UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Studies of exfoliated molybdenum disulfide catalyst in hydrocracking and hydroprocessing reactions Tye, Ching Thian 2006

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_2006-199152.pdf [ 15.09MB ]
Metadata
JSON: 831-1.0058992.json
JSON-LD: 831-1.0058992-ld.json
RDF/XML (Pretty): 831-1.0058992-rdf.xml
RDF/JSON: 831-1.0058992-rdf.json
Turtle: 831-1.0058992-turtle.txt
N-Triples: 831-1.0058992-rdf-ntriples.txt
Original Record: 831-1.0058992-source.json
Full Text
831-1.0058992-fulltext.txt
Citation
831-1.0058992.ris

Full Text

STUDIES OF E X F O L I A T E D M O L Y B D E N U M DISULFIDE C A T A L Y S T IN H Y D R O C R A C K I N G A N D H Y D R O P R O C E S S I N G REACTIONS by CHING THIAN T Y E B.Eng. (Chem) (Hons), Universiti Sains Malaysia, 1999 M.Sc. (Chem. Eng.), Universiti Sains Malaysia, 2001 A THESIS SUBMITTED IN PARTIAL F U L F I L M E N T OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in THE F A C U L T Y OF G R A D U A T E STUDIES (Chemical & Biological Engineering) THE UNIVERSITY OF BRITISH C O L U M B I A February 2006 © Ching Thian Tye, 2006 Abstract Studies of M 0 S 2 catalysts have demonstrated a close relationship between catalyst activity in hydroprocessing and the structure of the M 0 S 2 . Most reports have focused on hydrodesulfurization and hydrogenation and some of these remain somewhat controversial. Moreover, the effect of M 0 S 2 structure on heavy oil hydrocracking has not been well studied. In this work, exfoliated M 0 S 2 was used to study the catalyst structure-activity relationship in Cold Lake heavy oil hydrocracking and model compound hydroprocessing. The exfoliated M 0 S 2 is dispersed directly in the oil eliminating the influence of catalyst support, and the laminar structure of exfoliated M 0 S 2 provides an interesting framework to study the relationship between catalyst activity and structure. In heavy oil hydrocracking at 415°C, exfoliated M 0 S 2 had similar liquid and coke yields but exhibited a better quality liquid product, especially in asphaltene and microcarbon residue removal, than M 0 S 2 derived from molybdenum naphthenate (MoNaph). In addition, the feasibility of recycling coke was confirmed as more than 90 % of Mo resided in the coke after reaction and the spent catalyst-in-coke was active in Cold Lake heavy oil hydrocracking. The activity of exfoliated M 0 S 2 for hydroprocessing at 350°C was also determined using various model reactants. The reactants studied were naphthalene for hydrogenation, dibenzothiophene for hydrodesulfurization, quinoline and carbazole for hydrp-denitrogenation, and phenol for hydrodeoxygenation. Again, the resulting catalyst activity was compared to MoNaph derived M 0 S 2 . Exfoliated M 0 S 2 gave better overall hydrodesulfurization and hydrodeoxygenation as compared to MoNaph derived M 0 S 2 . In contrast, MoNaph derived M 0 S 2 showed higher hydrogenation and hydrodenitrogenation 11 activities. These results were a consequence of the morphological differences between the two catalysts in the model systems. The activity of exfoliated M 0 S 2 in hydrodesulfurization of dibenzothiophene was further compared to that of crystalline M 0 S 2 , MoNaph, and ammonium heptamolybdate derived M 0 S 2 . The prepared catalysts had significantly different morphologies, as described by the crystallite stack height and slab length. These data were used to estimate the fraction of rim and edge sites in the crystallites. The catalysts' selectivity for hydrogenolysis and hydrogenation reactions were shown to correlate with the fraction of rim or edge sites. Selectivity for hydrogenolysis increased as the fraction of edge sites increased. i i i Table of Contents Abstract ii Table of Contents iv List of Tables x List of Figures : xiv List of Abbreviations xxi Acknowledgement xxiv Chapter 1 Introduction 1 1.1 Objectives 7 Chapter 2 Literature Review 9 2.1 Terminology and Background 9 2.1.1 Heavy Oil and Residues 9 2.1.1.1 Asphaltene 10 2.1.1.2 Heteroatoms and Metals 11 2.1.2 Synthetic Crude Oil 12 2.1.3 Upgrading and Reactions 12 2.1.3.1 Thermal Cracking 13 2.1.3.2 Hydrocracking 14 2.1.3.3 Hydrodesufurization (HDS) 15 2.1.3.4 Hydrodenitrogenation (HDN) 15 2.1.3.5 Hydrodeoxygenation (HDO) 16 2.1.4 Catalyst Deactivation 17 2.2 Primary Upgrading 17 iv 2.2.1 Carbon Rejection 18 2.2.2 Hydrogen Addition/Hydrocracking 18 2.2.2.1 Fixed Bed Reactor 19 2.2.2.2 Ebullated Bed Reactor 19 2.2.2.3 Slurry Phase Reactor..... 20 2.2.3 Slurry Phase Process and Dispersed Catalyst 21 2.2.3.1 Hydrocracking Mechanism in the Presence of Dispersed Catalyst ....22 2.2.3.2 Molybdenum Based Catalyst 23 2.2.3.3 Active Species and Properties of Dispersed Catalyst 25 2.2.3.4 Multicomponent Catalyst 27 2.2.3.5 Effect of Reaction Variables 28 2.2.3.6 Catalyst Recycle 29 2.2.4 Exfoliated Catalyst 30 2.2.5 Hydrocracking of Different Cut Points 32 2.3 Secondary Upgrading/Hydroprocessing 33 2.3.1 Unsupported M 0 S 2 Structure-Activity Relationships 34 2.3.2 Use of Exfoliated Catalyst in Hydroprocessing 42 2.4 Summary...: 44 Chapter 3 Experimental 46 3.1 Catalysts Preparation 46 3.1.1 Preparation of Exfoliated MoS 2 47 3.2 Catalyst Activity Measurement- 48 3.2.1 Heavy Oil Hydrocracking Reaction 48 3.2.1.1 Feedstock 48 3.2.1.2 Reactor Setup 50 3.2.1.3 Hydrocracking Activity Test Procedure 50 3.2.1.4 Product Gas Analysis 53 3.2.1.5 Liquid Product Analysis 53 3.2.2 Model Compound Reactions 54 3.2.2.1 Reactor Setup 54 3.2.2.2 Hydroprocessing Activity Test Procedure 54 3.2.2.3 Liquid Product Analysis 56 3.3 Catalyst Characterization 56 3.3.1 X-ray Diffraction Analysis 57 3.3.2 Brunauer-Emmett-Teller (BET) Surface Area 58 3.3.3 Scanning Electron Microscopy-Energy Dispersive X-Ray Emission (SEM-EDX) 58 3.3.4 Transmission Electron Microscopy (TEM) 59 3.3.5 Diffuse Reflectance Infrared Fourier Transform Spectroscopy (DRIFTS)...59 Chapter 4 Exfoliated M0S2 Catalyst in Cold Lake Heavy Oil Hydrocracking 60 4.1 Properties of Exfoliated MoS 2 61 4.2 Catalyst Activity Studies 67 4.2.1 Activity Comparison 67 4.2.1.1 Effect of Sulfiding Reagent, H 2 S 70 4.2.2 Effect of Process Conditions on Hydrocracking Using Exfoliated M 0 S 2 72 4.2.2.1 Reaction Temperature 72 vi 4.2.2.2 Effect of Mo Concentration 73 4.2.2.3 Effect of Concentration and Temperature 75 4.2.3 Effect of Properties of Exfoliated MoS 2 77 4.2.3.1 Dispersion Concentration 77 4.2.3.2 Effect of Dispersing Solvent 79 4.2.4 Hydrocracking of Different Petroleum Fractions as Feedstock 81 4.3 Characterization of the Recovered Coke 85 4.3.1 BET Surface Area of the Recovered Coke 86 4.3.2 X R D Measurement of the Recovered Coke 86 4.3.3 Infrared Analysis 89 4.3.4 Metal Content Analysis 92 4.3.5 S E M and E D X of Coke 93 4.4 Catalyst Recycle 94 4.5 Summary..: 96 Chapter 5 Effect of MoS 2 Catalyst Morphology on Hydroprocessing Reactions 97 5.1 Catalyst Characterization 99 5.1.1 BET Surface Area 100 5.1.2 S E M - E D X 101 5.1.3 X R D 102 5.1.4 T E M Observation and Analysis 104 5.2 Hydrogenation of Naphthalene 109 5.3 Hydrodesulfurization of Dibenzothiophene 118 5.4 Hydrodenitrogenation 132 vii 5.4.1 Hydrodenitrogenation of Quinoline 132 5.4.2 Hydrodenitrogenation of Carbazole 139 5.5 Hydrodeoxygenation of phenol 149 5.6 Structure-activity Relationship 153 5.7 Activity Comparison between MoNaph Derived M0S2 and Exfoliated M0S2 . . 162 5.8 Summary 165 Chapter 6 Conclusions and Recommendations 166 6.1 Conclusions 166 6.2 Recommendations 168 Reference 170 Appendix A Methods Used for Heavy Oil Hydrocracking Liquid Analysis 180 Appendix B Experimental Results of Cold Lake Heavy Oil Hydrocracking Reactions 183 B-1 Summary of Results of Hydrocracking Activity Experiments 183 B-2 Error Analysis in Hydrocracking Reactions 188 B-3 Calculations 190 B-3.1 Sample of Calculation for Hydrogen Consumption 190 B-3.2 Determination of Average, Standard Error and % of Standard Error 194 B-3.3 Statistical Difference of Reactions 195 B-3.4 Variance Analysis: Comparison of H 2 S and Pure H2 in Hydrocracking Using MoNaph 197 B-3.5 Variance Analysis: Comparison of H2S and Pure H 2 in Hydrocracking Using Exfoliated MoS 2 199 B-3.6 Calculation of Total Heteroatom Removal for the Filtrate and Retentate...201 viii Appendix C Presence of Water in Model Compound Reactions 204 Appendix D Hydroprocessing Experiments 207 D-1 Summary of Hydroprocessing Reactions 207 D-2 Example of Calculation for Model Compound Experiments 208 D-3 Repeatability of Model Compound Experiments : 215 D-4 Calculated Data for Model Compound Experiments 221 Appendix E Sample Calculations of Rate Constants 257 E - l Example of Rate Constant Determination from Experimental Data 257 E-1.1 Example of 1 s t Order Rate Constant Determination 257 E-l .2 Sample Calculation for Zero Order Rate Constant Determination 259 E-2 Example of Apparent Activation Energy Calculation 261 E-3 Sample Calculation for Confidence Interval of Estimated Parameter 262 E-4 Plots of Experimental Data Versus Simulated Data 264 Appendix F Calculations in M0S2 Characterization and Analysis 266 F - l Example of Calculations from X R D Spetrum 266 F-2 MoS 2 Slab Length Calculations 267 F-3 Calculation of M0S2 Edge Sites, Rim Sites and Mo Dispersion 268 Appendix G Algorithm for Gaussian Newton-Raphson Parameter Estimation 271 ix List of Tables Table 1.1 Typical heavy oil and conventional light crude oil properties (Gray, 1994) 2 Table 3.1 Properties of different feedstock 49 Table 3.2 Test conditions 55 Table 4.1 Summary of M0S2 catalyst properties 63 Table 4.2 Cold Lake heavy oil hydrocracking using different M0S2 catalysts (600 ppm Mo, 415°C, initial H2 pressure of 3.5 MPa, reaction time of 1 hour) 69 Table 4.3 Effect of temperature on Cold Lake heavy oil hydrocracking using exfoliated M0S2 (600 ppm Mo, initial H2 pressure 3.5 MPa, reaction time of 1 hour) 73 Table 4.4 Effect of Mo concentration on Cold Lake heavy oil hydrocracking using exfoliated MoS 2 (415°C, initial H2 pressure of 3.5 MPa, reaction time of 1 hour). 75 Table 4.5 Effect of M0S2 concentration in water suspension during Cold Lake heavy oil hydrocracking using exfoliated MoS 2 (600 ppm Mo, 415°C, initial PL: pressure of 3.5 MPa, reaction time of 1 hour) 78 Table 4.6 Effect of solvent used to disperse exfoliated M0S2 on Cold Lake heavy oil hydrocracking (900 ppm Mo, 415°C, initial H 2 pressure of 3.5 MPa, reaction time of 1 hour) 80 Table 4.7 Different feedstock hydrocracking using exfoliated M0S2 (600 ppm Mo, 415°C, initial H2 pressure of 3.5 MPa, reaction time of 1 hour) 83 Table 4.8 Effect of feedstock in hydrocracking using exfoliated M0S2 (600 ppm Mo, 415°C, initial H2 pressure of 3.5 MPa, reaction time of 1 hour) 85 Table 4.9 BET surface area of coke recovered from different hydrocracking reactions (initial H2 pressure of 3.5 MPa and a reaction time of 1 hour) 87 Table 4.10 Ratio of Mo and V in coke recovered from different hydrocracking reactions (initial H2 pressure of 3.5 MPa and a reaction time of 1 hour) 92 Table 5.1 BET surface area of different catalysts recovered after reaction. (600 ppm Mo, 350°C, initial pressure 2.8 MPa, 5-hour reaction) 101 Table 5.2 Average stack height and average number of layers in crystallites from unused and recovered exfoliated M0S2 102 Table 5.3 Summary of M0S2 catalyst properties 109 Table 5.4 Comparison of hydrogenation of naphthalene using different M0S2 catalysts (600 ppm Mo, 350°C, initial pressure 2.8 MPa, 5-hr reaction time) 112 Table 5.5 Pseudo 1 s t order rate constant for hydrogenation of naphthalene at different temperature (initial pressure 2.8 MPa, 5-hr reaction time) 116 Table 5.6 Apparent activation energy corresponding to conversion of naphthalene to tetralin 116 Table 5.7 Comparison of hydrodesulfurization of DBT using different M0S2 catalysts (600 ppm Mo, 350°C, initial pressure 2.8 MPa, 5-hr reaction time) 122 Table 5.8 Estimated 1 s t order rate constants for the hydrodesulfurization of DBT (Figure 5.11) 127 Table 5.9 Comparison of hydrodenitrogenation of quinoline using MoNaph derived M0S2 and exfoliated M0S2 (600 ppm Mo, 350°C, initial pressure 2.8 MPa, 5-hr reaction time) 134 xi Table 5.10 Comparison of hydrodenitrogenation of carbazole using MoNaph derived M0S2 and exfoliated MoS 2 (600 ppm Mo, 350°C, initial pressure 2.8 MPa, 5-hr reaction time) 140 Table 5.11 Fitted rate constant for hydrodenitrogenation of carbazole using MoNaph derived MoS 2 and exfoliated MoS 2 at 350°C 148 Table 5.12 Comparison for hydrodeoxygenation of phenol using different MoS 2 catalysts (600 ppm Mo, 350°C, initial pressure 2.8 MPa, 5-hr reaction time) 151 Table 5.13 Summary of M o S 2 catalyst properties 154 Table 5.14 Dispersion of MoS 2 catalysts 156 Table 5.15 Rate constant and initial turn over frequency (TOF) for hydrodesulfurization of DBT over different MoS 2 catalysts at 350°C 157 Table 5.16 Rate constant and initial turn over frequency (TOF) for hydrodenitrogenation of carbazole over different MoS 2 catalysts at 350°C 158 Table 5.17 Turn over frequency for hydrodeoxygenation of phenol over different MoS 2 catalysts at 350°C 159 Table 5.18 Comparison of reactant conversion and heteroatom removal in different model reactions (600 ppm Mo, 350°C, initial pressure 2.8 MPa, 5-hr reaction time). .163 Table B - l Example of an experimental sheet 183 Table B-2 Summary of results of hydrocracking activity experiments 186 Table B-3 Liquid product analysis 188 Table B-4 Repeated experiments 189 Table B-5 Hydrocracking reaction with the reaction condition in Table B-4 189 Table B-6 Relative standard error % estimated for hydrocracking reactions 190 Table B-7 Response factor for gas components 192 xii Table B-8 Heteroatom, M C R and asphaltene in different heavy oil feedstock 201 Table B-9 Heteroatom removal, and M C R and asphaltene conversion in different heavy oil feedstock after catalytic hydrocracking reaction using exfoliated M0S2 (600 ppm Mo, 415°C, initial pressure of 3.5 MPa, 1-hour reaction time) 201 Table C - l Reactant conversion and heteroatom removal using exfoliated M0S2 in water (600 ppm Mo, 350°C, initial pressure 2.8 MPa, 5-hr reaction time) 204 Table C-2 Reactant removal and heteroatom removal using MoNaph with presence of 5-6% of water (600 ppm Mo, 350°C, initial pressure 2.8 MPa, 5-hr reaction time). ..205 Table D - l Summary of experiments for model compound hydroprocessing reactions (600 ppm Mo i f M0S2 catalyst was used, initial pressure 2.8 MPa, 5-hr reaction). ..207 Table D-2 cont'd 208 Table D-3 Results of hydrogenation of naphthalene 211 Table D-4 Results of hydrodesulfurization of DBT 213 Table D-5 Results of hydrodeoxygenation of phenol 214 xiii List of Figures Figure 2.1 Formation of coke from complex aromatics such as asphaltene (Speight, 1991). 14 Figure 2.2 Mechanism of hydrocracking (Kennepohl and Sanford, 1996) 23 Figure 2.3 A simplified scheme for catalyst recycling/regenerations of the slurry process (Del Bianco et a l , 1995) 30 Figure 2.4 Crystalline structure of M0S2 (a) overall structure; (b) (001) plane; (c) (100) plane and (d) (110) plane (Shimada, 2003) 35 Figure 2.5 Reaction network of dibenzothiophene (Daage and Chianelli, 1994) 37 Figure 2.6 Rim-edge model of a MoS 2 catalytic particle (Daage and Chianelli, 1994) 38 Figure 2.7 Schematic drawing of the catalyst structure: (Left) catalyst decomposed at lower temperatures and (Right) catalyst decomposed at higher temperatures. Solid arrows: hydrogenation active sites; dotted arrows: edge sites except rim sites (Iwataet al., 2001) 39 Figure 3.1 Schematic diagram of the slurry reactor setup 51 Figure 4.1 X-ray diffraction pattern of (a) untreated crystalline MoS 2 ; (b) MoS 2 after exfoliation dispersed in water; (c) MoS 2 after exfoliation and redispersed in decalin 62 Figure 4.2 T E M photo of (a) crystalline MoS 2 (b) exfoliated MoS 2 65 Figure 4.3 Slab length distributions for (a) crystalline MoS 2 (b) exfoliated MoS 2 66 Figure 4.4 Comparisons between coking, thermal cracking, catalytic hydrocracking using MoNaph and exfoliated MoS 2 (Reaction conditions: 430°C, initial H 2 pressure of 3.5 MPa, reaction time of 1 hour, 600 ppm Mo i f MoS 2 catalyst was used) 68 xiv Figure 4.5 Cold Lake heavy oil hydrocracking using M0S2 derived from MoNaph and exfoliated M0S2 (600 ppm Mo, 415°C, initial H 2 pressure of 3.5 MPa, reaction time of 1 hour) 71 Figure 4.6 Coke, liquid, gas yield and asphaltene conversion at different Mo concentrations and reaction temperatures in Cold Lake heavy oil hydrocracking using exfoliated M0S2 (initial H2 pressure of 3.5 MPa, reaction time of 1 hour) 76 Figure 4.7 Effect of decalin and water as solvent to disperse exfoliated M0S2 in Cold Lake heavy oil hydrocracking liquid product oil (900 ppm Mo, 415°C, initial H2 pressure of 3.5 MPa, reaction time of 1 hour) 80 Figure 4.8 Coke yield versus feedstock with different M C R in hydrocracking using exfoliated M0S2. (Reaction Conditions: 600 ppm Mo, 415°C, initial H 2 pressure of 3.5 MPa, reaction time of 1 hour; trend line) 84 Figure 4.9 X R D diffractograms for the coke recovered from catalytic hydrocracking of heavy oil with (a) 600 ppm Mo of exfoliated MoS 2 at 415°C; (b) 600 ppm Mo of MoNaph at 415°C; (c) 900 ppm Mo of exfoliated MoS 2 dispersed in decalin at 415°C; (d) 900 ppm Mo of exfoliated MoS 2 at 415°C and (e) 600 ppm Mo of exfoliated M0S2 at 430°C (initial H2 pressure of 3.5 MPa and a reaction time of 1 hour) 88 Figure 4.10 Infrared spectra for the coke recovered from catalytic hydrocracking of heavy oil with (a) 600 ppm Mo of exfoliated MoS 2 at 430°C; (b) 900 ppm Mo of exfoliated M o S 2 at 415°C; (c) 600 ppm Mo of exfoliated MoS 2 at 415°C; (d) 600 ppm Mo of exfoliated MoS 2 at 400°C; (e) retentate with 600 ppm Mo of exfoliated M o S 2 at 415°C; (f) 360 ppm Mo of exfoliated MoS 2 at 415°C (g) filtrate with 600 ppm Mo of exfoliated MoS 2 at 415°C; (h) 900 ppm Mo of xv exfoliated MoS 2 in decalin at 415°C; and (i) 600 ppm Mo of MoNaph at 415°C (initial H 2 pressure of 3.5 MPa and a reaction time of 1 hour) 90 Figure 4.11 S E M of coke recovered from hydrocracking using exfoliated MoS 2 . 93 Figure 4.12 Energy dispersive x-ray (EDX) mapping of coke collected after hydrocracking reaction 94 Figure 4.13 Activity of recycled MoS 2 during Cold Lake heavy oil hydrocracking (600 ppm Mo, 415°C, initial H 2 pressure of 3.5 MPa and a reaction time of 1 hour) 95 Figure 5.1 X R D diffractograms of the solid recovered after reactions using MoNaph precursor; (a) reaction without reactant with only solvent; (b) hydrogenation (c) hydrodenitrogenation and (d) hydrodesulfurization reactions (600 ppm Mo, 350°C, initial pressure 2.8 MPa, 5-hour reaction) 103 Figure 5.2 X R D diffracto grams of the solid recovered after hydrodesulfurization reaction using A H M precursor (600 ppm Mo, 350°C, initial pressure 2.8 MPa, 5-hour reaction) 103 Figure 5.3 X R D diffractograms of the solid recovered after reactions using exfoliated MoS 2 . (a) hydrodesulfurization (b) hydrogenation and (c) hydrodenitrogenation reactions (600 ppm Mo, 350°C, initial pressure 2.8 MPa, 5-hour reaction) 104 Figure 5.4(a) T E M of exfoliated MoS 2 recovered after reaction (hydrodesulfurization reaction at 350°C, initial pressure 2.8 MPa, 5-hour reaction 105 Figure 5.4 T E M of (b) MoNaph and, (c) A H M derived MoS 2 recovered after reaction (hydrodesulfurization reaction at 350°C, initial pressure 2.8 MPa, 5-hour reaction) 106 xvi Figure 5.5 Distribution of MoS 2 slab length for (a) MoNaph and (b) A H M derived MoS 2 reacovered after reactions (hydrodesulfurization reaction at 350°C, initial pressure 2.8 MPa, 5-hour reaction) 107 Figure 5.6 Distribution of the number of layers for MoNaph and A H M derived MoS 2 . (hydrodesulfurization reaction at 350°C, initial pressure 2.8 MPa, 5-hour reaction) 108 Figure 5.7 Naphthalene reaction scheme 110 Figure 5.8 Profile of product ratio using (a) exfoliated MoS 2 and (b) MoNaph derived MoS 2 in hydrogenation of naphthalene (600 ppm Mo, 350°C, initial pressure 2.8 MPa, 5-hr reaction time; trend line) I l l Figure 5.9 Plots of ln([NTLo]/[NTL]) versus time for hydrogenation of naphthalene with (a) no catalyst (thermal reaction); (b) MoNaph; and (c) exfoliated MoS 2 (600 ppm Mo, 350°C, initial pressure 2.8 MPa, 5-hr reaction time; fitted line) 114 Figure 5.10 Plots of In k versus 1000/7 for hydrogenation of naphthalene with (a) no catalyst (thermal reaction); (b) exfoliated MoS 2 and (c) MoNaph ( fitted line). 117 Figure 5.11 Reaction network for hydrodesulfurization of dibenzothiophene 118 Figure 5.12 Profile of product ratio using (a) crystalline MoS 2 and, (b) exfoliated MoS 2 in hydrodesulfurization of DBT (600 ppm Mo, 350°C, initial pressure 2.8 MPa, 5-hr reaction time; trend line) 120 Figure 5.13 Plots of ln([DBTo]/[DBT]) versus time for hydrodesulfurization of DBT with (a) crystalline MoS 2 ; (b) exfoliated MoS 2 ; (c) MoNaph; and (d) A H M derived MoS 2 (600 ppm Mo, 350°C, initial pressure 2.8 MPa, 5-hr reaction time; fitted line) 124 xvn Figure 5.14 Products yield versus conversion of DBT for (a) crystalline M0S2, (b) exfoliated MoS 2 , (c) MoNaph and (d) A H M derived MoS 2 (600 ppm Mo, 350°C initial pressure 2.8 MPa, 5-hr reaction time; fitted model) 125 Figure 5.15 Hydrodesulfurization versus conversion of DBT (600 ppm Mo, 350°C, initial pressure 2.8 MPa, 5-hr reaction time; fitted model) 130 Figure 5.16 Plots of In &DBT versus 1000/7" for hydrodesulfurization of DBT using (a) exfoliated M0S2 and (b) MoNaph derived MoS 2 ( fitted line) 131 Figure 5.17 Reaction network for hydrodenitrogenation of quinoline 133 Figure 5.18 Profile of product ratio using MoNaph derived MoS 2 in hydrodenitrogenation of quinoline (600 ppm Mo, 350°C, initial pressure 2.8 MPa, 5-hr reaction time; trend line) 135 Figure 5.19 Profile of product ratio using exfoliated MoS 2 in hydrodenitrogenation of quinoline (600 ppm Mo, 350°C, initial pressure 2.8 MPa, 5-hr reaction time; trend line) 136 Figure 5.20 Hydrodenitrogenation versus conversion of quinoline (600 ppm Mo, 350°C, initial pressure 2.8 MPa, 5-hr reaction time; trend line) 138 Figure 5.21 Plots of ln([QNLo]/[QNL]) versus time for hydrodenitrogenation of quinoline using (a) MoNaph derived MoS 2 and, (b) exfoliated MoS 2 (600 ppm Mo, 350°C, initial pressure 2.8 MPa, 5-hr reaction time; fitted line) 139 Figure 5.22 Plots of ln([CBZo]/[CBZ]) versus time for hydrodenitrogenation of carbazole using (a) MoNaph derived MoS 2 and, (b) exfoliated MoS 2 (600 ppm Mo, 350°C, initial pressure 2.8 MPa, 5-hr reaction time; fitted line) 141 Figure 5.23 Reaction network for hydrodenitrogenation of carbazole 142 xviii Figure 5.24 Profile of product ratio using (a) MoNaph and (b) exfoliated MoS 2 in hydrodenitrogenation of carbazole (600 ppm Mo, 350°C, initial pressure 2.8 MPa, 5-hr reaction time; trend line) 144 Figure 5.25 Products yield versus conversion of carbazole for (a) exfoliated MoS 2 , and (b) MoNaph derived MoS 2 (600 ppm Mo, at 350°C, initial pressure 2.8 MPa, 5-hr reaction time; fitted model) : 146 Figure 5.26 Comparison of hydrodenitrogenation over converted carbazole at 350°C ( fitted model) 147 Figure 5.27 Schematic diagram for hydrodenitrogenation of carbazole used in model fitting. 147 Figure 5.28 Hydrodeoxygenation routes of phenol '. 150 Figure 5.29 Plots of concentration profile for hydrodeoxygenation of phenol using (a) crystalline MoS 2 , (b) exfoliated MoS 2 , (c) MoNaph and (d) A H M derived MoS 2 (600 ppm Mo, 350°C, initial pressure 2.8 MPa, 5-hr reaction time; fitted line) 152 Figure 5.30 MoS 2 single layer slab model used in edge site calculation, n = number of edge side Mo; L- diagonal length of the hexagon 155 Figure 5.31 Ratio of k\lki and fcj/fe versus relative rim sites and edge sites for hydro-desulfurization of DBT ( trend line) 161 Figure C - l (a) Exfoliated M o S 2 in suspension in water (b) Exfoliated M o S 2 in the presence of organic solvent. A cross sectional view of the single M o S 2 layers, in which C represents an adsorbed organic molecule and OH" represents an adsorbed hydroxyl group (Divigalpitiya et al., 1989) 206 xix Figure D - l Calibration curves for (a) DBT and (b) quinoline to determine respective response factors 209 Figure E - l Plot of ln([DBT 0]/[DBT]) versus time in HDS of DBT using exfoliated MoS 2 at 350°C 258 Figure E-2 Concentration profile of phenol versus time in HDO reactions using exfoliated MoS 2 at350°C 260 Figure E-3 Plot of In k D B T versus 1000/T in HDS of DBT using MoNaph derived MoS 2.261 Figure E-4 Plots of experiment data versus simulated data for conversion of DBT (•) and yields of products (THDBT: A; BP:x; CHB+BCH:>K) in HDS of DBT using (a) crystalline MoS 2 ; (b) exfoliated MoS 2 ; (c) MoNaph derived MoS 2 ; (d) A H M derived M o S 2 at 350°C. 264 Figure E-5 Plots of simulated data versus experiment data for conversion of carbazole (•) and yields of products (THCBZ: A; CHB:x; CHCHe:«; gBCH:«) in hydrodenitrogenation of carbazole using (a) exfoliated MoS 2 and, (b) MoNaph derived MoS 2 at350°C. 265 Figure F - l X R D spectrum for peak corresponding to (002) plane in exfoliated MoS 2 dispersed in water 266 Figure G - l Algorithm for Gaussian Newton used in rate constant estimation 271 xx List of Abbreviations A H M Ammonium heptamolybdate A T T M Ammonium tetrathiomolybdate B C H Bicyclohexyl BET Brunauer-Emmett-Teller BP Biphenyl CBZ Carbazole CCR Conradson carbon residue CHB Cyclohexylbenzene CHCHe Cyclohexyl cyclohexene C P M C H Cyclopentylemethyl-cyclohexane DBT Dibenzothiophene D C L Decalin DHQ Decahydroquinoline DRIFT Diffuse Reflectance Infrared Fourier Transform Spectroscopy E D X Energy dispersive x-ray emission FID Flame ionization detector H C H Hexylcyclohexane H D N Hydrodenitrogenation HDO Hydrodeoxygenation HDS Hydrodesulfurization H Y D Hydrogenation XXI LHSV Liquid hourly space velocity M C R Microcarbon residue MoDTC Molybdenum dithiocarboxylate MoDTP Molybdenum dithiophosphate MoNaph Molybdenum naphthenate MoS 2 Molybdenum disulfide NCUT National Centre for Upgrading Technology N M M B Naphthylmethyl bibenzyl N T L Naphthalene OPA o-propylaniline OPCHA o-propylcyclohexylamine PB Propylbenzene PCH Propylcyclohexane PCHe Propylcyclohexene PHCBZ Perhydrocarbazole QNL Quinoline SCFE Supercritical fluid extraction S E M Scanning electron microscopy TCD Thermal conductivity detector T E M Transmission Electron Microscopy THCBZ Tetrahydrocarbazole THDBT Tetrahydrodibenzothiophene THQ Tetrahydroquinoline xxii TOF Turnover frequency TTL Tetralin X R D X-ray diffraction Acknowledgement First and foremost, I would like to thank my supervisor, Professor Kevin J. Smith for his guidance and support throughout my research. I would also like to thank Dr. Changchun Y u from the University of Petroleum, China for his help and inspiration during his stay at UBC. Special thanks also go to my thesis committee members Professor C. Jim Lim, Professor Tom Troczynski and Professor Madjid Mohseni for their advice and suggestions in completion of this thesis. My sincere thanks to Mary Mager from the Department of Metals and Materials Engineering and Lina Mandilao from the Wine Research Centre for their help in sample characterization. My appreciation also goes to the staff in the office, stores and workshop of Chemical & Biological Engineering for their cooperation and help. The fellowship from University Science of Malaysia for my Ph.D. study at UBC is greatfully acknowledged. I also thank Professor Subhash Bhatia and Professor Abdul Rahman Mohamad for their encouragement and support. I thank my parents, sisters, brothers and friends for their love and support throughout my study period. Lastly, I thank my colleagues in the catalysis group for their help. xxiv ^ ^ ^ ^ ^ ^ ^ ^ ^ s = = Chapter 1 = = = ^ = = = = = = = = ^ = Introduction Due to increasing world energy demand and decreasing conventional light crude oil reserves, mankind is constantly searching for alternative energy sources. Heavy crude oil can be converted into usable light oil and it is available in large reserves. Therefore, heavy oil has been predicted to be one of the major energy sources of the 21 s t century (Lanier, 1998; Moritis, 2004). In 2002, the Alberta Energy and Utilities Board reported that the largest known oil sand deposit on earth is in Alberta, Canada with an estimated 27.7 billion cubic meters of oil sands. It is foreseen that increasing quantities of heavy and non-conventional crude oils will need to be refined in the future as more oil sands deposits are mined to meet the energy demands of the world. The typical properties of heavy and light crude oil are listed in Table 1.1. Compared to conventional light crude oil, heavy oils are much more viscous and have lower hydrogen to carbon (H/C) ratio, higher concentrations of heteroatoms (S, N and O) and metals (Ni and V), and higher liquid volume of residue with boiling point (B.P.) higher than 525°C. A l l of the above properties pose challenges to existing oil refining technology for the conversion of heavy oils to usable light crude oil. Furthermore, stringent environmental regulations increase pressure on the refining industry to produce cleaner fuels, despite the use of a lower grade feedstock, heavy oil. To address the challenges mentioned above, the oil industry is continuously looking for better upgrading technologies that can maximize the production of commercially useful fractions, such as naphtha, kerosene, and gas oil from these heavy oils. 1 Table 1.1 Typical heavy oil and conventional light crude oil properties (Gray, 1994) Characteristic Heavy oil Light crude oil A P I * <20 38 H/C atomic ratio 1.4-1.6 1.8 Sulphur content, wt % 2-7 0.5 Nitrogen, wt % 0.2-0.7 0.1 Metals, ppm 100-1000 22 Viscosity, m2/s x 106at 40°C ~> 5000 5 Residue, B.P.>525°C, wt % >45 11 * A gravity unit (weight per unit of volume) of crude oil defined by American Petroleum Institute. The purpose of heavy oil upgrading is to convert or to crack the high molecular weight residue into distillable components, to increase the H/C ratio to 1.8 mol/mol needed for transportation fuels (Gray, 1994), and to reduce the heteroatoms and metals to acceptable levels. The two important processes used in heavy oil upgrading are primary upgrading and hydroprocessing. The primary upgrading step is where carbon-carbon bonds are broken. This step aims to maximize the percentage conversion of the residue to lighter products. Hydroprocessing or hydrotreating is a second step that aims to further upgrade the liquid oil, especially in heteroatom removal, such that the B.P. of the feed does not change appreciably. Coking and hydrocracking are two primary upgrading approaches used in the petroleum industry. Coking is a relatively economical, proven technology. However, coking produces large amounts of unmarketable coke. On the other hand, hydrocracking, 2 which is the focus of the present study, provides higher liquid yield and a better quality liquid product compared to coking. Hydrocracking processes often operate in the presence of a catalyst. The catalysts used are typically CoMo sulfides or NiMo sulfides supported on alumina. One of the operational problems for hydrocracking processes is that the catalysts are deactivated rapidly by the coke and heavy metals that are deposited during the process. Therefore, the use of conventional fixed-bed technologies is limited for common hydrocracking processes by the choice of feedstock and the severity of operation. There are other technologies, such as H-Oil and LC-Fining processes, which are able to process heavy oil at high severity by using ebullated bed technology. However, they are mechanically complex and use large amounts of expensive catalysts. In addition, the disposal of the spent catalyst may have significant environmental and economical impacts. The above-mentioned issues have led to the development of slurry reactor technology with the use of unsupported, dispersed catalysts for hydrocracking. In slurry processes, the active catalyst is slurried or suspended in the oil. The slurry technology combines the flexibility of coking with the high performance of hydrocracking. In slurry processes, a large variety of feedstocks can be processed with only minor changes in the process variables. Typically, finely dispersed catalysts in very low concentrations (0.05-0.3 wt%) are used in a slurry reactor to reduce the severity of the process and to improve the quality of the product (Del Bianco et al., 1995). Dispersed catalysts have a very small particle diameter (<1 um) as compared to that in ebullating bed (~ 0.8 mm) or trickle fixed bed (~ 1.5 mm) reactors (Dautzenberg and De Deken, 1984). The degree of dispersion of the catalysts is very important in slurry process as it strongly affects the catalyst activity. A well-dispersed catalyst should favour the rapid uptake of hydrogen, 3 which prevents free radical condensation among the heavy oil molecules that would lead to coke formation. A high degree of catalyst dispersion can be achieved by introducing water-or oil-soluble catalyst precursors to the feed (Panariti et al., 2000a). It is generally agreed that molybdenum based catalysts give the best performance in hydrocracking and hydroprocessing. M0S2 based catalysts are extensively used in commercial refineries. The application of dispersed M0S2 in heavy oil hydrocracking has been reported in a number of studies (Peureux et al., 1995; Rueda et al., 1997; Panariti et al., 2000a; Panariti et al., 2000b). Dispersed M0S2 catalysts are mostly based on oil-soluble organometallics such as molybdenum naphthenate (MoNaph), molybdenum dithiocarboxylate and molybdenum dithiophosphate. The active form of catalyst, M0S2, is generated in situ by thermally decomposing the precursor and reacting with sulfur. The sources of sulfur include sulfur originally present in the feedstock, or externally added suitable sulfur-containing compounds such as H 2 S and elemental sulfur (Kim et al., 1989). The M0S2 generated in situ is in the form of micrometer-sized or nanometer-sized particles. Although many studies focusing on the technical aspects of slurry processes have been reported over the past few years, there is no commercial slurry process operating in industry. This is partly due to the high cost of M0S2 catalyst, despite the use of low catalyst concentration. To tackle this problem, one approach is to further reduce the required amount of catalyst concentration (< 200 ppm) for the process while maintaining the overall hydrocracking performance. In this regard, a better understanding of the fundamental chemistry of the dispersed catalyst system is essential for catalyst performance enhancement and commercialization. In addition, the recycle of M0S2 catalyst is an important issue in catalyst development for an economic slurry hydrocracking process (Del Bianco etal., 1995). 4 On the other hand, the use of M0S2 based catalysts for hydrodesulfurization (HDS), hydrodenitrogenation (HDN) and hydrodeoxygenation (HDO) is well established in commercial hydroprocessing. Understanding the relationship between active sites and the structure of M0S2 catalysts is important when trying to improve catalyst performance and a number of studies have been reported on structure-activity relationships for M0S2 catalysts (Topsoe and Clausen, 1984; Hensen et al., 2001; Schweiger et al., 2002). Two types of active sites have been identified: one that promotes hydrogenation reactions and a second that promotes hydrogenolysis of the carbon-heteroatom bond of heterocyclic organic compounds present in the oil. The activity of M0S2 is generally accepted to be due to the edge planes of MoS 2 , whereas the basal plane is considered inert (Roxlo et al., 1986). However, the relationship between M0S2 structure and active sites for hydrogenolysis and hydrogenation during hydrodesulfurization has been somewhat controversial. For instance, using unsupported M0S2 crystallites, prepared with a different number of M0S2 layers, Daage and Chianelli (1994) proposed a rim-edge model in which hydrogenation of dibenzothiophene (DBT) to tetrahydrodibenzothiophene (THDBT) occurred exclusively on the top and bottom edge planes (rim sites) of M0S2 crystallites while the hydrogenolysis of DBT to biphenyl (BP) was catalyzed on all of the edge planes. On the other hand, Hensen et al. (2001) found that the effect of the number of M0S2 layers on selectivity depended on the reactant molecular size. Increasing M0S2 layer stacking increased the hydrogenation reaction as well as hydrogenolysis in the hydrodesulfurization of DBT, whereas hydrogenolysis in the hydrodesulfurization of thiophene was not a strong function of the number of M0S2 layers. Most of the studies related to M0S2 structure-activity relationships were carried out using model compounds and focused on hydrodesulfurization and hydrogenation reactions, 5 rather than hydrodenitrogenation and hydrodeoxygenation reactions. Furthermore, there is limited data on the effect of structure of M0S2 in heavy oil hydrocracking in a slurry process. M0S2 is a layered structure in which sheets of M0S2 are held together by weak van der waals forces between each layer. Exfoliation (Joensen et al., 1986) is a chemical treatment that transforms M0S2 crystallites into highly dispersed, single-layered M0S2 suspended in aqueous solution. Exfoliated M0S2 offers an opportunity to examine the M0S2 structural effects both in the heavy oil slurry hydrocracking process and in model compound reactions. In principle, layers of exfoliated M0S2 can be considered as a model of M0S2 rim sites suited for the study of the structure-activity relationships, independent of support or promoter effects. In previous studies, Bockrath and Parfitt (1995a) reported no significant differences were observed in hydrodesulfurization and hydrodenitrogenation of Hondo residue during hydrocracking with exfoliated and crystalline M0S2. An attempt by Del Valle et al. (1998) to compare the hydrodesulfurization activity of an unsupported M0S2 catalyst before and after exfoliation using DBT as reactant was not successful. Del Valle et al. (1998) suggested that the exfoliated catalysts were poisoned by the lithium used in the exfoliation process. In the present study, catalyst activity of exfoliated M0S2 was tested for both hydrocracking and hydroprocessing reactions. The study is conveniently divided into two parts. The first part of the study examines the effectiveness of exfoliated M0S2 as a dispersed catalyst for Cold Lake heavy oil hydrocracking. The second part of the study focused on the effects of M0S2 structure on hydrodesulfurization, hydrodenitrogenation, hydrodeoxygenation and hydrogenation reactions using model compounds as reactants. 6 In the first part of the study, the effects of catalyst properties and process conditions on the hydrocracking reactions in the presence of exfoliated M0S2 were studied. An investigation of catalyst recovery and recycle is included in this first part of the study. Although the catalyst may be used on a once-through basis, the possibility of catalyst recycle has been considered because of issues associated with catalyst cost, especially i f the Mo concentration in the feed oil is above 200 ppm (Del Bianco et al., 1994). In the second part of the study, the activity of exfoliated M0S2 catalyst is compared with M0S2 catalysts prepared by decomposition of oil soluble precursors, mainly MoNaph. The structure of M0S2 produced in this way is very different to that obtained by exfoliation. This may result in different catalyst activity. 1.1 Objectives The objectives of the present study are: 1. To verify the effectiveness of unsupported exfoliated M0S2 in the slurry hydrocracking process by comparing it to coking, thermal cracking and catalytic hydrocracking using MoNaph. 2. To study the effect of process conditions i.e. reaction temperature and Mo concentration on the hydrocracking reactions in the presence of exfoliated M0S2. 3. To study the effect of exfoliated MoS 2 properties, i.e. Mo concentration in the dispersion and dispersing solvent, on the hydrocracking reactions in the presence of exfoliated M0S2. 4. To evaluate the benefit of using different petroleum fractions as feed in the hydrocracking reactions with the presence of exfoliated M0S2. 7 5. To investigate the feasibility of spent catalyst recycling. 6. To study the effect of M0S2 structure on various hydroprocessing reactions using exfoliated M0S2 and model reactants including naphthalene for hydrogenation, dibenzothiophene (DBT) for hydrodesulfurization, quinoline and carbazole for hydrodenitrogenation, and phenol for hydrodeoxygenation. 7. To correlate the relationship between hydrogenation and hydrogenolysis in the model reactants to M0S2 rim sites and edge sites. 8 ™= Chapter 2 = Literature Review Chapter 2 is presented in three sections. The first part gives a brief overview and background to heavy oil upgrading and introduces important technical terminology. The second section is concerned with primary upgrading and includes a description of current industrial technology and an overview of recent developments in heavy oil hydrocracking using slurry reactor and dispersed metal sulfide catalysts. The third section deals with a review of hydroprocessing reactions, focusing on unsupported M0S2 catalyst structure-activity relationships. 2.1 Terminology and Background 2.1.1 Heavy Oil and Residues Crude oils that are used as the feedstock from which refineries produce a wide range of petroleum products, can be. categorized into conventional crude oil and heavy oil. Heavy oils are derived from various sources including tar sands, coal liquids and shale oils. Heavy oils have low hydrogen-to-carbon (H/C) atomic ratio ranging from 1.4 to 1.6 whereas the ratio in conventional crude oil is about 1.8 (Schumacher, 1982). Heavy oils have high viscosity (>102 mPa.s at 15.6°C) (Gray, 1994) with generally high levels of organic bonded sulfur, nitrogen and metals. In addition, heavy oils usually contain more than 45 wt% of residue, which makes them unsuitable for conventional refining. 9 Residue refers to the bottom product from a crude oil vacuum distillation column with an atmospheric equivalent boiling point (B.P.) above 525°C. Residues are fractions consisting mainly of polycyclic aromatic molecules, heteroatoms and asphaltene. Generally, heavy oil is defined as crude oil or petroleum residue with gravity below 20 API (Lanier, 1998). 2.1.1.1 Asphaltene Asphaltene, by definition, is a solubility class. Asphaltene is defined as all material in petroleum that is soluble in benzene or toluene and insoluble when treated in a large excess of heptane or pentane solvent (Gray, 1994). Asphaltene is believed to consist of condensed polynuclear aromatic rings with alkyl side chains and heteroatoms such as N , S, O, N i and V (Speight, 1991). The molecular mass of asphaltene has been reported to range from 2000 ± 500 Da (Wu et al., 1998). The lightest fraction of asphaltene is the most aromatic and has the lowest concentration of long side chains. The heaviest fraction has the lowest aromatic carbon fraction and the most aliphatic side chains. Asphaltene exists in crude oils as micelles that result from the interaction between asphaltene and resin (Gray, 1994; Andersen and Speight, 1999). Asphaltenes are responsible for the high viscosity of residue and act as coke precursors during heavy oil hydrocracking (Speight, 1991). Relationships between the asphaltene content of oil and other parameters have been widely studied. The asphaltene content was reported to be linearly related to the Conradson Carbon Residue (CCR) content. In addition, the gas yield and the residue conversion in heavy oil hydrocracking were also linearly related to the C C R conversion (Trasobares et al., 1998; Trasobares et al., 1999). 10 The reaction of the asphaltene fraction is important in hydrocracking of heavy oil. A key goal is to have the asphlatene crack to smaller molecules to increase the distillates in the product oil and to remove heteroatoms and metals. De-asphaltening is found to parallel V and N i removal (Callejas and Martinez, 2000). 2.1.1.2 Heteroatoms and Metals Heteroatoms are elements in crude oil other than carbon and hydrogen. In heavy oils, the heteroatoms S, N , O and heavy metals, especially V and N i are present. A typical concentration range in heavy oil is S: 2 - 7 wt%, N : 0.2 - 0.7 wt%, O: ~1 wt %, V : 100-1000 ppm and N i : 20 - 200 ppm (Gray, 1994). Sulfur is present in heavy oil as thiophene homologs and sulfides. Thiophene homologs can be resistant to further processing while sulfides are more easily removed. Nitrogen is present in two forms in heavy oil: nonbasic derivatives of pyrrole and basic derivatives of pyridine. Both types of nitrogen are highly resistant to removal. Oxygen is found as furan homologs, ethers and phenols. In heavy oil, metals exist as porphyrin metals and nonporphyrin metals. Porphyrin metals refer to metal chelated in porphyrin structures analogous to chlorophyll. Nonporphyrin metals are believed to be associated with the polar groups in the asphaltene. It is necessary to remove heteroatoms and metals since S, N and metals deactivate catalysts and O causes polymerization reactions that lead to gum formation. 11 2.1.2 Synthetic Crude Oil Heavy oil derived from different sources can be upgraded into a marketable and transportable product. These products, although varied in nature, are a mixture of hydrocarbons that resemble conventional crude oil and hence are termed synthetic crude oil (Speight, 1991). The synthetic crude oil that is produced from special upgrading processes can be refined in conventional refinery processes. 