UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Hydrogen spillover in a ceramic membrane reactor and packed-bed reactor Izadi-Najafabadi, Negar 2000

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

Item Metadata

Download

Media
831-ubc_2000-0086.pdf [ 5.6MB ]
Metadata
JSON: 831-1.0059003.json
JSON-LD: 831-1.0059003-ld.json
RDF/XML (Pretty): 831-1.0059003-rdf.xml
RDF/JSON: 831-1.0059003-rdf.json
Turtle: 831-1.0059003-turtle.txt
N-Triples: 831-1.0059003-rdf-ntriples.txt
Original Record: 831-1.0059003-source.json
Full Text
831-1.0059003-fulltext.txt
Citation
831-1.0059003.ris

Full Text

HYDROGEN SPILLOVER IN A CERAMIC MEMBRANE REACTOR AND PACKED-BED REACTOR By Negar Izadi-Najafabadi B . Sc. Sharif University of Technology  A THESIS SUBMITTED IN PARTIAL F U L F I L L M E N T OF T H E REQUIREMENTS FOR T H E D E G R E E O F M A S T E R OF A P P L I E D S C I E N C E  in T H E F A C U L T Y OF G R A D U A T E STUDIES D E P A R T M E N T OF CHEMICAL A N D B I O - R E S O U R C E ENGINEERING  We accept this thesis as conforming to the required standard  T H E UNIVERSITY OF BRITISH COLUMBIA  November 1999 © Negar Izadi-Najafabadi, 1999  In  presenting  degree  this  at the  thesis  in  University of  partial  fulfilment  British Columbia,  of  the  requirements  for  that permission  copying  granted  department  this or  thesis by  publication of this  for scholarly  his thesis  or  her  Department The University of British Columbia Vancouver, Canada  DE-6 (2/88)  MyJ  7  may be  representatives.  It  for financial gain shall not  permission.  Date  purposes  1999  advanced  I agree that the Library shall make it  freely available for reference and study. I further agree of  an  is  for extensive  by the head  understood  that  be allowed without  of  my  copying  or  my written  Abstract  The term spillover in heterogeneous catalysis is applied to the transport of active species from one surface to another, in which the second surface does not under the same conditions, sorb or form the active species. Since the first observation of spillover in the 1960's, there has been extensive research in the area of spillover, especially hydrogen spillover. Catalytic reactions such as hydrocracking are explained based on the spillover phenomenon. The role of the catalyst in hydrocracking is thought to be to provide spilt— over hydrogen species. Consequently, there is no need for direct contact between organic reactant(s) and catalyst, provided spilt-over hydrogen can have access to reactant by some means. In conventional hydrocracking systems, the organic reactant(s) and catalyst are in direct contact and this leads to coking and catalyst deactivation. Catalyst deactivation is a major concern in industry and different researchers are focusing on understanding and controlling coking and developing new coke-resistant catalysts. Based on the spillover phenomenon, one possible approach to reduce the impact of coking and catalyst deactivation is to separate the catalyst and organic reactant by a medium which can facilitate hydrogen spillover. A ceramic membrane reactor may be one possible configuration for separation of reactant and catalyst. The membrane tube would act as the separating medium while the ceramic material may facilitate hydrogen spillover. Furthermore, the ceramic material is resistant to the severe conditions present in hydrocracking reactions. The present study reports on an attempt to implement and assess this new approach of hydrocracking in a membrane reactor. In the first part of the work, diphenylmetliane(DPM) hydrocracking was investigated in a ceramic membrane reactor. The. catalyst (sulfided N i - M o / A ^ O a ) was separated from the D P M by means of a ceramic membrane tube. The results of different experiments showed the enhancement of the yields of products (benzene and toluene) in the presence of catalyst compared to the yields in the absence of catalyst, that was ascribed to hydrogen spillover. However, the transport of liquid D P M through the membrane and onto the 11  catalyst could not be eliminated and therefore, the extent of hydrogen spillover could not be quantified. To pursue the investigation of hydrogen spillover through the ceramic material of the membrane, and to achieve complete separation of catalyst and reactant, a catalytic reaction with solid reactant was also investigated. Coke was chosen as the solid reactant because it reacts with spilt-over hydrogen to produce methane. Hence for the second part of this work, amorphous silica-alumina was coked by thermal decomposition of propylene. Subsequently, the coked silica-alumina underwent temperature-programmed hydrogenation in the presence and absence of catalyst (2%Co-Si02). In the former case, the effect of separation of coked silica-alumina and catalyst by low surface area crushed ceramic material (the ceramic membrane used in the first part of the study) and high surface area dried silica-alumina, were investigated. The results confirmed that hydrogen spillover was responsible for at least part of the methane production. Also, hydrogen spillover decreased due to the separation of coked silica-alumina and catalyst by either the ceramic or the dried silica-alumina. Hydrogen transport by spillover through the ceramic membrane was confirmed, however, the extent of spillover was lower for the low surface area ceramic material than the high surface area dried silica-alumina.  in  Table of Contents  Abstract  *  u  List of Tables  «  List of Figures  x  Acknowledgement 1  u  xvi  Introduction  1  1.1  Motivation  1  1.2  Objectives  2  1.3  Approach  2  2 Literature Review 2.1  2.2  4  Spillover  4  2.1.1  Background  4  2.1.2  Spillover Distances  5  2.1.3  Spillover Rates  5  2.1.4  Methods for the Observation of Hydrogen Spillover  6  2.1.5  Nature of Spilt-over Hydrogen  8  2.1.6  Spillover in Applied Catalysis  9  Hydrocracking  10  2.2.1  Background  10  2.2.2  Classical Model of Hydroconversion  10  iv  2.2.3 2.3  2.4  2.5 3  Extension of the Classical Model of Hydroconversion  Coke  11 11  2.3.1  Background  11  2.3.2  Chemical and Physical characterization of Coke Deposits  13  2.3.3  Kinetics of Coke Formation  13  2.3.4  Spillover and Coking  14  Membrane Reactors  15  2.4.1  Background  15  2.4.2  Membrane Structure and Shape  15  2.4.3  Membrane Reactor Configurations  16  2.4.4  Coupling Catalyst and Membranes  16  2.4.5  Mechanism of Transport Through Inorganic Membranes  18  2.4.6  Major Applications of Inorganic Membrane Reactors  19  Summary of Literature Review  21  Hydro cracking Diphenylmethane in a Ceramic Membrane Reactor  23  3.1  Ceramic Membrane Reactor  24  3.2  Experimental Methods  27  3.2.1  Catalyst Sulfidation  27  3.2.2  Hydrocracking Reaction  28  3.2.3  Analysis of the Products  28  3.2.4  Safety Considerations of the Membrane Reactor  28  3.3  Results and Discussion of Hydrocracking Diphenylmethane in a Ceramic Membrane Reactor  29  3.3.1  29  Comparison of Thermal and Catalytic Experiments  v  3.4  3.3.2  Effect of Catalyst Immersion in D P M  33  3.3.3  Effect of Membrane Pore Size  35  3.3.4  Effect of Catalyst Location  36  Conclusions from Hydrocracking Diphenylmethane in a Ceramic Membrane Reactor  38  Effect of Hydrogen Spillover on Temperature-Programmed Hydrogenation of Coked Silica-Alumina  39  4.1  Apparatus  40  4.1.1  Drying and Coking Apparatus  40  4.1.2  Temperature-Programmed Reaction ( T P R ) Apparatus  41  4.1.3  Reactor  41  Experimental Methods  44  4.2.1  Drying and Coking  44  4.2.2  Catalyst Reduction  45  4.2.3  Temperature-Programmed Hydrogenation (TPH)  45  4.2.4  Temperature-Programmed Reaction (TPR) in Argon  46  4.2.5  Calibration  46  4.2  4.3  Results and Discussion of Temperature-Programmed Hydrogenation of Coked Silica-Alumina 4.3.1  46  Comparison of Non-catalytic (Thermal) and Catalytic TemperatureProgrammed Hydrogenation of Coked Silica-Alumina  4.3.2  Effect of Position of Catalyst and Coked Silica-Alumina on TemperatureProgrammed Hydrogenation  4.3.3  47  52  Effect of Hydrogen Spillover on Temperature-Programmed Hydrogenation of Coked Silica-Alumina  vi  52  4.3.4  4.4  Effect of Ceramic and Silica-Alumina on Temperature-Programmed Hydrogenation of Coked Silica-Alumina  57  4.3.5  Carbon Recovery  58  4.3.6  Test for Plug Flow  63  Conclusion of Effect of Hydrogen Spillover on Temperature-Programmed Hydrogenation of Coked Silica-Alumina  5 Conclusions & Recommendations for Future Work  65  68  5.1  Conclusions  68  5.2  Future Work  69  Bibliography  71  A Calibration of Gas Chromatograph  76  B Gas Chromatography of Feed (Diphenylmethane)  81  C Calculation of Amount of Dimethyl Disulfide(DMDS) for Soaking NiMo/Al 0 2  83  3  D Summary of Experiments in Hydrocracking Diphenylmethane in a Ceramic Membrane Reactor  84  D.l  84  Operating Conditions  D.2 Definition of Equations  84  D.3 Summary of Hydrocracking Experiments  84  E Calibration of Flow Controllers  93  F Calibration of Mass Spectrometer  103  G Thermogravimetric Analysis of Silica-Alumina  108  vii  H Calculation for Carbon Recovery I  109  Summary of Results for Temperature-Programmed Reaction of Coked Silica-Alumina  J  110  Summary of Temperature-Programmed Hydrogenation Profiles of Coked Silica—Alumina  122  K Summary of Temperature-Programmed Reaction Profiles of Coked Silica-  L  Alumina in Argon  139  Test for Plug Flow  143  L.l  143  Analysis of Tests for Plug Flow  M Surface Area Measurements  147  Vlll  List of Tables  3.1  Hydrocracking Diphenylmethane in a Membrane Reactor Compared to Thermal Conversion  3.2  31  Summary of Components Identified from Liquid Product Samples Generated at Reaction Conditions in Table 3.1  3.3  Hydrocracking Diphenylmethane in a Membrane Reactor with Catalyst on Shell-side and Tube-side  3.4  33  Summary of Components Identified froni Liquid Product Samples Generated at Reaction Conditions in Table 3.3  3.5  34  Hydrocracking Diphenylmethane in a Membrane Reactor with Pore Size  100A 3.6  32  and  1000A;  Catalyst on Tube-side and D P M on Shell-Side  37  Hydrocracking Diphenylmethane in a Membrane Reactor with Catalyst on Tube-side Placed at Different Levels; D P M on Shell-Side  37  4.1  Position of Valves for Different Experiments  43  4.2  Comparison of Products of Non-Catalytic (Thermal) and Catalytic Experiments  50  4.3  Effect of Coked Silica-Alumina and Catalyst Position on Methane Production 53  4.4  Effect of Hydrogen Spillover on Hydrogenation of Coked Silica-Alumina . . 56  4.5  Effect of Ceramic and Dried Silica-Alumina on Methane Production of Coked Silica-Alumina  61  4.6  Carbon Recovery in Different Experiments  62  4.7  Effect of Gap on Methane Production of Coked Silica-Alumina  64  A.l  Gas Chromatograph Standard Calibration Mixtures and Retention Time of Different Compounds  77 ix  D.l  Summary of Products in Catalytic Hydrocracking with Catalyst on TubeSide and D P M on Shell-Side  D.2 Summary of Products in Thermal Cracking  85 86  D.3 Summary of Products in Catalytic Hydrocracking with Catalyst on ShellSide and D P M on Shell-Side  87  D.4 Summary of Products in Catalytic Hydrocracking with Catalyst on ShellSide and D P M on Shell-Side  88  D.5 Summary of Products in Catalytic Hydrocracking with Catalyst on TubeSide and D P M on Shell-Side  89  D.6 Summary of Products in Catalytic. Hydrocracking with Catalyst on TubeSide and D P M on Shell-Side  90  D.7 Summary of Products in Catalytic Hydrocracking with Catalyst on TubeSide and D P M on Shell-Side  91  D. 8 Summary of Products in Catalytic Hydrocracking with Catalyst Elevated on Tube-Side and D P M on Shell-Side  92  E. l  Calibration of Mass Flow Controller for Argon  94  E.2  Calibration of Mass Flow Controller for Hydrogen  97  E.3 Calibration of Rotameter for Helium  99  E . 4 Calibration of Rotameter for Propylene  101  F. l  Intensity of Different Gases in Mass Spectrometer  104  1.1  Mass Balance  Ill  1.2  Summary of Results for Methane Production  113  1.3  Summary of Results for Propylene Production  115  1.4  Summary of Results for Ethane Production  117  1.5  Summary of Results for Ethylene Production  119  1.6  Summary of Medium of Separation for T P H of Coked Silica-Alumina . . . 121  L.I  Test for Plug Flow  144  xi  List of Figures  2.1  New Mechanism for Hydrocracking  12  2.2  Hydrogenation of Coke due to Hydrogen Spillover  14  2.3  Membrane Reactor Configurations  17  3.1  Process Flow Diagram of Ceramic Membrane Reactor  24  3.2  Hydrogen Flow in Ceramic Membrane Reactor  25  3.3  Detail of Membrane Holder for Reactor, Top Holder  27  3.4  Effect of Membrane Pore Size on Pressure Drop Required to Displace Liquid From Pores  36  4.1  Drying and Coking Apparatus  40  4.2  Process Flow Diagram  42  4.3  Reactor  44  4.4  Temperature-Programmed Hydrogenation of Coked Silica-Alumina . . . .  48  4.5  Temperature-Programmed Hydrogenation of Coked Silica-Alumina in the Presence of Catalyst (2%Co-Si0'2), Catalyst Placed on Top of the Coked Silica-Alumina  49  4.6  Comparison of Methane in Thermal and Catalytic Experiments  51  4.7  Effect of Position of Coked Silica-Alumina and Catalyst on Production  4.8  Methane Production from Coked Silica-alumina After T P R in Argon . . .  4.9  Methane Production from Coked Silica-Alumina in the Presence of Cata-  . . 54  lyst; After T P R in Argon; Catalyst and Coked Silica-Alumina were mixed 4.10 Effect of Ceramic on Methane Production of Coked Silica-Alumina  xn  . . . .  55  57 59  4.11 Effect of Dried Silica-Alumina on Methane Production of Coked SilicaAlumina  60  4.12 Possible Flow Direction in the Reactor Bed  65  A.l  Gas Chromatograph Calibration Graph for Benzene  78  A.2 Gas Chromatograph Calibration Graph for Toluene  79  A. 3 List of Program for Anaylsis of the Samples and Results of Gas Chromatography on Calibration Mixture #18  80  B. l  Results of Gas Chromatography on Feed ( D P M )  82  E.l  Calibration of Mass Flow Controller for Argon  95  E.2  Calibration of Mass Flow Controller for Argon  96  E.3 Calibration of Mass Flow Controller for Hydrogen  98  E.4 Calibration of Rotameter for Helium  100  E. 5 Calibration of Rotameter for Propylene  102  F. l  Calibration for Methane in Hydrogen and Argon  104  F.2  Calibration for Propylene in Hydrogen and Argon  105  F.3  Calibration for Methane in Argon  106  F. 4 Calibration for Propylene in Argon  107  G. l Thermogravimetric Analysis of Silica-Alumina  108  J.l  R97-2  122  J.2  R98-2  123  J.3  R99-2  124  J.4  R100-2  125  J.5  R102-2  126 xiii  J.6  R103-2  127  J.7  R104-2  128  J.8  R105-2  129  J.9  R109-2  130  J.10 R110-2  131  J.11 R l l l - 2  132  J.12 R112-2  133  J.13 R113-2  134  J.14 R115-2  135  .1.15 R117-2  136  J.16 R118-2  137  J.17 R119-2  138  K.l  R'104-2  139  K.2 R'105-2  140  K.3 R'112-2  141  K.4 R'118-2  142  L.l  E Curve for Inside Reactor Bed  145  L.2  E Curve for Tubings; Reactor was bypassed  146  M . l Summary of Analysis Options for Commercial N i - M o / A ^ O a Catalyst . . . 148 M.2 Isotherm Plot for Commercial N i - M o / A l 0 Catalyst 2  149  3  M.3 Incremental Pore Volume for Commercial N i - M o / A l 0 Catalyst  150  M.4 Summary Report for Commercial N i - M o / A l 0 Catalyst  151  M.5 Summary of Analysis Options for Silica-Alumina  152  M.6 Isotherm Plot for Silica-Alumina  153  2  2  xiv  3  3  M.7 Incremental Pore Volume for Silica-Alumina  154  M.8 Summary Report for Silica-Alumina  155  M.9 Summary of Analysis Options for Coked Silica-Alumina  156  M.10 Isotherm Plot for Coked Silica-Alumina  157  M . 11 Incremental Pore Volume for Coked Silica-Alumina  158  M . 12 Summary Report for Coked Silica-Alumina  159  xv  Acknowledgement  I would like to tliank my supervisor Dr. Kevin J . Smith for his support, guidance and encouragement throughout my research. I also would like to thank Dr. G . Chattopadhyaya for his instruction and assistance at the initial stage of my work. Special thanks must go to my colleague Dr. J . S. Soltan Mohammad Zadeh for his assistance and help during my studies. I would like to thank Dr. D. Posarac, the staff of Mechanical Workshop, Electrical Shop, Stores, Main Office of Chemical Engineering, the staff of Main Library of U B C and my colleagues: M . Simard, M . Chong Ping, P. Knezevich, J . Sawada, Dr. Q. Liu, N . Rezai, S. Kovacevic and A . Jamal for their assistance and support. The financial support of Syncrude Canada Limited and N S E R C are gratefully acknowledged. Finally, I would like to dedicate this thesis to my family for their love and support.  xvi  Chapter 1 Introduction  1.1  Motivation  The term spillover in heterogeneous catalysis is applied to the transport of active species from one surface to another, in which the second surface does not, under the same conditions, sorb or form the active species (Conner Jr. et al. (1986)). The first direct evidence for spillover was apparently obtained by Khoobiar (1964) during his investigation of the reduction of W O 3 on 0.5%Pt/Ai2O3. WO3 is known to be reduced by hydrogen above 200°C. Reduction of W O 3 leads to the formation of blue oxide, W4O11. In the presence of hydrogen, a mechanical mixture of 0.5%Pt/Ai2O and W O 3 , results in a color change at 3  room temperature, whereas for a mechanical mixture of A 1 0 and W O 3 , no color change 2  3  occurs. These observations were explained by virtue of hydrogen spillover. In other words, hydrogen adsorbed and dissociated on platinum, migrated via  AI2O3  onto W O 3 to take  part in the reduction. The migration from Pt to A I 2 O 3 is "spillover". Since the 1960's, the spillover phenomenon has been investigated in different catalytic systems especially in hydrogen atmospheres. These investigations have shown that in catalytic reactions such as hydroconversion and coke hydrogenation/oxidation, the spillover phenomenon occurs. The classical mechanistic model of hydroconversion reactions (e.g. hydrocracking) on bifunctional catalysts assumes dehydrogenation of a saturated hydrocarbon on the metal site of the catalyst, formation of a carbenium ion on the acidic site of the support and hydrogenation on the metal site. The transport between metal and support sites of the catalyst was postulated to occur via gas phase diffusion. The classical model was greatly accepted in literature. However, as will be discussed in Section 2 of Chapter 2, new results from hydroconversion experiments on layered catalysts, (e.g. P t / A ^ O s and H-erionite) can not be explained by the classical model. Hence, according to experimental results of different researchers, a new model was developed based on the hydrogen spillover 1  2  Introduction  phenomenon. The new model of hydroconversion assumes hydrogen dissociation on a metal site of the catalyst followed by spillover onto the support. The spilt-over hydrogen reacts with the organic reactant. Consequently, there is no need for direct contact between the organic reactant and catalyst, provided the spilt-over species can have access to the organic reactant. Direct contact between organic reactant and catalyst leads to coking and catalyst deactivation which is a major problem in industry. Coking can be partly avoided by hydrogen/oxygen spillover since spillover can play an important role in cleaning the surface of coked catalysts by removing and controlling byproduct coke. In other words, hydrogen/oxygen are activated on the metal sites of the catalyst and spillover onto the support to react with coke. The product of this reaction is C H or CO2. However, spillover cannot 4  completely prevent coking reactions due to the direct contact of organic reactant(s) and catalyst. The present study is based on the proposal that catalyst coking during hydrocracking can be avoided if the catalyst and organic reactant are separated by a membrane that facilitates the transport of hydrogen by spillover. A ceramic membrane reactor should provide the desired reactor configuration since separation of reactant and catalyst can be achieved by the ceramic membrane while the ceramic material may facilitate hydrogen spillover, needed for reaction.  1.2  Objectives  The objective of the present study was to demonstrate the transport of spilt-over hydrogen (produced on a suitable catalyst) through a ceramic membrane, and to show that these spilt-over species were reactive toward an organic reactant.  1.3  Approach  Two different experimental procedures were followed to meet the project objectives: • Hydrocracking Diphenylmethane(DPM) in a Ceramic Membrane Reactor: Unlike conventional hydrocracking systems, the model compound ( D P M ) was separated  Introduction  3  from the catalyst by a ceramic membrane tube. Results obtained in this configuration were compared with conventional hydrocracking. As to be discussed in Chapter 3, the membrane reactor configuration could not completely eliminate the contact between catalyst and liquid model compound. Hence the next catalytic system chosen for this study involved a solid reactant (coke) to ensure complete separation of reactant and catalyst by the ceramic. • Temperature-Programmed Hydrogenation of Coked Silica-Alumina: As to be discussed in Chapter 4, in this system the effect of hydrogen spillover on hydrogenation of coked silica-alumina was investigated in the presence and absence of a catalyst. In the former case, the effect of separation of coked silica-alumina from catalyst by a layer of crushed ceramic (used in the first part of the study) was investigated.  Chapter 2 Literature Review  The present thesis covers a wide range of different topics including hydrogen spillover, hydrocracking, coke formation and hydrogenation, and performance and operation of a ceramic membrane reactor. To provide some background information to the study, aspects of these topics that are relevant to the present work are briefly reviewed in the following sections.  2.1  Spillover  The spillover phenomenon was observed for the first time in the 1960's. Since then, many studies have been carried out to investigate spillover in different catalytic systems. However, due to the presence of hydrogen in many important catalytic reactions, hydrogen spillover was the focus of many researchers. In this section, the phenomenon of spillover, as it occurs in catalytic systems in general, is described and this is followed by a detailed discussion of hydrogen spillover.  2.1.1  Background  The phenomenon of spillover was defined at the First International Congress on spillover of adsorbed species: "Spillover involves the transport of an active species sorbed or formed on a first phase onto another phase that does not under the same condition sorb or form the species." The following comment was also added: "The result may be the reaction of these species on the second phase with other sorbing gases and/or reaction with, and/or activation of the second phase."