Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

A study of colloidal hydrocracking catalysts prepared in reverse micelles Hall, Andrew Graham 1996

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

Item Metadata

Download

Media
831-ubc_1996-0337.pdf [ 14.49MB ]
Metadata
JSON: 831-1.0058613.json
JSON-LD: 831-1.0058613-ld.json
RDF/XML (Pretty): 831-1.0058613-rdf.xml
RDF/JSON: 831-1.0058613-rdf.json
Turtle: 831-1.0058613-turtle.txt
N-Triples: 831-1.0058613-rdf-ntriples.txt
Original Record: 831-1.0058613-source.json
Full Text
831-1.0058613-fulltext.txt
Citation
831-1.0058613.ris

Full Text

A STUDY O F C O L L O I D A L H Y D R O C R A C K I N G C A T A L Y S T S P R E P A R E D IN R E V E R S E M I C E L L E S by ANDREW GRAHAM HALL B.Sc. The University of Cape Town, 1993 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF APPLIED SCIENCE in THE FACULTY OF GRADUATE STUDIES DEPARTMENT OF CHEMICAL ENGINEERING We accept this thesis as conforming to the^e^uired^tandard THE UNIVERSITY OF BRITISH COLUMBIA May, 1996 © Andrew G. Hall, 1996 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department The University of British Columbia Vancouver, Canada Date br Ta^ e. DE-6 (2/88) Abstract Slurry phase reactors employing dispersed metal sulfide catalysts offer the potential of high conversion with minimal coke yield for the hydrocracking of heavy oil. Currently, the dispersed catalysts are prepared by the addition of organometallic compounds to the heavy oil feed, but this synthesis technique offers little control over the size of the catalyst particles. Colloidal suspensions of a wide range of metals and metal compounds in hydrocarbons can be prepared from water-soluble metal salts dissolved within the water pools of a microemulsion. This synthesis technique allows for the simple control of particle size, a factor which together with their narrow size distribution makes these colloids potentially attractive as dispersed heavy oil hydrocracking catalysts. The synthesis of reduced metal and metal sulfide colloids in the water pools of the water/polyoxyethylene-4-laurylether/hexane microemulsion was investigated in the present study. The sizes of the reverse micelles were determined by dynamic light scattering (DLS), while the reduced Ni, Co and Fe colloids and the Ni, Co and Fe sulfide colloids were characterized by DLS, transmission electron spectroscopy (TEM), energy dispersive x-ray spectrometry (EDX) and x-ray photoelectron spectroscopy (XPS). The catalytic activity of the metal sulfide catalysts was also determined using the hydrocracking of diphelylmethane as a model reaction. The water: surfactant ratio (co) and metal ion concentration were found to be the key factors affecting the size of the reverse micelles in the water/PE4LE/hexane system. Monodisperse Co and Fe colloids with sizes ranging from 10-23 nm were prepared in the microemulsion system by the addition of N2FI4, and the size of the metal colloids was found to be directly related to co. ii Ni, Co and Fe sulfide colloids were prepared in reverse micelles using 5% H2S in H 2, and XPS analysis identified MS and CoS2 on the surface of these colloids. The metal sulfides proved difficult to characterize due to their extreme sensitivity to atmospheric oxygen. The Fe sulfide colloids oxidized readily, and could not be identified using XPS. The NiS and CoS2 colloids had average sizes of 67 and 71 nm respectively (as determined by TEM), and were more polydisperse than the reduced metal colloids prepared in the same system. The metal sulfide catalysts prepared in the water/polyoxyethylene-4-laurylether/hexane microemulsion were found to he less active for the hydrocracking of diphenylmethane than a dispersed catalyst prepared from the decomposition of Co naphthenate. The crystallite size was similar for both catalyst preparations (20-30 nm), suggesting that diffusion limitations may have controlled the rate of reaction in the case of the aggregated metal sulfide catalyst prepared in the microemulsion. iii Table of Contents Abstract ii List of Tables ix List of Figures xi Acknowledgements xiii CHAPTER 1 Introduction 1 1.1 Motivation for the study 1 1.2 Objectives of the present study 3 CHAPTER 2 Literature Review 5 2.1 Heavy oil upgrading 5 2.1.1 Introduction 5 2.1.2 Heavy oil upgrading processes 5 2.1.3 Catalytic hydrocracking of heavy oil 6 2.2 Dispersed catalysts for heavy oil hydrocracking 8 2.2.1 Introduction 8 2.2.2 Hydroconversion mechanisms in the presence of dispersed catalysts 9 2.2.3 The synthesis of dispersed metal catalysts 11 2.2.4 Effect of precursor solubility on catalyst activity 12 2.2.5 Size distribution of dispersed catalysts 12 2.2.6 Active species of dispersed catalysts 14 2.2.7 Disadvantages of dispersed catalysts derived from organometallic compounds 14 2.3 Preparation of metal colloids in reverse micelles 15 2.3.1 Introduction 15 2.3.2 Emulsions and microemulsions 15 2.3.3 Colloid formation in reverse micelles 16 2.3.4 Factors affecting the size of reverse micelles and colloids 20 2.3.5 Catalytic activity of colloids synthesized in reverse micelles ... 25 2.3.6 Preparation of dispersed hydrocracking catalysts in reverse micelles 27 3 Experimental Methods 28 3.1 Microemulsion preparation and colloid synthesis 28 3.1.1 Microemulsion preparation 28 3.1.2 Synthesis of reduced metal colloids using N2H4 29 3.1.3 Synthesis of metal sulfide colloids using dimethyl disulfide 29 3.1.4 Synthesis of metal sulfide colloids using H2S 30 3.2 Micelle characterization by dynamic light scattering (DLS) 31 3.3 Colloid characterization 32 3.3.1 Dynamic light Scattering (DLS) 32 3.3.2 Transmission electron microscopy (TEM) 33 3.3.3 Energy disperse x-ray spectrometry (EDX) 33 3.3.4 X-Ray diffraction (XRD) 34 3.3.5 X-Ray photoelectron spectroscopy (XPS) 35 3.3.6 BET surface area measurement 36 3.4 Activity Measurement 37 3.4.1 Reaction system 37 3.4.2 Modeling of the reaction system 37 3.4.3 Experimental apparatus and methods 38 3.4.4 Analysis of the activity test products using gas chromatography 40 3.4.5 Calculation of conversion 42 CHAPTER 4 Catalyst Preparation and Characterization 43 4.1 Microemulsion preparation 43 4.1.1 Introduction 43 4.1.2 Factors affecting the size of reverse micelles in the water/ PE4LE/hexane system. 44 4.2 The synthesis of reduced metal colloids in the water/PE4LE/ hexane system 49 4.2.1 Introduction 49 4.2.2 Characterization of reduced metal colloids 49 4.3 Synthesis of metal sulfide colloids with dimethyl disulfide 55 4.4 Synthesis of metal sulfide colloids with H2S 57 4.4.1 Introduction 57 4.4.2 Physical characterization of metal sulfide colloids using TEM andEDX 58 4.4.3 Characterization of metal sulfide colloids using XPS 64 4.4.4 Characterization of metal sulfide colloids using BET surface area measurement 69 4.5 Summary of the major findings on colloidal catalyst synthesis in reverse micelles 69 CHAPTER 5 Activity Measurements 71 5.1 Introduction 71 5.2 Phase equilibrium of the reaction system 71 5.3 Preliminary activity measurements 72 5.4 Activity measurements using recovered colloids 78 5.4.1 The use of recovered colloids 78 5.4.2 Experimental results 79 5.4.3 Catalyst characterization using XRD 82 5.5 Summary of the activity measurements performed 85 CHAPTER 6 Conclusions and Recommendations for Future Work 87 6.1 Conclusions 87 6.2 Recommendations and Future Work 89 REFERENCES 91 APPENDIX I 97 1.1 Microemulsions prepared 98 1.2 List of activity measurements 99 1.3 Modeling of the reaction system using ASPEN process simulation package 101 1.4 GC temperature program 102 1.5 GC standard calibration mixtures 103 1.6 GC calibration data and calibration equations 104 1.7 Equations for the calculation of conversion 108 APPENDIX H 109 2.1 TEM photographs of reduced Co and Fe colloids 110 2.2 EDX of reduced Co and Fe colloids 115 2.3 EDX spectrum of Ni sulfide prepared using DMDS 116 2.4 Synthesis of metal sulfide colloids in the water/PE4LE/hexane microemulsion using 5% H2S in H 2 117 2.5 TEM Photographs of metal sulfide colloids synthesized using 5% H2S in H 2 118 2.6 Explanation of terms used in Table 4.6 128 2.7 Standard binding energies of various Ni, Co and Fe species 129 2.8 BET surface area calculations 130 APPENDIX III 131 3.1 ASPEN simulation of the reaction system 132 3.2 Sample calculation of rate constant for activity measurements 133 3.3 Data for the preliminary activity measurements 134 3.4 Data for activity measurements with recovered colloids 146 3.5 GC traces of experiment #19 and decomposition of Co naphthenate 164 3.6 XRD spectra of the spent catalysts from the second set of activity measurements 165 List of Tables Table 2.1 Comparison between heavy oil hydrocracking reactor technologies 8 Table 2.2 Examples of metals and inorganic compounds prepared in reverse micelles . 17 Table 3.1 Calibration curves used to calculate the number of moles of benzene, toluene and DPM in the reactor product 42 Table 4.1 Hydrodynamic diameters of reverse micelles in the water/PE4LE/hexane system 45 Table 4.2 Hydrodynamic diameters of reverse micelles in the water/PE4LE/DHN system 45 Table 4.3 Reduced metal colloid sizes as determined by DLS 50 Table 4.4 Statistical analysis of the reduced Co colloid sizes in photographs #925, #926 and the reduced Fe colloid sizes in photograph #929 54 Table 4.5 Results of nickel sulfide colloid synthesis experiments using DMDS 56 Table 4.6 Simple statistical analysis of the Ni sulfide particle sizes in photographs #127,130, 133 and 134 and Co sulfide particle sizes in photographs #138 and 142 60 Table 5.1 Simulation results for the simplified reaction system 72 Table 5.2 Results of the preliminary activity measurements conducted in water/ PE4LE/decalin microemulsions 73 Table 5.3 Results of the additional activity measurement conducted in a water/PE4LE/ decalin microemulsions 78 Table 5.4 Results of the second set of catalytic activity measurements 80 Table 5.5 Summary of the catalyst characterization using XRD List of Figures Figure 2.1 Reaction mechanism in the presence of dispersed catalysts 10 Figure 2.2 Proposed mechanisms for metal boride formation in reverse micelles 19 Figure 2.3 Variation of the average radius (rM) of the water cores in the water/CTAB/ n-Hexanol microemulsion as a function of (a) water content and (b) Ni(H) concentration 21 Figure 2.4 Variation of the average diameter (d in A) of the nickel boride particles prepared in water/CTAB/n-Hexanol as a function of water content and Ni(U) concentration 24 Figure 3.1 Flow diagram of batch autoclave system used for colloid synthesis and activity measurements 30 Figure 4.1 Effect of co on the hydrodynamic diameters of reverse micelles in the water/PE4LE/hexane system 46 Figure 4.2 Effect of metal ion concentration within the water pool on the hydrodynamic diameter of reverse micelles in the water/PE4LE/hexane system. 48 Figure 4.3 Relationship between reverse micelle size and reduced metal colloid size 52 Figure 4.4 Size distribution of Co colloids present in TEM photographs #925 and #926 53 Figure 4.5 Size distribution of reduced Fe and Co colloids prepared in water/ PE4LE/hexane microemulsions 55 xi Figure 4.6 : Size distributions of nickel sulfide and cobalt sulfide colloids synthesized in experiments #30 and #31 60 Figure 4.7 EDX elemental analysis of a colloid from experiment #026 61 Figure 4.8 EDX elemental analysis of a colloid from experiment #027 61 Figure 4.9 EDX elemental analysis of a colloid from experiment #028 62 Figure 4.10 EDX elemental analysis of a colloid from experiment #029 62 Figure 4.11 Experiment #026 - XPS Scan of Ni 2p3 / 2 region 66 Figure 4.12 Experiment #026 - XPS Scan of S 2p region 66 Figure 4.13 Experiment #027 - XPS Scan of Co 2p3/2 region 67 Figure 4.14 Experiment #027 - XPS Scan of S 2p region 67 Figure 4.15 Experiment #028 and #029 - XPS Scan of Fe 2p3 / 2 region 68 Figure 4.16 Experiment #028 and #029 - XPS Scan of S 2p region 68 Figure 5.1 GC trace of a typical reaction feed for the preliminary activity measurements 75 Figure 5.2 GC trace of a mixture of benzene, toluene, decalin, PE4LE and decalin corresponding to 5% conversion of DPM to benzene and toluene...76 Figure 5.3 GC trace of a typical reaction product from the preliminary activity measurements 76 Acknowledgements First of all I would like to thank my supervisor Dr. Kevin Smith for his support, enthusiasm and guidance throughout my research. I would like to acknowledge the Alberta Department of Energy for their generous financial support of this project. Special thanks must also go to Dr. Barbara Frisken of the Department of Physics, Simon Fraser University for performing the light scattering measurements; and to Mary Mager of the Department of Metals and Materials Engineering, U.B.C. for helping me with the TEM work. I would also like to thank the fellow members of my research group for the valuable assistance which they have given me with my experimental work, and for creating a pleasant and stimulating work environment. In closing I would like to dedicate this thesis to my parents, for all the loving support, advice and encouragement which they have always given me over the years. xiii Chapter 1: Introduction 1.1 Motivation for the study A significant portion of the world's fossil fuel reserves exist as hydrocarbons referred to as heavy oils. Heavy oils are defined in terms of their physical properties as mixtures of hydrocarbons which have a viscosity of 102 - 105 mPa.s, a density of 934-1000 kg/m3 and an API gravity of 10-20. They originate from various sources including coal liquids, shale oils, oil sands and crude distillate bottoms. Canadian reserves of heavy oil in the Cold Lake and Athabasca regions of Alberta are estimated at between 2000-3000 billion barrels, compared to conventional oil reserves in the Middle East of 1500 billion barrels. Clearly these vast reserves of heavy oil in Alberta will supply an increasing portion of the worlds energy needs in the future, and it is essential that processes are developed to profitably pretreat these heavy oils. Heavy oils contain high levels of sulfur, metals and cokeable materials. The presence of these contaminants means that heavy oils must be upgraded to lighter products which can be pumped and processed more easily in conventional oil refineries (Del Bianco et al. 1993). The aim of the upgrading process is to increase the H:C ratio of the feed stock, as well as to remove the heteroatoms and metals (Laine and Trimm, 1982). The H:C ratio of heavy oil can be increased by either coking, or catalytic hydrocracking. Coking is currently the most common method used, but catalytic hydrocracking processes have gained increasing popularity in recent times due to their higher conversion (>90%), higher distillate yield and lower severity compared to coking processes (Dautzenberg and De Deken, 1985). Most catalytic hydrocracking processes utilize Co, W, Ni and Mo sulfide catalysts supported on alumina (Gates et al, 1979). However, commercial heavy oil hydrocracking l processes which employ these supported catalysts experience extremely high rates of catalyst deactivation due to the deposition of coke and metals in the catalyst pores (Oelderik et al., 1989; Ternan, 1983). These operational problems associated with supported catalysts have led to the development of 'slurry phase' reactors which use unsupported or 'dispersed phase' metal sulfide catalysts (Del Bianco et ai, 1993). The key advantage of dispersed catalysts is their ability to suppress coke formation. Since the catalyst particles are extremely small and highly dispersed within the heavy oil, the interp article distance is radically reduced. As a result, the catalyst particles are able to intercept and hydrogenate free radical intermediates which would otherwise condense to form coke (Bearden and Aldridge, 1981). In addition, the extremely small size (typically sub-micron) of dispersed catalysts results in a large increase in specific surface area compared to supported hydrocracking catalysts. Consequently, a much lower catalyst loading can be used in slurry reactors, which may result in decreased catalyst costs. Dispersed metal catalysts are most commonly prepared by the addition of oil soluble metal salts (organometallics) to the heavy oil feed of the slurry reactor. Although these organometallic precursors yield active hydrocracking catalysts, they provide little control over the size of the dispersed catalyst particles. In addition, the activity of dispersed catalysts is strongly dependent on the solubility of the organometallic additive in the heavy oil, and the choice of catalysts which can be prepared from organometallic precursors is thus fairly limited (Dabkowski et ai, 1991; Bearden and Aldridge, 1981; Hirschon and Wilson, 1989). Further understanding of the catalytic action of dispersed catalysts is hampered by the lack of control over particle size and the narrow selection of oil-soluble organometallics which are available. Consequently, there is a need to find a method for synthesizing dispersed catalysts in a hydrocarbon medium whereby the size of the catalyst particles can be easily controlled, and a large range of catalysts can be prepared. 2 A number of researchers have recently demonstrated that a wide range of metals and metal compounds dispersed in organic solvents can be prepared in the reverse micelles of a microemulsion (see section 2.3). Metal colloids prepared in reverse micelles are a potentially attractive alternative to catalysts prepared from the decomposition of organometallics for a number of reasons. Firstly, a number of studies have shown that the size of colloids synthesized within reverse micelles can be manipulated by simply changing the composition of the microemulsion (Ravet et al., 1984; Nagy et al, 1983; Modes and Lianos, 1989; Pileni et al., 1992). In addition, a far wider range of metal compounds, including bimetallic particles, can be synthesized in reverse micelles compared to the decomposition of organometallics. Finally, it is believed that colloidal catalysts prepared in microemulsions of oil-soluble organics can be very easily mixed with heavy oil and attain a far higher dispersion than colloidal catalysts from organometallics. 1.2 Objectives of the present study The present study focuses on the preparation and characterization of metal and metal sulfide colloids prepared in the water/polyoxyethylene-4-laurylether/hexane microemulsion, and the application of this colloidal catalyst to the hydrocracking of a model compound, diphenylmethane. The current research had a number of more specific objectives: 1. To determine the factors affecting the size of the reverse micelles in the water/ polyoxyethylene-4-laurylether/hexane microemulsion. 2. To establish whether a relationship exists between the size of the reverse micelles in the water/polyoxyethylene-4-laurylether/hexane system and the size of solid colloids synthesized within the reverse micelles. For this purpose, the synthesis of reduced Ni, Co and Fe colloids was initially studied as a model intramicellular synthesis reaction. 3 3. To determine a suitable technique for the synthesis of metal sulfide colloids in the water/polyoxyethylene-4-laurylether/hexane microemulsion, and whether the size of these metal sulfide colloids can be controlled in the microemulsion. I f possible, the metal sulfide species formed during the synthesis reaction were also to be identified. 4. Finally, the catalytic activity of the metal sulfide colloids synthesized in the water/polyoxyethylene-4-laurylether/hexane system was measured. The hydrocracking of a model compound diphenylmethane was studied for this purpose. The relative catalytic activity of these metal sulfide colloids synthesized in reverse micelles compared to catalysts prepared from organometallic compounds was of primary concern in these activity measurements. 4 Chapter 2 : Literature Review 2.1 Heavy oil upgrading 2.1.1 Introduction Heavy oil represents an extremely important energy source to both the Canadian and world economies. Currently, 21% of Canada's total petroleum needs are derived from Alberta oil sands (Lupien, 1995). On a global scale, Venezuelan and Canadian heavy oil reserves are approximately 4 times larger than known reserves of conventional crude oil in the Middle East, and there is little doubt that heavy oil will supply an increasing percentage of the worlds energy needs in the future. Consequently, the successful and profitable utilization of these heavy oil reserves is critical to the long term future of the petrochemical industry, and it is essential that processes are developed to handle and pre-treat these heavy feeds. 2.1.2 Heavy oil upgrading processes Heavy oils contain components which make processing in existing refineries extremely difficult (Del Bianco et al, 1993), including heteroatoms (sulfur, nitrogen, oxygen), metals (nickel and vanadium) and considerable amounts of cokeable materials (referred to as Conradson Carbon Residue, CCR). Consequently, heavy oil must be upgraded to convert the heavy feeds to lower boiling point products which can be pumped and processed more easily. The basic aim of the upgrading process is to increase the H:C ratio in the feed stock, though other process requirements include the reduction of overall molecular weight and the removal of metals and heteroatoms (Laine and Trimm, 1982). 5 The H:C ratio of heavy oil can be increased by using 2 different approaches: carbon rejection (coking), or hydrogen addition. Coking is a proven method currently favored by most US refiners, and involves removing carbon as coke via a pyrolysis reaction (Del Bianco et al, 1993). Despite their popularity, coking processes suffer from several disadvantages. Firstly, distillate yields are low since a large fraction of the feed is converted to gas and coke. Consequently, a higher crude intake is required to produce an equal quantity of light hydrocarbon product (Dautzenberg and De Deken, 1985). In addition, coking distillates are generally of poor quality (low CCR reduction and poor heteroatom and metal removal) and the large volumes of high sulfur, high metals coke produced present disposal problems (Del Bianco et al, 1993). 2.1.3 Catalytic hydrocracking of heavy oil Heavy oils can also be upgraded by the catalytic addition of hydrogen. Hydrogen is added to the heavy oil directly, removing inorganic materials and decreasing molecular weight by a hydrogenolysis/hydrocracking mechanism (Laine and Trimm, 1982). Considerable attention has recently been paid to catalytic hydrocracking processes, since they offer potential conversions of over 90%, including the removal of asphaltene coke precursors (Dautzenberg and De Deken, 1985). Hydrocracking processes are operated under less severe conditions than coking operations (430-460 °C compared to >500 °C in the case of coking), resulting in improved process economics in most cases (Sanford, 1996). Catalytic hydrocracking processes utilize metal sulfide catalysts to provide a source of hydrogen radicals (H») through the catalytic dissociation of H2. Hydrogen radicals (H«) promote bond cleavage reactions and control the retropolymerization processes, thus reducing the production of coke (Del Bianco et al, 1993). In most catalytic hydrocracking processes, combinations of the sulfides of Co, Mo, Ni and W, supported on 6 alumina are used (Weisser and Landa, 1970; Gates et al, 1979). Thomas and co-workers proposed the following mechanism for the generation of H« by iron sulfide (Thomas et al, 1982): FeS2 ->FeS + S S + H 2 -> H» + HS« HS«+H 2 ->H« +H 2S H2S + FeS -> FeS2 + H 2 Catalytic hydrocracking catalysts can be divided into two main groups: supported catalysts and unsupported or dispersed catalysts. At present, all commercial hydrocracking processes utilize supported catalysts in fixed bed reactors, with the exception of the LC-Fining (Van Driesen and Caspers, 1979) and the H-Oil (Eccles, 1982) processes which are based on ebullating bed technology. Commercial heavy oil hydrocracking processes employing supported catalysts experience unacceptably high rates of catalyst deactivation (Oelderik et al, 1989). Four factors contribute in varying degrees to the deactivation: the deposition of clay / mineral matter, the formation of coke on the catalyst surface, the deposition of metal sulfides in the catalyst pores, and thermal sintering of the catalyst (Ternan, 1983). In general, the deposition of metals and coke are the most critical factors leading to catalyst deactivation (Oelderik et al, 1989). In many cases the high rate of catalyst deactivation and high consumption of catalyst adversely affect the economic feasibility of processes using supported hydrocracking catalysts. 7 2.2 Dispersed catalysts for heavy oil hydrocracking 2.2.1 Introduction The chronic deactivation of supported heavy oil hydrocracking catalysts has prompted the development of processes using unsupported catalysts (Del Bianco et ai, 1993). These 'dispersed phase' catalysts are extremely small metal sulfide particles (typically sub-micron) dispersed within the heavy oil. The dispersed catalysts cannot be used in conventional fixed or ebullating bed reactors due to the combined problems of high pressure drops and increased bed plugging. Consequently, a new reactor technology referred to as the 'slurry phase' reactor has been developed to utilize the extremely small catalyst particles. A slurry phase reactor is essentially a 3 phase fluidized bed with the catalyst completely fluidized by the upward motion of the gas/liquid (Dautzenberg and De Deken, 1984). The reactor differs from a conventional fluidized bed in that the catalyst passes through the reactor with the feed, and in most cases it is not recovered or recycled back to the reactor. A comparison between the basic characteristics of fixed, ebullating and slurry phase reactors is given in Table 2.1. Table 2.1: Comparison between heavy oil hydrocracking reactor technologies (Dautzenberg and De Deken, 1984). Fixed Bed Ebullating Bed Slurry Phase Vol % Catalyst 60 40 1 Catalyst Size, mm 1.5 x3 0.8x3 .002 Particles/cm3 120 250 2.4 x 109 Interp article Distance, mm — 1.6 0.008 The use of dispersed phase catalysts offers several distinct advantages over supported catalysts for heavy oil hydrocracking: 8 1. Firstly, since the catalyst has no pore structure, deactivation resulting from pore diffusion limitations and metals deposition is not a problem (Bearden and Aldridge, 1981). 