New Self-healing Epoxy-based Polymers...mechanism of thermal self-healing in the prepared epoxy polymers. By different methods, including the determination of the change in glass transition

New Self-healing Epoxy-based Polymers

A thesis submitted in fulfilment of the requirements for the degree of

DOCTOR OF PHILOSOPHY

Nan Bai

M.S.Chem, Tsinghua University

B.S.Chem, Tsinghua University

Department of Materials Engineering

Monash University

Melbourne, Australia

July, 2014

Copyright Notices

Notice 1

Under the Copyright Act 1968, this thesis must be used only under the normal

conditions of scholarly fair dealing. In particular no results or conclusions should be

extracted from it, nor should it be copied or closely paraphrased in whole or in part

without the written consent of the author. Proper written acknowledgement should be

made for any assistance obtained from this thesis.

Notice 2

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

third-party content included in this thesis and have not knowingly added copyright

content to my work without the owner's permission.

i

Table of Contents

Table of Contents ...................................................................................... i

Declaration ............................................................................................... iv

Abstract ..................................................................................................... v

Acknowledgements ................................................................................. vii

Abbreviations........................................................................................... ix

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

1.1 Background ......................................................................................................... 1

1.2 Research objectives ............................................................................................. 3

1.3 Structure of the dissertation ................................................................................. 4

Chapter 2 Literature Review .................................................................. 6

2.1 Introduction ......................................................................................................... 6

2.2 Irreversible self-healing polymers ....................................................................... 7

2.2.1 Hollow glass fibres ....................................................................................... 7

2.2.2 Microcapsules............................................................................................... 9

2.2.3 Microvascular networks .............................................................................. 11

2.2.4 Linear polymer/network systems ............................................................... 13

2.3 Reversible self-healing polymers ...................................................................... 14

2.3.1 Diels-Alder reaction based systems ........................................................... 14

2.3.1.1 Furan-maleimide based system ........................................................................ 15

2.3.1.2 Dicyclopentadiene-based system ..................................................................... 30

2.3.1.3 Other DA-based systems .................................................................................. 33

2.3.2 Thiol-based systems ................................................................................... 36

2.3.3 Alkoxyamine-based systems ...................................................................... 39

ii

2.3.4 Photodimerisation-based systems .............................................................. 40

2.3.5 Hydrogen bonding systems ........................................................................ 42

2.3.6 Metal-ligand coordination systems ............................................................ 44

2.3.7 Guest-host systems ..................................................................................... 46

2.3.8 Ionic systems .............................................................................................. 48

2.4 Summary ........................................................................................................... 49

Chapter 3 Materials and Analytical Techniques ................................ 50

3.1 Materials ............................................................................................................ 50

3.2 Analytical techniques ........................................................................................ 51

3.2.1 Fourier transform infrared spectroscopy .................................................... 51

3.2.2 Near-infrared spectroscopy ........................................................................ 52

3.2.3 Differential scanning calorimetry ............................................................... 52

3.2.4 Dynamic mechanical analysis .................................................................... 53

3.2.5 Thermogravimetric analysis ....................................................................... 53

3.2.6 Swelling test ............................................................................................... 54

3.2.7 Gel permeation chromatography ................................................................ 54

3.2.8 Optical microscopy .................................................................................... 55

Chapter 4 Synthesis of Diamine Cross-linker ..................................... 56

4.1 Introduction ....................................................................................................... 56

4.2 Synthetic attempts to produce a diamine cross-linker with single DA unit ...... 56

4.3 Synthesis of diamine cross-linker with double DA units .................................. 63

4.4 Summary ........................................................................................................... 66

4.5 Experimental ..................................................................................................... 67

4.5.1 General ....................................................................................................... 67

4.5.2 Synthetic attempts via mono-maleimide intermediate ............................... 67

4.5.3 Synthesis of diamine cross-linker via bis-maleimide intermediate ............ 75

iii

Chapter 5 Preparation and Characterisation of the New Self-healing

Epoxy Polymers ...................................................................................... 79

5.1 Introduction ....................................................................................................... 79

5.2 Variation of the curing conditions for the preparation of thermosets ................ 79

5.2.1 Self-healing epoxy polymers prepared from DGEBA ............................... 79

5.2.2 Self-healing epoxy polymers prepared from TGAP .................................. 94

5.3 Characterisation of self-healing polymers....................................................... 102

5.3.1 Preparation of a control sample of epoxy polymer without DA unit ....... 102

5.3.2 The DGEBA-based self-healing epoxy polymer...................................... 103

5.3.3 The TGAP-based self-healing epoxy polymer ......................................... 109

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

Chapter 6 Self-healing Properties of the New Epoxy Polymers ...... 115

6.1 Introduction ...................................................................................................... 115

6.2 Study on DA and RDA reaction conditions ..................................................... 115

6.2.1 DA and RDA reactions in the diamine cross-linker .................................. 115

6.2.2 The conditions of the DA and RDA reactions in the cross-linked epoxy

polymers ............................................................................................................. 118

6.3 The thermal self-healing mechanism in the new epoxy system ...................... 127

6.4 Self-healing behaviour on the surface of polymer .......................................... 144

6.5 Summary ......................................................................................................... 152

Chapter 7 Conclusions and Future Work ......................................... 155

7.1 Conclusions ..................................................................................................... 155

7.2 Future work ..................................................................................................... 159

References ............................................................................................. 162

Publications ........................................................................................... 173

iv

Declaration

I hereby declare that this thesis contains no material which has been accepted for the

award of any other degree or diploma at any university or equivalent institution and

that, to the best of my knowledge and belief, this thesis contains no material

previously published or written by another person, except where due reference is

made in the text of the thesis.

Nan Bai

Department of Materials Engineering

Monash University

July 2014

v

Abstract

Cross-linked polymers with self-healing properties have become a very popular

polymer research field in recent times. In this thesis, a new type of cross-linked self-

healing polymer system based on the Diels-Alder (DA) system is developed, with

adduct of furan and maleimide groups as the cleavable and reformable segments being

incorporated into the amine-based cross-linker. Such a concept allows a large range of

commercially-available and widely-used epoxy resins to be used as monomers for the

production of self-healing polymers.

Initially, one DA unit was incorporated into the structure of diamine cross-linker and

its hydrochloride salt was obtained after several synthetic steps. However, the yield of

its neutralised free amine was very low and product was impure, even after many

methods were tried, including different neutralisation methods, different synthetic

routes and after increasing the carbon chain length of the target diamine in an attempt

to achieve better solubility. Given this issue, a new diamine cross-linker with two DA

units in the structure was designed, successfully synthesised and its structure was

verified by characterisations.

The successfully synthesised diamine cross-linker was used to cure two different

epoxy monomers, diglycidyl ether of bisphenol A (DGEBA) and triglycidyl p-amino

phenol (TGAP), yielding polymers with different crosslink densities. The curing

conditions were optimised based on near-infrared spectroscopy (NIR) and differential

scanning calorimetry (DSC) tests. It was necessary that the temperature of the curing

(polymerisation) cross-linked reaction should be lower than 80 oC, to minimise both

the Retro Diels-Alder (RDA) and other side reactions, which may irreversibly destroy

the DA unit. In addition, a high degree of cure is necessary for good mechanical

properties of the cross-linked polymer. The polymers cured using optimised

vi

conditions were then characterised by Fourier transform infrared spectroscopy

(FTIR), DSC, dynamic mechanical analysis (DMA) and thermogravimetric analysis

(TGA).

The conditions of DA and RDA reactions, both in the synthesised diamine cross-

linker itself, and in the cured epoxy polymers were investigated. The optimised

conditions, which were also the self-healing conditions, were used to study the

mechanism of thermal self-healing in the prepared epoxy polymers. By different

methods, including the determination of the change in glass transition and swelling

properties during the healing process, and the analysis of the components of

scissioned polymer at temperatures above RDA reaction temperature, the thermal self-

healing process of this type of epoxy polymer was understood.

The self-healing process on the surface of the prepared polymers was observed by

optical microscopy. Compared with the self-healing result from the control sample

without DA units, the healed surface of a self-healing polymer that had been scratched

demonstrated that the cleavage of the DA units and subsequent flow of the oligomer

were the basis of the self-healing process. The flow activation energy of the formed

oligomers during the healing process was calculated using the Arrhenius equation, and

was found to be a little higher than that of the linear polymer, likely due to its highly

branched structure.

vii

Acknowledgements

I would like to express my deepest appreciation to all those who provided me the

possibility to complete this thesis.

Firstly, I must express my special and sincere gratitude to my supervisors, Prof.

George Simon and Dr. Kei Saito, for the professional direction and continuous

supportive help throughout my PhD. Without your supervision and suggestions, I

cannot conquer the challenges and achieve the satisfying research results. I am very

grateful to have had your leadership and guidance during the past years’ scientific

research.

Thanks are also due to the Faculty of Engineering International Postgraduate

Research Scholarship (FEIPRS) and Monash Graduate Scholarship for financial

support.

Furthermore, I would like to acknowledge with much appreciation all the group

members for their friendship and assistance - especially Mr. Jack (Jiang) Wang (for

training on DSC and TGA), Mr. Hao Zhang, Dr. Priscilla Johnston, Dr. Gagan Kaur

(for training on GPC), Ms. Sepa Nanayakkara and Mr. Ashley Walker. And a big

thank you to all the technical and general staff in the Department of Materials

Engineering and the School of Chemistry in Monash University who have helped me

over the years - in particular, Dr. Jana Habsuda (for training on DMA), Mr. Silvio

Mattievich, Ms. Edna Tan, Ms. LeeAnn Hilborn, Ms. Michelle Laing, Dr. Peter

Nichols (for training on NMR), Dr. Boujemaa Moubaraki, Mr. Finlay Shanks (for

training and help on FTIR and NIR). Thanks to Dr. Yu Chen for the assistance and

training on the optical microscope. I am grateful for having been in such a friendly,

helpful and professional department with all the intelligent staff and PhD student

colleagues.

viii

Thanks to all my friends and alumni from Tsinghua University, for the wonderful time

and happiness shared together.

Last but not least, I would like to express my gratitude to my family. To my parents

and parents-in-law, thanks for your support and understanding. Thanks to Uncle

Zhang, Dr. Jinping Zhang, for your encouragement and suggestions over the past

years.

To my dear wife, Jia (Cathy) Cao, millions of thanks are still not enough for your

support, understanding, help, suggestions, encouragement and continuous love. I’m

very lucky to have you with me. The past years were quite hard for us, but we have

succeeded. I hope we will have a wonderful future.

ix

Abbreviations

δ Chemical shift

m Micrometre

Wavelength

ΔH

Enthalpy

oC Degrees Celsius

aq. Aqueous solution

ATR Attenuated total reflectance

ATRP Atom transfer radical polymerisation

B21C7 Benzo-21-crown-7

BIOP 2,6-Bis(benzimidazolyl)-4-oxypyridine

BHF Bis(hydroxymethyl)furan

BMD Bismaleimidodiphenylmethane

Boc Di-tert-butyl dicarbonate

BOP Benzotriazole-1-yl-oxy-tris(dimethylamino)phosphonium

hexafluorophosphate

CD Cyclodextrin

cm-1

Reciprocal centimetre

cm3 Cubic centimetre

calc. Calculated

CCD Charge-coupled device

CPD Cyclopentadiene

[CPH] Value of peak height

d Doublet (NMR); day (s)

DA Diels-Alder

DB24C8 Dibenzo[24]crown-8

DCDC Double cleavage drilled compression

DCPD Dicyclopentadiene

DFA Difurfuryl adipate

DGEBA Diglycidyl ether of bisphenol A

DGFA 1-(Furan-2-yl)-N,N-bis(oxiran-2-ylmethyl)methanamine

DMA Dynamic mechanical analysis

DMF Dimethylformamide

x

DMSO Dimethylsulfoxide

DNA Deoxyribonucleic acid

DPMBMI 1,1'-(Methylenebis(4,1-phenylene))bis(1H-pyrrole-2,5-dione)

DSC Differential scanning calorimetry

DTT Dithiothreitol

EEW Epoxide equivalent weight

EMAA Ethylene-co-methacrylic acid

ENB 5-Ethylidene-2-norbornene

eq. Equivalent

Et2O Diethyl ether

Et3N Triethylamine

EtOAc Ethyl acetate

ESI-MS Electrospray ionisation mass spectrometry

Fc Ferrocene

Fe Iron

FGE 2-((Oxiran-2-ylmethoxy)methyl)furan

FTIR Fourier transform infrared spectroscopy

g Gram

GPa Gigapascal

GPC Gel permeation chromatography

h Hour(s)

HCl Hydrogen chloride

HEDTA N-hydroxyethyl-ethylenediamine-triacetic acid

Hg Mercury

HMDS Hexamethyldisilazane

HMF 5-Hydroxymethylfurfural

HOPDMS Hydroxyl end-functionalised polydimethylsiloxane

Hz Hertz

J Joule

J Coupling constant

kJ Kilojoule

L Litre

m Multiplet (NMR)

M2 1,8-Bis(maleimido)-triethylene glycol

xi

MDPBM N,N’-Methylenediphenylbismaleimide

MHHPA 3a-Methylhexahydroisobenzofuran-1,3-dione

MHz Megahertz

min Minute (s)

mL Millilitre

mN Millinewton

Mn Number average molecular weight

mg Milligram

mm Millimetre

mmol Millimole

mol Mole

MPa Megapascal

Mw Weight average molecular weight

mW Milliwatt

Mw/Mn Polydispersity index

m/z Mass to charge ratio

NIR Near-infrared spectroscopy

NMR Nuclear magnetic resonance

PAA Poly(acrylic acid)

PAEI Poly(N-acetylethylenimine)

PA-F Furan-containing polyamide

PA-MI Maleimide-containing aromatic polyamide

PBA Poly(n-butyl acrylate)

PCL Poly(ε-caprolactone)

Pd Palladium

PDES Polydiethoxysiloxane

PEA Poly(ethylene adipate)

PEG Poly(ethylene glycol)

PEMAA Poly(ethylene-co-methacrylic acid)

PET Poly(ethylene terephthalate)

PFS Poly(2,5-furandimethylene succinate)

pH Logarithm of hydrogen ion

PK Polyketone

PMMA Poly(methyl methacrylate)

xii

Pt Platinum

R Alkyl or aryl group

RDA Retro Diels-Alder

ROMP Ring opening metathesis polymerisation

rpm Revolutions per minute

r.t. Room temperature

s Singlet (NMR); second (s)

SA Succinic acid

t Triplet (NMR)

Tc Crystallisation temperature

TCE 1,1,1-Tris-(cinnamoyloxymethyl)ethane

TCPD Tricyclopentadiene

TDS Thiuram disulfide

TEMPO (2,2,6,6-Tetramethylpiperidin-1-yl)oxy

tert Tertiary

TF Tris-furan

Tg Glass transition temperature

TGA Thermogravimetric analysis

TGAP Triglycidyl p-amino phenol

TGDDM N,N,N’,N’-Tetraglycidyl-4,4’-diaminodiphenylmethane

TLC Thin layer chromatography

Tm Melting transition temperature

TMI Tris-maleimide

TPy 2,2’:6’,2’’-Terpyridine

TTC Trithiocarbonate

UPy 2-Urido-4-pyrimidone

UV Ultraviolet

ν Wavenumber

W Watt

wt% Percentage by weight

Zn Zinc

Chapter 1 Introduction

1


1.1 Background

Polymers are increasingly used for more and more applications in modern society,

having many advantages, including low density, good processibility and good

chemical stability. As a result, polymers are becoming indispensable in today’s life,

and their annual consumption has reached billions of US dollars.1, 2

In a general sense, polymers have a finite lifetime, because microcracks can form

inside polymers due to accumulated high stresses and strains in service,2, 3

such as in

the flexing wings of an aircraft. Irreversible damage due to such microcracks can

significantly reduce the load-carrying capacity of polymers and shorten their lifetime.

Since these cracks are usually microscopic in size and deep inside the object, they are

hard to detect and repair. It would be desirable for polymers to have self-healing

properties, as do animals and plants.1-4

There are two main classes of polymers: thermoplastics and thermosets. Upon

heating, thermoplastic polymers become soft and deformable and can undergo flow.

They are effectively self-healing polymers, although such healing and flow processes

may lead to unacceptable distortion of the original plastic part. Highly cross-linked

thermosets are a very useful class of polymers and tend to have superior mechanical

properties over non-cross-linked (thermoplastic) polymers, including high modulus,

good thermal properties and solvent resistance.3 However, the cross-linked structure

of thermosets (such as epoxy resins) means that they cannot be healed simply by

reheating. Therefore other polymer designs or processes are necessary for thermosets

to have self-healing properties.


2

Several self-repairing strategies are being researched to make thermosets self-healing,

predominantly involving the storage of healing agents inside the polymer or the

incorporation of functional groups in the polymer structure to allow reversible

reactions to occur between chains.2 The former allows the maintenance of the good

mechanical properties of the original polymer, but it has limited healing efficiency

and does not allowed repeated healing cycles because the chemical agents can become

depleted. Therefore it is called irreversible self-healing polymer. The latter can

potentially self-heal multiple times, but requires external stimuli, and often

demonstrates reduced mechanical properties because the combination of functional

groups that needs to be incorporated into the polymer architecture changes the

original, advantageous molecular structure. Therefore it is called reversible self-

healing polymer.

It is the reversible self-healing polymer class that we target in this research, because

of its ability to be repeatedly healed. Much research is currently being undertaken in

this area, however, most of the polymer systems being investigated share a common

set of issues:

1. The self-healing polymers mentioned in the literature are often not suitable for

commercial applications because the structure and properties of the polymers are far

from that of the polymers required in industry and everyday life. Although the self-

healing polymers obtained show good, reversible properties, more research needs to

be undertaken to improve the properties so they align to the demands of the

commercial applications.

2. The mechanical properties of the healed materials are usually reduced compared to

the original samples, which is related to the extent of the healing efficiency of

reversible cycles. In addition, the time of reversible reaction cycles needs to be

improved, as many of the proposed polymer systems allow only a limited number of


3

healing cycles before irreversible changes happen, because of the various side

reactions that can occur during the healing process.

3. The synthesis of the monomers/oligomers containing reversible functional groups

is usually costly due to the expensive starting materials, multiple complex reaction

steps, and the significant efforts required in purification of the products. Strategies for

the synthesis of new but more cost effective materials is clearly desirable.

To date, the furan-maleimide-based system has been shown to be an efficient,

repairing chemical system that confers on polymers a self-healing ability. This moiety

has often been included in epoxy resin systems, which themselves have many

significant advantages, such as good mechanical properties, easy processing and small

volume shrinkage when polymerising. Epoxy resins have many applications in a

range of fields, including fibre-reinforced plastic materials, potting materials for

electronics, coatings, civil engineering applications and as general purpose

adhesives.5, 6

Clearly it is beneficial for such widely-used materials like epoxy resins

to be made repairable.

Based on the above considerations, our interest in self-healing polymers is focussed

on furan-maleimide-based epoxy polymers.

1.2 Research objectives

The aim of this thesis is to develop a new concept in self-healing, cross-linked

polymers based on the Diels-Alder (DA) system. The furan-maleimide-based

chemistry was used in this research with the mendable unit being incorporated in the

amine-based cross-linker unit. This is a new development, with previous work

involving incorporation of mendable groups (including the DA unit) in the epoxy

monomer. This new strategy means the amine can be used to cure a range (including


4

mixtures) of epoxy resins which are already commercially-available and widely-used.

Since the amine is usually the minority component of the reactive mixture, it is

desirable that this component contains the self-healing group. In addition, this would

allow some proportion of the required stoichiometric amine mixture to be the self-

healing amine, in combination with other commercial amines, further extending

possible molecular structures. The detailed technical objectives of the research are

given below:

1. Synthesis of a new diamine cross-linker with furan/maleimide units, using a

relatively simple synthetic scheme and methodology.

2. Preparation of self-healing epoxy network polymers based on commercial epoxy

resins and the diamine cross-linker, including optimisation of the curing conditions

and characterisation of the resultant cross-linked polymer. The degree of crosslink

density of the epoxy will also be varied by using either a bi- or tri-functional epoxy

resin.

3. Study of the reversible DA reaction in the polymer system and optimisation of the

conditions required to cause self-healing.

4. Exploration of the thermal mechanism of the self-healing process using different

characterisation methods.

5. Investigation of the self-healing process of these materials, as it occurs on the

surface of the system. This is relevant when they are used, for example, as a coating.

1.3 Structure of the dissertation

This thesis comprises six further chapters as described below.

Chapter 2 is a detailed review of the research to date on self-healing polymers with


5

different strategies and methods, involving either polymers which contain healing

agents to cause healing, or reversible self-healing polymers with different functional

groups which allow such a healing process.

Chapter 3 describes all the materials used for the synthesis of the diamine cross-

linker with DA units and the self-healing epoxy polymers. It also briefly describes the

instruments and techniques used in characterisation of the polymeric materials and the

self-healing mechanism in this polymer system.

Chapter 4 outlines the synthesis of the new diamine cross-linker with different

numbers of DA units, including the design of the synthetic route, trials of the synthetic

methods and characterisation of the products in each step.

Chapter 5 explores the curing process of the self-healing epoxy polymers with

different epoxy resins and optimises the curing conditions involved in the preparation

of the polymers, followed by their characterisation.

Chapter 6 studies the reversible DA reaction both in the diamine cross-linkers and in

the self-healing cross-linked polymer system. The thermal mechanism of the self-

healing behaviour is investigated using a variety of methods and the self-healing

process on the surface of the sample is demonstrated.

Chapter 7 draws conclusions about the research undertaken and recommends areas

for future research.

Chapter 2 Literature Review

6


2.1 Introduction

It is inevitable that thermosetting or cross-linked polymers are damaged and acquire

flaws or scratches, either internally or on the surface, during the lifetime of a product,

which may lead to a reduction in properties. If deeper within the material or on the

surface, this damage may lead to a degradation in mechanical properties,7 or if on the

surface of a material used as a coating, a lack of transparency, which may reduce the

functionality of the device for which the coating is required. Traditional healing or

repairing methodologies include welding, patching and in-situ curing of resins, or in

the case of coatings, polishing or buffing.8-10

However, these methodologies can lead

to substantial increases in maintenance costs and only the cracks on the surface are

readily able to be repaired, and often not with good results. Any damage deep in the

structure is both hard to observe and to repair.11

In the past decades, self-healing polymers have become an interesting and

contemporary field of research, and different strategies have been investigated to

produce cross-linked polymers with self-healing ability. In this chapter, two main

kinds of self-healing polymers, irreversible and reversible, are illustrated. The first is

based on the healing agents stored discretely within “reservoirs” in the polymer

system, which can be released by mechanical stimulus to fill the cracks and solidify.

The latter is based on the incorporation of chemical functional groups in healing,

often activated by external stimuli such as heat and light. Related research on these

different self-healing polymers is reviewed below.


7

2.2 Irreversible self-healing polymers

Traditional thermosetting polymers are unable to cleanly revert to their monomers,

oligomers or uncross-linked polymers in a reversible manner. In order to induce self-

healing ability within them, healing agents are used and stored in the polymers to

repair the cracks or damage autonomously or triggered by external stimuli, mimicking

the healing process in nature.12

This kind of irreversible self-healing polymer includes

polymer composites containing hollow glass fibres, microcapsules, microvascular and

linear polymers.

2.2.1 Hollow glass fibres

Fibres are usually used in polymer composites to reinforce strength for structural

applications, such as in bridges and racing cars. However, the main disadvantage of

such materials is that damage by impact can lead to premature failure in their

structural capability. The concept of healing agents incorporated into the hollow fibres

would allow the agents to be released if a crack propagates to damage both the

composite and the fibres. Of the two main classes of fibres used in composites for

structural enhancement, glass and carbon fibres, glass fibres are more readily useful

for self-healing use. They are currently the more common reinforcing component in

cross-linked polymers and it is easier to gain a desired and suitable structure for the

storage of the healing agents in them.13

Three ways have been used to date, separately or in combination, to store healing

agents, including resins like cyanoacrylate in one-part, the resin and the hardener are

separately placed into two different fibres which are incorporated into the matrix in an

orthogonal manner, and where resin is filled into fibres whilst the hardener exists in

microcapsules outside and close to the fibres (Figure 2.1).14


8

Figure 2.1 Schematic diagrams of self-healing polymer composites using hollow glass

fibres with three filling approaches.14

Dry15

and Li et al.16

pioneered the study of the use of healing agents stored in hollow

fibres to heal cracked engineered cementitious composites (bendable concrete).

Several chemicals were used as healing agents, including cyanoacrylate17

, ethyl

cyanoacrylate16

and methyl methacrylate18, 19

. Follow this concept, Motuku and co-

workers prepared self-healing polymer composites reinforced with different fibres to

contain healing agents and found that glass fibre was the most suitable for use in this

method.20

Bleay and his colleagues14

used small hollow glass fibres as structural

reinforcements and as the container for the healing agents to produce a self-healing

epoxy composite. The filling of healing agents into hollow glass fibres was developed

using a vacuum and it was found that a two-part system (separated resin and hardener)

was better to operate than a one-part system (self-cured healing agent). This is

because the fibre could become blocked by self-curing of the healing agent and

monomer in the one-part system during the filling process. Further research on the

production of self-healing composites with large-diameter hollow fibres containing

two-part healing resin was reported by Bond and co-workers.21

The healing efficiency


9

of freshly prepared samples at room temperature was 93 % based on flexural strength

testing. When the damaged sample was heated at 100 oC for 2 h, the efficiency was

reduced to 87 % because of the initiation of a premature curing process before healing

agents had dispersed into the crack efficiently.22, 23

The hollow fibre concept was

found to be able to heal the cracks and restore the strength, but this technique still

needs further improvement, such as an improved filling method.

2.2.2 Microcapsules

Hollow polymeric microcapsules can also be used to cause self-healing. Similar to

hollow glass fibres, microcapsules can release the healing agents into the crack when

the polymer is damaged and the crack ruptures the microcapsules. As before, the

healing agents then cure and solidify to repair the crack by polymerisation

(Figure 2.2).7

Figure 2.2 Process of self-healing using microcapsules with healing agents. (a) Crack occurs

within the polymer. (b) Ruptured microcapsules release the healing agents into the crack.

(c) The healing agents polymerise and heal the crack.7


10

The preparation of such microcapsules usually uses mini-emulsion polymerisation in

two liquid phases, oil and water.24

Brown et al. prepared microcapsules with different

diameters from 10 m to 1000 m and 160 nm to 220 nm in thickness.25

Blaiszik

et al. reported their method to prepare nano-sized microcapsules from 220 nm to

1.65 m upon ultrasonication.26

The often-used healing agent is dicyclopentadiene (DCPD), which heals the materials

via ring opening metathesis polymerisation (ROMP) (as shown in Scheme 2.1) with

first generation Grubbs’ catalyst ruthenium (IV), because of its broad application,

stability and rapid polymerisation at ambient conditions. In 2001, White and co-

workers first reported this system as a means to produce autonomous, self-healing

epoxy polymers, in which the healing agent DCPD was stored in microcapsules and

the catalyst was directly dispersed into the polymer.7 The average healing efficiency

of this self-healing polymer was 60 %, obtained from fracture toughness testing. The

same researchers also studied the self-healing behaviour of polymers with different

contents of microcapsule and catalyst and showed that the highest healing efficiency

of 90 % was from the polymer containing 5 wt% microcapsule and 2.5 wt% catalyst.27

Liu et al. reported a mixture of 5-ethylidene-2-norbornene (ENB) and DCPD with the

ratio of 1:3 by weight, and the self-healing agent was highly effective.28

Due to the

instability of the first generation Grubbs’ catalyst, wax was used to encapsulate the

catalyst and it was found able to maintain more than 90 % activity.29

With only

0.75 wt% of the catalyst, the polymer showed a 93 % healing efficiency. Different

kinds of Grubbs’ catalysts were investigated but no further improvement in healing

efficiency was found.30


11

Scheme 2.1 Ring opening metathesis polymerisation of DCPD with Grubbs’ catalyst.7

In addition to the DCPD/Grubbs’ catalyst system, other systems have been

investigated. Braun and co-workers reported self-healing polymers based on the

condensation reaction between hydroxyl end-functionalised polydimethylsiloxane

(HOPDMS) and polydiethoxysiloxane (PDES) with organotin catalyst.31

The fully-

autonomous self-healing behaviour gave 75 % efficiency.32

Zhang et al. reported a

new system with uncured epoxy resin as healing agent and hardener as catalyst.33

More than 100 % healing efficiency was obtained from the epoxy polymer with 10 %

epoxy resin in microcapsules and 2 % hardener mixed into the epoxy polymer,

because the bonding materials not only healed the cracks but also provided higher

fracture toughness to the damaged sites. Using a mercaptan hardener, an autonomous

self-healing epoxy polymer using epoxy/hardener system was produced by the same

group.34

The sample with 2.4 wt% of each microcapsule achieved a healing efficiency

of 100 % after 12 h at ambient temperature.

2.2.3 Microvascular networks

Both the hollow glass fibre and microcapsule strategies used for self-healing polymers

have a primary disadvantage, in that the healing at the same location can only happen

once. In order to allow multiple healing processes, a new biomimetic material with


12

microvascular networks was invented, in which the healing agents could flow like

blood to the crack and autonomously repair multiple times.