(Conner Jr. et al. (1986)). The first phase generating the active species is called the "initiator" or "activator" whereas the phase providing sites for the adsorption of active species is called the "acceptor". If the activator and acceptor are 4  Literature Review  5  in direct contact, primary spillover occurs but if there is an inert medium between them, secondary spillover occurs (Rozanov and Krylov (1997)). The spillover phenomenon is common in catalysis and different chemical species can show this kind of behavior, i.e. the atoms of hydrogen and its isotopes, oxygen, nitrogen, molecular fragments of C O may participate in spillover (Rozanov and Krylov (1997); Conner Jr. and Falconer (1995)). Spillover can take place from a metal to an oxide, from one metal to another metal (Peden and Goodman (1986)), from one oxide to another oxide (Delmon (1993)), from a metal oxide onto a metal (Kieken and Boudart (1992)), from metal sulfides to another metal sulfide (Chu and Schmidt (1993); Karroua et al. (1993)), carbon (Rodriguez and Baker (1993)) , and silica-alumina (Peden and Goodman (1986)).  2.1.2  Spillover Distances  The fact that split-over species can be transferred over long distances has not been universally accepted. The range of reported spillover distances varies from nanometers to centimeters (Lenz and Conner Jr. (1987)). The study of Lenz and Conner Jr. (1987) is a good example of spillover over large distances. In their work, spillover of hydrogen from P t / A l 0 3 to S i 0 was monitored by N M R and spilt-over hydrogen had to diffuse 12 cm 2  2  to reach the N M R probe. However, it was concluded that the hydroxyl group of Si02 facilitated transport of spilt-over hydrogen over large distances.  2.1.3  Spillover Rates  Many studies have shown that spillover does not occur by a single mechanism involving a single species. Different initiator and acceptor surfaces can result in substantially different spillover rates. The rates of the various processes involved in spillover can differ by orders of magnitude depending on the system and conditions. Importantly, spillover can be rapid. Many studies, have shown that hydrogen spillover can occur at room temperature and that split-over hydrogen atoms can be involved in reaction at room temperature (Conner Jr. and Falconer (1995)). One good example is the study by Ioannides and  6  Literature Review  Verykios (1993), of benzene hydrogenation at room temperature. In their study, hydrogen spilled over from R h to  AI2O3,  where it reacted with adsorbed benzene. A l l benzene was  hydrogenated after 15 minutes exposure to hydrogen at 298K.  2.1.4  Methods for the Observation of Hydrogen Spillover  According to Rozanov and Krylov (1997), there are seven methods for the observation of hydrogen spillover. These methods are based on the change observed in the prescence of spillover and should be applied with caution since the observations might be due to other effects rather than hydrogen spillover. Some of these methods are generally used for characterizing catalysts but as discussed herein, they can be used to prove the presence of hydrogen spillover as well. • Chemisorption of Hydrogen:  Chemisorption of hydrogen atoms on a metal  surface is a technique for metal surface area measurements but in certain cases, it may be applied for hydrogen spillover observation. Possible applications of this technique for the observation of hydrogen spillover are summarized as follows: — On supported metal catalysts, the results of hydrogen chemisorption are compared with other techniques such as adsorption of C O , broadening of X-ray diffractograms, electron microscopy, etc. If the results are similar except for hydrogen adsorption, it can be concluded that hydrogen spillover has occurred. — After oxygen adsorption, the surface is exposed to hydrogen and water will be produced which sorbs and remains on the carrier. In the presence of hydrogen spillover, the moles of oxygen remains constant but the molar ratio of hydrogen to oxygen exceeds 2. This method is called "oxygen-hydrogen titration". However, the possible effects of oxygen spillover and presence of water should be considered. — Gravimetric study of the adsorption of hydrogen and deuterium on the support and metal-support catalyst may reveal hydrogen spillover. The amount of hydrogen and deuterium adsorbed on the metal-support is much higher in the  Literature Review  <  case of hydrogen spillover than the amount capable of being adsorbed on the metal-support in the absence of hydrogen spillover (Roland et al. (1994)). • Reduction of Oxides: Spillover may affect the reduction of a reducible oxide (e.g. WO3, Mo03,CuO) when the oxide is mixed with a supported or unsupported metal. To study the reduction of oxides and the effect of hydrogen spillover, different methods such as gravimetric, differential thermal analysis, temperature-programmed reduction and microcalorimetric methods are employed. Due to spillover, the degree and the rate of reduction may increase or the reduction temperature may fall. Hence, reduction of oxides may be used as a confirmation of spillover. However, it is not easy to estimate the effect of spillover quantitatively due to following reasons: — The apparent activation energy for the reduction of oxides under spillover conditions does not differ from the noncatalytic reduction. Spillover increases the pre-exponential factor, — The degree of reduction depends on the extent of the interface between the metal catalyst and the oxide being reduced, — Water is a product which can either accelerate or retard both spillover and reduction. Hence, the effect of hydrogen spillover on reduction of oxides may be under/overestimated by other effects, explained above. • The influence of Spillover on Catalysis: Enhanced activity and selectivity are frequently observed in multiphase catalysts which can be due to spillover. There are a number of examples as explained by Rozanov and Krylov (1997) and the references in that paper. The most relevant example for the present study, is the hydrocracking of diphenylmethane reported by Stumbo et al. (1995). They observed that the rate of hydrocracking of diphenylmethane increased after the addition of Co-Mo/Si02 to silica-alumina by virtue of hydrogen spillover. However, they used isotopic methods (explained below) to provide more understanding about hydrogen spillover in the system.  Literature Review  6"  • Isotope Methods: One of the most common methods for the observation of hydrogen spillover is isotope exchange of deuterium or tritium with the - O H groups of the carrier. For the observation of isotope exchange, different techniques such as F T I R spectroscopy are utilized. • The Scavenger Method: Addition of substances which interact quantitatively with hydrogen (like perylene, anthracite) is another approach to identifying the presence of spillover. For detection, sensitive methods like electron paramagnetic resonance are used. • Temperature—Programmed Methods: Generally, in temperature-programmed techniques, desorption or production of species are monitored by different instruments (e.g. mass spectrometer) as the temperature increases linearly in time. B y plotting the concentrations of species versus time and temperature, it is possible to distinguish between different processes occurring on the surface at different rates. The area under the concentration versus time curve, represents total desorption or production and may be a useful information for comparison between different catalysts. The effect of hydrogen spillover on supported catalysts can be studied and compared with non-supported catalysts by T P D - T P R methods. On supported catalysts, due to hydrogen spillover, the amount of desorbed or produced species may increase or a new peak may appear in comparison with non-supported catalyst and the absence of hydrogen spillover. • Reverse Hydrogen Spillover: This term is frequently used to designate the reverse migration of hydrogen from the support to the metal. Hydrogen atoms from a hydrocarbon are adsorbed on the support. In the presence of a metal, hydrogen atoms are transferred to the metal to desorb as molecules. In this way, the process accelerates.  2.1.5  Nature of Spilt-over Hydrogen  The chemical nature of spilt-over hydrogen has been studied and a number of different species have been proposed: H atoms (Lenz et al. (1989)), H  +  ions (Levy and Boudart  9  Literature Review  (1974); Khoobiar et al. (1968)), ion pairs and H 3 species (Conner Jr. et al. (1986); Conner Jr. (1988)). Depending on the type of reaction and applied spectroscopic method, H atoms and H ions have been found in the same system under similar conditions (Roessner +  and Roland (1996); Roland et al. (1997)). Most recently, a new model proposed by Roland et al. (1997) assumes the coexistence of H atoms and H ions. This model explains that the physical nature of the spilt-over +  species, especially their charge, can only be described by considering their interaction with the solid. In other words, spilt-over hydrogen behaves as an electron donor which is located on the surface. Hence H ions and H atoms coexist on the surface of the catalyst +  but their ratio is determined by the electronic properties of adsorbate/solid system, i.e. electronic states, Fermi energy, concentration of the spilt-over species, density of the electronic states of the solid and temperature (Roessner and Roland (1996); Roland et al. (1997)). 2.1.6  Spillover in Applied Catalysis  Many important catalytic reactions involve spillover: hydrogenations, dehydrogenations, partial and total oxidations, hydrocarbon isomerization (Roessner and Roland (1996)), hydrocracking (Stumbo et al. (1995); Stumbo et al. (1997)), hydrodesulfurization, hydrodenitrogenation, gasification and hydrocarbon synthesis from syngas (Conner Jr. and Falconer (1995)). Spillover has also been implicated in the creation, maintenance, and regeneration of catalytic activity, including the removal or control of coke formation (Traffano and Parera (1986); Parera et al. (1983)). Although spillover is common in catalysis it is important to acknowledge that it is only one mechanistic step in a sequence of reactions(Conner Jr. and Falconer (1995)). In the proceeding sections, hydroconversion with an emphasis on hydrocracking, coking and the effect of hydrogen spillover for each will be discussed in detail.  Literature Review 2.2  10  Hydrocracking  2.2.1  Background  Hydroconversion reactions refer to a number of different catalytic reactions that occur in the presence of hydrogen. Hydrocracking refers to catalytic cracking of organic molecules in the presence of hydrogen under substantial pressure. Hydrocracking processes were developed in Germany and in England in the 1930's to supply aviation gasoline for military purposes. The catalyst consisted of various metal sulfides such as nickel or tungsten supported on HF-treated montmorillonite or silica-alumina. High pressures were required with these catalyst (e.g. over 20MPa), and the processes were uneconomic (Satterfield (1980)). Since then, different bifunctional catalysts (e.g. noble/non-noble metal on different zeolite supports) were developed in order to reduce the severity of the systems under hydrocracking. The balance between the metal sites and acidic support of bifunctional catalysts may be varied in order to process different feed stocks.  2.2.2  Classical Model of Hydroconversion  The classical model of hydroconversion which is widely accepted in the literature, was proposed by Mills et al. (1953) and by Weisz and Swegler (1957). They described the reaction mechanism by the independent action of different, physically distinct catalyst sites.  According to their model, hydroconversion of a hydrocarbon on a bifunctional  catalyst, occurs via the following steps : • Dehydrogenation of a saturated hydrocarbon on the metal site and production of an olefin • Formation of a carbenium ion on the acidic sites of the support • Hydrogenation on the metal site The transport between metal and support sites are postulated to occur via gas phase diffusion.  Based on their model, the only role of hydrogen is to prevent coking and  deactivation of the catalyst.  11  Literature Review  Roessner and Roland (1996) studied isomerization of n-hexane on a layered catalyst (Pt/Al2C«3 and H-erionite). They reported that activity and selectivity strongly depend on the existence of direct contact between the metal and support. This could not be explained by the classical model and gas phase diffusion of intermediates. Also, they found that the presence of hydrogen was essential even in the early stages of reaction, before coke formation on the catalyst. 2.2.3  Extension of the Classical Model of Hydroconversion  Based on the results on layered catalyst, H-D exchange and FTIR, Stumbo et al. (1995), Roessner and Roland (1996) and Stumbo et al. (1997) proposed a new model for hydroconversion which is based on hydrogen spillover: Hydrogen dissociates on the surface of the metal/metal sulfide and spills over onto the support where it creates Bronsted sites ( H ) . As a result a carbenium ion from the organic reactant will be formed. The +  carbenium ion reacts with other spilt-over hydrogen species. Figure 2.1 shows the steps in hydrocracking of diphenylmethane on the mechanical mixture of sulfided CoMo/Si02 and silica-alumina.  2.3  Coke  2.3.1  Background  Coking refers to the formation of carbonaceous deposits of higher hydrocarbons and coke itself (Baumgarten and Schuck (1997)). It occurs by carbon disproportionation, condensation and hydrogen abstraction reactions of adsorbed carbon-containing species such as hydrocarbons (Wolf and Alfani (1982)). Coke may contain high molecular-weight, polyaromatic molecules and considerable amounts of hydrogen. A representative empirical formula for coke is CHi and  CH .5 0  (Satterfield (1980)). Coking reactions are undesirable  side reactions which occur in many catalytic processes, leading to catalyst deactivation and many technical and economic problems. Understanding the chemistry of coke formation and its effect on catalyst deactivation provides insights into controlling the coking reactions. However, the kinetics and  Literature Review  H  2  Sulfided CoMo/ S i 0 «  2  2H*  Silica* — A l u m i n a • +. u  Figure 2.1: New Mechanism for Hydrocracking  13  Literature Review  mechanism of coking are rather complicated and depend on the type of reaction, feed composition, type of catalyst used and reactor environment. Hence, it is not possible to make general interpretations of the coking mechanism.  2.3.2  Chemical and Physical characterization of Coke Deposits  There are several experimental methods available to characterize the composition and structure of coke deposits.  Analysis of coke combustion products to determine C / H  ratio is the simplest method of characterization. C / H ratio is useful in the comparison of different types of coke and also to determine the effect of time on stream or coke aging. To identify different kinds of coke molecules, other techniques such as solvent extraction and pyrolytic mass spectrometry are employed. For detailed structural information of coke, different methods such as X-ray analysis, IR spectroscopy, T G A , electron microscopy and Auger electron spectroscopy are employed. Different coke characterization methods have indicated that coke deposits are complex structures and consist of monocyclic and polycyclic aromatic rings connected by aliphatic and alicyclic fragments. The relative proportion of each group depends on the reactioncatalyst system. The groups are interconnected, forming a crystalline pseudographitic structure and an amorphous phase. However, the degree of crystallinity and cross-linking varies with feed, catalyst and reaction conditions (Wolf and Alfani (1982)).  2.3.3  Kinetics of Coke Formation  For many catalytic systems, the weight percent of carbon deposited as the result of coking follows a simple relation: C  c  =  At"  where C is the weight percent of carbon deposited on the surface, t is the time on stream, c  and A and n are correlation constants that depend on feedstock, reaction, catalyst and operating conditions. This equation does not have any fundamental mechanistic basis and represents a simple two-parameter fit to the data valid in a narrow range of operating  14  Literature Review  conditions. However due to its simplicity and applicability to many different systems it is used in many kinetic studies of coke (Wolf and Alfani (1982)).  2.3.4  Spillover and Coking  It is known that hydrogen and oxygen spillover can play an important role for cleaning the surface of coked catalysts by removing and controlling byproduct coke. Hydrogen or oxygen from the metal spills over to the support where it reacts with the surface carbon. As a result, C H 4 or CO2 will be produced (Conner Jr. and Falconer (1995); Parera et al. (1983); Baumgarten and Schuck (1997)). Figure 2.2 shows the reaction of spilt-over hydrogen with coke deposited on the support of a metal supported catalyst. Parera et al. (1983) investigated the effect of hydrogen spillover in hydrogenation of coked alumina. In their work, alumina was coked by naphtha or methyl cyclopentane and underwent hydrogenation after addition of  AI2O3  or P t / A ^ O a . The hydrogenation results showed  that, due to hydrogen spillover, elimination of coke from coked alumina in the presence of Pt is higher than in the absence of Pt.  CH  Coke  4  Metal  Support  Figure 2.2: Hydrogenation of Coke due to Hydrogen Spillover  Literature Review 2.4  15  Membrane Reactors  2.4.1  Background  A membrane reactor is a particular type of multifunctional reactor where one or more chemical reactions, generally catalytically promoted, are carried out in the presence of a membrane. The membrane's permselectivity affects reactions, allowing improvements of either the achievable conversion (e.g. equilibrium reactions) or the selectivity toward intermediate products (e.g. consecutive reactions). During the 1960's, polymer membranes were popular but due to their modest temperature resistance, only low-temperature membrane reactors were studied, mostly in the biotechnology field. In the middle 1980's, inorganic membranes, both metallic and ceramic, were produced. This new generation of membranes could be applied to different catalytic processes due to their capability of withstanding high temperatures (200-600° C). Hence, application of membrane reactors for new processes in the chemical, petrochemical, pharmaceutical and environmental fields were proposed. However, presently there is no successful industrial application of inorganic-membrane reactors. This is due to several issues including membrane instability, insufficient permeability, insufficient permselectivity and the high cost of inorganic membranes. For ceramic membrane reactors, another issue is the difficulty in obtaining an effective seal between the metal body of the reactor and the ceramic membrane due to the different thermal expansion coefficients of these materials (Saracco and Specchia (1998)).  2.4.2  Membrane Structure and Shape  Inorganic membranes can be divided into two main categories: unsupported (symmetric) and supported (asymmetric). Symmetric membranes were produced first, and Vycor glass or solid-electrolyte are examples of this type of membrane. In general, the asymmetric structure is preferred due to proper balance among membrane permselectivity, permeability (the lower the permeability, the lower is the transmembrane flux at a given pressure difference) and mechanical strength(Hsieh (1991)).  16  Literature Review  The structure of supported membranes is generally made of two or three supporting porous layers plus a permselective top layer. The supporting layers possess a decreasing average pore size as the permselective layer is approached. This is done to minimize the overall pressure drop, with the obvious constraint that the permselective layers can not be supported directly on large pore-size supports; otherwise, massive formation of defects such as cracks or pinholes would take place(Bhave (1991)). Membrane geometry may be flat or tubular. The advantage of flat membranes is that, they can be stacked onto one another by interposition of corrugated plates. However, for the present flat membranes, membrane surface per unit volume is low (less than 30 m /m ) whereas for shell and tube configuration this number is 250 m / m . Hence the 2  3  2  3  shell and tube configuration seems to be more promising than flat membranes(Saracco and Specchia (1998)). 2.4.3  Membrane Reactor Configurations  As depicted in Figure 2.3, there are two common configurations for membrane reactors: • Countercurrent or cocurrent: These are typical flow patterns of tubular membranes(Saracco and Specchia (1998)). The choice of cocurrent or countercurrent, strongly depends on the particular reaction of interest (Mohan and Govind (1986); Mohan and Govind (1988b); Mohan and Govind (1988a)). • Stirred—chambers: This setup may be used for the flat membranes and is operated either in a batch or continuous mode and it is suitable for diffusion and reaction tests.  2.4.4  Coupling Catalyst and Membranes  In a catalytic-membrane reactor, membranes and catalyst have to be combined. There are five different configurations: • Fluidized Bed Membrane Reactor: In this configuration, the membrane is immersed in a fluidized bed of catalyst pellets. In this way, the most typical properties  17  Literature Review  •  M E M B R A N E  t  1  M E M B R A N E 1  Cocurrent  Stirred-Chamber  Countercurrent  Figure 2.3: Membrane Reactor Configurations of fluidized-bed reactors (good degree of mixing, high heat transfer coefficients, etc.) are coupled with the separation properties of the membrane (Adris et al. (1994)). However, the possible draw back might be strong abrasion of the membrane (Saracco and Specchia (1998)). • Packed Bed Membrane Reactor: The coupling of a permselective membrane with a packed bed of catalyst pellets has been one of the most widely studied membrane reactor setups. Generally, the catalyst is placed in the membrane tube but in some cases the permselective membrane tube was inserted at regularly spaced intervals into the packed bed of the catalyst pellets (Oertel et al. (1987)). • Catalyst—Deposited Membrane Reactor: The surface of the membrane is deposited with catalytic material. This setup is typical of solid-electrolyte membranes, where the catalyst is also playing the role of the electrode. Possible potential problems are. limitation of catalyst per unit membrane surface and poor electrical conductivity of some catalytic materials. • Catalytically Active Membrane Reactor: The membrane itself is catalytically  Literature Review  18  active. This can be achieved either by using catalytical active material such as Pd-alloys (Gryaznov (1992)) or by depositing membrane material with catalytic materials (Cannon and Hacskaylo (1992)). The undesired issues are low overall catalyst loads and difficulties in getting high surface areas per unit volume of membrane modules (Saracco and Specchia (1998)).  