development and characterization of an extractive-based ... · pei-yu kuo doctor of philosophy...

Development and Characterization of an

Extractive-based Bio-Epoxy Resin from Beetle-Infested

Lodgepole Pine (Pinus contorta var. latifolia) Bark

By

Pei-Yu Kuo

A thesis submitted in conformity with the requirements for the degree of Doctor of Philosophy

Faculty of Forestry University of Toronto

© Copyright by Pei-Yu Kuo 2016

ii

Development and Characterization of an Extractive-based

Bio-Epoxy Resin from Beetle-Infested Lodgepole Pine

(Pinus contorta var. latifolia) Bark

Pei-Yu Kuo

Doctor of Philosophy

Faculty of Forestry

University of Toronto

2016

Abstract

Deriving chemicals from renewable feedstock has become a necessity to reduce dependency

on petroleum, which release carbon dioxide when burned and aggravate the global warming

and ocean acidification. This work offers a potential alternative - bark extractives based epoxy

resin - to petro-based conventional epoxy. Our results showed successful epoxidation of bark

extractives after reaction with epichlorohydrin. The newly synthesized epoxy (E-epoxy) can

replace 50% of petroleum-based epoxy (P-epoxy) and the blend system displayed thermal

stability and tensile strength comparable to neat P-epoxy, which demonstrates a great promise

in using bark extractives as a substitute for BPA.

An examination of reaction parameters showed that the E-epoxy monomer can be synthesized

with high yield and reactivity using spray-dried extractives as substrates, a dioxane/water

combination as solvent, and tetrabutylammonium hydroxide as the ring-opening catalyst. An

examination of numerical parameters showed the maximum yield with minimum epoxy

equivalent weight was achieved after 4.5 hours reaction time with sodium hydroxide to

hydroxyl value molar ratio of 3.4 at a reaction temperature of 80 °C. The thermal properties of

E-epoxy were studied using TGA, FTIR, and Py-GC/MS, and a new thermal degradation

mechanism was proposed.

Additionally, nanocellulose fibres (NCFs) were incorporated to enhance E-epoxy’s mechanical

iii

performance. Based on an adjusted curing schedule, an E-epoxy/P-epoxy/NCFs composite

with high strength, ductility, thermal stability, and sustainability was developed. With 10% E-

epoxy, the toughness of neat epoxy resins improved 84 %; after incorporating NCFs, the tensile

strength and modulus of composites increased approximately two- and four-fold, respectively.

The maximum degradation peak of the composites was 24 °C higher than for neat epoxy resins.

Overall, bark extractives exhibit great promise to replace petro-based BPA; incorporating

NCFs into E-epoxy/P-epoxy blending system is an effective method to develop a strong and

sustainable bio-nanocomposite.

iv

Acknowledgement

I would like to express my sincere gratitude to my supervisors, Prof. Sain and Prof. Yan, for

their continuous efforts on mentoring me. Without their encouragement, support and guidance,

I would never have overcome many difficulties and completed my PhD. During the past five

years, they led me towards exciting research fields and showed me how to be a successful and

independent scientist. It has been my honour and privilege to work with both of them with their

trust and generosity.

I am also extremely indebted to my thesis committee members, Prof. Copper and Prof. Naguib,

and my thesis examiners, Prof. Dorgan and Prof. Singh, for the stimulating discussions and

their commitment to my professional development.

I would like to thank Andrew, Lindsey, and Mariam for many invaluable discussions on

editorial and scientific questions. I also want to acknowledge the helpful comments made by

Sankar, Shokouh, Nikhil, Javad and Zeen. Dmitry and George, were extremely helpful with

some technical aspects as well.

Particular thanks extend to my friends – Justin, Linghong and Rosanna for their kind

suggestions on editing. I am grateful to have had help from Robert, Shawn, Rujun, Lynn,

Luizmar, Crystal, Prashant, Sharon, Jieming, Rana, YiChun, Miroslava, Wendy, Julie,

Stephanie, Lukas, Julieta and colleagues from labs ES 2008 and ES 3004, for generously

sharing their time, discussions, and their enthusiasms with me.

Finally, I would like to thank my parents and my sisters. Their unconditional love and support

helped me through the toughest times. Without your support, I wouldn’t even be close to

finishing this journey. Thank you.

v

Table of Contents

Acknowledgement .............................................................................................................. iv

Table of Contents ................................................................................................................. v

List of Tables ...................................................................................................................... xi

List of Figures ................................................................................................................... xiii

List of Schemes ................................................................................................................. xvi

List of Abbreviations ...................................................................................................... xviii

Chapter 1 Introduction ......................................................................................................... 1

1.1 Motivation and Significance ..................................................................................... 1

1.2 Objectives .................................................................................................................. 3

1.3 Hypothesis ................................................................................................................. 3

1.4 Scope ......................................................................................................................... 3

Chapter 2 Literature Review ................................................................................................ 8

2.1 Conventional Epoxy Resins ...................................................................................... 8

2.1.1 Common Types of Epoxy Monomers and Curing Agents .............................. 8

2.1.2 Epoxy Monomer Synthesis ............................................................................. 9

2.1.3 Network Forming Mechanism ...................................................................... 11

2.2 Common Bio-based Epoxy Resins Derived from Renewable Resources ............... 15

2.2.1 Oil-based Epoxy Resins ................................................................................ 15

2.2.2 Saccharide-based Epoxy Resins ................................................................... 16

2.2.3 Lignin-based Epoxy Resins .......................................................................... 18

2.2.4 Terpene and Resin Acid-based Epoxy Resins .............................................. 18

2.2.5 Polyphenol-based Epoxy Resins ................................................................... 20

2.2.6 Wood-based Epoxy Resins ........................................................................... 21

2.3 Bark Extractives and Extraction Process ................................................................ 22

2.3.1 Chemical Composition of Pine Bark ............................................................ 22

vi

2.3.2 Extraction Technology .................................................................................. 23

2.4 Reinforced Bio-epoxy Resins ................................................................................. 26

2.4.1 Types of Reinforcements .............................................................................. 26

2.4.2 Cellulose Nanostructure ................................................................................ 27

2.4.3 Past and Current Research of Reinforced Petroleum-based Epoxy Resins

with Cellulose Nanostructure ..................................................................... 28

Chapter 3 Characterization of Bark Extractive-based Bio-epoxy Resins .......................... 35

3.1 Introduction ............................................................................................................. 36

3.2 Materials and Methods ............................................................................................ 37

3.2.1 Materials ....................................................................................................... 37

3.2.2 Methods ......................................................................................................... 38

3.2.2.1 Bark Extraction .............................................................................. 38

3.2.2.2 Synthesis of Bark Extractive Epoxy Resins ................................... 38

3.2.2.3 Curing of the Epoxy Monomers .................................................... 38

3.2.3 Characterization ............................................................................................ 39

3.2.3.1 Epoxy Equivalent Weight (EEW) .................................................. 39

3.2.3.2 Differential Scanning Calorimetry (DSC) ..................................... 39

3.2.3.3 Fourier Transform Infrared Spectroscopy (FTIR) ......................... 39

3.2.3.4 Thermogravimetric Analysis (TGA) .............................................. 39

3.2.3.5 Gel Permeation Chromatography (GPC) ....................................... 39

3.2.3.6 Nuclear Magnetic Resonance (NMR) ............................................ 40

3.2.3.7 Mechanical Universal Testing Machine ........................................ 40

3.3 Results and Discussion ............................................................................................ 40

3.3.1 Yield and Epoxy Equivalent Weight (EEW) ................................................ 40

3.3.2 Spectral Characterization of Bio-epoxy Resins ............................................ 41

3.3.3 Curing Behaviour .......................................................................................... 46

3.3.4 Mechanical Performance .............................................................................. 50

vii

3.3.5 Thermal Degradation and Thermal Stability ................................................ 51

3.4 Summary ................................................................................................................. 54

Chapter 4 Effects of Reaction Parameters on the Glycidyl Etherification of Bark

Extractives during Bio-epoxy Resin Synthesis ............................................................. 58

4.1 Introduction ............................................................................................................. 59


4.2.1 Materials ....................................................................................................... 61

4.2.2 Methods ......................................................................................................... 61

4.2.2.1 Extraction Procedure ...................................................................... 61

4.2.2.2 Synthesis Procedure ....................................................................... 61

4.2.2.3 Refining Procedure ........................................................................ 62


4.2.3.1 Hydroxyl Value (OHV) Determination ......................................... 62

4.2.3.2 Molecular Weight .......................................................................... 63

4.2.3.3 Epoxy Equivalent Weight (EEW) .................................................. 64

4.2.3.4 Structural Characterization ............................................................ 64

4.2.4 Experimental Design for Response Surface Methodology (RSM) ............... 65


4.3.1 Categorical Reaction Conditions .................................................................. 65

4.3.1.1 Effect of Substrates ........................................................................ 65

4.3.1.2 Effect of Solvents ........................................................................... 69

4.3.1.3 Effect of Catalysts .......................................................................... 72

4.3.2 Numerical Variables — Reaction Time, Reaction Temperature, and

NaOH/OHV Molar Ratio ........................................................................... 73

4.4 Summary ................................................................................................................. 79

Chapter 5 Thermal Degradation of Extractive-Based Bio-Epoxy Monomer and Network:

Kinetics and Mechanism ............................................................................................... 83

5.1 Introduction ............................................................................................................. 84

viii


5.2.1 Materials ....................................................................................................... 85

5.2.2 Extraction Procedure ..................................................................................... 85

5.2.3 Synthesis Procedure ...................................................................................... 85

5.2.4 Refining Procedure ....................................................................................... 85


5.2.5.1 Thermogravimetric Analysis (TGA) .............................................. 86

5.2.5.2 Fourier Transform Infrared Spectroscopy (FTIR) ......................... 86

5.2.5.3 Pyrolysis-Gas Chromatography/Mass Spectrometry (Py-GC/MS) 86


5.3.1 Thermal Stability of Monomers .................................................................... 87

5.3.1.1 Mass Loss of Monomers ................................................................ 87

5.3.1.2 Degradation Kinetics of Monomers ............................................... 89

5.3.1.3 FTIR Residual Analysis of Monomers after Isothermal Thermal

Degradation ................................................................................. 91

5.3.1.4 Py – GC/MS Analysis of Bark Extractives and Uncured E-epoxy

monomer ...................................................................................... 94

5.3.1.5 Proposed Degradation Mechanism for Uncured E-epoxy

monomer ...................................................................................... 97

5.3.2 Thermal Stability of the Cured Epoxy Networks ....................................... 101

5.3.2.1 Mass Loss and Derivative Mass Loss of the Cured Resins ......... 101

5.3.2.2 Degradation Kinetics of the Cured Networks .............................. 103

5.3.2.3 FTIR Residual Analysis of the Cured Resins after Isothermal

Thermal Degradation ................................................................. 104

5.3.2.4 Py-GC/MS Analysis of the Cured E-epoxy Resins ..................... 106

5.3.2.5 Degradation Mechanism .............................................................. 108

5.4 Summary ............................................................................................................... 112

Chapter 6 Influence of Nanocellulose Fibres (NCFs) on the Curing Behaviours of Epoxy

Resins .......................................................................................................................... 118

ix

6.1 Introduction ........................................................................................................... 119

6.2 Materials and Methods .......................................................................................... 120

6.2.1 Materials ..................................................................................................... 120

6.2.2 Methods ....................................................................................................... 121

6.2.3 Characterization .......................................................................................... 121

6.2.3.1 Differential Scanning Calorimetry (DSC) ................................... 121

6.2.3.2 Fourier Transform Infrared Spectroscopy (FTIR) ....................... 121

6.3 Results and Discussion .......................................................................................... 122

6.3.1 Curing Behaviour ........................................................................................ 122

6.3.2 Cure Kinetics .............................................................................................. 125

6.3.2.1 Model Selection ........................................................................... 125

6.3.2.2 Isothermal Versus Constant Heating Rate Run ........................... 127

6.3.2.3 The Kissinger Method .................................................................. 127

6.3.2.4 Model Free Method - Kissinger-Akahira-Sunose Method .......... 129

6.3.2.5 Model Fitting Method - Sestak-Berggren Model and Kamal

Model ......................................................................................... 130

6.3.2.6 Prediction of Conversion ............................................................. 132

6.4 Summary ............................................................................................................... 134

Chapter 7 Using Nanocellulose Fibres to Develop a High Performance Blended Petro-

Epoxy/Bio-Epoxy Composite ..................................................................................... 138

7.1 Introduction ........................................................................................................... 139

7.2 Materials and Methods .......................................................................................... 140

7.2.1 Materials ..................................................................................................... 140

7.2.2 Methods ....................................................................................................... 140

7.2.2.1 NCFs Paper Preparation ............................................................... 140

7.2.2.2 Composite Preparation ................................................................. 141

7.2.3 Characterization .......................................................................................... 141

7.2.3.1 Activation Energy (Ea) and Glass Transition Temperature (Tg) .. 141

x

7.2.3.2 Tensile Mechanical Properties ..................................................... 142

7.2.3.3 Morphology .................................................................................. 142

7.2.3.4 Thermal Stability ......................................................................... 142

7.3. Results and Discussion ......................................................................................... 142

7.3.1 The Effect of Nanocellulose Fibres on Activation Energies of Epoxy

Resins ....................................................................................................... 142

7.3.2 Mechanical Performance of Epoxy Resins and Epoxy/NCFs composites . 144

7.3.3 Morphological Characterization of NCFs Reinforced Epoxy Resins ......... 148

7.3.4 Thermal Stability ........................................................................................ 151

7.4 Summary ............................................................................................................... 153

Chapter 8 Conclusions, Contributions and Future Work ................................................. 157

8.1 Conclusions ........................................................................................................... 157

8.2 Scientific Contributions and Engineering Prospects ............................................. 158

8.3 Future Work .......................................................................................................... 159

Appendix A (Chapter 3) ................................................................................................... 161

Appendix B (Chapter 4) ................................................................................................... 163

Appendix C (Chapter 6) ................................................................................................... 164

xi

List of Tables

Table 2-1 Typical properties of some epoxy resins ............................................................. 9

Table 2-2 Advantages, disadvantages, and applications of common epoxy curing agents . 9

Table 2-3 Yield and major components of lodgepole pine bark extractives. .................... 24

Table 2-4 Extraction amounts of various extraction solvents and tree barks .................... 26

Table 3-1 Molecular weight and polydispersity of E-epoxy and L-epoxy ........................ 45

Table 3-2 TGA data of the cured epoxy resins .................................................................. 53

Table 4-1 Variables and their levels for Box–Behnken design. ........................................ 65

Table 4-2 OHV and molecular weight of two types of extractives. .................................. 66

Table 4-3 OHV from bark extractives using 31P-NMR. .................................................... 68

Table 4-4 Yield and EEW value of two types of E-epoxy monomers. .............................. 68

Table 4-5 Chemical and physical properties of common solvents for epoxy monomer

synthesis. ....................................................................................................................... 70

Table 4-7 Yield and EEW value of two types of epoxy monomers through two synthetic

paths. ............................................................................................................................. 73

Table 4-8 The B-B matrix and output responses. .............................................................. 75

Table 4-9 ANOVA results of the quadratic model for the yield of reaction. .................... 76

Table 4-10 ANOVA results of the quadratic model for the EEW value. .......................... 76

Table 4-11 Verification of the proposed optimal synthesis conditions. ............................ 78

Table 5-1 TGA data of the bio-based epoxy monomers. ................................................... 89

Table 5-2 Peak assignments of the FTIR spectra of monomers. ....................................... 93

Table 5-4 Peak assignments of the pyrolysis products of uncured bio-epoxy monomer as

detected by Py-GC-MS. ................................................................................................ 97

Table 5-5 TGA data of the cured epoxy resins. ............................................................... 103

Table 5-6 Peak assignments of the FTIR spectra of cured resins. ................................... 106

Table 5-7 Peak assignments of the cured bio-epoxy resins pyrolysis products detected by

Py-GC-MS................................................................................................................... 108

Table 6-1 Composition of samples .................................................................................. 121

xii

Table 6-2 Thermal characteristics of samples from the dynamic DSC analysis ............. 124

Table 6-3 Cure kinetics parameters by Kissinger Eq. and Sestak-Berggren empirical

Eq. ............................................................................................................................... 129

Table 6-4 Cure kinetics parameters by Kamal Equation ................................................. 132

Table 7-1 Sample compositions ....................................................................................... 141

Table 7-2 Tensile properties of neat epoxies ................................................................... 144

Table 7-3 Tensile strength and modulus of two types of reinforcement and two types of

resin compositions ....................................................................................................... 145

Table 7-4 Tensile strength and modulus of various E-epoxy replacements .................... 148

xiii

List of Figures

Fig. 2-1 TEM image of (a) nanocellulose whiskers and (b) nanocellulose fibres. ............ 28

Fig. 3-1 Effect of reaction temperature (a) and catalyst amount (b) on the product yield

and EEW value .............................................................................................................. 41

Fig. 3-2 FTIR spectra of (a) bark extractives, epoxidized bark extractives and commercial

epoxy resin, (b) bio-epoxy resins .................................................................................. 42

Fig. 3-3 Liquid state NMR spectrum of the E-epoxy monomers ...................................... 44

Fig. 3-4 Liquid state NMR spectrum of the bark extractives ............................................ 44

Fig. 3-5 Liquid state NMR spectrum of the L-epoxy monomers ...................................... 44

Fig. 3-6 Liquid state NMR spectrum of the C-epoxy monomers ...................................... 45

Fig. 3-7 GPC traces of uncured bio-epoxy resins .............................................................. 46

Fig. 3-8 Dependence of the activation energy on the extent of conversion evaluated from

non-isothermal DSC data .............................................................................................. 49

Fig. 3-9 Model-free prediction of isothermal cure from 140 to 170 °C using KAS method

(The experimental data are shown by line. The points correspond to the predicted

time) .............................................................................................................................. 49

Fig. 3-10 Evolution of oxirane functional group during curing......................................... 50

Fig. 3-11 Tensile strength and modulus of the neat P-epoxy, 10, 30, and 50% E-epoxy

replacement of P-epoxy ................................................................................................. 51

Fig. 3-12 Thermal degradation of the cured epoxy resins with various E-epoxy

replacements (0-50%) under air .................................................................................... 52

Fig. 3-13 Thermal stability of the cured P-epoxy, E-epoxy and L-epoxy under nitrogen

atmosphere .................................................................................................................... 53

Fig. 4-1 SEC traces of bark extractives. ............................................................................ 67

Fig. 4-2 31P-NMR analysis of (a) ODE and (b) SDE. ........................................................ 68

Fig. 4-3 13C-NMR spectra of E-epoxy monomers synthesized in various solvent

systems. ......................................................................................................................... 71

Fig. 4-4 Effect of catalyst types on functional groups of E-epoxy products observing by

FTIR (left) and NMR (right): (a) TBAH/NaOH and (b) twofold NaOH. ..................... 73

Fig. 4-5 Normal probability plot of (a) yield and (b) EEW value of E-epoxy monomer. . 76

xiv

Fig. 4-6 Response surface plots of various parameters on the product yield and reactivity :

(a) Effects of time and temperature, (b) Effects of temperature and NaOH/OHV ratio,

and (c) Effects of time and NaOH/OHV ratio. ............................................................. 78

Fig. 5-1 Thermal degradation of the P-epoxy monomer and E-epoxy monomer under

nitrogen environment by TGA, showing the (a) mass curves and (b) derivative of the

mass loss curves. ........................................................................................................... 87

Fig. 5-2 Deconvolution of the multi-peaks of the E-epoxy monomer DTGA results. ...... 88

Fig. 5-3 DTGA curves of the two model bio-based epoxy monomers: (a) lignin-based

monomer and (b) abietic acid-based monomer. ............................................................ 89

Fig. 5-4 Dependence of the Ea on the extent of conversion evaluated from non-isothermal

TGA data. ...................................................................................................................... 91

Fig. 5-5 FTIR spectra of the commercial monomer and bio-epoxy monomer after TGA in

nitrogen or air atmospheres at various temperatures: (a) P-epoxy under nitrogen, (b) P-

epoxy under air, (c) E-epoxy under nitrogen, (d) E-epoxy under air. ........................... 93

Fig. 5-6 Py-GC/MS chromatogram of bark extractives and uncured bio-epoxy

monomer ....................................................................................................................... 95

Fig.5-7 Thermal degradation of the P-epoxy resin and E-epoxy resin under nitrogen: (a)

mass loss curves and (b) derivative mass loss curves. ................................................ 102


TGA data. .................................................................................................................... 104

Fig. 5-9 FTIR spectra of the cured P-epoxy and cured E-epoxy after TGA in nitrogen or

air atomosphere at various temperatures: (a) P-epoxy under nitrogen, (b) P-epoxy

under air, (c) E-epoxy under nitrogen, (d) E-epoxy under air. .................................... 105

Fig. 5-10 Py-GC/MS chromatograms of the cured bio-epoxy resins. ............................. 108

Fig. 6-1 FTIR spectra obtained at different times for BiF-Aliph sample cured at 100°C.

(a) Spectrum from 4000-3000 cm-1, (b) Spectrum from 1000-500 cm-1 ..................... 123

Fig. 6-2 Model selection by (a) reaction rate and (b) Malek’s method ........................... 126

Fig. 6-3 Onset of diffusion control, w= 𝑙𝑛(𝑑𝛼/𝑑𝑡)/1 − 𝛼𝑛 − 𝑘1 .................................. 127


non-isothermal DSC data ............................................................................................ 130

Fig. 6-5 Model-free prediction of isothermal cure at 140°C and 170 °C from KAS

method. (The experimental data are shown by line. The points correspond to the

predicted time.) ........................................................................................................... 133

Fig. 7-1 Dependence of the 𝐸 𝑎 on the extent of conversion evaluated from non-

isothermal DSC data. .................................................................................................. 143

xv

Fig. 7-2 SEM images of cellulose nano-paper before oven-drying (a) and after oven-

drying (b) ..................................................................................................................... 145

Fig. 7-3 Tensile stress-strain curves of (a) P-epoxy and (b) 10%E-epoxy and their

reinforced composites ................................................................................................. 146

Fig. 7-4 Tensile properties of NCFs reinforced composites compared to literature data 147

Fig. 7-5 Tensile properties of NCFs reinforced composites compared to theoretical

data .............................................................................................................................. 148

Fig. 7-6 SEM images of fracture surface in cross-sections. P-epoxy/DNCF(a) and P-

epoxy/SNCF (b) showing the difference in nanopaper thickness caused by resin

penetration. A clear boundary was observed between DNCF and epoxy (c), but not at

SNCF/epoxy (d) .......................................................................................................... 149

Fig. 7-7 SEM images of fracture surface in cross-sections of P-Epoxy/SNCF (a) and

10%E-Epoxy/SNCF(b) showing that the diameter of nanocellulose fibres (white spots)

in 10%E-epoxy blending system is smaller compared to neat P-epoxy. .................... 150

Fig. 7-8 Thermal degradation of P-epoxy and P-epoxy/NCFs with various E-epoxy

replacement (a)(b) in nitrogen environment (c)(d) in air environment ....................... 152

xvi

List of Schemes

Scheme 1-1 Thesis overview and organization of chapters ................................................. 6

Scheme 2-1 Bisphenol A diglycidyl ether (BADGE) .......................................................... 8

Scheme 2-2 Mechanism of glycidyl etherification between phenolic compounds and

ECH ............................................................................................................................... 10

Scheme 2-3 Addition reaction of an epoxy resins with amine type curing agent .............. 12

Scheme 2-4 (a) Epoxidized palm oil, (b) epoxidized linseed oil ....................................... 15

Scheme 2-5 Synthesis of diglycidyl ether of isosorbide by various methods .................... 17

Scheme 2-6 A new synthesis method for epoxy monomer, using lactide as an example .. 17

Scheme 2-7 Three synthesis routes for vanillin-based epoxy resins ................................. 18

Scheme 2-8 Synthesis of epoxy monomer from abietic acid ............................................. 19

Scheme 2-9 Epoxidized imidodicarboxylic resinic of maleic anhydride (a) and

epoxidized ..................................................................................................................... 20

levopimaric acid (b) ........................................................................................................... 20

Scheme 2-10 Major composition of pine bark ................................................................... 23

Scheme 2-11 Major extractives based on the literature ..................................................... 25


ECH ............................................................................................................................... 61

Scheme 4-2 Mechanism of glycidyl etherification catalyzed by tetrabutylammonium

hydroxide (TBAH) ........................................................................................................ 72

Scheme 5-1 Three dominant thermal degradation mechanisms of epoxy monomer. ........ 99

Scheme 5-2 Proposed degradation mechanism of E-epoxy resins (The solid coloured

rounded rectangular with the compound numbers refer to the results from Py-GC/MS;

the yellow highlight circles refer to the results from FTIR and the rest parts are

supported by literature). .............................................................................................. 101

Scheme 5-3 Thermoxidation mechanism of epoxy monomer. ........................................ 101

Scheme 5-4 Three scission positions on vinylene ethers. ................................................ 109

Scheme 5-5 Three possible scission reactions on cured epoxy resins. ............................ 110

Scheme 5-6 Thermal oxidative mechanism based on the free radical mechanism. ......... 111

Scheme 5-7 Cope reaction proposed by Burton and Conley ........................................... 111

xvii

Scheme 6-1 Trimolecular transition state suggested by Smith. ....................................... 125

Scheme 7-1 Reaction between epoxy and hydroxyl groups: (a) catalyst effect and (b)

etherification. .............................................................................................................. 144

xviii

List of Abbreviations

A-epoxy Abietic acid based bio-epoxy

BADGE Bisphenol A diglycidyl ether

BFDGE Bisphenol F diglycidyl ether

BPA Bisphenol A

C-epoxy Cellulose based bio-epoxy

DCM Dichloromethane

DNCF Oven-dried nanocellulose fibre

DMBD Diethyl-methyl-benzene-diamine

DSC Differential scanning calorimetry

Ea Activation energy

ECH Epichlorohydrin

E-epoxy Bark extractive based bio-epoxy

EEW Epoxy equivalent weight

EHS Environment, health and safety

FTIR Fourier transform infrared spectroscopy

GPC Gel permeation chromatography

HTM Halpin-Tsai model

KAS Kissinger-Akahira-Sunose model

L-epoxy Lignin based bio-epoxy

MeOH Methanol

Mn Number average molecular weight

Mw Weight average molecular weight

NCFs Nanocellulose fibres

NMR Nuclear magnetic resonance

ODE Oven-dried bark extractives

OHV Hydroxyl value

PDI Polydispersity index

P-epoxy Petroleum based epoxy

PTC Phase transfer catalyst

Py-GC/MS Pyrolysis-gas chromatography-mass spectrometry

ROM Rule of mixture

RSM Response surface methodology

S-B Sestak-Berggren model

SNCF Solvent-exchanged nanocellulose fibre

SDE Spray-dried extractives

SEM Scanning electron microscope

SPI Solvent polarity index

xix

TBAH Tetrabutylammonium hydroxide

Tg Glass transition temperature

TGA Thermogravimetric analysis

Ts Statistic heat-resistant index temperature

1

Chapter 1 Introduction

1.1 Motivation and Significance

Epoxy resins are a class of high-performance thermosetting precursors, containing oxirane

groups that can be cured by a diversity of co-agents, such as amines and anhydrides, to form

three-dimensional networks. The cured epoxy resins, also called epoxies, hold a dominant

position in the polymer market - especially in the automotive, construction, electronic and

aerospace industries - due to their high strength, superior chemical-resistance, good

compatibility with other materials, and minimum shrinkage and volatility of by-products after

curing. The global demand of epoxy resin was estimated at roughly US$18.6 billion in 2013

and was forecasted to reach US$25.8 billion by 20181. This increase in value reflects the

increasing demand of epoxy resins in the global market over the coming years.

Additionally, the major component of epoxy resins, bisphenol A (BPA), is considered a toxic

substance, and is included in Schedule 1 of the Canadian Environmental Protection Act2 for

declaring the necessity to substitute BPA. In conjunction with the scarcity of crude oil,

replacing BPA with natural resources is evident, and thus many natural resources have been

explored to determine their feasibilities as alternatives. Vegetable oil3-5, liquefied wood6-7 and

lignin8-9 are considered as the most promising materials with which to synthesize bio-based

epoxy resins. Although epoxidized vegetable oils are already commercialized and commonly

used as plasticizers, the structure of vegetable oil, containing only aliphatic chains, limits its

mechanical strength and thermal stability10. Moreover, liquefied biomass contains more

aromatic structures compared to epoxidized oil, but the liquefaction process is energy-intensive

and uses toxic solvents like phenol11. Furthermore, lignin, a by-product of the paper industry,

has abundant aromatic structures, but its viscosity and the molecular weight create difficulties

during the manufacturing process. Therefore, finding a suitable natural resource to produce

bio-epoxy resins remains one of the most important areas of interest.

To address this issue, bark extractives have emerged as promising candidates to produce a new

type of bio-epoxy resins. Some of the major benefits of using tree bark include its abundance,

renewability, and its richness in phenolic compounds. The annual bark yield in Canada12 is as

much as 17 million m3; however, most tree bark is under-used, and is either left in the stands

after harvest or burned as low-efficiency fuels. Furthermore, the recent ongoing epidemic of

pine beetle infested trees provides an abundant resource to investigate the utility of tree bark.

2

Since early 1990, mountain pine beetle (MPB; Dendroctonus ponderosae) has destroyed more

than 19 million hectares of forest in B.C.13 and 36 million hectares of that in USA14. In attempts

to salvage these dead trees, the government of Canada has invested over $100 million to

develop new bio-derived products and the U. S. Forest Service has spent $30 million in

cellulosic ethanol plant, which converts beetle-infested biomass into bio-ethanol. In addition,

many technologies and researches are being applied and studied in order to find a suitable

solution to convert the waste bark into useful products and one of most promising application

is to convert bark into valuable chemicals.

Compared to wood, bark has more phenolic compounds such as tannins, flavonoids and lignin

oligomers, which can be developed into various polymers, including novolac15, resol16 type

phenol formaldehyde resins, and polyurethane17. Among all of the extraction processes,

alkaline extraction is able to dissolve most of the phenolic compounds18 and the remaining

sodium hydroxide can be used as catalyst to promote the synthesis procedure. For these reasons,

the potential of bark extractives should be exploited as a replacement for BPA. However, to

the best of our knowledge, few researches have explored the feasibility of developing bio-

epoxy resins from bark extractives.

The most common approach to synthesize epoxy monomers is to etherify the phenolic hydroxyl

groups using epichlorohydrin (ECH). The mechanism involves using the hydroxyl groups as

nucleophiles to attack carbons on ECH. Compared to other methods, such as double bond

oxidation, the ECH method is more cost-effective, industrially-preferred, and has a simpler

purification process19. Although many bio-epoxy resins have been produced through the ECH

procedure, bark extractives contain complex and diverse compounds resulting in irregular

reactivity. A suitable synthesis procedure must be formulated in order to ensure that bark

extractive-based epoxy (E-epoxy) monomer has a high yield and reactivity.

A potential bio-epoxy not only requires high yield and reactivity of monomer, but also demands

good mechanical and thermal properties of cured resin. As the literature indicates, many bio-

epoxy resins have been successfully synthesized, but their performance is often insufficient for

the high-performance material industry. In order to broaden applications of bio-epoxy resins,

this thesis presents an in-depth investigation into the thermal and mechanical properties, which

may provide some guidance for producing a bio-epoxy resin with superior quality.

3

1.2 Objectives

The overall objective of this study is to better understand the performance of extractive-based

epoxy resin (E-epoxy), and its feasibilities to replace petroleum-based epoxy resins (P-epoxy).

Each chapter of this present study was designed to reach the following sub-objectives:

1. Examining the feasibility of epoxidation reaction between bark extractives and ECH,

which leads to the development of an innovative bio-epoxy;

2. Studying the effect of synthesis conditions on the yield, reactivity and molecular

structures of E-epoxy;

3. Understanding the thermal stability and degradation kinetics of E-epoxy compared with

other types of epoxy resins;

4. Investigating the influence of NCFs on the curing behaviour of P-epoxy; and

5. Assessing the effect of NCFs on mechanical strength of E-epoxy/P-epoxy blended

system.

1.3 Hypothesis

The overall hypothesis of this study is that alkaline bark extractives contain a sufficient amount

of polyphenols to react with ECH, and produce novel bio-epoxy monomers.

1.4 Scope

This thesis outlines the characterization of a bio-epoxy resin produced from tree bark

extractives. Chapter 1 discusses the necessity to find an alternative to the toxic BPA and to

decrease the cost of BPA raw materials. In conjunction with the scarcity of petroleum resource,

it is evident there is a need to develop a new E-epoxy resin from natural materials. Specifically,

tree bark, a waste material from the lumber industry, is used in this study to produce low-

molecular-weight polyphenols as BPA substitutes. The goal of the thesis is to provide

information on deriving a bio-based epoxy resin from tree bark and characterize the newly

synthesized E-epoxy.

Chapter 2 provides a summary of the literature comparing the synthesis pathways of petro-

and bio-based epoxy resins, chemical compositions of bark, extraction technologies, and

reinforced bio-epoxy composites. In order to develop a bio-based material, it is essential to

4

study all the synthesis routes and understand the effects of synthesis conditions on the

properties of the final product. Furthermore, bark extractives have a variety of chemical

components including polysaccharides, degraded lignin, and terpenes, which show different

reaction rates toward ECH reagent. Thus, in order to predict the possible synthesis products, it

is critical to identify the chemical composition of bark extractives. Finally, one of the remaining

challenges from the literature is the insufficient mechanical strength of bio-based epoxy resin.

To address this issue, NCFs are applied as reinforcement to analyse the strengthening effect on

mechanical properties.

Throughout the following five chapters (Chapter 3-7), several aspects of E-epoxy resins were

analyzed. These chapters were written as self-contained manuscripts including Introduction

and Methods sections for each with some overlapping content. However, efforts were made to

minimize these reiterations.

Chapter 3 tests the feasibility of developing bio-epoxy resins from bark extractives. The

hypothesis of this study is that the polyphenols in bark extractives can react with

epichlorohydrin through a conventional glycidyl etherification. The first finding is that the

oxirane groups were successfully grafted on the bark extractives as confirmed by Fourier

transform infrared spectroscopy (FTIR) and nuclear magnetic resonance (NMR). In addition,

the molecular weight of the newly synthesized E-epoxy resin was significantly lower than that

of lignin-based epoxy resin. Lastly, the thermal and mechanical performances of the cured

blend system (E-epoxy/P-epoxy) were slightly lower than those of the petroleum-based epoxy

resins. The results in this chapter provide the basis of E-epoxy resin and indicate that more

work is required on understanding and improving the thermal and mechanical properties.

Chapter 4 investigates the effects of synthesis conditions on the yield and epoxy equivalent

weight (EEW) of the E-epoxy resin. According to the reaction mechanism, six parameters

(substrates, solvents, catalyst, reaction time, reaction temperature and catalyst amount) are

assumed as influential factors to cause significant differences on the production of E-epoxy

resin. The best combination of substrate/solvent/catalyst was spray-dried extractives/dioxane

with water/ phase transfer catalyst. The maximum yield with minimum epoxy equivalent

weight was achieved after 4.5 hours reaction time with sodium hydroxide to hydroxyl value

molar ratio of 3.4 at 80 °C reaction temperature. These results provide information that will

assist in establishing a synthesis protocol for E-epoxy resin, and also provide the guidance for

future scale-up production.

5

Chapter 5 further investigates the thermal stability of E-epoxy, and proposes a new thermal

degradation mechanism. As reported in Chapter 3, blending E-epoxy into petro-based epoxy

resin has a detrimental effect, which instigated us to explore the backbone structures of E-

epoxy. The major component identified by pyrolysis-gas chromatography-mass spectrometry

(Py-GC/MS) of E-epoxy resin was abietic acid, which has a mono-functionalized diterpene

structure. Other mono-functionalized compounds in E-epoxy were epimanool and steric acid,

which can decrease the crosslinking density and reduce the thermal stability due to their

aliphatic structures. In addition, a new thermal degradation mechanism was proposed to link

the observation from FTIR and volatile compounds from Py-GC/MS. The results provide some

insight into the thermal degradation mechanism of E-epoxy resin, which sheds light on the

thermal properties of E-epoxy and also provides directions for future thermal enhancement.

Chapter 6 explores the influences of NCFs on the curing behaviours of petroleum-based epoxy

resins. Prior to evaluating the mechanical enhancement of NCFs, it is necessary to examine the

effect of NCFs on the network forming behaviours. The catalyst effect of NCFs was observed

on the specific epoxy/amine system. Several models were applied to describe the curing

behaviour of epoxy/amine resin systems. We found that the Sestak-Berggren model better fitted

the experimental data than the Kamal model.

Chapter 7 combines the NCFs and P-epoxy/E-epoxy resins in order to increase the mechanical

performance of the blending system. The major hypothesis of this chapter is that incorporating

NCFs can significantly increase the mechanical performance of the blending system, while

enhancing the toughness by E-epoxy resin. A significant improvement on both tensile strength

and modulus of E-epoxy/P-epoxy blending system was observed as they were close to their

respective theoretical values. The toughness of 10% E-epoxy replacement was 80% higher than

the neat P-epoxy resin. These results show that adding NCFs into bio-epoxy resin can be an

efficient way to improve its mechanical performance without any chemical treatment.

The overall findings of the thesis demonstrate the potential of using bark extractives to replace

BPA and develop a bio-epoxy with superior mechanical and chemical properties. The final

chapter (Chapter 8) summarizes the results and conclusions of each chapter and the scientific

contributions of the work, and discusses the directions of future research. These include:

increasing the yield using different synthesis methods, examining the crosslinking density of

E-epoxy resin using NMR or DMA, and developing a fully bio-based epoxy resin with bio-

epichlorohydrin and bio-curing agent.

6

Scheme 1-1 Thesis overview and organization of chapters

7

References

1. A. M. Intelligence, Market report: global epoxy resin market, 2014.

2. M. Burgham and A. Sheffield, ed. E. C. a. H. Canada, Canada Gazette, Gatineau and

Ottawa, 2010, vol. 144, p. 21.

