Faculteit Ingenieurswetenschappen Chemische Proceskunde en Technische Chemie Laboratorium voor Petrochemische Techniek Directeur: Prof. Dr. Ir. Guy B. Marin
Single-event microkinetic modelling of the catalytic cracking of hydrocarbons over acid zeolite catalysts in the presence of coke formation
Author: Carmen M. Alonso Romero Promoters: Prof. Dr. Ir. G. B. Marin Prof. Dr. Lic. M.-F. Reyniers Coach: Ir. R. Van Borm
Thesis work submitted to obtain the degree of chemical engineer 2006 - 2007
FACULTEIT INGENIEURSWETENSCHAPPEN
Chemische Proceskunde en Technische Chemie Laboratorium voor Petrochemische Techniek Directeur: Prof. Dr. Ir. Guy B. Marin
Opleidingscommissie Scheikunde
Verklaring in verband met de toegankelijkheid van de scriptie
Ondergetekende, Carmen M. Alonso Romero afgestudeerd aan de UGent in het academiejaar 2006 - 2007en auteur van de scriptie met als titel:
Single-event microkinetic modelling of the catalytic cracking of hydrocarbons over acid zeolite catalysts in the presence of coke formation
verklaart hierbij: 1. dat hij/zij geopteerd heeft voor de hierna aangestipte mogelijkheid in verband met de consultatie van zijn/haar scriptie:
de scriptie mag steeds ter beschikking gesteld worden van elke aanvrager
de scriptie mag enkel ter beschikking gesteld worden met uitdrukkelijke, schriftelijke goedkeuring van de auteur
de scriptie mag ter beschikking gesteld worden van een aanvrager na een wachttijd van jaar
de scriptie mag nooit ter beschikking gesteld worden van een aanvrager
2. dat elke gebruiker te allen tijde gehouden is aan een correcte en volledige bronverwijzing
Gent, 20 august 2007
(Carmen M. Alonso Romero)
______Krijgslaan 281 S5, B-9000 Gent (Belgium) tel. +32 (0)9 264 45 16 • fax +32 (0)9 264 49 99 • GSM +32 (0)475 83 91 11 • e-mail: [email protected] http://www.tw12.ugent.be
Single-event microkinetic modelling of the catalytic cracking of hydrocarbons over acid zeolite catalysts in the presence of coke formation by Carmen M. Alonso Romero
Thesis work submitted to obtain the degree of chemical engineer Academic year: 2007
Universiteit Gent Faculteit Ingenieurswetenschappen
Promotors: Prof. Dr. Ir. G. B. Marin Prof. Dr. Lic. M.-F. Reyniers Coach: Ir. R. Van Borm
Overview
Fluid catalytic cracking is one of the major processes in refining. A point of interest concerning catalytic cracking is the formation and deposition of coke on the zeolite catalyst. Numerous efforts have been made to add knowledge of the FCC process.
In chapter 1 the Fluid catalytic cracking process and catalytic cracking catalyst will be discussed briefly. After that in chapter 2 coke formation and catalyst deactivation will be explained. Next, in chapter 3 both TEOM and recycle electrobalance reactor set up is described. Chapter 4, is devoted to explain SEMK (single-event microkinetic) modeling. The experimental results from the recycle electrobalance reactor of the catalytic cracking and coking of iso-octane and methylcyclohexane over different catalysts are presented in chapter 5. The experimental results obtained with TEOM reactor for the mixture (n-decane + methylcyclohexane + buthylcyclohexane/ 1-octene) over LZ-Y20 zeolite are in chapter 5 too. In chapter 6 the experimental data acquired with the TEOM reactor is used to estimate the kinetic parameters of the extended SEMK model to describe the cracking and coking process during the catalytic cracking of hydrocarbons. Finally, in chapter 7, the conclusions of the whole work performed are presented.
Keywords: catalytic cracking, single-event microkinetic modelling, Y zeolites, coke formation.
Single-event microkinetic modelling of the catalytic cracking of hydrocarbons over acid zeolite catalysts in the presence of coke formation
Carmen M. Alonso Romero
Supervisor: Rhona Van Borm
Abstract-The first part of this work is devoted to acquire The types of elementary steps of carbenium/carbonium ion experimental data to estimate the kinetic parameters of the chemistry most important in coke formation are mainly extended SEMK model to describe the cracking and coking alkylation, cyclization, (de)protonation, and hydride transfer. process during the catalytic cracking of hydrocarbons. The Cyclization does not increase the size of the carbenium ion experiments are performed in TEOM reactor and allow to study but leads to ring formation, a prerequisite for coke formation. the influence of process conditions on the kinetics of the cracking Transformation into aromatic rings is possible by a succession and coking process. In the second part the influence of process conditions on the kinetics of the cracking and coking is analyzed of deprotonation and hydride transfer steps [1]. for different zeolites using a recycle electrobalance reactor. Keywords-catalytic cracking, coke formation, single-event microkinetic modelling, Y-zeolites B. Single-event approach The single-event concept is based on the fundamental framework of transition state theory. This concept factors the I. INTRODUCTION structural contribution associated with a single elementary In recent petrochemical refineries fluid catalytic cracking step out of the elementary rate coefficient. The number of (FCC) is one of the principal processes applied to convert distinct configurations taken by a reactant and its transition crude oil into lighter more valuable transportation fuels. A state is related to changes in their symmetry numbers to point of interest concerning catalytic cracking is the formation account for all possible occurrences of identical single events and deposition of coke on the zeolite catalyst. The deposited making up the elementary step [2]. The elementary rate coke deactivates the catalyst leading to a decrease in coefficient k can be written as a function of the single event ~ hydrocarbon conversion and energetic efficiency of the FCC rate coefficient k : process. ~ To optimally operate an industrial FCC unit, understanding = eknk and quantifying the influence of feedstock composition, Where n is the number of single events, which is calculated process conditions and catalyst properties on the kinetics of e the catalytic cracking in the presence of coke formation is as the ratio between the global symmetry number of the indispensable. The kinetic modelling of the elementary reactant σ ,rgl and that of the transition stateσ gl,≠ . reactions occurring during the catalytic cracking of hydrocarbons is based on the single-event microkinetic III. EXPERIMENTAL RESULTS (SEMK) modelling approach, which has been developed at the “Laboratorium voor Petrochemische Techniek”. This work involves the use of a TEOM reactor, working at A. TEOM REACTOR different conditions of pressure, space time and temperatures With the experimental data of the catalytic cracking of a and using a mixture of n-decane+methylcyclohexane mixture n-decane + methylcyclohexane + butylcyclohexane / +butylcyclohexane/1-octene as feed over LZ-Y20. The 1-octene over LZ-Y20 zeolite in the TEOM reactor, the effect experimental data obtained allow to determine the influence of the operating conditions on conversion, product of process conditions on coke formation. distribution and coke formation is assessed. A quantitative In the second part of this work is based on experimental data analysis of the experimental data is very useful for the obtained with the recycle electrobalance reactor. The coke construction of the kinetic model. formation has been studied for two feeds, methylcyclohexane In figure 1 it is observed that n-decane conversion and iso-octane, different zeolites and different process dramatically drops during the first 20 minutes of reaction. conditions. This way, the influence of coke formation can be This occurs for all the components of the mixture and at all studied on different zeolites. reaction conditions applied. This is due to the deactivating effect of coke formation. It is generally known that the initial II. SINGLE-EVENT MICROKINETIC MODEL catalyst cracking activity is mainly associated to very strong Brønsted acid sites. The presence of strong acid sites accelerates the protolysis reactions, whose activation energy A. Coke formation in terms of elementary steps is typically high. Moreover, due to a higher carbenium ion Coke is defined as a carbenium ion with a size and structure lifetime adsorbed on strong acid sites than on weaker acid preventing its desorption, thus permanently covering the sites, propagation of the catalytic cracking reactions via active site(s) it was formed upon. hydride transfer is favored (Quintana-Solorzano, 2007).
