UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Iron promoted activated alumina for scavenging free oxygen in claus converters Bedrossian, Sevan 2004

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

Item Metadata

Download

Media
831-ubc_2004-0165.pdf [ 10.92MB ]
Metadata
JSON: 831-1.0058761.json
JSON-LD: 831-1.0058761-ld.json
RDF/XML (Pretty): 831-1.0058761-rdf.xml
RDF/JSON: 831-1.0058761-rdf.json
Turtle: 831-1.0058761-turtle.txt
N-Triples: 831-1.0058761-rdf-ntriples.txt
Original Record: 831-1.0058761-source.json
Full Text
831-1.0058761-fulltext.txt
Citation
831-1.0058761.ris

Full Text

IRON P R O M O T E D A C T I V A T E D A L U M I N A F O R S C A V E N G I N G F R E E O X Y G E N IN C L A U S C O N V E R T E R S by S E V A N B E D R O S S I A N B . A . S c , The U n i v e r s i t y o f Science & Techno logy , Tehran , 1996 A THESIS S U B M I T T E D IN P A R T I A L F U L F I L L M E N T O F T H E R E Q U I R E M E N T S F O R T H E D E G R E E O F M A S T E R O F APPLIED S C I E N C E i n T H E F A C U L T Y O F G R A D U A T E STUDIES D E P A R T M E N T O F C H E M I C A L & B I O L O G I C A L ENGINEERING W e accept this thesis as conforming to the required standard T H E U N I V E R S I T Y O F B R I T I S H C O L U M B I A M a r c h , 2004 © Sevan Bedross ian , 2004 Library Authorization In presenting this thesis in partial fulfillment 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. SEVAN BEDROSSIAN 31/3/2004 Name of Author (please print) Date (dd/mm/yyyy) Title of Thesis: IRON PROMOTED ACTIVATED ALUMINA FOR SCAVENGING F R E E OXYGEN IN CLAUS C O N V E R T E R S Degree: M.A.Sc Department of Chemical & Biological Engineering The University of British Columbia Vancouver, BC Canada Year: 2004 Abstract The Claus process is used to remove H2S from acid gas streams and is widely practiced in industry. The most common Claus catalyst in sulphur recovery units is activated alumina (y-AI2O3). However, it is well documented that excessive free oxygen in the Claus process stream deactivates the alumina catalyst. 0 2 surface reactions with S 0 2 lead to sulphation of the alumina surface, thereby inhibiting the conversion of H2S and CS2. One approach to minimizing deactivation by sulphate formation is to use a dual bed reactor configuration, where part of the conventional alumina catalyst is replaced with a catalyst that w i l l promote oxygen consumption and thereby limit sulphate formation. In the present study, Fe catalysts supported on alumina have been examined for oxygen removal. The catalysts were prepared by impregnating a commercial alumina support with solutions of Fe (NO3) 3.9H2O. Catalysts were characterized using X R D (X-ray Diffraction Spectroscopy) and D R I F T (Fourier Transmission Infrared Spectroscopy) as well as B E T surface area measurement. The stable phase of Fe on the alumina in the relatively rich H 2 S reaction gas, typical of Claus process streams, was verified to be FeS2- The effect of catalyst pre-treatment conditions on oxygen removal kinetics is also reported. Experiments showed that 0 2 consumption occurred on Fe 2 03 sites based on the reaction (a) and on FeS2 sites based on the sequential reaction (a) and (b). (a) H2S + ^-02^-Sn+H20 2 n (b) S + 02 -+S02 i i It was also concluded that the FeS 2/Y-Al 2C>3 catalyst is bifunctional due to activity of the alumina support towards the Claus reaction. Further study o f the O2 removal kinetics was therefore conducted on bulk FeS2 without support. B y assuming the stoichiometric order for reactants, the reaction rate constants on bulk FeS2 were determined using the H Y S Y S simulator and the rate of reaction (b) faster than reaction (a). Hence, FeS 2 provides an overall faster rate of O2 consumption than Fe2C>3. Also , a mechanism was postulated and the rate limiting steps that were consistent with the rate data were identified. i i i Table of contents Abstract ii Table of contents iv List of Tables ix List of Figures x List of Schemes xvii A C K N O W L E D G M E N T S xviii Chapter 1 Introduction 1.1 Background 1 1.1.1 Sulphur Recovery and Hydrogen Sulphide Conversion 1 1.1.2 Claus Process Description 1 1.1.3 Deactivation of Claus Catalyst 4 1.2 Motivation for the Study 6 1.3 Objectives of the Present Study 7 Chapter 2 Literature Review 2.1 Deactivation 9 2.1.1 Sulphur Fouling - Sulphur Deposition 9 2.1.2 Coke Deposition - Carbon Deposition 10 2.1.3 Loss in Surface Area - Thermal Aging 10 2.1.4 Sulphation 11 2.2 Effect of Oxygen on Sulphate Formation 12 2.3 Sulphate Formation on Active Alumina in Absence of Oxygen 21 2.4 Sulphation in Industrial Processes 22 2.4.1 Oxygen Leakage -Halan-Robb Gas Refinery Plant (Ontario) 22 2.4.2 Oxygen Leakage and Deactivation -iv Ram River Gas Plant (Alberta) 24 2.5 Industrial Oxygen Scavengers 25 2.6 Superclaus Process and the Role of Iron Promoted Metal Oxide Catalysts 27 2.7 Supported Metal Catalyst Preparation 30 2.7.1 Introduction 30 2.7.2 Impregnation Methods 31 2.7.3 Ion Exchange .32 2.7.4 Drying 32 2.7.5 Calcination 33 2.8 Preparation and Preconditioning of Fe/Al 203 Catalyst 33 2.9 Infrared Spectroscopy of the Surface of Metal Oxides Using Adsorbed, Probe Molecules 35 2.10 Summary 38 Chapter 3 Experimental Apparatus and Methods Safety note 40 3.1 Catalyst Preparation (Fe203/Al203 and FeS 2 /Al 2 0 3 ) 40 3.2 Catalyst Characterization 41 3.2.1 X R D (X-ray Diffraction) 41 3.2.2 Catalyst Crystallite Size Measurement 42 3.2.2 DRIFT (Fourier Transmission Infrared Spectroscopy) 46 3.2.4 B E T (Total Surface Area Measurement) 48 3.2.5 S E M (Scanning Electron Microscopy) 49 3.3 Experimental Setup for Catalyst Activity Test and Reaction Rate Study 50 3.3.1 Setup 50 3.3.2 Experimental Procedures 55 3.3.3 Calculation and Minimization of Transport Disguises 56 3.3.3.1 Temperature Gradients 57 3.3.3.2 Mass Transfer Gradients 57 v 3.3.4 Reactor Operation Regime 58 Chapter 4 Results and Discussion 4.1 Introduction 59 4.2 Porocel Fe-based Catalyst Performance Test 60 4.3 Characterization 70 4.3.1 Phase Characterization of Fe on AI2O3 using X R D 70 4.3.2 Phase Characterization of Fe on A1 2 0 3 using DRIFT 77 4.3.3 Study the Stability of the Catalyst using X R D 80 4.4 Role of FeS 2 (Metal Sulphide) and A1 2 0 3 (Catalyst Support) 81 4.5 Summary 84 Chapter 5 Reaction Kinetics and Mechanism 5.1 Introduction 86 5.2 Characterization of Active Iron Sulphide 86 5.3 Kinetic Study 89 5.3.1 Preliminary Test 89 5.3.2 Catalytic Activity Test 90 5.3.2.1 Influence of O2 Concentration 91 5.3.2.2 Influence of H 2 S Concentration 92 5.3.2.3 Influence of H 2 0 Concentration 96 5.3.2.4 Influence of Temperature 99 5.4 Reaction Kinetics 100 5.4.1 Data Analysis using Process Simulator 100 5.4.2 Optimization Method 102 5.4.3 Results 103 5.5 Mechanism of the Reaction 106 5.5.1 Adsorption 106 5.5.2 Surface Reactions 107 vi 5.6 Summary I l l Chapter 6 Conclusion and Recommendations 6.1 Conclusions 112 6.2 Recommendations 113 6.3 Future Work 114 Nomenclature 116 References 119 Appendices APPENDIX A S I M A D Z U GC-14A Operation 125 A.1 Carrier Gas Flow 125 A.2 T C D Temperature and Current 125 A.3 Temperature Program 126 A.4 Time Program 127 A.5 Peak Processing Parameters 127 APPENDIX B Summary of Porocel Fe-based Catalyst Performance Test 129 APPENDIX C Summary of Role of FeS 2 (Metal Sulphide) and A1 2 0 3 (Catalyst Support) 134 APPENDIX D Summary of Catalytic Activity Test (Pure FeS) 136 D . l Mass Flow Controllers & Feed Composition 136 D.2 Calibration and Base Area 138 D.3 Surface Area of 0 2 141 D.4 Reaction Kinetics Summary 143 APPENDIX E Summary of Statistical Analysis 158 v i i E . l Data Points Construction E.2 Goodness of the Model List of Tables Table 1.1 Typical stream composition in the Claus process; C O D = sulphur condenser; C N V = catalytic converter; concentration of gases are reported as mol/total mol; at the end of the table %H2S = percent mol per total mol concentration of the H2S at each stage 5 Table 4.1 The actual flow condition (% vol.) 61 Table 4.2 Activity of Fe-1 catalyst; 360 cc (STP) /min; 250°C; 1.5 g catalyst 61 Table 4.3 Activity of Fe-2 catalyst; 360 cc (STP) /min; 250°C; 1.5 g catalyst 61 Table 4.4 Activity of Fe-3 catalyst; 360 cc (STP) /min; 250°C; 1.5 g catalyst 61 Table 5.1 Experimental conditions used for studying the rate of O2 consumption; Weight = weight of pure FeS 89 Table 5.2 Reaction rate constants for each set of experimental data using H Y S Y S optimizer 103 Table 5.3 Summary of the statistical analysis conducted on the results in the range of experiments 103 List of Figures Figure 1.1 Typical Claus sulphur recovery unit's configuration 2 Figure 2.1 IR spectra of pure oxides: (a) ZrC>2, (b) T i 0 2 and (c) A1 2 0 3 after sulphation by S0 2 +0 2 at 723 K . (Lavalley etal., 2000) 14 Figure 2.2 Role of the Oxygen level on sulphur yield; feed contains 3 vol.% H 2 S, 1.5 vol.% S 0 2 ,30 vol.% H 2 0 , balance N 2 , 500 and 2000 ppm 0 2 at 250°C for 3 s contact time (Nedez et al., 1997) 18 Figure 2.3 IR spectra of promoted alumina after sulphation by S0 2 +0 2 at 723 K , (a) A1 2 0 3 ; (b) V/A1 2 0 3 ; (c) W/A1 2 0 3 ; (d) M o / A l 2 0 3 . (Laperdix et al., 2000). 19 Figure 2.4 IR spectra of promoted alumina after sulphation by S0 2 +0 2 at 723 K, (a) A1 2 0 3 ; (b) Na/Al 2 0 3 ; (c) C u / A l 2 0 3 ; (d) Fe /Al 2 0 3 ; (e) N i /A l 2 0 3 . (Laperdix etal., 2000) 19 Figure 2.5 Variation of sulphate species (relative to intensity of the band near 1380 cm'1) after introduction of H 2 S at increasing temperature, on A1 2 0 3 , Z r 0 2 and promoted aluminas. (Laperdix et al., 2000) 20 Figure 2.6 In-line burner performance of Gulf Canada Corporation's Hanlan-Robb, Ontario gas processing plant. Oxygen leakage has been shown before and after optimization. (Johnson et al., 1987) 23 x Figure 3.1 X - ray scattered by atoms in an ordered lattice interfere constructively in directions given by Bragg's law. Diffractograms are measured as a function of the angle 26. Rotation of the sample and source during the measurement enhances the number of particles that contribute to diffraction 42 Figure 3.2a Siemens thin film diffractometer; 29 = 3" (A: X-ray Source, B: Sample Holder and C: X-ray Detector) 43 Figure 3.2b Siemens thin film diffractometer; 28 = 70° 44 Figure 3.3 Diffraction pattern of the nanocrystalline iron sulphide on alumina support catalyst with 15% iron loading 45 Figure 3.4 A : Reaction Chamber; direction of reactants stream is specified on the picture 47 Figure 3.5 DRIFT setup used for catalyst surface study; C O , S 0 2 / N 2 and O2/N2 were introduced into the device using two separate mass flow controller; He and N 2 were introduced using two separate rotameter 48 Figure 3.6 Experimental setup for catalyst activity tests and reaction rate study 51 Figure 3.7 Sulphur trap and water drier; Quartz wool was used to trap the sulphur mist and prevent any sulphur carry over into the G C columns 52 Figure 3.8 Sulphur Trap 52 Figure 3.9 Water Drier 53 Figure 3.10a Valco valve before injection; loop sampling with two columns sequence reversal; Column 1 is Molecular Sieves 5A and column 2 is Poropak Q 54 Figure 3.10b Valco valve; injection position 54 Figure 4.1 Comparison of oxygen scavenging using Fe-2 and Fe-3 with and without preconditioning 63 Figure 4.2 Porocel iron oxide catalyst (Fe-3) before reaction 64 Figure 4.3 Porocel iron oxide catalyst (Fe-3) after reaction showed the presence of different phase of iron with black color on the catalyst 64 Figure 4.4 Porocel catalyst (Fe-3) after preconditioning showed that iron oxide phase in the bright red color transformed to black color 65 Figure 4.5 O2 scavenging activity over iron sulphide supported catalyst vs. Time; feed composition: 1.5vol. % H 2 S, 2.2 vol. %0 2 and 96.3 vol. % N 2 at 300°C temperature and 115 kPa pressure 66 Figure 4.6 0 2 scavenging activity over iron oxide supported catalyst vs. Time; feed composition: 1.5vol. % H 2 S, 2.2 vol. % 0 2 and 96.3 vol. % N 2 at 300°C temperature and 115 kPa pressure 67 Figure 4.7 Average Yield % of S and S 0 2 over iron sulphide supported catalyst vs. Time; feed composition: 1.5vol. % H 2 S, 2.2 vol. % 0 2 and 96.3 vol. % N 2 at 300°C temperature and 115 kPa pressure 68 Figure 4.8 Average Yield % of S and S 0 2 over iron oxide supported catalyst vs. Time; feed composition: 1.5vol. % H 2 S, 2.2 vol. % 0 2 and 96.3 vol. % N 2 at 300°C temperature and 115 kPa pressure 69 Figure 4.9 Average H 2 S conversion over iron oxide and iron sulphide catalysts vs. Time; feed composition: 1.5vol. % H 2 S, 2.2 vol. % 0 2 and 96.3 vol. % N 2 at 300°C temperature and 115 kPa pressure 70 Figure 4.10 X R D data for alumina supported iron oxide catalyst after calcination at 673 K in oxygen present environment; loading of the iron was 15 wt. % on the catalyst 72 Figure 4.11 X R D data for alumina supported iron oxide (15 wt. % Fe) after preconditioning in H 2 S / H 2 / N 2 stream at 673K; transformation of iron oxide species on the support to iron mono-sulphide 72 Figure 4.12 X R D data for alumina supported iron oxide (15 wt. % iron) catalyst after preconditioning in H 2 S / N 2 stream at 673K; Traces of S» (orthorhombic structure) fouling in the catalyst 74 Figure 4.13 Reduction-Sulphidation reaction guideline simulated using ASPEN PLUS software 76 Figure 4.14 Pattern of C O absorption on the four types of catalyst using DRIFT 78 Figure 4.15 Structure of Fe 203 and location of basic O 2" (blue = Fe; red = O) .79 Figure 4.16 Structure of FeS04 and location of basic O 2" (blue = Fe; red = O; yellow = S) 79 x i i i Figure 4.17 Alumina supported FeS catalyst after 4 hours reaction in H 2 S / 0 2 / N 2 (3/4.5/92.5 vol. %) feed stream 80 Figure 4.18 Performance of alumina support under Claus condition vs. Time; feed composition: 4 vol. % H 2 S, 2 vol.% S 0 2 and 1 vol.% 0 2 , balance N 2 at 300°C and 115 kPa 82 Figure 4.19 Performance of FeS 2/Al 203 catalyst under Claus condition in presence and absence of the 0 2 ; (see text for experimental conditions) 83 Figure 5.1 X R D data from bulk iron mono-sulphide (FeS) 87 Figure 5.2 X R D pattern of the FeS (5% Fe loading) on the A1 2 0 3 support 87 Figure 5.3 Surface of bulk FeS as seen by S E M 88 Figure 5.4 Effect of 0 2 feed inlet concentration on 0 2 reaction rate; feed composition was 6 vol. % H 2 S and 20% H 2 0 , balance N 2 ; Temperature was 270°C and pressure was 115 kPa 91 Figure 5.5 Effect of 0 2 inlet concentration on S 0 2 average selectivity and H 2 S average conversion on pure FeS 2 catalyst 92 Figure 5.6 Effect of H 2 S inlet concentration on 0 2 reaction rate; feed composition was 2 vol. % 0 2 and 20% H 2 0 at balance N 2 ; Temperature and pressure were 270°C and 115 kPa 93 Figure 5.7 Effect of H 2 S inlet concentration on S 0 2 average selectivity and H 2 S average conversion; feed composition was 2 vol. % 0 2 and 20% H 2 0 at balance N 2 ; Temperature and pressure were 270°C and 115 kPa 94 xiv Figure 5.8 Effect of H2S inlet concentration on O2 reaction rate. Feed composition was 2 vol. % O2 and 20% H2O at balance N2; Temperature was 280°C and pressure was 115 kPa 95 Figure 5.9 Effect of H2S inlet concentration on SO2 average selectivity and H 2 S average conversion; feed was 2 vol. % 0 2 and 20% H2O, balance N2; Temperature was 280°C and pressure was 115 kPa 95 Figure 5.10 Effect of presence and absence of water in the feed on O2 average reaction rate versus 0 2 concentration 96 Figure 5.11 Effect of presence and absence of water in the feed on SO2 average selectivity and H 2 S conversion versus O2 concentration....97 Figure 5.12 Effect of H2O concentration on O2 reaction rate in the feed contains 6 vol. % H 2 S and 2 vol. % O2 at balance N2; Temperature was 280°C and pressure was 115 kPa 98 Figure 5.13 Effect of H2O concentration on SO2 average selectivity and H 2 S conversion; feed was 6 vol. % H2S and 2 vol. % O2 at balance N 2 ; Temperature and pressure were 270°C and 115 kPa 99 Figure 5.14 Effect of temperature on O2 average reaction rate; feed was 6 vol. % H 2 S, 2 vol. % 0 2 and 20% H 2 0 , balance N 2 100 Figure 5.15 Experimental and simulated data; Effect of 0 2 feed inlet concentration on 0 2 and H 2 S conversion as well as SO2 selectivity 104 x v Figure 5.16 Experimental and simulated data; Effect of H2S feed inlet concentration on O2 and H 2 S conversion as well as SO2 selectivity 105 Figure 5.17 Experimental and simulated data; Effect feed temperature on 0 2 and H2S conversion as well as SO2 selectivity 105 x v i List of Schemes Scheme 2.1 Sulphate species in absence of water 13 Scheme 2.2 Sulphate species in presence of water 13 Scheme 2.3 Proposed mechanisms of sulfur formation on an alumina Claus catalyst. (Clark et al., 2000) 21 Scheme 2.4 Formation of surface carbonate 37 Scheme 2.5 Adsorption of C O on the surface of metal oxide 37 Scheme 4.1 Claus reaction happens on the support and active site but oxygen scavenging happens only on active iron sulphide site 84 xv i i ACKNOWLEDGMENTS I would like to thank my supervisor Dr. Kev in J. Smith for his support and guidance throughout my research. I also want to extend my appreciation to Dr. Ed . Luinstra from Porocel who assisted me during the initial stages of my research. M y sincere thanks to Dr. Chang-Chun Y u from Petroleum University of Beijing, China, for his assistance and instruction on D R I F T measurements. Special thanks to Dr. Dusco Posarac for his valuable guidance on using the H Y S Y S simulator. The financial support of N S E R C is also gratefully acknowledged. I would also like to thank the fellow members of my research group for assisting me with my experimental work, and for creating a pleasant work environment. M y sincere thanks, for her love, understanding and patience, to my wife, Armineh, and my gratuities for their kindness and encouragement, to my parents. In closing, I would like to dedicate this thesis to my brother, Ara , for all the support, advice and encouragement which he has always given me over the years. xv i i i Chapter 1 - Introduction Chapter 1- Introduction 1.1 Background 1.1.1 Sulphur Recovery and Hydrogen Sulphide Conversion Sulphur recovery refers to the conversion of hydrogen sulphide (H2S) to elemental sulphur (Ss). H2S is a by-product of natural gas processing and crude oi l refining in hydrodesulphurisation units (Goar, 1986). Yearly tonnages of H2S can be derived from recovered sulphur production figures: 14.4 mill ion ton from sour gas and 9.6 mil l ion ton from oi l refineries, worldwide (Phipps, 1993). Recovered sulphur accounts for 61 percent of total world sulphur production with an approximate growth rate of 1 percent annually (Morris et al., 1999). The most common H2S conversion method used in industry is the Claus process. Approximately 90 to 95 percent of recovered sulphur is produced by the Claus process. This process typically recovers 95 to 97 percent of the H2S in the feed stream. 1.1.2 Claus Process Description H2S, a by-product of crude o i l and natural gas processing, is recovered and converted to elemental sulphur by the Claus process. This process was invented in 1883 by Carl Friedrich Claus. During the 1930's the German chemical industry improved the original Claus process to its current state. Figure 1.1 shows a typical Claus sulphur recovery unit. The process consists of 1 Chapter 1 - Introduction multistage catalytic oxidation of H2S according to the following two reactions (Duncan, 1997; Capone, 1997). 3 H2S + - 0 2 -» S02 +H20 + heat Reaction 1.1 2H2S + S02+±-Sn+ 2H20 (n ~ 8) Reaction 1.2 The Claus reaction involves burning one-third of the H2S with air in a reactor furnace to form sulphur dioxide (SO2) according to Reaction 1.1. Air Boiler Furnace Claus Gas Converter 1 Condenser 1 Converter 2 Condenser 2 Converter 3 Condenser 3 Tail Gas rr Condenser 4 Liquid Sulphur Figure 1.1 Typical Claus sulphur recovery unit's configuration H 2 S concentration in the feed gas may vary between a minimum of 5 vol . % and maximum of 100 vol . %, normal variations being in the range of 10 to 90 vol . %. A challenge occurs i f several gas streams from different sources are to be processed and the Claus process has to handle both extremely low and extremely high hydrogen sulphide levels. The Claus process has 2 Chapter 1 - Introduction to be operated with pure oxygen in the first case (lean H2S) and air in the second case (rich H2S) and mixtures of pure oxygen and air under normal operating conditions (Lurgi Oel , 2003). The furnace normally operates at combustion chamber temperatures ranging from 980 to 1540°C with pressure rarely higher than 70 kPa. Before entering a sulphur condenser, hot gas from the combustion chamber is quenched in a waste heat boiler that generates high to medium pressure steam. About 80% of the heat released can be recovered as useful energy. Liquid sulphur from the condenser runs through a seal leg into a covered pit from which it is pumped to trucks or railcars for shipment to end users. Approximately 65 to 70% of the sulphur is recovered in this stage. Principal side reactions in the furnace include formation of C O S and CS2 from residual hydrocarbons present in the feed gas. The cooled gases exiting the condenser are then sent to the catalytic conversion stage o f the process. The remaining uncombusted r L S undergoes the Claus reaction (reacts with SO2) to form elemental sulphur according to Reaction 1.2. The catalytic reactors operate at lower temperature than furnace chamber, ranging from 200 to 350°C. Because the reaction is an equilibrium limited reaction, it is not possible for a Claus plant to convert all incoming sulphur compounds to elemental sulphur in a single stage catalytic converter. Therefore, two or more stages are used in series to recover the sulphur. Typically C O S and CS2 are also hydrolyzed in the first catalytic converter according to Reaction 1.3 and Reaction 1.4 (Schoofs, 1985). COS + H20<r> H2S + C02 ' Reaction 1.3 CS2+2H20<^2H2S + C02 Reaction 1.4 3 Chapter 1 - Introduction The first catalytic converter typically operates in the range of 270 to 350°C in order to hydrolyze C O S and CS2. The second and subsequent converters usually operate just above the dew point of sulphur vapour. The typical temperatures for second and third converters are 220 to 250°C and 195 to 210°C, respectively. They operate at decreasing temperatures in order to obtain a high sulphur yield (Nedez et al., 1996). If the sulphur recovery unit is located in a natural gas processing plant, the type o f reheat employed is typically either auxiliary burners or heat exchangers, with steam reheat being used occasionally. If the sulphur recovery unit is located in a crude o i l refinery, the typical reheat scheme uses 3000 to 5000 kPa steam. It is estimated that 95 to 97 percent overall recovery can be achieved depending on the number of catalytic reaction stages and reheating method used. Table 1.1 shows a typical stream composition through a Claus unit (Johnson et al, 1987). The tail gas, containing H2S, S 0 2 , sulphur vapour, and traces of other sulphur compounds formed in the combustion section escape with inert gases from the tail end of the plant. Thus, it is frequently necessary to follow the Claus unit with a tail gas clean up unit to achieve higher recovery. 1.1.3 Deactivation of Claus Catalyst The most widely used Claus catalyst in sulphur recovery units is non-promoted, spherical, activated alumina (J-AI2O3). Catalyst deactivation is the most frequent cause o f loss of Claus process performance. Deactivation is caused either by sulphur fouling, sulphation, coke deposition, or loss in surface area due to thermal sintering. 4 Chapter 1 - Introduction mol/mol Acid Gas COD 1 CNV1 COD 2 CNV 2 COD 3 CNV 3 COD 4 Outlet Inlet Outlet Inlet Outlet Inlet Outlet H 2S 100 Warn 35.69 38.94 8.37 9.18 1.91 3.82 1.69 so 2 0 m m 17.63 19.47 3.74 4.3 1.41 1.73 0.95 COS 0 1.11 1.17 0.06 0.06 0.1 0.11 0.1 C S 2 0 WBm 0.15 0.16 0.02 0.03 0.23 0.03 0.03 N 2 0 WEm 169.59 180.25 181.22 186.9 189.67 193.25 194.35 C 0 2 92.65 m m 76.62 81.36 83.5 86.61 86.46 88.34 89.54 H 2 0 6.52 • 9 59.3 61.55 92.12 92.89 100.16 100.73 102.86 Total 199.17 360.09 382.9 369.03 379.97 379.94 388.01 389.52 % H 2S 50.21% 9.91% 10.17% 2.27% 2.42% 0.50% 0.98% 0.43% Gulf Canada's Hanalan-Robb gas plant, Ontario, Canada Table 1.1 Typical stream composition in the Claus process; C O D = sulphur condenser; C N V = catalytic converter; concentration of gases are reported as mol/total mol; at the end of the table %H2S = percent mol per total mol concentration of the H2S at each stage. It is documented that excessive free oxygen in the process stream deactivates alumina catalyst by sulphation (Beveridge, 1999). In the presence of S 0 2 , the alumina catalyst surface quickly builds up a layer of S 0 3 (McHugh et al, 1998; Davydov et al, 2003) which can deactivate the catalyst through two mechanisms. The first occurs at temperatures 300 to 350 °C in which SO3 reacts directly with the catalyst and forms A1 2(S04)3, a large volume low density material that blocks active sites (Heck et al, 1995). The second mechanism is the chemisorption of SO3 on the catalytic active sites at lower temperature, which prevents those sites from further catalytic action because of inaccessibility of active surface sites due to geometric blockage, or changes in the structure of the catalytic surface. 5 Chapter 1 - Introduction Even small concentrations of free oxygen w i l l increase sulphation (McHugh et al, 1998; Johnson et al, 1987). Free oxygen enters into the feed stream from leakage caused by corroded, poorly designed or improperly operated burners (Goodboy et al, 1985). The main burner and reaction furnace combine to form the thermal reactor at the beginning of the Claus process. The main burner must perform the function of burning one third of the feed H2S to SO2 to satisfy the stoichiometric requirements of the Claus process and at the same time consume all of the oxygen in the combustion air (Fenderson et al, 1998). Evidence shows that in practice the amount of free oxygen leakage from burners in the influent of the first catalytic converter varies from 20 to 5,000 ppmv continuously and in some case up to 20,000 ppmv in pulses (Johnson et al, 1987; Tell ierera/ . , 1980). 