3. A. E. Gerbase, C. L. Petzhold and A. P. O. Costa, J Am Oil Chem Soc, 2002, 79, 797-802.

4. J. Nichols and E. Schipper, J Am Chem Soc, 1958, 80, 5711-5713.

5. P. B. Chakrawarti and V. Mehta, Indian J Technol, 1987, 25, 109-113.

6. T. Asano, M. Kobayashi, B. Tomita and M. Kajiyama, Holzforschung, 2007, 61, 14-18.

7. H. Pan, Renew Sust Energ Rev, 2011, 15, 3454-3463.

8. C. Sasaki, M. Wanaka, H. Takagi, S. Tamura, C. Asada and Y. Nakamura, Ind Crop

Prod, 2013, 43, 757-761.

9. E. Windeisen and G. Wegener, in Polymer science: A Comprehensive Reference, ed. M.

M. K. Matyjaszewski, Elsevier, Oxford, 2012, vol. 10, ch. 15, pp. 255-266.

10. J. M. Raquez, M. Deleglise, M. F. Lacrampe and P. Krawczak, Prog Polym Sci, 2010, 35,

487-509.

11. S. N. Cheng, I. D'cruz, M. C. Wang, M. Leitch and C. B. Xu, Energ Fuel, 2010, 24, 4659-

4667.

12. C. Xing, J. Deng, S. Y. Zhang, B. Riedl and A. Cloutier, Forest Prod J, 2006, 56, 64-69.

13. N. Staff, in University of Alberta News, University of Alberta, Edmonton, 2014.

14. A. Watts, in Climate News, 2015.

15. M. H. Alma and S. S. Kelley, Polym Degrad Stabil, 2000, 68, 413-418.

16. Y. Zhao, N. Yan and M. Feng, Int J Adhes Adhes, 2010, 30, 689-695.

17. J. D'Souza and N. Yan, Acs Sustain Chem Eng, 2013, 1, 534-540.

18. E. Aspe and K. Fernandez, Ind Crop Prod, 2011, 34, 838-844.

19. R. Auvergne, S. Caillol, G. David, B. Boutevin and J.-P. Pascault, Chemical Reviews,

2013, 114, 1082-1115.

8

Chapter 2 Literature Review

2.1 Conventional Epoxy Resins

In section 2.1, we briefly introduced the background information on conventional epoxy resins,

as a reference point with which to establish the foundation for preparing bio-epoxy resins. This

section includes discussion of the most common epoxy resins/curing agents, the synthesis path

of epoxy precursor, and the network formation. A full description of both the chemistry and

the properties of conventional epoxy can be found in several books and published peer-review

papers1-3.

2.1.1 Common Types of Epoxy Monomers and Curing Agents

The term epoxy refers to a broad group of reactive compounds, characterized by the presence

of an oxirane or epoxy group. This is represented by a three-member ring containing an oxygen

atom that is bonded with two carbon atoms. Epoxy groups are capable of forming a cross-

linked high molecular weight polymer. A general formula for an epoxy resin can be represented

by a linear polyether with terminal epoxy groups and interval secondary hydroxyl groups

(Scheme 2-1).

Scheme 2-1 Bisphenol A diglycidyl ether (BADGE) 2

The first commercial production of epoxy resin occurred simultaneously in Europe and in the

United States, beginning in the early 1940s. To date, several different types of epoxy resins -

bisphenol A epoxy, bisphenol F epoxy, novolacs epoxy, tetraphenolethane epoxy,

cycloaliphatic epoxy, and waterborne epoxy - have been developed (Table 2-1). Similarly,

curing agents come in many varieties, such as aliphatic amines, cycloaliphatic amines, aromatic

polyamines, carboxylic acid, anhydrides, etc. (Table 2-2). Thousands of combinations can

occur between epoxy monomers and curing agents, so the design and manufacture of epoxy

resins is complex, but also flexible.

9

In this study, Bisphenol F diglycidyl ether (BFDGE) was chosen due to its relative flexibility,

low viscosity and high reactivity. An aromatic amine - diethylmethylbenzenediamine

(DETDA) - was selected as the curing agent, which possesses high strength and modulus,

making it suitable for composites.

Table 2-1 Typical properties of some epoxy resins 4 n*1 EEW*2 (g/eq) Viscosity (cP) Functionality Applications

Bisphenol A 0.1-12 188-3200 11,000-15,000 1.9 General

Bisphenol F 0.15 165 2,500-5000 2.1 High fillers

Novolac 0.2-1.8 175-200 20,000-50,000 2.6-3.5 High thermal resistance environment

*1 n= number of repeat units

*2 EEW= Epoxy Equivalent Weight

Table 2-2 Advantages, disadvantages, and applications of common epoxy curing agents 2 Advantages Disadvantages Applications

Aliphatic

amine

Low viscosity

Low cost

Fast cure

Strong skin irritant

Poor bond strength above 80°C

High vapor pressure

Adhesives and sealants

Casting and encapsulation

Polyamide Loose mix ratio

Room temperature cure

Good bond strength and flexibility

High formulation cost

Low mechanical and chemical

properties

General-purpose adhesives

Casting and encapsulation

Aromatic

amine

High Tg

High mechanical performance

Moderate thermal and chemical

resistance

Long elevated-temperature

cures

Rigid

Composites

Adhesives

Electrical encapsulation

Anhydride Good thermal and chemical

resistance

Long elevated-temperature

cures

Insoluble in resin

Composites

Adhesives

Electrical encapsulation

Depending on the types of epoxy resins, the price can vary dramatically, usually from 1,700-

14,500 $/tonne5-6. The price of epoxy resins is 2-5 times higher than some thermosetting resins

such as polyester and phenol formaldehyde. Furthermore, the price of BPA is approximately

1,500-1,600 $/tonne7, which is 2.5 times higher than that of tannins (~600 $/tonne)8. Therefore,

using bark extractives as a replacement for BPA could result in substantial profits, which will

likely arouse interests within the chemical manufacturing industry.

2.1.2 Epoxy Monomer Synthesis

There are three major ways to produce commercial epoxy resins, including ECH etherification,

double-bond oxidation, and glycidyl methacrylate etherification. Since over 90% of

10

commercial epoxy resins are produced from ECH9, this thesis focuses on the synthesis

mechanism using ECH method.

The mechanism between hydroxyl groups and ECH is an SN2 reaction, and is known as

glycidylation, glycidyl etherification, O-alkylation, or the Williamson ether synthesis. As

shown in Scheme 2-2, the first reaction is an acid-base equilibrium (I) which forms phenoxides

that in turn promote the following SN2 reactions (II-IV). There are two possible SN2 reaction

mechanisms to reach an epoxide monomer: a direct substitution reaction (Mechanism A) and

a ring-opening/ring-closing reaction (Mechanism B). In mechanism A, a nucleophilic

phenolate ion attacks C3 on ECH and simultaneously breaks the C-Cl bond. Mechanism B has

two steps (III+IV): in the first step, the nucleophilic phenolate ion attacks C1 on ECH to open

the ring, forming an alcoholate in situ; in the second step, the alcoholate attacks the next carbon,

detaching a chloride and reconstructing an epoxy ring. Several studies reported that the

dehydrochlorination reaction (IV) is significantly faster than other reactions I/II/III 10-11.

Scheme 2-2 Mechanism of glycidyl etherification between phenolic compounds and ECH 1-2, 12

In addition to the major reactions (I-IV), side reactions result in the formation of low levels of

impurities, which decrease the epoxide content and affect the resin properties. The common

side reactions include (VI) abnormal addition, which forms compound 5 and brings difficulties

to dehydrochlorinate; (VII) incomplete dehydrochlorination, which results in residual

saponifiable or hydrolysable chloride, and further affects the reactivity of basic catalysts by

neutralizing such as tertiary amines; and (VIII) hydrolysis of epoxy groups, which generally

cause a small amount (0.1-5 %) of monohydrolyzed resins or α-glycol to be found in the

11

products. The occurrence of these reactions results in mono- and zero-functional epoxide

oligomers.

Based on this mechanism, two common methods were developed to synthesize liquid epoxy

resins. The first method is called the caustic coupling process (ring-opening process), which

uses only sodium hydroxide as a catalyst. Typically, to produce low molecular weight liquid

epoxy, the ECH/BPA ratio is controlled at 10:1, with 20-50 % NaOH(aq) slowly added into the

solution, which may take 2-3 hours; the temperature is maintained at approximately 50-100

°C2. The water content of the mixture is controlled between 0.3-2 %. After the reaction is

complete (the entire reaction can take 6−20 h at reflux1), the mixture usually separates into two

layers. The heavier aqueous layer is discarded, and the viscous product is washed with water

until the wash water is neutral.

The second method for synthesizing liquid epoxy resins is called the phase-transfer catalyst

(PTC) process. Different from the caustic coupling reaction, the PTC process uses ammonium

salts (e.g. triethylbenzyl-, trioctylbenzyl-, tributylbenzylammonium chlorides as well as

cetyltrimethylammonium bromide) to complete the coupling reaction, and adds sodium

hydroxide in order to dehydrochlorinate. In a typical PTC reaction, the ECH/BPA ratio can be

reduced from 10:1 to 3:113. The PTC catalyst is added around 2-10 wt% into a biphase solvent

system (e.g. dichloroform/water). Then, a concentrated (50%) aqueous solution of NaOH is

added dropwise within 2 hours, at room temperature. After introducing the entire amount of

NaOH, the solution is continuously stirred for 2 hours. Finally, the reaction is stopped by the

addition of water, and the organic phase is separated.

Based on the reaction scheme and the above discussion, three points should be noted: (1) the

dominant step is glycidyl etherification, which determines the reaction rate and properties of

the products; (2) pKa values of the substrate and types of catalyst determine the formation rate

of phenoxides; (3) water is the major source to form by-products during synthesis. Thus, it is

essential to choose the right catalyst, and remove water from the synthesis reaction system.

2.1.3 Network Forming Mechanism

Epoxy resins can be transformed from a liquid monomer into a tough and hard cross-linked

material through an anionic or cationic reaction with curing agents. In this study, amine type

of curing agent was selected due to its simplicity of reaction mechanism compared to anhydride

curing agent. The network forming mechanism between epoxy resins and amine type curing

12

agents is an anionic addition reaction, described in the following scheme:

Scheme 2-3 Addition reaction of an epoxy resins with amine type curing agent

The curing agent can be considered as a co-monomer in the polymerization reaction. The first

cross-linking reaction is to react with primary amines, where K1 is the rate constant for the

catalyzed reaction by hydroxyl groups. The reacted epoxy group then act as a hydroxyl group,

making the reaction autocatalytic, with the rate equation containing a rate constant K11. The

reaction mechanism of secondary amines is similar to that of the primary amines, and the rate

constants can be written as, K2, and K22. Other reactions may also occur, such as the reaction

between epoxy and hydroxyl groups (KOH) and epoxy itself (KE). If we assume only the

catalyzed reactions are important14, then the rate equations between epoxy groups with amines

could be:

-dE/dt = K1 [C0] [E][A1] + K11 [E][A1 ][P-OH] + K2 [C0][E][A2] + K22[E][A2][P-OH] Eq. (2-

1)

where [E], [A1], [A2], [C], and [P-OH] represent the concentrations of epoxy monomer, the

primary amine, the secondary amine, hydroxyl groups from catalyst and hydroxyl groups from

curing reaction. By controlling the concentrations of epoxy and curing agent, the rate constant

ratio between secondary and primary amine can be 0.514. Then, the resulting equation could be

written as

dα/dt = [κ1+κ2 α] (1-α)2 Hc Eq. (2-2)

where α is fractional conversion of epoxy groups, κ1=1/2(K1[C0][E0]), κ2=1/2(K11[E0]2), E0 is

the concentration of epoxy groups present initially and Hc is the Horie connection factor. This

model was also modified by Kamal15. The phenomenological Kamal model of cure kinetics is

one of the most frequently cited models in the literature on epoxy systems. This model was,

13

therefore, chosen in this study and is expressed by Eq. (2-3).

dα/dt = (k1 + k2 αm) (1 - α)n Eq.(2-3)

where m and n are the kinetic exponents of the reactions. To determine the parameters of Eq.

(2-3) m, n, k1, k2 were estimated without any constraints by fitting the experimental data using

a graphical method and nonlinear curve fitting.

Using the graphical method, the constant k1 in Eq. (2-3) can be calculated from the initial

reaction rate at α = 0. The kinetic constant is assumed to be of the Arrhenius form, k1 = A exp

(-E1/RT), where A, is the pre-exponential constant, and E1 is the activation energy. Thus, Eq.

(2-3) could be rewritten in the following form:

ln (dα/dt) = ln (k1 + k2αm) + n(1 – α) Eq.(2-4)

Except for the initial region, a plot of ln(dα/dt) versus ln(1-α) is expected to be linear with a

slope n. Eq. (2-4) thus be further rearranged to give:

ln [(dα/dt)/(1 - α)n - k1 ] = ln (k2) + m ln (α) Eq.(2-5)

The first term of Eq. (2-5) can be computed from the above estimated values of k1 and n. If the

left term of Eq. (2-5) is plotted against ln(α), a straight line is produced with a slope and

intercept approximated to m and k2, respectively. To obtain more precise values, an iterative

procedure is used until n, m, Ea1, Ea2 reach a convergent point.

Other common models with which to calculate the kinetic parameters include the Kissinger’s

equation16, the Sestak-Berggren (S-B) empirical equation17, the Kissinger-Akahira-Sunose's

(KAS) isoconversional equation18 and the Friedman differential isoconversional equation19.

The Kissinger’s equation is a common and simple method to calculate the activation energy

and pre-exponential factor using Eq. (2-6). However, it assumes the reaction is a simplified

first order reaction instead of an autocatalytic reaction, which may not well describe the curing

behaviour of epoxy resins accurately.

ln (β/Tp2 ) = ln (AR/Ea ) - (Ea /RTp) Eq.(2-6)

where β is the heating rate; Tp is the temperature at which the maximum heat flow rate occurs;

A is the pre-exponential factor, also called the Arrhenius factor; R is the gas constant; Ea is the

activation energy.

The Sestak-Berggren (S-B) empirical equation is a kinetic model introduced jointly by

14

Johnson-Mehl-Avrami (JMA) 20-21, and is shown as:

dα/dt = k αm(1 - α) n Eq.(2-7)

In kinetic analysis, it is generally assumed that the rate of reaction can be described by two

functions k(T) and f(α),

dα/dt = k(T)f(α) = A exp(-Ea/RT) f(α) Eq.(2-8)

where k(T) is the rate constant, and f(α) is the reaction model.

When the heating rate is constant, Eq. (2-8) can be rewritten as

dα/dT = A exp(-Ea/RT)f(α)/β Eq.(2-9)

As shown in Eq. (2-10), Friedman isoconversional methods offer a model-free estimation of

the activation energy. The fundamental assumption of the isoconversional model is that the

reaction rate is a function only of temperature:

(dln/dT)(dα/dt)α=E(a,α)/R Eq.(2-10)

where Ea,α is the activation energy at a given conversion.

KAS method is also an isoconversional method as shown in the Eq. (2-11), but is calculated in

an integral approach, which is suitable for integral data such as TGA mass loss. Compared to

Ozawa–Flynn–Wall method, the KAS method offers a significant improvement in the accuracy

of the Ea values.

ln (βi/Ta,i2 ) = Const - (Ea /RTa) Eq.(2-11)

where βi is the heating rate and Ta,i is the temperature to reach a given extent of conversion at

varying heating rate.

As mentioned above, these models consider the catalyst effect from hydroxyl groups. However,

the epoxy groups can also react with hydroxyl groups through an etherification reaction. As

such, it should be specially noted if any hydroxyl abundant materials are added into the epoxy

system.

15

2.2 Common Bio-based Epoxy Resins Derived from Renewable Resources

Many natural resources have been assessed as replacements for BPA. In this section, we outline

the properties and synthesis methods of various types of bio-based epoxy resins, and also

compare these to the petro-based epoxy resins.

2.2.1 Oil-based Epoxy Resins

Vegetable oils, such as linseed oil, soybean oil, and castor oil, appear to be excellent renewable

raw materials for thermosetting polymers, as they are expected to be inexpensive as well as

abundant, and possess low viscosity at room temperature. Vegetable oils have become the most

popular renewable resources in the chemical industry22. These plant oils have been widely

applied in polyurethane synthesis, and epoxidized plants oils have rapidly been developed.

Some bio-epoxy commercialized products are already available in the market such as Vikoflex

from Arkema Co. However, one major concern with oil-based bio-epoxy resins is their

mechanical performance for industrial and commercial uses.

To address this issue, some studies have developed the networks with high unsaturated fatty

acid triglycerides (e.g. linseed oil, Scheme 2-4) in order to increase the crosslinking density,

which in turn raise the value of Tg. For example, the Tg value increased from 50 °C to 133 °C23

as the crosslinking density got larger, using anhydride acid as the curing agent. Nevertheless,

compared to conventional epoxy networks, oil-based bio-epoxy still possess substandard

mechanical strength and Tg (184 °C)23. Thus, most of the industrial applications of oil-based

epoxy resins are limited to non-engineering fields such as coatings and general purpose

adhesives, or are used as reactive diluents to reduce the viscosity of petroleum-based epoxy

resins. Still, interesting developments remain to be discovered for the synthesis of fully bio-

based epoxy networks, with epoxy prepolymers and curing agents using vegetable oils.

(a)

(b)

Scheme 2-4 (a) Epoxidized palm oil, (b) epoxidized linseed oil

16

2.2.2 Saccharide-based Epoxy Resins

Carbohydrates are also a renewable resource with potential to be used as a petroleum-based

polymers alternative. Some sugar-based building blocks have been studied intensively, such as

monosaccharide (sorbitol), disaccharides (sucrose and isosorbide) and derived acid compounds

(lactic acid, succinic acid, itaconic acid, and levulinic acid). However, the hydroxyls on

saccharide compounds have high pKa values compared to phenolic hydroxyls, which means

the reactivity of alcohol groups toward ECH is lower than that of phenol groups. These alcohol

groups have to compete with the new generated hydroxyls from the reactions with ECH, which

results in incomplete dehydrochlorination reactions. Thus, saccharide-based epoxy resins

require a different technique to synthesize high quality products.

Sorbitol is a hydrogenized alcohol sugar of glucose and has been selected as one of the top 12

potential bio-based platform chemicals by the U.S Department of Energy24. Sorbitol

polyglycidyl ether has been produced as well-established commercial products such as

DENACOL™ and ERISYSTM. However, due to the incompleteness of the dehydrochlorination

reaction, these commercial products contain about 10−20 wt % chlorine, which embrittles the

epoxy network, and undergoes HCl formation1. Furthermore, compared to DGEBA, its Tg and

mechanical performance remain low25.

Sucrose, commonly known as table sugar, is a disaccharide composed of glucose and fructose

units, and is produced from sugarcane or beets, yielding an annual production of approximately

175 million tons26. Different from sorbitol, the epoxidized sucrose is usually synthesized using

double bond oxidation, which usually requires excess amounts of oxidative agents. A previous

study27 indicates that not all the oxidative agents produce satisfactory results. Peracetic acid

can offer a better ratio of substitution than magnesium peroxyphthalate, m-chlroperozybenzoic

acid, phosphotungstic acid with hydrogen peroxide and molybdenum hexacarbonyl with tert-

butyl hydroperoxide. The final functionality of sucrose-based epoxy monomer can reach 3.7

(octa-allyl sucrose)-7.2(crotonyl sucrose)27.

Isosorbide is prepared from sorbitol with an annual production capacity of 1.7 million tons28.

Isosorbide is also known to be one of the building blocks for chemical syntheses such as

polyesters or polycarbonates. Many reports29-30 detail different methods of preparing

epoxidized isosorbide, which are illustrated in the Scheme 2-5. However, the thermal stability

of epoxidized isosorbide can be an issue, as it contains large amount of oxygen in its ring

17

structures and two hydroxyl groups in each repeat unit. Based on the thumb rule of structure-

stability relationship, these oxgen-containing groups are weak links which can significantly

reduce the thermal stability of epoxidized isosorbide.

Scheme 2-5 Synthesis of diglycidyl ether of isosorbide by various methods 29-30

Scheme 2-6 A new synthesis method for epoxy monomer, using lactide as an example 31

Sugar-based acids are high-potential candidates for green chemistry, and their acid groups have

a strong tendency to react with ECH groups. Surprisingly, few research studies have focused

18

on developing epoxidized sugar-based acid. Assessing the literature, there is one interesting

potential, proposed by Pitet et al.31 as shown in Scheme 2-6.

2.2.3 Lignin-based Epoxy Resins

Lignin is the most abundant aromatic polymer, and has a highly branched and irregular

structure, which varies among the species or extraction technologies. Its basic building blocks

can be schematically simplified into “C9” units, comprised of phenolic moiety bearing three

aliphatic carbons. Based on the C9 unit, lignin possesses the advantage of both aliphatic and

phenolic hydroxyl groups, in variable proportions, which can be exploited to synthesize bio-

based polymers. However, lignin is usually disposed as a waste material or is seen as a low

energy intensive byproduct in the pulping industry. In fact, less than 5 % of lignin is used for

other purposes32. Currently, many companies and research institutions have developed and

patented epoxidized lignin-based resins or curing agents, including Hitachi Co., CIMV Co.,

Araco Co., and Industrial Technology Research Institute of Taiwan. The industrial lignin-based

epoxy resins have been well summarized in the literature33 and most of them require

pretreatments on lignin.

Due to the irregular structures of lignin, some studies33-35 focus on breaking down lignin into

chemically useful compounds, such as coumarylic acid, coniferic acid and sinapylic acid, and

vanillin. There are three major synthesis routes for vanillin-based epoxy resins (shown in

Scheme 2-7), of which all exhibit good thermal resistance.

Scheme 2-7 Three synthesis routes for vanillin-based epoxy resins

2.2.4 Terpene and Resin Acid-based Epoxy Resins

Terpenes, terpenoids, and resin acids are all important class of natural molecular compounds,

19

particularly abundant in pine trees. For the diterpene compounds, limonene, terpinolene, pinene,

terpinen and menthadiene have all been used as raw materials to produce epoxy resins and

curing agents. Among them, limonene is the most well-established diterpene source due to its

abundance. The world annual production of limonene is estimated at between 50 million and

75 million Kg1. However, monoepoxide is commonly found during the double bond

epoxidation process, which results in poor mechanical and thermal properties. Thus, epoxidized

limonene is commercially used as reactive diluents1.

Scheme 2-8 Synthesis of epoxy monomer from abietic acid

Resin acids can be divided into two groups: abietic- and pimaric-type resin acids with

characteristic hydrophenanthrene structures. The most common resin acids found in pine are

derived from three basic tricyclic carbon skeletons abietane, pimarane, and isopimarane. The

intrinsic acidity, rigidity, and renowned hydrophobicity, coupled with other chemical

properties, enable resin acids to be converted to a large number of derivatives36. Given their

structures, resin acids behave like aromatic or cycloaliphatic compounds in terms of rigidity.

Thus, they are good candidates for the preparation of rigid epoxy cross-linked polymers.

However, a previous study37 indicated that the reaction between abietic acids and ECH led to

mono-functionalized epoxy monomers (Scheme 2-8). In order to address this issue, some

studies38-39 modified abietic acid to maleoprimaric acid or levopimaric acid before reaction

with ECH (Scheme 2-9).

20

(a)

(b)

Scheme 2-9 Epoxidized imidodicarboxylic resinic of maleic anhydride (a) and

epoxidized

levopimaric acid (b) 38-39

2.2.5 Polyphenol-based Epoxy Resins

In the past five years, using polyphenols to produce bio-epoxy resins has received much

attention in academia. The most commonly studied polyphenol compounds include gallic acid,

catechin, curcumin, resveratrol, green tea tannins and cardanol.

Gallic acid can be easily isolated from gallotannins by oxidation reaction. The epoxidation

reaction between gallic acid and ECH was first reported by Tomita1, 40. The addition reaction

occurs at both the carboxylic acid group and at least one phenol group in the presence of an

ammonium type phase transfer catalyst. However, Tomita's results only achieved an average

epoxy functionality of 2 despite 4 per gallic acid. Nouailhas et al. proposed another synthetic

pathway leading to the functionality of gallic acid up to 3: gallic acid first reacts with allyl

bromide, followed by double bond oxidation by using m-chloroperbenzoic acid40. Recently,

Aouf et al. reported an adjusted synthesis condition allowing for the synthesis of

tetraepoxygallic acid41, which includes two steps: first, gallic acid reacts with ECH at 100 ºC

with the presence of benzyltriethylammonium chloride (BnEt3NCl) as the PTC catalyst for one

hour; secondly, the reaction temperature is reduced to 30 ºC and then NaOH(aq)/PTC are added

into the solution to react for 90 min.

Catechin was epoxidized either by reaction with ECH, or by alkylation with an unsaturated

halogenated compound followed by oxidation. A full characterization of the compounds shows

the presence of by-products with benzodioxane groups42, which then decreases the average

epoxy functionality. These byproducts result from an internal cyclization reaction between

phenolic alcohol in an ortho position and ECH after addition by an SN2 mechanism.

21

Some other polyphenols were investigated in the hopes to provide new biobased epoxy

monomers, such as curcumin and resveratrol43. The double bonds can be epoxidized by reaction

with m-chlorobenzoic acid, resulting in glycidyl ether compounds, thus enabling cross-linking

to provide epoxy cross-linked polymers.

The above compounds are both used in a pure state without any impurity. However, the

extraction of a single compound is very complex and not environmentally friendly, and so the

availability of these renewable resources is rather sparse. Currently, a new approach was tested,

using inexpensive green tea tannins44, which contained at least five different structures. The

resulting polymers showed an interesting Tg of between 140–190 °C, but no static mechanical

performance was reported.

The last example is of cardanol, which is a mixture of four malkylphenols differing by

unsaturation degree of aliphatic chain: 3% 3-(pentadecyl)phenol, 42% 3-(8Z-pentadecenyl)-

phenol, 17% 3-(8Z,11Z-pentadecadienyl)phenol, and 38% 3-(8Z,11Z,14Z-

pentadecatrienyl)phenol, with mainly cis conformation45. Cardanol is mainly produced from

the liquid of cashew nut shells (CNSL). The total production of CNSL approached 2.8 million

tons in 201046, and CNSL is one of the few major and economically-viable sources of naturally-

occurring phenols. CNSL can thus be regarded as a versatile and valuable raw material for

polymer production, and represents a good natural alternative to petrochemically derived

phenols. By a purifying distillation process, an industrial grade cardanol can be obtained (~90

%)47 in the form of yellow oil.

Due to its both aromatic and aliphatic structures, cardanol seems to be a promising candidate

for a substitution of petroleum-based phenol derivatives. Furthermore, cardanol enhances

coating properties, chemical and mechanical resistances, anticorrosion, and flexibility. A patent

by Cardolite Co. indicates that the epoxidized cardanol has been developed and the structure

of diepoxy cardanol can be optimized in order to get the best performance48.

2.2.6 Wood-based Epoxy Resins

For wood-based epoxy resins, all raw materials are derived from a liquefaction process such as

polyethylene glycol liquefaction49-51, and resorcinol liquefaction52. The synthesis routes follow

the ECH method. Their cured properties all seem slightly lower than petro-based epoxy resins,

such as Tg, shear strength, tensile/flexural strength and thermal resistance. An interesting

22

observation52 showed that wood-based epoxy resins had better compatibility with reinforced

fibre than the petroleum-based epoxy resins according to the composite fracture surfaces.

Considering the information above, we note that many individual components from bark

extractives such as catechin and resin acids have been used to synthesize new types of bio-

epoxy resins. However, no research has mentioned the feasibility of using bark extractives

mixture. Thus, this thesis study focuses on using bark extractives to replace BPA and develop

a novel bio-epoxy resin. The following section examines bark compositions in detail and selects

a suitable extraction process with a reasonable amount of polyphenols.

2.3 Bark Extractives and Extraction Process

Compared to wood, bark usually contains a higher percentage of extractives53-54, which are

essential phenolic resources for producing epoxies. In addition, many bark components have

been photo-degraded before timber harvesting that facilitates the subsequent extraction and

synthesis process. Thus, bark can be a suitable raw material to produce bio-based polymers.

The chemical composition of bark is diverse, significantly depending on species, growing

conditions, extraction methods, temperature, and particle size; thus obtaining analytical data

on bark samples is difficult. For that reason, the following discussion focuses only on the

chemical composition of lodgepole pine bark.

2.3.1 Chemical Composition of Pine Bark

Generally, the chemical composition of bark can be divided into four categories, cellulose,

hemicellulose, lignin and extractives (fatty acids, terpenes, resin acids, tannins, flavonoids,

lignans, stilbene etc.) as shown in Scheme 2-10. As reported by Huang and Yan55, lodgepole

pine bark (LPB) is composed of lignin (8.9 %), holocellulose (28.9 %), hexane, ethanol and

1% sodium hydroxide extractives (62.2 %). This thesis study used LPB as the raw material due

to its high concentration of extractives.

(a) Cellulose (b) Lignin unit - e.g. coniferyl alcohol

23

(c) Sugar from hemicellulose - e.g. xylose (d) Extractives - e.g. epimamool

Scheme 2-10 Major composition of pine bark

2.3.2 Extraction Technology

In order to gain access to target compounds in the bark with high yield, a suitable extraction

process is important. In the past few decades, numerous studies have conducted extraction with

traditional methods (solvent extraction), as well as novel techniques, such as ultrasound-

assisted extraction, microwave-assisted extraction, supercritical fluid extraction, and

accelerated solvent extraction. These techniques have been used to recover bioactive molecules

from plants to completely extract the compounds of interest without altering their properties;

however, these methods increase the cost of manufacturing and decrease the yield of extraction.

Thus, this thesis study only reviews the conventional solvent extraction process.

In the literature, six common solvent extraction fractions of pine bark have been examined:

hexane, benzene, acetone, methanol, water steam, and 1% sodium hydroxide. The extraction

compounds can vary significantly based on the solvent polarity. The polarity of hexane is only

0.1, but the extractive yield from LPB is around 15-16.4 %55-56 and the major compounds

include diterpene (>1,344 μg/g), fatty acid (> 432 μg/g) and wax ester (>248 μg/g). Among all

the diterpene, resin acids such as pimaral, andrographolide and dehydroabietal are the most

abundant categories56 (these structures are listed in the Scheme 2-11, a-c). Compared to other

organic solvents, hexane has the highest extractive yield from LPB.

Hergert57 studied lodgepole pine bark in the 1950s and demonstrated that it is richer in

flavonoids than the bark of other pine species. Five solvents were tested and their target

compounds and yield are as shown in Table 2-3. For the detailed chemical composition analysis,

the pine bark was pre-extracted with benzene, and then extracted with ether and acetone. It was

discovered that myricetin (Scheme 2-11d) accounts for 90 % of the total flavonoid fraction.

http://www.sciencedirect.com/science/article/pii/S0926669014002301#bib0150

24

Table 2-3 Yield and major components of lodgepole pine bark extractives.

Solvents Target Compounds Yield (%)

Petroleum Ether Wax, resin acids 5.53 ± 1.87

Benzene Wax, oxidized resin acids 2.43 ± 0.39

Ethyl Ether Flavonoids, phenolic wax 1.59 ± 1.10

Ethanol Phlobaphene 2.05 ± 0.24

Hot water Tannin and carbohydrates 16.1 ± 3.04

To increase the yield of diterpene, Rowe and Scroggins58 used a combination process to extract

lodgepole pine bark. The ground pine bark were first extracted by benzene and then followed

by 1 N NaOH (~ 4 wt.% NaOH), extracting 28.7 % of lodgepole pine bark, which was 10 times

higher than the previous study. Based on the total extractives, 20% was epimanool; 42.6 % was

free acids (linoleic acid and oleic acid were the two major compounds) and 9.4 % was combined

acids.

Shrimpton59 studied the three solvent systems (steam/acetone/methanol) on mountain pine

beetle infested wound response bark. Compared to the non-infested pine bark, the former

wound tissues contain more resin acid. Shrimpton determined that steam-volatile fraction

consisted mainly of monoterpenes; acetone fraction was dominant by resin acid; and methanol

fraction was mainly sugars. The reported compounds are in agreement with the results from

Rowe and Scroggins58 and Backlund et al.56. However, the yield of extractives in each fraction

was less than 7 %, which was lower than expected.

(a) Pimaral (b) Andrographolide

25

(c) Dehydroabietal (d) Myricetin

(e) Linoleic acid (f) Oleic acid

Scheme 2-11 Major extractives based on the literature

The last extraction technology to discuss is the low concentration alkaline solution. Due to its

high polarity and basic properties, alkaline solutions usually have the highest yield compared

to organic solvents. Based on the results reported by Huang and Yan55, 1% NaOH extraction

had the highest yield (35.5 %) compared to hexane and ethanol extraction processes (11.7-15.0

%, Table 2-4). Thus, beginning with an alkaline extraction of bark can increase the profit in a

large-scale extraction process due to its high yield60. Most importantly, alkaline treatment has

shown promise of achieving complete utilization of lignocelluloses with negligible

environmental impact61.

To examine the compounds in detail, the use of 1% NaOH can isolate condensed tannins, bark

lignin, low-molecular weight carbohydrates, suberin fragments, protein and alkaloids. The

number average molecular weight (Mn) of alkaline extracts is in the range of 1000–1500 Da62.

The composition and Mn of extracted substances depends on the alkaline extraction

26

conditions. Compared to lignin, the molecular weight of alkaline extraction is lower.

Furthermore, some polyphenols cannot be extracted with neutral organic solvents or hot water,

but are soluble in 1% NaOH at an elevated temperature. For the above reasons, 1 % NaOH

extraction process is an ideal approach to isolate the polyphenols from bark, and also facilitate

the epoxidation synthesis with ECH.

Table 2-4 Extraction amounts of various extraction solvents and tree barks 63 Bark Hexane

solubles (%)

Ethanol

solubles (%)

1% NaOH

solubles (%)

Formaldehyde-condensable

polyphenols in 1% NaOH

solubles

Lodgepole pine 15.0 11.7 35.5 8.3

Aspen 8.6 22.3 26.2 3.7

White birch 1.9 3.1 23.7 6.8

Sugar maple 2.3 5.5 20.5 2.6

Balsam fir 8.9 4.5 32.6 12.5

2.4 Reinforced Bio-epoxy Resins

Synthesis of bio-epoxy resins is the first step towards developing a green material. After

synthesis, the overall properties should also be considered in order to prepare green materials

with well-defined applications. Among all the properties, mechanical performance is one of the

most important criteria from an industry point of view. As mentioned above, most bio-epoxy

resins act as reactive diluents, which can improve toughness, reduce viscosity, and give better

wetting and impregnation of fibre, while it can also compensate some mechanical properties.

To address this issue, the most common approach to enhance the overall mechanical

performance of the epoxy resins is to add reinforcements.

2.4.1 Types of Reinforcements

In order to improve their properties, bio-based epoxy networks have been reinforced with

various materials, including glass fibre, natural fibres, nanoclays, and nanofibres. Glass fibre

is one of the most common reinforcing fibres for epoxy resins as several studies64-65 have shown

that glass fibres can significantly improve the tensile strength and modulus of bio-based epoxy

composites. However, previous studies have shown that glass fibres have higher environmental

impacts during the manufacturing process compared to the natural fibres. To address this issue,

many bio-fibres have been incorporated into bio-based epoxy resins, such as wheat straw66,

flax yarn/woven fabric67, corn stover68, regenerated cellulose69, hemp fibre62, banana fibre70,

and wood flour71. Compared to glass fibre, the reinforcement effect of natural fibre is slightly

27

lower. Among all the natural fibre reinforced biocomposites, flax yarn composites67 exhibit the

highest tensile strength of 298 MPa and flexural strength of 117 MPa. However, these micro-

scale fibres are opaque, which may limit their applications.

In the past two decades, nanocomposites have drawn much attention from academia and

industry. Nanoclay has occupied a large portion of nanocomposites research due to the low

cost of silicates. However, the mechanical improvement of nanoclays on bio-based epoxy resin

is limited owing to its layered structure. Most studies focus on its thermal resistance, catalytic

effect, and water resistance72-73. Currently, reinforcing bio-polymers NCFs have become more

popular. Masoodi et al.74 reported a bio-epoxy resin reinforced by NCFs. Although this study

provided a high amount of swelling test data, no mechanical improvements were observed in

this study. The potential of using NCFs thus requires further study.

2.4.2 Cellulose Nanostructure

Since 1985, the cellulosic nanostructure has been studied as reinforcement in nano-

composites75. Various terms were used to describe those cellulosic nanostructures, including

nano-whiskers, nano-crystals, micro-fibrils, micro-crystals, or micro-crystallites. These

terminologies may have led to some misunderstandings and ambiguities. Recently, several

review reports75-77 have sought to organize these descriptors. There are basically two families

of nanosized cellulosic structures, nanocellulose crystals and NCFs.

In morphology, NCFs are long and flexible with the length generally in the micrometer scale,

consisting of alternating crystalline and amorphous domains. Another noteworthy difference

between these two kinds of nanostructures is that NCFs display a web-like structure. Compared

with nanocellulose crystals, NCFs have a higher amorphous content and can be extracted via

mechanical processing of various natural materials. Nowadays, NCFs are usually obtained by

a homogenization or a refining process that displays high elastic modulus and high aspect ratio,

greater than 100, with diameters that range from 10-40 nm. In this study, NCFs were selected,

and their production process has been reported in other research78.

28

Fig. 2-1 TEM image of (a) nanocellulose whiskers and (b) nanocellulose fibres.

Obtained from Peng et al.79 and Wu80, respectively.

2.4.3 Past and Current Research of Reinforced Petroleum-based Epoxy Resins with Cellulose Nanostructure

Tang and Weder 81 reported that a simple solvent exchange process could efficiently isolate

nanocellulose whiskers from cotton and tunicate; then the media could be transferred from

water into organic solvent (dimethylformamide, DMF) suspension, and then mixed well with

BADGE epoxy resins. The whiskers content in this study varied between 4 to 24 % v/v. The

results indicated that adding nanocellulose whiskers could improve tensile storage moduli from

1.6GPa to 4.9 GPa. In addition, the study showed that the mechanical properties could be

predicted by a percolation model. Within the model’s framework, the tensile storage modulus

of the nanocomposite E’ can be expressed as

E′ =(1−2φ+φXr)Es

′ Er′ +(1−Xr)φEr

′2

(1−Xr)Er′ +(Xr−φ)Es

′ Eq. (2-11)

φ = Xr (Xr−Xc

1−Xc)

0.4

. Eq. (2-12)

where ψ is the volume fraction of whiskers that participates in the load transfer, Xr is the volume

fraction of the randomly oriented filler, Xc is the critical whisker percolation volume fraction

calculated by 0.7/A, where A is the aspect ratio of the filler, and E’s and E’r are the tensile

storage moduli of the neat polymer matrix and the reinforcing filler, respectively.

