Development of Bio-Based Phenol Formaldehyde Resol Resins Using Mountain Pine Beetle Infested Lodgepole Pine Barks

Development of Bio-based Phenol Formaldehyde Resol Resins Using Mountain Pine Beetle Infested Lodgepole Pine Barks

Yong Zhao

A thesis submitted in conformity with the requirements for the degree of Doctor of Philosophy Faculty of Forestry University of Toronto

Development of Bio-based Phenol Formaldehyde Resol Resins Using Mountain Pine Beetle Infested Lodgepole Pine Barks

Yong Zhao

Doctor of Philosophy

Faculty of Forestry

University of Toronto

2013

Abstract

Phenol formaldehyde (PF) resol resins have long been used widely as wood adhesives due to their excellent bonding performance, water resistance and durability. With the growing concern for fossil fuel depletion and climate change, there is a strong interest in exploring renewable biomass materials as substitutes for petroleum-based feedstock. Bark, rich in phenolic compounds, has demonstrated potential to partially substitute phenol in synthesizing bio-based PF resins.

In this study, acid-catalyzed phenol liquefaction and alkaline extraction were used to convert mountain pine beetle (MPB; Dendroctonus ponderosae) infested lodgepole pine (Pinus contorta) barks to phenol substitutes, liquefied bark and bark extractives. Two types of bio-based phenol formaldehyde (PF) resol resins, namely liquefied bark-PF resin and bark extractive-PF resins, were then synthesized and characterized.

It was found that acid-catalyzed phenol liquefaction and alkaline extraction were effective conversion methods to obtain phenol substitute with the maximum yield of 85% and 68%, respectively. The bio-based PF resol resins had higher molecular weights, higher

polydispersity indices, shorter gel times, and faster curing rates than the lab synthesized control PF resin without the bark components. Based on the lap-shear tests, the bio-based PF resol resins exhibited comparable wet and dry bonding strength to lab PF resin and commercial PF resin. The post-curing thermal stability of the bio-based PF resins was similar to the lab control PF resin.

The liquid-state 13C nuclear magnetic resonance (NMR) study revealed significant influences on the resin structures by the inclusion of the bark components. Methylene ether bridges, which were absent in the lab PF resin, were found in the bio-based PF resins. The bark components favored the formation of para-ortho methylene linkages in the bio- based bark extractive-PF resins. The liquefied bark-PF resin showed a higher ratio of para-para/ortho-para methylene link (-CH2-), a higher unsubstituted/substituted hydrogen

(-H/-CH2OH) ratio and a higher methylol/methylene (-CH2OH/-CH2-) ratio than the bark extractive-PF resin. Both tannin components of bark alkaline extractives and phenolated barks contributed to the acceleration of the curing rate of the bio-based resins.

This research demonstrated the promise of the bio-based PF resins containing either bark alkaline extractives or liquefied barks as environmentally friendly alternatives to PF adhesives derived solely from fossil fuel based phenol and proposed a novel higher value- added application of the largely available barks from the mountain pine beetle-infested lodgepole pine trees.

iii

Acknowledgements

First and foremost, I would like to express my sincere and profound gratitude to my supervisor, Prof. Ning Yan, for her outstanding and exceptional mentorship. With her inspirational guidance and encouragement, generous support and constant trust and patience, Prof. Yan brought me to the interesting and promising research areas, taught me how to do the research and helped me grow as a scientist. Her ever-lasting dedication to research, endless pursuit of excellence and rigorous working ethnic set up a perfect role model for those want to pursue their career in the academic area like me. I was extremely lucky to be her student and to work with her. Thanks to Prof. Yan, I had a fruitful and joyful PhD student life.

I would also like to give my heartfelt appreciations to my co-supervisor, Mr. Martin W. Feng, a senior research scientist, at FPInnovations-Wood Products Division, for his invaluable guidance, support, and encouragement during the course of my PhD study. I have benefited greatly from his advices.

I am very grateful to my supervisory committee, Prof. Paul Cooper, Prof. Bruce McKague, and Prof. Mohini Sain for their insightful suggestions and recommendations for my thesis study.

I would like to extend my thanks to Dr. Tim Burrow for his help with the NMR measurements and discussions; Dr. Arturo Rodriguez for the FTIR training and help; Dr. Robert Jeng for the MALDI-TOF training; Dr. Syed Abthagir Pitchai Mydeen for his kind, generous and consistent help on the TGA and DSC instruments and other lab work. Mr. Tony Ung for his kind technical support on using the hot press and lab-shear test. Mr. Gireesh Gupta and Dr. Zheng Chen for their generous support and useful discussions. Assistances by other lab members of Prof. Yan’s group at the Faculty of Forestry, University of Toronto are highly appreciated.

I would also like to especially thank Jiang Tang, Mingli Sun, Lei Shen, Feng He, Sheng Dai, Haizheng Zhong, Xuping Sun, Jingjing Li, Nan (Crystal) Wu, Jieming Chen, Xiao Han, Jing Wang, Myungjae Lee, Andrew Avsec, Lip Liew, Susan Frye, Luke Hall, Ethan

McBride, Dana Collins, Brooke Lehman, Adam Palmer, Adam Martin. Their friendship and encouragement are invaluable and important to me. They made my life at the University of Toronto so enjoyable and memorable.

I like to express my sincere thanks to my master thesis supervisor and some old friends in China, including, Prof. Benhua Fei, Prof. Songlin Yi, Prof. Xiaoxu Wang, Prof. Zhaohui Wang, Prof. Wenhua Lu, Prof. Haiqing Ren, Prof. Xia Lang, Prof. Yubo Chai, Ms. Dongkun Wang, Ms. Ying Liu, Ms. Yanhui Huang, Ms. Shuqin Zhang, Mr. Xiaoyu Lu, Mr. Zhenyu Wu and Mr. Xiaodong Zhou for their constant support and encouragement. Special thanks are given to my best friend, Yubo Chai, I am deeply indebt to him for all his help. Despite being far away, their constant support made my life in Canada much easier. I cannot imagine what my life would have been without their help. I will remember their kindness forever.

I would like to extend my utmost gratitude to my parents, my aunts and uncles, and my cousins as well as my other relatives. Their unconditional love, support and encouragement made all my accomplishments possible. To them, I dedicate this thesis.

Table of contents

Abstract ...... ii Acknowledgements ...... iv List of Tables ...... x List of Figures ...... xi List of Acronyms ...... xvi Chapter 1 Introduction ...... 1 1.1 Motivation and Significance ...... 1 1.2 Scope ...... 3 1.3 Hypotheses ...... 3 1.4 Objectives ...... 4 1.5 Thesis overview ...... 5 Chapter 2 Literature review ...... 7 2.1 Bark ...... 7 2.2 Chemical composition ...... 7 2.3 Bark phenolic compounds ...... 8 2.3.1 Lignin ...... 8 2.3.2 Tannins ...... 9 2.4 Reactivity of bark phenolic compounds towards formaldehyde ...... 11 2.4.1 Reactivity of lignin towards formaldehyde ...... 11 2.4.2 Reactivity of tannins ...... 12 2.5 Extraction and adhesive application of phenolic compounds from bark ...... 14 2.5.1 Motivation of using renewable biomass as the feedstock ...... 14 2.5.2 Extraction of phenolic compounds from bark for phenolic resin synthesis ..... 15 2.5.3 Adhesive application of bark phenolic compounds ...... 20 2.6 Mountain pine beetle infested lodgepole pine ...... 25

2.7 Conclusions ...... 26 Chapter 3 Experimental Approach ...... 28 3.1 Bark phenolic compounds extraction ...... 28 3.1.1 Bark phenol liquefaction ...... 28 3.1.2 Bark alkaline extraction ...... 30 3.2 Bio-based PF resin synthesis ...... 32 3.2.1 Liquefied bark-PF resin ...... 32 3.2.2 Bark extractive-PF resin ...... 32 3.2.3 Lab PF and commercial PF resins ...... 33 3.3 Characterization of bio-based PF resins ...... 33 3.3.1 pH, viscosity, solids content, gel time, Mw and Mw/Mn of the bio-based PF resins ...... 33 3.3.2 Differential scanning calorimetry (DSC) analysis of the bio-based PF resins . 34 3.3.3 Thermal gravimetric analysis (TGA) of the cured resins ...... 34 3.3.4 Evaluation of resins bonding strength ...... 37 3.3.5 Liquid-state 13C NMR measurement of bark phenolic compounds and bio-based PF resins ...... 38 Chapter 4 Bark Phenol Liquefaction ...... 39 4.1 Abstract ...... 39 4.2 Introduction ...... 39 4.3 Results and discussion ...... 42 4.3.1 Effect of reaction conditions on the liquefaction yield and free phenol content42 4.3.2 Effect of reaction conditions on the bark liquefaction residues ...... 51 4.3.3. Effect of reaction conditions on the morphology of the residues ...... 59 4.4 Conclusions ...... 64 Chapter 5 Characterization of Phenol Formaldehyde Resins Derived from Liquefied Lodgepole Pine Barks ...... 66 5.1 Abstract ...... 66

vii

5.2. Introduction ...... 66 5.3 Results and Discussion ...... 68 5.3.1 Basic properties of the liquefied bark-PF resins ...... 68 5.3.2 Adhesives curing behavior ...... 70 5.3.3 Thermal stability of the cured resins ...... 81 5.3.4 Lap shear test results ...... 82 5.4 Conclusions…………………………………...……………………….…………..83 Chapter 6 Bark Extractions and Bark Extractive-PF resins ...... 85 6.1 Abstract ...... 85 6.2 Introduction ...... 85 6.3 Results and Discussion ...... 88 6.3.1 Properties of the extractives ...... 88 6.3.2 Properties of the bark extractive-PF resins ...... 89 6.3.3 Adhesive curing behavior and curing kinetics ...... 90 6.3.4 Thermal stability ...... 94 6.3.5 Bonding strength ...... 96 6.3.6 FTIR of the bark extractives and bark extractive-PF resins ...... 98 6.4 Conclusions ...... 103 Chapter 7 Thermal Degradation Process of the Cured Bio-based PF Resins ...... 104 7.1 Abstract ...... 104 7.2 Introduction ...... 104 7.3 Results and discussion ...... 106 7.3.1Thermal stability of the bio-based PF resins ...... 106 7.3.2 Thermal degradation kinetics ...... 109 7.3.3 Structural changes of the bio-based PF resins during thermal degradation ... 114 7.4 Conclusions ...... 122 Chapter 8 Liquid-State 13C NMR Study of Bio-based PF Resins from Mountain Pine Beetle Infested Lodgepole Pine ...... 124

viii

8.1 Abstract ...... 124 8.2 Introduction ...... 124 8.3 Results and discussion ...... 126 8.3.1 Liquid-state 13C NMR spectrum of the bark alkaline extractives ...... 126 8.3.2 Liquid-state 13C NMR spectrum of liquefied bark ...... 129 8.3.3 Liquid-state 13C NMR spectrum of lab-made phenol formaldehyde resin ..... 131 8.3.4 Liquid-state 13C NMR spectrum of bark extractive-phenol formaldehyde resins ...... 133 8.3.5 Liquid-state 13C NMR spectrum of liquefied bark-phenol formaldehyde resins ...... 139 8.3.6 Relationship between the resin molecular structure and curing performance 141 8.4 Conclusions ...... 143 Chapter 9 Conclusions and Future work ...... 145 9.1 Major contributions ...... 145 9.2 Future work ...... 147 References ...... 149

List of Tables

Table 2.1 General chemical composition in wood and bark for hardwoods and softwoods [4] ...... 8 Table 4.1 Chemical composition of bark ...... 41 Table 5.1 Resin properties ...... 69 Table 5.2 Resin curing temperatures ...... 72 Table 5.3 Dynamic cure kinetic parameters for different resins ...... 73 Table 6.1 Molecular weight of the extractives ...... 89 Table 6.2 Properties of different resins (Lab PF: Laboratory made PF resin; Com PF: Commercial PF resin; BEPF: Bark extractive-PF resins made from alkaline extractives; BEPF (N): Bark extractive-PF resins made from acid-neutralized extractives) ...... 90 Table 6.3 Curing characteristics of bark extractive-PF resins made from alkaline extractives (BEPF) ...... 91 Table 6.4 Curing characteristics of bark extractive-PF resins made from acid-neutralized extractives (BEPF (N)) ...... 91 Table 6.5 Cure kinetic parameters for different resins (Lab PF: Laboratory made PF resin, Com PF: Commercial PF resin, BEPF: Bark extractives-PF resins made from alkaline extractives, BEPF (N): Bark extractives-PF resins made from acid-neutralized extractives, dα/dt: Conversion rate, r: Correlation coefficient) ...... 93 Table 6.6 Assignment of peaks in FTIR spectra ...... 102 Table 7.1 Weight loss of different resins during the thermal degradation process .... 109 Table 7.2 Activation energy of the tested resins calculated by the Kissinger method 114 Table 7.3 The peak assignment of the FTIR spectra of all tested resins ...... 119 Table 8.1 Assignment of chemical shifts for bark extractives [118,135,136] ...... 128 Table 8.2 Assignment of chemical shifts for liquefied bark [32-35] ...... 131 Table 8.3 Assignment of chemical shifts for PF resin and bio-based PF resins [123-127, 137] ...... 137 Table 8.4 The ratio of the relevant functional groups related to phenolic rings ...... 138 Table 8.5 Summary of the major different characteristics of resins in the 13C NMR spectrum ...... 140

List of Figures

Figure 1.1 Thesis overview and organization of thesis chapters ...... 6 Figure 2.1 Lignin precursors ...... 9 Figure 2.2 Flavonoid unit ...... 10 Figure 2.3 Wattle and Pine tannin unit ...... 11 Figure 2.4 Cross-linking of lignin by formaldehyde ...... 12 Figure 2.5 Reaction positions on the tannin molecules ...... 12 Figure 2.6 Sulfitation of tannins ...... 13 Figure 2.7 Fortification of tannin-based adhesives ...... 14 Figure 2.8 Typical liquefied products from acid-catalyzed phenol liquefaction of cellulose [32] ...... 18 Figure 2.9 Typical liquefied products from acid-catalyzed phenol liquefaction of lignin [33, 34] ...... 19 Figure 3.1 Lap shear specimen dimensions ...... 37 Figure 4.1 Effect of reaction temperature on the residue ratio ...... 43 (Phenol/bark ratio = 3, catalyst = 3 wt.% of phenol, reaction time = 60 min.) ...... 43 Figure 4. 2 Effect of reaction temperature on the free phenol content ...... 44 (Phenol/bark ratio = 3, catalyst = 3 wt. % of phenol, reaction time = 60 min.) ...... 44 Figure 4.3 Effect of phenol/bark ratio on the residue ratio ...... 45 (Reaction temperature=150°C, catalyst = 3 wt.% of phenol, reaction time =60 min.) . 45 Figure 4.4 Effect of phenol/bark ratio on the free phenol content ...... 46 (Reaction temperature=150°C, catalyst = 3 wt.% of phenol, reaction time =60 min.) . 46 Figure 4.5 Effect of reaction time on the residue ratio ...... 47 (Reaction temperature=150°C, catalyst =3 wt.% of phenol, phenol/bark ratio = 3) ..... 47 Figure 4.6 Effect of reaction time on the free phenol ...... 48

(Reaction temperature=150°C, catalyst = 3wt.% of phenol, phenol/bark ratio = 3) ..... 48 Figure 4.7 Effect of catalyst loading on the residue ratio ...... 49

(Reaction temperature=150°C, phenol/bark ratio = 3, reaction time = 60 min.) ...... 49 Figure 4.8 Effect of catalyst loading on the free phenol content ...... 50 (Reaction temperature=150°C, phenol/bark ratio = 3, reaction time =60 min.) ...... 50 Figure 4.9 Free phenol content of the liquefied bark fraction ...... 51 Figure 4.10 Chemical composition of residues from different reaction temperature .... 52 Phenol/bark ratio = 3, catalyst = 3 wt.% of phenol, reaction time = 60 min...... 52 Figure 4.11 FTIR of residues from different reaction temperatures ...... 53 (Phenol/bark ratio = 3, catalyst = 3 wt.% of phenol, reaction time = 60 min.) ...... 53 Figure 4.12 Chemical composition of the residues in relation to different phenol/bark ratios ...... 54 (Reaction temperature=150°C, catalyst = 3 wt.% of phenol, reaction time =60 min.) . 54 Figure 4.13 FTIR of residues from different phenol/bark ratios ...... 55 (Reaction temperature=150°C, catalyst = 3 wt.% of phenol, reaction time =60 min.) . 55 Figure 4.14 Chemical composition of residues from different reaction time ...... 56 (Reaction temperature=150°C, catalyst = 3 wt.% of phenol, phenol/bark ratio = 3) .... 56 Figure 4.15 FTIR of residues from different reaction time ...... 57 (Reaction temperature=150°C, catalyst = 3 wt.% of phenol, phenol/bark ratio = 3) .... 57 Figure 4.16 Chemical composition of the residues from different catalyst loadings .... 58 (Reaction temperature=150°C, phenol/bark ratio = 3, reaction time =60 min.) ...... 58 Figure 4.17. FTIR of residues from different catalyst loadings ...... 59 Figure 4.18 SEM of residues from different catalyst loadings ...... 60 (a) Original beetle infested pine bark. (b) Residues from bark liquefaction with 1% of catalyst loading. (c) Residues from bark liquefaction with 5% of catalyst loading. (d) Residues from bark liquefaction with 7% of catalyst loading. Reaction temperature=150°C, phenol/bark ratio = 3, reaction time =60 min...... 60 Figure 4.19 SEM of residues from different reaction temperatures ...... 61 (a) Residues from bark liquefaction at 120°C. (b) Residues from bark liquefaction at 150°C. (c) Residues from bark liquefaction at 180°C. Phenol/bark ratio = 3, catalyst = 3 wt.% of phenol, reaction time = 60 min...... 61 Figure 4.20 SEM of residues from different reaction time ...... 61

xii

(a) Residues from bark liquefaction at 30min. (b) Residues from bark liquefaction at 60min. (c) Residues from bark liquefaction in 120min. (d) Residues from bark liquefaction at 150min. Reaction temperature=150°C, catalyst = 3 wt.% of phenol, phenol/bark ratio = 3 ...... 61 Figure 4.21 SEM of residues from different phenol/bark ratios ...... 62 (a) Residues from bark liquefaction with phenol/bark ratio = 2. (b) Residues from bark liquefaction with phenol/bark ratio = 3. (c) Residues from bark liquefaction with phenol/bark ratio = 4. (d) Residues from bark liquefaction with phenol/bark ratio = 5. Reaction temperature=150°C, catalyst = 3 wt.% of phenol, reaction time =60 min. .... 62 Figure 4.22 TGA of beetle infested lodgepole pine bark ...... 63 Figure 4.23 TGA of residues from different catalyst loading ...... 64 (Reaction temperature=150°C, phenol/bark ratio = 3, reaction time =60 min.) ...... 64 Figure 5.1 Dynamic DSC curves of the liquefied non-infested lodgepole pine bark-PF resin (LGP-PF) ...... 70 Figure 5.2 Conversion rate as a function of the degree of conversion at various isothermal temperatures for the liquefied mountain pine beetle infested lodgepole pine bark-PF resin (LBI-PF) ...... 74 Figure 5.3 Conversion rate as a function of the degree of conversion at various isothermal temperatures for the liquefied non-infested lodgepole pine bark-PF resin (LGP-PF) ...... 75 Figure 5.4 Conversion rate as a function of the degree of conversion at various isothermal temperatures for the lab PF resin ...... 76 Figure 5.5 Conversion rate as a function of the degree of conversion at various isothermal temperatures for the commercial PF resin ...... 77 Figure 5.6 Conversion as a function of cure time at various isothermal temperatures for the liquefied mountain pine beetle infested lodgepole pine bark-PF resin (LBI-PF) .... 78 Figure 5.7 Conversion as a function of cure time at various isothermal temperatures for the liquefied non-infested lodgepole pine bark-PF resin (LGP-PF) ...... 79 Figure 5.8 Conversion as a function of cure time at various isothermal temperatures for the lab PF resin ...... 80 Figure 5.9 Conversion as a function of cure time at various isothermal temperatures for the commercial PF resin ...... 81 Figure 5.10 Thermal stability of different cured resins as measured by TGA ...... 82

xiii

Figure 5.11 Shear strength of lap shear specimens bonded with different types of adhesives (LBI-PF: Liquefied mountain pine beetle infested pine bark-PF resin; LGP- PF: Liquefied non-infested pine bark-PF resin; Com PF: Commercial PF resin) ...... 83 Figure 6.1 Thermal stability of post-cured BEPF resins (Lab PF: Laboratory made PF resin, Com PF: Commercial PF resin, BEPF: Bark extractive-PF resins made from alkaline extractives,) ...... 95 Figure 6.2 Thermal stability of post-cured BEPF (N) resins (Lab PF: Laboratory made PF resin, Com PF: Commercial PF resin, BEPF (N): Bark extractives-PF resins made from acid-neutralized extractives) ...... 96 Figure 6.3 Shear strength of lap shear specimens bonded with different types of adhesives (Lab PF: Laboratory made PF resin, Com PF: Commercial PF resin, BEPF: Bark extractive-PF resins made from alkaline extractives, BEPF (N): Bark extractive-PF resins made from acid-neutralized extractives) ...... 97 Figure 6.4 FTIR spectra of bark extractives ...... 99 Figure 6.5 FTIR spectra of the bark extractive-PF resins made from alkaline extractives (BEPF) ...... 100 Figure 6.6 FTIR spectra of the bark extractive-PF resins made from acid-neutralized extractives (BEPF (N)) ...... 101 Figure 7.1 Thermal degradation curves of the resins ...... 107 Figure 7.2 Derivative thermal degradation (DTG) curves of the resins ...... 108 Figure 7.3 Isoconversional plot of Flynn-Wall-Ozawa method for the bark extractive-PF resin ...... 110 Figure 7.4 Isoconversional plot of Flynn-Wall-Ozawa method for the liquefied bark-PF resin ...... 111 Figure 7.5 Isoconversional plot of Flynn-Wall-Ozawa method for the commercial PF resin ...... 111 Figure 7.6 Isoconversional plot of Flynn-Wall-Ozawa method for the lab PF resin .. 112 Figure 7.7 Activation energy of the tested resins at different conversion levels calculated by the Flynn-Wall-Ozawa method ...... 113 Figure 7.8 Activation energy of the tested resins at different conversion levels calculated by the Kissinger-Akahira-Sunose method ...... 113 Figure 7.9 FTIR of the bark extractive-PF resin at different degradation temperatures ...... 115 Figure 7.10 FTIR of the liquefied bark-PF resin at different degradation temperatures ...... 116

xiv

Figure 7.11 FTIR of the commercial PF resin at different degradation temperatures . 117 Figure 7.12 FTIR of the lab PF resin at different degradation temperatures ...... 118 Figure 7.13 Possible condensation reactions during the thermal degradation process 120 Figure 7.14 Possible oxidation reactions during the resin thermal degradation process ...... 121 Figure 7.15 Possible reactions for urea decomposition ...... 122 Figure 8.1 Liquid-state 13C NMR spectrum of the bark alkaline extractives ...... 127 Figure 8.2 Liquid-state 13C NMR spectrum of the liquefied bark ...... 130 Figure 8.3 Liquid-state 13C NMR spectrum of the lab PF resin ...... 132 Figure 8.4 Liquid-state 13C NMR spectrum of bark extractive-PF resin with 30 wt% phenol substitution rate by bark extractives ...... 134 Figure 8.5 Liquid-state 13C NMR spectrum of bark extractive-PF resin with 50 wt% phenol substitution rate by bark extractives ...... 135 Figure 8.6 Liquid-state 13C NMR spectrum of bark extractive-PF resin with 70 wt% phenol substitution rate by bark extractives ...... 136 Figure 8.7 Liquid-state 13C NMR spectrum of liquefied bark-PF resin ...... 139 Figure 8.8. Possible liquefied products from acid-catalyzed phenol liquefaction of bark ...... 143

List of Acronyms

ATR Attenuated total reflectance

BEPF Bark extractive phenol formaldehyde resin made from alkaline extractives

BEPF (N) Bark extractive phenol formaldehyde resin made from neutralized extractives

BSA Benzenesulfonic acid

BWT Boiling water test

COM PF Commercial phenol formaldehyde

DHB 2, 5-dihydroxybenzoic acid

DMSO-d6 Deuterated dimethyl sulfoxide

DSC Differential scanning calorimetry

EXO Exothermic

FTIR Fourier transfer infrared spectroscopy

GG Guaiacylglyerol-β-guaiacyl ether

GPC Gel permeation chromatography

HPLC High performance liquid chromatography

ICTAC International confederation for thermal analysis and calorimetry

LBI-PF Liquefied bark phenol formaldehyde resin made from mountain pine beetle infested lodgepole pine

LBPF Liquefied bark phenol formaldehyde

xvi

LGP-PF Liquefied bark phenol formaldehyde resin made from lodgepole pine without beetle infestation

LVL Laminated veneer lumber

MALDI-TOF/TOF Matrix-assisted laser desorption/ionization-time-of-flight/time-of- flight spectrometer

Mn Number average molecular weight

MPB Mountain pine beetle

MSA Methanesulfonic acid

Mw Weight average molecular weight

NDSA 1,5-naphthalenedisulfonic acid

NMR Nuclear magnetic resonance

NSA 1-naphthalenesulfonic acid

OSB Oriented strandboard

PF Phenol formaldehyde

PRF Phenol resorcinol formaldehyde

PTSA p-toluenesulfonic acid py-MBMS pyrolysis molecular weight beam spectrometer

RF Resorcinol formaldehyde

SA Sulfuric acid

SEM Scanning electronic microscopy

TGA Thermal gravimetric analysis

xvii

TMS Tetramethylsilane

TPF Tannin phenol formaldehyde

UF Urea formaldehyde

UV Ultraviolet

WSAD Water-soaking-and-drying

xviii

Chapter 1

Introduction

1.1 Motivation and Significance

Phenol formaldehyde (PF) resol resins, produced through addition and condensation reactions between phenol and formaldehyde under alkaline conditions, have been widely used as wood adhesives for production of many wood composites, such as plywood, oriented strand board (OSB), and laminated veneer lumber (LVL), due to their excellent bonding performance, water resistance and durability [1]. Commercially, PF resin has a large commercial market value: in 2010 PF resin markets were valued at $10 billion USD worldwide, with a $2.3 billion market in North America alone [2]. It is expected that the application of PF resol resins as wood adhesives will continue to hold the largest share of worldwide phenolic resins (~33%), and will drive phenolic resin consumption on a global scale [2].

Currently, the primary raw material used in commercial PF resins synthesis is phenol, a substance derived from fossil fuel resources (petroleum and coal). With growing environmental concerns for fossil fuel depletion and climate change, there is a strong interest in exploring renewable materials as alternative feedstock for PF resin production.

Tree barks, as a renewable non-food based biomass material, are generally available in large quantities as waste residues from forest mills. They are either disposed of directly or burned with other woody residues for heat recovery in mills. The heating value of bark is low in comparison to fossil fuels and the heating value drops dramatically when bark is wet. Similar to wood, bark consists mostly of cellulose, hemicellulose and lignin. But bark contains more extractives and phenolic compounds, such as tannins, resin acids, lipids, fatty acids, sterols, and terpenes than wood [3]. The high proportion of the phenolic compounds with the reactive phenolic moiety in bark, such as lignin and tannin, makes

bark a potentially attractive partial substitute to petroleum-based phenol for making PF resins. Furthermore, the polysaccharides in bark, such as cellulose and hemicellulose, are compatible with the phenolic resin structure suggesting bark components could offer greater flexibility and reduced rigidity in the resulting bio-based PF resins [1]. However, the highly species-dependent and variable chemical structure and reactivity of the bark phenolic compounds pose significant challenges to their commercial adaptations in adhesives.

Recently, due to the rampant mountain pine beetle (MPB; Dendroctonus ponderosae) infestation outbreaks in forests of western Canada, there is a large amount of beetle infested lodgepole pine resource available for utilization. At its peak outbreak in 2007, MPB affected approximately 10 million hectares of pine forests: an area representing about 15% of all forested lands in British Columbia. Lodgepole pine (Pinus contorta) trees are the main host of MPB, and therefore populations of this species experienced especially high mortality rates [62]. Given the large amount of readily available MPB-infested lodgepole pine bark biomass, converting or extracting bark to obtain phenol partial substitute in PF resin synthesis could be highly advantageous for the forest industry in both short and long terms. However, no previous research has investigated the adhesive application of the mountain pine beetle infested lodgepole pine barks.

In order to use barks from the mountain pine beetle infested lodgepole pine as part of raw materials for PF resin synthesis, suitable techniques for extracting the desirable bark components or converting bark components to phenol partial substitute need to be developed. In addition, an appropriate resin synthesis procedure has to be formulated to ensure the bio-based PF resins have comparable bonding performance and bonding characteristics to existing commercial PF resins. On the other hand, incorporating bark components into the PF resin synthesis, even only as partial replacement for petroleum- based phenol, can add significant complexity to the resin structure and affect the reactivity and performance of the resulting resins. Therefore, a better fundamental understanding of the impact of bark components on structures and properties of the bio-based PF resins is highly necessary.

1.2 Scope

In this thesis study, mountain pine beetle infested lodgepole pine bark was used as the raw material; acid-catalyzed phenol liquefaction and alkaline extraction were the two methods examined for obtaining the suitable phenol substitute from bark. The acquired liquefied bark and bark alkaline extractives were then used to partially replace petroleum-based phenol for making bio-based PF resol resins including both liquefied bark-PF resin and bark extractive-PF resin. The scope of the thesis research covers the following aspects:

1. To understand the influence of reaction conditions on the acid-catalyzed bark phenol liquefaction process and properties of the liquefied products.

2. To identify structural differences in the products obtained from the acid-catalyzed bark phenol liquefaction and bark alkaline extraction.

3. To investigate the impact of bark components on the structure and bonding and curing properties of the bio-based PF resol resins.

4. To relate the molecular structure characteristics of the bio-based PF resol resins to the resin curing performance

1.3 Hypotheses

The following hypotheses were addressed in this thesis:

1. Bark components can be used as or converted into suitable phenol substitute to partially replace petroleum-based phenol in PF resol resin formulation.

2. Both acid-catalyzed phenol liquefaction and alkaline extraction are effective methods to obtain desirable bark components, which could be used for formulating bio-based PF resol resins with good performance.

3. Bark components will not negatively affect the bonding strength of the resulting bio- based PF resol resin at appropriate phenol substitution levels, and instead, they can accelerate the resin curing rate due to the higher reactivity towards formaldehyde.