2.1.3 Upgrading and Reactions The purpose of heavy oil upgrading is to convert unconventional crude oil into synthetic crude oil that can be processed in a conventional refinery. Typically, there are two major stages in heavy oil upgrading, namely primary upgrading and hydroprocessing or secondary upgrading. Primary upgrading is the first reaction stage in the conversion of heavy oil and residue into distillates i.e. products with boiling point lower than 525°C. Primary upgrading involves massive carbon-carbon bond breaking that occurs under severe reaction conditions (400-550°C). The product oils from primary upgrading are further processed through hydroprocessing reactions to improve their quality, without significantly altering the boiling point of the oil. The hydroprocessing reactions focus on heteroatom removal at mild conditions compared to primary upgrading. Normally, catalysts are used in secondary hydroprocessing to provide selectivity for the desired reactions. The catalysts remove heteroatoms through hydrogenation and hydrogenolysis reactions. The chemistry involved during heavy oil upgrading is very complex. The following sections introduce the reactions occuring during heavy oil upgrading. 12 2.1.3.1 Thermal Cracking Thermal cracking reactions involve carbon-carbon bond rupture that is thermodynamically favoured at high temperature (>350°C). Thus, it is a phenomenon by which large oil molecules are converted into smaller molecules with lower boiling point. Thermal cracking is mainly due to free radical chain reactions. A free radical is an atom or group of atoms possessing an unpaired electron. For example, the chain reactions for M as parent compound, and R* as any smaller radical: Initiation: M -»• 2 R- (2.1) Propagation: Chain transfer R» + M —• R H + M» (2.2) p scission M« -+ R* + RCH=CH 2 (2.3) Termination: R* + R» —»• tar, coke, others (2.4) Reaction (2.3) involves breakage of the carbon-carbon bond P to the radical center to form an olefin and a smaller radical. During thermal cracking, large molecules are decomposed into smaller molecules as shown in reactions (2.1) and (2.2), which are the primary reactions. However, certain products may interact with one another to produce material with even higher molecular weight than the original feedstock. In heavy oil, fractions of the residue and asphaltene are believed to be the precursors of coke. It is generally assumed that the chemistry of coke formation involves radical condensation reactions to produce higher molecular weight, condensed aromatic species. The formation of coke from aromatics is depicted in Figure 2.1. 13 2.1.3.2 Hydrocracking Hydrocracking is a thermal cracking process occurring with hydrogenation. The process of hydrocracking generally occurs under high H 2 pressure (5-14 MPa) and high temperature (>350°C). During hydrocracking, polynuclear aromatics are readily hydrogenated and opened into lower boiling point materials. This suppresses the radical condensation and thereby reduces coke formation. The process is termed catalytic hydrocracking when a catalyst is applied to improve the reaction, otherwise it is normally referred to as thermal hydrocracking. Intermediates • Condensed aromatic I I Coke Alkyl side chains Figure 2.1 Formation of coke from complex aromatics such as asphaltene (Speight, 1991). 14 During hydrocracking, heteroatom removal also occurs simultaneously due to bond breakage. Some heteroatoms are released as gas such as H2S and some reside in the coke, especially the metals. 2.1.3.3 Hydrodesufurization (HDS) Hydrodesulfurization is a process involving a reaction with hydrogen to remove S compounds from crude oil. The catalyst used is normally sulfided C0M0 or N i M o / A i 2 0 3 . During hydrodesulfurization, S is removed as H2S. Examples of hydrodesulfurization reactions are: Hydrodesulfurization is important in heavy oil upgrading. The sulfur content in heavy oil is at least four times higher than that in conventional crude oil. Without proper treatment, S in the oil easily poisons catalysts and the S residue is released as SO2 upon combustion, which creates severe environmental problems. 2.1.3.4 Hydrodenitrogenation (HDN) Hydrodenitrogenation is a process involving hydrogenation to eliminate nitrogen from the crude oil. The presence of N in the oil deactivates the supported sulfided Mo-based catalyst and also inhibits the hydrodesulfurization reaction (Furimsky and Massoth, 2005). During a R-SH +H 2 -»• RH+H 2S Thiophene 15 hydrodenitrogenation reaction, N is removed as ammonia. An example of a hydrodenitrogenation reaction is shown as follows: H 2 Indole . C H 2 N H 2 C H 3 H 2 ,CU2 C H •3 + N H 3 Ethylbenzene 2.1.3.5 Hydrodeoxygenation (HDO) Likewise, hydrodeoxygenation is a process involving reaction with hydrogen to eliminate O from crude oil. A n example of HDO is: OH + H 2 Phenol Benzene + H 2 0 Hydrodeoxygenation is not as important as hydrodesulfurization and hydrodenitrogenation in heavy oil upgrading because O-containing compounds are not as harmful for catalysts and they are present in low concentration in most petroleum crude oil (< 2 wt%). Further, O is released as H 2 0 upon combustion, which does not have an environmental impact. 16 However, the O content can be more than 10 wt % in the case of synthetic crude derived from coal and biomass. Some O-compounds in the oil readily polymerize which induce instability during storage and poor performance upon combustion. 2.1.4 Catalyst Deactivation Catalyst deactivation owing to metal and coke deposition is a major problem in heavy oil hydrocracking processes. There are species within heavy oil that tend to form coke under hydrocracking conditions. Large amounts of coke (15 - 35 wt% of catalyst) accumulate rapidly on the catalyst during reaction. Rapid deposition of coke significantly reduces catalyst activity. Therefore, hydrcocracking reactions are normally operated at high hydrogen pressure and the lowest possible temperature to suppress coke formation and coke deposition on the catalyst. The deposition of N i , V and Fe on the catalyst is a consequence of demetallization reactions during hydrocracking. The amount of metal deposited on the catalyst (30 - 50 wt% of catalyst) can be greater than the amount of coke. The final loss of catalyst activity is usually associated with filling of pores in the catalyst by deposits of metal sulfide (Gray, 1994). 2.2 Primary Upgrading The objectives of primary upgrading include converting high molecular weight components to distillates, improving H/C atomic ratio and removing heteroatoms. Typically, there are two approaches to increase the H/C ratio: carbon-rejection and hydrogen-addition. The 17 choice between carbon-rejection and hydrogen-addition depends on environmental constraints, feedstock flexibility and product target. Carbon rejection processes are not only relatively economical, but are based on proven technology' that has high feedstock flexibility. However, they produce large amounts of undesirable coke, which corresponds to losses in distillable liquid product. On the other hand, the more expensive hydrogen-addition processes provide higher liquid yield and a better quality product compared to carbon rejection. 2.2.1 Carbon Rejection Carbon rejection is normally referred to as coking. The removal of elemental carbon or production of coke is affected by pyrolysis. The process involves heating the heavy oil at near atmospheric pressure without hydrogen addition to temperature > 500°C. A substantial part of the fuel is rejected from the oil in the form of coke. This results in the loss of potential liquid hydrocarbons. Two common coking technologies used in industry are delayed coking and fluid coking (Gray, 1994). 2.2.2 Hydrogen Addition/Hydrocracking Hydrogen is added to the oil molecules in heavy oil upgrading by hydrogen addition. Hydrogen addition processes are operated at a temperature range from 400 to 460°C and hydrogen pressure > 7 - 14 MPa. In heavy oil primary upgrading by hydrogen addition, the thermal mode of cracking dominates the process. Therefore, it is usually termed hydrocracking. Hydrocracking processes are normally carried out in the presence of 18 catalyst with the ultimate goal to mitigate coke formation in order to maximize residue conversion to lighter products. The catalysts used are typically CoMo sulfides or NiMo sulfides supported on Y-AI2O3. Higher liquid yields are obtained from hydrocracking processes, approximately 85 % compared to 70 % for coking (Sanford, 1994). Further, better liquid product quality is obtained, with lower C C R and heteroatom content, compared to coking. Fixed bed and ebullating reactors are the two common commercialized hydrocracking technologies used in upgrading heavy oil (Gray, 1994). Slurry phase reactor technology is under development. 2.2.2.1 Fixed Bed Reactor Fixed bed reactor processes have the advantage of ease of scale up and operation. The reactors operate in a downflow mode, with the liquid feed trickling downward over the solid catalyst cocurrent with the hydrogen gas. The main limitation of the fixed bed reactor is the accumulation of metal on the catalyst when a heavy feed is processed. The deposition of metal on the catalyst deactivates the catalyst and plugs the catalyst bed, which subsequently leads to a pressure build up in the reactor. The length of operation is dictated by the metal-holding capacity of the catalyst. Costly downtime is needed to replace the catalyst once its lifespan is over. Therefore, a fixed bed is not suitable for processing heavy oil with high metal content (Gray, 1994). 2.2.2.2 Ebullated Bed Reactor In an ebullated bed reactor, the catalyst is fluidized and remains suspended by the liquid phase in the reactor. The catalyst pellet diameter is usually < 1 mm. The ebullated bed 19 reactor is suitable for processing heavy oil at severe conditions (high temperature and pressure). H-Oil and LC-Fining are two licensed processes based on ebullated bed technology (Gray, 1994). Using ebullated bed technology, there are no catalyst or bed plugging problems and the liquid-recycle promotes good mixing within the reactor. In addition, fresh catalyst can be added and withdrawn continuously during operation without shutting down the reactor. However, the problem of catalyst deactivation and replacement of costly catalyst is still an important issue. 2.2.2.3 Slurry Phase Reactor The problems facing existing technology in coking and hydrocracking processes has lead to the development of slurry phase reactor technology that uses highly dispersed, unsupported catalyst. In slurry phase reactors, the active catalyst is mixed with the oil. The dispersed catalysts normally have a very small particle diameter (< 1 um) and very low catalyst concentrations in the oil are required. Over the years, different slurry phase processes have been developed and tested in bench-scale or pilot-scale process units. The distinctions among these processes are mostly in the type of catalyst being used as described in the following: - V E B A CombiCracking: developed by V E B A Oel. Using red mud powder, an iron based catalyst, as once through catalyst at a concentration of about 2 wt% (Niemann and Wenzel, 1993). - U - C A N Residcracking: developed in C A N M E T . The catalyst used is an inexpensive iron sulfate additive that is slurried into the feed at 0.5 - 3 wt% of total feed. The catalyst is used once through (Pruden et al., 1993). 20 - M-Coke (micrometallic-coke): developed by Exxon. This process uses phosphomolybdic acid and molybdenum naphthenate (MoNaph) as catalyst precursor. The active phase is formed in situ by thermal decomposition. The catalyst concentration is low (100 - 1000 ppm) due to high dispersion (Bearden and Aldrich, 1981). - High-conversion hydrocracking with homogeneous catalyst [(HC)3]: developed by Alberta Department of Energy. This process uses very low concentrations organometallic catalyst. The liquid product needs to be processed further (Lott and Lee, 2003). 2.2.3 Slurry Phase Process and Dispersed Catalyst Slurry processes combine the flexibility of a carbon rejection process and the high performance of a hydrogen addition process. In slurry processes, a large variety of heavy oil can be processed with only minor changes in process variables. Finely divided metal sulfides are normally used in the slurry reactor to promote hydrogen transfer and to prevent coke formation. This particular type of catalyst is referred to as dispersed catalyst. The potential of using dispersed catalysts, especially in heavy oil hydrocracking, is developing rapidly because of several advantages associated with the use of dispersed catalysts. The catalyst dispersion attainable in the oil is very high so that a very small concentration of catalyst, from 100 to a maximum of 10000 ppm by weight of catalyst metal, is potentially required in the feed. When the amount of catalyst used is sufficiently small (< 200 ppm), catalyst recovery and recycle steps are not essential (Del Bianco et al., 1995). Also, the external catalyst preparation steps involving solid impregnation and drying 21 are not required. Further, the highly dispersed, nano-sized unsupported catalysts can stay reactive much longer than the conventional supported catalysts. The conventional supported catalysts suffer from rapid deactivation due to the formation of surface deposits by coke and metals in heavy oil processing (Del Bianco et al., 1993; Del Bianco et al., 1995). 2.2.3.1 Hydrocracking Mechanism in the Presence of Dispersed Catalyst Heavy oil hydrocracking involves an extremely complicated chemistry due to the complex nature and the unknown molecular structure of different components present in the oil. In hydrocracking, the only feasible means to achieve carbon-carbon bond breakage in residues is via thermal reaction (Gray, 1994). The role of hydrogen was found not so much to stabilize the primary thermal reaction product but to change the course of the reaction to give a different ratio of product. A mechanism which describes the role of catalyst during hydrocracking has been proposed (Sanford, 1994; Sanford, 1995) and is shown in Figure 2.2. According to Sanford (1995), the main role of the catalyst is to activate hydrogen toward reaction with an aromatic carbon radical to give a cyclohexadienyl type intermediate (step III). Once the intermediate is formed, the molecule decomposes through a series of reactions to form distillate and gases. Hydrogen capping reactions (step IV) are not dominant in the hydrocracking mechanism as gaseous hydrogen was reported not incorporated into the residuum fraction. If the CCR forming molecule which is being decomposed contains sulfur in a thiophene type structure, then hydrodesulfurization (HDS) of the liquids could result with the production of hydrogen sulfide as one of the products. 22 With the presence of dispersed metal sulfide catalyst, solid coke formation (step II) would be reduced by capping radical intermediates with hydrogen (step IV). Hydrogen would be incorporated into the unreacted residue. Therefore, using a dispersed catalyst which is good in hydrogen transfer is very important in a slurry phase process. Thermal H , Hydrogenation catalyst 1 (VI) • <-< IC)] Thermal M - < Q I s-CCR H2/Catalyst i: (HI) No H 2 (II) Reactor Solids H Capping (Catalyst) (IV) H + (V) H2S+Gas+Distillate J% ,1 V K U 5 ' ' I U Figure 2.2 Mechanism of hydrocracking (Kennepohl and Sanford, 1996). 2.2.3.2 Molybdenum Based Catalyst Dispersed metal catalysts are most commonly prepared either by the addition of finely divided inorganic powders to the residue, or by the addition of water or oil soluble metal salts to the residue (Del Bianco et al., 1993). In the latter case, the active form of the catalyst is generated in situ as a slurry of micrometer-sized or nanometer-sized particles. The process is accomplished by thermally decomposing the precursor and adding sulfur, 23 either present in the feedstock or added as H2S, elemental sulfur, or suitable sulfur-containing compounds (Kim et al., 1989). Mo-based catalysts were reported to give the best performances in coke suppression and product upgrading compared to Fe, V , N i , Co, Cr, M n and Ti (Del Bianco et al., 1993; Panariti et al., 2000a). Molybdenum precursors that have been tested in hydrocracking include oil-soluble organometallics such as molybdenum naphthenate (MoNaph) (Bearden and Aldrich, 1981), molybdenum dithiocarboxylate (MoDTC), molybdenum dithiophosphate (Watanabe et al., 2002) and water-soluble ammonium phosphomolybdate and ammonium thiomolybdate (Alonso et al., 1998). Several'techniques have been developed to synthesize the micro-sized or nano-sized particles, including laser pyrolysis, microemulsions, and others. Using microemulsions is one technique to prepare dispersed catalysts. A range of colloidal metals and metal compounds, dispersed in organic solvents, were prepared in the reversed micelles of microemulsions, and the particle size of the colloid was manipulated by changing the composition of the microemulsions (Hall et al., 1998;Duangchan, 1998). Panariti et al. (2000a) studied a series of factors that might affect the catalyst activity and selectivity of different oil and water-soluble precursors and dispersed catalysts. These factors include precursor solubility, rate of activation and degree of dispersion. The results demonstrated that the best performance was obtained with microcrystalline molybdenite generated in situ from oil-soluble precursors, and the nature of the organic ligand did not play a very significant role in influencing the hydrogenation activity, provided that the ligand ensured good oil-solubility as well as thermal lability of the precursor. In addition, it was also reported that powdered materials gave results below expectation and inferior to oil-soluble compounds. 24 Some studies have been performed on the effect of solvent on the catalyst activity during the hydrocracking reaction. Duangchan (1998) compared the effect of solvents including n-hexane, decalin and toluene for dispersed catalysts prepared in reversed micelles using Cold Lake heavy oil. It was found that coke yield increased with solvent type in the order of decalin < toluene < hexane. However, better product quality in terms of asphaltene conversion, sulfur and M C R content were obtained for the system with hexane compared to that of decalin. Yoneyama and Song (1999) reported that Mo sulfide catalysts generated from ammonium tetrathiomolybdate (ATTM) in the presence of n-tridecane solvent with added water under hydrogen pressure (6.9 MPa) at 300-400°C are much more active than catalyst prepared from A T T M alone, especially when these catalysts were tested for C-C bond cleavage and naphthalene ring hydrogenation using the model compound 4-(1 -naphthylmethyl)bibenzyl. 2.2.3.3 Active Species and Properties of Dispersed Catalyst Many studies have examined the catalyst recovered at the end of reaction to determine its composition and surface properties (Liu et al., 1994; Fixari et al., 1994; Buker et al., 1997; Panariti et al., 2000a; Panariti et al., 2000b; Watanabe et al., 2002). Generally, MoS 2 > whether the catalyst is prepared in situ or ex situ, is accepted to be the active ingredient during the hydrocracking reaction. Molybdenum naphthenate (MoNaph) is known to be a promising catalyst precursor in heavy oil hydrocracking (Bearden and Aldrich, 1981). Fixari et al. (1994) made a comparative study between phosphomolybdic acid and MoNaph, used as catalyst precursor in Safaniya vacuum residue hydrocracking. From physical characterization data, it was 25 concluded that the higher activity using MoNaph was related to a better sulfidation step to M0S2 compared to phosphomolybdic acid. Peureux et al. (1995) characterized the solid coke produced after the hydrocracking of Safaniya residue with MoNaph. From high-resolution T E M , the average crystallite length of M0S2 was reported to be 5 nm and mainly single layers of M0S2 were reported with the absence of the (002) peak in X R D diffractograms. Liu et al. (1994) reported on the hydrocracking of Gudao residue using water-and oil-soluble molybdenum based catalysts. T E M results showed that the possible catalytic active phase generated from the dispersed Mo catalysts is the microcrystalline MoS 2 . The microcrystalline M0S2 appeared to be spherical shaped particles of less than 2 um in diameter (Liu et al., 1994). Buker et al. (1997) tested a series of oil soluble naphthenates of N i , Co, W and Mo as catalyst precursor for upgrading vacuum residues. Analysis of the solid residue by X R D revealed that, under hydrocracking conditions, the metal naphthenate precursors were transformed into NiS, CoS, M0S2 and WS2 microcrystallites of diameter 4-5.5 um and length of 4-6 um. Similarly, Panariti et al. (2000a) reported that Mo-based oil-soluble precursors generated a micro-crystalline sulfide that dispersed in the feedstock as irregular clusters with a mean diameter of 0.5-2 um; and that the crystallites consisted of a single layer or a few layers and had a crystal size of 20-40 A. Watanabe et al. (2002) reported that different activities of molybdenum dithiocarbamate (Mo-DTC) and molybdenum dithiophosphate (Mo-DTP) could be ascribed to the different extent of M0S2 formation from the complexes during reaction. Stronger sulfiding of Mo-DTP was expected to increase activity by completing the formation of MoS 2 . 26 2.2.3.4 Multicomponent Catalyst More promising new catalysts could stem from research on promoted or multicomponent systems. For instance, phosphorus is recognized as a secondary promoter in Ni-Mo supported catalysts and is used in several commercial hydroprocessing catalysts (Del Bianco et al., 1995). In the case of Mo-based dispersed catalyst, the presence of phosphorus has significantly enhanced hydrodemetallation especially of vanadium (Panariti et al., 2000a). Most of the' synergistic effects between Co and Mo dispersed catalysts were reported for hydrodesulfurization. Lee et al. (1996) reported that Co-Mo catalysts showed increased activity toward hydroprocessing reactions such as hydrodesulfurization and hydrodemetallation and the maximum promotion of hydroprocessing activity appeared at a catalyst composition corresponding to a Co/(Co+Mo) atomic ratio of 0.3. However, Mo alone produced the highest hydrocracking activity and this decreased with increasing Co content. Similar conclusions were reported for the Co-Mo system for hydrodesulfurization (Fixari et al.^ 1994; Panariti et al., 2000a). Buker et al. (1997) reported that Ni-Mo sulfide catalyst generated in situ was effective in hydrodesulfurization, hydrogenation and gave low coke formation in a batch reactor and a continuous flow of vacuum residue in a cascade of two reactors. However, a discouraging result was reported by Sato et al. (1999) for the additive N i . N i did not enhance the hydrogenation activities of the unsupported M0S2 catalyst prepared ex situ. The enhancement of the hydrogenation activity of unsupported M0S2 catalysts by N i is presumably more difficult than the promotion of hydrodesulfurization activity by Co. They have also performed the reaction of Arabian Heavy atmospheric residue over zeolite-supported catalyst and poor activity for residue conversion was also reported. 27 It seems that M0S2 alone is a good catalyst for hydrocracking. Synergistic or promoter effect of N i and Co to Mo on supported catalyst for hydrocracking was not observed in the unsupported dispersed catalyst. 2.2.3.5 Effect of Reaction Variables The reactions occurring during hydrocracking of heavy oil are mainly thermally initiated, free radical reactions. The conversion of the feedstock to distillate is thermally controlled. It is practically independent of catalyst concentration and hydrogen pressure (Liu et al., 1994; Sanford, 1994; Buker et al., 1997; Panariti et al., 2000b). Catalyst favors hydrogen uptake and affects the product quality but does not significantly influence the rate of vacuum residuum conversion (Del Bianco et al., 1994). It was proposed that during hydrocracking, hydrocarbon radicals are initially formed and are hydrogenated in the subsequent steps. Therefore, catalyst concentration has little effect on the conversion of residue (B.P. > 525°C) but enhances the selectivity for the generation of hydrocarbons with a higher H/C ratio and suppresses the formation of coke (Buker et al., 1997). Coke formation is very well controlled even at low catalyst concentration (-200 ppm). The catalyst inhibits polymerization of high molecular weight molecules and consequently the formation of the aggregated phase of the polycondensed asphaltenes is reduced (Liu et al., 1994). However, coke formation increased at very high catalyst concentrations (5,000 ppm of Mo) (Panariti et al., 2000b). Since the hydrocracking reaction is thermally controlled, at high temperatures (> 410°C) and high H2 pressure, where radical chain reactions are significant, relatively high catalyst concentrations (0.05 - 0.3%) are needed in order to suppress coke formation. At 28 lower temperatures (< 410°C), where the coke suppression is easier, catalyst concentration mainly influences the quality of the product. The dispersed Mo catalysts promote hydrogenation, hydrodesulfurization (up to 80 %) and hydrodemetallation (up to 99 %) considerably. Nickel and vanadium organo-metallic compounds present in the residue show a high reactivity even in the presence of very low molybdenum concentration (Panariti et al., 2000b). The activity of Mo towards hydrodenitrogenation (HDN) is less satisfactory (20-30%) (Del Bianco et al., 1995). 2.2.3.6 Catalyst Recycle Although the slurry process has the advantage of stable catalyst activity relative to a classical fixed-bed reactor, the cost of catalyst has a major impact on the economics of the process. In this light, a 'once through' catalyst scheme can only be considered for processes using very low catalyst concentrations (typically less than 200 ppm). For higher concentration, recycle of the catalyst is mandatory, and special devices must be designed for this purpose (Del Bianco et al., 1995). The above observation is still valid today. This is due to the increasing cost of molybdenum (USD 3 per kg in 1995 to USD 45 per kg in 2006) (American Metal Market, 2006), despite the high oil price (approximately USD 60 per barrel) in 2006 (Annual Energy Outlook, 2006). Therefore, the catalyst concentration that can be considered once through is far lower than 200 ppm. Figure 2.3 shows a theoretical scheme for a slurry phase technology. Fresh and recycled catalysts are continuously added while aged catalyst is purged or regenerated. A variety of separation methods have been suggested to recover the dispersed catalyst such as 29 decantation, filtration, centrifugation or concentrated via distillation in the vacuum bottoms of the product stream. However, regeneration of the deactivated dispersed catalyst as a result of the accumulation of coke and metals (nickel and vanadium) had not been paid much attention. The information on regeneration chemistry and engineering of hydroprocessing are available only for supported catalysts (Furimsky and Massoth, 1993). Regeneration at the moment is very complicated and expensive so that simpler and cheaper solutions are desirable. Therefore, the overall economy of the slurry process depends on developing an efficient way of recycling most of the catalyst. Catalyst make-up Residue H 2 Preheater Recycled catalyst Regenerated catalyst Slurry _» Separation w reactors w units Regenerator 0 2 "• Product •> Purged catalyst " • N i , V "•Flue gas Figure 2.3 A simplified scheme for catalyst recycling/regenerations of the slurry process (Del Bianco et al., 1995). 2.2.4 Exfoliated Catalyst Joensen et al. (1986) introduced an exclusive chemical method that enables one to produce single layers or two-dimensional material from bulk clusters of M0S2, called 'exfoliation'. 30 To exfoliate M0S2 powder into monolayers, M0S2 powder is first immersed in hexane containing n-butyllithium in excess (i.e. Li :Mo >1). Microscopically, the L i compound intercalates between the MoS 2 layers. Then, the L i intercalated M0S2 compound is washed to remove excess n-butyllithium. Finally, the MoS 2 is dried and subsequently exposed to water. Rapid evolution of gas is accompanied by an exfoliation of the M0S2, yielding single layers that remain in suspension. It is assumed that the reaction between water and the intercalated lithium forms hydrogen gas between the layers, and the expansion of this gas separates the M0S2 layer. As the reaction proceeds more deeply into each crystallite the layers become further separated. Eventually the layers become completely separated and remain suspended in the aqueous solution. Minimal long-range order has been confirmed by X R D analysis (Joensen et al., 1986). The M0S2 layers can be restacked by re-dispersion in different solvents (Miremadi et al., 1991). During the re-stacking process, other metals such as A l , Co, and N i can be intercalated into the M0S2 layers (Miremadi and Morrison, 1988). Organic molecules such as 1-hexane, benzene and n-butyl alcohol can also be included between layers of M0S2 (Divigalpitiya et al., 1989). The exfoliation method has been employed in preparing M0S2 catalyst for methanation by supporting the exfoliated M0S2 on AI2O3 and high activity was reported (Miremadi and Morrison, 1987; Scholz and Morrison, 1989). For heavy oil primary upgrading, exfoliated M0S2 catalysts had only been tested in Hondo residue hydrocracking (Bockrath and Parfitt, 1995). It was a preliminary study done at only one temperature (425°C) and one catalyst concentration (10 wt% MoS 2 ) . There was no significant difference reported in hydrodesulfurization and hydrodenitrogenation activity when using commercially available MoS 2 and exfoliated, restacked MoS 2 . The results did not show any consistent correlation between catalyst properties, such as surface area and 31 M0S2 stack height, and the hydrodesulfurization activity. However, promoted cobalt or nickel exfoliated M0S2 catalysts were found to promote desulfurization significantly (by 50 %) (Bockrath and Parfitt, 1995). 2.2.5 Hydrocracking of Different Cut Points Petroleum can be fractionated into different cuts. Petroleum at different cut points exhibits different characteristics such as different asphaltene contents and different heteroatom contents. For example, the supercritical fluid extraction (SCFE) of heavy oils revealed that the end cut contains most of the problematic species in heavy oil such as asphaltene, metals and microcarbon residue, and these species cannot readily be upgraded into useful products. In the coking process, most of the end cuts convert into coke. In view of this, Zhao et al. (2000) suggested that, in order to reduce coking and avoid catalyst deactivation, the heaviest portion of heavy oil should be removed prior to hydrocracking. In addition, based on the analysis of different narrow cuts of Athabasca bitumen, hydrocracking product residues and coking residues from SCFE, Chung and Xu (2001) reported that sulfur species converted in hydrocracking were mainly from the front-end of heavy oil. In contrast, only a small amount of sulfur reduction was found from the end-cut. Lai and Smith (2001) reported a method to remove asphaltene from Cold Lake heavy oil by ultrafiltration using ceramic monolith membranes with pore size 0.1 - 1.4 um. In the aforementioned study, up to 80 % of asphaltene was rejected from Cold Lake heavy oil. However, hydrocracking of these oil fractions has not been carried out. 32 2.3 Secondary Upgrading/Hydroprocessing Hydroprocessing or hydrotreating is a secondary treatment that further improves the quality of distillates from primary cracking processes. It is intended for selective removal of the heteroatoms, especially sulfur and nitrogen, from the feed with little conversion of the hydrocarbons. Therefore, hydroprocessing is normally operated at milder temperature (350 - 430°C compared to 400 - 550°C in primary upgrading), H 2 pressure of 2 - 7 MPa, with the presence of catalyst. Under these conditions, thermal cracking and catalyst poisoning are minimized. The typical catalysts used in commercial hydroprocessing are MoS 2 catalysts supported on A12C>3 and promoted with Co or N i . The materials are similar to that for catalytic hydrocracking, except that the provision of macropores in the alumina matrix is of less concern. Hydroprocessing a cracked distillate involves a combination of reactions occuring simultaneously, including hydrodesulfurization, hydrodenitrogenation, hydrodeoxygenation and hydrogenation of aromatics. Hydrodesulfurization has been the main concern of hydroprocessing in industrial refineries and research studies. This is because of the increasing use of heavy crude oils with high sulfur content and because of global reductions in the allowable sulfur content of automotive fuels (Plantenga and Leliveld, 2003). Hydrodenitrogenation has received much attention as well, since the presence of nitrogen-containing compounds affects the deep hydrodesulfurization process. Hydrodenitrogenation is more difficult since nitrogen-containing aromatic rings require hydrogenation prior to nitrogen removal, and hydrogenation of the nitrogen-containing aromatic rings is difficult because of steric and thermodynamic limitations (Trytten et al., 1990). Relatively, hydrodeoxygenation receives less attention compared to hydrodesulfurization and hydrodenitrogenation because of the low O content of these oils. 33 2.3.1 Unsupported M0S2 Structure-Activity Relationships M0S2 based catalysts are extensively used in hydrocracking and hydroprocessing of feeds containing heteroatoms such as nitrogen, sulfur and oxygen. Due to their widespread use, there have been a large number of experimental and theoretical studies on M0S2 catalysts. However, the exact chemical nature and the structure of active phase are still debated. The study of MoS 2 based catalyst structure-activity has attracted much attention over the past two decades. Different structural models associated with catalyst activity of molybdenum sulfide based catalysts have been proposed. The catalysts studied include supported M0S2 promoted with Ni(Co), supported M0S2 and unsupported M0S2 catalysts. M0S2 has a layered structure with a weak Van der Waal's interlay er bonding as shown in Figure 2.4 (a). Each layer can be described by a S-Mo-S atomic layer stacking. Within the layer, Mo ions are six coordinated by sulfur ions in a trigonal prismatic "MoSe" unit (Kasztelan et al., 1984). It is believed that the hexagonal structure is the most probable one for M0S2 particles (Kasztelan, 1990). The active sites in MoS 2 are generally accepted to be metallic sites with sulfur vacancies and they are closely associated with the edges of the layers of the crystal. The basal plane, with completely coordinated sulfur atoms, is inert. Roxlo et al. (1986) showed a direct correlation between the hydrodesulfurization catalytic activity of unsupported M0S2 platelets and powders with their respective density of edge plane sites. The infrared optical absorption due to the edge planes in MoS 2 platelets was found to be proportional to the hydrodesulfurization activity of DBT. A turnover frequency for DBT conversion of approximately 5 x 10" molecules/(site.s) in both amorphous and crystalline M0S2 was reported. Eijsbouts et al. (1993) have also shown a clear correlation between 34 hydrodesulfurization activity per gram and the number of Mo atoms at the corners and edges of M0S2 crystallites, reported as the M0S2 dispersion. High-activity commercial catalysts are found to have very high M0S2 dispersions. (c) (100) plane (edge plane) (d) (110) plane (edge plane) Figure 2.4 Crystalline structure of M0S2 (a) overall structure; (b) (001) plane; (c) (100) plane and (d) (110) plane (Shimada, 2003). The relationship between M0S2 structure and active sites for hydrogenolysis and hydrogenation during hydrodesulfurization has been the subject of numerous studies. The M0S2 models proposed account for the effect of structure and active sites by distinguishing between edge, corner and rim sites of the M0S2 crystallite. 35 Zmierczak et al. (1982) analyzed oxygen chemisorption on several Mo and CoMo catalysts on different supports that were evaluated for thiophene hydrodesulfurization and hexene hydrogenation. It was found that oxygen chemisorption did not correlate with hydrodesulfurization activity but there was a reasonable correlation with the hydrogenation activity. Later, Massoth et. al. (1984) proposed that the edge sites of M0S2 supported on AI2O3 are active for hydrogenation, whereas only the corner sites were active for hydrodesulfurization. Kasztelan et al.(1984) developed a series of geometrical models for promoted and unpromoted M0S2. Using unsupported M0S2, it was reported that hydrodesulfurization of thiophene and hydrogenation of propene, cyclohexene or toluene, as well as O2 chemisorption, only occurred on the edge sites of M0S2. The "Co-Mo-S" model proposed by Topsoe et al. (1984; 1986) has been widely accepted as the structural model for promoted M0S2 based catalysts. In this model, Co(Ni)-Mo-S is the active phase on supported and unsupported catalysts promoted by Co or Ni . Two types of Co-Mo-S phase were proposed to exist (Topsoe and Clausen, 1984) on supported catalyst. Co-Mo-S type (I) is highly dispersed, monolayer M0S2 slabs, with Co at the edge, which has strong interaction with the support. Co-Mo-S type (II) has weaker interaction with the support and is a fully sulfided, multiple layered M0S2 cluster with Co. Co-Mo-S type (II) is more catalytically active than Co-Mo-S type (I). These studies assumed that the active catalytic sites of M0S2 are only at the edge planes, i.e. (100) and (110) planes (Figure 2.4) of the M0S2 crystal. Vrinat et al. (1994) investigated the effect of M0S2 morphology on thiophene hydrodesulfurization over supported M0S2 catalysts without any promoters. Intrinsic activities per Mo atom were not directly related to the total number of edge Mo sites, i.e. 36 the activity could not be explained solely by differences in the length of the slabs of the crystallites. Furthermore, all activities followed a continuous decrease with an increase in the number of layers in the M0S2 microcrystallites, regardless of the support or the sulfidation temperature. They concluded that for supported M0S2 catalysts, only the top edge sites are catalytic active. Daage and Chianelli (1994) proposed a "rim-edge model" for unsupported M0S2 based on hydrodesulfurization of dibenzothiophene (DBT) (Figure 2.5). According to the rim-edge model, hydrogenation reactions are catalyzed predominantly by the top and bottom layers of the M0S2 slab (the r im sites) for large molecules like D B T , while both hydrogenation and hydrogenolysis are catalyzed on all the edge planes. The basal plane of MoS2is considered to be inert catalytically. R i m sites and edge sites are shown in Figure 2.6. It was presumed that when single or bi-layers occur, selectivity is dominated by the degree of coordinative unsaturation present. Smaller sulfur-containing molecules may not exhibit the rim-edge effect due to the absence of steric interference. hydrogenolysis S Dibenzothiophene (DBT) hydrogenation hydrogenolysis S Figure 2.5 Reaction network of dibenzothiophene (Daage and Chianelli, 1994) 37 Lateral dimension < • ;ht) Figure 2.6 Rim-edge model of a M0S2 catalytic particle (Daage and Chianelli, 1994) Iwata et al. (1998) investigated the catalyst activities and structural properties of commercial M0S2, M0S2 prepared from sulfidation of ammonium heptamolybdate and M0S2 from thermal decomposition of ammonium tetrathiomolybdate. The hydrodesul-furization of DBT and hydrogenation of 1 -methylenaphthalene were used as model test reactions. The activity of the MoS 2 catalysts was found not to be directly related to the BET surface area but dependent on the morphology. Highly bent multi-layered M0S2 structures were believed more catalytically active, while a well-crystallized M0S2 structure was more favourable for direct desulfurization. It was proposed that the bent basal planes possess some catalyst activity, although less than that of the edge planes. Later, Iwata et al. (2001) demonstrated that the hydrogenation rate of 1-methylnaphthalene on dispersed M0S2 was proportional to the number of rim sites and proposed an aggregated particle model based on the rim-edge model, which is depicted in Figure 2.7. Based on X R D analysis, the hydrogenation active sites of the M0S2 particles were found to be associated with the 'inflection' of the basal plane that corresponded to the 38 rim sites of the crystalline M0S2 microdomains. This explains the reason why curved slabs observed under T E M were catalytically active. A n aggregated catalyst slab, with smaller microdomains has more 'inflection' points compared to an aggregated catalysts slab with longer microdomains. Therefore, the growth of the microdomains should be suppressed in order to increase the number of 'inflection' points. Hensen et al. (2001) tested the hydrogenation and hydrodesulfurization activities corrected for the M o available at the edge planes on small (thiophene and toluene) and large (DBT) molecules, using supported M0S2 catalyst. It was found that the hydrogenation rate for thiophene and toluene increased with an increase in stacking degree of supported M0S2. On the other hand, the hydrodesulfurization rate of thiophene was not a strong function of the M0S2 stacking degree. However, they found that higher stacking degree improved the intrinsic D B T hydrodesulfurization activity as well . T E M _ _ T E M Figure 2.7 Schematic drawing of the catalyst structure: (Left) catalyst decomposed at lower temperatures and (Right) catalyst decomposed at higher temperatures. Solid arrows: hydrogenation active sites; dotted arrows: edge sites except rim sites (Iwataetal., 2001). 39 Vradman and Landau (2001) investigated the effect of stacking degree on the performance of N i promoted WS2 catalysts supported on alumina and silica. It is believed W-based catalysts, that are more active in hydrogenation compared with Mo-based catalysts, should assist in emphasizing the effect of stacking degree on hydrogenation activity. Higher activity in hydrogenation of toluene was obtained with multilayered WS2 slabs, which agreed with Hensen et al.'s (2001) results. Different performance between thin and thick W(Mo)S2 slabs was suggested to be attributed to geometric considerations. It is unlikely that W(Mo)S2 multilayered slab contains active sites different from a single layer or'thin slab since the layers are not chemically bonded. It is believed that multi-layered WS2 slabs provide higher density of multi vacancies compared with single-layered or thin slabs, facilitating the 7r-complexation of the aromatic ring. Devers et al. (2002) compared the hydrodesulfurization and hydrogenation activities of unsupported M0S2 with different morphologies obtained from hydrothermal and thermal decomposition of thiosalts. Hydrothermal synthesis was carried out by adding ammonium thiomolybdate to water in the presence of sulfur in the temperature range 200 to 300°C. Thiophene was used as the model reactant for hydrodesulfurization and tetralin was used for hydrogenation. The same order of magnitude in intrinsic activity was obtained using both hydrothermal sulfides and thermally produced sulfides, in spite of higher surface area and longer M0S2 slabs in hydrothermal M0S2 compared to that of thermally produced sulfides. It was suggested that a non-zero activity of curved basal plane as proposed by Iwata et al. (1998) might be the reason. Farag et. al. (2003b) investigated the activity of a series of bulk M0S2 catalysts synthesized from various precursors, using hydrodesulfurization of DBT as a model 40 reaction. A n interesting relationship was found between the average number of M0S2 stacked layers and the hydrodesulfurization selectivity. The hydrodesulfurization selectivity (ratio of the rate constant for direct desulfurization to the rate constant for hydrogenation) passed through a minimum at crystallite size of 4 nm and then increased with crystallite larger than 5 nm, irrespective of the starting precursor. Recently, activity per gram Mo of alumina supported M0S2 catalysts prepared by exfoliation was compared to that of traditionally prepared M0S2 catalyst, using the hydrodesulfurization of thiophene. Although the number of Mo atoms in the edge plane per gram of M0S2 catalyst prepared by exfoliation is 10 times lower than that of the standard catalyst, it was reported that similar activity per gram Mo was obtained. It was suggested that the hydrodesulfurization of thiophene can occur on the basal plane of M0S2 (Kochubey and Babenko, 2002; Kochubei et al., 2003). The fact that the active sites in M0S2 lie mainly on the edge planes is well recognized. However, the active site oh M0S2 that promotes hydrogenation or that promotes hydrogenolysis of the carbon-heteroatom bond of heterocyclic organic compounds remains more controversial. For supported catalysts, other than the dispersion and the structure of the active M0S2, the orientation of the active phase and the support also affect the catalyst activity and selectivity. Breysse et al. (2003) reviewed the effect of support on catalyst properties and activities, and Shimada (2003) reviewed the catalytic performance of MoS 2 catalysts focusing on the relationships at the interface between M0S2 cluster and the support. Supported M0S2 catalyst activity performance was strongly dependent on the morphology and orientation of M0S2 clusters on the supports. 