Lu et al. 82 discovered that microfibrillated cellulose (MFC) modified by a titanate coupling

agent can be successfully incorporated into BFDGE epoxy resins. The MFC was diluted with

water and then was solvent-exchanged with ethanol and acetone three times, respectively. After

29

solvent-exchange, the desired amounts of epoxy resin were added into MFC acetone

suspension. The tensile storage modulus of modified MFC/Epoxy composites increased by

approximately 20% compared with that of untreated samples. In addition, titanate modified

MFC showed a higher binding energy than silane modified MFC using XPS analysis.

Ohinish et al. 83 improved the tensile strength of UV-light curing epoxy from 8-17% by adding

MFC, which was solvent-exchanged by ethanol. The fibre loading amount was 0.5 w.t % and

the test specimens were cured by conventional UV curing technique. In addition, this report

mentioned that higher curing temperature can increase tensile strength and interfacial strength,

which was calculated by the average broken fibre length, nominal tensile strength and the

diameter of MFC.

Ruiz et al. 84 produced a new waterborne epoxy coating based on cellulose nano-crystals, which

contributed to an important improvement of the storage modulus of the composite. In addition,

the authors noted that the addition of cellulose dramatically increased the viscosity and that

increasing the amount of emulsifier, shear rate, and time of mixing could reduce the emulsion

droplet diameter.

Shimazaki et al. 85 used a NCFs membrane and then immersed it in BADGE epoxy resins to

produce a transparent and an excellent thermal conductive film. The in-plane thermal

conductivity of the nano-composite was over 1.0 W m-1 K-1, which was 3-5 times higher than

that of conventional transparent resins. The crystalline nature of the NCFs provided excellent

phonon pathways through the nano-composite. The tensile modules of films also increased

from 3.7 GPa (neat polymer) to 5 GPa.

30

References


2013, 114, 1082-1115.

2. E. Petrie, Epoxy adhesive formulations, McGraw-Hill, 2005.

3. J.-P. Pascault and R. J. J. Williams, Epoxy polymers : new materials and innovations,

Wiley-VCH, Weinheim, 2009.

4. J. D. Durig, Comparisons of epoxy technology for protective coatings and linings in

wastewater facilities, Houston, 1999.

5. A. E. Fletcher, Elsevier Advanced Technology Oxford, 1992.

6. C. Red, in High-Performance Composites, Ryan Delahanty, CW Compositesworld, 2014.

7. T. Orbichem, CHEM-NET FACTS CHEMICAL MARKET INSIGHT AND

FORESIGHT - ON A SINGLE PAGE- Bisphenol A,

http://www.orbichem.com/userfiles/CNF%20Samples/bpa_13_11.pdf.

8. V. Technical Research Centre of Finland, New products from bark to replace fossil

compounds: Adhesives and insulating foams from softwood bark tannins.,

http://www.sciencedaily.com/releases/2015/02/150223084051.htm.

9. M. Greenberg, R. Hamilton, S. Phillips and G. J. McCluskey, Occupational, Industrial,

and Environmental Toxicology, Mosby, Philadelphia, 1997.

10. H. Batzer and S. A. Zahir, J Appl Polym Sci, 1977, 21, 1843-1857.

11. N. S. Enikolopyan, M. A. Markevitch, L. S. Sakhonenko, S. Z. Rogovina and V. G.

Oshmyan, Journal of Polymer Science: Polymer Chemistry Edition, 1982, 20, 1231-1245.

12. B. Ellis, Chemistry and Technology of Epoxy Resins, Springer Netherlands, 1993.

13. J. Pielichowski and P. Czub, Angew Makromol Chem, 1997, 251, 1-12.

14. K. Horie, H. Hiura, M. Sawada, I. Mita and H. Kambe, J Polym Sci A1, 1970, 8, 1357-&.

15. S. Sourour and M. R. Kamal, Thermochim Acta, 1976, 14, 41-59.

16. H. E. Kissinger, J Res Nat Bur Stand, 1956, 57, 217-221.

17. J. B. Sestak, G., Thermochim, 1971, 274, 173-177.

18. T. A. a. T. Sunose, Method of determining activation deterioration constant of electrical

insulating materials, Chiba Institution, 1971.

19. H. L. Friedman, Journal of Polymer Science Part C: Polymer Symposia, 1964, 6, 183-

195.

20. M. Avrami, J Chem Phys, 1939, 7, 1103-1112.

31

21. M. Avrami, J Chem Phys, 1941, 9, 177-184.

22. T. W. Abraham and R. Hofer, in Polymer science: A comprehensive reference, ed. M. M.

K. Matyjaszewski, Elsevier, Oxford, 2012, vol. 10, ch. 3, pp. 15-58.

23. J. D. Earls, J. E. White, L. C. Lopez, Z. Lysenko, M. L. Dettloff and M. J. Null, Polymer,

2007, 48, 712-719.

24. T. Werpy, G. Petersen, A. B. Aden, J., J. Holladay, J. White, A. Manheim, D. Eliot, L.

Lasure and S. Jones, ed. U. S. D. o. Energy, U.S. Department of Energy, WASHINGTON

DC, 2004, vol. 1.

25. M. Shibata, S. Yoshihara, M. Yashiro and Y. Ohno, J Appl Polym Sci, 2013, 128, 2753-

2758.

26. U. S. D. o. Agriculture, ed. Agriculture, USDA, Washington, D.C., 2015.

27. N. D. Sachinvala, D. L. Winsor, R. K. Menescal, I. Ganjian, W. P. Niemczura and M. H.

Litt, Journal of Polymer Science Part A: Polymer Chemistry, 1998, 36, 2397-2413.

28. TransparencyMarketResearch, Global Sorbitol Market – Isosorbide, Propylene Glycol,

Glycerol & Other Downstream Opportunities, Applications (Toothpaste, Vitamin C,

Sweetener Etc.), Size, Share, Growth, Trends And Forecast 2012–2018, 2013.

29. M. Chrysanthos, J. Galy and J. P. Pascault, Polymer, 2011, 52, 3611-3620.

30. D. Achet, M. Delmas and A. Gaset, Biomass, 1986, 9, 247-254.

31. L. M. Pitet, S. B. Hait, T. J. Lanyk and D. M. Knauss, Macromolecules, 2007, 40, 2327-

2334.

32. M. Kleinert and T. Barth, Chemical Engineering & Technology, 2008, 31, 736-745.

33. T. Koike, Polymer Engineering & Science, 2012, 52, 701-717.

34. M. Ochi, T. Shiba, H. Takeuchi, M. Yoshizumi and M. Shimbo, Polymer, 1989, 30, 1079-

1084.

35. M. Fache, E. Darroman, V. Besse, R. Auvergne, S. Caillol and B. Boutevin, Green Chem,

2014, 16, 1987-1998.

36. H. Wang, X. Liu, B. Liu, J. Zhang and M. Xian, Polym Int, 2009, 58, 1435-1441.

37. H. H. Wang, B. Liu, X. Q. Liu, J. W. Zhang and M. Xian, Green Chem, 2008, 10, 1190-

1196.

38. A. M. Atta, R. Mansour, M. I. Abdou and A. M. Sayed, Polym Advan Technol, 2004, 15,

514-522.

39. F. Mustata and I. Bicu, Eur Polym J, 2010, 46, 1316-1327.

32

40. H. Nouailhas, C. Aouf, C. Le Guerneve, S. Caillol, B. Boutevin and H. Fulcrand, Journal

of Polymer Science Part A: Polymer Chemistry, 2011, 49, 2261-2270.

41. C. Aouf, H. Nouailhas, M. Fache, S. Caillol, B. Boutevin and H. Fulcrand, Eur Polym J,

2013, 49, 1185-1195.

42. C. Aouf, C. Le Guernevé, S. Caillol and H. Fulcrand, Tetrahedron, 2013, 69, 1345-1353.

43. F. Dasgupta, Google Patents, 2013.

44. S. Benyahya, C. Aouf, S. Caillol, B. Boutevin, J. P. Pascault and H. Fulcrand, Ind Crop

Prod, 2014, 53, 296-307.

45. M. Sultania, J. S. P. Rai and D. Srivastava, Journal of Hazardous Materials, 2011, 185,

1198-1204.

46. C. S. a. C. R. K. Prabakaran Golden Research Thoughts Journal 2014, 3.

47. P. Phani Kumar, R. Paramashivappa, P. J. Vithayathil, P. V. Subba Rao and A. Srinivasa

Rao, J Agr Food Chem, 2002, 50, 4705-4708.

48. Z. Dai and M. J. Chen, Google Patents, 2001.

49. M. Kobayashi, K. Tukamoto and B. Tomita, Holzforschung, 2000, 54, 93-97.


51. C. C. Wu and W. J. Lee, J Appl Polym Sci, 2010, 116, 2065-2073.

52. H. Kishi and A. Fujita, Environ Eng Manag J, 2008, 7, 517-523.

53. T. Filbakk, R. Jirjis, J. Nurmi and O. Hoibo, Biomass Bioenerg, 2011, 35, 3342-3349.

54. I. Miranda, J. Gominho and H. Pereira, Bioresources, 2012, 7, 4350-4361.

55. Z. Huang and N. Yan, Wood and fiber science, 2014, 46, 167-174.

56. I. Backlund, M. Arshadi, A. J. Hunt, C. R. McElroy, T. M. Attard and U. Bergsten, Ind

Crop Prod, 2014, 58, 220-229.

57. H. L. Hergert, J Org Chem, 1956, 21, 534-537.

58. J. W. Rowe and Scroggin.Jh, J Org Chem, 1964, 29, 1554-&.

59. Shrimpto.Dm, Can J Bot, 1973, 51, 527-535.

60. G. Vazquez, G. Antorrena and J. C. Parajo, Wood Sci Technol, 1987, 21, 155-166.

61. B. Xiao, X. F. Sun and R. C. Sun, Polym Degrad Stabil, 2001, 74, 307-319.

62. D. M. Fradinho, C. P. Neto, D. Evtuguin, F. C. Jorge, M. A. Irle, M. H. Gil and J. P. de

Jesus, Ind Crop Prod, 2002, 16, 23-32.

33

63. Z. Huang and N. Yan, Wood and fiber science, 2014, 46, 8.

64. S. N. Khot, J. J. Lascala, E. Can, S. S. Morye, G. I. Williams, G. R. Palmese, S. H.

Kusefoglu and R. P. Wool, J Appl Polym Sci, 2001, 82, 703-723.

65. Y. S. Lu and R. C. Larock, Macromol Mater Eng, 2007, 292, 1085-1094.

66. D. P. Pfister and R. C. Larock, J Appl Polym Sci, 2012, 123, 1392-1400.

67. X. S. Huang and A. Netravali, Compos Sci Technol, 2007, 67, 2005-2014.

68. D. P. Pfister and R. C. Larock, Bioresource Technol, 2010, 101, 6200-6206.

69. K. Adekunle, C. Patzelt, A. Kalantar and M. Skrifvars, J Appl Polym Sci, 2011, 122,

2855-2863.

70. C. Merlini, V. Soldi and G. M. O. Barra, Polym Test, 2011, 30, 833-840.

71. M. Mosiewicki, J. Borrajo and M. I. Aranguren, Polym Int, 2005, 54, 829-836.

72. Z. S. Liu, S. Z. Erhan and J. Y. Xu, Polymer, 2005, 46, 10119-10127.

73. Y. S. Lu and R. C. Larock, Biomacromolecules, 2006, 7, 2692-2700.

74. R. Masoodi, R. F. El-Hajjar, K. M. Pillai and R. Sabo, Mater Design, 2012, 36, 570-576.

75. S. J. Eichhorn, A. Dufresne, M. Aranguren, N. E. Marcovich, J. R. Capadona, S. J.

Rowan, C. Weder, W. Thielemans, M. Roman, S. Renneckar, W. Gindl, S. Veigel, J.

Keckes, H. Yano, K. Abe, M. Nogi, A. N. Nakagaito, A. Mangalam, J. Simonsen, A. S.

Benight, A. Bismarck, L. A. Berglund and T. Peijs, J Mater Sci, 2010, 45, 1-33.

76. S. J. Eichhorn, Soft Matter, 2011, 7, 303-315.

77. I. Siro and D. Plackett, Cellulose, 2010, 17, 459-494.

78. Q. X. Xu, J. Yi, X. F. Zhang and H. L. Zhang, Eur Polym J, 2008, 44, 2830-2837.

79. B. L. Peng, N. Dhar, H. L. Liu and K. C. Tam, Can J Chem Eng, 2011, 89, 1191-1206.

80. N. C. Wu, Master, University of Toronto, 2010.

81. L. M. Tang and C. Weder, Acs Appl Mater Inter, 2010, 2, 1073-1080.

82. J. Lu, T. Wang and L. T. Drzal, Compos Part a-Appl S, 2008, 39, 738-746.

83. Y. Ohnishi, T. Fujii and K. Okubo, Wit Trans Built Env, 2008, 97, 139-148.

84. M. M. Ruiz, J. Y. Cavaille, A. Dufresne, C. Graillat and J. F. Gerard, Macromol Symp,

2001, 169, 211-222.

85. Y. Shimazaki, Y. Miyazaki, Y. Takezawa, M. Nogi, K. Abe, S. Ifuku and H. Yano,

Biomacromolecules, 2007, 8, 2976-2978.

35

Chapter 3 Characterization of Bark Extractive-based Bio-epoxy

Resins

Abstract

This chapter outlines the synthesis and characterization of bio-based epoxy resins derived from

bark extractives. The resins were prepared at various temperatures and catalyst amounts to

determine an optimal for the yield and epoxy equivalent weight value. FTIR and NMR

techniques were used to characterize the chemical structures of E-epoxy monomers.

Measurement results indicated a successful epoxidation of bark extractives after reaction with

epichlorohydrin. GPC results revealed that the molecular weights and polydispersities of E-

epoxy monomers were lower than those of lignin epoxy (L-epoxy) monomers. The curing

kinetic parameters calculated with the Kissinger method and the Model-Free method showed

that E-epoxy had a lower curing activation energy value than petroleum-based epoxy (P-

epoxy). Compared to P-epoxy, E-epoxy/P-epoxy blending system displayed comparable tensile

strength and thermal stability. Research outcome demonstrated promises of using bark

extractives to synthesize epoxy resins replacing toxic BPA.

___________________________________________________________________________

A version of this chapter was published in Green Chemistry (2014) DOI: 10.1039/c4gc00459k

36

3.1 Introduction

Epoxy resins have widespread uses ranging from sealing integrated circuits to rocket coatings

due to their high strength, good compatibility with most materials, and great thermal stability.

However, with the recent awareness of toxicity associated with BPA, a key ingredient of epoxy

resin, there is a need to develop alternative safe raw materials for the resin synthesis. During

the past decade, researchers have explored several natural resources to replace BPA, such as

vegetable oil (soybean oil1-2, linseed oil3, palm oil4, castor oil5 and cashew nut shell liquid6),

simple polyols7, lignin8-9, rosin10, and liquefied biomass11-13. These polymers can reduce

dependency of the resin synthesis on petroleum resources and have the potential to be also

biodegradable. Among these alternatives, vegetable oil, liquefied biomass, and lignin have

emerged as the most promising materials, mainly owing to their commercial feasibility14.

Nevertheless, these renewable resource-derived polymers have limitations. For example, the

aliphatic structure of epoxidized oil results in low mechanical strength and poor thermal

stability15. In order to improve these properties, liquefied biomass, with a high aromatic ring

structure content, is used to enhance polymer mechanical and thermal performance16. However,

the liquefaction process is energy intensive and requires toxic solvents, like phenol17. Lignin,

the major waste material from the pulping industry, has been used to produce lignin-based

epoxy with good mechanical performance and thermal stability. However, the molecular

weight of lignin ranges from 2,000 Da to 50,000 Da18, leading to high viscosities that may

cause difficulties during manufacturing or use since a major application of epoxy resins is in

adhesives which require moderate viscosities in order to perfectly wet the substrate surfaces19.

On the other hand, bark extractives contain large amounts of low molecular weight polyphenols

and require less energy to extract. They have good potential to be used as alternative materials

to replace BPA. However, there is no reported work on bark extractive-based epoxy resin in

the literature.

Tree bark, a waste material of timber processing, is abundant, renewable, and rich in phenolic

compounds. Compared to wood, bark contains a higher amount of lignin and polyphenols (42-

55%)20. However, bark is usually considered to be a waste and is used as firelogs or burnt as

hog fuel, despite having a heating value two times lower than fuel oil21. Recently, various

polymers have been developed from bark, including novolac22 and resol-type23 phenol

formaldehyde resins (PF) as well as polyurethane24. Low molecular weight polyphenolic

compounds can be extracted from bark using various solvents, such as acetone, ethanol,

methanol, organic solvent mixtures, and 1 % NaOH(aq). Among these extraction, 1 % NaOH(aq)

37

was reported to be able to extract the phenolic compounds from barks with a high yield25. Most

importantly, the alkaline treatment has shown promise in achieving complete utilization of

lignocelluloses with little impact to the environment26. In addition, epoxy resins are synthesized

with epichlorohydrin under alkali condition, and thus, the remaining sodium hydroxide in the

bark extractives could also promote the epoxidation reaction. Although alkali treatment

dissolves a large amount of low molecular weight polyphenols, other fractions of bark are also

extracted that include lignin, cellulose, and hemicellulose. These components, which contain

large amount of hydroxyl groups, can also react with epichlorohydrin. Thus, in this study,

commercial lignin and cellulose were also included as model compounds to elucidate some of

the characterization results of bark extractive based epoxy resins.

The objective of this chapter is to develop a novel type of epoxy monomers from bark

extractives via a conventional epoxy resin synthesis process. The chemical structures of

uncured monomers were investigated using Fourier transform infrared spectroscopy (FTIR) as

well as nuclear magnetic resonance spectroscopy (NMR). Resin molecular weight and curing

behaviour were studied using gel permeation chromatography (GPC), and differential scanning

calorimetry (DSC), respectively. Finally, thermal stability and mechanical performance of the

cured bark-based resins were examined and compared to conventional BPA derived epoxy

resins using thermogravimetric analysis (TGA) and a universal mechanical tester.

3.2 Materials and Methods

3.2.1 Materials

Mountain pine beetle infested lodgepole pine (Pinus contorta) bark was provided by FP-

Innovations. Cellulose microcrystalline powder, epichlorohydrin (99%), dichloromethane, and

tetrabutylammonium hydroxide solution (1.0 M in methanol) were purchased from Sigma

Aldrich, ON, Canada. Sodium hydroxide (pellet), acetone (>99.5%) and 1,4-dioxane (99%)

were purchased from Caledon Laboratory Chemicals, ON, Canada. Lignin, protobind 1000,

was purchased from ATM India Ltd, Maharashtra, India. EPON 863 epoxy resins and

EPIKURE W amine type curing agent were supplied by Momentive Specialty Chemicals, OH,

USA.

38

3.2.2 Methods

3.2.2.1 Bark Extraction

Bark extractives were obtained from bark chips using 1% NaOH(aq) solution at 90 °C for two

hours with a 10 : 1 solvent/bark weight ratio. After filtering the residue fraction, the

concentrated extractives were spray-dried using a laboratory spray drier (Yamato GB210)

under the following conditions: inlet temperature of 160 °C, outlet temperature of 60 °C and

air pressure of 0.1 MPa.

3.2.2.2 Synthesis of Bark Extractive Epoxy Resins

E-epoxy was prepared following a similar method to that reported for commercially available

resins. Bark extractives (15 g), epichlorohydrin (150 g), solvent (1,4-dioxane, 150 g) and phase

transfer catalyst (tetrabutylammonium hydroxide solution, 2 mL) were placed in a round

bottom 3-neck glass flask and the temperature was raised to 60 °C while stirring. An excess of

epichlorohydrin was used in a mass ratio of bark extractives to epichlorohydrin of 1 to 10. The

experiments were carried out following a two-factor design, including three synthesis

temperatures (40 °C, 60 °C, 100 °C) and three sodium hydroxide amounts (0.5 mol, 1 mol, 2

mol). 50% w/w sodium hydroxide(aq) was then slowly added to the mixture using a pressure-

equalizing dropping funnel while stirring. The flask was kept at 60 °C for a total of 6 hours to

achieve the addition reaction of epichlorohydrin and the ring formation of epoxy groups. The

products were then diluted and washed with acetone, and the solution was filtered to remove

salt. The acetone and non-reacted epichlorohydrin in the filtered resin solution were evaporated

using a rotary evaporator at 120 °C under reduced pressure. Lignin-based epoxy resins and

cellulose-based epoxy resins were prepared following the same extraction and synthesis

procedure as that described. The replicate of each sample is one.

3.2.2.3 Curing of the Epoxy Monomers

Samples used for the physical testing were prepared by solvent casting and the amount of curing

agent was adjusted according to the resin’s epoxy equivalent weight (EEW) value. EPIKURE

W was used as a curing agent, and the resins were then placed in an oven for curing. The cure

profile was designed as follows: 80 °C for 1 hour followed by 121 °C for 1 hour and 177 °C

for 2 hours. Five samples were prepared and the average value of tensile properties was

reported.

39

3.2.3 Characterization

3.2.3.1 Epoxy Equivalent Weight (EEW)

The epoxy content of the synthesized resin was determined according to ASTM D1652. 0.2–

0.8 g of epoxy resin was placed in a 50 mL flask and dissolved in 10–15 mL of methylene

chloride. Crystal violet at 0.1% w/w 4,4′,4″-methylidynetris-(N,N-dimethylaniline) in glacial

acetic acid was used as the indicator. The solution was titrated with 0.1 N of hydrogen bromide

in acetic acid. The hydrogen bromide solution was standardized by 0.4 g of potassium hydrogen

phthalate each time before EEW determination.

3.2.3.2 Differential Scanning Calorimetry (DSC)

The curing behaviour and glass transition temperature (Tg) of the bio-epoxy resins were

evaluated with a differential scanning calorimeter (DSC) model Q 100 from TA Instruments

under a nitrogen atmosphere. Dynamic DSC measurements were carried out at a ramp rate of

5, 10, 15, and 20 °C min−1 in the temperature range of 30 to 300 °C to obtain curing heat-flow

curves of the liquid samples. Isothermal measurements were carried out from 140 to 170 °C.

After this, the cured samples were heated to 300 °C at 10 °C min−1 to obtain the Tg of the fully

cured samples.

3.2.3.3 Fourier Transform Infrared Spectroscopy (FTIR)

FTIR analysis was performed on a Bruker Tensor 27 spectrometer with a temperature controller

– I0977. Samples were sandwiched between two KBr pellets. All FTIR spectra were recorded

over 4000–400 cm−1 wave numbers at a resolution of 4 cm−1 with 28 scans.

3.2.3.4 Thermogravimetric Analysis (TGA)

Thermogravimetric analysis was performed using a TGA 2950 from 25 °C to 800 °C at a

heating rate of 10 °C min−1 under air and nitrogen flow.

3.2.3.5 Gel Permeation Chromatography (GPC)

Gel permeation chromatography (GPC) analysis was conducted on a Waters 2695 separation

module with a Waters 2998 photodiode array, a Waters 2414 refractive index detector, and two

Waters Styragel 5 μm, HR 4E 7.8 × 300 mm columns in series. The mobile phase used was

HPLC grade THF, at a flow rate of 1.2 mL min−1. The calibration curve was generated by

narrow disperse polystyrene standard from 156 000 Da to 580 Da in THF.

40

3.2.3.6 Nuclear Magnetic Resonance (NMR)

NMR spectra were obtained from a Varian NMR System 500 MHz spectrometer using a 5 mm

carbon detection probe. 50 mg of samples were dissolved in deuterated solvents in the presence

of 0.05% sodium 3-trimethylsilylpropionate-d4 (TMSP) as an internal reference. Heavy water

(D2O) was chosen to dissolve bark extractives and deuterated chloroform (CDCl3) was used to

dissolve the rest of the epoxidized samples. All spectra were referenced to the 1H and 13C

signals of TMSP at 0 ppm. The 1H spectrum was recorded at 25 °C after 526 scans. 30° pulse

flipping angle, 3.98 s acquisition time, and 1 s relaxation delay time were used. The 13C NMR-

1D NMR spectra were recorded at 25 °C after 1000 scans. 30° pulse flipping angle, 1.67 s

acquisition time, and 2 s relaxation delay time were used.

3.2.3.7 Mechanical Universal Testing Machine

The tensile properties of the cured epoxy resins were tested at room temperature using ASTM

type E tensile specimens. A standard computerized testing machine (Instron Model 20) was

used in accordance with the ASTM D-638 procedure with a load cell of 2 kN and a cross head

speed of 2.5 mm min−1. Five specimens were measured for each formulation and analyzed

using one way ANOVA.

3.3 Results and Discussion

3.3.1 Yield and Epoxy Equivalent Weight (EEW)

The efficiency of the addition reaction of the epoxy group was significantly influenced by the

reaction temperature and the amount of catalyst. Fig. 3-1 (a) shows the relationship between

the yield and the reaction temperature for the E-epoxy synthesis. When the synthesis

temperature increased, the yield of bio-resins also increased. However, higher temperatures

resulted in more side reactions, which led to difficulties during the subsequent washing steps.

In the solvent washing steps, acetone was first utilized to dissolve the crude products and

remove the by-product, salt. To complete the purification and remove the remaining unreacted

extractives, chloroform and water were then used for the liquid-liquid extraction. When the

reaction temperature was at 100 °C, the emulsion layer of liquid-liquid extraction became

thicker and remained longer due to the higher number of hydrolyzed epoxy rings. We also

found that the EEW values reduced as the heating temperature increased, likely due to a higher

substitution epoxidation.

The relationship between the yield and the catalyst amount is shown in Fig. 3-1 (b). When the

NaOH amount was equal to half of a mole, the epoxidation reaction proceeded minimally. In

41

contrast, when the catalyst molar amount was equal to two moles, the resins exhibited higher

yield values possibly due to hydrolysis reactions during the synthesis. Our results indicated that

the more suitable catalyst amount is 1 mole. Thus, based on the yield and EEW values, the

synthesis temperature was set at 60 °C and the catalyst amount at 1 mole. More discussion on

the effect of synthetic parameters on E-epoxy can be found in Chapter 4.

(Under 6 hours reaction time with 1 mol catalyst amount)

(Under 6 hours reaction time at 60 °C reaction temperature)

Fig. 3-1 Effect of reaction temperature (a) and catalyst amount (b) on the product yield and

EEW value

3.3.2 Spectral Characterization of Bio-epoxy Resins

Fig. 3-2 (a) shows the IR spectra of the bark extractives, the E-epoxy and the commercial epoxy

resin (P-epoxy). The characteristic absorption bands of bark extractives appear at 3417 cm-1

(OH stretching), 2929 cm-1(CH2 asymmetric), 2858 cm-1 (CH2 symmetric), 2715 cm-1

(aldehyde C-H), 1600 cm-1 /1426 cm-1 (C=C aromatic rings), 1097 cm-1 (C-O stretch in

hemicellulose), and 1017 cm-1 (C-H stretch in cellulose)27. After the bark extractives were

reacted with epichlorohydrin to form E-epoxy, epoxy functionalization was confirmed by the

observation of the absorption peak at 908 cm-1 (asymmetric) and 852 cm-1 (symmetric),

corresponding to the vibration of the epoxide functional group28-29. Compared to the bark

extractives, another significant difference was the lower OH stretching band intensity of E-

epoxy. Other characteristic peaks of E-epoxy include aromatic C-H stretch (3055 cm-1),

aldehyde carbonyl stretch (1724 cm-1), epoxy ring stretch (1252 cm-1), and C-O-C stretch (1101

cm-1). In addition, E-epoxy contained a higher percentage of alkanes and carbonyl groups than

P-epoxy, which could be attributed to the presence of lignin and polysaccharide in the bark

extractives.

To understand better the influence by the major components in the bark extractives, the IR

42

spectra of epoxidized lignin extractives (L-epoxy) and epoxidized cellulose extractives (C-

epoxy) are shown in Fig. 3-2 (b). L-epoxy showed strong aromatic stretch absorption bands at

3052 cm-1 (C-H stretch), 2263 cm-1 / 2033 cm-1 (Ring substitution pattern), and 1632 cm-1 /1501

cm-1 (Ring stretch). In contrast, C-epoxy had a long-chain band absorption at 710 cm-1 and a

strong alcohol vibration at 1113 cm-1 which also appeared on the E-epoxy IR spectrum.

(a) (b)

Fig. 3-2 FTIR spectra of (a) bark extractives, epoxidized bark extractives and commercial

epoxy resin, (b) bio-epoxy resins

Successful functionalization of the bark extractives was also supported through 1H NMR and

13C NMR spectroscopy. Epoxidized bark extractives (Fig. 3-3) showed fewer OH hydrogens

from shifts at 3.6 ppm to 3.9 ppm and showed the appearance of signals at chemical shifts of

2.58, 2.77, and 3.13 ppm representing the epoxy rings and 3.3 and 3.6 ppm from the protons

attached to an ether bond, differing from the bark extractives (Fig. 3-4). The chemical shifts at

9.8 ppm and 9.7 ppm were attributed to the protons attached on the aldehyde groups. Other

aromatic ring protons are shown at 6.3-7.8 ppm, with the signal at 6.8 ppm particularly assigned

to the aromatic protons in the guaiacyl unit30.

The difficulty of analyzing the E-epoxy by 1H NMR was mainly caused by overlapping signals;

therefore, 13C NMR was used for further investigation. In the 13C NMR spectra, the aromatic

structures from extractives and lignin were observed between 174.3 and 106.6 ppm as shown

in Fig. 3-3 (b). The chemical shift at 174.3 ppm belongs to the carbonyl group on the linkage

between catechin and gallic acid of the extractives31. The chemical shifts at 150.2-147.4 ppm,

represent C5 and C7 attached with phenolic hydroxyl groups on the A ring, while chemical

shifts at 146.7-144.6 ppm represent C4’ and C5’ attached with phenolic hydroxyl groups on the

43

B ring. Other aromatic carbons in flavonoids structures show their chemical shifts at 130 ppm

for the C1′, at 115-110 ppm for C8 (interflavonoid bond C4–C8) and at 105 ppm for the C6

(interflavonoid bond C4–C6)32. The absence of the C4 – C6 interflavonoid band at 95 ppm and

the presence of the C4 – C8 interflavonoid band at 105 ppm suggested that the units are

exclusively linked to C4–C8, a classical pattern for a procyanidin. When C4 connects with ring

A, its chemical shift would be observed at 36.5 ppm and when C4 is free, its chemical shift

would be located at 29 ppm33.

In addition, lignin fragments were observed in both the aromatic region (150-110 ppm) and the

aliphatic region (72-54 ppm). These peaks were also confirmed by comparing with the lignin-

based epoxy resin (Fig. 3-5). The major mono-lignin unit in pine bark is guaiacyl (G) and the

dilignin that commonly contains beta-O-4 bonding (48%) and biphenyl bonding (9.5-11%)34.

The chemical shifts at 150-145 ppm were assigned to the C3 and C4 on the G unit. The chemical

shifts at 134 ppm, 132 ppm, and 126 ppm correspond to C1, C5 (etherified), and C5 (non-

etherified), respectively. Other aromatic carbons in the lignin fraction include chemical shifts

at 123 ppm (C1 and C6 in φ–C(=O)C–C units), 113ppm (C5 in G units), and 110 ppm (C2 in G

units). In the aliphatic region, Cα (71.8 and 71.2 ppm), Cβ (129 and 54 ppm), and Cγ (63 and 62

ppm) were also observed30.

Furthermore, the chemical shifts at 44.1 ppm, 50.7 ppm, and 69.3 ppm are related to the epoxy

rings. Note that the absence of chemical shift at 45.2 ppm implies that there is an undetectable

level of residual or unreacted level of epichlorohydrin in our bio-resins. Therefore, the liquid-

state NMR confirmed that the bark extractives were successfully epoxidized with

epichlorohydrin and that by NMR the backbone structures are more similar to lignin than to

cellulose (Fig. 3-6).

44

(a) 1H NMR (b) 13C NMR

Fig. 3-3 Liquid state NMR spectrum of the E-epoxy monomers


Fig. 3-4 Liquid state NMR spectrum of the bark extractives


Fig. 3-5 Liquid state NMR spectrum of the L-epoxy monomers

45


Fig. 3-6 Liquid state NMR spectrum of the C-epoxy monomers

GPC was used to evaluate the molecular weights of the E-epoxy and L-epoxy. In Table 3-1,

the number average molecular weight (Mn) of E-epoxy is shown to be 588 Da and the

polydispersity to be 1.36, while that of L-epoxy are 1071 Da and 2.06, respectively. Compared

to L-epoxy, E-epoxy exhibited a lower molecular weight and narrower polydispersity, which

may offer more uniform performance of the cured resin. Note that the average molecular weight

of liquid type commercial epoxy is about 377 Da, which is smaller than both bio-based epoxy

resins.

Origin software (version 8.6; Microcal Software Inc., Northampton, MA) was used to separate

the multi-peaks in the GPC results. Three peaks were identified in E-epoxy and four peaks were

identified in L-epoxy. Comparing Fig. 3-7 (a) with (b), the first peak of L-epoxy is 3891 Da,

which is similar to the molecular weight reported for alkali lignin by other researchers26, 35. We

suggest that the first peak of E-epoxy (1071 Da) and the second peak of L-epoxy (1396 Da) are

from epoxidized lignin fragments with 5 or 6 monolignol monomers. The second peak of E-

epoxy (513 Da) may be an epoxidized catechin which reacted with four epichlorohydrin units.

This is a likely explanation since the peak of 528-529 Da is the most common repeating unit

shown in commercial tannins31. The third peak (345 Da/372 Da) and fourth peak (277 Da)

might be assigned to the epoxidized resin acid and epoxidized coniferyl alcohol, respectively.

Table 3-1 Molecular weight and polydispersity of E-epoxy and L-epoxy

Mn Mw PDI Peak1 Peak2 Peak3 Peak4

E-epoxy 588 (26) 796(21) 1.36 1071 513 345 -

L-epoxy 1071(43) 2210(39) 2.06 3891 1396 372 277

46

(a) E-epoxy monomers

(b) L-epoxy monomers

Fig. 3-7 GPC traces of uncured bio-epoxy resins

3.3.3 Curing Behaviour

DSC is a useful technique for studying cross-linking reactions of thermosetting epoxy resins.

Non-isothermal DSC analysis was chosen in this study to avoid diffusion control, which can

affect the calculation of kinetic parameters. As in many thermosetting polymers, diffusion and

vitrification also occur in epoxy systems. Performing the analysis with a constant heating rate

can improve segment mobility, which lessens the effect of these phenomena. The raw DSC

plots are shown in appendix A.

Among the various multiple heating rate methods, the Kissinger equation is the most

extensively used method to calculate the activation energy, which is expressed as:

𝑙𝑛(𝛽/𝑇𝑝2) = 𝑙𝑛(𝐴 × 𝑅/𝐸𝑎) − (𝐸𝑎/𝑅𝑇𝑝) Eq. (3 − 1)

where, β is the heating rate, Tp is the temperature at which the maximum heat flow rate occurs,

A is the pre-exponential factor (Arrhenius factor), R is the gas constant, and Ea is the activation

energy.

Using the Kissinger equation, the estimated Ea for commercial resin was 52.0 kJ/mol and Ea

for the bio-resins was 42.0 kJ/mol, both of which fall in the typical range of Ea for common

epoxy-amine polymerization systems (40-70 kJ/mol). The E-epoxy has a lower activation

energy and a higher reactivity compared to the commercial products, possibly due to the

abundant hydroxyl groups in the bark components enhancing the autocatalytic reactions in the

bio-epoxy system.

47

However, one of the limitations of the Kissinger equation is that the reaction mechanism is

assumed to be nth order. Considering that the real Ea value varies with the conversion rate, the

Kissinger-Akahira-Sunose's (KAS) isoconversional equation was applied to observe the

evolution of Ea with the conversion rate. The basic assumption of this analysis is that the

reaction rate at a constant conversion depends only on temperature. In kinetic analysis, it is

generally assumed that the rate of reaction can be described by two functions k(T) and f(α),

𝑑𝛼

𝑑𝑡= 𝑘(𝑇)𝑓(𝛼) = 𝐴 𝑒𝑥𝑝 (−

𝐸𝑎

𝑅𝑇) 𝑓(𝛼) Eq. (3 − 2)

where, k(T) is the rate constant, and f(α) is the reaction model. When the heating rate is constant,

Eq. (3-2) can be rewritten as

𝑑𝛼

𝑑𝑇= 𝐴 𝑒𝑥𝑝 (−

𝐸𝑎

𝑅𝑇) 𝑓(𝛼)/𝛽 Eq. (3 − 3)

As shown in Eq. (3-3), isoconversional methods offer an assumption-free estimation of the

activation energy. The fundamental assumption of the isoconversional model is that the

reaction rate is only a function of temperature:

𝑙𝑛 (𝛽𝑖/𝑇2𝑎,𝑖 ) = 𝐶𝑜𝑛𝑠𝑡 − (𝐸𝑎,α /𝑅𝑇𝑎) Eq. (3 − 4)

where Ea,α is the activation energy at a given conversion. Note that the KAS method offers a

significant improvement in the accuracy of the Ea value compared to the Ozawa–Flynn–Wall

method36.

Fig. 3-8 shows the values for Ea at different curing conversions, α, determined by the KAS

method for non-isothermal data. It is worth noting that the Ea of commercial resins varies within

a narrow range of 3.4 kJ/mol with respect to α. So the constant Ea values implied the diffusion

control resulted in negligible influence on the non-isothermal data of the test samples.

Compared to commercial products, bio-epoxy resins have a higher activation energy before 30

% conversion and a lower activation energy after 30 % conversion. The following two reasons

may account for this observation: First, as mentioned in the GPC section, the molecular weight

of bio-resins is higher than that of the commercial product that can result in a lower mobility

of the molecular chain movement and a higher activation energy at the initial stage. Second,

the viscosity of the reaction mixture decreases dramatically with increasing temperature, which

helps to promote the reaction at its final stage37-38. This non-constant activation energy trend

48

indicates that the non-isothermal reaction of bio-resins/amine curing agent probably follows a

multi-step mechanism38.