1.4 Objectives The overall objectives of this thesis are to utilize bark from mountain pine beetle infested lodgepole pine as a partial substitute of petroleum-based phenol, to formulate bio-based phenol formaldehyde resol resins using the bark components, and to obtain a mechanistic understanding of the impact of bark components on the structure and performance of the bio-based PF resol resins. The following specific set of objectives was identified:

1. To convert bark into suitable phenol substitutes by either acid-catalyzed phenol liquefaction or alkaline extraction

o To evaluate the effects of liquefaction reaction conditions on the liquefaction yield of bark phenol liquefaction, free phenol content in the liquefied bark fraction and liquefied residues.

o To evaluate the properties of the bark extractives from bark alkaline extraction.

o To investigate the characteristics of the liquefied bark and bark extractives.

2. To formulate and characterize the bio-based bark PF resol resins.

o To formulate the liquefied bark-PF resin using the liquefied bark fraction and to investigate the resin performance.

o To formulate the bark-extractive-PF resin using the bark alkaline extractives and to study the resin properties at various phenol substitution levels.

3. To investigate the relationship between the structure and performance of the bio-based bark PF resins

o To study the impact of bark components on the thermal stability and thermal degradation kinetics of the bio-based PF resins

o To study the molecular structure of bio-based PF resins in relation to the resin curing behavior.

1.5 Thesis overview

This thesis is divided into nine chapters. Chapter 1 introduces thesis scope and objectives. Chapter 2 reviews previous literature findings on adhesive application of relevant renewable biomass materials. Unexplored research areas and knowledge gaps related to bark utilization in adhesives are also summarized. Chapter 3 gives a detailed description of the materials, experimental methods, and equipment used in the study, including the acid- catalyzed bark phenol liquefaction and bark alkaline extraction methods and bio-based PF resol resins preparation and characterization techniques.

Chapter 4 and Chapter 5 focus on acid-catalyzed bark phenol liquefaction and characterization results of the liquefied bark-PF resin. Chapter 6 explores the application of bark alkaline extractives in the resin synthesis and the properties of the resulting bark- extractive PF resins. Chapter 7 investigates the thermal degradation characteristics of the bio-based bark PF resins. In Chapter 8, the relationship between structure and curing performance of the bio-based bark PF resins is discussed. Finally, Chapter 9 summarizes the main conclusions of this thesis and provides recommendations for future studies. The main thesis topics and their interrelationship are summarized in Figure 1.1.

• Introduction & background Introduction & literature review • Literature review (Chapter 1, Chapter 2) • Significance & objectives & hypothesis

Bark phenol liquefaction, alkaline extraction • Bark liquefaction and its influencing factors & products characterization • Bark alkaline extraction and bark extractives (Chapter 3, Chapter 4)

• Liquefied bark-PF resins synthesis and Bark-derived PF resins synthesis and characterization characterization • Bark extractive-PF resins synthesis and (Chapter 5, Chapter 6) characterization ( bonding strength, curing behavior…)

Fundamental understanding on the resins • Thermal degradation characteristics of bark- structure & performance derived PF resins (Chapter 7, Chapter 8) • Structural study of bark-derived PF resins (13C NMR)

Summary, conclusions and future work • Summary and conclusions (Chapter 9) • Future work

Figure 1.1 Thesis overview and organization of thesis chapters

Chapter 2

Literature review

2.1 Bark

Tree bark usually refers to all tissues external to and surrounding the vascular cambium of the tree. Generally it takes up about 9-15% of a typical log by volume [4]. In spite of its small proportion, bark plays an important role in a living tree. Bark has a complex anatomy and complicate chemical composition in order to maintain three main functions: (1) providing nutrient transport from the leaves to the rest of the tree, (2) protecting the sensitive inner cambium from desiccation, and (3) shielding from the environment as the primary defense of the tree against wildfire, mechanical injuries caused by heavy wind, and attacks by phytopathogens, phytophagous insects, larger animals, etc. [3]. In the wood industry, bark is a residue in forest mill operations and it is either abandoned or burned as a fuel.

2.2 Chemical composition

Bark has a similar chemical composition to wood, but it contains more extractives, a higher lignin content, and a smaller amount of holocellulose (Table 2.1). The bark extractives can be classified into polar and non-polar extractives. The amount of polar extractives, including flavonoids, phenolics, glycosides, tannins, sugars, etc, are generally three to five times more than that of the non-polar compounds, including waxes, resin acids, lipids, fatty acids, sterols, terpenes, etc. The bark chemical components can be separated into fractions with different polarity through sequential extraction using a series of organic solvents and hot water. To analyze lignin and polysaccharides, bark is usually first extracted using aqueous alkali solutions; followed by hydrolysis of the extractive-free bark by sulfuric acid to give Klason lignin and acid insoluble residue contents [3, 4].

Table 2.1 General chemical composition in wood and bark for hardwoods and softwoods [4]

Weight (%) Softwoods Hardwoods

Wood Bark Wood Bark

Lignin (%)* 25-30 40-55 18-25 40-50

Polysaccharides (%)* 66-72 30-48 74-80 32-45

Extractives (%) 2-9 2-25 2-5 5-10

Ash (%) * 0.2-0.6 Up to 20 0.2-0.6 Up to 20

* Based on extractive-free material

2.3 Bark phenolic compounds

2.3.1 Lignin

The polymeric phenolic part of bark is mostly composed of lignin, which is often described as a complex three-dimensional polymer comprised of various linked phenylpropenoid units. In nature, it has a primary role to bind cellulose chains within the ultra-structure of plant and wood fiber of the cell walls. Softwood and hardwood lignins may be distinguished by the presence of an additional methoxy group in the ortho-position of the phenyl rings [5].

Softwood lignin is composed mainly of guaiacyl units originating from the predominant precursor, trans-coniferyl alcohol (Figure 2.1 (a)), while hardwood lignin is composed of both guaiacyl and syringyl units derived from transconiferyl and trans-sinapyl alcohols (Figure 2.1 (b)), respectively. In general, hardwood bark lignins are mainly composed of syringyl, guaiacyl and small amounts of p-hydroxyphenyl nuclei while softwood bark lignins have quite similar composition of syringyl-guaiacyl ratio but differ in higher proportion of p-hydroxyphenyl units, which stem from trans-p-coumaryl alcohol (Figure 2.1 (c)) [3].

OH OH OH

OMe MeO OMe OH OH OH

(a) (b) (c)

Figure 2.1 Lignin precursors

2.3.2 Tannins

Tannins, with the molecular weights ranging from 500 to over 3000, are another important type of natural polyphenolic compounds present in a relatively large quantity in coniferous tree barks. The barks of some hardwood species, such as Quercus, Eucalyptus, Acacia, and Salix also contain a large amount of tannin extractives. Tannin could be classified into hydrolysable tannin and condensed tannin based on its structure and properties [1, 5].

2.3.2.1 Hydrolysable tannins

The hydrolysable tannins can be considered polyesters derived from glucose, which can be categorized into: (1) gallotannins, which release gallic acid and its derivatives when submitted to acid hydrolysis. (2) ellagitannins, which upon hydrolysis release ellagic acid and valonic acid. Caustic hydrolysis of resorcinolic tannin has been reported to cleave the inter-flavonoid bond and open the etherocyclic ring joining the A and B ring of the flavonoid unit (Figure 2.2). Acid hydrolysis has been shown to easily open the heterocyclic ring of polyflavonoids with the formation of carbonation, which is capable of reacting with another nucleophile present [1, 5].

(OH)

8 1 B HO O 7 OH A 2 3 6 5 4 OH (OH)

Figure 2.2 Flavonoid unit

2.3.2.2 Condensed tannins

The condensed tannins consist of flavonoid units, known as flavan-3ols (catechins) and flavan-3, 4-diols (leucoantocyanidins). With the average degree of condensation ranging from 4 to 12 flavonoid units, it is commonly present as polymer and does not undergo hydrolysis. The condensed tannins constitute more than 90% of the total world production of commercial tannins.

The main structure of tannin extractives from quebracho, mimosa (black wattle), hemlock and Douglas-fir bark is claimed to be constituted predominantly by four to six linked flavonoid units where the A-ring is of resorcinol type and B-ring of pyrogallol type units, with a small proportion of flavonoid units consisting of resorcinol A-ring and catechol B- ring. In pine (radiata, eliotae, taeda, aleppensis, sylvestris, patula, pinaster, etc) species, the flavonoid units are of phloroglucinol A-ring and catechol B-ring (catechin-type) typically linked by four to eight bonds in large proportion, with flavonoid units of phloroglucinol A-ring and phenol B-ring in a much lower proportion. The structure of the main polymeric constituents of wattle and pine tannins is shown in Figure 2.3 [1, 5].

(OH) (OH) OH OH B B HO O HO O (OH) OH A A

OH Wattle tannin unit Pine tannin unit Figure 2.3 Wattle and Pine tannin unit

Besides the flavonoid units, the non-tannin parts including carbohydrates, hydrocolloid gums and amino and imino acid fractions also exist in the bark tannin extractives. The hydrocolloid gums with hydrophilicity varying in concentration from 3 to 6% contribute significantly to the viscosity of the extractives in spite of their low concentration [1, 5].

2.4 Reactivity of bark phenolic compounds towards formaldehyde

2.4.1 Reactivity of lignin towards formaldehyde

Lignin can be cross-linked either intra- or inter-molecularly by either condensation or oxidative coupling. The reaction of lignin with formaldehyde is considered similar to phenol formaldehyde chemistry; methylolation and methylol condensation reactions occur during the reaction (Figure 2.4). However, the complexity of lignin structure reduces its reactivity. The presence of methoxy groups (adjacent to hydroxyls), steric hindrance and only 0.5 free position per C9 unit (phenylpropyl group) lead to higher reaction temperatures and longer cross-linking times. Most lignin sourced and used in adhesives is described as the spent sulfite liquor (SSL) waste products from pulp and paper mills [6], the reactivity of the lignin towards formaldehyde also highly depends on the lignosulfonate metal salt used (Na+, Mg2+, Ca2+ ) [7].

C 3 C 3 C3 C3 OMe + - H or OH OH CH2O MeO MeO CH2OH MeO CH2 OMe OR OR OR OH

Figure 2.4 Cross-linking of lignin by formaldehyde

2.4.2 Reactivity of tannins

Tannins from different sources exhibit different properties and reactivities, which are primarily attributed to the hydroxyl substitution patterns of different tannin extracts. The nucleophilic centers on the A-rings of flavonoid units tend to be more reactive than those found on the B-rings. Formaldehyde reacts with tannins to produce polymerization through methylene bridge linkages to reactive positions of the flavonoid molecules, mainly the A-rings. The reaction positions available on the A-rings are the 8-position of all the resorcinolic flavonoid units and the 6-position of all the phloroglucinolic flavonoid units (Figure 2.5) [1].

(OH) (OH)

OH OH

8 B B HO O HO O OH OH A A 6

Figure 2.5 Reaction positions on the tannin molecules

Resorcinolic A-rings show slightly lower reactivity towards formaldehyde compared to that of resorcinol. Phloroglucinolic A-rings behave instead as phloroglucinol. Pyrogallol or catechol B-rings are less reactive than A-rings, but they can be activated by anion formation at relatively high pH values. In general, only A-rings are used to crosslink the network of the tannin adhesives, B-rings only participate the reaction at high pH conditions (pH>10). The high reactivity of A-rings towards formaldehyde under strong

alkaline conditions will cause the short pot-lives of the resulting resins. In addition, due to their size and shape, tannin molecules become immobile at a low level of condensation with formaldehyde, the available reactive sites are too far apart from further methylene bridge formation. Meanwhile, the auto-condensation between tannin molecules could also occur. All of these will result in the incomplete polymerization and lead to weakness and brittleness of the resins.

OH HO OH OH OH HO O 2- 2x SO3 OH OH

NaSO3 SO3Na

HO OH OH HO O OH 2- OH SO3 OH OH

SO3Na

Figure 2.6 Sulfitation of tannins

Bridging agents with longer molecules, such as phenolic and aminoplastic resins could solve the problem of insufficient crosslinking by helping to bridge distances too large for interflavonoid methylene bridges. Tannin modification such as sulfitation (Figure 2.6), acid or alkaline hydrolysis to break down the hydrocolloid gums in tannin extractives can be also applied to improve the resulting resin properties [1, 8, 9]. Fortification or copolymerization with PF or UF (Figure 2.7) is found to be another practical approach to enhance resin performance [1, 8].

OH OH OH OH

HOH2C H2C CH2 CH2OH2 Tannin Tannin H2C CH2 Tannin

CH2OH CH2OH n n

HOH2CNHCONH CH2 NHCONH CH2 NHCONHCH2OH Tannin Tannin CH2 NHCONH CH2 Tannin n n

Figure 2.7 Fortification of tannin-based adhesives

2.5 Extraction and adhesive application of phenolic compounds from bark

2.5.1 Motivation of using renewable biomass as feedstock for adhesives

Phenol-formaldehyde (PF) resol resin, made from the addition and condensation reactions between phenol and formaldehyde under alkaline conditions, has been widely used as an adhesive in the wood products industry for a long time due to its good bond strength, water resistance and low initial viscosity [1]. Phenol used for PF resol resin synthesis is derived from fossil fuel resources (petroleum and coal). With the growing concern for fossil fuel depletion and environmental footprint, there is strong interest in exploring renewable resources as alternative feedstock to replace petroleum-based phenol for PF resin synthesis.

Lignin, tannin, and other phenolic compounds from renewable biomass materials, such as wood and agriculture residues, have been used for phenolic adhesive resin synthesis due to their similarity in structure to phenol [3]. While among them, bark is a non-food renewable biomass material, which is generally available in large quantity as waste residues from conversion of wood logs to various forest products. In mills, bark is usually mixed with other woody residues and used as hog fuel for heat recovery. But bark has a rather low heating value and its heating value drops sharply, particularly when wet. However, bark, with similar chemical composition to wood, is rich in extractives and phenolic compounds, such as lignin and tannins, which makes it attractive to be used as an alternative to petroleum-based phenol.

2.5.2 Extraction of phenolic compounds from bark for phenolic resin synthesis

The lignin and tannins could be directly used to partially replace 30–50% of petroleum- based phenol for synthesizing PF resins [10-14]. However, wood composites made using these types of adhesives have previously shown inconsistent and unsatisfactory performance due to the highly variable chemical structure of the lignin obtained through different origins and pulping methods, as well as the low reactivity of lignin and flavonoid B-ring units within tannins. Chemical modifications of lignin and tannin as well as suitable extraction methods are necessary for effectively improving the reactivity and producing the resins with more satisfactory and consistent performance [8, 9, 15-17].

The well-known methods used for the phenolic compounds extraction from biomass materials, especially bark, for phenolic resin synthesis could be classified into two groups: one is phenol liquefaction; the other is alkaline extraction.

2.5.2.1 Phenol liquefaction of bark and other biomass materials

Phenol liquefaction is found to be an effective method to convert biomass materials including those with a high lignin or polyphenol content to chemicals useful for phenolic resin synthesis [18]. During the liquefaction, phenol will not only dissolve the main components of the biomass materials but also react with the depolymerized lignin or cellulose [18-20]. The liquefied products are mostly phenolated during liquefaction and the phenolated products are very reactive with formaldehyde, which makes it an ideal material to partly replace petroleum-based phenol to synthesize liquefied biomass based phenol formaldehyde resins [19-21].

Reaction temperature, reaction time, catalysts types, catalyst loadings and biomass to phenol ratio will significantly affect the liquefaction process, yield and properties of the liquefied products.

(a) Reaction temperature and reaction time

Reaction temperature is one of the most important factors for biomass liquefaction. Alma et al [22-25] studied the liquefaction behavior of birch wood meal in phenol catalyzed by

acids and found that increasing the liquefaction temperature significantly reduced the residue content and increased the reactivity of the liquefied products (increased combined phenol). The reaction temperature also affects the molecular properties of the liquefied biomass. Lin et al [26] investigated the molecular weight and molecular weight distribution of the liquefied wood using Gel Permeation Chromatography (GPC). They found that higher temperature reduced the molecular weight and narrowed the molecular weight distribution curve. Reaction time is also an important parameter for biomass liquefaction. Increasing liquefaction time will also decrease the residue content and increase the combined phenol [22-25]. Lin [26] found that after a certain time, increasing reaction time would cause the recondensation between the decomposed products, which would also increase the molecular weight and residue ratio. This phenomenon is more obvious under low solvent ratio liquefaction conditions. Besides, the molecular characteristics, such as molecular weight and polydispersity of the liquefied products, are also time dependent and differ among various solvent ratios.

(b) Catalyst types and catalyst loadings

Acid, alkaline and salts can be applied as catalysts to improve liquefaction yield and lower liquefaction temperature. The catalysts will also render the liquefaction more oriented and produce more uniform liquefied products. The most commonly used acidic catalysts include sulfuric acid, hydrochloric acid [22, 23], phosphoric acid [27] and oxalic acid [24]. Acid catalysts can lower the liquefaction temperature more efficiently than alkaline catalysts such as NaOH and metallic salts. Organic acids such as formic acid [28-29], dibasic organic acids and organic anhydrate including maleic acid anhydride (MA), phthalic acid anhydride (PA), and trimellitic acid anhydride (1,2,4-Benzenetricarboxylic acid anhydride [30] were also used as catalysts for biomass liquefaction.

The effect of organic sulfonic acids, including sulfuric acid (SA), benzenesulfonic acid (BSA), methanesulfonic acid (MSA), 1,5-naphthalenedisulfonic acid (NDSA), 1- naphthalenesulfonic acid (NSA), and p-toluenesulfonic acid (PTSA), as catalysts during phenol liquefaction of Pinus radiata bark was investigated [31]. All the five organic sulfonic acids were very effective catalysts for the bark liquefaction in phenol with very

low residue ratio and played an important role in retarding the condensation reaction between phenol and degraded bark components during the liquefaction. While they have different effects on the properties of liquefied bark, such as combined phenol amount and molecular characteristics. PTSA was even effective for liquefaction at a solvent ratio (phenol/bark) as low as 2. The average molecular weights of the liquefied materials obtained from phenol-SA and phenol-PTSA systems were very similar at the same liquefaction time. The crystallinity of the residue when using PTSA as catalyst was slightly lower than that of the residue when using SA. The combined phenol content was higher using SA than using PTSA at the same liquefaction yield.

Alma [25] et al. studied the acid-catalyzed birch wood meal phenolation process. They concluded that increasing liquefaction solvent decreased the residue content and prevented the re-condensation of the decomposed compounds to some extent and increased the reactivity of the liquefied products (increase in combined phenol). Besides, the molecular characteristics of the liquefied products were also influenced by the biomass to solvent ratio. Higher solvent ratio restricted the broadening tendency of the liquefaction products towards larger molecular weight region during the acid-catalyzed liquefaction of birch wood in phenol [26].

(d) Mechanism of phenol liquefaction

The phenol liquefaction mechanism of wood, bark and other biomass materials with acid catalysts is still unclear. Model compounds such as cellobiose and guaiacylglyerol-β- guaiacyl ether (GG) were used to investigate the liquefaction mechanism [32-35]. In general, phenol reacted with the biomass and phenolated products are formed during the liquefaction.

Cellulose underwent degradation and dehydration into pyranose ring, followed by the reaction of the degraded products with phenol. Phenol was found to react with C1, C2, C3, C5 and C6 in cellulose or its degraded products except for C4 in the non-reducing end. The glucose was produced from degradation of cellulose via cello-oligo-saccharides and reacted with phenol in different forms with acid catalyst. Products such as tri(4-

hydroxylphenyl)methane (Figure 2.8 a), 1,1,2-tri(4-hydroxylphenyl)ethane (Figure 2.8 b) and other phenolated products (Figure 2.8 c) with their structural units highly resemble to methylolphenol typically present in conventional PF resins are formed during the liquefaction.

HO OH

HO OH CH

CH2

OH OH

(a) (b)

OH HO HO OH

CH CH CH CH 2 2 OH CH2

OH OH OH

(c)

Figure 2.8 Typical liquefied products from acid-catalyzed phenol liquefaction of cellulose [32]

Lignin underwent bond cleavages in both β-O-4 linkage and Cα-Cβ, and the condensation between phenol and GG occurred during liquefaction. The liquefaction of GG forms intermediates that are highly reactive and can further condense with each other or with phenol. The main liquefied products of GG such as diphenylmethanes and guaiacol were highly phenolated and could be used for phenolic resin synthesis.

OH OH

CH2 CH2

OH OMe OH

Figure 2.9 Typical liquefied products from acid-catalyzed phenol liquefaction of lignin [33, 34]

A study [36] using pyrolysis molecular weight beam spectrometer (py-MBMS) showed that during the sulfuric acid catalyzed phenol liquefaction of Calabrian pine (Pinus brutia Ten.), Lebanon cedar (Cedrus libani), acacia (Robinia pseudoacacia) and European chestnut bark, the formation of furfural from hemicellulose and hydroxymethyl furfural from cellulose was present in small amounts. The cleavages of the linkage of lignin- carbohydrate and ether bonds in the lignin side chain, especially benzyl ether bonds, occurred. Phenol is attached to lignin at the α-position on the side chain of lignin by C-C bonds, while for phenol the lignin moiety is in its ortho or para positions. This process caused the dissolution of bark powder into the reaction medium. Softwoods and hardwoods, both the starting bark materials and the solid residues remaining after liquefaction, have different py-MBMS spectra. Increasing the catalyst loading decreased the amount of carbohydrate and unmodified lignin, while increased the amount of phenolic fragments present in the liquefied products. There is a decrease in the carbohydrates and lignin structures and an increase in the amount of condensed structures in the residues as the concentration of the acid catalyst increased.

2.5.2.2 Alkaline extraction of bark

Compared with phenol liquefaction, alkaline extraction is another effective and relatively simple way to extract phenolic compounds from bark for phenolic resin synthesis. Tannin is one of the major products from the alkaline extraction of bark.

The extracts from various types of solvents including water are usually a mixture of tannins, sugars, pectin and lower molecular weight polyphenols [1, 4, 5, 38]. For a long time, tannins have been commercially extracted in South Africa and South America from various Acacia and Quebracho species. The commercial-scale extraction of other species including hemlock, spruce and pine species were also undertaken in other parts of the world. However, the low yield, high viscosity and variable extract quality limited its commercial production and adhesive application.

Many attempts were made to increase the extraction yield, modify the tannin viscosity and improve the product quality [5, 37]. Techniques included water extraction with addition of chemicals, solvent extraction, fractionation or extracts with ultrafiltration. Multi-stage squeeze extraction comprising two stage hot water extraction, one-stage NaOH aqueous extraction, and one-stage hot water washing with sulfitation of the last two stage of extracts was developed and found to effectively improve the tannin yield. In addition, the counter-current water/methanol and metabisulphite/urea solution, sulphite solution and vacuum belt sodium carbonate/sodium sulphite solution were also applied to improve the extraction yield, reduce the molecular weight and increase the reactivity of the extractives.

2.5.3 Adhesive application of bark phenolic compounds

2.5.3.1 Adhesive application of phenolated bark

Previous studies have shown that bark can be successfully liquefied in phenol with the aid of an acid catalyst [15, 38-41]. The liquefied products were further converted into phenolic resins for preparing moulding products or bark-based adhesives.

Alma et al [38] studied the phenolated barks of Calabrian pine (Pinus brutia Ten.), cedar (Cedruslibani), eucalyptus (Eucalyptus camaldulensis), acacia (Robinia pseudoacacia), Anatolia chestnut (Castanea sativa), and Turk oak (Quercus cerris) catalyzed by sulfuric acid and found that the mechanical properties of the molding materials from various species of phenolated bark were different. The tensile properties of the molding materials were affected by the catalyst loadings. Stress-strain behavior of the several bark-based molding materials was found to be similar to those of phenolated poplar wood and

commercial novolak resin based molding materials. Their modulus also increased with catalyst concentration increasing from 0.5 to 1-2% and then remained unchanged or increased slightly. The percentage tensile elongation went up from 1.23% to 1.57% when the catalyst concentration increased from 0.5 to 4%. The various bark-based molding materials (except for the acacia species) prepared at the catalyst concentration of 1-2% had quite comparable tensile properties to those of phenolated wood- and commercial novolak resin based molding materials.

In the research of thermal stability of novolak-type thermosetting resins made by the acid- catalyzed phenol liquefied bark of Calabrian pine (Pinus brutia) and Anatolia chestnut [39], it was found that under the same reaction condition, the residue, combined phenol content and thermal stability of the resin made by Calabrian pine were different from those made by Anatolia chestnut. The catalyst loadings significantly influenced the thermal stability of the resins. With increasing catalyst loading, the amount of bark residue after liquefaction decreased while the combined phenol content increased. The thermogravimetric weight losses and glass transition points (Tg) of the cured phenolated bark also increased with the increasing catalyst loading due to a higher cross-linking ratio. Tg of the cured phenolated bark was lower than that of the cured commercial novolak resin.

Lee et al. [40] liquefied different barks of Taiwan acacia and Chinese fir in phenol with either H2SO4 or HCl as the catayst. Then they prepared the liquefied bark-based resol resin and applied it to particleboard manufacturing. The resins made from H2SO4 catalyzed liquefied bark had a higher viscosity, a longer gel time at 135℃ and more higher molecular weight ingredients than the resin made from HCL catalyzed liquefied bark. They believed that strong acid promoted degradation of wood lignocelluloses and reduced their molecular weights, thus facilitating the combination of phenol with lignocellulose and accelerated the bark liquefaction. However, the strong acid caused the recondensation of lignin after initial degradation and increased the molecular weight of the liquefied products. Besides, the viscosity of liquefied bark-based resol resins was affected by bark species. The thermal properties of the resin, such as maximum temperature of exothermic peak and onset temperature, as well as mechanical properties

of particleboards bonded by these resins, including static bending strength and internal bonding strength, were also different between different bark species and catalysts used for liquefaction. Resol resins prepared from bark liquefied with H2SO4 as the catalyst had higher viscosity, while those with HCl as the catalyst had a higher maximum temperature and height of exothermic peak and a larger quantity of exothermic heat at condensation.

Santana et al [15] liquefied tannin and bark of black wattle in phenol with sulfuric acid as the catalyst. Tannin and bark replaced 33% of phenol in preparing resol-type plywood adhesives. The whole bark dissolved easier in phenol than tannin-free bark. Resins made from the whole bark gave better wood failure and higher shear strength than that made from tannin-free bark. Shear strengths and wood failures of Southern pine plywood prepared with liquefied tannin and bark resins were comparable to those of commercial phenol formaldehyde resins.

The influence of NaOH addition on the properties of bark-based phenol formaldehyde adhesives was also studied [41]. The level of NaOH addition had an important effect on various properties of bark-based phenol formaldehyde adhesives, such as gel time, free formaldehyde content, thermosetting peak temperature, molecular weight distribution as well as wet shear strength and free formaldehyde release from the bonded plywood. A two-step process of adding NaOH was found to be a better method in preparing bark- based PF adhesive.

2.5.3.2 Adhesive application of bark tannin extractives

The earliest attempts to produce wood adhesives from tannins started in the 1950s [5]. The use of tannins in structural gluing of glulam and finger-jointing emerged in the 1980s with two distinct tannin-based technologies evolving to displace resorcinol formaldehyde-based systems. Tannin extractives from Mangrove used for synthesizing tannin-formaldehyde resins were found to give a strong water resistant adhesive for bonding plywood [42]. Wattle tannin extractives were also reported to produce a waterproof adhesive [43, 44]. In addition, tannin extractives from Pinus radiata bark, ponderosa pine bark, larch pine, Acacia mollissima, western hemlock, spruce, and eucalyptus etc. were also applied for making bonding agent for various wood composites.

Kreibich and Hemingway blended southern pine bark extractives with resorcinol and heated the mixture under acidic condition. They thought that tannin contained in the extracts could be acid-catalyzed and cleaved to produce tannin-resorcinol-adduct, and this adduct could be applied to synthesize resorcinol phenol formaldehyde (RPF) copolymer resins. With this method, 60% of resorcinol could be replaced by this adduct, and the time needed for this copolymer resin preparation was only half of that for conventional resorcinol formaldehyde (RF) resin. These copolymer resins had cold setting properties and bonding strength resembling RF resins. In addition, they found that RPF copolymer resins blended with sulfonated southern pine bark extracts could cure at approximately 30- 60 min at the room temperature and the bonding properties would be invariable when 50% of RPF was substituted [45, 46].

Fechtal and Riedl prepared synthetic resins from the copolymerization of eucalyptus and Acacia mollissima bark extracts with phenol and formaldehyde, and applied these resins to particleboard manufacturing [47]. Lu and Shi prepared phenol-bark extracts-formaldehyde copolymer resin with 60% of phenol substituted with larch pine bark tannin [48]. Roffael et. al. used spruce bark tannin as the binder in the manufacturing of particleboard and fiberboard [49].

Vázquez et. al. added pine bark tannins to a resol prepared with a low formaldehyde content in order to obtain a high methylol group content. With a phenol substitution degree of 33%, it was possible to manufacture plywood of commercial quality [50]. Besides, they prepared phenol-bark extracts-formaldehyde copolymer resins using Pinus pinaster bark extracts for plywood manufacturing. They also compared the curing curves of tannin- phenol-formaldehyde resin measured by a Differential Scanning Calorimeter (DSC) and those of phenol formaldehyde resin. The cure kinetics of the two resins were established. It showed that tannin-phenol-formaldehyde resin cured faster than PF resin. The results implied that tannin-phenol-formaldehyde resin could achieve higher productivity and shorter press time than the conventional PF resin [51, 52].

However, poor and inconsistent bonding performance of tannin adhesives was reported in the literature. Poor wet strength, brittleness, and poor wood penetration of the tannin adhesives were attributed to factors including [1, 5, 8, 13, 14]: (1) the tannin molecules are

big and therefore cannot rotate freely around their backbone, which results in the observed inherent brittleness. (2) The high reactivity of tannin molecules causes premature cure. Consequently, the residual active centers become too far apart for formaldehyde molecules to bridge. The resulting incomplete crosslinking enhances loss of structural integrity in adverse environments. Fortification, hence copolymerization of tannins with various phenolic and amino plastic resins and other chemical modifications of tannins had to be applied for improving the tannin adhesives performance.

Sowunmi reported that poor mangrove tannin adhesives properties could be improved by treating tannin extractives with acetic anhydride and sodium hydroxide followed by modification with 20% resol-type PF resin addition [8]. Activation energy was lowered while the heat of reaction was increased, thus promoting a greater level of condensation and reducing the tendency to cure prematurely. Fortifying hydrolyzed mangrove tannin with 20% phenol or PF resin produced plywood with bond strength as good as those derived from some commercial PF resins.

Pichelin [53] used hexamine as the hardener to react with Mimosa tannin at pH 10. The results showed that it was a non-formaldehyde system. Under alkaline reaction condition, the hexamine decomposed slowly to the reactive iminoamino methylene bases intermediates and reacted with tannin rapidly. A long ambient temperature pot-life with fast hardening and no free formaldehyde emission adhesives was acquired. Full-scale structural beams satisfying the JIS A 5908 standard were produced with this kind of adhesive.

Mitsunaga et al. [17, 54, 55] treated the bark extractives with excessive phenol and boron trifluoride to cause phenolation. They found this method could open the pyran-ring of condensed tannin and broke down the intermolecular chain of flavonoid units, which reduced molecular weight and increased the activity of its molecular chain to promote the reaction of tannin with formaldehyde.

