UNIVERSITY OF CALGARY Study of the Pyrolysis of Straw Biomass
Transcription
UNIVERSITY OF CALGARY Study of the Pyrolysis of Straw Biomass
UNIVERSITY OF CALGARY Study of the Pyrolysis of Straw Biomass for Bio-oil Production and its Catalytic Upgrading by Aqsha Aqsha A THESIS SUBMITTED TO THE FACULTY OF GRADUATE STUDIES IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY IN CHEMICAL ENGINEERING GRADUATE PROGRAM CHEMICAL AND PETROLEUM ENGINEERING CALGARY, ALBERTA APRIL 2016 © Aqsha Aqsha 2016 Abstract This thesis presents a comprehensive study of the pyrolysis of lignocellulosic biomass (sawdust, wheat, oat, flax and barley straws). In the first part of the study, the pyrolysis of sawdust was observed using a thermogravimetric analyzer (TGA), in order to understand the devolatilization process and to obtain its global kinetic parameters. The influences of particle size, initial weight of the sample and heating rate on the devolatilization of sawdust particles were assessed. It was observed that the pyrolysis of sawdust differed significantly with variations in heating rate. As the heating rate increased, the char yield also increased. The kinetic parameters, including activation energy (E), frequency factor (k0) and order of reaction (n), for the two stages considered in the model were: E A2 = 79.53 (kJ/mol), EA3 = 60.71 (kJ/mol); k02 = 1.90 × 106 (1/min), k03 = 1.01 × 103 (1/min); n2 = 0.91, n3 = 1.78, respectively. In the second part of the study, the pyrolysis of several Canadian straw biomasses was studied using a TGA and a bench-scale horizontal fixed-bed reactor. The effects of various catalysts on product yields are discussed. When using zeolite catalysts, the bio-oil and bio-char yields of the straw pyrolysis increased to 46.44% and 38.77%, respectively, while the gas yield was decreased to 13.65%. The use of the catalyst zeolite ZY-SS had the most significant effect on overall biooil and bio-char yields, increasing the bio-oil yield by about 2% and the bio-char yield by 8%. A screening of different catalysts unveiled that Ni-Mo/TiO2 was the most active catalyst. The structure was investigated using Brunauer-Emmett-Teller (BET) surface area showed that the NiMo/TiO2 catalyst presented a higher surface area and more optimal mesopores than Ni-V/TiO 2. ii Based on the results, it appears that the significant hydrodeoxygenation (HDO) activity for both the catalysts attributed to a high dispersion of metals and acidic sites, which was affected by the interaction between the nickel and the titania support. The activation energy (E) values for guaiacol reactions over Ni-Mo/TiO2 and Ni-V/TiO2 were 93.6 and 98.4 kj/mol, respectively; and, the E values for anisole reactions over Ni-Mo/TiO 2 and Ni-V/TiO2 were 80.9 and 53.9 kj/mol, respectively. iii Acknowledgement Alhamdulillah. First and foremost, I would like to express my highest gratitude to Allah The Creator and Muhammad His Messenger, for providing me with the ability to think and to invent as well as His guidance throughout my entire life. I would also like to express my sincere gratitude to my supervisor, Dr. Nader Mahinpey, for his continuous support of my PhD study and research, and for his patience, advice and immense knowledge. His guidance has helped me throughout the research and writing of this thesis. I cannot imagine having a better supervisor and mentor for my PhD study. My sincere thanks also goes to Dr. Lakshmi Katta, Dr. Puli, Dr. Thila, Dr. Feridoun and Dr. Karami who provided me with opportunities to gain knowledge, as well as training during my early days in the laboratory. Without their valuable support, expertise and input, it would not have been possible to conduct this research. My sincere thanks goes to Mansour Tijani for his frequent help and discussions during the last few months of my research. I also want to mention my sense of gratitude to Ludivine Grass, Camilla Fernandes De Oliveira and Aurore Coussirat-Bourg for their contribution to the research during their internships in the laboratory. I thank my fellow lab mates (Ranjani, Hashem, Ehsan, Shuba, Alireza, Kavan, Bahareh, Arturo, Mansour, Crystal and Marlon) for the stimulating discussions, the long hours of working iv together before deadlines, and all the fun we have had over the last six years. I also thank my fellow internship students (Aruna, Rebecca, Andy, Ragav, Marco, Thais, Varada, Lucas, Felipe, Diego, Maria, Samantha, Yifei, Qinran, Tommy and Bruce), who spent their time with me during laboratory sessions. I also thank Spensix 3B, Badjitoe, Kakegawa, Soulmates, TK2000, Himatekers, Bukaka Fellow (Tongkeng) and all my teachers for their moral support and their interaction with me during my PhD study. Last but not the least, I would like to thank my wife Belladonna Maulianda, my son Nafi Rabbani and my family for supporting me spiritually throughout the writing of this thesis and my life in general. v Dedication I dedicate this work to my beloved wife, son and family. vi Table of Contents Abstract ........................................................................................................................... ii Acknowledgement ......................................................................................................... iv Dedication ...................................................................................................................... vi Table of Contents .......................................................................................................... vii List of Tables ...................................................................................................................x List of Figures and Illustrations .................................................................................... xii List of Symbols, Units, Chemical Formulas and Abbreviations ...................................xv CHAPTER 1: INTRODUCTION .......................................................................................1 1.1 Overview ....................................................................................................................1 1.2 Literature Review ......................................................................................................4 1.2.1 Introduction to Biomass ....................................................................................4 1.2.2 Pyrolysis of Lignocellulosic Biomass ...............................................................7 1.2.2.1 Slow (Conventional) Pyrolysis ................................................................8 1.2.2.1 Fast Pyrolysis ...........................................................................................9 1.2.3 Kinetic Study of Pyrolysis of Biomass ............................................................10 1.2.4 Pyrolysis Behaviors of Lignocellulosic Biomass ............................................11 1.2.5 Bio-Oil Production from Lignocellulosic Biomass .........................................12 1.2.6 Bio-Oil Upgrading ...........................................................................................13 1.3 Objectives ................................................................................................................15 1.4 Organization of the Thesis .......................................................................................16 1.5 Outcome of the Thesis .............................................................................................18 1.5.1 Published peer-reviewed journal papers and contribution: .............................18 1.5.2 Submitted peer-reviewed journal papers and contribution:.............................21 1.5.3 Published peer-reviewed conference proceedings and contribution: ..............22 1.6 References ................................................................................................................25 CHAPTER 2: RESEARCH METHODOLOGY AND SETUP .......................................30 2.1 Study of Biomass Pyrolysis and its Devolatilization Kinetics Using TGA.............31 2.1.1 Ultimate (Elemental Analysis) using Elemental Analyzer ..............................32 2.1.2 Proximate Analysis using TGA .......................................................................33 2.1.3 Thermal Degradation (Devolatilization) using TGA ......................................34 2.2 Bio-oil Production from Biomass Using Bench-scale Pyrolysis Reactor................35 2.3 Bio-oil Upgrading Using Catalyst ...........................................................................38 2.4 References ................................................................................................................39 CHAPTER 3: STUDY OF SAWDUST PYROLYSIS AND ITS DEVOLATILIZATION KINETICS ................................................................................................................40 3.1 Presentation of the Article .....................................................................................40 3.2 Abstract ..................................................................................................................41 3.3 Introduction ............................................................................................................42 3.4 Experimental Section .............................................................................................45 vii 3.4.1 Materials ..........................................................................................................45 3.4.2 Proximate Analysis ..........................................................................................45 3.4.3 Pyrolysis Experiment using TGA ....................................................................46 3.5 Results and Discussions .........................................................................................46 3.5.1 Ultimate and Proximate Analysis of Sawdust Sample ....................................46 3.5.2 Pyrolysis of Sawdust .......................................................................................49 3.5.2.1 Effect of Particle Size ............................................................................50 3.5.2.2 Effect of Initial Sample Weight .............................................................52 3.5.2.3 Effect of Heating Rate ...........................................................................54 3.5.2 Kinetics of Sawdust Pyrolysis .........................................................................56 3.6 Conclusions ............................................................................................................60 3.7 Acknowledgement .................................................................................................61 3.8 References ..............................................................................................................62 CHAPTER 4: CATALYTIC PYROLYSIS OF STRAW BIOMASSES (WHEAT, FLAX, OAT AND BARLEY STRAW) AND THE COMPARISON OF THEIR PRODUCT YIELDS ....................................................................................................................65 4.1 Presentation of the Article .....................................................................................65 4.2 Abstract ..................................................................................................................67 4.3 Introduction ..............................................................................................................68 4.3 Experimental Section .............................................................................................70 4.4.1 Raw Biomass Sample and catalysts ................................................................70 4.4.2 Proximate and Elemental Analysis ..................................................................71 4.4.3 Pyrolysis Experiment.......................................................................................72 4.4.4 Product Analysis and Characterization ...........................................................74 4.4 Results and Discussions .........................................................................................74 4.5.1 Proximate and Elemental Analysis ..................................................................74 4.5.2 Pyrolysis of Raw Biomass ...............................................................................78 4.5.3 Catalytic Pyrolysis of Raw Biomasses ............................................................81 4.5.4 Analysis of Pyrolysis Products ........................................................................81 4.5 Conclusions ............................................................................................................87 4.6 Acknowledgement .................................................................................................88 4.7 References ..............................................................................................................89 CHAPTER 5: CATALYTIC HYDRODEOXYGENATION OF GUAIACOL AS LIGNIN MODEL COMPONENT USING NI-MO/TIO2 AND NI-V/TIO2 CATALYSTS ...92 5.1 Presentation of the Article .....................................................................................92 5.2 Abstract ..................................................................................................................93 5.3 Graphical Abstract .................................................................................................94 5.4 Introduction ............................................................................................................95 5.5 Experimental ..........................................................................................................97 5.5.1 Catalyst Preparation.........................................................................................97 5.4.2 Characterization Studies ..................................................................................98 5.4.3 Reaction Study: HDO of Guaiacol ..................................................................99 5.6 Results and Discussions .......................................................................................100 5.5.1 Catalyst Properties .........................................................................................100 5.5.2 Catalytic Activity...........................................................................................110 viii 5.7 Conclusions ..........................................................................................................126 5.8 Acknowledgement ...............................................................................................127 5.9 References ............................................................................................................128 CHAPTER 6: UPGRADING OF ANISOLE COMPONENT BY NI-MO/TIO2 AND NIV/TIO2 CATALYSTS: SYNTHESIS, CHARACTERIZATION AND KINETIC MEASUREMENTS ................................................................................................131 6.1. Presentation of the Article ...................................................................................131 6.2. Abstract ................................................................................................................132 6.3 Introduction ............................................................................................................133 6.3. Experimental Section ...........................................................................................134 6.4.1 Catalysts Preparation .....................................................................................134 6.4.2 Characterization Studies ................................................................................135 6.4.3 Reaction Study: HDO of Anisole ..................................................................136 6.4. Results and Discussions .......................................................................................137 6.5.1 Structural Properties ......................................................................................137 6.5.2 Surface Analysis ............................................................................................140 6.5.3 Reaction and Kinetic Study of Anisole .........................................................145 6.5. Conclusions ..........................................................................................................154 6.6. Acknowledgement ...............................................................................................155 6.7. References ............................................................................................................157 CHAPTER 7: CONCLUSIONS AND SUGGESTIONS FOR FUTURE WORKS ......160 7.1. Relation between Chapters and Overall Achievement ........................................160 7.2. Conclusions ..........................................................................................................162 7.3. Suggestions for Future Work ...............................................................................165 APPENDIX A: LICENSE TO RE-PRINT ......................................................................166 A.1. License to re-print and re-use the content of WIT Publications (Chapter 3 & 4) 166 A.2. License to re-print of Chapter 3 – Study of Sawdust Pyrolysis and its Devolatilization Kinetics ................................................................................................................168 A.3. License to re-print of Chapter 5 – Catalytic Hydrodeoxygenation of Guaiacol as Lignin Model Component Using Ni-Mo/TiO2 and Ni-V/TiO2 Catalysts ............174 ix List of Tables Table 1.1 Typical elemental and water content of bio-oils from fast pyrolysis [4,5] ..................... 3 Table 1.2 Chemical composition of several types of biomass ........................................................ 6 Table 1.3 Typical properties of oils obtained through flash pyrolysis ......................................... 14 Table 3.1 Ultimate analysis of the sawdust in weight percentage ................................................ 46 Table 3.2 Proximate analysis of sawdust at different particle sizes ............................................. 48 Table 3.3 Char yield percentage for different values of particle size of the sample for sawdust pyrolysis with initial weight of 10 mg and heating rate of 10 C/min. ................................. 52 Table 3.4 Char yield percentage for different values of initial weight of the sample for sawdust pyrolysis with average particle size of 352 µm and heating rate of 10 C/min. ..... 52 Table 3.5 Char yield percentage for different values of heating rate of the sample for sawdust pyrolysis with average particle size of 352 µm and initial weight of 10 mg. ....................... 54 Table 3.6 Weight loss and ultimate devolatilization yield of the devolatilization events with average weight prior to devolatilization of 9.91 mg at different heating rates. .................... 58 Table 3.7 Kinetic parameters of the three events considered in the kinetic model. ..................... 59 Table 4.1 Chemical composition of wheat, oat, flax and barley straw samples ........................... 70 Table 4.2 Elemental composition and surface area data of the catalysts ...................................... 71 Table 4.3 Proximate and elemental analyses of biomasses .......................................................... 75 Table 4.4 Data comparison between raw biomass and its bio-char product ................................. 76 Table 4.5 Product yield comparison of the pyrolysis of wheat, oat, flax and barley straws using a TGA .......................................................................................................................... 78 Table 4.7 Elemental analysis of bio-char produced from each biomass sample .......................... 79 Table 4.8 Proximate and CHNS analysis of bio-char products after pyrolysis ............................ 82 Table 4.9 Water content of bio-oil produced from pyrolysis ........................................................ 83 Table 4.10 Heating value (MJ/kg) of biomasses and its products ................................................ 83 Table 4.11 GC-MS analysis of bio-oil produced from straw biomass with and without catalyst .................................................................................................................................. 85 x Table 4.12 Micro-GC analysis of gas produced from wheat straw biomass with and without catalyst .................................................................................................................................. 86 Table 5.1 Structural properties of metal–supported catalysts ..................................................... 104 Table 5.2 Structural properties of metal–supported catalysts ..................................................... 108 Table 6.1 Physico-chemical properties of Ni-Mo/TiO2 and Ni-Mo/TiO2 catalysts.................... 137 Table 6.2 Kinetic parameter of HDO reaction using Ni-Mo/TiO 2 and Ni-V/TiO2 .................... 154 xi List of Figures and Illustrations Fig. 1.1 Typical lignocellulosic biomass waste found in Canada (from left to right: wheat straw, oat straw, flax straw and barley straw) ......................................................................... 2 Fig. 1.2 Devolatilization profile of biomass components [22] ....................................................... 8 Fig. 2.1 Diagram of the experimental setup .................................................................................. 30 Fig. 2.2 Schematic of pyrolysis experiment using TGA............................................................... 31 Fig. 2.3 Schematic of Perkin Elmer 2400 Elemental Analyzer [1] .............................................. 32 Fig. 2.4 TGA profile of calcium oxalate [2] ................................................................................. 34 Fig. 2.5 Experimental setup of bench-scale pyrolysis .................................................................. 35 Fig. 3.1 TG profile for proximate analysis of sawdust for the particle size of 250 µm. ............... 47 Fig. 3.2 Classification of devolatilization stages under nitrogen atmosphere. ............................. 49 Fig. 3.3 Pyrolysis curve of sawdust at different particle sizes of sample with heating rate of 10 °C/min and initial weight of 10 mg of sample. (a) TG profile and (b) DTG profile. ....... 51 Fig. 3.4 Pyrolysis curve of sawdust at different initial weights of sample with heating rate of 10 °C/min and particle size of 352 µm. (a) TG profile and (b) DTG profile. ....................... 53 Fig. 3.5 Pyrolysis curve of sawdust at different heating rates with particle size of 325 µm and initial weight of 10 mg of sample. (a) TG profile and (b) DTG profile. ............................... 55 Fig. 3.6 Typical comparison of experimental data with predicted values at heating rate of 10 °C/min. .................................................................................................................................. 60 Fig. 4.1 Typical profile for proximate analysis using a TGA ....................................................... 72 Fig. 4.2 Thermogravimetric profiles of raw biomasses subject to pyrolysis. ............................... 73 Fig. 4.3 Location of straw biomasses and their biochar products on a Van Krevelen diagram. .. 77 Fig. 4.4 Product yields (wt%) from pyrolysis of straw biomass ................................................... 80 Fig. 4.5 Comparison of the product yields among the biomasses ................................................ 81 Fig. 5.1 X-ray diffraction patterns of anatase, Ni/TiO2, Ni-V/TiO2 and Ni-V/TiO2 before reduction ............................................................................................................................. 102 Fig. 5.2 X-ray diffraction patterns of anatase, Ni/TiO2, Ni-V/TiO2 and Ni-V/TiO2 after the reduction ............................................................................................................................. 103 xii Fig. 5.3 TEM image of TiO2, Ni/TiO2, Ni-V/TiO2 and Ni-Mo/TiO2 catalysts ........................ 105 Fig. 5.4 Adsorption isotherms of fresh and reduced (a) TiO2 , (b) Ni-Mo/TiO2 and (c) NiV/TiO2 catalysts. Inset: Pore size distributions................................................................... 108 Fig. 5.5 TPD-NH3 profiles of TiO2, Ni/TiO2, Ni-V/TiO2 and Ni-Mo/TiO2 after the reduction treatment ............................................................................................................................. 109 Fig. 5.6 Screening of catalysts. (a) Promoter effect: V, Mo, Cu, Fe. (b) Support effect: TiO 2, Al2O3, ZrO2, Fe2O3, SiO2 for HDO of guaiacol. Reaction conditions: 300 °C, 1 mL of guaiacol, H2 pressure of 100 psi, 5 h .................................................................................. 113 Fig. 5.7 Effect of catalyst on guaiacol conversion: Reaction conditions: 300 °C, 1 mL of guaiacol, H2 pressure of 100 psi, 5 h. ................................................................................. 115 Fig. 5.8 Effect of Ni loading (0, 5, 10 and 20 wt%) on HDO of guaiacol. Reaction conditions: 300 °C, 1 mL of guaiacol, H2 pressure of 100 psi, 5 h..................................... 116 Fig. 5.9 Effect of temperature on HDO of guaiacol. Reaction conditions: 1 mL of guaiacol , H2 pressure of 100 psi, 5 h .................................................................................................. 117 Fig. 5.10 Pressure dependence on HDO of guaiacol (a) Ni-Mo/TiO2 and (b) Ni-V/TiO2. Reaction conditions: 300 °C, 1 mL of guaiacol, 5 h ........................................................... 119 Fig. 5.11 Effect of reaction time on HDO of guaiacol (a) Ni-Mo/TiO 2 and (b) Ni-V/TiO2. Reaction conditions: 300 °C, 1 mL of guaiacol, H2 pressure of 100 psi ............................ 121 Fig. 5.12 Determination of order and rate constant by integral method: (a) Ni-Mo/TiO 2 and (b) Ni-V/TiO2. Reaction conditions: 300 °C, 1 mL of guaiacol, H2 pressure of 100 psi.... 123 Fig. 5.13 Effect of water on HDO of guaiacol: (a) Ni-Mo/TiO2 and (b) Ni-V/TiO2. Reaction conditions: 300 °C, 1 mL of guaiacol, H2 pressure of 100 psi, 5 h..................................... 126 Fig. 6.1 X-ray diffraction pattern of (a) TiO2, fresh Ni-Mo/TiO2 and Ni-Mo /TiO2 after the reduction and (b) TiO2, fresh Ni-V/TiO2 and Ni-V/TiO2 after the reduction. .................... 138 Fig. 6.2 TEM images of Ni-V/TiO2 and Ni-Mo/TiO2 catalysts before (fresh, F) and after reduction (R) ....................................................................................................................... 139 Fig. 6.3 Ni 2p spectra for Ni-Mo/TiO2 (fresh and reduced) and Ni-V/TiO2 (fresh and reduced)............................................................................................................................... 141 Fig. 6.4 Ti 2p spectra for Ni-Mo/TiO2 (fresh and reduced) and Ni-V/TiO2 (fresh and reduced) 142 Fig. 6.5 Core level spectra of O 1s for Ni-Mo/TiO 2 (fresh and reduced) and Ni-V/TiO2 (fresh and reduced) ........................................................................................................................ 143 Fig. 6.6 Mo 3d core level spectra for Ni-Mo/TiO2 (fresh and reduced) ..................................... 144 xiii Fig. 6.7 V 2p core level spectra for Ni-V/TiO2 (fresh and reduced) .......................................... 144 Fig. 6.8 Effect of temperature on HDO of anisole. Reaction conditions: 1 mL of anisole, H 2 pressure of 100 psi, 6 h and using Ni-Mo/TiO 2. ................................................................. 146 Fig. 6.9 Effect of temperature on HDO of anisole. Reaction conditions: 1 mL of anisole, H 2 pressure of 100 psi, 6 h and using Ni-V/TiO2. .................................................................... 146 Fig. 6.10 Effect of reaction time on HDO of anisole. Reaction conditions: 1 mL of anisole , H2 pressure of 100 psi, 300 °C and using Ni-Mo/TiO2....................................................... 147 Fig. 6.11 Effect of reaction time on HDO of anisole. Reaction conditions: 1 mL of anisole, H2 pressure of 100 psi, 300 °C and using Ni-V/TiO2. ........................................................ 147 Fig. 6.12 Plots of the of anisole concentration versus reaction time of HDO of anisole using Ni-Mo/TiO2 at (a) 250 °C, (b) 275 °C, (c) 300 °C and (d) 325 °C. .................................... 150 Fig. 6.13 Plot of the of logarithmic natural of rate constant versus 1/T of HDO of anisole using Ni-Mo/TiO2 to determine the kinetic parameter ....................................................... 151 Fig. 6.14 Fig. 6.12 Plot of the of anisole concentration versus reaction time of HDO of anisole using Ni-V/TiO2 (a) 250, (b) 275, (c) 300, (d) 325 °C temperatures ..................... 153 Fig. 6.15 Plot of the of logarithmic natural of rate constant versus 1/T of HDO of anisole using Ni-V/TiO2 to determine the kinetic parameter .......................................................... 154 Fig. 7.1 Relation between chapters ............................................................................................. 