Conversion of Petroleum Coke Into Valuable Products Using Catalytic and Non-Catalytic Oxy-Cracking Reaction

Home , Athabasca oil sands, Petroleum coke

University of Calgary PRISM: University of Calgary's Digital Repository

Graduate Studies The Vault: Electronic Theses and Dissertations

2018-04-20 Conversion of Petroleum Coke into Valuable Products using Catalytic and Non-Catalytic Oxy-Cracking Reaction

Manasrah, Abdallah Darweesh

Manasrah, A. D. (2018). Conversion of Petroleum Coke into Valuable Products using Catalytic and Non-Catalytic Oxy-Cracking Reaction. (Unpublished doctoral thesis). University of Calgary, Calgary, AB. doi:10.11575/PRISM/31818 http://hdl.handle.net/1880/106532 doctoral thesis

University of Calgary graduate students retain copyright ownership and moral rights for their thesis. You may use this material in any way that is permitted by the Copyright Act or through licensing that has been assigned to the document. For uses that are not allowable under copyright legislation or licensing, you are required to seek permission. Downloaded from PRISM: https://prism.ucalgary.ca UNIVERSITY OF CALGARY

Conversion of Petroleum Coke into Valuable Products using Catalytic and Non-Catalytic Oxy-

Cracking Reaction

Abdallah Darweesh Manasrah

A THESIS

SUBMITTED TO THE FACULTY OF GRADUATE STUDIES

IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE

DEGREE OF DOCTOR OF PHILOSOPHY

GRADUATE PROGRAM IN CHEMICAL AND PETROLEUM ENGINEERING

CALGARY, ALBERTA

APRIL, 2018

Abstract

Every year millions of tons petroleum coke (petcoke) is generated as a by-product from bitumen and heavy oil upgrading due to the increasing demand in energy. Petcoke is a carbonaceous solid consisting of polycyclic aromatic hydrocarbons with low hydrogen content, derived from the processing of oil sands and oil refineries. The upgrading and treating of petcoke typically include thermal techniques such as gasification and combustion. However, several challenges limit the effectiveness of these conventional processes such as sulfur and CO2 emissions as well as high energy and costs associated with low efficiency. Therefore, finding an alternative, efficient, environmentally-friendly and cost-effective technology to treat these massive amounts of petcoke is needed. In this study, an oxy-cracking technique, which is a combination of oxidation and cracking reactions, is introduced as an alternative approach for petcoke utilization. This oxy- cracking takes place in basic aqueous media, at mild operation temperatures (170-230 oC) and pressures (500-600 psi). The oxy-cracking reaction mechanism was investigated using Quinolin-

65 (Q-65) as a model molecule mimicking the residual feedstocks. Theoretical calaculations along with experimental reaction were carried out on Q-65 to explore the reaction pathways.

Consequently, several operating conditions on petcoke oxy-cracking were investigated, such as temperature, oxygen pressure, reaction time, particle size and mixing rate to optimize the solubility and selectivity of oxy-cracked products. To enhance the oxy-cracking reaction conversion, an in- house prepared copper-silicate catalyst was introduced and characterized using BET, SEM, FTIR and XRD techniques. The oxy-cracking technique successfully converted the petcoke into valuable products, particularly humic acids analogs with other functional groups such as carboxylic, carbonyl, and sulfonic acids, as confirmed by FTIR, XPS and NMR analyses, in addition to

minimal emission of CO2. Interestingly, based on the experimental findings, the metal contents in the obtained oxy-cracked products are significantly lower than that in the virgin petcoke.

Consequently, the heating value and oxidation behaviour of the oxy-cracked products was investigated using TGA. These results showed that the oxy-cracked petcoke is easier and faster to oxidize compared to the virgin petcoke, suggesting that the oxy-cracked petcoke could be an alternative-clean fuel for power generation.

Keywords: oxy-cracking, residual feedstock, petroleum coke, quinolin-65 (Q-65), copper- silicates, humic acids.

Acknowledgements

First and foremost, all praises to Allah (God) for the blessings and providing me with the strength, knowledge, ability and wisdom to achieve this research study and to persevere and complete it satisfactorily. Without his blessings, this achievement would not have been possible. This thesis also would not have been possible in its current form without supports and encouragements of several people.

I would like to express my heartfelt gratitude and sincere appreciation to my supervisor Dr.

Nashaat N. Nassar at the University of Calgary. Words are insufficient to appreciate and thank him for his whole hearted, exemplary guidance, involvement, inspiration, monitoring, constant encouragement, understanding, care and empathy during the entire period of research work.

My special thanks to Dr. Pedro Pereira (the committee advisor) for his support and invaluable advice. Having the opportunity to work in his well-equipped research labs. Sincere thanks are extended to my examination committee, Dr. Kevin Thurbide, Dr. Nancy Chen, Dr. Pedro Pereira

Almao and Dr. Jan Kopyscinski for their constructive comments and recommendations, helping to improve the quality of this dissertation.

My appreciation goes also to the research associates Mr. Lante Carbognani-Ortega, Dr. Gerardo

Vitale, Dr. Josefina Pérez, Dr. Azfar Hassan, and Dr. Carlos Scott, for their valuable assistance, feedback, experience and constant support.

I would also like to thank all members of Dr. Nassar Group for Nanotechnology Research, Amjad

El-Qanni, Ghada Nafie, Dr. Ismail Badran, Afif Hethnawi, Tatiana Montoya, Fahad Sagala,

Maysam Alnajjar, Mousa Taleb, Redhwan Al-Akbari, Arash Ostovar, and Robert Mustafin for providing a delightful and engaging environment through my years of study at the University of iii

Calgary. All members in the CAFE group, for all the great discussions, advice and feedback.

Particularly, Marianna Isabel Trujillo, for all her help with the NMR, Dr. Monica Bartolini, and

Ms. Josune Carbognani. I am also so thankful to the internship students; Redhwan Al-Akbari,

Brooke Mackay and Hadji Abbass, they had worked closely to my research area. My appreciation goes to the administrative team of the Chemical and Petroleum Engineering (CPE) graduate office,

Mrs. Suha Abusalim, Mr. Arthur and Jamie for their competence and professionalism.

I thankfully acknowledge the financial support of the Natural Sciences and Engineering Research

Council of Canada (NSERC), the Department of Chemical and Petroleum Engineering at the

Schulich School of Engineering at the University of Calgary.

Special thanks to my dear friend Nedal Marei, who has, in his own ways, motivation and help in the University of Calgary related work, care and moral support, assisting me as per his abilities, in whatever manner possible.

In addition, I am so grateful to my coffee-mates and friends; Dr. Hothifa Rojob and Dr. Ala Rajabi and to all my friends who have made my time in Canada more enjoyable than I could have imagined. I will always remember the great memories we had together.

My acknowledgement would be incomplete without thanking the biggest source of my strength, my family in Palestine, my parents, my brother, sisters, and parent’s in-law for their blessings, affection and support during my study. I am sincerely thankful and grateful to my beloved family in Canada; my wife Baraah, My daughter Lujain and my sons Yazed, and Tamim, for the love, confidence, encouragement and sustained support during this work. Finally, I would like to thank everybody who was behind the fulfillment of my thesis. Thank you all.

Dedication To my soulmate and love of my life “Baraah” To my little angles “Lujain, Yazed and Tamim” To my lovely parents which supported me a lot for being what I am in my life. To my brother and sisters. To my second family To my home country, the Holy Land; “Palestine”

Table of Contents

Abstract ...... i Acknowledgements ...... iii Dedication ...... v Table of Contents ...... vi List of Tables ...... x List of Figures and Illustrations ...... xi

Introduction ...... 1

1.1 Background ...... 1

1.2 Production of petcoke ...... 3 Upgrading coker...... 3 Delayed coking ...... 5

1.3 Types and properties of petcoke ...... 7

1.4 Current applications and challenges of petcoke ...... 9 Combustion for power generation ...... 10 Gasification ...... 11 Conversion to activated carbon ...... 12

1.5 Proposed solution - Oxy-cracking technique ...... 12 Oxy-cracking of petcoke ...... 13 Oxy-cracking reaction mechanism ...... 15 Catalytic oxy-cracking reaction ...... 16

1.6 Objectives...... 18

1.7 Thesis organization ...... 19

1.8 References ...... 21

Experimental and Theoretical Studies on Oxy-cracking of Quinolin-65 as a Model Molecule for Residual Feedstocks ...... 27

2.1 Abstract ...... 28

2.2 Introduction ...... 29

2.3 Experimental Work ...... 34 vi

2.3.1 Materials ...... 34 2.3.2 Experimental setup and procedure ...... 35 2.3.3 Characterization ...... 37 2.3.3.1 Fourier Transformed Infrared (FTIR) spectroscopy ...... 37 2.3.3.2 Total organic carbon (TOC) ...... 38 2.3.3.3 Gas chromatography (GC) ...... 38 2.3.4 Theoretical calculations ...... 39

2.4 Results and Discussion ...... 40 Experimental determination of the oxy-cracking reaction kinetics ...... 40 FTIR spectroscopy of virgin and oxy-cracked Q-65 ...... 48 Theoretical modeling ...... 50 2.4.3.1 Reactions of Q-65 with •OH radical ...... 50 2.4.3.2 Reactions of Q-65 with –OH anion...... 65

2.5 Conclusions ...... 67

2.6 Acknowledgments ...... 68

2.7 References ...... 69

Conversion of petroleum coke into valuable products using oxy-cracking techniques ...... 75

3.1 Abstract ...... 76

3.2 Introduction ...... 77

3.3 Experimental work ...... 81 Materials ...... 81 Experimental procedures and setup...... 82 Characterization ...... 83 3.3.3.1 FTIR analysis...... 83 3.3.3.2 Total organic carbon (TOC) analysis ...... 84 3.3.3.3 1H Nuclear magnetic resonance (NMR) spectroscopy ...... 85 3.3.3.4 Gas chromatography (GC) analysis ...... 85 3.3.3.5 X-ray photoelectron spectroscopy (XPS) ...... 85 3.3.3.6 Elemental analysis ...... 86

3.4 Results and Discussion ...... 86 Reaction kinetics ...... 86 Effects of operating conditions on petcoke oxy-cracking reaction ...... 96 3.4.2.1 Effect of the temperature ...... 99 3.4.2.2 Effect of reaction times ...... 101 vii

3.4.2.3 Effect of KOH ...... 102 Characterization results ...... 104 3.4.3.1 FTIR analysis for petcoke and oxy-cracking products...... 104 3.4.3.2 1H NMR analysis of the oxy-cracked petcoke ...... 108 3.4.3.3 XPS results of petcoke oxy-cracking ...... 110 Sulfur and metal analysis ...... 115

3.5 Conclusion ...... 117

3.6 Acknowledgment ...... 117

3.7 References ...... 118

Nanocrystalline copper silicate for catalytic oxy-cracking of petroleum coke ...... 124

4.1 Abstract ...... 125

4.2 Introduction: ...... 126

4.3 Materials and Methods ...... 130 Chemicals and reagents ...... 130 Synthesis of nanocrystalline copper-silicate material ...... 131 Catalyst characterizations ...... 132 Catalytic oxy-cracking of petcoke sample ...... 132 Characterization of oxy-cracking products and the spent catalyst ...... 134 Stability tests...... 135

4.4 Results and discussion ...... 136 Characterization study of the prepared catalyst...... 136 Catalytic activity and selectivity ...... 141 Reaction kinetics and mechanism...... 146 Leaching and stability tests of copper silicate ...... 157 FTIR analysis of the oxy-cracking products ...... 162

4.5 Conclusion ...... 165

4.6 Acknowledgments ...... 166

4.7 References ...... 166

A comparative study of thermal properties and heating values of oxy- cracked and virgin petroleum coke ...... 174 viii

5.1 Abstract: ...... 175

5.2 Introduction ...... 176

5.3 Experimental work ...... 179 Materials ...... 179 Preparation of oxy-cracking petcoke ...... 180 Thermogravimetric analysis ...... 181 Heating value measurements ...... 182 Elemental analysis ...... 183

5.4 Results and discussion ...... 184 Thermo-oxidative decomposition of virgin and oxy-cracked petcoke ....184 Heating values of virgin and oxy-cracked petcoke ...... 192

5.5 Conclusion ...... 197

5.6 Acknowledgment ...... 198

5.7 References ...... 198

Conclusions and Future Work ...... 202

8.1 Conclusions ...... 202

8.2 Recommendations ...... 205

Appendix A1 ...... 206 Supplementary Material ...... 206

S1. Characterization methods ...... 206 S1.1 1H Nuclear magnetic resonance (NMR) spectroscopy ...... 206 S1.2 X-ray photoelectron spectroscopy (XPS) ...... 206

S2. Characterization results ...... 207 S2.1 1H NMR spectroscopy of the oxy-cracked Q-65 ...... 207 S2.2 XPS of Q-65 before and after oxy-cracking ...... 209

S3. References ...... 214

Appendix A2 Effect of the pressure on the oxy-cracking reaction ...... 215

List of Tables

Table 1.1 The chemical composition of the green and calcined coke [50]...... 8

Table 2.1 Activation energies and frequency factors for Q-65 oxy-cracking...... 47

Table 2.2 Activation parameters and unimolecular rate constants for the reactions of •OH and–OH with Q-65b...... 66

Table 3.1 The chemical composition of the green petcoke sample considered in this study...... 82

Table 3.2 Determined values of oxy-cracking reaction constants...... 92

Table 3.3 Estimated activation energies and frequency factors of petcoke oxy-cracking...... 96

Table 3.4 Signal fitting for high-resolution spectra of species in original and oxy-cracked petcoke...... 112

Table 3.5 Elemental content in the virgin, oxy-cracked and residue petcoke at temperature 230 oC, pressure 750 psi and time 2 h...... 116

Table 4.1 Determined values of oxy-cracking reaction constants...... 152

Table 4.2 Estimated activation energies and frequency factors of petcoke oxy-cracking...... 152

Table 4.3 Estimated leached active metal (Cu) from the catalyst at different oxy-cracking reaction temperatures. Experimental conditions: catalyst dose, 0.10 g; reaction time, 1h…………...... 158

Table 5.1 The chemical composition of the virgin and oxy-cracked petcoke sample...... 180

Table 5.2 The thermal properties of the virgin and oxy-cracked petcoke ...... 187

Table 5.3 The proximate and ultimate analysis of the samples...... 194

Table 5.4 The heating values (HHV) for virgin and oxy-cracked petcoke samples...... 197

Table S 1 Fitting signal data of high-resolution spectra of species in virgin and oxy-cracked

Q65……………………………………………………………………………………………..211

List of Figures and Illustrations

Figure 1.1 Alberta oil sands upgrading coke inventory [15]...... 3

Figure 1.2 Schematic representation of a typical delayed coking unit [29]...... 6

Figure 1.3 The potential use of petcoke around the world [53]...... 9

Figure 1.4 Schematic representation of all possible products generated during the petcoke oxy-cracking reaction...... 15

Figure 2.1 Schematic illustration of the experimental setup (not to scale)...... 37

Figure 2.2 Triangular reaction scheme of Q-65 oxy-cracking. A is the Q-65, B is the soluble intermediate (TOC), C is CO2, and D is the inorganic carbon (IC)...... 41

Figure 2.3 Concentrations as a function of time during the Q-65 oxy-cracking at (a) 230 ºC, (b) 215 ºC, and (c) 200 ºC. The symbols represent experimental data, and the dotted lines are from the kinetics model (Equations (2.7-2.9)). Experimental operating conditions: Oxygen partial pressure: 750 psi, impeller speed: 1000 rpm, Q-65 amount: 0.05 g, KOH amount: 0.05 g, and water amount: 30 g...... 45

Figure 2.4 Arrhenius plots for the Q-65 oxy-cracking for each reaction pathway...... 47

Figure 2.5 FTIR spectra of virgin and oxy-cracked Q-65 at 230 °C and 2 h residence time...... 50

Figure 2.6 Optimized structure for the transition state of water formation (TS1). Grey atoms represent carbon, blue atoms represent nitrogen, white atoms represent hydrogen, yellow atoms represent sulfur and red atoms represent oxygen...... 53

Figure 2.8 Optimized structure for the transition states involved in the phenol I formation, a) TS2 and b) TS3. For the colour scheme, refer to Figure 2.6...... 56

Figure 2.10 Optimized structure for the transition states involved in the formation of phenol 2, a) TS4 and b) TS5. For the colour scheme, refer to Figure 2.6...... 58

Figure 2.11 Energy level diagrams for the phenol 2 formation mechanism initiated by •OH attack on Q-65b. Energy values represent the relative Gibbs free energies at 298 K (ZPE corrections included)...... 59

Figure 2.12 Optimized structure for the transition states involved in the formation of the carboxylic acid VII, a) TS6, b) TS7, and c) TS8. d) optimized structures for the acid VII. For the colour scheme, refer to Figure 2.6...... 63

Figure 2.13 Energy level diagrams for the phenol 3 formation mechanism initiated by •OH attack on phenol 2. Energy values represent the relative Gibbs free energies at 298 K (ZPE corrections included)...... 64

Figure 2.14 Energy level diagrams for the acid VII formation mechanism initiated by •OH attack on phenol 3. Energy values represent the relative Gibbs free energies at 298 K (ZPE corrections included)...... 64

Figure 3.1 Schematic representation of the experimental setup (not to scale)...... 83

Figure 3.2 Triangular reaction scheme of petcoke oxy-cracking, where A is the petcoke, B is the intermediates (desired products, TOC), and C: CO2 in the gas phase (CG) + CO2 in the liquid phase (carbonates IC)...... 88

Figure 3.3 Concentrations of A, B, and C as a function of reaction time at different reaction temperature 200 ºC, 215 ºC, and 230 ºC. The symbols represent experimental data, and the solid lines are the kinetics model (Eqs.3.12-3.14)...... 93

Figure 3.4 Arrhenius plots of petcoke oxy-cracking for each reaction pathway...... 96

Figure 3.5 Effect of mixing speed on the conversion of petcoke during oxy-cracking reaction (T = 215 oC, P = 750 psi and t = 2 h)...... 98

Figure 3.6 Effect of petcoke particle size on reaction conversion of petcoke (T = 215 oC, P = 750 psi and t = 2 h)...... 99

Figure 3.7 Effect of the reaction temperature on the selectivity and conversion of petcoke oxy-cracking (P = 750 psi and t = 1 h)...... 101

Figure 3.8 Reaction time effect on selectivity and conversion of petcoke oxy-cracking reaction (T = 180 oC and P = 750 psi)...... 102

Figure 3.9 Effect of KOH amounts on the selectivity and conversion of petcoke oxy- cracking reaction (T = 230 oC and P = 750 psi, time = 2 h)...... 104

Figure 3.10 FTIR spectra of the original petcoke, oxy-cracked products and residual petcoke at 230 °C and 2 h residence time...... 108 xii

1 Figure 3.11 H NMR spectra for oxy-cracked petcoke ran with D2O solvent. Signal frequencies for typical chemical structures appended...... 110

Figure 3.12 High-resolution XPS spectra of the deconvoluted C1s peak (a) before reaction, (b) after reaction...... 113

Figure 3.13 High-resolution XPS spectra of the deconvoluted O1s peak (a) before reaction, (b) after reaction...... 113

Figure 3.14 High-resolution XPS spectra of the deconvoluted N1s peak (a) before reaction, (b) after reaction...... 114

Figure 3.15 High-resolution XPS spectra of the deconvoluted S2p peak (a) before reaction, (b) after reaction...... 114

Figure 4.1 Schematic illustration of the experimental setup (not to scale) [10]...... 133

Figure 4.2 XRD powder patterns of copper-silicate cuprorivaite (blue line), the vertical lines (black) are the reference data for the cuprorivaite from COD database...... 137

Figure 4.3 The unit cell of the copper silicate cuprorivaite framework drawn with BIOVIA structure module [53], a) Unit cell of CaCuSi4O10 b) Side view of the surface (001) of CaCuSi4O10 and c) Top view of the surface (001) of CaCuSi4O10. Blue spheres represent copper atoms, yellow spheres are silicon atoms, red spheres are oxygen atoms and green spheres are calcium atoms...... 138

Figure 4.4 Nitrogen physisorption isotherms for copper-silicate...... 139

Figure 4.5 SEM images of copper-silicate material at different magnifications...... 140

Figure 4.6 Infrared spectroscopy of the prepared copper-silicate material...... 141

Figure 4.7 Effect of the reaction temperature on the selectivity and conversion of petcoke oxy-cracking (P = 750, t = 1 h, 1000 rpm and 0.10 g of catalyst)...... 143

Figure 4.8 Reaction time effect on selectivity and conversion of petcoke oxy-cracking reaction (T = 200 oC and P = 750 psi, 1000 rpm and 0.10 g of catalyst)...... 144

Figure 4.9 The square planar configuration of the copper atoms in the structure of CaCuSi4O10, the blue spheres are copper atoms and red ones are oxygen atoms...... 146

Figure 4.11 Arrhenius plots of petcoke oxy-cracking for each reaction pathway...... 150 xiii

Figure 4.12 Concentration profiles of A, B, and C as a function of reaction time at different reaction temperatures 185, 200, and 230 ºC under the presence of the Cu-silicate catalyst. The symbols represent experimental data, and the solid lines are the kinetics model (Eqs.4.5-4.7)...... 153

Figure 4.13 Schematic diagram for the proposed mechanism of petcoke oxy-cracking over the catalyst...... 156

Figure 4.14 The conversion and selectivity of B and C for three repeated cycles of Cu- silicates, 2 h, 200 oC...... 161

Figure 4.15 Overlays of the X-ray diffraction patterns of the fresh and regenerated catalysts. The red pattern is the regenerated catalysts, blue pattern is cuprorivaite and the green is wollastonite...... 162

Figure 4.16 FTIR spectra of the virgin petcoke, oxy-cracked products and the humic acid at 200 °C and 2 h residence time...... 165

Figure 5.1 Schematic diagram for the preparation of oxy-cracked petcoke sample...... 181

Figure 5.2. TG-DTA curve for the virgin petcoke, showing the ignition, peak and burnout temperatures...... 188

Figure 5.3. TG-DTA curve for oxy-cracked petcoke under air, showing the ignition, peak and burnout temperatures...... 189

Figure 5.4. The heat flow of virgin and oxy-cracked petcoke with temperature...... 190

Figure 5.5. The conversion percent (α) with temperature at heating rates of 5, 10 and 20 oC/min for a) virgin petcoke and b) oxy-cracked petcoke...... 191

Figure 5.6. Thermogravimetric analysis of the virgin and oxy-cracked petcoke at a heating rate of 10 oC/min, showing the M, MV, FC, and A...... 196

Figure S 1. 1H NMR spectra for a) virgin Q-65 ran with DCM, b) oxy-cracked Q-65 ran with

D2O solvent. Signal frequencies for typical chemical structures appended. 208

Figure S2. High-resolution XPS spectra with fitting of C1s (a) before reaction, (b) after reaction...... 212

Figure S3. High-resolution XPS spectra with fitting of O1s (a) before reaction, (b) after reaction...... 212

xiv

Figure S4. High-resolution XPS spectra with fitting of N1s (a) before reaction, (b) after reaction...... 213

Figure S5. High-resolution XPS spectra with fitting of S2p (a) before reaction, (b) after reaction...... 213

Scheme 2.1 Q-65 molecular structure [63]………………………………………………………35

• Scheme 2.2 The reaction mechanism initiated by OH attack on Q-65b to form H2O and phenol 1...... 52

Scheme 2.3 The formation mechanism of Phenol I, initiated by •OH attack on Q-65b...... 55

Scheme 2.4 The formation mechanism of Phenol II, initiated by the •OH attack on the LUMO of Q-65b...... 58

Scheme 2.5 The decomposition mechanism of Q-65b, initiated by the attack of •OH, and leading to the formation of Acid VII...... 62

Introduction

1.1 Background

The global demand for energy is rapidly increasing with the increased world population, modernization, and urbanization [1, 2]. Despite the variety of energy resources and the rise in technologies, the world still relies heavily on crude oil as the main energy source with a projected increased demand of 33% in 2035 [3, 4]. Therefore, there is an increasing demand to upgrade and recover unconventional crude oil, like bitumen and heavy oil, to meet current and future global energy demands.

Canada has the third largest oil reserves in the world after Venezuela and Saudi Arabia estimated at 171 billion barrels of Alberta’s oil sands reserves in unconventional oil and 4.7 billion barrels in conventional oil formations in 2015 [5, 6]. Oil sands are basically a bituminous mixture consisting of approximately 84% sand and clay, 11-12% bitumen and 5% water [7]. Bitumen is a highly viscous oil that cannot be pumped through the pipeline without heating or diluting [8].

Typically, after extracting the oil from the sand, the upgrading of bitumen is necessary to make it transportable through the pipeline and usable by conventional refineries [9]. The bitumen upgrading involves a coking process which is basically a thermal cracking unit in which the heavy hydrocarbons are heated to its thermal cracking temperature generating massive amounts of carbon-rich solid waste hydrocarbon byproduct known as petroleum coke (petcoke) [10, 11].

Petcoke is also generated as a by-product of refineries through either delayed or fluid coking processes at high temperature and pressure [12, 13].

The global production of petcoke has reached about 150 million metric tons per annum and is expected to increase in the future due to the progressively increasing heavier nature of the crudes

[13]. North America alone produces about 70% out of the total petcoke capacity [14]. In fact, based on the Alberta Energy Regulator (AER) statistics, nearly 106 million tons of solid petcoke was being stockpiled in Alberta alone as a by-product of the Canadian oil sands industries in 2016 [15,

16]. As shown in Figure 1.1, inventories remained constant from 1998 to 2000 due to higher on- site use of petcoke by the upgraders; however, this has been followed by a trend of rising stockpile inventories related to increased synthetic crude oil production at upgraders. This amount is projected to increase up to 500 million tons by 2040 if no significant application is found to utilize this by-product [17]. Consequently, approximately 90% of the produced petcoke have been accumulated and stored on-site for many years [15, 17]. Due to limited markets and minimal use for this commodity, the stockpile is growing at a rate of about 4 million tons a year [18, 19].

Therefore, several thermal techniques have been proposed for treating this residue material such as gasification and combustion to convert it into useful products [20, 21]. However, the effectiveness of these conventional processes is limited due to sulfur and CO2 emissions as well as high energy costs combined with low efficiency [22].

Figure 1.1 Alberta oil sands upgrading coke inventory [15].

1.2 Production of petcoke

The processing of conventional and unconventional oils to meet market or pipeline specifications typically involves several processes. Such an important process is “coking,” which is widely implemented in refineries and upgraders. As a result of the increased production of conventional and unconventional oils, large amounts of the byproduct of the coking process petcoke will be increased in Canada and worldwide [23]. The main two processes for producing petcoke are:

Upgrading coker

In Canada, bitumen is viscous at room temperature and acts much like cold molasses. Therefore, bitumen upgrading is necessary to isolate a lighter, more valuable, energy-rich oil product to make it transportable through the pipeline, and thus usable by conventional refineries [9]. The coking

process is basically a thermal cracking process in which the H/C atomic ratio of the product is increased by a carbon rejection mechanism [24]. Hence, multi-reactions that undergo a free radical mechanism are coupled together with cracking and polymerization reactions [11]. Cracking reactions usually produce gas and liquid products which are the most valuable ones, while radical polymerization reactions produce petroleum coke [25].

The main producers of petcoke in Canada are Syncrude Canada Ltd., Suncor Energy Inc., and

Canadian Natural Resources Ltd in Alberta near Fort McMurray. Additionally, a limited amount of petcoke is produced by Nexen from OrCrude gasification while Shell does not produce petcoke from the hydrocracking technology [15]. Approximately 50% of the total produced petcoke in

2016 came from Suncor’s operations while 20% from Syncrude, together with a production rate of 5 million tons per year [26]. Although Suncor and Syncrude are using a portion of the annual petcoke as site fuel 11% and 21%, respectively, the majority of the petcoke is stored on site as plies and remained in disuse [15]. However, the sheer volume of the produced petcoke is eventually exceeded the available space to store it. Over the life of oil sand projects in northern

Alberta, a massive amount of petcoke is produced [27].

Petcoke is regarded by the province as a potential future energy source, however, currently there is no economically feasible use or market for this material. This is mainly due to the high transport costs and the lack of technology for clean and efficient combustion, and thus is often disposed of in stockpiles. These stockpiles pose long-term problems, requiring huge land space, potentially leaching heavy metals, particulate matters, and other contents into the environment. Additionally, dust emissions from petcoke piles impose a serious threat to people in the vicinity [28]. As a result, oil sand operators face the challenge of providing long-term storage for this material in an

environmentally friendly manner that allows for its recovery and utilization for potential future applications.

Delayed coking

Another process producing petcoke is delayed coking which commonly occurs at a temperature range of 415-450 oC, as well as the fluid coking process which uses higher temperatures ranging from 480 to 565 oC [27]. Delayed coking is the most common and attractive process for upgrading the vacuum residue oil as it requires low capital investments in comparison with other upgrading processes. Figure 1.2 shows the schematic diagram of delayed coking process, where the main purpose of this process is maximizing the yield of distillate products and minimizing the coke production [29]. During this process, the feedstock is heated to its thermal cracking temperature in the furnace under high temperature and pressure conditions and finished in the coke drum after passing through the transfer lines [13]. Sequence and simultaneous reactions such as thermal cracking, condensation and polymerization occur in during the process. As cracking continues in the drum, refinery gases, lower boiling point products are produced in vapor phase and separate from the liquid and solids. Consequently, the solid particles start depositing in the drum till it is filled with the solidified coke. Then, the hot mixture from the furnace is switched to a second drum. During the filling of the second drum, the full drum is steamed to further reduce hydrocarbon content of the petcoke, and then the solid petcoke is cut down after quenching with high-pressure water nozzle, where it falls into a pit for reclamation to storage. Worth mentioning that the yield of distillate products does not reach above 40%, while coke production stands around 30% to 40%

[30]. Additionally, the delayed coker’s products require further hydrotreating to remove

heteroatoms and undesirable components such as olefins. Nevertheless, the delay coking process confronts in transportation and logistics and also produce a huge proportion of undesirable coke.

Figure 1.2 Schematic representation of a typical delayed coking unit [29].

Moreover, many research groups have looked at upgrading processes to convent petroleum residue, vacuum residue, solvent-deasphalter pitch, rich asphaltenes and resin fuel, to more desirable products like LPG, naphtha, diesel and gas oil [24]. These methods are highly dependent upon their thermal treatments such as carbon rejection and hydrogen addition [31]. Carbon rejection processes, for example, include pyrolysis (e.g., visbreaking, coking and rapid thermal pyrolyzer) [32], separation or extraction (e.g., solvent deasphalting process, SDA) [33] and cracking (e.g., VGO and VR FCC) [34] while the hydrogen addition process includes hydrocracking and steam cracking gasification [35, 36]. Worth noting that the carbon rejection processes have simple configurations compared to hydrogen addition processes and do not require

hydrogen injection and catalyst loading [24, 37]. However, carbon rejection produces lower yields and generates significant coke waste.

1.3 Types and properties of petcoke

The produced petcoke constitutes about 2% of the overall oil production which can be classified into calcined and green coke [38]. The latter one is the initial product of the coking process that can be used as fuel in metallurgical and gasification processes [39, 40]. Calcined coke, on the other hand, is produced by treating the green coke at a high temperature between 1200 and 1350 oC, which can be used to produce the carbon anodes in the aluminium industry [40, 41]. The calcination process usually improves the quality of petcoke by enhancing the C/H ratio, removing the ash and reducing the sulfur content, and thus minimizing the shrinkage during anode baking, and increasing coke strength [42, 43]. It is worth mentioning that the physical, structure, and morphology properties of petcoke strongly depend on the source of oil, type of coking processes and conditions [44]. Oil sands companies often use different refining methods to produce synthetic crude oil. This generates different types of coke similar in chemical composition, however, different in physical texture [38]. Suncor Energy Inc, for example, produces “delayed” coke by rapid heating of oil at certain temperature and pressure which has a sponge-like structure [45].

Generally, the delayed coke can be classified, based on its morphological characteristics, as shot, sponge or needle coke. Shot coke is always hard-solid with spherical shape while the sponge coke is dull and black with porous and amorphous structure and the needle coke is silver-gray, having crystalline structure [46]. In contrast, Syncrude’s coke is known as “fluid” coke and produced when oil is sprayed into steam injection through an instantaneous conversion from liquid oil to

solid coke via thermal cracking [47]. The produced coke displays a uniform, spherical shape with highly-graphitized layers and it is often described to have an onion-like structure [48].

Generally, the produced petcoke has a high carbon content (80-85wt%) consisting of polycyclic aromatic hydrocarbons with other heteroatoms, such as sulfur, nitrogen, and oxygen, and some metals are also presented [49]. It is worth noting that petcoke has a lower amount of ash, moisture, and volatiles matters compared to coal [25]. However, it contains high amounts of sulfur (4-8 wt%) and vanadium (~ 700 ppm), which can severely impact human and animal health [28]. Table 1.1 shows the chemical composition of two types of petcoke based on the petroleum feedstock and the coking methods employed [50].

Table 1.1 The chemical composition of the green and calcined coke [50].

Composition (wt%)

Green coke Calcined coke

Carbon 89.58 – 91.80 98.4

Hydrogen 3.71 – 5.04 0.14

Oxygen 1.30 – 2.14 0.02

Nitrogen 0.95–1.2 0.22

Sulfur 1.29 – 3.42 1.20

Ash 0.19 – 0.35 0.35

C/H atomic ratio 18:1 – 24:1 910:1

1.4 Current applications and challenges of petcoke

The coking process directly produces a green coke which can be exposed to an extra thermal process (calcination) to remove the volatile matter and then increase the carbon content [51]. In practical application, depending on its quality, the green petcoke is widely used in many industrial applications, such as a source of energy and carbon [52]. Figure 1.3 summarized the worldwide potential applications of petcoke [53]. Nearly 80% of the produced petcoke in world is used as a source of fuel in power plant electricity and cement kilns as demonstrated in Figure 1.3. The calcined petcoke can be also used as an energy source for lime production, brick and glass manufacturers due to the high carbon content. However, the main non-energy applications of petcoke are the manufacture of carbon anodes for the aluminium smelting industry, graphite electrodes for steel production and also nonferrous industries. For instance, in the smelting aluminium, the carbon anode is acting as an electrical conductor and a source of carbon in the electrolytic cell to reduce the alumina into aluminium metal using the Hall-Heroult process [54].

Figure 1.3 The potential use of petcoke around the world [53].

Combustion for power generation

Due to the high calorific value (30–35 MJ/kg), low-price, high availability, and low ash content, the petcoke is considered a valuable fuel resource for the future application particularly in North

America [55]. Currently, the use of petcoke in the energy sector for heat generation purposes is one of the most alternative application. Therefore, cement and power plants are currently the two greatest consumers of petcoke. A number of thermal power plants around the world, particularly in developing countries, are still combusting petcoke as a fuel especially with a high-sulfur content to produce steam for generating the electricity and heating the boilers [56]. As a consequence, burning virgin petcoke (without further modification) poses a number of environmental problems such as high SO2 and NOx emissions as well as issues of corrosion due to the presence of considerable amounts of sulfur, vanadium and nickel [57, 58]. Also, the behavior of petcoke in any combustion process is difficult to predict. This is because the physical characteristics of petcoke mainly depends on the coking conditions and particularly on the composition of feedstock from where it is derived [59, 60]. Even though the petcoke has a high heating value, it’s combustion in the power generation purpose is limited by the environmental regulations due to the large invariable emissions of sulfur dioxide (SO2) and nitrogen oxides (NOx) [27]. Additionally, the high potential of CO2 emission in burning petcoke is a major concern due to the high carbon content. Nevertheless, the current increasing use of natural gas as fuel in boilers for heat and power production has decreased the demand for petcoke. Therefore, new alternatives for utilization of petcoke are interested and of high amount importance.

Gasification

Gasification process was considered as an effective option for treating this massive amount of petcoke. The gasifying of petcoke offers refiners a variety of product slates mainly via gaseous products with a useful heating value such as syngas, which is primarily a mixture of carbon monoxide (CO) and hydrogen (H2) [21]. Typically, gasification involves series of exothermic chemical reaction of carbon with steam, air or oxygen at high temperatures (800-1800 °C) and pressures (145-1450 psi) to produce syngas, which can be used for either power generating or synthesis of chemicals, liquid and gaseous fuels [37]. In most cases, coal is the main feedstock for gasification; however, a large-scale of petcoke was introduced for the first time in the gasification unit by Wabash River Energy Ltd in 1995 [61]. As of today, the best-known gasification process for heavy residue fraction upgrading is the Flexicoking process developed by ExxonMobil [24].

However, given the uncertainty of the optimum gasifier configuration, the lack of industrial-scale gasifiers to confirm the feasibility and reliability of gasification process and the significant coke/ash content generated from the gasifier that still requires stockpiling or landfilling, the production of low-BTU waste gas gasification is neither a green nor a reliable method for solid- waste conversion [27, 62]. Additionally, the high production of CO2 during petcoke gasification is also another major environmental concern. Thus, currently there is no efficient solution for dealing with the remaining soot after gasification. Finally, considering the current price of natural gas and the recent increase in shale gas production, producing syngas from coke does not appear to be an economically viable solution for coke utilization.

Conversion to activated carbon

Apart from using petcoke as fuel for direct combustion or gasification, another potential use of petcoke is the production of activated carbon due to its high carbon content. Recently, many studies have demonstrated that petcoke can be activated thermally, physically and chemically in the presence of a reagent to produce activated carbon with a high specific surface area (~500-1500 m2/g) [63]. This prepared material can be used as an adsorbent or a catalyst support. As reported, the development of carbon-based catalysts has gained great interest due to the increased capacity in hydroprocessing to meet the high demands in the energy, heavier crude feeds and low-sulfur transportation fuel [17]. Again, the benefit of this carbon-based material comes from it is easily to gasify once it has reached the end of its life. However, the costs of generating activated carbon through either physical or chemical activation processes tend to be quite high, especially when the yield does not balance the generating costs, thus the total investment cost would be challenging for large-scale industries [64]. Moreover, the corrosive nature of the chemical activating agents has negative environmental impacts and often limits its application [65].

1.5 Proposed solution - Oxy-cracking technique

Despite all the attempts and processes that have been proposed in the industry for developing efficient and environmentally-friendly technologies to treat and reduce of petcoke, a massive amount of this material is still present. Therefore, researchers are still striving for developing an efficient, environmentally-friendly and costly effective technology to utilize this material. To overcome this challenge, the oxy-cracking process is introduced as an alternative technique to treat and convert these solid heavy hydrocarbons into valuable commodity products. This method is a combination of consecutive oxidation and cracking reactions that are taking place in an aqueous

alkaline medium at mild temperatures (170-230 °C) and pressures (500-750 psi), creating the name

“oxy-cracking” [66, 67]. At these reaction conditions, the organic compounds, in this case petcoke, are decomposed via free radical mechanisms to intermediates (e.g., carboxylic acids and other low molecular weight organic compounds) and a small amount of CO2. Recently, the oxy-cracking of petcoke was motivated by Ashtari et al. [68] for converting n-C7 asphaltenes into light commodity products. In fact, the concept of this process was inspired by oxidation and ozonolysis studies on solid heavy hydrocarbons such as asphaltenes, lignites and coal [69-71]. For instance, numerous researchers investigated the conversion of lignites and coal to humic substances in an acidic media using oxidizing agents such as KMNO4 and HNO3 [72]. As reported in several patents, producing the humic substances via pre-oxidizing of coal is industrially possible [70, 73]. Through this oxidation mechanism, coal is converted to humic acids, then treated with formaldehyde and alkali solution of bisulfate at 100 °C to form water-soluble products. Likewise, oxidation of asphaltenes in water under subcritical and supercritical conditions has been reported [74]. In this sense, the oxygen will react with the formed radicals and increase the reactivity of the oxidized species. The goal of water is to inhibit direct oxidation of asphaltenes to CO2, thus desulfurized products are obtained [75]. It was also reported that the reduction in the asphaltenes molecular weight was up to 50 % and aromatic rings cleavage was cracked up to 50 %. The obtained products have low boiling point and viscosity, which could be used as fuel oil or as a solvent, according to the authors

[76].

Oxy-cracking of petcoke

The oxy-cracking technique is believed to offer a new reaction pathway into an aqueous solution, thus solubilizing the petcoke as organic acids and the like. Such a new reaction pathway can be

achieved by changing the oxidation conditions, temperature and pressure, which will alter the mechanism and result in different product distributions. Consequently, via oxy-cracking conditions (i.e., temperatures 170-230 °C and pressures 500-750 psi), these solid heavy hydrocarbons (petcoke) do not completely oxidize to CO2 but instead form intermediate compounds as oxygenated hydrocarbons with different functionalities of carboxylic, sulfonic, and phenolic compounds [68, 77]. In this case, those partially oxidized intermediates become soluble in water via oxygen incorporation due to the polar functionalization of the aromatic edges and paraffinic terminal carbons. Therefore, the oxy-cracking reaction enhances the tendency of large heavy hydrocarbon molecules to disaggregate, making them more accessible to subsequent hydrogenation and cracking at lower temperatures [78]. The importance of this method is not only in converting the petcoke into valuable products but also could be a useful technique for petcoke demineralization and desulfurization [66]. Figure 1.4 shows the possible products of oxy-cracked petcoke as presented in the three phases. As seen, the gas phase remained as predominantly oxygen and an insignificant amount of CO2 while the liquid phase contained the oxygenated hydrocarbons like humic materials. Finally, the residue (solid phase) consisted of the minerals suggesting that the proposed technique could be employed for petcoke demineralization. The success of oxy- cracking reaction is evaluated by making the reaction favorable toward the desired products which strongly depends on the operating temperature and pressure as well as the reaction time and the basicity of the solution. Because of the high solubility and selectivity of oxy-cracked products in the alkaline medium and the insignificant amount of produced CO2, the oxy-cracking process could be considered an environmentally friendly one.

Gas phase

O2 CO2

Humic Acid

Oxy-cracking Liquid phase

O2, H2O,T & P solublized Fuel Petcoke petcoke

Solid phase K Fe Co S Ni V Mo Recovery Minerals

Figure 1.4 Schematic representation of all possible products generated during the petcoke oxy- cracking reaction.

Oxy-cracking reaction mechanism

Similar to the wet oxidation process, the oxy-cracking technique basically involves the generation of a free hydroxyl radical (•OH) which plays an essential role in the oxidation of hydrocarbons

[79]. Therefore, the oxy-cracking undergoes a parallel-consecutive reaction in which an oxidative decomposition took place in the first step producing different types of intermediates as oxygenated hydrocarbons. Those intermediates (desired products) formed in in the liquid phase are likely analogs of humic acids, carboxylic acids, and their corresponding salts, or other products [68, 77,

80, 81]. The general mechanism of such process is believed to involve the following primary initiation steps for generating the free radicals:

• • H2O + O2 → HO2 + OH Ea = 68.6 kcal/mol [79]

• • HO2 + H2O → H2O2 + OH Ea = 32.8 kcal/mol [79]

• • • • H2O + O → OH + OH Ea = 17 kcal/mol [82]

These initiation steps are followed by different propagation and termination pathways. As reported by many studies [79, 83, 84] indicated that a large range of intermediates and oxidized products may result from the above steps. After generating the HO•, the organic compounds are degraded producing intermediates by either adding to the double bond as in aromatic compounds and olefins or by hydrogen abstraction to form water as with alkanes or alcohols. The prevailing products in the liquid phase are alkyl hydroperoxides, which are quickly converted into hydroxylated hydrocarbons, ketones/aldehydes, and the like [79].

