UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Residue upgrading using dispersed catalysts prepared in reverse micelles Duangchan, Apinya 1998

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

Item Metadata

Download

Media
831-ubc_1998-271331.pdf [ 15.55MB ]
Metadata
JSON: 831-1.0058629.json
JSON-LD: 831-1.0058629-ld.json
RDF/XML (Pretty): 831-1.0058629-rdf.xml
RDF/JSON: 831-1.0058629-rdf.json
Turtle: 831-1.0058629-turtle.txt
N-Triples: 831-1.0058629-rdf-ntriples.txt
Original Record: 831-1.0058629-source.json
Full Text
831-1.0058629-fulltext.txt
Citation
831-1.0058629.ris

Full Text

RESIDUE UPGRADING USING DISPERSED CATALYSTS PREPARED IN REVERSE MICELLES by APINYA DUANGCHAN B.Sc. Chulalongkorn University, 1981 M.Sc. University of Alberta, 1994  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE STUDIES DEPARTMENT OF CHEMICAL AND BIO-RESOURCE ENGINEERING We accept this thesis as conforming to the required standard  THE UNIVERSITY OF BRITISH COLUMBIA April 1998 © Apinya Duangchan, 1998  In  presenting  degree freely  this  at the  thesis  in  partial  University  of  British Columbia, I agree that the  available for reference  copying  of  department  this or  publication of  thesis for by  his  or  DE-6 (2/88)  the  her  representatives.  of  7  IW  for  an  It  is  granted  advanced  Library shall make  agree that permission for  this thesis for financial gain shall not  /.PfrL3Q  requirements  scholarly purposes may be  The University of British Columbia Vancouver, Canada  D a t e  of  and study. I further  permission.  Department  fulfilment  by the  understood  that  it  extensive  head  of  copying  my or  be allowed without my written  Abstract  Previous studies of slurry phase reactors for residue oil upgrading used dispersed metal sulfide catalysts, prepared by  in situ  decomposition of either  FeS0  4  or the  naphthenate salts of Co, Mo and Ni, that were mixed with the residue oil. However, the size of the catalyst particles was not well established and a high dispersion of the catalysts in the oil was not always achieved. Recent studies have shown that catalysts prepared in reverse micelles can be well characterized and have particle sizes in the nm range. In the present study, catalysts prepared in reverse micelles were investigated as model catalysts for residue oil upgrading, to study the effects of metal type, particle size and dispersion on residue oil upgrading. Fe, Co and Mo catalysts were prepared in n-hexane/PE4LE (polyoxyethylene-4lauryl ether), decalin/PE4LE, toluene/DDAB (d^dodecyldimethylarnmonium bromide), and tetrahydrofuran/DDAB microemulsions, using nitrate salts of Fe and Co, and M0CI5. The metal catalysts were prepared by reduction of the metal ions with  N2H4.XH2O  or  L1BH4.  The size of metal particles prepared in the n-hexane/PE4LE microemulsions was measured using transmission electron microscopy (TEM). The size of the catalysts was in the range 5-8 nm and decreased in the order: Mo > Fe > Co. The size of Co particles prepared from  Co/PE4LE/n-hexane with various water/PE4LE ratios was in the range 5.1-5.9 nm, but no trend of increased particle size with increased water/PE4LE ratio was apparent from these measurements. sulfided  The catalysts prepared in microemulsions using different solvents were  in situ in  the reactor. The performance of the colloidal catalysts for residue oil  ii  upgrading was determined in a batch reactor operated at 430°C, in 5%H S/95%H , at an 2  2  initial pressure of 500 psig and a reaction time of 1 h. The dispersion of the metal catalysts in the residue oil was studied by mixing the residue oil with the metal catalysts prepared in reverse micelles. The amount of metal catalyst recovered in the asphaltene fraction of the residue oil was measured. A high catalyst content of the asphaltenes was taken as a measure of good dispersion of the catalysts in the asphaltenes. There was no significant difference in dispersion of metal catalysts prepared in hexane, decalin, toluene or tetrahydrofuran based microemulsions. Hence, any effect of microemulsion solvent on hydroconversion activity could be ascribed to the chemical properties of the solvents rather than catalyst dispersion effects. Residue oil hydroconversion using different metal catalysts prepared from different microemulsions using different solvents, showed a significant effect of metal type and solvent. Mo was superior to Fe and Co for suppression of coke and gas formation. Coke yield from decalin/PE4LE microemulsions was 5.8%, 6.2%, and 7.1% for Mo, Co and Fe, respectively. Mo provided the highest MCR and S conversions but lowest asphaltene conversion. Hence, it was concluded that Mo was a better hydrogenolysis catalyst whereas Fe and Co were better hydrogenation catalysts at the conditions of the present study. Hydrogen donor ability was important for residue upgrading and the choice of microemulsion solvent had a detectable effect on catalyst performance. Mixed CoMo and NiMo catalysts showed lower coke yield than the Co catalyst (5.0% for CoMo and NiMo compared to 6.2% for Co) and lower than that of Mo (5.8%). A synergistic effect of the CoMo and NiMo catalysts was observed for MCR conversion and coke yield. Cobalt  iii  naphthenate catalyst precursor gave higher conversions of MCR and S, and a comparable >525°C fraction conversion but gave lower asphaltene conversion than the Fe, Co, and Mo catalysts prepared in microemulsions.  iv  T a b l e  o f  C o n t e n t s  Abstract  ii  List of Tables  xi  List of Figures  xiii  Acknowledgements  xvii  Chapter 1  1  1.1 Background  1  1.2 Dispersed Catalysts for Slurry Phase Hydroconversion  4  1.3 Objectives of This Study  6  Chapter 2  7  Literature Review  7  2.1 Heavy Oils and Petroleum Residues  7  2.1.1 Asphaltenes  8  2.1.2 Metals and Heteroatoms  9  2.2 Hydroprocessing Reactions  11  2.2.1 Hydrodestilfurization (HDS)  11  2.2.2 Hydrodenitrogenation (HDN)  12  2.2.3 Hydrodemetallization (HDM)  13  2.2.4 Hydrocracking (HC)  13  2.3 Heavy Oil Upgrading (Primary Upgrading)  15  2.3.1 Carbon Rejection  15  2.3.2 Hydrogen Addition  16  2.3.2.1 Catalytic Hydrocracking  18  2.3.2.2 Residue Hydrocracking  18  2.4 Deactivation of Catalysts 2.4.1 Deposition of Coke  21 21  2.4.2 Deposition of Metal Sulfides 2.5 Heavy Oil Upgrading Processes  22 23  2.5.1 Fixed-Bed Trickle Reactors (FBR)  23  2.5.2 Ebullating-Bed Reactors (Back-Mixed Reactor, BMR)  24  2.5.3 Slurry-Phase Reactor  24  2.6 Dispersed Catalysts for Heavy Oil Hydrocracking 2.6.1 Introduction  26 26  2.6.2 Hydroconversion Mechanism in the Presence of Dispersed Catalysts  26  2.6.3 The Synthesis of Metal Catalysts  30  2.6.4 Effect of Precursor Solubility on Catalyst Activity  30  2.6.5 Active Species and Size Distribution of Dispersed Catalysts  32  2.7 Mixed Catalysts  33  2.8 Preparation of Metal Colloids in Reverse Micelles  34  2.8.1 Introduction  34  2.8.2 Emulsion and Microemulsion  35  2.8.3 Metallic Colloid Formation in Reverse Micelles  38  2.8.4 Factors Affecting the Size of Reverse Micelles and Colloids  40  2.8.4.1 Effect of Water Content  40  2.8.4.2 Effct of Metal Ion Concentration  42  2.8.4.3 Effect of Reducing Agent on the Size of the Particle  43  2.9 Catalytic Activity of Colloids Synthesized in Reverse Micelles  44  2.10 Summary  46  er 3  48  Experimental  48  3.1 Metal Catalyst Preparation in Reverse Micelles  48  3.1.1 Introduction  48  3.1.2Hexane  ,  50  3.1.3 Decalin  52  3.1.4 Toluene  54  3.1.5 THF  55  3.2 Preparation of Co Catalysts in Microemulsions with Different Water/Surfactant Volume Ratios 3.3 Colloid Characterization 3.3.1 Transmission Electron Microscopy 3.4 Dispersion of Colloidal Catalysts in Residue Oil 3.4.1 Procedure 3.5 Catalyst Activity Measurement  56 56 56 57 58 60  3.5.1 Experimental Setup  61  3.5.2 Hydroconversion Activity  63  3.5.3 Analysis of Gas Product  64  3.5.4 Determination of Coke Yield  66  3.5.5 Analysis of Liquid Products  66  3.5.5.1 Microcarbon Residue (MCR) Analysis  67  3.5.5.2 >525°C Fraction Analysis by Simulated Distillation of Petroleum Products  67  3.5.5.3 Asphaltene Analysis  68  3.5.5.4 Sulfur Analysis  69  er 4  70  Catalyst Preparation and Characterization  70  4.1 The Synthesis of Reduced Metal Colloids  70  4.1.1 Effect of Surfactant on Particle Size  71  4.1.2 Salting In and Out of Surfactants by Complexation with Inorganic Salts  72  4.1.3 Effect of Additives on HLB  73  4.1.4 Hard and Soft Acids and Bases  76  4.2 Characterization of the Metal Colloids  77  4.3 Characterization of Reduced Metal Colloids in Reverse Micelles with Different Water/Surfactant Volume Ratios 4.4 Summary Chapter 5  86 95 96  Activity Measurements  96  5.1 Introduction  96  5.2 Dispersion of Colloidal Catalysts in Residue Oil  96  5.2.1 Effect of Mixing and Settling Time  98  5.2.2.Effect of Solvents  101  5.3 Residue Upgrading Activity of Fe, Co and Mo Catalysts Prepared in Reverse Micelles  102  5.3.1 Effect of Catalyst on Gas Yield  103  5.3.2 Effect of Catalyst on Coke Yield  106  5.3.3 Effect of Catalyst on Liquid Product Quality  108  5.3.3.1 MCR(Microcarbon Residue) Conversion  Ill  5.3.3.2 >525°C Conversion (or Residue Conversion)  Ill  5.3.3.3 Sulfur Conversion  112  5.3.3.4 Asphaltene Conversion  112  5.4 Effect of Microemulsion Solvent on Catalyst Residue Upgrading Activity  116  5.4.1 Effect on Autoclave Pressure  116  5.4.2 Effect on Coke Yield  117  5.4.3 Effect on Liquid Product Quality  120  5.5 Catalyst Based on Organometallic Precursors  122 viii  5.5.1 Effect on Gas Formation  122  5.5.2 Effect on Coke Yield  124  5.5.3 Effect on Liquid Product Quality  126  5.6 Effect of Mixed Catalyst on Residue Upgrading  127  5.6.1 Effect on Gas Yield  128  5.6.2 Effect on Coke Yield  128  5.6.3 Effect on Liquid Product Quality  130  5.7 Summary Chapter 6  133 135  Conclusions and Recommendations  135  6.1 Conclusions  135  6.2 Recommendations  137  References  140  Appendix 1  151  1.1 Hydrocracking Activity Experiments  152  1.2 Summary of Results of Hydrocracking Activity Experiments  199  Appendix II Sample Calculations  201 202  2.1 Weight of H2/H2S Mixture in the Reactor Before Hydrocracking Reaction  202  2.2 Weight of Gas Outlet  203  2.3 The Amount of Metal in Residue  205  2.4 The Molar Ratio of Reducing Agent to Metal  207  Appendix HI  209  3.1 Comparison of Fe/Ni Ratio of Asphaltene Fraction Using Different Solvents  210  3.2 Comparison of Fe/V Ratio of Asphaltene Fraction Using Different ix  Solvents  213  3.3 Comparison of Coke Yield fromH-Donor Solvent, Organometallic Compound, Catalyst in Reverse Micelles, and Mixed Catalysts  216  3.4 Comparison of MCR Conversions from Fe, Co, Mo, CoMo, and NiMo Catalysts  220  3.5 Comparison of Sulfur Conversions from Fe, Co, Mo, CoMo, and NiMo Catalysts  225  3.6 Comparison of >525°C Conversions fromFe, Co, Mo, CoMo, and NiMo Catalysts  230  3.7 Comparison of Asphaltene Conversions from Fe, Co, Mo, CoMo, and NiMo Catalysts  233  3.8 Comparison of Coke/Asphaltene RatiosfromFe, Co, Mo, CoMo, and NiMo Catalysts  237  3.9 Determination of Number Average, Standard Deviation, Standard Error of Mean and %Standard Error of Mean  241  3.10 The Estimated Cost of Co prepared in Microemulsion Compared to Cobaltnaphthenate 3.11 Determination of Sample Size  241 242  List of Tables Table 3.1 Preparation of Fe, Co, and Mo Catalysts in Water/n-Hexane/PE4LE Microemulsions  51  Table 3.2 Preparation of Fe, Co, and Mo Catalysts in Water/DHN/PE4LE Microemulsion  53  Table 3.3 Composition of Ni, Co, and Mo Catalysts Used for NiMo and CoMo Catalyst Synthesis  54  Table 3.4 Preparation of Fe, Co and Mo Catalysts in Water/Toluene/DDAB Microemulsions  55  Table 3.5 Preparation of Fe, Co, and Mo Catalysts in Water/THF/DDAB Microemulsions  55  Table 3.6 Microemulsion with n-Hexane, Using Different Water/PE4LE Volume Ratios  56  Table 3.7 Properties of the Cold Lake Alberta Residue Oil Used in the Present Study  58  Table 3.8 Composition of Iron Colloidal Microemulsions  59  Table 4.1 The Number-Average Particle Size of Co Catalysts Prepared in Microemulsions with Different Water/Surfactant Ratios in the Co 2+  Solution/Hexane/PE4LE System  86  Table 5.1 Fe, Ni, V, Fe/Ni, and Fe/V Contents of Asphaltene Fraction of Residue Oil When Mixed with Fe/AOT/Ethanol/Toluene Microemulsion Catalyst at Different Mixing Times and a 2-h Settling Time  99  Table 5.2 Fe, Ni, V, Fe/Ni, and Fe/V Contents of Asphaltene Fraction of Residue Oil When Mixed with Fe/AOT/Ethanol/Toluene Microemulsions at Different Settling Times and a 2-h Mixing Time Table 5.3 Dispersion of Fe (Prepared in Different Solvents), Ni, V, Fe/Ni, and  100  Fe/V in Asphaltene Fraction of Residue Oil Following 2-h Mixing and 2-h Settling  101  Table 5.4 Gas Products From Residue Upgrading (50g Residue with 50 ml Metal Catalyst Prepared in Microemulsions, at 500 psig, 5%H S/95%H 2  2  Initial Partial Pressure, 430°C, and 1-h Reaction Time)  103  Table 5.5 Coke Yield From Residue Upgrading Using Metal Catalysts Prepared in DHN/PE4LE Microemulsions, at 500 psig, 430°C, 1 h  107  Table 5.6 Hydroconversion of Residue Oils With Catalysts in DHN/PE4LE Microemulsions, at 500 psig, 430°C, 1 h  109  Table 5.7 Average Product Analysis Data From Residue Oil Conversion with Catalysts in DHN/PE4LE Microemulsions at 500 psig, 430°C, 1 h  110  L i s t o f  F i g u r e s  Figure 1.1  Ebullating-Bed Reactor  3  Figure 2.1  Different Types of Molecules in Heavy Oils  7  Figure 2.2(a) Representation of an Asphaltene Structure from X-Ray Analysis  10  Figure 2.2(b) Asphaltene Structure  10  Figure 2.3  A Simplified Porphyrin Structure  11  Figure 2.4  Hyfaodesulfurization of Two Typical Non-Asphaltenic Fractions in Heavy Oil  Figure 2.5  12  Hydrodenitrogenation Reactions Involve the Intermediate Steps of Saturating the Nitrogen-Containing Ring  12  Figure 2.6  A Simple Representation of the Demetallization Reaction  13  Figure 2.7  The Hydrocracking Reactions of a Polycyclic Aromatic  14  Figure 2.8  A Model of Asphaltene Cracking  14  Figure 2.9  Processing of Heavy Crude Oil  16  Figure 2.10 Possible Pathways for Coking and Hydrocracking of a Residuum Molecule in Athabasca Bitumen  20  Figure 2.11 Metal Sulfide and Carbon Deposits on Sulfided Catalyst as a Function of Catalyst Age  22  Figure 2.12 Slurry-Phase Heavy Oil Upgrading Reaction Mechanism  27  Figure 2.13 M-Coke Conversion Mechanisms  28  Figure 2.14 Hydrocracking Mechanism  29  Figure 2.15 Schematic Illustration of Various Stages in the Growth of Ultrafine Particles in Microemulsions  40  Figure 3.1  Experimental Setup  62  Figure 4.1  TEM Micrograph of Co Particles in Hexane/PE4LE  79  Figure 4.2  TEM Micrograph of Fe Particles in Hexane/PE4LE  80  Figure 4.3  TEM Micrograph of Mo Particles in Hexane/PE4LE  Figure 4.4  Size Distribution of Mo, Fe and Co in Water/PE4LE/n-Hexane Microemulsions  Figure 4.5  81  82  Models of Solubilization of Ni(U), Co(II) and Fe(HI) Ions in the Inner Water Cores of Microemulsions  84  Figure 4.6  Micrograph of Co Prepared in 0.044 Water/PE4LE Ratio  87  Figure 4.7  Micrograph of Co Prepared in 0.055 Water/PE4LE Ratio  88  Figure 4.8  Micrograph of Co Prepared in 0.1 Water/PE4LE Ratio  89  Figure 4.9  Micrograph of Co Prepared in 0.2 Water/PE4LE Ratio  90  Figure 4.10  Particle Size Distribution of Co in Water/PE4LE/n-Hexane Microemulsions with Water/PE4LE Ratios = 0.2 and 0.1  Figure 4.11  Particle Size Distribution of Co in Water/PE4LE/n-Hexane Microemulsions with Water/PE4LE Ratios = 0.055 and 0.044  Figure 4.12  91  92  Effect of Water/PE4LE Ratio on the Hydrodynamic Diameters of Reverse Micelles and Colloidal Particles in the Water/PE4LE/nHexane System  Figure 5.1  The Autoclave Pressure vs Time During Residue Upgrading in DHN/PE4LE  Figure 5.2  108  Liquid Product from Residue Conversion with Catalysts in DHN/PE4LE Microemulsions at 500 psig, 430°C, 1 h  Figure 5.4  105  Coke Yield from Residue Upgrading Using Fe, Co and Mo Catalysts Prepared in DHN/PE4LE Microemulsions  Figure 5.3  93  110  Effect of Fe, Co and Mo Catalysts on Coke Yield and Coke Yield/Asphaltene Ratio on Residue Conversion with Catalysts Prepared in DHN/PE4LE Microemulsions at 500 psig, 430°C, 1 h  Figure 5.5  114  The Autoclave Pressure after 1-h Reaction Time from Residue Upgrading Using Catalysts Prepared from Microemulsions Based on xiv  Different Solvents Figure 5.6  117  Coke Yield after 1-h Reaction Time from Residue Upgrading Using Fe, Co and Mo Catalysts Prepared in Microemulsions Using Different Solvents  118  Figure 5.7(a) Effect of Solvent on Product Quality from Residue Conversion with Co/n-Hexane/PE4LE and with Co/DHN/PE4LE Catalyst at 500 psig, 430°C, 1 h  121  Figure 5.7(b) Effect of Solvent on Coke Yield and Coke Yield/Asphaltene Conversion from Residue Conversion with Co/n-Hexane/PE4LE and with Co/DHN/PE4LE Catalyst at 500 psig, 430°C, 1 h Figure 5.8  121  Autoclave Pressure vs Time During Hydroconversion of Residue Oil  Figure 5.9  Coke Yield from Residue Conversion at 500 psig, 430°C, 1 h  Figure 5.10  Liquid Product from Residue Conversion at 500 psig, 430°C, 1 h,  123 125  Using H-Donor Solvent, Catalyst from Reverse Micelle and Organometallic Compound Figure 5.11  127  The Autoclave Pressure vs Time During Residue Upgrading Using Different Types of Catalysts Prepared in DHN/PE4LE Microemulsions  Figure 5.12  129  Coke Yield in Mixed Catalysts from Residue Conversion at 500 psig, 430°C, 1 h  130  Figure 5.13(a) Liquid Product from Residue Conversion at 500 psig, 430°C,1 h, Using Co, Mo, CoMo and NiMo Catalysts Prepared in DHN/PE4LE Microemulsions  131  Figure 5.13(b) Effect of Co, Mo, CoMo, and NiMo on Coke Yield and Coke Yield/Asphaltene Ratio on Residue Conversion Using Catalysts Prepared in DHN/PE4LE Microemulsions at 500 psig, 430°C, xv  lh.  131  xvi  ACKNOWLEDGEMENT  I would like to thank my supervisor Prof. Kevin Smith for his support and guidance throughout my research. I also want to extend my appreciation to Dr. Goran Boskovic who assisted me at the initial stage of my research. Special thanks to Prof. Bruce Bowen, Prof. Collin Oloman, and Prof. J.S. Laskowski, my thesis committee for their criticism and advice; and to Dr. Miguel A. Romero for his assistance and valuable discussion. My sincere thanks to Mary Mager of the Department of Metals and Materials Engineering, U.B.C. for the instruction for TEM measurement. My appreciation is to the staff and my colleagues in the Chemical Engineering and Bio-Resources Department, Dr. J. Soltan Mohammad Zadeh, Mr. Issa Milad, Mr. Andrew G. Hall and Mr. Qingdong Lui. My gratitude for their kindness and encouragement, Venerable's Lee, Kambor, See, Luan, Wan, and Sona. My sincere thanks, for their love, understanding and patience, to my parents, my sisters and brothers, my relatives and Dr. Pipat Pichestapong. I would also like to thank for their encouragement and support, Mr. Putnam Barber, Dr. Valerie Lynch, Dr. Anak Iamaroon, Mr. Siegfried Gohlke, Mrs. Tossaporn Sariyont, and my colleagues at Kasetsart University in Bangkok, Thailand. I owe much gratitude to the Thailand-Canada Rattanakosin Scholarship Program for their financial support throughout my Ph.D. in Canada, to the Alberta Department of  xvii  Energy for theirfinancialsupport of this project and for providing Cold Lake residue, and to the Department of Chemical and Bio-Resource Engineering, U.B.C.  C  h  a  p  t  e  r 1  I n t r o d u c t i o n 1.1 Background The development of heavy oil processing technology has become more important due not only to me limited supply of crude oil, but also due to the increased exploitation of Canadian oil sands and heavy oils. In catalytic heavy oil upgrading processes, heavy oils are cracked or hydrogenated by both catalytic and thermal reactions to produce the desired distillates. The catalysts used are typically CoMo sulfides or NiMo sulfides supported on Y-AI2O3. The high content of asphaltenes and metals in heavy oils results in the deposition of coke and metal sulfides on the catalyst. The asphaltenes produce coke precursors, from which solid coke evolves. The conversion of the feedstock to useful products is decreased because of thefractionof the heavy oils that is converted to coke. In addition, the solid coke deposits on catalyst surfaces and adheres to the surfaces of the reactor and downstream equipment. The catalysts are rapidly deactivated by the deposition of coke and metal sulfides and the catalysts are difficult to regenerate since, although the coke may be removed by oxidation, the metal sulfides cannot. Many attempts have been made to overcome the problems associated with catalytic heavy oil upgrading processes. One approach is to design catalysts that have a high capacity for metals deposition while maintaining hy(frodesulfurization (HDS) and hydrodemetallization (HDM) activity (Oelderik et al, 1989; Dautzenberg and De Deken, 1985). Oelderik et al. (1989) studied the effect of catalyst structure on hydroconversion,  I  emphasizing the effect of catalyst pore size on conversion and deactivation rates. It was found that narrow pore, high surface area catalysts deactivated rapidly due to the deposition of metal sulfides at the external surface of the catalyst particle. However, the catalyst remained active for hydroconversion on the interior surface where metal sulfide deposition did not take place. Wide pore catalysts accommodated more metals, the pores were plugged slowly, and therefore, extended catalyst lifetimes were possible, but conversions were lower than those of catalysts with narrow pore size because of lower active surface area. These observations are supported by the results of Stanislaus et al. (1993). Other attempts to improve catalytic heavy oil upgrading processes have concentrated on reactor development. The conventional catalytic process for hydrodemetallization of heavy oil uses fixed bed reactors or ebullating-bed reactors (Figure 1.1). For a commercial fixed bed reactor processing metal-rich feedstocks such as Maya vacuum residue (metal concentrations: N i = 127 ppm, V = 635 ppm), the cycle length is less than four months (even when the best HDM catalysts are used). This is because one kilogram of catalyst has to adsorb metals of the order of 1 kg accompanied by about 1 kg of sulfur (the metals deposit as sulfided compounds such as NijSj) (Oelderik et al., 1989). Hence a large amount of down time is needed to replace the spent catalyst. The problem of short run length was partly solved by using an ebullating-bed reactor. In an ebullating-bed reactor, catalyst particles are suspended in the reactor by the upward velocity of the liquid feedstock and the hydrogen gas. Smaller catalyst particles are used in an ebullating-bed reactor (particle size approximately 1 mm) than a fixed-bed  2  reactor (particle size approximately 5 mm). Compared to afixed-bedreactor, the average reaction temperature is usually higher in the ebullating-bed reactor because of a high heat transfer rate. The rates of conversion are higher and the catalyst is replaced on-stream (Dautzenberg and De Deken, 1985). However, the capacity for metals uptake by the catalysts remainsfiniteand the catalyst life remains short. Catalyst deactivation problems may be reduced by using dispersed catalysts in a slurry-phase reactor. In this configuration small catalyst particles (< 1 urn diameter) are slurried with the oil and are transported through the reactor with the H and oil. The use 2  of dispersed catalysts offers several advantages since the small particle size and low concentration allow the particles to be used as a once-through catalyst (Del Bianco et al., 1993).  Vapor-liquid separator  Fresh catalyst  T t  T  Catalyst level (expanded)  t t  t  Catalyst level (settled)  t t  Feedstock plus hydrogen  Spent catalyst  Figure 1.1 Ebullating-bed reactor (Dautzenberg and De Deken, 1984).  3  1.2 Dispersed catalysts for slurry phase hydroconversion Catalysts reported for slurry type processes include FeSC>4, molybdenum naphthenate, and other organometallic compounds (Dabkowski et al., 1991). Nickel and molybdenum compounds are reportedly the most effective for the suppression of coke (Dabkowski et al., 1991). The performance of the catalyst is correlated with the solubility of the metal containing compounds in the oil. Hence organic complexes are normally more active than inorganic salts (Del Bianco et al., 1993). Other studies of slurry-reactor catalysts in the form of Fe(CO) and the naphthenate salts of Ni, Co and Mo have also 3  been reported (Lott et al., 1993). The oil-soluble metal compounds are well dispersed in the heavy oil. The amount of metal compound is in the range 0.01-5 wt%, based on the weight of the feedstock. The dispersed metal compounds decompose and react in situ with sulfur moieties to form metal sulfides. According to Lott et al. (1993) the metal sulfide particle size is of the order of 1 nm. The sulfide colloids accumulate on the surfaces of coke spheroids and interfere with coke precursor coalescence. Less coke is formed when using diluent that has the most hydrogen donor capability. Without added diluent, the coke adheres very strongly to the surfaces of the reactor and baffles. Moreover, continuously removing the highly volatile components from the reacting fluids in the reactor reduces coke formation. Spent catalyst from the slurry reactor remain in the pitch or coke residue of the process (Dautzenberg and De Deken, 1984; Dautzenberg and De Deken, 1985). Catalyst is either discarded or recycled back to the process. In some studies it has been found that the recycled catalysts have activity equivalent to fresh catalyst (Bearden and Aldridge,  4  1979). Even though dispersed catalysts have been successfully used to upgrade heavy oil (Lott et al., 1993), the generation of the active phase of the catalysts from organometallic precursors is not well understood. The sizes of the catalyst particles prior to reaction is not well defined and the dispersion of catalysts in the heavy oil is not necessarily very high and may depend on the solubility of the metal sulfide precursor. Anderson and Bockrath (1984) have revealed that even with oil-soluble organometallic precursors, quite large crystallites or agglomerates can be formed during reaction, and high dispersion is not obtained. Dabkowski et al. (1991) have also demonstrated that the solubility of the metal sulfide precursor plays an important role in determining the catalyst activity. An alternative to oil soluble organometallic catalyst precursors is to prepare catalysts in a microemulsion using water soluble metal salts dissolved in the water pools of reverse micelles in the microemulsion. A technique for making colloidal metals such as Pt, Pd, and Rh dispersed in an organic medium (herein referred to as the solvent) using a microemulsion is available (Boutonnet et al, 1982). The metal colloids are stable, highly dispersed in organic solvents, and have well-defined particle sizes that are reportedly in the range 3-5 nm. According to Hall (1996) and Pileni and Lisiecki (1993) catalysts with different particle sizes can be synthesized by varying the water/surfactant ratio and metal concentration of the microemulsion. In addition, preparation of mixed catalysts is possible by simply mixing the microemulsions of two catalysts prepared seperately. The present study is motivated by the fact that the size of both the metal and metal sulfide catalysts, obtained by sulfiding the metal catalysts prepared in a microemulsion, is very small (nm range), can be controlled, and can be well characterized. Moreover, the  5  colloidal metals prepared in a microemulsion are normally stable and well dispersed in the continuous on-containing solvent phase. Hence these catalysts should be good model catalysts in that their particle sizes will be controlled in the nm range and they will be well characterized. It is also expected that the dispersion of the colloidal metal catalysts in the heavy oil can be controlled by suitable choice of the solvent used for the microemulsion. Finally, the colloidal catalyst properties have not been reported for the heavy oil upgrading process and the importance of these properties for dispersed catalyst performance needs to be established.  1.3 Objectives of this study The objectives of the present study are as follows: (1) To prepare and characterize metal catalysts in the nm size range from water soluble salts using a water-in-oil-microemulsion. (2) To determine the activity of the catalysts prepared in microemulsions for upgrading residues from Cold Lake, Alberta. (3) To determine the effect of the microemulsion solvent on dispersion of the catalyst in the residue and on the catalyst activity. (4) To understand the relationship between activity and catalyst properties (e.g. size and metal type). (5) To investigate mixed catalysts such as CoMo and NiMo, prepared in microemulsions and to identify any synergistic effects of the mixed catalysts.  6  Chapter L i t e r a t u r e  2 R e v i e w  2.1 Heavy oils and petroleum residues Heavy oils are derived from various sources including coal liquids, shale oils, and tars. Heavy oils have low H/C atomic ratios ranging from 1.4 to 1.7 whereas the ratio in gasoline is about 2.1 (Oelderik et al., 1989; Laine and Trimm, 1982). Heavy oils have high viscosity (10 -10 mPa.s) (Gray, 1994) and contain a wide range of compounds. These 2  5  compounds include large paraffinic and polycyclic aromatic molecules, cyclic compounds containing N and S (Figure 2.1), organically bound metals and asphaltene micelles. The boiling points of heavy oils are usually higher than 350°C which make them unsuitable for conventional refining (Laine and Trimm, 1982).  Polycyclic aromatic  Heteroatomic cyclic compounds  Figure 2.1. Different types of molecules in heavy oils (Laine and Trimm, 1982).  Petroleum residue refers to the bottom product from a crude oil vacuum distillation column, having an atmospheric equivalent boiling point (BP) above 525°C. Many conventional crude oils contain 10-30% residue and these residues are usually used  7  to manufacture asphalt (Gray, 1994). Residues contain fractions with BP >525°C that consist mainly of polycyclic aromatic molecules, heteroatoms, and asphaltenes.  2.1.1 Asphaltenes In heavy oil upgrading research, much attention is paid to the asphaltene fraction since it is the key fraction to be cracked to remove heteroatoms and metals. In addition, the presence of asphaltenes is thought to be a major cause of coke formation during heavy oil processing. Asphaltenes are dark-brown to black friable solids that have no definite melting point. When heated they expand and then decompose, leaving a carbonaceous residue (Speight and Moschopedis, 1981). Asphaltenes are obtained from petroleums and bitumens by precipitation with a hydrocarbon solvent having a surface tension lower than 25xl0" N.m" at 25°C. Examples of such solvents include liquefied petroleum gases, low3  1  boiling petroleum naphthas, petroleum ether, pentane, isopentane and hexane (Speight and Moschopedis, 1981). Asphaltenes are soluble in liquids having a surface tension above 25 XlO" N.m such as pyridine, CS , CCU, and CeHe (Speight and Moschopedis, 1981; 3  -1  2  Bland and Davidson, 1967). Asphaltenes by definition are not a single molecular species but are a solubility class (Speight and Moschopedis, 1981). They are the fraction of petroleums that precipitate when one volume of petroleum is mixed with a minimum of forty volumes of a liquid hydrocarbon such as n-pentane, n-heptane or similar nonpolar solvent. The elemental analysis of asphaltenes precipitated using more than 40 volumes of n-pentane as  8  a precipitating medium, shows the amount of carbon to be 82 ± 3 wt%, hydrogen 8.1 ± 0.7 wt%, and the H/C atomic ratio equal to 1.15 ± 0.05 (Speight, 1978; Speight, 1979). It is believed that asphaltenes consist of sheets of condensed polynuclear aromatic ring systems bearing alkyl sidechains and alicyclic systems with heteroatoms scattered in different locations (Speight and Moschopedis, 1981). The number of rings may be as low as six and as high as fifteen to twenty in complex systems (Speight, 1994). The molecular mass of asphaltenes is in the range 600 to 140,000 g/mole (Dautzenberg and De Deken, 1985; Speight and Moschopedis, 1981). The asphaltenes form micellar structures in the heavy oil (Dautzenberg and De Deken, 1985). These micelles remain dispersed in the oil due to the presence of resins (Dautzenberg and De Deken, 1985). The reported average size of asphaltene micelles is in the range 6 to 9 nm (Dautzenberg and De Deken, 1985). Most of the metals and heteroatoms (Ni, V, S, N, and O) in heavy oils are concentrated in the asphaltenes (Hall and Herron, 1981). Not only must the Ni, V, S, N , and O components be removed during upgrading, but the H/C ratio must be increased as well. The structure of asphaltenes according to x-ray analyses is similar to that shown in Figure 2.2 (a) (Speight and Moschopedis, 1981), and results of spectroscopic and analytical techniques lead to the interpretation shown in Figure 2.2 (b) (Gray, 1994).  2.1.2 Metals and heteroatoms Metals present in heavy oil include V and Ni in the range 10-1000 ppm (Oelderik et al., 1989). Most of the metals are contained in metaUopoiphyrins (Elliot and Melchior, 1982). The porphyrins contain tetrapyrollic aromatic structures similar to the aromatic  9  sheets in asphaltenes. Therefore, they can be incorporated readily into the asphaltene micelles (Dautzenberg and De Deken, 1985). The V and N i complexes are square pyramids with four donor nitrogen atoms situated co-planar in the basal spaces (Dautzenberg and De Deken, 1985).  •fomatk sheets •liphjlic chalnt *Ny.  W ~ v ^ 3.6-3.8A(c/2)  W W W "  c/2 interlamellar distance L layer diameter a  L height of the unit cell N number of lamellae c  c  6- 16AIL.1  Figure 2.2 (a) Representation of an asphaltene structure from x-ray analysis (Speight and Moschopedis, 1981).  A. Crystallite; C. Particle; E. Weak link; G. Intracluster; I. Resin; K. Porphyrin;  B. Chain bundle D. Micelle F. Gap & hole H. Intercluster J. Single layer L. Metal (M)  Figure 2.2 (b) Asphaltene structure (Gray, 1994)  10  A simplified poiphyrin structure containing V is shown in Figure 2.3. The typical concentration of heteroatoms and metals in heavy oils is as follows (Gray, 1994): S 2-7 wt%, N 0.2-0.7 wt%, O ~1 wt%, V 100-1040 ppm, and Ni 20-200 ppm It is necessary to remove heteroatoms and metals since S, N , and metals deactivate catalysts and O causes polymerization reactions which lead to gum formation.  CiHs  Figure 2.3 A simplified porphyrin structure (Oelderik et al., 1989)  2.2 Hydroprocessing reactions The reactions that occur in hydroprocessing processes remove heteroatoms, metals, and reduce the molecular weight of the large hydrocarbon molecules.  2.2.1 Hyfaodesulfurization (HDS) HDS is an important step in the refining of crude oil (secondary upgrading) because S is poisonous to most catalysts in the refining process and residual S yields undesired SO upon combustion of the oil Common catalysts for HDS reactions are x  11  presulfided CoMo, NiMo or NiW supported on AI2O3. Typical HDDS reactions are shown in Figure 2.4.  +  2 ILS  H  ^  CH -CH-CH=CH 2  2  >  Butenes and Butane  Thiophene CHj C H  3  + H, HjS  B enzothiophene  Ethylbenzene  Figure 2.4 Hy<kodesulfurization of two typical non-asphaltenic fractions in heavy oil (Laine and Trimm, 1982).  2.2.2 Hydrodenitrogenation (HDN) HDN consumes more hydrogen than does HDS. The common catalysts for HDN reactions are presulfided NiMo or NiW on  AI2O3.  An example of the HDN reaction is  shown in Figure 2.5.  t> Pentane  Figure 2.5 Hydrodenitrogenation reactions involve the intermediate steps of saturating the nitrogen-containing ring (Laine and Trimm, 1982).  12  2.2.3 Hydrodemetallization (HDM) HDM is a very important reaction since heavy oils contain a large amount of V and Ni. These metals are in solution in the oil as dissolved porphyrins or in asphaltene micelles. HDM, similar to hydrocracking, is the hydroscission of the organic chain associated with the metal. The solubility of the metal complex in the oil is reduced during H D M and the metal sulfides produced from HDM will either precipitate on the sulfided catalyst (which can lead to rapid deactivation) or downstream of the sulfided catalyst. A simple demetallization reaction is shown in Figure 2.6. /  \ N  N  I  I  V=0  I  N  H  + 2H,S  > VS  2  +  I  +  HO  H I  N  Figure 2.6 A simple representation of the demetallization reaction (Dautzenberg and De Deken, 1985).  2.2.4 Hydrocracking (HC) The process of hydrocracking is the breakdown of large organic molecules to fragments as a result of cracking and hydrogenation. Common catalysts for HC reactions are sulfided NiMo or NiW. A typical pathway of HC reactions is shown in Figure 2.7. A model of hydrocracking of asphaltenes as proposed by Dautzenberg and De Deken (1985) is shown in Figure 2.8. Reaction "a" shows the disintegration of the asphaltene micelles and the removal of the metal component. The weak S and aliphatic bridges are broken and the metal is removed from the porphyrinic compound. Subsequently, the remaining  13  heteroatoms such as S are removed which will further reduce the molecular weight of the asphaltenes (reaction "b").  