1.2 Motivation for the Study In recent years, reduction of sulphur content in automobile gasoline has become an urgent goal of refineries. In 1991 the U S E P A (United States Environmental Protection Agency) proposed new legislation requiring a 90% reduction of sulphur content in automobile gasoline in the United States by the end of 2004. It was proposed to reduce the sulphur content from 300 ppm to 30 ppm by the end of 2004 ( U S E P A , 2003). On July 31, 2002 the Government of Canada also introduced new regulations to reduce the sulphur content in on-road diesel fuel by 95 percent by June 2006 ( current limit is 500 ppm) and sulphur content of gasoline to an average of 30 ppm by January1, 2005 (current limit is 300 ppm). Similar efforts are underway throughout the world. L o w sulphur fuels are key to the control of emissions from existing vehicles and enabling advanced control technology and fuel-efficient design for new automobiles. 6 Chapter 1 - Introduction This significant reduction of sulphur content in diesel and gasoline can be achieved only by increasing the capacity and efficiency of both desulphurization and sulphur recovery units in petroleum refineries. In order to improve the performance of recovery units, it is necessary to extend the active life of the catalyst by decreasing catalyst deactivation. A s mentioned above, deactivation o f catalyst due to sulphation is the most important problem in sulphur recovery units. One solution is to use alternative catalysts such as T i 0 2 instead of conventional alumina. Studies show that sulphate formation can be reduced on T i 0 2 compared to A l 2 0 3 . The problem, however, is that to produce mechanically strong catalysts from T i 0 2 , as required for fixed bed application, makes the catalyst very expensive: about 18 US$/kg (Geus, 1999), compared to 1.7 US$/kg for activated A 1 2 0 3 (Alcan Chemicals, 2004). Another feasible solution is to develop a low price guard catalyst that removes residual 0 2 and thereby protects the conventional alumina catalyst from deactivation due to sulphation. The latter hasn't been discussed in the literature and only two such industrial catalysts are available and in use. 1.3 Objectives of the Present Study The objective of the preset research is to understand the effect o f the properties of a Fe-alumina catalyst on its oxygen scavenging activity in a Claus reaction. Detailed objectives of the present study are: 1. To prepare iron doped activated alumina catalyst in the form of oxide and sulphide. 7 Chapter 1 - Introduction 2. To characterize the prepared catalysts in order to determine the phase of iron on the support using laboratory techniques such as X R D (X-ray Diffraction Spectroscopy), D R I F T (Fourier Transmission Infrared Spectroscopy), B E T (Total Surface Area Measurement) and S E M (Scanning Electron Microscopy). 3. To determine the activity of the catalyst for oxygen scavenging under Claus reaction conditions. 4 . To develop an empirical rate equation for oxygen scavenging with the catalyst. 5. To propose a mechanism that can explain the observed kinetics on the catalyst. 8 Chapter 2 - Literature Review Chapter 2 - Literature Review 2.1 Deactivation The most widely used Claus catalyst in sulphur recovery units is non-promoted spherical activated alumina (J-AI2O3). Catalyst deactivation is the most frequent cause of loss of Claus process performance. Catalyst deactivation may occur because of sulphur fouling, coke deposition, loss in surface area due to thermal sintering or catalyst sulphation due to oxygen leakage from the main burner and reheaters. The latter mechanism is usually the most important cause of catalyst deactivation in the Claus process. 2.1.1 Sulphur Fouling-Sulphur Deposition One of the common Claus catalyst deactivation mechanisms is pore plugging by means of sulphur capillary condensation. Mora (2000) showed that Si through Sg molecules, formed via the modified Claus reaction, can plug the catalyst pores at standard Claus converter conditions above the sulphur dew point. The elemental sulphur content of a used catalyst may be due to two mechanisms: adsorption and condensation. Kerr et al. (1997) reported that typical steady state sulphur concentrations on catalysts from first and second converters range between 3 to 10% by weight; for third and fourth converters the range is 10 to 15%. Adsorbed elemental sulphur is an unavoidable deactivation agent. 9 Chapter 2 - Literature Review Schoofs (1985) applied the Ke lv in equation, modified by the Cohan hypothesis, Equation 2.1 to sulphur condensation over alumina and showed that micropores (d< 10 A ) are the maximum Claus catalyst pore diameter that can be plugged by sulphur, in the temperature range 160 to360°C . , -4yV'Cos0* d = d t i / / 7 Equation 2.1 RT\n(p/p0) where, d = pore diameter (cm); y = liquid surface tension (dyn/cm); V = liquid molar volume (cm 3 / mol); 0* = contact angle ( 0 ) ; R = Universal gas constant (8.314 x 10 7 erg/mol K ) ; T = absolute temperature ( K ); p = vapour pressure of the condensation gas inside the pore(mm Hg); p0 = vapour pressure of the condensing gas over a planar surface (mm Hg). 2.1.2 Coke Deposition-Carbon Deposition A s a result of the front-end furnace, traces of hydrocarbon and amines undergo cracking and polymerization leading to carbon deposition on the Claus catalyst. Carbon deposition is an upstream phenomenon; therefore, more deposition likely occurs in the first converter than in the second or third converters (George, 1977). 2.1.3 Loss in Surface Area - Thermal Ageing High temperature thermal and hydrofhermal reactions in Claus reactors may lead to loss in surface area. Probably maximum permanent loss in surface area occurs during the regeneration 10 Chapter 2 - Literature Review process. Removal of condensed sulphur is generally carried out by raising the catalyst bed temperature to about 400°C in the first converter and 300°C in the second and third converters, and maintaining this temperature for 24 to 36 hours with dilute reactant. If carbon deposits are present, an oxidative burn-off is attempted by adding air to the feed gas flow. The catalyst temperature during the burn-off (regeneration) can exceed up to 980°C (George, 1977). 2.1.4 Sulphation It is well known that oxygen leakage into the Claus process deactivates the alumina catalyst by sulphation (Pearson, 1977). Sulphate formation on the surface of the Claus catalyst has an inhibiting effect upon the conversion of H2S, SO2, C O S and CS2. Sulphation has been the subject of many scientific works and industrial investigations. These works can be categorized with respect to their goals as: 1. Study of the effect of oxygen on sulphate formation on activated alumina catalyst. 2. Characterization of sulphate formation on activated alumina catalyst. 3. Rate and kinetic study of sulphate formation on the alumina catalyst. 4. Study of the effect of promoters on sulphate formation in alumina catalyst. 5. Study of the sulphation on the alternative catalyst to conventional alumina catalyst. 6. Investigation and optimization of oxygen leakage from burners in the Claus process. 7. Mechanistic studies of sulphate formation in presence and absence of oxygen. In the following sections these studies are described in more detail. 11 Chapter 2 - Literature Review 2.2 Effect of Oxygen on Sulphate Formation In the work of Pearson (1977), a typical feed to the third converter of the Claus process was simulated in the presence of O2. The feed composition was 1 vol . % H2S, 0.55 vol . % SO2, 0.1 vol . % C O S , 28 vol . % H 2 0 and 0.5 vol . % 0 2 the balance N 2 . This gas stream passed through a bed of fresh activated alumina for 50 hours at 220°C. Pearson showed that the C O S conversion declined from 100% to 35% after 5 hours and the conversion of H 2 S / S 0 2 declined steadily to 43.5% from 100% of theoretical value (equilibrium value) after 50 hours. A t the end of the run the alumina catalyst was found to contain 6.6% sulphate species ( A L ^ S O ^ ) . The same experiment was repeated at 270°C. It was observed that conversion of C O S declined after 20 hours. Pearson concluded that sulphate formation and reduction are likely to be occurring simultaneously and the net result is an equilibrium level of sulphate on activated alumina. This equilibrium is reduced by increasing the temperature. Pearson's work clearly showed the significance of O2 in the deactivation of the catalyst. Few authors have studied the nature of the species formed during alumina sulphation in the presence of SO2 and 0 2 . In a series of experimental studies, Saur et al. (1986) and Lavalley et al. (1996) characterized the sulphate species on the surface of activated alumina. SO2 adsorption on alumina and sulphation in presence of 0 2 was investigated by IR spectroscopy and thermogravimetric measurement by Lavalley et al. (1996). The experiment was conducted at temperatures quite similar to Claus process plant operating conditions (280 to 350°C). The sulphation was studied under various partial pressures of SO2 and 0 2 . In the experimental work, alumina samples were sulphated in SO2 and O2 gas after evacuation at 450°C. The spectrum of surface sulphate species showed a sharp IR band near 1380 cm' 1 and a broad one near 1045 cm' 1 . 12 Chapter 2 - Literature Review The first band had been assigned by isotopic exchange to a v(S = O) vibration and the second band to v(S - O) vibrations. These species were thermally stable up to 600°C. Modification of the IR spectrum in the presence of steam at room temperature showed the sensitivity of this structure to water. The IR band near 1380 cm' 1 was clearly observed i f the sample was well dehydroxylated 2 1 due to the adsorption on O " site and the band near 1045 cm" on the O H ' site of the alumina surface. Lavalley et al. (1993) also studied the effect of sulphation on Z r 0 2 , T i 0 2 and AI2O3. In these series of experiments, catalyst sulphation was provoked either by sulfate impregnation on the pure oxide or by injection of 0 2 pulses during the reaction. Deactivation of the catalyst was measured by the conversion level of C O S in the feed stream under typical Claus first converter conditions. Sulphate ions were introduced by dry impregnation of pure oxides with titrated aqueous solution of (NH4) 2S04. After pre-sulphation of the catalyst, the feed composition was set as 1 vol . % C O S and 2.2 vol . % H 2 0 and balance He and the temperature of reactor was 503 K . Sulphate species were well characterized by an IR band at 1380 cm' 1 and the amount was proportional to the IR intensity. Results showed that sulphate species were not reduced by H 2 S due to the presence o f water and their structures are different to those under titration conditions (absence of H 2 0 ) , but rather like that of hydrogen-sulphate. M - O \ M - O — S =0 M - O / Scheme 2.1 Sulphate species in absence o f water M - O H OH M - O - S ^ Scheme 2.2 Sulphate species in presence of water 13 Chapter 2 - Literature Review Loading pure oxides with sulphates decreased their catalytic activity. The degree of deactivation varied in the following order: AI2O3 » T i C ^ > Zr02. Introduction of oxygen using pulses in the gas feed during the reaction on pure oxides leads to the same effect as pre-sulphation. The IR spectra of samples after these experiments present characteristic bands of sulfate species already described. The amount of sulphation varied in the following order ZrC>2 > T i 0 2 > AI2O3 (see Figure 2.1) and this corresponds well to the observed deactivation. 1400 1300 1200 1100 1000 Wavenumbers (cm") Figure 2.1 IR spectra of pure oxides: (a) Z r 0 2 , (b) TiC>2 and (c) AI2O3 after sulphation by S 0 2 + 0 2 at 723 K . (Lavalley et al, 2000) Further Work by Lavalley et al. (2000) showed that sulphate formation by SO2 + O2 on T i 0 2 was more than Y-AI2O3, but reducibility of sulphate species by H 2 S on T i 0 2 was more than Y-AI2O3. Therefore the total net SO3 formation on the T i 0 2 was reported much less than Y-AI2O3 in the Claus conditions. 14 Chapter 2 - Literature Review Finally it was concluded that sulphation of metal oxides poisons basic sites and creates Bronsted acidity, thus explaining catalyst deactivation. Thermogravimetric experiments of alumina sulphation by SO2 and O2 at 350°C were conducted by Lavalley et al. (1996). Experimental results showed that SO2 and O2 adsorption gave a fast and important weight increase during the very first minutes and reached a steady state after 5 hours. After repeating the experiment under different partial pressure of O2 and SO2 it was concluded that the rate of sulphation corresponds to P^ and P j 0 ° . A kinetic model of alumina sulphation was proposed by the same authors. It was concluded that since mass variation depended o n P ^ 5 , the mechanism necessarily needs to include oxygen dissociation on the catalyst. But they found that it was difficult to specify on which type of sites the dissociation of O2 molecules occurs. Oxygen dissociation needs electrons and it was proposed as Reaction 2.1, where Kroger's (1974) notation was adopted. Oxygen vacancies are present on V-AI2O3 but free electrons have never been noted. Notice that when the oxide is promoted by such metals as platinum, palladium, or vanadium, oxygen dissociation can be increased and thus sulphation favoured. The elementary steps of the proposed mechanism were noted as Reaction 2.2, 2.3, 2.4 and 2.5; again Kroger's notation was adopted. (In this notation, ' refer to a net negative charge, x to a neutral defect, and • to a positive charge. V is a vacancy, V"d a tetrahedral A l vacancy, and Vd an oxygen vacancy and thus the V-AI2O3 is denoted as 2AfMl>(fo V"D Vd. 02 +2V6 + 4e' = 20* so2+ox0surf^so*0si 2s + 02 =20-s 15 Reaction 2.1 Reaction 2.2 Reaction 2.3 Chapter 2 - Literature Review 40surf + S Reaction 2.4 + MOj^ + 2 AlxAl +3<tf + V„+V0 = Al2(S04)3 +30iurf+2Aliurf Reaction 2.5 In this last step (Reaction 2.5), the sulphated surface layer of alumina contributes to the growth of the aluminum sulphate phase, while the structure elements of alumina, which were just beneath the sulphated surface layer, became surface oxygen and aluminum ions ready to be sulphated. In this mechanism, the nucleation of the sulphate phase was assumed to be instantaneous. Moreover, the extent of reaction area was always at equilibrium. Multiplying the first step by 3, the second by 3/2, and the third by 3, the linear combination of steps Reaction 2.2 to Reaction 2.5 obtained leads to the studied reaction: Two types o f sites are involved in this model. One of them Ogsulf, representing an oxygen ion of the alumina lattice, is located at the surface of the catalyst and exists in large quantity. SO2 is adsorbed on it by Reaction 2.2 as a sulphite-like species and thus oxidized into sulphate (Reaction 2.4). The other site, written as s (O2 dissociation site), is a catalytic site restored after sulphate species formation (Reaction 2.3 and Reaction 2.4). It was assumed that this site only exists in small amounts; the amount is assumed to be constant because of regeneration during Reaction 2.4. Work of Nedez et al. (1997) is another unique work which not only showed the effect of oxygen in production of the sulphate species on the surface of the catalyst but also the effect of O2 concentration on this phenomenon. Different catalysts were evaluated under two distinct Al203+3S02 +3/202 = Al2{SO,), Reaction 2.6 16 Chapter 2 - Literature Review experimental conditions. The conditions were highly representative of commercial operations. The first condition was : 320°C , 6 v o l . % H 2 S , 4 v o l . % S 0 2 , 1 v o l . % C S 2 , 30 v o l . % H 2 0 , balance N 2 , 1 0 to 2000 ppm 0 2 , 2 or 3 s contact time. The second set of conditions was: 250°C, 3 vo l .% H 2 S , 1.5 v o l . % S 0 2 , 30 v o l . % H 2 0 , balance N 2 , 500 to 2000 ppm 0 2 , with 2 or 3 s contact time. Three different types of alumina catalyst were used in the experiment: CR-3S (Axen Porocatalyse), Alumina X and Alumina Y . The catalysts differ mainly in N a 2 0 concentration and macroporosity with pore size between 0.1 and 1 urn. The concentration of N a 2 0 was 2100, 3235 and 2025 ppm and the ratio of V i / V 0 . i (pore volume ratio) was 0.83, 0.51 and 0.67 for C R -3S, Alumina X and Alumina Y , respectively. After 80 hours operation and under the second set of experimental conditions, these three catalysts showed significant deactivation. Sulphur yield relative to the 100% equilibrium amount, decreased to 84.5%, 73% and 71% for CR-3S , Alumina X and Y , respectively. In this work it was concluded that a minimum concentration of the N a 2 0 content (>1500 ppm) is needed for adequate activity, a level over 2500 ppm favors the sulphation of surface sites and this affects a particularly fast chemical deactivation. Also, the amount of oxygen in the feed stream affects the deactivation of the catalyst. Figure 2.2 shows one set of the experimental results by Nedez et al. (1997). The amount of promoter in the activated alumina catalyst can change the degree of sulphation. In agreement with the work of the Nedez et al. (1997), it has been observed that sulphate formation on activated alumina was increased by increasing the amount of N a doping (Piepluera/. , 1995). The very recent work of Laperdrix et al. (2000) studied the sulphation on different promoted activated aluminas and the role of H 2 S to reduce sulphated species. Sulphate species were formed by heating S 0 2 in excess of 0 2 at 723 K on AI2O3 and A l 2 0 3 promoted by Fe, Cu , 17 Chapter 2 - Literature Review N i , Na , M o , W or V . The results were studied by FT-IR spectroscopy. In addition, reducibility of sulphate species was studied on each catalyst as a function of temperature. The influence of water on the reduction was also studied in this work. 90 n 0 500 1000 1500 2000 2500 Oxygen(ppm) • CR -3S A Alumina X • Alumina Y Figure 2.2 Role of the Oxygen level on sulphur yield; feed contains 3 v o l . % H2S, 1.5 v o l . % S 0 2 , 30 v o l . % H 2 0 , balance N 2 , 500 and 2000 ppm 0 2 at 250°C for 3 s contact time (Nedez et al, 1997). Since all experiments were performed at the same conditions and the same alumina support, it was concluded that the effect of promoters toward certain surface sulphate species on the support could be divided into three categories. 1. M o , V , W : Lower amount of A l 2 0 3 sulphate species than on alumina, (see Figure 2.3) 2. Na : Few alumina surface sulphates, but formation of mixed and N a sulphate species. 3. Fe, Cu , N i : Larger amount of AI2O3 sulphate species than on alumina, (see Figure 2.4) 18 Chapter 2 - Literature Review 1500 1400 1300 1200 1100 1000 Wavenumbers (cm-1) Figure 2.3 IR spectra of promoted alumina after sulphation by S 0 2 + 0 2 at 723 K , (a) A 1 2 0 3 ; (b) V/AI2O3; (c) W / A 1 2 0 3 ; (d) M o / A l 2 0 3 (Lapardix et al, 2000) e d c b a . 1 1 1 1 1400 1300 1200 1100 10O0 Wavenumbers (cm-1) Figure 2.4 IR spectra of promoted alumina after sulphation by S 0 2 + 0 2 at 723 K , (a) AI2O3; (b) N a / A l 2 0 3 ; (c) C u / A l 2 0 3 ; (d) F e / A l 2 0 3 ; (e) N i / A l 2 0 3 (Lapardix et a/., 2000) 19 Chapter 2 - Literature Review The reducibility of the sulphated catalysts was studied using H2S at different temperatures. On promoted alumina catalysts, the reducibility of sulphate species was enhanced relative to that observed on pure alumina. The strongest effect was observed with copper. A decrease similar to that observed on CU/AI2O3 was observed on Fe /Al 2 03 and the influence of nickel was less pronounced (see Figure 2.5). Since industrial feeds may contain up to 30 percent of water, its effect on H2S sulphate reduction has been studied on all the sulphated catalysts. It has been found that the introduction of water in the presence of H 2 S promoted the reduction of sulphate species on all the sulphated catalysts. In the presence of water, covalent sulphate species on alumina are transformed into ionic species, which would be more easily reduced by H2S. Figure 2.5 Variation of sulphate species (relative to intensity of the band near 1380 cm"1) after introduction of H 2 S at increasing temperature, on AI2O3, Zr02 and promoted aluminas (Lapardix et al., 2000) ZxQt - * - Cu/Ai203 Ni/Alj03 - V - M0/AI2O3 = 0 WAtjOs - o - V/AJ2Oa 20 Chapter 2 - Literature Review 2.3 Sulphate Formation on Activated Alumina in Absence of Oxygen Clark et al. (2002) recently published a unique work on sulphate formation during H2S and SO2 conversion in anoxic environment. In this study, a mechanism for the conversion of H2S and SO2 over alumina was proposed, in which surface thiousulfate is formed and reacts with SO2 and H2S via a series of oxy-anions, to form sulphur and water. It was proposed that beside thiosulphate, sulphate is also formed as part of the catalytic cycle but the surface concentration is limited to some steady-state concentration by virtue of its reduction by H2S. SO, + 2H,S 3/8 S» + 2H 20 AM), A l j O , The Claus Reaction SOj + H 2 0 SO, 2 + 1/8 S s -+*s :o 3 2 + s o 2 SjOj2" + H2S HS 4 0 4 ~ + H2S •*> l H , S O j -> H S 4 0 4 + -+ [S31 + H,0 + HS 2 0,~ OR HSjO," + OH" " J -2H + SO, 2' (1) (2) (3) (4) (5) (6) OH HS 20," + OH S0 4 2' + H2S S0 4 : + H2S ions. 2 [$,] -* H,0 + SH Formation \ (7) (8) (9) .v = 6,7,8. 2* t predominant .speck's Scheme 2.3 Proposed mechanisms of sulfur formation on an alumina Claus catalyst (Clark et al., 2000) 21 Chapter 2 - Literature Review In their work, experiments were conducted on CR-3S commercial alumina provided by Axens Porocatalyes. This catalyst was chosen because of its high purity and low N a content. Clark et al. showed that sulphate and thiosulphate are formed on activated alumina as a result of the Claus reaction and the O2 is not required for the formation of these species, (see Scheme 2.3) In order to investigate the proposed mechanism, experiments were conducted at 220°C and 320°C with and without 3000 ppmv O2 added to a feed gas, representing the first and third converter. The experimental results demonstrated that 0 2 has little or no influence at 320°C, and sulphate levels are comparable with or without the added O2. However, at 220°C more sulphate was formed when O2 was present in the feed gas. Clark et al. concluded that sulphate build-up on the alumina in the third or a probably second converter was a consequence of both the Claus mechanism and the influence of free oxygen in the system. 2.4 Sulphation in Industrial Processes 2.4.1 Oxygen Leakage - Halan-Robb Gas Refinery Plant (Ontario) Work of Johnson et al. (1987) is one of the few studies at an actual industrial site related to oxygen leakage from the burners, and deactivation of the activated alumina due to sulphation. In this work a technique was developed to study the sulphation at G u l f Canada's Hanlan-Robb, Ontario, gas-processing plant. This site has an 1100 ton per day sulphur recovery facility consisting of two general stages. The first consists of a Claus unit with a furnace and three catalytic converters as well as three in-line burners for reheat of the feed stream. The second stage is a tail gas clean up unit. 22 Chapter 2 - Literature Review The major source of gas to the plant from the Hanlan field has a relatively poor H2S to CO2 ratio (0.84 to 1) and some unusual characteristics with a small amount of C6+ (aromatics). The equipment used to monitor the plant during this test included a continuous sample-handling system, an on-line gas chromatograph and a trace-oxygen monitor. Results from oxygen monitoring showed a significant amount of oxygen in the inlet stream. Correction was employed to the burners by increasing the acid gas to air ratio. Based on this correction, the plant was able to reduce the oxygen leakage to 30 ppm in the converter inlet which corresponds to 0.3 percent sulphate deposition on the catalyst, per day. 160 t , 0.3 0.4 0.5 0.6 0.7 0.8 A c i d F e e d G a s / A i r Burner No. 1 Burner No.2 - - 'Burner No.3 Figure 2.6 In-line burner performance of G u l f Canada Corporation's Hanlan-Robb, Ontario gas processing plant. Oxygen leakage has been shown before and after optimization. (Johnson et al, 1987) 23 Chapter 2 - Literature Review Figure 2.6 shows the oxygen leakage from each burner in the site before and after optimization. 2.4.2 Oxygen Leakage and Deactivation - Ram River Gas Refinery Plant (Alberta) Sulphate formation in Husky Oi l ' s Ram River Gas Plant Claus units was studied in the work of Beveridge (1989). The Ram River plant processes sour natural gas from 12 different main sources with a maximum H 2 S content of 39 vol . %. The plant produces 4600 tonnes per day sulphur as by product. The sulphur recovery facility consists of four separate Claus plants each with two catalytic converters and 1150 tonnes per day sulphur recovery capacity. The plant also has two separate tail gas processing Sulfreen units. The waste gas from tail gas process is passed through separate incinerators where the remaining sulphur components are oxidized to S 0 2 before being released through a common stack. The plant's Claus unit uses direct fired acid gas burners to control the converter outlet temperature. This method o f reheating is very common in natural gas refineries. The burners are operated with 25 to 30% excess H 2 S to insure complete 0 2 consumption. Stoichometric combustion would introduce large amounts of free oxygen into the system due to burner inefficiency. A study of the cause of deactivation of alumina catalyst and T i 0 2 used in this plant revealed that sulphation was the main source of deactivation. Burners were studied for oxygen leakage. Analysis of samples taken up from influent of catalytic converters confirmed that 30 to 70 ppm oxygen entered the process from the burners. With improved H 2 S / S 0 2 ratio control, and an oxygen monitoring program, the oxygen in the tail gas was reduced from 50 to 1600 ppm in 1976, to 30 to 100 ppm in 1984. This study clearly 24 Chapter 2 - Literature Review showed that oxygen leakage in the system is not completely preventable and the actual oxygen in the system probably is far more than the amount detected in the tail gas. A s a result of work conducted, the overall yearly recovery average was increased from 98.1% in 1980 to 98.7% in 1983. The Report did not contain any information about improvement of catalyst life time. 2.5 Industrial Oxygen Scavengers Probably the most complete information regarding the industrial use of oxygen scavengers is available in the form of a Patent (United States Patent 4,192,857-1980). The objective of this patent is to introduce an invention (catalyst) to overcome the disadvantage of the sensitivity o f Claus catalysts to oxygen. The new process, according the patent, thus enables the Claus process to be carried out at any appropriate temperature with practically constant high efficiency. The improvement, according to the invention, involves first passing the gas to be processed through a bed of deoxidation catalyst before subjecting it to the Claus reaction over the traditional catalyst for the latter. The deoxidation catalyst, according to the patent, is preferably a support such as alumina , bauxite, silica, a mixture of silica and alumina or a silico-aluminate containing a compound of Fe, N i , Co, Cu , or Zn . This compound may be an oxide, sulphide or an organic acid salt such as sulphate, nitrate, phosphate, acetate etc. The catalyst contains approximately 0.5 to 10 %, and preferably 2 to 6 % of metal by weight relative to the oxide catalyst. The support must have a specific area, determined by the B E T method, of at least 100 m 2 /g , and preferably greater than 150 m 2 /g. The pore volume must be 0.20 ml/g. The patent claimed that in the case of iron doped alumina, the conversion of oxygen in a feed gas with composition of 2.5 vol . % H2S, 1.5 vol . % 25 Chapter 2 - Literature Review S 0 2 , 27vol. % H 2 0 and 0.2vol. % 0 2 , balance N 2 can reach to 99.35, 99.4 and 99.8% at 250, 260 and 320°C respectively. One of the oxygen scavengers in industrial use is Axens Porocatalyse A M catalyst. According to the report, A M catalyst is an iron impregnated alumina used as a proactive layer in the converter to prevent the alumina from sulphation. In the Ram River gas refinery plant this catalyst has been used in all four second converters. Actual data from the site showed that this catalyst was able to reduce oxygen level from 1000 ppm to 50 ppm and 7 ppm with 1 and 1.5 second residence time, respectively. This catalyst is recommended for the second and third converters with 30 percent of total catalyst volume loading. In contrast, a poor performance was reported by Filatova et al. (2002) for A M catalyst in the Astrakhan (Russia) gas processing plant. Published results show that fresh A M catalyst has no activity toward converting oxygen , but the catalyst gains 100% activity (oxygen conversion) after 2.5 hours in a stream with composition of 2.5 v o l . % H 2 S , 1.25 v o l . % S 0 2 , 25 v o l . % H 2 0 , 3000 p p m 0 2 in balance N 2 at the temperature range of 200 to 270°C. Another similar catalyst available for industrial use is A l c o a S-100 S R catalyst. There is no published document available regarding the performance of this catalyst. 