Catalytic oxy-cracking reaction

According to our studies, the degree of petcoke conversion through oxy-cracking is highly influenced by the reaction temperature, residence time and the complexity of heavy waste hydrocarbons [66, 68]. Therefore, the high-energy consumption and capital investment may be required for the oxy-cracking operation. Additionally, although a relatively small amount of CO2 might be released through this process, any proposed technique must meet the environmental regulations. To meet these requirements, an effective catalyst is mandated in the oxy-cracking reactions for high selectivity, activity and stability [85, 86]. In recent years, many types of homogeneous and heterogeneous catalysts have been studied in the liquid phase oxidation particularly for wet air oxidation of organic-based pollutants present in wastewater [87, 88]. The 16

need for an extra separation step of a homogeneous catalyst has been the driving force for the development of active and stable heterogeneous catalysts for water phase applications and the reusability of the catalyst. Therefore, many researchers have studied the transition metals/oxides and noble metals catalysts in the wet air oxidation such as CuO, CoO, MnO, and ZnO [84]. Among of them, copper was found to be the most eminent catalyst in wet air oxidation due to its high activity, diminished electro-migration effect and low cost [89, 90]. Nickel, on the other hand, offers high catalytic activity in the hydrocarbons oxidation; however, it is prone to coke formation and sintering [91]. However, the reported metallic oxides, were suffering from deactivation during the oxidation reaction due to the metal leaching caused by the hot acidic media [92-94]. Thus, the activity and stability of these materials for the technological application might be lost under the reaction conditions because of carbon deposition [95-99]. Nevertheless, the noble metals catalyst when supported on alumina, silica, titania and zirconia or carbon materials, showed more promising catalytic activity than the transition metals. However, these noble metals are costly and sensitive to poisoning and surface deactivation [100-101].

Despite of the vigorous research in this field, there is still a need for an active, stable and cost- effective catalyst for industrial applications. Based on that, it is obvious that the developing of catalyst with high activity, selectivity, and stability is a current challenge in liquid-phase oxidation.

Therefore, it is believed that the incorporation of transition metal ions, especially the copper, into frameworks or cavities of ordered micro or mesoporous materials like silicates would be a promising and proposed route. This is because the naturally occurring silicates have proven to be the most attractive material in terms of adsorption and energy storage processes [102, 103].

Consequently, silicates were confirmed to be a promising candidate as a catalyst for the oxidation

reaction due to its chemical and thermal stability, great abundance, nontoxicity, environmentally- friendlier materials, and for its unique physical-chemical properties [103, 104]. Hence, this suggests that the using of copper-silicate materials for the petcoke oxy-cracking could be a promising strategy to develop an efficient catalyst for converting the solid waste heavy hydrocarbons into humic acid analogs. The principle of oxy-cracking reaction during catalytic and noncatalytic, as well as the reaction mechanism and application, will be addressed comprehensively in the upcoming chapters of this thesis.

1.6 Objectives

The main objective of this thesis is to develop an oxy-cracking technique as a new approach for converting petcoke into valuable commodity products. The specific objectives are:

1. Investigating the reaction mechanism and kinetics of the oxy-cracking process using

Quinolin-65 (Q-65) as a model molecule mimicking the solid heavy hydrocarbon. Confirm

the reaction mechanism; theoretically by applying the density functional theory (DFT) and

the second-order Møller–Plesset perturbation theory (MP2), and experimentally by

carrying out the reaction in a Parr reactor.

2. Converting the petcoke into valuable products using the catalytic and noncatalytic oxy-

cracking reaction in a batch reactor under mild operating conditions.

3. Optimizing the reaction conversion, solubility and selectivity of oxy-cracked products by

investigating several reaction conditions, such as temperature, oxygen pressure, reaction

time, particle size and mixing rate.

4. Characterizing the oxy-cracked products using FTIR, XPS and NMR analyses as well as

the insolubilized (residue) petcoke.

5. Developing the oxy-cracking reaction kinetics mechanism of petcoke using the generalized

“triangular” lumped kinetic model to determine the reaction kinetics parameters.

6. Synthesizing a nanocrystalline copper-silicate (CaCuSi4O10) material as an effective,

stable and environmental-friendly heterogeneous catalysts to enhance the selectivity and

conversion of the oxy-cracking reaction. The in-house prepared catalyst is characterized

using different techniques, namely: XRD, BET, SEM, and IR spectroscopy.

7. Validating the suitability of the prepared oxy-cracked petcoke as a feedstock for power

generation by comparing the heating value and thermo-oxidative decomposition of the oxy-

cracked products with the virgin petcoke.

1.7 Thesis organization

This is a manuscript-based thesis consisting of six chapters and two appendices including four journal articles. Two were published in peer-reviewed journals while the third one has been recently submitted and the fourth one is on its way to submission. The thesis chapters are as the following:

Chapter One provides a general introduction about petcoke, current technologies for petcoke handling, my proposed solution, and the objectives of the research.

Chapter Two explores the reaction mechanism and kinetics of the oxy-cracking technique experimentally and theoretically using the Quinolin-65 (Q-65) as a model molecule of the residual feedstock. The oxy-cracking reaction is experimentally performed in a Parr batch reactor.

Meanwhile, the theoretical study employed by computational modeling using the density functional theory (DFT) and the second-order Møller–Plesset perturbation theory (MP2) under the experimental conditions. 19

Chapter Three presents the oxy-cracking technique as an alternative approach to convert the real sample of heavy hydrocarbons (petcoke) into valuable commodity products. In this chapter, several operating conditions are investigated, such as temperature, oxygen pressure, reaction time, particle size and mixing rate to optimize the solubility and selectivity of oxy-cracked products with minimal emission of CO2. The kinetic modeling of petcoke oxy-cracking is provided based on the radical mechanism using the generalized “triangular” lumped kinetic model. Full characterizations of oxy-cracked products (desired products) and the residual solids are addressed here using FTIR,

NMR and XPS techniques. The metal and sulfur distributions in the oxy-cracked products are also investigated.

Chapter Four introduces a potential catalyst for the oxy-cracking reaction. Copper-silicate

(CaCuSi4O10) material was chosen to enhance the selectivity and conversion of the oxy-cracked products. The catalyst activity, selectivity and stability are investigated in a batch reactor at different reaction conditions. The leaching of copper ions from the catalyst in the solution as well as the reusability of catalyst are investigated. The oxy-cracked products are compared with a sample of commercial humic acid using IR-Spectroscopy.

Chapter Five introduces the application of oxy-cracked products as a new alternative fuel for power generation. The Chapter also presents the thermo-oxidative decomposition of oxy-cracked products using TGA analysis and measuring the heating value of the materials.

Chapter Six concludes the thesis with several suggested recommendations for possible continuation of research in this area.

Appendix A1 is included at the end of the thesis which provides supporting information for the first published paper with detailed NMR and XPS analysis for Chapter Two.

Appendix A2 includes the effect of pressure on the oxy-cracking reaction for Chapter Three.

It should be noted that this thesis is a manuscript-based, as such, there might be some repetitions between the chapters particularly in the introduction parts or in the experimental procedure, materials, and analyses sections.

1.8 References

[1] B. Liddle, Impact of population, age structure, and urbanization on carbon emissions/energy consumption: evidence from macro-level, cross-country analyses, Population and Environment 35(3) (2014) 286-304. [2] J. Conti, P. Holtberg, J. Diefenderfer, A. LaRose, J.T. Turnure, L. Westfall, International Energy Outlook 2016 With Projections to 2040, USDOE Energy Information Administration (EIA), Washington, DC (United States). Office of Energy Analysis, 2016. [3] IEA., World Energy Outlook 2017, Organisation for Economic Co-operation and Development, OECD2017. [4] I. Energy, C. Change, World Energy Outlook Special Report (2015), C) OECD/IEA (2017) 37. [5] M. Stewart, Removal of Organic and Inorganic Contaminants from Oil Sands Tailings using Carbon Based Adsorbents and Native Sediment, University of Alberta (Canada) 2013. [6] Y. Zhang, W. Shotyk, C. Zaccone, T. Noernberg, R. Pelletier, B. Bicalho, D.G. Froese, L. Davies, J.W. Martin, Airborne petcoke dust is a major source of polycyclic aromatic hydrocarbons in the Athabasca Oil Sands Region, Environmental science & technology 50(4) (2016) 1711-1720. [7] H. Ibrahim, Determination of the Calorific Value of Syrian Delayed Petroleum Coke, Int J Petrochem Sci Eng 1(3) (2016) 00012. [8] R.F. Meyer, E. Attanasi, Natural bitumen and extra-heavy oil, 2004 Survey of Energy Resources (Twentieth Edition), Elsevier 2004, 93-117. [9] L.F. Baker, The effects of petroleum coke amendments on macrophytes and aquatic invertebrates in northern Alberta, Canada constructed wetlands, Masters Abstracts International, 2007. [10] B.W. Gamson, Petroleum coke production, US Patent 3,684,697, 1972. [11] R.P. Dutta, W.C. McCaffrey, M.R. Gray, K. Muehlenbachs, Thermal cracking of Athabasca bitumen: influence of steam on reaction chemistry, Energy & fuels 14(3) (2000) 671-676. [12] H.G. Graf, H.R. Janssen, Process for improving product yields from delayed coking, US Patent 4,518,487, 1985. [13] K. Wisecarver, Delayed Coking, Springer Handbook of Petroleum Technology, Springer2017, 903-913. [14] W. Tao, Manging China’s Petroleum Coke Problem, (2015). [15] Alberta Energy Regulator, 2017. https://www.aer.ca/data-and-publications/statistical- reports/crude-bitumen-production. (Accessed 15 February 2018). [16] M.T. Ityokumbul, Experimental evaluation of molten caustic leaching of an oil sand coke residue, The Canadian Journal of Chemical Engineering 72(2) (1994) 370-374. 21

[17] S.C. Liew, Impacts of Vanadium and Coke Deposition on CO2 Gasification of Nickel Catalysts Supported on Activated Carbon from Petroleum Coke, University of Calgary, 2014. [18] P. Pourrezaei, A. Alpatova, P. Chelme-Ayala, L. Perez-Estrada, M. Jensen-Fontaine, X. Le, M.G. El-Din, Impact of petroleum coke characteristics on the adsorption of the organic fractions from oil sands process-affected water, International Journal of Environmental Science and Technology 11(7) (2014) 2037-2050. [19] H. Jang, T.H. Etsell, Morphological and mineralogical characterization of oil sands fly ash, Energy & fuels 19(5) (2005) 2121-2128. [20] A. González, N. Moreno, R. Navia, CO2 carbonation under aqueous conditions using petroleum coke combustion fly ash, Chemosphere 117 (2014) 139-143. [21] V. Nemanova, A. Abedini, T. Liliedahl, K. Engvall, Co-gasification of petroleum coke and biomass, Fuel 117 (2014) 870-875. [22] A.I.-E.a.E. Solutions, Executive Summary, 2017. http://eipa.alberta.ca/. (Accessed March 2017). [23] W. Zubot, M.D. MacKinnon, P. Chelme-Ayala, D.W. Smith, M.G. El-Din, Petroleum coke adsorption as a water management option for oil sands process-affected water, Science of the Total Environment 427 (2012) 364-372. [24] M.S. Rana, V. Samano, J. Ancheyta, J. Diaz, A review of recent advances on process technologies for upgrading of heavy oils and residua, Fuel 86(9) (2007) 1216-1231. [25] S.J. Yoon, Y. C. Choi, S. H. Lee, J. G. Lee, Thermogravimetric study of coal and petroleum coke for co-gasification, Korean Journal of Chemical Engineering 24(3) (2007) 512-517. [26] P.M. Fedorak, D.L. Coy, Oil sands cokes affect microbial activities, Fuel 85(12-13) (2006) 1642-1651. [27] J. Wang, E.J. Anthony, J.C. Abanades, Clean and efficient use of petroleum coke for combustion and power generation, Fuel 83(10) (2004) 1341-1348. [28] J.A. Caruso, K. Zhang, N.J. Schroeck, B. McCoy, S.P. McElmurry, Petroleum coke in the urban environment: A review of potential health effects, International journal of environmental research and public health 12(6) (2015) 6218-6231. [29] Delayed Coking Technology, 2017. https://www.scribd.com/document/353395529/Delayed- Coking-Technology. (Accessed 10 Januray 2018). [30] N. Cusher, P. Greenouch, J. Rolfe, J. Weiszmam, Handbook of Petroleum Refining Process, McGraw-Hill, New York, 1986. [31] A.N. Sawarkar, A.B. Pandit, S.D. Samant, J.B. Joshi, Petroleum residue upgrading via delayed coking: A review, The Canadian Journal of Chemical Engineering 85(1) (2007) 1-24. [32] F. Dautzenberg, J. De Deken, Reactor developments in hydrotreating and conversion of residues, Catalysis Reviews Science and Engineering 26(3-4) (1984) 421-444. [33] O.R. Koseoglu, A. Bourane, System and process for integrated oxidative desulfurization, desalting and deasphalting of hydrocarbon feedstocks, 2011/0,226,666, 2015. [34] J.G. Speight, New approaches to hydroprocessing, Catalysis Today 98(1-2) (2004) 55-60. [35] I. Mochida, X.Z. Zhao, K. Sakanishi, Catalyst deactivation during the hydrotreatment of asphaltene in an Australian brown coal liquid, Fuel 67(8) (1988) 1101-1105. [36] S. Maity, V. Perez, J. Ancheyta, M. Rana, G. Centeno, Effect of asphaltene contained in feed on deactivation of Maya crude hydrotreating catalyst, Petroleum Science and technology 25(1-2) (2007) 241-249. 22

[37] N.N. Nassar, A. Hassan, P. Pereira-Almao, Application of nanotechnology for heavy oil upgrading: Catalytic steam gasification/cracking of asphaltenes, Energy & Fuels 25(4) (2011) 1566-1570. [38] S. Öztürk, Improvement of anode properties by coke modification with additives, Université du Québec à Chicoutimi, 2017. [39] J. Menéndez, J. Pis, R. Alvarez, C. Barriocanal, E. Fuente, M. Diez, Characterization of petroleum coke as an additive in metallurgical cokemaking. Modification of thermoplastic properties of coal, Energy & fuels 10(6) (1996) 1262-1268. [40] F.J. Farago, D.G. Retallack, R.R. Sood, Calcination of coke, Google Patents, 1977. [41] A. Sarkar, D. Kocaefe, Y. Kocaefe, D. Bhattacharyay, D. Sarkar, B. Morais, Effect of crystallinity on the wettability of petroleum coke by coal tar pitch, Energy & Fuels 30(4) (2016) 3549-3558. [42] A. Gagnon, N. Backhouse, H. Darmstadt, E. Ryan, L. Dyer, D.G. Dixon, Impurity removal from petroleum coke, Light Metals 2013, Springer2016, 1057-1062. [43] L.P. Lossius, K.J. Neyrey, L.C. Edwards, Coke and Anode Desulfurization Studies, Essential Readings in Light Metals, Springer2016, 136-141. [44] E. Furimsky, Characterization of cokes from fluid/flexi-coking of heavy feeds, Fuel processing technology 67(3) (2000) 205-230. [45] B. Sarkar, Adsorption of Single-ring Model Naphthenic Acid from Oil Sands Tailings Pond Water Using Petroleum Coke-Derived Activated Carbon, University of Toronto (Canada), 2013. [46] S. Birghila, I.C. Popovici, A. Dumitru, Study on physical-chemical properties of petroleum cokes, Romanian Journal of Physics 56(7-8) (2011) 976-82. [47] H. Onder, E.A. Bagdoyan, Everything you always wanted to know about petroleum coke: a handbook, Allis Mineral Systems-Kenndy Van Saun1993. [48] C.C. Small, A.C. Ulrich, Z. Hashisho, Adsorption of acid extractable oil sands tailings organics onto raw and activated oil sands coke, Journal of Environmental Engineering 138(8) (2012) 833-840. [49] J.-H. Lv, X.-Y. Wei, Y.-H. Wang, T.-M. Wang, J. Liu, D.-D. Zhang, Z.-M. Zong, Characterization of condensed aromatics and heteroatomic species in Yanshan petroleum coke through ruthenium ion-catalyzed oxidation using three mass spectrometers, RSC Advances 6(66) (2016) 61758-61770. [50] A. Andrews, R.K. Lattanzio, Petroleum coke: industry and environmental issues, Congressional Research Service, Accessed on (2014). [51] M. Brym, A. Gagnon, C. Boulanger, D. Lepage, G. Savard, G. Bouchard, C. Lagace, A. Charette, Anode reactivity: Effect of coke calcination level, Light Metals (2009) 905-908. [52] S. Birghila, I.C. Popovici, Evaluation of the physical-chemical properties in petroleum coke, Advances in Petroleum Exploration and Development 6(2) (2013) 28-31. [53] American Fuel & Petrochemical Manufacturers 2015. http://www.afpm.org/policy-position- petroleum-coke. (Accessed 10 October 2016). [54] K. Grjotheim, C. Krohn, M. Malinovsky, K. Matiasovsky, J. Thonstad, Aluminum Electrolysis. The Chemistry of the Hall-Heroult Process, Aluminium- Verlag GmbH, Dusseldorf, Germany. 1977, 350 (1977). [55] S.H. Lee, C.S. Choi, Chemical activation of high sulfur petroleum cokes by alkali metal compounds, Fuel processing technology 64(1-3) (2000) 141-153. 23

[56] N. Berguerand, A. Lyngfelt, The use of petroleum coke as fuel in a 10 kWth chemical-looping combustor, International Journal of Greenhouse gas control 2(2) (2008) 169-179. [57] C. Cain-Borgman, Petroleum coke refueling an economic alternative for existing oil-fired boilers, Air Products and Chemicals, Inc., 2009. [58] K. Jayaraman, I. Gokalp, Gasification characteristics of petcoke and coal blended petcoke using thermogravimetry and mass spectrometry analysis, Applied Thermal Engineering 80 (2015) 10-19. [59] K.S. Milenkova, A.G. Borrego, D. Alvarez, R. Menéndez, H.I. Petersen, P. Rosenberg, Coal blending with petroleum coke in a pulverized-fuel power plant, Energy & fuels 19(2) (2005) 453- 458. [60] Q. He, R. Wang, W. Wang, R. Xu, B. Hu, Effect of particle size distribution of petroleum coke on the properties of petroleum coke–oil slurry, Fuel 90(9) (2011) 2896-2901. [61] C. Keeler, Operating experience at the Wabash River repowering project, Conference Paper, Gasification Technologies Conference, San Francisco, October, 2003, 12-15. [62] Y. Ando, K. Yoshikawa, M. Beck, H. Endo, Research and development of a low-BTU gas- driven engine for waste gasification and power generation, Energy 30(11-12) (2005) 2206-2218. [63] C. Small, Activation of delayed and fluid petroleum coke for the adsorption and removal of naphthenic acids from oil sands tailings pond water, (2011). [64] G. Stavropoulos, A. Zabaniotou, Minimizing activated carbons production cost, Fuel processing technology 90(7-8) (2009) 952-957. [65] A. Awoyemi, Understanding the adsorption of polycyclic aromatic hydrocarbons from aqueous phase onto activated carbon, 2011. [66] A.D. Manasrah, N.N. Nassar, L.C. Ortega, Conversion of petroleum coke into valuable products using oxy-cracking technique, Fuel 215 (2018) 865-878. [67] A.D. Manasrah, A. El-Qanni, I. Badran, L.C. Ortega, M.J. Perez-Zurita, N.N. Nassar, Experimental and theoretical studies on oxy-cracking of Quinolin-65 as a model molecule for residual feedstocks, Reaction Chemistry & Engineering 2(5) (2017) 703-719. [68] M. Ashtari, L. Carbognani Ortega, F. Lopez-Linares, A. Eldood, P. Pereira-Almao, New pathways for asphaltenes upgrading using the oxy-cracking process, Energy & Fuels 30(6) (2016) 4596-4608. [69] V. Platonov, A. Kudrya, S. Proskuryakov, Ozonolysis of asphaltenes from semicoking tar of G17 coal, Russ. J. Appl. Chem. 76(1) (2003) 148-152. [70] S.E. Moschopedis, Sulfomethylation of humic acids, lignites, and coals and products thereof, US Patent 3,352,902, 1967. [71] S.E. Moschopedis, J.G. Speight, Influence of metal salts on bitumen oxidation, Fuel 57(4) (1978) 235-240. [72] S.S. Fong, L. Seng, N.B. Majri, H.B. Mat, A comparative evaluation on the oxidative approaches for extraction of humic acids from low rank coal of Mukah, Sarawak, Journal of the Brazilian Chemical Society 18(1) (2007) 34-40. [73] V. Calemma, R. Rausa, Process for the production of regenerated humic acids from coal, U.S. Patent 4,788,360. 29 1988. [74] T. Sato, T. Adschiri, K. Arai, G.L. Rempel, F.T. Ng, Upgrading of asphalt with and without partial oxidation in supercritical water, Fuel 82(10) (2003) 1231-1239.

[75] M. Ashtari, New Pathways for Asphaltenes Upgrading via Oxy-cracking in Liquid Phase, University of Calgary, 2016. [76] W.P. Duyvesteyn, R.L. Morley, Oxidation of asphaltenes, U.S. Patent 8,529,687, 2013. [77] M. Ashtari, L. Carbognani, P. Pereira-Almao, Asphaltenes aqueous conversion to humic and fulvic analogs via oxy-cracking, Energy & Fuels 30(7) (2016) 5470-5482. [78] P. Haghighat, L. Carbognani, P. Pereira-Almao, Experimental study on catalytic hydroprocessing of solubilized asphaltene in water: A proof of concept to upgrade asphaltene in the aqueous phase, Energy & Fuels 30(4) (2016) 2904-2918. [79] L. Li, P. Chen, E.F. Gloyna, Generalized kinetic model for wet oxidation of organic compounds, AIChE Journal 37(11) (1991) 1687-1697. [80] R.C. Silva, J.R. Radović, F. Ahmed, U. Ehrmann, M. Brown, L. Carbognani Ortega, S. Larter, P. Pereira-Almao, T.B. Oldenburg, Characterization of acid-soluble oxidized asphaltenes by Fourier transform ion cyclotron resonance mass spectrometry: Insights on oxycracking processes and asphaltene structural features, Energy & Fuels 30(1) (2015) 171-179. [81] P. Haghighat, L.C. Ortega, P. Pereira-Almao, Experimental study on catalytic hydroprocessing of solubilized asphaltene in water: A proof of concept to upgrade asphaltene in the aqueous phase, Energy & Fuels 30(4) (2016) 2904-2918. [82] A. Lifshitz, J.V. Michael, Rate constants for the reaction, O+H2O→OH+OH, over the temperature range, 1500–2400 K, by the flash photolysis-shock tube technique: A further consideration of the back reaction, Symposium (International) on Combustion 23(1) (1991) 59-67. [83] S.K. Bhargava, J. Tardio, J. Prasad, K. Föger, D.B. Akolekar, S.C. Grocott, Wet oxidation and catalytic wet oxidation, Industrial & Engineering Chemistry Research 45(4) (2006) 1221- 1258. [84] J. Levec, A. Pintar, Catalytic wet-air oxidation processes: a review, Catalysis Today 124(3) (2007) 172-184. [85] J. Wang, W. Fu, X. He, S. Yang, W. Zhu, Catalytic wet air oxidation of phenol with functionalized carbon materials as catalysts: Reaction mechanism and pathway, Journal of Environmental Sciences 26(8) (2014) 1741-1749. [86] J.-C. Béziat, M. Besson, P. Gallezot, S. Durécu, Catalytic wet air oxidation on a Ru/TiO2 catalyst in a trickle-bed reactor, Industrial & engineering chemistry research 38(4) (1999) 1310- 1315. [87] X.C. Song, Y.F. Zheng, H.Y. Yin, Catalytic wet air oxidation of phenol over Co-doped Fe3O4 nanoparticles, Journal of nanoparticle research 15(8) (2013) 1856. [88] H. Wei, X. Yan, S. He, C. Sun, Catalytic wet air oxidation of pentachlorophenol over Ru/ZrO 2 and Ru/ZrSiO 2 catalysts, Catalysis today 201 (2013) 49-56. [89] A. Yabuki, S. Tanaka, Oxidation behavior of copper nanoparticles at low temperature, Materials Research Bulletin 46(12) (2011) 2323-2327. [90] A. Santos, P. Yustos, A. Quintanilla, F. Garcia-Ochoa, Kinetic model of wet oxidation of phenol at basic pH using a copper catalyst, Chemical engineering science 60(17) (2005) 4866- 4878. [91] S. Li, J. Gong, Strategies for improving the performance and stability of Ni-based catalysts for reforming reactions, Chemical Society Reviews 43(21) (2014) 7245-7256.

[92] N. Khanikar, K.G. Bhattacharyya, Cu (II)-kaolinite and Cu (II)-montmorillonite as catalysts for wet oxidative degradation of 2-chlorophenol, 4-chlorophenol and 2, 4-dichlorophenol, Chemical engineering journal 233 (2013) 88-97. [93] M. Parvas, M. Haghighi, S. Allahyari, Degradation of phenol via wet-air oxidation over CuO/CeO2–ZrO2 nanocatalyst synthesized employing ultrasound energy: physicochemical characterization and catalytic performance, Environmental technology 35(9) (2014) 1140-1149. [94] N. Habbache, N. Alane, S. Djerad, L. Tifouti, Leaching of copper oxide with different acid solutions, Chemical Engineering Journal 152(2) (2009) 503-508. [95] Y. Tu, Y. Xiong, S. Tian, L. Kong, C. Descorme, Catalytic wet air oxidation of 2-chlorophenol over sewage sludge-derived carbon-based catalysts, Journal of hazardous materials 276 (2014) 88- 96. [96] C.B. Maugans, A. Akgerman, Catalytic wet oxidation of phenol in a trickle bed reactor over a Pt/TiO 2 catalyst, Water research 37(2) (2003) 319-328. [97] Manasrah, A. D., Al-Mubaiyedh, U. A., Laui, T., Ben-Mansour, R., Al-Marri, M. J., Almanassra, I. W., & Atieh, M. A., Heat transfer enhancement of nanofluids using iron nanoparticles decorated carbon nanotubes. Applied Thermal Engineering, 107, (2016) 1008-1018. [98] Manasrah, A. D., Almanassra, I. W., Marei, N. N., Al-Mubaiyedh, U. A., Laoui, T., & Atieh, M. A., Surface modification of carbon nanotubes with copper oxide nanoparticles for heat transfer enhancement of nanofluids. RSC Advances, 8(4), (2018) 1791-1802. [99] S. Hamoudi, A. Sayari, K. Belkacemi, L. Bonneviot, F. Larachi, Catalytic wet oxidation of phenol over Pt x Ag 1− x MnO 2/CeO 2 catalysts, Catalysis today 62(4) (2000) 379-388. [100] L. Liotta, Catalytic oxidation of volatile organic compounds on supported noble metals, Applied Catalysis B: Environmental 100(3-4) (2010) 403-412. [101] Manasrah, A. D., Laoui, T., Zaidi, S. J., & Atieh, M. A., Effect of PEG functionalized carbon nanotubes on the enhancement of thermal and physical properties of nanofluids. Experimental Thermal and Fluid Science, 84, (2017) 231-241. [102] X. Wei, C. Tang, X. Wang, L. Zhou, Q. Wei, M. Yan, J. Sheng, P. Hu, B. Wang, L. Mai, Copper silicate hydrate hollow spheres constructed by nanotubes encapsulated in reduced graphene oxide as long-life lithium-ion battery anode, ACS applied materials & interfaces 7(48) (2015) 26572-26578. [103] A. De Andres, Naturally-occurring silicates as carriers for copper catalysts used in methanol conversion, Clays and Clay Minerals 40(2) (1992) 167-174. [104] R. Jin, Y. Yang, Y. Xing, L. Chen, S. Song, R. Jin, Facile synthesis and properties of hierarchical double-walled copper silicate hollow nanofibers assembled by nanotubes, ACS nano 8(4) (2014) 3664-3670.

Experimental and Theoretical Studies on Oxy-cracking of Quinolin-65 as a Model

Molecule for Residual Feedstocks

Graphical Abstract

Highlights ▪ Oxy-cracking process is proposed as a new technique for upgrading the residual feedstocks. ▪ Quinolin-65 has been selected as a model molecule mimicking the residual feedstocks. ▪ Oxy-cracking undergoes parallel consecutive reactions based on the experimental and theoretical studies. ▪ The Q-65 was oxy-cracked to aromatic intermediates with different families of organic acids and small amount of CO2.

This chapter is reproduced by permission of the Royal Society of Chemistry (RSC) from the following publication:

Manasrah, Abdallah D., Amjad El-Qanni, Ismail Badran, Lante Carbognani Ortega, M. Josefina Perez-Zurita, and Nashaat N. Nassar. "Experimental and theoretical studies on oxy-cracking of Quinolin-65 as a model molecule for residual feedstocks." Reaction Chemistry & Engineering 2, no. 5 (2017): 703-719. (DOI: 10.1039/C7RE00048K)

2.1 Abstract

Oxy-cracking is a combination of oxidation and cracking reactions for converting heavy hydrocarbons into commodity products with minimal emission of CO2. This reaction takes place in basic aqueous media, at mild operation temperatures (200−230 °C) and pressures (500-750 psi).

In this study, the main goal is to understand the oxy-cracking mechanism, involving oxidation and cracking reactions, of solid hydrocarbon represented by the model molecule Quinolin-65 (Q-65).

In the experimental part, the oxy-cracking reaction was performed in a Parr batch reactor operated at optimized oxygen partial pressure of 750 psi and temperatures between 200 °C and 230 °C. The reaction products were characterized by FTIR, TOC, GC, NMR, and XPS. We found that the main products are composed of organic carboxylic, phenolic, and carbonyl-containing compounds, with few amounts of inorganic carbons (IC). In the theoretical part of the study, a comprehensive computational modeling of Q-65 reactivity was performed using high level quantum theoretical calculations. The reactions studies indicated the attack of the hydroxyl radical (•OH) and hydroxide anion (–OH) on the Q-65 molecule. The theoretical study employed the density functional theory

(DFT) and the second-order Møller–Plesset perturbation theory (MP2) to study the reaction mechanisms under the same experimental conditions. Both, the theoretical and the experimental studies, confirmed the complexity of the reaction kinetics. The reaction kinetic results suggested that the Q-65 oxy-cracking reaction went through a parallel-consecutive reaction in which an oxidative decomposition took place in the first step producing different aromatic intermediates.

These intermediates were oxy-cracked consecutively into different organic acids and a small amount of CO2 gas.

2.2 Introduction

With the increased world population, modernization, and urbanization, the global demand for energy is projected to rise rapidly [1, 2]. Despite the variety of energy resources, the world still relies heavily on oil, which will inevitably deplete the world’s supply of conventional crude oil [3,

4]. Therefore, there is an increasing demand to upgrade and recover unconventional crude oil, like bitumen and heavy oil, to meet current and future global energy demands. Typically, processing of conventional and unconventional oils to meet market or pipeline specifications involves

“coking”, which is widely implemented in refineries and upgraders. This process generates a massive amount of solid-waste heavy hydrocarbon byproduct known as petroleum coke (petcoke)

[5]. Carbon is the most common species in these kinds of solid-wastes with many other metals and heteroatoms like sulpher, nitrogen, oxygen also present [6]. As a result of increased production of unconventional oils, asphaltenes and petcoke production will also increase in Canada and worldwide. Currently, there is no economically feasible use for these solid wastes and they are often disposed in stockpiles. In Canada, the stockpiles are growing at a rate of about 4 million tonnes a year [7, 8]. These stockpiles pose long-term problems, requiring huge land space, potentially leaching heavy metals, particulate matters, and other contents into the environment.

Thermal techniques have been proposed for treating these kinds of complex solid-waste heavy hydrocarbons and converting them into useful products. Gasification, a high-temperature process for converting carbonaceous solids into hydrogen and carbon monoxide gases, which subsequently could be used to produce liquid hydrocarbon through the Fischer-Tropsch process, is one common method [9-12]. As of today, the best-known gasification process for heavy residue fraction upgrading is the Flexicoking process developed by Exxon Mobil [13]. However, giving the

uncertainty of the optimum gasifier configuration, the lack of industrial scale gasifiers to confirm the feasibility and reliability of gasification process, the significant coke/ash content generated from the gasifier that still requires stockpiling or landfilling, and the production of low-BTU waste gas gasification is neither a green nor a reliable method for solid-waste conversion [10, 14, 15].

Combustion is another potential technique for energy production. Again, this technique has limited prospects due to sulfur and CO2 emissions [9, 16, 17].

Many researchers have investigated heavy hydrocarbons treatment processes, especially asphaltene-type waste, such as carbon rejection and hydrogen addition [17, 18]. Carbon rejection processes include pyrolysis (visbreaking, coking and rapid thermal pyrolyzer) [17-19], separation or extraction (solvent deasphalting process, SDA) [20, 21] and cracking (VGO and VR FCC) [18,

22]. Hydrocracking and steam cracking gasification are conventional processes of hydrogen addition [9, 23, 24]. These conventional hydroprocessing technologies still suffer from coke formation, high catalyst loading, and catalyst poisoning caused by metals deposition [23, 24].

Carbon rejection processes have simple configurations compared to hydrogen addition processes and do not require hydrogen injection and catalyst loading [18]. Nevertheless, carbon rejection produces lower yields and generates significant coke waste. Therefore, these processes were deemed to be unreliable technologies for the treatment of solid-waste heavy hydrocarbons, specifically due to their high-energy intensity and CO2 emissions [10, 14].

Advanced oxidation processes (AOPs), based on the creation of reactive species like hydroxyl radical (•OH), are reliable and capable of breaking down a wide range of complex heavy hydrocarbon compounds quickly and non-selectively [25-27]. One issue is the significant amount of gaseous CO2 produced, which requires further processing. Additionally, the solubilized organic

species required further separation and treatment before safe disposal. Therefore, AOPs cannot be the sole method for conversion of heavy waste hydrocarbon.

A number of researchers investigated the conversion of solid-waste heavy hydrocarbons (lignites and coal) to humic substances in acidic media using oxidizing agents such as KMNO4 and HNO3

[28, 29]. However, the process requires long reaction times for coal particles larger than 100 μm.

It was also found that the used acid was more expensive compared to the obtained products.

Moreover, the production of humic substances via coal oxidation was investigated industrially using a pre-oxidizing reaction [28-30]. Using this mechanism, coal is converted to humic acids, then treated with formaldehyde and alkali solution of bisulfate at 100 °C to form water-soluble products [31]. Although coal pre-oxidation was proposed in several patents, this process is expensive and also has low conversion [32-34].

Despite all the techniques that have been proposed for developing efficient and environmentally- friendly technologies for treating solid-waste heavy hydrocarbons, there are no industrially feasible options. The challenges that face the oil industry create a need for new environmentally- friendly and costly effective technologies to convert solid-waste heavy hydrocarbon into valuable products. To overcome this challenge, our research group has recently introduced an alternative technique to produce light commodity hydrocarbons from asphaltenes. This method involves a combination of two types of reactions in aqueous alkaline media, namely oxidation and cracking, creating the name “oxy-cracking” [35]. The oxy-cracking technique is inspired by asphaltenes oxidation and ozonolysis studies reported in previous works [31, 36-39]. This new oxy-cracking technique offers a pathway that creates valuable material from heavy oil and refinery waste.

Oxidized asphaltenes become soluble in water due to the polar functionalization of aromatic edges

and paraffinic terminal carbons via oxygen incorporation. Changing the oxidation conditions will alter the mechanism and result in different product distributions. It is well known that the hydroxyl radical (•OH) plays an essential role in the hydrocarbon wet oxidation process [40-42]. Such process is believed to involve the following primary initiation steps:

• • H2O + O2 → HO2 + OH Ea = 68.6 kcal/mol [43]

• • HO2 + H2O → H2O2 + OH Ea = 32.8 kcal/mol [43]

• • • • H2O + O → OH + OH Ea = 17 kcal/mol [44]

These initiation steps are followed by different propagation and termination pathways. As reported by many studies [25, 42, 43, 45], this indicated that a large range of intermediates and oxidized products may result from the above steps. The prevailing products in the liquid phase are alkylhydroperoxides, which are quickly converted into hydroxylated hydrocarbons and ketones/aldehydes [43]. Hence, via the oxy-cracking process, waste heavy hydrocarbons could be oxidized in an aqueous alkaline medium at mild temperatures (170-230 °C) and pressures (500-

750 psi) [35, 46, 47]. However, these heavy hydrocarbons do not oxidize completely to CO2, but instead form intermediate compounds that are soluble in water via oxygen incorporation. Thus, oxy-cracking enhances the tendency of large heavy hydrocarbon molecules to disaggregate, making them more accessible to subsequent hydrogenation and cracking at lower temperatures.

Those intermediates (desired products) are likely analogs of humic acids, carboxylic acids, and their corresponding salts, or other products [35, 46-48]. Because of the selectivity and high solubility of such complex hydrocarbons in an alkaline medium, the process is considered environmentally-friendly, as there is very little CO2 emitted.

Typically, these solid-waste heavy hydrocarbons, such as petcoke and asphaltenes, have no unique chemical identity and portray complex chemical structures that contain heteroatoms such as sulfur, nitrogen, and oxygen. However, some researchers have attempted to hypothesize model structures for asphaltenes, such as the Yen model or modified Yen model [49-57], based on physical and chemical methods. Understanding the reaction mechanism and the kinetics of such an innovative technique requires the chemical structure of the solid-waste heavy hydrocarbons to be known.

Finding a model molecule with a well-known structure that mimics the real structure and properties of solid-waste heavy hydrocarbons would clarify the oxy-cracking reaction mechanism. Several researchers propose Quinolin-65 (Q-65) as a model molecule for solid-waste heavy hydrocarbons

(like asphaltenes) because it has a large aromatic region in addition to side chains [58-60].

Furthermore, Q-65 has been confirmed to resemble asphaltene molecules regarding physical and chemical properties, thermal cracking behavior, and adsorption on solid surfaces [61, 62].

Recently, we studied the complete dry oxidation of Q-65 as a heavy-oil model molecule [63]. The research involved a comprehensive theoretical and thermogravimetric analysis of Q-65 in the absence of a catalyst. The study concluded that the thermo-oxidative decomposition of Q-65 is a complicated multi-step process involving different reaction pathways. The kinetic parameters from both the experimental and theoretical parts of the study strongly suggested catalyst incorporation.

Herein, we are interested in understanding the oxy-cracking of residual feedstocks using Q-65 as a model molecule. Therefore, the purposes of this study include: (1) carrying out an oxy-cracking reaction for Q-65 within a batch reactor under aqueous alkaline medium and mild operating conditions (i.e., temperatures varied between 200-230 °C and the operating pressure fixed at 750 psi) to keep the water in the liquid phase; (2) determining the experimental reaction kinetics

parameters and potential mechanism for the oxy-cracking of Q-65 in liquid phase using the generalized “triangular” lumped kinetic model for non-catalytic wet oxidation of organic compounds; and (3) carrying out computational modelling to understand the oxy-cracking mechanism using state-of-the-art ab-initio methods. The density functional theory (DFT) and the second-order Møller–Plesset perturbation theory (MP2) were implemented to study the oxy- cracking reaction mechanism under similar conditions to those in the experimental part. The theoretical study will not just validate the experimental findings, but also it serves twofold; first, it presents a new understanding for the oxy-cracking of Q-65 at the molecular level. Second, it provides vital kinetic parameters such as activation energies and Gibbs free-energies of reaction for different possible oxidation pathways.

2.3 Experimental Work

2.3.1 Materials

Quinolin-65 (Q-65) in powder form (C30H27NO2S, λmax = 565 nm, 80 wt% dye content) was purchased from Sigma-Aldrich (Ontario, Canada) and selected as a model compound mimicking the solid waste hydrocarbons. The chemical structure of Q-65 is shown in scheme 2.1 [63]. The

Q-65 was used as received without further purification as the impact of impurities on our results, from the application point of view, is minimal as real solid waste hydrocarbons, when oxy-cracked, will contain unknown chemical structures.

Potassium hydroxide (KOH, ACS reagent, ≥85%, pellets form) purchased from Sigma-Aldrich

(Ontario, Canada) was used to adjust the pH of reaction medium (deionized water) and help in solubilizing Q-65 in solution. Oxygen (99.9% ultrahigh purity) purchased from Praxair (Calgary,

Canada) was used as the oxidant gas.

Scheme 2.1 Q-65 molecular structure [63].

2.3.2 Experimental setup and procedure

The oxy-cracking experiments were carried out in the setup shown schematically in Figure 2.1. A

100 ml reactor vessel was used for the experiments (model number 4598, Parr Instrumental

Company, Moline, Il, USA). The reactor vessel was equipped with a heating oven connected to a temperature control loop, a pressure gauge and a mechanical stirrer connected to a speed controller.

The reactor is capable of handling pressures up to 1700 psi and temperatures up to 270 °C. In a typical experiment, 50 mg of Q-65 was charged into the reactor vessel containing 30 g of deionized water and a certain amount of KOH to keep the pH above 8.0 to help in solubilizing Q-65 and to avoid corrosion problems. The reaction was conducted at a stirring rate of 1000 rpm to minimize the interfacial mass resistance between the gas and liquid phases and to ensure uniform temperature and concentration profiles in the liquid phase. Leak tests were performed by pressurizing the reactor with O2 up to 1000 psi prior to fixing the operating pressure at 750 psi. The reactor was then heated to the desired temperature. Zero reaction time was set when the desired temperature 35

was attained. The reaction was carried out at different times, namely 30, 60, 90 and 120 min. The predetermined operating temperatures were chosen to be 200, 215, and 230 ºC. This temperature range was considered as an optimum range for favoring the Q-65 conversion with a minimum amount of CO2 production. At the end of each experiment, the reactor was cooled down to room temperature. Then, the gas was analyzed using gas chromatography, GC (SRI 8610C, SRI

Instruments), having thermal conductivity detection (TCD). The liquid layer containing the oxy- cracked Q-65 was carefully discharged from the reactor vessel for total organic carbon (TOC) analysis, and the reactor vessel and assembly were washed with soap and deionized water to remove any residue, therefore, the apparatus could be ready for another cycle of experiments. The oxy-cracked Q-65 was recovered by using an evaporator for further analysis by FTIR, NMR spectroscopy (see Appendix A1, S1.1), and XPS (see Appendix A1, S1.2) techniques to investigate the nature of the formed compounds. It is worth noting here that a minimum amount of solid residue (i.e., insolubilized Q-65) could only be collected at low reaction conversion, which was recovered by using an evaporator for the elemental analysis. The elemental analysis was performed in a combustion method using a PerkinElmer 2400 CHN (Waltham, Massachusetts, USA) analyzer to determine the amount of carbon, hydrogen, and nitrogen after the oxy-cracking reaction. All experiments were duplicated to confirm reproducibility.