Figure 2.7 The hydrocracking reactions of a polycyclic aromatic (Laine and Trimm, 1982)  —M-Metal (vanadium*) Aromatic sheet Aliphatic w^weak link (sulfur)  Asphaltene micelle  -y \  Figure 2.8 A model of asphaltene cracking (Dautzenberg and De Deken, 1985)  14  2.3 Heavy oil upgrading (primary upgrading) Heavy oils contain a large amount of cokeable materials (referred to as Conradson Carbon Residue, CCR), heteroatoms (O, N and S), and metals (Ni, V and Fe) which make it difficult to process in existing refineries (Del Bianco et al, 1993). Consequently, heavy oils need to be upgraded to increase the H/C ratio, to remove metals and heteroatoms to an acceptable level, and to lower the overall molecular weight (or lower the viscosity) so that the products can be pumped and processed more easily. There are two approaches to increase the H/C ratio: (i) by carbon rejection, which involves heating the oil at near atmospheric pressure, without hydrogen addition (coking processes), and (ii) by hydrogen addition, which involves heating at lower temperature and high hydrogen partial pressures (1600-2000 psig) (hydrocracking processes) (Sanford, 1994). The routes to upgrading heavy crude oil are shown in Figure 2.9.  2.3.1 Carbon rejection Carbon rejection can be done in two different ways. The first involves the production and removal of elemental carbon or coke by pyrolysis. The second is based on the total gasification of the crude, followed by removal of carbon as C 0 (Laine and 2  Trimm, 1982). A substantial part of the fuel is rejected in both methods, as gas or as coke, resulting in the loss of potential liquid hydrocarbons (Dautzenberg and De Deken, 1984).  15  Chemicals and fuels 4  Synthesis gas processing  Impurities removed as e.g. metal oxides, SOj, CO, H , COj N(X 2  gasification (removal of C as COj ) Carbon plus hydrocarbons  Heavy crude pyrolysis (removal of C as coke)  extraction  oils  hydrogenation/ hydrocracking (addition of hydrogen) Smaller hydrocarbons molecules  impurities removed as e.g. metals, H S, N H 3 2  downstream processing Chemicals and fuels  Figure 2.9 Processing of heavy crude oil (Laine and Trimm, 1982).  2.3.2 Hydrogen addition To upgrade heavy oil by hydrogen addition, hydrogen is added to the molecules directly (Laine and Trimm, 1982). Hydrogen addition processes are operated at milder conditions than coking operations (430-460°C compared to >500°C in the case of coking) and have higher liquid yields than coking process (about 85% compared to 70% in the case of coking) (Sanford, 1994), resulting in improved process economics in most cases (Sanford,  1996). This process generally accomplishes  significant  demetallization,  16  reduction, desulfurization, and viscosity reduction. The reactions occurring in hydrogen addition may be represented as follows: Cracking and hydrogen addition: R-R + Hj  =  2RH  Removal of O, N, S: ROH + H ,  RH + rLO  RN +  2H2  RH + NH,  RSH + Hj  RH + ILS  R-M + - H , = 2 ^  RH + M  Removal of metals:  Hydrocracking can be achieved by either catalytic or thermal cracking. In most industrial processes, thermal hydrocracking is used for conversion to lighter hydrocarbons, while catalysts are used to activate heteroatom and metal removal (Dautzenberg and De Deken, 1985). Catalytic treatment gives greater selectivity and, as a result, lower hydrogen consumption. However, the catalyst life is limited due to the deposition of coke and inorganic materials (Laine and Trimm, 1982). Catalytic treatment has the advantage of greater selectivity but the limited catalyst life is a disadvantage. Whether to select catalytic or non-catalytic methods to upgrade the heavy oils depends on the properties of the individual feedstock. In the present study attention is given to the catalytic method because of the higher liquid yield.  17  2.3.2.1 Catalytic hydrocracking There are two types of hydrocracking currently practiced in industry. The hydrocracking of heavy distillate obtained from straight-run refiriing or cracking operations is called distillate hydrocracking whereas hydrocracking the residue from straight-run refining is called residue hydrocracking and these processes are different. The residue hydrocracking process is operated at higher temperature than that of distillate hydrocracking. A different type of catalyst is used in residue hydrocracking due to the high content of metals and asphaltenes in the heavy feed. The thermal mode of cracking dominates in the residue hydrocracking process whereas catalytic reactions dominate the distillate hydrocracking process (Choudhary and Saraf, 1975).  2.3.2.2 Residue hydrocracking The hydrocracking of heavy oils involves simultaneous thermal and catalytic reactions. Heavy oil conversion is generally a thermal cracking of carbon-carbon bonds to form free radical molecules (Miki et al., 1983; Heck et al., 1992). In the presence of hydrogen or a hydrogen donor solvent, and a catalyst, the free radical molecules are stabilized by hydrogen addition, thus coke formation is suppressed (Sanford, 1994). At reaction temperature, for the first 30-40% residue oil conversion, the alkyl aromatic carbon-carbon bonds are thermally cracked, and the breaking of aromatic rings at the P position produces benzylic radicals and alkyl radicals (Smith and Savage, 1991; Sanford, 1994). The cracking of heavy oil results in paraffinic, olefinic, naphthenic, and aromatic molecules within the boiling range of C5 to 524°C, and the yields do not vary  18  greatly whether the hydrogen, hydrogen donor solvent, or the catalyst is present or absent (Strausz, 1989; Heck et al, 1992; Sanford, 1994). At higher levels of residue conversion, the homolysis of the condensed residue produces an aliphatic-carbon radical and an aromatic carbon radical. The aliphatic-carbon radical undergoes further cracking to produce distillate and gas. The aromatic-carbon radical reacts with other similar species and produces coke (Chen et al, 1991; Sanford, 1994). Hydrogen transfer to the reactive aromatic-carbon radical prevents the condensation reaction, resulting in less coke (Heck et al, 1992), and produces a carbonhydrogen bond and hydrogen radical. The hydrogen radical hydrogenates the aromatic ring and transforms it to a more crackable molecule which is further cracked to distillate and gas (Sanford, 1994). Thomas et al. (1982) proposed the generation of hydrogen radical (H») by iron sulfide catalyst as follows: FeS  ->  2  S + H -> 2  FeS + S H» + HS»  HS» + H - > H« + H S 2  2  H S + FeS -> FeS + H 2  2  2  The possible pathways of coking and hydrocracking of a residue molecule in Athabasca bitumen proposed by Sanford (1994) are shown in Figure 2.10.  19  Figure 2.10 Possible pathways for coking and hydrocracking of a residue molecule in Athabasca bitumen (Sanford, 1994).  20  2.4 Deactivation of catalysts Catalyst deactivation by the deposition of coke or metal sulfides is a major problem in residue upgrading. Since the feedstock contains many species which readily form coke under the conditions of reaction, hydrogenation of these species (to saturate double bonds or to favour hydrocracking) is in competition with their coke-forming tendencies. Usually hydrogen pressures are kept high and temperatures are kept as low as possible to minimize coke formation. The deposition of Ni, V and Fe on the sulfided catalyst occursfromthe HDM reaction.  2.4.1 Deposition of coke Coke is a highly carbonaceous material deposited on the sulfided catalyst surface. Coke deposition is a major problem in catalytic heavy oil upgrading, since the feedstock contains many species which readily form coke (Laine and Trimm, 1982). These include polycyclic aromatics, alkyl aromatics, and naphthenes. Various studies on catalyst deactivation (Oelderik et al., 1989; Oballa et al, 1993) have revealed that coke builds up rapidly on the sulfided catalyst during the first few hundred hours of hydroconversion until a steady state value is reached. At steady state, the rate of deposition of carbonaceous material by hydrocracking is matched by the rate of removal of coke by hydrogenation. The progressive deposition of coke and metals on the sulfided catalyst surface is illustrated in Figure 2.11.  21  Figure 2.11 Metal sulfide and carbon deposits on sulfided catalyst as a function of catalyst age (Oelderik et al, 1989).  2.4.2 Deposition of metal sulfides Frequent catalyst deactivation also results from the deposition of metal (as sulfides) on the surface of the catalyst. These metal sulfides are reaction products from the demetallization reactions which occur on the sulfided catalyst surface. The amount of metal sulfide deposited on the catalyst does not reach a steady state, for the following reasons: (i) the metal sulfides cannot escape from the sulfided catalyst surface, and continue to accumulate until the catalyst pores are completely plugged (Oelderik et al., 1989), and (ii) the metal deposition is an autocatalytic reaction. In other words, the deposited metal sulfides catalyze the decomposition of the metal organic complexes (Oelderik et al., 1989).  22  2.5 Heavy oil upgrading processes Heavy oil upgrading processes used today can be classified into four groups (Schuetze and Hofmann, 1984): thermal processes (visbreaking, coking), catalytic hydroprocesses (hydrocracking), non-catalytic hydroprocesses (hydrovisbreaking, donorsolvent processes), and catalytic cracking (fluid catalytic cracking (FCC)). At present most US refiners favour coking since it is a well studied and reliable technology with a relatively low capital cost (Del Bianco et al., 1993). There are fewer technological problems in processing the heavy feeds by coking; however, the distillate yields are low since a large portion of the heavy oil is converted to gas and coke. In addition, distillate quality is poor due to low CCR conversion and poor heteroatom and metals removal. The distillates are further treated over a fixed-bed catalyst to yield synthetic crude oil of acceptable quality for conventional refining (Bearden and Aldridge, 1981; Chen et al., 1988; Del Bianco et al., 1993).  2.5.1 Fixed-bed trickle reactors (FBR) Thefixed-bedtrickle reactor processes have been developed and licensed by Shell, Gulf, and Exxon. The processes are suitable for feed with low metal and heteroatom content. With heavy residual feedstocks, these processes suffer from bed plugging causing high pressure drop across the bed, especially if small catalyst particles are used (Dautzenberg and De Deken, 1984). Lifespan of the catalyst for heavy feed is short (4-6 months). The down time required to replace deactivated catalyst and the cost of catalyst are major drawbacks of thefixed-bedreactor technology.  23  2.5.2 Ebullating-bed reactors (Back-mixed reactor, BMR). H-Oil and LC-Fining hydrocracking processes (Choudhary and Saraf, 1975; Dautzenberg and De Deken, 1984) use ebullated-bed reactors to convert heavy residues and asphalts into lighter fractions. The solid catalyst particles are moving continuously with up-flowing heavy gas-oils. Catalyst and oil are well mixed by a circulation pump. Hydrogen gas bubbles through the catalyst suspension. The contact of catalyst, oil, and hydrogen is extremely efficient, the temperature in the reactor is uniform, and catalyst addition or withdrawal can occur in situ. Relatively small catalyst particles (about 1 mm) compared to those offixed-bedtrickle reactors (about 3 mm) can be applied without excessive pressure drop. However, the rapid deactivation of the catalyst remains a problem 2.5.3 Slurry-phase reactor To overcome the problems that occur due to deactivation of the supported catalysts, an emerging slurry-phase technology for heavy oil hydrocracking using dispersed catalyst has been developed. Slurry-phase technology, or "liquid-fluidized" technology, consists of a gas/liquid/soUd operation in which the catalyst is fluidized by the motion of both gas and liquid. The average catalyst particle size (<0.002 mm) is much smaller than the size of catalyst particles in fixed (1.5 mm length x 3 mm diameter) or ebullating beds (0.8 mm length x 3 mm diameter), which allows for much lower volume fraction loads of the catalyst (1% versus 40% in an ebullated bed) while providing the same catalyst surface  24  area. Slurry-phase operation solves the problem of reactor plugging which occurs in fixedbed operation. The large liquid-phase heat transfer coefficients and high liquid mass (i.e. heat capacity) diminish the risk of temperature excursions and make temperature control of slurry reactors fairly simple. The spent catalyst is included in the pitch or coke residue since the concentration of the catalyst particles is very low and the particle sizes are very small. Slurry phase processes have been developed by Exxon as the M-Coke process, by the Canadian Government as the CANMET process, by UOP as the Aurabon process, and by VEBA Oel as the VEBA Combi Cracking process. These processes have been tested in bench-scale or pilot plant units. The differences in these processes are mostly in the type of catalysts being used as described in the following: M-Coke (micrometallic-coke) Process: the metal compounds (e.g. aqueous phosphomolybdic acid, molybdenum naphthenate and carbonyl compounds) soluble or dispersed in the feed, are thermally decomposed "/'« situ" to the active form of the catalyst. The catalyst is an aggregate of metal sulfides and refractory carbonaceous material. The catalyst loading is 100-10,000 ppm CANMET Process: the catalyst used is pulverized coal impregnated with iron sulfate or other metallic salts. The catalyst loading, 0.5-5 wt% of total feed, is slurried into the oil feedstock and used as a once-through catalyst. Aurabon Process: finely divided vanadium sulfide powder is used as a catalyst and the used dispersed catalyst is recycled to the oil stream  25  VEBA Combi Cracking: red mud powder, an iron-based catalyst, is used as oncethrough catalyst at a concentration of about 2 wt%.  2.6 Dispersed catalysts for heavy oil hydrocracking  2.6.1 Introduction Catalysts used in hydrocracking processes can be classified into two groups: supported catalysts and unsupported catalysts. At present, all the commercial hydrocracking processes use supported catalysts and all commercialized hydrocracking units are fixed-bed processes except for two that are based on ebullated-bed technology, namely, commercial LC-Fining and H-Oil processes (Del Bianco et al., 1993). The key factors of the dispersed catalysts for residue hydrocracking are their small particle size and the high population of catalyst particles per unit volume of reacting oil. The inter-particle distance is an important factor since, during the hydrocracking reaction, the reactive free radical intermediates must be engaged immediately and hydrogenated to prevent the formation of coke and incompatible polymeric oils.  2.6.2 Hydroconversion mechanism in the presence of dispersed catalysts The hydrocracking of heavy oil using dispersed catalysts is based on thermal reactions of the heavy molecules that decompose to unstable fragments and free radicals. The function of dispersed catalysts (sometimes called additives) in the hydrocracking process is to introduce hydrogen to stabilize these crackedfragmentscreated by thermal  26  decomposition and to hydrogenate the free radicals, hence terminating the chain reactions which lead to coke formation. The approach to hydroconversion in a slurry-phase process is different from the conventional process in which large amounts of catalyst were used to directly attack the hydrocarbon feed molecules (reaction 1 in Figure 2.12). In slurry-phase processes, thermally-induced free-radical cracking reactions of asphaltenes and resins (reaction 2) occur, with subsequent hydrogenation of the unstable radicals to oil in the presence of hydrogen and a catalyst (reaction 3). It is very important that the catalyst be highly dispersed in the reacting oil in order to engage the free radical intermediates as quickly as possible at any point in the reactor to prevent further degradation to coke (reaction 4). The highly dispersed catalysts also act as nuclei for metal sulfide deposition.  Hydrogen-deficient large molecules (asphaltenes)  (1) H  :  Hydrogenated large molecules  Unstable-*" (2') Intermediates (radicals, olefins, etc.) *4) Catalyst "Oil" molecules  Coke  Figure 2.12 Slurry-phase heavy oil upgrading reaction mechanism (Dautzenberg and De Deken, 1984).  The M-coke hydroconversion mechanism (Figure 2.13) proposed by Bearden and Aldridge (1981) was also viewed as a thermally induced free radical cracking reaction of  27  heavy hydrocarbons. M-coke, a metal/coke combination, was formed in the feed by a chemical reaction between catalyst precursor and a portion of the feed (reaction A). Large molecules were thermally cracked to free radical molecules (reaction R i ) and hydrogenated to liquid oil by M-coke and hydrogen gas (reaction R ) rather than forming 2  coke via reaction R 3 . Conventional hydroconversion processes use substantial amounts of catalyst and depend more upon direct catalytic attack of the hydrocarbon molecules (reaction R 4 ) . Oil  Oil  H,  Large Molecules I  Radicals *  M M-Coke Figure 2.13 M-Coke conversion mechanisms. Ri= Thermal cracking. R = radical hydrogenation (inhibition). R = disproportionation, polymerization (coking). R4 = conventional hydroconversion. A = chemical interaction (Bearden and Aldridge, 1981). 2  3  Some portions of the heavy hydrocarbons, especially some of the coke precursors or Conradson carbon materials, may not be amenable to thermal cracking followed by hydrogenation. Those heavy molecules may need some hydrogen input via reaction R 4 before thermal cracking occurs. Thus, catalysts that are active for both reactions R and R4 2  are preferred. Mo was chosen in Bearden and Aldridge's (1981) study since they found a number of other catalysts that proved active for reaction R showed relatively poor 2  activity for R 4 .  28  A modified form of the hydrocracking mechanism proposed by Sanford (1994) was proposed recently by Kennepohl and Sanford (1996) as shown in Figure 2.14. According to this mechanism, large hydrocarbon molecules are thermally cracked to radical intermediates (reaction I) and form coke via reaction JJ. Catalyst activates hydrogen to react with an aromatic carbon radical and a cyclohexadienyl type intermediate is formed (reaction DI). This intermediate molecule decomposes through a series of reactions to form distillate and gases. The radical intermediate might be capped with hydrogen via reaction IV and further hydrogenated to the cyclohexadienyl type intermediate via reaction V . Reaction VI is the hydrogenation of large hydrocarbon molecules byfreshsupported catalyst at the early stages of the reaction before the catalyst is deactivated.  Thermal ». H S + Gas + Distillate 2  Figure 2.14 Hydrocracking mechanism proposed by Kennepohl and Sanford (1996).  29  2.6.3 The synthesis of metal catalysts The suitable metal constituents of the metal compounds are selected from the groups IVB, V B , VIB, VHB, and VIII. When the thermally decomposable metal compound is added to the hydrocarbon feedstock, it first disperses in the oil and subsequently is converted to a solid, non-colloidal catalyst (Bearden and Aldridge, 1979).  2.6.4 Effect of precursor solubility on catalyst activity It is important to study whether the nature of the organic portion of the catalyst precursor has any important influence on the activity of the dispersed catalysts. Some researchers have reported only small differences in catalytic activity of different catalyst precursors while others report a large effect of the precursor. Chen et al. (1989) studied the effect of different precursors in hquid phase hydrocracking of Athabasca bitumen. The precursors used in their study were the carboxylates of Ni, Co, Fe, and Sn, the naphthenates of Ni, Co, and Mo, and the acetylacetonates of Ni, Co, Mo, Fe and V. They reported a small effect of the precursors on the product quality and activity of the catalysts. For example, the asphaltene conversions were 52% and 49% for Ni naphthenate and Ni acetylacetonate, respectively, while conversions of 51% for Mo naphthenate and 53% for Mo acetyl acetonate were reported. Zhang et al. (1996) studied the effectiveness of dispersed catalysts in hydrocracking of coal liquefaction extract using different catalyst precursors (carbonyl, octoate, and naphthenate). The results showed that the molybdenum-based catalyst  30  precursors (Mo octoate, Mo naphthenate and Mo(CO) ) gave higher (up to 26%) 6  hydrocracking activity than the other catalyst precursors, e.g. W(CO) , Fe(CO) , Cr(CO) , 6  5  6  and N i octoate, in reducing high boiling-point material (BP >450°C) to lower-boiling fractions. Mo(CO)6 was the most active within the set of metal carbonyls. Similarly, Mo octoate was more effective than N i octoate. However, the differences among Mocontaining catalyst precursors (Mo octoate, Mo naphthenate and Mo(CO)e) were quite small (about 7% compared to 26% among different metal catalysts). Chen et al. (1989) and Zhang et al. (1996) reported small differences in catalyst activity for catalysts derivedfromcatalyst precursors with different organic functions, e.g. Mo octoate and Mo naphthenate. However, Bearden and Aldridge (1981) tested the activity of different molybdenum M-Coke precursors for the hydroconversion of heavy crude oil and their results showed major differences in activity depending on the precursor chosen. Among compounds of molybdenum, molybdenum naphthenate and Mo(CO)6 gave higher desulrurization activity, higher CCR conversion and lower coke yield than micronized M0O3 and M 0 S 2 powder. Curtis et al. (1987) examined the activity of different catalyst precursors in coprocessing: Co and Mo naphthenates and Mo and Ni octoates. Mo naphthenates gave better performance than Mo octoates with higher coal conversion (85.1% compared to 80.6%) and oil production (25.3% compared to 17.1%), and the two Mo catalysts showed higher levels of activity than the Co or Ni catalysts. In summary, the effect of different catalyst precursors on catalyst activity has been studied. Catalyst precursors with different types of metal showed different activities when  31  the same organic attachment was involved, whereas catalyst precursors with different organic fragments but with the same type of metal showed smaller differences in activities. Whether the different catalyst activities were due to the metal type, the particle size, or the dispersion of the catalysts is not well established. One of the goals of the present study was to use well characterized catalysts prepared in reverse micelles to provide a better understanding of the effect of particle size, metal type, and dispersion on catalyst activities.  2.6.5 Active species and size distribution of dispersed catalysts A number of different techniques have been used to characterize the active metal sulfide species and the size distribution of dispersed catalysts. Scanning electron microscopy (SEM) and transmission electron microscopy (TEM) are used to characterize the size of the bulk particle whereas crystallite size and the active species have been determined using X-ray diffraction (XRD), Mossbauer Spectroscopy, and Fourier transform infrared (FTIR) spectroscopy. Many studies using soluble catalyst precursors have demonstrated that the active form of the catalyst is a non-stoichiometric metal sulfide. Kim et al. (1989) characterized the active catalytic species generated from molybdenum naphthenate in the hydrogenation reaction of naphthalene at 380°C. The catalyst crystallites identified by X-ray diffraction were rhombohedral  M0S2,  2.6 nm in  size in the perpendicular direction and about 4.5 nm in the lateral direction.  32  Liu et al. (1994) studied XRD patterns of benzene insolubles formed during residue hydrocracking using oil-soluble molybdenum dithiocarboxylate, molybdenum dithiophosphate  and  water  soluble  ammonium phosphomolybdate,  ammonium  heptamolybdate and ammonium tetrathiomolybdate. The catalytically active phase generated from the Mo catalysts was microcrystalline  M0S2  and the average crystallite  size was 4-5 nm in the perpendicular direction, and 4-7 nm in the lateral direction. Moreover, it was found that the sulfur-containing compounds such as molybdenum dithiophosphate, or the existence of an external source of H S, were favourable in 2  converting the precursor to an active MoS phase. 2  Anderson and Bockrath (1984) and Herrick et al. (1990) reported that the active form of iron sulfide is pyrrhotite, a non-stoichiometric sulfide, Fei. S, where x = 0-0.2. x  The decomposition products of Fe(CO) used for coal/heavy oil co-processing had 5  crystalhte sizes ranging from 5-20 nm characterized by Mossbauer spectroscopy, and 1220.5 nm characterized by XRD. The bulk particle sizes were 100-1000 nm as determined by TEM analysis.  2.7 Mixed catalysts Synergism is sometimes observed when using a mixture of two or more metals. The two metals provide a complementary function resulting in better product quality. Moreover, the use of a second component could also reduce catalyst cost if the concentration of the more expensive component is reduced.  33  Lee et al. (1996) conducted atmospheric residual oil hydroconversion using a dispersed catalyst of Co and Mo in an expanded-bed reactor which was loaded with a certain amount of activated carbon granules. A synergistic effect in the conversions of asphaltenes and vanadium was observed. At a certain catalytic composition a maximum in conversion was observed. In their study, the optimum Co/(Co+Mo) atomic ratio was approximately 0.3. Chen et al. (1989) compared the activity of mixed NiMo catalyst with that of Ni and Mo separately (at the same concentration). The NiMo mixed catalyst was as effective as the Ni catalyst for asphaltene conversion, coke yield and light gas production, but was superior in the reduction of S and O in the asphaltene and maltene fraction.  2.8 Preparation of metal colloids in reverse micelles.  2.8.1 Introduction Even though the catalysts used in slurry-phase technology, for example, those derived from oil soluble Fe(CO)5, Ni naphthenate, Co naphthenate, and Mo naphthenate, are effective for coke suppression, they are difficult to control with respect to their solubility, dispersion in the residue oils, and size of the final metal sulfide catalyst. The sizes of these catalysts have been reported to have dimensions of the order of 1 nm (Lott et al., 1993; Kim et al, 1989; Liu et al, 1994). The new candidate catalysts for upgrading heavy oils in the present study are colloidal metal sulfide catalysts in the nm size range  34  prepared as rnicroemiilsions from water soluble metal salts. In this approach, better control of particle size and dispersion of catalyst in the oil is expected. At present, synthesis of nanometer size particles in reverse micelles is of great interest to scientists since nanoparticles (size range of 1-100 nm) possess a very large specific area and a large proportion of crystalline faces or edges with specific catalytic properties. It is difficult to obtain ultrafine and monodispersed nanoparticles by classical methods since most metallic and inorganic particles sinter easily during the reaction of their precursor species. The water-in-oil microemulsions have been shown to be ideal reaction media for the preparation of ultrafine metal particles (Boutonnet et ai, 1982).  2.8.2 Emulsion and microemulsion Prior to a discussion of microemulsions, a brief discussion of the term emulsion is necessary. If two immiscible liquids such as oil and water are mixed together, they will normally separate rapidly into two distinct phases of large dimension. However, if a surfactant (often called an emulsifying agent) is added to the components and the mixture is vigorously mixed in a blender or homogenizer, an emulsion is produced. The dispersed particles in the emulsion are spheres with particle size in the micrometer range with high polydispersity. The particles are usually visible in a light microscope, and the emulsion looks white when examined visually. Emulsions are twophase systems, with the surfactant adsorbed at the oil-water interface in an oriented monolayer, and they are thermodynamically unstable, eventually coalescing into oil and  35  water layers. Oil may be dispersed in water (o/w), or water may be dispersed in oil (w/o) depending on the specifics of the system, including the temperature (Hiemenz, 1986). Historically, the term microemulsion was applied to systems prepared by adding a fourth component, called a cosurfactant, and generally an alcohol of intermediate chain length, to an emulsion. The usual milky emulsion becomes transparent upon addition of the alcohol. The resulting system consists of either o/w or w/o dispersions with particle diameters in the 10- to 100-nm size range (Hiemenz, 1986). In fact microemulsions do not necessarily have a fourth component. The stabilizing system for a microemulsion must have the proper balance of hydrophile and lypophile character (HLB) which can be achieved in many ways: (1) using a combination of a hydrophilic (high HLB) surfactant (typically anionic) with a hydrophobic (low HLB) amphiphile (typically an alcohol of intermediate chain length) (2) using a single (usually double-chained) ionic surfactant, e.g. didodecyldimethyl ammonium bromide (DDAB) (3) using a single non-ionic surfactant of the poly (ethyleneglycol) alkyl ether type at appropriate temperatures, or (4) using no surfactant as such, but a short chain alcohol capable of strong hydrogen bonding with other molecules of itself and with water, e.g. propan-2-ol. By definition, water-in-oil (w/o) microemulsions are transparent mixtures of water, oil and a surfactant or cosurfactant. The mixture is thermodynamically stable and can form spontaneously (Hunter, 1989). Water-in-oil microemulsions can be visualized as water  36  droplets (diameter 5-50 nm) suspended in oil with the surfactant molecules adsorbed strongly at the interface between the oil and water, generating a barrier against coagulation or coalescence. The common mechanism of formation of a microemulsion involves a marked reduction in the interfacial tension between the oil and water, caused by the surfactant. The differences between microemulsions and emulsions are summarized as follows: The microemulsions are clear; emulsions are cloudy. Microemulsions form spontaneously; emulsions normally require vigorous stirring. Microemulsions are stable with respect to separation into their components; emulsions may have a degree of kinetic stability but ultimately coalesce into two major phases (Hiemenz, 1986). Many researchers have studied the structure and character of microemulsions. Microemulsions are dynamic structures (Lindman et al., 1982) in which there is a constant coalescence, breakdown, and deformation of microemulsion droplets (Leung et al., 1988). Results from nuclear magnetic resonance (NMR) and electron spin resonance (ESR) studies have shown that there is a constant and fast exchange of microemulsion components (e.g. surfactant and cosurfactant) between the interfacial film and the continuous phase (Zana and Lang, 1982). There is a rapid exchange between droplets through collisions and formation of "transient dimers." Such an exchange process and formation of dimers is relevant to the chemical reactions occurring in microemulsions. Microemulsions have low viscosity and can tolerate more added dispersed phase and still remain isotropic, with a minimal increase in droplet size, provided there is sufficient surfactant present for the extra adsorption.  37  2.8.3 Metallic colloid formation in reverse micelles To produce monodisperse, ultrafine particles in microemulsions, the aqueous cores of water-in-oil microemulsions are used as the medium for particle synthesis. In water-inoil (w/o) microemulsions, water is dispersed as nano-size droplets or water pools surrounded by a monolayer of surfactant molecules in the continuous hydrocarbon phase. (Note that in the terminology of the present study, the hydrocarbon phase is referred to as the organic solvent). The surfactant aggregates to form a reverse micelle. The principles of the formation of colloidal particles using microemulsions have been demonstrated by Ravet et al. (1987), Fendler (1987), and Leung et al. (1988) as follows. Firstly, metal salts are dissolved in the water pools of the w/o microemulsion. The water pools containing the dissolved metal ions exchange. Secondly, a reducing agent is introduced in one of two possible ways: either (a) solubilized separately into another w/o microemulsion and the metal ions and reducing agent are brought into contact through the energetic collisions between the microemulsion water pools containing the different reactants, or, alternatively, (b) the reducing agent is introduced directly into the microemulsion, which then diffuses to the water pools containing the metal ions. Thirdly, in the nucleation and growth step, a certain number of metal atoms are required to form a stable nucleus. Only those water pools which contain enough metal ions, reached by the reducing agent, can form a nucleus at the beginning of the reduction nucleation step. The nucleation step is always slower than 'the growth step' which results in a rapid rearrangement of the water pools. The metal ions brought into contact with existing nuclei participate in the growth process but no new nuclei are formed. The basic idea of the  38  growth of ultrafine particles in microemulsions was illustrated by Leung et al. (1988) as in Figure 2.15. In the nucleation process, the number of nuclei formed is proportional to the number of water pools (the amount of water) and depends on the integral of the gaussian distribution function of the number of metal ions per water pool. The lower limit of the integral equals the critical number of metal ions per water pool necessary for the formation of a stable nucleus. In addition, the number of nuclei depends on the diffusion ability of the reducing agent (increases with stronger and higher concentration of the reducing agent) in solution with respect to the rate of rearrangement of the water droplets. In summary the size of the catalyst particles prepared in reverse micelles is affected by the number of nuclei formed in the water pools at the very begirining of the reduction. The number of nuclei formed is a function of the number of water pools containing enough metal ions to form a stable nucleus and reached by the reducing agent. In other words, the size of the particle is affected by the micellar composition (water content), the metal salt concentration and the type and amount of reducing agent.  39  Step 1:  Step 2:  Solubilization of Reactant A  Contact of Different Reactants  (Reducing Agent)  Figure 2.15 Schematic illustration of various stages in the growth of ultrafine particles in microemulsions (Leung et al., 1988). 2.8.4 Factors affecting the size of reverse micelles and colloids  2.8.4.1 Effect of water content Lui and Straub (1984) found that the size of reverse micelles is dependent on the amount of water solubilized in the system As more water is added, the reverse micelles swell to accommodate the additional water; the ultimate result is a single aqueous drop surrounded by a surfactant membrane (Hunter, 1989).  40  Pileni et al. (1985) reported that, for the alkane/sodium bis(2-ethylhexyl) sulfosuccinate (AOT)/water system, the micelles were spheroidal with a linear variation of the droplet size with the amount of water solubilized in the system Robinson et al. (1979) found that the size of the reverse micelles of the AOT/water/heptane system was dependent upon the molar concentration ratio (R = [water]/[AOT]) of water to surfactant, not on the concentration of AOT at a constant R value. The radius of the aqueous cores was about 1 nm at R = 5 and about 2.5 nm at R = 20. Pileni and Lisiecki (1993) reported that the size of copper particles in copper/AOT/water-in-oil micelles increased from 2 to 10 nm upon increasing the water content CO (co = [water]/[AOT]) from 1 to 10. At CO > 10, the average size of the copper particles remained unchanged but the polydispersity increased. A change in the shape of the aggregates from spherical droplets (co = 2) to cylindrical aggregates (co = 4) was induced by increasing co. Nagy et al. (1983) also observed that the higher the water content in the Ni(JJ) solution/cetyltrimethylammonium bromide (CTAB)/n-hexanol microemulsions, the greater the dimensions of the water pools. They also reported that the average size of the N i boride particles increased with increased size of the water pools (or increasing water content). They explained this phenomenon by analyzing the nucleation process. As described in 2.8.3, the number of nuclei formed in the reverse micellar water pools is proportional to the ability of the reducing agent to (iiffuse into the microemulsion solution. At higher water content there is a lower concentration of reducing agent in the water pool  41  and hence a lower rate of diffusion. Results from their experiments confirmed that there was a linear decrease in the number of nuclei per micelle with an increase in water content. A low number of nucleation centers yields larger particles. For a constant metal ion concentration, as the number of nuclei formed decreases, the particle size increases.  