41 2.3.2 Use of Exfoliated Catalyst in Hydroprocessing Del Valle et al. (1994) compared the thiophene hydrodesulfurization activity of exfoliated M0S2 with unexfoliated M0S2 catalysts from different sources in a flow reactor. The thiophene/H2 was fed through the reactor at a flow rate of 100 ml/min, under 2 MPa and temperature 180 - 210°C with catalyst loading of 0.15 g. Exfoliated crystalline M0S2 was found to have hydrodesulfurization activity (2.7 x 10"7 mol/s.g) 3.5 times greater than crystalline MoS2(7.7 x 10"8 mol/s.g). The catalytic effect was obtained with no appreciable change in BET surface area. The increased catalyst activity of crystalline M0S2 obtained by exfoliation was explained by an increase of the edge sites that occurs due to the fragmentation from large to small particle. However, the exfoliated MoS 2 prepared by thermal decomposition of ammonium thiomolybdate had both activity and surface area reduced, suggesting that microporosity plays an important role in the MoS 2 catalyst. Curtis (1996) also reported a similar activity trend for thiophene hydro-desulfurization using exfoliated M0S2 and the original crystalline M0S2. Reactions were performed in a flow microreactor at 270°C, 101 kPa with 0.1-0.5 g of catalyst. The exfoliated M0S2 (1.19 x 10"9 mol/s.m2) was reported about twice as active as the original crystalline M0S2 (0.56 x 10" mol/s.m ). The difference in reactivity was suggested to be caused by a difference in the proportion of edge sites to rim sites as a function of crystallite size. In a later study by Del Valle et al. (1998), the catalyst activities of exfoliated M0S2 and unexfoliated MoS 2 catalysts from different sources for hydrodesulfurization of DBT in a batch reactor were compared. Reactions were carried out under 3.4 MPa of H 2 at 350°C with 8.8 g of DBT in 200 ml of decalin. However, the rate of DBT hydrodesulfurization 42 was reported to decrease for both the exfoliated M0S2 catalysts prepared from crystalline M0S2 and M0S2 prepared by thermal decomposition of ammonium thiomolybdate. The hydrogenation to hydrodesulfurization ratio also decreased when compared to MoS 2 catalyst without exfoliation. It was suggested that residual L i , introduced during the exfoliation, maybe responsible for the unexpectedly low hydrogenation activity of the exfoliated MoS 2 (Del Valle et al., 1998). A comparison has also been made between crystalline, exfoliated and electron irradiated WS2. The hydrodesulfurization of DBT (Galvan et al., 2000) was taken as a measure of catalyst activity. Reactions were carried out with 0.4 g of WS2 and 8.8 g of DBT in 200 ml decalin at 3.1 MPa H 2 pressure and 350°C. The catalyst activity of irradiated WS2 was found to be 2.4 times that of the exfoliated catalyst and 3.8 times that of crystalline WS2. The surface areas of the exfoliated and irradiated WS2 were nearly the same, but 30 % lower than that of crystalline WS2. The selectivity of exfoliated WS2 catalyst was similar to that of crystalline WS2. Electron irradiated WS2 was found to have better performance in hydrodesulfurization and this observation was attributed to the fracture of W S 2 crystallites, which creates new active surfaces by enhancing edge site concentration. The exfoliation method has also been used to prepare single layered M0S2 supported on alumina. The activity of supported exfoliated M0S2 was compared to that of supported multilayered (4 layers) M0S2 catalyst prepared from (NH4)2MoS4 in hydro-desulfurization reactions using thiophene and tetrahydrothiophene (Boone and Ekerdt, 2000). Reactant was fed to a reactor system by H2 being passed at a flow of 18 ml/min through a saturator containing liquid reactant (thiophene or tetrahydrothiophene) at 295 K and 1 atm. The reactor temperature varied from 250-400°C. The single layered exfoliated 43 M0S2, at the same Mo loading, was found to have similar activity and selectivity in the hydrodesulfurization reactions compared to multilayered M0S2 catalyst. Recently, a similar conclusion was also reported by Kochubey et al. (2003). Exfoliated catalyst has only been tested for its hydrodesulfurization activity using thiophene and DBT, but has not been used for hydrodenitrogenation and hydrode-oxygenation reactions. 2.4 Summary Development of heavy oil slurry hydrocracking processes is on-going. Dispersed catalysts are reported to provide promising capability for coke suppression in hydrocracking, and are known to play an important role in hydrogenation that leads to improved product quality. Although M0S2 has generally been accepted as an active catalyst in hydrocracking, the high M0S2 cost despite the use of low catalyst concentration makes a commercial slurry hydrocracking process uneconomical. Many studies have focused on finding a viable dispersed catalyst. Catalyst structure was reported to be related to catalyst activity in hydroprocessing reactions such as hydrodesulfurization (Daage and Chianelli, 1994). Therefore, catalyst structure could be a parameter that affects the catalyst activity during hydrocracking, but this has not been widely studied. There have been studies on the relationship between morphology and the activity of M0S2 catalyst using model reactants under hydroprocessing conditions. In particular, strong interest has been placed on the different genesis of hydrogenation and hydrodesulfurization active sites (Schweiger et a l , 2002; Chianelli et al., 2002). However, there is limited data on structure activity available for hydrodenitrogenation or hydrodeoxygenation reactions. 44 Moreover, most of the investigations on the structure and activities were on supported CoMoS and/or NiMoS catalyst or supported M0S2 (Eijsbouts, 1997; Sakashita and Yoneda, 1999; Sakashita et a l , 2000; Sakashita et al., 2001; Shimada, 2003). Only a limited number of studies have been carried out on the properties of bulk M0S2 and their effects on activity (Iwata et al., 1998; Calais et al., 1998; Del Valle et a l , 1998; Devers et al., 2002; Farag et al., 2003b). These studies have indicated that the catalyst activity is closely related to the edge sites and number of layers, in particular with sulfur vacancies formed on the edge sites, while the basal plane with completely coordinated sulfur atoms are thought to be inert. Unsupported exfoliated MoS 2 has potential as a catalyst for hydrocracking and hydroprocessing. As a model catalyst, it is well suited for the study of the dependence of activity on structural properties, independent of support and promoter effects. Exfoliated M0S2 catalysts, with their unique struture, have only been investigated for their hydrogenation activity in Hondo residue hydrocracking (Bockrath and Parfitt, 1995a), and model compound hydrodesulfurization reactions (Del Valle et al., 1994; Del Valle et al., 1998). Furthermore, there is no consistent correlation drawn between catalyst structure and activity from these results, and some reports indicate conflicting results. Further study of exfoliated MoS 2 in both hydroprocessing and hydrocracking reactions is therefore warranted. 45 Experimental Experiments were carried out to assess a series of M0S2 catalysts, prepared in different ways such that each catalyst had a different morphology. Hydrocracking of Cold Lake heavy oil and hydroprocessing of model compounds were examined using the catalysts dispersed in the oil. A slurry phase, stirred, batch reactor was used for all activity measurements. The present chapter describes the experimental procedures and methods used for preparation and characterization of the catalysts, and for determining their catalytic activities. The experimental section is conveniently divided into the preparation of M0S2 catalyst, the catalyst activity tests, product analysis and characterization of catalyst and recovered solid. 3.1 Catalysts Preparation Four different M0S2 catalysts were studied in the present work. Two catalysts were prepared by in situ decomposition of soluble Mo precursors: molybdenum naphthenate (MoNaph) (6.0 wt% Mo) from ICN Biomedicals Inc. (or from K & K Laboratories Inc. when specifically mentioned), and ammonium heptamolybdate tetrahydrate, (NH^eMoyC^ •4H2O (AHM) from Sigma-Aldrich. MoNaph was added directly to the reactor while A H M was first dissolved in 1 ml of distilled water before being added to the reactor. The Mo precursor dispersed in the reactor liquid and decomposed during the heat up period prior to 46 the start of the reaction period. Typically the reactor was heated at 5 or 10°C/min to the desired temperature in the presence of 5% H2S/ 95% H2 (see Section 3.2.1.3 and Section 3.2.2.2). The third catalyst was commercially available crystalline M0S2 powder (99%>) from Sigma-Aldrich, used without further treatment and the fourth catalyst was the exfoliated M0S2 prepared from the crystalline M0S2 powder. 3.1.1 Preparation of Exfoliated M0S2 M0S2 was exfoliated using a slightly modified method of Joensen et al. (1986). First, M0S2 powder (< 2 um, 99%, Sigma-Aldrich) was soaked in n-butyllithium (2.5 M in hexane, Sigma-Aldrich) such that the Li :Mo ratio was 1:1, to affect an intercalation of L i into the M0S2. The intercalation was done under Ar atmosphere in a glove box and the sample bottle containing M0S2 powder was shaken from time to time to ensure good mixing. The sample bottle was then sealed and left for at least 72 hours so that all the L i had intercalated into the M0S2 layers, and the intercalated M0S2 settled to the bottom of the sample bottle. The top layer solution was decanted and the Li-intercalated M0S2 material was subsequently exposed to water. Upon contact with water, rapid gas evolution occurred and the M0S2 formed a highly opaque suspension in water. The solution was sonicated for 30 min followed by 30 min of stirring. Subsequently, L i was removed from the solution by the following series of washing steps. The exfoliated M0S2 was separated from suspension using a centrifuge spinning at 3000 r.p.m. for 30 min. The supernatant was subsequently decanted and the collected MoS 2 material was redispersed by adding distilled water and stirred for another 47 30 min. The washing process was repeated until the solution pH was neutral confirming that all the L i had been removed. Normally, the washing step had to be repeated three times. The concentration of the exfoliated M0S2 solution was then adjusted to that required by adding an appropriate amount of distilled water. Typically, the exfoliated M0S2 aqueous solution was adjusted to a concentration of 2 wt% of M0S2 dispersed in water, unless otherwise stated. The exfoliated M0S2 would remain suspended in water for about two weeks. Once the age of the catalyst was more than 2 weeks, a new batch of catalyst was prepared. For the case when exfoliated M0S2 was dispersed in decalin, the same procedure was followed as for the exfoliated M0S2 dispersed in water. However, after the exfoliated M0S2 was washed to remove L i , the MoS 2 collected after centrifuging was washed again with isopropanol (99 %, Acros Organics) and stirred for 30 min. The washing process was repeated three times in isopropanol. The washing procedure was then repeated another three times, using decalin (98 %, Acros Organics) as the washing medium. Finally, the exfoliated M0S2 was adjusted to 2 wt% of M0S2 dispersed in decalin. Exfoliated M0S2 dispersed in decalin remained in suspension for two to three days. A new batch of catalyst was prepared for testing once the catalyst was more than 3 days old. 3.2 Catalyst Activity Measurement 3.2.1 Heavy Oil Hydrocracking Reaction 3.2.1.1 Feedstock The heavy oil used in the present study was Cold Lake Vacuum Bottoms provided by Esso Resources Canada. In addition, filtered oil and retentate derived from the ultra-filtration of 48 Cold Lake heavy oil using ceramic monolith membranes (Lai and Smith, 2001) with different properties were also used as heavy oil feed. The properties of these feedstocks are presented in Table 3.1. Table 3.1 Properties of different feedstock Particular Cold Lake Vacuum Bottoms Filtrate* Retentate* Carbon, wt% 80.5 84.6 83.7 Hydrogen, wt% 10.6 10.5 10.3 Sulfur, wt% 4.1 4.4 5.3 Nitrogen, wt% 0.34 0.30 0.47 Nickel, ppm 67.3 41.0 115.1 Vanadium, ppm 179.2 111.4 290.3 H/C atomic ratio 1.58 1.48 1.47 Asphaltene, wt% 17.9 7.0 . 27.0 Microcarbon residue, wt% 13.6 10.3 21.2 Boiling point fraction, wt% <177°C 0.0 0.9 0.0 177- 350 °C 17.1 16.7 13.3 350- 525 °C 27.3 33.6 27.9 > 525°C 55.7 48.8 58.8 *Cold Lake Vacuum Bottoms was filtered through a porous ceramic membrane (< 0.1 um pore size). The oil that passed through the filter is termed filtrate and that which did not pass through is termed retentate. 49 3.2.1.2 Reactor Setup An Autoclave Engineers high-pressure batch reactor (model Eze-Seal) was used for the catalyst evaluation in hydrocracking reactions. The reactor included a 300 ml stainless steel vessel, a removable electric furnace, motor driven magnetic stirrer, a safety valve, an internal water-cooling coil and a pressure transducer. A temperature controller controlled the electric furnace and a speed controller controlled the stirrer. A gas bomb was connected to the gas outlet line of the reactor in order to collect gas samples after reaction. The reactor setup is shown in Figure 3.1. 3.2.1.3 Hydrocracking Activity Test Procedure Prior to the hydrocracking reactions, the reactor was loaded with 80 g of Cold Lake heavy oil and the desired amount of catalyst or catalyst precursor. The Mo loading in the heavy oil ranged from 360 to 900 ppm. The reactor vessel was secured via 6 hexagonal bolts and enfolded with a removable electric furnace. The reactor was flushed with N 2 gas followed by 5% H2S / 95% H2 or pure H2 gas for at least 15 min, before being pressurized with the same gas to 3.5 MPa. The reactor was then heated to the desired reaction temperature at a rate of 5 °C/min. The stirrer speed was set at 350 rpm throughout the reaction, and tap water was used to cool the bearing of the magnetic stirrer. Once the reaction temperature (400 - 430°C) was achieved, it was maintained for 1 hour. The temperature of the furnace, the temperature inside the reactor vessel and the reactor pressure were recorded at 5 min intervals starting from the reactor heat up until the end of the 1 hour reaction period. After the reaction, the heating jacket was removed and 50 the reactor vessel was quenched to room temperature by passing cold water through the cooling coil in the reactor. The sealed autoclave was then left overnight. Water out Nj Gas X X - — X — > • < — HJ/HJS Speed Controller Temperature Controller Gas line Electric line Water line Rupture disc To fume hood 1& Thermocouple Gas vent Electronic Box Pressure Transducer Gas bomb Burst vent To vacuum pump Figure 3.1 Schematic diagram of the slurry reactor setup 51 The pressure after reaction of the reactor at room temperature was recorded in order to calculate the amount of gas in the reactor after the hydrocracking reaction. The gas in the vessel was collected in a gas bomb for analysis by gas chromatography before the reactor was depressurized and disassembled. Solid carbonaceous deposits on the impeller and reactor wall were removed by scraping with a spatula followed by washing with methylene chloride (CH2CI2). The coke produced during the hydrocracking reaction was defined as methylene chlorine insoluble material. The coke was separated from the liquid product or the washings liquid via vacuum filtration using a 3.0 urn pore size filter. The collected solid was washed with methylene chloride. The filtrate was refiltered through a 0.22 um pore size filter membrane. The filter membranes with coke were dried in an oven at 110°C overnight. Therefore, coke formed in a reaction corresponded to methylene chlorine insoluble solid recovered from both the reactor and the reactor washings. The reaction liquid product was recovered via extraction from methylene chloride (B.P. 40.7°C), by evaporating the solvent in a rotary evaporator set at 70°C. The amount of coke and liquid product recovered were weighed. With the amount of gas, liquid and coke recovered, an overall mass balance was calculated. The conditions of the catalyst activity test were based on the slurry phase hydrocracking process developed by the Alberta Department of Energy (Lott et al. 1993). In this process, the Liquid Hourly Space Velocity (LHSV) is 1 hr"1 and the operating temperature range is 430 - 460°C. The range of the temperature investigated in the present study was 400 - 430°C. This was due to the maximum operating temperature of the reactor used, which is 450°C, and to minimize the effect of thermal reactions in assessing catalyst hydrocracking activity. 52 3.2.1.4 Product Gas Analysis The main purpose of the product gas analysis was to quantify the amount of H2 gas consumed during reaction as well as to complete a mass balance before and after reaction. The hydrocarbon gas collected after reaction was analyzed by a Shimadzu gas chromatograph (model GC-14A), equipped with an integrator (Shimadzu CR501 Chromatopac) and a thermal conductivity detector (TCD). Columns suitable for separation of hydrocarbons and H2S (Poropak Q and molecular sieves) were used for the analysis. The detector current was set at 150 mA and the temperature was set at 180°C. A temperature program was used for the analysis. The column temperature was set at 30 °C initially for 10 min and then increased to 120°C at the rate of 20°C/min and maintained at this temperature for another 10 min. Helium gas was used as the carrier gas with a flow rate of 25 ml/min passing through the column. Gas was injected into the gas chromatography column using an auto injector. A pneumatic VICI Valco valve was used for the injection. The valve was set to have a switching period of 0.2 min. The gas sample was introduced into the inlet gas line using a 60 ml syringe and an injection pump with a flow rate set at 25 ml/min, the same as the carrier gas flow. Details of the product analysis and calculations can be found in Appendix B-3.1. 3.2.1.5 Liquid Product Analysis Liquid products recovered from reaction were analysed to determine the oil quality after reaction. The liquid products were analyzed for carbon, hydrogen, nitrogen and sulfur content, microcarbon residue (MCR), the boiling point distribution of products and the 53 metal content. These analyses were done at the National Centre for Upgrading Technology (NCUT), Alberta. Asphaltene content was determined according to the Syncrude asphaltene analytical method (Liu and Gunning, 1991). Appendix A provides a summary of the method used for each analysis. 3.2.2 Model Compound Reactions 3.2.2.1 Reactor Setup The reactor setup described in Section 3.2.1.2 was used for the model compound reactions. The reactions were performed with the catalyst particles freely suspended in the liquid phase and a stainless sampling tube permitted liquid samples to be withdrawn from the reactor during the course of the reaction. The estimated volume of this line was 1 ml. The reaction furnace and reactor temperature, pressure and stirrer speed were logged onto a computer in real-time. 3.2.2.2 Hydroprocessing Activity Test Procedure First, the reactor was loaded with 100 ml of the n-hexadecane solvent (99 %, Acros Organics), reactant (Table 3.2) and a suitable amount of catalyst or catalyst precursor to provide 600 ppm of Mo in the solvent during reaction. The amount of reactant added was chosen such that the heteroatom (S, N or O) concentration was in the range expected for a typical gas oil feedstock and is within the range normally used for hydroprocessing studies. The reactor was then sealed and purged with N 2 , followed by 5% H2S / 95% H2 gas for 15 min. 54 Table 3.2 Test conditions Reaction Reactant Initial concentration, wt% Hydrogenation Naphthalene 10.0 Hydrodesulfurization Dibenzothiophene 0.5 Hydrodenitrogenation Quinoline 0,2 Hydrodenitrogenation Carbazole 0.3 Hydrodeoxygenation Phenol 5.0 The stirrer speed was set at 1200 rpm in order to have full turbulent mixing in the reactor vessel. The reactor was pressurized in 5% H2S / 95% H2 to a pressure of 2.8 MPa initially and subsequently heated at a rate of 10°C/min to the reaction temperature (325 -375°C) and maintained at this temperature for 5 hours. Liquid samples were collected from time to time during the reaction. The liquid sampling line was purged by withdrawing 4 ml of liquid before collecting the zero time sample (0.5 ml). Another 7 samples were withdrawn during the course of the reaction over a 5-hour period. Samples were collected at 30 min intervals for the first 2 hours of reaction and at 1 -hour intervals for the next 3 hours of reaction. The liquid sampling line was always flushed with 2 ml of reacting mixture before sampling. After 5-hour reaction time, the reactor was quenched with cooling water. The reactor was depressurized and the liquid recovered after reaction. The catalyst activity measurements were based on a typical hydroprocessing reaction temperature range (300 - 400°C). The reactor pressure was in the range of 4 - 5 MPa at the reaction temperature. The reaction system was in the liquid phase at the reaction 55 temperature and pressure used. The solubility of H2 gas in the solvent was 0.10 mol/mol solvent, estimated at the reaction condition, and this was not affected significantly by the small pressure drop during reaction (Park et al., 1995; Ghosh et al., 2003). The solids present after reaction were collected by filtration, washed with pentane, vacuum dried for 3 hours at 100°C and then at 160°C for 2 hours before subsequent characterization. 3.2.2.3 Liquid Product Analysis Liquid samples were analyzed using a Shimadzu gas chromatograph (model GC-14A) with a AT-5 (30 m x 0.32 mm x Q.25 um) capillary column using a flame ionization detector (FID). A temperature program was used for the analysis. The column temperature was set at 60°C initially, and was ramped to 170°C at the rate of 5°C/min, followed by 10°C/min to 280°C. The identities of the products were determined by comparison with pure reference samples and by coupling gas chromatography and mass spectrometry analysis. Molar gas chromatography response factors were determined using pure samples. 3.3 Catalyst Characterization Different methods were used to characterize the fresh exfoliated catalyst, the coke recovered from hydrocracking reactions and the M0S2 catalyst recovered from model compound reactions. The main purpose of characterizing the fresh and recovered catalyst from model compound reactions was to establish the structure of the M0S2 crystallites. The fresh M0S2 and recovered catalysts were characterized by x-ray diffraction, BET surface 56 area, SEM-EDX, and transmission electron microscopy. The coke produced during hydrocracking reactions was also characterized through various methods to determine the physical and chemical properties of the coke. The characterization aimed to trace the fate of the dispersed catalyst used during the hydrocracking reaction. Standard sample preparation methods were used for each characterization method and are discussed in the following sections. 3.3.1 X-ray Diffraction Analysis X-ray diffraction (XRD) can be used to identify crystalline components present in a sample. In the present study, X R D was used to check for minimum long-range order in M0S2. The samples were dispersed in ethanol and air dried on a glass slide. The x-ray diffraction patterns were recorded with a Siemens D5000 powder diffractometer with power settings of 40 kV and 30 mA using Cu K a l radiation (k = 1.5406 A). The step-scan was taken over the range 29 from 5 to 70° in increments of 0.01°. Diffraction patterns derived from the diffactograms were compared with standard data files. The stack height of M0S2 was determined using the diffraction peak of the (002) M0S2 plane and the Scherrer equation: Dstack = 0.89Mj3cos.9) where Dstack is the average crystallite size of the M0S2 for the plane perpendicular to the basal plane; A, is the wavelength of Cu K a l (1.5406 A) , p is the measured full-width at half-maximum of the diffraction peak in radian. Acceptable estimates by this method are limited to stack heights from 3 - 200 nm. 57 3.3.2 Brunauer-Emmett-Teller (BET) Surface Area BET surface areas of the unused M0S2 catalysts, the coke recovered after hydrocracking and the recovered M0S2 catalysts from model compound reactions were determined. Samples for BET analysis must be in the solid phase. In order to analyze the BET surface area of exfoliated M0S2, the exfoliated M0S2 solution was dried at 90°C for 1 hour in a vacuum oven. After drying, the exfoliated M0S2 appears as silver, shiny flakes. The BET surface area of these flakes was determined. BET surface areas were measured by N2 adsorption at -196°C using a FlowSorb II 2300 Micromeritics analyzer. A 30% IS^/He mixture, fed at 15 ml/min was used for surface area measurement and a 95% N2/He mixture fed at 20 ml/min was used for total pore volume measurements. Samples were degassed at 150°C for 2 hours prior to measurement. The precision of the measurements were within ±10 %. 3.3.3 Scanning Electron Microscopy-Energy Dispersive X-Ray Emission (SEM-EDX) SEM-EDX was used to visualize the morphology of the coke, and the M0S2 dispersion and homogeneity in the coke. Energy dispersive x-ray emission was used to assist in composition analysis of the samples. The energy dispersive x-ray emission was done on a Hitachi S300N scanning electron microscope at 20 kV. The solid sample was used directly as specimen for the analysis. 58 3.3.4 Transmission Electron Microscopy (TEM) T E M was performed under high-resolution mode on a Hitachi H7600 electron microscope operated at 200 kV. The solid samples such as coke and recovered M0S2 catalyst, were ground to a fine powder with a pestle and mortar. The powder was dispersed in ethanol and deposited on a carbon-coated copper grid. The specimen was dried in air prior to analysis. For samples of exfoliated M0S2, the T E M specimen was prepared by dipping a carbon-coated grid into water diluted exfoliated M0S2 solution. The specimen was also dried in air prior to analysis. 3.3.5 Diffuse Reflectance Infrared Fourier Transform Spectroscopy (DRIFTS) DRIFTS was performed on a Bio Rad FTS 175 analyzer equipped with a mercury cadmium telluride detector (MCT). DRIFTS was used to analyze the coke recovered from reaction. DRIFTS is based on the diffusely reflected radiation from a solid sample. To relate band intensity linearly to concentration in DRIFTS, the relative reflectance spectrum was converted using the Kubelka-Munk (K-Munk) equation. The DRIFTS spectra were measured in the wave number range 700 - 4000 cm"1 with a resolution of 4 cm"1, 256 scan per spectrum. Finely ground coke powder samples were pressed into discs and loaded onto the sample holder with KBr as reference. The spectra were collected at room temperature. Different chemical functional groups present in the coke could be partially quantified by DRIFTS. This helped in clarifying the chemical properties of different cokes. 59 Exfoliated M o S 2 Catalyst in Cold Lake Heavy O i l Hydrocracking In this chapter, exfoliated M0S2 was studied for its catalytic hydrocracking activity of Cold Lake heavy oil. The evaluation was conducted to include the effect of different operating conditions, exfoliated M0S2 properties and different petroleum fractions as feed. In addition, the feasibility of recycling spent catalyst, which resides in coke after reaction, was investigated. In view of this, the results of the above studies are organized into four sections. Firstly, the presentation begins with a discussion on the properties of exfoliated M0S2, followed by the studies of catalyst activity. In the third section, the results of recovered coke characterization are presented. Finally, results of an investigation that examined the possibility of catalyst recycle are discussed. A version of this chapter has been published. Tye, CT. and Smith, KJ. (2004) Cold Lake Bitumen Upgrading Using Exfoliated MoS2. Catal. Lett. 95 (3-4): 203-209. 60 4.1 Properties of Exfoliated MoS2 M0S2 has a lamellar structure. The inter-lamellar van der Waals bonding between two adjacent layers of sulfur atoms is relatively weak and the lamellar structure can be separated into a number of layers through exfoliation. The method to exfoliate M0S2 crystallite was described in Section 3.1 of Chapter 3. In the present section, the morphology of the exfoliated M0S2 was examined. The M0S2 particle size, crystallite stack height and the number of layers in each crystallite were determined. The fresh exfoliated M0S2 was analyzed using X R D . Figure 4.1 shows the X R D diffractograms of: (a) untreated crystalline M0S2 before exfoliation, (b) exfoliated MoS 2 dispersed in water and (c) exfoliated M0S2 redispersed in decalin. A l l of the peaks for the untreated M0S2 are sharp and have high intensity especially the peak corresponding to the (002) plane. The average M0S2 crystallite particle size, estimated using Scherrer's equation applied to the (002) peak, was 35.8 nm for the untreated crystalline MoS 2 . After exfoliation in water (sample (b)), the (002) peak broadened and the estimated average particle size decreased to 4 nm. Correspondingly, the number of M0S2 layers per particle (Ns) decreased from about 57 layers to 6 layers as shown' in Table 4.1 (calculation in Appendix F-l) . The X R D data also indicate that the washed and dried catalyst was free of intercalated lithium hydrate following exfoliation since the inter-planar spacing (6.3 A) of the exfoliated M0S2 was much less than that expected (11.8 A) if lithium hydrate were intercalated (Divigalpitiya et al., 1989). The small increase in the average interplanar spacing from 6.2A (crystalline M0S2) to 6.3 A (exfoliated M0S2) is due to the presence of water between the M0S2 layers (Joensen et al., 1987). 61 40000 _ 30000 c u "I 20000 H ' CO £= Q) - 10000 o 4 o (a) (002) 10 (100) (101) (103) (004) (102) (006) (105) _A « 20 30 40 2 0 50 1600 i _ 1200 c 2 •e CO >> ID 800 - 400 10 20 30 40 2 9 50 Figure 4.1 X-ray diffraction pattern of (a) untreated crystalline M0S2; (b) M0S2 after exfoliation dispersed in water; (c) MoS 2 after exfoliation and redispersed in decalin. 6 2 Exfoliated M 0 S 2 remains in suspension for several days due to the formation of a hydroxylated M 0 S 2 surface but the M 0 S 2 will eventually restack to form multilayer particles (Joensen et al., 1987). When the exfoliated MoS 2 was re-dispersed in the non-polar organic solvent decalin, the period of suspension for MoS 2 was reduced. The X R D diffractogram of exfoliated M 0 S 2 dispersed in decalin is presented in Figure 4.1 (c). The estimated average particle size using Scherrer's equation and the (002) peak broadening was 5.5 nm, corresponding to 9 layers of M 0 S 2 in the average crystallite. Compared to the exfoliated M 0 S 2 dispersed in water, exfoliated M 0 S 2 dispersed in decalin has higher average stacking. The BET surface area of the original crystalline M 0 S 2 was 4.2 m2/g. The MoS 2 subjected to exfoliation and analyzed after drying showed a small increase in surface area to 7.8 m 2/g. The low surface area was consistent with the literature (Valle et al., 1994). Table 4.1 provides a summary of the properties of the exfoliated M 0 S 2 catalysts. Table 4.1 Summary of M 0 S 2 catalyst properties. Crystalline Exfoliated MoS 2 MoS 2 Water3 Decalin a Literatureb BET area, m lg 4.2 7.8 ±0.5 - 6.2 Slab length**, nm 560 ± 740 400 ± 460 400 ± 460 -Stack height, Dstack*, nm 35.8 4.0 5.5 14 Number of layers, Ns* 57 6 9 23 a Solvent in which exfoliated MoS2 was dispersed. b Data from Del Valle etal. (1998). * Dstack estimated using XRD line broadening of (002) peak; Ns estimated as the slab length divided by the inter-planar spacing. An example of calculation is given in Appendix F-l ** Estimated from TEM. 63 The morphology of the M0S2 particles was observed using high resolution T E M . Comparing exfoliated M0S2 to crystalline M0S2 in Figure 4.2, exfoliated M0S2 had slightly smaller average slab length and much lower stack height (Table 4.1). After exfoliation, the average slab length of M0S2 did not change appreciably. Exfoliation M0S2 remained a wide slab length distribution as the original crystalline M0S2 (Figure 4.3). From the slab length distribution shown in Figure 4.3, crystalline M0S2 was estimated to have an average slab length of 560 nm while the exfoliated M0S2 had a sheet morphology with an average slab length of 400 nm. A n E D X analysis showed that the S/Mo atomic ratio for exfoliated M0S2 was 2.6. . " , The characterization data for the exfoliated MoS 2 does not provide conclusive evidence for the presence of single layered M0S2 in the oil during reaction. The catalyst characterization is necessarily done ex situ, following a drying step that may result in partial restacking of the M0S2 (Joensen et al., 1987). A similar conclusion was drawn by Del Valle et al. (1994) in their attempts to identify single layered M o S 2 through T E M . However, the X R D analysis clearly shows that the M0S2 layers were dismantled during the exfoliation treatment. The data shows that exfoliated M0S2, dispersed in water or decalin consisted of particles with 6 - 9 M o S 2 layers in a large sheet (0.40 jam) morphology, compared to crystalline M0S2 that had 57 MoS 2 layers and an average slab length of 0.56 um. The morphology of these M0S2 particles is also very different from M0S2 derived from MoNaph, that had an average 10 nm slab length with 1 to 2 M0S2 layers, as presented in detail in Section 5.1. 64 Figure 4.2 T E M photo of (a) crystalline MoS 2 (b) exfoliated MoS 2 . 65 (a) 0 1 2 5 2 5 0 4 0 0 5 2 5 6 2 5 7 5 0 8 7 5 1 0 0 0 1 1 2 5 1 2 5 0 Slab length, nm (b) 10 0 1 2 5 2 5 0 3 7 5 5 0 0 6 2 5 7 5 0 8 7 5 1 0 0 0 1 1 2 5 1 2 5 0 Slab length, nm gure 4.3 Slab length distributions for (a) crystalline M0S2 (b) exfoliated M0S2 66 4.2 Catalyst Activity Studies The catalyst hydrocracking activity was studied using Cold Lake heavy oil. The catalytic hydrocracking reaction using exfoliated M0S2 was compared to a coking reaction (reaction under N 2 without catalyst), a thermal hydrocracking reaction (reaction under H 2 without catalyst) and catalytic hydrocracking with molybdenum naphthenate (MoNaph) or crystalline MoS 2 as catalyst. MoNaph is a precursor that generates MoS 2 in situ during the reaction in the presence of sulfur. It has been used as a catalyst precursor in heavy oil hydrocracking using slurry reactors (Bearden and Aldrich, 1981). A range of process conditions and exfoliated MoS 2 catalyst properties were also varied to determine their influence on the catalyst performance. The effect of feed oils with different asphaltene content on the hydrocracking reactions was also examined in the presence of exfoliated MoS 2 . Note that the exfoliated MoS 2 discussed in the rest of this chapter refers to exfoliated MoS 2 dispersed in water, except when otherwise noted. The heavy oil hydrocracking reactions were monitored by measuring coke yield, liquid yield, gas yield and the H 2 consumed in each batch reaction. The liquid product quality was monitored by measuring the product liquid H/C atomic ratio, residue conversion, asphaltene conversion, microcarbon residue (MCR) conversion and heteroatom removal. The heteroatom removal in the product liquid oil included nitrogen, sulfur and heavy metal (i.e. nickel and vanadium) removal. 4.2.1 Activity Comparison The catalytic hydrocracking of Cold Lake heavy oil using exfoliated MoS 2 was first compared with a coking reaction, a thermal hydrocracking reaction and catalytic 67 hydrocracking using MoNaph as precursor with 600 ppm Mo at 430°C. The coke yield and liquid yield of these reactions is compared in Figure 4.4. The coke yield decreased in the order: coking (14 ± 0.4 %) > thermal hydrocracking (9 ± 2 %) > hydrocracking using exfoliated M0S2 (6 %) > hydrocracking using MoNaph (4 ± 0.05 %). The liquid yield increased in the same order. As expected, both catalytic hydrocracking reactions had lower coke yield and higher liquid yield compared to the coking and thermal hydrocracking reactions. The repeatability analysis of the measurement of hydrocracking reaction is discussed in Appendix B-2. 100 • coke yield • liquid yield Coking Thermal Exfoliated MoNaph M o S 2 Reaction Figure 4.4 Comparisons between coking, thermal cracking, catalytic hydrocracking using MoNaph and exfoliated MoS 2 (Reaction conditions: 430°C, initial H2 pressure of 3.5 MPa, reaction time of 1 hour, 600 ppm Mo if MoS 2 catalyst was used). 68 Above 430°C, thermal cracking reactions dominate in heavy oil hydrocracking. In order to have a better comparison between catalytic reactions, the hydrocracking of Cold Lake heavy oil using exfoliated M0S2 was again compared to hydrocracking reactions using MoNaph and crystalline M0S2 at 415°C. The results are presented in Table 4.2. Table 4.2 Cold Lake heavy oil hydrocracking using different M0S2 catalysts (600 ppm Mo, 415°C, initial H2 pressure of 3.5 MPa, reaction time of 1 hour). Catalyst Crystalline MoNaph* Exfoliated MoS 2 MoS 2 Input gas H 2 H 2 S / H 2 H 2 Coke yield, wt% 2.87 0.98 1.05 ±0.06 Liquid yield, wt% .93.7 96.8 96.5 ±0.5 Gas yield, wt% 3.4 2.3 2.5 ±0.6 H2 consumptiori+++, wt% 34.7 50.0 53.1 ±5 .7 Residue conversion"1", wt% - 26.7 25.3 Liquid H/C atomic ratio+ + - 1.44 1.55 Liquid boiling point"1", wt% <177°C 3.5 4.0 177-350°C - 26.8 26.4 350-525°C - 28.9 28.0 >525 °C - 40.8 41.6 MoS2 derived from MoNaph in situ during reaction. With measurement error of ± 5 % With measurement error of ± 1 % Refer to Appendix B-3.1 for H 2 consumption calculations. The reaction using crystalline M0S2 had a much higher coke yield compared to the other two catalysts. The data show that reactions with exfoliated M0S2 and M0S2 derived 69 from MoNaph had very similar coke yields, liquid yields and residue conversions. However, the H/C ratio in the liquid product using exfoliated MoS 2 (1.55) was higher compared to that with MoNaph derived M0S2 (1.44). This result suggests that the exfoliated MoS 2 is more effective in hydrogenation during hydrocracking. In addition, as shown in Figure 4.5, the exfoliated MoS 2 had significantly higher activity for N and metals removal, and M C R and asphaltene conversion, compared to that of MoS 2 derived from MoNaph, suggesting that enhanced hydrogenation coupled with heteroatom hydrogenolysis (bond breakage caused by hydrogenation), occurred with the exfoliated catalyst. The enhanced hydrogenation in exfoliated MoS 2 compared to crystalline MoS 2 is suggested to be a consequence of the increased number of rim-edge sites associated with the exfoliated catalyst, consistent with the "rim-edge" model of Daage and Chianelli (1994). The reaction mechanism of Cold Lake heavy oil hydrocracking is very complex. The better liquid product quality obtained in the reaction using exfoliated MoS 2 compared to MoNaph derived MoS 2 in Cold Lake heavy oil hydrocracking is probably a consequence of a combination of hydrogenolysis and hydrogen transfer capability. 4.2.1.1 Effect of Sulfiding Reagent, H2S Note that in Table 4.2, 5 % H 2 S was present in the feed H 2 for the hydrocracking reaction using MoS 2 prepared in situ from MoNaph. The presence of H 2 S in the feed H 2 played a role in sulfiding the MoNaph precursor into MoS 2 . Further more, since the exfoliated MoS 2 was prepared ex situ, it was presumed that the sulfiding reagent would not contribute much to the reaction in this case. To confirm this assertion the hydrocracking reactions were 70 60 50 40 ™ 30 o E cu cm 20 10 • MoNaph • Exfoliated M o S 2 Error % S, ±1 N MCR Asphaltene V Ni ±1 ±2 ±10 ±2 ±2 Figure 4.5 Cold Lake heavy oil hydrocracking using M0S2 derived from MoNaph and exfoliated M0S2 (600 ppm Mo, 415°C, initial H 2 pressure of 3.5 MPa, reaction time of 1 hour). 71 conducted both in the presence of H2S/H2 and in pure H2 using exfoliated M0S2 catalyst. From the experimental data, it was found that there was no statistically significant difference (Appendix B-3.5) in coke, liquid and gas yields from Cold Lake heavy oil hydrocracking with exfoliated M0S2, either the presence of 5 % H2S in H2 or in pure H2. Therefore, subsequent hydrocracking reactions using exfoliated M0S2 as catalyst were carried out in pure H2 gas. 4.2.2 Effect of Process Conditions on Hydrocracking Using Exfoliated M0S2 Heavy oil hydrocracking requires severe reaction conditions and the catalyst activity is always related to process conditions. In the following sections, the effect of reaction temperature and the catalyst Mo concentration on Cold Lake heavy oil hydrocracking using exfoliated M o S 2 has been examined 4.2.2.1 Reaction Temperature The effect of reaction temperature on Cold Lake heavy oil hydrocracking using exfoliated M0S2 was investigated in the temperature range 400°C to 430°C. The lower temperature is typical of the on-set of thermal cracking, and the maximum represents conditions where thermal cracking reactions dominate. The results are presented in Table 4.3. With increasing temperature, the coke and gas yield increased and the liquid yield decreased. Hydrogen consumption also increased as did the sulfur and nitrogen removal, as expected. Chemical bond breaking especially breaking of carbon-carbon bonds are mainly thermal reactions (Gray, 1994). The breakage of carbon-carbon bonds yields free radicals that are highly reactive intermediates with an unpaired electron. The radicals can either lead to 7 2 distillate and process gas or undergo condensation reactions that yield coke. Therefore, at higher temperature, higher coke and gas yields were the result of more cracking and radical reactions. It is noted, however, that the M C R and asphaltene conversion reached a maximum at 415°C, and consequently, subsequent catalyst activity measurements were made at this temperature. Table 4.3 Effect of temperature on Cold Lake heavy oil hydrocracking using exfoliated M0S2 (600 ppm Mo, initial H2 pressure 3.5 MPa, reaction time of 1 hour). Temperature, °C 400 415 430 Coke yield, wt % 0.91 1.05 ±0.06 6.48 Liquid yield, wt % 97.5 96.5 ±0 .5 89.8 Gas yield, wt % 1.6 2.5 ±0 .6 3.2 H2 consumption4"4"1", wt % 27.2 53.1 ±5 .7 73.9 S removal4", wt% 14.2 28.1 30.8 N removal+, wt% 5.9 14.7 23.5 M C R conversion4"4", wt% 12.0 39.5 22.0 Asphaltene conversion, wt% 27.4 46.7 ±7 .1 45.7 With measurement error of ± 1 %; With measurement error of ± 2 % Refer to Appendix B-3.1 for H 2 consumption calculations. 4.2.2.2 Effect of Mo Concentration Catalyst concentration is an important process parameter in hydrocracking. It affects the coke yield and liquid product quality as well as the overall process economics (Panariti et 73 al., 2000b). The effect of exfoliated M0S2 concentration in Cold Lake heavy oil hydrocracking was studied at 415°C with Mo concentration of 360, 600 and 900 ppm Mo. In these experiments, increased volumes of exfoliated M0S2 (2 % MoS2-in-water suspension) were added to the oil to achieve the desired Mo concentration in the heavy oil. Consequently, the corresponding water content in the reaction mixture increased and was approximately 2.8, 4.7 and 6.8 wt% as the Mo content was set at 360, 600 and 900 ppm Mo, respectively. The effect of Mo concentration during hydrocracking was the main concern here, whereas the role of water is further examined and discussed in Section 4.2.3. The comparisons of hydrocracking reactions and product liquid qualities using different Mo concentrations of exfoliated M0S2 are shown in Table 4.4. Minimum coke was produced at a Mo concentration of 600 ppm. However, the coke yield was found to increase with further increases in Mo catalyst concentration. These results agree with those reported by Panariti et al. (2000b) and are explained by two phenomena taking place concurrently. The high level of hydrogenation of the heavy oil achieved in the presence of significant amounts of catalyst reduces asphaltene stability and this promotes coke formation. The M0S2 microcrystals dispersed in the oil may also be responsible for providing additional nucleation sites that lead to the precipitation of solids. Consequently, as the catalyst concentration increases, coke formation may prevail over coke suppression. The data in Table 4.4 also show that the liquid yield was not significantly affected by the M0S2 concentration. This observation agrees with the kinetic analysis of Panariti et al. (2000b) that since unsupported catalysts do not contain acidic functions, carbon-carbon bond cleavage reactions are thermally controlled. The H/C atomic ratio increased with higher Mo concentration, consistent with the increased of H2 consumption. The liquid product showed the best quality and the lowest coke yield at 600 ppm Mo in the oil. 74 Among the three Mo concentrations, the S, N and V removals, and M C R conversion were found to be highest at 600 ppm Mo. Table 4.4 Effect of Mo concentration on Cold Lake heavy oil hydrocracking using exfoliated M o S 2 (415°C, initial H 2 pressure of 3.5 MPa, reaction time of 1 hour). Mo concentration (ppm) 360* 600 900 Coke yield, wt% 1.55 1.05 ±0.06 1.76 ±0.12 Liquid yield, wt% 96.0 96.5 ±0 .5 96.0 ±0 .7 Gas yield, wt% 2.5 2.5 ±0 .6 2.3 ±0 .9 H 2 consumption""", wt% 54.4 53.1 ±5 .7 66.9 ± 13 Liquid product quality: Liquid H/C atomic ratio 1.49 1.55 1.61 S removal, wt% 18.8 28.1 26.9 N removal, wt% 5.9 14.7 11.8 M C R conversion, wt% 13.1 39.5 36.0 Asphaltene conversion, wt% 48.6 46.7 ±7 .1 39.4 ±7 .0 N i removal, wt% 13.5 14.3 .- 15.6 V removal, wt% 12.5 18.9 15.9 * Input gas used 5 %H2S/ 95 %H2. Refer to Appendix B-3.1 for H 2 consumption calculations. Note that the measurement error % for the H/C, S and N removal is ± 1 %; for MCR, Ni and V removal, the error is ± 2 %. 