Figure 2.1 Schematic illustration of the experimental setup (not to scale).

2.3.3 Characterization

2.3.3.1 Fourier Transformed Infrared (FTIR) spectroscopy

The oxy-cracked Q-65 in the aqueous layer was dried at 65 oC overnight in a vacuum oven to evaporate the water and recover the formed organic species. The solid was collected and characterized using a FTIR Bruker Tensor 27 instrument manufactured by Thermo Electron

Corporation with a smart diffuse reflectance attachment to carry out diffuse reflectance infrared

Fourier transform spectroscopy (DRIFTS) analysis. Initially, about 500 mg of potassium bromide

(KBr) powder was analyzed to define the background; then, approximately 5 mg of the oxy- cracked Q-65 sample dispersed in the 500 mg of KBr was analyzed. Virgin Q-65 was also analyzed

for comparison. The obtained spectra were averaged from 128 scans acquired with a resolution of

2 cm-1, ranging from 400 to 4000 cm-1.

2.3.3.2 Total organic carbon (TOC)

The aqueous sample containing the oxy-cracked Q-65 was centrifuged (5000 rpm and 10 min) to separate the remaining solid (unreacted and insolubilized species), followed by filtration (syringe filters with 0.5 µm pore size) to separate the suspended particles remaining in the water, in case there is any. The filtered samples were analyzed using a Shimadzu Total Organic Carbon Analyzer

(TOC-L CPH/CPN) to measure the total carbon (TC), total organic carbon (TOC), and inorganic carbon (IC) in the aqueous phase. Both TC and IC measurements were calibrated using standard solutions of potassium hydrogen phthalate and sodium hydrogen carbonate. Fifteen milliliters of the filtered samples were placed in standard TOC vials. Using the TOC software to control the system, the TC was automatically measured; after that, an acid was added to evolve CO2 from the sample to measure the remaining organic compounds, which were considered as TOC. All the measurements were taken for three samples from each experiment and the standard deviation was calculated and presented.

2.3.3.3 Gas chromatography (GC)

After the oxy-cracking reaction was completed and cooled down to room temperature, the compositional analysis of the produced gas was carried out with a GC (SRI 8610C Multiple Gas

#3 gas chromatograph SRI Instruments, Torrance, CA), provided with TCD detector and two packed columns connected in parallel (3′ molecular sieve/6′ Hayesep-D columns). GCs temperature holds on 42 °C for 10 min, then increases to 200 °C with a rate of 20 °C/min.

2.3.4 Theoretical calculations

All theoretical calculations in this work were performed using the Gaussian 09 program [64]. For the reactions involving the hydroxyl radical (•OH), geometry optimization and frequency calculations were performed using the hybrid density functional, B3LYP [65, 66] and the 6-

31+G(d) [67] basis set. For the reactions involving the hydroxide anion (–OH), the calculations were performed using the B97D3 functional with Grimme’s dispersion [68]. In addition, the –OH reactions were carried out in water by assigning (solvent water) in the scrf keyword in Gaussian.

Zero-point energies (ZPE) were scaled by a factor of 0.9748, as suggested by Scott and Radom

[69]. In order to obtain accurate energy values, single point energies for all species, including transition states, were requested after complete optimization at MP2/6-311+G(d,p) level of theory.

Exact details on locating the transition states and the corresponding intrinsic reaction coordinates

(IRC’s) are described in detail in previous works [63, 70, 71]. All transition state located in this work were confirmed to have one and only one imaginary frequency along the desired reaction coordinate, by plotting the corresponding IRC’s in both directions.

The activation and reaction parameters (enthalpy, entropy, and Gibbs free energy) were calculated as described in previous works [63, 70, 71]. The rate constants for the unimolecular and bimolecular reactions were obtained using the two following equations, respectively [72]:

 kBT  G‡ / RT kuni   e (2.1)  h 

‡ 2  kBT  S / R Ea  / RT ‡ kbimol  e   e e , Eabimol  H  2RT (2.2)  h 

where kB is Boltzmann constant, R is the universal gas constant, T is the temperature, h is Planck’s constant, ΔH‡, ΔS‡ and ΔG‡ are the enthalpy, entropy and Gibbs free energy of activation, respectively.

2.4 Results and Discussion

Experimental determination of the oxy-cracking reaction kinetics

The effects of temperature and residence time on Q-65 oxy-cracking were investigated to understand and validate the reaction kinetic mechanism. Other important parameters, such as the operating partial pressure, the stoichiometric ratio of Q-65 with water and KOH, and the impeller speed remained unchanged. Oxy-cracking is a gas-liquid reaction which includes various transport processes that might take place in series [35]. For this reason, preliminary experiments were conducted to optimize the operating oxygen partial pressure and the impeller speed. The reaction was not significantly affected by pressures above 750 psi, which was thus used as the optimum operating partial pressure. Therefore, we ensured that oxygen in an excess amount is present. The rate of the oxy-cracking was found to be independent of the impeller speed between 500 and 1500 rpm, indicating the absence of interfacial mass transfer resistance in the liquid phase. This agrees excellently with other oxy-cracking and wet oxidation studies [26, 35, 46, 47, 73, 74]. Hence, all experiments were carried out at the impeller speed of 1000 rpm.

After oxy-cracking of the Q-65, several soluble and insoluble intermediates were produced. Thus, the produced intermediates were considered as a lump component to describe the oxy-cracking reaction. Hence, the rate equations are developed based on the lumped concentrations of total organic carbon (TOC). A generalized “triangular” lumped kinetic model for non-catalytic wet oxidation of organic compounds in wastewater has been proposed by Li et al. [43]. It relies on the

assumption that some of the organic compounds present in the wastewater are directly oxidized to

CO2 and H2O, while the rest are converted to an intermediate product which is further oxidized.

Recently, Ashtari et al. [35] employed the “triangular” reaction pathway for asphaltenes oxy- cracking, where oxygen was used to convert asphaltenes to water-soluble material which contains carboxylic, sulfonic, and phenolic functionalities in addition to a minimum amount of CO2 gas.

The triangular reaction pathway was applied to the oxy-cracking of Q-65 while keeping the decomposition of alkylperoxo radicals under control to avoid the significant release of CO2 as shown in Figure 2.2.

Figure 2.2 Triangular reaction scheme of Q-65 oxy-cracking. 푨 is the Q-65, 푩 is the soluble intermediate (TOC), 푪 is CO2, and 푫 is the inorganic carbon (IC).

The mass balance was closed by conducting carbon-based balance before and after the oxy- cracking reaction of Q-65 as follows;

퐶푎푟푏표푛 푐표푛푡푒푛푡 𝑖푛 푡ℎ푒 표푟𝑖𝑔𝑖푛푎푙 푄 − 65 (퐶퐴표 ) =

퐶푎푟푏표푛 푐표푛푡푒푛푡 𝑖푛 푡ℎ푒 푙𝑖푞푢𝑖푑 푝ℎ푎푠푒 (퐶퐿 ) +

+ 퐶푎푟푏표푛 푐표푛푡푒푛푡 𝑖푛 푡ℎ푒 𝑔푎푠 푝ℎ푎푠푒 (퐶퐶) +

+ 퐶푎푟푏표푛 푐표푛푡푒푛푡 𝑖푛 푡ℎ푒 푠표푙𝑖푑 푟푒푠𝑖푑푢푒(퐶푅) (2.3)

퐶퐴표 was calculated by multiplying the mass of Q-65 by the fraction of carbon in the Q-65 and 퐶퐿was considered as TC. The overall carbon mass balance was closed to 98.21%, the leftover 1.79% was considered as a loss due to the experimental and instrumental errors. The latter percentage was added and presented alongside the TOC percentage error, which was less than 5% for all runs. The produced gas, most likely CO2, was analyzed online using GC. However, other gas contents were minimal, and thus, they were neglected. The ideal gas law was used to estimate the carbon content in gas phase (퐶퐶) using the following equation;

푃푉 퐶 = × 12 (2.4) 퐶 푅푇

where 푃 and 푇 are the pressure and temperature at the end of the reaction, respectively, 푉 is the

volume of the gas phase in the reactor vessel and 푅 is the ideal gas constant.

For the solid residue, the carbon content was estimated using elemental analysis; carbon content

after reaction (insolubilized Q-65) = (mass of solid residue) × (carbon wt%). Moreover, the

reaction conversion (푋) based on carbon mass was calculated as follows;

퐶 − 퐶 푋 = 퐴0 푅 (2.5) 퐶퐴0

퐶푅 is the carbon concentration in the solid residue. It is worth noting that the amount of Q-65 (퐶퐴)

for each residence time was calculated as follows;

퐶퐴 = 퐶퐴표(1 − 푋) (2.6) 42

The kinetics rate equations are presented by the set of the following three differential equations:

푑퐶 퐴 = −푟 = (푘 + 푘 )퐶2 (2.7) 푑푡 퐴 1 2 퐴

푑퐶 퐵 = +푟 = 푘 퐶2 − 푘 퐶 (2.8) 푑푡 퐵 2 퐴 3 퐵

푑퐶 퐶 = +푟 = 푘 퐶2 + 푘 퐶 (2.9) 푑푡 퐶 1 퐴 3 퐵 where 푡 is the reaction time, and 푘1, 푘2, and 푘3 are the rate constants for each reaction step (Figure

2.2). Mathematica (V10.2) was used to estimate the rate constant values (i.e., 푘1, 푘2, and 푘3) by fitting the kinetic results obtained experimentally to equations (2.7-2.9) and solving them simultaneously under initial conditions of 푡 = 0, where 퐶퐴 = 퐶퐴표, and 퐶퐵 = 퐶퐶 = 0. The proportional weighted sum-of-squares was minimized using Mathematica until all values of the correlation coefficient (R2) were very close to 1.0. Consequently, the values of activation energy were estimated using the Arrhenius equation as follows:

−퐸푎푖⁄푅푇 푘푖 = 퐴푖푒 (2.10) where 퐴 is the frequency factor, 퐸푎 is the activation energy, 𝑖 is the step of the reaction pathway

(1, 2, and 3), 푅 is the ideal gas constant, and 푇 is the temperature. Table 2.1 lists the estimated values of rate constants and activation energies.

Figure 2.3 shows the experimental concentration profiles of Q-65, intermediate compounds

(desired products), and CO2 expressed as a function of time during the oxy-cracking reaction at

200, 215, and 230 ºC, together with the kinetic model fit. The error bars shown in the figure represent the calculated standard deviation based on the TOC and GC measurements. The second- order kinetic model fit well to the experimental results agreeing with a similar study on catalytic

wet air oxidation of industrial effluents using the second-order lumped kinetic model [75, 76].

Moreover, this model described our reaction kinetics more accurately than the first-order kinetic model [35, 43].

Clearly, the reaction temperature is a key parameter in the oxy-cracking reaction of Q-65. By increasing the temperature, organic functionalities are oxy-cracked in water due to the partial oxy- cracking of Q-65 into oxygenated intermediates, however, some of these intermediates are decomposed oxidatively into CO2 and H2O. Interestingly, the highest temperature of 230 ºC appears to be the optimum temperature as the highest conversion, synchronized with low CO2 production occurred at that temperature (Figure 2.3a). This agrees well with similar studies reported for C7 asphaltenes oxy-cracking [35, 46] and C7 asphaltenes catalytic hydroprocessing

[73].

Figure 2.4 shows the Arrhenius plot for the Q-65 oxy-cracking reaction for each step of the proposed triangular reaction mechanism. As seen, Arrhenius behavior was observed in the temperature range 200-230 ºC. The values of activation energies and frequency factors are summarized in Table 2.1.

In our previous work regarding Q-65 thermo-oxidative decomposition, we proposed that the reaction begins with paraffinic chain loss, which was supported by both experimental and theoretical evidence [63]. In this study, a similar behavior was observed. Initially during the oxy- cracking reaction, the alkyl chain presumably breaks up, followed by the dissociation of the weak carbons and sulfur bonds. Afterwards, complete cracking of the molecule takes place to produce analogs of humic substances and a small fraction of CO2, as reported in other works [77-80]. A detailed insight into the reaction mechanism will be presented later in Section 2.5.3.

-1 The first reaction pathway (퐸푎1 = 9.8 kcal.mol ) involves a complete or deep oxidation reaction and might be favored to produce CO2 or demonstrates that Q-65 is partially soluble initially in the

-1 alkaline aqueous medium. In the second reaction pathway (퐸푎2 = 17.7 kcal.mol ), additional reactions are likely to occur due to the increased availability of aromatic moieties, forming heavier soluble and insoluble intermediates. This is also indicated by the higher value of the frequency

4 -1 -1 factor (9.3 × 10 L. mmol . s ). However, CO2 could be produced in the third reaction pathway by total oxidation of some solubilized intermediates. Remarkably, the activation energies of the

-1 first and third reaction pathways are similar (퐸푎3 = 10.9 kcal.mol ). A similar trend was reported by Ashtari et al. [35] for the oxy-cracking of C7 asphaltenes. However, the value of 퐸푎2 obtained in the previous work [35] was around 28.3 kcal.mol-1. The difference in activation energies could be attributed to the structural complexity of asphaltene molecules and aggregation. Hence,

asphaltene aggregates require more oxygen penetration during the oxy-cracking reaction to achieve the desired conversion and selectivity. However, Q-65 is more solubilized in water in comparison with asphaltene molecules.

Figure 2.4 Arrhenius plots for the Q-65 oxy-cracking for each reaction pathway.

Table 2.1 Activation energies and frequency factors for Q-65 oxy-cracking.

Activation energy (kcal.mol-1) Frequency factor (Unit)

-1 -1 푬풂ퟏ 9.8 ±0.225 15.6 (L.mmol .s )

4 -1 -1 푬풂ퟐ 17.7 ±0.945 9.3 × 10 (L.mmol .s )

-1 푬풂ퟑ 10.9 ±0.123 7.5 (s )

FTIR spectroscopy of virgin and oxy-cracked Q-65

Figure 2.5 shows the infrared spectra of the virgin and oxy-cracked Q-65 at 230 °C over 2 h reaction period. The presence of oxygen in the aqueous phase causes the Q-65 molecules to oxy- crack into organic species bearing carboxylic and phenolic functional groups. A suggested mechanism of such reaction will be further discussed (Section 2.5.3). It is worth mentioning here that KOH aids the solubilization of Q-65 by neutralizing the acidic species forming water soluble hydrocarbon salts. Many studies confirmed that the solubility of species containing oxygen increases in the basic media [81, 82]. Patil et al. [81] demonstrated that the solubility of lignin in water increased in the presence of NaOH [81]. Likewise, n-C7 asphaltene oxy-cracking showed that the alkaline medium enhanced the cracking and subsequent solubilization of n-C7 asphaltenes in water [35].

As shown in Figure 2.5, the featured bands of virgin Q-65 can be categorized into aliphatic and aromatic. For the aromatic bands, the C–H bending in 1,2-, 1,3-, and 1,4-disubstituted aromatics are detected at 750, 808, and 863 cm−1, respectively. These bands are associated with the out-of- plane bending C=C stretching vibration detected at 1602 cm−1 in the aromatic region. Furthermore, in the aliphatic region, the characteristic C–H stretching vibrations of alkyl groups were assigned at 2922 and 2852 cm−1. The bands at 1458 and 1375 cm−1 represent the C–H vibration of asymmetric and symmetric bending, respectively. The band corresponding to the C=O group in

Q-65 represented at a frequency of 1715 cm-1. No additional bands which may imply the presence of impurities of a very different nature of Q-65 where observed.

The FTIR spectrum of the oxy-cracked Q-65 was dramatically different than virgin Q-65 (Figure

2.5). The O-H band spanning between 3500-2500 cm-1 corresponds to -OH groups. This can

originate from the presence of alcohols and/or H-bond phenol groups. The presence of a carbonyl band at 1733 cm−1 was assigned to carboxylic groups. However, its contribution is negligible, indicating that these species are not important contributors to the spectrum, i.e., they are in salt forms. The most important doublet band centered at 1400 cm−1 was assigned to carboxylate anion stretching, suggesting that most carboxylic compounds are present as salts. The FTIR spectrum of oxy-cracked Q-65 shows the absence of the aromatic C=C ring stretching band (1600 cm-1) and out-of-plane bands (950−750 cm−1) as compared to the virgin Q-65, indicating the disappearance of most aromatic moieties. Also, the alkyl groups are hardly visible in the range 3030−2800 cm−1.

The IR band at 1137 cm−1 could be representative of sulfones (O=S=O) and those appearing at

~700 and 850 cm-1 might correspond to sulphonic acid bonds. Overall, a similar water-asphaltene solubilization mechanism reported recently [35, 46] is also applicable to this analysis. It has been reported by Calemma et al. [29] that the absorption bands between 1600 and 1800 cm−1 could belong to carboxyl groups, carbonyl groups, esters, aromatic esters, and ketones. This study showed that increasing the oxidation conditions activated the aromatic rings with hydroxyl groups.

As a result, phenol groups are converted to quinones and subsequently transformed to carbonyl groups through the rupture of benzene rings [29]. The above FTIR analysis for the virgin and oxy- cracked Q-65, demonstrates that phenolic and carboxylic functional groups are formed at 230 °C.

More experimental evidence for the formation of such groups can be seen in the Appendix A1.

104

863 2852 808 102 2922

100 1715 1375 770

1602 98

96 1733

94 1137

92 % Transmittance % 90 850 3200 700 88 Virgin Q-65 Oxy cracked Q-65 86 1400

4000 3500 3000 2500 2000 1500 1000 500 Wavenumber (cm-1)

Figure 2.5 FTIR spectra of virgin and oxy-cracked Q-65 at 230 °C and 2 h residence time.

Theoretical modeling

2.4.3.1 Reactions of Q-65 with •OH radical

As mentioned in the introduction section, the •OH radical plays an important role in the wet oxidation of the heavy waste hydrocarbon. Therefore, different reaction pathways for the model molecule, Q-65 with •OH radicals were explored. In our previous work [63], the aliphatic chain in

Q-65 was shortened to two carbon atoms, creating the new molecule “Q-65b”. Simplifying the structure in this manner has significantly sped up the calculations without affecting the accuracy

of the results. Scheme 2.2 shows the first attack of the •OH radical on Q-65b. In this reaction pathway, direct abstraction of aromatic hydrogen led to the formation of water and the radical intermediate I. This intermediate could be easily stabilized by another OH phenol I. The primary step for this pathway proceeded via the transition state TS1, which has a strong imaginary frequency of 1465.5i cm–1 corresponding to HO-H abstraction. An optimized structure of TS1 is shown in Figure 2.6. The energy level diagram of the hydrogen abstraction pathway is shown in

Figure 2.7. The diagram is constructed using relative Gibbs free energies calculated at MP2/6-

311+G(d,p)//B3LYP/6-31+G(d) level of theory. Notably, the activation barrier for this pathway is high (134.4 kcal/mol). Also, the first step is endogenic by 114.5 kcal/mol. This suggests that direct abstraction of H by •OH in Q-65b is highly unfavourable, in close agreement to previous studies on benzene [83, 84] whereas the abstraction channel is thought to be minor at low temperatures.

• Scheme 2.2 The reaction mechanism initiated by OH attack on Q-65b to form H2O and phenol 1.

150 134.4 114.5 TS1 Intermediate 1 + H O 100 2

0.0 0.0 0

Q65b + OH Intermediate 1 + OH

-210

-220 -224.2

Relative Gibbs Energies (kcal/mol) Energies Gibbs Relative -230 Phenol 1

-240

Figure 2.7 Energy level diagrams for the water formation mechanism initiated by the •OH attack on Q-65b. Energy values represent the relative Gibbs free energies at 298 K (ZPE corrections included). In contrast to the abstraction pathway, the direct attack of •OH on the aromatic carbon in Q-65b is more favourable. As depicted in Scheme 2.3, this reaction pathway proceeds first through the transition state TS2, leading to the formation of the radical intermediate II, followed by a hydrogen loss via TS3 to form phenol I. The optimized structure for both transitions states are shown in

Figure 2.8. The energy level diagram for this reaction pathway is plotted in Figure 2.9. In addition to the slightly lower barrier than the first pathway (128.5 kcal/mol), intermediate II is more stable than intermediate I, i.e., II lies 92.3 kcal/mol above Q-65b, whereas I lies 114.5 kcal/mol above

Q-65b, as mentioned earlier.

It is worth noting that similar attack of •OH on aromatic carbon in small aromatic compounds, such as toluene [85] and ethyl benzene [86], proceeds with very low barriers, and have exothermic reaction enthalpies. Suh et al. [85] calculated the reaction energies in the case of toluene to be - 54

17.5 kcal/mol at B3LYP/6-31G(d,p) level. The activation barrier was even negative at -3.2 kcal/mol [85]. The relatively high barrier and the large endothermic reaction energy of •OH attack on Q-65b can be explained by the high stability of the large aromatic system and the breakage of aromaticity caused by •OH attachment to the aromatic molecule.

Scheme 2.3 The formation mechanism of Phenol I, initiated by •OH attack on Q-65b.

a) TS2 b) TS3

Figure 2.8 Optimized structure for the transition states involved in the phenol I formation, a) TS2 and b) TS3. For the colour scheme, refer to Figure 2.6.

140 128.5 131.2 TS3 120 TS2

100 92.3 Intermediate II 80

20 0.0 -3.6 RelativeGibbs Energies (kcal/mol) 0 Phenol 1 + H Q65b + OH -20

Figure 2.9 Energy level diagrams for the phenol 1 formation mechanism initiated by •OH attack on Q-65b. Energy values represent the relative Gibbs free energies at 298 K (ZPE corrections included). The next reaction pathway involves the loss of the olefin chain in Q-65b, such loss can take place by unimolecular 1,3-H shift mechanism, as explained in our previous study [63]. Alternatively,

the loss can occur by a primary attack of •OH on the carbon adjacent to nitrogen in Q-65b, where the lowest unoccupied molecular orbital (LUMO) is centred. (Details on the HOMO and LUMO of the Q-65b molecule were presented in our recent work [63]). This attack forms intermediate III through the transition state TS3. III then can decompose into phenol II and an ethoxy radical, as depicted in Scheme 2.4. The optimized structures of the two transition states involved in this mechanism are shown in Figure 2.10. The proposed phenols suggested by this theoretical study

(such as I and II) are in agreement with our FTIR observations, as shown earlier in Section 2.4.2.

The attack of •OH on the LUMO of Q-65b has an energy barrier and Gibbs free reaction energy close to those of the previous two attacks discussed earlier, as seen in Figure 2.11. However,

Phenol II is more stable in relative to Phenol I, as clearly demonstrated from the two energy diagrams, Figure 2.9 and Figure 2.11.

Scheme 2.4 The formation mechanism of Phenol II, initiated by •OH attack on the LUMO of Q- 65b. a) TS4 b) TS5

Figure 2.10 Optimized structure for the transition states involved in the formation of phenol 2, a) TS4 and b) TS5. For the colour scheme, refer to Figure 2.6.

140 131.6 129.9 TS5 120 TS4

100 93.7 Intermediate III 80

20 0.0

RelativeGibbs Energies (kcal/mol) 0 -9.8 Q65b + OH Phenol 2 + OEth. -20

The next reaction pathway of interest is the attack of •OH on the carbon adjacent to the carbonyl group in Q-65b, as illustrated in Scheme 2.5. In this mechanism, we chose to start from the product of the previous mechanism, Phenol II, instead of the parent Q-65b module. This is based on the assumption that the course and the energies of the reactions would not be affected by this change.

We have successfully located the transition state TS6, leading to the formation of the radical intermediate IV, which undergoes a quick hydrogen shift to form V through the transition state

TS7, as shown in Scheme 2.5. Finally, intermediate V can be easily stabilized by an H radical to form Phenol III, which also contains an aldehyde group.

Other possible oxidations of Phenol III to form a carboxylic acid have been explored. As seen from Scheme 2.5, this can take place by the attack of an •OH radical to Phenol III, forming the intermediate VI, through a transition state TS8. Hydrogen loss converts intermediate VI into the carboxylic acid VII. The optimized structures of the transition states participating in this mechanism, along with acid VII are shown in Figure 2.12. The formation of the aldehyde group

(Phenol III) and the carboxylic acid shown in Scheme 2.5 are in agreement with the experimental results obtained in Section 2.4.

Figure 2.13 represents the energy level diagram for the reaction mechanism initiated by •OH radical attack on Phenol III. The first activation barrier in this mechanism is high (127.8 kcal/mol).

This suggests that this reaction pathway is not kinetically favourable, similarly to the previous two mechanisms. However, once this barrier is reached, the next barrier from the intermediate IV to V is relatively small (41.9 kcal/mol). In addition, the stabilization of V into Phenol III is very exothermic (-195.5 kcal/mol). Despite the fact that this reaction pathway is kinetically unfavourable, it is thermodynamically preferred, due to the huge Gibbs free energy of reaction 60

released by the formation of Phenol III. This is similar to the previous case where Phenol I is formed, as shown in Figure 2.7.

The energies associated with the formation of the carboxylic acid VII are plotted in Figure 2.14.

Similar to the previous reaction pathways, the primary step for this reaction is also unfavourable

‡ due to the relatively high Gibbs free energy of activation (∆퐺298) (127.8 kcal/mol). Since the whole reaction is exergonic by -17.5 kcal/mol, this reaction pathway is thermodynamically favourable.

Thus, different reaction pathways of the Q-65b molecule initiated by the •OH radical have been demonstrated. Although the role of the heteroatoms, S and N, in the reaction mechanism is not initially obvious, they help to create a location prone to free radical attack. It is known that direct attacks of free radicals on heteroatoms are not favourable [83, 87]. Thus, oxidation of heterocyclic compounds is more likely to occur from free radical attacks on carbon atoms adjacent to these heteroatoms. This prediction has been confirmed by performing calculations on the Q-65b molecule. It was discovered that attacks of •OH radicals on the heteroatoms, N and S, were highly energetic.

Scheme 2.5 The decomposition mechanism of Q-65b, initiated by the attack of •OH, and leading to the formation of Acid VII.

a) TS6 b) TS7

c) TS8 d) Acid VII

141.6 150 127.8 TS7 120.6 TS6 99.6 100 Intermediate V Intermediate IV 50 0.0 0.0 0 Intermediate V + H

Phenol 2 + OH

-170

-180

-190 -195.5 Relative Gibbs Energies (kcal/mol) Energies Gibbs Relative -200 Phenol 3

-210

160

140 134.3 TS8 120 102.3 100 Intermediate VI

20 0.0

0 Relative Gibbs Energies (kcal/mol) Energies Gibbs Relative Phenol 3 + OH -17.5 -20 Acid VII + H.

2.4.3.2 Reactions of Q-65 with –OH anion

Thus, it has been concluded in the previous section that the primary reactions of •OH radical with the Q-65b involve high activation barriers, under the level of theory used in this study.

Additionally, the primary steps for the reaction pathways were discovered to be endergonic. On the flip side, the experimental results obtained from the oxy-cracking of Q-65 involved low activation barriers, in the range of 10-17 kcal/mol. Since the oxy-cracking of Q-65 was performed under the experimental conditions of high pressure and using a basic solution with a water content of >10%, we considered the possibility that the reactions of Q-65 are actually initiated by reactions with the hydroxide anion (–OH). It was also noticed during our experiments on the Q-65 molecule that the pH was dropped from 13 to 12 by the end of the reaction cycle. This 10-fold decrease in the [HO-] concentration strongly suggests the involvement of the hydroxide anion in the oxy- cracking process. Therefore, we repeated the reaction mechanisms depicted in Scheme 2.4 and

Scheme 2.5 by replacing the •OH radical with HO–. In order to account for dispersion effects associated with the HO– anion, we employed the B97D3 density functional as explained in the theoretical calculations section. The reaction schemes were kept the same but with replacing the intermediates III to VI with the corresponding anions. Also, the transition states 4 to 8 were replaced with anions. The final products of the reactions, Phenols II and III, and the carboxylic

‡ acid VII were neutral. Interestingly, both the Gibbs free energy of activations (∆퐺298) and

0 • – reactions (∆퐺298) were dramatically dropped upon the replacement of OH with OH. Table 2.2 shows the difference between the reaction of •OH and–OH with Q-65b. The table also shows the rate constants calculated as per equations 2.1 and 2.2. It is pertinent to know that the operating pressure (ca. 750 psi) of our experiments is different than the pressure where the theoretical 65

calculations are based on (14.6 psi). However, further calculations under the same level of theory used in this work showed that increasing the pressure from 14.6 psi to 1470 psi had a negligible effect on the reaction energies.

Table 2.2 Activation parameters and unimolecular rate constants for the reactions of •OH and– OH with Q-65b.

Reactions with •OH Reactions with -OH

‡ ‡ k298 k298 ∆푮ퟐퟗퟖ ∆푮ퟐퟗퟖ

(kcal/mol) (s-1) (kcal/mol) (s-1)

TS1 134.4 1.6 ×10-86 TS2 128.5 3.3 ×10-82 TS3 38.9 1.7 ×10-16 TS4 131.6 1.9 ×10-84 16.3 7.1 TS5 36.1 2.0 ×10-14 5.0 1.4 ×109 TS6 127.8 1.8 ×10-82 23.8 2.2 ×10-5 TS7 41.9 1.0 ×10-18 27.4 4.9 ×10-8 TS8 134.3 1.9 ×10-86 18.0 0.41

Clearly, the activation barriers are much lower in the case of –OH. This leads to higher rate constants as Table 2.2 shows. Based on these findings, and because of the fact that the calculated rate constants obtained from the experimental part are closer to those obtained for –OH using DFT calculations, we propose that the oxy-cracking of Q-65 might also be initiated by –OH attacks on the aromatic molecule. This is supported by the fact that the reaction was performed under aqueous

basic conditions (pH >8). In addition, and since the activation barrier TS4 is the lowest among all other reaction barriers involved in this work, we propose that the –OH attack on the LUMO of Q-

65 is the most favourable reaction pathway. This observation is actually in agreement with our previous study on the thermal oxidation of Q-65 [63]. In addition, we cannot exclude the role of

O and O2 species in initiating the primary steps of oxidizing the Q-65 molecule. In our previous work [63], the kinetic and thermodynamic parameters obtained from the attack of singlet atomic oxygen (O 1D) on Q-65 were also close to those obtained from the experiment.

In order to fully understand why the theoretical kinetic parameters for the •OH reactions with Q-

65b obtained are different from the experimental one, a thermal and pressure analysis over the temperature range of 25 to 500 ºC and in the pressure range of 15 to 735 psi was conducted. Details on this procedure are explained elsewhere [63, 70, 71]. Surprisingly, and according to the results of this analysis, increasing the temperature or the pressure did not decrease the calculated values

‡ 0 of ∆퐺298 and ∆퐺298. Because the reactions studied in this work are bimolecular and form a single product, the results of the temperature and pressure analysis can be explained using the entropy approach. It is also noteworthy to mention that increasing the pressure above 750 psi in this experimental setup did not alter the reaction kinetics of Q-65b.

2.5 Conclusions

The oxy-cracking process of the solid-waste hydrocarbons model molecule, Quinolin-65 (Q-65), was experimentally investigated in a batch reactor at different operating temperatures 200, 215, and 230 ºC and a constant oxygen partial pressure of 750 psi. The reaction residence times ranged between 0 and 2 h to establish the reaction kinetics. The concentration of the solubilized organic compounds in the liquid phase was monitored by the total organic carbon (TOC), while the gas 67

products were analyzed online using gas chromatography (GC). The Q-65 was oxy-cracked consecutively into different pathways in which an oxidative decomposition took place in the first step producing different aromatic intermediates. The latter compounds were partially oxidized into different families of organic acids and a small amount of CO2. Furthermore, the oxy-cracked Q-

65 has been characterized using FTIR, NMR and XPS techniques. The results showed that various types of oxygenated hydrocarbons were produced, such as carboxylic acids, and phenolic compounds. To support our experimental findings, a theoretical study has been employed by using the density functional theory (DFT) and the second-order Møller–Plesset perturbation theory

(MP2) under the same proposed experimental conditions. The theoretical study succeeded in exploring many oxidation pathways of the Q-65 molecule initiated by hydroxyl radical (•OH).

However, the Gibbs free-energies of activation calculated from the •OH mechanisms were much higher than those obtained from the experiment. Since our experiments were carried out in a strong basic medium (pH=13), and the fact that the pH was dropped from 13 to 12 during the reaction, the theoretical calculations were performed on the hydroxide anion (-OH). The obtained activation energies were found to be much lower than those of the (•OH) radical, and closer to the experimental ones. This suggests that the hydroxide anions play a role in the oxy-cracking reaction under our experimental conditions. In brief, the novelty of this study is considered as a vital backbone for the real oxy-cracking application of petroleum residue which will be addressed in the near future.

2.6 Acknowledgments

The authors are grateful to the Natural Sciences and Engineering Research Council of Canada

(NSERC), the Department of Chemical and Petroleum Engineering at the Schulich School of 68

Engineering at the University of Calgary and Western Canada Research Grid. The authors acknowledge Marianna Trujillo for helping in the NMR analysis.

2.7 References

[1] M. Asif, T. Muneer, Energy supply, its demand and security issues for developed and emerging economies, Renewable and Sustainable Energy Reviews 11(7) (2007) 1388-1413. [2] B. Liddle, Impact of population, age structure, and urbanization on carbon emissions/energy consumption: evidence from macro-level, cross-country analyses, Population and Environment 35(3) (2014) 286-304. [3] S. Brinklov, E.K.V. Kalko, A. Surlykke, Intense echolocation calls from two 'whispering' bats, Artibeus jamaicensis and Macrophyllum macrophyllum (Phyllostomidae), Journal of Experimental Biology 212(1) (2009) 11-20. [4] M. Anandan, S. Ramaswamy, Global oil market: Macro economic scenario, Global Journal For Research Analysis 4(9) (2016). [5] W. Zubot, M.D. MacKinnon, P. Chelme-Ayala, D.W. Smith, M.G. El-Din, Petroleum coke adsorption as a water management option for oil sands process-affected water, Science of the Total Environment 427 (2012) 364-372. [6] J.-H. Lv, X. Wei, Y.-H. Wang, T.-M. Wang, J. Liu, D.-D. Zhang, Z.-M. Zong, Characterization of condensed aromatics and heteroatomic species in Yanshan petroleum coke through ruthenium ion-catalyzed oxidation using three mass spectrometers, RSC Advances 6 (2016) 61758-61770. [7] P. Pourrezaei, A. Alpatova, P. Chelme-Ayala, L. Perez-Estrada, M. Jensen-Fontaine, X. Le, M.G. El-Din, Impact of petroleum coke characteristics on the adsorption of the organic fractions from oil sands process-affected water, International Journal of Environmental Science and Technology 11(7) (2014) 2037-2050. [8] J.M. Hill, A. Karimi, M. Malekshahian, Characterization, gasification, activation, and potential uses for the millions of tonnes of petroleum coke produced in Canada each year, The Canadian Journal of Chemical Engineering 92(9) (2014) 1618-1626. [9] N.N. Nassar, A. Hassan, P. Pereira-Almao, Application of nanotechnology for heavy oil upgrading: Catalytic steam gasification/cracking of asphaltenes, Energy & Fuels 25(4) (2011) 1566-1570. [10] M. Malekshahian, J.M. Hill, Kinetic analysis of CO2 gasification of petroleum coke at high pressures, Energy & Fuels 25(9) (2011) 4043-4048. [11] Y. Zhang, S. Nagamori, S. Hinchiranan, T. Vitidsant, N. Tsubaki, Promotional effects of Al2O3 addition to Co/SiO2 catalysts for Fischer−Tropsch synthesis, Energy & Fuels 20(2) (2006) 417-421. [12] A.P. Steynberg, H.G. Nel, Clean coal conversion options using Fischer–Tropsch technology, Fuel 83(6) (2004) 765-770. [13] E. Mobil, FLEXICOKING™ Conversion Technology. http://www.exxonmobil.com/Apps/RefiningTechnologies/Files/sellsheet_06_flexicoking.pdf. (Accessed 6 February 2017). [14] B. Group, BP statistical review of world energy June 2011, 2011. http://www.bp.com/content/dam/bp- 69

country/de_de/PDFs/brochures/statistical_review_of_world_energy_full_report_2011.pdf. (Accessed October 15 2016). [15] C.E. Baukal, The john zink hamworthy combustion handbook: Volume 1-Fundamentals, CRC press2012. [16] J. Wang, E.J. Anthony, J.C. Abanades, Clean and efficient use of petroleum coke for combustion and power generation, Fuel 83(10) (2004) 1341-1348. [17] A.N. Sawarkar, A.B. Pandit, S.D. Samant, J.B. Joshi, Petroleum residue upgrading via delayed coking: A review, The canadian journal of chemical engineering 85(1) (2007) 1-24. [18] M.S. Rana, V. Samano, J. Ancheyta, J. Diaz, A review of recent advances on process technologies for upgrading of heavy oils and residua, Fuel 86(9) (2007) 1216-1231. [19] F.M. Dautzenberg, J.C. De Deken, Reactor developments in hydrotreating and conversion of residues, Catalysis Reviews 26(3-4) (1984) 421-444. [20] G.B. Brons, M. Siskin, K.O. Wrzeszczynski, Upgrading of bitumen asphaltenes by hot water treatment (C-2726), US 5,326,456, 1994. [21] O.R. Koseoglu, A. Bourane, System and process for integrated oxidative desulfurization, desalting and deasphalting of hydrocarbon feedstocks, US Patent, 2011/0,226,666, 2011. [22] J.G. Speight, New approaches to hydroprocessing, Catalysis Today 98(1) (2004) 55-60. [23] I. Mochida, X.Z. Zhao, K. Sakanishi, Catalyst deactivation during the hydrotreatment of asphaltene in an Australian brown coal liquid, Fuel 67(8) (1988) 1101-1105. [24] S. Maity, V. Perez, J. Ancheyta, M. Rana, G. Centeno, Effect of asphaltene contained in feed on deactivation of Maya crude hydrotreating catalyst, Petroleum Science and technology 25(1-2) (2007) 241-249. [25] J. Levec, A. Pintar, Catalytic wet-air oxidation processes: a review, Catalysis Today 124(3) (2007) 172-184. [26] R.V. Shende, J. Levec, Kinetics of wet oxidation of propionic and 3-hydroxypropionic acids, Industrial & engineering chemistry research 38(7) (1999) 2557-2563. [27] F. Luck, Wet air oxidation: past, present and future, Catalysis today 53(1) (1999) 81-91. [28] S.S. Fong, L. Seng, N.B. Majri, H.B. Mat, A comparative evaluation on the oxidative approaches for extraction of humic acids from low rank coal of Mukah, Sarawak, Journal of the Brazilian Chemical Society 18(1) (2007) 34-40. [29] V. Calemma, R. Rausa, Process for the production of regenerated humic acids from coal, US Patent 4,788,360, 1988. [30] E.M. Peña-Méndez, J. Havel, J. Patočka, Humic substances–compounds of still unknown structure: applications in agriculture, industry, environment, and biomedicine, J. Appl. Biomed 3(1) (2005) 13-24. [31] S.E. Moschopedis, Sulfomethylation of humic acids, lignites, and coals and products thereof, US Patent, 3,352,902, 1967. [32] V. Calemma, R. Rausa, R. Margarit, E. Girardi, FT-ir study of coal oxidation at low temperature, Fuel 67(6) (1988) 764-770. [33] B. Ignasiak, Production of metallurgical coke from oxidized caking coal, US Patent, 4,259,083, 1981. [34] E. Gorin, T.S. Robert, Size separation, preoxidation and fluidized low temperature carbonization of coal, US Patent, 3,047,472, 1962.

[35] M. Ashtari, L. Carbognani Ortega, F. Lopez-Linares, A. Eldood, P. Pereira-Almao, New pathways for asphaltenes upgrading using the oxy-cracking process, Energy & Fuels 30(6) (2016) 4596-4608. [36] V. Platonov, A. Kudrya, S. Proskuryakov, Ozonolysis of asphaltenes from semicoking tar of G17 coal, Russian journal of applied chemistry 76(1) (2003) 148-152. [37] G. Escobar, P. Patiño, S. Acevedo, O. Escobar, M.A. Ranaudo, J.C. Pereira, Interfacial properties of the products of ozonolysis of hamaca crude oil, Petroleum Science and Technology 19(1-2) (2001) 107-118. [38] S.E. Moschopedis, J.G. Speight, Chemical modification of bitumen heavy ends and their non- fuel uses, Shale Oil, Tar Sands, and Related Fuel Sources, American Chemical Society1976, pp. 144-152. [39] S.E. Moschopedis, J.G. Speight, Oxidative degradation of Athabasca asphaltenes, Fuel 50(2) (1971) 211-217. [40] F.J. Rivas, S.T. Kolaczkowski, F.J. Beltrán, D.B. McLurgh, Development of a model for the wet air oxidation of phenol based on a free radical mechanism, Chemical Engineering Science 53(14) (1998) 2575-2586. [41] M.N. Ingale, J.B. Joshi, V.V. Mahajani, M.K. Gada, Waste treatment of an aqueous waste stream from a cyclohexane oxidation unit, Process Safety and Environmental Protection 74(4) 265-272. [42] S.K. Bhargava, J. Tardio, J. Prasad, K. Föger, D.B. Akolekar, S.C. Grocott, Wet oxidation and catalytic wet oxidation, Industrial & Engineering Chemistry Research 45(4) (2006) 1221- 1258. [43] L. Li, P. Chen, E.F. Gloyna, Generalized kinetic model for wet oxidation of organic compounds, AIChE Journal 37(11) (1991) 1687-1697. [44] A. Lifshitz, J.V. Michael, Rate constants for the reaction, O+H2O→OH+OH, over the temperature range, 1500–2400 K, by the flash photolysis-shock tube technique: A further consideration of the back reaction, Symposium (International) on Combustion 23(1) (1991) 59-67. [45] S. Imamura, Catalytic and noncatalytic wet oxidation, Industrial & Engineering Chemistry Research 38(5) (1999) 1743-1753. [46] M. Ashtari, L. Carbognani, P. Pereira-Almao, Asphaltenes aqueous conversion to humic and fulvic analogs via oxy-cracking, Energy & Fuels 30(7) (2016) 5470-5482. [47] P. Haghighat, L.C. Ortega, P. Pereira-Almao, Experimental study on catalytic hydroprocessing of solubilized asphaltene in water: A proof of concept to upgrade asphaltene in the aqueous phase, Energy & Fuels 30(4) (2016) 2904-2918. [48] R.C. Silva, J.R. Radović, F. Ahmed, U. Ehrmann, M. Brown, L. Carbognani Ortega, S. Larter, P. Pereira-Almao, T.B. Oldenburg, Characterization of acid-soluble oxidized asphaltenes by Fourier transform ion cyclotron resonance mass spectrometry: Insights on oxycracking processes and asphaltene structural features, Energy & Fuels 30(1) (2015) 171-179. [49] N. Rodriguez, H. Marsh, E. Heintz, R. Sherwood, R. Baker, Oxidation studies of various petroleum cokes, Carbon 25(5) (1987) 629-635. [50] M. Legin-Kolar, D. Ugarković, Petroleum coke structure: Influence of feedstock composition, Carbon 31(2) (1993) 383-390.