161 xiv List of Symbols, Units, Chemical Formulas and Abbreviations Symbols [C] Weight percentage of carbon in biomass or compound [O] Weight percentage of oxygen in biomass or compound [H] [N] A C d dt dT E Ei H HHV k ki k0 k0i n ni pH Weight percentage of hydrogen in biomass or compound Weight percentage of nitrogen in biomass or compound Amount of ash from devolatilization process Amount of carbon from devolatilization process Diameter of particle (mm) Derivative of time variable Derivative of temperature variable Activation energy (J/mol) Activation energy of event i (J/mol) Heating rate High heating value Rate constant (1/min) Rate constant of event i Pre-exponential factor (1/min) Pre-exponential factor of event i Order of the reaction Order of the reaction from event i Acidity xv R Universal gas constant (J/Kmol) T Temperature (K) t V V(t) Vi(t) Vi * Time (min) Amount of volatile material from devolatilization process Mass fraction of volatile material yield up to time t Accumulated amount of released volatile material from event i to time t Mass fraction of volatile material yield up to t = ∞ xvi Greek Letters Kα X-ray spectral line λ Wavelength (nm) θ Angle of diffraction xvii Unit 1/min Unit of rate constant: one over minute °C Temperature unit: degree in Celsius °C/min Degree of Celsius per minute cm3 Unit of volume (space): cubic centimetre cm3/g Unit of pore volume: cubic centimetre per gram sample cP Unit of viscosity: centipoise eV Unit of energy: electron volt (1 eV = 1.602×10−19 joule) g/cm Unit of density: gram per unit volume of cubic centimetre kJ/mol Unit of activation energy: kilo Joule per mole kV Unit of electric potential: kilovolt m Unit of length: metre m2/g Unit of surface area: square metre per gram mg Unit of mass: milligram µm Unit of length: micrometre min Unit of time: minute MJ/kg Unit of energy density: mega Joule per kilogram weight ml Unit of liquid volume: millilitre mL/min Unit of gas flow rate: millilitre per minute mm Unit of length: millimetre h MP Unit of time: hour Mega Pascal xviii Mt Unit of mass: megatons nm Unit of length: nanometre PJ Unit of energy: peta Joule psi Unit of pressure: pound per square inch W Unit of power: watt (1 W = 1 J.s) Wt% Percentage of weight ° Unit of angle: degree xix Chemical Formulas Al2O3 Aluminum oxide catalyst support C=C Chemical double bond between carbon atom C1 Hydrocarbon with 1 atom carbon C2 Hydrocarbon with 2 atom carbon C2H4O2 Acetic Acid C4H6O2 4-Hydroxybutyric acid lactone C4H8O2 1-Hydroxy-2-butanone C5H4O2 Furfural C5H6O2 2-Furanmethanol C5H8O3 Acetoxyacetone C5H10O2 trans-1,3-Cyclopentanediol C5H10O2 Tetrahydro-2-furanmethanol C6H6O Phenol C6H8O 2-Methyl-2-cyclopenten-1-one C6H8O 3-Methyl-2-cyclopenten-1-one C6H8O2 2-Cyclopenten-1-one (C6H10O5)n Polysaccharide consisting of a linear chain of linked D-glucose units C6H12O2 4-Hydroxy-4-methyl-2-pentanone C7H8O o/m/p-Cresol C7H8O2 Guaiacol C8 Hydrocarbon with 8 atom carbon xx C8H10O2 4-Methylguaiacol (Creosol) C8H10O3 2,6-Dimethoxyphenol C9H12O2 4-Ethylguaiacol C20 Hydrocarbon with 20 atom carbon CH4 Methane gas C-O Chemical bond of carbon and oxygen atom Co Cobalt metal Co–Mo Cobalt and molybdenum active metal system Co2P Cobalt phosphide catalyst CO Carbon monoxide gas CO2 Carbon dioxide gas COS Carbonyl sulphide Cu Copper Cu(NO3)2·3H2O Cupper nitrate trihydrate Fe Iron Fe(NO3)3·9H2O Iron nitrate nonahydrate Fe2O3 Iron oxide catalyst support Fe2P Iron phosphide catalyst H2 Hydrogen gas H2O Water H2S Hydrogen sulphide La1−x.Ce.x NiO3 Perovskite system of lanthanum, cerium and nickel oxide xxi Mo Molybdenum MoP Molybdenum phosphide catalyst N2 Nitrogen gas N2O Nitrous oxide gas NH3 Ammonia (NH4)6Mo7O24 Ammonium heptamolybdate NH4VO3 Ammonium metavanadat Ni Nickel Ni-Mo Nickel-molybdenum active metal Ni-Mo/TiO2 Catalyst: nickel-molybdenum active metal on Titanium oxide support Ni-V Nickel-vanadium active metal Ni-V/TiO2 Catalyst: nickel-vanadium active metal on Titanium oxide support Ni2P Nickel phosphide catalyst NiMoO4 Nickel molybdate Ni(NO3)2·6H2O Nickel nitrate hexahydrate NiO-CuO Oxide form of nickel copper catalyst NiO-Fe2O3 Oxide form of nickel iron catalyst NiO-MoO3 Oxide form of nickel molybdenum catalyst NiO-V2O5 Oxide form of nickel vanadium catalyst NOX Nitrogen oxides compound O2 Oxygen gas OH- Hydroxide bond (group) xxii Pd Nobel metal palladium Pt Nobel metal platinum Rh Nobel metal rhodium SO2 Sulphur dioxide SOX Sulphur oxides compound TiO2 Titanium oxide catalyst support V Vanadium V2O5 Vanadium pentoxide WP Tungsten (wolfram) phosphide catalyst ZrO2 Zirconium oxide catalyst support xxiii Abbreviations BET Brunauer-Emmett-Teller C Carbon element C1 Hydrocarbon with 1 atom carbon C8 Hydrocarbon with 8 atom carbon C20 Hydrocarbon with 20 atom carbon CCD Charge-coupled device CDME Catechol dimethyl ether CHNS Carbon, hydrogen, nitrogen and sulphur DTG Derivative thermogravimetric EDX Energy-dispersive X-ray ETR Extended temperature range FC Fixed carbon FID Flame ionization detector GC Gas chromatography GC-FID Gas chromatography with a flame ionization detector GC-MS Gas chromatography / mass spectroscopy GHG Greenhouse gas H Hydrogen element HDO Hydrodeoxygenation reaction HHV Higher heating value ICDD The International Centre for Diffraction Data xxiv IUPAC International Union of Pure and Applied Chemistry KF Karl-fischer N Nitrogen element NOX Nitrogen oxides compound NSERC Natural Science and Engineering Research Council of Canada O Oxygen element PDF Portable document format pH Acidity scale PSD Pore size distributions Ref Reference S Sulphur element SOX Sulphur oxides compound TEM Transmission electron microscopy TG Thermogravimetric TGA Thermogravimetric analysis or thermogravimetric analyzer TPD Thermo-programmed desorption TPD-NH3 Thermo-programmed desorption of ammonia XPS X-ray photoelectron spectroscopy XRD X-ray Diffractometry or X-ray powder diffraction UHV Ultra-high vacuum UHP Ultra-high purity FWHM Full width at half maximum xxv CHAPTER 1: INTRODUCTION 1.1 Overview Fossil fuels, which are categorized as a non-renewable energy, have been identified as a primary source of greenhouse gas (GHG) emissions in our atmosphere. Based on a report from Environment Canada in 2013, 79% of total Canada’s GHG emissions come from carbon dioxide (CO2), 12% from methane (CH4) and about 7% from nitrous oxide (N2O) [1]. It is estimated that total Canadian GHG emissions from fuel combustion reached 537 megatons (Mt) in 2010, or roughly 1.8% of the world’s total GHG emissions [1]. Canada has been recognized as one of the largest fossil fuels producer in the world and consumed high amount of electricity per capita. Between 1990 and 2010, energy use in Canada increased by 22.3% or about 1.16% per year [1]. Based on data from Statistics Canada, fossil fuel consumption in Canada in 2010 represented approximately 69% of its total energy consumption (8479.1 PJ), including hydro, nuclear, solar, wind and biomass sources of energy, where the industrial sector accounted for the biggest share of energy used (38%), followed by transportation (31%), residential (16%), commercial/institutional (12%), and agricultural (3%) sectors [1]. The production and use of fossil fuels have raised environmental concerns. As a result, the production of renewable energy seems to be an interesting alternative to fulfill energy demand, while lowering environmental impact. 1 Fig. 1.1 Typical lignocellulosic biomass waste found in Canada (from left to right: wheat straw, oat straw, flax straw and barley straw) Canada is a country endowed with abundant and varied natural resources, including massive amounts of energy from sources such as petroleum, water, solar, wind and biomass. In 2015, based on the data from Natural Resources of Canada, Canada produced about 25 million tonnes wheat which also produced about the same amount of straw biomass. In addition, biomass productivity from plantation vary between 10-15 oven dry tonnes per hectare per year. The use of biomass or waste biomass, such as wheat straw, flax straw, barley straw or sawdust (Figure 1.1) in energy production provides several advantages from an environmental perspective. For example, the plant growth to generate biomass consumes CO2 from the atmosphere, which offsets the CO2 production created by the use of biomass as energy. The conversion of waste biomass into energy and value-added chemicals will help meet the growing energy demand, combat the increase in food crop prices, partially redirect biomass waste away from landfills and simultaneously minimize environmental damage through reduced GHG emissions into the atmosphere. There are several methods known to convert biomass into energy. One such method is pyrolysis, which is a thermal decomposition process with very limited amount of oxygen. The process is typically carried out at a temperature range of 400 °C to 700 °C and at a near-ambient pressure 2 [2,3]. During pyrolysis, the large molecule structure of the biomass is broken down, producing chemicals with smaller molecular structures. The process typically produces a hydrocarbon-rich gas mixture, an oil-like liquid (bio-oil), and a carbon-rich solid residue (bio-char). The gas product contains syngas (CO and H2) and CH4 and can be burned directly to produce heat or electricity, while the bio-char can either be used as fertilizer or to produce activated carbon. The bio-oil product, however, requires a further upgrading and/or separation processes, in order to be used directly. Although bio-oil is considered to be the most important pyrolysis product, due to its potential use as a liquid fuel alternative, its properties have a downside. As shown in Table 1.1, bio-oil is highly oxygenated [3,4], leading to a high viscosity, boiling point corrosiveness, as well as relatively low chemical and thermal stability. Its composition is mainly determined by the type of feedstock/biomass and the parameters of the pyrolysis process, such as temperature, heating rate and residence time [5]. For example, pyrolysis using a slow heating rate tends to produce more solid product than liquid product [6]. To overcome this problem, bio-oil can be upgraded using several methods including catalytic cracking and steam reforming [7]. Table 1.1 Typical elemental and water content of bio-oils from fast pyrolysis [4,5] Component Wt% Carbon 32-49 Oxygen 44-60 Water 15-30 Hydrogen 6-8.6 Nitrogen 0-0.2 3 In spite of the large amount of experimental data on the pyrolysis of various biomasses, there is still considerable disagreement on the specific mechanisms of the pyrolysis process required for the production of certain products, such as bio-oil and bio-char. For example, the detailed mechanisms and reaction routes of slow pyrolysis and fast pyrolysis remain unclear. The best methods to produce higher quality bio-oil and to upgrade bio-oil into a gasoline-like compound are likewise undetermined. As a result, an investigation of the kinetics of pyrolysis, the effects of several parameters on the pyrolysis product and its upgrading are needed, in order to answer these questions. 1.2 Literature Review 1.2.1 Introduction to Biomass When wood or lignocellulosic biomass is exposed to high temperatures, with or without the presence of oxygen, changes occur in its chemical structure. Pyrolysis is a process involving high temperatures that converts biomass into products, such as gaseous material, bio-oil and bio-char in the absence of oxygen [7-10]. During pyrolysis, the biomass undergoes a very complex set of reactions involving the formation of radicals in order to produce gaseous material, bio-oil and bio-char [9,10]. The composition and quality of the pyrolysis products depend on certain parameters, such as the reaction temperature, heating rate and residence time of the process [810]. It also depends on the type of biomass or, to be more specific, on the composition of the biomass. In Canada, biomass is widely available in the form of forestry and agricultural waste, including wheat straw, flax straw, barley straw and wood/sawdust.They are categorized as lignocellulosic 4 biomasses. In general, lignocellulosic biomass is composed of extractive components (oil, protein, sugar and others), ash and the three polymers that construct the cell walls: cellulose, hemicelluloses and lignin linked together by ether and ester bonds [11,12]. Cellulose is a polysaccharide consisting of a linear chain of linked D-glucose units as represented by the chemical formula (C6H10O5)n and is responsible for the production of flammable volatiles during thermal degradation at a temperature range of 200 °C to 450 °C. Cellulose has a high degree of polymerization (~10,000) and a large molecular weight (~500,000), which make it highly insoluble. It is the main component of wood, representing up to about 40% of wood’s dry weight [11-13]. Hemicellulose is also a polysaccharide, but with a branched and shorter chain of glucose units. Unlike the cellulose crystalline structure, hemicellulose is amorphous, easier to hydrolyze, and degrades within a smaller temperature range (200 °C to 260 °C). Lignin is randomly linked, amorphous, and composed of many phenolic groups. The degradation of lignin usually occurs in the temperature range of 225 °C to 450 °C, thereby producing a significant amount of phenolic compounds, given its nature [12,13]. The amount of cellulose, hemicellulose and lignin of a biomass can provide useful information to predict the product yield of a biomass. Biomass can be categorized based on its polymer composition or its proximate and elemental composition. In proximate analysis, the biomass is exposed to a drying temperature to determine its moisture content. This temperature is gradually increased to 900 °C in a nitrogen (N 2) atmosphere to obtain the volatile amount from its decomposition. The fixed carbon (FC) content 5 can also be determined through proximate analysis by replacing the N2 with oxygen (O2) to allow combustion, which produces ash as the resulting material. Elemental composition information, such as carbon, hydrogen, nitrogen, sulphur and oxygen (C, H, N, S and O) can be used to estimate the heating value of the biomass and to predict the quality of bio-oil that can be obtained through the pyrolysis process. In the elemental (i.e. ultimate) analysis, the biomass sample undergoes a complete combustion around 1000 °C and is then mixed, separated in the column and detected as an individual gas using a thermal conductivity detector. Table 1.2 shows the elemental composition of several types of biomass gathered from various sources. As can be seen, the volatile content of a biomass ranges from 75% to 86% of its total weight and contains as much as 47% oxygen. Table 1.2 Chemical composition of several types of biomass Species Pine wood Moisture Fir wood Beech wood 6.40 Oat straw 5.38 Flax straw 4.06 Rapeseed Wheat straw Rice straw Coconut shell Sugarcane bagasse Volatiles 16.81 80.49 19.42 74.80 18.30 74.48 77.57 75.84 4.00 86.04 5.58 79.82 13.61 11.26 16.07 FC 82.79 82.88 Spruce wood Pine bark Combustible 76.85 85.36 79.59 Ash C H 0.40 51.22 0.24 51.24 0.50 49.80 19.53 5.99 42.27 5.98 21.30 1.11 45.24 63.45 17.00 0.12 21.61 2.55 8.23 5.73 14.45 5.73 O [14] 42.68 [14] 42.70 [15] 45.20 [17] 0.15 43.77 53.21 5.98 0.52 43.48 52.25 8.06 3.91 35.78 5.84 0.18 40.78 4.34 58.14 5.95 6.20 0.12 0.40 Ref 42.64 6.07 50.93 6 0.12 S 50.01 9.54 3.38 6.02 N [14] [14] [16] 0.56 1.07 6.25 1.01 1.18 46.40 [17] 6.73 0.43 0.17 28.27 [18] 6.04 6.05 0.83 0.69 0.28 0.23 0.19 47.50 41.61 34.57 [17] [18] [18] 1.2.2 Pyrolysis of Lignocellulosic Biomass Pyrolysis is used to convert biomass to gas, liquid and solid products by heating it in the absence of oxygen. During pyrolysis, the biomass undergoes a very complex set of reactions involving the formation of radicals in order to produce gaseous material, bio-oil and bio-char [9,19]. In general, pyrolysis consists of two stages of devolatilization after the drying process. Primary devolatilization (less than 400 C) mainly devolatilizes hemicelluloses and celluloses, whereas secondary devolatilization (over 400 C) involves aromatization of components such as lignin [20]. Although pyrolysis is an old technology, it has been improved in recent decades and is becoming more attractive than many other thermochemical processes in order to produce liquid fuel and chemicals [21]. Pyrolysis involves many chemical reactions (cracking, elimination, addition, substitution and polymerization) and produces a large number of chemical components. Cracking will produce a smaller chain of hydrocarbons (ranging from C1 to C8), while polymerization will produce a longer chain of hydrocarbons (ranging up to C20). Since biomass contains around 45% oxygen, most of the pyrolysis products are phenolic and poly-alcohol compounds. 7 Fig. 1.2 Devolatilization profile of biomass components [22] Figure 1.2 shows the devolatilization profile of biomass components as the temperature increases. As can be seen, xylan and cellulose are more likely to decompose at lower and smaller ranges of temperature (200 °C to 400 °C) as compared to lignin (150 °C to 500 °C). Since most biomass consists of about 10-40% lignin, performing pyrolysis at a temperature of around 500 °C is more favorable in terms of a high yield from the decomposition of cellulose, hemicellulose and lignin. 1.2.2.1 Slow (Conventional) Pyrolysis Based on its heating rate, pyrolysis can be divided into slow pyrolysis and fast pyrolysis. Pyrolysis with a vapor residence time of 5-30 min is considered slow pyrolysis [3,13,23]. During slow pyrolysis, the intermediate end product can undergo a secondary reaction, forming a longer chain component that can lower the quality of bio-oil. Longer residence time also requires more 8 energy to process. Despite these disadvantages, slow pyrolysis may be easier to control and can produce more bio-char product, which can be suitable for soil enrichment. Nonetheless, the oil yield obtained from slow pyrolysis (25-50%) is lower than that from fast pyrolysis and may have a different chemical composition [3,13]. 1.2.2.1 Fast Pyrolysis In the fast pyrolysis process, the heating rate can be as high as 1000 °C/min. The typical residence time for fast pyrolysis is about 1-5 seconds with a reaction temperature range from 450 °C to 1000 °C [3,4]. The intent of fast pyrolysis is to limit the residence time of the liquid product in order to prevent a secondary reaction that can lead to the formation of a longer chain of hydrocarbon or a tar component. In addition to a fast heating rate and better control of reaction temperature, another key to prevent a secondary reaction is the use of rapid cooling. Using fast pyrolysis, a higher yield of liquid product can be obtained representing up to 75% of the weight of the biomass [3,23]. The general changes that occur during pyrolysis are: 1. Heat transfer from a heat source to increase the temperature of the solid feedstock; 2. Initiation of primary pyrolysis reactions at this higher temperature releases volatiles and forms char; 3. The flow of hot volatiles toward cooler solids results in heat transfer between hot volatiles and cooler unpyrolyzed feedstock; 4. Condensation of some of the volatiles in the cooler parts of the fuel, followed by secondary reactions, can produce tar; 9 5. Autocatalytic secondary pyrolysis reactions proceed, while primary pyrolytic reactions (item 2, above) occur simultaneously in competition; and, 6. Further thermal decomposition, reforming, water gas shift reactions, radical recombination and dehydrations can also occur, all of which are functions of the process residence time. 1.2.3 Kinetic Study of Pyrolysis of Biomass Many techniques can be used to study biomass pyrolysis to obtain kinetic parameters and to gain insight into the devolatilization process of biomass pyrolysis. One technique that has been widely used in the past is thermogravimetric analysis (TGA), which is a technique in which the mass of the biomass is measured as a function of temperature or time while the sample is subjected to a controlled atmosphere [24]. The mechanism of biomass thermal devolatilization, however, can vary depending on the components and composition of the biomass. TGA is a useful tool to study biomass devolatilization during pyrolysis. It provides data about biomass conversion as the temperature changes. The measurement of mass loss using TGA can yield information such as the thermal stability and kinetic parameters of the thermal events during pyrolysis. This kinetic information is crucial for the design of pyrolysis reactors and gasifiers [25,26]. Three stages can be distinguished during the heating process: drying, primary devolatilization and secondary devolatilization [27]. The kinetics of these stages can be determined by applying the Arrhenius equation to the separate slopes of constant mass devolatilization [28]. In addition, a nonlinear approach can be used to optimize the parameters. 10 1.2.4 Pyrolysis Behaviors of Lignocellulosic Biomass In the past few decades, many researchers have studied the pyrolysis behaviors of various biomasses and their kinetics. Caballero et al. investigated the pyrolysis kinetics of almond shells and olive pits and noted that the thermal devolatilization of lignocellulosic material can be differentiated into two global processes: devolatilization of hemicelluloses together with lignin, and cellulose devolatilization [29]. Researchers have also studied the influence of the heating rate on the pyrolysis of various biomasses. Altun et al. investigated the effect of particle size and heating rate on the pyrolysis of Silopi asphaltite [30]. They found that, by increasing the heating rate, the residue (bio-char and ash) at the end of pyrolysis also increased. In another experiment, Lapuerta et al. performed a non-isothermal pyrolysis of forestry waste using varying heating rates and showed that a curve fitting method corresponded well with the experimental data to reflect the kinetic parameters [31]. Kumar et al. studied the devolatilization of corn stover and concluded that the primary devolatilization was followed the kinetics of first-order reaction rate [32]. Chouchene et al. observed the effect of particle size on the pyrolysis of olive solid waste under nitrogen and oxidative atmospheres [33]. They found that samples smaller than 0.5 mm were the most reactive. They also found that, during oxidative pyrolysis under 10% oxygen (O 2), olive solid waste samples with diameters between 2 mm and 2.8 mm produced lower ash content and higher char quantity. 11 In another study, Lester et al. examined the pyrolysis of olive solid waste using three different particle sizes (53-75 mm, 106-150 m, and 150-212 mm) mixed with standard highly volatile coal with particle sizes in the range of 106-150 mm under nitrogen conditions [34]. They found that the particle size variations had a limited effect on the weight loss profile under the above conditions. Although many studies about biomass pyrolysis using thermogravimetric analysis have been reported, the effects of important parameters, such as particle size, heating rate and initial sample weight, have rarely been investigated for biomasses such as sawdust, oat straw or barley straw. This information will be vital in the design of reactors for larger scale biomass pyrolysis and in establishing global kinetic parameters for the devolatilization of lignocellulosic biomass. 1.2.5 Bio-Oil Production from Lignocellulosic Biomass Lignocellulosic biomass can be converted to bio-oil through pyrolysis followed by catalytic upgrading [3]. As previously mentioned, pyrolysis can be categorized as conventional pyrolysis (slow pyrolysis) and fast pyrolysis, depending on the operating conditions. In slow pyrolysis, biomass is heated up to 500 °C at a heating rate of less than 100 °C/min and a vapor residence time of between 5 and 30 min, resulting mainly in char production [35]. In fast pyrolysis, biomass is heated up to 500 °C, but at a faster heating rate (1000 °C/min and higher) and a vapor residence time of less than 2 seconds. Fast pyrolysis naturally favors bio-oil as the main product [36], unlike slow pyrolysis, which results in the production of bio-char. 12 Several chemical changes occur during pyrolysis. The primary pyrolysis reaction releases volatiles and concurrently forms char. During the next step, condensation can occur in the cooler parts of the system to form bio-oil. This process is naturally followed by secondary reactions that produce tar. Reforming, dehydration and polymerization can also occur, depending on the reaction time and the heating rate [3]. In the end, the volatile pyrolysis product can have more than 120 components, ranging from unspecified low molecular products contain phenolic OHgroups to cyclic and bi-cyclic products [37]. Current and ongoing research in pyrolysis is aimed at increasing the yield of bio-oil while increasing the quality of the oil through subsequent upgrading processes. This study researches the impact of pre-treating the sample prior to the pyrolysis process in order to obtain a better quantity and quality of bio-oil. 1.2.6 Bio-Oil Upgrading As observed in Table 1.3, bio-oil has interesting properties. It consists mostly of carbon (44-55 wt%) and oxygen (40-50 wt%). One of the advantages of bio-oil is its high carbon content, since this increases the high heating value of its combustion. Another advantage is its lower nitrogen and sulphur content, resulting in a lower amount of sulphur oxides (SOx) and nitrogen oxides (NOx) in the flue gas stream [29]. Bio-oil properties do, however, have some drawbacks: one of the main disadvantages is the poor stability of the oil due to its high oxygen content. Generally, the bio-oil from pyrolysis has a high oxygen (40-50 wt%) and water (20-50%) content [29,30]. High oxygen content in the oil causes condensation, which leads to high viscosity (Table 13 1.3). Consequently, bio-oil possesses a lower heating value compared to petroleum fuels, such as gasoline (45 MJ/kg), thereby demanding rigorous upgrading processes. Table 1.3 Typical properties of oils obtained through flash pyrolysis Parameter Value Yield (wt%) 50-75 Carbon (wt%) 44-55 Oxygen (wt%) 40-50 Hydrogen (wt%) 6-7 Nitrogen (wt%) Water (wt%) 0.0-0.2 20-30 Density (g/cm ) 1.2 3 Viscosity (cP. At 40 °C) Higher Heating Value (HHV, MJ/kg) 40-200 20 Upgrading is frequently carried out using common hydrotreating and synthetic zeolite catalysts [29,38,39]. The catalysts used in the upgrading process are selected based on their pore size, acidity and molecular sieving properties, in order to obtain maximum hydrocarbon yields [39,40]. Other than that hydrodeoxygenation (HDO) process can also be used in order to reduce oxygen content to produce fine chemicals such as phenol [40-43]. When using zeolites, although no hydrogen is required for the process, their capability for deoxygenation of lignocellulose phenolic compounds is limited due to poor hydrocarbon yield and easy decomposition by extensive coke deposits or dealumination by hydrolysis [44]. For better result, oxide metal catalyst is mixed with bio-oil and H2 gas is used as reaction gas during HDO [41,42]. Zhang et al. (2015) perform a fast pyrolysis using magnetic solid base catalyst K3PO4/Fe3O4 and successfully increased the yield of phenolic component about 1.5 times compared to non14 catalytic pyrolysis. They conclude that the catalyst would promote the decomposition of lignin to increase the phenolic yield [41]. 1.3 Objectives The scope of this research was to maximize the yield of bio-oil from the pyrolysis of various straw biomasses and to synthesize and utilize a catalyst during the post-treatment phase to increase the quality of the bio-oil. In particular, the study focused on the role of the catalyst in the upgrading process to enhance bio-oil properties to produce gasoline-like biofuel or valueadded chemicals. The hypothesis of this study was that the use of catalyst in the pyrolysis of the biomass would increase the yield and enhance the quality of the bio-oil produced by the pyrolysis and, at later stage, would also be effective for the upgrading of the bio-oil into value-added chemicals. This study investigated the kinetics of the pyrolysis process, the synthesis of the metal catalyst and the catalytic upgrading process of selected bio-oil components. In order to test the hypothesis, the following specific objectives were set: 1. Determine the mechanism and global kinetics of pyrolysis of lignocellulosic biomass under different parameters, such as heating rate, particle size and initial weight of the sample (Chapter 3). 2. Investigate the use of heterogeneous conventional solid acid catalysts (e.g. zeolite based catalysts) to enhance the yield and quality of bio-oil and to characterize the pyrolysis products (Chapter 4). 15 3. Study the application of oil hydrogenation catalysts (nickel, Ni, nickel-molybdenum, NiMo, nickel-vanadium, Ni-V, supported over alumina) in the presence of hydrogen on the guaiacol (bio-oil main component) upgrading process (Chapter 5). 4. Study the application and kinetics of oil hydrotreating and hydrogenation catalysts (Ni, Ni-Mo, Ni-V supported over alumina) in the presence of hydrogen on the anisole (bio-oil components) upgrading processes (Chapter 6). 