Sottos et al.35

produced a composite polymer with a microvascular morphology

produced by soft lithographic and direct-write assembly methods according to the

research of Stroock and Cabodi36

. DCPD was used as the healing agent and was

placed into the 200 m wide channels, whilst Grubbs’ catalyst was dispersed in the

epoxy coating that was placed on the top surface of the microvascular substrate

(Figure 2.3). This system could repair the crack multiple times at the same location,

with the healing agents refilled after each healing process and the average healing

efficiency remaining at about 40 %. The same researchers further developed this

method using a dual-ink structure to enhance healing ability, with 13 cycles of

damage and healing still leading to a material with good properties.37

Figure 2.3 Microvascular networks in self-healing polymers. (a) Human skin with

microvascular network structure. (b) 3D microvascular network in the polymer substrate. (c)

High-magnification cross-sectional image of the coating showing that cracks spread from the

surface to the microvascular openings at the interface (scale bar = 0.5 mm). (d) Optical image

of released healing agents cured with 2.5 wt% catalyst after cracks (scale bar = 5 mm).35


13

Williams and co-workers designed and manufactured a sandwich structural self-

healing material with horizontal and vertical channels filled with epoxy resin and

hardener, respectively.38

Sottos and his colleagues continued this method and used the

replenishment technique to accomplish multiple healings up to 16 times, with an

average healing efficiency of 60 %.39

However, insufficient mixing of epoxy resins

and hardeners in the crack limited the promotion of the healing efficiency.

2.2.4 Linear polymer/network systems

Linear polymers, which can melt and flow when the healing temperature is above the

melting point (or above the glass transition temperature for fully amorphous systems),

can be physically mixed with cross-linked polymers, and confer on the composites

self-healing properties based on a physical mechanism. Such a linear

polymer/network system maintains good mechanical properties with the necessary

thermoplastic polymer mobility at temperatures relevant to thermoplastics, which

allows self-healing.

Zoko and Takano originally prepared a glass/epoxy composite laminate containing a

particle-type thermosetting epoxy adhesive in 1999.40

This system could heal initial

damage by melting the epoxy particles at 120 oC to flow. Jones and Hayes filed a

patent which described the use of a “solid solution” of thermoplastic and thermoset

polymers for self-healing composite materials, showing good healing efficiency.41

In

their following publication, a thermosetting resin embedded with thermoplastic

healing agent was studied.42

The study showed that the sample’s optimal healing

efficiency was 70 %. This solid-state, self-healing resin system was also used as a

matrix for composites with reinforcing fibres.41, 43

Mather et al. employed

poly(-caprolactone) (PCL) as thermoplastic healing agent, which was mixed with

epoxy resin, and the mixture was cured to form a self-healing material.44

A similar

system was used in self-healing coatings and allowed healing of the damage, both


14

structurally and functionally.45

The same researchers subsequently used the partly

cross-linked PCL to form a linear/network PCL blend with shape memory assisted

self-healing properties.46

After treatment at temperatures above the melting transition

temperature (Tm), the blends containing 25 wt% or higher linear PCL showed 95 % or

higher self-healing efficiency and good shape memory properties.

2.3 Reversible self-healing polymers

2.3.1 Diels-Alder reaction based systems

Diels-Alder (DA)-based systems are widely used in self-healing polymers because the

DA moiety is thermally cleavable and remendable. The DA reaction is an organic

reaction between conjugated diene and substituted alkene, forming the substituted

cyclohexene system (Scheme 2.2). It requires very little energy to create the useful

cyclohexene ring with two new covalent bonds. This reaction was first reported by

Otto Paul Hermann Diels and Kurt Alder in 1928, and was named after them.47

According to the mechanism, only the cis conjugated diene can form the ring via DA

reaction, and cyclic dienes like furan and cyclopentadiene are much easier to react.

Most of the DA reactions are thermally-reversible, with the decomposition reaction of

the cyclic system called the Retro Diels-Alder (RDA) reaction (Scheme 2.2).

Scheme 2.2 General mechanism of Diels-Alder (DA) and Retro Diels-Alder (RDA) reaction.

To date, three of the most studied DA-based polymers include furan-maleimide-based

polymers, dicyclopentadiene-based polymers and anthracene-maleimide based

polymers.


15

2.3.1.1 Furan-maleimide based system

In 1969, Craven first incorporated adducts of furan and maleimide groups into

polymers to produce new materials with thermal reversibility.48

Since then, there has

been a large volume of published studies on furan and maleimide systems in the area

of thermally-reversible polymer research. Furan and maleimide groups can either be

attached to the thermoplastic polymer chain to achieve thermal cross-linking and

decross-linking, or be directly used for reversible DA polymerisation.2, 4

In 1979, Stevens and Jenkins successfully attached maleimide functional groups to the

main chain of polystyrene by maleimidomethylation, and studied several cross-linking

reactions, including DA cross-linking of pendant maleimide groups with furan

groups.49

Little further research was undertaken on the use of pendant furan and

maleimide groups via DA and RDA reactions for reversible cross-linking until the

early 1990s, when the topic of thermally-reversible polymer came to the fore.50

Furan

functional groups can be connected to one copolymer generated from styrene and

acryloyl chloride, or to the copolymer produced from styrene and maleic anhydride.

Thermally-reversible cross-linking with the maleimide-modified polystyrene in

acetone was studied on the copolymer with pendant furan groups because of its good

solubility, and the results showed that this reversible process, as shown in Scheme 2.3,

could be repeated five times. However, the RDA temperature of this reaction was as

high as 150 oC, which limited the application of such a system in polymers due to side

reactions and application conditions.

Saegusa et al. explored the reversible process between furan-modified

poly(N-acetylethylenimine) (PAEI) and maleimide-modified PAEI, and first prepared

a thermally-reversible polyoxazoline hydrogel (Scheme 2.4).51

They further studied the

gelation of the hydrogel by the DA reaction and subsequent cleavage by the RDA

reaction, proving that their furan-maleimide system was effective.52


16

Scheme 2.3 Gelation and reversal of gelation of the modified polystyrene and copolymer.50

Scheme 2.4 The cross-linking and decross-linking process of a thermally-reversible

polyoxazoline gel.51

Gandini et al. attempted to use N,N’-methylenediphenylbismaleimide (MDPBM), a

widely used commercial bis-maleimide, to study the reactivity of two series of furanic

random copolymers with furan group at different positions.53

In contrast to the

readily-achievable cross-linking of such polymers using the DA reaction, the highly

cross-linked gel did not show any indication of the RDA reaction, other than using

excessive 2-methylfuran as “trap” to catch the bis-maleimide and push the RDA

reaction to happen.54, 55

More investigations were undertaken and led to a new

research field in furan-maleimide-based polymers.56

Huglin et al. also worked on a copolymer which contained pendant furan groups and

was cross-linked with MDPBM. The kinetic process of the conversion of DA and

RDA between poly(styrene-co-furfuryl methacrylate) and MDPBM shown in

Scheme 2.5,57

was studied, and reaction rates, conversion ratios and activation energy

of the DA and RDA reactions were analysed.58, 59

The physical properties of the


17

synthesised cross-linked copolymer, such as glass transition temperature (Tg),

gel point, swelling and mechanical properties, were also investigated.60

Scheme 2.5 Representation of the reversible cross-linking process between poly(styrene-co-

furfuryl methacrylate) and MDPBM.57

While the furan-maleimide-based system has been used to cross-link the linear

polymer with pendant furan or maleimide groups, it has also been used to polymerise

multi-furan and multi-maleimide monomers. This research field originated from the

preparation of polyimides from the furan-maleimide-based system.61-64

Kuramoto and

co-workers reported a new, thermally-reversible polymer based on difurfuryl adipate

(DFA) and bismaleimidodiphenylmethane (BMD) (Scheme 2.6).65

The RDA cleavage

of the polymer occurred gradually at 90 oC, and DA reformation proceeded quickly at

60 oC. This recycle process was able to be repeated several times, although only 20 %

to 60 % of conversion was complete upon the cleaving and healing process.


18

Scheme 2.6 Thermal reversible polymer made from DFA and BMD.65

The furan-maleimide-based system has also been incorporated in dendritic

compounds. McElhanon and Wheeler prepared different generation dendrimers

(Figure 2.4).66

Whist the first generation dendrimer system fully recovered in

chemistry, the second generation and third generation were able to be healed only to

an efficiency of 40 % and 20 %, respectively, probably due to a combination of steric

and entropic influences. McElhanon and Wheeler also investigated the use of such a

furan-maleimide-based system in the preparation of epoxy resins (Figure 2.5) and

made a series of removable epoxy foams which can be dissolved using 1-butanol at

90 oC via RDA reaction.

67 Such epoxy resins have numerous potential applications in

encapsulation and adhesion, and can be used to partially heal cracks.


19

Figure 2.4 Structure of first-, second- and third-generation dendrimers.66

Figure 2.5 Structure of thermally-reversible furan-maleimide-based epoxy resins.67


20

Although many researchers have studied polymerisation based on furan-maleimide

system, bi-functional groups have been used and only linear polymers were formed

before 2002. Wudl’s group was the first to develop materials which were formed from

multi-functional groups of furan and maleimide to make highly cross-linked polymers

using the DA reaction.3 They also first tested the self-healing property of the thermal

reversible furan-maleimide-based cross-linked polymer. Monomers of tris-maleimide

(3M) and a tetra-furan (4F) were synthesised and used to form a polymeric material

after polymerisation at 80 oC for 3 h (Scheme 2.7). The resultant polymer was stable

and irreversible below 120 oC. After being heated at 130

oC for 25 min, 12 % of

adducts of furan and maleimide groups were scissioned. When the temperature of

treatment was increased to 150 oC, about 25 % of adducts was cleaved after 15 min.

The scissioned and separated furan and maleimide groups were able to fully reconnect

after being heated at 80 oC for 1 h. The cycle was successfully repeated 5 times,

which proved that the DA and RDA reactions were reversible in this polymeric

system. The mending efficiency was obtained by fracture toughness testing of

compact tension test specimens. The healing efficiency of the first mending compared

to the original sample was some 57 % using 150 oC as thermal treatment temperature.

When the temperature was reduced to 120 oC, the average efficiency reduced to 41 %.

The healing efficiency of the second mending compared to the first healed sample was

80 %, which was evidence that this material can be mended multiple times

successfully. From the experimental results, the mechanical properties of thermally-

reversible polymer, including the Young’s modulus of 4.72 GPa and the Poisson’s

ratio of 0.349, were similar to those of commercial cross-linked epoxy resins and

unsaturated polyesters. The polymer therefore had potential applications, such as self-

healing electronic packaging, to heal cracks caused by differences in thermal

expansion.


21

Scheme 2.7 DA and RDA reactions of the furan-maleimide-based self-healing polymer.3

In a similar fashion, two other remendable highly cross-linked polymers were

designed and prepared, in which the 4F monomer was retained, and the 3M monomer

was replaced by two different bis-maleimide monomers, 2ME and 2MEP

(Figure 2.6).68

Both 2ME and 2MEP showed lower melting points than 3M, but the

resultant polymers, 2ME4F and 2MEP4F, were still hard, colourless and transparent at

room temperature. The mechanical properties of 2MEP4F were similar to those of

3M4F, and both were stronger than those of 2ME4F, probably because of the presence

of the ethylenedioxy group in 2ME. For the same reason, the shape of the 2ME4F

sample changed before the temperature reached 110 oC and the healing efficiency of

2ME4F could not be studied. Using fracture toughness testing as the test for healing

efficiency, the first healing process (compared to the original sample) showed an

average recovery of 80 % after being heated at 115 oC for 30 min and 40

oC for 6 h.


22

The second healing process achieved a 78 % recovery, which again indicated that this

new polymer 2MEP4F also showed multi-healing behaviour with excellent efficiency.

Figure 2.6 Structure of 2ME and 2MEP.68

Plaisted and Nemat-Nasser examined the repeated fracturing and healing cycles of the

2MEP4F polymer using a new facture method, the double cleavage drilled

compression (DCDC) sample geometry.69

The apparatus is a long column of

rectangular cross section containing a through-thickness circular central hole with

small notches at its axial crowns. When a uniform axial compression is applied,

cracks initiate at the crowns and grow in an axially stable manner, along the mid-

plane of the sample (Figure 2.7). On the condition of the same length of the final

crack, the stresses to cause cracking in the healed sample and the original sample

were compared in order to obtain healing efficiency. This method has the unique

property that the cracked sample remains intact so that the failure surfaces remain

sufficiently close to allow subsequent healing. The results showed that the most

effective healing temperatures were between 85 o

C and 95 oC, for which full recovery

of fracture toughness could be obtained, and no dimensional change was found

following repetition four times. Although there are several limitations of this system,

including a low working temperature and very costly monomer synthesis, this ground-

breaking work gave new impetus to self-healing material research.


23

Figure 2.7 Left: schematic of DCDC sample geometry. The dotted lines represent the

location of the pre-crack and subsequent crack extension. Right: fracture and healing

sequence of a single sample. (A) Virgin sample with hole and pre-cracks visible; (B) sample

after first fracture event; (C) sample after first healing treatment; (D) sample after second

fracture event.69

Figure 2.8 Structure of TMI and TF.70

A number of related studies followed Plaisted and Nemat-Nasser’s work. Liu and

Hsieh used epoxy compounds as precursors for the preparation of tris-maleimide

(TMI) and tris-furan compounds (TF) (Figure 2.8).70

The polymer formed, after


24

treatment at 50 oC for 12 h, could be considered as an epoxy-based polymer and had

potential application in encapsulants and microelectronics due to the advantageous

properties of epoxy resins. The cracked sample was able to be recovered through the

DA adducts cleaving at 120 oC for 20 min, followed by chain reformation at 50

oC for

24 h, thus demonstrating self-healing ability.

Other research groups, such as that of Liu and his co-workers, modified a maleimide-

containing aromatic polyamide (PA-MI) with furan groups in order to achieve a furan-

containing polyamide (PA-F), and used it to produce cross-linked polymers

(PA-MI/PA-F) with the original PA-MI by DA reaction (Scheme 2.8).71

The film of the

cross-linked PA-MI/PA-F showed high toughness with Young’s modulus of 566 MPa,

an elongation at a break of 4.4 % and a tensile strength of 20 MPa. The self-healing

property was demonstrated by healing a notch on the surface, via the RDA reaction at

120 oC for 3 h and the DA reaction at 50

oC for 5 d. However, as observed under the

microscope, the healing was not complete due to the low mobility of the high

molecular polyamide chains.

Scheme 2.8 Thermally-reversible DA/RDA reactions between PA-MI and PA-F.71

In another system, Watanabe and Yoshie terminated poly(ethylene adipate) (PEA)

with furan groups and copolymerised the product PEA2F (Figure 2.9) with bis- and

tris-maleimides, 2M and 3M, to produce linear and cross-linked recyclable polymers,

PEA2F2M and PEA2F3M, respectively.72

Both of the polymers were able to be

converted to their constituent monomers by RDA reaction at 145 oC for 20 min, and


25

were able to be reformed by the DA reaction after treatment at 60 oC for 15 h. The

recyclability of the elastic PEA2F3M was up to 8 times, with the similar values of

tensile strength and elongation at breakage found after healing, with a slight reduction

in tensile modulus compared to that of the original sample.

Figure 2.9 Structure of 2M and PEA2F.72

In recent times, bis(hydroxymethyl)furan (BHF) (from reduction of biomass derived

5-hydroxymethylfurfural (HMF)) and succinic acid (SA) (from microbial

fermentation of sugars) were combined to synthesise linear

poly(2,5-furandimethylene succinate) (PFS), which was cross-linked via DA

cycloaddion with 1,8-bis(maleimido)-triethylene glycol (M2) to form a self-healing

polymer (Scheme 2.9).73

The mechanical properties of this bio-based polymer PFS/M2

could be varied by adjusting the molar ratio of the furan group to the maleimide

group. A damaged sample could be self-healed by putting the broken surfaces together

at room temperature without any external stimulus, or healing could be triggered by

an M2 solution at room temperature, when the molar ratio of furan to maleimide

groups in the polymer system was more than 3:1. The best healing efficiency obtained

from the PFS/M2 with a ratio of 6:1 was 75 % by contact at room temperature for 5 d

and 90 % by treatment with 70 mg/mL M2 solution. If there were more unreacted

furan groups in the polymer due to the higher ratio of furan groups to maleimide


26

groups, the molecular mobility of the polymer chains was higher, resulting in faster

healing of the polymer. However, the resultant poor mechanical properties of this high

ratio material limited its application. This was the first publication to refer to room

temperature self-healing of a furan-maleimide-based system, which proved that the

reversible DA/RDA reaction could happen slowly at room temperature, given a

sufficiently long reaction period.

Scheme 2.9 Preparation of bio-based polymer PFS/M2.73

Gotsmann and co-workers combined a furan-protected bis-maleimide derivative with

tris-furan for the synthesis of a cross-linked DA polymer, with the ability to switch

between a rigid and highly cross-linked state at low temperatures and a deformable

and fragmented state at high temperatures via reversible DA reaction (Scheme 2.10).74

At higher temperatures, the furan-maleimide adducts in the cross-linked DA polymer

were scissioned by the RDA reaction so that the polymer became soft and readily

deformable, whilst at lower temperatures adducts reformed by the DA reaction caused

the polymer to recover to be a hard and stable solid that could retain the changed

structure on the surface. Its ability to switch physical states between high and low

temperatures meant that it was considered to have potential applications in data

storage and lithography.74


27

Scheme 2.10 Preparation of the cross-linked DA polymer and reversible DA/RDA reactions

between a rigid and highly cross-linked state and a deformable, cleaved fragmented state.74

Tian et al. synthesised a new type of epoxy resin, 1-(furan-2-yl)-N,N-bis(oxiran-2-

ylmethyl)methanamine (DGFA) which contained a furan group in two reaction steps,

and used it to form a healable epoxy with bis-maleimide, 1,1'-(methylenebis(4,1-

phenylene))bis(1H-pyrrole-2,5-dione) (DPMBMI), and curing agent of anhydride,

3a-methylhexahydroisobenzofuran-1,3-dione (MHHPA) (as shown in Figure 2.10).75

In the cured DGFA polymer, thermally-stable linkages were formed from epoxide and

anhydride groups, and yielded polymers with similar mechanical properties to

commercial epoxy resins, with the thermally-reversible DA linkages formed from

furan and maleimide groups giving the polymer new self-healing properties that could

heal scratches through the DA and RDA reactions. DSC experiments demonstrated

that, although the endothermic enthalpy of the RDA reaction peak decreased slightly

over four cycles of healing, the recyclability of DA and RDA reaction in the cured

DGFA polymer was still observable in the DSC curve. Visible damage made by the

impact of an iron ball on the surface was largely healed after thermal treatments at

125 oC for 20 min and then 80

oC for 72 h, demonstrating that this polymer had a self-


28

healing ability though the cleavage and reconnection of the DA adducts within.

Following the same concept, Tian and his colleagues reduced the number of epoxide

groups and designed a new epoxy resin, 2-((oxiran-2-ylmethoxy)methyl)furan (FGE)

(as shown in Figure 2.10) with one furan and one epoxide group in the molecule to

increase self-healing efficiency.76

The two-step synthesised FGE was cured using the

same chemical method and a new polymer with greater reversibility was prepared.

Compared with fully reacted DGFA, the cured FGE polymer demonstrated greater

cleavage reversibility, as observed from the consistency of the endothermic enthalpy

over a number of cycles and the stability of the position of the RDA reaction peak.

Plaisted and Nemat-Nasser’s DCDC test method69

was used to quantify the efficiency

of the thermal self-healing process of both cured DGFA and cured FGE polymers.

After being heated at 110 oC for 20 min to open the DA adducts, and then 80

oC for

72 h to close the DA adducts, the healed sample of cured FGE showed an average

self-healing efficiency of 96 %, which was higher than that of cured DGFA (66 %).

Figure 2.10 Structures of DGFA, DPMBMI, MHHPA and FGE.75, 76

Zhang and co-workers converted polyketone to furan derivative (PK-furan) by the

Paal-Knorr reaction and cross-linked it with bis-maleimide to yield highly cross-

linked thermosets, the remendability of which allowed full recovery following

fracture using the RDA and DA reaction sequence (Scheme 2.11).77

The thermally-

reversible DA and RDA reactions in the cross-linked PK-furan could be easily

displayed by the gelation and degelation of a mixture of PK-furan and bis-maleimide

in DMSO solvent (Figure 2.11), which also implied the recycle ability of the thermoset


29

polymer. The Tg from DMA and the load to displacement behaviour from three-point

bending tests showed that self-healing efficiency was almost 100 % over three cycles

of breaking and healing.

Scheme 2.11 The DA and RDA reactions between PK-furan and bis-maleimide.77

Figure 2.11 (a) Gelation and degelation for PK-furan with bis-maleimide in DMSO solvent:

(1) initial mixture, (2) gel formation at 50 oC after 2 h, (3) back to soluble solution in 5 min

at 150 oC; (b) insoluble cross-linked PK-furan in DMSO solvent (upper) at 50

oC and fully

soluble solution (bottom) in 5 min at 150 oC.

77


30

2.3.1.2 Dicyclopentadiene-based system

This is another chemical system that has aroused much interest in the self-healing of

polymers. Cyclopentadiene (CPD) has a similar structure to that of cyclic conjugated

diene as furan, in which the methylene group takes the place of oxygen. Like furan,

CPD can also form a new cyclohexene ring with a substituted alkene or dienophile via

a DA reaction. In addition, CPD has an interesting advantage that it can also act as

dienenophile, which gives it the ability to produce dicyclopentadiene (DCPD)

undergoing a self-DA reaction.13

The DCPD system was discovered in the 1920s by

Staudinger and Bruson, who used CPD to prepare a polymer via self-DA

polymerisation.78, 79

In 1961, Stille and Plummer synthesised three different

homopolymers with high molecular weight from three bis-CPDs monomers

(Scheme 2.12).80

No research on the reversibility of the DCPD system was published until Kennedy

and Castner first reported the thermal reversible DA/RDA reaction in this cross-linked

system (Scheme 2.13).81, 82

Polymers with pendant CPD groups were prepared from

chlorine-substituted alkane polymers and dimethyl cyclopentadienyl aluminium. Two

cross-linking reactions of CPDs could occur in this polymer: cationic cross-linking

caused by the dimethylaluminium chloride generated from the substitution reaction,

and DA cross-linking, which occurred at room temperature. The DA cross-linking was

cleaved at 215 oC by the RDA reaction, and the insoluble polymer became soluble in

the presence of maleic anhydride, which trapped the CPD group. Research on the

synthesis and characterisation of CPD telechelic polyisobutylene, which had CPD

terminals to extend the polymer via the DA reaction, has also been reported.83-85


31

Scheme 2.12 Polymers prepared via self-DA reaction.80

Scheme 2.13 Polymers cross-linked by DA reaction and decross-linked by the RDA reaction

in the presence of maleic anhydride, allowing eventual dissolution.81, 82

In the following two decades, although different methods were used to attach CPD

groups to linear polymers (such as polystyrene86

, polyphosphazenes87

and different

chlorine-containing polymers88, 89

) or monomers (such as 1,4-dichlorobutane,

1,4-dibromobutane90

and epichlorohydrin91

), and the reversibility of the resultant

polymers was studied, no research on self-healing property of the CPD-based

polymers was reported until Wudl’s group published their results for remendable

materials92, 93

. A series of monomers with DCPD units were designed and synthesised.

This family of monomers, such as monomer 400, could melt at high temperature to

form two CPD groups by the RDA reaction, and were subsequently polymerised to


32

yield linear polymer segments by the DA reaction (Scheme 2.14). In addition, the

segments could crosslink to produce a single-component polymer network via a

second DA reaction, which formed a tricyclopentadiene (TCPD) structure

(Scheme 2.14). These hard, colourless and transparent cross-linked polymers could

self-heal cracks after treatment at about 120 oC for 20 min, with an average self-

healing efficiency of about 46 % based on a fracture test. In this polymer series,

additional oxygen was incorporated into the structure with different diols. The final

polymer was more brittle and difficult to heal because of the higher cross-linking

density.92

These CPD-based polymers also showed shape memory behaviour after

healing (Figure 2.12).

Scheme 2.14 Cross-linked polymers prepared from monomer 400 via DA reaction.92


33

Figure 2.12 Polymer 400 (a) damaged after compression testing and (b) recovered to its

original shape after healing.93

2.3.1.3 Other DA-based systems

Some other DA-based systems have been reported, including anthracene-based

polymers. The centre phenyl ring in anthracene plays a role similar to the furan group

as cyclic-conjugated diene and reacts with dienenophile to form adducts by DA

reaction. In 1979, Grigoras et al. first synthesised anthracene-based polymers from the

polymerisation of anthracene monomers on thermal treatment or in the presence of

Lewis acid as catalyst like titanium tetrachloride (Scheme 2.15).94

However, a

side-polymerisation process leading to the formation of anthracene-substituted

poly(methyl methacrylate) was also observed in the system.95

Scheme 2.15 DA polymerisation of (9-anthry1)methyl acrylate and methacrylate.94


34

Stevens designed and synthesised a new type of monomer, N-(2-anthryl) maleimide

(Figure 2.13a), which could yield high molecular weight polymer via DA reaction.96

Subsequently, Grigoras’ group followed the same concept and synthesised

N-(1-anthryl) maleimide (Figure 2.13b).97

Later, Grigoras reported the synthesis of a

bis-anthracene compound and copolymerised it with different bis-maleimides to form

new DA-type oligomers.98

Figure 2.13 Structure of (a) N-(2-anthryl) maleimide and (b) N-(1-anthryl) maleimide.96, 97

Although Grigoras and Stevens reported much research on the anthracene-based

system, the reversibility of this system was not mentioned. Collard and co-workers

reported an investigation of the thermally-reversible DA cross-linking of a modified

poly(ethylene terephthalate) (PET) copolymer, in which 2,6-anthracene units were

introduced (Scheme 2.16).99

Bis-maleimide was used to cross-link the copolymer at

125 oC for 1.5 h. Model experiments showed that the polymer network could only

break slowly at 250 oC by RDA reaction. However, since treatment at such high

temperatures for a long time also led to the decomposition of the copolymer, no good

reversible result was found. To date, no further investigations on the self-healing

property of anthracene-based polymer have been reported.

In addition to the anthracene-based system, Brand and Klapper published their

research on thermally-reversible DA copolymerisation between a bis-furan

(α,ω-bis(3-furylmethyl)-pentaethylene glycol) and a bis-acrylate oligomer

(α,ω-bis(trans-4,4,4-trifluorocrotonylethyl)-polyethylene glycol) (Scheme 2.17),100


35

which can be regarded as a simplified furan-maleimide-based system. The copolymer

underwent the cycle of cycloreversion reaction at 150 oC and cycloaddition reaction at

80 oC 12 times, without any side reactions under inert conditions. The viscosity of the

copolymer was controlled by changing the molecular weight and the degree of

polymerisation was adjusted by choosing appropriate reaction temperatures.

Scheme 2.16 Cross-linking of a modified PET with bis-maleimide.99

Scheme 2.17 Reversible DA polymerisation between bis-furan and bis-acrylate oligomer.100


36

2.3.2 Thiol-based systems

Two thiol groups can be combined to form disulfide bond by oxidation easily in a

polymer system, generating cross-linked polymers. This type of disulfide bridge can

also be cleaved, having the capability to return to two thiol groups by a reduction

reaction. This reversible process has therefore been used in polymer systems to

produce reversible cross-linked polymers.

Although thiol-containing polymers (also called mercaptan-containing polymers)

have been synthesised and researched from the 1950s,101, 102

until 1993, Chujo and his

colleagues were the first to report their investigation of the reversible reactions of the

thiol groups in polymer.103

A disulfide bridge was formed and cleaved in a redox-

reversible hydrogel prepared from a thiol-protected poly(N-acetylethylenimine)

(PAEI) by oxidation in air and then reduction using excess amounts of reducing

agents such as sodium borohydride in ethanol (Scheme 2.18). Tesoro and Sastri used

reversible disulfide bridges to synthesise a polyimide in the same year.104

Scheme 2.18 Reversible reactions between thiol groups in hydrogel by oxidation and

reduction.103

Tsarevsky and Matyjaszewski used the method of atom transfer radical

polymerisation (ATRP) and directly obtained a well-defined polystyrene with

reversible disulfide bonds at 90 oC, employing the diester of 2-bromopropionic acid


37

and bis(2-hydroxyethyl) disulfide as a bi-functional initiator and the copper (I)

bromide complex of pentamethyldiethylenetriamine as the catalyst.105

In the presence

of dithiothreitol (DTT), the disulfide bridges were reduced and cleaved to polystyrene

blocks with thiol terminals, which were oxidised by FeCl3 at 60 oC for 24 h to reform

the original polystyrene with high molecular weight (Scheme 2.19). Both the reducing

and oxidising processes were efficient and complete.

Scheme 2.19 Reversible reactions of disulfide bonds in polystyrene.105

Although thiol-based systems show good reversibility in experiments, due to the

dramatic difference between the conditions of reduction and oxidation and the

inconvenience of condition change, no self-healing polymers based on this system

using reduction and oxidation have yet been reported.

Recently, another property of disulfide groups, exchange reaction at moderate

temperature (Scheme 2.20), has been used to design and build a self-healing system.

Klumperman and co-workers employed this new self-healing concept to synthesise

cross-linked epoxies with different concentrations of disulfide groups.106

This kind of

rubbery thermoset showed a good healing ability using tensile testing for three healing

cycles, with the elongation at break of each healed sample being some 63 %, nearly

the same as that of the original sample (65 %). Matyjaszewski et al. synthesised a

poly(n-butyl acrylate)-based star polymer using ATRP and subsequently cross-linked

the arms by chain extension with disulfide-containing dimethacrylate.107

The disulfide

cross-linked star polymer underwent a redox process to cleave and re-crosslink,


38

yielding new star polymers with different disulfide cross-linkers. The resultant cross-

linked gel showed good self-healing properties at room temperature without any

external stimulus due to the disulfide exchange reaction.108

The healing capability was

determined by the relationship between initial film thickness and the width of crack.