2. Secondly, the extremely small catalyst particle size results in a considerable increase in specific surface area when compared to supported catalysts. Consequently, a much lower catalyst loading (<1000 ppm by weight) can be used, leading to decreased catalyst costs. 3. Since the catalyst is highly dispersed within the heavy oil (interparticle distances are very small), the ability of the catalyst particles to intercept and hydrogenate free radical intermediates is greatly increased. Consequently, dispersed catalysts are highly effective in suppressing coke formation (Del Bianco et al, 1993; Bearden and Aldridge, 1981). Proposed hydroconversion mechanisms in the presence of dispersed catalysts are discussed in section 2.2.2. 4. Aside from their catalytic role, the dispersed catalyst particles serve as nucleation sites for the small amounts of coke which do form. In this manner fouling of the reactor surfaces is prevented (Bearden and Aldridge, 1981). 5. There are several operational advantages associated with slurry reactors employing dispersed catalysts. Catalyst replenishment is simple, and the problems of reactor plugging inherent to fixed bed reactors are solved. In addition, large liquid-phase heat transfer coefficients make temperature control of slurry reactors fairly simple (Dautzenberg and De Deken, 1984). 2.2.2. Hydroconversion mechanisms in the presence of dispersed catalysts Several researchers have proposed conceptual hydroconversion mechanisms in the presence of dispersed catalysts (Dautzenberg and De Deken, 1984; Bearden and Aldridge, 1981). Such a proposed mechanism is shown in Figure 2.1. Heavy molecules such as asphaltenes and resins undergo a thermally induced free-radical cracking reaction to generate unstable 9 intermediates (reaction 2). The dispersed catalyst then hydrogenates the unstable radicals to yield oil (reaction 3), rather than allowing free radical coking reactions (e.g. polymerization) to occur via reaction 4. Several researches have used electron spin resonance (ESR) to demonstrate that dispersed catalysts derived from molybdenum naphthenate reduce the concentration of free radicals in heavy petroleum feed stocks (Rudnick, 1987; Varghese, 1986; Rudnick and Audeh, 1986). Consequently, the main catalytic effect of dispersed metal catalysts is to reduce the amount of coke formed during the hydrocracking reaction. Hydrogen Deficient large Molecules (asphaltenes) Catalyst Hydrogenated large molecules v(2) (2') Unstable Intermediates (Radicals, olefins etc.) Catalyst 'O i l ' Molecules (3) (4) Coke Figure 2.1 : Reaction mechanism in the presence of disperse catalysts (Dautzenberg and De Deken, 1984). Sanford et al. (1995) found that a similar mechanism also applies to supported heavy oil hydrocracking catalysts. The researchers found that supported metal catalysts lose their hydrotreating function very rapidly under hydrocracking conditions. In this context, hydrotreating refers to the reaction whereby an organic molecule (such as one containing an aromatic ring) is activated on one active site of the catalyst, whilst hydrogen is activated on another site, usually remote from the aromatic ring. The activated hydrogen then migrates to the activated aromatic and adds across a double bond, resulting in the saturation of that bond. Once the supported metal 1 0 catalysts lose their hydrotreating function, they essentially become a source of activated hydrogen, with a similar catalytic mechanism to that of dispersed catalysts. 2.2.3 The synthesis of dispersed metal catalysts A number of methods exist for the synthesis of sub-micron dispersed metal catalysts, including flame pyrolysis of organometallic compounds with ethylene (Andres et al, 1983; Hager et al, 1994), and low temperature vapor condensation (Bradley et al, 1991). However, these methods are complex and impractical for large scale industrial application. The most common synthesis method involves the direct addition of finely divided inorganic powders, water soluble or oil-soluble metal salts to the heavy oil feed (Del Bianco et al, 1993). These metal precursors include: • Inorganic metal compounds such as metal oxides, metal alloys, phosphomolybdic acid and ammonium tetrathiomolybdate (oil-insoluble). • Metal salts of organic amines and quaternary ammonium compounds (oil-insoluble). • Organometallic salts of organic acids such as naphfhenates and resinates (oil soluble). • Organomatallic compounds such as phthalocyanines, dithiocarbamates, dithiophosphates and alkyl molybdate (oil soluble). The metal constituents of these compounds are transition metals (Mo, Co, Ni, Fe, Cr etc.). In the case of oil-soluble metallic compounds, the active form of the catalyst is generated 'in-situ': the oil-soluble compound first dissolves within the oil and subsequently, under suitable thermal conditions, decomposes in the presence of H2S and other sulfur containing compounds in the heavy oil to yield metal sulfide catalyst particles (Del Bianco et ai, 1993). Researchers have found that the temperature profile (Utz et al, 1989; Lett et al, 1989), the method of addition (Liu et al, 1994), and the partial pressure of H2S (Derbyshire et al, 1986; Anderson and 11 Bockrath, 1984) can affect this decomposition reaction. However, these effects are highly complex and have not been comprehensively studied (Del Bianco et ai, 1993). 2.2.4 Effect of precursor solubility on catalyst activity A number of researchers have found that dispersed catalysts synthesized from oil soluble metal salts are more active than those resulting from oil insoluble compounds. Dabkowski and co-workers (1991) conducted hydrocracking experiments on Arabian heavy residues using dispersed catalysts derived from organometallic compounds of varying solubility. The researchers found that distillate yield, demetallization activity and desulfurization activity increased with increasing precursor solubility, whereas coke yield decreased with increasing solubility. Bearden and Aldridge (1981) conducted similar hydrocracking experiments on heavy crude oil using the M-COKE process, and found that catalysts from oil soluble Mo naphthenate and Mo(CO)6 gave higher desulfurization activity, higher CCR conversion and lower coke yields than oil-insoluble Mo0 3 and MoS2. Hirschon and Wilson (1989) found that dispersed catalysts derived from oil soluble Mo2(OAc)4 and Cp2Mo2(u-SH)2(u-S)2 gave higher conversion of coal to liquids than oil-insoluble tetrathiomolybdate impregnated from aqueous solution. 2.2.5 Size distribution of dispersed catalysts Researchers have studied the size distributions of dispersed metal catalysts derived from organometallic precursors using a number of different techniques. Bulk particle size was generally determined by microscopic techniques, including scanning electron microscopy (SEM) and transmission electron microscopy (TEM), whereas crystallite size was generally determined using x-ray diffraction (XRD) and Mossbauer Spectroscopy. Takemura and Okada (1988) investigated 12 the size of Ni crystallites resulting from the decomposition of nickel acetate in coal liquefaction feed. XRD analysis gave crystallite sizes ranging from 14.6 to 28.5 nm. The researchers also found that the Ni crystallite size increased with both decomposition temperature and nickel concentration. In addition the researchers found that the specific activity of the Ni catalyst increased linearly with crystallite diameter. Liu et al. (1994) investigated MoS2 synthesized in heavy oil from various oil-soluble and oil-insoluble Mo precursors. XRD analysis gave average crystallite sizes of 4-5 nm in the perpendicular direction, and 4-7 nm in the lateral direction, and the solubility of the metal salt did not have a dramatic effect on crystallite size. Bulk particle sizes (as determined by SEM and TEM) were found to be in the range of 1-2 um. Kim and co-workers characterized dispersed catalysts resulting from the decomposition of molybdenum naphthenate in coal liquefaction model systems (Kim et ai, 1989). XRD analysis gave crystallite sizes of 2.6 nm in the perpendicular direction and 4.5 nm in the lateral direction. The 'needle-shaped' crystallites displayed a so-called 'rag morphology' previously described by Pecoraro and Chianelli (1981). Bulk catalyst particle sizes were found to range from 50 to 250 nm, and the catalyst surface areas ranged from 150 to 200 m2/g. Herrick et al. (1990) investigated the decomposition products of Fe(CO)5 for coal/heavy oil co-processing. XRD analysis gave crystallite sizes ranging from 12-20.5 nm. Mossbauer spectroscopy gave crystallite sizes ranging from 5-20 nm. TEM analysis of the bulk catalyst gave particle sizes ranging from 100 to 1000 nm, and the particles displayed a granularity at the 10-50 nm scale. These studies show that oil soluble organometallics generally decompose to form very small crystallites in the 2-7 nm range. However, it would appear that the individual crystallites 13 aggregate to much larger bulk particles in the 50 - 3000 nm range. It is unclear whether this aggregation occurs under the reaction conditions, or later during cooling or sample preparation. 2.2.6. Active species of dispersed catalysts The active metal sulfide species of dispersed catalysts have been well characterized by a number of researchers using XRD, Mossbauer spectroscopy and Fourrier transform infra red (FTIR) spectroscopy. Liu et al. (1994), Kim et al. (1989), and Curtis and Pellegrino (1989) identified MoS2 as the active species formed from various molybdenum-based precursors. Anderson and Bockrath (1984) and Herrick et al. (1990) identified the non-stoichiometric sulfide Fei-xS as the active species formed from various iron-based organometallic compounds. 2.2.7. Disadvantages of dispersed catalysts derived from organometallic compounds Although organometallic precursors yield active catalysts for the hydrocracking of heavy oil, this synthesis technique suffers from a number of drawbacks. Firstly, there is little evidence in the literature to suggest that the catalyst particle size can be controlled when synthesized from these precursors. Hence, there is no information in the literature on the relationship between dispersed catalyst size and activity. In addition, the requirement for the organometallic precursors to be soluble in the heavy oil limits the choice of catalysts which can be synthesized from these precursors. Consequently, further understanding of the factors affecting the activity of dispersed catalysts is hampered by this apparent lack of control over particle size. Hence there is a need to find an alternate synthesis technique for dispersed catalysts in a hydrocarbon medium which allows for control over the catalyst particle size, and allows more flexibility with respect to the choice of metals which can be synthesized. 14 2.3 Preparation of metal colloids in reverse micelles 2.3.1 Introduction A novel technique for the preparation of metal colloids dispersed in organic solvents has been developed in recent years. Monodisperse suspensions of a wide range of metals and metal compounds have been prepared by the precipitation of water-soluble metal salts dissolved within the water pools of a microemulsion. This synthesis technique allows for the simple control of particle size, a factor which together with their narrow size distribution makes these colloids potentially attractive catalysts (Boutonnet et ai, 1991). 2.3.2 Emulsions and microemulsions An emulsion is generally denned as a mixture of two immiscible liquids wherein one of the liquids is broken into small particles and dispersed within the second liquid (Shaw, 1992). This produces a tremendous increase in the interfacial free energy of the system and makes emulsions thermodynamically unstable. This thermodynamic instability means that phase separation is usually rapid without the presence of an emulsifying agent to stabilize the system (Shaw, 1992). In nearly all emulsions, one of the phases is aqueous and the other a hydrocarbon or 'oil'. I f the oil is the dispersed phase, then the emulsion is termed 'oil-in-water' (OAV); and if the aqueous medium is the dispersed phase, then the emulsion is termed 'water-in-oil' (W/O). There are three types of emulsions based on the size of the dispersed particles (Rosen, 1978): (i) macroemulsions, opaque emulsions with particles larger than 400 nm; (ii) rniniemulsions, blue-white with particle sizes in the 100-400 nm range; (iii) microemulsions, transparent dispersions with particles smaller than 100 nm. A microemulsion is usually defined as an emulsion of water, oil and an amphiphile/surfactant which forms a single optically isotropic and thermodynamically stable liquid 15 solution (Danielsson and Lindman, 1981). The surfactant molecules adsorb to the interface between the immiscible liquids, and stabilize the dispersed phase by lowering the interfacial surface tension and free energy of the system (Shaw, 1992; Pillai et al, 1993). With the interfacial tension (yow) close to zero, microemulsions form spontaneously and are thermodynamically stable (Shaw, 1992). In addition, the surfactant layer decreases the rate of coalescence of the dispersed liquid particles by forming mechanical, steric and/or electrical barriers around them. Microemulsions are also classified as oil-in-water or water-in-oil, with surfactants self-assembled around droplets of the dispersed phase in either case. Surfactant-stabilized water pools in a hydrocarbon environment (i.e. O/W microemulsion) are often referred to as inverse or reverse micelles (Fendler, 1987). 2.3.3 Colloid formation in reverse micelles. Extremely small (<10 nm diameter) monodisperse metallic colloids can be prepared in-situ in the aqueous pools of W/O microemulsions (Fendler, 1987). A summary of examples cited in the literature is presented in Table 2.2. Water soluble metal salts are incorporated into the aqueous droplets of the microemulsion, and the metal/metal compound is then precipitated in the microheterogeneous environment by the addition of a suitable precipitation/reduction agent (Boutonnet et al, 1982; Martino et al, 1994). Confinement within the microenvironment of the reverse micelle limits particle growth (Fendler, 1987), since the surfactant interface provides a spatial constraint on the reaction volume (Martino et al, 1994). Following precipitation, the colloidal particles are stabilized against flocculation by the surfactant (Fendler, 1987; Henglein, 1989) and can remain dispersed in suspension for extended periods (Boutonnet et al, 1982). 16 Table 2.2 : Examples of metals and inorganic compounds prepared in reverse micelles. Authors Colloids Microemulsion System(s)1 Precipitation Agent Particle Size (nm) Boutonnet et al. (1991) Pt Water/PEGDE/Hexadecane N 2 H 4 3 ±0.5 Boutonnet et al. (1982) Pt, Rh, Pd, Ir Water/CTAB/Octanol Water/PEGDE/Hexane N.FL,/ H 2 Pt : 2.5-3 Rh: 3 Ir : 3 m&ietal. (1993) BaC03 Water/CTAB/Octane NH4CO3 5 - 15 Pileni and Lisiecki (1993) Cu Water/AOT/Alkane N 2 H 4 / NaBEU (in micelle) 3 - 28 Martino et al. (1991) Fe, Pd, FeS2 Water/DDAB/Toluene Water/CnEVOctane Water/Ph9E6/Cyclohexane LJBH4 Li2S Fe: 1.5±0.2 Pd: 1.8±0.2 FeS2: 3.110.1 Ravet et al. (1987) Ni2B, Co2B, Ni-Co-B Water/CTAB/n-Hexanol NaBFL 3 - 7 Nagy et al. (1983) Ni2B Water/CTAB/n-Hexanol NaBFLj 4 -7 Pileni et al. (1992) CdS Water/AOT/iso-Octane H 2S/ Na2S (in micelle) 2 -4 A number of different precipitation/reduction agents have been used to prepare colloids in reverse micelles, the choice of the particular reducing agent depending largely on the type of metal compound being synthesized. These agents can be divided into 3 broad categories: 1. Liquid reducing agents added directly to the microemulsions. Examples of liquid reducing agents reported in the literature include hydrazine hydrate N 2 H4 .xH 2 0 (Boutonnet et ai, 1982 and 1991), aqueous solutions of various metal borohydrides including LiBFL and NaBFL, 1 Note: PEGDE, Ci 2 E 4 and Ph 9E 6 represent the non-ionic surfactants pentaethyleneglycoldodecylether, butylethylene glycol n-dodecyl ether and polyoxyethylene(6) nonylphenol; CTAB, DDAB and AOT represent the ionic surfactants cetyltrimethylammonium bromide, didodecyldimethylammonium bromide and surfactant bis(2-ethylhexyl) sulfosuccinate respectively. 1 7 (Nagy et al, 1983; Ravet et al, 1984 and 1987; Martino et al, 1994) and aqueous solutions of lithium sulfide Li2S (Martino et al, 1994). 2. Gaseous precipitation agents added either in a static pressurized system or a flow system. Examples include H 2 (Boutonnet et al, 1982), H2S (Meyer et al, 1984; Lianos and Thomas, 1987; Petit et al, 1990; Pileni et al, 1992) and C0 2 (Kon-no et al, 1984; Kandori et al, 1986). 3. Reverse micelle entrapped reducing agents. The dispersed droplets in a microemulsion continuously collide, coalesce and decoalesce, resulting in the continuous exchange of solute content (Flecher et al, 1987). This exchange process can be utilized to react/precipitate different chemical species contained within different water pools. Two identical microemulsions containing the metal salt and the reducing agent dissolved within the water pools are mixed, resulting in the precipitation of solid colloids upon exchange of the reverse micelle contents. Examples of micelle entrapped reducing agents include NH4CO3 (Pillai et al, 1993), N 2H4.xH 20 (Pileni and Lisiecki, 1993) and Na2S (Petit and Pileni, 1988). Colloid formation in reverse micelles is a complex process involving interplay between nucleation, microcrystal formation, intermediate growth (Ostwald ripening), coagulation and flocculation, and no single mechanism has been agreed upon in the literature (Fendler, 1987). A number of researchers have proposed a simple statistical mechanism for crystal growth in reverse micelles (Ravet et al, 1987; Nagy et al, 1983). The proposed mechanism is based on a number of fundamental assumptions (Ravet et al, 1987): • Metal ions are statistically distributed through the water cores of a microemulsion according to a gaussian distribution (Void and Void, 1983). • A minimum number of metal ions are required to form a stable nucleus. • The nucleation step is always slower than the growth process. 18 A diagrammatic representation of the proposed mechanism for the formation of metal borides is given in Figure 2.2. | N a B H 4 / water Figure 2.2 : Proposed mechanisms for metal boride formation in reverse micelles (Ravet et ai, 1987; Nagy et ai, 1983). At the very beginning of the reduction process, nucleation occurs only in those water pools which contain more than the minimum number of metal ions required to form a stable nucleus. Following the initial nucleation step, additional ions are brought into contact with the existing nuclei through dynamic exchange between the water cores. These additional ions participate in the growth process, but no new nuclei are formed at this time since the nucleation process is much slower than particle growth. Consequently, the nuclei formed at the beginning of 19 the reduction grow at the same rate and ultimately form colloids of the same size (Ravet et al, 1987; Nagy et al, 1983). Cluster growth terminates when the colloids reach a critical size where the surface tension of the reverse micelle prevents further material entering and adding to the colloids (Steigerwald and Brus, 1989). According to this mechanism, the colloid size depends on a number of factors, including: the number of water cores containing enough ions to form stable nuclei, the rate of rearrangement of the system, and the ability of the reducing agent to diffuse into the solution and reach the water pools before system rearrangement. 2.3.4 Factors affecting the size of reverse micelles and colloids A number of factors affect the size of reverse micelles and the size of colloids synthesized therein. It is important to identify these factors since they allow one to manipulate the size (and therefore the surface area) of colloids synthesized using the reverse micelle technique. The most obvious factor which affects the size of reverse micelles is the water:surfactant ratio in the microemulsion (co). A number of researchers have found that the water pool size increases in a linear fashion with increasing co (Ravet et al, 1984; Kizling and Stenius, 1987; Nagy et al, 1983, Modes and Lianos, 1989; Pileni et al, 1992). This relationship is illustrated in Figure 2.3 (a) for the water/CTAB/n-Hexanol microemulsion. The observed trend has been found to apply to both ionic (Ravet et al, 1984) and non-ionic (Kizling and Stenius, 1987) surfactant systems. A number of systems have been well characterized; for example, Rw(A)=1.5*co for the water/AOT/Alkane system (Pileni et al, 1985), and Rw(A)=3.6+0.52*(% Water) for the water/CTAB/n-Hexanol system (Nagy et al, 1983). The trend of increasing water pool size with increasing co is relatively simple to explain. For a fixed amount of surfactant the interfacial area which can be stabilized by the surfactant is fixed. Consequently, new water pools cannot be 20 formed if co is increased, and the water pools must swell to accommodate extra water added to the system 0 4 8 12 16 0 0.2 0.4 0.6 0.8 Water (we ight %) [ N i " ] ( m o l a l / w a t e r ) Figure 2.3 : Variation of the average radius (rM) of the water cores in the water/CTAB/n-Hexanol microemulsion as a function of (a) water content and (b) Ni(II) concentration (Nagy et ai, 1982). Researchers have also found that the size of colloidal particles precipitated in reverse micelles increases with increasing CD , and this trend has been found to apply to all surfactant types and reducing agents (Ravet et ai, 1984; Nagy et al., 1983, Modes and Lianos, 1989; Pileni et al., 1992). This trend is illustrated in Figure 2.4. for Ni2B colloids synthesized in the water/CTAB/n-Hexanol microemulsion. Nagy et al. explained this trend in terms of their mechanism of colloid formation (discussed in section 2.3.3). Particle growth in a reverse micelle ceases when the particle reaches a critical size where the surface tension of the surfactant layer keeps the reverse micelle intact, thus preventing further material exchange between water pools. This implies that the critical nucleus size depends on the reverse micelle size, which in-turn depends on the water:surfactant ratio co (Steigerwald and Brus, 1989). Consequently, increasing co yields larger particles. 21 The second factor which affects reverse micelle size is the concentration of metal ions within the water pools. A number of researchers have found that water pool size increases with an increasing metal ion concentration, tending to a limiting water pool size at very high concentrations (Ravet et al., 1984; Nagy etal, 1983; Ravet et al., 1987). This trend is illustrated in Figure 2.3 (b). Ravet et al. explained this phenomenon for the case of Co(II) and Ni(H) ions in the ionic surfactant system water/CTAB/n-Hexanol (Ravet et ai, 1984). The authors found that one or two hexanol molecules participate in the co-ordination shells of tetrahedral Co(II) and Ni(II) complexes. Consequently, an increase in the concentration of metal ions in the water pool leads to an increase in number of hexanol molecules at the surfactant interface, and thus to an increase in the interfacial free energy of the system. An increase in the interfacial free energy of the system in turn leads to an increase in the reverse micelle size. The effect of metal ion concentration on the size of reverse micelles in microemulsions of non-ionic surfactants has not been directly reported in the literature. However, it is possible to deduce the effect of metal ion concentration in these systems by examining the effect of metal ions on the hydrophinc/hpophilic balance (HLB) of non-ionic polyoxyethylene (POE) based surfactants. Researchers have shown that different electrolytes affect the HLB of nonionic surfactants in different ways, depending on the nature of the electrolyte (Schick M . J., 1987). Schott found that nitrate salts of multivalent cations (e.g. Ni(N0 3) 2 , L1NO3, Mg(N03)2) form stable complexes with oxygen atoms in the POE chains of nonionic POE based surfactants, as shown below: R i R i 2+ Me 2 + + n O ^ -> [Me [ o ^ ]„ ] R 2 R 2 2 2 Schott found that this complex formation resulted in a so-called 'salting in' effect, raising the cloud point and HLB of the nonionic surfactant. The increase in I I LB of the surfactant was found to be proportional to the metal ion concentration in the system (Schott, 1973). The effect of the HLB on the size of reverse micelles of nonionic surfactants can be deduced by examining the effect of temperature on the HLB and reverse micelle size. A number of researchers have found that the HLB of POE based non-ionic surfactants decreases with increasing temperature (Schick, 1987; Shaw, 1992). Kizling and Stenius found that increasing temperature (decreasing HLB) leads to a decrease in the size of the reverse micelles in the nonionic surfactant system water/PEGDE/hexadecane (Kizling and Stenius, 1987). These two experimental observations imply that a decrease in HLB results in a decrease in reverse micelle size, and an increase in HLB leads to an increase in reverse micelle size. This in turn implies that increasing the concentration of certain metal ions (e.g. Ni(NC«3)2) within reverse micelles of nonionic surfactants will lead to an increase in the HLB of the surfactant molecules at the interface, thus leading to an increase in the size of the reverse micelles. The concentration of metal ions dissolved within reverse micelles also has an effect on the size of the colloidal particles precipitated within the reverse micelles. Ravet et al. (1984) and Nagy et al. (1983) studied the precipitation of Co2B and Ni2B particles in the water/CTAB/n-Hexanol system and found that a complex relationship exists between particle size and metal ion concentration (see Figure 2.4). 