[51] K. Akbarzadeh, H. Alboudwarej, W.Y. Svrcek, H.W. Yarranton, A generalized regular solution model for asphaltene precipitation from n-alkane diluted heavy oils and bitumens, Fluid Phase Equilibria 232(1) (2005) 159-170. [52] O. Castellano, R. Gimon, H. Soscun, Theoretical study of the σ–π and π–π interactions in heteroaromatic monocyclic molecular complexes of benzene, pyridine, and thiophene dimers: implications on the resin–asphaltene stability in crude oil, Energy & Fuels 25(6) (2011) 2526- 2541. [53] P.G. Redelius, The structure of asphaltenes in bitumen, Road Materials and Pavement Design 7(sup1) (2006) 143-162. [54] J.J. Adams, Asphaltene adsorption, a literature review, Energy & Fuels 28(5) (2014) 2831- 2856. [55] H. Wu, M.R. Kessler, Asphaltene: Structural characterization, molecular functionalization, and application as a low-cost filler in epoxy composites, RSC Advances 5(31) (2015) 24264- 24273. [56] O.C. Mullins, The modified Yen model, Energy & Fuels 24(4) (2010) 2179-2207. [57] H. Groenzin, O.C. Mullins, Asphaltene molecular size and structure, The Journal of Physical Chemistry A 103(50) (1999) 11237-11245. [58] F. López-Linares, L. Carbognani, M.F. González, C. Sosa-Stull, M. Figueras, P. Pereira- Almao, Quinolin-65 and Violanthrone-79 as model molecules for the kinetics of the adsorption of C7 Athabasca asphaltene on macroporous solid surfaces, Energy & Fuels 20(6) (2006) 2748-2750. [59] M.F. González, C.S. Stull, F. López-Linares, P. Pereira-Almao, Comparing asphaltene adsorption with model heavy molecules over macroporous solid surfaces, Energy & Fuels 21(1) (2007) 234-241. [60] N.N. Marei, N.N. Nassar, G. Vitale, The effect of the nanosize on surface properties of NiO nanoparticles for the adsorption of Quinolin-65, Physical Chemistry Chemical Physics 18(9) (2016) 6839-6849. [61] A. Devard, R. Pujro, G. de la Puente, U. Sedran, Hydrocarbon yield structure in the conversion of heavy model molecules (Quinolin-65) on fluidized catalytic cracking catalysts, Energy & Fuels 26(8) (2012) 5015-5019. [62] N.N. Marei, N.N. Nassar, G. Vitale, A. Hassan, M.J.P. Zurita, Effects of the size of NiO nanoparticles on the catalytic oxidation of Quinolin-65 as an asphaltene model compound, Fuel 207 (2017) 423-437. [63] I. Badran, N.N. Nassar, N.N. Marei, A. Hassan, Theoretical and thermogravimetric study on the thermo-oxidative decomposition of Quinolin-65 as an asphaltene model molecule, RSC Advances 6(59) (2016) 54418-54430. [64] R.A. Gaussian09, 1, MJ Frisch, GW Trucks, HB Schlegel, GE Scuseria, MA Robb, JR Cheeseman, G. Scalmani, V. Barone, B. Mennucci, GA Petersson et al., Gaussian, Inc., Wallingford CT (2009). [65] A.D. Becke, Density-functional exchange-energy approximation with correct asymptotic behavior, Physical Review A 38(6) (1988) 3098-3100. [66] J.P. Perdew, Density-functional approximation for the correlation energy of the inhomogeneous electron gas, Physical Review B 33(12) (1986) 8822-8824.

[67] G.A. Petersson, A. Bennett, T.G. Tensfeldt, M.A. Al‐Laham, W.A. Shirley, J. Mantzaris, A complete basis set model chemistry. I. The total energies of closed‐shell atoms and hydrides of the first‐row elements, The Journal of Chemical Physics 89(4) (1988) 2193-2218. [68] S. Grimme, J. Antony, S. Ehrlich, H. Krieg, A consistent and accurate ab initio parametrization of density functional dispersion correction (DFT-D) for the 94 elements H-Pu, The Journal of Chemical Physics 132(15) (2010) 154104. [69] A.P. Scott, L. Radom, Harmonic vibrational frequencies: An evaluation of Hartree−Fock, Møller−Plesset, quadratic configuration interaction, density functional theory, and semiempirical scale factors, The Journal of Physical Chemistry 100(41) (1996) 16502-16513. [70] I. Badran, Hot-wire chemical vapour deposition chemistry and kinetics of new precursors in the gas phase and on the wire surface, Department of Chemistry, University of Calgary, Calgary, Alberta, Canada, 2014, p. 310. [71] I. Badran, A. Rauk, Y.J. Shi, Theoretical study on the ring-opening of 1,3-disilacyclobutane and H2 elimination, The Journal of Physical Chemistry A 116(48) (2012) 11806-11816. [72] J.I. Steinfeld, J.S. Francisco, W.L. Hase, Chemical kinetics and dynamics, 2nd ed., Prentice Hall, Upper Saddle River, N.J., 1999. [73] P. Haghighat, L. Carbognani, P. Pereira-Almao, Kinetic study of preoxidized asphaltene hydroprocessing in aqueous phase, Energy & Fuels 30(7) (2016) 5617-5629. [74] J.H. Sinfelt, J. Wiley, Heterogeneous reactions: analysis, examples, and reactor design, Ann Arbor 1001 (1984) 48109. [75] A. Pintar, G. Berčič, M. Besson, P. Gallezot, Catalytic wet-air oxidation of industrial effluents: total mineralization of organics and lumped kinetic modelling, Applied Catalysis B: Environmental 47(3) (2004) 143-152. [76] G. Astarita, Lumping nonlinear kinetics: apparent overall order of reaction, AIChE journal 35(4) (1989) 529-532. [77] N.N. Nassar, C.A. Franco, T. Montoya, F.B. Cortés, A. Hassan, Effect of oxide support on Ni–Pd bimetallic nanocatalysts for steam gasification of nC 7 asphaltenes, Fuel 156 (2015) 110- 120. [78] N.N. Nassar, A. Hassan, P. Pereira-Almao, Thermogravimetric studies on catalytic effect of metal oxide nanoparticles on asphaltene pyrolysis under inert conditions, Journal of thermal analysis and calorimetry 110(3) (2011) 1327-1332. [79] Y. Hamedi Shokrlu, T. Babadagli, In-situ upgrading of heavy oil/bitumen during steam injection by use of metal nanoparticles: A study on in-situ catalysis and catalyst transportation, SPE Reservoir Evaluation & Engineering 16(03) (2013) 333-344. [80] M.N. Siddiqui, Catalytic pyrolysis of Arab heavy residue and effects on the chemistry of asphaltene, Journal of Analytical and Applied Pyrolysis 89(2) (2010) 278-285. [81] P.T. Patil, U. Armbruster, M. Richter, A. Martin, Heterogeneously catalyzed hydroprocessing of organosolv lignin in sub-and supercritical solvents, Energy & Fuels 25(10) (2011) 4713-4722. [82] H. Konnerth, J. Zhang, D. Ma, M.H. Prechtl, N. Yan, Base promoted hydrogenolysis of lignin model compounds and organosolv lignin over metal catalysts in water, Chemical Engineering Science 123 (2015) 155-163. [83] C. Barckholtz, T.A. Barckholtz, C.M. Hadad, A mechanistic study of the reactions of H, O (3P), and OH with monocyclic aromatic hydrocarbons by density functional theory, The Journal of Physical Chemistry A 105(1) (2001) 140-152. 73

[84] C.-Y. Lin, M.C. Lin, Thermal decomposition of methyl phenyl ether in shock waves: the kinetics of phenoxy radical reactions, The Journal of Physical Chemistry 90(3) (1986) 425-431. [85] I. Suh, D. Zhang, R. Zhang, L.T. Molina, M.J. Molina, Theoretical study of OH addition reaction to toluene, Chemical Physics Letters 364(5–6) (2002) 454-462. [86] M. Huang, Z. Wang, L. Hao, W. Zhang, Theoretical investigation on the mechanism and kinetics of OH radical with ethylbenzene, International Journal of Quantum Chemistry 111(12) (2011) 3125-3134. [87] A. Mellouki, G. Le Bras, H. Sidebottom, Kinetics and mechanisms of the oxidation of oxygenated organic compounds in the gas phase, Chemical Reviews 103(12) (2003) 5077-5096.

Conversion of petroleum coke into valuable products using oxy-cracking

techniques

Graphical Abstract

Highlights

• Oxy-cracking process is conducted as an alternative approach to petcoke utilization. • The process was optimized for high conversion, selectively and small amount of CO2. • The petcoke oxidation undergoes in parallel consecutive reactions. • The oxy-cracked petcoke was found to be carboxylic and phenolic acids and their salt. • The oxy-cracking technique could be employed for petcoke demineralization.

This chapter is reproduced by permission of Elsevier from the following publication:

Manasrah, Abdallah D., Nashaat N. Nassar, and Lante Carbognani Ortega. "Conversion of petroleum coke into valuable products using oxy-cracking technique." Fuel 215 (2018): 865-878. (DOI: 10.1016/j.fuel.2017.11.103)

3.1 Abstract

The global production of residual feedstock has reached 150 million metric tons per annum and is expected to increase in the future due to the progressively increasing heavier nature of the crudes.

Petroleum coke (petcoke), one of these residues, is a solid-rich carbon typically produced during the upgrading of heavy oil and delay coking of vacuum residue in the refinery. Finding an alternative technique to treat this massive amount of petcoke is highly needed as the conventional processes like gasification and combustion have limitations in terms of efficiency and environmental friendliness. In this study, an oxy-cracking technique, which is a combination of cracking and oxidation reactions, is conducted as an alternative approach for petcoke utilization.

The reaction is conducted in a Parr reactor where petcoke particles are solubilized in an aqueous alkaline medium and partially oxidized under mild operating temperature and pressure. Several operating conditions on petcoke oxy-cracking were investigated, such as temperature, oxygen pressure, reaction time, particle size and mixing rate to optimize the solubility and selectivity of oxy-cracked products. The results showed that the temperature and the residence time are the two major important parameters that affect the reaction conversion and selectivity. The experimental results enabled us to propose a reaction pathway based on the radical mechanism to describe the kinetic behavior of petcoke. Reaction kinetics indicated that petcoke oxidation undergoes a parallel-consecutive reaction in which an oxidative decomposition took place in the first step producing different oxidized intermediates. The oxy-cracked petcoke was characterized by FTIR,

XPS and NMR analyses. The oxy-cracked products were found to contain carboxylic, carbonyl, phenolic, and sulfonic functions. Moreover, the elemental analysis showed that most of the metals

remained in the residue, suggesting that the proposed technique could be employed for petcoke demineralization.

3.2 Introduction

Coking process is applied to upgrade bitumen, petroleum residue, vacuum residue, solvent- deasphalter pitch, rich asphaltenes and resin fuel, to more desirable products like LPG, naphtha, diesel and gas oil [1, 2]. During the coking process, fuel gas and petroleum coke (petcoke) are also produced [3-5]. This coking process is typically a thermal cracking process in which the H/C atomic ratio of the product is increased by a carbon rejection mechanism [6]. Hence, multi- reactions that undergo a free radical mechanism are coupled together with cracking and polymerization reactions [7]. Cracking reactions produce gas and liquid products which are the most valuable ones, while radical polymerization reactions produce petcoke [8]. This petcoke is a rock-like structure mainly made out of carbon, hydrogen, nitrogen, sulfur and some metals. The global production of petcoke has reached about 150 million metric tons per annum. The North

America alone produces about 70% out of the total petcoke capacity [9]. In 2014, Canada’s oil sands reserves amounted to approximately 166 billion barrels, the third largest after Venezuela and Saudi Arabia [10, 11]. In fact, in 2016, nearly 80 million tons of solid waste hydrocarbons

(petcoke) were being stockpiled in Alberta as a by-product of the Canadian oil sands industries

[9]. Consequently, approximately two-thirds of the produced petcoke have been accumulated and stored on-site for many years [12, 13]. Due to limited markets and minimal use for this commodity, the stockpile is growing at a rate of about 4 million tons a year [14, 15].

Petcoke is also generated in refineries through either delayed or fluid coking processes at high temperature and pressure. These coking processes are highly dependent upon their thermal

treatments [4]. Delayed coking commonly occurs at a temperature range of 415-450 oC, whereas fluid coking uses higher temperatures ranging from 480 to 565 oC [8, 16-18]. Generally, the produced petcoke has a high carbon content (80-85 wt%) consisting of polycyclic aromatic hydrocarbons with low hydrogen content. It is worth noting that petcoke has a lower amount of ash, moisture, and volatiles compared to coal [8]. However, it contains high amounts of sulfur (5-

7 wt%) and vanadium (~ 700 ppm), which can severely impact human and animal health [19].

Additionally, dust emissions from petcoke piles impose a serious threat to people in the vicinity

[19, 20]. Furthermore, petcoke can be classified into calcined and green coke, where the latter is the initial product of the coking process that can be used as fuel in metallurgical and gasification processes [20]. Calcined coke, on the other hand, is produced by treating the green coke at a high temperature between 1200 and 1350 oC [20, 21]. Depending upon its physical properties, structure, and morphology, the calcined coke can be formed as a sponge, shot and needle coke [22]. Oil sands companies often use different coking techniques [23]. Suncor Energy’s coking process, for example, produces “delayed” petcoke, which has a sponge-like structure [24]. In contrast,

Syncrude’s coke is known as “fluid” coke and it has highly-graphitized layers and is often described to have an onion-like structure [23, 25].

Petcoke has a high calorific value, as indicated from the literature at approximately ~37 MJ/kg, which makes it a valuable fuel resource for future application [26, 27]. Therefore, alternative uses of this material could be considered and of paramount importance. Recently, a number of researchers have investigated the potential use of petcoke as a precursor for production of activated carbon due to its high carbon content [5, 28, 29]. Meanwhile, many studies have demonstrated that petcoke can be thermally, physically and chemically activated, in the presence of a reagent to

produce activated carbon with the high specific surface area (~500-1500 m2/g) [28, 30-33].

However, the costs of generating activated carbon through either physical or chemical activation processes tend to be quite high, especially when the yield does not balance the generating costs, thus the total investment cost would be challenging for large-scale industries [34, 35]. Moreover, the corrosive nature of the chemical activating agents has negative environmental impacts and often limits its application [32].

Alternative applications have included the use of petcoke for heat generation purposes, such as combustion and gasification [15, 17, 36]. Hence, petcoke could be an alternative fuel in power generation due to its higher heating value and lower price than coal. Although petcoke is a potential combustible as a fuel source, combustion process would invariably produce large quantities of sulphur dioxide (SO2) [37]. Gasification process, on the other hand, was implemented to produce syngas and hydrogen at high temperatures (800-1800 °C) and pressures (145-1450 psi) [36, 38].

Although the gasification process captures more energy content, it is more capital intensive [39-

42]. Additionally, gasification is not deemed to be a reliable technology for the treatment of these waste hydrocarbons, primarily due to the decline in current gas prices and the recent increase in shale gas production [43, 44].

Despite all the attempts and processes that have been proposed in the industry to treat and reduce the amount of these waste hydrocarbons, a massive amount of petcoke is still present. Therefore, the development of an efficient and environmentally friendly technology is still a challenging and of paramount importance. Recently, our research group has introduced an oxy-cracking technique for converting heavy solid hydrocarbons into valuable products using a model residual feedstock exemplified by Quinolin-65 (Q-65) and n-C7 asphaltenes [40, 45]. This technique is a combination

of oxidation and cracking reactions in an aqueous alkaline media is inspired by asphaltenes oxidation and ozonolysis studies [46-51]. Through the oxy-cracking process, a new reaction pathway is offered in an aqueous alkaline medium, at mild temperatures (170-230 °C) and pressures (500-750 psi). Importantly, the oxy-cracking process has a high efficiency to convert the solid waste hydrocarbons into valuable light commodity products [40, 45]. The oxy-cracked materials (i.e., not completely oxidized) become soluble in water via oxygen incorporation due to the polar functionalization of the aromatic edges and paraffinic terminal carbons. Hence, the oxy- cracking undergoes a parallel-consecutive reaction in which an oxidative decomposition took place in the first step producing different aromatic intermediates. Those intermediates could be carboxylic acids and their corresponding salts, or other products [45, 52, 53]. Because of the high solubility and selectivity of oxy-cracked products in the alkaline medium and the insignificant amount of produced CO2, the oxy-cracking process could be considered an environmentally friendly one. Similarly, a theoretical and experimental study on oxy-cracking of the Quinolin-65

(Q-65) molecule as a model molecule for residual feedstocks has recently confirmed that the attacks of the hydroxyl radical (∙OH) plays an important role in the oxy-cracking reaction of Q-65 and n-C7 asphaltenes [40, 45].

Herein, in this study, we are applying the oxy-cracking technique as a new approach for converting petcoke into valuable commodity products by solubilizing it in water under alkaline conditions.

The maximum solubilization and selectivity of oxy-cracked petcoke (i.e., desired products) in alkaline media with minimal emission of CO2 are the main target. Eventually, at the end of the reaction, the oxy-cracked products are characterized using Fourier transformed infrared spectroscopy (FTIR), total organic carbon analysis (TOC), nuclear magnetic resonance (NMR)

spectroscopy, and X-ray photoelectron spectroscopy (XPS) techniques. The gas emissions generated during the oxy-cracking process are also idintifed using gas chromatography (GC). The non-reacted solid residue is also analyzed using elemental analysis and FTIR as well. Moreover, the generalized “triangular” lumped kinetic model [54] is used here to determine the reaction kinetics parameters and the potential mechanism for the oxy-cracking of petcoke. It is expected that this study opens a better outlook about the use of the oxy-cracking process in the oil industry, mainly in treating residual feedstock such as petcoke and the like.

3.3 Experimental work

Materials

A sample of green petcoke was obtained from Marathon Petroleum Company (Garyville, USA).

This black-solid sample has complex hydrocarbons structure which consists of polycyclic aromatic hydrocarbons (3-7 ring), such as benzopyrene. The sample was grinded and sieved to the particle size ranging between 53 and 710 µm. The elemental analysis for the petcoke sample was carried out using a PerkinElmer 2400 CHN analyzer (Waltham, Massachusetts, USA) for C, H, N contents and a Thermo Intrepid inductively coupled plasma-atomic emission spectroscopy (ICP-AES) for sulfur and metal contents. The chemical composition of the selected petcoke sample is listed in

Table 3.1.

KOH (ACS reagent, ≥85%, pellets form) purchased from Sigma-Aldrich (Ontario, Canada) was used to adjust the pH of the reaction medium (deionized water) and help in solubilizing the petcoke in water. Oxygen 99.9% ultrahigh purity purchased from Praxair (Calgary, Canada) was used as the oxidant gas.

Table 3.1 The chemical composition of the green petcoke sample considered in this study.

Composition C H N S V Fe Ni Mo Co O* wt% 84.48 3.81 1.55 4.46 0.08 0.06 0.03 0.01 0.15 5.37

* Estimated by difference.

Experimental procedures and setup

Figure 3.1 shows a schematic representation of the experimental setup. The setup consists of a 100 mL reactor vessel (model number 4598, Parr Instrumental Company, Moline, Il, USA). The reactor vessel is made of stainless steel SS-316 with 12 cm in length and 3.25 cm in diameter. The vessel was equipped with a heating oven connected to a temperature control loop, a pressure gauge and a mechanical stirrer with a speed controller. The reactor vessel is capable of handling pressures up to 1700 psi and temperatures up to 270 °C. The oxy-cracking experiments were carried out at temperatures from 150 to 250 oC and pressures up to 1000 psi. In a typical experiment, 1.0 g of solid petcoke sample was charged into the reactor vessel containing 20 g of deionized water and a specified amount of KOH. The pH of the reaction medium is kept above 8.0 by adding 1.0 g KOH to assist in solubilizing petcoke and to avoid corrosion problems. Leak tests were performed by pressurizing the reactor with O2 up to 1200 psi prior to fixing the operating pressure. Then, the mixer was set to 1000 rpm to minimize the interfacial mass resistance between the gas and liquid phase and to ensure uniform temperature and concentration profiles in the liquid phase. The reactor was then heated to the desired temperature. Once the desired pressure and temperature are attained, the zero-reaction time was considered. The reaction was carried out at different residence times, namely 15, 30, 45, 60, 120, 180 and 240 min. Several operating parameters were investigated to optimize the oxy-cracking reaction, such as temperature, reaction time, oxygen pressure, mixing

speed, particle size and amount of KOH. At the end of the reaction, the reactor was cooled down to room temperature. Then, the gas phase was analyzed using gas chromatography, GC (SRI

8610C, SRI Instruments). Afterwards, the liquid effluents were carefully discharged and filtered for total organic carbon (TOC) analysis. A small amount of presumably unreacted solid residue was collected at the bottom of the reactor vessel. The oxy-cracked and insolubilized (residual) petcoke were recovered using an evaporator (vacuum oven) for further analysis by Fourier transformed infrared spectroscopy (FTIR), nuclear magnetic resonance (NMR) spectroscopy and

X-ray photoelectron spectroscopy (XPS) to investigate the nature of the formed compounds. Also, elemental analysis was performed on the dried recovered solids.

Figure 3.1 Schematic representation of the experimental setup (not to scale).

Characterization

3.3.3.1 FTIR analysis

The chemical structure of the virgin petcoke sample, oxy-cracked (solubilized) and insoluble solid

(residue) were characterized with a Shimadzu IRAffinity-1S FTIR (Mandel, USA), provided with 83

a smart diffuse reflectance attachment to carry out diffuse reflectance infrared Fourier transform spectroscopy (DRIFTS) analysis. Initially, the background was defined by analyzing about 500 mg of pure potassium bromide (KBr) powder; then, approximately 5 mg of the petcoke sample dispersed in the 500 mg of KBr was analyzed. The IR spectra were obtained in the wave number ranging from 400 to 4000 cm-1; all the spectra were acquired as averages of 50 scans with a resolution of 4 cm-1. It is worth noting here, in case of oxy-cracked (solubilized) sample, the solidified organic species were collected by drying the solubilized petcoke in water overnight at

65 oC in a vacuum oven.

3.3.3.2 Total organic carbon (TOC) analysis

A Shimadzu total organic carbon analyzer (TOC-L CPH/CPN) was used to determine the carbon content of the solubilized organic and inorganic species present in the water. The TOC samples were prepared by centrifuging the solubilized species in a centrifuge (Eppendorf centrifuge 5804) at 5000 rpm and 15 min to separate the remaining solid (i.e., unreacted and insolubilized species).

The total carbon (TC), total organic carbon (TOC), and inorganic carbon (IC) of the aqueous phase were measured. Both TC and IC measurements were calibrated using standard solutions of potassium hydrogen phthalate and sodium hydrogen carbonate. Fifteen milliliters of the centrifuged solutions were placed in standard TOC vails. Using the TOC software to control the system, the TC was automatically measured. After that, an acid was added to evolve CO2 from the sample to measure the remaining organic compounds, which were considered as TOC. All the measurements were taken three times, and the average was used for the calculations with a 5% relative standard deviation.

3.3.3.3 1H Nuclear magnetic resonance (NMR) spectroscopy

The NMR spectrum of the oxy-cracked sample was determined with a Bruker 600 MHz spectrometer (4 mm BL4 liquid probe, cross-polarization program, and spin rate of 8k). The 1H

NMR spectrum was taken at 298 K using a D2O solvent with a pulse sequence zg30, a relaxation time of 2 s, and averaging 160 scans/run. The NMR spectrum was analyzed using the commercial

NMR simulator software (Mnova NMR) helping the assignment of most structure types available at different frequencies.

3.3.3.4 Gas chromatography (GC) analysis

The compositional analysis of the produced gases was carried out with a GC (SRI 8610C Multiple

Gas #3 gas chromatograph SRI Instruments, Torrance, CA). The GC was provided with a thermal conductivity detector (TCD) and two packed columns connected in parallel (3′ molecular sieve/6′

Hayesep-D columns). The molecular sieve column is used for permanent gases, while the

Hayesep-D column allows analysis for hydrocarbons up to C5. The gas analysis was carried out after the oxy-cracking reaction is completed and cooled down to room temperature. The GC measurements were repeated 5 times for each sample, and the average relative error was lower than 3%.

3.3.3.5 X-ray photoelectron spectroscopy (XPS)

The XPS analysis was conducted on the petcoke sample before and after reaction using an XPS

PHI VersaProbe 5000 spectrometer to provide information about the distribution of different atoms on the sample surface based on their binding energy. The oxy-cracked sample was collected after drying it in a vacuum oven at 65 oC overnight. XPS spectra give an additional information about the nature of bonds and component analysis [55]. The spectra were taken using a

monochromatic Al source (1486.6 eV) at 50 W and a beam diameter of 200.0 μm with a take off angle of 45°. The samples were pressed on double-sided tape and the spectra were taken with double neutralization. The sample sputtering protocol involved 20 min of Argon sputtering at 45°,

2 kV, 1.5 μA 2 × 2 (less than 10.5 nm/min). Calibration was performed with a SiO2/Si wafer having a SiO2 layer of 100 nm.

3.3.3.6 Elemental analysis

A combustion method using a PerkinElmer 2400 CHN analyzer (Waltham, Massachusetts, USA) was used for analyzing carbon, hydrogen, and nitrogen contents after and before oxy-cracking reaction. Both sulfur and nitrogen contents for organic materials were determined with an Antek

9000 system (Houston, TX, USA) by running toluene solutions (10 wt %/vol.). Calibration was performed with Accustandard IS-17368 (N) and Accustandard SCO-500x (S) standards.

For metal analysis (Fe, Ni, Co, Mo and V), the microwave assisted acid digestion procedure was used in a commercial unit model MARS 6 from CEM Corporation (Matthews, NC, USA) for digesting the solid residual samples. The system is provided with UltraPrep vessels of 100 mL capacity and a MARSXpress DuoTemp controller which was operated at a frequency of 2.45 GHz at 100% of full power (maximum of 1600 W). Sulfur and metal concentrations in the oxy-cracked and residual samples were determined by ICP−AES.

3.4 Results and Discussion

Reaction kinetics

The oxy-cracking reaction mechanism of heavy hydrocarbon compounds is very complex [56].

Even with a pure compound such as phenol, the high-temperature wet oxidation exact mechanism or reaction pathway has not been established yet [54, 56]. The wet oxidation of hydrocarbon

mixtures is much more complex than a single compound. Based on our recent findings on the oxy- cracking of model hydrocarbon compounds (Q-65), it was observed that the compound underwent a consecutive reaction in which an oxidative decomposition takes place in the first step producing different aromatic intermediates [45, 57]. The intermediates were oxy-cracked consecutively into different families of organic acids and small amounts of solubilized CO2 [56]. Similar reaction pathway has been proposed for oxy-cracking of n-C7 asphaltenes [40]. Using oxygen as an oxidizing agent, the asphaltenes was oxygenated and functionalized to different families of carboxylic, sulfonic, and phenolic compounds. This mechanism is inspired by the generalized kinetic model for wet air oxidation of organic compounds existing in wastewater which was proposed by Li et al. [54]. Accordingly, the oxy-cracking mechanism of petcoke is believed to obey similarly the triangular reaction pathway, as depicted in Figure 3.2, where petcoke slowly solubilized in water with small quantities of produced CO2 at early stages of the reaction. Under certain reaction conditions, longer residence time and higher temperature, the solubilized petcoke in water starts reacting with oxygen to produce CO2.

Considering the petcoke has a complex structure [58-60], many soluble and insoluble intermediates were produced during the reaction. The concentration of the intermediates (desired products) in the liquid phase was calculated based on carbon mass as the lumped total organic carbon (TOC) concentrations. However, the carbon content of initial feedstock was calculated using elemental analysis; carbon content before reaction (feedstock) = (mass of petcoke) × (carbon % in feed). The produced gas, most likely CO2, was analyzed online using GC. Other determined gas concentrations were very small, thus neglected. The reaction conversion based on carbon mass was calculated based on the following equation,

C −C Conversion, X = A0 R (3.1) CA0 where CAo is the carbon concentration of virgin petcoke before the reaction, CR is the residual carbon concentration (unreacted petcoke) that remains after the reaction. It is worth noting that the

numerator term (퐶퐴0 − 퐶푅) in Eq (1) represents the amount of carbon in the liquid phase as total carbon (푇퐶 = 푇푂퐶 + 퐼퐶) and the amount of carbon in the gas phase CO2 (CG). Hence,

CA0 − CR = (TOC) + (IC) + CG (3.2) considering the CO2 gas obeys the ideal gas behavior, then the carbon content in the gas phase (CG) could be calculated as follows;

PV C = 12 × (3.3) G RT

where, 푃 and 푇 are the pressure and temperature at the end of reaction, respectively. 푉 is the

volume of the gas phase in the reactor vessel and 푅 is the ideal gas constant.

The selectivity to produce the desired products (B) and CO2 (C) was calculated as follow,

(TOC) Selectivity to product B = (3.4) (TOC)+IC+CG

(IC+C ) Selectivity to product C = G (3.5) (TOC)+IC+CG

The kinetic rate equations for the oxy-cracking reaction in a batch reactor, as shown in Figure 3.2,

can be expressed by the set of the following three differential equations:

dC A = −r = (K + K )Cn1 (3.6) dt A 1 2 A

dC B = +r = K Cn1 − K Cn2 (3.7) dt B 2 A 3 B

dC C = +r = K Cn1 + K Cn2 (3.8) dt C 1 A 3 B

where,

−E1/RT m K1 = k′1e CO2 (3.9)

−E2/RT m K2 = k′2e CO2 (3.10)

−E3/RT m K3 = k′3e CO2 (3.11) where CA, CB, and CC are the carbon concentrations of original petcoke, desired products, and CO2, respectively. CO2 is the concentration of oxygen, n1, n2 and m are the reaction order of A, B and

O2, respectively. t is the reaction time, and K1, K2, and K3 are the reaction rate constants.

The reaction orders are experimentally determined to be first order for A and B, i.e., n1 = n2 =

1. Typically, the order of oxygen is either near zero (푚 = 0) or excess oxygen is used to reduce its effect on the reaction kinetics and enable hydrocarbon species (A and B) to be the limiting reactant [54, 61, 62]. Therefore, the oxygen terms will be considered as a constant, hence,

Equations 6 to 8 can be expressed as follows; dC A = −(K + K )C (3.12) dt 1 2 A dC B = K C − K C (3.13) dt 2 A 3 B dC C = K C + K C (3.14) dt 1 A 3 B

The kinetic parameters, i.e., K1, K2, and K3 were estimated using the Mathematica software

(V10.2) by fitting the experimental data to the differential equations (3.12-3.14) under the following initial conditions: at t = 0, CA = CAo, and CB = CC = 0. The proportional weighed sum- of-squares was minimized using the Mathematica until all values of the correlation coefficient (R2) were very close to 1.0. The kinetics experimental data were collected at three different temperatures of 200, 215, and 230 ºC and reaction times varying from 0 to 1 h. However, other important parameters, such as the operating partial pressure (750 psi), the mass ratio of petcoke to

KOH, and the impeller speed (1000 rpm) were all kept fixed. At these temperatures and reaction

times, the best range of conditions was selected to make the reaction favorable to the desired products. Indeed, at high temperatures (> 250 oC) and residence times (>2 h), combustion reaction becomes more favorable than oxy-cracking and more CO2 was produced [54]. However, low reaction conversions were obtained at low temperatures (<180 °C). The oxy-cracking reaction was not significantly affected by the oxygen partial pressure beyond 750 psi. Also, the oxy-cracking reaction rate was found to be independent of the impeller speed above 500 rpm, indicating there is no mass transfer limitation beyond this speed limit. These constrains and findings were acknowledged in the literature for oxy-cracking and wet air oxidation of hydrocarbon compounds

[40, 43, 45, 52, 53, 63, 64].

The estimated reaction constants of the petcoke oxy-cracking are presented in Table 3.2.

Consequently, the activation energies and frequency factors were estimated using Arrhenius equation based on the temperature and reaction constants as follow:

−Ei ′ RT Ki = kie (3.15)

′ where 푘푖 is the frequency factor for each step of the reaction, 퐸푖 is the activation energy, 𝑖 is the reaction step pathway (1, 2, and 3), 푅 is the ideal gas constant, and 푇 is the temperature.

Figure 3.3 compares the experimental data with the kinetic model for concentration profiles of petcoke (A), intermediate compounds (B), and CO2 (C) at three different temperatures of 200, 215, and 230 ºC as a function of time. Error bars shown in the figure represent the calculated standard deviation based on the TOC and GC measurements. Noticeably, the kinetic model showed an excellent agreement with the experimental results and described the proposed triangular reaction kinetics scheme accurately. It is clear that the reaction temperature is acting as a key parameter in the oxy-cracking reaction. Thus, at a higher temperature (i.e., 230 oC), the solubilization of oxy- 91

cracked compounds in water is increased and reached to the maximum concentration faster than at lower temperatures. Moreover, at a high reaction temperature, the produced CO2 in the gas phase is detected at the early stage of the reaction. Even at low reaction time, i.e., 15 min, the amount of produced CO2 is determinable by GC. This indicates that a direct reaction might be occurring between oxygen and petcoke to form CO2. These findings line up perfectly with our previous works on oxy-cracking of model compounds (Q-65) and asphaltenes [40, 45].

Table 3.2 Determined values of oxy-cracking reaction constants.

T K1 K2 K3 (°C) s-1 s-1 s-1 200 2.27 ×10-5 1.84 ×10-4 1.71 ×10-5 215 6.99 ×10-5 3.46 ×10-4 2.42 ×10-5 230 2.43 ×10-4 87.37 ×10-4 3.67 ×10-5

Figure 3.4 represents the Arrhenius plot of petcoke oxy-cracking reaction at three different reaction temperatures. By plotting ln(k) against 1/T, a good fitting was accomplished between Arrhenius equation and the experimental data, indicated by R2 values closed to 1. From the slope and intercept of the best-fit-line at each temperature, the values of activation energies and frequency factors of petcoke oxy-cracking were calculated and summarized in Table 3.3.

Moreover, the experimental and theoretical findings in our previous works on oxy-cracking [45] and thermo-oxidative decomposition [65] of Q-65 as a model molecule of heavy solid hydrocarbons confirmed that the reaction initiates by cracking the weakest bond in the alkyl chain to produce CO2. However, in case of Q-65 oxy-cracking reaction, after the alkyl chain initially cracked using hydroxyl radical (∙OH), the dissociation of the carbon bonds adjacent sulfur is consequently taken place. After that, the whole molecule is completely cracked to produce phenolic and carboxylic substances and a small amount of CO2. Accordingly, similar reaction mechanism and pathways could be considered here for the case of petcoke oxy-cracking. Hence, at the beginning of the reaction, an induction period is found in which there is small amount of

CO2 released. The reason of that refers to the oxygen which is initially incorporated into the petcoke/hydrocarbons molecules, after that a complete release of CO2 takes place. Nevertheless, this small amount of CO2 was noticed due to the short alkyl chains that present in the petcoke structures as confirmed in the FTIR results. This finding also supported by the highest value of activation energy (E1 = 39.46 kcal/mol) in the first reaction pathway. Additionally, the value of activation energy could be attributed to the complexity of aggregated structures of petcoke molecules [66-68]. Subsequently, petcoke aggregates require more oxygen penetration during the oxy-cracking reaction to achieve the desired conversion. Even though petcoke is not completely

dissolved in water at the beginning of the reaction, their nature is changed during the reaction due to the presence of KOH and partial oxidation. Therefore, petcoke particles were solubilized in water as oxygenated hydrocarbons analogs of carboxylic acids and the like. These findings were confirmed in the second reaction pathway where the activation energy E2 = 21.87 kcal/mol and

6 -1 a high value of the frequency factor 2.19 × 10 s . Consequently, CO2 could be produced in the third reaction pathway, E3 = 11.98 kcal/mol, by further reaction between solubilized aromatic moieties and oxygen. Although the activation energy for deep oxidation of petcoke to produce CO2 in the first reaction pathway (39.46 kcal/mol) is much higher than the one obtained in the third pathway (partial oxidation) (11.98 kcal/mol), the frequency factor in the first pathway (1.74 ×1012 s-1) is also higher than the third pathway (5.75 s-1). These findings also support that the rate consuming of petcoke into CO2 at the beginning of reaction could happen at the same rate of CO2 production from the oxidation of the organic compounds solubilized in water. In other words, the rate of forming and producing intermediate compounds (desired products) is more favorable than producing CO2 in both reaction pathways, as the activation energy (E2) was lower than E1 and the frequency factor (k′2) is higher than k′3. Similar trends were reported for oxy-cracking of model hydrocarbons Q-65 [45] and n-C7 asphaltenes [40].

Figure 3.4 Arrhenius plots of petcoke oxy-cracking for each reaction pathway. Table 3.3 Estimated activation energies and frequency factors of petcoke oxy-cracking.

Activation energy (kcal.mol-1) Frequency factor (s-1) 퐸 1 39.46 ±0.495 1.74 × 1012 퐸 2 21.87 ±0.532 2.19 × 106 퐸 3 11.95 ±0.981 5.75

Effects of operating conditions on petcoke oxy-cracking reaction

In this section, the effects of operating conditions such as temperature, residence time, oxygen partial pressure, amount of KOH, petcoke particle size and impeller speed were investigated. These parameters were optimized not only to maximize the reaction conversion and selectivity to produce the water-solubilized hydrocarbons (desired products) but also to minimize the amount of CO2 produced during the oxy-cracking reaction. Preliminary experiments were conducted to optimize

the operating oxygen partial pressure. The results revealed that the reaction conversion was not significantly affected by oxygen partial pressure beyond 750 psi (Appendix A2). Therefore, at this pressure and temperature range (180- 250 °C), the water exists only as a subcritical liquid. It is worth noting here that the oxygen is not utilized only in converting the petcoke to CO2 and water, but also a good proportion of oxygen consumed in converting petcoke to water-soluble materials.

Hence, at the given pressure, we ensured that oxygen is present in an excess amount.

The effect of mixing was investigated during the petcoke oxy-cracking reaction. High mixing speed should minimize the interfacial mass resistance between the gas and liquid phase, therefore enhancing the oxygen to transfer from the gas phase to the liquid phase. Additionally, the mixing speed helps to maintain a relatively uniform temperature and concentration profiles in the liquid phase [69-71]. The reaction conversion was evaluated by varying the mixing speed from 0 to 1000 rpm while fixing other parameters such as temperature (215 oC), oxygen pressure (750 psi) and reaction time (2 h). As seen in Figure 3.5, it is clear that when the mixing speed is below 500 rpm, a significant reduction in the reaction conversion occurred, thus the mass transfer region considered to be the controlling step. However, above 500 rpm, the effect is drastically reduced and there was no practically effect on the reaction conversion, i.e., the reaction region is the controlling step. Therefore, a high mixing speed would result advantageous for the aqueous phase by providing a well-mixed reactor content during the reaction. Thus, to avoid the mass transfer resistance the reaction was taking place in the turbulent region (i.e., Reynolds numbers, Re >

10000).

Figure 3.5 Effect of mixing speed on the conversion of petcoke during oxy-cracking reaction (T = 215 oC, P = 750 psi and t = 2 h).

The effects of petcoke particle sizes were also investigated on the conversion of oxy-cracking reaction. Different petcoke particle sizes ranging from 53 to 710 µm were tested to examine their effect on petcoke solubilization or any mass transfer limitations. Figure 3.6 shows the reaction conversion of petcoke oxy-cracking evaluated at different petcoke particle sizes, constant temperature (215 oC), mixing speed (1000 rpm), oxygen pressure (750 psi) and reaction time (2 h). As expected, no remarkable effect of particle size on the solubilization of petcoke can be seen.

This suggests that the mass transfer has a relatively minor impact on the studied range of particle sizes. Within average, the total petcoke conversion to desired products and CO2 was approximately constant (about 78.5%), and independent of particle size. These findings are valuable because grinding and manipulating very small particle sizes of petcoke are challenging and quite costly.

Therefore, the obtained results have relevance for the possible implementation of the process at large scale.

Figure 3.6 Effect of petcoke particle size on reaction conversion of petcoke (T = 215 oC, P = 750 psi and t = 2 h). 3.4.2.1 Effect of the temperature

The effect of the temperature on the conversion and selectively of the oxy-cracking reaction was investigated between 180 and 250 °C. Other important parameters such as the oxygen partial pressure was set to 750 psi to make sure the water was present in the subcritical state, mixing rate was 1000 rpm to prevent the liquid phase interfacial mass transfer resistance, and the residence time was 1 h. Based on the reaction kinetic results, the reaction performance improved by increasing the temperature. In fact, the reaction temperature is a key parameter in the oxy-cracking reaction. Thus, by increasing temperature (i.e., up to 250 oC) the solubilization of oxy-cracked compounds in water is increased. Although the solubilization of oxygenated hydrocarbons is

increased at a high temperature, the selectivity of producing CO2 gas is also increased. Hence, under longer reaction times the oxygenated intermediates further decomposed oxidatively to CO2 and H2O. Furthermore, the chemical transformation of oxygenated intermediates to CO2 appear to have complex mechanisms even at low reaction temperature.

Figure 3.7 shows the conversion and selectivity of oxy-cracking reaction at different reaction temperatures. It is clear that as the temperature increased the petcoke conversion to produce solubilized-hydrocarbons (B) is increased with a slight increase in CO2. However, the selectivity to produce the desired products (B) is slightly decreased with a further increase in temperature

(250 oC). Moreover, no reaction occurred at temperatures lower than 150 °C with the considered residence time. For instance, the reaction conversion was less than 30% when the temperature ranged from 150 to 180 °C at 1 h residence time. Interestingly, the highest conversion was obtained when the temperature ranged from 220 to 240 oC. Based on that, the optimum reaction temperature which provides the highest conversion and selectivity to B synchronized with a minimal amount of CO2 centers around 230 °C, as presented in Figure 3.7.

The results showed that the temperature has a great effect on the conversion and less effect on the selectivity to water solubilized products. Nevertheless, the oxy-cracking temperature is a key parameter not only on the conversion and the selectivity but also on the acidity of formed products.

Even the reaction rates are increased with temperature, the final TOC values of the desired products

(B) for temperature higher than 230 oC are practically constant after 1 h. The reason of that is due to the ability of the low-molecular weight chemical species that resist the oxidation process.

Another reason for this observation is the short life of free radicals (scavenging effects) since the presence of high strength carbonate (KOH) might kill some of the free radicals which otherwise

100

directly attack the other organic compounds as will be discussed in Section 3.4.2.3. Moreover, when the temperature is lower than 200 °C, the aqueous phase has an excess amount of free KOH

(i.e., pH >10). However, more acidic functional groups were produced at higher temperatures

(200−250 °C), and this is confirmed by lowering the values of pH for neutralization reactions to about 8.5. The pH reduction lines up perfectly with similar trends reported for n-C7 asphaltenes oxy-cracking [40, 52] and oxidizing organic compounds in subcritical water such as phenols, benzene, and other alcohols [71, 72].