2.4.5  Mechanism of Transport Through Inorganic Membranes  There are eight major mechanisms of transport through inorganic membranes which are discussed briefly (Saracco and Specchia (1998)). • Viscous Flow (Poiseuille Flow): This kind of flow takes place when the mean pore diameter of the membrane is larger than the mean free path of gas molecules (pore diameter higher than a few microns). As a result, collisions between different molecules are much more frequent than those between molecules and pore walls. Under these conditions, no separation between different molecules can be attained. • Knudsen Flow: As the pore dimensions of the membrane decreases (down to fractions of a micron) or the mean free path of gas molecules increases, which can be achieved by lowering the pressure or raising the temperature, the molecules collide more frequently with the pore walls of the membrane than with one another. In this regime, the permeating species flow throughout the membrane, almost independent of each other and transmembrane fluxes are proportional to the inverse square root of the molecular weight of the different gaseous compounds. Hence, the maximum separation factor between two different molecules becomes equal to the square root of the ratio of two molecular weights. • Surface Flow: According to this transport mechanism, one of the permeating molecules can preferentially physisorb on the pore walls (Kapoor et al. (1989)) while the. other molecules can migrate within the gas phase inside the pores. This regime is for pore sizes as small as a few nanometers. • Capillary Condensation: This regime occurs when one of the components can condense ( e.g. molecules with high molecular weight or high boiling point) within  19  Literature Review  the pores. As a result, the condensate fills the pores and then evaporates at the permeate side, where a low pressure is imposed (Kitao et al. (1991)). This mechanism is effective only for small pore sizes at relatively low temperatures. • Multilayer Diffusion:  In the case of strong molecule-surface interactions, this  regime occurs which is between the surface flow and capillary condensation regimes (Uhlorn et al. (1992)). • Molecular Sieving: This regime occurs when pore diameters are small (fractions of nanometer) enough to let only smaller molecules permeate while mechanically preventing the bigger ones from entering. If the pore dimension is monodispersed, high selectivity will be attained. The only limit for this regime is membrane stability which depends on temperature. • Atomic Transport: For metal membranes such as P d alloys or Ag, gas molecules such as oxygen and hydrogen, chemisorb and dissociate on one side of the membrane. Then the atoms dissolve in the metal matrix and diffuse toward the opposite side due to the concentration gradient across the membrane.  A t the permeate side,  atoms combine and desorb as molecules. Depending on temperature, pressure, and gas mixture composition, each of the above steps may become the rate determining step. In some dense materials such as Si02, gas diffusion across the membrane occurs in molecular form (Kim and Gavalas (1995)). • Ionic Transport: This regime is typical of solid electrolytes. The molecules undergo dissociative chemisorption. Atoms are ionized and transported through the crystalline lattice until they lose their charges, combine and desorb molecules at the permeate side.  2.4.6  Major Applications of Inorganic Membrane Reactors  In this section, the major applications of inorganic membrane reactors are reviewed briefly. More information is provided in Saracco and Specchia (1998) and the references in that article.  20  Literature Review  • Selective Removal of Reaction Product: This is the most popular application. As it can be understood from the name, one of the reaction products permeates through the membrane so that the equilibrium is shifted toward the product and consumption of reactants.  As a result, conversion per pass increases. However,  for this application, the permselectivity of the membrane is very important and it should be coupled with thermochemical stability and high permeability. The synthesis of such membranes is a difficult task for material scientists. • Coupling of Reactions: In this application, two reactions are coupled on different sides of the membrane. As a result, the heat of one reaction may be used by the other reaction (in the case of endothermic/ exothermic reactions) or the permeating species from one side may react with the other side and result in flux increase. In addition to this feature, membranes can play an important role in equilibrium limited reactions, as discussed above and in the activation of permeating species (refer to Section 2.4.5, Atomic Transport). Controlling operation of a coupling reaction membrane reactor is not an easy task. • Controlled Addition of Reactant: This is another interesting application of membrane reactors in which one of the reactants is supplied gradually along the reactor, through the membrane wall; whereas the other reactant is on the other side of the membrane. This application is useful for both permseletive and nonpermselective membranes. The only difference is that for nonpermselective membranes the reactant must be fed at the required purity for the reaction while for permselective membrane the reactant will be purified while permeating through membrane wall. This kind of application improves the selectivity and avoids premixing of reactants and consequent side reactions. •  Non-Permselective Membrane Reactor with Separate Feed of Reactants: In this application, two key reactants are fed on the opposite sides of the membrane and the reaction takes place inside the membrane. This application is useful for fast kinetic processes. The advantages of this setup are the same as those of "Controlled Addition of Reactant" with the added advantage of flexibility of controlling the  Literature Review  21  system by simple separate change of flow rates, concentrations and pressures of each reactant. • Enhancement of Reaction Selectivity toward Intermediate Products: In this application, the intermediate products are separated from the reactants thus avoiding further reaction of those intermediates. Since the molecules of the intermediate products are often bigger than the reactants or the products of the complete reactions, design of the membrane is critical.  2.5  Summary of Literature Review  In this chapter, spillover, hydrocracking, coke and membrane reactors were discussed to provide some background information for the present study. In the first section, spillover as it occurs in catalytic systems was discussed. As explained, spillover occurs in many catalytic reactions such as hydroconversion (e.g. hydrocracking) and hydrogenation of coked surfaces. In Section 2, hydroconversion reactions with a focus on hydrocracking were discussed. The classical model and the newly developed model of hydroconversion based on hydrogen spillover, were reviewed. In Section 3, coke was defined and the importance of spillover in coke removal was outlined. Section 4 was dedicated to membrane reactors and their application. Based on the definition of spillover and its role in hydroconversion reactions (e.g. hydrocracking), a novel reaction system can be proposed using a membrane reactor configuration. The role of catalyst in many catalytic systems is to provide spilt-over species to react with organic reactants. Hence, if the reactant and catalyst are separated by an inert medium which can facilitate spillover, the reaction can proceed. Membrane reactors can provide separation of reactant and catalyst while the spilt-over hydrogen can be facilitated by the membrane. Although membrane reactors have been used for a number of different purposes as explained in Section 4, there has been no report of their application based on the hydrogen spillover mechanism. In this study, a ceramic membrane reactor was used to study hydrocracking of diphenylmethane(DPM). The reactant (DPM) and the catalyst were separated by means of a ce-  Literature Review  22  ramie membrane tube. The details of this study are gathered in Chapter 3. In Chapter 4, temperature-programmed hydrogenation of coked silica-alumina and the effect of hydrogen spillover will be presented. In this chapter, several experiments will be discussed in which the catalyst and coked silica-alumina (the reactant) were separated by crushed ceramic (the same material used in the membrane reactor) to imitate the membrane reactor configuration.  Chapter 3 Hydrocracking Diphenylmethane in a Ceramic Membrane Reactor  As discussed in Section 2.1.6, spillover occurs in many catalytic reactions, including hydrocracking. Hydrogen spillover is one mechanistic step of hydrocracking that can partly prevent catalyst deactivation by reaction of spilt-over hydrogen with coke and its precursors on the surface of the catalyst (Section 2.3.4). However, due to the direct contact of catalyst and organic reactant(s), catalyst deactivation by coking is always present. As discussed in Section 2.2.3, the role of catalyst in hydrocracking reactions is to provide spilt-over hydrogen that subsequently cracks the organic molecule. In other words, there is no need for direct contact between the reactant and the catalyst, provided spilt-over hydrogen can reach the reactant. Membrane reactors, discussed in Section 2.4, are a type of multifunctional reactor where one or more chemical reactions, generally catalytically promoted, are carried out in the presence of a membrane. The membrane reactor configuration seems to be a suitable candidate for hydrocracking reactions since the reactant and catalyst could be separated by a membrane that permeates spilt-over hydrogen. In this way, deactivation of catalyst due to coke formation during hydrocracking could be avoided. To study a hydrocracking reaction in a membrane reactor, a ceramic membrane reactor was designed to facilitate transport of spilt-over hydrogen. For this study, diphenylmethane(DPM) was chosen as the model reactant and sulfided N i - M o / A l 0 3 was 2  used as the catalyst. In the following sections, the experimental apparatus and methods used in this study are described. Subsequently, the results of different experiments are presented and discussed. The conclusions from this part of the study are gathered in the last section.  23  Hydrocracking Diphenylmethane in a Ceramic Membrane Reactor 3.1  24  Ceramic Membrane Reactor  Figure 3.1 shows the process flow diagram of the membrane reactor setup that was designed by Dr. Wellington Kwok and Figure 3.2 shows the hydrogen flow pattern in the system. The reactor consisted of a cylindrical stainless steel vessel with an asymmetric Membralox® ceramic tube (purchased from US Filter, Pennsylvania) placed coaxially in the reactor vessel. The length of the ceramic tube was approximately 25 centimeter and the diameter was 1 centimeter. Gas flow from a cylinder, measured and controlled by a Brooks high-pressure mass flow controller (Model Brooks 5850, Rosemount Instrument) entered the reactor via the porous ceramic tube. Unreacted gas and products exited from the outlet of the reactor  Mass Flow Controller  "O"  -Xr TC  Temperature Controller  Flash Drum  Product Lines  Figure 3.1: Process Flow Diagram of Ceramic Membrane Reactor  Hydrocracking Diphenylmethane in a Ceramic Membrane Reactor  Note: H  s o  stands for spilt-over hydrogen.  Figure 3.2: Hydrogen Flow in Ceramic Membrane Reactor  25  Hydrocracking Diphenylmethane in a Ceramic Membrane Reactor  26  and were separated in a flash drum. The pressure of the flash drum was maintained by a pressure relief valve (Nupro) and the temperature was monitored by a temperature indicator (Omega, D P 371 Series) reading a K-type thermocouple installed from the top of the flash drum. The temperature of the reactor was measured by a K-type thermocouple and controlled by an Omega temperature controller (Omega, C N 76000 Series) which was connected to a radiant heater, supplied by Watlow. The pressure of the reactor was measured by a pressure transducer and controlled by an Omega controller (Omega, C N 76000 Series, PID) actuating on a Badger research control valve (Model 859 with E V A - 1 actuator, Badger Meter). For Safety reasons, a rupture disk with a maximum burst pressure of 3000psig at 400° C was purchased from Lamot and installed on separate tubing that connected the reactor vessel to vent. To avoid D P M solidification (melting point = 22-24° C), the outlet tubing of the reactor, flash drum and vent were heated by thermolyne heating tapes and the tape temperature was controlled by an Omega controller (Omega, C N 76000 Series). As shown in Figure 3.3, the reactor vessel was divided into the tube and shell-sides by two specially designed membrane holders which protected and sealed the membrane tube at either end of the reactor. The membrane holder consisted of a tapped, solid stainless steel cylinder in which the ceramic tube was placed. Sealing was achieved by means of braided graphite, compressed by a square-head threaded nut that screwed into the membrane holder. The holder assembly at the top of the reactor vessel had three outlets (for the reactor outlet, thermocouple and rupture disk) while the bottom had one outlet used for sample collection from the shell-side of the reactor. A second nut screwed on top of the square head nut to provide a connection between the ceramic membrane tube for gas inlet at the top of the reactor, and sample collection from the tube-side at the bottom of the reactor. To ensure sealing between these nuts, a copper gasket was placed around each square—head nut prior to placing of the second nut. The membrane holders were sealed by means of two threaded flanges tightened by eight stainless steel screws on either end of the reactor. Again, to ensure sealing between the holders and the reactor vessel, stainless steel graphite spiral wound gaskets (purchased from Wriason Seals Ltd.) were placed between the counterparts of each end of the reactor. To prevent fusing  27  Hydrocracking Diphenyhnethane in a Ceramic Membrane Reactor  Gas  L  Copper  L , ~  Membrane  GasketXJ Square Nut  Braided Graphite GaskeT  Threaded ge  Reactor  SS-Graphite Spiral wound Gasket  Wall  Membrane Tube  Figure 3.3: Detail of Membrane Holder for Reactor, Top Holder of threaded flanges and the reactor vessel due to high operating temperature, anti-seize compound was applied.  3.2 3.2.1  Experimental Methods C a t a l y s t Sulfidation  The catalyst used for these experiments was a commercial N i - M o / A ^ O s catalyst (Extrudate, 0.99 mm, Criterion Catalyst 1443b, CY/1.0), supplied by Syncrude Canada Limited. The catalyst was soaked in dimethyl disulfide ((DMDS), 98%, purchased from Aldrich) at room temperature overnight. The amount of DMDS was 1.5-5 times the stoichiometric amount required to sulfide the N i and Mo present on the catalyst (Appendix C). The wet  28  Hydrocracking Diphenylmethane in a Ceramic Membrane Reactor  catalyst was placed inside the ceramic tube of the reactor, heated to 250° C under nitrogen flow and held at that temperature for 2 hours. Subsequently, the reactor was cooled under nitrogen flow and the sulfided catalyst was kept in the reactor for the hydrocracking experiment.  3.2.2  Hydrocracking Reaction  In a typical experiment, about 5 g of presulfided catalyst was placed inside the ceramic tube and sulfided as explained in Section 3.2.1. After catalyst sulfidation, about 80 mL D P M (99%, purchased from Aldrich) was added to the shell-side of the reactor. The reactor was pressurized to 600psig under extra dry nitrogen (99.95%, supplied by Praxair) flow. Upon reaching this pressure, heating was started. Upon reaching the desired temperature (400° C), the gas was switched from nitrogen to 5 % H S / H (purchased 2  2  from Praxair) by means of a three-way valve. Gas flow was kept constant at 500sccm for both nitrogen and 5 % H S / H . The reaction was stopped after 1 hour and the reactor was 2  2  cooled in nitrogen. The liquid products were collected from the flash drum, the tube-side and shell-side of the reactor. A gas product sample was collected from flash drum for analysis.  3.2.3  A n a l y s i s of the P r o d u c t s  Liquid products were weighed to determine a mass balance. Liquid samples were analyzed qualitatively by G C - M S equipped with a 30m x 0.25//m capillary column coated with D B 17. Quantitative analysis was carried out by a Varian gas chromatograph ( Model 34000*) using a packed OV-17 column and flame ionization detector. Gas samples were analyzed by a HaysepQ column and thermal conductivity detector.  3.2.4  Safety Considerations of the M e m b r a n e R e a c t o r  As outlined in previous sections, the hydrocracking reaction of D P M was carried out at high temperature (i.e. 400° C), high pressure (i.e. GOOpsig) in the presence of corrosive gas ( 5 % H S / H ) . Under these severe operating conditions, sealing was an important and 2  2  I.  Hydrocracking Diphenylmethane in a Ceramic Membrane Reactor  29  critical task. Hence, to ensure that the reactor was leak free and the ends of the ceramic membrane tube were connected to the stainless steel housing, a number of different gaskets were carefully placed (explained in Section 3.1) and these were changed after each experiment. Prior to each experiment, the reactor was pressurized in nitrogen to lOOOpsig at room temperature and all the connections were checked for leaks by means of Snoop solution. The reactor setup was placed in an enclosed vented chamber to contain any accidental spill. Also, the cylinder of 5%H2S/H2 was placed in a separate enclosed vented chamber to avoid spreading this toxic gas in the case of a leak. For safety reasons, during operation, a hydrogen sulfide detector equipped with an alarm (Toxilog 1800 Series, supplied by Biosystem Inc.) and a fire extinguisher, were placed near the apparatus. The former was used to ensure the level of hydrogen sulfide was below the safety limit and the system was leak free. The latter was used in the case of a leak, since if liquid D P M were to contact the hot metal surface reactor, it would immediately ignite.  3.3  Results and Discussion of Hydrocracking Diphenylmethane in a Ceramic Membrane Reactor  To study D P M hydrocracking in a ceramic membrane reactor, different experiments were performed in the presence, and absence of catalyst. In the former case, the effect of catalyst immersion in D P M was compared with the case that catalyst and D P M were separated by the ceramic membrane. Also, the effect of membrane pore size and position of catalyst in the ceramic tube, on the hydrocracking products, was investigated. The experiments presented in Section 3.3.1 and Section 3.3.2 were carried out under the direct supervision and instruction of Dr. Goutam Chattopadhyaya.  3.3.1  Comparison of Thermal and Catalytic Experiments  To investigate the activation of hydrogen on sulfided catalyst and its transport through ceramic pores for reaction with D P M , thermal and catalytic experiments were carried out. For both experiments, D P M was placed in the shell-side of the reactor and the  Hydrocracking Diphenylmethane in a Ceramic Membrane Reactor  30  operating conditions were identical. For the catalytic experiment, about 5g of catalyst was placed inside the ceramic tube. The catalytic experiment was repeated and quantitative and qualitative analysis of the liquid products was performed. The major products of D P M hydrocracking in the presence of sulfided NI-M0/AI2O3 were benzene and toluene, confirming the results in Hattori et al. (1988). The D P M feed was analyzed by gas chromatography and the results showed that there was no benzene or toluene in the feed (Appendix B). A gas sample was collected and analyzed. In all cases, only trace amounts of methane were observed in the gas samples. Table 3.1 summarizes the results of the catalytic and thermal experiments. The reported average and standard deviation for mass balance percentage and weight percent of liquid recovery from different parts of the system (i.e. Shell-side, tube-side and separator) were calculated based on the mass balance and the weight percentage of liquid recovery obtained for each experiment. Consequently, the summation of average liquid recovery for each part of the system does not match the average mass balance. Table 3.2 shows a qualitative analysis of the liquid product of catalytic hydrocracking. As indicated in Table 3.2, benzylcyclohexane was another major product. However, the production of benzylcyclohexane was due to hydrogenation reaction catalyzed by the stainless steel reactor (Hattori et al. (1988)) and therefore the amount of benzylcyclohexane was not measured. The yields of the major products (benzene and toluene) in the catalytic experiments clearly show the activity of catalyst in comparison to the thermal experiment.  Also,  the results of the thermal experiment suggest that molecular hydrogen does not enhance hydrocracking. This is consistent with the mechanism of hydroconversion proposed by Stumbo et al. (1995); Roessner and Roland (1996) ; Stumbo et al. (1997). The mechanism is an extension of the classical model of hydroconversion and is based on hydrogen spillover. Hence, the catalytic experiments are consistent with hydrogen dissociation on the sulfided catalyst followed by spillover onto the support. However, in the ceramic membrane reactor, hydrogen was activated on the catalyst placed on the tube-side, whereas D P M was placed on the shell-side. There are a number of possibilities for reaction of spilt-over hydrogen with D P M in this configuration, namely: - Secondary spillover (Conner Jr. and Falconer (1995); Rozanov and Krylov (1997))  Hydrocracking Diphenylmethane in a Ceramic Membrane Reactor  31  Table 3.1: Hydrocracking Diphenylmethane in a Membrane Reactor Compared to Thermal Conversion. Temperature= 400°C, Pressure= 600psig, Gas Flow 500 seem of 5%H S in H , Membrane 2  2  Pore Size= lOOOA Thermal  Catalyst on Tube- Side D P M on Shell-Side  wt.% of Product Recovered from: Separator  2.6  1.5 ± 0 . 4  Tube-Side  36.4  6.2 ± 6 . 5  Shell-Side  61.0  92.3±6.9  99  96.7±0.4  Benzene  0.02  1.73±0.13  Toluene  0.02  1.22±0.60  Overall Mass Balance, wt.