In addition, an important aim of the curing study was to predict the curing behaviour of the bio-

epoxy resins. After obtaining the values for the kinetic parameters by the KSA method, we

plotted the predicted rate curve for isothermal data, as shown in Fig. 3-9. Compared to the

ASTM E698 method, several studies have indicated that the Vyazovkin method is more precise

and suitable for epoxy-amine curing kinetics prediction39-40. The Vyazovkin method uses the

following equations36:

𝑔(𝛼) = 𝑡𝛼𝐴𝛼𝑒𝑥𝑝(−𝐸𝑎,α/𝑅𝑇0) Eq. (3 − 5)

𝑔(𝛼) = (𝐴𝛼 /𝛽) ∫ 𝑒𝑥𝑝𝑇𝛼

𝑇0

(−𝐸𝑎,α

𝑅𝑇) 𝑑𝑇 Eq. (3 − 6)

Eq. (3-5) and (3-6) stand for the isothermal cure at Tc = T0 and non-isothermal cure at heating

rate β, respectively. To obtain tα, Eq. (3-6) can be divided by Eq. (3-5):

tα = [∫ expTα

T0

(−𝐸𝑎,α

𝑅𝑇) dT] /βexp (

−𝐸𝑎,α

𝑅𝑇0) Eq. (3 − 7)

Eq. (3-7) enables the determination of time for a given conversion. Fig. 3-9 shows the

isothermal epoxy cure from 140 to 170 °C as predicted by using the Ea dependence derived

earlier. The experimental results are shown as curves and the prediction values are shown as

dots. Since the Ea was offered at every 10 % of the conversion rate, the KAS prediction points

were also offered as discrete data. The overall model prediction gave satisfactory agreement

with the experimental isothermal measurements. The same observation was reported by other

groups39, 41.

49


non-isothermal DSC data

Fig. 3-9 Model-free prediction of isothermal cure from 140 to 170 °C using KAS method

(The experimental data are shown by line. The points correspond to the predicted time)

Furthermore, FTIR was used to monitor the concentration of epoxy functional groups during

the formation of networks, as shown in Fig. 3-10. The spectra were normalized to a reference

absorption at 1184 cm-1 that represents the C-C stretch of bisphenol42. The signal reduction at

908 cm-1 as well as at 880 cm-1 displayed the consumption of epoxy groups, whereas the signal

increase at 940 cm-1 revealed the formation of a new bond.

50

Fig. 3-10 Evolution of oxirane functional group during curing

3.3.4 Mechanical Performance

Bio-epoxy resins were blended with commercial epoxy resins (at 10 %, 30 %, and 50 % w/w

of E-epoxy) and then cured with the hardener EPIKURE W at 80 °C for 1 hour, 121 °C for 1

hour and 177 °C for 2 hours. The hardener EPIKURE W is an aromatic-type curing agent,

providing samples with low viscosity and high strength. Fig. 3-11 shows the tensile properties

of the cured 10 %, 30 %, and 50 % bio-epoxy resins, which display comparable properties to

commercial products. Based on one-way ANOVA, there is no significant difference in the

tensile strengths among the four groups (p=0.087).

Compared to the literature data, E-epoxy blending system (50% E-epoxy/50% P-epoxy) has a

higher tensile strength (63 MPa) than 50% epoxidized soybean oil-based epoxy (18 MPa)43.

The high strength of the bark-based epoxy resin is an attractive property of an alternatve

product. Moreover, compared to the long chain structure of epoxidized soybean oil, bark-based

epoxy has higher potential to be applied in the automobile or aerospace industries. The lower

tensile modulus of E-epoxy resin might be attributed to the bulky ring structures in extractives,

which decreases the crosslink density of the system44.

51

Fig. 3-11 Tensile strength and modulus of the neat P-epoxy, 10, 30, and 50% E-epoxy

replacement of P-epoxy

3.3.5 Thermal Degradation and Thermal Stability

TGA is used to evaluate thermal oxidation and thermal stability of the polymers. The

experiments were conducted under air and inert gas, respectively. Fig. 3-12 displays the TGA

plots obtained from testing commercial epoxy resins and bio-epoxy resins in air, while Fig. 3-

13 shows the results of TGA tests under a nitrogen atmosphere.

The commercial product has a strong degradation peak for thermal oxidation at about 382°C.

The bio-epoxy resins exhibited similar thermal degradation patterns. But bio-epoxy also shows

the presence of new peaks at temperatures of 310°C and 345 °C, which can be caused by phase

separation occurring in the network. The combustion process can be divided into three stages:

between 300-340 °C, 340-390 °C, and 480-600 °C. For the commercial samples and 10 % and

30 % bio-epoxy samples, the first stage showed a mass loss of 0.8-0.9 %, the second stage

showed 64-67 % mass loss and the last stage showed a mass loss of 32-35 %. However, for the

50 % bio-epoxy samples, the mass loss at the first stage increased to 17 % and at the second

stage decreased to 41 %, suggesting that more thermally sensitive compounds exist in the bio-

epoxy resins. Some other bio-based epoxy resins also displayed their degradation temperatures

at around 274-305 °C.7

To understand the thermal stability of the novel epoxy resins, the P-epoxy, E-epoxy and L-

epoxy were heated in a nitrogen atmosphere, the results of which are shown in Fig. 3-13. The

black line represents the mass loss of the cured resins. Compared to the petro-based epoxy, the

52

mass loss of bio-based epoxy was faster between 250 to 400°C. The mass loss curves were

transformed into their first derivative (green lines), which indicates the points of highest rate

of change in the mass loss. The commercial epoxies only showed one degradation peak at

392°C, which corresponds to the HCN bond breaking 45. In comparison, the E-epoxy exhibited

two degradation peaks. The first peak is at 349 °C, which is the same as the shoulder shown in

the curve for the L-epoxy, and the second peak was presented at 392°C, which is the same as

the degradation peak of HCN bonding. In addition, the statistic heat-resistant index temperature

(Ts) is given in Table 3-2. This value was determined from the temperatures at 5 % mass loss

(Td5) and 30 % mass loss (Td30) of the sample obtained by TGA, following Eq. (3-8):

Ts = 0.49(Td5 + 0.6(Td30 − Td5)) Eq. (3 − 8)

The values of the heat-resistance of the E-epoxy were similar to the L-epoxy but lower

compared to the commercial product. Several studies reported that bio-based resins exhibited

lower thermal degradation temperature compared to that of the petroleum resins44. Compared

to other bio-epoxy resins, however, E-epoxy showed comparable performance. Therefore,

additional refining steps, such as extraction of unreacted monomers, might be beneficial to

further improve thermal properties of the E-epoxy.

(a) Mass loss curves

(b) Deriv. mass loss curves

Fig. 3-12 Thermal degradation of the cured epoxy resins with various E-epoxy

replacements (0-50%) under air

53

Fig. 3-13 Thermal stability of the cured P-epoxy, E-epoxy and L-epoxy under nitrogen

atmosphere

Table 3-2 TGA data of the cured epoxy resins

Td onset (°C) Td5 (°C) Td30 (°C) Ts (°C) Char700 (%)

P-epoxy 369 353 383 182 12.2

E-epoxy 294 272 337 152 4.6

L-epoxy 297 277 346 156 18.1

54

3.4 Summary

This chapter examined the feasibility of using bark extractives to formulate bio-epoxy resins.

FTIR, NMR, and GPC methods were applied to identify the chemical structures and molecular

weights of the epoxies synthesized in this study. For uncured bio-epoxy monomers, the

absorption of epoxy groups was observed at 908 cm-1 in the FTIR spectrum and the same

groups were seen at 45 and 51 ppm in the 13C NMR spectrum. Both analyses indicated

successful glycidylation on bark extractives after reaction with epichlorohydrin. Furthermore,

the GPC results showed that E-epoxy has a lower molecular weight (588 Da) compared to the

L-epoxy (1071 Da). Based on the Kissinger kinetic analysis, the activation energy of the E-

epoxy is lower than that of the commercial products due to the abundant hydroxyl groups in

extractive promoting the autocatalytic network formation. The curing time of the bio-epoxy

resin was successfully predicted using non-isothermal data analyzed by the KAS method. For

the cured epoxy resins, mechanical performance and thermal properties showed that the novel

synthesized bio-epoxy resin blended with commercial epoxy displayed mostly comparable

properties as the commercial product. Therefore, bark extractives have great promise to be used

as the raw material to synthesize epoxy resins replacing toxic BPA.

55

References


2. H. Miyagawa, A. K. Mohanty, R. Burgueno, L. T. Drzal and M. Misra, J Polym Sci Pol

Phys, 2007, 45, 698-704.

3. A. Overeem, G. J. H. Buisman, J. T. P. Derksen, F. P. Cuperus, L. Molhoek, W. Grisnich

and C. Goemans, Ind Crop Prod, 1999, 10, 157-165.

4. S. Hirose, J Oil Palm Res, 2011, 23, 1110-1114.

5. T. R. Cuadrado and R. J. J. Williams, Polym Commun, 1989, 30, 239-240.

6. T. W. Abraham and R. Hofer, in Polymer science: A comprehensive reference, ed. M. M.

K. Matyjaszewski, Elsevier, Oxford, 2012, vol. 10, ch. 3, pp. 15-58.

7. D. Fourcade, B. S. Ritter, P. Walter, R. Schonfeld and R. Mulhaupt, Green Chem, 2013,

15, 910-918.

8. Y. Hasegawa, K. Shikinaka, Y. Katayama, S. Kajita, E. Masai, M. Nakamura, Y. Otsuka,

S. Ohara and K. Shigehara, Sen-I Gakkaishi, 2009, 65, 359-362.


Prod, 2013, 43, 757-761.

10. Q. Q. Ma, X. Q. Liu, R. Y. Zhang, J. Zhu and Y. H. Jiang, Green Chem, 2013, 15, 1300-

1310.

11. H. Kishi and A. Fujita, Environ Eng Manag J, 2008, 7, 517-523.

12. H. Kishi, A. Fujita, H. Miyazaki, S. Matsuda and A. Murakami, J Appl Polym Sci, 2006,

102, 2285-2292.


14. S. G. Tan and W. S. Chow, J Am Oil Chem Soc, 2011, 88, 915-923.

15. J. M. Raquez, M. Deleglise, M. F. Lacrampe and P. Krawczak, Prog Polym Sci, 2010, 35,

487-509.

16. T. Koike, Polym Eng Sci, 2012, 52, 701-717.

17. S. N. Cheng, I. D'cruz, M. C. Wang, M. Leitch and C. B. Xu, Energ Fuel, 2010, 24, 4659-

4667.



19. J. O. Wegner, ed. R. Höfer, Royal Society of Chemistry, Cambridge, 2009, ch. 11, pp.

425-435.

56

20. G. K. Gupta, Master Thesis, University of Toronto, 2009.

21. D. G. Briggs, University of Washington Institute of Forest Resources, Seattle, 1994, ch. 9,

pp. 106-110.

22. M. H. Alma and S. S. Kelley, Polym Degrad Stabil, 2000, 68, 413-418.

23. Y. Zhao, N. Yan and M. Feng, Int J Adhes Adhes, 2010, 30, 689-695.

24. J. D'Souza and N. Yan, Acs Sustain Chem Eng, 2013, 1, 534-540.

25. E. Aspe and K. Fernandez, Ind Crop Prod, 2011, 34, 838-844.

26. B. Xiao, X. F. Sun and R. C. Sun, Polym Degrad Stabil, 2001, 74, 307-319.

27. Y. Zhao, Ph.D. Thesis, University of Toronto, 2013.

28. S. Q. Ma, X. Q. Liu, Y. H. Jiang, Z. B. Tang, C. Z. Zhang and J. Zhu, Green Chem, 2013,

15, 245-254.

29. D. L. Pavia, G. M. Lampman and G. S. Kriz, in Introduction to Spectroscopy: A Guide

for Students of Organic Chemistry, Saunders (W.B.) Co Ltd, Philadelphia, 1979, ch. 2.

30. J. Z. Mao, L. M. Zhang and F. Xu, Cell Chem Technol, 2012, 46, 193-205.

31. P. Navarrete, A. Pizzi, H. Pasch, K. Rode and L. Delmotte, Ind Crop Prod, 2010, 32, 105-

111.

32. I. Wawer, M. Wolniak and K. Paradowska, Solid State Nuclear Magnetic Resonance,

2006, 30, 106-114.

33. K. Lorenz and C. M. Preston, Journal of Environmental Quality, 2002, 31, 431-437.

34. E. Adler, Wood Sci Technol, 1977, 11, 169-218.

35. S. Tsuda, K. Nakagawa, T. Oyama, A. Takahashi, Y. Okabe, H. Kagawa, S. Yamada and

Y. Okabe, J. Network Polym. Jpn., 2010, 31, 701-717.

36. S. Vyazovkin, A. K. Burnham, J. M. Criado, L. A. Perez-Maqueda, C. Popescu and N.

Sbirrazzuoli, Thermochim Acta, 2011, 520, 1-19.

37. G. Vazquez, J. Gonzalez-Alvarez, F. Lopez-Suevos, S. Freire and G. Antorrena, J Therm

Anal Calorim, 2002, 70, 19-28.

38. S. Vyazovkin and N. Sbirrazzuoli, Macromol Rapid Comm, 2006, 27, 1515-1532.

39. S. Vyazovkin and N. Sbirrazzuoli, Macromolecules, 1996, 29, 1867-1873.

40. V. P. Privalko, V. Y. Kramarenko, V. L. Sokol and A. M. Karateev, Polym Polym

Compos, 1998, 6, 331-336.

41. L. Gan, Z. J. Sun, Y. Z. Gu, M. Li and Z. G. Zhang, Acta Polym Sin, 2010, 8, 1016-1022.

57

42. D. W. Sohn and K. J. Ko, Korea Polym J, 1999, 7, 181-188.

43. X. J. Yang, S. H. Li, J. L. Xia, J. Song, K. Huang and M. Li, Bioresources, 2015, 10,

2130-2142.

44. X. Q. Liu, W. B. Xin and J. W. Zhang, Green Chem, 2009, 11, 1018-1025.

45. J. Madarasz and G. Pokol, J Therm Anal Calorim, 2007, 88, 329-336.

58

Chapter 4 Effects of Reaction Parameters on the Glycidyl

Etherification of Bark Extractives during Bio-epoxy Resin

Synthesis

Abstract

This chapter outlines the influence of a series of reaction conditions on the yield and reactivity

of the glycidyl etherification reaction of the bark extractive-based bio-epoxy monomer (E-

epoxy). To maximize the yield and epoxy content, the glycidylation reaction was examined

with various substrates, solvents, catalysts, time periods, reaction temperatures, and sodium

hydroxide/hydroxyl (NaOH/OHV) molar ratios. Spray-dried bark extractives were used as

substrates due to their higher hydroxyl group content and lower molecular weight compared to

the oven-dried bark extractives. A dioxane/water combination was selected from among four

solvents based on the yield and epoxy equivalent weight of the final product, and

tetrabutylammonium hydroxide was chosen as a ring-opening catalyst due to its effect of

suppressing hydrolysis. Furthermore, a response surface methodology (RSM) was applied to

find the effect on yield and reactivity of the E-epoxy monomer. Paramount to this the adequacy

of the RSM models and data were examined by the coefficients of determination and the

normalized probability plots of residuals. The maximum extent of conversion with minimum

epoxy equivalent weight was achieved after 4.5 hours reaction time with a NaOH/OHV molar

ratio of 3.4 at 80 °C reaction temperature.

___________________________________________________________________________

A version of this chapter was submitted to ACS Sustainable Chemistry & Engineering, Manuscript ID: sc-2015-

01098t

59

4.1 Introduction

Recent developments1-3 in bio-thermoset polymers suggest that renewable resources can

partially or fully replace petroleum-based polymers with comparable properties at reasonable

costs and with lower ecological impact. Among various types of bio-thermoset polymers, bio-

epoxy resins attract considerable attention from both academia and industry since the cost of

conventional epoxy resins is often higher than other thermosets such as polyester and

vinylesters. In addition, these environmentally friendly bio-polymers can reduce the usage of

bisphenol A (BPA), which is the compound used to synthesize 90% of the world's production

of petroleum-based epoxy resins3. The health concerns of BPA remain a subject for debate4-5;

however, there is a strong interest to replace BPA with natural resources. Various renewable

materials have been explored, including liquefied wood6-7, cashew nutshell liquid8, lignin9-11,

furan12-13, vanillin14-15 flavonoids16-17, and tree bark18.

Although many studies have reported promising results, there remains little systematic research

on synthetic routes19 for these bio-based epoxy resins. Currently, the most common synthesis

method of phenol-type epoxy is to etherify the hydroxyl groups using epichlorohydrin (ECH)

as shown in Scheme 4-1. The mechanism as a whole is an SN2 reaction, and is known as the

glycidyl etherification synthesis. Phenolic hydroxyl groups act as nucleophiles to attack the

alpha carbon on ECH. In contrast to BPA, the natural phenolic hydroxyl groups have

asymmetric structures and different acidity on their hydroxyl groups, which results in their own

distinctive reactivity. For example, in a standard synthetic condition, BPA usually has 85-95%

yield20, which is higher than gallic acid (68%)17 or catechin (48%)21. Furthermore, these bio-

derived epoxy monomers may have their epoxy content decreased along with their

benxodioxane derivatives19. For these reasons, it is important to select suitable reaction

conditions for each bio-based phenolic compound for epoxy resin synthesis.

Factors that can influence epoxy synthesis can be very diverse. Some key parameters are related

to substrate and solvent type, catalyst type and amount, and reaction time and temperature.

Bark extractives are used as the substrate for this study since our previous research18 has shown

that bark extractives can be a good candidate to replace BPA. Two types of substrates with

various hydroxyl values (OHV) and molecular weights were prepared using different drying

techniques. For a typical glycidylation reaction, a wide range of solvents have been reported

for both single-phase systems (methanol22, dimethylformamide23, dimethyl sulfoxide20,

hexane20, benzene20, water24, water/dioxane25-26, water/acetone27, water/ethanol27, as well as

60

water-toluene-butanol28), and biphasic solvent systems (water/dichloromethane20 and

water/hexane20, 29). The effects of solvent on the yield and reactivity could be further linked to

its solute solubility and pKa value. Another important factor must be considered in the selection

process is the environment, health, and safety impacts (EHS) of the solvents. A biphasic

solvent system generally requires phase transfer catalysts (PTCs) to accelerate the ring-opening

reaction for producing an epoxy monomer with low molecular weight, narrow polydispersity,

high yield, and elevated purity19. To subsequently close the ring, sodium hydroxide (NaOH) is

added to remove hydrogen and chlorine. Since the price of PTCs is higher than NaOH, the

other long-established option to catalyze the reaction is adding two-fold NaOH, which is

commonly practiced in industrial scale synthesis.

In addition, it is well known that numerical response variables (such as reaction time and

temperature) should not be investigated independently. Interactions may exist among the

variables and can render such univariate investigations meaningless. Response surface

methodology (RSM) is one of the most widely used methods to optimize the multiple process

parameters. According to the literature, the temperature was varied from 25-100 °C20, 22 for the

reaction liquid-type epoxy monomers, the reaction time ranged from 0.5 to 20 hours at reflux,

or 24 to 26 hours at room temperature3, 30, and the molar ratio between NaOH/OHV was in the

range of 0.12 to 1020, 24, 31.

The purpose of this chapter is to further investigate the reaction of extractive-based bio-epoxy

monomer (E-epoxy) and to identify an optimized set of synthetic conditions for E-epoxy

monomer with maximized yield and reactivity. Six parameters (substrates, solvents, catalysts,

time periods, temperatures, and NaOH/OHV ratios) were examined in this study. Two types of

extractives (spray-dried and oven-dried) were compared to observe the effect of OHV and

molecular weight on the yield and reactivity of E-epoxy. Four types of solvent systems

(water/dichloromethane, methanol, dimethylformamide, dimethyl sulfoxide, and

water/dioxane) were selected, and their target product structures were studied using nuclear

magnetic resonance (13C NMR). Tetrabutylammonium hydroxide (TBAH) was used as a PTC

instead of the two-fold alkaline addition, and the synthesized products were analyzed using

Fourier Transform Infrared Spectroscopy (FTIR) and 1H-NMR. In addition, three sets of

reaction variables (2-10 hour reaction time, 60-100 °C reaction temperatures and 1.4-4.2

NaOH/OHV molar ratios) for glycidylation reaction were optimized based on the yield and

reactivity of the resulting E-epoxy.

61


ECH

3, 30, 32


4.2.1 Materials

Bark chips of mountain pine beetle infested lodgepole pine (Pinus contorta var. latifolia) were

provided by FPInnovations. Epichlorohydrin (99%), CDCl3 (contains 0.03 % (v/v) TMS),

tetrabutylammonium hydroxide (1.0 M in methanol), 2-Chloro-4,4,5,5-tetramethyl-1,3,2-

dioxaphospholane (TMDP), chromium(III) acetylacetonate, cholesterol, phthalic anhydride,

crystal violet, potassium acid phthalate (primary standard grade), tetraethylammonium bromide

(anhydrous crystals), and perchloric acid (0.1 N in acetic acid) were purchased from Sigma-

Aldrich, Mississauga, ON, Canada. NaOH (pellet), acetone (>99.5%), methanol (>99.8%),

dioxane (99%), pyridine (99.8%), glacial acetic acid (>99.5%), dimethyl sulfoxide (reagent

grade), and chloroform (reagent grade) were supplied by Caledon Laboratory Chemicals,

Mississauga, ON, Canada. All chemicals were used as received without further purification.

4.2.2 Methods

4.2.2.1 Extraction Procedure

Bark chips (400 g) were immersed in 1% (w/w) NaOH in water with a 1:10 (solid:liquid) ratio

for 2 hours at 90 °C. Bark extractives were collected after filtration using Whatman #42 filter

papers and the liquid was dried in an oven at 103 °C for 24 hours or by using a spray-drier

operating in an inlet temperature of 150 °C and an outlet temperature of 80 °C with an air

pressure set at 0.1 MPa.

4.2.2.2 Synthesis Procedure

62

The synthetic route followed has been reported in detail in our previous research18 and was

adapted as circumstances required. Briefly, extractives (15 g) were mixed with epichlorohydrin

(150 g), solvent (50 mL) and PTC (2 mL) in a three-neck round-bottom flask and the

temperature of the mixture was raised to the reaction temperature between 40-100 °C

depending on each experimental run. NaOH (pellets) was added slowly to the solution to

catalyze the reaction. The synthesis proceeded for between 2 and 10 hours in a nitrogen

atmosphere.

4.2.2.3 Refining Procedure

The reaction products were diluted with excess acetone, and filtered to remove salt by-products.

Phase separation was then conducted on the resulting solution using water and chloroform in a

separatory funnel. The chloroform-miscible compounds were retained. The unreacted

epichlorohydrin and residual solvents were removed using a rotary evaporator at 120 °C under

reduced pressure via water aspirator.


4.2.3.1 Hydroxyl Value (OHV) Determination

The OHV of bark extractives was determined by both ASTM D4274 Test method B and 31P

NMR analysis. The two techniques are complementary. The ASTM method measures the total

hydroxyl amount of a sample, but this method cannot identify the types of hydroxyl groups,

which can be aliphatic, aromatic or carboxylic acids. The 31P NMR analysis indicates the types

of the hydroxyl groups, but do not necessary give accurate indicate of the total hydroxyl number

since some fraction of the bark extractives may not have totally dissolved in

pyridine/chloroform solvent used for the NMR analysis.

4.2.3.1.1 ASTM D4274 Test Method B

Bark extractives (1 g) and the phthalation reagent (25 mL) were heated at 100 °C for 2 hours.

The phthalation reagent was a solution of phthalic anhydride (113 g) in pyridine (700 mL).

After the reaction finished and the bottle cooled to room temperature, pyridine (50 mL) was

added, and then titrated with 0.5 N NaOH to its equivalence point. The calculation of hydroxyl

number is shown in Eq. (4-1).

Raw hydroxyl number = (B − A)(56.1)(N)/W………………………………………Eq. (4-1)

Where A and B are the milliliters at the equivalence point of the sample and blank, respectively;

N is normality of NaOH; and W is the weight of the sample.

63

Since the bark extractives are basic, the raw hydroxyl number has to be corrected by the

following steps: Bark extractives (1 g), pyridine (75 mL), and water (75 mL) were mixed, and

then the mixture was titrated by HCl to its equivalence point. After that, the solution was back

titrated with 0.1 N NaOH. A blank titration was titrated in the same manner. The alkalinity

correction was calculated using Eq. (4-1). Then, a corrected hydroxyl number was obtained as

follows:

Corrected hydroxyl number = raw hydroxyl number – alkalinity……………………Eq. (4-2)

The unit of hydroxyl number is mg KOH per gram of sample. In order to unify the units from

different OHV measuring methods, the corrected hydroxyl number was transformed as in Eq.

(4-3).

OHV = corrected hydroxyl number/ 56.1 (the mass of KOH)………………………Eq. (4-3)

4.2.3.1.2 31P Nuclear Magnetic Resonance (31P-NMR)

31P NMR analysis was done on an Agilent NMR System 500 MHz spectrometer using a 5 mm

oneNMR H/F probe with an acquisition time of 1.5 s, relaxation delay of 0.1 s, 30° pulse

flipping angle, and 2,000 scans. Samples were pre-treated based on the phosphitylation method

using 2-chloro-4,4,5,5-tetramethyl-1,3,2-dioxaphospholane (TMDP)33. The first step was to

develop a stock solution by mixing pyridine-deuterated chloroform (1:1.6, v/v), which was then

used to prepare a relaxation solution using chromium (III) acetylacetonate (5 mg/mL) as well

as a standard solution of cholesterol (5 mg/mL). Note that all the solutions were dried with

molecular sieves before use. The 31P NMR solutions were produced by mixing the sample (15

mg), relaxation solution (0.1 mL), standard solution (0.1 mL), and stock solution (0.8 mL).

Finally, TMDP (0.1 mL) was added, shaken vigorously, and then was transferred to an NMR

tube for analysis. All spectra were referenced to the (TMDP)2O peak at 132.2 ppm, which was

formed from the reaction of 1 mol of water with 2 mol of TMDP34 and had a peak for excess

TMDP (at 174.9 ppm) to ensure that all reactive species had been completely phosphitylated.

The OHV was integrated relative to the cholesterol standard.

4.2.3.2 Molecular Weight

The molecular weights of bark extractives were determined using a size exclusion column

(SEC) MCX 8-300 mm equipped with a Dionex DX600 ion chromatography and a UV detector

64

at 280 nm. The column was packed with sulfonated styrene-divinylbenzene copolymer-

network of 10 μm particles, and the separation range is from 100 Da to 70,000 Da. The mobile

phase used was 0.1 M NaOH, at a flow rate of 1 mL/min. The calibration curve was generated

by polystyrene sulfonate standard from 6,430 Da to 890 Da in 0.1 M NaOH. The concentration

of standards and samples were 1 mg/mL and all the samples were filtered before conducting

the experiments. Mn, Mw and polydispersity were calculated following ASTM D5296-11.

4.2.3.3 Epoxy Equivalent Weight (EEW)

EEW means the molecular weight of epoxy divided by its molar number of the oxirane. High

EEW value means that low concentration of epoxy group terminates at the molecular chain end.

The EEW of the synthesized resin was determined according to ASTM D1652-11. Epoxy resin

(0.4-2.3 g) was placed in a 50 mL flask and dissolved in 10-15 mL of chloroform. The

tetraethylammonium bromide reagent (10 mL) was added and the solution was mixed using a

magnetic stirring bar. Crystal violet at 0.1 % w/w in glacial acetic acid was used as the indicator

to determine the endpoints of titration. The solution was titrated with 0.1 N of perchloric acid

in acetic acid. The perchloric acid reagent was standardized by 0.4 g of potassium hydrogen

phthalate each time before EEW determination.

4.2.3.4 Structural Characterization

4.2.3.4.1 1H and 13C NMR

1H and 13C NMR spectra were obtained from an Agilent NMR System 500 MHz spectrometer

using a 5 mm Xsens Cold probe. Samples (50 mg) were dissolved in deuterated solvents in the

presence of 0.03% sodium 3-trimethylsilylpropionate-d4 (TMSP) as an internal reference.

Deuterated chloroform (CDCl3) was used to dissolve the epoxidized samples. All spectra were

referenced to the 1H and 13C signals of TMSP at 0 ppm. The 1H spectrum was recorded at 25

°C after 64 scans. The pulse flipping angle was 45°, with 4.5 s acquisition time and 1 s

relaxation delay time. The 13C NMR-1D NMR spectra were recorded at 25 °C after 1,000 scans.

The pulse flipping angle was 30°, with 2 s acquisition time and 1 s relaxation delay time.

65

4.2.3.4.2 Fourier Transform Infrared Spectroscopy (FTIR)

Samples were studied with a Bruker Tensor 27 spectrometer in an environmental chamber.

Liquid samples were placed between two KBr crystals, while solid samples were mixed with

KBr powder to form a pellet. All spectra were recorded over 4000-600 cm-1 at a resolution of

4 cm-1 with 32 scans.

4.2.4 Experimental Design for Response Surface Methodology (RSM)

Many studies have shown that RSM is more reliable than the one-variable-at-a-time (OVAT)

method. OVAT can lead to totally erroneous conclusions when parameters are dependent on

each other. In this case, the time, temperature, and catalyst used are all dependent parameters,

leading to the use of response surface analysis in this study. The process optimization for

maximizing the yield and reactivity of the E-epoxy monomers was carried out by Box–

Behnken design (BBD) with 3 variables. Each factor, or variable, is placed at one of three

equally spaced values, usually coded as -1, 0, +1. BBD requires only 17 experimental runs and

avoids the extreme treatment combinations compared to other RSM designs. The three

independent variables considered were the reaction time, reaction temperature, and

NaOH/hydroxyl group molar ratio. The ranges of these three parameters were chosen based on

our preliminary experiments and on the literature 19-20, 24, 35. The variables and their levels

selected for this study are shown in Table 4-1.

Table 4-1 Variables and their levels for Box–Behnken design.

Variables Term Level

-1 0 +1

Reaction time (hours) X1 2 4 6

Reaction temperature (°C) X2 60 80 100

NaOH/OHV molar ratio X3 1.4 2.8 4.2


4.3.1 Categorical Reaction Conditions

4.3.1.1 Effect of Substrates

Scheme 4-1 shows the phenolate ions act as nucleophiles to attack ECH, which transforms

hydroxyl groups to ether groups. Since the reaction initiates from hydroxyl groups, it is

necessary to measure the OHV of substrates. In addition, the reactivity of hydroxyl groups also

depends on their concentration and acidity. Higher acidity and lower pKa values can promote

the reaction rate. If the substrates do not contain sufficient hydroxyl groups, side reactions may

66

take place between ECH and a second hydroxyl group, which results in higher chlorine content

and more impurities. Therefore, an understanding at the molecular level of the reactivity of

phenolic monomers toward glycidylation represents a crucial step in the development of bio-

based epoxy monomers19.

In order to understand the effect of hydroxyl groups of substrates on the synthesis, two types

of alkaline bark extractives were prepared with different OHV and Mw by different drying

processes. The alkaline extraction process was chosen due to its high yield compared to other

extraction methods. In addition, the alkaline extractives are readily available, non-toxic, and

cost-effective. The most common drying method is oven drying, which is cost-effective and no

investment on special equipment. This drying process usually continues for 24 to 48 hours, and

it may decrease the reactivity of extracts. The spray-drying method can convert fluid into

uniform spherical shape powders in a short time, and it is widely applied on heat sensitive

materials such as milk or tea extracts.

The OHV and molecular weights of bark extractives were measured using the standard

esterification-phthalic method and SEC, respectively. Their results are shown in Table 4-2. The

OHV of spray-dried extractives (SDE) was 9.29 mmol/g, while the OHV of oven-dried

extractives (ODE) was only 7.66 mmol/g. SDE had significantly higher OHV than ODE (p <

0.01). In the literature, the OHV of oven-dried extractives varies from 367-440 mgKOH/g, equal

to 6.54-7.84 mmol/g36-37 akin to our ODE result. These oven-dried bark extractives may have

a further crosslinking reaction between these hydroxy groups due to a long period of

dehydration, which is consistent with the results from SEC. As shown in Fig. 4-1, the retention

time of ODE was shorter than SDE, which indicated the number molecular weight (Mn) and

the weight molecular weight (Mw) of ODE are both larger than that of SDE. To provide insight

into each type of hydroxy group present, a 31P NMR method was applied (Fig. 4-2).

Table 4-2 OHV and molecular weight of two types of extractives.

OHV (mmol/g) Mn Mw PDI Peak 1 Peak 2

ODE 7.66 ± 0.78 1,343 4,442 3.31 10,112 1,254

SDE 9.29 ± 0.90 1,196 3,979 3.32 9,115 1,010

67

(a) ODE

(b) SDE

Fig. 4-1 SEC traces of bark extractives.

The results from 31P NMR agreed with the investigation above that SDE had higher OHV than

ODE. However, the detected concentration of hydroxy groups using 31P NMR was much lower

than that using the esterification-phthalic method since the bark extractives were not fully

soluble in the pyridine/CDCl3 solution. Based on the soluble fraction, five types of hydroxy

groups are reported in Table 4-3, including aliphatic OH, condensed OH, guaiacyl OH, p-

hydroxyphenyl OH, and carboxylic OH. ODE exhibited low concentrations of carboxylic acid

hydroxy groups and aliphatic hydroxy groups, but its concentration of condensed OH was

approximately 14 times higher compared to SDE. The high concentration of condensed OH

was likely due to the prolonged dehydration process in the oven. In addition, these hydroxy

groups have their own distinct reactivities to ECH based on their pKa values and steric

hindrance. The pKa value of aliphatic hydroxyl is between 16 and 18; aromatic hydroxyl is

approximately 10; carbonyl hydroxyl is from 3.4 to 4.519, 38. Thus, carbonyl hydroxyl (e.g., in

resin acid or fatty acids) had the highest reactivity, the aromatic hydroxyl (e.g., in degraded

lignin or flavonoids) had moderate reactivity, and the aliphatic hydroxyl (e.g., in

polysaccharides) had the lowest reactivity, which has a similar probability as water to react

with ECH. Based on the above, SDE was assumed to be a better substrate than ODE.

68

Fig. 4-2 31P-NMR analysis of (a) ODE and (b) SDE.

Table 4-3 OHV from bark extractives using 31P-NMR.

OH (mmol/g) COOH

(mmol/g)

Total

(mmol/g) Aliphatic Condensed Guaiacyl P-hydroxyl

ODE 0.180 0.286 0.015 0.010 0.262 0.753

SDE 1.251 0.020 0.012 0.022 0.449 1.754

To confirm that SDE was indeed a better substrate than ODE, both extractives were used to

synthesize bio-epoxy monomers (Table 4-4). The yield of epoxy monomer from ODE was

lower than that from SDE (p = 0.03) because the high molecular weight of the ODE-based

monomer precipitated during the synthesis process. The average EEW value of the ODE-based

monomer was slightly higher than that of SDE-based monomer; however, there was no

statistical difference in the EEW values between the two types of drying technologies (p =

0.21). The EEW value indicates the concentration of epoxy groups on monomers, which

represents its reactivity for the subsequent crosslinking reactions. A lower EEW value

corresponds to a higher concentration of oxirane and higher reactivity.

Table 4-4 Yield and EEW value of two types of E-epoxy monomers.

Yield*1 (%) EEW (g/eq)

SDE-based monomer 48.1 ± 6.9 338 ± 38.4

ODE-based monomer 41.2 ± 4.9 366 ± 37.5

*1 The yield is estimated based on the Eq. (4)

* The replication of each experimental condition is three

69

According to the measurements using the esterification-phthalic method, the OH concentration

of SDE was 9.29 mmol/g. Assuming that all the hydroxyls react to ECH, which converts

hydroxy groups to 2-methyloxirane groups, the yield can then be calculated as follows:

Yield (%) = M/[B*(1+0.52)]*100………………………………………………….Eq. (4-4)

Where M is the weight of E-epoxy monomer, B is the weight of dry bark extractives, and 0.52

is the stoichiometric amount of 2-methyloxirane for 1 g of bark extractives. All of the yields

reported in this paper were calculated using the above equation.

4.3.1.2 Effect of Solvents

The function of a solvent is to dissolve reactants, assist in reaction processes, and mitigate EHS

impacts. In this chapter, the solute was alkaline bark extractives, which required high polarity

solvents to dissolve. Conventional BPA-based epoxy synthesis often adopts

water/dichloromethane (DCM) as the solvent system, but chlorinated solvents have the concern

of being potentially carcinogenic compounds. In addition, polar aprotic solvents seem

preferable for the entire process because the reactions follow an SN2 mechanism, which

requires aprotic solvents to avoid a solvation shell. The literature have indicated that many

solvents have been applied as reaction media in epoxy resin synthesis, including methanol

(MeOH)22, dimethylformamide(DMF)23, dimethyl sulfoxide (DMSO)20, hexane20, benzene20,

water24, water/dioxane25-26, water/acetone27, water/ethanol27, water-toluene-butanol28,

water/DCM20, and water/hexane20, 29. The properties of these solvents are listed in Table 4-5.

The solubility of bark extractive followed the solvent polarity index as shown in Table 4-5.

Solvents with higher polarity can dissolve more extractives. Based on the solubility results,

four solvent systems were selected to synthesize the E-epoxy monomer and the results are

reported in Table 4-6. The polarity scale is an overall measure of solvent strength and is a

composite of all types of solvent-solute interactions, except for dispersive interactions.

Among these solvents, the water/DCM combination showed the lowest yield (6.6 %) and

highest EEW value (4,647 g/eq). Its low yield and low reactivity were possibly caused by

secondary reaction of the hydrolysis of epoxy groups. Although water, which had little EHS

impact on the synthesis process, readily dissolves bark extractives and NaOH, the final product

with water as solvent was inadequate. Furthermore, a few difficulties emerged during the

refining processes. For instance, a thick emulsion layer of liquid-liquid separation was created,

70

which slowed the extraction of E-epoxy monomer from the raw products to a 24 hour long

process. In order to suppress the hydrolysis reactions, the use of organic solvents to synthesize

bio-epoxy resins is a widely accepted option.

Table 4-5 Chemical and physical properties of common solvents for epoxy monomer

synthesis.

Solvents SPI*1 Ex. sol. pKa BP(°C) Sol. in H2O EHSRA*2 Ref.