23 Figure 2.4 : Variation of the average diameter (d in A) of the nickel boride particles prepared in water/CTAB/n-Hexanol as a function of water content and Ni(II) concentration (Nagy etal, 1982). As illustrated in Figure 2.4, a minimum particle size was found, with particle size increasing with increasing ion concentration at high metal concentrations. Ravet et al explain this phenomena using the statistical mechanism for colloid formation in reverse micelles discussed in section 2.3.3. The size of particles formed in reverse micelles is inversely proportional to the number of nucleation sites formed at the beginning of the reaction. At low ion concentration, only a few water cores contain the minimum number of ions necessary to form a nucleus. Hence, few nuclei are formed at the very beginning of the reduction and the size of the particles formed is quite large. The researchers found that as ion concentration increases, the distribution of precursor ions in the reverse micelles changes, and the number of nuclei obtained by reduction increases faster than the total number of ions in the system. This results in a decrease in the particle size to a minimum. When more than 80% of the water cores have enough ions to form 24 stable nuclei, the number of nuclei formed remains quasi constant with increasing ion concentration. Hence, the size of the particles formed increases again. A number of other factors also affect the size of reverse micelles and the colloids synthesized therein. Kizling and Stenius (1987) found that the size of the reverse micelles in the non-ionic water/PEGDE/Hexadecane system decreased with increasing temperature between 22 and 38 °C. Further increases in temperature lead abruptly to phase separation. The observed trend was probably linked to changes in the hydrophobic/Upophobic balance (HLB) experienced by non-ionic surfactants at elevated temperatures (Overbeek et al., 1984). A number of researchers have also found that the amount of reducing agent affects the size of the colloids formed (Nagy et ai, 1983; Pileni et al, 1992). In general, the researchers found that larger particles were formed when the cation/anion ratio was « 1, and that the particle size decreased when the reducing agent was in excess. Finally, the type of micellar system used has an effect on the size of the solid particles precipitated in reverse micelles. The viscosity of the hydrocarbon continuous phase and the characteristics of the surfactant (e.g. hydrophobic/lipophobic balance) affect the morphology of the surfactant interface, the rate of exchange of reactants between water pools, and hence the size of the particles formed (Steigerwald and Brus, 1989). 2.3.5 Catalytic activity of colloids synthesized in reverse micelles The small size and monodisperse nature of dispersed particles prepared in reverse micelles make them attractive as potential catalysts. A number of researchers have investigated the use of colloidal metal/metal compounds synthesized in reverse micelles as catalysts for a number of hydrogenation and hydrogenolysis reactions. Nagy et al. (1983) and Ravet et al. (1984) investigated the hydrogenation of 1-heptene using Ni2B and Co2B colloids prepared in the water/CTAB/n-Hexanol microemulsion. The 25 reactions were performed in-situ in a mixture ethanol and a microemulsion containing the metal boride colloids. The researchers obtained quite different results for the different catalysts. The activity of the Ni2B colloids was found to be greater than Raney catalysts prepared in ethanol by 'conventional' methods. The presence of surfactant in the reaction system did not appear to affect the activity of the Ni2B catalyst (Nagy et al, 1983). However, the researchers found that the CTAB surfactant had a depleting effect on the catalytic activity of Co2B colloids synthesized in the same way. In this case 2 nm Co2B colloids showed lower activity for the hydrogenation of 1-heptene than 250 nm Co2B prepared in ethanol (Ravet et al, 1984). Boutonnet et al. investigated the hydrogenation and isomerization of 1-butene using platinum colloids in a microemulsion, and microemulsion-generated platinum supported on pumice (Boutonnet et al, 1986; Boutonnet et al, 1987). The researchers found that the deposition of microemulsion-generated Pt particles on a solid support provided an excellent catalyst for the hydrogenation and isomerization of 1-butene. However, considerably less catalytic activity was observed when the same reaction was performed in the microemulsion phase. The authors claim that the surfactant coating hindered the accessibility of 1-butene and hydrogen to the catalyst surface. Martino et al. (1994) investigated the catalytic activity of various metals prepared in microemulsions in three separate reactions: (i) hydrogenolysis of naphthylbibenzylmethane (a coal liquefaction model compound), (ii) hydropyrolysis of coaL and (iii) coal liquefaction. The researchers found that the presence of surfactant pyrolysis byproducts in reaction (i) resulted in a loss in catalyst activity. They attributed this loss in activity to the scavenging of hydrogen by the surfactant byproducts, as well as possible chemical and steric poisoning by the surfactant. The effect of surfactant was found to be less pronounced in reactions (ii) and (iii), and colloidal Pd 26 prepared in the water/DDAB/Toluene system was as active as commercial MoS2 for the hydropyrolysis of coal. In conclusion, the catalytic activity of colloidal metal particles synthesized in reverse micelles appears to be highly dependent on the particular reaction, type of metal and type of surfactant used, especially if the reaction is performed insitu in the microemulsion suspension. Evidence in the literature suggests that the presence of surfactant either depletes or has no effect on the catalytic effect of colloidal catalysts in reverse micelles, depending on the particular system studied. 2.3.6 Preparation of dispersed hydrocracking catalysts in reverse micelles The application of metal colloids prepared in reverse micelles as dispersed heavy oil hydrocracking catalysts holds a number of potential advantages over dispersed catalysts prepared from organometallics. Firstly, a number of researchers have demonstrated that the size of solid colloids prepared in reverse micelles can be controlled and manipulated, whereas there is little evidence in the literature to suggest that the size of catalyst particles synthesized from organometallics can be controlled. Consequently, the relationship between dispersed catalyst size and activity may potentially be investigated using dispersed catalysts prepared in reverse micelles. Secondly, the range of catalysts which can be prepared from organometallics is limited by the solubility of the organometallics in the heavy oil. No such restriction applies to colloidal catalysts prepared in microemulsions, and it is believed that these catalysts can be well dispersed or mixed within the heavy oil feed, especially if the organic component of the microemulsion is highly soluble in the heavy oil. 27 Chapter 3 : Experimental M e t h o d s 3.1 Mcroemulsion preparation and colloid synthesis 3.1.1 Microemulsion preparation The present work focused on the water-in-oil microemulsion water/polyoxyethylene-4-laurylether/n-hexane. Polyoxyethylene-4-laurylether (Sigma Chemical Company) and n-hexane (Aldrich Chemical Company, HPLC Grade) were used as received. Microemulsions were prepared with either cobalt nitrate (Co(N03)2.6H20, Sigma Chemical Company, 99.4%), nickel nitrate (Ni(N03)2.6H20, Aldrich Chemical Company) or ferric nitrate (Fe(NO)3.9H20, Fisher Scientific, 99.0 %) dissolved within the water pools of the microemulsions. Both the water: surfactant ratio (co) and the metal ion concentration in microemulsions were varied for each of the metals investigated. A full list of the microemulsions prepared is presented in Appendix 1.1. A number of water/polyoxyethylene-4-laurylether/decahydronaphthalene (Decahydronaphthalene, Sigma Chemical Company, 98%) microemulsions were also prepared for a prehminary series of activity measurements. This issue is discussed further in section 3.4.3. In a typical preparation, specified volumes of polyoxyethylene-4-laurylether (PE4LE) and hexane were pre-mixed. The appropriate volume of a metal nitrate solution of specified concentration was added to the PE4LE/hexane mixture. The system was equilibrated by mixing with a magnetic stirrer for a period of 12 hours. Metal salt solutions of varying concentration were prepared from stock solutions of known metal ion concentration. The stock solutions were assayed using atomic absorption by ACME Analytical Laboratories, Vancouver. The preparation procedure allowed for the precise control of the water: surfactant ratio in the microemulsion and metal ion concentration within the water pools of the microemulsion. After preparation, the microemulsions were stored in sealed bottles in a refrigerator at 3-4 °C. 3.1.2 Synthesis of reduced metal colloids Reduced nickel, cobalt and iron colloids were prepared in water/PE4LE/hexane microemulsions following the procedure used by Boutonnet et al. (1982). A typical preparation procedure involved adding hydrazine hydrate (N2FI4.XH2O, Aldrich Chemical Company, 55 wt% N 2 Ht) drop-wise to a microemulsion containing a metal salt dissolved within the water pools. A 5:1 molar ratio of [N 2FLi]: [Metal] was used to ensure the complete reduction of the metal species in the microemulsion (Martino et al, 1994). In addition, the size of colloids precipitated in reverse micelles is minimized when the reducing agent is in excess (Nagy et al., 1983; Pileni et al., 1992). The microemulsions were stored in sealed bottles following reduction. 3.1.3 Synthesis of metal sulfide colloids using dimethyl disulfide The high temperature synthesis of metal sulfide colloids in the water/PE4LE/hexane microemulsion using dimethyl disulfide (DMDS) was investigated in the early part of this study. These experiments were performed in a 300 mL Autoclave Engineers EZE-SEAL batch reactor, heated by a jacket furnace and temperature controller. The reactor contents were mixed by a magnetically driven impeller with an accompanying speed controller. A schematic diagram of the batch reactor system is given in Figure 3.1. A typical synthesis experiment involved adding dimethyl disulfide ((CH3)2S2, Aldrich Chemical Company, 98 %) equivalent to 3 wt% sulfur to 100 mL of a water/PE4LE/hexane microemulsion containing either nickel, cobalt or iron nitrate dissolved within the waterpools. The mixture was then placed in the reactor, the reactor was purged with N 2 (UHP, Medigas) and pressurized to 790 kPa (100 psig) with H 2 (UHP, 29 Medigas). The reactor was then heated to 210 °C at a temperature ramp rate of 1.5 °C/min, and held at this temperature for 2 hours. At the end of the reaction time, the jacket furnace was removed and the reactor was quenched using the internal water-cooled cooling coil. N, 5% H2S in H, H, To Vent Impeller 1 Jacket Furnace l A A A A l Cooling Coil Figure 3.1: Flow diagram of batch autoclave system used for colloid synthesis and activity measurements. 3.1.4 Synthesis of metal sulfide colloids using H2S Nickel, cobalt and iron sulfide colloids were prepared in water/PE4LE/hexane microemulsions using H2S at room temperature. In a typical experiment, 100 mL of a water/PE4LE/hexane microemulsion containing either nickeL cobalt or iron nitrate (dissolved within the waterpools) was placed in the reactor and the reactor was purged with nitrogen. The reactor was then pressurized to 7 MPa (1000 psig) with a 5% rnixture of H2S in H 2 (Medigas) and stirred at ± 500 rpm. The reaction proceeded for a specified reaction time of either 2 or 6 hours. At the end of the reaction, the reactor was vented and the metal sulfide suspension was removed. 30 3.2 Micelle characterization by dynamic light scattering (DLS) A number of techniques have been used to measure the size of reverse micelles in microemulsions, including dynamic light scattering (Boutonnet et al, 1982; Kizling and Stenius, 1987; Wilcoxon et al, 1993), small angle neutron scattering (Robinson et al, 1984), small angle x-ray scattering (Pileni et al, 1985) and 19F-NMR spectroscopy of labeled probe molecules in the micellar solution (Ravet et al, 1984; Ravet et al, 1985). In the present study, dynamic light scattering (DLS) was used to determine the size of the reverse micelles in the water/PE4LE/hexane microemulsion. The DLS measurements and analysis were performed by Dr. Barbara Frisken in the Department of Physics, Simon Fraser University. The DLS equipment consisted of a He/Ne laser (k = 633 nm) as a light source, a cell holder and an ALV 'Multi-u' autocorrelator. The microemulsion samples were diluted in hexane and placed in 20 mm vials in the cell holder. The method used by Kizling and Stenius (1987) and Chang and Kaler (1986) was followed to determine the size of the reverse micelles in solution. The normalized intensity correlation function g ( 2 )(i) was analyzed by the method of cumulants, and g(2)(x) was fit with a simple monomodal expression: g ( 2 ) (x) = B + Aexp(-2FT + u-c2) Where, T = delay time (s) A = Constant (depending on the geometry of the light scattering apparatus) B « 1 r = Average decay rate (1/s) T and p, were determined from the exponential regression, and the apparent diffusion coefficient of the scattering units (Da p p) was calculated using the equation: 31 Where, the scattering wave vector q is defined as follows: 4rcn 0 q = ~ S m ( - ) Here, 0 = Scattering angle (90°) A, = Wavelength of the scattered light = 633 nm Chang and Kaler (1986) have shown that the apparent diffusion coefficient D a p p ~ D 0 (intrinsic diffusion coefficient) for this type of system. Consequently, the hydrodynamic radii (rH) of the reverse micelles in the microemulsion were calculated directly from the Stokes-Einstein equation: kT rH = 6 ™ l D a p p Where, r\ = Solvent viscosity The parameter gives a measure of the relative variance of the monomodal distribution, a value of 0.05 or less indicating an acceptable degree of monodispersity The size of the reverse micelles in each microemulsion was measured 3 times and the average diameter was calculated for each sample. The standard deviation in the measured diameter was also calculated in each case. This standard deviation gave an estimate of the instrumentational error in the measurement of the reverse micelle diameters. 3.3 Colloid characterization 3.3.1 Dynamic light scattering (DLS) The particle sizes of the reduced metal colloids synthesized within reverse micelles were determined by dynamic light scattering (DLS). The experimental procedure was the same as that used to measure the size of the reverse micelles (described in section 3.2). However, the DLS 32 data were analyzed in a slightly different way. A multi-modal particle size distribution was assumed in the regression of the normalized intensity correlation function. Consequently, a distribution of particle sizes could be detected and characterized using DLS. 3.3.2 Transmission electron microscopy (TEM) The particle sizes of the reduced metal and metal sulfide colloids synthesized within reverse micelles were also determined by transmission electron microscopy (TEM). A Hitachi H-800 electron microscope operating in the transmission mode at 100 kV was used. TEM samples were prepared following the method used by Boutonnet et al. (1982) and Wilcoxon et al. (1993). Samples were deposited directly onto carbon coated Mo, W or Au grids by the application of 1-2 droplets (2-5 uL) of the conoid-containing microemulsion. The grids were then either examined as prepared, or dried in a vacuum oven at 120 °C and 2 torr for a period of 12 hours. The grids were examined using the electron microscope, and an attempt was made to examine and photograph many representative portions of the grid in order to determine the average particle size. The approximate average particle sizes were obtained from a histogram of the number vs. size of all the particles photographed. 3.3.3 Energy dispersive x-ray spectrometry (EDX) Elemental analysis of the particles observed by TEM was performed using an Ortec EEDS II energy disperse X-Ray spectrometer operating in tandem with the Hitachi H-800 electron microscope. The electron microscope was operated in the scanning transmission mode (STEM) for the EDX analysis. 33 3.3.4 X-Ray diffraction (XRD) Phase analysis of the metal sulfide colloids synthesized in the reverse micelles was performed by powder x-ray diffraction. A Siemens D5000 powder diffractometer using Cu Ka radiation (X = 1.5406 A) at 40 kV and 30 mA was used, and the analysis was performed at room temperature (22 °C). The metal sulfide colloids were extracted from the microemulsions following the method described by Boutonnet et al. (1991): Twice excess tetrahydrofuran (THF, Aldrich Chemical Company, HPLC grade) was first added to the microemulsions containing the metal sulfide colloids, causing the colloids to aggregate and settle from suspension over a period of approximately 12 hours. The colloid-free supernatant was decanted off and the settled colloids were washed 3-4 times with fresh THF to remove excess surfactant. The washed colloids were then dried in a vacuum oven at 22 °C and 2 torr for 12 hours. The dried samples were finally ground into a fine powder and mounted on an acrylic sample holder ready for XRD analysis. The metal sulfide species present were identified by comparing the XRD spectra obtained with standard spectra of various metal sulfide compounds. Siemens peak matching and identification software was also used to aid in the identification of the metal species present. The spent catalysts from the model compound activity tests were also analyzed using XRD, in order to identify the metal species present and determine the average crystallite sizes. The spent catalyst was allowed to settle in the liquid reactor product, and was then removed from the liquid using a pipette. The concentrated catalyst 'slurry' was then deposited drop-wise on a glass slide. The first stage of the sample preparation was performed under a nitrogen blanket in a glove box to minimize exposure to the ah. The samples were then transferred to a vacuum oven and dried at 22 °C and 2 torr for 12 hours to evaporate the liquid hydrocarbons. The average crystallite sizes were calculated from the peak width at half-height using the Debye-Scherrer equation: X ~ B.Cos© Where, B = Vw2 - w 2 t = Crystallite size (A) X = Wavelength (1.5406 A) 0 = Peak position (radians) W = Peak width at half height (radians) w = Instrumental peak broadening (radians) = 0.2 degrees 3.3.5 X-Ray photoelectron spectroscopy (XPS) The metal sulfide colloids synthesized in the reverse micelles were also analyzed using x-ray photoelectron spectroscopy (XPS). XPS allows one to identify the chemical species on the surface layer (several atoms thick) of a sample (Delgass et al., 1970). The XPS measurements were made by Dr. Philip Wong, in the Department of Chemistry, UBC. The XPS spectra were measured using a Leybold MAX-200 spectrometer with computer controlled data acquisition. Unmonochromatized Al Ka x-rays (hv = 1486.6 eV) were used. The system pressure was 1.28* 10"8 mbar during the data collection. The XPS system was calibrated using the C Is line of air-borne carbon (binding energy = 285.0 eV). Charging effects were also measured by the shift in the C Is line, and were automatically corrected for by the spectrometer software. Two samples were prepared for each of the metal sulfides analyzed by XPS. The first set was prepared and transported under a helium (UHP, Medigas) blanket in an attempt to prevent exposure to atmospheric oxygen. The samples were exposed to the air briefly when they were 35 removed from the inert atmosphere of the sample bottle and placed in the XPS sample holder. A typical sample preparation procedure involved depositing 100-200 droplets (0.5 to 1 mL) of the conoid-containing microemulsion onto a 1 cm2 square of glass fiber filter paper (glass micro-fiber filter paper, Watmann Company). The colloid impregnated square was then transported under a flowing He blanket to a vacuum oven and dried at room temperature (22 °C) and 2 torr for 12 hours. The vacuum oven was flushed with helium prior to evacuation to ensure an inert drying environment. After 12 hours, the dried sample was transported in an air-tight glass bottle to the XPS laboratory. The sample was pre-evacuated overnight at 10"4 torr to remove any rermining volatile matter adsorbed on the colloids, and then transferred to the spectrometer for analysis. The second set of samples was prepared using the same procedure, but the samples were exposed to the air during preparation. 3.3.6 BET surface area measurements The BET (Brunauer-Emmett Teller) surface area of a number of metal sulfide colloids synthesized in reverse micelles was determined by single point nitrogen adsorption/desorption using a Flowsorb 2300 analyzer. The metal sulfide colloids were recovered from the microemulsion using the method outlined in section 3.3.4. A pre-weighed sample of the recovered catalyst was then placed in a quartz sample tube and degassed under flowing N 2 at 130 °C for 20 minutes. The sample tube was then connected to the Flowsorb unit for the surface area measurements. The surface area of the sample was determined by the adsorption of N 2 from a 70:30 volumetric mixture of He and N 2 at -195 °C, and then by the subsequent desorption of the adsorbed N 2 monolayer at room temperature. The specific surface area of the sample was calculated by dividing the average of the adsorption and desorption surface area values by the weight of the sample. 36 3.4 Activity measurements 3.4.1 Reaction system A series of catalyst activity measurements was conducted to determine the catalytic activity of nickel, cobalt and iron sulfide colloids synthesized within the water/PEEL/hexane microemulsion. The hydrocracking of a model compound, diphenylmethane (DPM), was chosen as a model reaction for this purpose. A number of researchers have used this model reaction to determine the relative catalytic activity of various hydrocracking catalysts (Wei et al, 1992; Stenberg et al., 1983; Sweeny et al, 1987). These studies have shown that metal sulfides promote the cleavage of the alkyl-aromatic carbon-carbon bond in DPM, producing benzene and toluene as the reaction products: This reaction is attractive as a model compound reaction for a number of reasons. Firstly, the product distribution obtained allows one to differentiate between the hydrocracking and hydrogenation activity of the catalyst being tested. Secondly, the reactant and product molecules are well defined, and the concentration of these species can be easily determined by gas chromatography (GC). 3.4.2 Modeling of the reaction system Previous studies on the hydrocracking of DPM using metal sulfide catalysts have been conducted at temperatures ranging from 400 °C to 450 °C, and reaction pressures ranging from 10.4 to >20.8 MPa (Wei et ai, 1992; Stenberg et al., 1983; Sweeny et al, 1987). Given the physical limitations of the batch reactor used for the hydrocracking experiments (P < 20.8 MPa at 426 °C), it was important to determine the pressure and phase equihbrium of the reaction mixture 37 at the final reaction conditions. The ASPEN process simulation package version 5.5 (ASPEN Technology Inc.) was used for these calculations. A detailed description of the ASPEN simulation is given in Appendix 1.3. 3.4.3 Experimental apparatus and methods The activity measurements were conducted in the batch autoclave reactor described in section 3.1.3. Two shghtly different experimental methods were used for the activity measurements. A preliminary set of activity measurements was conducted in situ in water/PE4LE/decahydronaphthalene microemulsions containing metal sulfide catalyst colloids. A typical experiment involved placing 135 mL of a water/PE4LE/decalin microemulsion with a particular metal salt dissolved within the reverse micelles of the microemulsion into the reactor. The reactor was then pressurized with 5% H2S in H 2 at room temperature (22 °C) and 7 MPa for 2 hours to synthesize the metal sulfide catalyst colloids within the reverse micelles. Next, the reactor was vented, and 15 mL (15.015 g) of DPM (Aldrich Chemical Company, 99%) was added to the suspension of metal sulfide colloids in the microemulsion. The reactor was then repressurized to 2.2 MPa (300 psig) with 5% H2S in H 2 and heated to 430 °C at a temperature ramp rate of 5 °C/min (Autoclave Engineers recommend a temperature ramp rate of less than 7°C/min to avoid exposing the reactor vessel to excessive thermal shock). The reactor temperature was maintained at 430 °C for 2 hours. After 2 hours, the furnace was removed and the reactor was air-cooled until the temperature dropped to 250 °C (about 15 min). The reactor was then quenched using the internal water cooled cooling coil. The reactor vessel and reactor internals (impeller, cooling coil and thermocouple housing) were carefully cleaned between successive activity measurements in an attempt to remove any 38 catalyst adhering to the metal surfaces. The reactor vessel and internals were first cleaned with soap, water and steel wool to remove any solid deposits on the metal surfaces. The surfaces were then further cleaned using Brasso®, a petroleum based abrasive metal polish. Finally, the reactor and internals were rinsed with hexane and dried overnight at room temperature. A second set of activity measurements was conducted using a slightly different experimental procedure. Metal sulfide colloids were synthesized in a water/PE4LE/hexane microemulsion following the procedure outlined in section 3.1.4, and then extracted from the microemulsion and dried following the method outlined in section 3.3.4. The extracted metal sulfide colloids were then used as the catalyst in the DPM hydrocracking reaction. A typical experiment involved first mixing 120 mL of Decalin (Sigma Chemical Company, 98%) and 15.015g of Diphenylmethane. A 5 mL sample of the mixed reactor feed was then removed for analysis by gas chromatography (GC). A specified amount of the extracted catalyst powder was added to the rermining 120 mL of decalin/DPM, and this mixture was placed in the reactor. Presulfiding with 5% H2S in H 2 at room temperature (22 °C) and 7 MPa for 2 hours followed. The pressure was then reduced to ± 2.4 MPa (340 psig) and the reactor was heated to 400 °C at a temperature ramp rate of 5 °C/min using the electric jacket furnace. The reaction temperature was maintained at 400 °C for 3 hours. After 3 hours, the furnace was removed and the reactor was cooled following the procedure described above. As with the first set of activity measurements, the reactor and internals were cleaned between successive runs to remove residual catalyst. Activity measurements were also conducted with dispersed metal catalysts synthesized from organometallic precursors such as iron pentacarbonyl (Fe(CO)5, Aldrich Chemical Company) and Co naphthenate (Sigma Chemical Company, 8 wt% Co). The same procedure was used as that described above for the second set of activity experiments, with the organometallic 39 precursors replacing the colloidal metal sulfide catalyst. A full list of all the activity experiments conducted is presented in Appendix 1.2. 