Figure 3.7 Effect of the reaction temperature on the selectivity and conversion of petcoke oxy- cracking (P = 750 psi and t = 1 h).

3.4.2.2 Effect of reaction times

The effect of reaction time was also studied by varying the time from 15 min to 4 h under a constant pressure (750 psi), mixing speed (1000 rpm) and operating temperature of 180 °C. The effect of reaction time on the conversion and selectivity of oxy-cracking reaction is shown in Figure 3.8. It

101

is clear that the reaction conversion of petcoke to produce oxy-cracked hydrocarbons (B) and CO2 is significantly increased with time. However, the selectivity to product B is slightly decreased with further increase in time, and simultaneously the selectivity to product C slightly increased with time. The optimum reaction time could be considered when the conversion and selectivity of

B are high which was at time 2 h. Moreover, the effect of reaction time on the acidity of formed products has been noticed. By increasing the reaction time, the pH of the liquid phase decreased, thus more acidic compounds were produced.

Figure 3.8 Reaction time effect on selectivity and conversion of petcoke oxy-cracking reaction (T = 180 oC and P = 750 psi).

3.4.2.3 Effect of KOH

The oxy-cracking reaction was studied by changing the dosage of KOH from 0 to 2.5 g at constant temperature (230 oC), oxygen pressure (750 psi), reaction time (2 h) and mixing speed (1000 rpm).

Figure 3.9 shows the effect of KOH on the reaction conversion and selectivity to both B and C. As 102

seen, the conversion as well as the selectivity to B, significantly increased by increasing the amount of KOH and then slightly decreased by further increase of KOH dosage. However, the selectivity to C decreased by increasing the KOH amount. Thus, the optimal amount of KOH was chosen to be (1 g KOH/ 1 g petcoke) where the highest values of the reaction conversion and selectivity to

B were achieved and the lowest amount of CO2 was produced. In fact, KOH was needed here to prevent the corrosion effect caused by the high acidity species generated during the early oxidation stage of the process. The results showed that the KOH is also a key parameter for enhancing the solubilization of oxy-cracked materials, therefore increasing the reaction conversion, as well as the selectivity to the desired products. Some researchers [49, 52, 73] claimed that adding a base such as NaOH, KOH and NH4OH might have a catalytic effect on water-solubilized materials by neutralizing the acidic species and/or displacing the saponification equilibrium. For example, naphthenic acids dissolve in hot water at basic conditions formed soluble organic salts [74]. These findings have been confirmed by other studies [44, 75, 76] where the solubility of species containing oxygen increased in the basic media. Patil et al. [75] showed that the solubility of lignin in water increased in presence of NaOH. Likewise, Ashtari et al. [52] showed that the alkaline medium enhanced n-C7 asphaltene oxy-cracking and its subsequent solubilization in water. In this regard, 1.0 g of KOH per 1.0 g of petcoke was considered as the optimum value for the oxy- cracking experiments.

103

Figure 3.9 Effect of KOH amounts on the selectivity and conversion of petcoke oxy-cracking reaction (T = 230 oC and P = 750 psi, time = 2 h).

Characterization results

3.4.3.1 FTIR analysis for petcoke and oxy-cracking products

FTIR analysis is one of the most versatile techniques to understand the structures of such complicated organic mixtures like petroleum coke and heavy hydrocarbons. This is because FTIR has the ability to provide information about the functional groups based on their bond energies and orientation of atoms in space. The FTIR spectrum of the original petcoke was carried out and compared with the oxy-cracked product and the non-converted residue as well. Figure 3.10 shows the infrared spectra of the original petcoke, residual petcoke (non-soluble solid) and oxy-cracked petcoke solubilized fraction isolated from the reaction carried out at 230 °C and 2 h (i.e., the optimum conditions). It is evident from the figure that FTIR spectra of original petcoke and the oxy-cracked one are distinctly different from each other.

104

The spectrum of the original petcoke shows IR bands that can be assigned to the alkyls/aliphatic

(2850-3000 cm-1) and aromatic (~3040 cm-1 and 930-750 cm-1) regions. The presence of C-H bonds vibration out-of-plane in aromatics can be assigned to the 748, 804, and 860 cm-1 bands.

The corresponding C=C aromatic stretching vibration appears near 1580 cm-1, slightly below the typical 1600 frequency, thus believed conjugated with other groups like C=C region as reported in other studies [77-79]. However, for the oxy-cracked sample, the noticeable lower contribution from aromatic out plane bands is observed (930-750 cm-1). The transmittance at 3040 cm-1 due to aromatic C-H stretching vibrations can be found in the spectra for both the original and the oxy- cracked petcoke; however, much less important in the later.

Furthermore, in the aliphatic region, the presence of alkyl groups in the petcoke sample such as –

-1 -1 CH3, =CH2 and –CH2CH3 is evidenced by the bands around 2940 cm and 1380 cm which can be assigned to asymmetric and symmetric –C–H stretching and bending vibrations, respectively

[80]. The weak band at around 3500 cm-1 observed for the original petcoke can be assigned to free

O–H stretching vibration mode of hydroxyl functional groups. The broad-band spanning from about 2700 to 2000 cm-1 possibly corresponds to hydrogen bonded –OH functionalities. The presence of sulfoxide species in the original petcoke is assigned at the small band ∼1031 cm-1.

These findings are in accordance with many studies reported by Pruski et al. [81] and Michel et al.

[82] that showed the raw (green) petcoke comprises polynuclear aromatics with few alkyl chains as substituents and polynuclear aromatic molecules such as naphthalene, anthracene, benzopyrene, phenanthrene, coronene, triphenylene, and pyrene [67, 83].

The FTIR spectrum of the insoluble petcoke (solid residue after reaction) is also appended in

Figure 3.10. Worth mentioning here that small amount of solid residue could only be collected at

105

low reaction conversion for further characterizations. The structures of insolubilized solid material

(residue) was found very similar to the original petcoke according to their IR spectra, with some features changed due to the contribution of oxygenated functions. It is clear from the spectrum that at 3300−3700 cm-1 there is a higher contribution of OH groups in the remaining insolubilized solid compared with the original petcoke. Also, the C-O-C contributions (1363 cm-1) in the remaining solids was found less intense compared to the original petcoke which showed a broad-band spanning from about 1360-1100 cm-1. This later band can also be derived from the contribution of sulfones (centered in 1130 cm-1), in addition to other S-oxidized forms (sulfoxide at 1030 cm-1) with higher intensity compared with the original and oxy-cracked samples.

The FTIR spectrum of the oxy-cracked petcoke is dramatically different from that of raw petcoke

(Figure 3.10). It is worth noticing that a new significant band, appearing as an intense and broad peak in the range between 3300 and 3600 cm-1 corresponds to -O-H stretching vibration mode of hydroxyl functional groups [84]. This might be explained due to the presence of oxygen in the aqueous phase, thus the organic species of petcoke are oxy-cracked to oxygenated species bearing alcoholic, carboxylic and phenolic functional groups. Interestingly, the presence of carboxylate anion is observed as a doublet band centered at 1580 cm-1, indicating the presence of carboxylic salts. Free acids presence is also evidenced by the C=O band appearing at 1700 cm-1, thus some of the –OH observed in 3300-3600 cm-1 can be assigned to these free acids. Another important feature is the disappearance of most aromatic moieties in the region of out-of-plane bands (930−750 cm-

1), together with the important reduction of the aromatic C-H stretching at 3030 cm-1. Alkyl groups are somehow visible in the range of 3000−2850 cm−1, less contributing to the spectrum in comparison with the original sample and the unreacted solid. Moreover, the presence of esters

106

(∼1,850 cm-1) and aldehyde functions (∼2700 cm-1) are also possible [52, 85]. Carboxyl, esters, aromatic esters and ketones C=O functionalities could appear between 1600 and 1800 cm-1, thus all are feasible and not easily discriminated by the bands within this region of the spectrum. The

C-O-C and/or sulfonic bands (1360-1100 cm-1) in the oxy-cracked products are less intense compared to the original sample, as occurred with the insoluble solid. One of the most important features of the oxy-cracked sample is the broad-band spanning from about 2300-2800 cm-1; this can be a sign of important contribution of -CO3 (carbonates) to the sample which was isolated under basic conditions thus allowing for this possibility.

From the FTIR results, it could be concluded that the oxidized organic functional groups such as

- hydroxyl (-OH), carboxylic salts (O=C–O ), carboxylic acids (R-CO2H) and minor amounts of aldehyde/esters are formed during the oxy-cracking reaction. The functionalities identified by IR spectra of oxy-cracked petcoke are in accordance with the compounds found using XPS and NMR techniques as will be discussed in the next sections.

107

Figure 3.10 FTIR spectra of the original petcoke, oxy-cracked products and residual petcoke at 230 °C and 2 h residence time.

3.4.3.2 1H NMR analysis of the oxy-cracked petcoke

Nuclear magnetic resonance (NMR) spectroscopy is a very useful technique for identifying and analyzing organic compounds, and it is based on the magnetic properties of atomic nuclei. NMR analysis of oxy-cracked product was performed on a Bruker CFI 600 MHz spectrometer by dissolving the sample in deuterated water. The 1H NMR spectrum of the oxy-cracked sample produced at 230 oC and 2 h reaction time is shown in Figure 3.11. The NMR spectrum indicates that the oxy-cracked sample contains a significant quantity of aliphatic groups with chemical shifts in the range of 0–3 ppm. Methylene moieties (1.8 ppm) and methylenes bonded to the aromatic groups (2-2.7 ppm) can be present in the oxy-cracked sample as also confirmed by the FTIR results

[68, 86]. However, terminal methyl groups (at about 0.8 ppm) are not detectable as important signals in the oxy-cracked petcoke.

108

Moreover, the presence of the oxygenated functional groups such as alkoxy groups (probably methoxy, based on the sharp signals determined) are observed in the 3.7-4 ppm region. This is a strong indication, again in agreement with the FTIR and XPS results, that the oxy-cracked products are incorporated with oxygen producing typical oxygenated hydrocarbon compounds like ethers, acids and their salts. On the other hand, aromatic protons span chemical shifts in the range 6–9 ppm. These compounds could be diaromatic carboxylate salts molecules as assigned in the strong signal appearing around 8.5 ppm and methoxy-phenol type molecules (6.5 ppm) as well. The presence of carboxyl groups from carboxylic acids is supported by the small signals appearing around 10 ppm. From the preceding findings, it could be concluded that carboxyl derivatives and oxygenated hydrocarbons produced during oxy-cracking are the most significant fractions solubilized in water. These findings match well with the ones derived from the FTIR spectroscopy and XPS (Section 3.4.3.3) and are in good agreement with the results obtained by Manasrah et al.

[45] and Ashtari et al. [40] for the oxy-cracking of Q-65 and n-C7 asphaltenes, respectively.

109

1 Figure 3.11 H NMR spectra for oxy-cracked petcoke ran with D2O solvent. Signal frequencies for typical chemical structures appended.

3.4.3.3 XPS results of petcoke oxy-cracking

As shown previously in the FTIR analysis, Section 3.4.3.1, the chemical functionalities of petcoke before and after the reaction were identified. Through the XPS analysis, the atomic percentage of the presented elements and group functionalities on the surfaces of original and oxy-cracked products were also determined. Based on the FTIR and the elemental analysis results, the deconvolution of C1s, O1s, N1s and S2p signals along their positions was carried out [67]. Table

3.4 shows the atomic percentages of the main components, types and quantities of functional groups in both samples (i.e., petcoke after and before reaction). It is clear from the results that the original petcoke is mainly composed of carbon (88.75 at%), and a minor amount of some heteroatoms such as oxygen (8.65 at%), nitrogen (1.05 at%) and sulfur (1.55 at%). However, the

110

oxy-cracked sample showed a higher oxygen percentage (67.70 at%) and much lower carbon

(28.60 at%) and sulfur percentage (0.80 at%) compared with the original petcoke sample.

Figure 3.12a and b show the deconvoluted C1s spectra of petcoke before and after oxy-cracking.

The deconvolution of C1s signals was performed through centering the peaks for different functional groups at specific binding energy levels. It is clearly observed that the distribution of carbon species in the original petcoke was dramatically different than the one corresponding to the oxy-cracked sample. The C1s spectrum of original petcoke (Figure 3.12a), contains mainly four bond types (C=C), (C–C), (C–O) and (C=O) set to 283.79 eV, 284.80 eV, 286.34 eV, and 289.21 eV, respectively [67, 87]. The abundance of the 283.79 eV band (C=C) evidences that the petcoke sample contained a high amount of aromatic compounds and lower amount of oxygenated functionalities, as revealed by the FTIR and NMR analyses as well. However, the C1s spectrum of oxy-cracked sample (Figure 3.12b) shows the presence of similar signals as in the original petcoke with completely different intensities. Hence, the signal intensity attributed to the aromatic bonds

(C=C) is much decreased, while the abundance of oxygenated functions (O-C=O) was found very important. Figure 3.13 (oxygenates XPS) confirms the presence of carboxyl functions, as well as new C-OH, formed functionalities. The signal at 530.32 eV in both samples (i.e., original and oxy- cracked) attributed to the oxygen in C-O/O-C=O bonds which is higher by almost three times in oxy-cracked sample compared to the original petcoke. Interestingly enough, a distinctive signal at

532.77 eV in oxy-cracked sample (Figure 3.13b) is observed and attributed to oxygen in alcoholic groups (C-OH). This can be explained by the high degree of oxidation in petcoke during the oxy- cracking reaction. These findings are indeed in a good agreement with the results obtained from the

FTIR and NMR analyses (Figures 3.10 and 3.11).

111

On the other hand, the presence of heteroatoms such as nitrogen and sulfur are evidenced in Figures

3.14 and 3.15, respectively. Figure 3.14a and b show the N1s spectra for both samples. The spectra indicate the presence of pyridines (C-N=C) at 397.89 eV, which are naturally occurring. Similarly, the S2p doublet of petcoke was observed at 163.68 and 164.28 eV (Figure 3.15a). It indicated the presence of sulfur-containing functional groups such as thiophenics, sulphonic species (166.7 eV) and low contribution of sulphates (168 eV). Interestingly, lower contribution from thiophenics was observed in the oxy-cracked sample (Figure 3.15b) which is indicated by the relatively lower intensity of the S2p doublet. However, the sulphate contribution was found to be higher and particularly sulphonic species (166.7 eV) were found much more important in the oxy-cracked sample. Thus, it can be concluded that the sulfur compounds could exist in oxy-cracked sample, however, with a low contribution (25 % reduction) as will be shown in the next section.

Table 3.4 Signal fitting for high-resolution spectra of species in original and oxy-cracked petcoke. Before Reaction After Reaction Atomic Bond Bond Atomic Bond Bond Conc. assignment Conc. Conc. assignment Conc. (%) (%) (%) (%) C=C 70.66 C=C 16.18 C–C/C–H 16.05 C–C/C–H 43.09 C–O 8.91 C–O C1s 88.75 C=O 2.48 28.60 O-C=O 40.5 C-O 8.91 C-O, O=C C=O 2.48 OH C=O 40.5 O1s 8.65 67.70 C-OH C-N=C 1.02 C-N=C 0.05 N1s 1.05 2.90 C-S-C 0.90 C-S-C --- S2p 1.55 S-O 0.80 S-O

112

Figure 3.12 High-resolution XPS spectra of the deconvoluted C1s peak (a) before reaction, (b) after reaction.

Figure 3.13 High-resolution XPS spectra of the deconvoluted O1s peak (a) before reaction, (b) after reaction.

113

Figure 3.14 High-resolution XPS spectra of the deconvoluted N1s peak (a) before reaction, (b) after reaction.

Figure 3.15 High-resolution XPS spectra of the deconvoluted S2p peak (a) before reaction, (b) after reaction.

114

Sulfur and metal analysis

The content of sulfur and metals strongly depend on the nature and the coking process of crude oil, which can be found as organic and inorganic compounds [21, 88]. The sulfur compounds, for example, one of the most significant impurities in petcoke, could be attached to the carbon skeleton as thiophenes or to aromatic naphthenic molecules or between the aromatic sheets [21, 89]. On the other hand, the presence of metals, mainly nickel and vanadium, could occur as metal chelates or porphyrines as in the asphaltenes [90, 91]. However, other metals are not chemically bonded but intercalated in the petcoke structure, as mineral salts normally found as part of ashes [92, 93].

In this set of experiments, the element analysis of petcoke was taken into consideration during the oxy-cracking reaction to investigate the ability of this process to demineralize the petcoke. Table

3.5 shows the elemental analysis for the original petcoke sample (1000 mg), water solubilized products (oxy-cracked sample) and the remaining solids (residue) after the reaction that was carried out at 230 °C for 2 h. Worth to mention that these results were obtained at high reaction conversion, which means that the whole sample of petcoke was targeted to be solubilized in water.

As seen in the table, more carbon, hydrogen and nitrogen can be found in the liquid phase compared with residual solid. These findings indicate that 95% of petcoke being oxy-cracked and solubilized in water as discussed early in Section 3.4.3.1. Moreover, the primary heteroatoms and metals present in the original petcoke sample are sulfur and metals like vanadium, nickel, iron, cobalt and molybdenum. The produced hydrocarbons in the liquid phase were determined to contain some amounts of sulfur and metals. However, more iron, nickel, cobalt and molybdenum content can be observed in the residual solid compared with the liquid phase (oxy-cracked products). Also, more iron, and nickel were found in the residual solid compared with original

115

petcoke, the reason behind that is because of the acidic medium formed during the oxy-cracking reaction which caused corrosion to the reactor wall and impeller, thus metals would be leached out and increase their content in the residue. Interestingly, around 26% of sulfur remained in the residual solids, presumably as highly-fused sulfur aromatic rings and possible coprecipitated sulfates. It can be concluded from these findings that the nonsolublized solids (residue) contain a higher amount of metals compared with the oxy-cracked petcoke (solubilized). These findings suggest that the oxy-cracking process could be a useful technique for petcoke demineralization and desulfurization. Details and further investigation on the operating conditions for the demineralization and desulfurization of petcoke are beyond the scope of this study and will be addressed in a future publication.

Table 3.5 Elemental content in the virgin, oxy-cracked and residue petcoke at temperature 230 oC, pressure 750 psi and time 2 h. Elements Original Residual Liquid Total mass proportion (mg) (mg) (mg) % C 844.750 40.710 747.350 95.89 H 38.100 2.280 33.850 94.82 N 15.500 2.850 6.650 93.54 S 44.600 11.740 27.450 87.87 V 0.785 0.114 0.576 87.89 Ni 0.255 0.097 0.203 117.72 Fe 0.568 1.088 0.018 194.78 Mo 0.012 0.011 0.004 125.00 Co 0.151 0.044 0.018 41.10 Total 944.721 85.890 816.677 95.537

116

3.5 Conclusion

A new approach for petcoke conversion into valuable products was explored in this study by using oxy-cracking reaction which is operating at mild temperature and pressure in an aqueous alkaline medium. The reaction conditions were experimentally investigated in a batch reactor to optimize the highest conversion and selectivity to water-solubilized products with minimal amount of CO2 emission, where the optimal temperature and time were 230 oC and 2 h, respectively. The reaction kinetics was established at the residence times ranged between 0 and 2 h and different reaction temperatures 200, 215, and 230 oC. The kinetics results showed that the petcoke is oxy-cracked simultaneously into water-soluble species and CO2, with the consecutive reaction of soluble species into CO2. The concentration of the oxy-cracked petcoke in the liquid phase was measured as a lumped TOC, while CO2 was determined in gas products at the end of reaction using gas chromatography (GC) and inorganic carbon (IC). The oxygenated hydrocarbons (desired products) and the residual solids were characterized using FTIR, NMR and XPS techniques. The results showed that the main species solubilized in water were oxygenated hydrocarbons compounds and some organic acid such as carboxylic and sulfonic acids and their salts. The residual solids remaining after the reaction showed structures and functional groups similar to the original petcoke. Interestingly, most of the metals contents were found in the residual petcoke compared with the metals in the liquid phase. Therefore, the oxy-cracking technique can be considered as a new feasible process for conversion and demineralization/desulfurization of petroleum coke.

3.6 Acknowledgment

The authors are grateful to the Natural Sciences and Engineering Research Council of Canada

(NSERC), the Department of Chemical and Petroleum Engineering at the Schulich School of

117

Engineering at the University of Calgary. A special acknowledgement to Dr. Maha AbuHafeetha

Nassar for her help in drawing the schematic diagram of the experimental setup. Also, the authors acknowledge Dr. M. Josefina Perez-Zurita and Marianna Trujillo for helping in the XPS and NMR analyses.

3.7 References

[1] J.G. Speight, New approaches to hydroprocessing, Catalysis Today 98(1) (2004) 55-60. [2] C.S. Hsu, P. Robinson, Practical advances in petroleum processing, Springer Science & Business Media 2007. [3] E. Furimsky, Characterization of cokes from fluid/flexi-coking of heavy feeds, Fuel processing technology 67(3) (2000) 205-230. [4] E. Furimsky, Gasification of oil sand coke: Review, Fuel processing technology 56(3) (1998) 263-290. [5] W. Zubot, M.D. MacKinnon, P. Chelme-Ayala, D.W. Smith, M.G. El-Din, Petroleum coke adsorption as a water management option for oil sands process-affected water, Science of the Total Environment 427 (2012) 364-372. [6] M.S. Rana, V. Samano, J. Ancheyta, J. Diaz, A review of recent advances on process technologies for upgrading of heavy oils and residua, Fuel 86(9) (2007) 1216-1231. [7] R.P. Dutta, W.C. McCaffrey, M.R. Gray, K. Muehlenbachs, Thermal cracking of Athabasca bitumen: influence of steam on reaction chemistry, Energy & Fuels 14(3) (2000) 671-676. [8] S.J. Yoon, Y.-C. Choi, S.-H. Lee, J.-G. Lee, Thermogravimetric study of coal and petroleum coke for co-gasification, Korean Journal of Chemical Engineering 24(3) (2007) 512-517. [9] C.E.f.i. Peace, Manging China's petcoke problem, 2015. http://carnegieendowment.org/files/petcoke.pdf. (Accessed October 2016). [10] N.R. Canada, Oil Sands: A strategic resource for Canada, North America and the global market,2013.http://www.nrcan.gc.ca/sites/www.nrcan.gc.ca/files/energy/pdf/eneene/pubpub/pdf/ OS-brochure-eng.pdf. (Accessed 10 March 2017). [11] EPA, U.S. Environmental Protection Agency 2011. http://www.epa.gov/p2/pubs/p2policy/definitions.htm. (Accessed January 2016). [12] M.T. Ityokumbul, Experimental evaluation of molten caustic leaching of an oil sand coke residue, The Canadian Journal of Chemical Engineering 72(2) (1994) 370-374. [13] H. Jang, T.H. Etsell, Morphological and mineralogical characterization of oil sands fly ash, Energy & Fuels 19(5) (2005) 2121-2128. [14] P. Pourrezaei, A. Alpatova, P. Chelme-Ayala, L. Perez-Estrada, M. Jensen-Fontaine, X. Le, M.G. El-Din, Impact of petroleum coke characteristics on the adsorption of the organic fractions from oil sands process-affected water, International Journal of Environmental Science and Technology 11(7) (2014) 2037-2050. [15] J.M. Hill, A. Karimi, M. Malekshahian, Characterization, gasification, activation, and potential uses for the millions of tonnes of petroleum coke produced in Canada each year, The Canadian Journal of Chemical Engineering 92(9) (2014) 1618-1626. 118

[16] P.M. Fedorak, D.L. Coy, Oil sands cokes affect microbial activities, Fuel 85(12) (2006) 1642- 1651. [17] J. Wang, E.J. Anthony, J.C. Abanades, Clean and efficient use of petroleum coke for combustion and power generation, Fuel 83(10) (2004) 1341-1348. [18] P.J. Ellis, C.A. Paul, Tutorial: Petroleum coke calcining and uses of calcined petroleum coke, American Institute of Chemical Engineers 2000. [19] J.A. Caruso, K. Zhang, N.J. Schroeck, B. McCoy, S.P. McElmurry, Petroleum coke in the urban environment: a review of potential health effects, International journal of environmental research and public health 12(6) (2015) 6218-6231. [20] F.J. Farago, D.G. Retallack, R.R. Sood, Calcination of coke, US Patent 4,022,569, 1977. [21] H. Al-Haj-Ibrahim, B.I. Morsi, Desulfurization of petroleum coke: a review, Industrial & engineering chemistry research 31(8) (1992) 1835-1840. [22] L. Edwards, F. Vogt, M. Robinette, R. Love, A. Ross, M. McClung, R. Roush, W. Morgan, Use of shot coke as an anode raw material, Essential Readings in Light Metals: Electrode Technology for Aluminum Production, Volume 4 (2013) 36-41. [23] R. Dunbar, P. Eng, Canada’s Oil Sands-A World-Scale Hydrocarbon Resource, Strategy West Inc, 2007. [24] B. Sarkar, Adsorption of single-ring model naphthenic acid from oil sands tailings pond water using petroleum coke-derived activated carbon, University of Toronto, 2013. [25] E. Hall, E. Tollefson, Z. George, L. Schneider, Upgrading of delayed and fluid cokes from oil sand by desulfurization, The Canadian Journal of Chemical Engineering 60(3) (1982) 418-424. [26] S.H. Lee, C.S. Choi, Chemical activation of high sulfur petroleum cokes by alkali metal compounds, Fuel processing technology 64(1) (2000) 141-153. [27] N. Berguerand, A. Lyngfelt, The use of petroleum coke as fuel in a 10kWth chemical-looping combustor, International Journal of Greenhouse Gas Control 2(2) (2008) 169-179. [28] C. Small, Activation of delayed and fluid petroleum coke for the adsorption and removal of naphthenic acids from oil sands tailings pond water, University of Alberta 2011. [29] X. Li, Q. Zhang, L. Tang, P. Lu, F. Sun, L. Li, Catalytic ozonation of p-chlorobenzoic acid by activated carbon and nickel supported activated carbon prepared from petroleum coke, Journal of Hazardous Materials 163(1) (2009) 115-120. [30] W.A. Zubot, Removal of naphthenic acids from oil sands process water using petroleum coke, University of Alberta, 2011. [31] E.A. Morris, Modification of carbonaceous materials with sulfur and its impact on mercury capture and sorbent regeneration, University of Toronto, 2012. [32] A. Awoyemi, Understanding the adsorption of polycyclic aromatic hydrocarbons from aqueous phase onto activated carbon, University of Toronto, 2011. [33] A. Hethnawi, N.N. Nassar, A.D. Manasrah, G. Vitale, Polyethylenimine-functionalized pyroxene nanoparticles embedded on Diatomite for adsorptive removal of dye from textile wastewater in a fixed-bed column, Chemical Engineering Journal 320 (2017) 389-404. [34] G. Stavropoulos, A. Zabaniotou, Minimizing activated carbons production cost, Fuel processing technology 90(7) (2009) 952-957. [35] T.J. Bandosz, Activated carbon surfaces in environmental remediation, Academic press, New York USA, 2006.

119

[36] N.N. Nassar, A. Hassan, P. Pereira-Almao, Application of nanotechnology for heavy oil upgrading: catalytic steam gasification/cracking of asphaltenes, Energy & Fuels 25(4) (2011) 1566-1570. [37] A. González, N. Moreno, R. Navia, CO2 carbonation under aqueous conditions using petroleum coke combustion fly ash, Chemosphere 117 (2014) 139-143. [38] V. Nemanova, A. Abedini, T. Liliedahl, K. Engvall, Co-gasification of petroleum coke and biomass, Fuel 117 (2014) 870-875. [39] A.I.-E.a.E. Solutions, Executive Summary, 2017. http://eipa.alberta.ca/. (Accessed March 2017). [40] M. Ashtari, L. Carbognani Ortega, F. Lopez-Linares, A. Eldood, P. Pereira-Almao, New Pathways for Asphaltenes Upgrading Using the Oxy-Cracking Process, Energy & Fuels 30(6) (2016) 4596-4608. [41] M. Ashtari, New Pathways for Asphaltenes Upgrading via Oxy-cracking in Liquid Phase, University of Calgary, 2016. [42] L. Smith Moreno-Arciniegas, F.-E. Rodríguez-Corredor, L.-E. Afanador-Rey, J.-L. Grosso- Vargas, Syngas obtainment from the gasification of asphaltenes of the San Fernando crude oil, CT&F-Ciencia, Tecnología y Futuro 3(5) (2009) 189-202. [43] P. Haghighat, L. Carbognani Ortega, P. Pereira-Almao, Experimental Study on Catalytic Hydroprocessing of Solubilized Asphaltene in Water: A Proof of Concept To Upgrade Asphaltene in the Aqueous Phase, Energy & Fuels 30(4) (2016) 2904-2918. [44] P. Haghighat, Processing of Solubilized Asphaltene in Aqueous Media, University of Calgary, 2016. [45] Manasrah, A. D., El-Qanni, A., Badran, I., L. Carbognani, P, Perez-Zurita, M. J., & Nassar, N. N., Experimental and theoretical studies on oxy-cracking of Quinolin-65 as a model molecule for residual feedstocks Reaction Chemistry & Engineering 2 (2017 ) 703-719. [46] S. Moschopedis, J. Speight, Oxidative degradation of Athabasca asphaltenes, Fuel 50(2) (1971) 211-217. [47] V. Platonov, A. Kudrya, S. Proskuryakov, Ozonolysis of asphaltenes from semicoking tar of G17 coal, Russian journal of applied chemistry 76(1) (2003) 148-152. [48] S.E. Moschopedis, Sulfomethylation of humic acids, lignites, and coals and products thereof, US Patent 3,352,902, 1967. [49] S.E. Moschopedis, J.G. Speight, Influence of metal salts on bitumen oxidation, Fuel 57(4) (1978) 235-240. [50] G. Escobar, P. Patiño, S. Acevedo, O. Escobar, M.A. Ranaudo, J.C. Pereira, Interfacial properties of the products of ozonolysis of Hamaca crude oil, Petroleum science and technology 19(1-2) (2001) 107-118. [51] N.N. Marei, N.N. Nassar, M. Hmoudah, A. El‐Qanni, G. Vitale, A. Hassan, Nanosize effects of NiO nanosorbcats on adsorption and catalytic thermo‐oxidative decomposition of vacuum residue asphaltenes, The Canadian Journal of Chemical Engineering (2017). [52] M. Ashtari, L. Carbognani, P. Pereira-Almao, Asphaltenes Aqueous Conversion to Humic and Fulvic Analogs via Oxy-Cracking, Energy & Fuels 30(7) (2016) 5470-5482. [53] P. Haghighat, L. Carbognani Ortega, P. Pereira-Almao, Kinetic Study of Preoxidized Asphaltene Hydroprocessing in Aqueous Phase, Energy & Fuels 30(7) (2016) 5617-5629.

120

[54] L. Li, P. Chen, E.F. Gloyna, Generalized kinetic model for wet oxidation of organic compounds, AIChE journal 37(11) (1991) 1687-1697. [55] H. Estrade-Szwarckopf, XPS photoemission in carbonaceous materials: A “defect” peak beside the graphitic asymmetric peak, Carbon 42(8) (2004) 1713-1721. [56] Q. Zhang, K.T. Chuang, Lumped kinetic model for catalytic wet oxidation of organic compounds in industrial wastewater, AIChE journal 45(1) (1999) 145-150. [57] N.N. Marei, N.N. Nassar, G. Vitale, A. Hassan, M.J.P. Zurita, Effects of the size of NiO nanoparticles on the catalytic oxidation of Quinolin-65 as an asphaltene model compound, Fuel 207 (2017) 423-437. [58] K.S. Seshadri, E.W. Albaugh, J.D. Bacha, Characterization of needle coke feedstocks by magnetic resonance spectroscopy, Fuel 61(4) (1982) 336-340. [59] A. Sarkar, D. Kocaefe, Y. Kocaefe, D. Bhattacharyay, B. Morais, M. Pouliot, Characterization of petroleum coke and butts used in anode manufacturing in aluminum industry, Materials Science and Technology Conference and Exhibition 2013, MS and T 2013, 2014. [60] E.B. Brouwer, I. Moudrakovski, K.H. Chung, G. Pleizier, J.A. Ripmeester, Y. Deslandes, Magnetic resonance microimaging of petroleum coke, Energy & Fuels 13(5) (1999) 1109-1110. [61] J. Rivas, S.T. Kolaczkowski, F.J. Beltran, D.B. McLurgh, Degradation of maleic acid in a wet air oxidation environment in the presence and absence of a platinum catalyst, Applied Catalysis B: Environmental 22(4) (1999) 279-291. [62] K. Merchant, Studies of Heterogeneous Reactions, PhD Thesis, Department of Chemical Technology, University of Bombay, India, 1992. [63] R.V. Shende, J. Levec, Kinetics of wet oxidation of propionic and 3-hydroxypropionic acids, Industrial & engineering chemistry research 38(7) (1999) 2557-2563. [64] N. Rodriguez, H. Marsh, E. Heintz, R. Sherwood, R. Baker, Oxidation studies of various petroleum cokes, Carbon 25(5) (1987) 629-635. [65] I. Badran, N.N. Nassar, N.N. Marei, A. Hassan, Theoretical and thermogravimetric study on the thermo-oxidative decomposition of Quinolin-65 as an asphaltene model molecule, RSC Advances 6(59) (2016) 54418-54430. [66] J. Menéndez, J. Pis, R. Alvarez, C. Barriocanal, E. Fuente, M. Diez, Characterization of petroleum coke as an additive in metallurgical cokemaking. Modification of thermoplastic properties of coal, Energy & Fuels 10(6) (1996) 1262-1268. [67] A. Sarkar, D. Kocaefe, Y. Kocaefe, D. Bhattacharyay, D. Sarkar, B. Morais, Effect of Crystallinity on the Wettability of Petroleum Coke by Coal Tar Pitch, Energy & Fuels 30(4) (2016) 3549-3558. [68] A.T. Castro, NMR and FTIR characterization of petroleum residues: structural parameters and correlations, Journal of the Brazilian Chemical Society 17(6) (2006) 1181-1185. [69] S. Lin, S. Ho, Treatment of high-strength industrial wastewater by wet air oxidation a case study, Waste management 17(1) (1997) 71-78. [70] M. Parvas, M. Haghighi, S. Allahyari, Catalytic wet air oxidation of phenol over ultrasound- assisted synthesized Ni/CeO2–ZrO2 nanocatalyst used in wastewater treatment, Arabian Journal of Chemistry (2014). [71] J. Portela, E. Nebot, E.M. de la Ossa, Kinetic comparison between subcritical and supercritical water oxidation of phenol, Chemical Engineering Journal 81(1) (2001) 287-299.

121

[72] B.A. Abussaud, N. Ulkem, D. Berk, G.J. Kubes, Wet air oxidation of benzene, Industrial & Engineering Chemistry Research 47(13) (2008) 4325-4331. [73] G. Suppes, S. Roy, J. Ruckman, Impact of common salts on oxidation of alcohols and chlorinated hydrocarbons, AIChE journal 47(7) (2001) 1623-1631. [74] J.V. Headley, K.M. Peru, D.W. Mcmartin, M. Winkler, Determination of dissolved naphthenic acids in natural waters by using negative-ion electrospray mass spectrometry, Journal of AOAC International 85(1) (2002) 182-187. [75] P.T. Patil, U. Armbruster, M. Richter, A. Martin, Heterogeneously catalyzed hydroprocessing of organosolv lignin in sub-and supercritical solvents, Energy & Fuels 25(10) (2011) 4713-4722. [76] H. Konnerth, J. Zhang, D. Ma, M.H. Prechtl, N. Yan, Base promoted hydrogenolysis of lignin model compounds and organosolv lignin over metal catalysts in water, Chemical Engineering Science 123 (2015) 155-163. [77] L. Chunlan, X. Shaoping, G. Yixiong, L. Shuqin, L. Changhou, Effect of pre-carbonization of petroleum cokes on chemical activation process with KOH, Carbon 43(11) (2005) 2295-2301. [78] M. Akhter, J. Keifer, A. Chughtai, D. Smith, The absorption band at 1590 cm− 1 in the infrared spectrum of carbons, Carbon 23(5) (1985) 589-591. [79] B.J. Meldrum, C.H. Rochester, In situ infrared study of the surface oxidation of activated carbon in oxygen and carbon dioxide, Journal of the Chemical Society, Faraday Transactions 86(5) (1990) 861-865. [80] A.D. Manasrah, T. Laoui, S.J. Zaidi, M.A. Atieh, Effect of PEG functionalized carbon nanotubes on the enhancement of thermal and physical properties of nanofluids, Experimental Thermal and Fluid Science 84 (2017) 231-241. [81] M. Pruski, B. Gerstein, D. Michel, NMR of petroleum cokes II: Studies by high resolution solid state NMR of 1H and 13C, Carbon 32(1) (1994) 41-49. [82] D. Michel, M. Pruski, B. Gerstein, NMR of petroleum cokes I: relaxation studies and quantitative analysis of hydrogen by magnetic resonance, Carbon 32(1) (1994) 31-40. [83] U. Suriyapraphadilok, Characterization of coal-and petroleum-derived binder pitches and the interaction of pitch/coke mixtures in pre-baked carbon anodes, ProQuest2008. [84] A. Hethnawi, A.D. Manasrah, G. Vitale, N.N. Nassar, Fixed-bed column studies of total organic carbon removal from industrial wastewater by use of diatomite decorated with polyethylenimine-functionalized pyroxene nanoparticles, Journal of Colloid and Interface Science (2017). [85] J. Claine Petersen, A dual, sequential mechanism for the oxidation of petroleum asphalts, Petroleum science and technology 16(9-10) (1998) 1023-1059. [86] D. Grasso, Y.-p. Chin, W.J. Weber, Structural and behavioral characteristics of a commercial humic acid and natural dissolved aquatic organic matter, Chemosphere 21(10-11) (1990) 1181- 1197. [87] X. Huang, D. Kocaefe, Y. Kocaefe, D. Bhattacharyay, Wettability of bio-coke by coal tar pitch for its use in carbon anodes, Colloids and Surfaces A: Physicochemical and Engineering Aspects 490 (2016) 133-144. [88] A. Andrews, R.K. Lattanzio, Petroleum Coke: Industry and Environmental Issues, Congressional Research Service, 2014. [89] R.A. Meyers, Coal desulfurization, Marcel Dekker Inc., New York, 1977.

122

[90] H. Shlewit, M. Alibrahim, Extraction of sulfur and vanadium from petroleum coke by means of salt-roasting treatment, Fuel 85(5) (2006) 878-880. [91] J. Alvarado, M. Alvarez, A.R. Cristiano, L. Marcó, Extraction of vanadium from petroleum coke samples by means of microwave wet acid digestion, Fuel 69(1) (1990) 128-130. [92] P. Agarwal, D. Sharma, Studies on the Desulfurization of Petroleum Coke by Organorefining and Other Chemical and Biochemical Techniques Under Milder Ambient Pressure Conditions, Petroleum science and technology 29(14) (2011) 1482-1493. [93] P. Queneau, R. Hogsett, L. Beckstead, D. Barchers, Processing of petroleum coke for recovery of vanadium and nickel, Hydrometallurgy 22(1-2) (1989) 3-24.

123

Nanocrystalline copper silicate for catalytic oxy-cracking of petroleum coke

Graphical Abstract

O2 CO2

-OH K+

HO. -OH

+ . K O2 HO

ROOH + CO2

HO. O 2 ROOH K+ ROOK

CaCuSi4O10

Highlights

• A nanocrystalline copper-silicate was introduced as a catalyst for petcoke oxy-cracking. • The catalyst activity, selectivity and stability were investigated. • The oxy-cracked products were found to be humic acid analogs. • An insignificant amount of CO2 was released in the gas phase.

This chapter is part of a manuscript entitled " Nanocrystalline copper silicate for catalytic oxy- cracking of petroleum coke." by Manasrah, Abdallah D., Gerardo Vitale, Nashaat N. Nassar. to be submitted to Applied Catalysis B: Environmental

124

4.1 Abstract

Oxy-cracking has been developed recently as an effective technique for converting residual feedstocks, like petroleum coke (petcoke), into valuable commodity products. The method has shown interesting options regarding petcoke solubilization in water at mild temperatures and pressures. This offers a pathway for creating value from solid waste hydrocarbon. Oxidized petcoke becomes soluble in water due to the polar functionalization of aromatic edges and paraffinic terminal carbons via oxygen incorporation. This fact enhances the tendency of petcoke to disaggregate, crack and convert into humic acid analogs, making them a valuable product or more accessible to subsequent hydrogenation and cracking at lower temperatures. For practical purposes, and to favor high conversion and selectivity towards valuable product, energy consumption and capital investment should be minimized. In addition, low selectivity to CO2 emission is required to meet the global environmental regulations. Hence, addition of suitable heterogeneous catalysts to the oxy-cracking process could enhance the process conversion and selectivity. Therefore, in this study, a nanocrystalline copper-silicate (CaCuSi4O10) material was introduced to enhance the selectivity and conversion of the oxy-cracking reaction of petroleum coke. The nanocrystalline material was synthesized in-house and characterized using BET, SEM,

FTIR and XRD techniques. The catalytic activity of the nanocrystalline material was investigated by cracking the residual feedstock (petcoke) in the liquid phase. The results showed that the catalyst enables the reaction to occur at a lower temperature with higher conversion as compared with the non-catalyzed reaction. An insignificant amount of CO2 was formed in the gas and liquid phases at high temperature as confirmed by GC and TOC analyses, respectively. The triangular lump kinetics model was implemented to describe the reaction pathways. The oxy-cracked

125

products were found to be humic acid analogs with different contribution of the functionality groups such as carboxylic, carbonyl, and sulfonic acids as confirmed by FTIR analysis.