%: Products Yields, mole%  of hydrogen onto the ceramic membrane and subsequent transport to the D P M - Gas transfer of spilt-over hydrogen (Baumgarten et al. (1989)) - Transport of D P M through the membrane pores to the catalyst placed on the tubeside. D P M may flow through the membrane pores during loading of D P M by capillary flow discussed in Section 2.4.5. Also, D P M may diffuse as a consequence of the concentration profile between the shell-side and tube-side of the reactor. The results of catalytic hydrocracking show that most of the liquid product was recovered from the shell-side (~92%) but a significant amount of product (~6%) was collected from the tube-side of the reactor where the catalyst was located. Hence, the possibility that reaction between spilt-over hydrogen and D P M may have occurred on the tube-side of the reactor cannot be excluded and consequently, the reaction zone for D P M hydrocracking is not established.  Hydrocracking Diphenylmethane in a Ceramic Membrane Reactor  32  Table 3.2: Summary of Components Identified from Liquid Product Samples Generated at Reaction Conditions in Table 3.1. Note that the relative intensities do not correspond in direct proportion to concentration.  Relative Intensity Catalyst on Tube-Side Compound Identified  D P M on Shell-Side  Diphenylmethane (DPM)  100  Decane (Internal Standard)  10  Benzylcyclohexane  1.0  Toluene  0.86  Benzene  0.45  Diphenylenemethane  0.25  Decalin  0.41  Methyl-DPM  0.29  2-Methyl-DPM  0.26  4-Methyl-DPM  0.15  m-Benzyl-DPM  0.13  Triphenylmethane  0.15  33  Hydrocracking Diphenylmethane in a Ceramic Membrane Reactor 3.3.2  Effect of Catalyst Immersion in D P M  To understand the effect of the presence of D P M on the catalyst during hydrocracking, experiments were performed in which the catalyst and D P M were both placed on the shell-side of the reactor. The operating conditions were the same as in the previous experiments. Table 3.3 summarizes the results of the liquid product analysis. Table 3.4 shows a qualitative analysis of the liquid product of catalytic hydrocracking. Table 3.3: Hydrocracking Diphenylmethane in a Membrane Reactor with Catalyst on Shell-side and Tube—side Temperature= 400°C, Pressure= 600psig, Gas Flow 500 seem of 5%H S in H , Membrane 2  2  Pore Size= lOOOA Catalyst on Shell-Side  Catalyst on Tube- Side  D P M on Shell-Side  D P M on Shell-Side  96.8±1.1  96.7±0.4  Benzene  0.16±0.08  1.73±0.13  Toluene  0.18±0.01  1.22±0.60  Overall Mass Balance, wt.%: Products Yields, mole%  A gas sample was also collected and analyzed but in all cases, only trace amounts of methane were detected. The yields of products, in the case of separation of catalyst and D P M was clearly higher than the case where catalyst and D P M were both placed on the shell-side. In the latter case, the catalyst was immersed in the liquid and since there was no mixing, hydrogen presumably had to overcome a significant mass transfer resistance in order to be activated on the catalyst surface. Based on these results, it seems that diffusion of hydrogen through a liquid phase on the catalyst surface was the rate determining step. Also, it can be concluded that in the case that catalyst and D P M were separated by the membrane, if D P M was transported to the tube-side, then the amount of D P M did not result in complete wetting of the catalyst.  Hydrocracking Diphenylmethane in a Ceramic Membrane Reactor  34  Table 3.4: Summary of Components Identified from Liquid Product Samples Generated at Reaction Conditions in Table 3.3. Note that the relative intensities do not correspond in direct proportion to concentration.  Relative Intensity Catalyst on Shell-Side Compound Identified Diphenylmethane (DPM)  D P M on Shell-Side 902  Decane (Internal Standard)  10  Benzylcyclohexane  1.2  Toluene  0.73  Benzene  0.54  Diphenylenemethane  0.32  Decalin Methyl-DPM 2-Methyl-DPM 4-Methyl-DPM m-Benzyl-DPM Triphenylmethane  0.8  -  35  Hydrocracking Diphenylmethane in a Ceramic Membrane Reactor 3.3.3  Effect of Membrane Pore Size  As discussed, the results from previous sections suggest that one possibility for reaction of spilt-over hydrogen and DPM  is by transport of DPM  through the ceramic pores by  capillary forces to the catalyst on the tube-side of the reactor. It is known that at constant temperature, with a larger pore membrane, lower gas pressure drop is needed to overcome the capillary forces that draw liquid into the pores. The  relationship between the pressure drop and solid pore size is given by the following  equation: AP  = 4Scose/d  where A P is the pressure drop needed to overcome the capillary forces, S is the surface tension of the liquid, d is the solid pore size and 0 is the contact angle between the liquid and  solid surface. The contact angle for a liquid wetting the solid surface is between 0 and  90°.  The contact angle depends on the presence of contaminates and impurities of the  solid, whether the liquid is advancing over a dry surface or receding from a wet surface and  the extent of vibration of the droplet. Hence there is rarely a single-valued quantity  for  contact angle (Shaw (1992)). In Figure 3.4, the relationship between pressure drop  and  pore size is shown for the experimental system under study. For this calculation, it  was assumed that the contact angle between DPM the surface tension for DPM The  and the membrane surface is 45°  and  was obtained from Washburn et al. (1929).  data of Figure 3.4 suggest that under the same operating conditions, larger pore  membranes will require less of a gas pressure drop through the membrane to overcome the capillary force of DPM  to the tube-side and hence reduce the effect of DPM  to the catalyst. To evaluate the effect of DPM  transport  transport from the shell-side to the tube-  side of the reactor, experiments were carried out in which membranes with 100A  and  1000A pore size were employed. Table 3.5 compares the results of hydrocracking using 100A and 1000A pore size membranes in the membrane reactor. The  average percentage of liquid recovered from the tube-side of the membrane with  small pores (lOOA) is higher than the membrane with large pores (lOOOA), confirming  Hydrocracking Diphenylmethane in a Ceramic Membrane Reactor  1600  Membrane Pore Diameter (Angstrom)  Figure 3.4: Effect of Membrane Pore Size on Pressure Drop Required to Displace Liquid From Pores the theoretical calculations shown in Figure 3.4. Furthermore the yield of toluene for the 100A pore size membrane was lower than the 1000A pore size membrane.  3.3.4  Effect of Catalyst Location  As shown in previous sections, D P M may be transported through the membrane pores by capillary action and this effects the yield of products depending on the degree of catalyst wetting. In an attempt to minimize the catalyst wetting by D P M , the catalyst was placed on the tube-side above the liquid level of the shell-side. This was achieved by inserting a solid stainless steel rod inside the ceramic tube and placing the catalyst on top of the solid rod. Table 3.6 compares the result of this experiment with the standard catalytic experiment. The data show an increase in the yield of benzene and toluene in the case of the catalyst located above the liquid level, even though the amount of catalyst used in this  37  Hydrocracking Diphenylmethane in a Ceramic Membrane Reactor  Table 3.5: Hydrocracking Diphenylmethane in a Membrane Reactor with Pore Size  100A  and  lOOOA;  Catalyst on Tube-side and D P M on Shell-Side  Temperature= 400°C, Pressure= 600psig, Gas Flow 500 seem of 5%H S in H 2  100A  1000A  Separator  14.4±3.8  1.5±0.4  Tube-side  14.8±1.1  6.2 ± 6 . 5  Shell-side  70.8±4.9  92.3±6.9  Overall Mass Balance, wt.%:  88.7±7.5  96.7±0.4  Membrane Pore Size  2  wt.% of Product Recovered from:  Products Yields, mole% Benzene  1.59±0.54 1.73±0.13  Toluene  0.48±0.18  1.22±0.60  Table 3.6: Hydrocracking Diphenylmethane in a Membrane Reactor with Catalyst on Tube-side Placed at Different Levels; D P M on Shell-Side Temperature= 400°C, Pressure= 600psig, Gas Flow 500 seem of 5%H S in H , Membrane 2  Pore Size=  2  lOOOA  Catalyst above  Catalyst below  Liquid Level  Liquid Level  2.7  ~5.0  Separator  33  1.5±0.4  Tube-Side  6  6.2 ± 6 . 5  Shell-Side  61  92.3±6.9  Overall Mass Balance, wt.%:  92  96.7±0.4  Benzene  2.76  1.73±0.13  Toluene  4.27  1.22±0.60  Catalyst in Reactor, grams wt.% of Product Recovered from:  Products Yields, mole%  Hydrocracking Diphenylmethane in a Ceramic Membrane Reactor  38  case was almost half of the standard catalytic experiments. The results suggest that low D P M transport and hence less wetting of the catalyst gives higher yield of benzene and toluene.  3.4  Conclusions from Hydrocracking Diphenylmethane in a Ceramic Membrane Reactor  Cracking of D P M in a ceramic membrane reactor configured to separate the catalyst and the D P M reactant was carried out in the presence and absence of sulfided N1-M0/AI2O3. The yield of benzene and toluene from thermal cracking was very low, indicating that molecular hydrogen did not react with D P M , confirming the mechanism of hydrocracking reviewed in Section 2.2.3. The increased yield of benzene and toluene in the presence of catalyst was consistent with a hydrogen spillover mechanism. However, it was not clear whether spilt-over hydrogen permeated through the ceramic pores (via secondary spillover onto the ceramic and/or via gas phase transfer) or whether D P M was transported through the ceramic pores (due to capillary forces and/or concentration profiles on the sides of the membrane) to the catalyst. Different reactor configurations were investigated in an attempt to eliminate D P M transport to the tube-side but this possibility could not be completely excluded. Consequently, the reaction zone for D P M hydrocracking (i.e. the tube-side, shell-side or within the membrane pore) could not be clearly identified. The severe operating conditions needed for D P M hydrocracking caused many technical difficulties during the course of these experiments. Hence, other experiments were designed to simplify the experimental setup, eliminate the possibility of liquid transfer but still quantify the extent of hydrogen spillover through the ceramic membrane material. Chapter 4 describes these experiments.  Chapter 4 Effect of Hydrogen Spillover on Temperature—Programmed Hydrogenation of Coked Silica—Alumina  As reviewed in Section 2.3.4, spilt-over hydrogen reacts with coke deposited on the surface of a catalyst to produce methane. Indeed, hydrogen spillover can partly prevent coking in the course of catalytic reactions by reacting with coke and its precursors. However, due to the direct contact of organic reactant(s) with catalyst, coking and catalyst deactivation is always a problem. The configuration of a ceramic membrane reactor, as discussed in Chapter 3, may separate the reactant(s) and catalyst while maintaining catalytic activity by transport of spilt-over hydrogen through the ceramic membrane. However, the experimental work reported in Chapter 3, showed that the separation of liquid feed and catalyst by a porous ceramic membrane tube was not completely achievable. To obtain complete separation of reactant and catalyst by the ceramic membrane, and to further study the effect of hydrogen spillover in catalytic reactions, dried amorphous silica-alumina was coked artificially by propylene and this coked silica-alumina was used as a model reactant for hydrogenation. However, the chemical nature of coke is rather complicated and no simple chemical formula for coke exists. Coke includes cyclic and other organic groups depending on the organic reactant and the operating conditions under which the coke is produced (see Section 2.3). Therefore, the coked silica-alumina may undergo different hydrogenation reactions at different temperatures depending on the nature of the species which were formed during coking. Hence temperature-programmed methods provide better understanding of the reactivity of coked silica-alumina because in these methods the reaction is monitored while the temperature increases linearly with time. In this chapter, the experimental apparatus and methods used for drying and coking  39  Effect of Hydrogen Spillover on TPH of Coked  40  Silica-Alumina  of amorphous silica-alumina, and the temperature-programmed reaction of coked silicaalumina, are explained. Subsequently, results of different experiments are presented and discussed. In the last section, conclusions from this part of the work are drawn.  4.1 4.1.1  Apparatus Drying and Coking Apparatus  Figure 4.1 shows the flow diagram used for drying and coking of silica-alumina. The flowrate of gas from a cylinder was measured and controlled by a calibrated rotameter prior to the reactor inlet. The U-shape reactor was placed in a thermolyne furnace (Model 47900) purchased from V W R Scientific. The furnace was equipped with a thermocouple and temperature controller. The outlet of the reactor was connected to atmosphere.  Outlet to Atmosphere  H Quartz Reactor  Furnace Thermocouple  Furnace  Propylene  Helium  Figure 4.1: Drying and Coking Apparatus  Temperature Controller  Effect of Hydrogen Spillover on TPH of Coked 4.1.2  Silica-Alumina  41  Temperature-Programmed Reaction (TPR) Apparatus  As depicted in Figure 4.2, the apparatus consisted of a U-shape reactor, heated by a Lindberg furnace (Lindberg, Model 55301) equipped with an Omega programmable temperature controller (Model 2010), reading a K-type thermocouple placed i n the reactor. For safety reasons, an Omega pressure indicator was installed prior to the reactor inlet. Gases were measured and controlled by Brooks mass flow controllers (Models 5850E and 0154E). A 0.5 Lira Nupro filter and a Nupro check valve were placed before and after each mass flow controller sensor, respectively. Trace amounts of water were removed from the feed gases using moisture traps, (purchased from Chromatographic Specialties Inc. and Altech Associate Inc.) placed after the gas cylinder regulator. Since the system was rather complicated, rotameters were installed prior to the reactor inlet to provide visual confirmation of gas flow direction. The gas was analyzed by means of a quadropole mass spectrometer, purchased from Anglo Scientific Instruments. B y means of on-off and three-way valves, the apparatus was flexible and could be used for temperature-programmed reactions, catalyst reduction and calibration of the mass spectrometer. Table 4.1 summarizes the different valve positions for different experiments.  4.1.3  Reactor  A schematic diagram of the reactor is shown in Figure 4.3. The reactor was made of quartz glass by Canadian Scientific Glass Blowing to the specification given in Figure 4.3(a). For a typical experiment, the U tube of the reactor was filled with quartz wool, and the solid particles (catalyst, silica-alumina) were placed on top of the quartz wool in the 7mm diameter branch of the reactor, as shown in Figure 4.3(b). The 4mm diameter exit branch of the reactor was designed to facilitate rapid transport of the products to the analytical section of the system, thereby limiting the effects of secondary thermal reaction. For temperature-programmed reactions, a K-type thermocouple was placed on top of the quartz wool in the exit branch of the reactor to measure and control the temperature of the Lindberg furnace. To avoid breakage of the glass reactor by the heavy Swagelok connector fittings, special frames for both the T P R apparatus and the coking and drying apparatus were designed to carry the weight of these fittings. To facilitate  Effect of Hydrogen Spillover on TPH of Coked  Silica-Alumina  42  Effect of Hydrogen Spillover on TPH of Coked  43  Silica-Alumina co  fl fl H-=  -8 tH  fl o  CU  cu fl  , f l  u  r2 'cS O  cu  *tt  O  ci  Xi HJ  CU  co co cci  co cci  P< >,  fl O  .o  ttt  fl  cu  o o  P,  CQ t-i o  PQ O  H-3  o  <J cd cu  CS CU  tf  cu  o s-< c_>  H-3  CO  O.I C O  co co cci  l-l  co CU fl O  o  fl  1' CU  (-1  p-t  u  <u fl <u >> a, o  t—1  CM  .o  cu fl cu PH  o  n P-.  a, PQ o  OH  fl  o  fctt O  fl  O  i3 (U co  53  CQ l-l O  -u o  H-3 CJ ccj CU  cci CU  tf  tf  CU H-=  cu  6 o u  HJ  o <u  a, | C O  co co  CU l-l  (U  co cci  o  fl cu tf  fctt o o fctt oo o  .—i  cu u  H-=  •p  O  o  (-1  o  o u  tf  tf  co O  a  cci CU  ccj CU  <u  tf  O  PH  EH  o  ttt  (tt  oo  ofl  <u  .—i  fl fl u  o  o  to  H-=  o  cci  cu  tf  O  n o  +2  o cci cu  tf  O  cu  a o  A  CU  b O O  n  l-l H^>  cj cu  ft I  C O  co co  fl o  • --i  H^  o  cci CU •n  £  ^ s  u  tf  u CU  < 3  l-l  H-S  fl O b O IH  fl fl  co  O  CU  <—i OJ > !>  n '•>  > > >  cu cci i-i b O O  In Pi I  CU U fl  d <-<  H-2  cu  PH  xi H-=  PH ccj O  b O fl b O  l-i  fl Oi IH  t3 cu o CU  c fl o o  44  Effect of Hydrogen Spillover on TPH of Coked Silica-Alumina  attachment and detachment of the reactor and to avoid exposure of the reactor's contents to ambient, stainless steel Swagelok Quick connects were installed on the inlet and outlet tubes of the reactor.  7mm  4 mm H  J err  Thermocouple  39 cm  24 cm  Solid  >•  Quartz Wool  (b)  (a)  Figure 4.3: Reactor  4.2 4.2.1  Experimental Methods Drying and Coking  For a typical experiment, about 250 mg of silica-alumina (Silica-alumina, average particle size: 67/xm, catalyst support, grade 135, Aldrich Chemical Company Inc.) was placed in the quartz reactor on top of the quartz wool. After a leak test to confirm that the reactor was properly sealed, the sample was heated under 171 m L ( N T P ) / m i n helium (Grade UHP, 99.999%, supplied by Praxair) from 100°C to 600°C at a ramp rate of 10°C/min, followed by a soak time of 1 hour at 600°C. Upon completion of the drying, the three-way valve was switched from helium to propylene (Grade C P . , 99%, supplied by Praxair). The flowrate of propylene was set at 40 m L ( N T P ) / m i n for 15 minutes. Due to the high heat capacity of the furnace and to avoid aging of coke at high temperature, the reactor was  Effect of Hydrogen Spillover on TPH of Coked Silica-Alumina  45  immediately removed from the furnace using the Quick connects, and was placed in a dry nitrogen glove box to cool to ambient. The coke content of the silica-alumina was determined by Canadian Microanalytical Service Ltd.  4.2.2  Catalyst Reduction  About 260 mg of calcined 2%Co-Si02 (average particle size: 0.17mm), prepared by Soltan Mohammad Zadeh (1998) was placed in the quartz reactor on top of the quartz wool as described previously. Based on the procedure reported by Soltan Mohammad Zadeh (1998), the catalyst was reduced by heating in 150 m L / m i n hydrogen (Grade U H P , 99.999%, supplied by Praxair) from 30° C to 450° C at a ramp rate of 10°C/min and soaked at 450° C for 1 hour. Subsequently, the reduced catalyst was cooled to room temperature in hydrogen and the reactor containing the reduced catalyst was also placed in the dry nitrogen glove box.  4.2.3  Temperature—Programmed Hydrogenation ( T P H )  Prior to the hydrogenation experiment, addition of catalyst, ceramic or dried silicaalumina to the coked silica-alumina sample was carried out in the nitrogen glove box, as necessary. Subsequently, the reactor containing the sample was connected to the T P R frame. The reactor was removed from the nitrogen glove box and attached to the the T P R (Temperature-Programmed Reaction) apparatus. A mixture of 30%argon in hydrogen (Grade U H P , 99.999%, supplied by Praxair) was flowed at 30 m L ( N T P ) / m i n to the reactor at room temperature to remove any volatile matter. The exhaust gas was analyzed in 10 second intervals by the quadropole. mass spectrometer and recorded by a computer data logger. Upon stablizing the mass spectrometer, the reactor was heated from 40°C to 600° C at a ramp rate of 40°C/min. Subsequently, the reactor was cooled to 100° C in hydrogen and argon, and data were collected by the mass spectrometer during this period as well.  Effect of Hydrogen Spillover on TPH of Coked Silica-Alumina 4.2.4  46  Temperature-Programmed Reaction ( T P R ) in Argon  The experimental procedure was the same as for the T P H experiment, explained in Section 4.2.3, except that instead of 30 m L ( N T P ) / m i n argon and hydrogen, only 30 m L ( N T P ) / m i n argon flowed to the reactor.  4.2.5  Calibration  A t the end of each set of experiments, calibration of the mass spectrometer was performed. The position of the valves was set as explained in Table 4.1. A pulse of methane or propylene was injected in the argon and/or hydrogen flow, the flowrates of which were the same as for T P H and T P R experiments. Argon was used as an internal standard for determining total pressure of the gas mixture in the mass spectrometer vacuum chamber. Calibration factors for other components (ethane and ethylene) were estimated based on the methane calibration factor and the known intensity ratio of methane and each product, provided by the mass spectrometer supplier. In Appendix F the details of calibration are gathered.  4.3  Results and Discussion of Temperature-Programmed Hydrogenation of Coked Silica-Alumina  As reviewed in Section 2.3.4, hydrogenation of coked surfaces are enhanced by the presence of metal catalysts which provide spilt-over hydrogen species. Methane production is the result of reaction of spilt-over hydrogen with coke deposits. The main purpose of this study was to investigate the reaction of spilt-over hydrogen with coked silica-alumina, and to determine the possibility of transport of spilt-over hydrogen across a layer of the ceramic membrane material used in Chapter 3. However, due to the complex nature of the reaction system, different experiments had to be carried out to identify the sources of products. Hence, different experiments with and without catalyst were performed.  In the former case, studies on the effect of location of the  catalyst relative to coked silica-alumina were performed. To clarify the effect of hydrogen spillover on hydrogenation of coked silica-alumina, volatile matter was removed from the  Effect of Hydrogen Spillover on TPH of Coked  47  Silica-Alumina  coked silica-alumina by temperature-programmed reaction in argon and subsequently temperature-programmed hydrogenation in the presence and absence of catalyst were performed. The possibility of transport of spilt-over hydrogen across either a layer of ceramic or dried silica-alumina (Silica-alumina, average particle size: 67/zm, catalyst support, grade 135, Aldrich Chemical Company Inc.) was examined by placing a thin layer of crushed ceramic (the same ceramic material used for studies in Chapter 3) or dried silica-alumina between the coked silica-alumina and the catalyst.  