26 Chapter 2 - Literature Review 2.6 Superclaus Process and the Role of Iron Promoted Metal Oxide Catalysts The idea of using iron sulphide promoted on Y-AI2O3 as an effective active site for oxygen scavenging was adapted from some of the work conducted on Superclaus catalysts. The Superclaus process was introduced to industry in 1988 as an alternative to the conventional Claus tail gas1 processes (Lagas, 1993). The Superclaus unit uses a specially designed catalyst to promote the following non equilibrium, partial oxidation reaction of H2S in the presence of excess air: H2S + - 0 2 -> -S n + H20 Reaction 2.7 2 n In the Superclaus process the main objective is to minimize the oxidation of the H2S to SO2, in order to maximize the sulphur recovery. The complete oxidation reaction is: H2S + - 0 2 - > - S „ + H20 + 02 -> S02 + H20 Reaction 2.8 2 n According to Lagas et al. (1993) the performance of the Superclaus catalyst is determined by the kinetic control of the reaction sequence: ' The Claus tail gas is the stream coming out of the Claus unit which contains traces of H 2S, S0 2 and other sulphur compounds that still needs to be treated. The Mobil Oil Canada Lone Pine Creek Gas Plant adopted the Superclaus process for the first time in North America in 1990. 27 Chapter 2 - Literature Review H2S-^> S — S 0 2 Reaction 2.9 In this process, the catalyst is designed to minimize the rate constant ratio k^/ki.. In contrast, the idea behind the oxygen scavenging catalyst, discussed in Section 2.5 is to maximize the O2 consumption by increasing the H2S to SO2 oxidation. The formation of each SO2 molecule means that 1.5 molecules are being converted from the gas stream, which is 3 times more than the Supeclaus process. Therefore, in oxygen scavenging, the main objective is to increase the ratio of k.2/k\ in Reaction 2.9, in order to increase SO2 selectivity. Hence, the property of a catalyst which enhances the retro-Claus (SO2 formation) favours oxygen scavenging. The first generation Superclaus catalyst was alpha-alumina (a-Al 2 03) based, with an iron oxide and chromium oxide layer. Lavalley et al. (1999), in their work to make active catalysts for the Superclaus reaction, studied Fe/Cr. In the Superclaus process, H2S is oxidized directly by O2 at 230°C. In their work it is claimed that catalysts based on mixed iron-chromium oxides, deposited on a - A l 2 0 3 , are quite effective because this support has a lower specific area and larger pores than V-AI2O3, thus minimizing the likelihood of the retro-Claus reaction and oxidation of elemental sulphur. The importance of chromium has also been stressed: it favours S n selectivity and improves the resistance of the catalyst to deactivation (converting to sulphide). Supported Fe203, C r 2 0 3 and mixed iron-chromium oxides were tested in a flow of 1.2 vol . % H 2 S , 1 vol . % 0 2 in balance N 2 at 503 K . A l l samples were tested under similar conditions. Results showed that the initial conversion of H2S on Fe2C>3 was very high but decreased with time on stream, while the S n selectivity dropped rapidly as well . According to Berben et al. (1999) the iron (III) oxide covers some basic sites present on the surface of the alumina. Reaction to iron sulphide by exposure to an excess of hydrogen 28 Chapter 2 - Literature Review sulphide leads to recrystallization to more three dimensional species not completely covering the basic sites. A s a result, the selectivity to elemental sulphur drops after recrystallization of the iron oxide. Also , it was concluded that iron sulphide formation could enhance 0 2 consumption and thus increase SO2 formation. Lapaerdrix et al. (1999) also showed that over chromium oxide, conversion of H 2 S is also quite low (12%) but remains more stable over time. The mixed Fe/Cr = 0.5 oxides showed a better performance and selectivity toward elemental sulphur. In very similar work, Pieplu et al. (1995) explained the loss in sulphur selectivity observed on bulk iron or mixed iron-chromium oxide catalyst in the absence of steam in the feed, by the formation of an iron sulphide phase. The second generation Superclaus catalyst is silica (Si) based with only an Fe2C>3 layer. According to Lagas et al. (1993), iron oxide is considered as the active site in both first and second generation Superclaus catalysts. The silica (Si) based catalyst has a surface area of about 90m /g compared to the a-Al2C>3 based catalysts area of about 10 m /g. Increased surface area of the Si-based catalyst makes the Si-based version more active than the a -A l 2 03 type, thus the silica catalyst has a lower activation temperature than the CI-AI2O3 catalyst. A lower activation temperature means that a higher percentage of H 2 S can be introduced to the Superclaus unit and still minimize SO2 formation. In the work of Keller et al. (2001) the second generation of the Superclaus catalyst has been studied. In this work, selective oxidation of H2S over silicon carbide (SiC) supported iron catalyst was studied. According to the authors, the very exothermic nature of the H2S oxidation (ca. 70 °C temperature increase per percent of H2S converted in an adiabatic mode), could lead to a decrease in selectivity to elemental sulphur by the formation of SO2. The chemical inertness 29 Chapter 2 - Literature Review and the high thermal conductivity of S i C could make such a support a promising candidate for the substitution of oxide supports for over dewpoint H2S oxidation reactions. Results from the performance obtained on the Fe2C>3 ( 5 % ) / S i C catalyst as function of G H S V showed that SO2 was mainly produced according to the successive reaction between sulphur and oxygen and not directly from H2S. The very low acidity of S iC compared to traditional oxide supports could also explain the high sulphur selectivity obtained. Terbrde et al. ( 1 9 9 3 ) reported that the formation of sulphur radicals, which can react with molecular oxygen, could probably be attributed to the occurrence of highly acidic sites on the silica surface under the reaction conditions. The authors noted that addition of basic sodium diminished the formation of sulphur radicals. The catalytic oxidation of hydrogen sulphide to elemental sulphur was investigated by L i et al. ( 2 0 0 1 ) on four rare earth orthovanadates R e V C M ( R e was Ce, Y , L a or Sm) and three magnesium vanadates ( MgV206, Mg2V207 and Mg3V20g) . It was reported that at 2 6 0 ° C , in a flow of 3 v o l . % H2S, 5 vol.%) O2 in balance nitrogen, zero sulphur selectivity was obtained for YVO4 at 100%) H 2 S conversion, and for V2O5 at 96.3%) H 2 S conversion. It was concluded that zero selectivity indicates that all hydrogen sulphide reacted (or all sulfur formed) was converted completely to SO2. 2.7 Supported Metal Catalyst Preparation 2.7.1 Introduction The aim of the preparation of the catalytic materials that can be employed on an industrial scale is to obtain a product with high activity, selectivity and stability. In order to reach this 3 0 Chapter 2 - Literature Review objective the active phase - in our case metal - is usually deposited on the surface of a support, highly porous and thermostable materials with a high surface area and suitable mechanical strength. The common preparation methods of dispersed metal catalyst require a combination of different unit operations, which can be described as : ( i ) introduction of the metal precursor on the support by impregnation or ion- exchange, coprecipitation and deposition precipitation,( i i ) drying and calcination, and (iii) reduction. 2.7.2 Impregnation Methods According to the amount of solution used, two types of impregnation can be distinguished: one called "incipient wetness "or "dry" impregnation because the volume of the solution containing the precursor does not exceed the pore volume of the support. In the simplest case, the impregnation solution is sprayed onto the support which is maintained under stirring and absorbed H2O has been previously evacuated. In principle, this method appears to be simple, economic (especially when using solutions of costly active components) and able to give a reproducible metal loading which is however limited by the solubility of the metal precursor. However, when higher concentrations of the metal are required, solubility limitations can be overcome by carrying out consecutive (successive) impregnation steps. The second type of impregnation, called "wet" or "soaking", involves the use of an excess of solution with respect to the pore volume of the support. The mixture is left for a certain time under stirring and it is filtered and dried to recover the catalyst (Pinaa, 1998). 31 Chapter 2 - Literature Review 2.7.3 Ion Exchange Inorganic oxides such as AI2O3, SiC>2, T i 0 2 , M g O , which are commonly used as support materials, tend to polarize and to be surface charged once suspended in an aqueous solution. This charge wi l l be controlled by the p H of the solution according to: M-OH + H+A~ <r>M-OH2+A- Reaction2.10 M-OH + OH' <r>M-OH~ + H20 Reaction2.11 In acidic media Reaction 2.10 the adsorption surface site ( M - O H ) is positively charged and w i l l be covered by anions, while in basic media Reaction 2.11 the acidic surface site ( M - O H ) wi l l be negatively charged and covered by cations. For each oxide, a particular p H at which the surface wi l l not be charged wi l l then exist. This p H is called P Z C (zero point o f charge) or IEPS (isoelectric point) and for Y - A I 2 O 3 and 01-AI2O3 is reported to be 9.0 and 7.0, respectively. 2.7.4 Drying After impregnation, the material undergoes a drying treatment which is generally performed at temperatures between 80°C and 200°C in order to eliminate the solvent used in the previous impregnation step. Different variables such as the rate of heating, final temperature and time of treatment and type of atmosphere can influence the process and have to be selected according to the different systems. A simplified model which takes into account a uniform type of pores has been described by Kotter et al. (1979), showing how the active phase w i l l concentrate at the inner part of the 32 Chapter 2 - Literature Review particle or at the external surface as a function of the rate of drying. If the drying rate is very slow the evaporation of the solvent (usually water), which starts at the external surfaces, allows deeper diffusion of the salt into the liquid in the pore. This results in an increase of concentration of the solution in the inner pores. On the other hand, too high drying rates wi l l generate temperature gradients and w i l l force the solution towards the outer layer of the particles. In practice, the situation can be more complicated because we are dealing with a complex porous system (Fenelonov, 1979). 2.7.5 Calcination This treatment consists of heating the catalysts in oxidizing atmosphere at a temperature usually as high as or little higher than that encountered during the reaction. Calcination has the purpose o f decomposition o f the metal precursor with formation o f an oxide and removal o f cations or anions in the gas products which have been previously introduced. In the case of industrial production, calcinations are useful for the removal of extraneous materials, like binders or lubricants, which have been used during previous forming operations (extrusion, tablating, etc). 2.8 Preparation and Preconditioning of Fe/Al 2 0 3 Catalyst Iron promoted alumina is used in a number of different processes. Here some of the preparation and preconditioning methods used for FdAhO^, are reviewed. The reduction/oxidation of a high loading iron (15% Fe) catalyst supported on V-AI2O3 has been studied in the work of Thomson et al. (1991). The catalyst used in this study was a high 33 Chapter 2 - Literature Review loading iron catalyst supported on a high surface area Y-AI2O3 ( B E T area 224m /g and pore volume 0.631 cm 3/g). The catalyst was prepared by successive impregnation of a solution of Fe (N03)3.9H20 until a loading of 31.8%, based on the weight of iron, was achieved. After the support particles (100 - 200 mesh) were wetted with the solution, the catalyst was dried at 383 K for 24 hours and then calcined at 673 K in air for 24 hours. This cycle was repeated six times. The supported iron oxide was scanned by X R D after the calcination step and was found to be in the form of a-Fe 203 (hematite). However, Thomson also observed that there were noticeable shifts in the position of the supported hematite peaks when compared to those of unsupported hematite i.e. the lattice parameters of the supported hematite on Y-AI2O3 were smaller than in bulk hematite. They concluded that during the impregnation procedure it is likely that some of the A l in Y-AI2O3 was dissolved in the acidic iron nitrate solution and this dissolved A l is thought to be incorporated into the matrix of the hematite upon crystallization and calcination. This phenomenon was verified by the work of Vaishnava et al. (1985) who studied ¥qIA\2OT, by Mossbaur spectroscopy and found that the calcined Fe/Al203 was composed of Fe ions strongly interacting with the alumina surface to form a solid phase containing Fe, A l and O and 10%o bulk hematite. The S iC supported iron oxide catalyst used by Keller et al. (2001) was prepared by incipient wetness impregnation of the dry support (SiC) with an aqueous solution of Fe(N03)3 .9H 2 0. The resulting material was dried in an oven at 110°C for 2 hours and then calcined in air at 400°C for 2 hours in order to decompose the iron nitrate precursor and to form its corresponding oxide. In the work of L i et al. (2003) the Fe/y-Al 2 03 catalyst was made from two different precursors, iron nitrate and iron sulphate. Appropriate amounts of iron precursors FeS04.7H20 34 Chapter 2 - Literature Review or Fe (N03)3 .9H 2 0 in deionised water were mixed with y-ALiCb (0.01 urn in size and 99.99% in purity) for over 12 hours, followed by evaporation, drying and calcination at 500°C for 5 hours. Reduction and sulphidation of iron species on iron-supported Y-Zeolite was studied by temperature programmed reduction and sulphidation in the work of Inamura et al. (1993). In this work iron catalysts were made by the impregnation method from two different precursors, F e S 0 4 . 7 H 2 0 or F e ( N 0 3 ) 3 . 9 H 2 0 on the Y - Zeolite. Before T P R (Temperature Program Reduction) and TPS (Temperature Program Sulphidation) experiments, all catalysts were calcined in situ at 650 K for 2 hours in flowing dry air. After cooling the catalyst under A r flow, the flow was replaced with reacting gases 64.5 vol . % H 2 in balance A r for T P R and mixing streams of 64.5 vol . % H 2 and 5.44 vol . % H 2 S in balance A r for TPS . Quantitative T P R analysis of catalyst prepared, using Fe(N03)3 .9H 2 0 showed a broad consumption of the H 2 at 500 to 700 K and a sharp consumption at 1150-1200 K . TPS patterns showed that a slow adsorption of H 2 S during the isothermal period at 300 K , and a subsequent H 2 S evolution peak at 400 K (sharp) and 600 K (broad). The same sulphidation method was used in the work of Arnoldy et al. (1985) in order to study the temperature programmed sulphiding of M0O3/AI2O3. 2.9 Infrared Spectroscopy of the Surface of Metal Oxides Using Adsorbed, Probe Molecules The characterization methods generally applicable to supported metal oxide catalysts are reviewed in this section. Raman and infrared (IR) are complementary spectroscopies that are among the unique characterization techniques that have provided fundamental molecular level information about the surface properties of supported metal oxide catalysts. The IR information 35 Chapter 2 - Literature Review using probe molecules was used in the present study to identify the phase transformations of the oxygen scavenging catalyst. The intensities of IR absorption bands depend on the change in the dipole moment brought about by variations in the molecular geometry of the molecular vibration while the intensities of Raman bands depend on the change in polarizability associated with the vibration (Nakamoto, 1977). Consequently bonds possessing ionic character tend to give strong IR signals and bonds possessing covalent character tend give strong Raman signals. Preconditioning o f metal oxide prior to spectroscopy is an important procedure to eliminate the water absorbed on the surface of the catalyst. Under ambient conditions, the supported metal oxide catalyst contains a thin fi lm of water which hydrates the surface metal oxide species (Deo et al., 1991). The M = 0 and M - 0 vibrations o f the hydrated surface metal oxide species generally occur below 1000cm"1 which usually makes it difficult to detect these vibrations using IR spectroscopy. Dehydration shortens the terminal M = 0 bond and generally shifts this vibration to above 1000 cm"1 which makes it detectable in the IR (Wachs, 1996). The reason why it is often difficult to measure the metal oxide or metal sulphide vibrations of the catalytically active phase is that the frequencies are below 1000 cm" 1 where measurement is difficult due to absorption by the support. Therefore, gathering information in an indirect way, via hydroxyl groups on the support or via the adsorption of probe molecules such as C O is very common. C O adsorption can be used to characterize both acidic and basic sites of metal oxides. C O can interact with basic O " sites that lead, at 100 K , to the formation o f surface carbonate ( C 0 2 2 ' ) ions: 36 Chapter 2 - Literature Review O 2-C O t c C O - o -Scheme 2.4 Formation of surface carbonate where the C O molecule participates in the role of a Lewis acid. The carbonate species react with excess C O at temperatures higher than 100K giving rise to more complex species. IR spectra of such systems were studied by Coluccia et al. (1981) at room temperature who concluded that C O disproportionation occurs, leading to negatively charged polymeric structure like (CO)„ 2" and carbonate ions. The overall process may be described as: The band observed at high frequency (~ 2200cm"1) when carbon monoxide is weakly adsorbed on a highly oxidized zinc oxide sample has also received an alternative explanation from Amberg et al. (1965).These workers note the correlation between degree of preoxidation and the vibrational frequency of the adsorbed species and provide arguments for considering that the carbon monoxide is adsorbed via an ion-dipole interaction with surface oxide ions. The positive end of the C - 0 dipole is on the carbon atom. This atom, therefore, interacts with the negative surface oxide anions to give a species of the form: xO2- + (n + -)CO -> (CO)*- +-C02 2- Reaction 2.12 O 2- 5+ 5-C - 0 s Scheme 2.5 Adsorption of C O on the surface of metal oxide 37 Chapter 2 - Literature Review A weak band at about 2200cm"1 was also observed for carbon monoxide adsorbed on N i O -Si02 (Eischens, 1956). Also Davydov (1990) showed that C O can adsorb on the surface of iron oxide and this absorption can be enhanced by increasing the temperature. 2.10 Summary The Claus process is used to remove H2S from acid gas streams and is widely practiced in industry. The most common Claus catalyst in sulphur recovery units is activated Y-AI2O3. However, it is well documented that excessive free oxygen, leaked from the burners, in the Claus process stream deactivates the alumina catalyst by surface reaction with SO2 that leads to sulphation of the alumina. Sulphate formation on the surface of the catalyst has an inhibiting effect upon the conversion of H2S and this limits the catalyst activity. Sulphate formation on y-alumina was clearly observed by a sharp IR band near 1380 cm" 1 and a broad one near 1045 cm" 1 using FT-IR measurement. The intensity o f the band depends on microporosity and concentration o f Na20 as well as the type o f promoters on the AI2O3 catalyst. There are many reports describing different types of catalysts resistant to sulphation. Compared to Y-AI2O3, T i 0 2 catalyst deactivates much less. Studies showed that sulphate formation by SO2 + O2 on T i 0 2 was more than Y-AI2O3, but reducibility o f sulphate species by H 2 S on T i 0 2 was more than Y-AI2O3. Therefore the total net SO3 formation on the T i 0 2 was reported much less than Y-AI2O3 and this makes Ti02 an excellent catalyst for the Claus process, but 5 to 10 times more expensive than y - A l 2 0 3 . Another approach to minimizing deactivation by sulfate formation is to use a dual bed reactor configuration, where part of the conventional alumina catalyst is replaced with a catalyst that w i l l promote oxygen consumption and thereby limit sulfate formation. Two industrial 38 Chapter 2 - Literature Review studies based on this approach, using Axens Porocatalyse A M and A l c o a S-100 S R catalysts are described, but there are no journal publications available. In the present study, Fe based catalysts supported on alumina have been prepared and examined for oxygen removal. The effect of catalyst pre-treatment (sulphidation) conditions on oxygen removal, kinetics and mechanism of the reaction w i l l be reported. The idea of using Fe supported on Y-AI2O3 as an effective active site for oxygen scavenging was adapted from some of the work conducted on Superclaus catalysts. The Superclaus unit uses a specially designed catalyst to promote the following non equilibrium, partial oxidation of H2S in the presence of excess air: H2S + - 0 2 - > - S „ + H20 2 n In this process iron oxide based catalyst, supported on alumina, was used to promote S yield and H 2 S conversion. Superclaus catalyst is deactivated when the iron phase is changed from oxide to sulphide which promotes more oxygen consumption and S 0 2 selectivity. The conditions which make the Superclaus catalyst deactivate are ideal conditions for the oxygen scavenger catalyst. 39 Chapter 3 - Experimental Apparatus and Methods Chapter 3 - Experimental Apparatus and Methods Safety note: Work described in this thesis utilized H2S and SO2 gases. Both gases are very poisonous and should be handled in a ventilated, enclosed system equipped with gas detectors. Ventilation rates should be such that, in the event of a leak, H2S and S 0 2 cannot build up to concentrations > 100 ppmv. Gas detectors should be set to sound an alarm at concentrations of 10 ppmv. 3.1 Catalyst Preparation (Fe203/Al203 and FeS 2/Al 20 3) Catalysts were prepared at three different iron loadings. For the 2 wt. % and 5 wt. % Fe loading catalysts, the incipient wetness method was used and for the 15% Fe loading, successive incipient wetness impregnations were performed. The catalyst precursor was prepared by dissolving iron (III) nitrate nanohydrate 99.99+% [Fe (NO3) 3.9H2O] in distilled water (pH = 2). Alcan A A - 4 0 0 activated alumina, 48 mesh size, was used as support. The pore volume of the support was 0.5 cc/g. The support was loaded into a crucible and dried at 110 °C overnight. The aqueous solution was then evenly sprayed over the support using a 10 ml syringe. During the impregnation procedure it is likely that some of the A l in AI2O3 was dissolved in the acidic iron nitrate solution and this dissolved A l is thought to be incorporated into the matrix of the hematite (iron oxide) upon crystallization and calcination. This was verified by Vaishnava et al. (1985). The catalyst was dried in an oven at 110 °C over night and calcined in air at 400 °C for 2 h in order to decompose the iron nitrate precursor and to form its corresponding oxide. It was 40 Chapter 3 - Experimental Apparatus and Methods predicted that the catalyst would be active in the sulphided state, therefore activation was carried out by treating the oxide precursor in a mixture of H2S/N2 and H 2 (10 vol . % H2S, 40 vol . % H2 and balance nitrogen) at 400°C. The function o f the hydrogen is to prevent the decomposition of the relatively unstable H2S to elemental sulphur which would otherwise accumulate on the surface of the catalyst. Some of the experiments were carried out using pure FeS (99.9% purity) provided by Great Western Inorganics Company. 