1.4 Organization of the Thesis This thesis consist of seven chapters, organized consecutively in a paper-based format. Chapters 3 to 6 was written in a paper-based format and 2 of them have been published and the other 2 have been submitted to peer-reviewed journal paper. Brief description of each chapter is explained below. In Chapter 1, the author provides the overall background of the research project. A condensed literature review was written to describe few important definitions, finding and progress in the area of pyrolysis of biomass. In addition, the author also wrote the objective and the outcome of the research project in term of publications. Chapter 2 explains the research methodology that has been used by the author including the study of pyrolysis kinetics using TGA, the bench-scale reaction of pyrolysis and also the biooil upgrading process. In this chapter, the author present the experimental setups constructed based on the information from literature review. Chapter 3 presents the first published paper about the pyrolysis kinetics of pyrolysis of sawdust. In this publication, the author showed the methodology to obtain the kinetic parameters such as reaction order and the energy activation. The effect of several parameters 16 such heating rate, the initial weight and particle size was also discussed to gain the optimum condition for bench-scale pyrolysis process. In Chapter 4, the author wrote about the result from bench-scale pyrolysis of four different straw biomass (wheat, oat, flax and barley straw) and analyze the yield of the products including the bio-char, bio-oil and gaseous material. TGA instrument was also used to gather important information such as proximate analysis that can be used as an early prediction of the product yield that can be obtained during bench-scale process. Chapter 5 explains the catalyst synthesis, characterization and upgrading process of one of the main component of bio-oil, guaiacol. In this study, catalyst screening was performed under hydrodeoxygenation reaction using temperature range between 200-350 °C. Among various catalysts, Ni-Mo/TiO2 (prepared by impregnation technique) was chosen to understand the influence of reaction parameters, such as temperature, reaction time, H 2 pressure, and quantity of the catalyst. Chapter 6 discusses about upgrading process of other component of bio-oil, anisole. The reaction study was done on hydrodeoxygenation reaction of anisole using the same catalyst that previously used for guaiacol. The comparison between the upgrading of both component was shown in order to see the consistency of the process in reducing the oxygen content of bio-oil components. In addition, determination of kinetic parameter was also provided in this chapter. Chapter 7 provides the general conclusion that can be drawn from the result of each phase of the study. The author also offers few suggestions and future that can should be done in order to gather other important information that will be useful for the commercialization of the research project. 17 1.5 Outcome of the Thesis The author has published 2 peer-reviewed journal papers (chapter 1 and chapter 3), 3 peerreviewed international conference proceedings, and submitted a third (chapter 2) and fourth manuscript (chapter 4) as the first author. In addition, the author has published 3 peer-reviewed journal paper as co-author for related work in pyrolysis and torrefaction and currently submitted one manuscript as co-author in a work related to pyrolysis of municipal solid waste. A list of these publications follows: 1.5.1 Published peer-reviewed journal papers and contribution: 1. Aqsha, A., Mahinpey, N., Mani, T., Salak, F. & Murugan, P., Study of sawdust pyrolysis and its devolatilisation kinetics, The Canadian Journal of Chemical Engineering, 89(6), pp. 1451-1457, 2011. 2. Vincent, S.S., Mahinpey, N. & Aqsha. A., Mass transfer studies during CO2 gasification of torrefied and pyrolyzed chars, Energy, 67, pp. 319-327, 2014. 3. Aqsha, A., Katta, L. & Mahinpey, N., Catalytic Hydrodeoxygenation of Guaiacol as Lignin Model Component Using Ni-Mo/TiO2 and Ni-V/TiO2 Catalysts, Catalysis Letters, 145 (6), pp. 1351-1363, 2015. 4. Shi, H., Mahinpey, N., Aqsha, A., Silbermann, R., Characterization, thermochemical conversion studies, and heating value modeling of municipal solid waste, Waste Management, 48, pp. 34-47, 2016. 18 5. Tijani, M.M., Aqsha, A. & Mahinpey, N., Development of Oil-Spill Sorbent from Straw Biomass Waste: Experiments and Modeling Studies, Journal of Environmental Management, in press, doi:10.1016/j.jenvman.2016. 02.010, 2016. Candidate’s and other author’s contribution to Paper 1: The candidate designed the study with help from Mani, Salak and Murugan. The candidate searched the literature, selected the relevant articles, performing the experiment. The candidate analyzed and interpreted the data and prepared the manuscript with help from Mani and Murugan while Mahinpey review and critically revised the manuscript for important intellectual content. This study was completed under the guidance of the postdocs and supervisors. All authors critically revised the manuscripts and contributed intellectually to each work. Candidate’s and other author’s contribution to Paper 2: The candidate contribute in the design of the experiment and conducting the analysis of char product. The candidate prepared experimental setup and helping in the purchasing process of most of the apparatus. The candidate also involved in data analysis, derived the mass transfer model, and the review of the manuscript. All authors critically revised the manuscripts and contributed intellectually to each work. Candidate’s and other author’s contribution to Paper 3: The candidate designed the research and conducted the study with help from Katta and Tijani. Katta has trained the candidate in the catalyst synthesis and preparation and provided her thought on the type of analysis that were performed during the study. Tijani involved in helping the 19 candidate to setup the equipment and assist in the pyrolysis experiment. Candidate and Katta worked together in writing the manuscript and discussing the result. Katta provided her expertise in analyzing the result from charatcterization of the catalyst, while Tijani contributed in extracted TGA data and CHNS data. This study was completed under the guidance of the postdoc (Katta) and supervisors. All authors critically revised the manuscripts and contributed intellectually to each work. Candidate’s and other author’s contribution to Paper 4: The candidate was involved in the proposal stage of the study that was submitted to Alberta Innovates – Energy and Environment Solutions. The candidate contribute in the design of the experiment and conducting the sampling from a waste management facility (landfill) in Red Deer during summer 2013. The candidate also provide his help for data analysis and the review of the manuscript. All authors critically revised the manuscripts and contributed intellectually to each work. Candidate’s and other author’s contribution to Paper 5: The candidate was involve in the writing of NSERC-Engage proposal with ECCO Recycling Inc as industry partner. The candidate contribute in the design of the experiment and prepare the sample that was obtain from ECCO Recycling facility. The candidate provided the training of the instruments (CHNS, TGA and BET) to Tijani and also involved in data analysis related to absorbency value. All authors critically revised the manuscripts and contributed intellectually to each work. 20 1.5.2 Submitted peer-reviewed journal papers and contribution: 1. Aqsha, A., Tijani, M.M. & Mahinpey, N., Catalytic Pyrolysis of Straw Biomasses (Wheat, Flax, Oat and Barley Straw) and the Comparison of Their Product Yields, Energy Conversion and Management, submitted, 2016. 2. Aqsha, A., Katta, L., Tijani, M.M. & Mahinpey, N., Upgrading of Anisole Components by Ni-Mo/TiO2 and Ni-V/TiO2 Catalysts: Synthesis, Characterization and Kinetic Measurements, Energy & Fuel, submitted, 2016. 3. Shi, H., Mahinpey, N., Aqsha, A., Silbermann, R., Production of Value Added Chemicals from Waste Coffee Cups via Catalytic Pyrolysis, Catalysis Communications, submitted. Candidate’s and other author’s contribution to Paper 1: The candidate designed the study and prepare the experimental setup. The candidate searched the literature and performing the entire experiment with the help of Tijani. The candidate analyzed and interpreted the data and prepared the manuscript. Tijani was involved in data extraction related to bench-scale pyrolysis experiment and GC-MS analysis. This study was completed under the guidance of the postdocs and supervisors. All authors critically revised the manuscripts and contributed intellectually to each work. Candidate’s and other author’s contribution to Paper 2: The candidate designed the research and conducted the study with help from Katta and Tijani. Katta was involved in the early stage of the study, preparing the catalyst and analysing the XPS data obtained from Dr. Pereira Lab. Tijani was involved in data extraction from TEM instrument 21 that was located in foothills hospital. This study was completed under the guidance of the postdoc (Katta) and supervisors. All authors critically revised the manuscripts and contributed intellectually to each work. Candidate’s and other author’s contribution to Paper 3: The candidate was involved in the proposal stage of the study that was submitted to Alberta Innovates – Energy and Environment Solutions. The candidate contribute in the design of the experiment and conducting the sampling from a waste management facility (landfill) in Red Deer during summer 2013. The candidate also provide his help for data analysis and the review of the manuscript. All authors critically revised the manuscripts and contributed intellectually to each work. 1.5.3 Published peer-reviewed conference proceedings and contribution: 1. Aqsha, A., Kannaiyan, R., Mahinpey, N. & Martinuzzi R., Application of Subcritical Water on Bioenergy Conversion, 10th International Symposium on Supercritical Fluids (ISSF), Proceeding, San Fransisco, 2012 2. Aqsha, A., Tijani, M.M. & Mahinpey, N., Catalytic Pyrolysis of Straw Biomasses (Wheat, Flax, Oat And Barley Straw) And The Comparison Of Their Product Yields, Energy Production and Management in the 21st Century, WIT Press, pp. 1007-1016, 2014. 3. Aqsha, A., Mahinpey, N., Katta, L., Gras, L. & Lim, C.J., Synthesis of Novel Catalysts For Hydrodeoxygenation Of Bio-oil: Guaiacol As A Model Component, Energy & Sustainability V, WIT Press, pp. 489-498, 2014. Candidate’s and other author’s contribution to Paper 1: 22 The candidate designed the study and prepare the experimental setup of the glycerol decomposition using subcritical water experiment. The candidate searched the literature and performing the experiment related to glycerol decomposition. Kannaiyan was performing most of the experiment related to lignin decomposition and fungal pretreatment. The candidate and Kannaiyan analyzed and interpreted the data and prepared the manuscript. The candidate present the manuscript in the ISSF conference and collecting the feedback from the audience. This study was completed under the guidance of the both supervisors Dr. Mahinpey and Dr. Martinuzzi. All authors critically revised the manuscripts and contributed intellectually to each work. Candidate’s and other author’s contribution to Paper 2: The candidate designed the study and prepare the experimental setup. The candidate searched the literature and performing the entire experiment with the help of Tijani. The candidate analyzed and interpreted the data and prepared the manuscript. Tijani was involved in data extraction related to bench-scale pyrolysis experiment and GC-MS analysis. This study was completed under the guidance of the postdocs and supervisors. All authors critically revised the manuscripts and contributed intellectually to each work. Candidate’s and other author’s contribution to Paper 3: The candidate designed the research and conducted the experiment with help from Katta. Katta has trained the candidate in the catalyst synthesis and preparation and provided her thought on the type of analysis that were performed during the study. Candidate and Katta worked together 23 in writing the manuscript and discussing the result. Katta provided her expertise in analyzing the result from characterization of the catalyst. Gras was an internship student who assist on samople preparation during HDO reaction experiment. This study was completed under the guidance of the postdoc (Katta) and supervisors. All authors critically revised the manuscripts and contributed intellectually to each work. 24 1.6 References [1] Natural Resource of Canada, Energy Efficiency and Trends in Canada 1990 to 2010, pp. 142, 2011. [2] Demirbas, A., Ozturk, T. & Demirbas, M.F., Recovery of energy and chemicals from carbonaceous materials, Energy Sources Part A, 28, pp. 1473-1482, 2006. [3] Mohan, D., Pittman C.U. & Steele, P.H., Pyrolysis of wood/biomass for bio-oil: a critical review, Energy Fuels, 20, pp. 848-889, 2006. [4] Bridgwater, A.V., Meier, D. & Radlein, D., An overview of fast pyrolysis of biomass, Org. Geochem., 30, pp. 1479-1493, 1999. [5] Oasmaa, A. & Czernik, S., Fuel oil quality of biomass pyrolysis oils - state of the art for the end user, Energy Fuels, 13, pp. 914-921, 1999. [6] Garcia-Perez, M., Wang, X.S., Shen, J., Rhodes, M.J., Tian, F.J. & Lee, W.J., Fast pyrolysis of oil mallee woody biomass: effect of temperature on the yield and quality of pyrolysis products, Ind. Eng. Chem. Res., 47, pp. 1846-1854, 2008. [7] Zhang, Q., Chang, J., Wang, T. & Xu, Y., Upgrading Bio-oil over Different Solid Catalysts, Energy Fuels, 20, pp. 2717-2720, 2006. [8] Gupta, R.B. & Demirbas, A., Gasoline, Diesel and Ethanol Biofuels from Grasses and Plants, Cambridge University Press, pp. 140-157, 2010. [9] Yaman, S., Pyrolysis of biomass to produce fuels and chemical feedstocks. Energy Convers. Manag., 45, pp. 651-671, 2004. [10] Schniewind, A.P., Concise encyclopedia of wood and wood based materials. Pergamon Press, pp. 271-273. 1989. 25 [11] Higuchi, T., Biochemistry and molecular biology of wood, Springer, Berlin Heidelberg, New York, 1997. [12] Shafizadeh, F., Introduction to pyrolysis of biomass, J. Anal. Appl. Pyrol., 3, pp. 283-305, 1982. [13] Basu, P., Biomass Gasification and Pyrolysis Practical Design and Theory, Academic Press, Elsevier, pp. 33-38, 2010. [14] Peng, J.H., A Study of Softwood Torrefaction and Densification for Production of High Quality Wood Pellets, PhD Thesis of University of British Columbia, 2012. [15] Demirbas, A., Pyrolysis of ground beech wood in irregular heating rate conditions, J. of Anal. and App. Pyrolysis, 73(1), pp. 39-43, 2005. [16] Sensoz, S., Angin, D., Yorgun, S. and Kockar, O.M., Biooil Production from an Oilseed Crop: Fixed-Bed Pyrolysis of Rapeseed (Brassica napus L.), Energy Sources, 22(10), pp. 891-899, 2000. [17] Mani, T., Murugan, P. and Mahinpey, N., Pyrolysis of Oat Straw and the Comparison of the Product, Energy & Fuels, 25, pp. 2803-2807, 2011. [18] Tsai, W.T., Lee, M.K. and Chang, Y.M., Fast pyrolysis of rice straw, sugarcane bagasse and coconut shell in an induction-heating reactor, J. of Anal. and App. Pyrolysis, 76(1-2), pp. 230-237, 2006. [19] Mahinpey, N., Murugan, P., Mani, T. & Raina, R., Analysis of bio-oil, bio-gas and bio-char from pressurized pyrolysis of wheat straw using a tubular reactor. Energy Fuels, 23, pp. 2736-2742, 2009. [20] Fisher, T., Hajaligol, M., Waymack, B. & Kellogg, D., Pyrolysis behaviour and kinetics of biomass derived materials. J. Anal. Appl. Pyrolysis, 62, pp. 331-349, 2002. 26 [21] Scott, D.S. & Piskorz, J., The continuous flash pyrolysis of biomass, Can. J. Chem. Eng., 62, pp. 404-412, 1984. [22] Shafizadeh, F., Chemical Composition and Thermal Analysis of Cottonwood, Carbohydrate Res., 16, pp. 273-277, 1971. [23] Jahirul., M.I., Rasul, M.G., Chowdhury, A.A. & Ashwath, N., Biofuels Production through Biomass Pyrolysis - a Technological Review, Energies, 5, pp. 4952-5001, 2012. [24] Menczel, J.D. and Prime, R.B., Thermal analysis of polymers fundamental and applications. John Wiley & Sons: New Jersey, 2009. [25] Dirion, J.L., Reverte, C. and Cabassud, M., Kinetic parameter estimation from TGA: optimal design of TGA experiments. Chem. Eng. Res. Des., 86, pp. 618-625, 2008 [26] Kojima, T. Assavadakorn, P. and Furusawa, T., Measurement and evaluation of gasification kinetics of sawdust char with steam in an experimental fluidized bed. Fuel Process. Technol., 36, pp. 201-207, 1993. [27] Mani, T., Murugan, P., Abedi, J. and Mahinpey, N., Pyrolysis of wheat straw in a thermogravimetric analyzer: Effect of particle size and heating rate on devolatilization and estimation of global kinetics. Chem. Eng. Res., 478, pp. 1-7, 2010. [28] Gang, W. and Aimin, L., Thermal devolatilization and kinetics of mixtures of polylactic acid and Biomass during co-pyrolysis. Chin. J. Chem. Eng., 16(6), pp. 929-933, 2008. [29] Caballero, J.A., Conesa, J.A., Font, R. and Marcilla, A., Pyrolysis Kinetics of Almond Shells and Olive Stones Considering Their Organic Fractions, J. Anal. Appl. Pyrolysis, 42, pp. 159-175, 1997. [30] Altun, N.E., Hicyilmaz, C. and Kök, M.V., Effect of particle size and heating rate on the pyrolysis of Silopi asphaltite. J. Anal. Appl. Pyrolysis, 67, pp. 369-379, 2003. 27 [31] Lapuerta, M., Hernandez, J.J. and Rodriguez, J., Kinetics of devolatilization of forestry wastes from thermogravimetric analysis. Biomass Bioenergy, 27, pp. 385 - 391, 2004. [32] Kumar, A., Wang, L., Dzenis, Y.A., Jones, D.D. and Hanna, M.A., Thermogravimetric characterization of corn stover as gasification and pyrolysis feedstock. Biomass Bioenergy, 32, pp. 460-467, 2008. [33] Chouchene, A., Jeguirim, M., Khiari, B., Zagrouba, F. and Trouvé, G., Thermal devolatilization of olive solid waste: influence of particle size and oxygen concentration. Resour. Conserv. Recy., 54, pp. 271-277, 2010. [34] Lester, E., Gong, M. and Thompson, A., A method for apportionment in biomass/coal using thermogravimetric analysis. J. Anal. Appl. Pyrolysis, 80, pp. 111-7, 2007. [35] Bridgwater, A.V., Catalysis in thermal biomass conversion. Appl. Catal. A, 116, pp. 5-47, 1994. [36] Bridgwater, A.V., Renewable fuels and chemicals by thermal processing of biomass. Chem. Eng. J., 91, pp. 87-102, 2003. [37] Meier, D. & Faix, O., State of the art of applied fast pyrolysis of lignocellulosic materials a review, Biores. Tech., 68, pp. 71-77, 1999. [38] Adjaye, J.D. & Bakhshi, N.N., Production of hydrocarbons by catalytic upgrading of a fast pyrolysis bio-oil. Part I: Conversion over various catalysts, Fuel Process. Technol., 45, pp. 161-183, 1995. [39] Wang, H., Male, J. & Wang, Y., Recent Advances in Hydrotreating of Pyrolysis Bio-Oil and Its Oxygen-Containing Model Compounds, ACS Catalysis, 3(5), pp. 1047-1070, 2013. [40] Mu, W., Ben, H. X., Ragauskas, A. & Deng, Y. L., Lignin pyrolysis components and upgrading – technology review, Bioenergy Res., 6, pp. 1183-1204, 2013. 28 [41] Zhang, Z., Lu, Q., Ye, X., Li, W., Hu, B. & Dong, C., Production of phenolic-rich bio-oil from catalytic fast pyrolysis of biomass using magnetic solid based catalyst, Energy Convers.. & Manage., 106, pp. 1309-1317, 2015. [42] Kim, J. S., Production, separation and applications of phenolic-rich bio-oil – a review, Bioresour. Technol., 178, pp. 90-98, 2015. [43] Effendi, A., Gerhauser, H. & Bridgwater, A.V., Production of renewable phenolic resins by thermochemical conversion of biomass: a review, Renew. Sust. Energy Rev., 12, pp. 20922116, 2008. [44] Huber, G. W., Iborra, S. & Corma, A., Synthesis of transportation fuels from biomass: chemistry, catalysts, and engineering, Chem. Rev., 106(9), pp. 4044-4098, 2006. 29 CHAPTER 2: RESEARCH METHODOLOGY AND SETUP This research was divided into four different phases, as depicted in Figure 2.1. Phase 1 included the kinetic study of biomass pyrolysis using thermogravimetric analysis (TGA), while phase 2 consisted of the study of bio-oil production using bench-scale pyrolysis. Phase 3 involved catalyst synthesis and catalyst screening of the bio-oil component, and phase 4 encompassed upgrading the main components of the bio-oil, guaiacol and anisole, in order to reduce their oxygen content and transform the components into value-added chemicals. In this chapter, the pre-treatment method, pyrolysis experimental setup and bio-oil upgrading experimental scheme are explained in detail. Chemicals Fig. 2.1 Diagram of the experimental setup 30 2.1 Study of Biomass Pyrolysis and its Devolatilization Kinetics Using TGA Sawdust pyrolysis experiments were performed in the first phase of this project, subject to changes in various parameters including particle size, initial sample weight and heating rate. The devolatilization kinetics of biomass was studied using nonlinear optimization. Reaction kinetic parameters were obtained for the three different stages of pyrolysis. Samples used in this study were typical Canadian biomasses, such as sawdust (mix of wood), wheat straw, flax straw, barley straw and oat straw. For experimental purposes, the biomasses were ground in the laboratory mill and then sieved into different average particle sizes, varying from 75 to 1400 µm. In the first phase, the sawdust sample was analyzed using an elemental analyzer and then subjected to proximate analysis and the pyrolysis process using TGA (Figure 2.2). The TGA instrument was also used in every phase of the study in order to perform proximate analysis and pyrolysis to learn about the parameters affecting process product yields. Fig. 2.2 Schematic of pyrolysis experiment using TGA 31 2.1.1 Ultimate (Elemental Analysis) using Elemental Analyzer Ultimate analyses were performed using an elemental analyzer (2400 Elemental Analyzer, Perkin Elmer) in order to determine the carbon, hydrogen, nitrogen and sulphur (C, H, N and S) compositions of the biomass samples. The 2400 Elemental Analyzer consisted of four major zones, including the combustion, gas control, separation and detection zones (Fig. 2.3). The determination was based on the Pregl-Dhumas method, where samples were combusted (975 °C) in a pure oxygen environment and later reduced (500 °C) for the analysis of carbon dioxide (CO2), steam (H2O), nitrogen (N2) and sulphur dioxide (SO2) [1]. After combustion, the gases were captured in the mixing chamber, separated using the chromatography technique and later detected by a thermal conductivity detector. Prior to measurement, the instrument was prepared (filling the combustion tube, start-up, leak test) and calibrated using standard chemical cystine (29.99% C, 5.03% H, 11.66% N, 26.69% S). Fig. 2.3 Schematic of Perkin Elmer 2400 Elemental Analyzer [1] 32 The measurements were done using small sample quantities (about 2 mg). The samples were accurately weighed and then wrapped using small tin capsules purchased from the vendor. Prior to biomass analysis, few blank capsules were analyze to determine the instrument factor (kfactor) needed for the final calculation. A standard sample Cystine (C 6H12N2O4S2) was also measured and the result was compared to its theoretical C, H, N and S content. The measurements were repeated two to three times to obtain the average accurate value of 2 decimal points. 2.1.2 Proximate Analysis using TGA A proximate analysis of the sample was carried out using a TGA (STA6000, Perkin Elmer). Ceramic crucibles were used in order to minimize any thermal lag and to optimize heat transfer between thermocouples and crucibles. The instrument was equipped with a built-in mass flow controller in order to use certain flow rates of inert and reaction gas. In order to calibrate the instrument, a run using calcium oxalate (Fig. 2.3) was performed and matched with theoretical values provided by the vendor. The analysis of the biomass included the determination of its moisture, volatile material, fixed carbon (FC) and ash contents. 33 Fig. 2.4 TGA profile of calcium oxalate [2] About 10 mg of the sample were used for all the reactions. Initially, each sample was kept at 25 °C for 4 min in a N2 atmosphere and then heated to 110 °C with a N2 flow rate of 45 mL/min to determine its moisture content. A temperature of 110 °C was maintained for 5 min to achieve a constant weight before proceeding to the next step. The sample was heated to 900 °C at a heating rate of 80 °C/min, and this temperature was held for 5 min. In order to determine the FC content of the remaining solid (char), as well as its ash content, the carrier gas was switched from N 2 to air atmosphere at 900 °C and was held at this temperature for 7 min. The entire process was completed in about 32 min. 2.1.3 Thermal Degradation (Devolatilization) using TGA The non-isothermal experiments were performed at atmospheric pressure under an N 2 environment with a flow rate of 20 mL/min, beginning at room temperature and increasing to 800 °C. Pyrolysis was carried out by varying the heating rates from 5 to 20 °C/min, the particle size of the sample from 150 to 1400 µm and the initial weight of the sample from 5 to 20 mg. 34 Weight losses corresponding to increases in temperature were continuously recorded. In this study, all the experiments were replicated twice and had variations of less than 1%. 2.2 Bio-oil Production from Biomass Using Bench-scale Pyrolysis Reactor In order to obtain a significant yield of pyrolysis product, bench-scale pyrolysis was performed and analyzed. In this study, biomass was pyrolized on a bigger scale (10,000 times) than with the TGA. As can be seen in Figure 2.5, a stainless steel rod was used as a reactor to fit the available horizontal furnace and connect with the condenser, liquid collector and micro gas chromatography (micro-GC). Catalyst Powder Biomass Feedstock Stainless Steel Rod (Tubular Reactor) Heat Trace N2 Thermocouples Furnace Condenser Gas Filter Online Micro GC He Ice Bath Fig. 2.5 Experimental setup of bench-scale pyrolysis The feedstock included several types of biomass (wheat straw, flax straw, oat straw and barley straw) with particle sizes between 150 and 2000 m in the amount of 50-100 g for each 35 experiment. Before the experiments, the samples were ground in a laboratory mill and may have undergone subcritical pre-treatment and analysis using an elemental analyzer (2400 Elemental Analyzer, Perkin Elmer) to determine their composition. The bio-oil production through pyrolysis was performed in a horizontal stainless steel fixed-bed reactor (400 cm3 in volume) in an N2 atmosphere with ambient pressure. The reactor was connected to the fluid gas separator in order to collect the bio-oil while the gas was analyzed by micro-GC in real time to determine its composition. The reaction temperature varied between 300 and 500 C with reaction times of between 5 and 30 min. Pyrolysis product yields were determined gravimetrically by measuring the weight of the produced bio-oil and bio-char. The bio-oil was collected from the gas liquid separator, while the bio-char was the residual solid inside the reactor after the reaction was finished. The gas yield itself was determined by the difference between the total and combined weights of the produced bio-oil and bio-char. The experiments were carried out in series of four to study the effect of pyrolysis parameters (temperature, heating rate and reaction time) on the product yields and quality, in order to determine the optimal operating conditions that would produce the maximum yield and the best quality of bio-oil. The first series of experiments was performed to study the effect of the heating rate on the pyrolysis product yields and quality. These experiments were conducted using four different heating rates: 20, 50, 100 and 200 C/min. The reaction times varied according to the heating rates used in the experiments. The second series of experiments were performed to study the 36 effect of the particle size on the pyrolysis yields and quality. The particle sizes of the samples were 100-150, 250-300, 600-1000 and 1500-2000 m. Different particle sizes can affect the heat transfer rate during the reaction, which may impact the product yield and quality. The characterization of the bio-oil was performed using several instruments. The elemental ratio (C, H, N, S and oxygen, O) of the bio-oil was determined using an elemental analyzer (Perkin Elmer 2400 CHNS analyzer). The density of bio-oil was measured by pycnometer. In order to measure the viscosity of the bio-oil, a viscometer was used at room temperature (25 C). The pH of the bio-oil was measured using a pH meter to evaluate the corrosiveness of the bio-oil. The higher heating value (HHV) of the oil was either calculated based on the composition obtained from the ultimate analysis or was measured using a bomb calorimeter. The water content of the bio-oil was determined by applying the Karl-Fisher titration method. Since the bio-oil was presumed to contain more than a hundred chemicals, the analysis of the compounds was determined by gas chromatography / flame ionization detector (GC-FID) and gas chromatography / mass spectroscopy (GC-MS). The bio-char product produced from the pyrolysis was evaluated using an elemental analyzer in order to determine its composition and with Brunauer-Emmett-Teller (BET) surface area analysis to determine its porosity and surface area. The gaseous product was analyzed using a micro-GC to determine its composition (CO2, carbon monoxide, CO, hydrogen gas, H2, methane, CH4, N2). 37 2.3 Bio-oil Upgrading Using Catalyst The upgrading of the bio-oil was performed using two different catalysts (nickel molybdenum, Ni-Mo and nickel vanadium, Ni-V catalyst with titanium oxide base). The upgrading process was conducted in a batch reactor of a 30 cm3 size. The catalyst in the form of particles (size <150 m) was mixed with the bio-oil inside the reactor at certain ratios (1, 5, 10%). The reaction temperature varied between 250 C and 325 C. At the end of the reaction, the catalyst was removed from the liquid product, washed with hexane, dried at 100 C and was regenerated using hot air (over 500 C). The weight of the residual oil from the washing process was measured and analyzed. The liquid product was analyzed using GC-FID and GC-MS to detect the chemical compounds in the upgraded bio-oil. Several analyses, such as elemental analysis and density, viscosity, pH, HHV and water content measurements, were conducted with the same procedures as discussed previously. The gaseous product was analyzed using a micro-GC. 38 2.4 References [1] PerkinElmer, Organic Elemental Analysis, 2400 Series II CHNS/O Elemental Analysis Guide, 2011. [2] PerkinElmer, A beginner’s Guide, Thermogravimetric Analysis (TGA) STA 6000 Documentation, 2010. 