In addition,  several tests for investigating the flow pattern in the reactor were carried out.  4.3.1  Comparison of Non-catalytic (Thermal) and Catalytic  Temperature-  Programmed Hydrogenation of Coked Silica—Alumina To show the effect of hydrogen spillover on hydrogenation of coked silica-alumina, several experiments in the presence and absence of catalyst were performed. In the latter case, the catalyst (2%Co-Si02) was placed on top of the coked silica-alumina bed and these experiments were repeated to obtain a measure of the experimental repeatability. Figure 4.4 and Figure 4.5 show the products of non-catalytic (thermal) and catalytic temperature-programmed hydrogenation of coked silica-alumina, respectively. Given the limit on the number of products recorded by the data logger of the mass spectrometer, methane, propylene, ethane and ethylene were the only hydrocarbons monitored. As shown, ethylene, ethane and propylene production occurred in the same range of temperature whereas there were two regions for the production of methane. The first occurred in the same range of temperature as for the other products. The second region occurred at higher temperature (> 500°C) where the production of other products decreased to zero. A similar pattern occurred for all experiments including those to be discussed in later sections. Also, for the catalytic experiment, methane production between the first and second region did not drop to zero whereas for thermal (non-catalytic) experiment, methane production dropped to zero between the first and second region of methane production. Table 4.2 summarizes the results of the experiments in terms of the amount of product and T  m a x  (the temperature at which the maximum rate of production occurs),  with the standard deviations calculated from repeat experiments.  Effect of Hydrogen Spillover on TPH of Coked  Silica-Alumina  Figure 4.4: Temperature-Programmed Hydrogenation of Coked Silica-Alumina  Effect of Hydrogen Spillover on TPH of Coked Silica-Alumina  49  Figure 4.5: Temperature-Programmed Hydrogenation of Coked Silica-Alumina in the Presence of Catalyst (2%Co-Si0 ), Catalyst Placed on Top of the Coked Silica-Alumina 2  Effect of Hydrogen Spillover on TPH of Coked Silica-Alumina. d o CU CJ  3, CO  13 <u co  a  CM  _b0  Pi CU  3.  u  o o -H o>  CM O O  -H oo  CM  CO  I  o O  PH  X  a;  + = > cs O a  Pi JU  H  CT3  'o  a a,  OHO H  ed  6 t-i  E  cu H  o I  a  a  PH  a.  o u  PH  o H o  o  E  *S CO  bO  -»-=• O  ,—i  CD  PL, <+H  o a o CO  "in ccj  -H  co  -t-=  O  a a.  o h o  E  CM  CO CO  CO  >->  13 o -H IO CM  Oi  a  o  CO  _b0  -H Oi  CCJ  o  -H oo co  O <D CJ CH  CM  0) CO 0) tH  Oi  PH  I—I  <U  -H CM  -4-3  CM  -H  -H  CO  CO  CM  CM  -H  -H  t—I  -H  CO CO CM  -H co CM  O CJ  k CJ  <+H  O  0  o ci  u  CJ  13 13  O  O O  <+H  O  PI  Pi  .2  o  '+= cc3  H-=  a  Pi  bO  bO  CU  cu  cci ^  CJ  nd <U  CM  Oi  k  CJ  i—i  Oi  PI ci  PH  O  CT3  cd  a  CCJ  o I  a  o  bO u  CU PH  o  pi o t-l  Pi CO CU  u  >> M cu  a I  M O t-i PH  bO I  CU SH  (Ci  u O CJ  O  O  cu  PH  PI  api  13 k CJ  t-i ' d  >, id <u ^H  a a  b0 cd  O  u  PH  I  cu  <U  CJ  (H  o CU  ^H  PH  p! U  «3 CU PH  a  50  Effect of Hydrogen Spillover on TPH of Coked Silica-Alumina,  3.00E-03  n  2.50E-03 -\  0  100  .200  300  400  500  600  Temperature, °C Thermal in hydrogen, R97-2  Catalyst on top of coke, R98-2  Figure 4.6: Comparison of Methane in Thermal and Catalytic Experiments  Effect of Hydrogen Spillover on TPH of Coked  52  Silica-Alumina  A clear distinction between the catalytic and non-catalytic (thermal) production of methane is shown qualitatively in Figure 4.6 and quantitatively in Table 4.2. The T  max  of  methane for the catalytic experiment was unexpectedly higher than for the non-catalytic experiment. However, the catalyst promotes the reaction of less active carbon that occurs at higher temperature, resulting in a higher T - There were only slight increases in max  propylene and ethane production due to catalytic activity. Ethylene production was the same for non-catalytic (thermal) and catalytic experiments. As shown in Figure 4.4 and Figure 4.5, there was no significant production of ethane.  4.3.2  Effect of Position of Catalyst and Coked Silica-Alumina on TemperatureProgrammed Hydrogenation  To investigate the effect of the relative position of the catalyst and coked silica-alumina in the fixed bed reactor, experiments were carried out in which (a) the catalyst bed was placed on top of coked silica-alumina, (b) catalyst and coked silica-alumina were mixed, (c) coked silica-alumina was placed on top of the catalyst. Table 4.3 and Figure 4.7 compare the quantitative and qualitative results of these experiments, respectively. The maximum production of methane and ethane occurred when the coked silica-alumina was placed on top of the catalyst and the minimum production occurred when the catalyst was placed on top of the coked silica-alumina. Production of propylene followed the opposite trend of methane and ethane suggesting that some of the propylene that is produced from the coked silica-alumina, underwent hydrogenolysis reactions on the surface of the catalyst to produce methane and ethane. Given the standard deviation in ethylene production, no clear trend in ethylene production could be identified for the different catalyst and coked silica-alumina configurations.  4.3.3  Effect of Hydrogen Spillover on Temperature-Programmed Hydrogenation of Coked Silica-Alumina  As discussed in the previous section, it seemed that propylene underwent hydrogenolysis reactions on the surface of the catalyst, producing methane and ethane. However, the  Effect of Hydrogen Spillover on TPH of Coked  i3  cd fl  a  JS o  A  o  13  ft o  cd  CJ  9  _b£>  w  a  'o  M  CM  CO  41  41  o o  o o  i—1  CN  CO  i—i  41 I—1  C5  o  o  a  cu fl oj  co  o  HJ  w  A! O  'ft  13  s  41 r-  cd 1—1  CO CN  o co  41 t-  CN  o <u o  CO  co CU  cu  41  a  OJ  ^ft I  ft  (H  a  tH  PH  o  41  H o  CO  s  CM  O  fl  a  cd , d  H-=  3. O  co  41 oo  CO CO CM  o co  CM  CM  1—1  i—I  41  co  CM CO CM  CM  41  CO CM  41 CO CO CO  .a a Id Cd CJ  • *H CO  <HH  O fl fl  .  OJ  id' t-5  CU  fl  cr cu  1  CO  X  fa  fl o cci  o  13  cu X  fl o  cu  AS  o  o  ft  H-3  o  fl  Cd  Cd  .—.  a  A  CJ  CJ  13  <u oC J  * d  cu  cu  o  O CJ  <H-H  O  CJ  fl o  fl O  • —H  • f—i HJ  Id  fl cu  hO tH O  • —H  cd fl cu  H-=  13 >,  13  cd fl <u  bO O tH  13  bO O tH  >, J=l 13 cu  13  a a  CU  a acd  cu  afl  cd  CU  CJ  cd  "73  o  cu "o cj fl o tH  cu  xi  A  cd  buOO i-i  ft CO  ft  13  ogr  u fl  tH  ft  ft  tH  a  tH  bO O  ft  SH  fl cd u cu  ft  a  ,.—.,  CM  tH ft  I  1  Te:  cu  -d cu PQ -d <u  CU  'cd cd o  rod  •d  CU  CO  a  1-4  cd buO cu  CU  fl  ^3  cd  fl  cd fl  A  41  41  00  41  o co  CJ  HJ  fl  bO 41 f*  OJ  H-= cd  CN  cu  ft CU Xi  CU  xi  CM  3,  o  41  o  CO CU  on  'o  cd  fl  l-l  oj fl  CO  CU CJ  fl  cu  4) cn  Q  'cd H-= cd cu  CJ  41  CU  AS o  CO  •73  IO CM  o  cd O  O  41  CM  CM  a  >, T3  cd  o  ft o  H^ CO  to  41  o  CJ  H-= i—1  •8  cu  i3  CO  3.  _bO -H  "cd  13  cu o  O O  CO  >,  fl o CU  53  Silica-Alumina  CU cd  a fl  A  'ed  cd  o  I  tH  CU  fl  H^  cd  CU  13  cu  X  ft  a  OJ CU  >  tH fl  H-2  cd tH  CU  ft  a  Effect of Hydrogen Spillover on TPH of Coked  Silica-Alumina  54  1.60E-02 n  1.40E-O2 -\  Catalyst on top of coke, R98-2 Catalyst & coke were mixed, R113-2 —  - Coke on top of catalyst, R119-2  Figure 4.7: Effect of Position of Coked Silica-Alumina and Catalyst on Production source of propylene production was unknown. Propylene might be produced from coked silica-alumina and/or it might sorb on coked silica-alumina during the cooling period in the nitrogen glove box. However, the coked silica-alumina was purged at room temperature prior to hydrogenation and the temperature corresponding to the maximum production of propylene was rather high (about 243°C) for physical adsorption. Hence it seemed that chemisorption of propylene might have occurred . The other possible source was molecules of propylene trapped inside the pore structure of coke and/or silica-alumina. However, given the pore size of coked silica-alumina (about 50A, refer to Appendix M ) this option is not likely. The source of ethane and ethylene production seemed to be propylene, decomposition. Methane production occurred in two regions of temperature; the low temperature region corresponding to the production of propylene, ethylene and ethane whereas at high temperature (>500°C) only methane was produced. Hence it  55  Effect of Hydrogen Spillover on TPH of Coked Silica-Alumina  seemed that propylene decomposition mostly resulted in methane production at low temperature whereas the reaction of coked silica-alumina with hydrogen resulted in methane production at high temperature. Therefore, to investigate the role of hydrogen spillover and propylene decomposition in hydrogenation, several experiments were carried out in which the coked silica-alumina underwent temperature-programmed reaction ( T P R ) i n argon prior to temperatureprogrammed hydrogenation (TPH). As a result of the T P R , the volatile matter present from the coking step was removed. After T P R , the sample underwent T P H in the presence and absence of catalyst as before. Methane was the only product of hydrogenation in this reaction sequence as shown in Table 4.4.  3.00E-03 i  ,  2.50E-03 -  '  2.00E-03 •  •  o I ©  1  s  o g  • ' \  1.50E-03-  o <o u. •§  * • , 1.00E-03-  •  J  S 5.00E-04 -I  ',  J**  0  -**  100  200  —^  300  400  500  600  700  Temperature, ° C - - - Thermal in hydrogen; after pretreatment in argon, R104-2  Figure 4.8: Methane Production from Coked Silica-alumina After T P R in Argon Although the coked silica-alumina was aged during T P R in argon, which is known to reduce the reactivity of coke, it still reacted to produce methane.  Again, methane  production in the presence of catalyst was higher than from non-catalytic (thermal) hydrogenation, confirming the reaction of spilt-over hydrogen with coked silica-alumina.  Effect of Hydrogen Spillover on TPH of Coked  56  Silica-Alumina, J. u 13 CD  A)  O CJ  CU  11 1-fl  a  fl  <i cci  -1-3  PH E-I  CN  tH  .bO | t—I -H cu "o fl  11  CU  HJ  CO  :>> 13 bO 13 CJ !3 <HH o  fl o  O  CO  13  tf  CD  CJ  « »-H  CO  li  w  S3  o  -H oo oo  H^>  Ji  H  o E  o  Ofl  CN  o  tf  •8  bO |  A  0)  O tH  13  «fl o  CU  o  cu  Ii  fl  OH  O  CM  -H o co  I  tH  fl o bO  13  tH  PH  h a E  tH  CU  •-H  H  O  a  CM  A  OH  fl tH  13  >,  «  H-3  cj ctt  bO I 0)  fl .Xl  O  _ i  0 0  a  -H  A  li  13 i> o  H-3  CU  cc3 fl  o CN  I  tH  tf  PH  13 cu AS o CJ  <HH  fl  CM  <  CJ  O  -H oo tH  cu  cci  fl  tH  PH CU  cci fl  a a 13 k CJ  13 cii o  13  CO  CO bO O  CU  A  A  -J3 CJ I fl o 55  13 cu X  o  -*^» CJ cci  cu bO 13 tH tH cu OH H-3  o  fl  13  o IH  PH  o  ao tH  CJ  am me  bO  •  PH H  CO  cu CJ fl CD CO CD  tH  13  cu o  13  CJ  cu o  fl .2  fl .2  "H-=  cci  fl CU bO O tH  13 >> M 13  tH  CJ  H-3  cci  fl  CD bO O IH  13 >> 13 X(  am me  PI  co _b0 | o -H  am me  cci  fl O bO  b OO  b OO  PH 1  PH  PH  1  tH  I1  1  CU  tH  cci  fl  tH  H-i cci  H-3  CD P,  tH  CU P,  tH  CU P,  fl H-3 tH  fl fl  fl fl  HCD H  .m .  X  cu tH cu  tH  CD  tH  CU  tH  b OO tH  13  CU  fl  cci G  C CU ,—^  cci cci  J> X>  13 fl  »—1  o  cci cci fl  a  13 fl  cci  o  OH OH CCi O  57  Effect of Hydrogen Spillover on TPH of Coked Silica-Alumina  3.50E-03 -i 3.00E-O3 o c I  2.50E-03-  2 "5 c o  2.00E-03 -|  o o  1.00E-03  5.00E-04 •{ O.OOE+00  100  200  300  400  500  600  700  Temperature, ° C  - - - Catalytic hydrogenation; after pretreatment in argon, R118-2  Figure 4.9: Methane Production from Coked Silica-Alumina in the Presence of Catalyst; After T P R in Argon; Catalyst and Coked Silica-Alumina were mixed However, as shown qualitatively in Figure 4.8 and Figure 4.9, due to coke aging, methane production dropped in comparison to experiments in which the T P R in argon was omitted. Hence it is clear that coked silica-alumina and spilt-over hydrogen indeed reacted to produce methane. However, propylene decomposition was another source of methane production, specially for methane production at low temperature (about 250° C) .  4.3.4  Effect of Ceramic and Silica—Alumina on Temperature—Programmed Hydrogenation of Coked Silica-Alumina  The main purpose of this study was to investigate the reaction of spilt-over hydrogen with coked silica-alumina and to determine the possibility of transport of these species through the ceramic by gas phase transport and/or secondary spillover. However, due to the complex nature of coke, many experiments had to be carried out to clarify the effect of hydrogen spillover prior to the experiments discussed in this section. This section deals with those experiments in which the coked silica-alumina and catalyst were separated by a thin layer (about 5mm) of crushed ceramic (particle diameter 90-15G7/m). Since  Effect of Hydrogen Spillover on TPH of Coked Silica-Alumina  58  the surface area of the crushed ceramic is very low, some tests were performed with the ceramic layer replaced by high surface area, dried silica-alumina (Appendix M ) . Figure 4.10 and Figure 4.11 compare the results of these experiments qualitatively with the non-catalytic (thermal) and catalytic (catalyst placed on top of coked silica-alumina) experiments. Table 4.5 summarizes the results from these experiments. As shown in Figure 4.10 and Figure 4.11, when the coked silica-alumina and catalyst were separated by a layer of ceramic or dried silica-alumina, methane production was higher than the non-catalytic experiments, confirming the transport of spilt-over hydrogen through the ceramic and the dried silica-alumina. However, there was a decrease in methane production in comparison to the catalytic experiments (Table 4.5), where the catalyst was placed directly on top of the coked silica-alumina, suggesting the loss of spilt-over species. Also, the decrease in methane production due to the presence of dried silica-alumina was less than the crushed ceramic, suggesting that dried silica-alumina was a better medium for the transfer of spilt-over hydrogen. Propylene, ethylene and ethane production were approximately constant for these experiments.  4.3.5  Carbon Recovery  Table 4.6 summarizes the total amount of carbon deposited on the surface of dried silicaalumina during the coking step and the percentage of this carbon recovered during the course of temperature-programmed reaction. The average carbon content of the coked silica-alumina was 5.1±0.66 wt.%, as measured by Canadian Microanalytical Service Ltd . The amount of recovered carbon was calculated based on methane, ethane, propylene and ethylene production and was divided by the total amount of deposited carbon to obtain percentage of carbon recovery. The detailed calculations are gathered in Appendix H . As shown on Table 4.6, there was no significant difference in the percent carbon recovery between different non-catalytic and catalytic temperature programmed hydrogenations of coked silica-alumina. In the latter case, different positions of coked silicaalumina and catalyst, or the presence of an intermediate layer of crushed ceramic or dried silica-alumina, did not make any significant difference to the percentage carbon recovered. However, as discussed in previous sections, the distribution of products were  Effect of Hydrogen Spillover on TPH of Coked  Silica-Alumina.  3.00E-03 n  2.50E-03  c  to £ 2.00E-03 A v 1  g  1.50E-03  o S « o  1.00E-03  5.00E-04  •  0.00E+00 0  •  -  100  200  300  400  500  600  Temperature, ° C • - - Thermal in hydrogen, R97-2 Catalyst on top of coke, R98-2 Ceramic between coke & catalyst, R100-2  Figure 4.10: Effect of Ceramic on Methane Production of Coked Silica-Alumina  60  Effect of Hydrogen Spillover on TPH of Coked Silica-Alumina  3.00E-03  n  2.50E-03  ?  2.00E-03  g o  1.50E-03  CO  £ o  1.00E-03  5.00E-04  0.00E+00  J  — « * 100  200  300  400  500  600  Temperature, ° C Thermal in hydrogen, R97-2 •Catalyst on top of coke, R98-2 •Silica-alumina between coke & catalyst, R115-2  Figure 4.11: Effect of Dried Silica-Alumina on Methane Production of Coked SilicaAlumina  61  Effect of Hydrogen Spillover on TPH of Coked Silica-Alumina 13  - d  cd o  CD  CD A! O  AH O  fl  fl  CJ  13 CD Ai  £ H-=  CJ  A  <  CD  u  _b0  a  A  a.  CO  CD  o  o  CD fl <D  a  fl  .2 %J  CN O CD  -H <N CD  o  w  o fl  h a  S  13  o  i—< i—<  CD  -H lO co CD  CN  too o  AH  o  T—I  +1 to Tf co o -H t-  CN  Tf  CD  to CD CD  -H Tf  CN CD  Tf  -H i—I  CO  CN CD CD  OH O H-=>  00 i—1  O  CM  QJ  s OJ  fl o  cd fl  a  toO •  OH  o tH  PH  A  I  O  CD fl CD ;  a  Oi  -H Oi  oo i—1  -H CN  o> T-H  fl  CD co CD  1-  O H C  CM  Tf  co  oo  -H Tf  -H t—  CN  -H CN lO CN  _bfl  P 13  CD  s  A  "o a.  a  9  a  cd tH CD  o CJ  fcfl w Tf  CD  i—i  -f)  CD  i—i  i-H - CO  lO  h c  to  Tf  EH  CO CM  -H Oi CM  -fl oo CN  EH  a  lO  to Tf  'cd'  CM  cd fl  a  -H  13  CM  fl o cd  o  nd  .2 "tH  Q  I tH  OJ  O  13 H-=  cd  o I  fl o  CD P.  H-3  CJ fl  13  O  tH  PH  CD Ai  CD Ai O o  O  tJ *+H  a a  bO  cd  O tH PH  tH  bO O  tH  CD  PH  fl tH  I  CD tH fl  cd  cd cd fl  tH  CD  a a a  EH  'cd  OH  EH  fl  <u  tH  bO  13  O  i?  >, A CD  cd  O  13  13  H-3  bO  tH  tH  CD PH  cd fl CD  O  cd  ao  H^>  hO  a a  tH  • «H  cd fl CD  CD  fl o  O  H-3  >, A  O  fl  • *H  13  15  CJ  13  13  fl o  tH  H-=  tH  Cd CJ  «HH  I  CO  A  13  Cd CJ  cj  toO 13 H J  a  13  o  CJ  O  tH  A  cd fl  a  Ai  O  fl CD bO  Tf  a  13  <D Ai  cd  co  cd fl  Cd o  .2  cn  cd fl  fl 13  •d  i  < CO  H-=  <D  tH  PH  CD  A  A  fl  CT1  OH  OH  -H  t~ CM  tH  tH  CM  -fl  CD cj fl CD co <D  cd cj  CD  A  fl  oi 6  co  +1 -H  CD cj  H-»  CO  CD  CJ  13  13  CD  OH  CJ  CO  co  cd HJ cd CJ  fl  tH  OI  CD  o  CD CJ  CO CN  tH  CD  cd CJ  00  i o  tH  CO  -H  to  3  I  o  I  OH  CO  13 cd  CO  -H  i  'OH  cd fl  E  OJ  o  cd  to -H to  -fl  13 CD CJ cd  CD  CD  tH  H-5  A  13  fl  -H  1 — < r-  +1 -H r—I 00 to to CN  PH  rJfl  CD  CD  A  a  8  CD CD  CD CD  O  cd  CJ  tH  13  >, A  13  co >>  t I  cd Id cj 13 fl  cd cd fl  a "cd  cd cj  a a  CD  cd tH  bO O  tH  OH I  CD tH fl H—'  cd  cd H-=  cd o  13  fl cd cd fl  'a  tH  j*  OH  Id cd  CD  a EH  a  13  o  CD cd tH  bO O  tH  PH  J, tH fl H-=  cd tH  CD  OH  a EH  Effect of Hydrogen Spillover on TPH of Coked  62  Silica-Alumina  different for different experiments. There was no significant difference in carbon recovery percentage for different T P H experiments of coked silica-alumina due to the large contribution of hydrogenolysis reactions and small contribution of spilt-over hydrogen in methane production. The latter was in the range of error of measurements. Table 4.6: Carbon Recovery in Different Experiments Temperature-Programmed Hydrogenation  Total Deposited  wt.% Carbon  of Coked Silica-Alumina  Carbon (fimol)  Recovery  were mixed  900±3  9±2  Non-catalytic (Thermal)  898±6  9±2  900±4  10±1  892±3  8±2  903±4  10±2  834±29  11±0  900±4  14±3  898±0  0.2±0.1  903±4  1±0  Catalyst & coked silica-alumina  Catalyst was placed on top of coked silica-alumina Ceramic was placed between coked silica-alumina & catalyst Dried silica-alumina was placed between coked silica-alumina & catalyst Coked silica-alumina was placed on top of catalyst Temperature-Programmed Reaction of Coked Silica-Alumina in Argon T P H W after TPR( > in argon 2  Non-catalytic TPH* ) after T P R 1  ( 2 )  in argon  Catalyst & coked silica-alumina were mixed  Temperature-programmed hydrogenation of coked silica-alumina. (> Temperature—programmed reaction of coked silica-alumina. 2  Effect of Hydrogen Spillover on TPH of Coked Silica-Alumina 4.3.6  63  Test for Plug Flow  In Section 4.3.4 it was concluded that hydrogen spillover occurred based on the methane produced in a layered bed consisting of catalyst, ceramic membrane and coked silicaalumina. This conclusion is valid provided the possibility of back diffusion of volatile organics, produced from the coked silica-alumina can be excluded. Otherwise these organics could undergo hydrogenolysis reaction on the catalyst as illustrated in Figure 4.12(b). To determine the degree of backmixing in the reactor, and to obtain experimental evidence that back diffusion and subsequent reaction of organic did not occur, the following experiments were completed. Several tests were carried out in which the residence time distribution was measured using a pulse response method. The results were analyzed according to the method explained by Fogler (1992) and resulted in an average dispersion number=0.018 at 601°C, showing that the gas pattern in the reactor is close to plug flow. In Appendix L , details of these tests are gathered. To investigate the effect of deviation from plug flow on temperature-programmed hydrogenation of coked silica alumina, an experiment was performed in which the catalyst was elevated by means of quartz wool. In this way, there was about a 1cm gap between the coked silica-alumina and the quartz wool, and the height of the quartz wool was about Icm. Consequently, the distance between the coked silica-alumina and catalyst was about 2cm. Table 4.7 compares the results of this experiment with non-catalytic and catalytic experiments. The results of this test for the case in which there was a gap between the catalyst and coked silica-alumina were similar to those of non-catalytic experiments and quite different from the catalytic experiment in which the catalyst was placed directly on top of coked silica-alumina. Hence based on these experiment, it can be concluded that the deviation of plug flow in this system was not significant.  Effect of Hydrogen Spillover on TPH of Coked Silica-Alumina, <D  cd  Xi  d  0  <+H  o  d  "cd  PH  O  cd CJ  H-3 d  o nd  _b0 |  r d  Ii  W  CM O  oo o-H CM  00  a;  o o 41 oi  o  <  _b£> |  "o  to d  cu <u  CM  U  cd  cu  oo CO  l l  <d  •CO«—<  CO  d  X  d  0  "pH  CJ  CO •  o ^td  CM  X\  O  H-=>  T3 O O  i—t  CO CO  IO CO  'cd cd O  a  41 -H  CO  •a  "Id cd o  o  a E  JH  CO  oo -H -H m oo CN co IO CM  cd O  0)  a;  o  CM  CJ  d  cu  d  cu  CD  CO  d  i d  o  PH CD , d  'o  cu d cu  I  CM  -H  M  OI  "tf "tf  tH  E  CM  H  _b0|  H-=  CU  'o  •o -H  CM  CO  CM  d  -H co -tf  "tf .