active sites Conversion vs time (nC10) Catalyst Si/Al bulk Si/Al frame structure (mol NH3/kg) 100 W/F = 5 kgs/mol LZY20 H-USY 2,6 30 0,99 FAU 80 W/F = 10 kgs/mol CBV 720 H-USY 15 16 0,60 FAU 60 W/F = 15 kgs/mol W/F = 20 kgs/mol CBV 760 H-USY 30 100 0,23 FAU 40 W/F = 25 kgs/mol CBV 500 NH4-USY 2,6 3,9 1,50 FAU Conversion % Conversion 20 Table 1 Physical properties of the zeolites investigated. 0 0 2000 4000 6000 8000 10000 Time (s) of coke on the zeolites decreases in the following order: CBV
Figure 1 n-Decane conversion as a function of time for different space 500>CBV 720>LZY20>CBV 760. The Si/Al framework ratio times, at 753 K and 400 103 Pa. decreases in the same order. The high deposition of coke in CBV 500 might be due to a higher number of active sites (1.5 Conversion increases with increasing space time and mmol NH3/g). The same trend is observed for both feeds but temperature. Higher space time indicates a higher amount of the coke deposition is higher in case of iso-octane, because of acid sites available. It is well known that cracking reactions its higher reactivity. are favored at higher reaction temperatures. It also is important to point out that the coke formation was Conversion vs deposited coke found to be very rapid, especially in the first 20 minutes. A 80 sharp increase in coke content is observed. This indicates the 70 60 CBV500 existence of an autocatalytic effect, which explains the high 50 CBV720 40 CBV760 conversion during that period of time. It is observed also, that 30 LZY20 20 Conversion % the coke formation is higher when the space time increases as 10 was expected based on literature. It is shown in figure 2. 0 0.00 0.03 0.06 0.09 0.12 0.15 Deposited coke (kg coke/kg cat)
Deposited coke vs time Figure 4 Conversion as a function of deposited coke, T = 748 K, piC8 ~7 3 3.5E-02 W/F=5 Kgs/mol 10 Pa, W/F ~70 kg.s/mol, feed = iso-octane. 3.0E-02 W/F=15 kgs/mol 2.5E-02 W/F=20 kgs/mol 2.0E-02 IV. CONCLUSIONS 1.5E-02 W/F=25 kgs/mol 1.0E-02 coke/kg cat) 5.0E-03 The conversion of the mixture n- Deposited coke (kg 0.0E+00 0246810 decane+methylcyclohexane +butylcyclohexane/1-octene over Time (103s) LZ-Y20 zeolite improves with higher values of space time and temperature. In the TEOM reactor the deactivating effect Figure 2 Deposited coke as a function of time for different space times, 753 K and 400 103Pa total pressure. of coke on conversion occurs in the first 20 minutes of the reaction. This dramatic conversion drop indicates that coke Figure 3 represents the initial selectivities as a function of deactivation is selective to the strongest acid sites. conversion. This is very useful to evaluate the nature of the Coke deposition on Y zeolites increases with increasing formed reaction products. It is observed at every temperature framework Si/Al ratio of the zeolite: CBV500>CBV720> level that the selectivity of propane and isobutane increases LZY20>CBV760. It is also observed that when the reactor is with conversion in contrast to isobutene and propylene whose fed with iso-octane the conversion, the coking rate and the selectivity decreases. These results agree with the literature on deposited coke are higher compared to methylcyclohexane. cracking of n-decane/1-octene and butylcyclohexane/1-octene This indicates that iso-octane is more reactive than [3]. This trend is observed for both temperatures. methycyclohexane.
Selectivity vs Conversion V. FUTURE WORK
60 It would be interesting to work with different zeolites and a 50 Propane broader range of temperatures in the TEOM reactor to have a 40 Isobutane clear knowledge of the influence of the catalyst properties and 30 isobutene 20 isopentane the temperature on coke deposition. propylene 10 Selectivity, mol % For the recycle electrobalance reactor it is suggested to 0 020406080100 work with different partial pressures to know how this Conversion, mol% variable affects the coke deposition.
Figure 3 Initial product distribution as a function of nC10 + mchex + buchex initial conversion, at 723 K, P = 400 kPa and W/F = 5-25 kgcats/mol. REFERENCES [1] Moustafa T. M., Froment G. F. (2002) Kinetic modeling of coke B. RECYCLE ELECTROBALANCE REACTOR formation and deactivation in the catalytic cracking of vacuum gas oil, Ind. Eng. Chem. Res. 42 (41), 14-25. [2] Martinis J. M. Fromant G. F. (2006) Alkylation on solid acids. Part 2. The available experimental data allow to investigate the Single-event kinetic modeling. Ind. Eng. Chem. Res. 45 (3), 954-967. effect of acidity on methylcyclohexane and iso-octane [3] Quintana-Solorzano R. (2007) Single-event Microkinetics for coking in cracking. In table 1 the properties of the catalysts tested are catalytic cracking: Development and applications. Phd.Thesis at LPT, UGent (2007). represented. In figure 4 it is observed that the conversion decreases with deposited coke. With both feeds the deposition
i
TABLE OF CONTENTS
Chapter 1: Introduction ...... iii 1.1. Thesis content ...... 1 1.2. Justification...... 1 1.3. Fluid catalytic cracking...... 2 1.4. Typical process configuration of fluid catalytic cracking ...... 3 1.5. The catalytic cracking catalyst...... 7 1.6. Industrial feeds and typical products of the process...... 11 Chapter 2 : Coke formation and catalyst deactivation ...... 16 2.1. Coke definition and coke characterization...... 16 2.2. Coke nature...... 18 2.2.1. Insoluble coke...... 19 2.2.2. Soluble coke...... 19 2.3. Catalyst deactivation...... 20 2.3.1. Deactivation by poisoning ...... 21 2.3.2. Pore blockage...... 22 Chapter 3 : Procedures...... 24 3.1. Tapered element oscillating microbalance reactor ...... 24 3.1.1. TEOM® reactor: operating principle and characteristics...... 24 3.1.2. Description of the experimental set-up...... 28 3.1.2.1. Feed Section...... 29 3.1.2.2. Reaction Section ...... 29 3.1.2.3. Sampling and analysis Section ...... 30 3.2. Recycle electrobalance reactor ...... 32 3.2.1. Description of the recycle electrobalance reactor configuration ...... 32 3.3. Experimental data treatment ...... 33 3.3.1. Hydrocarbons outlet flow rates and carbon balances ...... 33 3.3.2. Conversion, yield and selectivity...... 35 3.3.3. Coke content determination...... 35 Chapter 4 : Single-Event Microkinetic modeling (SEMK) of fluid catalytic cracking ...... 37 4.1. Introduction...... 37 4.2. Carbenium ion chemistry of catalytic cracking ...... 38 4.3. Single-event microkinetics (SEMK)...... 42 4.4. Rate equations at single-event level...... 45 4.5. Relumping...... 45 4.6. Coke formation description via elementary reactions ...... 46 4.7. Relumped rate equations for coke formation...... 48 Chapter 5 : Experimental results ...... 50 5.1 Experimental results of the TEOM® Reactor ...... 51 5.1.1. Effect of operating conditions on activity...... 51 5.1.1.1 Influence of the space time ...... 51 5.1.1.2 Influence of the pressure...... 54 5.1.1.3. Influence of the temperature...... 55 5.1.2. Product distribution...... 57 5.1.3. Effect of operating conditions on coke formation ...... 62 ii
5.1.4. Conclusions...... 63 5.2 Experimental results of the recycle electrobalance reactor...... 65 5.2.1. Influence of the zeolite ...... 65 5.2.2. Influence of the feed ...... 68 5.2.3. Conclusions...... 69 Chapter 6 : Parameter estimation in presence of coke...... 71 6.1. Objective function...... 71 6.2. Statistics...... 73 6.3. Experimental reactor model equations ...... 74 6.4. Estimation of activation energies...... 76 6.5. Rate coefficients ...... 76 6.6. Results...... 79 6.7. Conclusions...... 82 Chapter 7 : Conclusions ...... 83
