNC clusters from approximately 1.56 ± 0.08 to 1.98 ± 0.01 as salt concentration

increases from 17 mM to 42 mM. The samples of these changes are shown in Figure 3.1A-D. The

fractal dimensions indicate that the addition of salt at lower concentrations produces less branchy

clusters, while a higher concentration of salt produces denser clusters. The recorded changes in the

cluster density are in line with AFM findings of Honorato-Rios et al. [194], reporting the pointier

CNC structures at lower salt concentrations. However, contrary to the above observation, the

known reaction limited aggregation (RLA) to diffusion-limited aggregation (DLA) transition

theories predict a decrease in fractal dimension with the addition of more counter ions in the

system. The fractal dimensions reported in the literature are lower for DLA than RLA [144].

Further studies are needed to explain the discrepancy.

Due to the anisotropic shape of the rod particles and their nano dimension size, the energy

barriers that disallow the aggregation of CNCs are not symmetric. As a result, the chance of the

end-to-end collision of CNCs is higher than side-to-side. However, side-to-side attachment is more

favored from the thermodynamic standpoint [195]. The dynamics of end-to-end attachment in real-

time has been recently imaged by Alivisatos and co-workers [195]. The authors depicted the

parallel alignment of CNCs before attaching at their extremities. This mechanism is kinetically

different from recent studies on linear assemblies of nanorods, which dictates non-favored end-to-

end attachment through non-uniform surface chemistries [196, 197].

Furthermore, clusters are found to be porous themselves as individual particles assume

random orientation in the space. Upon increasing the salt concentration, denser aggregates

appeared due to less strong repulsion forces among CNCs (Figure 3.1). Also, the morphology of

the porous clusters observed here is similar to the morphology of the gels obtained via the

hydrothermal gelation of CNC suspensions where gels are dried with supercritical carbon dioxide

67

(CO2) [43]. Desulfation happens in hydrothermal treatment of CNC, which accounts for the

decrease in surface charges of CNCs. The similarity between the structures obtained through

desulfation and gelation shows how the origin of aggregation which is DLVO forces. Also, it has

been shown that the larger the counter ion size, the larger the critical concentration for

aggregation.[198]


Figure 3.2 depicts the CLSM micrographs of the gradual structural developments in the

suspension morphology with the addition of MgCl2. From the CLSM viewpoint, in the absence of

salt, the uniform green color translates into the absence of aggregation. In fact, CNC suspensions

without salt are uniform, without any gel formation; thus, the fluorescence intensity is almost the

same for each pixel (Figure 3.2A). Upon the addition of a small amount of salt (8.5 mM), the first

sign of aggregation appeared in the composition (the bright spots in Figure 3.2C). A further

increase in MgCl2 concentration induces gelation in the system, which spans the entire

visualization cubes (Figure 3.2D-F). The green regions signify the presence of CNC gel structure

holding FB 28 fluorescence dye, whereas the dark parts indicate CNC-free regions. The results

show that the gelation evolution versus salt concentration follows a classical DLVO trend [199] in

which the salt addition gradually favors van der Waals forces over electrostatic forces. The CNC

gel structure rapidly evolves beyond a threshold salt concentration. The lack of further changes in

the microstructure following the addition of more salt (42 mM and above which are not shown

here) shows that the arrested phase (CNC) is somewhat uniform and insensitive to the salt content

variations.

68

Figure 3.2 Growth of CNC network at 15 g/L CNC at different contents of MgCl2: (A) 0 g, (B) 4.2 mM, (C) 8.5

mM, (D) 17 mM, (E) 21 mM and (F) 42 mM. The dimensions of the visualization cube are 100×1272×1272 µm3.

The 3-D confocal laser scanning microscopy (CLSM) images are rotated to obtain a better view of the gel hybrid

system. Resolution: 500 nm. Images were taken once the salt was added into the mixture. The ratio of salt/CNC

varies from 0 to 0.27

0 mM 4.2 mM

8.4 mM 16.8 mM

20mg

42 mM 21 mM

A B

C D

E F

250µm 250µm

250µm 250µm

250µm 250µm

69

To quantify Figure 3.2, the width of the distribution is characterized via normalization of standard

deviation in the following equation.

𝜎 = ⟨𝐴⟩−1√𝑁−1 ∑(𝐴𝑖 − ⟨𝐴⟩)2

𝑛

𝑖=1

3-1

Where A is the signal emitted by the sample, and 𝜎 denotes the degree of non-uniformity. The

justification for the accuracy of this formula can be sought in Ref [173]. Signal values were chosen

out of the boxes with sizes equal to the resolution of the images. N is the number of independent

boxes across the images. The value 𝜎 gives a rough indication of non-uniformity in the gel and is

plotted as a function of MgCl2 content (Figure 3.3). For each point, the degree of non-uniformity

is measured for three samples. The value of 𝜎 for MgCl2 increases because of the increase in salt

concentration, indicating the increased structural non-uniformity across the system.

Figure 3.3 Variation in 𝝈 as a function of MgCl2 concentration for CNC gels with 15 g/L CNC concentration

Using Schulz-Hardy rule, assuming that surface charge of CNCs is high, the ratio of critical

aggregation concentration for the MgCl2 system is roughly estimated to be 64 times smaller than

monovalent salt, NaCl (33 mM) [173]. The trend is consistent with Schultz-Hardy rule, i.e., MgCl2

causes gelation to happen sooner, even though a deviation can be observed in predictions (observed

70

8 mM in Figure 3.2C versus expected (33 mM/64) 0.5 mM and (33 mM/4) 8 mM for highly

charged and weakly charged particles, respectively. This deviation may be due to (1) system

contains particles that are not spherical , (2) assuming ionic radius to be negligible [200], and (3)

limited confocal micrograph resolution (300-500 nm). Therefore, the onset of the gelation can also

happen in salt concentrations below 8 mM.

Phan-Xuan et al. [201] recently reported that the onset of aggregation for MgCl2, measured

by turbidity, is one order of magnitude smaller than the aggregation onset in NaCl hybrid systems.

These results further validate the accuracy of confocal images versus conventional methods (for

instance, turbidity values) regarding the detection of perturbation onset and the overall monitoring

of the gelation and microstructure evolution.

3-3-3 Zeta potential and hydrodynamic radius

Figure 3.4 Semi-logarithmic variation of volume percentage as a function of the hydrodynamic radius of CNC.

Inset depicts gradual changes in zeta potential as the MgCl2/CNC ratio changes from 0 to 0.25.

71

The changes in zeta-potentials as a function of CNC concentration are also measured. The result

of these measurements as a function of the ratio of the MgCl2/CNC concentration is shown in the

inset of Figure 3.4. The CNC particles in the absence of any electrolyte show zeta potential values

of −64 mV, similar to values obtained by Boluk et al. [147] and Shafiei-Sabet et al. [136]. We have

used the procedure explained by Shaifie et al. [136] to find the sulfate half ester contents. Through

elemental analysis of SEM, the number of (SO3-) group per 100 anhydroglucose units can be

obtained based on the molecular formula of C6H10O5–(SO3)x and calculated from following

relationship S(wt%) = 100x×S× [6C + 10H +(5 + 3x)O+xS]-1 [202]. In our experiments, the EDAX

spectra of CNC show the 0.66 % sulfur content, which translates itself into 3.39 OSO3H per 100

glucose units. The descending pattern observed as a function of salt concentration is due to the

formation of diffuse layer and retraction of double-layer around each particle [203]. These results

show that the aggregation is expected even if the salt is added at low concentrations. Technically,

the charge or mobility of clusters could be assessed by electrophoresis measurements. However,

clusters are suspended in a poly-disperse sea of clusters and monomers, which makes the

determination of electrophoretic mobility quite challenging.

The results show that the volume percentage of the particles is shifted towards larger diameters. It

is noted that as time can impact gelation, measurement has been performed on the samples about

two hours following the introduction of salt. The results here are in line with the observations of

CLSM (see Figure 3.2) as the addition of salt increases the zeta potential values; hence, intensified

instability in the system.

3-3-4 Dynamics of CNC gel and its relevancy to eventual gel collapse

The gels of the CNC-NaCl hybrid system are shown through experiments to be dynamic at micro-

scale (particle level) [173]. Moreover, simulations also show that rearrangement of colloids after

gel formation makes the clusters more compact [204]. Fairly similarly, experimental studies show

that strands of gel over time becomes thicker [141, 155, 156]. For instance, empirical evidence

through the employment of CLSM showed microstructural changes inside the gel due to these

rearrangements [156-158]. Existing theories developed to explain these behavior such as Kramer’s

escape-time theory [155, 159, 160] and transient network theory [161]; are consistent with trends

we observed in CNC-salt hybrid suspension. Aligned with colloidal behavior seen out of CNC

72

gels, Zia et al. [140] recently reported that dynamics and rheology of fractal gels through

coarsening is under influence of rearrangements and the pace of it changes with changes in

interactions between particles. Mobility of the proven here can provide a scientific basis for

understanding gel collapse and self healing of CNC gel.

3-3-5 Gravity driven collapse of CNC gel

Coarsening might be the reason behind gel collapse [141, 160]. However, there is still debate on

what critical condition should the system reach for the collapse to happen. Poon et al. [141]

observed creation of multiple channels and corridors inside the gel before collapse of the gel.

Coarsening can push the system towards a brink of collapse at which continuous pores among

clusters is happening [165].

Two complex observation regarding collapse of a gel has been reported. Gradual type [161]

and delayed response before collapse. In second case, gel collapses after passage of certain time.

This period is under influence of cohesion and size of gel constituent particles [155, 166, 167].

Initially the system experiences attachment of single particles to branches of clusters while at later

times, particles slowly detach themselves from the clusters and migrate through the solvent [158,

168]. These integral changes induce structural reconstruction and unavoidable ultimate collapse

[158, 162, 169, 170].

In the present study, the dynamics of this process is monitored through recording time

evolution of 3D CLSM images, in a cube of 200×1272×1272 µm3. Figure 3.5 shows that the gel

network is distributed throughout the cell prior to the collapse. The collapse occurs due to the

dense CNC network is expected (1.6 vs. 1.0 g/cm3). The microscopic changes in the structure of

the network, which ended in gravity-driven collapse, are evident in Figure 3.5. In fact, the collapse

is expected due to the inability of the structure to sustain its weight [33]. A similar trend was also

observed for CNC concentration of 7.5 g/L.

Figure 3.6 shows the CLSM images of the colloid of 15 g/L CNC with 52 mM MgCl2 as a

function of time. Opposite to the results of Figure 3.6, the gel structure of CNC with 15 g/L

concentration does not collapse during the first 30 min time frame. The observed behavior is due

to the robustness of the colloid at 15 g/L. As a rule of thumb, collapse pace with increase in

73

concentration of CNC decreases. Teece et al. [156] reported identical observations for colloid-

polymer hybrid system with longer range of attraction forces. It is noteworthy to mention that

height in gravity direction is 200 µm, and collapse across the system does not happen evenly.

The different behavior observed between two different salt concentrations can be traced back

to the discussion presented by Solomon et al. [142] where one can show why suspension shows

more elasticity after increase in salt loading. After adjustments in ionic strength of the medium,

system will reach a zone in which interaction between particles is stronger than thermal motion

forces separating them. Hence, inhibition of particle movement is expected. These inhibitions is

due to effects related to excluded volume [205, 206] and/or pair interactions between CNC pairs

due to attractive van der Waals or depletion forces [207]. Interactions between CNCs due to

movement of individual CNCs can also play a significant role [208, 209]. The increase in particle

number density or strength of attractive interactions limits the mobility of rods to the orders of rod

radius. After reaching this threshold, system individual components do not move, which also

translates into more elasticity and non-ergodicity. Fractal gels are also system in which particles

stop moving altogether due to physical bonds. The behavior of particle gels and glasses is different

, which can be observed by techniques such as dynamic light scattering [210].

Considering the values reported earlier for MgCl2/CNC concentration ratio (increased

from 0.17 to 0.42) and fractal dimension (varies between 1.56 ± 0.08 to 1.98 ± 0.01), it is

concluded that fractal dimension between two Figure 3.5 and Figure 3.6 is almost identical

as both fractal dimensions in two figures should be close to 1.98 ± 0.01. Therefore, geometry

and compactness of clusters do not influence the rate of collapse in these two scenarios.

74

Figure 3.5 Gradual collapse of the gel network of the aqueous suspension of 5 g/L CNC with 52 mM MgCl2 content

in the time span of 30 min with intervals of 5 min. The dimension of the visualization box is 200×1272×1272 µm3,

and the resolution is 500 nm. The dispersion has a height of 3 mm, and the images are recorded at the height of

approximately 1 mm above the base of the cell

5 min

15 min 20 min

30 min25 min

10 min

250µm 250µm

250µm 250µm

250µm 250µm

75

Figure 3.6 Gradual collapse of the gel network of the aqueous suspension of 15 g/L CNC with 52 mM MgCl2 content

in a span of 30 min with intervals of 5 min. The dimension of the visualization box is 200×1272×1272µm3, and the

resolution is 500 nm. The dispersion has a height of 3 mm, and the images are recorded at the height of approximately

1 mm above the base of the cell

3-3-6 Molecular dynamic simulation

Molecular dynamic simulations are conducted to explore the magnitude of forces involved in the

CNC aggregation process. Specifically, we are interested in finding a correlation between this

numerical simulation and the experimental data discussed above through the parameter of the

salt/CNC concentration ratio. Literature-wise, the CNC molecular structure was first simulated

from Cellulose builder [211]. For the sake of simulation feasibility, the length of CNC rods was

reduced to 10 nm while maintaining the aspect ratio constant [201]. Each rod was modeled by five

10 min 5 min

20 min

30 min 25 min

15 min

250µm 250µm

250µm 250µm

250µm 250µm

76

single chains of cellulose molecules, giving a total length, L = 10 nm, and a diameter of d = 1 nm

(aspect ratio is 10). This is approximately 15 times smaller than experimental estimations. No

significant influences of the reduced size of the rods compared to experimental values were

observed upon simulating the same system with bigger lengths and diameters. 8912 water

molecules, reproducing the water density of 1.04 g/cm3, were packed in the system. The effect of

positive ion was simulated by adding Mg2+ and Cl- into the system with the concentrations of 25

mM, 50 mM, 100 mM, and 200 mM. The box size is 3 nm 8 nm 12 nm, as shown in Figure

3.7A with the periodic boundary conditions in three directions. The CNC rods are fixed in the Z

direction to avoid the influence of CNC shaking in different directions while they can still move

in the Y direction. Pcff force field is applied to the system [212]. In Figure 3.7, the free energy

between the rods is shown for several salt concentrations. The force field parameters considered

in this study are summarized in Table 3.1.

Figure 3.7 Molecular dynamic simulation. (A) Snapshot of two CNC rods in the sodium chloride solution. The rods

are fixed on the x and z-axis, and the rods are parallel (distance between two rods set at 3.5nm). The transparent

material is an aqueous solution in which the Na+ is yellow, and the Cl- is blue. In the CNC rod system, the carbon

atom is cyan, the oxygen atom is red, and the hydrogen atom is white. (B) Potential mean force (PMF) of two CNC

rods (T=298 K, P=0.1 MPa) in salts solutions of MgCl2.

3.0 3.1 3.2 3.3 3.4 3.5

0

5

10

15

20

25

30

PM

F (

Kcal/m

ol)

Distance (nm)

25mM MgCl2

50mM MgCl2

75mM MgCl2

100mM MgCl2

200mM MgCl2

A B

77

The PMF graphs, as shown in Figure 3.7B, are used to investigate the interaction between the two

CNC rods. The results of PMF show that the PMF of different ions concentration gravitates to

zero, where the distance between CNCs increases from 3.1 nm to 3.5 nm. However, adding Na+

and Cl- to the system changes the energy of the system and leads to the different behavior of CNC

interactions. For instance, when the ion concentration increases above 50 mM for NaCl solution,

the approaching rods receive less resistance. However, at lower concentrations, the two

approaching rods increase the PMF of the system more dramatically. In a nutshell, the results of

molecular dynamic simulation follow the logical mechanism expected for CNC rods with regards

to forces playing a role in salt solutions. Similar behavior is observed for the MgCl2-Water system.

For CNC rods dipped into a solution of MgCl2 with 25 mM concentration, the PMF values gravitate

towards infinity, which means that the two rods cannot be brought in closer. When the

concentration of Mg2+ is small (e.g., 25 mM), only a few ions can be absorbed by CNCs. Therefore,

the negative charge of CNC limits the distance that two rods can be brought to one another. Also,

the PMF value for the concentration of 50 mM for Mg2+ is similar to the PMF value of Na+ at 100

mM concentration, where they both produce quasi-similar trends at 3.1 nm.

In the simulations, the ratio of MgCl2/CNC, considering the number of cellulosic atoms and ions

in the system, varies between 0 to 0.0652 when MgCl2 concentration varies between 0 to 100 mM.

The simulation results show that when this ratio is around 0.05, the act of bringing the CNC closer

together becomes possible. Although the simulation results are not in close agreement with the

experimental data due to the simplification and scaling down in the model, the numerical results

demonstrated to be useful as a helpful guide towards comparing the effect of different amounts of

salt on forces that separate individual CNCs.

Table 3.1 Force field parameters employed in the system.

Atoms/ions Mass (g/mol) σ (nm) Ε (kJ/mol) Charge (e)

o*(H2O) 15.9904 0.3608 0.274 -0.7982

hw(H2O) 1.00797 0.1098 0.013 0.3991

coh(CNC) 12.01115 0.401 0.054 0.213

hc(CNC) 1.00797 0.2995 0.02 0.053

78

c1(CNC) 12.01115 0.401 0.054 0.08

oc(CNC) 15.9904 0.3535 0.24 -0.266

oh(CNC) 15.9904 0.3535 0.24 -0.5571

ho(CNC) 1.00797 0.1098 0.013 0.4241

c-(CNC) 12.01115 0.3908 0.12 0.2974

o-(CNC) 15.9904 0.3596 0.167 -0.5337

c2(CNC) 12.01115 0.401 0.054 0.027

Na+ 22.99 0.39624 0.738 1.0

Mg2+ 24.305 0.4053 0.040 2.0

Cl- 35.453 0.3915 0.305 -1.0

3-4 Conclusion

A systematic study on the CNC gelation in the presence of divalent ions was provided. It is

concluded that the addition of counterion, either Mg2+ or Na+, causes a sudden phase separation in

CNC-salt hybrid systems. It is also reported that the addition of salt created porous, self-similar

structures of CNCs, which are spanned in the whole visualization cube. However, in comparison

with CNC-NaCl systems, clustering occurred earlier and more intensely for CNC-MgCl2. The

shape and structure of CNC clusters in the presence of MgCl2 salt are revealed using TEM.

Calculated fractal dimensions show that the salt concentration affects the morphology of CNC

clusters in a way that at lower salt concentrations, less branchy clusters are formed. CLSM

micrographs complement information obtained via TEM, which depicts their power in monitoring

the colloidal behavior of CNC suspensions. Molecular dynamic simulations reveal the extent to

which two CNC rods in MgCl2 and NaCl solutions can be brought together. It is shown that when

the concentration of Mg2+ is small (e.g., 25 mM), only a few ions can be absorbed by CNCs.

Therefore, the negative charge of CNC limits the distance that the two rods can be brought to one

another. Furthermore, the collapse of CNC gels through by gravity is found to be more resilient at

the high concentrations of CNC while the rate of collapse is more rapid at low concentrations. For

future works, the protocols and techniques used in this work are suggested to be employed for

finding the difference in onset of gelation based on other divalent ions such as Ca2+.

79

CHAPTER 4: Nonlinear Viscoelastic Characterization of

Charged Cellulose Nanocrystal Network Structure in the

Presence of Salt in Aqueous Media8

The change in ionic strength of cellulose nanocrystal (CNC) suspensions is shown to contribute to

a respective change in colloidal behavior, such as stiffness and fractal gelation. In this study,

dynamic colloidal behavior and stability of aqueous CNC suspensions and their correlation with

nonlinear viscoelastic properties of the CNC gel structures in the presence of different

concentrations of sodium chloride (NaCl) salt were investigated. The microstructure of CNC/salt

suspensions/gels were investigated with a wide range of characterization technique. To obtain

further insight into the network structure of CNC/salt systems, for the first time, nonlinear rheology

of the suspensions/gels was analyzed to correlate macro-mechanical viscoelastic response of the

CNC/salt aqueous systems to structural changes as a response to strain. The intra-cycle

viscoelasticity explained utilizing qualitative Lissajous-Bowditch plots and quantitative nonlinear

parameters, demonstrates a strong dependence of the nonlinear response of the samples to salt

concentration, CNC concentration, and frequency of deformation. Higher intra-cycle nonlinearity

was observed upon increasing salt loading.

Graphical abstract

8 Characterization of Charged Cellulose Nanocrystal Network Structure in Presence of Salt in Aqueous Media, 1-15,

2020. Cellulose, In press.

ABM did the experimental design, data collection and interpretation, and manuscript preparation. MK contributed in

rheological measurement and interpretation.

80

4-1 Introduction

Cellulose nanocrystals (CNCs) are whisker shape particles [213]. CNCs, also besides being slender in

shape, have changeable levels of crystallinity [213]. To make a three-dimensional (3-D) structure out

of CNC, we need to gelify CNC particles. Aside from size and shape considerations, the Van der Waals

forces, surface charges, and hydrophobic and hydrogen interactions are other parameters playing

roles in governing the gelation of CNCs [138]. Experiments have shown that electrostatic forces

that keep particles separated can be adjusted through changes in surface charge density, such as

desulfation or annealing at high temperatures [43]. Accordingly, strengthening Van der Waals

interactions among CNCs over electrostatic repulsive forces can induce phase separation [173].

Entering to a very concentrated regime, through increasing CNC loading above 10 wt%, has also

been shown to induce concentration-dependent aggregation [214].

Similarly, self-similar structures of CNC gels can be formed by the addition of coagulants

such as salts or polymers into CNC suspension. Cherhal et al. [33] presented a study in which gel

formation happened after the introduction of NaCl into CNC suspension. In another study, Chau

et al. [48] experimentally showed that after increasing the ionic strength of suspensions, the

electrostatic repulsion among particles becomes weaker compared to attractive short-range forces

such as Van der Waals and hydrogen bonding. It was also claimed that stiffness of the gel is a

function of the charge number of salt and radii of the introduced ions. In another study, Uren a-

Benavides et al. [174] conjectured that the phase transition point for a CNC suspension coagulated

with ions occurs approximately one order of magnitude lower than the threshold of gelation for

pure CNC suspension. Despite the recent reports on the stability of various types of nano cellulose

1

2

b

1

2

b

200 µm 200 µm

1

2

b

1

2

b

200 µm

200 µm

200 µm 200 µm

Incre

asin

g S

alt

Co

nc.

81

[33, 147, 148, 151, 179, 182-185], to the best of our knowledge, no systematic study has been

reported in the literature targeting colloidal behavior and stability of hydrogels reported here.

More importantly, given one key application of CNC hydrogels in making scaffolds for tissue

engineering, rheological analysis of CNC suspensions and mechanical behavior of the formed gels

under external deformation is of paramount importance. For implanted scaffolds, mechanical

properties of hydrogels are vital for providing enough mechanical support to cells, particularly in

load-bearing tissues. The stiffness of hydrogels as the substrate for cell growth and the stresses

generated from the surrounding environment influences the fate, growth, and migration of different

cells [112, 113]. The fate of multipotent stem cells was shown to be dependent on relaxation and

retardation times of the scaffold [114]. Moreover, most tissues do not operate under a linear elastic

regime due to the heterogeneity and anisotropy in their ingredients, such as a combination of

different cell types and their distribution, directional expansion of cells, and composition of

extracellular matrices and structural proteins. Thus, characterizing the linear and nonlinear

viscoelastic behaviors of hydrogels is a key factor for engineering and implantation of tissues

[215]. Viscoelastic properties are impactful for hard tissues like bone [115], especially at low strain

rates and within the normal range of body frequency. To know how flawlessly the scaffold material

mimics the tissue being regenerated, their frequency and strain-dependent mechanical properties

should be evaluated in detail.

The linear viscoelastic rheological properties of CNC/salt gels have been previously studied.

Lenfant et al. [137] studied the linear viscoelastic response of electrically stabilized CNCs in the

presence of sodium and calcium ions. It was shown that CNC suspensions coagulate into gels at

20 mM salt concentration, whereas electrically stabilized CNC suspensions could tolerate a much

higher amount of salt prior to coagulation. In another study, Shafiei-Sabet et al. [136] reported that

for isotropic CNC suspensions, increasing the ionic strength of the system up to 5 mM NaCl via

weakening the electro-viscous effects and thus reduces the viscosity of CNC suspensions. Despite

the vast literature on the linear characterization of CNC/salt gels, the nonlinear viscoelastic

rheological properties of CNC/salt gels are yet largely unknown.

Hence, in this work, we performed an in-depth investigation of dynamic colloid behavior

and stability of CNC gelation as well as characterizing the mechanical response of CNC/salt

suspensions under different deformations (ranging from small to medium and large deformations).

82

The results revealed new aspects of CNC/salt systems that are not accessible via linear rheology

analysis. Viscoelastic properties of the CNC-salt suspensions are classified into two categories of

inter- and intra-cycles. We also used imaging techniques to interpret the correlation of rheological

properties with the stability of CNC/salt suspensions.


4-2-1 Materials

InnoTech Alberta was the provider of the CNC with reported length in the span of 100-200 nm

and a diameter of 5-15 nm. Based on the manufacturer datasheet, CNCs were extracted with acid

hydrolysis process, which causes negative charges to appear on CNCs.


CNC powder in DI water was sonicated, and a stock suspension with 3 wt% CNCs was prepared.

Suspensions were made with a pH of 6.8, measured using a Mettler Toledo Seven Compact pH-

meter (Mettler-Toledo 135 International Inc., Columbus, OH, USA). The ionic strength of the

suspension was adjusted via the addition of 200 mM NaCl. To make the final suspension with

desired concentrations, the samples after dilution with DI water were sonicated for 20 min. Ultra-

sonication (125 W Qsonica Sonicators Q125 Sonicator, Qsonica) was employed for suspending

CNCs in DI water. The ice bath treatment was done to prevent overheating on the surface of CNCs

as the surface charge of CNC particles is sensitive to temperature [50].


4-2-3-1 Scanning electron microscopy

The gel structure of CNC clusters and micro-morphology of the developed gel was evaluated using

XL30 Philips SEM. Prior to imaging with the SEM, the generated gels were freeze-dried using

83

liquid nitrogen. A small piece of the freeze-dried gel was mounted onto a silicon wafer. A layer of

gold was then sprayed on the samples, to limit electrostatic discharge.

4-2-3-2 Zeta potential measurements

Nano-Zetasizer (Malvern Instruments, Nano ZS, Malvern, UK) was employed to measure the zeta

potential and size of CNC particles suspended in DI water.

4-2-3-3 Rheology

Rheological measurements were performed using an Anton-Paar MCR 302 rheometer equipped

with a 50 mm diameter cone-plate geometry (cone angle of 1° and truncation of 101 μm). To reach

the desired stabilized morphology, a resting time of 10 min was used in the rheometer following

the CNC loading. The strain amplitude sweep experiment was carried out within the range of 0.1

to 1000.0 % and at an angular frequency of 1 rad/s on all samples to determine the linear

viscoelastic regime (LVR). Based on the results of this experiment, the strain amplitude of 1.0 %

was determined to be small enough to keep the deformation in the LVR. All experiments were

carried out at room temperature.

The rheometer was placed in a rigid and mechanically stable environment to minimize

mechanical noises and apply large amplitude oscillatory shear (LAOS) to the samples. To obtain

full waveform of shear stress and strain, the material was strained at constant frequency and

amplitude. LAOS data were collected after 5-6 cycles for each strain amplitude. Rheological tests

were performed on triplicates and quadruplicates for each sample.

4-2-4 Background

LAOS is a technique with expanding popularity among researchers [124, 126]. This approach is

destined to mark the onset of nonlinearities in complex materials. The LAOS tests involve

oscillation cycles at multiple strain amplitudes.

In the LAOS region, the sinusoidal input strain waveform is translated to a non-sinusoidal stress

response. There are various approaches to analyze the non-sinusoidal stress response, such as

Fourier transform rheology (FT- rheology) [127] and stress decomposition [128] methods. The

shear stress (𝜎) can be inscribed as in-phase and out-of-phase components of a time-domain

84

Fourier series of odd harmonics[129] being in steady-state condition for an oscillatory input strain

(𝛾(𝑡) = 𝛾0sin (𝜔𝑡)):

𝝈(𝒕) = 𝜸𝟎 ∑ [𝑮𝒏′ (𝝎, 𝜸𝟎) 𝐬𝐢𝐧 𝒏𝝎𝒕𝑵

𝒏=𝟏 + 𝑮′′𝒏(𝝎, 𝜸𝟎) 𝐜𝐨𝐬 𝒏𝝎𝒕] 4-1

In the above equation, 𝛾0 is strain amplitude, and 𝐺𝑛′ and 𝐺𝑛

′′ are amplitudes of n harmonics with

frequencies (nω). In the linear viscoelastic framework, the output stress waveform is the only

function of the first harmonic coefficients, 𝑛 = 1. The emergence of higher harmonics in the

resulting stress waveform depicts the appearance of nonlinear viscoelastic response, meaning that

the stress signal cannot be displayed by a simple sinusoidal waveform any longer. Furthermore,

𝐺′and 𝐺" lose their physical meaning in the nonlinear region, meaning another technique should

be implemented to explain the output stress signal.

FT- rheology is developed based on a sophisticated mathematical framework that is a powerful

technique to spot nonlinearities and higher-order harmonics in the stress waveform. However, it

cannot give a clear physical interpretation of higher-order harmonics and the resulting nonlinear

behaviors[130]. So, this method is insufficient to describe the material response. In 2008, Ewoldt

et al.[130] proposed novel measures based on the stress decomposition method introduced by Cho

et al.[128] in 2005 to give meaning to LAOS results.

Based on symmetric arguments proposed by Cho et al.,[128], the generic nonlinear stress

response (𝜎(𝑡)) can be decomposed into superposition of elastic and viscous stresses as below:

1) elastic stress component (𝜎′) as an odd function of normalized strain (𝑥(𝑡) = 𝛾(𝑡)

𝛾0 ),

2) viscous stress component (𝜎′′) as an odd function of the normalized strain rate (𝑦(𝑡) =

�� (𝑡)

��0). Thus, the total resulting stress can be described as the following:

𝜎(𝑡) = 𝜎′(𝑡) + 𝜎′′(𝑡). 4-2

Afterward, Ewoldt et al.[130] suggested a polynomial regression fit to the elastic (𝜎′) and viscous

(𝜎′′) lines. In their work, they argued the limitation of different polynomial basis functions, such

as Jacobi, Laguerre, Hermite, Chebyshev of the first and second kind, and Legendre. Considering

the mathematical and physical limitations, e.g., elastic(𝜎′) and viscous (𝜎′′) stresses are

85

orthogonal over a finite domain, they proposed that the set of Chebyshev polynomials of the first

kind is the best choice for fitting the output stress contributions. Then, the authors established a

physical interpretation of nonlinear viscoelasticity using Chebyshev coefficients.

Based on this method, a series of Chebyshev polynomials of the first kind in the orthogonal

space made up of the input strain and strain-rate can be used to represent the elastic 𝜎′ and viscous

𝜎′′stress components via the following equations:

𝜎′(𝑥: 𝜔, 𝛾0) = 𝛾0 ∑ 𝑒𝑛(𝜔, 𝛾0) 𝑇𝑛(𝑥), 4-3

𝜎′′(𝑥: 𝜔, 𝛾0) = 𝛾0 ∑ 𝑣𝑛(𝜔, 𝛾0) 𝑇𝑛(𝑦) 4-4

where �� =𝑥

𝛾0=

𝛾

𝛾0 and �� =

𝑦

𝛾0=

𝛾

��0

depicts the normalized version of strain and strain-rate, and 𝑇𝑛

symbolizes Chebyshev polynomials. ‘‘𝑒’’ and ‘‘𝑣’’ are elastic and viscous contributions and have

units of modulus (Pa) and viscosity (Pa.s-1), respectively.

The criteria for specification of the physical interpretation of the nonlinearity based on “𝑒”

and “𝑣” is defining the concavity of 𝜎′and 𝜎′′. As the magnitude of each Chebyshev coefficient

decays monotonically by increasing “𝑛”, the third-order Chebyshev coefficients (𝑒3 and 𝑣3)

determine the concavity of the elastic and viscous stress curves. According to these coefficients,

the following intra-cycle nonlinear behaviors can be observed: strain-stiffening (𝑒3 > 0), strain-

softening (𝑒3 < 0), shear-thickening (𝑣3 > 0) and shear-thinning (𝑣3 < 0)[130]. Moreover, the

nth-order Chebyshev coefficient and Fourier coefficients can be related to each other via the

following equations [130]:

𝑒𝑛 = 𝐺𝑛′ (−1)(𝑛−1)/2 4-5

𝑣1 =𝐺𝑛

′′

𝜔= 𝜂𝑛

′ 4-6

In the nonlinear regime, the measured dynamic moduli (𝐺1′and 𝐺1

′′) do not have a clear physical

meaning. Hence, using ‘’𝑒’’ and ‘’𝑣’’, Ewoldt et al.[130] defined local viscoelastic moduli and

viscosities to interpret the distorted stress signal. Hence, comparing the local viscoelastic moduli

(i.e., large-strain modulus (𝜎

𝛾|

𝛾=±𝛾0

≡ 𝐺𝐿′ ) and minimum-strain modulus (

𝑑𝜎

𝑑𝛾|

𝛾=0≡ 𝐺𝑀

′ )) can assist

to interpret intra-cycle elastic nonlinear behavior [130]. It is noted that both 𝐺𝑀′ and 𝐺𝐿

′ converge

to linear elastic modulus in the linear viscoelastic region, i.e., 𝐺𝑀′ =𝐺𝐿

′ =𝐺1′=𝐺′. Table 4.1 ands

Table 4.2 help to determine the nonlinearity based on the defined parameters. These elastic

86

measures have been used by Ewoldt et al. [130, 131] to develop a dimensionless index for

interpretation of intra-cycle elastic nonlinearity defined as:

S≡𝐺𝐿

′ −𝐺𝑀′

𝐺𝐿′

4-7

S (strain stiffening ratio) value equal to 0 corresponds to linear viscoelastic response, a positive S

indicates intra-cycle strain-stiffening behavior, and a negative S is indicative of intra-cycle strain-

softening. Like the above-mentioned elastic measures, viscous parameters have been introduced

to characterize intra-cycle viscous nonlinearity. In this context, a set of local dynamic viscosities

have been defined as minimum-rate dynamic viscosity 𝑑𝜎

𝑑��|

��=0≡ 𝜂′𝑀 and large-rate dynamic

viscosity 𝜎

��|

��=±��0

≡ 𝜂′𝐿[130, 131]. Similar to the elastic measures, in the linear regime, dynamic

viscosities converge to the linear real viscosity value 𝜂′ =𝐺"

𝜔, i.e., η'L=η'M=η'. The dimensionless

index for dissipative (viscous) intra-cycle nonlinearity has been proposed as:

T≡𝜂𝐿

′ −𝜂𝑀′

𝜂𝐿′

4-8

T=0 signifies linearity, T>0 implies intra-cycle shear-thickening, and T<0 corresponds to intra-

cycle shear-thinning behavior. It should be born in mind that there are other methods and

approaches, such as the sequence of physical processes [132] and intrinsic nonlinearity [133, 134],

which researchers used to interpret nonlinear data. Compared to the mentioned methods (e.g., FT-

rheology), the method that we used in this work provides us the physical interpretation of

nonlinearity with the aid of unambiguous material measures, which quantify nonlinear elastic and

viscous behavior, simultaneously. Thus, this method provides us with more substantial information

regarding the mechanism governing the microstructural changes under LAOS flow.