4.2.2.3 Effect of Concentration and Temperature Figure 4.6 shows the surface plot for the coke, liquid, and gas yield and the asphaltene conversion at different temperatures and Mo concentrations using exfoliated MoS 2 . It was 75 found that in the studied temperature and catalyst concentration range, there was a minimum coke yield and a maximum liquid yield at 415°C and 600 ppm. Generally, gas yield increased with reaction temperature and liquid yield reduce with temperature among the 3 temperatures studied. (ppm Mo) (ppm Mo) Cone. : Catalyst concentration Temp. : Temperature Figure 4.6 Coke, liquid, gas yield and asphaltene conversion at different Mo concentrations and reaction temperatures in Cold Lake heavy oil hydrocracking using exfoliated M0S2 (initial H2 pressure of 3.5 MPa, reaction time of 1 hour). 76 4.2.3 Effect of Properties of Exfoliated MoS 2 As described in Section 3.1.1, during the catalyst preparation, M0S2 was exfoliated by reacting with water and consequently the M0S2 was dispersed in water. The exfoliated M0S2 dispersed in water can be collected through centrifuging and redispersed into other solvents. In the present study, exfoliated M0S2 was redispersed in decalin. The concentration of exfoliated M0S2 in the solvent referred to herein as the dispersion M0S2 concentration, was then adjusted to the desired value, typically 2 wt% MoS 2 in the solvent. The dispersion concentration and dispersing solvent might affect the activity of the exfoliated M0S2 catalyst and consequently the effect of these properties were evaluated. 4.2.3.1 Dispersion Concentration The concentration of M0S2 in the exfoliated M0S2 dispersion might affect the degree of dispersion of the M0S2 in the heavy oil and this in turn might impact catalyst activity. To investigate this factor, exfoliated M0S2 was prepared with varying amounts of M0S2 dispersed in water. Experiments were carried out at a fixed Mo concentration of 600 ppm using exfoliated M0S2 with an M0S2 concentration of 2.0 and 8.4 wt% in water. In each experiment, the amount of dispersion added to the oil provided 600 ppm Mo in the reactant mixture, but the amount of water in the reactant oil was 4.7 and 1.0 wt%, respectively. Therefore, the effect of dispersion concentration also corresponded to the effect of water concentration. The data of Table 4.5 suggest that at the higher water concentration, the catalyst activity improved, yielding higher asphaltene conversions and reduced coke yield. 77 Table 4.5 Effect of M0S2 concentration in water suspension during Cold Lake heavy oil hydrocracking using exfoliated M0S2 (600 ppm Mo, 415°C, initial H 2 pressure of 3.5 MPa, reaction time of 1 hour). M0S2 concentration in water suspension, wt% 2.0 8.4 Water concentration in oil, wt% 4.7 1.0 Coke yield, wt% 1.05 ±0.06 1.61 Liquid yield, wt% 96.5 ± 0.5 95.3 Gas yield, wt% 2.5 ±0 .6 3.1 Asphaltene conversion, wt% 46.7 ±7 .1 23.6 At much higher concentrations, Yoneyama and Song (1999) also reported a beneficial effect of water. They reported that M0S2 catalysts generated from ammonium tetrathiomolybdate (ATTM) in the presence of n-tridecane solvent with added water under H2 pressure at 300-400°C, was much more active than catalyst prepared from A T T M alone. These catalysts were tested for carbon-carbon bond cleavage and naphthalene ring hydrogenation using the model compound 4-(l-naphthylmethyl)bibenzyl (NMBB). Although the results indicate that water promoted hydrogenation or hydrocracking reactions, the amount of water added in this system (the N M M B : Water weight ratio was 0.56) was much greater than the amount of water used in the present study. On the other hand, with increasing M0S2 dispersion concentration, this might reduce the degree of M0S2 catalyst dispersion within the oil which possibly reduce the catalyst activity and led to higher coke yield and lower asphaltene conversion. 78 4.2.3.2 Effect of Dispersing Solvent An important difference between the M0S2 prepared from MoNaph and that prepared by exfoliation is their respective dispersing medium. In this case, the exfoliated M0S2 was dispersed in water, while MoNaph is an oil soluble precursor that yields solid MoS 2 dispersed in the heavy oil. The solvent used during preparation of the exfoliated M0S2 may affect the degree of dispersion in the oil and the interaction between M0S2 particles and the heavy oil during reaction. Both factors may in turn influence the catalyst performance. Furthermore, different solvents were reported to have a significant effect on the product yield both in thermal hydrocracking (Rahimi and Gentzis, 2003) and catalytic hydrocracking (Duangchan , 1998), although the latter studies were carried out in up to 50% solvent. In the present study, exfoliated MoS 2 was dispersed in water and the non-polar organic solvent, decalin, prior to being added to the heavy oil. The effect that different dispersing solvents (water versus decalin) may have on activity was investigated at 900 ppm Mo and the results are compared in Table 4.6 and Figure 4.7. The exfoliated M0S2 dispersed in decalin gave a slightly higher coke yield as well as lower M C R removal compared to the exfoliated M0S2 dispersed in water. The results are consistent with the positive effect of water noted previously. In addition, the H/C atomic ratio of the product liquid oil was much lower (1.45) using exfoliated M0S2 dispersed in decalin compared to that of exfoliated M0S2 in water (1.61). However, in terms of the S, N , heavy metal and asphaltene removal, better results were obtained using exfoliated MoS 2 dispersed in decalin. 79 Table 4.6 Effect of solvent used to disperse exfoliated M0S2 on Cold Lake heavy oil hydrocracking (900 ppm Mo, 415°C, initial H 2 pressure of 3.5 MPa, reaction time of 1 hour). Exfoliated M0S2 dispersing medium Water Decalin Coke yield, wt% 1.76 ±0 .12 2.25 Liquid yield, wt% 96.0 ± 0 . 7 95.6 Gas yield, wt% 2.3 ± 0.9 2.1 H2 consumption, wt% 66.9 ± 13 45.3 Asphaltene conversion, wt% 39.4 ± 7 . 0 47.0 Liquid H/C atomic ratio 1.61 1.45 40 35 i 30 I 25 in 1— a> £ 20 o o 1 15 o E <u or 10 Error % S ±1 M C R ±2 • Decalin • Water V ±2 Figure 4.7 Effect of decalin and water as solvent to disperse exfoliated M0S2 in Cold Lake heavy oil hydrocracking liquid product oil (900 ppm Mo, 415°C, initial H2 pressure of 3.5 MPa, reaction time of 1 hour). 80 From the catalyst structure point of view, exfoliated M0S2 dispersed in decalin had a larger stack height than exfoliated M0S2 dispersed in water (Table 4.1) and the different results obtained in the hydrocracking reactions may be due to the difference in the number of M0S2 layers of the crystallite. In addition to that, exfoliated M0S2 dispersed in decalin may also have a better interaction between catalyst and heavy oil during reaction. As noted before, the increase in Mo concentration reported in Section 4.2.2.2, also corresponds to an increase in water concentration in the slurry. The study of dispersion concentration and dispersing solvent, showed that increasing water concentration or using water as catalyst dispersing solvent suppressed coke formation and increased M C R conversion. This suggests that the results obtained by increasing Mo concentration could be due to the increase in Mo concentration as well as an effect of water. 4.2.4 Hydrocracking of Different Petroleum Fractions as Feedstock The characteristics of heavy oil fractions change' with boiling point. Not surprisingly, the coke yield is reported to increase when cut point of the oil fraction increases (Rahimi et al., 2001). Chung & X u (2001) characterized the supercritical fluid extraction narrow cuts of Athabasca heavy oil and the heavy oil derived residues and reported that the conversion for most key species such as MCR, metals and N in hydrocracking is by partitioning. Sulfur species that convert in hydrocracking are mostly from the front-end of the heavy oil. There is only a small S reduction in the end-cut. End-cut that contains most of the problematic species is difficult to upgrade into useful product. In an effort to upgrade the end-cut of heavy oil, hydrocracking feedstock that was separated physically from the original heavy 81 oil has been proposed. In the present study, fractions of oil, generated by ultrafiltration according to molecular size (< 0.1 um) to yield filtrate and retentate with different properties (Lai and Smith, 2001), were used as feedstock in hydrocracking using exfoliated M0S2. Results obtained are expected to assist in identifying relationships between feedstock properties and catalyst performance. The catalyst activity of exfoliated M0S2 with the different feedstocks shown in Table 3.1, is given in Table 4.7. Coke yield had a linear relationship with the feedstock M C R content (Figure 4.8), which is consistent with the definition of M C R as a measure of propensity for coke formation. H2 consumption was also found to increase with feedstock asphaltene content, suggesting that H2 was used in the asphaltene cracking during hydrocracking. However, the residue conversions for different feedstock were within the range 21 to 25 wt%. The highest residue conversion was obtained for the original Cold Lake heavy oil while the lowest was for the filtrate. There was no direct relationship observed for the liquid H/C atomic ratio to the feedstock asphaltene conversion. H/C atomic ratio of hydrocracked filtrate and retentate was found to decrease, compared to the initial H/C atomic ratio. This shows that the hydrocracking reaction did not achieve its process objective of increasing H/C ratio of the oil. It is known that hydrocracking produces highly condensed aromatics. Dealkylation and aromatization reactions under hydrocracking conditions will produce coke and gases as ultimate products. Hence, H/C ratios of hydrocracked liquid product could be lower than that of the feedstock. 82 Table 4.7 Different feedstock hydrocracking using exfoliated M0S2 (600 ppm Mo, 415°C, initial H2 pressure of 3.5 MPa, reaction time of 1 hour). Feedstock Filtrate Cold Lake Retentate Feed properties: Asphaltene content, wt% 7.0 17.9 27.0 M C R 10.3 13.6 21.2 Product properties: Coke yield, wt% 0.44 1.05 ±0.06 2.38 Liquid yield, wt% 97.3 96.5 ± 0.5 94.6 Gas yield, wt% 2.2 2.5 ±0 .6 3.0 H2 consumption, wt % 42.2 53.1 ±5 .7 59.6 Residue conversion, wt% 21.7 25.3 23.4 Liquid H/C atomic ratio 1.54 1.55 1.43 Liquid boiling point, wt% — <177°C 4.9 4.0 4.4 177-350°C 26.4 26.4 24.5 350-525°C 30.5 28.0 26.1 >525 °C 38.2 41.6 45.1 * The oil that passed through the filter is termed filtrate and that which did not pass through is termed retentate. 83 2.5 j 2.0 -K 1-5-0) '>. « 1.0 -o o 0.5 -0.0 -0 5 10 15 20 25 Feed MCR content, wt % Figure 4.8 Coke yield versus feedstock with different M C R in hydrocracking using exfoliated M0S2. (Reaction Conditions: 600 ppm Mo, 415°C, initial H2 pressure of 3.5 MPa, reaction time of 1 hour; trend line). Table 4.8 shows the S, N and metal removal as well as asphaltene and M C R conversion of the different feedstocks. The highest metals removal was found in the hydrocracking reaction using the retentate, the heaviest feedstock. However, the highest M C R and asphaltene conversions were obtained for the original Cold Lake heavy oil. There was negligible asphaltene conversion observed in the reaction using the filtered oil. Similar S removal was obtained in all cases. Direct comparison of total heteroatom removal cannot be made using feed with different properties. A comparison was made based on the assumption that in ultrafiltration 60 wt% of the Cold Lake heavy oil reports to retentate and 40 wt% to fitrate (Lai and Smith, 2001). Considering the total heteroatom removal from both filtrate and retentate, based on the original feed oil (See Appendix B-3.6), higher metals removal, especially N i , was obtained (21.0 %) compared to direct hydrocracking of 84 Cold Lake heavy oil (14.3 %). Similar S and N removal was observed. This suggests that hydrocraking of Cold Lake heavy oil partitioned by ultrafiltration can improve heteroatom, especially metal removal. Table 4.8 Effect of feedstock in hydrocracking using exfoliated M0S2 (600 ppm Mo, 415°C, initial H2 pressure of 3.5 MPa, reaction time of 1 hour). Filtrate Retentate Total* Cold Lake S removal, wt% 29.7 31.2 30.7 28.1 N removal, wt% 10 14.9 13.4 14.7 M C R conversion, wt% 16 24.7 22.6 39.5 Asphaltene conversion, wt% 0 31.5 26.9 50.3 V removal, wt% 7.9 19.9 17.5 18.9 N i removal, wt% 6.1 24.5 21.0 14.3 * Total removal or conversion calculated from filtered oil and retentate based on 40/60 filtrate/retentate in lOOg basis of Cold Lake heavy oil (See Appendix B-3.6). 4.3 Characterization of the Recovered Coke The coke recovered from the hydrocracking reactions was characterized by BET surface area, X R D , infrared analysis, SEM-EDX and elemental analysis to assist in developing an understanding of the fate of the dispersed catalyst in slurry phase hydrocracking. The coke was analyzed for M0S2 content and characterized to determine the structure of the MoS 2 in the coke. The role of solvent on the M0S2 dispersion might also be deduced by determining the state of the M0S2 catalyst and its concentration in the coke after reaction. 85 4.3.1 BET Surface Area of the Recovered Coke The BET surface area of coke produced from reactions performed under different operating conditions are shown in Table 4.9. Basically, the coke produced during heavy oil hydrocracking exhibited low BET surface area (< 10 m2/g). This is consistent with data reported by Peureux et al. (1995). Coke recovered from hydrocracking at 415°C using MoNaph with a Mo concentration of 600 ppm had the highest surface area compared to that of crystalline M0S2 and exfoliated M0S2. Coke recovered from higher temperature reaction was found to have lower BET surface area. Further, the heavier the feedstock, the lower the coke BET surface area. This is clearly observed in the coke recovered from reaction using filtrate (6.3 m2/g) compared to original Cold Lake heavy oil (0.7 m2/g) and the retentate (< 0.1 m2/g). 4.3.2 XRD Measurement of the Recovered Coke Figure 4.9 shows the comparisons of the X R D patterns for the coke recovered from different experiments. The peak corresponding to M0S2, especially the plane (002) at 20 = 14°, was not observed in any of the samples. This means that the M0S2 added to the reactor had not gone through a restacking process during reaction. The most significant peak observed corresponded to graphite at 20 = 26.5°. In addition, peaks corresponding to iron sulfide and nickel sulfide in the coke were observed. 86 Table 4.9 BET surface area of coke recovered from different hydrocracking reactions (initial H2 pressure of 3.5 MPa and a reaction time of 1 hour). Run no. Catalyst Concentration, ppm Mo Reaction temperature, °C BET surface area, m2/g 32 Crystalline MoS 2 600 415 1.1 25 MoNaph 600 415 4.4 21* exMoS 2 -W 600 415 6.3 22** exMoS 2 -W 600 415 <0.1 24 exMoS 2-W 600 400 -11 exMoS 2-W 360 415 2.1 20 exMoS 2-W 600 415 0.7 23 exMoS 2-W 900 415 1.6 19 exMoS 2-D 900 415 1.1 26 exMoS 2 -W 600 430 0.4 14 - - 430 0.3 Note: exMoS2-W refer to exfoliated MoS2 dispersed in water; exMoS2-D refer to exfoliated MoS2 dispersed in decalin. * Using filtrate with 7% asphaltene in feedstock. ** Using retentate with 27% asphaltene in feedstock. 87 Figure 4.9 X R D diffractograms for the coke recovered from catalytic hydrocracking of heavy oil with (a) 600 ppm Mo of exfoliated MoS 2 at 415°C; (b) 600 ppm Mo of MoNaph at 415°C; (c) 900 ppm Mo of exfoliated MoS 2 dispersed in decalin at 415°C; (d) 900 ppm Mo of exfoliated MoS 2 at 415°C and (e) 600 ppm Mo of exfoliated MoS 2 at 430°C (initial H 2 pressure of 3.5 MPa and a reaction time of 1 hour). 88 4.3.3 Infrared Analysis Selected coke samples recovered from various catalytic hydrocracking experiments with heavy oil using different catalysts, solvents, operating conditions or feedstock with different asphaltene content, were analyzed using DRIFTS. The spectra are shown in Figure 4.10. The spectra provide more detailed information on the chemical character of the coke. Basically, all spectra exhibit a similar trend. There are two main frequency regions in which the detected bands occur (2800 - 3100 and 700 - 1800 cm"1). The two regions contain paraffmic and aromatic bands. The band at 1600 cm"1 is the so called coke band. In the range of the C-H stretching vibration in alkanes (2800 - 3000 cm"1), the bands occurring at 2855, 2924 and 2953 cm"1 correspond to the symmetric vibrations of CH2 and CH3, the asymmetric vibration of CH2, and the asymmetric vibration of CH3, respectively (Rozwadowski et al., 2001). These bands increase in intensity with both the reaction temperature (Figure 4.10 (d), (c), (a)) and the catalyst Mo concentration (Figure 4.10 (f), (c), (b)). They are also higher in coke recovered from reaction using exfoliated M0S2 in water then that of exfoliated M0S2 in decalin (Figure 4.10 (h)) as well as MoNaph (Figure 4.10 (i)). However, in terms of feedstock (Figure 4.10 (g), (c), (e)), the highest intensity was observed for coke recovered from reaction using the original Cold Lake heavy oil with asphaltene 17 %. The 3052 cm"1 band is due to the stretching vibration of C-H in polyalkenes or in aromatics. 89 c CO L a B CO c 3500 3250 3000 2750 2000 1500 1000 Wave number, cm" Figure 4.10 Infrared spectra for the coke recovered from catalytic hydrocracking of heavy oil with (a) 600 ppm Mo of exfoliated M o S 2 at 430°C; (b) 900 ppm Mo of exfoliated M o S 2 at 415°C; (c) 600 ppm Mo of exfoliated MoS 2 at 415°C; (d) 600 ppm Mo of exfoliated MoS 2 at 400°C; (e) retentate with 600 ppm Mo of exfoliated MoS 2 at 415°C; (f) 360 ppm Mo of exfoliated M o S 2 at 415°C (g) filtrate with 600 ppm Mo of exfoliated MoS 2 at 415°C; (h) 900 ppm Mo of exfoliated M o S 2 i n decalin at 415°C; and (i) 600 ppm Mo of MoNaph at 415°C (initial H 2 pressure of 3.5 MPa and a reaction time of 1 hour). 90 The band intensities in the range 700 - 1800 cm" increased with the paraffinic bands (2800 to 3000 cm"1). The bands at 1600 cm"1 are assigned to coke and attributed to the stretching vibration of C=C in microcrystalline graphitic structures, which are present in polycyclic aromatic compounds and might also constitute the carbonaceous deposits. The bands at 1460 and 1377 cm"1 are ascribed, respectively, to the asymmetric and symmetric bending vibrations of CH3. It is to be expected that the band due to the in-plane bending vibration of CH2 (at 1465 cm"1) is also present, but it might overlap with the 1460 cm"1 band (Rozwadowski et al., 2001). A relatively high band at 1460 cm"1 in all the spectra suggests that aliphatic groups connected with aromatic rings are present in a comparatively high amount. Although the results from the infrared analysis did not reveal information about the fate of the catalyst in the coke and the catalyst activity, it provided some information on the chemical characteristic of coke. Efforts have been made to correlate the infrared analysis results with the BET surface area or X R D , in order to help in elucidating the mechanism of coke formation in the heavy oil hydrocracking reactions. Comparing the catalyst used during reaction, coke from reactions using MoNaph derived M0S2 had larger BET surface area than that from exfoliated M0S2. In general, the coke from MoNaph is more porous, less crystalline and less graphitic compared to the coke from reactions using exfoliated M0S2. Other than the above-mentioned observations, there was no further correlation or conclusion that could be drawn from these data. 91 4.3.4 Metal Content Analysis The coke recovered from hydrocracking reactions was analyzed for Mo, Fe, N i and V content. The metal that resided in the coke included the metal removed from the oil after reaction. It was assumed that all metal distributed evenly in the coke. The results of the analysis are shown in Table 4.10. The amount of Fe and N i in the coke are not reported due to the fact that the amount of Fe and N i in the coke was found to be more than the total amount of original metal in the feed. In particular Fe was found to be few 100 times higher in the coke than feed oil. It is likely that much of the Fe came from the reactor during the process of recovering the coke from the reactor by scraping the reactor surface after reaction. Table 4.10 Ratio of Mo and V in coke recovered from different hydrocracking reactions (initial H 2 pressure of 3.5 MPa and a reaction time of 1 hour). Run Catalyst Concentration, Reaction Moc/Mofa v c/v f a no. . ppm Mo temperature, °C 25 MoNaph 600 415 0.90 0.38 20 exMoS 2 -W 600 415 0.95 0.09 22** exMoS 2 -W 600 415 0.92 0.12 21* exMoS 2 -W 600 415 0.94 0.05 11 exMoS 2 -W 360 415 1.0 0.13 23 exMoS 2 -W 900 415 0.96 0.19 19 exMoS 2-D 900 415 0.89 0.14 26 exMoS 2 -W 600 430 0.90 0.38 a Ratio of metal in coke to metal in feed with error ±2%. * Using filtrate oil with 7 % asphaltene in feedstock. ** Using retentate with 27 % asphaltene in feedstock. 92 The ratio of V in coke was not consistent in different reactions. It ranged from 5 to 40 %. However, from the analysis of all coke samples, more than 90 % of Mo in feed was found to report to the coke. In order to confirm this, the corresponding product liquid oils were also analyzed for the metal content. A l l liquid samples were analyzed as undetectable molybdenum (< 1 ppm) in the liquid. The same result was obtained for the reaction using MoNaph. Hence, it can be concluded that most of the Mo (> 90 %) in the feed resided in coke after the one-hour reaction. 4.3.5 SEM and EDX of Coke Figure 4.11 shows the S E M of coke recovered from reaction using exfoliated MoS 2 . The light coloured pieces were metal fragments from the reactor. In addition, an E D X mapping was also done on the coke to check the metal distribution. The result is shown in Figure 4.12. It was found that Mo and S were well distributed throughout the coke. Figure 4.11 S E M of coke recovered from hydrocracking using exfoliated MoS 2 . 93 Figure 4.12 Energy dispersive x-ray (EDX) mapping of coke collected after hydrocracking reaction. 4.4 Catalyst Recycle As already noted, the possibility of catalyst recycle in heavy oil upgrading using a slurry reactor may be an important consideration at high Mo concentrations (> 200 ppm) in the oil slurry. Consequently, the possibility of recycling the exfoliated M0S2 was examined. Analysis of the product oil and coke showed that > 90 % of the Mo resided in the coke after a batch reaction time of 1 hour. X R D and E D X analysis of the coke confirmed that the Mo was well distributed throughout the coke. To recover the used M0S2 catalyst from the solid coke is a challenge as the concentration involved is very low (< 0.5 wt%). Consequently, the coke produced from the different reactions was reused to test its catalytic activity and thereby simulate the recycling of the catalyst coke. Figure 4.13 reports the results obtained from heavy oil upgrading using the coke produced from a previous experiment using exfoliated M0S2 catalyst. The amount of coke 94 added to the oil was set to provide an equivalent 600 ppm Mo in the oil. The results are compared with an experiment in which coke produced from thermal hydrocracking of the same heavy oil with no detectable Mo content, was added to the Cold Lake feed. After 1 hour reaction in the batch reactor, the results clearly show that the recycled coke had significant activity, indicating that the M0S2 dispersed in the produced coke remains active for hydrocracking, albeit at a lower level than the original exfoliated MoS 2 , but higher than when the coke added was free of Mo. Although the coke yield remained low with the recycled exfoliated M0S2, the removal of S, M C R and asphaltene conversion all decreased. These results suggest that the exfoliated M0S2 remains effective for hydrogen transfer to limit coke production. However, since the M0S2 is dispersed within the coke, diffusion effects probably limit the catalyst effectiveness for hydrogenation and hydrodesulfurization reactions. coke yield Sulfur removal MCR removal asphaltene conversion Figure 4.13 Activity of recycled M0S2 during Cold Lake heavy oil hydrocracking (600 ppm Mo, 415°C, initial H2 pressure of 3.5 MPa and a reaction time of 1 hour). 95 4.5 Summary The potential of using, exfoliated M0S2 as a dispersed catalyst for heavy oil hydrocracking, specifically for Cold Lake heavy oil, was examined. It was found that the exfoliated M0S2 has comparable performance in coke yield to that of M0S2 derived from MoNaph at 415°C. In addition, exfoliated M0S2 produces better liquid product quality: better S, N and metals removal as well as M C R and asphaltene conversion than with M0S2 derived from MoNaph. It is believed that the different results obtained in these reactions was likely a consequence of different M o S 2 structure between the M0S2 derived from MoNaph and the exfoliated M0S2. Exfoliated MoS 2 was more effective in hydrogenation and carbon-heteroatom bond breakage than M0S2 derived from MoNaph. For the range of conditions examined, the best hydrocracking conditions for Cold Lake heavy oil using exfoliated M0S2 was found to be 415°C and 600 ppm Mo. In addition, after a one-hour batch reaction, more than 90% of the catalyst was found to be associated with the produced coke, and the spent catalyst-coke was shown to be active for heavy oil hydrocracking, suggesting that catalyst/coke recycle is possible for the M0S2 catalyst. The results show that exfoliated M0S2 could be an attractive dispersed catalyst for heavy oil hydrocracking. The concept of increased rim-edge sites associated with the exfoliated M0S2 that promotes hydrogenation reactions provides a suitable explanation of the observed activities of exfoliated M0S2 and MoS 2 prepared in situ from the decomposition of MoNaph. The structural effects of M0S2 is studied further using model compounds in Chapter 5. 96 Effect of MoS 2 Catalyst Morphology on Hydroprocessing Reactions New regulations for fuel quality, driven by environmental concerns, require an improvement in performance of hydroprocessing catalysts, especially in hydrode-sulfurization (Grange and Vanhaeren, 1997). Improvements in catalyst activity require knowledge of the relationships between catalyst morphology, structure and activity. Unsupported exfoliated M0S2 is a potentially good model catalyst since it allows one to study the structural effects of M0S2 on catalyst activity without interference from the catalyst support and catalyst promoters. M0S2 prepared by the decomposition of molybdenum naphthenate (MoNaph) in the presence of H2S was chosen as a catalyst for comparative study, since M0S2 derived from MoNaph has been shown to be an effective catalyst for upgrading the end products from Two papers on this chapter have been accepted for publication. Tye, CT. and Smith, K.J. (2006) Catalytic Activity of Exfoliated MoS2 in Hydrodesulfurization, Hydrodenitrogenation and Hydrogenation Reactions. Topics in Catalysis. In press. Tye, CT. and Smith, K.J. (2006) Hydrodesulfurization of Dibenzothiophene over Exfoliated MoS2 Catalyst. Catal. Today. Accepted. 97 coal liquid (Curtis and Pellegrino, 1989), and heavy oil (Bearden and Aldrich, 1981). M0S2 derived from MoNaph has often been utilized as a reference compound in studies of heavy oil hydrocracking catalysts (Del Bianco et al., 1993). In addition, M0S2 derived from MoNaph has been studied in various model compound reactions, including hydrogenation, hydrodesulfurization, hydrodenitrogenation and hydrodeoxygenation (Kim and Curtis, 1990; Ting et al., 1992; Rueda et al., 1997; Rueda et al., 2001). As demonstrated in Chapter 4, exfoliated M0S2 was found to give better liquid product quality in hydrocracking Cold Lake heavy oil compared to M0S2 derived from MoNaph. This indicates that exfoliated M0S2 and M0S2 derived from MoNaph have different activity in terms of hydrodesulfurization, hydrodenitrogenation as well as hydrogenation. In this chapter, a series of comparative studies have been conducted to evaluate the performance of M0S2 catalysts, prepared with varied morphology, in hydroprocessing reactions using model compounds. Four types of M0S2 catalysts were studied: exfoliated M0S2 and MoNaph derived M0S2 were tested in hydrogenation, hydrodesulfurization, hydrodenitrogenation and hydrodeoxygenation reactions. Crystalline M0S2 and ammonium heptamolybdate (AHM) derived M0S2 were examined for selected reactions. Note that the exfoliated M0S2, studied in all hydroprocessing reactions in this chapter, refers to exfoliated M o S 2 dispersed in decalin. Although exfoliated M0S2 dispersed in water was also tested in model compound reactions, water decreased the activity of M0S2. This could be partly due to the immiscibility of water and the organic solvent (n-hexadecane) used in the model compound reactions. The reduced interaction between exfoliated M0S2 dispersed in water and model reactant greatly reduced the catalyst performance. The effect of M0S2 dispersing solvent 98 was also studied in Cold Lake heavy oil hydrocracking and better performance in term of S, N , metal and asphaltene removal was observed using exfoliated M0S2 dispersed in decalin compared to exfoliated M0S2 dispersed in water (Section 4.2.3.2). Given low activity, the hydroprocessing results of exfoliated M0S2 dispersed in water using model compounds are given in Appendix C but are not discussed in this chapter. Hydrodesulfurization of dibenzothiophene (DBT), hydrodenitrogenation of quinoline and carbazole, hydrogenation of naphthalene as well as hydrodeoxygenation of phenol have been examined. The model compounds used in each case are representative of the constituents of petroleum that are pertinent to process modeling and catalyst development (Girgis and Gates, 1991; Furimsky, 2000; Mochida and Choi, 2004). The results from these experiments are expected to provide M0S2 structure-activity information on hydrogenation, hydrodesulfurization, hydrodenitrogenation and hydrodeoxygenation reactions. Although hydrodeoxygenation plays a less important role in typical heavy oil hydroprocessing reactions, it is of significant interest in processing oil derived from biomass (Furimsky, 2000) and interest in these reactions is increasing with an increase in the use of biomass as an alternative fuel. Model compound reactants provide insight into the heavy oil upgrading process while avoiding the complex nature of heavy oils. 5.1 Catalyst Characterization In order to study catalyst structure-activity relationships, the properties of the catalysts must be determined. The properties of unused, exfoliated M0S2 and crystalline M0S2 were reported in Table 4.1 and discussed in Chapter 4. In the case of M0S2 synthesized by in situ decomposition of MoNaph and ammonium heptamolybdate (AHM), the properties of the 99 solid material recovered from the reactor after reaction were determined. The solids recovered from reaction using exfoliated M0S2 were characterized in addition to the unused, exfoliated M0S2, as discussed in Section 4.1. Since the concentration of catalyst used in each reaction was very low (600 ppm), very small amounts (< 0.1 g) of catalyst were recovered after each reaction. There were cases in which the amount of recovered solid was insufficient for analysis, especially for those analyses that needed a minimum sample weight > 0.1 g. Nevertheless, representative recovered samples were analyzed. In what follows, BET surface area, SEM-EDX, X R D , and T E M analyses on recovered catalysts are reported and discussed. 5.1.1 B E T Surface Area BET surface areas for the catalysts recovered from different reactions are presented in Table 5.1. BET surface area analysis shows that the exfoliated M0S2 recovered from hydrodesulfurization and hydrodenitrogenation reactions had similar BET surface area (41 ± 6 m2/g and 37 ± 6 m2/g respectively). For the catalyst recovered from the decomposition of MoNaph, relatively higher BET surface area (250 - 500 m2/g) was observed, compared to the exfoliated M0S2. The high surface area of the MoNaph derived M0S2 was likely due to the formation of a carbonaceous phase with the M0S2 during the decomposition of MoNaph, which also preserved the high dispersion of M0S2 (Rueda et al., 1997). BET surface area for MoNaph derived M0S2 is correlated with the nature of the medium during reaction (Rueda et al., 2001). Higher BET surface area of the MoNaph derived M0S2 recovered from the hydrogenation reaction, compared to the hydrodesulfurization and hydrodenitrogenation reactions, could be due to the high concentration (10 wt%) of the 100 model compound (naphthalene) used for the hydrogenation reaction (< 1 wt% of reactant was used in hydrodesulfurization and hydrodenitrogenation reactions). In general, MoNaph derived M0S2 had a significantly larger BET surface area compared to the recovered exfoliated M0S2. Note that the BET surface area of A H M derived M0S2 was not determined since the amount of sample recovered after reaction was insufficient for BET analysis. Table 5.1 BET surface area of different catalysts recovered after reaction. (600 ppm Mo, 350°C, initial pressure 2.8 MPa, 5-hour reaction) Reaction Recovered exfoliated M0S2 MoNaph derived M0S2 (m2/g) (m2/g) 41 ± 6 250 ± 40 37 ± 6 300 ± 6 0 480 ± 5 0 5.1.2 SEM-EDX E D X analyses were performed on the recovered solid sample from various reactions with different M0S2 catalysts. Results from SEM-EDX confirmed the presence of Mo and S alone, with an average S/Mo atomic ratio of 2.5 ± 0.2. The average includes measurement of the recovered exfoliated M0S2, MoNaph derived M0S2 and A H M derived M0S2. Hydrodesulfurization Hydrodenitrogenation Hydrogenation 101 5.1.3 X R D Figure 5.1 shows the X R D diffractograms of the catalysts recovered from different reactions using the MoNaph as the catalyst precursor. Figure 5.2 presents the X R D diffractogram for A H M derived M0S2, recovered after a hydrodesulfurization reaction. MoNaph derived M0S2 and A H M derived M0S2 were more amorphous than the recovered exfoliated M0S2 shown in Figure 5.3 The average stack height of the M0S2 particles, estimated from the X R D line broadening of the peak corresponding to (002) plane at 14.4° was 1.8 nm (Figure 5.1) and 2.1 nm (Figure 5.2) for MoNaph derived MoS 2 and A H M derived M0S2, respectively. The recovered exfoliated M0S2 is compared to the unused exfoliated M0S2 in Table 5.2. The data suggest that some restacking occurred during reaction. The average number of M0S2 layers for the exfoliated M0S2, calculated from the stack height, increased from 9 to 14 after a 5-hour batch reaction. Table 5.2 Average stack height and average number of layers in crystallites from unused and recovered exfoliated M0S2. Exfoliated MoS 2 Unused Recovered after reaction* Stack height, nm 5.5 9.0 Number of layer 9 14 * Recovered after hydrodesulfurization reaction at 350°C, initial pressure 2.8 MPa, 5-hour reaction. 102 8000 Figure 5.1 X R D diffractograms of the solid recovered after reactions using MoNaph precursor; (a) reaction without reactant with only solvent; (b) hydrogenation (c) hydrodenitrogenation and (d) hydrodesulfurization reactions (600 ppm Mo, 350°C, initial pressure 2.8 MPa, 5-hour reaction). 1000 30 40 50 60 70 29 Figure 5.2 X R D diffractograms of the solid recovered after hydrodesulfurization reaction using A H M precursor (600 ppm Mo, 350°C, initial pressure 2.8 MPa, 5-hour reaction). 103 8000 H Figure 5.3 X R D diffractograms of the solid recovered after reactions using exfoliated M0S2 (a) hydrodesulfurization (b) hydrogenation and (c) hydrodenitrogenation reactions (600 ppm Mo, 350°C, initial pressure 2.8 MPa, 5-hour reaction). 5.1.4 TEM Observation and Analysis Selected samples of the M0S2 catalysts were observed using high resolution T E M . T E M micrographs • of the recovered catalysts are shown in Figure 5.4(a-c). A large lateral dimension or sheet morphology of exfoliated M0S2 is apparent and this morphology was retained after reaction (Figure 5.4(a)). The estimated slab length of the exfoliated M0S2 was an order of magnitude larger than that of MoNaph and A H M derived M0S2 (Figure 5.4(b-c)). MoNaph and A H M derived M0S2 exhibited similar slab lengths of about 10 nm. The average slab lengths of these two catalysts, shown in Table 5.3, were calculated from the distribution of the M0S2 slab lengths (Figure 5.5) determined from the T E M 104 micrographs. The distribution of the number of M0S2 layers for MoNaph and A H M derived M0S2 were also determined from the T E M micrographs and are shown in Figure 5.6. The average number of layers for MoNaph and A H M was 1.3 and 1.8, respectively. This result is lower than that estimated from the line-broadening of (002) peak of the X R D pattern and confirms that MoNaph and A H M lead to highly dispersed M0S2. An examination of the various catalyst properties showed that the catalyst before and after reaction exhibited similar properties. The characterization results of unused exfoliated M0S2, as discussed in Section 4.1 of Chapter 4 are similar to those of the exfoliated M0S2 recovered after reaction. 105 Figure 5.4 T E M of (b) MoNaph and, (c) A H M derived M 0 S 2 recovered after reaction (hydrodesulfurization reaction at 350°C, initial pressure 2.8 M P a , 5-hour reaction). 106 (a) 16 10 15 20 Length, nm (b) 12 10 15 20 Length, nm 25 Figure 5.5 Distribution of M0S2 slab length for (a) MoNaph and (b) A H M derived M0S2 reacovered after reactions (hydrodesulfurization reaction at 350°C, initial pressure 2.8 MPa, 5-hour reaction). 107 100 1 2 3 4 Number of MoS 2 layers Figure 5.6 Distribution of the number of layers for MoNaph and A H M derived M0S2. (hydrodesulfurization reaction at 350°C, initial pressure 2.8 MPa, 5-hour reaction). The catalyst characterization data can be summarized as follows. Four M0S2 catalysts (exfoliated M0S2, crystalline M0S2, MoNaph and A H M derived M0S2) were prepared each with different morphology. As discussed in Section 4.1, the exfoliated M0S2 had an average of 9 layers of M0S2 in a large sheet morphology (0.40 um) compared to the crystalline M0S2 with 57 layers and average slab length of 0.56 um. In contrast, M0S2 derived from MoNaph and A H M precursors had a slab length of approximately 10 nm with 1 to 2 layers of M0S2 in the crystallites. It is expected that X R D would slightly overestimate the stacking value since single layers were not accounted for by X R D whereas single layers can be visualized in T E M . A comparison of the properties of the M0S2 catalysts is made in Table 5.3. The activities of these catalysts are compared in the following sections. 108 Table 5.3 Summary of M0S2 catalyst properties Average Slab length, nm Stack height, nm Number of layers Crystalline MoS 2 560 ± 740 35.8 57 Exfoliated MoS 2 400 ± 460 5.5 9 MoNaph derived MoS2! 10 ± 10 1.8(0.8*) 2.8(1.3*) A H M derived M o S 2 b 9.3 ±9 .0 2.1(1.1*) 3.3 (1.8*) MoS2 synthesized by in situ decomposition of molybdenum naphthenate MoS2 synthesized by in situ decomposition of ammonium heptamolybdate From TEM analysis. 5.2 Hydrogenation of Naphthalene Hydrogenation is important in hydroprocessing as well as in primary upgrading. Recently, the reduction of aromatic compounds in fuels has become important, not only in lighter oil fractions, but also in middle distillates. Dispersed M0S2 catalysts were reported to be active in hydrogenation reactions especially in heavy oil upgrading (Panariti et al., 2000a). In the present study, the hydrogenation activities of exfoliated MoS 2 and MoNaph derived M0S2 were compared under hydroprocessing conditions (at 350°C). Naphthalene, an aromatic component with two fused rings, was used as the model reactant representing the aromatic components of middle distillate (Girgis and Gates, 1991). During the hydrogenation reaction, naphthalene is hydrogenated to tetralin and decalin, sequentially as shown in Figure 5.7. 109 Naphthalene (NTL) Tetralin (TTL) Decalin (DCL) Figure 5.7 Naphthalene reaction scheme. Figure 5.8 shows the product ratio profile for hydrogenation of naphthalene using exfoliated M0S2 and MoNaph derived M0S2. Both catalysts produced tetralin with little decalin (< 1 mol%). Therefore, tetralin was considered the only product in the hydrogenation of naphthalene and decalin was neglected in the rate constant calculation. M0S2 derived from MoNaph had been previously reported for the hydrogenation of naphthalene in a batch microreactor, with high concentration of catalyst (3000 ppm Mo) and reaction conditions relevant to coal liquefaction (380°C, 18.6 MPa) (Kim and Curtis, 1990). It was reported that MoNaph derived M0S2 promoted partial saturation of the multiring aromatic to hydroaromatic species but did not promote further saturation of the hydroaromatic. According to Kim and Curtis (1990), the catalytic reaction using MoNaph produced nearly complete naphthalene conversion, yielding tetralin as the primary product (92 %) and a small amount of decalin (5 %), compared to Ni-MoS2/Al203 that produced 9.5% tetralin and 86.6 %> decalin. The results of using MoNaph reported by Kim and Curtis (1990) are in agreement with those of the present study. 110 (a) o £ 0.6 g" 2 o 0.4 TD O A Naphthalene • Tetralin (b) 0.6 o -i—» • Naphthalene • Tetralin Figure 5.8 Profile of product ratio using (a) exfoliated M0S2 and (b) MoNaph derived M0S2 in hydrogenation of naphthalene (600 ppm Mo, 350°C, initial pressure 2.8 MPa, 5-hr reaction time; trend line). I l l Table 5.4 compares the conversion of naphthalene using exfoliated M0S2 catalyst and M0S2 derived from MoNaph with the thermal reaction (no catalyst). The M0S2 derived from MoNaph gave a higher conversion (35.8 %) after 5 hours reaction compared to the exfoliated M0S2 (27.0 %). This result showed that MoNaph derived M0S2 had a higher hydrogenation activity than exfoliated M0S2. Without catalyst, a naphthalene conversion of 17.6 % was obtained. Table 5.4 Comparison of hydrogenation of naphthalene using different M0S2 catalysts (600 ppm Mo, 350°C, initial pressure 2.8 MPa, 5-hr reaction time). Catalyst &NTL Conversion s - ' x l O 2 ml/(g Mo.s) x 102 % ++ No catalyst (thermal) 0.0010 ± 0.0004 - 17.6 Exfoliated MoS 2 - 5.74 ±0.96 27.0 MoNaph derived M o S 2 + - 7.7 ± 1.2 35.8 + MoNaph from ICN Biomedicals Inc. Data was extrapolated from 2-hour reaction. * Rate constant for disappearance of naphthalene, data are reported with 95% confidence limits. ** Naphthalene conversion with ±10 % error. Reaction rate constants for the disappearance of naphthalene, &NTL were calculated using the integrated equation for the first-order rate law. For a thermal reaction without catalyst, Eq. (5.1) was used: [M] K where [Mo] and [M\ denote the initial concentration and the concentration of reactant M, (in this case naphthalene) at time t, respectively. The rate constant for the disappearance of 112 M, KM has the unit of s"1 or hr"1. For a catalytic reaction, the rate constant was calculated using Eq. (5.2): m ^ = J f c „ . W . f (5.2) [M] where w is the weight of Mo (g), and the rate constant for the disappearance of M, KM has units of (g Mo-hrV'or (g Mo's)' 1 . The rate constant was determined from the slope of In versus t, as shown in Figure 5.9, and divided by the corresponding mass of Mo in catalytic reactions. Figure 5.9 shows that the experimental data fit the first order rate law well, with a correlation coefficient R 2 > 0.90. The rate constant for disappearance of naphthalene, & N T L , is reported in Table 5.4 for both the thermal and catalytic reactions. Hydrogenation of naphthalene in the presence of M0S2 catalysts was also carried out at different temperatures (325, 350 and 375°C). Note that the MoNaph used in all hydroprocessing experiments was from ICN Biomedicals Inc. (as mentioned in Section 2.5), except for the MoNaph used in studying the effect of temperature in the hydrogenation of naphthalene, which was supplied by K . & K Laboratories Inc. Although the conversion of naphthalene using catalyst precursors from different suppliers may be different, the mechanism of reaction and the effect of temperature during reaction should remain the same. The hydrogenation of tetralin to decalin was found negligible in reactions using either exfoliated M0S2 or MoNaph derived M0S2. However, tetralin could be dehydrogenated to naphthalene (Figure 5.7) since the reaction is reversible. Dehydrogenation of tetralin to naphthalene was reported at typical hydroprocessing 113 (a) (b) (c) Time, hr Figure 5.9 Plots of ln([NTLo]/[NTL]) versus time for hydrogenation of naphthalene with (a) no catalyst (thermal reaction); (b) MoNaph; and (c) exfoliated M0S2 (600 ppm Mo, 350°C, initial pressure 2.8 MPa, 5-hr reaction time; fitted line). 114 temperatures, usually greater than 340°C (Girgis and Gates, 1991). The lumped rate constants for hydrogenation, k\ and dehydrogenation, k.\ of naphthalene were calculated using the assumption that each step of the network in Figure 5.7 followed the first order rate law. Therefore, the kinetic equations used to simulate the reaction are as follows: - J P q T L J = k{[NTL] - k , [TTL] (5.3) dt _d\JTL\ = _ y t j N T L ] + £ ,[TTL] (5.4) dt where [NTL] and [TTL] refers to the concentration of naphthalene and tetralin, respectively. Note again that the production of decalin was ignored due to low concentration (< 1 mol%). The rate constants in the Eq.(5.3) and Eq.(5.4) were estimated using the Gauss-Newton optimization method. The details of the computation algorithm are given in Appendix G. The simulation results showed that dehydrogenation of tetralin to naphthalene only occurred to a significant extent at the reaction temperature of 375°C. L\ at 325°C and 350°C was negligible. The estimated rate constants (k\ and k.