Sustainable Alternative for Advanced Materials Liang

Home , Transalkylation

BIOBASED EPOXY COMPOSITES：SUSTAINABLE

ALTERNATIVE FOR ADVANCED MATERIALS

LIANG YUE

Submitted in partial fulfillment to the requirements for the degree of

Doctor of Philosophy

Department of Macromolecular Science and Engineering

CASE WESTERN RESERVE UNIVERSITY

May, 2018 CASE WESTERN RESERVE UNIVERSITY

SCHOOL OF GRADUATE STUDIES

We hereby approve the thesis/dissertation of

Liang Yue

candidate for the degree of Doctor of Philosophy *.

Committee Chair

Ica Manas-Zloczower

Committee Member

David Schiraldi

Committee Member

Donald Feke

Committee Member

Joao Maia

Date of Defense

11/17/2017

*We also certify that written approval has been obtained

for any proprietary material contained therein.

Table of Contents

List of Figures……………...……………………………………………...vii

List of Schemes………………………………………………….………….xi

List of Tables……………………………………………………………...xiii

Acknowledgement…...…………………………………..……………….xiv

Abstract…………….……………………………………..……………..xv

Chapter 1: Introduction…………………………………………………………………1

1.1 Thermoset Composites………………………………………………...………………2

1.1.1 Thermoset Resin…………………………………………………………….2

1.1.2 Reinforcement Materials for Epoxy Composites…………………………..10

1.2 Biobased Thermoset Composites…………………………………………………….12

1.2.1 Biobased Epoxy Resin……………………………………………………..12

1.2.2 Biobased Fillers……………………………………………………………17

1.3 Thesis Scope…………………………………………………………………………24

1.4 References…………………………………………………………………..………..26

Chapter 2: Biobased Epoxy Resin for Vacuum Infusion Processing……………..…31

2.1 Introduction…………………………………………………………………………..32

2.1.1 Vacuum Infusion Processing of Fiber Reinforced Epoxy Composites……32

2.1.2 Biobased Epoxy Resin……………………………………………………..33

2.1.3 Biobased Reactive Diluent…………………………………………………36

2.2 Experimental Section………………………………………………………………...41

2.2.1 Materials…………………………………………………………………...41

2.2.2 Preparation of Composites…………………………………………………41

2.2.3 Characterizations…………………………………………………………...43

2.3 Results and Discussion………………………………………………………………44

2.3.1 Chemorheology of Curing Study…………………………………………..44

2.3.2 Dynamic Mechanical Analysis of The composites………...……………….46

2.3.3 Mechanical Properties……………………………………………………...48

2.3.4 Morphology of Composites………………………………………………..51

2.3.5 Thermal Stability…………………………………………………………..52

2.4 Conclusions…………………………………………………………………………..53

2.5 References……………………………………………………………………………55

Chapter 3: Surface-Modified Cellulose Nanocrystals for Biobased Epoxy

Nanocomposites…………………………………………………………………………60

3.1 Introduction…………………………………………………………………………..61

iii

3.2 Experimental Section………………………………………………………………...65

3.2.1 Materials…………………………………………………………………...65

3.2.2 Preparation of Cellulose Nanocrystals……………………………………..65

3.2.3 Organosilanization of Cellulose Nanocrystals……………………………..66

3.2.4 Preparation of Cellulose Nanocrystals/Biobased Epoxy Composites……..67

3.2.5 Characterizations……………………………………………...... …………67

3.3 Results and Discussion………………………………………………………………69

3.3.1 Organosilanization of Cellulose Nanocrystals……………………………..69

3.3.2 Cellulose Nanocrystals/epoxy composites…………………………………76

3.4 Conclusions…………………………………………………………………………..85

3.5 References……………………………………………………………………………86

Chapter 4: Bacterial Cellulose Nanofiber Mats as Reinforcement for Epoxy-

Anhydride systems……………………………………………………………………...90

4.1 Introduction…………………………………………………………………………..91

4.2 DGEBA Reinforced with Bacterial Cellulose Mats…………………………………92

4.2.1 Experimental Section………………………………………………………92

4.2.1.1 Materials…………………………………………………………92

4.2.1.2 Biosynthesis of Bacterial Cellulose Mats………………...……...93

4.2.1.3 DGEBA/BC Mats Composites Preparation……………………..94

4.2.1.4 Characterizations…………………………………………………95

4.2.2 Results and Discussion…………………………………………………….96

4.2.3 Conclusions……………………………………………………………….100

4.3 DGEDP-ethyl Reinforced with Modified BC Mats………………………………...101

4.3.1 Experimental Sections……………………………………………………101

4.3.1.1 Materials………………………………………………………..101

4.3.1.2 Modification of BC Mats with TMSCL………………………...101

4.3.1.3 DGEDP-ethyl/BC Mats Composites Preparation………………102

4.3.1.4 Characterizations………………………………………………..103

4.3.2 Results and Discussion…………………………………………………...104

4.3.3 Conclusions……………………………………………………………….113

4.4 References…………………………………………………………………………..115

Chapter 5: Vitrimerization: A Novel Concept to Reprocess and Recycle Thermoset

Waste…………………………………………………………………………………...119

5.1 Introduction…………………………………………………………………………120

5.2 Concept of “Vitrimerization”……………………………………………………….122

5.3 Experimental Sections……………………………………………………………...124

5.3.1 Materials………………………………………………………………….124

5.3.2 The Vitrimerization Process………………………………………………124

5.4 Results and Discussion……………………………………………………………..125

5.5 Conclusions and Future Works……………………………………………………..127

5.6 References…………………………………………………………………………..129

Chapter 6: Conclusions and Future Works……………..…………………………..131

6.1 Conclusions…………………………………………………………………………132

6.2 Future Works……………………………………………………………………….133

Bibliography……………………………………………………………………...... ….134

List of Figures

Figure 1.1 U.S. Epoxy resin market volume by application, 2014 - 2024 (Kilo Tons)…….3

Figure 1.2 Evolution of the physical properties during the curing of typical epoxy resin as a function of the conversion of epoxy groups……………………………………………...5

Figure 1.3 Generalized time-temperature-transformation (TTT) cure diagram. A plot of the times to gelation and vitrification during isothermal cure vs. temperature delineates the regions of four distinct states of matter: liquid, gelled rubber, gelled glass, and ungelled glass…………………………………………………………………………………..……9

Figure 1.4 a) Rotational rheology at 25 °C of synthesized epoxy resins compared to the DGEBA. b) Viscosity as a function of ester length for the prepared resins with trend line for the diamond-shaped symbols and excluding the square symbol………………………………….15

Figure 1.5 a) Stress-strain curves of the cured DGEDP-ester epoxy resins compared to cured DGEBA b) Storage modulus (DMA) and Young’s modulus (tensile testing) of the cured materials……………………………………………………………………..…….16

Figure 1.6 a) Alpha transition temperature (peak of loss modulus) and glass transition temperature as a function of ester length, b) Tan(δ) as a function of temperature

………………………………………………………………………………………....17

Figure 1.7 TEM images of dried dispersion of cellulose nanocrystals derived from (a) tunicate, (b) bacterial (c) ramie and (d) sisal……………………………………..……….20

Figure 1.8 TEM images of cellulose nanofibers prepared by TEMPO-mediated oxidation of hardwood………………………………………………………………………….…..22

Figure 1.9 BC produced in static cultivation (a) and SEM image of BC (b)………..…….23

Figure 2.1 B75 Rotor blade manufactured by Siemens…………………………………..32

vii

Figure 2.2 (a) Rotational rheology at 25 °C of synthesized epoxy resins compared to the DGEBA. (b) Viscosity as a function of ester length for the prepared resins with trend line produced from diamond-shaped symbols and exludes the square symbol……………..…35

Figure 2.3 Average Newtonian viscosity as a function of percent GE in DGEDP-Pe mixtures at room temperature……………………………………………………………38

Figure 2.4 (a) Gel time as a function of the reactive diluent concentration at 80°C and (b) determining the time of frequency independence of tan δ for the 5 wt % GE composition………………………………………………………………………………39

Figure 2.5 (a) Storage modulus (25°C) and (b) peak of the loss modulus (alpha transition) being related to the glass transition temperature as a function of increasing glycidyl eugenol…………………………………………………………………………….…..…40

Figure2.6 (A) Schematic and photograph (B) illustrate the vacuum infusion process used herein……………………………………………………………………………………..43

Figure 2.7 Complex viscosity of epoxy resins and crosslinker as a function of time at 25 oC………………………………………………………………...………………………45

Figure 2.8 a) Storage modulus and b) Tan(δ) as a function of temperature for Hexion, DGEBA/15wt%GE and DGEDP/15wt%GE epoxy/glass fiber composites……………...47

Figure 2.9 Flexural stress-strain curves of the epoxy/glass fiber composites…………….49

Figure 2.10 Delamination resistance (R-curce) of the Hexion (A), DGEBA+GE (B), DGEDP-ethyl+GE (C), DGEDP-pentyl+GE (D)/Glass Fiber composites……………….50

Figure 2.11 SEM images of fracture surfaces for epoxy/glass fiber composites………....52

Figure 2.12 TGA curves for Hexion, DGEBA/GE and DGEDP/GE cured epoxy/glass fiber composites…………………………………………………………………..……53

Figure 3.1 1wt% modified CNC in Toluene, 1h sedimentation after 3min sonication...…71

viii

Figure 3.2 FTIR of non-modified and silane modified CNCs (a)CNC modified with MPS, (b)CNC modified with GPTMS, (c)CNC modified with PTMO, (d)CNC modified with APTMS…………………………………………………………………………….…….72

Figure 3.3 XPS spectra of CNC and APTMS modified CNC……………...…………….73

Figure 3.4 TGA of modified and unmodified CNC…………………………………...…74

Figure 3.5 TEM images (top) and XRD patterns (bottom) of CNC and APTMS modified CNC particles………………………………………………………………………...…..75

Figure 3.6 Images from optical microscopy under polarized light of biobased epoxy resin/CNC dispersions formed after mixing by ultrasonication, but before curing……….77

Figure 3.7 Storage modulus and complex viscosity as function of frequency of biobased epoxy resin/CNC dispersions formed after mixing by ultrasonication, but before curing…………………………………………………………………………………….79

Figure 3.8 DSC curves of cured CNC and APTMS modified CNC epoxy composites and Tg……………………………………………………………………………………...…81

Figure 3.9 Storage modulus and tan δ as a function of temperature of the epoxy composites. Unmodified CNCs are shown by dashed lines and APTMS modified CNCs are shown as solid lines………………………………………………………………………………...82

Figure 3.10 TGA of the epoxy composites. Unmodified CNCs are shown by dashed lines and APTMS modified CNCs are shown as solid lines……………………………………84

Figure 4.1 Biosynthesized BC mat water gel (left), freeze-dried BC mat (middle), morphology under SEM (right)…………………………………………………….….…94

Figure 4.2 Storage modulus versus temperature for the neat epoxy and the epoxy/BC composites……………………………………………………………………………….96

Figure 4.3 tan δ versus temperature for the neat epoxy and the epoxy/BC composites…………………………………………………………………………...…97

Figure 4.4 TGA of the neat epoxy and the epoxy/BC composites………………………98

Figure 4.5 FTIR of the neat epoxy and the epoxy/BC composites………………….……99

Figure 4.6 EDS results for BC (top) and modified BC (bottom)………………………104

Figure 4.7 XRD curves of the BC and modified BC…………………………...……….105

Figure 4.8 Tensile properties of neat DGEDP-ethyl, BC film and DGEDP-ethyl/BC composites with different BC volume fractions……………………………………...….107

Figure 4.9 Young`s modulus of the composites as a function of BC volume fraction….108

Figure 4.10 Storage modulus as function of temperature for DGEDP-ethyl, BC film and the composites………………………………………………………………………..…110

Figure 4.11 tan δ as the function of temperature of DGEDP-ethyl/BC composites…..…111

Figure 4.12 DSC curves of DGEDP-ethyl/BC composites……………………..…..…..111

Figure 4.13 TGA curves of DGEDP-ethyl, BC and the composites……………...……..112

Figure 5.1 Volume change of the cured epoxy during swelling in the solvent…………..125

Figure 5.2 Vitrimerized epoxy and original epoxy processed with extrusion and hot-press processing………………………………………………………………………………126

Figure 5.3 Shear stress relaxation experiments: normalized relaxation modulus as a function of time for vitrmerized epoxy at different temperatures………………………..127

List of Scheme

Scheme 1.1 Synthesis of DGEBA…………………………………………………...…….3

Scheme 1.2 Typical Epoxy–Amine Reactions ……………………………………...…….4

Scheme 1.3 Schematic representation of thermoset curing (from a to d) in step polymerization for the reaction between difunctional and trifunctional molecules………..7

Scheme 1.4 Epoxidation of allylated (a) Salicylic Acid and (b) 4-Hydroxybenzoic Acid………………………………………………………………………………………13

Scheme 1.5 Two-step synthetic pathway to synthesize n-alkyl diphenolate diglycidyl ethers that differ in the chain length of the n-alkanol ester moiety. The diglycidyl ethers of alkyl diphenolates (DGEDP-methyl ester; DGEDP-ethyl ester; DGEDP-n-butyl ester; DGEDP-n- pentyl ester) were prepared……………………………………….………….14

Scheme 1.6 Molecular structure of glucose, cellobiose and cellulose…………...……….18

Scheme 1.7 Microstructure of cellulose where crystalline and non-crystalline regions are shown…………………………………………………………………………………….19

Scheme 1.8 Oxidation of primary hydroxyls of cellulose to carboxylate groups by TEMPO/NaClO/NaClO2 system…………………………………………………...……21

Scheme 1.9 The arrangement of fibrils extruded from cell surface and their self-assembly process…………………………………………………………………………………....23

Scheme 2.1 Synthesis of glycidyl ether of eugenol………………………………………37

Scheme 3.1 Molecular structure of DGEDP-ethyl monomer…………………………….63

Scheme 3.2 Molecular structure of selected silanes used in this work……………………66

Scheme 3.3 Organosilanization of CNC…………………………………………………70

Scheme 4.1 Two-step preparation of epoxy/BC composites……………………..………95

Scheme 4.2 Acylation reactions of cellulose with anhydrides……………………....…..100

Scheme 4.3 Scheme of the BC surface modification…………………………………....102

Scheme 5.1 DGEBA/Fatty acid crosslinking reaction………………………………….123

xii

List of Tables

Table 1.1 Properties of graphene and carbon nanotubes summarized from reported literature…………………………………………………………………………….……11

Table 1.2 Glass transition temperature and tensile properties of biobased epoxy resin compare with DGEBA resin…………………………………………………….………..13

Table 1.3 Comparison of viscosities, experimental and theoretical epoxide equivalent weight of DGEDP………………………………………………………………………..15

Table 2.1 Comparison of storage modulus at 25oC and glass transition temperature (Tg) of epoxy/glass fiber composites…………………………………………………..….…..48

Table 2.2 Flexural properties of epoxy/glass fiber composites…………………………..49

Table 3.1 Ratios of elements for CNC and silane modified CNC from XPS……………..73

Table 3.2 Tg from DMA and DSC in addition to the storage modulus at 25 oC and 160 oC for the unmodified and APTMS-modified CNC/epoxy composites……………...………83

Table 4.1 Mechical and thermal properties of the DGEDP-ethyl, BC film and the composites. ET is the modulus obtained from tensile test, E30 and E180 are the storage moduli obtained from DMA at 30 oC and 180 oC, respectively. Tdmax is determined from TGA………………………………………………………………………...…………113

xiii

Acknowledgement

I would like to express my sincerest appreciation to my dear advisor, Professor Ica

Manas-Zloczower. Without your guidance, this thesis could not be possible. I feel lucky to be in the group for the past five years. Even I still don`t like the winter here, I enjoy my research and life in such a lovely group. I always appreciate the chance you give me to study here. Thank you for everything.

I would like to thank my committee members Prof. David Schiraldi, Prof. Joao

Maia and Prof. Donald Feke. It`s my great honor to have you in my defense committee.

I also would like to express my gratitude to those many talented people who helped me a lot during my research. Prof. Richard Gross and Dr. Anthony Maiorana in Rensselaer

Polytechnic Institue, whose did the synthesis of the DGEDP resin and the reactive diluent as well as the many helpful discussion. Prof. Philippe Dubois, Dr. Jean-Marie Raquez and

Dr. Farid Khelifa in University of Mons help me conduct my research in Belgium.

I would like to thank all the lovely members of our “university distinguished group”.

It`s my great pleasure to work with you here. I will always remember the enjoyable time we have together in and out of the lab. And I would also thank my friends in Kent Hale

Smith building. I wish you all could have a bright and prosperous future.

Last but not least, I would like to thank my family for offering me all the support, love and encouragement.

xiv

Biobased Epoxy Composites: Sustainable Alternative for Advanced Materials

Abstract

LIANG YUE

This thesis focuses on exploring the potential of applying biobased epoxy resins as a more sustainable alternative for the currently used petroleum-derived epoxy. This work demonstrates the feasibility of replacing petroleum based epoxy resins for vacuum infusion of fiberglass mats with a new biobased formulation with 85wt% of diglycidyl ethers of either ethyl or pentyl diphenolate esters (DGEDP) mixed with 15wt% glycidyl ether of eugenol (GE). This new formulation shows suitably viscosity and gelation time for composite vacuum infusion processing while maintaining and in some cases exceeding the mechanical and thermal properties of the petroleum based systems. Furthermore, biobased nanocomposites reinforced with amine modified cellulose nanocrystals displaying increased thermomechanical properties could further extend the application of biobased epoxy resins. Moreover, a bacterial based porous cellulose nanofiber network was applied as reinforcement for the biobased epoxy matrix. These biobased composites with high mechanical performance but low density were fabricated in a two-step conventional method by impregnation of the bacterial cellulose network with the resin mixture, subsequent hot pressing and curing. This research work demonstrates the potential of biobased epoxy resins as well as their composites as a sustainable choice in the future.

Additionally, another aspect of sustainability was addressed in this thesis by developing a method to reprocess and recycle thermoset materials based on the exchangeable transesterification reactions, characteristic of vitrimer chemistry. The epoxy permanent network becomes a dynamic network and could be processed again with conventional processing techniques like compression molding, extrusion and injection molding. The developed methodology promises recycling and reprocessing of waste thermoset materials.

CHAPTER 1

Introduction

1.1 Thermoset Composites

1.1.1 Thermoset Resin

Thermosets are polymers formed as chemically cross-linked networks. Different from thermoplastic polymers, which could melt and flow above a certain temperature, once the three-dimensional solid network is formed by chemical reaction of cure, thermosets polymers cannot melt and flow anymore. From the processing viewpoint, thermoplastics could be easily processed and reprocessed again above their melt temperature to shape into the application form with conventional processing methods like extrusion, hot press, etc., while thermoset polymers are shaped into their final application form during the curing process. Typical thermoset polymers include epoxy, polyurethane, acrylate, phenolic, unsaturated polyester and so on. Uncured thermoset resins are liquid mixtures of monomers. When reacting with the curing agent, the monomers start cross-linking resulting in an increase in viscosity and eventually when fully cross-linked become solid.

Because of the covalently bonded network, cured thermoset polymers generally possess excellent mechanical properties, dimensional stability, thermal stability and chemical resistance.1

Epoxy resin is the most widely used thermoset polymers. Approximately 70% of the market of thermosetting resins is epoxy resin.2 The term “epoxy” can be referred to a variety of molecules containing at least two epoxide groups. They have a broad range of applications from coatings, adhesives to matrix of structural composites. And the market volume is growing as shown in Figure 1.1.

Figure 1.1 U.S. Epoxy resin market volume by application, 2014 - 2024 (Kilo Tons)3

Diglycidyl Ether of Bisphenol A (DGEBA) is the most commonly used epoxy monomer.

The traditional route for synthesis DGEBA is using Bisphenol A (BPA) and epichlorohydrin in the presence of sodium hydroxide as shown in Scheme 1.1.4

Scheme 1.1 Synthesis of DGEBA

A variety of reactants can be used as the curing agent (also known as hardener) for epoxy.

For example, amines, acids, disulfides, anhydrides and amides represent the majority of

3 commercially available hardeners. Among them, the DGEBA-diamine systems are the most commonly used curing systems. A ring-opening addition reaction in the DGEBA- diamine system is shown in scheme 1.2. The kinetics of the curing is determined by the nucleophilic character of amine and the electrophilicity of the DGEBA monomer. The primary and secondary amines will react with the epoxide groups forming hydroxyl groups first.5 The hydroxyl groups also have a chance to react with the epoxide groups to form ether at a higher temperature.6

Scheme 1.2 Typical Epoxy–Amine Reactions7

Unlike the processing of thermoplastics, the thermoset curing involves chemical reactions between monomers and crosslinkers. During the curing of thermosets, the liquid state resin will gradually become solid, and there is a critical transition at some point which is

4 called the gel point. The viscosity of the resin during curing increases with conversion as illustrated in Figure 1.2. When reaching the gel point, the viscosity of the system becomes infinity and start showing a solid behavior with the appearance of elastic modulus. After the gel point, the system can no longer flow and loses its processability.

The time from the curing start to the time when the system gelled can be considered as the processing time window for the thermoset system. After the gelation, the elastic modulus of the system increases with conversion until a complete network is formed.

Figure 1.2 Evolution of the physical properties during the curing of typical epoxy resin as

a function of the conversion of epoxy groups.8

From a molecular level, as illustrated in Scheme 1.3, the curing reaction happens with nearby monomers to form branched bigger molecules (Scheme 1.3b). The molecular weight increases as the curing degree increases, resulting in increased viscosity. When

5 those larger molecules eventually become crosslinked and reach the gel point, the system forms an incompletely crosslinked network (Scheme 1.3c). At the end, when the system is fully cured, the molecular weight goes to infinity (Scheme 1.3d).

As discussed above, after the gel point the system transits from a liquid state to a solid state, and can no longer be reshaped again. When processing a thermoset product, the final part has to be shaped before gel point. This time is very important in many industrial processes. For example, resin transfer molding (RTM) is the conventional processing method to fabricate large engineering structural parts like wind turbine blades. During the processing, the premixed thermoset resin (containing hardener) will infuse into the fiber mats under vacuum. It is important to make sure that before the infusion is completed, the system is not gelled.

Scheme 1.3 Schematic representation of thermoset curing (from a to d) in step

polymerization for the reaction between difunctional and trifunctional molecules.

Glass transition (also called vitrification), is another important transition for thermoset materials when the system is transformed into a glassy state. It should be noticed that this transition could be either from the liquid state or the gel state into the glass state. During curing, vitrification could happen before or after gelation. It is a consequence of the reduction in polymer chain mobility. Unlike gelation, which totally depends on the conversion in the curing reaction independent of temperature, the vitrification is

7 temperature dependent. The temperature for glass transition (Tg) increases as the conversion increases. So the Tg at the fully cured state (where the conversion reaches the maximum) will be the maximum Tg for the system. In the glassy state, the reaction becomes diffusion controlled, and the chemical reaction rate will significantly decrease.