In order to provide the chance of exchange reaction, the Tg of this material should be

low enough to allow the polymer to flow slightly at room temperature.

Scheme 2.20 Interchange reaction between disulfide groups.106

Scheme 2.21 Mechanism of radical transfer reaction and radical crossover reaction.109


39

The same research group implanted the iniferter unit of thiuram disulfide (TDS) into

polyurethane and demonstrated the self-healing behaviour of cross-linked

polyurethane rubber at room temperature by visible light.109

The radical transfer

reaction among several TDSs and the radical crossover reaction between two TDSs

(Scheme 2.21) were considered to be the mechanism of self-healing. Following a

similar idea, trithiocarbonate (TTC) units, that can act as photoinitiators, were used to

cross-link poly(n-butyl acrylate) (PBA) and lead to the self-healing property of the

resultant polymer, so that the completely separated polymer pieces could form one

block upon exposure to ultraviolet irradiation at 330 nm.110

2.3.3 Alkoxyamine-based systems

Similar to thiol-based systems, alkoxyamine-based systems are also easy to recover

when the C-ON bonds break to form two radicals. The alkoxyamine unit has the

advantage of serving as the core of the reversible system and heals cracks or damage.

Otsuka and Takahara introduced alkoxyamine units to the polymer system, attaching

them to the side-chain and chain end. Through the cleavage of the C-ON bonds at

100 oC (Scheme 2.22), the pendant alkoxyamine groups can exchange with

alkoxyamine groups on other polymers to obtain graft copolymers111

or cross-linked

polymers112

. In the presence of an excess amount of alkoxyamine, the cross-linked

polymer converted to linear polymers after treatment at 100 oC for 24 h, which

showed that the cross-linking was reversible.

Scheme 2.22 Reversible reactions of alkoxyamine unit in polymers.112


40

Yuan et al. designed and synthesised a cross-linker containing alkoxyamine and

dimethacrylic ester, which polymerised with styrene to produce self-healing cross-

linked polystyrene.113

The dynamically-reversible C-ON covalent bonds in the cross-

link parts could frequently break and rapidly recombine to complete the exchange of

the crosslinking units upon heating, thus enabling the cracked polymer to recover over

a number of cycles. The average efficiency of healing was about 70 % in five cycles

of breaking and healing, when the molar ratio of styrene to cross-linker was 7.5:1 in

the cross-linked polymer. The possibility of the irreversible combination between two

methylene radicals decreased the self-healing efficiency and may limit the

applications, along with the reaction of radicals with unreacted monomers.

2.3.4 Photodimerisation-based systems

In comparison to using heat as a stimulus to heal, light is a highly desirable stimulus

because it is cheap, clean and easy to apply, although the depth to which it can

penetrate with sufficient intensity may always be an issue. The cycloaddition reaction,

for example, can be initiated by photo light, and cyclisation and cleavage can

therefore be achieved by employing different wavelengths of ultraviolet (UV) light,

which is the mechanism of photo self-healing in cross-linked polymers. Compounds

with reversible photodimerisation properties include coumarin, nitrocinnamate,

anthracene and cinnamic acid.

Saegusa’s group incorporated photosensitive coumarin groups containing -ketoester

into the polyoxazoline system, attached to the main chain.114

The modified

polyoxazoline crosslinked to form insoluble hydrogel by cyclobutane ring upon UV-

induced [2+2] cycloaddition of the pendant coumarins, with a 450 W high-pressure

Hg lamp (> 300 nm) for 7 h. The gel cleaved to soluble polyoxazoline via exposure

to UV from a low-pressure Hg lamp ( = 253 nm) (Scheme 2.23). Andreopoulos and

his colleagues modified a novel eight-branched poly(ethylene glycol) (PEG) with


41

nitrocinnamate groups.115

The PEG with pendant nitrocinnamate underwent photo

cycloaddition to produce hydrogels by 365 nm UV irradiation without any

photoinitiator or catalyst. The photoscissile gel was exposed to UV with 254 nm

wavelength and photocleaved. Subsequently, nitrocinnamate groups were replaced

with anthracene units, which exhibited high photosensitivity and rapid reversible

[4+4] photodimerisation.116

Coursan et al. attached anthracene groups to one end of

the polystyrene chain and investigated the reversible photodimerisation upon

alternative UV irradiation.117

The cycle of polystyrene dimerisation under UV

irradiation at 366 nm and dissociation at 280 nm occurred more than ten times without

any detectable degradation.

Scheme 2.23 Reversible photo-induced cross-link of polyoxazoline with pendant

coumarins.114

Chung and co-workers studied the self-healing properties of a newly designed

polymer cross-linked from 1,1,1-tris-(cinnamoyloxymethyl)ethane (TCE)

(Figure 2.14) by the [2+2] photocycloaddtion of cinnamoyl groups at 280 nm.118

The

hard and transparent cross-linked polymer was damaged and then exposed to UV

irradiation at 280 nm for 10 min to recover. Although the healing efficiency was only

14 % from the results of flexural strength testing, this is the first and only report of the

application of reversible photodimerisation in self-healing investigations to date.


42

Figure 2.14 Structure of 1,1,1-tris-(cinnamoyloxymethyl)ethane (TCE).118

2.3.5 Hydrogen bonding systems

A hydrogen bond is normally formed between hydrogen from the NH or OH groups

and highly electronegative atoms like nitrogen or oxygen by electromagnetic

attractive interaction. Although this electrostatic dipole-dipole interaction is weaker

than a covalent bond, it plays an important role in nature, such as the formation of the

double helical structure of deoxyribonucleic acid (DNA) and the stability of proteins.

The hydrogen bonding system has also been used to form linear polymers or cross-

linked polymers.119

Meijer and his colleagues employed the easily-synthesised 2-urido-4-pyrimidone

(UPy) as the linker attached at the end of the structure of polymer segments, and

prepared self-assembled linear and cross-linked polymers via quadruple hydrogen

bonding.120-127

The resultant complex could scission at about 90 oC and reform upon

cooling, demonstrating the ability to thermally self-heal (Scheme 2.24). Using the

same complex of UPys as cross-linker, a modified polycaprolactone used as coating

of the aluminium substrate, prepared by a company called SupraPolix, allowed

healing of scratches on the surface at 140 oC for 10 s.

128 The same company also

produced a hydrogel of PEG with UPy groups that allowed two distinct parts to be

connected together, upon applying pressure for several minutes.


43

Scheme 2.24 Reversible quadruple hydrogen bonding between UPy groups.127

Yagai et al. have reported the synthesis of binary supramolecular gels based on the

hydrogen bonding of bismelamine units with cyanurate or barbiturate units.129

The

cross-linked gels could turn completely to isotropic fluids after mechanical agitation

and recovered to the original state upon heating-cooling treatment. Rotello and co-

workers cross-linked diamidopyridine-attached polystyrene with bis-thymine via

hydrogen bonding, similar to base pairing in DNA.130, 131

The insoluble polymer was

dissolved at 50 oC and reformed when the temperature returned to room temperature

over multiple cycles.4

Figure 2.15 Structure of fatty diacid, fatty triacid, aminoethyl imidazolidone,

di(aminoethyl) urea and diamino tetraethyl triurea.132


44

Leibler and his colleagues synthesised supramolecular rubbers from the fatty acid

derivatives, diethylene triamine and urea.133

Five compounds (Figure 2.15) were

synthesised and associated together to form cross-linked rubbers in which both chains

and cross-links were connected by hydrogen bonding. When the material was cut or

fractured, it could recover to a consistent whole after bringing the pieces together for

some time at room temperature, even without heat.132

2.3.6 Metal-ligand coordination systems

In addition to hydrogen bonding, another main type of interaction to create

supramolecular polymers is the coordination between metal and ligand, which is

commonly seen in the structure of metal-containing compounds (especially transition

metals). Normally, several highly electronegative atoms such as nitrogen or oxygen

act as electron donors and provide lone electron pairs to bind to one central metal

atom or ion. This metal-ligand coordination system has been investigated in the

assembly of linear or cross-linked polymers.134

Schubert and his colleagues attached the ligand of 2,2’:6’,2’’-terpyridine (TPy) to the

end of diethylene glycol and used the Fe(II) as connecter to prepare water-soluble

reversible supramolecular polymer (Scheme 2.25).135

This Fe(II)/TPy coordination

complex could be dissociated by external factors such as high temperature or the

addition of strong chelating ligands like N-hydroxyethyl-ethylenediamine-triacetic

acid (HEDTA).

Scheme 2.25 Reversible metal-ligand coordination between Fe(II) and TPy.135


45

Rowan et al. employed another terdentate ligand 2,6-bis(benzimidazolyl)-4-

oxypyridine (BIOP) (Figure 2.16), which was attached to the end of three different

oligomers to coordinate a metal ion to produce linear supramolecular polymers.136, 137

Subsequently, the same group investigated the self-healing property of a cross-linked

supramolecular polymer consisting of poly(ethylene-co-butylene) ending with BIOP

and Zn(II).138

Cuts to a depth of 50 % to 70 % of the film thickness were fully healed

with 100 % healing efficiency for the material with 0.7 equivalent of Zn(II) upon

photothermal conversion of exposure to UV irradiation for two consecutive periods of

30 s via the cleavage, exchange and reform of the Zn(II)-BIOP coordination.

Figure 2.16 Structure of BIOP.136

The metal-ligand coordination between a bi-functional Pd(II) or Pt(II) pincer complex

and a pyridine group on the side-chain of PEG (Figure 2.17) was used to make

reversibly cross-linked PEG with self-healing property by Craig and his colleagues.139

The system could be cleaved and reformed with different applied stresses, so that it

could be applied as a stress-induced self-healing material to heal damage.

Figure 2.17 Structures of bi-functional Pd(II) or Pt(II) pincer complex and pyridine group

on the side-chain of PEG.139


46

Holten-Andersen et al. mimicked the process by which mussels make byssus and

designed a self-healing gel with pH-induced Fe(III)-catechol cross-links.140, 141

The

change of pH to deprotonate the hydroxyls in catechol could control the coordination

number of catechol to Fe(III) between 1 and 3, and changed the formation of polymer

to oligomer or cross-linked polymer. The gel which is damaged by shearing could

recover within 3 min after treatment.

2.3.7 Guest-host systems

Guest-host systems describe complexes formed depending on special structures of

large molecules such as cyclodextrin (CD) and crown ethers as hosts, which can hold

or trap small molecules or ions as guests by forces other than full covalent bonds.

Harada synthesised a supramolecular linear polymer based on the guest-host system

of two types of cyclodextrins.142

Later, his group added cyclodextrins on the side-

chain of poly(acrylic acid) (PAA) to produce the host polymers, PAA-CDs, which

formed cross-linked polymers, with the help of PAA with ferrocene (PAA-Fc) as guest

polymers.143

The cut parts could be rejoined to form one gel by holding pieces

together for 24 h or using redox agents.

Huang et al. prepared a poly(methyl methacrylate) (PMMA) polymer with pendent

dibenzo[24]crown-8 (DB24C8) groups to form two self-healing supramolecular gels

cross-linked by two different bis-ammoniums.144

To change the pH by adding a base

like triethylamine or an acid like trifluoroacetic acid, the bis-ammoniums were

neutralised to break the combination with DB24C8 or recovered to associate again

(Scheme 2.26). The self-healing property of the cross-linked gels was studied using

visual and rheological tests. The visible damage on the gels could recover in a short

time and both the storage and loss modulus of the healed samples were nearly the

same as the original samples. Additional work by the same group showed the


47

application of a similar system. The AA monomer with benzo-21-crown-7 (B21C7)

groups at each end was connected to the BB monomer with dialkylammonium salt at

each end to construct a linear supramolecular copolymer. Relying on the 1,2,3-triazole

groups in the BB monomer, the linear polymer could use these chemical

functionalities as ligands to coordinate with metal ions such as Pd(II), resulting in a

cross-linked supramolecular polymer with self-healing capacity.145

Because the

constitution of the polymer depended on non-covalent forces, the system responded to

multiple stimuli such as metallo-, thermo-, pH-, and cation-induced stimuli, resulting

in gel-sol transitions. This gel showed autonomous and rapid self-repair ability, and

the recovered sample maintained the same modulus and shape as the original sample.

Scheme 2.26 Reversible process in DB24C8/bis-ammoniums system.144


48

A novel idea for the use of - stacking interactions to prepare cross-linked

supramolecular polymers was reported by Hayes and co-workers.146

Two

-electron-poor diimides in chain-folding polydiimide trapped the -electron-rich

pyrene group at the end of polysiloxane, which was a thermal reversible process in the

polymeric system. A sample cut with a scalpel was heated to 115 oC and the damage

was barely visible.

2.3.8 Ionic systems

Many copolymers are able to incorporate ions within their structure. A copolymer, the

ion content of which is usually no more than 15 mol%, is called an ionomer. Ionic

interaction could allow reversible cross-linking in ionomers and endow them with

self-healing properties. DuPont originally researched and developed self-healing

copolymers, which included poly(ethylene-co-methacrylic acid) (PEMAA) with the

trade name of Nucrel® and sodium or zinc neutralised ethylene-co-methacrylic acid

(EMAA) ionomers with the trade name of Surlyn®.4

Kalista and co-workers investigated the self-healing ability of commercial products

from DuPont damaged by projectile puncture.147

The healing process had three stages.

After the pellet passed the film sample, the polymer was heated to about 98 oC above

the melting point of the sample during the puncture process, melted and elastically

closed the hole. The contacting, soft material repaired the damaged area by ionic

interactions during annealing. Finally, the continued interdiffusion, crystallisation and

relaxation of the polymer chains gradually increased the strength of the damage zone

to the original level. Other types of damage including sawing, cutting and nail

puncture was used to analyse the self-healing process and the results showed that the

healing of EMAA materials required energy such as heat generated from the damage

event like rapid sawing to give the damaged part sufficient energy for elasticity and

bonding.148

These results demonstrated that the sharp rise of the temperature at stage


49

one of the healing process was important, since it provided high elasticity and

mobility to the damaged part of the polymer, and was critical to the autonomous

healing of the ionic system. This concept has been well validated by further

studies.149-151

2.4 Summary

Different classes of self-healing polymers have been reviewed and discussed above.

Two main bases for self-healing has been outlined. The irreversible method is

designed using a physical mechanism, in which new monomers are used to heal

cracking or damage. The storage of the monomers and catalysts are mainly focused

and investigated. Hollow glass fibre, microcapsule and microvascular are used to

maintain the healing agents. The self-healing process starts with the damage to the

container. This self-healing is autonomous but only can happen once at the same site.

Reversible self-healing is based on the presence of particular chemical units which are

part of the molecular structure of the polymer. Normally, the unit has the ability to

cleave and reform. When the cleavable unit is on the crosslinking unit of the cross-

linked polymers, it can switch the polymer between being cross-linked and linear.

This chemical mechanism includes the reversible DA reaction, redox and interchange

of disulfide groups, reversible reactions of alkoxyamine unit, reversible photo-

induced crosslinking, interaction of molecules (hydrogen bonding, metal-ligand

coordination and guest-host bonding) and ionic bonding. This reversible process can

be controlled by heat, light, pH or redox processes, and by changing the conditions,

the self-healing behaviour can happen multi times. The techniques to characterise

self-healing ability include NMR to see the chemical restoration, fracture toughness

testing and double cleavage drilled compression to study the recovery of mechanical

properties, and microscopy to observe surface self-healing. Healing efficiency is most

often defined based on the results of testing of the recovery of mechanical properties.

Chapter 3 Materials and Analytical Techniques

50


In this chapter, all materials used in the research work in this thesis are listed. The

techniques used for chemical and physical characterisation of the resultant materials,

including their self-healing properties, are also described in detail.

3.1 Materials

The preparation of the new self-healing epoxy polymers in this study was based on

two epoxy resins with different structures and functionalities, the widely-used

bi-functional diglycidyl ether of bisphenol A (DGEBA), and tri-functional triglycidyl

p-amino phenol (TGAP) (Figure 3.1).

Figure 3.1 Structure of DGEBA and TGAP.

DGEBA (DER331), which is a clear, colourless and viscous liquid resin from the

reaction of epichlorohydrin and bisphenol A, was obtained commercially from ATL

Composites (Australia), with Mn of 370 g/mol and the epoxide equivalent weight

(EEW) of 185 g/eq.


51

TGAP, a light yellow, clear and low-viscosity liquid resin from the reaction of

epichlorohydrin and aromatic amines, was purchased from Sigma-Aldrich, with a

molecular weight of 277.32 g/mol and EEW of 92.44 g/eq.

The hardener with DA units was synthesised by the method shown in Chapter 4. The

commercial sources of the involved main starting materials are as follows:

1,8-Diaminooctane was purchased from Fluka; Maleic anhydride, furfurylamine and

di-tert-butyl dicarbonate were supplied by Sigma-Aldrich; all other reactants and

reagents were purchased from Merck and were used as received.

The hardener without DA units used to synthesise epoxy polymer, which was used for

comparison with epoxy polymer with DA units in Chapter 5 and 6, was

Ethacure 100 diamine (80 % 3,5-diethyl-2,4-toluenediamine and 20 % 3,5-diethyl-2,6

-toluenediamine) (Figure 3.2) from Sigma-Aldrich.

Figure 3.2 Structure of Ethacure 100 diamine.

3.2 Analytical techniques

3.2.1 Fourier transform infrared spectroscopy

Fourier transform infrared spectroscopy (FTIR) is a widely-used technique to achieve

chemical information about materials. When exposed to infrared radiation, the

molecules of materials absorb energy to produce vibrations and oscillations.

Molecules with different structures absorb light energy at different wavelengths, so


52

that information about chemical groups in molecule are identified from the FTIR

results.

In this study, the FTIR spectra from 4000 cm-1

to 600 cm-1

were recorded with a

Bruker EQUINOX 55 FTIR fitted with an MCT detector. The ATR sampler is a

Specac’s “Golden Gate” single bounce diamond ATR. The operating condition was

50 scans at a resolution of 4 cm-1

. The sample was cut from thin film samples and the

thickness was 0.5 mm.

3.2.2 Near-infrared spectroscopy

Near-infrared spectroscopy (NIR), as one form of FTIR, is based on molecular

vibrations upon radiation with light in the near-infrared region. Since the information

in the mid-infrared region can often be complicated and hard to analyse, NIR can be

used to investigate the chemical information from materials according to the

literature.152-154

In this study, NIR spectra were obtained from samples of some 0.8 mm thickness

using a Bruker EQUINOX 55 FTIR spectrometer with 32 scans at resolution of 8 cm-1

in the region 9000 cm-1

to 4000 cm-1

.

3.2.3 Differential scanning calorimetry

Differential scanning calorimetry (DSC) is a thermoanalytical technique used to

determine the thermal behaviour of materials. The technique works in the following

way: both the target sample and the reference sample are placed in aluminium pans

with covers at nearly the same temperature throughout the experiment and the sample

and reference are either scanned in temperature (heating or cooling) or held

isothermally. When the sample undergoes a physical transformation such as glass

transition or melting, more or less heat will need to flow to it than the reference to


53

maintain them both at the same temperature. From the record of the heat changes at

different temperatures, properties such as the glass transition temperature (Tg) of the

materials can be identified, as well as melting point Tm or crystallisation temperature

(Tc) on cooling.

In this study, all DSC experiments were performed with a Perkin-Elmer DSC-7 using

nitrogen purge (20 mL/min) and aluminium pans, with the weight of each sample

being about 10 mg.

3.2.4 Dynamic mechanical analysis

Dynamic mechanical analysis (DMA) is a thermomechanical technique used to study

and characterise polymers, especially their viscoelastic behaviour. Under applied

sinusoidal stress, the resultant sinusoidal strain in the materials is measured at

different temperatures (as temperature is scanned, for example) for a given frequency

of deformation. Many properties can be investigated, including Tg which can be

observed as a decrease in rigidity (Young’s modulus) or as a maximum of the

damping curve tan

In this work, the Tg was also obtained with a Perkin Elmer Dynamic Mechanical

Analyser DMA7 using penetration mode (probe tip placed on the surface) under an

inert atmosphere of helium (flow rate 35 cm3/min). The static force applied was

800 mN, the dynamic force was 500 mN, and the frequency was 1 Hz. The samples

for these experiments were prepared in a cylindrical silicon mould, with the diameter

being 5 mm and the thickness 2 mm.

3.2.5 Thermogravimetric analysis

Thermogravimetric analysis (TGA) is a method of thermal analysis, which

investigates the thermal stability of materials by measuring mass loss in the


54

atmosphere of a flowing inert gas as a function of the temperature increase at a

constant heating rate. TGA can also provide information about the physical and

chemical properties of materials.

In this study, thermal stability was tested by TGA with an EXSTAR TG/DTA6300 at a

heating rate of 10 oC/min under Argon atmosphere.

3.2.6 Swelling test

The swelling capacity of a polymer is the amount of liquid material that can be

absorbed. By using a standard swelling test, the swelling capacity can be measured to

study the internal structure of polymer, including its crosslink density and free

volume.

The samples exposed to solvents for the swelling test were thin films with a thickness

of 0.5 mm. The samples for the dissolution test were prepared in a small, cylindrical

silicon mould. The diameter was 3 mm and the height was 2 mm.

3.2.7 Gel permeation chromatography

Gel permeation chromatography (GPC) is a standard technique of chemical analysis

to determine the molecular weight distribution of linear and branched polymers. It is a

size-exclusion chromatographic method, which separates analytes on the basis of their

sizes. The polydispersity index and average molecular weight of a polymer can be

determined from this technique for the purpose of polymer characterisation.

In this study, the polydispersity indices and molecular weights of polymers were

determined by GPC, using a Tosoh Ecosec HLC-8320GPC equipped with both

refractive index and UV detectors (UV-detection, = 280 nm) using Tosoh alpha

4000 and 2500 columns. Dimethylformamide (DMF) containing lithium bromide

(10 mmol/L) was used as the solvent. Calibration curves were obtained using


55

polystyrene standards. The number average molecular weight (Mn) and the weight

average molecular weight (Mw) can be obtained, and the ratio of Mw/Mn represents the

polydispersity index.

3.2.8 Optical microscopy

Optical microscopy uses visible light and a system of lenses to magnify images of

small samples. The image from an optical microscope can be captured by normal

light-sensitive cameras such as charge-coupled device (CCD) cameras to generate a

micrograph. In this study, microscopy was used to determine the self-healing

properties of the samples.

Two different microscopes were used to capture images of the self-healing process.

The images of scratched and healed samples were recorded using a Nikon 80i

microscope and a Nikon DS-Fi1 CCD camera. The healing process of the scratched

surface was recorded using an Olympus BX60 microscope with 20X magnification

and a Canon Legria HFS20 camera. A Linkam THMS600 hot stage was used to heat

the sample under the microscope at a heating rate of 80 oC/min. The sample was

placed on a piece of glass, and the thickness of the sample was about 0.4 mm. The

scratch on the sample surface was usually some 50 m in width and 5 m in depth.

Chapter 4 Synthesis of Diamine Cross-linker

56


4.1 Introduction

In this chapter, the synthesis of structures of diamine cross-linker with DA units is

described. Initial attempts to prepare the diamine cross-linker with a single DA unit

were ineffective, as very low yields of the expected product were obtained. Later, an

alternate structure with double DA units in the diamine cross-linker was proposed,

and its synthesis was achieved with good levels of yield. The product was

characterised and shown to be of the desired structure.

4.2 Synthetic attempts to produce a diamine cross-linker

with single DA unit

The key aspect of this research has been the attempt to incorporate a DA unit, an

adduct of furan and maleimide moieties, within the structure of diamine cross-linker,

which can be cleaved and reconnected by DA/RDA reactions. The first synthetic

scheme proposed is shown in Scheme 4.1. Ethane-1,2-diamine was mono-Boc-

protected with di-tert-butyl dicarbonate (Boc) to give compound 1, using the method

reported by Skene et al.155

Compound 2 was obtained upon the reaction of maleic

anhydride and compound 1 to form maleimide group by condensation reaction, using

benzotriazole-1-yl-oxy-tris(dimethylamino)phosphonium hexafluorophosphate (BOP)

as catalyst, following the procedure reported by Liskamp.156

Subsequently, the Boc

group in compound 2 was deprotected in acid condition with HCl to yield

compound 3. Finally, with the reaction between furfurylamine and compound 3, the

diamine cross-linker with single DA unit (4) was prepared. This proposed scheme was

attempted, but unfortunately the yield of compound 3 in the Boc deprotection step


57

was found to be unreasonably low, and the final step from compound 3 to the desired

product 4 was unsuccessful, possibly because of some side reactions.

Scheme 4.1 Initial synthetic proposal of the diamine cross-linker.

An alternative method attempted was to form DA adduct from 2 and the

(furan-2-yl)methanamine before the deprotection of compound 2. However, an

unexpected product (structure shown in Figure 4.1) was obtained, probably due to the

Michael addition reaction, in which the carbon-carbon double bond in

,-unsaturated carbonyl compound readily reacts with the amine group. This

irreversible side reaction generated a Michael addition product, instead of forming the

desired DA units. Thus, a new strategy was required to prevent the occurrence of the

Michael addition reaction in the synthesis of a diamine cross-linker with DA units.

Figure 4.1 Structure of the unexpected product resulting from a Michael addition reaction

between amine and maleimide group.


58

Another route was proposed (Scheme 4.2), with the improvement involving the

protection of the amine group in (furan-2-yl)methanamine by Boc before its reaction

with compound 2. With this change, the side reaction was avoided. The hydrochloride

salt 7 was successfully synthesised and the total yield was 26 % over four steps. The

precise details of the synthetic method are illustrated in Section 4.5.2.

Scheme 4.2 Synthesis of diamine cross-linker with one furan-maleimide DA adduct from

ethane-1,2-diamine.

After compound 7 was obtained, the final step was the neutralisation of the

hydrochloride compound 7. Many methods were attempted to carry out this


59

neutralisation with the purpose of producing a pure free diamine cross-linker with a

single DA unit. Unfortunately, despite many attempts, the pure free diamine product

was not achieved. The 1H-NMR spectra showed that each product obtained was very

complex. The sodium chloride generated was irremovable from the product. Various

attempts were also tried to isolate the desired product from the mixture, but none were

successful. The following methods that were used are now outlined.

Initially, NaOH was used as a base, and CH2Cl2 was the organic solvent for

extraction. 1 mol/L NaOH solution was added to a solution of compound 7 (0.1 g) in

water (10 mL) to adjust the pH value to 8. The solution was then extracted with

CH2Cl2 (15 mL). The organic solution was tested using thin layer chromatography

(TLC), and no organic compound was found in the solution. The pH value in the

neutralisation was subsequently further adjusted to change the basicity from 10 to 14,

in attempts to obtain the desired product. However, TLC results showed no proof that

the desired product could be extracted from the basic solution using CH2Cl2.

Other weak bases, like triethylamine and ammonia, were also employed to neutralise

compound 7. Unfortunately, no expected product was extracted from the neutralised

solution using CH2Cl2. K2CO3 was also used to neutralise the solution of compound 7

in water, but this still did not resolve the issue.

Different organic solvents instead of CH2Cl2 were also tried in an attempt to extract

the product from the neutralised aqueous solution. Diethyl ether, hexane, ethyl

acetate, chloroform and tetrachloromethane were all not effective.

The water in the neutralised aqueous solution in the experiments above was then

removed by vacuum and each residue was tested by 1H-NMR, using D2O as solvent.

The results showed that the 1H-NMR spectra were complicated and not the expected

pure diamine product, however the product was possibly included in the residue. In


60

addition, the 1H-NMR results of the mixtures from different neutralisation conditions

varied.

Since no product was obtained from the neutralised solution using all the above

methods, it was thought that the product might have better solubility in water than in

organic solvents, and it might be feasible to collect the product by dissolving it in

organic solvents after the water was removed. In addition to the organic solvents used

previously for extraction, many other organic solvents which were soluble in water

could be used, including acetone, ethanol, acetonitrile and methanol. However, most

of the organic solvents were still not able to extract products from the residue.

Although the product may dissolve into solvents like ethanol or methanol because the

appearance of organic compounds in solvent was confirmed by TLC, the inorganic

salts were also soluble in this kind of solvent, and the 1H-NMR spectra of the product

were again complicated.

Base resin IRA-400(OH) was another candidate used to neutralise compound 7. The

mixture of compound 7 and the base resin in ethyl acetate was stirred for 2 h and then

filtered. The solvent was removed in vacuo and the residue was tested by 1H-NMR.

However, the spectrum remained complicated, and was not the desired free diamine

product.

From the NMR spectra of some of the mixtures above, it was possible that the desired

diamine was in the mixtures but mixed with other impurities. The poor solubility of

the target diamine in organic solvent was thus considered to be the main issue

preventing the purification of the product. Therefore, another diamine structure was

proposed, containing longer carbon chains between the two amine groups to achieve

better solubility of the desired diamine in organic solvents such as CH2Cl2, with the

desire that the free amine could be extracted out with an organic solvent after being

neutralised. An alternative structure of diamine cross-linker with furan-maleimide DA


61

adduct, compound 11, was therefore proposed, due to the unaccomplished

neutralisation of compound 7. 1,8-Diaminooctane was chosen as a starting material as

it would significantly increase the carbon number between the two amine moieties

from two to eight.