4-3-1 Morphological characterization of freeze-dried hydrogels under SEM

Figure 4.1 shows SEM images of instantly freeze-dried samples of 20 g/L and 30 g/L CNC mixed

with the salt content of 42.7 mM, 85.5 mM, and 172 mM, imaged at 250x and 1000x

magnifications. It appears that the morphology of samples varied at different salt and CNC

87

contents. Similar morphological changes were observed using confocal images (Figure 4.7). The

sample with 20 g/L CNC and 42.7 mM salt shows the finest pore size (Figure 4.1a). The salt

particles appeared brighter in backscattered images of the samples, have a homogenous

distribution across the images. The nature of brighter particles was verified with Energy-

Dispersive Spectroscopy (EDS), confirming the composition of sodium and chloride elements (not

shown here). Alternatively, the effect of salts on agglomeration and onset of the gelation in the

colloids can be quantified using transmission electron microscopy (TEM) images (readers are

referred to our previous work[173]).

b c

d e f

g h i

a

100 µm 50 µm

50 µm

50 µm

100 µm

50 µm

100 µm

100 µm

100 µm

100 µm

100 µm

50 µm

50 µm

50 µm

50 µm

88

Figure 4.1 Scanning electron microscopy (SEM) images of (a-b) 20 g/L CNC with 42.7 mM salt at 250x and 1000x

magnifications, respectively, (c-d) 20 g/L CNC with 85.5 mM salt, (e-f) 20 g/L with 172 mM salt, (g-h) 30 g/L with

42.7 mM salt, (i-j) 30 g/L CNC with 85.5 mM salt, (k-l) 30 g/L with 172 mM salt at 250x and 1000x magnifications,

respectively. All samples were freeze-dried out of 10 mL gelled suspension.

At the lower concentration (42.7 mM) of salt for both concentrations of CNC, phase separation

seems to occur at initial stages. It also stands out that for different concentrations of CNC (20 vs.

30 g/L), studied using SEM, the initial morphology development is different.

The analysis of TEM images of the same system (NaCl-CNC) [173] using the box-counting

method shows that the fractal dimension of CNC clusters varies from approximately 1.48 ± 0.06

to 1.87 ± 0.01 when the NaCl/CNC concentration ratio is increased from 0.05 to 0.5. Results of

Figure 4.1 shows that at joints, the fractal dimension follows the same trend of having pointier

junctions versus more compact junctions at higher salt concentrations. Variation in system ionic

strength causes a shift in the coagulation mechanism from diffusion-limited growth regime (DLA)

to reaction limited growth regime (RLA), as highlighted in our previous work [216]. CNC-salt

system is dynamic at the micro-level and undergoes compaction and sedimentation depending on

salt and CNC concentrations in a process, also known as coarsening. The effect of coarsening on

the fractal dimension has been studied through numerical simulations by Conti et al.[217] where

coarsening length scale and interfacial area of the fractal cluster have a power-law dependency on

time while the mass fractal dimension is shown to stay invariant.

A simple pore size measurement based on image J software also shows that the porosity changes

on average for CNC 20 g/L from 9.6 µm, 27 µm and 27.3 µm when salt loading changes from 42.7

mM to 172 mM, respectively. For CNC 30 g/L samples, porosity changes on average from 11.6

µm, 26 µm, and 27.8 µm when salt loading changes from 42.7 mM to 172 mM, respectively. These

results show that with an increase in salt concentration in both CNC loadings, the average porosity

j k l

50 µm

100 µm

50 µm

100 µm

50 µm

50 µm

89

increases. For these calculations, in the case of non-spherical pores, the first area was measured,

and then diameter was estimated accordingly.

In addition to the result of SEM, we monitored the evolution of CNC-salt structure using confocal

laser microscopy, where the results are presented in Figure 4.7 and Figure 4.8.

4-3-2 Zeta potential measurement

The zeta-potential was measured for CNC/salt suspension samples with the CNC concentration of

0.5 g/L and salt concentrations ranging from 0 to 102.7 mM (Figure 4.2). Since the gelation is

influenced by time, the zeta potential measurements were performed on the samples for about one

hour following the introduction of salt into the suspensions. The CNC particles in the absence of

any salt showed zeta potential values of −64 mV, akin to values obtained by Boluk et al. [147] and

Shafiei-Sabet et al. [136]. The formation of a diffuse layer and retraction of a double layer around

the particles is the primary reason for observing the descending trend of zeta potential as a result

of the increase in the salt concentration [203]. These results evidently show that the aggregation is

expected in the salted system, even in the presence of a small amount of salt concentration.

Since the system becomes cloudy at higher CNC concentrations, it is challenging to evaluate

the gel structure microscopically. However, similar changes in zeta potentials are expected for

higher CNC concentrations. The charge and mobility of clusters could be estimated by

electrophoresis measurement. Nonetheless, clusters are suspended in a poly-disperse ocean of

monomers and clusters, which complicates measuring the electrophoretic mobility of the mixture.

The discussion on the mobility of clusters and CNC monomers is out of the scope of this work and

reader are referred to the work of Groenewold et al. [218].

90

Figure 4.2 Changes in zeta potentials of CNC suspensions at a fixed concentration of 0.5 g/L CNC and as a function

of NaCl concentrations.

4-3-3 Linear viscoelastic behavior of CNC suspensions

The linear viscoelastic behavior of CNC solutions was evaluated using small amplitude oscillatory

shear (SAOS) tests (please see Figure 4.9 and Figure 4.10). The results of the linear viscoelastic

response confirm the gel-like behavior of the CNC aqueous systems in the presence of salt.

Moreover, comparing Figure 4.9 and Figure 4.10 reveals that increasing either CNC

concentration or salt concentration leads to an increase in the value of linear rheological parameters

(e.g., storage modulus (G'), loss modulus (G''), and complex viscosity (|𝜼*|)), verifying the

formation of a stronger microstructure. This is in line with imaging results, confirming gelation by

CNC clusters formation in the presence of salt. These results will be used in order to explain

nonlinear data.

4-3-4 Inter-cycle nonlinear viscoelastic behavior of CNC suspensions

Unlike SAOS flow, large amplitude oscillatory shear (LAOS) tests are not restricted to a narrow

strain range and, thus, provide further insights into the network of the suspensions or gels.

Moreover, LAOS experiments unravel the underpinning physical processes responsible for the

91

failure of the microstructure [219-224]. Figure 4.3 shows the strain amplitude dependence of

dynamic moduli (G and G) of the 20 g/L CNC samples containing different amounts of salt.

Figure 4.11 compares the strain dependency of the dynamic moduli of suspensions at different

CNC concentrations. Both storage and loss moduli are independent of the input strain amplitude

in the linear viscoelastic region (LVR) (Figure 4.3). In line with Figure 4.9 and Figure 4.10, the

value of the plateau moduli in LVR increases with increasing salt content. Moreover, the G > Ga

condition in small amplitudes indicates the dominance of the solid-like behavior for the 20 g/L

CNC samples as a result of gelation in the presence of salts. However, beyond the critical strain

amplitude c, i.e., strain at which linear to nonlinear viscoelastic behavior occurs, G features a

dramatic drop while G experiences a slight increase to reach a maximum value followed by a

dramatic decrease (overshoot).

Figure 4.3 Oscillatory amplitude sweep response of CNC 20 g/L suspensions containing different amount of salt

((a) 1.72, (b) 17.2, (c) 34.4, (d) 85.5, (e) 172 mM) for strain amplitudes of 𝛾0=0.1-1000% at an angular frequency

of 𝜔 =1 rad/s using a cone-plate geometry (at a truncation of 101 μm and cone tip angle of 1°) at 25˚C. (f) Critical

strain amplitude c (linear to nonlinear transition) and crossover strain amplitude T (solid to liquid transition).

0.01

0.1

1

10

100

1000

10000

0.1 1 10 100 1000

a

G"

G'

0.01

0.1

1

10

100

1000

10000

0.1 1 10 100 1000

b

0.01

0.1

1

10

100

1000

10000

0.1 1 10 100 1000

c

0.01

0.1

1

10

100

1000

10000

0.1 1 10 100 1000

d

0.01

0.1

1

10

100

1000

10000

0.1 1 10 100 1000

e

0 [%]

Dyn

am

ic M

od

uli

[P

a]

0

20

40

60

80

100

120

0

20

40

60

80

100

120

0 50 100 150 200

T

(%) c

(%

)

Salt Con. (mM)

f

92

One can expect four different types of viscoelastic nonlinearity upon exceeding the limit of the

linear viscoelastic region. Type I stream (strain-thinning) that is a common behavior observed in

polymer melts and solutions and occurs where both G and G decrease in the nonlinear region as

a result of reduced local drag by the alignment of network segments with the flow field; Type II

stream (strain-hardening) occurs where both G and G increase; Type III stream (weak strain

overshoot) occurs where G decreases while G first increases and then decreases; Type IV stream

(strong strain overshoot) occurs where G and G first increase followed by decreasing.

The drop in storage modulus of the CNC/salt samples at increasing strain amplitude is conjugated

with an increase in loss modulus, which confirms that CNC 20g/L-salt samples follow the type III

nonlinear viscoelastic behavior (Figure 4.3). The same behavior was observed for CNC 30g/L-

salt suspensions/gels (Figure 4.11). After the weak overshoot of G, the rate of the decrease in G

is more severe in comparison to G (Figure 4.3). Therefore, at crossover strain amplitude (T) at

each salt concentration, G becomes smaller than G, revealing solid-to-liquid transition as a result

of breakage in the CNC network structure at sufficiently large deformations.

Chen et al. [225] studied the rheological behavior of nanocrystalline cellulose (NCC) in the

aqueous solution of poly (vinyl alcohol) (PVA) with a flexible chain structure and carboxymethyl

cellulose (CMC) with semi-rigid chain structure. The authors observed type III behavior for

NCC/PVA systems while NCC/CMC showed type I behavior. They claimed that type III behavior

of NCC/PVA systems originates from flocculated structures of NCC particles by adsorption of

PVA chains and bridging effect. However, in our CNC systems, it is believed that the weak

overshoot is because of the formation of weak structural complexes in medium amplitude region

(shear-induced structures) [226]. This could be attributed to the decrease in the inter-particle

distances in the suspension/gel systems by applying the shear force, facilitating the process of

shear-induced network formation, leading to a larger number of load-bearing junctions (physical

bonds, e.g., ionic forces) during the transition into the nonlinear regime of deformation. Moreover,

increasing the content of the salt shifted the location of both c and T to lower strain amplitudes

(Figure 4.3f). This is attributed to the formation of a much stronger network upon increasing the

salt content, which yields at smaller deformations [219, 220]. In fact, increasing both salt

concentration and CNC concentration causes a shift in critical strain from linear to nonlinear

regimes to lower values. The dependency of elasticity of the structure has also been observed in

93

colloidal glasses[227]. This is attributable to the increase in attraction strength of bonded particles

responsible for bridging clusters [228].

Fractal dimension is a mathematical parameter that represents the compactness of clusters that

fabricates the overall fractal gel. In the morphology section, by comparing TEM images, the

connection between salt/CNC ratio and fractal dimension was deduced. It was found that increase

salt/CNC ratio makes the clusters more compact. Upon comparing trends in which critical strain

increases with the salt/CNC ratio, it can be conjectured that the fractal dimension is impactful on

the critical strain. As bond breakage in a gel starts with particles that link the clusters[227, 228]

together in a gel, it is expected that the geometry of the inter-cluster structures is important. In fact,

Shih et al. [229] proposed a scaling model relating the rheological properties of viscoelastic

systems to particle concentration and particularly the fractal dimension. Based on their definition,

in a strong-link regime, inter-cluster links are stronger than the intra-cluster links. They envisioned

the gelation process and the emergence of a solid-like rheological behavior in the frame of

aggregating fractal flocs. The model proposed by Shih et al. [229] connects mathematically c to

df. It is worth mentioning that align with Shih et al. [229] model, our data shows that increasing

concentration of CNC at constant salt values increases the elasticity of the system.

Wu et al.[230] proposed a scaling model to relate the structure of the colloidal gels to their elastic

parameters. In this context, the authors utilized storage modulus and the limit of linearity of

different systems to validate their rheological model. They defined the limit of linearity as “the

situation where the weakest bonds break, and the linear elastic behavior vanishes” and showed

that by increasing the loading of the nanofillers, the limit of the linearity increased for the weak-

link regime and decreased for the strong-link regime. Hence, considering the results of Figure 4.3f

and the above discussion, it can be concluded that CNC (20g/L) /salt systems can be considered

as a strong-link gel.

To get further insight into the network structure of the CNC/salt systems from the nonlinear

rheological point of view, we studied the intra-cycle viscoelastic parameters in the next sections.

4-3-5 Intra-cycle nonlinear viscoelastic parameters

94

In previous sections, we provided information regarding the inter-cycle viscoelastic behavior of

the CNC/salt systems in both linear and nonlinear viscoelastic regions. To complement the

nonlinear viscoelastic characterization of CNC/salt suspensions/gels and to find the origin of the

nonlinearity, the intra-cycle viscoelastic behavior of the samples is studied below. To our

knowledge, this is the first study systematically exploring the intra-cycle behavior of CNC/salt

systems at various CNC concentrations, salt concentrations, and frequencies.

Any intra-cycle nonlinearity can be discerned by excitation of higher harmonics in the output

stress waveform, leading to the divergence of local viscoelastic moduli (𝐺𝑀′ and 𝐺𝐿

′ ) and local

dynamic viscosities (𝜂𝑀′ and 𝜂𝐿

′ ) from each other. Figure 4.4 shows local viscoelastic moduli (𝐺𝑀′

and 𝐺𝐿′ ) and local dynamic viscosities (𝜂𝑀

′ and 𝜂𝐿′ ) as a function of input strain amplitude for CNC

20 g/L at two different salt contents (see Figure 4.11 for CNC 30 g/L suspensions/gels results).

Figure 4.4 Nonlinear viscoelastic measures of CNC 20 g/L-salt as a function of strain amplitude at an angular

frequency of 𝜔 =1 rad/s obtained using a cone-plate geometry (at a truncation of 101 μm and cone tip angle of 1°)

at 25˚C. (a) Dynamic viscosities (𝜂𝑀′ and 𝜂𝐿

′ ) and (b) local viscoelastic moduli (𝐺𝑀′ and 𝐺𝐿

′ ) for CNC 20 g/L- 85.5

mM salt. (c) Dynamic viscosities (𝜂𝑀′ and 𝜂𝐿

′ ) and (d) local viscoelastic moduli (𝐺𝑀′ and 𝐺𝐿

′ ) for CNC 20 g/L- 172

mM salt.

1

10

100

1000

10000

1 10 100 1000G' L

an

d G

' M[P

a]

0 [%]

Series1

Series2

G'MG'L

1

10

100

1000

10000

1 10 100 1000G' L

an

d G

' M[P

a]

0 [%]

1

10

100

1000

1 10 100 1000

' L

an

d

' M [P

a.s

]

0 [%]

Series1

Series2

'M 'L

1

10

100

1000

1 10 100 1000

' L

an

d

' M [P

a.s

]

0 [%]

a)CNC 20 g/L- 85.5 mM salt c)CNC 20 g/L- 172 mM salt

b)CNC 20 g/L- 85.5 mM salt d)CNC 20 g/L- 172 mM salt

95

Based on the results of Figure 4.3 and Figure 4.4, 𝐺𝑀′ and 𝐺𝐿

′ are equal and converge to the linear

storage modulus G in the LVR. Upon the initial departure of 𝐺𝑀′ and 𝐺𝐿

′ from LVR beyond the

critical strain amplitude, both 𝐺𝑀′ and 𝐺𝐿

′ decrease, indicative of inter-cycle strain softening

behavior (consistent with Figure 4.3). However, the rate of the decrease in 𝐺𝑀′ and 𝐺𝐿

′ is different,

leading to emergence of intra-cycle elastic nonlinearity. In this regard, the strain-stiffening ratio

(S) is employed to recognize the type and extent of intra-cycle elastic nonlinearity with a better

resolution.

Similarly, the relative behavior of 𝜂𝑀′ and 𝜂𝐿

′ provides valuable information about inter- and intra-

cycle dissipative feature of the suspension systems. Both 𝜂𝑀′ and 𝜂𝐿

′ are close to each other in LVR

and are strain independent. However, located in the nonlinear region, not only 𝜂𝑀′ and 𝜂𝐿

′ are no

longer equal but also behave differently (Figure 4.4a and c), making the interpretation of intra-

cycle dissipative (viscous) behavior of the samples challenging. Hence, for a simpler

representation of the viscous intra-cycle nonlinearity, we utilized the shear-thickening ratio (𝑇).

The type of intra-cycle nonlinearity can be easily discerned using S and T indices (see Table 4.1and

Table 4.2). Figure 4.5 shows these parameters at different deformations for CNC 20 g/L

suspensions and at two different salt contents (Figure 4.12 shows similar results for CNC 30 g/L).

In concert with Figure 4.3 and Figure 4.4, S and T are roughly zero in a linear framework, meaning

that no intra-cycle nonlinearity occurred in CNC/salt systems. S index increased and became

positive and showed a maximum value (Smax) in medium amplitude oscillatory shear (MAOS)

region (0 = 10-100%), followed by a dramatic decrease upon further increasing the deformation

towards the maximum strain amplitude (0 = 1000%). Therefore, the intra-cycle elastic

nonlinearity of CNC/salt suspensions is strain-stiffening behavior, whereas their inter-cycle elastic

nonlinearity was strain-softening (decreasing trend in both 𝐺𝑀′ and 𝐺𝐿

′ , see Figure 4.3 and Figure

4.4).

96

Figure 4.5 Elastic (S) and viscous (T) intra-cycle nonlinearity indices as a function of strain amplitude for CNC 20

g/L suspensions at 85.5 mM and 172 mM salt contents, measured using a cone-plate geometry (at a truncation of

101 μm and cone tip angle of 1°) at 25˚C and angular frequency of ω = 1 rad/s. Schematics show the state of the

systems in each regime.

Increasing the salt content shifted the value of S index to greater values (upward shift) in all the

probed strain amplitude window (i.e., more pronounced intra-cycle strain-stiffening behavior). The

intra-cycle strain-stiffening behavior in medium strain amplitudes mainly stems from breakage

and compression of the CNC agglomerates in response to the increasing deformation and shear

rate in one cycle. This collapse in the CNC agglomerates led to an increase in the surface area of

the CNC clusters which resulted in a shorter distance between the clusters and, consequently, the

formation of the higher number of active junctions contributing to the network [231] (shear-

induced network formation, this is in complete agreement with weak strain overshoot behavior

observed in Figure 4.3). This phenomenon takes place up to the strain amplitude at which Smax

occurs. Afterward, further addition of deformation destroys the 3D network by overcoming the

interactive forces among individual CNCs and CNC agglomerates and by the orientation of the

network elements, leading to a decrease in S value.

The drop in S value occurs approximately at a strain amplitude at which T becomes negative

(intra-cycle shear thinning behavior). The negative T index also shows a widespread rupture of the

network structure of CNCs and the reorientation of the rigid individual CNCs and agglomerates

parallel to the flow direction. However, prior to intra-cycle shear thinning behavior, the T index

slightly increases from roughly zero values to small positive values (intra-cycle shear-thickening

behavior) in the MAOS region (Figure 4.5). This behavior is more pronounced for CNC 30 g/L

-200

-100

0

100

200

-100

-50

0

50

100

1 10 100 1000

T [%]S [%]

0 [%]

85.5mM salt- 20g/L CNC-200

-100

0

100

200

-100

-50

0

50

100

1 10 100 1000

T [%]S [%]

0 [%]

172mM salt- 20g/L CNC

CNC Clusters

Densified CNC Clusters

Individual CNC

SAOS3D Network

MAOSDensified cluster

LAOSBroken network

and oriented clusters/CNCs

97

suspensions (Figure 4.12) and may be attributed to the breakage of the rigid CNC aggregates,

magnifying the dissipative feature observed during the transition into the nonlinear regime. In the

next section, we studied the effect of frequency of deformation on the intra-cycle viscoelastic

parameters.

4-3-6 Frequency dependence of intra-cycle viscoelastic parameters

The strain dependence of G and G for CNC 20 g/L-salt 85.5 mM at different angular frequencies

(ω = 0.5, 1.0, 5.0 and 10.0 rad/s) is shown in Figure 4.14. The CNC/salt suspensions showed type

III viscoelastic behavior at all frequencies of deformation (see Figure 4.14), indicating that the

type of the inter-cycle viscoelasticity of the CNC/salt systems is frequency invariant. However,

both S and T indices were significantly dependent on the frequency (Figure 4.6). By increasing

the frequency, the transition from intra-cycle shear-thickening to intra-cycle shear-thinning

behavior shifted to higher strain amplitudes (see Figure 4.6). Moreover, in the LAOS region

(0>100%), the absolute value of the T index decreased upon increasing the frequency. Also,

compared to the T index, the S index is shown to be more sensitive to the applied frequency (Figure

4.6). The value of Smax was approximately the same at frequencies of 0.5 and 1.0 rad/s. However,

upon increasing the frequency to 5.0 rad/s and 10.0 rad/s, Smax dramatically increased, and at larger

deformations, the intra-cycle strain-stiffening behavior switched to intra-cycle strain-softening

behavior. These results prove the extreme sensitivity of the intra-cycle nonlinearity of the CNC/salt

suspensions to the imposed frequency of the deformation. Since shear flow inhomogeneities might

have effects on flow kinematics and viscoelastic response at extremely large deformations

associated with high frequencies, the data points of S index at very large deformation (0>600%)

and large frequencies (5 rad/s and 10 rad/s) were not further discussed.

98

Figure 4.6 Elastic (S) and viscous (T) intra-cycle nonlinearity indices as a function of strain amplitude for CNC 20

g/L suspensions at 85.5 mM salt, measured using a cone-plate geometry (at a truncation of 101 μm and cone tip

angle of 1°) at 25˚C and angular frequency of (a) 0.5, (b) 1, (c) 5, and (d) 10 rad/s.

Based on the above discussions, studying the nonlinear viscoelastic parameters of CNC/salt

suspensions revealed a drastic difference between their inter- and intra-cycle viscoelastic

responses, which in turn can deliver more meaningful data regarding the microstructure of the

system. These results are not accessible via a simple strain sweep test, which presents purely

lumped viscoelastic functions providing an overall measure of inter-cycle nonlinearity. Moreover,

it was observed that intra-cycle nonlinear parameters are sensitive to any changes in microstructure

(e.g., increase in CNC or salt concentrations) and any changes in the condition of the tests (e.g.,

increase in the frequency of deformation). The same conclusion can be deduced based on

Lissajous-Bowditch plots (see Figure 4.15 and Figure 4.16 and the corresponding discussion).

4-4 Conclusion

This work provides a systematic study on the microstructure of cellulose nanocrystal (CNC) in

aqueous media in the presence of NaCl salt using a combination of imaging techniques and linear

and nonlinear rheological analyses. It is concluded that the addition of counter ion causes a sudden

phase separation in CNC-salt hybrid systems. The nonlinear rheology analysis was employed to

understand the relationship between the macro-mechanical response of the CNC/salt aqueous

systems and nano-scale structural properties. It is shown that nonlinear viscoelastic measures are

extensively sensitive to subtle changes in internal microstructures and yield valuable information

regarding the samples' network structure. For instance, the shape of the Lissajous-Bowditch plots

-200

-100

0

100

200

-100

-50

0

50

100

0.1 1 10 100 1000

T [%]S [%]

0 [%]

b

-200

-100

0

100

200

-100

-50

0

50

100

0.1 1 10 100 1000

T [%]S [%]

0 [%]

a

-200

-100

0

100

200

-100

-50

0

50

100

0.1 1 10 100 1000

T [%]S [%]

0 [%]

c

-200

-100

0

100

200

-100

-50

0

50

100

0.1 1 10 100 1000

T [%]S [%]

0 [%]

d

99

become more distorted upon increasing the salt concentration. Additionally, inter-cycle nonlinear

viscoelasticity of the CNC-salt samples shows type III viscoelastic behavior, while the elastic

intra-cycle viscoelasticity of the samples is strain-stiffening. Moreover, the intra-cycle viscoelastic

parameters demonstrated a strong dependency on the frequency of deformation. These cannot be

obtained with such clarity via rheological parameters obtained within the linear framework or a

simple strain amplitude test. The correlation between the microstructure and viscoelastic properties

under large deformations provides comprehensive guidance for the fabrication of high-

performance materials with precisely controllable microstructures and mechanical behaviors.

100

4-5 Supporting information (Chapter 4)


Micrographs captured by confocal laser scanning microscopy (CLSM) show gradual

microstructural changes in the suspension sample of 20 g/L CNC and with the addition of 0-172

mM NaCl (Figure 4.7). CNC suspensions in the absence of salt are homogeneous without any

agglomeration; thus the fluorescence intensity is almost constant for all pixels (Figure 4.7a). Upon

the addition of a small amount of salt (i.e., 42.7 mM), the first sign of aggregation appeared in the

composition (bright spots in Figure 4.7b). Further increase in salt concentration induces

perturbations to the system and covers the entire visualization cubes (Figure 4.7c and Figure

4.7d).

1

2

b

1

2

b

101

Figure 4.7 The growth of cellulose nanocrystal (CNC) network at 20 g/L CNC and at different contents of sodium

chloride (NaCl): (a) 0 mM, (b) 42.7 mM, (c) 85.5 mM, (d) 172 mM in deionized water (DI). The dimensions of the

visualization cube are 100×1000×1000 µm3. The three-dimensional (3D) confocal laser scanning microscopy

(CLSM) images are rotated to obtain a better view of the gel hybrid system. Images were taken immediately after

adding the salt into the mixture. Resolution: 500 nm

The green zones represent the existence of CNC-perturbed structures carrying FB 28 fluorescence

dye, whereas the dark zones indicate CNC-depleted zones. These results are in agreement with a

classical Derjaguin, Landau, Verwey, and Overbeek (DLVO) theory [199] in which the addition

of salt gradually empowers short-ranged van der Waals forces over long-ranged electrostatic

repulsive forces. The CNC gel structure grows rapidly beyond a threshold for salt concentration.

The lack of further variation in microstructure observed with the CLSM signifies that the fractalled

CNC phase has become insensitive to the variation in salt contents. Figure 4.8 shows similar

behavior for the CNC samples with 30 g/L concentration and in the presence of different salt

concentrations.

1

2

1

2

102

Figure 4.8 The growth of CNC network at 30 g/L CNC at different contents of NaCl salt: (a) 0 mM, (b) 42.7

mM, (c) 85.5 mM, (d) 172 mM salt. The dimensions of the visualization cube are 100×1272×1272 µm3. The

3D CLSM images are rotated to obtain a better view of the gel hybrid system. Images were taken immediately

after adding the salt into the CNC mixture. Resolution: 500 nm

4-5-2 Linear viscoelastic behavior of CNC suspensions

Figure 4.9 Linear viscoelastic characterization of CNC-salt solutions/gels for CNC concentration and at different

salt concentrations. (a) Storage modulus (G'), (b) Loss modulus (G''), and (c) Complex viscosity (|𝜼*|) of CNC/salt

solution at different salt concentrations for strain amplitudes of 1% using a cone-plate geometry (at a truncation of

101μm and cone tip angle of 1°) at 25˚C. (d) The storage modulus (G') versus salt concentration at an angular

frequency of 1 rad/s

Linear viscoelasticity of the CNC solutions was measured using oscillatory amplitude shear

(SAOS) tests (e.g., frequency sweep). The network structures and dynamics of CNC suspensions

that interact via pair potentials and experience Brownian motion are strongly interrelated to the

0.1

1

10

100

1000

10000

0.1 1 10 100 1000

G[P

a]

ω [rad/s]

a

1mg

5mg

10mg

20mg

50mg

100mg0.01

0.1

1

10

100

1000

10000

0.1 1 10 100 1000

G

[Pa]

ω [rad/s]

b

0.01

0.1

1

10

100

1000

10000

0.1 1 10 100 1000

| *

|[P

a.s

]

ω [rad/s]

b

1.72mM

8.5mM

17.2mM

34.2mM

85.5mM

172mM

0.01

0.1

1

10

100

1000

10000

0 50 100 150 200

G[P

a]

Salt Conc. (mM)

c

103

rheological parameters [232]. The results show that the storage modulus (G') of the CNC (20gr/L)-

salt system increases with the increase in salt concentration (Figure 4.9a), although the increase

in G' follows a descending trend ended up to a plateau (Figure 4.9d). Reaching to a plateau is an

indication of the formation of a strong gel. A similar behavior of G' dependency to salt

concentration was detected for the CNC gel with 30 gr/L CNC concentration (Figure 4.9). Also

changes in loss modulus (G'') against different salt concentrations follow the same trend as G'

(Figure 4.9b). The samples with low salt concentrations (<8.5mM) show a low G' since they are

in suspension/solution form still. However, for higher values of salt concentration, the system

depicts G'/G'' > 1 and is much less frequency dependent, which is usually an indication for the

formation of a self-supporting elastic gel (i.e., strong gel). These effects may be related to motility

inhibition among CNC particles at higher salt concentrations due to enabling of short attraction

forces over electrostatic forces.

Monitoring the complex viscosity (η*) of the samples as a function of salt concentrations shows

that the CNC/salt suspension systems have a shear thinning characteristic (Figure 4.9c). The

values of complex viscosity also increase with the increase in salt concentration.

The different behavior observed between two different salt concentrations can be traced back to

the discussion presented by Solomon et al. [142] where one can show why suspension shows more

elasticity after increase in salt loading. After adjustments in ionic strength of the medium, system

will reach a zone in which interaction between particles is stronger than thermal motion forces

separating them. Hence, inhibition of particle movement is expected. These inhibitions is due to

effects related to excluded volume [205, 206] and/or pair interactions between CNC pairs due to

attractive van der Waals or depletion forces [207]. Interactions between CNCs due to movement

of individual CNCs can also play a significant role [208, 209]. The increase in particle number

density or strength of attractive interactions limits the mobility of rods to the orders of rod radius.

After reaching this threshold, system individual components do not move, which also translates

into more elasticity and non-ergodicity. Fractal gels are also system in which particles stop moving

altogether due to physical bonds. The behavior of particle gels and glasses is different , which can

be observed by techniques such as dynamic light scattering [210].

104

4-5-3 The effect of Cellulose Nano Crystals (CNC) concentration of inter-cycle

viscoelastic behavior of CNC-salt suspensions

Figure 4.10 Oscillatory frequency sweep response of CNC solutions containing different amount of CNC (20 g/L

closed symbols and 30g/L open systems) and different salt concentrations at strain amplitudes of 𝜸𝟎= 1% using a

cone-plate geometry (with a truncation of 101μm and a cone tip angle of 1°) at 25˚C.

0.1

1

10

100

1000

10000

0.1 1 10 100 1000

G[P

a]

ω [rad/s]

a

20mg salt- 20gr/L CNC






34.2mM salt-20g/L CNC


172mM salt-20g/L CNC




0.1

1

10

100

1000

0.1 1 10 100 1000

G″

0 [%]

Dyn

am

ic M

od

uli

[P

a]

0.01

0.1

1

10

100

1000

10000

0.1 1 10 100 1000

Gʹ

20mg salt-20g/L CNC50mg salt-20g/L CNC100mg salt-20g/L CNC20mg salt-30g/L CNC50mg salt-30g/L CNC100mg salt-30g/L CNC

34.2mM salt-20g/L CNC85.5mM salt-

20g/L CNC172mM salt-

20g/L CNC34.2mM salt-30g/L CNC



105

Figure 4.11 Oscillatory amplitude sweep response of CNC solutions containing different amount of CNC (20

g/L closed symbols and 30 g/L open systems) and different salt concentrations for strain amplitudes of 𝜸𝟎= 0.1-

1000% at an angular frequency of 𝝎 = 1rad/s using a cone-plate geometry (with a truncation of 101μm and a

cone tip angle of 1°) at 25˚C.

Figure 4.12 Nonlinear viscoelastic measures () of CNC/salt suspensions with 30 g/L CNCs and two different

concentrations of a,c) 85.5 mM and b,d) 172 mM as a function of strain amplitude at an angular frequency of

𝝎 = 1rad/s using a cone-plate geometry (with a truncation of 101μm and a cone tip angle of 1°) at 25˚C.