2.8.4.2 Effect of metal ion concentration Nagy et al. (1983) reported that the effect of metal ion concentration on particle size is complex. The particle size reaches a rninimum at the optimum metal ion concentration. Ravet et al. (1987) explained this phenomenon by analyzing the nucleation process as follows: at low metal concentration, only few water pools contain the minimum number of ions required to form a nucleus, few nuclei are formed at the very begirining of the reduction, thus there is a small amount of nucleation centers and the size of the metal particles is relatively high. When the ion concentration increases, the number of nuclei formed by reduction increases, resulting in a decrease in the catalyst particle size. At higher concentration, when most of the water pools contain the minimum or more ions required to form the nuclei, the number of nuclei formed remains approximately constant with increasing ion concentration. Therefore, the size of the particles increases again. This was confirmed by the experiments of Nagy et al. (1983). Moreover, they found that in the water/cetyltrimemylammonium bromide (CTAB)/n-hexanol system at about 4xl0" molal 2  NiCl concentration, the number of water pools was independent of either the micellar 2  composition (water content, 8-20 wt%) or the total NiCl concentration. 2  42  2.8.4.3 Effect of reducing agent on the size of the particle. The number of nuclei formed in the nucleation process is affected by the rate of diffusion of the reducing agent which increases with reducing agent concentration (Ravet et al., 1987). However, a large excess of reducing agent does not cause any important effect. Nagy et al. (1983) obtained larger particles when they used a reducing agent (NaBH^ymetal salt (NaCl ) stoichiometric ratio below 3 and found no change of particle 2  size above that ratio. Wilcoxon et al. (1993) studied the effect of different reducing agents on the size of the metal particles. A stoichiometric 3-5 fold excess of the reducing agent was applied in each case to ensure complete reduction. Use of only stoichiometric amounts of these reducing agents slowed the initial reduction step and increased the final cluster polydispersity, as described earlier. They found an increase in the rate of reduction and obtained smaller colloids when they used stronger reducing agents, e.g. NaBHt. Therefore, it is necessary to provide a strong reducing agent in excess for the reduction step in order to increase the nucleation rate and consequently yield smaller particles. It is promising that the size of catalyst prepared in reverse micelles can be controlled and manipulated. In the present study, an initial goal was to produce metals with different particle sizes in the reverse micelles in order to study the effect of particle size on the catalytic activity. Use of micelles allows the size of the catalyst to be well characterized before the hydroconversion reaction takes place.  43  2.9 Catalytic activity of colloids synthesized in reverse micelles Many researchers have reported on the activities of catalysts synthesized in reverse micelles. Colloidal catalysts precipitated on supports or used as microemulsions for model compounds and in coal conversion processes have been described. However, there are no reports on the use of colloidal catalysts prepared in reverse micelles for heavy oil hydrocracking. Ravet et al. (1984) studied the hydrogenation of 1-heptene in a slurry-phasereactor at 22°C, 14.7 psig H , using monodisperse colloidal cobalt boride (Co B) prepared 2  2  in CTAB (cetyltrimemylammonium bromide)/1-hexanol/water reverse micelles, and Co B 2  catalyst prepared in ethanol (reducing CoCl with NaBFL, in ethanol). They found that the 2  CTAB surfactant had a depleting effect on the catalytic activity since the activity of Co B 2  colloids synthesized in CTAB (cetyltrimethylammonium bromide)/1-hexanol/water  (~2  nm diameter) was lower than a catalyst with greater particle size (diameter 250 nm) prepared in ethanol without the surfactant. Boutonnet et al. (1987) investigated the hydrogenation and isomerization of 1butene at a temperature of 20°C, 14.7 psig H , using platinum catalysts supported on 2  pumice  prepared  in  the  microemulsion  PEGDE  (pentaethyleneglycoldodecyl  ether)/hexadecane/H PtCl6.xH 0, in an ethanolic solution of HjPtCLj.Pt™, and by a 2  2  classical impregnation method. The Pt catalyst supported on pumice prepared in the microemulsion was washed with ethanol to remove surfactant and dried at 120°C. All the catalysts were reduced at 200°C under hydrogen flow prior to use. The catalysts prepared in the microemulsion (particle size 2.5 nm) were more active than those prepared in  44  ethanol (particle size 6 nm), and by classical impregnation methods. They observed no poisoning effect of surfactant in the catalytic reaction. Thus, by washing with alcohol followed by pretreatment of the catalyst by heating, surfactant was apparently completely removed. It is interesting to note that Boutonnet et al. (1987) used a nonionic surfactant which is not strongly adsorbed to the catalytic surface while Ravet et al. (1984) used CTAB (cetyltrimethylammonium bromide) in which decomposed N adsorbs strongly to the catalyst surface and poisons the catalyst. Martino et al. (1994) studied the catalytic activity of various metal catalysts synthesized in reverse micelle solutions in coal conversion processes at 400°C, 800 psig H , and compared the results with that of a commercial catalyst. Catalytic activity of the 2  metal catalysts was determined in model compound hydrogenolysis, coal hydropyrolysis, and coaltiquefactionreactions. The hydrogenolysis and coal liquefaction processes, which depend strongly on hydrogen transfer, were affected by the presence of the surfactants (hdodecyldimethyl ammonium bromide (DDAB), and butylethylene glycol n-dodecyl ether (C12E4)),  resulting in a loss of catalytic activity. The activity of a commercial catalyst was  far beyond that of the metal catalyst with surfactant. A moderate catalytic activity was observed when the catalyst was used in a form with no surfactant. They proposed that the loss of catalytic activity was due to hydrogen scavenging of the byproducts of surfactant disintegration at reaction conditions as well as chemical poisoning by surfactant. In coal hydropyrolysis, a process that does not depend on hydrogen transfer, more catalytic activity and less adverse effect of surfactant was observed.  45  Although the catalysts prepared in reverse micelles have small particle sizes, the above studies suggest that surfactant may negatively affect catalyst activity, depending on the type of surfactant. However, in some of these studies, reaction conditions were relatively mild. In the present work, reaction conditions are more severe and it is expected that at these conditions the surfactant will be decomposed. Nevertheless, in examining the effects of metal type or solvent on catalyst performance, the role of surfactant must not be ignored. In the present work, comparisons of this type are therefore restricted to systems with the same surfactant.  2.10 Summary In summary, conversion of residue is governed mainly by thermal reactions. Catalysts, however, play an important role in heteroatom and metal removal from asphaltenes, and in hydrogenation of the asphaltenes to more easily cracked molecules. Deactivation of catalysts is a major problem in heavy oil upgrading. The new approach to heavy oil upgrading, based on a slurry phase reactor, uses dispersed catalysts which have small particle size and low concentration. Although oil soluble catalysts have encouraging performance, the particle size before hydroconversion is not well established and the influence of dispersion on the catalyst performance is not clear. Catalysts prepared in reverse micelles have small particle size, in the nm range, and the organic solvent of the microemulsion has the potential to improve the dispersion of catalyst in the oil phase. The size of catalyst can be manipulated and measured before the hydroconversion reaction. This information is useful in studying the effect of particle size  46  on the catalyst hydroconversion activity. By using catalysts prepared in reverse micelles, the effect of metal type including mixed catalysts, solvent, and particle size on the catalyst activity can be obtained. This information will provide a better understanding of the role of catalysts in heavy oil upgrading.  47  Chapter 3 Experimental Experiments were carried out to assess the heavy oil upgrading performance of metal catalysts prepared in microemulsions using a slurry-phase reactor. The catalysts were prepared in a microemulsion in order to obtain nanometer size catalyst particles. The effect of metal catalyst properties and solvent on the hydrocracking activities of the catalysts was studied. The experiments were divided into three parts: firstly, preparation and characterization of the metal catalysts in the microemulsions was studied, secondly, a study of the solubility effect of solvents on the dispersion of catalysts in the residue oil was completed, and thirdly, activities of the metal catalysts for residue upgrading were determined. In the present chapter, the experimental procedure followed for each part is described.  3.1 Metal catalyst preparation in reverse micelles  3.1.1 Introduction The major components of a water-in-oil microemulsion are water, surfactant, and the oil phase (organic solvent). In the present work, the effect of catalyst properties (particle size, metal type) and solvent on the performance of catalysts prepared in microemulsions was of primary interest. In order to study these effects, stable microemulsions had to be prepared for different microemulsion compositions, different  48  metal types and different solvents. The choice of surfactant was therefore restricted to those that yielded stable microemulsions in each case, as described in detail in the following sections. In the present study the comparisons were made among catalysts that had the same surfactant, as described in Chapters 5 and 6. Besides compatibility between the type of surfactant and solvent, the type of metal salt and reducing agent were also chosen to ensure a stable microemulsion. The continuous oil phase (solvent) of the water-in-oil microemulsions used in the present work was either n-hexane (HPLC grade, Fisher Co.), decahydronaphthalene (decalin, DHN, Fisher Co.), toluene (Fisher Co.), or tetrahydrofuran (THF, Fisher Co.). Following previous studies published in the literature, microemulsions of nonpolar solvents were prepared using non-ionic surfactants. Previously, preparation of microemulsions from water/pentaethylene  glycol  dodecylether (PEGDE)Zhexane  (Boutonnet et al, 1982) and water/PEGDE/hexadecane (Kizling and Stenius, 1987) has been reported. The chemical formula of the surfactant PEGDE is ^ ^ ( C ^ F L O s O H . The aliphatic C12H25 group is hydrophobic and solubilizes in the hexane or hexadecane phase, while the hydrophilic group ((OC H4)sOH) solubilizes within the aqueous phase of the 2  microemulsion.  In the  present  work,  polyoxyethylene-4-lauryl  ether  (PE4LE,  Ci2H25(OC H4)40H, Sigma Chemical Co.), which has similar chemical structure to 2  PEGDE, was chosen for the n-hexane and DHN solvents, whereas the cationic surfactant didodecyldimethylammonium bromide (DDAB, [ C ^ C H z V ^ N ^ C H s ^ B r " , 98%, Aldrich Co.) was used as a surfactant for toluene and THF. The water soluble salts used in the  49  present study were nitrate salts of Fe (Fisher Co.) and Co (Sigma Co.), FeCl .6H 0 3  2  (Fisher Co.) and M0CI5 (Aldrich Co.). The preparation of the catalysts in microemulsions was carried out in an atmospheric environment at room temperature (22°C). The apparatus consisted of a 200 ml glass bottle with a sealed lid, a magnetic stirring bar, and a magnetic stirring motor. The water soluble salts, water, surfactant, and organic solvent were mixed (using the magnetic stirrer) in the sealed bottle for 15-20 minutes or until the water soluble salts were completely dissolved. Subsequently, the metal ions in the water pools of the microemulsion were reduced by adding specified amounts of hydrazine hydrate (IIJNNHJ.XHJO,  Aldrich Chemical Co., 55 wt%  N H4) 2  or LiBH, (2.0 M solution in  tetrahydrofuran, Aldrich Co.) to the microemulsion. The molar ratio of [NiELj]:[metal] and [LiBKU]:[metal] was approximately 5:1 to ensure the complete reduction of the metal species in the microemulsion. The resulting metal particles were surrounded by surfactant, preventing their coagulation.  3.1.2 Hexane The preparation of the water/n-hexane/PE4LE water-in-oil microemulsion followed the method used by Boutonnet et al. (1982). Nitrate salts of Fe and Co were dissolved in deionized water to make stock solutions that were assayed for metal concentration using atomic absorption (assay done by ACME Analytical Laboratories, Vancouver). The equivalent metal concentrations of the Fe  3+  and Co  2+  solutions were  154.25 gFe/L and 197.00 gCo/L, respectively. The Fe or Co solution was then mixed 3+  2+  50  with the n-hexane/PE4LE solution as described previously and subsequently, the Fe or +  Co was reduced by addition of the reducing agent. The water to PE4LE volume ratio of 2+  the microemulsions ranged from 0.044-0.056 and the PE4LE to n-hexane volume ratio was 0.05. The three-component mixtures were transparent and homogeneous. The compositions of the microemulsions prepared in n-hexane are shown in Table 3.1. The ammonium heptamolybdate salt ((NU4)6Mo7024.4H 0) when mixed with 2  water/n-hexane/PE4LE resulted in a cloudy solution. Hence MoCl , a very hygroscopic 5  salt, was used instead.  M0CI5  (0.062 g) was added directly to the n-hexane/PE4LE  mixture at room temperature (22°C) and mixed for about 10-15 minutes or until the salt dissolved completely. In this case no extra water was added to the system since water adsorbed from the outside environment was sufficient to dissolve the  M0CI5  during the  microemulsion preparation.  Table 3.1 Preparation of Fe, Co, and Mo catalysts in water/n-hexane/PE4LE microemulsions. n-Hexane  PE4LE  Metal  Reducing  pH  Colour  Agent  Before/After  Before/After  (ml)  (ml)  Type/Amount  Type/Volume  Reduction  Reduction  50  2.5  •Fe * soH0.14ml  N2IVO.il ml  1/6  Yellow/Brown  50  2.5  *Co sol /0.11 ml  NzHVO.ll ml  6/7  Pink/Yellow-Orange  50  2.5  MoCl /0.062 g (no water added)  L1BH4/O.6 ml  2/6  Yellow-Green/Black  3  2+  n  5  *Note: The equivalent metal concentrations of the Fe and Co sol" were 154.25 gFe/L and 197 gCo/L, respectively. +  +  51  As mentioned previously, the reducing agent also affects the stability of the microemulsion. The transparent yellow of the Fe -solution/n-hexane/PE4LE and the 3+  transparent pink of Co -solution/n-hexane/PE4LE turned transparent brown and orange, 2+  respectively, when reduced with N2H4.XH2O, but turned cloudy when reduced with L1BH4. The transparent yellow-green MoCi5/n-hexane/PE4LE microemulsion turned black when reduced with L1BH4 and cloudy when reduced with  N2H4.XH2O.  The cloudiness of the  mixture indicated that metal particles agglomerated, since the particle size was bigger and scattered more light. The clear transparent mixture was therefore preferred as an indication of small particle size. Therefore and Co salt solutions while  L1BH4  N2H4  was a suitable reducing agent for the Fe  was more suitable for the Mo salt solution in the  water/n-hexane/PE4LE system The microemulsions containing the reduced metal particles were stable for periods of months.  3.1.3 Decalin (DHN) PE4LE was chosen as the surfactant for DHN-based microemulsions. The metals were introduced using the same Fe and Co 3+  3.1.2.  L1BH4  2+  solutions and  M0CI5  as used in Section  was used to reduce the yellow Fe -solution/DHN/PE4LE, the pink Co 3+  2+  solution/DHN/PE4LE, and the green MoCl /DHN/PE4LE, all of which turned black after 5  reduction. In fact, the yellow Fe -solution/DHN/PE4LE, and the pink Co 3+  solution/DHN/PE4LE could also be reduced with  2+  N2H4.XH2O;  however, the stronger  reducing agent, LIBH4, was used to reduce all three metal ions. The compositions of the microemulsions prepared in DHN are shown in Table 3.2.  52  Table 3.2 Preparation of Fe, Co, and Mo catalysts in water/DHN/PE4LE microemulsion. DHN  PE4LE  Metal  LiBFL,  pH  Colour  (ml)  Before/ After  Before/After  Reduction  Reduction  (ml)  (ml)  type/amount  50  2.5  *Fe sot70.14ml  1.0  2/7  Light Yellow/Green-Black  50  2.5  *Co sol70.11 ml  1.0  5/8  Pink/Black  50  2.5  MoClj/0.062 g (no water added)  0.6  1/7  Green/Black  3+  2+  *Note: The equivalent metal concentrations of the Fe and Co sol" were 154.25 gFe/L and 197 gCo/L, respectively. +  +  Mixed metal catalysts were also prepared in DHN-based microemulsions by mixing microemulsions of the reduced metal. The composition of the single metal microemulsions is shown in Table 3.3. NiMo was prepared by mixing the Ni/1.25-ml PE4LE /25-ml DHN with the Mo/1.25-ml PE4LE /25-ml DHN microemulsion whereas CoMo was prepared by rnixing the Co/1.25-ml PE4LE /25-ml DHN with the Mo/1.25-ml PE4LE /25-ml DHN microemulsion. The metal concentration of NiMo and CoMo was 21.75-g metal/50-ml DHN. The total metal content of the mixed NiMo and CoMo catalysts was equivalent to that of the single metal catalysts prepared in the DHN-based microemulsions (i.e. 50 ml DHN and2.5mlPE4LE).  53  Table 3.3 Composition of Ni, Co, and Mo catalysts used for NiMo and CoMo catalyst synthesis.  DHN  PE4LE  Metal  L1BH4  (ml)  (ml)  (Type/Amount)  (ml)  25  1.25  * N i sol70.039 ml  0.5  25  1.25  *Co sol70.055 ml  0.5  25  1.25  MoCl /0.031 g (no water added)  0.3  *Note: Concentrations of N i respectively.  2+  2+  2+  5  and Co sol" were 275.64 gNi/L and 197 gCo/L,  3.1.4 Toluene The FeCl3.6H20/toluene/DDAB microemulsion preparation procedure given by Martino et al. (1994) was followed to prepare microemulsions using toluene as solvent. The transparent red-brown FeCl3.6H 0/toluene/DDAB turned colorless when reduced 2  with L1BH4. An attempt to use Fe(N03) 9H 0 in toluene was not successful. Not all the 3  2  Fe(N0 )3.9H 0 dissolved in the mixture. However, it was possible to use the Co nitrate 3  2  solution in toluene. The transparent light blue Co -solution/toluene/DDAB and the 2+  transparent yellow MoCls/toluene/DDAB turned black after reduction with  L1BH4.  The  compositions of the microemulsions prepared in toluene are shown in Table 3.4.  54  Table 3.4 Preparation of Fe, Co, and Mo catalysts in water/toluene/DDAB microemulsions. DDAB  Toluene  Metal  LiBfL,  pH  Colour  Before/After  Before/After  (ml)  (g)  Type/Amount  (ml)  Reduction  Reduction  50  2  FeCl .6H O/0.1049g (no water added)  1.0  5/7  Red-Brown/No Colour  50  2  Co *sol70.11ml  1.0  6/7  Light Blue/Black  50  2  MoClj/0.062 g (no water added)  0.6  1/6  Yellow/Black  3  2  2+  *Note: Concentration of Co  sol" was 197 gCo/L.  3.1.5 THF The catalyst preparation in THF was similar to that in toluene. However, in THF/DDAB, both FeCl .6H 0 and Fe nitrate solution gave transparent brown solutions, 3  2  but the latter was chosen for activity testing. They became transparent light yellow when reduced with L 1 B H 4 , and cloudy when reduced with N 2 H 4 . X H 2 O . The compositions of the microemulsions prepared in THF are shown in Table 3.5. Table 3.5 Preparation of Fe, Co, and Mo catalysts in water/THF/DDAB microemulsions. THF  DDAB  Metal  L1BH4  pH  Colour  Before/After  Before/After  (ml)  (g)  type/amount  (ml)  Reduction  Reduction  50  2  Fe** sol70.14ml  1.0  6/7  Red-Brown/Light Yellow  50  2  Co *sol70.11 ml  1.0  6/8  Light Blue/Black  50  2  M0CI5/O.O62 g  0.6  3/7  Yellow/Black  2+  (no water added)  *Note: Concentrations of Fe respectively.  and Co  sol" were 154.25 gFe/L and 197 gCo/L,  55  3.2 Preparation of Co catalysts in microemulsions with different water/surfactant volume ratios. In an attempt to produce catalysts with different particle sizes, metal colloids were prepared in microemulsions with different water/surfactant ratios as shown in Table 3.6. The amount of water was 0.11 ml and Co  2+  solution concentration was 3.3 M . The  amount of PE4LE was varied to obtain different water/PE4LE ratios and a specified volume of hexane was added to obtain transparent microemulsions. The procedure followed was the same as that described in Section 3.1.2.  Table 3.6 Microemulsions with hexane, using different water/PE4LE volume ratios. ([Co] = 3.3 M). Hexane (ml) 50 50 23 10  PE4LE (ml) 2.5 2.0 1.1 0.5  Water (ml) 0.11 0.11 0.11 0.11  Water/PE4LE 0.044 0.055 0.1 0.2  N H4 (ml) 0.11 0.11 0.11 0.11 2  3.3 Colloid characterization  3.3.1 Transmission electron microscopy The size of the metal particles present in the microemulsions after reduction with N2H4  or L1BH4 was measured using transmission electron microscopy (TEM). A Hitachi  H-800 electron microscope, operating in the transmission mode at 100 kV, was used for these analyses. The TEM specimens were prepared by depositing a 5 (il droplet of the microemulsion containing the reduced metal from a micropipette onto a gold grid coated  56  with carbon. The specimen was then dried in a vacuum oven at about 200°C and 2 torr for approximately 24 hours to remove water, organic solvent and surfactant. The size of 100-200 individual particles was measured from the TEM micrograph using a vernier caliper from which the number average size (d) and standard deviation (a) of the particles was calculated. The number of particles required to obtain a 95% confidence interval of width 1.0 (mean ± 0.5) and a standard deviation of 2.0 is 62 (Ott, 1988). The calculations of the number of particles required, d and a are shown in Appendix HI.  3.4 Dispersion of colloidal catalysts in residue oil Different solvents have different solubilities in residue oil which should result in different dispersions of the metal catalysts in the residue oil. For example, solvents that solubilize residue oil should have a higher amount of metal associated with the asphaltenes than other solvents that do not solubilize residue oil. In the present study, the amount of catalyst associated with the asphaltene portion of the residue is of interest, since asphaltenes are the major component to be hydrogenated or hydrocracked to obtain higher liquid yield. In addition, most of the heteroatoms and metals associated with the asphaltenes must be removed. A higher amount of catalyst associated with the asphaltene fractions suggests a higher possibility that the catalyst will enhance the hydrogenation and hydrocracking reactions. Cold Lake Alberta residue was used in the present study and the properties of the residue oil are shown in Table 3.7.  57  Table 3.7 Properties of the Cold Lake Alberta residue oil used in the present study. Property  Value  Units  Carbon Hydrogen Nitrogen Oxygen Sulfur Chlorine Ash Nickel Vanadium H/C atomic ratio Specific gravity at 15.6°C Viscosity at 60.0°C THF solids Asphaltenes Conradson carbon residue (CCR) Initial boiling point - 177°C 177 - 350°C 350- 525°C >525°C  83.1 9.4 0.69 n/a 5.75 0.04 0.05 111.2 285.5 1.35 1.038 210500 0.07 24.4 19.58 0.00 0.00 23.75 76.25  Weight% Weight% Weight% Weight% Weight% Weight% ppm ppm mPa.s Weight% Weight% Weight% Weight% Weight% Weight% Weight%  n/a = not available  3.4.1 Procedure Fe colloids were prepared in microemulsions using different solvents (n-hexane, decalin,  toluene,  and tetrahydrofuran),  the  same  surfactant  (Sodium bis(2-  emylhexyl)sulfosuccinate, AOT), and the same Fe concentration. AOT was chosen because it is applicable to all types of solvents used in the present study. Therefore, the effect of surfactant that might occur would be constant and the interpretation of the results  58  in Section 3.4 could be based on the type of solvent alone. In other experiments, the best performance of the catalysts was to be investigated, hence, more suitable surfactants that provide very clear and stable microemulsions for each solvent were chosen, e.g. PE4LE for n-hexane and for DHN, and DDAB for toluene and for THF. The composition of the iron colloids prepared for this part of the study are shown in Table 3.8.  Table 3.8 Composition of iron colloidal microemulsions. Components  Volume (ml)  AOT Ethanol Fe solution N H4 Solvent  2.5 0.5 0.12 0.1 50  1  2  2  3  CH^OjCHjCHCCjHiQH, 'AOT chemical formula, 2  3  N a + 0  > - < » -CftCH CH(QH0C H 2  4  !  Ethanol, a cosurfactant, was added to the microemulsion prepared in toluene only. n-hexane, DHN, toluene, or THF.  Residue oil (25 g) was mixed in a beaker with the microemulsion containing the reduced Fe and stirred for 2 h. n-Pentane (250 ml) was then added to the solution. Asphaltenes, which do not solubilize in n-pentane, were precipitated. The precipitate was left to settle for 2 h. Asphaltenes were then separated from the liquid oil using a vacuum filter (10-15 |im pore size) and washed with 250 ml of n-pentane. The precipitated asphaltenes were then dried at 105°C overnight. The amount of Fe, Ni, and V present in  59  the asphaltenes was determined using inductively coupled argon plasma (ICP) spectrometry and these analyses were done at the Alberta Research Council (ARC), Edmonton, Alberta. In each case, the hydrocarbon sample was diluted with neutral base oil and xylene. The solution was then directly aspirated into the ICP spectrometer. The emission intensities of the elements of interest (Ni, V, and Fe) were compared with standard solutions of known concentrations, previously analyzed. Since Ni and V were the major metals already present in the asphaltenes, the amount of Ni and V was measured as a standard for the amount of asphaltenes precipitated. Mixing times as well as settling times of 2, 4, and 6 h were examined. A blank experiment was also completed in which n-pentane was added to the microemulsion alone. No precipitation of the catalyst particles was observed, suggesting that, in the case of the residue oil/microemulsion mixture, the catalyst particles precipitate because of their coordination to the asphaltenes.  Furthermore, a thorough asphaltene  washing procedure in n-pentane was followed in an attempt to ensure that any catalyst particles not bound to the asphaltenes would not be retained by the filter.  3.5 Catalyst activity measurement An initial study of the activity of the catalysts based on diphenylmethane (DPM) hydrocracking to benzene and toluene, which was easy to monitor using gas chromatography, was performed by Hall (1996) following the procedure of Wei et al. (1992). In the present study the activity of the catalysts for residue upgrading was examined using Cold Lake residue as a feedstock.  60  3.5.1 Experimeiital setup The activity tests were performed in a 300 ml batch reactor (Autoclave Engineers). The reactor consisted of a 300 ml stainless steel vessel, a removable furnace, magnetic stirrer, cooling coil, safety valve, pressure gauge, speed controller, and temperature controller as shown in Figure 3.1. The reactor was pressurized using a 5%H S/95%H gas 2  2  mixture that was fed from gas cylinders to the reactor through appropriate ports in the reactor head.  61  Water outlet  Gas outlet  Water cooling  Thermocouple Safety valve  Speed controller  Temperature controller  Figure 3.1 Experimental setup for activity test.  62  3.5.2 Hydroconversion activity Residue oil (50 g) and 50 ml of microemulsion containing 21.75 mg of metal catalyst were placed in the reactor vessel. The metal loading in the residue oil was approximately 435 ppm The reactor vessel was secured to the body of the reactor via 6 hexagonal screws, and wrapped with a removable electric furnace. The reactor was purged with 5%H S/95%H before being pressurized with the same gas to 500 psig. The 2  2  H/ELS gas was used to ensure that the catalyst was sulfided in situ. Prior to heating the reactant mixture, it was important to cool the bearing of the magnetic stirrer using tap water to prevent hydrocarbon deposition between the stirrer shaft and the bearing. Since solid hydrocarbon deposition was not completely prevented, the stirring assembly had to be removed and cleaned occasionally. The stirrer speed was set at 350 rpm and the reactor vessel was heated to the reaction temperature (430°C) at a ramp rate 5°C/min. The heat-up time was approximately 1 h and 20 minutes and it can be considered equivalent to the hydrogenation step before hydrocracking in a continuous reactor. The final reaction temperature was maintained for a reaction time of 1 hour. During the heat-up and reaction periods, the temperatures of the furnace and reactant mixture in the reactor vessel, and the pressure of the autoclave were recorded at 5 min intervals until the 1 h-reaction time period was completed. After the reaction, the heating element was removed and the reactor vessel was first air-cooled to 300°C and then cooled to room temperature by passing cold water through the cooling coil in the reactor vessel. The sealed reactor was left overnight.  63  The pressure of the autoclave at room temperature was recorded to calculate the amount of gas in the reactor after the hydroconversion reaction. The gas in the reaction vessel was collected in a gas sample bag for analysis by gas chromatography (GC) before the reactor was depressurized and disassembled. The weight of reactant and products was recorded. Solid carbonaceous deposits that adhered to the reactor wall and the impeller were removed by scraping followed by washing with methylene chloride, CH C1 . The 2  2  wash liquid and coke were collected and separated using vacuum filtration. The amount of coke and liquid product recovered was weighed, so that an overall mass balance calculation could be completed. The conditions of the catalyst activity measurements were dictated by the potential application of these catalysts to the HC-3 slurry phase hydrocracking process developed by the Alberta Department of Energy (Lott et al, 1993). In the HC-3 process, the LHSV (Liquid Hourly Space Velocity) is 1 h" and the operating temperature is in the range of 1  430-460°C. Examination of the effect of different operating conditions was beyond the scope of the present study, given the lengthy experimental procedure to prepare and characterize the catalysts and to perform activity measurements using Cold Lake residue feedstock.  3.5.3 Analysis of gas product A Shimadzu gas chromatograph (model GC-8A), with a thermal conductivity detector (TCD) and equipped with a Shimadzu CR 601 Integrator was used for hydrocarbon gas analysis. The chromatograph column (6 ft x 1/8" x 0.085", SS) was  64  packed with Hayesep Q 80/100, which was suitable for separation of hydrocarbons and H S. The temperature of the column and detector was set at 50°C and 130°C, 2  respectively, and the TCD current was set at 90 mA. He carrier gas, used to get accurate hydrocarbon analyses, passed through the column at a flow rate of 30 ml/min and the column temperature was maintained at 50°C. A reference chromatogram was obtained by injection of 1 ml of a standard gas mixture (2.01 mol% CO, 10.0 mol%  CH4,  8.01 mol% C 0 , 2.02 mol% 2  C2H4,  2.00 mol%  C H6, 0.1 mol% C Hs, and balance He) into the GC column. The elution order of the 2  3  hydrocarbon gases determined from the reference chromatogram was: CO, CH4, C 0 , 2  C H4, 2  C H<5, and C3H8. 2  The response factor for each gas component (moles/area), based  on 3 repeat analyses was calculated. Response factors for air and H S were determined 2  similarly using a 1 ml air and 1 ml 5%H S/95%H mixture. CO and air showed the same 2  2  retention time, and H gas could not be quantified when He was used as a carrier gas. 2  Hence the gas components quantified by GC analysis were CH4, C 0 , C H4, C H6 and 2  2  2  H S; the other gases were quantified by difference and lumped as H , CO and air. 2  2  By injecting 1 ml of gas product from the reactor collected in a sample bag, the components present were identified by comparison of retention times. The average area from 3 repeat analyses of each identified peak was converted to the number of moles of component using the response factors. Hence the gas mixture molefractionwas obtained and, using the ideal gas law, the total number of moles of gas product in the reactor vessel was calculated from the pressure measured at the completion of the experiment. The number of moles of gas product multiplied by mole fraction of each gas component  65  yielded the number of moles of each component and this was readily converted to component mass. Hence the total mass of gas product was obtained. The mass of 5%H S/95%H in the reactor before reaction was calculated using a similar approach. The 2  2  gas product calculations are shown in Appendix U.  3.5.4 Determination of coke yield Products from the hydrocracking reaction consisted of solid coke, liquid oil, and gas. Coke was separated from the liquid product via vacuum filtration using a Buchnerglass-filter (10-15 |im) and washed with methylene chloride (CH C1 ). The filtrate was 2  2  refiltered through a filter membrane with pore size 0.2 |im. The filter membrane and the filter funnel with coke were dried in an oven at 105°C overnight. Washings from the impeller and the vessel product were filtered separately. Coke yield, expressed as a weight percent of the feed oil, was based on solid coke recovered from both the reactor and the reactor washings.  3.5.5 Analysis of liquid products Methylene chloride (BP 40.7°C) from the solids washing during filtration was necessarily mixed with the reactor liquid. It was subsequently removed from the liquid product using a rotary evaporator set at 55°C. The total liquid was weighed following CH C1 removal. The amount of liquid product at different boiling points (analysed using a 2  2  simulated distillation technique), microcarbon residue (MCR), and S, were determined by the Alberta Research Council. Asphaltene content was detennined in our laboratory  66  following the procedure developed by Syncrude (Liu and Gunning, 1991). Methylene chloride in the reactor washings was removed at temperatures up to 90°C to ensure that only a very small amount of CH2CI2 would be left in the product.  3.5.5.1 Microcarbon residue (MCR) analysis The MCR analysis followed the ASTM D4530 procedure. The liquid product sample was heated to 500°C in a glass vial under a N atmosphere for a specific time. The 2  sample was cracked and formed coke under these conditions. The volatile portion was removed with nitrogen and the remaining carbonaceous residue (coke, ash-forming constituents, and non-volatile constituents) was weighed and reported as "%microcarbon residue" of the original sample. MCR agrees well with CCR. MCR has some advantages over the CCR test method including improved control of test conditions, smaller sample size, less operator attention, and improved sample throughput. The efficiency of the hydrocracking reactions was reported in terms of MCR conversion as follows: % MCR Conversion = [wt-MCR] - [wt.MCR] [wt.MCR] Feed  Product  ^  Feed  3.5.5.2 >525°C analysis by simulated distillation of petroleum products The method used for this analysis follows the ASTM D2887 procedure. Liquid sample (2 ml) was injected into a gas chromatographic column 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 obtain percent  67  recovered at a given time. The time axis was converted to boiling point by comparison with a calibration curve generated under the same chromatographic conditions. The analysis results were reported as weight% removed at a particular temperature. An examplefromthe simulated distillation is as follows: Analysis results: %Weight removed versus Boiling Point (BP). %Weight removed 1 2 3 60 61 %Recovery %[wt.  BP >525°C]p uct rod  BP (°C) 40 75 88 525 536  = 65.4 = 40  o/o >525°C Conversion = [wt-BP > 525°C] -[wtBP > 525°C] [wtBP>525°C] Feed  Product  x l Q Q  Feed  3.5.5.3 Asphaltene analysis Liquid sample (2 g) were dissolved in 2 ml of toluene to which 80 ml of pentane were added. The sample mixture was kept in darkness and shaken for 5 minutes initially and then briefly shaken at 30 minutes intervals over an elapsed time of 2 hours. The asphaltene was collected on a medium pore (10-15 Lim)frittedglass filter, washed with npentane, dried at 105°C and weighed. % Asphaltene Conversion =  Asphaltene] - [wt. Asphaltene] [wt.Asphaltene] Feed  Product  x l Q Q  Feed  68  3.5.5.4 Sulfur analysis 200 mg of sample were burned in an induction-type furnace to yield SO2. The SO2 was passed into an absorber containing an acid solution of potassium iodide and starch indicator where it reduces iodine by the reaction: S 0 + I + 2H 0 -> H 2 S O 4 + 2HI 2  2  2  The iodate-iodide reaction which regenerates the iodine is KIO3  + 5KI + 6HC1 -> 3I + 6KC1 + 3H 0 2  2  The amount of potassium iodate necessary to deliver iodine and retain a slight blue color in the absorber is a measure of the sulfur content of the sample. % Sulfur Conversion = ^-Sulfur] - [wt. Sulfur] [wt. Sulfuric Feed  Product  ^  69  Chapter 4 Catalyst Preparation and Characterization 4.1 The synthesis of reduced metal colloids In heterogeneous catalysis, the catalyst activity is determined by the available catalyst surface area. In the case of solid dispersed catalysts, higher surface areas are achieved by reducing catalyst particle size. At present, there is much interest in using microemulsions as a novel medium for chemical reactions and for the synthesis of colloidal particles. In the present study, metal catalysts were synthesized in microemulsions using different types of solvent and the activity of these catalysts for hydrocracking of residue oil, was determined. The metal content in the activity experiments was fixed by maintaining the same metal concentration in each microemulsion. However, in order to obtain transparent microemulsions, the type of metal precursor, reducing agent and surfactant were varied depending on the solvent. In the following subsections a qualitative discussion of the reasons for the success in producing transparent microemulsions in different cases is presented, together with a discussion of the effects of surfactant, metal salt and reducing agent on the microemulsion. When using different metal precursors, the catalyst preparation procedure was modified slightly as described in Chapter  3.  