APPENDIX A: Operation of the set up………………………………………………………..85 1.Operational specifications of the TEOM 1500 PMA sensor unit...... 85 2.Procedure when loading and unloading the TEOM reactor……………………………...... 85 3.Checklist before performing an experiment ...... 87 4. Procedure when performing an experiment...... 88 4.1 Stop the regeneration program of the catalyst ...... 88 4.2 Set process conditions...... 88 4.3 Perform the GC-FID analyses...... 89 4.3.1 Online analysis: ...... 89 4.3.2 Offline analysis:...... 90 4.3.3 Storing a sample (for a later offline analysis):...... 90 4.4 Stopping procedure...... 90 4.5 Regeneration of the (deactivated) catalyst...... 91 5. Replacing the feed ...... 91 6. Additional information ...... 91
APPENDIX B: Experimental conditions of TEOM reactor………………………………….94
APPENDIX C: Experimental conditions of recycle electrobalance reactor………………...97 1. Experimental conditions for methylcyclohexane…………………………………………...…97 2. Experimental conditions for iso-octane……………………………………………………..…98
APPENDIX D: Calibration factors for the TEOM reactor…………………………………102
REFERENCES………………………………………………………………………………...105
iii
Abbreviations
Ãi aromatic with i carbons Ãi 13C-NMR Nuclear magnetic resonance for 13-carbon CF Calibration factor DME dimethylether EELS Electron Energy Loss Spectrometry EFAL Extraframework alumina EPR Electronic Energy Loss Spectrometry FCC Fluid catalytic cracking FCCU Fluid catalytic cracking unit FID Flame ionizez detector XRD X-ray difraction FTRI Fourier transformed infrared GC gas chromatography HF Heavy fuel HFAU Acid exchanged faujasite zeolite HPLC High performance liquid chromatography IR Infrared LCO Liquid cyclic oil LPT Liquefied petroleum gas MS Mass spectroscopy MFC Mass flow controller PI Pressure indicator PIC pressure indicator and controller PMA Pulse mass analyser alkanes-alkenes-cycloalkanes-aromatics referred to for PONA detailed REUSY Ultrastable Y zeolite exchanged with rate earths RSS Residual sumof squares in objective function SEMK Single-event microkinetic TCD Thermal conductivity detector TEOM Tapered element of the TEOM reactor TI Temperature indicator TIC Temperature indicator and control TSS Total sum of squares TST Transition state theory USY Ultra stable Y zeolite UV Ultra violet VGO Vacuum gas oil 129Xe- NMR Nuclear magnetic resonance for 129-Xe iv
FID Tapered element of the TEOM reactor List of symbols
A Preexponential factor s-1
C + -1 ,Ht Total concentration of acid sites on the catalyst mol(kgcat) - kgcoke(kgcats) 1 Cc Coke on catalyst E Activation energy kJ/mol f Frecuency Hz
Freg Calculated value for the stadistical F-test - Fº Molar flow rate mol mol(s)-1 h Planck’s constant Js H Enthalpy J/mol
°≠ Standard enthalpy difference between reactant and activated complex ΔH J/mol J/mol Ji Jacobian matrix for response i - kmol(kgcats) k Rate coeffient of a global reaction 1 - ~ kmol(kgcats) k Single-event rate coefficient 1
kBB Boltzmann constant J/K Mw Molecular mass Kg/mol ni,j Number of contributions present in component - n Order of reaction - ne Number of single events - nexp Number of experiments - nresp Number of responses - p Number of parameters to be estimated via regression -
Pt Total pressure kPa R Universal gas constant J/mol K 0 -1 rc Initial rate of coke formation kg(kgcats) -1 Rc Net production rate of coke kg(kgcats) ℜˆ 0 -1 ,kc Calculated average initial coking rate kg(kgcats) 0 ℜ -1 ,kc Experimental average initial coking rate for a given experiment k kg(kgcats) -1 Rij Net production rate of product j in experiment i, mol mol(kgcats) S entropy J(molK)-1 -1 Stransl Translational entropy J(molK) -1 Srot,ext Rotational external entropy J(molK) -1 Srot,int Rotational internal entropy J(molK) -1 Svib Vibrational entropy J(molK) v
~ -1 S rot int, Instrinsic rotational internal J(molK) ~ -1 S ,extrot Instrinsic rotational external entropy J(molK) mol (molc(alk) -1 Si Selectivity of product i converted) kg (kgc(alk) -1 Sc Coke selectivity fed) standardentropy difference between reactant and activated ΔS ,0 ≠ complex J(molK)-1 ~ Single-event standardentropy difference between reactant ΔS ,0 ≠ and activated complex J/(mol k) J(molK)-1 t Time on stream ks tcal t-student test calculated value - T Temperature K ˆ bV )( Covariance matrix -
Wjk Weight factor for experiment i and response k - W mass of the catalyst Kg -1 W/F ° space time kg.s/mol kgcats(mol) x molar fraction - X Conversion - y Molar fraction - mol (molmHC -1 -1 Yi Yield of product i mol (molmHC fed) fed) -1 yˆ ij Calculated yield of response j in experiment i mol(mol) y -1 ij Experimental yield of response j in experiment i mol(mol)
Greek symbols
β Vector optimal parameters θ+ fraction of Brönsted acid sites covered by carbenium ions + θ + + R1 fraction of acid sites covered by a R1 carbenium ion
σgl global symmetry
σin global symmetry number of the transition state
σext internal symmetry
r σ gl external symmetry
vi
Subscripts
alk alkilation alk_nucl nucleus alkilation alk_side side chain alkilation β Beta-scission c catalyst calc calculated cat catalyst c(alk) cycloalkane dep deprotonation carbenium exper experimental ext external i species hydrocarbon or lump i experimental int internal j group contribution j response species Lj lump j n carbon number r reaction reg regression rot rotational step step during the integrationof continuity equations tab trabulated value transl translational vib vibrational
Superscripts
~ Single-event º Initial, standard o reference state ˆ Average or calculated ˆ Intrinsic ≠ Activated complex or transition state + Carbenium ion Chapter 1: Introduction 1
Chapter 1: Introduction
1.1. Thesis content
This project involves the use of a TEOM reactor to acquire experimental data that will be used to estimate the kinetic parameters of the extended SEMK model to describe the cracking and coking process during the catalytic cracking of hydrocarbons.
The experiments performed will also allow to study the influence of process conditions on the kinetics of the cracking and coking process.
Moreover, the influence of process conditions on the kinetics of the cracking and coking process will also be studied for other zeolites tested by R. Van Borm using a recycle electrobalance reactor.
1.2. Justification In recent petrochemical refineries fluid catalytic cracking (FCC) is one of the principal processes applied to convert crude oil into lighter more valuable transportation fuels, gasoline being the most important part. A point of interest concerning catalytic cracking is the formation and deposition of coke on the zeolite catalyst. The deposited coke deactivates the catalyst leading to a decrease in hydrocarbon conversion and energetic efficiency of the FCC process. To deal with this coke phenomenon, it is necessary to use fluidized bed or riser technology to permit steady transfer of catalyst from the reactor to the regenerator, and, thus, operation in a permanent regime.
To optimally operate an industrial FCC unit, understanding and quantifying of the influence of feedstock composition, process conditions and catalyst properties on the kinetics of the catalytic cracking in the presence of coke formation is indispensable. The kinetic modelling of the reactions occurring during the catalytic cracking of hydrocarbons is based on the single-event Chapter 1: Introduction 2 microkinetic (SEMK) modeling approach, which has already been developed at the “Laboratorium voor Petrochemische Techniek”. Coke formation, like the main cracking reactions, was developed in terms of elementary steps, to limit the number of kinetic coefficients and to take advantage of analogies between coking steps and steps dealt with in the catalytic cracking.