Gelation is irreversible, while the vitrification is always reversible by changing the temperature. During the processing of a thermoset system, the curing temperature is always slightly higher than the Tg to achieve maximum reaction rate.

Figure 1.3 illustrates the isothermal time-temperature-transformation (TTT) cure diagram for a typical thermoset polymer like epoxy. The TTT cure diagram was introduced by

Enns and Gillham,9 to conveniently show transitions during curing, including gelation, vitrification and degradation. The times to gelation and vitrification as functions of curing temperature are plotted. Three critical temperatures are marked on the diagram: Tg0

(glass transition temperature before the system gelled), gel Tg (temperature of gelation and vitrification coincide) and Tg∞ (glass transition temperature of the fully cured system). Four possible phases during the curing are liquid, rubber, gelled glass and degraded polymer. This diagram could help define many parameters when processing with thermoset materials. Below Tg0 the reaction in the glassy state will be very slow.

Therefore below Tg0 could be ideal storage temperature for premixed resin system.

Between Tg0 and gel Tg, the resin system could undergo curing but remains in a liquid state for a longtime, which could help define the processing window for large part thermoset based composites, like wind turbine blades. Tg∞ is the glass transition temperature for the fully cured system. For high-temperature curing systems, they

8 undergo pre-curing at a temperature below Tg∞, and then a post-curing at slightly higher temperatures than Tg∞ to attain maximum curing degree and also avoid thermal degradation. A fully cured resin could achieve the maximum Tg, better mechanical properties, higher heat resistance and better chemical resistance.10

Figure 1.3 Generalized time-temperature-transformation (TTT) cure diagram. A plot of

the times to gelation and vitrification during isothermal cure vs. temperature delineates

the regions of four distinct states of matter: liquid, gelled rubber, gelled glass, and

ungelled glass.9

Besides thermally cured thermosets, there are also photocuring based resins. Photo-curing resins usually require less curing time.11,12 Those type of resins are mostly used in

9 biomedical applications and electron devices, since they can be cured in seconds. Also, special attention should be paid to a new kind of thermosets, the vitrimers, developed by

Leibler, which are totally different from the traditionally crosslinked thermoset resins.13

Vitrimers are crosslinked thermosets, but the crosslinks are based on exchangeable chemical bonds. Unlike the traditional thermoset network with permanent architecture, the vitrimer network is dynamic. Therefore, vitrimers could be reprocessed and recycled after being fully cured. About this kind of thermosets, there will be a more detailed discussion in Chapter 5.

1.1.2 Reinforcement Materials for Epoxy Composites

Polymer composites have been applied in many structural and nonstructural applications.

Comparing with metals, they are much more lightweight and have better corrosion resistance. There is an increasing number of high-performance polymer composites applied in different engineering field such as aerospace, automotive, construction and so on. As the most widely used matrix in polymer composites, the neat epoxy systems are rigid and brittle. To improve the toughness and strength of epoxy systems, various reinforcement materials have been used. Thermal and electrical conductive reinforcements are also widely employed to fabricate epoxy composites for functional materials.

The reinforcing effect of fillers in epoxy composites is complicated. A number of parameters such as mechanical properties, the size and aspect ratio of the filler, distribution of the filler in the matrix, filler matrix interfacial properties will affect the final mechanical properties of the composites. Mechanical properties are the primary

10 focus of this thesis. Reinforcement can be roughly categorized into three groups based on size/shape: particles, fibers and laminates or mats.

Among the many types of particle reinforcements, carbon nanofillers are widely used in epoxy composites as carbon nanotubes, carbon nanostructures, graphene and their derivatives.14,15,16 Table 1.1 summarizessome reported data of graphene and carbon nanotubes. Extensive research has been done on polymer nanocomposites using graphene and carbon nanotubes. Due to their high aspect ratio, percolation threshold for graphene and carbon nanotubes are usually very low. However there is always a dispersion issue when fabricating polymer nanocomposites. Van der Waals forces and strong π–π interactions in carbon nanofillers cause agglomeration in the polymer matrix.17 Many approaches have been developed to achieve good dispersion in a polymer matrix like surface modification or using dispersion agents.18

Table 1.1 Properties of graphene and carbon nanotubes summarized from reported literature.19 Young`s Tensile Electrical Thermal

modulus strength conductivity conductivity Graphene ~1 TPa 130±10 GPa ~3000 S/m ~5x103 W/mK

Carbon nanotubes ＞ 1 TPa 60~150 GPa ~7200 S/m 3500 W/mK

Although polymer nanocomposites could achieve high performance in mechanical properties, the issues of dispersion during processing and the relatively high cost of the carbon nanofillers limit their applications in large structural parts. Currently, the most commonly used polymer composites for the structural parts are the fiber reinforced polymer composites. More detailed introduction is given in Chapter 2.

1.2 Biobased Thermoset Composites

1.2.1 Biobased Epoxy Resin

The traditional epoxy resin is derived from petroleum product Bisphenol A (BPA).

However, in recent years the decreasing petroleum reserves and increasing environmental problems have concurred toward more sustainable development. Additionally, the potential health impact of using Bisphenol A, which is believed to increase the risk of cancer, has received a lot of attention.20,21 Therefore, the demand for alternative epoxy monomers from more sustainable and BPA free sources increased tremendously. There is an increasing interest in academia and industry in recent years to investigate the synthesis of epoxy monomers from biobased feedstock.

Many reports on the use of various renewable biobased resources to synthesize epoxy monomers have been published. Some raw materials used include plant oils, saccharides, polyphenols, lignin and lignin derivatives.22 Some of those reported biobased epoxy resin systems achieve comparable mechanical properties with the DGEBA resin.

For example, recently Robertson et al. reported epoxy resins derived from plant-based phenolic acids, which showed comparable mechanical properties with the DGEBA resin.23 After an allylation step of salicylic acid (SA) and 4-hydroxybenzoic acid (4HBA), the epoxidation is shown in Scheme 1.4. The yield is 70% to SA and 48% to 4HBA.

Scheme 1.4 Epoxidation of allylated (a) Salicylic Acid and (b) 4-Hydroxybenzoic Acid

The biobased phenolic acids derived epoxy resins exhibit comparable glass transition temperature, tensile strength and modulus with the petroleum-based DGEBA resin as shown in Table 1.2. However, these authors do not report on rheological properties, directly related to the processability of those types of biobased epoxy resins.

Table 1.2 Glass transition temperature and tensile properties of biobased epoxy resin compare with DGEBA resin.

Another example, Richard Gross et al. reported on a series biobased epoxy resin systems synthesized from esters of diphenolic acid. Attention will be focused on these type of biobased epoxy systems since they will be used in this research.

The diphenolic acid has a similar molecular structure with BPA, which may result in similar physical properties. The presence of an extra carboxylic acid group could allow

13 chemical modification to tune the properties further. The biobased epoxy resins based on diphenolic acid are synthesized as shown in Scheme 1.5. By variation of the ester side chain group (methyl, ethyl, butyl and pentyl), the properties of those diglycidyl ethers of alkyl diphenolates (DGEDP) based resins vary. The yields of the synthesized epoxy resins are from 85-97%.

DGEDP-n- pentyl ester) were prepared.24

These DGEDP based resins exhibit Newtonian behavior under rotational shearing (Figure

1.4), with viscosity values dependent on the ester chain length. Table 1.3 compares the viscosity of these DGEDP resins and DGEBA resin. With increasing ester chain length, the viscosity decreases. This is expected since longer side chain length increases the flexibility of the monomers. DGEDP-methyl has a very high viscosity at room temperature (792Pa.s). Such high viscosity is definitely not suitable for composite processing, but it may be ideal for adhesion applications. By increasing the ester chain

14 length to DGEDP-pentyl, the viscosity dramatically decreases (12Pa.s), close to DGEBA resin (4 Pa.s).

Figure 1.4: a) Rotational rheology at 25 °C of synthesized epoxy resins compared to the

DGEBA. b) Viscosity as a function of ester length for the prepared resins with trend line for the diamond-shaped symbols and excluding the square symbol.24

Table 1.3 Comparison of viscosities, experimental and theoretical epoxide equivalent weight of DGEDP.24

The synthesized DGEDP resins are cured with stoichiometric amounts of isophorone diamine. The mechanical properties of these resins are compared with DGEBA resin cured under similar conditions. Tensile test results are shown in Figure 1.5. There are very small differences in tensile strength and Young's modulus between DGEDP resins and DGEBA resins. DGEDP-methyl and DGEDP-ethyl exhibit even higher modulus than the DGEBA resin.

Figure 1.5: a) Stress-strain curves of the cured DGEDP-ester epoxy resins compared to cured DGEBA b) Storage modulus (DMA) and Young’s modulus (tensile testing) of the cured materials.24

The glass transition temperatures of those DGEDP resins show a trend of decreasing as the ester chain length increases. The DGEDP-ethyl has the highest glass transition temperature, which is comparable with the DGEBA resin. For many applications, higher glass transition temperature is preferred, but from the processing standpoint, lower viscosity is preferred, especially for vacuum infusion processing. These biobased

DGEDP resins could be chosen for different applications like adhesives, coatings or matrices for composites, based on processability and working conditions.

Figure 1.6 a) Alpha transition temperature (peak of loss modulus) and glass transition temperature as a function of ester length, b) Tan(δ) as a function of temperature.24

The DGEDP epoxy resins from renewable resources exhibit the potential to replace the petroleum-based DGEBA resin in the future. However, the disadvantages related to viscosity and glass transition temperature should also be further addressed.

1.2.2 Biobased Fillers

Cellulose is the most abundant renewable biopolymer available on earth. It widely exists in plants, but also in some marine animals like tunicates as well as in algae, fungi, bacteria and so on.25 The abundance of availability makes cellulose very low cost, and it`s also renewable and biodegradable. The abundance of surface hydroxyl groups, easily modified by chemicals,makes cellulose an ideal biobased raw material. Regardless of its

17 source, cellulose is a high molecular weight polymer made up of glucose units. The repeat segment in cellulose is known as cellobiose, which is a dimer of glucose linked by

β-(l-4)-glycosidic bonds as shown in Scheme 1.6.26

Scheme 1.6 Molecular structure of glucose, cellobiose and cellulose.26

In nature, due to the chain aggregation through hydrogen bonding, cellulose always exists as a crystalline structure. The crystalline state of cellulose is never complete. The presence of amorphous portion in cellulose varies from source to source but is often significant.27 Scheme 1.7 shows the microstructure of cellulose aggregation. The crystalline region - the cellulose nanocrystals, has attracted tremendous interest in academia due to its extraordinary properties.

Scheme 1.7 Microstructure of cellulose where crystalline and non-crystalline regions are shown.28

Much research has been done to study the isolation of cellulose nanocrystals. They can be extracted by mechanical or chemical treatments on various cellulose sources. Based on the treatment and the cellulose source, the resulting cellulose nanocrystals may have different crystallinity and aspect ratio. Chemical treatments are generally used in academic research to isolate the cellulose nanocrystals.

The most used approach to prepare the cellulose nanocrystals is based on acid hydrolysis.

The amorphous regions of cellulose are hydrolyzed in acid. The crystalline regions have more resistance to acid and preserve the structure. Sulfuric and hydrochloric acids are generally used in this processing. Raw cellulose materials like cotton, wood pulp, ramie fibers are dispersed in acid media under strictly controlled conditions including temperature, agitation and time.29 After hydrolysis, the resulting suspension is washed with excess water and dialyzed against distilled water to remove the acid completely. The rod-like nanocrystals are produced and could be stably dispersed in water. Figure 1.7 shows the dimensions of cellulose nanocrystals hydrolyzed from different sources.

Figure 1.7 TEM images of dried dispersion of cellulose nanocrystals derived from (a) tunicate, (b) bacterial (c) ramie and (d) sisal.26

Another well-established approach is TEMPO-mediated oxidation. This method is generally used for cellulose nanofiber preparation. Scheme 1.8 shows the oxidation reaction of primary alcohol groups of cellulose catalyzed by 2,2,6,6- tetramethylpiperidine-1-oxyl radical (TEMPO)-mediated system. After the oxidation reaction, significant amounts of carboxylate groups form. With a followed mechanical disintegration treatment step, those anionically-charged cellulose nanocrystals are easily isolated due to electrostatic repulsion and osmotic effects.30

Scheme 1.8 Oxidation of primary hydroxyls of cellulose to carboxylate groups by

31 TEMPO/NaClO/NaClO2 system.

Figure 1.8 shows the dimensions of hardwood cellulose nanofibers prepared by TEMPO- mediated oxidation. The advantage of this approach is that preparation is under moderate aqueous conditions and enables isolation of completely individualized cellulose nanofibers with at least a few microns’ length.32 Unlike the acid hydrolysis which can cause depolymerization of cellulose and results in low weight recovery ratios

(30~50%)33 ,this approach could achieve an yield as high as 90%.32 Cellulose nanofibers have much higher aspect ratio ( ＞ 100) for TEMPO-mediated oxidized hardwood cellulose.

Figure 1.8 TEM images of cellulose nanofibers prepared by TEMPO-mediated oxidation of hardwood.30

Besides chemical treatment of native cellulose from plants, another promising approach to obtain cellulose nanocrystals is biosynthesis from bacteria. Unlike the plant source native cellulose, bacteria based cellulose (BC) is produced with a very high degree of crystallinity and purity. BC is formed as continuous cellulose bundles by a self-assembly process. The process is illustrated in Scheme 1.9. The high crystallinity of BC fiber is up to 84-89%.34 The high crystallinity results in extraordinary mechanical properties.

Besides the high crystallinity, the continuous 3D network formed by the BC nanofibers as shown in Figure 1.9 is also very good for polymer composites fabrication.

Scheme 1.9 The arrangement of fibrils extruded from cell surface and their self- assembly process.35

Figure 1.9 BC produced in static cultivation (a) and SEM image of BC (b).35

Cellulose nanocrystals (or nanofibers) as discussed above are ideal biobased reinforcements for polymer nanocomposites due to their high surface area,

23 biodegradability, low density and low-cost. The highly crystallized structure accounts for their very good mechanical properties. The theoretical Young`s modulus for perfect crystalized cellulose is estimated to be 167.5 GPa, which is stronger than steel.36

Experimentally measured elastic moduli of cellulose nanocrystals from tunicate and cotton are reported to be 143GPa and 105GPa respectively.37,38 A large number of studies using cellulose nanocrystals as filler for polymer nanocomposites have been reported.29

1.3 Thesis Scope

This thesis aims to explore a sustainable replacement for traditional petroleum-based epoxy composites. Projects in this thesis are funded by NSF PIRE-RENEW grant. It is an international research program led by Case Western Reserve University and Rensselaer

Polytechnic Institute and focuses on using biobased materials to create sustainable replacements and improve material performance in high value-added, high-performance applicants for clean energy technologies, including solar and wind power.39

Chapter one gives an overview on traditional epoxy resin and epoxy composites as well as the biobased epoxy resins and reinforcements for composites. The synthesis and properties of DGEBA based biobased resins are discussed in detail. State of the art research progress in applying cellulose nanocrystals as reinforcement is also summarized.

Chapter two reports on the processability and mechanical properties of biobased epoxy resin systems formulated with biobased reactive diluent. This is first time reporting on biobased epoxy resin systems that have comparable initial viscosity and gelation time as well as mechanical properties with the commercial petroleum-derived resins for vacuum infusion processing.

Chapter three reports on using surface modified cellulose nanocrystals to improve the glass transition temperature for the DGEDP-ethyl resin. As the most suitable resin among the DGEDP type biobased resins for vacuum infusion processing, the DGEDP-ethyl formulated system exhibits comparable overall performance with the commercial product.

The only drawback is the slightly lower glass transition temperature. The strategy in this work was to use cellulose nanocrystals to prepare a nanocomposite with DGEDP-ethyl to achieve higher glass transition temperature.

Chapter four reports on using a porous bacterial cellulose nanofiber mat as reinforcement for DGEDP-ethyl resin. The highly crystalized bacterial cellulose has very low density and high mechanical properties. This work explores the potential of using this type of biobased mats to replace or partially replace the traditionally used reinforcement glass fiber mats in structural composites. The fabricated composites exhibit significant improvement in mechanical properties.

Chapter five reports on a new approach to reprocess and recycle thermoset materials based on the exchangeable transesterification reactions. A simple swelling procedure to infuse the catalyst into the thermoset network changes the permanent network into a dynamic vitrimer network, which is then processable and recyclable using conventional techniques.

The research presented in this thesis could shed some light on the potential application of biobased epoxy composites as a sustainable choice in the future. Also, the approach developed to reprocess and recycle crosslinked thermoset materials may help better deal with the thermoset waste in the future.

1.4 References

(1) Epoxy Resins: Chemistry and Technology, Second Edi.; May, C. A., Ed.

(2) Pham, H. Q.; Marks, M. J. Epoxy Resins. In Kirk-Othmer Encyclopedia of

Chemical Technology; John Wiley & Sons, Inc., 2000.

(3) grand view research. Epoxy Resin Market Size & Share, Industry Report, 2024.

(4) Dowd, R. T. General Chemistry of bisphenol A-based epoxy resins; Bruins, P. F.,

Ed.; John Wiley and Sons, 1968.

(5) Verchere, D.; Sautereau, H.; Pascault, J. P.; Riccardi, C. C.; Moschiar, S. M.;

Williams, R. J. J. Buildup of epoxycycloaliphatic amine networks. Kinetics,

vitrification, and gelation. Macromolecules 1990, 23 (3), 725–731.

(6) Mijovic, J.; Wijaya, J. Etherification reaction in epoxy-amine systems at high

temperature. Polymer (Guildf). 1994, 35 (12), 2683–2686.

(7) Patel, A.; Maiorana, A.; Yue, L.; Gross, R. A.; Manas-Zloczower, I. Curing

Kinetics of Biobased Epoxies for Tailored Applications. Macromolecules 2016, 49

(15), 5315–5324.

(8) H. H. Winter, et al. Techniques in Rheological Measurement; Collyer, A. ., Ed.;

Chapman and Hall: London, 1997.

(9) Enns, J. B.; Gillham, J. K. Time–temperature–transformation (TTT) cure diagram:

Modeling the cure behavior of thermosets. J. Appl. Polym. Sci. 1983, 28 (8), 2567–

2591.

(10) Jin, F.-L.; Park, S.-J. Thermal properties of epoxy resin/filler hybrid composites.

Polym. Degrad. Stab. 2012, 97 (11), 2148–2153.

(11) Morselli, D.; Bondioli, F.; Sangermano, M.; Messori, M. Photo-cured epoxy

networks reinforced with TiO2 in-situ generated by means of non-hydrolytic sol–

gel process. Polymer (Guildf). 2012, 53 (2), 283–290.

(12) Foix, D.; Ramis, X.; Serra, A.; Sangermano, M. UV generation of a

multifunctional hyperbranched thermal crosslinker to cure epoxy resins. Polymer

(Guildf). 2011, 52 (15), 3269–3276.

(13) Solid, L. D.; Hrubesh, L. W.; Chan, H. M.; Grenestedt, J. L.; Harmer, M. P.;

Caram, H. S.; Roy, S. K.; Handbook, P. T.; Raton, B.; Ashby, M. F.; et al. Silica-

Like Malleable Materials from. Science (80-. ). 2011, 334 (November), 965–968.

(14) Kim, J. A.; Seong, D. G.; Kang, T. J.; Youn, J. R. Effects of surface modification

on rheological and mechanical properties of CNT/epoxy composites. Carbon N. Y.

2006, 44 (10), 1898–1905.

(15) Naebe, M.; Wang, J.; Amini, A.; Khayyam, H.; Hameed, N.; Li, L. H.; Chen, Y.;

Fox, B. Mechanical Property and Structure of Covalent Functionalised

Graphene/Epoxy Nanocomposites. Sci. Rep. 2014, 4, 4375.

(16) Yue, L.; Pircheraghi, G.; Monemian, S. A.; Manas-Zloczower, I. Epoxy

composites with carbon nanotubes and graphene nanoplatelets – Dispersion and

synergy effects. Carbon N. Y. 2014, 78 (Supplement C), 268–278.

(17) Sengupta, R.; Bhattacharya, M.; Bandyopadhyay, S.; Bhowmick, A. K. A review

on the mechanical and electrical properties of graphite and modified graphite

reinforced polymer composites. Prog. Polym. Sci. 2011, 36 (5), 638–670.

(18) Loos, M. R.; Yang, J.; Feke, D. L.; Manas-Zloczower, I. Effect of block-

copolymer dispersants on properties of carbon nanotube/epoxy systems. Compos.

Sci. Technol. 2012, 72 (4), 482–488.

(19) Yue, L. Epoxy composites with hybrid carbon fillers, Case Western Reserve

University, 2014.

(20) Maffini, M. V; Rubin, B. S.; Sonnenschein, C.; Soto, A. M. Endocrine disruptors

and reproductive health: the case of bisphenol-A. Mol. Cell. Endocrinol. 2006,

254–255, 179–186.

(21) Chen, M.-Y.; Ike, M.; Fujita, M. Acute toxicity, mutagenicity, and estrogenicity of

bisphenol-A and other bisphenols. Environ. Toxicol. 2002, 17 (1), 80–86.

(22) Auvergne, R.; Caillol, S.; David, G.; Boutevin, B.; Pascault, J.-P. Biobased

Thermosetting Epoxy: Present and Future. Chem. Rev. 2014, 114 (2), 1082–1115.

(23) Yang, G.; Rohde, B. J.; Tesefay, H.; Robertson, M. L. Biorenewable Epoxy Resins

Derived from Plant-Based Phenolic Acids. ACS Sustain. Chem. Eng. 2016,

acssuschemeng.6b01343.

(24) Maiorana, A.; Spinella, S.; Gross, R. A. Bio-based alternative to the diglycidyl

ether of bisphenol A with controlled materials properties. Biomacromolecules

2015, 16 (3), 1021–1031.

(25) Moon, R. J.; Martini, A.; Nairn, J.; Simonsen, J.; Youngblood, J. Cellulose

nanomaterials review: structure{,} properties and nanocomposites. Chem. Soc. Rev.

2011, 40 (7), 3941–3994.

(26) Habibi, Y.; Lucia, L. A.; Rojas, O. J. Cellulose Nanocrystals: Chemistry, Self-

Assembly, and Applications. Chem. Rev. 2010, 110 (6), 3479–3500.

(27) Eichhorn, S. J. Cellulose nanowhiskers: promising materials for advanced

applications. Soft Matter 2011, 7 (2), 303–315.

(28) Westman, M. B. and G. Cellulose - Fundamental Aspects and Current Trends.