-<  "2 41 41 CO  t>-  f-H  CM iO  l l  0  0  'cd  "cd  d  ^ d  1 cd cj  cd cj  CM  d  r d  H-=  cd  PH  CU  , d  -»-=>  l l  o  o  (H  PH <U  V  CM  a;  o  -H  CU  PH  ed d  0  nd  O  o o  co  Oi  Oi  CO "tf CM  CO CO CM  -H 41  Cd  o  (H  bO  J>> OH cd  o  u  cd  O I d  o  .2  "ed* H-^» cd  o  PH  -»-= o d  - d  o  SH  PH  O  cd  d  d  O  b OO  cu  u  ( H  0 o SH CJ  X\  CU cd  O O <H-H  o d  H^>  cd d  (U  bO  o (H  Xi  x\  n  bO SH  O n I  CU  PH |  PH 1  SH C dU H-=  cd l-H  Id  H-3  cd CJ  S-H  h OO  1  CO  -d  b OO SH  PH  0  CD  id nd >>  nd >->  cd n cu PH  T3  .2 .2  cu  "cd  H-=  d  cu bO cd  cd H  'co  o  d  cd  0  "co nd  C+H  Id  cd o  r d  o o  O CJ  d  Id  i d  cu  CD  r M  0  am me  t>  CO  Oi  ina,  .2 "-u  CO  U  am me  d  OI *—I  ina  CU  CU PH d czJ  SH  I  CD SH  d - p  cd SH  C D PH d  Effect of Hydrogen Spillover on TPH of Coked 4.4  65  Silica-Alumina  Conclusion of Effect of Hydrogen Spillover on Temperature-Programmed Hydrogenation of Coked Silica-Alumina  To study the effect of hydrogen spillover on model catalytic reactions and to achieve separation of reactant and catalyst by the ceramic material used in Chapter 3, dried amorphous silica-alumina was coked by propylene and used as the reactant of the hydrogenation reaction. The average carbon percentage of coked samples was 5.1±0.66%. Due to the complex nature of coke, temperature-programmed hydrogenation (TPH) of coked silica-alumina in the presence and absence of catalyst was performed. For the T P H experiments in the presence of catalyst, the effect of relative position of coked silica-alumina and catalyst in the fixed bed was investigated. To clarify the effect of hydrogen spillover in the reaction of coked silica-alumina with hydrogen, coked silica-alumina underwent temperature-programmed reaction ( T P R ) in argon to remove volatile matter. Subsequently, the sample underwent T P H in the presence and absence of catalyst. The effect of separation of coked silica-alumina and catalyst by a layer (about 5mm) of crushed ceramic (the same material used in Chapter 3), or dried silica-alumina, were investigated  H + Ar  H + Ar  2  2  : Catalyst | —  • Catalyst j Coked |_ Silica-Alumina :  1  4=  J  Crushed : Ceramic  Spilt-over Hydrogen  (a)  I..........'  Coked Silica-Alumina  t  T-  Crushed Ceramic  Back Diffusion of Volatile Matter of Coked Silica-Alumina  (b)  Figure 4.12: Possible Flow Direction in the Reactor Bed  Effect of Hydrogen Spillover on TPH of Coked Silica-Alumina  66  in order to examine the possibility of transfer of spilt-over hydrogen on ceramic and dried silica-alumina. The gas flow pattern in the reactor was studied by residence time distribution using a response method. A T P H experiment was performed with a gap of 2cm between coked silica-alumina bed and catalyst to clarify the effect of gas flow pattern on the T P H reaction. The observations and conclusions from the experiments reported in Chapter 4 are summarized as follows: • Temperature-programmed hydrogenation of coked silica-alumina resulted in methane production in two regions. One at low temperature (243°C) and the other at high temperature (> 500°C). Other products (propylene, ethylene and ethane) were detected in the low temperature range. • There was a clear distinction between the amount of methane produced in catalytic and non-catalytic experiments. • Changing the relative positions of coked silica-alumina and catalyst in the fixed bed resulted in different product distributions. Methane and ethane production followed the opposite trend of propylene production, suggesting that some of the propylene underwent hydrogenolysis reactions to produce methane and ethane. Ethylene production was not affected by different positions of catalyst and coked silica-alumina. • Propylene production could be attributed to the coke deposit, physical and/or chemical adsorption during the coking reaction and/or the cooling period during coking. Given the temperature corresponding to maximum propylene production (about 250°C), it seemed that propylene physical adsorption was not likely. • T P H of samples which underwent T P R in argon resulted in methane production. Hence it was shown that coked silica-alumina reacted with spilt-over hydrogen and this reaction was a source of methane production. In these experiments, methane production in the presence of catalyst was higher than without catalyst.  Effect of Hydrogen Spillover on TPH of Coked Silica-Alumina  67  • Sepaxation of the coked silica-alumina and catalyst by a layer of crushed ceramic or dried silica-alumina resulted in reduction of methane production. The amount of methane produced in catalytic (with and without separation by ceramic or dried silica-alumina) and non-catalytic experiments followed the trend: catalyst on top of coked silica-alumina> dried silica-alumina between coked silica-alumina and catalyst > ceramic between coked silica-alumina and catalyst > non-catalytic (thermal). Hence, it seems that dried silica-alumina was a better medium for transfer of spilt-over hydrogen than the ceramic material. • There was no significant difference in carbon recovery percentage for different T P H experiments of coked silica-alumina due to the large contribution of hydrogenolysis reactions and small contribution of spilt-over hydrogen in methane production. The latter was in the range of the error of measurements for total carbon recovery. • The tests of flow pattern inside the reactor showed that there was a small deviation from ideal plug flow. However, the results of experiment in which there was a gap between the coked silica-alumina and catalyst were similar to a non-catalytic experiment and different from catalytic experiments where the catalyst was placed on top of the coked silica-alumina. Hence, the deviation from plug flow was not significant.  Chapter 5 Conclusions &c Recommendations for Future Work  5.1  Conclusions  The main objective of this study was to investigate the possibility of reaction of spilt-over hydrogen, produced on catalyst sites, with organic reactant which was separated from the catalyst by a ceramic membrane. Two catalytic systems were investigated to meet this objective. In the first part of the study, hydrocracking diphenylmethane (DPM) in a ceramic membrane reactor was investigated, in which the catalyst and organic reactant were separated by a ceramic membrane tube. The results of the experiments showed that separation of reactant and catalyst enhanced the product yields (benzene and toluene). However, contact between liquid and catalyst could not be eliminated by the porous ceramic membrane. Therefore, the extent of spillover in the ceramic membrane tube could not be quantified. To achieve complete separation of the reactant and catalyst, another catalytic system with solid reactant was studied. Since coke is solid and reacts with spilt-over hydrogen to produce methane, coke, was chosen as the reactant. To produce coke, amorphous silicaalumina was coked by thermal reaction of propylene. The coked silica-alumina underwent temperature-programmed hydrogenation in the presence and absence of catalyst. In the former case, the effect of separation of coked silica-alumina and catalyst by a layer of low surface area crushed ceramic (used in the first part of the study) on hydrogen spillover was investigated. Additional experiments were performed in which the catalyst and coked silica-alumina were separated by a layer of high surface area, dried amorphous silicaalumina. Based on these studies, the following conclusions can be drawn: • Based on experiments in the presence and absence of catalyst, it was shown that 68  69  Conclusions & Recommendations for Future Work  spilt-over hydrogen improved the yields of products for both systems under study. • For hydrocracking diphenylmethane in a ceramic membrane reactor, the effect of hydrogen spillover on ceramic material could not be quantified since it was not possible to eliminate the contact between reactant and catalyst. • For temperature-programmed hydrogenation of coked silica-alumina, the effect of hydrogen spillover on methane production was small. Methane production due to reaction of spilt-over hydrogen was about 1% of total carbon deposited on the surface of silica-alumina, whereas the total carbon recovery was about 10% of total deposited carbon on silica-alumina. • Based on T P H of coked silica-alumina in the presence of low surface area crushed ceramic, and high surface area silica-alumina, it can be concluded that a high surface area medium is more suitable for separation of reactant and catalyst.  5.2  Future Work  According to the results obtained, the following recommendations are suggested to improve the experimental technique and provide further insights into the phenomenon of hydrogen spillover: • As explained in Section 4.2.1, to avoid coke aging, the coked samples were placed in a nitrogen glove box at the end of the coking period. As a result, volatile matter present in coked silica-alumina reacted during temperature-programmed hydrogenation (TPH), complicating the reaction system under investigation. Hence it seems appropriate to study the effect of purging coked samples in an inert gas at coking temperature for different amounts of time on T P H profiles. • The catalyst for hydrocracking diphenylmethane was sulfided N i - M o / A l 0 3 whereas 2  for T P H of coked silica-alumina 2 % C o - S i 0 was used. Different catalysts provide 2  different amounts of spilt-over hydrogen. Hence it seems appropriate to investigate the effect of catalyst components (both metal and support) on hydrogen spillover and reaction with coked silica-alumina samples.  Conclusions k Recommendations for Future Work  70  • Medium of separation has a great impact on spillover. Therefore, the effect of medium component, surface area, presence of different groups such as hydroxyls and thickness should be investigated in T P H of coked silica-alumina. The above suggestions will provide information to choose the optimum catalyst and medium of separation for T P H of coked silica-alumina. It would be appropriate to use this information to modify the membrane reactor apparatus used in hydrocracking diphenylmethane to improve the product yield obtained in the present study.  Bibliography Adris, A . M . , C. J . Lim, and J . R. Grace (1994). The fluidized bed membrane reactor system: A pilot scale experimental study. Chem.Eng.Sci. ^P(24B), 5833-5843. Baumgarten, E . , C. Lentes-Wagner, and R. Wagner (1989). Hydrogen spillover through gas phase transport of hydrogen atoms. J. Catal. 117, 533-541. Baumgarten, E . and A . Schuck (1997). Investigations about coke burning and oxygen spillover. React. Kinet. Catal. Lett. 61(1), 3-12. Bhave, R. R. (1991). Inorganic Membranes synthesis, characteristics and applications. New York: Van Nostrand Reinhold. Cannon, K . C . and J . J . Hacskaylo (1992). Evaluation of palladium impregnation on the performance of a Vycor glass catalytic membrane reactor. J. Membr. Set. 65(3), 259-268. Chan, N . (1996). Asphaltene adsorption on ceramic membranes. B.A.Sc. Thesis, University of British Columbia, Vancouver, B . C . , Canada. Chu, X . and L . D . Schmidt (1993). Processes in M o S gasification. J. Catal. 144, 77-92. 2  Conner Jr., W . C. (1988). Spillover of hydrogen. In Z. Paal and P. G . Menon (Eds.), Hydrogen Effect in Catalysis: Fundamentals and Practical Applications (In Chem. Ind. Vol. SI), pp. 311-346. Marcel Dekker Inc., New York, N . Y . Conner Jr., W . C. and J . L . Falconer (1995). Spillover in heterogeneous catalysis. Chem. Rev. 35, 759-788. Conner Jr., W . C , G . M . Pajonk, and S. J . Teichner (1986). Spillover of sorbed species. Adv. Catal.  34(1),  1-79.  Delmon, B . (1993). New aspect of spillover effect in catalysis. In T. Inui, K . Fujimoto, T. Uchijima, and M . Masai (Eds.), The Control of Selectivity and Stability of Catalyst, Stud Surf Sci Catal, pp. 1-8. Elsevier.  71  72  BIBLIOGRAPHY  Fogler, H . S. (1992). Elements of chemical reaction engineering (Second ed.)., Chapter 13 and 14. Prentice-Hall Inc. Gryaznov, V . M . (1992). Platinum metals as components of catalyst-membrane systems. Plat. Met. Rev.  36(2),  70-79.  Hattori, H . , K . Yamshita, T . Tanabe, and K . Tanabe (1988). Catalytic features of iron, nickel, and molybdenum catalysts in hydrodenitrogenation and hydrocracking of model compounds of coal liquids. In M . J . Phillips (Ed.), Catalysis, Theory to Practice: Proceedings, & International Congress on Catalysis. h  Hsieh, H . P. (1991). Inorganic membrane reactors, a review. Catal. Rev.-Sci. Eng. 33(12), 1-70. Ioannides, T. and X . E . Verykios (1993). The interaction of benzene and toluene with rhodium dispersed on silica, alumina and titania carriers. J. Catal.  143(1), 175-186.  Kapoor, A . , R. T. Yang, and C. Wong (1989). Surface diffusion. Catal. Rev.-Sci. Eng.  31(1-2),  129-214.  Karroua, M . , P. G . Matralis, and B . Delmon (1993). Synergy between NiMoS and C o S in the hydrogenation. J. Catal. 139, 371-374. 9  8  Khoobiar, S. (1964). Particle to particle migration of H atoms on P t / A l 0 3 catalysts. 2  J. Phys. Chem.  68(12),  411-412.  Khoobiar, S., J . L . Carter, and P. J . Lucchesi (1968). The electronic properties of aluminum oxide and the chemisorption of water, hydrogen, and oxygen. J. Phys. Chem. 72(5), 1682-1688. Kieken, L. and M . Boudart (1992). Oxidation of Co on P d particles on o>Al 03; reverse 2  spillover. In L. Guczi, L. Solymosi, and P. Teteny (Eds.), New Frontiers in Catalysis: Proceedings of the 10 International Congress on Catalysis, pp. 1313-1324. Elsevier, th  Science Publishers B . V . K i m , S. and G . R. Gavalas (1995). Preparation of H -permselective silica membranes 2  by alternating reactant vapor deposition, hid. Eng. Chem. Res.  34(1),  168-176.  Kitao, S., M . Ishizaki, and M . Asaeda (1991). Permeation mechanism of water through  73  BIBLIOGRAPHY  fine porous ceramic membrane for separation of organic solvent/water mixtures. Key Eng. Mater. 61-62, 175-180. Lenz, D . H . and W . C. Conner Jr. (1987). Hydrogen spillover on silica: Ethylene hydrogenation and H - D exchange. J. Catal. 104, 288-298. 2  2  Lenz, D. H . , W . C . Conner Jr., and J . P. Fraissard (1989). Hydrogen spillover on silica. I l l Detection of spillover by proton N M R . J. Catal 117, 281-289. Levy, R. B . and M . Boudart (1974). The kinetics and mechanism of spillover. J. Catal. 32, 304-314. Mills, G . A . , H . Heinemann, T. H . Milliken, and A . G . Obald (1953). Catalytic mechanism. Ind. Eng. Chem.  45(1),  134-137.  Mohan, K . and R. Govind (1986). Analysis of a cocurrent membrane reactor. AIChE J. 32(12), 2083-2086. Mohan, K . and R. Govind (1988a). Analysis of equilibrium shift in isothermal reactors with a permselective wall. AIChE J.  34(9),  1493-1503.  Mohan, K . and R. Govind (1988b). Studies on a membrane reactor. Sep. Set. Technol. 55(12-13), 1715-1733. Oertel, M . , J . Sclunitz, W . Weirich, D. Jendryssek-Neumann, and R. Schulten (1987). Steam reforming of natural gas with integrated hydrogen separation for hydrogen production. Chem. Eng. Technol. 10, 248-255. Parera, J . M . , E . M . Trafanno, J . C. Musso, and C. L. Pieck (1983). Hydrogen and oxygen spillover on P t / A l 0 3 during naphtha reforming. In G . M . Pajonk, S. J. 2  Teichner, and J . E . Germain (Eds.), Spillover of Adsorbed Species, Stud Surf Sci Catal, Volume 17, Amsterdam. Elsevier. Peden, C. H . F . and D . W . Goodman (1986). Model studies of C u / R u bimetallic catalysts. Ind. Eng. Chem. Fundam. 25, 58-62. Rodriguez, N . M . and R. T. K . Baker (1993). Interaction of hydrogen with metal sulfide catalysts- direct observation of spillover. J. Catal. 140, 287-301.  74  BIBLIOGRAPHY  Roessner, F . and U . Roland (1996). Hydrogen spillover in bifunctional catalysis. J. Mol.Catal. A: Chem. 112, 401-412. Roland, U . , T. Braunschweig, and F . Roessner (1997). On the nature of spilt-over hydrogen. J. Mol.Catal. A: Chem. 127, 61-84. Roland, U . , H . G . Karge, and H . Winkler (1994). Hydrogen and deuterium adsorption on zeolite supported platinum. Evidence for hydrogen and deuterium spillover. In J . Weitkamp, H . P. Karge, and W . Holderich (Eds.), Zeolites and Related Micro-porous Material: State of the art. In Stud Surf Sci Catal, Volume 84, pp. 1239-1246. Elsevier, Science B . V . Rozanov, V . V . and O. V . Krylov (1997). Hydrogen spillover in heterogeneous catalysis. Russ Chem Rev 66(2), 107-119. Saracco, G . and V . Specchia (1998). Inorganic-membrane reactors. Chem Ind 71, 463500. Satterfield, N . C. (1980). Heterogeneous Catalysis in Practice. McGraw-Hill Inc. Shaw, D . J . (1992). Introduction to colloid and surface chemistry (Fourth ed.). Redwood Books, Trowbridge, Wiltshire: Butterworth Heinmann. Soltan Mohammad Zadeh, J . S. (1998). Methane homologation by the two-step cycle on Co catalysts. Ph. D. thesis, University of British Columbia, Vancouver, B . C . , Canada. Stumbo, A . A . , P. Grange, and B . Delmon (1995). Spillover hydrogen effect on amorphous hydrocracking catalysts. Catal. Lett. 31, 173-182. Stumbo, A . M . , P. Grange, and B . Delmon (1997). Creation of acidic sites by hydrogen spillover in model hydrocracking systems. In G . Froment, B . Delmon, and P. Grange (Eds.), Hydrotreatment and Hydrocracking of Oil Fractions, pp. 225-235. Elsevier, Science B . V . Traffano, E . M . and J . M . Parera (1986). Re influence on hydrogen spillover on PtR e / A l 0 . Appl. Catal. 28, 193-198. 2  3  Uhlorn, R. J . R., K . Keizer, and A . J . Burggraaf (1992). Gas transport and separation  BIBLIOGRAPHY  75  with ceramic membranes, Part I. Multilayer diffusion and capillary condensation. J. Membr. Sci. 66(2-3), 259-269. Washburn, E . W . , J . W . Clarence, and N . Ernest Dorsey (Eds.) (1929). International critical tables of numerical data, physics, chemistry and technology (First ed.), Volume IV. McGraw-Hill Book Company, Inc. Weisz, P. B . and E . W . Swegler (1957). Stepwise reaction on separate catalytic centers: Isomerization of saturated hydrocarbons. Science 126, 31-32. Wolf, E . E . and F . Alfani (1982). Catalysts deactivation by coking. Catal. Rev.-Sci.  Eng. 24(3), 329-371.  Appendix A Calibration of Gas Chromatograph  Table A . l shows the components of typical calibration mixtures. Decane was used as internal standard and 1 m L of it was added to all samples for analysis. The calibration graphs for these mixtures are plotted in Figure A . l and Figure A.2. Figure A.3 shows the list of gas chromatograph program and result of the anaylsis for calibration sample #18.  76  77  Calibration of Gas Chromatograph  Table A . l : Gas Chromatograph Standard Calibration Mixtures and Retention Time of Different Compounds  sample # 17 18 19  vol. Benzene micro litre 600 1200 1000  vol. Toluene micro litre 800 1200 1000  vol. DPM micro litre 5000 5000 5000  vol. C10 micro litre 1000 1000 1000  Sample # 17 18 19  Benzene(b) mqmole 6.8 13.5 11.3  Toluene(t) mqmole 7.6 11.3 9.4  DPM mqmole 30.0 30.0 30.0  Decane(dO) mqmole 5.1 5.1 5.1  Sample # 17 18 19  Benzene Area 212519 450024 412601  Toluene Area 336982 415572 376435  Decane Area 302304 233951 283115  DPM Area 2254643 1617534 1970272  Sample # 17 18 19  Retention Time Benzene 1.495 1.844 1.513  (min) Toluene 2.502 2.910 2.518  Retention Time Decane 4.706 5.351 4.728  (min) DPM 11.066 11.733 11.051  total vol. micro litre 7400 8400 8000 Mb/Mc10 Mt/McIO Mdpm/McIO 1.3 2.6 2.2 Ab/Ac10 0.70 1.92 1.46  1.5 2.2 1.8  5.9 5.9 5.9  At/AdO Adpm/Ac10 1.11 7.46 1.78 6.91 1.33 6.96  Calibration of Gas Chromatograph  Figure A . l : Gas Chromatograph Calibration Graph for Benzene  Calibration of Gas Chromatograph  Figure A.2: Gas Chromatograph Calibration Graph for Toluene  80  Calibration of Gas Chromatograph  VARIAN 3408 GAS CHRQSATKRAPH KETHOD 4 RUN 1113 TIKE 16:£8 61 CT 57 SAMPLE: RUH RODE: ANALYSIS OSCULATION TYPE: PERCENT  END  0  PEAK KO.  TII€ HffiE  RESULT CODE t!I OFFSET  1  1.944  16.5508 BB 6.6 e.eee  45B824  £  £.916  15.2836 BB 6.7 6.086  415572  3  5.351  8.6637 BB 9.6 6.688  £33551  4 11.733  59.4863 BB 1S.3 6.096  1617534  5 IE.823  6.6767 TS 6.688  £687  108.6808 6.688  £719169  CT  11 777  1E.823  CR UI  r i  CT  5.351 CR  TOTALS:  AREA COUNTS  DETECTED PEAKS: 5 REJECTED PEAKS: 6 fiKOUNT STANDARD: 1.6888888 RULTIPLIER: 1.6088888 DIVISOR: 1.6886886 KOISE: .8 OFFSET: -1 TIKE: RUH LOG EVENTS: 6.68 STATUS : INJECT DET B: FID ATT B: 16 RHS E" 8 A/Z B: YES COL T: 48"C INJ T: 250" C DET T: 356°C AUX T: 358 C UI: c 6.99 COL R: £8.6 C/iairi 2.24 til: 4 2.47 COL T: 79° C 3.56 Hi: 8 4.49 COL R: £6.e C/miri 5.47 COL T: 178"C 13.50 yi: 16 14.45 COL Ri 26.6'O'(dri £2.47 COL T: 336°C 32.5R STATUS El© C  1.844  3 T t T  t  ! i  CR INJECT FIO B 16X8 6.5 CM/M I K  0  c  i  1  Figure A.3: List of Program for Anaylsis of the Samples and Results of Gas Chromatography on Calibration Mixture. #18.  Appendix B Gas Chromatography of Feed (Diphenylmethane)  A mixture of 4mL of D P M and l m L of Decane were analyzed as shown in Figure B . l . The result of this analysis shows that there is no benzene or toluene in the feed. (The typical retention times of benzene and toluene are reported in Appendix A.)  81  Gas Chromatography of Feed(DPM)  Figure B . l : Results of Gas Chromatography on Feed ( D P M ) Sample # I repeat #1  AUG 1 ,1997  END  V*ARIAN 34133 GAS CHRO&ATOGRAPH tETHOD 4 RUN 1856 n«E 26:41 29 JAN 97 SAHPLEi RUN KOOE: ANALYSIS CALCULATION TYPE: PERCENT FEAK KO.  TIKE NAT'E  RESULT CODE WI OFFSET  1  4.558  12.2921 B8 11.6 6.638  339424  2  16.949  87.5493 BB 15.7 8.688  2417568  3  14.478  Q. 1585 BB 6.688  4377  166.6860 6.682  2761316  TOTALS:  AREA COUNTS  CT 21.475  14.478  t  £  DETECTED PEAKS: 4 REJECTED PEAKS: 1 AMOUNT STANDARD: 1.6838683 MULTIPLIER: 1.6668688 OIVISOR: 1.6888888 NOISE: 46.5 OFFSET: 7 TIKE: 6.68  1.68 £.48 4.50 4.92 9.48 11.64 14.56 22.48 32.56  RUN LOG EVENTS: STATUS INJECT DET B: FID ATT B: :6 RNG B: 8 A^Z B: YES Ul: 2 COL T: 46° C INJ T: 258'C DET T: 356" C AUX T: 358° C COL R: 26.6* C a i n COL T: 78° C COL R: £e.e C/Kiri Hi: 4 COL T: 178°C WI: 8 COL R: 2 6 . e C / E i n COL T: 338 C STATUS: : END =  c  C  CR  WI CT  .558  MilCR CT CR  INJECT FID B 16X8 6.5 CM/M 15*  Appendix C Calculation of Amount of Dimethyl Disulfide(DMDS) for Soaking N i - M o / A l 0 2  3  The amount of dimethyl disulfide(DMDS) for soaking the catalyst prior to sulfidation is calculated as follows: • Metal Content of Catalyst (approximate): N i : 3wt.%, Mo: 12wt.% • The amount of sulfur needed to form NiS and M0S2 for x grams of catalyst is calculated based on the weight percentage of each metal and their stoichiometry ratio of metal and sulfur. The stoichiometric amount of D M D S for sulfidation of 5 grams of catalyst is calculated as follows: Gram Sulfur Needed for Nickel Sulfidation = 0.03a Ni Igmol Ni lgmol S 32.06# S 5o catalyst x — x — — x —^——— x ^— * lg catalyst 58.70# Ni lgmol Ni lgmol S  v  (C.l) '  Gram. Sulfur Needed for Molybdenum Sulfidation = 0.12o Mo lgmol Mo lgmol S 32.06# S 5g catalyst x —,— x — — x ———— x ^— * * lg catalyst 95.94<7 Ni lgmol Ni lgmol S • The total amount of sulfur calculated above is 0.483gram.  v  (C.2) 1  Based on molecular  weight of D M D S (94.1894 g/gmol) and its specific gravity (1.046), the amount of D M D S to provide 0.483gram sulfur can be calculated:  Z\T 94.1894^  l9  DS  >< ^ lcm  = 0.011(,mo//cm ) DMDS ' 3  3  w  1  y  2gmol S OMlgmol DMDS Z2Mg S , „„„„ "TTTTTTT^ —-, 5 1 T~F = 0.705c75/cm DMDS lgmol DMDS lcm lgmol S Note: Chemical Formula of DMDS is ( C H ) S . n  x  x  r  7  n  r  n  l  3  3  3  2  (C.3) ' AS  C.4 '  v  2  0AS3gramS needed x ——~cT~~ ~ 0-685cm DMDS (stoichiometry) (C.5) 3  83  Appendix D Summary of Experiments in Hydrocracking Diphenylmethane in a Ceramic Membrane Reactor  D.l  Operating Conditions  Here is a summary of operating conditions for different hydrocracking experiments: • Reactor Temperature = 400° C • Reactor Pressure = 600psi • Gas Flow = 500 seem of 5% H S in H 2  2  • Wall Temperature = 250° C • Heating Tape Temperature = 55° C <.  