39 CHAPTER 3: STUDY OF SAWDUST PYROLYSIS AND ITS DEVOLATILIZATION KINETICS 3.1 Presentation of the Article This paper provides guidelines to determine the global kinetics of the pyrolysis of biomass and to understand the devolatilization process of biomass under pyrolysis temperatures. A thermogravimetric analyzer was used to perform the devolatilization process by varying three parameters: the particle size of the biomass (sawdust), the initial weight of the sample and the heating rate over a wide range of temperature (25 to 900 °C) under a nitrogen atmosphere. This was done to observe the different trends of devolatilization due to the composition of the biomass. Based on the results, the devolatilization process was modeled using a differential equation and later solved using an integration and nonlinear optimization numerical approach. As a result, the devolatilization process of the sawdust samples during pyrolysis was divided into three noninteracting stages. The first stage was identified as the moisture loss process, the second as the cellulose and hemicelluloses devolatilization, and the third as being dominated by lignin devolatilization. The reaction orders for stages 2 and 3 were 0.9 and 1.7, respectively. Based on the activation energy and reaction order data obtained from the model, stage 2 represented the cellulose and hemicellulose devolatilization, while the third stage represented a lignin devolatilization. In addition, ultimate and proximate analyses were performed in order to provide more information about elemental composition of the sample. 40 Study of sawdust pyrolysis and its devolatilization kinetics Aqsha Aqsha, Nader Mahinpey, Thilakavathi Mani, and Feridoun Salak Department of Chemical and Petroleum Engineering, Schulich School of Engineering, University of Calgary, Calgary, Canada, AB T2N 1N4. This article was published in The Canadian Journal of Chemical Engineering, 89(6), pp. 1451-1457, 2011. 3.2 Abstract Pyrolysis of sawdust was studied using a thermogravimetric analyzer (TGA) to understand the devolatilization process and to obtain its global kinetic parameters. The influences of particle size, initial weight of the sample and heating rate on the devolatilization of sawdust particles have been studied. Results from proximate analysis show that smaller particle size has more ash content compared to larger particle size. The TG and derivative TG curve for variation in particle size and initial weight of the sample showed significant difference in the third stage of the pyrolysis. In addition, the pyrolysis of sawdust differed significantly for variation in heating rate. As the heating rates increased, the char yield also increased. The devolatilization kinetics was studied considering different stages of pyrolysis. The kinetic parameters for thermal devolatilization of the sawdust were determined through a nonlinear optimization method of two independent parallel nth-order reaction models. The kinetic parameters such as activation energy, frequency factor and order of the reaction for the two stages considered in the model were: E2 = 79.53 (kJ/mol), E3 = 60.71 (kJ/mol); k02 = 1.90 × 106 (1/min), k03 = 1.01 × 103 (1/min); n2 = 0.91, n3 = 1.78, respectively. The results show good agreement between the proposed model and the experimental data of the sawdust pyrolysis. 41 3.3 Introduction Traditionally, using biomass as fuel by direct combustion is not as efficient and comfortable compared with the energy produced from crude oil or coke combustion due to the high amount of smoke and ash produced. Nowadays, biomass can be converted to various forms of energy through several different processes (chemical, biological and thermochemical) as an alternative to coal and oil. Pyrolysis is one of the thermochemical conversion processes that have been commonly used to convert biomass into bio-oil and bio-char [1]. Although pyrolysis is an old technology, it has been improved in recent decade and is becoming more attractive than many other thermochemical processes [2, 3]. Pyrolysis is used to convert biomass to solid, liquid and gas by heating it in the absence of oxygen. During pyrolysis, the biomass undergoes a very complex set of reactions involving the formation of radicals in order to produce bio-oil and bio-char [4]. In general, pyrolysis consists of two stages of devolatilization. Primary devolatilization (less than 400 °C) consists of mainly devolatilization of hemicelluloses, whereas secondary devolatilization (over 400 °C) involves an aromatization process of components such as lignin [5]. Many techniques have been used to study the biomass pyrolysis to obtain the kinetic parameters and to gain insight into the devolatilization process of biomass pyrolysis. However, thermogravimetric analysis (TGA) is the most widely used technique for biomass pyrolysis. TGA is a technique in which the mass of the biomass is measured as a function of temperature or time while the sample is subjected to a controlled atmosphere [6]. However, the mechanism of 42 thermal devolatilization of biomass could vary depending on the components and composition of the biomass. TGA is a useful tool to study biomass devolatilization during pyrolysis, since it provides data about biomass conversion as temperature changes. Measurement of mass loss from biomass using TGA can yield information such as thermal stability and kinetic parameters of the thermal events during pyrolysis. The kinetic information is crucial for the design of pyrolysis reactors and gasifiers [7, 8]. Three stages can be distinguished during the pyrolysis process that correspond to drying, primary devolatilization and secondary devolatilization [9]. The kinetics of these stages can be determined by the application of the Arrhenius equation corresponding to the separate slopes of constant mass devolatilization [10]. In addition, a nonlinear approach has been used to optimize the parameters. In the past few decades, pyrolysis behaviors of various biomasses and their kinetics have been studied by many researchers. Caballero et al. (1997) investigated the pyrolysis kinetics of almond shells and olive stones and noted that the thermal devolatilization of lignocellulosic material can be differentiated into two global processes [11]. The first is devolatilization of hemicelluloses together with lignin and the second is related to cellulose devolatilization. Researchers have also studied the influence of heating rate on pyrolysis of various biomasses. Lapuerta et al. (2004) performed non-isothermal pyrolysis of forestry wastes with varying heating rates and showed that an optimization method has good agreement with the experimental 43 data to obtain the kinetic parameters [12]. The devolatilization of corn stover has been studied by Kumar et al. (2008) and reported that the primary devolatilization follows the first-order reaction [13]. Chouchene et al. (2010) observed the effect of particle size on pyrolysis of olive solid waste under nitrogen and oxidative atmospheres [14]. The samples having a size lower than 0.5 mm was found to be more reactive compared to larger size particles. Low ash content and higher char quantity was produced from olive solid waste with diameters between 2 and 2.8 mm during oxidative pyrolysis under 10% O2 content [14]. In another study, Lester et al. (2007) examined olive solid waste pyrolysis of three different particle sizes (53 < d <75 mm, 106< d < 150 mm and 150 < d < 212 mm) mixed with a typical high volatile coal with a particle size of 106-150 mm under nitrogen conditions [15]. It was found that the particle size variations have a limited effect on the biomass weight loss under the conditions studied. Meanwhile, the study conducted by Chen et al. (2010) found that the increase of particle size of Beech wood under the explored range (350-800 µm) significantly decreases the apparent pyrolysis rate, and therefore increases the required time for completing pyrolysis reaction [16]. Although many studies about biomass pyrolysis using TGA have been reported, the effects of important parameters such as particle size, heating rate and initial sample weight have rarely been investigated for sawdust. In the present study, sawdust pyrolysis has been performed considering the effects of various parameters, such as particle size, sample initial weight and heating rate. The devolatilization kinetics of sawdust has also been studied using nonlinear optimization. Reaction kinetic parameters were obtained by considering the three different stages that occurred during pyrolysis. This research aims in investigating the pyrolysis characteristics 44 and kinetics of sawdust and the effect of several parameters on sawdust pyrolysis that are instrumental for scaling up the pyrolysis process. 3.4 Experimental Section 3.4.1 Materials The sawdust samples used in this study were the mixture of sawdust obtained from a waste wood recycling company. The sawdust was ground in the laboratory mill and then sieved into different average particle sizes varying from 75 to 1400 µm. 3.4.2 Proximate Analysis Proximate analysis of sawdust was carried out using a TGA (STA6000; Perkin Elmer). Ceramic crucibles were used in order to minimize any thermal lag and to optimize heat transfer between thermocouples and crucibles. The analysis of sawdust includes determination of its moisture, volatile material, fixed carbon and ash contents. About 10 mg of the sample was used for all the reactions. Initially, the sample was kept at 25 °C for 4 minutes in a nitrogen atmosphere and then heated to 110 °C with a nitrogen flow rate of 45 mL/min to determine its moisture content. The temperature of 110 °C was maintained for 5 minutes to achieve constant weight before proceeding to the next step. The sample was heated to 900 °C at a heating rate of 80 °C/min, and this temperature was held for 5 minutes. In order to find the fixed carbon content of the remaining solid (char), as well as ash content, the carrier gas was switched from nitrogen to air atmosphere at 900 °C and was held at this temperature for 7 minutes. The entire process was completed in 32 minutes. Ultimate analysis has been done by using Elemental Analyzer (2400 Elemental Analyzer, Perkin Elmer). 45 3.4.3 Pyrolysis Experiment using TGA The non-isothermal experiments were performed in atmospheric pressure under nitrogen environment with a flow rate of 20 mL/min from room temperature to 800 °C. Pyrolysis was carried out by varying the heating rates from 5 to 20 °C/min, the particle size of the sample from 150 to 1400 µm and the initial weight of the sample from 5 to 20 mg. Weight losses occurring in correspondence to temperature rises were continuously recorded with a computer working in coordination with the instrument. In this study, all the experiments were replicated twice with an uncertainty of less than 1%. 3.5 Results and Discussions 3.5.1 Ultimate and Proximate Analysis of Sawdust Sample Table 3.1 shows the ultimate analysis and the high heating value (HHV) of sawdust calculated based on the equation proposed by Demirbas [17]: HHV = {33.5[C] + 142.3[H]−15.4[O]−14.5[N]} × 10−2 (3.1) HHV, [C], [H], [O] and [N] represent the HHV, carbon, hydrogen, oxygen and nitrogen contents of material, respectively. The heating value reported in this work was within the range of the value (15 000-20 500 kJ/kg) reported by Demirbas for 16 different kinds of biomass [17]. The ultimate analysis corresponds to the average composition of CH 1.59N0.03S0.01. Table 3.1 Ultimate analysis of the sawdust in weight percentage Sawdust a C 49.28 H 6.59 N 1.70 Calculated from the composition of the sawdust 46 S 1.25 HHVa (kJ/kg) 19298 Figure 3.1 illustrates the TGA profile of the sawdust drying, devolatilization, consequent char formation, and finally, char oxidation for the proximate analysis. The first weight loss, around 110 °C, corresponds to the moisture loss of the sample. At temperatures below 110 °C, moisture content is reduced by up to 5% of the sample weight. The second weight reduction corresponding to increasing the temperature to 900 °C is attributed to devolatilization and pyrolysis of mainly organic materials. All the volatiles were evolved at 900 °C, and only the char remained. In this step, the yield of the volatile material is up to 67% of the initial weight of sawdust sample. The fixed carbon content can be found from the oxidation of the produced char by switching the carrier gas from nitrogen to air at 900 °C. In this step, the char is oxidized into carbon dioxide and carbon monoxide. The weight differences before and after switching the gas were used to determine the fixed carbon content, which is up to 18% of the weight sample. The remaining residue represents the ash content. 900 C Ash Fixed carbon Volatiles 25 C 110 C Moisture content Fig. 3.1 TG profile for proximate analysis of sawdust for the particle size of 250 µm. 47 The TGA profile of proximate analysis differed with the particle size of the sample. The volatile material increased as the particle size increased from 150 to 1400 µm as shown in Table 3.2. On the other hand, as the particle size increased, the ash content decreased. This might be caused by the distribution of biomass particle during grinding and sieving processes. When biomass was grounded, there is possibility that part of the biomass that contain the inorganic components (ash) was breakdown into smaller particle size. As a result, larger particles produced smaller amounts of ash. Similar results were reported by Mani et al. (2010) with particle size between 325 and 1350 µm of a wheat straw sample and also by Zanzi et al. (2002) in pyrolysis of olive waste and wood [9, 18]. In another study, Chouchene et al. (2010) observed comparable results with particle size between 2 and 2.8 mm of olive waste during oxidative pyrolysis, except that the volatile material decreased as particle size increased [14]. Chouchene et al. (2010) noted that this behavior might be caused by the larger heat propagation rate associated with smaller particle size, which could have, therefore, released a greater quantity of volatile material [14]. In addition, this may be due to the inorganic components separated from the lignocellulosic structure during size reduction of the sample and tend to accumulate in smaller size fractions [19]. Table 3.2 Proximate analysis of sawdust at different particle sizes Sieve size (µm) Average size (µm) 212-300 256 106-150 420-600 600-850 850-1000 1000-1400 Volatiles (%) Fixed carbon (%) Ash (%) <150 56.97 (±0.33) 17.65 (±0.23) 10.75 (±0.74) 505 70.13 (±0.02) 18.35 (±0.31) 4.01 (±0.06) 725 925 1200 60.01 (±0.05) 69.35 (±0.27) 69.18 (±0.03) 77.03 (±0.45) Uncertainty values are given in parentheses. 48 18.03 (±0.04) 17.63 (±0.33) 17.51 (±0.42) 18.15 (±0.16) 9.02 (±0.12) 4.61 (±0.40) 5.01 (±0.02) 1.01 (±0.21) 3.5.2 Pyrolysis of Sawdust As shown in Figure 3.2, several events can be distinguished during the thermal pyrolysis of the sawdust sample. Pyrolysis of sawdust involving thermal devolatilization consisted of a very complex set of reactions. The reactions can be represented as the sum of thermal devolatilization reactions of the individual components of lignin, cellulose and hemicellulose. As the temperature increased, the rate of the devolatilization process also increased. Generally, during pyrolysis, the moisture is removed initially at a temperature below 110°C [20]. Above 110 °C, the chemical bond breaks to release the volatile compounds. It has been reported that hemicelluloses degrades fast when compared to cellulose and lignin [21]. Above 200 °C, the celluloses and lignin started to break and released more volatiles upto 400 °C. Beyond 400 °C, the degradation rate is slow corresponding to the degradation of lignin components. Stage 2 Stage 1 Stage 3 Fig. 3.2 Classification of devolatilization stages under nitrogen atmosphere. 49 The derivative thermogravimetric (DTG) curve could assist in explaining the TG curve. In Figure 3.2, the spread of the reaction over temperatures of 25-900 °C appears as a relatively broad peak. The area under the peak represents the weight loss during the reaction. Thus, relative mass losses may be compared. When initial slow reactions are followed by fast reactions above 200 °C and then a second slow reaction initiates at 380 °C, there is a gradient change in the DTG curve. This information is used in this study to distinguish the different stages of each process and their respective reaction kinetics. 3.5.2.1 Effect of Particle Size The influence of the particle size was investigated for six different particle sizes of sawdust ranging between <150 and 1400 µm. As shown in Figure 3.3, particle size does not have a significant effect on the TG profile of the first and the second stages of sawdust pyrolysis. In contrast, as the particle size increased from 150 to 1400 µm, the char yield decreased (Table 3.3). This may be due to the amount of heat transferred to different sizes of particles. The samples consisting of larger particle may have had higher rate of heat transfer. As a result, during the second stage (primary devolatilization), larger amounts of sample were decomposed and less amounts of char remained at the end. Haykiri-Acma (2006) also obtained similar results for nonisothermal pyrolysis of hazelnut shell to study the role of particle size in the range between 150 and 1400 µm [22]. It has been found that the char yield is less for smaller particle size. A comparable result was also reported by Demirbas (2004) for the effect of temperature and particle size of various biomasses on char yield and reactivity [23]. Figure 3.3b shows DTG curves for different particle sizes in sawdust pyrolysis. 50 (a) (b) Fig. 3.3 Pyrolysis curve of sawdust at different particle sizes of sample with heating rate of 10 °C/min and initial weight of 10 mg of sample. (a) TG profile and (b) DTG profile. 51 Table 3.3 Char yield percentage for different values of particle size of the sample for sawdust pyrolysis with initial weight of 10 mg and heating rate of 10 C/min. Average particle size (µm) <150 256 505 725 925 1200 Char yield (%) 17.26 (±0.01) 15.16 (±0.05) 15.38 (±0.61) 15.94 (±0.03) 13.73 (±0.02) 11.85 (±0.66) Uncertainty values are given in parentheses. 3.5.2.2 Effect of Initial Sample Weight Figure 3.4 shows the influence of initial weight of the sample for sawdust pyrolysis. It is shown that the change in pyrolysis behavior was observed in the third stage due to the change of initial amount. According to the TG/DTG curves, it was clear that the char yield increased as the initial weight increased. However, initial weight did not have significant effect on the first and the second stages of the pyrolysis of the sawdust (Figure 3.4). The effect of the initial weight of the sample was mainly studied to understand the role of inert gas diffusion into the bed sample. By increasing the initial weight, the inert gas will encounter a higher diffusion resistance inside the bed. Since the bed sample was getting higher inside the crucible, the gas that act as the heating medium had to travel longer into the bulk and the heat transfer also would be affected adversely and thus more char is produced (Table 3.4). Table 3.4 Char yield percentage for different values of initial weight of the sample for sawdust pyrolysis with average particle size of 352 µm and heating rate of 10 C/min. Initial weight (mg) 5 10 15 20 Char yield (%) 9.54 (±0.39) 16.23 (±0.11) 18.13 (±0.33) 19.99 (±0.18) Uncertainty values are given in parentheses. 52 (a) (b) Fig. 3.4 Pyrolysis curve of sawdust at different initial weights of sample with heating rate of 10 °C/min and particle size of 352 µm. (a) TG profile and (b) DTG profile. 53 3.5.2.3 Effect of Heating Rate In this experiment, pyrolysis was carried out using four different heating rates. Figure 3.5a and b shows the TG and DTG curves for different heating rates from 5 to 50 °C/min. It was clear that heating rates have a significant effect on the TG/DTG profile of sawdust pyrolysis on all three stages. It can be seen that the TG curve shifted as the heating rate increased. As the heating rate increased the char yield also increased. At the lower heating rate, the heat transfer between the crucible and the sample were more efficient. This resulted in a proper drying and devolatilization during the second stage. As a result, fewer residues were produced at the slow heating rate. In contrast, at higher heating rate, the devolatilization process occurred sooner due to the increased rate of heat transfer between the crucible and the sample. Faster heating rates cause the primary devolatilization to complete rapidly because the temperature for secondary devolatilization has been reached rapidly. From the DTG curves for different heating rates (Figure 3.5b), it was concluded that the peak of the rate of devolatilization shifted to the higher value as heating rate increased. Since the faster heating rate led to less efficient heat transfer, the devolatilization rate increase faster than at lower heating rates, thus shifting the peak of the devolatilization rate. As a result, char yield percentage increased as the heating rates increased, as shown in Table 3.5. Table 3.5 Char yield percentage for different values of heating rate of the sample for sawdust pyrolysis with average particle size of 352 µm and initial weight of 10 mg. Heating rate (C/min) Char yield (%) 10 16.23 (±0.11) 50 20.30 (±0.13) 5 20 11.81 (±0.35) 22.17 (±1.55) Uncertainty values are given in parentheses. 54 (a) (b) Fig. 3.5 Pyrolysis curve of sawdust at different heating rates with particle size of 325 µm and initial weight of 10 mg of sample. (a) TG profile and (b) DTG profile. 55 3.5.2 Kinetics of Sawdust Pyrolysis In order to scale up the pyrolysis process, kinetic information is needed. Furthermore, kinetic information can provide a better understanding of the mechanisms of thermal processes of sawdust. The objective of the kinetics is to provide mathematical relationships between temperature and conversion. According to Figure 3.2, three visible events can be distinguished during the heating process of the sawdust sample. The first weight loss corresponds to the moisture release and the drying of the sample (30-110 °C). In a second event, between 110 and 400 °C, the lightest volatile compounds evolve and this event is usually called primary devolatilization. Finally, the last step corresponds to the secondary devolatilization, which happens above 400 °C. In this work, experimental results were correlated as a function of the yield of volatile material using a nonlinear model. The total volatile material can be obtained using the relation: V 1 (C A) (3.2) where C and A are the mass of fixed carbon and ash content, respectively. The final accumulated yield of each event is represented by V1*, V2* and V3*. In practice, once the fixed carbon and ash content is known, the final volatile material yield can be obtained as the remaining fraction. The sawdust pyrolysis was assumed as two independent parallel devolatilization reactions (mainly primary and secondary devolatilization), which are described by nth-order reactions as follows: ni dVi t k i Vi Vi t dt (3.3) 56 where Vi(t) is the accumulated mass of released volatile material from event i up to time t, Vi* is the ultimate yield of volatile material from event i up to t = ∞, ni is the reaction order of event i and ki is the rate constant of event i: E k i k 0i exp i RT (3.4) where k0i is the pre-exponential factor, Ei is the activation energy for event i, R is the gas constant and T is the temperature of the system. The total volatile material yield is computed by the addition of: V t Vi t (3.5) V Vi (3.6) 3 i 1 3 i 1 Since the process was held at constant heating rate (H), Equation (3.3) can be expressed as: dVi t V Vi t ki i dT H ni (3.7) Integration of Equation (3.7) gives: Vi Vi t Vi T k exp E i / RT 1 (1 ni ) 0i dT H 0 1 /(1 ni ) (3.8) Equation (3.8) contains an integral that can be solved by analytical solution. However, the analytical solution contain a special function, thus required complex mathematical solution to solve the problem. Various simplifications have been tried, and previous studies have presented 57 several approximations with varying complexity and accuracy [24]. In this study, the modified Donskoi approximation was used to replace the integral form in Equation (3.8) [25]. Vi Vi t Vi k 0i RT 2 exp E i / RT 1 (1 ni ) Ei H 1 /(1 ni ) (3.9) The kinetic parameters such as pre-exponential factor, activation energy and reaction order for each event can be obtained by a nonlinear regression method on the TG data of the sawdust pyrolysis. In this work, a nonlinear optimization method was used to obtain the kinetic parameters of pyrolysis of sawdust under different heating rates (5, 20 and 50 °C/min), which minimized the sum of square differences between experimental and predicted values of evolved volatile material in each stage (Table 3.6). In this process, GraphPad Prism was used to optimize the parameters. The code built in the GraphPad consisted of an optimization algorithm to minimize the error between the data and the model proposed. The challenge in this process was to have a good value for an initial estimate for the optimization. Table 3.6 Weight loss and ultimate devolatilization yield of the devolatilization events with average weight prior to devolatilization of 9.91 mg at different heating rates. Ultimate devolatilization yield Heating rates Weight after (C/min) devolatilization (mg) 5 (wt% of total mass loss) V1* 1.13 9.65 V2* 73.03 V3* 17.31 10 1.59 7.97 77.10 14.93 50 2.08 5.58 80.00 14.41 20 2.04 6.53 58 80.79 12.68 The kinetic parameters k0, E and n are listed in Table 3.7. The pre-exponential factor decreased from the order of 106 to 103, and the activation energy also decreased from 80 to 61 kJ/mol for stages 2 and 3 of the sawdust pyrolysis, respectively. The decrease in the activation energy might be caused by the reactions that happened in each stage. Lignin degradation may be dominant in the third stage, leading to lower activation energy. The order of the reaction of primary devolatilization is approaching 1 (n = 0.9) and for the secondary devolatilization it was 1.78. Comparison of reaction orders obtained from present work with the literature revealed that primary devolatilization (stage 2) represents the devolatilization of cellulose and hemicelluloses and the secondary devolatilization represents the degradation of lignin [26]. Table 3.7 Kinetic parameters of the three events considered in the kinetic model. Stage 1 Stage 2 Stage 3 k0 (min-1) 1.41 1.90 106 1.01 103 E (kJ/mol) 4.81 79.53 60.71 n 0.36 0.91 1.78 Figure 3.6 shows good agreement between the experimental and predicted value of accumulated volatile material yield curves (with respect to total volatile loss) at a heating rate of 10 °C/min. According to this result, it was clear that the devolatilization data predicted from this model was very similar to the experimental data obtained from TGA. In general, this model might be used to predict the TG curves for different heating rates. 59 Fig. 3.6 Typical comparison of experimental data with predicted values at heating rate of 10 °C/min. 3.6 Conclusions Pyrolysis of sawdust has been analyzed using TGA to study the effect of particle sizes, initial weights and heating rates. The results from the proximate analysis showed that the volatile material evolution was higher with larger particle sizes. In the pyrolysis process, the char yield decreased as the particle size of the sample became larger. On the other hand, reducing the particle size of the sample or increasing the initial weight or increasing the heating rate of pyrolysis process increased the char production. The devolatilization process of the sawdust samples during pyrolysis was divided into three noninteracting stages. The first stage was identified as the moisture loss process, the second as the cellulose and hemicelluloses devolatilization and the third as being dominated by lignin devolatilization. A modified devolatilization kinetic model was used to obtain the kinetic 60 parameters of the reactions. The reaction orders for the stages 2 and 3 were 0.9 and 1.7 respectively. Based on the activation energy and reaction order data obtained from the model, the stage 2 represented the cellulose and hemicellulose devolatilization while the third stage by lignin devolatilization. The result shows good agreement between the proposed model and the experimental data for different heating rates of the process. 3.7 Acknowledgement The authors wish to acknowledge the financial support of the Natural Science and Engineering Research Council of Canada (NSERC). 61 3.8 References [1] Mahinpey, N., Murugan, P., Mani, T. & Raina, R. Analysis of bio-oil, bio-gas and bio-char from pressurized pyrolysis of wheat straw using a tubular reactor, Energy Fuels, 23(5), pp. 2736-2742, 2009. [2] Mohan, D., Pittman, C. U. & Steele, P. H. Pyrolysis of wood/biomass for bio-oil: a critical review. Energy Fuels, 20, pp. 848-889, 2006. [3] Scott, D.S. & Piskorz, J. The continuous flash pyrolysis of biomass. Can. J. Chem. Eng., 62, pp. 404-412, 1984. [4] Yaman, S. Pyrolysis of biomass to produce fuels and chemical feedstocks. Energy Convers. Manag., 45, pp. 651-671, 2004. [5] Fisher, T., Hajaligol, M., Waymack, B. & Kellogg, D. Pyrolysis behaviour and kinetics of biomass derived materials. J. Anal. Appl. Pyrolysis, 62, pp. 331-349, 2002. [6] Menczel, J.D. & Prime, R.B. Thermal analysis of polymers fundamental and applications. John Wiley & Sons: New Jersey, 2009. [7] Dirion, J.L., Reverte, C. & Cabassud, M. Kinetic parameter estimation from TGA: optimal design of TGA experiments. Chem. Eng. Res. Des., 86, pp. 618-625, 2008. [8] Kojima, T., Assavadakorn, P. & Furusawa, T. Measurement and evaluation of gasification kinetics of sawdust char with steam in an experimental fluidized bed. Fuel Process. Technol., 36, pp. 201-207, 1993. [9] Mani, T., Murugan, P., Abedi, J. & Mahinpey, N. Pyrolysis of wheat straw in a thermogravimetric analyzer: Effect of particle size and heating rate on devolatilization and estimation of global kinetics. Chem. Eng. Res., 478, pp. 1-7, 2010. 