For example,  directly to the solvent/surfactant mixture while  Fe(N03)3  or  M0CI5  or FeCk was added  Co(N03)2  was dissolved in  deionized water before addition to the solvent/surfactant mixture. The small differences in  70  the preparation procedure should not change the microemulsion properties in any major way and the distribution of the metal ions in the microemulsions should be the same for all the preparation procedures. The results reported by Wilcoxon et al. (1993) suggested the use of LiBEL, as reducing agent rather than  N2H4.XH2O  in order to obtain smaller particle sizes. Hence, in  all the solvents except hexane, approximately 5 times the stoichiometric ratio of  L1BH4,  a  strong reducing agent, was used to ensure the complete reduction of the metal ions and to obtain the smallest particle sizes. In hexane, as mentioned in Section 3.1.2, use of L1BH4 as reducing agent yielded cloudy solutions and so  N2H4.XH2O  was used in this case. In  DHN, it was possible to use either L1BH4 or N2H4.XH2O as reducing agents for Fe , Co , +3  +2  and N i ion reduction, but LiBFL, was used. 2  4.1.1 Effect of surfactant on particle size In the present study, polyoxyethylene-4-lauryl ether, Ci2H25(OCH CH )40H 2  (PE4LE or  C12E4)  2  was used as a surfactant with hexane or DHN and transparent  microemulsions were obtained. However, when PE4LE was used with toluene or tetrahychofuran, there was salting out of the metal solution. Different types of surfactant were needed to replace the PE4LE when the solvent was toluene or THF. Other studies have shown that the hydrophile-lypophile balance in the microemulsion is critical for stability of metal colloids. Wilcoxon et al. (1993) reported that using a surfactant that is either too hydrophobic (e.g., Ci2H 5(OCH CH )30H, Ci E ) 2  2  2  2  3  or too hydrophilic (e.g., CnH^OC^CTJ^gOH, Ci E ) in octane (Cg) resulted in poor 2  8  71  stability of Au colloids. However, when using  (C12E3)  with a more hydrophobic solvent  such as Ci6, the colloids were stable (indicating that an optimal hydrocarbon number for a given surfactant QEj existed). In addition, increasing the hydrophobicity of the surfactant by using  C12E4  in place of CnEg in Cg, but keeping all other variables constant, reduced  the particle size distribution and resulted in smaller colloidal particles. In toluene and tetrahydrofuran, cationic d^dodecyldimefhylammonium bromide (DDAB), a relatively inexpensive commercially available surfactant, was used to replace the nonionic PE4LE surfactant. DDAB is soluble in aromatic solvents such as benzene, toluene, tetrahydrofuran, and xylene but not soluble in aliphatic hydrocarbons without the addition of large amounts of water or cosurfactants. Micelle size and polydispersity in the toluene/DDAB system should be very small. Martino et al. (1994) reported that the number-average diameter determined by T E M of Fe and Pd particles prepared in toluene/DDAB was 1.5 ± 0.2 nm and 1.8 ± 0.2 nm, respectively.  4.1.2 Salting in and out of surfactants by complexation with inorganic salts In Section 3.1.2 clear and stable microemulsions of Fe/n-hexane/PE4LE and Co/nhexane/PE4LE were obtained when Fe -solution/n-hexane/PE4LE and Co -solution/n3+  2+  hexane/PE4LE were reduced with N2H4.XH2O. However, these microemulsions were not stable and became cloudy when reduced with  L1BH4.  In contrast, the MoCWn-  hexane/PE4LE microemulsion turned black when reduced with  L1BH4  but cloudy when  reduced with N2H4.XH2O. The cloudy phenomenon occurs when the hydrophile-Upophile  72  balance (HLB ) of the nonionic surfactant is changed. Many studies (Schick, 1987; 1  Shinoda and Hirai, 1977) have reported that the HLB is affected by the addition of additives (e.g. electrolytes, alcohol, glycol) as described in the following sections.  4.1.3 Effect of additives on HLB The addition of electrolyte to nonionic microemulsions causes a change in the HLB via salting in or out of the hydrophilic group, and a change in cloud point (the temperature at which the microemulsion becomes cloudy) might result (Schick, 1987). In the present study, a reducing agent was added to the microemulsion to reduce the metal ion to zerovalent metal. Cloudiness was observed when an unsuitable reducing agent was used in the system (e.g.  L1BH4  for reduction of Co in the Co -solution/n-hexane/PE4LE system). 2+  2+  The addition of different additives to a microemulsion affects the micellemonomeric film equiUbrium and the final structure of the stabilizing film at the emulsion interface (Schick, 1987). The HLB value of a nonionic surfactant can be made more Upophilic upon addition of salting out electrolytes (Schick, 1987). Schott (1973) has shown that the salting in of polyoxyethylated surfactants by complex formation between ether linkages and metal ions, as in reaction 4.1, also occurs. RL Me2++ n  ,R2 \  - 2+  0 T ^ " Me 0  . \ R I . n-  (4.1)  HLB number is an empirical scale that expresses the amphiphilic nature of emulsifying agents. The least hydrophilic surfactants have the lowest HLB values. The optimum HLB number for forming an emulsion depends on the nature of the particular system. The HLB of the mixture of 20% sorbitan tristearate (HLB 2.1) and 80% polyoxyethylene sorbitan monostearate (HLB 14.9) is (0.2x2.l)+(0.8x14.9) = 12.3 (Shaw, 1986). 1  73  The competition for water of hydration between an electrolyte (Reaction 4.2 for a divalent cation) and a surfactant (reaction 4.3), is shown as follows:  Me + n W ^ L t  [Me(OH ) ]  2+  2  R l ^  2+  n  ( 4  2 )  R l ^  0 + 2H 0 ^ z ±  0.2H 0  2  R2^  2  (4.3)  R2^  Many studies have reported that the depression of cloud point or salting out from the interaction of inorganic electrolytes with polyethylene glycols is due to dehydration (Schott, 1973). Salting in of nonionic surfactants increases the cloud point. Nitrates of multivalent cations which form stable complexes with the oxygen atoms of the polyoxyethylene chain raise the cloud point and the HLB (Schott, 1973). Since ether groups act as ligands for almost all cations, the salting in should prevail over the salting out by dehydration. As discussed in Section 3.1.2, addition of Fe and Co nitrate salts in aqueous form to the n-hexane/PE4LE system, resulted in transparent microemulsions. However, when the n-hexane/PE4LE systems containing metal ion microemulsions were reduced with LiBFLt, the solution became cloudy. The effect of reducing agent and the cause of cloud point depression or salting out are proposed as follows: The addition of LiBFL, to the microemulsion provides BH4" and L i ions. The +  hydrolysis of BFLf yields metaborate ( B 0 ) and H (Ravet et al., 1987): 2  BH4+2H 0 2  ->  2  B0 " + 4H 2  2  74  Stable rmcroemulsions should be obtained when salting in (coordination of metal ions with the surfactant) dominates. Upon addition of LiBEL the coordination of metal ions (Fe or 3+  Co ) with the surfactant is impeded by the coordination of metaborate with Fe or Co 2+  3+  2+  and this diminishes the salting in of Fe or Co with the surfactant. At the same time, L i 3+  2+  +  (a hard acid, see Section 4.1.4) coordinates strongly to nitrate (a hard base, see Section 4.1.4), competing with its coordination to surfactant (less L i coordinates with surfactant +  and this favors salting out). In addition, the excess B H L V will compete with surfactant for H 0 of hydration. As a result, salting out is dominant, the HLB changes, and the cloud 2  point is depressed, explaining the cloudiness observed when LiBELj was added to the Fe  3+  and Co -solution/n-hexane/PE4LE microemulsions. 2+  In a similar way, the addition of  N2H4.XH2O  as a reducing agent to the metal ion  microemulsion (e.g. Fe and Co ) causes the reduction of metal ions, releases N gas, 3+  2+  2  and produces stable and clear microemulsions containing metal particles: -»  N + 4 H 0 + 4e"  Fe + 3e"  ->  Fe°  Co + 2e  -»  Co  N2H4  + 4OH-  3+  2+  2  2  0  ha this system surfactant was stabilized by hydration as in reaction (4.3). Hence no special effect was taking place and no salting out was observed, the microemulsion containing metal particles remained stable.  75  4.1.4 Hard and soft acids and bases In predicting the relative stabilities of complexes between the metal and ligands, the concept of acid-base interaction is applicable. Metal ions are empirically classified as 'hard acids' and 'soft acids' depending on the strength of their complexes with certain ligands. Type A (hard acid) metals are those with smaller ions from Groups I and II, the left hand side of the transition metals, those with high oxidation states, and those that form the most stable complexes with nitrogen and oxygen donors (e.g., ammonia, amines, water, ketones, and alcohols), and with F" and Cl". Examples of type A metals are H ", L i , 1  +  Na , Be , and Mg . Type A metals are small and not very polarizable. They prefer +  2+  2+  ligands that are also small and not very polarizable. The ligands are called 'hard bases', for example, NH , H 0 , OFT, N0 ", and Cl". Type B (soft acid) metals are thosefromthe right 3  2  3  hand side of the transition series, and transition metal complexes with low oxidation states e.g., Pd , Pt , Au , and [Fe(CO)5J. Type B metals are large and polarizable, and they 2+  2+  +  prefer the ligands that are large and polarizable. The ligands are called 'soft bases'; FT, CN", and CO are examples of soft bases. Hard acids prefer to react with hard bases, and soft acids react with soft bases (Lee, 1991). In Section 3.1.2 LiBKLt was used to reduce Mo  5+  in the n-hexane/PE4LE system.  Since all inorganic cations except Na , K , and NFL* form complexes with ethers (Schott, +  +  1973), L i , a strong acid, is likely to coordinate with the ether linkages of the surfactant and cause salting in. Cl" (from M0CI5) has the ability to act as a ligand for the central ion in competition with the ether linkages of the surfactants, decreasing the association of ether and cation (Fe or Mo ) and thus reducing the extent of salting in. Cl" sequesters metal 3+  5+  76  ions and competes for coordination to surfactant. Without the addition of L i , the salting +  out from Cl" dominates, as was the case when N2H4.XH2O was used as the reducing agent. Hence, as observed in the present work, the solution becomes cloudy. When hexane was replaced by DHN the surfactant was more solubilized in DHN and this raised the cloud point. Consequently, the system maintained stability at room temperature whether LiBFL, or  N2H4  was used as reducing agent. The stability of the  system is governed by the hydrophile-hp ophite balance (HLB) of surfactant, solvent, and additives. Surfactant molecules with a good HLB exhibit a strong solubilizing power. The system is unstable if the surfactant is too hydrophobic (e.g.  C12E3)  or too hycfrophilic (e.g.  Ci E ) (Wilcoxon et ai, 1993). 2  8  4.2 Characterization of the metal colloids The TEM micrographs of the Co, Fe, and Mo particles present in the metal/nhexane/PE4LE microemulsions after reduction are shown in Figures 4.1-4.3. The size distributions are shown in Figure 4.4. The calculationfromthe micrographs revealed that Co had a number-average particle size (d) of 5.3 ±0.1 run. Fe and Mo had larger d (6.6 ± 0.2 nm, and 7.7 ± 0.1 t i m , respectively). If one considers merely the effect of water/surfactant ratio, Mo should have the smallest d since no water was added in this case, and Fe should have the biggest d. The water/surfactant volume ratio of Fe was 1.3 compared to 1.0 for Co.  Clearly water/surfactant ratio was not the major factor  determining particle size. The different chemistry associated with each metal likely plays a  77  role and the different particle sizes may be due to, for example, different solvation of the ions at the interface and/or different rates of nucleation in the reduction process.  78  Figure 4.1 TEM micrograph of Co particles in hexane/PE4LE (d= 5.3 nm, a = 0.9 nm).  79  Figure 4.2 T E M micrograph o f Fe particles in hexane/PE4LE ( d = 6.6 nm, a = 1.9 nm)  Figure 4 . 3 T E M micrograph o f M o particles in hexane/PE4LE ( d = 7.7 nm, a = 1.6 nm).  81  Mo-Hexane: Avg. Particle Size 7.7 nm  Mo-Hexane:Avg. Particle Size 7.7 nm  Fe-Hexane:Avg. Particle Size 6.6 nm  Fe-Haxaie: Avg. Partide Sze 6.6 nm  25 3.5 45 5.5 6.5 7.5 &5 9.5 10.5 11.5 125 13.5 14.5  2.5 3.5 4.5 5.5 6.5 7.5 8.5 9.5 10.5 11.5 125  Partide Size (nm)  Partide Size (nm)  Co-Hexane:Avg. Particle Size 5.3 nm  Co-Hexaie: Avg. Particle Size 5.3 nm  o  1 1  25 3.5 4.5 5.5 6.5 7.5 a5 9.5 10.5 11.5 12.5 13.5 14.5  -+- -+- -+- -+- -+'..}  3.5  4.5  Particle Size (nm)  (a)  5.5  6.5  7.5  8.5  9.5  10.5 11.5 12.5  Particle Size (run)  (b)  Figure 4.4 Particle size distribution of Mo, Fe and Co in water/n-hexane/PE4LE microemulsions (a) histogram and (b) cumulative distribution curves.  82  The effect of solvation of different ions (Co , N i , and Fe ) on the size of metal boride particles synthesized in CTAB (cetyltrimemylammonium bromide)-l-hexanol-H 0 2  reverse micelle was studied by Ravet et al. (1987). The concentration and coordination of the metal ions within the differently sized, surfactant-entrapped water pools influenced the mechanism of metal boride formation in the reverse micelles. Ravet et al. (1987) studied the solubilization sites of ions in the water pools of the micelles using C-NMR. In 13  CTAB/l-hexanol/water, the N i  2+  and Co ions are located in the inner water cores of the 2+  micelles quite close to the interface as shown in Figure 4.5. On average, one hexanol molecule is included in the first coordination shell of Co ions, while one or more hexanol 2+  molecules participate in that of N i  2+  ions. The N i  2+  ions, being multiply coordinated with  hexanol at the interface, have lower mobility and hence the probability of collision between two reduced N i atoms, required to form a nucleus, is also lower. Consequently, the rate of nucleation is higher for cobalt boride than for nickel boride particles and Ravet et al. (1987) reported that the size of C02B was smaller than that of M2B. In the present study, considering the hard and soft acids and bases principle, M o  5+  is a harder acid than Fe or Co . This implies that bonding between Mo and surfactant is 3+  2+  stronger because the oxygen coordination site is a hard base. Therefore, there are more coordination sites between the surfactant and Mo and hence lower mobility of Mo compared to Co or Ni. The probability of collision between Mo atoms is less resulting in a lower nucleation rate and a larger particle size. In the periodic table the order of ionic radius is: Fe > Co 2+  2+  > N i . Fe is a harder acid than Co 2+  3+  2+  hence, similarly, the Fe  particles grow bigger than the Co particles.  83  Hexanol H 0 2  CTA*  Hexanol H2O  "** " H  2  C  (")  o  0 f " ' H 0  Hexanol  7 7 ? Hexanol / CTA*  2  ^Hexanol  Water core  Interface  Organic phase  Water core  Interface  Organic phase  Decanol  'Decanol Br"  \  •  Triton  Triton  \  H2O  — Decanol  H  •  j  C o (IJ)  2°  Triton  Triton  I  / Water core  .Decanol  Interface  Triton  Organic phase  Water core  C o (If)  Decanol  Organic phase  I  Triton  T r i t o n  Interface  Figure 4.5 Models of solubilization of Ni(H), Co(n), and Fe(m) ions in the inner water cores of microemulsions (Nagy et al., 1986).  84  Particle size data were obtained by TEM of the hexane based microemulsions only. The residue oil hydroconversion activity of the characterized Fe, Co, and Mo metal catalysts prepared in n-hexane based microemulsions was sufficient to deduce the effect of different metal types and particle sizes on residue upgrading. Since the catalysts prepared in hexane-based microemulsions were stable (for months) and gave good TEM micrographs, catalyst characterization was less problematic with these samples. The  average  metal particle size in n-hexane/PE4LE and DHN/PE4LE  microemulsions could be different since different solvents and reducing agents were used (N2H4.xH.2O was used in n-hexane/PE4LE and LiBEL; in DHN/PE4LE microemulsion). However, the relative size of Fe, Co, and Mo metal particles prepared in DHN should follow a similar trend as was obtained for hexane-based microemulsions, i.e. Mo > Fe > Co, since PE4LE surfactant was used in both microemulsions and the relative strength of interaction between the metal ions and the surfactant should be similar in hexane/PE4LE and DHN/PE4LE microemulsions. The study of hydroconversion activity using a single type of catalyst with different particle sizes would reveal the effect of surface area on catalytic activity. Co catalysts prepared using different water/surfactant volume ratios in order to obtain different particle sizes, by TEM, are described in the following section.  85  4.3 Characterization of reduced metal colloids in reverse micelles with different water/surfactant volume ratios. The water/surfactant volume ratio of the microemulsions was varied from 0.044 to 0.2 in order to obtain Co catalysts with different particle sizes, as shown in Table 4.1. The TEM micrographs of the reduced catalysts are shown in Figures 4.6-4.9 and the size distributions are shown in Figures 4.10 and 4.11. The Co particle size data from the present study and the hydrodynamic diameter of reverse micelles measured using lightscattering by Hall (1996) are shown in Figure 4.12. These data show that even though the diameter of the reverse micelles increased with increased water/surfactant volume ratio, the Co particles prepared in these micelles showed no trend in average particle size with water/surfactant ratio.  Table 4.1 The number-average particle size of Co catalysts prepared in microemulsions with different water/surfactant volume ratios in the Co -solution/hexane/PE4LE system ([Co ] = 3.3 mol/L). 2+  2+  Water/PE4LE  Average Particle Size ± Std Error of Mean  Volume Ratio  (nm)  0.044  5.3 ±0.1  0.055  5.1±0.1  0.1 0.2  5.9 ± 0.1 5.1 ± 0.1  86  Figure 4.6 Micrograph of Co prepared in 0.044 water/PE4LE volume ratio (d= 5.3 nm, a = 0.9 nm).  87  88  Figure 4.8 Micrograph of Co prepared in 0.1 water/PE4LE volume ratio (d= 5.9 nm, <y = 1.1 nm).  89  Figure 4.9 Micrograph of Co prepared in 0.2 water/PE4LE volume ratio (d= 5.1 nm, a = 0.8 nm).  90  Water/PE4LE: 0.2 Avg Particle Dia = 5.1 nm, Std Dev = 0.8 nm  Water/PE4LE: 0.2 Avg Particle Dia = 5.1 nm, Std Dev = 0.8  Water/PE4LE: 0.1 Avg Particle Dia = 5.9 nm, Std Dev = 1. lnm Water/PE4LE: 0.1 Avg Particle Dia = 5.9 nm, Std Dev. = 1.1 nm  25 2.5  3.5  4.5  5.5  6.5  7.5  8.5  3.5  4.5  5.5  6.5  7.5  &5 9.5  10.5  Particle Size (nm)  Particle Size (nm)  (a)  (b)  Figure 4.10 Particle size distribution of Co in water/n-hexane/PE4LE microemulsions with water/PE4LE volume ratios = 0.2 and 0.1 (a) histogram (b) cumulative distribution curve.  91  Water/PE4LE: 0.055 Avg Particle Dia = 5.1 nm, Std Dev =1.1 nm Water/PE4I£: 0.055  Particle Size (nm) Particle Size (nm)  Water/PE4LE: 0.044 Avg Particle Dia = 5.3 nm, Std Dev = 0.9 nm  Water/PE4LE: 0.044 Avg Particle Dia = 5.3 nm, Std Dev = 0.9 nm  1 OH  2.5 3.5 4.5 5.5 6.5 7.5 10.5  Particle Size (nm)  (a)  Particle Size (nm)  (b)  Figure 4.11 Particle size distribution of Co in water/n-hexane/PE4LE microemulsions with water/PE4LE volume ratios = 0.055 and 0.044 (a) histogram (b) cumulative distribution curve.  92  25 A Reverse micelle • Colloidal particle  20 + 15 + A  A  10 + 5+  0.03  0.06  0.09  -+-  -+-  0.12  0.15  0.18  0.21  Water/PE4LE Ratio  Figure 4.12 Effect of water/PE4LE volume ratio on the hydrodynamic diameters of reverse micelles and colloidal particles in the water/n-hexane/PE4LE system (temperature, 22°C, [Co ] in reverse micelle and in colloidal particle = 3.1, and 3.3 mol/L, respectively). The hydrodynamic diameters of the reverse micelles were obtained from Hall (1996). 2+  Many studies have reported different particle sizes at various water/surfactant ratios (Nagy et al., 1983 and Pileni and Lisiecki, 1993). Pileni and Lisiecki (1993) prepared Cu particles in alkane/sodium bis(2-ethylhexyl)sulfosuccinate (AOT)Avater using hydrazine as a reducing agent. By varying the water/surfactant ratio (co = [H 0]/[AOT]) 2  from co = 1 to co = 10, the particle size was increased from 2 to 10 nm. At water contents > 10, no changes in the particle size was observed, but the polydispersity increased. Pileni  93  and Lisiecki (1993) explained the increase in particle size upon increase of water/surfactant ratio in terms of the interfacial water structure as follows: at low water content, copper ions associated with surfactant molecules were not totally hydrated and few copper ions participated in the chemical reduction resulting in a smaller number of copper atoms to form a nucleus. An increase in water content induces an increase in the number of copper ions reduced by hydrazine, resulting in a higher number of copper atoms to form a nucleus and this favors the growth of the particles. At relatively high water contents, copper ions were totally hydrated and free water molecules were present. The difference in the energies between the hydration energy and the electrostatic interaction energy between the head polar groups of the surfactant and copper ions remained constant upon increasing the water content, and the particle size remained constant. The results of the present study and of Martino et al. (1994) showed no trend of increasing particle size when the water/surfactant ratio was increased. Martino et al. (1994) prepared Fe catalyst particles using FeCl .6H 0, a water soluble salt, 3  2  toluene/DDAB using LiBFLi as a reducing agent. The number-average diameter detennined by TEM of Fe was 1.5 ± 0.2 nm and the Fe particle size showed no discernible trend in the range of 0.001M-0.01M FeCl .6H 0 and 1-10 wt% DDAB in toluene. The 3  2  differences in the results of Co particle sizes at various water/PE4LE volume ratios using hydrazine as a reducing agent in the present study as well as the study of Martino et al. (1994) from that of Pileni and Liesiecki (1993), could come from differences in the microemulsions (e.g. surfactant), metal type, or the range of water/surfactant volume ratios studied. The interaction between Fe  3+  and DDAB in the study of Martino et al.  94  (1994) and the interaction between Co  and PE4LE in the present study, could be  different from the interaction between Cu and AOT. The differences in the energy +  between the hydration energy and the interaction energy between surfactant and metal ions in the present study and in the study of Martino et al. (1994) was probably unchanged upon increasing the amount of surfactant (or water/surfactant volume ratio) resulting in no effect of water/surfactant ratio on the particle growth for the range studied. Unfortunately, the variation in Co catalyst particle size was small and the hydroconversion activity of the catalysts prepared from different water/PE4LE volume ratios was therefore not investigated further. It is noteworthy that recently, Rueda et al. (1997) reported no correlation between the BET surface area of unsupported molybdenum sulfides and their catalytic activity in the reaction of thiophene hyd^odesulfurization.  4.4 Summary In summary, Fe, Co, and Mo catalysts were successfully prepared in four types of solvent; n-hexane, DHN, toluene, and tetrahydrofuran. In each case, suitable choice of the surfactant and reducing agent had to be made. The Fe, Co, and Mo particles prepared in nhexane were characterized by TEM. The average particle size of the catalysts was approximately 5-8 nm and decreased in the order: Mo > Fe > Co. The size of the Co catalysts prepared in microemulsions having different water/surfactant volume ratios (w = 0.044 to 0.2) was 5.1-5.9 nm. The results of the present study showed no clear trend of increasing particle size with increasing water/surfactant volume ratio.  95  Chapter 5 Activity Measurements 5.1 Introduction The second phase of the present research involved a comparative study of the catalytic activity of the Fe, Co and Mo catalysts prepared in reverse micelles and sulfided in situ  during hydrocracking. Cold Lake residue oil was used as the feedstock. In addition,  mixed catalysts of CoMo and NiMo were examined to evaluate possible synergistic effects between these metals. The activities of the catalysts prepared in reverse micelles were also compared to the activities of catalysts prepared by the decomposition of conventional organometallic catalyst precursors. In the present chapter, the results from these studies are described and discussed.  5.2 Dispersion of colloidal catalysts in residue oil. The microemulsion technique used in the present study to prepare the colloidal catalysts, resulted in metallic particles dispersed in an organic solvent. Prior to deterrriining the activity of these catalysts, it was important to establish the effect that the solvent may have on the dispersion of the catalyst when the microemulsion was mixed with the residue oil. The test procedure followed that described in Section 3.4. The catalyst dispersed in the microemulsion was mixed with the residue oil. The asphaltene fraction of the residue oil was then precipitated using n-pentane. If the metal catalyst was well dispersed and presumably strongly coordinated to the asphaltene fraction, it was expected that the  96  catalyst would precipitate with the asphaltenes. Conversely, the catalyst would be found in the filtrate during the filtration process if it were not coordinated to the asphaltenes. The pore size of the fritted glass funnel used to filter the asphaltenes was 10-15 |im. Hence, the catalyst particles (diameter < 10 nm) should not be filtered out of the oil if they were not strongly coordinated to the asphaltenes. Catalysts might be trapped in the asphaltene molecules during filtration, however, the solid asphaltenes remaining on the filter were washed several times with n-pentane to remove the pentane-soluble compounds. During washing the solid asphaltenes were also stirred manually in the filter funnel in order to remove the free catalyst particles. The anionic surfactant, AOT, was used for all solvents (n-hexane, DHN, toluene, and THF) so that solvent is the only variable to be considered in this set of experiments. Since AOT is a suitable surfactant for the non-polar solvent n-pentane, addition of npentane during asphaltene precipitation should not affect the microemulsion stability and the free catalyst particles should remain dispersed in the oil phase and not precipitate unless co-ordinated to the asphaltenes. Indeed, a blank experiment in which n-pentane was added to the Fe microemulsion showed no evidence of Fe precipitation. The catalysts used for residue upgrading in the present study were the transition metals Fe, Co, Mo and Ni. These transition metal atoms engage in dative bonding by sharing an electron pair that is localized on another atom Transition metal atoms with partially filled d orbitals require the additional electrons to fill the d-subshell following the 18-electron rule (other atoms generally follow the octet rule). The transition metal forms bonds with ligands possessing available lone pairs (e.g., R-O-R, R-S-H, and R-S-R), or to  97  ethylene, arornatics, and heterocycles which are the components of asphaltenes and residue oil. These molecules coordinate to the transition metal atoms either through their delocalized n molecular orbital electrons, or via localized electron lone pairs (Albert and Yates, 1987). The metal dispersion in the residue or in the asphaltenes may affect the hydrocracking activity of the catalysts, and the different solvents used in the microemulsion may result in different dispersions of metals in the residue or asphaltenes. It is expected that the more the metal catalysts coordinate to the asphaltenes, the higher the probability that hydrogen atom transfer to asphaltenes will occur. Fe was chosen as the catalyst for the catalyst dispersion tests because of its ease of preparation. Preparation of Mo was more difficult, primarily because of the very hygroscopic nature of the M0CI5 catalyst precursor.  5.2.1 Effect of mixing and settling time The residue oil was mixed with the microemulsion in a beaker and stirred to promote the coordination of the metal catalyst to asphaltenes. n-Pentane was added to the mixture and the asphaltenes, which do not solubilize in n-pentane, were precipitated. The precipitate was then left to settle. The mixing and settling times were set at 2, 4, and 6 hours to study their effect on the ability of the catalyst to coordinate to the asphaltenes. The resulting Fe, N i and V contents in the asphaltenes, at different mixing and settling times, are shown in Tables 5.1 and 5.2 for the toluene-based microemulsions. The percentage of Fe recovered in asphaltene was lowest in DHN (61%) and highest in THF  98  (92%). The amount of asphaltene precipitated was different when using different solvents, however, Ni and V, precipitated with the asphaltenes, were used as a reference to correct for this difference, basing the analysis on Fe/Ni and Fe/V ratios. An analysis of variance of the Fe/Ni and Fe/V ratios were performed and the calculations are shown in Appendix UI. The analysis of variance showed that there was no significant difference in the Fe/Ni and Fe/V ratios of the asphaltene fraction for the range of mixing and settling times studied. Thus, a 2-h mixing and 2-h settling time was sufficient for Fe to associate with the asphaltenes.  Table 5.1 Fe, Ni, V, Fe/Ni, and Fe/V contents of asphaltene fraction of residue oil when mixed with Fe/toluene/AOT/ethanol microemulsion catalyst at different mixing times and a 2-h settling time.  Solvents  Mixing Time  Metal Concentration in Asphaltene Fraction  (h)  (ppm) Fe  Ni  V  Fe/Ni  Fe/V  None  2  61  347  907  0.18  0.07  Toluene  2  2178  276  889  7.89  2.45  Toluene  2  2515  355  931  7.08  2.70  Toluene  4  2168  334  820  6.49  2.64  Toluene  6  2118  361  934  5.87  2.27  Note: The total amount of Fe added in the residue oil was equivalent to an Fe concentration in the asphaltenes of 3034 ppm. The repeatibility of the analytical procedure was tested in a sample product from one experiment (Fe-toluene, with 2-h mixing time and 4-h settling time). The sample was  99  divided into two portions and the Fe, Ni, and V contents were determined. The Fe/Ni of the two portions was 7.98 and 7.65, and the standard error of the mean was 2.1% whereas, the Fe/V of the two portions was 3.04 and 2.91 and the standard error of the mean was 2.1%. Table 5.2 Fe, Ni, V, Fe/Ni, and Fe/V contents of asphaltene fraction of residue oil when mixed with Fe/toluene/AOT/ethanol microemulsions at different settling times and a 2-h mixing time. Settling  Metal Concentration in Asphaltene Fraction  Time  (ppm)  (h)  Fe  Ni  V  Fe/Ni  Fe/V  2  2178  276  889  7.89  2.45  2  2515  355  931  7.08  2.70  4*  2921  366  961  7.98  3.04  4*  2840  371  975  7.65  2.91  6  2699  349  918  7.73  2.94  Note: * Sample product from the same experiment was divided into two portions and analysed. The repeatibility of the catalyst dispersion experiment was also determined by repeating the Fe dispersion experiment for catalysts prepared in both the toluene and hexane microemulsion with a 2-h mixing time and a 2-h settling time (Table 5.3). The standard error of the mean of the Fe/Ni ratio in hexane and toluene was 1.6% and 5.4%, respectively, whereas for Fe/V, values of 11.5% in hexane and 4.9% in toluene were obtained. In toluene, the experimental standard error of the mean (5.4% for Fe/Ni and  100  4.9% for Fe/V) was about 3% higher than the analysis standard error of the mean (2.1% for Fe/Ni and 2.1% for Fe/V). The calculation of the standard error of the mean is shown in Appendix III.  Table 5.3 Dispersion of Fe (prepared in different solvents), Ni, V, Fe/Ni, and Fe/V in asphaltenefractionof residue oil following 2-h mixing and 2-h settling. Solvents  Metal Concentration in Asphaltene Fraction (ppm) Fe  Ni  V  Fe/Ni  Fe/V  n-Hexane*  2485  343  1165  7.24  2.13  n-Hexane*  2499  356  930  7.02  2.69  DHN  1843  248  863  7.43  2.14  Toluene+  2178  276  889  7.89  2.45  Toluene+  2515  355  931  7.08  2.70  THF  2804  377  1007  7.44  2.78  * Replicate runs in hexane +Replicate runs in toluene 5.2.2 Effect of solvents The dispersion of the metal catalysts in the asphaltenes in different solvents was measured following 2-h mixing and 2-h settling. The results are shown in Table 5.3. An analysis of variance was performed to test, at the 95% level of significance, whether the differences among the Fe/Ni and Fe/V sample means were significant (Appendix HI). The results from this analysis showed that the differences in Fe, Fe/Ni and Fe/V ratios, among the different solvents, were not significant.  101  In conclusion, the difference in dispersion of catalyst metal due to the effect of different solvents was negligible. Hence any effect of solvent on the hydrocracking activity of the catalysts should be interpreted in terms of the chemical nature of the solvent rather than in terms of the influence of solvent on the metal dispersion in the residue oil.  5.3 Residue upgrading activity of Fe, Co, and Mo catalysts prepared in reverse micelles. The catalyst activity measurements were made in a batch reactor in which, for each test, 50 g of vacuum residue oil was charged into the autoclave together with the microemulsion based catalyst. The autoclave was sealed, purged in a flow of 5%H S/95%H and pressurized to 500 psig using the same gas mixture. The autoclave 2  2  was heated at a rate of 5°C/min until the temperature reached 430°C and was then held at this temperature for 1 h. During the heating and reaction period, the hydrogen sulfide reacted with the metal catalyst and this in situ sulfided catalyst promoted the hydrocracking of the residue. The autoclave pressure at different times in each experiment was recorded and is shown in Appendix I. The different solvents and catalysts resulted in differences in the total heating time to reach the reaction temperature and different final autoclave pressures after the 1 hour reaction time. The major differences resulted from the different physical (e.g. boiling points) and chemical properties of the different solvents used in preparing the microemulsions. In hexane (BP = 69°C), the time to reach 430°C was in the range 107-110 niinutes whereas in toluene (BP = 111°C) it was 80-82 minutes. The performances of the catalysts were measured in terms of gas yield, coke yield and the liquid product quality, i.e. the conversion of S, MCR, >525°C fraction, and  102  asphaltenes. The hydroconversion activities of the different catalysts in different solvents were determined, initially emphasizing the coke yield.  5.3.1 Effect of catalyst on gas yield The composition of the gas in terms of C H 4 ,  C 0 , H S, H and other  C2H4, C H 6 ,  2  2  2  2  undetected gases is shown in Table 5.4. For all catalysts, the major gas component was CH4,  followed by  C H6. 2  In general, the composition of the gas when using different  solvents and catalysts did not change in any major way when the reaction occurred at the same reaction conditions. One exception, however, was the higher C 0 content of the gas 2  produced when THF-based microemulsions were used. Table 5.4 Gas products from residue upgrading (50 g residue with 50 ml metal catalyst prepared in microemulsions, at 500 psig 5%H S/95%H initial partial pressure, 430°C, and 1-h reaction time). 2  Types of Microemulsions  Fe/Hexane/PE4LE Co/Hexane/PE4LE Mo/Hexane/PE4LE Fe/DHN/PE4LE Co/DHN/PE4LE Mo/DHN/PE4LE Fe/Toluene/DDAB Co/Toluene/DDAB Mo/Toluene/DDAB Fe/THF/DDAB Co/THF/DDAB Mo/THF/DDAB  2  CH4  Gas Product Composition (mol C2H4 C0 HS C H6  25.3 29.8 22.9 22.9 19.8 23.5 18.3 23.7 20.7 30.0 25.4 25.5  0.05 0.06 0.06 0.13 0.12 0.28 0.05 0.05 0.04 0.07 0.06 0.05  2  4.5 6.2 5.2 4.3 3.8 5.0 2.2 2.8 2.6 7.1 5.1 5.4  2  1.4 2.0 1.6 2.5 2.3 2.0 0.8 1.0 0.6 16.1 11.1 10.6  2  2.8 3.8 3.1 3.3 3.1 3.3 1.1 2.5 2.6 2.0 1.6 1.6  H & others* 65.9 58.2 67.0 66.9 70.9 66.0 77.5 69.9 73.5 44.8 56.8 56.9 2  * Obtained by difference (see Appendix II).  103  The formation of cracked hydrocarbon gases (CH4,  C H6) 2  plus H S, CO, and C 0 2  2  from the hydrocracking reactions tends to increase reactor pressure whereas hydrogenation and hydrogenolysis reactions together with the reaction of H S with metal 2  catalysts, tends to reduce the reactor pressure (Chen et al., 1988). Gaseous products such as CH4 and C H<; are less valuable than liquid products. The desired performance of the 2  catalyst is to convert the residue fraction (BP >525°C) to distillable liquids while at the same time minimizing gas and coke yield. Therefore, those catalysts which give the lowest final pressure after one-hour reaction time are considered superior because the low pressure is indicative of suppression of hydrocarbon gas formation and promotion of hydrogen incorporation into the liquid product (Chen et al., 1988). Thus the gas formation or the final pressure of the autoclave is one of several ways to observe the activity of the catalysts. However, to rank the activity of the catalysts, the quality of the liquid product and the coke yield must also be considered. The autoclave pressures after 1 h reaction at 430°C for Fe, Co, and Mo microemulsion catalysts in DHN are shown in Figure 5.1. The results of the present study showed that Mo was superior to Fe and Co, and Co was superior to Fe for suppression of gas formation or enhancement of H consumption, and 2  this trend was the same for each of the solvents considered.  