(29) Azizi Samir, M. A. S.; Alloin, F.; Dufresne, A. Review of Recent Research into

Cellulosic Whiskers, Their Properties and Their Application in Nanocomposite

Field. Biomacromolecules 2005, 6 (2), 612–626.

(30) Saito, T.; Kimura, S.; Nishiyama, Y.; Isogai, A. Cellulose Nanofibers Prepared by

TEMPO-Mediated Oxidation of Native Cellulose. Biomacromolecules 2007, 8 (8),

2485–2491.

(31) da Silva Perez, D.; Montanari, S.; Vignon, M. R. TEMPO-Mediated Oxidation of

Cellulose III. Biomacromolecules 2003, 4 (5), 1417–1425.

(32) Isogai, A.; Saito, T.; Fukuzumi, H. TEMPO-oxidized cellulose nanofibers.

Nanoscale 2011, 3 (1), 71–85.

(33) DONG, X. U. E. M. I. N.; REVOL, J.-F.; GRAY, D. G. Effect of microcrystallite

preparation conditions on the formation of colloid crystals of cellulose. Cellulose

1998, 5 (1), 19–32.

(34) Czaja, W.; Romanovicz, D.; Brown, R. malcolm. Structural investigations of

microbial cellulose produced in stationary and agitated culture. Cellulose 2004, 11

(3), 403–411.

(35) Huang, Y.; Zhu, C.; Yang, J.; Nie, Y.; Chen, C.; Sun, D. Recent advances in

bacterial cellulose. Cellulose 2014, 21 (1), 1–30.

(36) Tashiro, K.; Kobayashi, M. Theoretical evaluation of three-dimensional elastic

constants of native and regenerated celluloses: role of hydrogen bonds. Polymer

(Guildf). 1991, 32 (8), 1516–1526.

(37) Šturcová, A.; Davies, G. R.; Eichhorn, S. J. Elastic Modulus and Stress-Transfer

Properties of Tunicate Cellulose Whiskers. Biomacromolecules 2005, 6 (2), 1055–

1061.

(38) Rusli, R.; Eichhorn, S. J. Determination of the stiffness of cellulose nanowhiskers

and the fiber-matrix interface in a nanocomposite using Raman spectroscopy. Appl.

Phys. Lett. 2008, 93 (3), 33111.

(39) http://pire-renew.com.

CHAPTER 2

Biobased Epoxy Resin for Vacuum Infusion Processing

2.1 Introduction

2.1.1 Vacuum Infusion Processing of Fiber Reinforced Epoxy Composites

Vacuum infusion processing is the most commonly used manufacturing technique in the industry for fabricating fiber reinforced polymer composites. Vacuum infusion processing is ideal for efficient fabrication of large piece composites parts such as wind turbine blades.

Figure 2.1 is an example of a wind turbine blade (Siemens B75) fabricated with vacuum infusion processing by Siemens. It`s the longest wind turbine blade in the world with a length of 75metres. Such a large structural part is actually cast in one piece with just glass fiber mats and epoxy resin without any glue or joints to do the assembly. The vacuum infusion processing is processed under negative pressure, so the fiber mat can be highly compacted under pressure to reach very high fiber to resin ratio. It is also a very cost- effective method for manufacturing fiber reinforced thermoset composites with minimum resin waste.

Figure 2.1 B75 Rotor blade manufactured by Siemens.1

In the vacuum infusion processing, continuous fiber mats are layup on the mold of final shape and then sealed with a vacuum bag. The liquid resin is infused under vacuum into

32 the interspaces of the fiber mats. Once the entire piece is totally infused, the fiber-resin mixture is cured at a certain temperature under pressure. Fiber mats are the major component in the composites providing the ultra-high mechanical strength. The commonly commercial fiber products used for vacuum infusion processing are glass fibers, carbon fibers and Kevlar fibers. Among them, glass fiber mats are the most widely used reinforcement because of low cost, high tensile strength, high chemical resistance and excellent insulating properties. However, carbon fibers are considered the leading constituent reinforcement for high-performance applications. Carbon fibers have much better mechanical properties than glass fibers, especially the modulus could be as high as

1000GPa (E-glass fibers modulus is around 72GPa). Also, comparing the relative high density of glass fiber mats (E-glass fibers from Corning have a density of 2.54g/cc), the carbon fibers have lower density (T300 1.7g/cc). Lightweight is always an important consideration in applications for the aerospace industry. The disadvantage of carbon fibers is the price, limiting their application. Epoxy resin is the most commonly used polymer matrix for fiber reinforced composites. There are many commercial resin systems especially designed for vacuum infusion processing. The main requirements for processing are low initial viscosity and long enough gelation time.

2.1.2 Biobased Epoxy Resin

The need for polymer composites with low density and high strength is increasing across many industries such as aerospace, automotive and wind energy. In aerospace and automotive applications, the demand for lightweight materials is increasing since weight

23 reduction equates to lower fuel consumption and CO2 emissions . In wind energy,

33 polymer composites are the principle materials that enable production of longer and larger blades for transforming wind into electricity45. Wind energy is considered by many as a top candidate alternative energy source 67. Thus, in the foreseeable future, demand for high performance polymer composites is expected to grow.

Thermoset materials are usually used as matrix materials for fabrication of high performance lightweight polymer composites 891011. They generally have a crosslinked three-dimensional network, which leads to strong mechanical properties and high glass

12,13 transition temperatures (Tg) . Epoxy resins are the most prevalent of thermoset materials since, when properly designed, they can be processed at room temperature, provide good wettability to reinforcing fillers, high Tg, long lifetime and outstanding chemical resistance

14,15. Epoxy resins are used for a wide range of applications that include fiber-reinforced composite materials, adhesives, and coatings. The most common epoxy resin is the diglycidyl ether of bisphenol A (DGEBA) and oligomers thereof. DGEBA based epoxy resins are usually liquids with varied viscosity at room temperature, which enables their processing at low temperatures 1617.

However, the use of bisphenol A (BPA) is under review by the FDA for its potential to disrupt the human endocrine system, cause cancer, effect on the brain and more.18,19. BPA is also derived from petroleum and, therefore, is unsustainable since petroleum is a finite resource 20.

There is increased demand by the plastic industry to replace petroleum-derived feedstocks by readily renewable biobased feedstocks. Biobased epoxy resins with comparable properties to DGEBA are known that, if commercialized, would reduce BPA use. For example, starting with rosin, a DGEBA-free epoxy resin with comparable properties to

DGEBA resin was synthesized21. Also, epoxy resins from biobased polyols prepared by esterification with ethyl-4-hydroxybenzoate were reported with reduced modulus and Tg values relative to DGEBA 22. However, both of these are solids at room temperature, requiring a high processing temperature, which limits their applicability.

Recently, some of us reported the synthesis and characterization of a bio-based alternative to DGEBA.23 The diglycidyl ether of n-alkyl diphenolate esters (DGEDP) provides similar mechanical properties to those of DGEBA when cured with stoichiometric amounts of isophorone diamine. Furthermore, increasing the n-alkyl ester length leads to decreased epoxy resin viscosity and Tg of the final cured materials. As shown in Figure 2.2, with the increasing ester chain length, the viscosity changes from 792 Pa.s (DGEDP-methyl ester) to 12 Pa.s (DGEDP-n-pentyl ester). However, even for relatively longer n-alkyl esters, the epoxy resin viscosity (12 Pa.s, DGEDP-n-pentyl) is too high for vacuum infusion molding processes of continuous fibers.

Figure 2.2. Viscosity as a function of ester length for the prepared resins with trend line produced

from diamond-shaped symbols and exludes the square symbol.24

2.1.3 Biobased Reactive Diluent

Development of epoxy resins for infusion processing has been of great importance for the industry. Large pieces of engineering structural composites fabricated with high fiber content such as wind turbine blades require epoxy resins with very low viscosity.

Commercial resin products for the infusion application are usually based on the mixture of

DGEBA with low viscosity reactive diluents such as monofunctional aliphatic glycidyl esters and aliphatic or aromatic glycidyl ethers.2526 In addition to decreasing the viscosity of the system, the reactive diluent should also extend the gelation time of the resin to ensure long enough processing time to finish the infusion. Besides, the reactive diluent should also react with the hardener but be nonreactive with the resin under storage conditions. Many biobased epoxy resin systems have been reported, but there is little discussion about the processibility of these resins. Lack of information on processing may limit their commercialization.

A new kind of biobased reactive diluent has been recently developed by Maiorana and colleagues.24 It is a monofunctional glycidyl ether reactive diluent synthesized from eugenol, which is extracted from plant-based essential oils such as clove and nutmeg.2728

The biobased reactive diluent glycidyl ether of eugenol (GE) is synthesized in a single step processing as shown in Scheme 2.1. with an yield ranging from 85% to 90%.

Scheme 2.1 Synthesis of glycidyl ether of eugenol24

The main target use of GE is to decrease the viscosity of the biobased DGEDP epoxy resin.

Since the monofunctional glycidyl ether will react with the diamine hardener, the crosslinking density, as well as the physical properties of the cured resin, will possibly be different from the neat resin. To study how the reactive diluent will impact the resin properties, the GE was formulated with DGEDP-Pentyl at different ratios. DGEDP-Pentyl was chosen due to its lowest viscosity (12 Pa.s around) among the DGEDP based epoxy resins. From the processing viewpoint, it may be the most suitable resin for vacuum infusion processing. Figure 2.3 shows the viscosity of the formulated resin as function of

GE loading (5, 10, 15, 20, 30 and 50 wt% to DGEDP-Pentyl resin). GE could dramatically decrease the viscosity of DGEDP-Pentyl resin. 5 wt% GE could reduce the viscosity by

38%, 10 wt% could reduce it by 70% whereas 20 wt% will reduce the viscosity by 100% to about 1 Pa.s. The initial viscosities of these formulated resin systems are comparable with the commercial infusion resin system.

Figure 2.3 Average Newtonian viscosity as a function of percent GE in DGEDP-Pe mixtures at

room temperature.24

Gelation time is another important parameter for infusion processing. The hardener used is isophorone diamine and the stoichiometric ratio is based on the epoxied equivalent weights from both DGEDP-pentyl resin and the reactive diluent. Figure 2.4 shows the chemorheology study for the formulated resin systems with different GE loading. To speed up gelation tests were conducted at 80oC. The gel point is determined as the time at which

Tan(delta) is independent of frequency as shown in Figure 2.4b.

Figure 2.4 (a) Gel time as a function of the reactive diluent concentration at 80°C and (b)

determining the time of frequency independence of tan δ for the 5 wt % GE composition.24

Figure 2.4(a) shows the gel time is extended significantly as the GE loading increases allowing for better processability. Processability of the biobased DGEDP resin is improved by decreasing the system initial viscosity and extending the gelation time. The effect of GE on the mechanical and thermal properties of the cured formulated resin systems are studied with dynamic mechanical analysis. The storage modulus of formulated resin systems with different GE loadings at 25 oC is shown in Figure 2.5(a). Up to 30 wt% GE, the storage modulus remains similar (between 2 and 3 GPa).

Figure 2.5 (a) Storage modulus (25°C) and (b) peak of the loss modulus (alpha transition) being

related to the glass transition temperature as a function of increasing glycidyl eugenol.

The glass transition temperatures in Figure 2.5(b) are taken from the peak of the loss modulus correlated to the alpha transition temperature T α, which is closely related to the glass transition temperature Tg. The formulated resin systems exhibit a clear trend of decreasing Tg with increasing GE loading. As shown in Figure 2.5(b), the relation of Tg as the function of GE loading is nearly linear (slope=-1.36, R2=0.995). Unlike the storage modulus, the Tg of formulated resin systems is highly sensitive to the GE loading. Even

o with just 5 wt% GE, the Tg decreases by 6 C.

This work demonstrates the processability of bio-based epoxy resin formula in a composite system. Specifically, GE was used as the reactive diluent with DGEDP-ethyl and pentyl esters for an infusion system consisting of plain-woven glass fiber mats. Two biobased resin systems utilizing 15 wt% GE (DGEDP-ethyl and DGEDP-pentyl) and two benchmark systems (DGEBA with 15 wt% GE and a Hexion infusion resin) were studied.

A stoichiometric amount of isophorone diamine was used as a crosslinker for each resin system and the processing window was determined through chemorheology. The resins

40 were then utilized in a vacuum infusion molding process with plain woven glass fiber mats to fabricate a glass fiber/epoxy resin composite. Physical properties of the composites were evaluated by three point bend, fatigue life, and fracture toughness. Furthermore, the thermomechanical properties were assessed by dynamic mechanical analysis (DMA), differential scanning calorimetry (DSC), and thermogravimetric analysis (TGA). The interface between the different matrices and fillers was examined by scanning electron microscopy (SEM).

2.2 Experimental Section

2.2.1 Materials

Bio-based diglycidyl ethers of diphenolates (DGEDP) with ethyl (DGEDP-ethyl) and pentyl (DGEDP-pentyl) esters were synthesized from diphenolic acid according to a previous report23. The glycidyl ether of eugenol (GE) was also synthesized according to a previous report24. Hexion Inc. supplied a commercialized epoxy resin (L135i). Lab grade

Bisphenol A diglycidyl ether (DGEBA) and isophorone diamine (IPDA) were purchased from Sigma-Aldrich. Bi-directional E-glass fiber mats were purchased from Fiberglast.

2.2.2 Preparation of Composites

Epoxy Resin Preparation: Values of epoxide equivalent weights (EEW) for neat Hexion and DGEBA were those reported in the corresponding material data sheets. EEW’s of the neat biobased resin and reactive diluent were determined by titration according to ASTM

D1652 and the experimental values were previously reported23. Weights of reaction components for the formulated resin systems were determined using Eq. (1). Amine

41 hydrogen equivalent weights (AHEW) of the hardener were determined using Eq. (2). The parts by weight of hardener per hundred parts resin (phr) were determined using Eq. (3).

푇표푡푎푙 푊푒𝑖푔ℎ푡 퐸퐸푊 표푓 푀𝑖푥 = 푊푒𝑖푔ℎ푡 표푓 푅푒푠𝑖푛 퐴 푊푒𝑖푔ℎ푡 표푓 푅푒푠𝑖푛 퐵 (1) + 퐸퐸푊 표푓 퐴 퐸퐸푊 표푓 푅푒푠𝑖푛 퐵

푀표푙푒푐푢푙푎푟 푤푒𝑖푔ℎ푡 표푓 푎푚𝑖푛푒 퐴퐻퐸푊 = (2) 푛푢푚푏푒푟 표푓 푎푐푡𝑖푣푒 ℎ푦푑푟표푔푒푛푠

퐴퐻퐸푊 푥 100 푝ℎ푟 = (3) 퐸퐸푊

To have a comparable low resin viscosity as commercial Hexion resin that contains a reactive diluent and other additives, a biobased reactive diluent was formulated with the

DGEDP-ester epoxy resins and, as a reference, DGEBA epoxy resin. First, 15 wt% GE was mixed with two DGEDP-ester resins and DGEBA at 60 °C, respectively. After cooling to room temperature, formulated epoxy resin systems were mixed with IPDA according to the calculated phr. The prepared resins were degassed by centrifugation (2000 rpm for 1 minute).

Vacuum Infusion Processing: The epoxy/E-glass fiber mat composites were prepared by vacuum infusion processing. Twelve layers of E-glass fiber mats (20cm x 20cm) was covered with peel ply and sealed in a vacuum bag; the prepared epoxy resin was then infused into the mats under vacuum as shown in Figure 2.6. After infusion (average time for Hexion and DGEBA/GE were around 30 mins, DGEDP-ethyl/GE and DGEDP- pentyl/GE were around 40 mins), the composite was placed on a hot press under 2 tons pressure, cured at 80 oC for 4 hours and post cured at 160 oC for another 4 hours. Two spacer elements with a height of 3.5 mm were placed between the pressing plates to adjust

42 the thickness of the sample and control the fiber content (~75 wt%, for 12 layers E-glass fiber mats).

Figure2.6 (A) Schematic and photograph (B) illustrate the vacuum infusion process used herein.

2.2.3 Characterizations

Dynamic Mechanical Analysis (DMA): TA Instruments Q800, operating in dual cantilever mode with an amplitude of 10 μm, was used to determine the storage modulus

(Eʹ) and glass transition temperature (Tg, determined from the peak of the loss modulus) by scanning at 2 °C/min from 0 °C to 200 °C. All samples were run in triplicate.

Chemorheology: The viscosity of the epoxy resins was determined by rotational rheology at 25 °C using parallel plate geometry with 25 mm stainless steel plates. Shear rate sweep experiments from 1.25 - 1250 s-1 were conducted under steady state flow with an equilibration time of 1 min and a gap of 1 mm. Viscosity was reported as an average from the Newtonian region.

Bending Test: The test was conducted following ASTM D790 using an Instron 1011 universal tensile machine with a 10 kN capacity load cell and standard three point bending fixtures at a test speed of 2mm/min. Five specimens were tested for each sample.

Thermogravimetric Analysis (TGA): The thermal stability of prepared resins was studied by TGA using a TA Instruments Q500 with an aluminum pan. The samples were about 10 mg each and were run from room temperature to 800 °C at a heating rate of 10

°C per minute under constant N2 flow.

Scanning Electron Microscopy (SEM): SEM samples were prepared using specimens recovered from fracture toughness tests. Tests were conducted using SEM (JEOL JSM-

6510LV) at an operating voltage of 30 kV.

Mode-I Interlaminar Fracture Toughness Test: The Mode-I interlaminar fracture toughness (GIC) was measured in double cantilever beam (DCB) tests according to ASTM

D 5528 standard test method 29,30,31. Rectangular specimens (180 mm in length and 25 mm in width) were cut from the cured plane and a 50 mm initial crack was introduced by inserting a thin PTFE film in the mid-plane of the laminates during fabrication. Five specimens were tested at a crosshead speed of 5mm/min using a Zwick Universal testing machine. GIC is calculated by the modified beam theory using Eq. (4)

3푃훿 퐺 = (4) 1푐 2푏(푎+Δ) where: P is the critical load, δ is the load point displacement, b is the specimen width, a is the delamination length and Δ is determined experimentally by generating a least square plot of the cubic root of compliance C as a function of the crack length.

2.3 Results and Discussion

2.3.1 Chemorheology of Curing

Matrix viscosity is the most critical parameter for vacuum infusion processing with fiber mats 32. The processing window of an epoxy resin system depends on the resin gelation

44 time, which represents the ultimate pot life of a 2-part epoxy resin system and the arrest of flow 33. Low viscosity of the resin system results in good flow properties and fiber wetting

8,34. For fabrication of large, fiber reinforced composite parts like wind turbine blades, processing requires long infusion times to ensure a complete impregnation of the mats.

Figure 2.7 Complex viscosity of epoxy resins and crosslinker as a function of time at 25

Figure 2.7 displays the complex viscosity of different epoxy resins mixed with isophorone diamine at 25 oC as a function of time. The Hexion infusion epoxy resin system has the lowest initial viscosity (2.5 Pa.·s), whereas DGEDP-ethyl and DGEDP-pentyl have relatively higher initial viscosities (20 and 10 Pa.·s, respectively). Previous work demonstrated that addition of 15 wt% GE to the system results in significant viscosity reduction and gel time extension24. This reduction in viscosity and increase in gel time by addition of GE to DGEDP epoxy systems is a consequence of an extension in the molecular weight between crosslinks. Incorporating 15 wt% GE in DGEDP-ethyl results in both a

67.5% reduction in complex viscosity relative to neat DGEDP-ethyl(from ~20.4 Pa.s to

~6.6 Pa.s ) and a significant extension of gel time. Whereas neat DGEDP-ethyl resin undergoes a rapid rise in viscosity at room temperature within one hour after mixing with the hardener, addition of 15wt% GE to the system extended the gel time to more than 100 min, comparable to the Hexion infusion resin.

2.3.2 Dynamic Mechanical Analysis of Epoxy/E-Glass Fiber Mat composites

The thermomechanical properties of different epoxy/glass fiber composites were evaluated by dynamic mechanical analysis (DMA). The storage modulus and tan δ as a function of temperature are shown in Figures 2.8a and 2.8b, respectively, and the corresponding values are listed in Table 2.1. The storage modulus at 25 oC of the DGEDP-ethyl/15 wt% GE composite is 11% higher than that of the fabricated Hexion infusion resin. Furthermore, the storage moduli at 25 °C of both DGEDP-ethyl/GE and DGEDP-pentyl/GE composites are higher than that from DGEBA by 16.1 and 4.2%, respectively. According to a previous study, the ester side chain of DGEDP epoxy resins do not interrupt three dimensional network formation of the corresponding epoxy thermoset23. Since the EEW’s of

DGEDP/GE resins are higher than those of DGEBA and the Hexion infusion resins, it follows that DGEDP resins will form thermosets with relatively lower crosslink densities.

The question then is why do DGEDP/GE epoxy resin composites have higher modulus values than those prepared from Bisphenol A (BPA) based systems. This may be due to interactions of DGEDP/GE ester moieties with glass fibers rendering better interfacial adhesion between the matrix and mat.

Figure 2.8 a) Storage modulus and b) Tan(δ) as a function of temperature for Hexion,

DGEBA/15wt%GE and DGEDP/15wt%GE epoxy/glass fiber composites.

The dissipation factor tan δ is related to the matrix Tg. Composites produced with BPA based resins have higher Tg values than those from DGEDP/GE epoxy resins (Table 2.1).

The extent that increases in DGEDP ester chain length results in decreased cured epoxy

23 resin Tg is discussed elsewhere . Consequently, polymer chains of cured DGEDP based epoxy resins will more easily slide past each other when a force is applied at lower temperatures than for traditional BPA based systems. The Tg values of DEGDP-ethyl/GE and DGEGDP-pentyl/GE epoxy/glass fiber composites are 111 and 96 oC, respectively.

o Nevertheless, even 96 C is a sufficiently high Tg for most engineering applications, especially those where high working temperatures are not experienced.

o Table 2.1: Comparison of storage modulus at 25 C and glass transition temperature (Tg) of epoxy/glass fiber composites

DGEDP- DGEDP- Hexion DGEBA/GEa ethyl/GEa pentyl/GEa

Storage

modulus at 12.30 ± 0.03 11.80 ± 0.06 13.70 ± 0.04 12.30 ± 0.04

25oC (GPa)

Glass transition

temperature 138 140 111 96

(oC) a The GE content is 15wt%

2.3.3 Mechanical Properties

Figure 2.9 displays flexural stress-strain curves for epoxy/glass fiber composites with

DGEDP-ethyl/GE, DGEDP-pentyl/GE, Hexion and DGEBA/GE epoxy resins.

Corresponding values of flexural modulus (calculated from the linear region of the stress- strain curve), strength at break and strain at break are given in Table 2.2. The bio-based

DGEDP-ethyl/GE and DGEDP-pentyl/GE epoxy/glass fiber composites exhibit slightly higher flexural modulus and strength at break relative to those prepared from Hexion and

DGEBA/GE epoxy resins. Furthermore, the relative magnitudes of flexural modulus and strength are in the following order: DGEDP-pentylGE≈DGEDP- ethyl/GE >Hexion≈DGEBA/GE at 25 oC.