1.3. Fluid catalytic cracking One of the most important processes in petroleum refining is the catalytic cracking of heavy hydrocarbons to produce aggregated value liquid fuels (mainly gasoline and liquified petroleum gas). Since the middle of the 20th century, catalysis has undergone remarkable development both from the point of view of fundamental knowledge and from that of its applications. At the start of the 21st century, between 80 and 90 % of the products we use in our daily life have “seen” a catalyst at some point during their manufacture. The development of industrial catalysis in Europe in the last three decades has been especially influenced by a series of political and economic events. (Marcilly 2002). In Table 1.1 a brief history is shown of the important events in the development of FCC catalysts. Development Date
Aluminum chloride catalyst 1915 Activated clay catalysts (Houdry) 1928 Silica/alumina catalyst (Houdry) 1940 First commercial production of FCC catalyst (Davison) 1942 Commercial production of microspheroidal catalysts (Davison) 1948 Last run of powdered catalyst 1956 Commercial production of zeolite Y (Union Carbide) 1959 Introduction of zeolitic FCC catalyst (Mobil) 1962 Introduction of USY and REUSY (Davison) 1964 Introduction of combustion promotion (Mobil) 1974 Feed-added Ni passivation technology (Phillips) 1975 Introduction of ZSM-5 (Mobil) 1986 Introduction of Ni-tolerant matrix technologies (Davison) 1990 Introduction of SOx reduction technology (Amoco) 1992 Introduction of additives for gasoline sulfur reduction (Davison) 1995 Introduction of CSSN zeolite (Davison) 1996 Table1.1. Important dates in FCC catalyst development (Scherzer, 1990) Chapter 1: Introduction 3
In the 1960s, the introduction of zeolite Y revolutionized the process by increasing gasoline selectivity, adding almost US$ 2 of value per barrel of feed processed (Gambicki, 2000). In the 1980s, the introduction of zeolite ZSM-5 dramatically improved the ability of refiners to increase both gasoline octane number and the yields of light olefins. In the early 1990s, the introduction of new alumina technologies further increased the flexibility of the FCC unit to process heavier crude sources with a higher tolerance to the contaminant metals Ni and V. (Harding et al. 2001).
Although the FCC process has been researched for more than half a century, new and important developments continue to be made in several areas to optimize the process operation. Modern FCC units can take a wide variety of feedstocks and can adjust operating conditions to maximize production of gasoline, middle distillate (LCO) or light olefins to meet different market demands. The success of developing new FCC technology requires the integration of in-depth understanding of the underlying process science and innovation in engineering practices.
1.4. Typical process configuration of fluid catalytic cracking A fluid catalytic cracking unit consists basically of a hydrocarbon feeding section, a riser-reactor (including disengager and stripper), a regenerator-flue gas handling and a fractionation tower. The different parts can be observed in Figure 1.1.
The feedstock, a mixture of gas oils, is preheated up to 450-600K and injected at the reactor bottom using Venturi. The preheated feed contacts the hot catalyst (870-900 K, which flows from the regenerator) and evaporates.
In the riser, the cracking reactions take place. Lighter hydrocarbons are produced as main reaction products but at the same time coke is deposited on the catalyst surface leading to a severe catalyst activity loss. Products and particles concentration are a function of the riser position during the cracking. Typical dimensions and operating conditions of the riser and reactor are shown in Table 1.2.
Chapter 1: Introduction 4
Figure 1.1. Simplified scheme of the different sections that integrate a FCC unit. Main symbols: G-feed preheater, E-riser, C-disengager, RX-reactor, RG-regenerator, MF-main fractionator tower, J-air blower, G-feed preheater. The vessel containing part of the riser, the disengager and the stripper is tipically denoted as “reactor” (Han, 2004).
The catalyst is separated from the hydrocarbons when the mixture leaves the riser. This separation is done by cyclones in a vessel. The velocity of the solids-vapors mixture decreases when the diameter of the vessel increases, in this way the mixture is separated. During the trip up the riser, the cracking catalyst is "spent" by reactions which deposit coke on the catalyst and greatly reduce activity and selectivity. The "spent" catalyst is disengaged from the cracked hydrocarbon vapors and sent to a stripper where it is contacted with steam to remove hydrocarbons remaining in the catalyst pores.
Chapter 1: Introduction 5
Table1.2. Typical dimensions and operating conditions of the riser and reactor (Milhaccea, 1994).
The fluid-bed regenerator vessel is used to burn off the coke deposited on the catalyst surface as a byproduct of the cracking process. The regenerated enters again at the bottom of the riser, repeating the cycle. In table 1.3 the typical dimensions and operating conditions of the regenerator can be observed.
Table1.3. Typical dimensions and operating conditions of the regenerator of the FCC unit (Milhaccea, 1994). Chapter 1: Introduction 6
Although FCC is a process that has been commercially deployed for over 60 years, the technology continues to evolve to meet new challenges, which include processing more difficult feedstocks and meeting more stringent environmental regulations. Modern FCC units can process a wide variety of feedstocks and can adjust operating conditions to maximize the production of gasoline, middle distillate (LCO) or light olefins to meet different market demands.
The feed injection system is by far the most critical breakthrough of modern FCC reactor design. Three recent developments have made the feed injection system increasingly important (Wang, 2004):
• Due to the development of a highly active zeolite FCC catalyst, the reaction time has been shortened to a few seconds in the modern riser reactor. • The regenerator temperature is getting higher to achieve more complete catalyst regeneration. The typical modern riser top temperature is in the range of 510–566ºC, but typical regenerated catalyst temperature is much higher, in the range of 677–760ºC. Feed injection reduces thermal cracking reactions by cooling off the lower riser quickly through fast mixing and vaporization of the feed. • FCC feedstock is getting heavier, which makes feed vaporization more difficult.
The newest generation of side-entry FCC feed nozzles generates more uniform feed distribution as a result of better control of homogeneity of two-phase flow and vaporization at the nozzle exit using two phase choke flow. Some older FCC units still retain the original feed injection system located at the bottom of the riser (bottom-entry nozzles). A new generation of feed injection technology uses a similar side-entry vaporization mechanism. For catalyst circulation, the bottom-entry nozzles have the advantage of reducing pressure drop through the riser. This system also enables longer riser residence time if riser height is limited.
The newest generation of FCC feed nozzles also provides faster mixing with catalyst because of sudden expansion of two-phase choke flow at the nozzle exit, creating a strong suction to draw in catalyst. Commercial experience has shown that the feed injection angle also plays a significant role in catalyst mixing, impacting temperature profile in the riser. Chapter 1: Introduction 7
Although most modern FCC units have feed nozzles installed through riser shrouds at a fixed angle, a new feed nozzle design enables feed injection angle adjustment while using existing riser shrouds. This enables optimized mixing of feed and catalyst by adjusting the injection angle to optimize FCC performance.
Commercial FCC operation has confirmed that using the newest generation feed nozzles optimizes the temperature profile in the riser and substantially reduces dry gas, thereby increasing gasoline yield. These results are in line with the expectation that better feed injection design reduces thermal cracking reactions, which are the primary source of dry gas. As a result, catalytic cracking reactions are maximized and more desirable products are produced.
1.5. The catalytic cracking catalyst The refining of crude oil is a major industry which makes heavy use of zeolites in many parts of the refining process. Zeolite is derived from a Greek word meaning "boiling stones", since zeolites loose water from their porous structure when heated. They have many important industrial applications due to their unique architecture, especially in the petrochemical industry.
Over 600 zeolites are known, and new synthetic zeolites are developed and patented every year by large chemical companies.
Zeolites are kinetically trapped, thermodynamically metastable polymorphs of silica. They have zeolites have the general chemical formula SiO2 in which a fraction of the silicon may have been replaced by another element (aluminum, for example). Zeolites are synthesized from hydrolysable silicon sources (alkylorthosilicates, silicon chlorides and fumed silica, for example) under moderate temperatures and pressures and high pH to form a complex network of tetrahedral bonds. What distinguishes a zeolites from the more common forms of silica such as amorphous silica and quartz is that this network of tetrahedral building blocks results in a microscale pore system that is incorporated directly into the crystal structure. While quartz is the Chapter 1: Introduction 8 most thermodynamically stable structure for silica, a zeolite is generally stable enough to require high temperatures to overcome the activation energy necessary to rearrange to the equilibrium structure. This is represented in figure 1.2.