3.4.4 Analysis of the activity test products using gas chromatography. The reaction products from the activity measurements were identified by gas chromatography (GC). A Varian 3400-CX gas chromatograph with a flame ionization detector (FID) was used. The GC was equipped with a 3.2 mm x 2.1 m column (Chromatographic Specialties Inc.) packed with 5% OV-17 on Chromosorb WAW-DMCS. The temperature program used is given in Appendix 1.4. Helium (HHP, Medigas) was used as a carrier gas at a flowrate of 30 cm3/min. The GC analyses was performed by the manual injection of 1 ul of liquid reactor product into the column using a 10 ul mho-syringe. The micro-syringe was rinsed thoroughly with decalin between injections. Each sample was injected three times, and the average peak area of the three injections was used to calculate the concentration of each component in the sample. n-Decane (Fisher Scientific, 99.8%) was used as an internal standard to quantify the amounts of DPM, benzene and toluene in the reactor product. 3.611 mL of n-decane (approximately equivalent to a 36:1 volumetric ratio of decane:liquid product) was added to the liquid reaction products before the reactor was unloaded. The resulting mixture was then unloaded into a septum-capped glass bottle and stored at 4 °C in a refrigerator. A number of calibration curves were generated in order to relate the concentrations of benzene, toluene and DPM in the liquid reactor product to the amount of internal standard added to the sample. Standard calibration mixtures of n-decane (Fisher Scientific, 99.8%), DPM (Aldrich chemical company, 99%), decalin (Sigma chemical company, 98%), benzene (Fisher Scientific, A.C.S. reagent grade) and toluene (Fisher Scientific, A.C.S. reagent grade) were 40 prepared in proportions similar to those expected in the reactor product. In total, eight calibration standards were prepared corresponding to DPM conversions ranging from 0% to 20 %. A table of the compositions of the calibration standards is given in Appendix 1.5. Each calibration mixture was injected into the GC three times, and the average peak area of each of the components present was determined in each case. The calibration curve for each component n was generated by plotting: Area of Component n Peak Moles of Component n in Mixture v s Area of Decane Peak Moles of Decane in Mixture Two sets of calibration curves were prepared in order to determine the effect of continued use on the calibration of the column. The first set was prepared before the activity measurements were started, and the other set was prepared mid-way through the experimental work. The raw data and both sets of calibration curves are presented in Appendix 1.6. A least squares regression was performed on the data to determine the calibration equations from the data. The calibration equations were calculated in the following form: A: M: a* — + C A M ^•cio 1 T A C 1 0 Where, A = Area of component i Acio = Area of the n-decane internal standard Mj = # of moles of component i in the reactor product Mcio = # of moles of n-decane added to the reactor product The calibration equations obtained are presented in Table 3.1. The benzene and toluene calibration curves were both linear, with negligible scatter in the data points (R2 values of 0.998 and 0.999 respectively). The results of the recahbration indicate that the calibration of the GC with respect to benzene and toluene did not change with time. The 41 DPM calibration curves show more scatter than those of benzene and toluene, and the R2 value of the linear fit to the data was lower (R2 = 0.89). The lower apparent accuracy of the DPM calibration curves may be attributed to saturation of the FID detector, due to the high relative concentration of DPM in the GC injections. The recalibration results for DPM display the same scatter as the initial calibration. 3.4.5 Calculation of conversion The conversion for each activity measurement was calculated based on: i) the amount of DPM consumed in the reaction, ii) the amount of benzene produced, and hi) the amount of toluene produced. The equations used to calculate the conversions are given in Appendix 1.7. For each experiment, the number moles of DPM in the reactor feed was determined by GC analysis of the 5 mL sample taken from the reactor feed. The amounts of DPM, benzene and toluene in the reactor product were determined by GC analysis following the method described in section 3.4.4. Table 3.1: Calibration curves used to calculate the number of moles of benzene, toluene and DPM in the reactor product. Component a c R2 Benzene 0.5931 -0.0032 0.9984 Toluene 0.6971 -0.0057 0.9997 DPM 1.4872 -0.8807 0.8941 42 Chapter 4 : Catalyst Preparation and Characterization 4.1 Microemulsion preparation 4.1.1 Introduction A key advantage of preparing dispersed hydrocracking catalysts in reverse micelles is the proposed link between the size of the reverse micelles and the size of the colloidal catalyst synthesized therein. Consequently, it was important to first establish the factors which influence the size of the reverse micelles in the water/polyoxyethylene-4-lauryl ether/hexane system, as these factors may ultimately be used to exercise control over the catalyst particle size. Little information has been published in the literature on the water/polyoxyethylene-4-lauryl ether/hexane system. Polyoxyethylene-4-lauryl ether (or PE4LE) is a non-ionic surfactant with the chemical formula Ci2H25(OC2H4)4OH. The aliphatic C i 2 H 2 5 group is hydrophobic and solubilizes in the hexane phase, whilst the hydrophihc (OC2H4)4OH group solubilizes within the aqueous phase of the microemulsion. Some work has been published in the literature on the water/PEGDE/hexane system (Boutonnet et al, 1982), and the water/PEGDE/hexadecane system (Kizling and Stenius, 1987). Given the structural similarities between PE4LE (CnFL^OC^^OH) and PEGDE (Ci2H2 5(OC 2H4)5OH), it was expected that the water/PE4LE/hexane system would show similar characteristics to the water/PEGDE/hexane system. 4.1.2 Factors affecting the size of reverse micelles in the water/PE4LE/hexane system Previous studies on the preparation of microemulsions have shown that 3 key factors affect the size of the reverse micelles within the microemulsion, namely: the water: surfactant ratio 43 (co), the concentration of metal ions within the reverse micelle, and the type of metal dissolved within the reverse micelle (see section 2.3.4). Consequently, the effect of these 3 factors on the size of the reverse micelles in the water/PE4LE/hexane system was investigated. A number of microemulsions of varying co and metal ion concentration were prepared, with Ni, Fe or Co nitrates dissolved within the water pools. Ni, Co and Fe were chosen since combinations of these metals are commonly used as heavy oil hydrocracking catalysts. The hydrodynamic diameters of the reverse micelles were determined by dynamic light scattering (DLS). The DLS results are presented in Table 4.1. The average hydrodynamic diameter and the standard deviation in the measured diameter are quoted for each microemulsion (see section 3.2 for a more detailed description of these parameters). The parameter u/T2 gives an indication of the relative degree of polydispersity of the reverse micelles, a value of 0.05 or less indicating an acceptable degree of monodispersity. The hydrodynamic diameters of the reverse micelles were similar to those reported by Kizling and Stenius for the water/PEGDE/dodecane system (Kizling and Stenius, 1987). The researchers reported reverse micelle hydrodynamic diameters ranging from 12-18 nm (at 22 °C) for microemulsions containing 4-7 weight % water. In general, the reverse micelles in the water/PE4LE/hexane system displayed a fair degree of monodispersity. u/T2 (a measure of the variance in the diameters) was generally less than 0.1 for the microemulsions investigated, indicating a fairly narrow distribution of reverse micelle sizes. A number of water/PE4LE/decahydronaphthalene (DHN) microemulsions were also prepared for the preliminary set of activity measurements (see section 3.4.3). The hydrodynamic diameters of the reverse micelles of this system were determined by dynamic light scattering, and the results are presented in Table 4.2. The diameters of the reverse micelles in the 44 water/PE4LE/decahydronaphthalene microemulsions were slightly larger than reverse micelles of identical co and metal ion concentration in the water/PE4LE/hexane system. These slight differences in reverse micelle size were to he expected due to the differences in solubility of the surfactant in hexane and decahydronaphthalene. Table 4.1 : Hydrodynamic diameters of reverse micelles in the water/PE4LE/hexane system. # Metal dissolved CO Metal ion cone, in Avg. hydrodynamic u/T2 in reverse (vol/vol) reverse micelle diameter of rev. (Relative micelle (mol/L) micelle (nm) variance) 6 Ni 0.066 1.6 11.4 ± 0 . 2 0.05 1 Ni 0.066 3.1 12.9 ± 0 . 2 0.04 7 Ni 0.066 4.7 14.4 ± 0 . 2 0.1 12 Ni 0.099 3.1 20.4 ± 0 . 2 0.09 4 Co 0.033 3.1 2 Phases -8 Co 0.040 3.1 12.4 ± 0 . 2 0.07 10 Co 0.050 3.1 12.2 ± 0 . 2 -2 Co 0.066 3.1 13.2 ± 0 . 2 0.05 11 Co 0.083 3.1 15.8 ± 0 . 2 0.08 9 Co 0.099 3.1 21.8 ± 0 . 2 0.1 5 Co 0.132 3.1 2 Phases -13 Co 0.066 1.6 12.4 ± 0 . 4 0.08 14 Fe 0.066 1.6 13.0 ± 0 . 4 0.05 3 Fe 0.066 2.8 13.6 ± 0 . 2 0.08 15 Fe 0.099 2.8 18.8 ± 0 . 2 0.1 Table 4.2 : Hydrodynamic diameters of reverse micelles in the water/PE4LE/DHN system. # Metal dissolved in reverse micelle (0 (vol/vol) Metal ion cone, in reverse micelle (mol/L) Avg. hydrodynamic diameter of rev. micelle (nm) n/r2 (Relative variance) 17 Ni 0.066 3.1 14.2 ± 0 . 2 0.1 18 Co 0.066 3.1 14.2 ± 0 . 2 0.1 45 The effect of co (water: surfactant on a volumetric basis) on the size of the reverse micelles in the water/PE4LE/hexane system is shown in Figure 4.1. The data presented in Figure 4.1 indicate that the size of the reverse micelles increased with increasing co, and the trend was consistent for all three metals dissolved within the water pools. This result concurs with previous studies on a number of microemulsion systems, including water/CTAB/hexanol (Nagy et al, 1983), water/AOT/alkane (Pileni et al, 1985) and water/PEGDE/hexadecane (Kizling and Stenius, 1987). The sizes of the reverse micelles presented in Figure 4.1 do not display the precise linear dependence on co reported in previous studies. There is no obvious explanation for the lack of linearity in the data at this point. 1 CS s a Vi I 0.025 0.05 0.075 co 0.1 0.125 Figure 4.1 : Effect of© on the hydrodynamic diameters of reverse micelles in the water/ PE4LE/hexane system (temperature = 22 °C, [M 2 +] in water pool = 3.1 mol/L, 94.4 vol % hexane). It was also observed that stable, optically transparent microemulsions could only be prepared between co = 0.03 and co = 0.1. Phase separation occurred with microemulsions #4 and 46 #5 at values of© outside these limits. This observation agrees with previous studies conducted by Kizling and Stenius (1987) and Boutonnet et al. (1982) on non-ionic microemulsions. Boutonnet et al. found that a certain minimum amount of water was required to form clear, stable microemulsions in the water/PEGDE/hexane system with H2PtCl6 dissolved vvithin the water pools. This would explain the observed phase separation in the water/PE4LE/hexane system at values of © < 0.03. Kizling and Stenius found that a minimum concentration of PEGDE was required to form stable reverse micelles in the water/PEGDE/hexadecane system. The researchers claim that a minimum number of surfactant molecules is required to form a structure which can accommodate water between the ethylene oxide chains without the occurrence of extensive water/hydrocarbon contact. This would explain the phase separation at values of © > 0.1 (i.e. insufficient surfactant). The observed phase separation implies that the system imposes a natural limit on the size of the water pools which can be achieved. The data in Figure 4.1 indicate that the minimum water pool size for the water/PE4LE/hexane system was 12.2±0.2 nm at 22 °C. The effect of metal ion concentration within the reverse micelles on the hydrodynamic diameter of the reverse micelles is shown in Figure 4.2. A microemulsion with [Fe3+] = 4.5 mol/L could not be prepared since iron nitrate reached its solubility limit at about 3 mol/L. A cobalt microemulsion with [Co2+] = 4.5 mol/L was also not prepared. Nevertheless, the data in Figure 4.2 illustrate that the reverse micelle size increased with increasing metal ion concentration within the water pool. This result confirms the proposed relationship between metal ion concentration and reverse micelle size in nonionic surfactant systems discussed in section 2.3.4. It would appear that increasing the concentration of multivalent metal nitrates within the reverse micelles of a nonionic surfactant led to an increase in the HLB of the surfactant molecules at the interface, which in turn led to an increase in the size of the reverse micelles. It should be noted that 47 increasing the metal ion concentration over the range considered in this study did not have as strong an effect on reverse micelle size as increasing co (e.g. increasing metal ion concentration from 1.5 to 4.5 mol/L only resulted in an increase in reverse micelle size of 4 nm). I CD o I n i i 1 2 3 4 5 Metal ion concentration in water pool (mol/L) Figure 4.2 : Effect of metal ion concentration within the water pool on the hydrodynamic diameter of the reverse micelles in the water/PE4LE/hexane system (temperature 22 °C, co = 0.066, 94.4 vol % hexane). The data in Figure 4.2 also indicate that the type of metal species had some effect on the size of the reverse micelle. For all metal ion concentrations, the water pools containing Fe3+ were larger than those containing Ni 2 + and Co2+. This effect has not been reported in the literature for microemulsions of nonionic surfactants. However, the observed trend can be qualitatively explained using the same arguments used to relate metal ion concentration to reverse micelle size. Schott (1973) claimed that the 'salting-in' effectiveness of metal salts in POE based nonionic surfactants can be qualitatively correlated to the heat of hydration of the metal salt. Consequently, the differences in the heats of hydration of Ni(NC>3)2, Co(N03)2 and Fe(N03)3 could have led to 48 differences in the extent of 'salting-in' of the surfactant molecules at the interface of the reverse micelles. As discussed in section 2.3.4, these differences in the extent of 'salting-in' would have lead to differences in the HLB of the surfactants, which could account for the differences in reverse micelle size with metal species observed in Figure 4.2. 4.2 The synthesis of reduced metal colloids in the water/PE4LE/hexane system. 4.2.1 Introduction Once the factors influencing the size of the reverse micelles in the water/PE4LE/hexane system had been identified, it was necessary to investigate the relationship between reverse micelle size and the size of colloidal catalysts synthesized therein. The synthesis of reduced Ni, Co and Fe colloids was initially studied as a model intra-micellular precipitation reaction. Although zero valent metal is not the active species of dispersed heavy oil hydrocracking catalysts, the model intra-micellular reaction was easy to study, and the reduced metal colloids were stable and could be easily characterized using DLS. 4.2.2 Characterization of reduced metal colloids Reduced metal colloids were prepared in the microemulsions described in Table 4.1 by the addition of hydrazine (N 2H4.xH 20), using the method outlined in section 3.1.2. The reduction of the microemulsions containing Ni 2 + was characterized by a colour change from light green to colourless, whilst the Co2+ containing microemulsions changed from pink to orange. The Fe2+ microemulsions changed from a light yellow colour to dark orange upon reduction. In all cases, the suspensions of the reduced metal colloids were stable and optically transparent after reduction. The sizes of the reduced Ni, Co and Fe colloids synthesized using hydrazine were determined by dynamic light scattering (DLS). The results are presented in Table 4.3. The DLS 49 data were analyzed using both monomodal and multimodal particle size distributions to determine whether a distribution of particle sizes was present. Table 4.3 : Reduced metal colloid sizes as determined by DLS. # Metal CO Metal ion Concentration (mol/L) Monomodal Fit Multimodal Fit Particle Diameter (nm) u/r 2 D l (nm) D2 (nm) 6 Ni 0.066 1.6 12.2 ±0.2 0.06 1 Ni 0.066 3.1 126 ± 20 0.2 142.8 11.2 7 Ni 0.066 4.7 200 ± 40 0.17 248 14.4 12 Ni 0.099 3.1 100 ± 10 0.2 139 16.8 4 Co 0.033 3.1 2 Phase 8 Co 0.040 3.1 12.6 ±0.2 0.05 10 Co 0.050 3.1 12.6 ±0.4 0.1 2 Co 0.066 3.1 22 ± 1 0.2 11 Co 0.083 3.1 not measured 9 Co 0.099 3.1 23.4 ±0.2 0.1 5 Co 0.132 3.1 2 Phase 13 Co 0.066 1.6 12.2 ±0.4 0.03 14 Fe 0.066 1.6 15.2 ±0.2 0.05 3 Fe 0.066 2.8 17.8 ±0.2 0.05 15 Fe 0.099 2.8 20.6 ±0.2 0.1 In general, the size distributions of the reduced Co and Fe colloids were reasonably well described by a monomodal fit, with u/T2 < 0.1 in most cases. The sizes of the Co and Fe colloids ranged from 10 to 20 nm. By contrast, the reduced Ni size distributions were not well described by a monomodal fit (u/T2 « 0.2). The data from samples #1, #7 and #12 were better described by a bimodal size distribution, with both large particles > 100 nm, and smaller 10-20 nm colloids. It would appear that the 10-20 nm Ni colloids formed initially in the reverse micelles aggregated to larger 100-200 nm sized particles. Boutonnet et al. (1982) reported similar results for the synthesis of Pd, Rh and Ir colloids in water/PEGDE/hexane microemulsions using H 2. The 50 researchers were able to prepare 5 nm sized Pd colloids with H 2, but they found that the Rh and Ir colloids aggregated to larger particles which settled from suspension. The observed differences between the sizes of the reduced Ni and the reduced Co and Fe colloids may have been due to differences in the relative rates of reduction and nucleation during the reduction reaction. For example, a slow rate of reduction in the case of Ni may have allowed the Ni colloids to aggregate during the reduction reaction. The Ni colloids prepared in microemulsion #6 were an exception to the trend described above. In this case the reduced colloids were small (12.2 nm) and fairly monodisperse (u/T2 = 0.02). The small size of the reduced Ni colloids prepared in microemulsion #6 appears to suggest that the low concentration of metal ions in the water pool (1.6 mol/L) had a moderating effect on the apparently uncontrolled particle growth which occurred in the other Ni microemulsions. The relationship between the size of the reverse micelle and the size of the reduced metal colloids synthesized therein is presented in Figure 4.3. The data points representing microemulsions #1, #7 and #12 (extremely large Ni particle sizes) were not included in Figure 4.3 since the mechanism of colloid formation in these three cases was probably different, resulting in uncontrolled particle growth. Although there is some scatter in the data, a clear trend of increasing colloid size with increasing reverse micelle size is evident. This result concurs with a number of studies published in the literature which suggest that particle growth within reverse micelles is spatially constrained by the intrinsic size of the reverse micelle (Ravet et al, 1984; Nagy et al, 1983; Modes and Lianos, 1989; Pileni et al, 1992). The small amount of scatter in the data of Figure 4.3 may have been due to imperfect mixing/distribution of N2H4 during the reduction reaction. A number of researchers claim that a low [reducing ion]:[metal ion] ratio results in the formation of large colloids (Nagy et al, 1983; Pileni et al, 1992). Consequently, imperfect mixing of N2H4 could have resulted in local variations in the concentration of N 2 H 4 51 during the reduction reaction, which in turn could have lead to the formation of larger than expected colloids in some of the reduced samples. CD 1 CC Q O O O cu o • Ni A C O o Fe 10 15 20 Reverse Micelle Diameter (nm) 25 Figure 4.3 : Relationship between reverse micelle size and reduced metal colloid size. Since light scattering is an indirect method of determining particle size, some particle size measurements were made using transmission electron microscopy (TEM) to check the reduced metal colloid sizes given by DLS. TEM photographs of reduced Co and Fe colloids prepared in microemulsions #2 and #3 are presented in Appendix 2.1. The size distributions of the Co colloids present in TEM photographs #925 and #926 are given Figure 4.4. The two size distributions of the Co colloids in photographs #925 and #926 are fairly similar, both in terms of their general shape and the average particle diameter. This indicates that the colloid samples shown in the two photographs were representative of the larger colloid population. Table 4.4 gives the average sizes and standard deviations of the particles in each photograph. The size distributions are fairly narrow, with standard deviations (expressed as a percentage of the mean particle size) of 27% and 22% respectively. Martino et al. (1994) reported 8/d values of 5-15 % for Fe, Pd and FeS2 52 colloids prepared in the various microemulsion systems (see Table 2.2), whilst Boutonnet et al. (1991) reported a 8/d value of 15 % for Pt colloids prepared in the water/PEGDE/hexane system. Energy disperse x-ray (EDX) spectra of a typical reduced Co colloid in photo #925 and a typical reduced Fe colloid in photo #929 are shown in Appendix 2.2. A cobalt peak and an iron peak are present in each spectrum, though they are of low intensity due to the extremely small size of the colloids analyzed. Gold TEM grids were used for this analysis, accounting for the gold peak observed. Size Range (nm) • Photo 925 • Photo 926 Figure 4.4 Size distribution of Co colloids present in TEM photographs #925 and #926 53 Table 4.4 : Statistical analysis of the reduced Co colloid sizes in photographs #925, #926 and the reduced Fe colloid sizes in photograph #929. Photo # Number Average Particle Diameter d (nm) Standard Deviation 5 (nm) 5/d 925 (Co) 926 (Co) 8.7 8.4 2.3 1.9 0.27 0.22 929 (Fe) 10.7 3.1 0.29 Figure 4.5 shows the size distribution of the Fe colloids in photograph #929 and the size distribution of the Co colloids in photographs #925 and #926. The data in Figure 4.5 indicate that the Fe colloids synthesized in the reverse micelles of the water/PE4LE/hexane system were shghtly larger than the Co colloids synthesized in the same system. The sizes of the reverse micelles in the microemulsions used were very similar (d = 13.2 nm for Co, d = 13.6 nm for Fe). As with the reduced Ni particles, the slight differences in the sizes of the Co and Fe colloids may have been caused by differences in the relative rates of reduction and particle formation. It should be noted that DLS gave a shghtly larger average particle size compared to TEM for the same sample of reduced Co and Fe colloids. This discrepancy between the microscopic and light scattering results illustrates that these two techniques essentially measure different physical quantities. TEM provides a direct measurement of the colloid size only, whilst light scattering gives a measurement of the size of the metal colloid plus the layer of surfactant molecules which surrounds the colloids in suspension. If one assumes that PE4LE surfactant molecules are at least 2.6 nm in length, and that they surround the suspended colloids in a monolayer, then a net diameter of at least 13 nm is to be expected for the 8 nm colloid plus surfactant covering. 