Barbosa et al [9] applied chemical methods to modify tannins to improve the resulting adhesive performance. The flexible, long linear six carbon chains were introduced to open the tannin structure through partial esterification of the free hydroxyl groups with adipoyl

chloride, which acted as an internal plasticization of macromolecule and resulted in brittleness reduction and better distribution of forces throughout the laminate glue lines. The introduction of adipic ester segments between the hydroxyl groups of the compact tannin macromolecules could also help reduce brittleness.

In addition, metal ions have been applied for modifying the tannin extractives. Pizzi [13] reported the accelerating effect of bivalent metal ions observed in simple phenol- formaldehyde reactions was also observed in tannin-formaldehyde resins. Tannin steric hindrance prevented the retarding effect of trivalent metallic ions on tannin-formaldehyde reactions. Improved plywood strength and shorter pressing times were obtained by the addition of metal ions to tannin-formaldehyde resins.

Besides the traditional crosslinking agent, formaldehyde, other cross-linking agents were also used for the tannin adhesives synthesis. Furfural [1, 5] has been coupled with tannins to form an adhesive; glyoxal [56-57] was also found to be able to promote tannin crosslink due to its reactivity. The oxazolidines as the new type of formaldehyde donor and the cross-linker tris-nitromethane [58-60] have been employed to provide room temperature cure of tannin adhesives. In addition, the autocondensation of tannin could also give a satisfactory adhesive bond; however, the premature cure and incomplete crosslinking could easily occur and should be carefully avoid [8, 61].

2.6 Mountain pine beetle infested lodgepole pine

Lodgepole pine is one of the most important species in the interior British Columbia and western Alberta, Canada, which makes up 24 percent of the province’s total growing stock and 50 percent of the BC interior’s growing stock [62]. The bark of lodgepole pine is thicker (may be up to 2.5 cm near stump) with deep furrows. This bark has reddish-brown color.

The current outbreak of the mountain pine beetle infestation in the western provinces of Canada is a huge problem faced by the forest industry. This beetle attacks mostly mature (80-100 years old) lodgepole pine by feeding on the sapwood of the tree. The attack also has a symbiotic relationship with blue stain fungi that causes paralysis of immune system

of the trees [63]. According to the current estimates of BC Ministry of Forest and Range (MOFR), the infestation is expected to kill approximately 960 million m3 of tree by 2013 [64] and it is expected that at least 200-600 million m3 of wood would remain unharvested by 2013 [65]. This infestation will result in large-scale loss of jobs in the forest industry and is also a potential fire hazard. If the infested forest is not harvested, it will decay and release carbon to the atmosphere [66].

The cumulative effects of the fungi and beetle lead to chemical changes in wood and bark. Previous studies showed that the infested wood from lodgepole pine has lower lignin and hemicelluloses contents, as compared to the non-infested wood. The beetle infested lodgepole pine bark has a slightly higher holocellulose content, higher proportion of fatty and resin acids, lower lignin content, lower proportion of sterols, steryl esters and triglycerides than the bark from the non-infested lodgepole pine [67].

Exploring value-added applications of barks from mountain pine beetle infested lodgpole pine will be highly advantageous for forest industries and the environment.

2.7 Conclusions

This chapter reviewed previous research on obtaining phenol substitute from bark and other renewable biomass materials through acid-catalyzed phenol liquefaction and alkaline extraction. The application of the phenol partial substitute, such as liquefied bark and bark tannin extractives, in the PF resin system was also summarized. The conversion of bark into phenol substitution and their application in PF resol resins are very attractive not only because bark, a large available so-called waste material from forest industries, has been incorporated into the PF resin system and reduced the usage of petroleum-based phenol, but also because the inclusion of bark components in the PF resol resin system might modify and improve some resin properties.

Barks have been successfully liquefied in phenol using an acid catalyst or extracted by alkaline solutions for acquiring desirable phenolic compounds previously. The obtained liquefied bark and bark alkaline extractives were converted to bio-based PF resins and applied for manufacturing of plywood and particleboard with satisfactory performance.

However, there is no previous research studied the conversion of the mountain pine beetle infested lodgepole pine barks into phenol substitute for application in PF resin synthesis. How introduction of bark components in the resin synthesis would affect the molecular structure and properties of the resulting bio-based PF resins is still unclear and has not been explicitly studied. Whether beetle infestation would negatively affect performance of the resulting bio-based bark PF resins is unknown. Fundamental understanding of the role of bark components in the bio-based bark PF resins is still lacking in the existing literature.

The following chapters in this thesis describe the unexplored research areas related to the adhesive application of mountain pine beetle infested lodgepole pine barks and try to fulfill the knowledge gap. The acid-catalyzed phenol liquefaction and alkaline extraction of the mountain pine beetle infested lodgepole pine barks as the methods for obtaining the suitable phenol substitutes are investigated. The liquefied bark-PF resin and bark extractive-PF resins are formulated and characterized, respectively. The effect of bark components on the resin properties as well as the relationship between the resin structure and performance are explored.

Chapter 3

Experimental Approach

In this chapter, the experimental details for how to obtain phenol substitute using the mountain pine beetle infested lodgepole pine barks, how to make the bio-based PF resins using the acquired bark phenol substitutes and how to characterize the bark phenol substitutes and bio-based bark PF resins are summarized.

3.1 Bark phenolic compounds extraction

3.1.1 Bark phenol liquefaction

3.1.1.1 Liquefaction conditions

The bark powders (passed through No. 35 mesh screen) from the mountain pine beetle infested Lodgepole pine provided by FP-Innovations Forintek division were used as raw materials. The bark powders were dried in an oven at 105°C for 12h before the liquefaction experiments. Bark powders 10g, phenol in crystal form 20-50g, 96% sulfuric acid (used as the catalyst, loading rates: 1%, 2%, 3%, 5%, 7% based on the weight of phenol) were charged into a three-neck flask with a condenser. The reaction was carried out in a heating oil bath at 120°C, 150°C and 180°C under atmospheric pressure. The liquefaction reaction time was set as 30, 60, 90, 120, and 150 minutes, respectively. After the liquefaction reaction, the system was cooled to room temperature using cold water. Phenol, concentrated sulfuric acid (96%), methanol and other chemicals were reagent grade, purchased from Caledon Company without further purification.

3.1.1.2 Measurement of liquefaction yield

Liquefaction yield is obtained through the measurement of the residue ratio. The liquefied products were diluted by excessive methanol (around 150mL) and then filtered using Millipore filter paper. The methanol insoluble was dried in the oven at 105°C to a constant weight. The residue ratio was calculated using equation:

Wr R100%=× Wo (3-1)

Where: R is the residue ratio of the bark liquefaction reaction (%). Wr is the oven-dry weight of the residue after liquefaction (g). Wo is the original oven-dry weight of the bark before liquefaction (g).

3.1.1.3 Measurement of free phenol

Analysis of free phenol in liquefied bark was conducted using a Perkin Elmer HPLC with a UV detector set at a wavelength of 254nm. The sample (25 µl) was injected into the injection port with a 0.45 um PTFE pre-filter and was eluted with a mobile phase of 65% methanol and 35% water at a rate of 1.0 ml/minute through a C18 reverse phase column (4.6 mm x 250 mm). The free phenol peak appeared at approximately 1.8 minutes and was quantified by an external standard calibration in the concentration range of 50 to 2000 ppm (R2=0.99). Since there could be free phenol attached to the residues of bark liquefaction, so the free phenol content of the bark liquefaction in the liquefied bark fraction was calculated by the equation:

W FPh =Fph ×100% WW W o+ ph− R (3-2)

Where: FPh is the free phenol content in the liquefied bark fraction (%). WFph is the weight of free phenol in the liquefied bark fraction after the liquefaction (g). WR is the oven-dry weight of the residue after liquefaction (g). Wo is the original oven-dry weight of the bark before liquefaction (g). Wph is the weight of phenol before liquefaction.

3.1.1.4 Characterization of liquefied residues

(a) Chemical composition analysis

Chemical composition analysis of bark and liquefied bark residues include holocellulose, alpha-cellulose and lignin content. The extractive-free residues were first made according to ASTM D1105-96 [68]; holocellulose and α-cellulose contents were analyzed according

to Zobel et al.’s method [69]. Acid-insoluble lignin content was analyzed according to ASTM D1106-96 [70].

(b) Thermogravimetric analysis (TGA)

Bark and liquefied bark residues were first ground into fine powders which pass through No.40 mesh screen and remain on No.60 mesh screen. Around 10 mg of the sample was added to the platinum pan. Atmospheric air pressure was applied, and the temperature increased from room temperature to 700℃ at the heating rate of 20℃/min with nitrogen purging.

FTIR measurements of the liquefied residues were carried out using an FT-IR TENSOR 27 spectrometer (Bruker Optics, USA) having a frequency range of 4000cm-1 to 400cm-1 with KBr pellets. The FTIR measurement has a resolution of 4 cm-1.

(d) SEM

Scanning electronic microscopy (SEM, Hitachi S2500, Japan) was used to examine the morphology of the un-liquefied residues. Samples were sputter coated with gold before scanning.

3.1.2 Bark alkaline extraction

3.1.2.1 Alkaline extraction conditions

Air-dried powder sample (100 g) of mountain pine beetle-infested lodgepole pine bark that had passed through 0.500 mm sieve (35-mesh) was extracted by 500 mL of 1% NaOH aqueous solution in a boiling water bath for 2 h. Extraction was repeated two more times. Each extract was then poured through Whatman filter paper (110 mm Ø). The resulting alkaline filtrates from the three-stage extractions were combined.

The alkaline soluble fraction was separated into two parts: one was kept at pH 11.06, while the other portion was neutralized by 1M HCl to pH 7.0. Both fractions were then dried at 60°C on a hot plate to constant weights. The resulting solid extractives were

ground by mortar and pestle into powders for subsequent resin synthesis.

3.1.2.2 Measurement of the extraction yield

The alkaline insoluble bark residue was washed with copious amounts of 1% acetic acid aqueous solution and distilled water, and then air-dried to constant weight. The yield of the alkaline extractives was quantified as:

W !W Extractives(%) = 0 r "100 (3-3) W0

where, W0 is the air-dry weight of the bark before extraction and Wr is the air-dry weight of the residues after extraction.

3.1.2.3 Characterization of bark extractives

(a) Molecular characteristics

Weight average molecular weight (Mw), number average molecular weight (Mn) and polydispersity index (Mw/Mn) of the extractives were measured by a Matrix-Assisted Laser Desorption/Ionization-Time-of-Flight/Time-of-Flight spectrometer (MALDI- TOF/TOF, Applied Biosystems, Framingham, MA, USA). The mass spectra were acquired in low linear mode in a mass range from 60 to 2000 Da. The samples were prepared with DHB (2, 5-dihydroxybenzoic acid) as the matrix. The matrix solution was obtained by mixing the matrix material in 50% aqueous ethanol and 0.1% trifluoroacetic acid to give a 10mg/ml concentration. The matrix was premixed with the resin sample at a ratio of 4:1 (v/v), and then 1 µL of the mixture was added on the MALDI target and air- dried. Each spectrum was collected by the total number of ions from 500 laser pulses. The average molecular weight and polydispersity index were calculated using Mané et al.’s methods [71].

(b) Structure characterization

The FTIR measurements of the bark extractives were carried out using an FT-IR TENSOR 27 spectrometer (Bruker Optics, USA) having a frequency range of 4000cm-1 to 400cm-1.

The reactivity of the bark extractives was measured by the stiasny number, which was the formaldehyde condensable phenolic content. The stiasny number of the bark extractives was conducted by the following procedure. 100 mg of tested sample (bark extractives) was dissolved in 10 mL distilled water. 1 mL of 10M HCl and 2 mL formaldehyde solution (37%) were added into the dissolved sample. The reactants were heated and refluxed for 30 min. The reaction mixture was filtered while hot through a sintered glass filter (fine size). The precipitate was washed with hot water (5X10 mL) and dried in an oven. The yield of formaldehyde condensable phenolic compounds was expressed as a percentage of the weight of the starting material [72-73].

3.2 Bio-based PF resin synthesis

3.2.1 Liquefied bark-PF resin

Liquefied bark was acquired from the following liquefaction condition: lodgepole pine bark with or without mountain pine beetle infestation to phenol ratio is 1:3; sulfuric acid is used as the catalyst (3% of phenol weight). The liquefaction reaction was conducted at 150°C for 120 minutes. A calculated amount of the liquefied bark, 37% formaldehyde, and 40% sodium hydroxide (1/3 of total NaOH weight) were mixed in a three-neck flask. The reaction temperature was increased to 65°C within 30 minutes and was kept at 65°C for 10 minutes, followed by the addition of the remaining 2/3 of NaOH. The reaction mixture was then heated to 85°C and kept at 85°C for 60 minutes. After the reaction, the system was cooled to room temperature. No additional phenol was added during the liquefied bark-PF resin synthesis.

3.2.2 Bark extractive-PF resin

A calculated amount of bark extractives in powder form, phenol (crystal form), 37% formaldehyde, and 40% sodium hydroxide (1/3 of total NaOH weight) were mixed in a three-neck flask. The replacement levels of phenol by the bark extractives were 30%, 50% and 70% by weight. The reaction temperature increased from room temperature to 65s within 30 minutes and was kept at 65°C for 10 minutes in an oil bath, followed by the

addition of the remaining 2/3 of 40% NaOH. The reaction mixture was then heated to 85°C and kept at 85°C for 60 minutes. After the reaction, the reactor mass was cooled to room temperature.

3.2.3 Lab PF and commercial PF resins

In order to understand how the bark phenolic compounds affect the resulting resin properties, a laboratory made control PF resin (lab PF) without bark components was prepared by following exactly the same reaction steps used for the synthesis of bio-based PF resins. All the chemicals were purchased from Caledon Laboratory Chemicals, Canada, and used without further purification. A commercial PF resin, provided by the FP- Innovations, Forintek Division, for the face layers of oriented strandboard production was also used for comparison.

3.3 Characterization of bio-based PF resins

3.3.1 pH, viscosity, solids content, gel time, Mw and Mw/Mn of the bio-based PF resins pH values of the resins were measured at 25°C. The viscosity of the resins was measured using a Brookfield rotary viscometer with spindle 3 at 25°C. Solids content measurement was according to ASTM D 3529 [74]. Gel time was measured by charging 1g of resin into a 16-mm wide test tube and heating the test tube in an oil bath at 120±1°C. Gel time was defined as the time period from the immersion of the test tube into the oil bath to the beginning of the resin gelation (resin forming a string when a glass rod was lifted from the resin) An average value of three replicate measurements was reported.

The molecular characteristics of the PF resins, such as weight average molecular weight (Mw), number average molecular weight (Mn) and polydispersity index (Mw/Mn), were measured by a Matrix-Assisted Laser Desorption/Ionization-Time-of-Flight/Time-of- Flight spectrometer (MALDI-TOF/TOF, Applied Biosystems, Framingham, MA, USA). The detailed sample preparation, instrument parameters setup and data calculations are similar to those described in 3.1.2.3 (a) for bark extractives.

3.3.2 Differential scanning calorimetry (DSC) analysis of the bio-based PF resins

High-pressure pans (TA DSC-Q100, TA Instruments, USA) were used for investigating resin curing behavior. Dynamic scans were carried out at heating rates of 5°C /min, 10°C /min, 15°C /min and 20°C /min, respectively, starting from room temperature and increasing to 250°C.

Kissinger method was used to calculate the activation energy [75].

⎛ φ ⎞ E 1 ⎛ RA ⎞ ln⎜ ⎟ = − + ln⎜ ⎟ (3-4) ⎜ T 2 ⎟ R T E ⎝ p ⎠ p ⎝ ⎠ where, ϕ is the heating rate (K/s), Tp is the peak temperature (in Kelvin) at the given heating rate. A is the pre-exponential factor. R is the ideal gas constant and E is the activation energy.

For isothermal scans, a steady isothermal baseline was first established at the cure temperature using two empty sample pans. Five cure temperatures (110°C, 120°C, 130°C, 140°C and 150°C) were used in the isothermal DSC experiments. One replicate was done for each scan. Sample was chosen to be run at certain conditions randomly for checking reproducibility, and the maximum variation of the onset and peak temperature was found to be less than 1°C.

3.3.3 Thermal gravimetric analysis (TGA) of the cured resins

3.3.3.1 Thermal stability analysis

The bio-based PF, lab and commercial PF resins were cured in an oven at 80°C for 48 hours. The cured resins were ground into fine powders that were able to pass through a 100-mesh screen. About 10mg of each cured resin sample was added to a platinum pan and heated from the room temperature to 700°C at the rate of 10°C /min under N2 atmosphere using a thermal gravimetric analyzer (TGA, TA TGA-Q500, TA Instruments, USA).

3.3.3.2 Kinetic analysis of resin thermal degradation process

The kinetic parameters for the thermal degradation process of the PF resins were obtained using the thermogravimetric analysis employing multiple heating rates. It is known that the rate of thermal degradation can be expressed as:

d! = f (!)k(T) (3-5) dt where, α is degree of conversion, f (α) is the reaction model, k(T) is represented by an Arrhenius equation shown as:

E k(T) = Aexp(! ) (3-6) RT

A is pre-exponential factor, E is the activation energy, T is temperature, R is gas constant.

Meanwhile, degree of conversion is given by the weight loss during the thermal degradation process and expressed as:

m ! m ! = 0 (3-7) m0 ! m" in which, m0 is the initial mass of the sample, m is the sample mass at certain reaction time/temperature, m∞ is the final mass of the sample after degradation.

Since different heating rates, β, were applied, so the equation (3-5) could be expressed as:

d! d! dT d! = = " (3-8) dt dT dt dT

Combining the equations (3-5), (3-6), and (3-8), we get:

d! d! E = " = f (!)Aexp(! ) (3-9) dt dT RT

Rearranging the equation (3-9) and integrating from T0, the initial temperature that corresponds to α0, to Tw, the temperature when certain weight loss occurred, we have:

! d! A Tw E = exp(" )dT (3-10) ! !T !0 f (a) " 0 RT

In order to determine the kinetic parameters of the thermal degradation process for the PF resins, Flynn-Wall-Ozawa [76, 77], Kissinger-Akahira-Sunose [78] and Kissinger [78] methods, which were previously used to obtain thermal degradation kinetics of PF adhesives and recommended by International Confederation for Thermal Analysis and Calorimetry (ICTAC) kinetics committee for performing kinetics computation on thermal analysis data were used.

For the Flynn-Wall-Ozawa method, we have the expression:

AE E log(!) = log( " )! 2.315! 0.4567 " (3-11) Rg(") RT g (α) is the integral of reaction model.

For the Kissinger-Akahira-Sunose method, the expression is given as:

!i E" ln( 2 ) = Const ! (3-12) T",i RT

th βi is the heating rate under i temperature program. Tα,i is the temperature at which extent of conversion α is reached under ith temperature program.

For the Kissinger method, the expression is:

! AR E ln( 2 ) = ln( )! (3-13) Tp E RTp

Tp is the peak temperature obtained from the DTG curves of the tested resins at different heating rates.

3.3.3.3 Structural changes of the resins during thermal degradation

The resin samples after degradation at temperatures of 200°C, 400°C, 600°C, 800°C were collected and analyzed using Fourier Transform Infrared Spectroscopy (FTIR). The FTIR

measurement was carried out using a FT-IR TENSOR 27 spectrometer with ATR attachment (Bruker Optics, USA) having a spectra range of 4000cm-1 to 400cm- 1. Elemental analysis was carried out using a 2400 Series II Carbon Hydrogen Nitrogen (CHNS) Elemental Analyzer.

3.3.4 Evaluation of resins bonding strength

3.3.4.1 Lap-shear sample preparation

Poplar veneer was cut into strips (3 mm thick, 25.4 mm wide, and 108 mm long) with the length direction parallel to wood grain. The veneers were conditioned at 20°C and 50% relative humidity for at least one week before usage. The two layered poplar wood veneer specimens were bonded using the bio-based PF resins, lab PF resin and commercial PF resin. The adhesives were applied on one side of the poplar strip in an area of 25.4 mm×25.4 mm in size. The spread rate of the adhesives was 0.025-0.035g/cm2 on the solid basis. The adhesive-coated area of the poplar strip was then overlapped with an uncoated poplar strip, which had no adhesive. The resulting two-layered lap shear specimen was hot pressed at 160°C under the thickness control of 4.5mm for 3 minutes. After cooling and conditioning, the specimen were tested for shear strength on a Zwick universal test machine (Zwick/Z100, Zwick Roell group, Germany) following the standard lap shear test methods as described in ASTM D5868 [79]. The crosshead speed was 1.3mm/min. The average values based on a minimum of 10 replicates were reported. T-test was conducted to check for the statistical difference within the results.

25.4mm 3 mm 25.4mm

108mm

Figure 3.1 Lap shear specimen dimensions

3.3.4.2 Water resistance test

The lap shear specimens were subjected to both a water-soaking-and-drying (WSAD) test and a boiling water test (BWT), according to voluntary standard PSl-95 published by the US Department of Commerce through the Engineered Wood Association, Tacoma, WA [80]. For the WSAD test, the specimens were soaked in water at room temperature for 24h, and then were dried in a fume hood at room temperature for 24 h, followed by the shear strength measurements. For the BWT test, the specimens were first boiled in water for 4 h, and then dried for 20 h at 63±2°C. After the drying, the specimens were boiled in water again for 4 h, and then cooled down with tap water, followed by the shear strength measurements when they were still wet. The shear strength obtained using this method was defined as BWT/wet strength.

3.3.5 Liquid-state 13C NMR measurements

The liquefied bark, bark extractives, lab PF resin, liquefied bark-PF resin and bark 13 extractive-PF resins were first dissolved in DMSO-d6. The liquid-state C NMR spectra of these samples were recorded using a Unity 500 spectrometer under the following conditions: the pulse angle of 60 degrees (8.3 µs), a relaxation delay of 10s and with gated Waltz-16 1H decoupling during the acquisition period. About 400 scans were accumulated for each spectrum. The 13C chemical shifts were measured using tetramethylsilane (TMS) as the internal standard.

Chapter 4

Bark Phenol Liquefaction

4.1 Abstract

In this chapter, mountain pine beetle infested lodgepole pine (MPB) bark was liquefied in phenol using sulfuric acid as a catalyst. Effects of reaction conditions on the liquefaction yield, free phenol content in the liquefied bark fraction, chemical composition and morphology of the unliquefied residues were discussed. Results showed that liquefaction conditions, such as reaction time, reaction temperature, phenol/bark ratio, and catalyst loading, had significant effects on the liquefaction yield, free phenol content of the liquefied bark fraction and the properties of the unliquefied bark residues. Higher reaction temperature and prolonged reaction time not only promoted the degradation of the bark components during the liquefaction but also induced recondensation reactions among the degraded bark components. Higher phenol/bark ratio increased the bark liquefaction yield by retarding the recondensation reactions among the degraded bark components. The liquefaction yield, free phenol content in the liquefied bark fraction and residues properties were found to be mostly affected by the catalyst loading, followed by the reaction temperature, reaction time and phenol/bark ratio. Acid-catalyzed phenol liquefaction is an effective method to convert bark from the mountain pine beetle infested lodgepole pine into liquefied bark as an alternative phenolic feedstock for PF resol resin formulation.

4.2 Introduction

With the growing concern for the rapid depletion of non-renewable fossil fuel resources, developing technologies that can convert renewable biomass to fuel, energy, and chemicals becomes increasingly important [18]. Among various biomass conversion techniques, liquefaction has been reported to be an effective method for obtaining useful chemicals products [19, 20]. Acid-catalyzed phenol liquefaction of wood and bark is a known method for producing liquefied products suitable as raw materials to synthesize

more environmentally friendly phenolic resins for, thermo-molding and wood adhesives applications [22-35].

Similar to wood, bark is a renewable biomass material, but bark is also rich in polyphenols and has a lower cost. Previous studies [38-41] on acid-catalyzed phenol liquefaction of barks have shown that liquefaction reaction conditions, such as reaction temperature, reaction time, and catalyst loading, could significantly affect the residue ratio of the reaction, the free phenol content in the liquefied bark, and performance properties of the products made from the liquefied bark. However, no previous liquefaction work has used lodgepole pine barks even though it has been shown wood species is an important factor determining the final liquefaction products [22-35].

Past research on acid-catalyzed phenol liquefaction has also shown that during the liquefaction process, various components of the biomass, such as cellulose, lignin and hemicellulose, were both degraded and phenolated [22-26]. The highly phenolated liquefied products were highly reactive; as a result, recondensation reactions among the liquefied products could easily occur to reduce the final liquefaction yield. Study [81] on the wood liquefaction residues indicated that phenol not only dissolved cellulose, hemicellulose and lignin, but also reacted with the intermediate degraded products. Increase in phenol to wood ratio reduced the lignin content but increased the hemicellulose and cellulose contents in the solid residues. Higher phenol to wood ratio, however, was found to retard the recondensation reactions among the liquefied products and increased the final liquefaction yield [22-26, 81].

Meanwhile, due to the recent rampant mountain pine beetle infestation outbreaks in the western provinces of Canada, there is a large amount of beetle infested lodgepole pine wood resources available for utilization. It would be highly beneficial to explore phenol liquefaction as a method to convert these beetle infested pine barks to useful feedstock for making wood adhesives. Additionally, mountain pine beetle infestation was found to affect the wood chemical composition, physical and mechanical properties [62-67]. The chemical composition of mountain pine beetle infested lodgepole pine barks and non- infested lodgepole pine barks are shown in Table 4.1. There was no obvious difference in organic solvent extractives between the two types of barks. The beetle infested lodgepole

pine bark had a higher amount of 1% NaOH soluble than the non-infested lodgepole pine bark. Generally, 1% NaOH soluble of woody material is an indication of the degree of fungal decay, or degradation by heat, light, oxidation and etc. Since the hot alkali removed low molecular weight carbohydrates mainly consisted of hemicellulose and degraded cellulose in the bark, the beetle infested lodgepole pine bark gave a higher amount of hot alkaline extractives as expected. Comparing with the non-infested lodgepole pine bark, the beetle infested lodgepole pine bark had a lower α-cellulose and Klason lignin contents, which indicated that some α-cellulose and lignin had also decayed due to beetle infestation. The beetle infested barks also had a higher amount of hot water extractives, perhaps related to the immune response of the tree to the beetle attack. How these changes would affect the liquefaction behavior and properties of the liquefied fractions of the beetle infested lodgepole pine barks are still unknown.

Table 4.1 Chemical composition of bark

Beetle Infested Lodgepole Pine Non-infested % Bark Lodgepole Pine Bark Ethanol-Toluene extractives 17.7 (0.5) ** 18.2 (0.4) Dicholormethane extractives 14.2 (0.5) 14.7 (0.8) Hot water soluble * 7.5 (1.2) 3.9 (1.2) 1% NaOH solubles 68.1 (0.5) 61.9 (0.3) Holocellulose 46.7 (1.2) 46.5 (0.4) α-cellulose 20.5 (1.1) 24.9 (3.2) Klason lignin 42.6 (0.7) 45.1 (0.6) Ash 4.0 (1.8) 4.1 (2.8) * After Ethanol-Toluene extraction.

** Number in bracket is standard deviation.

Therefore, in this chapter, mountain pine beetle infested lodgepole pine barks were liquefied in phenol with an acid catalyst. The effects of different liquefaction reaction conditions on the liquefaction yield, free phenol content of the liquefied bark fraction and properties of the unliquefied residues were investigated to better understand the efficacy of phenol liquefaction as a conversion technique for these beetle infested pine barks.

A version of this chapter was accepted for publication in Current Organic Chemistry (2012).

4.3 Results and discussion

4.3.1 Effect of reaction conditions on the liquefaction yield and free phenol content

In previous studies, reaction conditions were found to significantly affect wood liquefaction yield, combined phenol content of the liquefied wood [22-26]. Liquefaction yield was evaluated by measuring the residues ratio of the liquefaction reaction instead.

4.3.1.1 Reaction temperature

When the reaction temperature increased from 120°C to 150°C, the residue ratio decreased significantly. When the reaction temperature increased from 150°C to 180°C, the residue ratio showed no further significant decrease, rather, reached a constant value. It suggested 150°C as a suitable temperature for the bark liquefaction reaction. Similar trend on the effect of reaction temperature on the residue ratio was observed for birch and Chinese tallow wood phenol liquefaction previously [22, 81]. It was found that liquefied wood reacted at 180°C had a slightly lower residue content than that at 150°C and the changes were not significant. This observation was attributed to the secondary condensation reactions by the degraded cellulose and lignin previously, which could also be an explanation for the observed plateau at 180°C in this study.

Residue ratio (%) Residue 10

0 100 120 140 160 180

Reaction temperature (℃)

Figure 4.1 Effect of reaction temperature on the residue ratio

(Phenol/bark ratio = 3, catalyst = 3 wt.% of phenol, reaction time = 60 min.)

From Figure 4.2, it could be seen that reaction temperature also significantly affected the amount of free phenol in the liquefied bark fraction. With the increasing temperature, the amount of free phenol decreased significantly. The higher the reaction temperature, the more phenol reacted with the bark components during the liquefaction with a lesser amount of free phenol being recovered after the reaction.

Free phenol (%) Free 20

0 100 120 140 160 180 Reaction temperature (℃)

Figure 4. 2 Effect of reaction temperature on the free phenol content

(Phenol/bark ratio = 3, catalyst = 3 wt. % of phenol, reaction time = 60 min.)

4.3.1.2 Phenol/bark ratio

Phenol/bark ratio also affected liquefaction yield and free phenol content of the liquefied bark fraction. As is shown in Figure 4.3, increasing phenol/bark ratio decreased the residue content. When the phenol/bark ratio was increased from 2 to 3, significant reduction in the residue ratio was observed. When the phenol/bark ratio exceeded 3, even though the residue ratio decreased further, the level of reduction was not as large. A similar effect of the phenol/bark ratio on the residue ratio was found for birch wood liquefaction in phenol in a previous study [24]. Liquefaction and recondensation were given as the competing reactions that resulted in such a trend. Higher amount of solvents used in the liquefaction retarded the recondensation reactions among degraded components by blocking the reactive sites of the liquefied wood components and preventing the formation of new larger molecular weight insoluble residues [26].

15 Residue ratio (%) Residue 10

0 1 2 3 4 5 6 Phenol/bark ratio

Figure 4.3 Effect of phenol/bark ratio on the residue ratio

(Reaction temperature=150°C, catalyst = 3 wt.% of phenol, reaction time =60 min.)

The amount of free phenol in liquefied bark increased gradually with the increasing phenol/bark ratio (shown in Figure 4.4). Higher phenol/bark ratio promoted the liquefaction and retarded the secondary condensation reaction of the decomposed components to give more free phenol after the liquefaction.

20 Free phenol (%) Free

0 1 2 3 4 5 6 Phenol to bark ratio

Figure 4.4 Effect of phenol/bark ratio on the free phenol content

(Reaction temperature=150°C, catalyst = 3 wt.% of phenol, reaction time =60 min.)