Figure1.2. Schematic of the energetics of zeolite formation. (www.engineering.purdue.edu)
While any number of elements can be substituted for silicon in the zeolite framework (Al, Ga, Ti, Zr, V, Fe, Sn, for example). The addition of aluminum, a tri-valent cation, into the framework results in a negative charge in the crystal that must be compensated for with a positive charge. This charge compensation can be completed with any cation that is available; however, in order to be of catalytic interest that cation should be hydrogen. The standard synthesis techniques generally use sodium for charge compensation. The completed sodium form of the zeolite is then stirred in a solution containing ammonium ions which replace the sodium ions at the cation exchange sites in the zeolite. Once all of the sodium has been replaced by ammonium, the zeolite is dried and calcined to liberate ammonia. The resultant zeolite is now in its protonic or solid acid form.
Acid catalyzed reactions are ubiquitous in organic synthesis. From ring closure to isomerization, all of these reactions benefit from the presence of an acid promoter. In the past, many of these reactions had to be carried out by mixing reactants in the presence of potent mineral acids. In the case of the cracking of petroleum, raw feed stock was pressurized and heated causing large hydrocarbons to break into lighter more useful compounds. The introduction of the solid acid Y Chapter 1: Introduction 9 zeolite and Fluid Catalytic Cracking (FCC) technology revolutionized this process by replacing the inefficient free radical based homogeneous thermal process (which produced excessive amounts of coke and light gases) with a heterogeneous process based on carbocation chemistry reminiscent of the liquid phase. In addition to the solid acid nature of the catalyst, the pore system can be vital in the manipulation of the selectivity of acid catalyzed reactions. While liquid phase acid catalyzed reactions allow the formation of essentially any transition state, the introduction of a pore system containing the acid site requires the transition state to be able to fit in the pore system. Any transition states that are too bulky are not formed and thus selectivity can be improved by choosing solid acids with pore systems that are sterically incompatible with by-product transition states. The most commonly cited example of this would be the formation of some types of coke. The transition states for some forms of coke are too large to form within the pore system, therefore these types of coke are not formed.
Figure 1.3 - Brönsted and Lewis acid sites (Stöcker, 2005).
In the case of Y zeolite which has a very low silicon to aluminum ratio, the structure is not very thermally stable. In order to get Y zeolite to remain stable at elevated temperatures, a significant amount of the aluminum must be removed resulting in USY zeolite (Ultra Stable Y Zeolite). It may seem contradictory to concentrate on the removal of aluminum when aluminum is the cause of the exploitable chemical properties of the material; however, there are five major reasons to Chapter 1: Introduction 10 consider dealumination: enhanced thermal stability, increased pore size, increased hydrophobicity, generation of Lewis acidity and elimination of Brønsted acidity.
Zeolites with high aluminum contents such as Y zeolite can also have their pore systems sufficiently affected by dealumination by removing enough of the aluminum in the pore wall to create larger pores capable of accommodating molecules that were previously too large to enter the original pore system. In terms of hydrophobicity, the more aluminum sites and hence Brønsted acid sites, the more hydrophilic the zeolite. If a reaction requires the migration of a very hydrophobic molecule into the pore system, then dealumination would result in more hydrophobic zeolite and better reaction environment. While aluminum incorporation in the framework results in Brønsted acidity, extraframework aluminum (or more accurately aluminum oxyhydroxides) in the pore system result in Lewis acidity. Reactions that benefit from Lewis acidity would benefit from zeolites containing this extraframework aluminum. The elimination of Brønsted acid sites can serve two primary purposes: decrease in overall acidity or decrease in external acidity. In the cases where a zeolite is just too acidic for the reaction to proceed efficiently (possibly destroying the desired product before it can leave the pore system), reduction of the Brønsted acid sites via dealumination is warranted. While the pore system of a zeolite in combination with the acid properties results in acid catalyzed reactions with selectivity determined by what transition states can fit in the pore system, a complication is that there are acid sites on the external surface of the zeolite which are accessible to molecules not in the pore system. These molecules will react unselectively since there is no pore wall to prevent unwanted transition states from forming. As a result, selective dealumination of the external surface would benefit reaction selectivity by forcing the reaction to only occur within the confines of the pore system. Chapter 1: Introduction 11
Figure 1.4 Y-zeolite (www.iza-structure.org)
Y zeolite was first synthesized in the sodium form in 1964 by Union Carbide. (Breck, 1964). Intended for use as an adsorbent, Y zeolite eventually found a home as the heterogeneous catalyst for the Fluidized Catalytic Cracking (FCC) of hydrocarbons. The morphology is that of faujasite (FAU) with 12 membered rings forming a three dimensional pore system of 7.4 Å. The overall crystallographic space group is Fd3m with a cubic lattice parameter of about 24.3 Å. Due to the symmetry contained in the structure, there is only one T-site (the unique tetrahedral building blocks that make up the structure). Y zeolite is not a particularly siliceous zeolite with Si/Al ratio in the range of 1 to 25. This high aluminum content allows Y zeolite to be synthesized rather easily without the need for an organic structure directing agent.
1.6. Industrial feeds and typical products of the process
The use of thermal cracking units to convert gas oils into naphtha dates from before 1920. These units produced small quantities of unstable naphthas and large amounts of by-product coke. While they succeeded in providing a small increase in gasoline yields, it was the commercialization of the fluid catalytic cracking process in 1942 that really established the foundation of modern petroleum refining. The process not only provided a highly efficient means Chapter 1: Introduction 12 of converting high-boiling gas oils into naphtha to meet the rising demand for high-octane gasoline, but it also represented a breakthrough in catalyst technology.
The thermal cracking process functioned largely in accordance with the free-radical theory of molecular transformation. Under conditions of extreme heat, the bond between carbon atoms in a hydrocarbon molecule can be broken, thus generating a hydrocarbon group with an unpaired electron. This negatively charged molecule, called a free radical, enters into reactions with other hydrocarbons, continually producing other free radicals via the transfer of negatively charged hydride ions (H-). Thus, a chain reaction is established that leads to a reduction in molecular size, or “cracking,” of components of the original feedstock.
The use of a catalyst in the cracking reaction increases the yield of high-quality products under much less severe operating conditions than in thermal cracking. Several complex reactions are involved, but the principal mechanism by which long-chain hydrocarbons are cracked into lighter products can be explained by the carbenium ion theory. According to this theory, a catalyst promotes the removal of a negatively charged hydride ion from a paraffin compound or the addition of a positively charged proton (H+) to an olefin compound. This results in the formation of a carbenium ion, a positively charged molecule that has only a very short life as an intermediate compound which transfers the positive charge through the hydrocarbon. These transfers continue as hydrocarbon compounds come into contact with active sites on the surface of the catalyst that promote the continued addition of protons or removal of hydride ions. The result is a weakening of carbon-carbon bonds in many of the hydrocarbon molecules and a consequent cracking into smaller compounds. (Quintana-Solorzano, 2007)
Olefins crack more readily than paraffins, since their double carbon-carbon bonds are more friable under reaction conditions. Isoparaffins and naphthenes crack more readily than normal paraffins, which in turn crack faster than aromatics. In fact, aromatic ring compounds are very resistant to cracking. They readily deactivate fluid cracking catalysts by blocking the active sites of the catalyst. Table 1.4 illustrates many of the principal reactions that are believed to occur in fluid catalytic cracking unit reactors (www.britannica.com). The reactions postulated for olefin compounds apply principally to intermediate products within the reactor system, since the olefin content of catalytic cracking feedstock is usually very low. Chapter 1: Introduction 13
Table 1.4 Principal reactions in FCC (www.britannica.com)
Each type of hydrocarbons reacts under catalytic cracking conditions in certain definite ways. The major difference among hydrocarbons of a particular type lies in their crackability or extent of conversion for a given set of operating conditions. (Bollas, 2004). In all cases, for each type of molecule, increasing the molecular weight or carbon number increases the crackability. A variety of primary and secondary reactions take place during catalytic cracking. These include chain rupture, isomerization, cyclization, dehydrogenation, polymerization, hydrogen transfer, and condensation. Hence, the result of cracking even a simple molecule such as a normal paraffin is complex. Chapter 1: Introduction 14
Normal paraffins crack mostly to olefins and paraffins and give fair yields of very light gasoline (mostly C5 and C6 hydrocarbons). The normal paraffins are fairly difficult to crack. The reaction rates and products of paraffin cracking are determined by the molecular size and structure. Paraffinic molecules containing tertiary carbon atoms crack most readily, whereas quaternary carbon atoms are more resistant to cracking.