104  1600  200 +  0  -I  0  1  1  20  1  1  40  1  1  60  1  1  1  80  1  100  1  1  120  1  1  140  1  1 160  Time (mm)  Figure 5.1 The autoclave pressure vs time during residue upgrading in DHN/PE4LE (500 psig, 430°C, 1 h). The heating rate was 5°C/min and the reactor reached 430°C after 82 minutes.  105  5.3.2 Effect of catalyst on coke yield. Low coke yield (or dichloromethane-insoluble matter) reflects a superior ability of the catalyst to transfer hydrogen to unsaturated hydrocarbon fragments, reducing condensation reactions that lead to coke. The coke yields were measured using Cold Lake residue reacted at 430°C with an initial 5%H S/95%H pressure of 500 psig and a reaction 2  2  time of 1 hour. The Fe, Co, and Mo catalysts prepared in reverse micelles with various solvents were examined. The mechanism of residue hydrocracking is mainly a thermal process as described in Section 2.6.2. Carbon-to-carbon bonds are cracked thermally to form free radical molecules which are stabilized by hydrogen addition. Without catalyst, the aromaticcarbon radical reacts with other similar species to produce coke. Catalysts transform hydrogen gas to hydrogen radical (H») which stabilizes the aromatic-carbon radical and prevents coke formation (Sanford, 1994). Catalysts differ slightly in their abilities to promote bond cleavage and hydrogenation reactions (Chen et al, 1988). Some molecules are not readily thermally cracked; so eventually, these molecules form coke. Catalysts that promote cleavage reactions will remove heteroatoms from the residue molecules and hydrogenate the resulting products to produce more easily hydrocracked cycloparaffins, resulting in less coke and higher liquid yields. The coke yields (Table 5.5 and Figure 5.3) obtained using different catalysts prepared in reverse micelles using DFTN/PE4LE, increased in the order Mo < Co < Fe. The analysis of these data via an analysis of variance following Fisher's least significant  106  difference (LSD) (Appendix HI), revealed that at the 90% level of significance, the coke yield from the Mo/DHN/PE4LE was significantly lower than that from Co/DHN/PE4LE catalyst. Furthermore, the coke yield from Fe/DHN/PE4LE was significantly greater than that obtained from Co/DHN/PE4LE and Mo/DHN/PE4LE. These results are supported by the results of Chen et al. (1988) and Bearden and Aldridge (1979). Chen et al. (1988) used naphthenate salts as catalyst precursors and reported that coke yield increased in the order: Mo < Co whereas Bearden and Aldridge (1979) used iron naphthenate, cobalt resinate, and molybdenum resinate as catalyst precursors and reported a coke yield order as follows: Mo < Co < Fe. Note that the trend in coke yield with the 3 catalysts was similar to that of the autoclave pressure results discussed in Section 5.3.1.  Table 5.5 Coke yield from residue upgrading using metal catalysts prepared in DHN/PE4LE microemulsions, at 500 psig, 430°C, 1 h. Experiment #  Catalyst  7 21 24 8 19 23 6 17  Fe Fe Fe Co Co Co Mo Mo  Coke Yield (%) 7.4 7.0 7.0 6.6 6.0 6.0 5.8 5.8  Coke Yield (%) Mean ± Std Error  7.1±0.1  6.2 ±0.2 5.8 ±0.0  107  Fe  Go  Figure 5.2 Coke yield from residue upgrading using Fe, Co, and Mo catalysts prepared in DHN/PE4LE microemulsions, at 500 psig, 430°C, 1 h. From the gas and coke formation results, it may be concluded that Mo was the best catalyst for suppression of gas and coke, both of which are undesirable products. The high performance of Mo for the suppression of gas and coke formation compared to Fe, was in agreement with previous studies in which organometallic precursors were used for catalyst preparation.  5.3.3 Effect of catalyst on liquid product quality The liquid yield reported in Table 5.6 includes both the organic solvent used in the microemulsions and the liquid products from residue upgrading. During the disassembly of the reactor vessel, the liquid was exposed to the atmosphere and some volatile products escaped from the reactor vessel. Solvents with higher boiling points were more likely to  108  remain with the liquid products. As seen in Table 5.6, the liquid yields from Fe, Co and Mo in DFTN/PE4LE microemulsions were in the range 89.4-93.4%. Beside the desirable high liquid yield, the quality of the liquid product is also important. The liquid product should have low values of MCR, >525°C residue fraction, S, and asphaltenes and this is achieved by high conversion of these properties. The analyses of the liquid products obtained from different catalysts are shown in Table 5.7. The average values of the experimental results from different types of catalyst are plotted in Figure 5.3.  Table 5.6 Hydroconversion of residue oils with catalysts in DHN/PE4LE microemulsions at 500 psig, 430°C, 1 h. Experiment #  Catalysts  7 21 24 8 19 23 6 17  Fe Fe Fe Co Co Co Mo Mo  Note: Liquid yield (%)  =  Liquid Yield (%) 89.4 92.7 92.7 91.8 93.4 92.2 92.5 93.1  Liquid Yield (%) Mean ± Std Error  91.6 ± 1.1  92.5 ±0.5 92.8 ±0.3  the weight of liquid yield xlOO residue oil + microemulsion  109  Table 5.7 Average product analysis data from residue oil conversion with catalysts in DHN/PE4LE microemulsions at 500 psig, 430°C, 1 h.  Catalysts  % Conversion Mean ± Std Error MCR >525° C  S Fe Co Mo  40.1 ±0.6 47.3 ± 1.7 53.4 ±3.3  50.4 49.6 ± 1.6 47.7  Asphaltenes  30.3 ± 1.0 31.8 ± 1.4 34.5 ±0.6  a  a  51.9± 1.4 55.8 ± 1.3 38.0 + 0.6  The analysis was not repeated.  100 90  m Fe/DHN/PE4LE • Co/DHN/PE4LE  80  • Mo/DHN/PE4LE  70 60  53.4 47.3  50 40  30.3  49.6  47.7  -1:  40.1 31.8  55.8 50.4  34.5  38.0  IS  30 20  1SS  \  ytisi wi  10  %  :  tiiiiiiii  0 MCR  >525C  Asphaltenes  Figure 5.3 Liquid product from residue conversion with catalysts in DETN/PE4LE microemulsions at 500 psig, 430°C, 1 h.  110  5.3.3.1 MCR (Microcarbon residue) conversion or CCR (Conradson carbon residue). Microcarbon residue, based on ASTM method D-4530, is a measure of the cokeformation tendency (during pyrolysis in nitrogen) of an organic material. CCR is defined as the weight percent of the coke which forms on destructive distillation. In many studies CCR values are reported and they are interchangeable with MCR values. The results from Table 5.7 and Figure 5.3 showed that the MCR conversion of Mo was 12% higher than that of Fe and 8% higher than that of Co and these differences are above the range of the analysis error (4-6%). The difference in the M C R conversion between Co and Fe was about 5% and this is in the range of the analysis error. However, the ranking of these average values also happens to duplicate the rankings obtained from gas and coke yields. The analysis of MCR conversion via an analysis of variance following Fisher's least significant difference (LSD) (Appendix HI), showed that at the 90% level of significance, the MCR conversion with Mo was not significantly higher than with Co catalyst, but it was higher than with Fe. The high MCR conversion of Mo was consistent with its low coke yield reported in Section 5.3.2, due to the ability of Mo to convert CCR materials to non-CCR materials.  5.3.3.2 >525°C conversion (or residue conversion). Residue conversion is shown in Figure 5.3. There was no difference in hydrocracking activities among the catalysts based on an analysis of variance (shown in Appendix HI). Conceivably, the residue conversion was dominated by thermal cracking and the effect of different catalysts was not significant.  Ill  5.3.3.3 Sulfur conversion Sulfur poisons catalysts in downstream hydrotreating processes, and sulfur in fuel oil pollutes the environment. The sulfur compounds in residue consist of aliphatic sulfur species and thiophene compounds, and these exhibit very different reactivities. The aliphatic sulfur species are readily converted by thermal reactions while the conversion of thiophene compounds requires catalytic hydrodesullurization via hydrogenolysis and hydrogenation. From Figure 5.3, among Fe, Co, and Mo catalysts, Mo showed the highest hydrodesulfurization activity. The analysis of variance following Fisher's least significant difference (LSD) (Appendix UI), showed that at the 90% level of significance, the S conversion of Mo was significantly higher than that of Co catalyst, and both Mo and Co gave higher S conversion than Fe. The high hydrodesulfurization activity of Mo was supported by Kennepohl and Sanford (1996) and Lui et al. (1994) who indicated that M0S2  was an excellent hydrodesulfurization (HDS) catalyst and could promote the  destruction of heterocyclic ring systems at S centers.  5.3.3.4 Asphaltene conversion In terms of asphaltene conversion to liquid product the order of asphaltene conversion shown in Figure 5.3 is: Co > Fe > Mo. The differences in asphaltene conversion data among the catalysts was also confirmed by the analysis of variance and Fisher's least significant difference, at the 90% level of sigriificance, shown in Appendix  112  m. The results of this study revealed that the asphaltene conversion in the hquid product was lower with Mo than with Fe and Co catalysts. Catalysts do not only affect the asphaltene conversion but also the hydrogen pressure. The promotion of hydrogenation reactions by catalysts is a consequence of hydrogen transfer from gas phase to the asphaltenes via the catalyst. Some of the aromatic compounds are hydrogenated and the products are more readily thermally cracked than the aromatics. When the hydrogen pressure increases, there is an increase in the probability that the catalyst will contact the hydrogen gas and transfer hydrogen to the asphaltenes. Chen et al. (1988) found that the asphaltene content after hydrocracking of Athabasca bitumen decreased when the hydrogen pressure increased. In the present study, the initial hydrogen pressure was kept constant, thus there was no major effect of pressure on the asphaltene conversion. However, Chen et cr/.'s result implies that catalysts with good hydrogenation activity reduce asphaltene content. Catalysts behave differently for hydrogenation reactions. Catalysts that give low asphaltene content in the residue upgrading product reflect a high hydrogenation activity. Figure 5.3 shows the quality of the liquid products obtained when using Fe, Co, and Mo catalysts prepared in DHN/PE4LE. Among the catalysts studied, Mo showed the highest activity for MCR and S conversions but the lowest activity for asphaltene conversion. The low asphaltene conversion in the presence of Mo indicated that most of the residue oil was converted to non-CCR molecules but the converted molecules were not completely hydrogenated to pentane-soluble compounds, resulting in high asphaltene content in the liquid product. The low gas pressure for Mo indicated that most of the  113  hydrogen incorporated with the residue oil and the cracked product yielded liquid product. The experimental data confirmed the consumption of hydrogen but the amount could not be accurately quantified by the experimental procedures used here. The low asphaltene but high MCR and S conversions for Mo indicated that Mo was active for heteroatom removal and H transfer to stabilize the aromatic radicals, but not as active as Fe and Co for hydrogenating the aromatic radicals to undergo further transformations to nonpentane-soluble compounds. The ratio of coke yield/asphaltene conversion for Fe, Co, and Mo catalysts is presented in Figure 5.4 and the lowest value is desired since this implies asphaltene conversion to distillable liquid rather than to coke. Even though Mo gave the lowest coke yield, the ratio of coke/asphaltene conversion for Mo (Figure 5.4) was higher than for Fe or Co at the 90% level of significance.  18 16 14 -12 -|  io +  o  8 --  15.3  • Fe/DHN/PE4LE • Co/DHN/PE4LE • Mo/DHN/PE4LE  13.8  11 1  7.1 5.8  6 -4 -2 -0 Coke  Coke/Asphaltene  Figure 5.4 Effect of Fe, Co, and Mo catalysts on coke yield and coke yield/asphaltene ratio on residue conversion with catalysts prepared in DHN/PE4LE microemulsions at 500 psig, 430°C, 1 h.  114  The results of this study on the activity of Mo are also supported by the study of Curtis et al. (1987). Curtis et al. (1987) studied the coprocessing of coal with heavy residue and in comparing Mo naphthenate to Co naphthenate catalyst precursors, showed that Mo naphthenate gave higher coal conversion, higher oil production, yielding lower insoluble organic matter (methylene chloride insoluble or coke) but lower asphaltenes conversion than that of Co naphthenate. The results of the present study suggest that Mo was active for both hydrogen transfer to the radical molecules via reaction R2 (Figure 2.14) and to large hydrocarbon molecules via reaction R 4 (Bearden and Aldridge, 1981). Fe and Co were active for reaction R but showed relatively poor activity for reaction R 4 . Some of the coke 2  precursors or Conradson carbon materials that were not convertable by thermal cracking followed by hydrogenation, required some hydrogen input via reaction R 4 before undergoing cracking. Mo converted these coke precursors to non-coke precursors via reaction R 4 and this supplemented reaction R , resulting in less coke formed than with Fe 2  or Co. The high S conversion of Mo catalyst showed that Mo was an active hyckodesulfurization catalyst and could promote the destruction of heterocyclic ring systems at S centers as indicated by Dennepohl and Sanford (1996). The poor asphaltene conversion but high MCR and S conversions of Mo reflects the fact that Mo promotes hydrodesulfurization of heavy molecules and stabilizes the aromatic radicals via hydrogen capping. However, Mo is inefficient in promoting hydrogen transfer to the aromatic molecules to yield hydroaromatic compounds that can be further thermally hydrocracked to pentane-soluble products. The possibility of improving the  115  asphaltene conversion of the Mo catalyst, by nibring Mo with a catalyst that promotes hydrogenation reactions, such as Co or Ni, is discussed in Section 5.6.3.  5.4 Effect of microemulsion solvent on catalyst residue upgrading activity  5.4.1 Effect on autoclave pressure The catalysts used in the present study were colloidal particles stabilized with surfactant and dispersed in organic solvents. The relatively low boiling point of the organic solvents resulted in increased pressure as the reactor temperature increased, in addition to the increased pressure that is a result of gas production. The presence of the organic solvent dictated that the initial pressure of 5%H S/95%H in the reactor be set at 500 psig, 2  2  lower than that of other studies (Chen et al., 1988; Sanford and Chung, 1991). Indeed, the pressure increase when using catalysts prepared in tetrahydrofuran was so high that some gas had to be released during the reaction to maintain the autoclave pressure below the maximum allowable (about 2500 psig, at 430°C). Hence, the catalysts prepared in THF are not compared to those prepared in n-hexane, DHN, and toluene. The effect of low boiling point solvent on the final reactor pressure is shown in Figure 5.5. Hexane (BP 69°C) solvent gave the highest pressure (up to 2300 psig), whereas toluene (BP 111°C) and DHN (cis & trans, BP 189-191°C) gave lower pressures (up to 1800 psig and 1500 psig), respectively.  116  2500  3  n-Hexane/PE4LE  Toluene/DDAB  DHN/PE4LE  Solvent/Surfactant  Figure 5.5 The autoclave pressure after 1-h reaction time (500 psig, 430°C, 1 h) from residue upgrading using catalysts prepared from microemulsions based on different solvents.  5.4.2 Effect on coke yield The catalysts of the present study were prepared in reverse micelles and dispersed in different solvents, some of which are known to facilitate hydrogen transfer. Hence, the coke yields are likely to be very dependent on the solvent used in the catalyst synthesis. Figure 5.6 shows a relatively small effect of metal type on coke yield in each solvent, but shows a more significant effect of different solvents. For each solvent except THF, the coke yield increased in the order Mo < Co < Fe. The (lifferences in coke yield among Fe, Co and Mo were significant, based on the analysis done for DHN and discussed in Section 5.3.2.  117  50 42.8  45 •40 ••  • Fe  35 •-  • Co  42.4  43.0  • Mo O "3 25 +  19.6  15 + 10  12.2 7.1  6.2  5.8  5 + 0 DHN/PE4LE  Toluene/DDAB  n-Hexane/PE4LE  THF/DDAB  Solvent/Surfactant  Figure 5.6 Coke yield after 1-h reaction timefromresidue upgrading (500 psig, 430°C, 1 h) using Fe, Co and Mo catalysts prepared in microemulsions using different solvents.  As discussed in Section 5.2, the effect of solvent was not a consequence of different catalyst dispersions in the residue oil with different solvents. The data of Figure 5.6 showed that for each metal type, coke yield increased with solvent type in the order DUN < toluene < hexane < THF. The differences in coke yields among n-hexane/PE4LE, DHN/PE4LE, toluene/DDAB, and THF/DDAB were much greater than the differences among Fe, Co, and Mo catalysts prepared in the DHN/PE4LE microemulsions. In the latter case the differences were shown to be significant and hence, the differences among the solvents should also be significant. Hexane is a paraffinic solvent which, at certain concentrations, precipitates asphaltenes from the residue oil solution. When asphaltenes are separated in a second  118  phase, the coke precursors are in close contact and tend to combine with each other to form coke (Lott et al, 1993; Shaw, 1988). In the present work asphaltene likely phase seperated upon addition of hexane at room temperature. Upon heating to reaction temperature hexane would vaporize. However, because the asphaltenes had phase seperated coke formation would likely be enhanced. DHN is a known hydrogen-donor solvent that can transfer hydrogen to the thermally generated organic radicals and would remain in the liquid phase at reaction conditions. Hence, the coke precursors are hydrogenated prior to forming coke, reducing coke yield. Even though toluene does not provide a hydrogen transfer capability, it promotes solubility of the asphaltenes in the oil and hence reduces coke formation. Coke yield in hexane was higher than in toluene; coke yield in DHN was the lowest. Tetrahy<kofuran, a ring compound containing one oxygen atom, was also used as a solvent for the microemulsions. At reaction temperature and with the assistance of metal catalyst, the C-0 bond of tetrahydrofuran is weakened and the molecule becomes a diradical which is very active. This diradical molecule will react with the other radicals (chain propagation) or coke precursors to form coke, or it may react with hydrogen atoms to produce butanol and butanal. Butanal may react with hydrogen to produce propane and CO, and CO may react with O to produce C 0 . A high C 0 content was observed when 2  2  using catalysts prepared in tetrahydrofuran based microemulsions (see Table 5.4). Coke yield in tetrahydrofuran was very high (42-43%) compared to that obtained with other solvents (5-23%). These results imply that the chemical nature of the solvent used in the  119  microemulsion is very important in residue upgrading and the H-donor capability is the most important property deterniining coke yield. It should be noted that in comparing the effects of different solvents, the surfactant had to be changed from PE4LE to obtain stable microemulsions with toluene and THF. In comparing the effects of DHN versus toluene, therefore, the differences in surfactant (PE4LE versus DDAB) may be important, as discussed in Chapter 2. However, the trends in coke yield as a function of catalyst type are the same among the solvents, and for each catalyst type, the effect of solvents is well explained by the chemical properties of the solvents as detailed above. Hence, it seems that although the effect of different surfactants cannot be excludedfromthese results, their effect seems relatively minor compared to the effect of solvent type and metal type.  5.4.3 Effect on liquid product quality Figure 5.6 shows that the type of solvent has an important effect on coke yield. The investigation of the liquid products from residue upgrading in hexane/PE4LE and in DHN/PE4LE (Figure 5.7 (a) and (b)) revealed a dramatic difference in quality as well. The quality of the liquid product (MCR, >525°C, S, and asphaltene conversions) in hexane/PE4LE was better than that in DHN/PE4LE. Moreover, the coke/asphaltene ratio in hexane was also higher than that in DHN/PE4LE. This was due to the fact that in hexane, most of the highly aromatic, heavy components are rejected as coke (coke yield ranged from 19.6% to 22.4% compared to 5.8-7.1% in DHN), leaving the liquid product with lower S, MCR, >525°C, and asphaltene contents.  120  Figure 5.7 (a) Effect of solvent on product quality from residue conversion with Co/nhexane/PE4LE and with Co/DHN/PE4LE catalysts at 500 psig, 430°C, 1 h.  30 BCo/Hexane/PE4LE • Co/DHN/PE4LE  25 +  23.9  20.8 20 4-  g o  11.1 10 6.2 5  Coke  Coke/Asphaltene  Figure 5.7 (b) Effect of solvent on coke yield and coke yield/asphaltene conversion from residue conversion with Co/n-hexane/PE4LE and with Co/DHN/PE4LE catalysts at 500 psig, 430°C, 1 h.  121  5.5 Catalysts based on organometallic precursors As has been discussed in Chapter 2, one of the more conventional approaches to preparing dispersed catalysts for residue upgrading is to use organometallic precursors that are soluble in residue oil. The precursors decompose under the reaction conditions to produce, in situ, the dispersed catalyst. In this section, the activity of cobalt naphthenate is compared to the Co catalysts prepared in a reverse micelle.  5.5.1 Effect on gas formation The effect of the organometallic catalyst on the suppression of gas can be seen by comparing the autoclave pressure in the absence of metal catalyst with that obtained in the presence of cobalt naphthenate (Figure 5.8). In both experiments, DHN was added to the feed so that direct comparisons could be made with the catalysts prepared in reverse micelles, in which DHN was used as the solvent and PE4LE as a surfactant of the microemulsion. The autoclave pressure in the absence of catalyst (DHN) was higher than in the presence of catalyst (DHN + cobaltnaphthenate). The other evidence that showed the effectiveness of the catalyst to suppress gas formation was the lower final autoclave pressure obtained with cobalt naphthenate catalyst precursor with no DHN (experiment # 3, Appendix I) compared to that of thermal hydrocracking (experiment # 1, Appendix I) (no catalyst and H-donor solvent). The autoclave pressure when using cobalt naphthenate as catalyst precursor was lower than when Co prepared in a reverse micelle was used. These differences are thought to arise from the surfactant. Surfactant is believed to affect the autoclave pressure in two  122  ways. Firstly, the long hydrocarbon tail of surfactant is likely cracked to small hydrocarbon molecules at reaction conditions and this increases autoclave pressure. Secondly, there is evidence (Martino et al, 1994 and Ravet et al, 1984) that surfactants (e.g.didodecyldimethylammonium bromide (DDAB) and cetyltrimethylammonium bromide (CTAB)) poison the catalysts and lower catalyst hydrogenation activity.  1600 -p 1500 1400 --  -I—i—i—i—I—i—i—i—i—i—i—i—i—i—i—i—i—i—i—<—i—i—'—i—i—i—i—i—i—I—i—i—i—I  400 0  20  40  60  80 Time (min)  100  120  140  160  Figure 5.8 Autoclave pressure vs time during hydroconversion of residue oil (500 psig, 5 C/min heating rate, 430°C, 1 h). 0  123  5.5.2 Effect on coke yield In the present study, coke yield (Figure 5.9) from thermal hydrocracking of the Cold Lake residue alone was 15.0 ± 0.0% and this was reduced to 12.6% in the presence of the catalyst precursor cobalt naphthenate. Coke suppression was much decreased (5.9%) when DHN, a H-donor solvent, was added, and further reduced to 5.0% when cobalt naphthenate together with DHN were used. The lower coke yield with cobalt naphthenate than with DHN was due to the coke suppression activity of Co metal catalyst derived from the decomposition of the catalyst precursor, cobalt naphthenate. Similarly, Co metal catalyst from Co/DHN/PE4LE should also have the ability to lower the coke yield. However, the coke yield with Co/DHN/PE4LE (6.2%) was higher than with DHN (5.9%). The higher coke yield with Co/DHN/PE4LE compared with pure DHN may be a consequence of deactivation of the Co catalyst by surfactant (PE4LE) as discussed in Chapter 2. Alternatively, as will be discussed in Section 5.5.3 this may reflect a change in selectivity of the hydroconversion reactions that occur in the presence of Co, that result in higher coke yield.  124  Thermal  Cobalt naphthenate  DHN  Cobalt Co/DHN/PE4LE naphthenate/DHN  Figure 5.9 Coke yieldfromresidue conversion at 500 psig, 430°C, 1 h.  The coke yield from cobalt naphthenate dissolved in DHN was 5.0 ± 0.0% compared to 6.2 ± 0.2% for the Co/DHN/PE4LE. The Co catalyst prepared in the microemulsion was less active for coke suppression than the Co catalyst prepared by decomposition of the Co naphthenate catalyst precursor, and this difference may also be due to the effect of the surfactant. Martino et al. (1994) reported a decrease in the activity of  a  Shell  324  catalyst  (NiMo/Al20 ) 3  doped  with  cationic  surfactant  ((hdodecyldimemylammonium bromide, DDAB). They suggested that in the presence of surfactant, catalytic activity was hindered by two mechanisms. Surfactant chemically poisoned the catalyst, and byproducts of surfactant disintegration at reaction conditions behaved like a hydrogen acceptor and scavenged hydrogen from hydrogen-donating solvents. Thus, less hydrogen transfer to the residue occurred and this indirectly lowered  125  the activity of the catalyst. The latter mechanism is important in hydrogenolysis reactions and liquefaction as these processes depend strongly on hydrogen transfer. Two types of surfactant were used in the residue upgrading of the present study, the  nonionic,  polyoxyethylene-4-lauryl  ether  and  didodecyldimethylammonium bromide. The nonionic surfactant  the  cationic,  was most likely  hydrocracked to small hydrocarbon molecules and the byproducts of surfactant disintegration at reaction conditions scavenged some hydrogen atoms (Martino et al., 1994). The cationic surfactant was disintegrated at high temperature and released a basic nitrogen compound, which is a poison to the hydrogenation catalyst (Gray, 1994; Sundaram et al., 1988; Stanislaus and Cooper, 1994) as well as a H scavenger.  5.5.3 Effect on liquid product quality (MCR, S, > 525°C, and asphaltene conversions). Cobalt naphthenate/DHN gave higher conversions of MCR and S, and a comparable >525°C fraction conversion, but lower asphaltene conversion than the colloidal Co/DHN/PE4LE as shown in Figure 5.10. The hydroxyl groups of the PE4LE, as well as water molecules in the microemulsions, possibly exchanged their hydrogen atoms to give an abundance of hydrogen which should preferentially promote the hydrogenation reaction. The data of Figure 5.1 show a change in catalyst selectivity that suggests the difference between the Cobalt naphthenate/DHN and the Co/DHN/PE4LE catalysts may not due to the PE4LE alone. Other factors such, as the properties of the resulting Co particles, may also be important.  126  MCR  >525C  Asphaltenes  Figure 5.10 Liquid product from residue conversion at 500 psig, 430°C, 1 h, using Hdonor solvent, catalystfromreverse micelle, and organometallic compound.  5.6 Effect of mixed catalyst on residue upgrading In residue upgrading the major reactions enhanced by catalysts are hydrogenation and hydrogenolysis and there is no single catalyst that has high activity for both reactions. It is well known that the presence of a second metal can increase the catalyst activity more than the sum of the activities of the single catalysts based on binary sulfides. Many researchers have studied the use of mixed catalyst and synergism has been observed. Synergism in atmospheric residue hydrotreatment (Lee et al, 1996) and Athabasca bitumen upgrading experiments (Chen et al, 1989) have been observed for mixed CoMo (mixture of cobalt and molybdenum naphthenates) and NiMo (mixture of nickel and molybdenum naphthenates) catalysts. In these cases, it is supposed that the two metals  127  provide a complementary function. Furthermore, the use of a second component could reduce catalyst cost if the concentration of a more expensive component can be reduced. In the present work, mixed catalysts could be readily prepared by simply mixing the microemulsion containing the reduced metals (Section 3.1.3 ). The present section reports on the activity of these mixed catalysts for residue oil upgrading.  5.6.1 Effect on gas yield The final autoclave pressures after 1-h reaction at 430°C using the CoMo and NiMo mixed catalysts are shown in Figure 5.11. The pressure using the mixed catalysts was comparable to that of Mo and lower than that of Co. The final autoclave pressure of CoMo was slightly lower than that of NiMo and this result was consistent with the results of Sanford and Chung (1991). These investigators used commercially available catalysts of NiMo and CoMo supported on Y-Al 0 for catalytic hydrocracking of 400-g bitumen at an 2  3  initial H pressure of 800 psig, and a reaction temperature of 370°C. 2  5.6.2 Effect on coke yield The efficiency of Mo for coke suppression was promoted when Mo was mixed with N i or Co, which are efficient for hydrogenation, as shown in Figure 5.12. The analysis of variance and the Fisher's least significant difference (LSD) procedure, at the 90% level of significance, showed that coke yield from Co and Mo was higher than that from NiMo and CoMo. The optimum ratio of NiMo or CoMo depends on the  128  composition of the residue. For example, a high ratio of Mo is needed in residue with high heteroatoms when the removal of the heteroatoms is important.  1300  9 0 0 -I  80  1  1 90  1  1 100  1  1 110  •  1 120  1  1 130  1  1 1 40  •  1 150  Time (min)  Figure 5.11 The autoclave pressure vs time during residue upgrading using different types of catalysts prepared in DHN/PE4LE microemulsions (500 psig, 430°C, 1 h).  129  Co/DHN/PE4LE  Mo/DHN/PE4LE CoMo/DHN/PE4LE NiMo/DHN/PE4LE  Figure 5.12 Coke yield in mixed catalystsfromresidue conversion at 500 psig, 430°C, lh.  5.6.3 Effect on liquid product quality (MCR, S, >525°C, and asphaltene conversions) The quality of the liquid product using Co, Mo, CoMo and NiMo mixed catalysts is shown in Figure 5.13 (a) and (b). The results show that Mo had higher MCR and S conversions than that of Co and the synergistic effect on MCR conversion was observed for the mixed CoMo catalyst at 90% level of significance. The S conversion of CoMo was higher than that of NiMo as indicated by Gray (1994) and apparently consistent with the results of the present study shown in Figure 5.13 (a).  130  100 90 + 80 470 -§  60  IMo/DHN/PE4LE • Co/DHN/PE4LE •CoMo/DHN/PE4LE H NiMo/DHN/PE4LE  4-  55.0 54.4  53.4  MCR CT  (Std dev)  8  50.2  >525C  1.9  5 5  49.6,  2.9  0.1  Asphaltene  2.1  Figure 5.13 (a) Liquid product from residue conversion at 500 psig, 430°C, 1 h, using Co, Mo, CoMo, and NiMo catalysts prepared in DHN/PE4LE microemulsions. 18 16  • • • •  14 12  %  10 8 6 4 2 0  Mo/DHN/PE4LE Co/DHN/PE4LE CoMo/DHN/PE4LE NiMo/DHN/PE4LE  EE 5.8  6.2  5.2  Coke  CT  (Std dev)  0.2  5.2  Coke/Asphaltene  0.8  Figure 5.13 (b) Effect of Co, Mo, CoMo, and NiMo on coke yield and coke yield/asphaltene ratio on residue conversion using catalysts prepared in DHN/PE4LE microemulsions at 500 psig, 430°C, 1 h.  131  The analysis of variance at the 90% level of significance showed no differences in the >525°C fraction conversion but showed significant differences in asphaltene conversion (Appendix DT). This implies that residue conversion was mainlyfromthermal cracking and the effect of catalysts was small. The low hydrogenation activity of Mo compared to Co catalyst was shown by the low asphaltene conversion in Figure 5.13 (a) and the conversion was improved with the addition of Ni and Co. Moreover, the improvement in the hydroconversion activity of NiMo and CoMo catalysts was reflected in the coke/asphaltene ratio as shown in Figure 5.13 (b). The lower asphaltene conversion of CoMo than that of NiMo in the present study is consistent with the lower hydrogenation activity of CoMo compared to that of NiMo, indicated by Meille et al. (1997). Fisher's least significant difference (LSD) was not able to identify a statistically sigriificant difference in asphaltene conversion between CoMo and NiMo catalysts mainly because of the small number of data. An increased number of replicated experiments would increase the level of confidence in the average values. However, the consistency of the experimental results of the present study with other studies, support the trends observed in the present study. In the contact synergy model proposed by Delmon (1979), MoS and Co S are 2  9  8  assumed to exist as separate crystallites adjacent to each other, and the role of the promoter (CogSg) is to activate and transfer hydrogen atoms to MoS . These spilled-over 2  hydrogen atoms initiate reduction on the MoS surface yielding S vacancies, which would 2  be the catalytically active sites.  132  Topsoe et al. (1981) found a unique form of sulfided Co, the "CoMoS" phase which correlated with activity, in both supported and unsupported CoMo catalysts. For CoMo catalysts it was suggested Co should be located at the edge of the MoS  2  crystallites, or between the layers near the edge. The main idea is that at the edge, sulfur atoms which upon leaving cause vacancies, are shared by Co and Mo. The strength of the metal-sulfur bonds which lead to vacancies are intermediate (not too strong or too weak) and this is necessary for high activity. In the present study, the CoMo and NiMo catalysts were obtained by combining individual metal catalysts previously prepared in different microemulsions. The mixed catalysts consist of both metals in separate particles surrounded by surfactant dispersed throughout the organic phase. The improvement in hydroconversion activity of Mo when mixed with N i or Co observed in this study is well explained by the contact synergy model rather than the Topsoe model. The metals are most likely not in contact with each other but rather behave as synergistic pairs. Co and N i are good hydrogenation catalysts which supply hydrogen atoms to MoS , creating more vacancies at the edge of M0S2 and making 2  M0S2 more active.  5.7 Summary In summary, organic solvents used in colloid preparation showed a greater effect on residue conversion than the catalyst metal type. The effect of organic solvents was due to their chemical nature since the effect of solvent on catalyst dispersion was not significant. Mo showed high activity for coke and gas suppression, S and M C R  133  conversions but showed low asphaltene conversion at the conditions studied (500 psig, 430°C, 1 h). This indicated that Mo catalyst was a better hytfrodesulfurization but poorer hydrogenation catalyst than Fe or Co. The high activity of Mo was not due to particle size effects since it was shown that Mo particles (d =7.7 nm) were larger than Fe (d = 6.6 nm) or Co (d =5.3 nm) prepared in reverse micelles, and the same catalytic metal mass was used in each reaction. The synergistic effect of CoMo and NiMo mixed catalysts was observed for coke suppression and MCR conversion at 90% level of significance. Co catalysts prepared using organometallic compounds were superior to those prepared in reverse micelles except in terms of asphaltene conversion. The differences in performance between these two catalysts may be due to the effect of surfactant present in the catalysts prepared in reverse micelles, or may be due to differences in the properties of the catalyst particles resultingfromthe two methods of preparation.  