4.3.1.3 Reaction time

In general, reaction time is a key parameter for biomass liquefaction. According to previous work [22-26] the liquefaction reaction of wood occurred rapidly at the early stage of reaction and the residue ratio decreased continuously with a prolonged reaction time. But after a certain time, with increasing reaction time the residue ratio increased again probably due to recondensation among the decomposed products.

In this study, the effect of reaction time on bark liquefaction is shown in Figure 4.5. The process of bark liquefaction had a similar characteristic to wood liquefaction. Within the first 30 minutes of liquefaction, 60% to 70% of bark components were dissolved in the liquefaction solvent. The residue ratio slowly decreased when the reaction time increased from 30 to 120 min. It was also clearly shown in Figure 4.3 that after the residue ratio reached a minimum value; it started to increase as the reaction time was further prolonged. Probably with the increase in the reaction time, the residue ratio increased again due to recondensation reactions among the liquefied components.

10 Residue ratio (%) Residue

0 20 40 60 80 100 120 140 160

Reaction time (min)

Figure 4.5 Effect of reaction time on the residue ratio

(Reaction temperature=150°C, catalyst =3 wt.% of phenol, phenol/bark ratio = 3)

The effect of reaction time on the free phenol content of the liquefied bark fraction is shown in Figure 4.6. It was obvious that by increasing the reaction time from 30 min to 150 min, the amount of free phenol decreased gradually in the liquefied bark. With a prolonged reaction time, bark had gradually dissolved in the phenol; an increasing amount of phenol had reacted with the degraded components of bark and led to a lower amount of free phenol after the liquefaction.

Free phenol (%) Free 30

20 20 40 60 80 100 120 140 160 Reaction time (min)

Figure 4.6 Effect of reaction time on the free phenol content

(Reaction temperature=150°C, catalyst = 3wt.% of phenol, phenol/bark ratio = 3)

4.3.1.4 Catalyst loading

The effect of catalyst loading on the residue ratio is shown in Figure 4.7. When the catalyst loading increased from 1% to 3%, the residue ratio decreased significantly. When the catalyst loading increased further from 3% to 7%, the residue ratio decreased slowly and reached a plateau. Higher catalyst loading not only promoted the extent of liquefaction but also accelerated the hydrolysis and degradation of the long chain and crystalline cellulose, which was the most resistant component to acid-catalyzed liquefaction in phenol. Lin et al. [26] studied the molecular characteristics of birch wood liquefaction in phenol catalyzed by acid and found that the effects of catalyst loading on liquefaction was similar to the reaction time. Higher catalyst loading promoted the decomposition of the wood components during the liquefaction, but after it reached a saturation value, recondensation occurred at rather low phenol/wood ratio. Although increasing catalyst loading could effectively reduce the residue ratio and improve the bark liquefaction yield, the excessive amount of strong acid would have inevitably led to other various negative effects, such as corrosion of equipment and environmental pollution. So 3% of catalyst loading was chosen for the subsequent bark liquefaction reactions in this study.

20 Residue ratio (%) Residue

0 0 1 2 3 4 5 6 7 8 Catalyst loading (%)

Figure 4.7 Effect of catalyst loading on the residue ratio

(Reaction temperature=150°C, phenol/bark ratio = 3, reaction time = 60 min.)

Catalyst loading played a significant role in affecting the amount of free phenol in the liquefied bark faction (shown in Figure 4.8). When the catalyst loading increased from 1% to 3%, the amount of free phenol decreased significantly. As the catalyst loading exceeded 3%, the amount of free phenol decreased further but more gradually. Higher amount of catalyst loading promoted the extent of bark liquefaction in two aspects, which includes promoting the degradation of bark components and the reaction of phenol with the decomposed products during the liquefaction. Similar results were observed for wood liquefaction [22-26].

100 90 80

70 60 50 40

Free phenol (%) Free 30 20 10 0 0 1 2 3 4 5 6 7 8 Catalyst loading (%)

Figure 4.8 Effect of catalyst loading on the free phenol content

(Reaction temperature=150°C, phenol/bark ratio = 3, reaction time =60 min.)

4.3.1.5 Free phenol content in relation to bark liquefaction conditions

The changes in free phenol content of the liquefied bark fraction as a function of the liquefaction conditions are shown in Figure 4.9. The free phenol content highly depended on the reaction conditions, including reaction temperature, reaction time, catalyst loading and phenol/bark ratio as well as the amount of phenol used for the liquefaction. Among these parameters studied, catalyst loading was the most effective factor affecting the free phenol content of the liquefied bark fraction, followed by the reaction temperature, reaction time and phenol/bark ratio.

100 Beetle infested pine bark

1%1%, 150℃,60min, ReactionReaction time time P/B=3 80 CatalystCatalyst loading loading ReactionReaction temperature temperature 120℃, 3%, Phenol/BarkPhenol/Bark ratio ratio 120 2%, 150℃,60min, P/B=3, 60min℃ 2% 60 P/B=3 P/B=5, 30min30min, 3%, 150℃, P/B=3 P/B=4, 3%, P/B=5 P/B=4 3%, 60min, 150℃ 60min, 3%, 60min, 150 , P/B=3 P/B=2, 3%, ℃ 150℃ P/B=2 90min,90min 3%, 150℃, P/B=3 ℃ 40 60min, 150 120min,120min 3%, 150℃, P/B=3 150min 150min, 3%, 150℃, P/B=3 180℃, 3%,180℃ 5%,5% 150℃,60min, P/B=3, 60min P/B=3 7%,7% 150℃,60min, Free Free phenol content (%) 20 P/B=3

0 15 20 25 30 35 40 45 50 55

Total phenol weight (g)

Figure 4.9 Free phenol content of the liquefied bark fraction

4.3.2 Effect of reaction conditions on the bark liquefaction residues

The residue ratio of bark liquefaction was significantly affected by the reaction time, reaction temperature, catalyst loading level, and phenol/bark ratio. The liquefaction residues were consisted of both unliquefied solid biomass and recondensed products newly formed during the liquefaction process. The tendency for recondensation among the liquefied products was dependent on the liquefaction conditions. For acid-catalyzed phenol liquefaction, the phenol acted not only as a liquefaction solvent but also as a cross- linker for the degraded components [35, 81].

4.3.2.1 Reaction temperature

Figure 4.10 shows the effects of reaction temperature on the chemical composition of the unliquefied residues. With the reaction temperature increased from 120°C to 150°C, the lignin content decreased that resulted in a higher percentage of holocellulose in the residue. When the temperature was further increased from 150°C to 180°C, although the lignin

content was further decreased, the amount of holocellulose did not show obvious increase probably due to the thermal degradation of hemicellulose at around 180°C as indicated by a slight increase in the alpha cellulose ratio. It was also reported that recondensation reaction among decomposed components might occur at 180°C, which could have also affected the chemical composition of the residues [81].

70 Holocellulose Alpha-cellulose Lignin

(%) 30

0 120 150 180 Reaction temperature (℃)

Figure 4.10 Chemical composition of residues from different reaction temperature

Phenol/bark ratio = 3, catalyst = 3 wt.% of phenol, reaction time = 60 min.

The FTIR spectra of the residues from different reaction temperatures are shown in Figure 4.11. The peak assignments were conducted based on the published literatures [35, 81].

180°C

150°C

Transmittance Transmittance 1713 C=O

675 CH 1364,1317 phenolic OH

and G ring of lignin

120°C

2919 CH2

1485 C=C 1590 phenolic 1054,1028, C-O, C-H 3358 OH ring in carbohydrates 1153 CH 3500 3000 2500 2000 1500 1000 500 Wavenumber cm-1

Figure 4.11 FTIR of residues from different reaction temperatures

(Phenol/bark ratio = 3, catalyst = 3 wt.% of phenol, reaction time = 60 min.)

The intensity of the peak at 1364cm-1 due to phenolic hydroxyl groups decreased with the increasing reaction temperature. The peak at 1054cm-1 representing the C-O deformation in secondary alcohols and aliphatic esters of carbohydrates weakened when the reaction temperature increased from 120°C to 180°C. These changes indicated that with the increasing reaction temperature, the bark components such as lignin, cellulose and hemicellulose experienced more degradation. However, the intensity of the peak at 1590cm-1 and 1485 cm-1 attributed to aromatic skeletal vibration decreased when reaction temperature increased from 120℃ to 150°C, and increased again when the temperature increased to 180°C. It could be attributed to the recondensation reaction among decomposed components, which were also reported by previous literatures [26,81].

4.3.2.2 Phenol/bark ratio

The effect of phenol/bark ratio on the chemical composition of the residue is shown in Figure 4.12. When the phenol/bark ratio was increased from 2 to 3, the lignin content decreased while the holocellulose content increased remarkably. Due to the similar phenolic structures of the lignin and phenol, lignin would be more easily dissolved in phenol than cellulose. Increasing the phenol/bark ratio promoted the removal of the lignin in the liquefaction process that resulted in the relative increase in the holocellulose content. When the phenol/bark ratio was further increased from 3 to 5, the amount of holocellulose, alpha-cellulose and lignin all decreased, which could be due to the fact that when the phenol/bark ratio exceeded a certain amount it would not only promote the dissolution of lignin, hemicellulose and cellulose but also could retard the recondensation reactions among the degraded components.

70 Holocellulose Alpha-cellulose Lignin 60

40 ) % ( 30

0 2 3 4 5 Phenol/Bark

Figure 4.12 Chemical composition of the residues in relation to different phenol/bark ratios

(Reaction temperature=150°C, catalyst = 3 wt.% of phenol, reaction time =60 min.)

Phenol/bark ratio also affected the characteristics of FTIR spectra of the residues (shown in Figure 4.13). The intensity of the peaks at 1054cm-1 and 1028cm-1 representing the C-O deformation in secondary alcohols and aliphatic esters of carbohydrates, C-O deformation in primary alcohols and the peaks at 1364cm-1 and 1317cm-1 attributed to phenolic hydroxyl groups and G-rings of lignin significantly decreased when phenol/bark ratio exceeded 3, which indicated that the hemicellulose, cellulose and lignin had undergone more degradation under a higher phenol/bark ratio. The peaks at 675cm-1 and 594cm-1 representing C-H of benzene ring became stronger with the increasing phenol/bark ratio, which could be attributed to more reacted phenol. It would also be the reason why the free phenol content of the liquefied bark fraction gradually increased with the increasing phenol/bark ratio.

Phenol/bark ratio = 5!

Phenol/bark ratio = 4!

1713 C=O

675,594 C-H

Phenol/bark ratio = 3! Transmittance Transmittance

1364,1317, phenolic

OH,G ring of lignin

Phenol/bark ratio = 2 ! 1153 CH

2919, 2854, CH 2 1054, 1028, C-O, C-

H in carbohydrates 3358 OH! 1590,1510, phenolic rings 3500 3000 2500 2000 1500 1000 500 Wavenumber cm-1 Figure 4.13 FTIR of residues from different phenol/bark ratios

(Reaction temperature=150°C, catalyst = 3 wt.% of phenol, reaction time =60 min.)

4.3.2.3 Reaction time

As shown in Figure 4.14, the reaction time also significantly affected the chemical components of the residues after bark liquefaction. With the reaction time prolonged from 30min to 120min, holocelluose, alpha-cellulose and lignin contents all decreased. When the reaction time was higher than 120 min, the holocelluose, alpha-cellulose and lignin contents increased again, which was quite similar to the relationship between the residue ratio and reaction time as discussed before. After the reaction time reached higher than 120min, significant recondensation reactions had occurred and as a result, the holocellulose and lignin contents of the residues increased.

70 Holocellulose Alpha-cellulose Lignin 60

) % ( 30

0 30 60 120 150 Reaction time (min)

Figure 4.14 Chemical composition of residues from different reaction time

(Reaction temperature=150°C, catalyst = 3 wt.% of phenol, phenol/bark ratio = 3)

The differences in the FTIR spectra of the residues acquired from different reaction times are shown in Figure 4.15. The peak intensity at 1364cm-1 and 1317cm-1 representing G- rings of lignin decreased with the reaction time prolonged from 30 min to 150min. The intensity of the peak at 1054cm-1 and 1028cm-1 attributed to the C-O deformation in secondary alcohols and aliphatic esters of carbohydrates and C-O deformation in primary alcohols also decreased as the reaction time increased from 30 min to 120min. When the

reaction time was further prolonged to 150 min, the intensity of the peak at 1054cm-1 increased again. The changes of those characteristic peaks were consistent with the changes in the residue ratio and residue chemical composition discussed in previous sections.

150min!

120min!

1713 C=O

675, CH

60min! Transmittance Transmittance

1364,1317, phenolic

OH,G ring of lignin

30min!

2919, 2854, CH2

1054, 1028, C-O, C- 1590,1510, phenolic rings H in carbohydrates 3358 OH!

3500 3000 2500 2000 1500 1000 500 Wavenumber cm-1 Figure 4.15 FTIR of residues from different reaction time

(Reaction temperature=150°C, catalyst = 3 wt.% of phenol, phenol/bark ratio = 3)

4.3.2.4 Catalyst loading

The effect of catalyst loading on the chemical component of the residue is shown in Figure 4.16. Higher catalyst loading promoted the liquefaction process and the dissolution of lignin, hemicellulose and cellulose. The acid catalyst not only catalyzed the liquefaction reaction but also played a significant role in the acceleration of hydrolysis of cellulose and

hemicellulose. Higher catalyst loading helped dissolving a higher amount of lignin during the liquefaction and left lower amounts of holocellulose and alpha-cellulose in the residues.

70 Holocellulose Alpha-cellulose Lignin 60

) % ( 30

0 1 3 5 7 Catalyst loading (%)

Figure 4.16 Chemical composition of the residues from different catalyst loadings

(Reaction temperature=150°C, phenol/bark ratio = 3, reaction time =60 min.)

Figure 4.17 shows the differences in the FTIR spectra of the residues from various catalyst loadings in liquefaction. The intensity of the peaks at 1054cm-1 and 1028cm-1 representing the C-O deformation in secondary alcohols and aliphatic esters of carbohydrates, C-O deformation in primary alcohols and the peaks at 1364cm-1 and 1317cm-1 attributed to phenolic hydroxyl groups and G-rings of lignin all significantly decreased with the increasing catalyst loading. The peaks at 675cm-1 and 594cm-1 representing C-H in the benzene ring increased with the increasing catalyst loading, which could be attributed to the fact that higher catalyst loading promoted the reaction of phenol with the liquefied products as well as recondensation among the degradation products.

7% catalyst!

5% catalyst!

3% catalyst! Transmittance Transmittance

1364,1317, phenolic

OH,G ring of lignin 675, 594, CH

1% catalyst!

1054, 1028, C-O, C- 1153 CH H in carbohydrates 2919, 2854, CH2

1590,1510, phenolic rings 3358 OH!

3500 3000 2500 2000 1500 1000 500 Wavenumber cm-1 Figure 4.17. FTIR of residues from different catalyst loadings

(Reaction temperature=150°C, phenol/bark ratio = 3, reaction time =60 min.)

4.3.3. Effect of reaction conditions on the morphology of the residues

From Figure 4.18 we can see that the morphology of the residues after liquefaction with 1% catalyst loading was quite similar to the original bark. The bundles of cellulose fibers and micro fibers could still be clearly seen. It could be explained as that liquefaction carried out under low catalyst loading conditions would dissolve lignin preferably, which was the most sensitive part to phenol liquefaction. While the cellulose, considered as the most resistant part to phenol liquefaction, was left to form a larger portion of the residues. When the catalyst loading increased, the morphology of the residues changed significantly. Small particles were the main portion of the residues, indicating that the higher catalyst loading promoted the liquefaction process by accelerating the lignin, hemicellulose and cellulose dissolution, the long chain cellulose was broken down; only small particles

consisting of the degraded cellulose and lignin or recondensed compounds were left after the liquefaction. These results were consistent with the findings from the chemical compositional analysis of the residues, which showed that the cellulose and lignin contents of the residues decreased with the increasing catalyst loading used in the liquefaction.

! ! ! !

(a) (b) (c) (d)

Figure 4.18 SEM of residues from different catalyst loadings

(a) Original beetle infested pine bark. (b) Residues from bark liquefaction with 1% of catalyst loading. (c) Residues from bark liquefaction with 5% of catalyst loading. (d) Residues from bark liquefaction with 7% of catalyst loading. Reaction temperature=150°C, phenol/bark ratio = 3, reaction time =60 min.

Although reaction temperature, reaction time and phenol/bark ratio affected the residue ratio in bark liquefaction, free phenol content in the liquefied bark fraction and chemical composition of the residues, significant differences in the morphology of the residues acquired from different reaction temperatures, reaction times and phenol/bark ratios were not observed (Figure 4.19 to Figure 4.21).

! ! !

(a) (b) (c)

Figure 4.19 SEM of residues from different reaction temperatures

(a) Residues from bark liquefaction at 120°C. (b) Residues from bark liquefaction at 150°C. (c) Residues from bark liquefaction at 180°C. Phenol/bark ratio = 3, catalyst = 3 wt.% of phenol, reaction time = 60 min.

! ! ! !

(a) (b) (c) (d)

Figure 4.20 SEM of residues from different reaction time

! ! ! !

(a) (b) (c) (d)

Figure 4.21 SEM of residues from different phenol/bark ratios

(a) Residues from bark liquefaction with phenol/bark ratio = 2. (b) Residues from bark liquefaction with phenol/bark ratio = 3. (c) Residues from bark liquefaction with phenol/bark ratio = 4. (d) Residues from bark liquefaction with phenol/bark ratio = 5. Reaction temperature=150°C, catalyst = 3 wt.% of phenol, reaction time =60 min.

4.3.3.4 TGA comparisons

(a) TGA of bark before liquefaction

Figure 4.22 shows the thermal degradation of the beetle infested lodgepole pine bark. The three main components cellulose, hemicellulose and lignin underwent the decomposition between 200°C to 700°C and the peaks representing different chemical components of the bark were reflected in the TGA curve. Drastic weight of loss started approximately around 200°C, due to the degradation of hemicellulose and cellulose of the bark. Lignin began to decompose at around 350°C because of its relatively stable aromatic rings in the structure. The shoulder peak at around 200-320°C of the DTG curve was caused by the thermal degradation of hemicellulose and cellulose of the bark, whereas lignin was mainly responsible for the peak at around 350-500°C. A small peak of the DTG curves at around 600°C could be attributed to the burning of the char formed from lignin during the thermal degradation process.

100 0.6 1 st 0.5 Deriv. Weight /Temperature (%/°C ) 80 Weight loss

Deriv. weight loss 0.4

60 0.3

0.2

Weight (%) Weight 40

0.1 20 0

0 -0.1 0 100 200 300 400 500 600 700 Temperature (℃) !

Figure 4.22 TGA of beetle infested lodgepole pine bark

(b) TGA of the liquefied residues

During the liquefaction, the bark chemical components, cellulose, hemicellulose and lignin, underwent degradation and phenolation; the changes of the chemical composition of liquefied residues with liquefaction conditions were discussed above. These changes also affected the thermal stability of the liquefied residues. Similarity in the shape from the TGA curves of bark and liquefied residues from various liquefaction conditions was observed. However, the remaining weights of liquefied residues after the thermal degradation were different and affected by the liquefaction conditions. The TGA curves and remaining weight of the liquefied residues from low phenol/bark ratio (phenol/bark=2), less reaction time (30min), lower reaction temperature (120°C), and low catalyst loading (1%) were more similar to the original bark before liquefaction. It indicated that under those conditions, less bark chemical components were liquefied. With increasing phenol/bark ratio, reaction temperature, reaction time and catalyst loading,

more bark chemical components degraded during liquefaction, less cellulose, hemicellulose and lignin were left in the residues, the relative content of the ash content in the residues increased, and therefore, the remaining weight of the liquefied residues increased.

100

1% catalyst 3% catalyst

Weight (%) Weight 40 5% catalyst 7% catalyst Bark

0 0 100 200 300 400 500 600 700 Temperature (℃)

Figure 4.23 TGA of residues from different catalyst loading

(Reaction temperature=150°C, phenol/bark ratio = 3, reaction time =60 min.)

4.4 Conclusions

In this chapter, mountain pine beetle infested lodgepole pine barks were liquefied in phenol using sulfuric acid as a catalyst. Results showed that reaction conditions, such as reaction time, temperature, phenol/bark ratio, and catalyst loading, had significant effects on the liquefaction yield, the free phenol content in the liquefied fraction and the properties of the residues from bark liquefaction.

Increasing phenol/bark ratio and catalyst loading decreased the residue ratio, more cellulose were converted to methanol soluble compounds while fewer amounts of cellulose short fibers and lignin were left in the residues. Increasing the reaction temperature from 120°C to 150°C reduced the residue ratio. After the temperature reached 150°C, the residue ratio leveled off even as the temperature was further increased. The residue ratio decreased continuously with a longer reaction time, but after 120 min, residue ratio increased again. Higher reaction temperature and prolonged reaction time promoted the degradation of the bark components during liquefaction but induced recondensation reactions among the degraded bark components. The free phenol content increased with the increasing phenol/bark ratio while decreased with the increasing catalyst loading, reaction temperature and reaction time. Catalyst loading was the most effective factor affecting the liquefaction yield, free phenol content of the liquefied bark fraction and residues properties, followed by the reaction temperature, reaction time and phenol/bark ratio.

Chapter 5

Characterization of Phenol Formaldehyde Resins Derived from Liquefied Lodgepole Pine Barks

5.1 Abstract

In this chapter, barks from the lodgepole pine (Pinus contorta Dougl.) with and without the mountain pine beetle (MPB, Dendroctonus ponderosae Hopkins) infestation were liquefied in phenol with sulfuric acid. The liquefied portions of the bark were used to synthesize liquefied bark-PF (LBPF) adhesive resins under alkaline conditions without further addition of phenol. In comparison to a commercial phenol-formaldehyde (PF) resin and a lab PF resin, the LBPF resins were found to have higher average molecular weights, higher polydispersity indices and shorter gel times. The viscosities of the LBPF resins were higher than the viscosity of the lab PF resin, but lower than the viscosity of the commercial PF resin. Isothermal DSC tests indicated that all resins exhibited both nth- order and autocatalytic cure mechanisms. The post-curing thermal stability of the LBPF resins was similar to that of the lab PF resin at higher temperatures, but differed significantly from that of the commercial PF resin. All these resins had similar dry bonding strengths; the LBPF resins showed the highest wet bonding strengths. Beetle infestation was shown to have no negative effect on the bonding properties of the LBPF resins.

5.2. Introduction

Bark is generally available in large quantities as waste residues from conversion of wood logs to various forest products. In mills, bark is usually mixed with other woody residues and used as hog fuel for heat recovery. However, bark has a rather low heating value and its heating value drops sharply particularly when wet. Therefore, exploring other value-

added, more cost- and energy-efficient uses for bark would be highly advantageous for the forest mills.

Bark has a similar chemical composition to wood except that it contains more extractives and phenolic compounds [3]. Previous research has shown that lignin and bark extractives, such as tannin, could partially replace 30%~50% of petroleum derived phenol, for synthesizing PF resins [10-15]. However, wood composites made using these types of adhesives have previously shown inconsistent and unsatisfactory performance due to the highly variable chemical structure of the lignin obtained through different origins and methods, as well as the low reactivity of lignin and flavonoids B-ring units within tannins. Chemical modification of lignin or tannin, such as liquefaction or phenolation, was found to be able to effectively improve the reactivity and produce resins with more satisfactory performance [10-12, 17].

Previous studies [15, 38-41] have shown that bark can be successfully liquefied in the presence of phenol using an acid catalyst. The liquefied products can be subsequently converted to liquefied bark-PF (LBPF) resins to be used for manufacturing plywood and particleboard. It was found that the mechanical properties of the wood composite products bonded with the LBPF adhesives were comparable to those bonded with commercial PF resins. However, no studies have reported on the curing kinetics and curing behavior of LBPF resins.

Recently, due to rampant mountain pine beetle (Dendroctonus ponderosae Hopkins) infestation outbreaks in the western provinces of Canada, there is a large amount of beetle infested lodgepole pine (Pinus contorta Dougl.) resources that is available for utilization. Liquefaction of bark from the beetle-infested lodgepole pine and conversion of the liquefied bark to LBPF resins as wood adhesives could be one of the attractive possibilities. However, the impact of beetle infestation on the suitability of bark as a raw material for resin synthesis is unknown.

In chapter 4, the bark from the mountain pine beetle infested lodgepole pine was successfully liquefied in phenol using sulfuric acid as the catalyst. The effects of reaction conditions such as reaction temperature, reaction time, catalyst loading and phenol/bark

ratio on the liquefaction yield and liquefied bark and residues were studied. In this chapter, the liquefied bark acquired from reaction temperature 150°C, reaction time 120 min, phenol/bark ratio 3 and catalyst loading 3% was used as the alternative phenolic feedstock for the liquefied bark-PF (LBPF) resin formulation. Without the need to remove any excess phenol and without the need to add additional phenol, the liquefied bark mixtures was reacted with formaldehyde under alkaline conditions to give liquefied bark-PF resin (denoted as LBI-PF), which were then tested and compared with a lab PF resin and a commercial PF resin. Meanwhile, bark from the lodgepole pine without mountain pine beetle infestation will be also liquefied and converted to liquefied bark-PF resin (denoted as LGP-PF) following exactly the same steps of LBI-PF resin formulation, with the aim of checking the suitability of the beetle infested lodgepople pine as a raw material for resin synthesis.

A version of this chapter was published in International Journal of Adhesion & Adhesives (2010).

5.3 Results and Discussion

5.3.1 Basic properties of the liquefied bark-PF resins

The properties of the synthesized resins are shown in Table 5.1. The pH values for the liquefied bark-PF resins and the lab PF resin were similar. The solids content of liquefied mountain pine beetle infested pine bark-PF (LBI-PF) and liquefied non-infested pine bark- PF (LGP-PF) resins were 52.97% and 52.35%, respectively. The solids content of the bio- based resins was lower than the solids content of the commercial PF resin (59.00%), but was higher than the solids content of the lab PF resin (48.87%). It is known that commercial PF resins for OSB contain significant amounts of urea. This is why it has a very high solids content.

Even though the same synthesis procedure was followed, the liquefied bark-PF resins had higher viscosities, higher molecular weights (Mw and Mn) and higher polydispersity indices than the lab PF resin. This could be the result of the presence of some larger molecular compounds, i.e. degraded bark components in the liquefied bark fraction. The

low viscosity, low Mw and low polydispersity index of the lab PF resin (25 cp) indicated that this resin had a low degree of polymerization, which also explained why its solids content was lower than those of liquefied bark-PF resins.

The liquefied bark-PF resins had the shortest gel time when compared to the lab and commercial PF resins. The shorter gel time does not mean that liquefied bark-PF resins have faster cure rates than the PF resins. As can be seen in the following Section 5.3.2, the commercial PF had the lowest peak temperature and hence probably the highest reactivity. Empirically, gel time can sometimes give indications about resin cure rates but conclusions should not be drawn on gel time alone, such as demonstrated in this study. Phenolic resin gelation can sometimes be a physical phenomenon (reversible gelation) instead of a chemical one (irreversible gelation).

Table 5.1 Resin properties

pH Solids Viscosity Gel time Mn (Da) Mw (Da) Mw/Mn content (cp) at 120°C (%) (s)

Com PF 11.16 59.00 200 172 2.12×102 3.86×102 1.82

LBI-PF 12.07 52.97 125 152 3.26×102 6.19×102 1.89

LGP-PF 11.96 52.35 150 161 4.37×102 8.49×102 1.95

Lab PF 11.93 48.87 25 173 2.59×102 3.27×102 1.25

(LBI-PF: Liquefied mountain pine beetle infested lodgepole pine bark-PF resin; LGP-PF: Liquefied non-infested lodgepole pine bark-PF resin, Lab PF: Laboratory made PF resin; Com PF: Commercial PF resin.)

5.3.2 Adhesives curing behavior

Figure 5.1 Dynamic DSC curves of the liquefied non-infested lodgepole pine bark-PF resin (LGP-PF)

The DSC curves of the liquefied non-infested lodgepole pine bark-PF resins at different heating rates had a single exothermic peak (Figure 5.1). The same was true for all other types of resins (curves not shown). These curves were consistent with what were typically observed for PF resins [82]. The DSC curves also seemed to suggest that the curing process of the resins was mainly dominated by condensation reactions. The addition reactions for the formations of methylol compounds might have largely completed during the resin synthesis stage.

The onset and peak temperatures of the resins are listed in Table 5.2. With an increasing heating rate, both the onset temperature and peak temperature shifted to higher

temperature. Since the actual cure temperatures should be independent of the heating rate, peak and onset temperatures were extrapolated to the heating rate of zero for comparison [83, 84]. The extrapolated onset temperatures for both the liquefied mountain pine beetle infested lodgepole pine bark-PF resin and liquefied non-infested lodgepole pine bark-PF resin were found to be 99°C. The extrapolated onset temperatures for the liquefied bark- PF resins were similar to the commercial PF resin, but were slightly higher than the lab PF resin. It indicated that the lab PF resin was more reactive at the lower temperatures.

The extrapolated peak temperatures at 0°C /min heating rate for the liquefied mountain pine beetle infested pine bark-PF resin and liquefied non-infested pine bark-PF were 134°C and 133°C, respectively. The extrapolated peak temperatures of the commercial PF resin and lab PF resin were 128°C and 136°C, respectively. It implied that the commercial PF resin had a higher reactivity than the liquefied bark-PF resins and lab PF resin at higher temperatures. The lab PF resin was more reactive at lower temperature while the LBPF resins were more reactive at the higher temperature, and the commercial PF being the most reactive resin at the higher temperatures. The degraded bark components might have contributed to the different reactivity among different types of resins at different temperatures. Since the detailed composition for the commercial PF resin is unknown to us, it would not be too meaningful to speculate the mechanism associated with the commercial PF resin.

Table 5.2 Resin curing temperatures

LBI-PF LGP-PF Commercial PF Lab PF

Heating rate Onset TempPeak Temp Onset Temp Peak Temp Onset Temp Peak Temp Onset Temp Peak Temp

(°C /min) (°C) (°C) (°C) (°C) (°C) (°C) (°C) (°C)

0 * 99 134 99 133 98 128 96 136

5 104 141 105 140 106 134 104 144

10 113 151 114 151 114 144 113 156

15 118 160 119 159 122 151 121 165

20 123 165 125 165 129 157 129 171

* Extrapolated values from the intercept of the plots of the onset temperatures and peak temperatures versus the heating rate.

LBI-PF: Liquefied mountain pine beetle infested lodgepole pine bark-PF resin;

LGP-PF: Liquefied non-infested lodgepole pine bark-PF resin

The activation energies were calculated using the Kissinger equation and are given in Table 5.3. Other dynamic parameters, including the kinetic equations derived from the dynamic DSC analysis, are also shown in Table 5.3. Compared with the lab PF resin, the LBPF resins had higher activation energies, which seemed to suggest that the degraded bark components in the liquefied bark fraction might have made the resin more difficult to cure. One possible mechanism could be related to the lower molecular mobility of these components to form crosslinking. Even though the cure activation energy was higher for the LBPF resins, the reactive sites on the bark components might have shortened the gel time and accelerated the curing process of the resins at the higher curing temperatures where cure reaction was mainly controlled by diffusion. The commercial PF resin had the highest activation energy while its value of the pre-exponential factor was the highest among the different types of resins. The latter indicated that the commercial PF resin had the fastest curing rate.