Figure 2.9 Flexural stress-strain curves of the epoxy/glass fiber composites.

Table 2.2: Flexural properties of epoxy/glass fiber composites

Sample Strength at Strain at break Modulus Mode-I break interlaminar (%) (GPa) fracture (MPa) toughness, GIC J/m2

Hexion 324.8 ± 23.3 2.5±0.2 14.76±0.32 524±23

DGEBA+GE 334.8 ± 18.1 2.7±0.3 14.18±0.41 503±19

DGEDP-ethyl 356.9 ±1 1.8 2.6±0.3 15.47±0.29 587±22 +GE

DGEDP-pentyl 350.0 ± 17.4 2.5±0.1 15.78±0.36 509±26 +GE

Figure 2.10 Delamination resistance (R-curce) of the Hexion (A), DGEBA+GE (B), DGEDP-ethyl+GE (C), DGEDP-pentyl+GE (D)/Glass Fiber composites.

For the DCB test, multiple crack propagation loading cycles were conducted and the results were used to determine Mode-I interlaminar fracture toughness. The critical load P and critical displacement δ recorded for each cycle of load-displacement curves were used in

Eq. (4) to determine GIC. These data were used to construct the delamination resistance plots (R-curve) for each specimen tested. These plots are shown for different resins in

Figure 2.10. The fracture toughness GIC values were calculated based on averages for the plateau region in the R-curve. The data reported in Table 2.2 are average values and standard error based on three tests conducted for each of the composite materials. During the test, mat debonding from the matrix leads to macro-cracks that propagate along the

50 interface until failure. Hence, it follows that the GIC of fiber reinforced polymer composites is governed by the adhesion strength at the fiber/matrix interface. 30,35,36,37,38

Structures of the glass fiber mats as well as the matrix to fiber ratio will also affect the composite fracture toughness but these variables are fixed for the epoxy/fiber glass composites studied herein. The GIC of DGEDP-ethyl/GE is slightly higher than both

Hexion and DGEBA/GE epoxy/glass fiber composites (by 12% and 17%, respectively).

Furthermore, GIC’s of DGEDP-pentyl/GE and DGEBA/GE are equivalent.

2.3.4 Morphology of Composites

Interfacial debonding will result in failure of stress transfer between the matrix and fiber.

Fracture surface SEM images in Figure 2.11 illustrate that all the tested epoxy/glass composite systems have substantial chemical affinity between fibers and matrices.

Furthermore, the SEM images provide no evidence for either fiber pull out from the matrix and delamination between fibers. The lack of debonding between the matrix and glass fibers for DGEDP-based systems further indicates good adhesion.

Figure 2.11 SEM images of fracture surfaces for epoxy/glass fiber composites.

2.3.5 Thermal Stability

The thermal stability of Hexion, DGEBA/GE, DGEDP-ethyl/GE and DGEDP-pentyl/GE epoxy/glass composite samples were studied by thermogravimetric analysis (TGA). TGA weight loss curves as a function of temperature in Figure 2.12 show a one-step degradation mechanism. To highlight differences in thermal stability, the onset of degradation at 5% weight loss (Td5%) was determined. The epoxy/glass composite with the highest Td5% was prepared using Hexion. Values of Td5% for DGEDP-ethyl/GE and DGEDP-pentyl/GE are

32 and 16 oC below that of the Hexion system. The small difference in thermal stability between DGEBA and DGEDP-ester epoxy/glass composites is consistent with a previous report by our team where the onset of degradation at 10% weight loss (Td10%) for cured

DGEDP-ester epoxy resins is only 15 oC below that of cured DGEBA23. Hence, the

52 thermostability of the biobased epoxy resins when processed into epoxy/glass composites has suitable thermal stability for most engineering applications.

Figure 2.12 TGA curves for Hexion, DGEBA/GE and DGEDP/GE cured epoxy/glass fiber composites.

2.4 Conclusions

The glycidyl ether of eugenol (GE) was used as reactive diluent since it effectively lowers the viscosity of DGEDP-ester epoxy resins to enable their processing by vacuum infusion for epoxy/E-glass fiber mat composites.24 As incorporation of GE in DGEDP-ester composites lowers its Tg, 15 wt% GE was selected as it is the minimum GE content needed to attain suitable viscosity reductions and gel time extension. Comparison of the thermomechanical properties of epoxy resin/glass fiber composites prepared using two biobased (DGEDP-ethyl/GE and DGEDP-pentyl/GE) and two benchmark fossil carbon based epoxy resin systems (DGEBA/GE and Hexion) was performed. The storage modulus at 25 oC of the DGEDP-ethyl/GE composite is 11% higher than that fabricated with the

Hexion infusion resin. While composites produced from Hexion and DGEBA/GE have

o higher Tg’s (138 and 140 C, respectively) than those of DGEDP-ethyl and DGEDP-pentyl

o (111 and 96 C, respectively) epoxy/glass fiber composites, the Tg’s of the DEGDP-ester composites are suitable for most engineering applications. Bio-based DGEDP-ethyl and

DGEDP-pentyl composites exhibit better flexural modulus than those from Hexion and

DGEBA/GE. Furthermore, DGEDP-ethyl/GE composites have higher interlaminar fracture toughness than Hexion and DGEBA composites. These results indicate that the

DGEDP-ethyl epoxy resin provides stronger adhesion at the fiber/matrix interface relative to the BPA epoxy resins studied herein. Future work will further interrogate the origin of the high adhesion between DGEDP-esters and glass fibers. Finally, TGA temperature scans of the biobased and BPA-based epoxy resins all show a one-step thermal decomposition process. The onset of decomposition at 5% weight loss for DGEDP-ethyl/GE and DGEDP- pentyl/GE are 32 and 16 oC below that of the Hexion system. However, the thermostability of the biobased epoxy resins when processed into epoxy/glass composites is still suitable for most engineering applications. We therefore conclude that biobased DGEDP-ester/GE epoxy resins have excellent potential for use as sustainable alternatives to BPA-based epoxy resins for engineering applications.

2.6 References:

1. https://www.siemens.com/press/pool/de/feature/2014/energy/2014-03-hull/fact-

sheet-b75-rotor-blade-e.pdf.

2. Song, Y. S., Youn, J. R. & Gutowski, T. G. Life cycle energy analysis of fiber-

reinforced composites. Compos. Part A Appl. Sci. Manuf. 40, 1257–1265 (2009).

3. Witik, R. A., Payet, J., Michaud, V., Ludwig, C. & Månson, J.-A. E. Assessing the

life cycle costs and environmental performance of lightweight materials in

automobile applications. Compos. Part A Appl. Sci. Manuf. 42, 1694–1709 (2011).

4. EWEA. Wind energy – The facts, a guide to the technology, economics and future

of wind power . (Earthscan Publications Ltd. Belgium , 2009).

5. Loos, M. R., Yang, J., Feke, D. L. & Manas-Zloczower, I. Effect of block-

copolymer dispersants on properties of carbon nanotube/epoxy systems. Compos.

Sci. Technol. 72, 482–488 (2012).

6. Islam, M. R., Mekhilef, S. & Saidur, R. Progress and recent trends of wind energy

technology. Renew. Sustain. Energy Rev. 21, 456–468 (2013).

7. Leung, D. Y. C. & Yang, Y. Wind energy development and its environmental

impact: A review. Renew. Sustain. Energy Rev. 16, 1031–1039 (2012).

8. Yue, L., Pircheraghi, G., Monemian, S. A. & Manas-Zloczower, I. Epoxy

composites with carbon nanotubes and graphene nanoplatelets – Dispersion and

synergy effects. Carbon N. Y. 78, 268–278 (2014).

9. Srivastava, S. K. & Singh, I. P. Hybrid epoxy nanocomposites: lightweight

materials for structural applications. Polym J 44, 334–339 (2012).

10. Xie, Y., Hill, C. A. S., Xiao, Z., Militz, H. & Mai, C. Silane coupling agents used

for natural fiber/polymer composites: A review. Compos. Part A Appl. Sci. Manuf.

41, 806–819 (2010).

11. Al-Saleh, M. H. & Sundararaj, U. Review of the mechanical properties of carbon

nanofiber/polymer composites. Compos. Part A Appl. Sci. Manuf. 42, 2126–2142

(2011).

12. Pilato, L. A. & Michno, M. J. Acute toxicity, mutagenicity, and estrogenicity of

bisphenol-A and other bisphenols. Adv. Compos. Mater.

13. Commarieu, B. et al. Ultrahigh Tg Epoxy Thermosets Based on Insertion

Polynorbornenes. Macromolecules 49, 920–925 (2016).

14. Auvergne, R., Caillol, S., David, G., Boutevin, B. & Pascault, J.-P. Biobased

Thermosetting Epoxy: Present and Future. Chem. Rev. 114, 1082–1115 (2014).

15. Tang, L.-C. et al. The effect of graphene dispersion on the mechanical properties

of graphene/epoxy composites. Carbon N. Y. 60, 16–27 (2013).

16. Raquez, J.-M., Deléglise, M., Lacrampe, M.-F. & Krawczak, P. Thermosetting

(bio)materials derived from renewable resources: A critical review. Prog. Polym.

Sci. 35, 487–509 (2010).

17. Pascault, J. P. & Williams, R. J. J. In Epoxy Polymers: New Materials and

Innovations . (Wiley VCH: Weinheim , 2010).

18. Chen, M.-Y., Ike, M. & Fujita, M. Acute toxicity, mutagenicity, and estrogenicity

of bisphenol-A and other bisphenols. Environ. Toxicol. 17, 80–86 (2002).

19. Maffini, M. V, Rubin, B. S., Sonnenschein, C. & Soto, A. M. Endocrine disruptors

and reproductive health: the case of bisphenol-A. Mol. Cell. Endocrinol. 254–255,

179–86 (2006).

20. Jenck, J. F., Agterberg, F. & Droescher, M. J. Products and processes for a

sustainable chemical industry: a review of achievements and prospects. Green

Chem. 6, 544–556 (2004).

21. Mantzaridis, C. et al. Rosin acid oligomers as precursors of DGEBA-free epoxy

resins. Green Chem. 15, 3091–3098 (2013).

22. Fourcade, D., Ritter, B. S., Walter, P., Schonfeld, R. & Mulhaupt, R. Renewable

resource-based epoxy resins derived from multifunctional poly(4-

hydroxybenzoates). Green Chem. 15, 910–918 (2013).

23. Maiorana, A., Spinella, S. & Gross, R. A. Bio-based alternative to the diglycidyl

ether of bisphenol A with controlled materials properties. Biomacromolecules 16,

1021–31 (2015).

24. Maiorana, A., Yue, L., Manas-Zloczower, I. & Gross, R. Structure–property

relationships of a bio-based reactive diluent in a bio-based epoxy resin. J. Appl.

Polym. Sci. n/a-n/a (2016). doi:10.1002/app.43635

25. Hwang, P. Y.; Alto, P.; City, R. U.S. Patent 3488404. (1973).

26. Hove, C. L. F. D.; Heymans, D.; Steinbrecher, C.; Kotlweska, A.; Sand, R. V. U.S.

Patent 2014/0316030 A1. (2014).

27. Lugemwa, F. N. Extraction of Betulin, Trimyristin, Eugenol and Carnosic Acid

Using Water-Organic Solvent Mixtures. Molecules 17, (2012).

28. Mehmet, A., Ertas, M., Nitz, S. & Kollmannsberger, H. Chemical composition and

content of essential oil from the bud of cultivated Turkish clove (Syzygium

aromaticum L.). BioResources 2, (2007).

29. Kostagiannakopoulou, C., Loutas, T. H., Sotiriadis, G., Markou, A. &

Kostopoulos, V. On the interlaminar fracture toughness of carbon fiber composites

enhanced with graphene nano-species. Compos. Sci. Technol. 118, 217–225

(2015).

30. Silva, H., Ferreira, J. A. M., Capela, C. & Richardson, M. O. W. Mixed Mode

interlayer fracture of glass fiber/nano-enhanced epoxy composites. Compos. Part

A Appl. Sci. Manuf. 64, 211–222 (2014).

31. Zhang, H., Liu, Y., Kuwata, M., Bilotti, E. & Peijs, T. Improved fracture

toughness and integrated damage sensing capability by spray coated CNTs on

carbon fibre prepreg. Compos. Part A Appl. Sci. Manuf. 70, 102–110 (2015).

32. Meier, R., Kahraman, I., Seyhan, A. T., Zaremba, S. & Drechsler, K. Evaluating

vibration assisted vacuum infusion processing of hexagonal boron nitride sheet

modified carbon fabric /epoxy composites in terms of interlaminar shear strength

and void content. Compos. Sci. Technol. 128, 94–103 (2016).

33. Laik, S., Galy, J., Gérard, J.-F., Monti, M. & Camino, G. Fire behaviour and

morphology of epoxy matrices designed for composite materials processed by

infusion. Polym. Degrad. Stab. 127, 44–55 (2016).

34. Yuexin, D., Zhaoyuan, T., Yan, Z. & Jing, S. Compression Responses of Preform

in Vacuum Infusion Process. Chinese J. Aeronaut. 21, 370–377 (2008).

35. Borrego, L. P., Costa, J. D. M., Ferreira, J. A. M. & Silva, H. Fatigue behaviour of

glass fibre reinforced epoxy composites enhanced with nanoparticles. Compos.

Part B Eng. 62, 65–72 (2014).

36. Flore, D. & Wegener, K. Modelling the mean stress effect on fatigue life of fibre

reinforced plastics. Int. J. Fatigue 82, 689–699 (2016).

37. Harris, B. Fatigue in composites science and technology of the fatigue response of

fibre-reinforced plastics . (Elsevier Science Ltd, Boca Raton , 2003).

38. Talreja, R. & Singh, C. V. Damage and failure of composite materials .

(Cambridge University Press, Cambridge, New York , 2012).

CHAPTER 3

Surface-modified Cellulose Nanocrystals for Biobased Epoxy Nanocomposites

3.1 Introduction

Epoxy resins are among the most versatile thermoset polymers available and are used in a wide range of applications. They represent the matrix of choice for high strength composites, corrosion resistant coatings, structural adhesives and electronic materials.

Wind turbine blades, printed circuit boards, and structural adhesives used in automobile assembly all utilize epoxy resins primarily due to their high glass transition temperature, elastic modulus, toughness, and adjustable viscosity.1

In the epoxy industry, Bisphenol A and Novolacs represent benchmark materials due to their ability to impart desired physical and thermal properties in cured resins. Bisphenol A has recently come under extensive criticism due to its potential to disrupt the endocrine system.2–4 Novolacs are manufactured from phenol/formaldehyde and can contain high amounts of Bisphenol F. Furthermore, both bisphenol A and novolacs are primarily derived from petroleum feedstocks and are ultimately unsustainable.

Recent advances in biobased epoxy resins have yielded an extensive array of sustainable platform chemicals that could provide a route to petroleum feedstock replacement. Most notable of the biobased epoxy resin platforms are epoxidized vegetable oils that are already used as ingredients for protective coatings.[5,6] Other platforms are epoxidized tree rosins, diglycidyl ethers and esters of furans, cashew nutshell liquid based epoxy resins, multifunctional glycidyl flavonoids, diglycidyl ferulates, and diglycidyl diphenolates7,8

One challenge in utilizing biobased resins is that their inherent structure can lead to decreases in final polymer properties such as glass transition temperature and modulus.9–13

Often, the performance of biobased resins is not benchmarked against current commercial systems that leads to a lack of information on their potential for commercialization .14

The biobased DGEDP epoxy resins as discussed in Chapter 2 demonstrate the potential to be alternative to the currently used petroleum-based epoxy resin for vacuum infusion processing. We have compared the formulated biobased epoxy resin systems (DGEDP- ethyl and DGEDP-pentyl with 15 wt% biobased reactive diluent GE) with the commercial resin system from Hexion. Both the initial viscosity and gelation time are comparable with the commercial resin system. Moreover, the mechanical properties of the glass fiber reinforced composites are even better than the commercial resin system, which we believe is due to the extra ester chain groups from the DGEDP resin resulting in better adhesion between the resin matrix and fiber surface by formation of hydrogen bonding. The drawback of the formulated biobased resin systems is that the glass transition temperatures are relatively lower than the DGEBA type commercial resin system. The commercial resin

o o system has a Tg of 138 C, whereas the Tg for DGEDP-ethyl/GE resin system is 111 C and for the DGEDP-pentyl/GE resin system is 96oC. Such temperatures are still high enough to be applied in most engineering applications, but we still would like to explore the potential of further improving the Tg. From our previous study,15 DGEDP-ethyl is the most suitable candidate resin among the DGEDP type biobased resins for vacuum infusion processing due to its relatively low viscosity and comparable mechanical performance.

Thus, it is chosen for further study in this work. The molecular structure of DGEDP-ethyl is shown in scheme 3.1. It has similar structure with DGEBA but has an extra ester side chain.

CH2CH3

Scheme 3.1 Molecular structure of DGEDP-ethyl monomer.

Recently, advances in high modulus nanofillers such as graphene, carbon nanotubes, and silica have shown promise in producing materials that exceed the thermal and physical properties of the neat thermoset matrix.16 However, the industrial adoption of these fillers is relatively difficult due to production costs, inherent dangers associated with their use, and difficulty in obtaining good dispersions due to surface chemistry.

Cellulose nanocrystals (CNCs), extracted from abundant cellulose, represent a nanofiller that is sustainable, biodegradable, and has excellent strength-to-weight ratio, and can be used as reinforcement for epoxy composites.17,18 In order to fully realize the potential performance of CNCs as nano-building blocks in polymeric systems, it is important to modify the hydroxyl rich CNC surface in order to achieve filler separation and individualization within hydrophobic matrices. Extensive attention has been given to CNC modification methodologies such as grafting hydrophobic polymers from and to their surface. Modification techniques often utilize reactions of CNC hydroxyl groups to perform esterification, silanization, and epoxide ring opening19–21

In the absence of suitable catalysts and reaction conditions, the nucleophilicity of CNC hydroxyl groups is low. Consequently, ring-opening of epoxide groups by CNC hydroxyl

63 moieties proceeds slowly. In contrast, primary aliphatic amines are much better nucleophiles for epoxy ring-opening reactions. Recently, Dubois and coworkers reported the silanization of CNC surfaces to introduce a wide variety of functional groups. This led to the improved dispersion of these modified CNCs in polylactide.22 Indeed, organosilanization chemistry offers a mild method that can be carried out under aqueous conditions without protection-deprotection chemistry, and that enables the introduction of primary amines to CNC surfaces.

We hypothesized that amine functionalization of CNC through the use of amino-silanes would allow for ring-opening of epoxy groups during their dispersion in epoxy resins. Such ring-opening reactions will result in CNC surface chemistry that is identical to the matrix and will lead to well-dispersed CNC-epoxy nanocomposites. This dispersion method would also avoid the use of organic solvents. Based on earlier work on CNC epoxy thermoset nanocomposites, we further hypothesized that good dispersion of modified

CNCs will lead to improvements in nano-reinforced composite thermomechanical properties at lower nanofiber loading than had previously been reported.23

Herein we report the preparation of amine modified CNC utilizing 3-

(aminopropyl)trimethoxysilane (APTMS). CNC surface modification was assessed through spectroscopic techniques including X-ray Photoelectron Spectroscopy (XPS) and

Fourier Transform Infrared (FTIR). Amine functionalized CNCs at various concentrations were dispersed in a biobased epoxy derived from diphenolic acid and the dispersion quality was assessed through rheological and microscopic techniques. The amine functionalized

CNC showed improved dispersion and better mechanical and thermal properties in the cured nanocomposites relative to unmodified CNC epoxy composites. To further support

64 the choice of APTMS, other silane-modifications with 3-(trimethoxysilyl)propyl methacrylate (MPS), diethoxy(3-glycidyloxypropyl)methylsilane (GPTMS) and trimethoxy(propyl)silane (PTMO) were performed following the same procedure and their dispersion performance was also studied.

3.2 Experimental Sections

3.2.1 Materials

Biobased epoxy resin DGEDP-ethyl (diglycidyl ether of diphenolate ethyl ester) was synthesized as previously reported.24 3-Aminopropyltrimethoxysilane (APTMS), 3-

(Trimethoxysilyl)propyl methacrylate (MPS), Diethoxy(3-glycidyloxypropyl) methylsilane (GPTMS) and Trimethoxy(propyl)silane (PTMO) and isophorone diamine (IPDA) were purchased from Sigma-Aldrich. Pure ramie fibers obtained from

Stucken Melchers GmbH & Co. (Germany). All chemicals were used as received.

3.2.2 Preparation of Cellulose Nanocrystals

CNCs were extracted from ramie fibers as described in a previous paper.23 In summary, 80 g purified ramie fibers were cut into small pieces and treated with 1 L of 4% NaOH solution at 80 oC for 2h to remove any residual hemicellulose or lignin. These fibers were then acid hydrolyzed in 800 mL sulfuric acid solution (65 wt %) at 55 °C for 30 min under continuous mechanical stirring. Then the suspension was thoroughly washed with deionized water until neutrality, dialyzed against deionized water for a few days and then filtered through a sintered glass to remove non-hydrolyzed fibers. Dry CNCs were recovered by freeze- drying.

3.2.3 Organosilanization of CNC

The objective of surface modification is to render the CNC more compatible or have covalent bonding with the biobased DGEDP-ethyl resin to develop fully biobased composites. Four silane with different functional groups were selected as shown in scheme

3.2. MPS and PTMO modified CNC would be more hydrophobic, and GPTMS and

APTMS modified CNC could participate in the crosslinking reaction in the epoxy diamine curing system.

Scheme 3.2 Molecular structure of selected silanes used in this work.

The procedure for all silane modification are based on the previous optimization study.22

As an example APTMS (100 mM) was dissolved in an ethanol/water mixture (80/20 w/w,

100 mL) and the solution pH was maintained at about 4 with a citrate buffer (10 mM). This solution was maintained at ambient temperature with magnetic stirring for 2 hours.25,26

CNCs (0.5 g) were then added to this solution and the mixture was maintained at ambient temperature with magnetic stirring for an additional 2 h. The APTMS grafted CNCs were washed with water by centrifugation and recovered as a solid residue after freeze-drying.

Subsequently, the APTMS grafted CNCs were maintained at 110 oC for 16 h under vacuum.

3.2.4 Preparation of CNC/epoxy composites

APTMS modified CNCs were dispersed in the biobased epoxy resin (DGEDP-ethyl) at concentrations of 1, 5 and 10 wt.% by ultrasonication (10 g batch, 2 min at 60 °C, 10s on/off pause mode with an amplitude of 45%). Then, equimolar ratio of IPDA hardener was added and mixed well. This mixture was centrifuged to remove trapped air and then transferred to Teflon coated stainless steel molds at room temperature till gelling occurs. The samples were further cured under compression molding at constant pressure (2 T) for 4 h at 80 °C and then another 4 h at 160 °C. Composites with unmodified CNCs and all the modified

CNCs were prepared following an identical method as described above.