Scheme 4.3 Synthesis of diamine cross-linker with one furan-maleimide DA adduct from

1,8-diaminooctane.

A similar synthetic route as used for the synthesis of diamine 7 was adopted for 11.

Unexpectedly, the yield of the second step of forming the maleimide 9 was only 12 %,

which would significantly increase the cost of synthesis. To improve its yield, zinc


62

chloride and hexamethyldisilazane (HMDS) were used as catalysts instead of BOP,

inspired by Toru’s published work using them to catalyse the reaction forming

maleimide.157

The revised synthetic route is shown in Scheme 4.3 and the total yield to

hydrochloride salt, compound 11, was 62 %, over four steps. The precise details of the

synthetic method are illustrated in Section 4.5.2.

After obtaining the hydrochloride salt 11, 1 mol/L NaOH solution was used to

neutralise it and CH2Cl2 was employed for extraction. Unfortunately, no organic

compounds were obtained with CH2Cl2 solvent. The methods mentioned above, with

different bases, different extraction solvents, and dissolving in solvents after removing

water, were again attempted. However, once more no improved result was gained and

the situation was similar to that of neutralisation of hydrochloride salt 7.

In order to avoid the formation of hydrochloride salt, which could not be neutralised

and extracted, a new method to deprotect the Boc groups in compound 10, instead of

using HCl/Et2O, was attempted. Acid resins can be used to deprotect the Boc groups,

according to the literature.158

Amberlyst 15 was selected as the acid resin due to its

wide usage, and the method reported by Bergbreiter was adopted.158

Amberlyst 15 (0.5 g) was added to a solution of compound 10 (0.1 g) in CH2Cl2

(10 mL). The mixture was stirred at room temperature overnight. TLC was used to

monitor the reaction. When compound 10 was all consumed, as detected by TCL, the

mixture was filtered. The resin was washed with CH2Cl2, tetrahydrofuran and

methanol, successively. The resin was added to a solution of ammonia in methanol

(10 mL) and the mixture was stirred for 1 h before being filtered for absorption of the

amine from the acid resin. The solvent was removed, and the residue was tested by

1H-NMR. Unfortunately it was a complicated mixture, possibly including the

expected product and the yield was very low.


63

To ensure that impurity was not introduced from the resin, Amberlyst 15 was cleaned

before use, according to the following method.158

The acid resin was soaked in

methanol for 24 h, washed with methanol and then neutralised with a solution of

saturated ammonia in methanol. The resin was then acidified with 3 mol/L HCl in

50 % methanol and rinsed with methanol, tetrahydrofuran and then CH2Cl2. The

cleaned resin was used to deprotect compound 10 again. Still no improvement was

observed and the result was similar to that above, a complicated mixture.

Many methods were tried to obtain the desired diamine cross-linker, including

neutralisation of compound 7 or compound 11. Unfortunately, no suitable method was

found. The high hydrophilicity of the desired diamine cross-linker may be the one of

the reasons that it is hard to be extracted from the water layer using organic solvents.

This makes it difficult to purify because the inorganic salts like sodium chloride can

not be removed from the water layer. The other possible reason is the instability of the

desired diamine cross-linker. The compound may be decomposed during the

neutralisation process.

4.3 Synthesis of diamine cross-linker with double DA units

Due to the difficulties in synthesising the diamine cross-linker with a single DA unit,

a structure of the diamine cross-linker with two DA units was proposed. Scheme 4.4

outlines the synthetic route towards diamine 14.

1,8-Diaminooctane was treated with maleic anhydride under the catalysis of acetic

anhydride, nick (II) acetate and triethylamine in refluxed acetone to yield

compound 12, using the method reported by Kossmehl.159

Furfurylamine was

protected by Boc group to generate compound 5. The reaction of compounds 5 and 12

was refluxed in ethyl acetate to yield the Boc-protected diamine 13, being a mixture

of exo- and endo-isomers. After two Boc groups were deprotected, the resultant


64

hydrochloride salt was neutralised using NaOH solution and the desired diamine 14

was obtained with a total yield of 41 % over four steps. It was a very viscous liquid

with a brown colour and a mixture of exo- and endo-isomers, containing the same

structure as compound 13 (Figure 4.2). From the integration in 1H-NMR spectrum of

diamine 14, the ratio of exo/endo could be deduced as 6.7:1, with the exo-isomer

being the main product.160-163

Although the endo-isomer is initially formed in the

reaction due to its kinetic stability, the thermodynamic stability of exo- means that

most of the endo-isomer transform to the exo- as the reaction proceeds. However, a

pure exo- or endo-isomer is unnecessary to produce cross-linkers capable of self-

healing, because both of the isomers can undergo DA and RDA reactions. This is the

first time, to our knowledge, that this kind of diamine cross-linker 14 has been

synthesised and reported. The greater concentration of DA units in the structure

increased the hydrophobicity of the compound 14 so that it is easy to dissolve in the

organic solvent and to be obtained.


65

Scheme 4.4 Synthesis of diamine cross-linker with two DA units.

Figure 4.2 Structure of endo- and exo-isomers


66

4.4 Summary

In this chapter, the synthetic routes and methods of the synthesis of diamine cross-

linker with DA units, adducts of furan and maleimide groups, were investigated.

Initially, feasible procedures to synthesise a diamine cross-linker with single DA unit

were explored. After several attempts at different synthetic routes, the hydrochloride

salt of the desired diamine was successfully obtained. Many methods were tried to

neutralise the hydrochloride salt to achieve the pure product, including using different

bases, changing the pH value of the solution, employing different organic solvents,

and purifying the residue from the water solution, but none were successful. The

NMR spectra generally showed that there was organic compound in the generated

residue and that the desired free diamine product probably existed, but its poor

solubility in organic solvent was the key problem leading to difficulty in extracting

the free diamine. In order to increase its solubility in organic solvents, another target

structure was proposed, with the carbon number between two maleimide groups in the

diamine cross-linker increased to eight. A similar synthetic route was applied, with an

improvement in the step involving the formation of the maleimide to optimise the

yield. The hydrochloride salt of this desired final product was obtained. However, the

result of trying to neutralise the hydrochloride salt was still unsatisfactory and no pure

free diamine was obtained. Since no feasible procedure could be developed after

many attempts, the structure of the initial target diamine was modified.

A new diamine cross-linker was designed with two DA units in the structure, and this

was synthesised successfully. The synthetic route was optimised and the target

diamine cross-linker was obtained with a 41 % yield over four synthetic steps. This

type of diamine cross-linker with DA units is the first of its type to be synthesised and

reported.


67

4.5 Experimental

4.5.1 General

1H-NMR spectra were carried out on a Bruker DRX 400 and chemical shifts were

calibrated by CDCl3 (δ = 7.26), D2O (δ = 4.79) or DMSO-d6 (δ = 2.5).164

13

C NMR

spectra were recorded on a Bruker DRX 400 and chemical shifts were calibrated by

the signal of CDCl3 (δ = 77.16) or DMSO-d6 (δ = 39.52).164

Mass spectra were

recorded on a Micro mass platform. Fourier transform infrared spectra were recorded

with a Bruker EQUINOX 55 FTIR fitted with an MCT detector. The ATR sampler

was a Specac’s “Golden Gate” single bounce diamond ATR. The operating condition

involved 50 scans, leading to a resolution of 4 cm-1

.

4.5.2 Synthetic attempts via mono-maleimide intermediate

1) Synthesis of tert-butyl (2-aminoethyl)carbamate (1).

According to the literature155

, a solution of di-tert-butyldicarbonate (3.26 g, 15 mmol)

in CH2Cl2 (30 mL) was added dropwise to a solution of ethane-1,2-diamine (10 mL,

150 mmol) in CH2Cl2 (10 mL). The resulting reaction mixture was stirred at room

temperature for 20 h. The CH2Cl2 was then removed in vacuum. The residue was

dissolved in ethyl acetate (EtOAc) (30 mL), washed with a saturated solution of

Na2CO3 (2 × 20 mL), dried over Na2SO4, and then concentrated in vacuum to give the

desired monoamine as a white oil (1.98 g , 83 %).

1H NMR (400 MHz, CDCl3) δ 4.87 (s, 1H), 3.16 (dd, J = 11.8 Hz, 5.9 Hz, 2H),

2.81-2.76 (m, 2H), 1.44 (s, 9H), 1.32-1.22 (m, 2H).


68

FTIR ν (cm-1

) 3355, 3313, 2975, 2933, 2871, 1683, 1523, 1365, 1274, 1250, 1167.

2) Synthesis of tert-butyl (2-(2,5-dioxo-2,5-dihydro-1H-pyrrol-1-yl)ethyl)carbamate

(2).


, monoamine 1 (3.2 g, 20 mmol) and Et3N (2.0 g,

20 mmol) were dissolved in Et2O (40 mL) at 0 oC. A solution of maleic anhydride

(1.96 g, 20 mmol) in Et2O (40 mL) was added dropwise, and the reaction mixture was

stirred for 4 h while being allowed to reach room temperature. The precipitated crude

intermediate triethylammonium salt was filtered off and dissolved in CH2Cl2

(100 mL). Additional Et3N (4.0 g, 40 mmol) and Benzotriazole-1-yl-oxy-tris-

(dimethylamino)-phosphonium hexafluoro-phosphate (BOP) (8.84 g, 20 mmol) were

added, and stirring was continued for 2 h. CH2Cl2 was removed in vacuo and the

resulting dark brown residue was redissolved in EtOAc (200 mL), and then washed

with HCl aq. (1 mol/L, 2 × 100 mL), saturated NaHCO3 aq. (100 mL), and brine

(100 mL), successively. After drying over Na2SO4, the mixture was filtered and

concentrated in vacuo to give a dark brown, oily residue. Purification by flash

chromatography (silica gel, hexane/EtOAc 1:1) yielded the product as a white,

crystalline solid (1.85 g, 39 %).

1H NMR (400 MHz, CDCl3) δ 6.71 (s, 2H), 4.71 (s, 1H), 3.66 (dd, J = 6.4 Hz, 4.9 Hz,

2H), 3.33 (dd, J = 10.9 Hz, 5.8 Hz, 2H), 1.41 (s, 9H).

FTIR ν (cm-1

) 3345, 3087, 2979, 2933, 1701, 1679, 1516, 1432, 1403, 1359, 1288,


69

1251, 1166, 943, 842, 690.

Melting point: 122 - 123 oC.

3) Synthesis of tert-butyl (furan-2-ylmethyl)carbamate (5).

A solution of di-tert-butyldicarbonate (1.12 g, 5.13 mmol) in CH2Cl2 (10 mL) was

added dropwise to a solution of (furan-2-yl)methanamine (0.5 g, 5.13 mmol) in

CH2Cl2 (5 mL). The resulting reaction mixture was stirred at room temperature for

20 h. The solvent was then removed and the residue was taken up in EtOAc (15 mL),

washed with Na2CO3 aq. (2 × 10 mL), dried over Na2SO4, filtered and concentrated to

give the product as oil (0.941 g, 93 %).

MS (m/z): [M + Na]+ calc. for C10H15NO3Na 220.0944, found 220.0948.

1H NMR (400 MHz, CDCl3) δ 7.33 (dd, J = 1.9 Hz, 0.8 Hz, 1H), 6.30 (dd, J = 3.2 Hz,

1.9 Hz, 1H), 6.19 (d, J = 2.9 Hz, 1H), 4.28 (d, J = 5.4 Hz, 2H), 1.44 (s, 9H).

FTIR ν (cm-1

) 3339, 2978, 2932, 2361, 1698, 1507, 1391, 1367, 1272, 1250, 1166,

737.

4) Synthesis of tert-butyl ((2-(2-((tert-butoxycarbonyl)amino)ethyl)-1,3-dioxo-

1,2,3,3a,7,7a-hexahydro-4H-4,7-epoxyisoindol-4-yl)methyl)carbamate (6).


70

A solution of compound 2 (0.2 g, 0.83 mmol) and 5 (0.164 g, 0.83 mmol) in EtOAc

(8 mL) was refluxed overnight and the solvent was then removed in vacuo. The

residue was purified by flash chromatography (silica gel, hexane/EtOAc 1:1) to give

the end product as white solid (0.3 g, 82 %). Compound 6 is a mixture of two isomers,

endo- and exo-. The endo-isomer appeared as the main product.

MS (m/z): [M + H]+ calc. for C21H32N3O7 438.2235, found 438.2232.

1H NMR (400 MHz, CDCl3) δ 6.56 (d, J = 5.7 Hz, 1H), 6.50 (d, J = 4.6 Hz, 1H), 6.43

(d, J = 5.9 Hz, 1H), 6.35 (d, J = 5.8 Hz, 1H), 5.26 (dd, J = 5.5 Hz, 1.6 Hz, 1H), 5.20

(d, J = 1.6 Hz, 1H), 5.01 (s, 1H), 4.70 (s, 1H), 4.60 (s, 1H), 4.01 (dd, J = 13.9 Hz,

6.5 Hz, 1H), 3.83 (dd, J = 15.0 Hz, 6.6 Hz, 1H), 3.78-3.61 (m, 3H), 3.56 (dt,

J = 13.5 Hz, 5.1 Hz, 1H), 3.48 (t, J = 5.7 Hz, 1H), 3.41-3.15 (m, 3H), 2.97 (d,

J = 6.4 Hz, 1H), 2.89 (d, J = 6.4 Hz, 1H), 1.52-1.38 (m, 26H).

FTIR ν (cm-1

) 3369, 2983, 2931, 1702, 1683, 1527, 1392, 1365, 1248, 1160, 960, 864.


5) Synthesis of 2-(2-aminoethyl)-4-(aminomethyl)-3a,4,7,7a-tetrahydro-1H-4,7-

epoxyisoindole-1,3(2H)-dione hydrochloride salt (7).


71

Compound 6 (0.171 g, 0.39 mmol) was dissolved in CH2Cl2 (4 mL). A saturated

solution of hydrogen chloride in Et2O (6 mL) was added. The mixture was stirred

overnight. The solvent was then removed in vacuo to yield the product as a white

solid (0.12 g, 99 %). Since compound 6 is a mixture, compound 7 is also a mixture of

two isomers.

MS (m/z): [M – 2HCl + H]+ calc. for C11H16N3O3 238.1186, found 238.1187.

1H NMR (400 MHz, D2O) δ 6.93 (d, J = 5.7 Hz, 1H), 6.80 (d, J = 5.5 Hz, 1H), 6.68

(d, J = 5.7 Hz, 1H), 6.62 (d, J = 5.8 Hz, 1H), 5.59 (d, J = 5.4 Hz, 1H), 5.51 (s, 1H),

4.09 (dd, J = 17.8 Hz, 12.9 Hz, 2H), 3.97 (t, J = 6.0 Hz, 2H), 3.83 (dd, J = 14.2 Hz,

8.4 Hz, 1H), 3.68 (d, J = 14.3 Hz, 1H), 3.49 (d, J = 6.4 Hz, 1H), 3.43 (d, J = 6.4 Hz,

1H), 3.36 (t, J = 5.9 Hz, 2H), 3.25 (t, J = 6.2 Hz, 1H).

13C NMR (400 MHz, D2O) δ 178.1, 176.6, 176.3, 176.1, 140.1, 137.7, 134.6, 132.8,

88.0, 87.9, 81.3, 79.6, 50.5, 49.0, 48.1, 47.8, 39.9, 38.5, 37.5, 37.1, 36.3, 35.8.

FTIR ν (cm-1

) 3353, 2925, 2854, 1685, 1519, 1412, 1298, 1261, 1143, 748, 719.


6) Synthesis of tert-butyl (8-aminooctyl)carbamate (8).


72


, a solution of di-tert-butyldicarbonate (2 g, 9.2 mmol) in

CH2Cl2 (20 mL) was added dropwise to a solution of 1,8-diaminooctane (13.2 g,

91.6 mmol) in CH2Cl2 (130 mL). The resulting reaction mixture was stirred at room

temperature overnight. CH2Cl2 was then removed in vacuo. The residue was taken up

in EtOAc (150 mL), washed with a saturated solution of Na2CO3 (2 × 150 mL), dried

over Na2SO4, and concentrated in vacuo. The residue was purified by distillation to

give the product as clear oil (1.69 g, 69 %).

1H NMR (400 MHz, CDCl3) δ 4.51 (s, 1H), 3.09 (dd, J = 13.1 Hz, 6.5 Hz, 2H), 2.66

(t, J = 7.0 Hz, 2H), 1.53-1.35 (m, 13H), 1.29 (s, 9H), 1.22 (s, 2H).

FTIR ν (cm-1

) 3353, 2975, 2920, 2850, 1681, 1520, 1363, 1275, 1248, 1167.

7) Synthesis of tert-butyl (8-(2,5-dioxo-2,5-dihydro-1H-pyrrol-1-yl)octyl)carbamate

(9).

Monoamine 8 (3.41 g, 14 mmol) in toluene (75 mL) was added dropwise to a solution

of maleic anhydride (1.37 g, 14 mmol) in toluene (25 mL), followed by adding Et3N

(2.12 g, 21 mmol). The reaction mixture was stirred for 2 h, and then zinc chloride

(1.9 g, 14 mmol) and a solution of HMDS (3.38 g, 21 mmol) in toluene (30 mL) were

added. After refluxing for 2 h, the solvent was removed in vacuo. Et2O (100 mL) was

added to the residue, and the solution was filtered and concentrated to give the crude

product as a light yellow solid (4.52 g, 100 %), which was used directly for further

reaction.


73

MS (m/z): [M + Na]+ calc. for C17H28N2O4Na 347.1941, found 347.1939.

1H NMR (400 MHz, CDCl3) δ 6.66 (s, 2H), 4.51 (s, 1H), 3.48 (t, J = 7.3 Hz, 2H), 3.07

(dd, J = 12.4 Hz, 6.1 Hz, 2H), 1.63-1.49 (m, 2H), 1.42 (s, 11H), 1.26 (s, 8H).

FTIR ν (cm-1

) 3354, 3087, 2974, 2925, 2854, 1683, 1520, 1413, 1363, 1248, 1167,

1122, 995, 839, 696.


8) Synthesis of tert-butyl (8-(4-(((tert-butoxycarbonyl)amino)methyl)-1,3-dioxo-

1,3,3a,4,7,7a-hexahydro-2H-4,7-epoxyisoindol-2-yl)octyl)carbamate (10).

A solution of compound 9 (1.0 g, 3.1 mmol) and 5 (5.4 g, 27.7 mmol) in EtOAc

(80 mL) was refluxed overnight. The solvent was then removed in vacuo. The residue

was purified by flash chromatography (silica gel, hexane/EtOAc 1:1) to give the

product as a white solid (1.59 g, 99 %). The compound 10 is a mixture of two

isomers, endo- and exo-, with the endo-isomer being the main product.


1H NMR (400 MHz, CDCl3) δ 6.54 (d, J = 5.7 Hz, 1H), 6.49 (d, J = 5.5 Hz, 1H), 6.40

(d, J = 5.7 Hz, 1H), 6.32 (d, J = 5.8 Hz, 1H), 5.25 (dd, J = 5.4 Hz, 1.6 Hz, 1H), 5.20

(d, J = 1.6 Hz, 1H), 5.04 (s, 1H), 4.50 (s, 1H), 3.99 (dd, J = 13.4 Hz, 7.4 Hz, 1H), 3.80

(dd, J = 14.8 Hz, 6.7 Hz, 1H), 3.63 (ddd, J = 13.2 Hz, 11.2 Hz, 5.9 Hz, 1H), 3.45 (t,


74

J = 7.3 Hz, 2H), 3.34-3.24 (m, 1H), 3.20 (d, J = 7.7 Hz, 1H), 3.08 (dd, J = 12.8 Hz,

6.3 Hz, 2H), 2.94 (d, J = 6.5 Hz, 1H), 2.86 (d, J = 6.4 Hz, 1H), 1.61-1.36 (m, 29H),

1.32-1.14 (m, 10H).

FTIR ν (cm-1

) 3379, 2979, 2931, 2856, 1700, 1683, 1521, 1398, 1365, 1248, 1161,

960, 862.


9) Synthesis of 4-(aminomethyl)-2-(8-aminooctyl)-3a,4,7,7a-tetrahydro-1H-4,7-

epoxyisoindole-1,3(2H)-dione hydrochloride salt (11).

Compound 10 (0.1 g, 0.19 mmol) was dissolved in CH2Cl2 (4 mL). A saturated

solution of hydrogen chloride in Et2O (6 mL) was added. The mixture was stirred

overnight. The solvent was then removed in vacuo to give the product as a white solid

(0.056 g, 91 %), which is a mixture of endo- and exo-isomers.

MS (m/z): [M – 2HCl + H]+ calc. for C17H28N3O3 322.2125, found 322.2125.

1H NMR (400 MHz, D2O) δ 6.90-6.76 (m, 1H), 6.69 (d, J = 7.3 Hz, 1H), 6.61 (dd, J =

5.7 Hz, 3.7 Hz, 1H), 6.52 (d, J = 5.8 Hz, 1H), 5.50 (dd, J = 5.5 Hz, 1.7 Hz, 1H), 5.37

(dd, J = 7.2 Hz, 1.8 Hz, 1H), 4.01 (t, J = 9.1 Hz, 1H), 3.97-3.90 (m, 1H), 3.76 (d,

J = 14.3 Hz, 1H), 3.65 (d, J = 7.5 Hz, 1H), 3.58-3.44 (m, 3H), 3.42-3.29 (m, 1H), 3.27

(d, J = 6.4 Hz, 1H), 3.02 (t, J = 7.6 Hz, 2H), 2.65-2.41 (m, 1H), 2.34-2.19 (m, 1H),

2.17-2.06 (m, 1H), 1.77-1.62 (m, 3H), 1.62-1.50 (m, 2H), 1.49-1.09 (m, 10H).


75

FTIR ν (cm-1

) 3388, 2916, 2848, 1693, 1599, 1508, 1398, 1342, 1176, 1143, 864, 829.


4.5.3 Synthesis of diamine cross-linker via bis-maleimide

intermediate

1) Synthesis of tert-butyl (furan-2-ylmethyl)carbamate (5)

Furfurylamine (11.7 g, 0.12 mol) in CH2Cl2 (100 mL) was added under stirring to

di-tert-butyl dicarbonate (28.9 g, 0.132 mol) in CH2Cl2 (100 mL) at 0 oC. The solution

was stirred at room temperature for 24 h. After the reaction, CH2Cl2 was removed in

vacuo and EtOAc was added to the residue. The solution was washed with saturated

Na2CO3 solution (2 × 30 mL), dried over anhydrous Na2SO4, filtered, and

concentrated to yield compound 5 (26.2 g, 99 %) as yellow crystals.

MS (m/z): [M + Na]+ calc. for C10H15NO3Na 220.0944, found 220.0948.

1H NMR (400 MHz, CDCl3) δ 7.34 (d, J = 1.9 Hz, 1H), 6.30 (dd, J = 3.2 Hz, 1.9 Hz,

1H), 6.20 (d, J = 2.9 Hz, 1H), 4.83 (s, 1H), 4.29 (d, J = 5.4 Hz, 2H), 1.45 (s, 9H).

13C NMR (100 MHz, CDCl3) δ 155.72, 152.21, 142.18, 110.48, 107.05, 37.90, 28.52.

FTIR ν (cm-1

) 3338, 2978, 2932, 2362, 1698, 1507, 1392, 1367, 1273, 1250, 1166,

736.


76

2) Synthesis of 1,1'-(octane-1,8-diyl)bis(1H-pyrrole-2,5-dione) (12)

1,8-Diaminooctane (4.0 g, 27.7 mmol) in CH2Cl2 (60 mL) was added under stirring to

maleic anhydride (6.0 g, 61.2 mmol) in CH2Cl2 (60 mL) at 0 oC and the mixture was

stirred overnight at room temperature. CH2Cl2 was then removed in vacuo and the

residue was dissolved in acetone (120 mL), and acetic anhydride (11.45 g,

112.2 mmol), nick (II) acetate (0.14 g, 0.56 mmol) and Et3N (1.58 mL, 11.4 mmol)

were added to the solution under stirring. The mixture was heated at 65 oC for 2 d,

and then cooled to room temperature. Acetone was removed in vacuo. The residue

was poured over ice water (400 mL) and stirred for 30 min. The mixture was filtered

and the solid residue was washed with water (2 × 50 mL). After air-drying at room

temperature, the crude product was purified by flash chromatography (silica gel,

hexane/EtOAc 3:1) to yield compound 12 (4.36 g, 52 %) as a white powder.159


1H NMR (400 MHz, CDCl3) δ 6.67 (s, 4H), 3.49 (t, J = 7.2, 4H), 1.67-1.46 (m, 4H),

1.38-1.14 (m, 8H).

13C NMR (100 MHz, CDCl3) δ 171.00, 134.18, 134.18, 38.00, 29.06, 28.60, 26.74.

FTIR ν (cm-1

) 3109, 3090, 2941, 2912, 2363, 2339, 1700, 1423, 1374, 1131, 841,702.



77

3) Synthesis of di-tert-butyl ((2,2'-(octane-1,8-diyl)bis(1,3-dioxo-2,3,3a,4,7,7a-

hexahydro-1H-4,7-epoxyisoindole-4,2-diyl))bis(methylene))dicarbamate (13)

A solution of compound 1 (14.13 g, 71.71 mmol) and 12 (4.36 g, 14.34 mmol) in

EtOAc (130 mL) was heated at 78 oC overnight. The solution was then cooled and

concentrated in vacuo, and the residue was purified by flash chromatography

(silica gel, hexane/EtOAc 2:1) to give compound 13 (6.63 g, 81 %) as a white solid.

Compound 13 is a mixture of exo- and endo-isomers (ratio of exo/endo = 7.5:1 as

calculated from the integration of 1H-NMR data).


Exo-isomer: 1H NMR (400 MHz, CDCl3) δ 6.53 (d, J = 5.6 Hz, 2H), 6.49 (d,

J = 5.6 Hz, 2H), 5.27 (bs, 2H), 5.19 (d, J = 1.6 Hz, 2H), 3.79 (dd, J = 14.8 Hz, 6.7 Hz,

2H), 3.67 (dd, J = 14.7 Hz, 6.4 Hz, 2H), 3.45 (t, J = 7.3 Hz, 4H), 2.94 (d, J = 6.5 Hz,

2H), 2.85 (d, J = 6.4 Hz, 2H), 1.57-1.48 (m, 4H), 1.45 (s, 18H), 1.31-1.10 (m, 8H).

13C NMR (100 MHz, CDCl3) δ 176.08, 175.39, 156.17, 138.85, 137.08, 101.88,

91.07, 80.89, 79.78, 50.30, 48.25, 40.03, 39.07, 28.95, 28.52, 27.60, 26.56.

FTIR ν (cm-1

) 3366, 2978, 2933, 2857, 1766, 1693, 1526, 1435, 1399, 1291, 1247,

1164, 975, 862, 737.



78

4) Synthesis of 2,2'-(octane-1,8-diyl)bis(4-(aminomethyl)-3a,4,7,7a-tetrahydro-1H-

4,7-epoxyisoindole-1,3(2H)-dione) (14)

Concentrated HCl aqueous solution (37 %, 50 mL) was added dropwise to the

solution of compound 13 (6.63 g, 9.5 mmol) in acetone (50 mL) under stirring at 0 oC.

The mixture was stirred at room temperature for 2 h. Acetone was then removed in

vacuo and the residue was washed with CHCl3 (3 × 30 mL) to remove any unreacted

starting material. Water (100 mL) was then added to the residue and 0.5 mol/L NaOH

solution was added dropwise until pH reached 9. CHCl3 (3 × 100 mL) was used to

extract, and then the organic solvents were combined, and concentrated to give

diamine 14 (4.63 g, 98 %) as a liquid, with the ratio of exo/endo at 6.7:1, as

determined by 1H-NMR.


Exo-isomer: 1H NMR (400 MHz, CDCl3) δ 6.54 (d, J = 5.7 Hz, 2H), 6.50 (dd,

J = 5.7 Hz, 1.6 Hz, 2H), 5.20 (d, J = 1.7 Hz, 2H), 3.43 (t, J = 8.0 Hz, 4H), 3.34-3.11

(m, 4H), 2.94 (d, J = 6.4 Hz, 2H), 2.83 (d, J = 6.4 Hz, 2H), 1.55-1.39 (m, 4H),

1.24 (s, 8H).

13C NMR (100 MHz, CDCl3) δ 176.27, 175.36, 138.57, 137.08, 92.45, 91.97, 82.99,

80.80, 50.46, 48.06, 41.46, 39.74, 28.94, 27.59, 26.55.

FTIR ν (cm-1

) 3373, 2932, 2857, 1768, 1692, 1439, 1402, 1153, 731.

Chapter 5 Preparation and Characterisation of the New Self-healing Epoxy Polymers

79

Chapter 5 Preparation and Characterisation of

the New Self-healing Epoxy Polymers

5.1 Introduction

In this chapter, self-healing polymers consisting of diamine cross-linker with double

DA units and different epoxy resins were formed under different curing conditions to

optimise the preparation process. Since the DA adducts are scissioned at temperatures

greater than 100 oC, the curing temperature had to be sufficiently low to avoid this

scission, whilst still encouraging a sufficiently high degree of cure. Near-infrared

spectroscopy (NIR) and differential scanning calorimetry (DSC) were used to

investigate the curing process. The final, cured epoxy polymer was then characterised

using Fourier transform infrared spectroscopy (FTIR), DSC, dynamic mechanical

analysis (DMA) and thermogravimetric analysis (TGA) to obtain its mechanical

properties, including glass transition temperature and thermal stability.