Table 5.3 Dynamic cure kinetic parameters for different resins

E (kJ/mol) A (s-1) Kinetic Equation r

dα 90.849400 LBI-PF 78.12 2.08X109 =2.08× 10 exp(- )(1-α ) 0.99 dt T

dα 90.759338 LGP-PF 77.60 1.85X109 =1.85× 10 exp(- )(1-α ) 0.99 dt T

dα 80.648450 Lab PF 70.22 1.59X108 =1.59× 10 exp(- )(1-α ) 0.99 dt T

Commercial dα 109895 0.56 82.23 1.18X1010 =1.18× 10 exp(- )(1-α ) 0.99 PF dt T r: correlation coefficient

Figures 5.2 to 5.5 show the resin conversion rate versus the degree of conversion under different isothermal temperatures. It is evident from Figures 5.2 to 5.5 that the curing reactions for all resins followed an nth-order kinetic mechanism when the conversion rate was below 10%. When the conversion rate reached above 20%, the cure reactions became autocatalytic. The autocatalytic cure mechanism dominated at higher curing temperatures. The change in the curing mechanism as a function of temperature was also observed for some commercial PF resins before [83, 84].

0.025

0.02 110℃ 120℃ 130℃ 140℃ 150℃ 0.015

0.01 Conversion rate(/s) Conversion

0.005

0 0 10 20 30 40 50 60 70 80 90 100 Conversion(%)

Figure 5.2 Conversion rate as a function of the degree of conversion at various isothermal temperatures for the liquefied mountain pine beetle infested lodgepole pine bark-PF resin (LBI-PF)

0.02

110℃ 120℃ 0.015 130℃ 140℃ 150℃

0.01 Conversion rate (/s) rate Conversion 0.005

0 0 10 20 30 40 50 60 70 80 90 100 Conversion (%)

Figure 5.3 Conversion rate as a function of the degree of conversion at various isothermal temperatures for the liquefied non-infested lodgepole pine bark-PF resin (LGP-PF)

0.016

0.014 110℃ 0.012 120℃ 130℃ 140℃ 0.01 150℃

0.008

0.006 Conversion rate (/s) rate Conversion 0.004

0.002

0 0 10 20 30 40 50 60 70 80 90 100 Conversion (%)

Figure 5.4 Conversion rate as a function of the degree of conversion at various isothermal temperatures for the lab PF resin

0.05

0.045 110℃ 0.04 120℃ 0.035 130℃ 140℃ 0.03 150℃

0.025

0.02

Conversion rate (/s) rate Conversion 0.015

0.01

0.005

0 0 10 20 30 40 50 60 70 80 90 100 Conversion (%)

Figure 5.5 Conversion rate as a function of the degree of conversion at various isothermal temperatures for the commercial PF resin

Conversion rates of different resins as a function of time at various isothermal temperatures are shown in Figures 5.6 to 5.9. All the resins reached about 60% of conversion within 25 minutes at the five isothermal temperatures. With the increasing isothermal cure temperature, the cure reaction reached the same level of conversion at a shorter time. The shape of the curves changed with different isothermal cure temperatures. The shape change of the curves for LBI-PF resin was similar to the lab PF resin, but was different from the change of LGP-PF and commercial PF resins. These changes were probably due to the change in the cure mechanism. At high cure temperature, the autocatalytic mechanism might have dominated the cure reaction.

100

50 ℃ 40 110 120℃ Conversion(%) 30 130℃ 140℃ 20 150℃

0 0 5 10 15 20 25 30 35 40 Time(min)

Figure 5.6 Conversion as a function of cure time at various isothermal temperatures for the liquefied mountain pine beetle infested lodgepole pine bark-PF resin (LBI-PF)

100

50 110℃ 40 120℃

Conversion(%) 130℃ 30 140℃ 150℃ 20

0 0 5 10 15 20 25 30 35 40 Time(min)

Figure 5.7 Conversion as a function of cure time at various isothermal temperatures for the liquefied non-infested lodgepole pine bark-PF resin (LGP-PF)

100

40 110℃ Conversion (%) Conversion 120℃ 30 130℃ 20 140℃ 150℃ 10

0 0 5 10 15 20 25 30 35 40 Time (min)

Figure 5.8 Conversion as a function of cure time at various isothermal temperatures for the lab PF resin

100

40 110℃ 120℃ Conversion (%) Conversion 30 130℃ 140℃ 150℃ 20

0 0 5 10 15 20 25 30 35 40 Time (min)

Figure 5.9 Conversion as a function of cure time at various isothermal temperatures for the commercial PF resin

5.3.3 Thermal stability of the cured resins

The thermal degradation process of the commercial PF resin differed significantly from those of the lab PF and LBPF resins, especially when the temperature was higher than 200°C (Figure 5.10). This could be due to the fact that commercial PF resin contained a significant amount of urea that is known to decompose more easily at high temperatures. The lab PF resin and LBPF resins had more similar thermal degradation processes. The lab PF resin had a slightly higher amount of weight retained after thermal degradation compared with the LBPF resins.

Figure 5.10 Thermal stability of different cured resins as measured by TGA

5.3.4 Lap shear test results

Results of the lap-shear tests of two-layer specimens bonded with four different types of resins are shown in Figure 5.11. The specimens bonded with LBPF resins had similar dry shear strength values as the specimens glued by the commercial and lab PF resins. After the water-soaking-and-drying (WSAD) treatment and the boiling water treatment (BWT/wet), all samples showed no delamination. The LBPF resins gave higher lap shear strength than the commercial and lab PF resins after both the WSAD and BWT/wet tests. According to the t-test, there is no significant difference between the dry shear strength of the specimens bonded with LBPF and commercial and lab PF resins at the level of p=0.05. While the wet shear strength of the samples bonded with LBPF was significantly higher than those bonded with commercial and lab PF at the level of p=0.05. Some previous studies on condensed tannins-polyethylenimine resins and bark extractives-diisocyanate

mixtures also showed good wet bonding strength in resulting wood composites [16, 85]. The researchers attributed the good bonding properties to the extractives and to the effects from the catechol moiety of the tannins in the bark, serving as the wet strength agent in wood composites, which is consistent with the findings in this study. Furthermore, the bonding properties of the liquefied bark-PF resins did not seem to be negatively affected by the beetle infestation of the lodgepole pine bark.

5.4 Conclusions

The bark-phenol-formaldehyde resins synthesized with liquefied barks of beetle-infested lodgepole pine bark and non-infested lodgepole pine bark were found to have higher average molecular weights, higher polydispersity indices and shorter gel times than Lab PF resin and commercial PF resin. The cure activation energy for the LBPF resins was lower than that of the commercial PF resin but was higher than the cure activation energy

of the lab PF resin. The curing behavior of all these resins followed both nth-order and autocatalytic mechanisms. The cured LBPF resins had a more similar thermal stability as the cured lab PF resin, which was different from the commercial PF resin. LBPF adhesive resins showed similar dry bonding strength and better wet bonding strength than those of PF adhesive resins. Beetle infestation was shown to have no negative effect on the bonding properties of the LBPF resins.

Chapter 6

Bark Extractions and Bark Extractive-PF resins

6.1 Abstract

In this study, phenol-formaldehyde resins derived from bark extractives were synthesized and characterized. Bark of lodgepole pine (Pinus contorta Dougl.) infested by mountain pine beetle (Dendroctonus ponderosae Hopkins) was first extracted with 1% NaOH. The bark extractives with and without acid neutralization were then dried to the solid state. The neutralized and non-neutralized extractives were used to partially replace petroleum-based phenol for synthesizing bark extractives-phenol-formaldehyde resins. In comparison with a commercial phenol-formaldehyde (PF) resin and a laboratory made PF resin (Lab PF), the bark extractive-PF (BEPF) resins were found to have higher molecular weights, higher viscosities and shorter gel times. Acid neutralization of the bark extractives increased the molecular weight of the extractives and modified the performance and curing behavior of the resulting bark extractive-PF resins. Bark extractive-PF resins showed a similar level of post-cured thermal stability to that of the lab PF at higher temperatures, but they differed significantly from that of the commercial PF resin. The bark extractive-PF resins made from both neutralized and non-neutralized extractives at 30% replacement of phenol (by weight) exhibited similar dry and wet bond strengths to the commercial PF resin. At 50% substitution level, bark extractive-PF resins had dry and wet bond strength similar to the lab PF resin. Our findings suggest that alkaline extractives from mountain pine beetle- infested lodgepole pine bark are suitable for partially substituting phenol in the synthesis of phenolic resin for use in wood adhesives.

6.2 Introduction

Tannins, possessing polyphenolic molecular structures, have higher chemical reactivity towards formaldehyde than phenol and could be also used to partially replace phenol for synthesizing phenolic resins [86, 87]. One of the earliest attempts to produce tannins as

wood adhesives was reported in 1958 [5] and used radiata pine bark as the source. Previous research has shown successful application and satisfactory performance of tannin adhesives in particleboard and plywood manufacturing [13, 16, 42, 46-49, 51]. The most widely used industrial tannins are obtained from the wood of quebracho tree (Schinopsis balansae, Argentina) and from the bark of mimosa (Acacia mearnsii, Brazil and South Africa) [5]. Although various attempts were made to produce wood adhesives from tannin extracts of barks of different pine and spruce species, such as radiata pine (Pinus radiata D. Don), mature pine species (Pinus caribaea, Pinus elliottii, Pinus pinaster and Pinus sylvestris) and spruce species (Picea abies), the production of tannin from these tree species has not generally been successful on a commercial scale [5, 37]. The low extract yields, low reactivity of the extracts (low stiasny values, i.e. percentage of formaldehyde condensable components), excessive viscosity and variable extract quality were found to be the factors affecting the adhesives application [5, 86, 87]. The limited availability and high costs of the tannin extracts are also factors restricting their commercialization [5].

Various methods have been tried to improve the yield of tannin containing extractives and make it more desirable for PF resin synthesis. Multi-stage extractions of bark by alkali of different concentrations were able to improve the extractives yields [5, 8, 86-90]. Separation of the high molecular weight components such as highly polymerized procyanidin polymers from the extractives and sulfitation can be carried out before the resin synthesis. Sulfitation and phenolation processes can be used to reduce the molecular weight and increase the reactivity of the extractives [86-90].

Bark is a renewable biomass feedstock generally available in large quantities as waste residues from conversion of wood logs to various forest products. Bark has a similar chemical composition to wood [3], but it contains more extractives such as fatty acids, terpenes, suberins and phenolic compounds such tannins, which make it possible to utilize bark as a substitution for phenol in PF resin synthesis. It has been reported in the previous studies and previous chapters that bark can be liquefied in phenol and the liquefied bark can react with formaldehyde to give an adhesive resin that has bond strength comparable to those of commercial PF resins [40, 41, 91].

The phenolic compounds of bark extractives have been used as accelerators of PF resins in particleboard and plywood production due to their ability to minimize gelation and shorten press time [50-52, 59]. The resins derived from the copolymerization of Eucalyptus, Acacia mollissima and Pinus pinaster bark extracts with phenol and formaldehyde have been successfully applied to particleboard and plywood manufacturing [8, 92, 93]. Resorcinol-formaldehyde resins with up to 60 weight percentage of resorcinol replaced by bark extractives were successfully prepared and found to have excellent bond strength and cold-setting capability [93].

Recently, due to rampant mountain pine beetle (Dendroctonus ponderosae Hopkins) infestation in the western provinces of Canada, there is a large amount of diseased lodgepole pine (Pinus contorta Dougl.) available for utilization. Using the bark extractives from these affected lodgepole pine trees to partially replace phenol for the production of wood adhesives could be one of the attractive possibilities for generating value from this resource.

Compared with phenol liquefaction, bark alkaline extraction is a simpler and more straightforward method for obtaining alternative phenolic feedstock for PF resin synthesis. The extraction conditions are much milder than the phenol liquefaction. No strong acids or strong alkaline and no catalysts are needed for the extraction. If the PF resins made from the bark alkaline extractives from the mountain pine beetle infested lodgepole pine have similar bonding performance and curing behavior as the liquefied bark-PF resin, bark alkaline extraction will be more attractive for the future industrial scale-up and application.

In chapter 4 and chapter 5, we investigated the acid-catalyzed phenol liquefaction of the bark from mountain pine beetle infested lodgepole pine; the liquefied bark was subsequently used as an alternative feedstock of petroleum-based phenol for liquefied bark-PF resin synthesis. It was found that the wet bonding strength of liquefied bark-PF resin was better than that of the commercial PF resin, while the dry bonding strength was similar. The liquefied bark components affected the curing behavior of the resulting liquefied bark-PF resin. Acid-catalyzed phenol liquefaction is an effective conversion method for obtaining alternative phenolic feedstock for PF resins synthesis.

Therefore, in this chapter, bark from mountain pine beetle infested lodgepole pine is extracted with 1% sodium hydroxide solution. The extractives with and without acid neutralization are then dried to a solid state, which are then used to partially replace phenol for reacting with formaldehyde under alkaline condition to give bark extractives- phenol-formaldehyde resins. The resulting bark extractive-PF resins are tested and compared with a lab PF resin and a commercial PF resin.

A version of this chapter was published in Journal of Adhesion Science and Technology (2012).

6.3 Results and Discussion

6.3.1 Properties of the extractives

Alkaline-soluble extractives are expected to contain mainly polyphenolic compounds such as tannins along with other low molecular weight carbohydrates such as hemicellulose and degraded cellulose in the bark. The average yield of the 1% NaOH-extractives from mountain pine beetle-infested lodgepole pine bark calculated from equation 1 was 68.06%. The molecular weights of the alkaline extractives are shown in Table 6.1. The acid- neutralized extractives had higher molecular weights and polydispersity indices than the alkaline extractives. The molecular weights of the alkaline extractives decreased and their polydispersity indices increased after drying, indicating degradation and/or hydrolysis of the extractives during the drying process. These results were consistent with those reported by previous researchers that found tannins molecules could undergo hydrolysis and degradation under both alkaline and acidic conditions [1, 5, 8, 14, 37, 89, 90]. The increase in molecular weights and polydispersity indices of the acid-neutralized (pH 7) extractives after drying could be explained by the self-condensation of the extractive molecules during the drying process under pH-neutral condition [8, 90]. Compared with alkaline condition, neutral pH might cause auto-condensation and less hydrolysis of the extractives, and as a result, the molecular weight of the extractives was increased [8, 89, 90]. The pH-neutralization and drying process affected the molecular weight of the bark extractives. The reactivity of the bark extractives is not affected by the neutralization.

Table 6.1 Molecular weight of the extractives

Mn (Da) Mw (Da) Mw/Mn Stiasny number (%)

Alkaline Extractives (wet) 1.76×103 2.99×103 1.70 -

Alkaline Extractives (dry) 1.19×103 2.21×103 1.85 37.93

Neutralized Extractives (wet) 1.82×103 3.68×103 2.03 -

Neutralized Extractives (dry) 2.82×103 6.85×103 2.43 37.50

6.3.2 Properties of the bark extractive-PF resins

The solids contents of bark extractive-PF resins ranged from 46.44% to 49.05%, which were less than that of the commercial PF resin (Table 6.2). The pH values of the bark extractive-PF resins made from neutralized extractives (denoted as BEPF (N)) were slightly lower than of those resins made from alkaline extractives (denoted as BEPF). Even though the same synthesis procedure was followed, bark extractive-PF resins had higher viscosities, molecular weights (Mw and Mn) and polydispersity indices than the lab PF resin. This may be the result of the presence of some larger molecules in the bark extractives. The bark extractive-PF resins showed shorter gel times compared to those of the lab and commercial PF resins. Gel time is commonly used by the adhesives industry as an indirect indicator of the rate of polymerization.

At the same phenol-replacement level, the BEPF (N) resins had higher molecular weights, higher polydispersity indices, higher viscosity and shorter gel times than BEPF resins. This might be related to the degradation and/or hydrolysis of the alkaline extractives during the drying process, which led to the lower molecular weights and lower viscosities of the resulting BEPF resins [5, 8, 89, 90]. For the BEPF (N) resins, extra caution was needed during their synthesis to avoid the sudden drastic increase in the viscosity and subsequent gelation. The lab PF was synthesized only from phenol and formaldehyde in the presence of 40% NaOH solution without other additives or subsequent dehydration,

which resulted in a relatively low solids content and viscosity.

The average molecular weights of the bark extractive-PF resins were found to be lower than those of the starting bark extractives (Table 6.2). This is understandable because smaller molecules, such as formaldehyde and phenol, react with each other to form oligomers and low molecular weight polymers. Additionally, the bark extractives could have further hydrolyzed during the resin cooking process due to the alkaline reaction conditions.

Table 6.2 Properties of different resins

Viscosity Solids Gel time at pH at 25°C Mn (Da) Mw (Da) Mw/Mn content (%) 120 °C (s) (cP) 30% BEPF 11.68 47.90 800 58 3.82×102 7.76×102 2.03 50% BEPF 11.88 48.24 700 61 3.07×102 8.72×102 2.84 70% BEPF 11.84 46.44 450 65 3.63×102 9.43×102 2.60 30% BEPF (N) 11.05 49.05 900 49 6.63×102 1.18×103 1.76 50% BEPF (N) 11.25 48.91 800 57 7.05×102 1.19×103 1.68 70% BEPF (N) 10.72 46.50 650 68 4.08×102 1.03×103 2.52 Lab PF 11.93 48.87 25 173 2.59×102 3.27×102 1.25 Com PF 11.16 59.00 200 172 2.12×102 3.86×102 1.82

(Lab PF: Laboratory made PF resin; Com PF: Commercial PF resin; BEPF: Bark extractive-PF resins made from alkaline extractives; BEPF (N): Bark extractive-PF resins made from acid-neutralized extractives)

6.3.3 Adhesive curing behavior and curing kinetics

At 30, 50 and 70 by weight percentage replacement of phenol, the DSC curves for all bark extractive-PF resins had a single exothermic peak at different heating rates (curves not shown). This was consistent with what we had observed for other PF resins and for liquefied bark-PF resins in the previous studies [41, 75, 91]. The single exothermic peak in the DSC curves suggested that the curing process of the bark extractive-PF resins was mainly dominated by the condensation reactions. The addition reactions for the formation

of methylol compounds may have been largely completed during the resin synthesis stage.

The onset and peak temperatures of the bark extractive-PF (BEPF and BEPF (N)) resins both shifted to higher temperatures with increasing heating rate (Tables 6.3 and 6.4). The actual cure temperatures should be independent of the heating rate; thus, peak and onset temperatures were extrapolated to zero heating rate for comparison [75, 83, 84].

Table 6.3 Curing characteristics of bark extractive-PF resins made from alkaline extractives (BEPF)

30% BEPF 50% BEPF 70% BEPF Heating rate Onset temp. Peak temp. Onset temp. Peak temp. Onset temp. Peak temp. (°C/min) (°C) (°C) (°C) (°C) (°C ) (°C) 0 96 138 99 136 97 139 5 100 145 102 141 102 145 10 111 156 112 155 114 156 15 114 164 119 160 119 162 20 120 169 120 166 123 169

Table 6.4 Curing characteristics of bark extractive-PF resins made from acid-neutralized extractives (BEPF (N))

30% BEPF (N) 50% BEPF (N) 70% BEPF (N) Heating rate Onset temp. Peak temp. Onset temp. Peak temp. Onset temp. Peak temp. (°C/min) (°C) (°C) (°C) (°C) (°C) (°C) 0 98 128 99 121 105 146 5 103 136 106 123 117 149 10 111 146 115 134 126 157 15 118 155 122 135 138 160 20 121 161 129 138 149 164

The extrapolated onset temperatures for the bark extractive-PF resins (Tables 6.3 and 6.4) at different replacement levels of phenol (except 70% BEPF (N)) were similar to that of

the commercial PF resin (98°C) but slightly higher than the lab PF resin (95°C). This suggested that the lab PF resin was more reactive at low temperatures.

The extrapolated peak temperatures for BEPF resins (Table 6.3) at 30%, 50% and 70% replacements of phenol were similar to that of the lab PF (136°C) but higher than that of the commercial PF resin (127°C). This result implied that the commercial PF resin was more reactive at higher temperatures than the lab PF resin and BEPF resins.

The extrapolated peak temperatures for the BEPF (N) resins at 30% and 50% replacements of phenol were lower than that of the lab PF resin but close to that of the commercial PF resin (Table 6.4). This suggested that the acid-neutralization of the alkaline bark extractives and the subsequent drying at elevated temperatures might have helped increase the reactivity of the resulting BEPF (N) resins if the phenol replacement level did not exceed 50%. The extrapolated peak temperature for BEPF (N) resin at 70% substitution level was much higher than for others, which implied that the presence of more neutralized extractives with large molecular weight would reduce the reactivity of the resulting bark extractive-PF resins possibly due to the immobility and inaccessible reactive sites on the extractive molecules [8, 89, 90].

Table 6.5 shows the activation energies calculated using the Kissinger equation, as well as other kinetic parameters, including pre-exponential factors and the kinetic equations. Compared with the lab PF resin, the bark extractive-PF resins had higher activation energies, which seemed to suggest that the bark extractives might have made the resins more difficult to cure. One possible mechanism could be related to the lower molecular mobility of the bark extractives, which made it difficult for the resulting resins to form cross-links [1, 13, 16, 94].

Table 6.5 Cure kinetic parameters for different resins

E (kJ/mol) A (s-1) Kinetic Equation r

9 d! 9 9625 0.85 30% BEPF 79.98 2.78×10 = 2.78!10 exp(" )(1"!) 0.99 dt T

9 d! 9 9555 0.65 50% BEPF 79.40 2.73×10 = 2.73!10 exp(" )(1"!) 0.99 dt T

9 d! 9 10034 0.58 70% BEPF 83.38 7.55×10 = 7.55!10 exp(" )(1"!) 0.99 dt T

30% BEPF 9 d! 9 8872 0.82 73.73 7.17×10 = 7.17!10 exp(" )(1"!) 0.99 (N) dt T

50% BEPF 14 d! 14 13970 1.05 116.09 8.15×10 = 8.15!10 exp(" )(1"!) 0.99 (N) dt T

70% BEPF 17 d! 17 17385 0.84 144.47 3.45×10 = 3.45!10 exp(" )(1"!) 0.99 (N) dt T

8 d! 8 8450 0.64 Lab PF 70.22 1.59×10 =1.59 !10 exp(" )(1"!) 0.99 dt T

10 d! 10 9895 0.56 Com PF 82.23 1.18×10 =1.18!10 exp(" )(1"!) 0.99 dt T

(Lab PF: Laboratory made PF resin, Com PF: Commercial PF resin, BEPF: Bark extractives-PF resins made from alkaline extractives, BEPF (N): Bark extractives-PF resins made from acid-neutralized extractives, dα/dt: Conversion rate, r: Correlation coefficient)

The BEPF resins with 30%, 50% and 70% replacements of phenol and the BEPF (N) resin with 30% replacement of phenol by weight had higher curing rates than lab PF but lower curing rates than the commercial PF resin. The BEPF (N) resins with 50% and 70% replacements of phenol had much higher activation energies and pre-exponential factors than other bark extractive-PF resins. A possible reason for the higher activation energy is that the neutralized extractives with higher molecular weight could have blocked the movement of the molecules and restricted further reaction with the cross-linker [6, 8]. The reactive sites on the extractives and increased self-condensation of the extractives might have contributed to the relatively high pre-exponential factors and faster curing rates of the

resulting resins at these replacement levels [75, 83, 84, 91]. However, premature cure of the neutralized resins might occur causing incomplete cross-linking [1, 8, 89, 90]. The neutralization of the extractives before the drying process during the bark extraction step significantly affected the curing behavior of the resulting bark extractive-PF resins.

6.3.4 Thermal stability

The thermal degradation process of the commercial PF resin differed significantly from those of the lab PF, BEPF resins with 30%, 50% and 70% phenol replacements and BEPF (N) resins with 30% and 50% phenol replacements, especially when the temperature was higher than 200°C (Figure 6.1 and 6.2). This could be due to the fact that the commercial PF resin contained a significant amount of urea, which is known to decompose more easily at high temperatures. The lab PF resin, BEPF and BEPF (N) resins at 30% and 50% replacements of phenol showed similar thermal degradation behaviors. The BEPF (N) resin made with 70% replacement of phenol underwent more extensive thermal degradation at higher temperature, which could be due to insufficient cross-linking caused by the immobility of the large molecules of the extractives [8, 90]. The lab PF resin had a similar amount of weight retained after thermal degradation as the BEPF resins and BEPF (N) resins, except for the one made from neutralized extractives at 70% weight replacement of phenol.

100 Extractives 30% BEPF 90 50% BEPF 70% BEPF Com PF 80 Lab PF

70 Weight (%) Weight

40 0 100 200 300 400 500 600 700 Temperature (°C)

Figure 6.1 Thermal stability of post-cured BEPF resins

(Lab PF: Laboratory made PF resin, Com PF: Commercial PF resin, BEPF: Bark extractive-PF resins made from alkaline extractives,)

Figure 6.2 Thermal stability of post-cured BEPF (N) resins

(Lab PF: Laboratory made PF resin, Com PF: Commercial PF resin, BEPF (N): Bark extractives-PF resins made from acid-neutralized extractives)

6.3.5 Bonding strength

Based on the statistical t-test, at the level of p=0.05, the specimens bonded with both BEPF and BEPF (N) resins made with 30% replacement of phenol had dry shear strength values similar to those of specimens bonded with the commercial and lab PF resins (Figure 6.3). After the water-soaking-and-drying treatment and the boiling water treatment, no delamination occurred in any of the specimens. The BEPF and BEPF (N) resins made with 30% replacement of phenol gave similar lap shear strengths as the commercial and lab PF resins in both the WSAD and BWT/W tests. At 50% substitution level, the BEPF and BEPF (N) resins had similar dry and wet bond strengths as the lab PF resin. When the weight replacement of phenol increased to 70%, the dry, WSAD and BTW/W shear strength values of the specimens decreased significantly. The wet bond strength of the BEPF (N) resin was higher than that of BEPF resin at 70% replacement of phenol. Attempts were made to perform 100% replacement of phenol, but the bonded

specimens had very low dry shear strength; i.e. 1.87 MPa (±0.32) for the BEPF resin and 1.16 MPa (±0.28) for the BEPF (N) resin. After water-soaking and boiling, all the specimens bonded by the resins made at 100% phenol-replacement level were delaminated. These suggested that such bark extractives could replace up to 50% of phenol to make bark extractive-PF resins without significantly reducing the bond strength. Good wet bond strength of the bark extractive-PF resins at 30% and 50% replacements of phenol could be attributed to the catechol moiety of the tannins in the bark extractives [8, 85]. The neutralization of extractives before drying was found to have no significant effect on the bonding properties of the bark extractive-PF resins made at 30% and 50% replacements of phenol, but adjusting the pH to neutral level apparently improved the wet bond strength of the bark extractive-PF resins made at 70% replacement of phenol.

Figure 6.3 Shear strength of lap shear specimens bonded with different types of adhesives

(Lab PF: Laboratory made PF resin, Com PF: Commercial PF resin, BEPF: Bark extractive-PF resins made from alkaline extractives, BEPF (N): Bark extractive-PF resins made from acid-neutralized extractives)

6.3.6 FTIR of the bark extractives and bark extractive-PF resins

The FTIR spectra of the bark extractives and bark extractive-PF resins are shown in Figures 6.4 to 6.6. The assignments of the peaks are given in Table 6.6. According to the published literature [95, 96], the intense and broad band between 3400 cm-1 to 3500 cm-1 indicated the presence of -OH groups in large quantities in the bark extractives. The small -1 peak at 2900 cm from the CH2 groups was observed in bark extractives. The peak around 1608 cm-1 was assigned to the aromatic ring vibrations, indicating the existence of the phenolic compounds in the bark extractives. The peak at 1456 cm-1 was due to the C=C bonds in aromatic rings and O-H in-plane deformation as well as CH3 and CH2 asymmetrical stretchings. The peaks at 1036 cm-1 and 1085 cm-1 were assigned to the C-H and C-O stretchings in the cellulose and hemicellulose. The peaks at 817 cm-1 and 731 cm- 1 were the C-H out-of-plane and stretching associated with two adjacent hydrogen atoms as well as in-plane methylene rocking in straight chain lipids. The intensity of the peak at 1456 cm-1 in alkaline extractives increased significantly and the peak at 817 cm-1 appeared after drying, which indicated the occurrence of extractives hydrolysis and structure change during the drying process.

Figure 6.4 FTIR spectra of bark extractives

Figure 6.5 FTIR spectra of the bark extractive-PF resins made from alkaline extractives (BEPF)

100

Figure 6.6 FTIR spectra of the bark extractive-PF resins made from acid-neutralized extractives (BEPF (N))

For the bark extractive-PF resins, the intense and broad band between 3400 cm-1 to 3500 cm-1 was attributed to the -OH groups in the resins. The small peak at 2950cm-1 was -1 -1 -1 assigned to the CH2 groups. The peaks at 1606 cm , 1481 cm and 1456 cm were due to the vibration of aromatic rings and CH3 and CH2 asymmetrical stretchings. The peak at -1 1380 cm corresponded to the phenol O-H in-plane bending and CH3 symmetrical bending. The peak at 1282 cm-1 corresponded to the C-C-O asymmetric stretching, and the peaks at 1193 cm-1 and 1045 cm-1 were attributed to the C-H in-plane deformations. The -1 peak at 1019 cm representing C-O single bond stretching vibrations of -CH2OH group was observed in the bark extractive-PF resins. The peak at 826 cm-1 was due to C-H stretching. When replacement of phenol increased to 70%, there was a significant reduction in the intensity of the peaks at 1481 cm-1, 1456 cm-1 and 1380 cm-1 for both BEPF resin and BEPF (N) resin. The intensity of the peak at 1282 cm-1 from the C-C-O

101

asymmetric stretching also decreased, while the intensity of the peak at 1411 cm-1 (representing the C-C skeletal stretching) increased significantly. These results could be attributed to the self-condensation of the extractives, which may cause insufficient cross- linking for resin cure, higher curing activation energy and decreased bonding properties of the bark extractive-PF resins made at 70% replacement level.