3.2 Catalyst Characterization 3.2.1 X R D (X-ray Diffraction Spectroscopy) X R D (X-ray Diffraction) was used to identify the bulk crystalline phases present in the catalyst to monitor bulk transformations in the catalyst, and to estimate the catalyst crystallite size. X-rays have wavelengths in the Angestrom range and they are sufficiently energetic to penetrate into solids and are therefore well suited to probe their internal structure. X-ray diffraction is elastic scattering of X-ray photons by atoms in a periodic lattice. The scattered monochromatic X-rays that are in phase, give constructive interference. Figure 3.1 illustrates how diffraction of X-rays by crystal planes allows one to derive lattice spacing by using the Bragg relation (Niemantsverdriet, 1995). 2dsin0 = n * X Equation 3.1 41 Chapter 3 - Experimental Apparatus and Methods Bragg's Law Figure 3.1 X - ray scattered by atoms in an ordered lattice interfere constructively in directions given by Bragg's law. Diffractograms are measured as a function of the angle 29. Rotation o f the sample and source during the measurement enhances the number of particles that contribute to diffraction. In this equation, X is the wavelength of the X-ray, 0 the scattering angle, and n* is an integer representing the order of the diffraction peak (usually n* = 1). Bragg's Law is one o f most important laws used for interpreting X-ray diffraction data. B y measuring the angle 29, under which constructively interfering X-rays leave the crystal, the Bragg relation (Equation 3.1) gives the corresponding lattice spacing, which is characteristic of a certain compound. In this study, phase analysis of the supported catalyst and bulk species was performed by X-ray diffraction. Samples were ground using a mortar and pestle in order to make a fine powder. The powder was mixed with ethanol. The mixture was evenly spread on a sample holder (glass slide) and dried in open atmosphere at room temperature for approximately 10 min to make a thin polycrystalline layer of material. After drying, the slide was placed into the diffractometer. The X R D experiment was conducted on the samples, using a Siemens diffractometer with C u K a as X-ray source (k = 1.54 A and energy of 8.04 keV). The diffractometer was set on 40.0 m A and 40.0 k V for all the samples. Figure 3.2a and Figure 3.2b show the X R D equipment used in this study. The X R D pattern o f the powder samples is measured with a movable X-ray source and sample holder as well as a stationary detector, which 42 Chapter 3 - Experimental Apparatus and Methods scans the intensity o f the diffracted radiat ion as a function o f the angle 28 between the i n c o m i n g and diffracting beams. A standard program was used to step 20 f rom 3 to 70 degree (see Figures 3.2a and 3.2b). Figure 3.2a S iemens th in f i l m diffractometer; 29 = 3° ( A : X - r a y Source, B: Sample Ho lde r and C : X - r a y Detector) 43 Chapter 3 - Experimental Apparatus and Methods F i g u r e 3.2b Siemens th in film diffractometer; 29 = 70° R a w X R D data was interpreted us ing EVA software. EVA software contains a vast spectra data base for chemica l components (organic and inorganic) . B y u s ing the program, the sample X R D spectra were compared to the spectra o f probable components spectra i n the data base and hence the bu lk components present were identified by f ind ing a match between the spectra. 3.2.2 Catalyst Crystallite Size Measurement X R D patterns f rom samples were also used for catalyst crystal l i te s ize measurement. F o r example, F igure 3.3 shows the X - r a y diffraction pattern o f the nanocrysta l l ine FeS2 supported catalyst w h i c h exhibi ts s ignif icant l ine broadening. The extent o f b roadening is described by /?, w h i c h is the fu l l w i d t h at h a l f m a x i m u m intensity o f the peak. 44 Chapter 3 - Experimental Apparatus and Methods 600 500 400 ? c o 300 o • 200 100 29 = 33.16 t 31 31.5 32 32.5 33 33.5 34 34.5 35 26 Figure 3.3 Diffraction pattern of the nanocrystalline iron sulphide on alumina support catalyst with 15% iron loading After the value of the /? (in radians) was corrected for instrumental broadening, it was substituted into Scherrer's equation: L = K'X ficosO Equation 3.2 where L is the dimension of the particle in the direction perpendicular to the reflecting plane, X is the X-ray wavelength, 0 is the diffraction angle and AT' is a constant which is often taken as 1 (Niemantsverdriet, 1995). 45 Chapter 3 - Experimental Apparatus and Methods 3.2.3 DRIFT (Fourier Transmission Infrared Spectroscopy) The often used acronym D R I F T or DRIFTS stands for diffuse reflectance infrared Fourier transform spectroscopy. The diffusely scattered radiation is collected by an ellipsoidal mirror and focussed on the detector. The infrared absorption spectrum is described by the Kubelka-Munk function: K" (\-RJ2 S 2R„ where K" = is the absorption coefficient, a function of the frequency v S = is the scattering coefficient i?oo= is the reflectivity of a sample of infinite thickness, measured as a function of v Equation 3.3 In diffuse reflectance mode, samples can be measured as loose powders, with the advantages that not only is the tedious preparation of wafers avoided, but also the gas diffusion limitation associated with tightly pressed samples is avoided. Infrared spectroscopy provides information about the behaviour of the catalyst and pure species surface during the reaction. In this project experiments were conducted in order to study the following: • Production of SO3 on the activated alumina. • Phase study of supported FeS2, FeS, Fe2C>3, FeSC>4 and pure FeS2. Catalysts were ground using a mortar and pestle to the particle size > 180 (am before being placed on the sample holder in the reaction chamber (see Figure 3.4). 46 Chapter 3 - Experimental Apparatus and Methods F i g u r e 3.4 A : Reac t i on Chamber ; direct ion o f reactants stream is specif ied o n the picture. The spectra were measured us ing a Biorad FTS175 spectrometer for D R I F T , w i t h a heated reaction chamber at a resolut ion o f 4 c m ' 1 . The pre-treatment o f the catalyst was started by heating the catalyst f r o m r o o m temperature to 2 5 0 ° C i n v a c u u m (~10" 3 mbar) for 1 hour. T h i s procedure was done to ensure that the surface was c lean and adsorbed contaminates (CO2 and H2O) were desorbed. T h e catalyst was cooled d o w n to r o o m temperature under H e f low. In order to study the effect o f sulphation, Alcan AA-400 act ivated a l u m i n a was used. The premixed reactants SO2/N2 and 0 2 /N2 were del ivered through two separate 5850E Brookes mass f l o w controllers. Eff luent gas was vented into the fume hood . T w o series o f experiments were conducted o n act ivated a lumina . In the first set, SO2/N2 was used as reactant and i n the second set SO2/N2 and 0 2 / N 2 was used as reactants. The D R I F T spectra were measured i n each case and compared w i t h the results f rom literature. T o study the surface properties o f the catalyst, four supported catalysts were compared and D R I F T experiments were conducted on them using C O as probe gas. T h e probe gas ( C O ) was 47 Chapter 3 - Experimental Apparatus and Methods introduced into the reaction chamber (using a 5850E Brookes mass flow controller) at room temperature for 1 hour and at atmospheric pressure. After this, the reaction chamber was evacuated to eliminate free species in the gas phase and to ensure that the spectra were associated with species chemisorbed on the surface of the catalyst. Spectra were measured and compared with spectra obtained with an empty chamber or with catalyst in the absence of the probe gas. Figure 3.5 shows the D R I F T setup used in this study. >Vent Figure 3.5 D R I F T setup used for catalyst surface study; C O , S 0 2 / N 2 and 0 2 / N 2 were introduced into the device using two separate mass flow controllers; He and N 2 were introduced using two separate rotameters. 3.2.4 BET (Total Surface Area Measurement) The surface area and pore size measurements were conducted using a Micromeritics FlowSorb 2300, by determining the quantity of a N 2 gas that adsorbs as a single layer of molecules, so called monolayer, on a catalyst surface. This adsorption is done at or near the boiling point of the N 2 gas. Under this condition, the area covered by each N 2 gas molecule, was 48 Chapter 3 - Experimental Apparatus and Methods known within relatively narrow limits. The area of the catalyst sample was calculated directly from the number of adsorbed N 2 molecules, which is derived from the gas quantity at the prescribed conditions, and the area occupied by each. For B E T measurement a mixture of N 2 / H e (30 vol . % / 70 vol . %) was used. For this N 2 and He mixture, conditions are most favourable for the formation of a monolayer of adsorbed N 2 at atmospheric pressure and -176°C, the temperature of liquid N 2 . 3.2.5 SEM (Scanning Electron Microscopy) In order to confirm the catalyst crystallite size, bulk FeS was also studied by S E M . S E M (scanning electron microscopy) was conducted on the samples using a Hitachi S4700 microscope. Bulk FeS was coated first by a thin layer of gold using a sputter coater machine and after that were placed into the S E M sample holder and loaded into the microscope following a standard procedure. 4 9 Chapter 3 - Experimental Apparatus and Methods 3.3 Experimental Setup for Catalyst Activity Test and Reaction Rate Study 3.3.1 Setup Catalyst activity was measured in the experimental apparatus shown in Figure 3.6. The setup included a fixed-bed down-flow reactor, 9 mm ID and 400 mm length, made of quartz. The quartz tube was connected to a header at the top and a sulphur condenser at the bottom using Swagelok 3/8"stainless steel (SS 316) fittings. To make the reactor gas tight and prevent the quartz reactor from breaking at the connections, graphite ferrules (3/8" ID graphite ferrules) were used for both ends. The reactor was placed inside a Lindberg tube-furnace that had an automatic temperature controller and the temperature was measured using two K-type thermocouples. One thermocouple was mounted at the top o f reactor in a position close to the surface of the catalyst bed, to measure the feed temperature, and the second was located at the bottom of reactor in a position to measure the actual bed temperature. The furnace controller used the feed temperature as the measuring parameter. To remove sulphur from the product stream and prevent any further reaction between sulphur and water vapour (Claus reverse reaction), a stainless steel (SS 316) condenser was installed immediately after the reactor. The temperature of the sulphur condenser was maintained at 130°C using a heating tape connected to a voltage controller and monitored by a K-type thermocouple attached to the body of the vessel (sulphur condenser). In order to remove water vapour from the product stream, a P2O5 drier was installed after the sulphur condenser. P2O5 is a strong water absorbent and is inert during the reaction. Quartz 50 Chapter 3 - Experimental Apparatus and Methods wool was placed in the sulphur trap and water drier to trap any sulphur mist which was not collected by the sulphur condenser (see Figure 3.7). The flowrates of the premixed reactant streams EfeS/l^, SO2/N2 or H2, O2/N2 and pure N2 were controlled by four separate 5850E Brooks mass flow controllers. The reactants were mixed before the reactor to provide the required feed gas composition. A n Omega pressure gauge in the range of 0-30 psi was placed before the reactor to monitor pressure in the line and in the reactor. Catalyst bed P205 Drier Figure 3.6 Experimental setup for catalyst activity tests and reaction rate study 51 Chapter 3 - Experimental Apparatus and Methods F i g u r e 3.7 Su lphur trap and water drier; Quar tz w o o l was used to trap the sulphur mis t and prevent any sulphur carry over into the G C co lumns . F i g u r e 3.8 Sulphur Trap 52 Chapter 3 - Experimental Apparatus and Methods F i g u r e 3.9 Water D r i e r Water was injected into the feed gas us ing a Harvard Apparatus syringe pump ( M o d e l 55-1144) and f lowed through a l / 1 6 " O D stainless steel tube p laced into the tube-furnace and connected to the header o f the reactor. Gas sampl ing was performed at 30 m i n intervals us ing a Valco automatic va lve attached to the G C w i t h a 500 u l s ampl ing loop. Ana lyses o f the feed and product were performed using a GC-14A Shimadzu gas chromatograph equipped w i t h dual co lumns and a thermal conduct iv i ty detector ( T C D ) . Separat ion o f 0 2 and N 2 was performed o n a Molecular Sieves 5A c o l u m n and H 2 S and S 0 2 on a Poropak Q c o l u m n using H e as the carrier gas i n both co lumns and a suitable temperature program, (see A p p e n d i x A ) 53 Chapter 3 - Experimental Apparatus and Methods S A M P L E i COLUMN 4r 2 DETECTOR F i g u r e 3.10a V a l c o va lve before injection; loop sampl ing w i t h two co lumns sequence reversal; C o l u m n 1 is Molecular Sieves 5A and c o l u m n 2 is Poropak Q. SAMPLE SAMPLE LOOP C O L U M N t VENT/ WASTE CARRIER/ MOBILE PHASE DETECTOR F i g u r e 3.10b V a l c o va lve ; inject ion pos i t ion 54 Chapter 3 - Experimental Apparatus and Methods The exit gas from the reactor and the G C passed through a trap containing a concentrated N a O H solution and vented into a fume hood. In order to meet safety regulations the setup and toxic gas cylinders (H2S/N2 and SO2/N2 gas cylinders) were placed inside the fume hood. K Dragor H2S detector was mounted in the fume hood and was set to sound an alarm at 10 ppmv H2S concentrations. 3.3.2 Experimental Procedures • GC-14A Shimadzu Gas chromatograph was turned on 1.5 hours prior to the actual experiment in order to stabilize the output signal. • The catalyst mass was measured using a Fischer Scientific - GRAM ATIC balance with accuracy of 0.0001 g and placed into the reactor using quartz wool as support at the bottom of catalyst bed. • To ensure the system was gas tight, the system was pressurized to 170 - 184 kPa (10-12 psig) using pure N 2 and the pressure was monitored for 30 min. If there was no decrease in pressure after this time period, the system pressure was released to vent and the vent line connected to a bubble meter. • B y using a bubble meter and stop watch the effluent flow rate o f gas was measured and compared to influent flow rate. • The reactor was heated at a rate of 10°C/min to the set point and at the same time flushed with N2. • The sulphur trap and the header of the reactor were heated to 130±1°C and 120±2°C respectively. 55 Chapter 3 - Experimental Apparatus and Methods • The feed gas was introduced to the system using the mass flow controllers. Harvard Apparatus syringe pump was set and turned on to provide the vapour in the feed stream. • The pressure of the system was set at about 115 kPa (2 psig) using the needle valve after the drier. This valve distributed the product gas to the G C and to the vent. B y changing the ratio of these two streams the pressure of the reactor was set to the desired value. • Sampling o f the gas was conducted in 30 min intervals using the automatic Valco sampling valve. • A t the end of each experiment the reactor temperature was increased to 350±1°C and flushed with N2 to vaporize any sulphur deposition on the catalyst. 3.3.3 Calculation and Minimization of Transport Disguises Heat and mass transfer effects, caused by intrareactor, interphase, or interparticle gradients can disguise the results obtained from catalyst tests. Therefore the following calculations and assumptions have been considered to minimize these effects in the present study. Heat transfer gradients are classified by their level of importance as: • Intrareactor (axial) temperature gradient • Intrareactor (radial) temperature gradient • Interphase temperature gradient • Interparticle temperature gradient Mass transfer gradients are classified by their level of importance as: • Intraparticle concentration gradients (Internal diffusion) • Interphase concentration gradients (External diffusion) 56 Chapter 3 - Experimental Apparatus and Methods 3.3.3.1 Temperature Gradients To minimize the axial temperature gradient in the reactor a low concentration of the reactants in the feed stream was used (in most cases the feed stream contained 70 to 90% inert gas). Also a K - type thermocouple was placed in the catalyst bed to monitor the temperature rise in the bed. Based on experimental temperature observations it was concluded that axial temperature rise was negligible and the assumption of an isothermal reactor was valid. Due to the small particle size (48 mesh) and the very high space velocity (40,000 to 70,000 h"1) the intrareactor (radial), interphase and interparticle temperature gradients were considered negligible as well . 3.3.3.2 Mass Transfer Gradients To ensure that the flow regime in the reactor was plug flow the rule of thumb — > 50 dP was considered, where L is catalytic bed height and dP is the particle diameter. L _ 1 . 9 x l Q - 2 dP ~ 2 . 9 7 x 1 0 ' For supported catalyst — = ——— = 64 and L 1 9x 10 2 For pure FeS catalyst — = — — = 475 which both met the above criterion. d„ 4 x l 0 - 5 57 Chapter 3 - Experimental Apparatus and Methods 3.3.4 Reactor Operating Regime High conversion (conversions more than 5%) and the magnitude of error for this measurement associated with low 0 2 concentration did not allow operating the reactor under differential mode. Therefore the reactor was operated under integral mode. In differential mode, the definition of the rate is: A C AW/F and for the integral mode: Equation 3.4 dC r = Equation 3.5 dWIF M where C is the concentration of O2 , W is the weight of catalyst and F the feed mole flowrate to the reactor. In the integral mode the concentration of the reactants changed through the reactor. Measuring the concentration of reactants at each point inside the catalytic bed was not feasible. Therefore, all conversion, selectivity and rate of reactions were calculated based on the inlet and the outlet reactants concentration to the reactor (catalytic bed). Consequently, the conversion, selectivity and rate o f reactions are reported as average (apparent) measurements rather than point measurements in the figures and plots of the following chapters. For example average rate of O2 consumption and H2S conversion were defined as: Average 0 2 Reaction Rate (mol / s.g) = (Molar F low 0 2 In - Mola r F low 0 2 Out) / (Weight of Catalyst) Average H 2S Conversion % = (Molar F low H 2 S In - Molar flow H 2 S Out) / (Molar F low H 2 S In) x 100 Average S 0 2 Selectivity % = (Molar F low S 0 2 Out) / (Molar F low H 2 S In - Mola r flow H 2 S Out) x 100 Note, however, that when using H Y S Y S simulator to estimate the rate constants, the change in concentration of reactants through the reactor was respected and calculations were based on complete integral mode. 58 Chapter 4 - Results and Discussion Chapter 4 - Results and Discussion 4.1 Introduction The present study was inspired from a test of the O2 scavenging capability of three Fe-based catalysts. The catalysts were provided by Porocel and the experimental conditions were chosen based on information also provided by Porocel. Results from these tests raised questions regarding the active phase or phases of the catalysts, the reaction mechanism and the reaction kinetics. To answer these questions the study evolved in six phases, the results from one phase leading to the work carried out in subsequent phases. In the first phase, Fe oxides (Porocel catalysts) were tested for oxygen scavenging. Experiments revealed an unsatisfactory performance of the Porocel Fe-oxide catalysts. Also , it was observed that in a few cases, the Fe phase changed from oxide to sulphide during the experiment and increased the catalyst oxygen scavenging capability. From the literature review and based on these initial results, it was speculated that Fe sulphide was a better catalyst for O2 scavenging than Fe oxide. Therefore, in the second phase of the present work, Porocel Fe catalyst was preconditioned in an H2/H2S gas stream and the resulting catalyst was tested against the Porocel Fe-oxide catalyst. In the third phase, in order to study the state of Fe on the alumina support and to determine the changes that occurred during reaction and pre-treatment, Fe based catalyst on the V-AI2O3 was prepared at different Fe loadings. Characterization of the prepared catalysts was conducted using DRIFT, X R D , B E T and S E M measurements, before and after preconditioning and Claus reaction. 59 Chapter 4 - Results and Discussion In the fourth phase, performance of the prepared catalyst (Fe-sulphide on Y-AI2O3) for O2 scavenging was investigated. It was concluded that because some reactions occurred on the y-AI2O3 support, the interpretation of results was difficult in respect of the Fe-sulphide active phase. Therefore, to eliminate the effect of Y-AI2O3, it was decided to investigate bulk Fe-sulphide as the catalyst instead of a supported Fe-sulphide catalyst. In the fifth phase of the work, the effect of reactant concentration on the activity of the bulk Fe-sulphide was determined. The rate of 0 2 consumption was measured while the concentration of O2, H 2 S and H 2 0 were varied. Also , the effect of temperature on the rate of O2 consumption was investigated. Experiments were conducted at three different temperatures within the range of commercial conditions. In the final phase, the results from the experimental work on the bulk Fe-sulphide were used to construct O2 consumption kinetic expressions and also to develop a mechanism to explain the kinetics of the reactions. Results and discussion of the work conducted through each of these six phases are described in the following subsections. 4.2 Porocel Fe-based Catalyst Performance Test The catalysts and experimental conditions were based on information provided by Porocel. Porocel provided three Fe-oxide commercial catalysts, Fe-1, Fe-2 and Fe-3. The reactor was maintained at 250 ± 1°C in the initial set of experiments and at 280 ± 1 °C for the supplemental set of experiments. The reactor feed gas was mixed just prior to entering the reactor. A controlled flow of a pre-mixed H2S/N2 (15.16% H2S by volume) gas was mixed with a controlled flow o f pre-mixed SO2/O2/N2 (5.10% S 0 2 and 0.74% O2 by volume) and water was injected into the S02/0 2/N2 gas stream to make a feed with gas composition as presented in 60 Chapter 4 - Results and Discussion Table 4.1. Total gas volumetric flow was 360 cc (STP)/min in the first set of data and 110 cc (STPVmin in the supplemental set of experiments. H 2 S so2 o2 N 2 H 2 0 4.5% 2.25% 0.33% 66.78% 26.13-27.03% Table 4.1 The actual flow condition (% vol.) Experimental results in Tables 4.2, 4.3 and 4.4 showed a low oxygen scavenging activity for all three Fe-based catalysts vs. time. Time Conversion, Pressure (min) o2 H 2 S so2 (kPa) 15 11 75 63 225 45 13 53 44 225 75 13 46 39 225 105 18 41 35 274 Table 4.2 Activi ty o f Fe-1 catalyst; 360 cc (STP) /min; 250°C; 1.5 g catalyst Time < Conversion, Pressure (min) o2 H 2 S so2 (kPa) 15 14 72 53 232 45 9 56 48 232 75 9 50 44 232 105 10 50 44 232 Table 4.3 Activi ty of Fe-2 catalyst; 360 cc (STPVmin; 250°C; 1.5 g catalyst Time < Conversion, Pressure (min) o2 H 2 S so2 (kPa) 15 37 93 81 222 45 33 80 65 239 75 12 79 58 253 105 14 60 50 274 Table 4.4 Activi ty o f Fe-3 catalyst; 360 cc (STPVmin; 250°C; 1.5 g catalyst 61 Chapter 4 - Results and Discussion The apparent low activity of the Fe based catalysts reported above may arise from two possible sources. In the first instance, as noted above, the reactor pressure was relatively high during these tests and at the reaction temperatures, condensation of S onto the catalyst particles was possible. Apart from visual evidence for the presence of S on the catalyst after reaction, the low surface area of the used catalyst is also indicative of sulphur deposition. For example the B E T surface area of the Fe-2 decreased from 230 m 2 /g to 68 m 2 /g . In the second instance, the catalysts were tested without any pre-treatment and the reactor was operated for the relatively short period of 2 to 2.5 hours time-on-stream. Consequently, the initial state of the iron, which from the color of the catalyst appeared to be Fe203, may not change during reaction, and indeed the appearance of the used catalyst was the same as that of the fresh catalyst. Consequently, i f Fe203 was not the active phase for the oxidation reaction, then the conversions would be low, as observed. In an attempt to address the above two issues, the tests of the Fe-2 and Fe-3 catalysts were repeated under conditions where the reactor pressure was approximately 115 kPa. In addition, the reactor tests were operated for a period of about 4 to 6 hours time-on-stream, and finally, the effect of a 3 hour pre-treatment of the catalyst in a H2S/H2/N2 mixture at 673K was determined. These supplemental activity tests were also done at lower space velocity (4400 h"1) using a total gas flowrate of 110 cc(STP)/min versus a G H S V of 14,400h_ 1 and a flowrate of 360 cc(STP)/min used in the initial activity tests. In addition, higher reaction temperatures (280°C versus 250°C) were used. The more severe reaction conditions were chosen to obtain higher oxygen conversions that were more easily quantified. The conversions of the feed gas components at these conditions in the empty reactor were 24% O2, 4% H2S and 1% SO2. Pre-conditioning o f the catalysts, or reduction-sulphidation, refers to the simultaneous reduction-sulphidation of the catalyst by a H2S/H2/N2 (10/40/50 vol . %) mixture reacting over 62 Chapter 4 - Results and Discussion the oxide catalyst for 3 h at 400°C. The activity of the catalysts treated in this way was compared to that obtained without pre-treatment and the results are summarized in Figure 4.1. O Fe3-Without Precond. AFe3-Wi th Precond. •Fe2-Wi thout Precond. • F e 2 - W i t h Precond. Figure 4.1 Comparison of oxygen scavenging using Fe-2 and Fe-3 with and without preconditioning The data show that pre-conditioning the catalysts increased the oxygen conversion by 5 to 10 percentage points, and in all cases the catalysts were much more active than the empty reactor. In addition, in this second set o f experiments, the catalyst activity remained well above the thermal reaction activity even after 6 hours time-on-stream. This observation suggests that 63 Chapter 4 - Results and Discussion the higher reaction pressure o f the in i t ia l experiments increased S deposi t ion on the catalyst w h i c h in turn resulted in a loss o f catalyst act ivi ty, whereas in the supplemental experiments conducted at l o w pressure, no such rapid deposi t ion occurred and hence a severe loss in act ivi ty was not observed. The catalyst also showed evidence o f i ron sulphide format ion i n the case o f the precondi t ioned material as shown in Figures 4.2, 4.3 and 4.4. The used catalyst that was not pre-condi t ioned, showed signs o f the presence o f red i ron oxide , and was s imi la r in co lor to that o f the unreacted catalyst. 