4.2 Introduction:

Millions of tonnes of residual feedstocks like petroleum coke, hereinafter referred to as petcoke, are generated annually as a by-product from bitumen and heavy oil upgrading or even from delayed and fluid coking processes in the petroleum refineries [1]. Petcoke is typically stockpiled at the refineries or upgraders because limited markets exist for this commodity. Upgrading and treatment of these residual materials typically include thermal processes like pyrolysis, gasification and combustion [2-5]. However, these conventional processes have some limitations; [6] including its high sulfur and CO2 emissions, high energy consumptions and low conversion efficiency as well as high capital cost [7, 8]. We have recently developed an inexpensive condition for high solubilisation of petcoke in relatively low proportion of water using the so called oxy-cracking method. This method will serve as an alternative technique for converting the residual feedstocks into commodity chemicals [9-11]. The oxy-cracking is a combination of oxidation and cracking reactions in an aqueous alkaline media that occur at mild temperatures (170-230 °C) and pressures

(500-750 psi). The importance of this process is not only the high efficiency, conversion and selectivity but also less amount of CO2 generated during the reaction due to the new reaction pathway offered in an aqueous alkaline medium [10, 11]. Through this reaction, a new reaction pathway is established by partially oxidizing and solubilizing the heavy hydrocarbons in the basic solution. The solubility of oxy-cracked materials is a result of the polar functionality of the aromatic edges and paraffinic terminal carbons that are incorporated with oxygen [10, 11]. Those solubilized compounds were found to be analogs of carboxylic and humic acids and their 126

corresponding salts [10, 12]. However, some hydrocarbons are completely oxidized to CO2 through a parallel consecutive-reaction. The reaction mechanism of such a new technique has been recently developed by our group experimentally and theoretically using Quinolin-65 (Q-65) as a model molecule for the residual feedstocks [11]. Accordingly, the oxy-cracking reaction mechanism is initiated by either the hydroxyl radical (•OH) or the hydroxide anion (−OH) to produce different families of organic acids and a small amount of CO2. Subsequently, the noncatalytic oxy-cracking of petcoke was implemented in another study at mild temperatures (170-

230 °C) and pressures (500-750 psi) [10]. The degree of petcoke conversion through oxy-cracking was highly influenced by reaction temperature, residence time and the complexity of residual feedstocks [9, 10]. Therefore, a high-energy consumption and capital investment may be required for the oxy-cracking operation. Although a relatively small amount of CO2 is released through this process, the proposed technique will help to meet the recent Paris Agreement guidelines [13]. To achieve that goal (i.e., reduction of CO2 emission), heterogeneous catalysts are mandated in the oxy-cracking reactions for high selectivity, activity and stability [14, 15]. Additionally, the selectivity of CO2 emission is acknowledged to be dictated by operational conditions and the catalyst type [16, 17]. Therefore, many researchers have studied the transition metals/oxides and noble metals catalysts in the wet air oxidation [18-21]. However, the metallic oxides, such as CuO,

CoO, MnO, and ZnO, were suffering from deactivation during the oxidation reaction due to the metal leaching caused by the hot acidic media [22-25]. Thus, the activity and stability of these materials for the technological application might be lost under the reaction conditions because of carbon deposition [26-29]. The noble metal catalysts, on the other hand, when supported on alumina, silica, titania and zirconia or carbon-based materials showed more promising catalytic

127

activity than the transition metals [30-32]. However, these noble metals are costly in addition to being sensitive to poisoning and surface deactivation [33-35].

To facilitate the oxy-cracking process, the oxygen transfer from the dissolved phase to the active sites of the catalyst must be rapid in order to create the reaction radicals followed by partial oxidation while preventing or minimizing CO2 formation [10, 11]. To meet these requirements, effective materials in the form of nanoparticles are proposed. Nanoparticles not only possess an active surface area, but also selective surface reactivity, rapid ion delivery, dispersibility, low diffusional resistance and intrinsic reactivity [36, 37]. Active metals are often found to be impregnated on supports, embedded in amorphous or crystalline structures or organized as a core- shell arrangement [38-40]. Occasionally, the metal-based compounds suffer from coke deposition and sintering due to agglomeration of the active sites, and leaching of active metals may occur as a result of the acidic conditions created during the oxidation [41, 42]. Therefore, developing a novel solid catalyst with high activity, selectivity, and stability is a current challenge in liquid- phase oxidation. To this end, the incorporation of transition metal ions into frameworks or cavities of materials like silicates would be a promising and proposed route. One possible strategy to address this approach is to anchor copper ions into the silicate framework by designing a homogeneously distributed molecular structure with isolated active sites over a solid matrix.

Compared to metal and metal oxide catalysts, the naturally occurring silicates have proven to be the most attractive material in terms of adsorption and energy storage processes [43, 44].

Consequently, silicates were confirmed to be promising candidates as catalysts for the oxidation reaction due to its chemical and thermal stability, great abundance, nontoxicity, environment- friendliness, and its unique physical-chemical properties [44-46]. Our research group has recently

128

proven that an iron-silicate pyroxene has a great catalytic activity towards asphaltenes oxidation and cracking [47]. On the other hand, copper is one of the most eminent catalyst in wet air oxidation due to its high activity, diminished electro-migration effect and cost effectiveness [48-

50]. Nickel, for example, offers high catalytic activity in the hydrocarbon oxidation; however, it is prone to coke formation and sintering [51]. In addition, the ability of copper to switch between two oxidation states (1+ and 2+) and the large concentration of vacancies in the solid solution promotes high redox activity, which enables a high tolerance against coking [52-55]. From another viewpoint, a crystalline copper silicate like cuprorivaite (CaCuSi4O10) with a known thermally stable three-dimensional framework may be suitable as a catalyst for the intended application if a nanocrystalline material can be produced. In this way, enough copper should be expose on the surface but within a stable nanocrystalline framework preventing the deactivation of copper active sites at high oxidation temperature without destruction of the nanocrystalline framework. In addition, dispersion of nanocatalysts is crucial in order to minimize mass transfer limitation and catalytic deactivation due to the fact that dispersed nanocatalysts act as quasi-homogeneous catalysts [47, 56, 57].

As far as the authors are aware, this is the first study to examine the application of a nanocrystalline copper-silicate material as a catalyst for oxy-cracking of petcoke in an alkaline medium. Herein, we prepared Cu-silicate (CaCuSi4O10) catalyst in-house using the co-precipitation synthetic route and posterior thermal treatment. The prepared catalyst was characterized before and after oxy- cracking reaction using XRD, SEM, BET, and FTIR. The activity of the catalyst was investigated through the oxy-cracking process which was carried out in a batch reactor under aqueous alkaline medium and mild operating conditions for maximum solubility and selectivity of petcoke.

129

Additionally, zero emission of CO2 is an objective for the proposed oxy-cracking process. The oxy-cracking conversion and selectivity were measured using the total organic carbon analysis

(TOC) while the gas emissions were characterized using gas chromatography (GC). The catalytic oxy-cracking reaction mechanism was developed based on the radical mechanism. The oxy- cracked products were characterized using the FTIR. The present study suggests that using the nanocrystalline copper-silicate materials for the petcoke oxy-cracking could be a promising strategy to develop an efficient catalyst for converting the residual feedstocks into commodity chemicals like humic acid analogs.

4.3 Materials and Methods

Chemicals and reagents

For a typical preparation of the copper-silicate (CaCuSi4O10) nanocrystalline material, the following chemicals and reagents were used: 70 wt% purity nitric acid (HNO3, Sigma Aldrich,

Ontario, Canada); copper(II) acetate (Cu(OOCCH3)2.H2O, Sigma Aldrich, Ontario, Canada); sodium silicate (27 wt.% SiO2, 10.8 wt.% Na2O, Sigma Aldrich, Ontario, Canada), calcium hydroxide (Ca(OH)2, Sigma Aldrich, Ontario, Canada); 99% purity sodium hydroxide (NaOH,

VWR, Ontario, Canada); and sulfuric acid (95–98% purity, Sigma Aldrich, Ontario, Canada) was used for the catalyst regeneration. For the oxy-cracking reaction, green petcoke sample (Marathon

Petroleum Company, Garyville, USA) was grinded and sieved to a particle size of 53 to 710 µm and used as the source of residual feedstocks. Potassium hydroxide (KOH, ACS reagent, ≥85%,

Sigma-Aldrich, Ontario, Canada) was used to adjust the pH of the reaction medium. Ultra-high purity oxygen (99.9%, Praxair, Calgary, Canada) was used as the oxidant gas.

130

In addition, potassium bromide (KBr, Sigma-Aldrich, Ontario, Canada) was used for the infrared analysis. Ultra high purity nitrogen (99.9%, Praxair, Calgary, Canada) was used for the surface area measurements of the prepared material. The carrier gas for the GC was helium (99.9% ultrahigh purity, Praxair, Calgary, Canada). Commercial humic acid (53680 humic acid, Sigma-

Aldrich, Ontario, Canada) was used and characterized for comparison purposes, with oxy-cracked products. All chemicals and reagents were used as received without any further purification.

Synthesis of nanocrystalline copper-silicate material

The copper-silicate (CaCuSi4O10) material was synthesized using a simple co-precipitation method followed by a thermal treatment. An acidic solution was prepared by dissolving 12 ml of nitric acid into 600 ml deionized water with magnetic stirring (300 rpm) followed by the addition of

10.254 g copper(II) acetate. After complete dissolution of the copper in the acid solution, ~ 45.492 g of sodium silicate was carefully added to the solution with agitation for 5 min until a homogenized solution was achieved. Subsequently, a blue gel formed when the pH was increased to 8.0-8.5 by the addition of NaOH pellets under magnetic stirring (300 rpm). The blue gel was allowed to stand for 10 min in order to ensure that pH was stable in the range of 8.0-8.5. The solution was then filtered and washed using copious amounts of de-ionized water under vacuum at room temperature in order to remove excess salts. After thorough washing, the filtered product was allowed to stand at room temperature by passing air through it for ~15 min under vacuum suction. Approximately 3.762 g of calcium hydroxide was added to the wet cake and mixed gently until a homogeneous and pale blue smooth paste was obtained. The pale blue paste was dried in an oven overnight at 200 oC. The dried product was grinded using a mortar and pestle, and calcined in an oven at 850 C for 3 h in a muffle furnace with a heating ramp of 10 C/min. The furnace

131

was then cooled down to room temperature, and the powdered Cu-silicate with nanocrystalline domain sizes was obtained.

Catalyst characterizations

The crystalline phases of the prepared and spent catalysts were characterized using X-ray diffraction (XRD) Ultima III Multi-Purpose Diffraction System (Rigaku Corp., The Woodlands,

TX) with Cu Kα radiation operating at 40 kV and 44 mA. The scan range was 3–90° 2θ using a

0.05° degree and a counting time of 0.2 degree/min. The crystalline domain sizes of the prepared materials were determined using the Scherrer equation as implemented in the PDXL software.

The textural properties and surface areas of the prepared catalyst were measured using the

Brunauer-Emmett-Teller (BET) method. This was accomplished by performing nitrogen physisorption at -196 ºC using TriStar II 3020, Micromeritics Corporate, Norcross, GA. The test sample was previously outgassed at 150 ºC under N2 flow overnight before analysis to remove the moisture. Scanning electron microscopy (SEM) was used to visualize the surface morphology of the prepared materials. A field emission Quanta 250 SEM manufactured by FEI was used, with an accelerating voltage of 20 kV and a spot size of 3.0 to view the morphology of the samples. The tested sample was prepared by taping a very small quantity of the powder over a carbon tape holder and releasing the excess and loose particles. Finally, the molecular bonds in the prepared catalyst were identified using a Shimadzu IRAffinity-1S FTIR (Mandel, USA).

Catalytic oxy-cracking of petcoke sample

Petcoke oxy-cracking experiments were carried out in a 100 ml stainless steel vessel (model number 4598, Parr Instrumental Company, Moline, Il, USA) as shown schematically in Figure 4.1.

The vessel was equipped with a heating oven connected to a temperature control loop, a pressure

132

gauge and a mechanical stirrer with a speed controller. In a typical experiment, 1.0 g of solid petcoke sample and a predetermined amount of catalyst (0.10 g CaCuSi4O10) were charged into the reactor vessel containing 20.0 g of deionized water and 1.0 g of KOH before heating up the reaction vessel to the required temperature. It is noteworthy that a specific amount KOH is required here in order to increase the pH of the solution, and thus, enhancing the solubility of petcoke and avoid the potential corrosion problems. Prior to heating, the reaction vessel was leak tested by sealing and pressurizing the vessel with O2. The reaction solution was then heated to the desired temperature, with the stirring speed set at 1000 rpm. A high mixing speed was used here in order to minimize the interfacial mass transfer resistance between the gas and liquid phase and to ensure uniform temperature and concentration profiles in the liquid phase. Once the set temperature was reached, the reaction time zero is defined. The reaction experiments were carried out at different times (15, 30, 60, 90 and 120 min), temperatures (150-250 oC) and at a constant pressure (750 psi).

Figure 4.1 Schematic illustration of the experimental setup (not to scale) [10]. 133

Characterization of oxy-cracking products and the spent catalyst

At the end of the reaction, the reactor was cooled down and connected to the GC (SRI 8610C,

Torrance, CA) to analyze the released gases. The GC is provided with a thermal conductivity detector (TCD) and two packed columns connected in parallel (3′ molecular sieve/6′ Hayesep-D columns). Afterwards, the liquid phase was carefully withdrawn and filtered for total organic carbon (TOC) analysis using a Shimadzu Total Organic Carbon Analyzer (TOC-L CPH/CPN). All the measurements in TOC and GC were taken respectively three and five times, and the average was used for the calculations with a 5% relative standard deviation.

The oxy-cracked products were recovered by drying in a vacuum oven overnight at 65 oC and characterized using FTIR. A Shimadzu IRAffinity-1S FTIR (Mandel, USA), provided with a smart diffuse reflectance attachment to carry out diffuse reflectance infrared Fourier transform spectroscopy (DRIFTS) analysis, was used. Initially, the background was defined by analyzing

~500 mg of potassium bromide (KBr), then ~ 5 mg of the sample dispersed into the 500 mg of

KBr and analyzed together. The IR spectra were obtained in the wavenumber ranging from 400 to

4000 cm-1, then the spectra were acquired as averages of 50 scans with a resolution of 4 cm-1. A small amount of presumably unreacted solid residue and the used catalyst were collected at the bottom of the reactor vessel. The residual materials (e.g., spent catalyst, minerals and insolubilized petcoke) were recovered and dried using vacuum oven at 65 oC for XRD analysis. Additionally, the metal analysis was performed for the liquid phase to detect any leaching from the catalyst. The leached metal concentrations in the samples (Cu) was analyzed at ALS Environmental

Laboratories (Alberta, Canada) using Inductively Coupled Plasma Mass Spectroscopy (Dissolved

Metals in Water method by CRC ICPMS). The carbon, hydrogen, and nitrogen contents after the

134

oxy-cracking reaction were analysed using a PerkinElmer 2400 CHN analyzer (Waltham,

Massachusetts, USA).

The oxy-cracking reaction conversion was determined from the carbon mass using the following

equation,

C −C Conversion, X = A0 R (4.1) CA0

where CAo is the carbon concentration of virgin petcoke before the reaction and CR is the carbon

concentration in the residual petcoke. Moreover, the selectivity of the desired products (B) and

CO2 (C) were calculated as follow;

(TOC) Selectivity to product B = (4.2) (TOC)+IC+CG

(IC+C ) Selectivity to product C = G (4.3) (TOC)+IC+CG where CG is the carbon content in the gas phase, which was calculated using ideal gas law as follow;

12 PV C = (4.4) G RT

where 푃 and 푇 are the pressure and temperature at the end of reaction, respectively. 푉 is the

volume of the gas phase in the reactor vessel and 푅 is the ideal gas constant.

Stability tests

The stability of the copper-silicate catalyst in the heterogeneous oxy-cracking reaction was

considered in this study. A small amount of the residual material which contains the spent catalyst

and the insolubilized petcoke were collected at the bottom of the reactor vessel. The residual

materials were recovered by filtering the solution after the catalytic tests and drying at 65 C

overnight in a vacuum oven to remove the residual water. The dried sample was then washed with

a diluted sulfuric acid (< 3%) solution to remove unwanted metals such as K, Ni, and Fe that 135

remained in residual materials after reaction, and filtered again. The filtrate sample was calcined at 600 C for 6 h in order to remove any organic species that may have been adsorbed on the material. The spent catalyst was then reused for several cycles of oxy-cracking reaction after further analysis by XRD.

4.4 Results and discussion

Characterization study of the prepared catalyst

The structure of the prepared copper-silicate material has been carefully analyzed by several characterization techniques. The XRD was employed to identify the framework structure. As shown in Figure 4.2, the good intensity of the signals in the XRD patterns implies a well crystallized material. The XRD pattern of this material was matched perfectly with the pdf card

01-085-0158 (cuprorivaite) of the Crystallographic Open Database (COD database) included within the PDXL software (Integrated X-ray powder diffraction software). Additionally, the broad signals clearly indicate the formation of a nanocrystalline copper silicate.

° As displayed in Figure 4.3, the CuCaSi4O10 has a tetragonal crystal structure (훼 = 훽 = 훾 = 90 ) with space group P4/ncc where its lattice constants are: 푎 = 푏 = 7.3017 퐴°, 푐 = 15.1303 퐴° accordingly with a unit cell volume equal to 806.7 퐴°3. It is clear that the metal and ligand oxygen atoms lie in the (001) crystal plane (the XY plane) along with the [00l] Z molecular direction [58].

The Si-centered tetrahedra are parallel to (001) and linked to form two Si8O20 sheets within the height of one cell and each tetrahedron has one unshared corner. The presence of Cu atom sites has a centrosymmetric, planar ligand environment of (D4h) symmetry, which is formed by four oxygens from the unshared corners; such coordination is super-stable and characteristic of divalent

Cu [59]. On the other hand, the Ca atoms are found to be situated in 8-fold coordination midway 136

between sheets [60]. Moreover, the crystalline domain sizes of the prepared material were estimated from the most intense peaks using Scherrer’s equation which is implemented in the

PDXL software where the average domain size was ~93 nm. It is noteworthy that the positions and relative intensities of the diffraction peaks of the synthesized pure-phase sample are in good agreement with previously reported studies on copper silicate, also known as Egyptian blue [61-

64].

Figure 4.2 XRD powder patterns of copper-silicate cuprorivaite (blue line), the vertical lines (black) are the reference data for the cuprorivaite from COD database.

137

o The textural properties of the prepared catalyst were investigated by N2 physisorption at -196 C using the BET analysis. Figure 4.4 shows the nitrogen physisorption isotherms of the CaCuSi4O10 which can be classified as Type II curves based on the IUPAC classification. Worth noting here that the curve clearly indicates the absence of any microporosity in the prepared materials as the isotherm starts from zero without any sudden jump in the Y-axis at p/po  0. Additionally, the estimated specific surface area was 0.63 m2/g by applying the BET method in the range of relative pressures (p/po) between 0.03 and 0.3 and assuming a value of 0.162 nm2 for the cross-section of adsorbed nitrogen molecules at -196 oC.

138

Figure 4.4 Nitrogen physisorption isotherms for copper-silicate. Also, the SEM was used to analyze the surface morphology of the catalyst. Figure 4.5 shows SEM images for the surface of copper-silicates at different magnifications. It clearly suggests that the prepared copper-silicate material has characteristic steps, ridges, and terraces on the surface of the prepared nanocrystalline material. The powders were made up of coarse crystals of cuprorivaite with different shapes and sizes and the size of these particles are in the scale of a few microns (10 and 30 µm). The SEM images revealed that the synthesized CaCuSi4O10 has a nonporous structure with large grains of fused micronic-crystals.

139

Figure 4.5 SEM images of copper-silicate material at different magnifications.

Figure 4.6 shows the infrared spectrum of the prepared copper-silicate material. The IR-spectrum displays characteristic bands lying mainly in the region between 1400 and 400 cm-1 which are attributed to the asymmetric and symmetric stretching vibrations of Si−O−Si and Si−O−Cu, and the bending vibration of O−Si−O and O−Cu−O. Additionally, the silicate band 1085 cm-1 was clearly shifted down which indicates the formation of Si-O-Cu bond and provides evidence for the incorporation of copper metal in the silicate framework structure [65].

As expected, the presence of the water molecules bound to the surface of copper-silicates is presented here which ascribed by the -OH stretching bands at 3637 cm-1 and -OH bending band at

140

1640 cm-1 [65]. However, the small band centered at around 665 cm-1 is related to the bending vibration of -OH that might be located in the tetrahedral position shared by four Cu atoms [66].

Figure 4.6 Infrared spectroscopy of the prepared copper-silicate material.

Catalytic activity and selectivity

The catalytic activity and selectivity for oxy-cracking of petcoke over CaCuSi4O10 catalyst were investigated in a batch reaction. Based on wet air oxidation studies, the reaction rate depends on many factors such as temperature, catalyst loading and solution pH [67, 68]. The results in terms of petcoke oxy-cracking conversion and selectivity to produce both of intermediates (desired products, B) and CO2 (C) are presented in Figures 4.7 and 4.8. The batch reaction experiments were carried out by varying the temperature from 150 to 250 °C while keeping the rest of the reaction conditions constant (oxygen partial pressure 750 psi, stirring speed 1000 rpm, residence 141

time 1 h, and 0.10 g of catalyst). As seen in Figure 4.7, the rate of the oxy-cracking reaction conversion is significantly increased upon raising the reaction temperature. Thus, at 250 oC, ~97% petcoke conversion was reached after 1 h over the CaCuSi4O10 catalyst. This high conversion can be attributed to more free radical species, such as hydroxyl (•OH), hydroperoxyl (•OOH) and alkylperoxy (ROO●), being generated in the solution at elevated temperatures [67]. Interestingly, the reaction conversion was more than 45 % even at low temperatures of 175 oC, which is promising when compared to the non-catalytic oxy-cracking of petcoke as no reaction has been observed to occur at that temperature [10]. Although a high reaction temperature (250 oC) is not favourable in the oxy-cracking process without a catalyst, no remarkable amount of CO2 was observed in the presence of the proposed Cu-silicate catalyst of the present study. Interestingly enough, the selectivity to produce the desired products (B) was 99% even at the low reaction temperature of 150 oC and reaction time of 15 min.

142

Figure 4.7 Effect of the reaction temperature on the selectivity and conversion of petcoke oxy- cracking (P = 750, t = 1 h, 1000 rpm and 0.10 g of catalyst).

The effect of reaction times on catalytic activity and selectivity were also studied by varying the time from 15 to 120 min keeping the temperature at 200 °C as shown in Figure 4.8. As expected, the rate of reaction conversion is significantly increased with time. A full reaction conversion was obtained after 2 h at 185 oC, whereas only 1 h was required at 200 oC and 0.5 h at 250 oC as will be shown in the next section. Interestingly, the selectivity to product B is almost constant and reaching 99% with time. Even at the longest reaction time of 120 min, the amount of produced

CO2 is not detectable by GC. Nevertheless, the produced CO2 may be trapped in the aqueous basic solution (pH >8) in the form of carbonate and bicarbonate. These findings indicate that the Cu- silicates catalyst possesses superior activity and selectivity compared with other catalysts proposed

143

in the literature for wet oxidation reaction such as MnO2/ CeO2, Ru, Pt and Ru/TiO2, Mn−Ce-oxide and Perovskite catalysts LaBO3 (B= Cu, Fe, Mn, Co, Ni) [29, 36, 69, 70].

Figure 4.8 Reaction time effect on selectivity and conversion of petcoke oxy-cracking reaction (T = 200 oC and P = 750 psi, 1000 rpm and 0.10 g of catalyst).

Moreover, the activity of Cu-silicates in the oxy-cracking reaction mostly relates to Cu+2 characteristic in the silicate frameworks. As shown in Figure 4.9, the Cu atom have four coordinated atoms of oxygen as for Si. Thus, the square planar configuration allows d-orbitals to take part in the reaction. Hence, anchoring Cu+2 in the silicate framework using our synthesis method leads to produce a material with nanocrystalline domain size, and thereby increases the number of active sites which have a benefit in activating the petcoke. Additionally, it was previously suggested that the larger the surface area of a catalyst the higher the catalyst activity

144

for the oxidation reaction. However, this was not the case in this study. The activity of oxy- cracking reaction over CaCuSi4O10 was found to be higher than those with high surface area as reported in the literature for wet air oxidation [19, 20, 22, 71]. These findings suggest that the surface area is not only the sole determining factor for the catalytic activity in the oxy-cracking reaction but also the types of active site on the surface of catalyst [72]. Similar findings were reported by Wang et al. [14] for the oxidation of phenol in aqueous phase over functionalized carbon fibre catalyst that exhibited higher activity compared with functionalized graphite, where the latter has higher surface area. Nevertheless, the basicity of the catalyst is another major role to improve the catalytic activity. Therefore, involvement of a calcium ion in the cuproravite structure acts as a basic aid to assist the active site of the catalyst surface by bringing more reactant molecules, thus enhancing the catalyst performance.

Moreover, the influence of pH solution on the conversion and selectivity of oxy-cracking reaction has been previously investigated [10]. It was concluded that the high pH is recommended for the high reaction rate of oxidation. The reasons for that are under alkaline condition (pH >8) the possibility of hydroxyl radical’s formation is increased, and more produced CO2 will be trapped.

Therefore, this further supports the presence of free radicals in the oxidation reactions [67, 73].

Thus, during the oxy-cracking experiment 1.0 g of KOH per 1.0 g of petcoke was considered as the optimum value. These findings also agree well with Nassar et al. [74] that a basic catalyst has lower oxidation temperature for asphaltenes oxidation compared with acidic one. The reaction rates of oxidation and selectivity have been shown to be greatly affected by solution pH. This leads to the conclusion that pH has a significant effect not only on the catalytic activity but also on the

145

stability and leaching of the active phase from the catalyst as will be discussed in the following sections.

Figure 4.9 The square planar configuration of the copper atoms in the structure of CaCuSi4O10, the blue spheres are copper atoms and red ones are oxygen atoms.

Reaction kinetics and mechanism

The catalytic performance of the Cu-silicate material on the petcoke oxy-cracking was studied in the presence of oxygen as an oxidant. The kinetic experimental data was collected at temperatures of 185 ºC, 200 ºC, and 230 ºC and reaction times varying from 0 to 2 h. It was previously concluded that at severe reaction (i.e., temperatures > 250 oC and residence times >2 h), the combustion reaction is more favorable than oxy-cracking and possibility of production of CO2 is significant

[10]. Additionally, the reaction conversions were found to be low at temperatures less than 185 °C

[10]. Moreover, in a typical oxy-cracking reaction, the solubility of oxygen in the aqueous solution is increased with pressure, which favors the oxy-cracking. However, oxygen partial pressure beyond 750 psi did not significantly affect the reaction; and hence was kept constant at that value for the whole sets of experiments. Stirring is very important as well for favoring the interaction

146

between oxygen and petcoke. Here, no mass transfer limitation was observed when the impeller speed operated above 500 rpm. Therefore, the impeller speed was fixed to 1000 rpm during all the reaction runs. The mass ratio of petcoke to KOH was fixed to 1:1, this is where the highest conversion and selectivity were obtained [10]. These reaction conditions (i.e., temperature, oxygen pressure, the amount of KOH and mixing speed) were found to be the best reaction conditions for the production of oxy-cracked desired products [10, 11].

In previous works on the oxy-cracking of petcoke [10], Quinolin-65 (Q-65) [11] and asphaltenes

[9], the kinetic models were determined based on carbon mass as a lumped total organic carbon

(TOC) concentration. This is due to the fact that these heavy solid hydrocarbons have a complex structure, and thus many soluble and insoluble intermediates were produced during the reaction

[75]. On the other hand, the CO2 was found to be the most produced gas in the gas phase as determined by the online GC analyzer. The triangular reaction pathway, as depicted in Figure 4.10, has been recently developed by our research group to describe the mechanism of such reaction using a lumped kinetics model [10, 11]. This reaction pathway was successful in describing the oxy-cracking of heavy hydrocarbons such as asphaltenes [9] and petcoke [10]. The rate of reaction is derived from the lumped concentrations of all organic materials in the water as total organic carbon (TOC).

147

Figure 4.10 Triangular reaction scheme of petcoke oxy-cracking, where A is the petcoke, B is the intermediates (oxy-cracked products, TOC), and C: CO2 in the gas phase (CG) + CO2 in the liquid phase (carbonates IC). The net reaction rates based on the free radical (•OH) mechanism, shown in Figure 4.10, can be described from the elementary reactions and are expressed as follows; dC A = −(K + K )C (4.5) dt 1 2 A dC B = K C − K C (4.6) dt 2 A 3 B dC C = K C + K C (4.7) dt 1 A 3 B where CA, CB, and CC are the carbon concentrations of original petcoke, oxy-cracked products, and

CO2, respectively. K1, K2, and K3 are the reaction rate constants. Based on this model, the reaction orders of petcoke (A) and oxy-cracked products (B) were determined to be first-order while for the oxygen is considered to be zero-order. Excess oxygen is used to reduce its effect on the reaction kinetics and enable the hydrocarbon species to be the limiting reactant [76, 77]. The reaction kinetic parameters of oxy-cracking (i.e., K1, K2, and K3) presented in Equations (4.5-4.7) were estimated by fitting the kinetics experimental data using the Mathematica software (V10.2). The following 148

initial conditions were considered; at the beginning of reaction (t = 0), the carbon concentration in petcoke (CA) is equal to the amount of carbon before reaction (CAo) based on the elemental analysis. However, the concentrations of both oxy-cracked materials (CB) and CO2 (CC) were zero at the beginning of the reaction. The resultant reaction constants of the petcoke oxy-cracking are presented in Table 4.1. Consequently, the dependence of the reaction kinetic constants on the reaction temperature obeys the Arrhenius equation. Thus, the apparent activation energies and frequency factors are summarized in Table 4.2 and estimated as follows:

−Ei ′ RT Ki = kie (4.8)

4.11 represents the Arrhenius plot of petcoke oxy-cracking reaction at the three reaction temperatures. The three curves are approximately linear with the correlation coefficient values close to 1.

149

Figure 4.11 Arrhenius plots of petcoke oxy-cracking for each reaction pathway.

Comparison of our previous kinetic results for oxy-cracking reaction without a catalyst [10] with the ones using the proposed catalyst, suggests that the reaction rate in the second pathway proceeds favorably towards the intermediates which are the desired products (oxy-cracked products). By using our catalyst this rate is much faster than that without catalyst. These findings are also supported by the low value of activation energy in the presence of a catalyst which is 25% less than that in the absence of a catalyst [10]. Interestingly, the reaction rate for forming CO2 in either reaction pathways (1 or 3) with presence of catalyst is lower than that without it. As shown in

Figure 4.12, the concentration profiles of petcoke (A), oxy-cracked compounds (B), and CO2 (C)

150

at the three reaction temperatures as a function of time fit well with the proposed kinetic model. It is wroth noting that the error bars in Figure 4.12 represent the calculated standard deviation based on the TOC and GC measurements. Petcoke was not directly oxidized to CO2 but partially oxidized to intermediates as phenolic and carboxylic substances being produced through a hydroxyl radical

(•OH) attacks. However, insignificant amount of CO2 was noticed at the beginning of the reaction due to the deep oxidation of the short alkyl chains that are left over on the petcoke structures after the coking process. Thus, a low activation energy (E1 = 15.40 kcal/mol) in the first reaction pathway is expected. This insignificant amount of CO2 is related to the shortest induction period which might be required to reach a high enough concentration of catalyst in the liquid phase in order to incorporate oxygen into the hydrocarbon molecules. The oxidation of the hydrocarbons over copper-silicate catalyst demonstrates that a complex reaction took place in the liquid phase which can be attributed to the complexity of petcoke aggregates. In this case, the dissociation of the carbon bonds adjacent to heteroatoms such as sulfur, oxygen and nitrogen took place, respectively [11]. This fact is supported by the low value of the activation energy in the second

4 -1 reaction pathway (E2 = 17.00 kcal/mol) and high value of frequency factor (2.36 × 10 s ). The low activation energy (E2) value supports the conclusion that polymerization reactions might be involved and the formation of (•OH) radicals over the catalyst is the rate limiting step [39].

Interestingly, the apparent activation energies for oxidation of phenolic compounds in water solution over copper oxide have been reported to be in the range of 16.75 to 40.63 kcal/mol in wet air oxidation [70, 78, 79]. Eventually, an insignificant amount of CO2 as carbonates and bicarbonates (pH~8.5-9.8) might be formed in the third reaction pathway due to the further reaction between the solubilized hydrocarbons and oxygen. However, this reaction pathway requires higher

151

activation energy (E3 = 28.10 kcal/mol) for producing the CO2 compared to the first reaction pathway. These findings suggest that the path of conversion of petcoke into CO2 is favored at the beginning of the reaction, which is associated with a higher rate of reaction compared with the oxidation of solubilized organic compounds in water. Moreover, the results also show that even though the reaction rates are increased with temperature, the final TOC values of oxy-cracked compounds (B) for a temperature higher than 200 oC are practically constant after 1 h. The reason for this is presumably due to the ability of these formed short-chain organic species to resist the oxidation process [80]. Another reason for this observation is the short life of free radicals

(scavenging effects) since the presence of strong basic solution KOH would destroy some of the free radicals which would otherwise directly attack the other organic compounds.

Table 4.1 Determined values of oxy-cracking reaction constants.

T K1 K2 K3 (°C) s-1 s-1 s-1

185 -5 -4 -6 8.21 ×10 1.67 ×10 1.63 ×10

200 -4 -4 -6 1.07 ×10 3.85 ×10 9.61 ×10

230 -4 -4 -5 3.56×10 9.25 ×10 2.69 ×10

Table 4.2 Estimated activation energies and frequency factors of petcoke oxy-cracking.

Activation energy (kcal.mol-1) Frequency factor (s-1)

퐸 3 1 15.40 ±0.235 1.65 ×10

퐸 4 2 17.01 ± 0.632 2.36 ×10

퐸 7 3 28.10 ±0.781 4.56 ×10

152

153

Furthermore, the reaction mechanism of the oxidation reaction in the liquid phase over a heterogeneous catalyst are rather complex and not well understood. Many researchers have concluded that the oxidation under subcritical liquid conditions involves chain reactions in which oxygen and hydroxyl, hydroperoxyl and organic hydroperoxy free radicals are participated during the reaction [81, 82]. Accordingly, Accordingly, the mechanism of oxy-cracking reaction over the produced Cu-silicate catalyst is believed to follow the wet air oxidation mechanisms and undergoes several mechanistic steps as presented schematically in Figure 4.13. It is worth mentioning here that the solid petcoke particles float in water as chunks and masses due to the hydrophobicity effects. However, the presence of KOH in the aqueous medium is playing a role in dispersing the petcoke particles through a saponification-like reaction. After the petcoke particles (RH) being dispersed homogeneously in the alkaline medium, and in the presence of oxygen, the oxy-cracking reaction takes place into several steps on the catalyst surface. It is believed that the oxygen molecule will be diffused to the surface of the catalyst suspended in the liquid phase. The role of the catalyst here is to activate the reactants, and thus transferring electron to initiate the free radicals [83]. Subsequently, the adsorbed oxygen will oxy-crack the petcoke at the active site and convert it to oxygenated hydrocarbons that are soluble in water, due to the polar functionalization of aromatic edges and paraffinic terminal carbons via oxygen incorporation. The promotion of free radical reactions involves the catalyst introduction in catalytic cycles through the reduction-oxidation homolytic reactions of Cu(+2)/Cu+1 [84-86]. The free-radical reaction continues while the catalytic donates or accepts electron to form the organic hydroperoxide (ROOH). Then, the intermediates (ROOH) will be degraded to produce the oxygenated hydrocarbons such as peroxy radicals (ROO●), which can later react with the petcoke

154

(RH) and produce more organic hydroperoxides (ROOH) [76]. The later produced compounds may be further oxidized into carbon dioxide, water and other products. It is important to note that the copper in the prepared nanocrystalline CaCuSi4O10 has a buffering capacity for the oxygen on its surface in which the alternation between two oxidation states (Cu+2/Cu+1) and forming the oxygen vacancies can easily occur at the reaction conditions used in this study. These oxygen vacancies have enough potential to transfer more oxygen through the lattice [87]. Eventually, the termination reaction takes place when the generated radicals disappeared and were consumed by reacting with K ions from bicarbonate/carbonate (~110 ppm at 230 oC) that were formed during the reaction or by recombining themselves if they did not attack the organic species [73, 88].

Therefore, at the end of reaction, three phases are obtained. The gas phase remained as predominantly oxygen and an insignificant amount of CO2 while the liquid phase contained the oxygenated hydrocarbons like humic analogs which will be discussed in Section 3.5. Finally, the residue (solid phase) consisted of the minerals and the spent catalyst and some unreacted residue.

More details about the minerals content in the petcoke before and after reaction have been addressed in our previous work [10].

155

O2 CO2

-OH K+

HO. -OH

+ . K O2 HO

ROOH + CO2

HO. O 2 ROOH K+ ROOK

CaCuSi4O10

Figure 4.13 Schematic diagram for the proposed mechanism of petcoke oxy-cracking over the catalyst.

156

Leaching and stability tests of copper silicate

The metal analysis for copper before and after the reaction is presented in Table 4.3. The data is reported in terms of the percentage and concentration of copper leached with respect to the initial amount present in the catalysts (CaCuSi4O10) at various reaction temperatures. Indeed, it can be noticed that the percentage of active metal (Cu) leached to the aqueous solution increased with reaction temperature. The leaching of copper from the catalyst was detected in the range 2 to 3 wt% of the original total amount of Cu, during the oxy-cracking reaction. Therefore, our catalyst

(CaCuSi4O10) seems to be stable enough in the aqueous solutions where the maximum leaching percentage was less than 3% (26 ppm) from the original Cu amount at elevated temperatures.

Hence, the stable structure of Cu-silicates may have inhibited the leaching of Cu+. Generally, the leaching of copper into the aqueous solution is most likely due to the effect of solution pH and the interaction of copper with intermediate organic compounds. It was reported that the leaching of active metal components is one of the main issues causing the catalyst deactivation during catalytic wet air oxidation [71, 89]. Thus, a lot of efforts have been spent to control and overcome the catalyst deactivation by developing an active and stable catalyst [49, 55, 90]. Xu et al. [90], for example, detected the leaching of the copper ~32 ppm under wet oxidation of phenolic compounds over Cu0.5-xFexZn0.5Al2O4. However, a significant reduction in the catalyst activity was observed due to the deposition of carbonaceous materials. Likewise, Gomes et al. [91] reported that the leaching of copper ranged between 21 and 29 ppm for aniline oxidation. In another study, around

19-28% of Cu was leached out during the wet oxidation of ferulic acid using Cu/Mn/AC [92].

Nevertheless, it was demonstrated that the pH of the solution plays a crucial role in the leaching of a catalyst. Hence, a direct relationship between copper leaching and decreasing pH was observed

157

[54]. To overcome the leaching of these active metals, one can add a basic solution such as NaOH,

KOH and NH4OH into the aqueous solution [49, 93, 94]. Consequently, in this present study, an excess amount of KOH (i.e., pH >10) was added to the aqueous solution at the beginning of the reaction. Even at the end of the reaction, the pH was still higher than 8.5 which is due to the formation of acidic functional groups that consumed part of the original KOH. It is believed that the effect of KOH is not only to enhance the solubility of oxy-cracked materials as mentioned earlier in the previous sections but also to maintain the solution in basic condition, thus, preventing the leaching of copper. This finding corroborates with a wet oxidation study of phenol using a

CuO/Al2O3 catalyst in a solution containing bicarbonate buffer (solution pH of 8) [95]. In that study, it was observed that the use of sodium bicarbonate buffer was needed to prevent the leaching of copper and produce less toxic compounds than the phenol such as catechol, hydroquinone, and p-benzoquinone. As reported, in case of the hot-acidic aqueous solution, the leaching of copper from the catalyst surface is increased allowing the reaction of produced acids, especially the carboxylic acid, with copper oxide to give organometallic compounds [38, 55].

Table 4.3 Estimated leached active metal (Cu) from the catalyst at different oxy-cracking reaction temperatures. Experimental conditions: catalyst dose, 0.10 g; reaction time, 1 h.

Active Metal (%) in the Leaching at reaction temperature (oC)

metals prepared catalyst 170 200 230

wt% ppm wt% ppm wt% ppm

Cu 16.90 1.75 15.30 2.45 21.50 2.98 26.14

158

The stability and reusability of any catalyst are very essential for industrial application. To evaluate the long-term stability of our catalyst (CaCuSi4O10), a number of successive cycles were conducted for petcoke oxy-cracking. The catalyst was separated from the reaction mixture at the end of the reaction and washed as described in Section 4.3.6. The activity of the recycled catalyst was determined by carrying out the oxy-cracking at a temperature of 200 oC and residence time of 2 h.

Other operating conditions were kept constants for all experiment runs for testing the recyclability of the spent catalyst. The catalyst activity in terms of reaction conversion and selectivity of both

B and C of the three consecutive experiments is presented in Figure 4.14. As shown, the selectivity to produce B reached nearly 98% after three cycles of reaction except that the third cycle showed an insignificant downward trend as compared with the first run which is seen not to decrease, within the experimental error. Likewise, the selectivity of C is slightly increased with each run, which may due to a part of the framework collapsing during reaction and regeneration; however, the trend is within the experimental error. Interestingly, after three runs, the reaction conversion was still stable and maintained at 92%, 90% and 87% for three consecutive runs, respectively, which supports the successful reusability and stability of the CaCuSi4O10 as a catalyst. These findings are expected because the copper-silicate, which is also known as Egyptian Blue, is a very stable pigment and resistant to most acids and alkaline environment [96, 97]. Additionally, copper- silicate shows resistance to fading even under strong light and can still be observed in Egyptian historical relics which have been exposed outdoor for thousands of years without losing their color

[98-100].

These findings were also confirmed by XRD analysis of the spent catalyst as shown in Figure 4.15.

The figure shows that the XRD patterns following regeneration after the first cycle and the fresh

159

catalyst (cuprorivaite). The main catalyst structure (cuproravite) remained unchanged; however, amorphous material can be observed together with traces of wollastonite. The XRD pattern of the regenerated catalyst matches perfectly with the pdf card no. 01-085-0158 (cuprorivaite) of the

Crystallographic Open Database (COD). However, some small traces of wollastonite (Ca3Si3O9) was observed based on the pdf card no. 1011227. It was reported that the Egyptian Blue pigment consists of CaCuSi4O10 with variable amounts of wollastonite (CaSiO3), high amount of Cu oxides and cuprite (Cu2O) [62]. An important difference in crystalline domain sizes were observed between the fresh (93 nm) and regenerated (44 nm) catalyst which may be due to the disaggregation of some crystallites and/or a new ordering of crystalline was obtained under the reaction conditions. Additionally, the intensities of the crystallographic phase in the regenerated catalyst are lower than for the fresh catalyst. This may be explained by first, a possible contamination with amorphous materials coming from the petcoke; second, amorphous material coming from the partially destruction of the cuproravite structure and third, a mixture of both.

160

Figure 4.14 The conversion and selectivity of B and C for three repeated cycles of Cu-silicates, 2 h, 200 oC.

161

FTIR analysis of the oxy-cracking products

The infrared spectra of the original petcoke and the oxy-cracked compounds (i.e., solubilized fractions) isolated from the reaction that was carried out at 200 °C and 2 h are shown in Figure

4.16. As seen, for the original petcoke spectrum, the alkyls/aliphatic and aromatic regions assigned at (2850-3000 cm-1) and (~3040 cm-1 and 930-742 cm-1), respectively. The aromatic stretching vibration of C=C appears at around 1600 cm-1, which could be conjugated with other groups. The out-of-plane C-H bonds vibration in the aromatic range is assigned at 804, and 860 cm-1 bands.

The C-H stretching vibration due to the aromatic appears at 3040 cm-1. Moreover, the presence of

-1 alkyl groups such as –CH3, =CH2 and –CH2CH3 is evidenced by the bands around 2910 cm and

162

1380 cm-1 which can be assigned to asymmetric and symmetric –C–H stretching and bending vibrations, respectively. The possibility of –OH functionalities (3500 cm-1) is present, which seems to be interacting through hydrogen bonding as the signals are very broad spanning from about

2700 to 2000 cm-1. The presence of heteroatoms such as sulfur in form of sulfoxide species can be assigned to the small band ∼1031 cm-1. It can be concluded that the petcoke has a high contribution of polynuclear aromatic hydrocarbons and few contributions of aliphatic chains with some heteroatoms such as sulfur, nitrogen and oxygen [101, 102]. More detail about the structure and nature characteristics of the used petcoke can be found in our previous work [10].