54 50 0-6 6-8 8-10 10-12 12-14 14-16 16+ Size Range (nm) • Reduced Co • Reduced Fe Figure 4.5 : Size distributions of reduced Fe and Co colloids prepared in water/PE4LE/hexane microemulsions 4.3 Synthesis of metal sulfide colloids using dimethyl disulfide. Once the synthesis of reduced metal colloids in the water/PE4LE/hexane system was demonstrated, the focus of the research shifted to the synthesis of metal sulfide colloids. A number of researchers have shown that metal sulfides are the active form of dispersed heavy oil hydrocracking catalysts (see section 2.2.6). Consequently, the direct intramicellular synthesis of metal sulfide colloids is of more practical significance than the synthesis of zero valent metal colloids. The synthesis of nickel sulfide colloids in the water/PE4LE/hexane system using dimethyl disulfide ((CH3)2S2) was initially investigated. Dimethyl disulfide (DMDS) is an organic sulfiding agent commonly used to presulfide supported metal catalysts used in fixed bed heavy oil 55 hy(kocracking processes. The results of the synthesis reactions with DMDS are presented in Table 4.5. An EDX spectrum of the solid reaction products are presented in Appendix 2.3. Table 4.5 : Results of the nickel sulfide colloid synthesis experiments using DMDS. Average Particle Size Elements detected by EDX 150 210 1 mm 1 mm Ni Ni, S(2:l ratio) The results presented in Table 4.5 indicate that DMDS was not an effective sulfiding agent below its decomposition temperature of 207 °C. No sulfur was detected by EDX analysis in the solid product of the synthesis reaction performed at 150 °C, but sulfur was detected in the product of the reaction at 210 °C. Severe particle aggregation was observed in both the experiments, and the solid particles obtained were generally in the rnillimeter size range or larger. This aggregation was probably due to the phase separation of the microemulsion at the elevated reaction temperatures, resulting in the formation of bulk aqueous and organic phases. The temperature sensitivity of microemulsions of non-ionic polyoxyethylene based surfactants has been reported by a number of researchers (Overbeek et ai, 1984; Kizling and Stenius, 1987; Shaw, 1993). Kizling and Stenius found that phase separation occurred in the water/PEGDE/hexadecane system at temperatures ranging from 32 °C to 42 °C, depending on the water content of the microemulsion. Phase separation at elevated temperature in microemulsions of non-ionic surfactants is thought to be due to a decrease in the hydrophihc-hpophihc balance (HLB) of the surfactant (Overbeek et ai, 1984) . The hydration of the lyophihc poly(ethylene oxide) groups decreases with increasing temperature, resulting in the surfactant becoming less hydrophilic - i.e. HLB decreases. Phase separation finally occurs when the solubility of the 56 surfactant in the aqueous phase decreases to a point where the surfactant can no longer occupy the interface between the aqueous and organic phases. The temperature sensitivity of microemulsions (as illustrated in the results presented in Table 4.5) has a number of important implications for the synthesis of colloidal hydrocracking catalysts in reverse micelles. Organic sulfiding agents such as DMDS which require high reaction temperatures are completely unsuitable for the synthesis of nm sized metal sulfide colloids in reverse micelles. Clearly, any precipitation reaction to prepare metal sulfide colloids in reverse micelles must take place at a relatively low temperature (< 30 °C) to maintain the structural integrity of the microemulsion. This temperature restriction has important implications for the synthesis of this type of catalyst in industrial settings. 4.4 Synthesis of metal sulfide colloids using H2S 4.4.1 Introduction The results of the catalyst synthesis experiments with DMDS clearly indicated that any precipitation reaction in reverse micelles of non-ionic surfactants must be conducted close to ambient temperature (20-30 °C). Consequently, the focus of the research shifted to finding a suitable low-temperature sulfiding agent. Two sulfiding agents have been used by previous researchers to prepare metal sulfide colloids in reverse micelles at room temperature, namely: H2S, and reverse micelle entrapped S2" ions (see Table 2.2). Metal sulfide synthesis using reverse micelle entrapped S2" is a controlled and well characterized technique, but it is only suited to laboratory scale preparations. Colloid preparation with H2S gas is presumably a more robust technique, and thus more suited to large industrial scale applications. Consequently, the preparation of metal sulfide colloids was investigated using 5% H2S in H 2. A 5% mixture of H2S in H 2 is a typical gas composition found in a refinery H 2 stream. 57 Sulfides of Ni, Co and Fe were prepared in reverse micelles of the water/PE4LE/hexane system following the procedure outlined in section 3.1.4. Details of the experiments performed are presented in Appendix 2.4. Fairly lengthy reaction times of 2 and 6 hours were used to account for the low rates of reaction expected at room temperature. Published data on the solubility of H2S in hexane (Fogg, 1988) indicate that the mole fraction of H2S in the liquid phase XH2S > 0.04 at the reaction conditions (22 °C, partial pressure of H2S = 300 kPa). Consequently, external mass transfer limitations were assumed to be negligible in the reaction system. The colloids synthesized in experiments #26, 27, 28 and 29 were characterized by EDX and x-ray photoelectron spectroscopy (XPS) in an attempt to determine the metal sulfide species formed during the synthesis reaction. Experiments #30, 31 and 32 were repeats of experiments #26, 27 and 29, and the colloids synthesized in these experiments were characterized by TEM in order to establish the size distributions of the particles formed. 4.4.2 Physical characterization of metal sulfide colloids using TEM and EDX. The colloids synthesized in experiments #30, 31 and 32 were characterized by TEM in order to establish the size distributions of the particles formed. The synthesis of nickel sulfide colloids in the microemulsions containing Ni 2 + (experiments #26 and #30) was characterized by a colour change from light green to dark green/black, whilst the Co2+ containing microemulsions (experiments #27 and 31) changed from pink to dark purple/black . The Fe2+ microemulsions (experiments #28, 29 and 32) changed from a light yellow colour to dark yellowback. The metal sulfide suspensions were unstable when exposed to ah, and generally changed colour from clear-black to the colour of the original microemulsion over a period of 12 hours. The colour change was probably due to reoxidation of the metal sulfide colloids by atmospheric oxygen. Pileni et al. (1992) found that CdS colloids prepared in the water/AOT/iso-octane system 58 reoxidized (photocorroded) in the presence of oxygen and light according to the following reaction: hv CdS+202 -» CdS04 -> Cd2 + +S0 2 -The sensitivity of the metal sulfide suspensions meant that the colloid sizes could not be determined using dynamic light scattering. Consequently, TEM was the only method used to determine the size of the metal sulfide colloids. Photographs of the metal sulfide colloids synthesized are shown in Appendix 2.5. The size distributions of the nickel sulfide and cobalt sulfide colloids synthesized in experiments #30 and 31 are presented in Figure 4.6. Unfortunately, decent size distributions of the Fe sulfide colloids synthesized in experiment #32 could not be obtained. The size distribution of the nickel sulfide colloids was taken from the colloids in photographs #127, 130, 132 and 134 (194 particles), whilst the size distribution of the cobalt sulfide colloids was taken from the colloids in photographs #138 and 142 (622 particles). A simple statistical analysis of the Ni sulfide and Co sulfide particle sizes is given in Table 4.6. An explanation of the terms in Table 4.6 is given in Appendix 2.6. The surface area average particle diameters (d S A ) of the colloids were calculated for purposes of comparison with the results of the BET surface area measurements (as discussed in section 4.4.4). EDX spectra of typical colloids seen in Appendix 2.5 are presented in Figures 4.7 to 4.10. The preparation of metal sulfide colloids with H2S appears to have been accompanied by some degree of aggregation or uncontrolled particle growth. In all cases, the metal sulfide colloids were far larger than the zero-valent metal colloids prepared in identical microemulsions (see Table 4.3). The nickel sulfide and cobalt sulfide colloids were also significantly larger than the CdS 59 colloids prepared by Pileni et al. (1992) in the ionic surfactant system water/AOT/iso-octane using 100% H2S (d » 4 nm). 50 " 1 0-25 25-50 50-75 75-100 100-125 125-150 150+ Size Range (nm) • Nickel Sulfide (Experiment #30) • Cobalt Sulfide (Experiment #31) Figure 4.6 : Size distributions of nickel sulfide and cobalt sulfide colloids synthesized in experiments #30 and #31 (Microemulsions: co = 0.066, [M2+] = 3.1 mol/L). Table 4.6 : Simple statistical analysis of the Ni sulfide particle sizes in photographs #127, 130, 132 and 134 and the Co sulfide particle sizes in photographs #138 and 142. NiS Exp.#30 CoS2 Exp. #31 Number Average Particle 67 71 Diameter d n (nm) Number Standard Deviation 31 32 6„ (nm) Average Surface Area 17069 18268 SA (nm2) Surface Area Average Particle 74 76 Diameter d S A (nm) 60 1 C/3 t O > 13 ti Figure 4.7 : 0.8 A 0.6 A 0.4 0.2 H 0 0 S K a N i K a 2.5 5 keV 7.5 10 EDX elemental analysis of a colloid from experiment #026 (Ni microemulsion, 7 MPa 5% H 2 S in H 2 , 22 °C, 6 hours) CD I 13 Figure 4.8 : EDX elemental analysis of a colloid from experiment #027 (Co microemulsion, 7 MPa 5% H 2 S in H 2 , 22 °C, 6 hours) 61 1 0.8 -0 -I , , 1 H 0 2.5 5 7.5 10 keV Figure 4.9 : E D X elemental analysis of a colloid from experiment #028 (Fe microemulsion, 7 MPa 5% H 2 S in H 2 , 22 °C, 2 hours) 1 0.8 H V5 a I Pi 0.6 H 0.4 H 0.2 SKa 2.5 FeKa 5 — i — 7.5 - r 10 keV Figure 4.10 : E D X elemental analysis of a colloid from experiment #029 (Fe microemulsion, 7 MPa 5% H 2 S in H 2 , 22 °C, 6 hours) 62 The size distributions of the nickel sulfide and cobalt sulfide particles presented in Figure 4.6 indicate that the particle sizes ranged from 25 to 150 nm in both cases. The Co sulfide particles displayed a distinct bimodal distribution, with a large number of smaller colloids in the 25-50 nm range, and larger particles in the 75 to 125 nm range. This trend is clearly apparent in photograph #141. The Ni sulfide particles did not display a distinct bimodal distribution, and most of the nickel sulfide colloids were in the 25-75 nm size range. Photograph #131 shows a magnified image of 100-120 nm Ni sulfide particles. It would appear that the particles display a granularity at the 10 nm scale, indicating that the particles may be made up of smaller 10 nm colloids. This hypothesis is supported by the presence of very faint 10-15 nm sized particles in photograph #131. In conclusion, the results indicated that the synthesis of nickel sulfide and cobalt sulfide in the reverse micelles of the water/PE4LE/hexane microemulsion was accompanied by aggregation and excessive particle growth beyond the confines of the reverse micelles. As in the case of the synthesis of reduced Ni colloids with hydrazine, the observed aggregation may have been due to a low rate of reaction, which would have given the colloids sufficient time to interact and coalesce. The differences between the size distributions of the Ni sulfide and Co sulfide colloids may also have been due to differences in the rate of reaction between the two metals. The EDX results indicated the presence of metal-sulfur containing compounds in each of the samples analyzed. The relative areas of the sulfur and metal peaks indicated that the suhur:metal ratio in the bulk was > 1. However, it was not possible to further identify the metal sulfide compounds present in each sample from the EDX results alone. In addition, since EDX analysis only identifies elements heavier than sodium (n = 23), it was not possible to identify the presence of other metal compounds (e.g. metal oxides) in the samples. Consequently, another 63 analysis technique was required to adequately characterize the metal sulfide compounds formed using H2S. 4.4.3 Characterization of metal sulfide colloids using XPS X-Ray Photoelectron Spectroscopy (XPS) was used in an attempt to identify the metal sulfide species prepared in the water/PE4LE/hexane system using H2S. It should be noted that XPS is a surface sensitive technique, and tells very little about the nature of the bulk material. The effect of exposure to air after preparation was also investigated using XPS. As explained in section 3.3.5, two samples were prepared for each of the metal sulfides analyzed by XPS. The first set was prepared and transported under a helium blanket in an attempt to prevent exposure to atmospheric oxygen. The second set of samples was prepared using the same procedure, but the samples were exposed to the air during preparation. The XPS results are presented in Figures 4.11 to 4.16. Standard binding energies of various Ni, Co and Fe compounds are presented in Appendix 2.7 for reference. Figures 4.11 and 4.12 show the XPS spectra of samples prepared from experiment #026 (H2S sulfiding of a Ni microemulsion). The spectrum of the sample prepared under He indicates that a mixture of NiS and NiO was present on the colloid surfaces after sulfiding. The presence of NiO is difficult to explain, given the lack of oxygen in the reacting system. It is believed that the surfaces of the NiS colloids were partially oxidized by exposure to 0 2 present in the He supply, or by atmospheric oxygen from the accidental exposure to air during sample preparation. The differences between the spectra of the 'air-free' and 'air-exposed' samples further illustrate the air sensitivity of the NiS colloid surfaces. Both the NiS peak at 853 eV (Figure 4.11) and the S2" peak at 162 eV (Figure 4.12) are absent from the spectra of the air-exposed sample, indicating that the surfaces of the NiS colloids oxidized to NiO upon exposure to the ah. 64 Figures 4.13 and 4.14 show the XPS spectra of samples prepared from experiment #027 (H2S sulfiding of a Co microemulsion). The peak at 780.5 eV in spectrum (a) of Figure 4.13 indicates the presence of a Cobalt-Oxygen compound (either CoO or Co(OH)2 ) on the surface of the sulfided colloids. However, the peak at 162.7 eV in spectrum (a) in Figure 4.14 clearly indicates the presence of sulfur as S2" (i.e. sulfur in a metal-sulfur bond), suggesting that CoS2 was present on the particle surface. Exposure to air had little effect on the Co 2p3 / 2 spectrum, whilst the S 2p spectrum showed a clear shift from a Metal-Sulfur bond to a Metal-S04 bond. From these spectra, it would appear that both cobalt oxide and cobalt sulfide were initially present on the surface of the colloids. The S 2p spectra indicate that the cobalt sulfide species on the surface of the colloids was oxidized to CoS04 upon exposure to the air. Figures 4.15 and 4.16 show the XPS spectra of samples prepared under He from experiments #028 and #029 (H2S sulfiding of Fe microemulsion). In both cases, the surface species appeared to be either Fe203 or Fe304, and the S 2p spectrum was flat in both cases (the S 2p spectrum of experiment #026 was included in Figure 4.16 for purposes of comparison). It would appear that the surfaces of the sulfided iron colloids were completely oxidized to Fe203 or Fe304 during the sample preparation procedure. 65 Figure 4.11 : Experiment #026 - XPS Scan of Ni 2p3/2 region, (a) Sample prepared under He, (b) Sample exposed to air. 157 159 161 163 165 167 169 171 173 Binding Energy (eV) Figure 4.12 : Experiment #026 - XPS Scan of S 2p region, (a) Sample prepared under He, (b) Sample exposed to air. 66 765 775 785 795 805 815 825 Binding Energy (eV) Figure 4.13 : Experiment #027 - XPS Scan of Co 2p3 / 2 region, (a) Sample prepared under He, (b) Sample exposed to air. 157 159 161 163 165 167 169 171 173 Binding Energy (eV) Figure 4.14 : Experiment #027 - XPS Scan of S 2p region, (a) Sample prepared under He (b) Sample exposed to air. 0.0 705 710 715 720 725 730 Binding Energy (eV) Figure 4.15 : Experiments #028 and #029 - XPS Scan of Fe 2p3 / 2 region, (a) #029, sample under He, (b) #028, sample under He. Figure 4.16 : Experiments #028 and #029 - XPS Scan of S 2p region, (a) S 2p scan from Figure 4.11 (b) #029, sample under He, (c) #028, sample under He. 68 4.4.4 Characterization of metal sulfide colloids using BET surface area measurement The surface area of a sample of cobalt sulfide colloids prepared in an experiment identical to experiment #31 was measured using N 2 adsorption. As mentioned before, the metal sulfide colloids were recovered from the microemulsion using tetrahydrofuran, following the method outlined in section 3.3.4. The surface area of the sample was found to be 25 m2/g from both the N 2 adsorption and desorption measurements. Assuming that the sample was made up of non-porous spheres, this specific surface area corresponded to a average particle size of 56 nm. (details of the surface area calculations and calculations of average particle size are presented in appendix 2.8). The specific surface area as determined the BET surface area measurement (25 m2/g) was slightly larger than the specific surface area calculated from the TEM measurements1 (18 m2/g). This indicates that the CoS2 colloids had some internal porosity which resulted in the higher specific surface area as determined by N 2 adsorption/desorption. 4.5 Summary of the major findings on colloidal catalyst preparation in reverse micelles The experimental work outlined thus far has described the synthesis of colloidal heavy oil hydrocracking catalysts in the reverse micelles of the water/PE4LE/hexane microemulsion. The water: surfactant ratio (co) and the metal ion concentration in the reverse micelle were identified as the key factors affecting the size of the reverse micelles in the water/PE4LE/hexane system. The smallest reverse micelles (11.4±0.2 nm in diameter) were obtained by minimizing both co and the metal ion concentration. A direct relationship was demonstrated between reverse micelle size and the size of reduced metal colloids synthesized in the reverse micelle, and fairly monodisperse Co and Fe 1 TEM Measurements of CoS2 particle size (see section 4.4.2): The S. A. average particle size of 76 nm for the recovered CoS2 corrresponded to a specific surface area of 18 m2/g. 69 colloids were prepared with sizes ranging from 10-23 nm. This result suggests that the size of reduced metal colloids prepared in the water/PE4LE/hexane system can be controlled by simply adjusting the water to surfactant ratio (co) in the microemulsion. The reduced Ni aggregated to much larger 100-200 nm sized particles, indicating that the rate of reduction also affects the size of solid colloids synthesized in microemulsions. The water/PE4LE/hexane system was found to be sensitive to increases in temperature, indicating that the synthesis of colloidal catalysts in reverse micelles must be performed at room temperature (20-30 °C). Ni, Co and Fe sulfide colloids were prepared in the water/PE4LE/hexane system using H2S at room temperature (22 °C), but this synthesis method was characterized by particle growth beyond the confines of the microemulsion. The size of the Ni and Co sulfide colloids (as determined by TEM) ranged from 25 to 150 nm, and were smaller than similar bulk catalyst particles prepared by the decomposition of organometallic compounds (see section 2.2.5). BET surface area measurements confirmed the particle size measurements made by TEM. Bulk EDX analysis of the metal sulfide colloids showed strong S peaks in all three metal species. However, the exact metal sulfide species prepared in the water/PE4LE/hexane system using H2S could not be identified using EDX. XPS analysis identified NiS and CoS2 species on the surfaces of the Ni and Co sulfide samples. However, the XPS analysis also revealed that the metal sulfide colloids were extremely sensitive to oxygen. Strong peaks of NiO, CoO/Co(OH)2 and Fe203/Fe304 were identified in each of the samples respectively. 70 Chapter 5 : Activity Measurements 5.1 Introduction The final phase of the research involved an investigation of the catalytic activity of the metal sulfide colloids prepared by the reverse micelle technique. The relative activity of these metal sulfide colloids, compared to metal sulfide catalyst particles prepared by the decomposition of organometallic compounds, was of primary concern in this investigation. A series of catalyst activity measurements was conducted to determine the hydrocracking activity of the catalysts. A s mentioned in section 3.4.1, the hydrocracking of a model compound diphenylmethane (DPM) was chosen for this purpose. Two sets of activity measurements were conducted, the first set using metal sulfide colloids in the microemulsion, and the second set using metal sulfide colloids recovered from the microemulsion. The spent catalyst from the activity measurements was characterized using X R D in an attempt to determine the active species present in each case. 5.2 Phase equihbrium of the reaction system The phase equihbrium of the reaction system was modeled using the A S P E N simulation package in order to determine the pressure and vapor/liquid distribution of the components in the reactor at the final conditions. A simplified reaction system of decahycfronaphmalene/H2S/H2 was used in the simulation. Water/PE4LE/decahydronaphthalene (or decalin) microemulsions were used for the first set of activity measurements since decalin is less volatile than hexane, and thus more likely to remain in the hquid phase at the reaction conditions. In addition, decalin is an inert diluent commonly used in D P M hydrocracking experiments (Wei et ai, 1992). D P M and P E 4 L E were not included in the simulation since the A S P E N component database did not have the 71 required thermodynamic parameters for these compounds. Details of the ASPEN simulations are presented in Appendix 3.1. A summary of the simulation results using the Redlich-Kwong-Soave (RKS) equation of state is presented in Table 5.1. According to the simulation results, the reaction mixture became supercritical between 405 and 410 °C. However, the presence of DPM and PE4LE (which have a lower relative volatility than decalin) in the reaction mixture probably raises the critical temperature of the system. The ASPEN simulation was only performed once the first set of activity measurements had been conducted, and prompted a reduction in reaction temperature from 430 °C to 400 °C. Table 5.1: Simulation results for the simplified reaction system (135 ml Decalin, 165 ml 5% H2S in H 2, initial pressure = 2.2 MPa (314 psi)) Temperature (°C) Final Pressure (psi) Fraction of decalin in liquid phase 390 1055 0.95 400 1100 0.97 405 1130 0.98 410 System supercritical 5.3. Preliminary activity measurements A preliminary set of activity measurements was conducted in water/PE4LE/decalin microemulsions containing metal sulfide colloids. As explained in section 5.2, decalin microemulsions were used for the activity measurements since decalin is less volatile than hexane, and thus more likely to remain in the liquid phase at the reaction conditions. The sizes of the reverse micelles in the water/PE4LE/decalin system were determined by dynamic light scattering (DLS), and the results are presented in Table 4.2. The diameters of the reverse micelles in the water/PE4LE/decalin microemulsions (#17 and #18 in Table 4.2) were similar to those of water/PE4LE/hexane microemulsions of identical composition (#1 and #2 in Table 4.1). 72 Consequently it is believed that the metal sulfide colloids prepared in the decalin system (and used in the preliminary activity measurements) were of similar size to the metal sulfide colloids prepared in the hexane system. The results of the preliminary activity measurements are presented in Table 5.2. The conversions based on the number of moles of benzene formed were generally found to be the same as the conversions based on the number of moles of toluene formed. Consequently, the conversions are reported on the basis of the number of moles of benzene or toluene formed (B,T). The conversions are also reported on the basis the number of moles of DPM converted during the reaction. The first order rate constants based on the conversion with respect to benzene/toluene are also presented for each experiment, since the first order rate constants allow for the direct comparison of experiments with different metal loadings. A sample calculation of the rate constant for experiment #8 (in Table 5.2) is given in Appendix 3.2. Details of each activity measurement, including the temperature/pressure profiles and GC analyses are presented in Appendix 3.3. Table 5.2 : Results of the preliminary activity measurements conducted in water/PE4LE/ decalin microemulsions (vol. microemulsion = 120 mL co = 0.066, mass DPM = 15.015g, initial pressure = 2.2 MPa, reaction temperature = 430°C, reaction time = 2 hours) Exp. # Catalyst Metal Loading (ppm) Sulfiding Time (hours) % Conv. (B,T basis) % Conv. (DPM basis) First order rate constant based on B, T basis (cm3/g/h) 8 Co sulfide 620 2 5 8 49 9 Co sulfide 620 6 6 8 51 11 Blank _ 2 5 12 -microemulsion1 13 Fe sulfide 520 6 4 12 43 14 Fe sulfide 2 520 6 4 9 47 1 Thermal cracking 2 elemental S added corresponding to a 4:1 mol ratio S:Fe 73 The cobalt sulfide (experiments #8 and #9) and iron sulfide (experiments #13 and #14) colloids prepared in water/PE4LE/decalin microemulsions did not show significant catalytic activity when compared to the thermal cracking experiment conducted with a microemulsion containing only water (experiment #11). The differences of 2-3% in the conversion based on DPM are not considered to be significant, and can probably be attributed to inaccuracies in the DPM calibration curve used to calculate the concentration of DPM (as discussed in section 3.4.4). The differences between the conversion based on DPM and the conversion based on toluene are not readily explained. The benzene and toluene may have cracked to lighter products during the reaction, accounting for the lower-than-expected concentration of these compounds in the liquid reaction products. The general inactivity of the metal sulfide catalysts in the preliminary activity measurements may have been due to a 'poisoning' effect caused by byproducts of surfactant pyrolysis. As mentioned in section 2.3.5, Martino et al. (1994) found that surfactant pyrolysis byproducts resulted in a loss of catalyst activity in the hydrogenolysis of naphthylbibenzylmethane (NBM) using both FeS2 colloids prepared in a water/PE4LE/octane microemulsion1, and Fe colloids prepared in water/DDAB/toluene. The authors attributed this loss in activity to the scavenging of hydrogen by the surfactant byproducts, as well as possible chemical and steric poisoning by the surfactant. In addition, the authors found that the catalytic activity of a commercial catalyst (Shell 324) for the hydrogenolysis of NBM decreased with increased doping with the surfactant DDAB. The formation of surfactant byproducts during the preliminary activity measurements is clearly illustrated in Figures 5.1, 5.2 and 5.3. Figure 5.1 shows a GC trace of a typical reaction 1 The colloidal FeS2 catalyst was freeze dried and used as particle-embedded surfactant powder in the activity measurements. 