Naphthenes and isoparaffins tend to crack at about the same rate, but the product distributions are very different. Naphthenes produce relatively little gas and give excellent yields of gasoline. The gasoline is of better quality than that from paraffin cracking and contains appreciable quantities of aromatics, resulting from dehydrogenation of naphthenic rings.
Aromatics crack in several ways. The benzene ring is practically impossible to crack. Condensed-ring aromatics without side chains are converted to a limited extent, but almost entirely to coke. Alkylaromatics with side chains containing at least three carbon atoms crack extensively by shearing off the entire side chain. With long side chains, secondary reactions will occur, resulting in products similar to those from the cracking of olefins and paraffins. Generally, more aromatic feeds give poorer FCC yields. A contributing factor to this general trend is that, as the number of ring structures in the feed increases, the likelihood increases that dehydrogenation from contaminant metals will cause multi-ring aromatics to form, leading to condensation and coking of the catalyst. The molecular structure of the aromatic hydrocarbons is another important issue regarding their crackability. The distribution of aromatics among the degree of condensation clearly affects the rate of cracking. As the number of rings in a polynuclear aromatic molecule increases, the rate of cracking decreases, although the aromatic content appears to remain the same. (Lerner et al., 1997). The net result of the catalytic cracking of aromatic hydrocarbons is moderate yields of gas, very little gasoline, large quantities of very aromatic cycle stock, and high coke yields.
Olefins seldom appear in catalytic cracking feeds, but their reactions are of interest because they are the primary products of other cracking reactions. Olefins heavier than about C6 are extremely reactive. The products of olefin cracking are primarily propylene and butenes, along with butanes Chapter 1: Introduction 15 from secondary reactions. Some polymerization and cyclization takes place in olefin cracking to produce a small amount of cycle stock and fairly high coke yields.
Non-hydrocarbon contaminants, such as nitrogen, iron, nickel, vanadium, and copper compounds, act as poisons to cracking catalysts. Nitrogen, being basic, reacts with the acid centers on the catalyst and lowers the catalyst activity. However, the basicity of nitrogen compounds at cracking conditions can vary widely (Plank et al., 1955).Thus, the total nitrogen content is considered as a better indicator for cracking inhibition by basic nitrogen compounds (Service, 1960). The metals deposit and accumulate on the catalyst and cause a reduction in throughput by increasing coke formation. Finally, the sulfur content of an FCC feed has no major effect on the crackability of the feed, but it strongly affects the product distribution and quality (Service, 1960).
At first natural silica-alumina clays were used as catalysts, but by the mid-1970s zeolitic and molecular sieve-based catalysts became common. Zeolitic catalysts give more selective yields of products while reducing the formation of gas and coke.
The yield of light products (with boiling points less than 220° C) is usually reported as the conversion level for the unit. Conversion levels average about 60 to 70 percent in Europe and Asia and in excess of 80 percent in many catalytic cracking units in the United States. About one- third of the product yield consists of fuel gas and other gaseous hydrocarbons. Half of this is usually propylene and butylene, which are important feedstocks for the polymerization and alkylation processes. The largest volume is usually cracked naphtha, an important gasoline blend stock with an octane number of 90 to 94. The lower conversion units of Europe and Asia produce comparatively more distillate oil and less naphtha and light hydrocarbons. Chapter 2: Coke Formation and Catalyst Deactivation 16
Chapter 2 : Coke formation and catalyst deactivation
2.1. Coke definition and coke characterization
Zeolites are widely used as shape-selective acid catalysts in the petrochemical industry, and their deactivation due to coking has received considerable research attention over the years due to their commercial importance (Paweewan et al., 1997).
Hydrocarbon reactions over acidic zeolite catalysts are accompanied by the formation of heavy, low-boiling point, high-molar-mass by products that deposit on the surface of the catalyst and cause deactivation (Froment, 1997). These secondary byproducts remain trapped either in the zeolite pores or on the external surface of the crystallites. The term coke is used for the mixture of these numerous components that remain at ambient conditions after the removal of the reaction mixture and the term coke precursors for components intermediate in the complex reaction network that forms coke. The formation of coke occurs through a sequence of consecutive reactions consisting of alkene oligomerization and/or alkylation, cyclization and rearrangement of coke precursor molecules. Typical structures of coke precursors, and coke molecules are represented in figure 2.1.
Figure 2.1 Typical structures of coke precursors and coke molecules (Reyniers et al., 2000). Chapter 2: Coke Formation and Catalyst Deactivation 17
Flow conditions, mass and heat transfer play a significant role in the formation of coke. The formation and composition of carbonaceous compounds and thus the deactivation effect are determined by the network structure of pores. The role played by the size and shape of zeolite pores can be even more important than chemical properties such as strength and density of active sites (Guisnet et al., 1997).
In the conventional FCC process with a zeolite Y-based catalyst, a significant portion of the feedstock is converted into coke, i.e., typical coke yields range from 3 to 7% (Quintana- Solorzano, 2007). The heat produced by the combustion of coke is used in various ways: (i) to heat the feed to the reaction temperature, (ii) to provide energy to the endothermic cracking reactions, (iii) to heat the combustion air and (iv) the coke on the spent catalyst to the regenerator temperature, (v) to supply the heat lost from the reactor/regenerator, and (vi) to heat the steam to exit temperature (Wilson, 1997).
Although the understanding of catalyst deactivation by coke has seen substantial progress, detailed characterization of the deposits continue being complicated. The chemical identity of such coke components can be determined through various spectroscopic techniques such as EPR, C-NMR, Xe-NMR, UV and FTIR, and complemented with XRD, EELS and laser Raman spectroscopy. FTIR is also useful to determine the effect of coke deposits on the acid properties of the catalyst (Quintana-Solorzano, 2007). Most of these techniques operate in static mode but the deposition of carbonaceous materials can also be followed by in situ IR spectroscopy followed by on-line gas chromatography (GC).
However, in most instances, these approaches require total separation of coke from the catalyst, involving the destruction of the catalyst by HF followed by the dissolution of the soluble coke compounds in an organic solvent (Guisnet, 1989 and Koon, 2000). This is illustrated in figure 2.2. Chapter 2: Coke Formation and Catalyst Deactivation 18
Figure 2.2 Procedure for the separation of soluble and non soluble fractions from a coked zeolite by using HF solution (Quintana-Solorzano, 2007).
The heavy fraction corresponding to the insoluble coke is often only characterized by the average H/C ratio. As a consequence, little information on structure-property relationships exists despite the great number of physicochemical characterization tools applied.
2.2. Coke nature Catalysts are of considerable technological and economic importance to most chemical reactions, especially in the petrochemical industry. In general, the deactivation of catalysts results from the formation of coke on the catalysts during the reaction process.
Generally, coke is composed of polynuclear aromatics, polyaromatics, polyolefins, pre-graphite or graphite, the exact composition depending on the chemical reactions, catalysts and reaction temperature. The pre-separation of coke is a necessary step so that its composition can be determined by GC.
Coke is separated from catalysts by dissolving the catalysts in acids or bases. Then the coke is
divided into soluble coke and insoluble coke by extraction with C2H4Cl2.
Chapter 2: Coke Formation and Catalyst Deactivation 19
2.2.1. Insoluble coke The strong influence of the reactant composition on the coke nature is no longer observed for the non-soluble coke fraction, which is typically formed at high temperatures. Compared with low temperature coke, insoluble coke has a predominant polyaromatic character, lower H/C ratio, higher basic character and its composition is not strongly related to the reactant hydrocarbon nature. The pore geometry of the zeolite is also a determining parameter on the rate and nature of the insoluble coke formed. As the temperature increases, coke molecules evolve to bulkier structures (Quintana-Solorzano, 2007).
Appleby and coworkers (Guisnet et al., 1997) demonstrated that alkenes and polyaromatic hydrocarbons have large tendencies to form coke during the catalytic cracking of industrial feeds at temperatures between 723-773 K through a complex series of cyclisation, hydride transfer and condensation reactions.