4-5-4 Intra-cycle nonlinear viscoelastic parameters for 30g/L CNC suspension

0.1

1

10

100

1000

10000

1 10 100 1000

G' L

an

d G

' M [P

a]

0 [%]

1

10

100

1000

1 10 100 1000 ' L

an

d

' M [P

a.s

]0 [%]

1

10

100

1000

10000

1 10 100 1000

G' L

an

d G

' M [P

a]

0 [%]

1

10

100

1000

1 10 100 1000

' L

an

d

' M [P

a.s

]

0 [%]

85.5mM salt- 30g/L CNC 172mM salt- 30g/L CNC

-200

-100

0

100

200

-100

-50

0

50

100

1 10 100 1000

T [%]S [%]

0 [%]

85.5mM salt- 30g/L CNC-200

-100

0

100

200

-100

-50

0

50

100

1 10 100 1000

T [%]S [%]

0 [%]

172mM salt- 30g/L CNC

106

Figure 4.13 Elastic (S) and viscous (T) intra-cycle nonlinearity indices as a function of strain amplitude for

30g/L CNC solutions and at a) 85.5 and b) 172 mM salt using a cone-plate geometry (with a truncation of 101μm

and a cone tip angle of 1°) at 25˚C and angular frequency of ω=1rad/s.

4-5-5 Effect of frequency on inter-cycle response of 20 g/L CNC suspension

containing 85.5 mM salt

Figure 4.14 Oscillatory amplitude sweep response of CNC 20 g/L solutions containing at salt 85.5 mM for strain

amplitudes of 𝜸𝟎=0.1-1000% at different angular frequencies using a cone-plate geometry (with a truncation of

101μm and a cone tip angle of 1°) at room temperature.

4-5-6 Lissajous-Bowditch plots of CNC/salt suspensions at 20 g/L CNC

suspensions at different salt concentrations

The plot of instantaneous stresses τ(t) against strains γ(t) (often called Lissajous−Bowditch plots)

for different strain amplitudes (γ0) allows one to follow the actual network response to each loading

cycle during the oscillatory shear testing. For purely elastic materials in the linear region, the

response is expected to be completely in phase (δ = 0), and, as a result, the Lissajous−Bowditch

plot is presented in a line with a slope of G′. The response of purely viscous materials is expected

to be out of phase by δ = π/2, and the Lissajous− Bowditch plot becomes a circle (see our previous

works for more information about Lissajous plots [219, 220, 226]). Viscoelastic materials like

polymer networks exhibit both elastic and viscous properties, and, thus, the Lissajous−Bowditch

plot is expected to be a perfect ellipse in linear region, with the magnitude of the complex modulus

0.01

0.1

1

10

100

1000

10000

0.1 1 10 100 1000

ω=0.5rad/s

G'

G"

0.01

0.1

1

10

100

1000

10000

0.1 1 10 100 1000

ω=1rad/s

0.01

0.1

1

10

100

1000

10000

0.1 1 10 100 1000

ω=10rad/s

0.01

0.1

1

10

100

1000

10000

0.1 1 10 100 1000

ω=15rad/s

Dyn

am

ic M

od

uli [

Pa

]

0 [%]

107

|G*| as the slope of the semimajor axis. With the aid of Lissajous−Bowditch plots, one can visually

determine the emergence of nonlinearity, type of nonlinearity, and extent of nonlinear viscoelastic

behavior. Hence, Lissajous−Bowditch curves facilitate the qualitative assessment of nonlinear

behavior and deliver information about the type of intra-cycle nonlinear viscoelastic behavior.

In this regard, as mentioned earlier, the Lissajous plots are a simple ellipse in LVR, and this is

because the output stress signal of the viscoelastic materials only includes the first harmonic

coefficient in LVR. The emergence of any nonlinearity (due to the excitation of higher harmonics

in the output shear stress waveform) in the viscoelastic character of the material by surpassing

LVR can be identified by any distortion in ellipsoidal pattern of the Lissajous plots.

The Lissajous plots of the CNC 20 g/L suspensions at different salt concentrations are shown

in Figure 4.15 (Figure 4.16 shows the Lissajous loops of CNC 30 g/L at different salt

concentrations). For easier comparison, the stress, strain, and strain rate were normalized with

respect to their maximum values. As can be seen in Figure 4.15, the materials response in the LVR

(e.g., 0=1%) is a typical ellipsoidal Lissajous plots. However, by increasing the strain amplitude

to 0 = 40%, the Lissajous plots deviated from ellipsoidal shape and became distorted, signaling

the occurrence of intra-cycle nonlinearity in the suspensions. Moreover, the deviation in the

linearity of the Lissajous plots became more pronounced as a result of the increase in salt

concentration (reinforcing the network). This reveals the sensitivity of the Lissajous plots to any

changes in the microstructure. We observed similar nonlinear viscoelastic behavior in our previous

works, where samples with a stronger network demonstrated higher nonlinearity [219, 220].

108

Elastic Projection Viscous Projection

Figure 4.15 Dimensionless Lissajous-Bowditch loops for CNC 20 g/L solutions containing (a, b) 17.2, (c, d) 85.5,

and (e, f) 172mM salt, measured using cone-plate geometry (with a truncation of 101μm and a cone tip angle of

1°) at 25˚C. Projections on the elastic (𝝉 - 𝜸) and viscous (𝝉 - 𝒅𝜸

𝒅𝒕) planes are presented at strain amplitudes of 𝜸𝟎 =

1, 40, 100, and 250% and at an angular frequency of 𝝎 = 1rad/s

Interestingly, the viscous Lissajous plots of CNC 20 g/L suspension containing 17.2 mM salt

showed a self-intersection in the total shear stress curve at the strain amplitude of 0 = 100%. The

-1.5

-1

-0.5

0

0.5

1

1.5

-1.5 -1 -0.5 0 0.5 1 1.5

() n

orm

.

()norm.

c

-1.5

-1

-0.5

0

0.5

1

1.5

-1.5 -1 -0.5 0 0.5 1 1.5

() n

orm

.

(d/dt)norm.

d

-1.5

-1

-0.5

0

0.5

1

1.5

-1.5 -1 -0.5 0 0.5 1 1.5

() n

orm

.

()norm.

e

-1.5

-1

-0.5

0

0.5

1

1.5

-1.5 -1 -0.5 0 0.5 1 1.5

() n

orm

.

(d/dt)norm.

f

Salt

85.5

mM

Salt

172m

M

-1.5

-1

-0.5

0

0.5

1

1.5

-1.5 -1 -0.5 0 0.5 1 1.5

() n

orm

.

(d/dt)norm.

b

Salt

17.2

mM

-1.5

-1

-0.5

0

0.5

1

1.5

-1.5 -1 -0.5 0 0.5 1 1.5

() n

orm

.

()norm.

a

1%40%100%250%

109

self-intersection became more obvious at the strain amplitude of 0 = 250%. Several reports have

observed the emergence of self-intersection under LAOS deformation for different systems (e.g.,

polymer solutions and polymer nanocomposites) [220, 233]. Self-intersection is accompanied by

a maximum value in elastic projection, which signifies the existence of an overshoot in shear stress

response in that cycle, like the stress overshoot response during the start-up of the steady shear

flow. This response can be correlated to the network rupture, followed by the flow of the samples

at sufficiently large strains. The Lissajous shapes become more complicated, and the shape of the

elastic Lissajous plots is much closer to a rectangle than an ellipse at higher salt concentrations,

corresponding to an elastoviscoplastic material. Hence, upon increasing the salt content, the elastic

Lissajous plots deform closer to the rectangle shape, which can be interpreted as a greater extent

of nonlinearity upon reinforcing the CNC network structure in aqueous media.

4-5-7 Lissajous-Bowditch plots of CNC/salt suspensions at 30 g/L CNC

suspensions at different salt concentrations

Sa

lt 8

5.5

mM

Salt

172m

M

-1.5

-1

-0.5

0

0.5

1

1.5

-1.5 -1 -0.5 0 0.5 1 1.5

() n

orm

.

(d/dt)norm.

d

-1.5

-1

-0.5

0

0.5

1

1.5

-1.5 -1 -0.5 0 0.5 1 1.5

() n

orm

.

()norm.

c

-1.5

-1

-0.5

0

0.5

1

1.5

-1.5 -1 -0.5 0 0.5 1 1.5

() n

orm

.

(d/dt)norm.

b

-1.5

-1

-0.5

0

0.5

1

1.5

-1.5 -1 -0.5 0 0.5 1 1.5

() n

orm

.

()norm.

a

110

Figure 4.16 Dimensionless Lissajous-Bowditch loops for CNC 30 g/SL suspension containing a, b) 85.5, c, d) 172

mM salt using cone-plate geometry (with a truncation of 101μm and a cone tip angle of 1°) at 25˚C. Projections on

the elastic (𝝉 - 𝜸) and viscous (𝝉 - 𝒅𝜸

𝒅𝒕) planes are presented at strain amplitudes of 𝜸𝟎 = 1, 40, 100, and 250% and

an angular frequency of 𝝎 =1 rad/s.

4-5-8 Nonlinear parameters

The nonlinear elastic parameters are summarized in where it provides a guideline for recognizing

the type of elastic nonlinearity based on the defined parameters.

Table 4.1 Characterizing the elastic nonlinearity in response to imposed large amplitude oscillatory shear (LAOS)

deformation.

Elastic parameters Inter-cycle Intra-cycle

Both 𝐺𝑀′ and 𝐺𝐿

′ increase by increasing 𝛾0 Inter-cycle strain stiffening -

Both 𝐺𝑀′ and 𝐺𝐿

′ decrease by increasing 𝛾0 Inter-cycle strain-softening -

S > 0 at each strain - Intra-cycle strain stiffening

S = 0 Linear elastic behavior Linear elastic behavior

S < 0 at each strain - Intra-cycle strain softening

𝑒3 > 0 - Intra-cycle strain stiffening

𝑒3 = 0 Linear elastic behavior Linear elastic behavior

𝑒3 < 0 - Intra-cycle strain softening

The viscous nonlinear parameters are summarized in Table 4.2, where it provides a guideline for

recognizing the type of viscous nonlinearity based on the defined parameters.

111

Table 4.2 Characterizing viscous nonlinearity in response to imposed LAOS deformation.

Viscous parameters Inter-cycle Intra-cycle

Both 𝜂𝑀′ and 𝜂𝐿

′ increase by increasing 𝛾0 Inter-cycle shear thickening -

Both 𝜂𝑀′ and 𝜂𝐿

′ decrease by increasing 𝛾0 Inter-cycle shear thinning -

T > 0 at each strain - Intra-cycle shear thickening

T = 0 Linear viscous behavior Linear elastic behavior

T < 0 at each strain - Intra-cycle shear thinning

𝑣3 > 0 - Intra-cycle shear thickening

𝑣3 = 0 Linear viscous behavior Linear elastic behavior

𝑣3 < 0 - Intra-cycle shear thinning

112

CHAPTER 5: Viscoelastic properties of poly (vinyl alcohol)

hydrogels with cellulose nanocrystals fabricated through

NaCl addition 9

This work characterizes the viscoelastic and mechanical behavior of hybrid cellulose nanofibers

(CNC)/ Polyvinyl Alcohol (PVA)/NaCl hydrogel. Average porosity values measured by scanning

electron microscopy shows 23±5, 69±4, 72±7, 73±8, and 72±10 µm for samples with 10 g/L, 15

g/L, 20 g/L, 25 g/L, and 30 g/L CNC, respectively. In all samples with the CNCs concentration

higher than 10 g/L, we observed a wide distribution and almost equal average pore size. It depicted

that the pore size on average for the hybrid hydrogel with 5 wt% PVA is no longer dependent on

CNC concentration for CNC concentrations above 15 g/L. Transmission electron microscopy was

employed to characterize the distribution of CNCs inside the PVA matrix. Small and large

amplitude oscillatory shear measurements were performed on CNC-loaded PVA hydrogels to

understand the microstructure and viscoelastic behavior. At CNC loadings near and less than the

percolation threshold, 15 g/L, a polymer-CNC network was detectable via TEM imaging and

according to the plateau value on the storage modulus. Another jump in storage modulus was

observed where a connection among n CNC fibers was established. At these loadings, a CNC

network was present in the hydrogel, indicated by limited CNC loading dependence of the storage

modulus. The results of Large Amplitude Oscillatory Shear (LAOS) showed a viscoelastic

behavior for the hybrid material, demonstrating its potential for applications in articular cartilage

repair. To explore the interaction between CNC-CNC and CNC-PVA, the Foglar-Tucker model,

designed to account for modeling the orientation of CNC particles within the matrix, was fitted on

stress-overshoot experiments. The present study on the CNC-PVA hydrogels opens avenues for

further developing advanced materials for biomedical and energy applications.

9 Moud, A. A.; Kamkar, M.; Sanati-Nezhad, A.; Hejazi, S. H., Sundararaj. U.T, CHAPTER 5: Viscoelastic

properties of poly (vinyl alcohol) hydrogels with cellulose nanocrystals fabricated through NaCl addition. To be

submitted.

ABM did the experimental design, data collection and interpretation, and manuscript preparation. MK contributed in

rheological measurement and interpretation.

113

Graphical abstract

5-1 Introduction

Poly (vinyl alcohol) (PVA) is a synthetically derived, water-soluble, non-toxic, and biodegradable

polymer [234]. It is the most widely manufactured water-soluble polymer in the market[235]. PVA

has been used in a variety of different industries, such as medical, packaging, food industry, and

paper making[236]. Reported application for PVA are in important field of biomedical applications

such as tissue-mimicking, cell culturing, and body implants [89, 237]. PVA has been combined

with different fibers and micro-sized cellulosic materials such as Cellulose nanocrystals (CNCs).

Due to hydrophilicity of CNCs, CNC-PVA hydrogels present a potentially sustainable option for

materials development and design as these hydrogels potentially can be light-weight,

biocompatible, and biodegradable materials.

The hydrophilicity of CNCs stems from the structure of cellulose that owns hydroxyl groups on

its main repeating unit. There are other advantages that CNC brings to the table in addition to its

similar chemistry to water and PVA. For instance, cellulose nanocrystals (CNCs) are nanoscale,

114

abundant naturally derived fibers from plant sources ranging from wood, linen as well as

microorganisms such as bacteria, tunicate, and algae [2, 238, 239]. CNCs also possess strength

and Young’s modulus as high as approximately 7.5 GPa[240] and 100–220 GPa[2, 241, 242],

respectively. Considering their mechanical properties8 with relatively low density and the potential

of being naturally driven make CNCs an ideal renewable source for making mechanically robust

hydrogels. CNCs' geometry also aids further enhancement of mechanical properties due to their

slender shape. Indeed, it has been reported that CNC fibers depending on source and method of

production can have widths changing between 3 nm and 50 nm and lengths changing between 50

nm to 300 nm [2, 50, 241, 243, 244].

Application-wise, the combination of cellulose-PVA and nanocellulose-PVA materials for a

variety of applications has been established previously. In the literature, there are reports of

combining CNCs with PVA for aim of producing hydrogels with high mechanical integrity [62,

235, 242, 244-250]. The interactions between PVA and CNCs due to hydrogel bonding and

intermolecular interactions enforces the hydrogels to own satisfactory mechanical properties[57].

For instance, increasing CNC concentration from a low amount of 0.1 to 0.2 wt% relative to total

gel weight has increased the gel storage modulus of the composite from 3.8 kPa to 14.3 kPa[74].

Upon addition of only 3 wt% micro-fibrillated cellulose into a PVA polymer[57], tensile strength

and modulus of the composite hydrogel enhanced for about 13% and 34%, respectively.

Tissue engineering has emerged as a suitable destination for injectable hydrogels [251]. PVA-CNC

solution-processed hydrogel seems to fit well within this category as both CNC and PVA are

biocompatible, abundant, and with low-cost production. Hydrogels, after being injected and during

injections, need to meet certain criteria. For example, the extent of healing, timing, critical strain,

non-linear behavior, the extent of respond to multivariant strain amplitude levels, and cyclic

loading is crucial for design. Precise rheological characterization is needed for design of injectable

hydrogels before employment in the field of interest. Here we propose fabrication of a hydrogel

with PVA and CNC with the addition of salt as a gelifying agent. We aim to characterize the

hydrogel using SEM, TEM, linear and non-linear rheology, and compression testing.

This work also inspects the influence of CNCs on rheological behavior of PVA while NaCl is

added as gelling agent. The LAOS’s behavior of the PVA-CNC system has come under scrutiny

in this report. Moreover, solution processing of PVA-CNC and un-crosslinked version of PVA has

115

been chosen since this hydrogel is destined to be used as an injectable scaffold for tissue

engineering. TEM and SEM of the hydrogels were also employed for the aim of complete

characterization of hydrogels and fabrication of a hydrogel with repeatable properties. TEM

images also made understanding the rheological behavior of the gel deeper. Tissue engineering

demands the hydrogel to reconstruct itself after being injected, and this feature requires a physical

gel.


5-2-1 Materials and Materials preparation

InnoTech Alberta provided the CNC with measured length in the span of 100-300 nm and a

diameter of 10-40 nm. According to the method of CNC production, extracted rods from the

natural cellulosic source had negative charges due to the acid hydrolysis process. PVA was

purchased from Sigma Aldrich with a reported molecular weight of 130,000 g/mol.

CNC powder was sonicated in DI water for 10 mins, and a stock suspension with a concentration

of 6wt% was made. At the end of sonication, initial murky suspension got cleared, and the level

of suspension transparency varied with CNC concentration. Suspension of CNC alone pH

measurement showed a value of 6.8. The measurement was done employing a Mettler Toledo

Seven Compact pH-meter (Mettler-Toledo 135 International Inc., Columbus, OH, USA). Ultra-

sonication (125 W Qsonica Sonicators Q125 Sonicator, Qsonica) was employed for suspending

CNCs in DI water. It was found based on zeta potential values that 1000 joule per gram of CNC

is enough to obtain a fully dispersed suspension. The ice bath treatment was carried out during

sonication to mitigate the risk of overheating on the surface of CNCs as the surface charge of CNC

particles is vulnerable to elevated temperatures [50].

The protocol for making suspension of CNC in PVA was (1) dissolving PVA in DI water under

the condition of constant stirring for one hour under the temperature of 80 C° to make a solution

of PVA in water with 5 wt% concentration. To make sure there is no PVA crystal chunk in the

solution, the solution was placed under DLS measurements. The result showed no peak related to

PVA chunks being present in the system. (2) Addition of sonicated CNCs into the mix. (3) To

116

make the final hydrogels, CNC with desired concentrations was added into the PVA-DI water

solution and was sonicated for extra 5 mins (4) in the final stage; NaCl with a concentration of

17.24 mM was added to promote gelation of CNC-PVA hydrogel. In the entirety of the paper, as

salt concentration and PVA did not change, we refer to the sample only with regards to changes in

CNC concentration. We observed that addition of 17.24 mM salt is enough to cause gelation in

CNC-PVA-salt samples visually through vial inversion test. Addition of 5 wt% PVA deemed

enough to show impact of addition of PVA on mechanical properties of CNC hydrogel.


5-2-2-1 Scanning electron microscopy

The gel structure of CNC-PVA hydrogels and their micro-morphology was investigated by using

XL30 Philips SEM. Prior to each imaging session, the generated gels were freeze-dried under the

influence of liquid nitrogen. A small piece of freeze-dried hydrogel was placed carefully onto a

silicon wafer. A sheet of gold was then sprayed all over the surface of the sample, to mitigate

electrostatic discharge.

5-2-2-2 Transmission electron microscopy

The TEM analysis of the CNCs was carried out on a Tecnai TF20 G2 FEG-TEM (FEI, Hillsboro,

Oregon, USA) at 200 kV acceleration voltage with a standard single-tilt holder. The images were

taken with a Gatan UltraScan 4000 CCD (Gatan, Pleasanton, California, USA) at 2048×2048

pixels. For the TEM analysis of the CNC, the droplets of CNC-PVA suspensions were placed on

a holey carbon-coated Cu TEM grid and dried at room condition.

5-2-2-3 Rheology

Rheological measurements were performed using an Anton-Paar MCR 302 rheometer equipped

with 50 mm diameter cone-plate geometry (cone angle of 1° and truncation of 101 μm). To reach

the desired stabilized morphology, a resting time of 20 min was used in the rheometer following

the CNC-PVA hydrogel loading. The strain amplitude sweep experiment was carried out within

the range of 0.1 to 1000.0 % and at an angular frequency of 1 rad/s on all samples to determine the

117

linear viscoelastic regime (LVR). Based on the results of this experiment, the strain amplitude of

1% was determined to be small enough to keep the deformation in the LVR. All experiments were

carried out at room temperature.

The rheometer was placed in a rigid and mechanically stable environment to minimize

mechanical noises and apply large amplitude oscillatory shear (LAOS) to the samples. To obtain

full waveform of shear stress and strain, the material was strained at constant frequency and

amplitude. LAOS’s data were collected after 5-6 cycles for each strain amplitude. Rheological

tests were performed on triplicates and quadruplicates for each sample.

5-2-2-4 Compression tests

Compression testing was done at the rate of 10 mm/mins with bose 3200 device. The 3200 Series

III compression test equipment can be utilized for 225 N or 450 N maximum force capacity. The

system is equipped with a bandwidth, which is also able to carry out tests cyclically till 300 Hz

and 200 Hz for analyzation of mechanical properties.


5-3-1 Morphological characterization of freeze-dried hydrogels under SEM

Probing pore size with Image J reveals pore size distribution across samples. 23, 69, 72, 73, 72 µm

for samples with 10 g/L, 15 g/L, 20 g/L, 25 g/L and 30 g/L CNC, respectively. In all samples, after

10 g/L CNC loading, we observe a wide distribution in pore size and almost equal average pore

sizes. The trend in the average pore size versus the CNC/PVA ratio has been depicted on Figure

5.1e. This observation reveals that after the addition of 15 g/L CNC into 5wt% PVA, the average

pore sizes its not dependent on CNC loading anymore.

b c d a

118

Figure 5.1 Scanning Electron Microscopy (SEM) micrographs of cellulose nanocrystals (CNC)- poly (vinyl alcohol)

hydrogels (PVA) freeze dried samples: a) (CNC 10 g/L), b) (CNC 15 g/L), c) (CNC 25 g/L), d) (CNC 30 g/L) at

magnification of 100x. e) The average pore size of samples as a function of CNC concentration

5-3-2 Morphological characterization of CNC-PVA colloids

The morphology of CNC can be studied by various microscopic techniques such as transmission

electron microscopy (TEM) and scanning electron microscopy (SEM). Morphology evaluation is

a useful tool for inspection of surface morphology, state of constituent’s distribution and more

importantly evaluation of size on CNC particles. In the case of having high aspect ratio particles,

, the probability formation of a network between the CNC particles will be high; therefore, it will

yield hydrogels with improved mechanical properties.

TEM images of CNC-PVA hydrogels at different concentrations are depicted in Figure 5.2

Consistently, with an increase in the concentration of CNC, the corresponding TEM images also

become more crowded with particles. Another interesting observation is the full dispersion of CNC

inside PVA, showing that sonication inside the PVA solution was enough to evenly disperse and

distribute CNCs. If we compare TEM images associated with the CNC-NaCl system published

previously[216], we understand that gelation in the presence of PVA does not happen as intensely

as in naked CNC-NaCl systems. Based on the perfect distribution and dispersion of CNCs

observed here, we can conclude that the mechanical properties of the CNC-PVA system should

e

119

also be isotropic in nature. Using Image J, we analyze the TEM images where the average length

and diameter of CNCs are estimated as 150±10 nm and 11±3 nm, respectively. In the report by

Huan et al., 9 ± 3 nm and 90 ± 28 nm was measured for CNC diameter and length, and the average

aspect ratio was reported as 10 [252]. In their article, the method of CNC production shared

similarities to this work.

Figure 5.2 Distribution of CNC particles embedded in PVA at different concentrations and at 0.5 µm resolutions

respectively: a-d (CNC 10-15-25-30 g/L)

The results of morphological investigations by microscopy-based techniques here are similar to the

results elsewhere [50, 253-257]. As expected, based on TEM images, the morphology of individual

CNCs is also a spindle shape.

5-3-3 Rheological characterization of CNC-PVA samples

To comprehend the impact of various CNCs on the viscoelastic traits of hydrogels, oscillatory

measurements were carried out at 25◦C in hydrogels with the concentration of CNCs equal to 10,

15, 20, 25, and 30 g/L. In Figure 5.3, the G′ and G′′ are drawn as a function of frequency for

hydrogels within the linear deformation range, and they are illustrated in Figure 5.3a-b. Two

abrupt increases exist in storage modulus curves versus frequency as a function of CNC

concentration. These abrupt increases can be associated with CNC-PVA network formation (15g/L

till 25g/L), and direct CNC-CNC interaction in the PVA matrix after concentration reaches to

30g/L. In all concentrations, there is a network that makes the behavior of hydrogel elastic (G′ >

G′′). At high frequencies where G′(ω) > G′′(ω), a more dominant elastic character was observed,

which depicted a typical solid-like character.

a b c d

120

Figure 5.3 a) Storage, b) loss modulus and c) complex viscosity as a function of frequency at strain amplitude of γ0 =

1%. d) Storage (solid symbols) and loss moduli (open symbols) as a function of strain amplitude at angular frequency

of ω = 1rad/s.

Figure 5.4 Storage modulus as a function CNC concentration. Values are extracted from Figure 5.3.

We are well positioned above the rheological percolation threshold for all samples as storage

modulus-frequency data, even for 10 g/L CNC sample. The same trend can be observed for 15

g/L, 20 g/L, 25 g/L, and 30 g/L samples. Overall, it can be perceived that the viscoelasticity of

these hydrogels followed the order of 30>25>20>15>10 g/L CNC reinforced PVA hydrogels.

Figure 5.4 shows the effect of the addition of CNC into the PVA matrix. After adding 3 times

1

10

100

1000

0.01 0.1 1 10 100

G' (P

a)

Frequency (Hz)

CNC 10g/l

CNC 15g/l

CNC 20g/l

CNC 25g/l

CNC 30g/l

1

10

100

1000

0.01 0.1 1 10 100

G"

(P

a)

Frequency (Hz)

0.01

0.1

1

10

100

1000

10000

0.01 0.1 1 10 100

|η*

| (P

a)

Frequency (Hz)

a b c

0.1

1

10

100

1000

0.01 0.1 1 10 100 1000

G' &

G"

(P

a)

γ ( )

d

121

more CNC into the system, hydrogel displays more solid-like behavior and 40 times larger storage

modulus (391 vs. 11). Jumps in storage modulus values are quite noticeable between 11 and 103

and between 103 to 391. The ratio of storage modulus values is about 100 times more, and the

second jump is about 40 times stronger.

Since the sudden increase in values of storage modulus at higher CNC loading showed more solid-

like behavior, this result suggests that the rheological response was correlated to the structuring

(structural formation) of CNCs in the hydrogels. These observations showed additional insight into

structure and dynamic of the hydrogels. The same trend in data can also be observed in Figure

5.3c that shows complex viscosity variation as a function of frequency. Looking at the structure,

the data obtained here shows that network was present in all CNC concentrations except for 10

g/L sample. As it was expected, viscosity and elasticity of hydrogel increased with increase in

CNC content. However, trends in storage modulus versus frequency showed behavior is dependent

on CNC loadings. For hydrogels with concentration of higher than 10 g/L, it can be estimated that

network of polymers through entanglements is connecting CNCs. Specially it can be seen storage

modulus still increases with increase in CNC loading [173]. Polymer network connecting CNCs

played a role in increasing storage modulus at or below percolation point, while CNC network

played its role at higher CNC loadings. Finally, transition from linear to non-linear regime is shown

in Figure 5.3d. In presence of PVA molecules the addition of more CNCs caused the threshold of

transition from linear to non-linear to decrease. This result was also seen in our previous work

[258].

According to experimental observations [124], the LAOS behavior of fluids with complex

behavior can be categorized into several type of behavior. (1) strain thinning (storage modulus and

loss modulus decreasing); (2), strain hardening (storage modulus and loss modulus increasing);

(3), weak strain overshoot (storage modulus decreasing, and loss modulus increasing continued by

decreasing); (4), strong strain overshoot (storage modulus and loss modulus increasing continued

by decreasing). In our work, PVA-CNC shows type 3 behavior. The type 3 trend for the CNC-

PVA system means that both the creation and loss in the structure of the hydrogel increase with

the strain amplitude, but the rate of destruction happens faster than creation. The rate of junction

fabrication is enough for network formation (or any microstructure occurring because of these

interactions), yielding strain hardening behavior, while the rate of loss in the structure becomes

122

more dominant at larger strain amplitudes. Thus, the overshoot observed can be regarded as the

point of equality of creation and destruction rate of microstructural junctions inside the hydrogel.

Although not as common as type I behavior, the similar response has been observed in other

complex fluids such as di-block copolymers or surfactant containing solutions[259], dough [260,

261] or xanthan gum [262], emulsions [263, 264], silica suspension [265, 266], and alumina

particles suspended in PDSM under imposition of an electric field [267]. Whittle and Dickinson

[268] via generating three-dimensional gel of particles with Brownian dynamic simulation of soft

particles with spherical shapes predicted strain hardening behavior. In our previous report [216],

we examined the storage modulus of CNC-NaCl gel without PVA and found storage modulus of

20 g/L and 30 g/L CNC samples equal to 15 Pa and 25 Pa, respectively. In comparison to this

work, with the aid inclusion of PVA, hydrogel mechanical properties have improved greatly to

91.3 and 344 Pa for the same samples. This simple comparison shows how PVA chains indeed

strengthen and improve the mechanical properties of the hydrogel.

5-3-4 Storage modulus-recovery relationship

It is well established that storage modulus 𝐺′is sensitive to the microstructural changes within

hydrogel [269]. To probe the pace and magnitude of restructuring and elastic recovery of the PVA-

CNC hydrogel network, storage modulus was monitored immediately after test of start-up shear.

According to a recent study [270], the recovery of samples during this period can be divided into

two distinct stages. In the first step, fast recovery observed in the curves is attributable to

restructuring of the network. Therefore, the initial intensification of modulus is due to the joining

of the existing clusters and aggregation of un-oriented nanofillers to form a 3D network structure.

The restructuring in this stage is mainly controlled by particle-particle and particle-matrix

interactions. In the second stage, particle motions controlled by Brownian diffusion, the orientation

distribution of fillers returns to its original isotropic state. Figure 5.5 shows the result of the storage

modulus recovery versus annealing time for PVA-CNC hydrogels. In Figure 5.5, samples with

lower CNC loadings, i.e., 10-15 g/L, show a smooth increase in the storage modulus. For the

sample containing 20g/L CNC, the storage modulus does not increase as much compared to the

sample with 30 g/L CNC. The ascending trend of the modulus in this test is not only a function of

nanofiller concentration but also a function of the dispersion state of the nanofillers. It is interesting

123

that the ultimate recovery after 1500 seconds is close to the original value that we obtained during

frequency sweep measurements (366 vs. 386). This recovery shows gel at higher CNC

concentration is fully recoverable. The faster rate of recovery for this hydrogel can also be

attributed to the second jump we record in the storage modulus versus frequency. It was

demonstrated that the second jump was attributable to CNC-CNCs junctions. Breaking these

junctions in shear means orientation that will take less time in comparison to polymeric chains

relaxation to original isotropic orientation. However, glancing over data for the sample with 10

g/L CNC shows the lowest recovery (3.21 vs. 141). The recovery rate for destined applications is

important as it shows how fast the system heals. To provide further insight into the network

healing, the recovery rate of the networks were also investigated for 10 g/L and 30 g/L samples at

different strain amplitudes of 1% and 30% for period of time of 200 second each and we followed

this cycle for each strain amplitude 2 times results are shown in Figure 5.11. Results showed both

samples can recover however 10 g/L samples showed less potential. Herein, we monitored storage

modulus recovery as a sign of structural healing as a function of time for 10 g/L and 30 g/L

samples. The protocol of experiment was application of 1% amplitude strain for 10000 seconds

and then resting for 386 seconds. Results are shown in Figure 5.12. Consistent with results

observed earlier structural reconstruction happens inside the hydrogel, however rate of

reconstruction is faster for higher CNC loadings.

Figure 5.5 a) Flow curve (�� = 1s-1) and b) reconstruction of PVA-CNC hybrid hydrogel network after breakage as a

function of time (γ0 = 1% and ω = 1rad/s).

5-3-5 Lissajous-Bowditch plots of CNC/salt suspensions at different CNC and

salt concentrations

0.01

0.1

1

10

100

1000

0 50 100 150 200

η(P

a.s

)

Shear rate (s-1)

CNC 30g/lCNC 25g/lCNC 20g/lCNC 15g/lCNC 10g/l

0

100

200

300

400

0 500 1000 1500

G' (P

a)

Time (s)

CNC 30g/l

CNC 25g/l

CNC 20g/l

CNC 15g/l

CNC 10g/l

a b

124

The plot of instantaneous stresses τ(t) against strains γ(t) (often called Lissajous−Bowditch plots)

for different strain amplitudes (γ0) allows one to follow the actual network response to each loading

cycle during the oscillatory shear testing. For purely elastic materials in the linear region, the

response is expected to be completely in phase (δ = 0), and, as a result, the Lissajous−Bowditch

plot is presented in line with a slope of G′. The response of purely viscous materials is expected to

be out of phase by δ = π/2 (see our previous works for more information about Lissajous plots

[219, 220, 226]). Viscoelastic materials like polymer networks exhibit both elastic and viscous

properties, and, thus, the waveform forms a simple ellipse signaling the linear response, and the

slope of the semimajor axis represents the magnitude of the complex modulus |G*|. The ellipse

shape in LVR is because the output stress signal of the viscoelastic materials only includes the first

harmonic coefficient in LVR. The emergence of any nonlinearity (due to the excitation of higher

harmonics in the output shear stress waveform) in viscoelastic character of the material by

surpassing LVR can be identified by any distortion in ellipsoidal pattern of the Lissajous plots.

Therefore, with the aid of Lissajous−Bowditch (LB) plots, one can visually recognize the

transformation from a simple ellipse, representing the linear response to a complex nonlinear

shape, implying nonlinear behavior. Hence, LB plots facilitate the qualitative assessment of intra-

cycle nonlinear behavior and deliver information about the emergence of nonlinearity, type of

nonlinearity, and extent of nonlinear viscoelastic behavior.