Water 9.0 Exc. 15.8 100 N/A - 39

DMSO 7.2 Good 35.1 189 M*4 - 39

Methanol 6.6 Good 15.5 64.6 M - 39

DMF 6.4 Good -0.30*3 153 M ** 39

Acetone 5.4 Fair 19.7 56.2 M - 39

Ethanol 5.2 Good 15.9 78.5 M - 39

Dioxane 4.8 Fair 2.1*3 101.1 M ** 39

DCM 3.2 Poor - 39.8 1.32 **

Benzene 2.7 Poor 43 80.1 0.18 ** 40

Hexane 0.06 Poor 60.0 69.0 0.06 ** 39

*1 SPI = Solvent Polarity Index; *2 EHSRA=EHS regulatory alerts; *3 was measured by its conjugated acid, *4 M=

Miscible

Two types of single-solvent systems were examined, including MeOH (polar protic) and

DMSO (polar aprotic). The results showed an improvement of yield, which increased to 28.4

% and 58.8%, respectively. The EEW from MeOH system had the lowest (231 g/eq) among

the four systems, and the EEW value from DMSO media was 949 g/eq. Although MeOH, as a

polar protic solvent, can create a solvation shell around the extractives, the results from the

MeOH system showed an increase in both the yield and epoxy content. However, the pKa value

of MeOH is 15.5, so there is a possibility that MeOH can join the glycidyl etherification. From

the 13C NMR results (Fig. 4-3), a significant chemical shift at 60 ppm was observed, which

indicated that MeOH may have reacted with epichlorohydrin to produce 2-methoxyoxirane. To

address this issue, DMSO was chosen to improve the yield of final products and avoid the side

reactions between solvent and ECH. However, the boiling temperature of DMSO is 189 °C,

which requires a high refining temperature (140 °C in this study). At high temperatures, further

condensation reactions can occur and be promoted to form a high-molecular-weight E-epoxy

monomer with elevated EEW values.

71

Table 4-6 Yield and EEW value of bio-epoxy monomer synthesized in different media.

Solvents Yield (%) EEW (g/eq)

Water/DCM 6.6 ± 3.2 4,647 ± 743

MeOH 28.4 ± 8.2 231 ± 76

DMSO 58.8 ± 7.5 949 ± 46

Water/Dioxane 48.1 ± 6.9 338 ± 38

*1 All the synthesis were conducted for six hours with 4.2 NaOH/OHV ratios at 60 °C

The last solvent system was water/dioxane combination, which can dissolve more extractives

than a single water phase. The yield of the final product was 48.1 % and the EEW value was

338 g/eq. Although its yield and EEW values were only second best among the four types of

solvents, the advantage of using water/dioxane combination was its well-balanced properties,

including its moderate boiling temperature (approximately 100 °C), no side reactions observed,

and stability in a high pH environment. Furthermore, the concentration of oxirane peaks on E-

epoxy monomer was more significant than any above solvent system according to the 13C NMR

results (Fig. 4-3). For these reasons, the water/dioxane combination was chosen for this study

and for the following experiment.

Fig. 4-3 13C-NMR spectra of E-epoxy monomers synthesized in various solvent systems.

72

4.3.1.3 Effect of Catalysts

In Scheme 4-1, NaOH was used to promote two reactions: forming phenoxide ions and

dehydrohalogenating the intermediates. The conventional approach is to add NaOH into the

synthesis solution at two stages, which is superior to adding all of the NaOH at once initially.

However, this method provides an average yield and moderate purity with non-negligible

amount of byproducts such as 1-chloro-3-aryloxypropan-2-ols. The remaining chlorine can be

up to 30 wt% of the epoxy resins3. In addition, when these NaOH do not bond to form

phenoxides or withdraw chlorines, these free nucleophiles at high temperatures can further

stimulate side reactions.

To increase the yield and prevent side reactions, phase transfer catalyst (PTCs) were introduced

to synthesize low-molecular-weight epoxy resins in 198041. Compared to the conventional

approach, the PTC method reduces the risk of hydrolyzing desired products and saves the

amount of ECH consumed by NaOH20. Although the PTC method seems promising in the

process of epoxy synthesis, only a small number of studies focus on this field because the cost

of PTC can be 5-10 times higher than NaOH. Among the many types of PTC, quaternary

ammonium salts (QAS) are the most widely used catalysts for epoxy synthesis. The mechanism

of QAS is to shuttle OH between an alkaline aqueous phase and an immiscible organic phase

by the quaternary ammonium cation as shown in Scheme 4-2.

Scheme 4-2 Mechanism of glycidyl etherification catalyzed by tetrabutylammonium

hydroxide (TBAH)

14, 42

Even though water/dioxane was not a biphasic system, the results showed that adding TBAH

can improve yield and increase the epoxy content as shown in Table 4-7. It is likely because

TBAH can prevent side reactions such as oligomerization and hydrolysis. Our FTIR and NMR

results showed that the concentration of hydroxy groups of bio-epoxy resins under PTC

synthetic path was significantly lower than that under twofold NaOH addition reactions (Fig.

73

4-4). According to the ratios of two NMR peak areas (5.3 ppm/3.1 ppm), the concentration of

secondary hydroxy groups was twice less in TBAH method than in twofold NaOH synthesis.

Therefore, TBAH can effectively increase the yield of the E-epoxy monomer and decrease the

possibility of hydrolysis reactions during synthesis.

Table 4-7 Yield and EEW value of two types of epoxy monomers through two synthetic

paths.

Yield (%) EEW (g/eq) 2°-OH / Epoxy ratio

TBAH + NaOH 48.1 ± 6.9 338 ± 38 0.27

NaOH + NaOH 15.1 ± 4.8 642 ± 27 0.55

*Three samples for each synthesis routes were prepared to get average and standard deviation values of yield and EEW

Fig. 4-4 Effect of catalyst types on functional groups of E-epoxy products observing by

FTIR (left) and NMR (right): (a) TBAH/NaOH and (b) twofold NaOH.

4.3.2 Numerical Variables — Reaction Time, Reaction Temperature, and NaOH/OHV Molar Ratio

Continuous variables were analyzed using response surface methodology (RSM) due to their

dependency on one another. Incorporating the B-B Design, 17 samples were tested with three

levels for each factor (NaOH/OHV molar ratio, reaction temperature, and reaction time), and

the response parameters were the yield of the glycidylation reaction and epoxy equivalent

weight of the final samples.

The span of each synthetic parameter was selected based on accepted methods as reported in

the literature. For reaction temperature, it is known that increasing temperature can accelerate

the overall reaction rate; however, side reactions can be promoted simultaneously such as

oligomerization and hydrolysis. Previous studies have reported that the reaction temperatures

74

of liquid form epoxy monomers range from 25-100 °C, whereas in bio-epoxy synthesis20-21 the

temperatures are between 30-104 °C24. Below or beyond this range, product yield and reactivity

are both decreased. Generally, if the temperature is raised above 100 °C, the final products

transform from liquid to solid form. In addition, increasing the temperature above 110 °C can

cause thermal degradation for thermally sensitive compounds such as flavonoids43. For these

reasons, we chose 40-100 °C as our reaction temperature range.

According to the literature, the reported reaction time of conventional epoxy is in the range of

0.5-20 hours3 and for bio-epoxy synthesis the time ranges from 1.5-8 hours24. Most studies

agree that the total amount of glycidyl substrate (monomers and oligomers) increases over time.

In contrast, the influence of reaction time on the risk of hydrolyzing products is still uncertain22.

In order to maximize the yield of monomer and avoid hydrolysis, time periods were selected

based on a correlation between reaction time and reaction temperature, which was established

based on the literature data24. In respect of the reaction temperature range, 2-10 hour reaction

time periods were selected for this study.

The last parameter considered in this study was the NaOH/OHV ratio, which was susceptible

to having more significant effects than reaction time on yield and epoxy content. With

insufficient NaOH, certain amounts of unreacted polyphenols will remain and decrease the

yield. With excess NaOH, ECH can be hydrolyzed and wasted29. According to the

stoichiometry, the mole ratio of NaOH to OHV on the substrate should be 1:1. However, some

studies have shown that excess amount of base can accelerate the reaction20. In an ordinary bio-

epoxy resins synthesis, the NaOH/OHV ratio is in the range of 0.12-4.56, 24. For this reason,

1.4-4.2 molar ratios were adopted for this study.

75

Table 4-8 The B-B matrix and output responses.

X1 Temperature X2 Time X3 NaOH/Substrate Y1 Yield Y2 EEW

(°C) (hour) molar ratio (%) (g/eq)

60 2 2.8 10.5 351

60 6 1.4 15.2 899

60 6 4.2 34.2 382

60 10 2.8 21.2 317

80 2 1.4 22.5 449

80 2 4.2 51.0 398

80 6 2.8 54.5 348

80 6 2.8 51.0 398

80 6 2.8 54.0 468

80 6 2.8 57.5 444

80 6 2.8 59.0 368

80 10 1.4 38.9 693

80 10 4.2 50.5 1625

100 2 2.8 36.4 498

100 6 1.4 33.2 1045

100 6 4.2 64.0 2240

100 10 2.8 37.9 1771

The yield and the EEW values of the E-epoxy monomer based on B-B design are shown in

Table 4-9. The ranges of yields and EEW values were 10.5-64.0 % and 317-2,240 g/eq,

respectively. Although the yield can be up to 64 % based on the dry weight of spray-dried bark

extractives, the yield of bio-epoxy resin based on the dry bark chips was only between 2-14 %.

In addition, based on our previous research18, the average molecular weight of bio-epoxy resins

was 796 Da. Thus, the functionality per molecule was approximately 0.36-2.51.

Prior to analyzing the influence of the parameters, the model adequacy was examined and

results are listed in Table 4-9 and Table 4-10. The responses were fitted with various functions

such as linear, two-factor interaction (2FI), and quadratic models in order to find the best-fit

model. Among the models, the quadratic model responded best to both the yield and EEW

values according to their coefficients of determination (R2 = 0.96 and R2 = 0.98, respectively)

and their p-values of lack of fit (p = 0.71 and p = 0.09, respectively).

76

Table 4-9 ANOVA results of the quadratic model for the yield of reaction.

Sum of Squares df Mean Square F Value p-value

Model 4177 9 464.17 54.28 <0.0001

Residual 59.86 7 8.55 - -

Lack of Fit 16.16 3 5.39 0.49 0.71

Pure Error 43.70 4 10.93 - -

Total 4237 16 - - -

p-value of lack of fit : 0.0051 (linear model) and 0.0029 (2FI model).

Table 4-10 ANOVA results of the quadratic model for the EEW value.

Sum of Squares df Mean Square F Value p-value

Model 5.5E+6 9 6.1E+5 100.4 <0.0001

Residual 4.2E+4 7 6.0E+3 - -

Lack of Fit 3.2E+4 3 1.1E+4 4.2 0.09

Pure Error 1.0E+4 4 2.5E+3 - -

Total 5.5E+6 16 - - -

p-value of lack of fit : 0.0002 (linear model) and 0.0004 (2FI model).

Table 4-9 and Table 4-10 show that the model adequacy matched requirements at a 5%

statistical significance level. The second step was to identify the data distribution by the normal

probability plot of residuals (Fig. 4-5), which demonstrated that no further data transformation

was required, since all of the data points were approximately linear. Other residual diagnoses

are provided in the Appendix B.

(a)

(b)

Fig. 4-5 Normal probability plot of (a) yield and (b) EEW value of E-epoxy monomer.

Thus, the response surface plots were generated to illustrate the influence of the continuous

reaction parameters as shown in Fig. 4-6.We found that all three parameters were significant

to the yield (p < 0.01) by ANOVA analysis. Reaction temperature and NaOH/OHV molar ratio

77

had the highest effects on yield compared to reaction time. Temperature and time both followed

a quadratic path, as shown in the 3D plots (Fig. 4-6). At higher temperatures, the yield decrease

was likely due to further oligomer formation between E-epoxy monomers. Many studies have

shown that oligomers will solidify when they have very high molecular weight29, 44. These

oligomers can precipitate with salt to the bottom of a reaction flask, which might be left out of

subsequent work up and decrease the total yield. As opposed to time and temperature, the yield

increased with an increasing NaOH/OHV molar ratio, which had a positive linear effect. It is

well known that the rate determining step in a glycidylation reaction is reaction II in Scheme

4-145. More catalyst can promote the first equilibrium reaction, accelerate reaction II, and

increase the total yield. A previous study24 also showed that the yield of a reaction generally

increased with the increase of NaOH/OHV ratio and that the optimal NaOH/OHV ratio was

1.5. According to the ANOVA analysis, we found the final model to predict the yield is as

follows:

Yield (Y1) = 55.1 + 11.3X1 + 3.58X2 + 11.3X3– 16.4X12

– 12.2X22 …………………Eq. (4-5)

Based on a confidence interval of 99%, the significant terms are X1, X2, X3, X12, and X2

2. Since

the constants of Xn2 in yield model are negative, the parabola open down and there is maximum

vertex point for extent of conversion.

On the right side of Fig. 4-6, the influence of reaction conditions on EEW values is shown.

Unlike the results for yield, those for EEW showed different trends at different levels. If we

fixed two parameters at a relatively low level, the last parameter followed a quadratic path due

to incomplete dehydrochlorination or propagation. On the other hand, in an opposite situation,

the last parameter showed a linear positive relationship. Previous studies have indicated that at

lower temperatures, the effects of NaOH is not as significant as at high temperatures24. The

final model to predict the EEW is as follows:

EEW (Y2) = 405.2 + 450.63 X1 + 338.75X2 + 194.88X3 + 326.75X1X2 + 428X1X3 + 245.75

X2X3 + 339.65X12

+ 396.65X32………………………………….……………………Eq. (4-6)

Based on a confidence interval of 99%, all of the terms are significant except for X22. The

constants of X12 and X3

2 are positive in the EEW model so there is a minimum of the value of

EEW.

To match the maximum point of yield and the minimum EEW value, we can reach an optimal

point for both equations. The optimal conditions were 80°C, 4.5 hour and 3.4 NaOH/OHV

78

molar ratio. To validate the accuracy of this regression equation, supplementary experiments

were conducted under these optimal conditions. The results listed in Table 4-11 confirmed the

accuracy of both models

(a)

(b)

(c)

Fig. 4-6 Response surface plots of various parameters on the product yield and reactivity :

(a) Effects of time and temperature, (b) Effects of temperature and NaOH/OHV ratio, and

(c) Effects of time and NaOH/OHV ratio.

Table 4-11 Verification of the proposed optimal synthesis conditions.

Yield (%) EEW (g/eq)

Predicated values 59.2 667

Experimental values 59.0 ± 1.8 594 ± 52

79

4.4 Summary

In this chapter, we tested three categorical parameters (substrates, solvents, and catalysts) and

three numerical parameters (reaction time periods, reaction temperatures, and NaOH/OHV

ratios). Our results showed spray-dried extractives are a good substrate, with both high

hydroxyl content (9.29 mmol/g) and low molecular weight (Mw = 3,979 Da). Among four types

of solvent systems, a water/dioxane combination was selected for this study owing to its

balanced properties on each perspective such as yield (48.1 %), reactivity (EEW = 338 g/eq),

boiling temperature (approximately 100°C), no side reactions observed, and stability in a high

pH environment. To avoid hydrolysis reactions occurring, tetrabutylammonium hydroxide

(TBAH) was added at the first stage as a ring-opening catalyst. According to our 1H NMR

results, the use of TBAH can significantly decrease the hydrolyzed by-products. To optimize

the yield and reactivity, the experiment should be conducted for 4.5 hours with 3.4 NaOH/OHV

molar ratio at 80 °C reaction temperature.

80

References

1. J. M. Raquez, M. Deléglise, M. F. Lacrampe and P. Krawczak, Prog Polym Sci, 2010, 35,

487-509.

2. T. Koike, Polymer Engineering & Science, 2012, 52, 701-717.


2013, 114, 1082-1115.

4. L. L. Ferreira, R. Couto and P. J. Oliveira, European Journal of Clinical Investigation,

2015, 45, 32-36.

5. E. EFSA Panel on Food Contact Materials, Flavourings and Processing Aids (CEF),

EFSA Journal, 2015, 13.

6. H. Kishi, A. Fujita, H. Miyazaki, S. Matsuda and A. Murakami, J Appl Polym Sci, 2006,

102, 2285-2292.


8. S. Gunasekaran, T. B. Gorczyca and H. S. Cole, Google Patents, 2003.






Prod, 2013, 43, 757-761.

12. F. S. Hu, J. J. La Scala, J. M. Sadler and G. R. Palmese, Macromolecules, 2014, 47, 3332-

3342.

13. J. Deng, X. Liu, C. Li, Y. Jiang and J. Zhu, Rsc Adv, 2015, 5, 15930-15939.

14. M. Fache, E. Darroman, V. Besse, R. Auvergne, S. Caillol and B. Boutevin, Green Chem,

2014, 16, 1987-1998.

15. M. Fache, R. Auvergne, B. Boutevin and S. Caillol, Eur Polym J.


Prod, 2014, 53, 296-307.

17. C. Aouf, H. Nouailhas, M. Fache, S. Caillol, B. Boutevin and H. Fulcrand, Eur Polym J,

2013, 49, 1185-1195.

18. P. Y. Kuo, M. Sain and N. Yan, Green Chem, 2014, 16, 3483 - 3493.

19. C. Aouf, C. Le Guernevé, S. Caillol and H. Fulcrand, Tetrahedron, 2013, 69, 1345-1353.

81

20. J. Pielichowski and P. Czub, Angew Makromol Chem, 1997, 251, 1-12.

21. H. Nouailhas, C. Aouf, C. Le Guerneve, S. Caillol, B. Boutevin and H. Fulcrand, J Polym

Sci Pol Chem, 2011, 49, 2261-2270.

22. H. Batzer and S. A. Zahir, J Appl Polym Sci, 1977, 21, 1843-1857.

23. C. S. Wang and Z. K. Liao, Polym Bull, 1991, 25, 559-565.

24. F. Ferdosian, Z. S. Yuan, M. Anderson and C. Xu, Rsc Adv, 2014, 4, 31745-31753.

25. A. K. Banthia, D. Lunsford, D. C. Webster and J. E. Mcgrath, J Macromol Sci Chem,

1981, A15, 943-966.

26. G. Jungang, J Appl Polym Sci, 1993, 48, 237-241.

27. S. Z. Rogovina, O. B. Salamatina, M. A. Markevich, A. I. Nepomnyashchii and N. S.

Enikolopyan, B Acad Sci Ussr Ch+, 1975, 24, 1626-1629.

28. M. A. Markevich, A. I. Kuzaev, L. S. Sakhonenko, L. V. Vladimorov, V. A. Pekarskii

and M. G. Brusilovskii, Vysokomol Soedin a+, 1984, 26, 615-622.

29. H.-C. Kang, B. Lee, J. Yoon and M. Yoon, J Amer Oil Chem Soc, 2001, 78, 423-429.

30. E. Petrie, Epoxy adhesive formulations, McGraw-Hill, 2005.

31. H. Kishi, Y. Akamatsu, M. Noguchi, A. Fujita, S. Matsuda and H. Nishida, J Appl Polym

Sci, 2011, 120, 745-751.

32. B. Ellis, Chemistry and Technology of Epoxy Resins, Springer Netherlands, 1993.

33. A. Granata and D. S. Argyropoulos, J Agr Food Chem, 1995, 43, 1538-1544.

34. A. W. T. King, I. Kilpelainen, S. Heikkinen, P. Jarvi and D. S. Argyropoulos,


35. N. S. Enikolopyan, M. A. Markevitch, L. S. Sakhonenko, S. Z. Rogovina and V. G.

Oshmyan, Journal of Polymer Science: Polymer Chemistry Edition, 1982, 20, 1231-1245.

36. S. Hartman and N. Y. Mahopac, Manufacture of Polyurethanes from bark, Champion

International Corporation, Stamford, Conn. ,UNITED STATES Pat., 1977.

37. K. Kemppainen, M. Siika-Aho, S. Pattathil, S. Giovando and K. Kruus, Ind Crop Prod,

2014, 52, 158-168.

38. W. Bradley, J. Forrest and O. Stephenson, J Chem Soc, 1951, 1589-1598.

39. V. T. Wyatt, J Appl Polym Sci, 2014, 131.

40. A. J. Gordon and R. A. Ford, The Chemist's Companion: A Handbook of Practical Data,

Techniques, and References, John Wiley & Sons Inc., New York, 1973.

82

41. A. K. Banthia, D. Lunsford, D. C. Webster and J. E. McGrath, Journal of

Macromolecular Science: Part A - Chemistry, 1981, 15, 943-966.

42. D. Mandal and J. A. Gladysz, Tetrahedron, 2010, 66, 1070-1077.

43. J. H. Bridson, Master, University of Waikato 2007.

44. D. R. Medjitov, L. G. Shode and G. M. Tseitlin, Polym Bull, 1998, 40, 509-516.

45. G. Rokicki and W. Kuran, Polymer, 1980, 21, 1069-1072.

83

Chapter 5 Thermal Degradation of Extractive-Based Bio-Epoxy

Monomer and Network: Kinetics and Mechanism

Abstract

In order to broaden the applications of bio-epoxy resins in high performance sector, an

understanding of thermal behaviour of these environmentally-friendly biopolymers is essential.

This chapter investigates the thermal degradation mechanism of a bio-epoxy resin (E-epoxy)

derived from bark extractives in comparison with a petroleum-based epoxy resin. The

thermogravimetric analysis (TGA) results show that the activation energy of E-epoxy varied

significantly with the extent of degradation indicating a multistage degradation mechanism

involving a variety of compounds. According to Fourier transform infrared spectroscopy

(FTIR) analysis, the dehydration and crosslinking reactions occurred at low temperatures,

while the Claisen chain rearrangement and chain-scission reactions dominated at high

temperatures. The pyrolysis-gas chromatography-mass spectrometry (Py-GC/MS) results show

that a significant amount of methyl abieta-8,11,13-trien-18-oate, diethyl phthalate, 2,2'-

isopropylidenebis(3,5-dimethylbenzofuran), and epimanool were detected in the bio-epoxy

resins. The newly proposed degradation mechanism of bio-epoxy resins based on structural

illustration through FTIR and Py-GC/MS can provide guidance for design of high performance

bio-based epoxies.

___________________________________________________________________________

A version of this chapter has been submitted to Journal of Analytical and Applied Pyrolysis,

JAAP-D-15-00326

84

5.1 Introduction

Epoxy resins have versatile molecular structures and well-established properties to match many

applications, ranging from traditional wood finishing to solar cell coating. However, due to the

health concerns of bisphenol A (BPA), a major ingredient of epoxy resins, there is a growing

interest in finding alternative synthetic precursors from renewable resources. In the last decade,

significant efforts have been made to substitute BPA with natural products such as vegetable

oils1-5, liquefied biomass6-7, lignin8-10, and rosin11-12. These bio-based polymers not only reduce

dependency on petroleum resources, but are also compatible with existing processes of epoxy

synthesis.

However, the use of bio-based epoxy resins at high temperatures limits their applications13-16.

For example, the aerospace industry has used epoxy resins for making fuselage parts for over

forty years, and the service temperature in that application is between 115°C and 311°C17. Most

current bio-based epoxy resins cannot withstand these high temperatures. In order to expand

the range of applications of bio-based epoxy resins, thermal stability must be considered as a

necessary factor. Nevertheless, there has been limited research focused on systematic studying

of thermal degradation behaviour of bio-based epoxy resins18.

Thermal degradation is widely studied via thermogravimetric analysis (TGA), which

characterizes the chemically stable temperature of materials and provides an understanding of

the degradation kinetics. Kinetic analysis of the dynamic TGA data can offer an estimation of

activation energy as a function of the conversion rate, and can also predict the degradation time

in an isothermal mode. As recommended by the ICTAC Kinetic Committee19, the integral

isoconversional method – Kissinger-Akahira-Sunose (KAS) – provides a more accurate

measurement than the Kissinger method or the Ozawa-Flynn-Wall method. However, the

chemical components of bio-epoxy resins are complex that cause difficulties on TGA data

interpretation.

In order to provide insight into the thermal degradation process and better link molecular

structure to thermal stability, the use of model compounds is commonly adopted to infer the

activation energy of epoxy resins20-22. Other techniques such as pyrolysis–gas

chromatography–mass spectrometry (Py-GC/MS) and FTIR can also provide degradation

information by detecting degraded volatile gas and chemical changes in the degraded solid

residuals, respectively. A detailed understanding of the decomposition pathways and

degradation by-products can offer directions for developing heat resistant bio-based epoxy

resin products.

85

The objective of this chapter was to investigate the thermal degradation mechanisms of a bark

extractive-based bio-epoxy resin that has shown to have competitive mechanical strength and

thermal performance to petroleum derived epoxy resin23. The apparent heat stability and mass

loss as a function of temperature were measured using TGA. The correlation between thermal

stability and molecular structure was investigated through the model compound approach,

FTIR and Py-GC/MS analyses. A new thermal degradation mechanism was proposed based on

the findings and literature.


5.2.1 Materials

Bark chips from lodgepole pine, Pinus contorta Douglas (Pinales: Pinaceae), infested by

mountain pine beetle were provided by FPInnovations. Model compounds (abietic acid and

alkali lignin), epichlorohydrin, and 1.0 M tetrabutylammonium hydroxide in methanol were

purchased from Sigma-Aldrich, Mississauga, ON, Canada. Sodium hydroxide, acetone,

dioxane, and chloroform were supplied by Caledon Laboratory Chemicals, Mississauga, ON,

Canada. All chemicals were used as received without further purification.

5.2.2 Extraction Procedure

Bark chips (400 g) and model compounds (50 g) were immersed in 1% (w/w) NaOH in water

with a solid to liquid ratio of 1:10 for 2 hours at 90 °C. Bark extractives and model compounds

extractives were collected after filtration using cellulose grade 42 filters supplied by Whatman

and the extractives were dried using a Yamato GB210 spray-drier operating under an inlet

temperature of 150 °C, an outlet temperature of 80 °C, and an air pressure of 0.1 MPa.

5.2.3 Synthesis Procedure

The synthesis method followed was described in detail in our previous research23. Briefly, dry

extractives (15 g) were combined with epichlorohydrin (150 g), tetrabutylammonium

hydroxide (2 mL), and dioxane (150 g) in a three-neck round-bottom flask and temperature

was raised to 60 °C. Sodium hydroxide solution (50% w/w) was slowly added in the solution

to catalyze the reaction. The synthesis proceeded for 6 hours under a nitrogen atmosphere.

5.2.4 Refining Procedure

The reaction products were diluted with excess acetone and filtered to remove salt by-products.

Phase separation was further conducted on the resulting solution using water and chloroform

in a separatory funnel. The chloroform-miscible compounds were retained. The unreacted

epichlorohydrin and residual solvents were removed using a rotary evaporator in an oil bath to

86

120 °C under reduced pressure.


5.2.5.1 Thermogravimetric Analysis (TGA)

The thermal stability of epoxy resins was investigated using a thermal gravimetric analyzer

(TGA-Q500, TA Instruments, USA). Approximately 10-15 mg of sample was weighed in a

platinum pan and operated under a continuous flux of air or nitrogen (60 cm3/min). Under

dynamic analysis, all of the samples were heated from room temperature to 700 °C with 5, 10,

15, and 20 °C/min ramp; under isothermal conditions, samples were heated under 100, 200,

300, 400, or 500 °C for 90 min, respectively.


Degraded epoxy residuals were studied using a Bruker Tensor 27 spectrometer in an

environmental chamber. Liquid samples were placed between two KBr crystal discs, while

solid samples were mixed with KBr powder to form a pellet. All spectra were recorded between

4000-600 cm-1 at a resolution of 4 cm-1 with 32 scans.

5.2.5.3 Pyrolysis-Gas Chromatography/Mass Spectrometry (Py-GC/MS)

The apparatus of Py-GC/MS consists of a multi-shot pyrolysis (PY-3030D Model, Frontier Lab

Ltd., Japan) coupled to a gas chromatography (MS5973 Model, Agilent Technologies, USA).

In each test, approximately 0.1-0.2 mg of samples were decomposed into volatile compounds

or fragments. The entire pyrolysis process was divided into three heating stages, 70 ºC-300 ºC,

300 ºC-450 ºC and 450 ºC-550 ºC with 20 ºC/min ramp. The carrier gas was helium at a flow

rate of 13.3 ml/min. The transfer line temperature was set at 320 ºC to keep the chemicals

volatile during transport from the pyrolyzer to the GC column. The stationary phase of the GC

column was dimethyl polysiloxane (UA-1 model, Frontier Lab Ltd., Japan). The GC

temperature was initially held at 50 ºC for 1 minute and then ramped to 320 ºC at 20 ºC /min

with a soak for 5 minute. The mass spectra were measured using electron impact ionization

energy of 70 eV. The mass detector was focused on 230 m/z at scan rate of 2 scans per second.

The data were searched using the NIST MS library.


TGA is used primarily to determine the composition of materials and to predict their thermal

stability at elevated temperatures in inert/air environment. Many researchers have reported that

thermal stability of epoxy resins is mainly affected by their monomeric structure. Therefore,

87

we looked first at the thermal stability of the monomers and then at the thermal stability of the

cured resins with network structures.

5.3.1 Thermal Stability of Monomers

5.3.1.1 Mass Loss of Monomers

The mass losses as a function of temperature for petroleum-based epoxy (P-epoxy) monomer

and extractive-based epoxy (E-epoxy) monomer and their derivatives are shown in Fig. 5-1 (a)

and (b). The P-epoxy monomer was used as the reference and its degradation occurred from

150 °C to 450 °C. There are three major degradation peaks in the derived mass loss curve of

P-epoxy, 307 °C (79%), 354 °C (12%), and 413 °C (9%). By contrast, E-epoxy monomer

degradation initiated earlier starting from 50 °C, and there is no single identifiable peak. The

continuous degradation process of the E-epoxy resins was likely caused by the complexity of

the bark extractive components. This broad decomposition pattern also implies that the E-

epoxy resin contains not only aromatic extractive compounds, but also aliphatic fractions,

which reduce the overall thermal stability.

Fig. 5-1 Thermal degradation of the P-epoxy monomer and E-epoxy monomer under nitrogen

environment by TGA, showing the (a) mass curves and (b) derivative of the mass loss curves.

In order to separate the broad peaks of E-epoxy monomer in the derivative TGA (DTGA)

results, Origin (version 8.6; Microcal Software Inc., Northampton, MA) was used to identify

six major peaks from the E-epoxy resins, shown in Fig. 5-2. The R2 of this fitting model is

0.999 and the residuals sum of square is only 0.06, which is lower than five-peaks fitting (0.14)

and seven-peaks fitting (0.21). Compared to Lorentzian and Voigt models, we found that Gauss

curves fit our experimental data best. These peaks, ranging from low to high temperatures, were

at 157 °C (24.3%), 242°C (26.2%), 280 °C (9.9%), 318 °C (14.2%), 352 °C (6.6%) and 391 °C

(18.8%), respectively. The percentage shown in the parentheses are the ratios of the single peak

88

area to the total area.

Fig. 5-2 Deconvolution of the multi-peaks of the E-epoxy monomer DTGA results.

To understand these deconvoluted peaks of the E-epoxy monomer, abietic acid-based epoxy

(abbreviated as A-epoxy) and alkaline lignin-based epoxy (abbreviated as L-epoxy) were

synthesized as model compounds, and their peaks were compared with the E-epoxy. Abietic

acid and alkaline lignin represent two major extractive categories, resin acid and degraded

lignin, respectively. Fig. 5-3 shows that A-epoxy had one major decomposition temperature at

285 °C and L-epoxy had two major decomposition temperatures at high temperature (358 °C

and 394 °C). The degradation temperature patterns of the model compounds suggest that the

fifth and sixth peak of the E-epoxy resins were from degraded lignin and the third peak was

from abietic acid. The heat-resistant index temperature (Ts) is given in Table 5-1. This value

was determined from the temperatures at 5% mass loss (Td5) and 30% mass loss (Td30) of the

sample as obtained by TGA, following Eq. (5-1):

Ts = 0.49 (Td5 + 0.6 (Td30 − Td5)) Eq. (5-1)

89

Table 5-1 TGA data of the bio-based epoxy monomers.

Td onset (°C) Td max(°C) Td5(°C) Td30(°C) Ts(°C) Char600 (%)

E-epoxy 158 261 113 228 89.2 4.0

L-epoxy 138 395 85 180 69.6 34.0

A-epoxy 228 285 92 242 89.2 3.0

The E-epoxy monomer had the same heat-resistance value as the A-epoxy monomer, which

implies that the major constituent of the E-epoxy monomer might have a similar structure to

abietic acid. In addition, the L-epoxy has a generally broad decomposition region akin to E-

epoxy. The degradation of the propanoid side chain of lignin, accounting for more than half of

the inter-unit linkages, occurs at a relatively low temperature between 230-350 °C24-25. For the

aromatic part, the degradation begins around 400 °C26.

(a)

(b)

Fig. 5-3 DTGA curves of the two model bio-based epoxy monomers:

(a) lignin-based monomer and (b) abietic acid-based monomer.

5.3.1.2 Degradation Kinetics of Monomers

The kinetic characterization of thermal degradation is important for a better understanding of

the structure-property relationship based on the activation energy (Ea) and the pre-exponential

factor. To avoid any premature assumptions of the reaction mechanism and to reduce the noise,

the Kissinger–Akahira–Sunose's (KAS) isoconversional equation was applied when observing

the evolution of the Ea at a given conversion rate. Note that the KAS method offers significant

improvement in the accuracy of the Ea value compared to the Ozawa–Flynn–Wall method19.

The basic assumption of this analysis is that the reaction rate at a constant conversion depends

only on temperature. In a kinetic analysis, it is generally assumed that the reaction rate can be

described by two functions, k(T) and f(α),

90

𝑑𝛼/𝑑𝑡=k(𝑇)𝑓(𝛼)=𝐴𝑒𝑥𝑝(−𝐸𝑎/𝑅𝑇)𝑓(𝛼) Eq. (5-2)

where k(T) is the rate constant and f(α) is the reaction model. When the heating rate (𝛽) is

constant, Eq. (5-2) can be rewritten as

𝑑𝛼/𝑑𝑇=(−𝐸𝑎/𝑅𝑇)𝑓(𝛼)/𝛽 Eq. (5-3)

As shown in Eq. (5-3), the isoconversional method offers an assumption-free estimate of the

Ea.

(𝑑𝑙𝑛/𝑑𝑇)(𝑑𝛼/𝑑𝑡)𝛼=𝐸𝑎,α/𝑅 Eq. (5-4)

where Ea,α is the activation energy at a given conversion. Note that the KAS method offers

significant improvement in the accuracy of the Ea,α value compared to the Ozawa–Flynn–Wall

method19.

It was found that the Ea of the P-epoxy monomer remained at a constant value of approximately

105 kJ/mol before 70% conversion, and then it abruptly increased to 200 kJ/mol (Fig. 5-4a).

The growing change in Ea was likely due to the carbonized bonds in the degraded molecules.

Compared to the P-epoxy monomer, the E-epoxy monomer shows larger changes in Ea and a

higher overall Ea after 60% conversion. The Ea of the E-epoxy monomer shows a two-stage

increase, which in turn suggests a complex multi-step degradation mechanism. Owing to its

multiplicity, neither nth order nor autocatalytic reactions were used to calculate the value of Ea.

Please note the average activation energy of E-epoxy (121 kJ/mol) is higher than that of P-

epoxy (118 kJ/mol). In addition, to better understand the degradation mechanism, the Ea of A-

epoxy and L-epoxy were also investigated, as shown in Fig. 5-4b. Due to its simpler structure,

A-epoxy had a more consistent Ea compared to E-epoxy resin and L-epoxy resins. Among the

three epoxy resins, the Ea of E-epoxy changes the most with temperature, from 64 kJ/mol to

240 kJ/mol, due to its complex structures. Because of the high viscosity of L-epoxy which

hindered the solvent drying process, the Ea can also be influenced by the remaining volatilized

solvent, which decreased the activation energy at 40% conversion27. However, since thermal

kinetic analysis is not species specific, it is difficult to draw a solid conclusion about which

molecules are in the resins. For this reason, we further explored the residuals after degradation

using FTIR.

91

(a)

(b)


TGA data.

5.3.1.3 FTIR Residual Analysis of Monomers after Isothermal Thermal Degradation

The structural changes of monomers at different thermal degradation stages were characterized

by FTIR as shown in Fig. 5-5. The tentative assignments of the peaks observed in the diverse

temperature spectra of the monomers are reported in Table 5-2.

P-epoxy monomer was relatively stable compared to E-epoxy monomer, which contained a

high percentage of alkane (2928 cm-1 and 2854 cm-1) and carbonyl groups (1724 cm-1). Based

on FTIR analysis, no functional groups were found to be significantly changed after P-epoxy

was heated up to 200 °C in a nitrogen environment. Note that the total mass loss of P-epoxy

had already reached 10% at 200 °C, but the only difference in the FTIR was the slight increase

of hydroxyl group absorption band caused by hydrolyzed ether groups. When the temperature

increased further, the intensity of epoxy group absorption band decreased until it vanished at

400 °C. As the P-epoxy monomer was heated up to 300 °C which gave a mass loss around

80%, a new carbonyl absorption band was observed at 1721 cm-1, which was assigned to the

isomerization of ether bonds.

Different from the FTIR spectrum of P-epoxy, the carbonyl absorption band of E-epoxy was

not caused by thermal degradation. The high concentration of carbonyl groups is a specific

feature of the bark alkaline extractives, owing to the oxidation reaction between polyflavonoids

and air during the extraction process 28-29. Furthermore, E-epoxy showed several changes in the

functional groups even at 200 °C, including an increase in the intensity of a carbon-carbon

double bond absorption (1642 cm-1), which indicates a dehydration reaction was occurring, and

a decrease of epoxy absorption (1101 cm-1, 912 cm-1, and 841 cm-1). When the temperature

92

reached 400 °C, the E-epoxy appeared to have carboxylic acid absorption from 3300 to 2500

cm-1. There are two general points noted in both monomers when tested under a nitrogen

atmosphere: (a) the intensity of the hydroxyl group absorption peak progressively increased

with the increasing temperature, and (b) at 400 °C the total mass loss of both monomers reached

90%, at that point most of the absorption features were removed from the fingerprint region.