3.2.5 Characterizations

Infrared Spectroscopy: Fourier transform infrared (FTIR) measurements were performed using a Bruker Tensor 17 spectrometer from Bio-RAD using a spectral width ranging from

500 to 4000 cm–1, a resolution of 4 cm–1 and an accumulation of 32 scans.

XRD: Analyses by XRD were performed using a powder diffractometer Siemens D 5000 using Cu Kα radiation at room temperature in the range of 2θ = 5–30° at a scanning rate of

2°/min.

TEM: TEM images of CNC particles were recorded using a Philips CM200 with an acceleration voltage of 20 kV.

XPS: XPS analysis was performed with an Axis Ultra spectrometer (Kratos Analytical).

Samples were irradiated with monochromated X-rays (Al Kα, 1486.6 eV). The X-ray source was used under standard conditions with an operating pressure of 10 -8 Torr.

Photoelectrons were analyzed from a selected area (700 μm by 300 μm) with a take-off- angle of 90°.

Optical Microscopy: After sonication, CNC/epoxy suspensions were observed between crossed polarizers with a Leica DMRXP optical microscope in transmission mode. Images were obtained using an ICT/P polarizer and a 360 analyzer with a neutral density filter and recorded with a QImaging QICAM digital camera.

Differential Scanning Calorimetry (DSC): DSC measurements were performed by using a DSC Q100 differential scanning calorimeter from TA Instruments with Tzero aluminum pans. Each sample was heated from 0 to 250 °C at a heating rate of 5 °C/min under constant nitrogen flow. The glass transition temperature, Tg, was taken as the midpoint of the drop in heat capacity during the heating cycle.

Thermogravimetric Analysis (TGA): TGA measurements were performed using a TGA

Q500 apparatus from TA Instruments with aluminum pans. Samples of about 10 mg were heated from room temperature up to 800 °C at 10 °C/min under nitrogen flow.

Dynamic Mechanical Analysis (DMA): Mechanical behaviors were analyzed under ambient atmosphere using a Q800 DMTA apparatus from TA Instruments in tensile mode.

The measurements were carried out at a constant frequency of 1 Hz, strain amplitude of

0.05% over a temperature range of 20–180 °C. Samples were prepared by cutting strips from the films with a width of 5 mm.

Rheology: Rheological properties of CNC/epoxy suspensions were obtained at 25 °C on a

TA Instruments ARES G2 rheometer with 25 mm parallel plate geometry and a 1 mm gap size. Frequency sweeps from 0.1 to 100 rad/s at strain level of 0.1% were conducted within the linear viscoelastic regime.

3.3 Results and Discussion

3.3.1 Organosilanization of CNC

Modification of CNCs by salinization is believed to occur as shown in the Scheme 3.3. In the presence of excess ethanol, and pH values about 4, silane hydrolysis occurs.26 Due to the large amount of free hydroxyl groups on CNC surfaces, silanols form stable hydrogen bonds with the CNCs. Annealing under vacuum at 110 °C promotes further condensation and formation of a well-defined polysiloxane layer onto the CNC surface.25,27,28

Scheme 3.3 Organosilanization of CNC.

After organosilanization, the surface properties of CNC are totally changed. Figure 3.1 shows the suspensions of the surface modified CNCs dispersed in toluene (at 1 wt %) with

3min sonication and after letting it stand for 1h. The CNCs modified with MPS, PTMO and GPTMS, could stably disperse in the solvent due to their hydrophobic surface characteristics. The unmodified CNCs and the APTMS modified CNCs did not form stable suspensions in toluene.

Figure 3.1 1wt% modified CNC in Toluene, 1h sedimentation after 3min sonication.

FTIR spectra of the different CNCs are shown in Figure 3.2. For unmodified CNC, the broad peak ranging from 3024 cm-1 to 3691 cm-1 is due to CNC free O-H stretching vibrations. For the MPS modified CNC, the presence of C=C double bond from the MPS derivative at 1640 cm-1 is clearly observed and marked on Figure 3.2a. For the GPTMS modified CNC, the epoxy ring C-O-C identical peak at 909 cm-1 and 846cm-1 from

GPTMS are clearly observed and marked on Figure 3.2b. Also, for the PTMO modified

CNC, C-H3 peak (2800~300 cm-1 with several distinct peaks) from PTMO are observed and marked on Figure 3.2c. The N-H in-plane-bend vibrational bands with both peaks at

1560 cm-1 and 1640 cm-1 in Figure 3.2d also indicate the presence of primary amine groups on the APTMS modified CNC (CNC-APTMS), when comparing with the only peak at

1640cm-1 for unmodified CNC which comes from the stretch of C=C bonding. 22,29 The

71 formation of Si-O-C and Si-O-Si by condensation of the hydroxyl groups were not easily observed, due to the strong cellulose C-O-C vibration (1050-1170 cm-1). 27

Figure 3.2 FTIR of non-modified and silane modified CNCs (a)CNC modified with MPS,

(b)CNC modified with GPTMS, (c)CNC modified with PTMO, (d)CNC modified with

APTMS.

The presence of functional groups is also confirmed by XPS analysis. For example, Figure

3.3 shows the presence of a N1s peak for the APTMS modified CNC, which was not observed for non-modified CNCs. Table 3.1 lists changes in the elements after modification for all CNCs and an increase of Si amount for the surface modified ones

72 supports grafting to CNCs. Also the surface O/C ratio of modified and non-modified CNCs is different. The decrease in the O/C ratio upon modification is due to the carbon-richer moieties on the filler surface.

Figure 3.3 XPS spectra of CNC and APTMS modified CNC.

Table 3.1 Ratios of elements for CNC and silane modified CNC from XPS.

C O Si N

CNC 55.7 44.3

CNC-APTMS 57.1 39.4 1.8 1.6

CNC-MPS 60.6 34.9 4.9

CNC-GPTMS 53.8 39 7.2

CNC-PTMO 56 35.9 8

Figure 3.4 shows the TGA results for all CNCs. The thermal stability of CNC increases by silane modification (for example for the APTMS modified CNC Td-onset increases from

287oC to 352oC) due to the formation of a dense polysiloxane layer on CNC surfaces which delays thermal decomposition. Also the differences on residues above 600 oC are due to the molecular weight of the different silanes. Since all modifications were performed at the same concentration (100mM), the amount of residue reflects the weight usage: MPS

24.835g, PTMO 16.43g, GTPMS 23.6g, APTMS 17.93g.

Figure 3.4 TGA of modified and unmodified CNC.

TEM images (Figure 3.5) display individual CNC nanoparticles for both modified and pristine CNC samples. Unmodified CNC shows typical dimensions for CNC extracted from ramie fibers (length ~250 nm, diameter ~10 nm) as previously reported. 22The average diameter of the APTMS modified CNC (~25 nm) is slightly higher than those of unmodified CNC. This morphological change of CNC-APTMS is attributed to the formation of a polysiloxane layer on CNC surface. Indeed, similar results have been reported for other silane modified CNCs.22

The impact of silanization on CNC crystallinity was evaluated by XRD. The diffraction profiles of unmodified and modified CNCs are shown in Figure 3.5 (bottom). The XRD curve of unmodified CNCs is typical of a cellulose type 1 pattern, with characteristic diffraction peaks at 15o and 22o on the 2θ scale. The XRD patterns show both crystalline and amorphous regions in the form of diffraction bands and a diffuse pattern. The XRD patterns for unmodified and CNC-APTMS are similar, indicating that the crystalline structure of the CNC is not affected by surface modification. Crystallinity degree calculated

30 from Segal equation (Ic=(I22.6-I18)/I22.6) are 78.3% for CNC and 78.1% for CNC-APTMS.

Figure 3.5 TEM images (top) and XRD patterns (bottom) of CNC and APTMS modified

CNC particles.

3.3.2 CNC/epoxy composites

The dispersion state of CNCs in the biobased epoxy resin after mixing by ultrasonication, but before curing, was evaluated with optical microscopy. The anisotropic CNC crystalline structure can be distinguished from the isotropic epoxy resin under polarized light in transmission mode.29 Therefore, observation of large birefringent domains would indicate

CNC aggregation. Figure 3.6 shows micrographs of suspensions after sonication at different concentrations. For unmodified CNC/epoxy suspensions at all concentrations, birefringent domains on the order of 100 microns are observed, indicating that unmodified

CNCs form large aggregates in the epoxy matrix. In contrast, MPS, PTMO and GPTMS modified CNCs all show improved dispersion comparing with the unmodified CNCs. And among them, MPS and PTMO modified CNCs show less aggregation at high concentration comparing with the GPTMS modified CNC. The optical micrographs indicate that the filler hydrophobic surface can improve CNC dispersion in the DGEDP-ethyl resin.

Moreover, APTMS modified CNCs show little aggregation with almost no birefringent domains in the micrographs at any of the concentrations studied. Since mixing by ultrasonication of epoxy resin and CNC-APTMS resulted in an exotherm that was evident by an increase in dispersion temperature, we believe that some fraction of APTMS- modified CNCs react with DEGDP-ethyl epoxy groups enabling better dispersion within the matrix. In contrast, unmodified CNCs having only hydroxyl groups which are not

76 sufficiently nucleophilic to react with DEGDP-ethyl will form aggregates within the matrix as observed by optical microscopy.

Figure 3.6 Images from optical microscopy under polarized light of biobased epoxy

resin/CNC dispersions formed after mixing by ultrasonication, but before curing.

The CNC dispersion state in biobased epoxy resin was also investigated by performing frequency sweep rheological measurements. The dependence of complex viscosity (η∗) and storage modulus (G’) on the oscillatory frequency (ω) are shown in Figure 3.7. For the neat biobased epoxy monomer, the complex viscosity remains constant through a wide range of frequencies, indicative of Newtonian behavior. For dispersions containing unmodified CNCs, the complex viscosity follows a similar trend as that of the neat epoxy resin with relatively small increases at 5 and 10 wt.% filler. However, the complex viscosity of resin mixtures with MPS, PTMO and GPTMS modified CNCs show shear thinning behavior at 10 wt% loading. Also at low frequencies, a plateau region in storage modulus is observed, which indicates the formation of filler network within the resin, due to improved dispersion.

Moreover, the APTMS modified CNC dispersions show a more significantly different behavior with pronounced shear thinning. The storage modulus for dispersions containing

5 and 10 wt% CNC-APTMS display solid-like behavior with a plateau region at low frequencies. This is indicative of filler network formation as a result of good dispersion. In the case of unmodified CNC dispersions, solid-like behavior is not observed at all concentrations, highlighting the absence of a filler network and a lack of good particle dispersion. The slight decrease in viscosity from around 49 to 47 Pa.s for the 1 wt % unmodified CNC sample is within the instrument error.

Figure 3.7 Storage modulus and complex viscosity as function of frequency of biobased epoxy resin/CNC dispersions formed after mixing by ultrasonication, but before curing.

The optical microscopy images and the rheology studies on the DGEDP-ethyl resin mixture with different modified CNCs show consistent results. The surface modification improves

CNC dispersion in the resin. Among all surface modifications, APTMS modified CNCs show the best results for dispersion quality. Even at 10 wt% filler concentration no significant aggregation could be observed. Also 5 wt% APTMS modified CNCs show network formation within the resin, while the hydrophobic modified CNCs (MPS, GPTMS and PTMO) show network formation only at 10 wt%. Consequently, the properties of the CNC-APTMS/epoxy composites were further investigated.

The glass transition temperature (Tg) of the biobased epoxy composites after curing was studied by DSC. Figure 3.8 displays first-heat DSC traces of neat biobased epoxy and epoxy/CNC nanocomposites with various CNC contents. No significant exothermic peaks were observed on heating which indicates that all samples were cured to the maximum degree of conversion before testing. The Tg value was determined by the midpoint of the drop observed in the heat flow versus temperature curve. Tg of the neat epoxy is 126.4°C, while that of nanocomposites increased with CNC loading. For unmodified CNC composites, small increases in Tg were observed for samples loaded with 1 and 5 wt.% of

CNC, while the Tg increased by 3.5°C at 10 wt.% CNC. The CNC-APTMS epoxy composites display relatively larger increases in Tg, especially at 5 and 10 wt.% (6.3 and

9.9 °C , respectively). The effect of nanofillers on epoxy resin Tg is attributed to a reduction in polymer chain mobility.16 It is believed that CNCs act as physical interlocking points in

80 the cured matrix restraining chain mobility.31 For unmodified CNCs, nanoparticle agglomeration limits this “restriction effect”. The ability of APTMS-modified CNC surface primary amine moieties to react with DEGDP-ethyl epoxy groups resulting in stronger interfacial interactions between the nanofiller and matrix creating additional barriers that restrict chain motion and cause Tg to shift to higher temperatures.

Figure 3.8 DSC curves of cured CNC and APTMS modified CNC epoxy composites and

The thermo-mechanical properties of the biobased epoxy nanocomposites were investigated by DMA and Figure 3.9 shows the storage modulus and alpha transition temperatures (tan δ, associated to Tg) of the nanocomposites. The neat resin and nanocomposites both show typical amorphous thermoset behavior: below Tg the storage modules decreases slightly with increasing temperature and above Tg the storage modulus rapidly drops. Both unmodified and APTMS modified CNC nanocomposites display higher storage modulus. Below Tg, the increase in storage modulus for the nanocomposites at the same filler loading was similar for the unmodified and APTMS modified CNCs.

However, above Tg, the storage modulus of CNC-APTMS nanocomposites was significantly higher than for the unmodified CNC nanocomposites. For example, at 160 oC, the storage modulus of 10wt% APTMS modified CNCs nanocomposite is 151.5 MPa, more than 7 times larger compared to the neat resin (19.5MPa), whereas, with 10wt% unmodified CNCs, the storage modulus was only 37.3MPa. The significant reinforcement of CNC-APTMS nanocomposites above Tg is attributed to the well dispersed rigid nanofiller in the matrix. Also, considering the presence of amine functional groups on modified CNCs, a potential reaction between filler and the biobased epoxy could also facilitate the reinforcement. Glass transition temperatures from tan δ curves show the same trend as that from DSC. A comparison of Tg values from DMA and DSC is shown in Table

3.2, along with the tensile storage modulus at 25 °C and 160 oC for all samples. These results further confirm the strong interactions between APTMS-CNCs and the biobased

DGEDP-ethyl resin.

Figure 3.9 Storage modulus and tan() as a function of temperature of the epoxy composites. Unmodified CNCs are shown by dashed lines and APTMS modified CNCs

are shown as solid lines.

o o Table 3.2 Tg from DMA and DSC in addition to the storage modulus at 25 C and 160 C for the unmodified and APTMS-modified CNC/epoxy composites.

Tg from Storage modulus at Storage modulus Tg from DMA Samples DSC 25oC at 160oC oC oC MPa MPa

Neat DGEDP-ethyl 129.2 126.4 2405±68 19.5±3.6

1 wt 129.4 127.5 2765±55 25.8±2.9

CNC-APTMS 5 wt 137.9 132.7 3265±63 44.4±5.5

10 wt 139.8 136.3 3492±72 151.5±37.4

1 wt 129 128.4 2786±64 23.5±3.2

Unmodified CNC 5 wt 133.2 126.8 2915±86 26.1±4.5

10 wt 135.2 129.9 3309±82 37.3±5.9

Thermal stability of the composites was studied with TGA as shown in Figure 3.10 and

Table 3.3. The thermal stability of the composites increased by comparison with the neat biobased epoxy. As the CNC content increases, the temperature at the maximum weight loss rate of the composites shifts to higher temperature. Similar results have been previously reported.29,32

Figure 3.10 TGA of the epoxy composites. Unmodified CNCs are shown by dashed lines

and APTMS modified CNCs are shown as solid lines.

Table 3.3 Decomposition temperature of the epoxy composites compare with neat

DGEDP-ethyl.

Samples Td 5% (oC) Td max (oC)

Neat DGEDP-ethyl 304.2 365.4

1 wt 311.8 365.1

CNC-APTMS 5 wt 322 381.9

10 wt 323.2 383.5

1 wt 311.6 365.7

Unmodified CNC 5 wt 313.2 365.7

10 wt 312.8 382.8

3.4 Conclusions

Free hydroxyl groups at the surface of CNCs derived from ramie fibers were modified by silanization with APTMS. XPS analysis provided definitive evidence that APTMS modification resulted in CNC fiber surfaces decorated with amines. Improved dispersion quality of CNC-APTMS relative to unmodified CNCs was evident from optical micrographs obtained under polarized light for the biobased epoxy resin/CNCs dispersions before curing. Furthermore, increases in Tg for 5 and 10% loaded CNC-APTMS nanocomposites were substantially higher than for the systems containing unmodified

CNCs. This provides further evidence that APTMS modification of CNCs enables physical interlocking points in the cured matrix restraining chain mobility. The improved dispersion of CNC-APTMS is attributed to the reaction of amino groups with DGEDP-ethyl epoxy groups increasing compatibility between filler and matrix components. In addition, rheology measurements demonstrate that the storage modulus for dispersions containing 5 and 10 wt% CNC-APTMS display solid-like behavior with a plateau region at low frequencies, indicative of filler network formation. Good dispersion of the CNC-APTMS as well as stronger filler-matrix interactions contributes to the significant improvement in storage modulus above Tg.

3.5 References

1. Auvergne, R., Caillol, S., David, G., Boutevin, B. & Pascault, J.-P. Biobased

Thermosetting Epoxy: Present and Future. Chem. Rev. 114, 1082–1115 (2014).

2. Maffini, M. V, Rubin, B. S., Sonnenschein, C. & Soto, A. M. Endocrine disruptors

and reproductive health: The case of bisphenol-A. Mol. Cell. Endocrinol. 254–255,

179–186 (2006).

3. Chen, M.-Y., Ike, M. & Fujita, M. Acute toxicity, mutagenicity, and estrogenicity

of bisphenol-A and other bisphenols. Environ. Toxicol. 17, 80–86 (2002).

4. O’Connor, J. C. & Chapin, R. E. Critical evaluation of observed adverse effects of

endocrine active substances on reproduction and development, the immune

system, and the nervous system. Pure Appl. Chem. 75, 2099–2123 (2003).

5. Stemmelen, M. et al. A Fully Biobased Epoxy Resin from Vegetable Oils: From

the Synthesis of the Precursors by Thiol-ene Reaction to the Study of the Final

Material. J. Polym. Sci. Part A Polym. Chem. 49, 2434 (2011).

6. Patel, A., Maiorana, A., Yue, L., Gross, R. A. & Manas-Zloczower, I. Curing

Kinetics of Biobased Epoxies for Tailored Applications. Macromolecules 49,

5315–5324 (2016).

7. Brocas, A.-L. et al. Epoxidized rosin acids as co-precursors for epoxy resins. Des.

Monomers Polym. 17, 301–310 (2014).

8. Sultania, M., Rai, J. S. P. & Srivastava, D. Process modeling, optimization and

analysis of esterification reaction of cashew nut shell liquid (CNSL)-derived epoxy

resin using response surface methodology. J. Hazard. Mater. 185, 1198–1204

(2011).

9. Maiorana, A., Yue, L., Manas-Zloczower, I. & Gross, R. Structure–property

relationships of a bio-based reactive diluent in a bio-based epoxy resin. J. Appl.

Polym. Sci. n/a-n/a (2016). doi:10.1002/app.43635

10. Kadam, A., Pawar, M., Yemul, O., Thamke, V. & Kodam, K. Biodegradable

biobased epoxy resin from karanja oil. Polymer (Guildf). 72, 82–92 (2015).

11. Hernandez, E. D., Bassett, A. W., Sadler, J. M., La Scala, J. J. & Stanzione, J. F.

Synthesis and Characterization of Bio-based Epoxy Resins Derived from Vanillyl

Alcohol. ACS Sustain. Chem. Eng. 4, 4328–4339 (2016).

12. Liu, R. et al. Bio-based epoxy-anhydride thermosets from six-armed linoleic acid-

derived epoxy resin. RSC Adv. 6, 52549–52555 (2016).

13. Garrison, M. D. & Harvey, B. G. Bio-based hydrophobic epoxy-amine networks

derived from renewable terpenoids. J. Appl. Polym. Sci. 133, n/a-n/a (2016).

14. Baroncini, E. A., Kumar Yadav, S., Palmese, G. R. & Stanzione, J. F. Recent

advances in bio-based epoxy resins and bio-based epoxy curing agents. J. Appl.

Polym. Sci. 133, n/a-n/a (2016).

15. Yue, L., Maiorana, A., Patel, A., Gross, R. & Manas-Zloczower, I. A sustainable

alternative to current epoxy resin matrices for vacuum infusion molding. Compos.

Part A Appl. Sci. Manuf. 100, 269–274 (2017).

16. Yue, L., Pircheraghi, G., Monemian, S. A. & Manas-Zloczower, I. Epoxy

composites with carbon nanotubes and graphene nanoplatelets – Dispersion and

synergy effects. Carbon N. Y. 78, 268–278 (2014).

17. Tang, L. & Weder, C. Cellulose Whisker/Epoxy Resin Nanocomposites. ACS

Appl. Mater. Interfaces 2, 1073–1080 (2010).

18. Peng, S. X., Shrestha, S., Yoo, Y. & Youngblood, J. P. Enhanced dispersion and

properties of a two-component epoxy nanocomposite using surface modified

cellulose nanocrystals. Polymer (Guildf). 112, 359–368 (2017).

19. Singh, M., Kaushik, A. & Ahuja, D. Surface functionalization of nanofibrillated

cellulose extracted from wheat straw: Effect of process parameters. Carbohydr.

Polym. 150, 48–56 (2016).

20. Hu, W., Chen, S., Yang, J., Li, Z. & Wang, H. Functionalized bacterial cellulose

derivatives and nanocomposites. Carbohydr. Polym. 101, 1043–1060 (2014).

21. Yu, H.-Y., Qin, Z.-Y., Yan, C.-F. & Yao, J.-M. Green Nanocomposites Based on

Functionalized Cellulose Nanocrystals: A Study on the Relationship between

Interfacial Interaction and Property Enhancement. ACS Sustain. Chem. Eng. 2,

875–886 (2014).

22. Raquez, J.-M. et al. Surface-modification of cellulose nanowhiskers and their use

as nanoreinforcers into polylactide: A sustainably-integrated approach. Compos.

Sci. Technol. 72, 544–549 (2012).

23. Khelifa, F., Habibi, Y., Bonnaud, L. & Dubois, P. Epoxy Monomers Cured by

High Cellulosic Nanocrystal Loading. ACS Appl. Mater. Interfaces 8, 10535–

10544 (2016).