1 1 SjB? O — 3 - = r o —EE CO - = 4* -== • F i g u r e 4.2 Poroce l i ron oxide catalyst (Fe-3) before react ion » I — I F i g u r e 4.3 Poroce l i ron oxide catalyst (Fe-3) after reaction showed the presence o f different phase o f i ron w i th black color on the catalyst. 64 Chapter 4 - Results and Discussion I ' 1 F i g u r e 4.4 Poroce l catalyst (Fe-3) after precondi t ioning showed that i ron oxide phase i n the bright red color transformed to b lack color . Results showed that precondi t ioning o f the Poroce l Fe-based catalyst increased oxygen scavenging. P r i m a r i l y it was concluded that precondi t ioning o f the i ron ox ide in the stream o f H2S/H2/N2 changed the phase o f i ron probably to i ron sulphide. B a s e d on literature reports (see Chapter 2) there is evidence that suggests that O2 consumpt ion over i ron sulphide occurs by React ion 4.1 and 4.2 whereas over i ron oxide only Reac t ion 4.1 occurs: H2S + - 0 2 < - > - £ „ +H20 Reac t ion4 .1 2 n S + <92 -> S02 Reac t ion 4.2 Therefore, based on the s toichiometry o f the reactions, it was speculated that the iron sulphide was able to convert about three times more oxygen compared to the i ron ox ide catalyst. Fo r this reason i ron sulphide and i ron oxide supported on y -a lumina were investigated for the removal o f oxygen and both have been found to be active for selective convers ion o f oxygen in 65 Chapter 4 - Results and Discussion the Claus process. The iron sulphide catalyst was prepared by preconditioning the Fe-3 catalyst in H 2 S / H 2 / N 2 (10/40/50 vol . %). A performance test of the oxidation of H 2 S by 0 2 on the iron sulphide supported catalyst at temperature range between 220°C and 320°C was carried out. In the first series, 0.3 vol . % of oxygen was used in the feed composition. In the course of the experiments iron oxide catalyst was converted to iron sulphide because of temperature (300°C) and the strong reducing environment (higher concentration of H 2 S than 0 2 ) . To overcome this problem in the second series of experiments oxygen was increased to 2.23 vol . % in the feed to create an oxidizing environment and maintain a stable iron oxide phase. Performance tests at 300°C showed that for the feed stream o f 1.5vol. % H 2 S , 2.2 vol . % 0 2 and 96.3 vol . % N 2 the iron sulphide catalyst converted 94% of the oxygen compared to iron oxide with 3 7 % 0 2 conversion. 1 * A I I ~L I - p — i i 1 60 120 180 240 Time(min) F i g u r e 4.5 0 2 scavenging activity over iron sulphide supported catalyst vs. Time; feed composition: 1.5vol. % H 2 S , 2.2 vol . % 0 2 and 96.3 vol . % N 2 at 300°C temperature and 115 kPa pressure 90 80 •2 70 > c o u ffi 2 > < 60 50 40 30 -20 -10 66 Chapter 4 - Results and Discussion Figure 4.5 shows 0 2 conversion of > 90% over the iron sulphide supported catalyst. The iron oxide catalyst (Figure 4.6) showed an initial high conversion of O2 but the conversion of O2 decreased to about 38% after 150 min. The relative performance of the catalysts was in very good agreement with the assumption that iron sulphide supported catalyst can convert three times more oxygen than iron oxide supported catalyst (see Reaction 4.1 and 4.2). 100 1 90 80 c O 70 l > 60 c 0 u 50 O 40 o> 2 30 a> 20 10 0 60 120 180 Time(min) 240 300 360 F i g u r e 4.6 O2 scavenging activity over iron oxide supported catalyst vs. Time; feed composition: 1.5vol. % H 2 S , 2.2 vol . % 0 2 and 96.3 vol . % N 2 at 300°C temperature and 115 kPa pressure Iron sulphide showed significantly better catalytic performance than the iron oxide. Catalytic reaction of H 2 S with 0 2 on the surface of iron sulphide catalyst is in the favor of S 0 2 production and high O2 consumption. In contrast, the surface of iron oxide catalyst is in favor of 67 Chapter 4 - Results and Discussion S production and lower O2 consumption. Figure 4.7 and Figure 4.8 show the S and SO2 yield versus time using iron oxide and iron sulphide supported catalyst, respectively. From the results, it was concluded that iron sulphide supported catalyst was a more active catalyst than iron oxide for oxygen consumption. The higher activity is due to the conversion of H2S to S 0 2 in sequential reactions - Reaction 4.1 and Reaction 4.2, whereas the iron oxide supported catalyst converted H2S to S and water vapor via Reaction 4.1. 2 > a> O) ra > < 60 120 T i m e ( m i n ) • S 0 2 ; • S 180 240 F i g u r e 4.7 Average Y i e l d % of S and S 0 2 over iron sulphide supported catalyst vs. Time; feed composition: 1.5vol. % H 2 S , 2.2 vol . % 0 2 and 96.3 vol . % N 2 at 300°C temperature and 115 kPa pressure 68 Chapter 4 - Results and Discussion 2 a> '> a O) « 0) > < 60 120 180 T i m e ( m i n ) • S 0 2 ; • S 240 300 360 Figure 4.8 Average Y i e l d % of S and S 0 2 over iron oxide supported catalyst vs. Time ; feed composition: 1.5vol. % H 2 S , 2.2 vol . % 0 2 and 96.3 vol . % N 2 at 300°C temperature and 115 kPa pressure For both cases conversion of the H 2 S was reported around 100%. Figure 4.9 shows the H 2 S conversion over both iron oxide and iron sulphide catalyst. 69 Chapter 4 - Results and Discussion 100 80 S 60 > c o u a> 40 o> re i _ < 20 • H2S conversion on iron oxide • H2S conversion on iron sulphide 60 120 180 T i m e ( m i n ) 240 300 360 F i g u r e 4.9 Average H2S conversion over iron oxide and iron sulphide catalysts vs. Time; feed composition: 1.5vol. % H 2 S , 2.2 vol . % 0 2 and 96.3 vol . % N 2 at 300°C temperature and 115 kPa pressure 4.3 Characterization 4.3.1 Phase Characterization of Fe on AI2O3 using XRD The iron supported alumina catalyst was prepared based on the preparation procedure described in Chapter 3. X R D (X-ray Diffraction) was used to determine the phase composition of the solid. In order to enhance the intensity of the diffraction pattern, catalysts with an iron loading of 15 wt. % on the alumina was used in all characterization experiments*. •Note: For confidentiality reasons Fe-1, Fe-2 and Fe-3 catalysts provided by Porocel could not be characterized. 70 Chapter 4 - Results and Discussion The alumina supported iron oxide scanned by X R D after calcination (at 673 K ) was found to be in the form o f Y-Fe203 (magnetite) with rhombohedral structure whereas unsupported iron oxidized to a-Fe203 (hematite) at the same temperature. According to Thomson et al. (1991) the reason that y-Fe203 formed during oxidation of the supported iron, is due to impurities in the iron compounds which, in this case, are alumina ions. During the impregnation procedure it is likely that some of the A l in Y-AI2O3 was dissolved in the acidic iron nitrate solution and this dissolved A l is thought to be incorporated into the matrix of the iron oxide. It is known that the y-Fe2C»3 spinel lattice is thermodynamically stabilized by reticule impurities and by other crystalline imperfections (Vaishnava, 1985). The crystal chemical transformation mechanism is proposed from iron oxide to iron sulphide as follows: Calcination at 673 K and iron oxide formation can be proposed as: 2Fei+ (N03Y +°2 > Feu -O2- -Fei+ + NO,(g) Reaction4.3 Figure 4.10 shows the X R D data for alumina supported iron oxide after calcination in an oxygen environment at 673 K for 4 hours. The X R D pattern shows Fe2C»3 as the active phase of iron on the alumina support. Preconditioning (reduction-sulphidation) of the iron oxide precursor was conducted at 673 K using H 2 S / H 2 / N 2 (10/40/50 vol . %). Figure 4.11 shows the X R D data for Fe2C>3/ 7-alumina after preconditioning, indicating that the Fe2C>3 phase was transformed to F e x S y (Pyrthotite) with x = 0.90-0.95 and y = 1 with hexagonal structure. 71 Chapter 4 - Results and Discussion F i g u r e 4.10 X R D data for alumina supported iron oxide catalyst after calcination at 673 K in oxygen present environment; loading of the iron was 15 wt. % on the catalyst 500 F i g u r e 4.11 X R D data for alumina supported iron oxide (15 wt. % Fe) after preconditioning in H2S/H2/N2 stream at 673K; transformation of iron oxide species on the support to iron mono-sulphide 72 Chapter 4 - Results and Discussion The mechanism of transformation was proposed as follows. This mechanism was also proposed by Inamura et al. (1993) to explain the sulphidation o f iron oxide precursor on Y -Zeolite. In this mechanism Fe "(*) is still bonded to some framework oxygen atoms. Fe3+ (OH)- +1/2 H2 -> Fe2+ (*) + H20 Fei+ (SHY +1/2 H2 Fe2+ (*) + H2S Fe2+(*) + H2S -> FexSY +H20 X = 0.95,F = 1 or combination of following reactions: Fe2+ +H2 -+Fe0+H2O Fe° +H2S^FexSY + H2 X = 0 .9 -0 .95 ,7 = 1 Reaction 4.4 Reaction 4.5 Reaction 4.6 Reaction 4.7 Reaction 4.8 Reaction 4.9 In order to study the effect of H2S and H2 on the reduction-sulphidation reaction, the catalyst (Fe203/ y-alumina) was preconditioned in the absence of H 2 in a stream of H2S/N2 (10/90 vol. %). Figure 4.12 shows the X R D data for the catalyst after this preconditioning. In the absence of hydrogen at 673K, Fe203/ y-alumina catalyst was transformed to FeS2/ y-alumina. The mechanism of this transformation was proposed as follows: Fei+ - O2- - Fe3+ + H,S -> Feu - S2~ - Fe3+ + H,0 Reaction 4.10 73 Chapter 4 - Results and Discussion In this mechanism, Fe species are not reduced and sulphur replaces oxygen in the lattice. Results also showed that traces of Sg (sulphur) were present in the catalyst which was likely due to the instability of H2S. H2S is not a very stable molecule and can dissociate into H 2 and S2 (Arnoldy et al., 1985) and S2 species combine to form Ss that fouls the catalyst. This reaction, however, is suppressed for the most part by the presence o f H 2 in the preconditioning (sulphiding) mixture. 500 20 30 40 50 60 70 26 Figure 4.12 X R D data for alumina supported iron oxide (15 wt. % iron) catalyst after preconditioning in H2S/N2 stream at 673K; Traces of Sg (orthorhombic structure) fouling in the catalyst These observations are in the very good agreement with the result from the preconditioning of the catalyst in the presence of H2 (see Section 4.2) which showed no traces of Sg in the catalyst. Also it may be concluded that H 2 is the key species that determines the transformation of the iron oxide phase to FeS or FeS2 on the support. 74 Chapter 4 - Results and Discussion Different sulphiding reactions can take place at different temperatures and accurate study of the effect of H2S and H2 would require a study by T P R (Temperature Programmed Reaction) which was beyond the scope of the present research. However, a general guide concerning the stability of various phases was constructed using ASPEN PLUS simulation software using Gibbs reactor and performing an equilibrium calculation at atmospheric pressure. In this simulation, iron oxide was defined as pure species and the concentration of H2S and H2 in the feed stream at balance N2 as well as temperature were varied. The results of the simulation are provided in Figure 4.13. In the following the sulphiding reactions w i l l be discussed by temperature region. (i) High temperature sulphiding: Above 1100 K , sulphiding results in mono-sulphide iron (FeS) at any concentration of the H2S. This result is independent of the concentration of H2 in the feed stream. (ii) Medium temperature sulphiding: In the range of 650 to 1100 K the active phase is very sensitive to the presence of H2. FeS2 is converted to FeS even at a low concentration of H2. (iii) L o w temperature sulphiding: Below 650 K , Fe in the form FeS2 is converted to FeS i n the presence of H2, but the effect of H2 concentration is not as strong as in medium temperature range. Cahi l l et al. (2000) proposed two primary pathways for transformations of iron mono-sulphide to iron di-sulphide: 2FeS + 2H+ -> FeS2 + Fe2+ + H2 (g) Reaction 4.11 75 Chapter 4 - Results and Discussion 1200 1100 4 1000 H 900 2 800 £ 700 600 500 400 High - temperature sulphiding Medium - temperature sulphiding F e S 2 Low- temperature sulphiding 15% H 2 S 10% H 2 S 5% H 2 S 10 20 30 40 50 60 H 2 Concentrator^ vol.%) 70 Figure 4.13 Reduction-Sulphidation reaction guideline simulated using A S P E N P L U S software at atmospheric pressure; lines = phase transfer border Such a reaction was suggested by Berner (1970), and examined further by Schoonen and Barnes (1991), Furukawa and Barns (1995) and most recently, by W i l k i n and Barnes (1996). A second proposed pathway by which iron mono-sulphide (FeS) transforms to iron di-sulphide (FeS 2) is via the H 2 S route (Taylor et al. 1979; Wachtershauser, 1988; Drobner et al, 1990; 76 Chapter 4 - Results and Discussion Rickard, 1997) in which H2S is the oxidant. The latter mechanism is valid for converting FeS to FeS2 during the reaction. FeS + H2S -> FeS2 + H2 (g) Reaction 4.12 Based on the proposed mechanism H2 prevents the formation of FeS2 and this is in very good agreement with both the experimental results reported here, in the absence and presence of H 2 , and the phase diagram presented as Figure 4.13. 4.3.2 Phase Characterization of Fe on AI2O3 using DRIFT In order to study the phase of iron present on the supported catalyst after calcination and preconditioning, as well as during reaction under Claus conditions, and to identify the possibility of catalyst poisoning, another characterization procedure was followed. DRIFTS of C O chemisorption on four types of catalyst were conducted: 1) Sample 1: Iron oxide catalyst after calcination (Fe203 / alumina) with 15 weight% iron loading. 2) Sample 2: Iron oxide catalyst after reduction-sulphidation using H2S/H2/N2 stream (FeS/alumina). 3) Sample 3: Iron sulphide catalyst after 4 hours reaction in H2S/O2/N2 (3/5/92 vol . %) stream at 573 K (FeS2 / alumina). 4) Sample 4: Sulphated catalyst prepared using FeS04 precursor (FeS04/alumina). 77 Chapter 4 - Results and Discussion 0.3 2300 2200 2100 2000 1900 F i g u r e 4.14 Pattern of CO absorption on the four types of catalyst using D R I F T On Fe203 and FeS04, adsorption of CO at room temperature resulted in a broad adsorption band with two clear peaks at 2005 and 2027 cm"1 (see Figure 4.14). CO interacted with basic O2" sites to lead, at room temperature, to the formation of surface 'carbonite', CO22" and this appeared as a broad peak in the range of 1850 to 2050 cm"1. In contrast, in the absence of O 2' , there was no adsorption of CO observed in the case of supported F e S and FeS2. This result showed that C O can be used to characterize the catalyst before and after preconditioning and also it can be easily used to identify catalyst poisoning by the presence of F e S 0 4 . Figure 4.15 and Figure 4.16 show the structure of Fe 2 0 3 and FeS04 and the location of basic O " in the structure. 78 Chapter 4 - Results and Discussion F i g u r e 4.15 Structure o f F e 2 0 3 and locat ion o f basic 0 2 "(b lue = Fe ; red = O ) F i g u r e 4.16 Structure o f F e S 0 4 and locat ion o f basic O 2 " (blue = Fe ; red = O ; y e l l o w = S) 79 Chapter 4 - Results and Discussion The results from Sample 3 showed that the iron sulphide catalyst, operated under Claus conditions and in presence of oxygen, was stable and didn't convert to iron sulphate. 4.3.3 Study of the Stability of Catalyst using XRD In order to study the stability of the catalyst in the presence of oxygen, experiments were conducted on alumina supported FeS. In these series of experiments, preconditioned catalysts were tested using H 2 S / 0 2 / N 2 (3/4.5/92.5 vol . %) feed stream. After 4 hours at 300°C, the catalysts were characterized by X R D and the diffraction pattern o f the catalyst showed that FeS 2 was still present. Figure 4.17 shows the X R D pattern of FeS after 4 hours reaction that had transformed to FeS 2 . 500 20 30 40 50 60 70 26 F i g u r e 4.17 Alumina supported FeS catalyst after 4 hours reaction in H 2 S / 0 2 / N 2 (3/4.5/92.5 vol . %) feed stream at 300°C 80 Chapter 4 - Results and Discussion 4.4 Role of FeS2 (Metal Sulphide) and A1 20 3 (Catalyst Support) The catalysts in the present study were prepared as iron precursor on alumina support (see Chapter 3). Alumina is a well known active catalyst for the Claus reaction (Reaction 4.13). Therefore, it was speculated that H2S oxidation reactions (Reaction 4.1 and 4.2) occurred on the active iron sites whereas the Claus reaction (Reaction 4.13) occurred on the alumina support. Consequently the catalyst is bifunctional. 2H2S + S02<^-Sn + 2H2 O Reaction 4.13 n To confirm the proposed bifunctional activity o f the catalyst, separate tests of the support (alumina) and FeS2/alumina (with 5% iron loading on the catalyst) were conducted under Claus reaction conditions. The results with alumina are shown in Figure 4.18, obtained using 0.1179 g of the alumina support in 4 vol . % H 2 S , 2 v o l . % S 0 2 and 1 v o l . % 0 2 , balance N 2 at 300°C and 115 kPa. The gas hourly space velocity was kept at 68,000 h" 1. Alumina alone showed little activity for the conversion of 0 2 (approximately 20% conversion) but high activity for the Claus reaction and conversion of the H 2 S and S 0 2 , according to Reaction 4.13. Conversion of H 2 S and S 0 2 initially was 53% and 52%, but decreases to reach a steady state conversion of 46% and 42%), respectively, after 2 hours. Oxygen conversion was approximately 20% and it was assumed that part of this conversion was due to thermal reactions with H 2 S . 81 Chapter 4 - Results and Discussion 100 90 80 70 -| c o 1 60 0 > 8 50 a o> a 40 > < 30 20 A • S0 2 • H 2 S ; A 0 2 * A • A A • A • A * A A • • A A A A A A A 30 60 90 120 150 Time(min) 180 210 240 270 Figure 4.18 Performance of alumina support under Claus condition vs. Time; feed composition: 4 vol . % H 2 S , 2 v o l . % S 0 2 and 1 v o l . % 0 2 , balance N 2 at 300°C and 115 kPa; measurement error was ± 5% In the second series of experiment 0.1193 g of the FeS 2 on alumina support was tested under two separate conditions. In the first, using a feed stream of 4 vo l . % H 2 S and 2 vol . % S 0 2 , balance N 2 at 300°C and pressure of 115 kPa and in the second series, 4 vol . % H 2 S , 2 vol. % S 0 2 and 1 vol . % 0 2 , balance N 2 at 300°C and pressure of 115 kPa. Figure 4.19 shows the results of the experiments. In the experiment using F e S 2 / A l 2 0 3 catalyst it was observed that in the absence of oxygen, average conversion of H 2 S and S 0 2 reached a steady state level of 62% and 77%, respectively. B y introducing 0 2 into the feed stream, the conversion of H 2 S decreased to 59% and S 0 2 to 25%. This drop in S 0 2 conversion was because of the production of S 0 2 based on Reaction 4.2. A t the same time, 0 2 conversion was reported as 78%). Based on these results it is suggested that 82 Chapter 4 - Results and Discussion Reaction 4.13 occurs primarily on the alumina support and Reaction 4.1 and 4.2 occur mainly on the FeS 2 . Schematic 4.1 shows how the species react on the surface of the alumina supported FeS 2 (see Appendix C for detail calculation). 100 i 90 80 70 C o 'i 60 > c o 50 O <D OJ 40 E > < 30 20 10 0 Claus process - without 0 2 Claus process - with 0 2 A • A A A A A A • • • • • • • • A A A A A S 0 2 ; • H 2 S ; A 0 2 30 60 90 120 150 180 Time(min) 210 240 270 300 330 F i g u r e 4.19 Performance of F e S 2 / A l 2 0 3 catalyst under Claus condition in presence and absence o f the 0 2 ; (see text for experimental conditions); measurement error was ± 2 % In the present research the objective was to study the rate of Reaction 4.1 and 4.2 on FeS 2 sites, therefore, in order to disguise the bifunctional behaviour of the catalyst, bulk FeS 2 was used as the catalyst by preconditioning bulk FeS. 83 Chapter 4 - Results and Discussion H 2 S H 2 S + 0 2 so 2 1 / . S + H 2 0 / FeS 2 A 1 2 0 3 Scheme 4.1 Claus reaction occurs on the alumina (A1 2 0 3 ) and FeS 2 active site, 4.5 Summary In this chapter evolution of the present study was described and the result of the work in this chapter can be summarized as: 1. Commercial Fe-oxide catalysts (Porocel) showed little activity for 0 2 consumption under the Claus reaction conditions. 2. Activity of the Fe-oxide catalyst for 0 2 scavenging improved by preconditioning using H 2 S / H 2 / N 2 gas stream. 3. Fe based catalyst supported on A l 2 0 3 was prepared at different Fe loadings. Prepared catalysts were characterized by X R D and D R I F T methods before and after whereas 0 2 scavenging occurs on the active FeS 2 site. 84 Chapter 4 - Results and Discussion preconditioning in H2S/H2/N2 gas mixture. The possibility o f Fe phase change was suggested and a general guide to the phase change was simulated. 4. Characterization of the catalyst was developed using D R I F T and C O as a probe gas. Results showed that C O can be used to identify the Fe phase change from oxide (O2") to sulphide (S"2) and sulphide to sulphate (SO42"). 5. FeS2 is stable under typical Claus conditions when exposed to O2 leakage (0-5 vol . %). 6. FeS2 supported on the AI2O3 was considered as a bifunctional catalyst, due to reaction o f O2 consumption mainly on FeS2 sites and Claus reaction mainly on the AI2O3 sites. 85 Chapter 5 - Reaction Kinetics & Mechanism Chapter 5 - Reaction Kinetics & Mechanism 5.1 Introduction The purpose of the work covered in this chapter was to study the rate of oxygen consumption on FeS2 at typical Claus reaction conditions. Initially, experiments were performed using a laboratory scale, plug flow reactor (see Chapter 3) with a feed composition of O2, H 2 S , H2O and N2 in order to study the effect of concentration of each reactant on O2 consumption rate. Later, it was decided to use the kinetic data to develop a simple power-law reaction rate expression for O2 consumption. Also , a mechanistic approach was used to explain the kinetics of the reaction. 5.2 Characterization of the Active Iron Sulphide In the previous section, it was concluded that FeS on AI2O3 support was converted to FeS 2 in the course of the Claus reaction and in the presence of a H2S r ich feed. Therefore, in the present study, FeS is the initial form of the catalyst and FeS2 is the active site in respect of the reactions. In order to eliminate the bifunctional phenomena of the supported catalyst shown in the previous section, bulk FeS was used. The particle size of the bulk FeS and supported FeS were compared using X R D and S E M measurements. Figure 5.1 and Figure 5.2 show the X R D pattern of bulk FeS and supported FeS (5% Fe loading) used for particle size measurement. 86 Chapter 5 - Reaction Kinetics & Mechanism 87 Chapter 5 - Reaction Kinetics & Mechanism U s i n g Scherrer ' s equation (see Chapter 3) and extent o f broadening at 29 = 43° , the particle size for bu lk FeS and supported F e S were calculated as 300 A (30.0 nm) and 106 A(10 .6 nm), respectively. A l s o , S E M o f the bu lk F e S shown i n F igure 5.3 conf i rms that F e S was constructed o f smal l size particles w i t h partial porosity. F i g u r e 5.3 Surface o f bu lk F e S as seen b y S E M 8K Chapter 5 - Reaction Kinetics & Mechanism 5.3 Kinetic Study 5.3.1 Preliminary Test The kinetics of Reaction 5.1a and Reaction 5.2b: H2S + - 0 2 ->-S„ + H20 Reaction5.1a 2 n S + 02^> S02 Reaction 5.1b were studied on bulk FeS2. Operating conditions were set in the range that represented typical Claus catalytic converter reaction conditions. The O2 reaction rate was studied by varying H2S and H2O concentrations in the range of 6 to 8 vol . % and 10 to 30 vol . %, respectively, the balance being inert N2. To minimise the measurement error, an O2 concentration range of 1 to 4 % was used rather than the typical range in Claus converters reported (<1%). Three temperatures (250, 270 and 290°C) were used to study the effect of temperature on reaction rate constants. Table 5.1 shows the actual experimental conditions used in this study. 89 Chapter 5 - Reaction Kinetics & Mechanism Feed Inlet Concentration (vol. %) Temperature °C(±1) Pressure kPa(±5) Weight g(±0.02) Exp. Code o2 H 2S H 2 0 N 2 1 6 20 73 270 115 0.23 E01 2 6 20 72 270 115 0.23 E02 3 6 20 71 270 115 0.23 E03 4 6 20 70 270 115 0.23 E04 2 6 20 72 270 115 0.23 E05 2 8 20 70 270 115 0.23 E06 2 10 20 68 270 115 0.23 E07 2 6 10 82 270 115 0.23 E08 2 6 20 72 270 115 0.23 E09 2 6 30 62 270 115 0.23 E10 2 6 20 72 250 115 0.23 E11 2 6 20 72 270 115 0.23 E12 2 6 20 72 290 115 0.23 E13 Table 5.1 Experimental conditions used for studying the rate of O2 consumption; Weight = weight of bulk FeS 5.3.2 Catalytic Activity Test Once the operating conditions shown in Table 5.1 were established, a series of experiments was completed in which concentration of one of the reactants (H2S, O2 or H2O) was changed, leaving the others constant. The effect of the reverse reaction in Reaction 5.1a was not considered because the reaction was always operated under conditions far from equilibrium. Also, sulphur was removed from the effluent of the catalytic bed, using a sulphur trap to eliminate any undesired reaction between H2O and O2 (see Appendix D for detailed calculations for this Section). 90 Chapter 5 - Reaction Kinetics & Mechanism 5.3.2.1 Influence of O2 Concentration The O2 average reaction rate obtained on the bulk FeS2 as a function of O2 concentration in the feed is presented i n Figure 5.4. In this experiment the feed composition was H2S 6 vol . %, H 2 0 20 vol . %, balance N2 while varying the O2 concentration at 1, 2, 3 and 4 vol . %. Temperature of the reactor was 270°C and pressure was 115 kPa. Results show that 0 2 reaction rate increased with increasing O2 concentration in the feed. The SO2 selectivity and H2S conversion in this series o f experiments are presented in Figure 5.5. Results show that increasing the O2 concentration increased both the SO2 selectivity and H2S conversion. • • • 1 1 1 1 0 1 2 3 4 5 0 2 inlet feed concentration (vol.