As expected, the FTIR spectrum of the oxy-cracked products is dramatically different than the original petcoke. The oxy-cracked products obtained at 200 °C were compared with a sample of commercial humic acid as shown in the figure. As seen, the IR spectrum of oxy-cracked products with the presence of the catalyst resembles the one obtained for the commercial humic acids. The broad band spanning from 3700 to 2500 cm-1 indicates the presence of OH groups in both samples

(i.e., oxy-cracked and commercial humic acid). An important feature is the intense and broad peak appearing between 3318 and 3503 cm-1 which correspond to -O-H stretching vibration mode of hydroxyl functional groups. These functionality groups are formed due to the presence of oxygen in the aqueous phase and are related to oxygenated species such as carboxylic functional groups.

The presence of carboxylic acids (C=O) are evident as indicated by the double band centred at

1710 cm-1 in both samples; however, they are more observed in the oxy-cracked products as compared with the commercial humic acid. Interestingly, a complete cracking in the aromatic species is evidenced in the oxy-cracked sample not only by the disappearance of aromatic moieties in the region of out-of-plane bands (930−750 cm-1) but also the reduction of the aromatic C-H

163

stretching at 3030 cm-1. On the other hand, the alkyl groups are no longer visible in the range of

3000−2850 cm−1 in the oxy-cracked sample indicating the complete oxy-cracking of petcoke at that reaction temperature. Moreover, another important feature present in the oxy-cracked sample is the sharp band at ~1842 cm-1 suggesting a possibility of carbonyl compounds such as lactones and esters; however, this band did not appear in the commercial humic acid sample. Calemma et al. [103] reported that the absorption bands between 1600 and 1800 cm−1 in the humic acid obtained by coal oxidation could belong to carbonyl and carboxyl bands. Another important difference between the oxy-cracked sample and the commercial humic acid is the band at 1215 cm-1 corresponding to the presence of sulfur as sulfone compounds, this is due to the high concentration of sulfur in the sample (∼6% sulfur). However, no sulfur compounds were reported in the literature for the humic acids produced from coal oxidation [103, 104].

Based on these results, it can be concluded that the oxy-cracking product characteristics are similar to humic analogs, but with some sulfur content. Therefore, the product consists of primarily oxidized organic functional groups such as hydroxyl (-OH), carboxylic salts (O=C–O-), carboxylic acids (R-CO2H) and minor amounts of esters. These functionalities identified by IR spectra over the catalyst are in accordance with the compounds found in our previous work without a catalyst, although no humic acids analogs were found in the absence of the catalyst. These findings suggest the copper-silicates catalyst is more selective toward producing humic acids analogs under the tested conditions, than that without catalyst.

164

Figure 4.16 FTIR spectra of the virgin petcoke, oxy-cracked products and the humic acid at 200 °C and 2 h residence time.

4.5 Conclusion

A copper-silicate nanocrystalline material was successfully synthesized in-house and successfully tested as a new catalyst for petcoke oxy-cracking reaction. The catalyst activity and selectivity were experimentally investigated in a batch reactor at different reaction conditions. More importantly, a high reaction conversion and high selectivity to water solubilized products with almost zero emission of CO2 were obtained even at high reaction temperatures which meets the

Canadian Environmental Regulation [105]. Furthermore, the triangular lumped kinetics model was successfully employed to describe the oxy-cracking reaction based on the hydroxyl radical

165

mechanism. In this model, the petcoke is oxy-cracked simultaneously into water soluble species and CO2 with the consecutive reaction of soluble species into CO2. Interestingly, the catalyst was found to be stable enough in the aqueous solutions where the leaching percentage was less than 3 wt% of the whole Cu, even at elevated temperatures. After being reused thrice, the CaCuSi4O10 still retained its catalytic activity. The oxy-cracked compounds solubilized in water during the reaction were characterized using FTIR and the main species were carboxylic, carbonyl, phenolic, and sulfonic functions which are responsible for humic acid compounds. The excellent catalytic activity, selectivity, stability and environmental friendly nature of copper-silicates under the mild operating conditions reveal its promise as a high-performance catalyst in the oxy-cracking process.

4.6 Acknowledgments

The authors are grateful to the Natural Sciences and Engineering Research Council of Canada

(NSERC), the Department of Chemical and Petroleum Engineering at the Schulich School of

Engineering at the University of Calgary. A great acknowledgement to Dr. Christophrer Debuhr for providing access to the instrumentation facility for analytical electron microscope in the

Geoscience Department. Thanks to Mr. Lante Carbognani Ortega for helping in the FTIR analysis and to Dr. Amjad El-Qanni for assisting in drawing the reaction mechanism.

4.7 References

[1] B.N. Murthy, A.N. Sawarkar, N.A. Deshmukh, T. Mathew, J.B. Joshi, Petroleum coke gasification: A review, The Canadian Journal of Chemical Engineering 92(3) (2014) 441-468. [2] M.S. Rana, V. Samano, J. Ancheyta, J. Diaz, A review of recent advances on process technologies for upgrading of heavy oils and residua, Fuel 86(9) (2007) 1216-1231. [3] R.P. Dutta, W.C. McCaffrey, M.R. Gray, K. Muehlenbachs, Thermal cracking of Athabasca bitumen: influence of steam on reaction chemistry, Energy & fuels 14(3) (2000) 671-676. [4] E. Hall, E. Tollefson, Z. George, L. Schneider, Upgrading of delayed and fluid cokes from oil sand by desulfurization, The Canadian Journal of Chemical Engineering 60(3) (1982) 418-424.

166

[5] N.N. Nassar, A. Hassan, P. Pereira-Almao, Application of nanotechnology for heavy oil upgrading: Catalytic steam gasification/cracking of asphaltenes, Energy & Fuels 25(4) (2011) 1566-1570. [6] J.-P. Lange, R. Schoonebeek, P. Mercera, F. Van Breukelen, Oxycracking of hydrocarbons: chemistry, technology and economic potential, Applied Catalysis A: General 283(1) (2005) 243- 253. [7] J. Wang, E.J. Anthony, J.C. Abanades, Clean and efficient use of petroleum coke for combustion and power generation, Fuel 83(10) (2004) 1341-1348. [8] L. Smith Moreno-Arciniegas, F.-E. Rodríguez-Corredor, L.-E. Afanador-Rey, J.-L. Grosso- Vargas, Syngas obtainment from the gasification of asphaltenes of the San Fernando crude oil, CT&F-Ciencia, Tecnología y Futuro 3(5) (2009) 189-202. [9] M. Ashtari, L. Carbognani Ortega, F. Lopez-Linares, A. Eldood, P. Pereira-Almao, New pathways for asphaltenes upgrading using the oxy-cracking process, Energy & Fuels 30(6) (2016) 4596-4608. [10] A.D. Manasrah, N.N. Nassar, L.C. Ortega, Conversion of petroleum coke into valuable products using oxy-cracking technique, Fuel 215 (2018) 865-878. [11] A. Manasrah, A. El-Qanni, I. Badran, L.C. Ortega, M.J. Perez-Zurita, N.N. Nassar, Experimental and theoretical studies on oxy-cracking of Quinolin-65 as a model molecule for residual feedstocks, Reaction Chemistry & Engineering 2.5 (2017) 703-719. [12] M. Ashtari, L. Carbognani, P. Pereira-Almao, Asphaltenes aqueous conversion to humic and fulvic analogs via oxy-cracking, Energy & Fuels 30(7) (2016) 5470-5482. [13] J. Rogelj, M. Den Elzen, N. Höhne, T. Fransen, H. Fekete, H. Winkler, R. Schaeffer, F. Sha, K. Riahi, M. Meinshausen, Paris Agreement climate proposals need a boost to keep warming well below 2 C, Nature 534(7609) (2016) 631-639. [14] J. Wang, W. Fu, X. He, S. Yang, W. Zhu, Catalytic wet air oxidation of phenol with functionalized carbon materials as catalysts: Reaction mechanism and pathway, Journal of Environmental Sciences 26(8) (2014) 1741-1749. [15] J.-C. Béziat, M. Besson, P. Gallezot, S. Durécu, Catalytic wet air oxidation on a Ru/TiO2 catalyst in a trickle-bed reactor, Industrial & engineering chemistry research 38(4) (1999) 1310- 1315. [16] A. Luna, L. Rojas, D. Melo, M. Benachour, J. de Sousa, Total catalytic wet oxidation of phenol and its chlorinated derivates with MnO2/CeO2 catalyst in a slurry, Brazilian Journal of Chemical Engineering 26(3) (2009) 493-502. [17] G. Jing, M. Luan, T. Chen, Progress of catalytic wet air oxidation technology, Arabian Journal of Chemistry 9 (2016) S1208-S1213. [18] J. Levec, A. Pintar, Catalytic oxidation of aqueous solutions of organics. An effective method for removal of toxic pollutants from waste waters, Catalysis today 24(1-2) (1995) 51-58. [19] X.C. Song, Y.F. Zheng, H.Y. Yin, Catalytic wet air oxidation of phenol over Co-doped Fe3O4 nanoparticles, Journal of nanoparticle research 15(8) (2013) 1856. [20] H. Wei, X. Yan, S. He, C. Sun, Catalytic wet air oxidation of pentachlorophenol over Ru/ZrO2 and Ru/ZrSiO2 catalysts, Catalysis today 201 (2013) 49-56. [21] A. Santos, P. Yustos, A. Quintanilla, F. Garcıa-Ochoa, Lower toxicity route in catalytic wet oxidation of phenol at basic pH by using bicarbonate media, Applied Catalysis B: Environmental 53(3) (2004) 181-194. 167

[22] J. Levec, A. Pintar, Catalytic wet-air oxidation processes: a review, Catalysis Today 124(3) (2007) 172-184. [23] N. Khanikar, K.G. Bhattacharyya, Cu(II)-kaolinite and Cu(II)-montmorillonite as catalysts for wet oxidative degradation of 2-chlorophenol, 4-chlorophenol and 2, 4-dichlorophenol, Chemical engineering journal 233 (2013) 88-97. [24] M. Parvas, M. Haghighi, S. Allahyari, Degradation of phenol via wet-air oxidation over CuO/CeO2–ZrO2 nanocatalyst synthesized employing ultrasound energy: physicochemical characterization and catalytic performance, Environmental technology 35(9) (2014) 1140-1149. [25] N. Habbache, N. Alane, S. Djerad, L. Tifouti, Leaching of copper oxide with different acid solutions, Chemical Engineering Journal 152(2) (2009) 503-508. [26] Y. Tu, Y. Xiong, S. Tian, L. Kong, C. Descorme, Catalytic wet air oxidation of 2-chlorophenol over sewage sludge-derived carbon-based catalysts, Journal of hazardous materials 276 (2014) 88- 96. [27] C.B. Maugans, A. Akgerman, Catalytic wet oxidation of phenol in a trickle bed reactor over a Pt/TiO2 catalyst, Water research 37(2) (2003) 319-328. [28] Z. Masende, B. Kuster, K. Ptasinski, F. Janssen, J. Katima, J. Schouten, Platinum catalysed wet oxidation of phenol in a stirred slurry reactor: A practical operation window, Applied Catalysis B: Environmental 41(3) (2003) 247-267. [29] S. Hamoudi, A. Sayari, K. Belkacemi, L. Bonneviot, F. Larachi, Catalytic wet oxidation of phenol over Pt x Ag 1−x MnO2/CeO2 catalysts, Catalysis today 62(4) (2000) 379-388. [30] A.B. Ayusheev, O.P. Taran, I.A. Seryak, O.Y. Podyacheva, C. Descorme, M. Besson, L.S. Kibis, A.I. Boronin, A.I. Romanenko, Z.R. Ismagilov, Ruthenium nanoparticles supported on nitrogen-doped carbon nanofibers for the catalytic wet air oxidation of phenol, Applied Catalysis B: Environmental 146 (2014) 177-185. [31] D.K. Lee, J.S. Cho, T. Yu, Y.S. Lee, J.W. Choe, S.S. Lee, Kinetic modeling and dynamic simulation for the catalytic wet air oxidation of aqueous ammonia to molecular nitrogen, Korean Journal of Chemical Engineering 33(11) (2016) 3109-3114. [32] D. Duprez, F. Delanoe, J. Barbier, P. Isnard, G. Blanchard, Catalytic oxidation of organic compounds in aqueous media, Catalysis today 29(1) (1996) 317-322. [33] S. Kolaczkowski, P. Plucinski, F. Beltran, F. Rivas, D. McLurgh, Wet air oxidation: a review of process technologies and aspects in reactor design, Chemical Engineering Journal 73(2) (1999) 143-160. [34] S. Royer, D. Duprez, F. Can, X. Courtois, C. Batiot-Dupeyrat, S. Laassiri, H. Alamdari, Perovskites as substitutes of noble metals for heterogeneous catalysis: dream or reality, Chemical reviews 114(20) (2014) 10292-10368. [35] L. Liotta, Catalytic oxidation of volatile organic compounds on supported noble metals, Applied Catalysis B: Environmental 100(3) (2010) 403-412. [36] E. Szabados, G. Sági, F. Somodi, B. Maróti, D. Srankó, A. Tungler, Wet air oxidation of paracetamol over precious metal/Ti mesh monolith catalyst, Journal of Industrial and Engineering Chemistry 46 (2017) 364-372. [37] A. El-Qanni, N.N. Nassar, G. Vitale, Experimental and computational modeling studies on silica-embedded NiO/MgO nanoparticles for adsorptive removal of organic pollutants from wastewater, RSC Advances 7(23) (2017) 14021-14038.

168

[38] N. Inchaurrondo, P. Massa, R. Fenoglio, J. Font, P. Haure, Efficient catalytic wet peroxide oxidation of phenol at moderate temperature using a high-load supported copper catalyst, Chemical engineering journal 198 (2012) 426-434. [39] O.P. Taran, S.A. Yashnik, A.B. Ayusheev, A.S. Piskun, R.V. Prihod’ko, Z.R. Ismagilov, V.V. Goncharuk, V.N. Parmon, Cu-containing MFI zeolites as catalysts for wet peroxide oxidation of formic acid as model organic contaminant, Applied Catalysis B: Environmental 140 (2013) 506- 515. [40] R.N. Royan, A.B. Sulong, H. Suherman, J. Sahari, Effect of wet oxidation on the dispersion and electrical properties of multi-walled carbon nanotubes/epoxy nanocomposites, Key Engineering Materials, Trans Tech Publ, 2011, pp. 162-166. [41] S.K. Bhargava, J. Tardio, J. Prasad, K. Föger, D.B. Akolekar, S.C. Grocott, Wet oxidation and catalytic wet oxidation, Industrial & engineering chemistry research 45(4) (2006) 1221-1258. [42] S. Roy, A.K. Saroha, Ceria promoted γ-Al2O3 supported platinum catalyst for catalytic wet air oxidation of oxalic acid: kinetics and catalyst deactivation, RSC Advances 4(100) (2014) 56838-56847. [43] X. Wei, C. Tang, X. Wang, L. Zhou, Q. Wei, M. Yan, J. Sheng, P. Hu, B. Wang, L. Mai, Copper silicate hydrate hollow spheres constructed by nanotubes encapsulated in reduced graphene oxide as long-life lithium-ion battery anode, ACS applied materials & interfaces 7(48) (2015) 26572-26578. [44] A. DE ANDRES, Naturally-occurring silicates as carriers for copper catalysts used in methanol conversion, Clays and Clay Minerals 40(2) (1992) 167-174. [45] R.S.d. Cruz, J.M.d.S. Silva, U. Arnold, M.S. Sercheli, U. Schuchardt, Copper containing silicates as catalysts for liquid phase cyclohexane oxidation, Journal of the Brazilian Chemical Society 13(2) (2002) 170-176. [46] R. Jin, Y. Yang, Y. Xing, L. Chen, S. Song, R. Jin, Facile synthesis and properties of hierarchical double-walled copper silicate hollow nanofibers assembled by nanotubes, ACS nano 8(4) (2014) 3664-3670. [47] M. Hmoudah, Nano-Pyroxene for Adsorption and Catalytic Thermal Decomposition of Visbroken Residue Asphaltenes, University of Calgary, 2016. [48] A. Yabuki, S. Tanaka, Oxidation behavior of copper nanoparticles at low temperature, Materials Research Bulletin 46(12) (2011) 2323-2327. [49] A. Santos, P. Yustos, A. Quintanilla, F. Garcia-Ochoa, Kinetic model of wet oxidation of phenol at basic pH using a copper catalyst, Chemical engineering science 60(17) (2005) 4866- 4878. [50] M. Timofeeva, S.T. Khankhasaeva, E. Talsi, V. Panchenko, A. Golovin, E.T. Dashinamzhilova, S. Tsybulya, The effect of Fe/Cu ratio in the synthesis of mixed Fe, Cu, Al- clays used as catalysts in phenol peroxide oxidation, Applied Catalysis B: Environmental 90(3) (2009) 618-627. [51] S. Li, J. Gong, Strategies for improving the performance and stability of Ni-based catalysts for reforming reactions, Chemical Society Reviews 43(21) (2014) 7245-7256. [52] N.L. Pacioni, V. Filippenko, N. Presseau, J. Scaiano, Oxidation of copper nanoparticles in water: mechanistic insights revealed by oxygen uptake and spectroscopic methods, Dalton Transactions 42(16) (2013) 5832-5838.

169

[53] S. Jain, A. Jain, V. Devra, Copper nanoparticles catalyzed oxidation of threonine by peroxomonosulfate, Journal of Saudi Chemical Society (2016). [54] F. Arena, R. Giovenco, T. Torre, A. Venuto, A. Parmaliana, Activity and resistance to leaching of Cu-based catalysts in the wet oxidation of phenol, Applied Catalysis B: Environmental 45(1) (2003) 51-62. [55] A. Santos, P. Yustos, A. Quintanilla, G. Ruiz, F. Garcia-Ochoa, Study of the copper leaching in the wet oxidation of phenol with CuO-based catalysts: causes and effects, Applied Catalysis B: Environmental 61(3) (2005) 323-333. [56] L. Prati, A. Villa, Gold colloids: from quasi-homogeneous to heterogeneous catalytic systems, Accounts of chemical research 47(3) (2013) 855-863. [57] C.-x. Xiao, H.-z. Wang, X.-d. Mu, Y. Kou, Ionic-liquid-like copolymer stabilized nanocatalysts in ionic liquids I. Platinum catalyzed selective hydrogenation of o- chloronitrobenzene, Journal of Catalysis 250(1) (2007) 25-32. [58] R.J. Ford, M.A. Hitchman, Single crystal electronic and EPR spectra of CaCuSi4O10, a synthetic silicate containing copper (II) in a four-coordinate, planar ligand environment, Inorganica Chimica Acta 33 (1979) L167-L170. [59] A. Pabst, Structures of some tetragonal sheet silicates, Acta crystallographica 12(10) (1959) 733-739. [60] Y.-J. Li, S. Ye, C.-H. Wang, X.-M. Wang, Q.-Y. Zhang, Temperature-dependent near- 2+ infrared emission of highly concentrated Cu in CaCuSi4O10 phosphor, Journal of Materials Chemistry C 2(48) (2014) 10395-10402. [61] Q. Jiang, Q. Shi, H. Xu, J. Li, J. Dong, Hydrothermal Synthesis of Pure‐Phase Copper Silicate Na2Cu2Si4O11·2H2O with Ammonia as Complexing Agent, European Journal of Inorganic Chemistry 2011(13) (2011) 2112-2117. [62] A. Panagopoulou, K. Karanasios, G. Xanthopoulou, Ancient Egyptian Blue (CaCuSi4O10) Pigment by Modern Solution Combustion Synthesis Method, Eurasian Chemico-Technological Journal 18(1) (2016) 31-37. [63] D. Johnson-McDaniel, T.T. Salguero, Exfoliation of Egyptian blue and Han blue, two alkali earth copper silicate-based pigments, Journal of visualized experiments: JoVE (86) (2014). [64] T.T. Salguero, D. Johnson-McDaniel, C.A. Barrett, A. Sharafi, R. Weimar, T. Blevins, Nanoscience of Metal Silicate-Based Pigments, MRS Online Proceedings Library Archive 1618 (2014) 161-166. [65] M. Srinivas, P. Srinivasu, S.K. Bhargava, M.L. Kantam, Direct synthesis of two-dimensional mesoporous copper silicate as an efficient catalyst for synthesis of propargylamines, Catalysis today 208 (2013) 66-71. [66] G. Zhan, H.C. Zeng, Topological Transformations of Core–Shell Precursors to Hierarchically Hollow Assemblages of Copper Silicate Nanotubes, ACS applied materials & interfaces 9(42) (2017) 37210-37218. [67] A. Tehrani-Bagha, M. Gharagozlou, F. Emami, Catalytic wet peroxide oxidation of a reactive dye by magnetic copper ferrite nanoparticles, Journal of Environmental Chemical Engineering 4(2) (2016) 1530-1536. [68] A.K. Saroha, Treatment of industrial organic raffinate containing pyridine and its derivatives by coupling of catalytic wet air oxidation and biological processes, Journal of Cleaner Production (2017). 170

[69] R. Levi, M. Milman, M. Landau, A. Brenner, M. Herskowitz, Catalytic wet air oxidation of aniline with nanocasted Mn−Ce-oxide catalyst, Environmental science & technology 42(14) (2008) 5165-5170. [70] O.P. Taran, A.B. Ayusheev, O.L. Ogorodnikova, I.P. Prosvirin, L.A. Isupova, V.N. Parmon, Perovskite-like catalysts LaBO3 (B= Cu, Fe, Mn, Co, Ni) for wet peroxide oxidation of phenol, Applied Catalysis B: Environmental 180 (2016) 86-93. [71] S. Roy, M. Vashishtha, A. Saroha, Catalytic wet air oxidation of oxalic acid using platinum catalysts in bubble column reactor: a review, Journal of Engineering Science & Technology Review 3(1) (2010). [72] Y. Zhan, H. Li, Y. Chen, Copper hydroxyphosphate as catalyst for the wet hydrogen peroxide oxidation of azo dyes, Journal of hazardous materials 180(1) (2010) 481-485. [73] H. Jani, Catalytic wet air oxidation of pulp and paper mills effluent, (2009). [74] N.N. Nassar, A. Hassan, P. Pereira-Almao, Effect of surface acidity and basicity of aluminas on asphaltene adsorption and oxidation, Journal of colloid and interface science 360(1) (2011) 233-238. [75] K.S. Seshadri, E.W. Albaugh, J.D. Bacha, Characterization of needle coke feedstocks by magnetic resonance spectroscopy, Fuel 61(4) (1982) 336-340. [76] L. Li, P. Chen, E.F. Gloyna, Generalized kinetic model for wet oxidation of organic compounds, AIChE Journal 37(11) (1991) 1687-1697. [77] J. Rivas, S.T. Kolaczkowski, F.J. Beltran, D.B. McLurgh, Degradation of maleic acid in a wet air oxidation environment in the presence and absence of a platinum catalyst, Applied Catalysis B: Environmental 22(4) (1999) 279-291. [78] J.R. Katzer, Catalytic oxidation of phenol in aqueous solution over copper oxide, Industrial & Engineering Chemistry Fundamentals 17(3) (1978) 234-234. [79] R.R. Zapico, P. Marín, F.V. Díez, S. Ordóñez, Influence of operation conditions on the copper-catalysed homogeneous wet oxidation of phenol: development of a kinetic model, Chemical Engineering Journal 270 (2015) 122-132. [80] S. Lin, S. Ho, C. Wu, Kinetic and performance characteristics of wet air oxidation of high- concentration wastewater, Industrial & engineering chemistry research 35(1) (1996) 307-314. [81] H. Debellefontaine, M. Chakchouk, J. Foussard, D. Tissot, P. Striolo, Treatment of organic aqueous wastes: wet air oxidation and wet peroxide oxidation, Environmental pollution 92(2) (1996) 155-164. [82] M. Kurian, V. Remya, C. Kunjachan, Catalytic wet oxidation of chlorinated organics at mild conditions over iron doped nanoceria, Catalysis Communications (2017). [83] R.V. Shende, V.V. Mahajani, Kinetics of wet air oxidation of glyoxalic acid and oxalic acid, Industrial & engineering chemistry research 33(12) (1994) 3125-3130. [84] J.L. Urrea, S. Collado, A. Laca, M. Díaz, Wet oxidation of activated sludge: Transformations and mechanisms, Journal of environmental management 146 (2014) 251-259. [85] S. Kumar, C. Sharma, NiO–CeO2 nano-catalysts: Synthesis, characterization and application in catalytic wet air oxidation of wastewater, Materials Express 5(5) (2015) 419-428. [86] J. Mikulová, J. Barbier, S. Rossignol, D. Mesnard, D. Duprez, C. Kappenstein, Wet air oxidation of acetic acid over platinum catalysts supported on cerium-based materials: Influence of metal and oxide crystallite size, Journal of catalysis 251(1) (2007) 172-181.

171

[87] C. Zhang, A. Michaelides, D.A. King, S.J. Jenkins, Oxygen vacancy clusters on ceria: decisive role of cerium f electrons, Physical Review B 79(7) (2009) 075433. [88] M. Ingale, J. Joshi, V. Mahajani, M. Gada, Waste treatment of an aqueous waste stream from a cyclohexane oxidation unit: a case study, Process safety and environmental protection 74(4) (1996) 265-272. [89] S. Keav, A. Martin, J. Barbier, D. Duprez, Deactivation and reactivation of noble metal catalysts tested in the catalytic wet air oxidation of phenol, Catalysis Today 151(1) (2010) 143- 147. [90] A. Xu, M. Yang, H. Du, C. Sun, Influence of partial replacement of Cu by Fe on the CWO of phenol in the Cu0.5− x Fex Zn0.5 Al2O4 spinel catalysts, Catalysis Communications 7(8) (2006) 513- 517. [91] H. Gomes, P. Selvam, S. Dapurkar, J. Figueiredo, J. Faria, Transition metal (Cu, Cr, and V) modified MCM-41 for the catalytic wet air oxidation of aniline, Microporous and mesoporous materials 86(1) (2005) 287-294. [92] B.R. Yadav, A. Garg, Catalytic wet oxidation of ferulic acid (a lignin model compound) in the presence of non-noble metal based catalysts at mild conditions, Chemical Engineering Journal 252 (2014) 185-193. [93] Y. Pi, M. Ernst, J.-C. Schrotter, Effect of phosphate buffer upon CuO/Al2O3 and Cu (II) catalyzed ozonation of oxalic acid solution, Ozone Science & Engineering 25(5) (2003) 393-397. [94] A. Xu, M. Yang, R. Qiao, H. Du, C. Sun, Activity and leaching features of zinc–aluminum ferrites in catalytic wet oxidation of phenol, Journal of hazardous materials 147(1) (2007) 449- 456. [95] A. Santos, P. Yustos, A. Quintanilla, F. Garcia-Ochoa, Influence of pH on the wet oxidation of phenol with copper catalyst, Topics in catalysis 33(1) (2005) 181-192. [96] G. Accorsi, G. Verri, M. Bolognesi, N. Armaroli, C. Clementi, C. Miliani, A. Romani, The exceptional near-infrared luminescence properties of cuprorivaite (Egyptian blue), Chemical Communications (23) (2009) 3392-3394. [97] J. Riederer, Egyptian blue, Artists’ pigments: a handbook of their history and characteristics 3 (1997) 23-46. [98] D.A. Scott, M. Dennis, N. Khandekar, J. Keeney, D. Carson, L.S. Dodd, An Egyptian cartonnage of the Graeco-Roman period, Studies in Conservation 48(1) (2003) 41-56. [99] M. Ibrahim Abadir, Experimental Study of the Egyptian blue degradation by the copper chloride cancer, Universitat Politècnica de Catalunya, 2014. [100] V. Daniels, R. Stacey, A. Middleton, The blackening of paint containing Egyptian blue, Studies in conservation 49(4) (2004) 217-230. [101] A. Sarkar, D. Kocaefe, Y. Kocaefe, D. Bhattacharyay, D. Sarkar, B. Morais, Effect of crystallinity on the wettability of petroleum coke by coal tar pitch, Energy & Fuels 30(4) (2016) 3549-3558. [102] U. Suriyapraphadilok, Characterization of coal-and petroleum-derived binder pitches and the interaction of pitch/coke mixtures in pre-baked carbon anodes, ProQuest2008. [103] V. Calemma, P. Iwanski, R. Rausa, E. Girardi, Changes in coal structure accompanying the formation of regenerated humic acids during air oxidation, Fuel 73(5) (1994) 700-707.

172

[104] L. Madronová, J. Kozler, J. Čežı́ková, J.r. Novák, P. Janoš, Humic acids from coal of the North-Bohemia coal field: III. Metal-binding properties of humic acids—measurements in a column arrangement, Reactive and Functional Polymers 47(2) (2001) 119-123. [105] E.a.C.C. Canada, Regulations to Reduce Carbon Dioxide Emissions from Coal-fired Electricity, 2015. https://www.ec.gc.ca/lcpe-cepa/eng/regulations/detailReg.cfm?intReg=209. December, 2017.

173

A comparative study of thermal properties and heating values of oxy-

cracked and virgin petroleum coke

Graphical Abstract

Highlights

• The oxy-cracked petcoke was proposed as an alternative clean fuel for power generation. • The thermo-oxidative behaviour of the oxy-cracked and virgin petcoke was investigated. • The heating values of virgin and oxy-cracked petcoke were calculated. • The combustion reaction characteristics were also investigated.

This chapter is part of a manuscript entitled "A comparative study of thermal properties and heating values of oxy-cracked and virgin petroleum coke." by Manasrah, Abdallah D., Azfar Hassan, Nashaat N. Nassar. to be submitted to Fuel

174

5.1 Abstract:

Petroleum coke (petcoke) is a challenging fuel in terms of its complicity, high sulfur, nitrogen and low volatile content, and a source of undesirable emission when used for energy generation.

Recently, we envisaged a new method to convert the petcoke into a clean combustion fuel by reducing the sulfur and nitrogen contents using the oxy-cracking technique and consequently increase the reactivity and combustibility of oxygenated petcoke. This study continues our previous work and aims at comparing the heating value and thermos-oxidative behaviour of the oxy-cracked and virgin petcoke using thermogravimetric analysis (TGA). For this purpose, a sample of petcoke was oxy-cracked at 200 oC and 750 psi in a Parr reactor following our previous work. TGA analysis showed that the oxy-cracked petcoke is easier and faster to oxidize as compared to the virgin petcoke. We found out that there was also a significant improvement in the combustion performance parameters of the oxy-cracked petcoke such as ignition, peak and burnout temperatures. Moreover, the heating value of oxy-cracked petcoke is relatively similar to one of the virgin petcoke where the nitrogen and sulfur content in the oxy-cracked petcoke is much lower than that of virgin petcoke.

175

5.2 Introduction

Petroleum coke (petcoke) is a carbonaceous solid consisting of polycyclic aromatic hydrocarbons with low hydrogen content, derived from the processing of oil sands and other oil industries

(delayed coking in the refineries) [1, 2]. The production of petcoke has soared significantly over the years due to constant increase in energy demand [3, 4]. This petcoke production is expected to grow continuously with increasing supply of heavy crude oil around the world [5, 6]. Due to the high availability of petcoke, it sells as a substitute fuel for low-rank coals for power generation in cement, iron and steel industries [7-9]. This is because of the low-price, high availability, low ash content (0.1–0.3 wt%) and high heating value of the petcoke [9, 10]. The heating value of petcoke is reported in the range of 30–35 MJ/kg, which is higher than that of coal by 30% [11]. However, petcoke usually contains appreciable amounts of sulfur and metallic impurities such as vanadium, nickel, and iron [12]. The sulfur and metal content strongly depend on the nature and the coking processes of crude oil, which can be found as organic and inorganic compounds [13]. The sulfur compounds, for example, one of the most significant impurities, could be attached to the carbon skeleton as thiophenes or to side chains of aromatic naphthenic molecules or between the aromatic sheets [14]. On the other hand, the presence of metals, mainly nickel and vanadium, could occur as metal chelates or porphynines as in the asphaltenes [15]. Other metals are not chemically bonded but intercalated in the petcoke structure, as mineral salts normally found as part of ashes [16].

Nonetheless, a number of thermal power plants around the world are using petcoke as a fuel, especially with a high-sulfur content, to produce steam for generating the electricity and heating the boilers [17, 18]. As a consequence, burning virgin petcoke (without further modification) poses a number of environmental problems such as high SO2 and NOX emissions as well as issues of

176

corrosion due to the presence of considerable amounts of sulfur, vanadium and nickel [19]. In addition to high SO2 and NOx emissions, production of CO2 while burning petcoke is a major environmental concern [20]. Also, the virgin petcoke is difficult to ignite due to its low volatile matter content [21]. Therefore, petcoke is usually blended in different ratios with other fuels, such as coal, before burning in boilers to meet emission compliance, simplified ignition and increased flame stability [22, 23].

To meet future environmental regulations, many researchers are trying to blend petcoke with coal in different proportions or trapping and recovering the CO2 during the burning processes [2, 24].

The cement industry, for example, consumes a large portion of fuel-grade petcoke to combust in kilns where up to 50% of petcoke can be added to the fuel mixture. This blending has to be carefully considered due to the effects of high sulfur and vanadium content in the concrete quality.

High sulfur content can crack the cement and cause plugging or fouling of the preheater, while the high vanadium content (above 500 ppm) can weaken the cement strength [25, 26]. Additionally, the presence of vanadium in petcoke during oxidation, produces vanadium oxide (V2O5), that causes slagging in the boiler pipes [27, 28]. Generally, fuel-grade coke is hard to grind (high

Hardgrove Grindability Index, HGI), thus it usually undergoes the physical change in size (<75

µm) to enhance combustion efficiency [22, 25]. Typically, the pulverization of coke to fine powder increases its surface area and thereby increases its combustibility by decreasing the combustion time of each particle [29]. The behavior of petcoke in any combustion process is difficult to predict.

This is because the physical characteristics of petcoke mainly depends on the coking conditions and particularly on the composition of feedstock from where it is derived [30, 31].

177

Taking into consideration the aforementioned issues for petcoke, herein we envisage a new route to process this material for its use in energy generation. Oxy-cracking of petcoke via sequential oxidation and cracking reactions under alkaline conditions might be a promising technique to generate a new type of material that can be used as a fuel, with low sulfur, nitrogen and metal contents. The oxy-cracking process was recently reported as an alternative-practical approach for partially oxidizing and solubilizing the petcoke into alkaline media with minimal emission of CO2 at mild reaction conditions (170-230 °C and 750 psi) [32-34]. Oxy-cracking technique revealed to be an effective process for converting the petcoke into new products via oxygen incorporation using the free radical mechanism. These oxy-cracked materials were found to contain lower molecular weight compounds that we believe are easier to oxidize [32]. Additionally, it has been suggested in our previous work that the oxy-cracking process might be an effective route for demineralization and desulfurization of feedstocks [32, 34]. The solubilized petcoke obtained by the oxy-cracking contained low metals and sulfur contents compared with virgin petcoke as reported in our previous work [32].

Furthermore, it is important in any thermal power plant to determine the calorific value of the coke/coal to assess its quality when used as a fuel. A number of studies have been performed on the measurement of the calorific value of fuel [35, 36]. One of the common methods is using the adiabatic bomb calorimeter. However, this method of analysis is cost intensive and requires a sophisticated equipment. On the other hand, the calorific value of petcoke was found to be significantly affected not only by the volatile matter (VM), moisture (M) and ash (A) contents of petcoke but more importantly on the fixed carbon (FC) and hydrogen content of the material [11].

It is easier and more cost effective to determine the calorific value based on the proximate analysis

178

using TGA as compared with Bomb calorimeter [37]. Many correlations were originally proposed

to predict the heating value of for coal that can be also applied for the petcoke [11, 38, 39]. This is

because of the similarity of chemical compositions between the petcoke and coal.

Herein, the main purpose of this study is to explore the thermal behavior of these oxy-cracked

products and compare them with the virgin material. This approach has the potential to open the

doors to the use of the oxy-cracked petcoke as an environmentally friendly and cost-effective fuel

for clean energy generation, in comparison with virgin petcoke or coal. Therefore, preliminary

oxidation experiments were carried out on the virgin and oxy-cracked petcoke using

thermogravimetric analysis (TGA) to investigate the thermal and calorific properties of these

materials. Moreover, the oxidation reaction of any fuel using TGA will reflect upon the

combustion reaction characteristics. Hence, the advantage of TGA analysis is that it gives a quick

assessment of fuel values, such as the ignition (푇퐼퐺), peak (푇푚) and burnout (푇퐵) temperatures. To

the best of our knowledge, this is the first attempt that has been made in this work to evaluate the

heating value of virgin and oxy-cracked petcoke using different correlations.

5.3 Experimental work

Materials

A sample of green petcoke obtained from Marathon Petroleum Company (Garyville, USA) was grinded and sieved to the particle size ranging between 710 and 53 µm. The chemical composition of the considered virgin petcoke and the oxy-cracked petcoke samples are summarized in Table 5.1, that was also reported in our previous work [32]. A sample of oxy-cracked petcoke, which is mainly the solubilized organic species, was collected from the alkaline solution and solidified after washing

179

with acid and drying in a vacuum oven overnight at 65 °C. HCl (37%, ACS reagent, Sigma Aldrich,

Ontario, Canada) was used for washing the oxy-cracked sample.

Table 5.1 The chemical composition of the virgin and oxy-cracked petcoke sample.

Composition C H N S V Fe Ni Mo Co K O* (wt%) Virgin 84.48 3.81 1.55 4.46 0.08 0.06 0.03 0.01 0.15 0.00 5.37 petcoke Oxy-cracked 62.31 2.68 1.10 1.32 0.04 0.01 0.01 0.00 0.00 3.54 28.99 petcoke * Estimated by the difference. Preparation of oxy-cracking petcoke

A sample of oxy-cracked petcoke was obtained after oxy-cracking the virgin petcoke following the procedure developed in our previous study [32]. Figure 5.1 shows a schematic representation for performing oxy-cracked petcoke. In brief, the reaction was carried out in a PARR batch reactor

(model number 4598, Parr Instrumental Company, Moline, Il, USA) by mixing 1.0 g of original petcoke with 20 ml of deionized water under alkaline conditions (pH~13). The high value of pH is needed for helping in solubilizing petcoke and avoiding corrosion problems within the reactor.

During the reaction, the petcoke sample was oxy-cracked at 200 oC, with oxygen pressure of 750 psi, for 120 min. The mixer speed was set at 1000 rpm, to avoid the mass transfer limitations. At the end of the reaction, the liquid effluent was carefully discharged and filtered in a centrifuge

(Eppendorf centrifuge 5804) at 5000 rpm for 15 min to separate the remaining solid (i.e., unreacted and/or insolubilized species). Worth mentioning here that the pH of the obtained liquid solution after reaction ranged between 8 and 10 based on the reaction conditions. The pH was measured using a Mettler Toledo pH meter (Mississauga, Canada). Afterwards, few drops of HCl (37%) was added to the black liquid solution (solubilized petcoke in water) until the pH of the solution 180

decreased to ~6. Consequently, the solid particles were allowed to settle for 3 h which were then separated by centrifuging and decanting the supernatant solution. It was found that the settled black solid contains most of the organic products based on TOC measurements (~90% of total TOC).

However, small amount of hydrocarbons (<10% of total TOC) remained soluble in the supernatant.

These remained hydrocarbons could be recovered by allowing the solution to settle for 48 h. Then, these settled-solid hydrocarbons were centrifuged and washed twice with 5% HCl solution to remove the excess K left over after reaction. Eventually, the collected solidified organic species were dried in a vacuum oven overnight at 65 oC before TGA analysis.

Figure 5.1 Schematic diagram for the preparation of oxy-cracked petcoke sample.

Thermogravimetric analysis

The virgin petcoke and the washed oxy-cracked petcoke were subjected to thermal oxidation using a thermogravimetric TGA/DSC analyzer (SDT Q600, TA Instruments, Inc., New Castle, DE). As for the oxidation study, samples of ~5 mg of both materials were heated up from room temperature to 800 oC with a heating rate of 10 oC/min under the air flow of 100 cm3/ min. The TGA results, weight loss (TG) and weight loss rate (DTG) profiles, were analyzed to determine the combustion

181

performance parameters (i.e., ignition, peak, and burnout temperatures). These parameters significantly depend on the chemical composition of the fuel and can be calculated by the intersection method [40]. The ignition temperature is calculated at the point where the TG peak, which is the point of initial devolatilization after the sample was dried, and the tangent line of the mass loss profile are intersected. The peak temperature is determined at the maximum DTG peak.

Eventually, the burnout temperature is calculated at the intersection point between the two tangent lines; the first line is tangent to the mass loss profile at the point where the DTG peak occurs and the second line is tangent to the point where the weight loss is unchanged. It is also approximated by the temperature where weight loss of the sample reaches to ~1%/min at the terminal phase of the DTG profile [41].

Heating value measurements

The high heating value (HHV) of virgin and oxy-cracked petcoke was determined by proximate

analysis using TGA. The moisture (M) content and volatile matter (VM) were estimated by heating

up the sample from room temperature to 500 oC under nitrogen atmosphere flowing at a rate of

100 ml/min. The fixed carbon (FC) and the ash (A) content (residue) were obtained by continuing

heating the sample from 500 to 800 oC at a heating rate of 10 ∘C/min under air flow, passing at a

flow rate of 100 ml/min over the sample. The change in the sample weight was monitored until

there was no further change in weight. After estimating the values for of the aforementioned

properties (i.e., M, VM, FC, and A), the heating value was calculated using differently proposed

correlations. Schuster [42], for example, proposed a correlation for the coal-based material on the

volatile matter (VM) only. This formula is limited for the ash free sample and can be expressed

as;

182

−3 퐻퐻푉 = 4.183 × 10 × (800 + 푉푚 × (70 − 1.65 × 푉푚)) (5.1)

However, a correlation for estimating the calorific values based on the ash content as well as the volatile matter (VM) was derived by Kucukbayrak et al. [43] for lignites. In this equation (Eq. 5.2) the polynomial function of (VM) and ash (A) was assumed to estimate the HHV in MJ/kg, given as follows;

−3 2 퐻퐻푉 = 76.56 − 1.3(푉푚 + 퐴) + 7.03 × 10 (푉푚 + 퐴) (5.2)

Another correlation for calculating of HHV (KJ/kg) was proposed by Cordero et al. [44] based on proximate analysis of charcoals and lignocellulosics. This equation is expressed mainly on both

VM and FC content in the dry basis.

퐻퐻푉 = 354.3퐹푐 + 170.8푉푚 (5.3)

Likewise, Parikh et al. [45] considered the volatile matter, fixed carbon, and ash content for determining the HHV (MJ/kg) for coals. This equation most likely fits well for the fuel with high fixed carbon contents as it is the most dominating factor and expressed as follow;

퐻퐻푉 = 0.3536퐹푐 + 0.1559푉푚 − 0.0078퐴 (5.4)

Recently, Majumder et al. [38] developed a new correlation based on proximate analysis data for predicting HHV of coal. The proposed model involves the effects of all the major variables affecting the HHV such as the ash, moisture, volatile matter and fixed carbon as follow.