The LB plots of the CNC/PVA/salt hydrogel at different CNC concentrations and various

amplitudes (γ0 = 1, 7, 10, 40%, see arrows in Figure 5.3d) are shown in Figure 5.6. As can be

seen in Figure 5.6, the LB loops in Short Amplitude Oscillatory Shear (SAOS) region (γ0 = 1%)

are ellipses in both viscous and elastic projections. The narrow LB loops in elastic projection in

concert with wide ellipses in viscous projection in SAOS region verify the dominant elastic

response of these samples due to gel-like structures. However, the narrow elastic LB loops rotate

counter-clockwise as the CNC concentration increases, indicative of higher elasticity as a direct

consequence of the formation of a stronger network. These results are in complete agreement with

the results of Figure 5.3 and Figure 5.4.

125

By increasing the amplitude of deformation, the LB plots deviated from ellipsoidal shape and

became distorted, indicative of occurrence of intra-cycle nonlinearity in the systems. For instance,

the elastic LB plots look like curvilinear parallelograms at larger deformations. As the CNC

concentration increases, the viscous LB plots become more distorted and elastic LB plots become

more boxed-shaped. This reveals the sensitivity of the LB plots to any changes in the

microstructure. We observed similar nonlinear viscoelastic behavior in our previous works where

samples with a stronger network demonstrated higher nonlinearity [219, 220, 258].

Figure 5.6 Lissajous Bowditch plots: a) stress versus strain b) stress versus strain rate for CNC-PVA/salt

hydrogels at different CNC contents and strain amplitudes of γ0 = 1, 7, 10, 40% and angular frequency of

ω=1rad/s.

-6

-4

-2

0

2

4

6

-0.015 0 0.015

τra

w(P

a)

γ'raw (-)

-6

-4

-2

0

2

4

6

-0.015 0 0.015

τra

w(P

a)

γraw (-)

-15

-10

-5

0

5

10

15

-0.13 0 0.13

τra

w(P

a)

γ'raw (-)

-15

-10

-5

0

5

10

15

-0.13 0 0.13

τra

w(P

a)

γraw (-)

-15

-10

-5

0

5

10

15

-0.5 0 0.5

τra

w(P

a)

γ'raw (-)

-15

-10

-5

0

5

10

15

-0.5 0 0.5

τra

w(P

a)

γraw (-)

γ0

= 1

%γ

0=

10

%γ

0=

40

%

Viscous Projection Elastic Projection

-15

-10

-5

0

5

10

15

-0.1 0 0.1

τra

w(P

a)

γ'raw (-)

-15

-10

-5

0

5

10

15

-0.1 0 0.1

τra

w(P

a)

γraw (-)

γ0

= 7

%

126

As it can be seen in Figure 5.6, after increasing the deformation to γ0 = 7%, self-intersections

appeared in the viscous LB plots of samples containing CNC greater than 15 g/L. Several reports

have observed the emergence of self-intersection under LAOS deformation for different systems

(e.g., polymer solutions and polymer nanocomposites) [220, 233]. Self-intersection is

accompanied by a maximum value in elastic projection in the same quadrant of the deformation

cycle, which signifies existence of an overshoot in shear stress response, similar to the stress

overshoot in the start-up of the steady shear flow. This response can be correlated to the network

rupture followed by flow of the samples at sufficiently large strains. As the rheometer plate reaches

to the γ0 in each cycle of deformation (starting point of the arrows), the flow direction is reversed,

and the samples are deformed. Since these gel-like samples behave elastically, the stress growths

linearly with accumulation of the deformation and reaches a local maximum followed by a

decrease with further increasing the shear rate. As mentioned earlier, this behavior is due to the

breakage and yielding of the network structure and samples start to flow beyond this point

(maximum stress, overshoot). This behavior will be studied in more details in the following

section. It should be borne in mind that contrary to the start-up of steady shear flow, LAOS is a

periodic flow. Hence, network structures which collapse irreversibly in the period of (time scale

of) each cycle of oscillation would not show self-intersection and secondary loops.

Interestingly, the secondary loop observed in viscous LB plots of samples containing 20 g/L and

25 g/L CNC at strain amplitudes of γ0 = 7 and 10% disappeared as the deformation further

increased to γ0 = 40%. However, secondary loops became more pronounced for 30 g/L at γ0 = 40%.

Considering the constant frequency of deformation in each cycle, the samples experienced higher

shear rates at larger strain amplitude resulting in widespread structural collapse in samples. This

rationalizes the inability of the samples containing 20 g/L and 25 g/L CNC to structural recovery

once deformed beyond a critical strain. This is further confirmed as the viscous LB plots of samples

containing 10, 15, 20, 25 g/L CNC became narrower at γ0 = 40%, indication of development of

fully viscous flow. The difference in nonlinear viscoelastic of CNC/PVA/salt hydrogels at high

and low CNC concentrations will be further discussed in the following section.

5-3-6 The sequence of physical processes

127

Elastic LB plots in Figure 5.7 provide further insight into the rich phenomenology of structural

evolution of PVA-CNC-salt systems under oscillatory flow at multiple strain amplitudes. Figure

5.7and b show the linear and nonlinear elastic LB plots of the hydrogels at two different

concentrations of CNC (10 and 30 g/L). At first glance, it can be understood that system containing

higher amount of CNC falls into nonlinear regions at smaller strain amplitude. That is, the LB plot

of PVA-CNC (10 g/L)-salt at γ0 =10% is ellipsoidal shape while a high distortion in LB plot of

PVA-CNC (30 g/L)-salt at γ0 =5% is observable. Moreover, in nonlinear framework, each system

follows a different scenario as discussed below.

The observed nonlinear behavior in Figure 5.7 is studied here with the aid of sequence of physical

processes method [132]. Beginning at zero shear stress in nonlinear region (see circle in Figure

5.7c) the systems are strained and, hence, the network structures in the systems are deformed in a

linear elastic fashion. This process continues until the structures in each system deform beyond

their yield strain (see square in Figure 5.7c). At this point, the static yielding of system containing

30 g/L CNC is associated with a bump immediately after the initially elastically build up of stress

(Figure 5.7b). While no overshoot was observed for systems with 10 g/L CNC (Figure 5.7a).

From now on, solid state corresponds to PVA-CNC (30 g/L)-salt and we address PVA-CNC (10

g/L)-salt as soft state.

Similar behavior is observable for the solid state in the other spatial direction and another

overshoot occurs as the flow reverses. Hence, the network structures of the solid state are assumed

to break as the yield stress is exceeded and reform when the deformation rate is instantaneously

zero at γ = ± γ0. The straining-yielding-flowing-reformation behavior signals a fast reversible

network reformation for solid state which stems from inter-particle forces and the associated

particle dynamics due to Brownian motion continually regenerating the network structure. Since

this network reformation is time dependent and the frequency of deformation is constant in our

LAOS experiments, the static yield stress is strain dependent. That is, the static yield stress

decreases as the amplitude of deformation increases from γ0 =10 to 19% (follow the blue arrow in

Figure 5.7b). In fact, flow of the material takes a finite time to relax. This is in line with the

findings of Rogers et al. [132] in which they mentioned that the stress overshoot associated with

static yield stress takes smaller and smaller portion of the LB plots upon increasing the amplitude

of deformation.

128

It should be mention that the static yield stress of the solid state disappeared upon further increasing

the strain amplitude (for instance see Figure 6d showing the LB plot of solid state at γ0 =100%).

The open gray circles in Figure 5.7c and d compare the data obtained for power-law fit by

regression to the real data (black line) of solid state at strain amplitudes of γ0 =10 and 100%. As

can be seen in Figure 5.7c, the power-law fails to describe the pre-yielding waveform (before

static yield stress) of the solid state at γ0 =10%. However, it acceptably fits the post-yielding (after

the static yield stress) portion of waveform. This analysis reveals that the deviation of the power-

law from the waveform in pre-yielding region is because an elastic deformation takes palace in

this region and the good fit in post-yielding region in because a viscous flowing response occurs.

The power-law accurately describe the behavior of the waveform at sufficiently large deformations

(see Figure 5.7d). In fact, no seeing the static yield stress at higher level of deformation shows

that network restructuring in the system is time dependent for the 30 g/L CNC. The systems

experience higher shar rate at larger strain amplitudes leading to a more collapse structure (higher

accumulated stresses). Therefore, there exist higher level of stress to relax in the same amount of

time at large deformation. As a result, there is no sign of elastic deformation and static yield stress

at strain amplitude of 100%. That is, a larger portion of waveform comes from power-law fluid

response.

The power-law index found to be n = 0.47 and 0.74 for strain amplitude of 10 and 100%,

respectively. It should be born in mind that for n = 1 the waveform is expected to be a simple sine

with a single harmonic in the frequency domain and n equal to zero would result in waveform

containing many higher harmonics (higher nonlinearity). Hence, the decreasing trend of power-

law index upon increasing the strain amplitude is against our expectation (i.e., we expect to observe

higher nonlinearity upon increasing the amplitude of the deformation). However, this behavior can

be rationalized by the fact that at sufficiently high shear rates associated with large deformations,

the junctions between the elements of the network structure is entirely lost and since there is not

enough time for structural reformation, the solid state loses its viscoelastic character. Thus, it

behaves more of a fully viscous fluid than a viscoelastic hydrogel. As a result, the viscoelastic

nonlinearity vanishes.

Hence, analyzing the stress wave form in both linear and nonlinear region with the aid of LB plots

helped us to evaluate the structural differences in the samples and provided us more information

129

on structural features of the samples. These results were not identifiable in linear rheological data

or lumped viscoelastic data provide by a routine strain amplitude test. This confirms the vigorous

sensitivity of the intra-cycle local viscoelastic measures to any changes in the microstructure.

Figure 5.7 Lissajous-Bowditch plots of a) CNC(10 g/L)/PVA/salt at strain amplitudes of γ0 = 10, 14, and 19%, b)

CNC(30 g/L)/PVA/salt at strain amplitudes of γ0 = 3.5, 5, 10, 14, and 19%. c) and d) open gray circles represent

power-law flow response of CNC(30g/L)/PVA/salt (solid line, in corresponding to a strain amplitude of c) 10%

and d) 100% and angular frequency of ω=1 rad/s, raw waveform as a black solid line). The direction of traversal is

indicated by the dashed arrows.

5-3-7 Overshoot during a start-up experiment

-15

-10

-5

0

5

10

15

-0.21 -0.14 -0.07 0 0.07 0.14 0.21

τra

w (P

a)

γraw (-)

3.5%

-20

-15

-10

-5

0

5

10

15

20

-1.2 -0.8 -0.4 0 0.4 0.8 1.2

τra

w (P

a)

γraw (-)

-15

-10

-5

0

5

10

15

-0.16 -0.08 0 0.08 0.16

τra

w (P

a)

γraw (-)

-1.5

-1

-0.5

0

0.5

1

1.5

-0.21 -0.14 -0.07 0 0.07 0.14 0.21

τra

w (P

a)

γraw (-)

19% 14%

10%

CN

C 1

0g

/l

CN

C 3

0g

/l

10% 100%

CNC 30g/l CNC 30g/l

a

b

c) d)

130

Rheology of filled polymers due to intricate interaction between particle-particle and particle-

polymer and cluster break up and migration is complex. Hence, investigating the rheological

properties of suspension of polymer-nanoparticles is not trivial. Migration and rotation of particles

inside the filled polymers impacts their rheology. The relationship between flow and nano particles

orientation can be studied by non-liner rheological experiments. Polymers filled with

axisymmetric particles such as CNCs, when sheared between two plates, show a stress or viscosity

overshoot. The first study on such interactions has been reported for dilute suspensions wherein

interaction between particles are not considered. Extensions on the modeling of particle

interactions with isotropic orientation in Newtonian [271] and non Newtonian[272] fluids, for

ellipsoids with high aspect ratio have been also reported. For semi-dilute suspensions, Simple

models, based on the experimental observations[273], to predict the interactions between particles

with a shortened the period of rotation[274] and late with the inclusion of migration and

diffusion[275] are introduced.

5-3-8 Modeling of CNC orientation based on Folgar-Tucker based models

Figure 5.8 Transient shear stress of the CNC-PVA hydrogel measured at a shear rate of 1/s and 25 C°.

The relationship between flow and rotation of rod particles can be assessed with non-liner

rheological tests. Figure 5.8 shows stress versus strain relationship for the highest CNC loaded

131

(30 g/L, solid sate) hydrogel and the lowest CNC loading (10 g/L, soft state). The goal of this test

is to reveal the difference between network structure in solid and soft states even further. Results

showed that samples with 30 g/L CNC displayed more energy consumption (higher stress at equal

shear rates) compared to sample with 10 g/L CNC. Hence, quite aligned with the results of previous

sections, the sample with a higher level of CNC loadings exhibits a stronger network and therefore

reaches higher stress overshoot compared to the sample with 10 g/L CNC. According to Doi and

Edwards theory[40], the concentration regime can be divided to dilute, semi-dilute, concentrated,

and nematic regimes. In the dilute region, free rotation of particles without interaction with

neighboring particles is possible. In the semi-dilute region, rotation of particles without interacting

with the particles in their vicinity is not possible. In the concentrated regime, i.e. higher number

density per unit of volume, translational and rotational movements of particle both happens with

difficulty. Jeffery [102] first formulated equations relating orientation to the flow field for rotation

of short fibers. Later for concentred suspension, Koch [274] introduced a simple model for the

prediction of interaction between particles. This model does not consider interactions among

particles and assumes that the matrix is of non-Newtonian type. Other authors improved this model

until Folgar and Tucker [275] (also known as FT model) added a term into the equation for

considering effect of interactions between particles. Equation 1 displays this equation:

��𝟐 =𝑫𝒂𝟐

𝑫𝒕=

𝟏

𝟐 (𝛀𝒂𝟐 − 𝒂𝟐𝛀) +

𝟏

𝟐𝝀(��𝒂𝟐 + 𝒂𝟐�� − 𝟐��: 𝒂𝟒) 5-1

+2𝐶𝐼 ��(𝐼 − 3𝑎2)

5-2

In this equation 𝑎2 is second order orientations tensor, Ω is vorticity tensor, �� is rate of deformation

tensor, 𝑎4 is fourth order orientations tensor, 𝐼 is identity matrix, 𝐶𝐼 interaction parameter, ��

effective shear rate, and 𝜆 shape factor that goes to 1 for long slender particles. The definition of

these parameters can be found in Advani and Tucker articles [109].

Equation 1 contain fourth order orientation tensor that should be approximated using a closure

approximation. There are many approximations for fourth-order tensor which the simplest one is

quadratic closure estimation[276] as follow:

132

𝑎4𝑞

⟺ 𝑎𝑖𝑗𝑘𝑙𝑞

= 𝑎𝑖𝑗𝑎𝑘𝑙

5-3

This approximation is not as accurate as the other closures, however it helps with prediction of

reasonable results in all flows and, for the present article, has the advantage of mathematical

simplicity. For obtaining the transient stress after application of shear, Jeffery [102], hand [277]

and Lipscomb [278] suggested following equation:

σ = −𝑃𝐼 + 𝜂𝑚�� + 𝜂𝑚𝜙{𝜇1�� + 𝜇2𝛾: 𝑎4}

5-4

In this equation, 𝜂𝑚 is the viscosity of the matrix and 𝜇1 and 𝜇2 are rheological constants. Usually

𝜇1 takes the value of 2. For using this equation, two parameters of 𝜇2 and 𝐶𝐼 should be used as

fitting parameters which are related to concentration and aspect ratio of particles, respectively.

Considering a greater 𝐶𝐼 (meaning stronger interactions between particles), 𝑎11 at steady state

condition would take a lower value. Moreover, Jeffery model gives complete orientation

possibility to particles at extremely high strains. However, FT model through considering particle-

particle interactions, does not allow particles to get fully oriented.

In Figure 5.9 a-b Folgar-Tucker prediction with quadratic closure approximation on experimental

data has been shown. Naturally, this closure contains slight error at isotropic state. There are more

advanced closure approximations that are not focus of this article and therefore will not be used

here.

According to FT model, orientation of asymmetric particles happens fast and changing interaction

fitting parameters does not help with this issue. This problem was also observed in report of

Letwimolnun et al. [279] for the case of nanoclay dispersion in polypropylene. Moreover, it can

also be seen that according to this equation, interaction parameter is higher for higher CNC

loadings (CI=0.006 vs 0.002). Slower kinetic of PVA-CNC can be attributed to special hydrogen

bonds interaction between CNC-CNC and CNC-PVA and interactions of PVA-CNC. For

overcoming this problem, FT equation can be multiplied at a pre-factor called “k” that causes the

model to impose less strain onto CNCs.

133

Figure 5.9 Fitting FT equation on experimental data a) CNC 30 g/L using 𝜇2 = 4300 𝐶𝐼 = 0.006 and quadratic

approximation. b) CNC 10 g/L using 𝜇2 = 2700 𝐶𝐼 = 0.002 and quadratic approximation at shear rate of 1/s.

c-d) Fitting SRF equations on experimental data for CNC 30 g/L using 𝜇2 = 4300 𝐶𝐼 = 0.006 and quadratic

approximation under k values of 0.7, 0.8 and 0.85 at two magnifications.

For overcoming this problem, the FT equation can be multiplied at a pre-factor called k that causes

the model to impose less strain onto CNCs.

��2 =𝐷𝑎2

𝐷𝑡=

1

2 (𝑘) (Ω𝑎2 − 𝑎2Ω) + (𝑘)[

1

2𝜆 (��𝑎2 + 𝑎2�� − 2��: 𝑎4) + 2 𝐶𝐼 ��(𝐼 − 3𝑎2)]

5-5

In the above equation strain reduction factor causes, the strain that are imposed on to particles gets

asymmetric. Implementation of parameter k; it changes between 0 and 1; ensures that fibers orient

slowly. While the SRF model is not objective, it does provide a good match to experimental rheological

data in transient simple shear flow [280-282], and to experimental fiber orientation data in simple shear

injection moldings[283]. As can be seen, the application of this factor (See Figure 5.9), causes the

kinetic of Folgar-Tucker to get slower but still it is not able to predict the experimental behavior

fully. In addition, other disadvantages of this model are shift in peak location, pace of achieving

stress overshoot and change in answer due to change in point of reference. In a nutshell, it can be

said even though application of aforesaid changes causes the improvements in the model, but the

134

answers are not satisfactory. Even though these models provide satisfactory results for suspension

contains long fibers, the results for smaller particles like CNC is not good enough. Extensive

interactions between CNC-CNC, CNC-PVA and higher surface area of CNC in comparison to

longer fibers is very impactful. Also, in these investigations high aspect ratio of CNC causes

physical interaction with PVA chains. As a result, force field must expend higher energy to orient

the fibers. Another point that should be considered in these equations is the spindle shape of CNCs

that upon consideration in the model, might improve the answers. Molecular methods for

considering these peculiar behaviors of CNC has been used but for solving these equation intensive

computer calculations is needed. Nonetheless, considering all these short comings, the model is

still able to provide orientation level at different shear fields.

5-3-9 Mechanical properties of CNC-PVA hydrogels

Figure 5.10 Stress vs. Strain relationship for CNC-PVA freeze-dried samples under compression test

The mechanical properties such as compressive of hydrogels are important parameters, which

intensely affects their applications. It stands out that sequence of orientation observed in rheology

part can be seen here as well. The stress–strain curves of the hydrogels with different amount of

CNCs are shown in Figure 5.10. All hydrogels depicted curves in shape of “J” which is a

manifestation of materials with high compressive strength [284]. Maximum points in graphs shows

the point at which material has collapsed. The compressive strength of the hydrogels increased

135

with an increase in the amount of CNCs. The CNC-PVA sample with 35 g/L CNC showed the

highest compressive stress at 3.4 MPa, which was due to the stiff chains in the strong pore wall.

The values obtained here for compressive stress of the prepared hydrogels was dramatically higher

that values reported in the literature for other PVA-hydrogels. It appears that for reinforcement

purposes for current PVA 5 wt% concentration CNC loading should be higher than 25 g/L.

5-4 Conclusion

CNC-PVA hydrogels at high CNC loadings up to 30 g/L obtained using a relatively simple water-

based processing method without aid of harmful chemical additives. The imaging analysis

suggested that CNC has two types of network inside PVA-salt, at low at concentration of CNC (10

g/L), PVA mediated network formation through bridging between individual CNCs and clusters,

and at high loading of CNC (30 g/L) loading, CNC-CNC and cluster-cluster direct contacts act as

load-bearing junction and contributed in network formation. We referred to the latter

microstructure as double penetrated network structure. These findings were further validated by a

wide range of linear and nonlinear rheological techniques. In terms of linear rheology, two jumps

(i.e., sudden increase) in the values of rheological parameters were observed upon increasing the

CNC loading, signaling two different mechanism for network formation. This is in complete

agreement with imaging analysis. Nonlinear rheology of the systems was studied via rotational

(e.g., flow curve and start-up flow) and oscillatory tests (large amplitude oscillatory shear flow

(LAOS)). LAOS results were analyzed by adopting stress decomposition method with the aid of

Lissajous-Bowditch (LB) plots. In terms of nonlinear viscoelasticity, although a same inter-cycle

behavior (type III) was observed at all CNC loadings, the systems exhibit drastically different

intra-cycle behavior. The extensive discussion based on LB plots by utilizing sequence of physical

processes approach reveals emergence of static yield stress in intra-cycle response of the systems

at high CNC loadings. This approach also enabled used to scrutinize the structural evolution of the

samples in response to different deformations, providing us complementary information about the

microstructure, which is not accessible via linear rheological tests or typical strain sweep tests.

Result of compressive strength measurements showed that all hydrogels depicted curves in shape

136

of “J” which is a manifestation of materials with high compressive strength. Example of

applications of the developed hydrogel can be in the field of cartilage repair.

137


5-5-1 Mechanical properties of CNC-PVA hydrogels

Rheology tests were performed to measure elastic response of the self-healing hydrogel[284,

285].The strain amplitude sweeps are shown in Figure 5.11. The hydrogel samples at two CNC

concentrations of 10 g/L and 30 g/L were subject to an alternate level of strains of 1% and 30% in

repeated cycles to investigate their elastic response. Figure 5.11 shows that the storage modulus,

for both CNC concentrations, decreases by one and two orders of magnitude for 10 g/L and 30 g/L

of CNC, respectively. However, when the strain level decreased back to 1%, the storage modulus

almost recovered the initial values and the hydrogel established its initial strength, which suggest

the structure can depict recoverability. Moreover, the instance change in storage modulus values

upon reduction of shear, depicts strain induced restructuring in the hydrogel. Although level of

returning to the original strength was more pronounced for 30 g/L sample (0.89 vs 0.74). In the

Brownian dynamic simulations of Moghimi et al [286]. it was shown that structure heals faster to

rearrangement of particles at low shear rates. It was observed that for low strain amplitude,

oscillatory shear induces short range rearrangements inside the structure that increase number of

junctions as a function of time, however the process does not change the microstructure

significantly. In this shear regime, shear helps the structures to heal more quickly.

Figure 5.11 a) recovery of 10 g/L sample at strain of 1% and 30% b) recovery of 30 g/L sample at strain amplitude

of 1% and 30%.

1

10

100

1000

0 250 500 750 1000

Sto

rag

e M

od

ulu

s (

Pa

)

Time (s)

Strain 30% Strain 30%

Strain 1% Strain 1% Strain 1%

1

10

100

1000

0 250 500 750 1000

Sto

rag

e M

od

ulu

s (

Pa

)

Time (s)

Strain 1%

Strain 1%

Strain 1%

Strain 30%Strain 30%

a b

138

The hydrogel also goes under cyclic loadings, in which storage modulus recovery was measured

in three repeating cycles. Figure 5.12 illustrates the storage modulus as a function of time for the

two samples of 10 g/L and 30 g/L where the samples go through 1% strain for duration of 1000

seconds followed by resting for 386 seconds. This cycle is repeated for three times where after

each period, the sample gains its original shear viscosity indicating the healing ability of both

hydrogels.

Figure 5.12 a) Storage modulus recovery of 10 g/L sample at strain amplitude of 1% after 3 cyclic strain- storage

modulus recovery b) storage modulus recovery of 30 g/L sample at strain amplitude of 1% after 3 cyclic strain- storage

modulus recovery

139

CHAPTER 6: Self-healing and Collapse in CNC-based Gels

and Suspensions10

CNC hydrogels, while mechanically weak, have unique properties such as high-water contents,

flexibility, and biocompatibility. One of the requirements of CNC hydrogels is to have

mechanically stability and be self-healable. Herein, using Fluorescence recovery after

photobleaching (FRAP) analysis, we assess the gel stability by quantifying the gel collapse and

the level of self-healing of CNC gels with different CNC and NaCl concentrations. We use the

mean signal intensity obtained by confocal laser scanning microscopy to measure the signal loss

of samples made of CNC at 6 g/L, 10 g/L, and 30 g/L concentrations and as a function of initial

gel heights and NaCl loadings. Samples at low CNC concentrations (6 g/L and 10 g/L) experience

a stronger collapse rate under gravity than the rate observed at high CNC concentration (30 g/L).

FRAP is used to qualitatively demonstrate the self-healing ability of samples at various CNC/NaCl

loadings. It is reported that increasing the CNC concentrations hinders the particle mobility and

thus impedes the self-healing process. FRAP recovery analysis shows that when the ratio of

NaCl/CNC increases beyond 0.1, the mobility of the ensemble of CNC particles becomes severely

restricted.

10 Moud, A. A.; Sanati-Nezhad, A.; Hejazi, S. H., Self-healing and collapse in CNC-based gels and suspensions. To

be submitted.

140

Graphical abstract

6-1 Introduction

Hydrogels belong to an important class of materials made in a three-dimension network swollen

in water. Examples of hydrogels in various applications include jelly made of polysaccharides

[287]; contact lenses fabricated out of silicone [288]; and cells existing in the human body, which

are linked through a three-dimensional hydrogel of collagen [289]. These hydrogels display

mechanical properties of fluids and solids simultaneously (i.e., they display viscoelastic

properties).

Hydrogel can be fabricated via different way such as construction from tiny particles through

aggregation or chemical reactions that leads to chemical links [290]. Coagulation, a method of

micro or nanoparticles self-assembly into three-dimensional structures, starts from having a

suspension of particles. Colloids, a terminology limited to suspension with the particle range of 1

to 1000 nm, can coagulate into gel fractals in the presence of attractive interactions. The gel

network, during after formation and during the time of employment, is under the influence of

gravitational stress. Indeed, gels, after being fully formed can be still unstable at low filler loading

141

or weak attraction levels and can dissociate into fragments descending to the bottom of containers.

Moreover, the formation of the gel is also affected by gravity, which can impose a size-limiting

strain on the evolving aggregates [291] or push sedimentation before a network can form [165].

Indeed, the collapse of the gel network, made from different materials and under dissimilar

conditions, can happen for a wide range of gelation mechanisms and attraction levels [165, 166,

292-296]. The mechanical stability of gels under gravity is important as it can affect the gel

properties and morphology.

Qualitatively, collapse may happen at a constant rate or at a rate that decreases with time,

interchangeably. Gel collapse undergoes through slow initial compaction followed by a significant

restructuring and rapid sedimentation, which is ultimately transitioned to slow compaction

reaching a steady height. In the case of gels with tunable inter-particle attractions, the collapse

dynamics can jump from steady sedimentation to three-stage sedimentation as the inter-particle

attraction is decreased [162]. In the literature, the collapse has been attributed to the weak

attractions that influence the network aging properties, such as the network response towards

gravity over time [166, 292, 296, 297] or to the rheological characteristics for strong attractions

[298].

For gels formed with short heights and high-volume fractions (i.e., low porous gels), the gel height

decreases exponentially in time until reaching a steady height determined by the balance of

gravitational stress and network elastic stress. The rate of collapse initially, is a decreasing function

of volume fraction, which can be modeled using Darcy’s law for the collapse of porous materials

[162]. Modeling in the past has been done to connect macroscopic velocity of the fluid flow

through the gel and local displacement of the solid network along the gravitational axis using darcy

law. Peddireddy et al. [138] studied the gelation kinetics and the network structure of CNCs in an

aqueous solution. First, they reported that the CNC gel grows through fractal aggregates until it

reaches to percolation. They observed the macroscopic sedimentation for CNC concentrations less

than 4 g/L, due to the gravitational stress with a change the in the height of the network rate of

collapse for NaCl concentrations more than 50 mM had a little dependency on the NaCl content,

while it was faster for lower CNC concentrations.

Aside from the effect of gravity on the gel formation and stability, the gel self-healing mechanisms

have been studied [85]. The dynamics of particles in a gel network can be efficiently characterized

142

by the diffusion coefficient, quantified through the fluorescence recovery after photobleaching

(FRAP) analysis module of CLSM. In a FRAP experiment, a fluorescent species is irreversibly

photo-bleached in either a circular or rectangle shaped region of interest (ROI). Thus, one can

record the exchange of particles between the bleached and unbleached regimes and correlate that

to the translational diffusion of particles. Bruggen et al. [299] experimentally measured the self-

diffusion in isotropic dispersion of colloidal rods of bohemit with the length and diameter of 325

nm and 46 nm. The dependency of the translational diffusion coefficient on concentration was

evaluated based on a FRAP protocol for volume fractions up to 0.22. It was shown that the

translational diffusion coefficient is a linear changes with volume fraction, up to roughly 0.14,

however, at higher loading levels of particles, the diffusion coefficient decrease to 3 percent of its

values at infinite dilution of particles. Seifert and Oppermann[300] also used FRAP analysis to

measure the diffusion coefficient of PMMA microsphere dyed with rhodamine. FRAP is shown to

be capable of measuring diffusion coefficients in rapidly diffusing systems[300]. The use of FRAP

for measuring diffusion of particles in a solvent with various viscosities has been also

validated[301]. Karvinen et al. [302] studied the FRAP of fluorescein-labeled dextrans, where the

mobilities of different hydrogels are distinguished. FRAP analysis is shown to have the ability to

assess the diffusion accurately in different media [303, 304]. FRAP is a perfect method due to

versatility and accuracy to study gel healing.

In the present study, we study the gel networks experiencing creeping sedimentation at three

distinctive cellulose nanocrystals (CNC) concentrations of 6 g/L, 10 g/L, and 30 g/L. We use the

aqueous solutions of charged-stabilized CNC, where the CNC particles are rod-shaped with the

measured hydrodynamic diameter of 205 nm. To initiate aggregation, we add a monovalent salt,

NaCl, to a final concentration of 86.2 mM. CNC particles experience a strong van der Waals

attraction, when the ionic strength is this high, thus the particle undergo diffusion-limited cluster

aggregation (DLCA) [289] primarily.

We (i) evaluate the diffusion of CNC in suspension and clusters and (ii) quantify the CNC-based

gel collapse at different CNC and NaCl concentrations. We use confocal laser scanning

microscopy (CLSM) to perform FRAP analysis and measure the gel collapse. To the best of our

knowledge, this paper is the first extensive report on probing the diffusion of CNCs inside the gel

and suspensions of CNCs in dilute, semi-dilute, and concentrated regimes. Finally, we show that

there is a connection between the zeta potential, immobile particle percentage, and storage

143

modulus in the CNC hydrogel. This finding paves the way for optimized engineering of the

hydrogel, with the balanced healing ability and mechanical properties. Quantification of the gel

collapse behavior of CNC gel and its self-healing property is critical in many applications,

including water and air filters, oil spill sponges, and tissue engineering [305-307].


6-2-1 Materials

CNC, with the reported length of 100-200 nm and diameter of 5-15 nm, is supplied by InnoTech

Alberta. Based on the manufacturer datasheet, CNCs are extracted with acid hydrolysis process

and have negative charges. NaCl and FB28 were purchased from Sigma Aldrich. Fluorescent FB

28 fluorescent dye that binds strongly to cellulose in general [173] is used here to find the location

of CNCs.


We employ Ultra-sonication (125 W Qsonica Sonicators Q125 Sonicator, Qsonica) for suspending

CNCs in DI water for 10 mins. To prevent overheating on the surface of CNCs, the sonication is

done in an ice bath (i.e., a bath with constant 0 °C) as the surface charge of CNC particles is

sensitive to temperature [50, 308]. The level of dispersion is monitored through tracking zeta

potential values with dynamic light scattering (DLS). It is found that supplying 1000 joule per

gram of CNC is sufficient for a complete sonication of CNCs that brings down initial CNC clusters

to their individualistic sizes. We use the pH-meter (Mettler-Toledo 135 International Inc.,

Columbus, OH, USA) to measure pH of CNC-water system to be 6.8. The ionic strength of CNC

suspension is changed through the addition of a concentred 200 mM of NaCl.

We examine the binding of FB28 dye to CNC particles by sonicating them in DI-water first and

then let the suspension flow through a filter paper (Whatman™ Quantitative filter Paper) with the

pore size of 100 nm. The approach is adopted from previous works where the binding of FB28 to

chitin is determined [309-311]. We expect that, in the case of no binding, due to its size, the FB28

can easily pass through the filter, while CNC particles get trapped. We rinse the filter paper several

times with deionized water and then inspect the paper under UV light and under confocal

144

microscopy. Figure 6.15 shows that after filtration, the separated CNCs on the filter has

maintained FB28 (i.e., FB28 binds to CNC).


6-2-3-1 Dynamic light scattering

Nano-Zetasizer (Malvern Instruments, Nano ZS, Malvern, UK) is employed to measure the size

of CNC particles suspended in DI water. A He-Ne laser (Spectra Physics 2020, with the

wavelength of 𝜆 = 633 𝑛𝑚) and a backscatter detection system at 173° is used to capture the

dynamics of CNCs. The backscattering detection system, which eliminates the multiple scattering

phenomena of scattered light, allows for measuring the translational diffusion of CNCs of highly

concentrated samples [312]. As opposed to detection at 90°, at a high scattering angle, the

contribution of rotational diffusion can also be neglected, and the translation diffusion can be

estimated [312]. For cellulose, the refractive index, and the extinction coefficient at laser

wavelength of 632.8 nm are, respectively, 1.46869 and 0. For more accurate measurements the

refractive index of the material must be give as an input to DLS device. The relationship between

the refractive index (𝑛) of cellulose and wavelength (𝜆) of incident light can be found as [313]:

𝑛2 − 1 =1.124𝜆2

𝜆2 − 0.011087 6-1

Results of DLS studies are used to obtain an approximate quantitative assessment of cluster sizes

in CNC hydrogel.