5.2 Variation of the curing conditions for the preparation of

thermosets

5.2.1 Self-healing epoxy polymers prepared from DGEBA

DGEBA, a common and widely used epoxy monomer, was used as epoxy resin to be

cured by the synthesised diamine cross-linker with double DA units. The mixture of

diamine cross-linker and DGEBA was prepared in the following steps. Firstly, to

reduce the possibility of cure reaction before the NIR experiment, the diamine cross-

linker and DGEBA were mixed at room temperature. The mixture then needed to be


80

degassed under vacuum at room temperature. The molar ratio of the NH and epoxide

groups was 1:1, which equated to the molar ratio of diamine cross-linker and DGEBA

of 1:2. The curing reaction of diamine cross-linker and DGEBA is shown in

Scheme 5.1.

Scheme 5.1 Curing reaction of the diamine cross-linker with two DA adducts (14) and

DGEBA.

To determine suitable curing conditions, the degassed mixture was heated at different

temperatures for varying lengths of time. Since one amine group can react with two

epoxide groups, one diamine cross-linker molecule is able to react with four epoxide

groups to form a cross-linked epoxy polymer. Since the final epoxy polymer was

insoluble in any solvents at room temperature, near-infrared spectroscopy was chosen

as the method to determine the degree of cure of the epoxy polymer sample. The NIR

spectra of DGEBA and diamine cross-linker are shown in Figure 5.1. From previous


81

work152-154

, the FTIR peaks of the DGEBA spectrum can be readily assigned. The

prominent peak at 4532 cm-1

corresponds to the epoxide group and the two peaks at

4624 cm-1

and 4678 cm-1

reflect the phenyl group in the DGEBA spectrum. The two

peaks at 4952 cm-1

and 6532 cm-1

in the diamine cross-linker spectrum relate to the

amine groups. These peaks are distinct in each of the components, and thus can be

seen in the combination of the two. The concentration of phenyl groups remains

constant during the curing process and the strength of the peaks at 4624 cm-1

and

4678 cm-1

therefore remains constant.

Figure 5.1 The NIR spectra of DGEBA and diamine cross-linker.

Since the samples used in the experiments had slightly different thicknesses, internal

standards were necessary to allow the comparison of different spectra. The peaks at

4624 cm-1

and 4678 cm-1

were chosen as these internal standards because of their

unchanged magnitude during cure. The NIR spectra of the mixtures of DGEBA and


82

diamine cross-linker before and after curing are shown in Figure 5.2. The spectrum of

the uncured mixture showed a strong peak at 4532 cm-1

for the epoxide group and two

peaks at 4952 cm-1

and 6532 cm-1

for the amine group. After the sample was cured at

60 oC for 12 h, for example, the 4532 cm

-1 peak became very weak and the two peaks

at 4952 cm-1

and 6532 cm-1

were much weakened in the spectrum of the cured

mixture. At the same time, a peak at 6981 cm-1

appeared corresponding to hydroxyl

groups. These changes clearly demonstrated that the curing reaction between epoxide

and amine groups was occurring, and that hydroxyl groups were generated as a result

of the epoxy-amine reaction.

Figure 5.2 The NIR spectra of (a) an unreacted mixture of DGEBA and diamine cross-linker,

and (b) a mixture of DGEBA and diamine cross-linker, cured at 60 oC for 12 h.

Since the epoxide group peak and phenyl group peak both appeared in the range

4400 cm-1

to 4800 cm-1

, this section of the spectra was the focus of the NIR


83

experiments. During treatment such as heating at 60 oC, the epoxide group peak at

4532 cm-1

was reduced with reaction time (Figure 5.3), and this peak was used to

monitor the progress of the curing reaction. As mentioned above, the peaks at

4624 cm-1

and 4678 cm-1

remained the same during cure and either could be used as

an internal standard. The peak at 4624 cm-1

was chosen as the internal standard

because it had greater magnitude than the 4678 cm-1

peak.

Figure 5.3 The NIR spectra of the mixture of DGEBA and diamine cross-linker heated at

60 oC, for different lengths of reaction time.

After the spectra were obtained, peak-fitting software (Igor Pro 6.2) was used to

determine the height of the epoxide group peak and the internal standard (Figure 5.4),

since it was able to deconvolute the spectrum.


84

Figure 5.4 The NIR spectrum was peak-fitted to obtain the peak height in Igor Pro 6.2.

The degree of cure, , was calculated using the follow equation:

0

0

][][

][][

ISEP

ISEP

t

t

(Equation 5.1)

where [EP]0 and [EP]t are the peak height of epoxide group peak at 4532 cm-1

at time

t = 0 and t, respectively; [IS]0 and [IS]t are the peak height of internal standard peak at

4624 cm-1

at time t = 0 and t, respectively.

After calculation using Equation 5.1 from the NIR results, the degree of cure as a

function of reaction time for different cure temperatures from 60 oC to 100

oC is

shown in Figure 5.5. For the first few hours, the degree of cure at higher temperatures

was greater than that at lower temperatures, as expected. However, after 3 h, the

degrees of cure at 80 o

C and 90 oC were greater than others. After 6 h, the degree of

cure for 80 oC samples was the greatest. From 12 h to 24 h, the increasing rate of the

degree of cure at each temperature became negligible and this part of the curve

became almost flat. Finally, the degree of cure at 60 oC and 70

oC became greater than

that at 90 oC and a little lower than that at 80

oC. The highest degree of cure was


85

achieved at about 83 % after curing at 80 oC for 24 h.

Figure 5.5 Degree of cure (%) against reaction time at different temperatures, as determined

from NIR analysis.

This phenomenon was likely due to the competition between the RDA reaction and

the Michael addition reaction (Scheme 5.2). At the beginning of the curing process,

only the reaction between the epoxide units and amine groups occurred, with the

result that higher temperatures led to a higher degree of cure. After 1 h, the RDA

reaction was able to occur in the samples cured at the highest temperature (100 oC).

The DA adducts of furan and maleimide groups in the diamine cross-linker were

cleaved and the maleimide groups appeared in the system. As the concentration of

maleimide groups increased, the unreacted amine groups were able to react with the

maleimide groups via the Michael addition reaction.166

Consequently, the

concentration of amine groups was reduced, and a greater number of epoxide groups

remained unreacted in the system, leading to a reduced degree of cure. The DA and


86

RDA reactions are reversible, and the rate of the RDA reaction depends on how close

the reaction temperature is to 120 oC, whilst the Michael addition reaction can occur

above 70 oC

166. This means that the RDA reaction readily occurs in addition to the

Michael addition reaction at 100 oC. Thus, after 3 h of reaction, the degree of cure at

100 oC was lower than that at 80

oC and 90

oC. The RDA reaction and Michael

addition reaction also occurred in the sample heated at 90 oC, likewise leading to a

reduced degree of cure. Without the influence of the two reactions, the degree of cure

of samples cured at 60 oC and 70

oC exceeded that of those held at 90

oC and 100

oC,

and almost reached the degree of cure of samples reacted at 80 oC.

Scheme 5.2 Chemical schema of the DA and RDA reaction and Michael addition reactions.

The influence of the RDA reaction and the Michael addition reaction can be clearly

seen in Figure 5.6. The uncured mixture was heated at different temperatures for 12 h.

The degree of cure increased when the temperature was increased from 60 o

C to

80 oC. However, for cure temperatures above 80

oC, the degree of cure decreased. For


87

the highest temperature (130 oC), the degree of cure was 10 % less than that for

samples reacted at 60 oC. This result clearly showed that at high cure temperatures

close to 120 oC, the RDA reaction occurred, thereby producing the maleimide group

that could react by Michael addition. A greater concentration of epoxide groups

remained unreacted and a lower degree of cure of the sample resulted.

Figure 5.6 The degree of cure at different temperatures after 12 h treatment, as calculated

from NIR data.

Due to the occurrence of the Michael addition reaction, the optimal condition of cure

that is sought should meet at least two prerequisites. First, the heating temperature

should be lower than 80 oC, so that the possibility of the cleavage of DA adducts by

the RDA reaction is as little as possible. Second, the degree of cure at that condition

should be as high as possible. It can be seen in Figure 5.5 and Figure 5.6 that the

influence of the RDA reaction and the Michael addition reaction was quite low when

the cure temperature was below 80 oC. These figures also show that the highest


88

degree of cure, about 83 %, could be obtained for samples cured at 80 oC after 12 h or

24 h reaction time. However, from Figure 5.6, it appears that the difference in the

degree of cure between 60 oC and 80

oC was only 2 %. Thus the degree of cure

obtained from the condition of curing the sample at 60 oC for 12 h was acceptable.

Importantly, 60 oC is significantly lower than the RDA reaction temperature of 120

oC

and the Michael addition reaction temperature of 70 oC. Therefore, 60

oC was

considered as the most suitable cure temperature, with a preferred curing time of 12 h.

The effect of post-curing of self-healing polymers was also investigated by treating

samples at high temperatures (>100 oC) for 2 h. For these samples, the materials were

cured at 60 oC for 12 h and then post-cured at various higher temperatures for a

further 2 h. The resulting NIR spectra are shown in Figure 5.7. For post-curing

temperatures of 100 o

C and 130 oC, the shape of the two spectra was basically

unchanged, compared to those before post-curing. When the temperature of the post-

cure increased to 160 oC, the epoxide group peak at 4532 cm

-1 became smaller, which

showed that the post-curing process was partially effective. The curves of post-curing

at 190 oC and 220

oC were similar, indicating that the degree of cure was stable when

the temperature was above 190 oC. The change in the degree of cure in the post-curing

process for different temperatures, which was obtained from the calculation using

Equation 5.1 from the NIR results, can be clearly seen in Figure 5.8. For post-curing

with the temperature at 100 o

C and 130 oC, the degree of cure increased less than

0.5 %, still being about 81 %, compared to that before post-curing. When the post-

curing temperature was as high as 160 oC, the increase in the degree of cure was

obvious. It is possible that, although most of the DA units cleaved and the cross-

linked polymer was reduced to branched polymers or oligomers at 120 oC, the

mobility of the branched polymers or oligomers was significant at this temperature,

which enabled the unreacted or partly reacted epoxy monomer to post-cure with

amine, and the degree of cure increased. However, the increase in the degree of cure


89

was only 6 %. When the post-cure temperature was above 190 oC, the degree of cure

was stable at 92 %, some 10 % more than before being post-cured. However, any

cleaved DA units have the possibility of being consumed at these high temperatures

due to side reactions such as the polymerisation of maleimide groups.167, 168

Post-curing of the epoxy samples was found not to greatly improve the degree of cure

at temperatures below 160 oC. Rather, it caused increased scission of the DA units and

resulted in the degradation of the cross-linked polymer. Thus, post-curing was

considered not to be a suitable reaction condition in this research.

Figure 5.7 The NIR spectra of samples post-cured for 2 h at various temperatures, for an

epoxy sample initially reacted at 60 oC for 12 h.


90

Figure 5.8 The degree of cure of post-cured samples which were initially reacted at 60 oC for

12 h and then post-cured at various temperatures for 2 h.

Differential scanning calorimetry (DSC) was also used to analyse the curing process.

The samples of the uncured mixture of DGEBA and diamine cross-linker were heated

at different temperatures for 12 h and then scanned by DSC, with the results shown in

Figure 5.9. In the thermogram of the sample cured at 60 oC for 12 h, in addition to a

step change at about 61 oC due to the glass transition, there was a significant

endothermic peak at 120 oC related to the occurrence of the endothermic RDA

reaction of the DA adducts in the system. With an increase in cure temperature, it can

be seen that the endothermic peak of the RDA reaction became smaller and was found

not to be present for samples cured at 120 oC for 12 h. This was indicative of the fact

that additional RDA reaction occurred when the sample was cured at higher

temperatures. When the curing temperature was 120 oC, the RDA reaction was

complete, as was as the curing reaction, and no RDA exotherm was observed. From


91

the DSC results, it is therefore reasonable to deduce that a cure temperature lower

than 80 oC is necessary to avoid the occurrence of the RDA reaction. In addition, the

spectra of (a) and (b) in Figure 5.9 look similar, indicating that the degree of cure of

the mixture cured at 60 oC for 12 h was similar to that of the material cured at 80

oC

for 12 h. Taking this into account, 60 oC for 12 h is considered to be the optimal cure

condition.

Figure 5.9 DSC scans of mixtures cured for different temperatures, all for 12 h: (a) 60 oC,

(b) 80 oC, (c) 100

oC and (d) 120

oC. The heating rate of the thermograms is 5

oC/min.

The DSC results of samples cured at a particular temperature, such as 60 oC, for

different lengths of time are shown in Figure 5.10. For a mixture which has not

already been reacted before the thermal scan, the DSC curve (a) shows a large


92

exothermic peak from 50 o

C to 180 oC. Since both the curing and RDA reaction

occurred during the DSC thermal scan of that sample, this exothermic peak is likely

due to the overlap of an exothermic peak of the epoxy/amine curing reaction and the

endothermic peak of the RDA reaction.

Most reaction enthalpies (ΔH) of the epoxy/amine reaction lie between 103 kJ/mol

and 110 kJ/mol,169, 170

and the reaction enthalpy of RDA reaction in the furan-

maleimide system is 83 kJ/mol.76

The molar ratio of epoxy/amine and

furan/maleimide was 2:1 in the uncured mixture. The highest degree of cure measured

by NIR was 83 %. The RDA reaction percentage, calculated from 1H-NMR tests of

cycle of DA and RDA reactions of Boc-protected diamine cross-linker in

Section 6.2.1, was 95 %. Taking all of the above factors into account, the reaction

enthalpy of the epoxy/amine addition could be determined theoretically to be more

than twice that of the RDA reaction after the calculations, which explained the fact

that no clear peak relating to the RDA reaction can be observed in curve (a).

With an increase in reaction time at 60 oC, the endothermic peak of RDA reaction

becomes apparent in curve (c) and continues to increase in magnitude with reaction

time. After 12 h the RDA reaction peak became constant, which can be explained as

follows. During curing at 60 oC, reaction between the epoxide and amine groups

occurred. After longer curing times, the number of unreacted epoxide groups in the

system during the DSC test was reduced, leading to a significant increase in the

strength of the RDA reaction, in comparison to the curing reaction. In addition, the

magnitude of the RDA reaction peak did not change much at longer cure times, since

the curing reaction had largely ceased. The glass transition temperature (Tg) could also

be observed in the thermograms, and rose from 42 o

C to 61 oC as the curing time

increased, eventually becoming constant after 12 h. It appears that for curing at 60 oC,

the reaction process is largely complete after 12 h.


93

Figure 5.10 DSC thermal scans of samples reacted at 60 oC for (a) 0 h, (b) 1 h, (c) 3 h, (d) 6

h, (e) 9 h, (f) 12 h, (g) 24 h. The heating rate of the thermal scan is 5 oC/min.

The DSC results above were in accordance with the results from NIR, and they again

confirmed that 60 oC for 12 h was a suitable curing condition for this epoxy-amine

system. Thus, with this optimal curing condition, the following standard preparation

process for the new self-healing epoxy polymer based on DGEBA can be summarised

as follows:


94

Diamine cross-linker with double DA units (14) (1.21 g, 2.43 mmol) and DGEBA

(1.8 g, 4.86 mmol) were directly mixed and degassed in vacuo (0.5 mm Hg) at room

temperature for 3 h. The mixture was then cured at 60 oC for 12 h to yield the cross-

linked epoxy polymer.

5.2.2 Self-healing epoxy polymers prepared from TGAP

Scheme 5.3 The curing reaction of the diamine cross-linker with two DA adducts (14) and

TGAP.


95

In order to achieve more epoxy polymer with high cross-linked density, TGAP with

high functionality (tri-functional, compared to DGEBA being bi-functional) was

cross-linked using the same diamine cross-linker with the double DA units. The

curing process was also investigated using the same method as for DGEBA and the

cross-linker. TGAP was blended with the diamine cross-linker at room temperature to

avoid the occurrence of any curing before testing. A 1:1 molar ratio of epoxide groups

and NH groups equated to a 4:3 molar ratio of TGAP and the cross-linker. The

chemical reaction of the curing process is shown in Scheme 5.3.

In order to determine a suitable and optimal curing condition, the method used for

studying the DGEBA-based cross-linked epoxy polymer was also employed for the

TGAP-based system. Since more epoxide groups were in the structure of TGAP, the

epoxy polymer produced from TGAP had a higher crosslink density than that formed

from DGEBA. The NIR spectra of TGAP and diamine cross-linker are shown in

Figure 5.11. Based on the literature171

, the strong peak at 4524 cm-1

reflects the

epoxide group and the two peaks at 4628 cm-1

and 4682 cm-1

are due to the phenyl

group in the TGAP spectrum. Compared to the peaks in the DGEBA spectrum, the

epoxide peak is more dominant, although the phenyl group peak is weaker, since there

are more epoxide groups and fewer phenyl groups (per unit mass) in the structure of

TGAP (compared to DGEBA). The two peaks at 4952 cm-1

and 6532 cm-1

in the

diamine cross-linker spectrum reflect the amine groups. No peak appeared at

4524 cm-1

in the diamine spectrum. Therefore, in the spectrum of the mixture of

TGAP and diamine cross-linker, that peak relates only to the epoxide group. The

phenyl group peaks at 4628 cm-1

and 4682 cm-1

do not change during the curing

reaction between TGAP and diamine cross-linker.


96

Figure 5.11 The NIR spectra of the TGAP and diamine cross-linker.

The NIR spectra of the mixture of TGAP and diamine cross-linker before and after

curing are shown in Figure 5.12. The mixture was heated at 60 oC for 12 h. The figure

shows that the epoxide group peak at 4524 cm-1

is reduced and the two amine peaks at

4952 cm-1

and 6532 cm-1

are much weaker than those in the spectrum of the uncured

mixture. The appearance of the hydroxyl group peak at 6981 cm-1

, as also observed in

the experiment with DGEBA, further demonstrates the occurrence of the TGAP

curing reaction. As before, an internal standard peak is required for the analysis of the

DGEBA-based epoxy polymer. Although the peaks related to the phenyl groups at

4628 cm-1

and 4682 cm-1

are comparatively smaller than in the case of DGEBA

because of the higher concentration of the epoxide groups in the TGAP monomer,

they can still be used as the internal standard. The larger peak at 4628 cm-1

was

chosen as the internal standard.


97

Figure 5.12 The NIR spectra of the mixture of TGAP monomer and diamine cross-linker

before and after curing at 60 oC for 12 h.

Figure 5.13 The NIR spectra of the mixture of TGAP and diamine cross-linker, after being

heated at 60 oC for different lengths of time.


98

The spectral range from 4400 cm-1

to 4800 cm-1

was used to observe the change of the

epoxide group peak during curing in the NIR testing experiments. In Figure 5.13, the

epoxide group peak at 4524 cm-1

became weaker after longer curing times at 60 oC.

Based on the NIR data, the degree of cure was calculated using Equation 5.1. The

peak heights of the epoxide group peak at 4524 cm-1

and the internal standard peak at

4628 cm-1

employed in the calculation were obtained from curve fitting using Igor

Pro 6.2, as discussed previously.

The uncured mixture of TGAP and diamine cross-linker was cured at different

temperatures from 60 oC to 100

oC, as a function of curing time. In Figure 5.14, it can

be seen that in the early part of the reaction, the degree of cure at higher temperatures

increased faster than that at lower temperatures. With the increasing curing time, the

rate of cure at higher temperatures significantly reduced compared to that at the

beginning of the cure process. The degree of cure at lower temperatures was similar to

that at higher temperatures after 12 h, and became greater for longer times. The

highest degree of cure was about 85 % after curing at 80 oC for 24 h.

This trend was quite similar to that observed for the DGEBA-based epoxy, and can

also be explained by the competition between the RDA reaction and the Michael

addition reaction. At the beginning of the cure process, the reaction between epoxide

groups and amine groups dominates for all cure temperatures. After about 1 h of cure,

the RDA reaction starts when the temperature is above 80 oC, and the DA adducts of

furan and maleimide groups begin to cleave and maleimide moieties appear. With the

increase in curing time, the concentration of maleimide groups increases, resulting in

the possibility of Michael addition between the maleimide group and the unreacted

amine groups. This causes the concentration of amine groups to decrease and

unreacted epoxide groups remain, and the cure rate decreases as a result. For cure

temperatures above 80 oC, the final degree of cure was lower.


99

Figure 5.14 The degree of cure of the uncured mixture of TGAP and diamine cross-linker

cured at different temperatures for different lengths of time.

Figure 5.15 The degree of reaction at different temperatures after 24 h cure temperatures

determined from FTIR data.


100

Figure 5.15 shows the result of competition between the RDA reaction and the

Michael addition reaction. The uncured sample was heated at different temperatures

for 24 h. With the variation in curing temperature from 60 o

C to 80 oC, the degree of

cure increased. When the cure temperature was greater than 80 oC, the ultimate

(long time) degree of conversion is reduced. The higher the cure temperature, the

lower degree of cure achieved. This result was similar to that of DGEBA-based epoxy

polymer and clearly demonstrated that the RDA and Michael addition reactions

influence the final degree of cure for high cure temperatures.

Due to the possible occurrence of the Michael addition reaction, the choice of the

curing temperature was crucial. It should not be too high to allow the RDA reaction

and Michael addition reaction to both occur and damage the self-healing system, nor

be too low to make the time of cure of TGAP and diamine cross-linker too long, or to

achieve a too low degree of cure. Figure 5.14 and Figure 5.15 indicate that, the

influence of the RDA reaction and the Michael addition reaction was modest when the

temperature of treatment was less than 80 oC. The highest degree of cure, about 85 %,

was obtained by the curing condition of 80 oC for 24 h. However, the degree of cure

obtained from treatment at 60 oC for 24 h was only 3 % lower than the highest cure

achieved shown in Figure 5.15. Therefore, the degree of cure for samples cured at the

lower 60 oC temperature for 24 h was acceptable. The cure temperature of 60

oC is

both sufficiently far away from the RDA reaction temperature of 120 oC and is lower

than the problematic Michael addition reaction temperature of 70 oC. As mentioned in

the discussion of the post-curing for DGEBA-based epoxy polymers, the post-curing

process in the tri-functional system was not able to achieve a much higher degree of

cure, but rather encouraged further side reactions.


101

Figure 5.16 DSC thermograms of samples cured at 60 oC for (a) 0 h, (b) 1 h, (c) 3 h, (d) 6 h,

(e) 9 h, (f) 12 h, and (g) 24 h. The heating scan rate is 5 oC/min.

As with the DGEBA system, DSC was used to investigate the curing process of

TGAP and diamine cross-linker. Figure 5.16 shows the results of uncured mixtures

treated at 60 oC for different lengths of time. Curve (a) of a temperature scan of the

uncured mixture shows a huge exothermic peak at a temperature range from 50 oC to

170 oC, which was a result of the influence of the epoxy curing reaction and RDA

reaction, both occurring in the same temperature range. As mentioned before, since


102

the reaction enthalpy of the epoxy/amine curing reaction is more than twice that of the

RDA reaction, the RDA reaction peak is not visible in this thermogram. With a greater

length of time of curing at 60 oC, the endothermic peak of the RDA reaction appears

after 3 h in Curve (c), becoming more obvious as the curing reaction process

diminishes in comparison. The Tg is first visible in Curve (d) and increases from 51 oC

to 72 oC for extended curing times. The DSC results show that no more cure at 60

oC

is visible after 24 h.

From the results of NIR and DSC experiments, the suitable condition for curing of the

epoxy polymer based on TGAP was 60 oC for 24 h. The preparation process of the

new self-healing epoxy polymer consisting of TGAP used can therefore be

summarised as follows:

Diamine cross-linker (14) (1.5 g, 3 mmol) and TGAP (1.11 g, 4 mmol) were mixed

without the use of any solvents and degassed in vacuo (0.5 mm Hg) at room

temperature for 3 h. The uncured mixture was then heated at 60 oC for 24 h to produce

the cross-linked, epoxy polymer.

5.3 Characterisation of self-healing polymers

5.3.1 Preparation of a control sample of epoxy polymer without DA

unit

In order to compare the difference between the epoxy polymer with and without a DA

unit, a control sample of epoxy polymer was prepared. This was a combination of

DGEBA and a widely-used diamine curing agent, Ethacure 100. This control sample

was prepared using the following procedure, and was used to compare the self-healing

systems in terms of thermal stability (measured using TGA) and self-healing on

scratches.


103

Ethacure 100 (0.048 g, 0.27 mmol) and DGEBA (0.2 g, 0.54 mmol) were mixed

directly and degassed in vacuo (0.5 mm Hg) for 3 h. The mixture was then cured at

60 oC for 10 h before being heated at 140

oC for 2 h to further post-cure the sample.

FTIR was used to characterise this control sample, and the data are as follows:

FTIR ν (cm-1

) 3375 (OH), 2962 (CH3), 2928 (CH2), 2870 (CH2), 1607 (phenyl),

1507 (phenyl), 1459 (phenyl), 1231 (OH), 1180 (OH), 1033 (R-O-R), 826 (phenyl).

Peak-fitting software (Igor Pro 6.2) was used to analyse the NIR data and the degree

of cure was then calculated, using the height change of the epoxide group peak at

4532 cm-1

as a function of cure. The phenyl group peak at 4624 cm-1

was used as the

internal standard. The final degree of cure was 83 %, which was similar to that of

DGEBA-based self-healing epoxy polymers.

5.3.2 The DGEBA-based self-healing epoxy polymer

The DGEBA-based self-healing epoxy polymer was prepared using the optimal

curing condition, 60 oC for 12 h. FTIR was used for the characterisation of this cross-

linked polymer. The FTIR spectrum of the DGEBA-based epoxy polymer is shown in

Figure 5.17 and the data are as follows:

FTIR ν (cm-1

) 3431 (OH), 2927 (CH2), 2858 (CH2), 1691 (C=O), 1604 (phenyl),

1506 (phenyl), 1400 (furyl), 1238 (OH), 1180 (OH), 1032 (R-O-R), 827 (phenyl).


104

Figure 5.17 The FTIR spectrum of the DGEBA-based self-healing epoxy polymer.

DSC was used to determine the Tg of the DGEBA-based self-healing epoxy polymer

cured at 60 oC for 12 h, and the result is shown in Figure 5.18. The Tg appears at about

61 oC, which is not high because of the flexible long carbon chain in the molecular

structure of that polymer and the degree of cure (81 %). A large peak occurs from

85 oC to 150

oC, with its peak at 117

oC, which represents the endothermic RDA

reaction of DA adducts and clearly demonstrates that many DA units remain

uncleaved by the curing process. Additional post-curing is observed in the DSC

testing when the temperature is above 200 oC, due to the 19 % uncured or partly cured

epoxy monomer in the system.


105

Figure 5.18 DSC result of the DGEBA-based epoxy polymer cured at 60 oC for 12 h, with

the heating rate of 5 oC/min.

DMA was also employed to analyse the DGEBA-based epoxy polymer. In this

experiment, a probe tip was placed on the top of a rectangular sample with a static

force and a dynamic force applied. The probe position was the distance between the

holder of the sample and the probe touching the top of the sample, which meant that it

actually was the height of the sample. In Figure 5.19, a height change of the cured

sample as a function of temperature can be clearly observed. As the temperature

increased, the height initially changed little at the first stage due the sample being

cross-linked. As the temperature reached the Tg range of the sample, the height

reduced slightly as the material became more flexible and could deform under stress

from the probe. However, it was still able to broadly maintain the same height. With

the temperature close to 100 oC, the sample height started to decrease dramatically,

with the sample dimensions completely destroyed at 115 oC, the height being only


106

some 0.1 mm. Clearly, at this temperature the DA units were cleaved and the cross-

linked structure was destroyed, allowing the sample to flow. This is therefore a very

useful, straightforward method to observe the effects of chain scission due to the RDA

reaction.

Figure 5.19 Probe height as a function of temperature measured by DMA of the cured

DGEBA-based self-healing epoxy polymer, with a heating rate of 2 oC/min.

In addition, the Tg was tested by DMA experiment. With the increased temperature,

the resultant sinusoidal strain changed at the glass transition range of the sample

under applied sinusoidal stress, since the sample became soft. The maximum of tan

from the ratio of storage modulus and loss modulus is related to the Tg of the sample.

In Figure 5.20, the Tg of the epoxy polymer sample is 81.7 oC. This value is 20

oC

higher than that from the DSC result, as expected, because of the higher frequency,

1 Hz, used in the dynamic mechanical technique, compared to the much lower

frequency due to heat flow occurring in the calorimeter.


107

Figure 5.20 The Tg of the cured DGEBA-based epoxy polymer obtained from DMA testing.

The thermal stability of the DGEBA-based epoxy polymer was investigated by TGA.

From Figure 5.21, it can be observed that some 6 % of the original weight of the cross-

linked epoxy polymer is lost below 300 oC. Since the degree of cure of the sample

was 81 %, there was still 19 % of unreacted or partly reacted epoxy monomer inside

the system. When the temperature of the sample in the TGA rose above 120 oC during

the TGA testing, the RDA reaction of the DA unit occurred and crosslink was

effectively cleaved to change the epoxy polymer to a branched polymer or oligomer.

Some proportion of epoxy monomer was free and likely thermally released to

contribute to the 6 % loss, whilst the other remaining material remained and continued

to cure.