24. Maiorana, A., Spinella, S. & Gross, R. A. Bio-based alternative to the diglycidyl

ether of bisphenol A with controlled materials properties. Biomacromolecules 16,

1021–31 (2015).

25. Singha, A. S. & Rana, A. K. Effect of Aminopropyltriethoxysilane (APS)

Treatment on Properties of Mercerized Lignocellulosic Grewia optiva Fiber. J.

Polym. Environ. 21, 141–150 (2012).

26. Abdelmouleh, M., Boufi, S., ben Salah, A., Belgacem, M. N. & Gandini, A.

Interaction of Silane Coupling Agents with Cellulose. Langmuir 18, 3203–3208

(2002).

27. Fernandes, S. C. M. et al. Bioinspired antimicrobial and biocompatible bacterial

cellulose membranes obtained by surface functionalization with aminoalkyl

groups. ACS Appl. Mater. Interfaces 5, 3290–3297 (2013).

28. Salon, M.-C. B. et al. Studies of interactions between silane coupling agents and

cellulose fibers with liquid and solid-state NMR. Magn. Reson. Chem. 45, 473–83

(2007).

29. Xu, S., Girouard, N., Schueneman, G., Shofner, M. L. & Meredith, J. C.

Mechanical and thermal properties of waterborne epoxy composites containing

cellulose nanocrystals. Polymer (Guildf). 54, 6589–6598 (2013).

30. Mwaikambo, L. Y. & Ansell, M. P. Chemical modification of hemp, sisal, jute,

and kapok fibers by alkalization. J. Appl. Polym. Sci. 84, 2222–2234 (2002).

31. Pickering, K. L., Efendy, M. G. A. & Le, T. M. A review of recent developments

in natural fibre composites and their mechanical performance. Compos. Part A

Appl. Sci. Manuf. 83, 98–112 (2016).

32. Saba, N. et al. Thermal and dynamic mechanical properties of cellulose nanofibers

reinforced epoxy composites. International Journal of Biological Macromolecules

102, (2017).

CHAPTER 4

Bacterial Cellulose Nanofiber Mats as Reinforcement for Epoxy-Anhydride systems

4.1 Introduction

Epoxy resins are the most commonly used thermoset polymers and have been used in a wide range of applications from coatings, adhesives to composite materials. They typically have high glass transition temperature, good chemical resistance, low density and high strength as provided by the cross-linked network structure. Thus, they are often used as the matrix for fiber reinforced lightweight structural composites, and widely applied in wind energy, automobile, and aviation industry. Traditionally DGEBA type epoxy resin is synthesized from the unrenewable petroleum-based resource. With the decreasing availability of petroleum resources and increasing environmental concerns related to the petroleum industry, it is of great importance today to develop alternative materials from renewable resources. Therefore, developing epoxy resin derived from renewable biomass becomes attractive in both academia and industry.

Recently, a new type of biobased epoxy resins, diglycidyl ethers of diphenolates (DGEDP), has been developed. The DGEDP biobased epoxy resins have tunable molecular structure and properties.1 Furthermore, this type of DGEDP resins exhibit comparable qualities with the commercial DGEBA resin. Good mechanical properties and processability make them a promising sustainable alternative for fabrication of fiber reinforced composites with the conventional vacuum infusion method.2

Combining a biobased epoxy resin with renewable reinforcement to develop high performance biobased structural composites will be of great importance for engineering applications. Nanocellulose, as highly crystalized nanoparticles or nanofibers, can be isolated from various renewable sources, including wood, cotton, ramie fibers, bacterial and some marine animals. It displays excellent mechanical properties. The theoretical

Young`s modulus for perfectly crystalized nanocellulose is established to be 167.5 GPa.3

Furthermore, the abundance of availability makes cellulose very low cost. Thus it becomes attractive as biobased reinforcement, and a variety of research has been done to apply nanocellulose as reinforcement for polymer composites. Bacterial cellulose (BC) nanofiber has an experimental Young`s modulus of 78 GPa, which is even higher than the traditionally used reinforcement E-glass fibers （72GPa） in lightweight composites.4

Moreover, the density of the BC (1.25 g/cm3) is only half of the glass fiber (2.5 g/cm3 ).

A problem in fabricating nanocellulose reinforced polymer composites is dispersion.

With large surface area and high aspect ratio, the percolation threshold for the nanocellulose is low. A tendency of aggregation of the nanocellulose at high volume fractions makes it difficult to process. An approach of achieving high volume fractions of filler is the solvent exchange method. However, at filler high volume fractions, the nanocellulose and resin mixture becomes a paste, which is very difficult to cast.

The BC network formed with continuous nanofibers is a pre-percolated network which could easily achieve high volume fraction while avoiding the dispersion problems. Here, we report on the fabrication and characterization of a biobased nanocomposite made with biobased DGEDP epoxy resin and BC nanofiber as reinforcement.

4.2 DGEBA Reinforced with Bacterial Cellulose Mats

4.2.1 Experimental Section

4.2.1.1 Materials

Bisphenol A diglycidyl ether (DGEBA) resin with epoxide equivalent weight 174,

Hexahydro-4-methylphthalic anhydride (MHHPA) and 1-methylimidazole (1-MI) are purchased from Sigma-Aldrich.

4.2.1.2 Biosynthesis of Bacterial Cellulose mats

Bacterial cellulose mat (BC-mat) was produced by cultivation of strain G. xylinus ATCC

700178 in Hestrin-Schramm medium (0.5% peptone, 0.5% yeast extract, 0.12% citric acid)5,6 and 4.0% mannitol. Culture pH was adjusted to 5.0 before autoclaving using acetic acid. Mannitol was autoclaved separately. Medium components and culture containers were autoclaved at 121°C for 15 min. Stock culture was prepared from 5-day grown culture, and 1 ml aliquots of culture were mixed with sterile glycerol (20%) and kept at -80°C. Seed cultures were prepared using full content of glycerol tube to incubate 50-ml fresh HS- medium and 2.0% mannitol in 250-ml Erlenmeyer flasks. Seed cultures were incubated at

30°C for 5 days. BC-mats were produced in sterile dishes filled with 800 ml HS medium.

Dishes were inoculated with 10 vol.% of seed-culture broth, then incubated at 30°C for 3 weeks. For the removal of microorganism and remaining culture medium, the BC membrane was harvested and repeatedly rinsed in 0.5M NaOH for 2 days, washed under distilled water for two days, cut into small pieces and freeze-dried as shown in Figure 4.1.

Bacterial cellulose mats were produced at RPI in Prof. Gross lab.

Figure 4.1 Biosynthesized BC mat water gel (left), freeze-dried BC mat (middle),

morphology under SEM (right)

4.2.1.3 DGEBA/BC mats composites preparation

The MHHPA anhydride curing agent catalyzed by 1-MI is a very latency curing system for epoxy. The reaction at room temperature is extremely slow, the mixture could remain in flow able liquid state for more than 48 hours, which was more than enough to complete the infusion of BC mats. Scheme 4.1 describes the two-step processing of the composite preparation. First, DGEBA resin was well mixed with the anhydrides curing agent

(MHHPA) at stoichiometric ratio with 1 phr 1-MI as accelerator. Pre-weighted modified

BC mats were immersed in the resin mixture. Then, the mixture was placed in a vacuum oven at room temperature for 8 hours, to ensure the complete infusion of resin into the porous BC mats. After a full infusion, the BC mats turned into a transparent gel-like state

(as shown in Scheme 4.1). The DGEBA/BC gel was then placed between Teflon sheets under a hot press at 80oC. At this temperature, the viscosity of the resin mixture became very low. A very slight pressure was applied to press the liquid resin mixture out of the BC network to achieve the desired BC loading in the final composites. Then the composites were cured at 80 oC for 4 hours followed by a post curing at 160 oC for another 4 hours.

Composites with 10 wt%, 15 wt%, 25wt% and 50 wt% BC network loading were prepared by this method.

Scheme 4.1 Two-step preparation of epoxy/BC composites

4.2.1.4 Characterizations

Dynamic Mechanical Analysis (DMA): TA Instruments Q800, operating in tensile mode with a constant frequency of 1 Hz, strain amplitude of 0.05%, was used to determine the storage modulus (Eʹ) and glass transition temperature (Tg, determined from the peak of the loss modulus) by scanning at 5 °C/min from 0 °C to 200 °C. Samples were prepared by cutting strips from the films with a width of 5mm.

Fourier-transform infrared spectroscopy (FTIR): The functional groups of cured samples were investigated with Perkin Elmer Spectrum Two in a range between 4000 and

400 cm-1.

4.2.2 Results and Discussion

The thermo-mechanical properties of the DGEBA/BC composites were analyzed by dynamic mechanical analysis. Figure 4.2 shows the tensile storage modulus of the prepared composites with different BC loadings as a function of temperature. The DGEBA matrix displays typical behavior of amorphous thermoset polymer. At the temperature below Tg, the storage modulus decreased only slightly with the increasing temperature, but above Tg, the storage modulus dropped dramatically to a rubbery state. A clear trend of increasing storage modulus with the increasing BC loading was observed. At 30 oC, the storage modulus increased more than 3 times from 2.2 GPa for the neat DGEBA to 7.6 GPa for the composites with 50 wt% BC. Above Tg, the reinforcement was more significant. At 180 oC, the storage modulus increased more than 130 times from 0.014 GPa for the neat

DGEBA to 1.88 GPa for the composites with 50 wt% BC.

Figure 4.2 Storage modulus versus temperature for the neat epoxy and the epoxy/BC

composites

The tan δ peak corresponding to the Tg shifts to lower temperature with the increasing BC loading. The Tg decreased from 153 oC for neat DGEBA to 131 oC for composites with 50 wt% BC. Similar phenomena of Tg decreasing in epoxy/cellulose nanocrystal composites

7,8,9,10 as well as in other polymermatrices11,12 were reported previously. The reason for the

Tg decrease is not well explained in the reported literature. A possible explanation discussed in one of the reports was the presence of unreacted epoxides lowering the average cross-link density, which results in lower Tg.7

Figure 4.3 tan δ versus temperature for the neat epoxy and the epoxy/BC composites

Figure 4.4 shows the thermal degradation curves obtained from TGA. DGEBA/BC composites decompose at a lower temperature than neat DGEBA. The decomposition temperatures of the DGEBA/BC composites decrease with increasing BC loading. If indeed the average cross-link densities of the composites are lower than neat DGEBA, then a change in thermal degradation is expected.

Figure 4.4 TGA of the neat epoxy and the epoxy/BC composites

FTIR was conducted to analyze the possible reactions. In the DGEBA/anhydride curing system with the catalyst imidazole, the ideal stoichiometry of epoxides to anhydrides should be 1:1, considering all epoxide groups only react with anhydrides. However, besides anhydride-epoxy copolymerization, there is also epoxy homopolymerization in this curing system.13,14 Therefore there should be excess anhydrides left in the cured system. This is confirmed by the FTIR in Figure 4.5, where in the neat DGEBA a C=O peak from unreacted anhydride was observed. Meanwhile, in all DGEBA/BC composites, this peak disappeared. The reason is that the surface free hydroxyl groups from BC react with some anhydride.

Figure 4.5 FTIR of the neat epoxy and the epoxy/BC composites

The reaction of hydroxyl groups from cellulose and anhydrides has been well studied in the literature. And it has been reported as a well-established approach for surface modification of cellulose.15 Scheme 4.2 shows the acylation of cellulose with succinic, maleic and phthalic anhydrides. These reactions happen under mild conditions without any solvent or catalyst. During the curing process at 160 oC, besides the epoxy-anhydride cross- link reaction, the cellulose-anhydride reaction also occurs in the same time. The disappearance of anhydride peak on FTIR confirms that part of the anhydride is consumed by BC, resulting in lower cross-link density of the curing system.

Scheme 4.2 Acylation reactions of cellulose with anhydrides.15

4.2.3 Conclusions

DGEBA/BC composites are successfully prepared by infusion of the resin into the 3D porous BC network, subsequent hot-press and curing. The MHHPA anhydride curing agent catalyzed by 1-MI could provide long enough processing time to complete the resin infusion. The composites reinforced with BC display significant mechanical reinforcement, especially above the Tg. However, due to the reaction of anhydride and surface free

100 hydroxyl groups from BC, the composites exhibit lower Tg and thermal degradation temperatures compared to neat DGEBA.

4.3 DGEDP-ethyl Reinforced with Modified BC mats

The target of this project is to fabricate BC reinforced biobased DGEDP composites. To minimize the cellulose-anhydride reaction, and also help the resin infusion processing, the surface hydroxyl groups of BC are modified with chlorotrimethylsilane. After modification, there will be less hydroxyl groups on the BC surface. At the same time, the hydrophilic surface of the native BC becomes more hydrophobic, which would facilitate the infusion of the hydrophobic resin.

4.3.1 Experimental Sections

4.3.1.1 Materials

Biobased epoxy resin DGEDP-ethyl (diglycidyl ether of diphenolate ethyl ester) was synthesized as previously reported.1 Hexahydro-4-methylphthalic anhydride (MHHPA), 1- methylimidazole (1-MI) and chlorotrimethylsilane (TMSCl) (≥99%) was purchased from

Sigma Aldrich, trimethylamine (≥99.5%) was purchased from EMD Millipore Corporation,

Koptec Pure Ethanol-200 Proof was purchased from Decon Labs, Inc., and dichloromethane was purchased from Macron Fine Chemicals (dried over 4Å MS for 24 h prior to use). All other reagents and solvents were used as received without further purification and deionized water was used in all experiments.

4.3.1.2 Modification of BC mats with TMSCL

101

The BC mats were modified with TMSCl following literature procedures as shown in

Scheme 4.3.16,17 To a 500 mL round bottomed flask equipped with magnetic stirrer 200 mL dichloromethane and 4.8 mL trimethylamine were added and stirred thoroughly for 10 min.

After that 6.0 mL TMSCl were added and stirred for another 10 min followed by the addition of highly thick bacterial cellulose mats. Then, the flask was attached to a water condenser and Ar inlet, and refluxed at 60 oC for about 4 h. Afterward the reaction flask was cooled to room temperature and the reaction mixture was quenched by addition of small amounts of EtOH. Subsequently the liquids were decanted and the BC mats were transferred into a beaker with EtOH (100 mL) and washed 3 times (3x100 mL, solvent changed every 30 min). These mats were washed with DI water (3x100 mL) and freeze dried for 48 h.

Scheme 4.3 Scheme of the BC surface modification.

4.3.1.3 DGEDP-ethyl/BC mats composites preparation

The MHHPA anhydride curing agent catalyzed by 1-MI is a very latency curing system for epoxy. The reaction at room temperature is extremely slowand the mixture could remain in liquid state for more than 48 hours, which was more than enough to complete the infusion of BC mats. DGEDP-ethyl resin was well mixed with the anhydride curing agent

(MHHPA) at stoichiometric ratio and with 1 phr 1-MI as accelerator. Pre-weighted

102 modified BC mats were immersed in the resin mixture. Then, the mixture was placed in vacuum oven at room temperature for 8 hours, to ensure the complete infusion of resin into the porous BC mats. After complete infusion, the BC mats changed into a transparent gel- like state. The DGEDP-ethyl/BC gel was then placed between Teflon sheets in a hot press at 80oC. At this temperature, the viscosity of the resin mixture became very low. Slight pressure was applied to press the liquid resin mixture out of the BC network in order to achieve the desired BC loading in the final composite. These composites were cured at 80 oC for 4 hours, followed by a post curing at 160 oC for another 4 hours. Composites with

5 %, 10%, 20% and 50 % v/v BC loading were prepared.

4.3.1.4 Characterizations

Tensile Test: The test was conducted with a Zwick-Roell Z0.5 tensile Instron 1011 universal tensile machine with a 10 kN capacity load cell and standard three point bending fixtures at a test speed of 2mm/min. Five specimens were tested for each sample.

Thermogravimetric Analysis (TGA): The thermal stability of the prepared resins was studied by TGA using a TA Instruments Q500 with an aluminum pan. The samples were about 10 mg each and were run from room temperature to 600 °C at a heating rate of 10 °C per minute under constant N2 flow.

103

Energy-Dispersive X-ray Spectroscopy (EDS or EDX). The relative elemental composition of unmodified and modified BC mats was determined by EDS or EDX spectra under same testing conditions with an accelerating voltage of 10 kV and platinum coating.

4.3.2 Results and Discussion

Surface modification of BC mats

Figure 4.6 EDS results for BC (top) and modified BC (bottom)

104

To investigate the trimethylsilylation of bacterial cellulose, EDS spectra were performed.

The unmodified BC mat displayed only carbon (63.23%) and oxygen (36.76%) peaks without the silicon peak. For the trimethylsilylated BC mat, the silicon peak appears along with carbon and oxygen with relative atomic % of 0.40, 37.03 and 62.58% respectively.

16 The degree of substitution (DS) was calculated using equation 4.1. where 퐶푆푖 and 퐶푐 are the atomic percentages of carbon and silica respectively. The calculated DS is 0.039, and since this is less than the 0.4, based on the theory of Heux et al. the conclusion is that the modification has occurred only on the surface of the cellulose nanofibers and not in the interior.17

퐶 DS = 푆푖 Equ.4.1 (퐶푐−3퐶푆푖)/6

Figure 4.7 XRD curves of the BC and modified BC

The impact of modifcation on BC crystallinity was evaluated by XRD. The diffraction profiles of unmodified and modified BC are shown in Figure 4.7. The degree of cystallinity

105 calculated from Segal equation (Ic=(I002-Iamorphous)/Iamorphous) is 92% for both unmodified and modified BC, which indicates also the limited amount of modification occurring only at the surface of the BC nanofiber and not affecting BC crystallinity.

Tensile Properties of DGEDP-ethyl/BC mats composites

Mechanical properties of the biobased DGEDP-ethyl resin/BC composites were evaluated using the uniaxial tensile test. Young's modulus, tensile strength and the elongation break of the composites (with 5 %, 10%, 20% and 30% volume fraction) and the neat DGEDP- ethyl and BC dry film are shown in Figure 4.8. The composites reinforced with 30 % v/v

BC show an increase in Young`s modulus (8.8 GPa) and tensile strength (84MPa) by comparison with the neat DGEDP-ethyl resin (1.22 GPa modulus and 60 MPa tensile strength). A clear trend of increasing Young`s modulus as the BC loading increases was observed. All composites exhibited a more rigid behavior with a dramatically lower elongation at break by comparison with the neat resin. It should be pointed out that the BC film tested was prepared by drying the BC gel in air. Previous studies show that BC gels dried under heat and pressure could achieve lower porosity and better mechanical properties.18,19

106

Figure 4.8 Tensile properties of neat DGEDP-ethyl, BC film and DGEDP-ethyl/BC

composites with different BC volume fractions.

The Young`s modulus of randomly oriented fiber composites can be modeled using the

Cox-Krenchel model, developed based on the classical shear-lag theory.20,21 Based on this model the Young’s modulus of the composite, Ecomposite, can be calculated as:

Ecomposite= η0ηLvfEf + (1-vf)Em Equ.4.2 where Ef is the fiber modulus, Em is the matrix modulus, η0 and ηL are efficiency factors corresponding to fiber orientation and length respectively, and vf is the fiber volume fraction. Here Em=1.22 GPa as measured in the tensile test, and Ef =78±17 GPa as measured by Guhados etc. by atomic force microscopy.4

107

The orientation factor of BC fibers can be considered as the value of in-plane isotropic orientation η0=0.375. Considering BC to be a percolated network formed with continuous fibers, the length efficiency of the BC fibers is assumed to be unity. The Cox-Krenchel model is plotted as the black line in Figure 4.9

Figure 4.9 Young`s modulus of the composites as a function of BC volume fraction.

Using a simple rule of mixture, the composite Young’s modulus can be written as:

Ecomposite= vfEr + (1-vf)Em Equ.4.3

Here Er is the bulk Young`s modulus of the neat BC film. The reported values for BC film range from 10 to 35 GPa.22 In this work the Young`s modulus of the BC film prepared by air drying is 16.8GPa as measured in the tensile test. Using the experimental values for the matrix and the BC film the rule of mixture results are shown as the red line in Figure 4.9.

These results suggest that the BC mats used in our composites have a higher modulus than

108 the dried BC film and the calculated value η0Ef for the modulus of the BC film could give a more accurate prediction.

Thermo- Mechanical Properties of DGEDP-ethyl/BC mats composites

The thermo-mechanical properties of the DGEDP-ethyl/BC biobased composites were investigated by dynamic mechanical analysis and Figure 4.10 shows the storage modulus of the composites with different BC loadings as a function of temperature. The biobased matrix DGEDP-ethyl displays typical behavior of amorphous thermoset polymer. At temperatures below Tg, the storage modulus decreases only slightly with increasing temperature, but above Tg, the storage modulus drops dramatically. A clear trend of increasing storage modulus with increasing BC loading was observed. At 30 oC, the storage modulus increased more than 3 times from 2.27 GPa for the neat DGEDP-ethyl to 7.7 GPa for the composites with 30% v/v BC. Above Tg, the reinforcement was more significant.

At 180 oC, the storage modulus increased more than 100 times from 0.024 GPa for the neat

DGEBA to 2.47 GPa for the composites with 30% v/v BC. The data are summarized in

Table 4.1.

109

Figure 4.10 Storage modulus as function of temperature for DGEDP-ethyl, BC film and

the composites

Figure 4.11 is the tan  peak corresponding to the Tg as a function of temperature.

Interestingly for 5% v/v BC composites, the Tg is slightly increased, but then the Tg start decreasing with further increasing the volume fraction of BC. The Tg data obtained from

DSC as shown in Figure 4.12 exhibit a similar trend. Analogous phenomena of changes in

Tg for epoxy/cellulose nanocrystal composites were reported previously. Tang8, Omrani23 and Saba24 reported that in epoxy/cellulose nanocrystal systems, Tg increases slightly at low filler concentration and decreases at higher filler concentration. Mathew25 reported that in tunicate cellulose nanowhisker reinforced sorbitol plasticized starch nanocomposites, the Tg was found to slightly increase at low filler concentration but decrease at higher loadings.5 It should be also pointed out, that regardless of the expectations for an enhancement effect on Tg, quite a few authors did not observe improvement in Tg in nanocellulose polymer composites.9,7,11,10,12,26,27,28,29,30

110

Figure 4.11 tan() as the function of temperature of DGEDP-ethyl/BC composites.

Figure 4.12 DSC curves of DGEDP-ethyl/BC composites.

111

Thermal stability of DGEDP-ethyl, BC and their composites was studied with thermal gravimetric analysis. Decomposition of DGEDP-ethyl occurs between 350 and 500 oC, whereas the degradation of BC is in a lower range between 300 to 400 oC. Thus, the decomposition temperature of DGEDP-ethyl/BC composites was expected to be between neat DGEDP-ethyl and BC. The results shown in Figure 4.13 confirm this hypothesis. The temperature at the maximum weight loss rate slightly decreases as the BC loading increases.