6-2-3-2 Confocal laser scanning microscopy

We use FB28 dye for the imaging of CNCs. The concentration of the dye needs to be selected

below a threshold value to neither influence the behavior of CNCs nor the gel. In our experiments,

for samples with 10 g/L CNC concentration, zeta potential with addition of FB28 up to 500 ppm,

does not change. However, this concentration of dye is sufficient to provide enough fluorescence

to CNCs [172]. Physical mixing of the dye with CNCs for roughly 30 mins for incubation of the

dye to CNCs is needed to be done in a dark environment. CLSM monitoring is performed for

suspensions at different NaCl and CNC concentrations. CLSM measurements are carried out with

an inverted Nikon confocal microscope (Ti-A1R) equipped with the apochromatic lens objectives

of 10X (NA=0.45) and 20X NA=0.75) providing the resolutions of 500 and 300 nm, respectively.

145

For 10X optical lens, numerical aperture and working distance has been reported 0.45 and 4 and

for 20x optical lens 0.75 and 1 mm, respectively. The microscope’s galvanometer-based scanner

enables achieving high-resolution images up to 4096 x 4096 pixels.

Conventionally, in a FRAP experiment, a fluorescent species is irreversibly photo-bleached in

either a circular or rectangle region of interest (ROI). Thus, one can record the exchange of

surrounding unbleached particles and bleached particles in ROI at a pace that is controlled by the

mobility and interaction parameters involved between the medium-particle and particle-particles.

This phenomenon leads to a recovery of the bleached region. If we assume that all particles are

100% mobile, the recovery after passage of some time will be complete. However, the recovery

will not be always 100% and some of the labelled particles in ROI are immobile; they neither

contribute to overall recovery nor they give up their site for other unbleached particles. Throughout

this work, we designated them as immobile particles or immobile particle percentage. Confocal

laser scanning microscopy (CLSM) is usually employed to track the temporal evolution of the

recovery rate, using the same laser for capturing images and bleaching but operating at different

intensity levels. Subsequently, analyzing information can be done by fitting a model onto

fluorescent recovery curves. The underlying assumptions for the FRAP model, as well as how it

fits vary significantly between different approaches, but eventually it boils down to fitting the

recovery rate indicating how species diffuse into the bleached region [314].

Conventionally, about ten separate FRAPs are taken and averaged out to generate a single FRAP

recovery curve. To mitigate this issue, 10–30 adjacent points in the slower part of the curve are

averaged. For all sets of FRAP measurements, we use a laser with a wavelength of 405 nm under

100% laser power for one loop, for a duration of ~1-second stimulation. Depending on the recovery

rate of the samples, different acquisition timing is selected. We select offset to be zero and adjust

the gain to obtain the best resolution. In some cases, the 2x line averaging is also used to reduce

the noise. To be able to capture the diffusion of CNC rods, two frames/sec is selected, and the size

of the visualization cube is set at 512 𝜇𝑚 × 512 𝜇𝑚. Pinhole with 1.8 airy units is selected to get

an optical sectioning value of ~ 16.25 𝜇𝑚. Diffraction-limited axial dimension is key for optical

sectioning. Depending on the numerical aperture and pinhole airy unit, the optical slice thickness

in CLSM can reach thicknesses as low as 0.5 micrometers. Moreover, the pixel size of 0.29 𝜇𝑚

and the ROI with the size of 10 𝜇𝑚 are chosen. For all measurements, the PL APO 10x (NA=0.45)

146

lens is used. For FRAP, the ratio of FB28/CNC concentration is fixed at 4×10-5 for all CNC

samples. One 𝜇𝑚 steps in the z-direction are taken to generate a 3-D micrograph of the gel and the

CNC in NaCl-free suspension. In all experiments, the samples are placed between two coverslip

glasses. Other information such as the protocol used and equations are discussed in the

supplementary information section.

Figure 6.1 Gaussian bleached area (circular) immediately after bleaching in the sample of CNC with 45 g/L

concentration and 20 mM NaCl

Figure 6.1 shows the fluorescent intensity immediately after bleaching of the sample with CNC

at 45 g/L concentration and 20 mM NaCl for two selected ROIs with 10 and 20 µm diameter.

Considering the width of bleached ROI with respect to the rest of the plane and the fact that the

suspension or the gel is monitored between two glass covers, the system can be considered

uniformly bleached and 2-dimensional (See Figure 6.1).

For the gravity collapse experiments, a circular ROI is selected with a nominal diameter of 500

µm. The optical section is 8.21 µm with an optical resolution of 500 nm. The pixel size is 2.32 µm

with the pixel dwell time of 2.18 µs. The respective scan size is 512×512 µm, and the graphs are

captured with PL APO 10x (NA=0.45) under the line averaging of 2X with the pace of two frames

per second. Laser power is kept constant throughout the whole experiment at 10%. Each point is

147

repeated 3-6 times under the duration of the 240-second experiment. Pinhole is kept constant at

1.8 airy unit in all experiments.

6-3 Results and discussion

6-3-1 Confocal imaging accuracy verification

We first evaluate the accuracy of CLSM images. We conduct the calibration experiment using

FluoSpheres™ Polystyrene Microspheres with a diameter of 1.0 µm and yellow-green fluorescent

(505/515), which is typically employed for tracer studies. The particle pictures were taken with PL

APO 10x, 20x, and 40x and their corresponding size distribution is shown in Figure 6.2. The size

distributions of 1.46 ± 0.44 µm, 1.3 ± 0.41 µm and 1.12 ± 0.15 µm are evaluated for PL APO 10x,

20x and 40x, respectively. The trend shows an improvement in the precision of measurements

when lenses with higher magnifications are chosen. Optical resolution given by microscope for PL

APO 10x, 20x and 40x are 0.69, 0.41 and 0.27 µm, respectively.

PL APO 10x PL APO 20x PL APO 40x

148

Figure 6.2 Distribution of Polystyrene microparticle sizes and their respective CLSM micrographs presented at 5 µm

scale bar.

6-3-2 Relationship between the CLSM signal and the CNC concentration

We establish a direct correlation between the measured signal strength and the CNC concentration

for image quantification. The signal (A) is correlated to the fluorescence concentration, which the

signal absolute values is under influence of gain and intensity of the laser. The probable values of

A can range from 0 to 4050 in integer steps for the images. Generally, it is recommended, as

intensity changes with the depth of structure visualization, to adjust the set of gain and laser

intensity in a manner that the whole illumination 2-D or 3-D space is efficiently covered by signal

distribution and to avoid the pixel saturation. To relate signal to the CNC concentration, we

determine f(A) for NaCl free systems labeled with FB28. In practice, the number of dye molecules

fluctuates in the plane of focus; hence, the average fluorescence intensity spatially changes.

Fluctuations in signal strength can be defined by following equation:

149

σ = ⟨A⟩−1√n−1 ∑(Ai − ⟨A⟩)2

n

i=1

6-2

Where ⟨A⟩ is the spatially averaged signal and 𝐴𝑖 is the value in pixel 𝑖. The smallest amount of 𝜎

is strictly determined by the intensity fluctuations of the dye. It will decrease as a function of the

illuminated volume if the volume gets bigger or the duration of illumination becomes longer.

Typically, illumination volumes are strictly dependent on the optics and will change upon changes

in the lenses; however, it is invariant with respect to the pixel size. The duration of illuminations

is dependents on scanning speed, i.e., frames taken per unit of time. 𝜎 value can be diminished

through increasing the FB28 concentration or through averaging values of repeated experiments.

For the case of repeated experiments, care should be taken to keep the dye-CNC concentration

spatially identical for the duration of the experiments. For the case of solutions with alike dye

concentrations, the use of similar gain and laser intensity values is not possible. Therefore, it is

crucial to develop a relationship between these factors. Pragmatically, the most accurate way is

determination of the absolute values of signal through monitoring a standard solution, akin to

scattering experiments.

Here, we use FB28 solutions with known concentrations as the standard. The FB28 concentration

needs to be sufficiently high to give a significant signal, but not so high where it can modify the

structure of CNC gels and their colloidal behavior. The possible influence of FB28 on the structure

is investigated by measuring the turbidity of gels formed containing different amounts of FB28.

We find that the structure of CNC gels is not significantly influenced by FB28 for concentrations

up to 500 ppm (g/g). Unless specified, we have used the CNC/FB28 concentration equal to 0.01

for obtaining the calibration curve for all experiments. The signal and its standard deviation are

the identical for solutions containing FB28 at CNC concentrations up to 40 g/L. This shows that

gluing to CNC does not influence the fluorescent property of FB28. The identical experimental

observation is made after gel formation. Figure 6.3 shows the average signal intensity as a function

of the CNC concentration for a concentration ratio of CNC to FB28 equal to 0.01. The mean value

of intensity extracted from Figure 6.3 is 489, 833, and 2432 a.u for 6 g/L, 10 g/L, and 30 g/L

sample, respectively. A linear relationship is found A and C. Practically, establishing a connection

between emission intensities and the CNC concentration through a calibration curve is possible

using the same solvent (water here), temperature, and pH. Determination of CNC concentration

150

can be done through comparing signal values of unknown samples to the calibration curve or in

some cases extrapolation.

Figure 6.3 A linear relationship between the mean signal value and the concentration of CNC. The concentration ratio

CNC/FB28 for all samples is set at 0.01. The measurement is done at 1 mm above the base of the petri-dish.

A similar linear relationship is also reported for the case of globular protein b-lactoglobulin

concentration and confocal microscopy signals [149]. The CNC concentration-intensity data in

Figure 6.3 indicates that the number of CNC in each window of observation of confocal for 35

g/L should be 1.75 times the 20 g/L. In our experiments, we observe this phenomenon as signal

ratio that holds between 35 g/L and 20 g/L and is approximately 1.75. Due to the gravity collapse

or the recovery of photobleached area we can assign the mean signal value per gram of CNC and

later use it to find the concentration changes. There are varieties of parameters such as the depth

that laser has travelled, the concentration of dye, and the quantum yield of dye that can influence

the signal observed in fluorescence spectroscopy. Anything that can quench the process of

transition states of the dye molecules during excitation-emission can impact the measured intensity

[315].

151

When discussing intensity, Beer’s Law should be employed (following equation), in which I is the

intensity of signal, ε is molar absorptivity associated with the solution, b is the path travelled by

the laser, C is the concentration of FB28 here, and P0 is the power of the laser [316].

I = 2.303 K′ε b C P0 6-3

𝐾′ is a constant which depends on many factors, including the geometry and FB28 quantum yield.

The intensity fluctuates linearly with b, C and P0. Increasing laser power, means higher number of

dye molecules will receive photons and more received photons means higher intensity of emission

at constant quantum yield.

The utility of equation 6-3 at absorbance of 0.05 or higher is lost as the there will be negative

deviation from the standard curve. Another unique condition of fluorophore containing solutions

is that ground-state molecules can reabsorb emitted photons and get excited. However, upon

calibration of the solution containing fluorophore, establishing a connection between the signal

intensity and the concentration is possible and accurate. The results of this connection have been

shown in Figure 6.3. The ascendant trend of signal strength at constant laser power, with the

increase in CNC concentrations, depicts a perfect distribution of fluorescent agents across the

system (data not shown here).

6-3-3 Quantitative analysis of CLSM images of CNC gels

Since 𝐴 is proportional to 𝐶, the CNC concentration distribution 𝑓(𝐶) can be obtained from the

signal distribution 𝑓(𝐴). Figure 6.4 depicts the variation of CNC-NaCl volume fractions as a

function of NaCl concentration. Figure 6.4 indicates the modulation in the structural

inhomogeneity of CNC after the NaCl addition. Results depict that a region with a lower CNC

concentration tends to occupy more space as we increase the NaCl concentration. Clusters grow

till they reach a certain size, as 𝑅𝑐~𝑎𝜑1/(3−𝑑𝑓), at which point they span space, and form a gel

[317]. In this equation, 𝑅𝑐 is the radius of the cluster with the fractal dimension of df and 𝑎 is the

size of primary particles.

152

Figure 6.4 CNC concentration distribution for original suspension of CNC 5g/L with the addition of NaCl at 10, 15,

20, 25, and 30 mM

6-3-4 Dynamicity of CNC gel and eventual gel collapse

In our previous study [173], we demonstrate that gels of the CNC-NaCl hybrid system are dynamic

at the micro-level. Simulations have also shown that clusters are continuously rearranged,

compacted, and form denser structures [216]. One offshoot of dynamicity is the collapse of CNC

gel. In the following, we formulate a framework that can tell, with time, the rate of gel collapse at

different height levels. As the pace of fall is sensitive to collapse timing, the location of the focal

plane, and the initial gel height, all measurement are done under identical conditions. Moreover,

we normalized all graphs with respect to the initial measuring intensity at time zero.

Figure 6.5a-b shows the time evolution of normalized mean signal intensity representing the gel

collapse with respect to changes in CNC and NaCl concentrations. In the case of zero NaCl, the

signal intensity fluctuates over 240 seconds for CNC 6 g/L, which can be due to the Brownian

motion of particles that disallow particles to settle. Once the NaCl concentration is increased to

153

17.24 and 25.86 mM, the mean intensity decays in time due to the appearance of bigger clusters,

which are resulted from the screening of negative charges. The size of the cluster at this stage is

big enough to cause shifts in the behavior of the particles. Moreover, it appears that increased

amounts of NaCl of CNC enhance the rate of collapse. Unlike the low CNC concentration (6 g/L),

the addition of NaCl to a high CNC concentration (30 g/L) decreases the pace of structural decay

(Figure 6.5b). This can be due to gel formation at high CNC concentrations. The final normalized

mean intensities after 240 s for samples of 8.62, 17.24, 25.86, and 43.1 mM of NaCl are 0.93, 0.89,

0.87 and 0.85, respectively reflecting that NaCl increases gels resistance towards gravity effects.

Figure 6.5 Mean signal intensity decay for the CNC concentrations of (a) 6 g/L and (b) 30 g/L at 0-43.1 mM NaCl

loadings. Results are captured at 1 mm location above the base of the petri dish (the initial gel height is 5mm) with

a 10x apo lens (NA=0.45) and at the timing of 5 min after gelation.

Bartlett et al. [162] show that gels, fabricated via weak physical bonds, settle under their own

weight, following two distinct regimes. For an initial lag time, the formation of a space-spanning

network resists compaction. This solid-like behavior persists only for a limited time. However, our

experiments show a constant decline in mean signal intensity, suggesting that the macro-scale

observations [162] in microstructural changes cannot be simply extended to the micro-scale.

154

For 30 g/L, we previously conducted rheological measurements where it is shown that after adding

8.5 mM NaCl into the system, gelation is reached as 𝐺′ > 𝐺′′ over the entire frequency range.

Therefore, we expect to have a gel at high CNC concentrations and all the NaCl concentrations

studied here. The decreasing rate of gel collapse as the NaCl concentration increases depicts that

the gel becomes sturdier, which is consistent with the previously reported 𝐺′ value and is an

increasing function of NaCl concentration until it reaches a plateau.

Figure 6.6 Mean signal intensity decay for 30 g/L CNC concentration and 43.1 mM NaCl concentration at different

gel initial heights. Results were captured at a 1 mm location above the base of the petri dish with a 10x apo lens

(NA=0.45) 5 min after the introduction of NaCl.

It is noteworthy to mention that there are slow changes with respect to time in the average signal

intensity across the samples. The gel compaction rate might be different at the macro level versus

the micro-level. Harich et al. [318] report that the speed of microscopic collapse for poly-

methylmethacrylate dispersed in cis-decalin is 8 µm/s, which is an order of magnitude higher than

the speed of the macroscopic collapse 0.6 µm/s. Manely et al. [319] show that the macro scale

height of gel exponentially decays in time, where the gel collapses are faster at the higher initial

155

gel height. Our results, as shown in Figure 6.6, also confirms the increase of the collapse rate as

the initial height of gel increases.

Using the relationship established between the CNC concentration and the mean signal intensity,

the collapse rates of CNC in terms of grams are estimated and shown in Figure 6.7. The addition

of NaCl slows down the gel collapse for CNC 30 g/L sample, as shown with a decrease in the CNC

loss concentration as the NaCl concentration increase (Figure 6.7a). This agrees with the result of

rheology in which the storage modulus of the gel increases with the NaCl loading. Furthermore,

the CNC loss concentration is an increasing function of NaCl concentrations for the low CNC

concentrations of 6 g/L and 10 g/L, which is attributed to the increasing size of CNC clusters with

the increase in NaCl concentration.

Figure 6.7 Average loss of CNC out of the control box over 400 s period of the experiment for CNC 30 g/L (top)

CNC 10 g/L (middle) and CNC 6 g/L (bottom).

Dynamic light scattering is utilized to assess the sizes of clusters after the NaCl addition. Figure

6.8 shows that with the addition of only 0-70 mM NaCl into the suspension of CNCs with 0.5 g/L

concentration, the z-average sizes of CNC grew from 125 nm to 400 nm. These results roughly

show what sizes of clusters are expected in the gel.

156

Figure 6.8 Depiction of changes in cluster size and mobility of CNCs in the CNC-DI water suspension system on

semi-logarithmtic scale. The minimum in the z-factor can be due to the retraction of double layer.

6-3-5 Dynamics characterization of CNC clusters in gel using FRAP

We use FRAP to analyze the extent of CNC particle mobility in the suspension. FRAP analysis

can be easily quantified as a function of recovery (𝜏1/2), defined as the period required for region

of interest bleached location to recover halfway through between the original and intensity at

steady state [320-323]. This protocol is simple as 𝜏1/2 can be easily read from recovery curves

[320-323]. Note that the half time recovery can be affected by parameters such as size and shape

of the ROI and the protocol used for bleaching [314]; hence, for sake of comparison across studies

it cannot be used. Contrary to the half time recovery, 𝐷𝑟𝑛 provides a quantitative assess of particle

movement through diffusion [324, 325]. Accurate. estimates of diffusion are also a vital starting

step for reaction-diffusion molecular analysis [314, 326]

We utilize the two-dimensional (2D) FRAP equations developed by Axelrod for a Gaussian laser

[324] and by Soumpasis [325] for uniform circular laser bleaching [325]. The model relates 𝐷𝑟𝑛,

𝜏1/2 and 𝑟𝑛 for an isotropic diffusion system as:

𝐷𝑟𝑛 = 0.224𝑟𝑛

2

𝜏1/2

6-4

157

In this equation, 𝑟𝑛 is the radius of the bleached area, and the coefficient 0.224 has been

numerically determined. Herein, we follow the protocol provided by Kang et al. [314] to adjust

equation 6-3 for the gaussian bleach profile, as suggested by equation 6-5. The details of their

approach have been briefly described in the supplementary information.

𝐷𝑟𝑛 =𝑟𝑛

2 + 𝑟𝑒2

8𝜏1/2

6-5

t=0s t=10s t=15s

Figure 6.9 (a-c) Frap recovery curves for CNC with a concentration of 6g/L at 0, 20 mM, and 86.2 mM NaCl loadings.

(d-e) Temporal-spatial CLSM 3-D images of the sample with 6g/L CNC captured at 0, 10, and 15 seconds after

bleaching. Visualization box size: 148.347 µm×148.347 µm, resolution 500 nm.

158

Figure 6.10 (a-f) FRAP recovery curves of samples with the CNC concentrations of 45 g/L and 30 g/L at various

concentrations of NaCl (0, 17.2, 34.4, 51.7,70 mM)

Figure 6.11 (a-c) Variation of diffusion coefficient as a function of NaCl loadings (0, 8.62, 20, 86.2 mM)

depicted on left-hand side and immobile particle percentage as a function of NaCl loadings (0, 8.62, 20, 86.2

mM) depicted on right-hand side of each figure

159

Figure 6.12 Time series of FRAP done on CNC 45 g/L sample with PL APO 10x optic (NA=0.45). The sample

immobile fraction is 0% after 40 seconds. Scale bar;10 µm resolution 500 nm; and ROI size 10 µm.

160

Figure 6.13 The FRAP recovery of the sample with 45 g/L CNC and 20 mM NaCl that over a span of 8.4 min has

not healed. The height of the 3-D image shows changes in time that were continuously captured. The visualization

box size is 140 µm ×140µm and resolution is 500 nm.

To use Equation 6-5, one must determine 𝑟𝑒 as the effective radius of the bleached profile. Effective

radius determination happens through sketching bleaching profile and obtaining 𝑟𝑒 as the radius at

14% of bleaching depth. For obtaining values of 𝐷𝑟𝑛 in Figure 6.9-11equation 10 is used. Figures

6.10 (a-c) visually shows the FRAP experiments for 6 g/L sample over a period of 15 seconds.

Figures 6.10(d-f) depicts the FRAP recovery curves for a dilute suspension of CNC with 6 g/L

concentration at different NaCl concentrations. As expected, increasing NaCl concentration retards

the mobility of CNCs and increases the population of immobile particles in clusters. In fact, upon

introducing NaCl, 𝐷𝑟𝑛 decreases from 5.5±1.4 to 0.4±0.28 𝜇𝑚2/𝑠 as NaCl loading increases from

0 to 86.2 mM. Similar results are observed in other CNC concentrations; however, the trend is

different. Figure 6.11 depicts the FRAP recovery data for 45 g/L and 30 g/L CNC with various

concentrations of NaCl. 𝐷𝑟𝑛 is 0.3±0.15 𝜇𝑚2/𝑠 based on 10 repetitions at 3 different points for

CNC 30 g/L with 70 mM NaCl. Moreover, it appears that the percentage of mobile particles also

decreases when the NaCl/CNC concentration ratio passes a certain threshold. Inspecting the values

of 𝐷𝑟𝑛 in multiple locations reveals the existence of inhomogeneity in the gel, with regards to

diffusion (See Figure 6.11, graphs in each set of CNC and NaCl concentration has been sketched

in 3 points). This is attributable to non-uniform distribution of NaCl during the process of gel

formation. Jonasson et al. [327] report that gels are heterogeneous at the microscale, and therefore,

the local diffusion properties can vary with the position. The difference in the diffusion coefficients

observed at different CNC concentrations (i.e., 6 g/L, 10 g/L, and 30 g/L) can also be due to

different degrees of mobility inside the gel network with different level of fractal dimensions.

Immobile particle percentage depicted in Figure 6.11 can be calculated as (𝐹𝑖 − 𝐹𝑠𝑠)/(𝐹𝑖 − 𝐹0)

where 𝐹𝑖, 𝐹0, and 𝐹𝑠𝑠 are the normalized intensities before bleaching, immediately after bleaching,

and at the steady-state condition. The fraction of immobile particles for cases in Figure 6.11 is

measured to be 0-0.91 for different sets of CNC and NaCl loadings. For all CNC loadings, there

is a threshold of NaCl concentration, after which the immobile particle percentage increases to

about 90% and 𝐷𝑟𝑛 decreases. The NaCl concentration lies around 10 mM for 6 g/L of CNC

concentration and about 20 mM for the CNC concentrations of 10 g/L and 20 g/L.

161

Figure 6.12 and 13 show the healing process after bleaching for CNC at the concentration of 45

g/L without NaCl and with 20 mM NaCl, respectively. 2-D images of Figure 6.12 show the

changes in the x-y plane while the 3-D images of Figure 6.14 illustrate the changes in the x-y

plane while times in terms of height changes. The recovery for the case of no added NaCl is fast

(about 20 seconds), while the addition of NaCl considerably slows down the recovery to more than

eight minutes.

As reported in Figure 6.11, the diffusivity of CNC slows down as the CNC concentration increases

from 6 g/L to 30 g/L (compare diffusion coefficient values across the images d-f). This is expected,

as an increase in the concentration of CNC rods hinders the CNC particle movements. Gelation

can occur at 50 g/L due to the increase in concentration.

The concentration (volume fraction) of particles in suspension (𝜑) and the aspect ratio (length to

diameter) of the rod particles (𝑟 = 𝑎/𝑏) be employed as a criterion to distinguish three regimes of

diluted, semi-diluted, and concentrated networks. Generally, for concentrations, 𝜑 << (1/𝑟2),

particles structurally or dynamically do not sense one another; thus, the network is dilute. When

1/𝑟2 << 𝜑 << (1/𝑟), each particle makes a few contacts with neighboring particles, and the

network is semi-dilute. Finally, when the concentration is high 𝜑 >> (1/𝑟), the rod rotation is

Figure 6.14 (a) Measured diffusion coefficients normalized by the 𝐷𝑖𝑛𝑓𝑖𝑛𝑖𝑡𝑒 , as a function of(𝑎/𝑏)2𝜑. Note that

(𝑎/𝑏)2𝜑 is proportional to the number of rods in the volume 𝐿3 with L the length of the rods. The results of Brownian

dynamics and Edwards-Evans equation are also given as a function of concentration. (b) universality graph that

connects data of zeta potential, storage modulus, and immobile fraction obtained through FRAP analysis. The lines

in the above graphs of FRAP data, zeta potential values, and storage modulus are drawn as a guide to the eye.

162

limited by particles in its vicinity. Models applicable to motion of single rods can be employed to

find out about effect of particle geometry and concentration on changes in dynamic and mobility

of particles. The Doi and Edwards [328] tube model can enable the researcher to forecast the

diffusion based on concentrations of CNCs. The tube model suggests that, upon increasing the

concentrations of rods, initially particles lose their ability to move along the vector perpendicular

to their main axis. This is a reasonable assumption for semi-dilute regime and high aspect ratio

combination. Following that, Edwards and Evans develop a Green’s function formalism to

estimate the dependency of mobility parallel to the main axis on concentration using aspect ratio

of b/a where b is the diameter, and a is the length of particles. Model [296] considered the

diffusion of a rod particles moving through a network of rigid rods. Excluded volume of

neighboring particles in their model slowed down the particle translation motion at volume

fractions of 𝜑~(1/𝑟) . In addition to this model [296], we also use the fitted model to the results

of Brownian dynamics simulations for hard spherocylinders, as reported by Lowen [329].

Figure 6.14a shows the diffusion coefficient normalized by the diffusion coefficient at infinite

dilution (here at low CNC concentration of 0.5 g/L) as a function of (𝑎/𝑏)2𝜑 in diluted and semi-

diluted regimes. In this graph, we also plot the predictions from the Brownian dynamics

simulations of Lowen [329] and the Edwards and Evans theories. The FRAP analysis and both

theoretical models show that the diffusion is a decreasing function of particle concentrations. The

models overestimate the diffusion values at diluted concentrations.

Figure 6.14b illustrates the universality graphs in which the zeta potentials are connected to the

immobile particle percentages and the storage moduli. The region, when there is a dip in absolute

values of zeta potential, is accompanied with a surge in the immobile particle percentage and the

storage moduli. This graph depicts the similarities in the trends of parameters as a function of

NaCl/CNC concentration fraction that manifests CNC dynamics inside the gel.

6-4 Conclusion

We studied the dynamics of CNC gel and suspension under gravity for various CNC

concentrations and salt loadings. To quantify the fluorescence images, we established a

relationship between the signal strength and the FB28/CNC concentrations. We evaluated the

sample collapse using CLSM mean signal intensity where the signal loss, due to the falling of CNC

163

clusters, were quantified for CNC at concentrations of 6 g/L, 10 g/L, and 30 g/L sample, and at

various NaCl loadings. Results showed that at low CNC concentration, the signal loss increases in

time, while it decreases at the 30 g/L CNC concentration. These opposing trends could be due to

the gelation of CNC-NaCl samples at high CNC concentrations. We performed FRAP analysis to

probe the diffusion of CNCs at a different level of concentrations (i.e., dilute and semi-dilute

regimes). After the addition of extra CNCs into a dilute system, the rate of particle diffusion slows

down. Furthermore, we measured the rate of CNC particle diffusion in gels and its dependency on

CNC and NaCl concentrations. FRAP experiments also showed that the system is dynamic at all

CNC and NaCl loadings and can be healed regardless of used concentrations. We finally illustrated

the connection among immobile particle percentage, zeta potential, and storage modulus through

CNC/NaCl concentration ratio. The findings of the present study can pave the path for engineering

hydrogel with an optimized level of healing and mechanical properties.

164


6-5-1 Theory of FRAP

FRAP as a common confocal tool for assessing dynamics of entities in situ, holds an important

role among other methods, such as single-particle tracking, fluorescence correlation spectroscopy,

and image correlation spectroscopy [330-335]. While each technique has its own weaknesses and

points of strength, FRAP has many advantages over alternative methods. For instance, FRAP

enables the researcher to find mobile and immobile percentage of particles or dye molecules which

is not accessible through other techniques.

Protocol for performing FRAP

The following are the general guideline that are key for the FRAP data acquisition, presented here.

(1) Set the image in confocal software with an appropriate zoom to the zone of interest. With

CNC gel, a 1024 × 1024-pixel image at 10X Apo lens electronic zoom is initially a good

point for starting the FRAP experiment.

(2) Employing ROI selection tool, trace a circular region for bleaching in 2-D or 3-D

visualization space.

(3) A general rule of thumb to establish a stable fluorescence baseline before initializing FRAP

is to gather 3 pre-bleach images if taking pictures at a rate of roughly 1 frame/second is

used.

(4) The number of repetitions that are required will hinges on intensity of laser, pace of

scanning, the photo-stability of the fluorescent dye, and the quickness of diffusion of the

material under investigation.

(5) After choosing a bleach laser (Nikon provides 4 options) and set it at the maximum power.

(6) The bleaching phase must be sufficiently short to limit recovery of ROI during bleaching.

As a rule of thumb, it is recommended that the total bleaching time be at least 15 times

shorter than the period of recovery [336].

165

6-5-3 Dye binding to CNCs

Figure 6.15 Filtered CNC-FB28 in condition (a) Before and (b) after exposure to UV light

Aromatic molecules have tendency to absorb to the surface of carbohydrate-based polymers due

to van der Waals forces (CH–p interactions) and the hydrophobic effects. These CA interactions

is the main reason behind bindings onto CNC in the aqueous systems. As electrostatic interactions

, van der Waals interactions and hydrophobic effects are all reported in the literature as the possible

mechanism of dye adsorption onto cellulose, it is possible they are all in our system responsible

for dye attachment on CNCs. Nonetheless, the adhesion between dye and CNC is strong enough

to resist the effect of rinsing and centrifugation.

6-5-4 Additional considerations

To facilitate analysis, recovery curves can be normalized in one of two ways. Firstly, the values

get converted to make the pre-bleached intensity value equal to 1. This modification allows for

making comparisons between samples with varying levels of brightness as it easily describes the

data as ratio of its original value. Secondly, for simplification purposes, the curve can be fully

normalized so that the initial intensity becomes one and the intensity at the time of bleaching equal

to 0. This can represent the data as a proportion of recovery with 1 depicting full recovery and 0,

no recovery. If measurements are accessible for each bleach set-up, a background and a suitable

reference region will be selected to characterize unintentional bleaching region, a typical approach

that is called the double normalization:

166

𝑁𝑜𝑟𝑚(𝑡) =𝑅𝑒𝑓𝑝𝑟𝑒−𝑏𝑙𝑒𝑎𝑐ℎ

𝑟𝑒𝑓(𝑡)

𝐹𝑅𝐴𝑃(𝑡)

𝐹𝑅𝐴𝑃𝑝𝑟𝑒−𝑏𝑙𝑒𝑎𝑐ℎ

6-6

167

CHAPTER 7: Summary and conclusion

In this thesis, CNC particles were used to generate novel functional biomaterials. The highlights

of our findings in terms of CNC gelation, mechanical properties of CNC hydrogels, and the gel

healing characteristics are as follows.

A. CNC gelation

1. CNC in the presence of a threshold amount of salt (whether monovalent or divalent)

aggregates. The state of CNC aggregation in the presence of sodium chloride (NaCl) was

monitored using TEM and CLSM. CLSM micrographs revealed patterns in CNC clusters

with the presence of regions with both colloid-rich and colloid-poor patterns. Moreover, a

dynamic structure for gels, continually rearranging over the course of time, was recorded.

Zeta potential data, coupled with CLSM images, confirmed the impact of NaCl on the gel

formation of CNCs.

2. SEM images showed that the gel mesh size could be influenced by a variation in the CNC

concentration with constant NaCl content. An increase in the concentration from 7.5 g/L

to 15 g/L led to a decrease in the mesh size, from 1.4 µm to 1.2 µm.

3. The divalent salt (MgCl2) was found to be stronger at pushing the zeta potential values

towards lower absolute values than monovalent salt (NaCl). According to the Schulze-

Hardy rule, and assuming CNCs as highly charged particles, the ratio of critical aggregation

concentration for MgCl2 system was found to be approximately 64 times smaller than the

value obtained for the case of NaCl. The observed trend was well predicted by the Schulze-

Hardy rule, even though a deviation was observed in the predicted aggregation onset for

highly charged and weakly charged particles, respectively. This deviation might be due to:

(i) dealing with non-spherical particles in the present system, (ii) neglecting ionic radius,

and/or (iii) limited confocal micrograph resolution (300-500 nm).

4. A global parameter (i.e., NaCl/CNC ratio) was found to be capable of linking between the

zeta potential values and the onset of gelation for different CNC and NaCl concentrations.

5. It was found that sonicating the coagulated samples shattered CNC clusters, and if NaCl is

added, the resulting shattered clusters would not aggregate again.

6. Molecular dynamic simulations showed that two CNCs rods could be brought to as close

as of 3.0-3.5 nm to another but was heavily influenced by the type and quantity of salt

168

(NaCl and MgCl2). The ratio of MgCl2/CNC at 0.05 was found to be critical for CNC

particles to approach each other.

7. It was found that fractal dimensions obtained through TEM and CLSM image processing

could return similar values, although the values of fractal dimension could vary with the

chosen intensity for confocal imaging.

8. The fractal dimension in 2-D estimated based on TEM and CLSM micrographs was found

to be an increasing function of NaCl /CNC concentration ratio.

9. No connection was found between the fractal dimension and the experimental rate of

collapse. We obtained equal fractal dimensions as a function MgCl2/CNC ratio for two

cases of 5 g/L CNC and 15 g/L CNC with 52 mM MgCl2. However, the rate of gel collapse

was different.

B. Rheology and mechanical properties of CNC hydrogels

1. The dynamic colloidal behavior and the stability of aqueous CNC suspensions and their

correlation with the nonlinear viscoelastic properties of the CNC gel structures in the

presence of different NaCl concentrations were investigated. The change in ionic strength

of cellulose nanocrystal (CNC) suspensions contributed to the respective colloidal

behavior, such as stiffness and fractal gelation.