On the other hand, the decomposition process in air was faster and showed differences when

compared to that in a nitrogen environment. The effect of oxygen led to changes in both the

reaction rate and kinetic order of the reaction. Even at 200 °C, two increasing absorption bands

were observed when heating in air, including the hydroxyl groups at 3448 cm-1 and carbonyl

group at 1655 cm-1, as shown in Fig. 5-5b. Many studies have indicated that hydroperoxide

groups are produced during heating of epoxy in an air environment, which in turn form a strong

carboxylic acid absorption band ranging from 3500 to 2500 cm-1 above or equal to 300 °C30.

These acid compounds might be formed from the degraded phenolic compounds. In E-epoxy,

the most significant absorption band after degradation at 100 °C was at 757 cm-1, which was

likely associated with the residual chlorine on the hydrolyzed products. The effect of the

residual chlorine on the thermal stability was found to be negative due to its low bond energy31.

At 200 °C, the growing absorptions at 940, 880, and 700 cm-1 suggest asymmetric and

symmetric carbon double bonds32. Most aliphatic C-H absorption bands had disappeared at 300

°C, and the carboxylic acid absorption appeared at 400 °C. Furthermore, there is no residual

char observed at the end of the degradation process, which is common for most epoxy resins33.

93

Fig. 5-5 FTIR spectra of the commercial monomer and bio-epoxy monomer after TGA in

nitrogen or air atmospheres at various temperatures: (a) P-epoxy under nitrogen, (b) P-

epoxy under air, (c) E-epoxy under nitrogen, (d) E-epoxy under air.

Table 5-2 Peak assignments of the FTIR spectra of monomers.

Wavenumber (cm-1) Peak assignments Ref.

3640-3530 υ (Ph-O-H) 34

3570-3200 υ (O-H); Hydrogen Bonded 33-34

3130-3006 υ (Ar-H) 33-34

2980-2780 υ C-(C-H) and υ C=(C-H2) 33

3300-2500 υ (O-H), Carboxylic acid 34

1740-1694 υ (C=O), Aldehyde or acid 33

1642-1580/1510-1450 υ (C=C-C), Aromatic ring 35

1475-1450/1380-1360 δ (C-H2) 33

1250-1237 υ (Ar-O-C) 35

1176-1169/1113-1108 δ Ar-(C-H) 33

1135-1024 υ C-(C-O) 33

915-905 υ (C-O), oxirane 34, 36

839-822 υ (C-O-C), oxirane 34, 36

814-807 δ (Ar-H) 34

783-621 γ (Ar-CH2) 34

757 υ (C-Cl) 35

υ:stretching; δ: in-plane bending; γ:out-plane bending

94

5.3.1.4 Py – GC/MS Analysis of Bark Extractives and Uncured E-epoxy monomer

Py-GC/MS can be used to analyze both small molecules and three-dimensional high molecular

weight polymers, including epoxy monomers and cured epoxies. In this section, the backbone

structure and thermal degradation mechanisms of E-epoxy monomer were identified and

proposed via Py-GC/MS.

The pyrolysis process for bark extractives was separated into two stages, 70-300 °C and 300-

550 °C (Fig. 5-6) in order to analyze the degradation reactions and offer some insights into

degradation mechanism. At lower temperatures, degraded compounds from major components

(holocellulose, lignin, fatty acid, terpenoids and polyphenols) of bark extractives were mostly

found. The first two low-molecular-weight volatiles were acetone and acetol, which are

common pyrolysates from cellulose37. The compound 3 was likely from degraded lignin. The

compound 4 was alcoholized epimanool, which was a diterpene compound and is typically

found in lodgepole pine barks38. The last and most abundant component at low temperature

stage was betulin. In birch, the content of betulin can be up to 30% of the dry weight of the

extraction39. At higher temperature stage, identifying the original compound is difficult due to

the chain rearrangement and cyclization reactions, which in turn generated abundant

naphthalene and phenanthrene.

Bark extractives (I) 70-300 °C

E-epoxy monomer (I) 70-300 °C

95

Bark extractives (II) 300-550 °C

E-epoxy monomer (II) 300-450 °C

E-epoxy monomer (III) 450-550 °C

Fig. 5-6 Py-GC/MS chromatogram of bark extractives and uncured bio-epoxy monomer

The E-epoxy contains a variety of pyrolysates, including 20 recognizable compounds based on

the Py-GC/MS spectra (Fig. 5-6). There were three stages for E-epoxy monomer pyrolysis: 70-

300 °C, 300-450 °C and 450 °C-550 °C. At lower pyrolysis temperatures, dehydroabietic acid

and BPA were identified as major products. Consistent with the TGA results, abietic acid is

one of the primary components of E-epoxy. Moreover, abietic acid can be increased in infested

lodgepole pine as a defense mechanism against mountain pine beetle attack40. Therefore, it is

highly possible that abietic acid is one of the major backbone structures of E-epoxy. The

bisphenol A/F/P structure seen in the Py-GC/MS spectra may come from reconstructed lignin

fractions or a recombination between acetone and phenols. It is very unlikely that the bio-based

epoxy resin contains some bisphenol A and F during synthesis. These compounds were

generated as byproducts during pyrolysis. Some by-products produced from synthesis process,

such as hydrolyzed epichlorohydrin and polymerized epichlorohydrin were also observed. In

96

the second temperature range (300-450 °C), BPA was still the dominant product, following by

dimethyl benzene and phenanthrene. Dimethyl benzene and phenanthrene are common

pyrolysates from lignin41. The existence of the dimethyl benzene suggests that lignin is another

important backbone structure of the E-epoxy monomer. In addition, para-dioxane diethanol

was detected, of which similar derivatives have been previously reported13. The last by-product

seen was allyl alcohol, one of the three most important degradation gases from oxirane

compounds42. Allyl alcohol is a more abundant by-product from oxirane than acetone, due to

the asymmetry of the aliphatic segment43. In the temperature range (above 450 °C), compound

32 appeared as the major product due to the large amount of sestererpenes in bark extractives38.

Other observed compounds were cycloheptatriene, xylene, trimethyl-benzene, naphthalene and

phenanthrene.

Table 5-3 Peak assignments of the pyrolysis products of bark extractives as detected by Py-

GC-MS.

Peak # Compounds Mw formula Main fragments m/z

1 Acetone 58.08 C3H6O 43, 58, 27

2 Acetol 74.08 C3H6O2 43, 31, 74

3 2-Methoxy-4-vinylphenol 150.17 C9H10O2 135, 107, 51

4

4-(3-Hydroxy-3-methylpentyl)-

3,4a,8,8-tetramethyldecahydro-

1-naphthalenol

310.51 C20H38O2 109, 189, 69

5 Betulin 442.72 C30H50O2 189, 207, 411

6 Toluene 92.14 C7H8 91, 65, 39

7 Hexamethylcyclotrisiloxane 222.46 C6H18O3Si3 207, 96, 191

8 p-Xylene 106.17 C8H10 91, 106, 77

9 Diethylbenzene 134.22 C10H14 119, 105, 134

10 Naphthalene,1,7-Dimethyl 156.22 C12H12 156, 141, 115

11 Phenanthrene, 3,6-Dimethyl 206.28 C16H14 206, 191, 102

12 Phenanthrene, methyl isopropyl 234.34 C18H18 219, 234, 204

97

Table 5-4 Peak assignments of the pyrolysis products of uncured bio-epoxy monomer as

detected by Py-GC-MS.


13 α-Chlorohydrin 110.54 C3H7ClO2 44, 61, 79

14 n-Butyl glycidyl ether 130.19 C7H14O2 29, 57, 41

15 1,2,4-Butanetriol 106.12 C4H10O3 75, 31, 45

16 2,2'-[1,2-Ethanediylbis

(oxymethylene)]dioxirane 174.19 C8H14O4 57, 31, 45

17 2,2'-Isopropylidenebis(3,5-

dimethylbenzofuran) 332.44 C23H24O2 317, 171, 210

18 Bisphenol P 346.19 C24H26O2 331, 158, 135

19 Bisphenol F 200.23 C13H12O2 107, 200, 77

20 Bisphenol A 228.29 C15H16O2 213, 119, 91

21 Methyl abieta-8,11,13-trien-18-oate 314.46 C21H30O2 239, 299, 43

22 Allyl alcohol 58.08 C3H6O 57, 29

23 1,4-Dimethyl-2-ethylbenzene 134.22 C10H14 119, 134, 91

24 1,4-Dioxane-2,6-diyldimethanol 148.16 C6H12O4 57, 31, 117

25 Bisphenol A 228.29 C15H16O2 213, 119, 91

26 Phenanthrene, methyl isopropyl 234.36 C18H18 219, 102

27 Cycloheptatriene 92.14 C7H8 91, 65, 39

28 p-Xylene 106.17 C8H10 91, 106, 77

29 1,2,4-Trimethylbenzene 120.19 C9H12 105, 120, 77

30 Naphthalene, 2,6-Dimethyl 156.22 C12H12 156, 115, 77

31 Phenanthrene, 2,7-Dimethyl 206.29 C16H14 206, 102

32

3-isopropyl-6a,10b-dimethyl-8-(2-oxo-2-

phenylethyl)dodecahydro-7H-

benzo[f]chromen-7-one

396.57 C26H36O3 105, 207, 77

5.3.1.5 Proposed Degradation Mechanism for Uncured E-epoxy monomer

Many studies have proposed possible degradation mechanisms for epoxy resins, both under

inert and air conditions42, 44. In general, there are two categories of reactions during thermal

decomposition: chain rearrangement reactions and cleavage reactions.

The chain rearrangement in epoxy systems is known to occur over a wide temperature range

98

through common rearrangement reactions such as isomerization and Claisen rearrangement43.

The isomerization of the epoxy group to an aldehyde occurs at around 300 °C44-45, and during

this reaction carbon dioxide is released from the carboxyl groups, water from the hydroxyl

groups, carbon monoxide from the weakly bound oxygen groups, and hydrogen from the

aliphatic and methoxy groups45. The Claisen rearrangement is expected to happen at higher

temperatures, and during this reaction allyl vinyl ether is converted to γ, δ-unsaturated

carbonyl46.

For the cleavage reactions, they can be divided into two categories: non-chain scission reactions

and chain scission reactions. The two ordinary non-chain scission reactions of epoxy systems

are dehydration, and dehydrogenation47 while the three typical chain scission reactions are

named based on the position at which they occur: random-chain scission, side-chain scission,

and end-chain scission48. Other possible reactions are etherification, crosslinking reactions and

cyclization reactions, which occur along with the aforementioned reactions. For example, a

chain broken by random-chain scission can immediately crosslink with other broken chains, or

it may lose small fragments from secondary reactions.

There are several thermal decomposition theories in the literature that link degradation

reactions of uncured epoxy resins with degradation by-products are summarized in Scheme 5-

135, 42, 44-45. Based on the starting reactions, we named the following three theories for epoxy

resins including chain rearrangement mechanism, etherification mechanism, and ring-opening

mechanism. The rearrangement reaction mechanism occurs initially followed by a cleavage

reaction, which produces the principal volatile compound of conventional epoxy - acrolein.

The second mechanism begins with an etherification reaction and then the products are

dehydrated or rearranged, followed by cleavage reactions, which produces allyl alcohol,

acetone, and simple phenolic compounds. With a cyclization reaction, the final products can

be relatively high molecular weight compounds, which may remain as residuals. The last

mechanism includes first a ring-opening reaction, followed by a rearrangement or a cleavage

reaction, and then produces allyl alcohol and acetone.

99

Scheme 5-1 Three dominant thermal degradation mechanisms of epoxy monomer. 35, 42, 44-45

These theories and our findings on both P-epoxy and E-epoxy are in good agreement, which

form the foundation of our newly proposed mechanism for uncured bio-epoxy monomer

(Scheme 5-2). Some important findings based on our FTIR and Py-GC/MS results are

summarized below: (1) In the low temperature ( ≤ 200 °C), there was an increasing amount of

hydroxyl groups in FTIR spectra and the E-epoxy changed from liquid to solid form, which

was likely caused by the etherification. As stated by the second theory, the first degradation

reaction is etherification at the very beginning. Previous study32 showed that the initial

temperature of crosslinking reaction begins from 220°C, which is close to our observed

temperature (200°C). (2) A strong chloride peak appeared in the FTIR absorption at 200 °C,

and volatile chlorine derivatives were detected by Py-GC/MS. (3) Double bond absorption

bands were observed at 200 °C using FTIR. Likely these unsaturated bonds were created from

dehydration reactions. Most studies36, 41, 43 agree that the dehydration reaction is dominant and

necessary for inducing further scission reactions. A thermal breakdown process is believed to

be initiated at the weakest link along the chain and an unsaturated bond can weaken the adjacent

C-O bond and initiate the scission reaction from there. These observations are consistent with

the theory. (4) The intensity of the aldehyde and alkane absorption bands dramatically

decreased on heating from 200 to 300 °C, which suggests decarbonylation and fragment

100

cleavage, respectively. These scission reactions can create small volatile gases such as acrolein,

allyl alcohol, and acetone. Allyl alcohol was detected by Py-GC/MS. Some studies showed that

these volatile products were mostly produced at 280 °C32. Additionally, these principal volatile

products can degrade further to other low molecular weight gases, such as monoxides, ethylene,

methane, and carbon dioxide. Although these volatile gases could be detected, the dominant

decomposition compounds were phenol and cresols based on the results of Py-GC/MS. (5)

From 350 to 550 °C, rearrangement reactions and cyclization were both observed on the lignin

unit and extractives. Compiling the above information, we propose the following mechanism

to describe the degradation behaviour of E-epoxy (the epoxidized dilignol and epoxidized

abietic acid were chosen as the example):

We applied the same concept developed from the theory 2 and 3 in our proposed mechanism.

The top reaction is a ring-opening reaction to open the epoxide group and produce allyl alcohol

(Compound 22 detected by Py-GC/MS). The second reaction is used to explain the chlorine

containing pyrolysates (Compound 13-15) since small amounts of chlorine were detected. The

third reaction is based on chain scission reaction and Claisen chain rearrangement to produce

xylene (Compound 28) and 1, 2, 4, - trimethylbenzene (Compound 29). The fourth reaction is

an etherification to further crosslink the bio-epoxy monomer and form a network structure,

which changes the sample form from liquid to solid. In the Scheme 2(V), we proposed A-epoxy

is one major component of E-epoxy. Its decomposition reactions include cleavage and

aromatization. In the Py-GC/MS analysis, compound 21 was detected at first heating stage (70-

300 °C) and compound 26 was detected at second heating stage (300-450 °C).

101

Scheme 5-2 Proposed degradation mechanism of E-epoxy resins (The solid coloured rounded

rectangular with the compound numbers refer to the results from Py-GC/MS; the yellow

highlight circles refer to the results from FTIR and the rest parts are supported by literature).

In an air environment, the theory mainly focuses on peroxide oxidation since oxygen can easily

initiate free radical reactions49, which was found using electron spin resonance during the

thermal degradation49. Free radicals can be detected even at low temperatures (e.g. 20 °C)50.

An increased number of free radicals are formed by the thermal cleavage of bonds, and then

they react with oxygen to yield peroxides or hydroperoxides. Park and Blount51 first suggested

the two mechanisms for these reactions, which were later modified by Lee44. The Scheme

below shows the yield of carbon dioxide, vinylphenol, and unsaturated acid. Based on the FTIR

residual analysis, acid absorption bands were observed at approximately 300 °C in an air

environment, the trials for which were run at a lower temperature than those under an inert

environment.

Scheme 5-3 Thermoxidation mechanism of epoxy

monomer.

44, 51

5.3.2 Thermal Stability of the Cured Epoxy Networks

5.3.2.1 Mass Loss and Derivative Mass Loss of the Cured Resins

102

In order to form a 3D network, epoxy monomers were cured with diethyl-methyl-benzene-

diamine (DMBD). Generally, the thermal stability of epoxy resins is mainly affected by the

structure of the curing agents. DMBD is an aromatic structure hardener, which was chosen in

this study to improve the thermal stability. Compared to uncured E-epoxy monomer, its

statistical heat resistance index temperature increased by 63 °C, which was also more than the

increase of P-epoxy (54 °C) with DMBD added. On the other hand, the crosslinking density is

a debatable factor affecting thermal stability shown by many studies52-55. Therefore, the effect

of crosslinking density on thermal resistance was not considered here.

In the thermal stability measurements, the most significant difference between the two samples

is at the initial stage, as shown in Fig. 5-7. The temperature of the onset degradation of E-epoxy

is 86 °C, much lower than that of P-epoxy, which can be explained by the aliphatic chains from

degraded lignin and diterpene. Overall, we found the thermal performance of P-epoxy is more

stable than that of E-epoxy. A single sharp derivative decomposition temperature of P-epoxy

appeared at 382 °C, while two derivative degradation peaks of cured E-epoxy resins were seen

at 330 °C and 392 °C. The second peak, at 392°C, is close to the degradation peak of HCN

bonds. In addition, since epoxy is rigid and brittle, the surface of epoxy resins is usually full of

cracks. These cracks can contribute more continued burning54, which is why the residuals of

epoxy resins were less than other thermoset resins.

Fig.5-7 Thermal degradation of the P-epoxy resin and E-epoxy resin under nitrogen:

(a) mass loss curves and (b) derivative mass loss curves.

103

Table 5-5 TGA data of the cured epoxy resins.

Td onset(°C) Td max(°C) Td5(°C) Td30(°C) Ts(°C) Char600 (%)

Cured E-epoxy 287 392 272 337 152 4.6

Cured P-epoxy 373 387 364 387 185 10.7

5.3.2.2 Degradation Kinetics of the Cured Networks

Following a similar analysis for epoxy monomer, these TGA data were converted by eq. 5-2 to

5-4 to obtain the value of Ea. To date, the kinetic models of cured epoxy resins are still under

debate56-58. Therefore, the model-free method, KAS, was applied in this study to obtain Ea

against the conversion rate. The Ea of these cured epoxy resins were found to be higher

compared to those of the monomers. For P-epoxy resin, its Ea kept a constant value (180-200

kJ/mol) before it reached 70% conversion, and then, it dramatically increased to 269 kJ/mol at

80% conversion. Compared to P-epoxy resins, cured E-epoxy exhibits lower Ea, except

between 50-65% conversion. Before 50% conversion, the Ea value of E-epoxy multi-stage

increased from 148 kJ to 192 kJ. Between 50% and 65% conversion, the Ea of E-epoxy was

higher than that of P-epoxy, which indicates that E-epoxy is relatively stable at certain

conversion stage, but the overall activation energy of E-epoxy (198 kJ/mol) is lower than that

of P-epoxy (211 kJ/mol). Moreover, the Ea value was dynamic in E-epoxy resins, which is

associated with the existence of various weak linkages that give rise to different mechanisms

of initiation. In the literature, the value for Ea can span from 63 kJ/mol to 267 kJ/ mol depending

on the components of the epoxy/amine systems59.

104

Fig. 5-8 Dependence of the Ea on the extent of conversion evaluated from non-

isothermal TGA data.

5.3.2.3 FTIR Residual Analysis of the Cured Resins after Isothermal Thermal

Degradation

Compared to the uncured epoxy monomer, large numbers of hydroxyl groups were observed

at room temperature. These were formed through the curing process. For the cured P-epoxy,

the major change of functional groups began at 300 °C. At 300 °C, the residuals showed fewer

hydroxyl groups (3570-3200 cm-1), fewer ether groups (1237-1108 cm-1/ 915 cm-1), more

amine groups (3400-3000 cm-1), more alkene groups (3032-2966 cm-1), and more carbonyl

groups (1740-1700 cm-1) than before heating. As mentioned before, dehydration reactions

occurred in the early stages of heating, which induced further scission reactions. Owing to the

decreased number of hydroxyl groups, the amine absorption band became apparent. Intact ether

absorption bands affirm that the C-N bond was weaker than the C-O linkage. At 400 °C, a

shoulder due to an amine absorption band was observed next to that of a hydroxyl group at

1628 cm-1. Note that there was no acid peak observed even at the high temperatures.

The cured E-epoxy started to degrade around 200 °C. From 200 to 400 °C, more free phenolic

compounds and carbonyl aldehydes were observed. Since E-epoxy has plenty of OH bonds that

can react with epoxy rings and form ether bonds, there were strong ether absorption bands seen

at 1300-1000 cm-1. Furthermore, bio-epoxy shows a strong absorption at 1000-650 cm-1, which

suggests the existence of aromatic C=C bonds35. E-epoxy has a strong acid peak that formed at

400 °C, and in air the acid peak showed up at 300 °C.

105

Fig. 5-9 FTIR spectra of the cured P-epoxy and cured E-epoxy after TGA in nitrogen or air

atomosphere at various temperatures: (a) P-epoxy under nitrogen, (b) P-epoxy under air, (c) E-

epoxy under nitrogen, (d) E-epoxy under air.

106

Table 5-6 Peak assignments of the FTIR spectra of cured resins.

Wavenumber (cm-1) Peak assignments Ref.

3645-3350 υ (O-H), Non-bonded hydroxyl; υ(Ph-O-H) 34

3570-3200 υ (O-H); Hydrogen Bonded 33

3400-3000 υ (N-H) 34

3032-2966 υ (C=C-H) 33, 35

2980-2780 υ (C-H) and υ(CH2) 33

1740-1700 υ (C=O), ketoaldehyde 33, 60

1650-1590 δ (N-H) 34, 61

1600/1505/1450 υ(Aromatic C=C) 35, 60

1237/1172/1108 υ (Ether C-O-C) 35

1033 δ (C-N) 33

915 δ (Oxirane C-O-C) 34

819/753 δ (Ar-H) 33, 60

876/819 δ (C-O-O-C) 34

721 δ (C-Cl) 35

5.3.2.4 Py-GC/MS Analysis of the Cured E-epoxy Resins

Once epoxy monomers are cured with an amine hardener, their networks are essentially

insoluble in any solvent, which makes many analytical techniques inaccessible for

characterization. To address this issue, pyrolysis analysis was used to characterize these

intractable materials. According to the literature20-21, 62, the major pyrolysis products of cured

epoxy resins are bisphenol A, dimethylaniline, isopropylphenol, acetaldehyde, cresols,

ethylene oxide, and allyl alcohol63. Other minor degradation products are quite complex and

more than 20 compounds have been reported63. This complexity is caused by random bond

breaking and forming during thermal degradation, which suggests that Py-GC/MS may have

challenges in offering precise molecular identification.

The major degradation products of E-epoxy resins between 70 °C to 300 °C were aniline,

epimanool, remaining surfactant, diethyl phthalate, and chlorobutane. Different to the E-epoxy

monomer, the principal products seen in Py-GC/MS were from the curing agent (DMBD). The

structure of DMBD had been rearranged from meta-substitution to para-substitution, which

implies the chain cleavages and isomerization reactions happened simultaneously. At low

107

heating degradation temperature, one major backbone structure of the E-epoxy monomer is

epimanool. However, epimanool and abietic acid are both mono-functional molecules, so some

part of the E-epoxy should contain linear fragments instead of a complete network. This can

partially explain the inferior thermal stabilities of E-epoxy. Diethyl phthalate is the other major

pyrolysate, which is a common degradation product from lignin64.

Between 300 °C and 450 °C, acrolein was found in E-epoxy resin and it is one of the principal

degradation gases in conventional epoxy resins63. The second major decomposition product

was 2-methyl-4-vinylphenol, likely breaking away from lignin41. At the last stage of pyrolysis

(450-550 °C), high amounts of stearic acid were observed, which is usually among the top five

most abundant components of extractives in lodgepole pine38. Previous study has already

confirmed that the amount of stearic acid is increased after attack by the mountain pine beetle40.

(I) 70-300 °C

(II) 300-450 °C

108

(III) 450-550 °C

Fig. 5-10 Py-GC/MS chromatograms of the cured bio-epoxy resins.

Table 5-7 Peak assignments of the cured bio-epoxy resins pyrolysis products detected by Py-

GC-MS.


33 1-Chlorobutane 92.57 C4H9Cl 56,41,27

34 Tributylamine 185.35 C₁₂H₂₇N 142,100,41

35 Diethyl phthalate 222.24 C12H14O4 149, 177,76

36 N4,N4-Diethyl-2-methyl-1,4-

benzenediamine 178.27 C11H18N2 163, 134,77

37 Epimanool 290.48 C20H34O 137, 81, 95

38 Acrolein 50.06 C3H4O 56, 27



41 Pentamethylaniline 163.26 C11H17N 163, 91, 131


benzenediamine 178.27 C11H18N2 163, 178, 134


benzenediamine 178.27 C11H18N2 163, 178, 134

44 Toluene 92.14 C7H8 91, 65, 39

45 Pentamethylaniline 163.26 C11H17N 163, 148, 91

46 2,3-Bis[(trimethylsilyl)oxy]propyl

(9Z,12Z)-9,12-octadecadienoate 498.89 C27H54O4Si2 73, 207, 147

5.3.2.5 Degradation Mechanism

Although the degradation behaviour of the cured epoxy resins has been extensively studied

since 195025, 27, 35, 44-45, 51, 65, a debate on the degradation mechanism has continued for about

two decades. A nearly complete framework surrounding the mechanism was proposed around

197035. In this theory, the thermal decomposition of epoxy starts from the dehydration of the

secondary alcohol, and is then followed by different scission reactions, including allyl-nitrogen

scission, aryl-methylene scission, and allyl-oxygen scission. As mentioned in the monomer

section, the next reactions include further crosslinking and Claisen rearrangements, which lead

109

to assorted phenols during thermal degradation.

Scheme 5-4 Three scission positions on vinylene ethers. (modified from Peterson-Jones66)

Among these scission reactions, allyl-nitrogen scission is most likely to occur. The weakest

bond among the monomers is the C-O bond. After reacting with the amine curing agent, the

weakest bond changed from C-O bond to C-N bond. Within a cured epoxy network, the C-N

bond dissociation energy is 65 kJ/mol, which is lower than that of the C-O linkage (70 kJ/mol)

33. Additionally, dehydration is a major reaction during degradation61 and it usually occurs next

to an allyl-nitrogen bond, which increases the probability of nitrogen scission reaction. These

nitrogen chain scission can release volatile compounds, such as ammonia and methylamine32.

However, due to the low sensitivity of early instruments in the past, the detection of polar amine

groups was an issue for early studies43, 45, which were unable to detect amine products during

thermal degradation. Consequently, many studies44, 62 incorrectly assumed that nitrogen

fragments stay as residuals on the epoxy and undergo charring. In our Py-GC/MS results, large

amounts of amine curing agents were detected at 350 °C. Allyl-oxygen scission can increase

the concentration of hydroxyl group as shown in the FTIR section and the splitting fraction can

cyclize for the further charring, which also contributes to the thermal stability.

(I) Allyl-Nitrogen Scission and Crosslinking

110

(II) Allyl-Oxygen Scission and Cyclization

(III) Aryl-Methylene Scission and Claisen Chain Rearrangement

Scheme 5-5 Three possible scission reactions on cured epoxy resins. (modified from

Peterson-Jones66)

There are two major theories that explain the thermal oxidative decomposition mechanism in

epoxy/amine systems, including (1) a free radical mechanism, namely peroxide theory in

Scheme 5-6, and (2) the Cope reaction in Scheme 5-7. Most researchers today agree that the

free radical mechanism dominates thermal oxidative degradation. In this mechanism, the epoxy

first forms an unstable double bond via a dehydration reaction and then is peroxidized at the

alpha carbon hydrogen to form an acid group. The acid group is subsequently dehydrated and

attacked by an active hydrogen to break the C-N bond. The above description matches our

FTIR results that also show an increase in the number of acid groups and amine groups over

time. In addition, no nitrile linkages were found between 1550 cm-1-1475 cm-1 or 1360 cm-1-

1290 cm-1 regions, which does not fit with the Cope reaction mechanism. Although Burton67

claimed that the free radical theory is not accurate because there was no effect from the addition

of antioxidants in epoxy/amine system, there was similarly no direct evidence to show that the

Cope reaction dominates the thermal oxidation reaction.

111

Scheme 5-6 Thermal oxidative mechanism based on the free radical mechanism. 35, 65

Scheme 5-7 Cope reaction proposed by Burton and Conley 67-68

112

5.4 Summary

This chapter investigated the thermal behaviour and degradation compounds of E-epoxy

monomer and its cured network. TGA, FTIR, and Py-GC/MS provide some insight into its

lower thermal performance in comparison with that of P-epoxy resin. For uncured E-epoxy

monomer, DTGA data shows a similar pattern to alkali lignin and abietic acid. Py-GC/MS

detected significant amounts of dehydroabietic acid and lignin fractions at relatively low

degradation temperatures (70-350 °C). Moreover, other extractive compounds from bark

skeletal structures during the degradation at these temperatures were identified as

cycloheptatriene, and diterpene. Based on the FTIR results, dehydration reactions occurred

when bio-epoxy monomer was heated up to 200 °C. There were also some carboxylic

compounds at high degradation temperatures (300-400 °C). On the other hand, cured E-epoxy

resin shows two distinct degradation peaks in the DTGA analysis, with one of the peaks

appearing at 392 °C, which represents the degradation of HCN bonds. Furthermore, Py-GC/MS

provided more detailed analysis of the cured E-epoxy resin compositions, such as diethyl

phathalic, epimanool, methoxy-vinyl phenol, and stearic acid. Therefore, the substandard

thermal performance of E-epoxy resin can be explained by the presence of these

monofunctionalized diterpene and fatty acid molecules.

113

References

1. B. S. Rao and A. Palanisamy, Eur Polym J, 2013, 49, 2365-2376.

2. S. Hirose, J Oil Palm Res, 2011, 23, 1110-1114.

3. H. Miyagawa, A. K. Mohanty, R. Burgueno, L. T. Drzal and M. Misra, J Polym Sci Pol Phys,

2007, 45, 698-704.


5. T. R. Cuadrado and R. J. J. Williams, Polym Commun, 1989, 30, 239-240.



8. C. Sasaki, M. Wanaka, H. Takagi, S. Tamura, C. Asada and Y. Nakamura, Ind Crop Prod,

2013, 43, 757-761.

9. E. Windeisen and G. Wegener, in Polymer science: A Comprehensive Reference, ed. M. M.

K. Matyjaszewski, ELSEVIER, Oxford, 2012, vol. 10.



11. Q. Q. Ma, X. Q. Liu, R. Y. Zhang, J. Zhu and Y. H. Jiang, Green Chem, 2013, 15, 1300-

1310.

12. X. Q. Liu, W. B. Xin and J. W. Zhang, Green Chem, 2009, 11, 1018-1025.

13. H. Nouailhas, C. Aouf, C. Le Guerneve, S. Caillol, B. Boutevin and H. Fulcrand, J Polym

Sci Pol Chem, 2011, 49, 2261-2270.

14. J. L. Qin, H. Z. Liu, P. Zhang, M. Wolcott and J. W. Zhang, Polym Int, 2014, 63, 760-765.

15. J. L. Chen, F. L. Jin and S. J. Park, Macromol Res, 2010, 18, 862-867.

16. S. J. Park, F. L. Jin and J. R. Lee, Mat Sci Eng a-Struct, 2004, 374, 109-114.

17. Q. Guo, ed., Thermosets: structure, properties and applications, Woodhead publishing

limited, 2012.

18. X. Feng, A. East, W. Hammond, Z. Ophir, Y. Zhang and M. Jaffe, J Therm Anal Calorim,

2012, 109, 1267-1275.



20. Paterson.Jc, V. A. Percy, R. G. F. Giles and A. M. Stephen, J Appl Polym Sci, 1973, 17,

1867-1876.

114

21. Paterson.Jc, V. A. Percy, R. G. F. Giles and A. M. Stephen, J Appl Polym Sci, 1973, 17,

1877-1887.

22. Y. L. Wang, X. Y. Wang, L. M. Liu and X. Y. Peng, J Mol Model, 2009, 15, 1043-1049.

23. P. Y. Kuo, M. Sain and N. Yan, Green Chem, 2014, 16, 3483 - 3493.

24. R. Wittkowski, J. Ruther, H. Drinda and F. Rafieitaghanaki, Acs Symposium Series, 1992,

490, 232-243.

25. W. Fiddler, W. E. Parker, Wasserma.Ae and R. C. Doerr, J Agr Food Chem, 1967, 15, 757-

&.

26. M. Brebu and C. Vasile, Cell Chem Technol, 2010, 44, 353-363.

27. D. W. Levi, L. Reich and H. T. Lee, Polymer Engineering & Science, 1965, 5, 135–141.

28. R. W. Hemingway and G. W. Mcgraw, J Liq Chromatogr, 1978, 1, 163-179.

29. G. Vázquez, G. Antorrena and J. C. Parajó, Wood Sci Technol, 1987, 21, 155-166.

30. N. Mailhot, S. Morlat-Theias, M. Ouahioune and J. L. Gardette, Macromol Chem Physic,

2005, 206, 575-584.

31. V. V. Korshak and S. V. Vinogradova, Russian Chemical Reviews, 1968, 37, 885-906.

32. N. Grassie, M. I. Guy and N. H. Tennent, Polym Degrad Stabil, 1985, 12, 65-91.

33. P. Musto, G. Ragosta, P. Russo and L. Mascia, Macromol Chem Physic, 2001, 202, 3445-

3458.

34. J. Coates, Appl Spectrosc Rev, 1998, 33, 267-425.

35. M. A. Keenan and D. A. Smith, J Appl Polym Sci, 1967, 11, 1009-&.

36. M. G.González, J. C. Cabanelas and J. Baselga, in Infrared Spectroscopy - Materials

Science, Engineering and Technology, ed. T. M. Theophanides, InTech, 2012, ch. 13.

37. Dekui Shen, Rui Xiao, Sai Gu and H. Zhang, in The Overview of Thermal Decomposition

of Cellulose in Lignocellulosic Biomass, eds. T. v. d. Ven and J. Kadla, InTech, Rijeka,

2013.

38. I. Backlund, M. Arshadi, A. J. Hunt, C. R. McElroy, T. M. Attard and U. Bergsten, Ind

Crop Prod, 2014, 58, 220-229.

39. B. Green, M. D. Bentley, B. Y. Chung, N. G. Lynch and B. L. Jensen, J Chem Educ, 2007,

84, 1985.

40. D. M. Shrimpton, Canadian Journal of Botany, 1973, 51, 527-534.

41. C. A. Mullen and A. A. Boateng, Fuel Process Technol, 2010, 91, 1446-1458.


115

43. D. P. Bishop and D. A. Smith, J Appl Polym Sci, 1970, 14, 205-&.

44. L. H. Lee, J Appl Polym Sci, 1965, 9, 1981-&.

45. L. H. Lee, J Polym Sci Part A, 1965, 3, 859-&.

46. L. Claisen and E. Tietze, Berichte der deutschen chemischen Gesellschaft (A and B Series),

1925, 58, 275-281.

47. J. C. Paterson-Jones, J Appl Polym Sci, 1975, 19, 1539-1547.

48. S. o. F. P. Engineers, ed. P. J. DiNenno, National Fire Protection Association Quincy, 2002,

ch. 7, pp. 110-131.

49. A. P. Meleshevich, R. Y. Strakovskaya and L. M. Kovalenko, Sov Atom Energy+, 1987,

62, 225-227.

50. C. E. Yuan, M. Q. Zhang and M. Z. Rong, J Mater Chem A, 2014, 2, 6558-6566.

51. W. R. R. Park and J. Blount, Ind Eng Chem, 1957, 49, 1897-1902.

52. T. Dyakonov, P. J. Mann, Y. Chen and W. T. K. Stevenson, Polym Degrad Stabil, 1996,

54, 67-83.

53. M. Iji and Y. Kiuchi, Polym Advan Technol, 2001, 12, 393-406.

54. S. V. Levchik and E. D. Weil, Polym Int, 2004, 53, 1901-1929.

55. L. Reich, Journal of Polymer Science: Macromolecular Reviews, 1968, 3, 49-112.

56. A. I. Lesnikovich and S. V. Levchik, J Therm Anal, 1983, 27, 89-94.

57. L. J. Xia, L. Zuo, X. L. Wang, D. P. Lu and R. Guan, J Adhes Sci Technol, 2014, 28, 1792-

1807.

58. L. Barral, J. Cano, J. Lopez, I. Lopez-Bueno, P. Nogueira, M. J. Abad and C. Ramirez, Eur

Polym J, 2000, 36, 1231-1240.

59. V. Bellenger, E. Fontaine, A. Fleishmann, J. Saporito and J. Verdu, Polym Degrad Stabil,

1984, 9, 195-208.

60. T. Sugita, Journal of Polymer Science Part C: Polymer Symposia, 1968, 23, 765-774.


62. Leisegan.Ec and A. M. Stephen, J Appl Polym Sci, 1970, 14, 1961-&.

63. S. Tsuge, H. Ohtani and C. Watanabe, Pyrolysis - GC/MS Data Book of Synthetic Polymers,

Elsevier, 2011.

64. J. Cho, S. Chu, P. J. Dauenhauer and G. W. Huber, Green Chem, 2012, 14, 428-439.

65. M. B. Neiman, A. S. Strizhkova, I. I. Levantovskaia, B. M. Kovarskaia, M. S. Akutin and

116

L. I. Golubenkova, J Polym Sci, 1962, 56, 383-&.

66. T. M. Paterson, R. K. Smalley and H. Suschitzky, Synthesis-Stuttgart, 1975, 187-189.

67. B. L. Burton, J Appl Polym Sci, 1993, 47, 1821-1837.

68. R. T. Conley, Thermal Stability of Polymers, Marcel Dekker Inc, New York, 1970.

118

Chapter 6 Influence of Nanocellulose Fibres (NCFs) on the

Curing Behaviours of Epoxy Resins

Abstract

Epoxy composites were prepared using diglycidyl ether bisphenol F and water-dilutable

diglycidyl ether bisphenol A with curing agents, polyoxypropylenediamine and

diethylmethylbenzenediamine, in water or dimethylformamide as a solvent. The influence of

NCFs and solvents on curing kinetics of epoxy composites was investigated. Curing kinetic

parameters were calculated using the model-fitting methods and the isoconversional method.

Among these, the Sestak–Berggren equation best fit the experimental data. Results indicated

that dimethylformamide decreased the reaction rate, whereas water revealed the opposite

pattern. NCFs catalyzed the reaction between bisphenol F resins and the aromatic curing agent.