The control sample of cross-linked epoxy polymer cured with Ethacure 100 diamine,

an aromatic amine without DA adducts, was also tested using TGA for comparison of


108

its thermal stability (the degree of cure of that sample was 83 % conversion). The

weight loss of the control sample was only 2 % for temperatures below 300 oC,

because the crosslink component in this system cannot break and most of the uncured

or partly cured epoxy monomer cannot be released. Compared with the result of the

control sample without DA units, the stability of DGEBA-based epoxy polymer with

DA units decreased a little in the temperature range from 200 oC to 300

oC. However,

the cross-linker was stable throughout the temperature range of the healing process

under 150 oC, which will be discussed in Chapter 6. The weight of the self-healing

epoxy polymer reduced to 60 % at about 370 oC. This weight loss was studied by

El Gouri and co-workers, who suggested that this was caused by the release of some

volatile organic segment compounds.172

The segment could be the 1,3-di-phenoxy

isopropanol chain extender from the epoxy chain scission at temperature above

320 oC, as discovered by Dyakonov and his colleagues.

173 The weight of the control

sample was lost at temperatures above 350 oC for the same reason.

Figure 5.21 TGA results of the cured DGEBA-based cross-linked epoxy polymer with DA

units, and the control sample of epoxy polymer without DA units.


109

5.3.3 The TGAP-based self-healing epoxy polymer

The TGAP-based self-healing epoxy polymer was produced using the optimal curing

condition, 60 oC for 24 h. FTIR was employed to analyse the polymer structure. The

FTIR spectrum of the TGAP-based epoxy polymer is shown in Figure 5.22 and the

data are as follows:

FTIR ν (cm-1

) 3436 (OH), 2927 (CH2), 2856 (CH2), 1687 (C=O), 1510 (phenyl),

1402 (furyl), 1228 (OH), 1035 (R-O-R), 816 (phenyl), 748 (phenyl).

Figure 5.22 The FTIR spectrum of the TGAP-based self-healing epoxy polymer.

The Tg of the cured TGAP-based self-healing epoxy polymer, as obtained using DSC,

was about 72 oC (Figure 5.23), some 10

oC higher than that of the DGEBA-based

epoxy polymer, because the length of the TGAP molecule was shorter and more

multifunctional than that of DGEBA, and thus the cross-linked structure was tighter


110

and the crosslink density of the TGAP-based epoxy higher. The large peak from 85 oC

to 130 oC, with the peak at 116

oC, possessed the same features as the curve of the

DGEBA-based sample, and reflected the RDA reaction in the epoxy system which

occurred during the DSC thermal scan. It appeared that a small percentage of RDA

reaction occurred in the curing process of the TGAP-based epoxy polymer. From

175 oC, post-curing commenced and yielded an exothermic peak, because the

cleavage of the DA unit by RDA reaction made the 18 % uncured or partly cured

epoxy monomer free to flow and react with the remaining amine.

Figure 5.23 DSC thermal scan of the TGAP-based epoxy polymer cured at 60 oC for 24 h,

with the DSC heating rate being 5 oC/min.

DMA was also used to obtain the Tg of the TGAP-based epoxy polymer, with the peak

at 92.4 oC of the tan in Figure 5.24 representing the Tg of the epoxy polymer sample.

As before, the Tg was some 20 oC higher than that obtained from the DSC scan.


111

Figure 5.24 Tg of the cured TGAP-based epoxy polymer obtained from DMA testing.

TGA was also used to analyse the thermal stability of the TGAP-based self-healing

epoxy polymer, with the result shown in Figure 5.25. The weight loss of the cross-

linked epoxy polymer largely commenced at 130 oC, reducing to 93 % by 270

oC. The

RDA reaction occurred at 130 oC to break the DA units in the crosslink component of

the epoxy polymer, which resulted in the rapid decrease in crosslink density of the

polymer. Like the DGEBA-based epoxy polymer, since the degree of cure of the

TGAP-based polymer sample was 82 %, 18 % of unreacted or partly reacted epoxy

monomer remained in the system. Due to the scissioning of the epoxy polymer

structure by the RDA reaction, this material was free to move. Some proportion

reacted with the free amine remaining in the system and continued to cure, while

some proportion were thermally released, which caused 7 % of the weight lost. For

temperatures above 270 oC, the weight reduced sharply, which meant that the main


112

chain of the TGAP-based epoxy polymer became unstable and started to decompose

at that temperature. Compared with the TGA trace of DGEBA-based epoxy polymer,

the scissioning temperature of the TGAP-based polymer (270 oC) was lower than that

of the DGEBA-based epoxy polymer (300 oC). However, due to its higher crosslink

density, the breakdown rate of the TGAP-based polymer was lower than that of the

DGEBA-based epoxy polymer, as can be seen in Figure 5.25. This result is consistent

with the research of Petal.174

Figure 5.25 TGA results of the cured TGAP-based cross-linked epoxy polymer and of the

cured DGEBA-based cross-linked epoxy polymer.

5.4 Summary

In this chapter, the preparation and characterisation of the cross-linked, self-healing

epoxy polymers from epoxy monomers and the synthesised diamine cross-linker with


113

double DA units have been described and discussed. Two kinds of epoxy monomers

with different levels of functionalities, DGEBA and TGAP, were used to prepare

epoxy polymers with different crosslink densities, all using a stoichiometric ratio of

NH groups and epoxide groups of 1:1.

The optimal curing conditions were determined by analysing the mixture of epoxy

monomer and diamine cross-linker. The data from NIR were used to calculate the

degree of cure of each sample cured under different conditions. It was found that

curing temperatures above 80 oC were not desirable, because of the occurrence of the

Michael addition reaction together with the RDA reaction. After factors relating to

temperature and degree of cure were taken into account, the most suitable curing

conditions for each were determined: (i) 60 oC for 12 h for the DGEBA-based epoxy

polymer, with the final degree of cure being 81 %, and (ii) 60 oC for 24 h for the

TGAP-based epoxy polymer, with the resulting degree of cure being 82 %.

DSC temperature scanning was used to observe the different processes for the

samples cured for different lengths of time, at different temperatures. The DSC

thermograms clearly showed that the peak corresponding to the RDA reaction

disappeared after the sample was cured above 80 oC for 12 h, which showed that the

RDA reaction occurred during the curing process under these conditions. The results

of the DSC scans correlated with results from NIR and confirmed that the optimal

conditions previously determined were appropriate for the curing process.

The self-healing cross-linked epoxy polymers were characterised by FTIR, DSC,

DMA and TGA. Using standard conditions, the Tg of the polymers were found (using

DSC) to be 61 oC for DGEBA-based polymer and 72

oC for TGAP-based polymer.

DMA was also employed to determine the Tg, and temperatures some 20 oC higher

than that from DSC were observed, due to the higher frequency of the DMA

technique. The thermal stability of the samples was investigated by TGA. Given that


114

the RDA reaction occurs and the DA units become cleaved, a greater percentage of

weight loss occurs in the self-healing epoxy polymer at temperatures less than 300 oC,

compared to that of the control sample without DA units. However, importantly, the

self-healing polymer was thermally stable in the temperature range of the healing

process under 150 oC, which provided the conditions for more detailed studies into

self-healing presented in the next chapter.

Chapter 6 Self-healing Properties of the New Epoxy Polymers

115

Chapter 6 Self-healing Properties of the New

Epoxy Polymers

6.1 Introduction

In this chapter, self-healing properties based on the reversible DA reaction are

investigated. DA and RDA reactions in the diamine cross-linker and in the polymer

system are demonstrated by experiments using nuclear magnetic resonance (NMR)

and Fourier transform infrared spectroscopy (FTIR), respectively. The conditions of

reversible DA reactions are used to explore the thermal mechanism of the self-healing

behaviour by different methods, including dynamic mechanical analysis (DMA),

swelling and shape change. The healing process on the surface of the sample at

different temperatures is also demonstrated using optical microscopy. In addition, a

method for the determination of the flow activation energy of the healing species is

proposed and used to calculate the value for this system.

6.2 Study on DA and RDA reaction conditions

6.2.1 DA and RDA reactions in the diamine cross-linker

Two DA adducts of furan and maleimide group as healable units were placed into the

structure of diamine cross-linker (14) to achieve self-healing properties of the final

cross-linked epoxy polymer. Since the chemical basis of the self-healing process of

the polymer is the DA and RDA reactions in the diamine cross-linker, it is first

necessary to confirm if the thermal cycling of DA and RDA reactions occur when this

moiety is placed in the cross-linker. Since the final epoxy polymer is cross-linked and

cannot be dissolved in any solvents at room temperature, use could not be made of


116

available characterization methods such as solution NMR. Therefore, in order to

determine whether the DA and RDA reactions could occur in the diamine cross-linker

itself, NMR tests were undertaken on the cross-linker dissolved in a solvent.

According to the literature,75, 76

the temperatures of the DA and RDA reactions are

75 oC and 120

oC, respectively. Therefore, these temperatures were used in the NMR

test of the cycle of DA and RDA reactions in this phase of the study. Since the carbon-

carbon double bond in the maleimide group, which appears after RDA reaction, is

readily able to react with the amine group by the Michael addition reaction and thus

inhibit the subsequent DA reaction when the temperature is above 70 oC, the Boc-

protected diamine cross-linker (13) was used instead of the free diamine cross-linker

(14) in the NMR experiments (Scheme 6.1). DMSO-d6 was chosen as the solvent for

the test, due to its high boiling point of 189 oC and thus its low volatility at the DA

and RDA reaction temperatures.

Scheme 6.1 DA and RDA reaction scheme for the Boc-protected diamine cross-linker (13)

The 1H-NMR test of compound 13 was first run at room temperature, and the sample

was then heated at 120 oC for 30 min, quenched by quickly cooling to room

temperature and remeasured. The sample was subsequently treated at 75 oC for 20 h,


117

cooled to room temperature and reanalysed. The 1H-NMR spectra are shown in

Figure 6.1. Peaks of 1, 2 and 3 relate to the DA adduct and the peaks of 1’, 2’ and 3’

are ascribed to the furan group in the RDA product, compound 5. In the NMR

spectrum of compound 13, only peaks of 1, 2 and 3 were observed (Figure 6.1a). After

the sample was heated at 120 oC for 30 min, peaks of 1, 2 and 3 disappeared and

peaks of 1’, 2’ and 3’ appeared (Figure 6.1b). This demonstrated that compound 5 was

generated due to the occurrence of the RDA reaction. The sample was then heated at

75 oC for 20 h and the peaks of 1, 2 and 3 once again increased in strength in the

NMR spectrum (Figure 6.1c). This proved that the DA reaction had occurred and that

the DA units of the furan and maleimide groups reformed.

Figure 6.1 1H-NMR spectra of compound 13 (a) before heating; (b) heated at 120

oC for

30 min and quenched quickly to room temperature; (c) heated at 120 oC for 30 min, then at

75 oC for 20 h and cooled to room temperature.


118

The rate of the reaction was, nonetheless, low. This may be because the tests were

carried out in an NMR tube and the solution could not be stirred during the treatment

of the sample, which may reduce the rate of the DA reaction. In addition, the low

concentration of compound 13 in the DMSO-d6 may also contributes to the slow

reaction rate, and it took some 20 h to obtain a better NMR spectrum showing the DA

reaction (Figure 6.1c). It would be expected that the DA reaction would occur at a

more rapid rate inside the epoxy polymer because the furan and maleimide groups are

constrained to be closer together, rather than in dilute solution.

6.2.2 The conditions of the DA and RDA reactions in the cross-linked

epoxy polymers

Since the cross-linked epoxy polymers are insoluble in any solvent, the NMR

technique could not be used to study the DA and RDA reactions in the system. FTIR

was found to be the best choice to investigate the DA and RDA reactions in such

epoxy polymers.

For the DGEBA-based epoxy polymer, 140 oC was selected as the temperature for

cleaving the DA units in the cross-link unit because it was above the RDA reaction

temperature of 120 oC, but not sufficiently high to cause side reactions. A DA reaction

temperature of 75 oC was used to reform the DA adducts in the system. The FTIR

spectra of the uncured and cured DGEBA-based epoxy polymer are shown in

Figure 6.2. The peak at 694 cm-1

is the peak of C-H bond which is part of the C=C in

the maleimide ring (Scheme 6.1).175

The height of this peak was used to determine the

number of the free maleimide groups in the polymer, and the change in peak height

allowed the measurement of the occurrence of DA and RDA reactions. The peak at

1691 cm-1

for the maleimide carbonyl group was constant during the DA and RDA

reactions, allowing it to be used as an internal standard for quantification.175


119

Figure 6.2 FTIR spectra of (a) the sample of cured DGEBA-based epoxy polymer; (b) the

sample heated at 140 oC for 30 min and quenched quickly to room temperature; (c) after that,

the sample treated at 75 oC for 5 h and cooled to room temperature.

In the FTIR spectra of the cured sample (Figure 6.2a), the peak at 694 cm-1

was weak

because most of the maleimide groups had joined with the furan groups to form DA

units. The DGEBA-based epoxy polymer sample was then heated at 140 oC for

30 min and quickly quenched to room temperature. The height of the peak at 694 cm-1

subsequently increased in magnitude by a factor of three (Figure 6.2b) compared to

that in the spectrum of the original sample (Figure 6.2a), because the occurrence of the

RDA reaction caused the DA units to cleave and caused an increase in the number of

maleimide groups. The sample was then heated at 75 oC for 5 h and cooled to room

temperature. The peak at 694 cm-1

reduced in height (Figure 6.2c), becoming similar

to that in the spectrum of the original sample (Figure 6.2a). This result indicated that

the concentration of maleimide groups returned to a level similar to that in the original


120

sample, demonstrating the almost total regeneration of the DA units by the DA

reaction, as measured by FTIR analysis. This showed that with the temperature

conditions outlined above, the cycle of DA and RDA reactions in the cured DGEBA-

based epoxy polymer could be achieved and monitored using infrared spectroscopy.

The cycle was repeated three times and the results obtained were similar.

In order to find the most suitable conditions for undertaking the DA and RDA

reactions in the cross-linked polymer system in the solid state, two series of

experiments were designed and carried out. Since the height of the peak at 694 cm-1

changes during the DA and RDA reactions, it can be used to quantify the number of

maleimide groups in the system and thus monitor the extent of the DA/RDA reaction.

To compare the different spectra, an internal standard was used, which is the

maleimide carbonyl group peak at 1691 cm-1

. The value of peak height [CPH] was

thus quantified using the follow equation:

o

o

t

t

ISMM

ISMM

CPH

][][

][][

(Equation 6.1)

where [MM]o and [MM]t are the peak height of the maleimide group peak at 694 cm-1

at the original state of the cured sample and time t, respectively; [IS]o and [IS]t are the

peak height of the internal standard at the original state (cured sample) and time t,

respectively. For the RDA reaction process, a greater numerical value of [CPH] is

indicative of more maleimide in the system, which means the degree of RDA reaction

is greater. For the DA reaction process, a smaller numerical value of [CPH] represents

less maleimide content in the system, which means the degree of DA reaction is much

higher.


121

Figure 6.3 The change in infrared peak height at 694 cm-1

as a function of annealing time for

different temperatures, for the DGEBA-based epoxy polymer.

Figure 6.3 shows the height change of the peak at 694 cm-1

after calculation using

Equation 6.1 from the FTIR results of the cured sample of DGEBA-based epoxy

polymer, after heating for different time periods at 110 oC, 120

oC, 130

oC, 140

oC and

150 oC, respectively. Each sample was quickly quenched to room temperature from

the heated condition and then tested. For temperatures lower than or equal to the RDA

reaction temperature, 120 oC, the highest [CPH] was found to be 2.5. When the

temperature reached 130 oC, a value of 3.5 for [CPH] was obtained after treatment for

30 min, which showed that the rate of RDA reaction increased to produce more

maleimide groups at temperatures greater than 120 oC. With the increase of

temperature, the time required for obtaining high values of [CPH] decreased. Three

conditions, 130 oC for 30 min, 140

oC for 20 min and 150

oC for 10 min, were found,

which were all able to lead to high values of [CPH]. After being heated at high


122

temperatures for 2 h, the [CPH] dropped to 2. This indicated that side reactions of the

double bond in the maleimide, for example polymerisation of the maleimide,53, 168

were able to occur at these temperatures and thus there was a decrease in the amount

of free maleimide groups. A shorter heating time was therefore required to avoid the

occurrence of unwanted side reactions. When comparing the three conditions,

although 150 oC was the highest temperature and probably caused the greatest amount

of side reaction, the shorter time required for healing in that condition would allow

less chance for the side reaction to occur, which was more important. Therefore, the

heating condition of 150 oC for 10 min was chosen as the most suitable condition for

the RDA reaction in the solid state of the DGEBA-based cross-linked epoxy polymer.

After being heated at 150 oC for 10 min and quenched to room temperature to cause

substantial cleavage of the DA unit in the cross-link part to occur and be retained

("frozen in"), cured samples of DGEBA-based epoxy polymer were then treated for

different lengths of time at 50 oC, 60

oC and 70

oC, respectively, to find the most

suitable condition for causing the DA reaction in the system. The results after

calculation using Equation 6.1 are shown in Figure 6.4. For 50 oC, the value of [CPH]

could not be reduced to the original level after 20 h treatment, while both 60 oC and

70 oC caused the value of [CPH] to return to 1 after heat treatment for 5 h, which

corresponded to the concentration of the maleimide groups being reduced to a similar

level to that in the original cured sample. Because the lower temperature reduced the

degree of side reactions, the condition of 60 oC for 5 h was finally chosen as the most

suitable condition for the DA reaction to reform the crosslink.


123

Figure 6.4 The variation in peak height at 694 cm-1

of the FTIR spectra of the DGEBA-based

epoxy polymer heated at 150 oC for 10 min against heating time at different temperatures.

In summary, from the two series of experiments described above, the most suitable

conditions for RDA and DA reactions in the cured sample of DGEBA-based epoxy

polymer were found to be 150 oC for 10 min and 60

oC for 5 h. Using these

conditions, the cross-linked polymer was able to cleave and recover up to at least

3 times without any obvious change in the FTIR result. This condition was also used

in the exploration of the self-healing process described in Section 6.3.

For epoxy polymers generated from the tri-functional TGAP epoxy monomer and

diamine cross-linker, the conditions used above for the DA and RDA reactions were

applied to investigate the cleavage and recrosslink reactions for this more highly

cross-linked epoxy system. The FTIR results are shown in Figure 6.5. The peak at

694 cm-1

is once again due to the free maleimide ring and the peak at 1687 cm-1

for


124

the maleimide carbonyl group was again used as an internal standard. The 694 cm-1

was relatively weak in the spectrum of a cured sample (Figure 6.5a), which proved

that most of the maleimide groups were in the form of DA adducts and there was little

free maleimide in the cured sample. After the polymer sample was treated at 150 oC

for 10 min and quenched quickly to room temperature, the height of the peak

(Figure 6.5b) increased by a factor of three, compared with that of the sample before

such treatment (Figure 6.5a), indicating that more maleimide groups existed in the

sample because the DA units were cleaved by the RDA reaction. After the sample was

treated at 60 oC for 5 h and cooled to room temperature, the height of the peak

(Figure 6.5c) decreased to the level found in the original sample (Figure 6.5a), which

showed that the DA reaction occurred and the DA units reformed in the TGAP

system, as was observed in the DGEBA-based epoxy network.

Figure 6.5 FTIR spectra of (a) the sample of cured TGAP-based epoxy polymer; (b) the

sample heated at 150 oC for 10 min and quenched quickly to room temperature; (c) following

(b), the sample was treated at 60 oC for 5 h and cooled to room temperature.


125

As for the DGEBA-based epoxy polymer, the same method was used to find the

optimal conditions to induce the DA and RDA reactions in the TGAP-based epoxy

polymer. The height change of the peak at 694 cm-1

was followed during the

experiment and the maleimide carbonyl group peak at 1687 cm-1

was employed as an

internal standard to compare different spectra.

Figure 6.6 The change in infrared peak height at 694 cm-1

as a function of annealing time for

different temperatures, for the TGAP-based epoxy polymer.

The height change of the maleimide peak, after the cured sample was heated at

different temperatures for different lengths of time, is shown in Figure 6.6, with the

data calculated using Equation 6.1 from the FTIR results. The value of [CPH]

obtained at a temperature less than 120 oC was low. With the temperature held above

120 oC, the rate of RDA reaction increased, resulting in the production of more

maleimide groups and a greater value of [CPH]. For the DGEBA-based epoxy


126

polymer, three conditions, 130 oC for 30 min, 140

oC for 20 min and 150

oC for

10 min, all demonstrated the ability to bring about high values of [CPH] compared

with other conditions. Due to the occurrence of side reactions of maleimide groups,

such as polymerisation of the maleimide which can occur if the sample is held at high

temperatures for an extended time, the shortest annealing time was considered to be

the best condition to prevent side reactions from occurring. The annealing condition

of 150 oC for 10 min was selected as the most suitable condition for the RDA reaction

in the TGAP-based epoxy polymer.

To find the most suitable condition for the DA reaction, the cured sample of TGAP-

based epoxy polymer was heated at 150 oC for 10 min and quenched rapidly to room

temperature to generate a sample with substantial cleavage of the DA unit in the

cross-link unit. The sample was then heated at different temperatures for different

lengths of time. The results after calculation using Equation 6.1 are shown in

Figure 6.7. It was found that both 60 oC and 70

oC could cause the values of [CPH] to

be reduced to 1, which meant that the concentration of maleimide groups had returned

to the value found in the original cured sample, demonstrating that the sample was

fully recovered in terms of scissioning and reformation. However, due to the higher

cross-link density in the TGAP-based epoxy polymer, the recovery time (7 h) was

necessarily longer than that used for the DGEBA-based epoxy polymer. With the

lower temperature, the condition of 60 oC for 7 h was selected as the most suitable

condition for the DA reaction to reform the crosslink component of the epoxy cross-

linked network.


127

Figure 6.7 Peak height change at 694 cm-1

of the sample of TGAP-based epoxy polymer

heated at 150 oC for 10 min compared with the original sample, as a function of heating time

at different temperatures, as characterised using FTIR.

From the results of the two series of experiments described above, the conditions of

150 oC for 10 min and 60

oC for 7 h were finally confirmed for the RDA and DA

reactions in the cured sample of TGAP-based epoxy polymer. These conditions were

used in the exploration of the self-healing process described in the following section.

6.3 The thermal self-healing mechanism in the new epoxy

system

The healing process of the DA-based self-healing epoxy polymer can be divided into

three components: scission of the cross-linker by the RDA reaction; the flow of

mobile material; and the subsequent reformation of the network by the DA reaction.


128

To prove and understand the mechanism of the thermal self-healing process in the

DA-based epoxy polymer, a range of characterisation experiments were undertaken,

including DMA, swelling tests, dissolution behaviour, GPC and shape change. The

previously-determined conditions for the DA and RDA reactions in Section 6.2.2

were used as the healing conditions to treat the samples for these experiments. For the

DGEBA-based epoxy polymer, the conditions were 150 oC for 10 min and 60

oC for

5 h.

DMA tests were used to determine the Tg of the samples before, during and after the

healing process, to provide information about the state of the materials at key times,

particularly in relation to the crosslink density, where a higher Tg would be indicative

of a greater covalent linkage between chains.

The tan traces of the DGEBA-based epoxy polymer samples produced by different

thermal conditions are shown in Figure 6.8. The samples were treated under different

conditions, quenched quickly to room temperature and tested. For the original cured

sample, the Tg was 81.7 oC, as mentioned in Chapter 5. After the cured sample was

heated at 150 oC for 10 min, the Tg dropped to 72.1

oC. This indicated that some parts

of the crosslink were cleaved as a result of the RDA reaction and the cross-linked

epoxy polymer was scissioned, leading to an increase in molecular mobility.

For the following series of experiments, the cured sample was first heated at 150 oC

for 10 min to cause the RDA cleavage in the cross-linker, then cooled to room

temperature, and subsequently heated for varying annealing times at 60 oC to allow

the occurrences of DA reactions to reform the crosslink. After annealing at 60 oC for

1 h, the Tg increased to 76.2 oC due to the reformation of the crosslink. With the

increase in heating time, the Tg increased to 79.3 oC after 3 h and finally reached

83.1 oC after 5 h. This demonstrated that the concentration of the reformed crosslinks

from DA reaction increased, and the sample was finally sufficiently recovered to be in


129

the same crosslink state as the original cured sample. The final Tg was (within

experimental error) similar to the original value of 81.7 oC. This experiment

demonstrated that the cured sample could be scissioned and subsequently returned to

the original state during annealing.

Figure 6.8 The tan traces of the (a) cured DGEBA-based epoxy polymer, (b) cured sample

heated at 150 oC for 10 min, (c) cured sample heated at 150

oC for 10 min and then heated at

60 oC for 1 h, (d) cured sample heated at 150

oC for 10 min and then heated at 60

oC for 3 h,

(e) cured sample annealed at 150 oC for 10 min and then heated at 60

oC for 5 h, from DMA.

The heating rate of the DMA scan is 2 oC/min and the measuring frequency is 1 Hz.

The swelling properties of thermoplastic and thermosetting polymers are intrinsically

different, making this a potent method for better understanding the microstructure of

the polymer, before and after scission and crosslink reformation. The percentage of

weight increase of samples [WI] is determined using the equation as follows:


130

%1000

0

W

WWWI t (Equation 6.2)

where W0 and Wt are the weight of the sample before and after the swelling test,

respectively. After being treated under different conditions, the samples were placed

in acetone for 7 d. The samples were subsequently removed, blotted with tissue paper

and weighed.

Figure 6.9 Swelling test of (a) cured DGEBA-based epoxy polymer, (b) cured sample

annealed at 150 oC for 10 min, (c) cured sample annealed at 150

oC for 10 min and then

heated at 60 oC for 1 h, (d) cured sample annealed at 150

oC for 10 min and then heated at

60 oC for 3 h, (e) cured sample annealed at 150


oC for 5 h.

The results of swelling tests of samples exposed to different scission and reformation

conditions, calculated using Equation 6.2, are shown in Figure 6.9. In each condition, a

new cured sample was used. For the original cured sample, the weight increase was

26 %. For the sample treated at 150 oC for 10 min before the swelling tests, the weight


131

increase reduced to 1.9 %, probably because DA units in the epoxy polymer became

disconnected and the non-cross-linked components did not have the same ability to

hold the solvent as the cross-linked one. It was also possible that some non-cross-

linked portions were lost. The 1.9 % weight increment may result from some cross-

linked polymer left in the material due to incomplete scissioning. When the partly

degraded sample was heated at 60 oC, the swelling ability recovered rapidly and

reached the original level after 5 h. This showed that the molecular architecture was

able to recover to the same degree as the original cured sample.

However, it was possible that any fully scissioned and decoupled monomer, linear or

branched materials, that were dissoluble in the solvent, resulted in the loss of material

and influenced the swelling result. Therefore, after the swelling test, samples were

dried at 50 oC overnight in vacuo to fully evaporate the solvent contained within the

samples and their weights were measured again. The percentage of the weight

decrement of the samples is shown in Figure 6.10. The cured sample lost 10.4 % of the

original weight after swelling test, which might have a contribution from some 20 %

uncured materials. The sample heated at 150 oC for 10 min lost 31.1 % in weight after

exposure to solvent, which was three times more than that of the cured sample. It was

clear that the cleavage of DA units during the heating resulted in the scissioning of the

cross-linked sample and some components in the sample were dissolved. With the

increase in annealing time at 60 oC, the percentage of weight decrement was reduced

to the level of the original sample, which meant the reformation of the cleaved DA

units occurred as a function of heating time.


132

Figure 6.10 The decrement of the weight change of the sample described in Figure 6.9, after

being dried, compared with the weight of the original sample.

Because the weight of the original samples decreased during the tests, the dried

weight of the sample should be used to calculate the weight increment in swelling

tests. The results calculated in this manner are shown in Figure 6.11 and differences

between the samples can be clearly observed. All of the weight increments were

above 40 % due to the additional internal volume inside the samples due to the loss of

incompletely cured components in the test. The sample which was heated at 150 oC

for 10 min showed the greatest increase because the cleavage of the DA units in the

crosslink part by the RDA reaction changed the structure and reduced its crosslink

density compared with other samples. With longer treatments at 60 oC, the weight

increment reduced to a level similar to that of the original sample because the DA

reaction reformed the cross-linked structure.


133

Figure 6.11 Swelling test results of the samples described in Figure 6.9, based on the dried

sample weight.

The solubility of the cured epoxy at temperatures greater than the RDA reaction

temperature was studied to demonstrate the occurrence and consequence of the RDA

reaction. The cured, cross-linked sample of DGEBA-based epoxy polymer was placed

in DMSO solvent and the mixture was heated at 140 oC, above the 120

oC temperature

at which the RDA reaction occurred. After 3 min using a stirring speed of 1000 rpm,

the 35 mg cured sample was fully dissolved in the solvent. This phenomenon showed

that the RDA reaction cleaved the DA units in the crosslink moiety, and the cross-

linked polymer scissioned to low molecular weight, branched polymers or oligomers,

which could subsequently be dissolved in DMSO.

In order to further analyse the components of the dissolved polymer in the organic

solvent, GPC was used with DMF as the solvent. One cured sample was placed into


134

DMF solvent at 140 oC, and was completely dissolved after 2 min. Maintaining the

same dissolving time, other cured samples with similar weights were placed in DMF

at different temperatures of 130 oC, 120

oC and 110

oC, respectively. After 2 min, the

undissolved residue was filtered and dried at 100 oC overnight in vacuo to remove the

solvent. The weight was then measured and the results are shown in Figure 6.12. For

the solvent temperature of 130 oC, the sample could not be totally dissolved, with

some 3.4 % of the original weight left. When the temperature of the solvent was

reduced, the amount of the sample remaining increased. For the temperature of

110 oC, which is lower than the RDA reaction temperature of 120

oC, more than 30 %

of the original weight remained. This demonstrated that the rate of the RDA reaction

decreased with reduced heating temperature.