62 [10] Gang, W. & Aimin, L. Thermal devolatilization and kinetics of mixtures of polylactic acid and Biomass during copyrolysis. Chin. J. Chem. Eng., 16(6), pp. 929-933, 2008. [11] Caballero, J.A., Conesa, R.J.A. & Marcilla, A., Pyrolysis kinetics of almond shells and olive stones considering their organic fractions. J. Anal. Appl. Pyrolysis, 42, pp. 159-175, 1997. [12] Lapuerta, M., Hernandez, J.J. & Rodriguez, J., Kinetics of devolatilisation of forestry wastes from thermogravimetric analysis. Biomass Bioenergy, 27, pp. 385 - 391, 2004. [13] Kumar, A., Wang, L., Dzenis, Y.A., Jones, D.D. & Hanna, M.A., Thermogravimetric characterization of corn stover as gasification and pyrolysis feedstock. Biomass Bioenergy, 32, pp. 460-467, 2008. [14] Chouchene, A., Jeguirim, M., Khiari, B., Zagrouba, F. & Trouvé, G., Thermal devolatilization of olive solid waste: influence of particle size and oxygen concentration. Resour. Conserv. Recy., 54, pp. 271-277, 2010. [15] Lester E., Gong M. & Thompson A., A method for apportionment in biomass/coal using thermogravimetric analysis. J. Anal. Appl. Pyrolysis, 80, pp. 111-7, 2007. [16] Chen, L., Dupont, C., Salvador, S., Boissonnet, G., & Schweich, D., Influence of Particle Size, Reactor Temperature and Gas Phase Reactions on Fast Pyrolysis of Beech Wood, Int. J. Chem. Reactor Eng., 8(A10), pp. 1-19, 2010. [17] Demirbas, A., Calculation of higher heating values of biomass fuels. Fuel, 76(5), pp. 431434. [18] Zanzi, R., Sjostrom, K. & Bjornbom, E., Rapid Pyrolysis of Agricultural Residues at High Temperature, Biomass Bioenergy, 23, pp. 357-366, 2002. 63 [19] Bridgeman, T.G., Darvel, L.I., Jones, J.M., Williams, P.T., Fahmi, R., Bridgwater, A.V., Barraclough, T., Shield, I., Yates, N., Thain S.C. & Donnison, I.S., Influence of Particle Size on the Analytical and Chemical Properties of Two Energy Crops, Fuel, 86, pp. 60-72, 2007. [20] Kim, S.S., Kim, J., Park, Y.H. and Park, Y.K., Pyrolysis Kinetics and Decomposition Characteristics of Pine Trees, Bioresour. Technol., 101, pp. 9797-9802, 2010. [21] Becidan, M., Skreiberg, Ø. & Hustad, J.E., Products Distribution and Gas Release in Pyrolysis of Thermally Thick Biomass Residues Samples, J. Anal. Appl. Pyrolysis, 78, pp. 207-213, 2007. [22] Haykiri-Acma, H., The role of particle size in the non-isothermal pyrolysis of hazelnut shell. J. Anal. Appl. Pyrolysis, 75, pp. 211-216, 2006. [23] Demirbas, A., Effects of temperature and particle size on bio-char yield from pyrolysis of agricultural residues. J. Anal. Appl. Pyrolysis, 72, pp. 243-248, 2004. [24] Cai, J., Yao, F., Yi, W. & He, F., New temperature integral approximation for nonisothermal kinetics. AIChE J., 52, pp. 1554-1557, 2006. [25] Donskoi, E. & McElwain, D.L.S., Approximate modelling of coal pyrolysis. Fuel, 78, pp. 825-835, 1999. [26] Caballero, J.A., Conesa, R.J.A. and Marcilla, A., Pyrolysis Kinetics of Almond Shells and Olive Stones Considering Their Organic Fractions, J. Anal. Appl. Pyrolysis, 42, pp. 159175, 1997. 64 CHAPTER 4: CATALYTIC PYROLYSIS OF STRAW BIOMASSES (WHEAT, FLAX, OAT AND BARLEY STRAW) AND THE COMPARISON OF THEIR PRODUCT YIELDS 4.1 Presentation of the Article In this work, the pyrolysis of four different straw biomasses (wheat, oat, flax and barley straws) was studied using a bench-scale pyrolysis system that was built in the University of Calgary’s engineering machine shop and laboratory. The reactor was a batch-type horizontal reactor enclosed by a horizontal split tubular furnace and consisted of a stainless-steel tube with a length of 535 mm and an inside diameter of 22.5 mm. Thermocouples were placed within the reactor system to measure the furnace, sample, reactor and outlet temperatures. Pyrolysis creates three end-products: bio-char in the form of solid powder, bio-oil in the form of black liquid, and gaseous materials. In order to determine the yield of each product (solid, liquid and gas products), the reactor was connected to a condenser and liquid trap to recover the bio-oil. The bio-char remained in the reactor, and the gaseous product yield was calculated by subtraction. In this series of experiments, a zeolite catalyst was used in order to observe the effect of the catalyst during pyrolysis. Bio-char was recovered from the reactor and later analyzed using a thermogravimetric analyzer for proximate analysis, the CHNS/O (carbon, hydrogen, nitrogen, sulphur and oxygen) analyzer for elemental analysis, a calorimeter for determination of the heating value and a BrunauerEmmett-Teller system (BET Micromeritics ASAP 2020) for surface area measurement. 65 Bio-oil was analyzed using gas chromatography-mass spectrometry (GC-MS HP 6890) to determine its chemical composition and later underwent elemental analysis and heating value measurements. In addition, a Karl-Fischer titrator (KF Methrohm 870 Titrino) was used to determine the water content of the bio-oil. 66 Catalytic Pyrolysis of Straw Biomasses (Wheat, Flax, Oat and Barley Straw) and the Comparison of their Product Yields A. Aqsha, M.M. Tijani & N. Mahinpey Department of Chemical and Petroleum Engineering, Schulich School of Engineering, The University of Calgary, Calgary, Canada, AB T2N 1N4 This article has been submitted to Energy Conversion and Management peer-reviewed journal in February 2016. 4.2 Abstract In this study, the pyrolysis of several Canadian straw biomasses was studied using a thermogravimetric analyzer and a bench-scale horizontal fixed-bed reactor, in order to better understand the devolatilization process and to obtain information about the product yields of these biomasses. The straw biomasses were converted through pyrolysis performed in a fixedbed reactor at temperatures of 500 °C to study the influence of the feedstock on product yields. The effects of various catalysts on product yields are also discussed. When using zeolite catalysts, the bio-oil and bio-char yields of the straw pyrolysis were increased to 46.44% and 38.77%, respectively, while the gas yield was decreased to 13.65%. The use of catalyst zeolite ZY-SS had the most significant effect on overall bio-oil and bio-char yields, increasing the biooil yield by about 2% and the bio-char yield by 8%. This catalyst had the most significant effect on the pyrolysis of flax straw, where the bio-oil yield was increased to 46.44%. In the pyrolysis of oat straw, the use of the catalyst consistently decreased the gas yield; however, the bio-oil yield increased the most significantly (43.32%) with the use of catalyst zeolite ZY-NS. The use of zeolite ZY-NS also increased the bio-oil yield during the pyrolysis of barley straw (43.03%). 67 Keywords: Biomass, Biofuels, devolatilization process, kinetic of pyrolysis, proximate analysis, pyrolysis of biomass. 4.3 Introduction Using several different processes (chemical, biological and thermochemical), biomass can be converted into various types of chemicals that can be used either as alternatives to fossil fuel, such as gasoline and petrodiesel, or as value-added chemicals. Lignocellulosic straw biomasses (wheat, flax, oat and barley) are considered promising renewable feedstock sources for the production of chemicals and biofuels, and they are widely available in the Canadian Prairies. The use of biomass or waste biomass, such as wheat straw, flax straw, barley straw, or sawdust, for energy production will provide several advantages from an environmental perspective. Pyrolysis is a thermochemical process that can be used to break down certain materials into a liquid product called bio-oil [1]. Although pyrolysis is an old technology, it is becoming more attractive than other thermochemical processes, as a result of significant progress in the past decade [2, 3]. During pyrolysis, the applied heat decomposes the materials into a volatile compound (condensed and uncondensed material) that will later produce bio-char, bio-oil, and gaseous material. Similar to wood, straw biomass from wheat, flax, oat, and barley is primarily composed of cellulose, lignin, and hemicellulose [4]. Each straw biomass has a different composition that is suspected to have a significant effect on the product yield and composition [5]. The temperature 68 of the reaction, heating rate, particle size, and residence time also have considerable influence on the properties of the product and its composition [6, 7]. For instance, Haykiri-Acma [8] suggested that increasing the heating rate to 50 °C/min leads to higher conversion rates due to the better mass and heat transfer inside the reactor [9]. In addition, a temperature of 500 °C should be used to maximize the liquid oil yield [10]. If these parameters are not met, secondary reactions may occur, thereby decreasing the oil yield and leading to a higher yield of gas and char [11]. Recently, the use of pyrolysis to produce fine chemicals that have low oxygen content such as phenol has gained some attention [12-14]. It was suspected that low lignin content in biomass caused the low production of phenolic compound [12]. Zhang et al. (2015) perform a fast pyrolysis using magnetic solid base catalyst K3PO4/Fe3O4 and successfully increased the yield of phenolic component about 1.5 times compared to non-catalytic pyrolysis. They conclude that the catalyst would promote the decomposition of lignin to increase the phenolic yield [15]. Other researchers also proposed the high density polyethylene (HDPE) material to be used in copyrolysis of almond shell to produce bio-oil containing higher hydrogen to carbon and oxygen ratio [16]. They successfully reduced oxygen content up to 86% and produced bio-oil with H/C ratios close to transport fuel. However providing pure and dry HDPE for co-pyrolysis might be difficult due to the waste management system that most of the time combined this waste type with other municipal solid waste. In this research, the results of experimental studies on the pyrolysis of several Canadian straw biomasses (wheat, flax, oat, and barley) are described. The pyrolysis of the straw biomasses was 69 performed in a fixed-bed reactor at temperatures of 500 °C to study the influence of the feedstock on product yields. The effects of various catalysts on the product yields were also investigated. To evaluate the effects of the zeolite catalyst on the straws, three forms of zeolite catalysts were used (silica-sodium-based solution, silica-based solution and H-form of the silicabased solution), which were named as ZY-NS, ZY-SS and HZY-SS, respectively. The results from the comparison of biomasses provide useful information about the suitability of each biomass for particular end products, such as bio-char, bio-oil, and gaseous material. 4.3 Experimental Section 4.4.1 Raw Biomass Sample and catalysts The straw biomass samples (wheat, flax, oat, and barley) used in this study were obtained from a farm in Saskatchewan, Canada. The straw biomasses were ground and sieved to obtain an average particle size of less than 1 mm. The chemical composition of the samples according to various documented sources is set out in Table 4.1 while the elemental composition of the catalysts and their surface area are presented on Table 4.2 Table 4.1 Chemical composition of wheat, oat, flax and barley straw samples Cellulose Hemicellulose Lignin a Wheat Straw Oat Straw Flax Straw Barley Straw 34.2a 37.6a 53.0b 33.8c 13.9a 12.9a 24.0b 13.8c 23.7a 23.3a 13.0b 21.9c Adapa et al [17]; b Buranov and Mazza [18]; c Mussatto and Teixeira [19] 70 Table 4.2 Elemental composition and surface area data of the catalysts Si/Al Si (%wt) Al (%wt) Na (%wt) BET (m2/g) ZY-NS ZY-SS HZY-SS 1.67 2.26 19.1 15.7 31.9 13.1 645 2.26 35.5 41.29 12.0 10.8 656 18.2 437 4.4.2 Proximate and Elemental Analysis Proximate analyses of the straw biomasses were performed using a thermogravimetric analyzer (TGA, NETZSCH TG 209 F1 Libra). All experiments consisted of three steps: drying, devolatilization in a nitrogen inert atmosphere, and combustion with air (Figure 4.1). About 10 mg of each sample with a particle size of 1 mm were used for this experiment. Initially, each biomass sample was kept at 25 °C for 4 min in a nitrogen atmosphere. The moisture content was determined from the weight loss when the sample was heated at a rate of 85 °C/min up to 110 °C and held constant for 5 min. Devolatilization began once the temperature reached 110 °C, and a heating rate of 80 °C/min was maintained up to 900 °C. The final temperature was held constant for 7 min in an air atmosphere to allow for complete combustion. The flow rate of the nitrogen and air was maintained at 45 mL/min. 71 Fig. 4.1 Typical profile for proximate analysis using a TGA Elemental analyses were performed using a 2400 carbon, hydrogen, nitrogen, sulphur, oxygen (CHNS/O) analyzer (2400 elemental analyzer, Perkin-Elmer). The combustion time was about 5 minutes, and the packing of the combustion tube consisted of chromium oxidizer, silver tungstate on magnesium oxide, and silver vanadate. The solid sample was placed in a tin capsule, while the liquid was placed in a sealed aluminium capsule. 4.4.3 Pyrolysis Experiment Pyrolysis experiments were carried out using the TGA and a batch-type horizontal reactor enclosed within a horizontal split tubular furnace. The reactor consisted of a stainless-steel tube with a length of 535 mm and an inside diameter of 22.5 mm. Thermocouples were placed within the reactor system to measure the temperatures of the furnace, pyrolysis reactor, and outlet. Nitrogen was used as the inert gas with a constant flow rate of 50 mL/min throughout the process. During the experiments, each sample was heated at a rate of 100 °C/min to a final temperature of 500 °C at atmospheric pressure. 72 At the end of the reactor, volatile products (gas and liquid) were isolated using a separator that was submerged in a 0 °C liquid water-ice chamber. The liquid product (bio-oil) was collected at the end of the reaction and measured using a laboratory scale. Fig. 4.2 Thermogravimetric profiles of raw biomasses subject to pyrolysis. For each experiment, 10 g of the straw biomass were loaded into the stainless steel reactor. Each experiment was performed with the same weight at a temperature of 500 °C for about 1 hour. The product yields were calculated using the weight data of the bio-char, bio-oil and gas products, and the initial biomass sample. All experiments were performed at least twice, and the uncertainty associated with the data was within 2%. Figure 4.2 provides the profiles of the raw biomasses subject to pyrolysis using a TGA. 73 4.4.4 Product Analysis and Characterization Pyrolysis creates three end products: bio-char in the form of solid powder, bio-oil in the form of black liquid, and gaseous materials. Bio-char was recovered from the reactor and later analyzed using the TGA for the proximate analyses, the CHNS/O analyzer for elemental analyses, a calorimeter for determination of the heating values, and a Brunauer–Emmett–Teller system (BET Micromeritics ASAP 2020) for surface area measurements. Bio-oils were analyzed using gas chromatography / mass spectrometry (GC-MS HP 6890) to determine its chemical composition and also underwent elemental analyses and heating value measurements. In addition, a Karl-Fischer titrator (KF Methrohm 870 Titrino) was used to determine the water contents of the bio-oils. 4.4 Results and Discussions 4.5.1 Proximate and Elemental Analysis The results of the proximate and elemental analyses of all four biomass samples are presented in Table 4.3. These results demonstrate that the moisture content of the biomasses was in the range of 3-5%, the volatile content about 74-76%, the fixed carbon in the range of 14-19%, and the ash content below 5%. The data from these analyses of condensable (bio-oil) and non-condensable (gas) materials produced during pyrolysis can be used to predict the composition of products derived from pyrolysis. It was predicted that the combined amount of bio-oil and gaseous materials produced through pyrolysis would be approximately 75%. Based on the elemental analysis data, the bio-oil 74 product would be expected to contain a high amount of oxygen. The elemental analyses also proved that the sulfur contents of the biomasses were very small (~1%), while the oxygen contents were about 45% (calculated by difference). Both values are common among straw and wood biomasses. Table 4.3 Proximate and elemental analyses of biomasses Proximate Analysis (wt%) Wheat Straw Oat Straw Flax Straw Barley Straw 5.30 4.38 4.75 2.69 Moisture Volatile 75.88 74.04 74.80 75.64 Ash 4.69 3.21 1.12 4.38 Fixed Carbon Elemental Analysis (wt%) 14.12 18.37 19.32 17.28 C 43.64 43.26 46.76 44.83 N 0.68 0.88 1.11 0.93 H S O* 5.82 6.12 1.08 1.14 44.09 45.39 6.34 1.14 43.53 6.25 1.18 42.43 * Weight of O was obtained by subtracting the sum of the CHNS and ash contents from 100% (percentage). Table 4.4 presents the data comparison of the various raw biomasses and their bio-char products using proximate and elemental analyses. The volatile matter, as a percentage of biomass, was reduced to as low as 18.23 % after pyrolysis (bio-char). The proximate analyses revealed that the bio-chars were composed of about 66.03% fixed carbon and 22.2% volatile matter. The high amount of carbon increased the energy density of the char, which is reflected in the heating value data measured with the calorimeter. 75 The elemental analyses demonstrated that the bio-chars had carbon compositions in the range of 66.42% (oat straw) to 72.61% (barley straw), representing increases about 53.5% and 62% from their original values, respectively. The oxygen content was greatly reduced from 44.09% to a value as low as 13.46% (wheat straw). These results suggest that most of the oxygen transformed into liquid and gaseous products, while most carbon remained as solid products. Table 4.4 Data comparison between raw biomass and its bio-char product Biomass Sample Proximate Analysis (wt%) Moisture Raw wheat straw 5.30 Raw oat Raw flax straw straw 4.38 4.75 Raw barley straw 2.69 Wheat Flax Barley Oat straw straw biostraw bio- straw biobio-char char char char 3.67 4.93 2.83 2.26 Volatile Matter 75.88 74.04 74.80 75.64 26.65 20.78 18.23 23.14 Ash 4.69 3.21 1.12 4.38 12.13 11.22 4.55 5.52 Fixed Carbon Elemental Analysis (wt%) 14.12 18.37 19.32 17.28 57.55 63.08 74.40 69.09 C 43.64 43.26 46.76 44.83 67.87 66.42 71.73 72.61 N 0.68 0.88 1.11 0.93 3.02 1.26 2.32 1.73 H S Oa 5.82 1.08 6.12 1.14 6.34 1.14 6.25 1.18 2.97 0.55 2.34 2.80 0.49 0.61 3.09 0.62 44.09 45.39 43.53 42.43 13.46 20.28 17.99 16.44 H/C Molar Ratio 1.59 1.69 1.62 1.66 0.52 0.42 0.47 0.51 O/C Molar Ratio 0.76 0.79 0.70 0.71 0.15 0.23 0.19 0.17 C3.6H5.8O2.8 C3.6H6.1O2.8 C3.9H6.3O2.7 C3.7H6.2O2.7 C5.7H2.9O0.8 C5.5H2.3O1.1 C6.0H2.8O1.1 C6.0H3.1O1.0 17.97 17.51 18.62 17.95 23.40 23.99 27.67 27.70 2.438 1.535 2.579 1.889 6.056 1.991 1.661 2.287 H/Ceff Molar Ratio = (H-2×O)/C Empirical Formula Heating Value (MJ/kg, dry) Surface Area (m2/g, dry basis) a 0.07 0.11 0.22 0.24 0.22 0.01 0.09 O was obtained by subtracting the sum of the CHNS and ash contents from 100% (percentage). 76 0.17 Fig. 4.3 Location of straw biomasses and their biochar products on a Van Krevelen diagram. The hydrogen/carbon (H/C) molar ratio, effective hydrogen/carbon (H/Ceff) molar ratio and oxygen/carbon (O/C) molar ratio of each raw material and bio-char product were also calculated to investigate whether the raw samples and bio-char products had any similarities. It was found that the H/C ratios of the raw biomasses and bio-chars were around 1.6 and 0.5, respectively. According to the Van Krevelen diagram (Figure 4.3), the straw biomass samples (wheat, oat, flax, and barley) were located in the biomass zone, while bio-char products were located below the coal zone. This suggests that the produced bio-char products had higher heating values than those of the raw biomasses and had compositions that were closer to those of coal. In addition, the empirical analysis showed a notable reduction in hydrogen, which suggests the formation of methane (CH4), H2 and H2O during pyrolysis, which was confirmed by the gas analysis and the Karl Fischer titration of bio-oil. 77 4.5.2 Pyrolysis of Raw Biomass The three end products of the pyrolysis process were collected and measured to determine the yield of each product. The bio-char product was gathered from the reactor in the form of black solid powder, bio-oil was collected from the condenser unit connected to the reactor, and gaseous product was determined by the difference. A comparison can be made between the volatile content produced (Table 4.5) during pyrolysis using a TGA and the combined yields of the bio-oil and gas from each sample (Table 4.6) produced using a bench-scale reactor. It was shown that the yields of both products during pyrolysis were close to the volatile content data of each sample with a discrepancy of about 79%. The total bio-oil and gas yields were always below the volatile content produced in proximate analyses. Table 4.5 Product yield comparison of the pyrolysis of wheat, oat, flax and barley straws using a TGA Yield Moisture Wheat Straw Oat Straw Flax Straw Barley Straw Volatiles Char + Ash 5.6 68.8 25.6 4.6 69.1 26.3 4.5 74.4 21.1 3.6 73.2 22.6 Table 4.6 Product yield comparison of the pyrolysis of wheat, flax, oat and barley straws using bench scale reactor Yield Bio-char Bio-oil Gaseous product Wheat Straw Oat Straw Flax Straw Barley Straw 36.2 40.5 23.3 32.7 40.0 27.3 78 32.1 33.5 27.7 30.7 40.2 35.8 The results suggest that the volatile products (liquid and gas) produced during pyrolysis contained a higher oxygen percentage (less carbon in total) compared to the volatile materials produced from the proximate analysis procedure. This discrepancy may have been a result of differing temperature conditions of the processes and can also be explained by analyzing the biochar product resulting from pyrolysis as shown in Table 4.5. The elemental analyses of the biochar produced during pyrolysis still contained about 2-3% hydrogen and as high as 20% oxygen. The hydrogen, oxygen, and carbon in the char can be used to produce more volatile materials (bio-oil and gaseous product) with the use of a catalyst, a faster heating rate, or perhaps a different pyrolysis temperature. The lower sulfur content of the bio-chars (Table 4.7) compared to those of the raw biomass samples (Table 4.3) and the GC-MS analysis results suggest that half of the sulfur may have been converted to sulfur dioxide (SO2) by oxidation, hydrogen sulfide (H2S) by reduction, and carbonyl sulfide (COS). In a previous study, Shao et al. [20] concluded that SO 2, H2S and CO gases can be produced under pyrolysis in the temperature range of 200°C to 500°C). Table 4.7 Elemental analysis of bio-char produced from each biomass sample Composition Wheat Straw Oat Straw Flax Straw Barley Straw Char Char Char Char C 67.87 66.42 71.73 72.61 N 3.02 1.26 2.32 1.73 H S 2.97 0.55 2.34 0.49 79 2.8 0.61 3.09 0.62 The various yields of the pyrolysis products, such as liquid, gas and char, from all biomasses are presented in Figure 4.4. The wheat straw biomass produced the maximum bio-oil yield of 40.2%, while the maximum gas yield of 30.9% was obtained from barley straw. The yields of the biochar products, were in the range of 32.1% to 36.2%. The difference between the product yields of each biomass may have been influenced by the cellulose, hemicellulose and lignin compositions of the raw biomasses (Table 4.1). However, there were no clear trends between the structural compositions of the biomasses and the product yields. Gas Fig. 4.4 Product yields (wt%) from pyrolysis of straw biomass The combined amount of bio-oil and gas products did not exceed the amount of volatile compounds that were produced during proximate analysis (Table 4.3) and from TGA pyrolysis (Table 4.5). Nonetheless, the ash composition data from the proximate analyses suggest that the higher the ash content of the biomass resulted in less volatile produced during pyrolysis. This may indicate that inorganic compound within biomass (ash) will not have a direct impact on the cracking of biomass into volatiles. 80 4.5.3 Catalytic Pyrolysis of Raw Biomasses Three different catalysts were used to examine their respective effects on product yields. As shown in Figure 4.5, the use of catalyst ZY-SS (#2) had the most significant effect in increasing the yields of bio-oil (by about 2%) and bio-char (by up to 8%). The use of catalyst #2 also had the most significant effect on the pyrolysis of flax straw, increasing its bio-oil yield up to 46.4%. In the pyrolysis of oat straw, the use of a catalyst decreased the gas yield; however, the yield of bio-oil had the highest increase (43.3%) with the use of catalyst ZY-NS (#1). The use of catalyst #1 also increased the bio-oil yield during the pyrolysis of barley straw (43.0%). Gas Fig. 4.5 Comparison of the product yields among the biomasses 4.5.4 Analysis of Pyrolysis Products Proximate and CHNS analyses were performed of the bio-char products from pyrolysis. The compositions of the proximate analyses and CHNS compositions between raw biomasses and bio-chars from pyrolysis reactions using different feedstocks and catalysts were compared. As can be seen in Table 4.8, bio-chars contained higher percentages of carbon. The pyrolysis broke 81 down the chemical bonding of biomasses and released more hydrogen and oxygen from their chemical structure, forming volatile material that later condensed as bio-oil, with the remainder as non-condensable gases. The carbon percentage increased from about 44% up to about 72% (char from barley straw). Table 4.8 Proximate and CHNS analysis of bio-char products after pyrolysis Sample Catalyst Moisture Volatile Wheat straw - 3.67 26.65 Flax straw - 2.83 18.23 Oat straw Barley straw - - 4.93 2.26 20.78 23.14 Fixed Carbon Ash C H N S 57.55 12.13 67.87 2.97 3.02 0.55 74.40 4.55 63.08 69.09 11.22 66.42 2.34 1.26 0.49 5.52 71.73 2.80 2.32 0.61 72.61 3.09 1.73 0.62 Wheat straw #1 2.11 13.99 53.78 30.11 61.34 2.79 1.19 0.46 Flax straw #1 1.83 7.50 75.41 15.25 62.89 2.77 2.03 0.57 Wheat straw #2 1.79 13.18 54.49 30.55 59.82 2.87 1.29 0.61 Flax straw #2 1.49 13.97 72.97 11.57 60.12 2.63 1.48 0.45 Wheat straw #3 1.64 8.08 58.67 26.62 67.32 2.46 1.14 0.52 #3 0.84 14.18 68.40 Oat straw Barley straw Oat straw Barley straw Oat straw Flax straw Barley straw #1 #1 #2 #2 #3 #3 3.44 2.02 2.82 2.35 8.70 20.87 14.10 20.97 2.60 13.32 2.29 24.90 68.77 62.61 64.70 61.78 52.44 59.23 19.09 64.11 2.29 1.06 0.56 14.51 65.38 3.12 2.02 0.63 18.38 64.36 2.37 1.15 0.54 14.90 64.93 3.33 1.40 0.70 31.64 62.44 2.55 1.00 0.59 16.58 58.67 2.23 1.66 0.42 13.58 69.27 2.97 1.28 0.62 Bio-oil produced from pyrolysis often contains water in the range of 15% to 30% [20, 21]. The high amount of water is considered one drawback of pyrolytic oil. In our study, the water contents of the bio-oils produced from pyrolysis were determined using a Karl-Fischer titration system and are presented in Table 4.9. The water contents of the products of all reactions ranged 82 from 11.66% to 17.56%. When there was no additional catalyst present during the reaction, water formed during the reaction and increased the initial value of water content of the biomass, ranging from 2.75 to 5.3% to about 14.2% to 17.6%. The use of a catalyst, however, showed very little influence on water formation and only reduced it by about 2%. Table 4.9 Water content of bio-oil produced from pyrolysis Sample Wheat straw No Catalyst Oat straw Flax straw Barley straw 17.6 Water content of bio-oil (% weight) KS23-ZY-NaSi KS20-Sisol HZY Sisol 15.5 15.4 15.5 17.1 16.4 14.2 16.5 13.1 11.7 15.7 12.1 12.5 16.0 14.6 12.3 Table 4.10 Heating value (MJ/kg) of biomasses and its products Sample Wheat straw Oat straw Flax straw Barley straw No catalyst 23.40 23.99 27.67 27.70 KS20-Sisol 21.48 22.97 25.27 24.64 Raw biomass Bio-char product KS23-ZY-NaSi HZY Sisol Bio-oil product 17.97 21.43 21.57 17.51 21.44 21.44 18.62 23.54 26.51 17.95 24.68 26.20 No catalyst 27.16 27.08 27.39 27.47 KS20-Sisol 27.26 27.66 27.42 27.93 KS23-ZY-NaSi HZY Sisol 27.07 26.97 27.73 27.13 83 27.39 27.41 27.47 27.93 Table 4.10 presents a clear comparison of the heating values among the raw biomasses and their pyrolysis products. As can be seen for all of the samples, the heating values of the bio-char and bio-oil products were higher (up to a 55% increase) compared to their original heating values (raw biomass). However, when the heating value of the product from the non-catalytic pyrolysis was compared to the product from the catalytic pyrolysis, there was no significant increase or reduction. The GC-MS analyses showed that most of the chemicals produced from pyrolysis contained 1, 2 or 3 oxygen atoms. By performing a further upgrading reaction, such as hydrodeoxygenation, the amount of oxygen can be reduced from these chemicals. It was also observed that most components that were produced contained an aromatic ring (benzene, furan and cyclopentane) and had carbon atoms ranging from C2 to C8. Table 4.11 shows 18 components that were produced in all reactions, with or without catalyst; however, there was a variation in terms of their concentrations (% area). For example, a few components, including 1-hydroxy-2-butanone, furfural, acetoxyacetone, 2-cyclopenten-1-one, guaiacol and 2,6-dimethoxyphenol, had higher concentrations, ranging from 4% to 19% while the rest of the components had concentrations of mostly below 2%. Catalyst ZY-SS (#2) doubled and almost tripled the production of furfural in the pyrolysis of wheat and flax straws, respectively. 