___________________________________________________________________________

A version of this chapter has been published in European Polymer Journal (2013) 49:3778-

3787

119

6.1 Introduction

The push towards bio-based materials has been evident in the past decade. Numerous efforts

have been focused on developing green materials for various engineering applications,

including building supplies, automotive components as well as electronic devices. NCFs, one

of the representative bio-materials, has attracted a rapid growth of interest in both academia

and industry owing to its excellent engineering properties such as stiffness, Young‘s modulus,

and thermal stability. The high potential of NCFs enables NCFs-based materials to reduce the

usage of petro-based polymers. Although the usage of NCFs would provide significant interest

in terms of mechanical and thermal properties, the water sensitive nature of NCFs caused by

the presence of plentiful hydroxyl groups has limited its application. In order to increase the

usage of NCFs, the incorporation of NCFs with various thermoplastic 1-5 and thermosetting

polymers 6-10 has been developed for tailoring valuable properties, such as excellent optical

characteristics, high thermal conductivity as well as a low absorption of water .

Among polymer matrices, epoxy resins exhibit the lowest shrinkage and the highest diversity

of applications from rocket casings to dental fillings. Epoxies are an important class of high

performance thermosetting polymers and their favorable characteristics result in a global

market of $ 5.5 billion(USD) in 2011 11. The incorporation of nano-fillers into epoxy resins has

shown an excellent combination of high stiffness, strength, and good fatigue resistance.

However, a major drawback of fibre-reinforced epoxy is their incompatible interface, which

has led to a wide body of work in improving the adhesion of cellulose-based fibres with

conventional epoxy resins. Of the several methods proposed, solvent exchange as well as

surface modification 12-16, and water-based resins 17, are the most commonly used modification

to improve adhesion between the fibres and resins. Solvent treated nanocellulose fibres possess

lower surface energy 18, and thus the aggregation of fibres significantly decreased. Of the

commonly used solvents for NCFs 14, 19, dimethylformamide (DMF) was applied in this

research to clarify the solvent effect on NCFs/epoxy composites. In addition, due to the high

cost and long processing time of solvent exchange, water-based epoxy resins appear as a new

alternative, which can emulsify in water, and circumvents using organic solvents. Water-based

epoxy resins have good viscosity control which is an integral part of high surface area nano-

sized materials. These techniques, which were compared in this research, generally lead to a

successful adhesion improvement between the fibres and resins.

120

Although much attention has been given to the influence of these treatments on mechanical

properties 20-23, little is known about their influence on the curing behaviour of epoxy resins.

Understanding curing process is extremely important to obtain consistent products with the

desired physical and mechanical properties. As well, the curing process of epoxy resins can be

significantly affected by these interface treatments. For example, silane treated glass fibers

composites have lower activation energy than that with untreated fibers composites 24. Thus,

understanding the curing behaviour is necessary for manufacturing of NCFs/epoxy composites.

The curing kinetics information is useful in selecting the time-temperature curing regime, and

in controlling the structure of the 3-D network. From a theoretical point of view, many

researchers have proposed modeling techniques regarding curing kinetics of epoxy-based

composite systems 25. Modeling selection depends upon reaction types and resin structures.

The Kissinger equation is widely used to describe simplified first order reactions 26, while the

Sestak-Berggren empirical equation 27, and the Kamal equation 28 are commonly used to

describe autocatalytic systems, which consider hydroxyl groups as a catalyst. Recent kinetic

studies have focused on comparing model-fitting methods with isoconversional methods,

sometimes called model-free methods. The isoconversional methods follow the activation

energy as a function of the extent of reaction conversion without reaction mechanism

assumptions. Among these model-free methods, Kissinger-Akahira-Sunose method (KAS

method) is a common and accurate method that is widely used to describe the curing kinetic of

epoxy-amine systems. The methods mentioned above were compared to select a model that

more accurately describes the epoxy-amine systems curing behaviour and with a better fit to

the experimental data.

The objective of this chapter is to investigate the curing behaviour of NCFs/epoxy composite

systems. The influence of the type of epoxy, curing agent types and solvents was observed.

Significant changes in curing kinetic parameters were observed by means of differential

scanning calorimeter (DSC) and Fourier transform infrared spectroscopy (FTIR). The

correlation between NCFs/solvent addition and activation energy was described by the

Kissinger equation, the Sestak-Berggren equation, the Kamal equation and the KAS equation.


6.2.1 Materials

Two types of epoxy resins (EPON863, WD 510) and two types of amine hardeners (EPIKURE

121

3274, EPIKURE W) were supplied by Momentive Specialty Chemicals, OH, USA. EPON863

is a bisphenol F type epoxy resins, and its epoxy equivalent weight ranges 165-174 g/eq. WD

510 is a water-dilutable bisphenol A type epoxy resin, and its epoxy equivalent weight ranges

190-205 g/eq. EPIKURE 3274 (aliphatic type) and EPIKURE W (aromatic type), were

employed as diamine curing agents. Their major components are polyoxypropylenediamine

(POPDA) and diethylmethylbenzenediamine (DETDA), respectively and their amine hydrogen

equivalent weights are 76 g/eq and 45 g/eq, respectively. Dimethylformamide (DMF) was

purchased from Sigma–Aldrich Canada. These chemicals were used as received without further

purification. NCFs were lab-made and their diameter is between 5 and 60 nm. Additional

information on nanocellulose fibres preparation and characterization has been reported

elsewhere29.

6.2.2 Methods

The samples were prepared according to the formulations in Table 6-1. Prior to composite

preparation, the epoxy resin, curing agent, and nanocellulose fibres (2% w/w) were mixed for

5 min to ensure a homogeneous distribution. The composites were stored in desiccators in order

to prevent moisture absorption.


6.2.3.1 Differential Scanning Calorimetry (DSC)

The curing behaviour and glass transition temperature (Tg) of the epoxy composites were

evaluated with a differential scanning calorimeter (DSC) model Q 100 from TA instruments.

Dynamic DSC measurements were carried out at a ramp rate of 5, 10, 15, and 20°C/min from

30 to 300°C to obtain the curing heat-flow curves of liquid samples. Then, the cured samples

were heated up to 300°C at 10°C /min to obtain the Tg of the fully cured samples.


FTIR was performed on a Bruker Tensor 27 spectrometer with temperature controller – I0977

and was carried out with the samples between two KBr pellets. All FTIR absorption spectra

were recorded over 4000–400 cm-1 wavenumbers at a resolution of 4 cm-1 with 28 scans.

Table 6-1 Composition of samples

Samples*1 Resins Curing agents Solvents NCF*2

122

BiF-Aliph 69% BiF 31% POPDA N/A 0%

BiF-Aliph-DMF 35% BiF 15% POPDA 50% DMF 0%

BiF-Aliph-DMF-NCFs 33% BiF 15% POPDA 50% DMF 2%

BiF-Arom 79% BiF 21% DETDA N/A 0%

BiF-Arom-DMF 40% BiF 10% DETDA 50% DMF 0%

BiF-Arom-DMF-NCFs 38% BiF 10% DETDA 50% DMF 2%

BiF-Arom-NCFs 78% BiF 20% DETDA N/A 2%

WD-Aliph 72% WD 28% POPDA N/A 0%

WD-Aliph-Water 36% WD 14% POPDA 50% Water 0%

WD-Aliph-Water-NCFs 35% WD 13% POPDA 50% Water 2%

WD-Arom 81% WD 19% DETDA N/A 0%

WD-Arom-Water 41% WD 9% DETDA 50% Water 0%

WD-Arom-Water-NCFs 39% WD 9% DETDA 50% Water 2%

*1 Bisphenol F epoxy resins abbreviated as BiF; water-dilutable bisphenol A type epoxy resins abbreviated as WD;

aliphatic curing agents abbreviated as Aliph; aromatic curing agents abbreviated as Arom. *2 2% NCFs was calculated based the dry weight percentage


6.3.1 Curing Behaviour

The polymerization process can be monitored by FTIR and DSC. FTIR was used to monitor

the concentration of functional groups as the reaction proceeds via the evolution of

concentration of amino groups, hydroxyl groups or epoxide groups. The active hydrogen of the

amino groups was consumed by the epoxide group of the resin and formed hydroxyl groups

through the ring opening reaction. As shown in Fig. 6-1(a), it was observed that amino groups

(3380-3310 cm-1) decreased and hydroxyl groups (3600-3400 cm-1) increased. However, the

absorption peaks of amino and hydroxyl groups tend to overlap and affect the precision of the

calculation of kinetic parameters. Thus, in this study, the curing process was observed using

the epoxide absorption at 915 cm-1 (Fig. 6-1(b)). The signal reduction at 915 cm-1 shows the

consumption of epoxy groups. The spectra were normalized to the reference absorption at 1184

cm-1(not shown) that represents the C-C stretch at bisphenol 30.

123

Fig. 6-1 FTIR spectra obtained at different times for BiF-Aliph sample cured at 100°C.

(a) Spectrum from 4000-3000 cm-1, (b) Spectrum from 1000-500 cm-1

DSC is another useful technique for studying crosslinking reactions of thermosetting epoxy

resins. Each non-isothermal curve can provide onset cure temperature, Tonset, peak cure

temperature, Tp, Tg, percentage of conversion at peak-temperature and total reaction heat, △H.

The Tonset and Tp were both extrapolated temperature at heating rate=0. Table 6-2 showed the

influence of solvent (DMF or water) and NCFs on curing behaviours.

In the blank resins (without fibre and solvent), WD systems have higher Tg and △H than BiF

systems but both △H of blank resins falls in the typical range (100-118 kJ/mol), which is close

to the heat of epoxy ring opening 31. After the addition of DMF, Tonset and Tp were increased in

both BiF/Aliph and BiF/Arom systems but the addition of water decreased Tonset and Tp in

WD/Aliph and WD/Arom systems. This indicates DMF retards the reactions, but water acts as

a catalyst. Some research also indicated that the time of Tp was delayed and heat of reaction

decreased due to the addition of solvents, like acetone, tetrahydrofuran, as well as toluene in

an open system 32-33. The cure reaction was significantly accelerated with the addition of 3%

of water, which catalyzed the reaction of primary amines with epoxides to form secondary

amines 32-33.

124

Table 6-2 Thermal characteristics of samples from the dynamic DSC analysis

Samples Tonset (°C) Tp (°C) Tg (°C) △H (J．g-1) α at Tp (%)

BiF-Aliph 49 (0.94) 82 (0.99) 59±1.7 512 47

BiF-Aliph-DMF 67 (0.95) 112 (0.98) 69±2.8 359 34

BiF-Aliph-DMF-NCFs 48 (0.96) 121 (0.97) 69±3.9 494 25

BiF-Arom 119 (0.98) 156 (0.97) 133±2.5 606 50

BiF-Arom-DMF 156 (0.98) 240 (0.97) 140±3.7 523 54

BiF-Arom-DMF-NCFs 117 (0.94) 166 (0.88) 136±4.8 558 54

BiF-Arom-NCFs 115 (0.98) 140 (0.97) 140±3.9 630 56

WD-Aliph 61 (0.99) 85 (0.98) 67±1.7 557 48

WD-Aliph-Water 38 (0.99) 77 (0.95) 78±4.3 196 35

WD-Aliph-Water-NCFs 60 (0.98) 98 (0.99) 79±3.5 347 48

WD-Arom 95 (0.94) 134 (0.96) 161±6.1 568 48

WD-Arom-Water 87 (0.71) 123 (0.99) 131±8.3 392 45

WD-Arom-Water-NCFs 141 (0.96) 198 (0.96) 129±4.0 335 50

*Values in the bracket are the correlation coefficient.

Many researchers already concluded that hydroxyl-containing solvents, such as alcohols and

phenol, have a catalyzed activation and can decrease Ea for curing reactions of epoxy resins

[12, 35]. In this case, the epoxy ring carbon atom becomes more sensitive to nucleophilic attack.

The reaction proceeds through a trimolecular transition state initially suggested by Smith and

presented in Scheme 6-1 34. However, most of the studies focus on alcohol liquid compounds.

Little is known about the influences of hydroxyl-containing solid samples, such as

nanocellulose fibres. Thus, we applied the NCFs to investigate the curing behaviour of

epoxy/amine systems. It is notable that etherification can also happen between an epoxide ring

and a hydroxyl group of NCFs. Nevertheless, since we have not found significant evidence to

confirm an etherification reaction in our epoxy system via FTIR, we will focus on only the

catalytic effect of hydroxyl groups on epoxide rings.

125

Scheme 6-1 Trimolecular transition state suggested by Smith. 34

In Table 6-2, the addition of NCFs in BiF/Aliph, BiF/Arom and WD/Aliph systems decreased

Tonset which demonstrated its catalytic effect. Nevertheless, the Tp showed the opposite trend

due to the addition of solvents. Some researchers reported low amount of NCFs can decrease

Ea on the epoxy-amine cure as observed by a progressive decrease of exothermic peak enthalpy

and activation energy. In our case, no decrease of Tp was found but the Ea value of

NCFs/Epoxy/Amine system did decrease significantly, which will be discussed later.

Using MANOVA analysis, the Tonset and Tp showed the curing agent type is dominantly

affecting the curing behaviour (p<0.0001) compared to the resin type (p=0.011). In addition,

introduction of additives (solvent or NCFs) significantly influence the Tonset and Tp values

(p=0.005). Therefore, further investigations are necessary to clarify the effects of additives on

resin systems curing behaviours, which will be discussed later.

In thermosetting polymers, Tg is dependent on the crosslink density because the mobility of

chain segments is restricted due to the network structure. As shown in Table 6-2, Tg values was

increased after the addition of solvent or NCFs. A possible reason is that solvent can prevent

the verification occurring and NCFs can catalyze the curing reactions so both of the Tg values

increased slightly.

6.3.2 Cure Kinetics

6.3.2.1 Model Selection

As mentioned earlier, some studies analyzed kinetic parameters either by a nth order model or

autocatalyzed model of epoxy/amine systems 35. Generally, nth order models are preferred

when large amounts of hydroxyl groups are observed in the initial stage 23. To examine the

suitability of a reaction model, two methods are commonly used in epoxy/amine systems. The

126

simplest way is to plot the reaction rate versus the degree of conversion. As shown in Fig. 6-

2(a), although NCFs contained a high amount of hydroxyl groups, the curing system followed

autocatalytic reactions because the maximum reaction rate occurred when the degree of

conversion was close to 15% instead of at time = 0. After adding 2% NCFs and/or solvents into

resins, the cure behaviour can still be described by an autocatalytic model indicating that NCFs

did not affect the curing mechanism of epoxy/amine systems. This is corroborated by other

researchers that showed the autocatalytic character of epoxy-amine reactions is still preserved

despite the presence of hydroxyl groups 36.

Fig. 6-2 Model selection by (a) reaction rate and (b) Malek’s method

Secondly, a statistical method 37 was proposed to analyze the non-isothermal experimental data

for a thermally stimulated physical or chemical process. Two functions, y(α) and z(α), are used

to determine an appropriate kinetic model. The y(α) and z(α) functions can be expressed in

terms of,

y(α) = (𝑑𝛼/𝑑𝑡)exp(x) Eq. (6 − 1)

𝑧(𝛼) = 𝜋(𝑥) (𝑑𝛼

𝑑𝑡) (𝑇/𝛽) Eq. (6 − 2)

where x is the reduced activation energy (Ea/RT), and π(x) is related to the temperature integral,

which can be approximated well using the equation by Senum and Yang 38

𝜋(𝑥) = (𝑥3 + 18𝑥2 + 88𝑥 + 96)/(𝑥4 + 20𝑥3 + 120𝑥2 + 240𝑥 + 120) Eq. (6 − 3)

By plotting normalized y(α) and z(α), the shape of the function f(α) was obtained, as shown in

Fig. 6-2(b). The y(α) and z(α) can be used as a diagnostic tool for kinetic model determination.

127

The y(α) function has a maximum αM ≠0 and the z(α) has a maximum at αP =0.51. Based on

the selection guide 37, the empirical SB model and JMA model (exponent >1) can give the best

description of the data.

6.3.2.2 Isothermal Versus Constant Heating Rate Run

Diffusion control affects the calculation of kinetic parameters because it will increase the Ea

value which is assumed to be constant. As in many thermosetting polymers, diffusion and

vitrification occur in epoxy systems. Performing the analysis with a constant heating rate can

improve segment mobility, which lessens the effect of these phenomena. In Fig. 6-3, both

isothermal and dynamic modes were used to determine when the systems become diffusion

controlled by observing the onset of curve from linear to nonlinear. When the reaction

temperature is isothermal, the onset of diffusion control occurred at 30 - 40 % degree of

conversion, but in dynamic mode at 80 - 90 % of conversion. Thus, in the following kinetic

study, we used a non-isothermal analysis to determine the kinetic parameters to avoid the

influence from diffusion.

Fig. 6-3 Onset of diffusion control, w= 𝑙𝑛[(𝑑𝛼/𝑑𝑡)/(1 − 𝛼)𝑛 − 𝑘1]

6.3.2.3 The Kissinger Method

Among the various multiple heating rate methods, Kissinger equation is the most extensively

used method to calculate the activation energy. Based on the Kissinger equation for BiF/Aliph

and BiF/Arom systems, the Ea for the blank epoxy resins (without solvent and fibre) was 41.0

and 52.0 kJ/mol, which fell in the typical range of Ea for common epoxy-amine polymerization

systems (40–70 kJ/mol) 39. The addition of DMF only caused negligible effects on the Ea value,

44.3 and 50.5 kJ/mol, because DMF is a polar aprotic solvent which have good dielectric

128

constants and polarity and won’t form the solvent shell. The ring-opening mechanism of

epoxy/amine systems follows certain type of SN2 mechanism and DMF is commonly applied

in SN2 reactions to provide a naked nucleophile. Also, the catalyst effect of NCFs in terms of

providing proton will not be affected by the solvent.

Furthermore, the activation energy was dramatically affected by the addition of NCFs. After

the addition of NCFs, the Ea values of NCFs composites (40.3 and 37.6 kJ/mol) dropped lower

than the blank epoxy composites, which again reveals the catalytic influence of NCFs. When

compared to the NCFs in the neat system, DMF-NCFs can decrease the Ea values even more.

A possible reason is DMF may stabilize the trimolecular transition state mentioned before.

For water dispersible systems, the Ea for the blank WD resins (without solvent and fibre) was

44.8 and 35.4 kJ/mol, respectively. After adding water into the resin systems, the Ea value

dramatically decreased because water is an ideal proton provider to decrease the Ea value (29.1

kJ/mol) and also water can emulsify the resins and trap the resins and curing agent in a restricted

volume which sped up the reaction rate in early stage but reduced the efficiency in the later

stage. However, for the WD/Arom system, the activation energy increased by approximately

50 % which may be caused by the aromatic amine not being compatible with the water. Water

is a stronger catalyst than NCFs so the influence of NCFs became negligible in WD systems.

Therefore, the Ea of the BiF system decreased by adding NCFs and the Ea of WD systems

decreased by adding water.

129

Table 6-3 Cure kinetics parameters by Kissinger Eq. and Sestak-Berggren empirical Eq.

Kissinger Eq. Sestak-Berggren empirical Eq.

Samples Ea (kJ/mol) A (s-1) k (s-1) m n overall

BiF-Aliph 41.0 1.69E+03 2.99E-03 0.70 0.81 1.51

BiF-Aliph-DMF 44.3 2.04E+03 1.25E-03 0.44 0.83 1.27

BiF-Aliph-DMF-NCFs 40.3 3.87E+02 8.60E-04 0.42 0.78 1.20

BiF-Arom 52.0 2.75E+03 4.34E-04 0.72 0.60 1.32

BiF-Arom-DMF 50.5 4.71E+02 1.44E-04 0.88 1.11 1.99

BiF-Arom-DMF-NCFs 37.6 2.41E+01 2.91E-04 0.75 0.58 1.33

BiF-Arom-NCFs 44.0 4.13E+02 7.25E-04 0.97 0.83 1.80

WD-Aliph 44.8 6.99E+03 3.63E-03 0.63 0.66 1.29

WD-Aliph-Water 29.1 3.95E+01 3.27E-03 0.45 1.13 1.58

WD-Aliph-Water-NCFs 37.3 4.60E+02 2.69E-03 0.67 0.96 1.63

WD-Arom 35.4 4.88E+01 1.14E-03 0.62 0.67 1.29

WD-Arom-Water 54.0 1.69E+04 1.46E-03 0.67 0.53 1.20

WD-Arom-Water-NCFs 51.2 4.11E+02 8.25E-05 0.73 0.60 1.33

6.3.2.4 Model Free Method - Kissinger-Akahira-Sunose Method

The limitations of the Kissinger equation include that the reaction mechanism is assumed to be

nth order and the reaction kinetics follows single-step kinetic. Considering that the real Ea value

depends upon the conversion rate, the KAS equation was applied to observe the evolution of

the Ea value with conversion rate. Note that compared to the Ozawa–Flynn–Wall method, the

KAS method offers a significant improvement in the accuracy of the Ea value 40.

Fig. 6-4 exhibits the Ea at the different curing conversion α, determined by the KAS method

for non-isothermal data. It is noteworthy that the Ea of blank resins varies in a narrow range of

3.4 kJ/mol with respect to α, compared to the Ea range when solvent and NCFs are introduced.

The thermal curve also showed that the peaks in four blank resins are symmetric and uniformly

spaced which implied that the Ea values are constant. In addition, the constant Ea values implied

that the diffusion control had negligible influence on non-isothermal data of the blank resins.

Even though the influence of diffusion control in non-isothermal data started from 80-90 %

conversion by observing the onset of curve from linear to nonlinear, the transition from

chemical to diffusion control is very gradual in the KAS method.

130

Fig. 6-4 Dependence of the activation energy on the extent of conversion evaluated from non-

isothermal DSC data

After adding solvent, the increase of Ea was observed in BiF/Aliph, BiF/Arom and WD/Arom

systems. In the beginning of WD/Aliph system, the Ea value is lower than the blank resins due

to the catalyst effect of water. However, the Ea values increase continuously until 60 %

conversion, which may be due to the solvent shell effect inhibiting the network forming. After

combination with NCFs, the Ea showed different trends in the two curing agents. The NCFs

decreased the Ea in Arom systems but increased or remained constant in Aliph systems. This

implies NCFs can act as a catalyst when combine with low reactive curing agent in an epoxy

system. Compared to solvent/NCFs systems, dried NCFs decrease the Ea value (26 kJ/mol)

similar to that of solvent/NCFs systems initially but increased significantly because of the rise

of viscosity. The dramatic increase of viscosity has been reported in several researches 15.

6.3.2.5 Model Fitting Method - Sestak-Berggren Model and Kamal Model

Unlike isoconversional methods, the model fitting methods are capable of identifying multi-

step reaction models suitable for the description of complex kinetics. As we mentioned earlier,

after adding solvent or NCFs, peaks and shoulders are observed in the Ea curve, it is likely that

131

a process dominated by a multi-step model. In addition, according to the Malek’s method [37],

the empirical S-B model can give the best description of the data and Kamal is the most

common method to analyze autocatalytic reactions in literature.

According to the S-B equation (Table 6-3), the overall reaction order of BiF and WD system

ranged 1.20-1.99 and 1.20-1.63, respectively which lie inside the range reported in the literature

37. Compared to the S-B equation, the overall reaction order from Kamal equation is lower

(Table 6-4). In Kamal equation, two Ea were obtained, Ea1 representing the early-stage

autocatalytic reaction and Ea2, which affects the reaction after the initial autocatalytic stage. In

table 4, the Ea1 increased after adding the DMF but due to the presence of NCFs as an external

catalyst in epoxy/amine systems, the Ea1 decreased in BiF systems. The two activation energies

correspond to two rate constants k1 and k2. k1 values dropped in the order of 5 to 15 times

compared to k2, so k1 values were highly affected by the presence of solvent. k1 is the catalyst

rate constant, attributed to catalyst species initially present, such as impurities, fibres or

particles. k2 is autocatalyst rate constant that refers to catalyst species produced by reactions.

k2 values increased up to 4 times after adding NCFs in the BiF/Arom system but not in the

BiF/Aliph system, so NCFs behaved as a catalyst in the BiF/Arom system but as an inhibitor

in the BiF/Aliph system.

132

Table 6-4 Cure kinetics parameters by Kamal Equation

Samples

Ea1 Ea2 k1 k2

m n Overall (kJ/mol) (s-1)

BiF-Aliph 83.72 14.82 6.73E-03 1.20E+00 0.59 0.60 1.19

BiF-Aliph-DMF 86.54 18.00 5.33E-04 1.13E+00 0.41 0.45 0.86

BiF-Aliph-DMF-NCFs 80.03 67.18 1.42E-04 7.41E-02 0.49 0.48 0.97

BiF-Arom 67.54 69.13 2.43E-04 8.32E-02 0.57 0.49 1.06

BiF-Arom-DMF 113.40 49.49 7.03E-06 4.18E-02 0.60 0.51 1.11

BiF-Arom-DMF-NCFs 65.11 32.14 2.54E-04 1.26E-01 0.65 0.55 1.20

BiF-Arom-NCFs 62.05 31.59 1.34E-04 3.37E-02 0.69 0.66 1.35

WD-Aliph 90.80 27.94 6.39E-03 6.42E-01 0.56 0.49 1.05

WD-Aliph-Water 25.70 58.36 4.63E-03 2.04E+00 0.66 0.61 1.27

WD-Aliph-Water-NCFs 95.96 45.33 5.24E-03 1.17E+00 0.52 0.55 1.07

WD-Arom 88.78 46.22 7.21E-04 3.03E-01 0.52 0.48 1.00

WD-Arom-Water 136.01 57.02 5.75E-03 3.55E-01 0.53 0.61 1.14

WD-Arom-Water-NCFs 107.91 49.66 3.54E-06 2.86E-02 0.59 0.50 1.09

6.3.2.6 Prediction of Conversion

An important aim of this study was to predict the curing behaviour for epoxy resins. After

obtaining the values of kinetic parameters, we plotted the predicted rate curve for the different

systems in supplementary materials. The S-B model better fitted the experimental data,

especially at low conversion rates. However, due to solvent evaporation, blank epoxy/amine

systems are accurately predicted compare to solvent and NCFs systems. By comparison, the

reaction rate predicted by the S-B model is in good accordance with the experimental rate,

which substantiates that the S-B model is able to give the reliable kinetic rate for describing

the advancement of non-isothermal reactions.

Vyazokin and Lesnikovich 41 proposed a model-free method to predict the isothermal kinetics

from non-isothermal data. The method uses the following equations,

𝑔(𝛼) = 𝑡𝛼𝐴𝛼𝑒𝑥𝑝(−𝐸𝑎/𝑅𝑇0) Eq. (6 − 4)

𝑔(𝛼) = (𝐴𝛼 /𝛽) ∫ 𝑒𝑥𝑝𝑇𝛼

𝑇0

(−𝐸𝑎/𝑅𝑇)𝑑𝑇 Eq. (6 − 5)

133

Eq. (6-4) and (6-5) stand for isothermal cure at Tc=T0 and non-isothermal cure at heating rate

β, respectively. To obtain tα, Eq. (6-6) can be divided by Eq. (6-5).

tα = [∫ expTα

T0

(−𝐸𝑎/𝑅𝑇)dT] /βexp(−𝐸𝑎/𝑅𝑇0) Eq. (6 − 6)

Eq. (6-6) enables one to determine the time at a given conversion. Fig. 6-5 showed the

isothermal epoxy cure at 140 and 170 °C as predicted by using the Ea dependence derived

earlier. We can see the model prediction using the KAS method gave satisfactory agreement

with the experimental isothermal measurements before 80% conversion rate.

Fig. 6-5 Model-free prediction of isothermal cure at 140°C and 170 °C from KAS method.

(The experimental data are shown by line. The points correspond to the predicted time.)

134

6.4 Summary

The curing kinetics of four resins systems (BiF/Aliph, BiF/Arom, WD/Aliph and WD/Arom)

in the presence of diluents (DMF and water) and NCFs were examined by DSC. The evolution

of the maximum reaction rates revealed that the autocatalytic model is better suited than the

nth order model to describe the kinetic behaviour of the resin systems. Also, the non-isothermal

data can avoid the influence of diffusion control. The Ea values derived from the Kissinger

equation, KAS equation and the Kamal equation decreased in the presence of NCFs. However,

the opposite was observed after the addition of DMF. In other words, the catalyst effect of

NCFs was observed in bisphenol F systems; but, no catalytic effect of NCFs was observed in

water-diluable systems because the presence of water dramatically decreased the Ea value in

WD/Aliph systems. Compared to Kamal model, the S–B model was in better agreement with

the experimental data. In addition, the model-free approach could provide the isothermal curing

time of the epoxy system with NCFs addition.

135

References

1. Q. Cheng, S. Q. Wang, T. G. Rials and S. H. Lee, Cellulose, 2007, 14, 593-602.

2. J. Lu, T. Wang and L. T. Drzal, Compos Part a-Appl S, 2008, 39, 738-746.

3. A. N. Nakagaito, A. Fujimura, T. Sakai, Y. Hama and H. Yano, Compos Sci Technol,

2009, 69, 1293-1297.

4. M. Sain and K. Oksman, Cellulose Nanocomposites: Processing, Characterization, and

Properties, 2006, 938, 2-8.

5. G. Siqueira, J. Bras and A. Dufresne, Biomacromolecules, 2009, 10, 425-432.

6. M. Henriksson and L. A. Berglund, J Appl Polym Sci, 2007, 106, 2817-2824.

7. S. Iwamoto, A. N. Nakagaito and H. Yano, Appl Phys a-Mater, 2007, 89, 461-466.

8. A. N. Nakagaito and H. Yano, Cellulose, 2008, 15, 323-331.

9. M. O. Seydibeyoglu and K. Oksman, Compos Sci Technol, 2008, 68, 908-914.

10. Y. Shimazaki, Y. Miyazaki, Y. Takezawa, M. Nogi, K. Abe, S. Ifuku and H. Yano,


11. A. M. Intelligence, Global Epoxy Resin Market By Application & Geography – Forecasts

up to 2017, 2012.

12. A. Sbiai, A. Maazouz, E. Fleury, H. Sautereau and H. Kaddami, Bioresources, 2010, 5,

672-689.

13. J. E. Ehlers, N. G. Rondan, L. K. Huynh, H. Pham, M. Marks and T. N. Truong,

Macromolecules, 2007, 40, 4370-4377.

14. N. E. Marcovich, M. L. Auad, N. E. Bellesi, S. R. Nutt and M. I. Aranguren, J Mater Res,

2006, 21, 870-881.

15. C. Gousse, H. Chanzy, M. L. Cerrada and E. Fleury, Polymer, 2004, 45, 1569-1575.

16. M. Tizzotti, A. Charlot, E. Fleury, M. Stenzel and J. Bernard, Macromol Rapid Comm,

2010, 31, 1751-1772.

17. M. M. Ruiz, J. Y. Cavaille, A. Dufresne, C. Graillat and J. F. Gerard, Macromol Symp,

2001, 169, 211-222.

18. S. Noorani, J. Simonsen and S. Atre, Cellulose, 2007, 14, 577-584.

19. S. Fujisawa, Y. Okita, T. Saito, E. Togawa and A. Isogai, Cellulose, 2011, 18, 1191-1199.

20. A. N. Netravali, X. Huang and K. Mizuta, Adv Compos Mater, 2007, 16, 269-282.


136

22. H. Q. Liao, Y. Q. Wu, M. Y. Wu, X. R. Zhan and H. Q. Liu, Cellulose, 2012, 19, 111-

119.

23. A. Omrani, L. C. Simon and A. A. Rostami, Mat Sci Eng a-Struct, 2008, 490, 131-137.

24. D. Olmos, A. J. Aznar, J. Baselga and J. Gonzalez-Benito, J Colloid Interf Sci, 2003, 267,

117-126.

25. A. Yousefi, P. G. Lafleur and R. Gauvin, Polym Composite, 1997, 18, 157-168.

26. H. E. Kissinger, J Res Nat Bur Stand, 1956, 57, 217-221.

27. J. B. Sestak, G., Thermochim, 1971, 274, 173-177.

28. S. Sourour and M. R. Kamal, Thermochim Acta, 1976, 14, 41-59.

29. S. Janardhnan and M. M. Sain, Bioresources, 2006, 1, 176-188.

30. D. W. Sohn and K. J. Ko, Korea Polym J, 1999, 7, 181-188.

31. B. A. Rozenberg, 1986, Advances in Polymer Science Volume 75, 113-165.

32. S. G. Hong and C. S. Wu, Thermochim Acta, 1998, 316, 167-175.

33. S. Choi, A. P. Janisse, C. H. Liu and E. P. Douglas, J Polym Sci Pol Chem, 2011, 49,

4650-4659.

34. I. T. Smith, Polymer, 1961, 2, 95-108.

35. J. M. Barton, Adv Polym Sci, 1985, 72, 111-154.

36. F. X. Perrin, T. M. H. Nguyen and J. L. Vernet, Macromol Chem Physic, 2007, 208, 718-

729.

37. J. Malek, Thermochim Acta, 1992, 200, 257-269.

38. G. I. Senum and R. T. Yang, J Therm Anal, 1977, 11, 445-449.

39. N. A. Stjohn and G. A. George, Prog Polym Sci, 1994, 19, 755-795.



41. S. Vyazovkin and C. A. Wight, Int Rev Phys Chem, 1998, 17, 407-433.

138

Chapter 7 Using Nanocellulose Fibres to Develop a High

Performance Blended Petro-Epoxy/Bio-Epoxy Composite

Abstract

In order to develop a strong and ductile bio-based nanocomposite, the present chapter

incorporates 25 wt.% nanocellulose fibres (NCFs) and 10-30 wt.% bio-epoxy monomer (E-

epoxy) into commercial epoxy resin and outlines its mechanical and thermal characterization.

With 10% E-epoxy, the toughness of neat epoxy resins improved 84 %; after embedding NCFs,

the tensile strength and modulus of composites increased approximately twofold and fourfold,

respectively. In a nitrogen environment, the maximum degradation peak of the P-epoxy

composites was detected 24 °C higher than that of neat P-epoxy. The morphological analyses

demonstrated that both P-epoxy and E-epoxy fully covered the NCFs, indicating a superior

interface between them. The average of fibre cluster diameters in commercial resin composites

was significantly larger than that in E-epoxy composites, indicating that E-epoxy can better

penetrate into the NCFs clusters, and form a strong bond with the reinforcements. Thus, this

study provides a unique combination (E-epoxy/P-epoxy/NCFs) of superior mechanical

performance, high thermal stability, and good environmental sustainability.

139

7.1 Introduction

Deriving chemicals from bio-resources is an inevitable trend since the issue of future petroleum

shortage is pressing. Many studies have reported new building blocks to replace conventional

polymers1-2. One current academic research interest concerns substituting BPA with natural

compounds, such as saccharides,3, vegetable oils4, polyphenols5, lignin6 and liquefied

biomass7. BPA is the raw material for epoxy resins and polycarbonates, and has potential risks

to human health8. Currently, the available commercial bio-based epoxy resins include sorbitol

polyglycidyl ether (DENACOL™ and ERISYSTM) and oil-based epoxy resins (Vikoflex®,

Super Sap™, and Senso™). Although many bio-based epoxies seem promising, owing to their

sustainability, their mechanical properties are inadequate for the high-performance material

industry.

In order to address this issue, introducing high strength reinforcements into the epoxy resins is

the most common approach. In conjunction with the concept of green materials, NCFs emerged

as the most promising bio-fibre to enhance the mechanical properties of the matrix. NCFs are

renewable, lightweight, cost-effective, have high aspect ratio, and high specific

strength/modulus (100-160 GPa9), and can be obtained from various resources, such as algae10,

tunicate11, bacteria12, agricultural residuals13, and wood14. Beyond these features, previous

studies indicate that NCFs can shorten the curing process of epoxy resins,15 and that their

mildly-acidic surface shows potential to react with amine type epoxy curing agents16. Thus,

NCFs can be an ideal reinforcement for epoxy systems.

However, due to the high viscosity of NCFs suspension, most studies17-19 add low

concentrations of NCFs into epoxy matrices, which result in a limited improvement on

mechanical performance20-22. To increase the fibre loading, Yano's and Berglund's research

teams applied NCFs paper into epoxy resins, which significantly improved the mechanical

performance of P-epoxy23-24. In addition, several studies used silane25, titanate25, and (2,2,6,6-

tetramethylpiperidin-1-yl) oxyl (TEMPO)18 to pretreat fibres in order to improve the

mechanical performance of P-epoxy. However, the use of NCFs in bio-epoxy resins is hardly

found,18 and to the best of our knowledge, no mechanical improvement of NCFs/bio-based

epoxy is reported. Furthermore, no effect of NCFs on blended P-epoxy/E-epoxy system was

ever investigated.

140

The major advantage of using a resin blend system is its increased toughness compared to

conventional epoxy resins, because the added monomer may reduce the crosslinking density

by inserting into the initial repeating units, or they may form a second network that can create

phase separation. Furthermore, the presence of polar groups in the second phase monomer can

significantly improve the adhesion properties of epoxy resins26. Previous studies indicate that

bio-epoxy resins can improve toughness and adhesion properties27-28. Among all the bio-epoxy

resins, the E-epoxy presents low cost, high mechanical performance, high thermal resistance,

and contains significant amounts of hydroxyls29, as a potentially inexpensive alternative epoxy

resin system. Therefore, the potential impact of adding E-epoxy is a new field requiring more

detailed investigations.

The aim of this chapter is to better understand the reinforcement effect of NCFs on an

epoxy/bio-epoxy blend system. The curing behaviour of both epoxy resins was studied by

differential scanning calorimetry (DSC). The interface between fibre/matrix and fibre

diameters were investigated using a scanning electron microscope (SEM), and the mechanical

property was evaluated using a universal tester. The thermal stabilities of neat resins and

composites were further analyzed using thermogravimetric analysis (TGA).


7.2.1 Materials

NCFs and E-epoxy monomer were both lab-made, and more information regarding preparation

and characterization has been reported elsewhere29-30. Note that the NCF used in this study is

non-acid hydrolyzed cellulose, which is more flexible to form physical entanglements than

nanocellulose whiskers. The commercial epoxy monomer, Epon Resin 863, was used as a

control sample, and its epoxy equivalent weight is 165 - 173 g/eq. Epikure W was applied to

cure epoxy monomer, and its major component is diethylmethylbenzenediamine (DETDA).

The epoxy monomer and the curing agent were obtained from Momentive Specialty Chemicals,

OH, USA. Acetone (> 99.5%) was purchased from Caledon Laboratory Chemicals, ON,

Canada.