퐻퐻푉 = 0.35퐹푐 + 0.33푉푚 − 0.11푀 − 0.03퐴 (5.5)

Elemental analysis

A PerkinElmer 2400 CHN analyzer (Waltham, Massachusetts, USA) was used for analyzing carbon, hydrogen, and nitrogen contents for virgin and oxy-cracked petcoke samples using combustion method. The sulfur content was determined with an Antek 9000 system (Houston, TX, 183

USA) calibrated with Accustandard SCO-500x (S) standards and running toluene solutions (10 wt.

%/vol.). The metal contents in the virgin and oxy-cracked petcoke samples were analyzed at ALS

Environmental Laboratories (Alberta, Canada) using Inductively Coupled Plasma Mass

Spectroscopy (Dissolved Metals in Water method by CRC ICPMS).

5.4 Results and discussion

Thermo-oxidative decomposition of virgin and oxy-cracked petcoke

Thermo-oxidative decomposition of virgin and oxy-cracked petcoke was performed to evaluate the thermal degradation behavior under air. Figures 5.2 and 5.3 show the rate of mass loss (TG) and the derivative of rate of mass loss (DTG) profiles under oxidation by air from room temperature to 800 oC at a heating rate of 10 oC/min for the virgin and oxy-cracked petcoke, respectively. It is clear from the profiles (Figure 5.2) that the oxidation of petcoke sample occurs at a temperature around 540 oC which is evidenced by the presence of an exothermic symmetric peak beyond 540 oC as shown in Figure 5.4. Worth noting that there is an initial increase in mass loss for the virgin petcoke sample (Figure 5.2) which is due to the adsorption of oxygen [46]. As expected, the oxy-cracked sample lost a higher percentage of its original weight and more quickly than the virgin petcoke sample at the early oxidation stage, which might be explained due to the high content of volatile matter present in the oxy-cracked sample. As shown in Figure 5.3, the oxy- cracked sample is competely oxidized with max rate at 475 oC, which is lower than that of the virgin petocke. This shows that the oxy-cracked petcoke can be oxidaized earlier than the virgin petcoke under similar oxidation conditions. This is also evident in the heat flow profiles of the two samples shown in Figure 5.4. From Figure 5.4, it is clear that the oxidation of the oxy-cracked sample occurs earlier than the virgin petcoke. This shift to lower oxidation temperatures in oxy-

184

cracked petcoke might be related to the low molecular weight compounds present as volatile

matters that formed during the oxy-cracking reaction. Thus, enhancing the reactivity of oxy-

cracked petcoke as the virgin petcoke has a low reactivity.

Figure 5.5 shows the plot of conversion degree (α) against temperature for non-isothermal

oxidation at three heating rates (5, 10, and 20 oC/min). The degree of conversion (α) is the fraction

of reactant decomposed at a specific temperature and is defined in terms of the mass change or the

mass of volatile generated [47]. The conversion percent ratio or the extent of reaction of petcoke

and oxy-cracked samples was estimated by Eq. (5.6) 푚 − 푚 훼 = 표 푡 (5.6) 푚표 − 푚∞ where 푚표 is the initial sample mass, 푚푡 is the sample mass at any time and 푚∞ is the final sample mass.

It is clear from the figure that as the heating rate decreased, the thermo-oxidative decomposition is shifted gradually to the lower temperature for both samples. Interestingly, the decomposition temperature of the oxy-cracked sample is much lower than virgin petcoke at any heating rate. At low heating rate (5 oC/min), for example, to obtain a 50% conversion of virgin petcoke a temperature of 498 oC is required while a temperature of 445 oC is needed for oxy-cracked one to obtain the same conversion. This significant decrease in reaction temperature again shows that the oxy-cracked sample is easier to oxidize as compared to the virgin petcoke. At a temperature lower than 430 oC, about 30% conversion is obtained for the oxy-cracked sample whereas no conversion is observed in petcoke at that temperature. This high conversion in the oxy-cracked sample at that temperature is attributed to vaporization of volatile matter that was formed during the oxy-cracking reaction. 185

It is worth noting that the slope of the oxy-cracked sample changes during the first half of the reaction process, as shown in Figure 5.5b, indicating that multiple reaction mechanisms are taken place during the oxidation reaction. This is in contrast to the slope profile of the virgin petcoke which shows that the oxidation is happening by one mechanism.

It has been reported that the ignition (푇퐼퐺), peak (푇푚) and burnout (푇퐵) temperatures for the fuel are the most important characteristic parameters to determine the combustion performance [48]. These parameters are significantly dependent on the chemical composition of the fuel [49]. Hence, the ignition temperature, the temperature at which a sudden decrease in mass loss on the DTG curves, indicates how easily the fuel is ignited [40]. The peak temperature and its corresponding rate of mass loss are determined at the maximum rate of mass loss. These parameters (i.e., 푇푚 and its mass loss) indicate the combustibility and reactivity of the fuel, where fuel with low value of 푇푚 temperature can easily ignite and react. Burnout temperature, on the other hand, is defined as the temperature at which the mass of the sample remains constant without any change during the combustion process [50]. Table 5.2 shows the determined values of these key combustion parameters extracted from Figures 5.3 and 5.4 for both virgin and oxy-cracked samples, respectively. As expected, these combustion parameters were found to be low in the case of oxy- cracked sample during the oxidation. In particular, the initial degradation temperature (ignition temperature) of the oxy-cracked sample is significantly reduced by 13% as compared to the virgin petcoke. This is due to the high content of the volatile matters in the oxy-cracked sample. The higher ignition temperature of petcoke sample could be attributed to the higher nitrogen content, which retards ignition of volatiles and reactions at the material surface [51]. The reactivity of the fuel is usually evaluated by the peak temperature; the higher the temperature indicating the lower

186

the reactivity [41]. Interestingly, the peak temperature was found to be low for the oxy-cracked sample, and thus exhibiting presence of more reactive compounds. As expected, a significant difference was observed in burnout temperature between virgin petcoke and oxy-cracked one reducing the burnout times of the fuel [18].

Based on these results, it can be concluded that the oxy-cracked products are more reactive, efficient, less pollutant and easier to oxidize than the virgin petcoke. This can be ascribed to the high content of volatile matter (VM) that formed in the oxy-cracked sample which has a significant effect on the heating value measurement as will be discussed in the next section.

Table 5.2 The thermal properties of the virgin and oxy-cracked petcoke Ignition Temp., Peak Temp., Burnout Temp.,

o o o (푻푰푮), C (푻풎), C (푻푩), C

Virgin petcoke 480 535 590

Oxy-cracked 420 475 508

187

100 1.2

1.0 80

0.8 60

0.6

0.4 Weight Loss (%) Loss Weight

20 (%/°C) Weight Deriv. 0.2

0 0.0

100 200 300 400 500 600 700 800 Temperature (°C)

Figure 5.2. TG-DTA curve for the virgin petcoke, showing the ignition, peak and burnout temperatures.

188

Figure 5.3. TG-DTA curve for oxy-cracked petcoke under air, showing the ignition, peak and burnout temperatures.

189

virgin petcoke oxy-cracked 30

10 Heat Flow (W/g) HeatFlow

100 200 300 400 500 600 700 800 Temperature (°C)

Figure 5.4. The heat flow of virgin and oxy-cracked petcoke with temperature.

190

Figure 5.5. The conversion percent (휶) with temperature at heating rates of 5, 10 and 20 oC/min for a) virgin petcoke and b) oxy-cracked petcoke.

191

Heating values of virgin and oxy-cracked petcoke

The heating values (HHV) were experimentally determined based on the amount of volatile matter

(VM), moisture (M), ash (A) and fixed carbon (FC) contents in oxy-cracked and virgin petcoke samples extracted from Figure 5.6. Figure 5.6 shows the profiles for the % mass loss with the increase in the temperature for virgin petcoke as well as for oxy-cracked petcoke obtained up to

750 oC. As shown in the figure, during the heating of the samples, the first stage in mass loss corresponds to the drying step (~200 oC) where the moisture is evaporated from the samples. The second region is the devolatilization stage (200-500 oC) where the volatiles are removed. It is worth noting that the first two stages are obtained under pyrolysis process where the (M) and (VM) contents are determined. Moreover, the combustion stage is taking place between 500-630 oC, where the loss of heavier hydrocarbons (total carbon) occurs. The final stage relates to the residual combustion stage (ash, >630 oC) where the combustion process has nearly ended [48]. The combustion and residual combustion stages were obtained under oxidation with air where the total fixed carbon (FC) and ash (A) content are estimated.

Typical proximate and ultimate analysis of petcoke and oxy-cracked samples are summarized in

Table 5.3. It is clear that the VM content of the oxy-cracked sample is significantly higher than the virgin petcoke sample. The ash content (A) was found to be less than 4% in oxy-cracked sample which was found to contain mainly potassium (K) left over after reaction while the ash content in the virgin petcoke contained metals such as iron, nickel, vanadium and cobalt. It was reported that the moisture (M) and ash content (A) have a significant effect on the quality of the fuel [52, 53].

Hence, the high value of M and A contents led to decrease in the calorific value of the fuel which then increases the operational costs. The high ash content might cause fouling and agglomeration

192

behavior during combustion, thus influences the burning rate. On the other hand, the high proportion of fixed carbon (FC) and volatile matter (VM) are playing a crucial role in chemical energy stored in the fuel. Although the FC is high in the petcoke, the oxy-cracked sample has a high value of VM. Thus, the high ratio of VM/FC was observed in the oxy-cracked sample indicating high availability of energy in the fuel. As for the ultimate analysis, the oxy-cracked sample compared to virgin petcoke has a lower than average carbon, sulfur, hydrogen and nitrogen content and a higher content of oxygen. The possible reason for this change in the chemical compositions could be explained by oxy-cracking reaction conditions. The reaction temperature and time were proven to have great effect on the solubilization of petcoke in water [32]. Under such conditions, more oxygenated hydrocarbons were produced, and thus high reaction conversion

(~ 94%) was obtained under the optimal conditions. However, a small amount of solid residue

(unreacted petcoke) was left over after the reaction, where the sulfur, nitrogen and metals are the major components of this residua [54].

193

Table 5.3 The proximate and ultimate analysis of the samples. Proximate analysis (wt%) Virgin petcoke Oxy-cracked petcoke

Volatile material (VM) 1.99 20.79 Moisture (M) 0.007 5.38 Ash (A) 2.68 3.60 Fixed Carbon (FC) 95.32 70.23 VM/FC ratio 0.021 0.29 Ultimate analysis, (wt%) C 84.40 62.31 H 3.81 2.68 N 1.55 1.10 S 4.46 1.32

Accordingly, the HHV for the samples were estimated using the proximate correlations (Equations

5.1-5.5) and presented in Table 5.4. It can be seen that the estimated HHV of the virgin petcoke sample is in range of (30-37MJ/kg) using any of these correlations which are in good agreement with the reported values of petcoke [55, 56]. This led us to conclude that the HHV correlations proposed for the coal can be accurately applied to any fuel like petcoke. Consequently, the HHV values of oxy-cracked products were estimated using the aforementioned correlations. As seen in

Table 5.4, the calculated HHV for the oxy-cracked sample is in the range of (28-31 MJ/kg).

However, the HHV estimated for the oxy-cracked by Schuster equation [42] is rather higher than expected (37 MJ/kg) as this correlation is only a function of VM, indicating the use of this equation is not valid for the petcoke. The reduction in HHV of oxy-cracked petcoke can be explained by the high oxygen content. Hence, the oxy-cracked product-structure contains high oxygen in the form 194

of carboxyl and phenolic compounds which plays an important role in chemical reactivity and oxidation. As reported, the higher carbon and hydrogen content, the higher the calorific value of the material [57]. Thus, a higher FC content and lower VM content were observed in the virgin petcoke sample as compared to the oxy-cracked sample. Even though the HHV of petcoke is relatively higher than the oxy-cracked sample, the nitrogen and sulfur content in the oxy-cracked sample are much lower, and thus low atmospheric gaseous emissions. Interestingly, the HHVs of the oxy-cracked products were found to be higher than that for the ranked-coals (9.50 – 27 MJ/kg) as previously reported [39, 58]. Therefore, the use of oxy-cracked products as fuel appears probable for power generation, by co-firing, pyrolysis or gasification into energy.

Overall, the properties of oxy-cracked products as fuel in this study are in good agreement with the typical reported values for fuel-grade petcoke and/or coal as the HHV was significantly high compared to other solid fuels in literature [56, 59]. These findings suggest that the oxy-cracked petcoke may be used as a potential fuel source of clean energy, due to it is low content of sulfur and nitrogen and its easiness to ignite.

195

Figure 5.6. Thermogravimetric analysis of the virgin and oxy-cracked petcoke at a heating rate of 10 oC/min, showing the M, MV, FC, and A.

196

Table 5.4 The heating values (HHV) for virgin and oxy-cracked petcoke samples.

Heating value, Correlation (MJ/kg) Virgin petcoke Oxy-cracked petcoke Equation 5.1 34.02 36.57 Equation 5.2 32.80 27.88 Equation 5.3 34.11 28.43 Equation 5.4 34.00 28.05 Equation 5.5 33.94 30.74

5.5 Conclusion

This study investigated the potential use of oxy-cracked petcoke as a fuel by thermally studying their oxidation, combustion properties and measuring the calorific value as well. The thermogravimetric analysis was used to evaluate the thermal degradation behavior of the virgin and oxy-cracked petcoke. The oxidation of oxy-cracked petcoke occurs at 475 °C which is lower than that of virgin petcoke where the oxidation is occurred at 540 °C. The heating values were estimated by the proximate analysis using different correlations. The results also indicate that oxy-cracked products contain a high proportion of volatile compounds and significantly high calorific heating value (~30

MJ/kg) for future energy applications. Furthermore, these findings suggest that the oxy-cracked petcoke will exhibit high reactivity comparable to the other fuels. This study also has determined that oxy-cracking process is a promising technique capable of transforming the petcoke into an

197

environmental-friendly fuel for thermal applications not only due to its combustion behavior, but also the low promotion of sulfur, nitrogen and metals.

5.6 Acknowledgment

The authors are grateful to the Natural Sciences and Engineering Research Council of Canada

(NSERC), the Department of Chemical and Petroleum Engineering at the Schulich School of

Engineering at the University of Calgary.

5.7 References

[1] E. Furimsky, Gasification of oil sand coke, Fuel Processing Technology, 56 (1998) 263-290. [2] J. Wang, E.J. Anthony, J.C. Abanades, Clean and efficient use of petroleum coke for combustion and power generation, Fuel, 83 (2004) 1341-1348. [3] B.N. Murthy, A.N. Sawarkar, N.A. Deshmukh, T. Mathew, J.B. Joshi, Petroleum coke gasification: A review, The Canadian Journal of Chemical Engineering, 92 (2014) 441-468. [4] J.D. Craddock, T.D. Rantell, J.C. Hower, D.T. Whitlow, J. Wiseman, M.C. Weisenberger, Anode coke from coal–A low cost approach, Fuel, 187 (2017) 229-241. [5] J. Li, J. Zhao, L. Zhang, T. Liu, Y. Fang, Transformation behavior of vanadium in petroleum coke during high temperature CO2 gasification, Fuel, 194 (2017) 83-90. [6] G.A. Núñez, M.a.I. Briceño, C. Gómez, T. Asa, H. Farzan, S. Hu, D.D. Joseph, Colloidal petcoke-in-water suspensions as fuels for power generation, Energy & Fuels, 26 (2012) 7147- 7154. [7] Y.K. Salkuyeh, T.A. Adams II, Integrated petroleum coke and natural gas polygeneration process with zero carbon emissions, Energy, 91 (2015) 479-490. [8] T. Chmielniak, M. Sciazko, Co-gasification of biomass and coal for methanol synthesis, Applied energy, 74 (2003) 393-403. [9] C. Eswaraiah, Grinding characteristics of coal and petroleum coke/coal blends on utilization for combustion, Particulate Science and Technology, 34 (2016) 223-228. [10] L. Ren, R. Wei, Y. Gao, Co-gasification reactivity of petcoke and coal at high temperature, Fuel, 190 (2017) 245-252. [11] H. Ibrahim, Determination of the Calorific Value of Syrian Delayed Petroleum Coke, Int J Petrochem Sci Eng, 1 (2016) 00012. [12] E. Hall, E. Tollefson, Z. George, L. Schneider, Upgrading of delayed and fluid cokes from oil sand by desulfurization, The Canadian Journal of Chemical Engineering, 60 (1982) 418-424. [13] H. Al-Haj-Ibrahim, B.I. Morsi, Desulfurization of petroleum coke: a review, Industrial & engineering chemistry research, 31 (1992) 1835-1840. [14] H. Shlewit, M. Alibrahim, Extraction of sulfur and vanadium from petroleum coke by means of salt-roasting treatment, Fuel, 85 (2006) 878-880. [15] P. Queneau, R. Hogsett, L. Beckstead, D. Barchers, Processing of petroleum coke for recovery of vanadium and nickel, Hydrometallurgy, 22 (1989) 3-24. 198

[16] J. Xiao, Y. Zhang, Q. Zhong, F. Li, J. Huang, B. Wang, Reduction and Desulfurization of Petroleum Coke in Ammonia and Their Thermodynamics, Energy & Fuels, 30 (2016) 3385-3391. [17] Q. Zhou, Q. Zhao, G. Zhou, H. Wang, T. Xu, S. Hui, Comparison of combustion characteristics of petroleum coke and coal in one-dimensional furnace, Frontiers of Energy and Power Engineering in China, 4 (2010) 436-442. [18] N. Berguerand, A. Lyngfelt, The use of petroleum coke as fuel in a 10 kWth chemical-looping combustor, International Journal of Greenhouse gas control, 2 (2008) 169-179. [19] N.S. Yuzbasi, N. Selçuk, Air and oxy-fuel combustion behaviour of petcoke/lignite blends, Fuel, 92 (2012) 137-144. [20] J. Wang, E.J. Anthony, Clean combustion of solid fuels, Applied Energy, 85 (2008) 73-79. [21] R.W. Bryers, Utilization of petroleum coke and petroleum coke/coal blends as a means of steam raising, Fuel Processing Technology, 44 (1995) 121-141. [22] C. Cain-Borgman, Petroleum coke refueling an economic alternative for existing oil-fired boilers, Air Products and Chemicals, Inc., 2009. [23] K. Jayaraman, I. Gokalp, Gasification characteristics of petcoke and coal blended petcoke using thermogravimetry and mass spectrometry analysis, Applied Thermal Engineering, 80 (2015) 10-19. [24] M. Koffi, R.J. da Silva, Contral of the SOx Emissions During the Pertoleum Coke Combustion in Fluidized Bed, 19th International Congress of Mechanical Engineering, (2007). [25] J. Lee, J. Baker, J. Rolle, R. Lierena, Comparison of fuel properties of petroleum cokes and coals used in power generation, Amer Chemical SOC, 1999, pp. 810-810. [26] F.A. Amorim, B. Welz, A.C. Costa, F.G. Lepri, M.G.R. Vale, S.L. Ferreira, Determination of vanadium in petroleum and petroleum products using atomic spectrometric techniques, Talanta, 72 (2007) 349-359. [27] J. Li, J. Zhao, L. Zhang, X. Dai, Y. Fang, Predicting the vanadium speciation during petroleum coke combustion by thermodynamic equilibrium calculation, Journal of Thermal Analysis and Calorimetry, 130 (2017) 2239-2247. [28] S.J. Yoon, Y. C. Choi, S.-H. Lee, J. G. Lee, Thermogravimetric study of coal and petroleum coke for co-gasification, Korean Journal of Chemical Engineering, 24 (2007) 512-517. [29] R.D. Jarvis, Handling and combustion of ground petroleum coke in the pulp and paper industry, Matrix Industrial Systems, Inc., 2010. [30] K.S. Milenkova, A.G. Borrego, D. Alvarez, R. Menéndez, H.I. Petersen, P. Rosenberg, Coal blending with petroleum coke in a pulverized-fuel power plant, Energy & fuels, 19 (2005) 453- 458. [31] Q. He, R. Wang, W. Wang, R. Xu, B. Hu, Effect of particle size distribution of petroleum coke on the properties of petroleum coke–oil slurry, Fuel, 90 (2011) 2896-2901. [32] A.D. Manasrah, N.N. Nassar, L.C. Ortega, Conversion of petroleum coke into valuable products using oxy-cracking technique, Fuel, 215 (2018) 865-878. [33] A. Manasrah, A. El-Qanni, I. Badran, L.C. Ortega, M.J. Perez-Zurita, N.N. Nassar, Experimental and theoretical studies on oxy-cracking of Quinolin-65 as a model molecule for residual feedstocks, Reaction Chemistry & Engineering, 2.5 (2017) 703-719. [34] M. Ashtari, L. Carbognani Ortega, F. Lopez-Linares, A. Eldood, P. Pereira-Almao, New pathways for asphaltenes upgrading using the oxy-cracking process, Energy & Fuels, 30 (2016) 4596-4608. 199

[35] A. Demirbaş, Calculation of higher heating values of biomass fuels, Fuel, 76 (1997) 431-434. [36] C. Sheng, J. Azevedo, Estimating the higher heating value of biomass fuels from basic analysis data, Biomass and Bioenergy, 28 (2005) 499-507. [37] S. Mishra, S. Panda, N. Pradhan, D. Satapathy, S.K. Biswal, B.K. Mishra, Insights into DBT biodegradation by a native Rhodococcus strain and its sulphur removal efficacy for two Indian coals and calcined pet coke, International Biodeterioration & Biodegradation, 120 (2017) 124-134. [38] A. Majumder, R. Jain, P. Banerjee, J. Barnwal, Development of a new proximate analysis based correlation to predict calorific value of coal, Fuel, 87 (2008) 3077-3081. [39] M. Chukwu, C. Folayan, G. Pam, D. Obada, Characterization of some Nigerian coals for power generation, Journal of Combustion, 2016 (2016). [40] J. J. Lu, W.-H. Chen, Investigation on the ignition and burnout temperatures of bamboo and sugarcane bagasse by thermogravimetric analysis, Applied Energy, 160 (2015) 49-57. [41] S.G. Sahu, P. Sarkar, N. Chakraborty, A. Adak, Thermogravimetric assessment of combustion characteristics of blends of a coal with different biomass chars, Fuel Processing Technology, 91 (2010) 369-378. [42] V. Schuster, Uber die Berechnung des Heizwertes von Kohlen aus der Immediatezusammensetzung, Brennst Chem, 32 (1951) 19-20. [43] S. Küçükbayrak, B. Dürüs, A.E. Meríçboyu, E. Kadiog̈ lu, Estimation of calorific values of Turkish lignites, Fuel, 70 (1991) 979-981. [44] T. Cordero, F. Marquez, J. Rodriguez-Mirasol, J. Rodriguez, Predicting heating values of lignocellulosics and carbonaceous materials from proximate analysis, Fuel, 80 (2001) 1567-1571. [45] J. Parikh, S. Channiwala, G. Ghosal, A correlation for calculating HHV from proximate analysis of solid fuels, Fuel, 84 (2005) 487-494. [46] S. Salvador, J.-M. Commandre, B. Stanmore, Reaction rates for the oxidation of highly sulphurised petroleum cokes: the influence of thermogravimetric conditions and some coke properties☆, Fuel, 82 (2003) 715-720. [47] N.N. Nassar, A. Hassan, P. Pereira-Almao, Comparative oxidation of adsorbed asphaltenes onto transition metal oxide nanoparticles, Colloids and surfaces A: Physicochemical and Engineering aspects, 384 (2011) 145-149. [48] S.A. El-Sayed, M. Khairy, Effect of heating rate on the chemical kinetics of different biomass pyrolysis materials, Biofuels, 6 (2015) 157-170. [49] F.A. López, T.A. Centeno, I. García-Díaz, F.J. Alguacil, Textural and fuel characteristics of the chars produced by the pyrolysis of waste wood, and the properties of activated carbons prepared from them, Journal of Analytical and Applied Pyrolysis, 104 (2013) 551-558. [50] X. Zeng, S. Zheng, H. Zhou, Q. Fang, C. Lou, Char burnout characteristics of five coals below and above ash flow temperature: TG, SEM, and EDS analysis, Applied Thermal Engineering, 103 (2016) 1156-1163. [51] N. Whitely, R. Ozao, Y. Cao, W.-P. Pan, Multi-utilization of chicken litter as a biomass source. Part II. Pyrolysis, Energy & fuels, 20 (2006) 2666-2671. [52] J.M. Sweeten, K. Annamalai, B. Thien, L.A. McDonald, Co-firing of coal and cattle feedlot biomass (FB) fuels. Part I. Feedlot biomass (cattle manure) fuel quality and characteristics☆, Fuel, 82 (2003) 1167-1182.

200

[53] Z. Liu, A. Quek, R. Balasubramanian, Preparation and characterization of fuel pellets from woody biomass, agro-residues and their corresponding hydrochars, Applied Energy, 113 (2014) 1315-1322. [54] M. Ashtari, L. Carbognani, P. Pereira-Almao, Asphaltenes aqueous conversion to humic and fulvic analogs via oxy-cracking, Energy & Fuels, 30 (2016) 5470-5482. [55] Y.M. Elkasabi, A.A. Boateng, C.A. Mullen, Methods of Producing Calcined Coke From Bio- oil And Calcined Coke Produced Thereby, U.S. Patent Application 14/972,881, 2016. [56] A. Demirbas, N. Ak, A. Aslan, N. Sen, Calculation of higher heating values of hydrocarbon compounds and fatty acids, Petroleum Science and Technology, (2018) 1-6. [57] R.H. White, Effect of lignin content and extractives on the higher heating value of wood, Wood and fiber science, 19 (2007) 446-452. [58] Q. Feng, J. Zhang, X. Zhang, S. Wen, Proximate analysis based prediction of gross calorific value of coals: a comparison of support vector machine, alternating conditional expectation and artificial neural network, Fuel Processing Technology, 129 (2015) 120-129. [59] R.G. Etter, Production and use of a premium fuel grade petroleum coke, U.S. Patent 6,168,709, 2001.

201

Conclusions and Future Work

Each of the paper-based chapters included its detailed conclusions. This section contains a summary of the key conclusions for the research presented in this thesis along with some recommendations for future work in this area of research.

8.1 Conclusions

A novel technique for solid heavy hydrocarbons conversion in aqueous media was proposed, where the feasibility of producing valued-added products mostly in forms of humic acid analogs were verified. The main objective of this thesis was successfully accomplished by exploring a new approach for petcoke conversion into valuable products using the oxy-cracking technique. The oxy-cracking is a result of combining oxidation and cracking reactions in an aqueous alkaline medium under mild temperature and pressure. The concept of oxy-cracking was investigated experimentally and theoretically on Quinolin-65 (Q-65) as a model molecule of residual feedstock, as presented in Chapter Two. The experimental study was conducted in a batch Parr reactor at different operating temperatures (200-230 ºC) and constant oxygen partial pressure of 750 psi while the reaction times ranged between 0 and 2 h. The concentration of the solubilized organic compounds in the liquid phase was monitored by the total organic carbon (TOC), while the gas products were analyzed online using gas chromatography (GC). The experimental results enabled us to establish the reaction kinetics based on the lumped kinetics model. Thus, the Q-65 was oxy- cracked consecutively into different pathways in which an oxidative decomposition took place in the first step producing different aromatic intermediates. The latter compounds were partially oxidized into different families of organic acids and a small amount of CO2. These findings were also confirmed theoretically and supported by the computational modeling under the same

202

proposed experimental conditions. The theoretical study succeeded in exploring many oxidation pathways of the Q-65 molecule initiated by hydroxyl radical (•OH) and the hydroxide anion (-OH) as well. Moreover, in the real application of petroleum residue, the noncatalytic oxy-cracking reaction of petcoke was conducted in a batch reactor, where petcoke particles are solubilized in an aqueous alkaline medium and partially oxidized under mild operating temperature and pressure.

Several operating conditions were investigated, such as temperature, oxygen pressure, reaction time, particle size and mixing rate to optimize the solubility and selectivity of oxy-cracked products. The results showed that the temperature and the residence time are the two major important parameters that affect the reaction conversion and selectivity where the highest

o conversion and selectivity along with minimal amount of CO2 emission were obtained at 230 C and 2 h. Also, to avoid mass transfer limitations the mixing rate was kept above 500 rpm. The triangular reaction with a lumped kinetics model has successfully described the kinetics of petcoke oxy-cracking reaction based on the hydroxyl radical mechanism. Therefore, the petcoke is oxy- cracked simultaneously into water-soluble species and CO2, with the consecutive reaction of soluble species into CO2. Comprehensive characterization was carried out on water solubilized petcoke (desired products) and the residual solids using FTIR, NMR and XPS techniques.

According to FTIR results, the main species solubilized in water were oxygenated hydrocarbons compounds and some organic acid such as carboxylic and sulfonic acids and their salts. These results were also confirmed by NMR and XPS analysis. Additionally, the FTIR results of the residual solids remaining after the reaction showed that the structures and functional groups are similar to the original petcoke. Interestingly, most of the metals contents were found in the residual petcoke compared with the metals in the liquid phase. Therefore, the oxy-cracking technique can

203

be considered as a new feasible process for conversion and demineralization/desulfurization of petcoke.

New catalyst was proposed to enhance the reaction conversion and selectivity of the oxy-cracking as well as to reduce the required energy. A nanocrystalline copper-silicate material was tested for the petcoke oxy-cracking, as presented in Chapter Four. Upon synthesis, the properties of the catalyst were successfully characterized using XRD, BET, SEM, and IR spectroscopy. The catalyst activity and selectivity were experimentally investigated in a batch reactor at different reaction conditions. More importantly, a high reaction conversion and high selectivity to water solubilized products with almost zero emission of CO2 were obtained. Interestingly, the leaching percentage was less than 3% of the whole Cu even at elevated temperatures. After being reused three times, the material still retained satisfactory for its catalytic activity. The solubilized petcoke in water, by catalytic oxy-cracking, were found to contain carboxylic, carbonyl, phenolic, and sulfonic functions which responsible for humic acids analogs based on the FTIR results. Comparing with the commercial humic acid, there were many similarities except the higher presence of sulfonic compounds in the oxy-cracked petcoke.

The application of oxy-cracked products, from the noncatalytic reaction, as a new alternative fuel for power generation was investigated by measuring the heating value and the thermo-oxidative decomposition analysis using TGA as presented in Chapter Five. The heating value results indicated that oxy-cracked products contain a high proportion of volatile compounds and significantly high calorific heating value (~30MJ/kg). Regardless of the heating value of oxy- cracked products being relatively lower than the virgin petcoke, the nitrogen and sulfur content in the oxy-cracked products is much lower than that in virgin petcoke, thus low atmospheric gaseous

204

emissions. Additionally, the oxidation results showed that the oxy-cracked product is easier and faster to oxidize compared to the virgin petcoke. These findings suggest that the use of oxy-cracked product as a fuel for power generation would be safer, more efficient, and less polluting than the virgin petcoke.

8.2 Recommendations

The main target of this thesis was to convert the residual feedstock such as petcoke into valuable products. However, in order to have the oxy-cracking as a real process, it would be highly recommended running the oxy-cracking reaction in a continuous process using CSTR, packed bed and BFR as well as investigating the optimal operating conditions for the high yield.

Another factor to consider is the investigation of the operating conditions for demineralization of petcoke and desulfurization using oxy-cracking technique.

Moreover, conducting a full economic study assessment for scaling up this process to ensure its cost effectiveness is essential to feasible applications.

Finally, another application in the oxy-cracked products is to convert them into carbon nanotubes

(CNT) to be used in as a nanofluids for enhanced oil recovery. Our group has already started investigating this approach due to its high potential application in the area of enhanced oil recovery.

205

Appendix A1

Supplementary Material

S1. Characterization methods

S1.1 1H Nuclear magnetic resonance (NMR) spectroscopy

The NMR spectra of oxy-cracked Q-65 were determined by a Bruker 600 MHz spectrometer (4 mm BL4 liquid probe, cross-polarization program, and spin rate of 8k). 1H NMR spectra were

o 1 taken at 25 C with D2O solvent and a 4 mm internal diameter probe. H NMR spectra were collected with a pulse sequence of zg30, a relaxation time of 2s, and averaging of 160 scans/run.

Virgin Q-65 was also analyzed for comparison. The NMR spectrum was analyzed using the commercial NMR simulator software (Mnova NMR) which assigns most structure types available at different frequencies.

S1.2 X-ray photoelectron spectroscopy (XPS)

After drying the oxy-cracked Q-65 in a vacuum oven at 65 oC overnight, the solid was characterized using an XPS PHI VersaProbe 5000 spectrometer to determine the species present and to quantify the amount of carbon, oxygen, nitrogen, and sulfur on the surface. The spectra were taken using a monochromatic Al source (1486.6 eV) at 50 W and a beam diameter of 200.0

μm with a takeoff angle of 45°. The samples were pressed on double-sided tape, and the spectra were taken with double neutralization. The binding energies were reported relative to C1s at 284.8 eV. The sample sputtering protocol involved 20 min of Argon sputtering at 45°, 2 kV, 1.5 μA 2 ×

2 (10.5 nm/min). Calibration for sputtering purposes was performed with SiO2/Si wafer having a

206

SiO2 layer of 100 nm. The high-resolution spectra of C1s, O1s, N1s and S2p were fitted using

MultiPak software developed by Physical Electronics.

S2. Characterization results

S2.1 1H NMR spectroscopy of the oxy-cracked Q-65

The oxy-cracking reaction was further confirmed by using high-resolution NMR for analyzing the virgin and oxy-cracked Q-65. The 1H NMR spectrum of virgin Q-65 is shown in Figure S1a. The

1H NMR spectrum shows well-resolved signals for aromatic protons (6.5-9 ppm) which can be attributed to aromatics rings. The aliphatic hydrogen signals (0.5-4.6 ppm) were divided into three types of protons; terminal methyl groups (0.5-1.0 ppm) and internal methylene groups (1.0-1.85 ppm), in addition to α and β protons to oxygen atoms appear at 4.6 ppm and 2.8 ppm, respectively.

For the 1H-NMR spectrum of oxy-cracked Q-65 at high conversion (230 oC and 2 h), dramatic differences were found as shown in Figure S1b. The spectrum presents typical functional chemical shifts such as alkyl methylene (Al-CH3) groups (1.9 ppm), methylene bonded to aromatic (Ar-

CH2) groups (2.7 ppm) and aromatic molecules (~7.0−8.5 ppm) as well as methyl groups from aromatic esters (Ar-COOCH3) (2.3 ppm), and methoxy groups (Ar-O-CH3) bonded to monoaromatics (4.0 ppm). Potassium carboxylates salt type molecules were assigned to the strong signal appearing around 8.2 ppm. From the preceding findings, production of carboxylates and phenol derivatives during oxy-cracking is again confirmed, as it was formerly evidenced by FTIR and XPS spectroscopy. These NMR signals most likely represent the oxygenated products resulted from oxy-cracking of Q-65 such as presence of acids and their salts. These findings match with the information derived from FTIR spectroscopy and XPS, especially for the production of

207

carboxylic acids during oxy-cracking reaction, similar to those results obtained by Ashtari et al.

[1] for oxy-cracking of C7 asphaltenes.

Aliphatic

{

Aromatic

{

10 8 6 4 2 0 ppm

Ar-CH2

{

Ar-CH3

Ar-COOCH{ 3

Ar-OCH3

{ {

Ar-CH2

10 8 6 4 2 0 ppm

Figure S 1. 1H NMR spectra for a) virgin Q-65 ran with DCM, b) oxy-cracked Q-65 ran with D2O solvent. Signal frequencies for typical chemical structures appended.

208

S2.2 XPS of Q-65 before and after oxy-cracking

Table S1 shows the results of the quantitative XPS analysis along with the positions of the C1s,

O1s, N1s and S2p signals of virgin Q-65 and oxy-cracked Q-65. The virgin Q-65 is composed

primarily of carbon atoms (91.4 at%) and a minor amount of heteroatoms (3.9 at% oxygen, 2.5

at% nitrogen, and 2.2 at% sulfur). Minor differences with the expected values can be related with

the presence of the impurities in the sample.

The results show that the oxy-cracked Q-65 has a higher proportion of oxygen atoms (86 at%) and

much lower proportion of carbon atoms (12.6 at%) compared with the virgin Q-65.

Figure S2a and b show the XPS spectra of C1s for both samples (virgin and oxy-cracked Q-65). It

can be clearly observed that the distribution of carbon species dramatically changes when Q-65 is

submitted to oxy-cracking reaction. Figure S2a shows the C1s spectrum of virgin Q-65. As seen,

four signals at 283.79 eV, 284.80 eV, 286.34 eV, and 289.21 eV can be observed, which were

attributed to carbon bonds (C=C) with high intensity, (C–C/C–H), (C–O) and (C=O), respectively

[2, 3]. The C1s spectrum of the oxy-cracked Q-65 (Figure S2b) shows the same four signals but

the distribution of species is completely different. The signal corresponding to the aromatic C=C

bonds (283.79 eV) decreased dramatically and the signal at 284.90 eV which is attributed to C-

H/C-C remains unchanged. The most aromatic (C=C) bonds have been oxygenated and formed

either carboxylic bonds or C-OH bonds. Indeed, this result is in a good agreement with what was

observed in the FTIR of oxy-cracked Q-65.

Figure S3a and b show the O1s XPS spectra for both virgin and oxy-cracked Q-65. Virgin Q-65 spectrum (Figure S3a) shows two signals at 530.32 eV and 532.06 eV attributed to oxygen in

209

carboxylic groups (C=O) and oxygen in C–O bonds, respectively. The oxy-cracked Q-65 (Figure

S3b) shows a different spectrum. In addition to the two bonds observed in the virgin Q-65, a distinctive signal at 532.77 eV can be observed and attributed to oxygen in hydroxyls bonds (O-H)

[4]. Additionally, the intensity of the signals in the oxy-cracked Q-65 are almost four times higher than that in the virgin Q-65. This confirms the high degree of oxidation of Q-65, in agreement with the observed behavior of the C1s spectrum.

Figure S4 shows the N1s XPS spectra. The observed signal at 397.89 eV of virgin Q-65 (Figure

S4a) is assigned to nitrogen with pyridinic (C-N=C) [5]. However, no signals were observed in the region of N1s in the oxy-cracked Q-65 (Figure S4b) but noise. Similarly, the S2p doublet of virgin

Q-65 was observed at 162.67 eV and 163.88 eV. This doublet is assigned to thiophenic species, as shown in Figure S5a [6]. However, Figure S5b shows no signals for sulfur after the oxy-cracking reaction of Q-65. Thus, we can conclude from both Figures S4b and S5b that little amounts of sulfur and nitrogen species in the solubilized Q-65 were present.

210

Table S 1 Fitting signal data of high-resolution spectra of species in virgin and oxy-cracked Q- 65. Before Reaction After Reaction

Atomic Atomic

Conc. Position Bond Conc. Position Bond

(%) (eV) assignment (%) (eV) assignment

283.79 C=C 284.90 C–C/C–H

284.80 C–C/C–H 286.04 C–O

286.34 C–O 289.71 O-C=O

C1s 91.40 289.21 C=O 12.60

530.32 C-O 530.36 C-O, O=C-

532.06 C=O 531.63 OH

532.77 C=O

O1s 3.90 86.00 C-OH

397.89 C-N=C ------

N1s 2.50 0.80

162.67 C-S-C ------

S2p 2.20 163.88 S-O 0.60

211

Figure S2. High-resolution XPS spectra with fitting of C1s (a) before reaction, (b) after reaction.

Figure S3. High-resolution XPS spectra with fitting of O1s (a) before reaction, (b) after reaction.

212

Figure S4. High-resolution XPS spectra with fitting of N1s (a) before reaction, (b) after reaction.

Figure S5. High-resolution XPS spectra with fitting of S2p (a) before reaction, (b) after reaction.

213

S3. References

[1] M. Ashtari, L. Carbognani Ortega, F. Lopez-Linares, A. Eldood, P. Pereira-Almao, New Pathways for Asphaltenes Upgrading Using the Oxy-Cracking Process, Energy & Fuels, 30 (2016) 4596-4608. [2] T. Ramanathan, F. Fisher, R. Ruoff, L. Brinson, Amino-functionalized carbon nanotubes for binding to polymers and biological systems, Chemistry of Materials, 17 (2005) 1290-1295. [3] Y.-L. Huang, H.-W. Tien, C.-C.M. Ma, S.-Y. Yang, S.-Y. Wu, H.-Y. Liu, Y.-W. Mai, Effect of extended polymer chains on properties of transparent graphene nanosheets conductive film, Journal of Materials Chemistry, 21 (2011) 18236-18241. [4] J. Wang, C. Li, L. Zhang, G. Que, Z. Li, The properties of asphaltenes and their interaction with amphiphiles, Energy Fuels, 23 (2009) 3625-3631. [5] S. Kelemen, M. Gorbaty, P. Kwiatek, Quantification of nitrogen forms in Argonne premium coals, Energy & Fuels, 8 (1994) 896-906. [6] W. Abdallah, S. Taylor, Surface characterization of adsorbed asphaltene on a stainless steel surface, Nuclear Instruments and Methods in Physics Research Section B: Beam Interactions with Materials and Atoms, 258 (2007) 213-217.

214

Appendix A2 Effect of the pressure on oxy-cracking reaction

Figure S 6 Effect of pressure on the conversion of petcoke during oxy-cracking reaction at t = 2 h and 1 g of KOH.

215

Statement of Contribution for the thesis titled “Conversion of Petroleum Coke into Valuable Products using Catalytic and Noncatalytic Oxy-Cracking Reaction” The co‐authors listed below have certified that: 1. They meet the criteria for authorship where they have participated in the conception, design and planning, execution, or interpretation of the results, of at least that part of the publications in the thesis in their field of expertise; 2. They take public responsibility for their part of the publication, except for the first author who accepts overall responsibility for the publication; 3. There are no other authors of the publications according to these criteria; and 4. They grant permission to the use of the publications in the student’s (first author) thesis Author Contribution Signature

Abdallah D. Manasrah Developing a research plan, performing the experimental setup, analysis, interpreting the results and writing the first draft of the manuscripts Nashaat N. Nassar Principal investigator and supervisor of the work

Amjad El-Qani Helping in conducting the oxy-cracking reaction of Q-65 and analyzing the kinetic modeling in the first paper Ismail Badran Contributing in the theoretical calculations of oxy- cracking the Q-65 in the first paper Lante Carbognani Interpreting the FTIR and NMR results of the oxy- cracked products in the first and second papers Josefina Perez-Zurita Assisting in the XPS analysis and interpreting the XPS results in the first paper Gerardo Vitale Helping in the preparation of the catalyst and XRD analysis and in revising the first draft of the third paper Azfar Hassan Contributing in the TGA analysis, interpreting and revising the results in the fourth paper

______Dr. Nashaat N. Nassar Date Supervisor

216

217