84 Table 4.11 GC-MS analysis of bio-oil produced from straw biomass with and without catalyst Sample & catalyst in the reaction Sample and catalyst used in the pyrolysis Wheat Oat Flax Barley Wheat Oat Flax Barley Wheat Oat Flax Barley Wheat Oat Flax Barley C2H4O2 1.28 2.89 1.06 1.74 1.58 1.95 0.86 1.45 - - - - 1.10 1.89 0.94 1.03 trans-1,3-Cyclopentanediol C5H10O2 0.43 2.59 1.47 1.93 1.70 1.81 0.80 1.00 2.57 1.85 0.71 0.70 0.85 4-Hydroxy-4-methyl-2pentanone C6H12O2 0.94 3.68 2.65 1.93 1.61 Acetoxyacetone C5H8O3 4.10 5.22 5.43 Chemical Formula 1-Hydroxy-2-butanone C4H8O2 Acetic Acid - - 5.52 11.07 9.13 10.64 2.74 5.44 C6H8O 1.81 C6H8O Tetrahydro-2-furanmethanol o/m/p-Cresol Furfural 2-Furanmethanol 2-Methyl-2-cyclopenten-1-one C5H4O2 4-Ethylguaiacol 2,6-Dimethoxyphenol 8.22 7.65 8.32 5.16 8.09 11.09 4.25 10.16 19.26 5.16 16.91 2.79 1.39 1.86 6.17 7.58 4.08 6.69 15.41 11.31 7.96 6.48 8.15 4.81 1.31 6.14 3.04 4.16 2.51 3.01 5.56 7.67 5.50 5.82 3.65 5.19 3.54 2.40 2.66 2.35 2.58 1.81 2.31 3.39 1.97 3.95 3.52 2.13 2.18 1.79 2.14 2.64 3.04 2.45 2.50 3.22 2.88 1.94 2.71 2.95 3.25 2.80 2.24 2.78 C5H10O2 1.58 3.83 3.11 2.07 1.96 2.72 2.44 4.00 3.21 2.42 3.63 2.16 2.47 C7H8O 5.14 1.76 3.54 C8H10O2 2.03 1.39 C8H10O3 6.85 C6H8O2 C7H8O2 C9H12O2 1.75 3.51 5.42 9.88 1.09 4.71 4.61 4.68 2.05 4.35 5.69 9.54 4.86 with catalyst #3 (HZY-SS) 3.26 C5H6O2 C6H6O 4-Methylguaiacol (Creosol) 7.55 with catalyst #2 (ZY-SS) 12.81 Phenol Guaiacol 9.23 with catalyst #1 (ZY-NS) 11.81 C4H6O2 2-Cyclopenten-1-one 5.49 - 6.32 4-Hydroxybutyric acid lactone 3-Methyl-2-cyclopenten-1-one - 4.29 0.98 3.90 2.56 3.28 2.24 3.97 0.54 3.62 3.21 3.67 2.63 2.23 3.63 1.88 3.37 1.57 3.02 3.39 5.51 7.22 5.66 4.49 5.32 9.13 15.59 7.75 10.89 6.76 11.24 2.55 2.23 1.52 1.55 1.96 2.10 10.06 3.96 1.97 2.77 1.38 8.39 3.33 4.40 1.52 0.39 6.58 7.49 85 1.22 3.85 2.44 3.11 6.62 7.72 1.19 2.75 1.29 2.76 5.28 2.26 5.60 6.94 1.78 2.63 4.26 3.93 6.92 11.14 5.40 16.23 1.84 3.21 1.65 3.95 3.09 4.12 3.30 1.32 4.01 4.08 5.48 2.84 3.55 1.48 1.20 9.62 4.19 5.01 3.69 6.60 4.09 3.96 5.54 1.96 6.36 2.84 4.13 5.60 1.36 4.16 4.14 3.88 4.97 2.45 3.16 3.93 4.32 5.49 0.96 1.60 0.90 5.13 6.89 3.31 9.94 8.80 2.16 6.40 0.92 3.44 2.37 1.88 4.97 4.07 3.88 2.30 6.27 3.66 2.42 2.87 1.25 7.51 10.08 1.95 1.31 6.79 0.94 2.96 6.26 2.05 6.55 Table 4.12 Micro-GC analysis of gas produced from wheat straw biomass with and without catalyst H2 CH4 CO CO2 No catalyst 0.9% 1.2% 28.8% 68.1% KS20-Sisol 1.8% 1.4% 34.7% 61.3% KS23-ZY-NaSi HZY Sisol 1.7% 1.3% 1.5% 1.5% 31.0% 30.6% 64.9% 65.5% The comparison of hydrogen (H2), methane (CH4), carbon monoxide (CO) and carbon dioxide (CO2) compositions of the pyrolysis reactions of wheat straw are shown in Table 4.12. While the production of H2 and CH4 are only about 3% of the total amount, CO and CO2 accounted for most of the produced gases (around 96%). The use of catalyst has changed the distribution of gas compositions as can be seen in Table 4.12. The production of H 2 gas increases when catalyst was used. The increase of CH4 gas about 11.5% to about 24.6% was also noticed for the catalytic reaction. The increase of H2 and CH4 gases can be related to the composition of bio-oil and also the yield of bio-char. It can be seen from Table 4.11, when the catalyst was used, there is an increase in most oxygen rich compound. When rich oxygen compounds was formed, cracking of cellulose or lignin molecule will release more H2 and CH4 in the gas phase. In addition, there is also an increase of bio-char yield when catalyst was used (Fig. 4.5). From the analysis of the products, we found that there are no significant change on C, H and O composition of bio-char. However interesting result was found on bio-oil composition produced from pyrolysis with and without catalyst. The use of zeolite based catalyst increased the production of rich-oxygenated compound (with benzene and cyclopentane ring) that we were 86 trying to avoid despite the increase of the bio-oil yield. The use of a catalyst, however, showed very little influence on water formation and only reduced it by about 2%. From calorimeter measurement, the heating values of the bio-char and bio-oil products were higher (up to a 55% increase) compared to their original heating values (raw biomass). From gas product analysis, we found that there is an increase of H2 and CH4 gas when catalyst was used that can be related to the composition of bio-oil and also the yield of bio-char. 4.5 Conclusions The results of the proximate analyses of raw biomasses can be used to predict product yields produced during pyrolysis, with the total yields of bio-oil and gaseous product consistently lower than the volatile content. The ash composition values from the proximate analyses suggest that a higher ash content suppressed the production of gaseous product. The pyrolysis of four different biomasses was performed both with and without a catalyst. It was observed that the bio-oil yield during the pyrolysis increased with the use of a zeolite-based catalyst. The use of a zeolite ZYSS catalyst showed the most significant effect in increasing the bio-oil yield during the pyrolysis of wheat and flax straws, increasing its bio-oil yield up to 46.4%, whereas the use of zeolite ZYNS showed the most significant effect during the pyrolysis of oat and barley straws. From GCMS analysis of bio-oil, it was found that catalyst ZY-SS (#2) doubled and almost tripled the production of furfural in the pyrolysis of wheat and flax straws. 87 4.6 Acknowledgement The authors wish to acknowledge the financial support of the Natural Sciences and Engineering Research Council of Canada (NSERC) and the Ministry of Higher Education of the Kingdom of Saudi Arabia. We would like to thank Aurore Coussirat-Bourg for her contribution in the experimental part of this study. 88 4.7 References [1] Mahinpey, N., Murugan, P., Mani, T. & Raina, R., Analysis of bio-oil, bio-gas and bio-char from pressurized pyrolysis of wheat straw using a tubular reactor. Energy and Fuels, 23, pp. 2736- 2742, 2009. [2] Scott, D.S. & Piskorz, J., The continuous flash pyrolysis of biomass. Canadian Journal of Chemical Engineering, 62, pp. 404-412, 1984. [3] Mohan, D., Pittman, C.U. & Steele, P.H., Pyrolysis of Wood/Biomass for Bio-Oil: A Critical Review. Energy Fuels, 20, pp. 848-889, 2006. [4] Bicho, P.A. & Stumborg, M.A., Wheat straw: a viable fibre source for Canada? Pulp & Paper Canada, pp. 146-149, 1998. [5] Wang, S., Guo, X., Wang, K. & Luo, Z., Influence of the interaction of components on the pyrolysis behavior of biomass. Journal of Analytical and Applied Pyrolysis, 91(1), pp. 183189, 2011. [6] Beis, S.H., Onay, Ö. & Koçkar, Ö.M., Fixed-bed pyrolysis of safflower seed: influence of pyrolysis parameters on product yields and compositions. Renewable Energy, 26 (1), pp. 21-32, 2002. [7] Vamvuka, D., Bio-oil, solid and gaseous biofuels from biomass pyrolysis processes - an overview. International Journal of Energy Research, 35(10), pp. 835-862, 2011. [8] Haykiri-Acma, H., Yaman, S. & Kucukbayrak, S., Effect of heating rate on the pyrolysis yields of rapeseed. Renewable Energy, 31(6), pp. 803-810, 2006. [9] Biagini, E., Fantei, A. & Tognotti, L., Effect of the heating rate on the devolatilization of biomass residues. Thermochimica Acta, 472(1-2), pp. 55-63, 2008. 89 [10] Mani, T., Murugan, P., Abedi, J. & Mahinpey, N., Pyrolysis of Wheat Straw in a Thermogravimetric Analyzer: Effect of Particle Size and Heating Rate on Devolatilisation and Estimation of Global Kinetics. Chemical Engineering Research and Design, 478, pp. 1-7, 2010. [11] Hosoya, T., Kawamoto, H. & Saka, S. Pyrolysis behaviors of wood and its constituent polymers at gasification temperature. Journal of Analytical and Applied Pyrolysis, 78 (2), pp. 328-336, 2007. [12] Mu, W., Ben, H. X., Ragauskas, A. & Deng, Y. L., Lignin pyrolysis components and upgrading – technology review, Bioenergy Res., 6, pp. 1183-1204, 2013. [13] Kim, J. S., Production, separation and applications of phenolic-rich bio-oil – a review, Bioresour. Technol., 178, pp. 90-98, 2015. [14] Effendi, A., Gerhauser, H. & Bridgwater, A.V., Production of renewable phenolic resins by thermochemical conversion of biomass: a review, Renew. Sust. Energy Rev., 12, pp. 20922116, 2008. [15] Zhang, Z., Lu, Q., Ye, X., Li, W., Hu, B. & Dong, C., Production of phenolic-rich bio-oil from catalytic fast pyrolysis of biomass using magnetic solid based catalyst, Energy Convers.. & Manage., 106, pp. 1309-1317, 2015. [16] Önal, E., Uzun, B. B. & Pütün, A. E., Bio-oil production via co-pyrolysis of almond shell as biomass and high density polyethylene, Energy Convers. Manage., 78, pp. 704–710, 2014. [17] Adapa, P., Tabil, L. & Schoenau, G., Compaction characteristics of barley, canola, oat and wheat straw. Biosystems Engineering, 104, pp. 335-344, 2009. 90 [18] Buranov, A. U. & Mazza, G., Lignin in straw of herbaceous crops. Industrial Crops and Products, 28, pp. 237-259, 2008. [19] Mussatto, S. I. & Teixeira, J.A., Lignocellulose as raw material in fermentation processes. Current Research, Technology and Education Topics in Applied Microbiology and Microbial Biotechnology, pp. 897-907, 2010. [20] Oasmaa, A. & Czernik, S., Fuel Oil Quality of Biomass Pyrolysis Oils - State of the Art for the End Users. Energy & Fuels, 13, pp. 914-921, 1999. [21] Bridgewater, A. V., Review of fast pyrolysis of biomass and product upgrading. Biomass and Bioenergy, 38, pp. 68-94, 2012. 91 CHAPTER 5: CATALYTIC HYDRODEOXYGENATION OF GUAIACOL AS LIGNIN MODEL COMPONENT USING NI-MO/TIO2 AND NI-V/TIO2 CATALYSTS 5.1 Presentation of the Article In order to understand the behavior of the bio-oil component during the upgrading process, a hydrodeoxygenation (HDO) reaction was performed for a main bio-oil component called guaiacol. In this work, we focused on catalyst screening and the distribution of the product formed by the reaction. The objective of the study was to find a suitable catalyst that could be used to transform highly oxygenated components such as guaiacol into lesser oxygenated compounds, such as benzene, cyclohexane and phenol. The catalysts were prepared using an incipient wetness impregnation technique with an aqueous solution of the respective metal precursors. The experiments were performed in a 30 mL batch reactor made of stainless steel (Swagelok). Reaction parameters such as temperature, reaction time and hydrogen gas to guaiacol ratio were varied using the optimized catalyst. Blank reactions (without hydrogen, without catalyst, without hydrogen and catalyst, and bare support) were also carried out under identical conditions. 92 Catalytic Hydrodeoxygenation of Guaiacol as Lignin Model Component Using Ni-Mo/TiO2 and Ni-V/TiO2 Catalysts Aqsha Aqsha , Lakshmi Katta , Nader Mahinpey Department of Chemical and Petroleum Engineering, Schulich School of Engineering, The University of Calgary, Calgary, Canada, AB T2N 1N4 This article was published in Catalysis Letters, 145(6), pp. 1351-1363, 2015. 5.2 Abstract Catalytic hydrodeoxygenation of guaiacol (2-methoxy phenol), an oxygen rich lignin model compound, has been investigated aiming at the elucidation of active catalysts for production of deoxygenated products. All catalysts were synthesized by impregnation technique. Screening of different catalysts unveiled Ni-Mo/TiO2 as the most active catalyst. Ni-Mo/TiO2 was compared with Ni-V/TiO2 to understand the influence of promoter (Mo and V); nevertheless, Ni-V/TiO 2 produced no positive effect compared to Ni-Mo/TiO2. Structural investigation was performed using XRD, TEM, BET surface area and TPD measurements. As evidenced by XRD and TEM, all catalysts composed of nano-sized particles. Ni-Mo/TiO 2 showed small sized finely dispersed particles, on the other hand large particles and loss of dispersion noted for Ni-V/TiO 2 and Ni/TiO2, respectively. BET surface area measurement depicted Ni-Mo/TiO2 catalyst presented high surface area with optimal mesopores than Ni-V/TiO 2. The NH3-TPD data revealed that NiMo/TiO2 acidic strength is higher than the Ni-V/TiO2. Influence of catalysts and reaction variables were investigated using both Ni-Mo/TiO2 and Ni-V/TiO2 to determine their potential 93 role on guaiacol conversion. High activity of Ni-Mo/TiO 2 during wide range of conditions could be attributed to: (i) small sized and finely dispersed active metal particles, (ii) more mesopores, and (iii) high acidic strength. Water as a solvent, showed no impact on Ni-V/TiO 2 performance, while on Ni-Mo/TiO2, guaiacol conversion and phenol selectivity inclined to decrease. 5.3 Graphical Abstract Keywords: Hydrodeoxygenation, Lignin, Ni-based catalyst, Impregnation, Guaiacol 94 5.4 Introduction Due to depletion in petroleum resources combined with increasing energy demand and inevitable environmental concerns from fossil fuels, it is vital to develop economical and energy-efficient processes for the sustainable fuel production. To achieve this, several approaches have dragged preponderance attention [1–6]. Utilization of biomass to produce bio-fuels is one of the simplest and cheapest methodologies for the development of sustainable energy [7]. Nonetheless, presence of excess oxygen in liquid products derived thermochemical processes of lignocellulose biomass imparts poor stability and low heating values, which made these oils unsuitable to be used as transportation fuel [8]. Bio-oils can be upgraded into liquid transportation fuel by numerous processes: hydrotreating (hydrodeoxygenation), zeolite cracking, and emulsification with diesel fuel [9]. Amongst, catalytic hydrodeoxygenation (HDO: hydrogenolysis of C-O linkage in presence of hydrogen) of bio fuels became a convincing approach to mimic them as crude oils [2, 10]. HDO pathway normally requires a bi-functional catalyst consisting of a reducible site (active metals) and an acidic center (support or metal-support interface) together facilitate the hydrogenation and a proper alignment for the hydrogenolysis processes (also catalyzes dehydration, alkylation, isomerization and condensation reactions) [11]. Recently, there has been much progress in HDO of pyrolysis bio oils employing different types of catalysts such as, zeolites [12], sulfided metal systems (Co, Co–Mo, Ni–Mo, etc.) [13–16], noble metals (Pd, Rh, Pt, etc.) [16–20] and metal phosphides [21] nevertheless, some of these systems exposed with potential limitations. In case of zeolites, although no hydrogen is required for the process, their 95 capability for deoxygenation of lignocellulose phenolic compounds is limited due to poor hydrocarbon yield and easy decomposition by extensive coke deposits or dealumination by hydrolysis [2, 16]. Sulfide (Ni-Mo, Co-Mo) catalysts require H 2S to maintain HDO activity which contaminates bio-oil components with sulphur and potentially problematic due to waterinduced catalyst deactivation and coke accumulation, while noble metals are expensive and nonabundant [22]. Exploring inexpensive metals combinations (at least one of the metals are active for hydrogenation) together with acidic support could alleviate deoxygenation activity [11, 23– 25]. It is worth noting that HDO processes of both continuous and batch reactions are majorly carried out under the conditions typical of crude oil hydrotreatment (50 bar pressure of pure H 2). Bridgwater et al. estimated that 62 kg of H2 is required to deoxygenate per tonne of bio-oil [26]. Escalona et al. studied La1−x Ce x NiO3 perovskite reactivity using 5 MP H2 pressure that yields cyclohexanol and methyl cyclohexanol as major products [27]. However, it is desirable to avoid hydrogenation of aromatics in the bio-oils since the increase in H2 consumption would decrease the octane number. Also, minimizing hydrogen demand is important for reducing the cost of biomass derived liquids. Reports on production of aromatics utilizing the mild operational conditions hitherto are scarce and to be reconciled. Olcese et al. [28] studied effect of pyrolysis gas at atmospheric pressure using guaiacol, in which phenol and cresol were identified as major products analogous to our observation. Wu et al. studied HDO of guaiacol at ambient conditions deploying nickel phosphide catalysts [29]. Zhao et al. [25] illustrated HDO of guaiacol under 96 ambient pressure utilizing silica-supported transition metal phosphide catalysts (i.e., Ni 2P, Fe2P, MoP, Co2P, and WP) led to phenol and benzene selectively. The development of water tolerant catalysts is another critical challenge as water is a major byproduct in the bio-oil production [30]. Therefore, studying the conversion of guaiacol in presence of water is more practical than in organic solvents. Based on the aforementioned background, current study is focused on HDO of guaiacol carried at moderated pressures and temperatures (100 psi and 300 °C). Guaiacol is an important building block consensus preferred choice as a prototype compound that represent lignin and many of its derivatives [31]. The objectives of the current work are: (i) screening of several catalysts varying metals and support to achieve prominent catalyst, (ii) relating the structural behavior of catalysts with catalytic performance to understand structure–activity relationships, (iii) studying the effect of reaction conditions, and (iv) understanding the effect of water presence on catalytic performance. 5.5 Experimental 5.5.1 Catalyst Preparation All the chemicals were purchased from Sigma-Aldrich and Alfa-Aesar. The catalysts were prepared by incipient wetness impregnation technique with an aqueous solution of the respective metal precursor: adequate amounts of nickel nitrate hexahydrate (Ni(NO 3)2·6H2O), ammonium heptamolybdate ((NH4)6Mo7O24), Ammonium metavanadate (NH4VO3), cupper nitrate trihydrate (Cu(NO3)2·3H2O) and iron nitrate nonahydrate (Fe(NO3)3·9H2O) are used to obtain nickel (Ni, 10 wt%), molybdenum (Mo, 10 wt%), vanadium (V, 10 wt%), copper (Cu, 10 wt%) and iron 97 (Fe, 10 wt%) oxides, respectively. First, suspension of commercial support in the metal precursor solution was stirred on a hot plate at 70 °C until complete evaporation attained. Impregnated sample was oven dried at 110 °C for 12 h. The dried samples were subsequently calcined in air at 500 °C for 3 h to obtain final catalysts NiO-MoO3 (Ni-Mo), NiO-V2O5 (Ni-V), NiO-CuO (NiCu), NiO-Fe2O3 (Ni–Fe) deposited onto ZrO2, TiO2, Al2O3, Fe2O3 supports. 5.4.2 Characterization Studies An energy-dispersive X-ray (EDX) analyzer was used for the elemental analysis of the samples. The EDX analyzer attached to the scanning electron microscope allowed for the determination of the elemental composition at any point of interest on the catalyst surface. Powder X-ray diffraction data were acquired on a Rigaku Multiflex X-ray Diffractometer equipped with Cu Kα radiation (λ = 0.1548 nm) to identify the phase constitutions in the samples. The step size and the time per step were respectively, fixed at 0.02° and 1 s in the range 2θ of 20°–80°. Crystallite size is measured using Scherrer’s equation with the help of highest intense peaks. Crystalline phases were identified using reference data ICDD PDF files. The BET surface area, pore volume and pore size distribution of the samples were determined using ASAP 2020 equipment. Prior to the experiment, samples were degassed at 200 °C for 2 h to remove any surface adsorbed residual moisture. The TEM studies were made on a Tecnai F20 instrument equipped with a slow-scan CCD camera and at an accelerating voltage of 200 kV. Samples for TEM were prepared by dispersing them ultrasonically in ethyl alcohol and after dispersion a droplet was deposited on a copper grid supporting a perforated carbon film and allowed to dry. The dried specimen is used for the examination under the microscope. NH3-TPD data was collected on a Micromeritics 98 Autochem 2910 Automated System. Prior to the analysis, the samples (approximately 50 mg) were treated with pure Helium gas at 200 °C for 2 h. After treatment the temperature was lowered to 100 °C and the sample was saturated with anhydrous NH3 (3.99 % in He) at a flow rate of 30 mL min−1 for 1 h. Subsequently, the samples were purged again with 30 mL min −1 of pure helium (UHP helium 99.999 %) for 2 h in order to remove any weakly bound (physisorbed) NH3 and then sample temperature was reduced to 50 °C. Temperature was raised to 800 °C at 5 °C min−1 ramping rate and then desorbed gas was monitored after baseline established. 5.4.3 Reaction Study: HDO of Guaiacol HDO of guaiacol is carried out in a batch reactor. All the samples are pre-activated at 350 °C for 2 h at 5 °C min−1 under one atmosphere H2 pressure. The experiments were performed in a 30 mL batch reactor of Swagelok made from stainless steel. 10 wt% catalyst to guaiacol were loaded in the reactor. The mixture was then heated to desired temperature using subcritical bath in the requisite hydrogen atmosphere. After completing the reaction, the reactor was placed in a water bath to bring the reactor to room temperature (~25 °C). The reaction mixture was subsequently analyzed by GC–MS (Agilent 7890) equipped with a flame ionization detector (FID). HP-5MS capillary column (diameter 0.25 mm, thickness 25 µm, length 30 m) was used for separation and the column temperature was 100 °C. The major products detected in GCMS were: cyclohexanol, cyclohexanone, phenol, anisole, cresol, catechol dimethyl ether (CDME), creosol and catechol. Reaction parameters such as, temperature, reaction time and H 2/guaiacol ratio are varied using optimized catalyst. Blank reactions (without hydrogen, without catalyst, without hydrogen and catalyst, and bare support) are also carried under identical conditions. 99 Conversion was checked to be negligible with sole guaiacol (3.8 %) disclosing the important role of catalyst and hydrogen. We also tested non reduced sample to check the effect of reduction on catalytic performance which is found to be lower than the reduced samples. All the experiments were at least performed in duplicate and the results were well reproducible within ±3 % range of conversion. Conversion, selectivity and yield were calculated on the basis of the following Eqs. (1–3) [32]: Conversion (%) = [(guaiacolin − guaiacolout) / guaiacolin] × 100 (5.1) Selectivity (%) = (the mass of specific products/the mass of all products) × 100 (5.2) Yield (%) = Conversion (%) × Selectivity (%) / 100 (5.3) 5.6 Results and Discussions 5.5.1 Catalyst Properties Wide-angle XRD results of both fresh and reduced Ni-Mo/TiO 2, Ni-V/TiO2 and Ni/TiO2 samples were compiled in Figs. 5.1 and 5.2, respectively. For comparison, spectrum of titania support was also incorporated in the X-ray diffraction patterns of metal catalyst. The XRD patterns of all the samples along with pristine titania exhibited intense peaks at 2θ of 25.2°, 36.9°, 37.8°, 38.5°, 48.1°, 53.8°, 55.1°, 62.6°, 68.7°, 70.3°, 75.0° indexed as (101), (103), (004), (112), (200), (105), (211), (204), (116), (200), (215) reflections of TiO2 anatase (ICDD 75-1537) [33] and weak peaks centered at 2θ of 27.4°, 36.1°, 39.1°, 41.2°, 56.6° indexed as (110), (101), (200), (111), (220) corresponding to rutile phase of TiO2 (ICDD 82-0514) [34]. Compared to pure titania, metal deposited titania peaks slightly shifted towards left side indicating that high ionic radius Ni 100 metal could slightly incorporated into the support oxide. Deposited titania peaks are much sharper and intensified than the pure titania implying sintering of the particle during the deposition/calcination treatment. Crystallite size for bare titania is 17.4 nm (Table 5.1), after the deposition crystallite size of titania for Ni/TiO 2, Ni-Mo/TiO2, and Ni-V/TiO2 found to be slightly increased except for Ni-Mo/TiO2 and maintained same with reduction treatment. Fresh samples X-ray reflections at 2θ of 37.2°, 43.3°, and 62.9° corresponding d spacing 2.412, 2.088, 1.477 Å are characteristic of cubic NiO phase (ICDD 78-0423) [35]. MoO3 diffraction patterns corresponding to the orthorhombic phase were observed at 2θ of 23.6°, 26.8°, 52.3° and NiMoO 4 phase formation was identified at 2θ of 28.8°, 32.5°, and 43.9° [36, 37]. Peaks at 2θ of 20.7°, 22.1°, 26.6°, 31.5°, 34.8° are characteristic of V2O5 (ICDD 72-0598). Peaks displayed at 2θ of 44.2°, 51.8°, 76.3° for reduced catalysts (Fig. 5.2) corresponding d spacing 2.034, 1.762, 1.246 Å are typical of Ni metal (ICDD 04-0850) [38]. 101 * VNiOO ; 30 Ni/TiO 2 * # TiO 2 40 50 2Theta ( ) 0 60 )A (2 15 16 )A 00 )A (1 (2 )A X (2 04 10 01 ) (1 (1 20 ^* Ni-V/TiO 2 4 (1 05 ) (2 A 11 ) A (2 20 )R )R X * )A X XX ^# 00 ^ 5 ^ NiMoO (2 # 2 (101) R (103) A (004) A A ((112) 20 ( 1 0) R 11 )R # A Intensity (a.u.) X # MoO 3 ; Ni-Mo/TiO 2 70 Fig. 5.1 X-ray diffraction patterns of anatase, Ni/TiO2, Ni-V/TiO2 and Ni-V/TiO2 before reduction The crystallite size data of the Ni metal calculated using the diffraction line at 2θ of 44.2° is presented in Table 5.1. As given in the table, the crystallite size of Ni metal found to be much smaller for Ni-Mo/TiO2 (8.2 nm) followed by Ni-V/TiO2 (10.3 nm) and Ni/TiO2 (16.3 nm). Based on crystallite size data it is clear that addition of promoter (Mo and V) prevents the agglomeration of Ni metal and promotes the dispersion, this observation is also supported by TEM and BET discussed in subsequent section. No peaks due to Mo and V (metal form) for the 102 reduced samples (Fig. 5.2) in the investigated region indicate either they are in the amorphous state or present as highly dispersed small crystallites beyond the XRD detection. Ni-Mo/TiO 2 Ni-V/TiO 2 $ Ni 40 50 2Theta ( ) 0 60 )A (2 15 16 (2 ) A 00 )A (1 )A 04 (2 5) A 11 (2 ) A 20 )R (2 (1 0 )A 00 (2 (101) R (103) A (004) A (112) A (2 00 (1 ) R 11 )R )R 10 )A 01 (1 (1 30 TiO 2 $ $ Intensity (a.u.) 20 Ni/TiO 2 70 Fig. 5.2 X-ray diffraction patterns of anatase, Ni/TiO2, Ni-V/TiO2 and Ni-V/TiO2 after the reduction TEM study was conducted to see the distribution of the metal particles over the support oxide (Fig. 5.3). For comparison, TEM of bare titania is also presented. All the TEM images show the catalyst material composed of highly crystalline nanosized particles of homogenous distribution. Pure TiO2 image shows the particles ranging between 20 and 30 nm with cubic and hexagonal 103 shapes. After metals (Ni, Mo, V) deposition, enlarged support particles formed without any clear shape. Ni-Mo/TiO2 and Ni-V/TiO2 images showed at least more than one different sized particles other than titania deduced two kinds of metals presence. All the metals particles are spherical in shape and homogenously distributed around the titania particle. It is clear from TEM of Ni/TiO 2 sample that metals strongly attached to titania support and formed agglomerated clusters, when V and Mo added to the Ni/TiO2, uniformly distributed particles around titania support are noticed. Particles of Ni-Mo/TiO2 are markedly smaller and smooth, while Ni-V/TiO2 showed bigger particles and composed of coarse boundaries. Few moiré patterns (indicated with arrows) appeared due to lattice mismatches between the layers. Table 5.1 Structural properties of metal–supported catalysts Elemental crystallite size (nm)a Sample TiO2 Ni-V/TiO2 Ni-Mo/TiO2 a Ni metal Anatase 10.3 18.9 - 8.2 17.4 16.8 Composition (wt %)b Ni Mo/V 12.4 9.1 - 11.4 - 11.6 Total acidity (peak area)c 0.41 1.10 1.13 XRD Sherrer’s equation; bEDX analysis; cTPD-NH3 desorption 104 20 nm Ni/TiO2 TiO2 Ni-Mo/TiO2 Ni-V/TiO2 Fig. 5.3 TEM image of TiO2, Ni/TiO2, Ni-V/TiO2 and Ni-Mo/TiO2 catalysts The elemental compositions of Ni-Mo/TiO 2 and Ni-V/TiO2 catalysts are presented in Table 5.1. As shown in the table, the actual metal loading is close to the theoretical value of 10 %. Nitrogen adsorption–desorption isotherms of Ni-Mo/TiO2 and Ni-V/TiO2 catalysts along with bare support under fresh and reduced conditions are shown in Fig. 5.4. Pore size distributions (PSD) are presented as inset in the Fig. 5.4. All samples exhibited type IV isotherm (IUPAC classification) 105 with a plateau at high relative pressure indicating the mesoporous nature of the materials with limited microporosity, which is also supported by PSD. The changes in specific surface area, pore volume and average pore diameter before and after the reduction treatment are presented in Table 5.2. The reduction treatment of TiO2 support has no influence on surface area, nevertheless an increment was noted for Ni-Mo/TiO2 and Ni-V/TiO2. Compared to bare support, pore volume and pore diameter are higher for metal deposited support similar to the observation noted by Yang et al. [39]. Increased amount of micro (13 m2/g) and macropores (0.18 cm3/g), respectively manifested on Ni-Mo/TiO2 and Ni-V/TiO2 after reduction treatment. It has been reported that mesopores of the catalysts would improve mass and heat transfer in reactions, and, therefore, micro and macropores are of less significance in the conversion of guaiacol [40]. An increase in mesopore volume is observed for both Ni-V/TiO2 and Ni-Mo/TiO2, although average pore volume (exception for Ni-V/TiO2) and pore diameter decreased after the reduction. Increase in surface area of Ni-V/TiO2 and Ni-Mo/TiO2 under reduced conditions could be ascribed to an increase in the mesopore volume. Though Ni-Mo/TiO 2 (0.28 cm3/g) has same pore volume (meso + macro) to that of Ni-V/TiO2 under reduced conditions, Ni-Mo/TiO2 attained large amount of mesopores (0.14 cm3/g) as confirmed from the Fig. 5.4b, c, and Table 5.2. Pretreatment on Ni-Mo/TiO2 lead to an increase in micro and mesopores at the cost of macropores indicate shrinking of the pores and thus making easy metal accessibility. The decrease in the average pore volume comprehends the close proximity between the guaiacol, metal sites and acidic sites for effective guaiacol hydrodeoxygenation [30]. 