7.2.2 Methods

7.2.2.1 NCFs Paper Preparation

10.7g gel-state NCFs (2.5 wt. %) was diluted to reach 0.1 wt. % concentration, and bubbles

were removed from the solution using an ultrasonic bath. To form a paper shape, the suspension

was vacuum-filtered through a Supor-100 membrane filter (pore size = 0.1 µm, Pall

141

Corporation) in a Buchner funnel. After filtration, the wet films were carefully removed from

the membrane, and wet-pressed following a standard paper-drying process. Briefly, the wet

film was placed between two filter membranes and compressed by a dryer with pressure of

approximately 50 psi for seven minutes. The NCFs paper was then removed from the

membrane, and dried in the oven (DNCF) or immersed in acetone overnight (SNCF) in order

to remove all the water, prior to the composite preparation.

7.2.2.2 Composite Preparation

The sample compositions are listed in Table 7-1. All the samples were cut into a dog-bone

shape for tensile testing (ASTMD 638-5) from the DNCF and SNCF and were immersed in

epoxy resins overnight. To reduce the viscosity of the epoxy resins, the mixture of the resin

monomer/curing agent was diluted with acetone. Next, the samples were removed from the

bath, and cured in an oven with the following curing schedule: 1h at 65˚C, 1h at 80˚C, 1h at

121˚C, and 2h at 177˚C for post-cure. All the samples were prepared with one single layer of

NCFs.

Table 7-1 Sample compositions

Monomers

(ratios of P-epoxy : E-epoxy)

Curing

agents

NCFs*1

P-epoxy 79 % (100:0) 21 % -

10%E-epoxy 79 % (90:10) 17 % -

P-epoxy/DNCF 59 % (100:0) 16 % DNCF 25 %

10%E-epoxy/DNCF 62 % (90:10) 13 % DNCF 25 %

P-epoxy/SNCF 59 % (100:0) 16 % SNCF 25 %

10%E-epoxy/SNCF 62 % (90:10) 13 % SNCF 25 %

20%E-epoxy/SNCF 64 % (80:20) 11 % SNCF 25 %

30%E-epoxy/SNCF 66 % (70:30) 9 % SNCF 25 %

*1 DNCF = Oven-dried nanocellulose fibre; SNCF = Solvent-exchanged nanocellulose fibre


7.2.3.1 Activation Energy (Ea) and Glass Transition Temperature (Tg)

The Ea and Tg of the epoxy composites were evaluated using a DSC model Q 100 from TA

instruments. Dynamic DSC measurements were carried-out at a ramp rate of 5, 10, 15, and 20

°C/min, from 30 to 300 °C in order to obtain the curing heat-flow curves of liquid samples.

The cured samples were heated to 300 °C at 10°C /min in order to detect the Tg of neat epoxy

and composites.

142

7.2.3.2 Tensile Mechanical Properties

The tensile tests of the specimens were performed using an Instron machine, Model 3367,

equipped with a 2 kN load cell. Samples were cut into dog-bone shaped specimens using an

ASTM specimen-cutting die. The sample thickness was determined using a bench top model

micrometer. The span was 25 mm, with a speed of 2.5 mm/min. Five specimens were measured

for each composite type.

7.2.3.3 Morphology

SEM images were used to study the morphology of NCFs and its composites. All of the images

were acquired using a JEOL JSM6610-Lv with an accelerating voltage 10-15 kV. To visualize

the distinct features of each sample, NCFs papers were frozen by liquid nitrogen to maintain

their surface structure, and avoid cell collapse. Composite samples were examined on their

cross-sections, which were broken by tension force. All the samples were coated with gold.

7.2.3.4 Thermal Stability

The thermal stability of epoxy resins and composites were investigated using a TGA, Q500,

TA Instruments, USA. Approximately 10-15 mg of sample was weighed in a platinum pan,

and operated under a continuous flux of air or nitrogen (60 cm3/min). Under dynamic analysis,

all of the samples were heated from room temperature to 700 °C, with 10 °C/min ramp.

7.3. Results and Discussion

7.3.1 The Effect of Nanocellulose Fibres on Activation Energies of Epoxy Resins

The kinetic characterization of curing behaviour is important to better understand the influence

of NCFs surfaces on network-forming, based on the activation energy (Ea) and the pre-

exponential factor. To avoid any premature assumptions of the reaction mechanism, and to

reduce noise, the Friedman differential isoconversional method was applied to observe the

evolution of the Ea at a given conversion rate. Compared to the Kissinger-Akahira-Sunose and

Ozawa–Flynn–Wall method, the Friedman method offers a significant improvement in the

accuracy of the Ea value31-32. The basic assumption of this analysis is that the reaction rate at a

constant conversion depends only on temperature. In a kinetic analysis, it is generally assumed

that the reaction rate can be described by two functions, k(T) and f(α),

𝑑𝛼/𝑑𝑡 = k(𝑇)𝑓(𝛼) = 𝐴𝑒𝑥𝑝(−𝐸𝑎/𝑅𝑇)𝑓(𝛼) Eq. (7-1)

where k(T) is the rate constant, and f(α) is the reaction model. When the heating rate (𝛽) is

constant, Eq. (7-1) can be rewritten as:

143

𝛽 (𝑑𝛼/𝑑𝑇) = 𝐴𝑒𝑥𝑝 (−𝐸𝑎/𝑅𝑇)(𝛼) Eq. (7-2)

Then, this equation can be easily taken the logarithmic derivative of the reaction rate at

conversion equal to constant:

ln [𝛽(𝑑𝛼/𝑑𝑇)]= ln (𝐴) −𝐸𝑎, α/𝑅𝑇 Eq. (7-3)

where 𝐸𝑎, α is the activation energy at a given conversion.

Fig. 7-1 exhibits the 𝐸𝑎 at the different curing conversion α, for both neat epoxy (solid lines)

and NCFs reinforced epoxy (dotted lines). Since the NCFs contained around 98 % acetone, the

overall 𝐸𝑎 was expected to increase due to the dilution effects on the concentration of the

reactants. Nevertheless, the DSC results showed that the 𝐸𝑎 decreased as a result of the catalytic

effect of NCFs. For the P-epoxy, the 𝐸𝑎 was decreased from 60.6 to 36.6 kJ/mol, while the 𝐸𝑎

of E-epoxy was decreased from 45.4 to 35.8 kJ/mol. The E-epoxy monomer had a lower 𝐸𝑎

than the P-epoxy monomer because of the existing hydroxyl groups in the E-epoxy resins,

which have been studied using FTIR and NMR in our previous research29. The DSC results

indicate that NCFs can act as a catalyst for both epoxy resins to reduce the Ea to approximate

36 kJ/mol.

In addition, when there are excess amounts of epoxy groups, the reaction between epoxy and

hydroxyl group can occur at elevated temperatures as shown in Scheme 7-1. The possibility of

the etherification has been reported by other reports which can explain the good interface

between resins and fibres.

Fig. 7-1 Dependence of the 𝐸 𝑎 on the extent of conversion evaluated from non-isothermal

144

DSC data.

Scheme 7-1 Reaction between epoxy and hydroxyl groups: (a) catalyst effect and (b)

etherification.

7.3.2 Mechanical Performance of Epoxy Resins and Epoxy/NCFs composites

The average tensile strength, modulus, and toughness of neat P-epoxy and 10% E-epoxy are

summarized in Table 7-2. In this study, toughness was estimated by the total area under stress-

strain. After adding 10 % E-epoxy, the tensile strength decreased slightly, from 76 MPa to 70

MPa, but the strain improved from 6.6 % to 12.3 %, which led to a toughness improvement to

84 %. Student pair t-test analysis of P-epoxy and 10%E-epoxy revealed that the differences in

tensile strength/ modulus were not statistically significant (p=0.087/0.111), but there is a

significant increase in the toughness for 10%E-epoxy resins (p=0.005). The improvement in

the toughness suggests that more energy was absorbed prior to fracture which reduced the

brittle cracking.

Table 7-2 Tensile properties of neat epoxies

Tensile Strength (MPa) Tensile Modulus (GPa) Strain (%) Toughness* (MJ/m3)

P-epoxy 76.0 ± 2.88 1.96 ± 0.06 6.6 2.34 ± 0.02

10%E-epoxy 70.2 ± 3.18 1.24 ± 0.14 12.3 4.43 ± 0.06

*= 5% statistically significant difference

E-epoxy monomer acted as a toughener, which improved elongation, but slightly compensated

the strength and modulus. To address this issue, the use of NCFs should increase the tensile

strength, and maintain the toughness. As described in Table 7-3, the strong reinforcement effect

of NCFs was observed in both epoxies, even with DNCF. Note that the fibre loading is only 25

wt. %. Based on the percolation theory, the percolation threshold is less than 6 vol. %33, but

between 6 vol. % to 18 vol. % is not considered as a dense nanopaper network34. Thus, the

145

fibre loading of this study was set at 25 wt. % (equal to 20.1-23.5 vol. % depending on the

NCFs density assumptions). The improvements of tensile strength of P-epoxy and 10%E-epoxy

after incorporating DNCF were 6.0 % and 8.7 %, respectively. With a SNCF, the increase of

tensile strength of P-epoxy and 10%E-epoxy reached 47.4 % and 88.0 %, which is significantly

higher than the reinforcement from DNCF. This is likely due to the surface difference of NCFs,

as shown in Fig. 7-2. The NCFs before the oven-drying process possessed a fibrous structure;

after oven-drying it exhibited a compact and smooth surface. The nano-scale cellulose tends to

collapse by capillary force during the water evaporation, and the deformed condition is fixed

by hydrogen bonds that form between hydroxyl groups of the cellulose35-36, thus producing a

dense surface that restrains the penetration of resins into the NCFs network. DNCF showed an

impenetrable property, but E-epoxy/DNCF exhibited a three-fold and four-fold increase in

tensile modulus, which is significantly higher than SNCF. One of the possible reasons is that

the epoxy/SNCF composites created many voids inside the structure during the solvent

evaporation process, which decreased the modulus of epoxy/SNCF composites. Moreover,

previous research has indicated that the porosity of composites can significantly decrease the

modulus16.

Table 7-3 Tensile strength and modulus of two types of reinforcement and two types of resin

compositions

DNCF SNCF

Strength (MPa) Modulus (GPa) Strength (MPa) Modulus (GPa)

P-epoxy 80.5 ± 8.00 (↑6.0 %) 7.70 ± 0.80 (↑293 %) 112 ± 3.83 (↑47.4 %) 4.19 ± 0.97 (↑114 %)

10% E-epoxy 76.3 ± 9.31 (↑8.7 %) 6.18 ± 1.42 (↑398 %) 132 ± 2.39 (↑88.0 %) 4.94 ± 0.26 (↑298 %)

(a)

(b)

Fig. 7-2 SEM images of cellulose nano-paper before oven-drying (a) and

after oven-drying (b)

146

The stress-strain curves of P-epoxy/SNCF composites show different patterns to 10%E-

epoxy/SNCF composites (Fig. 7-3). First of all, there is a slope change for P-epoxy/SNCF

composite at 2.3 % strain. The reason for this slope change might be the result of debonding

between fibre/matrix interface. Some studies37-38 indicate that before the ultimate strength is

achieved, the composites will reach a stress level where the resin will begin to crack away from

the reinforcing fibre not aligned with the applied load, and these cracks will spread through the

resin matrix. This is known as ‘transverse micro-cracking,’ and although the composite has not

completely failed at this point, the breakdown process has commenced. Secondly, compared to

neat P-epoxy, P-epoxy/SNCF composites showed a very brittle fracture, without yielding or

strain hardening regions, owing to the poor interface between fibre/matrix. By contrast, 10%

E-epoxy/SNCF avoids the micro-cracking problem, which indicates a superior adhesion

bonding between fibres and E-epoxy.

Fig. 7-3 Tensile stress-strain curves of (a) P-epoxy and (b) 10%E-epoxy and their reinforced

composites

According to the literature16, 19, 22-24, 39-40, the average tensile strength and modulus of

epoxy/NCFs composites are 85 MPa and 4.6 GPa, respectively (NCFs from a mechanical

breakdown process). Our findings show a better tensile strength, and a comparable tensile

modulus compared to the literature average (Fig. 7-4, our experimental results were labeled as

red triangle marks). Among the literature, Yano’s research (the points at the right-upper corner

of Fig. 7-4) showed the highest mechanical performance (350 MPa, 14 GPa). There are two

possible reasons to explain this difference. First, our NCFs contain some hemi-celluloses (>10

wt. %), which is an intrinsic amorphous polymer, and can aggregate cellulose fibrils, resulting

a detrimental effect on the composite strength. Secondly, the epoxy resin adopted in Yano’s

research is UV-curable epoxy resin, which might exhibit different properties to conventional

147

epoxy resin. However, his research is not fully explained as the composite strength (350 MPa)

is higher than NCF reinforcement (260 MPa).

Fig. 7-4 Tensile properties of NCFs reinforced composites compared to literature data 16, 19, 22-24, 39-40

In addition, our experimental data was compared to the prediction value based on the rules of

mixture (ROM). A previous review study34 states that the mechanical properties can be better

described by the ROM than other micromechanical models, such as Halpin-Tsai model and

Cox-Krenchel model. In order to verify this, the predicted tensile properties of the

nanocomposites were plotted using the ROM equations:

Ecomposite = VfEf + (1-Vf)Em………………………………………………………….Eq. (7-4)

σcomposite = Vfσf + (1-Vf) σm………………………………………………………….Eq. (7-5)

where Ecomposite and σcomposite represent the predicted tensile modulus and strength, respectively.

Vf denotes the fibre volume fraction; Ef, σf, Em and σm corresponds to the tensile modulus and

tensile strength of NCFs paper and matrix, respectively.

In the prediction section, we applied three tensile strength and modulus of fibre, including the

neat NCFs paper tested in this study (ROM 1, 145 ± 14.1 MPa and 9.5 ± 1.68 GPa), the highest

reported tensile properties of nanopapers (ROM 2, 243 Mpa and 14.9 GPa),24 and the highest

reported tensile strength and modulus of single plant fibre (ROM 3, 1 GPa and 100 GPa)41,

respectively. Our experimental result is higher than the first prediction (ROM1), and close to

the second prediction model (ROM2). Furthermore, the use of single plant fibre in the

prediction model (ROM3) is far higher than the experimental data, since the nanofibres tend to

be curved and randomly oriented in the plane, which results in a significant strength difference

between a single fibre and NCFs film. In order to eliminate the influence of randomly-oriented

fibres, Halpin-Tsai model (HTM) was applied to the single plant fibre data. Nevertheless, the

148

predicted value of HTM is still 1.2 to 2.5 times higher than the experimental data in tensile

strength and modulus, respectively. Therefore, the ROM based on the highest reported NCFs

paper properties might be a better choice to predict experimental results.

Fig. 7-5 Tensile properties of NCFs reinforced composites compared to theoretical data

The greatest toughness was achieved by adding 10 % E-epoxy. With more E-epoxy, the tensile

strength and toughness decreased. Many studies have shown the optimal amount to add is

approximately 10-20 %26, 42. There are two possible reasons to explain the improved toughness

using E-epoxy. The first reason is that the E-epoxy resin reduces the crosslinking density of

the network by inserting some large molecules into the network structure. Since the network is

loosened by the E-epoxy, the Tg also decreased from 114 °C to 107 °C. The second possible

reason for why E-epoxy can improve toughness results from forming an interpenetrating

network (IPN), which increases the size of physical cross-linked domains. When the first

network is more tightly cross-linked than the second, the second network can offer some room

to allow a micro-deformation.

Table 7-4 Tensile strength and modulus of various E-epoxy replacements

Tensile strength

(MPa)

Tensile Modulus

(GPa)

Toughness

(MJ/m3)

Tg

(°C)

10%E-epoxy/SNCF 132 ± 2.39a 4.94 ± 0.26a 4.36 ± 0.23a 114 ± 1.0a

20%E-epoxy/SNCF 100 ± 14.0b 4.52 ± 0.06a 2.17 ± 0.04b 112 ± 0.0b

30%E-epoxy/SNCF 91.0 ± 5.00b 3.55 ± 0.10b 2.14 ± 0.26b 107 ± 0.5c

* The statistical analysis is based on Scheffé's method

7.3.3 Morphological Characterization of NCFs Reinforced Epoxy Resins

The impact of using DNCF and SNCF is shown in Fig. 7-6. The oven-dried NCFs film

composites show a laminated structure in the SEM images of P-epoxy/DNCF composites.

149

There are plenty of layers of NCFs in the middle, and the thickness is approximately 20 μm.

The P-epoxy/SNCF composites show a homogeneous middle layer, and the thickness of the

NCFs swells to 90-100 μm due to the resin penetration (Fig. 7-6 b). With closer observation of

the tensile fracture cross-section (Fig. 7-6 c, 7-6 d), the fibrous and layer structure of NCFs

paper are still obvious, while the SNCF paper composites show a matrix-covered appearance,

with many white spots, which is possibly the fracture surface of single or bundle of NCFs. As

reported in many research studies34, the individual fractured NCFs appear as white dots in SEM

images. Their extremely short pullout length indicate strong interfacial adhesion, and strong

molecular interactions, probably because the surface hydroxyls of NCFs react with the oxirane

group through an etherification reaction, or due to the secondary bonding between these

hydroxyls.

Fig. 7-6 SEM images of fracture surface in cross-sections. P-epoxy/DNCF(a) and P-

epoxy/SNCF (b) showing the difference in nanopaper thickness caused by resin

penetration. A clear boundary was observed between DNCF and epoxy (c), but not at

SNCF/epoxy (d)

150

The influence of adding E-epoxy resins is depicted in Fig. 7-7. There is no visual difference

between these two composites. The diameter of each white spot was calculated by ImageJ (1.48

version, National Institutes of Health, USA). The dot diameters of P-epoxy/SNCF composites

ranged from 83nm to 502 nm with the average of 253 nm ± 83 nm (n=50). The dot diameters

of 10%E-epoxy/SNCF ranged from 72 nm to 626 nm with the average of 195 nm ± 117 nm

(n=50), which is significantly smaller than that of P-epoxy/SNCF (p=0.009). This implies that

the E-epoxy resins can wet and penetrate NCFs better than P-epoxy resins; this finding also

demonstrates the superior mechanical performance of 10% E-epoxy/SNCF composites.

Fig. 7-7 SEM images of fracture surface in cross-sections of P-Epoxy/SNCF (a) and 10%E-

Epoxy/SNCF(b) showing that the diameter of nanocellulose fibres (white spots) in 10%E-

epoxy blending system is smaller compared to neat P-epoxy.

151

7.3.4 Thermal Stability

The mass loss as a function of temperature for P-epoxy/SNCF and E-epoxy/SNCF are shown

in Fig. 8. The maximum degradation temperature of reference P-epoxy sample was 387°C;

after incorporating NCF, the maximum degradation temperature increased to 411°C, which is

likely due to formation of porous structures, or associated with strong interactions between

fibres and the matrix43. The porous structures are formed after thermal degradation of the NCF,

and the remaining matrix forms a porous structure that can decelerate thermal conductivity.

However, due to a lower thermal stability value of NCF compared to epoxy resins, the initial

decomposition temperature of composites decreased from 373 °C to 337 °C. According to

cellulose decomposition mechanism, cellulose dehydrates between 200 and 280 °C; the

extensive degradation of cellulose occurs when the temperature is more than 300 °C44. Thus,

the heat-resistance index temperature (Ts) of P-epoxy and P-epoxy/SNCF is 185°C and 181°C,

respectively. Adding E-epoxy slightly decreased thermal resistance, which formed a

degradation shoulder at around 356 °C. Based on our previous report, the first shoulder of E-

epoxy resin is around 330 °C29, which is due to the thermal sensitive compounds, such as abietic

acid and fatty acid. The second degradation temperature is around 416 °C, which is slightly

higher than the P-epoxy resin due to the cross-linked lignin.

In an air environment (Fig. 7c, 7d), the decomposition patterns seem more complex than in a

nitrogen environment, as the former involves oxidative reactions. There are three major peaks

for P-epoxy decomposition, while there are five peaks for P-epoxy/SNCF composites. From

300-500 °C, these degradation peaks of P-epoxy/SNCF all overlapped, and the onset

temperature is the same as that of P-epoxy resins. However, the last degradation peak of P-

epoxy/SNCF composites is 64 °C higher than that of neat P-epoxy resins. Adding E-epoxy

resins increased the degradation ratio at low temperature range (300-350 °C), but the high

temperature regions (550-650 °C) are almost intact. Furthermore, no cellulose degradation

peaks can be distinguished from the overlapping peaks.

152

Fig. 7-8 Thermal degradation of P-epoxy and P-epoxy/NCFs with various E-epoxy

replacement (a)(b) in nitrogen environment (c)(d) in air environment

153

7.4 Summary

This chapter investigated the influence of incorporating nanocellulose fibres on epoxy/bio-

epoxy blending systems. Adding 10 wt.% bio-epoxy resins improved toughness significantly

from 2.34 to 4.48 MJ/m3, while adding solvent-exchanged nanocellulose fibres (25 wt.%)

increased the tensile strength and modulus by 88 % and 298 %, respectively. Although E-epoxy

resin can improve toughness, excess amounts of E-epoxy can have an adverse effect on its

mechanical performance and Tg. Consequently, the highest mechanical strength was obtained

with 10% E-epoxy replacement. Compared to data in research literature, our mechanical

performance is above the average line (85 MPa for tensile strength and 4.6GPa for tensile

modulus), and close to the prediction values using the rule of mixture; this indicates a

satisfactory interface between fibre and the matrix. Based on SEM results, the resin/SNCF

demonstrates a good interface compared to the resin/DNCF. The poor interface of

epoxy/DNCF can be explained by the collapsed fibrous structure during water evaporation.

The fibre diameter of P-epoxy/SNCF is around 253 nm, and the fibre diameter of blending

system is around 195 nm, which implies that E-epoxy resins have a better penetrating ability

than P-epoxy. In addition, incorporating NCF can postpone the maximum degradation rate;

however, the initial degradation temperature decreases. Thus, this chapter successfully

provides a strong and ductile NCF reinforced epoxy/ bio-epoxy composite, without using any

chemical surface pretreatment on NCF.

154

References

1. E. de Jong, A. Higson, P. Walsh and M. Wellisch, Biofuel Bioprod Bior, 2012, 6, 606-

624.

2. A. J. Ragauskas, C. K. Williams, B. H. Davison, G. Britovsek, J. Cairney, C. A. Eckert,

W. J. Frederick, J. P. Hallett, D. J. Leak, C. L. Liotta, J. R. Mielenz, R. Murphy, R.

Templer and T. Tschaplinski, Science, 2006, 311, 484-489.

3. F. S. Hu, J. J. La Scala, J. M. Sadler and G. R. Palmese, Macromolecules, 2014, 47, 3332-

3342.

4. S. Gunasekaran, T. B. Gorczyca and H. S. Cole, Google Patents, 2003.


Prod, 2014, 53, 296-307.


Prod, 2013, 43, 757-761.

7. H. Kishi, Y. Akamatsu, M. Noguchi, A. Fujita, S. Matsuda and H. Nishida, J Appl Polym

Sci, 2011, 120, 745-751.

8. L. L. Ferreira, R. Couto and P. J. Oliveira, European Journal of Clinical Investigation,

2015, 45, 32-36.

9. S. Tanpichai, F. Quero, M. Nogi, H. Yano, R. J. Young, T. Lindstrom, W. W. Sampson

and S. J. Eichhorn, Biomacromolecules, 2012, 13, 1340-1349.

10. S. J. Hanley, J. Giasson, J. F. Revol and D. G. Gray, Polymer, 1992, 33, 4639-4642.

11. P. Terech, L. Chazeau and J. Y. Cavaille, Macromolecules, 1999, 32, 1872-1875.

12. M. Grunert and W. T. Winter, J Polym Environ, 2002, 10, 27-30.

13. A. Alemdar and M. Sain, Bioresource Technol, 2008, 99, 1664-1671.

14. S. Janardhnan and M. Sain, Int J Polym Sci, 2011.

15. P. Y. Kuo, N. Yan and M. Sain, Eur Polym J, 2013, 49, 3778-3787.

16. K. Y. Lee, T. Tammelin, K. Schulfter, H. Kiiskinen, J. Samela and A. Bismarck, Acs Appl

Mater Inter, 2012, 4, 4078-4086.

17. M. Matos Ruiz, J. Y. Cavaillé, A. Dufresne, J. F. Gérard and C. Graillat, Compos

Interface, 2000, 7, 117.


19. M. Nuruddin, T. H. Mahdi, M. V. Hosur and S. Jeelani, presented in part at the ASC 29th

Technical Conference, UC San Diego, 2014.

155

20. M. Matos Ruiz, J. Y. Cavaillé, A. Dufresne, J. F. Gérard and C. Graillat, Compos

Interface, 2000, 7, 117-131.

21. L. M. Tang and C. Weder, Acs Appl Mater Inter, 2010, 2, 1073-1080.

22. H. Q. Liao, Y. Q. Wu, M. Y. Wu and H. Q. Liu, Polym Composite, 2011, 32, 837-845.

23. H. Yano, J. Sugiyama, A. N. Nakagaito, M. Nogi, T. Matsuura, M. Hikita and K. Handa,

Adv Mater, 2005, 17, 153-+.

24. F. Ansari, S. Galland, M. Johansson, C. J. G. Plummer and L. A. Berglund, Compos Part

a-Appl S, 2014, 63, 35-44.

25. J. Lu, P. Askeland and L. T. Drzal, Polymer, 2008, 49, 1285-1296.

26. D. Ratna, J Adhes Sci Technol, 2003, 17, 1655-1668.

27. F. Isabelle, M. Patrick and D. Stoil, in Interpenetrating Polymer Networks, American

Chemical Society, 1994, vol. 239, ch. 26, pp. 539-556.

28. L. Di Landro and G. Janszen, Composites Part B: Engineering, 2014, 67, 220-226.

29. P. Y. Kuo, M. Sain and N. Yan, Green Chem, 2014, 16, 3483-3493.

30. N. C. Wu, Master, University of Toronto, 2010.



32. A. Ortega, Thermochim Acta, 2008, 474, 81-86.

33. G. Siqueira, J. Bras and A. Dufresne, Polymers-Basel, 2010, 2, 728-765.

34. K. Y. Lee, Y. Aitomaki, L. A. Berglund, K. Oksman and A. Bismarck, Compos Sci

Technol, 2014, 105, 15-27.

35. H. Yano and S. Nakahara, J Mater Sci, 2004, 39, 1635-1638.

36. M. Nogi, S. Iwamoto, A. N. Nakagaito and H. Yano, Abstr Pap Am Chem S, 2009, 237.

37. D. J. Kim, S. H. Kang and T. H. Ahn, Materials, 2014, 7, 508-526.

38. P. Palmero, L. Montanaro, H. Reveron and J. Chevalier, Materials, 2014, 7, 5012-5037.

39. D. M. Bruce, R. N. Hobson, J. W. Farrent and D. G. Hepworth, Composites Part A:

Applied Science and Manufacturing, 2005, 36, 1486-1493.

40. H. Liao, Y. Wu, M. Wu, X. Zhan and H. Liu, Cellulose, 2012, 19, 111-119.

41. W. S. Beck, Bioscience, 1983, 33, 662-662.

42. D. Ratna and A. K. Banthia, J Adhes Sci Technol, 2000, 14, 15-25.

156

43. K. A. Sindhu, R. Prasanth and V. K. Thakur, in Medical Applications of Cellulose and Its

Derivatives: Present and Future, ed. V. K. Thakur, John Wiley & Sons, Inc, Hoboken,

2014.

44. Y. Peng, D. Gardner, Y. Han, A. Kiziltas, Z. Cai and M. Tshabalala, Cellulose, 2013, 20,

2379-2392.

157

Chapter 8 Conclusions, Contributions and Future Work

This chapter summarizes the characterization of extractive-based epoxy resins, presents our

scientific contribution to the field of bio-epoxy resins and engineering prospects of our newly

developed materials, and provides some future directions for further investigations or scale-up

applications.

8.1 Conclusions

The necessary transition from fossil-based BPA towards the development of sustainable

building blocks has become evident. The renewable alkaline bark extractives are uniquely

suited to replace BPA due to their akin polyphenol structures. This thesis endeavours to provide

a comprehensive and systematic study on the development and characterization of bio-based

epoxy resins from tree bark and improves the yield, reactivity, thermal resistance, and

mechanical properties of bio-epoxy resin in order to enlarge its future applications.

The major findings are summarized as follows:

1. The epoxidation reaction occurred successfully between bark extractives and ECH as

confirmed by FTIR and NMR.

2. The activation energy of extractive-based epoxy resin was lower than petroleum-based

epoxy resin, which implies extractive-based epoxy resin demands less energy to trigger the

reaction.

3. To produce E-epoxy efficiently, the best combination of reaction parameters is using

spray-dried extractives as substrates, a dioxane/water combination as solvent, and

tetrabutylammonium hydroxide as a ring-opening catalyst. The maximum extent of

conversion with minimum epoxy equivalent weight was achieved after 4.5 hours reaction

time with 3.4 sodium hydroxide/hydroxyl value molar ratios at 80 °C reaction temperature.

4. The degradation kinetic parameters show that the activation energy of bio-epoxy was a

dynamic value, indicating a multistage degradation mechanism or a variety of components.

5. FTIR spectra show that dehydration and crosslinking reactions occurred at low

temperatures while more Claisen chain rearrangement and chain-scission reactions occurred

at high temperatures.

158

6. Py-GC/MS results exhibit that the major skeletal structures were methyl abieta-8,11,13-

trien-18-oate, diethyl phthalate, 2,2'-isopropylidenebis(3,5-dimethylbenzofuran), and

epimanool.

7.NCFs catalyzed the reaction between commercial bisphenol F resins and the aromatic

amine curing agent.

8. The curing behaviours of NCFs catalyzed epoxy resins can be better fitted with S-B model

than Kamal model.

9. After incorporating NCFs, the tensile strength and modulus of the blending system

(10%E-epoxy/90%P-epoxy) increased 88 % and 298 %, respectively.

10. The maximum thermal degradation peak of the composites (NCFs/P-epoxy or NCFs/P-

epoxy/E-epoxy) was detected 24 °C higher than both neat epoxy resins.

8.2 Scientific Contributions and Engineering Prospects

1. Our research sheds new light on the reaction mechanism between bark extractives and ECH

and is the first study to develop a bio-epoxy resin directly from bark extractives. The novel

extractive-based epoxy resin can replace up to 50% conventional epoxy resins and the

blending system displays competitive mechanical and thermal properties compared to

commercial epoxy resins. Different to other bio-based epoxy resins, extractive-based epoxy

resin does not require complex and toxic extraction processing. Furthermore, this resin is a

nonfood-based biomaterial which is less likely to impact on consumer price in developing

countries compared to oil- or sugar-based biomaterials. Thus, our study provides a highly

potential renewable raw material to produce epoxy resins.

2. This thesis identifies the influential factors of E-epoxy preparation that reveals insight into

the glycidylation reaction between ECH and renewable biomass. Our research is one of the

first studies investigating both categorical and numerical reaction parameters on yield,

reactivity and molecular structures of bio-epoxy resins. Our research shows that hydroxyl

group amounts can significantly affect the yield of products; solvents greatly control the

reactivity and functionality of extractive-based epoxy resin; catalysts dominate the

possibility of side reactions. For the numerical parameters, temperatures are more influential

on yield and reactivity compared to time and catalyst amount.

159

3. Our work establishes a new and fundamental thermal degradation mechanism of bio-epoxy

resins, which may lead to a promising thermal enhancement. Additionally, the thermal

approach applied in this study can avoid a time-consuming separation process and obtain

informative results. As mono-epoxidized resin acid and mono-epoxidized fatty acid were

detected in this research, it also clearly explained why the thermal performance of

extractive-based epoxy resins was slightly lower than commercial products.

4. This study develops a unique composite (E-epoxy/P-epoxy/NCFs) of high mechanical

strength, superior thermal stability, and long-term sustainability. This composite is an

innovative development, which well balances the strength and toughness, and to the best of

our knowledge, no research has ever combined bio-epoxy resin and NCFs together with a

tensile strength above 100 MPa. Compared to data in research literature, the mechanical

performance of our composite is above the average line, and close to the prediction values

using the rule of mixture, indicating a satisfactory interface between fibres and the matrix.

The toughness is significantly improved by adding extractive-based epoxy resin and the

tensile strength and modulus was enhanced by adding nanocellulose fibres. In addition,

incorporating nanocellulose can postpone the maximum degradation rate. Lastly, no

chemical surface pretreatment is required on nanocellulose, which adheres to the principle

of green chemistry.

8.3 Future Work

Future work in the research area of bio-based epoxy resins should focus on further improving

performance of the extractive-based epoxy resin to facilitate their future applications and

commercialization. Key areas proposed for future works are outlined as follows:

1. It would be worthwhile to conduct a life cycle assessment (LCA) on bio-based epoxy

resins as it can offer a cradle-to-cradle examination on energy consumption, carbon dioxide

emission and environmental impacts. Based on the results from LCA, the sustainability and

toxicity of our bio-based epoxy resins can be studied and further improved.

2. It will be of beneficial to improve the extent of conversion of E-epoxy production. The

first method is to change the synthesis route from ECH method to a double bond oxidation.

In ECH method, some hydroxyl groups cannot be fully converted into epoxy groups. To

address this issue, we can apply 3-bromoprop-1-ene to covert hydroxyl groups into double

structure and then reacts to hydrogen peroxide to obtain epoxy groups. This method not only

160

increase the yield, but also offer some insight into the components which cannot be

transformed through ECH method. The second method is to choose a strong and non-

nucleophilic base as a ring-closing catalyst. In this thesis, we used NaOH to finish the ring-

closing reaction, but NaOH only reacts with the compounds with pKa value smaller than

15.8 and also partially hydrolyzed the final product. Thus, we can use N,N-

diisopropylethylamine (DIEA) instead of NaOH. DIEA can react with any compound with

pKa value lower than 35 and also decrease the side reactions.

3. Crosslinking density is an essential factor for the mechanical performance of

thermosetting polymers. It would be of great interest to investigate the crosslinking density

by swelling test, DMA measurements and NMR. The crosslinking density of bio-epoxy

resins through various curing schedule should be compared. The unlinked fraction or

oligomers can be detected by a solvent extraction process. We recommend to conduct a 40

hour extraction under reflux with dichloromethane/ tetrahydrofuran in soxhlet extractor.

These extracted compounds can offer some insight into the mono-functionalized compound

and possible side reactions, which also help to explain the inferior thermal property of

extractive-based epoxy resin. In addition, ESR technology should be applied on the

detection of free radicals. Identifying the free radical types and generating temperature can

definitely offer shed a light on thermal oxidation degradation.

4. It would be of great importance to develop a fully bio-based epoxy/fiber

nanocomposite, which combines bio-phenol, bio-epichlorhydrin, bio-curing agent and bio-

nanofibres. In our study, we successfully developed a bio-phenol but we did not combine

our bio-epoxy with bio-epichlorohydrin, which is available from Dow chemical Company

and Solvay Company. In addition, there are several bio-curing agents developed from

natural resources, such as malapimaric anhydride, which already shows great potential to

replace conventional anhydride type curing agent. Less attention has been focused on the

fully bio-based composites due to their often low strength, but our study have shown its

thermal and mechanical strength can be improved. The most important thing is that a fully

bio-based composite can significantly reduce the dependency of petroleum resource and

carbon print.

161

Appendix A (Chapter 3)

Appx.1 and 2 show the original DSC plots before any data manipulation and their thermal

characteristics were listed in the Appx.3. These information were used to support the section

3.3.3.

Appx. 1 DSC dynamic scans of P-Epoxy resins with amine curing agent at the heating rates

of 5, 10, 15 and 20 ℃min-1

Appx. 2 DSC dynamic scans of E-Epoxy resins with amine curing agent at the heating rates

of 5, 10, 15 and 20 ℃min-1

162

Appx.3 Thermal characteristics of samples from the dynamic DSC analysis

Epoxy

resins

Heating rate

(℃/min)

Onset

temperature (℃)

Peak temperature

(℃)

Kissinger

Activation

energy (kJ/mol)

P-Epoxy 5 130.7 174.9 52.0 kJ/mol

10 141.5 195.2

15 157.2 210.2

20 164.0 215.9

E-Epoxy 5 104.3 156.2 42.3 kJ/mol

10 116.3 177.1

15 129.8 192.9

20 133.2 203.0

163

Appendix B (Chapter 4)

The following figures show adequacy of the response surface analysis model in chapter 4,

including residual analysis (Appx. 4-5) and the comparison between predicted and actual

values (Appx. 6).

Yield EEW

Appx. 4 Residuals versus predicted

Yield

EEW

Appx. 5 Residuals versus run

Yield

EEW

Appx. 6 Predicted versus actual

Design-Expert?Softwareyield

Color points by value ofyield:

64

10.5

Run Number

Exter

nally

Stud

entiz

ed R

esidu

als

Residuals vs. Run

-6.00

-4.00

-2.00

0.00

2.00

4.00

6.00

1 3 5 7 9 11 13 15 17

Design-Expert?SoftwareEEW

Color points by value ofEEW:

2240

317

Run Number

Exte

rnall

y Stu

dent

ized

Resid

uals

Residuals vs. Run

-6.00

-4.00

-2.00

0.00

2.00

4.00

6.00

1 3 5 7 9 11 13 15 17

Design-Expert?Softwareyield

Color points by value ofyield:

64

10.5

Actual

Pred

icte

d

Predicted vs. Actual

0

10

20

30

40

50

60

70

0 10 20 30 40 50 60 70

Design-Expert?SoftwareEEW

Color points by value ofEEW:

2240

317

Actual

Pred

icte

d

Predicted vs. Actual

0

500

1000

1500

2000

2500

0 500 1000 1500 2000 2500

164

(a) BiF-Aliph

BiF-Aliph-DMF

(c)BiF-Aliph-DMF-NCFs

Ap

pen

dix

C (C

ha

pter 6

)

(d) BiF-Arom

(e) BiF-Arom-DMF

(f) BiF-Arom-DMF-NCFs

Appx. 7 Curve of reaction rate versus conversion rate with Kamal model and S-B model predictions (cont’)

165

(g) WD-Aliph

(h) WD-Aliph-Water

(i) WD-Aliph-Water-NCFs

(j) WD-Arom

(j) WD-Arom-Water

(k) WD-Arom-Water-NCFs

Appx. 7 Curve of reaction rate versus conversion rate with Kamal model and S-B model predictions

development and characterization of an extractive-based ... · pei-yu kuo doctor of philosophy...

Documents