74 feed (microemulsion plus DPM) for the preliminary activity measurements. The decalin, PE4LE (surfactant) and DPM peaks are clearly identified2. Figure 5.2 shows the GC trace of a mixture of benzene, toluene, decalin, PE4LE and DPM corresponding to a 5% theoretical conversion of DPM to benzene and toluene. Figure 5.3 shows the GC trace of an actual reaction product from the preliminary activity measurements (experiment #13). Clearly, a number of peaks other than those of the theoretical reaction products (benzene and toluene) are present in Figure 5.3. These peaks probably represent the pyrolysis byproducts of the surfactant PE4LE. In addition, the surfactant peak (retention time « 7.5 min) in Figure 5.3 is far smaller than the corresponding peak in Figure 5.1, indicating that the surfactant was converted to other byproducts during the reaction. s Retention time (min) Figure 5.1: GC trace of a typical reaction feed for preliminary activity measurements. (DETN = decalin, PE4LE = polyoxyethylene-4-lauryl ether, DPM = diphenylmethane). 2 Decalin has 2 peaks corresponding to the cis- and trans- isomers. 75 D H N DPM J DHN PE4LB - r 10 Retention time (min) Figure 5.2 : GC trace of a mixture of Benzene (B), Toluene (T), Decalin (DHN), PE4LE and DPM corresponding to 5% conversion of DPM to benzene and toluene. S 10 Retention time (min) Figure 5.3 : GC trace of a typical reaction product from the preliminary activity measurements (B = benzene, T = toluene, DHN = decalin, DPM = diphenylmethane). 76 The results in Table 5.2 also indicate that increasing the catalyst synthesis time ('sulfiding time' in Table 5.2.) from 2 to 6 hours had little effect on the activity of the Co sulfide colloidal catalyst (compare experiments #8 and #9). This result indicates that either: i) increasing the colloid synthesis time has little effect on the physical characteristics or the yield of the colloidal catalyst particles, or ii) the 'poisoning' effect of the surfactant or surfactant byproducts is strong enough to mask any effect that increased catalyst synthesis time might have on catalyst activity. The addition of extra elemental sulfur (experiment #14) also had little effect on the catalytic activity of the Fe sulfide catalyst synthesized in the water/PE4LE/decalin system. Two previous studies by Wei and co-workers have shown that the addition of extra elemental sulfur has a promotional effect on the hydrocracking of diphenyhnethane (Wei et al, 1992) and di(l-naphthyl) methane (Wei et al, 1993) with FeS2 catalysts. This provides further evidence that the 'poisoning' effect of the surfactant/surfactant byproducts over-rides any catalytic effects in the reaction system. Given the similarities between the reaction system studied by Martino et al. and the system studied in the present work, it seems likely that the presence of surfactant in the DPM reaction system had a similar 'poisoning' effect as that reported by Martino et al. (1994). However, the results of the ASPEN simulation indicated that the reaction system may have been supercritical at 430 °C, which also could have resulted in the apparent lack of catalytic activity observed for the Co sulfide and Fe sulfide colloids in this reaction system Clearly, the reason for the apparent lack of catalytic activity could not be determined from the results outlined in Table 5.2 alone. An additional experiment (#22) was conducted with cobalt sulfide colloids in a water/PE4LE/decalin microemulsion at 400 °C, a temperature which was below the critical point of the system according to the ASPEN simulation results. The results of experiment #22 are presented in Table 77 5.3. The results of a thermal cracking experiment at 400 °C are also included in Table 5.3 for comparison. Table 5.3 : Results of the additional activity measurement conducted in water/PE4LE/ decalin microemulsion (vol. microemulsion = 120 ml, co = 0.066, mass DPM = 15.015g, initial pressure = 2.2 MPa, reaction temperature = 400°C, reaction time = 3 hours) Exp. Catalyst Metal Sulfiding % Conv. % Conv. First order rate constant # Loading Time (B, T basis) (DPM basis) based on B, T basis (ppm) (hours) (cm3/g/h) 22 Co 620 6 1 7 4 microemulsion 15 Thermal - 6 1 6 -Again the cobalt sulfide catalyst did not show significant catalytic activity compared to the thermal cracking experiment (experiment #15). Since the activity of the cobalt sulfide catalyst relative to thermal cracking did not increase when the temperature was reduced from 430 °C to 400 °C, it would appear that the lack of catalytic activity observed at 430 °C was not due to the system being in a supercritical state. It would appear that the apparent lack of catalytic activity observed for the Co sulfide and Fe sulfide colloids in this reaction system (as shown in Table 5.2) may be attributed to the presence of surfactant byproducts in the reaction system. 5.4 Activity measurements using recovered colloids 5.4.1 The use of recovered colloids As discussed in section 5.3, the presence of surfactant and surfactant pyrolysis byproducts made it impossible to perform meaningful catalytic activity measurements with colloidal metal sulfide catalysts in the microemulsions. Consequently, the metal sulfide colloids were recovered from suspension in order to remove the surfactant, and the recovered colloids were then used in a 78 second set of activity measurements. As explained in section 3.4.3, the colloidal catalysts were prepared in water/PE4LE/hexane microemulsions using H2S, and then recovered from suspension using tetrahycfrofuran (THF). The metal sulfide colloids aggregated during the extraction process with THF. However, a BET surface area measurement of a sample of recovered colloids revealed that the catalyst still had a large specific surface area (see section 4.4.4). Consequently, the recovered catalysts were probably composed of loose aggregates of the original 25-125 nm sized metal sulfide colloids prepared in the reverse micelles. The iron sulfide colloids oxidized completely to orange-coloured iron oxide (Fe203 or Fe304) during the recovery process. Hence, Fe sulfide colloids were not recovered from the microemulsions and no activity measurements were conducted with Fe sulfide. The second set of activity measurements was conducted at a lower temperature of 400 °C. The reduction in reaction temperature was prompted by the results of the ASPEN simulation of components in the reactor, which indicated that a liquid phase may not have been present in the reactor at 430 °C. The reaction time was increased from 2 to 3 hours to compensate for the lower reaction rates expected at 400 °C. 5.4.2 Experimental results The results of the second set of activity measurements are presented in Table 5.4. Conversions are again reported on the basis of the number of moles of benzene and/or toluene formed, as well as the number of moles of DPM converted during the reaction. The first order rate constant based on the conversion with respect to benzene/toluene is also presented for each experiment. Details of each activity measurement, including temperature/pressure profiles and GC analyses are presented in Appendix 3.4. 79 Table 5.4 : Results of the second set of catalytic activity measurements. (liquid volume = 130 ml initial pressure 5% H2S in H 2 = 2.4 MPa, reaction temperature = 400 °C, reaction time = 3 hours) Exp. # Catalyst Metal Loading (ppm) C O AP during reaction (psi) % Conv. (B,T basis) % Conv. (DPM basis) Rate constant based on B, T conversion (cm3/g/h) 15 Thermal - - 50 1 6 -16 Fe(CO)5 926 - 50 1 6 2.5 17 Recovered NiS 655 0.066 40 1 9 2.3 18 Fe(CO)5 1000 - 50 1 4 3.8 19 Co naphthenate 1000 - 150 15 15 58.4 20 Recovered CoS2 1000 0.099 100 3 3 9.3 23 Recovered CoS2 2000 0.066 50 1 4 1.7 The first order rate constants based on the number of moles of benzene/toluene formed during the reaction were used as the basis for comparison of the relative activities of the various catalysts. The rate constants were calculated using the conversion based on the number of moles of benzene/toluene formed during the reaction since the production of benzene and toluene reflected the true hydrocracking activity of the catalyst. In addition, the calibration of the GC for the detection of benzene and toluene was more accurate (R2=0.99) than the calibration for the detection of DPM (R2=0.89). As with the preliminary activity measurements, the conversion based on the number of moles of DPM converted was generally greater than the conversion based on the production of benzene and toluene. This result tends to suggest that other by-products were produced during the reaction. However, no other significant products were found by GC and MS (mass spectroscopy) analysis of the reaction products. Some gas (CH, or C2He) may have been produced during the reactions, accounting for the increase in pressure over the course of the 80 reaction. The increase in pressure may also have been due to the addition of benzene and toluene to the vapor phase over the course of the reaction. Moreover, the conversion based on benzene was generally the same as the conversion based on toluene. This would tend to indicate that benzene and toluene were stable reaction products, and were not selectively cracked to lighter components. The catalyst obtained from the decomposition of iron pentacarbonyl (experiments #16 and #18) showed little activity over and above the base thermal activity (experiment #15) for the hydrocracking of DPM. The difference between the results of runs #16 and #18 was within the error associated with the determination of DPM concentration by GC. The active catalyst obtained from the decomposition of cobalt naphthenate showed the greatest activity for the hydrocracking of DPM. The relatively high conversion was accompanied by a large increase in pressure (150 psi) over the course of the reaction, which indicates that the benzene and toluene products went into the vapor phase at the reaction conditions. In addition, the conversion based on DPM concentration was equal to the conversion based on the production of benzene and toluene, indicating that little or no by-products were formed. An additional experiment (experiment #21, not shown) was conducted to determine whether benzene or toluene were produced by the decomposition of Co naphthenate in decalin at the reaction conditions. GC traces of the reaction product of experiment #19 and the products of the decomposition of Co naphthenate are presented in Appendix 3.5. The peaks of benzene, toluene, n-decane (the internal GC standard), decalin and DPM are labeled. Clearly, no benzene and toluene are evident in the product of the decomposition of Co naphthenate, indicating that these compounds were not produced by the decomposition of Co naphthenate. The main decomposition products were found at short retention times (0.46 min and 0.998 min), and at retention times from 5.7 min to 5.9 min. 81 The recovered MS catalyst (experiment #17) showed little activity over and above thermal cracking. The recovered C0S2 catalyst in experiment #20 showed some activity, which was again reflected in an increased pressure rise of 100 psi over the course of the reaction. However, the second batch of CoS2 catalyst showed little activity. The reason for the difference between the activity of the cobalt and nickel sulfide catalysts synthesized by the microemulsion technique and the catalyst from the decomposition of cobalt naphthenate is not obvious. This matter is discussed further in section 5.4.3. 5.4.3 Catalyst characterization using XRD The spent catalysts from the second set of activity measurements were characterized by x-ray diffraction (XRD) in order to determine the active species and the average crystallite sizes of the catalysts. A sample of the fresh C0S2 catalyst recovered from the microemulsion was also analyzed by XRD before it was used in experiment #23. The preparation of the XRD samples was conducted in an air-free envhonment in order to minimize exposure to atmospheric oxygen. The results of the XRD analyses are presented in Table 5.5. The XRD spectra of the samples are presented in Appendix 3.6. Fei.xS was identified as the active species resulting from the decomposition of Fe(CO)5 during the activity measurements. This result is consistent with previous studies which also identified the non-stoichiometric sulfide Fei-xS as the active species formed by the decomposition of iron-based organometallic compounds in sulfur rich hydrocarbon environments (Anderson and Bockrath, 1984; Herrick et al., 1990). CogFeSg was identified as the active species resulting from the decomposition of Co naphthenate (spent catalyst from experiment #19). The presence of iron in the catalyst was not 82 expected, and this iron contaminant may have originated from the stainless steel of the reactor and reactor internals. Table 5.5 : Summary of the catalyst characterization using XRD. Experiment / Catalyst Sample Species Identified Crystal Plane Peak Position (2©) Relative peak intensity Crystallite Size (nm) #18 - Spent Catalyst (Fe(CO)5 precursor) Fe!.xS 2 0 0 29.82 0.60 > 150 1 2 0 6 33.70 0.66 > 150 1 2 0 12 43.60 1 75 2 2 0 53.00 0.43 47 #19 - Spent Catalyst (Co naphthenate precursor) CosFeSs 1 1 1 15.50 0.32 32 3 1 1 29.79 0.96 20 2 2 2 31.22 0.33 27 5 1 1 47.63 0.35 29 4 4 0 52.13 1 21 #20 - Spent Catalyst C04S3 / Co8FeS8 1 1 1 15.50 0.25 not measured 3 1 1 29.84 0.95 68 2 2 2 31.62 1 > 150 1 5 1 1 47.65 0.29 33 4 4 0 52.12 0.75 32 #23 - Fresh Catalyst Recovered CoS2 Amorphous - - - -#23 - Spent Catalyst Co8FeS8 1 1 1 15.51 0.17 22 3 1 1 29.98 0.88 29 2 2 2 31.27 0.31 26 5 1 1 47.74 0.37 27 4 4 0 52.19 1 22 1 No peak broadening was detected over-and-above normal instrumentational broadening in these samples. The estimate of a minimum crystallite size of 150 nm in these samples was based on the assumption that the peak widths can be read off the XRD spectra with an accuracy of +0.01 degrees. The spectrum of the spent catalyst from experiment #20 differed slightly from the spectrum of the catalyst from experiment #19. Firstly, the peak at 20 = 15.5° was slightly smaller than in the spectra of the catalyst from #19. This peak resulted from the presence of Fe in the crystal matrix, which indicates that the spent catalyst from experiment #20 had a lower level of iron contamination than the spent catalyst from experiment #19. Secondly, the peak at 20 «31.6° 83 was more intense and shifted slightly to the right compared to the corresponding peak in the spectrum of the spent catalyst from experiment #19. A good match was more difficult to find for the spent catalyst from experiment #20, the data fit the standard spectra for CogFeSg and C04S3 equally well. The XRD spectrum of the spent catalysts formed from the recovered C0S2 in experiment #23 was very similar to the spectrum of the active species resulting from the decomposition of Co naphthenate (spent catalyst from experiment #19). Again the relative intensity of the peak at 20=15.5° was smaller than in the spectrum of the catalyst from exp. #19. In conclusion, it would appear that the spent catalysts from experiments #19, 20 and 23 were of very similar crystalline form. The main difference between the catalysts appears to have been in the amount of iron contaminant incorporated in the crystal lattices (as indicated by differences in the relative heights of the peak at 20=15.5°). No comparable results could be generated for the NiS catalyst since sufficient spent catalyst could not be collected from the reactor product of activity experiment #17 to obtain a XRD spectrum XRD analysis revealed that the cobalt sulfide catalyst recovered from the microemulsion (fresh catalyst for experiment #23) was amorphous. No sharp peaks were identified (see XRD spectrum in Appendix 3.6), and the single broad peak present in the XRD spectrum was characteristic of an amorphous sohd. Clearly, the cobalt sulfide catalysts in experiments #20 and #23 underwent a phase transition from amorphous C0S2 to crystalline CosFeSg during the reaction. This phase transition was accompanied by the formation of 20-30 nm sized crystallites (particles) in the CogFeSg catalyst (see Table 5.5). It is interesting to compare the results of the activity measurements (as presented in Table 5.4) with the results of the XRD analysis of the spent catalysts (Table 5.5). Assuming that the catalysts were present as dispersed single crystallites (particles) at the reaction conditions, one 84 would expect the catalyst activity to have been inversely proportional to the crystallite size. For example, Takemura and Okada (1988) found that the specific activity of dispersed nickel catalysts for coal liquefaction increased with decreasing Ni crystallite size. However, the results in Table 5.5 indicate that there was no clear relationship between the catalyst activity and crystallite size. The crystallite size of the CogFeSg catalyst resulting from the decomposition of Co naphthenate (experiment #19) was small (20-30 nm), and the catalyst displayed high activity (k = 58 cm3/g/h) as was to be expected. However, the crystallite size of the spent catalyst from experiment #23 was also small, but the catalyst displayed little catalytic activity. In addition, the crystallite size of the spent catalyst from experiment #20 was slightly larger (30-68 nm), but the catalyst was more active than the catalyst from experiment #23. The lack of a strong correlation between catalyst activity and crystallite size indicates that the rate of reaction was not dependent on the available surface area of catalyst. This suggests that the macro structure of the catalyst introduced diffusion limitations which controlled the rate of reaction. For example, it is possible that the crystallites of the catalysts in experiments #18, #20 and #23 aggregated into agglomerates, either during the extraction process from the microemulsion, or by exposure to the elevated reaction conditions. Diffusion limitations resulting from the internal porosity would then account for the low activity of the catalyst in experiments #18, #20 and #23. Likewise, it is feasible that the catalyst in experiment #19 was present as dispersed single crystallites which did not suffer from diffusion limitations. 5.5 Summary of the activity measurements performed The metal sulfide colloids synthesized in reverse micelles showed little catalytic activity for the hydrocracking of DPM when the activity measurements were conducted in water/PE4LE/decalin microemulsions. This lack of activity can be attributed to a poisoning effect 85 by the surfactant and surfactant pyrolysis byproducts (identified by GC). Martino et al. (1994) reported a similar poisoning effect in N B M hydrogenolysis reactions performed in water/CnE^octane and water/DDAB/toluene microemulsions containing FeS 2 and Fe colloidal catalysts. A second set of activity measurements was conducted with a number of dispersed catalysts prepared from organometallic additives, and the metal sulfide colloids extracted from the microemulsions. The catalyst obtained from the decomposition of Co naphthenate showed the greatest activity for the hydrocracking of DPM. The active species was identified as CogFeSg, with an average crystallite size of 26 nm. The active species formed from microemulsion-extracted CoS 2 were also identified as CogFeSg, but these catalysts were found to be less active than the CogFeSg from the decomposition of Co naphthenate. The lack of a correlation between the crystallite size and the catalyst activity suggests that the catalysts from the microemulsion-extracted CoS 2 were present as aggregates with internal porosity which introduced rate controlling diffusion limitations. It would appear that the extraction process made it difficult to measure the intrinsic activity of the metal sulfide prepared in reverse micelles. 86 Chapter 6 : Conclusions and Recommendations for Future Work 6.1 Conclusions The synthesis of colloidal heavy oil hydrocracking catalysts in the reverse micelles of the water/PE4LE/hexane microemulsion was investigated in the present study. The factors affecting the size of the reverse micelles were identified, and the synthesis of reduced Ni, Co and Fe colloids and the corresponding metal sulfide colloids in the microemulsion was completed. The metal sulfide species synthesized in the microemulsion with 5% H2S in H 2 were characterized by TEM, EDX and XPS. Finally, the catalytic activity of the metal sulfide colloids synthesized in the water/PE4LE/hexane microemulsion was investigated using the hydrocracking of a model compound, diphenylmethane. The principle observations and conclusions resulting from this research are listed below: 1. The water: surfactant ratio (co) and metal ion concentration were identified as the key factors affecting the size of the reverse micelles in the water/PE4LE/hexane microemulsion. Reverse micelles ranging in diameter from 11.4 to 21.8 nm were obtained by varying co between 0.04 and 0.09. 2. A direct relationship was found between the reverse micelle size and the size of reduced Co and Fe colloids synthesized by the addition of N2FL, to water soluble Co and Fe salts dissolved in the water pools of the reverse micelles. The reduced Co and Fe colloids were monodisperse with sizes ranging between 10 and 23 nm, depending on co. Reduced Ni formed larger 100-200 nm aggregates, suggesting that the rate of reduction also had an effect on the size of metal particles prepared in reverse micelles. 3. High temperature colloid synthesis using dimethyl disulfide was not possible due to the instability of the microemulsion at elevated temperatures. Nickel, cobalt and iron sulfide colloids were prepared in water/PE4LE/hexane microemulsions using 5% H2S in H 2 at room temperature (22°C). The Ni sulfide and Co sulfide colloids had average sizes of 67 and 71 nm respectively (as determined by TEM), but were more polydisperse than the reduced metal colloids prepared in the same system. Again this result suggests that a low rate of reaction leads to particle growth beyond the confines of the reverse micelles. 4. XPS analysis of the sulfided colloids identified NiS and CoS2 as the species formed with 5% H2S in H2. The XPS analysis illustrated the extreme difficulty encountered when trying to characterize nanometer sized colloids without altering the physical and chemical nature of the colloids during the sample preparation and analysis. The metal sulfide colloids prepared were extremely sensitive to atmospheric oxygen, and strong peaks of NiO, CoO/Co(OH)2 and Fe 20 3/Fe 30 4 were detected. 5. The metal sulfide catalysts prepared in reverse micelles showed little activity for the hydrocracking of a model compound DPM when the activity measurements were performed in the microemulsions containing the metal sulfide colloids. This inactivity was ascribed to the poisoning of the catalyst by products of the pyrolysis of the surfactant formed during the reaction. The metal sulfide catalysts were extracted from the microemulsion to avoid the problem of the surfactant, and the extracted catalysts were used in a second set of activity measurements. The catalyst obtained from the decomposition of cobalt naphthenate showed greater activity for the hydrocracking of DPM than NiS and CoS2 colloids extracted from the microemulsion. 6. The spent catalysts from the activity measurements were characterized by XRD. The active species resulting from Co naphthenate and the extracted CoS2 were both identified as CosFeSs. The iron contamination in the catalysts probably resulted from corrosion of the reactor surfaces. No correlation was found between the crystallite size and the catalyst activity, which suggests that the catalysts from the microemulsion-extracted C0S2 were present as aggregates with internal porosity which introduced rate controlling diffusion limitations. It would appear that the extraction process made it difficult to measure the intrinsic activity of the metal sulfide prepared in reverse micelles. 6.2 Recommendations and Future work The results of this study indicate that the size of reduced Co and Fe colloids synthesized in the water/PE4LE/hexane microemulsion can be controlled by manipulating the size of the reverse micelles. However Ni, Co and Fe sulfide colloids synthesized in the same system with 5% H2S in H 2 grew beyond the confines of the reverse micelles. Consequently, reduced metal colloids should be used as catalysts for future heavy oil hydrocracking experiments since one can retain control over the size of the catalyst particles added to the heavy oil. In this case the reduced metal colloids would be sulfided by naturally occurring sulfur in the heavy oil. Likewise, the colloidal catalysts synthesized in microemulsions should not be extracted from the microemulsions before use. The colloids appear to aggregate during the extraction process, decreasing the dispersion of the catalyst in the liquid phase, and introducing mass transfer effects which make it difficult to determine the intrinsic activity of the colloidal catalyst. More specifically, the following future work is recommended to further expand on the results of this thesis: 1. Ni, Co and Fe colloids should be prepared in other microemulsions besides the water/PE4LE/hexane system. Other systems which form stable reverse micelles over a wider range of co values, and form water pools smaller than 12 nm should be investigated. Examples 8 9 may include: water/CTAB/alkane, water/AOT/alkane, water/DDAB/toluene and water/Ph9E6/ cyclohexane. Investigation of these systems will also allow for the determination of the effect of the solubility of the organic phase of the microemulsion in the heavy oil on the degree dispersion of the colloidal catalyst in the heavy oil. 2. A wider range of metals, possibly including bimetallic colloids (see Ravet et ai, 1987) should be prepared by the reverse micelle technique and used as hydrocracking catalysts. 