106 200 1.4x10 -3 -3 TiO2 reduced 1.0x10 3 Pore Volume (cm /g) 3 Quantity Absorbed (cm /g) -3 150 100 8.0x10 -4 6.0x10 -4 4.0x10 -4 2.0x10 -4 0.0 0 40 80 0.0 6.0x10 -4 5.0x10 -4 3 Quantity Absorbed (cm /g) 150 4.0x10 -4 3.0x10 -4 2.0x10 -4 100 1.0x10 -4 0.0 160 0.4 0.6 200 Relative Pressure (p/p ) 0 0.8 Ni-Mo/TiO2 fresh 0 40 80 Pore diameter (nm) 120 1.0 (b) Ni-Mo/TiO2 reduced 3 Pore volume (cm /g) 200 0.2 120 Pore diameter (nm) 50 0 (a) TiO2 fresh 1.2x10 160 50 0 0.0 0.2 0.4 0.6 Relative Pressure (p/p ) 107 0 0.8 1.0 6.0x10 -4 5.0x10 -4 4.0x10 -4 3.0x10 -4 2.0x10 -4 1.0x10 -4 Ni-V/TiO2 fresh 3 Quantity Absorbed (cm /g) 150 100 (C) Ni-V/TiO2 reduced 3 Pore volume (cm /g) 200 0 40 80 120 Pore diameter (nm) 50 0 0.0 0.2 0.4 0.6 Relative Pressure (p/p ) 0 0.8 1.0 Fig. 5.4 Adsorption isotherms of fresh and reduced (a) TiO 2 , (b) Ni-Mo/TiO2 and (c) NiV/TiO2 catalysts. Inset: Pore size distributions Table 5.2 Structural properties of metal–supported catalysts Sample Surface area (m2/g)d TiO2 51 Ni-Mo/TiO2 43 Ni-V/TiO2 42 TiO2 49 Ni-Mo/TiO2 58 Ni-V/TiO2 51 Total pore volume (cm3/g)d Mesopore volume (cm3/g)d Fresh catalysts Average pore diameter (nm)d 0.16 0.08 12.5 0.30 0.12 28.1 0.27 0.09 Reduced catalysts 25.6 0.13 0.08 11.2 0.28 0.14 19.5 0.28 0.10 108 22.2 d BET analysis Ni-Mo/TiO2 Ni-V/TiO2 Ni/TiO2 TCD Signal (a.u.) TiO2 100 200 300 400 500 Temperature ( C) 0 600 700 Fig. 5.5 TPD-NH3 profiles of TiO2, Ni/TiO2, Ni-V/TiO2 and Ni-Mo/TiO2 after the reduction treatment In order to compare the acidity of samples, thermo-programmed desorption of ammonia (TPDNH3) profiles of all the samples after reduction are presented in Fig. 5.5, while the relative peak areas are tabulated in Table 5.1. The low (T < 200 °C) and high temperature (250–450 and 600 °C) desorbed peaks designated as weak, medium, and strong acid sites. The desorption profiles of both TiO2 and Ni/TiO2 reveal that a moderate amount of weak acid sites, as evidenced by the small desorption peaks centered at around 200 and 300 °C. For Ni-V/TiO 2 and Ni-Mo/TiO2, the desorption peaks at 200 °C amplified, and additionally two new peaks at 400 and 600 °C developed with similar desorption pattern. The desorption peaks of Ni-V/TiO 2 and Ni-Mo/TiO2 109 apparently demonstrates Ni, Mo and V incorporation together increases the overall strength of acidic sites [41]. From the table it could be realized that the order of acidic strength follows the trend: TiO2 < Ni/TiO2 < Ni-V/TiO2 < Ni-Mo/TiO2. It has been reported that the strong acidic sites are pivotal for the reaction wherein, guaiacol molecule could first adsorb on acidic sites through C-O bond and then strong acidic sites favors hydrogenolysis and dehydration reactions [42]. Based on acidic strength, it is probable that Ni-Mo/TiO2 could be an efficient HDO catalyst. 5.5.2 Catalytic Activity Under present reaction conditions, all the catalysts lead more than ten products. We only presented the species accounting for more than 1 % of the total products. Guaiacol conversion products are mainly determined by methoxy, hydroxyl and benzene ring. Based on literature and product distribution a reaction network is proposed in Scheme 1 [15]. As shown in the scheme, there are four major reactions pathways realized in the HDO of guaiacol: (1) direct hydrogenation of aromatic ring, (2) demethoxylation, (3) demethylation, and (4) transalkylation directing to a variety of products such as cyclohexanone, cyclohexanol, anisole, phenol, cresol, catechol, CDME and creosol. Trace amounts of benzene, cyclohexane, 1,2-cyclohexane diol, methyl cyclohexanone, methoxy cyclohexanone and cyclohexene are also noted. Demethylation of guaiacol furnishes catechol (Eq. 1(a) in Scheme 1) [43]. Guaiacol can transform to phenol in two routes: demethylation of guaiacol to catechol and then deoxygenation to phenol or direct demethoxylation of guaiacol to phenol (Eq. 1(b) and (c) in Scheme 1) [44]. 110 OCH3 OH OH (a) +H2 -CH4 OH (b) OH +H2 -H2O (1) (c) +H2 OCH3 OH OCH3 OH -CH3OH +H2 OCH3 (2) -H2O OH Inter CH3 transfer CH3 + -H2O OCH3 HO OH H3CO OCH3 OH OCH3 OH OH +3H2 O +2H2 +2H2 +H2 +H2 +H2 OH (4) (5) O -CH3OH O OH OH -CH3OH OCH3 (3) CH3 OCH3 + OCH3 Intra CH3 transfer OCH3 OH OH (6) OH (7) Scheme 1. Schematic representation of possible reaction pathways of HDO of guaiacol 111 The methoxy group that removed during the formation of phenol was identified in GCMS as methanol. Anisole is formed by the deoxygenation to remove OH group of guaiacol (Eq. 2 in Scheme 1) [45]. Intermolecular methyl transfer followed by deoxygenation results in o- and mcresol (Eq. 3 in Scheme 1). Intramolecular methyl transfer leads to catechol, CDME and creosol (methyl transfer to oxygen to form CDME and methyl transfer to carbon to form o-, m-, pcreosol) (Eq. 4 in Scheme 1). The transalkylated products are obtained by the acidic sites [46]. Hydrogenation of guaiacol/phenol leads to cyclohexanol and cyclohexanone (Eqs. 5, 6, 7 in Scheme 1). No observation of aromatic ring condensed products such as bicycles under present conditions confirming catalysts are mute to polymerization of guaiacol. Screening of catalysts varying promoter (V, Mo, Cu, Fe) and support (TiO 2, Al2O3, ZrO2, Fe2O3, SiO2) under nominal conditions (300 °C, 5 h and 100 psi hydrogen) towards HDO of guaiacol are presented in Fig. 5.6a, b. Promoter effect on Ni supported ZrO2 is presented in Fig. 6a. Among various bimetallic combinations Ni-Mo and Ni-V are prominent and showed guaiacol conversion ~42 and 23 %, respectively with major deoxygenated products. Effect of support is evaluated utilizing Ni-Mo deposited TiO2, Al2O3, ZrO2, Fe2O3, and SiO2 supports, presented in Fig. 5.6b. Irrespective of the type of the support, conversion found to be similar and ranged between 43 and 47 % except for Ni-Mo/Fe2O3. The activity data disclosed that Ni-Mo and Ni-V bimetalic clusters and TiO2 support combination presented high conversion with prominent low oxygenated products, whilst rest of the catalysts showed major amount of undesirable products. 112 Guaiacol conversion Product selectivity with one oxygen Product selectivity with two oxygens Conversion / Selectivity (%) 80 60 40 20 0 80 Conversion / Selectivity (%) (a) Ni-V/ZrO 2 (b) Ni-Mo/ZrO 2 Ni-Cu/ZrO 2 Ni-Fe/ZrO 2 Guaiacol conversion Product selectivity with one oxygen Product selectivity with two oxygens 60 40 20 0 Ni-Mo/TiO 2 Ni-Mo/ZrO 2 Ni-Mo/Al2O 3 Ni-Mo/Fe2O 3 Ni-Mo/SiO 2 Fig. 5.6 Screening of catalysts. (a) Promoter effect: V, Mo, Cu, Fe. (b) Support effect: TiO 2, Al2O3, ZrO2, Fe2O3, SiO2 for HDO of guaiacol. Reaction conditions: 300 °C, 1 mL of guaiacol, H2 pressure of 100 psi, 5 h 113 Reaction data prompted us to consider Ni-Mo/TiO 2 and Ni-V/TiO2 catalysts for further investigation in order to study the guaiacol reactivity under variable conditions to optimize the catalytic performance. Products distribution is greatly affected by the choice of catalyst: phenol and creosol are the main products while cyclohexanone, anisole, catechol and CDME are the minor products over Ni-Mo/TiO2. Cyclohexanol is additionally noticed on Ni-V/TiO2 catalyst. Formation of cyclohexanol, cyclohexanone, phenol and anisole indicates a competition between hydrogenation of C=C aromatic system and hydrogenolysis of C-O, which is controlled by the mode of adsorption of guaiacol on the catalyst. Phenol is viewed as major product under our reaction conditions on all the catalysts [47]. Removal of oxygen from phenol is thermodynamically more difficult than the other oxygenates [48]. Similar observation was noted by Bui et al. [13] while studying catalytic conversion of guaiacol using Al 2O3, TiO2 and ZrO2 oxides, wherein TiO2 support formed phenol as major product. The influence of catalyst and metal on guaiacol conversion is presented for Ni-Mo/TiO 2 in Fig. 5.7. The guaiacol conversion in the absence of catalyst and on the support is fairly low (10 and 15 %, respectively). Since the support is barely active, the conversion is ascribable to the metal phase. In order to understand metallic effect we have compared guaiacol conversion of NiMo/TiO2 with Ni/TiO2 and Mo/TiO2. As seen from the figure, bimetallic Ni-Mo catalyst gave noticeably better results than mono metallic Ni and Mo catalysts and the result was not just an average of mono metallic catalysts. Ni/TiO2, Mo/TiO2 and Ni-Mo/TiO2 systems yield phenol of 5.2, 5.8 and 13.6 %, respectively. Additionally, Ni-Mo/TiO 2 showed meager portions of catechol and CDME compared to the respective mono metallic catalysts. Enhanced performance of Ni114 Mo/TiO2 with reference to Ni/TiO2 and Mo/TiO2 signify the feasible bimetallic interaction. Effect of Ni metal loading (0, 5, 10 and 20 %) was provided in Fig. 5.8. Catalytic activity with absence or low Ni metal loadings is sluggish due to inadequate active sites, the performance raised with Ni loading suggesting the homogeneous distribution of metals with corresponding increase in the active sites. Subsequently, activity dropped down at higher Ni loadings (20 wt%). Similar observation was noted by Sepúlveda et al. [47] wherein, guaiacol activity decreased with increased Re loading (>0.6 atoms per nm2). The poor performance at high metal loadings could be accounted for loss of active sites due to Ni metal aggregates. 50 Conversion (%) 40 30 20 10 0 Blank TiO 2 Ni/TiO 2 Catalyst(s) Mo/TiO 2 Ni-Mo/TiO 2 Fig. 5.7 Effect of catalyst on guaiacol conversion: Reaction conditions: 300 °C, 1 mL of guaiacol, H2 pressure of 100 psi, 5 h. 115 Conversion (%) 50 40 15 Conversion Cycloehexanone Anisole Phenol Cresol CDME Creosol Catechol 12 9 Yeild (%) 60 30 6 20 3 10 0 10Mo/TiO 2 5Ni-10Mo/TiO 2 10Ni-10Mo/TiO 2 Ni metal loading (wt%) 20Ni-10Mo/TiO 2 0 Fig. 5.8 Effect of Ni loading (0, 5, 10 and 20 wt%) on HDO of guaiacol. Reaction conditions: 300 °C, 1 mL of guaiacol, H2 pressure of 100 psi, 5 h The dependence of the guaiacol conversion on Ni-Mo/TiO 2 and Ni-V/TiO2 at different temperatures was displayed in Fig. 5.9. From the figure it is very clear that the deoxygenation strongly depends on the reaction temperature. As expected the conversion is very low at 200 °C and increased in slant with temperature up to 300 °C for both the catalysts. A dramatic increase is noticed for Ni-Mo/TiO2 from 300 °C and reached about 96.5 % at 350 °C. Ni-V/TiO 2 showed negligible conversion until 300 °C and then a sudden rise is noticed at 350 °C with about 85.9 % conversion. 116 (a) Conversion (%) 80 60 Conversion Cyclohexanone Anisole Phenol Cresol CDME Creosol Catechol Xylenol 40 Ni-Mo/TiO 2 30 Selectivity (%) 100 20 40 10 20 100 200 (b) Conversion (%) 80 60 250 300 350 Tem perature ( C) 0 40 Ni-V/TiO2 Conversion Cycloehexanol Cycloehexanone Anisole Phenol Cresol CDME Creosol Catechol Xylenol 30 20 40 10 20 0 0 Selectivity (%) 0 200 250 300 Temperature ( C) 0 350 0 Fig. 5.9 Effect of temperature on HDO of guaiacol. Reaction conditions: 1 mL of guaiacol , H2 pressure of 100 psi, 5 h 117 The product distribution also exhibited noticeable temperature dependence, the selectivity of deoxygenated products at 200 °C quite low and then increased accordingly with temperature. The selectivity of catechol and CDME decreased with temperature for Ni-Mo/TiO 2 while those are slightly elevated for Ni-V/TiO2. At 350 °C dimethyl phenol appeared for both the catalysts. It is clear from the figure that between the two catalysts Ni-Mo/TiO 2 showed promising selectivity towards phenol and cresol. A crucial role is demonstrated by H2 as a co-reactant for the removal of oxygen from the guaiacol. To illustrate the importance of H2, guaiacol conversion and product selectivity at different mole ratios of H2/guaiacol on Ni-Mo/TiO2 and Ni-V/TiO2 catalysts are presented in Fig. 5.10. Initially, the guaiacol conversion was 19.1 and 34.6 % for Ni-Mo/TiO 2 and Ni-V/TiO2, respectively. Under low hydrogen pressures, the high oxygenated transalkylated products catechol, CDME and creosol yields predominated specifically on Ni-V/TiO 2, and phenol and other low oxygenated products are low. This could be due to in the absence of hydrogen, methyl transfer and/demethylation prevails. With an increase in H2 pressure deoxygenated products intensified at the expense of catechol and CDME due to high H2 coverage. The products data underline the importance of H2 and points out the advantage of high pressure operation to obtain complete deoxygenated products. 118 (a) 60 Conversion (%) 50 40 60 Ni-Mo/TiO 2 Conversion Cyclohexanone Anisole Phenol Cresol CDME Creosol Catechol 50 40 Selectivity (%) 70 30 30 20 20 10 10 70 Conversion (%) 60 0 (b) 50 40 1.0 0.5 Hydrogen/Guaiacol mole ratio Conversion Cyclohexanol Cyclohexanone Anisole Phenol Cresol CDME Creosol Catechol 2.0 Ni-V/TiO 2 60 50 40 30 30 20 20 10 10 0 0 Selectivity (%) 0 0 1.0 0.5 Hydrogen/Guaiacol mole ratio 2.0 0 Fig. 5.10 Pressure dependence on HDO of guaiacol (a) Ni-Mo/TiO 2 and (b) Ni-V/TiO2. Reaction conditions: 300 °C, 1 mL of guaiacol, 5 h 119 Consistent with previous reports, with increase in the H 2 pressure the selectivity for the formation of deoxygenated products raised conversely, the major oxygenated products are prone to decrease [46, 49]. With raise in H2 pressure the guaiacol conversion (34.0–40.9 for Ni-V/TiO2 and 22 to 67.2 % for Ni-Mo/TiO2) and phenol selectivity (10–25 % for Ni-V/TiO2 and 9.3–38 % for Ni-Mo/TiO2) slowly raised. Catechol (21.4–1.9 % for Ni-V/TiO2; 32.1–0 % for Ni-Mo/TiO2) and CDME (39.3–6.8 for Ni-V/TiO2; 32–13.9 % for Ni-Mo/TiO2) selectivities abruptly decreased. At 2.0 molar ratio of H2/guaiacol, cyclohexanol and cyclohexanone increased to 16 and 10 %, respectively, with a meager decrease in phenol for Ni-V/TiO 2. This means cyclohexanol and cyclohexanone formation initiated by both phenol and ring saturation of guaiacol (Reaction 5, 6, 7 in Scheme 5.1). A stable propensity is observed for creosol selectivity for both the catalysts. Anisole and cresol are quite stable for Ni-V/TiO 2 and a meager increase is noted for the Ni-Mo/TiO2. The result of desired products with a raise in hydrogen pressure could be due to the involvement of H2 in the equilibrium between the phenol, catechol and CDME. High hydrogen pressure leads to a shift of catechol and CDME to phenol (and to subsequent cyclohexanol and cyclohexanone) and low H2 pressure results the opposite effect. Figure 5.11 shows guaiacol conversion and evolution of products with time over Ni-Mo/TiO 2 and Ni-V/TiO2. Conversion increased for both the catalysts with increase in the reaction time without much change in the products selectivity. As increase in the reaction time the conversion continued to be increased and reached a stable point within 8 h. About 42 % guaiacol conversion was noticed on Ni-V/TiO2 and a remarkable increase in conversion of ~66 % was obtained on Ni-Mo/TiO2. All the products have shown a stable behavior on both the catalysts. 120 100 Ni-Mo/TiO 2 Conversion Cycloehexanone Anisole Phenol Cresol CDME Catechol Creosol Conversion/ Selectivity (%) 80 60 40 20 0 100 2 4 6 8 Reaction time (h) 10 Ni-V/TiO 2 60 40 12 (b) Conversion Cycloehexanone Anisole Phenol Cresol CDME Creosol Catechol 80 Conversion/ Selectivity (%) (a) 20 0 2 4 6 8 Reaction time (h) 10 12 Fig. 5.11 Effect of reaction time on HDO of guaiacol (a) Ni-Mo/TiO 2 and (b) Ni-V/TiO2. Reaction conditions: 300 °C, 1 mL of guaiacol, H 2 pressure of 100 psi 121 In order to determine the reaction order and rate constant of the reaction, the activity data (concentration and time) fitted to power-law equation. As can be seen in Fig. 5.12, the slope of a plot of 1/CA as a function of time is linear with slope of k, thus HDO of guaiacol global kinetics followed the second order mechanism. The rate constant of Ni-Mo/TiO 2 is three times higher than the Ni-V/TiO2 and the high rate constant for Ni-Mo/TiO2 explains its high reactivity over Ni-V/TiO2. Ni-Mo/TiO2 has shown better performance over Ni-V/TiO2 at variable conditions, although the latter has favored cyclohexanol formation. The support phase is same for both the catalysts, therefore the activity difference could be originated from the metal phase and its prominent interaction with support. The facile guaiacol conversion on Ni-Mo/TiO 2 compared to Ni-V/TiO2 appears to be due to the synergetic effect between acid sites and structural properties. As indicated by XRD, BET and TPD the high activity of Ni-Mo/TiO 2 catalyst could be correlated with the strength of acidic sites, greater mesopore volume and presence of highly dispersed small sized metal active sites. 122 400 Ni-Mo/TiO2 (a) 1/Cguaiacol 300 200 y = 18.883x + 112.86 R² = 0.9897 100 0 0 250 4 6 8 Reaction Time (h) 10 12 (b) Ni-V/TiO2 200 1/Cguaiacol 2 150 100 y = 5.9974x + 121.66 R² = 0.9308 50 0 0 2 4 6 8 Reaction Time (h) 10 12 Fig. 5.12 Determination of order and rate constant by integral method: (a) Ni-Mo/TiO 2 and (b) Ni-V/TiO2. Reaction conditions: 300 °C, 1 mL of guaiacol, H 2 pressure of 100 psi. 123 Effect of water on guaiacol conversion is used to simulate the water content effect on the catalytic performance in bio-oil. These experiments were carried out under the same conditions as those performed without the water. Effect of water on guaiacol conversion and phenol/catechol ratio at 0, 0.5, 1.0, and 1.5 mL for Ni-Mo/TiO 2 and Ni-V/TiO2 is presented in Fig. 5.13. Remarkably, water has great impact on the overall conversion and product distribution on Ni-Mo/TiO2. Guaiacol conversion and products such as, phenol, cresol, anisole decreased and catechol increased for Ni-Mo/TiO2 with the addition of water indicating minimization of deoxygenation, demethoxylation and transalkylation pathways. Conversion and phenol/catechol ratio for Ni-Mo/TiO2 decreased from 43 to 20 and 1.26 to 0.45 %, respectively. Olcese et al. and Leiva et al. also noticed similar observation [28, 50]. However, water presence has shown no significant influence on guaiacol conversion or phenol/catechol ratio for Ni-V/TiO 2. To understand the water intolerance of Ni-Mo/TiO2, we performed identical reactions using Mo/TiO2 catalyst and the results are presented in Fig. 5.13c. Apparently, as water addition increases, a decrease in the guaiacol conversion (20 to 14 %) and phenol selectivity, and an increase in the catechol selectivity (phenol/catechol ratio 1.11 to 0.12) are noticed similar to that of Ni-Mo/TiO2 catalyst. Halász et al. stated that at 300 °C water molecules dissociatively adsorbs onto MoO3 and thus diminishes the number of Lewis acidic sites [51, 52]. Based on this observation and our experimental evidence, it is confirmed that in presence of water, MoO 3 phase is the reason for the deactivation of Ni-Mo/TiO2 catalyst. 124 125 Fig. 5.13 Effect of water on HDO of guaiacol: (a) Ni-Mo/TiO 2 and (b) Ni-V/TiO2. Reaction conditions: 300 °C, 1 mL of guaiacol, H2 pressure of 100 psi, 5 h 5.7 Conclusions HDO of guaiacol was successfully studied using Ni-Mo/TiO 2 and Ni-V/TiO2 catalysts under relatively low pressure (100 psi) and moderate temperatures (200–350 °C). Ni (10 wt%) − Mo (10 wt%)/TiO2 showed prominent results with reference to un-promoted Ni/TiO2, Mo/TiO2 and variable Ni metal loadings. Incorporation of Mo and V to Ni/TiO 2 not only enhanced dispersion of metal species, but also improved the catalytic performance. Variations in the product distribution and conversions were noted for both Ni-Mo/TiO 2 and Ni-V/TiO2 catalysts under different reaction conditions, which could be attributed to differences in the dissemination of active sites: metallic and acidic sites with the interference of promoter Mo and V, and particle 126 size distribution. With increase in temperature and H2/guaiacol mole ratio, both catalysts showed improved conversion and selectivity. In particular, Ni-Mo/TiO 2 showed better selectivity for phenol and cresol, while Ni-V/TiO2 showed prominent cyclohexanol along with phenol. Well dispersed small sized metal particles, high surface area, mesopore structure and more acidic sites could be the reason for the high activity of Ni-Mo/TiO2. A steady behavior is observed for timeon-stream study during 12 h for both catalysts confirm their stability. In aqueous medium, NiV/TiO2 maintained the guaiacol conversion and phenol/catechol ratio than Ni-Mo/TiO 2 proved its tolerance to water presence. 5.8 Acknowledgement Authors are grateful to the financial support of the Natural Science and Engineering Research Council of Canada (NSERC). Lakshmi Katta thank university's Eyes High Program for the Postdoctoral Fellowship. The authors also thank Ludivine Gras for her contribution in the reaction part. 127 5.9 References [1] Serrano-Ruiz JC, Dumesic JA (2011) Energy Environ Sci 4:83 [2] Huber GW, Iborra S, Corma A (2006) Chem Rev 106:4044 [3] Elliott DC (2007) Energy Fuels 21:1792 [4] Choudhary TV, Phillips CB (2011) Appl Catal A 397:1 [5] Bridgwater AV, Peacocke GVC (2000) Renew Sustain Energy Rev 4:1 [6] Wang H, Male J, Wang Y (2013) ACS Catal 3:1047 [7] Anex RP, Aden A, Kazi FK, Fortman J, Swanson RM, Wright MM, Satrio JA, Brown RC, Daugaard DE, Platon A, Kothandaraman G, Hsu DD, Dutta A (2010) Fuel 89(Supplement 1):S29 [8] Ruddy DA, Schaidle JA, Ferrell Iii JR, Wang J, Moens L, Hensley JE (2014) Green Chem 16:454 [9] Zhao C, Kou Y, Lemonidou AA, Li X, Lercher JA (2009) Angew Chem 121:4047 [10] Zakzeski J, Bruijnincx PCA, Jongerius AL, Weckhuysen BM (2010) Chem Rev 110:3552 [11] Hong Y-K, Lee D-W, Eom H-J, Lee K-Y (2014) J Mol Catal A 392:241 [12] Zhao C, Lercher JA (2012) Angew Chem Int Ed 51:5935 [13] Bui VN, Laurenti D, Delichère P, Geantet C (2011) Appl Catal B: Environ 101:246 [14] Bui VN, Laurenti D, Afanasiev P, Geantet C (2011) Appl Catal B: Environ 101:239 [15] Lin Y-C, Li C-L, Wan H-P, Lee H-T, Liu C-F (2011) Energy Fuels 25:890 [16] Zhu X, Lobban LL, Mallinson RG, Resasco DE (2011) J Catal 281:21 [17] Gutierrez A, Kaila RK, Honkela ML, Slioor R, Krause AOI (2009) Catal Today 147:239 [18] Nimmanwudipong T, Runnebaum R, Block D, Gates B (2011) Catal Lett 141:779 128 [19] Elliott DC, Hart TR (2008) Energy Fuels 23:631 [20] Ohta H, Kobayashi H, Hara K, Fukuoka A (2011) Chem Commun 47:12209 [21] Zhao HY, Li D, Bui P, Oyama ST (2011) Appl Catal A 391:305 [22] Yan N, Yuan Y, Dykeman R, Kou Y, Dyson PJ (2010) Angew Chem Int Ed 49:5549 [23] Olcese RN, Bettahar M, Petitjean D, Malaman B, Giovanella F, Dufour A (2012) Appl Catal B 115–116:63 [24] Wang X, Rinaldi R (2012) ChemSusChem 5:1335 [25] He Z, Wang X (2012) Hydrodeoxygenation of model compounds and catalytic systems for pyrolysis bio-oils upgrading. Catal Sustain Energy 1:28–52 [26] Bridgwater AV (1996) Catal Today 29:285 [27] Escalona N, Aranzaez W, Leiva K, Martínez N, Pecchi G (2014) Appl Catal A 481:1 [28] Olcese R, Bettahar MM, Malaman B, Ghanbaja J, Tibavizco L, Petitjean D, Dufour A (2013) Appl Catal B 129:528 [29] Wu S-K, Lai P-C, Lin Y-C (2014) Catal Lett 144:878 [30] Liu C, Wang H, Karim AM, Sun J, Wang Y (2014) Chem Soc Rev 43:7594 [31] Saidi M, Samimi F, Karimipourfard D, Nimmanwudipong T, Gates BC, Rahimpour MR (2014) Energy Environ Sci 7:103 [32] Yue C-J, Zhang Q-Y, Gu L-P, Su Y, Zhu S-P (2014) Asia Pac J Chem Eng 9:581 [33] Bathe SR, Patil PS (2014) J Mater 2014:5 [34] Ricci PC, Carbonaro CM, Stagi L, Salis M, Casu A, Enzo S, Delogu F (2013) J Phys Chem C 117:7850 [35] Nogueira NAS, da Silva EB, Jardim PM, Sasaki JM (2007) Mater Lett 61:4743 129 [36] Selvaraj M, Shanthi K, Maheswari R, Ramanathan A (2014) Energy Fuels 28:2598 [37] Chandra Mouli K, Mohanty S, Hu Y, Dalai A, Adjaye J (2013) Catal Today 207:133 [38] Cheng FY, Chen J, Gou XL (2006) Adv Mater 18:2561 [39] Yang Y, Ochoa-Hernández C, de la Peña O’Shea VA, Pizarro P, Coronado JM, Serrano DP (2014) Appl Catal B 145:91 [40] Kim SD, Baek SC, Lee Y-J, Jun K-W, Kim MJ, Yoo IS (2006) Appl Catal A 309:139 [41] Fang K, Ren J, Sun Y (2005) J Mol Catal A 229:51 [42] Zhang X, Zhang Q, Chen L, Xu Y, Wang T, Ma L (2014) Chin J Catal 35:302 [43] González-Borja MÁ, Resasco DE (2011) Energy Fuels 25:4155 [44] Sun J, Karim AM, Zhang H, Kovarik L, Li XS, Hensley AJ, McEwen J-S, Wang Y (2013) J Catal 306:47 [45] Nimmanwudipong T, Aydin C, Lu J, Runnebaum R, Brodwater K, Browning N, Block D, Gates B (2012) Catal Lett 142:1190 [46] Olcese RN, Bettahar M, Petitjean D, Malaman B, Giovanella F, Dufour A (2012) Appl Catal B 115–116:63 [47] Sepúlveda C, García R, Reyes P, Ghampson IT, Fierro JLG, Laurenti D, Vrinat M, Escalona N (2014) Appl Catal A 475:427 [48] Furimsky E (2000) Appl Catal A 199:147 [49] Jin S, Xiao Z, Li C, Chen X, Wang L, Xing J, Li W, Liang C (2014) Catal Today 234:125 [50] Leiva K, Sepúlveda C, García R, Fierro JLG, Escalona N (2014) Catal Commun 53:33 [51] Halász I, Gáti G (1979) React Kinet Catal Lett 12:411 [52] Prasomsri T, Nimmanwudipong T, Roman-Leshkov Y (2013) Energy Environ Sci 6:1732 130 CHAPTER 6: UPGRADING OF ANISOLE COMPONENT BY NI-MO/TIO 2 AND NIV/TIO2 CATALYSTS: SYNTHESIS, CHARACTERIZATION AND KINETIC MEASUREMENTS 6.1. Presentation of the Article In this chapter, the results from the reaction and kinetic studies of the hydrodeoxygenation of guaiacol and anisole are presented. The hydrodeoxygenation of anisole was performed using two different catalysts previously tested in the hydrodeoxygenation of guaiacol. Various reaction temperatures (250, 275, 300 and 325 °C) and reaction times (2, 4, 6, 8 hours) were used in order to determine the selectivity of the reaction. All of the samples were pre-activated at 350 °C for two hours at 5 °C min–1 under atmospheric H2 pressure. The experiments were performed in a 30 mL batch reactor made of stainless steel (Swagelok). Additional analysis of the catalyst using an X-ray photoelectron spectroscopy instrument was performed to provide a better understanding of the catalyst surface. Finally, a kinetic study of both the hydrodeoxygenation of guaiacol and anisole were performed using the integral method. As a result, the reaction order, rate constant and energy activation of the reactions were obtained. 131 Upgrading of Anisole Components by Ni-Mo/TiO2 and Ni-V/TiO2 Catalysts: Synthesis, Characterization and Kinetic Measurements Aqsha, A., Katta, L., Tijani, M.M. & Mahinpey, N. Department of Chemical and Petroleum Engineering, Schulich School of Engineering, The University of Calgary, Calgary, Canada, AB T2N 1N4 This article has been submitted to Energy & Fuels in January 2016. 6.2. Abstract This work reports the catalytic hydrodeoxygenation (HDO) of model components (anisole) of lignocellulose biomass over Mo and V promoted Ni deposited titania support. The physicochemical properties of each material were elucidated, and reactions were studied over a wide range of temperatures to enable examination of kinetic parameters. Crystallite sizes measured using X-ray powder diffraction were corroborated well with the transmission electron microscopy results. The presence of Ni, Ti, Mo and V species was confirmed through X-ray photoelectron spectroscopy (XPS). Significant HDO activity for both the catalysts could be attributed to high dispersions of metals and acidic sites, which were affected by interaction between the Ni metal and the titania support. The higher activity of Ni-Mo/TiO 2 may have resulted from the high Ni/Ti surface atomic ratio, which was confirmed by XPS. The major product of the anisole HDO reactions was phenol. Interestingly, a considerable amount of benzene and cyclohexane were also noticed in the anisole HDO reaction. The activation energy 132 values for anisole reactions over Ni-Mo/TiO2 and Ni-V/TiO2 were 80.9 and 53.9 kj/mol, respectively. Keywords: Anisole, bio-oil, catalyst, hydrodeoxygenation, model components. 6.3 Introduction Sustainable alternative energies are viable solutions for the replacement of fossil fuels and can greatly reduce the CO2 emissions associated with the transportation sector. Biomass, as an energy alternative, can be effectively converted into bio-oil via several physical, thermal and biological processes. Bio-oil has high potential as liquid fuel, as it contains less nitrogen and sulphur than fossil fuel. However, the produced bio-oil possesses chemical and thermal instability, due to the presence of a significant amount of oxygen [1-3]. Catalytic hydrodeoxygenation (HDO), a chemical method, is a feasible approach for the upgrading of lignin derived bio-oils [4-7]. Lignin, consisting largely of aromatic linkages in polymeric units, occupies a major part of the lignocellulose biomass; however, lignin-derived bio-oils have received less research attention than cellulosic feedstock [8,9]. In the current investigation, anisole have been chosen as representatives of compounds containing hydroxyl and methoxy groups, both of which are abundant in lignocellulose. Although there have been many reports presented on the HDO of lignin model components, studies on the use of anisole as probes under a single roof have been limited. 133 Precious metal based catalysts and zeolites have been used for a wide variety of applications, including the hydrogenolysis of oxygenated compounds [10-18]. However, a number of studies have disclosed that cheaper ions, such as transition metal combinations, can be effective as precious metal catalysts [19-21. Details about the reactivity over different bimetal catalysts and operating conditions using guaiacol have been presented in our previous work [22]. It is also vital to determine kinetic parameters, such as the rate constant and activation energy, in order to understand the complete reaction network. A pseudo first-order kinetic model has been used by several researchers to devise the kinetic parameters for bio-oil. The objectives of this study were the upgrading of anisole using nickel and molybdenum (Ni-Mo) and nickel and vanadium (Ni-V) supported on titania as catalysts, which are quite active in the removal of oxygen based on previous observations, and the provision of a quantitative foundation by determining the kinetics of those reactions. 6.3. Experimental Section 6.4.1 Catalysts Preparation The catalysts used in this work were prepared by incipient wetness impregnation. Requisite amounts of nickel nitrate hexahydrate (Ni(NO3)2·6H2O), ammonium heptamolybdate ((NH4)6Mo7O24), and ammonium metavanadate (NH4VO3) were used to obtain nickel (Ni, 10 wt%), molybdenum (Mo, 10 wt%), and vanadium (V, 10 wt%) oxides, respectively. All the chemicals were purchased from Sigma-Aldrich or Alfa Aesar. 