%) Figure 5.4 Effect of O2 feed inlet concentration on O2 reaction rate; feed composition was 6 vol . % H2S and 20% H 2 0 , balance N 2 ; temperature was 270°C and pressure was 115 kPa 91 Chapter 5 - Reaction Kinetics & Mechanism The results are in very good agreement with the stoichiometric order of the O2 in Reaction 5.1a and Reaction 5.1b. Assuming the O2 order in Reaction 5.1a and 5.1b is 0.5 and 1, respectively, the rate of Reaction 5.1a and hence H 2 S conversion would increase with O2 concentration. Furthermore SO2 selectivity would increase with O2 concentration in the feed. 0 2 Inlet feed concentration (vol. %) Figure 5.5 Effect of O2 inlet concentrations on SO2 average selectivity and H2S average conversion on bulk FeS2 catalyst 5.3.2.2 Influence of H2S Concentration The O2 average reaction rate obtained on the bulk FeS2 as a function of H2S concentration in the feed is presented in Figure 5.6. In this experiment the feed composition was 0 2 2 vol . %, H2O 20 vol . %, balance N2 while varying the H 2 S at 6, 8 and 10 vol . %. The temperature of the reactor was 270°C and pressure 115 kPa. Increasing the H2S concentration increased the O2 92 Chapter 5 - Reaction Kinetics & Mechanism average conversion by enhancing the rate of Reaction 5.1a. Even though O2 is being consumed by both Reaction 5.1a and 5.1b, the amount of 0 2 consumption in Reaction 5.1b is twice that consumed in Reaction 5.1a. Therefore, increasing the H2S concentration increased the amount of 0 2 consumed i n the Reaction 5.1a but at the same time decreased 0 2 concentration available for Reaction 5.1b. 4.0E-04 3.0E-04 H2S inlet feed concentartion(vol.%) F i g u r e 5.6 Effect o f H2S inlet concentration on O2 reaction rate; feed composition was 2 vol . % O2 and 20% H2O at balance N 2 ; Temperature and pressure were 270°C and 115 kPa Hence, when the concentration of H 2 S is high (>6%) the overall O2 conversion is not changed significantly by excess H 2 S . This was concluded by comparing the rate of O2 consumption by varying 0 2 concentration (Figure 5.4) and H 2 S concentration (Figure 5.6). Figure 5.7 shows that the selectivity to SO2 decreased with increasing H 2 S concentration which indicates that the rate o f Reaction 5.1b decreased. 93 Chapter 5 - Reaction Kinetics & Mechanism * • I A S 0 2 Selectivity; • H 2 S Conversion 1 1 — 1 1 1 6 7 8 9 10 11 H2S inlet feed concentration (vol. %) F i g u r e 5.7 Effect of H2S inlet concentration on SO2 average selectivity and H2S average conversion; feed composition was 2 vol . % O2 and 20% H2O at balance N2; Temperature and pressure were 270°C and 115 kPa Similar results were obtained in another set of experiments in which the feed composition was O2 2 vol . %, H2O 20 vol . %, balance N2 while varying the H2S at 5, 6 and 7 vol . %>. The temperature of the reactor was set at 280°C. Except for the result obtained at 4 % H2S, it was also observed that increasing H2S concentration increased O2 conversion (see Figure 5.8). 94 Chapter 5 - Reaction Kinetics & Mechanism 3.E-03 O.E+00 H2S inlet feed concentration (vol. %) Figure 5.8 Effect of H 2 S inlet concentration on 0 2 reaction rate. Feed composition was 2 vol . % 0 2 and 20% H 2 0 at balance N 2 ; Temperature was 280°C and pressure was 115 kPa 80 70 2 60 | | 50 00 & 40 > v u * 30 « 8, 20 2 0) < 10 A S 0 2 Selectivity; • H 2 S Conversion A 5 6 7 H 2S inlet feed concentartion (vol. %) Figure 5.9 Effect of H 2 S inlet concentration on S 0 2 average selectivity and H 2 S average conversion; feed was 2 vol . % 0 2 and 20% H 2 0 , balance N 2 ; Temperature was 280°C and pressure was 115 kPa 95 Chapter 5 - Reaction Kinetics & Mechanism Also the average selectivity of SO2 decreased at constant average conversion of H2S (see Figure 5.9). It was concluded that overall O2 rate of reaction was determined by the sum of the rate of Reaction 5.1a which increases by increasing the H2S concentration and Reaction 5.1b which decreases by increasing the H2S concentration. 5.3.2.3 Influence of H2O Concentration A rapid and strong deactivation of the catalyst with a decrease in O2 conversion in the presence of H2O, was observed. The negative effect of H2O observed in this work can be explained by the competitive adsorption of H2O on the surface of the catalyst. The competition for the active site (bulk FeS2) between H2O and other reactants limits the adsorption of O2, H2S and S, and consequently, decreases the rate of O2 consumption. • • • • • with H 2 0 ; • without H 2 0 1 1 1 2 3 0 2 inlet feed concentration (vol.%) Figure 5.10 Effect of presence and absence of water in the feed on O2 average reaction rate versus O2 concentration 5.0E-04 1 o E. 4.0E-04 -| ,0 3.0E-04 H ra u o <S 2.0E-04-| a> ai ra 1. > 1.0E-04 < 0.0E+00 0 96 Chapter 5 - Reaction Kinetics & Mechanism Experiments were conducted to study the reaction rate in the absence and presence of H2O. The feed composition in this experiment was H2S 4 vol . %, O2 1 and 2 vol . % and H2O 0 and 20 vol . %, balance N 2 . The temperature of the reactor was 280°C. Figure 5.10 shows that at very low O2 concentration the effect of H2O is very small, but by increasing the concentration of O2, the negative effect of water is more obvious. The results showed (see Figure 5.11) that competitive adsorption between H2O and S (sulphur) is greater than between H2O and H 2 S . Similar results were reported in the work of Keller et al. (2001) on SiC-supported iron catalyst. It was observed that in the presence of H2O, SO2 selectivity decreased drastically but at the same time the conversion of H2S remained the same. Here the effect of H2O on Reaction 5.1b is more significant than on Reaction 5.1a. 90 1 80 • 70 -c 0 •g 0 60 ->0 0 50 • 08 sctivit 40 • i 30 • E a > 20 -< 10 -0 • A 0 1 2 3 0 2 concentration in the feed (vol.%) • S 0 2 Selectivity in 20% H 2 0 ; A S 0 2 Selectivity in 0% H 2 0 • H 2 S Conversion in 20% H 2 0 ; 0 H 2 S Conversion in 0% H 2 0 F i g u r e 5.11 Effect of presence and absence of water in the feed on SO2 average selectivity and H2S conversion versus O2 concentration 97 Chapter 5 - Reaction Kinetics & Mechanism Based on the obtained results it was concluded that the S (sulphur) formed in Reaction 5.1a leaves the surface of the catalyst and in the second step re-adsorbs on the catalyst during Reaction 5.1b. The significant competitive adsorption between H2O and S occurs during this step. 4.0E-04 3.0E-04 H 20 inlet feed concentarton (vol.%) Figure 5.12 Effect of H 2 0 concentration on 0 2 reaction rate in the feed contains 6 vol . % H 2 S and 2 vol . % 0 2 at balance N 2 ; Temperature was 280°C and pressure was 115 kPa In another set of experiments, the effect of H 2 0 concentration in the gas feed was studied. In this experiment the feed composition was H 2 S 6 vol . % and 0 2 2 vol . % with varying H 2 0 at 10, 20 and 30 vol . %, balance N 2 . The temperature of the reactor was 270°C and the pressure maintained at 115 kPa. The results showed that when H 2 0 concentration was higher than 10 vol . % in the feed, the decrease in the conversion was relatively small. Strong adsorption of H 2 0 on FeS 2 can saturate the surface of the catalyst and poison the 98 Chapter 5 - Reaction Kinetics & Mechanism active sites even in very small concentrations (< 5 H2O %) therefore; by increasing the H2O concentration from 10 to 30 vol . % H2O, the change in the reaction rate is not significant (see Figure 5.12). Also , again the result showed that despite the small effect on H2S conversion, the effect on SO2 selectivity in this case is very obvious (see Figure 5.13). 10 15 20 H20 concentration in the feed (vol.%) r 25 30 35 Figure 5.13 Effect of H2O concentration on SO2 average selectivity and H 2 S conversion; feed was 6 vol . % H 2 S and 2 vol . % O2 at balance N2; Temperature and pressure were 270°C and 115 kPa. 5.3.2.4 Influence of Temperature In order to study the reaction rate constants, a series of experiments was conducted at three different temperatures. These three different temperatures were set in the range that represented 99 Chapter 5 - Reaction Kinetics & Mechanism typical Claus first and second converters (see Chapter 1). Figure 5.14 shows that by increasing the temperature the average reaction rate of 0 2 increased. i f 250 260 270 280 290 300 Temperature (°C) Figure 5.14 Effect of temperature on O2 average reaction rate; feed was 6 vol . % H2S, 2 vol . % 0 2 and 20% H 2 0 , balance N 2 5.4 Reaction Kinetics O) c 1 o E. 2.0E-04 & o (0 <b <S 1.0E-04 ] o o> 2 a> > < 0.0E+00 4 240 5.4.1 Data Analysis using Process Simulator Today's process simulators present a collection of subroutines representing unit operations that can be easily assembled to a desired plant configuration. They contain component databases and property packages, can accommodate design specifications, provide sensitivity analyses, and formal optimization routines. In the recent years using process simulators to study the rate of 100 Chapter 5 - Reaction Kinetics & Mechanism reactions based on laboratory data is becoming more common. Work of Posarac et al. (2003) shows the advantages of using process simulators. The process simulator used in this study was H Y S Y S Plant™ 3.0.1 of Hyprotech, Ltd because of its relatively accurate calculation for packed bed reactor with complex kinetics, calculation of component properties using advanced property packages, as well as detailed handling of heat transfer. Using the process simulator had the advantage of being able to examine a non-isothermal reactor. In the course of the experiments, the temperature recorded by the second independent thermocouple (see Chapter 3) was applied to the reactor in the simulation. The key assumptions in the development of the simulation and analysis of data were: 1. The gas flow velocity, concentration and temperature distribution along the bed radius was uniform. 2. The processes of mass transfer inside the particles were quasi-steady-state. 3. Temperature gradients inside the catalyst particle were absent. 4. Decreases in reaction rate due to the blockage of active sites by condensed sulphur and poisoning of the support site by sulphation, were absent. 5. Reaction was irreversible. The validity of the plug flow assumption was previously discussed in Chapter 3. Also , since the apparatus operates at high temperatures and close to atmospheric pressure (—115 kPa) the assumption of ideal gas behaviour was also valid with compressibility factors greater than 0.99. Also , it was assumed that the stoichiometric order of the reactants in the reaction was valid and the results were used to find the values for reaction constants for each set o f experimental data. A plug flow reactor was defined in the simulator. Dimensions of the reactor were defined based on 101 Chapter 5 - Reaction Kinetics & Mechanism the volume of the catalyst bed in the experimental work. The rate of Reaction 5.1a and 5.1b were presented in the simulator as two power law expressions. -El - rD2 = kme RT C0o52CH2S Rate of 0 2 consumption based on Reaction 5.1a -El - rQ2 = k02e RT CQ2CS% Rate of O2 consumption based on Reaction 5.1b For this analysis three sets of data, which were discussed in Section 5.3.1 were used. In the analysis, Set 1 refers to a series of experimental data in which the O2 concentration was varied (Exp. E01, E02, E03 and E04), Set 2 refers to the experimental data in which the H2S was varied (Exp. E05, E06 and E07) and Set 3 refers to experimental data in which the temperature of the feed was varied (Exp. E l 1, E l 2 and E l 3 ) . 5.4.2 Optimization Method The H Y S Y S optimizer and spreadsheet were utilized to find the optimum values of k o i , ko2> E i and E2 by minimizing an objective function. The objective function was based on differences between the experimental and simulated results. The simulated results were calculated using an integral packed bed reactor, defined in the simulator, and an iteration subroutine carried out by the optimizer. Flowrates of each reactant, temperature, pressure and the volume of the bed were defined for each set of experiments. The objective function chosen was the sum of the square o f differences between the experimental and simulated average O2 and H2S conversions, as wel l as the difference between the experimental and simulated average SO2 selectivity, without any weighting factors due to, similarity of the magnitude of the conversion and selectivity values. The objective function was defined as follows: 102 Chapter 5 - Reaction Kinetics & Mechanism 0 F = E(*PJ-*».;)2 Equation 5.1 where xp} was the predicted average conversions ( 0 2 and H2S) and selectivity (SO2), xm 7 was the experimental (measured) gas average conversion (O2 and H2S) and selectivity (SO2). 5.4.3 Results Table 5.2 shows the reaction rate constants optimized by the simulator for each set of experimental data including objective function (OF) values. Experiment EifkJ/mol) ko2 E2(kJ/mol) OF Set1 4.32 x 108 66.77 3.78 x 10 1 0 58.23 1743.5 Set 2 5.48* 107 44.30 6.84 x 107 39.81 2005 Set 3 5.96 x 107 67.55 4.54 x 109 63.54 282.9 Table 5.2 Reaction rate constants for each set of experimental data using H Y S Y S optimizer Figure 5.15, 5.16 and 5.17 show the experimental and simulated data. In order to validate the results and compare the model prediction and experimental data, statistical analysis was conducted on the results. Table 5.3 shows the summary of the analysis (see Appendix E) . Experiment Std. Error R2 F Fcritical (a = 0.05) Set1 13.83 0.81 38.11 2.55 Set 2 16.93 0.63 14.52 2.55 Set 3 5.01 0.89 70.65 2.55 Table 5.3 Summary o f the statistical analysis conducted on the results in the range of experiments 103 Chapter 5 - Reaction Kinetics & Mechanism In Table 5.3 parameters which indicate the goodness o f the model fitted to experimental 9 9 9 data are R'-and F. The R* measure ranges from 0 to 1 and as a model more ideal, the R values approaches 1 (0 represents a complete lack of fit), the standard error decreases toward zero, and the F value tends toward infinity. If the value of F is lower than the value of Fcrj ticai, the model is not valid. From the results of Table 5.3 it was concluded that the calculation based on the experiments of Set 3 produced the most valid rate constants. Results shown in Table 5.2 for the 0 2 reaction rate constants was constructed by assuming stoichiometric order of reactants in Reaction 5.1a and 5.1b. Optimization of the order of reactants hasn't been considered in the present studies, due to limited experimental data. 0 1 2 3 4 5 0 2 inlet feed concentartlon (vol.%) • O2 Exp. Conv.;AH.2S Exp. Conv.;* SO2 Exp. Selec. . . . . O 2 S im. Conv.; H2S Sim. Conv.; SO2 Sim. Selec. Figure 5.15 Experimental and simulated data; Effect o f O2 feed inlet concentration on O2 and H2S conversion as well as SO2 selectivity 104 Chapter 5 - Reaction Kinetics & Mechanism HjS inlet feed concentration (vol.%) • 0 2 Exp. C o n v . ; A H 2 S Exp. Conv.;« S 0 2 Exp. Selec. . . . . 0 2 Sim. Conv.; H 2 S Sim. Conv.; S 0 2 Sim. Selec. Figure 5.16 Experimental and simulated data; Effect of H 2 S feed inlet concentration on 0 2 and H 2 S conversion as well as S 0 2 selectivity 270 Reactor tempreture (°C) 0 2 Exp. C o n v . ; A H 2 S Exp. Conv.;* S 0 2 Exp. Selec. . 0 2 Sim. Conv.; H 2 S Sim. Conv.; S 0 2 Sim. Selec. Figure 5.17 Experimental and simulated data; Effect feed temperature on 0 2 and H 2 S conversion as well as S 0 2 selectivity 1 0 5 Chapter 5 - Reaction Kinetics & Mechanism 5.5 Mechanism of the Reaction Knowing the form of the rate law from the previous section, here, a mechanism was postulated and the limiting steps that were consistent with the rate data and the assumptions made. 5.5.1 Adsorption Chemisorption of the reactants is a necessary part of a heterogeneous catalytic reaction. In the proposed mechanism the oxygen adsorbs on the FeS2 dissociatively. H2S and H2O adsorb on the surface of FeS2 molecularly. In this section, SI w i l l represent an active site; alone it denotes a vacant site. The combination of SI with another molecule or atom implies that the species is adsorbed on the site S I . The total molar concentration of active sites per unit mass of catalyst is labelled as Ct and the molar concentration of vacant sites is presented as CV- Hence: 02+2S\*+20.Sl Equation 5.2 H2S + SI o H2S.S\ Equation 5.3 H20 + SI <-> H2O.S\ Equation 5.4 Each adsorption step is considered as an elementary reaction and species concentration in the gas phase are replaced by their respective partial pressure. The rate expression for each species as given in Equation 5.2, 5.3 and 5.4 is: -r02= k5P"Cv - k_5Cosl Equation 5.5 -rms= k6PH2SCv - k.6CH2SSl Equation 5.6 106 Chapter 5 - Reaction Kinetics & Mechanism RHIO~ ^I^HIO^V k_nCH20S] Equation 5.7 A t equilibrium, the net rate of adsorption equals zero. The right hand side of the rate of adsorption for Equation 5.5, 5.6 and 5.7 was set equal to zero and the concentration of each species adsorbed on the surface of FeS2 was calculated as: Cosx = K5P°25CV Equation 5.8 Cms -K6PH2SCV Equation 5.9 CH2O =K1PH2C>CV Equation 5.10 where the surface reaction equilibrium constant KA was defined as: K-A 5.5.2 Surface Reactions In order to write the surface reactions, a sequence of reaction steps was assumed. Also , an assumption was made that each individual step, as shown in the preceding steps, was reversible, except for the postulated limiting steps. Finally, non rate-limiting steps were used to eliminate all coverage-dependent terms. The sequence of the surface Reaction 5.1a and 5.1b are presented as follows. Limit ing steps are labelled as RDS1 and RDS2. H 2 S . S ; + 0.57 —» H S . S i + HO.S1 HS.S7 + HO.S;<-> S.S1 + H 2 0 . 5 7 S.S1 <r+S + .Sl (RDS1) Equation 5.11 Equation 5.12 Equation 5.13 107 Chapter 5 - Reaction Kinetics & Mechanism S.S1 + O.Sl<~* SO.S7 + .SI so. si + o.si—> S02.S7 + SI so2.si <-» S0 2 + .SI H2O.S7 H 20 + .57 (RDS2) Equation 5.14 Equation 5.15 Equation 5.16 Equation 5.17 Equation 5.11 and 5.15 were considered the rate limiting steps in the overall reaction sequence and the rate of each step is: Equation 5.18 rRDS\ ~ KxlCH2S£]C0£\ rRDS2 ~ K]sCsosxCosx Equation 5.19 With RDS1 and RDS2 as limiting steps, specific reaction rate ks o f the other no-limiting steps is large by comparison and hence it was assumed, that, ^ 0 Equation 5.20 Solving Equation 5.13 f o r C s.s\ P C S.SI ~ Equation 5.21 B y using Equation 5.21 and 5.8, solving Equation 5.14 for C so.si • ^SO.Sl ~ v r S r 0 2 ^ V Equation 5.22 by substituting Equation 5.8, 5.9, 5.21 and 5.22 in Equation 5.18and 5.19 and solving: Equation 5.23 r — Y Y V P p0 - 5 / ^ 2 rRDSl ~ ' / v l l ' I Y 5 - ' v 6 - r W 2 S - r 0 2 V ' RDS2 KnKl4K5K5 3 r s r 0 2 ^ V K, Equation 5.24 13 The concentration of vacant sites, Cy, was eliminated from Equation 5.23 and 5.24 by utilizing the site balance to give the total concentration of sites Ct which is assumed constant. total sites = vacant sites + occupied sites 108 Chapter 5 - Reaction Kinetics & Mechanism Since O2, H2S, H2O and intermediate species are adsorbed on the surface, the total concentration of sites is: C, =CV+C02 S l + CH2S S 1 + CH2Q S l + intermediates Equation 5.25 by assuming that the concentration of the intermediate species was negligible and also assuming a relatively high concentration of H2O on the sites compared to the low concentration of H 2 S and O2, the total concentration of sites was simplified to : C, =Cy+CH20Sl Equation 5.26 by substituting Equation 5.26 into Equation 5.23 and 5.24 : = KnK5K6PH2SP£C? Equation 5.27 rmsi = KuKnKlPsPo2^ Equation 5.28 Kn(\ + K1PH2Q) The proposed mechanism and obtained equations for RDS1 and RDS1 were consistent with the assumed Reaction 5.1a and 5.1b, respectively. The rate of RDS1 is 0.5 order in respect of the O2 partial pressure, which was also assumed in Reaction 5.1a. Also , the rate of R D S 2 is first order in respect of the O2 partial pressure which was assumed in Reaction 5.1b. The order of H 2 S and S partial pressure also were consistent with the assumed orders o f Reaction 5.1a and 5.1b, respectively. Equation 5.27 and 5.28 show that the effect of H 2 0 on the RDS1 is higher than R D S 2 . This conclusion was consistent with the observed results, which showed that the effect of H2O on SO2 selectivity was higher than H 2 S conversion. It was observed (see Section 5.3.2.3) that H2O decreases the reaction rate. This phenomenon is more obvious at low concentration of H2O and when the concentration of H2O exceeds 5 vol . % the effect remains constant. 109 Chapter 5 - Reaction Kinetics & Mechanism In the Claus process the H2O concentration at first and second catalytic converters is about 20 to 30 vol . %. Therefore, the effect of H2O concentration change has a negligible effect on the overall reaction rate and the Equation 5.27 and 5.28 can be simplified by assuming that the denominator is constant and independent of H2O partial pressure. = K"K>K'*»™Po>C< E q u a t i o n 5 2 9 K21 K,, K, -, K 2 Pc Pn-> Cf rRDS2 = M 1 2 5 s 0 2 ' Equation 5.30 K2& Equation 5.29 and 5.30 can be simplified further, hence: rRDs\ = K29PH2spm o r Equation5.31 rRos\ = K^HisC-m. Equation 5.32 where Ki2 = K29 x ( R T ) x ( R T ) 0.5 rRDS2 = KwPsPo2 o r Equation 5.33 rRosi = K3*CsCo2 Equation 5.34 where Ku = K3Q x ( R T ) x ( R T ) The K32 and K34 were evaluated in Section 5.4.3 using H Y S Y S optimizer for three separate sets of experiments as: KJ2 = kme RT Equation 5.35 zh. Ki4 = k02e RT Equation 5.36 110 Chapter 5 - Reaction Kinetics & Mechanism 5.6 Summary The work in Chapter 5 can be summarized as: 1- Experiments were conducted on bulk FeS2 in order to study the reaction kinetics. 2 - Increasing the O2 concentration increases the rate of reaction (O2 consumption) on FeS2. 3- Increasing the H2S concentration increases the rate of reaction (O2 consumption) on FeS2. 4- Increasing the H 2 0 concentration decreases the rate of reaction (O2 consumption) on FeS2. H 2 0 changes the reaction rate significantly at low concentration of H2O but this effect on the rate of O2 consumption remains approximately unchanged when the H2O saturates all the active sites. 5- A t low concentration of the reactants (O2 < 5 vol . % and H2S < 10 vol . %) the reaction kinetic on FeS2 was expressed as two power law equations corresponding to Reaction 5.1a and 5.1b as: -E\ _ r —lr p RT C0$C '02 ~ ^OX* 02^HIS -E2 '02 ~ n ' 0 2 e ^02^S& 6- The values for rate contents were estimated using the H Y S Y S simulator and validity of the values were investigated by statistical analysis. I l l Chapter 6 - Conclusions and Recommendations Chapter 6 - Conclusions and Recommendations 6.1 Conclusions Fe-based catalyst supported on AI2O3 was considered as a guard catalyst in Claus converters to prevent the deactivation of AI2O3 catalyst due to sulphation. In the present study, Fe-based catalysts were prepared at different Fe-loadings and activity of the catalysts was studied for O2 removal. The following conclusions can be drawn from the present study: 1. FeS2/ AI2O3 catalyst is more active than Fe203/Al203 catalyst for O2 removal in Claus conditions. Experiments showed that O2 consumption occurred on the Fe203 sites based on the reaction (a) and on FeS2 sites based on the sequential reaction (a) and (b). (a) H2S + ^02-^-Sn+H20 2 n (b) S + 02->S02 2. It was suggested that during the reaction, mainly 0 2 consumption occurred on FeS2 sites and Claus reaction on AI2O3 sites. Therefore, FeS2/ AI2O3 can be considered a bifunctional catalyst. 3. Reaction kinetics was studied on bulk FeS2 and a reaction rate expression was constructed based on a stoichiometric assumption for order of reactants. Statistical analysis showed a good agreement for the constructed model with koi= 5.96 x 10 7 , E i= 67.55 (kJ/mol) and ko 2 = 4.54 x 10 9 and E 2 = 63.54 (kJ/mol) and the experimental results. - E l „ - U pRTfO.Sf-> '01 ~ ^0\^ y-'02y~'H2S -E2 _ r - k P RT C C '02 ~ A ' 0 2 e ^02^S& 112 Chapter 6 - Conclusions and Recommendations The suggested values are valid only for O2 consumption on the bulk FeS2 and under the experimental conditions used in the present work. 4. A Mechanism of the reaction and the rate limiting steps were postulated that were consistent with the rate data. 6.2 Recommendations Experimental work in the present study encountered some difficulties due to the nature of the reaction. Heterogeneous catalytic reactions with a multiphase effluent sometimes made the experimental work difficult to conduct. The following are some recommendations for improved experimental work. 1. One of the most important problems, in the course of experiments, was to avoid the condensation of effluent sulphur in the down flow of the reactor. Consequently, the effluent line from the reactor to the sulphur trap was heated using a heating tape. Also , the sulphur trap was installed at a minimum distance from the reactor in order to minimize the possibility of sulphur condensation. It is recommended that a furnace (heating box) is installed at the end of the reactor and that the sulphur trap and effluent lines, including the valves and connections, are installed inside this furnace. 2. The Lindberg tube-furnace used in this work or any other similar tube furnaces have difficulty in maintaining a uniform temperature along the reactor. The problem arises because of the natural draft of air flow from the bottom of the reactor up along the furnace tube. The air draft can cause significant heat transfer disguises along the reactor. It is recommended that a proper isolation method be used at the top and bottom of the furnace to prevent this problem. 113 Chapter 6 - Conclusions and Recommendations 3. In this work, conversion of O2 was calculated using an indirect method based on the concentration of SO2 and H2O produced (see Appendix D - Section D.2). Interference of the O2 peak with an unknown peak (see Appendix A - Section A.3) and also, inconsistency o f the O2 peak from the effluent were two important reasons to use indirect methods for O2 measurement. It is recommended to investigate this problem and use a direct measurement of the O2 concentration in future work. 6.3 Future Work In order to further investigate the preconditioning of the catalyst and the rate of reaction on the supported catalyst as well as stability, additional studies are suggested. 1. Preconditioning method used in the present study was based on the work of Inamura et al. (1993). It is suggested to conduct T P R and TPS experiments on the catalysts with different Fe loadings to investigate the pattern of H2 and H2S consumption at different temperatures. 2. In this work Fe phase change was studied using D R I F T experiments and C O as probe gas. It is recommended in future work to study the use of other probe gases such as N O . 3. Conditions for the rate of reaction study on the pure FeS2 in the present work (Table 5.1) were prescribed initially in order to qualitatively investigate the effect o f each reactant on the overall reaction rate. Therefore it is suggested to design a set of experiments with longer period of time and larger concentration range for each reactant. 