2. The nonlinear rheology of the suspensions/gels was used to correlate the macro-mechanical

viscoelastic response of the CNC/NaCl aqueous systems to the nano-scale structural

features. The intra-cycle viscoelasticity, explained by Lissajous-Bowditch plots and

quantitative nonlinear parameters, demonstrated a strong dependence of the nonlinear

response of the samples to NaCl concentration. Increasing in NaCl loading led to

observations of higher intra-cycle nonlinearity.

3. The effect of shear at disrupting CNC-PVA-NaCl gel was found to be reversible over a

long period of time for all CNC loadings. For composite with higher loadings of CNC, full

structure recovery occurred at around ~1500 seconds.

4. The SEM image evaluation showed that the porosity was an increasing function of NaCl

concentration.

169

5. After loading 15 g/L CNC, wide distribution in the pore size and almost equal average pore

sizes were observed in PVA-CNC samples. Moreover, after the addition of 15 g/L CNC

into 5 wt% PVA, the average pore sizes were no longer dependent on CNC loading.

6. The recovery of the storage modulus of the gel was found to be the highest when CNC

concentration was 30 g/L. Recovery of storage modulus for samples with a quantity of

CNC lower than 30 g/L revealed slower recovery.

7. As the NaCl concentration increased from 1.72 mM to 172 mM, the storage modulus (G')

and the loss modulus (G'') of the CNC (20 and 30 gr/l)-NaCl system followed three

regimes: an initial increase, followed by a descending trend, which ended up in a plateau

region. Reaching a plateau was an indication of the formation of a strong gel. Samples with

low NaCl concentrations (< 8.5mM) showed a low G' as they were still in the form of a

solution/suspension and had not yet formed a solid hydrogel network. However, for higher

values of NaCl concentration, the system depicted G'/G'' > 1, which was much less

frequency dependent signaling the formation of a self-supporting elastic gel (i.e., strong

gel).

8. The CNC/NaCl suspension systems had a shear-thinning characteristic. The values of

complex viscosity increased with an increase in NaCl concentration.

9. The increasing influence of packing or attractive interactions as a result of the increase in

ionic strength of the medium could limit the retardation of individual rods to the scales on

the order of rods diameter. At this point, arrested dynamics were reached, which also

translated into more elasticity and non-ergodicity.

10. The rheological characterization of PVA/CNC revealed the existence of two types of

networks, a polymer mediated CNC network at low CNC loadings and a CNC network at

high CNC loadings. The transition between these two networks was related to the CNC

percolation threshold.

11. The mechanical testing on PVA/CNC aerogels showed that all hydrogels depicted curves

in the shape of “J”, the index of materials with high compressive strength (334). The

compressive strength of the hydrogels increased with an increase in the amount of CNCs.

The CNC-PVA sample with 35 g/L CNC showed the highest compressive stress at 3.4

MPa, which was due to the stiff chains in the strong pore wall.

C. CNC gel healing assessment

170

1. The diffusion of CNCs at different concentrations (i.e., dilute and semi-dilute regimes) was

evaluated using fluorescence recovery after photobleaching (FRAP) images. The rate of

diffusion of particles due to crowding with other particles slowed down as the

concentration of CNC increased. Furthermore, the CNC diffusion inside the gel was a

decreasing function of CNC and NaCl concentrations. It was shown that the gel dynamicity

was shown to change across the gel. The system was found to be dynamic at all CNC and

NaCl loadings and had the ability to heal, regardless of probed concentrations.

2. We illustrated the connection among immobile particle percentage, zeta potential, and

storage modulus through CNC/NaCl concentration ratio.

7-1 Future works

1) One can use confocal laser scanning microscopy images to visualize the healing process

after 3-D printing the CNC gel.

2) One can use borax for cross-linking of CNC-PVA hydrogel, in order to improve its

mechanical properties and tune it based on the degree of cross-linking. Borax is a chemical

that, depending on the amount used, can cross-link PVA chains in a facile manner.

3) Anisotropic hydrogels can be used for culturing nerve cells. Axonal cells have a very

peculiar linear shape that requires a scaffold that is both robust mechanically and is also

able to accommodate these cells.

4) Another technique currently available to capture the dynamics of CNCs in suspension and

gel is raster imaging. Employing this technique might lead to a better reading of CNC

mobility in the hydrogel.

5) A setback of the methodology used in this thesis is non-uniform gelation due to the method

of mixing of NaCl with CNC suspension. One can use slow dissolving salts to overcome

this issue. It will be beneficial to see how adding a slow dissolving salt, which provides

ample time for uniform gelation, will influence the mechanical properties of the gel.

6) The Folgar-Tucker formulation can be employed to study gel recovery by rheology. Folgar-

Tucker orientation equation relates shear at the macro-level to the micro orientation of

nanoparticles. As the storage modulus recovery is related to how storage modulus builds

up after shear, it might be useful to observe how to gel recovery characteristic time is

related to the time scale required for particles to reorient themselves.

171

7) One can take advantage of molecular dynamic simulation to probe the accuracy of the

Schulze-Hardy rule on differentiating the impact of MgCl2 and NaCl on CNC gelation in

this empirical equation.

8) Free volume in the gel can affect the diffusion coefficient of CNCs in the gel. FRAP can

be used here to study these effects on CNC dynamics in the gel. Swelling in the hydrogel

can influence free volume inside the gel and loosen up CNC gel. Therefore, we expect to

see different diffusion values inside the gel.

9) FRAP recovery curves can be used to study the effect of shear in gels. Different shear rate

values will orient CNC differently, and therefore diffusion for such cases might be

different.

10) It would be interesting to investigate the changes of CNC zeta potentials before and after

rinsing. In earlier chapters, using zeta potential values, we showed how adding salt causes

the zeta potential values to go toward lower absolute values. It is interesting to see how

CNC zeta potential values will change if the system gets filtered and rinsed with water a

couple of times.

11) It would be beneficial to find salt and dyed CNC interaction, in terms of zeta potential. In

Chapter 2 we showed that up to 500 ppm of dye does not change zeta potential values of

CNCs. The question remains on how the presence of dye on CNCs will affect zeta potential

in the presence of NaCl or MgCl2.

12) Another fluorophore can be used in combination with FB28 to study the migration of

another set of particles in the CNC gel. Using a different fluorophore, we might be able to

find to what extent the second group of CNCs can migrate into the original gel.

13) In 3-D printing applications, healing of breakage points and T-sections are interesting to

study; these points give mechanical integrity to the gel. Diffusion across the boundary at

which the gel breakage happens will give vital information about the healing rate of CNC

or CNC-polymer hydrogel, a phenomenon that can be extensively studied with FRAP.

14) Storage modulus recovery versus time for CNC-PVA gel can be studied as a function of

temperature. In earlier chapters, we studied storage modulus recovery as a function of time

for PVA-CNC hydrogels. We expect to see that an increase in temperature expedites the

healing rate of the hydrogel.

172

15) The porosity of hydrogel can be studied using CLSM and rheometry for the cases of

unsheared samples and sheared samples. It is expected that shear will likely distort CNC

gel pores. The rate of recovery of the pores as a function of time can be an interesting

subject of investigation.

16) Hydrolysis condition and post treatment will alter the level of sulfation and consequently

surface charge of CNCs. Therefore, more prolong surface treatment of CNC will alter

initial magnitude of zeta potential of CNCs. It would be interesting to investigate how

changing surface charges and surface charge distributions can alter zeta potential of CNCs.

17) In the present study, the ionic strength effects have not been considered with respect to

CNC surface charge. Note that, the conclusions made throughout this thesis on ionic

strength effects are without considering the surface charge density/distribution of the

starting CNCs. Moreover, surface chemistry of CNC has a strong affect on both the

rheology and interactions between nanoparticles during gelation or interactions between

CNCs and the polymer in the hydrogel systems. Future studies are needed to unravel the

response of hydrogels investigated in the present studies with respect to the surface

chemistry of CNC. It is interesting to observe how changing surface charge and

distribution of charges would change final properties of hydrogels made with CNCs.

Moreover, pragmatically, effect of charge distribution of CNC on fractal dimension can

also be investigated. The way half ester sulfate groups are decorating CNC can influence

the way CNCs are influenced by electrostatic repulsions. Therefore, it is of interest to

assess the effect of hydrolysis condition on CNCs and subsequently study fractal dimension

development because of salt addition via TEM.

18) It would be interesting to see how charge distribution affects zeta potential. The

investigation on correlation of charge distribution on rod particles and its influence on

particle movement and orientation can happen for CNCs. The model for theoretically

relating zeta potential for uneven charge distribution on CNCs can be developed.

19) Fibers with diamagnetic anisotropy align under static magnetic fields. Cellulose fibers can

undergo magnetic alignment. It has been reported that static magnetic field can align chiral

nematic axis of CNCs in the field direction. It can be interesting to observe how this

property of CNCs in combination to gelation can be used to manufacture ultra-strong CNC

173

made fibers. It can be envisioned that all CNCs would orient along the same direction and

this would help with enhancement of the reinforcement of mechanical properties.

20) In this thesis, due to uncertainty into whether FB28 preferentially attach to CNC or PVA

alone, the CLSM imaging was not performed for CNC-PVA hydrogel. However, it is

recommended due to similar chemistry of CNC and PVA, for the purpose of CLSM

imaging, each of the constituent tagged using covalent bonding with appropriate

fluorescent dyes.

21) As CNC-polymer hydrogel is planned to be used as a scaffold for tissue engineering, water

processing methodology is necessary. Therefore, addition of polymer will be narrowed

down to water soluble polymers such as PVA and polyethylene oxide (PEO). The criterion

on using PVA in the present study is its proven biocompatibility. Moreover, PVA is a cheap

abundant polymer that can make hydrogels made with CNC attractive. It would be

interesting to observe how addition of PEO will alter hydrogel properties in comparison to

PVA.

22) Upon breaking the gel structure with shear cluster and cluster links of CNCs gets broken

and initial CNCs do not get impacted. Studying cluster morphology after breakage of

structure with techniques such as BET, SEM, TEM and AFM would be interesting.

23) Supramolecuar interactions between CNCs and PVA in hydrogel can be explored. Also it

would be interesting to see how concentration and temperature of CNC hydrogel will

govern the competition for establishing hydrogen bonds between PVA-PVA, PVA-CNC,

CNC-CNC, PVA-water and CNC-water molecules. Taking into consideration

supramolecular interactions and competiton between constituent of the gel on final

mechanical properties and healing of the hydrogel can be an interesting future work.

174

Copyrights

SPRINGER NATURE LICENSE

TERMS AND CONDITIONS

Feb 08, 2020

This Agreement between University of Calgary -- Aref Abbasi Moud ("You") and Springer

Nature ("Springer Nature") consists of your license details and the terms and conditions

provided by Springer Nature and Copyright Clearance Center.

License Number 4764441178329

License date Feb 08, 2020

Licensed Content Publisher Springer Nature

Licensed Content Publication Cellulose

Licensed Content Title Cellulose nanocrystal structure in the

presence of salts

Licensed Content Author Aref Abbasi Moud et al

Licensed Content Date Sep 11, 2019

Type of Use Thesis/Dissertation

Requestor type non-commercial (non-profit)

Format print and electronic

Portion full article/chapter

Will you be translating? no

Circulation/distribution 1 - 29

Author of this Springer Nature content yes

Title Dr

Institution name University of Calgary

Expected presentation date May 2020

Requestor Location


2204 Cherokee drive calgary

12 street SW calgary

calgary, AB T2L 0X5

Canada

Attn: University of Calgary

Total 0.00 USD

Terms and Conditions

175

Springer Nature Customer Service Centre GmbH


This agreement sets out the terms and conditions of the licence (the Licence) between you

and Springer Nature Customer Service Centre GmbH (the Licensor). By clicking 'accept'

and completing the transaction for the material (Licensed Material), you also confirm your

acceptance of these terms and conditions.

1. Grant of License

1. The Licensor grants you a personal, non-exclusive, non-transferable, world-wide

licence to reproduce the Licensed Material for the purpose specified in your

order only. Licences are granted for the specific use requested in the order and

for no other use, subject to the conditions below.

2. The Licensor warrants that it has, to the best of its knowledge, the rights to

license reuse of the Licensed Material. However, you should ensure that the

material you are requesting is original to the Licensor and does not carry the

copyright of another entity (as credited in the published version).

3. If the credit line on any part of the material you have requested indicates that it

was reprinted or adapted with permission from another source, then you should

also seek permission from that source to reuse the material.

2. Scope of Licence

1. You may only use the Licensed Content in the manner and to the extent

permitted by these Ts&Cs and any applicable laws.

2. A separate licence may be required for any additional use of the Licensed

Material, e.g. where a licence has been purchased for print only use, separate

permission must be obtained for electronic re-use. Similarly, a licence is only

valid in the language selected and does not apply for editions in other languages

unless additional translation rights have been granted separately in the licence.

Any content owned by third parties are expressly excluded from the licence.

3. Similarly, rights for additional components such as custom editions and

derivatives require additional permission and may be subject to an additional

fee. Please apply

to [email protected]/[email protected]

m for these rights.

4. Where permission has been granted free of charge for material in print,

permission may also be granted for any electronic version of that work, provided

mailto:[email protected]



176

that the material is incidental to your work as a whole and that the electronic

version is essentially equivalent to, or substitutes for, the print version.

5. An alternative scope of licence may apply to signatories of the STM Permissions

Guidelines, as amended from time to time.

Duration of Licence

1. A licence for is valid from the date of purchase ('Licence Date') at the end of the

relevant period in the below table:

Scope of Licence Duration of Licence

Post on a website 12 months

Presentations 12 months

Books and journals Lifetime of the edition in the language purchased

Acknowledgement

1. The Licensor's permission must be acknowledged next to the Licenced Material in print.

In electronic form, this acknowledgement must be visible at the same time as the

figures/tables/illustrations or abstract, and must be hyperlinked to the journal/book's

homepage. Our required acknowledgement format is in the Appendix below.

Restrictions on use

1. Use of the Licensed Material may be permitted for incidental promotional use and

minor editing privileges e.g. minor adaptations of single figures, changes of format,

colour and/or style where the adaptation is credited as set out in Appendix 1 below. Any

other changes including but not limited to, cropping, adapting, omitting material that

affect the meaning, intention or moral rights of the author are strictly prohibited.

2. You must not use any Licensed Material as part of any design or trademark.

3. Licensed Material may be used in Open Access Publications (OAP) before publication

by Springer Nature, but any Licensed Material must be removed from OAP sites prior to

final publication.

Ownership of Rights

1. Licensed Material remains the property of either Licensor or the relevant third party and

any rights not explicitly granted herein are expressly reserved.

http://www.stm-assoc.org/intellectual-property/permissions/permissions-guidelines/


177

Warranty

IN NO EVENT SHALL LICENSOR BE LIABLE TO YOU OR ANY OTHER PARTY OR

ANY OTHER PERSON OR FOR ANY SPECIAL, CONSEQUENTIAL, INCIDENTAL OR

INDIRECT DAMAGES, HOWEVER CAUSED, ARISING OUT OF OR IN CONNECTION

WITH THE DOWNLOADING, VIEWING OR USE OF THE MATERIALS REGARDLESS

OF THE FORM OF ACTION, WHETHER FOR BREACH OF CONTRACT, BREACH OF

WARRANTY, TORT, NEGLIGENCE, INFRINGEMENT OR OTHERWISE (INCLUDING,

WITHOUT LIMITATION, DAMAGES BASED ON LOSS OF PROFITS, DATA, FILES,

USE, BUSINESS OPPORTUNITY OR CLAIMS OF THIRD PARTIES), AND

WHETHER OR NOT THE PARTY HAS BEEN ADVISED OF THE POSSIBILITY OF

SUCH DAMAGES. THIS LIMITATION SHALL APPLY NOTWITHSTANDING ANY

FAILURE OF ESSENTIAL PURPOSE OF ANY LIMITED REMEDY PROVIDED HEREIN.

Limitations

1. BOOKS ONLY:Where 'reuse in a dissertation/thesis' has been selected the following

terms apply: Print rights of the final author's accepted manuscript (for clarity, NOT the

published version) for up to 100 copies, electronic rights for use only on a personal

website or institutional repository as defined by the Sherpa guideline

(www.sherpa.ac.uk/romeo/).

Termination and Cancellation

1. Licences will expire after the period shown in Clause 3 (above).

2. Licensee reserves the right to terminate the Licence in the event that payment is not

received in full or if there has been a breach of this agreement by you.

Appendix 1 — Acknowledgements:

For Journal Content:

Reprinted by permission from [the Licensor]: [Journal Publisher (e.g.

Nature/Springer/Palgrave)] [JOURNAL NAME] [REFERENCE CITATION (Article

name, Author(s) Name), [COPYRIGHT] (year of publication)

For Advance Online Publication papers:



name, Author(s) Name), [COPYRIGHT] (year of publication), advance online

publication, day month year (doi: 10.1038/sj.[JOURNAL ACRONYM].)

For Adaptations/Translations:

Adapted/Translated by permission from [the Licensor]: [Journal Publisher (e.g.

http://www.sherpa.ac.uk/romeo/

178



Note: For any republication from the British Journal of Cancer, the following

credit line style applies:

Reprinted/adapted/translated by permission from [the Licensor]: on behalf of Cancer

Research UK: : [Journal Publisher (e.g. Nature/Springer/Palgrave)] [JOURNAL

NAME] [REFERENCE CITATION (Article name, Author(s) Name),

[COPYRIGHT] (year of publication)


Reprinted by permission from The [the Licensor]: on behalf of Cancer Research UK:

[Journal Publisher (e.g. Nature/Springer/Palgrave)] [JOURNAL NAME]

[REFERENCE CITATION (Article name, Author(s) Name), [COPYRIGHT] (year

of publication), advance online publication, day month year (doi:

10.1038/sj.[JOURNAL ACRONYM])

For Book content:

Reprinted/adapted by permission from [the Licensor]: [Book Publisher (e.g. Palgrave

Macmillan, Springer etc) [Book Title] by [Book author(s)] [COPYRIGHT] (year of

publication)

Other Conditions:

Version 1.2

Questions? [email protected] or +1-855-239-3415 (toll free in the US) or +1-

978-646-2777.

SPRINGER NATURE LICENSE

TERMS AND CONDITIONS

Jun 15, 2020


179

This Agreement between University of Calgary -- Aref Abbasi Moud ("You") and Springer

Nature ("Springer Nature") consists of your license details and the terms and conditions

provided by Springer Nature and Copyright Clearance Center.

License Number 4850170917768

License date Jun 15, 2020

Licensed Content Publisher Springer Nature

Licensed Content Publication Cellulose

Licensed Content Title

Nonlinear viscoelastic characterization of charged

cellulose nanocrystal network structure in the

presence of salt in aqueous media

Licensed Content Author Aref Abbasi Moud et al

Licensed Content Date Apr 24, 2020

Type of Use Thesis/Dissertation

Requestor type non-commercial (non-profit)

Format print and electronic

Portion full article/chapter

Will you be translating? no

Circulation/distribution 1 - 29

Author of this Springer Nature content yes

Title Dr Aref Abbasi Moud

Institution name University of Calgary

Expected presentation date Jun 2020

Order reference number 1

Requestor Location


2204 Cherokee drive calgary

12 street SW calgary

calgary, AB T2L 0X5

Canada

Attn: University of Calgary

Total 0.00 USD


Springer Nature Customer Service Centre GmbH


This agreement sets out the terms and conditions of the licence (the Licence) between you

and Springer Nature Customer Service Centre GmbH (the Licensor). By clicking 'accept'

180

and completing the transaction for the material (Licensed Material), you also confirm your

acceptance of these terms and conditions.

1. Grant of License

1. The Licensor grants you a personal, non-exclusive, non-transferable, world-wide

licence to reproduce the Licensed Material for the purpose specified in your

order only. Licences are granted for the specific use requested in the order and

for no other use, subject to the conditions below.

2. The Licensor warrants that it has, to the best of its knowledge, the rights to

license reuse of the Licensed Material. However, you should ensure that the

material you are requesting is original to the Licensor and does not carry the

copyright of another entity (as credited in the published version).

3. If the credit line on any part of the material you have requested indicates that it

was reprinted or adapted with permission from another source, then you should

also seek permission from that source to reuse the material.

2. Scope of Licence

1. You may only use the Licensed Content in the manner and to the extent

permitted by these Ts&Cs and any applicable laws.

2. A separate licence may be required for any additional use of the Licensed

Material, e.g. where a licence has been purchased for print only use, separate

permission must be obtained for electronic re-use. Similarly, a licence is only

valid in the language selected and does not apply for editions in other languages

unless additional translation rights have been granted separately in the licence.

Any content owned by third parties are expressly excluded from the licence.

3. Similarly, rights for additional components such as custom editions and

derivatives require additional permission and may be subject to an additional fee.

Please apply

to [email protected]/[email protected]

m for these rights.

4. Where permission has been granted free of charge for material in print,

permission may also be granted for any electronic version of that work, provided

that the material is incidental to your work as a whole and that the electronic

version is essentially equivalent to, or substitutes for, the print version.




181

5. An alternative scope of licence may apply to signatories of the STM Permissions

Guidelines, as amended from time to time.

Duration of Licence

1. A licence for is valid from the date of purchase ('Licence Date') at the end of the

relevant period in the below table:

Scope of Licence Duration of Licence

Post on a website 12 months

Presentations 12 months

Books and journals Lifetime of the edition in the language purchased

Acknowledgement

1. The Licensor's permission must be acknowledged next to the Licenced Material in print.

In electronic form, this acknowledgement must be visible at the same time as the

figures/tables/illustrations or abstract, and must be hyperlinked to the journal/book's

homepage. Our required acknowledgement format is in the Appendix below.

Restrictions on use

1. Use of the Licensed Material may be permitted for incidental promotional use and minor

editing privileges e.g. minor adaptations of single figures, changes of format, colour

and/or style where the adaptation is credited as set out in Appendix 1 below. Any other

changes including but not limited to, cropping, adapting, omitting material that affect the

meaning, intention or moral rights of the author are strictly prohibited.

2. You must not use any Licensed Material as part of any design or trademark.

3. Licensed Material may be used in Open Access Publications (OAP) before publication

by Springer Nature, but any Licensed Material must be removed from OAP sites prior to

final publication.

Ownership of Rights



182

1. Licensed Material remains the property of either Licensor or the relevant third party and

any rights not explicitly granted herein are expressly reserved.

Warranty

IN NO EVENT SHALL LICENSOR BE LIABLE TO YOU OR ANY OTHER PARTY OR

ANY OTHER PERSON OR FOR ANY SPECIAL, CONSEQUENTIAL, INCIDENTAL OR

INDIRECT DAMAGES, HOWEVER CAUSED, ARISING OUT OF OR IN CONNECTION

WITH THE DOWNLOADING, VIEWING OR USE OF THE MATERIALS REGARDLESS

OF THE FORM OF ACTION, WHETHER FOR BREACH OF CONTRACT, BREACH OF

WARRANTY, TORT, NEGLIGENCE, INFRINGEMENT OR OTHERWISE (INCLUDING,

WITHOUT LIMITATION, DAMAGES BASED ON LOSS OF PROFITS, DATA, FILES,

USE, BUSINESS OPPORTUNITY OR CLAIMS OF THIRD PARTIES), AND

WHETHER OR NOT THE PARTY HAS BEEN ADVISED OF THE POSSIBILITY OF

SUCH DAMAGES. THIS LIMITATION SHALL APPLY NOTWITHSTANDING ANY

FAILURE OF ESSENTIAL PURPOSE OF ANY LIMITED REMEDY PROVIDED HEREIN.

Limitations

1. BOOKS ONLY:Where 'reuse in a dissertation/thesis' has been selected the following

terms apply: Print rights of the final author's accepted manuscript (for clarity, NOT the

published version) for up to 100 copies, electronic rights for use only on a personal

website or institutional repository as defined by the Sherpa guideline

(www.sherpa.ac.uk/romeo/).

Termination and Cancellation

1. Licences will expire after the period shown in Clause 3 (above).

2. Licensee reserves the right to terminate the Licence in the event that payment is not

received in full or if there has been a breach of this agreement by you.

Appendix 1 — Acknowledgements:

For Journal Content:


http://www.sherpa.ac.uk/romeo/

183






name, Author(s) Name), [COPYRIGHT] (year of publication), advance online

publication, day month year (doi: 10.1038/sj.[JOURNAL ACRONYM].)

For Adaptations/Translations:

Adapted/Translated by permission from [the Licensor]: [Journal Publisher (e.g.



Note: For any republication from the British Journal of Cancer, the following

credit line style applies:

Reprinted/adapted/translated by permission from [the Licensor]: on behalf of Cancer

Research UK: : [Journal Publisher (e.g. Nature/Springer/Palgrave)] [JOURNAL

NAME] [REFERENCE CITATION (Article name, Author(s) Name),

[COPYRIGHT] (year of publication)


Reprinted by permission from The [the Licensor]: on behalf of Cancer Research UK:

[Journal Publisher (e.g. Nature/Springer/Palgrave)] [JOURNAL NAME]

[REFERENCE CITATION (Article name, Author(s) Name), [COPYRIGHT] (year

of publication), advance online publication, day month year (doi:

10.1038/sj.[JOURNAL ACRONYM])

For Book content:

Reprinted/adapted by permission from [the Licensor]: [Book Publisher (e.g. Palgrave

Macmillan, Springer etc) [Book Title] by [Book author(s)] [COPYRIGHT] (year of

publication)

Other Conditions:

Version 1.2

Questions? [email protected] or +1-855-239-3415 (toll free in the US) or +1-

978-646-2777.


184

References

[1] A. Dufresne, Nanocellulose: from nature to high performance tailored materials, Walter de Gruyter

GmbH & Co KG2017.

[2] R.J. Moon, A. Martini, J. Nairn, J. Simonsen, J. Youngblood, Cellulose nanomaterials review:

structure, properties and nanocomposites, Chemical Society Reviews 40(7) (2011) 3941-3994.

[3] Y. Habibi, L.A. Lucia, O.J. Rojas, Cellulose nanocrystals: chemistry, self-assembly, and applications,

Chemical reviews 110(6) (2010) 3479-3500.

[4] S. Kalia, B. Kaith, I. Kaur, Cellulose fibers: bio-and nano-polymer composites: green chemistry and

technology, Springer Science & Business Media2011.

[5] J.-M. Lehn, Supramolecular chemistry, Science 260(5115) (1993) 1762-1764.

[6] J.J. Elsner, M. Zilberman, Novel antibiotic-eluting wound dressings: An in vitro study and

engineering aspects in the dressing's design, Journal of tissue viability 19(2) (2010) 54-66.

[7] A. Song, A.A. Rane, K.L. Christman, Antibacterial and cell-adhesive polypeptide and poly (ethylene

glycol) hydrogel as a potential scaffold for wound healing, Acta biomaterialia 8(1) (2012) 41-50.

[8] G. Siqueira, J. Bras, A.J.B. Dufresne, Cellulose whiskers versus microfibrils: influence of the nature

of the nanoparticle and its surface functionalization on the thermal and mechanical properties of

nanocomposites, 10(2) (2008) 425-432.

[9] C. Zhou, Q. Wu, T. Lei, I.I. Negulescu, Adsorption kinetic and equilibrium studies for methylene blue

dye by partially hydrolyzed polyacrylamide/cellulose nanocrystal nanocomposite hydrogels, Chemical

Engineering Journal 251 (2014) 17-24.

[10] I. David, McGraw-Hill Yearbook of Science and Technology, Soil Science 98(4) (1964) 278.

[11] B.L. Peng, N. Dhar, H. Liu, K. Tam, Chemistry and applications of nanocrystalline cellulose and its

derivatives: a nanotechnology perspective, The Canadian Journal of Chemical Engineering 89(5) (2011)

1191-1206.

[12] S.J. Eichhorn, A. Dufresne, M. Aranguren, N. Marcovich, J. Capadona, S. Rowan, C. Weder, W.

Thielemans, M. Roman, S. Renneckar, Current international research into cellulose nanofibres and

nanocomposites, Journal of materials science 45(1) (2010) 1-33.

[13] G. Chinga-Carrasco, Cellulose fibres, nanofibrils and microfibrils: the morphological sequence of

MFC components from a plant physiology and fibre technology point of view, Nanoscale research letters

6(1) (2011) 417.

[14] J. Simonsen, Bio-based nanocomposties: challenges and opportunities, [Online]. Available:

http://people.forestry.oregonstate.edu/john-simonsen/sites/develd7.forestry.oregonstate.edu.john-

simonsen/files/Nanocomposites.pdf. .

[15] M. Matus, Super-Durable Material Made from Wood Waste is Stronger, Cheaper,and Lighter Than

Kevlar, [Online]. Available: http://inhabitat.com/superdurable-material-made-from-wood-waste-is-

stronger-cheaper-and-lighter-thankevlar/. ([Accessed: 05-June-2014]).

[16] M.C.B. de Figueirêdo, M. de Freitas Rosa, C.M.L. Ugaya, M.d.S.M. de Souza, A.C.C. da Silva

Braid, L.F.L. de Melo, Life cycle assessment of cellulose nanowhiskers, Journal of Cleaner Production 35

(2012) 130-139.

[17] M.R. Mauricio, P.G. da Costa, S.K. Haraguchi, M.R. Guilherme, E.C. Muniz, A.F. Rubira, Synthesis

of a microhydrogel composite from cellulose nanowhiskers and starch for drug delivery, Carbohydrate

polymers 115 (2015) 715-722.

[18] S. Montanari, M. Roumani, L. Heux, M.R. Vignon, Topochemistry of carboxylated cellulose

nanocrystals resulting from TEMPO-mediated oxidation, Macromolecules 38(5) (2005) 1665-1671.

[19] N. Drogat, R. Granet, C. Le Morvan, G. Bégaud-Grimaud, P. Krausz, V. Sol, Chlorin-PEI-labeled

cellulose nanocrystals: synthesis, characterization and potential application in PDT, Bioorganic &

medicinal chemistry letters 22(11) (2012) 3648-3652.

http://people.forestry.oregonstate.edu/john-simonsen/sites/develd7.forestry.oregonstate.edu.john-simonsen/files/Nanocomposites.pdf

http://people.forestry.oregonstate.edu/john-simonsen/sites/develd7.forestry.oregonstate.edu.john-simonsen/files/Nanocomposites.pdf

http://inhabitat.com/superdurable-material-made-from-wood-waste-is-stronger-cheaper-and-lighter-thankevlar/

http://inhabitat.com/superdurable-material-made-from-wood-waste-is-stronger-cheaper-and-lighter-thankevlar/

185

[20] J.K. Jackson, K. Letchford, B.Z. Wasserman, L. Ye, W.Y. Hamad, H.M. Burt, The use of

nanocrystalline cellulose for the binding and controlled release of drugs, International journal of

nanomedicine 6 (2011) 321.

[21] B. O′Connor, NCC: Environmental Health and Safety Update. 2012 TAPPI, Nanotechnology

Conference for Renewable Materials, Montréal, QC, 2012.

[22] O. Kristiina, B. Alexander, Handbook Of Green Materials: Processing Technologies, Properties And

Applications (In 4 Volumes), World Scientific2014.

[23] M. Börjesson, G. Westman, Crystalline nanocellulose—preparation, modification, and properties,

Cellulose-fundamental aspects and current trends, InTech2015.

[24] S. Yamanaka, K. Watanabe, N. Kitamura, M. Iguchi, S. Mitsuhashi, Y. Nishi, M. Uryu, The

structure and mechanical properties of sheets prepared from bacterial cellulose, Journal of Materials

Science 24(9) (1989) 3141-3145.

[25] C. Endes, S. Camarero-Espinosa, S. Mueller, E. Foster, A. Petri-Fink, B. Rothen-Rutishauser, C.

Weder, M. Clift, A critical review of the current knowledge regarding the biological impact of

nanocellulose, Journal of nanobiotechnology 14(1) (2016) 78.

[26] J. George, S.J.N. Sabapathi, science, applications, Cellulose nanocrystals: synthesis, functional

properties, and applications, 8 (2015) 45.

[27] M. Pereda, A. Dufresne, M.I. Aranguren, N.E.J.C.p. Marcovich, Polyelectrolyte films based on

chitosan/olive oil and reinforced with cellulose nanocrystals, 101 (2014) 1018-1026.

[28] H.-M. Ng, L.T. Sin, T.-T. Tee, S.-T. Bee, D. Hui, C.-Y. Low, A.J.C.P.B.E. Rahmat, Extraction of

cellulose nanocrystals from plant sources for application as reinforcing agent in polymers, 75 (2015) 176-

200.

[29] A. Iwatake, M. Nogi, H.J.C.S. Yano, Technology, Cellulose nanofiber-reinforced polylactic acid,

68(9) (2008) 2103-2106.

[30] N. Savadekar, S.J.C.P. Mhaske, Synthesis of nano cellulose fibers and effect on thermoplastics starch

based films, 89(1) (2012) 146-151.

[31] N. Lin, A.J.E.P.J. Dufresne, Nanocellulose in biomedicine: Current status and future prospect, 59

(2014) 302-325.

[32] G. Lenfant, M.-C. Heuzey, T.G.M. van de Ven, P.J. Carreau, A comparative study of ECNC and

CNC suspensions: effect of salt on rheological properties, Rheologica Acta 56(1) (2017) 51-62.

[33] F. Cherhal, F. Cousin, I. Capron, Influence of charge density and ionic strength on the aggregation

process of cellulose nanocrystals in aqueous suspension, as revealed by small-angle neutron scattering,

Langmuir 31(20) (2015) 5596-5602.

[34] T. Witten Jr, L.M. Sander, Diffusion-limited aggregation, a kinetic critical phenomenon, Physical

review letters 47(19) (1981) 1400.

[35] P. Meakin, Diffusion-controlled deposition on fibers and surfaces, Physical Review A 27(5) (1983)

2616.

[36] P. Bolhuis, D. Frenkel, Tracing the phase boundaries of hard spherocylinders, The Journal of

chemical physics 106(2) (1997) 666-687.

[37] D. Frenkel, H. Lekkerkerker, A. Stroobants, Thermodynamic stability of a smectic phase in a system

of hard rods, Nature 332(6167) (1988) 822.