\) are summarized in Table 5.5. The apparent activation energy for the hydrogenation of naphthalene was calculated using the Arrhenius equation: k = A-e-{EalRT) (5.5) where k is the rate constant; A is the Arrhenius constant; Ea, denotes the apparent activation energy; R represents the gas constant and T is the reaction temperature. Ea was determined from the slope of the graph In k versus MT plotted in Figure 5.10. The estimated Ea values reported in Table 5.6 show that MoNaph and exfoliated MoS2had similar Ea values. This suggests that the rate-determining step during hydrogenation of naphthalene over the 115 temperature range considered, is thermally controlled and is not significantly affected by the catalyst structure. Table 5.5 Pseudo 1 s t order rate constant for hydrogenation of naphthalene at different temperature (initial pressure 2.8 MPa, 5-hr reaction time). Thermal Exfoliated MoS 2 MoNaph* derived MoS 2 Rate constant** s"1 xlO" 5 ml/(g Mo.s) x 102 ml/(g Mo.s) x 10 ~ kx at 325°C 0.63 ± 0.04 1.98 ±0.12 0.84 ± 0.04 350°C 1.00 ±0.40 5.56 ±0.40 2.43 ±0.12 375°C 2.73 ±0.14 9.31 ±0.49 3.18 ±0.15 k.i at 375°C + + - 13.8 ± 1.9 0.87 ±0.20 * MoNaph from K & K Laboratories Inc. ** Data are reported with 95% confidence limits ++ k.i at 325 °C and 350°C was not significant within 95% confidence limits, therefore they are not reported. Table 5.6 Apparent activation energy corresponding to conversion of naphthalene to tetralin. Reaction Apparent activation energy, Ea, kJ/mol Thermal reaction 94 ± 22 Exfoliated MoS 2 99 ± 1 8 MoNaph 86 ± 27 ** Data are reported with 95% confidence limits 116 1.50 1.55 1.60 1000/7-, K"1 1.65 1.70 1.50 1.55 1.60 1000/7-, K"1 1.65 1.70 ( C ) 1.50 1.55 1.60 1.65 1.70 1000/7", K-1 Figure 5.10 Plots of In k versus 1000/r for hydrogenation of naphthalene with (a) no catalyst (thermal reaction); (b) exfoliated M0S2 and (c) MoNaph ( fitted line). 117 5.3 Hydrodesulfurization of Dibenzothiophene Dibenzothiophene (DBT) is an organosulfur compound present in heavy oi l that is less reactive than thiols, sulfides and disulfides (Gates and Topsoe, 1997). For hydro-desulfurization reactions, D B T was chosen as a model reactant since the reaction mechanism is well established (Vasudevan and Fierro, 1996; Gates and Topsoe, 1997; Mochida and Choi , 2004). In addition, D B T allows the study of sulfur removal through different parallel pathways. One pathway is by direct sulfur-removal or hydrogenolysis leading to biphenyl, the other is by hydrogenation of an aromatic ring prior to sulfur removal. During hydrodesulfurization, sulfur is removed as H2S. The hydrodesulfurization reaction network of D B T is depicted in Figure 5.11. Biphenyl (BP) Direct S-removal k\ Dibenzothiophene Cyclohexylbenzene Bicyclohexyl Tetrahydro-dibenzothiophene (THDBT) Hydrogenation S Figure 5.11 Reaction network for hydrodesulfurization of dibenzothiophene. 118 In the present study, DBT was added to hexadecane solvent to provide 900 ppm of S in the oil, within the concentration range that is normally present in gas oil. Crystalline M0S2, exfoliated MoS 2 , MoNaph and A H M derived M0S2 were tested for their catalyst activities in hydrodesulfurization of DBT. The major products of DBT conversion from all the catalysts were tetrahydro-dibenzothiophene (THDBT), biphenyl (BP), cyclohexylbenzene (CHB), and bicyclohexyl (BCH). This observation is in agreement with the literature (Daage and Chianelli, 1994; Del Valle et al., 1998; Devers et al., 2002). BP is a product of direct desulfurization of DBT, whereas THDBT is a product of DBT hydrogenation. CHB could be produced either by hydrogenolysis of THDBT or by hydrogenation of BP. B C H is a consequent hydrogenation product of CHB. Plots of product ratio (moles product/total moles of reactant plus products) as a function of time for various M0S2 catalysts at 350°C are shown in Figure 5.12 (a-d). When different M0S2 catalysts were used, different DBT conversions were observed. Crystalline MoS 2 (Figure 5.12 (a)) yielded BP, THDBT and CHB, whereas B C H was also detected in the experiment using exfoliated M0S2 (Figure 5.13 (b)). After a 5-hr batch reaction using exfoliated M0S2, the products in descending order were BP, CHB, THDBT and B C H . Figure 5.12 (c) shows that THDBT was the major product over the 5-hr batch reaction with MoNaph derived M0S2. Other products included BP and CHB. For the reaction using A H M derived M0S2, almost 100% of DBT was converted after a 5-hr batch reaction and the products detected were mainly the desulfurized products CHB, B C H and BP (Figure 5.12(d)). 119 (a) (b) 0.20 0.16 0.00 j§ 0 • THDBT _ • BP A CHB -XBCH O DBT Jm* 2 3 Time, hr 1.0 0.8 o E 0.4 E CQ Q 0.2 0.0 Figure 5.12 Profile of product ratio using (a) crystalline M0S2 and, (b) exfoliated M0S2 in hydrodesulfurization of DBT (600 ppm Mo, 350°C, initial pressure 2.8 MPa, 5-hr reaction time; trend line). 120 Time, hr gure 5.12 Profile of product ratio using (c) MoNaph and, (d) A H M derived M0S2 in hydrodesulfurization of DBT (600 ppm Mo, 350°C, initial pressure 2.8 MPa, 5-hr reaction time; trend line). 121 Table 5.7 compares the DBT conversion, hydrodesulfurization (HDS) activity and estimated rate constant for the disappearance of DBT, &DBT, on the 4 different M0S2 catalysts at 350°C and an initial pressure of 2.8 MPa. After a 5-hr batch reaction, the order of DBT conversion achieved was crystalline M0S2 (12.6%) < MoNaph derived M0S2 (15.9%) < exfoliated MoS 2 (35.3%) < A H M derived MoS 2 (99.5%). However, the order in terms of hydrodesulfurization activity (HDS, %) was MoNaph derived M0S2 (3.3%) < crystalline MoS 2 (8.9%) < exfoliated MoS 2 (25.8%) < A H M derived MoS 2 (97.6%). Hydrodesulfurization activity is defined as the total yield of desulfurized products (Appendix D-3). M0S2 catalyst derived from MoNaph provided DBT conversion comparable to crystalline M0S2. DBT conversion using exfoliated M0S2 was almost double that obtained when using MoNaph derived M0S2. Of all the catalysts examined, A H M derived catalyst gave the highest DBT conversion after 5 hour reaction. Table 5.7 Comparison of hydrodesulfurization of DBT using different M0S2 catalysts (600 ppm Mo, 350°C, initial pressure 2.8 MPa, 5-hr reaction time). Catalyst/precursor * D B T * * , ml/(g Mo- s ) x l 0 2 Conversion, % HDS, % Crystalline MoS 2 2.27 ±0.38 12.6 8.9 Exfoliated MoS 2 7.24 ± 1.32 35.3 25.8 MoNaph 2.94 ± 0.64 15.9 3.3 A H M 88.3 ± 13.5 99.5 97.6 ** Data are reported with 95% confidence limits Note that the DBT conversion for respective reaction are with ± 10 % error. 122 The DBT concentration profile was found to follow first order reaction kinetics with a correlation coefficient R 2 > 0.95 (Figure 5.13). First order was assumed and the rate constant for the disappearance of DBT, &DBT was determined from the slope of the plot of In- versus time for each catalyst as shown in Figure 5.13. [DBT] The rate constants for disappearance of DBT are reported in Table 5.7. Under similar reaction conditions, the rate constant for hydrodesulfurization of DBT in the 7 7 presence of unsupported M0S2 catalysts has been reported in the range 1.9 x 10" - 6.7 x 10" mol/(g MoS2-s) at 350°C (Del Valle et al., 1998). In order to make a comparison, a conversion factor was used. The rate constant reported was calculated to be in the range of 1.3 x 10"3 - 4.7 x 10"3 ml/(g Mo-s). The rate constants in the present study were 9 9 significantly higher in the range from 2.27 x 10" to 88.3 x 10" ml/(g Mo-s). Figure 5.14 shows the product yields as a function of DBT conversion for each of the four catalysts. For exfoliated M0S2 and crystalline M0S2, the yield of BP was greater than the yield of THDBT for all DBT conversions (Figure 5.14 (a) and (b)). In contrast, for MoNaph and A H M derived M0S2, the THDBT yield was greater than the BP yield at the same DBT conversion (below 30%)(Figure 5.14 (c) and (d)). Furthermore, at the same DBT conversion, the yield of BP decreased in the order of crystalline M0S2 > exfoliated M0S2 > MoNaph derived M0S2 - A H M derived M0S2. Figure 5.14 also shows that all catalysts exhibited a similar profile for the total yield of the final products,-i.e. CHB and B C H (increasing with increasing DBT conversion). These results demonstrate that the four M0S2 catalysts have different reaction kinetics for DBT conversion. 123 Figure 5.13 Plots of ln([DBTo]/[DBT]) versus time for hydrodesulfurization of DBT with (a) crystalline MoS 2 ; (b) exfoliated MoS 2 ; (c) MoNaph; and (d) A H M derived M0S2 (600 ppm Mo, 350°C, initial pressure 2.8 MPa, 5-hr reaction time; fitted line). 124 (a) (b) 0 5 10 15 20 0 20 40 60 80 100 Conversion of DBT, % Conversion of DBT, % Figure 5.14 Products yield versus conversion of DBT for (a) crystalline M0S2, (b) exfoliated MoS 2 , (c) MoNaph and (d) A H M derived MoS 2 (600 ppm Mo, 350°C initial pressure 2.8 MPa, 5-hr reaction time; fitted model). 125 To clarify the kinetics of the hydrodesulfurization of DBT, the lumped rate constants of the reaction steps shown in Figure 5.11 were estimated for each of the catalysts studied. Each step was assumed to follow first order kinetics. Since the amount of hydrogen in the reaction vessel was in excess, the dependency of the reaction kinetics on H2 pressure can be neglected. Note that although the first order assumption is oversimplified for hydrodesulfurization of DBT, this assumption has been widely used in comparative catalyst studies (Iwata et al., 1998; Farag et al., 2003b), as in the present work. The kinetic equations used in the simulation are shown in Eq.(5.6) to (5.10). The rate constants in the ordinary differential equations (k/s) were estimated using the Gauss-Newton optimization method (Appendix G). The estimated rate constants are summarized in Table 5.8. _ 4 D B T ] = ( ^ + ^ ) [ D B T ] ( 5 6 ) dt _4THDBT] = ^ [ T H D B T ] _ ^ [ D B T ] (5.7) dt _d\BP] = _ h [ D m ] + ^ [ B p ] ( 5 .8) dt _ JTCHB] = _ ^ [ T H D B T ] + £ 5 [CHB] - £ 4 [BP] (5.9) dt _ 4 B C H ] ( 5 1 0 ) dt 5 where [] indicates concentration of the indicated compound. 126 Table 5.8 Estimated I s order rate constants for the hydrodesulfurization of D B T (Figure 5.11). —i Rate constant MoNaph Exfoliated M 0 S 2 Crystalline A H M M o S 2 ml/(gMo.s) 325 °C 350 °C 375 °C 325 °C 350 °C 375 °C 350 °C 350°C ki 0.0043 0.0047 0.0480 0.0046 0.0386 0.0963 0.0120 0.169 ± 0 . 0 0 0 7 ± 0 . 0 0 1 1 ± 0 . 0 0 3 9 ±0.0011 ± 0 . 0 0 3 2 ± 0 . 0 0 5 2 ± 0 . 0 0 0 9 ± 0 . 0 1 2 k2 0.0052 0.0192 0.0522 0.0053 0.0302 0.0613 0.0112 0.425 ± 0 . 0 0 0 7 ± 0 . 0 0 1 8 ± 0 . 0 0 3 2 ±0.0005 ± 0 . 0 0 2 2 ± 0 . 0 0 2 7 ± 0 . 0 0 0 9 ± 0 . 0 3 1 h 0.0507 0.388 0.0484 0.268 0.714 0.1080 2.191 ± 0 . 0 0 5 4 ± 0 . 0 1 4 ±0.0027 ± 0 . 0 1 1 ± 0 . 0 1 3 ± 0 . 0 0 3 8 ± 0 . 0 6 0 £ 4 0.0401 0.066 0.0258 0.114 0.0972 0.0182 0.098 ± 0 . 0 0 5 7 ± 0 . 0 1 2 ± 0 . 0 0 2 7 ± 0 . 0 1 1 ± 0 . 0 0 8 1 ± 0 . 0 0 3 8 ± 0 . 0 2 3 ki 0.123 0.124 0.1073 0.226 ± 0 . 0 1 3 " ± 0 . 0 1 1 ± 0 . 0 0 9 0 • ± 0 . 0 3 2 ki/k2 0.82 0.24 . , 0.92 0.87 1.28 1.57 1.07 0.40 fyk_ - • 2.64 7.44 9.11 8.88 11.7 9.64.- 5.15 Table 5.8 presents all the estimated rate constants (kj) for each step of the DBT hydrodesulfurization network for the four catalysts at 350°C. For MoNaph and exfoliated M0S2 rate constants at 325°C and 375°C are also reported. Comparison of the hydrodesulfurization of DBT over the four catalysts of the present study was made at 350°C. From Table 5.8, one observes that most of the h for THDBT was a magnitude higher than k\ and k2 for DBT conversion, for all the catalysts. This may be due to the geometric asymmetry of THDBT that facilitates the interaction with the catalyst (Farag et al., 2003b). However, the selectivity trends for the 4 catalysts, especially direct desulfurization of DBT versus hydrogenation can be categorized into two groups that show similar trends. The MoNaph and A H M derived M0S2 have similar selectivity as does the exfoliated M0S2 and crystalline M0S2. Hydrodesulfurization of DBT proceeds through direct desulfurization by hydrogenolysis or through hydrogenation reactions, corresponding to the rate constants k\ and fo, respectively (Table 5.8). For M0S2 catalysts derived from both MoNaph and A H M precursors, the low values of k\/k\ show that DBT conversion was favoured by the aromatic ring hydrogenation pathway in Figure 5.11. The ratio k\/k_ for the exfoliated M0S2 (k\/k2 = 1.28) and the crystalline M0S2 (k\/k2 = 1.07) were of similar magnitude and close to unity, indicating that DBT conversion proceeds through both paths (k\ and k_) in parallel with similar selectivity towards direct desulfurization via hydrogenolysis to BP and hydrogenation to THDBT. These characteristics also explain the higher THDBT yield obtained when using MoNaph and A H M derived M0S2 and the higher BP yield obtained when using exfoliated M0S2 and crystalline M0S2 as catalyst (Figure 5.14). Note that the curves plotted in Figure 5.14 correspond to the calculated yields using the fitted kinetic parameters. 128 Exfoliated M0S2 and crystalline M0S2 had a higher kx/k_ ratio (ky/k_ = 8.88 and 9.64, respectively) than MoNaph and A H M derived M0S2 (^3/^2= 2.64 and 5.15, respectively). The ratio of ki/k_ compares the hydrogenolysis activity for THDBT conversion to CHB (£3) and the hydrogenation of DBT to THDBT (k_). The ratios k\/k_ and tyki show that exfoliated M0S2 and crystalline M0S2 had a better hydrogenolysis activity than the MoNaph and A H M derived M0S2, implying that crystalline M0S2 has more hydrogenolysis sites than poorly crystalline or amorphous M0S2. Higher hydrogenolysis activity of crystalline M0S2 compared to that of A H M derived MoS2was also reported by Iwata et al. (1998). For all the catalysts, CHB was produced mainly through hydrogenolysis (A3) rather than hydrogenation of BP ( £ 4 ) . This catalytic feature is in agreement with that reported in the literature (Daage and Chianelli, 1994; Breysse et al., 2001; Farag et al., 2003b). In general, the kinetic analysis has demonstrated that exfoliated M0S2 and crystalline M0S2 had a similar selectivity trend, in favour of direct desulfurization to BP. However, the exfoliated M0S2 had more than twice the DBT conversion than the crystalline M0S2. MoNaph and A H M derived M0S2 were more oriented to hydrogenation in hydrodesulfurization reactions of DBT. Among the four catalysts, A H M gave much larger hydrogenation and hydrogenolysis rate constants (ml/(g Mo.s)) than the other catalysts. However, when comparing the hydrodesulfurization activity of these different M0S2 catalysts at the same DBT conversion, as shown in Figure 5.15, crystalline M0S2 gave the highest sulfur removal. Crystalline M0S2 was selective to direct desulfurization, although total DBT conversion was low. This result is in agreement with Iwata et. al. (1998) who reported that well crystallized M0S2 was more favorable than armorphous M0S2 for direct desulfurization during the reaction of DBT. 129 50 Conversion of DBT, % Figure 5.15 Hydrodesulfurization versus conversion of DBT (600 ppm Mo, 350°C, initial pressure 2.8 MPa, 5-hr reaction time; fitted model). The apparent activation energy of DBT hydrodesulfurization, calculated using the Arrhenius equation, was also determined for exfoliated M0S2 and MoNaph derived M0S2. The data are shown in Figure 5.16 and Ea determined for exfoliated M0S2 and MoNaph was 186 ± 68 kJ/mol and 149 ± 27 kJ/mol, respectively. 130 (a) 1.5 1.55 1.6 1.65 1000/7", K"1 (b) c 1.55 1.6 1.65 1000/7, K"1 1.7 Figure 5.16 Plots of In £DBT versus 1000/T for hydrodesulfurization of DBT using (a) exfoliated M0S2 and (b) MoNaph derived M0S2 ( fitted line). 131 5.4 Hydrodenitrogenation Hydrodenitrogenation is required mostly to minimize catalyst poisoning in subsequent hydroprocessing of catalysts. The presence of N-compounds inhibits hydrodesulfurization and other reactions due to their preferential adsorption on catalytic sites. Hydro-denitrogenation is also required to meet N O x emission standards from combustion, in order to minimize air pollution (Satterfield, 1991; Furimsky, 2005). Nitrogen is generally considered harder to remove than sulfur or oxygen. In heavy oil, nitrogen is present predominantly in heterocyclic aromatic compounds having five-membered pyrollic or six-membered pyridinic N rings, which are difficult to denitrogenate. Heterocyclic nitrogen compounds are therefore often used in the study of hydrodenitrogenation. The heterocyclic nitrogen compounds are classified as basic and non-basic, and both types of nitrogen are highly resistant to removal. In the present study, hydrodenitrogenation of quinoline, a basic nitrogen compound, and carbazole, a non-basic nitrogen compound, were studied using MoNaph derived M0S2 and exfoliated M0S2. 5.4.1 Hydrodenitrogenation of Quinoline Quinoline is a six-membered ring heterocyclic nitrogen compound that is often used as a model compound for hydrodenitrogenation reactions. This is due to its bicyclic molecular structure, where all the reactions that take place in an industrial hydrodenitrogenation process also occur in the hydrodenitrogenation of quinoline. The reactions include C-N bond cleavage, hydrogenation of an aromatic heterocyclic ring and hydrogenation of a phenyl ring. The reaction mechanism of quinoline has been widely studied using promoted and supported catalysts such as N i - M o / A l 2 0 3 and Co-Mo/AI2O3 (Satterfield and Smith, 132 1986; Girgis and Gates, 1991; Prins et a l , 1997; Massoth and Kim, 2003). In the present study, hydrodenitrogenation of quinoline was evaluated using unsupported M0S2 dispersed catalysts. Figure 5.17 shows the major reaction paths for the hydrodenitrogenation of quinoline (Kim and Curtis, 1990; Prins et al., 1997). There are two ways to remove nitrogen from quinoline, one via 1,2,3,4 tetrahydroquinoline (THQ1) and o-propylaniline (OPA) and the other via 5,6,7,8 tetrahydroquinoline (THQ5) and decahydroquinoline (DHQ). Quinoline (QNL) 1,2,3,4 tetrahydro-quinoline (THQ1) o-propylaniline (OPA) 5,6,7,8 tetrahydro-quinoline (THQ5) Decahydroquinoline* (DHQ) o-propylcyclo-hexylamine* (OPCHA) Propylbenzene (PB) Propylcyclohexene rPCHe^ ) * product not detected Propylcyclohexane (PCH) Figure 5.17 Reaction network for hydrodenitrogenation of quinoline. 133 The performance of MoNaph derived M0S2 and exfoliated MoS 2 in hydro-denitrogenation of quinoline is summarized in Table 5.9. Conversion of quinoline and hydrodenitrogenation with MoNaph derived MoS 2 and exfoliated MoS 2 was 100%, 69% and 77%, 28%, respectively. The hydrodenitrogenation activity (HDN, %) is defined as the total yield of denitrogenated product. Table 5.9 Comparison of hydrodenitrogenation of quinoline using MoNaph derived MoS 2 and exfoliated MoS 2 (600 ppm Mo, 350°C, initial pressure 2.8 MPa, 5-hr reaction time). Catalyst £ Q N L \ rnl/(g Mo- s) xlO Conversion, % HDN, % MoNaph derived MoS 2 4.22 ± 0.22 100 68.9 Exfoliated MoS 2 3.12 ± 0.51 76.5 28.2 Rate constants for disappearance of quinoline, &QNL, data are reported with 95% confidence limits. Note that the quinoline conversion is with ± 10 % error. Figure 5.18 shows the product ratio profile of hydrodenitrogenation of quinoline using MoNaph as catalyst precursor. Complete conversion of quinoline was achieved after a 5-hour batch reaction. The major products detected in the reaction were THQ1, THQ5, OPA, propylcyclohexene (PCHe) and propylcyclohexane (PCH). Although propylbenzene (PB) was also detected, the quantity was <1 mol%. 134 0.8 0.7 * O Q N L • THQ1 X T H Q 5 • O P A A PCHe + P C H 0 2 3 4 5 Time, hr Figure 5.18 Profile of product ratio using MoNaph derived MoS2in hydrodenitrogenation of quinoline (600 ppm Mo, 350°C, initial pressure 2.8 MPa, 5-hr reaction time; trend line). At the beginning of the reaction, some THQ1 was formed when the reaction temperature reached 350°C. THQ1 then converted to other products and its quantity reduced gradually as the reaction proceeded. It was noticed that even at a 60 % conversion of quinoline (QNL) after 2 hours reaction, THQ1 was still the major product. This means that the conversion of THQ1 is a relatively slow step in the reaction. In contrast, no THQ5 was detected when the reaction system reached 350°C. THQ5 continued to form faster than it reacted during the first hour of reaction. Meanwhile, OPA increased throughout the reaction and remained the second major product at the end of the reaction period. OPA was the only N-containing product remaining after the 5-hour reaction. This implies that the conversion of OPA is also a slow step in the reaction. In other words, the 135 hydrodenitrogenation of quinoline via THQ1 and OPA were both relatively slow steps in the formation of non-N-containing products, and this observation is in agreement with the literature (Rangwala et al., 1990; Massoth and Kim, 2003). PCHe and PC H increased accordingly with time. PCHe reached a maximum after 2 hours and was then hydrogenated quickly into PCH. By that time, THQ1 dropped to a similar level as PCHe. The sudden increase in the hydrogenation of PCHe to PCH suggests that the reaction was strongly inhibited by reactant and the intermediate products especially THQ1. The catalytic reaction of quinoline with MoNaph produced PCH as the primary product with a yield of 68 %. 0.8 Time, hr Figure 5.19 Profile of product ratio using exfoliated M0S2 in hydrodenitrogenation of quinoline (600 ppm Mo, 350°C, initial pressure 2.8 MPa, 5-hr reaction time; trend line). 136 When the exfoliated M o S 2 was used (Figure 5.19), formation of OPA as well as PCHe and PCH increased slightly. It is noted that reactions with MoNaph and exfoliated M0S2 had a similar ratio of quinoline to THQ1 at time zero. However, the hydrodenitroge-nation obtained using exfoliated M0S2 was lower than that of MoNaph derived M0S2 after a batch reacion time of 5-hour. The final two major products were THQ1 and OPA. Under the reaction conditions used in the present study, all species shown in Figure 5.17 were detected in the liquid product after reaction, except DHQ and OPCHA. Nevertheless, DHQ and OPCHA are considered reactive intermediates in accordance with the literature (Satterfield and Smith, 1986; Kim and Curtis, 1990). K im and Curtis (1990) reported that PB was not hydrogenated to PCHe or P C H using MoNaph as catalyst precursor. Therefore, it was proposed that the main precursor for PCHe is OPCHA, a reactive intermediate which was produced from OPA or DHQ. Although DHQ was not detected in the reaction conditions of this study, its presence as a reactive intermediate is widely accepted in the hydrodenitrogenation network of quinoline (Satterfield and Smith, 1986; Kim and Curtis, 1990; Girgis and Gates, 1991; Prins et al., 1997; Massoth and Kim, 2003). Thus, hydrodenitrogenation of quinoline took place by two pathways. The first, through denitrogenation after both rings were fully saturated to yield PCHe and PCH. The second, through hydrogenation followed by denitrogenation without saturation of the ring to produce PB'. Under the reaction conditions of the present study, the second pathway was unlikely since PB was not a major product and there were only trace amounts of PB detected at the end of reaction. Figure 5.20 shows that using MoNaph derived M0S2, a higher hydrodenitrogenation compared to exfoliated M0S2 at the same quinoline conversion, was observed. In addition, 137 C-N cleavage and hydrogenation of PCHe to PC H over exfoliated M0S2 was found to be lower than with MoNaph derived M0S2 catalyst. Lower aromatic ring hydrogenation activity, which corresponded to lower quinoline conversion, was observed previously (Satterfield and Carter, 1981). Lower hydrodenitrogenation observed with exfoliated M0S2 was believed to be due to a lack of active sites for hydrogenation in exfoliated M0S2 compared to MoNaph. 100 f; 80 o E I 60 -I TO c cu ro ~ 40 c 0) 2 "O >. I 20 A 0 & O MoNaph derived M 0 S 2 • Exfoliated M 0 S 2 O / \ — n , 20 40 60 80 Quinoline conversion, mol % 100 Figure 5.20 Hydrodenitrogenation versus conversion of quinoline (600 ppm Mo, 350°C, initial pressure 2.8 MPa, 5-hr reaction time; trend line). Hydrodenitrogenation of quinoline is a complex reaction. The present results do not give sufficient data for the determination of rate constants in the network. However, a pseudo first order assumption was used to simulate the conversion of quinoline for comparison purposes. The rate constant for the disappearance of quinoline, £QNL was 138 presented in Table 5.9. Figure 5.21 shows that the experimental data from reaction using MoNaph fitted the pseudo 1 s t order rate law assumption fairly well. However, this is not the case for the reactions using exfoliated M0S2. (a) 2.5 (b) 2.0 Figure 5.21 Plots of ln([QNLo]/[QNL]) versus time for hydrodenitrogenation of quinoline using (a) MoNaph derived M0S2 and, (b) exfoliated M0S2 (600 ppm Mo, 350°C, initial pressure 2.8 MPa, 5-hr reaction time; fitted line). 5.4.2 Hydrodenitrogenation of Carbazole Carbazole a five-membered heterocycle, non-basic, nitrogen compound, has been used as a model reactant for hydrodenitrogenation (Nagai et al., 1998; Nagai et al., 2000; Szymanska et al., 2003). This class of nitrogen compound is non-basic because the extra pair of electrons of the nitrogen heteroatom is involved in the 7t cloud of the ring and is therefore not readily available for interacting with acids. It is not only difficult to remove N from this class of compounds (Landau, 1997), but they also poison the catalyst for hydrodesulfurization (Ho, 2003). There is relatively less information available for 139 hydrodenitrogenation of carbazole than quinoline. M0S2 catalysts tested for hydrodenitro-genation of carbazole were mostly supported and promoted. The activities of exfoliated M0S2 and the M0S2 derived from MoNaph for the hydrodenitrogenation of carbazole measured in the present work are summarized in Table 5.10. Similar to hydrodenitrogenation of quinoline, after a 5-hour reaction, the MoNaph derived M0S2 gave a higher carbazole conversion (37.7 %) and hydrodenitrogenation activity (30.4 %) compared to the exfoliated MoS 2 (32.6 % and 20.4 %, respectively). The carbazole conversion first order rate constants of both catalysts were derived from Figure 5.22. Table 5.10 Comparison of hydrodenitrogenation of carbazole using MoNaph derived M0S2 and exfoliated MoS 2 (600 ppm Mo, 350°C, initial pressure 2.8 MPa, 5-hr reaction time). Catalyst £ C B Z , ml/(g Mo.s )x l0 2 Conversion, % H D N % MoNaph derived M0S2 8.92 ±0.37 37.7 30.4 Exfoliated MoS 2 7.57 ±0.42 32.6 20.4 Rate constants, &CBZ> a r e reported with 95% confidence limits. Note that the carbazole conversion for respective reaction are with ± 10 % error. Products obtained in the hydrodenitrogenation of carbazole are divided into three groups: N-containing compounds, direct denitrogenation products from C-N hydro-genolysis and are the products from route 1, 2 and 3 in Figure 5.23, and some side reaction products. The overall reaction pathway for hydrodenitrogenation of carbazole proposed by Nagai et al. (1988) is referred to herein as shown in Figure 5.23 (Furimsky, 2005). 140 (a) (b) Time, hr Time, hr Figure 5.22 Plots of ln([CBZo]/[CBZ]) versus time for hydrodenitrogenation of carbazole using (a) MoNaph derived M0S2 and, (b) exfoliated M0S2 (600 ppm Mo, 350°C, initial pressure 2.8 MPa, 5-hr reaction time; fitted line). The hydrogenated nitrogen product, tetrahydrocarbazole (THCBZ) was detected in a significant amount using both MoNaph derived M0S2 and exfoliated M0S2 catalysts. Successively hydrogenated carbazole compounds such as hexahydrocarbazole (HHCBZ), octahydrocarbazole (OHCBZ) and perhydrocarbazole (PHCBZ) were not detected under the current operating conditions due to their high reactivity, in agreement with the work of Nagai et al. (1988). The denitrogenation products detected were cyclohexylbenzene (CHB), cyclohexyl-cyclohexene (CHCHe) and bicyclohexyl (BCH). In addition, there were some side reactions such as isomerization of B C H to hexylcyclohexane (HCH) and cyclopentyl-methyl-cyclohexane (CPMCH). Hydrocracking of one of the saturated rings in B C H resulted in the formation of H C H . These isomerization products were also reported in hydrodenitrogenation of carbazole over bulk molybdenum carbide in the presence of 50 ppm sulfur in the feed at 360°C and 6 MPa (Szymanska et al., 2003). 141 Carbazole (CBZ) Biphenyl (BP) 1,2,3,4-tetrahydrocarbazole (THCBZ) * 1,2,3,4,4a,9a-octahydro- 3-cyclohexyl-cyclohexene carbazole (OHCBZ) (CHCHe) II 3 O p -H *Perhydrocarbazole (PHCBZ) * product not detected Figure 5.23 Reaction network for hydrodenitrogenation of carbazole. cyclopentylmethyl-cyclohexane (CPMCH) 142 The distribution of products (mole ratio) from hydrodenitrogenation of carbazole using MoNaph derived MoS 2 is shown in Figure 5.24 (a). The formation of THCBZ from hydrogenation of carbazole was faster than its consumption in C-N hydrogenolysis for denitrogenation products at the start of the reaction. However, when the molar ratio of carbazole in the solvent dropped more than 20 %, the formation of denitrogenation and isomerization products increased significantly. During this period, the rate of conversion of THCBZ was more than its formation rate. After a 5-hour batch reaction with MoNaph, the major products obtained were H C H (11 mol%) and C P M C H (10 mol%), followed by THCBZ (9 mol%), B C H (6 mol%), CHB (1 mol%) and CHCHe (1 mol%). Figure 5.24(b) shows the profile of the product ratio distribution of hydrodenitro-genation of carbazole over exfoliated M0S2. THCBZ remained the major product throughout the 5-hour reaction. THCBZ formed faster than its conversion in the first one and a half hours of the reaction and increased linearly at the start of the reaction. THCBZ then remained almost constant at 13 mol% for the rest of the reaction. This may be explained by the similar rate of conversion of THDBZ to products and the rate of THCBZ formation from hydrogenation of carbazole. The products obtained after 5-hour reaction in descending order were T H D B Z (13 mol%), C P M C H (7 mol%), H C H (6 mol%), B C H (5 mol%), CHB (2 mol%) and CHCHe (1 mol%). 143 Time, hr Time, hr gure 5.24 Profile of product ratio using (a) MoNaph and (b) exfoliated M0S2 in hydrodenitrogenation of carbazole (600 ppm Mo, 350°C, initial pressure 2.8 MPa, 5-hr reaction time; trend line). 144 Note that biphenyl (BP) was not detected in the products and a similar observation was made over M02C catalyst (Szymanska et al., 2003), and nitrided and sulfided M0/AI2O3 (Nagai et al., 2000; Massoth and Kim, 2003). The mechanism in Figure 5.23 indicates that the hydrogenation of carbazole is required before C-N bond hydrogenolysis can occur regardless of the type of catalyst. Direct denitrogenation through hydrogenolysis was not observed with carbazole over the two M0S2 catalysts studied. The product yields as a function of carbazole conversion are reported in Figure 5.25 for both catalysts and the data show that both catalysts gave a similar trend in product yield versus carbazole conversion. Figure 5.26 compares the hydrodenitrogenation of carbazole over different catalysts at the same carbazole conversion at 350°C. Similar trends for exfoliated M0S2 and MoNaph derived M0S2 on carbazole conversion were found. Based on the obtained products, the reaction network of carbazole was further simplified into Figure 5.27. Employing the simplified reaction network, the kinetic constants for hydrodenitrogenation of carbazole over MoNaph and exfoliated M0S2 were estimated. The fitted rate constants were then used to simulate the reaction network. It is noted that for the sake of comparison, B C H , H C H and C P M C H were grouped into one lump sum and this group is denoted as gBCH. 145 Conversion of carbazole, mol % gure 5.25 Products yield versus conversion of carbazole for (a) exfoliated M0S2, and (b) MoNaph derived MoS 2 (600 ppm Mo, at 350°C, initial pressure 2.8 MPa, 5-hr reaction time; fitted model). 146 35 Conversion of carbazole, % Figure 5.26 Comparison of hydrodenitrogenation over converted carbazole at 350°C ( fitted model). C B Z _ THCBZ k2 CHB gBCH B C H + H C H + C P M C H CHCHe Figure 5.27 Schematic diagram for hydrodenitrogenation of carbazole used in model fitting. 147 Likewise, a pseudo first order rate law was assumed for all reactions. The kinetic equations for hydrodenitrogenation of carbazole are given in Eq.(5.11) to Eq.(5'.15). d[CBZ] = _ k [ C B Z ] ( 5 U ) dt d[THCBZ] = _ ^ + ^ + k ^ [ m C B Z ] + k[[CBZ] (5.12) dt d ^ C H B ^ = k3 [THDBT] (5.13) dt d[CHCHe] = K [ m D B T ] ( 5 1 4 ) dt d[gB^m=k2[THDBT} (5.15) The simulated results are summarized in Table 5.11. MoNaph derived M0S2 gave higher rate constants in most of the reaction steps compared to the exfoliated M0S2. Table 5.11 Fitted rate constant for hydrodenitrogenation of carbazole using MoNaph derived M o S 2 and exfoliated MoS 2 at 350°C. Rate constant, ml/g Mo.s MoNaph derived M0S2 Exfoliated MoS 2 ki 0.091 ±0.003 0.078 ± 0.003 k2 0.460 ±0.018 0.340 ±0.017 k3 0.027 ±0.013 0.044 ±0.013 k4 0.023 ±0.016 0.017 ±0.015 k_/k\ 5.05 4.36 MoNaph derived M0S2 gave slightly better reactant conversions and hydrode-nitrogenation in both hydrodenitrogenation of quinoline and carbozole compared to exfoliated M0S2. It was found that some observations made during the hydro-148 denitrogenation of quinoline and carbazole were similar. N removal occurs only after the hydrogenation of at least one ring of the N-compound. Direct denitrogenation through hydrogenolysis, as occured in hydrodesulfurization of DBT, was not observed in carbazole. Subsequently, the mechanism of C-N bond cleavage was also different from hydrogenolysis of C-S bond in hydrodesulfurization of DBT. A similar ratio for ki/k\ which corresponds to the rate constant of the major product, gBCH, to that of the hydrogenation of carbazole was obtained for both catalysts with different structure. This suggests that C-N hydrogenolysis in quinoline and carbazole are not as sensitive to M0S2 structure as C-S hydrogenolysis in DBT. 5.5 Hydrodeoxygenation of phenol The importance of hydrodeoxygenation which occurs during hydroprocessing depends on the origin of the feed. Hydrodeoxygenation receives little attention in conventional crude processing as the oxygen content is typically less than 2 wt%. Further, during hydrodeoxygenation, oxygen in the feed is converted into H2O, which is environmentally benign. Hydrodeoxygenation is important in the case of synthetic crude derived from coal, oil shale and especially biomass with oxygen content more than 10 wt%. Oxygenated compounds have been connected to the instability of fuel which may lead to poor performance during fuel combustion (Furimsky, 2000). Phenolic and furanic oxygen are highly refractory oxygenated compounds identified in petroleum and coal derived oil. In the present study, phenol was used as a model oxygen compound for the hydro-deoxygenation reaction. 149 The reaction network of hydrodeoxygenation of phenol is shown in Figure 5.28. Hydrodeoxygenation of phenol was found to proceed via two parallel pathways (Furimsky, 2000). Phenol can be deoxygenated to benzene through direct hydrogenolysis of the C-0 bond or to cyclohexane or cyclohexane with combined hydrogenation of the aromatic ring before hydrogenolysis of the C-0 bond. Note that the conversion of phenol is equivalent to the hydrodeoxygenation activity. Four catalysts including exfoliated M0S2, crystalline M0S2, MoNaph and A H M derived M0S2 were used in the study of hydrodeoxygenation of phenol. The main product of reaction was benzene regardless of which catalyst was used, suggesting that the major path for hydrodeoxygenation of phenol was through the direct hydrogenolysis of the C-0 bond. In the present study, cyclohexane and cyclohexene were negligible, with the amount detected less than 1 mol%. OH Direct hydrogenolysis route Phenol Benzene Combined hydrogenation-hydrogenolysis route + Cyclohexane Cyclohexene Figure 5.28 Hydrodeoxygenation routes of phenol. Table 5.12 summarizes the results from the four catalysts studied. The phenol conversion, in descending order, was crystalline MoS 2 (5 %) < MoNaph (11%) < exfoliated 150 M0S2 (15 %) < A H M (19 %). The lowest conversion was obtained over crystalline M0S2 and the highest hydrodeoxygenation was obtained in the reaction over A H M derived M0S2 as in the hydrodesulfurization reaction of DBT. Figure 5.29 shows that the concentration profiles drop linearly with reaction time. A zero order rate law was therefore used to calculate the rate constant for the phenol conversion. The rate constant is summarized in Table 5.12. Not many studies have been reported on phenol hydrodeoxygenation. A rate constant reported for the reaction of phenol to benzene in hexadecane over Ni-Mo/AbCb at 275°C was 1.09 x 10"6 mol/(g (Mo+Ni).s) (Chon and Allen, 1991). Table 5.12 Comparison for hydrodeoxygenation of phenol using different M0S2 catalysts (600 ppm Mo, 350°C, initial pressure 2.8 MPa, 5-hr reaction time). Catalyst k, mol/(g Mo-s) xlO Conversion , % Crystalline M o S 2 0.36 ±0.01 fTo Exfoliated MoS 2 1.09 ± 0.08 15.0 MoNaph derived MoS 2 0.82 ± 0.02 11.1 A H M derived M o S 2 1.35 ±0.13 19.3 Data are reported with 95% confidence limits The phenol conversion for respective reaction are estimated with ± 10 % error. Compared to the other hydroprocessing reactions studied in previous sections, hydrodeoxygenation of phenol is a simple, straightforward reaction that involves hydrogenolysis of the C-0 bond. It is suggested that C-0 hydrogenolysis in phenol occurred at the same sites as the C-S hydrogenolysis in hydrodesulfurization of DBT. The order of phenol conversion with catalyst was also found similar to that of DBT conversion, i.e. crystalline M0S2 < MoNaph derived MoS 2 < exfoliated MoS 2 < A H M derived MoS 2 . 151 Figure 5.29 Plots of concentration profile for hydrodeoxygenation of phenol using (a) crystalline MoS 2 , (b) exfoliated MoS 2 , (c) MoNaph and (d) A H M derived M o S 2 (600 ppm Mo, 350°C, initial pressure 2.8 MPa, 5-hr reaction time; fitted line). 152 5.6 Structure-activity Relationship Over the years, efforts have been made to identify the type of active sites, especially for C-S hydrogenolysis and hydrogenation, and their relation to the structure of the M0S2 crystallite. It is well accepted that most of the reactions take place at the M0S2 edge surface (Roxlo et al., 1986). However, several issues remain ambiguous, such as how many types of active sites are available, where these active sites are located and how the reactant molecules are adsorbed on the catalyst surface. Studies of M0S2 structure-activity relationships have been done on supported M0S2, promoted and supported M0S2 as well as unsupported M0S2 catalysts (Topsoe and Clausen, 1986; Hensen et al., 2001). The purpose of using unsupported M0S2 is to avoid the interference and complexity of support or promoter effects (Daage and Chianelli, 1994; Iwata et al., 1998; Iwata et al., 2001; Farag et al., 2003b). These studies correlated the hydrogenation and hydrogenolysis activities with M0S2 structure in the hydrodesulfurization of DBT. In the present study, unsupported M0S2 derived from different sources with different structures has been investigated. Hydrogenation and hydrogenolysis activities have been related to structure in hydrodesulfurization of DBT. The structure-activity relationship was also examined for hydrodenitrogenation of carbazole and hydrodeoxygenation of phenol. Four M0S2 catalysts were used in the present study. From various characterization data, it was concluded that the M0S2 catalysts exhibited different structure, as discussed in Section 5.1 and shown again in Table 5.13. Furthermore, the four catalysts were shown to have different activity toward each of the hydroprocessing reactions examined and the catalyst BET surface area does not account for the differences in activity. As previously 153 mentioned, active sites of M0S2 are on the edge surface, the catalysts with different particle size should have a different number of edge sites that are associated with active sites. Although each of the Mo atoms at the edge surface may not be catalytically active, the total edge sites should be directly proportional to the total active sites. With the known properties of M0S2 catalysts, the edge site density of each catalyst can be estimated. Table 5.13 Summary of M0S2 catalyst properties Crystalline Exfoliated MoNaph A H M M0S2 M0S2 derived M0S2 derived M0S2 Average Slab length, nm Stack heights, nm Number of layers 560 ±740 35.8 57 400 ± 460 5.5 9 10 ± 10 1.8 (0.8*) 2.8(1.3*) 9.3 ±9 .0 2.1(1.1*) 3.3 (1.8*) * Analyzed by TEM. The calculation of the edge sites was based on the method given by Kasztelan et al. (1984). The assumptions used for the calculation are as the follows: (1) The M0S2 slabs were assumed to be in the shape of a perfect hexagon i.e. each side of the hexagon has the same number of Mo atoms as shown in Figure 5.30. (2) A l l the slabs of the catalyst have the same diagonal length. (3) A l l Mo atoms at the edge sites are considered to be the active sites. (4) Catalyst activity is regarded as uniform at the M0S2 edge surface, which is an approximation in view of the difference in activity for corner and edge sites. (5) The basal planes of M0S2 slabs are not catalytically active. 154 The parameter needed for the calculation is the slab length of the M0S2 particle. The number of edge side Mo atoms, n of a single slab was estimated using the formula L = 3.2 (2n-l) [A] where L denotes the average slab length (Kasztelan et al., 1984). Having obtained n, the total, edge and basal Mo atoms can be calculated. The detailed calculations of M0S2 dispersion can be found in Appendix F-3. The M0S2 dispersion was defined as the ratio of Mo at the edge surface to the total number of Mo atoms used. The calculated dispersion of the different M0S2 catalysts of the present study are presented in Table 5.14 Figure 5.30 M0S2 single layer slab model used in edge site calculation, n = number of edge The difference in activity of the different M0S2 catalysts was thought to be due to the differences in Mo dispersion, corresponding to the different number of catalyst active sites. For example, the higher activity per gram Mo of the exfoliated M0S2 compared to crystalline M0S2 in hydrodesulfurization of DBT was due to a higher number of active site per gram of exfoliated M0S2 than that in crystalline M0S2. The number of active sites of the exfoliated M0S2 was expected to increase through exfoliation. The estimated rate constants were corrected for the amount of Mo at the edge surface of the M0S2 crystallites using the n side Mo; L= diagonal length of the hexagon. 155 Mo dispersion given in Table 5.14. The corrected reaction rate constants and the initial turnover frequency (TOF), determined for the hydrodesulfurization of DBT over the different catalysts, are summarized in Table 5.15. The TOF is defined as the number of molecules reacting per active site per unit time and the initial TOF was calculated from the estimated rate constant and the initial concentration of reactant used in the reactor. For a non-structure sensitive reaction, the catalyst should give the same TOF at the same operating condition, independent of the catalyst structure. Table 5.14 Dispersion of M0S2 catalysts. Catalyst • Dispersion, /MO, % Crystalline MoS 2 023 Exfoliated MoS 2 0.32 MoNaph derived MoS 2 9.72 A H M derived MoS 2 10.71 Roxlo et al. (1986) reported a TOF for DBT conversion of approximately 5 x 1 0 " DBT molecules/(molecule Mo.s) on both amorphous and crystalline MoS 2 at 350°C. In the present study the TOFs for hydrodesulfurization of DBT over 4 different MoS 2 catalysts are of similar magnitude, except for the MoNaph derived MoS 2 catalyst (Table 5.15). 