Table 6.6 Assignment of peaks in FTIR spectra

Bark extractives Bark extractive-PF resins Wavenumber Assignment Wavenumber Assignment (cm-1) (cm-1)

3400~3500 hydroxyl group, H- 3400~3500 hydroxyl group, H-bonded bonded OH stretching OH stretching

2920~2950 CH2 stretching

2920 CH2 stretching 1606 aromatic rings vibration

1608 aromatic rings vibration 1481, 1456 C=C bonds in aromatic rings, CH3 and CH2 asymmetrical stretching

1411 C-C skeletal stretching C=C bonds in aromatic 1456 rings, OH in-plane deformation, CH3 and CH2 asymmetrical 1380 phenol O-H in-plane bend, stretching CH3 symmetrical bending

1085, 1036 C-H and C-O stretching C-C-O asymmetric in the cellulose and 1282,1193,1045 stretching and C-H in hemicellulose plane deformations

C-H out-of-plane, 817, 731 stretching associated with two adjacent C-O single bond hydrogen atoms, in- 1019 stretching vibrations of - plane methylene CH2OH group rocking absorption in straight chain lipids C-H out-of-plane, 992, 857, 826 stretching associated with two adjacent hydrogen atoms, in-plane methylene rocking in straight chain lipids

102

6.4 Conclusions

Bark of lodgepole pine (Pinus contorta Dougl.) infested by mountain pine beetle (Dendroctonus ponderosae Hopkins) was extracted with 1% NaOH. The bark alkaline extractives with and without acid neutralization were used to partially replace petroleum- based phenol for synthesizing bark extractives-PF resins. The bark extractive-PF resins were found to have higher molecular weights, higher viscosity and shorter gel times than the commercial PF resin and lab PF resin. Acid-neutralization of bark extractives before drying increased the average molecular weight of the extractives, and affected the properties and curing behavior of the resulting bark extractive-PF resins. The bark extractive-PF resins exhibited higher curing rate than the lab PF resin. The post-cured thermal stability of most bark extractive-PF resins was similar to that of the lab PF resin at high temperatures, but differed significantly from that of the commercial PF resin. The bark extractive-PF resins made from both alkaline and neutralized extractives at 30% replacement of phenol exhibited similar dry and wet bond strengths as those of commercial PF resin and lab PF resin. At 50% substitution level, the bark extractive-PF resins had similar dry and wet bond strengths as the lab PF resin. Extractives obtained from the 1% NaOH extraction of bark from mountain pine beetle-infested lodgepole pine were found to be a suitable material to partially replace phenol for the synthesis of bark extractive-PF resins as wood adhesives.

103

Chapter 7

Thermal Degradation Process of the Cured Bio-based PF Resins

7.1 Abstract

The thermal stability and thermal degradation kinetics of the bio-based PF resins were investigated by thermogravimetric analysis (TGA). The structural changes of the resins during the thermal degradation at different stages were studied by the Fourier Transform Infrared Spectroscopy (FTIR). Results showed that the post-curing thermal stability of the bio-based PF resins was similar to that of the lab PF resin but differed significantly from that of the commercial PF resin. Bio-based PF resin made from bark alkaline extractives exhibited different thermal stability and different thermal degradation kinetics from the bio-based PF resin made from phenol liquefied bark. The bio-based PF resin made from bark alkaline extractives had better thermal stability than the bio-based PF resin made from phenol liquefied bark. The structural changes and difference of the bio-based PF resins, lab PF and commercial PF resins during the thermal degradation were also clearly identified by FTIR.

7.2 Introduction

Phenol formaldehyde (PF) resins have been widely used as adhesives, coatings, thermal insulation materials, and molding compounds due to their good mechanical properties and heat resistance [1]. Generally, phenol and formaldehyde go through addition and condensation reactions under either alkaline or acid conditions to form an either cross- linked thermoset or linear thermoplastic polymer. The structure and properties of the polymer depend highly on the reaction conditions.

Thermal stability of a thermoset polymer is highly dependent on its structure and cross- linking density [97, 98]. It is an important property of polymer durability. Various studies

104

have investigated the thermal degradation mechanisms for PF resins [99-103]. It has been suggested that the thermal degradation of PF resins follows an auto-oxidization process that mainly includes three stages. In the first thermal degradation stage, additional cross- linking was formed between the remaining un-reacted functional groups in the cured PF resins. In the second thermal degradation stage, the level of cross-linking was reduced with release of degraded by-products, such as methane, hydrogen, carbon monoxide, small oligomer, and water. In the third thermal degradation stage, the cured resin network collapsed followed by the carbonization and graphitization processes [99, 100].

Thermogravimetric analysis (TGA) that measures a sample’s mass loss as a function of temperature and/or time is a common technique for characterizing polymer thermal degradation behavior [101-103]. Kinetic studies during the thermal degradation process have been used to acquire fundamental understanding of the structural changes in phenolic resins [99]. Depending on whether a single or multiple-heating rate is used, technique used to obtain the kinetic parameters is termed either as a differential or an integral method [78, 104]. It is believed that multiple heating rates method gives more reliable results than the single heating rate method with smaller experimental errors [10]. Meanwhile, the “model- free” isoconversional method is a widely adopted technique for deriving relevant kinetic parameters using thermogravimetric (TG) and differential thermogravimetric (DTG) curves measured at different heating rates that makes no presumption about the reaction function and reaction order [76-78, 104-107].

Previous studies [108-110] have also shown that the thermal stability and thermal degradation kinetics of the PF resins were significantly affected by the resin synthesis conditions. For example, it was found that the activation energy of the resol resins with various formaldehyde to phenol molar ratios decreased sharply at first and then remained almost constant during the thermal decomposition process. However, the activation energy of the novolac resins with certain formaldehyde to phenol molar ratios were almost unchanged in the whole degradation process, while that of novolac resins with some other formaldehyde to phenol molar ratios gradually decreased [110].

The inclusion of other materials, such as lignin and cellulose fibers, in the PF resin

105

synthesis has been shown to modify the thermal stability and degradation kinetics of the resulting resins [111-113]. Lignin-phenol-formaldehyde (LPF) resol resins were found to have higher thermal stability under both nitrogen and air conditions than PF resins with different degradation characteristics [111]. Meanwhile, LPF novolac resins were found to be less stable at lower temperatures than PF resins, but they were comparable as PF resins at higher temperatures [112]. The addition of cellulose fibers to PF resol resins reduced thermal resistance of the resulting resol resins [113].

In chapters 4, 5, 6, phenolic compounds from bark were used to partially substitute petroleum-derived phenol in the PF resol resin synthesis [91, 114]. In these studies, bark phenolic compounds were extracted by either acid-catalyzed phenol liquefaction or alkaline extraction and used for formulating bio-based PF resins. The resulting bio-based- PF resins, with both liquefied bark and bark alkaline extractives, were found to have comparable bonding strengths to a commercial PF resin. But the activation energy and pre-exponential factors of the bio-based PF resins in curing differed from those of commercial PF resins suggesting potential structural variations in these cured adhesives. Even though stability of the cured adhesives is an important consideration in applications, no previous study has investigated the impact of bark components on the thermal stability and degradation kinetics of the resulting resins. In this chapter, the thermal stability and thermal degradation behavior and kinetics of the cured bio-based PF resins are investigated. The liquefied bark-PF resin and bark extractive-PF resin made from mountain pine beetle infested lodgepole pine with similar bonding strength and curing behavior will be used in this study; a lab made PF and a commercial OSB face PF resin will be used as comparisons.

A version of this chapter has been accepted for publication in Thermochimica Acta (2012).

7.3 Results and discussion

7.3.1Thermal stability of the bio-based PF resins

The thermal degradation curves of the different resins are shown in figure 7.1. When the temperature increased from room temperature to 200°C, the weight loss of the cured resins

106

was mainly caused by the evaporation of moisture, dehydration, as well as the degradation of the small molecular weight compounds. The evolution of water could be attributed to the condensation reaction between the residual methylol groups and phenolic OH groups. Water elimination might lead to the formation of new crosslinks [99, 100].

Figure 7.1 Thermal degradation curves of the resins

From 200°C to 400°C, the resins underwent further degradation. The release of excess phenol, aldehyde, short oligomers and water were main contributors to the weight loss. When the temperature increased from 400°C to 600°C, major polymer decomposition took place [99,100]. The weight loss was the result of the formation of products, such as CO,

CO2, benzaldehyde, phenol, and its associated polymers, with random chain scission and the initial formation of char [100-103]. After the temperature increased above 600°C, the degradation continued and a carbon-like structure (char) was gradually formed, generating carbon monoxide as byproduct. It was also considered to be the carbonization process.

107

Figure 7.2 Derivative thermal degradation (DTG) curves of the resins

It is clearly shown that the thermal degradation process of the commercial PF resin differed significantly from those of the lab PF and bio-based PF resins, especially when the temperature was higher than 200°C. Two distinct derivative thermal degradation (DTG) peaks (shown in figure 7.2) were observed within the temperature range of 200°C to 400°C for the commercial PF resin. The reason for that could be the fact that usually a significant amount of urea and other additives are present in the commercial PF for controlling the viscosity and solids content. These additives would degrade at a relatively lower temperature. Elemental analysis revealed that a nitrogen content of 13% was in the cured commercial PF resin and it decreased to 4% when the temperature increased from the room temperature to 400°C.

The weight losses of all the tested resins during the whole thermal degradation process are shown in Table 7.1. The total weight loss of the commercial PF resin from room temperature to 800°C was higher than those of the lab PF resin and bio-based PF resins (include both liquefied bark-PF resin and bark extractive-PF resin). The bark extractive PF resin had the highest residual weight above 800°C, followed by the lab PF resin, liquefied bark-PF resin and commercial PF resin.

108

The inclusion of various additives, such as urea and starch, in the commercial PF mainly contributed to its higher weight loss (35.3%) than the lab PF (20.7%), liquefied bark-PF resin (22.7%) and bark extractive-PF resin (18.7%) from room temperature to 400°C. When the temperature increased from 400°C to 600°C, the bio-based PF resins had slightly lower weight losses than the lab PF and commercial PF resin. From 600°C to 800°C, the weight loss of the commercial PF resin was the lowest among all the test resins, valued at 5.0%. The lab PF and bark-extractive PF resin had similar weight loss during that temperature range, 8.1% and 8.7% respectively; while the liquefied bark-PF resin had the highest weight loss, i.e.12.1%. The reason could be that the liquefied bark components with phenolic side chains or branch structures in the liquefied bark-PF resin were thermally stable at lower temperature started to degrade.

The difference in the weight loss between the bark extractive-PF resin and liquefied bark- PF resin could be attributed to the difference in chemical composition between the liquefied bark and bark extractives as well as the difference in the resin structures. The structural study of the liquefied bark-PF resin and bark extractive-PF resins will be discussed in Chapter 8.

Table 7.1 Weight losses of different resins during the thermal degradation process

Weight loss at different temperature range (%) Total weight loss (%)

Resin Type RT~200°C 200°C~400°C 400°C~600°C 600°C~800°C RT~800°C

Commercial PF 13.2 22.1 14.9 5.0 55.2

Liquefied Bark-PF 9.2 13.5 12.8 12.1 47.6

Bark extractive-PF 7.5 11.2 14.0 8.7 41.4

Lab-PF 10.3 10.4 14.8 8.1 43.6

7.3.2 Thermal degradation kinetics

The dependence of the activation energy on the extent of degradation provides information on how the polymer structure affects the degradation kinetics [78, 104]. The linear

109

relationship between log β and -1/T of the tested resins at different stages of thermal degradation based on the Flynn-Wall-Ozawa method are shown in Figures 7.3 to 7.6. 2 2 Similar linear relationship between ln (β/Tp ) and -1/Tp (Kissinger method), ln (β/Tα,i ) and -1/T (Kissinger-Akahira-Sunose method) were also obtained (Figures not shown). It should be noted that the data for the bio-based PF resins fitted well with the kinetic model during all stages of the thermal degradation process. However, in the early stage (weight loss <20%) the data for the commercial PF and lab PF resins had poorer fit than the later stages. It could be attributed to the difference in mechanisms at different conversion levels [110, 111].

Figure 7.3 Isoconversional plot of Flynn-Wall-Ozawa method for the bark extractive-PF resin

110

Figure 7.4 Isoconversional plot of Flynn-Wall-Ozawa method for the liquefied bark-PF resin

Figure 7.5 Isoconversional plot of Flynn-Wall-Ozawa method for the commercial PF resin

111

Figure 7.6 Isoconversional plot of Flynn-Wall-Ozawa method for the lab PF resin

The activation energy (E) of the tested resins at different stages of thermal degradation calculated by the Flynn-Wall-Ozawa method and Kissinger-Akahira-Sunose method from the slope is shown in Figure 7.7 and Figure 7.8. The activation energy calculated by the two methods had the same trend for all the tested resins.

In the early stage (conversion < 20%), the commercial PF resin had the lowest activation energy; it could be explained by the fact that the commercial PF resin contained a significant amount of urea and other additives, which would easily decompose at lower temperatures. For all the test resins, the activation energy generally increased before the conversion reached 60%. The increased activation energy indicated the gradual degradation of the main network of the resins.

The highest activation energy of the bio-based PF resins, the lab PF resin and commercial PF resin was observed when the conversion rate was higher than 70%. It could be due to the fact that in the late thermal degradation stage, their structures became more ordered and required more energy to break down. The decomposition mechanism changed when the conversion was higher than 70% [110-112]. Degradation and formation of new cross-

112

links occurred at the same time [99, 110, 112].

Figure 7.7 Activation energy of the tested resins at different conversion levels calculated by the Flynn-Wall-Ozawa method

Figure 7.8 Activation energy of the tested resins at different conversion levels calculated by the Kissinger-Akahira-Sunose method

The bark extractive-PF resin exhibited the highest activation energy compared with the liquefied bark-PF resin, lab PF resin and commercial PF resins in most of the stages of the

113

thermal degradation process. It indicated that the cured bark extractive-PF resin was more thermally stable than the cured liquefied bark-PF. This is also consistent with the results of different weight loss between the bark extractive-PF resin and liquefied bark-PF resin shown before. The bark components in the bio-based PF resins did affect their post-cured crosslink density. The difference in the chemical composition and structures between the bark extractive-PF and liquefied bark-PF resin could contribute to their different thermal stability and thermal degradation behavior.

The activation energy of all the tested resins calculated by the Kissinger method by 2 plotting ln (β/Tp ) and 1/T obtained from their DTG curves was shown in Table 7.2. It showed that the activation energy of the bark extractive-PF resin was higher than that of the rest of the resins in most of the cases, the same as what was obtained by the Flynn- Wall-Ozawa method and Kissinger-Akahira-Sunose method. For the commercial PF resin, the activation energy of the second peak was lower than the other peaks, which could be attributed to urea. The results and trends calculated by the Kissinger method was also consistent with the results calculated by the Flynn-Wall-Ozawa method and Kissinger- Akahira-Sunose.

Table 7.2 Activation energy of the tested resins calculated by the Kissinger method

Activation energy (kJ/mol) Peak 1 Peak 2 Peak 3 Peak 4 Peak 5 Bark extractive-PF 110.96 (0.98) 254.75 (0.94) 232.96 (0.98) NA NA Liquefied bark-PF 102.74 (0.97) 206.02 (0.93) 222.12 (0.97) NA NA Commercial PF 106.71 (0.99) 103.21 (0.99) 138,84 (0.99) 118.05 (0.99) 189.24 (0.97) Lab PF 97.29 (0.97) 263.19 (0.98) 219.98 (0.97) NA NA *Values in the bracket are the correlation coefficient. * NA: not available

7.3.3 Structural changes of the bio-based PF resins during thermal degradation

The structural changes of the four types of the cured resins at different thermal degradation stages were characterized by FTIR. The FTIR spectra of the tested resins are shown from Figure 7.9 to 7.12.

114

Bark extractive-PF

1353 phenol OH 200°C

2848, 2914, CH2

400°C

1650,1600, C=O, benzene rings 1249, 1199, C-O Transmittance [%] [%] Transmittance 600°C

800°C 1740 C=O 3400 OH 879 benzene ring

1467 benzene ring

3500 3000 2500 2000 1500 1000 500 -1 Wavenumber cm

Figure 7.9 FTIR of the bark extractive-PF resin at different degradation temperatures

115

Liquefied bark PF

1230 C-O 200°C

400°C

2848, 2914, CH2 1650, 1600, C=O, benzene ring Transmittance [%] [%] Transmittance 600°C

1141, C-O

800°C 1740 C=O 3400 OH 879 benzene ring

1467 benzene ring

3500 3000 2500 2000 1500 1000 500 -1 Wavenumber cm

Figure 7.10 FTIR of the liquefied bark-PF resin at different degradation temperatures

116

Commercial PF

1353 OH on benzene ring 200°C

2227, 2167, CN, NCO 1213, 1170 C-O 400°C 1650 C=O

2848, 2914, CH2 3350, OH 1585 benzene Transmittance [%] [%] Transmittance ring 1213C-O 600°C

879 benzene ring

800°C 1740 C=O

1467 benzene ring

3500 3000 2500 2000 1500 1000 500 Wavenumber cm-1

Figure 7.11 FTIR of the commercial PF resin at different degradation temperatures

117

Lab PF

200°C

2848, 2914, CH2 400°C 1650,1600, C=O, benzene rings

1253,1203, 1136, C-O Transmittance [%] [%] Transmittance 600°C

800°C

3429, OH 1740 C=O 879 benzene ring

1467 benzene ring

3500 3000 2500 2000 1500 1000 500 -1 Wavenumber cm

Figure 7.12 FTIR of the lab PF resin at different degradation temperatures

According to previous studies [95, 100, 111], the peaks assignments were given in Table 7.3. The intensities and bands of the functional groups of the tested resins changed with the increased thermal degradation temperature.

118

Table 7.3 Peak assignment of the FTIR spectra of all tested resins

Wavenumber (cm-1) Peak assignment

3350-3500 phenolic OH strech

2848-2914 aliphatic CH2 asymmetric stretch 2164, 2227 CN, NCO groups 1740,1650,1600 carbonyl groups, benzene ring 1587 benzene ring 1467 benzene ring 1435 aliphatic CH scissor bending 2 1353 phenolic OH in-plane deformation 1230, 1191,1147 alkyl-phenol C-O stretch 1050 aromatic CH in-plane deformation 1006 aromatic linked CH rock and/or aromatic CH 3 879 polysubstituted aromatic ring

783 CH2 out-of-plane ring deformation 621 CH2 out-of-plane ring deformation

The intensity of the OH peak at around 3350cm-1 to 3500cm-1 for all the tested resins decreased when temperature was increased from room temperature to 400°C. This is due to the condensation process involving either residual methylol groups or the phenolic hydroxyl groups and forming of the new cross-linking. Possible reactions were shown in Figure 7.13 [99, 100]. The OH peak intensity at 1353 cm-1 also decreased due to the same mechanism. When the temperature was further increased to above 600°C, the intensity of the OH peak decreased significantly and the peak at 1353 cm-1 disappeared. The main structure of the resins changed to a poly-aromatic structure [99]. The intensity of the -1 -1 peaks at 2914cm and 2848 cm due to aliphatic CH2 asymmetric stretch and aliphatic

CH2 symmetric stretch, respectively, decreased with increasing thermal degradation temperature.

OH OH OH OH CH2OH H CH2 -H2O +

119

OH + HO O

HO HO

OH + CH2 CH + H2O

OH OH

Figure 7.13 Possible condensation reactions during the thermal degradation process

The peak at 1740cm-1 due to the carboxylic groups was observed in the bark extractive-PF resin, liquefied bark-PF resin and lab PF resin when the thermal degradation temperature was 400°C. It also appeared in all the tested resins when the thermal degradation increased up to 800°C. The intensity of the peak was weak, but its appearance was considered as an evidence for oxidation during the resin thermal degradation under the inert atmosphere [99, 100, 111,]. The possible reactions were shown in Figure 7.14.

OH OH OH OH CH H C 2 2 HOH2C +OH +OH

-H2O

OH OH OH OHC HOOC +OH

+ CO2

120

O O OH OH OH CH 2 CH C

OH OH OH

Figure 7.14 Possible oxidation reactions during the resin thermal degradation process

The carbonyl groups usually appearing at 1650 cm-1 as a shoulder peak next to 1600 cm-1 could be observed in all the resins when the thermal degradation temperature was 200°C. And the intensity of the peak decreased when the temperature increased. The intensity of the peak of benzene rings at 1467 cm-1 increased for all the resins during the thermal degradation process. The intensity of the peak assigned to the methylene bridges at 1435 cm-1 decreased when the temperature increased reflecting the occurrence of the resin network decomposition.

The peak of alkyl-phenol C-O stretching at 1213 cm-1 and 1170 cm-1 decreased with increasing thermal degradation temperature. It disappeared when the temperature was above 600°C. The alkyl-phenol C-O stretching peak of the liquefied bark-PF resin still present even when the temperature was 800°C.

For the bark extractive-PF resin, the peak at 1251 cm-1 decreased at 200°C but increased at 400°C, indicating formation of ethers. While for the lab PF and commercial PF resin, peak at 1251 cm-1 was absent in the original spectrum and appeared when the thermal degradation temperature was 400°C. It also implied that the oxidation occurred during the resin thermal degradation process.

The intensity of the peak at 879 cm-1 assigned to the poly-substituted aromatic ring increased with the increasing thermal degradation temperature. The peak of CH2 out-of- plane ring deformation at 783 cm-1 was observed in all the tested resins when the thermal degradation temperature was 800°C.

When the thermal degradation temperature was lower than 400°C, the polymer network of

121

the phenolic resins remained essentially intact, while when the thermal degradation temperature was above 600°C, dramatic changes were seen in the FTIR spectra of all the resins due to the collapse of the network to form polyaromatic domains. It is consistent with what were reported for PF resins [99, 111].

It is also worthy of noting that the difference in the FTIR spectrum of the commercial PF resins when the thermal degradation temperature was at 200°C and 400°C. According to the previous study [115], the peaks at 2227 cm-1 and 2167 cm-1, attributed to CN triple bond in the cyanamide and the –NCO groups were observed when the thermal degradation temperature was 200°C. The intensity of the peak at 2167cm-1 decreased significantly while the intensity of the peak at 2227 cm-1 increased when the temperature increased to 400°C. Both peaks disappeared when the thermal degradation temperature increased to 600°C. The reason for the change could be due to the thermal degradation of urea in the commercial PF resin. The possible reactions that were reported before are shown in Figure 7.15.

Figure 7.15 Possible reactions for urea decomposition

7.4 Conclusions

The thermal stability and thermal degradation kinetics of the phenol formaldehyde resins derived from mountain pine beetle (Dendroctonus ponderosae Hopkins) infested lodgepole pine (Pinus contorta Dougl.) barks were investigated by thermogravimetric analysis (TGA). The structural changes of the resins during the thermal degradation at different stages were studied by Fourier transform infrared spectroscopy (FTIR). Results

122

showed that the post-curing thermal stability of the bio-based PF resins was similar to that of the lab PF resin but differed significantly from that of the commercial PF resin. Bio- based PF resin exhibited different thermal stability and thermal degradation kinetics from the phenol liquefied bio-based PF resin. The bark extractives-PF resin had better thermal stability than the liquefied bark-PF resin. The structural changes and the differences in structures among the bio-based PF resins, lab PF and commercial PF resin during the thermal degradation process were clearly identified by the FTIR analysis. The commercial PF resin had different FTIR spectra characteristics from the bio-based PF resins and lab PF resin when the thermal degradation temperature was at 200°C and 400°C. The thermal degradation mechanism involving oxidation was observed for the bio-based PF resins.

123

Chapter 8

Liquid-State 13C NMR Study of Bio-based Bark PF Resins from Mountain Pine Beetle Infested Lodgepole pine

8.1 Abstract

In this study, two types of bio-based PF resins, namely liquefied bark-PF resin and bark extractive-PF resin, were synthesized from acid-catalyzed phenol-liquefied bark and bark alkaline extractives, respectively. The bio-based resins were characterized for their chemical compositions and molecular structures using the liquid-state 13C nuclear magnetic resonance (NMR) technique. The results indicated that the introduction of bark components (either as liquefied bark or as bark extractives) to the phenolic resin synthesis affected resin structures and curing performance. Methylene ether bridges were found in the bio-based bark PF resins. Bark components made the formation of para-ortho methylene linkage more favorable in bio-based bark PF resins than in lab PF resins. Molecular structures of the liquefied bark-PF resin differed significantly from those of the bark extractive-PF resins. The liquefied bark-PF resin showed a higher ratio of para- para/ortho-para methylene link (-CH2-), a higher unsubstituted/substituted hydrogen (-H/-

CH2OH) ratio and a higher methylol/methylene (-CH2OH/-CH2-) ratio than the bark extractive-PF resin. The tannin components of the bark extractives accelerated the curing rate of the resulting bark extractive-PF resin. The bark extractives made the ortho position of phenol react more favorably with formaldehyde than the para position. The liquefied bark with phenolated structures had more reactive sites towards formaldehyde than the bark extractives and accelerated the curing rate of the resulting liquefied bark-PF resin.

8.2 Introduction

Phenol formaldehyde (PF) resins, obtained through reacting phenol with formaldehyde, have been widely used as adhesives, coatings, thermal insulation materials, and molding

124

compounds due to their good mechanical properties and heat resistance [1]. With the increasing concern for fossil fuel depletion and carbon footprint, there is a strong global interest to explore renewable resources as alternative feedstock for making PF resins. Bark and bark extractives including phenolic compounds have been successfully used to partially substitute petroleum-based phenol in the resin synthesis [86, 87, 91, 114, 116, 117]. In studies shown in chapter 4, 5, 6, both phenol liquefied bark-PF resol resin [91] and alkaline bark extractive-PF resol resins [114] were synthesized. It was found that the wet bonding strength of the liquefied bark-PF resin was higher than that of a commercial PF resin while the dry bonding strength was similar. Meanwhile, the bark extractive-PF resins with 50wt.% phenol replacement by bark extractives exhibited similar dry and wet bonding strength as those of the lab-made control PF (lab PF) resin without bark components. The curing rate of the liquefied bark-PF resin was less than that of the commercial PF resin but higher than that of the lab-made control PF resin. The curing behavior and curing kinetics of the both types of bio-based PF resins at various phenol substitution levels differed significantly from those of the lab PF resin.

Liquid 13C nuclear magnetic resonance (NMR) spectroscopy has been successfully applied to characterize the composition of phenol liquefied cellulose, phenol liquefied cellulose and lignin model compounds, bark tannin extractives, and PF resins [32-35, 118-127]. The NMR studies on the phenol liquefaction of cellulose, cellobiose and guaiacylglycerol- β-guaiacyl ether (GG) showed that phenol reacted with these compounds during the liquefaction [32-35]. It was found that compounds formed during cellulose liquefaction have structural units highly resembling the conventional PF resins such as methylolphenol [32]. The liquefaction of GG forms intermediates that are highly reactive and can further condense with each other or with phenol. The main liquefied products of GG such as diphenylmethanes and guaiacol were highly phenolated and could be used for phenolic resin synthesis [33, 34]. The reactivity of those liquefied products would differ from phenol in the resin formulation and would affect the reactivity and adhesion of the resulting adhesives. The NMR studies on the commercial bark tannin extractives such as black wattle tannin extracts, mimosa tannin extracts, quebracho tannin extracts, pine tannin extracts and bark sodium sulfite extractives from radiata pine and maritime pine have identified and differentiated the structure and degree of polymerization of

125

polyflavonoid tannin [118-122]. The structure and reactivity of tannins varied significantly among tree species and extraction methods. Numerous NMR studies on PF resins in the literature provided qualitative or quantitative information on the resin’s reaction mechanism, structure, and composition [123-127]. It was generally believe that the methylene link was the dominant structure in these resins and the no methylene-ether linkage was reported for the uncured resol resins under the alkaline conditions [128, 129]. Methylene ether bridges between phenol rings were not stable and could be eliminated under the alkaline conditions [128, 129]. In addition, 13C NMR also provided useful information for investigating the cure-accelerated PF resin and PUF resins [129-134].

No previous studies used 13C NMR to elucidate the compositional characteristics of phenol liquefied bark and bark alkaline extractives especially those from the mountain pine beetle infested lodgepole pine as well as the resulting liquefied bark-PF and bark extractive-PF resins. Thus, in this study in order to better understand the reaction mechanism and curing behavior of these bio-based PF resins, the liquid-state 13C NMR and DSC have been used to measure the structural features and curing characteristics of the bio-based PF resins.

A version of this chapter has been accepted for publication in ACS Sustainable Chemistry & Engineering (2012).

8.3 Results and discussion

8.3.1 Liquid-state 13C NMR spectrum of the bark alkaline extractives

The liquid-state 13C NMR spectrum of bark alkaline extractives is shown in Figure 8.1. According to previous studies, the assignment of the chemical shifts is listed in Table 8.1. The chemical shifts observed in the region around 150ppm, 145ppm and 133ppm were primarily associated with polyphenolic procyanidin and prodelphinidin tannins and lignins [122].

126

Bark alkaline extractives

(OH) CH2

5' CH OH 6' CH 8 1 4' 1' B 3' 4 HO 7 9 O 5 O HO 2 OH 1 2' HO 2 A 3 3 OH OH 6 OCH OH 3 5 10 4 OH (OH) O O

Figure 8.1 Liquid-state 13C NMR spectrum of the bark alkaline extractives

The chemical shift at 181 ppm was attributed to the quinone structures due to the oxidation of phenolic hydroxyl groups. The chemical shift at 174 ppm belonged to either catechin or epicatechin gallate, suggesting the C=O bond of a gallic residue linked to catechin or epicatechin gallate [118].

The tannin flavonoid structure existing in the bark alkaline extractives was shown as the chemical shifts at 150 ppm representing C5, C7 attached to the phenolic –OH group on the flavonoid A-ring, while shifts at 140-145 ppm representing the C3’ and C4’ on the B-ring, shifts at 131 ppm, 116-118 ppm, 110-115 ppm and 105 ppm for the C1’, C5’ and C2’, C4-C8 and C4-C6 interflavonoid bond respectively. The chemical shift at 68 ppm was associated with C3 located in the chain interior and upper chain-ending positions. The chemical shift at 29 ppm was assigned to C4 [134]. The intensity of this chemical shift was affected by the oligomeric carbohydrates in the bark extractives. The chemical shift at 36.5 ppm belonged to the C4 involved in the interflavonoid bond, while the chemical shift

127

at 19.5 ppm was attributed to the free C4. These two peaks indicate the degree of polymerization of tannins in the bark extractives. It can be seen that the intensity of the chemical shift at 110-115 ppm was higher than that at 105 ppm, suggesting the tannins in the bark alkaline extractives from the mountain pine beetle infested lodgepole pine was mainly procyanidian. The intermediate intensity of the chemical shift at 116 ppm indicated the existence of the pyrogallol-type B-ring and catechol B-ring in the tannins from the alkaline pine bark extractives. The chemical shift of C1’ at 130-132 ppm for the catechol B-rings and 132-135 ppm for the pyrogallol B-rings also supported the coexistence of the catechol B rings and pyrogallol B-rings in the tannins from the bark alkaline extractives.