The change is not significant comparing 420.2 oC for the composites with 30 % v/v BC loading with 425.1 oC for the neat DGEDP-ethyl resin.

Figure 4.13 TGA curves of DGEDP-ethyl, BC and the composites

112

Table 4.1Mechical and thermal properties of the DGEDP-ethyl, BC film and the

composites. ET is the modulus obtained from tensile test, E30 and E180 are the storage

o o moduli obtained from DMA at 30 C and 180 C, respectively. Tdmax is determined from

TGA.

E E E Tg Tg Tdmax T 30 180 DMA DSC GPa GPa GPa oC oC oC DGEDP- 1.22±0.41 2.27 0.024 152.2 136.5 425.1 ethyl

BC film 16.40±0.84 15.80 15.6 \ \ 378.4

5 % v/v 2.50±0.42 2.96 0.21 157.7 138.6 427.9

10% v/v 3.90±0.53 3.94 0.45 144.9 128.9 425.2

20% v/v 7.20±0.71 6.33 1.72 143.8 127.2 424.4

30% v/v 8.80±0.98 7.70 2.47 141.9 123.3 420.2

4.3.3 Conclusions

High-performance biobased nanocomposites of biobased epoxy resin DGEDP-ethyl and

BC nanofiber network as reinforcement were fabricated by impregnation of BC network with the resin mixture, subsequent hot-pressing and curing. This processing method could easily achieve high cellulose nanofiber volume fractions (30 % v/v and higher).

Hydrophobic modification of the BC network can improve the resin impregnation and also reduce the possible reaction between BC and anhydride. The BC network exhibits excellent reinforcing properties for the biobased epoxy resin. Both Young`s modulus and storage

113 modulus of the composites were significantly higher by comparison with the neat resin.

Furthermore, the storage modulus of the composites at high temperature increased more than 100 times compared with the neat resin.

114

4.4 References

1. Maiorana, A., Spinella, S. & Gross, R. A. Bio-based alternative to the diglycidyl

ether of bisphenol A with controlled materials properties. Biomacromolecules 16,

1021–31 (2015).

2. Yue, L., Maiorana, A., Patel, A., Gross, R. & Manas-Zloczower, I. A sustainable

alternative to current epoxy resin matrices for vacuum infusion molding. Compos.

Part A Appl. Sci. Manuf. 100, 269–274 (2017).

3. Tashiro, K. & Kobayashi, M. Theoretical evaluation of three-dimensional elastic

constants of native and regenerated celluloses: role of hydrogen bonds. Polymer

(Guildf). 32, 1516–1526 (1991).

4. Guhados, G., Wan, W. & Hutter, J. L. Measurement of the Elastic Modulus of

Single Bacterial Cellulose Fibers Using Atomic Force Microscopy. Langmuir 21,

6642–6646 (2005).

5. Hestrin, S. & Schramm, M. Synthesis of cellulose by Acetobacter xylinum. 2.

Preparation of freeze-dried cells capable of polymerizing glucose to cellulose.

Biochem. J. 58, 345–352 (1954).

6. Ruka, D. R., Simon, G. P. & Dean, K. M. Altering the growth conditions of

Gluconacetobacter xylinus to maximize the yield of bacterial cellulose.

Carbohydr. Polym. 89, 613–622 (2012).

7. Ansari, F., Galland, S., Johansson, M., Plummer, C. J. G. & Berglund, L. A.

Cellulose nanofiber network for moisture stable, strong and ductile biocomposites

and increased epoxy curing rate. Compos. Part A Appl. Sci. Manuf. 63, 35–44

(2014).

115

8. Tang, L. & Weder, C. Cellulose Whisker/Epoxy Resin Nanocomposites. ACS

Appl. Mater. Interfaces 2, 1073–1080 (2010).

9. Lu, J., Askeland, P. & Drzal, L. T. Surface modification of microfibrillated

cellulose for epoxy composite applications. Polymer (Guildf). 49, 1285–1296

(2008).

10. Margem, F. M., Monteiro, S. N., Bravo Neto, J., Rodriguez, R. J. S. & Soares, B.

G. The dynamic-mechanical behavior of epoxy matrix composites reinforced with

ramie fibers. MatÃ\copyrightria (Rio Janeiro) 15, 164–171 (2010).

11. Raquez, J.-M. et al. Surface-modification of cellulose nanowhiskers and their use

as nanoreinforcers into polylactide: A sustainably-integrated approach. Compos.

Sci. Technol. 72, 544–549 (2012).

12. Walsh, C. C. and P. ALIGNED NANOCOMPOSITES FROM CELLULOSE

NANOCRYSTALS BY ELECTROSPINNING WITH A SOLUBLE POLYMER

FOLLOWED BY THERMOSET IMPREGNATION Craig Clemons and Phil

Walsh , USDA Forest Service , Forest Products Laboratory , Madison , WI

Electrospinning and Composite Pr. in ANTEC (2017).

13. Bouillon, N., Pascault, J. P. & Tighzert, L. Influence of different imidazole

catalysts on epoxy-anhydride copolymerization and on their network properties. J.

Appl. Polym. Sci. 38, 2103–2113 (1989).

14. Park, W. H., Lee, J. K. & Kwon, K. J. Cure Behavior of an Epoxy-Anhydride-

Imidazole System. Polymer Journal 28, 407–411 (1996).

15. Cellulose - Fundamental Aspects and Current Trends. (2015).

16. Sai, H. et al. Surface Modification of Bacterial Cellulose Aerogels’ Web-like

116

Skeleton for Oil/Water Separation. ACS Appl. Mater. Interfaces 7, 7373–7381

(2015).

17. Fumagalli, M., Ouhab, D., Boisseau, S. M. & Heux, L. Versatile Gas-Phase

Reactions for Surface to Bulk Esterification of Cellulose Microfibrils Aerogels.

Biomacromolecules 14, 3246–3255 (2013).

18. Retegi, A. et al. Bacterial cellulose films with controlled microstructure–

mechanical property relationships. Cellulose 17, 661–669 (2010).

19. Iguchi, M., Yamanaka, S. & Budhiono, A. Bacterial cellulose—a masterpiece of

nature’s arts. J. Mater. Sci. 35, 261–270 (2000).

20. Lee, K. Y., Ho, K. K. C., Schlufter, K. & Bismarck, A. Hierarchical composites

reinforced with robust short sisal fibre preforms utilising bacterial cellulose as

binder. Compos. Sci. Technol. 72, 1479–1486 (2012).

21. Lee, K. Y., Aitomäki, Y., Berglund, L. A., Oksman, K. & Bismarck, A. On the use

of nanocellulose as reinforcement in polymer matrix composites. Compos. Sci.

Technol. 105, 15–27 (2014).

22. Moon, R. J., Martini, A., Nairn, J., Simonsen, J. & Youngblood, J. Cellulose

nanomaterials review: structure{,} properties and nanocomposites. Chem. Soc.

Rev. 40, 3941–3994 (2011).

23. Omrani, A., Simon, L. C. & Rostami, A. A. Influences of cellulose nanofiber on

the epoxy network formation. Mater. Sci. Eng. A 490, 131–137 (2008).

24. Saba, N. et al. Thermal and dynamic mechanical properties of cellulose nanofibers

reinforced epoxy composites. International Journal of Biological Macromolecules

102, (2017).

117

25. Mathew, A. P. & Dufresne, A. Morphological Investigation of Nanocomposites

from Sorbitol Plasticized Starch and Tunicin Whiskers. Biomacromolecules 3,

609–617 (2002).

26. Lu, J., Wang, T. & Drzal, L. T. Preparation and properties of microfibrillated

cellulose polyvinyl alcohol composite materials. Compos. Part A Appl. Sci. Manuf.

39, 738–746 (2008).

27. Azizi Samir, M. A. S., Alloin, F., Sanchez, J.-Y. & Dufresne, A. Cellulose

nanocrystals reinforced poly(oxyethylene). Polymer (Guildf). 45, 4149–4157

(2004).

28. Grunert, M. & Winter, W. T. Nanocomposites of Cellulose Acetate Butyrate

Reinforced with Cellulose Nanocrystals. J. Polym. Environ. 10, 27–30 (2002).

29. Peng, S. X., Moon, R. J. & Youngblood, J. P. Design and characterization of

cellulose nanocrystal-enhanced epoxy hardeners. Green Mater. 2, 193–205 (2014).

30. Petersson, L., Mathew, A. P. & Oksman, K. Dispersion and properties of cellulose

nanowhiskers and layered silicates in cellulose acetate butyrate nanocomposites. J.

Appl. Polym. Sci. 112, 2001–2009 (2009).

118

CHAPTER 5

Vitrimerization: A Novel Concept to Reprocess and Recycle Thermoset Waste

119

5.1 Introduction

Thermosets are covalently cross-linked networks that, unlike thermoplastics cannot be reprocessed by melting or dissolution in any solvents. These networks can exhibit a glass transition temperature (Tg) lower than the designed application service temperature

(elastomer/rubber) or higher than the service temperature (thermoset resins). Thermoset rubbers find applications in the automotive industry (interiors, bumpers), biomedical devices, bedding, furniture, packaging, gaskets, O-rings and so on.

Usually, thermoset resins show significant benefits in comparison with thermoplastics exhibiting dimensional stability, high mechanical properties, high thermal/creep/ and chemical resistance, as well as durability. This class of polymers maintain their structural strength, thermal and electrical resistance characteristics even after prolonged use.

Thermosets find many industrial uses as coatings, adhesives and fiber reinforced composites for many high-tech applications. High stiffness and strength in combination with their light weight make them play a vital role in composites for clean energy production, (e.g. wind turbine blades, hydrokinetic power generation, support structures for solar systems and their encapsulations, and geothermal energy production) and in the manufacturing of lighter vehicles (automotive, airplanes, trains, boats and aerospace) for reduced fuel consumption. There are many other applications requiring high structural strength and durability, thermal and corrosion resistance, such as structural materials for buildings, pipelines, industrial equipment and/or their components for instance heat exchangers, light-emitting diode lenses, fly-wheels for electricity grid stability, containers, or off-shore structures in which composite materials are the material of choice.

120

Despite the great features of thermosets, they are typically produced in relatively small quantities and are very expensive. Their main advantage, which is retaining a lasting and intractable three- dimensional structure, creates also a crucial disadvantage since they cannot be recycled or reprocessed.

A very interesting strategy to induce re-formability and healing in chemically crosslinked polymer networks is by using exchangeable chemical bonds that will lead to a dynamic cross-linked network.1,2,3 Polymeric systems containing such exchangeable bonds are covalent adaptable networks (CANs).4 Depending on the exchange mechanism,

CANs can be classified to two different categories. The first category comprises networks where the exchange mechanism of crosslinks is dissociative. For such networks, most of the cross-links break under certain conditions (temperature, UV-light exposure, pH,) and re-form again with a change in the conditions. This type of adaptive networks show a sudden and significant decrease in the viscosity, with breaking of the cross-linking bonds.5In the second category, the mechanism of crosslinking is associative and a crosslinking bond does not break until a new bond forms, which makes the network permanent and dynamic.6,7,8

Vitrimers are polymeric associative CANs that have permanent networks and demonstrate a gradual viscosity decrease upon heating, which is a distinctive character of vitreous silica.9 By definition, vitrimers are polymeric networks made with covalent crosslinking. The crosslinking bonds of such networks have an associative nature which results in the ability of material to change its topology via exchange reactions. These increasing temperature and provides malleability to the network. The viscosity of vitrimers is governed by the chemical exchange reaction at elevated temperatures and like

121 silica and unlike dissociative networks and thermoplastics, decreases gradually. Vitrimers maintain permanent network at all temperatures until degradation, and they can swell but not dissolve in specific solvents. However, swelling ratios are higher for these networks in comparison with the non-dynamic ones.

Due to importance of polymer recycling, scientists are interested in the concept of designing materials based on cradle-to-cradle life cycle. Dynamic networks offer the opportunity to design materials considering the cradle-to-cradle concept. We propose a process to recycle thermosets coined as “vitrimerization” by changing them into dynamic networks with the use of an appropriate catalyst solution which will turn them into vitrimers. Thermosets will be swollen in a solution of an appropriate catalyst to allow the catalyst to diffuse into the network. Upon removal of the solvent, the catalyst will facilitate the occurrence of exchange reactions at elevated temperature rendering the system a dynamic network. Thus, the thermoset becomes recyclable and healable for many times. Vitrimers have been studied extensively since their inception. The world of thermosets has changed dramatically. New thermosets created will be able to be recycled and reprocessed post its intended use. However, the already generated thermoset waste still poses a significant problem. This resarch attempts to provide a solution to present thermoset waste.

5.2 Concept of “Vitrimerization”

The reprocessable and recyclable vitrimer materials are based on dynamic associative exchange reactions. So far, there are several well studied reactions which could be dynamic associative exchange reactions with specific catalyst including carboxylate

122 transesterification,10,11 transamination of vinylogous urethanes,12 transalkylation of triazolium salts13 and so on. The pioneering work of Leibler et al. demonstrated the concept of vitrimers by using transesterification within epoxy and acid/anhydride systems.8 Once there are free hydroxyl and ester groups in the network, then the exchangeable transesterification reactions could be promoted by the catalyst. Based on this concept, the new concept of “vitrimerization” is initiated. For a crosslinked thermoset network, if there are free hydroxyl and ester groups in the network, by introducing an appropriate catalyst, the network could become vitrimer-like network. This concept can be further promoted to other kind of dynamic associative reactions. Once the thermoset network turns into vitrimer-like network, it becomes reprocessable and recyclable.

To proof the concept of “vitrimerization”, the transesterification reaction in epoxy networks is investigated. The epoxy and fatty acid (a mixture of tricarboxylic and dicarboxylic fatty acids) crosslinked network is chosen due to the accessible free hydroxyl and ester groups. The crosslink reaction is shown as Scheme 5.1.

Scheme 5.1 DGEBA/Fatty acid crosslinking reaction.14 reedited

123

The approach is to swell the cured thermoset with a catalyst solution, and allow the catalyst to diffuse into the network, then remove the solvent. If the concept works, the vitrmierized thermoset can behave like vitrimer materials, thus reprocessable.

5.3 Experimental Sections

5.3.1 Materials

Bisphenol A diglycidyl ether (DGEBA) resin with epoxide equivalent weight 174 and the catalysit Tin (II) 2-ethylhexanoate (Sn(Oct)2) are purchased from Sigma-Aldrich. The mixture of fatty acid monomers Pripol 1040 was provided by Uniqema Inc. All reagents and solvents were used as received without further purification.

5.3.2 The Vitrimerization Process

The DGEBA and fatty acids were well mixed based on stoichiometry (epoxy group/acid group 1:1), and kept at 160 oC until fully cured. The cured sample was cut into small cubes (approximately 2mm*2mm*2mm) and immersed into a dichloromethane solvent with 20 wt% Tin (II) 2-ethylhexanoate (Sn(Oct)2) catalyst for 48 hours at room temperature with magnetic stirring. The samples were then washed with ethanol to remove the catalyst from the surface and dried in a vacuum oven for 24 hours at 80oC.

The volume change during swelling was recorded as shown in Figure 5.1. After 24 hours at room temperature, the volume expanded around 1.6 times. This process allows the small catalyst molecules to diffuse into the epoxy network. It should be pointed out, that swelling behavior could vary with solvent, swelling time and temperature. Here dichloromethane was chosen because, first it could swell the epoxy network, second, it

124 could dissolve the catalyst very well, and last it has a low boiling point (39.6 oC), thus facilitating removal after use.

Figure 5.1 Volume change of the cured epoxy during swelling in the solvent.

5.4 Results and Discussion

The vitrimerized epoxy after complete solvent removal was then processed with conventional polymer processing techniques including extrusion and hot pressing. Figure

5.2 shows epoxy processed with extrusion at 250 oC. The neat epoxy could not be extruded. However, the vitrimerized epoxy could be successfully extruded into one piece of tensile bar. The same phenomenon was observed for the hot-press processing. The epoxy small cuts were placed on the hot-press plate at 250 oC and after 10 minutes preheat, a 5-ton pressure was applied and kept for 15 minutes. The original epoxy cubes were pressed into powder, whereas the vitrimerized epoxy cubes were pressed into a film as shown in Figure 5.2. This result indicates the concept of vitrimerization works. The vitrimerized epoxy cubes were welded into one piece due to the transesterification reaction occurring at the interface, which could bond the separated parts together.

125

Figure 5.2 Vitrimerized epoxy and original epoxy processed with extrusion and hot-press

processing

One important characteristic of vitrimer materials is the stress-relaxation behavior. The dynamic networks with exchangeable reactions enable topology rearrangements which allow the system to relax stresses. Shear stress relaxation experiments were conducted with the vitrimerized epoxy. The film samples were prepared with hot-press as described above. Figure 5.3 shows the time and temperature dependent stress-relaxation curves of the vitrimerized epoxy. Based on Maxwell model for viscoelastic materials, the relaxation time was determined at 0.37 (1/e) of the normalized relaxation modulus.12 A vitrimer-like stress-relaxation behavior was observed for the vitrimerized epoxy.

Compared with the original epoxy sample which has a permanent crosslinked network and did not show stress-relaxation within the experimental time scale at 200 oC, the normalized relaxation modulus of the vitrimerized epoxy decreases from 1 to 0.37 in approximately 33min (~2000s). Also, the relaxation rate, which depends on the

126 transesterification reaction rate, is faster at higher temperatures. The stress-relaxation test also indicates the validity of the vitrimerization concept. The vitrimerized epoxy exhibits a vitrimer-like behavior with dynamic network at high temperatures.

Figure 5.3 Shear stress relaxation experiments: normalized relaxation modulus as a

function of time for vitrmerized epoxy at different temperatures.

5.5 Conclusions and Future Works

This chapter introduces the new concept of vitrimerization of crosslinked thermoset networks, which could turn thermosets into vitrimer-like materials reprocessable with conventional processing techniques. The concept was proofed with the epoxy fatty acid crosslinked network by using the transesterification reaction. The preliminary study indicates that by simple swelling the thermoset in a catalyst solution and allowing the catalyst to diffuse into the network, the epoxy network changes into a vitrimer-like dynamic network. This procedure could also work for epoxy anhydride networks,

127 polyurethane networks and other thermoset polyester networks containing free hydroxyl groups and ester groups.

This concept could be further explored in the future. More research could be done to optimize the vitrimerization conditions. Equally important is to study material properties after vitrimerization.

128

5.6 References

1. Lehn, J.-M. Dynamers: dynamic molecular and supramolecular polymers. Prog.

Polym. Sci. 30, 814–831 (2005).

2. Rowan, S. J., Cantrill, S. J., Cousins, G. R. L., Sanders, J. K. M. & Stoddart, J. F.

Dynamic Covalent Chemistry. Angew. Chemie Int. Ed. 41, 898–952 (2002).

3. Wojtecki, R. J., Meador, M. A. & Rowan, S. J. Using the dynamic bond to access

macroscopically responsive structurally dynamic polymers. Nat Mater 10, 14–27

(2011).

4. Kloxin, C. J., Scott, T. F., Adzima, B. J. & Bowman, C. N. Covalent Adaptable

Networks (CANs): A Unique Paradigm in Cross-Linked Polymers.

Macromolecules 43, 2643–2653 (2010).

5. Denissen, W., Winne, J. M. & Du Prez, F. E. Vitrimers: permanent organic

networks with glass-like fluidity. Chem. Sci. 7, 30–38 (2016).

6. Kloxin, C. J., Scott, T. F., Park, H. Y. & Bowman, C. N. Mechanophotopatterning

on a Photoresponsive Elastomer. Adv. Mater. 23, 1977–1981 (2011).

7. Scott, T. F., Schneider, A. D., Cook, W. D. & Bowman, C. N. Photoinduced

Plasticity in Cross-Linked Polymers. Science (80-. ). 308, 1615–1617 (2005).

8. Solid, L. D. et al. Silica-Like Malleable Materials from. Science (80-. ). 334, 965–

968 (2011).

9. Angell, C. A. Formation of Glasses from Liquids and Biopolymers. Science (80-. ).

267, 1924–1935 (1995).

129

10. Brutman, J. P., Delgado, P. A. & Hillmyer, M. A. Polylactide Vitrimers. ACS

Macro Lett. 3, 607–610 (2014).

11. Capelot, M., Montarnal, D., Tournilhac, F. & Leibler, L. Metal-catalyzed

transesterification for healing and assembling of thermosets. J. Am. Chem. Soc.

134, 7664–7667 (2012).

12. Denissen, W. et al. Vinylogous Urethane Vitrimers. Adv. Funct. Mater. 25, 2451–

2457 (2015).

13. Hendriks, B., Waelkens, J., Winne, J. M. & Du Prez, F. E. Poly(thioether)

Vitrimers via Transalkylation of Trialkylsulfonium Salts. ACS Macro Lett. 6, 930–

934 (2017).

14. Yu, K., Taynton, P., Zhang, W., Dunn, M. L. & Qi, H. J. Influence of

stoichiometry on the glass transition and bond exchange reactions in epoxy

thermoset polymers. RSC Adv. 4, 48682–48690 (2014).

130

CHAPTER 6

Conclusions and Future Works

131

6.1 Conclusions

This thesis aims at exploring the feasibility of replacing the currently used petroleum- derived epoxy resins. The biobased formulation with DGEDP resin and glycidyl ether of eugenol as the reactive diluent demonstrates comparable processability for vacuum infusion processing while maintaining and in some cases exceeding the mechanical performance with commercial resin system.

Furthermore, biobased nanocomposites reinforced with amine modified cellulose nanocrystals achieved improved thermal-mechanical properties and could further extend the application of the biobased epoxy resin. The reaction of amino groups with DGEDP- ethyl epoxy groups increased compatibility between filler and matrix components. Good dispersion of the nanofillers, as well as stronger filler-matrix interactions, contributed to significant improvement in storage modulus above Tg.

Moreover, high-performance biobased nanocomposites composed of biobased epoxy resin DGEDP-ethyl and bacterial cellulose nanofiber network as reinforcement were fabricated with a two-step conventional method by impregnation of bacterial cellulose network with the resin mixture, subsequent hot pressing and curing. High nanofiber volume content bio composites with low density but significantly improved mechanical properties were achieved. The overall performance of the biobased epoxy resins as well as their composites demonstrates their feasibility as a more sustainable choice for the future.

The proof of concept work of vitrimerization clearly showed the reprocessability of the epoxy thermoset. The epoxy network became vitrimer-like network and could be processed again with conventional processing techniques like compression molding,

132 extrusion and injection molding. The developed methodology may assure recycling and reprocessing of waste thermoset materials, which could further improve sustainability.

6.2 Future Works

This thesis is focused on mechanical performance of the biobased DGEDP epoxy resin as well as their composites for their potential application as structural materials. However, the tunable molecular structure makes it possible for these biobased DGEDP resins to be applied in other areas like coatings, adhesives or functional materials. For example, the

DGEDP-methyl with high viscosity could be suitable for adhesives.