156 Table 5.15 Rate constant and initial turn over frequency (TOF) for hydrodesulfurization of DBT over different MoS 2 catalysts at 350°C. Rate constant (ml/mol Mo e.s) xlO" 2 * Crystalline MoS 2 Exfoliated MoS 2 MoNaph derived MoS 2 A H M derived M0S2 k-DBT 9.51 21.7 0.29 7.91 h 5.02 11.5 0.05 1.51 k2 4.69 9.04 0.19 3.81 h 45.2 80.3 0.50 19.6 7.62 34.3 0.40 0.88 £5 - 37.2 - 2.03 ki/k2 1.07 1.28 0.24 0.40 h/k_ 9.64 8.88 2.64 5.15 Initial TOF** 1.98 4.36 0.06 1.65 Moe refers to the number of Mo atoms available at the edge sites. Unit of TOF is molecule DBT/(Moe.s) x 102 The TOF of the MoNaph derived catalyst was about two orders of magnitude lower than that observed on the other catalysts. Previous studies have shown that M0S2 derived from MoNaph has a significant amount of carbonaceous material associated with the M0S2 (Bearden and Aldrich, 1981; Rueda et al., 1997). In one study, 11 wt% of carbon was reportedly present in the MoNaph derived M0S2 catalyst (Rueda et al., 2001). In the present study, large amounts of solid were recovered after reaction using MoNaph, compared to the other catalysts with the same Mo feed concentration, suggesting the presence of carbonaceous material. Most likely the low activity (or TOF) of this catalyst was a 157 consequence of limited access to the catalyst active sites because of the carbon associated with the M0S2. To validate this assertion, M0S2 prepared from A H M that generated small M0S2 particles of similar size to that obtained from MoNaph, but without the carbonaceous material, was tested. Consistently, the A H M derived M0S2 exhibited much higher activity than the MoNaph derived M0S2 in hydrodesulfurization of DBT (95.8 %) and gave TOF comparable to the crystalline. M0S2 and exfoliated M0S2 (Table 5.15). Similar to the hydrodesulfurization reaction, A H M derived M0S2 exhibited higher activity than MoNaph derived MoS 2 in hydrodeoxygenation of phenol. The lower magnitude TOF of MoNaph derived M0S2 compared to exfoliated M0S2 was also observed in the hydrodenitrogenation of carbazole, in which hydrogenation is a requirement prior to the cleavage of C-N bonds. The corrected rate constant and initial TOF for hydrodenitrogenation of carbazole are reported in Table 5.16. Table 5.16 Rate constant and initial turn over frequency (TOF) for hydrodenitrogenation of carbazole over different MoS 2 catalysts at 350°C. Rate constant (ml/mol Mo e.s) x 10"3 * Initial TOF** Catalyst : '. . : : &CBZ k\ k2 h k4 xlO Exfoliated MoS 2 227 235 102 1.31 (L52 Z51 MoNaph derived MoS 2 0.09 0.09 0.45 0.03 0.02 0.10 * Moe refer to the number of Mo atoms available at the edge sites. ** Unit of TOF is molecule carbazole/(Moe.s) For hydrodeoxygenation of phenol, a different pattern of TOF for phenol over different M0S2 catalysts was observed (Table 5.17). The phenol TOF over MoNaph and A H M derived M0S2 were similar, ~ l x l 0 " 2 phenol molecule/(sites.s), whereas on crystalline 158 M0S2 and exfoliated M0S2, the TOF was an order magnitude higher than over MoNaph and A H M derived M0S2. Recalling that hydrodeoxygenation of phenol involved mainly hydrogenolysis of C-0 bond, the phenol TOF represents the comparison of C-0 bond cleavage via hydrogenolysis for different catalysts. The TOF data of phenol imply that crystalline M0S2 and exfoliated M0S2 have a higher number of active sites for hydrogenolysis in C-0 bond cleavage. This observation is consistent with the rate constant k\ and £3 determined for the hydrodesulfurization of DBT (Table 5.15), corresponding to the hydrogenolysis reaction steps. The data suggest that crystalline M0S2 has more hydrogenolysis sites than amorphous M0S2. Higher hydrogenolysis product in hydrodesulfurization of DBT with crystalline M0S2 compared to that of A H M derived M0S2, under the same conversion, was also reported by Iwata et. al. (1998). Table 5.17 Turn over frequency for hydrodeoxygenation of phenol over different M0S2 catalysts at 350°C. Catalyst TOFPhenoi,rnolecule/(Moe.s) x 10 Crystalline M0S2 15.2 Exfoliated MoS 2 32.6 MoNaph derived M0S2 0.9 A H M derived MoS 2 1.3 From the catalyst structural point of view, there are two important differences between the crystalline M0S2 and exfoliated M0S2, and the MoNaph and A H M derived 159 M0S2. First, the two former catalysts have larger slab lengths, corresponding to a large basal plane, compared to the latter two catalysts. Secondly, the former catalysts have a higher degree of stacking of M0S2 layers in the crystallites. Since the basal plane is widely accepted to be inert with few active sites, the difference in selectivity among these catalysts is likely due to the different number of M0S2 layers in the catalyst particle. In the literature, various models have been proposed to account for the effect of structure on active sites by distinguishing between edge, corner and rim sites of the M0S2 crystallite (Daage and Chianelli, 1994). Knowing the crystallite dimensions (Table 5.13) the fraction of rim or edge sites can be estimated using the geometric arguments of Kasztelan (1984). Figure 5.31 plots the k\lki ratio and the kj/ki ratio as a function of the edge or rim sites, for the 4 catalysts studied herein. Both ratios show an increasing trend with an increase in the fraction of edge sites (or alternatively a decreasing trend with increasing fraction of rim sites). Noting that both rate constant ratios (&1/&2 and tyki) are a measure of the rate of hydrogenolysis to the rate of hydrogenation occurring on the catalysts, it can be concluded, in agreement with the model of Daage and Chianelli (1994) that hydrogenation is favored on rim sites. Also note that the correlation presented here shows a trend over a wide range of catalyst morphologies, in which both stack height and slab length have been varied. In the rim-edge model (Daage and Chianelli, 1994), a straightforward relation between the M0S2 layers and selectivity was reported. The M0S2 catalysts with different structure used in the current study were prepared through very different methods and the average size of different M0S2 particles size was varied over a wide range from 10 to 600 nm. Further, reactions were carried out in the presence of H2S, which is reported to adversely affect hydrodesulfurization (Farag et a l , 2003a; Farag et al., 2003c). 160 12 10 8 T J C CD =s 6 M— o o 4 •= to a. O kyik2 • k^k2 G O 0.2 0.4 Relative rim sites 0.6 0.8 c CO JC o g CO rr 0.2 0.4 0.6 0.8 Relative edge sites 1.2 Figure 5.31 Ratio of k\lk2 and k^lki versus relative rim sites and edge sites for hydro-desulfurization of DBT ( trend line). 161 Compared to crystalline M0S2, the rate constant for the conversion of DBT, ADBT and conversion of phenol were slightly higher over exfoliated M0S2. A possible reason could be the presence of some active sites on the basal planes after exfoliation in addition to the edge planes sites. Activity of the basal plane in hydrodesulfurization of thiophene over M0S2/AI2O3 prepared by exfoliation method has been reported previously (Kochubei et al., 2003; Kochubey et al., 2004). The structure of M0S2 was found distorted after exfoliation (Gordon et al., 2002). The catalyst prepared via exfoliation with 10 times larger particle diameter compared to the standard catalyst was found to have a close or even higher activity (Kochubey and Babenko, 2002). It was suggested that there was molecular adsorption on the basal plane of exfoliated M0S2 especially for hydrogenolysis. Nonetheless, the activity on the basal planes is lower than that of the edge site. 5.7 Activity Comparison between MoNaph Derived M0S2 and Exfoliated M0S2 The performance of the MoNaph derived M0S2 and exfoliated M0S2 in hydrogenation, hydrodesulfurization, hydrodenitrogenation and hydrodeoxygenation reactions using model reactants is summarized in Table.5.18. The results show that the M0S2 derived from MoNaph had a slightly higher activity for hydrogenation reactions compared to the exfoliated M0S2. The naphthalene conversion was higher on the MoNaph derived catalyst than the exfoliated M0S2 and the same trend was observed in the hydrodenitrogenation of quinoline and carbazole. In the hydrodenitrogenation case, C-N cleavage only occurred after hydrogenation of at least one aromatic ring. Since the hydrogenation reaction was favored on the MoNaph derived M0S2, conversion in hydrodenitrogenation was higher on MoNaph derived M0S2 compared to the exfoliated M0S2. 162 Table 5.18 Comparison of reactant conversion and heteroatom removal in different model reactions (600 ppm Mo, 350°C, initial pressure 2.8 MPa, 5-hr reaction time). MoNaph Exfoliated MoS 2 Reaction Reactant Conversion, S , N , 0 Conversion, S , N , 0 % removal, % % removal, % H Y D Naphthalene 35.8 - 27.0 -HDS Dibenzothiophene 15.9 3.3 35.3 25.8 H D N Quinoline 100 68.9 76.5 28.2 H D N Carbazole 37.7 30.4 32.6 20.4 HDO Phenol 11.1 11.1 15.0 15.0 Note that the reactant conversions for respective reaction are with ± 10 % error. The exfoliated M0S2, however, had a higher activity in hydrodesulfurization of DBT and hydrodeoxygenation of phenol compared to that of the M0S2 derived from MoNaph. DBT conversion occurs through two parallel reaction paths: direct desulfurization •or hydrogenation of the aromartic ring followed by C-S hydrogenolysis. Since the exfoliated M0S2 had a high activity for the direct desulfurization of DBT to yield BP, compared to the MoNaph derived M0S2 that yielded THDBT, the former catalyst had overall higher hydrodesulfurization activity. Hydrodenitrogenation of quinoline requires ring hydrogenation before N removal. Although carbazole has the same skeletal structure as DBT, nitrogen removal only occurs after at least one of the adjacent rings has been hydrogenated. For HDO of phenol, oxygen removal from phenol proceeded through direct hydrogenolysis of C-0 without ring saturation. Higher phenol conversion was obtained using exfoliated M0S2 than MoNaph derived M0S2. These results showed that exfoliated M0S2 has better hydrogenolysis activity than MoNaph derived M0S2. As a result, the better hydrogenation activity in MoNaph derived M0S2 clearly explained the trend of results 163 shown in Table 5.18. Alternatively, when ring saturation is not a requirement for carbon heteroatom cleavage, comparable or even better performance was obtained by using exfoliated M0S2. The comparison of the similar catalysts for Cold Lake bitumen hydrocracking at 415°C and 600 ppm Mo has been reported in Chapter 4. In the comparison, the exfoliated M0S2, dispersed in water rather than decalin, had a higher removal of S, N , M C R and asphaltenes, than the M0S2 derived from MoNaph. The effect of dispersing solvent on the exfoliated M0S2 in Cold Lake bitumen hydrocracking was also examined at a Mo concentration of 900 ppm Mo. A higher S, N , metal and asphaltene removal was obtained when the exfoliated M0S2 was dispersed in decalin rather than water. Consequently, it can be concluded that exfoliated M0S2 dispersed in decalin would give better heteroatom removal in Cold Lake bitumen hydrocracking at 415°C and 600 ppm Mo, than MoNaph derived M0S2. In hydrocracking, the catalyst performance is associated with the ability of the catalyst to transfer hydrogen to prevent condensation among carbon radicals. The performance of the exfoliated M0S2 is associated with its morphology. In the case of single model compound reactants, the differences in catalyst performance follows previous suggestions that the hydrogenation occurs preferentially on rim-edge sites. Since the MoNaph derived M0S2 morphology yields a greater fraction of rim-edge sites than the exfoliated M0S2, the former catalyst has higher hydrodenitrogenation and hydrogenation activity. Hydrogenolysis of C-S occurs preferentially on edge sites and since the exfoliated M0S2 has a greater fraction of edge sites, this catalyst has higher hydrodesulfurization activity than that of the MoNaph derived M0S2. However, the differences in the catalyst performance during hydrocracking is more complex. In this case, thermal reactions 164 dominate and hydrogen transfer to cap free radicals becomes important. The results from the present study show that the exfoliated M0S2 is better suited for this Cold Lake hydrocracking probably because of a combination of hydrogenolysis and hydrogen transfer capability. Although the exfoliated M0S2 had lower hydrodenitrogenation activity than the M0S2 derived from MoNaph in the case of carbazole as the reactant, at higher temperature with Cold Lake bitumen, N removal occurred to a higher degree on the exfoliated M0S2. 5.8 Summary The catalyst activity of exfoliated M0S2 (in decalin) and MoNaph derived M0S2 were compared in hydrogenation of naphthalene, hydrodesulfurization of DBT, hydro-denitrogenation of quinoline and carbazole, and hydrodeoxygenation of phenol. The results provided a comprehensive quantitative comparison of hydroprocessing reactions between exfoliated M0S2 and MoNaph. In general, exfoliated M0S2 was found to favour the hydrogenolysis pathway during reaction while MoNaph derived M0S2 favoured the hydrogenation pathway. In addition, the activity of another two catalysts: crystalline M0S2 and A H M derived M0S2, which exhibit different M0S2 structure, were tested for the study of M0S2 structure-activity relations. The structure-activity relation study was performed using the geometrical model for active site of M0S2 proposed by Kasztelan et al. (1984). The selectivity of hydrogenation and hydrogenolysis in hydrodesulfurization DBT was found to be related to the rim and edge sites, in agreement to the rim-edge model for M0S2 particles. 165 Conclusions and Recommendations 6.1 Conclusions Exfoliated M0S2 has been studied as a dispersed catalyst for Cold Lake bitumen hydrocracking in a slurry reactor. At 415°C, exfoliated M0S2 was found to be effective in hydrogen transfer during hydrocracking and had similar coke yield as obtained with MoNaph derived M0S2. Exfoliated M0S2 also provided better liquid product quality in terms of hydrodesulfurization, hydrodenitrogenation, asphaltene conversion, M C R removal and metal removal, compared to that of M0S2 derived from MoNaph. Within the range of temperatures studied (400-430°C), the optimum temperature to minimize coke yield was found to be 415°C. There was also an optimum Mo concentration of 600 ppm Mo for the lowest coke yield and the best liquid product quality. The liquid yield was not significantly affected by the M0S2 concentration, in agreement with the fact that unsupported catalysts do not contain acidic functions and the carbon-carbon bond cleavage reactions are thermally controlled. Exfoliated M0S2 was also redispersed in an organic solvent, decalin. The slight increase in coke yield in decalin, compared to that in water, was mainly due to an increase in M0S2 stack height. In addition, the presence of water was identified as a positive contributing factor in M C R removal in heavy oil hydrocracking. New experimental work has also been done to explore the possibility of recycling the dispersed catalyst. After hydroprocessing for one hour in a batch reactor, more than 166 90% of the Mo catalyst resided in the coke. The hydrocracking performance of this recovered catalyst-coke was found to be satisfactory, suggesting that recycling the catalyst-coke in the slurry process is possible, making the slurry process more economically viable. Hydrodesulfurization of DBT was examined over a series of M0S2 catalysts each with different morphology. It was found that a change in the morphology of M0S2 affects the product selectivity in hydrodesulfurization of DBT. The activity of the catalysts toward hydrogenolysis of C-S bonds versus hydrogenation of DBT and THDBT is shown to correlate with the fraction of edge sites present in the crystallites. Selectivity to hydrogenolysis increased as the fraction of edge sites increased. This is in agreement with the rim-edge model of Daage and Chianelli (1994). The catalyst activity of exfoliated M0S2 and MoNaph derived M0S2 were compared for the hydrogenation of naphthalene, hydrodesulfurization of DBT, hydrodenitrogenation of quinoline and carbazole, and hydrodeoxygenation of phenol. The results provide a quantitative comparison between exfoliated M0S2 and M0S2 prepared from MoNaph. Exfoliated M0S2 was found to give a better overall hydrodesulfurization and hydro-deoxygenation compared to MoNaph derived M0S2 catalyst, whereas MoNaph derived M0S2 was found to give higher hydrogenation and hydrodenitrogenation activity. These results are thought to be a consequence of the morphology of the two catalysts. Hydrogenation is favoured on M0S2 catalysts with more rim sites. Results from the model compound studies show that a simple change in morphology: M0S2 crystallite size and stack height appears to have a profound influence on the reactions, especially in selectivity of DBT hydrodesulfurization, regardless of the catalyst preparation method. The major contribution of the present study is that a better insight and understanding on the use of exfoliated M0S2 as a heavy oil upgrading catalyst is provided. 167 The study correlates the catalyst activity and structural properties of unsupported M0S2 catalysts prepared by different methods over a wide range of morphologies. Primarily, selectivity to hydrogenation and hydrogenolysis in the DBT reaction was found to correlate with the quantity of M0S2 rim sites and edge sites, respectively. 6.2 Recommendations Some future research directions and recommendations for improved catalyst design in hydrocracking and hydroprocessing are outlined below. • Exfoliation starting from smaller size of M0S2 The exfoliated M0S2 in the present study was prepared from commercial crystalline M0S2, which is of much larger size (~1 jam) compared to M0S2 derived from precursors such as MoNaph or A H M (-0.01 um). It is recommended to exfoliate M0S2 of different particle sizes and compare their activity to M0S2 derived from oil soluble precursors. • Co and N i promoter in exfoliated M0S2 The exfoliated M0S2 used in the present study was unpromoted. It is recommended to study the activity of exfoliated M0S2 promoted by Co or N i . Success in promoting the exfoliated dispersed M0S2 by Co or N i could lead to higher catalyst activity. • Catalyst recycle study The potential of recycling the dispersed unsupported exfoliated M0S2 has been illustrated. It would be beneficial to the development of the slurry reactor to investigate how to increase the activity of the M0S2 catalyst within the coke. • Active sites for hydrodenitrogenation Higher hydrodenitrogenation was obtained over MoNaph derived M0S2 compared to exfoliated M0S2. Delmon (1995) suggested that there is a different type of active site on 168 M0S2 catalyst for hydrodenitrogenation, other than hydrogenation and hydrogenolysis proposed for hydrodesulfurization. Cleavage of C-N from saturated N compounds can occur via P elimination or nucleophilic substitution. A further study on the effect of catalyst characteristics on this type of active site might help to provide an insight into hydrodenitrogenation reactions. • Effect of H2S in reactions H2S is present in hydrocracking and hydroprocessing reactions in commercial plants. Although the presence of H2S was reported to have different effects on hydro-desulfurization (Farag et al., 2003a; Farag et al., 2003c), hydrodenitrogenation (Satterfield et al., 1985; Callant et al., 1995) and hydrodeoxygenation (Viljava et al., 2000) reactions, the effect of H2S towards these reactions remains unclear. • Reactions in mixed reactants In the present study, each model compound reaction was carried out with an individual model reactant. The catalyst activity should also be tested using mixed reactants. The presence of different molecules may inhibit other reactions and thereby affect the overall reactivity. Reactions carried out with mixed reactants will more closely simulate the environment in heavy oil hydroprocessing. 169 Reference Alberta Energy Utility Board, Alberta's Reserves 2002 and Supply/Demand Outlook 2003-2012, Statistical Series 2003-98. (2002) Alonso G., Del Valle M . , Cruz J., Licea-Claverie A . , Petranovskii V . , and Fuentes S. Preparation of M0S2 and WS2 catalysts by in situ decomposition of ammonium thiosalts. Catalysis Letters, 52 (1-2), 55-61 (1998) American Metal Market, Free market price guide. 114 (8-3), 11 (2006). Andersen S.I. and Speight J.G. Thermodynamic models for asphaltene solubility and precipitation. Journal of Petroleum Science & Engineering, 22 (1-3), 53-66 (1999) Annual Energy Outlook 2006, Department of Energy, United States (2006) Bearden R. and Aldrich C L . Novel Catalyst and Process to Upgrade Heavy Oils. Energy Progress, 1 (1-4), 44-48 (1981) Bockrath B.C. and Parfitt D.S. Application of exfoliation techniques to the preparation of M0S2 liquefaction catalysts. Coal Science and Technology, 24 (2), 1343-1346 (1995) Boone W.P. and Ekerdt J.G. Hydrodesulfurization studies with a single-layer molybdenum disulfide catalyst. Journal of Catalysis, 193 (1), 96-102 (2000) Breysse M . , Afanasiev P., Geantet C , and Vrinat M . Overview of support effects in hydrotreating catalysts. Catalysis Today, 86 (1-4), 5-16 (2003) Breysse M . , Berhault G., Kasztelan S., Lacroix M . , Mauge F., and Perot G. New aspects of catalytic functions on sulfide catalysts. Catalysis Today, 66 (1), 15-22 (2001) Buker K. , Berndt H. , Lucke B., and Kotowski W. Highly dispersed metal sulfide catalysts for the hydroconversion of vacuum distillation residues. Hydrotreatment and Hydrocracking of Oil Fractions, 106, 523-527 (1997) Calais C , Matsubayashi N . , Geantet C , Yoshimura Y. , Shimada H. , Nishijima A. , Lacroix M . , and Breysse M . Crystallite size determination of highly dispersed unsupported M0S2 catalysts. Journal of Catalysis, 174 (2), 130-141 (1998) Callant M . , Holder K . A . , Grange P., and Delmon B. Effect of the H 2 S and H-2 Partial Pressures on the Hydrodenitrogenation (HDN) of Aniline and Indole Over A Nimop-Gamma-Al203 Catalyst. Bulletin des Societes Chimiques Beiges, 104 (4-5), 245-251 (1995) Callejas M . A . and Martinez M.T. Hydroprocessing of a Maya residue. 1. Intrinsic kinetics of asphaltene removal reactions. Energy & Fuels, 14 (6), 1304-1308 (2000) 170 Chianelli R.R., Berhault G., Raybaud P., Kasztelan S., Hafner J., and Toulhoat H. Periodic trends in hydrodesulfurization: in support of the Sabatier principle. Applied Catalysis A-General, 227 (1-2), 83-96 (2002) Chon S. and Allen D.T. Catalytic Hydroprocessing of Chlorophenols. AIChE Journal, 37 (11), 1730-1732(1991) Chung K . H . and X u C M . Narrow-cut characterization reveals resid process chemistry. Fuel, 80 (8), 1165-1177(2001) Curtis C W . and Pellegrino J.L. Activity and Selectivity of Three Molybdenum Catalysts for Coal Liquefaction Reactions. Energy & Fuels, 3, 160-168 (1989) Curtis, M . D. Hydrodesulfurization Catalysts and Catalyst Models Based on Mo-Co-S Clusters and Exfoliated MoS 2 , in "Transition Metal Sulfur Chemistry", (Steifel, E. I.; Matsumoto, K. , Eds.), ACS Symposium Series 653, American Chemical Society, Washington, D. C , 154-175 (1996). Daage M . and Chianelli R.R. Structure-Function relations in Molybdenum Sulphide Catalysts: The "Rim-Edge" Model. Journal of Catalysis, 149, 414-427 (1994) Dautzenberg F . M . and De Deken J.C. Reactor Developments in Hydrotreating and Conversion of Residues. Catalysis Review Science Engineering, 26 (3 & 4), 421-444 (1984) Del Bianco A. , Panariti N . , Dicarlo S., Beltrame P.L., and Carniti P. New Developments in Deep Hydroconversion of Heavy Oil Residues with Dispersed Catalysts 2. Kinetic Aspects of Reaction. Energy & Fuels, 8 (3), 593-597 (1994) Del Bianco A. , Panariti N . , Dicarlo S., Elmouchnino J., Fixari B. , and Leperchec P. Thermocatalytic Hydroconversion of Heavy Petroleum Cuts with Dispersed Catalyst. Applied Catalysis A-General, 94 (1), 1-16 (1993) Del Bianco A. , Panariti N . , and Marchionna M . Upgrading Heavy Oil Using Slurry Processes. Chemtech, 25 (11), 35-43 (1995) Del Valle M , Avalos-Borja M , Cruz J, and Fuentes S . Exfoliation of M0S2 catalysts: structural and catalytic changes. 351 [Materials Research Society Symposium Proceedings (Molecularly Designed Ultrafme/Nanostructured Materials)], 287-292. 1994. Del Valle M . , Cruz-Reyes J., Avalos-Borja M . , and Fuentes S. Hydrodesulfurization activity of M0S2 catalysts modified by chemical exfoliation. Catalysis Letters, 54 (1-2), 59-63 (1998) Delmon B. Selectivity in HDS, HDN, HDO and Hydrocracking Contribution of Remote-Control and Other New Concepts. Bulletin des Societes Chimiques Beiges, 104 (4-5), 173-187 (1995) Devers E., Afanasiev P., Jouguet B., and Vrinat M . Hydrothermal syntheses and catalytic properties of dispersed molybdenum sulfides. Catalysis Letters, 82 (1-2), 13-17 (2002) 171 Divigalpitiya W.M.R., Frindt R.F., and Morrison S.R. Incusion Systems of Organic Molecules in Restaked Single-Layer Molybdenum Disulphide. Science, 246 (4928), 369-371 (1989) Duangchan A . Residue Upgrading Using Dispersed Catalysts prepared in Reverse Micelles. 1998. University of British Columbia. Englezos P and Ka, Applied Parameter Estimation for Chemical Engineers (Chemical Industries). Marcel Dekker Inc., New York, (2001) Eijsbouts S. Review On the flexibility of the active phase in hydrotreating catalysts. Applied Catalysis A: General, 158, 53-92 (1997) Eijsbouts S., Heinerman J.J.L., and Elzerman H.J.W. M0S2 Structures in High-Activity Hydrotreating Catalysts .1. Semiquantitative Method for Evaluation of Transmission Electron-Microscopy Results - Correlations Between Hydrodesulfurization and Hydrodenitrogenation Activities and M0S2 Dispersion. Applied Catalysis A-General, 105 (1), 53-68 (1993) Farag H. , Sakanishi K. , Kouzu M . , Matsumura A. , Sugimoto Y . , and Saito I. Investigation of the influence of H2S on hydrodesulfurization of dibenzothiophene over a bulk M0S2 catalyst. Industrial & Engineering Chemistry Research, 42 (2), 306-310 (2003a) Farag H. , Sakanishi K. , Kouzu M . , Matsumura A. , Sugimoto Y . , and Salto I. Dibenzothiophene hydrodesulfurization over synthesized M0S2 catalysts. Journal of Molecular Catalysis A-Chemical, 206 (1-2), 399-408 (2003b) Farag H. , Sakanishi K. , Kouzu M . , Matsumura A. , Sugimoto Y . , and Salto I. Dual character of H2S as promoter and inhibitor for hydrodesulfurization of dibenzothiophene. Catalysis Communications, 4 (7), 321-326 (2003c) Fixari B., Peureux S., Elmouchnino J., Leperchec P., Vrinat M . , and Morel F. New Developments in Deep Hydroconversion of Heavy Oil Residues with Dispersed Catalysts 1. Effect of Metals and Experimental Conditions. Energy & Fuels, 8 (3), 588-592 (1994) Furimsky E. Catalytic hydrodeoxygenation. Applied Catalysis A-General, 199 (2), 147-190 (2000) Furimsky E. Hydrodenitrogenation of petroleum. Catalysis Reviews-Science and Engineering, 47 (3), 297-489 (2005) Furimsky E. and Massoth F.E. Regeneration of Hydroprocessing Catalysts. Catalysis Today, 17 (4), 535-& (1993) Galvan D.H., Alonso G., Rangel R., Del Valle M . , Adem E., and Fuentes S. Catalytic study of WS2 undergoing electron irradiation. Journal of Catalysis, 189 (2), 263-268 (2000) 172 Gates B.C. and Topsoe H. Reactivities in deep catalytic hydrodesulfurization: Challenges, opportunities, and the importance of 4-methyldibenzothiophene and 4,6-dimethyldibenzothiophene. Polyhedron, 16 (18), 3213-3217 (1997) Ghosh A. , Chapman W.G., and French R.N. Gas solubility in hydrocarbons - a SAFT-based approach. Fluid Phase Equilibria, 209 (2), 229-243 (2003) Girgis M.J . and Gates B.C. Reactivities, Reaction Networks, and Kinetics in High-Pressure Catalytic Hydroprocessing. Industrial & Engineering Chemistry Research, 30 (9), 2021-2058 (1991) Gordon R.A., Yang D., Crozier E.D., Jiang D.T., and Frindt R.F. Structures of exfoliated single layers of WS2, M0S2, and MoSe2 in aqueous suspension. Physical Review B: Condensed Matter and Materials Physics, 65 (12), 125407-1-125407/9 (2002) Grange P. and Vanhaeren X . Hydrotreating catalysts, an old story with new challenges. Catalysis Today, 36 (4), 375-391 (1997) Gray M R , Upgrading Petroleum Residues and Heavy Oils. Marcel Dekker, Inc., New York, (1994) Hall A .G . , Duangchan A. , and Smith K.J . Characterization of dispersed hydroprocessing catalysts prepared in reversed micelles. Canadian Journal of Chemical Engineering, 76 (4), 744-752 (1998) Hensen E.J.M., Kooyman P.J., van der Meer Y. , van der Kraan A . M . , de Beer V.H.J. , van Veen J.A.R., and van Santen R.A. The relation between morphology and hydrotreating activity for supported M0S2 particles. Journal of Catalysis, 199 (2), 224-235 (2001) Ho T.C. Inhibiting effects in hydrodesulfurization of 4,6-diethyldibenzothiophene. Journal of Catalysis, 219 (2), 442-451 (2003) Iwata Y. , Araki Y . , Honna K. , Mik i Y. , Sato K., and Shimada H" Hydrogenation active sites of unsupported molybdenum sulfide catalyst for hydro-processing heavy oils. Catal Today, 65, 335-341 (2001) Iwata Y. , Sato K. , Yoneda T., Mik i Y. , Sugimoto Y. , Nishijima A. , and Shimada H. Catalytic functionality of unsupported molybdenum sulfide catalysts prepared with different methods. Catalysis Today, 45 (1-4), 353-359 (1998) Joensen P., Crozier E.D., Alberding N . , and Frindt R.F. A study of single-layer and restacked M0S2 by x-ray diffraction and x-ray absorption spectroscopy. Journal of Physic C: Solid State Physic, 20, 4043-4053 (1987) Joensen P., Frindt R.F., and Morrison S.R. Single-layer M0S2. Material Research Bulletin, 21, 457-461 (1986) 173 Kasztelan S. A Descriptive Model of Surface Sites on M0S2 (WS2) Particles. Langmuir, 6 (3), 590-595 (1990) Kasztelan S., Toulhoat PL, Grimblot J., and Bonnelle J.P. A Geometrical Model of the Active Phase of Hydrotreating Catalysts. Applied Catalysis, 13, 127-159 (1984) Kennepohl D. and Sanford E. Conversion of Athabasca bitumen with dispersed and supported Mo-based catalysts as a function of dispersed catalyst concentration. Energy & Fuels, 10 (1), 229-234 (1996) Kim H. , Curtis C.W., Cronauer D.C., and Saykowsky D.J. American Chemical Society Div Fuel Chemical Preprint, 34 (4), 1431 (1989) Kim H.G. and Curtis C W . Reaction Pathways of Model Coprocessing Systems Using Molybdenum Naphthenate and Excess Sulfur. Energy & Fuels, 4 (2), 206-214 (1990) Kochubei D.L, Rogov V . A . , Babenko V.P., Bogdanov S.V., and Zaikovskii V.I. Structure and thiophene hydrodesulfurization activity of M0S2/AI2O3 catalysts. Kinetics and Catalysis, 44(1), 135-140(2003) Kochubey D.L and Babenko V.P. Structure of MoS2-based catalysts for hydrodesulfurization prepared via exfoliation. Reaction Kinetics and Catalysis Letters, 77 (2), 237-243 (2002) Kochubey D.L, Rogov V . A . , and Babenko V.P. Thiophene hydrodesulfurization activity of M0S2/AI2O3 catalysts prepared via exfoliation. The effect of cobalt. Reaction Kinetics and Catalysis Letters, 83 (1), 181-186 (2004) Lai W . C and Smith K.J . Heavy oil microfiltration using ceramic monolith membranes. Fuel, 80 (8), 1121-1130 (2001) Landau M . V . Deep hydrotreating of middle distillates from crude and shale oils. Catalysis Today, 36 (4), 393-429 (1997) Lanier D. Heavy oil - A major energy source for the 21st century. UNITAR Centre for Heavy Crude and Tar Sands 1998, 1-5 (1998) Lee D.K., Yoon W.L., and Woo S.I. Hydrotreatment of an atmospheric residual oil over the dispersed cobalt and molybdenum catalysts in a carbon expanded- bed reactor. Fuel, 75 (10), 1186-1192 (1996) Liu C.G., Zhou J.S., Que G.H., Liang W.J., and Zhu Y.J . Hydrocracking of Gudao Residue with Dispersed-Phase Mo Catalyst. Fuel, 73 (9), 1544-1550 (1994) Liu JK and Gunning H.E. Syncrude Analytical Methods Manual for Bitumen Upgrading. Alberta oil Sands Technology and Research Authority, Edmonton, Alberta, (1991) Lott R and Lee L . K . Upgrading of Heavy Crude Oils and Residues with (HC)s Hydrocracking Technology. 2003. New Orleans. 2003 AIChE annual meeting. 174 Lott R.K., Lee L .K . , and Lai W.C. Alberta Oil Sands Technology and Research Authority. [2,025,125]. 1993. Canada. Massoth F.E. and K i m S.C. Kinetics of the H D N of quinoline under vapor-phase conditions. Industrial & Engineering Chemistry Research, 42 (5), 1011-1022 (2003) Massoth F.E., Muralidhar G., and Shabtai J. Catalytic Functionalities of Supported Sulfides .2. Effect of Support on Mo Dispersion. Journal of Catalysis, 85 (1), 53-62 (1984) Miremadi B.K. , Cowan T., and Morrison S.R. New structures from exfoliated M0S2. Journal of Applied Physics, 69 (9), 6373-6379 (1991) Miremadi B .K. and Morrison S.R. High Activity Catalyst from Exfoliated M0S2. Journal of Catalysis, 103, 334-345 (1987) Miremadi B.K. and Morrison S.R. Aluminium Inclusions and Precipitates in Restacked Exfoliated MoS 2 . Journal of Catalysis, 112, 418-426 (1988) Mochida I. and Choi K . H . An overview of hydrodesulfurization and hydrodenitrogenation. Journal of the Japan Petroleum Institute, 47 (3), 145-163 (2004) Moritis G. Oil sands drive Canada's oil production growth. Oil & Gas Journal, 102 (21), 43 (2004) Nagai M . , Goto Y . , Irisawa A. , and Omi S. Catalytic activity and surface properties of nitrided molybdena-alumina for carbazole hydrodenitrogenation. Journal of Catalysis, 191 (1), 128-137(2000) Nagai M . , Goto Y. , Uchino O., and Omi S. TPD and X R D studies of molybdenum nitride and its activity for hydrodenitrogenation of carbazole. Catalysis Today, 43 (3-4), 249-259 (1998) Nagai M . , Masunaga T., and Hanaoka N . Hydrodenitrogenation of Carbazole on A M0/AI2O3 Catalyst - Effects of Sulfiding and Sulfur-Compounds. Energy & Fuels, 2 (5), 645-651 (1988) Niemann K. and Wenzel F. The Veba-Combi-Cracking-Technology - A n Update. Fuel Processing Technology, 35 (1-2), 1-20 (1993) Panariti N . , Del Bianco A. , Del Piero G., and Marchionna M . Petroleum residue upgrading with dispersed catalysts Part 1. Catalysts activity and selectivity. Applied Catalysis A-General, 204 (2), 203-213 (2000a) Panariti N . , Del Bianco A. , Del Piero G., Marchionna M . , and Carniti P. Petroleum residue upgrading with dispersed catalysts Part 2. Effect of operating conditions. Applied Catalysis A-General, 204 (2), 215-222 (2000b) 175 Park J.K., Robinson R.L.Jr., and Gasem K . A . M . Solubilities of Hydrogen in Heavy Normal Paraffins at Temperatures from 323.2 to 423.2 K and Pressures to 17.4 MPa. Journal of Chemical & Engineering Data, 40 (1), 241-244 (1995) Peureux S., Bonnamy S., Fixari B., Lambert F., Leperchec P., Pepindonat B., and Vrinat M . Deep Hydroconversion of Heavy Oil Residues with Dispersed Catalysts - Analysis of the Transformation. Bulletin des Societes Chimiques Beiges, 104 (4-5), 359-366 (1995) Plantenga F.L. and Leliveld R.G. Sulfur in fuels: more stringent sulfur specifications for fuels are driving innovation. Applied Catalysis A-General, 248 (1-2), 1-7 (2003) Prins R., Jian M . , and Flechsenhar M . Mechanism and kinetics of hydrodenitrogenation. Polyhedron, 16 (18), 3235-3246 (1997) Pruden B B , Muir G, and Skripek M . The CANMET Hydrocracking Process: Recent Developments. Edmonton, Alberta, Canada. Oil Sands Our Petroleum Future Conference. 1993. Rahimi P., Gentzis T., Taylor E., Carson D., Nowlan V. , and Cotte E. The impact of cut point on the processability of Athabasca bitumen. Fuel, 80 (8), 1147-1154 (2001) Rahimi P .M. and Gentzis T. Thermal hydrocracking of Cold Lake vacuum bottoms asphaltenes and their subcomponents. Fuel Processing Technology, 80 (1), 69-79 (2003) Rangwala H.A., Lana I.G.D., Otto F.D., Yeniova H. , and Alnuaimi K . Influence of Catalyst Properties and Operating-Conditions on Hydrodenitrogenation of Quinoline. Energy & Fuels, 4 (5), 599-604(1990) Roxlo C.B., Daage M . , Ruppert A.F. , and Chianelli R.R. Optical-Absorption and Catalytic Activity of Molybdenum Sulfide Edge Surfaces. Journal of Catalysis, 100 (1), 176-184 (1986) Roxlo C.B., Deckman H.W., Gland J., Cameron S.D., and Chianelli R.R. Edge Surfaces in Lithographically Textured Molybdenum-Disulfide. Science, 235 (4796), 1629-1631 (1987) Rozwadowski M . , Lezanska M . , Wloch J., Erdmann K. , Golembiewski R., and Kornatowski J. Investigation of coke deposits on Al -MCM-41. Chemistry of Materials, 13 (5), 1609-1616 (2001) Rueda N . , Bacaud R., Lanteri P., and Vrinat M . Factorial design for the evaluation of the influence of preparation parameters upon the properties of dispersed molybdenum sulfide catalysts. Applied Catalysis A-General, 215 (1-2), 81-89 (2001) Rueda N . , Bacaud R., and Vrinat M . Highly dispersed, nonsupported molybdenum sulfides. Journal of Catalysis, 169 (1), 404-406 (1997) 176 Sakashita Y. , Araki Y . , Honna K. , and Shimada H. Orientation and morphology of molybdenum sulfide catalysts supported on titania particles, observed by using high- resolution electron microscopy. Applied Catalysis A-General, 197 (2), 247-253 (2000) Sakashita Y . , Araki Y . , and Shimada H. Effects of surface orientation of alumina supports on the catalytic functionality of molybdenum sulfide catalysts. Applied Catalysis A-General, 215(1-2), 101-110(2001) Sakashita Y . and Yoneda T. Orientation of M0S2 clusters supported on two kinds of gamma-AI2O3 single crystal surfaces with different indices. Journal of Catalysis, 185 (2), 487-495 (1999) Sanford E.C. Molecular Approach to Understanding Residuum Conversion. Industrial & Engineering Chemistry Research, 33 (1), 109-117(1994) Sanford E.C. Conradson Carbon Residue Conversion During Hydrocracking of Athabasca Bitumen - Catalyst Mechanism and Deactivation. Energy & Fuels, 9 (3), 549-559 (1995) Sato K. , Iwata Y . , Honna K. , Araki Y. , Toneda T., Mik i Y . , and Shimada H. Proposal of new dispersed catalysts for the primary upgrading of residual oils. Science and Technology in Catalysis 1998,121, 411-414 (1999) Satterfield C.N. Heterogeneous Catalysis In Industrial Practice, 2nd Edition. McGraw-Hill, Inc., (1991) Satterfield C.N. and Carter D.L. Effects of Water-Vapor on the Catalytic Hydrodenitrogenation of Quinoline. Industrial & Engineering Chemistry Process Design and Development, 20 (3), 538-540(1981) Satterfield C.N. and Smith C M . Effect of Water on the Catalytic Reaction Network of Quinoline Hydrodenitrogenation. Industrial & Engineering Chemistry Process Design and Development, 25 (4), 942-949 (1986) Satterfield C.N., Smith C M . , and Ingalls M . Catalytic Hydrodenitrogenation of Quinoline -Effect of Water and Hydrogen-Sulfide. Industrial & Engineering Chemistry Process Design and Development, 24 (4), 1000-1004 (1985) Scholz G.A. and Morrison S.R. Methanation on Exfoliated and Supported Mos2. Canadian Journal of Chemistry-Revue Canadienne de Chimie, 67 (5), 862-866 (1989) Schumacher M . M . Heavy Oil and Tar Sands Rcovery and Upgrading. In: Schumacher M M (ed) Noyes Data Corporation, New Jersey, (1982) Schweiger H. , Raybaud P., Kresse G., and Toulhoat H. Shape and edge sites modifications of M0S2 catalytic nanoparticles induced by working conditions: A theoretical study. Journal of Catalysis, 207 (1), 76-87 (2002) 177 Shimada H . Morphology and orientation of M0S2 clusters on AI2O3 and TiCh supports and their effect on catalytic performance. Catalysis Today, 86 (1-4), 17-29 (2003) Speight J .G. The Chemistry and Technology of Petroleum. Marcel Dekker, Inc., New York, (1991) Szymanska A . Lewandowski M . , Sayag C , and Djega-Mariadassou G . Kinetic study of the hydrodenitrogenation of carbazole over bulk molybdenum carbide. Journal of Catalysis, 218(1), 24-31 (2003) Ting P.S., Curtis C . W . , and Cronauer D . C . Reactions of Model Compounds in the Presence of Pyrene and M0S2 to Simulate Coprocessing. Energy & Fuels, 6 (4), 511-518 (1992) Topsoe PL, Candia R., Topsoe N . Y . , and Clausen B .S . On the State of the Co-Mo-S Model . Bulletin des Societes Chimiques Beiges, 93 (8-9), 783-806 (1984) Topsoe H . and Clausen B .S . Importance of Co-Mo-S Type Structures in Hydrodesulfurization. Catalysis Reviews-Science and Engineering, 26 (3-4), 395-420 (1984) Topsoe H . and Clausen B .S . Active-Sites and Support Effects in Hydrodesulfurization Catalysts. Applied Catalysis, 25 (1-2), 273-293 (1986) Trasobares S., Callejas M . A . , Benito A . M . , Martinez M . T . , Severin D. , and Brouwer L . Kinetics of Conradson carbon residue conversion in the catalytic hydroprocessing of a Maya residue. Industrial & Engineering Chemistry Research, 37 (1), 11-17 (1998) Trasobares S., Callejas M . A . , Benito A . M . , Martinez M . T . , Severin D . , and Brouwer L . Upgrading of a petroleum residue. Kinetics of Conradson carbon residue conversion. Industrial & Engineering Chemistry Research, 38 (3), 938-943 (1999) Triola M F , Elementary statistics. Addison Wesley Longman, Inc, (1999) Trytten L . C . , Gray M . R . , and Sanford E . C . Hydroprocessing of Narrow-Boil ing Gas O i l Fractions - Dependence of Reaction-Kinetics on Molecular-Weight. Industrial & Engineering Chemistry Research, 29 (5), 725-730 (1990) Valle M . D . , Avalos-Borja M , Cruz J, and Fuentes S. Exfoliation of M0S2 catalysts: structural and catalytic changes. 351 [Materials Research Society Symposium Proceedings (Molecularly Designed Ultrafine/Nanostructured Materials)], 287-292. 1994. Vasudevan P.T. and Fierro J . L . G . A review of deep hydrodesulfurization catalysis. Catalysis Reviews-Science and Engineering, 38 (2), 161-188 (1996) Vil java T.R., Komulainen R.S. , and Krause A.O. I . Effect of H 2 S on the stability of C0M0/AI2O3 catalysts during hydrodeoxygenation. Catalysis Today, 60 (1-2), 83-92 (2000) Vradman L . and Landau M . V . Structure-function relations in supported N i - W sulfide hydrogenation catalysts. Catalysis Letters, 77 (1-3), 47-54 (2001) 178 Vrinat M . , Breysse M , Geantet C. , Ramirez J., and Massoth F. Effect of M0S2 Morphology on the H D S Activi ty of Hydrotreating Catalysts. Catalysis Letters, 26 (1-2), 25-35 (1994) Watanabe I., Otake M . , Yoshimoto M . , Sakanishi K . , Korai Y . , and Mochida I. Behaviors o f o i l -soluble molybdenum complexes to form very fine M d S 2 particles in vacuum residue. Fuel, 81 (11-12), 1515-1520 (2002) W u J.Z., Prausnitz J . M . and Firoozabadi A . , Molecular-Thermodynamic Framework for Asphaltene - O i l Equilibria. AIChE Journal, 44 (1998) 1188. Yoneyama Y . and Song C.S . A new method for preparing highly active unsupported M o sulfide. Catalytic activity for hydrogenolysis of 4 - ( l - naphthylmethyl)bibenzyl. Catalysis Today, 50(1), 19-27(1999) Zhao S., Kotlyar L .S . , Woods J.R., Sparks B . D . , Hardacre K . , and Chung K . H . Molecular transformation of Athabasca bitumen end-cuts during coking and hydrocracking. Fuel, 80 (8), 1155-1163 (2001) Zmierczak W. , Muralidhar G. , and Massoth F .E . Studies on Molybdena Catalysts .11. Oxygen-Chemisorption on Sulfided Catalysts. Journal of Catalysis, 11 (2), 432-438 (1982) 179 Methods Used for Heavy O i l Hydrocracking Liquid Analysis CHN analysis The carbon, hydrogen and nitrogen (CHN) content analysis was based on A S T M D5373. The sample was combusted at elevated temperature in an atmosphere of purified oxygen. The sample was converted into carbon dioxide, water vapor, nitrogen oxides and ash, respectively. Then, quantity of the gases was determined in an appropriate reference gas stream. Sulfur content analysis A S T M D4239 method was used to analyze the sulfur content of test samples. The sample was burned in a tube furnace at 1350°C in a stream of oxygen to oxidize the sulfur. The combustion products were absorbed in an iodine-contained aqueous solution. When sulfur dioxide is scrubbed by the diluent, the trace iodine reduces to iodide, thus causing an increase in resistance. The change in resistance was detected. Iodine titrant was then added proportionally to the reaction vessel until the trace excess of iodine was replenished and the solution resistance was back to its initial level. Finally, the sulfur content of the sample was calculated from the volume of titrant used. 180 Microcarban Residue (MCR) Analysis The M C R analysis followed the A S T M D4530 method. The liquid sample was heated in a glass vial under a nitrogen blanket at 500°C for 20 min. The sample cracked and formed coke under these conditions. The volatile portion was removed with nitrogen and the residue was weighted after cooling and reported as Microcarbon Residue (MCR) of the original sample. M C R is an alternative to Conradson Carbon Residue (CCR). M C R has some advantages over C C R test in terms of better precision, smaller sample size, less time consuming and improved sample throughput. The percent M C R conversion is then calculated using the following equation. o/o M C R Conversion = [wt-MCR] F e e d - [wt . M C R ] P r o d u c t x ^ [wt .MCR] F e e d where [wt .MCR] F e e d and [wt .MCR] P r o d u c t denote M C R weight in feed and product, respectively. Simulated Distillation of Petroleum Product The method used followed A S T M D2887 procedure. A known quantity of liquid sample was injected into a gas chromatography where the hydrocarbons were separated by boiling point. The area under the chromatogram was measured as a function of time and the accumulated area was normalized to 100% to obtained percent recovered at a given time. The time axis was converted to boiling point by referring to a calibration curve generated under the same chromatographic conditions. The analysis results were reported as the weight % removed at a particular temperature. 181 Asphaltene analysis A 2 g liquid sample was dissolved in 2 ml toluene. Then 80 ml of n-pentane were added. The sample mixture was kept in darkness and was shaken for 5 minutes at 30 minutes intervals over an elapsed time of 2 hours. The resulting precipitate was then collected by a medium pore (10-15mm) Buchner filter funnel under a slight vacuum, washed with n-pentane, dried at 110°C and weighted as asphaltene (Liu and Gunning, 1991). The percentage of asphaltene conversion could then be determined from the following equation. „, A , , . [wt. of asphaltene]F- A - [wt .of asphaltene]pmduc. i n n % Asphaltene conversion = J F e e d L x 100 (A-2) [wt. of asphaltene] F e e d 182 Appendix B Experimental Results of Cold Lake Heavy Oil Hydrocracking Reactions B - l Summary of Results of Hydrocracking Activity Experiments Table B - l Example of an experimental sheet Experiment 28 Hydrocracking condition Feedstock: Cold Lake heavy oil Catalyst: ex-MoS2-W Concentration: 900 ppm Mo Input gas: H 2 Initial reactor pressure: 500psig Reaction temperature: 415°C Heating rate: 5°C/min Stirrer speed: 350 rpm Reaction period: 1 hr Gas input: Pressure (psig) 500 Mole 0.3119 Temperature (°C) 20 Weight (g) 0.6288 Volume (ml) 220 Experimental data: Time Reactor Furnace Pressure Speed of (min) temperature (°C) temperature (°C) (psig) impeller (rpm) 0 20 19 525 363 5 49 69 588 362 10 79 96 629 363 15 106 117 660 364 20 132 131 694 363 30 159 153 730 363 35 181 164 755 361 40 208 185 811 362 45 232 202 886 360 50 257 221 1017 360 55 282 241 1261 362 60 308 262 1495 367 65 332 281 1658 360 70 353 300 1832 362 75 379 324 1937 361 80 400 341 1937 362 85 414 350 1936 362 90 416 349 1936 368 183 95 415 350 1936 363 100 413 353 1936 363 105 413 351 1936 363 110 415 351 1935 364 115 416 350 1935 365 120 416 349 1935 362 125 416 349 1935 356 130 415 349 1935 356 135 416 349 1935 363 140 416 349 1935 365 145 415 348 1935 362 150 232 196 638 370 Stop Cool down reactor Gas output: Pressure (psig): 336 - ' Temperature (°C): 22 GC analysis: Component Response factor GC area from gas injection Mole fraction Weight (g) 1 2 3 Average CH 4 4.37x10"' 412673 412567 409346 411529 0.1800 0.6266 C 2H 4 1.58xl0"b 48533 48103 47723 48120 0.0761 0.4630 C 2H 6 3.43xl0-7 184047 183044 180904 182665 0.0627 0.4090 C 3H 8 1.05x10"' 168804 166808 165339 166984 0.0175 0.1678 C 4H 8 l.OlxlO"7 25517 19323 19106 21315 0.0022 0.0263 C4H10 8.33x10"" 32365 31262 32942 32190 0.0027 0.0338 H2S 1.74x10-' 245880 238834 238383 241032 0.0421 0.3110 N 2 2.33x10"' 172793 175457 196588 181613 0.0424 -H 2 0.5745 0.2513 Total 1 2.2887 Mass balance Mass before reaction: Mass before reaction Heavy oil: 80.02 g Coke recovered: 1.45 g Catalyst: 5.55 g Product liquid: 81.35 g Gas input: 0.63 g Product gas: 2.29 g Total: 86.20 g Total: 85.09 g % Coke yield = (1.45 / 85.09) x 100 = 1.70 % Liquid yield = (81.35 / (85.09) ><100 = 95.60 % Gas yield = (2.29 / 85.09) ><100 = 2.69 % Recovery = (85.09/86.20) xl00 = 98.71 184 Note that the yield percentage is defined as follows: r-^r wt. of X recovered after reaction . . . % Yield of X = : — x l O O (B-l) Total weight recovered after reaction where X denotes either gas, liquid or coke. Summary of all the heavy oil hydrocracking reaction are summarized in Table B-2. 