D.2  D e f i n i t i o n of Equations  Liquid Recovery = Liquid Drained from each Section / Total Liquid Product  Yield of Benzene or Toluene = Total Moles of Benzene or Toluene / Moles of DPM,„  Liquid Mass Balance = (Mass of DPM,-,, - Mass of D P M  Note: D P M stands for Diphenylmethane.  D.3  S u m m a r y of H y d r o c r a c k i n g E x p e r i m e n t s  84  o u t  ) / Mass of D P M ,  n  85  Summary of Experiments in Hydrocracking DPM..  Table D . l : Summary of Products in Catalytic Hydrocracking with Catalyst on Tube—Side and D P M on Shell-Side Date of Experiment: March 4"', 1997 Membrane Pore Size =  1000A  Catalyst = 4.2 gram D P M , = 69.98 gram = 0.4160 gmole n  Reaction Time = 60 minutes Mass Balance, wt.% = 97  Section  Liquid Recovery  Benzene  Toluene  gram  wt.%  gmole  gmole  Flash Drum  0.8335  1.2  1.60E-4  1.49E-4  Tube-Side  1.099  1.6  6.86E-5  1.01E-4  Shell-Side  66.1774  97.2  7.13E-3  7.05E-3  Yield(mole%) Benzene  1.64  Toluene  1.63  Summary of Experiments in Hydrocracking DPM..  Table D.2: Summary of Products in Thermal Cracking Date of Experiment: April 7"', Membrane Pore Size = DPM,  n  =  74.75 gram  1997  1000A =  0.44 gmole  Reaction Time = 60 minutes Mass Balance, wt.% — 99  Section  Liquid Recovery gram  Flash Drum Tube-Side Shell-Side  wt.%  1.9026 2.6 26.7827 36.4 41.1628 61.0  Benzene  Toluene  gmole  gmole  7.55E-5 3.84E-4 0 5.16E-5 o 0  Yield(mole%) Benzene Toluene  0.02 0.02  Note: 1.8262 gram liquid product was collected by paper towel.  87  Summary of Experiments in Hydrocracking DPM..  Table D.3: Summary of Products in Catalytic Hydrocracking with Catalyst on Shell-Side and D P M on Shell-Side Date of Experiment: April 14"', 1997 Membrane Pore Size =  1000A  Catalyst = about 5 gram D P M ,  n  = 61.495 gram = 0.3655 gmole  Reaction Time = 60 minutes Mass Balance, wt.% = 96  Section  Liquid Recovery  Benzene  Toluene  gram  wt.%  gmole  gmole  Flash Drum  3.2573  5.5  3.83E-5  1.19E-4  Tube-Side  51.4  87.8  5.93E-4  3.22E-4  Shell-Side  3.0298  6.7  1.76E-4  1.81E-4  Yield(mole%) Benzene  0.22  Toluene  0.17  Note I: 0.8843 gram liquid product was collected by paper towel. Note II: Ceramic tube was broken in the course of reaction.  88  Summary of Experiments in Hydrocracking DPM..  Table D.4: Summary of Products in Catalytic Hydrocracking with Catalyst on Shell-Side and D P M on Shell-Side Date of Experiment: April 24"\ 1997 Membrane Pore Size = 1000A Catalyst = about 5 gram DPM;„ = 61.195 gram = 0.3637 gmole Reaction Time = 60 minutes Mass Balance, wt.% = 98  Section  Liquid Recovery  Benzene  Toluene  gram  wt.%  gmole  gmole  Flash Drum  8.9591  14.9  3.52E-5  3.32E-4  Tube-Side  4.6825  7.8  5.55E-5  7.44E-5  Shell-Side  46.5  77.3  2.65E-4  2.56E-4  Yield(mole%) Benzene  0.10  Toluene  0.18  89  Summary of Experiments in Hydrocracking DPM..  Table D.5: Summary of Products in Catalytic Hydrocracking with Catalyst on Tube-Side and D P M on Shell-Side Date of Experiment: May 5"', 1997 Membrane Pore Size = 1000A Catalyst = about 5 gram DPM,„ = 72.3 gram = 0.4 gmole Reaction Time = 57 minutes Mass Balance, wt.% = 96  Section  Liquid Recovery  Benzene  Toluene  gram  wt.%  gmole  gmole  Flash Drum  1.2783  1.8  3.55E-5  4.95E-5  Tube-Side  7.5  10.8  1.21E-3  1.01E-3  Shell-Side  60.9283  87.4  3.05E-3  2.41E-3  Yield(mole%) Benzene  1.82  Toluene  0.81  90  Summary of Experiments in Hydrocracking DPM..  Table D.6: Summary of Products in Catalytic Hydrocracking with Catalyst on Tube—Side and D P M on Shell-Side Date of Experiment: July 14"', 1997 Membrane Pore Size = 100A Catalyst = 5.2610 gram DPM,-,, = 78.78 gram = 0.47 gmole Reaction Time = 60 minutes Mass Balance, wt.% = 83  Section  Liquid Recovery  Benzene  Toluene  gram  wt.%  gmole  gmole  Flash Drum  7.70  11.72  5.58E-3  2.08E-3  Tube-Side  9.20  14.00  2.65E-3  4.90E-4  Shell-Side  48.80  74.28  1.83E-3  5.3E-4  Yield(mole%) Benzene  2.13  Toluene  0.66  91  Summary of Experiments in Hydrocracking DPM..  Table D.7: Summary of Products in Catalytic Hydrocracking with Catalyst on Tube-Side and D P M on Shell-Side Date of Experiment: August 7"', 1997 Membrane Pore Size = 100A Catalyst = 5.0388 gram DPM,-„ = 80.0788 gram = 0.48 gmole Reaction Time = 60 minutes Mass Balance, wt.% = 94  Section  Liquid Recovery  Benzene  Toluene  gram  wt.%  gmole  gmole  Flash Drum  12.82  17.04  1.80E-3  7.80E-4  Tube-Side  11.74  15.06  2.37E-3  4.80E-4  Shell-Side  50.68  67.36  8.00E-4  1.60E-4  Yield(mole%) Benzene  1.03  Toluene  0.30  92  Summary of Experiments in Hydrocracking DPM..  Table D.8: Summary of Products in Catalytic Hydrocracking with Catalyst Elevated on Tube-Side and D P M on Shell-Side Date of Experiment: September 30"', 1997 Membrane Pore Size = 1000A Catalyst = 2.7011 gram DPM,„ = 80 gram = 0.48 gmole Reaction Time = 60 minutes Mass Balance, wt.% — 92  Section  Liquid Recovery  Benzene  Toluene  gram  wt.%  gmole  gmole  Flash Drum  24.06  32.56  12.66E-3  10.88E-3  Tube-Side  4.79  6.48  4.90E-4  1.30E-4  Shell-Side  45.06  60.96  7.36E-3  2.24E-3  Yield(mole%) Benzene  4.27  Toluene  2.76  Appendix E Calibration of Flow Controllers  The mass flow controllers and rotameters were calibrated at room temperature and pressure, using a soap bubble flowmeter. To obtain an accurate measurement, each flow was measured three times and the average values summarized in tables, were used to determine the calibration factors. In the following pages the measured data and calibration curves are gathered.  93  Calibration of Flow Controllers  Table E . l : Calibration of Mass Flow Controller for Argon  P.109705 11-May-99  CALIBRATION OF FLOW CONTROLLER ARGON Bubble Flowmeter Time(sec) Volume(mlit) 23.4 10 23.37 10 23.4 10 AVG  23.39  Bubble Flowmeter Time(sec) Volume(mlit) 17.85 20 17.72 20 17.69 20 AVG  17.75  Bubble Flowmeter Time(sec) Volume(mlit) 15.15 30 15.1 30 15.12 30 AVG  15.12  SUMMARY Bubble Flowmeter Time(sec) Volume(mlit) 23.39 10 17.75333 20 15.12 30  Flow(mlit/min) 25.64 25.67 25.64  Flow Controller % Flow(mlit/min) 202 28.57 202 28.57 202 28.57  25.65  Flow(mlit/min) 6723 67.72 67.83  Flow Controller % Flow(mlit/min) 50.1 70.85 50.1 70.85 50.1 70.85  67.59  Flow(mlit/min) 118.81 119.21 119.05  Flow Controller % Flow(mlit/min) 85.4 120.77 85.4 120.77 85.4 120.77  119.02  Flow(mlit/min) 25.65 67.59 119.02  Flow Controller % Flow(mlit/min) 20.2 28.57 50.1 70.85 85.4 120.77  Calibration of Flow Controllers  y = 1.0259x R = 05979 J  140  T  20  0-1 0.00  I 20.00  -  -  -  "I 40.00  —r60.00  •  1 80.00  H 100.00  1 120.00  Bubble Flowmeter, mL(NTP)/mln  Figure E . l : Calibration of Mass Flow Controller for Argon  1 140.00  Calibration of Flow Controllers  y = 0.7255X R = 0.9979 2  100 •]  90  60  70 •  60 •  Flow Conti  p SO  40  30  20 -  10 •  0 0.00  1 20.00  1  (  1  40.00  60.00  80.00  \ — 100.00  :  h  !  120.00  Bubble Flowmeter, mL(NTP)/mln  Figure E.2: Calibration of Mass Flow Controller for Argon  140.1  Calibration of Flow Controllers  Table E.2: Calibration of Mass Flow Controller for Hydrogen  P.109705 11-May-99  CALIBRATION OF FLOW CONTROLLER HYDROGEN Bubble Flowmeter Time(sec) Volume(mlit) 37.37 20 37.35 20 37.22 20 AVG  37.31  Bubble Flowmeter Time(sec) Volume(mlit) 11.75 10 11.9 10 11.65 10 AVG  11.77  Bubble Flowmeter Time(sec) Volume(mlit) 20.22 20 20.25 20 20.31 20 AVG  20.26  Bubble Flowmeter Time(sec) volume(mlit) 17.66 20 17.66 20 17.75 20 AVG  17.69  SUMMARY Bubble Flowmeter Time(sec) Volume(mlit) 37.31 20 10 11.77 20 20.23 20 17.69  FIow(mlit/min) 32.11 32.13 32.24  Flow Controller Flow(mlit/min) 20 20 20  32.16  Flow(mlit/min) 51.06 50.42 51.50  Flow Controller Flow(mlit/min) 40 40 40  51.00  Flow(mlit/min) 59.35 59.26 59.08  Flow Controller Flow(mlit/min) 50 50 50  59.23  Flow(mlit/min) 67.95 67.95 67.61 67.84  Flow(mlit/min) 32.16 51 59.23 67.84 0  Flow Controller Flow(mlit/min) 60 60 60  Flow Controller Flow(mlit/min) 20 40 50 60 0  Calibration of Flow Controllers  Figure E.3: Calibration of Mass Flow Controller for Hydrogen  Calibration of Flow Controllers  Table E.3: Calibration of Rotameter for Helium  P.110457b 21-JUI-99  CALIBRATION OF ROTAMETER (A) For Drying Silica-Alumina Bubble Flowmeter Flow(mlit/min) Time(sec) Volume(mlit) 20.54 29.21 10 20.64 29.07 10 20.32 29.53 10 AVG  29.27  10  20.50  Bubble Flowmeter Flow(mlit/min) Time(sec) Volume(mlit) 40.32 14.88 10 40.62 14.77 10 40.35 14.87 10 AVG  14.84  10  40.43  Bubble Flowmeter Flow(mlit/min) Time(sec) Volume(mlit) 99.17 18.15 30 99.50 18.09 30 99.01 18.18 30 AVG  18.14  30  99.23  Bubble Flowmeter Flow(mlit/min) Time(sec) Volume(mlit) 190.48 12.60 40 190.17 12.62 40 190.63 12.59 40 AVG  12.60  40  190.43  Rotameter 10 10 10 10  Rotameter 20 20 20 20  Rotameter 40 40 40 40  Rotameter 60 60 60 60  Summary Bubble Flowmeter Flow(mlit/min) Time(sec) Volume(mlit) 20.50 29.27 10 40.43 14.84 10 99.23 18.14 30 190.43 12.60 40  Rotameter 10 20 40 60  Calibration of Flow Controllers  Figure E.4: Calibration of Rotameter for Helium  Calibration of Flow Controllers  Table E.4: Calibration of Rotameter for Propylene  P.110457b 21-Jul-99  CALIBRATION OF ROTAMETER (A) For Coking Dried Silica-Alumina Bubble Flowmeter Time(sec) Volume(mlit) 13.97 10 14.12 10 14.12 10 AVG  14.07  10  Bubble Flowmeter Time(sec) Volume(mlit) 21.54 10 21.62 10 21.66 10 AVG  21.61  10  Bubble Flowmeter Time(sec) Volume(mlit) 14.60 20 14.63 20 14.72 20 AVG  14.65  20  Bubble Flowmeter Time(sec) Volume(mlit) 9.03 20 9.13 20 9.09 20 AVG  9.08  20  Rotameter Flow(mlit/min) 42.95 42.49 42.49  10 10 10  42.65  10  Rotameter Flow(mlit/min) 27.86 27.75 27.70  5 5 5  27.77  5  Rotameter Flow(mlit/min) 82.19 82.02 81.52  20 20 20  81.91  20  Rotameter Flow(mlit/min) 132.89 131.43 132.01  30 30 30  132.11  30  Summary Bubble Flowmeter Time(sec) Volume(mlit) 21.61 10 14.07 10 14.65 20 9.08 20  Rotameter Flow(mlit/min) 27.77 42.65 81.91 132.11  5 10 20 30  Calibration of Flow Controllers  Figure E.5: Calibration of Rotameter for Propylene  Appendix F Calibration of Mass Spectrometer  The calibration procedure is explained in Section 4.2.5. Figure F . l , Figure F.2, Figure F.3 and Figure F.4 show typical response of mass spectrometer (MS) to injection of methane and propylene in hydrogen and/or argon. Each peak corresponds to 250/xL injection of methane or propylene. To calculate response factor the following steps were followed: 1. Total Pressure of Mass Spectrometer Chamber = M S Signal for Argon / Mole Fraction of Argon / Intensity of Argon provided by supplier (Refer to Table F . l ) . 2. Mole Fraction of each Component = (MS Signal for each Component / Total Pressure) — Baseline of each Component The baseline for each component is calculated as follows: Arithmetic Average of Initial Data or (MS Signal for each Component / Total Pressure) 3. Plot mole fraction of desired component versus time and integrate the area under the curve by trapezoid rule. 4. Response Factor (RF) = Total Moles Injected / Area under Mole Fraction-Time curve 5. Response Factor for Ethane & Ethylene =  RF^fethane  x Intensity of Ethane or  Ethylene / Intensity of Methane Intensities were all provided by Mass Spectrometer Supplier (Refer to Table F . l ) . 6. To calculate the amount of each product for a given experiment the following equation was used: Moles of Product = R F of that product x Area under Mole Fraction-Time Curve 103  Calibration of Mass Spectrometer Mole fraction- time curve is calculated as outlined above (steps 1 to 3).  Table F . l : Intensity of Different Gases in Mass Spectrometer Gas  Intensity(%)  Methane(15)  86  Ethane(27)  24  Ethylene(30)  64  Argon (40)  87  Area under the C u r v e = 0.03543  4.00E-02  n  3.50E-02  2.50E-02  S  2.00E-02  1.00E-02  0.0OE+OO 6  8 Time, mln  Figure F . l : Calibration for Methane in Hydrogen and Argon  Calibration of Mass Spectrometer  Area Under the Curve = 0.01181  MOE-02 - i  Time, min  Figure F.2: Calibration for Propylene in Hydrogen and Argon  Calibration of Mass Spectrometer  Figure F.3: Calibration for Methane in Argon  107  Calibration of Mass Spectrometer  Area under the Curve = 0.00747  Figure F.4: Calibration for Propylene in Argon  Appendix G Thermogravimetric Analysis of Silica—Alumina  About 22 mg of silica-alumina underwent thermogravimetric analysis in a T A Instrument Inc. TA2000 thermal analysis system equipped with a TG51 thermogravimetric analyzer to measure the weight loss as the temperature was increased in 58 m L ( N T P ) / m i n helium. The temperature was raised from 30°C to 600°C at a ramp rate of 10°C/min and the sample was soaked at 600° C for 1 hour. The weight loss of sample was about 17%.  Figure G . l : Thermogravimetric Analysis of Silica-Alumina Sample: Size: Method: Comment:  F i l e : GTGA1NEGAR.1 O p e r a t o r : Negar Run D a t e : 5 - J u l - 9 9 0 & 38  S i l i c a Alumina E2.1920 mg He h e a t He  100  ao  o  20  60 Time (min)  108  80  100 120 G e n e r a l V 4 . 1 C DuPont 2000  Appendix H Calculation for Carbon Recovery  Carbon Recovery, wt.% = 100 x (Total Carbon Recovered (mole)/ Total Carbon Deposited on Silica-Alumina (mole))  Total Carbon Recovered (mole) = (Methane (mole) x 1) + (Ethane (mole) x 2) + (Ethylene (mole) x 2) + (Propylene (mole) x 3)  Total Carbon Deposited on Silica-Alumina = Average Carbon Content of SilicaAlumina (%) x Weight of Dried Silica-Alumina  Average Carbon Content = 5.1 wt.% measured by Canadian Microanalytical Service Ltd.  Weight of Dried Silica-Alumina = 0.87 (measured by T G A , Refer to Appendix G) x Original Weight  109  Appendix I Summary of Results for Temperature—Programmed Reaction of Coked Silica-Alumina  In this Section, the material and results for Temperature-Programmed Hydrogenation of Coked Silica-Alumina are summarized.  110  Summary of Results for TPH of Coked Silica-Alumina  la lo  * £ £ £ III S I " » «|?« s £ m* ° 8 8 g|  g  g < CO £  g 1  eo  2  o  o  v  « 5"  s sP a£  a s co |  5 se §  5  sfeje 1 1  o- Si J= £  « c « .c UJ  * N IO (0 CM Jo CM  I e > ^ o  E  S  O CO »- f g  CD  rE S oS oo qo f c o o o o o f c o o o o ^ q o o o o ' ^ : dddd^ o o ' o o d ^ o o o o ^ o o o o o c M  « LU  « o o -  I  - at  £  l | | "5  CN  8 S S 8 «"" £  •S o  CM «  e S  .  - J*- *- CO  8 8 o *  I*  CM " «>  O B CJ Q  g UJ O in in '  t o co j r  <  3 CM  <D CM CM CM odd  «titUi  CO £ CM CM d d v  C  ?  CM CM d d  CM* s ftas 2 g. £ » 2 oc oc oc * n  "g "S E E o c s> a > < < CO CO "8 "8 XL O O  T o ir-r- CM -r- » CO  >i«si^™>l C5  W  < CO  1 o  <<  o o " o _o  Iff sis t  s  t  IS I  |S| © C o  O O  f If  O O  i «  CO CO TJ T>  2S 8 8 CX Q. O O C C O O "8 "8 O O  "o " o  «_ < D Q. O. « n at aj  8  U  112  Summary of Results for TPH of Coked Silica-Alumina  * * * * ££ £ £ £ £  * * *!  * * J5 o  8  •s t j s g  0  0  0  S  CO CO  •£  <M  fj  0 ' *  £ £ £ g -  < C  to s\  CO IO CO e8  s s  C  • E  o  •  IE  o  .2  8 S 8 S 8 g g  1° I e R I 8 8 a o  »  a  N  n  o -E E » ; ;  d  o o  *- o - o o" o  o  I  o a? o ci o o  V  TS  |  K  r- o  o  o  g  ?  CO Oi  CD <D  o *—  o  N  d d  S 2 d d  r |  ~  ci  o  o  3  » S  f-  2 CO (D *" Jy|  Pi  o o  CM  E 3  :y  8  I is  3 5 8 S  3  (A IO IO to <M CJ CJ CJ  IO Ift  8 8 IO to CJ CJ  dddd  2S J e  9  CJ CD P» o a:  oe "  CM CM CM CM -Si  sis  E  ??i  1 1 1  ffl  i l  »- <o  1  8 «  9  1  c o  •i  c o  c o  c o  c E cr E  £I  a> o) ir o «I 1«  < < CO CO  11  g:  < co  f lo  * o  Summary of Results for TPH of Coked Silica-Alumina  > CD  ill  Ii i  !?  W  11  W  «  r  838~g  1  11 E  In: £ €  i»8  !85~*  E  ???  ?? id  1  u i UJ t u a> r- i o  Ui  S CO  cIS^J  uj  S S  o.  CO  i  9 9  Ui tu  r-  S a  d d  CMIOCD S CM,IOCM , CV)  « £!  CM CM CM  ID IO  o o  CM CM  9  u> £ 8 3  |5 ^  UI  "IS <£> K  CM CM  "  88i  11 HI  CM O  CM CM CM CM  "t CO V  S  w  UJ u i  CD o CO  ! | o |  B  ,??  9 9 9  9 9  S  UJ UJ  CONS CO CM CO  id  CM CJ ui  IIS  8  S  n  3s  O O O  i >  " B  - CM IO > £ i oi C M, i E S  CM IO  o o  , CM CM  oc t r  6 d> o o  cn  8  se •g * o o •0 «e  -& * s ~ s 8 S ? g i 8 -a © •»  i p i «  p  i  0  «  o  •B 15  & 8-  I I  11  ||  s «  o o  5  •si if s 11 Hi  s E E  ^,  £  113  Summary of Results for TPH  of Coked Silica-Alumina  "E  s  "> S  f i  II  i-  **  I  o  T  f- CO CO to CM CM  fl  u  <u a A  S  5  ??  Ui o> v  ,9  *~  co  °  o  LU  LL)  Ui  ??  UJ  1  "S 8 2  uiin9 CO co cd  B  '  o  •  fl cn 8 3  fl  .+»9 ao o  to  v  !  Tt  oi  aa  UO  IO  co  v  O  O  o  O  CM CM A US o *-  j  9 9 9  UJ  UJ  UJ  5  3 n i - io  o  1 1 1l  UJ  9  9  ut  jj  Ei  •8  2  i  |  "  to to to IO CM CM CM. CM o o o o  fc  CC  CO CM  99  UJ  UJ  ii  li  lifii Sis  to  to  «S_CM.  o  o  o  I CM CM CM . CM uS CO , »~ O »-  CM CM CM d> o •-  9  UJ  • to  3 ai  iiiiff*  £  8  W  CO CM CO CO  • CO CO Cb CO IO  §CO C£O C8O « £® I CO CM CM CM CM  Is . 8 8  |  ????  UJ o r-  UJ r-.  CM  ! to  OJ  9  9  O 03 <D CO  00  CO  CM CM  CM CM  a  i i i i i  CM CM CM  t-l  n3  iili  1  P-,  l-l  li  1  A  O  iiii  o  CM CM  CC  1«  S  8  ss  •» «o  < <  CO CO TJ "© e  •8 o  •  * o  S 6 6 S S  I* £> E»  2*  re e s  ---- |  CC CC CC CC  %  o_ o_ o_ o_  r"  CC  CC  £ «  8 *  ? < a.  a.  CO  Summary of Results for TPH of Coked Silica-Alumina  > r-  s  r F! S e ,,  u) z z z  2  |« « o ^  ^  a i A i i ) j f <<< W ^ <M »fi IZZZ  o •o  •—<  o  CM  I « io w  T3  O  j*;  "  3 8 S R 2 S laCOO ON  *5  Cft CM S T-  «>  W *- K CD «£| | d CO (O r  t-i  PM  cu  ft S 8 «io , C 8 !  a  (U 4  >,  ko co co CC £  UJ UJ  •  PU O  CO  u  7  PH tn  2  co CU  IN  n  &  CO 9  UJ UJ OI CO CM CO K CO  CO CM CM lUJ UJ Ui  I CM CM '9 9 I UJ UJ  ui  I  '- • UJ UJ F- <o  ,,CM TCMC*OCMn IO  s»  ^  IO  IO  1  3 5 In , CM CM CM  .,1 i s i f l l  £ 3*. E  LU  p?9 9  CM CM CM CM  1 2  f?99  io CM CM o" d U1  hr CM 3 ko I A in CM CM CM  hr  M  r-  KM CM id d l Oi CM A A ; o o  Fee <  ^3  CO  X  E  K  t  « •  I  Iii  - s?  t» 5  < <  Oi </> •u XI  • O *o ©  $£ ss  1?  "B. "O.  Il  O O  1*  -a  115  Summary of Results for TPH of Coked  N  Silica-Alumina,  I  ^1  z z z z  .2 o  I  s  v  to  r - -* O  ^  OI  CO CM  » «o r-  zrzz  o  u  P-.  If- W CO <D O jSI | o U> •» CO {J S)|<  cu  ei  CP ,—i :>>  a, o  ITJ-  I"  CO <D C M  E  PL,  1|<0 CO  (-1  • 9 9  Jul  ,o  n n B w 9 9 9 9 UJ UJ UJ UJ 0 ) 0 0)10  CO CO 9 9 UJ UJ  < <|  O) N O) O C M C M CO  UJ  lol  Is sz z z z  )|CM C M  •?  9  C M C^  CU  ki ui ui  lUJ UJ  Eh  I®  tf  S  UJ  CM  lUJ LULU LU  . 8|gg  Id d d d  sas8h  6 m tu  Ia  IS  ii  1  3«  | c . $ | f sis i  ?|f  M C M « " 5 C  do  32  *  :T f  .,JS  ii  £1  CO N CO U5  C M C M C M C M  id d  Id 6 ci d CM CM CM CM  " » CM *0 • kO r- O 1-  o  00 CM I ICM C M  ml  C M CJ C Tr C M l  O  u  io  id d  T- <  ir 5~ £ 8 8 o "o  '1!  8 X  c  mi  CO CO  N  £51 *  9 m « •  . £ . £ J= J=  <<  c  • « cc fi  <r ir or <r  « « g: co o. p. 2 • 3  O  Summary of Results for TPH of Coked  a  .2 o  fl o tH  PH aj fl cc3  A  w t-l  .o  CO  tf cc3  A CO  Silica-Alumina  Summary of Results for TPH of Coked Silica-Alumina  0> CO «O S T - r-  o »- o «~ j  W O  CO  O  j  f V  «T~|  O  oi d '  CM  iill  Iii Id8  liii  iii  118  8  d|  fl  .2  U) CD N (V , , to. io u> o  v=  u fl id o  o d d d "*  oi ui <o O  SH  O  O  CO  d d d  d  n  ;S  88 8 8  o o o o o o r- d d d  d|  PH  0)  fl  23!  ^o5? o «- q q J. d d o o co d o o o '  d d d d d d t-  3?  33  3333  UJ UJ  UJ LU  cci  A  wfl  8.5 *- ci  SH  co o CD  tf  3  5 ^ IS  a a  A CO CU  fl fl  • rt H-=  fl o o  .1  3  S S8  S  S3  SI 2 SIs*  !  ????„>.8+ 8+ 8+8+ N f o di o to £ S £ 8 8 uu tu s  UJ UJ UJ UJ  "fc  UJ UJ  8 8 d  S 8 ir  IT  CM Oj  o  o  s  L i  to to CJ CVJ  d  d  iiii  ii  Ji IO to OI Oi OI CM  CD lOl IO IOI CM CMCM CM I  d d d d  d d  d  d|  d  OC oc 8  as  o  13 -o  lac  oc  CM Ol|<  1:  6c oc  £  cell  ss 8-8c  o co co oc  EH  li i i  UJ UJ UJ O O CO  S © OJ OI CM CO  r-  O  8 8 a to  vi n  IS < <  CO CO  ^ "S o o  ,o  o  CO O l CC CO  a) aj to eo  ^ _c _c _c CC CC CC CC  o g « rape  lEi|  E co.  ;  « « cc c I ccc t *l  I  no.  Hp 1  1  !=  «  al  =o "S° l " H i 5 2  • CO  OI  o o| I  Summary of Results for TPH of Coked Silica-Alumina  g o o. 5 t-. <o  CM  5$  £ e  0  o  X uj  b 8 s - 5 §1 1? 8 *" ^ I  O  0 O  u Pw  cu  PI  3  UJ  b  a a  UJ  SS  _  5 " £  oj  CU  M  ???  w  Ut U ) CD tri  J*?9 Jul UJ l i O Cxi  o  3  3  a  PI CO  X  UJ  CD U>  8  99?  Lit UI UJ O f N  6  )fflO . hs s; 8 t 5  fs  SR  res em  UJ CM CM  dd  fit k  < < e>  fS  *l  sP  5 si  ho ' »  [CM CM CM  S •? <o  UJ CM CM  O O  <lf <M  ICM CM  £  d d> loo  I E  TJ •  E E o «5  $ *  .< < k o co  rs  |o O (-8 «S  * l  iiriti  1  X U) CM CM dd ci  X _><  ^pl  H  CM CM  i i  M tO CD k- w » I  &o  UJ CM CM  If ss do  3  UJ  IA UJ'  c6 o»  co CU  p=:  JUJIUJ UJ  LU UI CM eg  LU UI UJ NlSiJ  -1-5  11 8-81*  Ril  I < o e fB.U  ss 5:  X  I i III?  II  Summary of Results for TPH of Coked  Silica-Alumina  T e  5*  E  o» «A f -  „.  . ^ Kz n  fl .2  a  : 8 8|  z z z  - d  o'l  o o d  o o o o d d  o | o I o l  o  op  s E  fl o  <<< j z z z z  3 3 3 2!  (H  P4  CD fl CU *>> ^fl  ,o  ????  ft*  O* r- «  p9 111 UI  fl  Tin  CO  hi] UI UI UI  o  CO tf  a aA  re?  Jo o ©j o d o l Id  si s IS S ? | |<z z< z< z< pS 6| ^ lo o  E  >  o  RJU LU  o o  d  Is si  CO  -cT  CO fl  V Ifl CO CM  lil IA Of OI p d  eg o i  .3  •+ *t m to ifl IA Ui u> oi oi o i o i d o d o oi oi oi o i  v= A  | E  o cj  Six •AOI <OAi lI Id o ' l  OI OI c  Ifctefcoc  88 o *o  8-8-  3  8  i  ll  SS  S S 6 S  S 8  .£ c c C  ho *©  ro.ii ii  a « « a  ICC OC CC CC  O. 0_ 0_ 0_  So  Summary of Results for TPH of Coked Silica-Alumina  121  Table 1.6: Summary of Medium of Separation for T P H of Coked Silica-Alumina  Description Thermal, 5 m m ceramic on top of coke 5 m m ceramic between coke, cat 3-5 m m ceramic between coke, cat 3-5mm dried S A between coke, cat 5 m m dried S A between coke, cat  Run # 103 100 109 101 115  Medium of Separation Ceramic(g) Silica-Alumina(g) 0.3083 0.3144 0.2950 0.1313 0.1244 -  Note: The silica-alumina was dried in a ceramic crucible placed in V W R Scientific furnace. The sample was heated from 100°C to 600°C at a ramp rate of 10°C/min and it was soaked for 1 hour at this temperature. Consequently, the dried sample was placed in nitrogen glove box to cool to ambient. This drying process was carried out at the same time the silica-alumina for coking was dried in the reactor.  