Table 8.1 Assignment of chemical shifts for bark extractives [118,135,136]

Chemical shifts (ppm) Assignment 181 C=O in quinone structures 174 Catechin or epicatechin gallate 166 CH-γ in p-coumarate ester 150 C5, C7 on tannin phenolic A ring 140-145 C3', C4' on tannin phenolic B ring 133-135 C-1 in guaiacyl and syringyl unit 131 C1' on tannin phenolic B ring 122 C6 in the guaiacyl units 116-118 C5' on tannin phenolic B ring 110-115 C4-C8 interflavonoind bond 105 C4-C6 interflavonoid bond 72.4 CH-α in the β-O-4’ (erythro) guaiacyl 71.6 CH-γ in β-β’ units and CH-α in the β-O-4 (threo) guaiacyl 69.6 CH-4 in xylose non-reducing end unit 68 C3 on the tannin phenolic A ring 62.9 Cγ in the β-O-4 structures 55.6 methoxyl groups 54.6 CH-β in β-β’ units 36.5 C4 interflavonoid bond 19.5 Free C4 on tannin phenolic A ring

In addition, lignin fragments in the bark extractives were also present. The chemical shift at 55.6 ppm was assigned to the methoxyl groups in the lignin. The chemical shift at 62.9 ppm corresponded to Cγ in the β-O-4 structures. The chemical shift at 122 ppm was related to the C6 in the guaiacyl units. The chemical shifts at 133 to 135 ppm were

128

attributed to C1 in guaiacyl and syringyl units. The chemical shifts at 72.4 and 71.6 ppm were assigned to CH-α in the β-O-4’ (erythro) guaiacyl, CH-γ in β-β’ units and CH- α in the β-O-4 (threo) guaiacyl, respectively. The chemical shift at 54.6 ppm was CH-β in β-β’ units. Hemicellulose structures could also be seen in the bark extractives. The chemical shift at 69.6 ppm was the result of CH-4 in xylose non-reducing end unit [135, 136]. The assignment of chemical shifts for the bark alkaline extractives is summarized in Table 8.1

Therefore, the liquid-state 13C NMR spectra have confirmed that the bark alkaline extractives from mountain pine beetle infested lodgepole pine contained tannin, degraded lignin and degraded hemicellulose in their composition. The tannin and degraded lignin have molecular structures suitable for the phenolic resin synthesis.

8.3.2 Liquid-state 13C NMR spectrum of liquefied bark

The liquid-state 13C NMR spectrum of the phenol-liquefied bark is shown in Figure 8.2. The chemical shifts were assigned based on previous publications [32-35]. Multiple peaks within the ranges of 157-159, 129-131, 119-122 and 115-117 ppm were observed in the liquid-state 13C NMR spectrum of the liquefied bark. It indicated that there were both free phenol and phenolated bark components with combined phenol structures in the liquefied bark, since free phenol is expected to be represented by chemical shifts of C-OH at 158.5 ppm, ortho-carbon at 115.9 ppm, meta-carbon at 130.1 ppm, and para-carbon at 121.3 ppm. The chemical shift at 174.83 ppm arose from the C=O group of the glucuronic acid indicating that it is esterified to lignin. The chemical shifts representing the C-β in β-O-4’ and C-γ in β-O-4’ at 86.0 and 60.1 ppm were not observed in the liquefied bark, implying that the β-aryl ether structure of lignin was split during the liquefaction. The chemical shift at 72.5 ppm was from C-α in β-O-4’ (erythro) guaiacyl. The chemical shift at 55.5 ppm was the methoxyl groups (-OCH3). Guaiacylglyerol-α-phenyl-β-guaiacy structure in the liquefied bark was observed with C-γ at 63.1 ppm, guaiacyl unit from 110 ppm to 150 ppm with C1 in etherified guaiacyl unit at 134.78 ppm, C4 in nonetherified guaiacyl unit at 145.25 ppm, C4 in etherified guaiacyl unit in β-5’ at 146.51 ppm. The phenylcoumaranes, benzocyclobutanes, triphenylethanes, diphenylmethanes and guaiacol

129

structures and fragments were also observed in the liquefied bark, which were formed mainly by lignin during the liquefaction.

Figure 8.2 Liquid-state 13C NMR spectrum of the liquefied bark

Phenylglucopyranoside, 2-glucopyranosylphenol, 4-glucopyranosylphenol and 4-(2- hydroxylphenyl)-4-(4-hydroxylphenyl)-1,2,3-butanetriol derived from the cellulose liquefaction were observed in the liquefied bark. But the intensity of the chemical shifts from 73 ppm to 80 ppm representing the -CH-OH and -CH2OH in the polysaccharide were very weak, suggesting that these compounds have been converted into tri(4- hydroxylphenyl)methane, 1,1,2-tri(4-hydroxylphenyl)ethane (strong chemical shifts at 150~160 ppm and 120~130 ppm representing the phenolic structures in these products were observed) and other phenolated products during the liquefaction. The assignment of chemical shifts for the liquefied bark is summarized in Table 8.2.

130

Overall, the process of bark liquefaction in phenol appeared to be similar to wood liquefaction in phenol. The bark components underwent degradation and chemical bond formation/cleavage during the liquefaction, followed by reactions of the degraded products with phenol to form phenolated barks. Compared with bark alkaline extractives, more phenolic structures were found in the liquefied bark.

Table 8.2 Assignment of chemical shifts for liquefied bark [32-35]

Chemical shifts (ppm) Assignment 174.83 glucouronic acid 157-159 C-OH on the phenolic ring 146.51 C4 in etherified guaiacyl unit in β-5’ 145.25 C4 in nonetherified guaiacyl unit 134-135 C1 in etherified guaiacyl and syringyl unit 129-131 meta-carbons on the phenolic ring 119-122 para-carbon on the phenolic ring 115-117 ortho-carbon on the phenolic ring 110-130 guaiacyl unit 86 C-β in β-O-4’

73-80 -CH-OH and -CH2OH in the polysaccharide 72.5 C-α in β-O-4’ (erythro) guaiacyl 63.1 C-γ in guaiacylglyerol-α-phenyl-β-guaiacyl structure 60.1 C-γ in β-O-4’ 55.5 methoxyl groups

8.3.3 Liquid-state 13C NMR spectrum of lab-made phenol formaldehyde resin

The liquid-state 13C NMR spectrum of the lab-made PF resin is shown in Figure 8.3. The chemical shifts were assigned to the corresponding functional groups based on previous literatures on model compounds [123-127].

The chemical shifts at 153.5-163.9 ppm were assigned to phenoxy carbons. The chemical shifts of para alkylated phenolic groups were between 158.0 and 159.5 ppm, which showed higher intensity than those of chemical shifts of ortho alkylated phenolic groups between 153.5 and 157.2 ppm. The chemical shifts of phenoxy carbons varied with the pH of the resin due to the inducting polarization of the phenoxy groups.

131

The chemical shift at 48.7 ppm was due to the methanol (CH3OH) in the formaldehyde solution as a stabilizer. The methanol could also be formed during the resin synthesis from the Cannizzaro reaction of formaldehyde. The chemical shift at 55.2 ppm was assigned to the methyl carbon in hemiformal (CH3OCH2OH), with the methylene carbon chemical shift at 89.3 ppm (CH3OCH2OH).

Figure 8.3 Liquid-state 13C NMR spectrum of the lab PF resin

The chemical shift at 81.7 ppm was related to the unreacted formaldehyde in the PF resins. The small chemical shifts around 86-87 ppm belonged to formaldehyde oligomers. The addition reaction products of ortho-hemiformal and para-hemiformal have the chemical shift at 89.3 ppm for the carbon connected to the hydroxyl (Ph-CH2OCH2OH) and at 63.1-

63.7 ppm for the carbon connected to the phenolic ring (Ph-CH2OCH2OH). The chemical shift at 92.1-93.7 ppm for the ortho- and para-Ph-CH2OCH2OCH2OH was also observed. No peaks were found between 69 and 73 ppm, indicating that the methylene ether bridges were not formed between the phenolic rings during the synthesis of PF resins under the

132

experimental conditions. This was consistent with previously published results [127-129]. Ortho-ortho link of methylene groups between phenols at 29.2-29.6 ppm were not observed in the spectrum (Ph-CH2-Ph), while the ortho-para and para-para link of methylene groups at 34.6-35.0 ppm and 39.1-41.0 ppm were observed. It is understandable since the ortho-ortho methylene links are not favored in the presence of sodium hydroxide [127, 128]. The peaks at 114-116 ppm of unsubstituted ortho aromatic carbons of phenol and the peaks at 120-124 ppm of unsubstituted para aromatic carbons of phenol were observed. The peaks at 60.4-64.7 ppm represented the methylol groups (Ph-CH2OH) in the PF resins, with the ortho link at 60.4-62.8 ppm and para link at 63.1-64.7 ppm.

8.3.4 Liquid-state 13C NMR spectrum of bark extractive-phenol formaldehyde resins

The difference in the liquid-state 13C NMR spectra between the bark-extractive PF resin and the lab PF resin could be observed clearly. The incorporation of bark extractives in the formulation of the phenolic resins introduced new peaks (shown in Figures 8.4 to 8.6).

133

Figure 8.4 Liquid-state 13C NMR spectrum of bark extractive-PF resin with 30 wt% phenol substitution rate by bark extractives

The appearance of the new chemical shifts at 29.3-30.9 ppm, 28.8~29.1 ppm, 23.8~24.9 ppm in the bark extractive-PF resins could be attributed to the methylene groups between phenol and the tannin-B ring (29.3-30.9 ppm), and between phenol and the tannin-A ring (phloroglucinol ring, 28.8~29.1, 23.9~24.9 ppm) [13, 120, 121, 137,140], which were not observed in the lab PF. These peaks provided evidence that the tannins components (both A ring and B ring) in the bark extractives reacted with the formaldehyde during the resin synthesis. In general, tannin-A ring is more reactive than tannin-B ring towards formaldehyde; however, the reactivity of tannin-B ring could be increased when the pH of the reaction system is higher than 10 [1]. The intensity of these peaks increased with the increasing amount of bark extractives in the resin synthesis.

Previous studies reported that the phenolic compounds of bark extractives could be used as accelerators for PF resins in particleboard and plywood production due to their ability to minimize gelation and shorten press time [92, 138, 139]. Previously published DSC results have found that the introduction of bark extractives to the PF resin synthesis affected the curing behavior and curing kinetics of the resulting bark extractive-PF resins [91, 114]. The bark extractive-PF resins had faster curing rates than the lab PF resin did. The reaction between the tannin components in the bark alkaline extractives and formaldehyde could be one of the reasons for the acceleration of the curing rates of the bark extractives- PF resins.

134

Figure 8.5 Liquid-state 13C NMR spectrum of bark extractive-PF resin with 50 wt% phenol substitution rate by bark extractives

The intensities of the chemical shifts at 60~65 ppm and 35~37 ppm, which were attributed respectively to methylol groups and methylene groups adjacent to phenol rings, were weaker in the bark extractives-PF resins than in the lab PF resins. With the increasing level of phenol substitution by bark extractives in the resin synthesis, the intensities of these chemical shifts in the bark extractives-PF resins decreased. This is likely due to the fact that, as more phenol was replaced by the bark extractives, less methylol groups and methylene groups can be formed between phenols. Meanwhile, the structures of bark extractives could also introduce steric hindrance to the reactions between phenol and formaldehyde. A small chemical shift at 71~73 ppm, which indicated the existence of the methylene ether bridges, was observed in the spectrum of bark extractives-PF resins but not in the spectrum of the lab PF resin.

135

Figure 8.6 Liquid-state 13C NMR spectrum of bark extractive-PF resin with 70 wt% phenol substitution rate by bark extractives

In the spectra of the bark extractive-PF resins with different levels of phenol substitution, the intensities of chemical shifts for the unsubstituted para carbons of phenol at 120-124 ppm were lower than those of the lab PF resin. The chemical shifts at 115-116 ppm, which correspond to the unsubstituted ortho aromatic carbons of phenol, were absent. This suggested that the ortho aromatic carbons of phenol had completely reacted after the introduction of bark extractives to the PF resin. Similar observations have been reported in the literature for the PF resol resins and the reason was that the ortho substitution was more favored by the addition of curing accelerators [127,129]. Therefore, bark alkaline extractives have the potential to speed up the resin curing reaction.

The methoxyl group at 54.3 ppm from the bark extractives was clearly observed in the bark extractive-PF resins, and the intensity increased with the increasing amount of bark extractives (phenol replacement ratio) in the resin synthesis. The intensity of chemical

136

shift at 89.3 ppm of the Ph-CH2OCH2OH increased significantly in the bark extractive-PF resins.

The intensities of the chemical shifts at 150~160 ppm due to phenoxy groups decreased significantly in the spectra of bark extractive-PF resins when compared to the spectrum of the lab PF resin. These chemical shifts were almost invisible in the spectrum of bark extractive-PF resins with 70 % phenol replacement. The chemical shifts at 166~168 ppm, which represent carbonyl groups in bark extractive-PF resin increased in intensities with the increasing level of phenol substitution. The carbonyl groups in the lab PF resins were formed by the oxidation of the phenolic rings to quinone structures during and after the resin synthesis. The carbonyl groups in the bark extractive-PF resins may have originated from the bark extractives, or the oxidation of the phenolic rings or both. The assignment of chemical shifts for PF resin and bark extractive-PF resins is summarized in Table 8.3.

Table 8.3 Assignment of chemical shifts for PF resin and bio-based PF resins [123-127, 137]

Chemical shifts (ppm) Assignment 166-168 carbonyl C=O 153.5-163.9 phenoxy carbons 153.5-157.2 phenoxy, alkylated in ortho position 158.0-159.5 phenoxy, alkylated in para position 129.0-130.4 substituted para aromatic carbons of phenol 126.0-128.1 substituted ortho aromatic carbons of phenol 120.0-124.0 unsubstituted para aromatic carbons of phenol 114.0-116.6 unsubstituted ortho aromatic carbons of phenol

86-93.7 formaldehyde oligomers,Ph-CH2OCH2OH 81.7 unreacted formaldehyde in the resins 69.0-73.0 phenolic methylene ether bridges 63.1-64.7 para methylol 60.4-62.8 ortho methylol 54.3 methoxyl groups 48.7-50 methanol 39.1-41 para-para methylene linkage 34.6-35.0 ortho-para methylene linkage 29.3-29.6 ortho-ortho methylene linkage 29.3-30.9 methylene linkdage between phenol and tannin-B ring 28.8~29.1, 23.9~24.9 methylene linkdage between phenol and tannin-A ring

137

Table 8.4 shows the ratios of the relevant functional groups related to phenolic rings. In comparison with the lab PF resin, bark extractive-PF resins had a higher level of para/ortho methylol group substitution and a lower amount of para-para/ortho-para methylene bridges than the lab PF resin. For the lab PF resin, the substitution occurred more at the para positions than at the ortho positions and para methylol groups reacted more easily with other para positions to form para-para links. When the amount of the free para positions decreased, the reaction with ortho positions increased and led to the decrease in the para/ortho ratio of methylol and the para-para/ ortho-para links. However, the bark extractives allowed the ortho position of phenol to react more favorably with formaldehyde than the para position, and the bark extractives also resulted in more ortho- para methylene linkage formation in the bark extractive-PF resins than the lab PF resin, which yielded a lower ratio of para-para/ortho-para methylene links in the bark extractive- PF resins.

The methylol/methylene ratio and unsubstituted/substituted hydrogen (-H/-CH2OH) ratio of the bark-extractive PF resins were lower than those of the lab PF resin. These two ratios decreased with increasing phenol substitution level by bark extractives. The addition of bark extractives reduced the unsubstituted hydrogen, and methylol/methylene ratio of the resulting bark extractive-PF resins; as a result it could contribute to the acceleration of the curing rate, which was consistent with the previously reported DSC results [91, 114].

Table 8.4 The ratio of the relevant functional groups related to phenolic rings

Sample para/ortho p-p/o-p link Unsubstituted/substituted Methylol/methylene

(-CH2OH) (-CH2-) hydrogen(-H/-CH2OH) (-CH2OH/-CH2-) Lab PF 0.18 1.97 0.18 3.76 Bark extractive- 2.04 0.57 0.16 1.56 PF 30 Bark extractive- 0.98 0.82 0.14 1.44 PF 50 Bark extractive- 0.31 0.74 0.11 1.36 PF 70 Liquefied bark- 0.73 1.17 0.42 3.68 PF

138

8.3.5 Liquid-state 13C NMR spectrum of liquefied bark-phenol formaldehyde resins

The liquid-state 13C NMR spectrum of liquefied bark-PF resin is shown in Figure 7. The intensities of the chemical shifts of the phenoxy groups at 150-160 ppm were similar to those of bark extractive-PF resins, which were weaker than those of the lab PF resin. The chemical shifts at 164~167 ppm represent carbonyl groups of liquefied bark-PF resin. The methoxy group from the liquefied bark components was clearly observed at 55.5 ppm in the spectrum of the liquefied bark-PF resin. The chemical shifts at 71~76 ppm represent the C-α in the β-O-4’ guaiacyl unit as well as the methylene ether bridges.

Figure 8.7 Liquid-state 13C NMR spectrum of liquefied bark-PF resin

Compared with bark extractive-PF resins, the intensities of the chemical shifts at 60~65 ppm and 35~37 ppm attributed to methylol groups and methylene groups were stronger in the liquefied bark-PF resin. It is reasonable to expect that the newly formed phenolated products are generally highly reactive and easily react with formaldehyde to form

139

methylol groups. However, the intensities of the chemical shifts of methylol groups and methylene groups in the liquefied bark-PF resin were weaker than those of the lab PF resin, indicating that the liquefied bark had retarded the reaction of phenol with formaldehyde probably due to steric hindrance. The chemical shifts at 120~130 ppm of the substituted and unsubstituted phenolic rings in the liquefied bark-PF resin had higher intensities than those of the bark extractive-PF resins probably due to phenolation of the bark components. Besides para-para and ortho-para methylene link at 34~35 ppm and 40~41 ppm, the ortho-ortho methylene link with a lower intensity at 29.3 ppm was also observed for the liquefied bark PF resin, which was absence from the bark extractive-PF resins and the lab PF resin. Another difference between the bark extractive-PF resins and the liquefied bark- PF resin was the presence of the chemical shifts at 115-116 ppm for the unsubstituted ortho carbons of phenol in the liquefied bark-PF resin. It could also conclude that the phenolic rings attached to degraded bark components during liquefaction were mostly at their para position. The major different characteristics of the bark extractive-PF resins, liquefied bark-PF resin and lab PF resins are summarized in Table 8.5.

Table 8.5 Summary of the major spectral differences of resins

Resins Major differences based on liquid-state 13C NMR spectra

Lab PF no methylene ether linkage, no ortho-ortho methylene linkage

Bark extractive-PF methylene ether linkage and methoxy group were observed, no unsubstituted ortho aromatic carbon of phenol,

Liquefied bark-PF methylene ether linkage, guaiacyl units, methoxy group, ortho- ortho methylene linkage, unsubstituted ortho aromatic carbon of phenol

These differences could have been caused by the difference in the structure of liquefied

140

bark and bark extractives. Bark after phenolation likely possessed a large amount of phenolic rings, such as tri (4-hydroxylphenyl)methane, 1,1,2-tri(4-hydroxylphenyl)ethane and etc. These products were highly reactive and could further condense with formaldehyde, phenol or with each other. During the condensation reaction with formaldehyde, ortho-ortho methylene links were formed. The liquefied bark could be considered as a curing accelerator to the liquefied bark-PF resin since it promoted the ortho substitution [127, 129]. Meanwhile, due to the steric hindrance of the phenolated products, the ortho position on the phenolic rings could not be easily accessed; as a result, unsubstituted ortho aromatic carbons of phenol were observed in the liquefied bark-PF resin.

The liquefied bark-PF resins had a higher para/ortho (-CH2OH) methylol group substitution and a lower para-para/ortho-para methylene bridge ratio than the lab PF resin. It indicated that the liquefied bark also enhanced the formation of ortho-para methylene linkage in the liquefied bark-PF resin faster comparing to the lab PF resin. The liquefied bark-PF resins had a higher para-para/ortho-para methylene bridge ratio than the bark extractive-PF resin due to its higher steric hindrance effects, i.e. ortho position of the phenolic rings was blocked or hard to be accessed. The methylol/methylene ratio of the liquefied bark-PF resin was lower than that of the lab PF resin but higher than that of the bark extractive-PF resins, while the unsubstituted/substituted hydrogen (-H/-CH2OH) ratio of the liquefied bark-PF resin was higher than that of the lab PF and bark extractive-PF resins. It also supported the fact that the liquefied bark had more reactive sites toward formaldehyde than the bark extractives but the steric hindrance caused by its structure might have slowed down the addition and condensation reactions.

8.3.6 Relationship between the resin molecular structure and curing performance

Previous DSC results in chapter 5 and 6 showed that the bio-based PF resins had higher curing activation energies and faster curing rates than the lab PF resin that had no bark [91, 114]. The curing activation energy for the lab PF resin, liquefied bark-PF resin, bark extractive-PF with 30 wt.% phenol substitution level, bark extractive-PF resin with 50 wt.% phenol substitution level and bark extractive-PF with 70 wt.% phenol substitution

141

level was 70.22 kJ/mol, 78.12 kJ/mol, 79.98 kJ/mol, 79.40 kJ/mol and 83.38 kJ/mol, respectively. The pre-exponential factor indicating the curing rate for the lab PF resin, liquefied bark-PF resin, bark extractive-PF with 30 wt% phenol substitution level, bark extractive-PF resin with 50 wt% phenol substitution level and bark extractive-PF with 70 wt% phenol substitution level was 1.59×108s-1, 2.98×109s-1, 2.78×109s-1, 2.73×109s-1, 7.55×109s-1, respectively. The curing enthalpy for the lab PF resin, liquefied bark-PF resin, bark extractive-PF with 30 wt% phenol substitution level, bark extractive-PF resin with 50 wt% phenol substitution level and bark extractive-PF with 70 wt% phenol substitution level was 343.35 J/g, 219.58 J/g, 237.60 J/g, 182.48 J/g, 136.73 J/g, respectively. Combining these DSC measurements with the liquid-state 13C NMR results in this study, it could be seen that the inclusion of bark components either in the form of phenol liquefied bark or in the form of bark extractives in the PF resins affected the resins’ molecular structure and curing behavior.

In comparison with the lab PF resin, the faster curing rates indicated by the pre- exponential factor of bark extractive-PF resins were highly likely due to the tannin compounds in the bark extractives. The phloroglucinol structure of tannin A rings had a higher reactivity towards formaldehyde than phenol. It would accelerate the curing rate of the resins [92,138,139,140]. Besides, the addition of bark extractives to the resins synthesis facilitated ortho substitution of phenol instead of the para substitution. It also promoted more formation of para-ortho methylene links than para-para methylene links and reduced the unsubstituted hydrogen. All these factors could contribute to the faster curing rate of the resulting bark extractive-PF resins.

For the liquefied bark-PF resin, phenolated products (Figure 8.8) of degraded lignin (diphenylmethanes and guaiacol) and degraded cellulose (methylolphenols) as well as other fragments with newly generated multi-phenolic rings in the structure were more reactive towards formaldehyde, and could help the resin cure faster. The phenolated bark in the resin synthesis favored the formation of ortho-ortho methylene linkage. The phenolated bark also allowed the formation of para-ortho methylene link more than para- para methylene link. The higher para-para/ortho-para methylene bridges ratio, higher methylol/methylene ratio and higher unsubstituted/substituted hydrogen (-H/-CH2OH)

142

ratio of the liquefied bark-PF resin indicated the existing steric hindrance effect of the liquefied bark, which could retard the resins’ curing reactions.

OH HO HO OH OH OH OH

CH CH CH 2 CH2 CH2 CH2 OH CH2

OH OH OMe OH OH OH OH

Figure 8.8. Possible liquefied products from acid-catalyzed phenol liquefaction of bark

Additionally, the larger molecules of the liquefied bark and bark extractives with lower molecular mobility could make it difficult for the resulting resins to form cross-links [8, 13, 32-35, 94]; therefore, the curing activation energies of the bio-based PF resins was higher than that of the lab PF resin. However, the reactive sites on the liquefied bark or bark extractives might still have potential to accelerate the curing process of resins at higher temperatures. Besides, the self-condensation of the liquefied bark components and bark extractives could also contributed to the relatively high pre-exponential factors and faster curing rates of the bio-based PF resins despite of having higher curing activation energies. The addition of bark components either in the form of liquefied bark or in the form of bark extractives to the PF resol resin formulation reduced the curing enthalpy of the resulting resins. Overall, the introduction of bark components to the PF resins accelerated the curing rate of the resulting bio-based PF resins.

8.4 Conclusions

The introduction of bark components to the phenol formaldehyde resin synthesis affected the resins structures and curing characteristics. Bio-based PF resins exhibited higher curing activation energies, higher curing pre-exponential factors and lower curing enthalpy than the lab PF resin. Methylene ether bridges were observed in the bio-based PF resins. Bark components were incorporated into the resulting resin structures as reactants.

Tannin structures of mainly procyanidin type, consisted of phloroglucinol A-ring, catechol

143

B-ring and pyrogallol B-ring, were observed in the alkaline extractives of mountain pine beetle infested lodgepole pine bark. Lignin and hemicellulose fragments were also observed in the bark alkaline extractives. The tannin components in the bark alkaline extractives reacted with formaldehyde during the phenolic resin synthesis and accelerated the curing rate of the resulting resins. The addition of bark alkaline extractives to the phenolic resin synthesis made the ortho substitution of phenol more favorable than the para substitution. The bark extractive-PF resins had a higher para/ortho methylol group substitution, a lower ratio of para-para/ortho-para methylene bridges and a lower ratio of unsubstituted/substituted hydrogen (-H/-CH2OH) than the lab PF resin.

Phenolated products of degraded lignin and degraded cellulose such as triphenylethanes, diphenylmethanes and guaiacol and fragments with new phenolic rings in the structures were observed in the phenol-liquefied bark. The phenolic rings attached to the degraded bark components during liquefaction were mostly at the para-position. The phenolated products can accelerate the curing rate of the liquefied bark-PF resin. Ortho-ortho methylene bridges and unsubstituted ortho position of the phenolic rings were observed in the liquefied bark-PF resin. The liquefied bark-PF resins had a higher ratio of para/ortho

(-CH2OH) methylol group substitution, a higher ratio of unsubstituted/substituted hydrogen (-H/-CH2OH) and a lower ratio of para-para/ortho-para methylene bridges than the lab PF resin. The liquefied bark-PF resins had a higher ratio of para-para/ortho-para methylene bridges, a higher methylol/methylene ratio and a higher ratio of unsubstituted/substituted hydrogen (-H/-CH2OH) than bark extractive-PF resin. Phenol- liquefied bark had more reactive sites toward formaldehyde than bark alkaline extractives but the steric hindrance could have retarded the addition and condensation reactions.

144

Chapter 9

Conclusions and Future work

9.1 Major contributions

Starting with the acid-catalyzed phenol liquefaction and alkaline extraction of mountain pine beetle infested lodgepole pine bark, followed by the formulation and characterization of bio-based PF resins including liquefied bark-PF resin and bark extractive-PF resins, this thesis describes a comprehensive and systematic study on the development and characterization of bark PF resol resins containing mountain pine beetle infested lodgepole pine bark components.

The major findings of this thesis are as follows:

1. Acid-catalyzed phenol liquefaction is an effective method to convert bark from the mountain pine beetle infested lodgepole pine into liquefied bark, which could be used as an alternative phenolic feedstock for PF resol resin formulation. The liquefaction yield, properties of the liquefied bark and unliquefied residues are highly dependent on the liquefaction conditions. Catalyst loading was the most effective factors affecting the liquefaction yield and free phenol content of the liquefied bark fraction, followed by the reaction temperature, reaction time and phenol/bark ratio.

2. The liquefied bark-PF resins made from the mountain pine beetle infested lodgepole pine have higher average molecular weights, higher polydispersity indices and shorter gel times the lab control PF and commercial PF resins. The viscosities of the liquefied bark-PF resins were higher than the viscosity of the lab PF resin, but lower than the viscosity of the commercial PF resin. The liquefied bark-PF resin exhibited both nth-order and autocatalytic cure mechanisms. All these resins had similar dry bonding strengths; the liquefied bark-PF resins showed the highest wet bonding strengths. Beetle infestation was shown to have no negative effect on the bonding properties of the liquefied bark PF resins.

145

3. Alkaline (1% NaOH) extraction is another suitable and effective method to obtain desirable bark extractives including phenolic compounds from the mountain pine beetle infested lodgepole pine barks, to partially substitute petroleum-based phenol in the synthesis of phenolic resol resin. The bark extractive-PF (BEPF) resins were found to have higher molecular weights, higher viscosities and shorter gel times than the lab PF and commercial PF resins. Acid neutralization of the bark extractives increased the molecular weight of the extractives and modified performance and curing behavior of the resulting bark extractive-PF resins. The bark extractive-PF resins made from both neutralized and non-neutralized extractives at 30% replacement of phenol (by weight) exhibited similar dry and wet bond strength to the commercial PF resin. At 50% substitution level, bark extractive-PF resins had dry and wet bond strengths similar to the lab PF resin.

4. The post-curing thermal stability of the bio-based PF resins was similar to that of the lab PF resin but differed significantly from that of the commercial PF resin. The bio-based PF resin made from bark alkaline extractives had better thermal stability than the bio-based PF resin made from phenol liquefied bark. The thermal degradation kinetics for the liquefied bark-PF resin and bark extractive-PF resins were different. Thermal degradation mechanism involving oxidation was observed for the bio-based PF resins.

5. The introduction of bark components (either as liquefied bark or as bark extractives) to the PF resins synthesis affected resin’s molecular structure and curing. Bark components were incorporated into the resulting resin structures as reactants. Methylene ether bridges were found in the bio-based PF resins. Bark components made the formation of ortho-para methylene bridge of bio-based PF resins faster than lab PF resins without bark components. Molecular structures of the liquefied bark-PF resin differed significantly from those of bark extractive-PF resins. The liquefied bark-PF resin had a higher para-para/para-ortho methylene

link (-CH2-), a higher unsubstituted/substituted hydrogen (-H/-CH2OH) ratio and a

higher methylol/methylene (-CH2OH/-CH2-) ratio than the bark extractive-PF resin. The tannin components of the bark extractives accelerated the curing rate of the

146

resulting bark extractive-PF resin. The bark extractives made the ortho position of phenol react more favorably with formaldehyde than the para position. The liquefied bark with phenolated structures had more reactive sites towards formaldehyde than the bark extractives; it also accelerated the curing rate of the liquefied bark-PF resin.

Both phenol liquefied bark and bark alkaline extractives from the mountain pine beetle infested lodgepole pine barks are suitable to be used as an alternative to petroleum-based phenol in the formulation of bio-based phenolic resol resins for use as wood adhesives. The bio-based PF resins had comparable bonding strength and curing performance to the commercial PF resins with good promises for industrial utilization. The conversion methods of obtaining phenol substitutes and their application to bio-based phenolic resol resins formulation are also applicable to non-infested lodgepole pine barks and other tree species.