185 Table B-2 Summary of results of hydrocracking activity experiments No Catalyst Mo concentration, ppm Temperature, °C Gas input 1 - - 430 N 2 2 - - 430 N 2 3 - - 430 95 %H2/5%H2S 4 MoNaph 600 430 95 %H2/5%H2S 5 exMoS2-W 360 430 95 %H2/5%H2S 6 exMoS2-W 900 430 95 %H2/5%H2S 7 exMoS2-W 360 430 95 %H2/5%H2S 8 MoNaph 600 430 95 %H2/5%H2S 9 . - 430 95 %H2/5%H2S 10 * exMoS2-W • 360 400 95 %H2/5%H2S 11 exMoS2-W 360 415 95 %H2/5%H2S 12 exMoS2-W 900 400 95 %H2/5%H2S 13 exMoS2-W 900 415 95 %H2/5%H2S 14 - - 430 H 2 15 exMoS2-W 360 430 H 2 16 MoNaph 600 430 H 2 17 exMoS2-W 360 430 H 2 18 MoNaph 600 430 H 2 19 exMoS2-D 900 415 H 2 20 exMoS2-W 600 415 H 2 21* exMoS2-W 600 415 H 2 22** exMoS2-W 600 415 H 2 23 exMoS2-W 900 415 H 2 24 exMoS2-W 600 400 H 2 25 MoNaph 600 415 95 %H2/5%H2S 26 exMoS2-W 600 430 H 2 27 exMoS2-W (8.37%) 900 415 H 2 28 exMoS2-W 900 415 H 2 29 exMoS2-W 600 415 H 2 30 Catalyst coke 600 415 H 2 31 Thermal coke - 415 H 2 32 Crystalline MoS2 600 415 H 2 Note: exfoliated MoS2 in water: exMoS2-W; exfoliated MoS2 in decalin: exMoS2-D * Using filtrate with 7% asphaltene; **Using retentate with 27% asphaltene 186 Table B-2 cont'd No Coke yield, Liquid yield, Gas yield, H 2 Asphaltene Recovery, % % % consumed, % conversion, % % 1 13.84 86.16 - - 33.39 77.27 2 13.43 86.57 - - 35.34 82.77 3 8.02 87.55 4.43 36.63 46.68 98.65 4 3.90 92.60 3.50 100.00 45.67 93.50 5 5.88 88.26 5.86 83.27 49.64 93.04 6 2.92 88.25 4.37 90.80 54.16 98.89 7 6.71 88.54 4.75 75.05 60.80 99.44 8 3.85 92.69 3.46 98.01 59.63 99.49 9 9.59 84.40 , 6.00, . 58.16 52.54 98.94 10 2.69 95.62 1.69 37.19 42.49 103.31 11 1.55 95.99 2.46 54.35 48.63 99.15 12 1.40 97.03 1.57 43.71 39.98 102.10 13 1.80 92.74 2.75 73.06 49.64 97.60 14 8.24 86.60 5.16 41.79 45.67 99.62 15 5.50 90.04 4.47 66.95 47.24 98.89 16 4.75 90.99 4.25 67.08 38.86 97.47 17 6.66 87.45 4.79 57.79 38.58 90.37 18 4.60 91.58 3.82 61.41 47.07 92.12 19 2.25 95.63 2.11 45.30 46.96 99.04 20 1.08 96.72 2.19 56.06 50.30 99.05 21* 0.44 97.33 2.24 42.22 0.00 99.43 22** 2.38 94.64 2.98 59.6 31.48 99.36 23 1.82 96.36 1.81 73.94 35.79 92.10 24 0.91 97.46 1.63 27.2 27.41 97.46 25 0.98 96.76 2.26 49.95 34.39 98.77 26 6.48 89.80 3.21 73.91 45.67 94.67 27 1.61 95.33 3.06 41.51 23.62 99.55 28 1.70 95.60 2.69 59.76 42.94 98.71 29 1.02 96.21 2.78 50.21 43.05 98.69 30 1.15 94.87 2.78 51.68 30.21 99.48 31 2.70 92.72 3.37 37.42 21.83 99.29 32 2.87 93.74 3.40 34.71 24.62 98.19 187 The result of the liquid product analysis was summarized in Table B-3. The conversion and removal of heteroatom or unwanted composition are defined as the following: % Conversion/removal of X = [ w t " X ^ ~ L W t X J f t o d u c t x l 0 f J ( B _ 2 ) where X refers to either M C R , asphaltene, metals, S or N . Table B-3 Liquid product analysis No Residue H/C Conversion (wt %) >525°C S N MCR Ni V Fe 11 - 1.49 18.83 5.88 13.05 13.52 12.50 24.19 19 1.45 29.58 17.65 24.04 23.18 23.88 -20 25.3 1.55 28.12 14.71 39.53 14.26 18.92 -21 21.7 1.54 29.68 10.0 16.0 6.10 7.90 63.49 22 23.4 1.43 31.18 14.89 24.66 24.25 19.88 97.12 23 - 1.61 26.89 11.76 35.99 15.60 15.85 54.84 24 - 1.48 14.18 5.88 11.95 - - -25 26.7 1.44 22.25 2.94 15.71 7.13 10.55 46.77 26 - 1.39 30.81 23.53 21.98 - - -30 - 1.54 10.27 - 10.40 - - -31 - 1.50 9.78 - 0.22 - - -Error % for C,H,N,S = ± 1.0 %; MCR = ± 1-2 %; Ni, V, Fe = ± 2 % B-2 Error Analysis in Hydrocracking Reactions Hydrocracking experiments were repeated randomly under different conditions including temperature, concentration, gas input and catalyst. The analysis of the resulting experimental data yielded the estimates of experimental errors associated with the reported data. Overall, there were 8 sets of repeated experiments as shown in Table B-4. 188 Table B-4 Repeated experiments Run no. Catalyst Mo concentration, ppm Temperature, °C Gas input 1 & 2 - - 430 N 2 3 & 9 - - 430 95 %H2/5%H2S 4 & 8 MoNaph 600 430 95 %H2/5%H2S 5 & 7 ExMoS2-W 360 430 95 %H2/5%H2S 15&17 MoNaph 600 430 H 2 16& 18 ExMoS2-W 360 430 H 2 20&29 ExMoS2-W 600 415 H 2 23 &28 ExMoS2-W 900 415 H 2 Table B-5 shows the activity measurements of the repeated hydrocracking reactions in Table B-4 with 95% confidence intervals. Table B-5 Hydrocracking reaction with the reaction condition in Table B-4. Run no. Coke yield, % Liquid yield, % Gas yield, % H 2 consumption, % Asphaltene conversion, % 1 &2 13.63 ± 0.41 86.16 ±0.41 - - 34.4 ± 1.9 3 &9 8.99 ± 1.89 85.98 ±3.09 5.22 ± 1.54 47.40 ±21.1 49.6 ±5.7 4&8 3.88 ±0.05 92.65 ± 0.09 3.48 ±0.04 99.01 ± 1.95 52.7 ± 13.7 5&7 6.30 + 0.81 88.40 ± 0.27 5.31 ± 1.09 79.16 ±8.06 55.2 ± 10.9 15&17 6.08 ± 1.14 88.77 ±2.50 4.63 ±0.31 62.4 ± 9.0 42.9 ±8.5 16& 18 4.68 ±0.15 91.29 ±0.57 4.04 ± 0.42 64.25 ±5.56 43.0 ±8.05 20&29 1.05 ±0.06 96.47 ± 0.50 2.49 ± 0.58 53.14 ±5.73 46.7 ± 7.06 23 &28 1.76 ±0.12 95.98 ± 0.74 2.25 ± 0.86 66.9+ 13.9 39.4 ±7.01 From the 8 sets of repeated experiments shown in Table B-4, the standard errors % were calculated and the errors associated with the conversion and the yield data were as presented in Table B-6. Note that the standard % is associated with the variability arising 189 from experiment. Small standard error % implies less variability and high reproducibility. Typically, this value should be kept not greater than 15%. From these reactions, the standard error % of coke yield was ± 5% for reactions at 415°C and, in some cases, it reached 15% at 430°C. For liquid yield, the standard error was consistently less than 3%. This result is significant since the liquid product constitutes about 95% of the total product. On the other hand, the standard error of gas yield ranged from 1% to 30%. The obvious variability could be owing to some minimal leakage in the reactor. For asphaltene conversion, the average standard error was found be around 14%. Table B-6 Relative standard error % estimated for hydrocracking reactions. Run no. Coke yield, % Liquid yield, % Gas yield, % H 2 consumption, % Asphaltene conversion, % ; i & 2 2.15 0.34 - - 4.01 3 & 9 15.2 2.59 21.3 32.12 8.35 4 & 8 0.91 0.07 0.81 1.42 18.8 5 & 7 9.32 0.22 14.8 7.34 14.3 15&17 13.5 2.03 4.89 10.4 14.3 16& 18 2.34 0.45 7.56 6.24 13.5 20&29 4.01 0.37 16.8 7.79 10.9 23 &28 4.82 0.56 27.7 15.0 12.8 B-3 Calculations B-3.1 Sample of Calculation for Hydrogen Consumption In this sample of calculation, data of experiment 11 is used. Weight of gas input in the reactor before hydrocracking reaction In this case, gas input was 95%H2/5%H2S in the reactor before hydrocraking reaction and 190 the gas mixture was assumed to follow the ideal gas law. Density of heavy oil is approximately l.Og/cm Weight of heavy oil = 80.0 g Volume of heavy oil - 80.0g/1.0g/cm3 = 80.0 cm 3 Volume of reactor = 300 cm 3 3 3 3 Therefore, the volume of input gas mixture in the reactor = 300 cm - 80 cm = 220 cm Before heating up the reactor, assuming that the room temperature = 18°C = 29IK and, reactor vessel pressure = 506 psig From ideal gas law n = PV/RT where n = no. of g-moles in gas phase P = pressure, atm V = volume, cm R = gas constant, 82.06 atm.cm /mol.K T = temperature, K Therefore, mole of H2 = 0.95 n t ot ai - 0.3014 moles mole of H 2 S = 0.05 n t otai = 0.0159 moles The composition of H2/H2S mixture is 95% of H2 and 5% of H2S Molecular mass of H2, M W H j = 2.02 g/mole Molecular mass of H 2 S, M W H j S = 34.08 g/mole Hence, weight of the H2/H2S mixture inlet = mole of H 2 x M W H j + mole of H 2 S x M W H z S = 0.3014 x 2.02 + 0.0159 x 34.08 = 0.6075 + 0.5406= 1.15 g 191 Response factor It was expected that the composition of the gas collected after reaction would be composed mainly of light hydrocarbons ( Q - C 5 ) and H2S, the balance being unreacted H2. Reference gas chromatograms of these gases were obtained by injecting a standard gas with known composition of the gas into the gas chromatograph. The response factors for each gas component (mole %/area), based on 3 repeated analyses were calculated. H2 gas could not be detected because He gas was used as carrier gas. For this reason, the gas components that could be quantified by gas chromatography were CH4, C2H4 ; C2H6, C3H8, C4H8, C 4 H 1 0 , C5H12 and H2S. The H2 gas was determined by the difference. When the gas product collected in the gas bomb from the reaction was injected into the gas chromatograph, the components of the gas were identified by comparing the retention times of the peaks. Each analysis took 45 min. The average area of each peak from at least 2 repeated analyses was converted into mole fraction using the response factors. The response factors used are summarized in Table B-7. Table B-7 Response factor for gas components Retention time (min) Gas component Mole fraction per area 1.220 CH 4 4.37 xlO"' 2.665 C 2 H 4 1.58 x l 0 _ b 7.653 C 2 H 6 3.43 xlO"' 15.361 CsHg 1.05 xlO"' 21.827 C 4 H 8 1.01 xlO" 7 23.704 C4H10 8.33 xlO" 8 11.124 H2S 2.27 xlO"' 14.150 N 2 2.33 xlO" 7 192 Weight of gas output after hydrocracking reaction The total volume of gas output was assumed to be equal to that of the gas input. After cooling the reactor, P = 332 psig = 22.59atm, T = 19°C = 292K, V = 220cm3 Mole fraction of each component in the output gas was calculated. For example mole fraction of CH4 = Average area of CH4 x response factor of CH4 = 308426 x 4.37 x l O - 7 = 0.1349 After mole fraction of each gas was calculated, they were normalized to total gas without N2. There is presumably no N2 in the gas outlet. N2 in the air was mixed to the gas product. Then the partial pressure of each component was calculated. For example partial pressure of CH4 in the output gas PCH4 = normalized mole fraction of C H 4 x total pressure = 0.1495 x 22.59 atm =3.3771 atm Mole of CH4, n C H4 = ( 3 - 3 7 7 1 X 2 2 0 ) = 0.0310 mole (82.05)(292) Molecular mass of CH4, M W C H ( = 16.04 g/mole Weight of CH 4=0.497g Summation of weight for all the gas component = total weight of gas output = 2.05 g 193 The calculations for all the gas components are as follows: Gas component Average area Mole fraction per area Mole fraction, yi Normalized Yi Partial pressure, Pj (atm) Mole of component, Weight (g) C H 4 308426 4.37 * 107 0.1349 0.1495 3.3771 0.0310 0.4974 C 2 H 4 26186 1.58 xlO" 6 0.0414 0.0459 1.0362 0.0095 0.2669 C 2 H 6 146042 3.43 xlO"' 0.0501 0.0555 1.2545 0.0115 0.3464 C3H8 123005 1.05 x 10 ' 0.0129 0.0143 0.3235 0.0030 0.1310 C 4 H 8 21462 1.01 xlO" 7 0.0022 0.0024 0.0544 0.0005 0.0280 C 4 H i 0 31625 8.33 x 10 8 0.0026 0.0029 0.0659 0.0006 0.0352 H2S 263416 2.27 x 10"7 0.0599 0.0663 1.4968 .0.0137 0.4684 N 2 418108 2.33 x 10 ' 0.0975 - - - -Subtotal 0.4015 - 7.6084 0.0699 1.7732 H 2 0.5985 0.6632 14.9828 0.1376 0.2773 Total 1 22.5912 0.2074 2.0505 Therefore, H2 consumption during hydrocracking reaction = H 2 i n p u t - H 2 output x l 0 Q % H 2 input = 0 . 6 0 7 5 - 0 . 2 7 7 3 x l 0 0 % 0.6075 = 54.35 % B-3.2 Determination of Average, Standard Error and % of Standard Error The average of sample, dj was calculated as follows: where, a, = sample i n = number of sample a,==^- (B-3) 194 The standard error, S.E. was calculated as follows: S.E.= n-1 (B-4) The percent of standard error was calculated as follows: S . £ x l 0 0 %S.E.= (B-5) B-3.3 Statistical Difference of Reactions Variance analysis: Comparison of different reactions The coke yield from different hydrocracking reactions at 430°C were as follows: No. Treatment Coke yield %, x; Sample size, \\\ Ti = Sxi SSi = Ex" 1 Coking 13.84, 13.43 2 27.27 371.78 2 Thermal 8.02, 9.95 2 17.97 163.32 3 MoNaph 3.90,3.85 2 7.75 30.03 4 ExMoS 2 -W 6.48 1 6.48 41.99 Total N = In; =7 T = SIxj =59.47 ESxi2=607.12 Note: No. of treatment, k = 4 Analysis of variance (Triola, 1999) was performed to test at 0.05 level of significance, whether the differences among the sample mean was significant. Hypothesis: Null hypothesis, H 0 : \i\ = \i2 = M3 = M4 Alternative hypothesis, H A : The ja's are not all equal. Level of significance (a): a = 0.05 195 Criterion: reject the null hypothesis i f F test > 9.2766, the value of Fo.os for k-1 =4-1 =3 and N-k = 7-4 = 3 degree of freedom, where F is to be determined by a variance analysis. Otherwise, H 0 is accepted or reserved judgement. Calculations: Total sum of squares, S S T = ££x, 2 - T 2 / N = 607.12-(59.47)2(1/ 7) =101.9679 Treatment sum of squares, S S T R = Z(Tf7iii) - T 7 N = (27.27)2(l/2) +(17.97)2(l/2) +(7.75)2(l/2) +(6.48)2 -(59.47)2(1/ 7) = 100.0181 Error sum of squares, S S E = S S T - SS- r r = 101.9679-100.0181 =1.9498 Then, treatment mean square, M S T R = S S T r / ( k - l ) =33.3394 Error mean square, M S E = SS E / (N-k) = 0.6499 and, F = M S T R / M S E = 51.2974 A l l these results are shown in the following table. Table of analysis of variance for different reactions Source of variation Sum of square, Degree of Mean F SS freedom, df square Treatments (reaction) 100.0181 3 (k-1) 33.3394 51.2974 Error 1.9498 3 (N-k) 0.6499 Total 101.9679 6 196 Decision Since F = 51.2974 > F(o.o5,3,3) = 9.2766, the null hypothesis must be rejected. It can be concluded that there is a difference in the coke yield of different reactions; one of the reaction coke yields differ from the rest of the reactions at 95% level of confidence. B-3.4 Variance Analysis: Comparison of H2S and Pure H2 in Hydrocracking Using MoNaph The coke yields from different hydrocracking reactions using MoNaph at 430°C were tabulated as follows: No. Treatment (input gas) Coke yield %, x. Sample size, nj Tj = Zxj SSi = Ex" 1 5%H2S in H 2 3.90,3.85 2 7.75 30.03 2 PureH 2 4.75,4.32 2 9.08 41.30 Total N = En; = 4 T = EExj =16.83 £Ex"= 71.34 Note: No. of treatment, k = 2 Analysis of variance (Triola 1997) was performed to test at 0.05 level of significance, whether the differences among the sample mean was significant. Hypothesis: Null hypothesis, H 0 : ui = u.2 Alternative hypothesis, H A : The u's are not equal. Level of significance (a): 0.05 197 Criterion: reject the null hypothesis i f F test > 18.513, the value of Fo.os for k-1 =2-1 = 1 and N-k = 4-2 = 2 degree of freedom at level of 0.05 significance, where F is to be determined by a variance analysis. Otherwise, H 0 is accepted or reserved judgement. Calculations Total sum of squares, S S T = £ Z x 2 - T 2 / N = 71.34 - (16.83)2(1/ 4) = 0.5353 2 2 Treatment sum of squares, SSr r = = £(Tj /nj) - T /N = (7.75)2(l/2) +(9.08)2(l/2) - (16.83)2(1/ 4) = 0.4413 Error sum of squares, S S E = S S T - S S i r = 0.5353-0.4413 = 0.0940 Then, treatment mean square, M S i r = SST r /(k-l) = 0.4413 Error mean square, M S E = SS E /(N-k) = 0.0470 and, F = M S T r / M S E = 9.3815 A l l these results are shown in the following table: Source of variation Sum of Degree of Mean F square, SS freedom, df square Treatments (input gas) 0.4413 1 (k-1) 0.4413 9.3815 Error 0.0941 2 (N-k) 0.0470 Total 0.5353 3 Decision Since F = 9.3815 < F(o.o5,i,2) = 18.513, the null hypothesis was accepted at 95% level of confidence. It can be concluded that the presence of H2S doesn't significantly affect the activity in hydrocracking reaction using MoNaph with 5% H2S in the input gas. 198 However, when the analysis of variance was performed at 0.1 level of significance, (For level of 0.1 significance, reject the null hypothesis i f F test > 8.5263, otherwise H 0 i s accepted) the null hypothesis was rejected at 90% level of confidence, since F = 9.3815 > F(o.i,i,2) = 8.5263. B-3.5 Variance Analysis: Comparison of H2S and Pure H2 in Hydrocracking Using Exfoliated MoS 2 The coke yields from different hydrocracking using exfoliated M0S2 at 430°C were shown as follows: No. Treatment (input gas) Coke yield %, x; Sample size, n\ Tj = Exj SSi = Sxf" 1 5%H 2S in H 2 5.88,6.71 2 12.59 79.60 2 Pure H 2 5.50, 6.66 2 12.16 74.61 Total N = Enj = 4 T=E£xj =24.75 I 2 V = 154.20 Note: No. of treatment, k = 2 Analysis of variance (Triola 1997) was performed to test at 0.05 level of significance, whether the differences among the sample mean was significant. Hypothesis: Null hypothesis, H 0 : ui = U2 Alternative hypothesis, H A : The (a.'s are not equal. Level of significance (a): 0.05 Criterion: reject the null hypothesis if F test > 18.513, the value of Fn.05 for k-1 =2-1 = 1 and N-k = 4-2 = 2 degree of freedom at level of 0.05 significance, where F is to be determined by a variance analysis. Otherwise, H 0 is accepted or reserved judgement. 199 Calculations Total sum of squares, SS T = I I x 2 - T 2 / N = 154.20 - (24.75)2(1/ 4) = 1.0635 Treatment sum of squares, S S i r = = £(T 2/hj) - T 2 / N = (12.59)2(l/2) +(12,16)2(l/2) - (24.75)2(1/ 4) = 0.0462 Error sum of squares, S S E = S S T - SS-rr = 1.0635-0.0462 = 1.0173 Then, treatment mean square, MSir = SSir/Ck-l) = 0.0462 Error mean square, M S E = SSE/(N-k) = 0.5086 and, F = M S T r / M S E = 0.0909 A l l these results are shown in the following table. Source of variation Sum of square, SS Degree of freedom, df Mean square F Treatments (input gas) 0.0462 1 (k-1) 0.0462 0.0909 Error 1.0173 2 (N-k) 0.5086 Total 1.0635 3 Decision Since F = 0.0909 < F (0.o5,i,2) = 18.513, the null hypothesis was accepted at 95% level of confidence. It can be concluded that the presence of H2S doesn't significantly affect the activity in hydrocracking reaction using MoNaph with 5% H2S in the input gas. When the analysis of variance was performed at 0.1 level of significance, (For level of 0.1 significance, reject the null hypothesis i f F test > 8.5263, otherwise H 0 i s accepted) the null hypothesis was also accepted at 90% level of confidence, since F = 0.0909 < Fm. 1,1,2) = 8.5263. 200 B-3.6 Calculation of Total Heteroatom Removal for the Filtrate and Retentate Table B-8 shows the composition of heteroatom and undesired component in original Cold Lake heavy oil, the filtrate and retentate and Table B-9 shows the removal or conversion of these components after a one hour hydrocracking reaction. Table B-8 Heteroatom, M C R and asphaltene in different heavy oil feedstock. Feed S, wt% N , wt% MCR, wt% Asphl*, wt% V , ppm, N i , ppm Cold Lake 4.09 0.34 13.56 17 179.2 67.3 Filtrate 4.38 0.3 10.31 7 111.4 41 Retentate 5.26 0.47 21.17 27 290.3 115.1 *Asphl refer to asphaltene Note that filtrate and retentate were the product from previous study (Lai and Smith, 2001). Table B-9 Heteroatom removal, and M C R and asphaltene conversion in different heavy oil feedstock after catalytic hydrocracking reaction using exfoliated M0S2 (600 ppm Mo, 415°C, initial pressure of 3.5 MPa, 1-hour reaction time). Feed S, wt% N , wt% MCR, wt% Asphl*, wt% V , ppm, N i , ppm Cold Lake 28.1 14.7 39.5 50.3 18.9 14.3 Filtrate 29.7 10 16 0 7.9 6.1 Retentate 31.2 14.9 24.7 31.5 19.9 24.5 *Asphl refer to asphaltene For Cold Lake heavy oil, it is assumed that, after reaction, 40 wt% went to the filtered oil and 60 wt% is the retentate (based on the residue in the oil). Calculation of total heteroatom 201 removal was based on the basis of lOOg of Cold Lake heavy oil. This implies that lOOg Cold Lake heavy oil would give 40g of filtered oil and 60g of retentate. The amount of heteroatom removed and undesired component converted in 40g of filtered oil and 60g of retentate are given as follows: Feed s,g • N , g MCR, g asphl, g v , g N i , g lOOg Cold Lake 1.1493 0.0500 5.3562 8.5510 0.003387 0.000962 40g filtered oil 0.5203 0.0120 0.6598 0 0.000352 0.000100 60g retentate 0.9847 0.0420 3.1374 5.1030 0.003466 0.001692 Total removed 1.5050 0.0540 3.7972 5.1030 0.003818 0.001792 Total removal, % 30.7 13.4 22.6 26.9 17.5 21.0 The total amount of heteroatom removed from 40g filtered oil and 60g of retentate were then compared to lOOg of Cold Lake heavy oil. The total removal for each heteroatom and undesired component was calculated for example for S total removal from 40g filtered oil and 60g retentate . _050 x l 0 0 % 0.4(4.38)+ 0.6(5.26) = 30.7 % 202 ( a) Feed with different asphaltene % 900 i 800 -0 20 40 60 80 100 Concentration, wt% (b) Boiling curve after reaction from different feed 900 "j 1 800 0 20 40 60 80 100 Concentration, wt% gure B- l Simulated boiling curve for (a) feed heavy oil with different asphaltene content and (b) boiling curves for these feed after hydrocracking reaction (EXM0S2-W at 600 ppm Mo, 415°C, initial pressure of 3.5 MPa, 1-hr reaction). 203 Presence of Water in Model Compound Reactions The catalyst activity of exfoliated M0S2 dispersed in water has been tested in hydrogenation of naphthalene, hydrodesulfurization of DBT, hydrodenitrogenation in quinoline and carbazole, and hydrodeoxygenation of phenol (Table C- l ) . As shown in Table 5.20, in model compound systems, exfoliated M0S2 dispersed in water was found to have much lower reactant conversion and heteroatom removal compared to exfoliated M0S2 dispersed in decalin. In the aforementioned systems, all the reactions seemed to be inhibited. The conversion of reactant was found even lower than that of the thermal reaction in hydrogenation of naphthalene. A thermal reaction provided 16% conversion of naphthalene; only 3.4% conversion was achieved when using exfoliated M0S2 dispersed in water as catalyst. Table C - l Reactant conversion and heteroatom removal using exfoliated M0S2 in water (600 ppm Mo, 350°C, initial pressure 2.8 MPa, 5-hr reaction time). Reaction Reactant Conversion, % Heteroatom removal, % Hydrogenation naphthalene 3.4 -Hydrodesulfurization DBT 1.7 1.2 Hydrodenitrogenation quinoline 68.6 4.52 Hydrodenitrogenation carbazole 15.9 1.24 Hydrodeoxygenation phenol 1.1 1.1 204 When exfoliated M0S2 dispersed in water was used in catalytic hydrocracking of Cold Lake heavy oil reaction, encouraging result was obtained. Better heteroatom removal was observed comparing to that of MoNaph at 415°C. The positive effect of water has also being observed via the increasing amount of water during a catalytic hydrocracking of Cold Lake reaction over exfoliated M0S2 dispersed in water (Section 4.2.3.1). However, this phenomenon was not observed in the hydroprocessing reactions using model compound in the present study. The presence of water not only had no positive effect, but also suppressed the overall reaction regardless of model compounds used. The negative effect of water in model compound reaction was also observed in the catalytic reactions using MoNaph with the presence of similar amount of water. The reactant removal and hetroatom removal in the reaction using MoNaph with the presence of 5 - 6 % of water are presented in Table C-2. It was believed that the presence of water tends to segregate the M0S2 catalyst from the reactant molecules due to the immiscibility of water and the organic reactant medium. Therefore, the diffusion of reactant molecule to the surface of catalyst was greatly reduced. Table C-2 Reactant removal and heteroatom removal using MoNaph with presence of 5-6% of water (600 ppm Mo, 350°C, initial pressure 2.8 MPa, 5-hr reaction time). Reaction Reactant Conversion, % Heteroatom removal, % Hydrodesulfurization DBT 13.3 6.7 Hydrodenitrogenation quinoline 78.2 40.0 Hydrodenitrogenation carbazole 18.9 4.8 205 The poor performance of exfoliated M0S2 dispersed in water could also be explained using the model of exfoliated M0S2 proposed by Divigalpitiya et al. (1989). It was believed that after exfoliation, the M0S2 layers remained in suspension in the water are due to surrounding by negative charge hydroxyl group (Miremadi et al., 1991) as shown in Figure C - l and therefore repelled each other. The OH" groups at the edge sites were assumed to be more tightly bound than those at the basal planes because the basal planes are inert compared with the edge planes (Roxlo et al., 1986; Roxlo et al., 1987). During a reaction, mixing with water immiscible organic medium probably induced displacement of the OH" groups on the basal plane by the organic molecule. However, most of the OH" on the edge planes, which is believed to be where the active sites located, retained. Consequently, the catalyzing purpose of M0S2 was not achieved in the case of exfoliated M0S2 dispersed in water. Therefore, a reaction was uncatalyzed but inhibited by water when exfoliated M0S2 dispersed in water was used. As a result, the positive effect of water observed in hydrocracking reaction was not observed and could not be explained via the model system studied. (a) (b) H H H H H Exfoliated O O O O O MoS 2 layer c C C C C HOI .", 1011 HQS i O H 0 0 0 0 0 c c c c c H H H H H Figure C - l (a) Exfoliated M0S2 in suspension in water (b) Exfoliated M0S2 in the presence of organic solvent. A cross sectional view of the single M0S2 layers, in which C represents an adsorbed organic molecule and OH" represents an adsorbed hydroxyl group (Divigalpitiya et al., 1989). 206 Hydroprocessing Experiments D-l Summary of Hydroprocessing Reactions Table D - l Summary of experiments for model compound hydroprocessing reactions (600 ppm Mo if M0S2 catalyst was used, initial pressure 2.8 MPa, 5-hr reaction). No. M-Reactant Catalyst Temperature, °C Conversion, % Heteroatom removal, % 1 Naphthalene - 325 & 350 - -4 Naphthalene - 375 39.2 -3 Naphthalene MoNaph 350 35.8 -7 Naphthalene ExMoS2-W 350 3.45 -2 Naphthalene ExMoS2-D 350 27.0 -5 Naphthalene ExMoS2-D 325 12.2 -6 Naphthalene ExMoS2-D 375 34.3 -11 Naphthalene ExMoS2-D 375 38.1 -9 Naphthalene MoNaph* 350 74.1 -8 Naphthalene MoNaph* 325 38.9 -10 Naphthalene MoNaph* 375 70.6 -13 Dibenzothiophene MoNaph 350 15.9 3.3 18 Dibenzothiophene MoNaph 325 5.55 1.3 19 Dibenzothiophene MoNaph 375 43.1 34.0 15 Dibenzothiophene MoNaph+water 350 13.3 6.7 14 Dibenzothiophene ExMoS2-W 350 1.7 1.2 12 Dibenzothiophene ExMoS2-D 350 35.3 25.8 21 Dibenzothiophene ExMoS2-D 350 31.0 24.9 16 Dibenzothiophene ExMoS2-D 325 6.5 3.7 17 Dibenzothiophene ExMoS2-D 375 60.5 54.3 20 Dibenzothiophene Crystalline MoS2 350 12.6 8.9 22 Dibenzothiophene AHM 350 99.5 97.4 24 Quinoline MoNaph 350 100 68.9 26 Quinoline MoNaph+water 350 78.2 40.0 207 Table D-2 cont'd 25 Quinoline ExMoS2-W 350 68.6 4.52 23 Quinoline ExMoS2-D 350 76.5 28.2 29 Carbazole MoNaph 350 37.7 30.4 30 Carbazole MoNaph+water 350 18.9 4.8 28 Carbazole ExMoS2-W 350 15.9 1.2 27 Carbazole ExMoS2-D 350 32.6 20.4 31 Carbazole ExMoS2-D 350 30.0 16.2 34 Phenol MoNaph 350 11.1 11.1 33 Phenol ExMoS2-W 350 1.2 1.2 32 Phenol ExMoS2-D 350 15.0 15.0 35 Phenol Crystalline MoS2 350 4.7 4.7 36 Phenol AHM 350 19.3 19.3 Note that all the MoNaph used were from IBC medical Inc.^ except MoNaph* from K&K Lab. Inc In the above table, the % Conversion and heteroatom removal are computed using the following equations: [mole of X ] F e e d - [mole of X ] P r o d u c t % Conversion of X = xlOO [mole of X ] F e e d where X denotes the name of the corresponding reactant. % Heteroatom removal (HDS, HDN, HDO) _ Mole of molecules detected without heteroatom except the solvent Mole of feed reactant or total of unreacted reactant and products (D-l) xlOO (D-2) D-2 Example of Calculation for Model Compound Experiments Response Factors Used in the Model Compound G C Analysis Response factor of reference component was obtained by injecting known composition of 0.05 u,L pure liquid sample into gas chromatography (GC). For each specific component, three samples of different concentrations were injected. The response factor was determined from the slope of line plotted for concentration versus GC area. 208 For example response factors of DBT and quinoline were determined from the slope of plots in Figure D - l . Response factor of DBT = 1.5 x 109 mol/cc per area Response factor of quinoline = 2.3 x 109 mol/cc per area 0 5000 10000 15000 20000 25000 30000 35000 x = GC area 5.0 0 5000 10000 15000 20000 25000 x = GC area Figure D - l Calibration curves for (a) DBT and (b) quinoline to determine respective response factors. 209 Sample of Calculation for Hydrogenation of Naphthalene Data of experiment M-3 (26/08/04) was used (hydrogenation of naphthalene using MoNaph derived M0S2. Concentration of naphthalene and tetralin of samples collected at different periods of reaction were calculated. For example, Concentration of naphthalene at time zero (when reactor reached reaction temperature) = average GC area x response factor of naphthalene = 287190 x 2.0 xlO" 9 mol/cc per area = 5.74 xlO" 4mol/cc Conversion of naphthalene after 5 hr reaction (t) _ [Naphtahlene]t=Q - [Naphthalene]^ ™ X 1 \)\) [Naphthalene]t=0 =____xm% 5.74 = 35.8% Mole ratio of naphthalene after 5 hr reaction (t) _ [Naphtahlene], [Total naphthalene and products], 3.69 3.69 + 1.97 : 0.652 210 Recovery after 5 hr reaction (t) _ [Total molecule], [Total molecule] t = 0 - ^ • " " " x l O O t t 5.76xl0" 4 -= 98.26 mol% Results of calculation for each component at each stage were summarized in the following table: Table D-3 Results of hydrogenation of naphthalene Time, Naphthalene Tetralin Total hr Average mol/cc conversion, % mol ratio Average mol/cc mol ratio mol/cc, area x lO 4 area x lO 4 xlO 4 0 287190 5.74 0.00 0.9965 995 0.02 0.0035 5.76 0.5 279229 5.58 2.77 0.9706 8465 0.17 0.0294 5.75 1 269148 5.38 6.28 0.9377 17885 0.36 0.0623 5.74 1.5 258769 5.18 9.90 0.9022 28042 0.56 0.0978 5.74 2 250033 5.00 12.94 0.8635 39534 0.79 0.1365 5.79 3 223600 4.47 22.14 0.7853 61149 1.22 0.2147 5.69 4 205813 4.12 28.34 0.7157 81749 1.63 0.2843 5.75 5 184505 3.69 35.76 0.6520 98470 1.97 0.3480 5.66 Sample of calculation for HDS of DBT Data of experiment M-12 (19/10/04) was used (HDS of DBT using exMoS 2-D). Four products were detected for HDS of DBT. For example concentration of DBT at time zero = average GC area x response factor of DBT = 13349 x 1.5 xlO" 9 mol/cc per area = 2.01 xlO' 4 mol/cc 211 Conversion of DBT after 5 hr reaction (t) J D B T U - r D B T ] , ^ [DBT], . 0 - ^ ' - " " x l O W t 2.01 = 35.3% Mole ratio of DBT after 5 hr reaction (t) [ D B T ] , [Total DBT and products], 1.30 1.94 = 0.67 Heteroatom removal-HDS _ [corrected total moles],=0 - [total S compound], [corrected total moles],=0 - " • 9 4 - ° ° 1 7 ) - L 4 3 x 100K (1.94-0.017) = 25.6% Recovery after 5 hr reaction (t) _ [Total molecule], [Total molecule] t = 0 - ' • " " " ' - . x l O O * 2.02 x l0~ 5 = 96.04 mol% 212 Results of calculation for each component at each stage were summarized in the following table: Table D-4 Results of hydrodesulfurization of DBT Time, hr DBT THDBT S compound, mol/cc, xlO 5 Avg. area mol/cc xlO 5 Conv., % mol ratio Avg. area mol/cc x lO 6 mol ratio 0 13379 2.01 0.00 0.9917 0 0.00 0.0000 2.01 0.5 13137 1.97 1.81 0.9753 114 0.17 0.0084 1.99 1 12820 1.92 4.18 0.9451 284 0.43 0.0210 1.97 1.5 12177 1.83 8.98 0.9090 449 0.67 0.0335 1.89 2 11634 1.75 13.04 0.8755 592 0.89 0.0445 1.83 3 10578 1.59 20.94 0.8000 761 1.14 0.0576 1.70 4 9805 1.47 26.71 0.7375 844 1.27 0.0635 1.60 5 8650 1.30 35.34 0.6693 856 1.28 0.0662 1.43 Time BP CHB BCH Total hr Avg. mol/cc mol Avg. mol/cc mol Avg. mol/cc mol mol/cc area x lO 6 ratio area xlO 6 ratio area x lO 7 ratio xlO 5 0 90 0.17 0.0083 0 0.00 0.0000 0 0 0.0000 2.02 0.5 176 0.33 0.0163 0 0.00 0.0000 0 0 0.0000 2.02 1 312 0.58 0.0287 57 0.11 0.0052 0 0 0.0000 2.03 1.5 472 0.88 0.0439 146 0.27 0.0136 0 0 0.0000 2.01 2 609 1.14 0.0572 243 0.45 0.0228 0 0 0.0000 1.99 3 904 1.69 0.0852 510 0.95 0.0481 113 1.82 0.0092 1.98 4 1177 2.20 0.1104 791 1.48 0.0741 164 2.87 0.0144 1.99 5 1419 2.65 0.1369 1091 2.04 0.1052 241 4.35 0.0224 1.94 For components such as carbazole and phenol are not completely dissolved in hexadecane at the room temperature. The concentrations of these components in the solvent after reaction were determined by subtracting the total products from feed concentration before reaction. Sample of calculation for HDO of phenol Using hydrodeoxygenation of phenol as an example (data of experiment M-32 is used). Concentrations of products were first calculated. In this case, benzene is the only product. Concentration of benzene was calculated, for example, at time zero 213 = average GC area x response factor of benzene = 382 x 4.00 x lO" 9 mol/cc per area = 1.53 xlO - 6 mol/cc With initial concentration of phenol = 4.02 xlO" mol/cc, concentration of phenol at time zero = initial concentration of phenol - concentration of total products (e.g. benzene) = 4.02 xlO" 4mol/cc - 1.53 xlO" 6mol/cc = 4.01 xlO" 4 mol/cc These steps were repeated for all samples collected at different period during reaction and the results of calculation were summarized in Table D-5. Table D-5 Results of hydrodeoxygenation of phenol Time, hr Average area Benzene, mol/cc xlO5 Phenol mol/cc xlO4 conversion, % 0 382 0.15 4.01 0.00 0.5 2738 1.10 3.91 2.35 1 4634 1.85 3.84 4.24 1.5 6745 2.70 3.75 6.35 2 8540 3.42 3.68 8.14 3 11138 4.46 3.58 10.73 4 13437 5.37 3.49 13.03 5 15379 6.15 3.41 14.96 214 D-3 Repeatability of Model Compound Experiments Selected experiment of model compound reactions were repeated to check for the experimental error. Repeatability of hydrogenation of naphthalene (Experiment M-6 & M-l l ) Reactant: naphthalene (10 wt%) Solvent: hexadecane (100ml) Catalyst: exMoS 2-D Mo concentration: 600 ppm Input gas: 5%H 2S/95%H 2 Initial reactor pressure: 2.8 MPa Reaction temperature: 375°C Heating rate: 10°C/min Stirrer speed: 1200 rpm Reaction period: 5 hr Time, hr Mol ratio of naphthalene 1 2 Average S.E. % S.E. 0 0.9948 0.9923 0.9935 1.83 x l 0 " J 0.18 0.5 0.9576 0.9463 0.9520 8.02 x l 0 - J 0.84 1 0.9216 0.9015 0.9116 1.42 x\0'z 1.56 1.5 0.8864 0.8569 0.8716 2.09 x lO - 2 2.40 2 0.8504 0.8168 0.8336 2.38 xlO" 2 2.86 3 0.7801 0.7375 0.7588 3.01 x\0-2 3.97 4 0.7380 0.6760 0.7070 4.39 xlO" 2 6.21 5 0.6949 0.6228 0.6588 5.10 x l 0 " z 7.74 Time, hr Mol ratio of tetralin 1 2 Average S.E. % S.E. 0 0.0052 0.0077 0.0065 1.83 x l 0 " J . 28.15 0.5 0.0424 0.0537 0.0480 8.02 x l f j - J 16.71 1 0.0784 0.0985 0.0884 1.42 xlO'2 16.06 1.5 0.1136 0.1431 0.1284 2.09 xlO" 2 16.28 2 0.1496 0.1832 0.1664 2.38 xl(V* 14.30 3 0.2199 0.2625 0.2412 3.01 xlO" 2 12.48 4 0.2620 0.3240 0.2930 4.39 x lO - 2 14.98 5 0.3051 0.3772 0.3412 5.10 xlQ- 2 14.95 Conversion: 1 2 Average S.E. % S.E. 34.27 38.08 36.18 2.69 7.47 215 Repeatability of HDS of DBT Experiment M-12 and M-21 Reactant: DBT Heteroatom: 900ppm S Solvent: hexadecane Volume: 100ml Catalyst: exMoS 2-D Concentration: 600 ppm Mo Input gas: 5%H 2S/95%H 2 Initial reactor pressure: 2.8 MPa Reaction temperature: 350°C Heating rate: 10°C/min Stirrer speed: 1200 rpm Reaction period: 5 hr Product liquid analysis: Time, hr Mol ratio of DBT 1 2 Average S.E. % S.E. 0 0.9917 0.9918 0.9918 7.07x10"5 0.01 0.5 0.9753 0.9662 0.9708 6.43x10"' 0.66 1 0.9451 0.9405 0.9428 3.25x10"' 0.34 1.5 0.9090 0.9166 0.9128 5.37x10"' 0.59 2 0.8755 0.8916 0.8836 1.14xl0" 2 1.29 3 0.8000 0.8351 0.8176 2.48xl0" 2 3.03 4 0.7375 0.7754 0.7565 2.68xl0" 2 3.54 5 0.6693 0.7115 0.6904 2.98 xlO" 2 4.32 Time, hr Mol ratio of THDBT 1 2 Average S.E. x lO ' % S.E. 0 0.0000 0.0000 0.0000 0.00 0 0.5 0.0084 0.0176 0.0130 6.51 50.08* 1 0.0210 0.0263 0.0237 3.75 15.82 1.5 0.0335 0.0322 0.0329 0.92 2.79 2 0.0445 0.0373 0.0409 5.09 12.44 3 0.0576 0.0504 0.0540 5.09 9.43 4 0.0635 0.0599 0.0617 2.55 4.13 5 0.0662 0.0622 0.0642 2.83 4.41 216 Time, hr Mol ratio o fBP 1 2 Average S.E. % S.E. 0 0.0083 0.0082 0.0083 7.07 x 1 0 3 0.85 0.5 0.0163 0.0161 0.0162 1.41 xlO" 4 0.87 1 0.0287 0.0288 0.0288 7.07 xlO" 3 0.25 1.5 0.0439 0.0427 0.0433 8.49 xlO" 4 1.96 2 0.0572 0.0555 0.0564 1.20 x l 0 " J 2.13 3 0.0852 0.0805 0.0829 3.32 x l 0 " J 4.00 4 0.1104 0.1053 0.1079 3.61 x l 0 " J 3.35 5 0.1369 0.1355 0.1362 9.90 xlO" 4 0.73 Time, hr Mol ratio of CHB 1 2 Average S.E. % S.E. 0 0.0000 0.0000 0.0000 0 0 0.5 0.0000 0.0000 0.0000 0 0 1 0.0052 0.0045 0.0049 4.95 xlO" 4 10.10 1.5 0.0136 0.0086 0.0111 3.54 x l0 " J 31.89 2 0.0228 0.0156 0.0192 5.09 x l0 " J 26.51 3 0.0481 0.0341 0.0411 9.90 x l 0 " j 24.09 4 0.0741 0.0594 0.0668 1.04 xlO" 2 15.57 5 0.1052 0.0908 0.0980 1.02 xlO" 2 10.41 Time, hr Mol ratio of B C H 1 2 Average S.E. % S.E. 0 0.0000 0.0000 0.0000 0 0 0.5 0.0000 0.0000 0.0000 0 0 1 0.0000 0.0000 0.0000 0 0 1.5 0.0000 0.0000 0.0000 0 0 2 0.0000 0.0048 0.0024 3.39 xlO"-3 141.3* 3 0.0092 0.0090 0.0091 1.41 xlO" 4 1.55 4 0.0144 0.0162 0.0153 1.27 xlO" J 8.30 5 0.0224 0.0224 0.0224 0 0.00 Conversion 1 2 Average S.E. % S.E. 35.34 31.00 33.17 3.07 9.26 Activity of HDS 1 2 Average S.E. % S.E. 25.8 24.9 25.35 6.36X10" 1 2.51 217 Repeatability of HDN of carbazole (Experiment M-27 & M-21) Reactant: carbazole Heteroatom: 200ppm N Solvent: hexadecane Volume: 100ml Catalyst: exMoS 2-D Concentration: 600 ppm Mo Input gas: 5%H 2S/95%H 2 Initial reactor pressure: 2.8 MPa Reaction temperature: 350°C Heating rate: 10°C/min Stirrer speed: 1200 rpm Reaction period: 5 hr Time, hr Mol ratio o f C B Z 1 2 Average S.E. % S.E. 0 0.9881 1.000 0.9941 8.41 x i 0 " j 0.85 0.5 0.9299 0.941 0.9355 7.85 x l 0 _ J 0.84 1 0.8966 0.913 0.9048 1.16 xlO" 2 1.28 1.5 0.8430 0.890 0.8665 3.32 x\0'z 3.83 2 0.8209 0.868 0.8445 3.33 xlO" 2 3.94 3 0.7701 0.825 0.7976 3.88 xlO'2 4.86 4 0.7161 0.773 0.7446 4.02 x l 0 _ / 5.40 5 0.6664 0.700 0.6832 2.38 x\Q-z 3.48 Time, hr Mol ratio of THCBZ 1 2 Average S.E. % S.E. 0 0.0119 0.000 0.0060 8.41 x i 0 " J 140.1* 0.5 0.0653 0.059 0.0622 4.45 x l 0 " J 7.15 1 0.0902 0.081 0.0856 6.51 x l 0 " J 7.61 1.5 0.1155 0.094 0.1048 1.52 xlO" 2 14.50 2 0.1215 0.099 0.1103 1.59 x\o-z 14.42 3 0.1295 0.109 0.1193 1.45 x\0-z 12.15 4 0.1315 0.118 0.1248 9.55 x l 0 _ J 7.65 5 0.1300 0.138 0.1340 5.66 x l 0 " j 4.22 Time, hr Mol ratio of CHB 1 2 Average S.E. % S.E. 0 0.0000 0.000 0.0000 0 0 0.5 0.0000 0.000 0.0000 0 0 1 0.0000 0.000 0.0000 0 0 1.5 0.0091 0.000 0.0046 6.43 x l 0 " J 139.8* 2 0.0105 0.006 0.0083 3.18 xl0" J 38.31 3 0.0150 0.009 0.0120 4.24 xKT* 35.33 4 0.0175 0.014 0.0158 2.47 x l 0 " J 15.63 5 0.0215 0.020 0.0208 1.06 x l 0 _ J 5.10 218 Time, hr Mol ratio of CHCHe 1 2 Average S.E. % S.E. 0 0.0000 0.000 0.0000 0 0 0.5 0.0000 0.000 0.0000 0 0 1 0.0000 0.000 0.0000 0 0 1.5 0.0000 0.000 0.0000 0 0 2 0.0000 0.000 0.0000 0 0 3 0.0000 0.000 0.0000 0 0 4 0.0083 0.014 0.0112 4.03 x l 0 ' j 35.98 5 0.0116 0.022 0.0168 7.35 x l ( r J 43.75 Time, hr Mol ratio of B C H 1 2 Average S.E. x l 0 J % S.E. 0 0.0000 0.000 0.0000 0 0 0.5 0.0048 0.000 0.0024 3.39 141.25* 1 0.0091 0.006 0.0076 2.19 28.82 1.5 0.0136 0.010 0.0118 2.55 21.61 2 0.0172 0.014 0.0156 2.26 14.49 3 0.0259 0.019 0.0225 4.88 21.69 4 0.0354 0.030 0.0327 3.82 11.68 5 0.0459 0.039 0.0425 4.88 11.48 * Time, hr ' Mol ratio of C P M C H 1 2 Average S.E. x l 0 j % S.E. 0 0.0000 0.000 0.0000 0 0 0.5 0.0000 0.000 0.0000 0 0 1 0.0041 0.000 0.0021 2.90 138.10* 1.5 0.0110 0.005 0.0080 4.24 53.00 2 0.0176 0.013 0.0153 3.25 21.24 3 0.0317 0.029 0.0304 1.91 6.28 4 0.0484 0.044 0.0462 3.11 6.73 5 0.0657 0.063 0.0644 1.91 2.97 Time, hr Mol ratio of H C H 1 2 Average S.E. % S.E. 0 0.0000 0.000 0.0000 0 0 0.5 0.0000 0.000 0.0000 0 0 1 0.0000 0.000 0.0000 0 0 1.5 0.0078 0.000 0.0039 5.52 x l 0 _ J 141.54* 2 0.0124 0.000 0.0062 8.77 x l0 " J 141.45* 3 0.0278 0.009 0.0184 1.33 xlO" 2 72.28 4 0.0428 0.018 0.0304 1.75 xlO" 2 57.57 5 0.0589 0.034 0.0465 1.76 xlO" 2 37.85 219 Conversion (mol %) 1 2 Average S.E. % S.E. 32.56 30.00 31.28 1.81 5.79 Activity of H D N (mol %) 1 2 Average S.E. % S.E. 20.4 16.2 18.3 2.97 16.23 Note that in the above results, there are few cases where the % S.E. is excessively large (those marked with *). It is believed that these large %S.E values were neither caused by the poor reproducibility of the experiments nor by the poor reacting system consistency. In deed, as can be observed from the above data compilation that they occurs during the first appearance of the intermediate products. For example, in the case of HDS of DBT, the excessive large %S.E. values were associated with the first appearance for THBDT at the sampling time of 0.5hr and B C H at the sampling time of 2hr, respectively. Similar observations also found in the H D N of carbazole system. 220 D-4 Calculated Data for Model Compound Experiments In what follows, information such as reactant concentration and product concentration at each sampling time for all the experiments using various model compound systems are tabulated. The final product recovery is summarized at the end of each tabulated result. Note that all the MoNaph used was from ICN Biomedicals Inc. except those specifically mentioned from K & K Lab Inc. Experiment M-l Reaction: Hydrogenation of naphthalene Reactant: naphthalene (10 wt%) Solvent: hexadecane (100ml) Catalyst: - Concentration: -Input4gas: 5%H 2S/95%H 2 Initial reactor pressure: 2.8 MPa Reaction temperature: 325°C & 350°C Heating rate: 10°C/min Stirrer speed: 1200 rpm Reaction period: 2 hr at 325°C & 2 hr at 350°C Product liquid analysis: Time, hr Temp., °C Pressure, psig Naphthalene Tetralin Total mol/cc mol/cc Conv., % mol ratio mol/cc mol ratio 0 325 568 6.39x10-4 0.00 0.9996 2.67x10-' 0.0004 6.39xl0' 4 0.5 325 600 6.32x10"" 1.17 0.9946 3.45xl0" 6 0.0054 6.35xl0"4 1 325 580 6.25 x lO - 4 2.24 0.9892 6.84xl0" b 0.0108 6.32xl0"4 1.5 325 557 6.17xl0- 4 3.46 0.9833 1 .05x10° 0.0167 6.27x10"4 2 325 526 6.11xl0- 4 4.35 0.9773 1.42x10° 0.0227 6.25 x lO - 4 0 350 630 6.04 xlO" 4 0.00 0.9752 1.54x10° 0.0248 6.19x10"" 0.5 350 610 5.94xl0"4 1.64 0.9582 2.59x10° 0.0418 6.20x10"" 1 350 591 5.87xl0"4 2.81 0.9434 3.52x10° 0.0566 6.22x10"" 1.5 350 577 5.65xl0"4 6.47 0.9275 4.42x10° 0.0725 6.09x10"" 2 350 552 5.65xl0"4 6.42 0.9113 5.50x10° 0.0887 6.20x10"" Average 6.25x10"" Recovery after 5 hr reaction = 97.0 mol % 221 Experiment M-2 Reaction: Hydrogenation of naphthalene Reactant Catalyst Input gas Reaction temperature Stirrer speed Solvent Concentration Initial reactor pressure Heating rate Reaction period Naphthalene (10 wt%) exMoS 2-D 5%H 2S/95%H 2 350°C 1200 rpm hexadecane (100ml) 600 ppm Mo 2.8 MPa 10°C/min 5hr Product liquid analysis: Time, hr Pressure, psig Naphthalene Tetralin Total mol/cc mol/cc conversion, % mol ratio mol/cc mol ratio 0 630 5.84xl0 - 4 0.00 0.9964 2.09 xl0" b 0.0036 5.86 xlO" 4 0.5 611 5.60xl0"4 4.05 0.9715 1.64 xlO" 5 0.0285 5.77 xlO" 4 1 589 5.55xl0"4 4.85 0.9445 3.26 xlO" 5 0.0555 5.88 xlO" 4 1.5 570 5.17xl0"4 11.44 0.9189 4.56 xlO" 5 0.0811 5.63 xlO" 4 2 554 . 5.08xl0"4 13.04 0.8918 6.16 xlO" 5 0.1082 5.69 xlO" 4 3 533 4.87xl0" 4 16.60 0.8458 8.88 xlO" 5 0.1542 5.76 xlO" 4 4 510 4.61X10"4 21.01 0.8059 1.11 xlO" 4 0.1941 5.72 xlO" 4 5 485 4.26xl0" 4 27.02 0.7713 1.26 xlO" 4 0.2287 5.52 xlO" 4 Average 5.73 x 10 4 Recovery after 5 hr reaction = 94.2 mol% 222 Experiment MS Reaction: Hydrogenation of naphthalene Reactant Catalyst Input gas Reaction temperature Stirrer speed Solvent Concentration Initial reactor pressure Heating rate Reaction period naphthalene (10wt%) MoNaph 5%H 2S/95%H 2 350°C 1200 rpm hexadecane (100ml) 600 ppm Mo 2.8 MPa 10°C/min 5hr Product liquid analysis: Time, hr Pressure, psig Naphthalene Tetralin Total mol/cc mol/cc conversion, % mol ratio mol/cc mol ratio 0 675 5.74 xlO" 4 0.00 0.9965 1.99 xlO" 6 0.0035 5.76 xlO" 4 0.5 656 5.58 xlO" 4 2.77 0.9706 1.69 xlO"' 0.0294 5.75 xlO" 4 1 633 5.38 xlO" 4 6.28 0.9377 3.58 xlO" 5 0.0623 5.74 xlO" 4 1.5 608 5.18 xlO" 4 9.90 0.9022 5.61 xlO" 5 0.0978 5.74 xlO" 4 2 582 5.00 xlO" 4 12.94 0.8635 7.91 xlO" 5 0.1365 5.79 xlO" 4 3 542 4.47 xlO" 4 22.14 0.7853 1.22 xlO" 4 0.2147 5.69 xlO" 4 4 520 4.12 xlO" 4 28.34 0.7157 1.63 xlO" 4 0.2843 5.75 xlO" 4 5 474 3.69 xlO" 4 35.76 0.6520 1.97 xlO" 4 0.3480 5.66 xlO" 4 Average 5.74 xlO" 4 Recovery after 5 hr reaction = 98.3 mol% 223 Experiment M-4 Reaction: Hydrogenation of naphthalene Reactant Catalyst Input gas Reaction temperature Stirrer speed Solvent Mo concentration Initial reactor pressure Heating rate Reaction period naphthalene (10wt%) 5%H 2S/95%H 2 375°C 1200 rpm hexadecane (100ml) 2.8 MPa 10°C/min 5hr Product liquid analysis: Time, hr Pressure, psig Naphthalene Tetralin, mol/cc Decalin, mol/cc Total mol/cc mol/cc conversion, % 0 611 6.27 xlO" 4 0.00 2.76 x 106 1.18 xlO"' 6.30 x lO - 4 0.5 593 6.04 xlO" 4 3.64 2.87 x l O 0 1.44 xlO" 7 6.33 xlO" 4 1 572 5.77 xlO" 4 7.92 5.76 xlO" 5 1.72 xlO" 7 6.35 xlO" 4 1.5 552 5.39 x10 4 14.04 7.99 xlO" 5 2.95 xlO"' 6.19 x lO - 4 2 523 5.17 xlO" 4 17.58 1.04 x lO ' 4 4.16 x 107 6.21 xlO" 4 3 499 4.70 x lO - 4 25.06 1.48 xlO" 4 3.63 x 107 6.18 xlO" 4 4 472 4.21 xlO" 4 32.82 1.85 x 104 5.37 xlO" 7 6.06 xlO" 4 5 453 3.81 xlO" 4 39.17 2.10 xlO" 4 8.14 x 107 5.92 xlO" 4 Average 6.19 xlO" 4 Recovery after 5 hr reaction = 94.0 mol% 224 Experiment M-5 Reaction: Hydrogenation of naphthalene Reactant Catalyst Input gas Reaction temperature Stirrer speed Solvent Mo concentration Initial reactor pressure Heating rate Reaction period naphthalene (10wt%) exMoS 2-D 5%H 2S/95%H 2 325°C 1200 rpm hexadecane (100ml) 600 ppm 2.8 MPa 10°C/min 5hr Product liquid analysis: Time, Pressure, Naphthalene Tetralin, Total mol/cc hr psig mol/cc conversion, % mol/cc 0 630 5.70 x 10 4 0.00 8.50 xlO" 7 5.71 xlO" 4 0.5 607 5.69 xlO" 4 0.22 6.13 x 106 5.75 xlO" 4 1 579 5.58 xlO" 4 2.05 1.17 xlO" 5 5.70 xlO" 4 1.5 541 5.55 xlO" 4 2.63 1.82 xlO" 5 5.73 xlO" 4 2 514 5.49 xlO" 4 3.74 2.44 xlO" 5 5.73 xlO" 4 3 488 5.33 xlO" 4 6.48 3.72 xlO" 5 5.70 xlO" 4 4 454 5.22 xlO" 4 8.39 4.98 xlO" 5 5.72 xlO" 4 5 412 5.00 xlO" 4 12.24 5.96 xlO" 5 5.60 xlO" 4 Average 5.71 xlO" 4 Recovery after 5 hr reaction = 98.1 mol% 225 v Experiment M-6 Reaction: Hydrogenation of Reactant Catalyst Input gas Reaction temperature Stirrer speed Solvent Mo concentration Initial reactor pressure Heating rate Reaction period iphthalene : naphthalene (10 wt%) : exMoS 2-D : 5%H 2S/95%H 2 :375°C : 1200 rpm : hexadecane (100ml) : 600 ppm : 2.8 MPa : 10°C/min : 5hr Product liquid analysis: Time, hr Pressure, psig . Naphthalene Tetralin, mol/cc Total mol/cc xlO4 mol/cc xlO4 conversion, % 0 640 5.61 0.00 2.91 xl0-b 5.64 0.5 614 5.38 4.09 2.38 xlO-5 5.62 1 553 5.15 8.17 4.38 xlO"5 5.59 1.5 495 4.87 13.08 6.24 x 105 5.50 2 440 4.56 18.74 8.02 xlO"' 5.36 3 407 4.08 27.21 1.15 xlO"4 5.24 4 . 