4. The modeling of the kinetic expressions was conducted based on the assumption that the stoichiometric order of the reactants in the reaction was valid and the results were used to find the values for reaction constants for each set of experimental data. The result was 114 Chapter 6 - Conclusions and Recommendations that the model had good agreement in the low concentration range but it had relatively poor validation at the high concentration. Therefore, it is recommended that the order of the reactants are verified by further experiments and modeling using Langmuir-Hinshelwood approach over a wider range of process conditions to confirm the model validity. 5. Construct a method to combine the rate of reaction (Claus reaction) on the AI2O3 support with the rate of O2 consumption on pure FeS2 proposed in Chapter 5, in order to investigate the rate of the reaction on the supported FeS2 catalyst. The rate of the Claus reaction on the AI2O3 support is investigated in the work of Birkholz et al. (1987) by accounting for the reversibility of the reaction. This w i l l allow a study of bifunctional catalysts and the influence o f the interaction between support and active phase on the overall reaction rate. 115 Nomenclature Nomenclature Particle diameter L Ration of length of reactor to particle diameter dp Net negative charge • Positive charge C Gas concentration d Catalyst pore diameter Ei, E2 Activation Energy F Feed mass flowrate K Reaction equilibrium rate constant (Kj;j is Equation number) K: Constant K" Absorption coefficient ki, k2 Reaction rate constant koi, ko2 Preexponential factor or frequency factor L Dimension of the particle M Metal atom n Number of C O in polymeric structure n* A n integer (usually n* =1) in Bragg's L a w Equation Os Surface oxide P Pore vapor pressure P(02) Partial pressure of the O2 in the feed P(S02) Partial pressure of the S 0 2 in the feed Po Surface vapour pressure Nomenclature R Roo RE S Sl T u V V V"D VO.I V , VQ W x Greek Letters P Y 8-8+ e* e x Universal gas constant Reflectivity of a sample of infinite thickness Rare earth element Scattering coefficient Active site Absolute temperature Gas velocity Vacancy Liquid molar volume Tetrahedral A l vacancy Pore volume smaller than 0.1 pm Pore volume smaller than 1 um Oxygen vacancy Catalyst weight Neutral defect Extent of broadening Liqu id surface tension Negative charge Positive charge Contact angle Diffraction or Scattering angle X-ray wavelength 117 Nomenclature Gas viscosity V frequency v (S=0) Electron vibration when S = 0 v (S-O) Electron vibration when S-0 P Gas density Abbreviations B E T Total Surface Measurement C N V Catalytic Converter in Claus process C O D Sulphur Condenser in Claus process DRIFT or D R I F T S Fourier Transmission Infrared Spectroscopy FT-IR Fourier Transmission Infrared Spectroscopy G C Gas Chromatograph G H S V Gas Hourly Space Velocity ID Inside Diameter L H Langmuir-Hinshelwood approach L i n (Counts) Intensity counted on linear scale OF Objective Function R D S Rate Determining Step S E M Scanning Electron Microscopy T C D Thermal Conductivity Detector X R D X-ray Diffraction Spectroscopy References References: 1. Aloco Industrial Chemicals, 16010 Baker's Point Lane, Suite 250, Houston, Texas, U S A , 2003. 2. Amberg, C . H . ; Seanor, D . A . , 'Proceedings o f the Third International Congress on Catalysis', North -Hol land , Amsterdam, p.450,1965. 3. Arnoldy, P.; Heijkant, V . D . A . M . ; Bok, D . G . D . ; Moul i jn , J. A . , 'Temperature-Programmed Sulfiding M o 0 3 / A l 2 0 3 Catalysts', Journal o f Catalysis 92,35-55(1985). 4. Axens North America Inc., 1800 St. James Place, Suite 500, Houston, Texas, U S A , 2003. 5. Berben, P.; Geus, J. ; Brink, V a n De P.; Terorde, R., Sulphur Seminar, Amesterdam, The Netherlands,1999. 6. Beveridge, B., 'Sulphur Recovery Optimization of Claus/Sulfureen Units ' , Canterra Energy L T D , technical Report, Calgary, Alberta, Canada, 2002. 7. Birkholz, K . O. R.; Behie, L . A . ; Dalla Lana, I. G . ; Can, J. , Chem. Eng., 65(1987) 778-784. 8. Cahi l l , C . L . ; Benning, L . G . ; Barnes, H . L . ; Parise, J. B . , 'In situ time-resolved X-ray diffraction o f iron sulfides during hydrothermal pyrite growth', Chemical Geology 167(2000) 53-63. 9. Capone, M . , in: Kroschwitz J .K. , Howe-Grant M . (Eds.), Encyclopaedia o f Chemical Technology, V o l . 23, Wiley, New York, 1997, p. 432. 10. Clark, P .D. ; Dowling, N. I . ; Huang, M . ; Okemona, O.; Butl in, G .D . ; Hou, R.; Kijlstra, W . S.; ' Studies o f sulfate formation during the conversion of H 2 S and S 0 2 to sulfur over activated alumina', Applied Catalysis A : General 235 (2002) 61-69. 11. Coluccia, S.; Garrone, E . ; Guglielminotti, E . ; Zecchina, A . , Journal o f Chemical S o c , Farad, Trans, 1, 77(1981) 1063. 119 References 12. Davydov, A . A . , 'Infrared Spectroscopy of Adsorbed Species on the Surface of Transition Metal Oxides' , John Wiley and Sons Ltd, 1990. 13. Davydov, A . A . ; Shepot'ko, M . L . , Kinetics and Catalysis, Vol .44, No . 2, 2003, pp 260-267. 14. Deo, G . ; Wachs, I. E . , Journal of Physical Chemistry, 95(1991)5889. 15. Duncan, M . P . , in: Kroschwitz J .K. , Howe-Grant M . (Eds.), Encyclopaedia of Chemical Technology, V o l . 23, Wiley, N e w York, 1997, p. 276. 16. Eischens, R. P.; Pl iskin W . A . , Advanced Catalysis, 9 662(1956). 17. Fenderson, S., 'Improving Claus Sulphur Recovery Unit Reliability through Engineering Design', presented at Brimstone Engineering Sulphur Recovery Symposium, September 15-18,1998, Texas, U S A . 18. Fenelonov, V . B . ; Neimark, A . V . ; Kheifets, L . I . ; Samakhov, A . A . , in: Delmon B . , Grange P., Jacobs P . A , Poncelet G . (Eds.), 'Preparation of Catalysts II ' , Elsevier, Amsterdam, 1979, p. 232. 19. Filatova, O. Ye . ; Kislenko, N . N . ; Morgun, L . V . ; Makhoshvil i , Y u . A . , Sulphur 2002, 26-31 October 2002, Austria, Vienna. 20. Fogler, H . Scott, 'Elemental of Chemical Reaction Engineering', Third Edition, Prentice Hal l PTR, 2000. 21. George, Z . M . , 'Poisoning and Regeneration of Claus Alumina Catalysts', presented at The 5 t h Canadian Symposium on Catalysis, Calgary, Alberta, October 26-27,1977. 22. Geus, J.; Berben, P.; Brink, Van De P.; Terorde, R., Sulphur Seminar, Amesterdam, The Netherlands,1999. 120 References 23. Geus, J .W., in: V . Cortes Corberan, S. V i c . Bel lon (Eds.), Proceedings o f the 2nd Word Congress and 4th European Workshop Meeting on N e w Developments in Selective Oxidation, 1993, Paper No . P54. 24. Goar, B . , 'Sulphur Recovery Technology', Energy Progress, Vol .6 (2):71-75, June 1986. 25. Goodboy, K . P.; Downing J. C ; Fleming, H . L . , O i l & Gas Journal, 89,Nov. 4,1985. 26. Inamura, K . ; Iwamoto, R.; lino, A . ; Takyu, T., 'Reduction and Sulfidation Properties of Iron Species in Iron-Supported Y-Zeolite by Temperature-Programmed Reduction and Sulfiding', Journal of Catalysis 142, 274-288(1993). 27. Johnson, A . ; Edwards, T. R.; Mil ler , M . F., 'Techniques curb catalyst deactivation at Sulphur plant', O i l & Gas J., Oct 26 (1987)33. 28. Keller N . ; Pham-Huu, C ; Ledoux, J. M . , 'Continues process for selective oxidation of H2S over SiC-supported iron catalysts into elemental sulfur above its dewpoint' , Applied Catalysis A : General 217(2001) 205-217. 29. Keller, N . ; Pham-Huu, C ; Ledoux, J. M . , 'Continuous process for selective oxidation of H 2 S over SiC-supported iron catalyst into elemental sulfur above its dewpoint', Applied Catalysis A : General 217 (2001) 205-217. 30. Kerr R., Paskall H . G . , Ballash N . , Energ. Proc. Can. 46(1977) 40-51. 31. Kotter, M . ; Riekert, L . , in: Delmon, B . ; Grange, P.; Delmon, B . (Eds.), 'Preparation o f Catalysts IT, Elsevier, Amsterdam, 1979, p. 51. 32. Kroger, F. A . , in 'The Chemistry of Imperfect Crystals' 2 n d edition, North- Holland, Amestrdam, 1974. 33. Lagas, J .A. ; Nasato, E . ; MacDougall , R.S. , 'Benefits o f Retrofitting with Second Generation Superclaus Catalyst', presented in Canadian Gas Processors Association, 2 n d Quarterly Meeting, Calgary, Alberta, 1993. 121 References 34. Laperdrix, E . ; Costentin, G . ; Guyen, N . N . ; Saur, O.; Lavalley, J. C , Journal of Catalysis 187,385-391(1999). 35. Laperdrix, E . ; Sahibed-dine, A . ; Costentin, G . ; Saur, O.; Bensitel, M . ; Nedez, C ; Mohammed Saad, A . B . ; Lavalley, J.C., 'Reduction of sulfate species by H 2 S on different metal oxides and promoted aluminas', Applied Cat. B : Environmental. 26(2000) 71-80. 36. Lavalley, J .C., 'Infrared spectrometric studies of the surface basicity of metal oxides and zeolites using adsorbed probe molecules', Catalysis Today 27(1996)377-401. 37. Lavalley, J .C.; Bachelier, J.; Aboulayt, A . ; Legendre, O.; Luck, F. , 'Act ivi ty o f different metal oxides towards C O S hydrolysis. Effect of S 0 2 and sulphation', Catalysis Today, 17 (1993) 55-62. 38. Lavalley, J .C.; Laperdix, E . ; Sahibed-dine A . , G . ; Costentin Saur, O.; Benstiel, M . ; Nedez, C. ; , Mohamed Saad, A . B . , 'Reduction of sulphate species by H 2 S on different metal oxides and promoted aluminas', Applied Catalysis B : Environmental 26 (2000) 71-80. 39. Lavalley, J .C.; Pieplu, A . ; Saur, O.; Pijolet, M . ; Legendre, O., ' A Kinetic Model for Alumina Sulphation', Journal of. Catalysis. 159, 394-400 (1996). 40. L i , K . - T . ; Ch i , Z . - H . , 'Selective oxidation of hydrogen sulfide on rare earth orthovanadates and magnesium vanadates', Appl ied Catalysis A : General 206 (2001) 197-203. 41. L i , K . - T . ; Hung Y . - C , 'Hydrogenation of sulfur dioxide to hydrogen sulfide over Fe/y-A 1 2 0 3 catalysts', Applied Catalysis B : Environmental (2003) 13-20. 42. Lurgi Oel , Sulfur Recovery, technical bulletin, 2003. 43. McHugh, T.; Luinstra, E . ; Clark, P., Alberta Sulphur Research Ltd, Sulphur 98,1998. 122 References 44. Mora, R . L . , 'Sulphur condensation influence in Claus catalyst performance', Journal of Hazardous Materials, A 7 9 (2000) 103-115. 45. Morris, R. J.; Bery, de C , 'The world sulphur situation and the emerging role o f Latin America ' , presented at the 1999 IF A Production and International Trade Conference, Buenous Aires, Argentina, October 13-14,1999. 46. Nakamoto, K . , 'Infrared and Raman Spectra of Inorganic and Coordination Compounds', 4 t h edition, Wiley, N e w York, 1977. 47. Nedez, C ; Ray, J .L. , 'Understanding Claus Catalyst Deactivation Mechanism: Optimization of Alumina Using Physisco-chemical Parameters', Elseiver Science B . V . , and Catalyst Deactivation 1997, 251-258. 48. Nedez, C ; Ray, J -L. , Catalysis Today, 27(1996) 49-53. 49. Niemantsverdriet, J. W. , Spectroscopy in Catalysis, V C H , (1995) 50. Pearson, M . J . , 'Relationship between Claus activity and properties of aged alumina catalysis', presented at The 5 t h Canadian Symposium on Catalysis, Calgary, Alberta, October 26-27,1977. 51. Phipps, Jr. A . J . , , 'Changes in world energy markets and the outlook on recovered sulphur', presented at the sulphur institute's 14 t h Phosphate-Sulphur Symposium, January 24-26,1993, Boca Raton, Florida. 52. Pieplu, A . , Ph.D. Thesis, Caen, France, 1994. 53. Pieplu, A . ; Janin, A . ; Saur, O.; Lavalley, J.C., in :Centi G . , et. al. (Eds.), Processing's o f the 1st World Congress on Environmental Catalysis, 1995, p. 303. 54. Pinna, F., 'Supported Metal Catalysts Preparation', Catalysis Today, 41(1998) 129-137. 123 References 55. Posarac, D . ; Hudgins, R. R.; Jaree, A . , 'Analysis of Steady-State Temperature Profiles from a Laboratory Reactor by Means of a Process Simulator', The Canadian Journal of Chemical Engineering, 81(2003)103-109. 56. Schoofs, G . R., 'Sulfur Condensation in Claus Catalyst', A L C O A Technical Bulletin, Feb. 1985. 57. Schoofs, G . R., 'Sulphur Condensation in Claus Catalyst', Hydrocarbon Process, (1985) 71-73, February. 58. Taraj, P.; Taraj, J., 'Preparation, characterization and sintering behaviour of spherical iron oxide doped alumina particles', Acta Materialia 50(2002) 5-12. 59. Tellier, J., United States Patent 4,192,857. March 11,1980. 60. Terorde, R . J . A . M . ; Jong de., M . ; Crombag, M . J . D . ; V a n den, B . ; Di l len van, A . J . ; Geus, J.W., Catalysis Today 17(1993)217. 61. Thomson, J. W. ; Jung, H . , 'Reduction/ Oxidation of a High Loading Iron Oxide Catalyst', Journal of Catalysis 128, 218 - 230 (1991). 62. U .S . Environmental Protection Agency, Background Report, AP-42 Section 5.18 sulphur recovery, 2003. 63. United States Patent 4,192,857, Tellier et al, March 11,1980. 64. U S E P A http://www.epa.gov/region04/oeapages/intergov/sg060499.htm, 2003. 65. Vaishnava, P. P.; Ktorides, P. I.; Montano P. A . ; Mbadcam. K . J.; Melson, G . A . , Journal of Catalysis 301(1985). 66. Wachs, E . I.; 'Raman and IR studies of surface metal oxide species on oxide supports: Supported metal oxide catalysts', Catalysis Today 27 (1996) 437-455. 124 Appendix A APPENDIX A - SIMADZU GC-14A Operation A. 1 Carrier Gas Flow Three different kinds o f gases can be used as carrier gas, He, N 2 or H 2 . For the present study He (UHP grade; Praxair) was used as carrier gas for its high resolving power and less time needed for analysis. A s the T C D is a density detector, the carrier gas flowing into the T C D should be maintained constant and to obtain separation on the column in high efficiency, selection of proper flow rate of the carrier gas is essential. For He the flow rate in the range 20 -40 cm/sec is proper in terms of linear velocity or for the column used in the present setup, the flow arte of 25 - 50 ml/min. The flow was measured directly at the column outlet using a soap film flow meter. For more information see Gas Chromatograph GC 14A Instruction Manual, SHIMADZU, 11-8, 1989. A.2 TCD Temperature and Current T C D temperature should set 20 to 50°C higher than the column oven temperature (final temperature in temperature elevation). In the setup in present study the final and the T C D temperature was set to 120°C and 180°C, respectively. Current flowing in the T C D filaments was set from the keyboard and it was fixed at D E T No.4. Current can be set from 0mA to 200 m A , the upper limit of current depends on the carrier 125 Appendix A gas type and T C D thermostatic oven temperature. For T C D thermostatic temperature of 180 and He as carrier gas the T C D current was set to 150 m A . For more information see Gas Chromatograph GC 14A Instruction Manual, SHIMADZU, 11-9, 1989. A. 3 Temperature Program Heating of the column and oven were executed by a desired multi-step heating program. The heating program was used mainly for two reasons. During the experiment some H2O molecules escaped the P2O5 drier and entered to G C columns. In order to flash out the H2O, oven and column was heated up to 120°C. H 2 S N 2 f S 0 2 35 °C 3 min H 2 0 A-8 min 120 °C 6 min Scheme A . l Location of peaks and temperature program in the present study 126 Appendix A Second, to be able to take the sample each 30 min the column was heated up to speed up the appearance of SO2 peak. A.4 Time Program A time program consists of a set of programmed parameters to be changed after the program start. Program allowed maintaining an acceptable base line for each peak and eliminates the undesired peaks. T I M E P R O G R A M 0 . 0 5 L . O N : B . On 1 B . O F F 2 . 9 L . O F F 3 . 5 5 L . ON 3 . 6 A T T E N (0 ) = 8 3 . 7 L . O F F 6 L . ON 7 A T T E N (0 ) = 4 9 . 7 L . O F F 1 0 . 7 L . ON 1 3 . 5 L . O F F 1 4 . 5 L . ON For more information see CR501 Chromatopac Instruction, SHIMADZU, 16-1, 1989. A. 5 Peak Processing Parameters The entire peak processing parameters, including peak detection, correction of the baseline drift, separation of unresolved peak, and peak area measurement, was carried out according to peak processing parameters. The M E T H O D 41 was used instead o f the default M E T H O D 21 in order to observe the base line for each peak and made proper corrections using time program. 127 Appendix A A N A L Y S I S P A R A M E T E R F I L E W I D T H 5 S L O P E 7 0 D R I F T 0 M I N . A R E A 1 0 0 0 0 0 T . D B L 0 S T O P . T M 6 5 5 A T T E N 4 S P E E D 1 0 M E T H O D $ 4 1 F O R M A T $ 0 S P L . W T 1 0 0 I S . W T 1 For more in format ion see CR501 Chromatopac Instruction, SHIMADZU, 7-1, 1989. 128 Appendix B APPENDIX B - Summary of Porocel Fe-based Catalyst Performance Test Catalyst Type Porocel Fe-1 Date of Exp. April 2/2002 Catalyst Weight(g) 1.46 Reactor Temperature(°C) 250 Feed Calibration Time Peak Area (min) o2 N 2 H 2 S so2 30 10877 2916862 222770 195475 60 11437 2917806 221428 191522 90 11591 2922447 222470 192764 Average 11302 2919038 222223 193254 Time Actual Area Normalized Area Pressure (min) o2 N 2 H 2S S 0 2 o2 H 2S S 0 2 kPa 15 16619 4824626 91877 119453 10055 55588 72273 225 45 16236 4809666 171888 177891 9854 104321 107964 225 75 15636 4661911 192464 189744 9790 120511 118808 225 105 14222 4462438 199007 190613 9303 130178 124687 273 Time Conversion % (min) o2 H 2S S 0 2 15 11 75 63 45 13 53 44 75 13 46 39 105 18 41 35 129 Appendix B Catalyst Type Porocel Fe-2 Date of Exp. April 5/2002 Catalyst Weight(g) 1.48 Reactor Temperature(°C) 250 Feed Calibration Time Peak Area (min) o2 N 2 H 2 S so2 30 18432 4414728 342110 303628 60 18049 4404479 333040 297105 90 Average 18241 4409604 337575 300367 Time Actual Area Normalized Area Pressure (min) o2 N 2 H 2S S 0 2 o2 H 2S so2 kPa 15 15764 4405838 95940 139563 15777 96022 139682 232 45 16574 4402731 147477 156110 16600 147707 156354 232 75 16237 4325101 165851 163837 16554 169091 167038 232 105 16571 4430515 170833 169887 16493 170027 169085 232 Time Conversion % (min) o2 H 2S S 0 2 15 14 72 53 45 9 56 48 75 9 50 44 105 10 50 44 130 Appendix B Catalyst Type Porocel Fe-3 Date of Exp. March 21/2002 Catalyst Weight(g) 1.50 Reactor Temperature(°C) 250 Feed Calibration Time Peak Area (min) o2 N2 H2S so 2 30 10877 2916862 222770 195475 60 11437 2917806 221428 191522 90 11591 2922447 222470 192764 Average 11302 2919038 222223 193254 Time Actual Area Normalized Area Pressure (min) o 2 N 2 H2S S 0 2 o 2 H2S S 0 2 kPa 15 11365 4695089 23924 59629 7066 14874 37073 225 45 12340 4758463 73714 111537 7570 45219 68421 239 75 15831 4652959 73354 129861 9932 46019 81468 253 105 15674 4727688 144797 156545 9678 89403 96656 273 Time Conversion % (min) S 0 2 15 37 93 81 45 33 80 65 75 12 79 58 105 14 60 50 131 Appendix B Catalyst Type Porocel Fe-3 Date of Exp. Aug 21/2002 Catalyst Weight(g) 1.00 Reactor Temperature(°C) 300 Feed Calibration S 0 2 Calibration Time Peak Area Peak Area (min) o 2 N2 H 2 S N2 so 2 30 90264 4393404 124556 4449071 181280 60 93517 4521351 128836 4450230 185812 90 91172 4414522 126801 Average 91651 4443092 126731 4449651 183546 Feed Calibration S 0 2 Calibration Flowrate(ml/min) 0 2 N 2 H 2S Flowrate(mol/min) 0 2 N 2 H 2S Flowrate (ml/min) S 0 2 N 2 Flowrate(mol/min) S 0 2 N 2 3.99 172.01 2.72 1.78E-04 7.68E-03 1.21 E-04 2.72 171.2 1.21 E-04 7.64E-03 Time Actual Area Normalized Area Pressure (min) o 2 N 2 H 2S so 2 o 2 H 2 S S 0 2 kPa 30 5674 4409313 0 143042 5717 0 144351 115 60 18892 4497881 0 106649 18662 0 105505 115 90 38531 4469962 0 51455 38299 0 51221 115 120 48627 4444815 2072 21681 48608 2071 21705 115 150 52423 4447678 3256 7825 52369 3253 7828 115 180 54673 4451424 4741 9769 54571 4732 9765 115 210 55920 4423256 4313 4941 56171 4332 4970 115 240 56605 4376767 5737 0 57463 5824 0 115 270 57455 4385437 6428 0 58210 6513 0 115 300 57944 4423682 6999 0 58198 7030 0 115 330 57946 4385434 7313 0 58708 7409 0 115 Time mol/min mol/min mol/min Conversion % Selectivity % (min) o 2 H 2 S S 0 2 o 2 H 2S S so 2 30 1.11E-05 0.00E+00 9.55E-05 94 100 21 79 60 3.63E-05 0.00E+00 6.98E-05 80 100 43 57 90 7.44E-05 0.00E+00 3.39E-05 58 100 72 28 120 9.45E-05 1.98E-06 1.44E-05 47 98 88 12 150 1.02E-04 3.12E-06 5.18E-06 43 97 96 4 180 1.06E-04 4.53E-06 6.46E-06 40 96 94 6 210 1.09E-04 4.15E-06 3.29E-06 39 97 97 3 240 1.12E-04 5.58E-06 O.OOE+00 37 95 100 0 270 1.13E-04 6.24E-06 0.00E+00 36 95 100 0 300 1.13E-04 6.74E-06 0.00E+00 37 94 100 0 330 1.14E-04 7.10E-06 0.00E+00 36 94 100 0 132 Appendix B Catalyst Type Porocel Fe-3 Date of Exp. Aug 23/ 2002 Catalyst Weight(g) 1.00 Reactor Temperature(°C) 300 Preconditioned (Fe-sulphide) Feed Calibration S 0 2 Calibration Time Peak Area Peak Area (min) o 2 N 2 H 2 S N 2 s o 2 30 90264 4393404 124556 4449071 181280 60 93517 4521351 128836 4450230 185812 90 91172 4414522 126801 Average 91651 4443092 126731 4449651 183546 Feed Calibration S 0 2 Calibration Flowrate(ml/min) 0 2 N 2 H 2S Flowrate(mol/min) 0 2 N 2 H 2S Flowrate (ml/min) S 0 2 N 2 Flowrate(mol/min) S 0 2 N 2 3.99 172.01 2.72 1.78E-04 7.68E-03 1.21E-04 2.72 171.2 1.21E-04 7.64E-03 Time Actual Area Normalized Area Pressure (min) o 2 N 2 H 2 S s o 2 o 2 H 2 S S 0 2 kPa 30 3198 4430948 0 147288 3207 0 147910 115 60 4444 4476488 0 145674 4411 0 144801 115 90 5969 4432887 0 148282 5983 0 148843 115 120 6875 4500560 0 147929 6787 0 146256 115 150 7137 4455176 0 132973 7118 0 132808 115 180 6868 4415586 0 147754 6911 0 148894 115 210 7009 4564988 0 146375 6822 0 142677 115 240 270 300 330 Time mol/min mol/min mol/min Conversion % Selectivity % (min) o 2 H 2S S 0 2 o 2 H 2S S S 0 2 30 6.23E-06 0.00E+00 9.79E-05 97 100 19 81 60 8.57E-06 0.00E+00 9.58E-05 95 100 21 79 90 1.16E-05 0.00E+00 9.85E-05 93 100 19 81 120 1.32E-05 0.00E+00 9.68E-05 93 100 20 80 150 1.38E-05 0.00E+00 8.79E-05 92 100 28 72 180 1.34E-05 O.OOE+00 9.85E-05 92 100 19 81 210 1.33E-05 0.00E+00 9.44E-05 93 100 22 78 133 Appendix C APPENDIX C - Summary of Role of FeS2 (Metal Sulphide) and A1 20 3 (Catalyst Support) Catalyst Type A l 2 0 3 Date of Exp. April 22/2002 Catalyst Weight(g) 0.1179 Reactor 300 Temperature(°C) Feed Calibration Time Pick Area (min) o2 N 2 H 2 S so2 30 25502 3267042 241980 150496 60 19786 3461196 246017 160475 Average 22644 3364119 243999 155486 Feed(ml/min)-STP Feed(mol/min)-STP o 2 N 2 H2S so 2 o 2 N 2 H2S so 2 1.78 165.54 7.12 3.56 7.28E-05 6.77E-03 2.91 E-04 1.46E-04 Time Actual Area Normalized Area Pressure (min) o 2 N 2 H 2S S 0 2 o 2 H 2S S 0 2 kPa 30 18034 3372338 114056 74233 17990 113778 74052 122 60 18276 3416304 126475 92615 17997 124543 91200 120 90 15398 3534650 129901 87073 14655 123634 82872 120 120 17081 3477942 131463 90736 16522 127161 87766 120 150 18617 3418789 132888 93637 18319 130763 92140 120 180 18807 3411046 134086 98127 18548 132241 96777 120 210 18441 3432548 134202 87887 18073 131527 86135 120 240 18537 3435966 135112 98865 18149 132287 96798 120 Time Conversion % c i nve rs ion % Selectivity % (min) o 2 H 2S so 2 s, H 2 0 o 2 H2S S 0 2 S S02 30 5.78E-05 1.36E-04 6.93E-05 2.77E-05 1.55E-04 21 53 52 55 45 60 5.79E-05 1.49E-04 8.54E-05 2.57E-05 1.43E-04 21 49 41 40 60 90 4.71 E-05 1.48E-04 7.76E-05 2.67E-05 1.44E-04 35 49 47 46 54 120 5.31 E-05 1.52E-04 8.22E-05 2.61 E-05 1.39E-04 27 48 44 41 59 150 5.89E-05 1.56E-04 8.63E-05 2.56E-05 1.35E-04 19 46 41 36 64 180 5.96E-05 1.58E-04 9.06E-05 2.51 E-05 1.33E-04 18 46 38 32 68 210 5.81 E-05 1.57E-04 8.07E-05 2.63E-05 1.34E-04 20 46 45 40 60 240 5.83E-05 1.58E-04 9.06E-05 2.51 E-05 1.33E-04 20 46 38 32 68 134 Appendix C Catalyst Type FeS 2 /AI 2 0 3 Date of Exp. April 23/2002 Catalyst Weight(g) 0.1193 Reactor 300 Temperature(°C) Feed Calibration Pick Area O z N 2 H 2 S S 0 2 Time (min) 30 12772 3444039 256104 169531 60 11761 3578580 248982 165378 90 16185 3421688 253639 157529 120 9215 3567928 258170 172149 Average 12483 3503059 254224 166147 Feed(ml /min)-STP Feed(mol/min)-STP o2 N 2 H 2S S 0 2 o2 N 2 H 2S so2 1.78 165.54 7.12 3.56 7.28E-05 6.77E-03 2.91 E-04 1.46E-04 Time Actual Area Normalized Area Pressure (min) o2 N 2 H 2S S 0 2 o2 H 2S S 0 2 kPa 30 60 0 3658234 77773 40410 0 74474 38696 120 90 0 3607909 83956 46683 0 81516 45326 120 120 0 3602641 86985 49793 0 84581 48417 120 150 0 3636706 88462 45867 0 85211 44181 120 180 0 3686792 88601 49191 0 84186 46740 120 210 3023 3591416 102955 122602 2949 100422 119586 120 240 3009 3600080 106239 131646 2928 103376 128098 120 270 3071 3569657 108181 123322 3014 106163 121021 120 300 2378 3638844 109915 131073 2289 105813 126182 120 Time Conversion % Conversion % Selectivity % (min) so2 s S 0 2 30 60 0.00E+00 8.53E-05 3.39E-05 3.22E-05 2.06E-04 100 71 77 84 16 90 0.00E+00 9.34E-05 3.97E-05 3.14E-05 1.98E-04 100 68 73 80 20 120 0.00E+00 9.69E-05 4.24E-05 3.11E-05 1.94E-04 100 67 71 78 22 150 0.00E+00 9.76E-05 3.87E-05 3.16E-05 1.94E-04 100 66 73 80 20 180 O.OOE+OO 9.64E-05 4.10E-05 3.13E-05 1.95E-04 100 67 72 79 21 210 1.72E-05 1.15E-04 1.05E-04 2.33E-05 1.76E-04 76 60 28 41 59 240 1.71E-05 1.18E-04 1.12E-04 2.24E-05 1.73E-04 77 59 23 35 65 270 1.76E-05 1.22E-04 1.06E-04 2.31 E-05 1.70E-04 76 58 27 37 63 300 1.34E-05 1.21 E-04 1.11E-04 2.26E-05 1.70E-04 82 58 24 35 65 135 Appendix D APPENDIX D - Summary of Catalytic Activity Test (Bulk FeS) D. 1 Mass Flow Controllers & Feed Composition Premixed reactants H2S/N2, S0 2 /N2 and pure gases N2, H2 as well as O2/N2 (Extra Dry Ai r ) was supplied by Praxair Company. The flowrates of the premixed reactants streams F k S / ^ SO2/N2, O2/N2 and pure N2 or H2 were controlled by four separate 5850E Brookes mass flow controllers with " D " type connector. Mass flow controllers were connected to a 5878 Brooks control box with four output channels. Each channel provided output signal to operate the mass flow control valves in the range of 0-100%. Because some of mass flow controllers were operated for a gas other than the gas it was calibrated with, a correction factor was used for each controller. For more detail information about calculation of correction factor, see " I N S T A L A T I O N A N D O P E R A T I N G I N S T R U C T I O S , Brooks Mass F low Controller, Model 5850 E , Issue 7, August 1997". Table D . l shows the mass flow controllers specifications. Flow Controller Flow Range (SCCM) Calibration Gas Actual Gas Correction Factor MFC1 0-200 N 2 H 2 S/N 2 0.973 MFC2 0-20 C H 4 0 2 /N 2 1.237 MFC3 0-500 NH 3 N 2 1.272 MFC4 0-200 N 2 S0 2 /N 2 0.947 Table D . l Mass Flow Controllers' specifications ( M F C ) 136 Appendix D The reactants were mixed together before the reactor to provide the required feed gas composition. Controllers' output signal was calculated using a spread sheet in order to provide the desired gas composition at the entrance of catalytic bed. Table D.2 shows the spread sheet with a sample calculation for a feed with composition of H2S 6 vol . %, SO2 0 vol . %, 0 2 1 vol . %, H 2 0 20 vol . % and balance N 2 . The desired flowrate at the entrance of catalytic bed was 200 ml/min at 270°C. Because the gases and the mass flow controllers were in the room temperature the flow rate of each gas was calculated at ambient temperature and pressure. Ambient temperature(°C) 21 Reactor flowrate(ml/min) 200 Reactor temperature(°C) 270 Reactor pressure(kPa) 115 Feed Composition H 2S vol. % 6 S 0 2 vol. % 0 0 2 vol. % 1 H 2 0 vol.