Appendix J Summary of Temperature-Programmed Hydrogenation Profiles of Coked Silica-Alumina  2.50E-03  0.00E+00  0.00E+O0 300  200  400  Temperature, °C Methane, R97-2 -Propylene, R97-2  -Ethane. R97-2 -Ethylene, R97-2  Figure J . l : R97-2  122  Summary of TPH Profiles of Coked  123  Silica-Alumina  4.00E-O3  3.00E-03  2.50E-03 H  2.00E-03  | * u- c  1.S0E-03  1.00E-O3  O.OOE+00  O.OOE+OO  300  400  Temperature, *C —a—Ethane. R98-2 — —Ethylene. R98-2  - - - Methane, R98-2 Propylene. R98-2  Figure J.2: R98-2  Summary of TPH Profiles of Coked  Silica-Alumina  Figure J.3: R99-2  124  Summary of TPH Profiles of Coked  125  Silica-Alumina  4.S0E-03  2.50E-03  4.00E-03 2.00E-03 A  o g c sz  f Sill *  1.S0E-03  •S 1.00E-03 i  5.00E-O4 \ S.00E-04 O.OOE+00  O.OOE+00 200  300  400  500  Temperature, *C -Ethane, R100-2 -Ethylene, R100-2  Methane. R100-2 -Propylene. R100-2  Figure J.4: R100-2  Summary of TPH Profiles of Coked  0  100  200  126  Silica-Alumina  300  400  600  600  Temperature, *C - ^ - E t h a n e , R102-2 — —Ethylene, R102-2  - - - Methane, R102-2 Propylene. R102-2  Figure J.5: R102-2  700  Summary of TPH Profiles of Coked  Silica-Alumina  Figure J.6: R103-2  127  Summary of TPH Profiles of Coked  128  Silica-Alumina  3.S0E-03 -  3.00E-03 •  _  Z.50E-03 •  o  e  O CO  3= c 8 I 2.00E-O3 • U. 4D tt» o  =  E  1.S0E-03 •  1.00E-O3 •  5.0OE-O4 • o.ooE+oo 200  300  400  500  Temperature, *C |- - - Methane, R104-2]  Figure J.7: R104-2  -  600  - • • 700  Summary of TPH Profiles of Coked  Silica-Alumina  2.50E-03 n  2.00E-03 A  o 1.50E-03 c o co  z: U  1= CO  <o x: 1.00E-03  5.00E-04 •  O.OOE+00 300  400  Temperature, C - - Methane. R105-2 |  Figure J.8: R105-2  Summary of TPH Profiles of Coked  Silica-Alumina  Figure J.9: R109-2  130  Summary of TPH Profiles of Coked  10  n  Silica-Alumina  200  300  400  500  Temperature, *C —€>-Ethane, R110-2 — —Ethylene, R110-2  - - - Methane, R110-2 Propylene, R110-2  Figure J.10: R110-2  600  700  Summary of TPH Profiles of Coked  0  100  132  Silica-Alumina  200  300  400  500  Temperature, °C —•—Ethane, R111-2 — —Ethylene, R111-2  - - - Methane. R111-2 Propylene, R111-2  Figure J . l l : R l l l - 2  600  700  Summary of TPH Profiles of Coked  Silica-Alumina  6.00E-04 -i  S.OOE-04  4.00E-04 H  o © 8 J 3.00E-04 u. o  2.00E-04  r* 1.00E-04  %  # i «  f  >  * <  O.OOE+00 300  400  Temperature, °C • - - Methane, R112-21  Figure  J.12: R112-2  Summary of TPH Profiles of Coked  Silica-Alumina  Figure J.13: R113-2  134  Summary of TPH Profiles of Coked  135  Silica-Alumina  + 3.S0E-O3  0  100  200  300  400  600  Temperature, *C —a— y£thane.R115-2 — —yEthylene,R11S-2  - - - yCH4,R115-2 — y C 3 H 6 , R115-2  Figure J.14: R115-2  600  700  Summary of TPH Profiles of Coked Silica-Alumina,  Figure J.15: R117-2  136  Summary of TPH Profiles of Coked  Silica-Alumina  3.SOE-03  3.00E-03  2.S0E-O3  2.00E-03  o  1.50E-03  1.00E-03 f* S.OOE-04  O.OOE+00  —f—  100  300  400  Temperature, °C [ - • - Methane. R118-2 ]  Figure J.16: R118-2  Summary of TPH Profiles of Coked  0  100  138  Silica-Alumina  200  300  400  600  Temperature, *C —•—Ethane, R119-2 — —Ethylene, R119-2  - - - Methane, R119-2 Propylene, R119-2  Figure J.17: R119-2  600  700  Appendix K Summary of Temperature—Programmed Reaction Profiles of Coked Silica-Alumina in Argon  1.60E-03 -i  1.40E-03  1.00E-03  « *? 8.00E-04  II  S I 6.00E-04 H  4.00E-04 1  2.00E-CM  0.00E+00  O.OOE+00 300  400  Temperature, *C Methane, R'104-2 -Propylene, R'104-2  -Ethane, R'104-2 -Ethylene, R'104-2  Figure K . l : R'104-2  139  Summary of TPR Profiles of Coked Silica-Alumina  140  in Argon  1.20E-03  T  3.50E-O3  8.00E-O4 A  P * 6.00E-04 J  4.00E-04 H  2.00E-04  O.OOE+00 300  400  Temperature, °C Methane. R'105-2 -Propylene, R'105-2  -Ethane, RM0S-2 -Ethylene, R'105-2  R'105-2.XLS  Figure K.2: R'105-2  Summary of TPR Profiles of Coked Silica-Alumina in Argon  Figure K.3: R'112-2  141  Summary of TPR Profiles of Coked Silica-Alumina  in Argon  Figure K.4: R'118-2  142  Appendix L Test for Plug Flow  To investigate the pattern of gas flow in the reactor, several pulse flow response test were performed at 601° C. About 250 mg of silica-alumina was placed in the reactor and dried according to the drying procedure explained in Section 4.2.1. The dried silica-alumina was attached to the T P R setup. A mixture of 30 m L ( N T P ) / m i n hydrogen and argon (30% argon in hydrogen) flowed to the reactor. Pulses of methane were injected into the reactor and the exhaust gas was monitored by mass spectrometer. The results were analyzed based on the method outlined in Fogler (1992).  L.l  Analysis of Tests for Plug Flow  The tracer (methane) was injected into the packed bed of silica-alumina at a location more than two or three particle diameters downstream from the entrance and measured some distance upstream from the exit. Hence, the open-open vessel boundary conditions were applied (Fogler (1992)). It was assumed that the tracer was an ideal inert pulse, since at 601° C physical adsorption of methane does not occur. The following equations were used for calculations respectively: Residence-Time Distribution Function (E(t)) E(t) = Mole Fraction of Tracer (Methane)/Total  Total Amount of Injected Tracer — I Jo  Amount of Injected  Mole Fraction dt  Tracer(L.l)  (L-2)  Mole fraction of tracer was calculated as outlined in Appendix F . rOO  Mean Residence Time (t ) = I tE(t)dt Jo m  143  (L.3)  144  Test for Plug Flow /•oo  Variance (a ) = / (t Jo 1 Skewness (s ) = —  (LA)  t ) E(t)dt  2  2  m  f°° (t -  3  (L.5)  t ) E(t)dt 3  m  Jo  CT2  Peclet Number (Pe ) is calculated from the following equation: r  a tl  2 Pe  2  8  r  (L.6)  + Pel  Dispersion Number =  (L.7)  Pe  r  (L.8)  Residence Time (r) =  The results of these tests are summarized in Table L . l . Sample graphs are depicted in Figure L . l and Figure L.2. Table L . l : Test for Plug Flow tm  ± c  Pe  r  Dispersion No.  (sec)  T  s  (sec)  (sec) /  3  3  Inside Reactor, T = 601° C 95±13  105  0.009  93  148  88±15  68  0.015  85  84  79±17  46  0.022  76  79  78±19  38  0.026  74  184  60±14  42  0.024  58  128  64±14  48  0.021  61  107  Reactor was bypassed  2  Test for Plug Flow  E Curve T = 601°C Per = 105  350  Time, sec  Figure L . l : E Curve for Inside Reactor Bed  Test for Plug Flow  E Curve Reactor was b y p a s s e d Per=48  (b) 0.05 •] 0.045 0.04 0.035  50  100  150 t,  200  sec  Figure L.2: E Curve for Tubings; Reactor was bypassed  250  Appendix M Surface Area Measurements  Surface area and pore volume of commercial N i - M o / A l 0 3 catalyst, silica-alumina, coked 2  silica-alumina and crushed ceramic were measured by a Micromeritics A S A P 2010, V3.02 C (Accelerated Surface Area and Porosimetry System). The system is operated by computer and the data are analyzed based on the chosen options. The measured surface area of crushed ceramic (Pore Diameter=  1000A;  Particle Di-  ameter= 90-150/xm) was very small ( B E T Surface Area= 0.1447m /g, Langmuir Surface 2  Area= 0.1715m /g). The Micromeritics A S A P 2010 apparatus can not operate accurately 2  for such low surface area. However, the surface area of ceramic tubes (Pore Diameter=  1000A)  was measured by Chan (1996), using mercury porosimetry. The total area ob-  tained by this method was 0.082 m / g . From Micromeritics and mercury porosimetry 2  measurements, it is clear that the surface area of ceramic membrane is much lower that the dried silica-alumina (Refer to Figure M.8). Summary of Analysis Options, Isotherm Plot, Incremental Pore Volume and Summary Report are listed respectively for commercial N i - M o / A ^ O a catalyst, silica-alumina and coked silica-alumina:  147  Surface Area Measurements  *  Figure M . l : Summary of Analysis Options for Commercial N i - M o / A l 0 Catalyst 2  ASAP 2010 V3.02  Page 22  S e r i a l # 325  Unit 1  C  Sample I d : c l e a n c a t . ( a d s - d e s ) O p e r a t o r I d : negar Submitter I d : F i l e Name: C:\...\NEGAR\CAT1 AD.SMP Started: Completed: Report Time: Sample Weight: Warm F r e e s p a c e : Equil. Interval:  06/10/97 11:35:54 06/11/97 00:45:15 11/22/99 13:35:05 0.6634 g 27.9656 cm* 5 sees Options  Analysis Adsorptive: Maximum m a n i f o l d p r e s s u r e : Non-ideality factor: Density conversion factor: Therm, t r a n . hard-sphere diameter: Molecular c r o s s - s e c t i o n a l area:  Min.  Analysis Adsorptive: A n a l y s i s Bath: Thermal C o r r e c t i o n : Smoothed P r e s s u r e s : C o l d Freespace: Low P r e s s u r e Dose:  Report Nitrogen 925.00 mmHg 0.000066 0.0015468 3.860 A 0.162 nm*  F a s t e v a c u a t i o n : Yes U n r e s t r i c e d evac from: 5.0 mmHg Leak t e s t : No Leak t e s t d u r a t i o n : 120 sees E v a c u a t i o n time: 0.1 hours B a c k f i l l Gas: A n a l y s i s E q u i l i b r a t i o n i n t e r v a l : 5 sees Maximum volume increment: No T a r g e t t o l e r a n c e : 5.0fto r5.0 mmHg e q u i l . d e l a y a t P/Po >= 0.995: 600 sees Free space group: Lower dewar a f t e r f r e e space: E v a c u a t i o n time: Leak t e s t : Leak t e s t d u r a t i o n :  Measured No 0.1 hours No 180 sees  p r e s s u r e d o s i n g : No 0.00 em'/g STP Dose amount: 0.00 hours Minimum e q u i l i b r a t i o n d e l a y : Maximum e q u i l i b r a t i o n d e l a y : 999.00 hours Low  Po type: Measured Temperature type: E n t e r e d Temperature: 77.35 K Measurement i n t e r v a l : 120 minutes I n s i d e diameter  o f sample tube: 9.530 mm  3  N2 77.35 K Ko No 90.3312 cm" None  Surface Area Measurements  Figure M.2: Isotherm Plot for Commercial N i - M o / A l 0 2  ASAP 2010 V3.02  Catalyst  S e r i a l ff 325  Unit 1  C  3  Page  Sample I d : c l e a n c a t . ( a d s - d e s ) O p e r a t o r I d : negar Submitter I d : F i l e Name: C: \. . . \NEGAR\CAT1_AD. SMP Started: Completed: Report Time: Sample Weight: Warm F r e e s p a c e : Equil. Interval:  06/10/97 11:35:S4 06/11/97 00:45:1S 11/22/99 13:35:05 0.6634 g 27.9656 cm» 5 sees  Analysis Adsorptive: A n a l y s i s Bath: Thermal C o r r e c t i o n : Smoothed P r e s s u r e s : Cold Freespace: Low P r e s s u r e Dose:  N2 77.35 K No No 90.3312 cm None  1  Isotherm Plot + Adsorption O Desorption 350  300  250 t-  200 a>  •oe CO  <  CD  E  150  100  0.0  0.1  0.2  0.3  0.4  0.5  0.6  Relative Pressure (P/Po)  0.7  0.8  0.9  1.0  Surface Area Measurements  Figure M.3: Incremental Pore Volume for Commercial N i - M o / A l 0 Catalyst 2  3  M i c r o m e r i t i c s Instrument C o r p o r a t i o n Unit 1  DET Plus(TM) VI.02 ( 2010 )  Page 9  Sample I d : c l e a n c a t . ( a d s - d e s ) O p e r a t o r I d : negar Submitter I d : F i l e : C:\..ANEGAR\CAT1AD.SMP Started: Completed: Report Time: Sample Weight: Warm F r e e s p a c e :  Analysis Adsorptive: A n a l y s i s Bath: Thermal C o r r e c t i o n : Equil. Interval: C o l d Freespace:  06/10/97 11:35:54 06/11/97 00:45:15 11/22/99 13:33:53 0.6634 g 27.9656 cm'  N2 77.35 K No 5 sees 90.3312 cm"  P o r o s i t y D i s t r i b u t i o n by O r i g i n a l D e n s i t y F u n c t i o n a l Theory Model: N2 6 77K on Carbon, S l i t Pores Method: Non-negative R e g u l a r ! z a t i o n ; No Smoothing Volume T o t a l Volume Area T o t a l Area  in in in in  Pores Pores Pores Pores  < <= > >=  14.83 2525.70 2525.70 14.83  A A A A  Incremental Surface A r e a vs. + Incremental Surface A r e a  0.02551 0.50268 0.007 132.728  cm»/g em'/g m'/g m»/g  P o r e Width  1e+02 Pore Width (Angstroms)  1e+03  Surface Area Measurements  Figure M.4: Summary Report for Commercial Ni-Mo/Al2C*3 Catalyst  ASAP 2010 V3.02  C  Unit 1  S e r i a l # 325  Page 23  Sample I d : c l e a n c a t . ( a d s - d e s ) O p e r a t o r I d : negar Submitter I d : F i l e Name: C:\...\NEGAR\CAT1_AD.SMP Started: Completed: Report Time: Sample H e i g h t : Warm F r e e s p a c e : Equil. Interval:  06/10/97 11:35:54 06/11/97 00:45:15 11/22/99 13:3S:0S 0.6634 g 27.9656 cm* 5 sees  Analysis Adsorptive: A n a l y s i s Bath: Thermal C o r r e c t i o n : Smoothed P r e s s u r e s : Cold Freespace: Low P r e s s u r e Dose:  N2 77.35 K No No 90.3312 cm' None  Summary Report Area  S i n g l e P o i n t S u r f a c e A r e a a t P/Po 0.19919780 :  267.0328  m /g  BET S u r f a c e A r e a :  276.2954  m»/g  Langmuir S u r f a c e A r e a :  381.7218  m'/g  BJH A d s o r p t i o n Cumulative S u r f a c e A r e a o f pores between 17.000000 and 3000.000000 A Diameter:  325.3111  m'/g  BJH D e s o r p t i o n Cumulative S u r f a c e A r e a o f pores between 17.000000 and 3000.000000 A Diameter:  391.3083  m'/g  J  Volume  S i n g l e P o i n t T o t a l Pore Volume o f pores l e s s than 729.6939 A Diameter a t P/Po 0.97274390:  0.519385 cm /g  BJH A d s o r p t i o n Cumulative Pore Volume o f pores between 17.000000 and 3000.000000 A Diameter:  0.546693 c m V g  BJH D e s o r p t i o n Cumulative Pore Volume o f pores between 17.000000 and 3000.000000 A Diameter:  0.548853  J  em'/g  Pore S i z e  Average  Pore Diameter  (4V/A by BET):  75.1926  A  BJH A d s o r p t i o n Average  Pore Diameter  (4V/A) :  67.2209  A  BJH D e s o r p t i o n Average  Pore Diameter  (4V/A):  56.1044  A  Surface Area Measurements  Figure M.5: Summary of Analysis Options for Silica-Alumina  Full Report Set ASAP 2010 V3.02 C  Unit 1  Serial 8 325  Sample Id: Silica-Alumina Operator Id: Negar Submitter Id: File Name: C:\...\NEGAR\SIL-ALUM.SMF Started: Completed: Report Time: Sample Height: Warm Freespace: Equil. Interval:  02/10/99 17:41:02 02/11/99 02:36:22 02/11/99 08:59:01 0.3100 g 29.4947 cm' 5 sees  Analysis Adsorptive: N2 Analysis Bath: 77.35 K Thermal Correction: No Smoothed Pressures: No Cold Freespace: 94.4358 Low Pressure Dose: None  Options Report Analysis Adsorptive: Maximum manifold pressure: Non-ideality factor: Density conversion factor: Therm, tran. hard-sphere diameter: Molecular cross-sectional area: Fast evacuation: Unrestriced evac from: Leak test: Leak test duration: Evacuation time: Backfill Gas: Equilibration interval: Maximum volume increment: Target tolerance: Min. equil. delay at P/Po >•= 0.995:  Nitrogen 925.00 mmHg 0.000066 0.0015468 3.860 A 0.162 run' Yes 5.0 mmHg Yes 120 sees 0.5 hours Analysis 5 sees No 5.0 t or 5.0 mmHg 600 sees  Free space group: Measured Lower dewar after free space: Yes Evacuation time: 0.5 hours Leak test: Yes Leak test duration: 180 sees Low pressure dosing: No Dose amount: 0.00 cra'/g STP Minimum equilibration delay: 0.00 hours Maximum equilibration delay: 999.00 hours Po type: Measured Temperature type: Entered Temperature: 77.35 K Measurement interval: 120 minutes Inside diameter of sample tube: 9.530 mm  Page  Surface Area Measurements  153  Figure M.6: Isotherm Plot for Silica-Alumina F u l l Report Set ASAP 2010 V3.02  C  U n  it 1  Serial « 325  Page  Sample Id: Silica-Alumina Operator Id: Negar Submitter Id: F i l e Name: C:\... \NEGAR\SIL-ALUM.SMP Started: Completed: Report Time: Sample Weight: Warm Freespace: E q u i l . Interval:  + Adsorption  02/10/99 17:41:02 02/11/99 02:36:22 02/11/99 08:59:01 0.3100 g 29.4947 cm' 5 sees  Analysis Adsorptive: N2 Analysis Bath: 77.35 K Thermal Correction: No Smoothed Pressures: No Cold Freespace: 94.4358 cm Low Pressure Dose: None  J  Isotherm Plot  /  /, w CO  300  /  •e o  tn XI  <  200  150  100  50  0.9 Relative Pressure (P/Po)  1.0  3  Surface Area Measurements  154  Figure M.7: Incremental Pore Volume for Silica-Alumina Micromeritics Instrument Corporation Unit 1  DFT Plus(TM) VI.02 ( 2010 )  Page12  Sample Id: Silica-Alumina Operator Id: Negar Submitter Id: F i l e : C:\..ANEGAR\SIL-AIAJM.SMP Started: Completed: Report Time: Sample Weight: Warm Freespace:  02/10/99 17:41:02 02/11/99 02:36:22 02/11/99 09:46:S2 0.3100 g 29.4947 cm*  Analysis Adsorptive: Analysis Bath: Thermal Correction: E q u i l . Interval: Cold Freespace:  N2 77.35 K No 5 sees 94.4358 cm  1  Porosity Distribution by Original Density Functional Theory Model: N2 6 77K on Carbon, S l i t Pores Method: Non-negative Regularization; No Smoothing Volume Total Volume Area Total Area  0.050  in in in in  Pores Pores Pores Pores  < <= > >•=  14.83 2339.13 2339.13 14.83  A A A A  0.05472 0.69860 1.654 262.084  cm»/g cm»/g m»/g * m'/g  Incremental Pore Volume vs. Pore Width + Incremental Pore Volume  0.045 0.040 0.035 CD 0.030 E a o > e 0.025 o  ~ 0.020 CO E " 0.015 0.010 0.005 0.000  1e+02 Pore Width (Angstroms)  1e+03  Surface Area Measurements  Figure M.8: Summary Report for Silica-Alumina  Full ASAP 2010  V3.02  C  Report  Unit  Set  1  Serial  Sample I d : m i x t u r e o f 4 batches o f coked O p e r a t o r I d : Negar Submitter Id: F i l e Name: C : \ . . . \ N E G A R \ C O K E H I X . S M P Started: Completed: Report Time: Sample H e i g h t : Harm F r e e s p a c e : Equil. Interval:  07/06/99 09:52:16 07/06/99 16:28:23 07/07/99 09:22:56 0.2248 g 2 9 . 6 9 4 9 cm* 5 sees  « 325  26  SA  Analysis Adsorptive: Analysis Bath: Thermal C o r r e c t i o n : Smoothed P r e s s u r e s : Cold Freespace: Low P r e s s u r e D o s e :  Summary  page  N2 77.35 K NoNo 9 4 . 1 2 2 2 cm* None  Report  Area  Single  Point  477.7084  m«/g  505.1171  m»/g  708.2262  m'/g  BJH A d s o r p t i o n C u m u l a t i v e S u r f a c e A r e a o f p o r e s between 17.000000 and 3000.000000 A Diameter:  581.0299  m /g  BJH D e s o r p t i o n C u m u l a t i v e S u r f a c e A r e a o f p o r e s between 17.000000 and 3000.000000 A Diameter:  652.9397  m'/g  BET S u r f a c e Langmuir  Surface  Area  at  P/Po  0.199807S2  :  Area:  Surface  Area:  l  Volume  Single  P o i n t T o t a l Pore Volume o f p o r e s l e s s than 26SS.6S15 A Diameter a t P/Po 0.99267784:  BJH A d s o r p t i o n C u m u l a t i v e Pore Volume o f pores between 17.000000 and 3000.000000 A Diameter: BJH  D e s o r p t i o n C u m u l a t i v e Pore Volume o f pores between 17.000000 and 3000.000000 A Diameter: Pore  Average  Pore  Diameter  (4V/A by  0.667504  em'/g  0.695826  em'/g  0.691912  cra'/g  Size  BET):  52.8594  A  BJH A d s o r p t i o n A v e r a g e  Pore  Diameter  (4V/A):  47.9030  A  BJH D e s o r p t i o n A v e r a g e  Pore  Diameter  (4V/A) :  42.3875  A  Surface Area Measurements  Figure M.9: Summary of Analysis Options for Coked Silica-Alumina  Full ASAP  2010 V 3 . 0 2  Unit  C  Report  Set  1  Serial  # 325  P a g e 2S  Sample I d : m i x t u r e o f 4 batches o f coked SA O p e r a t o r I d : Negar Submitter I d : F i l e Name: C : \ . . . \ N E G A R \ C O K E M I X . S M P Started: Completed: Report Time: Sample H e i g h t : Warm F r e e s p a c e : Equil. Interval:  07/06/99 09:52:16 07/06/99 16:28:23 07/07/99 09:22:S6 0.2248 g 2 9 . 6 9 4 9 cm» 5 sees  Options Analysis Adsorptive: Maximum m a n i f o l d p r e s s u r e : Non-ideality factor: Density conversion factor: Therm. t r a n . hard-sphere diameter: Molecular cross-sectional area:  Min.  Fast evacuation: U n r e s t r i c e d evac from: Leak t e s t : Leak t e s t d u r a t i o n : Evacuation time: B a c k f i l l Gas: Equilibration interval: Maximum v o l u m e i n c r e m e n t : Target tolerance: e q u i l . d e l a y a t P / P o >= 0 . 9 9 S :  Analysis Adsorptive: Analysis Bath: Thermal C o r r e c t i o n : Smoothed P r e s s u r e s : Cold Freespace: Low P r e s s u r e D o s e :  Report Nitrogen 9 2 5 . 0 0 mmHg 0.000066 0.0015468 3.860 A 0 . 1 6 2 nm« Yes 5 . 0 mmHg Yes 120 s e e s 0.5 hours Analysis 5 sees No S.O 4 o r 5 . 0 mmHg 600 s e e s  Free space group: Measured a f t e r f r e e s p a c e : No Evacuation time: 0.1 hours L e a k t e s t : No L e a k t e s t d u r a t i o n : 180 s e e s  Lower dewar  Low p r e s s u r e d o s i n g : No Dose amount: 0.00 e m ' / g STP Minimum e q u i l i b r a t i o n d e l a y : 0.00 hours Maximum e q u i l i b r a t i o n d e l a y : 9 9 9 . 0 0 h o u r s Po t y p e : M e a s u r e d Temperature type: Entered 77.3S K Temperature: M e a s u r e m e n t i n t e r v a l : 120 m i n u t e s Inside  diameter  o f sample  tube:  9 . 5 3 0 mm  N2 77.35 K No No 9 4 . 1 2 2 2 cm» None  Surface Area Measurements  Figure M.10: Isotherm Plot for Coked Silica-Alumina  Full ASAP 2010  V3.02  C  Unit  Report 1  Set Serial  Sample I d : m i x t u r e o f 4 batches o f coked O p e r a t o r I d : Negar Submitter Id: F i l e Name: C : \ . . . \ N E G A R \ C O K E M T X . S M P Started: Completed: Report Time: Sample H e i g h t : Harm F r e e s p a c e : Equil. Interval:  07/06/99 09:52:16 07/06/99 16:28:23 07/07/99 09:22:56 0.2248 g 2 9 . 6 9 4 9 cm* 5 sees  # 325  Page  SA  Analysis Adsorptive: Analysis Bath: Thermal C o r r e c t i o n : Smoothed P r e s s u r e s : Cold Freespace: Low P r e s s u r e D o s e :  N2 77.3S K No No 94.1222 cm None  1  Surface Area Measurements  Figure M . l l : Incremental Pore Volume for Coked Silica-Alumina  Micromeritics DET P l u s ( T M )  VI.02  { 2010  Sample I d : Operator Id: Submitter Id: File: Started: Completed: Report Timej Sample K e i g h t : Warm F r e e s p a c e :  Instrument  )  Corporation Unit  mixture Negar  of  4 batches  of  coked  1  Pagel2  SA  C : \ . . . \ N E G A R \ C O K E M I X . SMP  07/06/99 09:52:16 07/06/99 16:28:23 07/07/99 10:52:47 0.2248 g 2 9 . 6 9 4 9 cm*  Analysis Adsorptive: Analysis Bath: Thermal C o r r e c t i o n : Equil. Interval: Cold Freespace:  N2 77.35 K Ko 5 sees 94.1222 cm'  Porosity  D i s t r i b u t i o n by O r i g i n a l Density Functional Theory M o d e l : N2 G 7 7 K o n C a r b o n , S l i t P o r e s M e t h o d : N o n - n e g a t i v e R e g u l a r i z a t i o n ; No S m o o t h i n g  Volume Volume Area Total Area  Total  i i i i  n n n n  Pores Pores Pores Pores  < <= > >=  14.83 2727.29 2727.29 14.83  A A A A  0.03290 0.62731 0.000 246.562  cm*/g cm*/g m'/g m'/g  Incremental Pore Volume vs. Pore Width + Incremental Pore Volume 0.045  0.040  ! i  J_L  0.005  0.000 1e+02  Pore Width (Angstroms)  1e+03  Surface Area Measurements  Figure M.12:  Summary Report for Coked Silica-Alumina  Full ASAP 2010  V3.02  C  Sample Operator Submitter File  Unit Id: Id: Id:  Report  Set  1  Serial  «  32S  Page  23  Silica-Alumina Negar  Name:  C : \ . . .\NEGAR\SIL-ALUM.SMP  Started: Completed: Report Time: Sample Weight: Warm F r e e s p a c e : Equil. Interval:  02/10/99 17:41:02 02/11/99 02:36:22 02/11/99 08:59:01 0.3100 g 29.4947 cm 5 sees  Analysis Adsorptive: Analysis Bath: Thermal C o r r e c t i o n : Smoothed P r e s s u r e s :  1  Cold Freespace: Low P r e s s u r e D o s e :  Summary  N2 77.35 No No  K  94.4358 None  cm  1  Report  Area  Single  Point  BET S u r f a c e Langmuir  Surface  Area at  P/Po  0.19967543  :  Area:  Surface  Area:  BJH A d s o r p t i o n C u m u l a t i v e S u r f a c e A r e a o f p o r e s between 17.000000 and 3000.000000 A Diameter: BJH D e s o r p t i o n Cumulative Surface A r e a o f pores between 17.000000 and 3000.000000 A Diameter:  556.45S4  m'/g  585.1687  m'/g  818.008S  m'/g  650.77S8  m'/g  722.4857  m'/g  Volume  Single  P o i n t T o t a l Pore Volume o f p o r e s l e s s t h a n 2000.7097 A Diameter a t P/Po 0.99024468:  BJH A d s o r p t i o n Cumulative Pore Volume o f pores between 17.000000 and 3000.000000 A Diameter: BJH D e s o r p t i o n C u m u l a t i v e Pore Volume o f pores between 17.000000 and 3000.000000 A Diameter: Pore  Average  Pore  Diameter  (4V/A by  BET):  Pore Diameter  (4V/A) :  BJH  Pore  (4V/A) :  Diameter  em'/g  0.788460  em'/g  0.782719  craVg  Size  BJH A d s o r p t i o n A v e r a g e Desorption Average  0.751849  51.3936  A  48.4628  A  43.3348  A  

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.831.1-0059003/manifest

Comment

Related Items