399 3.86 31.15 1.37 xlO"4 5.23 5 393 3.69 34.27 1.62 xlO"4 5.27 Average 5.43 Recovery after 5 hr reaction = 93.4 mol% 226 Experiment M-7 Reaction: Hydrogenation of naphthalene Reactant Catalyst Input gas Reaction temperature Stirrer speed Solvent Mo concentration Initial reactor pressure Heating rate Reaction period naphthalene (10 wt%) exMoS 2-W 5%H 2S/95%H 2 350°C 1200 rpm hexadecane (100ml) 600 ppm 2.8 MPa 10°C/min 5hr Product liquid analysis: Time, hr Pressure, psig Naphthalene Tetralin, mol/cc Total mol/cc xlO 4 mol/cc x lO 4 conversion, % 0 937 5.75 0.00 5.78 xlO" 8 5.75 0.5 944 5.75 0.04 8.85 xlO" 7 5.76 1 938 5.72 0.58 1.83 xlO" 6 5.74 1.5 913 5.71 0.78 2.92 xlO" 6 5.74 2 905 5.68 1.18 4.01 xlO" 6 5.72 3 886 5.63 2.13 6.50 xlO" 6 5.69 4 865 5.59 2.76 9.11 xlO" 6 5.68 5 844 5.55 3.45 1.18 xlO" 5 5.67 Average 5.72 Recovery after 5 hr reaction = 98.6 mol % 227 Experiment M-8 Reaction: Hydrogenation of naphthalene Reactant Catalyst Input gas Reaction temperature Stirrer speed Solvent Mo concentration Initial reactor pressure Heating rate Reaction period naphthalene (10wt%) MoNaph ( K & K Lab. Inc.) 5%H 2S/95%H 2 325°C 1200 rpm hexadecane (100ml) 600 ppm 2.8 MPa 10°C/min 5hr Product liquid analysis: Time, hr Pressure, psig Naphthalene Tetralin, mol/cc Total mole, mol/cc x lO 4 mol/cc x lO 4 conversion, % 0 570 5.60 0.00 1.28 xlO" 6 5.61 0.5 549 5.32 4.97 2.09xl0" 5 5.53 1 503 5.04 9.96 5.86xl0"5 5.63 1.5 490 4.83 13.64 6.30x10° 5.46 2 468 4.56 18.62 7.97x10° 5.35 3 446 4.22 24.64 1.19xl0"4 5.41 4 416 3.85 31.30 1.59xl0"4 5.44 5 406 3.42 38.93 1.93x10"" 5.35 Average 5.47 Recovery after 5 hr reaction = 95.4 mol% 228 Experiment M-9 Reaction: Hydrogenation of naphthalene Reactant Catalyst Input gas Reaction temperature Stirrer speed Solvent Mo concentration Initial reactor pressure Heating rate Reaction period naphthalene (10wt%) MoNaph ( K & K Lab. Inc.) 5%H 2S/95%H 2 350°C 1200 rpm hexadecane (100ml) 600 ppm 2.8 MPa 10°C/min 5hr Product liquid analysis: Time, hr Pressure, psig Naphthalene Tetralin, mol/cc Decalin, mol/cc Total mol/cc xlO 4 mol/cc x 104 conversion, % 0 592 6.12 0.00 7.00 xlO" 6 0 6.19 0.5 524 5.41 11.55 9.34 xlO" 5 1.74 xlO"' 6.34 1 504 4.77 21.97 1.45 xlO" 4 3.95 xlO" 7 6.23 1.5 496 4.19 31.44 2.05 xlO" 4 1.36 xlO" 6 6.25 2 500 3.59 41.27 2.50 x 104 1.66 xlO" 6 6.11 3 570 2.60 57.45 3.26 xlO" 4 2.47 xlO" 6 5.89 4 516 2.00 67.25 3.80 xlO" 4 2.68 xlO" 6 5.83 5 459 1.58 74.11 4.16 xlO" 4 4.54 xlO" 6 5.79 Average 6.08 xlO" 4 Recovery after 5 hr reaction = 93.5 mol% 229 Experiment M-10 Reaction: Hydrogenation of naphthalene Reactant Catalyst Input gas Reaction temperature Stirrer speed Solvent Mo concentration Initial reactor pressure Heating rate Reaction period naphthalene (10 wt%) MoNaph ( K & K Lab. Inc.) 5%H 2S/95%H 2 375°C 1200 rpm hexadecane (100ml) 600 ppm 2.8 MPa 10°C/min 5hr Product liquid analysis: Time, hr Pressure, psig Naphthalene Tetralin, mol/cc Decalin, mol/cc Total mol/cc x lO 4 mol/cc x lO 4 conversion, % 0 595 5.49 0.00 1.34 xlO" 5 1.69 xlO"' 5.62 0.5 559 4.77 13.06 8.48 xlO" 5 3.51 xlO" 7 5.62 1 493 4.09 25.42 1.45 xlO" 4 6.60 xlO"' 5.55 1.5 551 3.61 34.15 2.12 xlO" 4 1.72 xl0" b 5.75 2 528 2.93 46.54 2.55 xlO" 4 1.61 xlO" 6 5.50 3 448 2.27 58.66 3.10 xlO" 4 2.98 xlO" 6 5.40 4 468 1.90 65.47 3.45 xlO" 4 4.59 x 10 6 5.39 5 475 1.61 70.64 3.59 xlO" 4 5.79 xlO" 6 5.26 Average 5.51 x 10 4 Recovery after 5 hr reaction = 93.6 mol% 230 Experiment M-ll Reaction: Hydrogenation of Reactant Catalyst Input gas Reaction temperature Stirrer speed Solvent Mo concentration Initial reactor pressure Heating rate Reaction period iphthalene : naphthalene (10 wt%) : exMoS 2-D : 5%H 2S/95%H 2 :375°C : 1200 rpm : hexadecane (100ml) : 600 ppm : 2.8 MPa : 10°C/min : 5hr Product liquid analysis: Time, hr Pressure, psig Naphthalene Tetralin, mol/cc Total mol/cc mol/cc x lO 4 conversion, % 0 607 6.19 0.00 4.83xl0 - 6 6.23 x 10 4 0.5 587 5.85 5.46 3.32x10° 6.18x10"" 1 533 5.54 10.48 6.05x10° 6.14xl0"4 1.5 549 5.28 14.66 8.82x10° 6.16x10"" 2 546 4.95 19.98 l . l l x l O " 4 6.06xl0"4 3 451 4.44 28.15 1.58xl0"4 6.03xl0"4 4 517 4.11 33.62 1.97xl0 - 4 6.08xl0"4 5 500 3.83 38.08 2.32xl0" 4 6.15xl0"4 Average 6.13xl0"4 Recovery after 5 hr reaction = 98.7 mol% 231 Experiment M-12 Reaction: HDS of DBT Reactant: DBT Solvent: hexadecane Catalyst: exMoS 2-D Input gas: 5%H 2S/95%H 2 Reaction temperature: 350°C Stirrer speed: 1200 rpm Heteroatom: 900 ppm S Volume: 100 ml Concentration: 600 ppm Mo Initial reactor pressure: 2.8 MPa Heating rate: 10°C/min Reaction period: 5 hr Product liquid analysis: Time, Pressure, DBT THDBT S compound, mol/cc x lO 5 hr psig mol/cc x lO 5 conversion, % mol ratio mol/cc mol ratio 0 647 2.01 0.00 0.9917 0 0.0000 2.01 0.5 641 1.97 1.81 0.9753 1.71 xlO"' 0.0084 1.99 1 645 1.92 4.18 0.9451 4.26 x lO - 7 0.0210 1.97 1.5 634 1.83 8.98 0.9090 6.73 xlO" 7 0.0335 1.89 2 625 1.75 13.04 0.8755 8.87 xlO" 7 0.0445 1.83 3 613 1.59 20.94 0.8000 1.14 xlO" 6 0.0576 1.70 4 597 1.47 26.71 0.7375 1.27 xlO" 6 0.0635 1.60 5 587 1.30 35.34 0.6693 1.28 xlO" 6 0.0662 1.43 Time, hr BP CHB BCH Total mol/cc xlO 5 mol/cc mol ratio mol/cc mol ratio mol/cc mol ratio 0 1.68 xlO" 7 0.0083 0 0.0000 0 0.0000 2.02 0.5 3.29 xlO" 7 0.0163 0 0.0000 0 0.0000 2.02 1 5.84 xlO" 7 0.0287 1.06x10"' 0.0052 0 0.0000 2.03 1.5 8.82 xlO" 7 0.0439 2.74x10"' 0.0136 0 0.0000 2.01 2 1.14 xlO" 6 0.0572 4.54x10"' 0.0228 0 0.0000 1.99 3 1.69 x 10 6 0.0852 9.53x10"' 0.0481 1.82 xlO"' 0.0092 1.98 4 2.20 xlO" 6 0.1104 1.48 xlO" 6 0.0741 2.87 xlO" 7 0.0144 1.99 5 2.65 xlO" 6 0.1369 2.04xl0" 6 0.1052 4.35 xlO" 7 0.0224 1.94 Recovery after 5 hr reaction = 96.0 mol% 232 Experiment M-13 Reaction: HDS of DBT Reactant: DBT Solvent: hexadecane Catalyst: MoNaph Input gas: 5%H 2S/95%H 2 Reaction temperature: 350°C Stirrer speed: 1200 rpm Heteroatom: 900 ppm S Volume: 100 ml Concentration: 600 ppm Mo Initial reactor pressure: 2.8 MPa Heating rate: 10°C/min Reaction period: 5 hr Product liquid analysis: Time, hr Pressure, psig DBT THDBT S compound, mol/cc xlO5 mol/cc xlO5 conv., % mol ratio mol/cc mol ratio 0 667 2.11 0.00 0.9920 0 0.0000 2.11 0.5 634 2.04 3.48 0.9873 0 0.0000 2.04 1 601 2.04 3.31 0.9772 1.82 xlO"7 0.0087 2.06 1.5 567 2.02 4.07 0.9606 4.99 xlO"7 0.0237 2.07 2 538 2.00 5.20 0.9429 8.28 xlO"7 0.0390 2.08 3 494 1.95 7.42 0.9188 1.16 xlO"6 0.0544 2.07 4 447 1.85 12.16 0.8886 1.54 xlO"6 0.0738 2.01 5 396 1.77 15.94 0.8713 1.79 xlO"6 0.0877 1.95 Time, hr BP CHB BCH Total mol/ccxlO5 mol/ccxlO7 mol ratio mol/cc mol ratio mol/cc mol ratio 0 1.70 0.0080 0 0.0000 - - 2.13 0.5 2.61 0.0127 0 0.0000 - - 2.06 1 2.94 0.0141 0 0.0000 - - 2.09 1.5 3.32 0.0157 0 0.0000 - - 2.11 2 3.84 0.0181 0 0.0000 - - 2.12 3 4.88 0.0230 8.35x10-* 0.0039 - - 2.13 4 5.89 0.0282 1.94 xlO-7 0.0093 - - 2.09 5 5.84 0.0287 2.51x 107 0.0123 - - 2.04 Average 2.09 Recovery after 5 hr reaction = 95.8 mol% 233 Experiment M-14 Reaction: H D S of D B T Reactant: D B T Solvent: hexadecane Catalyst: e x M o S 2 - W Input gas: 5 % H 2 S / 9 5 % H 2 Reaction temperature: 350°C Stirrer speed: 1200 rpm Heteroatom: 900 ppm S Volume: 100ml Concentration: 600 ppm M o Initial reactor pressure: 2.8 M P a Heating rate: 10°C/min Reaction period: 5 hr Product liquid analysis: Time, Pressure, DBT THDBT S compound, hr psig mol/ccxlO5 conv., % mol ratio mol/cc mol ratio mol/cc x 105 0 923 2.00 0.00 0.9923 - - 2.00 0.5 893 1.99 0.24 0.9918 - - 1.99 1 893 2.00 -0.44 0.9917 - - 2.00 1.5 865 1.99 0.26 0.9904 - - 1.99 2 849 1.99 0.49 0.9900 - - 1.99 3 840 1.97 1.31 0.9859 - - 1.97 4 805 1.97 1.38 0.9832 - - 1.97 5 790 1.96 1.66 0.9808 - - 1.96 Time, BP CHB BCH Total hr mol/ccxlO7 mol ratio mol/cc mol ratio mol/cc mol ratio mol/ccxlO5 0 1.55 0.0077 0 0.0000 - - 2.01 0.5 1.64 0.0082 0 0.0000 - - 2.01 1 1.68 0.0083 0 0.0000 - - 2.02 1.5 1.94 0.0096 0 0.0000 - - 2.01 2 2.01 0.0100 0 0.0000 - - 2.01 3 2.37 0.0119 4.40 x IO"8 0.0022 - - 2.00 4 2.95 0.0147 4.1 lx IO"8 0.0021 - - 2.00 5 3.30 0.0165 5.40 x IO"8 0.0027 - - 2.00 Average 2.01 Recovery after 5 hr reaction = 99.5 mol% 234 Experiment M-15 Reaction: HDS of DBT Reactant: DBT Solvent: hexadecane Catalyst: MoNaph + water Input gas: 5%H 2S/95%H 2 Reaction temperature: 350°C Stirrer speed: 1200 rpm Heteroatom: 900 ppm S Volume: 100ml Concentration: 600 ppm Mo Initial reactor pressure: 2.8 MPa Heating rate: 10°C/min Reaction period: 5 hr Product liquid analysis: Time, nr Pressure, psig DBT THDBT S compound, mol/cc x lO 5 mol/cc xlO 5 con v., % mol ratio mol/cc mol ratio 0 665 1.99 0.00 0.9915 0 0.0000 1.99 0.5 671 1.95 1.59 0.9861 0 0.0000 1.95 1 674 1.96 1.19 0.9828 0 0.0000 1.96 1.5 658 1.93 2.55 0.9794 0 0.0000 1.93 2 656 1.94 2.46 0.9757 0 0.0000 1.94 3 614 1.92 3.33 0.9482 2.84 xlO" 7 0.0140 1.95 4 585 1.74 12.50 0.8852 1.13 xlO"" 0.0577 1.85 5 501 1.72 13.29 0.8586 1.34 xl '0 ' 6 0.0666 1.86 Time, hr BP CHB BCH Total mol/cc xlO 5 mol/ccxlO7 mol ratio mol/cc mol ratio mol/cc mol ratio 0 1.70 0.0085 0 0.0000 - - 2.00 0.5 2.76 0.0139 0 0.0000 - - 1.98 1 3.44 0.0172 0 0.0000 - - 2.00 1.5 4.06 0.0206 0 0.0000 - - 1.98 2 4.82 0.0243 0 0.0000 - - 1.98 3 6.69 0.0331 9.54xl0"8 0.0047 - - 2.02 4 9.04 0.0460 2.16X10-7 0.0110 - - 1.96 5 11.2 0.0560 3.78xl0 _ / 0.0188 - - 2.01 Average 1.99 Recovery after 5 hr reaction -100% 235 Experiment M-l 6 Reaction: HDS of DBT Reactant: DBT Solvent: hexadecane Catalyst: exMoS 2-D Input gas: 5%H 2S/95%H 2 Reaction temperature: 325°C Stirrer speed: 1200 rpm Heteroatom: 900 ppm S Volume: 100ml Concentration: 600 ppm Mo Initial reactor pressure: 2.8 MPa Heating rate: 10°C/min Reaction period: 5 hr Product liquid analysis: Time, hr Pressure, psig DBT THDBT S compound, mol/ccxlO5 mol/cc xlO5 conv., % mol ratio mol/cc xlO7 mol ratio 0 603 2.10 0.00 1.0000 0.94 0.0000 2.10 0.5 570 2.08 1.17 0.9970 0.95 0.0000 2.08 1 541 2.08 1.07 0.9930 1.42 0.0023 2.09 1.5 513 2.09 0.48 0.9885 1.93 0.0047 2.10 2 484 2.10 0.43 0.9777 3.46 0.0118 2.12 3 452 2.02 3.95 0.9611 4.81 0.0184 2.06 4 428 2.03 3.38 0.9523 5.38 0.0208 2.08 5 400 1.97 6.49 0.9375 6.37 0.0259 2.02 Time, hr BP CHB BCH Total mol/cc xlO5 mol/cc xlO7 mol ratio mol/cc mol ratio mol/cc mol ratio 0 1.51 0.0000 0 0.0000 - - 2.10 0.5 2.13 0.0030 0 0.0000 - - 2.09 1 2.50 0.0047 0 0.0000 - - 2.10 1.5 2.95 0.0068 0 0.0000 - - 2.12 2 3.78 0.0106 0 0.0000 - - 2.14 3 5.11 0.0171 7.17 xlO"8 0.0034 - - 2.10 4 6.18 0.0219 1.07 x 10 ' 0.0050 - - 2.14 5 7.59 0.0289 1.62 x 10"7 0.0077 - - 2.10 Average 2.01 Recovery after 5 hr reaction -100% 236 Experiment M-17 Reaction: HDS of DBT Reactant: DBT Solvent: hexadecane Catalyst: exMoS 2-D Input gas: 5%H 2S/95%H 2 Reaction temperature: 375°C Stirrer speed: 1200 rpm Heteroatom: 900 ppm S Volume: 100 ml Concentration: 600 ppm Mo Initial reactor pressure: 2.8 MPa Heating rate: 10°C/min Reaction period: 5 hr Product liquid analysis: Time, hr Pressure, psig DBT THDBT S compound, mol/cc x 105 mol/cc conv., % mol ratio mol/cc mol ratio 0 699 2.13x10° 0.00 0.9841 8.33 xlO" 8 0.0038 2.14 0.5 704 2.04x10° 4.43 0.9291 5.86x10"' 0.0267 2.10 1 700 1.90x10° 11.13 0.8663 8.97x10"' 0.0410 1.99 1.5 686 1.73x10° 18.75 0.8046 1.13xl0"6 0.0523 1.85 2 674 1.58x10° 25.96 0.7387 1.30xl0"6 0.0609 1.71 3 656 1.32x10° 38.01 0.6228 1.29xl0"6 0.0607 1.45 4 642 1.07x10° 49.85 0.5038 1.16xl0"6 0.0546 1.19 5 629 8.43 x 10 6 60.50 0.4053 9.57xl0" 7 0.0460 0.94 Time, hr BP CHB BCH Total mol/cc xlO 5 mol/cc mol ratio mol/cc mol ratio mol/cc mol ratio 0 2.14xl0" 7 0.0098 0.00 0.0000 4.78xl0" 8 0.0022 2.17 0.5 7.60x10"' 0.0346 1.11x10"' 0.0051 9.92xl0" 8 0.0045 2.20 1 1.42 xlO" 6 0.0651 4.60x10"' 0.0210 1.45x10"' 0.0066 2.19 1.5 2.01 x 10 6 0.0933 8.55x10"' 0.0397 2.19x10"' 0.0102 2.16 2 2.71 x 10 6 0.1268 1.31 xlO" 6 0.0615 2.59x10"' 0.0121 2.14 3 3.99 xlO" 6 0.1878 2.30xl0" 6 0.1084 4.31x10"' 0.0203 2.12 4 5.18xl0" 6 0.2439 3.39xl0"6 0.1595 8.11x10"' 0.0382 2.12 5 6.07x10"6 0.2917 4.31 xlO" 6 0.2073 1.03 xlO" 6 0.0497 2.08 Average 2.15 Recovery after 5 hr reaction = 95.9 mol% 237 Experiment M-18 Reaction: HDS of DBT Reactant: DBT Solvent: hexadecane Catalyst: MoNaph Input gas: 5%H 2S/95%H 2 Reaction temperature: 325 °C Stirrer speed: 1200 rpm Heteroatom: 900 ppm S Volume: 100ml Concentration: 600 ppm Mo Initial reactor pressure: 2.8 MPa Heating rate: 10°C/min Reaction period: 5 hr Product liquid analysis: Time, hr Pressure psig DBT THDBT S compound, mol/cc xlO 5 mol/cc xlO 5 conv., % mol ratio mol/cc mol ratio 0 596 2.13 0 0.9947 0 0.0000 2.13 0.5 576 2.10 1.52 0.9914 0 0.0000 2.10 1 551 2.09 1.92 0.9874 6.07xl0 - 8 0.0029 2.10 1.5 539 2.10 1.45 0.9837 1.29x10"' 0.0060 2.11 2 514 2.10 1.62 0.9759 2.63x10-' 0.0122 2.12 3 488 2.05 3.99 0.9702 3.37xl0" 7 0.0160 2.08 4 453 2.03 4.54 0.9596 5.02x10"' 0.0237 2.09 5 430 2.01 5.55 0.9590 4.84x10"' 0.0230 2.06 Time, hr BP CHB BCH Total mol/cc xlO 5 mol/cc x lO 7 mol ratio mol/cc mol ratio mol/cc mol ratio 0 1.14 0.0053 - - - - 2.14 0.5 1.82 0.0086 - - - - 2.12 1 2.06 0.0097 - - - - 2.12 1.5 2.20 0.0103 - - - - 2.14 2 2.54 0.0118 - - - - 2.15 3 2.91 0.0138 - - - - 2.11 4 3.55 0.0167 - - - - 2.12 5 3.77 0.0180 - - - - 2.10 Average 2.12 Recovery after 5 hr reaction = 98.1 mol% 238 Experiment M-19 Reaction: HDS of DBT Reactant: DBT Solvent: hexadecane Catalyst: MoNaph Input gas: 5%H 2S/95%H 2 . Reaction temperature: 375 °C Stirrer speed: 1200 rpm Heteroatom: 900 ppm S Volume: 100 ml Concentration: 600 ppm Mo Initial reactor pressure: 2.8 MPa Heating rate: 10°C/min Reaction period: 5 hr Product liquid analysis: Time, hr Pressure psig DBT THDBT S compound, mol/cc xlO 5 mol/cc xlO 5 conv., % mol ratio mol/cc mol ratio 0 663 2.12 0 0.9911 0 0.0000 2.12 0.5 636 1.98 6.80 0.9550 4.96x10"' 0.0240 2.03 1 614 1.92 9.36 0.9082 9.22xl0 - 7 0.0435 2.02 1.5 575 1.90 10.39 0.8677 1.29xl0"b 0.0589 2.03 2 551 1.76 16.95 0.8243 1.52xl0-6 0.0710 1.91 3 497 1.56 26.55 0.7381 1.70xl0-6 0.0806 1.73 4 456 1.37 35.28 0.6586 1.64x10"" 0.0788 1.54 5 419 1.21 43.12 0.5886 1.34xl0"6 0.0651 1.34 Time, hr BP CHB BCH Total mol/cc xlO 5 mol/cc mol ratio mol/cc mol ratio mol/cc mol ratio 0 1.91xl0"7 0.0089 0 0.0000 0 0.0000 2.14 0.5 4.35x10"' 0.0210 0 0.0000 0 0.0000 2.07 1 8.74x10"' 0.0413 1.47x10"' 0.0069 0 0.0000 2.12 1.5 1.16xl0"6 0.0528 3.73x10"' 0.0170 7.83xl0"8 0.0036 2.19 2 1.52xl0"6 0.0711 6.41x10"' 0.0300 7.68x10"" 0.0036 2.14 3 2.33xl0" 6 0.1101 1.29xl0"6 0.0609 2.18x10"' 0.0103 2.11 4 3.09x10"" 0.1484 2.07x10"" 0.0992 3.14x10"' 0.0150 2.09 5 3.89x10"" 0.1897 2.71x10"" 0.1321 5.03x10"' 0.0245 2.05 Average 2.11 Recovery after 5 hr reaction = 95.8 mol% 239 Experiment M-20 Reaction: HDS of DBT Reactant: DBT Solvent: hexadecane Catalyst: crystalline M0S2 Input gas: 5%H 2S/95%H 2 Reaction temperature: 350°C Stirrer speed: 1200 rpm Heteroatom: 900 ppm S Volume: 100ml Concentration: 600 ppm Mo Initial reactor pressure: 2.8 MPa Heating rate: 10°C/min Reaction period: 5 hr Product liquid analysis: Time, hr Pressure psig DBT THDBT S compound, mol/cc x l o5 mol/cc x lO 5 conv., % mol ratio mol/cc mol ratio 0 683 2.09 0.00 0.9806 2.79* 10-'' 0.0131 2.12 0.5 655 2.08 0.61 0.9694 3.96x10-' 0.0185 2.12 1 629 2.06 1.26 0.9571 5.40xl0"7 0.0251 2.12 1.5 591 2.03 2.83 0.9459 5.70x10-' 0.0266 2.09 2 567 1.99 4.58 0.9333 6.79x10"' 0.0318 2.06 3 537 1.95 6.56 0.9076 8.33xl0"7 0.0387 2.03 4 476 1.89 9.40 0.8812 9.66x10-' 0.0450 1.99 5 439 1.83 12.57 0.8573 1.05 x 10 6 0.0494 1.93 Time, hr BP CHB BCH Total mol/cc xlO 5 mol/cc x lO 7 mol ratio mol/cc mol ratio mol/cc mol ratio 0 1.34 0.0063 0 0.0000 - - 2.13 0.5 2.58 0.0120 0 0.0000 - - 2.14 1 3.84 0.0178 0 0.0000 - - 2.15 1.5 5.05 0.0235 8.60xl0"8 0.0040 - - 2.15 2 6.51 0.0305 9.56xl0"8 0.0045 - - 2.14 3 9.25 0.0430 2.29x10"' 0.0107 - - 2.15 4 11.7 0.0547 4.12x10"' 0.0192 - - 2.15 5 14.1 0.0664 5.73 xlO" 7 0.0269 - - 2.13 Average 2.14 Recovery after 5 hr reaction -100 mol% 240 Experiment M-21 Reaction: HDS of DBT Reactant: DBT Solvent: hexadecane Catalyst: exMoS 2-D Input gas: 5%H 2S/95%H 2 Reaction temperature: 350°C Stirrer speed: 1200 rpm Heteroatom: 900ppm S Volume: 100ml Concentration: 600 ppm Mo Initial reactor pressure: 2.8 MPa Heating rate: 10°C/min Reaction period: 5 hr Product liquid analysis: Time, hr Pressure, psig DBT THDBT S compound, mol/cc x lO 5 mol/cc xlO 5 conv., % mol ratio mol/cc mol ratio 0 676 2.11 0.00 0.9918 0 0.0000 2.11 0.5 684 2.07 2.04 0.9662 3.78x10"' 0.0176 2.11 1 672 1.99 5.77 0.9405 5.57x10"' 0.0263 2.05 1.5 667 1.94 8.38 0.9166 6.80x10"' 0.0322 2.00 2 650 1.87 11.58 0.8916 7.83xl0" 7 0.0373 1.95 3 638 1.73 18.23 0.8351 1.04 xlO" 6 0.0504 1.83 4 628 1.60 24.47 0.7754 1.23 xlO" 6 0.0599 1.72 5 619 1.46 31.00 0.7115 1.28 xlO" 6 0.0622 1.59 Time, hr BP CHB BCH Total mol/cc xlO 5 mol/cc mol ratio mol/cc mol ratio mol/cc mol ratio 0 1.75x10"' 0.0082 0 0.0000 0 0.0000 2.13 0.5 3.46x10"' 0.0161 0 0.0000 0 0.0000 2.14 1 6.10x10"' 0.0288 9.48xl0"8 0.0045 0 0.0000 2.12 1.5 9.02x10"' 0.0427 1.81x10"' 0.0086 0 0.0000 2.11 2 1.16xl0"6 0.0555 3.27x10"' 0.0156 1.00x10"' 0.0048 2.10 3 1.67 xlO" 6 0.0805 7.05x10"' 0.0341 1.86x10"' 0.0090 2.07 4 2.17xl0" 6 0.1053 1.22 xlO" 6 0.0594 3.33x10"' 0.0162 2.06 5 2.78xl0" 6 0.1355 1.86xl0"6 0.0908 4.59x10"' 0.0224 2.05 Average 2.10 Recovery after 5 hr reaction = 96.2 mol% . 241 Experiment M-22 Reaction: HDS of DBT Reactant: DBT Solvent: hexadecane Catalyst: A H M Input gas: 5%H 2S/95%H 2 Reaction temperature: 350°C Stirrer speed: 1200 rpm Heteroatom: 900 ppm S Volume: 100 ml Concentration: 600 ppm Mo Initial reactor pressure: 2.8 MPa Heating rate: 10°C/min Reaction period: 5 hr Product liquid analysis: Time, hr Pressure, psig DBT THDBT S compound, mol/cc mol/cc conv., % mol ratio mol/cc mol ratio 0 697 2.09*10-' 0 0.9789 0 0.0000 2.09x10"' 0.5 673 1.45x10"' 30.48 0.7140 2.51xl0" b 0.1232 1.70x10° 1 656 1.01x10"' 51.78 0.5034 2.48xl0" 6 0.1239 1.26x10° 1.5 637 6.79xl0" 6 67.51 0.3409 2.04xl0" 6 0.1023 8.83xl0"b •2 614 4.18xl0" 6 80.00 0.2149 1.26xl0"b 0.0646 5.44xl0' 6 2.5 582 2.60x10"" 87.57 0.1355 8.09xl0"7 0.0422 3.41 xlO" 6 3 562 1.36X10' 6 93.51 0.0703 4.30x10"' 0.0223 1.79X10"6 4 532 3.88xl0" 7 98.15 0.0208 1.71 xlO" 7 0.0092 5.59x10"' 5 512 1.13x10"' 99.46 0.0060 1.09 xlO" 7 0.0059 2.22x10"' Time, BP CHB BCH Total mol/cc hr mol/cc mol ratio mol/cc x lO 6 mol ratio mol/cc mol ratio xlO 5 0 2.38xl0" 7 0.0112 0 0.0000 0 0.0099 2.13 0.5 1.41xl0"b 0.0694 1.29 0.0633 4.01 xlO" 7 0.0301 2.03 1 2.47xl0" 6 0.1236 3.33 0.1666 1.44 xlO" 6 0.0826 2.00 1.5 3.24 xlO" 6 0.1628 5.08 0.2552 2.55xl0- 6 0.1388 1.99 2 3.83xl0" 6 0.1969 6.61 0.3400 3.36xl0" 6 0.1835 1.94 2.5 4.23 xlO" 6 0.2207 7.57 0.3953 3.74xl0- 6 0.2062 1.92 3 4.46xl0" 6 0.2307 8.14 0.4217 4.71 xlO" 6 0.2550 1.93 4 4.44x10"6 0.2382 8.33 0.4469 5.10xl0"6 0.2849 1.86 5 4.10xl0" 6 0.2201 8.00 0.4290 . 6.11X10'6 0.3389 1.86 Average 1.98 Recovery after 5 hr reaction = 87.3 mol% 242 Experiment M-23 Reaction: Hydrodenitrogenation of quinoline Reactant: quinoline Solvent: hexadecane Catalyst: exMoS 2-D Input gas: 5%H 2S/95%H 2 Reaction temperature: 350°C Stirrer speed: 1200 rpm Heteroatom: 200ppm N Volume: 100ml Concentration: 600 ppm Mo Initial reactor pressure: 2.8 MPa Heating rate: 10°C/min Reaction period: 5 hr Product liquid analysis: Time, hr Pressure, psig Quinoline THQ1 THQ5 mol/cc conv., % mol ratio mol/cc x lO 7 mol ratio mol/cc xlO 7 mol ratio 0 640 3.52xl0" 6 0.00 0.3257 1.37 0.6615 1.37 0.0127 0.5 653 2.47xl0" 6 29.90 0.2426 4.19 0.6814 4.19 0.0412 1 646 1.82xl0"6 48.25 0.1896 5.70 0.6651 5.70 0.0594 1.5 638 1.53xl0"6 56.46 0.1676 6.61 0.6289 6.61 0.0724 2 627 1.42 xlO" 6 59.72 0.1623 6.85 0.5808 6.85 0.0785 3 615 1.23 xlO" 6 64.90 0.1517 7.12 0.4840 7.12 " 0.0875 4 604 9.42 xlO" 7 73.22 0.1273 6.21 0.4054 6.21 0.0839 5 593 8.25x10"' 76.53 0.1206 5.30 0.3049 5.30 0.0775 Time, hr o-propylaniline PCHe P C H Total mol/cc mol/cc mol ratio mol/cc mol ratio mol/cc xlO7 mol ratio 0 0 0.0000 0 0.0000 0 0.0000 l.osxio"5 0.5 3.54x10"' 0.0349 0 0.0000 0 0.0000 1.02X10"3 1 5.33x10"' 0.0555 1.67x10"' 0.0174 1.25 0.0131 9.60xl0"6 1.5 6.95x10"' 0.0760 2.98x10"' 0.0326 2.06 0.0226 9.14xl0"6 2 8.49x10"' 0.0973 4.28x10"' 0.0490 2.80 0.0321 8.73xl0"6 3 1.12x10"" 0.1381 6.46x10"' 0.0794 4.83 0.0594 8.14xl0"6 4 1.31 xlO" 6 0.1767 8.27xl0"7 0.1117 7.04 0.0951 7.40xl0"6 5 1.47 xlO" 6 0.2152 l .OlxlO" 6 0.1469 9.23 0.1349 6.84x10'" Average 8.85x10"" Recovery after 5 hr reaction = 63.3 mol% 243 Experiment M-24 Reaction: H D N of quinoline Reactant: Quinoline Solvent: hexadecane Catalyst: MoNaph Input gas: 5%H 2S/95%H 2 Reaction temperature: 350°C Stirrer speed: 1200 rpm Heteroatom: 200 ppm N Volume: 100 ml Concentration: 600 ppm Mo Initial reactor pressure: 2.8 MPa Heating rate: 10°C/min Reaction period: 5 hr Product liquid analysis: Time, hr Pressure, psig Quinoline THQ1 THQ5 mol/cc conv., % mol/cc mol/cc mol ratio mol/cc mol ratio 0 695 2.63 xlO" 6 0.00 0.2735 6.82xl0"6 0.7086 1.73xl0"7 0.0000 0.5 676 2.43xl0" 6 7.77 0.2840 4.62xl0" 6 0.5399 5.73x10"' 0.0671 1 658 1.82xl0' 6 30.71 0.2352 3.58xl0"6 0.4618 6.42xl0"7 0.0828 1.5 639 1.25 xlO" 6 52.57 0.1728 2.88xl0" 6 0.3984 6.09x10"' 0.0843 2 619 1.03 xlO" 6 60.85 0.1523 2.12xl0" 6 0.3139 5.64x10"' 0.0834 3 591 6.67x10"' 74.67 0.1137 7.89x10"' 0.1344 3.54x10"' 0.0603 4 562 4.12xl0" 7 84.35 0.0776 9.83xl0" 8 0.0185 1.73x10"' 0.0326 5 537 0 100.00 0.0000 0 0 0 0.0000 Time, hr opropylaniline PCHe PCH Total mol/cc xlO 6 mol/cc mol ratio mol/cc mol ratio mol/cc mol ratio 0 0 0.0000 0 0.0000 0 0.0000 9.63 0.5 6.50x10"' 0.0761 2.82xl0" 7 0.0330 0 0.0000 8.55 1 9.68x10"' 0.1248 4.68x10"' 0.0603 2.73x10"' 0.0351 7.76 1.5 1.28xl0"6 0.1772 7.29x10"'' 0.1009 4.80x10"' 0.0664 7.23 2 1.41xl0"b 0.2088 8.89x10"' 0.1313 7.46xl0" 7 0.1103 6.77 3 1.65xl0"6 0.2814 0 0.0000 2.41 xlO" 6 0.4102 5.87 4 1.69 xlO" 6 0.3172 0 0.0000 2.94xl0" 6 0.5541 5.31 5 1.50xl0"6 0.3108 0 0.0000 3.32xl0" 6 0.6892 4.81 Average 6.99 Recovery after 5 hr reaction = 49.9 mol% 244 Experiment M-25 Reaction: Hydrodenitrogenation of quinoline Reactant: quinoline Solvent: hexadecane Catalyst: exMoS 2 -W Input gas: 5%H 2S/95%H 2 Reaction temperature: 350°C Stirrer speed: 1200 rpm Heteroatom: 200 ppm N Volume: 100ml Concentration: 600 ppm Mo Initial reactor pressure: 2.8 MPa Heating rate: 10°C/min Reaction period: 5 hr Product liquid analysis: Time, hr Pressure, psig Quinoline THQ1 THQ5 mol/cc conv., % mol/cc mol/cc x lO 6 mol ratio mol/cc mol ratio 0 960 7.09x10"" 0.00 0.6358 4.06 0.3642 0 0.0000 0.5 943 3.84xl0- 6 45.89 0.3685 6.34 0.6086 2.39x10"' 0.0230 1 957 3.07xl0" 6 56.77 0.3100 6.31 0.6379 4.58x10"'' 0.0463 1.5 923 3.04xl0" 6 57.11 0.3145 5.85 0.6051 5.95 xlO" 7 0.0615 2 913 3.01xl0" 6 57.62 0.3152 5.46 0.5721 8.37x10"' 0.0878 3 913 2.77xl0" 6 61.02 0.3108 4.62 0.5187 1.06xl0"6 0.1194 4 891 2.45xl0" 6 65.49 0.2942 4.08 0.4898 1.13xl0"6 0.1358 5 894 2.23xl0- 6 68.55 0.2733 3.64 0.4457 1.45 xlO" 6 0.1773 Time, hr o-propylaniline PCHe PCH Total mol/cc mol/cc mol ratio mol/cc mol ratio mol/cc mol ratio 0 0 0.0000 0 0.0000 0 0.0000 1.12xl0"5 0.5 0 0.0000 0 0.0000 0 0.0000 1.04 xlO"' 1 5.71xl0"8 0.0058 0 0.0000 0 0.0000 9.89xl0"6 . 1.5 1.83x10"' 0.0189 0 0.0000 0 0.0000 9.67xl0" 6 ' 2 2.38x10"' 0.0250 0 0.0000 0 0.0000 9.54xl0"6 3 3.20 xlO" 7 0.0360 1.35x10"' 0.0151 0 0.0000 8.90xl0"6 4 4.02x10"' 0.0483 1.66x10"' 0.0200 9.94xl0" 8 0.0119 8.32xl0"6 5 4.77x10"' 0.0584 2.58x10"' 0.0316 1.11x10"' 0.0136 8.16xl0"6 Average 9.51 xlO" 6 Recovery after 5 hr reaction = 72.9 mol% 245 Experiment M-26 Reaction: H D N of quinoline Reactant: Quinoline Solvent: hexadecane Catalyst: MoNaph + water Input gas: 5%H 2S/95%H 2 Reaction temperature: 350°C Stirrer speed: 1200 rpm Heteroatom: 200 ppm N Volume: 100 ml Concentration: 600 ppm Mo Initial reactor pressure: 2.8 MPa Heating rate: 10°C/min Reaction period: 5 hr Product liquid analysis: Time, hr Pressure, psig Quinoline THQ1 THQ5 mol/cc conv., % mol/cc mol/cc mol ratio mol/cc mol ratio 0 937 4.87x10-" 0.00 0.4874 4.95x10"" 0.4953 1.72x10"' 0.0172 0.5 944 3.81xl0"6 21.43 0.3830 4.32x10"" 0.4341 1.28x10"" 0.1289 ' 1 938 2.84xl0"b 37.65 0.3039 3.73x10"" 0.3989 1.83x10"" 0.1961 1.5 913 2.39xl0"6 45.51 0.2656 3.12x10'" 0.3471 2.06x10"" 0.2287 2 905 1.89x10-" 53.69 0.2257 2.62x10"" 0.3133 1.98x10"" 0.2361 3 886 1.46x10-" 59.30 0.1984 1.49x10"" 0.2031 1.76x10"" 0.2388 4 865 9.74xl0"7 69.15 0.1504 9.50x10"' 0.1468 1.32x10"" 0.2037 5 844 5.48x10"' 78.23 0.1061 4.82xl0"7 0.0934 7.63x10"' 0.1477 Time, hr o-propylaniline PCHe PCH Total mol/cc xlO" mol/cc mol ratio mol/cc x lO 7 mol ratio mol/cc mol ratio 0 0 0.0000 0 0.0000 0 0.0000 9.99 0.5 4.09x10-' 0.0411 1.28 0.0129 0 0.0000 9.95 1 6.30x10"' 0.0674 3.15 0.0337 0 0.0000 9.35 1.5 8.57xl0"7 0.0953 3.74 0.0415 1.96x10"' 0.0218 9.00 2 1.03x10"" 0.1229 5.24 0.0626 3.30x10"' 0.0394 8.37 3 1.25x10"" 0.1705 7.32 0.0995 6.61x10"' 0.0898 7.36 4 1.38x10"" 0.2125 7.79 0.1203 1.08x10"" 0.1663 6.48 5 1.31x10"" 0.2529 7.66 0.1484 1.30x10"" 0.2515 5.16 * Average 8.21 Recovery after 5 hr reaction = 51.7 mol% 246 Experiment M-27 Reaction: H D N of carbazole Reactant: carbazole Solvent: hexadecane Catalyst: exMoS 2-D Input gas: 5%H 2S/95%H 2 Reaction temperature: 350°C Stirrer speed: 1200 rpm Heteroatom: 200 ppm N Volume: 100 ml Concentration: 600 ppm Mo Initial reactor pressure: 2.8 MPa Heating rate: 10°C/min Reaction period: 5 hr Product liquid analysis: Time, hr Pressure psig THCBZ CHB CHCHe mol/cc mol ratio mol/cc mol ratio mol/cc mol ratio 0 648 1.33x10"' 0.0119 0.00 0.0000 0.00 0.0000 0.5 657 7.31x10"' 0.0653 0.00 0.0000 0.00 0.0000 1 643 l .OlxlO" 6 0.0902 0.00 0.0000 0.00 0.0000 1.5 643 1.29 xlO" 6 0.1155 1.02 xlO" 7 0.0091 0.00 0.0000 2 632 1.36xl0"6 0.1215 1.18x10"' 0.0105 0.00 0.0000 3 618 1.45 x 10 6 0.1295 1.68xl0"7 0.0150 0.00 0.0000 4 614 1.47 xlO" 6 0.1315 1.96x10"' 0.0175 9.33xl0" 8 0.0083 5 604 1.45 xlO" 6 0.1300 2.41 xlO" 7 0.0215 1.30xl0"7 0.0116 Time, hr BCH CPMCH HCH Total product mol/cc mol/cc mol ratio mol/cc mol ratio mol/cc mol ratio 0 0.00 0.0000 0.00 0.0000 0.00 0.0000 1.33x10"' 0.5 5.34X10-8 0.0048 0.00 0.0000 0.00 0.0000 7.85x10"' 1 1.02x10"' 0.0091 4.64xl0" 8 0.0041 0.00 0.0000 1.16xl0"6 1.5 1.52x10"' 0.0136 1.23x10"' 0.0110 8.77xl0"8 0.0078 1.76xl0"6 2 1.92x10"' 0.0172 1.96x10"' 0.0176 1.39x10"' 0.0124 2.00xl0" 6 3 2.90x10"' 0.0259 3.54x10"' 0.0317 3.11x10"' 0.0278 2.57xl0" 6 4 3.96x10"' 0.0354 5.41x10"' 0.0484 4.79xl0" 7 0.0428 3.18xl0"6 5 5.14x10"' 0.0459 7.36x10"' 0.0657 6.59x10"' 0.0589 3.73xl0"6 Time, hr Carbazole mol/cc mol ratio Conversion, % 0 1.11x10° 0.9881 0.00 0.5 1.04 xlO" 5 0.9299 5.89 1 l.OOxlO"5 0.8966 9.26 1.5 9.43 xlO" 6 0.8430 14.69 2 9.19xl0" 6 0.8209 16.92 3 8.62xl0"6 0.7701 22.06 4 8.01xl0"6 0.7161 27.53 5 7.46xl0" 6 0.6664 32.56 H D N after 5 hr reaction = 20.4 mol% 247 Experiment M-28 Reaction: H D N of carbazole Reactant: carbazole Solvent: hexadecane Catalyst: exMoS2-W Input gas: 5%H2S/95%H2 Reaction temperature: 350°C Stirrer speed: 1200 rpm Heteroatom: 200 ppm N Volume: 100ml Concentration: 600 ppm Mo Initial reactor pressure: 2.8 MPa Heating rate: 10°C/min Reaction period: 5 hr Product liquid analysis: Time, hr Pressure psig THCBZ CHB BCH mol/cc mol ratio mol/cc mol ratio mol/cc mol ratio 0 884 0.00 0.0000 0.00 0.0000 0.00 0.0000 0.5 875 0.00 0.0000 0.00 0.0000 0.00 0.0000 1 881 1.36x10"' 0.0117 8.10xl0"8 0.0000 8.10xl0"8 0.0070 1.5 870 3.91x10-' 0.0336 8.32xl0"8 0.0000 8.32xl0"8 0.0072 2 838 5.06x10"' 0.0436 6.79xl0"8 0.0000 6.79xl0" s 0.0058 3 817 8.79x10"' 0.0756 1.12x10"' 0.0053 1.12x10"' 0.0097 4 813 1.05 xlO" 6 0.0905 1.09xl0"7 0.0055 1.09 xlO" 7 0.0093 5 798 1.71xl0"6 0.1470 9.84xl0"8 0.0039 9.84xl0" 8 0.0085 Time, hr Total product mol/cc Carbazole mol/cc mol ratio Conversion, % 0 0.00 1.16x10° 1.0000 0.00 0.5 0.00 1.16x10° 1.0000 0.00 1 2.17x10"' 1.14x10° 0.9814 1.86 1.5 4.74xl0" 7 1.11x10"' 0.9592 4.08 2 5.74xl0"7 1.10x10° 0.9506 4.94 3 1.05 xlO" 6 1.06x10° 0.9095 9.05 4 1.22 xlO" 6 1.04x10° 0.8947 10.53 5 1.85 xlO" 6 9.77xl0" 6 0.8406 15.94 H D N after 5 hr reaction = 1.24 % 248 Experiment M-29 Reaction: HDN of carbazole Reactant: carbazole Solvent: hexadecane Catalyst: MoNaph Input gas: 5%H2S/95%H2 Reaction temperature: 350°C Stirrer speed: 1200 rpm Heteroatom: 200ppm N Volume: 100ml Concentration: 600 ppm Mo Initial reactor pressure: 2.8 MPa Heating rate: 10°C/min Reaction period: 5 hr Product liquid analysis: Time, hr Pressure psig THCBZ CHB CHCHe mol/cc mol ratio mol/cc mol ratio mol/cc mol ratio 0 625 1.16x10"' 0.0102 0.00 0.0000 0.00 0.0000 0.5 617 4.84xl0" 7 0.0426 0.00 0.0000 0.00 0.0000 1 621 8.55x10"' 0.0753 0.00 0.0000 0.00 0.0000 1.5 604 1.18x10"" 0.1039 0.00 0.0000 7.83xl0"8 0.0069 2 595 1.39x10"" 0.1226 9.16xl0"8 0.0081 8.42xl0"8 0.0074 3 580 1.63x10"" 0.1431 1.32x10"' 0.0116 9.35xl0" 8 0.0082 4 572 1.36x10"" 0.1199 1.51x10-' 0.0132 1.02x10"' 0.0089 5 566 1.06x10"" 0.0929 1.29 xlO" 7 0.0113 1.52x10"' 0.0133 Time, hr BCH CPMCH HCH Total product mol/cc mol/cc mol ratio mol/cc mol ratio mol/cc mol ratio 0 0.00 0.0000 0.00 0.0000 0.00 0.0000 1.16x10"' 0.5 0.00 0.0000 0.00 0.0000 0.00 0.0000 4.84x10"' 1 6.65x10"' 0.0059 0.00 0.0000 0.00 0.0000 9.22x10"' 1.5 9.78xl0" 8 0.0086 9.19xl0"8 0.0081 8.65xl0"8 0.0076 1.53x10"" 2 1.71x10"' 0.0151 1.97x10"' 0.0173 2.08x10"' 0.0183 2.15x10"" 3 3.13x10"' 0.0276 5.06x10"' 0.0445 5.57x10"' 0.0490 3.23x10"" 4 4.84x10"' 0.0426 8.22x10"' 0.0724 9.14x10"' 0.0804 3.83x10"" 5 6.46x10"' 0.0569 1.15x10"" 0.1011 1.26x10"" 0.1112 4.39x10"" Time, hr Carbazole mol/cc Conversion, % mol ratio 0 1.14x10° 0.00 1.0000 0.5 1.10x10° 3.24 0.9676 1 1.06x10° 7.09 0.9291 1.5 9.94x10"" 12.48 0.8752 2 9.33x10"" 17.86 0.8214 3 8.25x10"" 27.38 0.7262 4 7.64x10"" 32.73 0.6727 5 7.08x10"" 37.65 0.6235 HDN after 5 hr reaction = 30.37 % 249 Experiment M-30 Reaction: H D N of carbazole Reactant: carbazole Solvent: hexadecane Catalyst: MoNaph + water Input gas: 5%H2S/95%H2 Reaction temperature: 350°C Stirrer speed: 1200 rpm Heteroatom: 200ppm N Volume: 100ml Concentration: 600 ppm Mo Initial reactor pressure: 2.8 MPa Heating rate: 10°C/min Reaction period: 5 hr Product liquid analysis: Time, hr Pressure psig THCBZ CHCHe mol/cc mol ratio mol/cc mol ratio 0 880 0.00 0.0000 0.00 0.00 0.5 879 1.52x10"' 0.0128 0.00 0.00 1 854 3.02 xlO" 7 0.0254 0.00 0.00 1.5 874 6.84x10"' 0.0574 0.00 0.00 2 848 7.94xl0" 7 0.0667 0.00 0.00 3 830 1.17xl0"b 0.0984 0.00 0.0000 • 4 807 1.63 xlO" 6 0.1374 7.91 x 10 8 0.0066 5 814 1.68xl0"6 0.1415 7.46xl0"8 0.0063 Time, hr BCH CPMCH HCH Total product mol/cc mol/cc mol ratio mol/cc mol ratio mol/cc mol ratio 0 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.5 0.00 0.00 0.00 0.00 0.00 0.00 1.52x10"' 1 0.00 0.00 0.00 0.00 0.00 0.00 3.02x10"' 1.5 . 0.00 0.00 0.00 0.00 0.00 0.00 6.84x10"' 2 0.00 0.00 0.00 0.00 0.00 0.00 7.94x10"7 3 5.18xl0"8 0.00 1.08xl0"7 0.0090 0.00 0.00 1.33 xlO"6 4 8.70xl0"8 0.0073 1.16x10"' 0.0097 0.00 0.00 1.92 xlO"6 5 1.33x10"' 0.0111 2.12xl0"7 0.0178 1.49x10"' 0.0125 2.25 xlO"6 Time, hr Carbazole mol/cc Conversion, % mol ratio 0 1.19X10"5 0.00 1.0000 0.5 1.18xl0"5 1.28 0.9872 1 1.16X10"' 2.54 0.9746 1.5 1.12xl0"5 ' 5.74 0.9426 2 1:1 ix io° 6.67 0.9333 3 1.06 xlO"5 11.18 0.8882 4 9.99xl0"6 16.10 0.8390 5 9.65xl0"6 18.92 0.8108 H D N after 5 hr reaction = 4.77% 250 Experiment M-31 Reaction: H D N of carbazole Reactant: carbazole Solvent: hexadecane Catalyst: exMoS2-D Input gas: 5 %H2S/95 %H 2 Reaction temperature: 350°C Stirrer speed: 1200 rpm Heteroatom: 200 ppm N Volume: 100 ml Concentration: 600 ppm Mo Initial reactor pressure: 2.8 MPa Heating rate: 10°C/min Reaction period: 5 hr Product liquid analysis: Time, hr Pressure psig THCBZ CHB CHCHe mol/cc mol ratio mol/cc mol ratio mol/cc mol ratio 0 621 0.00 0.000 0.00 0.000 0.00 0.000 0.5 532 6.86xl0" 7 0.059 0.00 0.000 0.00 0.000 1 522 9.33x10"' 0.081 0.00 0.000 0.00 0.000 1.5 517 1.08x10"" 0.094 0.00 0.000 0.00 0.000 2 570 1.14xl0"6 0.099 6.90xl0"8 0.006 0.00 0.000 3 580 1.26x10"" 0.109 1.07x10"' 0.009 0.00 0.000 4 584 1.37x10"" 0.118 1.57xl0"7 0.014 3.61 x 10 8 0.014 5 487 1.59x10"" 0.138 2.29xl0" 7 0.020 7.74xl0"8 0.022 Time, hr BCH CPMCH HCH Total product mol/cc mol/cc mol ratio mol/cc mol ratio mol/cc mol ratio 0 0.00 0.000 0.00 0.000 0.00 0.000 0 0.5 0.00 0.000 0.00 0.000 0.00 0.000 6.86x10"' 1 6.80xl0" 8 0.006 0.00 0.000 0.00 0.000 1.00x10"" 1.5 1.19x10"' 0.010 6.20xl0"8 0.005 0.00 0.000 1.26x10"" 2 1.64x10"' 0.014 1.53x10"' 0.013 0.00 0.000 1.53x10"" 3 2.17xl0" 7 0.019 3.33x10"' 0.029 1.05 xlO" 7 0.009 2.02x10"" 4 3.41x10"' 0.030 5.09x10"' 0.044 2.13x10"' 0.018 2.62x10"" 5 4.44x10"' 0.039 7.25x10"' 0.063 3.90x10"' 0.034 3.46x10"" Time, hr Carbazole mol/cc mol ratio Conversion, % 0 1.15x10° 1.000 0.00 0.5 1.08x10° 0.941 5.95 1 1.05x10° 0.913 8.67 1.5 1.03x10° 0.890 10.96 2 1.00x10° 0.868 13.24 3 9.51x10"" 0.825 17.54 4 8.91x10"" 0.773 22.72 5 8.07x10"" 0.700 30.00 H D N after 5 hr reaction = 16.2 mol % 251 Experiment M-32 Reaction: Hydrodeoxygenation of phenol Reactant Solvent Catalyst Input gas Reaction temperature Stirrer speed Weight Volume Mo concentration Initial reactor pressure Heating rate Reaction period Phenol hexadecane exMoS 2-D 5%H 2S/95%H 2 350°C 1200 rpm 5 wt% in solvent 100 ml 600 ppm Mo 2.8 MPa 10°C/min 5hr Product liquid analysis: Time, Pressure,ps Phenol Benzene, hr ig mol/cc xlO4 conversion, % mol/cc xlO5 0 668 4.01 0.00 0.15 0.5 651 3.91 2.35 1.10 1 621 3.84 4.24 1.85 1.5 586 3.75 6.35 2.70 2 553 3.68 8.14 3.42 3 496 3.58 10.73 4.46 4 444 3.49 13.03 5.37 5 400 3.41 14.96 6.15 HDO after 5 hr reaction =15.0 mol% 252 Experiment M-33 Reaction: Hydrodeoxygenation of phenol Reactant Solvent Catalyst Input gas Reaction temperature Stirrer speed Weight Volume Mo concentration Initial reactor pressure Heating rate Reaction period Phenol hexadecane exMoS 2-W 5%H 2S/95%H 2 350°C 1200 rpm 5 wt% in solvent 100 ml 600 ppm Mo 2.8 MPa 10°C/min 5 hr Product liquid analysis: Time, Pressure, Phenol Benzene, hr psig mol/cc *104 conversion, % mol/cc 0 1239 4.52 0.00 3.06 xlO"' 0.5 1443 4.52 0.04 4.75 xlO"' 1 1408 4.52 0.12 8.53 xlO"' 1.5 1384 4.51 0.23 1.36 xlO" 6 2 1355 4.51 0.34 1.86 xlO" 6 3 1318 4.50 0.59 2.98 xlO" 6 4 1283 4.48 0.90 4.36 xlO" 6 5 1190 4.47 1.15 5.49 xlO" 6 HDO after 5 hr reaction =1.15 mol% 253 Experiment M-34 Reaction: Hydrodeoxygenation of phenol Reactant Solvent Catalyst Input gas Reaction temperature Stirrer speed Weight Volume Mo concentration Initial reactor pressure Heating rate Reaction period Phenol hexadecane MoNaph 5%H 2S/95%H 2 350°C 1200 rpm 5 wt% in solvent 100 ml 600 ppm Mo 2.8 MPa 10°C/min 5hr Product liquid analysis: Time, Pressure, Phenol Benzene, hr Psig mol/cc xlO4 conversion, % mol/cc 0 671 4.03 0 8.72 xlO"7 0.5 665 3.98 1.20 5.71 xlO-" 1 651 3.95 1.94 8.69 xl0"b 1.5 637 3.90 .3.22 1.38 xlO"5 2 623 3.86 4.10 1.74 xlO"5 3 607 3.78 6.21 2.59 xlO"5 4 596 3.67 8.81 3.63 xlO"5 5 571 3.58 11.13 4.57 xlO"5 HDO after 5 hr reaction =11.1 mol% 254 Experiment MS 5 Reaction: Hydrodeoxygenation of phenol Reactant Solvent Catalyst Input gas Reaction temperature Stirrer speed Weight Volume Mo concentration Initial reactor pressure Heating rate Reaction period Phenol hexadecane crystalline M0S2 5%H 2S/95%H 2 350°C 1200 rpm 5 wt% in solvent 100 ml 600 ppm Mo 2.8 MPa 10°C/min 5hr Product liquid analysis: Time, Pressure, Phenol Benzene, hr psig mol/cc x lO 4 conversion, % mol/cc 0 703 4.03 0 2.85 x lO - ' 0.5 673 4.01 0.53 2.42 x 10"b 1 609 3.98 1.17 5.00 xlO" 6 1.5 603 3.97 1.53 6.44 xlO" 6 2 579 3.94 2.09 8.71 xlO" 6 3 563 3.92 2.74 1.13 xlO" 5 4 547 3.87 3.85 1.58 xlO"' 5 . 541 3.84 4.65 1.90 x 10 5 HDO after 5 hr reaction = 4.65 mol% 255 Experiment MS 6 Reaction: Hydrodeoxygenation of phenol Reactant Solvent Catalyst Input gas Reaction temperature Stirrer speed Weight Volume Mo concentration Initial reactor pressure Heating rate Reaction period Phenol hexadecane A H M derived M0S2 5%H 2S/95%H 2 350°C 1200 rpm 5 wt% in solvent 100ml 600 ppm Mo 2.8 MPa 10°C/min 5hr Product liquid analysis: Time, Pressure, Phenol Benzene, hr psig mol/cc x lO 4 conversion, % mol/cc 0 797 4.02 0 1.94 xlO' 6 0.5 803 3.85 4.33 1.94 xlO' 5 1 716 3.78 6.16 2.67 xlO' 5 1.5 797 3.64 9.65 4.08 xlO"5 2 774 3.56 11.56 4.85 xlO' 5 3 752 3.45 14.34 5.97 xlO" 5 4 732 3.34 17.13 7.09 xlO" 5 5 712 3.25 19.34 7.98 xlO" 5 HDO after 5 hr reaction = 19.3 mol% 256 Sample Calculations of Rate Constants E - l Example of Rate Constant Determination from Experimental Data E - l . l Example of 1st Order Rate Constant Determination Data of experiment M-12 was used as the example (HDS of DBT at 350°C using exMoS2-Design equation for liquid phase batch reactor catalytic reaction. where [DBT] = concentration of DBT, mol/cc t = time, hr fDBT = rate of disappearing of DBT, mol/(cc-g Mo-hr) w = weight of catalyst used in the reaction, g Mo With the assumption that the reaction rate follows the 1 s t order rate law: D). d[DBT] dt = r, w (E-l) T*DBT ' = -A: [DBT] (E-2) where k,= rate constant, (g Mo-hr) -i d[DBT] dt = -k'[DBT] • w (E-3) Integration of Eq.(E-3), 257 \x\[DBT] - ln[DBT0] = -k'-w • t [DBT] ln [DBT 0 ] / [DBT] versus time was then plotted as shown in Figure E - l . (E-4) 0.5 Figure E - l Plot of ln( [DBT 0 ] / [DBT]) versus time in H D S of D B T using exfoliated M o S 2 a t 350°C. From the fitted line, l n I ^ U = 0.0794-f [DBT] Hence, k'-w = 0.0794 Catalyst used during reaction = 640 ppm M o = 0.00064 g o f M o / g of solvent Density of solvent (hexadecane) at 350°C = 0.5042 g/cc (E-5) 258 Therefore, k' = 0-0794 hr x _ 1 hr 0.00064 g M o /g s o l v e n t x 0.5042 g s o l v e n t/cc 3600 s = 6.83 xiO" 2cc/(gMo- s) E-1.2 Sample Calculation for Zero Order Rate Constant Determination Data of experiment M-32 was used as the example (HDO of phenol at 350°C using exMoS 2-D). Design equation for liquid phase batch reactor catalytic reaction. d[phenol] , — 5 — r * - w ( E - 6 ) where [phenol]= concentration of phenol, mol/cc t = time, hr '"phenol' = rate of disappearing of phenol, mol/(cc-g Mo-hr) w = weight of catalyst used in the reaction, g Mo With the assumption that rate follow the zero order rate law: '"phenol' = - k' (E-7) where k'= rate constant, mol/(cc-g Mo-hr) • ^[phenol] _ — Integration of Eq.(E-3), iphenol f lhenol„ ^ [ P h e n 0 1 ] = "^Phenol ' W d t [phenol]-[phenolo] = -phenol'"W-r [phenol] = -&Phenoi'"W-f + [phenolo] (E-9) 259 Concentration of phenol versus time was plotted as shown in Figure E-2. From the fitted line, [phenol] =-0.1186r +3.9614 (E-10) Hence, ^phenol' mW= 0.1186 o 4 2 o 3.0 -I , , 1 , 1 0 1 2 3 4 5 t = time, hr Figure E-2 Concentration profile of phenol versus time in HDO reactions using exfoliated MoS2at350°C. Catalyst used during reaction = 600 ppm Mo = 0.0006 g of Mo/ g of solvent Density of solvent (hexadecane) at 350°C = 0.5042 g/cc _ . 0.1186xlO"5mol/cc hr lhr 1 hereiore, k = x 0.0006 g M o /g s o l v e n t x 0.5042 g s o l v e n t/cc 3600 s = 1.09 xlO -5mol/(g Mo-s) 260 E-2 Example of Apparent Activation Energy Calculation Arrhenius equation: k = A• exp (-EaIRT) (E-l 1) Where, k = rate constant A =Arrhenius constant Ea = activation energy, J/mol R = gas constant = 8.3144 J/mol.K T = reactions temperature, K When In K_BT was plotted versus 1000/T, a straight line can be fitted as shown in Figure E-3. -13 -14 £ -15 1 M -16 -17 y =-17.94x + 14.077 R2 = 0.9885 1.5 1.55 1.6 1.65 x = 1000/T, K"1 1.7 Figure E-3 Plot of In k D B T versus 1000/T in HDS of DBT using MoNaph derived MoS 2 . The fitted straight line has an equation of v = -17.94x+ 14.077 (E-12) This is compatible with the following equation from Eq.(E-l 1): In k= In A-Ea/RT (E-l 3) 261 From Eq.(E-12) and Eq.(E-13): -Ea/R = -17.94/1000 Ea = 149 kJ/mol E-3 Sample Calculation for Confidence Interval of Estimated Parameter Data of experiment M-2 was used as an example to compute the rate constant for disappearance of naphthalene. Experimental Data: Time (hr) 0 0.5 1.0 1.5 2.0 3.0 4.0 5.0 NTL (mol/cc) 5.84E-4 5.60E-4 5.55E-4 5.17E-4 5.08E-4 4.87E-4 4.61E-4 4.26E-4 Recall that the rate of disappearance can be written as: d C ± = -kC" dt Where Ca , k and n denote reactant concentration, rate constant and reaction order, respectively. Assuming that the reaction is of first order, i.e. n = 1, the above equation can be recast into the following form: In W o V C a J = kt Clearly, the rate constant can be computed from the graph ln(C a 0 / Ca) vs t. 9 1 From the excel spread sheet, the fitted k is found to be 6.25 x 10" hr" or 5.74 x 10 ml/g.Mo.s. 262 The 95% confident interval is calculated as follows: Step 1: The residual is computed using the following equation: where yt and yt denote the ln(C o 0 ICa) derived from the experimental data and that obtained from the fitted curve at time /-th, respectively. Step 2: The standard error of the fitted k can then be calculated using the following standard ordinary least square formula: SE(k) = where n, tit and t represent the sample size, time i-th and the average of time, respectively. In this case, the resulting standard error of A: is 3.4 x 10-3 hr"1 or 3.2 x 10"3 ml/g.Mo.s. Using the two tail student-t distribution, the "student's t value for 95% confident interval with 6 degree of freedom is 3. Therefore, the resulting 95% confident interval for the estimated k can be calculated as follows: ±t0 025 6SE(k) = ±3x3.2x10^ = ± 9 . 6 x l 0 - 3 ml/g.Mo.s. 263 E-4 Plots of Experimental Data Versus Simulated Data Exp. data Exp. data Figure E-4 Plots of experiment data versus simulated data for conversion of DBT (•) and yields of products (THDBT: A; BP:x; CHB+BCH:)K) in HDS of DBT using (a) crystalline MoS 2 ; (b) exfoliated MoS 2 ; (c) MoNaph derived MoS 2 ; (d) A H M derived MoS 2 at 350°C. 264 Exp. data Figure E-5 Plots of simulated data versus experiment data for conversion of carbazole (•) and yields of products (THCBZ: • ; CHB:x; CHCHe:«; gBCH:») in hydrodenitrogenation of carbazole using (a) exfoliated M0S2 and, (b) MoNaph derived M o S 2 at 350°C. 265 ^ = = = = Appendix F ______________===_______=_= Calculations in MoS2 Characterization and Analysis F - l Example of Calculations from X R D Spetrum 600 Figure F - l X R D spectrum for peak corresponding to (002) plane in exfoliated M0S2 dispersed in water. Inter-planar spacing of each M0S2 layers, d, was calculated using Bragg's law from the broadening of peak 20 =14.04: 2 d sin 0 = X d = 0.1541 = 6.3 A 2 sin 7.2. K 180. (F-l) 266 Stack height, D s t a ck of M0S2 particle was determined from Scherrer equation: Dstack = 0.894/(0 cos 6) (F-2) where X = x-ray wave length, 0.1541 nm and, P = full-width at half maximum. From Figure F - l , the peak corresponding to (002) plane of M o S 2 (29 =14.04) in the X R D spectrum, p = Pright- P i e f t = 15.2 -13.2 = 2.00 0.89x0.1541 So Dstack 2 sin 14.04 n x 180 = 4.0nm Therefore, the M0S2 particle has average number of layer, Ns = D stack/^ = 4.0/0.63 ~ 6 layers F-2 MoS 2 Slab Length Calculations The M0S2 slab length was determined from T E M micrographs (Hensen et al., 2001). From different micrographs, the length (/) of at least 100 M0S2 fringes was measured. The average slab length of M o S 2 particle, L was determined by the following equation: L = __li-y' (F-3) i=\..x where yt being the frequency of U out of x number of total fringes calculated. 267 F-3 Calculation of M0S2 Edge Sites, Rim Sites and Mo Dispersion The calculations of M0S2 edge sites and rim sites was based on the assumption that the M0S2 slab was a perfect hexagon, as presented by Kasztelan et al. (1984). The equations used for a single slab are presented as follows: Relations between the average slab length (Z) and the number of edge site Mo atoms, n: Z = 3.2(2«-l) A (F-4) Total number of Mo: 3n2-3n+l (F-5) Number of Mo at the edge: 6n-6 (F-6) An example of active sites calculations for crystalline M0S2 is given here: Given the Mo added into each experiments = 4.75 x l fT 5 mol = 4.75 x lO - 5 mol x 6.023 x l O 2 3 = 2.86 x l O 2 0 atom Mo From T E M analysis, the average slab length (L) for crystalline M0S2 was 560nm. Hence, from Eq.(F-4), « = (Z/3.2+l)/2 = 876 Total number of Mo for a single slab, M T = 3n2-3n+\ = 2296875 From X R D analysis, crystalline M0S2 has average layers, Ns = 57. Therefore, single M0S2 cluster = Ns X M T = 57 x 2296875 = 1.31 x 108 atom Mo Total number of Mo add in Total number of M0S2 cluster : Number of Mo in a single cluster = 2.86 x l 0 2 0 / 1 . 3 1 x l 0 8 2.18 x l O 1 2 268 For a single slab, number of Mo at the edge = 6n-6 = 5583 For a single cluster, number of Mo at the edge = 5583 x 57 (Ns) = 3.18 x 10 5 Total number of Mo at the edge = (3.18 xl0 5 )x(2.18 x l 0 1 2 ) = 6.95 x l O 1 7 The Mo dispersion,/MO, was defined as: _ Total number of Mo at the edge and rim sites Total number of Mo atoms The following equation gives/MO in terms of n and x, x being the total number of slabs: /MO=2> , -6 /2> , 2 -3# I i . +1 i=\..x i=\..x For crystalline M0S2, / M O = 6.95 x l O 1 7 / atom Mo) / (2.86 x l O 2 0 atom Mo) = 0.23 % Mo atom at the edge located at the top and bottom layers of a cluster is considered as rim sites. The ultimate number of Mo at the edge sites, M E , are defined as total number of Mo at the edge planes other than rim sites, = 5583 x (57-2) x (2.18 x l O 1 2 ) = 6.70 x l O 1 7 and the total number of Mo at the rim sites, M R , = 6.95 x l O 1 7 - 6.70 x l 0 1 7 = 2.44 x l O 1 6 _ . . . . M R Relative rim sites = Total number of Mo atom at the edge sites = (2.44 x l0 1 6 a tom Mo) / (6.95 x l O 1 7 atom Mo) = 0.035 269 Relative edge sites = Total number of Mo atom at the edge sites = (6.70 x l O 1 7 atom Mo) / (6.95 x l O 1 7 atom Mo) = 0.964 270 Algorithm for Gaussian Newton-Raphson Parameter Estimation Algorithm used for solving kinetic equations in different hydroprocessing reactions are given in Figure G - l . The detail of each step in the calculation please refer to Englezos and Ka,(200T). C Start 3 Initialization k, error, toi, maxlte ii Experimental Data .t,y, No Output Results >j ' Fitted values Statistical property ^ Stop ^ • Jacobian matrix Solve ODE dz/dt lilllll ill E l l l l l i l l l (User defined) A and B matrices Generate Sensitivity matnx and formulating matrices A and B Marquardt Modification Apply Marquardt modification to matnx A Lino Soarch Update u LU Decomposition Solve for Ak Update k Update k and compute •' '•' error Warning «Maximum: iteration iii fcXi 01 J stop ) Figure G - l Algorithm for Gaussian Newton used in rate constant estimation 271 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.831.1-0058992/manifest

Comment

Related Items