134 The catalysts were prepared by first stirring a suspension of the commercial support into the metal precursor solution on a hot plate at 70°C until complete evaporation was attained. The impregnated sample was then oven dried at 110°C for 12 h. The dried samples were subsequently calcined in air at 500°C for 3 h to obtain the final catalysts – NiO-MoO 3 (Ni-Mo) and NiO-V2O5 (Ni-V) – deposited onto titanium dioxide (TiO 2) support 6.4.2 Characterization Studies Powder X-ray diffraction (XRD) data were acquired using a Rigaku Multiflex X-ray Diffractometer equipped with Cu−Kα radiation (λ = 0.1548 nm) to identify the phase constitutions in the samples. The step size and time per step were fixed at 0.02° and 1 s, respectively, in the range of 2θ from 20° to 80°. The crystallite size was measured using the Scherrer equation, with aid of the highest intense peaks. Crystalline phases were identified using reference data files (PDF) from the International Centre for Diffraction Data (ICDD). The transmission electron microscopy (TEM) studies were made on a Tecnai F20 instrument equipped with a slow-scan CCD (charge-coupled device) camera and at an accelerating voltage of 200 kV. Samples for TEM were prepared by dispersing them ultrasonically in ethyl alcohol. After dispersion, a droplet was deposited on a copper grid supporting a perforated carbon film and allowed to dry. The dried specimen was used for examination under the microscope. For the X-ray photoelectron spectroscopy (XPS) measurements, each sample was mounted and pumped down from atmospheric pressure to ultra-high vacuum (UHV) pressure of 10 -7-10-8 Pa. 135 A monochromatic aluminium X-ray source was used. Wide energy surveys were carried out (0 – 1,200 eV) with pass energy of 187.85 eV to identify the elements in the sample using an X-ray setting of 200 µm 50W/15KV. After identifying the elements present in the sample, an acquisition with a smaller energy window (10 – 30 eV with pass energy of 23.5 eV) was performed to obtain high-energy resolution spectra of each element requested, in order to determine the chemical environment and oxidation states. The reference binding energy was chosen as the C1s level of the adventitious carbon present at 284.6 eV. 6.4.3 Reaction Study: HDO of Anisole The HDO experiments of anisole were carried out in a 30 mL batch reactor from stainless steel (from Swagelok). All the samples were pre-activated at 350°C for 2 h at 5°C min–1 under one atmosphere of hydrogen (H2) pressure. Ten wt% of the catalyst to the reactant (anisole) was loaded in the reactor. The mixture was then heated to the desired temperature using a subcritical bath in the requisite H2 atmosphere. After completing the reaction, the reactor was placed in a water bath to bring it to room temperature (~25 °C). The reaction mixture was subsequently analyzed with a gas chromatography / mass spectrometry system (GC-MS, Agilent 7890) equipped with a flame ionization detector (FID). An HP-5MS capillary column (diameter of 0.25 mm, thickness of 25 µm, and length of 30 m) was used for separation at a column temperature of 100°C. The major products detected with GC-MS for the HDO of anisole were phenol, benzene, cyclohexane, 136 methoxycyclohexane, cyclohexanol, cyclohexanone and methylanisole. All the experiments were performed at least twice, and the results were reproducible within a range of conversion of ± 3%. 6.4. Results and Discussions 6.5.1 Structural Properties To determine the crystal phase and size, XRD measurements were carried out. The XRD results of Ni-Mo/TiO2 and Ni-V/TiO2 under fresh and reduced conditions are presented in Figure 6.1. For comparison, TiO2 anatase as purchased is also presented. The sharp well-defined peaks in the diffraction patterns correspond to crystalline anatase and rutile phases. The diameters of the Ni particles estimated from the full width at half maximum (FWHM) of the most intense reflections using the Scherrer equation are displayed in Table 6.1 for all the catalysts. These data reveal that the average sizes of the Ni metal active phase for the Ni-Mo/TiO 2 and Ni-V/TiO2 catalysts were 8.3 and 10.2 nm, respectively. No indications of Ni2O2 and Ni2O3 were found within the investigated range. Table 6.1 Physico-chemical properties of Ni-Mo/TiO 2 and Ni-Mo/TiO2 catalysts Catalyst Crystallite sizea (nm) S.A.b (m2/g) - 42 Ni-V/TiO2-F Ni-V/TiO2-R 10.9 Ni-Mo/TiO2-F Ni-Mo/TiO2-R 51 - a 43 8.2 58 b c XRD: BET SA: XPS analysis 137 Ni/Tic atomic ratio 0.107 0.093 0.123 0.131 (a) Intensity (a.u.) NiMo/TiO2-R NiMo/TiO2-F TiO2 20 30 40 50 2Theta ( ) 0 60 70 80 Intensity (a.u.) (b) NiV/TiO2-R NiV/TiO2-F TiO2 20 30 40 50 2Theta ( ) 0 60 70 80 90 Fig. 6.1 X-ray diffraction pattern of (a) TiO 2, fresh Ni-Mo/TiO2 and Ni-Mo /TiO2 after the reduction and (b) TiO2, fresh Ni-V/TiO2 and Ni-V/TiO2 after the reduction. 138 Representative TEM images for both Ni-Mo/TiO2 and Ni-V/TiO2 under both fresh and reduced conditions are presented in Figure 6.2. The size distributions were obtained by counting particles. In all cases, the size of the metal was similar to that of the anatase phase, as confirmed by TEM. Both metallic and oxidic phases were in close proximity, facilitating the metal-support interactions and thus HDO reactivity. Ni-Mo/TiO2-R Ni-Mo/TiO2 -F Ni-V/TiO2-F Ni-V/TiO2-R Fig. 6.2 TEM images of Ni-V/TiO2 and Ni-Mo/TiO2 catalysts before (fresh, F) and after reduction (R) 139 6.5.2 Surface Analysis XPS was used to elucidate the chemical composition and valence state of the metal components at the surface. Elements identified in the given spectra were Ni, Ti, oxygen (O), V, Mo, and carbon (C), which was due to adventitious carbons. Peaks in all the spectra were attenuated for the reduced samples. Interestingly, in all the spectra (Ni 2p, Ti 2p, Mo 3d and V 2p), the high binding energy peaks FWHM were much broader and thus shorter than the low binding energy peaks. Ni 2p is a complex spectrum with a mixture of core level and satellite peaks. Ni 2p spectra for Ni-Mo/TiO2 (fresh and reduced) and Ni-V/TiO2 (fresh and reduced) were deconvoluted by Gaussian curves representing Ni 2p3/2 and Ni 2p1/2 spin orbit levels, as presented in Figure 6.3. All the samples showed major component peaks at 873.2 and 855.8 eV (BE = 17.4 eV) corresponding to Ni 2p1/2 and Ni 2p3/2, respectively, due to Ni2+. Moreover, related satellite peaks (shake-up features) were observed at ~6 eV (two broad peaks at 880.01 and 861.9 eV) higher in the binding energy region. Figure 6.4 shows core level spectra of Ti 2p of all the four samples. The two component peaks were observed at binding energy values approximately 5.7 eV apart at 458.79 and 464.48 eV. These two peaks were assigned to Ti 2p3/2 and Ti 2p1/2. The energy separation was due to the expected oxidation state of Ti4+, which is consistent with the literature reports. No indications of peaks due to reduce Ti (Ti3+) were noted for the reduced samples spectra [23]. 140 Ni/Ti surface atomic ratios are presented in Table 6.1. It is obvious that the Ni/Ti was greatly influenced by the promoter and treatment conditions. The Mo promoter enriched the Ni amount over the surface compared to the V promoter. The reduction treatment of Ni-Mo/TiO 2 further enhanced the surface Ni; however, the Ni amount decreased over Ni-V/TiO 2 with reduction treatment. Ni 2p 3/2 Ni-Mo/TiO 2-R Ni-Mo/TiO 2-F Ni-V/TiO 2-R Ni-V/TiO 2-F Ni 2p 1/2 Ni 2p 3/2 satellite C/S satellite Ni 2p 1/2 Ni 2p 900 890 880 870 860 Binding energy (eV) 850 840 Fig. 6.3 Ni 2p spectra for Ni-Mo/TiO 2 (fresh and reduced) and Ni-V/TiO 2 (fresh and reduced) 141 Ti 2p Ni-Mo/TiO2-R Ni-Mo/TiO2-F Ni-V/TiO2-R C/S Ni-V/TiO2-F Ti 2p1/2 470 465 Ti 2p3/2 460 Binding energy (eV) 455 450 Fig. 6.4 Ti 2p spectra for Ni-Mo/TiO2 (fresh and reduced) and Ni-V/TiO 2 (fresh and reduced) The core level spectra of O 1s are presented in Figure 6.5. The asymmetric O 1s spectra were combinations of two component peaks (530.04 and 531.2 eV). The low binding energy peak appeared due to Ti-O, and the high energy peak can be attributed to M-O. The Mo 3d core level spectra for Ni-Mo/TiO2 (fresh and reduced) are presented in Figure 6.6, which indicates that there were two different Mo species present on the surface. Both fresh and reduced samples showed major 3d orbital doublets at 232.5 and 235.7 eV (BE = 3.2 eV), due to the inert Mo6+ state in MoO3 or NiMoO4. 142 C/S Ni-Mo/TiO2-F 522 528 Ni-Mo/TiO2-R 534 540 522 Ni-V/TiO2-F 522 528 528 534 540 534 540 Ni-V/TiO2-R 534 540 522 Binding energy (eV) 528 Fig. 6.5 Core level spectra of O 1s for Ni-Mo/TiO 2 (fresh and reduced) and Ni-V/TiO2 (fresh and reduced) Reduced Ni-Mo/TiO2 showed two shoulder peaks at 231.2 and 233.8 eV, which can be attributed to Mo5+. V 2p core level spectra for Ni-V/TiO2 (fresh and reduced) sample are presented in Figure 6.7. V 2p exhibited two symmetric peaks with a splitting of ~7.5 eV. The binding energies of V 2p3/2 and V 2p1/2 peaks were observed at 516.9 and 524.4 eV, respectively, which are characteristic of stoichiometric V2O5 and agreed well with values in the literature [24]. 143 Mo Mo 6+ 3d 3/2 Mo N i-M o/T iO 2 -R 3d 5/2 Mo 5+ C/S Mo Mo 3d 3/2 24 5 6+ 240 6+ 2 35 5+ 6+ 3d 5/2 2 30 N i-M o /T iO 2 -F B inding energy (eV ) 22 5 22 0 Fig. 6.6 Mo 3d core level spectra for Ni-Mo/TiO2 (fresh and reduced) Ni-V/TiO 2-R V 2p C/S Ni-V/TiO 2-F V 2p 1/2 525 V 2p 3/2 520 515 Binding energy (eV) 510 Fig. 6.7 V 2p core level spectra for Ni-V/TiO 2 (fresh and reduced) 144 6.5.3 Reaction and Kinetic Study of Anisole HDO reactions of anisole were carried out in a batch reactor, and their global kinetic analysis using the integral method was used to calculate kinetic parameters, such as the reaction order, rate constant, and energy activation. Results from the HDO of anisole can be seen in Figures 6.8 to 6.10. Under the reaction conditions, it was found that the reaction produced more than 10 different products; of these products, we present the species that accounted for more than 1% of the total products. As shown in Figures 6.8 and 6.9, nine major products were detected by GC-MS, including benzene, cyclohexane, toluene, methoxycyclohexane, cyclohexanol, phenol, o,m,p-methylanisole, oisoprophylanisole, and p-isoprophylanisole. The anisole transformation can be achieved by various reaction pathways, such as dealkylation, demethylation, deoxygenation, hydrogenation, isomerization, and even ring opening reaction. The effects of reaction temperature under the controlled conditions using Ni-Mo/TiO 2 and NiV/TiO2 were clearly seen for benzene, cyclohexane, phenol, and isoprophylanisole. As the temperature increased, the production of benzene and phenol also increased; however, benzene started to decrease beyond 275°C, perhaps due to competition from the reaction to produce more phenol. 145 Fig. 6.8 Effect of temperature on HDO of anisole. Reaction conditions: 1 mL of anisole, H 2 pressure of 100 psi, 6 h and using Ni-Mo/TiO2. Fig. 6.9 Effect of temperature on HDO of anisole. Reaction conditions: 1 mL of anisole, H 2 pressure of 100 psi, 6 h and using Ni-V/TiO2. 146 Fig. 6.10 Effect of reaction time on HDO of anisole. Reaction conditions: 1 mL of anisole , H2 pressure of 100 psi, 300 °C and using Ni-Mo/TiO2. Fig. 6.11 Effect of reaction time on HDO of anisole. Reaction conditions: 1 mL of anisole, H2 pressure of 100 psi, 300 °C and using Ni-V/TiO 2. 147 Based on the possible reaction pathways, anisole could have been transformed into benzene via direct deoxygenation. Subsequently, benzene may have undergone hydrogenation and converted to cyclohexane. The presence of H2 gas is vital to provide sufficient hydrogen bonding needed for anisole transformation. The removal of methyl from anisole leads to phenol and further produces toluene by hydrogenation. Conversion pathways of anisole most likely began with hydrogenation of the benzene ring of anisole, forming methoxycyclohexane, followed by the removal of methyl group to form cyclohexanol. If the reaction continued to undergo the deoxygenation, cyclohexane would have been produced. Similar observations were also mentioned by Jin et al. when performing HDO of anisole using a Ni-based catalyst [25]. From the graph in Figure 6.9, we observed that the formation of benzene was higher at lower temperatures (250 and 275 C) and started to decrease at higher temperatures (300 and 325 C), which may have been caused by the hydrogenation and methylation step. Aside from varying the temperature of the reaction, effects of reaction time over HDO of anisole were also studied. With both Ni-Mo/TiO2 and Ni-V/TiO2, the conversion increased from 18.7% (2 hours) up to 59.9% (8 hours) and from 14.8% (2 hours) up to 46.1% (8 hours), respectively (Figures 6.10 and 6.11). It was noticed that the use of Ni-Mo/TiO 2 favored the formation of phenol. A similar observation was noted by Bui et al. when studying the catalytic conversion of guaiacol using aluminum oxide (Al2O3), TiO2 and zirconium dioxide (ZrO2), where the TiO2 support formed phenol as the major product. It was also found that using a longer reaction time 148 favored benzene and cyclohexane-type molecules without any alkyl or alcohol branches. It is clear that the product distributions were greatly affected by the use of catalyst. (a) (b) 149 (c) (d) Fig. 6.12 Plots of the of anisole concentration versus reaction time of HDO of anisole using Ni-Mo/TiO2 at (a) 250 °C, (b) 275 °C, (c) 300 °C and (d) 325 °C. 150 In order to obtain kinetic parameters, HDO of anisole was performed under four different temperature conditions and reaction times. For every reaction temperature, three different graphs were plotted, in order to determine the reaction order of the reaction using the integral method. Second-order reaction trends were found to match the data obtained, as shown in Figures 6.12 and 6.14 for Ni-Mo/TiO2 and Ni-V/TiO2, respectively. Since the reaction order had been determined, the rate laws of the HDO of anisole could be written and rearranged as equations of line. The plot of the logarithmic natural of the rate constant (ln k) and 1 over temperature (1/T) is shown in Figures 6.13 and 6.15 for Ni-Mo/TiO 2 and Ni-V/TiO2, respectively. Using a linear regression method, k and the energy activation (EA) can be determined. The values of the kinetic parameters for HDO reactions using both catalysts is presented in Table 6.2. Fig. 6.13 Plot of the of logarithmic natural of rate constant versus 1/T of HDO of anisole using Ni-Mo/TiO2 to determine the kinetic parameter 151 (a) (b) 152 (c) (d) Fig. 6.14 Fig. 6.12 Plot of the of anisole concentration versus reaction time of HDO of anisole using Ni-V/TiO2 (a) 250, (b) 275, (c) 300, (d) 325 °C temperatures 153 Fig. 6.15 Plot of the of logarithmic natural of rate constant versus 1/T of HDO of anisole using Ni-V/TiO2 to determine the kinetic parameter Table 6.2 Kinetic parameter of HDO reaction using Ni-Mo/TiO 2 and Ni-V/TiO2 Catalyst k (L/mol hr) EA (kJ/mol) Ni-V/TiO2 6481.6 53.89 Ni-Mo/TiO2 9735.2 80.93 6.5.Conclusions The HDO of anisole was carried out in a batch reactor using Ni-Mo/TiO2 and Ni-V/TiO2 catalysts under relatively low pressure (100 psi) and moderate temperatures (250 to 325 °C). Using XPS analysis, we found that the peaks in all the spectra were attenuated for the reduced 154 samples. Interestingly, the high binding energy peaks FWHM in all spectra (Ni 2p, Ti 2p, Mo 3d and V 2p) were much broader and thus shorter than the low binding energy peaks. We suggest that differences in the dissemination of active sites (metallic and acidic sites) with the interference of promoters Mo and V and that particle size distribution in both Ni-Mo/TiO 2 and Ni-V/TiO2 catalysts contributed to product distribution and conversions during the HDO reaction. Nine major products were detected by GC-MS including benzene, cyclohexane, toluene, methoxycyclohexane, cyclohexanol, phenol, o,m,p-methylanisole, o-isoprophylanisole and pisoprophylanisole. With longer reaction times for Ni-Mo/TiO 2 and Ni-V/TiO2, the conversion increased from 18.7% (2 h) up to 59.9% (8 h) and from 14.8% (2 h) up to 46.1% (8 h), respectively. The high activity of Ni- Mo/TiO2 can be explained by the well-dispersed small sized metal particles, high surface area, mesopore structure and more acidic sites of the catalyst. A kinetic analysis revealed that the reaction order was 2nd order in both reactions (using Ni-Mo/TiO2 and Ni-V/TiO2); and, the activation energy values for anisole reactions over Ni-Mo/TiO 2 and NiV/TiO2 were 80.9 and 53.9 kj/mol, respectively. 6.6. Acknowledgement The authors wish to acknowledge the financial support of the Natural Sciences and Engineering Research Council of Canada (NSERC) and Ministry of Higher Education in Kingdom of Saudi 155 Arabia, and would like to thank Diego Costa and Camilla Fernandes de Oleivera for their contribution in the experimental part of this study. 156 6.7. References [1] Yaman, S. Pyrolysis of biomass to produce fuels and chemical feedstocks, Energy Convers. Manag., 45, pp. 651-671, 2004. [2] Fisher, T., Hajaligol, M., Waymack, B. & Kellogg, D. Pyrolysis behaviour and kinetics of biomass derived materials, J. Anal. Appl. Pyrolysis, 62, pp. 331-349, 2002. [3] Ruddy, D. A., Schaidle, J. A., Ferrell III, J.R., Wang, J., Moens, L. & Hensley, J. E., Recent advances in heterogeneous catalysts for bio-oil upgrading via ex situ catalytic fast pyrolysis: catalyst development through the study of model components, Green Chem, 16, pp. 454-490, 2014. [4] Bridgwater, A. V. & Peacocke, G. V. C., Fast pyrolysis processes for biomass, Renew. Sustain. Energy, 4(1), pp. 1-73, 2000. [5] Huber, G. W., Iborra, S. & Corma, A., Synthesis of transportation fuels from biomass: chemistry, catalysts, and engineering, Chem. Rev., 106, pp. 4044-4098, 2006. [6] Serrano-Ruiz, J. C. & Dumesic, J. A., Catalytic routes for the conversion of biomass into liquid hydrocarbon transportation fuels, Energy Environ. Sci., 4(83), pp. 83-99, 2011. [7] Elliott, D. C., Historical Developments in Hydroprocessing Bio-oils, Energy & Fuels, 21, pp. 1792 – 1815, 2007. [8] Bridgewater, A. V., Review of fast pyrolysis of biomass and product upgrading. Biomass and Bioenergy, 38, pp. 68-94, 2012. [9] Mohan, D., Pittman, C. U. & Steele, P. H. Pyrolysis of wood/biomass for bio-oil: a critical review. Energy & Fuels, 20, pp. 848-889, 2006. 157 [10] Zhao, C. & Lercher, J. A., Upgrading Pyrolysis Oil over Ni/HZSM-5 by Cascade Reactions, Angewante Chemie, 51(24), pp. 5935-5940, 2012. [11] Bui, V.N., Laurenti, D., Afanasiev, P. & Geantet, C., Hydrodeoxygenation of guaiacol with CoMo catalysts. Part I: Promoting effectof cobalt on HDO selectivity and activity, Appl. Catal. B: Environ., 101, pp 239-245, 2011. [12] Bui, V.N., Laurenti, D., Afanasiev, P. & Geantet, C., Hydrodeoxygenation of guaiacol Part II: Support effect for CoMoS catalysts on HDO activity and selectivity, Appl. Catal. B: Environ., 101, pp 246-255, 2011. [13] Lin, Y. C., Li, C. L., Wan, H. P., Lee, H. T. & Liu, C. F., Catalytic Hydrodeoxygenation of Guaiacol on Rh-Based and Sulfided CoMo and NiMo Catalysts, Energy & Fuels, 25, pp. 890-896, 2011. [14] Zhu, X., Lobban, L. L., Mallinson, R. G. & Resasco, D. E., Bifunctional transalkylation and hydrodeoxygenation of anisole over a Pt/HBeta catalyst, J Catal., 281, pp. 21-29, 2011. [15] Gutierrez, A., Kaila, R. K., Honkela, M. L., Slioor, R. & Krause, A. O. I., Hydrodeoxygenation of guaiacol on noble metal catalysts, Catal. Today, 147, pp. 239-246, 2009. [16] Nimmanwudipong, T., Runnebaum, R., Block, D. & Gates, B., Catalytic Reactions of Guaiacol: Reaction Network and Evidence of Oxygen Removal in Reactions with Hydrogen, Catal. Lett., 141, pp. 779-783, 2011. [17] Elliott, D. C. & Hart, T. R., Catalytic Hydroprocessing of Chemical Models for Bio-oil, Energy & Fuels, 23, pp. 631-637, 2009. 158 [18] Ohta, H., Kobayashi, H., Hara, K. & Fukuoka, A., Hydrodeoxygenation of phenols as lignin models under acid-free conditions with carbon-supported platinum catalysts, Chem. Comm., 47, pp. 12209-12211, 2011. [19] Olcese, R. N., Bettahar, M., Petitjean, D., Malaman, B., Giovanella, F. & Dufour, A., Gasphase hydrodeoxygenation of guaiacol over Fe/SiO2 catalyst, Appl. Catal. B: Env.,63, pp. 63-73, 2012. [20] Wang, X., Rinaldi, R., Solvent effects on the hydrogenolysis of diphenyl ether with Raney nickel and their implications for the conversion of lignin, ChemSusChem, 5, pp. 1455-1456, 2012. [21] He, Z., Wang, X., Hydrodeoxygenation of model compounds and catalytic systems for pyrolysis bio-oils upgrading, Catal. Sustain. Energy, 1, pp. 28–52, 2012. [22] Aqsha, A., Katta, L., & Mahinpey, N., Catalytic Hydrodeoxygenation of Guaiacol as Lignin Model Component Using Ni-Mo/TiO2 and Ni-V/TiO2 Catalysts, Catal. Lett., 145(6), pp. 1351-1363, 2015. [23] Kruse, N. & Chenakin, S., XPS characterization of Au/TiO 2 catalysts: Binding energy assessment and irradiation effects, App. Catal. A: General, 391, pp. 367-376, 2011. [24] Valenzuela, R. X., Fierro, J. L. G., Corberan, V. C. & Mamedov, E. A., Ethane oxidehydrogenation selectivity and reducibility of mixed NiVSb oxides, Catal. Lett., 40, pp. 223-228 – 1996. [25] Jin, S., Xiao, Z., Li, C., Chen, X., Wang, L., Xing, J., Li, W. & Liang, C., Catalytic hydrodeoxygenation of anisole as lignin model compound over supported nickel catalyst, Catal. Today, 234, pp. 125-132, 2014. 159 CHAPTER 7: CONCLUSIONS AND SUGGESTIONS FOR FUTURE WORKS 7.1. Relation between Chapters and Overall Achievement In this research, three topics related to pyrolysis have been studied. The first topic is the kinetic study of pyrolysis using thermogravimetric analysis (TGA) and the effect of several parameters (particle size, heating rate, and initial weight). We performed the pyrolysis of sawdust biomass at the beginning of the study and from the result we found that the reaction order of the pyrolysis is around 1 (Chapter 3). We also be able to predict the yield of bio-oil, biogas and biochar through the experiment that will be useful for the next stage of the research. Using the same technique, we also pyrolysed the other type of straw biomasses (wheat, oat, flax and barley). In Chapter 4, pyrolysis of straw biomasses has been scale-up to bench scale size experiment (approximately 200 times of TGA scale ) in order to be able to collect certain amount of the products that can be analysed and characterized. The setup was designed using stainless steel horizontal rod surrounded by heating element of a furnace. From the experiment, the yield data of pyrolysis product (biochar, bio-oil and gaseous product) were obtained and follow the prediction data that were obtained using the method that was discussed in Chapter 3. Bio-oil analysis that were obtained in the previous show the oxygen content were high and therefore need un upgrading process to reduce the oxygen content. In the last two chapters (chapter 5 and 6), bio-oil upgrading component were performed using catalytic hydrdeoxygenation (HDO) process. Two of common bio-oil components (guaiacol and anisole) 160 were chosen to represent the majority of phenol type component in the bio-oil and few catalysts were synthesized and screened to achieve high conversion rate and yield. Screening of different catalysts unveiled Ni-Mo/TiO2 as the most active catalyst for the HDO of guaiacol and anisole. The relation between chapters can be seen in the diagram below. Pyrolysis Research Study Pyrolysis using TGA (Chapter 3) Biomass analysis Kinetic analysis Pyrolysis in bench-scale reactor size (Chapter 4) Bio-oil upgrading (Chapter 5 and 6) Reactor design and setup Catalyst synthesis Pyrolysis of straw biomass HDO of Guaiacol & Anisole Kinetic study of HDO reaction Fig. 7.1 Relation between chapters 161 7.2. Conclusions The main objective of this study was to maximize the yield of bio-oil from the pyrolysis of various straw biomasses and to synthesize a catalyst for utilization during the post-treatment phase to increase the quality of the bio-oil. In order to meet this objective, the study was divided into three different parts. In the first part of the study, pyrolysis of sawdust biomass was performed using a thermogravimetric analyzer in order to understand the degradation process of biomass and its kinetics. Several parameters were varied during the experiment to understand their effect. In the second part of the research, pyrolysis of several straw biomasses were conducted in a bench-scale reactor based on the kinetics data obtained previously using a TGA. Three different products (bio-char, bio-oil and gaseous material) were obtained from the process and analyzed using several instruments to measure their heating value and composition. Catalysts were also used in this stage to enhance the yield and quality of the bio-oil based on their respective composition. In the last stage, an upgrading process was conducted on the main components of the bio-oil, i.e. guaiacol and anisole. A hydrodeoxygenation (HDO) reaction was selected with the aim of reducing the oxygen content of bio-oil components to produce lesser oxygenated components such as benzene and cyclohexane. Prior to the reaction, several different metal active catalysts were synthesized using a wet impregnation method and further screened to determine the most efficient catalyst for the HDO reaction. 162 Two catalysts, Ni-Mo/TiO2 and Ni-V/TiO2, were used for both the HDO of guaiacol and anisole to study their activity. Temperature, reaction time and H2/guaiacol molar ratio were varied to study their respective effect on the conversion and selectivity of the HDO reaction. From these studies, we have made several conclusions: The results from the proximate analysis showed that the volatile material evolution was higher with larger particle sizes. In the pyrolysis process, the char yield decreased as the particle size of the sample increased. Reducing the particle size of the sample or increasing the initial weight or heating rate of the pyrolysis process increased char production. The devolatilization process of the sawdust samples during pyrolysis was divided into three non-interacting stages. The first stage was identified as the moisture loss process, the second as cellulose and hemicelluloses devolatilization and the third as being dominated by lignin devolatilization. A modified devolatilization kinetic model was used to obtain the kinetic parameters of the reactions. The reaction orders for the stages 2 and 3 were 0.9 and 1.7, respectively. The results of proximate analyses on raw biomasses can be used to predict the product yield produced during pyrolysis. The total yields of bio-oil and gas were consistently lower than the volatile content produced in the proximate analysis. It was observed that the bio-oil yield during the pyrolysis increased with the use of a zeolite-based catalyst. 163 The HDO of guaiacol was successfully studied using Ni-Mo/TiO 2 and Ni-V/TiO2 catalysts under relatively low pressure (100 psi) and moderate temperatures (200 to 350 °C). Ni (10 wt%)/TiO2 and Mo (10 wt%)/TiO2 showed promising results with reference to unpromoted Ni/TiO2, Mo/TiO2 and variable Ni metal loadings. Incorporation of Mo and V to Ni/TiO2 not only enhanced dispersion of metal species, but also improved the catalytic performance. Variations in the product distribution and conversions were noted for both Ni-Mo/TiO 2 and Ni-V/TiO2 catalysts under different reaction conditions, which may have been attributed to differences in the dissemination of active sites (metallic and acidic sites) with the interference of promoters Mo and V and particle size distribution. With increases in temperature and H2/guaiacol mole ratio, both catalysts showed improved conversion and selectivity. In particular, Ni-Mo/TiO2 showed better selectivity for phenol and cresol, while NiV/TiO2 exhibited more cyclohexanol along with phenol. Well-dispersed, small-sized metal particles, high surface area, mesopore structure and more acidic sites may have been the reasons for the high activity of Ni-Mo/TiO2. In an aqueous medium, Ni-V/TiO2 maintained a better guaiacol conversion and phenol/catechol ratio than Ni-Mo/TiO2 and proved its tolerance to the presence of water. 164 7.3. Suggestions for Future Work Suggestions for future work include: 1) Pressure is another important parameter for pyrolysis. Pyrolysis in higher pressure conditions may have a positive effect on bio-oil production. In order to study this effect, a high-pressure reactor is needed. 2) Nitrogen gas is used in most pyrolysis. Pyrolysis using hydrogen gas (H2) or methane (CH4) should be performed to investigate whether the use of these gases can provide hydrogen and methyl group sources that can alter the product composition of the pyrolysis. There is a possibility that the use of H 2 or CH4 may significantly reduce the oxygen content of bio-oil, hence increasing the quality of the bio-oil. 3) An HDO reaction on two or more bio-oil components (simultaneously) should be studied to learn their interaction and to mimic the real composition of bio-oil. 4) It would also be interesting to find out what impact mixed gases (H 2 and CH4) may have during an HDO reaction of the bio-oil component. 5) Use of combination of feedstock can be done between biomass and rich hydrocarbon waste material such as plastic in order to increase the carbon and hydrogen ratio over oxygen in bio-oil product. 165 APPENDIX A: LICENSE TO RE-PRINT A.1. License to re-print and re-use the content of WIT Publications (Chapter 3 & 4) 166 167 A.2. License to re-print of Chapter 3 – Study of Sawdust Pyrolysis and its Devolatilization Kinetics 168 169 170 171 172 173 A.3. License to re-print of Chapter 5 – Catalytic Hydrodeoxygenation of Guaiacol as Lignin Model Component Using Ni-Mo/TiO 2 and Ni-V/TiO2 Catalysts 174 175 176 177 178