Figure 6.12 The residual weight of cured sample of DGEBA-based epoxy polymer after

being placed in DMF with (a) 140 oC, (b) 130

oC, (c) 120

oC and (d) 110

oC for 2 min.


135

The solution above was analysed using GPC and the results are shown in Figure 6.13.

The shapes of curves are similar, with a significant, broad peak appearing at 14.5 min

and a sharp peak at 17 min. The Mw and ratio of Mw/Mn of each peak is listed in

Table 6.1. The Mw of the sharp peak in each chromatogram is the same, some

500 g/mol with Mw/Mn of 1.0, which demonstrates that this peak represents a small

molecular fragment. However, the broader peak has a variable Mw value from

4 000 g/mol to 12 000 g/mol, increasing with the decrease of the temperature of DMF

from 140 oC to 110

oC. This indicates that the DA units were cleaved and the cross-

linked polymer was scissioned to smaller branched polymers or oligomers, not simply

to linear polymers with long chain. With the decrease in the annealing temperature,

fewer DA units were cleaved and larger oligomeric fragments resulted. This result is

further evidence that the epoxy network in the polymer becomes scissioned and small

oligomers are formed, which are the key materials flowing during the thermal self-

healing process.

Figure 6.13 GPC chromatograms (obtained at room temperature) of dissolved components

of scissioned DGEBA-based epoxy polymer in DMF obtained at different temperatures.


136

Table 6.1 The Mw and Mw/Mn of each peak in each GPC chromatogram in Figure 6.13.

Temperature (oC) Mw (g/mol) Mw/Mn

140 4 162 1.61

494 1.03

130 4 235 1.70

486 1.03

120 4 865 1.98

496 1.03

110 12 185 3.39

496 1.03

The effects of heating were also studied visually in macroscopic samples. A disk

sample of size 5 mm was heated without any applied force at 120 oC, 130

oC and

140 oC, for 30 min. The resultant shape change is shown in Figure 6.14. For the

sample heated at 120 oC, the shape appears the same as the original and does not vary.

A slight change in shape occurred for the sample heated at 130 oC but the difference is

still not obvious, whilst the object collapses completely when held at 140 oC after

30 min. It is clear that the DA units are sufficiently cleaved by the RDA reaction at

140 oC, such that the cross-linked network of the epoxy polymer is largely scissioned

and reduced to the branched small polymers or oligomers mentioned above, whose

molecular mobility and ability to flow under their own weight by gravity are apparent.

Figure 6.14 Shape change of the cured sample after being heated for 30 min at (a) room

temperature, (b) 120 oC, (c) 130

oC and (d) 140

oC.


137

From the above experiments, it is clear that the DA units in the crosslink part are

sufficiently cleaved at temperatures above 140 oC to lead to the production of

branched small polymers or oligomers which are able to flow under their own forces.

The more highly cross-linked epoxy polymer system based on the tri-functional

TGAP monomer was also studied using the series of experiments above. For the

TGAP-based epoxy polymer, however, the heating conditions were 150 oC for 10 min

and 60 oC for 7 h.

DMA was used to observe the Tg change under the heating conditions above, which

indicates the change in the crosslink density. The tan traces of the TGAP-based

epoxy polymer samples treated under different thermal conditions are shown in

Figure 6.15. The samples were heated and then quenched quickly to room temperature

and tested. The Tg of the original cured sample was 92.4 oC, which was higher than

that of the DGEBA-based polymer because of its higher crosslink density. After the

cured sample was treated at 150 oC for 10 min, the Tg decreased to 81.9

oC due to the

cleavage of some DA units in the crosslinks by the RDA reaction, the cross-linked

sample becoming more flexible and the molecular units more mobile. The scissioned

sample was then heated at 60 oC for different lengths of time. The DA reaction

reformed the DA units and the Tg was found to increase to 84.6 o

C after 1 h. As the

time of heating increased, the Tg also rose with greater DA reactions occurring. The Tg

was 88.1 oC after 3 h, 90.5

oC after 5 h and finally 93.2

oC after 7 h. The final Tg was,

within experimental error, the same as that of the original sample, being some

92.4 oC. As with the DGEBA-based system previously described, this indicated that

the sample was able to recover to the same state as the original cured sample.


138

Figure 6.15 The tan traces of (a) cured TGAP-based epoxy polymer, (b) sample heated at

150 oC for 10 min, (c) sample heated at 150


oC for 1 h,

(d) sample heated at 150 oC for 10 min and then heated at 60

oC for 3 h, (e) sample heated at

150 oC for 10 min and then heated at 60

oC for 5 h, (f) sample heated at 150

oC for 10 min and

then heated at 60 oC for 7 h, from DMA. The heating rate of the DMA scan is 2

oC/min and

the measuring frequency is 1 Hz.

The swelling tests undertaken for DGEBA system were used to understand the

microstructure of the TGAP-based epoxy polymer during the healing process. From

the results of the DGEBA-based epoxy polymer, it was known that the sample would

lose weight after the swelling test, because the uncured monomers or the decoupled

oligomers arising from scissioning could be dissolved into the solvent. The first

parameter measured was the weight decrement after the swelling test. In this case, the

samples were dried at 50 oC overnight in vacuo after the swelling test, and their

weights were measured. The weight decrement of the samples is shown in Figure 6.16.

For the TGAP-based epoxy polymer, the percentage of the weight loss was 5.4 %,


139

which was lower than that of the DGEBA-based epoxy polymer (10.4 %) due to the

higher crosslink density. The sample treated at 150 oC for 10 min lost 24.6 %, which

was four times more than that of the cured sample, because the RDA reaction enabled

the DA units to scission, and thus more small, soluble branched polymers or

oligomers were dissolved. After the sample was held at 60 oC for several hours, the

percentage decrease of the sample weight reduced. Finally, after 7 h, the weight

decrement returned to the original level. This showed that the DA units were

gradually reformed, and the structure of the sample recovered to be similar to that of

the original sample.

Figure 6.16 The change in weight of (a) TGAP-based epoxy polymer, (b) sample heated at

150 oC for 10 min, (c) sample heated at 150


oC for 1 h,

(d) sample heated at 150 oC for 10 min and then heated at 60

oC for 3 h, (e) sample heated at

150 oC for 10 min and then heated at 60

oC for 5 h, (f) sample heated at 150

oC for 10 min and

then heated at 60 oC for 7 h, after being placed into acetone for 7 d, dried. All of these were

compared with the weight of the original sample.


140

Based on the dried weight of the samples, the results of the swelling test are shown in

Figure 6.17, with the data calculated using Equation 6.2. The weight increment of the

original cured sample was 24.8 %. After the sample was heated at 150 oC for 10 min,

the weight increment increased to 32 %, because the scission of the DA units occurred

during the treatment and the crosslink density reduced, leaving more free volume for

ingress of the solvent. When the sample with the lower crosslink density structure was

heated at 60 oC, the swelling ability of the sample decreased and reduced to 24.9 %

after 7 h, the same as that of the original cured sample. These results demonstrate that

the swelling ability of the sample can be restored and the structure of the original

sample recovered. Although a longer time was required, the results of the experiment

were the same as those for the lower crosslink density DGEBA-based epoxy polymer.

Figure 6.17 Swelling test results of the sample described in Figure 6.16, based on the final

dried sample weight.


141

The solubility of the cured sample of TGAP-based epoxy polymer at high

temperatures was also investigated using the same method as that for the DGEBA-

based epoxy polymer. The cured sample was placed into DMF solvent at 140 oC.

After 4 min, it was fully dissolved, which demonstrated that the DA units in the

crosslink were fully scissioned and that the cross-linked polymer was reduced to

branched small polymers or oligomers at the temperature above the RDA reaction

temperature of 120 oC. Nonetheless, it took twice as long (compared to the DGEBA

system) to dissolve the sample due to the higher crosslink density and the more DA

units in the TGAP-based sample.

Figure 6.18 The residual weight of cured sample of TGAP-based epoxy polymer after being

placed in DMF at (a) 140 oC, (b) 130

oC, (c) 120

oC and (d) 110

oC, all for 4 min.

Held in solvent for the same time (4 min), other cured samples were placed in DMF at

different temperatures of 130 oC, 120

oC and 110

oC. After 4 min the solution was


142

filtered and the residue was dried at 100 oC overnight in vacuo. Figure 6.18

demonstrates the residue weight. Some 10 % of the original weight remained when

the temperature was reduced to 130 oC. For treatments at 120

oC, more than 60 % of

material remained undissolved. This value increased to over 90 %, when the

temperature reduced to 110 oC, showing that the rate of the RDA reaction decreased

rapidly when the temperature was lowered. Compared with the results for the

DGEBA-based sample, higher temperatures and/or greater lengths of time were

required to effectively scission the DA units and decouple the architecture of the

cross-linked sample based on TGAP because of the more complicated polymer

structure.

GPC was used to analyse the dissolved component in the solutions obtained and the

resultant GPC chromatograms are shown in Figure 6.19. Three large peaks overlapped

in the range from 13 min to 17 min and the shapes of the curves were similar for the

different temperatures. Table 6.2 lists the Mw and the ratio of Mw/Mn of each peak. For

the second and third peak, the Mw in each chromatogram are similar, being about

1500 g/mol with Mw/Mn of 1.1 and about 570 g/mol with Mw/Mn of 1.0. These two

peaks thus appear to represent small molecular fragments separated from the polymer

by the RDA reaction. The Mw of the first peak in each curve is different, from

7 000 g/mol to 30 000 g/mol, along with the decrease of the solvent temperature from

140 oC to 110

oC. This demonstrates that fewer DA units were cleaved and thus larger

branched polymers or oligomers remained in samples when the temperature reduced.

Although the molecular weight is greater than that of DGEBA-based epoxy polymer,

the epoxy network was clearly broken down by the RDA reaction, and small

oligomeric species were produced.


143

Figure 6.19 GPC chromatograms (obtained at room temperature) of dissolved components of

scissioned TGAP-based epoxy polymer in DMF, undertaken at different temperatures.

Table 6.2 The Mw and Mw/Mn of each peak in each GPC chromatogram in Figure 6.19.

Temperature (oC) Mw (g/mol) Mw/Mn

140

6 817 1.35

1 555 1.10

583 1.05

130

12 616 2.12

1 555 1.13

560 1.05

120

27 150 3.65

1 522 1.10

585 1.04

110

31 370 4.37

1 490 1.11

560 1.04


144

The above experiments again proved that the DA units in the crosslink were

scissioned to deconstruct the cross-linked epoxy polymer by the RDA reaction, but

that they could reform to recover the cross-linked polymer structure through the DA

reaction.

6.4 Self-healing behaviour on the surface of polymer

The self-healing property of the DGEBA-based cross-linked epoxy polymer with DA

units was clearly demonstrated using optical microscopy. The self-healing behaviour

on the surface of the DGEBA-based polymer sample is shown in Figure 6.20.

Figure 6.20 Photographs of (a) scratches made on the surface of the cross-linked sample of

DGEBA-based epoxy polymer, (b) partially-healed sample after being heated at 140 oC for

3 min, (c) fully-healed sample after being heated at 140 oC for 30 min.


145

A scratch was made on the sample surface using a razor blade, the scratch being some

40 m in width (Figure 6.20a). After being heated at 140 oC for 3 min, the sample was

partially healed and the width of the scratch decreased to about 15 m (Figure 6.20b).

The scratch disappeared and became completely healed after heating at 140 oC for 30

min (Figure 6.20c). The fully-healed sample was then heated at 75 oC for 5 h to reform

the crosslink in the polymer. The round bubble inside the sample and the right edge of

the sample can be used to locate the original position of the scratch. This experiment

thus made clear that the DA units in the damaged part of the sample were cleaved by

RDA reaction at high temperature, and the resulting branched small polymers or

oligomers were able to flow to heal the scratch.

In order to prove the dominant influence of the DA units in the self-healing process, a

control sample of cross-linked epoxy polymer cured with Ethacure 100, which did not

contain a DA-based unit, was used to run the same experiment for comparison. The

result is shown in Figure 6.21.

Figure 6.21 Photographs of control sample of epoxy polymer cross-linked with Ethacure 100

diamine (no DA adducts) (a) scratches made on the surface, (b) sample after heating at 140 oC

for 30 min.


146

It can be seen that the scratch was slightly changed as a result of the enthalpy change

during the heating process after the sample was heated at 140 oC for 30 min. Although

this influence helped the scratch to heal a little, it was not able to fully heal the scratch

and cause it to not be visible. Thus, the epoxy polymer cured with the synthesised

diamine cross-linker with DA units clearly showed self-healing behaviour. This

recovery process was found to occur in the same location on the polymer sample for

3 times without any fatigue or diminution of healing potential, which showed the

multi-self-healing nature of the cross-linked epoxy polymer with DA units.

The effect of different temperature conditions on the self-healing ability of the

DGEBA-based epoxy polymer sample to repair the scratched surface was studied by

recording the healing process using a video camera attached to an optical microscope.

The scratch on the sample for each test was similar in width and depth. The result for

the temperature condition of 110 oC is shown in Figure 6.22.

Figure 6.22 Images of a scratch in the DGEBA-based epoxy polymer sample heated at

110 oC for (a) 0 min, (b) 0.16 min, (c) 1.66 min and (d) 4.45 min.


147

A scratch was made on the sample’s surface with some 50 m in width and some 5

m in depth (Figure 6.22a). The scratch changed and reduced to some 30 m in width

after the sample was heated at 110 oC for 0.16 min (Figure 6.22b). After 1.66 min, the

scratch became very shallow (Figure 6.22c). Finally, it disappeared and healed after

4.45 min (Figure 6.22d). For temperatures of 120 oC and 130

oC, the healing process

was similar to that of 110 oC, but the healing time reduced, since more energy was

provided by the higher temperatures to accelerate the healing process.

The result of healing time at different temperatures is shown in Table 6.3. Compared

with the RDA condition of 150 oC for 10 min obtained from FTIR experiments in

Section 6.2.2, the healing temperature was lower and the healing time much shorter in

the surface self-healing process, possibly due to the fact that only the material on the

surface of the sample needed to be scissioned and flow (as opposed to the bulk

sample). Therefore the scratch on the surface was healed when the temperature was

close to or above the RDA reaction temperature of 120 oC. For certain healing

temperatures, the healing time depended on the width and depth of the scratch, which

meant that bigger scratches needed more time to recover, and vice versa.

Table 6.3 Healing times for different annealing temperatures for the DGEBA-based epoxy

polymer. The scratch was some 50 m in width and some 5 m in depth.

Temperature (oC) Healing time (min)

110 4.45

120 1

130 0.16

From the thermal self-healing mechanism, it was clear that the small branched

polymers or oligomers produced by the RDA reaction flowed and healed the scratches


148

during heating. The healing time could thus be considered as the time for the

oligomers to flow sufficiently to fill 50 m scratches. In the above experiments, the

healing time was clearly influenced by the temperature and the data in Table 6.3 can

therefore be used to calculate the flow activation energy of the scissioned material.

We proposed to use the Arrhenius equation as follows to calculate the activation

energy of flow, a reasonable assumption given the relatively low molecular weights of

the components:176

t

B

RT

EAk a

exp (Equation 6.3)

where k is the flow speed; A is a constant; Ea is the flow activation energy; R is the

gas constant; T is the heating temperature; B is the flowing distance of the material,

which can be seen as a constant in this experiment because of the certain width of the

scratch; t is the healing time (1/t is healing rate). The Equation 6.3 can be rewritten as

Equation 6.4 as follows:

B

A

RT

E

t

a ln1

ln (Equation 6.4)

Using the data in Table 6.3, ln(1/t) against 1/T is shown in Figure 6.23. After linear

fitting, the result is as follows:

482.6025326 xy (Equation 6.5)

From Equation 6.4 and Equation 6.5, the flow activation energy is calculated as

follows:

molkJmolJREa /210/31441.82532625326


149

Figure 6.23 Arrhenius plots of 1/t ( a measure of the healing rate) as a function of 1/T for

sample of DGEBA-based epoxy polymer

The activation energy of flow of linear polymers is usually in the range of 20 kJ/mol

to 150 kJ/mol.177, 178

The result above is a little higher than this range, probably

because the oligomer resulting from the RDA reaction was highly branched so that it

could not flow like linear polymers. In addition, only three temperature points were

taken over a limited temperature range. Furthermore, the flow represented by this

equation was quiescent flow that filled the scratch on the surface of the polymer, as

opposed to a forced flow such as that due to a rheological test, and thus surface

tension and other factors could influence the result. Nonetheless, the result provides

some confirmation of the mechanism of self-healing, due to self-diffusion of the

cleaved components.

The self-healing properties of the epoxy polymer based on TGAP were also

investigated using a similar method. The result for the temperature at 130 oC is shown


150

in Figure 6.24. The scratch on the sample was some 40 m in width and 5 m in depth

(Figure 6.24a). After being heated at 130 oC for 0.16 min, the scratch changed to some

10 m in width (Figure 6.24b). There was only a little scratch left after 1.66 min

(Figure 6.24c) and it was hard to see the scratch after 5.7 min exposure (Figure 6.24d),

which meant it completely healed. The scratch on the sample used for tests at different

annealing temperatures was similar in width and depth. Although the healing

processes at temperatures of 140 oC and 150

oC were similar to that described above,

the times for healing to occur were shortened because the DA unit could readily

cleave and the oligomeric materials readily flow at the higher temperatures. The

healing temperature and time for the sample of TGAP-based polymer were higher

than that of DGEBA-based polymer due to a greater concentration of DA units and a

more highly cross-linked system. Table 6.4 shows the required healing times for the

different temperatures.

Figure 6.24 Micrographs of a scratch in the TGAP-based epoxy polymer sample heated at

130 oC for (a) 0 min, (b) 0.16 min, (c) 1.66 min and (d) 5.73 min.


151

Table 6.4 Healing time at different temperatures for the TGAP-based epoxy polymer. The

scratch was some 40 m in width and some 5 m in depth.

Temperature (oC) Healing time (min)

130 5.73

140 1.15

150 0.16

The data in Table 6.4 can be used to calculate the activation energy of flow of the de-

cross-linked material obtained from the degradation of the TGAP-based epoxy

polymer sample, using the Arrhenius equation as before (Equation 6.4). Figure 6.25

shows ln(1/t) against 1/T. After linear fitting, the result is as follows:

852.6830128 xy (Equation 6.6)

From Equation 6.4 and Equation 6.6, the activation energy of flow is calculated as

follows:

molkJmolJREa /250/31441.83012830128

The flow activation energy of the epoxy polymer based on TGAP was higher than that

of DGEBA, because the resultant oligomer was larger and more likely branched than

that of DGEBA-based polymer. The GPC results in Section 6.3 clearly show that the

molecular weight of the oligomer from TGAP-based polymer was greater than that

from DGEBA-based polymers. Thus, a greater energy of activation is required for the

oligomers and fragments from the TGAP-based polymer to flow and heal the surface

scratch of the sample, and higher healing temperatures are required.


152

Figure 6.25 Arrhenius plots of 1/t (rate of healing) as a function of 1/T for the sample of

TGAP-based epoxy polymer

6.5 Summary

In this chapter, the key feature of the thermal self-healing process, the DA and RDA

reactions of the DA units, was studied. Before investigating reacted epoxy monomer,

the diamine crosslinking unit with two DA units was studied to follow the occurrence

of the DA and RDA reactions by NMR. Due to the occurrence of Michael addition as

a side reaction, the neat diamine cross-linker itself was not suitable for the test. Boc-

protected diamine cross-linker was therefore used to analyse the DA and RDA

reactions. The NMR result clearly showed that the RDA reaction of the DA unit

scissioned the adduct at 120 oC and the DA reaction enabled it to reform at 75

oC.

This cycle could be repeated at least 3 times, which demonstrated that the DA and

RDA reactions in the diamine cross-linker were reversible.


153

Once the insoluble, cross-linked epoxy polymer was prepared, FTIR was chosen as

the best choice to investigate the chemical groups which tracked the DA and RDA

reactions. With the help of the maleimide carbonyl group peak as an internal standard,

the height changes of the maleimide group peak under different temperature

conditions were calculated and compared. For the bi-functional DGEBA-based epoxy

polymer, the most suitable conditions of the RDA and DA reactions were 150 oC for

10 min and 60 oC for 5 h. For the tri-functional TGAP-based epoxy polymer, the

optimal conditions chosen for the RDA and DA reactions were 150 oC for 10 min and

60 oC for 7 h.

Using these optimal conditions, the mechanism of the thermal self-healing was

studied. In the DMA test, the glass transition of the epoxy polymer reduced after the

sample was annealed at the temperature appropriate for causing the RDA reaction,

and the sample partly scissioned to small, branched polymer units or to oligomers

with higher thermal mobility. After the partly-scissioned sample was heated under the

DA reaction condition, the Tg returned to the value of the original sample due to the

restoration of the cross-linked structure. In the swelling test, more free volume was

produced within the cross-linked structure of the network polymer upon the cleavage

of the DA units by RDA reaction and the swelling ability of the sample increased. The

structure was able to be recovered by the restoration of the DA units when the sample

was held at the temperature relevant to the DA reaction and this sample’s swelling

ability reduced to the level of the original polymer. The scissioned cross-linked

polymer sample was able to be dissolved in organic solvent like DMSO or DMF at

temperatures close to or above the RDA reaction temperature of 120 oC. At lower

temperatures, dissolution was more difficult, and more residues remained following

this treatment. From the GPC result of the dissolved part of the polymer, it could be

seen that larger polymer segments remained in the solution at lower temperatures, due

to the reduced activity of the scissioning process. Further proof of scissioning could


154

be seen in a simple experiment, in which the cross-linked polymer sample was

observed to macroscopically flow and change shape under its own weight at

sufficiently high temperatures. All of the above results verified that the mechanism of

the thermal self-healing process consisted of three aspects: cleavage of the DA units

in the cross-linker by the RDA reaction, the flow of the resulting small branched

polymers or oligomers, and the subsequent reformation of the cross-linked network by

the DA reaction.

The results of the self-healing process were readily observable on the surface of the

sample using optical microscopy. A scratch made with a razor blade was visually fully

healed at temperatures close to or above 120 oC. Although the enthalpy change during

such heating could assist scratch healing to a very limited extent, it was clear by

comparison with a non-self-healable control sample, that the scratch was not healed

completely without DA units in its structure. Since in the case of scratch healing only

the surface of the polymer needed to cleave and reform, the healing time in this

situation was shorter than that obtained from the FTIR experiment, where a more bulk

healing was required.

The activation energy of flow of the scissioned material was calculated from the rate

of scratch healing of the sample. The value was similar to that of linear polymers,

albeit a little greater, probably due to its highly branched structure and the limited

temperature range that could be used to obtain the value. The healing temperatures

and the activation energy of flow/healing for the tri-functional TGAP-based epoxy

polymer was found to be greater than that of the DGEBA-based epoxy polymer, due

to the greater concentration of DA units in the polymer and its higher cross-linked

density.

Chapter 7 Conclusions and Future Work

155


7.1 Conclusions

Self-healing is a contemporary area of research in polymer science that is attracting

much attention. A range of self-healing strategies have been reported and used in

different applications. Thermosetting or cross-linked polymers particularly benefit

from this strategy because of the restricted chain mobility which does not allow

scratches or cracks to be filled by surrounding material, even if heated. Two main

strategies have been reported to date to introduce a self-healing ability, either

incorporating healing agents that release reactants and heal, or by building functional

chemical groups into the molecular structure that can be stimulated in some way to

cleave, allowing a degree of polymer flow, and then to be reformed again to allow the

original properties to be regained. Since there is a need for reversible and repeatable

healing, the latter method is more effective in potential applications. In addition,

because no second phase is incorporated, the material can remain clear, which can be

useful for coatings. In this research, for these reasons, we investigated the latter

approach.

Using the well-known furan-maleimide-based system, a diamine cross-linker with

thermally-reversible Diels-Alder (DA) units incorporated was proposed and

synthesised. The units could cleave via the Retro Diels-Alder (RDA) reaction and

reform by the DA reaction. Previous work in which the DA unit had been used in

epoxy resins had placed this moiety within the epoxy monomer itself. Our approach

was to place it within the cross-linker, and thus a range of widely-used epoxy

monomers or indeed mixtures of them (as often occurs in mixes in the aerospace

industry) could be used.


156

A diamine cross-linker with a single DA unit was attempted first. Unfortunately, after

many trials, including modifying the chemical structure of starting materials, using

different synthetic schema and applying a variety of experimental methods, the yield

of the expected product was too low and impure to allow useful application. A

synthetic route of the diamine cross-linker with two DA units was designed and the

target compound was successfully obtained via four reaction steps with a total yield of

41 %.

The above-mentioned diamine cross-linker with two DA units was then used to cure

two well-known epoxy monomers in wide commercial use with different

functionalities, in order to produce self-healing epoxy polymers with different

crosslink densities. The curing process of the epoxy polymers was investigated using

near-infrared spectroscopy (NIR) and thermally by differential scanning calorimetry

(DSC). Because the DA units were added to the structure of the cross-linker before

the curing reaction, high curing temperatures would initiate the occurrence of the

RDA reaction and thus allow the possibility of the Michael reaction between the

amine group and the maleimide group, thereby resulting in loss of amine reactant as

well as irreversible loss of the crosslinking agent. The appropriate curing conditions,

which not only result in as high a degree of cure as possible, but also use as low a

temperature as possible to prevent this side reaction from happening, were determined

following a series of experiments. The curing condition of DGEBA and the diamine

was found to be optimal at 60 oC for 12 h with a degree of cure of 81 %, whilst that of

TGAP was at 60 oC for 24 h with a degree of cure of 82 %. These epoxy polymers

were subsequently characterised using Fourier transform infrared spectroscopy

(FTIR), DSC, dynamic mechanical analysis (DMA) and thermogravimetric analysis

(TGA).


157

The DA and RDA reactions of the DA units in diamine cross-linker alone were first

examined using NMR in solution. These results showed the RDA reaction (unit

opening) occurred at 120 oC and the DA reaction (unit closing) at 75

oC. Because of

the insolubility of the cross-linked polymer, the DA and RDA reactions in the reacted

thermosetting polymer system could not be undertaken in the dissolved state by

NMR. However they could be detected using FTIR by monitoring the change in the

maleimide peak at 694 cm-1

. From the results of two series of experiments, suitable

conditions of DA and RDA reactions for both the bi- and tri-functional systems were

determined. In the DGEBA-based and TGAP-based systems, the most suitable

condition for the RDA reaction was found to be the same, 150 oC for 10 min.

However, for the same temperature condition of 60 oC, because of the high crosslink

density, the time for the DA reaction in the TGAP-based system to progress to a

similar degree was 2 h longer than that for the DGEBA-based system. Subsequent

self-healing testing using different characterisation methods was based on these

conditions.

The self-healing behaviour of the two cross-linked epoxy polymers was demonstrated

using a range of methods, including measuring the glass transition temperature (Tg)

using DMA, swelling tests, dissolution behaviour (at different temperatures) and

shape change. In the DMA tests, the Tg of the DGEBA-based epoxy polymer

decreased from 81.7 oC to 72.1

oC after treatment in the RDA reaction condition

(150 oC for 10 min). When the sample was heated in the DA reaction condition (60

oC

for 5 h), the Tg returned to the original value. The Tg of the TGAP-based epoxy

polymer also underwent a similar process during treatment in RDA and DA reaction

conditions. This demonstrated clearly that the cross-linked structure scissioned and

reformed. In the swelling tests, the cross-linked polymer sample lost non-cross-linked

material after it was heated under the RDA reaction conditions, due to the cleavage of

the DA units by the RDA reaction. If the sample underwent treatment to encourage


158

the DA reaction, the swelling ability recovered to the original level, which proved that

the structure of the sample was similar to that of the original sample. It was found that

the cross-linked sample could be dissolved in a solvent like DMSO or DMF after

being heated at 140 oC for a few minutes. With a decrease in the heating temperature,

more dissolution time was needed. Based on GPC measurements of the dissolved

component of the polymer, higher temperatures caused the molecular weight of the

dissolved component (oligomer) to have a lower molecular weight. For the DGEBA-

based epoxy polymer, the Mw of the dissolved component was about 12 000 g/mol

with Mw/Mn of 3.4 after heating at 110 oC for 2 min. When the temperature changed to

140 oC, the Mw decreased to about 4 000 g/mol with Mw/Mn of 1.6. The shape change

of the cross-linked polymer could also be observed at higher temperatures, which

indicated that the cross-linked system was cleaved and the resultant material able to

flow under its own weight. For the DGEBA-based epoxy polymer, the shape of the

sample was completely changed when the temperature became 140 oC.

The mechanism of the thermal self-healing was thus shown by the above results, and

the three stages of the self-healing process are clear. The DA units in the cross-linker

were broken under the condition of the RDA reaction. The segment of the cross-

linked polymer, branched small polymer or oligomer, was able to flow to fill the

scratches. Finally, the DA units reformed to restore the cross-linked structure of the

polymer by DA reaction.