[38] D. Frenkel, B. Mulder, The hard ellipsoid-of-revolution fluid: I. Monte Carlo simulations, Molecular

physics 55(5) (1985) 1171-1192.

[39] E. Grelet, Hexagonal order in crystalline and columnar phases of hard rods, Physical review letters

100(16) (2008) 168301.

[40] M. Doi, S.F. Edwards, The theory of polymer dynamics, oxford university press1988.

[41] M.J. Green, A.N.G. Parra-Vasquez, N. Behabtu, M. Pasquali, Modeling the phase behavior of

polydisperse rigid rods with attractive interactions with applications to single-walled carbon nanotubes in

superacids, The Journal of chemical physics 131(8) (2009) 084901.

[42] L. Lewis, S.G. Hatzikiriakos, W.Y. Hamad, M.J. MacLachlan, Freeze–Thaw Gelation of Cellulose

Nanocrystals, ACS Macro Letters 8(5) (2019) 486-491.

186

[43] L. Lewis, M. Derakhshandeh, S.G. Hatzikiriakos, W.Y. Hamad, M.J. MacLachlan, Hydrothermal

gelation of aqueous cellulose nanocrystal suspensions, Biomacromolecules 17(8) (2016) 2747-2754.

[44] F.B. Birkner, J.J. Morgan, Polymer flocculation kinetics of dilute colloidal suspensions, Journal‐

American Water Works Association 60(2) (1968) 175-191.

[45] C. Wilson, Gelation in colloid–polymer mixtures, Faraday Discussions 101 (1995) 65-76.

[46] E.E. Ureña-Benavides, G. Ao, V.A. Davis, C.L. Kitchens, Rheology and phase behavior of lyotropic

cellulose nanocrystal suspensions, Macromolecules 44(22) (2011) 8990-8998.

[47] A.E. Way, L. Hsu, K. Shanmuganathan, C. Weder, S.J. Rowan, pH-responsive cellulose nanocrystal

gels and nanocomposites, ACS Macro Letters 1(8) (2012) 1001-1006.

[48] M. Chau, S.E. Sriskandha, D. Pichugin, H.l. Therien-Aubin, D. Nykypanchuk, G.g. Chauve, M.

Methot, J. Bouchard, O. Gang, E. Kumacheva, Ion-mediated gelation of aqueous suspensions of cellulose

nanocrystals, Biomacromolecules 16(8) (2015) 2455-2462.

[49] Z. Hu, E.D. Cranston, R. Ng, R. Pelton, Tuning cellulose nanocrystal gelation with polysaccharides

and surfactants, Langmuir 30(10) (2014) 2684-2692.

[50] S. Beck-Candanedo, M. Roman, D.G. Gray, Effect of reaction conditions on the properties and

behavior of wood cellulose nanocrystal suspensions, Biomacromolecules 6(2) (2005) 1048-1054.

[51] M. Chau, K.J. De France, B. Kopera, V.R. Machado, S. Rosenfeldt, L. Reyes, K.J. Chan, S. Forster,

E.D. Cranston, T. Hoare, Composite hydrogels with tunable anisotropic morphologies and mechanical

properties, Chemistry of Materials 28(10) (2016) 3406-3415.

[52] A. Dorris, D.G. Gray, Gelation of cellulose nanocrystal suspensions in glycerol, Cellulose 19(3)

(2012) 687-694.

[53] D. Chen, D. Lawton, M. Thompson, Q. Liu, Biocomposites reinforced with cellulose nanocrystals

derived from potato peel waste, Carbohydrate polymers 90(1) (2012) 709-716.

[54] T. Abitbol, T. Johnstone, T.M. Quinn, D.G. Gray, Reinforcement with cellulose nanocrystals of poly

(vinyl alcohol) hydrogels prepared by cyclic freezing and thawing, Soft Matter 7(6) (2011) 2373-2379.

[55] S. Tanpichai, K. Oksman, Cross-linked nanocomposite hydrogels based on cellulose nanocrystals

and PVA: Mechanical properties and creep recovery, Composites Part A: Applied Science and

Manufacturing 88 (2016) 226-233.

[56] W. Li, D. Wang, W. Yang, Y. Song, Compressive mechanical properties and microstructure of

PVA–HA hydrogels for cartilage repair, RSC advances 6(24) (2016) 20166-20172.

[57] K. Song, W. Zhu, X. Li, Z. Yu, A novel mechanical robust, self-healing and shape memory hydrogel

based on PVA reinforced by cellulose nanocrystal, Materials Letters 260 (2020) 126884.

[58] S. Butylina, S. Geng, K. Oksman, Properties of as-prepared and freeze-dried hydrogels made from

poly (vinyl alcohol) and cellulose nanocrystals using freeze-thaw technique, European Polymer Journal

81 (2016) 386-396.

[59] J. You, J. Cao, Y. Zhao, L. Zhang, J. Zhou, Y. Chen, Improved mechanical properties and sustained

release behavior of cationic cellulose nanocrystals reinforeced cationic cellulose injectable hydrogels,

Biomacromolecules 17(9) (2016) 2839-2848.

[60] J.S. Gonzalez, L.N. Ludueña, A. Ponce, V.A. Alvarez, Poly (vinyl alcohol)/cellulose nanowhiskers

nanocomposite hydrogels for potential wound dressings, Materials Science and Engineering: C 34 (2014)

54-61.

[61] N. Lin, A. Geze, D. Wouessidjewe, J. Huang, A. Dufresne, Biocompatible double-membrane

hydrogels from cationic cellulose nanocrystals and anionic alginate as complexing drugs codelivery, ACS

applied materials & interfaces 8(11) (2016) 6880-6889.

[62] A. Mihranyan, Viscoelastic properties of cross-linked polyvinyl alcohol and surface-oxidized

cellulose whisker hydrogels, Cellulose 20(3) (2013) 1369-1376.

[63] T. Ben Shalom, Y. Nevo, D. Leibler, Z. Shtein, C. Azerraf, S. Lapidot, O. Shoseyov, Cellulose

nanocrystals (CNCs) induced crystallization of polyvinyl alcohol (PVA) super performing nanocomposite

films, Macromolecular bioscience 19(3) (2019) 1800347.

187

[64] T.S. Anirudhan, S.R. Rejeena, Poly (acrylic acid‐co‐acrylamide‐co‐2‐acrylamido‐2‐methyl‐1‐

propanesulfonic acid)‐grafted nanocellulose/poly (vinyl alcohol) composite for the in vitro

gastrointestinal release of amoxicillin, Journal of Applied Polymer Science 131(17) (2014).

[65] S.Y. Ooi, I. Ahmad, M.C.I.M. Amin, Cellulose nanocrystals extracted from rice husks as a

reinforcing material in gelatin hydrogels for use in controlled drug delivery systems, Industrial Crops and

Products 93 (2016) 227-234.

[66] S.A.A. Ghavimi, E.S. Lungren, T.J. Faulkner, M.A. Josselet, Y. Wu, Y. Sun, F.M. Pfeiffer, C.L.

Goldstein, C. Wan, B.D. Ulery, Inductive co-crosslinking of cellulose nanocrystal/chitosan hydrogels for

the treatment of vertebral compression fractures, International journal of biological macromolecules 130

(2019) 88-98.

[67] P. Bertsch, L. Schneider, G. Bovone, M.W. Tibbitt, P. Fischer, S. Gstohl, Injectable Biocompatible

Hydrogels from Cellulose Nanocrystals for Locally Targeted Sustained Drug Release, ACS Applied

Materials & Interfaces 11(42) (2019) 38578-38585.

[68] R. Nigmatullin, R. Harniman, V. Gabrielli, J.C. Munoz-Garcia, Y.Z. Khimyak, J.s. Angulo, S.J.

Eichhorn, Mechanically robust gels formed from hydrophobized cellulose nanocrystals, ACS applied

materials & interfaces 10(23) (2018) 19318-19322.

[69] H. Liu, C. Li, B. Wang, X. Sui, L. Wang, X. Yan, H. Xu, L. Zhang, Y. Zhong, Z. Mao, Self-healing

and injectable polysaccharide hydrogels with tunable mechanical properties, Cellulose 25(1) (2018) 559-

571.

[70] A. Khabibullin, M. Alizadehgiashi, N. Khuu, E. Prince, M. Tebbe, E. Kumacheva, Injectable shear-

thinning fluorescent hydrogel formed by cellulose nanocrystals and graphene quantum dots, Langmuir

33(43) (2017) 12344-12350.

[71] J. Tang, M.U. Javaid, C. Pan, G. Yu, R.M. Berry, K.C. Tam, Self-healing stimuli-responsive

cellulose nanocrystal hydrogels, Carbohydrate Polymers 229 (2020) 115486.

[72] Q. Wu, Y. Meng, S. Wang, Y. Li, S. Fu, L. Ma, D. Harper, Rheological behavior of cellulose

nanocrystal suspension: Influence of concentration and aspect ratio, Journal of Applied Polymer Science

131(15) (2014).

[73] C. Zhou, Q. Wu, Y. Yue, Q. Zhang, Application of rod-shaped cellulose nanocrystals in

polyacrylamide hydrogels, Journal of colloid and interface science 353(1) (2011) 116-123.

[74] J.R. McKee, S. Hietala, J. Seitsonen, J. Laine, E. Kontturi, O. Ikkala, Thermoresponsive

nanocellulose hydrogels with tunable mechanical properties, ACS Macro Letters 3(3) (2014) 266-270.

[75] A.-L. Oechsle, L. Lewis, W.Y. Hamad, S.G. Hatzikiriakos, M.J. MacLachlan, CO2-switchable

cellulose nanocrystal hydrogels, Chemistry of Materials 30(2) (2018) 376-385.

[76] M. Talantikite, A. Gourlay, S.L. Gall, B. Cathala, Influence of Xyloglucan Molar Mass on

Rheological Properties of Cellulose Nanocrystal/Xyloglucan Hydrogels, Journal of Renewable Materials

7(12) (2019) 1381-1390.

[77] K. Hou, Y. Li, Y. Liu, R. Zhang, B.S. Hsiao, M. Zhu, Continuous fabrication of cellulose

nanocrystal/poly (ethylene glycol) diacrylate hydrogel fiber from nanocomposite dispersion: rheology,

preparation and characterization, Polymer 123 (2017) 55-64.

[78] S. Shafiei-Sabet, W.Y. Hamad, S.G. Hatzikiriakos, Rheology of nanocrystalline cellulose aqueous

suspensions, Langmuir 28(49) (2012) 17124-17133.

[79] J. Wang, A. Chiappone, I. Roppolo, F. Shao, E. Fantino, M. Lorusso, D. Rentsch, K. Dietliker, C.F.

Pirri, H. Grützmacher, All‐in‐One Cellulose Nanocrystals for 3D Printing of Nanocomposite Hydrogels,

Angewandte Chemie International Edition 57(9) (2018) 2353-2356.

[80] H. Huang, X. Wang, J. Yu, Y. Chen, H. Ji, Y. Zhang, F. Rehfeldt, Y. Wang, K. Zhang, Liquid-

Behaviors-Assisted Fabrication of Multidimensional Birefringent Materials from Dynamic Hybrid

Hydrogels, ACS nano 13(4) (2019) 3867-3874.

[81] K.M. Rao, A. Kumar, S.S. Han, Polysaccharide based bionanocomposite hydrogels reinforced with

cellulose nanocrystals: drug release and biocompatibility analyses, International journal of biological

macromolecules 101 (2017) 165-171.

188

[82] M. Ghorbani, L. Roshangar, Construction of collagen/nanocrystalline cellulose based-hydrogel

scaffolds: synthesis, characterization, and mechanical properties evaluation, International Journal of

Polymeric Materials and Polymeric Biomaterials (2019) 1-7.

[83] J. Han, T. Lei, Q. Wu, Facile preparation of mouldable polyvinyl alcohol-borax hydrogels reinforced

by well-dispersed cellulose nanoparticles: physical, viscoelastic and mechanical properties, Cellulose

20(6) (2013) 2947-2958.

[84] C. García-Astrain, K. González, T. Gurrea, O. Guaresti, I. Algar, A. Eceiza, N. Gabilondo,

Maleimide-grafted cellulose nanocrystals as cross-linkers for bionanocomposite hydrogels, Carbohydrate

polymers 149 (2016) 94-101.

[85] D.L. Taylor, M. in het Panhuis, Self‐healing hydrogels, Advanced Materials 28(41) (2016) 9060-

9093.

[86] J. Pawley, Handbook of biological confocal microscopy, Springer Science & Business Media2006.

[87] F. Fergg, F. Keil, H. Quader, Investigations of the microscopic structure of poly (vinyl alcohol)

hydrogels by confocal laser scanning microscopy, Colloid and Polymer Science 279(1) (2001) 61-67.

[88] I. Savina, M. Tuncel, A. Tuncel, I.Y. Galaev, B. Mattiasson, Analysis of polymer grafted inside the

porous hydrogel using confocal laser scanning microscopy, eXPRESS Polymer Letters 1(4) (2007) 189-

196.

[89] T. Koyano, N. Minoura, M. Nagura, K.i. Kobayashi, Attachment and growth of cultured fibroblast

cells on PVA/chitosan‐blended hydrogels, Journal of Biomedical Materials Research: An Official Journal

of The Society for Biomaterials, The Japanese Society for Biomaterials, and the Australian Society for

Biomaterials 39(3) (1998) 486-490.

[90] C.B. Highley, C.B. Rodell, J.A. Burdick, Direct 3D printing of shear‐thinning hydrogels into self‐

healing hydrogels, Advanced Materials 27(34) (2015) 5075-5079.

[91] K.E. Crompton, R.J. Prankerd, D.M. Paganin, T.F. Scott, M.K. Horne, D.I. Finkelstein, K.A. Gross,

J.S. Forsythe, Morphology and gelation of thermosensitive chitosan hydrogels, Biophysical chemistry

117(1) (2005) 47-53.

[92] S. Gorman, M. Tunney, P. Keane, K. Van Bladel, B. Bley, Characterization and assessment of a

novel poly (ethylene oxide)/polyurethane composite hydrogel (Aquavene®) as a ureteral stent

biomaterial, Journal of Biomedical Materials Research: An Official Journal of The Society for

Biomaterials, The Japanese Society for Biomaterials, and the Australian Society for Biomaterials 39(4)

(1998) 642-649.

[93] K. Yasuda, J.P. Gong, Y. Katsuyama, A. Nakayama, Y. Tanabe, E. Kondo, M. Ueno, Y. Osada,

Biomechanical properties of high-toughness double network hydrogels, Biomaterials 26(21) (2005) 4468-

4475.

[94] Z. Gong, Y. Yang, Q. Ren, X. Chen, Z. Shao, Injectable thixotropic hydrogel comprising regenerated

silk fibroin and hydroxypropylcellulose, Soft Matter 8(10) (2012) 2875-2883.

[95] M. Auty, M. Fenelon, T. Guinee, C. Mullins, D. Mulvihill, Dynamic confocal scanning laser

microscopy methods for studying milk protein gelation and cheese melting, Scanning 21(5) (1999) 299-

304.

[96] P.J. Skrzeszewska, J. Sprakel, F.A. de Wolf, R. Fokkink, M.A. Cohen Stuart, J. van der Gucht,

Fracture and self-healing in a well-defined self-assembled polymer network, Macromolecules 43(7)

(2010) 3542-3548.

[97] X. Yu, X. Cao, L. Chen, H. Lan, B. Liu, T. Yi, Thixotropic and self-healing triggered reversible

rheology switching in a peptide-based organogel with a cross-linked nano-ring pattern, Soft Matter 8(12)

(2012) 3329-3334.

[98] J. Yuan, X. Fang, L. Zhang, G. Hong, Y. Lin, Q. Zheng, Y. Xu, Y. Ruan, W. Weng, H. Xia, Multi-

responsive self-healing metallo-supramolecular gels based on “click” ligand, Journal of Materials

Chemistry 22(23) (2012) 11515-11522.

[99] F. Herbst, S. Seiffert, W.H. Binder, Dynamic supramolecular poly (isobutylene) s for self-healing

materials, Polymer Chemistry 3(11) (2012) 3084-3092.

189

[100] C. Tucker, S.G. Advani, Processing of short-fiber systems, Composite Materials Series (1994) 147-

147.

[101] F. Dupret, V. Verleye, Advances in the Flow and Rheology of Non-Newtonian Fluids, Modelling

the flow of fiber suspensions in narrow gaps, Rheology series. Amsterdam: Elsevier (1999) 1347-1398.

[102] G.B. Jeffery, The motion of ellipsoidal particles immersed in a viscous fluid, Proceedings of the

Royal Society of London. Series A, Containing papers of a mathematical and physical character 102(715)

(1922) 161-179.

[103] L. Leal, E. Hinch, The effect of weak Brownian rotations on particles in shear flow, Journal of

Fluid Mechanics 46(4) (1971) 685-703.

[104] E. Hinch, L. Leal, The effect of Brownian motion on the rheological properties of a suspension of

non-spherical particles, Journal of Fluid Mechanics 52(4) (1972) 683-712.

[105] E. Shaqfeh, D.L. Koch, The effect of hydrodynamic interactions on the orientation of axisymmetric

particles flowing through a fixed bed of spheres or fibers, The Physics of fluids 31(4) (1988) 728-743.

[106] R.S. Bay, C.L. Tucker III, Fiber orientation in simple injection moldings. Part II: Experimental

results, Polymer composites 13(4) (1992) 332-341.

[107] J.H. Phelps, C.L. Tucker III, An anisotropic rotary diffusion model for fiber orientation in short-and

long-fiber thermoplastics, Journal of Non-Newtonian Fluid Mechanics 156(3) (2009) 165-176.

[108] N. Phan-Thien, X.-J. Fan, R. Tanner, R. Zheng, Folgar–Tucker constant for a fibre suspension in a

Newtonian fluid, Journal of Non-Newtonian Fluid Mechanics 103(2-3) (2002) 251-260.

[109] S.G. Advani, C.L. Tucker III, The use of tensors to describe and predict fiber orientation in short

fiber composites, Journal of rheology 31(8) (1987) 751-784.

[110] R.S. Bay, Fiber orientation in injection-molded composites: a comparison of theory and experiment,

University of Illinois at Urbana-Champaign, 1991.

[111] S. Shafiei Sabet, Shear rheology of cellulose nanocrystal (CNC) aqueous suspensions, University of

British Columbia, 2013.

[112] H. Lv, L. Li, M. Sun, Y. Zhang, L. Chen, Y. Rong, Y. Li, Mechanism of regulation of stem cell

differentiation by matrix stiffness, Stem cell research & therapy 6(1) (2015) 103.

[113] J.H. Wen, L.G. Vincent, A. Fuhrmann, Y.S. Choi, K.C. Hribar, H. Taylor-Weiner, S. Chen, A.J.

Engler, Interplay of matrix stiffness and protein tethering in stem cell differentiation, Nature materials

13(10) (2014) 979.

[114] O. Chaudhuri, L. Gu, D. Klumpers, M. Darnell, S.A. Bencherif, J.C. Weaver, N. Huebsch, H.-p.

Lee, E. Lippens, G.N. Duda, Hydrogels with tunable stress relaxation regulate stem cell fate and activity,

Nature materials 15(3) (2016) 326.

[115] S.C. Cowin, Bone mechanics handbook, CRC press2001.

[116] Y. Chen, W. Xu, W. Liu, G. Zeng, Responsiveness, swelling, and mechanical properties of PNIPA

nanocomposite hydrogels reinforced by nanocellulose, Journal of Materials Research 30(11) (2015) 1797-

1807.

[117] M.A. Karaaslan, M.A. Tshabalala, D.J. Yelle, G. Buschle-Diller, Nanoreinforced biocompatible

hydrogels from wood hemicelluloses and cellulose whiskers, Carbohydrate Polymers 86(1) (2011) 192-

201.

[118] K.J. De France, K.J. Chan, E.D. Cranston, T. Hoare, Enhanced mechanical properties in cellulose

nanocrystal–poly (oligoethylene glycol methacrylate) injectable nanocomposite hydrogels through control

of physical and chemical cross-linking, Biomacromolecules 17(2) (2016) 649-660.

[119] J. Yang, C.-R. Han, J.-F. Duan, F. Xu, R.-C. Sun, Mechanical and viscoelastic properties of

cellulose nanocrystals reinforced poly (ethylene glycol) nanocomposite hydrogels, ACS applied materials

& interfaces 5(8) (2013) 3199-3207.

[120] K.J. Le Goff, C. Gaillard, W. Helbert, C. Garnier, T. Aubry, Rheological study of reinforcement of

agarose hydrogels by cellulose nanowhiskers, Carbohydrate polymers 116 (2015) 117-123.

[121] J.R. McKee, E.A. Appel, J. Seitsonen, E. Kontturi, O.A. Scherman, O. Ikkala, Healable, stable and

stiff hydrogels: combining conflicting properties using dynamic and selective three‐component

190

recognition with reinforcing cellulose nanorods, Advanced Functional Materials 24(18) (2014) 2706-

2713.

[122] Y. Wang, L. Chen, Impacts of nanowhisker on formation kinetics and properties of all-cellulose

composite gels, Carbohydrate polymers 83(4) (2011) 1937-1946.

[123] J. Yang, C. Han, Mechanically viscoelastic properties of cellulose nanocrystals skeleton reinforced

hierarchical composite hydrogels, ACS applied materials & interfaces 8(38) (2016) 25621-25630.

[124] K. Hyun, M. Wilhelm, C.O. Klein, K.S. Cho, J.G. Nam, K.H. Ahn, S.J. Lee, R.H. Ewoldt, G.H.

McKinley, A review of nonlinear oscillatory shear tests: Analysis and application of large amplitude

oscillatory shear (LAOS), Progress in Polymer Science 36(12) (2011) 1697-1753.

[125] W. Cox, E. Merz, Correlation of dynamic and steady flow viscosities, Journal of Polymer Science

28(118) (1958) 619-622.

[126] C. Saengow, A.J. Giacomin, Review of nonlinear oscillatory shear flow notations and

presentations: Polymeric liquids, Current opinion in colloid & interface science (2019).

[127] M. Wilhelm, P. Reinheimer, M. Ortseifer, High sensitivity Fourier-transform rheology, Rheologica

Acta 38(4) (1999) 349-356.

[128] K.S. Cho, K. Hyun, K.H. Ahn, S.J. Lee, A geometrical interpretation of large amplitude oscillatory

shear response, Journal of rheology 49(3) (2005) 747-758.

[129] A.J. Giacomin, J.M. Dealy, Large-amplitude oscillatory shear, Techniques in rheological

measurement, Springer1993, pp. 99-121.

[130] R.H. Ewoldt, A. Hosoi, G.H. McKinley, New measures for characterizing nonlinear viscoelasticity

in large amplitude oscillatory shear, Journal of Rheology 52(6) (2008) 1427-1458.

[131] R.H. Ewoldt, P. Winter, J. Maxey, G.H. McKinley, Large amplitude oscillatory shear of

pseudoplastic and elastoviscoplastic materials, Rheologica acta 49(2) (2010) 191-212.

[132] S.A. Rogers, B.M. Erwin, D. Vlassopoulos, M. Cloitre, A sequence of physical processes

determined and quantified in LAOS: Application to a yield stress fluid, Journal of Rheology 55(2) (2011)

435-458.

[133] K. Hyun, W. Kim, A new non-linear parameter Q from FT-Rheology under nonlinear dynamic

oscillatory shear for polymer melts system, Korea-Australia Rheology Journal 23(4) (2011) 227-235.

[134] R. Salehiyan, Y. Yoo, W.J. Choi, K. Hyun, Characterization of morphologies of compatibilized

polypropylene/polystyrene blends with nanoparticles via nonlinear rheological properties from FT-

rheology, Macromolecules 47(12) (2014) 4066-4076.

[135] R. Prathapan, R. Thapa, G. Garnier, R.F. Tabor, Modulating the zeta potential of cellulose

nanocrystals using salts and surfactants, Colloids and Surfaces A: Physicochemical and Engineering

Aspects 509 (2016) 11-18.

[136] S. Shafiei-Sabet, W. Hamad, S. Hatzikiriakos, Ionic strength effects on the microstructure and shear

rheology of cellulose nanocrystal suspensions, Cellulose 21(5) (2014) 3347-3359.

[137] G. Lenfant, M.-C. Heuzey, T.G. van de Ven, P.J. Carreau, A comparative study of ECNC and CNC

suspensions: effect of salt on rheological properties, Rheologica Acta 56(1) (2017) 51-62.

[138] K.R. Peddireddy, I. Capron, T. Nicolai, L. Benyahia, Gelation Kinetics and Network Structure of

Cellulose Nanocrystals in Aqueous Solution, Biomacromolecules 17(10) (2016) 3298-3304.

[139] M. Owczarz, A.C. Motta, M. Morbidelli, P. Arosio, A Colloidal Description of Intermolecular

Interactions Driving Fibril–Fibril Aggregation of a Model Amphiphilic Peptide, Langmuir 31(27) (2015)

7590-7600.

[140] R.N. Zia, B.J. Landrum, W.B. Russel, A micro-mechanical study of coarsening and rheology of

colloidal gels: Cage building, cage hopping, and Smoluchowski's ratchet, Journal of Rheology 58(5)

(2014) 1121-1157.

[141] W.C. Poon, L. Starrs, S. Meeker, A. Moussaid, R.M. Evans, P. Pusey, M. Robins, Delayed

sedimentation of transient gels in colloid–polymer mixtures: dark-field observation, rheology and

dynamic light scattering studies, Faraday Discussions 112 (1999) 143-154.

[142] M.J. Solomon, P.T. Spicer, Microstructural regimes of colloidal rod suspensions, gels, and glasses,

Soft Matter 6(7) (2010) 1391-1400.

191

[143] P.A. Buining, C. Pathmamanoharan, J.B.H. Jansen, H.N. Lekkerkerker, Preparation of colloidal

boehmite needles by hydrothermal treatment of aluminum alkoxide precursors, Journal of the American

Ceramic Society 74(6) (1991) 1303-1307.

[144] A. Mohraz, D.B. Moler, R.M. Ziff, M.J. Solomon, Effect of monomer geometry on the fractal

structure of colloidal rod aggregates, Physical review letters 92(15) (2004) 155503.

[145] B. Derjaguin, L. Landau, Theory of the stability of strongly charged lyophobic sols and of the

adhesion of strongly charged particles in solutions of electrolytes, Acta physicochim. URSS 14(6) (1941)

633-662.

[146] E.J.W. Verwey, J.T.G. Overbeek, J.T.G. Overbeek, Theory of the stability of lyophobic colloids,

Courier Corporation1999.

[147] Y. Boluk, R. Lahiji, L. Zhao, M.T. McDermott, Suspension viscosities and shape parameter of

cellulose nanocrystals (CNC), Colloids and Surfaces A: Physicochemical and Engineering Aspects 377(1)

(2011) 297-303.

[148] J. Araki, Electrostatic or steric?–preparations and characterizations of well-dispersed systems

containing rod-like nanowhiskers of crystalline polysaccharides, Soft Matter 9(16) (2013) 4125-4141.

[149] K. Ako, D. Durand, T. Nicolai, L. Becu, Quantitative analysis of confocal laser scanning

microscopy images of heat-set globular protein gels, Food Hydrocolloids 23(4) (2009) 1111-1119.

[150] D.C. Grahame, The electrical double layer and the theory of electrocapillarity, Chemical reviews

41(3) (1947) 441-501.

[151] L. Zhong, S. Fu, X. Peng, H. Zhan, R. Sun, Colloidal stability of negatively charged cellulose

nanocrystalline in aqueous systems, Carbohydrate polymers 90(1) (2012) 644-649.

[152] G. Quezada, R.I. Jeldres, C. Goñi, P.G. Toledo, A.D. Stickland, P.J. Scales, Viscoelastic behaviour

of flocculated silica sediments in concentrated monovalent chloride salt solutions, Minerals Engineering

110 (2017) 131-138.

[153] A. McKenzie, STRUCTURE AND PROPERTIES OF PAPER. 18. RETENTION OF WET-END

ADDITIVES, Appita 21(4) (1968) 104-&.

[154] J.-Q. Hu, Q. Chen, Z.-X. Xie, G.-B. Han, R.-H. Wang, B. Ren, Y. Zhang, Z.-L. Yang, Z.-Q. Tian, A

simple and effective route for the synthesis of crystalline silver nanorods and nanowires, Advanced

Functional Materials 14(2) (2004) 183-189.

[155] R. Buscall, T.H. Choudhury, M.A. Faers, J.W. Goodwin, P.A. Luckham, S.J. Partridge, Towards

rationalising collapse times for the delayed sedimentation of weakly-aggregated colloidal gels, Soft

Matter 5(7) (2009) 1345-1349.

[156] L.J. Teece, M.A. Faers, P. Bartlett, Ageing and collapse in gels with long-range attractions, Soft

Matter 7(4) (2011) 1341-1351.

[157] N. Verhaegh, H. Lekkerkerker, Phase transitions in colloidal suspensions, International School of

Physics Enrico Fermi, IOS Press, 1997, pp. 347-381.

[158] I. Zhang, C.P. Royall, M.A. Faers, P. Bartlett, Phase separation dynamics in colloid–polymer

mixtures: the effect of interaction range, Soft Matter 9(6) (2013) 2076-2084.

[159] Y.-L. Chen, V. Kobelev, K. Schweizer, Barrier hopping, viscous flow, and kinetic gelation in

particle-polymer suspensions, Physical Review E 71(4) (2005) 041405.

[160] V. Gopalakrishnan, K. Schweizer, C. Zukoski, Linking single particle rearrangements to delayed

collapse times in transient depletion gels, Journal of Physics: Condensed Matter 18(50) (2006) 11531.

[161] A. Potanin, W. Russel, Fractal model of consolidation of weakly aggregated colloidal dispersions,

Physical Review E 53(4) (1996) 3702.

[162] P. Bartlett, L.J. Teece, M.A. Faers, Sudden collapse of a colloidal gel, Physical Review E 85(2)

(2012) 021404.

[163] R.J. d’Arjuzon, W. Frith, J.R. Melrose, Brownian dynamics simulations of aging colloidal gels,


[164] A. Fluerasu, A. Moussaïd, A. Madsen, A. Schofield, Slow dynamics and aging in colloidal gels

studied by x-ray photon correlation spectroscopy, Physical Review E 76(1) (2007) 010401.

192

[165] C. Allain, M. Cloitre, M. Wafra, Aggregation and sedimentation in colloidal suspensions, Physical

review letters 74(8) (1995) 1478.

[166] L. Starrs, W. Poon, D. Hibberd, M. Robins, Collapse of transient gels in colloid-polymer mixtures,

Journal of Physics: Condensed Matter 14(10) (2002) 2485.

[167] A. Parker, P.A. Gunning, K. Ng, M.M. Robins, How does xanthan stabilise salad dressing?, Food

Hydrocolloids 9(4) (1995) 333-342.

[168] C.P. Royall, S.R. Williams, T. Ohtsuka, H. Tanaka, Direct observation of a local structural

mechanism for dynamic arrest, Nature materials 7(7) (2008) 556.

[169] S. Buzzaccaro, E. Secchi, G. Brambilla, R. Piazza, L. Cipelletti, Equilibrium concentration profiles

and sedimentation kinetics of colloidal gels under gravitational stress, Journal of Physics: Condensed

Matter 24(28) (2012) 284103.

[170] S.W. Kamp, M.L. Kilfoil, Universal behaviour in the mechanical properties of weakly aggregated

colloidal particles, Soft Matter 5(12) (2009) 2438-2447.

[171] J. Plášek, B. Hošková, Solvatochromic effect in the optical spectra of calcofluor and its relation to

fluorescent staining of yeast cell walls, Journal of fluorescence 20(1) (2010) 343-352.

[172] Y.P. Zhang, V.P. Chodavarapu, A.G. Kirk, M.P. Andrews, Nanocrystalline cellulose for covert

optical encryption, Journal of Nanophotonics 6(1) (2012) 063516.

[173] A.A. Moud, M. Arjmand, N. Yan, A.S. Nezhad, S.H. Hejazi, Colloidal behavior of cellulose

nanocrystals in presence of sodium chloride, ChemistrySelect 3(17) (2018) 4969-4978.

[174] E.E.U. Benavides, Cellulose nanocrystals properties and applications in renewable nanocomposites,

Clemson University, 2011.

[175] J.T.G. Overbeek, Recent developments in the understanding of colloid stability, Plenary and Invited

Lectures, Elsevier1977, pp. 431-445.

[176] M. Sano, J. Okamura, S. Shinkai, Colloidal nature of single-walled carbon nanotubes in electrolyte

solution: the Schulze− Hardy rule, Langmuir 17(22) (2001) 7172-7173.

[177] J.C. Stendahl, M.S. Rao, M.O. Guler, S.I. Stupp, Intermolecular forces in the self‐assembly of

peptide amphiphile nanofibers, Advanced Functional Materials 16(4) (2006) 499-508.

[178] B.G. Rånby, Fibrous macromolecular systems. Cellulose and muscle. The colloidal properties of

cellulose micelles, Discussions of the Faraday Society 11 (1951) 158-164.

[179] L. Wågberg, G. Decher, M. Norgren, T. Lindström, M. Ankerfors, K. Axnäs, The build-up of

polyelectrolyte multilayers of microfibrillated cellulose and cationic polyelectrolytes, Langmuir 24(3)

(2008) 784-795.

[180] T. Lindström, C. Soremark, Flocculation of cellulosic dispersions with alginates in the presence of

divalent metal ions, Journal of Colloid and Interface Science 55(1) (1976) 69-72.

[181] S. Kratohvil, G.E. Janauer, E. Matijević, Coagulation of microcrystalline cellulose dispersions,

Journal of colloid and interface science 29(2) (1969) 187-193.

[182] M.M. de Souza Lima, R. Borsali, Static and dynamic light scattering from polyelectrolyte

microcrystal cellulose, Langmuir 18(4) (2002) 992-996.

[183] X.M. Dong, J.-F. Revol, D.G. Gray, Effect of microcrystallite preparation conditions on the

formation of colloid crystals of cellulose, Cellulose 5(1) (1998) 19-32.

[184] S. Elazzouzi-Hafraoui, J.-L. Putaux, L. Heux, Self-assembling and chiral nematic properties of

organophilic cellulose nanocrystals, The Journal of Physical Chemistry B 113(32) (2009) 11069-11075.