134  Chapter 6 Conclusions and Recommendations 6.1 Conclusions A series of metal catalysts (Fe, Co, and Mo) were successfully prepared in microemulsions, using different types of organic solvents (n-hexane, DHN, toluene, and tetrahydrofuran). To obtain a stable microemulsion the hydrophile-lypophile balance (HLB) of the system and, hence, the surfactant must be suitable. The addition of metal salt and reducing agent can cause salting in or salting out and disturb the stability of the microemulsion. Hence careful choice of metal salt and reducing agent was necessary to obtain stable microemulsions. Catalysts prepared from microemulsions with different metal types had shghtly different particle sizes and size distributions in the range 5-8 nm. The average particle size decreased in the order: Mo > Fe > Co. It is proposed that the most important factors that led to different metal size were the solvation of the metal ions at the surfactant-water interface and the rate of nucleation during the reduction process. Some previous studies have reported that bigger particles were obtained as the water to surfactant ratio of the microemulsion increased; however, in the present study, Mo, which was prepared in a microemulsion with the lowest water to surfactant ratio, showed the largest particle size. Furthermore, a series of Co catalysts prepared in microemulsions with varying water to surfactant ratio showed no important effect on particle size.  135  Prior to studying the hydroconversion activity of the catalysts, the dispersion in residue oil of the catalyst prepared in different solvents was determined in order to investigate the effect of the solvents on the dispersion of the catalysts. The statistical data analysis showed no difference of metal catalyst dispersion in the residue oil using microemulsions based on different solvents. Hence it may be concluded that any effect of microemulsion solvent on hydroconversion activity must be due to the chemical properties of the solvent rather than catalyst dispersion effects. The hydroconversion activity of the catalysts reflected in coke yield data showed that solvent had a greater impact on the coke yield than the metal type. H-donor ability of the solvent was the most important factor for coke suppression. In spite of the bigger particle size of Mo compared to Fe and Co, Mo showed the highest activity for conversion of MCR, S and >525°C fraction, and had the lowest coke yield and gas formation. The greater activity of Mo relative to Fe and Co was therefore due to its chemical nature, not particle size. Mo showed good hydrocracking but low hydrogenation activities since the asphaltene conversion in Mo was below that of Fe and Co. An improvement in asphaltene conversion with Mo was observed when Mo was mixed with Co or Ni. The apparent synergistic effect of the mixed CoMo and NiMo catalyst was observed in the reduced coke yield and in enhanced MCR conversion. CoMo catalyst showed the best performance among the catalysts studied (Fe, Co, and Mo). The present study showed a synergistic effect by mixing two separated metal particles whereas in some studies, Co was found integrated in the M0S2 structure of CoMo catalysts. The  136  structure of the CoMo mixed catalyst after  in situ  sulfiding was not investigated and it is  beyond the scope of the present study. Comparing the Fe, Co, and Mo catalysts prepared in DHN/PE4LE microemulsions with the more conventional cobalt naphthenate catalyst precursor, showed the latter to have a lower coke yield and higher conversions of MCR, S and 525°C fraction, but lower asphaltene conversion than the Fe and Co/DHN/PE4LE microemulsion catalysts. The decrease in activity suggests the possibility of catalyst poisoning by the PE4LE. Moreover, other factors, such as differences in catalyst properties, may also be important. In the present study, even though catalysts prepared from microemulsions did not show greater activity than the same catalyst prepared from the corresponding organometallic compound, they served as useful model catalysts since the size of the solid metal particles before the hydroconversion reaction was well characterized.  6.2 Recommendations Even though catalysts prepared in microemulsions had small particle size and were well dispersed in the oil phase, they are not attractive as commercial catalysts at the present time. The hydroconversion activity and the cost of the catalyst prepared in microemulsions was not superior to that of the conventional organometallic compound catalyst (Appendix DI). The present study points to an effect of surfactant (PE4LE) on catalyst activity but the exact mechanism is not known. However, the addition of various surfactants to the cobalt naphthenate precursor and/or to DHN would clearly show the  137  effect of each surfactant on both the activity and selectivity of the catalyst. It is recommended that such tests be completed in any future work. The hydroconversion activity of the catalyst prepared in microemulsions could be improved by reducing catalyst particle size and by reducing any possible effects of surfactant. The effect of surfactant could be reduced by using surfactant that adsorbs less to the catalyst surface or using less surfactant.  1. Using surfactant that provides suitable HLB for the system Microemulsions with suitable hydrophile-lypophile balance (HLB) provide small catalyst particles. By varying the surfactant with different HLB values one can identify a suitable surfactant which provides smallest particle size in the nm range for that particular system (Wilcoxon et al., 1993).  2. Preparing the catalyst at higher temperature to reduce surfactant content It is possible to prepare metal catalysts with a smaller amount of surfactant by preparing the catalyst in a microemulsion at higher temperature. The metal particles prepared at higher temperature were found to be smaller and monodisperse (Wilcoxon et al.,1993). After high temperature reduction of the catalyst in the microemulsion, the amount of surfactant in the microemulsion can be reduced by lowering the microemulsion temperature to room temperature. The surfactant will be phase separated. By removing the excess surfactant in the lower phase one should obtain catalyst dispersed in the organic solvent with less surfactant (Wilcoxon et al., 1993). The reduced amount of surfactant also  138  reduces the cost of the catalyst prepared in the microemulsion. The cost can be further reduced if a light end product is used as the solvent for the microemulsion and if the reducing metal ions are not reduced by a reducing agent but are sulfided  in situ  under  reaction condition. The effect of this approach to catalyst synthesis on catalyst activity needs to be determined.  3. Using a suitable surfactant that gives a less negative effect on catalytic activity A surfactant that does not block the accessibility of hydrogen gas and hydrocarbon molecules, does not poison the catalyst and is not easily disintegrated needs to be identified. Experiments should be conducted with likely candidates to establish their effectiveness.  139  References  Albert, M.R. and Yates, Jr. J.T., The Surface Scientist's Guide to Organometallic  Chemistry, American Chemical Society, Washington, DC, 1987.  Anderson, R.R. and Bockrath, B.C., Fuel, 1984, 63 (3), p. 329-333.  Bearden, R. and Aldridge, C.L., U.S. Patent 4,134,825 to Exxon Research and  Engineering Co., 1979, January 16.  Bearden, R. and Aldridge, C.L., Energy Progress, 1981, 1 (1-4), p. 44-48.  Bland, W.F. and Davidson, R.L., Eds., Petroleum Processing Handbook, Mc  Graw HilL New York, 1967, p. 3.  Boutonnet, M., Kizhng, J., and Stenius, P., Colloid and Surfaces, 1982, 5, p. 209225.  Boutonnet, M., Kizhng, J., Mintsa-Eya, V., Choplin, A., Touroude, R., Make, G., and Stenius, P., J. of Catalysis, 1987, 103, p. 95-104.  140  Chen, H.H., Montgomery, D.S., and Strausz, O.P., AOSTRA Journal of Research, 1988, 4, p. 45-57.  Chen, H.H., Montgomery, D.S., and Strausz, O.P., AOSTRA Journal of Research, 1988, 4, p. 143-152.  Chen, H.H., Montgomery, D.S., and Strausz, O.P., AOSTRA Journal of Research, 1989, 5, p. 33-48.  Chen, H.H., Moljelsky, J.D., Payzant, E.M., Lown, E.M., Henry. D., Wallace, D., and Strausz, O.P., AOSTRA Journal of Research, 1991, 7, p. 17-35.  Choudhary, N. and Saraf, D.N., Ind. Eng. Chem., Prod, Res. Dev., 1975, 14, 2, p. 74-83.  Curtis, C.W., Tsai, K J . , and Guin, J.A., Ind. Eng. Chem. Res., 1987, 26, p. 12-18.  Dabkowski, M.J., Shih, S.S., and Albinson, K R , AIChE Symposium Series, 1991, 87 (282), p. 53-61.  Dautzenberg, F.M. and De Deken, J.C, Catal. Rev.-Sci. Eng., 1984, 26 (3-4), p. 421-444.  141  Dautzenberg, F.M. and De Deken, J.C, In Proceedings of Symposium on Developments in Hydrodemetallization Catalysts, A.C.S. Div Petroleum Chem  Miami Beach Meeting, 1985, April 28-May 3, p. 8-20.  Del Bianco, A , Panariti N., Di Carlo, S., Elmouchnino, J., Fixari, B., and Le Perchec, P., Applied Catalysis A: General, 1993, 94, p.  1-16.  Delmon, B., in Proceeding of the Third International Conference on the  Chemistry and Uses ofMolydenum, Ann Arbor, Michigan, 1979, August 19-23, p.  73-85.  Elliot, J.J. and Melchior, M.T., Kirk-Othmer Encyclopedia of Chem. Tech., 3rd  ed., John Wiley and Sons, New York, 1982, 17, p. 119.  Fendler, J.H., Chem. Rev., 1987, 87, p. 877-899.  Freund, J.E. and Smith, K M . , Statistics: A First Course, 4th ed., Prentice-Hall  Inc., New Jersey, 1986.  Gates, B.C., Katzer, J.R., Schuit, G.C.C., Chemistry of Catalytic Processes,  McGraw-HilL New York, 1979.  142  Gray, K M . , Upgrading Petroleum Residues and Heavy Oils, Marcel Dekker Inc.,  New York, 1994.  Hall, A.G., M.A.Sc. Thesis, University of British Columbia, Vancouver, BC,  Canada, May, 1996.  HalL G. and Herron, S.P., in Chem. of Asphaltenes, J.W. Bunger and N.C. L i ,  eds., Adv. Chem Ser., 1981, 195, p. 137.  Heck, K H . , Rankel, L.A., and Di Guiseppi, F.T., Fuel Processing Technology,  1992, 30, p. 69-81.  Heimenz, P . C , Principles of Colloid and Surface Chemistry, 2nd ed., Marcel  Dekker, New York, 1986.  Herrick, D.E., Tierney, J.W., and Wender, I., Energy & Fuels, 1990, 4, p. 231-  236.  Hirschon, A S . and Wilson Jr., K B . , Am Chem Soc Prep DivFuel Chem, 1989, 34  (3), p. 881-885.  143  Hunter, K J., Foundations of Colloid Science, Vol II, Oxford University Press, New York, 1989.  Johnson, R and Bhattacharyya, G., Statistics: Principles and Methods, John Wiley & Sons, New York, 1985.  Kennepohl, D. and Sanford, Energy & Fuel, 1996, 10, p. 229-234.  Kim, H, Curtis, C.W., Cronauer, D . C , and Sajkowski, D.J., A.C.S. Prep DivFuel Chem, 1989, 34 (4), p. 1431-1438.  Kizling, J. and Stenius, P., J. of Colloid and Interface Science, 1987, 118 (2), p.  482-492.  Laine, J. and Trimm, D.L., J. Chem Tech. Biotechnol, 1982, 32, p. 813-833.  Lee, J.D., Concise Inorganic Chemistry, 4th ed., New York, 1991, p. 267-268.  Lee, D . K , Yoon, W.L., and Woo, S.L, Fuel, 1996, 75 (10), p. 1186-1192.  Lindman, B., Kamenka, N., Brun, B., and Nilson, P.G., Microemulsions , Robb, I.D., ed., Plenum Press, New York, 1982, p. 115.  144  Leung, R., Hou, M.J., and Shah, D.O., Surfactant Sci. Ser., 1988, 28, p. 315-367.  Liu, C , Zhou, J., Que, G., Liang, VV., and Zhu, Y., Fuel, 1994, 73 (9), p. 15441550.  Liu, J.K. and Gunning, H.E., Syncrude Analytical Methods Manual for Bitumen  Upgrading, AOSTRA Technical Publication Series #11, Alberta Oil Sands Technology and Research Authority, Edmonton, Alberta, 1991.  Lott, R . K , Lee, L.K., and Cyr, T., Canadian Patent no. 2,025, J25 to Alberta Oil Sands Technology and Research Authority, Canada, 1990-1993.  Lui, S. and Straub, B.R., Reverse Micelle, Plenum Press, New York, 1984.  Martino, A , Wilcoxon, J.P., and Kawola, J.S., Energy & Fuels, 1994, 8, p. 12891295.  Meille, V., Schulz, E., Lemaire, M., and Vrinat, M., J. of Catalysis, 1997, 170, p. 29-36.  Miki, Y., Yamadaya, S., Oba, M., and Sugimoto, Y., J. Catalysis., 1983, 83, p. 371-383.  145  Mochida, I., Zhao, X.Z., and Skanishi, K , Ind. Eng. Chem. Res., 1990, 29, p.  334-337.  Nagy, J.B., Gourgue, A., and Derouane, E.G., Preparation of Catalysts III,  Elsevier, Amsterdam, 1983, p. 193-200.  Nagy, J.B., Deroune, E.J., Gourgue, A , Lufhnpadio, N., Ravet, I., Verfailhe, J.P., Proceedings of the Sixth International Symposium on Surfactants in Solution-  Modern Aspects, New Delhi, India, 1986, Aug 18-22.  Newson, E., A.C.S. Reprints, Division of Fuel Chemistry, 1972, 17 (2).  Oballa, M . , Wong, C , and Krzywicki, A., Presentation to AIChE Spring National  Meeting, Houston, 1993.  Oelderik, J.M., Sie, S.T., and Bode, D, Applied Catalysis, 1989, 47, p. 1-24.  Ott, L., An Introduction to Statistical Methods and Data Analysis, 3rd ed., PWS-  Kent Pubhshing Co., Boston, 1988.  Pileni, M.P., Zemb, T., and Petit, C.J., Chem Phys. Lett., 1985, 118, p. 414.  146  Pileni, M.P. and Lisiecki, I., Colloids and Surfaces, 1993, 80 (1), p. 63-68.  Ravet, I., Gourgue, A., Gabelica, Z., and Nagy, J.B., Proceedings of International Congress on Catalysis (VIII), Berlin, 1984, 4, p. 871-878.  Ravet, I., Nagy, J.B., and Derouane, E.G., Preparation of Catalysts IV, Elsevier, Amsterdam, 1987, p. 505-517.  Robinson, B.H., Steytler, D . C , and Fletcher, P.D.I, Surface Active Agents, Society of Chemical Industry, A Symposium held at Nottingham University, England, 1979, September 26-28.  Rueda, N , Bacaud, R., and Vrinat, M., J. of Catalysis, 1997, 169, p. 404-406.  Sanford, E . C , Ind. Eng. Chem. Res., 1994, 33, p. 109-117.  Sanford, E . C , Canadian Chemical News, 1996, 48 (1), p. 21-23.  Sanford, E.C. and Chung, K.H., AOSTRA Journal of Research, 1991, 7, p. 37.  Schick, M.J., Nonionic Surfactants, Marcel Dekker, Inc., New York, 1987.  147  Schinoda, K and Hirai, T., J Phys. Chem., 1977, 81, p. 1842.  Schott, H., Journal of Colloid and Interface Science, 1973, 43 (1), p. 150-155.  Schuetze, B. and Hofinann, H., Hydrocarbon Process, 1984, 63 (2), p. 75.  Shaw, D.J., Introduction to Colloid and Surface Chemistry, 3rd ed., Butterworth & Co., 1986.  Shaw, J.M., Gaikwad, K R , Stowe, D.A., Fuel, 1988, 67, p. 1554-1559.  Smith, C M . and Savage, P.E., AIChEJ,  1991, 37, p. 1613-1624.  Speight, J.G., The Structure of Petroleum Asphaltenes, in Formation Series, Alberta Research Council, 1978, 81.  Speight, J.G., 178th Nat. Meet. Am. Chem. Soc, Washington, D . C , 1979, Sept.  Speight, J.G., Presentation to Symposium on the Role of Asphaltenes in Petroleum Exploration,  13-18, p. 200.  Production and Refining,  San Diego, CA, 1994, March  Speight, J.G. and Moschopedis, S.E., in Chem. of Asphaltenes, J.W. Bunger and N.C. L i , eds., Adv. Chem. Ser., 1981, 195, p. 1.  Stanislaus, A. and Cooper, B.H., Catal. Rev.-Sci. Eng., 1994, 36 (1), p. 75-123.  Stanislaus, A., Halabi, M.A., and Khan, Z., Presentation toAIChE Spring National Meeting, Houston, 1993.  Strausz, O.P., Bitumen and Heavy Oil Chemistry. In AOSTRA Technical Handbook on Oil Sands, Bitumens and Heavy Oils; Hepler, L.G., Hsi, C , Eds.; AOSTRA;  Edmonton, 1989.  Sundaram, K M . , Katzer, J.K, andBischoff, K B , Chem. Eng. Comm., 1988, 71, p. 53-71.  Thomas, M.G., Padrick, T.D., and StohL F.V., Fuel, 1982, 61, p. 761.  Topsoe, H., Clausen, B.S., Candia, K , Wivel, C , and Morey, S., J. Catal, 1981, 68 (2), p. 433.  Wei, X.Y., Ogata, E., Zong, Z.M., and Niki, E., Energy & Fuels, 1992, 6, p. 868869.  149  Weisser, P. andLanda, S., Sulfide Catalysts, Their Properties and Applications, Pergamon Press, New York, 1970.  Wilcoxon, J.P., Williamson, R X . , and Baughman, R., J. Chem. Phys., 1993, 98 (12), p. 9933-9950.  Zana, R. and Lang, J., in Solution Behavior of Surfactants, Vol. 2 (Mittal, K.L. andFendler, E.J., eds.), Plenum Press, New York,, 1982, p. 1195.  Zhang, S.F., Herod, A.A., and Kandiyoti, R., Fuel, 1996, 76 (1), p. 39-49.  150  Appendix I  1.1 Hydrocracking activity experiments 1.2 Summary of results of hydrocracking activity experiments  151  1.1 Hydrocracking activity experiments Experiment 1 Thermal hydrocracking 1. Hydrocracking condition Initial pressure of H and H S in the reactor 500 psig. Reaction temperature 430°C for 1 h. Rate of heating 5°C/min. 2  Time (h:min) 0:0 0:10 0:20 0:30 0:40 0:50 1:0 1:10 1:20 1:28 2:28 stop & cool  2  Reaction vessel temperature (°C) 17 31 83 142 192 242 292 342 392 430 430 19  Furnace temperature (°C) 18 52 101 203 232 279 317 365 411 446 429  -  Pressure (psig) 500 500 590 675 710 790 840 900 960 1030 1260 500  Speed of impeller (rpm) 0 0 333 336 319 336 335 335 332 334 336  -  2. GC Analysis Gas inlet Pressure (psig) 500  Volume (ml) 220  Temperature (K) 290  Gas Outlet Component Standard moles/area Area/lml gas injection Average Moles/lml gas injection Mole traction Weight (g)  co  Weight (g) 1.132143  n (moles) 0.314484  CH, 2.41x10""  C H, 1.62x10""  CjHe 1.25x10""  1.59x10"  HS 1.95x10""  Air 2.93x10""  311288 331432 311954 318224.7 7.65x10"*  2338 2327 2247 2304 3.73x10  334362 328718 321019 328033 4.09x10"*  7213 7032 10192 8145.667 1.3xl0"  307106 301986 295945 301679 5.89x10*  180090 185138 204689 189972.3 5.57x10"*  1.82xl0"  4.16xl0"  0.212463  0.001034  0.113423  0.003604  0.163584  -  0.505892  1.000000  1.061738  0.009047  1.06276  0.04953  1.737135  -  0.31601  4.23622  2  s  2  2  7  H -  Total  2  -  5  5  3. Mass balance Mass before reaction Weight of residue 80 g. Weight of H and H S in the reactor before reaction 1.1 g. Mass after reaction Weight of coke 12.0 g/80 g residue = 7.5 g/50 g residue. Weight of liquid hydrocarbon 63.9 g. Weight of gas mixture 4.2 g. % Liquid yield = (63.9/80) x 100 = 79.9 % Recovery = (80.1/81.1) x 100 = 98.8 2  2  152  Experiment 2 Hydrocracking of residue using Co/DDAB/toluene microemulsion 1. Hydrocracking condition Initial pressure of H and H S in the reactor 500 psig. Reaction temperature 430°C for 1 h. Rate of heating 5°C/min. 2  Time (h:min) 0:0 0:10 0:20 0:30 0:40 0:50 1:0 1:10 1:20 1:21 2:21 stop & cool  2  Reaction vessel temperature (°C) 22 76 128 176 226 275 325 375 425 430 430 19  Furnace temperature (°C) 23 101 155 193 235 278 320 365 426 432 425 19  Pressure (psig) 500 500 520 575 640 710 850 1000 1325 1340 1690 450  Speed of impeller (rpm) 353 371 361 355 353 351 350 348 347 349 356 355  2. GC Analysis Gas inlet Pressure (psig) 500  Volume (ml) 203  Temperature (K) 295  n (moles) 0.285265  Weight (g) 1.026953  Gas Outlet Component Standard moles/area Area/1 ml gas injection  Average Moles/lml gas injection Mole fraction Weight (g)  CH, 2.41x10""  C H, 1.62x10"  CjH« 1.25x10""  1.59x10"  co  HS 1.95x10"  Air 2.93x10""  H -  297770 360275 351916 353777 368721 368040 344412 349273 8.4x10-"  1021 1188 1207 1217 1216 1261 1099 1172.714 1.9xl0"  73017 84595 82098 79466 84785 82226 75772 80279.86 lxlO"  20731 24526 17364 32082 22151 18642 19832 22189.71 3.54x10"'  41710 48442 47550 45641 45167 43295 40049 44550.57 8.7x10"  172656 211898 220499 214078 225488 225062 207568 211035.6 6.19xl0"  2.48xl0"  4.16xl0"  0.237261  0.000536  0.028242  0.009989  0.024579  -  0.699393  1.000000  0.984635  0.003891  0.219761  0.114004  0.216755  -  0.362812  1.901857  2  8  6  2  2  7  6  Total  2  5  5  153  3. Mass balance Mass before reaction Residue 50.0 g. Weight of residue and microemulsion of Co in toluene 91.0 Weight of H and H S in the reactor before reaction 1.0 g. 2  2  Mass after reaction Weight of coke 7.1 g. Weight of liquid hydrocarbon 78.8 g. Weight of gas mixture 1.9 g. % Liquid yield = (78.8/91) x 100 % Recovery = (87.8/92) x 100  = 86.6 = 95.4  Experiment 3 Hydrocracking of residue using cobaltnaphthenate 1. Hydrocracking condition Initial pressure of H and H S in the reactor 500 psi. Reaction temperature 430°C for 1 h. Rate of heating 5°C/min. 2  2  Reaction vessel temperature (°C) 23 82 132 182 235 278 328 375 430 430 20  Time (h:min) 0:0 0:10 0:20 0:30 0:40 0:50 1:0 1:10 1:11 2:11 stop & cool  Furnace temperature (°C) 22 116 181 245 304 357 410 450 462 425 -  Pressure (psig) 500 n/r n/r n/r n/r n/r n/r n/r n/r n/r 350  Speed of impeller (rpm) 358 352. 347 348 347 343 340 339 352 352 352  n/r = not recorded 2. GC Analysis Gas inlet Pressure (psig) 500  Gas Outlet Component Standard moles/area Area/1 ml gas injection Average Moles/1 ml gas injection Mole fraction Weight (g)  Volume (ml) 250  ca,  n (moles) 0.350124  Temperature (K) 296  co  Weight (g) 1.260448  2.41x10"  1.62x10""  CH4  CA 1.25x10"  1.59x10"  HS 1.95x10""  Air 2.93x10""  H -  342552 367241 372576 360789.7 8.68xl0"  1800 1962 1945 1902.333 3.08x10-"  183136 195395 190490 189673.6 2.36xl0  7071 7180 12356 8869 1.41xl0"  141666 153070 148330 147688.7 2.89xl0"  120315 126615 131450 126126.7 3.7x10'  2.38x10"  4.16xl0"  0.228981  0.000812  0.062343  0.00373  0.076127  -  0.628007  1.000000  0.907119  0.005629  0.463075  0.040639  0.640859  -  0.310985  2.368305  2  6  -6  2  2  7  6  6  Total  2  5  5  3. Mass balance Mass before reaction Residue 50 g Weight of residue and Co naphthenate 50.3 g Weight of H and H S in the reactor before reaction 1.3 g. Mass after reaction Weight of coke 6.3 g Weight of liquid hydrocarbon 38.5 g Weight of gas mixture 2.4 g. % Liquid yield = (38.5/50.3) x 100 = 76.5 % Recovery = (47.2/51.6) x 100 = 91.5 2  2  155  Experiment 4 Hydrocracking of residue using Mo/DDAB/toluene microemulsion 1. Hydrocracking condition Initial pressure of H and H S in the reactor 500 psi. Reaction temperature 430°C for 1 h. Rate of heating 5°C/min. 2  Time (h:min) 0:0 0:10 0:20 0:30 0:40 0:50 1:0 1:10 1:20 1:22 2:22 stop & cool  2  Reaction vessel temperature (°C) 18 76 120 170 220 270 320 370 420 430 430 25  Furnace temperature (°C) 18 105 145 187 228 269 310 353 398 408 -  Pressure (psig) 500 575 620 690 760 860 1000 1110 1250 1290 425  Speed of impeller (rpm) 357 361 353 357 359 358 358 357 354 358 -  2. GC Analysis Gas inlet Pressure (psig) 500  Gas Outlet Component Standard moles/area Area/1 ml gas injection Average Moles/1 ml gas injection Mole fraction Weight (g)  Volume (ml) 250  ca,  Temperature (K) 291  n (moles) 0.289186  Weight (g) 1.041069 Total  2.41x10""  1.62x10"  CjHo 1.25x10""  C0 1.59x10"  HS 1.95x10"  Air 2.93x10""  B -  326766 305443 295173 321919 312325.3  80489 74078 70642 79893 76275.5 9.5xl0"  14228 13313 14412 15698 14412.75 2.3xl0"  50070 46276 43927 50074 47586.75 9.3x10"  186763 174889 157552 203850 180763.5 5.3x10"*  -  7.51X10-  1056 866 872 987 945.25 1.53x10""  2.67xl0'  4.16xl0"  0.206973  0.000421  0.026177  0.00633  0.025612  -  0.734487  1.000000  0.794888  0.002831  0.188503  0.06685  0.209021  -  0.352604  1.614698  C2H4  5  7  2  7  2  7  2  5  5  3. Mass balance Mass before reaction Residue 50 g. Weight of residue and microemulsion of Mo in toluene 91.0 g. Weight of H and H S in the reactor before reaction 1.0 g. Mass after reaction Weight of coke 6.1 g. Weight of liquid hydrocarbon 78.5 g. Weight of gas mixture 1.6 g. % Liquid yield = (78.5/91.0) x 100 = 86.3 % Recovery = (86.2/92.0) x 100 = 93.7 2  2  156  Experiment 5 Hydrocracking of residue using Fe/DDAB/toluene microemulsion. 1. Hydrocracking condition Initial pressure of H and H S in the reactor 500 psi. Reaction temperature 430°C for 1 h. Rate of heating 5°C/min. 2  Time (h:min) 0:0 0:10 0:20 0:30 0:40 0:50 1:0 1:10 1:20 2:20 stop & cool  2  Reaction vessel temperature (°C) 31 89 147 179 230 280 330 380 430 430 19  Furnace temperature (°C) 28 115 180 191 236 277 320 366 421 -  -  Pressure (psig) 500 550 625 650 750 850 960 1200 1500 1750 510  Speed of impeller (rpm) 350 357 356 354 355 355 356 356 358 367 -  2. GC Analysis Gas inlet Pressure (psig) 500  Volume (ml) 203  Temperature (K) 304  n (moles) 0.276819  Weight (g) 0.99655  Gas Outlet Component Standard moles/area Area/lml gas injection Average Moles/lml gas injection Mole fraction Weight (g)  CH, 2.41x10""  1.62x10"  C H« 1.25x10"  C0 1.59x10""  HS 1.95x10""  Air 2.93x10"  286669 274115 255686 262816 269821.5 6.49x10"  1068 1082 956 1004 1027.5 1.66x10  67543 65037 60271 61708 63639.75 7.93x10"'  22288 17755 14263 15468 17443.5 2.78x10"'  20189 20018 19503 19520 19807.5 3.87x10"'  213966 219578 192824 211884 209563 6.15x10"  2.75xl0"  4.16xl0"  0.183066  0.000469  0.022361  0.007843  0.010915  -  0.775346  1.000000  0.861024  0.003859  0.197197  0.101445  0.109087  -  0.455841  1.728453  C2H4  2  s  2  2  H -  Total  2  -  5  5  3. Mass balance Mass before reaction Residue 50 g. Weight of residue and microemulsion of Fe in toluene 91.0 g. Weight of H and H S in the reactor before reaction 1.0 g. Mass after reaction Weight of coke 7.4 g. Weight of liquid hydrocarbon 80.8 g. Weight of gas mixture 1.7 g. % Liquid yield = (80.8/91.0) x 100 = 88.8 % Recovery = (89.9/92.0) x 100 = 97.7 2  2  157  Experiment 6 Hydrocracking of residue using Mo/PE4LE/DHN microemulsion 1. Hydrocracking condition Initial pressure of H and H S in the reactor 500 psi. Reaction temperature 430°C for 1 h. Rate of heating 5°C/min. 2  2  Reaction vessel temperature (°C) 20 83 124 174 224 269 319 369 419 430 430 20  Time (h:min) 0:0 0:10 0:20 0:30 0:40 0:50 1:0 1:10 1:20 1:22 2:22 stop & cool  Furnace temperature (°C) 20 116 170 201 241 281 325 368 411 421 -  Pressure Opsig) 500 n/r n/r n/r n/r n/r n/r n/r n/r n/r n/r 400  Speed of impeller (rpm) 365 362 358 355 354 354 355 355 356 356 -  2. GC Analysis Gas inlet Pressure (psig) 500  Gas Outlet Component Standard moles/area Area/lml gas injection Average Moles/1 ml gas injection Mole fraction Weight (g)  Volume (ml) 200  ca,  Temperature (K) 293  Weight (g) 1.018683  n (moles) 0.282967  2.41x10""  1.62x10""  c a,  CA 1.25x10"  1.59x10'"  co  HS 1.95x10""  Air 2.93x10""  H -  317981 300515 312623 319832 312737.8 7.52X10"  5487 5217 5473 5674 5462.75 8.84x10""  129998 125070 129846 132775 129422.3 1.61x10"*  41105 35091 39757 42620 39643.25 6.32xl0"  54433 52511 55185 56251 54595 1.07xl0"  272606 290905 339137 395265 324478.3 9.52xl0"  2.12xl0"  4.16xl0"  0.234472  0.002754  0.050252  0.019697  0.033244  -  0.659581  1.000000  0.849254  0.017458  0.341272  0.196193  0.255867  -  0.298624  1.958668  2  6  2  2  7  6  6  Total  2  5  5  3. Mass balance Mass before reaction Residue 50 g. Weight of residue and microemulsion of Mo in decalin 95.5 g. Weight of H and H S in the reactor before reaction 1.0 g. Mass after reaction Weight of coke 2.9 g. Weight of liquid hydrocarbon 88.3 g. Weight of gas mixture 2.0 g. % Liquid yield = (88.3/95.5) x 100 = 92.5 % Recovery = (93.2/96.5) x 100 = 96.6 2  2  158  Experiment 7 Hydrocracking of residue using Fe/PE4LE/DHN microemulsion 1. Hydrocracking condition Initial pressure of H and H S in the reactor 500 psi. Reaction temperature 430°C for 1 h. Rate of heating 5°C/min. 2  Time (h:min) 0:0 0:10 0:20 0:30 0:40 0:50 1:0 1:10 1:20 1:22 2:22 stop & cool  2  Reaction vessel temperature (°C) 20 66 120 170 214 270 320 370 420 430 430 15  Furnace temperature (°C) 21 85 156 195 232 282 327 372 415 424 412 24  Pressure (psig) 500 n/r n/r n/r n/r n/r n/r n/r n/r n/r n/r 400  2. GC Analysis Gas inlet Pressure (psig,) 500  Volume (ml) 200  Component Standard moles/area Area/lml gas injection Average Moles/lml gas injection Mole fraction Weight (g)  n (moles) 0.282967  Temperature (K) 293  Gas Outlet  co  Speed of impeller (rpm) 358 356 362 361 358 357 355 354 354 353 354 354  Weight (g) 1.018683  CK, 2.41x10-"  1.25x10-"  1.59x10-"  HS 1.95x10"  Air 2.93x10""  H  1.62x10-"  298257 320053 335721 337493 322881 7.77x10-"  2616 2772 2805 2585 2694.5 4.36xl0"  112893 119115 122630 117293 117982.8 1.47x10-"  41914 46540 52365 67819 52159.5 8.31xl0"  52950 57433 58679 56892 56488.5 1.1x10""  242248 265636 275861 264290 262008.8 7.68x10""  2.27xl0"  4.16xl0"  0.229  0.001285  0.043336  0.024516  0.032539  -  0.669324  1.000000  0.843835  0.008288  0.299411  0.248431  0.254788  -  0.308296  1.963049  C2H4  8  2  2  7  Total  2  -  5  5  3. Mass balance Mass before reaction Residue 50 g. Weight of residue and microemulsion of Fe in decalin 95.3 g. Weight of H and H S in the reactor before reaction 1.0 g. Mass after reaction Weight of coke 3.7 g Weight of liquid hydrocarbon 85.2 g. Weight of gas mixture 2.0 g. % Liquid yield = (85.2/95.3) x 100 = 89.4 % Recovery = (90.9/96.3) x 100 = 94.4 2  2  159  Experiment 8 Hydrocracking of residue using Co/PE4LE/DHN microemulsion l.Hydrocracking condition Initial pressure of H and H S in the reactor 500 psi. Reaction temperature 430°C for 1 h. Rate of heating 5°C/min. 2  2  Reaction vessel temperature (°C) 19 76 107 170 223 271 321 371 421 430 430 20  Time (h:min) 0:0 0:10 0:20 0:30 0:40 0:50 1:0 1:10 1:20 1:22 2:22 stop & cool  Furnace temperature (°C) 19 91 164 203 242 283 327 372 417 426 414 30  Pressure (psi) 500 560 610 640 660 700 750 810 900 940 1200 410  Speed of impeller (rpm) 353 352 354 358 350 353 352 353 352 354 355 358  2. GC Analysis Gas inlet Pressure (psig) 500  Volume (ml) 198  Gas Outlet Component Standard moles/area Area/lml gas injection Average Moles/lml gas injection Mole fraction Weight (g)  Temperature (K) 292  CH6  co  Weight (g) 1.01195  n (moles) 0.281097  Total  1.25x10'"  1.59x10'"  HS 1.95x10"  Air 2.93x10"  H -  101822 96021 100105 114258 103051.5 1.28X10  47317 41389 47906 56496 48277 7.7xl0'  53755 51048 52265 60485 54388.25 1.06xl0'  257279 247061 242158 285737 258058.8 7.57xl0"  -  6.74X10*  2611 2439 2599 2823 2618 4.23xl0'  2.41xl0'  4.16xl0"  0.19793  0.001244  0.037722  0.022614  0.031222  -  0.709267  1.000000  0.727473  0.008004  0.25996  0.228568  0.243853  -  0.325856  1.793714  CH, 2.41x10'"  C H, 1.62x10'"  277040 255789 273336 313941 280026.5  2  8  -  2  -6  2  2  7  6  6  2  5  5  3. Mass balance Mass before reaction Residue 50 g Weight of residue and microemulsion of Co in decalin 97.0 g Weight of H and H S in the reactor before reaction 1.0 g. Mass after reaction Weight of coke 3.3 g Weight of liquid hydrocarbon 89.0 g Weight of gas mixture 1.8 g. % Liquid yield = (89.0/97.0) x 100 = 91.8 % Recovery = (94.1/98.0) x 100 = 96.0 2  2  160  Experiment 9 Hydrocracking of residue using Mo/PE4LE/n-hexane microemulsion 1. Hydrocracking condition Initial pressure of H and H S in the reactor 500 psi. Reaction temperature 430°C for 1 h. Rate of heating 5°C/min. 2  2  Time (h:min) 0:0 0:10 0:20 0:30 0:40 0:50 1:0 1:10 1:20 1:30 1:40 1:50 2:50 stop&cool  Reaction vessel temperature (°C) 17 69 123 168 218 268 317 355 368 413 418 430 430 19 GC Anal ysis  2. Gas inlet  Pressure (psig) 500  Furnace temperature (°C) 17 101 153 185 229 270 314 363 427 462 481 463 456 20  Volume (ml) 201  Temperature (K) 290  Gas Outlet Component Standard moles/area Area/lml gas injection  Average Moles/lml gas injection Mole fraction Weight (g)  Pressure (psig) 500 580 640 700 800 980 1100 1310 1450 1580 1750 1850 2080 380  co  Speed of impeller (rpm) 358 351 345 352 353 353 348 352 353 355 360 358 358  Weight (g) 1.034367  n (moles) 0.287324  CH, 2.41x10""  1.62x10""  CjH* 1.25x10""  1.59x10"  HS 1.95x10"  Air 2.93x10""  H -  347370 297494 341526 361585 354528 340500.6 8.19x10""  1424 1214 1258 1523 1423 1368.4 2.21xl0"  154898 128977 148073 160041 158346 150067 1.87X10"  38535 32692 36851 34701 38958 36347.4 5.79xl0"  59526 49907 56644 60470 60278 57365 1.12x10"  206949 174087 200120 217654 205719 200905.8 5.89x10"  -  2.39xl0"  4.16xl0"  0.229377  0.00062  0.052354  0.016227  0.031385  -  0.670037  1.000000  0.795923  0.003764  0.340622  0.15484  0.231422  -  0.290623  1.817194  C2H4  8  6  2  2  7  Total  2  5  5  3. Mass balance Mass before reaction Residue 50 g. Weight of residue and microemulsion of Mo in n-hexane 82.3 g. Weight of H and H S in the reactor before reaction 1.0 g. Mass after reaction Weight of coke 9.8 g. Weight of liquid hydrocarbon 63.9 g. Weight of gas mixture 1.8 g. % Liquid yield = (63.9/82.3) x 100 = 77.6 % Recovery = (75.5/83.3) x 100 = 90.6 2  2  161  Experiment 10 Hydrocracking of residue using Mo/DDAB/THF microemulsion. 1. Hydrocracking condition Initial pressure of H and H S in the reactor 500 psi. Reaction temperature 430°C for 1 h. Rate of heating 5°C/min. 2  2  Time Reaction vessel Speed of impeller Furnace Pressure (h:min) (rpm) temperature (°C) temperature (°C) (psig) 0:0 18 18 500 0:10 70 88 580 351 0:20 124 149 620 350 0:30 168 179 700 349 0:40 217 218 810 348 0:50 267 261 990 348 315 351 1:0 306 1190 366 1:10 367 1450 351 1:20 401 424 1720 350 1:30 398 447 2380 349 1:33 398 447 2500 349 stop * 1:46 350 380 2200 344 1:50 366 410 2310 364 2:04 398 465 361 keep 2500 2:10 418 455 360 keep 2500 2:22 423 457 352 keep 2500 2:31 430 463 354 keep 2500 3:31 430 354 keep 2500 stop&cool 21 22 * stop furnace and release some gases to reduce pressurefrom2500 to 2100 psi  2. GC Analysis Gas inlet Pressure (psig) 500  Volume (ml) 200  Temperature (K) 291  n (moles) 0.284912  Weight (g) 1.025684  Gas Outlet Component Standard moles/area Area/lml gas injection  Average Moles/1ml gas injection Mole fraction Weight (g)  CH, 2.41x10""  C2H4  C2H6  1.25x10""  C0 1.59x10"  HS 1.95x10""  Air 2.93x10""  H  1.62x10""  374994 426159 324421 342395 341908 392763 367106.7 8.83x10"*  1113 1329 1159 1096 1070 1273 1173.333 1.9x10""  153101 175586 133510 141308 139630 161397 150755.3 1.88x10"*  232223 266108 202901 214075 213129 246032 229078 3.65x10"*  27422 32273 24584 25916 25195 29393 27463.83 5.36x10"'  241770 267511 208219 225443 227780 261949 238778.7 7x10"*  -  0.25524  0.000549  0.054283  0.105551  0.015508  -  -  -  -  -  2  2  -  Total  2  -  1.97xl0"  4.16xl0"  0.568869  1.000000  -  -  5  5  162  3. Mass balance Mass before reaction Residue 50 g. Weight of residue and microemulsion of Mo in THF 94.7 g. Weight of H and H S in the reactor before reaction 1.0 g. Mass after reaction Weight of coke 21.5 g. Weight of liquid hydrocarbon - g. Weight of gas mixture - g (some gases were released during reaction period). % Liquid yield = % Recovery = 2  2  163  Experiment 11 Hydrocracking of residue using Co/PE4LE/n-hexane microemulsion. 1. Hydrocracking condition Initial pressure of H and H S in the reactor 500 psig. Reaction temperature 430°C for 1 h. 2  2  Rate of heating 5°C/min. Time Reaction vessel (h:min) temperature (°C) 0:0 21 0:10 71 0:20 121 0:30 171 0:40 221 0:50 271 1:0 314 1:10 357 1:20 398 1:30 407 1:40 423 1:47 430 2:47 430 stop&cool 22 2. GC Anal ysis Gas inlet Pressure (psig) 500  Furnace temperature (°C) 22 96 146 185 228 272 319 410 454 472 462 466 461 22  Volume (ml) 199  Pressure OjJsig) 500 550 600 700 810 1000 1190 1480 1680 1810 1990 2020 2200 400  Temperature (K) 294  Speed of impeller (rpm) 355 351 357 359 360 358 359 358 357 359 362 358 -  Weight (g) 1.010142  n (moles) 0.280595  Gas Outlet Component Standard moles/area Area/lml gas injection  Average Moles/lml gas injection Mole fraction Weight (g)  co  Total  CH, 2.41x10""  C H, 1.62x10""  1.25x10"  1.59x10""  HS 1.95x10""  Air 2.93x10""  H -  424013 448298 424108 443600 452141 438432 1.05xl0"  1362 1378 1406 1369 1410 1385 2.24xl0"  168309 176740 172044 178027 178768 174777.6 2.18xl0"  41718 47437 42037 45039 47055 44657.2 7.12xl0"  66047 69645 68002 69867 70709 68854 1.35xl0"  196706 214260 205725 215006 224380 211215.4 6.19xl0"  -  2.06xl0-  4.16xl0"  0.29787  0.000633  0.061496  0.020107  0.037993  -  0.581902  1.000000  1.066209  0.003964  0.412725  0.19792  0.288984  -  0.26036  2.230162  2  5  8  CJHS  6  2  2  7  6  6  2  5  5  3. Mass balance Mass before reaction Residue 50 g. Weight of residue and microemulsion of Co in n-hexane 84.0 g. Weight of H and H S in the reactor before reaction 1.0 g. Mass after reaction Weight of coke 9.9 g. Weight of liquid hydrocarbon 55.5 g. Weight of gas mixture 2.2 g. % Liquid yield = (55.5/84) x 100 = 66.1 (liquid came out with gas) % Recovery - (67.6/85.0) x 100 = 79.5 2  2  164  Experiment 12 Hydrocracking of residue using Fe/PE4LE/n-hexane microemulsion. 1. Hydrocracking condition Initial pressure of H and H S in the reactor 500 psig. Reaction temperature 430°C for 1 h. Rate of heating 5°C/min. 2  Time (h:min) 0:0 0:10 0:20 0:30 0:40 0:50 1:0 1:10 1:20 1:30 1:40 1:47 2:47 stop&cool  Reaction vessel temperature (°C) 16 67 115 165 215 265 315 364 382 405 421 430 430 17  2. GC Anal ysis Gas inlet Pressure (psig) 500  Gas Outlet Component Standard moles/area Area/lml gas injection  Average Moles/lml gas injection Mole fraction Weight (a)  2  Furnace temperature (°C) 16 95 140 182 225 271 315 403 452 482 470 474 467 22  Volume (ml) 200  Pressure (psig) 500 590 620 710 820 980 110 1390 1520 1800 2000 2100 2300 400  Temperature (K) 289  co  Speed of impeller (rpm) 360 353 351 348 350 351 349 349 350 351 350 351 355  Weight (g) 1.032782  n (moles) 0.286884  Total  CH) 2.41x10""  C H, 1.62x10"  C Hs 1.25x10"  HS 1.95x10'"  Air 2.93x10"  H  1.59x10"  343443 374810 409807 372766 381600 376485.2 9.06x10"*  1016 1025 1102 1067 1080 1058 1.71x10  120439 128961 141581 127114 130107 129640.4 1.61xl0"  26692 31777 33111 32404 31256 31048 4.95x10"'  46177 49575 55026 50762 52311 50770.2 9.92x10"'  162297 190547 217723 219732 210471 200154 5.87xl0"  -  2.36x10"'  4.16xl0'  0.253461  0.000479  0.0452  0.013852  0.02776  -  0.659247  1.000000  0.92753  0.003068  0.310138  0.139403  0.215871  -  0.301561  1.897569  2  s  2  6  2  2  6  2  -  5  3. Mass balance Mass before reaction Residue 50 g Weight of residue and microemulsion of Fe in n-hexane 83.4 g Weight of H and H S in the reactor before reaction 1.0 g. Mass after reaction Weight of coke 11.2 g Weight of liquid hydrocarbon 64.1 g Weight of gas mixture 1.9 g. % Liquid yield = (64.1/83.4) x 100 = 76.9 % Recovery = (77.2/84.4) x 100 = 91.5 2  2  165  Experiment 13 Hydrocracking of residue using Co/DDAB/THF microemulsion. 1. Hydrocracking condition Initial pressure of H and H S in the reactor 500 psig. Reaction temperature 430°C for 1 h. Rate of heating 5°C/min. 2  2  Reaction vessel temperature (°C) 18 69 117 167 216 266 316 365 399 394 422 430 430 17 2. GC Anal ysis Time (h:min) 0:0 0:10 0:20 0:30 0:40 0:50 1:0 1:10 1:20 1:30 1:40 1:49 2:49 stop&cool  Furnace temperature (°C) 18 88 142 185 228 276 329 391 450 483 483 489 481 23  Pressure (psig) 300 350 400 490 590 730 950 1200 1550 2500 keep 2500 keep 2500 keep 2500 500  Speed of impeller (rpm) 356 353 351 351 348 348 350 347 345 347 348 346 350 -  Gas inlet Pressure (psig) 300  Volume (ml) 200  Gas Outlet Component Standard moles/area Area/lml gas injection  Average Moles/lml gas injection Mole fraction Weight (g)  Temperature (K) 291  CH6  co  Weight (g) 0.61541  n (moles) 0.170947  CH, 2.41x10"  1.62x10"  1.25x10"  1.59x10""  HS 1.95x10"  Air 2.93x10"  H -  299762 370638 365307 368874 349243 350764.8 8.44xl0"  1181 1443 1247 1393 1339 1320.6 2.14xl0"  115474 143246 141583 143373 135318 135798.8 1.69xl0-  198787 245354 241828 244143 231635 232349.4 3.7xl0"  22219 27935 27907 27753 26740 26510.8 5.18xl0"  230688 298603 295453 308129 287195 284013.6 8.33xl0"  -  1.89xl0"  4.16xl0"  0.253602  0.000642  0.050847  0.111328  0.015567  -  0.568014  1.000000  -  -  -  -  -  -  -  -  6  8  2  6  2  2  6  7  6  Total  2  5  5  3. Mass balance Mass before reaction Heavy oil 50 g Weight of heavy oil and microemulsion of Co in THF 93.7 g Weight of H and H S in the reactor before reaction 0.6 g. Mass after reaction Weight of coke 21.2 g Weight of liquid hydrocarbon - g Weight of gas mixture - g (some gases were released during reaction period). % Liquid yield = % Recovery = 2  2  166  Experiment 14 Hydrocracking of residue using Fe/DDAB/THF microemulsion. 