3. The activity of reduced metal colloids synthesized in reverse micelles should be determined for the hydrocracking of heavy oil. The poisoning effect of surfactant byproducts identified in the model compound activity tests may be a serious disadvantage of preparing colloidal catalysts in reverse micelles, and it must be determined whether a similar poisoning effect occurs in heavy oil hydrocracking tests. 4. The spent catalyst form the heavy oil hydrocracking tests must be recovered and characterized. This will indicate whether the size of the original reduced metal colloid added to the heavy oil has any effect on the size of the active catalyst particle present at the reaction conditions. The relationship between crystallite size and hydrocracking activity should be investigated, as this could shed more light on the catalytic action of dispersed catalysts in heavy oil hydrocracking. 90 References Anderson R. R, Bockrafh B. C, Fuel, 1984, 63 (3), p. 329-333. Andres M., Charcosset H., Chiche P., Davignon L., Djega-Mariadassou G., Joly J., Pregermain S.,Fuel, 1983, 62, p. 69-72. Bearden R., Aldridge C. L., Energy Progress, 1981, 1, 1-4, p. 44-48. Boutonnet M , Kizling J., Stenius P., Colloids and Surfaces, 1982, 5, p. 209-225. Boutonnet M., Kizling J., Touroude K, Make G., Stenius P., Appl. Catal, 1986, 20, p. 163. Boutonnet M., Kizling J., Mintsa-Eya V., Choplin A., Touroude R., Maire G., Stenius P., Journal of Catalysis, 1987, 103, p. 95-104. Boutonnet M., Kizling I , Touroude R., Maire G., Stenius P., Catalysis Letters, 1991, 9, p. 347-354. Bradley J.S., Millar J. M., Hill E. W., Behal S., Journal of Catalysis, 1991, 129, p. 530-539. Chang N. J. and Kaler E. W., Langmuir, 1986, 2, p. 184-190. Curtis C. W., Pellegrino J. L., Energy & Fuels, 1989, 3 (2), p. 160-168. DabkowskiM. J., Shih S. S., Albinson K K, A.I.ChE. Symposium Series, 1991, 87 (282), p. 53-58. Danielsson I., Lindman B., Colloids and Surfaces, 1981, 3, p. 391-392. 91 Dautzenberg F. M., De Deken J. C, Catal. Rev. -Sci. Eng., 1984, 26 (3&4), p. 421-444. Dautzenberg F. M., De Deken J. C, In proceedings of Symposium on Developments in Hydrodemetallization Catalysts, A.C.S. Miami Beach Meeting, April 28 -May 3,1985, p. 8-20. Del Bianco A., Panariti N., Di Carlo S., Elmouchnino J., Fixari B., Le Perchec P., Applied Catalysis A: General, 1993, 94, p. 1-16. DelgassW. N., Hughes T. K, Fadley C. S., Catal. Rev., 1970, 4, p. 179-219. Derbyshire F., Davis A., Epstein M., Stansberry P., Fuel, 1986, 65, p. 1233-1239. EcclesR. M., Oil and Gas Journal, 1982, 80, p. 121. Fendler J., Chern. Rev., 1987, 87, p. 877-899. Flecher P. D. I., Howe A. M., Robinson B. H., J. Chern Soc.,Faraday Trans. 1, 1987, 83, p.985-1006. Fogg P. G. T., Solubility Data Ser., 32, 1988. Gates B.C., Katzer J.R., Schuit G.C.C., Chemistry of Catalytic Processes, McGraw - HilL New York, 1979. Hager G. T., Bi X. X., Eklund P. C, Givens E. N., Derbyshire F. J., Energy & Fuels, 1994, 8, p. 88-93. Henglein A., Chern. Rev., 1989, 89, p. 1861-1873. Herrick D. E., Tierney J. W., Wender I., Energy & Fuels, 1990, 4, p. 231-236. 92 Hirschon A. S., Wilson R. B., Am. Chern. Soc., Prep. Div Fuel Chern., 1989, 34 (3), p. 881-885. Kandori K, Kon-no K., Kitahara A., Fujiwara M., Tamura T., Proceedings of the Sixth International Symposium on Surfactants in Solution-Modern Aspects, New Delhi, India, Aug 18-22, 1986. Kon-no K, Koide M., Kitahara A., Nippon Kagaku Kaishi, 1984, 6, p. 815. KimH., Curtis C. W., Cronauer D. G., Sajkowski D. J., Am. Chern. Soc., Prep. Div Fuel Chern., 1989, 34 (4), p. 1431-1438. Kizling J., Stenius P., Journal of Colloid and Interface Science, 1987, 118 (2), p. 482-492. Laine J., TrimmD., J. Chern. Tech. Biotechnol., 1982, 32, p. 813-833. Lett R. G., Cugini A. V., Utz B R., Kranstman D., Cillo D. L., Jin G. T., VI Korea-U.S.A. Joint Workshop on Coal Utilization Technology, October 17-18, 1989. Lianos P., Thomas J. K, Journal of Colloid and Interface Science, 1987, 117 (2), p. 505-512. Liu C, Zhou J., Que G., Liang W., Zhu Y, Fuel, 1994, 73, 9, p. 1545-1550. Lupien, M., Syncrude - Facts and Figures, 1995, July, Syncrude Canada Ltd. Martino A., Wilcoxon J. P., Kawola J. S., Energy & Fuels, 1994, 8, p. 1289-1295. Meyer M., Wallberg C, Kurihara K, Fendler J., J. Chern Soc, Chern Commun., 1984, p. 90-91. Modes S., Lianos P., J. Phys. Chern., 1989, 93, p. 5854-5859. 93 Nagy J. B., Gourgue A., Derouane E. G., Preparation of Catalysts III, 1983, Elsevier, Amsterdam, p. 193-200. Oelderik J.M., Sie S.T., Bode D., Applied Catalysis, 1989, 47, p. 1-24. Overbeek J. T. G., de Bruyn P. L., Verkoekcx F., 'Microemulsions', in Surfactants, Tandros T. F., Ed., Academic Press, London, 1984, p. 111-132. Pecoraro T. A., Chianelli R. R., J. Catal, 1981, 67, p. 430. Petit C, Pileni M. P., J. Phys. Chern., 1988, 92, p. 2282-2286. Petit C, Lixon P., Pileni M. P., J. Phys. Chern., 1990, 94, p. 1598-1603. Pileni M. P., Lisiecki I., Colloids and Surfaces A: Physiochemical and Engineering Aspects, 1993, 80, p. 63-68. Pileni M. P., Motte L., Petit C, Chern. Mater., 1992, 4, p. 338-345. Pileni M. P., Zemb T., Petit C. J., Chern. Phys. Lett, 1985, 118, p. 414. Pillai V., Kumar P., Multani M., Shah D., Colloids and Surfaces A: Physiochemical and Engineering Aspects, 1993, 80, p. 69-75. Ravet I., Gourge A., Gabelica Z., Nagy J. B., Proceedings of the VIII International Congress on Catalysis, Berlin 1984, 4, p. 871-878. Ravet I., Nagy J. B., Derouane E. G., Preparation of Catalysts, 1987, Elsevier, Amsterdam, p. 193-200. 94 Robinson B. H., Topraciaglu C, Dore J., Chieux P., J. Chern. Soc. Faraday Trans. I, 1984, 80, p. 13. Rosen M. J., Surfactants and Interfacial Phenomena, John Wiley & Sons Inc., New York, 1978. RudnickL. R., Audeh S.,Am. Chern. Soc, Prep., Div. Pet. Chern., 1986, 31 (3), p. 686. Rudnick L. R., US Patent 4 642 175, 1987. Sanford E. C, Canadian Chemical News, 1996, 48 (1), p. 21-23. Sanford E. C, Steer J. G., Muehlenbachs K, Gray M. K, Energy and Fuels, 1995, 9, p. 928-935. Schott H., Journal of Colloid and Interface Science, 1973, 43 (1), p. 150-155. Schick M. J., Nonionic Surfactants, Marcel Dekker Inc., New York, 1987, p. 528-530. Shaw D. J., Colloid and Surface Chemistry, Butterworth-Heinemann Ltd., 4th. Ed., Oxford., 1992. Steigerwald M. L., Brus L. E., Annu. Rev. Mater. Sci., 1989, 19, p. 471-495. Stenberg V. I., Ogawa T. O., Willson W. G., Miller D., Fuel, 1983, 62, p. 1487. Sweeny P. G., Stenberg V. I., HeiR. D., Montano P. A., Fuel, 1987, 66, p. 532-41. Takemura Y., Okada K, Fuel, 1988, 67, p. 1549-1553. Ternan M., Canadian Journal of Chemical Engineering, 1983, 61, p. 689-696. Thomas M. G., Padrick T. D., StohlF. V., Fuel, 1982, 61, p. 716. 95 Utz B. K, Cugini A. V., Frammel E. A , Am. Chern. Soc, Prep. Div Fuel Chern., 1989, 34 (4), p. 1423. Van Driesen K P., Caspers J., Hydrocarbon Processing, 1979, 58 (5), p. 107. Varghese P., US Patent 4 581 127, 1986. Void R. D., Void M. L., Colloid and interface chemistry, Addison Wesley, London, 1983, p. 181-186. Weisser P., Landa S., Sulphide Catalysts, their Properties and Applications, Pergamon Press, New York, 1970. Wilcoxon J. P., Williamson R. L., Baughman R., J. Chern. Phys., 1993, 98 (12), p. 9933-9950. Wei X., Ogata E., Zong Z., Niki E., Energy & Fuels, 1992, 9, p. 868-9. WeiX., Ogata E., Zong Z., NikiE., Fuel, 1993, 72 (11), p. 1547-1552. 9 6 Appendices Appendix 1 1.1 Microemulsions prepared 1.2 List of activity measurements. 1.3 Modeling of the reaction system using ASPEN process simulation package. 1.4 GC temperature program 1.5 GC standard calibration mixtures 1.6 GC calibration data and calibration equations 1.7 Equations for the calculation of conversion Appendix 1.1 : Microemulsions Prepared Water/PE4LE/Hexane Microemulsions # Metal salt in Water Pool CO (vol/vol) Metal Concentration in water pool (mol/1) Microemulsion Status 1 Ni(N0 3) 2 0.066 3.1 OK 2 Co(N03)2 0.066 3.1 OK 3 Fe(N03)3 0.066 2.8 OK 4 Co(N03)2 0.033 3.1 Cloudy 5 Co(N03)2 0.132 3.1 2 Phase 6 Ni(N0 3) 2 0.066 1.6 OK 7 Ni(N0 3) 2 0.066 4.7 OK 8 Co(N03)2 0.040 3.1 OK 9 Co(N03)2 0.099 3.1 OK 10 Co(N03)2 0.050 3.1 OK 11 Co(N03)2 0.083 3.1 OK 12 Ni(N0 3) 2 0.099 3.1 OK 13 Co(N03)2 0.066 1.6 OK 14 Fe(N03)3 0.066 1.6 OK 15 Fe(N03)3 0.099 2.8 OK 16 Ni(N0 3) 2 0.066 3.1 OK 21 Ni(N0 3) 2 0.040 3.1 OK 22 Fe(N03)3 0.040 2.8 OK Water/PE4LE/Decahydronaphthalene Microemulsions: # Metal salt in Water Pool w (vol/vol) Metal Concentration in microemulsion (mol/1) Microemulsion Status 17 Ni(N0 3) 2 0.066 3.1 OK 18 Co(N03)2 0.066 3.1 OK 19 Fe(N03)3 0.066 2.8 OK 20 Ni(N0 3) 2 0.066 3.1 OK Appendix Appendix 1.2 : List of Activity Measurements Activity Tests conducted in Microemulsions (method 1) Exp # Reactor Charge (Volumetric Ratio) Catalyst (Metal Loading) Liq.. Vol. (ml) Sulfiding Cond. (Time/Temp) IP . (psig) Reaction Temp (°C) Reaction Time (h) 1 Decalin / DPM 8:1 ™ 90 2 h / 25 °C 1000 430 1 2 Decalin / DPM 8:1 Fe(CO)3 (620 ppm) 90 2 h / 25 °C 1000 430 1 3 M.E. #1/ DPM 8:1 Ni Sulfide (620 ppm) 90 2 h / 25 °C 1000 430 1 8 M.E. #2/DPM 8:1 Co Sulfide (620 ppm) 135 2 h / 25 °C 300 430 2 9 M.E. #2/ DPM 8:1 Co Sulfide (620 ppm) 135 6 h / 25 °C 300 430 2 11 Blank M.E. / DPM 8:1 135 2 h / 25 °C 300 430 2 13 M.E. #3/ DPM 8:1 Fe Sulfide (520 ppm) 135 6 h / 25 °C 300 430 2 141 M.E.#3/ DPM 8:1 Fe Sulfide (520 ppm) 135 6 h / 25 °C 300 430 2 22 M.E.#2/DPM 8:1 Co Sulfide (620 ppm) 135 6 h / 25 °C 360 400 3 1 elemental S added Appendix 1.2 Activity Tests conducted with extracted catalyst (method 2) Exp # Reactor Charge (Volumetric Ratio) Catalyst (Metal Loading) Liq. Vol. (ml) Sulfiding Cond. (Time/Temp) LP. (psig) Reaction Temp (°C) Reaction Time (h) 15 DHN/DPM 8:1 ~ 130 340 400 3 16 DHN/DPM 8:1 Fe(CO)5 926 ppm Fe 130 2 h / 25 °C 325 400 3 17 DHN/DPM 8:1 Extracted NiS 655 ppm Ni 130 2 h / 25 °C 300 400 3 18 DHN/DPM 8:1 Fe(CO)5 1000 ppm 130 2 h / 25 °C 340 400 3 19 DHN/DPM 8:1 Co Naphthenate 1000 ppm Co 130 2 h / 25 °C 340 400 3 20 DHN/DPM 8:1 Extracted CoS2 1000 ppm Co 130 2 h / 25 °C 360 400 3 21 DHN Co Naphthenate 1000 ppm Co 130 2 h / 25 °C 340 400 3 23 DHN/DPM 8:1 Extracted CoS2 2000 ppm Co 130 2 h / 25 °C 350 400 3 100 Appendix 1.3 Appendix 1.3 : Modeling of the reaction system using ASPEN process simulation package. The ASPEN process simulation package version 5.5 (ASPEN Technology Inc.) was used to determine the pressure and phase equihbrium of the reaction mixture at the final reaction conditions for these calculations. In other words, for a particular set of initial conditions (Ti, Pi), the final pressure (P2) and the vapor/liquid distribution (x/y) at the final reaction conditions (T 2 ) were required. The ASPEN process simulation package version 5.5 (ASPEN Technology Inc.) was used for these calculations. ASPEN cannot be used to simulate static or batch systems. Consequently, an arbitrary time dimension was added to the physical parameters of the system to fulfill this simulation specification (e.g. reactor volume ml was simulated as a flow in ml/s). The ASPEN simulation was performed in 2 steps: Firstly, n (the number of moles of 5% H2S in H 2 present at the initial conditions (Ti, Pi)) was calculated using a simple mixer block. Once n had been calculated, the reaction system under the initial conditions (Ti, Pi) was fully characterized. This reaction system was then entered into the flashcurve option of ASPEN, and a series of flash calculations were performed at various combinations of T 2 and P2. The particular pressure P2 which gave a final volume V 2 = 300 ml defined the vapor/hquid equihbrium of the reaction system under the reaction conditions. This calculation routine was performed for values of T 2 ranging between 400 and 430 °C. The SRK equation of state was used for the simulation. n moles 5% H2S in H 2 120 ml Decalin 15 ml DPM Heat x moles Vapor y ml Liquid Initial Conditions : Ti = 20 °C Pi = 300 psig Total Vol. = 300 ml Final Conditions T 2 P2 Total Vol. = V 2 101 Appendix Appendix 1.4 : Gas Chromatograph Temperature Program 40 °C 1 min Ramp to 70 °C @ 50 °C/min 70 °C 2 min Ramp to 170 °C @ 50 °C/min 170 °C 5 min Ramp to 270 °C @ 50 °C/min 270 °C 2 min Appendix 1.5 Appendix 1.5 : GC standard calibration mixtures # DPM Vol. Benzene Vol. Toluene Vol. DPM Vol. DHN Vol. C10 Total Vol. Conv. (%) (ml) (ml) (ml) (ml) (ml) (ml) 1 0 0 0 4000 32 1 37 2 2.5 53 63 3900 32 1 37.02 3 5 106 127 3800 32 1 37.03 4 7.5 159 190 3700 32 1 37.05 5 10 211 253 3600 32 1 37.06 6 12.5 264 317 3500 32 1 37.08 7 15 317 380 3400 32 1 37.1 8 20 423 506 3200 32 1 37.13 103 Appendix 1.6 Appendix 1.6 : GC Calibration data and calibration curves Initial Calibration Order of Calibration standand injections: 8, 1, 4, 6, 3 Sample # Repea t# Benzene Area Toluene Area C10 Area D H N Area D P M Area AB/Acio AT/ACIO ADPM/ACIO ADHN/ACIO 1 1 0 0 994562 42498980 6299802 0 0 6.334 42.731 2 0 0 954371 40011637 5784536 0 0 6.061 41.925 3 0 0 936148 39129864 5674284 0 0 6.061 41.799 AVG. 0 0 961694 40546827 5919541 0 0 6.152 42.152 3 1 131124 149099 948618 39189217 5412334 0.138 0.157 5.705 41.312 2 128301 141420 903110 38190713 5326533 0.142 0.157 5.898 42.288 3 136947 152619 972630 39366026 5313893 0.141 0.157 5.463 40.474 AVG. 132124 147713 941453 38915318.7 5350920 0.140 0.157 5.689 41.358 4 1 193053 226375 973945 39721228 5079145 0.198 0.232 5.215 40.784 2 189188 218785 914556 37921009 4886789 0.207 0.239 5.343 41.464 3 183054 214140 929337 38302532 4999489 0.197 0.230 5.380 41.215 AVG. 188432 219767 939279 38648256 4988474 0.201 0.234 5.313 41.154 6 1 315506 378987 942871 38632335 4725070 0.335 0.402 5.011 40.973 2 301867 364227 907939 37861915 4649583 0.332 0.401 5.121 41.701 3 299522 363172 905632 37832584 4631505 0.331 0.401 5.114 41.775 AVG. 305632 368795 918814 38108945 4668719 0.333 0.401 5.082 41.483 8 1 592329 686158 1067880 44398543 4985691 0.555 0.643 4.669 41.576 2 563569 654291 1020812 43195072 4876173 0.552 0.641 4.777 42.314 3 545502 634454 996294 42630601 4791228 0.548 0.637 4.809 42.789 AVG. 567133 658301 1028329 43408072 4884364 0.551 0.640 4.752 42.227 104 Appendix 1.6 Calibration Check - 4/01/96 Order of Injection: 3, 8, 4 Sample # Repeat # Benzene Area Toluene Area C10 Area D H N Area D P M Area AB/ACIO Ar/Acio ADPM/ACIO ADHN/ACIO 3 1 95717 114907 745441 30630101 4373046 0.128 0.154 5.866 41.090 2 95700 114673 719507 29503444 3771662 0.133 0.159 5.242 41.005 3 91201 110149 701520 28925611 3966629 0.130 0.157 5.654 41.233 AVG. 94206 113243 722156 29686385.3 4037112 0.130 0.157 5.588 41.109 4 1 139084 169371 731842 29241412 4116917 0.190 0.231 5.625 39.956 2 140090 170345 733431 29550967 4153986 0.191 0.232 5.664 40.291 3 141129 171971 740297 29552596 4078272 0.191 0.232 5.509 39.920 AVG. 140101 170562 735190 29448325 4116392 0.191 0.232 5.599 40.056 8 1 370625 446417 714311 29056086 3519256 0.519 0.625 4.927 40.677 2 375488 453134 722033 29483768 3645268 0.520 0.628 5.049 40.834 3 378019 453162 682805 27978764 2501202 0.554 0.664 3.663 40.976 AVG. 374711 450904 706383 28839539.3 3221909 0.531 0.639 4.546 40.829 Calibration Curves - Initial and Recalibration Sample MB/MCIO AB/ACIO AB/ACIO MT/M«O AJ/ACIO AT/ACIO MDPM/MCIO ADPM/ACIO ADPM/ACIO # Initial Calibr. Recal. Initial Calibr. Recal. Initial Calibr. Recal. 3 0.232 0.138 0.128 0.232 0.157 0.154 4.407 5.705 5.866 0.232 0.142 0.133 0.232 0.157 0.159 4.407 5.898 5.242 0.232 0.141 0.130 0.232 0.157 0.157 4.407 5.463 5.654 4 0.348 0.198 0.190 0.348 0.232 0.231 4.291 5.215 5.625 0.348 0.207 0.191 0.348 0.239 0.232 4.291 5.343 5.664 0.348 0.197 0.191 0.348 0.23 0.232 4.291 5.38 5.509 8 0.928 0.555 0.519 0.928 0.643 0.625 3.711 4.669 4.927 0.928 0.552 0.520 0.928 0.641 0.628 3.711 4.777 5.049 0.928 0.548 0.554 0.928 0.637 0.664 3.711 4.809 3.663 105 Benzene Calibration Curves 0.2 0.4 0.6 0.8 MB/Mcio o Initial Calibration • Recalibration 0.2 Toluene Calibration Curves 0.4 0.8 0.6 MT/Mcio o Initial Calibration • Recalibration Appendix 1.6 5 4 < 4.5 4 3.5 3.7 DPM Calibration Curves a i 6 • o 8 o • • • e • 3.9 4.3 4.1 M D P M / M C I O o Initial Calibration • Recalibration 4.5 107 Appendix 1.7 : Equations for the calculation of conversion Appendix 1.7 % Conversion (based on Benzene) = # moles of Benzene in Reactor Product . 100 # moles of DPM in Reactor Feed % Conversion (based on Toluene) = # moles of Toluene in Reactor Product . 100 # moles of DPM in Reactor Feed % Conversion (based on DPM) = # moles of DPM in Feed - # moles of DPM in Product .100 # moles of DPM in Reactor Feed 108 Appendix 2 2.1 TEM Photographs of reduced Co and Fe colloids 2.2 EDX of reduced Co and Fe colloids 2.3 EDX spectrum of Ni sulfide prepared using DMDS 2.4 Synthesis of metal sulfide colloids in the water/PE4LE/hexane microemulsion using 5% H2S in H 2 2.5 TEM Photographs of metal sulfide colloids synthesized using 5% H2S in H 2 2.6 Explanation of terms used in Table 4.6 2.7 Standard binding energies of various Ni, Co and Fe species 2.8 BET surface area calculations Appendix 2.1 Appendix 2.1 TEM photographs of reduced Co and Fe colloids no Appendix 2.1 111 Appendix 2.1 Photograph #926 : Reduced Co colloids at 100 000* magnification (1 m m = 4 nm) 112 Appendix 2.1 Photograph #928 : Reduced Fe colloids at 50 000* magnification (1 mm = 8 nm) 113 Appendix Photograph #929 : Reduced Fe colloids at 100 000* magnification (1 mm = 4 nm) Appendix 2.2 : EDX spectra of reduced Co and Fe colloids EDX Spectrum of a typical reduced Fe colloid Appendix 2.3 : EDX spectrum of Ni sulfide prepared using DMDS 1.0 0.0 2.5 5.0 7.5 10.0 KeV Appendix 2.4 Appendix 2.4 : Synthesis of metal sulfide colloids in the water/PE4LE/hexane microemulsion using 5% H2S in H2. (T = 22 °C, co=0.066, [Ni2 +] in water pool = 3 mol/L, [Co2 ] = 3 mol/L, [Fe3 ] = 2.75 mol/L) Exp. # Metal / Microemulsion # Pressure (MPa) Time (hours) Particle Size by TEM Elements by EDX Surface species by XPS 26 Ni (Batch #1) 7 6 No Yes Yes 27 Co (Batch #2) 7 6 No Yes Yes 28 Fe (Batch #3) 7 2 No Yes Yes 29 Fe (Batch #3) 7 6 No Yes Yes 30 Ni (Batch #1) 7 6 Yes Yes No 31 Co (Batch #2) 7 6 Yes Yes No 32 Fe (Batch #3) 7 6 No Yes No 117 Appendix 2.5 Appendix 2.5 : TEM photographs of metal sulfide colloids synthesized using 5% H2S in H 2 118 Appendix 2 5 Photograph #127 : N i sulfide colloids at 20 000* magnification (1mm = 20 nm) Appendix 2.5 Photograph #130 : Ni sulfide colloids at 20 000* magnification (1mm = 20 nm) Photograph #131 : Ni sulfide colloids at 100 000* magnification (1mm = 4 nm) Appendix 2.5 Photograph #132 : Ni sulfide colloids at 20 000* magnification (1mm = 20 nm) Appendix 2.5 122 Appendix 2.5 Photograph #134 : Ni sulfide colloids at 30 000* magnification (1mm = 13 nm) 123 Photograph #138 : Co sulfide colloids at 20 000* magnification (1mm = 20 nm) Appendix 2.5 124 Photograph #139 Appendix 2.5 Co sulfide colloids at 20 000* magnification (1mm = 20 nm) Photograph #141 : Co sulfide colloids at 20 000* magnification (lmrn = 20 nm) Appendix 2.5 127 Appendix 2.6 : Explanation of terms used in Table 4.6 Appendix 2.6 The number average particle size dnwas calculated as follows: n n Where, d; = diameter of particle i n = number of particles The number standard deviation 8„ was calculated as follows: n - 1 The surface areas of the metal sulfide colloids in the photographs were calculated (assuming spherical particles) and the average surface area SA was calculated as follows: ——- ESA ; SA = L Where, SAj = surface area of particle i The surface area average particle size d S A was calculated from SA as follows: d S A = SA 71 128 Appendix Appendix 2.7 : Standard binding energies of various Ni, Co and Fe species Compound Metal 2p 3 / 2 S2p ZnS - 161.7 S 2 0 3 - 162.5 Na 2 S0 4 - 169.1 FeS 706.7 162.7 FeO 709.6 Fc^Os 710.9 CoS2 778.1 162.8 CoO 780.4 CoCOrTh 781.3 NiS 853 162.3 NiO 855.9, 861.4 N i 2 0 3 856 NiSO, 857.0 169.4 Appendix 2.8 Appendix 2.8 : BET surface area calculations 1. Calculation of Specific Surface Area Mass of Sample = 0.053 g Surface Area (adsorption) =1.33 m2 Surface Area (desorption) =1.33 m2 .'. Specific Surface Area = 1.33/0.053 = 25 m2/g 2. Calculation of mean particle size For Spherical non-porous particles, r (particle radius) = 3 (Specific Surface Area*p) Assuming spherical non-porous particles of CoS2 present, p = 4.26*106 g/m3 .\r = 3 4.26*106 . 25 = 28.2nm .\ d = 56.4 nm 130 Appendix 3 3.1 ASPEN simulation of the reaction system 3.2 Sample calculation of rate constant for activity measurements 3.3 Data for the preliminary activity measurements 3.4 Data for activity measurements with recovered colloids 3.5 GC traces of experiment #19 and decomposition of Co naphthenate 3.6 XRD spectra of the spent catalysts from the second set of activity measurements Appendix 3.1 Appendix 3.1: ASPEN simulation of the reaction system 1. Determination of the number of moles of each component at the initial conditions (22 °C, 314 psi): The system was modeled as a simple mixer, as follows: .-. The DESIGN SPEC block of ASPEN was used to adjust the flowrate of stream (2) such that V = 300 ml/s. .*. The reaction system was : 2.7 kmol/h Decalin 0.54kmol/hH2 0.03 kmol/h H2S 2. The reaction system determined in part 1. above was entered into the FLASH CURVE option of ASPEN at the final reaction temperature T (varied from 390 °C to 410 °C) . The V/L distribution was then determined for a range of pressures. The particular pressure P which resulted in V= 300 ml/s defined the system at temperature T. (1) 135 ml/s Decalin (3) Mixed Stream Vol. flow = V ml/s (2) 5% HjS in Hj 132 Appendix 3.2 : Sample calculation of rate constant for activity measurements Rate equation for batch reactor: 1-x = exp(-k.x) Where, x = Conversion k = Rate Constant (cm3/g/h) x = Space time _ w.t v w = Mass of catalyst (g) t = Reaction time (h) v = Liquid volume (cm3) .•. For experiment #8 (Table 5.2), v = 135 cm3 w = 0.076 g t = 2h x = 0.05 (conversion w.r.t. benzene or toluene) .\k = - ln( l -0.05) 0.076*2 135 = 49 cm3/g/h Appendix Appendix 3.3 : Data for preliminary activity measurements Appendix 3.3 Experiment #8 Summary: Co Microemulsion cracking of DPM Sulphiding Time @ 1000 psig 2 h Reaction Time 2 h Reaction Temperature 430 °C Total Volume 135 ml 1. Reactor Charge: 120 ml Unreduced Co/PELE/Decalin. 15 ml Diphenylmethane 300 psig IP 5% H 2S in H 2 2. Reactor Preparation: 3. Nitrogen Purging: 4. Catalyst Synthesis: 2.1. Reactor and internals cleaned with soap and water, rinsed with water 2.2. Reactor and internals scoured with 'Brasso' 2.3. Reactor rinsed with hexane (2 hours with stirring) Reactor purged with N2 before and after run. Reactor charge including DPM presulphided in 5% H2S/H2 for 2 hours, (with stirring) 5. Reaction Profiles: Time to reach 430 °C Air cooling time 430 °C to 250 °C Water cooling Time 250 °C to 15 °C Total Reaction Time 90.5 min 24 min - min - min Time (min) Temp. (°C) Furnace Temp Pressure (psig) Stirring (rpm) 0 18 18 300 996 11 79 99 400 1000 20 124 148 450 996 30 174 193 460 991 40 225 240 500 993 50 274 288 550 994 60 324 338 650 996 70 373 395 800 996 80 409 454 1180 996 90 428 492 1490 990 90.5 430 493 1500 990 210 430 444 1700 215 392 1350 220 345 1150 225 300 1000 230 277 900 234 250 135 Appendix 3.3 6. GC Results: Calibration Curves : # Moles of component i in 1 \xl = m*(Area of component i) + c m c Benzene 2.09E-13 3.23E-9 Toluene 1.99E-13 2.47E-9 DPM 8.91E-14 4.37E-8 Reaction Product Injection # Component Area Moles in 1 ml 1 Benzene 99697 2.41E-08 Toluene 158632 3.40E-08 DPM 6257603 6.01E-07 Decalin 38845118 2 Benzene 103546 2.49E-08 Toluene 165933 3.55E-08 DPM 6439298 6.17E-07 Decalin 39988426 3 Benzene 102189 2.46E-08 Toluene 164479 3.52E-08 DPM 6302058 6.05E-07 Decalin 39258855 Conversions based on Injections #1, #2 and #3 # moles of DPM before 6.61E-07 cracking Average # moles of DPM 6.08E-07 after cracking Average # moles of 2.45E-08 Benzene after cracking Average # moles of 3.49E-08 Toluene after cracking DPM Conversion based 8 on DPM balance DPM Conversion based 5.3 on Toluene formed 136 Appendix 3.3 Experiment #9 Summary: Co Microemulsion cracking of DPM Sulphiding Time @ 1000 psig Reaction Time Reaction Temperature Total Volume 6h 2h 430 °C 135 ml 1. Reactor Charge: 2. Reactor Preparation: 3. Nitrogen Purging: 4. Catalyst Synthesis: 5. Reaction Profiles: 120 ml 15 ml 320 psig IP Unreduced Co/PELE/Decalin Diphenylmethane 5% H 2S in H 2 2.1. Reactor and internals cleaned with soap and water, rinsed with water 2.2. Reactor and internals scoured with 'Brasso' 2.3. Reactor rinsed with hexane (2 hours with stirring) Reactor purged with N 2 before and after run. Reactor charge including DPM presulphided in 5% H 2 S/H 2 for 6 hours. Time to reach 430 °C Air cooling time 430 °C to 250 °C Water cooling Time 250 °C to 15 °C Total Reaction Time 90 min 25 min 30 min 265 min Time (min) Temp. (°C) Furnace Temp Pressure (psig) rpm Stirring 0 21 22 320 1007 10 76 98 400 1017 20 124 151 450 1011 30 177 197 460 1012 40 226 242 500 1008 50 276 290 550 1007 60 326 340 625 1004 70 375 398 775 1005 80 411 463 1150 1004 90 430 490 1450 1015 210 430 445 1525 1007 215 381 1200 220 325 960 225 303 900 230 272 800 235 250 750 240 70 400 245 37 350 255 23 340 265 14 300 137 Appendix 3.3 6. GC Results: Calibration Curves : # Moles of component i in 1 ul = m*(Area of component i) + c m c Benzene 2.09E-13 3.23E-9 Toluene 1.99E-13 2.47E-9 DPM 8.91E-14 4.37E-8 Reaction Product Injection # Component Area Moles in 1 ml 1 Benzene 105059 2.52E-08 Toluene 169433 3.62E-08 DPM 6403035 6.14E-07 Decalin 39772695 2 Benzene 105670 2.53E-08 Toluene 169502 3.62E-08 DPM 6370335 6.11E-07 Decalin 39441154 3 Benzene 105224 2.52E-08 Toluene 168940 3.61E-08 DPM 6311489 6.06E-07 Decalin 38906733 Conversions based on Injections #1, #2 and #3 # moles of DPM before cracking 6.61E-07 Average # moles of DPM after cracking 6.11E-07 Average # moles of Benzene after cracking 2.52E-08 Average # moles of Toluene after cracking 3.62E-08 DPM Conversion based on DPM balance 7.7 DPM Conversion based oh Toluene formed 5.5 138 Experiment #11 Summary: Thermal Cracking with Blank Microemulsion Sulphiding Time @ 1000 psig 2 h Reaction Time 2 h Reaction Temperature 430 °C Total Volume 135 ml 1. Reactor Charge: 120 ml Water/PELE/Decalin 15.015 g DPM 335 psig IP 5%H 2 SinH 2 2. Reactor Preparation: 3. Nitrogen Purging: 4. A311Presulphiding: 5. Reaction Profiles: 2.1. Reactor and internals cleaned with soap and water, rinsed with water 2.2. Reactor and internals scoured with 'Brasso' 2.3. Reactor rinsed with hexane (2 hours with stirring) Reactor purged with N 2 before and after run. Reactor charge including