Insoluble coke has been investigated by means of 13C MAS-NMR and C/H elemental analysis; however, it appears to be impossible to determine the individual composition of insoluble coke.
2.2.2. Soluble coke The formation of this type of coke has been generally associated to relatively low temperature, i.e. typically below 473 K, and has been reported to be intrinsically independent of the zeolite composition or structure and extremely dependent on the reactant nature.
Soluble coke, which is typically formed during the acid cracking of relatively simple alkenes, cycloalkenes, cycloalkanes or aromatics, is characterized by a relatively high content of cyclic and/or branched oligomeric structures formed via polymerization and condensation reactions.
The components of soluble coke may be identified individually by means of gas chromatography-mass spectrometry (GC-MS) analysis. As a result, the analysis of soluble coke may provide more important information about coke on catalysts than that of insoluble coke. Chapter 2: Coke Formation and Catalyst Deactivation 20
Soluble coke on catalysts may be converted into insoluble coke by further dehydrogenation and condensation reactions with increasing reaction time or reaction temperature. However, the composition of soluble coke cannot be determined completely by normal temperature GC, because some of the components are not sufficiently volatile and they are usually retained in the pores. In addition, some components of coke, which can be determined by GC, cannot be extracted from coke by C2H4Cl2. On the other hand, it is more reasonable to divide coke on catalysts into volatile coke and non-volatile coke, as most catalytic reactions are gas–solid phase reactions.
2.3. Catalyst deactivation Catalyst deactivation, the loss over time of catalytic activity or selectivity, is a problem of great economical concern in application of commercial catalytic processes. Catalyst deactivation is attributed to interaction between the catalyst and the impurities present in a process effluent in which the catalyst is used. Any chemical or physical interaction that reduces catalyst activity or selectivity is classified as catalyst deactivation phenomena. In general, deactivation leads to a shortened catalyst lifetime, and the replacement of an aged catalyst to a new one is determined by the industrial processes for which the catalyst is used. Industrial catalytic deactivation can range from short term to several years. Given that reduced catalyst lifetime has a strong negative impact on the process economics improved catalyst lifetime is of great commercial value. The causes of catalyst deactivation can be grouped into: chemical deactivation through reversible or irreversible poisoning; physical deactivation through fouling; thermal deactivation through sintering; loss of active material by vaporization; and mechanical deactivation through attrition or erosion (Petersen et al., 1987, Bartholomew, 2001, Forzatti et al., 1999, Chen et al., 1992, Moulijn et al., 2001).
Catalyst deactivation in the processes studied is mainly poisoning of the catalysts by impurities in the effluent gas.
Understanding of the catalyst deactivation mechanism and impact on catalyst performance is vital for further optimization of catalyst structures as well as physical and chemical properties with respect to tailoring crystal structures to resist deactivation. Chapter 2: Coke Formation and Catalyst Deactivation 21
2.3.1. Deactivation by poisoning Poisoning is defined as a loss of catalytic activity due to the chemisorption of impurities on the active sites of the catalyst. Usually, a distinction is made between poisons and inhibitors. Poisons are substances that interact very strongly and irreversibly with the active sites, whereas the adsorption of inhibitors on the catalyst surface is weak and reversible. In the latter case, the catalytic activity can be at least partly restored by regeneration. This irreversible/reversible or permanent/temporary nature of deactivation and the regeneration possibility of a catalyst are the main differences between poisoning and inhibition (Butt & Petersen, 1988, Forzatti & Lietti, 1999). However, the distinction between permanent and temporary poisoning is not always so clear, since strong poisons at low temperatures may be less harmful in high-temperature applications (Moulijn et al., 2001). Catalyst poisons can also be classified as selective or non- selective. The description of a poison as selective or non-selective is related to the nature of the surface and the degree of interaction of the poison with the surface. A poison can also be selective in one reaction, but not in another (Butt & Petersen, 1988).
Catalytic converters are poisoned by the impurities in fuel and lubrication oil, or by shavings from the exhaust tailpipe. Even low levels of impurities are enough to cover the active sites and decrease the performance of a catalytic converter. It follows that the analysis of poisoned catalysts may be complicated since the content of poison of a fully deactivated catalyst can be as low as 0.1 wt-% or even less (Forzatti & Lietti, 1999). Lead (Pb), sulfur (S), phosphorus (P), zinc (Zn), calcium (Ca), and magnesium (Mg) compounds are typical catalyst poisons (Liu & Park, 1993)
Three limiting models are identificated in deactivation by poisoning: (a) uniform site poisoning, (b) selective site poisoning, and (c) pore mouth poisoning. (Quintana-Solorzano, 2007)
In uniform site poisoning there is a homogeneous deactivating effect of the poison on the acid sites due to the zeolite activation which is assumed to occur uniformly throughout the particle. Uniform poisoning typically occurs when the diffusion rate of the poison on the zeolite is large compared to the rate constant of the coverage reaction.
There is selective site poisoning when some sites are more active than others and, hence, the poison deactivates the existing sites to a different extent. Chapter 2: Coke Formation and Catalyst Deactivation 22
In pore mouth poisoning, it is assumed that all the acid sites are uniformly distributed on the catalyst but the reaction rate is much more rapid than the diffusion rate. Due to the high reaction rate, the active acid sites near the external surface of the crystal contribute largely to the observed acidity, while those at the catalyst pores have a poor contribution. In this mode, the deactivation starts at the catalyst surface where the pore mouths are located. Thus the deactivation progresses inward. Towards the center of the particle the deactivated zone increases and the hydrocarbon molecules have to displace longer distances for reaching an active site, which reflects in an activity decay. The activity of the zeolite declines more rapidly than the number of deactivated sites and the diffusion decreases as the pores become blocked.
2.3.2. Pore blockage
Fouling covers all phenomena where the surface is covered with a deposit, e.g. with combustion residues such as soot or with mechanical wear. Coke formation is the most widely known form of fouling (it is even used as a synonym for fouling). Coke formation is not very clearly defined. There are probably as many mechanisms of coke formation as there are reactions and catalysts where this phenomenon is encountered. During the coke formation, carbonaceous residues cover the active surface sites, and decrease the active surface area. First, this block out the active compounds to reach the surface sites, and second, the amount of coke might be so large that carbon deposits block the internal pores in the catalyst. In many cases, hydrocarbons and aromatic materials are primarily responsible for coke formation. Among these other deactivation mechanisms, pore blocking is probably one of the most important mechanisms. Pore blocking is often connected to coke formation, and when the amount of coke is high on the catalyst’s surface, it may be possible for the coke itself to block off the pore structure (Butt & Petersen, 1988, Mouljin et al., 2001).
This deactivation mode largely depends on the pore topology of the zeolite. Apart from the amount of coke, the severity of the deactivation by coke depends on its localization, viz., pores, channels, cavities, outer surface, etc. and, hence, depending on the zeolite nature the deactivation effect may vary (Guisnet et al., 1989). Chapter 2: Coke Formation and Catalyst Deactivation 23
Guisnet and coworkers (Guisnet et al., 1997) proposed that three different deactivation modes occur within the deactivation by pore blockage. They referred to site coverage as mode 1, while for pore blockage three different modes denoted as mode 2, mode 3 and mode 4 were defined.
Figure 2.3 Schematic representations of the four modes of catalyst deactivation by coke (Magnoux et al., 1997).
In mode 2, the sites of cavities of channel intersections cannot be accessed due to the presence of an adsorbed coke molecule, namely, the acid site is inaccessible essentially by chemical reasons. Mode 3 is associated to the blockage of the access of channels and cavities, i.e. the internal pore volume, where coke molecules are basically formed via growth. Whereas mode 4 is designated to the blockage of the acid sites in channels and cavities where no coke molecules are located, i.e. it is produced by the steric effect of coke growth.
Chapter 3: Procedures 24
Chapter 3 : Procedures
3.1. Tapered element oscillating microbalance reactor The experimental work consisted of performing experiments aimed at studying catalytic cracking and coking reactions simultaneously in an inertial microbalance reactor commercially referred to as TEOM® Series 1500 Pulse Mass Analyzer.