1. Hydrocracking condition Initial pressure of H and H S in the reactor 500 psig. Reaction temperature 430°C for 1 h. Rate of heating 5°C/min. 2  Reaction vessel temperature (°C) 16 72 118 167 217 267 316 367 399 397 427 430 430 15  Time (h:min) 0:0 0:10 0:20 0:30 0:40 0:50 1:0 1:10 1:20 1:30 1:40 1:47 2:47 stop&cool  2. GC Analysis Gas inlet Pressure (psig) 300  Gas Outlet Comp on ait Standard moles/area Area/lml gas injection  Average Moles/lml gas injection Mole fraction Weight (g)  2  Furnace temperature (°C) 16 91 138 180 224 271 323 385 441 478 471 474 464 15  Volume (ml) 200  Pressure (psig) 300 340 390 450 580 740 990 1220 1590 keep 2500 keep 2500 keep 2500 keep 2500 320  Temperature (K) 289  Speed of impeller (rpm) 352 350 354 353 354 353 353 353 352 355 357 359 359 351  n (moles) 0.17213  Weight (g) 0.619669  CR, 2.41xl0  -11  1.62x10"  c as 1.25x10"  co 1.59x10""  HS 1.95x10"" 2  Air 2.93x10""  H  265809 453540 444681 431411 398860.3 9.59x10"  876 1571 1590 1533 1392.5 2.25x10""  122064 203574 201602 197780 181255 2.26xl0"  216252 365427 357755 351360 322698.5 5.14xl0"  21127 35701 34917 35477 31805.5 6.21xl0"  230399 376011 368837 341285 329133 9.65x10"  -  1.43xl0"  4.16xl0"  0.300319  0.000705  0.070678  0.161022  0.01945  -  0.447826  1.000000  -  -  -  -  -  -  -  -  C2H4  6  2  6  2  6  7  6  -  Total  2  5  5  3. Mass balance Mass before reaction Residue 50 g. Weight of residue and microemulsion of Fe in THF 94.8 g. Weight of H and H S in the reactor before reaction 0.6 g. Mass after reaction Weight of coke 21.4 g. Weight of liquid hydrocarbon - g. Weight of gas mixture - g (some gases were released during reaction period). % Liquid yield = % Recovery = 2  2  167  Experiment 15 Hydrocracking of residue using cobaltnaphthenate/DHN. 1. Hydrocracking condition Initial pressure of H and H S in the reactor 500 psig. Reaction temperature 430°C for 1 h. Rate of heating 5°C/min. 2  2  Reaction vessel temperature (°C) 16 81 118 168 217 267 317 367 417 430 430 17  Time (h:min) 0:0 0:10 0:20 0:30 0:40 0:50 1:0 1:10 1:20 1:22 2:22 stop&cool  Furnace temperature (°C) 16 114 148 193 235 277 320 367 412 424 410 21  Pressure (psig) 500 600 630 680 730 790 840 920 1000 1010 1130 390  Speed of impeller (rpm) 355 361 356 355 358 354 364 367 363 366 371 370  2. GC Analysis Gas inlet Pressure (psig) 500  Volume (ml) 202  Temperature (K) 289  Gas Outlet Component Standard moles/area Area/lml gas injection Average Moles/lml gas injection Mole fraction Weight (g)  co  Weight (g) 1.04311  n (moles) 0.289753  CH, 2.41x10"  C H, 1.62x10"  C Hs 1.25x10"  1.59x10"  HS 1.95x10"  Air 2.93x10""  H -  273156 275593 263881 270001 270657.8  75681 75553 71170 73409 73953.25 9.21xl0"  23304 24298 19041 31083 24431.5 3.89xl0"  74941 74556 70285 72312 73023.5 1.43x10""  236510 229458 213542 224602 226028  -  6.51X10"  1307 1318 1236 1280 1285.25 2.08x10""  6.63X10"  2.57x10"  4.16xl0"  0.186169  0.000595  0.026344  0.011137  0.040794  -  0.734962  1.000000  0.670887  0.003749  0.178  0.110366  0.31239  -  0.331068  1.60646  2  6  2  7  2  2  7  6  Total  2  5  5  3. Mass balance Mass before reaction Residue 50 g Weight of residue and cobaltnaphthenate in decalin 93.2 g Weight of H and H S in the reactor before reaction 1.0 g. Mass after reaction Weight of coke 2.5 g. Weight of liquid hydrocarbon - g. Weight of gas mixture 1.6 g. % Liquid yield = (some liquid came out during releasing gases) % Recovery 2  2  168  Experiment 16  Thermal cracking II  1. Hydrocracking condition Initial pressure of H and H S in the reactor 500 psig. Reaction temperature 430°C for 1 h. Rate of heating 5°C/min. 2  Time (h:min) 0:0 0:5 0:10 0:15 0:20 0:25 0:30 0:35 0:40 0:45 0:50 0:55 1:00 1:05 1:10 1:14 1:15 1:20 1:25 1:30 1:35 1:40 1:45 1:50 1:55 2:00 2:05 2:10 2:14 stop&cool  2  Reaction vessel temperature (°C) 18 77 91 126 141 168 192 216 243 267 291 317 342 365 430 436 425 436 431 431 430 430 430 430 430 430 430 430 20  Furnace temperature (°C) 18 113 132 183 196 230 257 284 318 342 368 396 422 449 456 452 424 430 427 426 425 426 426 426 428 429 429 429  Pressure (psig) 500 600 620 680 700 730 780 800 850 880 910 950 990 1030 1100 1100 1130 1200 1270 1280 1300 1300 1300 1320 1350 1350 1370 1380 500  Speed of impeller (rpm) 353 353 364 354 357 341 349 359 352 358 353 354 351 357 357 353 356 355 357 356 353 353 352 358 360 356 356 369  169  2. GC Analysis Gas inlet Pressure (psig) 500  Volume (ml) 250  n (moles) 0.35614  Temperature (K) 291  Weight (g) 1.282105-  Gas Outlet Component Standard moles/area Area/lml gas injection  Average Moles/1 ml gas injection Mole fraction Weight (g)  Total  CH, 2.41x10""  C H, 1.62x10"  C H5 1.25x10"  co 1.59x10'"  HS 1.95x10"  Air 2.93x10"  H -  237839 302556 320310 287284 319395 284921 292050.8 7.02x10"  1547 1310 1184 1333 1189 1083 1274.333 2.06xl0"  123395 121044 129073 121479 118766 113262 121169.8 1.51x10"  5476 26322 23759 31787 26704 23663 22951.83 3.66x10"  100655 94664 102268 98873 95397 91633 97248.33 1.9x10""  197546 219517 237697 231062 232142 216610 222429 6.52x10""  -  2.43x10°  4.16x10°  0.200279  0.000588  0.043033  0.010431  0.054163  -  0.691505  1.000000  1.133452  0.005821  0.456639  0.162338  0.651376  -  0.489184  2.898808  2  6  8  2  2  7  2  2  3. Mass balance Mass before reaction Residue 50 g Weight of H and H S in the reactor before reaction 1.3 g. 2  2  Mass after reaction Weight of coke 7.5 g Weight of liquid hydrocarbon 37.7 g Weight of gas mixture 2.9 g. % Liquid yield = (37.7/50.0) x 100 % Recovery = (48.1/51.3) x 100  = 75.4 = 93.8  170  Experiment 17 Hydrocracking of residue using Mo/PE4LE/DHN microemulsion. 1 .Hydrocracking condition Initial pressure of H and H S in the reactor 500 psig. Reaction temperature 430°C for 1 h. Rate of heating 5°C/min. 2  Time (h:min) 0:0 0:5 0:10 0:15 0:20 0:25 0:30 0:35 0:40 0:45 0:50 0:55 1:00 1:05 1:10 1:15 1:20 1:22 1:25 1:30 1:35 1:40 1:45 1:50 1:55 2:00 2:05 2:10 2:15 2:20 2:22 stop&cool  2  Reaction vessel temperature (°C) 18 48 81 93 120 144 170 194 219 244 269 294 319 344 369 394 419 430 435 430 430 430 430 430 430 430 430 430 430 430 430 20  Furnace temperature (°C) 18 73 108 120 146 170 193 212 235 257 279 300 322 344 366 389 412 423 421 410 409 408 409 410 409 409 409 409 409 410 410 30  Pressure (psig) 500 570 590 600 620 650 680 700 730 750 790 810 850 890 920 970 1000 1030 1080 1080 1080 1100 1120 1150 1170 1190 1200 1200 1200 1220 1220 420  Speed of impeller (rpm) 362 360 357 365 365 364 358 353 353 354 354 357 357 359 361 360 361 363 363 362 365 363 364 365 363 363 365 364 364 364 366  2. GC Analysis Gas inlet Pressure (psig) 500  Volume (ml) 199  Temperature (K) 291  n (moles) 0.283488  Weight (g) 1.020556  Gas Outlet Component Standard moles/area Area/lml gas injection  Average Moles/lml gas injection Mole fraction Weight (g)  CR, 2.41x10"  C2H4  C H6  1.62x10""  1.25x10""  C0 1.59x10""  HS 1.95x10"  Air 2.93x10"  269625 309278 336661 328136 393100 327360 7.87x10"  12533 6140 6454 6138 7401 7733.2 1.25xl0"  107715 120025 131555 121067 147615 125595.4 1.56x10""  67051 30724 40372 39042 50267 45491.2 7.25x10"'  72676 80257 83302 81495 97456 83037.2 1.62x10""  220978 270945 280825 271699 320553 273000 8.01xl0"  2.17xl0"  4.16xl0"  0.234405  0.003724  0.046574  0.021587  0.04829  -  0.64542  1.000000  0.887003  0.02466  0.330451  0.224638  0.388309  -  0.305289  2.16035  6  2  7  2  2  6  H -  Total  2  -  5  5  3. Mass balance Mass before reaction Residue 50 g Weight of residue and microemulsion of Mo in decalin 96.1 g Weight of H and H S in the reactor before reaction 1.0 g. Mass after reaction Weight of coke 2.9 g Weight of liquid hydrocarbon 89.5 g Weight of gas mixture 2.2 g. % Liquid yield = (89.5/96.1) x 100 = 93.1 % Recovery = (94.6/97.1) x 100 = 97.4 2  2  172  Experiment 18 Hydrocracking of residue using Co/PE4LE/DHN microemulsion. 1. Hydrocracking condition Initial pressure of H and H S in the reactor 500 psig. Reaction temperature 430°C for 1 h. Rate of heating 5°C/rnin. 2  Time (h:min) 0:00 0:05 0:10 0:15 0:20 0:25 0:30 0:35 0:40 0:45 0:50 0:55 1:00 1:05 1:10 1:15 1:20 1:25 1:30 1:35 1:40 1:43 1:45 1:50 1:55 2:00 2:05 2:10 2:15 2:20 2:25 2:30 2:35 2:40 2:43 stop&cool  2  Reaction vessel temperature (°C) 20 54 77 95 121 146 170 195 220 245 270 295 319 329 347 326 351 374 407 422 426 430 430 430 430 430 430 430 430 430 430 430 430 430 430 17  Furnace temperature (°C) 20 67 97 115 145 168 186 207 229 252 274 297 322 367 413 439 456 475 488 472 472 471 470 467 467 466 466 466 466 466 465 465 465 465 465  Pressure (Psig) 500 530 550 580 620 680 700 750 810 890 960 1040 1120 1380 1400 1480 1580 1700 1880 1980 2020 2050 2080 2100 2100 2100 2100 2100 2100 2100 2120 2120 2120 2120 2120 400  Speed of impeller (rpm) 360 361 357 356 354 354 355 355 356 357 357 355 353 351 352 354 352 352 353 351 353 353 353 353 352 354 352 351 352 350 352 354 353 353 -  2. GC Analysis Gas inlet Pressure (psig) 500  Volume (ml) 200  Gas Outlet Component CR, Standard 2.41x10"" moles/area Area/lml 522436 gas injection 537245 474224 Average 511301.7 Moles/lml 1.23xl0" gas injection Mole 0.356411 fraction Weight (g) 1.304268 5  Temperature (K) 293  n (moles) 0.282967  Weight (g) 1.018683  CR, CjRs 1.62x10" 1.25x10""  co 1.59x10""  HS 1.95x10"" 2  Air 2.93x10"  1593 1799 1558 1650 2.67xl0"  185028 191085 164857 180323.3 2.25xl0"  51917 58202 50709 53609.33 8.55x10"  87521 89369 78681 85190.33 1.66xl0"  247155 241969 236309 241811 7.09x10""  1.74x10"' 4.16x10"'  0.000773  0.065097  0.024765  0.048229  -  0.504725 1.000000  0.004953  0.446659  0.249222  0.375046  -  0.230877 2.611026  2  8  6  2  7  6  H -  2  Total  -  3. Mass balance Mass before reaction Residue 50 g Weight of residue and microemulsion of Co in hexane 83.3 g Weight of H2 and H2S in the reactor before reaction 1.0 g. Mass after reaction Weight of coke 10.6 g Weight of Uquid hydrocarbon 63.3 g Weight of gas mixture 2.6 g. % Liquid yield - (63.3/83.3) x 100 = 76.0 % Recovery = (76.5/84.3) x 100 = 90.7  174  Experiment 19 Hydrocracking of residue using Co/PE4LE/DHN microemulsion. 1. Hydrocracking condition Initial pressure of H and H S in the reactor 500 psig. Reaction temperature 430°C for 1 h. Rate of heating 5°C/min. 2  Time (h:min) 0:00 0:05 0:10 0:15 0:20 0:25 0:30 0:35 0:40 0:45 0:50 0:55 1:00 1:05 1:10 1:15 1:20 1:23 1:25 1:30 1:35 1:40 1:43 1:45 1:50 1:55 2:00 2:05 2:10 2:15 2:20 2:23 stop&cool  2  Reaction vessel temperature (°C) 15 42 76 91 116 140 166 190 224 241 266 291 316 341 365 390 415 430 435 430 430 430 430 430 430 430 430 430 430 430 430 430 10  Furnace temperature (°C) 15 69 107 123 149 172 196 214 244 258 279 300 321 343 365 388 411 424 424 411 410 410 410 410 410 410 411 410 410 410 410 410 -  Pressure (psig) 500 560 580 600 620 650 680 700 730 780 800 820 880 900 950 1000 1030 1090 1120 1150 1180 1200 1220 1250 1280 1300 1300 1320 1330 1350 1350 1350 450  Speed of impeller (rpm)  -  360 356 353 354 360 359 359 356 356 359 361 361 359 359 357 356 357 360 357 366 361 361 361 360 361 358 361 362 361 361 361 -  2. GC Analysis Gas inlet Pressure (psig) 500  Volume (ml) 198  Temperature (K) 288  Gas Outlet Component Standard moles/area Area/lml gas injection  Average Moles/lml gas injection Mole fraction Weight (g)  co  Weight (g) 1.026005  n (moles) 0.285001  CH, 2.41x10"  C H, 1.62x10'"  C Hs 1.25x10"  1.59x10""  HS 1.95x10""  Air 2.93x10"  315260 311980 305380 308929 301322 308574.2 7.42xl0"  2140 2045 2014 2132 2136 2093.4 3.39x10  108842 108923 105273 107092 102874 106600.8 1.33xl0"  83817 65063 68093 71633 75082 72737.6 1.16xlO"  71294 70853 67771 69529 66852 69259.8 1.35xl0'  207568 199136 198774 214653 191861 202398.4 5.94x10"  2.44xl0"  4.16xl0"  0.208125  0.00095  0.037236  0.032512  0.03794  -  0.683238  1.000000  0.86924  0.00694  0.291592  0.373419  0.336718  -  0.356695  2.234604  2  6  s  2  6  2  2  7  6  6  H  Total  2  -  . -  5  5  3. Mass balance Mass before reaction Residue 50 g Weight of residue and microemulsion of Co in decalin 96.4 g Weight of H and H S in the reactor before reaction 1.0 g. Mass after reaction Weight of coke 3.0 g Weight of liquid hydrocarbon 90.0 g Weight of gas mixture 2.2 g. % Liquid yield = (90.0/96.4) x 100 = 93.4 % Recovery = (95.2/97.4) x 100 = 97.7 2  2  176  Experiment 20 Hydrocracking of residue using cobaltnaphthenate/DHN. 1. Hydrocracking condition Initial pressure of H and H S in the reactor 500 psig. Reaction temperature 430°C for 1 h. Rate of heating 5°C/min. 2  Time (h:min) 0:00 0:05 0:10 0:15 0:20 0:25 0:30 0:35 0:40 0:45 0:50 0:55 1:00 1:05 1:10 1:15 1:20 1:23 1:25 1:30 1:35 1:40 1:45 1:50 1:55 2:00 2:05 2:10 2:15 2:20 2:23 stop&cool  2  Reaction vessel temperature (°C) 12 41 71 88 109 137 170 186 216 237 262 287 312 338 362 387 412 430 437 430 431 430 430 430 430 430 430 430 430 430 430 12  Furnace temperature (°C) 12 64 101 123 149 176 208 222 245 261 282 302 324 350 370 392 415 434 435 413 417 413 413 413 413 413 413 414 414 414 415 -  Pressure (psig) 500 540 580 600 630 650 700 700 740 750 790 800 830 880 900 950 1000 1030 1050 1050 1080 1100 1120 1150 1170 1190 1200 1210 1220 1230 1230 450  Speed of impeller (rpm) -  -  -  -  2. Mass balance Mass before reaction Residue 50 g Weight of residue and microemulsion of cobaltnaphthenate in decalin 93.4 g Weight of H and H S in the reactor before reaction 1.0 g. Mass after reaction Weight of coke 2.5 g Weight of liquid hydrocarbon 86.4 g Weight of gas mixture - g. % Liquid yield = (86.4/93.4) x 100 = 92.5 % Recovery = 2  2  178  Experiment 21 Hydrocracking of residue using Fe/PE4LE/DHN microemulsion. 1. Hydrocracking condition Initial pressure of H and H S in the reactor 500 psig. Reaction temperature 430°C for 1 h. Rate of heating 5°C/min. 2  Time (h:min) 0:00 0:05 0:10 0:15 0:20 0:25 0:30 0:35 0:40 0:45 0:50 0:55 1:00 1:05 1:10 1:15 1:20 1:23 1:25 1:30 1:35 1:40 1:45 1:50 1:55 2:00 2:05 2:10 2:15 2:20 2:23 stop&cool  2  Reaction vessel temperature (°C) 14 40 81 93 117 139 165 189 214 239 264 290 314 339 364 389 414 430 435 430 430 430 430 430 430 430 430 430 430 430 10  Furnace temperature (°C) 67 111 120 144 165 189 208 230 253 274 295 315 337 358 381 403 418 419 405 403 404 404 404 404 404 405 405 405 405 405 -  Pressure (psig) 500 550 580 600 620 640 680 700 730 780 800 820 860 900 950 1010 1100 1180 1200 1240 1280 1300 1330 1370 1400 1400 1420 1440 1450 1460 450  Speed of impeller (rpm) 352 350 351 349 351 349 348 349 347 349 348 349 349 349 347 350 352 353 351 352 351 353 353 353 352 351 353 353 -  179  2. GC Analysis Gas inlet Pressure (psig) 500  Volume (ml) 198  Temperature (K) 287  n (moles) 0.285994  Weight (g) 1.029579  Gas Outlet Component Standard moles/area Area/lml gas injection  Average Moles/lml gas injection Mole fraction Weight (g)  Total  CK, 2.41x10"  C2H4  CH<5  1.62x10"  1.25x10"  C0 1.59x10"  HS 1.95x10""  Air 2.93x10""  H -  305998 325336 321740 329265 320584.8 7.71xl0"  2063 2215 2269 2316 2215.75 3.58x10""  99921 102789 108107 103900 103679.3 1.29xl0"  65857 68649 77558 79389 72863.25 1.16xl0"  62329 61277 62820 63375 62450.25 1.22xl0"  288300 254704 259071 267358.3 7.84xl0"  -  2.23xl0"  4.16xl0"  0.228428  0.001062  0.038259  0.034406  0.03614  -  0.661705  1.000000  0.954037  0.00776  0.299605  0.395174  0.320746  -  0.345453  2.322775  6  2  2  6  6  2  6  6  2  5  5  3. Mass balance Mass before reaction Residue 50 g Weight of residue and microemulsion of Fe in decalin 96.3 g Weight of H and H S in the reactor before reaction 1.0 g. Mass after reaction Weight of coke 3.5 g Weight of liquid hydrocarbon 89.3 g Weight of gas mixture 2.3 g. 2  2  % Liquid yield = (89.3/96.3) x 100 % Recovery = (95.1/97.3) x 100  = 92.7 = 97.7  180  Experiment 22 Hydrocracking of residue using DHN. 1. Hydrocracking condition Initial pressure of H and H S in the reactor 500 psig. Reaction temperature 430°C for 1 h. Rate of heating 5°C/min. 2  Time (h:min) 0:00 0:05 0:10 0:15 0:20 0:25 0:30 0:35 0:40 0:45 0:50 0:55 1:00 1:05 1:10 1:15 1:20 1:23 1:30 1:35 1:40 1:45 1:50 1:55 2:00 2:05 2:10 2:15 2:20 2:23 stop&cool  2  Reaction vessel temperature (°C) 13 42 76 90 103 155 161 189 213 239 264 289 314 338 364 389 414 430 430 430 430 430 430 430 430 430 430 430 430 430 9  Furnace temperature (°C) 13 66 104 115 160 187 181 205 227 251 273 295 317 339 362 385 407 423 412 409 408 409 409 409 409 409 409 409 409 409  Pressure (psig) 500 550 590 600 630 680 680 700 710 740 780 800 830 890 910 950 1010 1060 1100 1100 1140 1140 1200 1200 1220 1220 1250 1250 1280 1280  Speed of impeller (rpm) 350 348 350 352 347 349 349 349 349 350 349 349 349 348 349 350 350 349 350 350 351 351 352 352 351 350 350 350  500  181  2. GC Analysis Gas inlet Pressure (psig) 500  Volume (ml) 202  Temperature (K) 286  Gas Outlet Component Standard moles/area Area/lml gas injection  Average Moles/lml gas injection Mole fraction Weight (g)  co  n (moles) 0.292792  Weight (g) 1.054052  CH, 2.41x10"  1.62x10-"  1.25x10"  1.59x10"  HS 1.95x10"  Air 2.93x10"  184824 244823 215981 227643 186977 234855 215850.5 5.19xl0"  699 933 964 811 742 852 833.5 1.35X10"  48729 68458 54071 62451 52392 64813 58485.67 7.29x10"'  14380 19556 21405 16261 13735 22684 18003.5 2.87x10"'  46456 60268 52645 58803 53052 60910 55355.67 1.08X10'  238771 268754 257570 121559 440098 225173 258654.2 7.59xl0"  2.67xl0"  4.16xl0"  0.152647  0.000396  0.02142  0.008438  0.031794  -  0.785305  1.000000  0.725245  0.003296  0.190816  0.110242  0.320995  -  0.466385  1.816979  6  8  2  2  6  6  H -  Total  2  5  5  3. Mass balance Mass before reaction Residue 50 g. Weight of residue and decalin 94.0 g. Weight of H and H S in the reactor before reaction 1. lg. Mass after reaction Weight of coke 2.9 g Weight of liquid hydrocarbon 87.4 g. Weight of gas mixture 1.8 g. % Liquid yield = (87.4/94.0) x 100 = 93.0 % Recovery = (92.1/95.1) x 100 = 96.8 2  2  182  Experiment 23 Hydrocracking of residue using Co/PE4LE/DHN microemulsion. 1. Hydrocracking condition Initial pressure of H and H S in the reactor 500 psig. Reaction temperature 430°C for 1 h. Rate of heating 5°C/min. 2  Time (h:min) 0:00 0:05 0:10 0:15 0:20 0:25 0:30 0:35 0:40 0:45 0:50 0:55 1:00 1:05 1:10 1:15 1:20 1:23 1:30 1:35 1:40 1:45 1:50 1:55 2:00 2:05 2:10 2:15 2:20 2:23 stop&cool  2  Reaction vessel temperature (°C) 16 46 70 91 115 146 164 191 216 241 266 291 316 341 366 391 416 430 430 430 430 430 430 430 430 430 430 430 430 430 13  Furnace temperature (°C) 16 63 95 120 161 178 190 213 236 257 278 296 319 340 363 386 409 423 410 409 409 409 410 410 410 410 410 410 410 410  Pressure (psig) 500 530 550 590 610 640 660 690 700 720 780 800 830 880 900 930 990 1030 1090 1100 1130 1170 1190 1200 1220 1250 1260 1290 1300 1300 420  Speed of impeller (rpm) 347 345 345 343 347 347 347 348 348 352 351 355 354 352 356 356 357 366 361 362 363 363 365 363 363 362 362 362 362  2. GC Analysis Gas inlet Pressure (psig) 500  Gas Outlet Component Standard moles/area Area/lml gas injection Average Moles/lml gas injection Mole fraction Weight (g)  Volume (ml) 199  ca,  c a,  Temperature (K) 289  co  Weight (g) 1.027618  n (moles) 0.28545  2.41x10""  1.62x10"  CA 1.25x10""  1.59x10"  HS 1.95x10"  Air 2.93x10""  231089 258025 265220 251444.7 6.05x10"*  167019 2314 68173 79168.7 1.28x10""  187730 85937 158290 143985.7 1.79x10"*  98984 54806 145864 99884.7 1.59x10"*  171143 49637 74289 98356.3 1.92x10"*  139943 217150 172807 176633.3 5.18x10"*  2.38xl0"  4.16xl0"  0.166074  0.035165  0.049251  0.04372  0.05276  -  0.65303  1.000000  0.643817  0.238565  0.357992  0.466096  0.434637  -  0.316449  2.457557  2  2  2  H -  Total  2  -  5  5  3. Mass balance Mass before reaction Residue 50 g. Weight of residue and microemulsion of Co in decalin 95.8 g. Weight of H and H S in the reactor before reaction 1.0 g. Mass after reaction Weight of coke 3.0 g. Weight of liquid hydrocarbon 88.3 g. Weight of gas mixture 2.5 g. % Liquid yield = (88.3/95.8) x 100 = 92.2 % Recovery = (93.8/96.8) x 100 = 96.9 2  2  184  Experiment 24 Hydrocracking of residue using Fe/PE4LE/DHN microemulsion. 1. Hydrocracking condition Initial pressure of H and H S in the reactor 500 psig. Reaction temperature 430°C for 1 h. Rate of heating 5°C/min. 2  2  Time (h:min)  Reaction vessel temperature (°C)  Furnace temperature (°C)  0:00 0:05 0:10 0:15 0:20 0:25 0:30 0:35 0:40 0:45 0:50 0:55 1:00 1:05 1:10 1:15 1:20 1:22 1:30 1:35 1:40 1:45 1:50 1:55 2:00 2:05 2:10 2:15 2:20 2:23  IS 48 74 92 119 144 169 194 218 243 268 294 318 343 368 393 418 430 430 430 430 430 430 430 430 430 430 430 430 430  17 66 100 123 152 177 196 218 240 261 281 303 323 345 367 390 413 424 411 409 410 410 410 410 410 410 410 410 410 410  Pressure Opsig) 500 530 550 590 600 630 670 690 710 740 780 810 850 900 950 1000 1100 1180 1260 1290 1310 1350 1380 1400 1400 1420 1440 1450 1480 1480  stop&cool 9  500  Speed of impeller (rpm) 353 359 358 360 360 361 359 360 360 360 360 361 360 360 360 360 360 363 363 362 362 363 363 362 362 363 362 363 363 363  2. GC Analysis Gas inlet Pressure (psig) 500  Volume (ml) 200  Gas Outlet Component Standard moles/area Area/lml gas injection Average Moles/lml gas injection Mole fraction Weight (g)  n (moles) 0.284912  Temperature (K) 291  co  Weight (g) 1.025684  Total  CH, 2.41x10""  C H, 1.62x10-"  C Hs 1.25x10-"  1.59x10"  HS 1.95x10"  Air 2.93x10-"  H -  273618 341409 442149 286869 336011.3 8.08x10"*  1854 2435 2927 2197 2353.25 3.81xl0-  84687 100160 129890 99699 103609 1.29x10*  69226 66129 114267 57646 76817 1.22x10"*  49745 59188 77007 48358 58574.5 1.14x10-*  209803 283306 351468 224647 267306 7.84x10"*  -  2.2xl0"  4.16xl0"  0.239409  0.001128  0.038231  0.036272  0.033895  -  0.651065  1.000000  1.126201  0.009282  0.337205  0.46922  0.338825  -  0.382833  2.663567  2  8  2  2  2  2  -  5  5  3. Mass balance Mass before reaction Residue 50 g Weight of residue and microemulsion of Fe in decalin 95.8 g Weight of H and H S in the reactor before reaction 1.0 g. Mass after reaction Weight of coke 3.5 g Weight of liquid hydrocarbon 88.8 g Weight of gas rruxture 2.7 g. % Liquid yield = (88.8/95.8) x 100 = 92.7 % Recovery = (95/96.8) x 100 = 98.1 2  2  186  Experiment 25  Hydrocracking of residue using cobaltnaphthenate/DHN.  1. Hydrocracking condition Initial pressure of H and H S in the reactor 500 psig. Reaction temperature 430°C for 1 h. Rate of heating 5°C/min. 2  Time (h:min) 0:00 0:05 0:10 0:15 0:20 0:25 0:30 0:35 0:40 0:45 0:50 0:55 1:00 1:05 1:10 1:15 1:20 1:23 1:30 1:35 1:40 1:45 1:50 1:55 2:00 2:05 2:10 2:15 2:20 2:23 stop&cool  2  Reaction vessel temperature (°C) 11 38 75 86 112 136 163 186 212 237 262 287 312 336 362 387 412 430 430 430 430 430 430 430 430 430 430 430 430 430  Furnace temperature (°C) 11 64 108 117 144 168 192 210 234 255 275 296 315 338 361 383 406 423 410 409 408 409 409 409 410 410 410 410 410 410  Pressure (psig) 500 540 580 590 610 650 680 700 720 740 780 820 840 880 910 960 1000 1020 1020 1020 1020 1030 1050 1090 1100 1110 1120 1130 1150 1150  14  410  400  Speed of impeller (rpm) 347 348 347 345 343 344 345 345 343 344 344 343 344 346 346 348 347 351 350 349 350 349 348 348 348 348 348 348  2. GC Analysis Gas inlet Pressure (psig) 500  Volume (ml) 202  Gas Outlet Component Standard moles/area Area/lml gas injection Average Moles/lml gas injection Mole fraction Weight (g)  Weight (g) 1.061475  n (moles) 0.294854  Temperature (K) 284  CK, 2.41x10"  1.62x10"  CH4  C H« 1.25x10"  1.59x10""  HS 1.95x10"  Air 2.93x10"  315802 181630 366490 287974.1 6.93x10"*  1828 1530 2150 1836.033 2.97x10  93638 64152 109970 89253.45 1.11x10"*  25319 10109 34915 23447.67 3.74xl0"  101203 67993 120026 96407.13 1.88x10*  209496 310025 276143 265221.5 7.78x10"*  2.35xl0"  4.16xl0"  0.204812  0.000878  0.032874  0.011052  0.055687  -  0.694697  1.000000  0.764906  0.005739  0.230204  0.113504  0.441947  -  0.324309  1.880610  2  2  s  co  2  2  7  H -  Total  2  -  5  5  3. Mass balance Mass before reaction Residue 50 g. Weight of cobaltnaphthenate 0.2716 g. Weight of residue and cobaltnaphthenate in decalin 93.6 g. Weight of H and H S in the reactor before reaction 1.1 g. Mass after reaction Weight of coke 2.5 g Weight of liquid hydrocarbon 88.8 g. Weight of gas mixture 1.9 g. % Liquid yield = (88.8/93.6) x 100 = 94.9 % Recovery = (93.2/94.7) x 100 = 98.4 2  2  188  Experiment 26 Hydrocracking of residue using DHN 1. Hydrocracking condition Initial pressure of H and H S in the reactor 500 psig. Reaction temperature 430°C for 1 h. Rate of heating 5°C/min. 2  Time (h:min) 0:00 0:05 0:10 0:15 0:20 0:25 0:30 0:35 0:40 0:45 0:50 0:55 1:00 1:05 1:10 1:15 1:20 1:22 1:30 1:35 1:40 1:45 1:50 1:55 2:00 2:05 2:10 2:15 2:20 2:22 stop&cool  2  Reaction vessel temperature (°C) 21 49 80 95 121 146 171 195 221 246 271 296 321 346 371 396 421 430 430 430 430 430 430 430 430 430 430 430 430 430 17  Furnace temperature (°C) 20 70 108 122 147 171 191 212 235 257 279 300 322 343 364 385 409 417 404 403 403 403 403 403 403 403 403 403 403 403  Pressure (psig)  500 550 580 600 620 650 680 700 710 740 780 800 830 880 910 970 1020 1040 1080 1100 1120 1150 1180 1200 1210 1220 1250 1270 1280 1280 430  Speed of impeller (rpm) 345 345 341 341 342 340 341 340 339 346 348 347 347 347 347 349 351 353 352 353 352 351 353 355 354 354 352 352 352  2. GC Analysis Gas inlet Pressure (psig) 500  Volume (ml) 202  n (moles) 0.284825  Temperature (K) 294  Weight (g) 1.02537  Gas Outlet Component Standard moles/area Area/lml gas injection Average Moles/lml gas injection Mole fraction Weight (g)  CK, 2.41x10""  1.62x10"  1.25x10""  C0 1.59x10""  HS 1.95x10"  Air 2.93x10"  140467 74527 83351 99448.3 2.39xl0"  1186 795 872 951 1.54x10  74341 61329 66526 67398.67 8.4x10"'  1875 1898 1924 1899 3.03x10'"  78462 58543 63111 98356.3 1.3xl0"  223631 1194994 1353756 924127 2.71xl0"  9.92xl0"  4.16xl0"  0.165007  0.001061  0.057915  0.002088  0.08989  -  0.684038  1.000000  0.655615  0.007378  0.431456  0.022816  0.758957  -  0.339732  2.215954  C2H4  6  2  s  2  6  5  H -  Total  2  -  6  5  3. Mass balance Mass before reaction Residue 50 g. Weight of residue and decalin 92.8 g. Weight of H and H S in the reactor before reaction 1.0 g. Mass after reaction Weight of coke 3.0 g. Weight of liquid hydrocarbon 87.5 g. Weight of gas mixture 2.2 g. % Liquid yield = (87.5/92.8) x 100 = 94.3 % Recovery = (92.7/93.8) x 100 = 98.8 2  2  190  Experiment 27 Hydrocracking of residue using NiMo/PE4LE/DHN microemulsion. 1. Hydrocracking condition Initial pressure of H and H S in the reactor 500 psig. Reaction temperature 430°C for 1 h. Rate of heating 5°C/min. 2  Time (h:min) 0:00 0:05 0:10 0:15 0:20 0:25 0:30 0:35 0:40 0:45 0:50 0:55 1:00 1:05 1:10 1:15 1:20 1:23 1:30 1:35 1:40 1:45 1:50 1:55 2:00 2:05 2:10 2:15 2:20 2:23 stop&cool  2  Reaction vessel temperature (°C) 16 45 71 90 117 140 166 191 216 241 266 291 316 341 366 391 416 430 430 430 430 430 430 430 430 430 430 430 430 430 14  Furnace temperature (°C) 16 63 97 118 149 169 192 211 232 253 274 295 313 334 357 379 401 414 401 400 400 401 402 402 402 402 402 402 402 402  Pressure (psig) 500 540 570 600 620 650 680 700 720 750 780 820 850 880 920 950 1000 1030 1080 1080 1100 1100 1130 1150 1180 1190 1200 1200 1220 1220  Speed of impeller (rpm)  311 351 345 344 341 339 339 339 337 338 338 339 338 337 336 337 338 337 337 337 341 359 358 357 356 358 359 359 359  400  191  2. GC Analysis Gas inlet Pressure (psig) 500  Volume (ml) 201  Temperature (K) 289  Gas Outlet Component Standard moles/area Area/lml gas injection Average Moles/1 ml gas injection Mole fraction Weight (g)  co  CH, 2.41x10""  C H, 1.62x10""  C Hs 1.25x10"  1.59x10"  231984 280057 252531 202092 241666 5.81X10"  2208 2116 2122 2006 2113 3.42x10"*  179199 228777 101196 82237 147852.5 1.84xl0"  89791 104586 105197 50801 87593.75 1.4xl0"  0.160942  0.000946  0.050994  0.598091  0.006154  0.355316  2  6  2  6  n (moles) 0.288318  Weight (g) 1.037946  H -  Total  HS 1.95x10"  Air 2.93x10"  101207 218982 64450 52089 109182 2.13X10"  177943 208662 175722 185144 186867.8 5.48x10"  2.49xl0"  4.16xl0  0.038659  0.059054  -  0.689405  1.000000  0.395077  0.466345  -  0.320245  2.141229  2  2  6  6  6  2  -  5  -5  3. Mass balance Mass before reaction Residue 50 g Weight of residue and microemulsion of Ni/Mo in decalin 94.0 g Weight of H and H S in the reactor before reaction 1.0 g. Mass after reaction Weight of coke 2.6 g Weight of liquid hydrocarbon 87.3 g Weight of gas mixture 2.1 g. % Liquid yield = (87.3/94.0) x 100 = 92.9 % Recovery = (92.0/95.0) x 100 = 96.8 2  2  192  Experiment 28 Hydrocracking of residue using CoMo/PE4LE/DHN microemulsion. 1. Hydrocracking condition Initial pressure of H and H S in the reactor 500 psig. Reaction temperature 430°C for 1 h. Rate of heating 5°C/min. 2  Time (h:min) 0:00 0:05 0:10 0:15 0:20 0:25 0:30 0:35 0:40 0:45 0:50 0:55 1:00 1:05 1:10 1:15 1:20 1:23 1:30 1:35 1:40 1:45 1:50 1:55 2:00 2:05 2:10 2:15 2:20 2:23 stop&cool  2  Reaction vessel temperature (°C) 12 37 79 87 114 137 163 186 213 237 262 287 312 337 362 387 412 430 430 430 430 430 430 430 430 430 430 430 430 430 13  Furnace temperature (°C) 12 64 113 117 144 166 191 208 233 255 275 296 316 337 359 381 402 419 407 405 405 405 407 407 406 407 407 407 407 407  Pressure (psig) 500 550 600 600 630 670 680 700 720 750 780 800 830 880 900 940 1000 1030 1080 1080 1080 1100 1110 1130 1150 1180 1190 1200 1200 1200  Speed of impeller (rpm) 343 340 346 345 345 344 350 353 347 349 349 349 350 350 349 349 349 350 351 351 350 351 350 351 353 355 351 351  430  193  2. GC Analysis Gas inlet Pressure (psig) 500  Volume (ml) 200  Gas Outlet Component Standard moles/area Area/lml gas injection Average Moles/1 ml gas injection Mole fraction Weight (g)  Temperature (K) 285  CH6  co  n (moles) 0.29091  HS 1.95x10"  Air 2.93x10"  H -  60616 42247 138547 256571 124495.3  107044 119433 118681 172813 129492.8 3.8xl0"  -  CK, 2.41x10"  C H, 1.62x10"  1.25x10"  1.59xl0  188094 173670 187448 241534 197686.5 4.75xl0-  1978 1892 1901 1805 1894 3.06x10*  104584 65283 140988 157052 116976.8 1.46xl0"  63679 40279 130045 100470 83618.3 1.33xl0"  2.43X10"  0.125793  0.000811  0.038549  0.035262  0.064339  0.501778  0.005658  0.288316  0.386806  0.545372  2  6  2  6  2  2  -11  6  Weight (g) 1.047277  6  6  -  Total  2  -  2.78xl0~  4.16xl0"  0.735247  1.000000  0.366606  2.094536  5  5  3. Mass balance Mass before reaction Residue 50 g. Weight of residue and microemulsion of Co/Mo in decalin 94.7 g. Weight of H and H S in the reactor before reaction 1.0 g. Mass after reaction Weight of coke 2.6 g. Weight of liquid hydrocarbon 88.0 g. Weight of gas mixture 2.1 g. % Liquid yield = (88.0/94.7) x 100 = 92.9 % Recovery = (92.7/95.7) x 100 = 96.9 2  2  194  Experiment 29 Hydrocracking of residue using DHN. 1. Hydrocracking condition Initial pressure of H and H S in the reactor 500 psig. Reaction temperature 430°C for 1 h. Rate of heating 5°C/min. 2  Time (h:min) 0:00 0:05 0:10 0:15 0:20 0:25 0:30 0:35 0:40 0:45 0:50 0:55 1:00 1:05 1:10 1:15 1:20 1:23 1:30 1:35 1:40 1:45 1:50 1:55 2:00 2:05 2:10 2:15 2:20 2:22 stop&cool  2  Reaction vessel temperature (°C) 16 46 76 89 118 141 167 190 217 241 266 291 316 341 366 391 416 430 430 430 430 430 430 430 430 430 430 430 430 430 10  Furnace temperature (°C) 16 68 106 119 149 171 194 212 236 257 279 300 320 341 363 384 406 405 404 405 404 405 405 405 405 405 405 405 405 405  Pressure (psig) 500 550 580 590 610 630 660 680 700 720 750 780 810 840 890 920 980 1020 1050 1080 1100 1120 1150 1180 1180 1200 1210 1220 1230 1230 400  Speed of impeller (rpm)  334 334 331 341 338 339 339 339 339 338 339 339 339 337 338 338 342 342 343 344 346 345 342 342 342 342 341 342 342  2. GC Analysis Gas inlet Pressure (psig) 500  Gas Outlet Component Standard moles/area Area/lml gas injection Average Moles/lml gas injection Mole fraction Weight (g)  Volume (ml) 202  CH4  Temperature (K) 289  c a3  co  n (moles) 0.289753  Weight (g) 1.04311  CH, 2.41x10"  1.62x10"  1.25x10"  1.59x10"  HS 1.95x10"  Air 2.93x10"  H -  223051 256210 269209 261552 252505.5 6.07x10""  1095 1159 1093 1361 1177 1.9xl0-  61814 66085 70826 71540 67566.3 8.42x10"'  26406 21221 37226 16285 25284.5 4.03x10"'  60283 63268 68987 69040 65394.5 1.28x10""  147163 153244 191334 173794 166383.8 4.88x10""  -  2.81x10"'  4.16x10"'  0.165409  0.000519  0.022922  0.010977  0.034792  -  0.765382  1.000000  0.626483  0.003437  0.162781  0.114327  0.280017  -  0.362357  1.549402  2  8  2  2  2  2  Total  3. Mass balance Mass before reaction Residue 50 g. Weight of residue and decalin 93.0 g. Weight of H and H S in the reactor before reaction 1.0 g. 2  2  Mass after reaction Weight of coke 2.9 g Weight of liquid hydrocarbon 87.6 g Weight of gas mixture 1.5 g % Liquid yield = (87.6/93.0) x 100 = 94.2 % Recovery = (92.0/94.0) x 100 = 97.9  196  Experiment 30 Hydrocracking of residue using Co/PE4LE/n-hexane microemulsion. 1. Hydrocracking condition Initial pressure of H and H S in the reactor 500 psig. Reaction temperature 430°C for 1 h. Rate of heating 5°C/min. 2  Time (h:min) 0:00 0:05 0:10 0:15 0:20 0:25 0:30 0:35 0:40 0:45 0:50 0:55 1:00 1:05 1:10 1:15 1:20 1:25 1:30 1:35 1:40 1:45 1:50 1:55 2:00 2:05 2:10 2:15 2:20 2:25 2:30 2:35 2:40 2:45 2:50 stop&cool  2  Reaction vessel temperature (°C) 22 56 75 98 125 160 176 202 222 248 272 297 320 347 371 377 393 412 422 421 425 427 430 430 430 429 430 430 430 429 430 430 430 430 430 15  Furnace temperature (°C) 20 65 94 119 170 187 189 215 234 258 280 305 334 373 406 434 453 469 473 481 486 482 481 478 477 477 476 476 476 475 476 475 475 475 475  Pressure (psig) 500 530 550 580 650 680 700 730 780 830 900 1000 1180 1300 1410 1450 1550 1650 1720 1820 2000 2020 2100 2150 2180 2200 2220 2250 2270 2280 2280 2300 2300 2300 2310  Speed of impeller (rpm) 351 357 353 355 354 355 356 353 354 354 354 353 350 350 349 350 351 351 352 353 352 351 355 357 359 356 359 358 358 357 358 358 360 359 359  340  197  2. GC Analysis Gas inlet Pressure (psig) 500  Volume (ml) 200  Temperature (K) 295  Weight (g) 1.011776  n (moles) 0.281049  Gas Outlet Component Standard moles/area Area/lml gas injection Average Moles/lml gas injection Mole fraction Weight (g)  CjH* 1.62x10'"  C H« 1.25x10""  C0 1.59x10"  HS 1.95x10""  Air 2.93x10""  384207 606310 554876 181397 431697.5 1.04xl0"  1690 1578 1657 1698 1655.75 2.68x10""  152746 304100 283219 70578 202660.8 2.52xl0"  33424 92509 102571 15623 61031.75 9.73x10"  76153 159683 252433 36884 131288.3 2.56xl0"  196146 248935 296945 106865 212222.8 6.22x10"  1.89X10"  4.16xl0"  0.29354  0.000757  0.071366  0.027502  0.072504  -  0.534331  1.000000  0.919407  0.00415  0.419114  0.236887  0.482569  -  0.2092  2.271326  5  2  6  2  1  2  6  6  H -  Total  CH, 2.41x10"  2  -  5  5  3. Mass balance Mass before reaction Residue 50 g. Weight of residue and decalin 83.0 g. Weight of H and H S in the reactor before reaction 1.0 g. 2  2  Mass after reaction Weight of coke 10.2 g. Weight of Uquid hydrocarbon 65.8 g. Weight of gas mixture 2.3 g. % Liquid yield = (65.8/83.0) x 100 = 79.3 % Recovery = (78.3/84.0) x 100 = 93.2  198  TJ" T T  «o ^ ON  '  »<o ro  rf  t  *  •  cocococococococov-ito»n  r- c* vo OOOOOvTtr-rHOr-Ov O C O o vdoo"t-*Tf"r-"odvdr-"vd Ov  ON ON  co rf  *nro co «o ro  Tf ' C O C O C O C O C O C O C O C O C O C O C O CO C O C O CO C O CO TfTfTfTfTfTfTfTfTfrfTf Tf rf rf  T T vo T f h  t  h  . ..  ' oo' vo r- c-* «o co  1  O N O N O N O N O N O N O N O N O N  O N O N O N  ON  ON  Tf Tf  ro ro •  rf rf rf  ON  •  Tf Tf  o <ot<oo\MHONMn«o  Tj-  CM  a  < o o « o «n C N "o 00 oo o\ ' O N VO Tf" C * O N C O Tf vo vd Tf « n v-» vo vo vo c~ vo  coodvovdvdvSoodvSco vo r ^ v o c ~ v o v o v o t ^ v o v o C "  co  « o *n » o  oo oo oo n >-H o <o n N O r f r-uoo co • vo co oo ci C O Tf* H r O C O O N C O C O C O C O C O C O « * O T f ' vS co T f (  H  «n  o vq vq vo  \ D  ci cirt  8  ON  16.4 5.4  "O  ON  od o tS*n vo  62.8 68.9 73.5 50.6 69.2  |  (%)  Recovery  TT  91.5 • | 96.8 96.9 98.8 93.8  rt - co co ro a 5 a co co ro co ro  •8  Tf  u-i vd  rt  O VO  ) t  _,oc<  00  H M Mo  o vo o  Tf °v *o Tf o ro ro O  | 1 rt  *o TJ-"  d < o f c< ^ o o c o c n o o o o o o r ^ O N v q vd vo" K «o c i w o c i T f c o c o c o r o T f c i  ci oo v q rt ci ci ci t-^  , ro o n «o  rt Tf «0  rt  r<  ON  MS vd »o ro |  s  o  O  rt  11.3 4.3  ^  16.1 11.1 10.6  «n 00 O O C * . ^ « O T f V O f O _ C O T f ' - < d rf o r-4 ci ci ci •-" O O ^ c i c o ' r o c i c o T f , - " O  R o  O N 00 f- • O  I O  T T  •3  |  o oo o o o o o o  0.08 0.09 0.08 0.10 0.06  r- vo *n o oo o o o  »o vo 0 0 0 0 vo t H i O M H H M O N V O o oo o o o o o oo o o o o o o o o o r o o 00 VO T f ON  vS  Ct  CO  c< vq C O --" ci c i ci  ON  »o«orf  ON C O » O V N O N O O O N O O O O V O V O  ci i r C l cs H ON"  00  i v d v d c i c i c o o N O v d o d H H f H C t f i r n M ' - H  oo cn i o q n r-; oo c i wo  1  vq  ~ci^cicici«-"cici'-'  t  o co co" C * C I C *  ON  o cs  ci ci  CO 00  cCO  t>  o  «-< c* cs  i> o \ v q  . ..  Tf  ri ri  c< ov Tf C <  C O  O vd  CO ON  c~ c-  VO  O C 0 < S T f C - C - 0 0 T f < S  K cnrf'^cKHcs^coci  l >  O N O N O N O O O N O N O N O N O N  vi ON «0  r - 00  ' ci Tf ci CO od ON ON ON O N  O A O O \ r ^ ^ " O f O O O v i ^ v i o \ 0 \  00  Tf  cicocicoro'cocororocicicicici H  vo  CO  vd  vd 0 0 00  r-.  •  i  M  v d M N  i  "O  to  s  I  ci  •a  ci  a •8 o  6  I  ro vq vo  (1 vd  Fe-Hexane Co-Hexane Co-Hexane 2nd Co-Hexane 3rd Mo-Hexane Pure DHN Pure DHN 2nd Pure DHN 3rd Fe-DHN Fe-DHN 2nd Fe-DHN 3rd Co-DHN Co-DHN 2nd Co-DHN 3rd Cobaltnaph-DHN Cobaltnaph-DHN 2nd Cobaltnaph-DHN 3rd Mo-DHN Mo-DHN 2nd Fe-Toluene Co-Toluene Mo-Toluene Fe-THF Co-THF Mo-THF  r-  76.5 92.9 92.9 79.9 75.4  (%)  (g) 11.2 9.9 10.6 10.2 9.8  VO *  1  NiMo-DHN CoMo-DHN Thermal hydrocracking Thermal hydrocracking 2nd  Liquid yield Coke yield Types of microemulsions  ON  00  o  |  (g)  Gas yield  CO  30.0 25.4 25.5 22.9 16.1 12.6 21.3 20.0  tS o <n  i-l O C ^ V O O N ^ r ^ T f ^ O N f O » 0 0 ' O  v  r  t  ( >  *0  C l T f Tf C O  3  o ro r- oo rt v° rA C i  1  199  Liquid product Analysis Run no. 12 11 19 31 9 23 27 30 7 22 25 8 20 24 16 21 26 6 18 5 2 4 14 13 10 3 28 29 1 17  Types of microemulsions Fe-Hexane Co-Hexane Co-Hexane 2nd Co-Hexane 3rd Mo-Hexane Pure DHN Pure DHN 2nd Pure DHN 3rd Fe-DHN Fe-DHN 2nd Fe-DHN 3rd Co-DHN Co-DHN 2nd Co-DHN 3rd Cobaltnaph-DHN Cobaltnaph-DHN 2nd Cobaltnaph-DHN 3rd Mo-DHN Mo-DHN 2nd Fe-Toluene Co-Toluene Mo-Toluene Fe-THF Co-THF Mo-THF Cobaltnaphthenate Ni/Mo-DHN Co/Mo-DHN Thermal hydrocracking Thermal hydrocracking 2nd  s  70.3 62.0 41.5 41.2 40.0 39.2 47.6 50.1 44.3 57.4 56.7 50.1 40.2 42.9 55.7  Conversion (%) >525°C 66.0 50.1 50.4 48.0 51.2 50.4 47.7 54.4 55.0 50.4  MCR 88.1 82.9 32.7 31.7 30.7 28.5 32.8 33.6 29.1 46.5 33.9 35.1 40.8 39.9 45.0  Asphaltene 87.5 86.6 43.2 48.1 50.4 50.4 54.8 50.6 56.3 57.7 53.4 42.3 38.6 37.4 50.2 48.6 -  Coke yield Catalysts Fe-Hexane Co-Hexane Mo-Hexane Fe-DHN Co-DHN Mo-DHN Ni/Mo-DHN Co/Mo-DHN Conaph-DHN DHN Fe-Toluene Co-Toluene Mo-Toluene Fe-THF Co-THF Mo-THF Thermal hydrocracking Cobaltnaphthenate  Coke yield (%), mean ± std error 22.4 20.8 ± 0.40 19.6 7.14 ±0.14 6.20 ± 0.20 5.80 ±0.00 5.20 5.20 5.00 ± 0.00 5.86 + 0.06 14.8 14.2 12.2 42.8 42.4 43.0 15.0 ±0.0 12.6  200  Appendix U  2. Sample calculations 2.1 Weight of H 2 / H 2 S mixture in the reactor before hydrocracking reaction 2.2 Weight of gas outlet 2.3 The amount of metal in residue 2.4 The molar ratio of reducing agent to metal  201  2. Sample calculations 2.1 Weight of H2/H2S mixture in the reactor before hydrocracking reaction 2.1.1 Number of moles of H /H S in the reactor before hydrocracking reaction 2  2  We assume that the gaseous mixture obeys ideal gas law. Using data of the experiment of hydrocracking of residue with Mo in toluene microemulsion at 430°C Density of heavy oil is approximately  1.0  g/cm  Density of toluene is  0.867 g/cm  3  3  Weight of Mo-in-toluene microemulsion before reaction  = 41.0 g = 41.0/0.867  Volume of Mo-in-toluene microemulsion before reaction  = 47.3 cm  3  Note: Since toluene is a major component in microemulsion, the density of the microemulsion is approximately equal to the density of toluene. Total volume of liquid reactants  = 50 + 47 = 97 cm  Volume of H ^ S in reactor vessel  = 300 - 97 = 203 cm  Pressure  = 500 psi  3  18°C = 273 + 18 = 291 K  Temperature From ideal gas law where  3  PV/RT  n  n  = no. of g-moles in gas phase  P  = pressure, atm  V  = volume, cm  R  = gas constant  3  = 82.05 cm .atm.g-mole" .K"-1 3  T  1  = temperature, K  202  _ (500/14.7)(203) (82.05)(291)  11  = 0.28919 g-moles 2.1.2 Weight of H 2 / H 2 S in the reactor before the hydrocracking reaction The composition of  H2/H2S  Molecular mass of H , 2  Weight of  H2/H2S  2  = 34  MWH2S  mixture  2  =2  MWH2  Molecular mass of H S, 2  mixture is 95% of H and 5% of H S  = mole fraction of H x M W H 2 2  X Hmixtu +  mole fraction of H S x M W H 2 S 2  re  X n ixtu m  re  = 0.95 x 2 x 0.28919 + 0.05 x 34 x 0.28919 = 1.04107 g 2.2 Weight of gas outlet 2.2.1 No. of moles per area of known composition of reference gas mixture 1 ml of reference gas was injected in GC. The retention times and areas of the gases were recorded. The injection was repeated at least three times and the average value of the area was used in the other calculation. Knowing the retention time each peak was identified for a specific gas component and the no.of moles per area was calculated. Retention time (min) 0.430 0.585 1.028 1.905 2.430 3.398 Note:  Gas component Air* CH, COz C2H4 C2H6  HS  No. of moles/area 2.93 x lO 2.41 x 10" 1.59 xlO 1.62 x lO 1.25 x 10" -11  u  -11  -11  11  1.95 xlO" The peak of H is not able to identify from GC using He as a carrier gas. We assume the remaining gas is H * N and O2 are not separated in the GC and emerge as a single peak. 2  11  2  2  2  203  2.2.2 No. of moles of gases after the hydrocracking reaction The volume of gas outlet was assumed to be equal to that of the gas inlet. Using datafromhydrocracking of heavy oil in Mo-toluene microemulsion After cooling, P = 425 psi, T = 25°C, V = 203 cm  3  n = PV/RT _ (425/14.7)(203) (82.05)(298) Total no.of moles of gas inlet  = 0.240034 g-moles  2.2.3 No. of moles of each gas in 1 ml of GC injection Gas component  Average area  Air CH, CCb  No.of moles in 1 ml  No.of moles/area  180763.50 312325.30 14412.75 945.25 76275.50 47586.75  5.30 x 10" 2.93 x 10" 7.51 x 10" 2.41 x 10" 2.30 xlO" 1.59 xlO" C2H4 1.53 x 10" 1.62 x 10" Q2H6 9.50 x 10" 1.25 x 10" HS 9.30 xlO" 1.95 xlO' Total 1.49 x 10" Note: There is presumably no air in the gas outlet. However, while taking gas mixture from the reactor to gas sample bag and injection of gas sample to GC, some air is mixed to the gas product. Therefore, it is necessary to know the amount of air in the gas injection to GC and subtract the amount of air from the gas product. 2  6  11  6  11  7  11  8  11  7  11  7  11  5  For ideal gas at 1 aim, 20°C, and 1 cm  3  n = PV/RT = (l)(l)/(82.05)(293) = 4.16xl0"  5  No. of H gas in 1 cm of gas mixture = 4.16xl0" - 1.49xl0" = 2.67x10 3  5  5  2  204  2.2.4 Mass of gas outlet (without air) Total no.of moles of gas outlet from 2.2 = 0.24003 gmoles No.of moles of CH4 in gas outlet = molefractionof CH4 x total no.of gas = 0.20697 x 0.24003 = 0.04968 g-moles Weight of CH4  = no.of mole x MW = 0.04968 x 16 = 0.79488 g  Total mass of gas outlet  = summation of mass of all gases = 1.61470 g  Gas composition CH, CO2 C2H4 C2H6  HS H 2  2  No.of moles 7.51 x 10" 2.30 xlO 1.53 x 10" 9.50 x 10" 9.30 xlO" 2.67 xlO" 6  -7  8  7  7  5  Mole fraction  No.of moles of gas outlet  MW  Mass (g)  0.20697 0.00633 0.00042 0.02618 0.02561 0.73449  0.04968 0.00152 0.00010 0.00628 0.00615 0.17630  16 44 28 30 34 2  0.79489 0.06685 0.00283 0.18850 0.20902 0.35260  Total  1.61469  2.3. The amount of metal in residue (ppm). Concentration of Co solution =  197 g/1  Concentration of Fe solution =  154.25 g/1  Concentration of Ni solution =  275.64 g/1  2.3.1 ppm Fe in residue 2.3.1.1 Using Fe solution  205  0.14 ml of Fe solution  154.25gFe x 0.14 ml sol" 1000ml sol n  0.02160 gFe 0.02160gFe 50 g residue  ppm Fe in residue  431.9 xlO"  6  432 ppm 2.3.1.2 Using FeCl .6H 0 3  2  0.1049 g ofFeCl .6H 0 3  2  = 0.1049gFeCl .6H Ox 3  2  5 5.847 gFe 270.32gFeCl .6H O 3  2  0.02167 gFe ppm Fe in residue  0.02167 gFe 50 g residue 433.4x10^ 433 ppm  2.3.2 ppm of Co in residue 2.3.2.1 Using Co solution 0.11 ml of Co solution  197 gCo xO.llmlsor 1000 ml sol n  0.02167 gCo ppm Co in residue  0.02167 gCo 50 g residue 433.4X10"  6  433 ppm  206  2.3.2.2 Using cobaltnaphthenate (8 wt% Co) 8gCo  0.2714 g of Co-naphthenate  lOOgCo-naph  x0.2714gCo-naph  0.02171 gCo 0.02171gCo 50g heavy oil  ppm Co in residue  434.2x10"* 434 ppm  <-  2.3.3 ppm of Mo in residue 0.0620 g of MoCl  0.0620gMoCl x  5  5  95.94gMo 273.24 gMoCl  5  0.02177 gMo 0.02177 gMo 50 g residue  ppm Mo in residue  435.4x10^ =  435 ppm  <—  2.4. The molar ratio of reducing agent to metal 2.4.1 Using N2H4.XH2O as a reducing agent (55 wt% N2H4 in the solution, sp.gr 1.029, MW 32.05). 0.11mlN H4.xH2O 2  =  0.11x 1.029  0.1132 g N2H4.XH2O  =  0.1132gN H .xH Ox 2  4  2  55gN H 2  100gN H .xH O 2  0.0623 g  4  4  2  N2H4  207  lmole 0.0623gN,H x 4  2  32.05g  4  1.9424 x 10" mole 3  Fe 0.0216 g  =  0.0216 gFe x 1/55.847  =  3.8668 x 10" mole  Co 0.0217 g  =  0.0217 gCo x 1/58.933  =  3.6771 x 10" mole  Mo 0.0218 g =  0.0218 gMo x 1/95.94  =  2.2691 x 10" mole  4  4  4  Molar ratio of N H4/Fe  = 1.9424 x 10 /3.8668 x 10" =  5.0  Molar ratio of NzKL/Co  = 1.9424 x 10" /3.6771 x 10" =  5.3 <-  2  3  4  3  4  <-  Note: Mo was reduced using LiBFLi only 2.4.2 Using LiBFLt as a reducing agent ( 2 M LiBFLt solution) 1 ml LiBFLt  = 1 x 2 mole of LiBIL/lOOO ml solution  = 2.0 x 10" mole 3  0.6 ml LiBFLi = 0.6 x 2 mole of LiBHVlOOO ml solution  = 1.2 x 10" mole 3  Molar ratio of LffiKU(1.0 ml)/Fe = 2 x 10" /3.8668 x 10" 3  =5.2  <-  =  <-  4  Molar ratio of LiBILtO-O ml)/Co = 2 x 10" /3.6771 x 10" 3  4  Molar ratio of LiBH4(0.6 mlVMo = 1.2 x 10" /2.2691 x 10" = 3  4  5.4 5.3  <-  208  Appendix HI  3.1  Comparison of Fe/Ni ratio of asphaltene fraction using different solvents.  3.2  Comparison of Fe/V ratio of asphaltene fraction using different solvents.  3.3  Comparison of coke yield from H-donor solvent, organometallic compound, catalyst in reverse micelles, and mixed catalysts.  3.4  Comparison of MCR conversionsfromFe, Co, Mo, CoMo and NiMo catalysts.  3.5  Comparison of S conversionsfromFe, Co, Mo, CoMo and NiMo catalysts.  3.6  Comparison of >525°C conversionsfromFe, Co, Mo, CoMo, and NiMo catalysts.  3.7  Comparison of asphaltene conversionsfromFe, Co, Mo, CoMo and NiMo catalysts.  3.8  Comparison of coke/asphaltene ratios from Fe, Co, Mo, CoMo and NiMo catalysts.  3.9  Determination of number average, standard deviation, standard error of mean, and % standard error of mean.  3.10  The estimated cost of Co prepared in microemulsion compared to cobalt naphthenate.  3.11  Determination of sample size.  209  3.1 Data analysis: Comparison of Fe/Ni ratio of asphaltene fraction using different solvents. The results of Fe/Ni ratio from different solvents were as follows: Treatment (Solvent)  Fe/Ni, x;  Sample Size,  T; = Xx;  nj  n-Hexane (1) Decalin (2) Toluene (3) Tetrahydrofuran (4) No. of Treatment, k = 4  7.24, 7.02 7.43 7.89, 7.08 7.44  2 1 2 1 N = 2>; = 6  14.26 7.43 14.97 7.44 T=I5>i = 44.10  An analysis of variance (Freund and Smith, 1986) was performed to test at the 0. 05 level of significance, whether the differences among the sample means was significant.  1. Hypotheses  Null hypothesis, ELy. \ii = \i = (13 = 2  Alternative hypothesis, H : The pi's are not all equal. A  2. Level of significance (a) a = 0.05  3. Criterion Reject the null hypothesis if F (test statistic) > 19.16, the value of  F.5 0  0  for k - 1 =  4-1 = 3 and N - k = 6 - 4 = 2 degrees offreedom,where F is to be determined by an analysis of variance. Otherwise, accept Ho or reserve judgment.  210  4. Calculations T, = 14.26, T = 7.43, T = 14.97, T = 7.44, T = 44.10, and X i x = 324.64. 2  2  3  4  Then, substituting these totals together with n, = 2, n = 1, n = 2,114 = 1, and N = 6 into 2  3  the computing formulas for the sums of squares, we get total sum of squares, SST  = X X x - T /N = 324.64 - (l/6)(44.1) = 0.50, 2  2  2  treatment sum of squares (variation among sample means, SS(Tr)) = E(Ti /ni) - T /N 2  2  = (1/2)(14.26) + (7.43) + (1/2)(14.97) +(7.44) - (l/6)(44.10) = 0.15, 2  2  2  2  2  and error sum of squares (variation within the sample), SSE = SST - SS(Tr) = 0.50-0.15 = 0.35 Then, treatment mean square, MS(Tr) = SS(Tr)/(k-l) = 0.15/3 = 0.05, error mean square, MSE = SSE/(N-k) = 0.35/2 = 0.18, and F = MS(Tr)/MSE = 0.05/0.18 = 0.28. All these results are shown in the following analysis-of-variance (AOV) table:  Table 3.2 AOV Table for the Fe/Ni ratio of asphaltene fraction. Source of Variation  Treatment (Solvent) Error Total  Sum of Squares (SS) 0.15 0.35 0.50  Degrees of Freedom (df)  Mean Square (MS)  F  3 2 5  0.05 0.18  0.28  211  5. Decision Since F = 0.28 does not exceed 19.16, the null hypothesis is preserved; we conclude that there is no difference in the Fe/Ni ratio among the solvents at 95% level of confidence.  212  3.2 Data analysis: Comparison of Fe/V ratio of asphaltene fraction using different solvents. The results of Fe/Ni ratio from different solvents were as follows: Treatment (Solvent) n-Hexane (1) Decalin (2) Toluene (3) Tetrahydrofuran (4) No. of Treatment, k = 4  Fe/V, x; Sample Size, n; 2.13,2.69 2 2.14 1 2.45, 2.70 2 2.78 1 N = 2>i = 6  T; = Ex; 4.82 2.14 5.15 2.78  T=SZxi= 14.89  An analysis of variance (Freund and Smith, 1986) was performed to test at the 0. 05 level of significance, whether the differences among the sample means was significant.  1. Hypotheses  Null hypothesis, FL,: p.i = p.2 = p^ = m Alternative hypothesis, H : The pi's are not all equal. A  2. Level of significance (a) a = 0.05  3. Criterion Reject the null hypothesis if F (test statistic) > 19.16, the value of  F 5 0 0  for k - 1 =  4-1 = 3 and N - k = 6 - 4 = 2 degrees of freedom, where F is to be determined by an analysis of variance. Otherwise, accept FL, or reserve judgment.  213  4. Calculations Ti = 4.82, T = 2.14, T = 5.15, T = 2.78, T = 14.89, and U x 2  3  4  2  = 37.37. Then,  substituting these totals together with ni = 2, n = 1, n = 2, IL, = 1, and N = 6 into the 2  3  computing formulas for the sums of squares, we get = EEx - T /N= 37.37 - (1/6)(14.89) = 0.42, 2  total sum of squares, SST  2  2  treatment sum of squares (variation among sample means, SS(Tr)) = I (Ti /ni) - T /N 2  2  = (l/2)(4.82) + (2.14) + (1/2)(5.15) +(2.78) - (1/6)(14.89) = 0.23, 2  2  2  2  2  and error sum of squares (variation within the sample), SSE  = SST - SS(Tr) = 0.42- 0.23 = 0.19  Then, treatment mean square, MS(Tr) = SS(Tr)/(k-l) = 0.23/3 = 0.08, error mean square, MSE = SSE/(N-k) = 0.19/2 = 0.09, and F = MS(Tr)/MSE = 0.08/0.09 = 0.83. All these results are shown in the following analysis-of-variance (AOV) table:  Table 3.3 AOV Table for the Fe/V ratio of asphaltene fraction. Source of Variation Treatment (Solvent) Error Total  Sum of Squares (SS) 0.23 0.19 0.42  Degrees of Freedom (df) 3 2 5  Mean Square (MS) 0.08 0.09  F 0.83  214  5. Decision Since F = 0.83 does not exceed 19.16, the null hypothesis is preserved; we conclude that there is no difference in the Fe/V ratio among the solvents at 95% level of confidence.  215  3.3 Data analysis: Comparison of coke yield from H-donor solvent, organometallic compound, catalyst in reverse micelles, and mixed catalysts. To compare the activity of different catalysts in residue oil upgrading, the experiment of each type of catalyst was repeated three times, except the experiment of Mo catalyst was repeated two times and no repetition for CoMo and NiMo. The results of coke yield from different catalysts were as follows:  Treatment (Catalyst) Fe-DHN (1) Co-DHN (2) Cobaltnaphthenate-DHN (3) Mo-DHN (4) DHN (5) CoMo (6) NiMo (7) No. of Treatment, k = 7  Coke Yield (%), x  ;  Ti = Xx;  Sample Size, nj 3 3 3 2 3 1 1  7.4, 7.0, 7.0 6.6, 6.0, 6.0 5.0, 5.0, 5.0 5.8, 5.8 5.8, 6.0, 5.8 5.2 5.2  N =Ini=16  21.4 18.6 15.0 11.6 17.6 5.2 5.2  T=XIx;=94.6  An analysis of variance (Freund and Smith, 1986) was performed to test at the 0.05 level of significance, whether the differences among the sample means (Ui) are signhicant.  1. Hypotheses  Null hypothesis, H,,: \±\ = [l = \lz = \±4 = \is 2  Alternative hypothesis, H : The (i's are not all equal. A  2. Level of significance (a) a = 0.05  216  3. Criterion  Reject the null hypothesis if F (test statistic) > 3.37, the value of F .os for k - 1 = 70  1 = 6 and N - k = 1 6 - 7 = 9 degrees offreedom,where F is to be deteirnined by an analysis of variance. Otherwise, accept Ho or reserve judgment.  4. Calculations Tx = 21.4, T = 18.6, T = 15.0, T = 11.6, T = 17.6, T = T = 5.2, T = 94.6, and 2  XXx  2  3  4  5  6  7  = 567.96. Then, substituting these totals together with n i = 3, n = 3, n = 3, at = 2, 2  3  n = 3, n6 = 1, n = 1 and N = 16 into the computing formulas for the sums of squares, we 5  7  get total sum of squares, SST =  XXx 2  T /N  = 567.96 - (1/16)(94.6) = 8.64,  2  2  treatment sum of squares (variation among sample means, SS(Tr)) = X ( T  2  /nj)  - T /N 2  = (1/3)(21.4) + (1/3)(18.6) + (1/3)(15.0) + (1/2)(11.6) + 2  2  2  2  (1/3)(17.6) + (5.2) +(5.2) - (1/16)(94.6) = 8.26 2  2  2  2  and error sum of squares (variation within the sample), SSE  = SST - SS(Tr) = 8.64 - 8.26 = 0.38  Then, treatment mean square, MS(Tr) = SS(Tr)/(k-l) = 8.26/6 = 1.38, error mean square, MSE = SSE/(N-k) = 0.38/9 = 0.04, and F = MS(Tr)/MSE = 1.38/0.04 = 33.20. All these results are shown in the following analysis-of-variance (AOV) table:  217  Table 3.4 AOV Table for the coke yield. Source of Variation Treatments (Catalysts) Error Total  Sum of Squares (SS) 8.26 0.38 8.64  Degrees of Freedom (df) 6 9 15  Mean Square (MS) 1.38 0.04  F 33.20  5. Decision Since F = 33.20 exceeds 3.37, the null hypothesis must be rejected; in other words, we conclude that there is a difference in the activities among the catalysts, at least one of the catalyst means differsfromthe rest.  6. Determine the difference between means Using Fisher's least significant difference (LSD) to locate significant differences among catalysts when samples are not equal (Ott, 1988; Johnson and Bhattacharyya, 1985). The Fisher's least significant difference (LSD) procedure makes use of the quantity  LSD=t  a/  W  II t  ±  + J-  In,  j)  n  = the critical t value  S^, = error mean square (MSE) in AOV table = 0.04 a  = . 1 (we specified the experimentwise error rate at 90% level)  n; and nj = the respective sample sizesfromtreatment i and j We find to.05  = 1.833 (degree offreedom= 9)  When comparing between n = 2 and n = 3  218  LSD = t, S w — +  —  a/2-  = 1.833, .04|- + - I = .34 V  \2  3.  and comparing between n = 3 and n = 3 1  LSD=t jS  2 —  a/  Vi n  1 +  — n  „ .04x2 = .833, = .30 n  jJ  Comparing between n = 1 and n = 2, we get LSD = 0.46 Comparing between n = 1 and n = 3, we get LSD = 0.43 We rank the sample means from lowest to highest. Catalysts Sample Mean  3 (n=3) 5.00  6,7 (n=l) 5.20  4(n = 2) 5.80  2(n = 3) 6.20  5(n = 3) 5.86  Kn = 3) 7.14  Coke yield at 90% level of significance Conaph Catalyst  3^  CoMo, NiMo Mo 6J7  4  DHN Co 5  2  Fe 1  Conaph < CoMo, NiMo < Mo < DHN < Co < Fe  219  3.4 Data analysis: Comparison of MCR conversions from Fe. Co. Mo. CoMo. and NiMo catalysts. To  compare the MCR conversions from different  catalysts in residue  hydroconversion, the experiment of Fe and Co was repeated three times, except the experiment of Mo catalyst was repeated two times and no repetition for CoMo and NiMo catalysts. The results of MCR conversions from different catalysts were as follows: Treatment (Catalyst)  MCR Conversion (%),  Fe-DHN (1) Co-DHN (2) Mo-DHN (3) CoMo (4) NiMo (5) No. of Treatment, k = 5  31.7, 30.7, 28.5 32.8, 33.6, 29.1 33.9, 35.1 39.9 40.8  Sample Size, n;  T; = 2>i  3 3 2 1 1  90.9 95.5 69.0 39.9 40.8  N = 2>i=10  T=IExi=336.1  X;  An analysis of variance (Freund and Smith, 1986) was performed to test at the 0.05 level of significance, whether the differences among the sample means was significant.  1. Hypotheses  Null hypothesis, FL; \i\ = [i = Ha = m = Us 2  Alternative hypothesis, H : The it's are not all equal. A  2. Level of significance (a) a = 0.05  220  3. Criterion Reject the null hypothesis if F (test statistic) > 5.19, the value of Fo.os for k - 1 = 51 = 4 and N - k = 1 0 - 5 = 5 degrees of freedom, where F is to be determined by an analysis of variance. Otherwise, accept Ff, or reserve judgment.  4. Calculations Tj = 90.9, T = 95.5, T = 69.0, T = 39.9, T = 40.8, T = 336.1, and U x 2  3  4  = 11449.08.  2  5  Then, substituting these totals together with ni = 3, n = 3, n = 2, n4 = 1, n = 1, and N 2  3  5  10 into the computing formulas for the sums of squares, we get total sum of squares, SST = YLx -T /N 2  2  = 11449.08- (1/10)(336.1) = 152.75, 2  treatment sum of squares (variation among sample means, SS(Tr)) = X(T/nj) - T /N 2  2  = (l/3)(90.9) + (l/3)(95.5) + (l/2)(69.0) + (39.9) 2  2  2  2  + (40.8) -(l/10)(336.1) = 135.09 2  2  and error sum of squares (variation within the sample), SSE  = SST - SS(Tr) = 152.75 - 135.09= 17.66  Then, treatment mean square, MS(Tr) = SS(Tr)/(k-l) = 135.09/4 = 33.77, error mean square, MSE = SSE/(N-k) =17.66/5 = 3.53, and F = MS(Tr)/MSE =33.77/3.53 = 9.56. All these results are shown in the following analysis-of-variance (AOV) table:  221  Table 3.5 AOV Table for the MCR conversions. Source of Variation 135.09 17.66 152.75  Treatment (Catalysts) Error Total  Degrees of Freedom (df) 4 5 9  F 33.77 3.53  9.56  5. Decision Since F = 9.56 exceeds 5.19, the null hypothesis is reserved; we conclude that there is difference in the MCR conversion among the catalysts at 95% level of confidence.  6. Determine the difference between means Using Fisher's least significant difference (LSD) to locate significant differences among catalysts when samples are not equal (Ott, 1988; Johnson and Bhattacharyya, 1985). The Fisher's least significant difference (LSD) procedure makes use of the quantity  t  = the critical t value  S^ = error mean square (MSE) in AOV table = 3.53 a  = . 1 and .2 (we specified the experimentwise error rate at 90% and 80% level)  n; and nj = the respective sample sizesfromtreatment i and j We  find  to.os = 2.015, t .i = 1.476 (degree offreedom= 5) 0  When comparing between n = 1 and n = 2  2.015,3.53 1 +  222  and comparing between n = 2 and n = 2, we get LSD = 3.79 Comparing between n = 1 and n = 1, we get LSD = 5.36 n  1 1 2 1 2 3  LSD  j  1 2 2 3 3 3  to.05  to.i  5.36 4.64 3.79 4.37 3.46 3.09  3.92 3.40 2.77 3.20 2.53 2.27  We rank the sample means from lowest to highest. Catalysts Sample Mean  1(n=3) 30.3  2 (n=3) 31.8  3(n = 2) 34.5  4(n=l) 39.9  5(n=l) 40.8  MCR conversions at 80% level of significance  Catalyst  t1 0  Fe  Co  Mo  CoMo  NiMo  1  2  3  4  5  Fe < Co < Mo < CoMo < NiMo  MCR conversion from Mo was higer than that from Fe and Co catalysts but less than that from CoMo and NiMo catalysts at 80% level of confidence.  MCR conversions at 90% level of significance Catalyst  Fe  Co  Mo  CoMo  NiMo  1  2  3  4  5  223  to.05  Fe < Co, Fe < Mo, CoMo, and NiMo, Co < Mo < CoMo < NiMo.  The MCR conversion from CoMo and NiMo was higher than that of Fe, Co, and Mo catalysts but there was no significant difference in MCR conversion between CoMo and NiMo at 90% level of confidence.  224  3.5 Data analysis: Comparison of S conversions from Fe. Co. Mo. CoMo. and NiMo catalysts. To compare the S conversions from different catalysts in residue hydroprocessing, the experiment of Fe and Co was repeated three times, except the experiment of Mo catalyst was repeated two times and no repetition for CoMo and NiMo catalysts. The results of S conversions from different catalysts were as follows: Treatment (Catalyst) Fe-DHN (1) Co-DHN (2) Mo-DHN (3) CoMo (4) NiMo (5) No. of Treatment, k = 5  S Conversion  (%), x;  41.2, 40.0, 39.2 47.6, 50.1,44.3 56.7, 50.1 42.9 40.2  Sample Size, n 3 3 2 1 1  T; = Ex;  ;  120.4 142.0 106.8 42.9 40.2  N = I»i=10 T=ZIxi=452.3  An analysis of variance (Freund and Smith, 1986) was performed to test at the 0. 05 level of significance, whether the differences among the sample means was significant. 1. Hypotheses  Null hypothesis, Ho: \x,\ = \i2 = P-3 = M4 = 1^5 Alternative hypothesis, H : The (i's are not all equal. A  2. Level of significance a = 0.05 3. Criterion Reject the null hypothesis if F (test statistic) > 5.19, the value of Fo.os for k - 1 = 5-1 = 4 and N - k = 1 0 - 5 = 5 degrees offreedom,where F is to be determined by an analysis of variance. Otherwise, accept Ho or reserve judgment.  225  4. Calculations Tj = 120.4, T = 142.0, T = 106.8, T = 42.9, T = 40.2, T = 452.3, and X i x = 20752.8. 2  2  3  4  5  Then, substituting these totals together with ni = 3, n = 3, n = 2, n4 = 1, n = 1, and N = 2  3  5  10 into the computing formulas for the sums of squares, we get total sum of squares, SST = U x - T /N 2  2  = 20752.80 - (1/10)(452.3) = 296.18 2  treatment sum of squares (variation among sample means, SS(Tr)) = X(T /nj) - T /N 2  2  = (1/3)(120.4) + (1/3)(142.0) + (1/2)(106.8) + (42.9) 2  2  2  2  + (40.2) - (1/10)(452.3) = 255.39 2  2  and error sum of squares (variation within the sample), SSE = SST - SS(Tr) = 296.18- 255.39 = 40.79 Then, treatment mean square, MS(Tr) = SS(Tr)/(k-l) = 255.39/4 = 63.85, error mean square, MSE = SSE/(N-k) = 40.79/5 = 8.16, and F = MS(Tr)/MSE = 63.85/8.16 = 7.83. All these results are shown in the following analysis-of-variance (AOV) table:  Table 3.6 AOV Table for the S conversions Source of Variation Treatment (Catalysts) Error Total  Sum of Squares (SS) 255.39 40.79 296.18  Degrees of Freedom (df) 4 5 9  Mean Square (MS) 63.85 8.16  F 7.83  226  5. Decision Since F = 7.83 exceeds 5.19, the null hypothesis must be rejected; in other words, we conclude that there is a difference in the MCR conversions among the catalysts at 95% level of significance. 6. Determine the difference between means Using Fisher's least significant difference (LSD) to locate significant differences among catalysts when samples are not equal (Ott, 1988; Johnson and Bhattacharyya, 1985). The Fisher's least significant difference (LSD) procedure makes use of the quantity  t  = the critical t value  S^ = error mean square (MSE) in AOV table = 8.16 a = . 1 and .05 (we specified the experimentwise error rate at 80% and 90% level) n; and nj = the respective sample sizesfromtreatment i and j We  find  to.i = 1.476, to.os = 2.015 (degree offreedom= 5)  When comparing between n = 1 and n = 2  LSD = t  and comparing between n = 2 and n = 2, we get LSD = 4.2 Comparing between n = 1 and n = 1, we get LSD = 6.0  227  ni  n  1 1 2 1 2 3  LS D  J  1 2 2 3 3 3  to.05  to.i  8.14 7.05 5.76 6.65 5.25 4.70  5.96 5.16 4.22 4.87 3.85 3.44  We rank the siimple meansfromlowe;st to highest. Catalysts Sample Mean  1 (n = 3) 40.3  5 (n = 1) 40.2  4(n=l) 42.9  2(n = 3) 47.3  3(n = 2) 53.4  S Conversion at 80% level of significance Catalyst  Fe 1  NiMo CoMo Co 5 4 2  Mo 3  Fe< NiMo < CoMo, Fe and NiMo < Co and Mo, CoMo < Co < Mo Mo had the highest hydrodesulfurization activity among Fe, Co, CoMo, and NiMo catalysts at 80% level of confidence. S Conversion at 90% level of significance Catalyst  Fe 1  NiMo CoMo Co 5 4 2  Mo 3  Fe< NiMo < CoMo, Fe and NiMo < Co and Mo, CoMo < Co < Mo.  228  Mo had the highest hydrodesulfurization activity among Fe, Co, CoMo, and NiMo catalysts at 90% level of confidence.  229  3.6 Data analysis: Comparison of >525°C conversions from Fe. Co. Mo. CoMo. and NiMo catalysts. To compare the >525°Cfractionconversions from different catalysts in residue oil hydroprocessing, the experiment of Fe and Co was repeated two times, and no repetition for Mo, CoMo and NiMo. The results of >525°C fraction conversions of different catalysts were as follows: Treatment (Catalyst)  >525°C Conversion (%),  Sample Size, n;  T; = Sxj  1 2 1 1 1 N = »i=6  50.4 99.2 47.7 55.0 54.4 T=IIx;=306.7  X;  Fe-DHN (1) Co-DHN (2) Mo-DHN (3) CoMo (4) NiMo (5) No. of Treatment, k = 5  50.4 48.0, 51.2 47.7 55.0 54.4  An analysis of variance (Freund and Smith, 1986) was performed to test at the 0.05 level of significance, whether the differences among the sample means was significant.  1. Hypotheses  Null hypothesis, Ho: \i = \x = |i3 = pu = Us y  2  Alternative hypothesis, H : The ix's are not all equal. A  2. Level of significance a = 0.05  230  3. Criterion Reject the null hypothesis if F (test statistic) > 225, the value of F0.05 for k - 1 = 51 = 4 and N - k = 6- 5 = l degrees offreedom,where F is to be determined by an analysis of variance. Otherwise, accept Ho or reserve judgment.  4. Calculations Ti = 50.4, T = 99.2, T = 47.7, T = 55.0, T = 54.4, T = 306.7, and £ £ x = 15725.25. 2  2  3  4  5  Then, substituting these totals together with ni = 1, n = 2, n = 1, at = 1, n = 1, and N = 2  3  5  6 into the computing formulas for the sums of squares, we get total sum of squares, SST = X i x - T /N = 15725.25 - (l/6)(306.7) = 47.77 2  2  2  treatment sum of squares (variation among sample means, SS(Tr)) = Z(Ti /n )- T /N 2  2  ;  =(50.4) + (l/2)(99.2) + (47.7) + (55.0) 2  2  2  2  + (54.4) - (l/6)(306.7) = 42.65 2  2  and error sum of squares (variation within the sample), SSE = SST - SS(Tr) = 47.77-42.65 = 5.12 Then, treatment mean square, MS(Tr) = SS(Tr)/(k-l) = 42.65/4 = 10.66, error mean square, MSE = SSE/(N-k) = 5.12/1 = 5.12, and F = MS(Tr)/MSE = 10.66/5.12 = 2.08. All these results are shown in the following analysis-of-variance (AOV) table: Table 3.7 AOV Table for the >525°Cfractionconversions Source of Variation Treatment (Catalysts) Error Total  Sum of Squares (SS) 42.65 5.12 47.77  Degrees of Mean Square Freedom (df) (MS) 10.66 4 1 5.12 5  F 2.08  231  5. Decision Since F = 2.08 does not exceed 225, the null hypothesis is preserved; we conclude that there is no difference in the >525°C fraction conversions among the catalysts at 95% level of confidence.  232  3.7 Data analysis: Comparison of asphaltene conversions from Fe, Co, Mo, CoMo. and NiMo catalysts. To compare the asphaltene conversions from different catalysts in residue oil hydroconversion, the experiment of Fe and Co was repeated three times, except the experiment of Mo catalyst was repeated two times and no repetition for CoMo and NiMo catalysts. The results of asphaltene conversions from different catalysts were as follows: Treatment (Catalyst)  Asphaltene Conversion (%),  Fe-DHN (1) Co-DHN (2) Mo-DHN (3) CoMo (4) NiMo (5) No. of Treatment, k = 5  50.4, 54.8, 50.6 56.3, 57.7, 53.4 38.6, 37.4 48.6 50.2  Sample Size, n;  T; =  5>i  Xj  3 3 2 1 1  155.8 167.4 76 48.6 50.2  N = >>i=10  T=IXx;= 498.0  An analysis of variance (Freund and Smith, 1986) was performed to test at the 0.05 level of significance, whether the differences among the sample means was significant.  1. Hypotheses  Null hypothesis, Ho: \i\ = \i = |i3 = \U = |^ts 2  Alternative hypothesis, H : The |i's are not all equal. A  2. Level of significance a = 0.05  233  3. Criterion Reject the null hypothesis if F (test statistic) > 5.19, the value of Fo.os for k - 1 = 51 = 4 and N - k = 1 0 - 5 = 5 degrees of freedom, where F is to be determined by an analysis of variance. Otherwise, accept Ho or reserve judgment. 4. Calculations Ti = 155.8, T = 167.4, T = 76.0, T = 48.6, T = 50.2, T = 498.0, and XXx = 25225.89. 2  2  3  4  5  Then, substituting these totals together with ni = 3, n = 3, n = 2, o» = 1, n = 1, and N 2  3  5  10 into the computing formulas for the sums of squares, we get total sum of squares, SST = U x - T /N 2  2  = 25225.89- (1/10)(498.0) = 424.49 2  treatment sum of squares (variation among sample means, SS(Tr)) = X(T /nj) - T /N 2  2  = (1/3)(155.8) + (1/3)(167.4) + (l/2)(76.0) + (48.6) 2  2  2  2  + (50.2) -(l/10)(498.0) =401.85 2  2  and error sum of squares (variation within the sample), SSE = SST - SS(Tr) = 424.49- 401.85 = 22.64 Then, treatment mean square, MS(Tr) = SS(Tr)/(k-l) = 401.85/4 = 100.46, error mean square, MSE = SSE/(N-k) = 22.64/5 = 4.53, and F = MS(Tr)/MSE = 100.46/4.53 = 22.19. All these results are shown in the following analysis-of-variance (AOV) table:  Table 3.8 AOV Table for the asphaltene conversions Source of Variation Treatment (Catalysts) Error Total  Sum of Squares (SS) 401.85 22.64 424.49  Degrees of Freedom (df) 4 5 9  Mean Square (MS) 100.46 4.53  F 22.19  234  5. Decision Since F = 22.19 exceeds 5.19, the null hypothesis must be rejected; in other words, we conclude that there is a difference in the asphaltene conversions among the catalysts, at least one of the catalyst means differs from the rest. 6. Determine the difference between means Using Fisher's least significant difference (LSD) to locate significant differences among catalysts when samples are not equal (Ott, 1988; Johnson and Bhattacharyya, 1985). The Fisher's least significant difference (LSD) procedure makes use of the quantity  t  = the critical t value  S^ = error mean square (MSE) in AOV table = 4.53 a = .2 and. 1 (we specified the experimentwise error rate at 80% and 90% level) n; and nj = the respective sample sizesfromtreatment i and j We  find  to.05 =  2.015, t  01  = 1.476 (degree offreedom= 5)  When comparing between n = 1 and n = 2  LSD=t. a/2  ( 1  +  1 n  = 2.015, 4.53 1 +  and comparing between n = 2 and n = 2, we get LSD = 4.29 Comparing between n = 1 and n = 1, we get LSD = 6.06  235  n  1 1 2 1 2 3  LSD  j  1 2 2 3 3 3  to.05  to.i  6.06 5.25 4.29 4.95 3.91 3.50  4.44 3.85 3.14 3.63 2.87 2.56  We rank the sample means from lowest to highest. 3(n = 2) Catalysts Sample Mean 38.0  4(n=l) 48.6  5(n=l) 50.2  Kn = 3) 51.9  2(n = 3) 55.8  Asphaltene conversion at 80% level of significance. Catalyst  Mo 3  CoMo NiMo 4 5  Fe 1  Co 2  Mo < CoMo < NiMo < Fe < Co  The asphaltene conversion from Co catalyst was the highest and that from Mo was the lowest at 80% level of confidence. Asphaltene conversion at 90% level of significance. Catalyst  Mo 3  CoMo NiMo 4 5  Fe 1  Co 2  Mo < CoMo < NiMo < Fe < Co  The asphaltene conversionfromMo catalyst was the lowest at 90% level of confidence.  236  3.8 Data analysis: Comparison of coke/asphaltene ratios from Fe, Co, Mo, CoMo. and NiMo catalysts. To compare the coke/asphaltene ratios from different catalysts in residue oil hydroconversion, the experiment of Fe and Co was repeated three times, except the experiment of Mo catalyst was repeated two times and no repetition for CoMo and NiMo catalysts. The results of coke/asphaltene ratios from different catalysts were as follows: Treatment (Catalyst) Fe(l) Co-DHN (2) Mo-DHN (3) CoMo (4) NiMo (5) No.of Treatment, k = 5  Coke/Asphalt ene  Sample Size,  Ratio (%), x; 14.68, 12.77, 13.83 11.72, 10.40, 11.23 15.02, 15.51 10.70 10.36  ns 3 3 2 1 1  N = Ini=10  Ti = Zxi 41.29 33.36 30.53 10.70 10.36 T=ZXx; = 126.23  An analysis of variance (Freund and Smith, 1986) was performed to test at the 0. 05 level of significance, whether the differences among the sample means was significant.  1. Hypotheses  Null hypothesis, Ho: p.i = | l = 2  =  pU = p.5  Alternative hypothesis, H : The pi's are not all equal. A  2. Level of significance a = 0.05  237  3. Criterion Reject the null hypothesis if F (test statistic) > 5.19, the value of F0.05 for k - 1 = 51 = 4 and N - k = 1 0 - 5 = 5 degrees of freedom, where F is to be determined by an analysis of variance. Otherwise, accept Ho or reserve judgment. 4. Calculations Ti = 41.29, T = 33.36, T = 30.53, T = 10.70, T = 10.36, T = 126.23, and X i x = 2  2  3  4  5  1629.80. Then, substituting these totals together with ni = 3, n = 3, n = 2,114 = 1, n = 1 2  3  5  and N = 10 into the computing formulas for the sums of squares, we get total sum of squares, SST = H x - T /N 2  2  = 1629.80 - (1/10)(126.23) = 36.37 2  treatment sum of squares (variation among sample means, SS(Tr)) = I(T /hj)- T /N 2  2  = (1/3)(41.29) (l/3)(33.36) + (l/2)(30.53) +(10.70) + (10.36) 2  2  2  2  2  - (1/10)(1629.80) =33.53 2  and error sum of squares (variation within the sample), SSE = SST - SS(Tr) = 36.37- 33.53 = 2.84 Then, treatment mean square, MS(Tr) = SS(Tr)/(k-l) = 33.53/4 = 8.38, error mean square, MSE = SSE/(N-k) = 2.84/5 = 0.57, and F = MS(Tr)/MSE = 8.38/0.57 = 14.74. All these results are shown in the following analysis-of-variance (AOV) table:  Table 3.9 AOV Table for the coke yield/asphaltene ratios Source of Variation Treatment (Catalysts) Error Total  Sum of Squares (SS) 33.53 2.84 36.37  Degrees of Freedom (df) 4 5 9  Mean Square (MS) 8.38 0.57  F 14.74  238  5. Decision  Since F = 14.74 exceeds 5.19, the null hypothesis must be rejected; in other words, we conclude that there is a difference in the coke/asphaltene ratios among the catalysts, at least one of the catalyst means differs from the rest. 6. Determine the difference between means  Using Fisher's least significant difference (LSD) to locate significant differences among catalysts when samples are not equal (Ott, 1988; Johnson and Bhattacharyya, 1985). The Fisher's least sigriificant difference (LSD) procedure makes use of the quantity  +  t  —  = the critical t value  S^ = error mean square (MSE) in AOV table = 0.569 a = . 1 and .2 (we specified the experimentwise error rate at 90% and 80% level) n; and nj = the respective sample sizesfromtreatment i and j We  find  to.05 =  2.015, to.i = 1.476 (degree offreedom= 5)  When comparing between n = 1 and n = 2  LSD=t  and comparing between n = 2 and n = 2, we get LSD =1.11 Comparing between n = 1 and n = 1, we get LSD = 1.57  239  n  1 1 2 1 2 3  LSD  j  1 2 2 3 3 3  to.05  to.i  2.15 1.86 1.52 1.75 1.39 1.24  1.57 1.36 1.11 1.29 1.02 0.91  We rank the sample meansfromlowest to highest. Catalysts Sample Mean  5(n=l) 10.36  4(n=l) 10.7  2(n = 3) 11.12  l ( n = 3) 13.76  3(n = 2) 15.27  Coke/asphaltene ratio at 80% level of significance. Catalyst  NiMo 5  CoMo Co 4 2  Fe 1  Mo 3  NiMo < CoMo < Co < Fe < Mo There was no synergistic effect of CoMo catalyst on coke/asphaltene ratio at 80% level of confidence. Coke/asphaltene ratio at 90% level of significance. Catalyst  NiMo 5  CoMo Co 4 2  Fe 1  Mo 3  NiMo < CoMo < Co < Fe < Mo  The result was similar to that at 80% level of confidence. There was no synergistic effect of CoMo catalyst on coke/asphaltene ratio at 90% level of confidence.  240  3.9 Determination of number average, standard deviation, standard error of mean, and %standard error of mean. The number average of sample d was calculated as follows: n  n Where,  d; = sample i n = number of sample  The number standard deviation 8 was calculated as follows: n  y  n-1  The standard error of mean, S.E. was calculated as follows: S.E. =  4= Vn  The percent standard error of mean was calculated as follows: %S.E.= - 1 S  E  X 1 0  °  3.10 The estimated cost of Co prepared in microemulsion compared to cobalt naphthenate  Cdn$/Unit  Cdn$  Cobalt Naphthenate (Sigma Co.)  Amount Used 0.271875 g  33.30/237.5 g  0.0381  PE4LE (Sigma Co.) LiBH, (Sigma Co.) Co(N0 ) .6H 0 (Sigma Co.)  2.5 ml 1.0 ml 0.1074 g  57.80/1053 ml 164.70/800 ml 94.90/500 g  0.1372 0.2059 0.0204 0.3635  Chemicals  3  2  2  241  3.11 Determination of sample size (Ott. 1988). Sample size required for a 100(l-a)% confidence interval for (i (population mean) of the form y (sample mean) ± E: n  _ (Z  f c  2  a/2  E  2  where n  = sample size required  Z /2 a  = a value of Z having a tail area of a/2 to its right. = 1.96 at 95% of confidence  a  = standard deviation = 2 (value of sample standard deviation, s)  2E  = the width of interval or tolerable error (specified value) =1.0  n  - < >'< >' (0.5) 1%  2  2  =62  Even though a is estimated by the sample standard deviation s, the formula is still a very good approximation for large sample sizes. As a very rough rule, this formula can be used when n is larger than 30.  242  

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:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0058629/manifest

Comment

Related Items