9.2 Future work

Further work could be made in the research area of bio-based PF resins to further improve performance of the bio-based PF resins to facilitate their future applications and commercialization. Key areas proposed for further work are outlined as follows:

1. It could be worthwhile to explore barks from other tree species, such as non- infested lodgepole pine, Douglas-fir, spruce, sugar maple, aspen, oak, etc., as feedstock. These barks are also largely available from Canadian forest industries. Both single bark species and multiple bark species mixtures are worth to be investigated.

2. It would be of great interest to investigate the adhesive application of the liquefied barks acquired from other liquefaction conditions, the bark alkaline extractives from other extraction conditions, as well as bark phenolic compounds or related fractions from other extraction techniques, such as autoclave extraction or liquefaction, microwave assisted extraction or liquefaction, supercritical phenol liquefaction and extraction, other solvent liquefactions/extraction etc. The bark

147

pretreatment, liquefied barks or bark alkaline extractives fractionation and separation before resin formulation, as well as the application of different fractionated portions from liquefied bark and bark extractives could also be studied.

3. It would be beneficial to further optimize the formulation recipes and synthetic conditions for the bio-based PF resol resins to improve their performance. Different molar ratios of the raw materials, various phenol substitution levels, liquefied bark from various liquefaction conditions, bark extractives from various extraction conditions, different cooking temperature and cooking time etc. should also be further researched.

4. It would be of great importance to investigate free formaldehyde emission of the bio-based PF resol resins and conduct economic analysis and life cycle assessment on this novel green adhesive.

5. It is also interesting to explore and develop bio-based novolac resins using barks for other types of applications.

148

References

1. Pizzi, A., Wood Adhesives Chemistry and Technology, Vol.1. Marcel Dekker, New York, (1993)

2. http://www.ihs.com/products/chemical/planning/ceh/phenolic-resins.aspx

3. Hon, D.N.S, Shiraishi, N., Wood and Cellulose Chemistry, Second ed., Rev. and Expanded, Marcel Dekker. Inc., New York, pp. 243-274. (2000)

4. Harkin, J.M. and Rowe, J.W., Bark and its possible uses, Forest Research Note-091, 4-056-9 71, Forest Products Laboratory, Forest Service, U.S. Department of Agriculture (1971)

5. Van Langenberg, K., Grigsby, W., Ryan, G., Green adhesives: options for the Australian industry-summary of recent research into green adhesives from renewable materials and identification of those that are closet to commercial uptake, Forest & Wood Products for Australia, PNB158-0910, June (2010)

6. Nitmz, H.H., (1983) “ Pizzi, A., Wood adhesives, Ed. Pizzi, A., Marcel Dekker, New York, pp. 248-287. (1983)

7. Truter, P., Pizzi, A., Vermaas, H., Cold-setting wood adhesives from kraft hardwood lignin, Journal of applied polymer science, 51, (7), 1319-1322. (1994)

8. Sowunmi, S., Ebewele, R.O., Conner, A.H., River, B.H., Fortified Mangrove Tannin- Based Plywood Adhesive, Journal of applied polymer science, 62, 577-584. (1996)

9. Barbosa, A.P., Mano, E.B., Andrade, C.T., Tannin-based resins modified to reduce the wood adhesive brittleness, Forest products journal, 50, (9), 89. (2000)

10. Çetin, N.S., Özmen, N., Use of organosolv lignin in phenol-formaldehyde resins for particleboard production I. Organosolv lignin modified resins, International journal of adhesive and adhesion, 22, 477-480. (2002)

11. Çetin, N.S., Özmen, N., Use of organosolv lignin in phenol-formaldehyde resins for particleboard production II. Particleboard production and properties, International journal of adhesive and adhesion, 22, 481-486. (2002)

149

12. Khan, M.A., Ashraf, S.M., Malhotra, V.D., Development and characterization of a wood adhesive using bagasse lignin, International journal of adhesive and adhesion, 24, 485-493. (2004)

13. Pizzi, A., Phenolic and tannin-based adhesive resins by reaction of coordinated metal ligands - 2. Tannin adhesive preparation, characteristics and application, Journal of applied polymer science, 24 (5), 1257-1268. (1979b)

14. Pizzi, A., Scharfetter, H.O., Chemistry and development of tannin-based adhesives for exterior plywood, Journal of applied polymer science, 22 (6), 1745-1761. (1978a)

15. Santana, M.A.E., Baumann, M.G.D., Conner, A.H., Utilization of black wattle bark and tannin liquefied in phenol in the preparation of resol-type adhesives, Lignocellulosic-plastics composites 325-341. (1997) Downloadable from http://www.fpl.fs.fed.us/documnts/pdf1997/santa97a.pdf

16. Grigoriou, A.H., Bark extractives from Pinus halepensis mill fortified with polymeric diisocyanate for exterior grade particleboards, Holz als Roh-und Werkstoff, 55, 269- 274. (1997)

17. Mitsunaga, T. Kondo,O. Abe I., The phenolations of bark extracts in the presence of boron trifluoride and the reactivities of the products with formaldehyde, Mokuzai Gakkaishi, 41 (2), 200‒205. (1995)

18. Demirbas, A., Mechanism of liquefaction and pyrolysis reactions of biomass, Energy conversion and management, 41, 633-646. (2000)

19. Demirbas, A., Biomass resource facilities and biomass conversion processing for fuels and chemicals. Energy conversion and management, 42, 1357-1378. (2001)

20. Demirbas, A., Thermochemical conversion of biomass to liquid products in the aqueous medium. Energy sources. 27, 1235-1243. (2005)

21. Holmbom, B., Hemming, J., Eckerman, C., Value-added products from barks of Nordic wood species using bioconversion and chemical technology. VVT report. 127 (2007)

150

22. Alma, M. H., Yoshioka, M., Yao Y., Some characterizations of hydrochloric acid catalyzed phenolated wood-based materials, Mokuzai Gakkaishi, 41, 741-748. (1995a)

23. Alma, M. H., Yoshioka, M., Yao Y., Preparation and characterization of the phenolated wood using hydrochloric acid as a catalyst, Wood science and technology, 30, 39-47, (1995)

24. Alma, M. H., Yoshioka, M., Yao Y., Phenolation of wood using oxalic acid as a catalyst: Effects of temperature and hydrochloric acid addition, Journal of applied polymer science, 61, 675-683. (1996)

25. Alma, M. H., Yoshioka, M., Yao Y., Preparation of sulfuric acid-catalyzed phenloated wood resin, Wood science and technology, 32, 297-308. (1998)

26. Lin, L. Z., Yao, Y., Yoshioka, M., Molecular weights and molecular weight distributions of liquefied wood obtained by acid-catalyzed phenolysis, Journal of applied polymer science, 351~357. (1996)

27. Lin, L. Z., Yoshioka, M., Yao, Y., Shiraishi, N., Liquefaction of wood in the presence of phenol using phosphoric acid as a catalyst and the flow properties of the liquefied wood, Journal of applied polymer science, 52, 1629-1636. (1994)

28. Maldas, D., Shiraishi, N., Liquefaction of biomass in the presence of phenol and water using alkalies and salts as the catlyast, Biomass and bioenergy, 12, 273-279. (1997)

29. Yilgin, M., Pehlivan, D., Poplar wood-water slurry liquefaction in the presence of formic acid catalyst, Energy conversion and management. 45, 2687-2696. (2004)

30. Krzan, A., Kunaver, M., Tisler, V., Wood liquefaction using dibasic organic acids and glycols, Acta chimica slovenica, 52, 253-258. (2005)

31. Mun, S. P., Gilmour, I. A., Jordan, P. J., Effect of organic sulfonic acids as catalysts during phenol liquefaction of pinus radiate bark. Journal of industrial engineering and chemistry, 12, 720-726. (2006)

32. Lin, L. Z., Yao, Y., Yoshioka, M., Liquefaction mechanism of cellulose in the presence of phenol under acid catalysis, Carbohydrate polymers, 57, 123-129. (2004)

151

33. Lin, L. Z., Yao, Y., Shiraishi, N., Liquefaction mechanism of β-O-4 lignin model compound in the presence of phenol under acid catalysis. Part 1. Identification of the reaction products, Holzforschung, 55, 617-624. (2001)

34. Lin, L. Z., Nakagame, S. J., Yao, Y., Liquefaction mechanism of β-O-4 lignin model compound in the presence of phenol under acid catalysis. Part 2. Reaction behavior and pathways, Holzforschung, 55, 625-630. (2001)

35. Zhang, Y.C., Ikeda, A., Hori, N., Takemura, A., Ono, H., Yamada, T., Characterization of liquefied product from cellulose with phenol in the presence of sulfuric acid, Bioresource technology, 97, 313-221. (2006)

36. Alma, M. H., Kelley, S. S., The application of pyrolysis-molecular beam mass spectrometry for characterization of bark phenolysis products, Biomass and bioenergy, 22, 411-419. (2002)

37. Li, J., Maplesden, F., Commercial production of tannins from radiate pine bark fro wood adhesives, IPENZ transactions, 25, 46-52. (1998)

38. Alma, M. H., Kelley,S. S., Bektas, I., The tensile properties of molding products obtained by the condensation of various tree barks and phenol by using sulfuric acid as catalyst, Journal of materials science letters, 19, 1517-1520. (2000)

39. Alma, M. H., Kelley, S. S., Thermal stability of novolak-type thermosettings made by condensation of bark and phenol, Polymer degradation and stability, 68, 413-418. (2000)

40. Lee, W. J., Liu, C. T., Preparation of liquefied bark-based resol resins and its application to particleboard, Journal of applied polymer science, 87, 1837-1841. (2003)

41. Gao, Z. H., Yuan, J. L., Wang, X. M., Phenolated larch-bark formaldehyde adhesivewith multiple additions of sodium hydroxide, Pigment and resin technology, 36, 279-285. (2007)

42. Brandt, T.G., Mangrove tannin-formaldehyde resins as hot press plywood adhesives, Tectona XLII, 137-150. (1953)

152

43. Plomley, K.F., Tannin-formaldehyde adhesives, CSIRO Division of Forest Products Technology Paper No. 46, Melbourne, Australia, 16-19. (1966)

44. Plomley, K.F., Slashevski, A., Waterproof particleboard, CSIRO Forest Product Newsletter, No. 363, July. (1969)

45. Kreibich, R.E., Hemingway, R.W., Condensed tannin-resorcinol adducts in laminating adhesives, Forest product journal, 35, (3), 23-25. (1985)

46. Kreibich, R.E., Hemingway, R.W., Condensed tannin-sulfonate derivatives in cold setting wood laminated adhesives, Forest product journal, 37, (2), 43-46. (1987)

47. Fechtal, M., Riedl, B., Use of eucalyptus and Acacia mollossima bark extra- formaldehyde adhesives in particleboard manufacture, Holzforschung, 47 (4), 349- 357. (1993)

48. Lu, Y., Shi. Q., Larch tannin adhesive for particleboard, Holz als Roh-und Werkstoff, 53, 17-19. (1995)

49. Roffael, E., Use of spruce tannin as a binder in particleboards and medium density fiberboards (MDF), Holz als Roh-und Werkstoff, 58, 301-305. (2000)

50. Vázquez, G., Preparation of wood adhesives by polycondensation of phenolic acids from Pinus pinaster bark with resols, Holz als Roh-und Werkstoff, 47, 491-494. (1989)

51. Vázquez, G., Curing. kinetics of tannin-phenol-formaldehyde adhesives as determined by DSC, Journal of thermal analysis and calorimetry, 70, 19-28. (2002)

52. Vázquez, G., Curing process of phenol-urea-formaldehyde-tannin (PUFT) adhesives Kinetic studies by DSC and DMA, Journal of thermal analysis and calorimetry, 82, 143-149. (2005)

53. Pichelin, F., Structural beams from thick wood panels bonded industrially with formaldehyde free tannin adhesives, Forest products journal, 56, 5. (2006)

54. Mitsunaga, T., Abe, I., Nogami, T., The formation of ring opened compound from (+)-catechin by BF3 catalyst, Mokuzai Gakkaishi, 39, (3), 328-332. (1993)

153

55. Mitsunaga, T., Abe, I., Ohara, S., The chemical structure and stereochemistry of phenol adduct compound derived from (+)-catechin upon phenolation, Mokuzai Gakkaishi, 40, (1), 100-106. (1994)

56. Lei, H., Pizzi, A., Du, G., Environmentally friendly mixed tannin/lignin wood resins, Journal of applied polymer science, 107, (1), 203-209. (2008)

57. Ballerini, A., Despres, A., Pizzi, A, Non-toxic, zero emission tannin-glyoxal adhesives for wood panels, Holz Roh Werkstoff, 63, (6), 477-478. (2005)

58. Patent “WO9909083”. Oxazolidine hardeners for bonding wood articles with resorcinol resins,

59. Trosa, A., Pizzi, A., A no-aldehyde emission hardener for tannin-based wood adhesives for exterior panels, Holz Roh Werkstoff, 59, (4), 266-271. (2001)

60. Kim, S., Lee, Y. K., Kim, H. J., Lee, H.H., Physico-mechanical properties of particleboards bonded with pine and wattle tannin-based adhesives, Journal of Adhesion Science and Technology, 17, (14), 1863-1875. (2003)

61. Pizzi, A., Meikleham, N., Dombo, B., Roll, W., Autocondensation-based, zero emission, tannin adhesives for particleboard, Holz als Roh- und Werkstoff, 53, (1), 201-204. (1995)

62. Kumar, A., A conceptual comparison of bioenergy options for using mountain pine beetle infested wood in Western Canada, Bioresource technology, 100, 387-399. (2009)

63. Pacific Forestry Centre, 2005. Mountain pine beetle. Insect-host interactions, Outbreak origin. Canadian Forest Service, Natural Resources Canada. http://www.pfc.forestry.ca/entomology/mpb/outbreak/interactions_e.html

64. Clark, B., Pine beetle crosses rockies. Vancouver Sun, A1-A2. (2005)

65. Ministry of Forests and Range (MoFR), Timber supply and the mountain pine beetle infestation in British Columbia, Forest Analysis Branch, British Columbia, October. (2003)

66. Cote, W.A., Young, R.J., Risse, K.B., Costanza, A.F., Tonelli, J.P., Lenocker, C., A

154

carbon balance method for paper and wood products. Environmental Pollution, 116, (Suppl. 1), S1-S6. (2002)

67. Safranyik, L., Wilson, B., The mountain pine beetle: A synthesis of biology, management, and impacts on lodgepole pine”. Canadian Forest Service, Natural Resource Canada, (2006).

68. ASTM D1105-96 Standard test method for preparation of extractive-free wood. (2007)

69. Zobel, B.J., Stonecypher, R., Browne, C., Kellison, R.C., 1966. Variation and inheritance of cellulose in southern pines. TAPPI 49, 383-387.

70. ASTM D1106-96 Standard test method for acid-insoluble lignin in wood.

71. Mané, C., Sommerer, N., Yalcin, T., Cheynier, V., Cole, R.B., Fulcrand, H., Assessment of the molecular weight distribution of tannin fractions through MALDI- TOF MS analysis of protein-tannin complexes, Analytical Chemistry, 79, 2239-2248. (2007)

72. Hillis, W.E., G. Urbach, Reaction of polyphenols with formaldehyde, Journal of Applied Chemistry, 9, 665-673. (1959)

73. Garro Galvez, J.M., Riedl,B., Conner, A.H., Analytical studies on tara tannins, Holzforschung, 51, 235-243.(1997)

74. ASTM D3529/D3529M-97 (2003) e1

75. Kissenger, H.E., Reaction kinetics in differential thermal analysis, Analytical chemistry, 29, 1702-1706, (1957)

76. Flynn, J.H., Wall, L.A., General treatment of the thermogravimetry of polymers, Journal of Physical Chemistry, 70, 487-523. (1966)

77. Ozawa, T., A new method of analyzing thermogravimetric data, Bulletin of the Chemical Society of Japan, 38, 1881-1886. (1965)

78. Vyazovkin, S., Burnham, A.K., Criado, J.M., Pérez-Maqueda, L.A., Popescu, C., Sbirrazzuoli, N., ICTAC Kinetics Committee recommendations for performing kinetic computations on thermal analysis data, Thermochimica Acta, 520, 1-19. (2011)

155

79. ASTM D5868 - 01(2008)

80. PS1-95 Construction and Industrial Plywood http://www.plumcreek.com/Portals/0/downloads/productInfo/V995.pdf

81. Pan, H., Shupe, T.F., Hse, C.Y., Characterization of liquefied wood residues from different liquefaction conditions, Journal of Applied Polymer Science, 105, 3739- 3746. (2007)

82. He, G.B., Riedl, B., A. Aït-Kadi, Model-free kinetics: Curing behavior of phenol formaldehyde resins by differential scanning calorimetry, Journal of applied polymer science, 87, 433-440. (2003)

83. Lei, Y., Wu, Q., Lian, K., Cure kinetics of aqueous phenol-formaldehyde resins used for oriented strandboard manufacturing: Analytical technique, Journal of applied polymer science, 100, 1642-1650. (2006)

84. Lei, Y., Wu, Q., Cure kinetics of aqueous phenol-formaldehyde resins used for oriented strandboard manufacturing: Effects of wood flour, Journal of applied polymer science, 102, 3774-3781. (2006)

85. Li, K., Geng, X., Simonsen, J., Karchesy, J., Novel wood adhesives from condensed tannins and polyethylenimine, International journal of adhesion and adhesives, 24, 327-333. (2004)

86. Yazaki, Y., Collins, P.J., Wood adhesives based on tannin extracts from barks of some pine and spruce species, Holz als Roh- und Werkstoff, 52, 307-310. (1994)

87. Yazaki, Y., Collins, P.J., Wood adhesives from pinus radiate bark, Holz als Roh- und Werkstoff, 52, 185-190. (1994)

88. Ferreira, É.S., Lelis, R.C.C., Brito, E.O., Iwakiri, S., Use of tannin from pinus oocarpa bark for manufacture of plywood, Proceedings of the 51st International Convention of Society of Wood Science and Technology, Nov 10-12, Concepción CHILE, (2008)

89. Weissmann, G., Study on Pine bark extracts, International journal of adhesion and adhesives, 31-35 (1983)

156

90. Fradinho, D. M., Pascoal Neto, C., Evtuguin, D., Jorge, F.C., Irle, M.A., Gil M. H., Pedrosa de Jesus, J., Chemical characterization of bark and of alkaline bark extracts from maritime pine grown in Portugal, Industrial crops and products, 16, 23-32 (2002)

91. Zhao, Y., Yan, N., Feng, M., Characterization of phenol–formaldehyde resins derived from liquefied lodgepolepine barks. International journal of adhesion and adhesives. 30. 689-695. (2010)

92. Fechtal, M., Riedl, B., Use of Eucalyptus and Acacia mollissima bark extract- formaldehyde adhesives in particleboard manufacture, Holzforschung, 47, 349-357. (1993)

93. Lee, W. J., Lan, W. C., Properties of resorcinol–tannin–formaldehyde copolymer resins prepared from the bark extracts of Taiwan acacia and China fir, Bioresource Technology, 97, 257-264. (2006)

94. Pizzi, A., Merlin, M., A new class of tannin adhesives for exterior particleboard, International journal of adhesion and adhesives, 1, 261-264. (1981)

95. Poljanšek, I., Krajnc, M., Characterization of phenol-formaldehyde prepolymer resins by inline FT-IR spectroscopy, Acta Chimica Slovenica, 52, 238-244. (2005)

96. Arasaretnam, S., Karunanayake, L., Synthesis, characterization, and metal adsorption properties of tannin–phenol–formaldehyde resins produced using tannin from dried fruit of Terminalia chebula (Aralu), Journal of applied polymer science, 115, 1081- 1088. (2010)

97. Erä, V. A., Mattila, A., Thermal analysis of thermosetting resins, Journal of Thermal Analysis, 10, 461-469. (1976)

98. Manley, T.R., Thermal analysis of polymers. Pure and Applied Chemistry, 61, 1353- 1360. (1989)

99. Chen, Y.F., Chen, Z.B., Xiao, S.Y., Liu, H.B., A novel thermal degradation mechanism of phenol-formaldehyde type resins, Thermochimica Acta, 47, 639-643. (2008)

157

100. Costa, L. L., Montelera, R.D., Camino, G., Weil, E.D., Pearce, E.M., Structure- charring relationship in phenol formaldehyde type resins, Polymer degradation and stability, 56, 23-35. (1997)

101. Chetan, M.S., Ghadage, R.S., Rajan, C.R., Gunjikar, V.G., Ponrathnam, S., Thermolysis of orthonovolacs. Part 1. Phenol-formaldehyde and m-cresol formaldehyde resins, Thermochimica Acta, 228, 261-270. (1993)

102. Chetan, M.S., Ghadage, R.S., Rajan, C.R., Gunjikar, V.G., Ponrathnam, S., Thermolysis of orthonovolacs. Part 2. Phenol-formaldehyde and α-naphthol- formaldehyde resins, Journal of applied polymer science, 50, 685-692. (1993)

103. Shulman, G.P., Lochte, H.W., Thermal degradation of polymers. 2. mass spectrometric thermal analysis of phenol-formaldehyde polycondensates. Journal of applied polymer science, 10, 619-635. (1996)

104. Popescu, C., Integral method to analyze the kinetics of heterogeneous reactions under non-isothermal conditions A variant on the Ozawa-Flynn-Wall method, Thermochimica Acta, 285, 309-323. (1996)

105. Vyazovkin, S., Modification of the integral isoconversional method to account for variation in the activation energy, Journal of computational chemistry, 22, 178-183. (2001)

106. Órfão, J.J.M., Martins, F.G., Kinetic analysis of thermogravimetric data obtained under linear temperature programming - A method based on calculations of the temperature integral by interpolation, Thermochimica Acta, 390,195-211. (2002)

107. Gyulai, G., Greenhow, E.J., A new integral method for the kinetic analysis of thermogrametric data, Journal of Thermal Analysis, 6, 279-291. (1974)

108. Rao, M.R., Alwan, S., Scariah, K.J., Sastri, K.S., Thermochemical characterization of phenolic resins-Thermogravimetric and pyrolysis-GC studies, Journal of thermal analysis, 49, 261-268. (1997)

109. Rao, M.R., Rao, B.S.M., Rajan, C.R., Ghadage, R.S., Thermal degradation kinetics of phenol-crotonaldehyde resins, Polymer degradation and stability, 61, 283-288. (1998)

158

110. Lee, Y.K., Kim, D.J., Kim, H.J., Hwang, T.S., Rafailovich, M., Sokolov, J., Activation energy and curing behavior of resol- and novolac-type phenolic resins by differential scanning calorimetry and thermogravimetric analysis, Journal of applied polymer science, 89, 2589-2596. (2003)

111. Alonso, M.V., Oliet, M., Domínguez, J.C., Rojo, E., Rodríguez, F.. Thermal degradation of lignin-phenol-formaldehyde and phenol-formaldehyde resol resins. Structural changes, thermal stability, and kinetics, Journal of thermal analysis and calorimetry, 105, 349-356. (2011)

112. Pérez, J.M., Fernández, A., Thermal stability and pyrolysis kinetics of lignin- phenol-formaldehyde resins, Journal of applied polymer science, DOI 10.1002/app.34817. (2011)

113. Zárate, C.N., Aranguren, M.I., Reboredo, M.M., Thermal degradation of a phenolic resin, vegetable fibers, and derived composites, Journal of applied polymer science, 107, 2977-2985. (2008)

114. Zhao, Y., Yan, N., Feng, M., Bark extractives-based phenol-formaldehyde resins from beetle-infested lodgepole pine, Journal of Adhesion Science and Technology, 2012, DOI: 10.1080/01694243.2012.697689

115. Schaber, P.M., Colson, J., Higgins, S., Thielen, D., Anspach, B., Brauer, J., Thermal decomposition (pyrolysis) of urea in an open reaction vessel, Thermochimica Acta, 424,131-142. (2004)

116. Jenkin, D.J., Adhesives from Pinus radiata bark extractives, Journal of adhesion, 16, 299-310. (1984)

117. Stefani, P. M., Peña, C., Ruseckaite, R.A., Piter, J.C., Mondragon, I., Processing conditions analysis of Eucalyptus globulus plywood bonded with resol-tannin adhesives, Bioresource technology, 99, 5977-580. (2008)

118. Navarrete, P., Pizzi, A., Pasch, H., Rode, K., Delmotte, L., MALDI-TOF and 13C NMR characterization of martime pine industrial tannin extract. Industrial Crops and Products, 32, 105-110. (2010)

119. Thompson, D., Pizzi, A., Simple 13C NMR methods for quantitative determinations

159

of polyflavonoid tannin characteristics, Journal of applied polymer science, 55, 107- 112. (1995)

120. Pizzi, A., Stephanou, A., A comparative 13C NMR study of polyflavonoid tannin extracts for phenolic polycondensates. Journal of applied polymer science, 50, 2105- 2013. (1993)

121. Pizzi, A., Stephanou, A., A 13C NMR study of polyflavonoid tannin adhesive intermediates. II. Colloidal state reactions, Journal of applied polymer science, 51: 2125-2130. (1994)

122. Grigsby, W.J., Hill, S.J., Mcintosh, C.D., NMR estimation of extractables from bark: Analysis method for quantifying tannin extraction from bark, Journal of wood chemistry and technology, 23, 179-195. (2003)

123. Werstler, D.D., Quantitative 13C NMR characterization of aqueous formaldehyde resins: 1 Phenol-formaldehyde resins, Polymer, 27, 750-760. (1986)

124. Grenier-Loustalot, M., Larroque, S., Grande, D., Grenier, P., Bedel, D., Phenolic resins: influence of catalyst type on reaction mechanisms and kinetics, Polymer, 37, (8), 1363-1369. (1996)

125. Smit, R., Pizzi, A., Schutte, C.J., Paul, S.O., The structure of some phenol- formaldehyde condensates for wood adhesives, Journal of macromolecule science A, 26, (6), 825-841. (1989)

126. Kim, M.G., Amos, L.W., Barnes, E., Study of the reaction rates and structures of a phenol-formaldehyde resol resin by carbon-13 NMR and gel permeation chromatography, Industrial engineering chemistry research, 29, (10), 2032-2037. (1990)

127. Luukko, P., Alvila, L., Holopainen, T., Rainio, J., Pakkanen, T.T., Effect of alkalinity on the structure of phenol-formaldehyde resol resins, Journal of applied polymer science, 82,(1), 258-262. (2001)

128. He, G.B., Yan, N., 13C NMR study on structure, composition and curing behavior of phenol-urea-formaldehyde resol resins, Polymer, 45, 6813-6122. (2004)

160

129. Lei, H., Pizzi, A., Despres, A., Pasch, H., Du, G.B., Ester acceleration mechanisms in phenol-formaldehyde resin adhesives. Journal of applied polymer science, 100, 3075-3093. (2001)

130. Park, B.D., Ried, L.B., 13C NMR study on cure-accelerated phenol-formaldehyde resins with carbonates, Journal of applied polymer science, 77, 841-51. (2000)

131. Fan, D.B., Chang, J.M., Li, J., Mao, A., Zhang, L.T., 13C NMR study on the structure of phenol-urea-formaldehyde resins prepared by methylolureas and phenol, Journal of applied polymer science, 112, 2195-2200. (2009)

132. Fan, D.B., Chang, J.M., Li, J.Z., On the structure and cure acceleration of phenol- urea-formaldehyde resins with different catalysts. European polymer journal, 45, 2849-2857. (2009)

133. Vázquez, G., López-Suevos, F., Villar-Garea, A., González-Alvarez, J., Antorrena, G., 13C NMR analysis of phenol-urea-formaldehyde prepolymers and phenol-urea- formaldehyde-tannin adhesives, Journal of adhesion science and technology, 18, (13), 1529-1543. (2004)

134. Lorenz, K., Preston, C.M., Characterization of high-tannin fractions from humus by 13C cross-polarization and magic angle spinning nuclear magnetic resonance. Journal of environmental quality, 31, 431-436. (2002)

135. Sun, X.F., Sun, R.C., Fowler, P., Baird, M.S., Extraction and characterization of original lignin and hemicelluloses from wheat straw. Journal of agriculture and food chemistry, 53, 860-870. (2005)

136. Vázquez, G., Antorrena, G., González, J., FTIR, 1H and 13C NMR characterization of acetosolv-solubilized Pine and Eucalyptus lignins, Holzforschung, 51, 158-166. (1997)

137. Werstler, D.D., Quantitative 13C NMR characterization of aqueous formaldehyde resins: 2 Resorcinol-formaldehyde resins, Polymer, 27, 757-762. (1982)

138. Vázquez. G., González-Alvarez, J., Antorrena, G., Curing of a phenol- formaldehyde-tannin adhesives in the presence of wood. Journal of thermal analysis and calorimetry, 84, 651-654. (2006)

161

139. Chen, C.M., Copolymer resins of bark and agricultural residue extracts with phenol and formaldehyde: 40 percent weight of phenol replacement, Forest product journal, 32, 14-18. (1982)

140. Pizzi, A., Horak, R. M., Ferreira, D. Roux, D. G., Condensates of phenol, resorcinol, phloroglucinol and pyrogallol as model compounds of flavonoid A- and B- rings with formaldehyde, Journal of applied polymer science, 24, 1571-1578. (1979)

162

Appendix

MALDI-TOF-TOF spectra of bio-based bark PF resins

Figure 1. MALDI-TOF-TOF spectra of liquefied bark-PF resin made from mountain pine beetle infested lodgepole pine

Figure 2. MALDI-TOF-TOF spectra of liquefied bark-PF resin made from non-infested lodgepole pine

163

Figure 3. MALDI-TOF-TOF spectra of Laboratory made control PF resin

Figure 4. MALDI-TOF-TOF spectra of bark alkaline extractives

164

Figure 5. MALDI-TOF-TOF spectra of bark extractive-PF resin with 30 wt.% phenol substitution rate by bark extractives

Figure 6. MALDI-TOF-TOF spectra of bark extractive-PF resin with 50 wt.% phenol substitution rate by bark extractives

165

Figure 7. MALDI-TOF-TOF spectra of bark extractive-PF resin with 70 wt.% phenol substitution rate by bark extractives

166

Thermal degradation of residues from different liquefaction conditions

100

P/B=2

Weight (%) Weight 40 P/B=3 P/B=4 P/B=5 Bark 20

0 0 100 200 300 400 500 600 700

Temperature (℃)

Figure 8. TGA of residues from different phenol/bark ratio

(Reaction temperature=150°C, catalyst = 3 wt.% of phenol, reaction time =60 min.)

167

100

Temperature 120℃

Weight (%) Weight 40 Temperature 150℃ Temperature 180℃

Bark 20

0 0 100 200 300 400 500 600 700

Temperature (℃)

Figure 9. TGA of residues from different reaction temperature

(Reaction time=60 min, catalyst = 3 wt.% of phenol, phenol/bark ratio = 3)

168

100

Reaction time 30min Weight (%) Weight 40 Reaction time 60min Reaction time 120min Reaction time 150min 20 Bark

0 0 100 200 300 400 500 600 700

Temperature (℃)

Figure 10. TGA of residues from different reaction time

(Reaction temperature=150°C, catalyst = 3 wt.% of phenol, phenol/bark ratio = 3)

169