The self-healing of a micrometer-sized scratch on the surface of the sample was also

studied using optical microscopy. The scratch was able to heal fully at temperatures

close to or above the RDA reaction temperature, 120 oC. The scratched control sample

without DA units was heated at 140 oC for 30 min but the scratch healed only a little

by enthalpy change and did not fully disappear based on visual inspection. This

comparison proved that the cleavage of the DA units and subsequent flow of the


159

oligomer were the basis of the self-healing process. The healing temperature of the

TGAP-based epoxy polymer was higher than that of the DGEBA-based epoxy

polymer, because it had more DA units and a higher cross-linked density. According

to the Arrhenius equation, the flow activation energy of the oligomer obtained from

the decomposition of epoxy polymer can be calculated using the data on healing

temperature and time. The result was 210 kJ/mol for the DGEBA-based epoxy

polymer and 250 kJ/mol for the TGAP-based one. Both of the results were higher

than the range of normal linear polymers due to the likely highly branched structure of

the cleaved units.

7.2 Future work

Self-healing epoxy-based polymers have potential applications in coatings and

construction materials. However, more research is required to further improve the

self-healing conditions and make them more readily applicable.

Other epoxy monomers could also be cured by this diamine cross-linker with two DA

units. The epoxy monomer could have, for example, more epoxide groups like

N,N,N’,N’-tetraglycidyl-4,4’-diaminodiphenylmethane (TGDDM), which has

tetraglycidyl moieties to produce mendable epoxy polymers with a higher cross-

linked density.

The structure of the diamine cross-linker could thus be adjusted to modify the

physical and mechanical properties of the final, cross-linked polymer. Shorter carbon

chains or phenyl groups could be built in the diamine cross-linker between two DA

units, which may decrease the mobility of the cross-linker, resulting in higher Tg of

the cured cross-linked epoxy polymer. However, high Tg of polymer would reduce the

mobility of polymer and inhibit the DA reaction, which is problem that needs to be

considered when designing monomers. Care would also need to be taken to observe


160

how the resulting cure temperature was related in magnitude to the DA and RDA

reaction temperatures. The number of DA units in the cross-linker could be increased

to give it greater functionality and thus form more highly cross-linked thermosets with

greater rigidity, for example. The structure of furan and maleimide group could also

be altered to modify the DA and RDA reaction temperatures, such as using methylene

instead of oxygen in the furan group or using methine instead of nitrogen in the

maleimide group.

The curing process of the self-healing polymer could be further studied by

demonstrating the concentration changes of the primary and secondary amine groups

during the reaction. They would be compared with the epoxy group conversion to

check the self-consistency of the reaction. The time-temperature-transformation

behaviour of the new self-healing system would be investigated to obtain a higher

level understanding and masterminding of the phenomenon. It could be interesting

that the system with RDA reactions would show novel features compared to the

standard epoxy system.

Various nanoparticles, which could gain heat energy through different techniques,

could also be dispersed in these self-healing polymers. The final composite could

change the method to trigger the healing process from direct heating to different ways

such as ultraviolet (UV) irradiation and exposure to alternating magnetic fields. For

the UV-initiator, titanium oxide is a well-known photo-thermal catalyst under UV

light, which could cause the temperature increase of the sample to initiate self-

healing. Nanoparticles of titanium oxide like P25, which have an average particle size

of 25 nm, should be able to produce heat energy from UV light (the photothermal

effect). Alternatively, the inclusion of magnetic particles such as magnetite offer the

possibility of applying alternating magnetic fields for heating within the particles by

relaxational and hysteresis losses. This heat in turn can be distributed within the


161

polymer (provided the particles are well dispersed) and cause the temperature of the

sample to rise, leading to self-healing. By changing the strength of the magnetic field,

the temperature of the sample could thus be varied and controlled. Different kinds of

nanoparticles with different sizes could be studied to optimise the self-healing

behaviour of the cross-linked polymers mentioned above.

References

162

References

1. Y. C. Yuan, T. Yin, M. Z. Rong and M. Q. Zhang, Express Polym. Lett., 2008,

2, 238-250.

2. S. Bergman and F. Wudl, Self Healing Materials, Springer, Dordrecht, 2007,

pp. 45-68.

3. X. Chen, M. A. Dam, K. Ono, A. Mal, H. Shen, S. R. Nutt, K. Sheran and F.

Wudl, Science, 2002, 295, 1698-1702.

4. S. D. Bergman and F. Wudl, J. Mater. Chem., 2008, 18, 41-62.

5. N. Bai, K. Saito and G. P. Simon, Polym. Chem., 2013, 4, 724-730.

6. S. H. Goodman, Handbook of Thermoset Plastics, Noyes Publications,

Berkshire, 1998, ch. 6, pp. 193-268.

7. S. R. White, N. R. Sottos, P. H. Geubelle, J. S. Moore, M. R. Kessler, S. R.

Sriram, E. N. Brown and S. Viswanathan, Nature, 2001, 409, 794-797.

8. A. Yousefpour, M. Hojjati and J. P. Immarigeon, J. Thermoplast. Compos.

Mater., 2004, 17, 303-341.

9. J. Paul and R. Jones, Eng. Fract. Mech., 1992, 41, 127-141.

10. K. B. Zimmerman and D. Liu, Exp. Mech., 1996, 36, 142-147.

11. D. Y. Wu, S. Meure and D. Solomon, Prog. Polym. Sci., 2008, 33, 479-522.

12. R. S. Trask, H. R. Williams and I. P. Bond, Bioinsp. Biomim., 2007, 2, P1-P9.

13. E. B. Murphy and F. Wudl, Prog. Polym. Sci., 2010, 35, 223-251.

14. S. M. Bleay, C. B. Loader, V. J. Hawyes, L. Humberstone and P. T. Curtis,

Composites Part A, 2001, 32, 1767-1776.

15. C. Dry, Compos. Struct., 1996, 35, 263-269.

16. V. C. Li, Y. M. Lim and Y. W. Chan, Composites Part B, 1998, 29, 819-827.

17. C. Dry, Int. J. Mod. Phys. B, 1992, 6, 2763-2771.

18. C. Dry, Smart Mater. Struct., 1994, 3, 118-123.

References

163

19. C. Dry and W. McMillan, Smart Mater. Struct., 1996, 5, 297-300.

20. M. Motuku, U. K. Vaidya and G. M. Janowski, Smart Mater. Struct., 1999, 8,

623-638.

21. J. W. C. Pang and I. P. Bond, Compos. Sci. Technol., 2005, 65, 1791-1799.

22. R. S. Trask and I. P. Bond, Smart Mater. Struct., 2006, 15, 704-710.

23. R. S. Trask, G. J. Williams and I. P. Bond, J. R. Soc. Interface, 2007, 4, 363-

371.

24. J. M. Asua, Prog. Polym. Sci., 2002, 27, 1283-1346.

25. E. N. Brown, M. R. Kessler, N. R. Sottos and S. R. White, J. Microencapsul.,

2003, 20, 719-730.

26. B. J. Blaiszik, N. R. Sottos and S. R. White, Compos. Sci. Technol., 2008, 68,

978-986.

27. E. N. Brown, N. R. Sottos and S. R. White, Exp. Mech., 2002, 42, 372-379.

28. X. Liu, J. K. Lee, S. H. Yoon and M. R. Kessler, J. Appl. Polym. Sci., 2006,

101, 1266-1272.

29. J. D. Rule, E. N. Brown, N. R. Sottos, S. R. White and J. S. Moore, Adv.

Mater., 2005, 17, 205-208.

30. G. O. Wilson, M. M. Caruso, N. T. Reimer, S. R. White, N. R. Sottos and J. S.

Moore, Chem. Mater., 2008, 20, 3288-3297.

31. S. H. Cho, H. M. Andersson, S. R. White, N. R. Sottos and P. V. Braun, Adv.

Mater., 2006, 18, 997-1000.

32. M. W. Keller, S. R. White and N. R. Sottos, Adv. Funct. Mater., 2007, 17,

2399-2404.

33. T. Yin, M. Z. Rong, M. Q. Zhang and G. C. Yang, Compos. Sci. Technol.,

2007, 67, 201-212.

34. Y. C. Yuan, M. Z. Rong, M. Q. Zhang, J. Chen, G. C. Yang and X. M. Li,

Macromolecules, 2008, 41, 5197-5202.

35. K. S. Toohey, N. R. Sottos, J. A. Lewis, J. S. Moore and S. R. White, Nature

Mater., 2007, 6, 581-585.

References

164

36. A. D. Stroock and M. Cabodi, MRS Bull., 2006, 31, 114-119.

37. A. R. Hamilton, N. R. Sottos and S. R. White, Adv. Mater., 2010, 22, 5159-

5163.

38. H. R. Williams, R. S. Trask and I. P. Bond, Smart Mater. Struct., 2007, 16,

1198-1207.

39. K. S. Toohey, C. J. Hansen, J. A. Lewis, S. R. White and N. R. Sottos, Adv.

Funct. Mater., 2009, 19, 1399-1405.

40. M. Zako and N. Takano, J. Intell. Mater. Syst. Struct., 1999, 10, 836-841.

41. F. W. Jones and S. A. Hayes, WO Pat., WO 2005/066244 A2, 2005.

42. S. A. Hayes, F. R. Jones, K. Marshiya and W. Zhang, Composites Part A,

2007, 38, 1116-1120.

43. S. A. Hayes, W. Zhang, M. Branthwaite and F. R. Jones, J. R. Soc. Interface,

2007, 4, 381-387.

44. X. Luo, R. Ou, D. E. Eberly, A. Singhal, W. Viratyaporn and P. T. Mather, ACS

Appl. Mater. Inter., 2009, 1, 612-620.

45. X. Luo and P. T. Mather, ACS Macro Letters, 2013, 2, 152-156.

46. E. D. Rodriguez, X. Luo and P. T. Mather, ACS Appl. Mater. Inter., 2011, 3,

152-161.

47. O. Diels and K. Alder, Liebigs Ann. Chem., 1928, 460, 98-122.

48. J. M. Craven, US Pat., US 3435003, 1969.

49. M. P. Stevens and A. D. Jenkins, J. Polym. Sci. Polym. Chem. Ed., 1979, 17,

3675-3685.

50. S. A. Canary and M. P. Stevens, J. Polym. Sci., Part A: Polym. Chem., 1992,

30, 1755-1760.

51. Y. Chujo, K. Sada and T. Saegusa, Macromolecules, 1990, 23, 2636-2641.

52. Y. Imai, H. Itoh, K. Naka and Y. Chujo, Macromolecules, 2000, 33, 4343-

4346.

53. H. Laita, S. Boufi and A. Gandini, Eur. Polym. J., 1997, 33, 1203-1211.

References

165

54. C. Goussé, A. Gandini and P. Hodge, Macromolecules, 1998, 31, 314-321.

55. R. Gheneim, C. Perez-Berumen and A. Gandini, Macromolecules, 2002, 35,

7246-7253.

56. C. Goussé and A. Gandini, Polym. Bull., 1998, 40, 389-394.

57. E. Goiti, M. B. Huglin and J. M. Rego, Polymer, 2001, 42, 10187-10193.

58. E. Goiti, F. Heatley, M. B. Huglin and J. M. Rego, Eur. Polym. J., 2004, 40,

1451-1460.

59. E. Goiti, M. B. Huglin and J. M. Rego, Macromol. Rapid Commun., 2003, 24,

692-696.

60. E. Goiti, M. B. Huglin and J. M. Rego, Eur. Polym. J., 2004, 40, 219-226.

61. G. C. Tesoro and V. R. Sastri, Ind. Eng. Chem. Prod. Res. Dev., 1986, 25, 444-

448.

62. X. He, V. R. Sastri and G. C. Tesoro, Makromol. Chem., Rapid Commun.,

1988, 9, 191-194.

63. C. D. Diakoumakos and J. A. Mikroyannidis, J. Polym. Sci., Part A: Polym.

Chem., 1992, 30, 2559-2567.

64. C. D. Diakoumakos and J. A. Mikroyannidis, Eur. Polym. J., 1994, 30, 465-

472.

65. N. Kuramoto, K. Hayashi and K. Nagai, J. Polym. Sci., Part A: Polym. Chem.,

1994, 32, 2501-2504.

66. J. R. McElhanon and D. R. Wheeler, Org. Lett., 2001, 3, 2681-2683.

67. J. R. McElhanon, E. M. Russick, D. R. Wheeler, D. A. Loy and J. H. Aubert, J.

Appl. Polym. Sci., 2002, 85, 1496-1502.

68. X. X. Chen, F. Wudl, A. K. Mal, H. B. Shen and S. R. Nutt, Macromolecules,

2003, 36, 1802-1807.

69. T. A. Plaisted and S. Nemat-Nasser, Acta Mater., 2007, 55, 5684-5696.

70. Y. L. Liu and C. Y. Hsieh, J. Polym. Sci., Part A: Polym. Chem., 2006, 44,

905-913.

References

166

71. Y. L. Liu and Y. W. Chen, Macromol. Chem. Phys., 2007, 208, 224-232.

72. M. Watanabe and N. Yoshie, Polymer, 2006, 47, 4946-4952.

73. C. Zeng, H. Seino, J. Ren, K. Hatanaka and N. Yoshie, Macromolecules, 2013,

46, 1794-1802.

74. B. Gotsmann, U. Duerig, J. Frommer and C. J. Hawker, Adv. Funct. Mater.,

2006, 16, 1499-1505.

75. Q. Tian, Y. C. Yuan, M. Z. Rong and M. Q. Zhang, J. Mater. Chem., 2009, 19,

1289-1296.

76. Q. A. Tian, M. Z. Rong, M. Q. Zhang and Y. C. Yuan, Polym. Int., 2010, 59,

1339-1345.

77. Y. Zhang, A. A. Broekhuis and F. Picchioni, Macromolecules, 2009, 42, 1906-

1912.

78. H. Staudinger and H. A. Bruson, Justus Liebigs Ann. Chem., 1926, 447, 97-

110.

79. H. Staudinger and H. A. Bruson, Justus Liebigs Ann. Chem., 1926, 447, 110-

122.

80. J. K. Stille and L. Plummer, J. Org. Chem., 1961, 26, 4026-4029.

81. J. P. Kennedy and K. F. Castner, J. Polym. Sci. Polym. Chem. Ed., 1979, 17,

2039-2054.

82. J. P. Kennedy and K. F. Castner, J. Polym. Sci. Polym. Chem. Ed., 1979, 17,

2055-2070.

83. J. P. Kennedy, G. M. Carlson and K. Riebel, Polym. Bull., 1983, 9, 268-275.

84. J. P. Kennedy and G. M. Carlson, J. Polym. Sci. Polym. Chem. Ed., 1983, 21,

2973-2986.

85. J. P. Kennedy and G. M. Carlson, J. Polym. Sci. Polym. Chem. Ed., 1983, 21,

3551-3561.

86. M. Miura, F. Akutsu, T. Usui, Y. Ikebukuro and K. Nagakubo, Makromol.

Chem., 1985, 186, 473-481.

87. J. C. Salamone, Y. Chung, S. B. Clough and A. C. Watterson, J. Polym. Sci.,

References

167

Part A: Polym. Chem., 1988, 26, 2923-2939.

88. X. Chen and E. Ruckenstein, J. Polym. Sci., Part A: Polym. Chem., 1999, 37,

4390-4401.

89. E. Ruckenstein and X. Chen, J. Polym. Sci., Part A: Polym. Chem., 2000, 38,

818-825.

90. X. Chen and E. Ruckenstein, J. Polym. Sci., Part A: Polym. Chem., 2000, 38,

1662-1672.

91. X. Chen and S. Jiao, Acta Polym. Sin., 1999, 5, 564-570.

92. E. B. Murphy, E. Bolanos, C. Schaffner-Hamann, F. Wudl, S. R. Nutt and M.

L. Auad, Macromolecules, 2008, 41, 5203-5209.

93. J. S. Park, K. Takahashi, Z. H. Guo, Y. Wang, E. Bolanos, C. Hamann-

Schaffner, E. Murphy, F. Wudl and H. T. Hahn, J. Compos. Mater., 2008, 42,

2869-2881.

94. S. Dumitrescu, M. Grigoras and A. Natansohn, J. Polym. Sci. Polym. Lett. Ed.,

1979, 17, 553-559.

95. C. I. Simionescu, M. Grigoras and V. Bărboiu, J. Polym. Sci. Polym. Chem.

Ed., 1985, 23, 2089-2098.

96. M. P. Stevens, J. Polym. Sci. Polym. Lett. Ed., 1984, 22, 467-471.

97. C. I. Simionescu and M. Grigoras, J. Polym. Sci., Part C: Polym. Lett., 1990,

28, 39-45.

98. M. Grigoras and G. Colotin, Polym. Int., 2001, 50, 1375-1378.

99. J. R. Jones, C. L. Liotta, D. M. Collard and D. A. Schiraldi, Macromolecules,

1999, 32, 5786-5792.

100. T. Brand and M. Klapper, Des. Monomers Polym., 1999, 2, 287-309.

101. C. Lee and W. H. Daly, Fortschritte der Hochpolymeren-Forschung, Springer,

Berlin Heidelberg, 1974, vol. 15, ch. 3, pp. 61-90.

102. G. Wulff and I. Schulze, Angew. Chem. Int. Ed. Engl., 1978, 17, 537-538.

103. Y. Chujo, K. Sada, A. Naka, R. Nomura and T. Saegusa, Macromolecules,

1993, 26, 883-887.

References

168

104. G. C. Tesoro and V. R. Sastri, US Pat., US 5260411, 1993.

105. N. V. Tsarevsky and K. Matyjaszewski, Macromolecules, 2002, 35, 9009-

9014.

106. J. Canadell, H. Goossens and B. Klumperman, Macromolecules, 2011, 44,

2536-2541.

107. J. Kamada, K. Koynov, C. Corten, A. Juhari, J. A. Yoon, M. W. Urban, A. C.

Balazs and K. Matyjaszewski, Macromolecules, 2010, 43, 4133-4139.

108. J. A. Yoon, J. Kamada, K. Koynov, J. Mohin, R. Nicolaÿ, Y. Zhang, A. C.

Balazs, T. Kowalewski and K. Matyjaszewski, Macromolecules, 2011, 45,

142-149.

109. Y. Amamoto, H. Otsuka, A. Takahara and K. Matyjaszewski, Adv. Mater.,

2012, 24, 3975-3980.

110. Y. Amamoto, J. Kamada, H. Otsuka, A. Takahara and K. Matyjaszewski,

Angew. Chem., 2011, 123, 1698-1701.

111. Y. Higaki, H. Otsuka and A. Takahara, Macromolecules, 2004, 37, 1696-1701.

112. Y. Higaki, H. Otsuka and A. Takahara, Macromolecules, 2006, 39, 2121-2125.

113. C. Yuan, M. Z. Rong, M. Q. Zhang, Z. P. Zhang and Y. C. Yuan, Chem. Mater.,

2011, 23, 5076-5081.

114. Y. Chujo, K. Sada and T. Saegusa, Macromolecules, 1990, 23, 2693-2697.

115. Y. Zheng, F. M. Andreopoulos, M. Micic, Q. Huo, S. M. Pham and R. M.

Leblanc, Adv. Funct. Mater., 2001, 11, 37-40.

116. Y. Zheng, M. Micic, S. V. Mello, M. Mabrouki, F. M. Andreopoulos, V.

Konka, S. M. Pham and R. M. Leblanc, Macromolecules, 2002, 35, 5228-

5234.

117. M. Coursan, J. P. Desvergne and A. Deffieux, Macromol. Chem. Phys., 1996,

197, 1599-1608.

118. C. M. Chung, Y. S. Roh, S. Y. Cho and J. G. Kim, Chem. Mater., 2004, 16,

3982-3984.

119. V. Berl, M. Schmutz, M. J. Krische, R. G. Khoury and J. M. Lehn, Chem. Eur.

J., 2002, 8, 1227-1244.

References

169

120. R. P. Sijbesma, F. H. Beijer, L. Brunsveld, B. J. B. Folmer, J. H. K. K.

Hirschberg, R. F. M. Lange, J. K. L. Lowe and E. W. Meijer, Science, 1997,

278, 1601-1604.

121. J. H. K. K. Hirschberg, F. H. Beijer, H. A. van Aert, P. C. M. M. Magusin, R.

P. Sijbesma and E. W. Meijer, Macromolecules, 1999, 32, 2696-2705.

122. R. F. M. Lange, M. Van Gurp and E. W. Meijer, J. Polym. Sci., Part A: Polym.

Chem., 1999, 37, 3657-3670.

123. B. J. B. Folmer, R. P. Sijbesma, R. M. Versteegen, J. A. J. van der Rijt and E.

W. Meijer, Adv. Mater., 2000, 12, 874-878.

124. A. El-ghayoury, E. Peeters, A. P. H. J. Schenning and E. W. Meijer, Chem.

Commun., 2000, 1969-1970.

125. J. H. K. K. Hirschberg, L. Brunsveld, A. Ramzi, J. A. J. M. Vekemans, R. P.

Sijbesma and E. W. Meijer, Nature, 2000, 407, 167-170.

126. A. P. H. J. Schenning, P. Jonkheijm, E. Peeters and E. W. Meijer, J. Am. Chem.

Soc., 2000, 123, 409-416.

127. A. W. Bosman, R. P. Sijbesma and E. W. Meijer, Mater. Today, 2004, 7, 34-39.

128. G. M. L. van Gemert, J. W. Peeters, S. H. M. Söntjens, H. M. Janssen and A.

W. Bosman, Macromol. Chem. Phys., 2012, 213, 234-242.

129. S. Yagai, M. Higashi, T. Karatsu and A. Kitamura, Chem. Mater., 2004, 16,

3582-3585.

130. F. Ilhan, T. H. Galow, M. Gray, G. Clavier and V. M. Rotello, J. Am. Chem.

Soc., 2000, 122, 5895-5896.

131. F. Ilhan, M. Gray and V. M. Rotello, Macromolecules, 2001, 34, 2597-2601.

132. P. Cordier, F. Tournilhac, C. Soulie-Ziakovic and L. Leibler, Nature, 2008,

451, 977-980.

133. D. Montarnal, P. Cordier, C. Soulié-Ziakovic, F. Tournilhac and L. Leibler, J.

Polym. Sci., Part A: Polym. Chem., 2008, 46, 7925-7936.

134. U. S. Schubert and C. Eschbaumer, Angew. Chem. Int. Ed., 2002, 41, 2892-

2926.

135. S. Schmatloch, M. F. González and U. S. Schubert, Macromol. Rapid

References

170

Commun., 2002, 23, 957-961.

136. J. B. Beck, J. M. Ineman and S. J. Rowan, Macromolecules, 2005, 38, 5060-

5068.

137. P. K. Iyer, J. B. Beck, C. Weder and S. J. Rowan, Chem. Commun., 2005, 319-

321.

138. M. Burnworth, L. Tang, J. R. Kumpfer, A. J. Duncan, F. L. Beyer, G. L. Fiore,

S. J. Rowan and C. Weder, Nature, 2011, 472, 334-337.

139. F. R. Kersey, D. M. Loveless and S. L. Craig, J. R. Soc. Interface, 2007, 4,

373-380.

140. M. J. Harrington, A. Masic, N. Holten-Andersen, J. H. Waite and P. Fratzl,

Science, 2010, 328, 216-220.

141. N. Holten-Andersen, M. J. Harrington, H. Birkedal, B. P. Lee, P. B.

Messersmith, K. Y. C. Lee and J. H. Waite, Proc. Natl. Acad. Sci., 2011, 108,

2651-2655.

142. A. Harada, J. Polym. Sci., Part A: Polym. Chem., 2006, 44, 5113-5119.

143. M. Nakahata, Y. Takashima, H. Yamaguchi and A. Harada, Nature Commun.,

2011, 2, 511.

144. M. Zhang, D. Xu, X. Yan, J. Chen, S. Dong, B. Zheng and F. Huang, Angew.

Chem., 2012, 124, 7117-7121.

145. X. Yan, D. Xu, J. Chen, M. Zhang, B. Hu, Y. Yu and F. Huang, Polym. Chem.,

2013, 4, 3312-3322.

146. S. Burattini, H. M. Colquhoun, B. W. Greenland and W. Hayes, Faraday

Discuss., 2009, 143, 251-264.

147. S. J. Kalista, T. C. Ward and Z. Oyetunji, Mech. Adv. Mater. Struc., 2007, 14,

391-397.

148. S. J. Kalista and T. C. Ward, J. R. Soc. Interface, 2007, 4, 405-411.

149. R. J. Varley and S. van der Zwaag, Acta Mater., 2008, 56, 5737-5750.

150. R. J. Varley and S. van der Zwaag, Polym. Int., 2010, 59, 1031-1038.

151. R. John Varley and S. van der Zwaag, Polym. Test., 2008, 27, 11-19.

References

171

152. B. G. Min, Z. H. Stachurski, J. H. Hodgkin and G. R. Heath, Polymer, 1993,

34, 3620-3627.

153. G. Lachenal, A. Pierre and N. Poisson, Micron, 1996, 27, 329-334.

154. N. Poisson, G. Lachenal and H. Sautereau, Vib. Spectrosc, 1996, 12, 237-247.

155. J. Guy, K. Caron, S. Dufresne, S. W. Michnick, Skene and J. W. Keillor, J. Am.

Chem. Soc., 2007, 129, 11969-11977.

156. P. v. d. Veken, E. H. C. Dirksen, E. Ruijter, R. C. Elgersma, A. J. R. Heck, D.

T. S. Rijkers, M. Slijper and R. M. J. Liskamp, ChemBioChem, 2005, 6, 2271-

2280.

157. P. Y. Reddy, S. Kondo, T. Toru and Y. Ueno, J. Org. Chem., 1997, 62, 2652-

2654.

158. Y. S. Liu, C. Zhao, D. E. Bergbreiter and D. Romo, J. Org. Chem., 1998, 63,

3471-3473.

159. G. Kossmehl, H. I. Nagel and A. Pahl, Angew. Makromol. Chem., 1995, 227,

139-157.

160. C. A. G. Haasnoot, F. Deleeuw and C. Altona, Tetrahedron, 1980, 36, 2783-

2792.

161. F. Benito-Lopez, A. J. Mettelalj, R. J. M. Egberink, A. H. Velders, R. M.

Tiggelaar, H. Gardeniers, D. N. Reinhoudt and W. Verboom, Chim. Oggi-

Chem. Today, 2010, 28, 56-59.

162. J. H. Cooley and R. V. Williams, J. Chem. Educ., 1997, 74, 582-585.

163. L. Rulíšek, P. Šebek, Z. Havlas, R. Hrabal, P. Čapek and A. Svatoš, J. Org.

Chem., 2005, 70, 6295-6302.

164. H. E. Gottlieb, V. Kotlyar and A. Nudelman, J. Org. Chem., 1997, 62, 7512-

7515.

165. C. Dardonville, C. Fernandez-Fernandez, S. L. Gibbons, G. J. Ryan, N.

Jagerovic, A. M. Gabilondo, J. J. Meana and L. F. Callado, Biorg. Med. Chem.,

2006, 14, 6570-6580.

166. A. M. Scutaru, M. Wenzel and R. Gust, Eur. J. Med. Chem., 2011, 46, 1604-

1615.

References

172

167. H. Block, S. M. Walker and R. Groves, Polymer, 1972, 13, 527-535.

168. G. Scheltjens, M. M. Diaz, J. Brancart, G. Van Assche and B. Van Mele, React.

Funct. Polym., 2013, 73, 413-420.

169. R. J. Varley, J. H. Hodgkin, D. G. Hawthorne and G. P. Simon, J. Appl. Polym.

Sci., 1996, 60, 2251-2263.

170. H. Stutz and J. Mertes, J. Polym. Sci., Part A: Polym. Chem., 1993, 31, 2031-

2037.

171. R. J. Varley, G. R. Heath, D. G. Hawthorne, J. H. Hodgkin and G. P. Simon,

Polymer, 1995, 36, 1347-1355.

172. M. El Gouri, A. El Bachiri, S. E. Hegazi, M. Rafik and A. El Harfi, Polym.

Degrad. Stab., 2009, 94, 2101-2106.

173. T. Dyakonov, P. J. Mann, Y. Chen and W. T. K. Stevenson, Polym. Degrad.

Stab., 1996, 54, 67-83.

174. R. D. Patel, R. G. Patel and V. S. Patel, Thermochim. Acta, 1988, 128, 149-

156.

175. S. Magana, A. Zerroukhi, C. Jegat and N. Mignard, React. Funct. Polym.,

2010, 70, 442-448.

176. S. Y. Gu, K. Zhang, J. Ren and H. Zhan, Carbohydr. Polym., 2008, 74, 79-85.

177. V. P. Privalko and Y. S. Lipatov, J. Polym. Sci. Polym. Phys. Ed., 1976, 14,

1725-1727.

178. E. A. Collins and C. A. Krier, Trans. Soc. Rheol., 1967, 11, 225-242.

Publications

173

Publications

The following publications resulted from the research conducted during the author’s

Doctor of Philosophy candidature:

(1) N. Bai, K. Saito, G.P. Simon, “Synthesis of a diamine cross-linker containing

Diels-Alder adducts to produce self-healing thermosetting epoxy polymer from a

widely used epoxy monomer”, Polymer Chemistry 2013, 4, 724-730.

(2) N. Bai, G.P. Simon, K. Saito, “Investigation of the thermal self-healing mechanism

in a cross-linked epoxy system”, RSC Adv., 2013, 3, 20699-20707.

(3) N. Bai, G.P. Simon, K. Saito, “Epoxy Resins”, PCT/AU2013/000518, 2012.

Documents

New Self-healing Epoxy-based Polymers...mechanism of thermal self-healing in the prepared epoxy polymers. By different methods, including the determination of the change in glass transition