[185] J.-F. Revol, On the cross-sectional shape of cellulose crystallites in Valonia ventricosa,

Carbohydrate Polymers 2(2) (1982) 123-134.

[186] W. Chen, S. Enck, J.L. Price, D.L. Powers, E.T. Powers, C.-H. Wong, H.J. Dyson, J.W. Kelly,

Structural and energetic basis of carbohydrate–aromatic packing interactions in proteins, Journal of the

American Chemical Society 135(26) (2013) 9877-9884.

[187] J.L. Asensio, A. Ardá, F.J. Canada, J.s. Jimenez-Barbero, Carbohydrate–aromatic interactions,

Accounts of chemical research 46(4) (2012) 946-954.

[188] S. Plimpton, Fast parallel algorithms for short-range molecular dynamics, Journal of Computational

Physics 117(1) (1995) 1-19.

193

[189] W. Humphrey, A. Dalke, K. Schulten, VMD: visual molecular dynamics, J Mol Graph 14(1) (1996)

33-38.

[190] R. Kjellander, H. Greberg, Mechanisms behind concentration profiles illustrated by charge and

concentration distributions around ions in double layers, Journal of Electroanalytical Chemistry 450(2)

(1998) 233-251.

[191] G. Fiorin, M.L. Klein, J. Hénin, Using collective variables to drive molecular dynamics

simulations, Molecular Physics 111(22-23) (2013) 3345-3362.

[192] H.A. Lorentz, Ueber die Anwendung des Satzes vom Virial in der kinetischen Theorie der Gase,

Annalen Der Physik 248(1) (2010) 127-136.

[193] T. Hagiwara, H. Kumagai, K. Nakamura, Fractal analysis of aggregates formed by heating dilute

BSA solutions using light scattering methods, Bioscience, biotechnology, and biochemistry 60(11) (1996)

1757-1763.

[194] C. Honorato-Rios, C. Lehr, C. Schütz, R. Sanctuary, M.A. Osipov, J. Baller, J.P. Lagerwall,

Fractionation of cellulose nanocrystals: enhancing liquid crystal ordering without promoting gelation,

NPG Asia Materials (2018) 1.

[195] Q. Chen, H. Cho, K. Manthiram, M. Yoshida, X. Ye, A.P. Alivisatos, Interaction potentials of

anisotropic nanocrystals from the trajectory sampling of particle motion using in situ liquid phase

transmission electron microscopy, ACS central science 1(1) (2015) 33-39.

[196] K. Liu, Z. Nie, N. Zhao, W. Li, M. Rubinstein, E. Kumacheva, Step-growth polymerization of

inorganic nanoparticles, science 329(5988) (2010) 197-200.

[197] K.J. Bishop, Nanoscale Self-Assembly: Seeing Is Understanding, ACS Publications, 2015.

[198] A. Moncho-Jordá, F. Martınez-López, R. Hidalgo-Alvarez, The effect of the salt concentration and

counterion valence on the aggregation of latex particles at the air/water interface, Journal of colloid and

interface science 249(2) (2002) 405-411.

[199] J. Gregory, The effect of cationic polymers on the colloidal stability of latex particles, Journal of

Colloid and Interface Science 55(1) (1976) 35-44.

[200] M. Sano, A. Kamino, S. Shinkai, Critical Coagulation of Langmuir Monolayers: 2D Schulze−

Hardy Rule, The Journal of Physical Chemistry B 104(44) (2000) 10339-10347.

[201] T. Phan-Xuan, A. Thuresson, M. Skepö, A. Labrador, R. Bordes, A. Matic, Aggregation behavior

of aqueous cellulose nanocrystals: the effect of inorganic salts, Cellulose 23(6) (2016) 3653-3663.

[202] W.Y. Hamad, T.Q. Hu, Structure–process–yield interrelations in nanocrystalline cellulose

extraction, The Canadian Journal of Chemical Engineering 88(3) (2010) 392-402.

[203] J. Israelachvili, Intermolecular and Surface Forces 3rd edn (Burlington, MA: Academic), (2011).

[204] A.M. Puertas, M. Fuchs, M.E. Cates, Aging in attraction-driven colloidal glasses, Physical Review

E 75(3) (2007) 031401.

[205] F. Höfling, T. Munk, E. Frey, T. Franosch, Critical dynamics of ballistic and Brownian particles in

a heterogeneous environment, The Journal of chemical physics 128(16) (2008) 164517.

[206] T. Bauer, F. Höfling, T. Munk, E. Frey, T. Franosch, The localization transition of the two-

dimensional Lorentz model, The European Physical Journal Special Topics 189(1) (2010) 103-118.

[207] T. Schilling, S. Jungblut, M.A. Miller, Depletion-induced percolation in networks of nanorods,

Physical review letters 98(10) (2007) 108303.

[208] J.K. Dhont, W.J. Briels, Viscoelasticity of suspensions of long, rigid rods, Colloids and surfaces A:

Physicochemical and engineering aspects 213(2-3) (2003) 131-156.

[209] C.A. Stover, D.L. Koch, C. Cohen, Observations of fibre orientation in simple shear flow of semi-

dilute suspensions, Journal of Fluid Mechanics 238 (1992) 277-296.

[210] P.N. Pusey, W. Van Megen, Dynamic light scattering by non-ergodic media, Physica A: Statistical

Mechanics and its Applications 157(2) (1989) 705-741.

[211] T.C. Gomes, M.S. Skaf, Cellulose-builder: a toolkit for building crystalline structures of cellulose,

Journal of Computational Chemistry 33(14) (2012) 1338-1346.

[212] H. Sun, S.J. Mumby, J.R. Maple, A.T. Hagler, An ab Initio CFF93 All-Atom Force Field for

Polycarbonates, Journal of the American Chemical Society 116(7) (1994) 2978-2987.

194

[213] M. Mariano, N. El Kissi, A. Dufresne, Cellulose nanocrystals and related nanocomposites: review

of some properties and challenges, Journal of Polymer Science Part B: Polymer Physics 52(12) (2014)

791-806.

[214] H. Oguzlu, C. Danumah, Y. Boluk, Colloidal behavior of aqueous cellulose nanocrystal

suspensions, Current opinion in colloid & interface science 29 (2017) 46-56.

[215] A. Vedadghavami, F. Minooei, M.H. Mohammadi, S. Khetani, A.R. Kolahchi, S. Mashayekhan, A.

Sanati-Nezhad, Manufacturing of hydrogel biomaterials with controlled mechanical properties for tissue

engineering applications, Acta biomaterialia 62 (2017) 42-63.

[216] A.A. Moud, M. Arjmand, J. Liu, Y. Yang, A. Sanati-Nezhad, S.H. Hejazi, Cellulose nanocrystal

structure in the presence of salts, Cellulose 1-15.

[217] M. Conti, B. Meerson, P.V. Sasorov, Breakdown of scale invariance in the phase ordering of fractal

clusters, Physical review letters 80(21) (1998) 4693.

[218] J. Groenewold, T. Zhang, W.K. Kegel, Electrophoresis in charge-stabilized colloidal cluster phases,

The Journal of Physical Chemistry B 115(22) (2011) 7264-7267.

[219] M. Kamkar, E. Aliabadian, A. Shayesteh Zeraati, U. Sundararaj, Application of nonlinear rheology

to assess the effect of secondary nanofiller on network structure of hybrid polymer nanocomposites,

Physics of Fluids 30(2) (2018) 023102.

[220] M. Kamkar, S. Sadeghi, M. Arjmand, U. Sundararaj, Structural Characterization of CVD Custom-

Synthesized Carbon Nanotube/Polymer Nanocomposites in Large-Amplitude Oscillatory Shear (LAOS)

Mode: Effect of Dispersion Characteristics in Confined Geometries, Macromolecules (2019).

[221] M. Kamkar, S.N. Sultana, S.P. Pawar, A. Eshraghian, E. Erfanian, U. Sundararaj, The Key Role of

Processing in Tuning Nonlinear Viscoelastic Properties and Microwave Absorption in CNT-based

Polymer Nanocomposites, Materials Today Communications (2020) 101010.

[222] T.B. Goudoulas, N. Germann, Nonlinear rheological behavior of gelatin gels: In situ gels and

individual layers, Journal of colloid and interface science 553 (2019) 746-757.

[223] T.B. Goudoulas, N. Germann, Nonlinear rheological behavior of gelatin gels: In situ gels and

individual gel layers filled with hard particles, Journal of colloid and interface science 556 (2019) 1-11.

[224] R. Salehiyan, H.Y. Song, W.J. Choi, K. Hyun, Characterization of effects of silica nanoparticles on

(80/20) PP/PS blends via nonlinear rheological properties from Fourier transform rheology,

Macromolecules 48(13) (2015) 4669-4679.

[225] Y. Chen, C. Xu, J. Huang, D. Wu, Q. Lv, Rheological properties of nanocrystalline cellulose

suspensions, Carbohydrate polymers 157 (2017) 303-310.

[226] E. Aliabadian, M. Kamkar, Z. Chen, U. Sundararaj, Prevention of network destruction of partially

hydrolyzed polyacrylamide (HPAM): Effects of salt, temperature, and fumed silica nanoparticles, Physics

of Fluids 31(1) (2019) 013104.

[227] A. Zaccone, H. Wu, E. Del Gado, Elasticity of arrested short-ranged attractive colloids:

Homogeneous and heterogeneous glasses, Physical review letters 103(20) (2009) 208301.

[228] M. Laurati, S. Egelhaaf, G. Petekidis, Nonlinear rheology of colloidal gels with intermediate

volume fraction, Journal of Rheology 55(3) (2011) 673-706.

[229] W.-H. Shih, W.Y. Shih, S.-I. Kim, J. Liu, I.A. Aksay, Scaling behavior of the elastic properties of

colloidal gels, Physical review A 42(8) (1990) 4772.

[230] H. Wu, M. Morbidelli, A model relating structure of colloidal gels to their elastic properties,

Langmuir 17(4) (2001) 1030-1036.

[231] E. Aliabadian, S. Sadeghi, M. Kamkar, Z. Chen, U. Sundararaj, Rheology of fumed silica

nanoparticles/partially hydrolyzed polyacrylamide aqueous solutions under small and large amplitude

oscillatory shear deformations, Journal of Rheology 62(5) (2018) 1197-1216.

[232] T.G. Mason, D. Weitz, Optical measurements of frequency-dependent linear viscoelastic moduli of

complex fluids, Physical review letters 74(7) (1995) 1250.

[233] T.B. Goudoulas, N. Germann, Concentration effect on the nonlinear measures of dense

polyethylene oxide solutions under large amplitude oscillatory shear, Journal of Rheology 62(5) (2018)

1299-1317.

195

[234] C. Gao, G.Y. Xiong, H.L. Luo, K.J. Ren, Y. Huang, Y.Z. Wan, Dynamic interaction between the

growing Ca–P minerals and bacterial cellulose nanofibers during early biomineralization process,

Cellulose 17(2) (2010) 365-373.

[235] M. Roohani, Y. Habibi, N.M. Belgacem, G. Ebrahim, A.N. Karimi, A. Dufresne, Cellulose

whiskers reinforced polyvinyl alcohol copolymers nanocomposites, European polymer journal 44(8)

(2008) 2489-2498.

[236] E. Chiellini, A. Corti, S. D'Antone, R. Solaro, Biodegradation of poly (vinyl alcohol) based

materials, Progress in Polymer science 28(6) (2003) 963-1014.

[237] S. Jiang, S. Liu, W. Feng, PVA hydrogel properties for biomedical application, Journal of the

mechanical behavior of biomedical materials 4(7) (2011) 1228-1233.

[238] H.A. Khalil, Y. Davoudpour, M.N. Islam, A. Mustapha, K. Sudesh, R. Dungani, M. Jawaid,

Production and modification of nanofibrillated cellulose using various mechanical processes: a review,

Carbohydrate polymers 99 (2014) 649-665.

[239] E.C. Ramires, A. Dufresne, A review of cellulose nanocrystals and nanocomposites, Tappi J 10(4)

(2011) 9-16.

[240] A.B. Reising, R.J. Moon, J.P. Youngblood, Effect of particle alignment on mechanical properties of

neat cellulose nanocrystal films, Journal of Science & Technology for Forest Products and Processes:

Volume 2, Number 6, 2012; pp. 32-41. 2(6) (2012) 32-41.

[241] R.R. Lahiji, X. Xu, R. Reifenberger, A. Raman, A. Rudie, R.J. Moon, Atomic force microscopy

characterization of cellulose nanocrystals, Langmuir 26(6) (2010) 4480-4488.

[242] H.A. Silvério, W.P.F. Neto, N.O. Dantas, D. Pasquini, Extraction and characterization of cellulose

nanocrystals from corncob for application as reinforcing agent in nanocomposites, Industrial Crops and

Products 44 (2013) 427-436.

[243] X.M. Dong, D.G. Gray, Effect of counterions on ordered phase formation in suspensions of charged

rodlike cellulose crystallites, Langmuir 13(8) (1997) 2404-2409.

[244] T. Abitbol, H.S. Marway, S.A. Kedzior, X. Yang, A. Franey, D.G. Gray, E.D. Cranston, Hybrid

fluorescent nanoparticles from quantum dots coupled to cellulose nanocrystals, Cellulose 24(3) (2017)

1287-1293.

[245] E. Fortunati, F. Luzi, D. Puglia, A. Terenzi, M. Vercellino, L. Visai, C. Santulli, L. Torre, J. Kenny,

Ternary PVA nanocomposites containing cellulose nanocrystals from different sources and silver

particles: Part II, Carbohydrate polymers 97(2) (2013) 837-848.

[246] A.N. Frone, I. Chiulan, D.M. Panaitescu, C.A. Nicolae, M. Ghiurea, A.-M. Galan, Isolation of

cellulose nanocrystals from plum seed shells, structural and morphological characterization, Materials

Letters 194 (2017) 160-163.

[247] A. Pakzad, J. Simonsen, R.S. Yassar, Gradient of nanomechanical properties in the interphase of

cellulose nanocrystal composites, Composites Science and Technology 72(2) (2012) 314-319.

[248] M.S. Peresin, Y. Habibi, J.O. Zoppe, J.J. Pawlak, O.J. Rojas, Nanofiber composites of polyvinyl

alcohol and cellulose nanocrystals: manufacture and characterization, Biomacromolecules 11(3) (2010)

674-681.

[249] A.J. Uddin, J. Araki, Y. Gotoh, Characterization of the poly (vinyl alcohol)/cellulose whisker gel

spun fibers, Composites Part A: Applied Science and Manufacturing 42(7) (2011) 741-747.

[250] H. Yang, M.N. Alam, T.G. van de Ven, Highly charged nanocrystalline cellulose and

dicarboxylated cellulose from periodate and chlorite oxidized cellulose fibers, Cellulose 20(4) (2013)

1865-1875.

[251] J. Baier Leach, K.A. Bivens, C.W. Patrick Jr, C.E. Schmidt, Photocrosslinked hyaluronic acid

hydrogels: natural, biodegradable tissue engineering scaffolds, Biotechnology and bioengineering 82(5)

(2003) 578-589.

[252] S. Huan, L. Bai, G. Liu, W. Cheng, G. Han, Electrospun nanofibrous composites of polystyrene and

cellulose nanocrystals: manufacture and characterization, RSC Advances 5(63) (2015) 50756-50766.

[253] N.L.G. de Rodriguez, W. Thielemans, A. Dufresne, Sisal cellulose whiskers reinforced polyvinyl

acetate nanocomposites, Cellulose 13(3) (2006) 261-270.

196

[254] S. Elazzouzi-Hafraoui, Y. Nishiyama, J.-L. Putaux, L. Heux, F. Dubreuil, C. Rochas, The shape and

size distribution of crystalline nanoparticles prepared by acid hydrolysis of native cellulose,

Biomacromolecules 9(1) (2008) 57-65.

[255] M. Rosa, E. Medeiros, J. Malmonge, K. Gregorski, D. Wood, L. Mattoso, G. Glenn, W. Orts, S.

Imam, Cellulose nanowhiskers from coconut husk fibers: Effect of preparation conditions on their thermal

and morphological behavior, Carbohydrate Polymers 81(1) (2010) 83-92.

[256] G. Siqueira, J. Bras, A. Dufresne, Luffa cylindrica as a lignocellulosic source of fiber,

microfibrillated cellulose and cellulose nanocrystals, BioResources 5(2) (2010) 727-740.

[257] G. Tonoli, E. Teixeira, A. Corrêa, J. Marconcini, L. Caixeta, M. Pereira-da-Silva, L. Mattoso,

Cellulose micro/nanofibres from Eucalyptus kraft pulp: preparation and properties, Carbohydrate

polymers 89(1) (2012) 80-88.

[258] A. Abbasi Moud, M. Kamkar, A. Sanati-Nezhad, S.H. Hejazi, U. Sundararaj, Nonlinear viscoelastic

characterization of charged cellulose nanocrystal network structure in the presence of salt in aqueous

media, Cellulose (2020).

[259] C. Daniel, I.W. Hamley, M. Wilhelm, W. Mingvanish, Non-linear rheology of a face-centred cubic

phase in a diblock copolymer gel, Rheologica acta 40(1) (2001) 39-48.

[260] N. Phan-Thien, M. Newberry, R.I. Tanner, Non-linear oscillatory flow of a soft solid-like

viscoelastic material, Journal of non-newtonian fluid mechanics 92(1) (2000) 67-80.

[261] N. Phan-Thien, M. Safari-Ardi, Linear viscoelastic properties of flour–water doughs at different

water concentrations, Journal of Non-Newtonian Fluid Mechanics 74(1-3) (1998) 137-150.

[262] K. Hyun, S.H. Kim, K.H. Ahn, S.J. Lee, Large amplitude oscillatory shear as a way to classify the

complex fluids, Journal of Non-Newtonian Fluid Mechanics 107(1-3) (2002) 51-65.

[263] C. Bower, C. Gallegos, M. Mackley, J. Madiedo, The rheological and microstructural

characterisation of the non-linear flow behaviour of concentrated oil-in-water emulsions, Rheologica acta

38(2) (1999) 145-159.

[264] T. Mason, M.-D. Lacasse, G.S. Grest, D. Levine, J. Bibette, D. Weitz, Osmotic pressure and

viscoelastic shear moduli of concentrated emulsions, Physical Review E 56(3) (1997) 3150.

[265] F. Yziquel, P. Carreau, P. Tanguy, Non-linear viscoelastic behavior of fumed silica suspensions,

Rheologica acta 38(1) (1999) 14-25.

[266] S.R. Raghavan, S.A. Khan, Shear-thickening response of fumed silica suspensions under steady and

oscillatory shear, Journal of colloid and interface science 185(1) (1997) 57-67.

[267] M. Parthasarathy, D.J. Klingenberg, Large amplitude oscillatory shear of ER suspensions, Journal

of non-newtonian fluid mechanics 81(1-2) (1999) 83-104.

[268] M. Whittle, E. Dickinson, Large deformation rheological behaviour of a model particle gel, Journal

of the Chemical Society, Faraday Transactions 94(16) (1998) 2453-2462.

[269] U. Fumio, Y. Hiroshi, N. Kumiko, N. Sachihiko, S. Kenji, M. Yasunori, Swelling and mechanical

properties of poly (vinyl alcohol) hydrogels, International journal of pharmaceutics 58(2) (1990) 135-142.

[270] B. Ranjbar, H. Nazockdast, Shear flow‐induced orientation and structural recovery of multiwalled

carbon nanotube in poly (ethylene oxide) matrix, Journal of Applied Polymer Science 132(15) (2015).

[271] S.M. Dinh, R.C. Armstrong, A rheological equation of state for semiconcentrated fiber suspensions,

Journal of Rheology 28(3) (1984) 207-227.

[272] A. Gibson, S. Toll, Mechanics of the squeeze flow of planar fibre suspensions, Journal of non-

newtonian fluid mechanics 82(1) (1999) 1-24.

[273] S. Ranganathan, S. Advani, Fiber–fiber interactions in homogeneous flows of nondilute

suspensions, Journal of Rheology 35(8) (1991) 1499-1522.

[274] D.L. Koch, A model for orientational diffusion in fiber suspensions, Physics of Fluids 7(8) (1995)

2086-2088.

[275] F. Folgar, C.L. Tucker III, Orientation behavior of fibers in concentrated suspensions, Journal of

reinforced plastics and composites 3(2) (1984) 98-119.

197

[276] M. Doi, Molecular dynamics and rheological properties of concentrated solutions of rodlike

polymers in isotropic and liquid crystalline phases, Journal of Polymer Science: Polymer Physics Edition

19(2) (1981) 229-243.

[277] G.L. Hand, A theory of dilute suspensions, Archive for Rational Mechanics and Analysis 7(1)

(1961) 81-86.

[278] G. Lipscomb II, M.M. Denn, D. Hur, D.V. Boger, The flow of fiber suspensions in complex

geometries, Journal of Non-Newtonian Fluid Mechanics 26(3) (1988) 297-325.

[279] W. Letwimolnun, B. Vergnes, G. Ausias, P.J. Carreau, Stress overshoots of organoclay

nanocomposites in transient shear flow, Journal of non-newtonian fluid mechanics 141(2-3) (2007) 167-

179.

[280] M. Sepehr, P.J. Carreau, M. Moan, G. Ausias, Rheological properties of short fiber model

suspensions, Journal of Rheology 48(5) (2004) 1023-1048.

[281] M. Sepehr, G. Ausias, P.J. Carreau, Rheological properties of short fiber filled polypropylene in

transient shear flow, Journal of Non-Newtonian Fluid Mechanics 123(1) (2004) 19-32.

[282] M. Sepehr, P.J. Carreau, M. Grmela, G. Ausias, P. Lafleur, Comparison of rheological properties of

fiber suspensions with model predictions, Journal of polymer engineering 24(6) (2004) 579-610.

[283] H.M. Huynh, Improved fiber orientation predictions for injection-molded composites, University of

Illinois at Urbana-Champaign, 2001.

[284] Y. Zhang, L. Tao, S. Li, Y. Wei, Synthesis of multiresponsive and dynamic chitosan-based

hydrogels for controlled release of bioactive molecules, Biomacromolecules 12(8) (2011) 2894-2901.

[285] Q. Wang, J.L. Mynar, M. Yoshida, E. Lee, M. Lee, K. Okuro, K. Kinbara, T. Aida, High-water-

content mouldable hydrogels by mixing clay and a dendritic molecular binder, Nature 463(7279) (2010)

339-343.

[286] E. Moghimi, A.R. Jacob, N. Koumakis, G. Petekidis, Colloidal gels tuned by oscillatory shear, Soft

Matter 13(12) (2017) 2371-2383.

[287] P. Tomasik, Chemical and functional properties of food saccharides, CRC Press2003.

[288] P.B. Morgan, N. Efron, M. Helland, M. Itoi, D. Jones, J.J. Nichols, E. van der Worp, C.A. Woods,

Twenty first century trends in silicone hydrogel contact lens fitting: an international perspective, Contact

Lens and Anterior Eye 33(4) (2010) 196-198.

[289] H. Lodish, A. Berk, C.A. Kaiser, M. Krieger, M.P. Scott, A. Bretscher, H. Ploegh, P. Matsudaira,

Molecular cell biology, Macmillan2008.

[290] E.M. Ahmed, Hydrogel: Preparation, characterization, and applications: A review, Journal of

advanced research 6(2) (2015) 105-121.

[291] S. Manley, L. Cipelletti, V. Trappe, A. Bailey, R.J. Christianson, U. Gasser, V. Prasad, P. Segre, M.

Doherty, S. Sankaran, Limits to gelation in colloidal aggregation, Physical review letters 93(10) (2004)

108302.

[292] T.B. Blijdenstein, E. van der Linden, T. van Vliet, G.A. van Aken, Scaling behavior of delayed

demixing, rheology, and microstructure of emulsions flocculated by depletion and bridging, Langmuir

20(26) (2004) 11321-11328.

[293] G. Glasrud, R. Navarrete, L. Scriven, C. Macosko, Settling behaviors of iron oxide suspensions,

AIChE journal 39(4) (1993) 560-568.

[294] N.A. Verhaegh, D. Asnaghi, H.N. Lekkerkerker, Transient gels in colloid–polymer mixtures studied

with fluorescence confocal scanning laser microscopy, Physica A: Statistical Mechanics and its

Applications 264(1-2) (1999) 64-74.

[295] J.R. Weeks, J.S. van Duijneveldt, B. Vincent, Formation and collapse of gels of sterically stabilized

colloidal particles, Journal of Physics: Condensed Matter 12(46) (2000) 9599.

[296] R.M.L. Evans, L. Starrs, Emergence of a stress transmission length-scale in transient gels, Journal

of Physics: Condensed Matter 14(10) (2002) 2507.

[297] M.L. Kilfoil, E.E. Pashkovski, J.A. Masters, D. Weitz, Dynamics of weakly aggregated colloidal

particles, Philosophical Transactions of the Royal Society of London. Series A: Mathematical, Physical

and Engineering Sciences 361(1805) (2003) 753-766.

198

[298] D. Senis, C. Allain, Scaling analysis of sediment equilibrium in aggregated colloidal suspensions,


[299] M. Van Bruggen, H. Lekkerkerker, J. Dhont, Long-time translational self-diffusion in isotropic

dispersions of colloidal rods, Physical review E 56(4) (1997) 4394.

[300] S. Seiffert, W. Oppermann, Systematic evaluation of FRAP experiments performed in a confocal

laser scanning microscope, Journal of microscopy 220(1) (2005) 20-30.

[301] K. Braeckmans, L. Peeters, N.N. Sanders, S.C. De Smedt, J. Demeester, Three-dimensional

fluorescence recovery after photobleaching with the confocal scanning laser microscope, Biophysical

journal 85(4) (2003) 2240-2252.

[302] J. Karvinen, T.O. Ihalainen, M.T. Calejo, I. Jönkkäri, M. Kellomäki, Characterization of the

microstructure of hydrazone crosslinked polysaccharide-based hydrogels through rheological and

diffusion studies, Materials Science and Engineering: C 94 (2019) 1056-1066.

[303] D. Sustr, C. Duschl, D. Volodkin, A FRAP-based evaluation of protein diffusion in polyelectrolyte

multilayers, European Polymer Journal 68 (2015) 665-670.

[304] C. Picart, J. Mutterer, Y. Arntz, J.C. Voegel, P. Schaaf, B. Senger, Application of fluorescence

recovery after photobleaching to diffusion of a polyelectrolyte in a multilayer film, Microscopy research

and technique 66(1) (2005) 43-57.

[305] M.M. Pérez-Madrigal, J.E. Shaw, M.C. Arno, J.A. Hoyland, S.M. Richardson, A.P. Dove, Robust

alginate/hyaluronic acid thiol–yne click-hydrogel scaffolds with superior mechanical performance and

stability for load-bearing soft tissue engineering, Biomaterials science 8(1) (2020) 405-412.

[306] X. Yue, W. Li, Z. Li, F. Qiu, J. Pan, T. Zhang, Laminated superwetting aerogel/membrane

composite with large pore sizes for efficient separation of surfactant-stabilized water-in-oil emulsions,

Chemical Engineering Science 215 (2020) 115450.

[307] K. Li, Q. Luo, J. Xu, K. Li, W.-w. Zhang, L. Liu, J. Ma, H. Zhang, A Novel Freeze-drying-free

Strategy to Fabricate a Biobased Tough Aerogel for Separation of Oil/Water Mixtures, Journal of

Agricultural and Food Chemistry (2020).

[308] P. Pusey, Photon correlation and light beating spectroscopy, Plenum Press, New York (1974).

[309] H. Ehrlich, O.V. Kaluzhnaya, E. Brunner, M.V. Tsurkan, A. Ereskovsky, M. Ilan, K.R. Tabachnick,

V.V. Bazhenov, S. Paasch, M. Kammer, Identification and first insights into the structure and

biosynthesis of chitin from the freshwater sponge Spongilla lacustris, Journal of Structural Biology

183(3) (2013) 474-483.

[310] H. Ehrlich, M. Maldonado, K.d. Spindler, C. Eckert, T. Hanke, R. Born, C. Goebel, P. Simon, S.

Heinemann, H. Worch, First evidence of chitin as a component of the skeletal fibers of marine sponges.

Part I. Verongidae (Demospongia: Porifera), Journal of Experimental Zoology Part B: Molecular and

Developmental Evolution 308(4) (2007) 347-356.

[311] E. Brunner, H. Ehrlich, P. Schupp, R. Hedrich, S. Hunoldt, M. Kammer, S. Machill, S. Paasch, V.

Bazhenov, D. Kurek, Chitin-based scaffolds are an integral part of the skeleton of the marine demosponge

Ianthella basta, Journal of Structural Biology 168(3) (2009) 539-547.

[312] S. Harding, Protein Hydrodynamics. Protein: a comprehensive treatise, JAI Press, Greenwich, 1999.

[313] N. Sultanova, S. Kasarova, I. Nikolov, Dispersion proper ties of optical polymers, Acta Physica

Polonica-Series A General Physics 116(4) (2009) 585.

[314] M. Kang, C.A. Day, E. DiBenedetto, A.K. Kenworthy, A quantitative approach to analyze binding

diffusion kinetics by confocal FRAP, Biophysical journal 99(9) (2010) 2737-2747.

[315] M. Kasha, Collisional perturbation of spin‐orbital coupling and the mechanism of fluorescence

quenching. A visual demonstration of the perturbation, The Journal of Chemical Physics 20(1) (1952) 71-

74.

[316] B. Herman, Fluorescence microscopy, Current protocols in cell biology (1) (1998) 4.2. 1-4.2. 10.

[317] M. Carpineti, M. Giglio, Spinodal-type dynamics in fractal aggregation of colloidal clusters,


[318] R. Harich, T. Blythe, M. Hermes, E. Zaccarelli, A. Sederman, L.F. Gladden, W.C. Poon,

Gravitational collapse of depletion-induced colloidal gels, Soft Matter 12(19) (2016) 4300-4308.

199

[319] S. Manley, J. Skotheim, L. Mahadevan, D.A. Weitz, Gravitational collapse of colloidal gels,


[320] E. Salmon, R. Leslie, W. Saxton, M. Karow, J. McIntosh, Spindle microtubule dynamics in sea

urchin embryos: analysis using a fluorescein-labeled tubulin and measurements of fluorescence

redistribution after laser photobleaching, The Journal of cell biology 99(6) (1984) 2165-2174.

[321] J. Stricker, P. Maddox, E. Salmon, H.P. Erickson, Rapid assembly dynamics of the Escherichia coli

FtsZ-ring demonstrated by fluorescence recovery after photobleaching, Proceedings of the National

Academy of Sciences 99(5) (2002) 3171-3175.

[322] D.K. Cheerambathur, G. Civelekoglu-Scholey, I. Brust-Mascher, P. Sommi, A. Mogilner, J.M.

Scholey, Quantitative analysis of an anaphase B switch: predicted role for a microtubule catastrophe

gradient, The Journal of cell biology 177(6) (2007) 995-1004.

[323] P. Lajoie, E.A. Partridge, G. Guay, J.G. Goetz, J. Pawling, A. Lagana, B. Joshi, J.W. Dennis, I.R.

Nabi, Plasma membrane domain organization regulates EGFR signaling in tumor cells, The Journal of

cell biology 179(2) (2007) 341-356.

[324] D. Axelrod, D. Koppel, J. Schlessinger, E. Elson, W.W. Webb, Mobility measurement by analysis

of fluorescence photobleaching recovery kinetics, Biophysical journal 16(9) (1976) 1055.

[325] D. Soumpasis, Theoretical analysis of fluorescence photobleaching recovery experiments,

Biophysical journal 41(1) (1983) 95-97.

[326] F. Mueller, D. Mazza, T.J. Stasevich, J.G. McNally, FRAP and kinetic modeling in the analysis of

nuclear protein dynamics: what do we really know?, Current opinion in cell biology 22(3) (2010) 403-

411.

[327] J.K. Jonasson, N. Loren, P. Olofsson, M. Nyden, M. Rudemo, A pixel‐based likelihood framework

for analysis of fluorescence recovery after photobleaching data, Journal of microscopy 232(2) (2008) 260-

269.

[328] S. Edwards, K. Evans, Dynamics of highly entangled rod-like molecules, Journal of the Chemical

Society, Faraday Transactions 2: Molecular and Chemical Physics 78(1) (1982) 113-121.

[329] H. Löwen, Brownian dynamics of hard spherocylinders, Physical Review E 50(2) (1994) 1232.

[330] J. Lippincott-Schwartz, G.H. Patterson, Development and use of fluorescent protein markers in

living cells, Science 300(5616) (2003) 87-91.

[331] E. Haustein, P. Schwille, Fluorescence correlation spectroscopy: novel variations of an established

technique, Annu. Rev. Biophys. Biomol. Struct. 36 (2007) 151-169.

[332] D.L. Kolin, P.W. Wiseman, Advances in image correlation spectroscopy: measuring number

densities, aggregation states, and dynamics of fluorescently labeled macromolecules in cells, Cell

biochemistry and biophysics 49(3) (2007) 141-164.

[333] E. Toprak, P.R. Selvin, New fluorescent tools for watching nanometer-scale conformational

changes of single molecules, Annu. Rev. Biophys. Biomol. Struct. 36 (2007) 349-369.

[334] D. Alcor, G. Gouzer, A. Triller, Single‐particle tracking methods for the study of membrane

receptors dynamics, European Journal of Neuroscience 30(6) (2009) 987-997.

[335] A. Bancaud, S. Huet, G. Rabut, J. Ellenberg, Fluorescence perturbation techniques to study

mobility and molecular dynamics of proteins in live cells: FRAP, photoactivation, photoconversion, and

FLIP, Cold Spring Harbor Protocols 2010(12) (2010) pdb. top90.

[336] R.D. Phair, P. Scaffidi, C. Elbi, J. Vecerová, A. Dey, K. Ozato, D.T. Brown, G. Hager, M. Bustin,

T. Misteli, Global nature of dynamic protein-chromatin interactions in vivo: three-dimensional genome

scanning and dynamic interaction networks of chromatin proteins, Molecular and cellular biology 24(14)

(2004) 6393-6402.

gel development using cellulose nanocrystals

Documents