DPM presulphided in 5% H 2 S/H 2 for 2 hours Time to reach 430 °C Air cooling time 430 °C to 250 °C Water cooling Time 250 °C to 15 °C Total Reaction Time 91.5 min 20 min 8 min - min Time (min) Temp. (°C) Furnace Temp Pressure (psig) rpm Stirring 0 24 24 335 1004 10 75 95 400 1010 21.5 132 156 460 1009 30 175 195 500 1010 40 225 242 510 1008 50 275 288 600 1004 60 325 337 700 1003 70 374 393 870 1000 80 410 460 1300 997 90 426 488 1600 999 91.5 430 490 1640 993 160 430 443 1550 1019 211.5 430 449 1550 995 221.5 326 1040 226.5 291 920 231.5 262 850 233.5 250 840 236.5 110 530 238.5 51 450 241.5 30 410 End 20 400 Appendix 3.3 6. GC Results: Calibration Curves : # Moles of component i in 1 (xl = m*(Area of component i) + c m c Benzene 2.09E-13 3.23E-9 Toluene 1.99E-13 2.47E-9 DPM 8.91E-14 4.37E-8 Reactor Feed Injection # Component Retention Time Area Moles in 1 ml 1 Decalin 5.323 41520945 Surfactant 7.291 90352 DPM 8.07 6593892 6.31E-07 2 Decalin 5.312 40063288 Surfactant 7.325 76382 DPM 8.057 6340212 6.09E-07 3 Decalin 5.337 43820740 Surfactant 7.317 92634 DPM 8.084 7047160 6.72E-07 Reaction Product Injection # Component Area Moles in 1 ml 1 Benzene 100391 2.42E-08 Toluene 154417 3.32E-08 DPM 5882786 5.68E-07 Decalin 36157816 2 Benzene 109730 2.62E-08 Toluene 169529 3.62E-08 DPM 6467442 6.20E-07 Decalin 39582279 3 Benzene 96909 2.35E-08 Toluene 149876 3.23E-08 DPM 5737077 5.55E-07 Decalin 35757466 140 Conversions based on Injections #land #3 # moles of DPM before cracking (from GC) 6.37E-07 Average # moles of DPM after cracking 5.61E-07 Average # moles of Benzene after cracking 2.39E-08 Average # moles of Toluene after cracking 3.28E-08 DPM Conversion based on GC results 11.9 DPM Conversion based on Toluene formed 5.1 Appendix 3.3 Experiment #13 Summary: Fe Microemulsion cracking of DPM • Sulphiding Time 6 h Reaction Time 2 h Reaction Temperature 430 °C Total Volume 135 ml 1. Reactor Charge: 120 ml Unreduced Fe/PELE/Decalin 15.015 g DPM 300 psig IP 5% H 2S in H 2 2. Reactor Preparation: 2.1. Reactor and internals cleaned with soap and water, rinsed with water 2.2. Reactor and internals scoured with 'Brasso' 2.3. Reactor rinsed with hexane (2 hours with stirring) 3. Nitrogen Purging: 4. Catalyst Synthesis: Reactor purged with N 2 before and after run. Reactor charge including DPM presulphided in 5% H 2 S/H 2 for 6 hours (with stirring) 5. Reaction Profiles: Time to reach 430 °C 90 min Air cooling time 430 °C to 250 °C 20 min Water cooling Time 250 °C to 15 °C min Total Reaction Time - min Time (min) Temp. (°C) Furnace Temp Pressure (psig) rpm Stirring 0 21 22 300 1005 10 77 101 400 1005 20 133 165 450 30 183 207 480 40 234 255 510 50 283 299 590 60 333 349 690 70 382 410 890 80 412 466 1300 90 430 487 1600 102 431 461 1660 135 430 457 1700 150 430 457 1700 180 430 458 1710 210 430 458 1750 215 377 1300 220 344 1150 225 297 1010 230 263 950 235 90 500 142 Appendix 3.3 6. GC Results: Calibration Curves : # Moles of component i in 1 (0.1 = m*(Area of component i) + c m c Benzene 2.09E-13 3.23E-9 Toluene 1.99E-13 2.47E-9 DPM 8.91E-14 4.37E-8 Reaction Product Injection # Component Area Moles in 1 ml 1 Benzene 61035 1.60E-08 Toluene 121334 2.66E-08 DPM 5992375 5.78E-07 Decalin 37445883 2 Benzene 57518 1.53E-08 Toluene 117148 2.58E-08 DPM 5754914 5.57E-07 Decalin 36228767 3 Benzene 60292 1.58E-08 Toluene 122260 2.68E-08 DPM 6020792 5.80E-07 Decalin 37672308 Conversions based on Injections #land #3 # moles of DPM before 6.61E-07 cracking (from mass) Average # moles of DPM 5.79E-07 after cracking Average # moles of 1.59E-08 Benzene after cracking Average # moles of 2.67E-08 Toluene after cracking DPM Conversion based 12.4 on DPM balance DPM Conversion based 4 on Toluene formed 143 Appendix 3.3 Experiment #14 Summary: Fe Microemulsion cracking of DPM with extra Sulphur added Sulphiding Time 6h Reaction Time 2h Reaction Temperature 430 °C Total Volume 135 ml 1. Reactor Charge: 120 ml 15.015 g 300 psig IP 0.173 g Unreduced Fe/PELE/Decalin DPM 5% H 2S in H 2 Elemental Sulphur (4:1 molar ratio S:Fe) 2. Reactor Preparation: 3. Nitrogen Purging: 4. Catalyst Synthesis: 5. Reaction Profiles: 2.1. Reactor and internals cleaned with soap arid water, rinsed with water 2.2. Reactor and internals scoured with 'Brasso' 2.3. Reactor rinsed with hexane (2 hours with stirring) Reactor purged with N 2 before and after run. Reactor charge including DPM presulphided in 5% H 2 S/H 2 for 6 hours (with stirring) Time to reach 430 °C Air cooling time 430 °C to 250 °C Water cooling Time 250 °C to 15 °C Total Reaction Time 91 min ' 24 min 15 min - min Time (min) Temp. (°C) Furnace Temp Pressure (psig) rpm Stirring 0 25 25 300 1005 10 79 100 350 20 129 151 400 30 179 196 440 40 229 242 480 50 279 288 510 60 329 338 600 70 378 393 800 80 411 455 1200 90 425 478 1450 91 430 478 1450 211 430 451 1650 220 339 1100 225 301 1000 230 272 900 235 250 850 240 72 450 245 37 400 250 25 380 144 Appendix 3.3 6. GC Results: ( Calibration Curves : # Moles of component i in 1 ul = m*(Area of component i) + c m c Benzene 2.09E-13 3.23E-9 Toluene 1.99E-13 2.47E-9 DPM 8.91E-14 4.37E-8 Reaction Product Injection # Component Area Moles in 1 ml 1 Benzene 82239 2.04E-08 Toluene 135553 2.95E-08 DPM 6363651 6.11E-07 Decalin 39000863 2 Benzene 74722 1.89E-08 Toluene 121175 2.66E-08 DPM 5735454 5.55E-07 Decalin 35766537 3 Benzene 80570 2.01E-08 Toluene 129503 2.82E-08 DPM 6176983 5.94E-07 Decalin 38077492 Conversions based on Injections # land #3 # moles of DPM before 6.61E-07 cracking (from mass) Average # moles of DPM 6.02E-07 after cracking Average # moles of 2.02E-08 Benzene after cracking Average # moles of 2.88E-08 Toluene after cracking DPM Conversion based 8.9 on DPM balance DPM Conversion based 4.4 on Toluene formed 145 Appendix Appendix 3.4 : Data from activity measurements with recovered colloids Appendix 3.4 Experiment #15 Summary: 1. Reactor Charge : 2. Reactor Preparation: Thermal Cracking Sulphiding Time Reaction Time Reaction Temperature Total Volume 130 ml 340 psig IP Oh 3h 400 °C 135 ml Reactant Mixture (120 ml Decalin, 15.015g DPM) 5% H 2S in H 2 2.1. Reactor /internals cleaned with soap and water, then rinsed 2.2. Reactor and internals scoured with 'Brasso' 2.3. Reactor rinsed with hexane (2 hours with stirring) 3. Nitrogen Purging: 4. Pressure Testing: Reactor purged with N 2 before and after run. Reactor pressure tested @ 1000 psig for 30 min No obvoius leaks / pressure decrease 5. Reaction Profiles: Time to reach 400 °C 77 min Air cooling time 400 °C to 20 min 250 °C Water cooling Time 25 0 °C to 15 °C 31 min Total Reaction Time 180 min Pressure Rise @ 400 °C 5 0 psi Overall Pressure Rise 0 psi Time (min) Temp. (°C) Pressure (psig) rpm Stirring Notes 0 22 340 700 10 80 390 712 25 147 450 698 35 194 500 693 45 250 540 700 55 295 600 710 75 390 850 700 77 400 910 701 Reaction Temp. 90 400 940 708 257 400 960 710 Air Cooling Started 267 315 725 720 277 250 600 723 Water Cooling Started 282 71 400 718 387 37 360 719 308 18 340 - Water Cooling Stopped 147 Appendix 3.4 6. GC Results: Reactor Feed Sample: 1 2 3 Average CIO 985545 1019525 1024387 1009819 Decalin 36119681 36550496 36510708 36393628 DPM 5415974 5391745 5418128 5408616 A D P M / A C I O 5.495 5.288 5.289 5.358 A D E C A L I N / A C I O 36.649 35.851 35.642 36.047 Reaction Product: 1 2 3 Average Benzene 14254 13508 13130 13631 Toluene 19028 17916 18073 18339 C10 965336 915797 917725 932953 Decalin 35393582 34246205 34236266 34625351 DPM 5392734 5176039 5189307 5252693 A B E N Z E N E / A C I O 0.015 0.015 0.014 0.015 A T O L U E N E / A C I O 0.02 0.02 0.02 0.02 A D P M / A C I O 5.586 5.652 5.655 5.631 A D E C A L I N / A C I O 36.665 37.395 37.306 37.122 7. GC Analysis: Calibration Curves: Mi/MCIO = (l/m)*Ai/AC10 - c/m m c Benzene 0.5931 -0.0032 Toulene 0.6971 -0.0057 DPM 1.4872 -0.8807 Feed Sample Reactor Feed Reactor Product Vol. CIO (ml) 0.139 3.611 Moles CIO 0.00071 0.01853 Moles Benzene 0.00056 Moles Toulene 0.00067 Moles DPM by GC 0.00299 0.08626 0.08112 Theoretical Moles DPM 0.00331 0.08595 % Conversion (Benz. ref.) 0.6 % Conversion (Tol. ref.) 0.8 % Conversion (DPM ref.) 6 148 Appendix 3.4 Experiment #16 Summary: Fe(CO)j Cracking @ 1000 ppm Fe SulphidingTime@ 1000 psig 2 h Reaction Time 3 h Reaction Temperature 400 °C Total Volume 135 ml 1. Reactor Charge : 130 ml Reactant Mixture (120 ml Decalin, 15.015g DPM) 325 psig IP 5%H 2 SinH 2 254 ul Fe(CO)5 2: Reactor Preparation: 2.1. Reactor /internals cleaned with soap and water, then rinsed 2.2. Reactor and internals scoured with 'Brasso' 2.3. Reactor rinsed with hexane (2 hours with stirring) 3. Nitrogen Purging: Reactor purged with N 2 before and after run. 4. Pressure Testing: Reactor pressure tested @ 1000 psig for 2 h during Sulphiding Small pressure decrease (1000 - 975 psig) 5. Reaction Profiles: Time to reach 400 °C 78 min Air cooling time 400 °C to 17 min 250 °C Water cooling Time 25 0 °C to 15 °C 20 min Total Reaction Time 180 min Pressure Rise @ 400 °C 50 psi Overall Pressure Rise 25 psi Time (min) Temp. (°C) Pressure (psig) rpm Stirring Notes 0 23 325 721 10 68 360 723 21 122 426 718 30 163 460 714 40 213 500 714 50 263 540 711 60 314 650 715 70 364 790 709 78 400 950 712 Reaction Temp. 258 400 1000 715 Air Cooling Started 263 339 800 722 268 293 700 726 273 262 650 727 275 250 640 726 Water Cooling Started 295 18 350 - Water Cooling Stopped 149 Appendix 3.4 6. GC Results: Reactor Feed Sample: 1 2 3 Average CIO 1038115 1074322 1049790 1054076 Decalin 39492939 39276443 38779450 39182944 DPM 5690263 5677645 5642867 5670258 A D P M / A C I O 5.481 5.285 5.375 5.38 A D E C A U N / A C I O 38.043 36.559 36.94 37.181 Reaction Product: 1 2 3 Average Benzene 14311 13998 13909 14073 Toluene 16159 18386 17164 17236 C10 1039945 1021124 1014252 1025107 Decalin 38299378 38203249 37893746 38132124 DPM 5593604 5663763 5681375 5646247 A B E N 2 E N E / A C I O 0.014 0.014 0.014 0.014 A T O L U E N E / A C I O 0.016 0.018 0.017 0.017 A D P M / A C I O 5.379 5.547 5.602 5.509 A D E C A U N / A C I O 36.828 37.413 37.361 37.201 7. GC Analysis: Calibration Curves: Mi/MCIO = (l/m)*Ai/AC10 - c/m m c Benzene 0.5931 -0.0032 Toulene 0.6971 -0.0057 DPM 1.4872 -0.8807 Feed Sample Reactor Feed Reactor Product Vol. CIO (ml) 0.139 3.661 Moles CIO 0.00071 0.01878 Moles Benzene 0.00054 Moles Toulene 0.00061 Moles DPM by GC 0.003 0.08625 0.0807 Theoretical Moles DPM 0.00331 0.08595 % Conversion (Benz. ref.) 0.6 % Conversion (Tol. ref.) 0.7 % Conversion (DPM ref.) 6.4 150 Appendix 3.4 Experiment #17 Summary: Recovered NiS Cracking @ 655 ppm Ni Sulphiding Time @ 1000 psig 2 h Reaction Time 3 h Reaction Temperature 400 °C Total Volume 135 ml 1. Reactor Charge : 130 ml Reactant Mixture (120 ml Decalin, 15.015g DPM) 300 psig IP 5% H 2S in H 2 0.3481 g Recovered NiS 2. Reactor Preparation: 2.1. Reactor /internals cleaned with soap and water, then rinsed 2.2. Reactor and internals scoured with 'Brasso' 2.3. Reactor rinsed with hexane (2 hours with stirring) 3. Nitrogen Purging: Reactor purged with N 2 before and after run. 4. Pressure Testing: Reactor pressure tested @ 1000 psig for 2 h during Sulphiding No pressure decrease observed 5. Reaction Profiles: Time to reach 400 °C 77 min Air cooling time 400 °C to 20 min 250 °C Water cooling Time 250 °C to 15 °C 13 min Total Reaction Time 180 min Pressure Rise @ 400 °C 40 psi Overall Pressure Rise 40 psi Time (min) Temp. (°C) Pressure (psig) rpm Stirring Notes 0 22 300 722 10 68 350 721 20 117 400 722 32 180 450 723 40 218 490 721 50 265 ,540 721 60 315 600 721 70 365 740 720 77 400 920 718 Reaction Temp. 257 400 960 723 Air Cooling Started 262 356 800 731 267 312 700 735 272 275 640 737 277 250 600 735 Water Cooling Started 290 32 340 - Water Cooling Stopped 151 Appendix 3.4 6. GC Results: Reactor Feed Sample: 1 2 3 Average CIO 1056615 1046479 1098626 1067240 Decalin 37247151 37233793 38357069 37612671 DPM 5596447 5705342 5803819 5701869 ADPM/ACIO 5.297 5.452 5.283 5.344 ADECALIN/ACIO 35.251 35.58 34.914 35.248 Reaction Product: 1 2 3 Average Benzene 8400 8856 8071 8442 Toluene 13413 14400 13481 13765 C10 946416 979846 1041281 989181 Decalin 34550790 35024041 36344210 35306347 DPM 5202088 5259779 5622184 5361350 ABENZENE/ACIO 0.009 0.009 0.008 0.009 ATOLUENE/ACIO 0.014 0.015 0.013 0.014 ADPM/ACIO 5.497 5.368 5.399 5.421 ADECAIJN/ACIO 36.507 35.744 34.903 35.718 7. GC Analysis: Calibration Curves: Mi/MCIO = (l/m)*Ai/AC10 - c/m m c Benzene 0.5931 -0.0032 Toulene 0.6971 -0.0057 DPM 1.4872 -0.8807 Feed Sample Reactor Feed Reactor Product Vol. CIO (ml) 0.139 3.611 Moles C10 0.00071 0.01853 Moles Benzene 0.00037 Moles Toulene 0.00052 Moles DPM by GC 0.00298 0.08627 0.07851 Theoretical Moles DPM 0.00331 0.08595 % Conversion (Benz. ref.) 0.4 % Conversion (Tol. ref.) 0.6 % Conversion (DPM ref.) 9 152 Appendix 3.4 Experiment #18 Summary: Fe(CO)5 @ 1000 ppm Fe (Repeat of Experiment #16) Sulphiding Time @ 1000 psig 2h Reaction Time 3 h Reaction Temperature 400 °C Total Volume 135 ml 1. Reactor Charge : 130 ml Reactant Mixture (120 ml Decalin, 15.015g DPM) 340 psig IP 5%H 2SinH2 275 ul Fe(CO)3 2. Reactor Preparation: 2.1. Reactor /internals cleaned with soap and water, then rinsed 2.2. Reactor and internals scoured with 'Brasso' 2.3. Reactor rinsed with hexane (2 hours with stirring) 3. Nitrogen Purging: Reactor purged with N 2 before and after run. 4. Pressure Testing: Pressure tested @ 1000 psig for during sulphiding No pressure decrease observed 5. Reaction Profiles: Time to reach 400 °C 77 min Air cooling time 400 °C to 17 min 250 °C Water cooling Time 250 °C to 15 °C 18 min Total Reaction Time 180 min Pressure Rise @ 400 °C 50 psi Overall Pressure Rise 0 psi Time (min) Temp. (°C) Pressure (psig) rpm Stirring Notes 0 20 340 709 10 69 400 702 20 116 450 701 30 165 490 700 40 216 550 701 50 266 600 700 60 315 690 697 70 365 810 699 77 400 990 699 Reaction Temp. 257 400 1040 702 Air Cooling Started 262 339 850 711 267 297 760 711 272 262 600 710 274 250 Water Cooling Started 277 125 500 707 282 60 430 707 292 20 340 - Water Cooling Stopped 153 Appendix 6. GC Results: Reactor Feed Sample: 1 2 3 Average CIO 739232 895610 815900 816914 Decalin 33073940 38068173 34572937 35238350 DPM 4982167 5768849 5056495 5269170 ADPM/ACIO 6.74 6.441 6.197 6.459 ADECALIN/ACIO 44.741 42.505 42.374 43.207 Reaction Product: 1 2 3 Average Benzene 19598 18400 17520 18506 Toluene 21797 20671 21289 21252 C10 788787. 758817 713411 753672 Decalin 28017178 27054419 25764932 26945510 DPM 4580654 4297111 4118708 4332158 ABENZENE/ACIO 0.025 0.024 0.025 0.025 ATOLUENE/ACIO 0.028 0.027 0.03 0.028 ADPM/ACIO 5.807 5.663 5.773 5.748 ADECALIN/ACIO 35.519 35.653 36.115 35.763 7. GC Analysis: Calibration Curves: Mi/MC 10 = (1/m)*Ai/AC 10 - c/m m c Benzene 0.5931 -0.0032 Toulene 0.6971 -0.0057 DPM 1.4872 -0.8807 Feed Sample Reactor Feed Reactor Product Vol. C10 (ml) 0.139 3.611 Moles C10 0.00071 0.01853 Moles Benzene 0.00087 Moles Toulene 0.0009 Moles DPM by GC 0.00352 0.08573 0.08258 Theoretical Moles DPM 0.00331 0.08595 % Conversion (Benz. ref.) 1 % Conversion (Tol. ref.) 1.1 % Conversion (DPM ref.) 3.7 Appendix 3.4 Experiment #19 Summary: 1. Reactor Charge : Cobalt Naphthenate @1000ppm Co Sulphiding Time @ 1000 psig Reaction Time Reaction Temperature Total Volume 130 ml 340 psig IP 1.475 g 2h 3 h 400 °C 135 ml Reactant Mixture (120 ml Decalin, 15.015g DPM) 5% H 2S in H 2 Cobalt Naphthenate (8% Co) 2. Reactor Preparation: 3. Nitrogen Purging: 4. Pressure Testing: 5. Reaction Profiles: 2.1. Reactor /internals cleaned with soap and water, then rinsed 2.2. Reactor and internals scoured with 'Brasso' 2.3. Reactor rinsed with hexane (2 hours with stirring) Reactor purged with N 2 before and after run. Reactor pressure tested @ 1000 psig for 2 h during Sulphiding No pressure decrease observed Time to reach 400 °C Air cooling time 400 °C to 250 °C Water cooling Time 250 °C to 15 °C Total Reaction Time Pressure Rise @ 400 °C Overall Pressure Rise 77 min 18 min 17 min 180 min 150 psi 50 psi Time (min) Temp. (C) Pressure (psig) rpm Stirring Notes 0 20 340 716 TO 68 390 713 20 113 440 709 30 164 460 706 40 214 500 719 50 265 550 719 60 314 640 719 70 364 790 716 77 400 950 719 Reaction Temp. 257 400 1100 721 Air Cooling Started 262 344 900 724 267 298 800 724 272 267 740 725 275 250 700 724 Water Cooling Started 277 131 500 728 292 18 390 728 Water Cooling Stopped 155 Appendix 3.4 6. GC Results: Reactor Feed Sample: 1 2 3 Average CIO 868359 863831 840645 857612 Decalin 37992928 38119872 37226146 37779649 DPM 5782060 5864828 5725924 5790937 ADPM/ACIO 6.659 6.789 6.811 6.753 ADECALIN/ACIO 43.753 44.129 44.283 44.055 Reaction Product: 1 2 3 Average Benzene 308662 315151 297559 307124 Toluene 358091 365047 342429 355189 C10 778496 793545 750393 774145 Decalin 28290020 28735905 27415395 28147107 DPM 3883433 3941921 3759559 3861638 ABENZENE/ACIO 0.396 0.397 0.397 0.397 ATOLDENE/ACIO 0.46 0.46 0.456 0.459 ADPM/ACIO 4.988 4.967 5.01 4.989 ADECALW/ACIO 36.339 36.212 36.535 36.362 7. GC Analysis: Calibration Curves: Mi/MCIO = (l/m)*Ai/AC10 - c/m m c Benzene 0.5931 -0.0032 Toulene 0.6971 -0.0057 DPM 1.4872 -0.8807 Feed Sample Reactor Feed Reactor Product Vol. CIO (ml) 0.139 3.611 Moles CIO 0.00071 0.01853 Moles Benzene 0.01249 Moles Toulene 0.01234 Moles DPM by GC 0.00366 0.08559 0.07312 Theoretical Moles DPM 0.00331 0.08595 % Conversion (Benz. ref.) 14.6 % Conversion (Tol. ref.) 14.4 % Conversion (DPM ref.) 14.6 156 Appendix 3.4 Experiment #20 Summary: 1. Reactor Charge : Recovered C0S2 @ 1000 ppm Co Sulphiding Time @ 1000 psig Reaction Time Reaction Temperature Total Volume 130 ml 360 psig LP 0.523 g 2h 3h 400 °C 135 ml Reactant Mixture 5% H 2S in H 2 Recovered C0S2 (120 ml Decalin, 15.015g DPM) 2. Reactor Preparation: 2.1. Reactor /internals cleaned with soap and water, then rinsed 2.2. Reactor and internals scoured with 'Brasso' 2.3. Reactor rinsed with hexane (2 hours with stirring) 3. Nitrogen Purging: Reactor purged with N 2 before and after run. 4. Pressure Testing: Reactor pressure tested @ 1000 psig for 2 h during Sulphiding No pressure decrease observed 5. Reaction Profiles: Time to reach 400 °C 77 min Air cooling time 400 °C to 18 min 250 °C Water cooling Time 250 °C to 15 °C 17 min Total Reaction Time 180 min Pressure Rise @ 400 °C 100 psi Overall Pressure Rise 0 psi Time (min) Temp. (°C) Pressure (psig) rpm Stirring Notes 0 18 360 715 10 70 440 718 20 120 490 715 30 170 510 714 40 220 560 714 50 270 640 714 60 320 740 711 70 369 860 713 77 400 1000 716 Reaction Temp. 257 400 1100 718 Air Cooling Started 262 342 900 726 267 306 800 726 272 267 750 726 275 250 700 728 Water Cooling Started 277 129 500 728 292 19 360 729 Water Cooling Stopped 157 Appendix 3.4 6. GC Results: Reactor Feed Sample: 1 2 3 Average CIO 853651 879035 848613 860433 Decalin 37347250 38699657 37153824 37733577 DPM 5693173 5965816 5653750 5770913 ADPM/ACIO 6.669 6.787 6.662 6.706 ADECAUN/ACIO 43.75 44.025 43.782 43.852 Reaction Product: 1 2 3 Average Benzene 50483 53714 52201 52133 Toluene 58562 62974 61311 60949 C10 779895 830132 800811 803613 Decalin 28758832 30338733 29283992 29460519 DPM 4517299 4894702 4659430 4690477 ABENZENE/ACIO 0.065 0.065 0.065 0.065 ATOLUENE/ACIO 0.075 0.076 0.077 0.076 ADPM/ACIO 5.792 5.896 5.818 5.836 ADECAUN/ACIO 36.875 36.547 36.568 36.663 7. GC Analysis: Calibration Curves: Mi/MCIO = (l/m)*Ai/AC10 - c/m m c Benzene 0.5931 -0.0032 Toulene 0.6971 -0.0057 DPM 1.4872 -0.8807 Feed Sample Reactor Feed Reactor Product Vol. CIO (ml) 0.139 3.611 Moles CIO 0.00071 0.01853 Moles Benzene 0.00213 Moles Toulene 0.00217 Moles DPM by GC 0.00364 0.08561 0.08367 Theoretical Moles DPM 0.00331 0.08595 % Conversion (Benz. ref.) 2.5 % Conversion (Tol. ref.) 2.5 % Conversion (DPM ref.) 2.3 158 Appendix 3.4 Experiment #21 Summary: Co Naphthenate @ 1000 ppm Co, No DPM Sulphiding Time @ 1000 psig 2 h Reaction Time 3 h Reaction Temperature 400 °C Total Volume 135 ml 1. Reactor Charge : 2. Reactor Preparation: 3. Nitrogen Purging: 4. Pressure Testing: 5. Reaction Profiles: Time to reach 400 °C Air cooling time 400 °C to 250 °C Water cooling Time 250 °C to Total Reaction Time Pressure Rise @ 400 °C Overall Pressure Rise 120 ml Decalin 340 psig LP 5%H 2 SinH 2 1.475 g Cobalt Naphthenate (8% Co) 2.1. Reactor /internals cleaned with soap and water, then rinsed 2.2. Reactor and internals scoured with 'Brasso' 2.3. Reactor rinsed with hexane (2 hours with stirring) Reactor purged with N 2 before and after run. Reactor pressure tested @ 1000 psig for 2 h during Sulphiding 40 psi pressure decrease observed (leak in thermocouple well) 15 °C 75 min min min 180 min 250 psi 100 psi Time (min) Temp. (°C) Pressure (psig) rpm Stirring Notes 0 22 340 708 11 83 420 710 31 183 500 719 41 233 540 716 51 284 600 717 61 333 700 713 71 381 890 709 75 400 990 717 Reaction Temp. 255 400 1240 724 Air Cooling Started - 20 440 -6. GC Results: Negligible Benzene/Toluene in reactor product Therefore B/T products in Experiment #19 due to catalytic cracking of DPM with sulphided Co catalyst 159 Appendix 3.4 Experiment #22 Summary: Co Microemulsion cracking of DPM Sulphiding Time @ 1000 psig 6 h Reaction Time 3 h Reaction Temperature 400 °C Total Volume 135 ml 1. Reactor Charge: 120 ml Unreduced Co/PELE/Decalin 15 ml Diphenylmethane 360 psig IP 5%H 2 SinH 2 2. Reactor Preparation: 2.1. Reactor /internals cleaned with soap and water, then rinsed 2.2. Reactor and internals scoured with 'Brasso' 2.3. Reactor rinsed with hexane (2 hours with stirring) 3. Nitrogen Purging: Reactor purged with N 2 before and after run. 4. Pressure Testing: Reactor pressure tested @ 1000 psig for 6 h during Sulphiding 50 psi pressure decrease observed 5. Reaction Profiles: Time to reach 400 °C 75 min Air cooling time 400 °C to - min 250 °C Water cooling Time 250 °C to 15 °C - min Total Reaction Time 180 min Pressure Rise @ 400 °C 160 psi Overall Pressure Rise -60 psi Time (min) Temp. (°C) Pressure (psig) rpm Stirring Notes 0 18 360 660 10 76 400 711 20 130 450 710 32 189 490 702 40 228 540 769 50 278 590 767 60 328 660 768 70 378 790 769 75 400 890 775 Reaction Temp. 255 400 1050 779 Air Cooling Started - 17 300 -160 Appendix 3.4 6. GC Results: Reactor Feed Sample: No GC analysis possible due to Co in Microemulsion Reaction Product: 1 2 3 Average Benzene 14005 16921 13434 14787 Toluene 24269 30230 24178 26226 CIO 1296257 1611836 1262819 1390304 Decalin 44285884 51866371 41777247 45976501 DPM 5618212 6921936 5357194 5965781 ABENZENE/ACIO 0.011 0.010 0.011 0.011 ATOLUENE/ACIO 0.019 0.019 0.019 0.019 ADPM/ACIO 4.334 4.294 4.242 4.290 ADECAUN/ACIO 34.164 32.178 33.083 33.142 7. GC Analysis: Calibration Curves: Mi/MCIO = (l/m)*Ai/AC10 - c/m m c Benzene 0.5931 -0.0032 Toulene 0.6971 -0.0057 DPM 1.4872 -0.8807 Feed Sample Reactor Feed Reactor Product Vol. C10 (ml) - 3.611 Moles C10 - 0.01853 Moles Benzene - 0.00043 Moles Toulene - 0.00065 Moles DPM by GC - - 0.06442 Theoretical Moles DPM - 0.08925 % Conversion (Benz. ref.) 0.5 % Conversion (Tol. ref.) 0.7 % Conversion (DPM ref.) 27.8 161 Appendix 3.4 Experiment #23 Summary: 1. Reactor Charge : Recovered C0S2 @ 2000 ppm Co Sulphiding Time @ 1000 psig Reaction Time Reaction Temperature Total Volume 130 ml 350 psig LP 0.4017 g 2h 3h 400 °C 135 ml Reactant Mixture (120 ml Decalin, 15.015g DPM) 5% H 2S in H 2 Recovered C0S2 2. Reactor Preparation: 3. Nitrogen Purging: 4. Pressure Testing: 5. Reaction Profiles: 2.1. Reactor /internals cleaned with soap and water, then rinsed 2.2. Reactor and internals scoured with 'Brasso' 2.3. Reactor rinsed with hexane (2 hours with stirring) Reactor purged with N 2 before and after run. Reactor pressure tested @ 1000 psig for 2 h during Sulphiding No pressure decrease observed Time to reach 400 °C Air cooling time 400 °C to 250 °C Water cooling Time 250 °C to 15 °C Total Reaction Time Pressure Rise @ 400 °C Overall Pressure Rise 75 min 17 min 23 min 180 min 50 psi n.a. psi Time (min) Temp. (°C) Pressure (psig) rpm Stirring Notes 0 19 n.a. (est. 350) 728 Pressure gague not 10 75 n.a. 739 Working 20 124 n.a. 742 30 174 n.a. 741 40 224 n.a. 740 50 274 n.a. 739 60 324 n.a. 737 70 374 n.a. 736 75 400 925 741 Reaction Temp. 255 400 975 752 Air Cooling Started 260 342 975 758 265 299 970 758 270 261 950 760 272 250 950 Water Cooling Started 275 101 650 295 17 400 Water Cooling Stopped 162 Appendix 3.4 6. GC Results: Reactor Feed Sample: 1 2 3 Average CIO 944192 949507 965048 952916 Decalin 45083965 44402572 45260195 44915577 DPM 4280117 4423399 4517067 4406861 ADPM/ACIO 4.533 4.659 4.681 4.624 ADECAUN/ACIO 47.749 46.764 46.899 47.137 Reaction Product: 1 2 3 Average Benzene 22228 21217 20300 21248 Toluene 28745 27768 26968 27827 C10 1040431 993293 949890 994538 Decalin 44058125 42510483 41006977 42525195 DPM 4598841 4396039 4215317 4403399 ABENZENE/ACIO 0.021 0.021 0.021 0.021 ATOLUENE/ACIO 0.028 0.028 0.028 0.028 ADPM/ACIO 4.420 4.426 4.438 4.428 ADECAUN/ACIO 42.346 42.798 43.170 42.771 7. GC Analysis: Calibration Curves: Mi/MCIO = (l/m)*Ai/AC10 - c/m m c Benzene 0.5931 -0.0032 Toulene 0.6971 -0.0057 DPM 1.4872 -0.8807 Feed Sample Reactor Feed Reactor Product Vol. CIO (ml) 0.139 3.611 Moles CIO 0.00071 0.01853 Moles Benzene 0.00077 Moles Toulene 0.00090 Moles DPM by GC 0.00264 0.08661 Theoretical Moles DPM 0.00331 0.08595 % Conversion (Benz. ref.) 0.9 % Conversion (Tol. ref.) 1.0 % Conversion (DPM ref.) 4.2 163 Appendix 3.5 Appendix 3.5 : GC traces of experiment #19 and decomposition of Co naphthenate t/i 8 Retention time (min) GC trace of Experiment #19 : Reaction product of activity test using Co Naphthenate (CIO is the n-decane GC internal standard). rji 8 10 Retention time (min) GC trace of the product of the decomposition of Co Naphthenate in decalin at 400 °C (LP 5 % H2S in H 2 = 300 psig, reaction time = 3 hours). 164 Appendix 3.6 Appendix 3.6 : XRD spectra of the spent catalysts from the second set of activity measurements. XRD spectrum of spent catalyst from experiment #19 (Co Naphthenate precursor). 165 Appendix 3.6 200 150 H 0 -I 1 1 1 1 1 1 0 10 20 30 40 50 60 2 0 X R D spectrum of spent catalyst from experiment #20 (raw data). X R D spectrum of spent catalyst from experiment #20 (baseline sutracted). 166 Appendix 3.6 167 

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.831.1-0058613/manifest

Comment

Related Items