% 20 N 2 vol. % 73 Reactants F low Rate ( at ambient temperature) Total Feed (ml/min) 123.0 H 2S (ml/min) 7.38 S 0 2 (ml/min) 0.00 0 2 (ml/min) 1.23 H 2 0 (ml/min)_Vapour 24.60 H 2 0 (ml/min)_Liquid 1.25 N 2 (ml/min)_Total 89.81 N2(ml/min)_Pure 43.87 Table D.2 Calculation of reactants flow rate 137 Appendix D The flow rate of each reactants was used to the output signal was determined. Output Signal Calculat ion Correction factor (MFC1) 0.973 Output signal % (MFC1) 25.0 Correction factor (MFC2) 1.237 Output signal % (MFC2) 23.7 Correction factor (MFC3) 1.272 Output signal % (MFC3) 6.9 Correction factor (MFC4) 0.947 Output signal % (MFC4) 0.0 Table D.3 Calculation of the output signal; ( M F C ) = Mass F low Controller D.2 Calibration and Base Area In order to calibration of the peak surface area, before conducting each experiment, the feed was by-passed the reactor and it was introduced to G C . Samples were taken from the feed and the peak surface areas of each reactant were recorded in a table. To measure the SO2 produced in the course o f experiment, same procedure was conducted on the mixture of N2/SO2. The premixed N 2 / S 0 2 and pure N2 with a known concentration were introduced to the G C and the samples were taken from the mixture. Table D.4 shows the sample feed calibration and SO2 calibration peak surface areas. 138 Appendix D Feed Calibration S 0 2 Calibration Time Peak Area Peak Area (min) o2 N2 H2S N2 so2 30 20832 3526895 529964 3765320 265757 60 6294 3278533 533584 3465762 282796 90 6548 3297226 536244 3743376 294588 120 3734660 300169 Average 11225 3367551 533264 3677280 285828 T a b l e D.4 Calibration of reactants During the actual run the gas samples were taken from the effluent of the reactor. Gases were cooled down to room temperature before being introduced to G C . The actual surface area were recorded in the table corresponded to sampling time. In order to use the surface area o f the actual peaks, first the surface areas were normalized using N 2 surface area (concentration) as base gas. Normalized Peak Area = ( N 2 Calibration Area I N 2 Actual Area) x (Reactant Actual Area) Time Actual Area Normalized Area Pressure (min) N 2 H 2S S 0 2 H 2S S 0 2 kPa 30 3613284 457358.66 0 426255 0 115 60 3543989 460901.43 25521 437955 26481 115 90 3557286 463598.9 23114 438872 23894 115 120 3579303 475091.77 27510 446985 28263 115 T a b l e D.5 Normalization of the actual area using N 2 as a base gas 139 Appendix D The normalized area was used to calculate the molar flowrate of reactant at effluent of catalytic bed. Table D.6 shows the molar flow arte of reactants in the feed stream (except H2O) and also molar flowrate of SO2 calibration. Feed Calibration S 0 2 Calibration Flowrate(ml/min) 0 2 N 2 H 2S Flowrate(mol/min) 0 2 N 2 H 2S Flowrate (ml/min) S 0 2 N 2 Flowrate(mol/min) S 0 2 N 2 1.23 89.80 7.38 5.03E-05 3.67E-03 3.02E-04 3.69 119.33 1.51 E-04 4.88E-03 T a b l e D.6 Molar flow rate of reactants in feed stream and calibration of SO2 Using the flowrate of reactants at influent and effluent of catalytic bed the conversion and selectivity were calculated as follow: Time mol/min mol/min mol/min mol/min Conversion % Selectivity % (min) H2S S 0 2 s . H 2 0* o 2 H2S S S 0 2 30 2.41 E-04 0.00E+00 7.57E-06 6.06E-05 60.2 20 100 0 60 2.48E-04 1.40E-05 5.00E-06 5.39E-05 81.4 18 74 26 90 2.48E-04 1.26E-05 5.10E-06 5.34E-05 78.2 18 76 24 120 2.53E-04 1.49E-05 4.24E-06 4.88E-05 78.2 16 69 31 Average STDEV 79.3 73 2f 1 9 09 3 5 3 5 T a b l e D .7 Molar flowrate, conversion and selectivity calculations Using the flowrate of reactants at influent and effluent of catalytic bed the conversion and selectivity were calculated as follow: 140 Appendix D mol Ss produced = (—) x (mol H2S converted - mol S02 produced) 8 mol 02 converted = ( — ) x mol Ss produced + (2) x mol SO2 produced 16 The conversion and selectivity were calculated as follows: 0 2 Conversion % = (Molar F low 0 2 In - [0.5 x (Molar F low H 2 S In - Molar flow H 2 S Out) + (Molar flow S 0 2 Out)]) / (Molar F low 0 2 In) x 100 H 2 S Conversion % = (Molar F low H 2 S In - Molar flow Ff 2S Out)/ (Molar F low H 2 S In) xlOO S 0 2 Selectivity % = (Molar flow S 0 2 Out)/ (Molar F low H 2 S In - Molar flow H 2 S Out) xlOO S 8 Selectivity % = 100 - S 0 2 Selectivity % D.3 Surface Area of O2 Preliminary work showed that using the same direct method to calculate the conversion for 0 2 had significant error. Results showed the fluctuation in measurement of area for 0 2 and the nature of error was due to low concentration of 0 2 and absorption of 0 2 in the system and in the G C columns. Figure D . l shows the area obtained from G C for 0 2 from a feed gas stream contains 20 vol . % 0 2 and 80 vol . % N 2 . 141 Appendix D 400000 380000 360000 340000 320000 n 2 300000 280000 260000 240000 -\ 220000 200000 • • • i r 0 1 2 3 4 5 6 7 8 Sample No. F i g u r e D . l Surface A r e a for 0 2 peak for seven sampl ing at constant concentration 142 Appendix D D.4 Reaction Kinetics Summary Feed Inlet Concentration (vol. %) Temperature °C(±1) Pressure kPa(±5) Weight g(±0.02) Exp. Code o 2 H2S H 2 0 N 2 1 6 20 73 270 115 0.23 E01 2 6 20 72 270 115 0.23 E02 3 6 20 71 270 115 0.23 E03 4 6 20 70 270 115 0.23 E04 2 6 20 72 270 115 0.23 E05 2 8 20 70 270 115 0.23 E06 2 10 20 68 270 115 0.23 E07 2 6 10 82 270 115 0.23 E08 2 6 20 72 270 115 0.23 E09 2 6 30 62 270 115 0.23 E10 2 6 20 72 250 115 0.23 E11 2 6 20 72 270 115 0.23 E12 2 6 20 72 290 115 0.23 E13 T a b l e D . 8 Exper imenta l condi t ions 143 Appendix D Exp. Flowrate(mol/min) Cat. Weight 0 2 Rate Code 0 2 In 0 2 Out 0 2 Diff. (g) (mol/min-g) E01 5.0E-05 1.0E-05 3.9E-05 0.2434 1.62E-04 E02 9.9E-05 1.7E-06 9.8E-05 0.2434 4.01 E-04 E03 1.5E-04 0.0E+00 1.5E-04 0.2434 6.03E-04 E04 2.0E-04 1.4E-05 1.8E-04 0.2434 7.58E-04 E05 1.0E-04 3.4E-05 6.7E-05 0.2150 3.1 OE-04 E06 1 .OE-04 3.0E-05 7.0E-05 0.2150 3.27E-04 E07 9.9E-05 2.1 E-05 7.8E-05 0.2150 3.63E-04 E08 1.04E-04 1.60E-05 8.76E-05 0.2272 3.85E-04 E09 1.02E-04 1.99E-05 8.23E-05 0.2272 3.62E-04 E10 9.76E-05 2.30E-05 7.46E-05 0.2272 3.28E-04 E11 1 .OE-04 6.8E-05 3.4E-05 0.2150 1.60E-04 E12 9.9E-05 5.2E-05 4.7E-05 0.2150 2.20E-04 E13 9.6E-05 4.0E-05 5.6E-05 0.2150 2.60E-04 T a b l e D . 9 0 2 consumpt ion rate i n each exper imental cond i t ion 144 Appendix D Experiment E01 Code Date of Exp. July 16th/2003 Catalyst Weight(g) 0.2434 Reactor Temperature(°C) 270 Feed Calibration Peak Area 0 2 N 2 H 2 S S 0 2 Calibration Peak Area N 2 S 0 2 Time (min) 30 20832 3526895 529964 3765320 265757 60 6294 3278533 533584 3465762 282796 90 6548 3297226 536244 3743376 294588 120 3734660 300169 Average 11225 3367551 533264 3677280 285828 Feed Calibration S 0 2 Calibration Flowrate(ml/min) 0 2 N 2 H 2S Flowrate(mol/min) 0 2 N 2 H 2S Flowrate (ml/min) S 0 2 N 2 Flowrate(mol/min) S 0 2 N 2 1.23 89.80 7.38 5.03E-05 3.67E-03 3.02E-04 3.69 119.33 1.51 E-04 4.88E-03 Time Actual Area Normalized Area Pressure (min) N 2 H 2S s o 2 H 2S S 0 2 kPa 30 3613284 457358.66 0 426255 0 115 60 3543989 460901.43 25521 437955 26481 115 90 3557286 463598.9 23114 438872 23894 115 120 3579303 475091.77 27510 446985 28263 115 Time mol/min mol/min mol/min mol/min Conversion % Selectivity % (min) H2S S 0 2 s„ H 20* o 2 H2S S S 0 2 30 2.41 E-04 0.00E+00 7.57E-06 6.06E-05 60.2 20 100 0 60 2.48E-04 1.40E-05 5.00E-06 5.39E-05 81.4 18 74 26 90 2.48E-04 1.26E-05 5.10E-06 5.34E-05 78.2 18 76 24 120 2.53E-04 1.49E-05 4.24E-06 4.88E-05 78.2 16 69 31 Average 79.3 17 73 " . 27 STDEV 1 9 0.9 3.5 3.5 H 2 0* = produced by reaction 145 Appendix D Experiment E02 Code Date of Exp. July 16th/2003 Catalyst Weight(g) 0.2434 Reactor Temperature(°C) 270 Feed Calibration Peak Area 0 2 N 2 H 2 S S0 2 Calibration Peak Area N 2 S 0 2 Time (min) 30 29108 3462245 532491 3765320 265757 60 24126 3443791 535637 3465762 282796 90 35098 3475593 533654 3743376 294588 120 3734660 300169 Average 29444 3460543 533927 3677280 285828 Feed Calibration S 0 2 Calibration Flowrate(ml/min) 0 2 N 2 H2S Flowrate(mol/min) 0 2 N 2 H2S Flowrate (ml/min) S0 2 N 2 Flowrate(mol/min) S0 2 N 2 2.46 88.57 7.38 1.01 E-04 3.62E-03 3.02E-04 3.69 119.33 1.51 E-04 4.88E-03 Time (min) s o 2 Normalized Area H2S S 0 2 Pressure kPa 30 3542269 370319 92955 361775 96498 115 60 3555877 372970 95971 362971 99248 115 90 3515228 373604 89128 367792 93237 115 120 3589907 374965 92942 361453 95204 115 Time mol/min mol/min mol/min mol/min H2S S0 2 S, H20* Conversion % Selectivity % (min) 0 2 H2S s s o 2 30 2.05E-04 5.10E-05 5.80E-06 9.73E-05 99.0 32 48 52 60 2.05E-04 5.24E-05 5.53E-06 9.66E-05 100.1 32 46 54 90 2.08E-04 4.92E-05 5.59E-06 9.39E-05 95.6 31 48 52 120 2.04E-04 5.03E-05 5.90E-06 9.75E-05 98.4 32 48 52 Average 98.3 I 32 47 53 STDEV 1.9 0.5 1.1 1.1 H 2 0* = produced by reaction 146 Appendix D Experiment Code E03 Date of Exp. July 16th/2003 Catalyst Weight(g) 0.2434 Reactor Temperature(°C) 270 Feed Calibration Peak Area 0 2 N 2 H 2 S SOi Calibration Peak Area N 2 S 0 2 Time (min) 30 27895 3355163 534597 3765320 265757 60 6294 3278533 533584 3465762 282796 90 6548 3297226 536244 3743376 294588 120 3734660 300169 Average 13579 3310307 534808 3677280 285828 Feed Calibration S 0 2 Calibration Flowrate(ml/min) Flowrate(mol/min) Flowrate (ml/min) Flowrate(mol/min) o 2 N 2 H 2S o 2 N 2 H 2S S 0 2 N 2 S 0 2 N 2 3.69 87.33 7.38 1.51 E-04 3.57E-03 3.02E-04 3.69 119.33 1.51 E-04 4.88E-03 Time (min) SC-2 Normalized Area H 2S S 0 2 Pressure kPa 30 3590020 319465 162118 294574 166058 115 60 3584603 324821 166393 299966 170695 115 90 3524112 329548 154667 309555 161389 115 120 3559975 334719 160828 311245 166127 115 Time mol/min mol/min mol/min mol/min H 2S S 0 2 S 8 H 2Q* Conversion % Selectivity % (min) 0 2 H 2S S S 0 2 30 1.66E-04 8.77E-05 5.99E-06 1.36E-04 103.0 45 35 65 60 1.69E-04 9.01E-05 5.30E-06 1.33E-04 103.6 44 32 68 90 1.75E-04 8.52E-05 5.24E-06 1.27E-04 98.6 42 33 67 120 1.76E-04 8.77E-05 4.81 E-06 1.26E-04 99.9 42 30 70 Average 101.3 1 " 4 3 , 33 67 ...;.^ >::;.s.:.; •-•:..'y~~. STDEV 2.4 1.5 2.0 1 2.0 i H 2 0* = produced by reaction 147 Appendix D Experiment Code E04 Date of Exp. July 16th/2003 Catalyst Weight(g) 0.2434 Reactor Temperature(°C) 270 Feed Calibration Peak Area 0 2 N 2 H 2 S SC-2 Calibration Peak Area N 2 S 0 2 Time (min) 30 20832 3526895 529964 3765320 265757 60 6294 3278533 533584 3465762 282796 90 6548 3297226 536244 3743376 294588 120 3734660 300169 Average 11225 3367551 533264 3677280 285828 Feed Calibration S 0 2 Calibration Flowrate(ml/min) 0 2 N 2 H 2S Flowrate(mol/min) 0 2 N 2 H 2S Flowrate (ml/min) S 0 2 N 2 Flowrate(mol/min) S 0 2 N 2 4.92 86.11 7.38 2.01 E-04 3.52E-03 3.02E-04 3.69 119.33 1.51 E-04 4.88E-03 Time (min) S 0 2 Normalized Area H 2S S 0 2 Pressure kPa 30 3526082 294172 216853 280946 226152 115 60 3555853 300330 204390 284426 211370 115 90 3552844 314904 190037 298481 196693 115 120 3608990 265878 244508 248091 249135 115 Time mol/min mol/min mol/min mol/min H 2S S 0 2 S 8 H 2 0* Conversion % Selectivity % (min) 0 2 H 2S S S 0 2 30 1.59E-04 1.19E-04 2.93E-06 1.43E-04 94.8 47 16 84 60 1.61 E-04 1.12E-04 3.66E-06 1.41 E-04 90.5 47 21 79 90 1.69E-04 1.04E-04 3.63E-06 1.33E-04 84.6 44 22 78 120 1.40E-04 1.32E-04 3.73E-06 1.61 E-04 105.5 53 19 81 Average 93.8 48 19 81 STDEV 8.8 4.0 2.4 2.4 H 2 0* = produced by reaction 148 Appendix D Experiment Code E05 Date of Exp. July 24th/2003 Catalyst Weight(g) 0.215 Reactor TemperaturefC) 270 Feed Calibration Peak Area 0 2 N 2 H 2 S S 0 2 Calibration Peak Area N 2 S 0 2 Time (min) 30 29108 3462245 532491 3765320 265757 60 24126 3443791 535637 3465762 282796 90 35098 3475593 533654 3743376 294588 120 3734660 300169 Average 29444 3460543 533927 3677280 285828 Feed Calibration S0 2 Calibration Flowrate(ml/min) Flowrate(mol/min) Flowrate (ml/min) Flowrate(mol/min) o 2 N 2 H 2S o 2 N 2 H 2S S 0 2 N 2 S 0 2 N 2 2.46 88.57 7.38 1.01 E-04 3.62E-03 3.02E-04 3.69 119.33 1.51 E-04 4.88E-03 Time (min) S 0 2 Normalized Area H 2S S 0 2 Pressure kPa 30 3522886 379730 45960 373010 47974 115 60 3566577 393417 47115 381721 48577 115 90 3509619 400484 44878 394884 47022 115 120 3521864 402540 44392 395531 46351 115 Time mol/min mol/min mol/min mol/min H 2S S 0 2 S„ H 2 0* Conversion % Selectivity % (min) 0 2 H 2S S S 0 2 30 2.11 E-04 2.53E-05 8.20E-06 9.10E-05 70.4 30 72 28 60 2.16E-04 2.56E-05 7.55E-06 8.60E-05 68.3 29 70 30 90 2.23E-04 2.48E-05 6.72E-06 7.86E-05 63.7 26 68 32 120 2.24E-04 2.45E-05 6.72E-06 7.82E-05 63.2 26 69 31 Average 66.4 28 70 30 STDEV 3.5 2.0 1.7 | 1.7 H 2 0* = produced by reaction 149 Appendix D Experiment Code E06 Date of Exp. July 24th/2003 Catalyst Weight(g) 0.215 Reactor TemperaturefC) 270 Feed Calibration Peak Area 0 2 N 2 H 2 S S 0 2 Calibration Peak Area N 2 S 0 2 Time (min) 30 27273 3404982 730971 3765320 265757 60 7194 2966001 737672 3465762 282796 90 21818 3346042 737975 3743376 294588 120 3734660 300169 Average 18762 3239008 735539 3677280 285828 Feed Calibration SO2 Calibration Flowrate(ml/min) 0 2 N 2 H 2S Flowrate(mol/min) 0 2 N 2 H 2S Flowrate (ml/min) S 0 2 N 2 Flowrate(mol/min) S 0 2 N 2 2.46 86.10 9.84 1.01 E-04 3.52E-03 4.02E-04 3.69 119.33 1.51 E-04 4.88E-03 Time (min) so2 Normalized Area H 2S S 0 2 Pressure kPa 30 3439025 562839 25352 530104 27108 115 60 3417134 564503 30482 535077 32803 115 90 3424549 562843 23517 532348 25253 115 120 115 Time mol/min mol/min mol/min mol/min H 2S S 0 2 S„ H 2 0* Conversion % Selectivity % (min) 0 2 H 2S S so2 30 2.90E-04 1.43E-05 1.23E-05 1.12E-04 70.1 28 87 13 60 2.93E-04 1.73E-05 1.15E-05 1.1 OE-04 71.7 27 84 16 90 2.91 E-04 1.33E-05 1.22E-05 1.11 E-04 68.5 • 28 88 12 120 Average 70.1 | 28 86 14 STDEV 1.6 0.3 2.0 2.0 H 2 0* = produced by reaction 150 Appendix D Experiment Code E07 Date of Exp. July 24th/2003 Catalyst Weight(g) 0.215 Reactor Temperature(°C) 270 feed Calibration Peak Area 0 2 N 2 H 2 S S 0 2 Calibration Peak Area N 2 S 0 2 Time (min) 30 52518 3232910 812634 3765320 265757 60 28055 3280358 816956 3465762 282796 90 58389 3208218 818048 3743376 294588 120 33153 3273753 819499 3734660 300169 Average 43029 3248810 909540* 3677280 285828 Feed Calibration S 0 2 Calibration Flowrate(ml/min) 0 2 N 2 H 2S Flowrate(mol/min) 0 2 N 2 H 2S Flowrate (ml/min) S 0 2 N 2 Flowrate(mol/min) S 0 2 N 2 2.46 85.67 12.30 1.01 E-04 3.50E-03 5.03E-04 3.69 119.33 1.51 E-04 4.88E-03 Time (min) so2 Normalized Area H 2S S 0 2 Pressure kPa 30 3355037 600000 11603 581003 12717 115 60 3378175 707819 22293 680713 24267 115 90 3328902 708509 22557 691463 24918 115 120 3357092 704371 12923 681652 14156 115 Time mol/min mol/min mol/min mol/min H 2S S 0 2 S , H 2 0* Conversion % Selectivity % (min) 0 2 H 2S S S 0 2 30 3.21 E-04 6.71 E-06 2.19E-05 1.82E-04 97.0 36 96 4 60 3.76E-04 1.28E-05 1.42E-05 1.27E-04 75.6 25 90 10 90 3.82E-04 1.32E-05 1.34E-05 1.21 E-04 73.0 24 89 11 120 3.77E-04 7.47E-06 1.48E-05 1.26E-04 70.1 25 94 6 Average 78.9 28 92 8 STDEV 12.2 5.7 34 3.4 H 2 0* = produced by reaction 151 Appendix D Experiment Code E08 Date of Exp. July 21th/2003 Catalyst Weight(g) 0.2272 Reactor TemperaturefC) 290 Feed Calibration Peak Area 0 2 N 2 H 2 S S 0 2 Calibration Peak Area N 2 S 0 2 Time (min) 30 30661 3506607 466177 3765320 265757 60 20277 3449040 468018 3465762 282796 90 21387 3464030 466897 3743376 294588 120 3734660 300169 Average 24108 3473226 467031 3677280 285828 Feed Calibration S 0 2 Calibration Flowrate(ml/min) 0 2 N 2 H 2S Flowrate(mol/min) 0 2 N 2 H 2S Flowrate (ml/min) S 0 2 N 2 Flowrate(mol/min) S 0 2 N 2 2.46 100.87 7.38 1.01 E-04 4.13E-03 3.02E-04 3.69 119.33 1.51 E-04 4.88E-03 Time (min) SOj Normalized Area H 2S S 0 2 Pressure kPa 30 3588750 348829 84343 337600 86424 115 60 3567984 351117 82725 341792 85259 115 90 3585728 352504 76720 341444 78679 115 120 3565039 353822 82596 344710 85196 115 Time mol/min mol/min mol/min mol/min H 2S S 0 2 S 8 H 2 0* Conversion % Selectivity % (min) 0 2 H 2S s so2 30 2.18E-04 4.56E-05 4.75E-06 8.36E-05 86.9 28 45 55 60 2.21 E-04 4.50E-05 4.49E-06 8.09E-05 85.0 27 44 56 90 2.21 E-04 4.15E-05 4.95E-06 8.12E-05 81.6 27 49 51 120 2.23E-04 4.50E-05 4.26E-06 7.91 E-05 84.0 26 43 57 Average 84.4 27 45 55 STDEV 2.2 1 0.6 2.5 2.5 H 2 0* = produced by reaction 152 Appendix D Experiment Code E09 Date of Exp. July 21th/2003 Catalyst Weight(g) 0.2272 Reactor Temperature(°C) 290 Feed Calibration Peak Area 0 2 N 2 H 2 S S 0 2 Calibration Peak Area N 2 S 0 2 Time (min) 30 29108 3462245 532491 3765320 265757 60 24126 3443791 535637 3465762 282796 90 35098 3475593 533654 3743376 294588 120 3734660 300169 Average 29444 3460543 533927 3677280 285828 Feed Calibration S 0 2 Calibration Flowrate(ml/min) 0 2 N 2 H 2S Flowrate(mol/min) 0 2 N 2 H 2S Flowrate (ml/min) S 0 2 N 2 Flowrate(mol/min) S 0 2 N 2 2.46 88.57 7.38 1.01 E-04 3.62E-03 3.02E-04 3.69 119.33 1.51 E-04 4.88E-03 Time (min) so 2 Normalized Area H 2S S 0 2 Pressure kPa 30 3474627 394326 67682 11749 387897 115 60 3477048 385396 66550 6126 392728 115 90 3513093 387735 74116 5981 383567 115 120 3549944 390917 72109 7530 381935 115 Time mol/min mol/min mol/min mol/min H 2S S 0 2 S 8 H 2 0* Conversion % Selectivity % (min) 0 2 H 2S S S 0 2 30 2.22E-04 3.78E-05 5.25E-06 7.98E-05 77.3 26 53 47 60 2.17E-04 3.72E-05 5.98E-06 8.50E-05 79.2 28 56 44 90 2.16E-04 4.10E-05 5.62E-06 8.59E-05 83.4 28 52 48 120 2.15E-04 3.94E-05 5.87E-06 8.64E-05 82.1 29 54 46 Average 80.5 28 54 46 STDEV 2.8 ! 1.0 1 8 18 H 2 0* = produced by reaction 153 Appendix D Experiment E10 Code Date of Exp. July 21th/2003 Catalyst Weight(g) 0.2272 Reactor Temperature(°C) 290 Feed Calibration Peak Area 0 2 N 2 H 2 S S 0 2 Calibration Peak Area N 2 S 0 2 Time (min) 30 20066 3298719 593213 3765320 265757 60 27855 3361934 592521 3465762 282796 90 36251 3386932 592226 3743376 294588 120 3734660 300169 Average 28057 3349195 592653 3677280 285828 Feed Calibration S Q 2 Calibration Flowrate(ml/min) 0 2 N 2 H 2S Flowrate(mol/min) 0 2 N 2 H 2S Flowrate (ml/min) S 0 2 N 2 Flowrate(mol/min) S 0 2 N 2 2.46 76.27 7.38 1.01 E-04 3.12E-03 3.02E-04 3.69 119.33 1.51 E-04 4.88E-03 Time (min) S 0 2 Normalized Area H 2S SQ 2 Pressure kPa 30 3459574 438239 70447 424257 74880 115 60 3497020 438704 64046 420159 67347 115 90 3440107 437479 62512 425918 66822 115 120 3426822 434334 53868 424495 57805 115 150 3450537 435391 60555 422604 64534 115 Time mol/min mol/min mol/min mol/min H 2S S 0 2 S 8 H 2 0* Conversion % Selectivity % (min) 0 2 H 2S S so2 30 2.16E-04 3.95E-05 5.78E-06 8.58E-05 81.9 28 54 46 60 2.14E-04 3.56E-05 6.54E-06 8.78E-05 79.0 29 60 40 90 2.17E-04 3.53E-05 6.20E-06 8.49E-05 77.3 28 58 42 120 2.16E-04 3.05E-05 6.89E-06 8.56E-05 72.9 28 64 36 150 2.15E-04 3.41 E-05 6.57E-06 8.66E-05 76.9 29 61 39 Average 76.5 29 61 39 liliiiiiiMil STDEV 26 04 26 2.6 154 Appendix D Experiment Code E11 Date of Exp. August 1th/2003 Catalyst Weight(g) 0.215 Reactor Temperature(°C) 250 Feed Calibration Peak Area 0 2 N 2 H 2 S S 0 2 Calibration Peak Area N 2 S 0 2 Time (min) 30 29108 3462245 532491 3765320 265757 60 24126 3443791 535637 3465762 282796 90 35098 3475593 533654 3743376 294588 120 3734660 300169 Average 29444 3460543 533927 3677280 285828 Feed Calibration S 0 2 Calibration Flowrate(ml/min) 0 2 N 2 H 2S Flowrate(mol/min) 0 2 N 2 H 2S Flowrate (mi/min) S 0 2 N 2 Flowrate(mol/min) S 0 2 N 2 2.55 91.93 7.66 1.04E-04 3.76E-03 3.13E-04 3.69 119.33 1.51 E-04 4.88E-03 Time (min) S 0 2 Normalized Area H 2S S Q 2 Pressure kPa 30 3279120 413440 0 436314 0 115 60 3479377 428295 9927 425977 10492 115 90 3465518 431564 17658 430944 18737 115 120 3397180 434292 12138 442392 13139 115 Time mol/min mol/min mol/min mol/min H 2S S 0 2 S 8 H 2 0* Conversion % Selectivity % (min) 0 2 H 2S S S 0 2 30 2.56E-04 0.00E+00 7.16E-06 5.73E-05 27.5 18 100 0 60 2.50E-04 5.54E-06 7.23E-06 6.33E-05 35.7 20 91 9 90 2.53E-04 9.89E-06 6.32E-06 6.04E-05 38.5 19 84 16 120 2.60E-04 6.94E-06 5.85E-06 5.37E-05 32.4 17 87 13 Average 33.5 I ' 19 * 90 10 STDEV 4.7 1 3 7.1 H 2 0* = produced by reaction 155 Appendix D Experiment Code E12 Date of Exp. August 1th/2003 Catalyst Weight(g) 0.215 Reactor TemperaturefC) 270 Feed Calibration Peak Area 0 2 N 2 H 2 S S 0 2 Calibration Peak Area N 2 S 0 2 Time (min) 30 29108 3462245 532491 3765320 265757 60 24126 3443791 535637 3465762 282796 90 35098 3475593 533654 3743376 294588 120 3734660 300169 Average 29444 3460543 533927 3677280 285828 Feed Calibration S 0 2 Calibration Flowrate(ml/min) 0 2 N 2 H 2S Flowrate(mol/min) 0 2 N 2 H 2S Flowrate (ml/min) S 0 2 N 2 Flowrate(mol/min) S 0 2 N 2 2.46 88.57 7.38 1.01 E-04 3.62E-03 3.02E-04 3.69 119.33 1.51 E-04 4.88E-03 Time (min) S 0 2 Normalized Area H 2S S 0 2 Pressure kPa 30 3340382 398083 9867 412403 10862 115 60 3476396 403458 23718 401618 25089 115 90 3473299 405304 13019 403815 13784 115 120 3440710 408966 24539 411323 26226 115 Time mol/min mol/min mol/min mol/min H 2S S 0 2 S 8 H 2 0* Conversion % Selectivity % (min) 0 2 H 2S S S 0 2 30 2.33E-04 5.74E-06 7.87E-06 6.87E-05 39.8 23 92 8 60 2.27E-04 1.32E-05 7.69E-06 7.48E-05 50.3 25 82 18 90 2.28E-04 7.28E-06 8.28E-06 7.36E-05 43.8 24 90 10 120 2.33E-04 1.38E-05 6.93E-06 6.93E-05 48.2 23 80 20 Average 45.5 24 86 14 STDEV 4.7 j 1.0 5.7 5.7 H 2 0* = produced by reaction 156 Appendix D Experiment Code E13 Date of Exp. August 1th/2003 Catalyst Weight(g) 0.215 Reactor TemperaturefC) 290 Feed Calibration Peak Area <D2 N 2 H 2 S S 0 2 Calibration Peak Area N 2 S 0 2 Time (min) 30 29108 3462245 532491 3765320 265757 60 24126 3443791 535637 3465762 282796 90 35098 3475593 533654 3743376 294588 120 3734660 300169 Average 29444 3460543 533927 3677280 285828 Feed Calibration S 0 2 Calibration Flowrate(ml/min) 0 2 N 2 H 2S Flowrate(mol/min) 0 2 N 2 H 2S Flowrate (ml/min) SC-2 N 2 Flowrate(mol/min) S 0 2 N 2 2.37 85.42 7.12 9.69E-05 3.49E-03 2.91 E-04 3.69 119.33 1.51 E-04 4.88E-03 Time (min) S 0 2 Normalized Area H 2S S 0 2 Pressure kPa 30 3255610 391097 38614 415716 43615 115 60 3279530 391721 38347 413342 42998 115 90 3373396 392662 40160 402806 43778 115 120 3273542 388975 29978 411195 33675 115 Time mol/min mol/min mol/min mol/min H 2S S 0 2 S 8 H 2 0* Conversion % Selectivity % (min) 0 2 H 2S S S 0 2 30 2.27E-04 2.30E-05 5.18E-06 6.45E-05 57.0 22 64 36 60 2.25E-04 2.27E-05 5.38E-06 6.58E-05 57.3 23 65 35 90 2.20E-04 2.31 E-05 6.05E-06 7.15E-05 60.7 25 68 32 120 2.24E-04 1.78E-05 6.14E-06 6.69E-05 52.9 23 73 27 Average 57.0 23 68 u STDEV 3.2 1.1 4.1 • 4.1 H 2 0* = produced by reaction 157 Appendix E APPENDIX E - Summary of the Statistical Analysis E.l Data Points Construction Polynomial curves were fitted on the actual experimental and simulated data for each set of experiments using Origin software. Fitted curves were used to predict some extra data points in order to conducted statistical analysis. In the following tables A , B , C and D are: Also , the actual data points are highlighted in grey shade in order to be distinguished from predicted data points. E.2 Goodness of the Model Goodness of the model and predicted rate constants was investigated as following. In the following description, SSM is the sum of squares about the mean, SSE is the sum of squares errors (residuals), n is the total number of data values, and m is the number of coefficients in the model. DOF, the degree of freedom, is n-m. y = A + Bx + Cx2+ Dx3 Experimental y Simulated Experimental ^ Experimntal _Mean 158 Appendix E Coefficient of Determination(R-squared) SSM Standard Error (Root MSE) Std.Err.= . : S S E DOF F-Test (F-statistic) (SSM - SSE) l(m -1) F = -SSE I DOF A s a fitted m o d e l becomes more ideal , the R 2 values approach 1, the standard error decreases toward zero, and F tends toward inf ini ty . 159 Appendix E Statistical Analysis - Experiment Set 1 Experimental Simulation o2 H 2S S 0 2 o2 H 2S S 0 2 A 3.400E+01 -1.065E-13 -2.300E+01 1.000E+02 5.700E+00 -3.300E+00 B 6.150E+01 1.733E+01 6.600E+01 0.000E+00 1.878E+01 3.337E+01 C -1.740E+01 -3.550E-14 -1.800E+01 O.OOOE+00 -2.400E+00 -5.600E+00 D 1.450E+00 -3.333E-01 2.000E+00 0.000E+00 2.167E-01 4.333E-01 02(vol.%) Conv. Conv. Selec. Conv. Conv. Selec. 0.50 60.6 8.6 5.8 100.0 14.5 12.0 1.00 79 6 17 0 27.0 100.0 22.3 24.9 L50 92 0 24 9 42 3 100.0 29.2 35.6 2.00 99.0 32 0 53.0 100.0 35.4 44.5 2.50 101 7 38 1 60.8 100.0 41.0 51.9 3.00 101 1 43 0 67.0 100.0 46.3 58.1 3.50 98 3 46 4 73.3 100.0 51.3 63.5 4.00 94 4 48 0 81.0 100.0 56.3 68.3 4.50 90.5 47.6 91.8 100.0 61.4 72.9 5.00 87.8 45.0 107.0 100.0 66.7 77.7 Sum 904.8 350.5 608.8 1000.0 424.5 509.4 Mean(Exp.) 62.14 SSE 4969.97 SSM 26824.91 DOF 26.00 R2 0.81 Std Error 13.83 F 38.11 Fcritical(a = 0.05) 2.55 160 Appendix E Statistical Analysis - Experiment Set 2 Experimental o2 Simulation H 2S S 0 2 o2 H 2S S 0 2 A 8.590E+01 2.800E+01 5.380E+02 -1.700E+00 2.740E+01 4.680E+01 B -7.075E+00 0.000E+00 -1.155E+02 1.468E+01 -5.000E-02 -2.000E+00 C 6.375E-01 0.000E+00 6.250E+00 -6.125E-01 -5.000E-02 -4.441 E-16 D 0.000E+00 0.000E+00 0.000E+00 0.000E+00 0.000E+00 0.000E+00 H2S(vol.%) Conv. Conv. Selec. Conv. Conv. Selec. 5.50 66.3 28.0 91.8 60.5 25.6 35.8 6.00 66.4 28.0 70.0 64.3 25.3 34.8 -6.50 66.8 28.0 51.3 67.8 25.0 33.8 7.00 67.6 28.0 35.8 71.0 24.6 32.8 7.50 68.7 28.0 _ 23.3 73.9 24.2 31.8 8.00 70.1 28.0 14.0 76.5 23 8 30.8 8.50 71.8 28.0 7.8 78.8 23.4 29.8 9.00 73.9 28.0 4.8 80.8 22.9 28.8 76.2 28.0 4.8 82.4 22.4 27.8 10.00 78.9 28.0 8.0 83.8 21.9 26.8 "I Sum 706.7 280.0 311.6 739.8 239.1 313.0 Mean(Exp.) 43.28 SSE 7448.75 SSM 19928.16 DOF 26.00 0.63 Std Error 16.93 F 14.52 Fcriticalio. = 0.05) 2.55 161 Appendix £ Statistical Analysis - Experiment Set 3 Experimental < Simulation o 2 H 2S S 0 2 o 2 H2S S 0 2 A -3.849E+02 -5.498E+02 7.788E+02 1.018E+02 -3.613E+00 -1.670E+02 B 2.580E+00 4.150E+00 -6.200E+00 -1.390E+00 -2.100E-01 1.165E+00 C -3.620E-03 -7.500E-03 1.250E-02 4.370E-03 1.130E-03 -1.750E-03 D 0.000E+00 0.000E+00 0.000E+00 0.000E+00 0.000E+00 0.000E+00 Temp. Conv. Conv. Selec. Conv. Conv. Selec. 245.00 29.9 16.8 10.1 23.5 12.8 13.4 250.00 33.8 19.0 10.0 27.4 14.5 14.9 255.00 37.6 20.8 10.6 31.5 16.3 16.3 260.00 41.2 22.3 11.8 35.8 18.2 17.6 265.00 44.5 23.3 13.6 40.2 20.1 18.9 270.00 47.8 24.0 16.0 45.0 22.1 20.0 275.00 50.8 24.3 19.1 50.0 24.1 21.1 280.00 53.7 24.3 22.8 55.2 26.2 22.0 285.00 56.3 23.8 27.1 60.6 28.3 22.9 290.00 58.8 23.0 32.0 66.2 30.5 23.7 Total Sum 454.3 221.6 172.8 435.4 213.0 190.8 Mean(Exp) 28.29 SSE 653.81 SSM 5983.87 DOF 26.00 0.89 Std Error 5.01 F 70.65 Fcritical(a = 0.05) 2.55 162 

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.831.1-0058761/manifest

Comment

Related Items