Another potential field worth further exploring is the biobased vitrimer materials.

Vitrimer materials are a new area, very promising in the future due to their reprocessable and recyclable features. The biobased epoxy resins as well as other biobased thermoset resins can be potentially vitrimers to increase sustainability from start to end. DGEDP type epoxy monomers with the extra ester groups, may promote the transesterification reaction in the dynamic network.

Last but not least is the research on vitrimerization processing for reprocessing of the thermoset waste. There is huge amount of thermoset waste produced every day. The non- degradable and non-processable waste creates environmental problems. The vitrimerization concept demonstrated the possibility of reprocessing thermosets, but there is still more work to make it feasible for commercial application. The processing conditions could be further optimized. So far, only the polyester type thermoset network which can undergo transesterification has been studied. Rubber would be a potential new direction to further study the chemistry and design for vitrimerization.

133

Bibliography

Auvergne, R.; Caillol, S.; David, G.; Boutevin, B.; Pascault, J.-P. Biobased Thermosetting Epoxy: Present and Future. Chem. Rev. 2014, 114 (2), 1082–1115.

Al-Saleh, M. H. & Sundararaj, U. Review of the mechanical properties of carbon nanofiber/polymer composites. Compos. Part A Appl. Sci. Manuf. 42, 2126–2142 (2011).

Ansari, F., Galland, S., Johansson, M., Plummer, C. J. G. & Berglund, L. A. Cellulose nanofiber network for moisture stable, strong and ductile biocomposites and increased epoxy curing rate. Compos. Part A Appl. Sci. Manuf. 63, 35–44 (2014).

Azizi Samir, M. A. S., Alloin, F., Sanchez, J.-Y. & Dufresne, A. Cellulose nanocrystals reinforced poly(oxyethylene). Polymer. 45, 4149–4157 (2004).

Azizi Samir, M. A. S.; Alloin, F.; Dufresne, A. Review of Recent Research into Cellulosic Whiskers, Their Properties and Their Application in Nanocomposite Field. Biomacromolecules 2005, 6 (2), 612–626.

Angell, C. A. Formation of Glasses from Liquids and Biopolymers. Science . 267, 1924– 1935 (1995).

Bouillon, N., Pascault, J. P. & Tighzert, L. Influence of different imidazole catalysts on epoxy-anhydride copolymerization and on their network properties. J. Appl. Polym. Sci. 38, 2103–2113 (1989).

Borrego, L. P., Costa, J. D. M., Ferreira, J. A. M. & Silva, H. Fatigue behaviour of glass fibre reinforced epoxy composites enhanced with nanoparticles. Compos. Part B Eng. 62, 65–72 (2014).

134

Brutman, J. P., Delgado, P. A. & Hillmyer, M. A. Polylactide Vitrimers. ACS Macro Lett. 3, 607–610 (2014).

Capelot, M., Montarnal, D., Tournilhac, F. & Leibler, L. Metal-catalyzed transesterification for healing and assembling of thermosets. J. Am. Chem. Soc. 134, 7664– 7667 (2012).

Chen, M.-Y., Ike, M. & Fujita, M. Acute toxicity, mutagenicity, and estrogenicity of bisphenol-A and other bisphenols. Environ. Toxicol. 17, 80–86 (2002). Commarieu, B. et al. Ultrahigh Tg Epoxy Thermosets Based on Insertion Polynorbornenes. Macromolecules 49, 920–925 (2016).

Czaja, W.; Romanovicz, D.; Brown, R. malcolm. Structural investigations of microbial cellulose produced in stationary and agitated culture. Cellulose 2004, 11 (3), 403–411.

Dowd, R. T. General Chemistry of bisphenol A-based epoxy resins; Bruins, P. F., Ed.; John Wiley and Sons, 1968.

Dong, X. M.; Revol, J.F.; Gray, D. G. Effect of microcrystallite preparation conditions on the formation of colloid crystals of cellulose. Cellulose 1998, 5 (1), 19–32.

Da Silva Perez, D.; Montanari, S.; Vignon, M. R. TEMPO-Mediated Oxidation of Cellulose III. Biomacromolecules 2003, 4 (5), 1417–1425.

Denissen, W., Winne, J. M. & Du Prez, F. E. Vitrimers: permanent organic networks with glass-like fluidity. Chem. Sci. 7, 30–38 (2016).

Denissen, W. et al. Vinylogous Urethane Vitrimers. Adv. Funct. Mater. 25, 2451–2457 (2015).

135

Enns, J. B.; Gillham, J. K. Time–temperature–transformation (TTT) cure diagram: Modeling the cure behavior of thermosets. J. Appl. Polym. Sci. 1983, 28 (8), 2567–2591.

Epoxy Resins: Chemistry and Technology, Second Edi.; May, C. A., Ed.

Eichhorn, S. J. Cellulose nanowhiskers: promising materials for advanced applications. Soft Matter 2011, 7 (2), 303–315.

EWEA. Wind energy – The facts, a guide to the technology, economics and future of wind power . (Earthscan Publications Ltd. Belgium , 2009).

Foix, D.; Ramis, X.; Serra, A.; Sangermano, M. UV generation of a multifunctional hyperbranched thermal crosslinker to cure epoxy resins. Polymer. 2011, 52 (15), 3269– 3276.

Fumagalli, M., Ouhab, D., Boisseau, S. M. & Heux, L. Versatile Gas-Phase Reactions for Surface to Bulk Esterification of Cellulose Microfibrils Aerogels. Biomacromolecules 14, 3246–3255 (2013).

Fourcade, D., Ritter, B. S., Walter, P., Schonfeld, R. & Mulhaupt, R. Renewable resource- based epoxy resins derived from multifunctional poly(4-hydroxybenzoates). Green Chem. 15, 910–918 (2013).

Flore, D. & Wegener, K. Modelling the mean stress effect on fatigue life of fibre reinforced plastics. Int. J. Fatigue 82, 689–699 (2016).

Guhados, G., Wan, W. & Hutter, J. L. Measurement of the Elastic Modulus of Single Bacterial Cellulose Fibers Using Atomic Force Microscopy. Langmuir 21, 6642–6646 (2005).

Grunert, M. & Winter, W. T. Nanocomposites of Cellulose Acetate Butyrate Reinforced

136 with Cellulose Nanocrystals. J. Polym. Environ. 10, 27–30 (2002).

Grand view research. Epoxy Resin Market Size & Share, Industry Report, 2024.

H. H. Winter, et al. Techniques in Rheological Measurement; Collyer, A. ., Ed.; Chapman and Hall: London, 1997.

Habibi, Y.; Lucia, L. A.; Rojas, O. J. Cellulose Nanocrystals: Chemistry, Self-Assembly, and Applications. Chem. Rev. 2010, 110 (6), 3479–3500.

Huang, Y.; Zhu, C.; Yang, J.; Nie, Y.; Chen, C.; Sun, D. Recent advances in bacterial cellulose. Cellulose 2014, 21 (1), 1–30.

Hendriks, B., Waelkens, J., Winne, J. M. & Du Prez, F. E. Poly(thioether) Vitrimers via Transalkylation of Trialkylsulfonium Salts. ACS Macro Lett. 6, 930–934 (2017).

Hwang, P. Y.; Alto, P.; City, R. U.S. Patent 3488404. (1973).

Hove, C. L. F. D.; Heymans, D.; Steinbrecher, C.; Kotlweska, A.; Sand, R. V. U.S. Patent 2014/0316030 A1. (2014).

Harris, B. Fatigue in composites science and technology of the fatigue response of fibre- reinforced plastics . (Elsevier Science Ltd, Boca Raton , 2003).

Hestrin, S. & Schramm, M. Synthesis of cellulose by Acetobacter xylinum. 2. Preparation of freeze-dried cells capable of polymerizing glucose to cellulose. Biochem. J. 58, 345– 352 (1954). http://pire-renew.com.

137 https://www.siemens.com/press/pool/de/feature/2014/energy/2014-03-hull/fact-sheet- b75-rotor-blade-e.pdf.

Isogai, A.; Saito, T.; Fukuzumi, H. TEMPO-oxidized cellulose nanofibers. Nanoscale 2011, 3 (1), 71–85.

Iguchi, M., Yamanaka, S. & Budhiono, A. Bacterial cellulose—a masterpiece of nature’s arts. J. Mater. Sci. 35, 261–270 (2000).

Islam, M. R., Mekhilef, S. & Saidur, R. Progress and recent trends of wind energy technology. Renew. Sustain. Energy Rev. 21, 456–468 (2013).

Jenck, J. F., Agterberg, F. & Droescher, M. J. Products and processes for a sustainable chemical industry: a review of achievements and prospects. Green Chem. 6, 544–556 (2004).

Jin, F.-L.; Park, S.-J. Thermal properties of epoxy resin/filler hybrid composites. Polym. Degrad. Stab. 2012, 97 (11), 2148–2153.

Kloxin, C. J., Scott, T. F., Adzima, B. J. & Bowman, C. N. Covalent Adaptable Networks (CANs): A Unique Paradigm in Cross-Linked Polymers. Macromolecules 43, 2643–2653 (2010).

Kloxin, C. J., Scott, T. F., Park, H. Y. & Bowman, C. N. Mechanophotopatterning on a Photoresponsive Elastomer. Adv. Mater. 23, 1977–1981 (2011).

Kostagiannakopoulou, C., Loutas, T. H., Sotiriadis, G., Markou, A. & Kostopoulos, V. On the interlaminar fracture toughness of carbon fiber composites enhanced with graphene nano-species. Compos. Sci. Technol. 118, 217–225 (2015).

138

Kim, J. A.; Seong, D. G.; Kang, T. J.; Youn, J. R. Effects of surface modification on rheological and mechanical properties of CNT/epoxy composites. Carbon N. Y. 2006, 44 (10), 1898–1905.

Loos, M. R.; Yang, J.; Feke, D. L.; Manas-Zloczower, I. Effect of block-copolymer dispersants on properties of carbon nanotube/epoxy systems. Compos. Sci. Technol. 2012, 72 (4), 482–488.

Leung, D. Y. C. & Yang, Y. Wind energy development and its environmental impact: A review. Renew. Sustain. Energy Rev. 16, 1031–1039 (2012).

Laik, S., Galy, J., Gérard, J.-F., Monti, M. & Camino, G. Fire behaviour and morphology of epoxy matrices designed for composite materials processed by infusion. Polym. Degrad. Stab. 127, 44–55 (2016).

Lu, J., Askeland, P. & Drzal, L. T. Surface modification of microfibrillated cellulose for epoxy composite applications. Polymer. 49, 1285–1296 (2008).

Lee, K. Y., Ho, K. K. C., Schlufter, K. & Bismarck, A. Hierarchical composites reinforced with robust short sisal fibre preforms utilising bacterial cellulose as binder. Compos. Sci. Technol. 72, 1479–1486 (2012).

Lee, K. Y., Aitomäki, Y., Berglund, L. A., Oksman, K. & Bismarck, A. On the use of nanocellulose as reinforcement in polymer matrix composites. Compos. Sci. Technol. 105, 15–27 (2014).

Lugemwa, F. N. Extraction of Betulin, Trimyristin, Eugenol and Carnosic Acid Using Water-Organic Solvent Mixtures. Molecules 17, (2012).

Lu, J., Wang, T. & Drzal, L. T. Preparation and properties of microfibrillated cellulose polyvinyl alcohol composite materials. Compos. Part A Appl. Sci. Manuf. 39, 738–746

139

(2008).

Lehn, J.-M. Dynamers: dynamic molecular and supramolecular polymers. Prog. Polym. Sci. 30, 814–831 (2005).

Moon, R. J., Martini, A., Nairn, J., Simonsen, J. & Youngblood, J. Cellulose nanomaterials review: structure, properties and nanocomposites. Chem. Soc. Rev. 40, 3941–3994 (2011).

Maiorana, A.; Spinella, S.; Gross, R. A. Bio-based alternative to the diglycidyl ether of bisphenol A with controlled materials properties. Biomacromolecules 2015, 16 (3), 1021– 1031.

Maffini, M. V; Rubin, B. S.; Sonnenschein, C.; Soto, A. M. Endocrine disruptors and reproductive health: the case of bisphenol-A. Mol. Cell. Endocrinol. 2006, 254–255, 179– 186.

Mantzaridis, C. et al. Rosin acid oligomers as precursors of DGEBA-free epoxy resins. Green Chem. 15, 3091–3098 (2013).

Maiorana, A., Yue, L., Manas-Zloczower, I. & Gross, R. Structure–property relationships of a bio-based reactive diluent in a bio-based epoxy resin. J. Appl. Polym. Sci. 2016

Mehmet, A., Ertas, M., Nitz, S. & Kollmannsberger, H. Chemical composition and content of essential oil from the bud of cultivated Turkish clove (Syzygium aromaticum L.). BioResources 2, (2007).

Meier, R., Kahraman, I., Seyhan, A. T., Zaremba, S. & Drechsler, K. Evaluating vibration assisted vacuum infusion processing of hexagonal boron nitride sheet modified carbon fabric /epoxy composites in terms of interlaminar shear strength and void content. Compos. Sci. Technol. 128, 94–103 (2016).

140

Margem, F. M., Monteiro, S. N., Bravo Neto, J., Rodriguez, R. J. S. & Soares, B. G. The dynamic-mechanical behavior of epoxy matrix composites reinforced with ramie fibers. MatÃ\copyrightria (Rio Janeiro) 15, 164–171 (2010).

Mathew, A. P. & Dufresne, A. Morphological Investigation of Nanocomposites from Sorbitol Plasticized Starch and Tunicin Whiskers. Biomacromolecules 3, 609–617 (2002).

Mijovic, J.; Wijaya, J. Etherification reaction in epoxy-amine systems at high temperature. Polymer. 1994, 35 (12), 2683–2686.

Morselli, D.; Bondioli, F.; Sangermano, M.; Messori, M. Photo-cured epoxy networks reinforced with TiO2 in-situ generated by means of non-hydrolytic sol–gel process. Polymer. 2012, 53 (2), 283–290.

Naebe, M.; Wang, J.; Amini, A.; Khayyam, H.; Hameed, N.; Li, L. H.; Chen, Y.; Fox, B. Mechanical Property and Structure of Covalent Functionalised Graphene/Epoxy Nanocomposites. Sci. Rep. 2014, 4, 4375.

Omrani, A., Simon, L. C. & Rostami, A. A. Influences of cellulose nanofiber on the epoxy network formation. Mater. Sci. Eng. A 490, 131–137 (2008).

Pascault, J. P. & Williams, R. J. J. In Epoxy Polymers: New Materials and Innovations . (Wiley VCH: Weinheim , 2010).

Pilato, L. A. & Michno, M. J. Acute toxicity, mutagenicity, and estrogenicity of bisphenol- A and other bisphenols. Adv. Compos. Mater.

Patel, A.; Maiorana, A.; Yue, L.; Gross, R. A.; Manas-Zloczower, I. Curing Kinetics of Biobased Epoxies for Tailored Applications. Macromolecules 2016, 49 (15), 5315–5324.

141

Pham, H. Q.; Marks, M. J. Epoxy Resins. In Kirk-Othmer Encyclopedia of Chemical Technology; John Wiley & Sons, Inc., 2000.

Peng, S. X., Moon, R. J. & Youngblood, J. P. Design and characterization of cellulose nanocrystal-enhanced epoxy hardeners. Green Mater. 2, 193–205 (2014).

Petersson, L., Mathew, A. P. & Oksman, K. Dispersion and properties of cellulose nanowhiskers and layered silicates in cellulose acetate butyrate nanocomposites. J. Appl. Polym. Sci. 112, 2001–2009 (2009).

Park, W. H., Lee, J. K. & Kwon, K. J. Cure Behavior of an Epoxy-Anhydride-Imidazole System. Polymer Journal 28, 407–411 (1996).

Rusli, R.; Eichhorn, S. J. Determination of the stiffness of cellulose nanowhiskers and the fiber-matrix interface in a nanocomposite using Raman spectroscopy. Appl. Phys. Lett. 2008, 93 (3), 33111.

Rowan, S. J., Cantrill, S. J., Cousins, G. R. L., Sanders, J. K. M. & Stoddart, J. F. Dynamic Covalent Chemistry. Angew. Chemie Int. Ed. 41, 898–952 (2002).

Raquez, J.-M. et al. Surface-modification of cellulose nanowhiskers and their use as nanoreinforcers into polylactide: A sustainably-integrated approach. Compos. Sci. Technol. 72, 544–549 (2012).

Ruka, D. R., Simon, G. P. & Dean, K. M. Altering the growth conditions of Gluconacetobacter xylinus to maximize the yield of bacterial cellulose. Carbohydr. Polym. 89, 613–622 (2012).

Raquez, J.-M., Deléglise, M., Lacrampe, M.-F. & Krawczak, P. Thermosetting (bio)materials derived from renewable resources: A critical review. Prog. Polym. Sci. 35, 487–509 (2010).

142

Retegi, A. et al. Bacterial cellulose films with controlled microstructure–mechanical property relationships. Cellulose 17, 661–669 (2010).

Silva, H., Ferreira, J. A. M., Capela, C. & Richardson, M. O. W. Mixed Mode interlayer fracture of glass fiber/nano-enhanced epoxy composites. Compos. Part A Appl. Sci. Manuf. 64, 211–222 (2014).

Song, Y. S., Youn, J. R. & Gutowski, T. G. Life cycle energy analysis of fiber-reinforced composites. Compos. Part A Appl. Sci. Manuf. 40, 1257–1265 (2009).

Saito, T.; Kimura, S.; Nishiyama, Y.; Isogai, A. Cellulose Nanofibers Prepared by TEMPO-Mediated Oxidation of Native Cellulose. Biomacromolecules 2007, 8 (8), 2485– 2491.

Sturcová, A.; Davies, G. R.; Eichhorn, S. J. Elastic Modulus and Stress-Transfer Properties of Tunicate Cellulose Whiskers. Biomacromolecules 2005, 6 (2), 1055–1061.

Srivastava, S. K. & Singh, I. P. Hybrid epoxy nanocomposites: lightweight materials for structural applications. Polym J 44, 334–339 (2012).

Sai, H. et al. Surface Modification of Bacterial Cellulose Aerogels’ Web-like Skeleton for Oil/Water Separation. ACS Appl. Mater. Interfaces 7, 7373–7381 (2015).

Saba, N. et al. Thermal and dynamic mechanical properties of cellulose nanofibers reinforced epoxy composites. International Journal of Biological Macromolecules 102, (2017).

Scott, T. F., Schneider, A. D., Cook, W. D. & Bowman, C. N. Photoinduced Plasticity in Cross-Linked Polymers. Science. 308, 1615–1617 (2005).

143

Sengupta, R.; Bhattacharya, M.; Bandyopadhyay, S.; Bhowmick, A. K. A review on the mechanical and electrical properties of graphite and modified graphite reinforced polymer composites. Prog. Polym. Sci. 2011, 36 (5), 638–670.

Solid, L. D.; Hrubesh, L. W.; Chan, H. M.; Grenestedt, J. L.; Harmer, M. P.; Caram, H. S.; Roy, S. K.; Handbook, P. T.; Raton, B.; Ashby, M. F.; et al. Silica-Like Malleable Materials from. Science. 2011, 334 (November), 965–968.

Tang, L.-C. et al. The effect of graphene dispersion on the mechanical properties of graphene/epoxy composites. Carbon. 60, 16–27 (2013).

Tashiro, K.; Kobayashi, M. Theoretical evaluation of three-dimensional elastic constants of native and regenerated celluloses: role of hydrogen bonds. Polymer. 1991, 32 (8), 1516– 1526.

Talreja, R. & Singh, C. V. Damage and failure of composite materials . (Cambridge University Press, Cambridge, New York , 2012).

Tashiro, K. & Kobayashi, M. Theoretical evaluation of three-dimensional elastic constants of native and regenerated celluloses: role of hydrogen bonds. Polymer. 32, 1516–1526 (1991).

Verchere, D.; Sautereau, H.; Pascault, J. P.; Riccardi, C. C.; Moschiar, S. M.; Williams, R. J. J. Buildup of epoxycycloaliphatic amine networks. Kinetics, vitrification, and gelation. Macromolecules 1990, 23 (3), 725–731.

Westman, M. B. and G. Cellulose - Fundamental Aspects and Current Trends. 2015.

Witik, R. A., Payet, J., Michaud, V., Ludwig, C. & Månson, J.-A. E. Assessing the life cycle costs and environmental performance of lightweight materials in automobile applications. Compos. Part A Appl. Sci. Manuf. 42, 1694–1709 (2011).

144

Wojtecki, R. J., Meador, M. A. & Rowan, S. J. Using the dynamic bond to access macroscopically responsive structurally dynamic polymers. Nat Mater 10, 14–27 (2011).

Walsh, C. C. and P. Allgned nanocomposites from cellulose nanocrystals by electrospinning with a soluble polymer followed by thermoset impregnation, Craig Clemons and Phil Walsh , USDA Forest Service , Forest Products Laboratory , Madison , WI Electrospinning and Composite Pr. in ANTEC (2017).

Xie, Y., Hill, C. A. S., Xiao, Z., Militz, H. & Mai, C. Silane coupling agents used for natural fiber/polymer composites: A review. Compos. Part A Appl. Sci. Manuf. 41, 806–819 (2010).

Yu, K., Taynton, P., Zhang, W., Dunn, M. L. & Qi, H. J. Influence of stoichiometry on the glass transition and bond exchange reactions in epoxy thermoset polymers. RSC Adv. 4, 48682–48690 (2014).

Yue, L., Maiorana, A., Patel, A., Gross, R. & Manas-Zloczower, I. A sustainable alternative to current epoxy resin matrices for vacuum infusion molding. Compos. Part A Appl. Sci. Manuf. 100, 269–274 (2017).

Yuexin, D., Zhaoyuan, T., Yan, Z. & Jing, S. Compression Responses of Preform in Vacuum Infusion Process. Chinese J. Aeronaut. 21, 370–377 (2008).

Yue, L., Pircheraghi, G., Monemian, S. A. & Manas-Zloczower, I. Epoxy composites with carbon nanotubes and graphene nanoplatelets – Dispersion and synergy effects. Carbon. 78, 268–278 (2014).

Yang, G.; Rohde, B. J.; Tesefay, H.; Robertson, M. L. Biorenewable Epoxy Resins Derived from Plant-Based Phenolic Acids. ACS Sustain. Chem. Eng. 2016

145

Yue, L. Epoxy composites with hybrid carbon fillers, Case Western Reserve University, 2014.

Zhang, H., Liu, Y., Kuwata, M., Bilotti, E. & Peijs, T. Improved fracture toughness and integrated damage sensing capability by spray coated CNTs on carbon fibre prepreg. Compos. Part A Appl. Sci. Manuf. 70, 102–110 (2015).

146