The catalytic cracking of hydrocarbons on solid acid catalysts is accompanied by coke formation that deactivates the catalyst and changes its selectivity. The measuring of the rate of coke deposition is required for assessing its deactivation effect on the catalyst performance. Conventional microbalance reactors permit a simultaneous study of the main and coking reactions as a function of the coke content, nevertheless, there are some drawbacks detected during its operation (Chen et al., 1996, Hershkowitz et al., 1993): • A major portion of the fed gases do not flow through the catalyst bed • True space velocities cannot be determined • Very small changes in the catalyst bed cannot be accurately assessed The development of the new microbalance reactor system represented a solution to the problems cited above.
3.1.1. TEOM® reactor: operating principle and characteristics A Rupprecht & Patashnick TEOM 1500 Pulse Mass Analyzer (100 mg sample volume) was used in an experimental setup designed for the measurement of equilibrium adsorption in microporous materials such as zeolites. The active element of the TEOM consists of three main hardware components, viz., the sensor unit, the control unit and the interface computer. The sensor unit is composed by the tapered element, the optic components and heating parts, as illustrated in figure 3.1. Chapter 3: Procedures 25
Figure 3.1. Sensor unit including the tapered element, heat parts, optics, inlet and outlet gas positions, and thermocouples (Quintana-Solorzano, 2007).
The tapered element, made in Pyrex glass, contains at the lower the test material in its catalyst bed. The so-called TEOM® reactor is a fixed bed micro-reactor with a high-resolution microbalance that measures, in situ and in real time, changes in mass during gas-solid interactions. The catalyst load can be varied from 5-100 mg, depending on the sample bed size (Quintana-Solorzano, 2007).
The TEOM® reactor can work at high temperatures, i.e. up to 823 K, but at low-moderate pressures, up to 1000 kPa. In table 3.1 some of the temperature and pressure specifications of the TEOM® reactor are shown.
Chapter 3: Procedures 26
Table 3.1 Operational specifications of TEOM 1500 PMA sensor unit (Quintana-Solorzano, 2007).
A feedback system maintains the oscillation of the tapered tube, which natural frequency will change with mass variation. Based on the operating principle of the TEOM, the total mass uptake consists of the amount adsorbed and the mass change caused by the change of the gas density in the tapered tube.
If the TEOM is oscillating at the start of the experiment with the frequency of f0 and exhibits the frequency f1 after a mass uptake, the total mass uptake can be obtained as a function of f0, f1, and the spring constant K0 (Zhu et al., 1998). ⎛ 11 ⎞ (3.1) KMMM ⎜ −=Δ+Δ=Δ ⎟ a g 0 ⎜ 2 2 ⎟ ⎝ 1 ff 2 ⎠
ΔMa is the adsorbed mass and ΔMg is the change in the gas density determined by measuring a mass change during a blank run under the same conditions. From eq. 3.1 it follows that a relative mass change is measured with respect to the conditions at the beginning of the experiment. At the start of every experiment the mass is set to zero, and this corresponds with a frequency f0. This frequency depends slightly on the amount of sample in the tapered element and is on the order of
70 Hz. The spring constant K0, is a weak function of the temperature.
The control of the TEOM® 1500 PMA is performed by two microprocessors. The one in the control unit executes a modular code that manages incoming and outcoming analogous and Chapter 3: Procedures 27 digital signals. The other processor, which is written in LabVIEW® for Windows® software, monitors and controls the entire micro reactor system. Numerous valves, mass flow controllers and other gas preparation and cleanup devices, as well as downstream equipment, can be integrated into the system and controlled from the system’s interface. Modifying the LabVIEW- based software for the incorporation of additional hardware can be as easy as pulling up icons and positioning them for use. In figure 3.2 is shown the control scream of the TEOM reactor setup.
Figure 3.2 Control screen of the TEOM reactor setup in LabVIEW (www.rpco.com).
Furthermore, the profiles of the main operation variables, e.g., temperature, pressure, accumulated mass, etc can be visualized as a function of reaction time in graphics during the operation and numerically stored on the hard disk of the computer by means of ASCII files, i.e. *.raw, *.DT1 and *.DT2 (Quintana-Solorzano, 2007).
Chapter 3: Procedures 28
3.1.2. Description of the experimental set-up The TEOM sep-up can be divided in three different sections: feed section, reaction section and analysis/sampling section. It is represented in figure 3.3.
Figure 3.3 Flow diagram of the TEOM setup installed at the LPT, Ghent University (Quintana-Solorzano, 2007).
Chapter 3: Procedures 29
3.1.2.1. Feed Section In the feed section different feed lines can be distinguished, for the liquid hydrocarbons, helium, air and dimethyl ether.
The helium feed line has three ramifications, as helium is used as diluting gas for the liquid hydrocarbons, reactor startup gas and inert for the purge line. Air is used for regeneration of the catalyst. Dimethyl ether (DME) is used as internal standard enabling the calculation of the outlet molar flow rates of the different hydrocarbons. In the figure 3.3 the lines of nitrogen, hydrogen air and helium to the chromatograph are represented too. Each feed line is provided with an on/off valve, a check valve, a Kobold manometer (0-16 bar) and a Brooks 5850E mass flow controller.
The liquid hydrocarbons are fed by a Shimadzu LC-A10 HPLC pump and are mixed with helium and evaporated. The hydrocarbons/helium mixture is sent to a four-way valve during the stabilization stage (four-way valve in the OFF position) or to the reactor during the reaction stage (four-way valve in the ON position). The temperatures of the evaporator/mixer and the flow of helium must be controlled to ensure a proper vaporization of the hydrocarbons and a constant composition of the mixture. The total pressure in the stabilizing line and the reactor has to be controlled too and is done with manual back pressure regulators.
3.1.2.2. Reaction Section The reaction section consists of the sensor unit of the TEOM® 1500 PMA (the tapered element, heating parts and optic components). The TEOM reactor operates isothermally and is equipped with two thermocouples that indicate the temperature at the reactor wall and the reactor outlet. To ensure a good quality of frequency signal, the catalyst must be perfectly packed between two layers of quartz wool. It is represented in figure 3.4.
Chapter 3: Procedures 30
Figure 3.4 Catalyst package, gas inlet and outlet flow direction and optics position in the TEOM reactor (Quintana-Solorzano, 2007).
The total pressure in the reactor is adjusted by back pressure regulators. When the experimental conditions are stable, the reaction stage begins by switching the four-way valve to the ON position; it means that the reaction mixture is directed towards the reactor and the inert to the vent.
The effluent line from the sensor unit comprises a manual back pressure regulator, a mixing element for the exit gases and the internal standard (DME), a manometer, a metering valve and a condenser. The manual back pressure regulator is used to control the total pressure in the reactor. The mixing element mixes the effluent gases and the DME which is used as internal standard. The metering valve is used to adjust the slight overpressure needed to force the effluent stream through the analysis section.
3.1.2.3. Sampling and analysis Section The analysis section consists of a HP gas chromatograph (GC) equipped with a FID detector and a PONA capillary column. The carrier gas flow rate of the chromatograph is about 1 ml/min. The PONA (50 m x 0.20 mm x 0.5 µm) contains dimethylpolysiloxane as stable phase and is used for the separation of hydrocarbons. A split inlet was utilized since the capillary column only takes a very small part of the sample for the analysis. An efficient separation of the hydrocarbons in the range C1-C4 was archived by cooling down the column to 223 K with liquid nitrogen. The Chapter 3: Procedures 31 temperature was maintained 1 min in the oven at 223 K and was increased at a rate of 6 K/min up to 473. Hydrogen was not analyzed. The integration, processing and storage of the GC signals were carried out by means of the GC-software ATLAS 2000 V.4.85.
The sampling and analysis section contains a six-way valve and a ten-way valve in a heated box as it is shown in figure 3.5.
The on-line mode is when the sample is injected directly to the GC. If the sample is stored in one of the nine sampling loops of the ten-way valve, it is to be analyzed in the so-called off-line mode. In this way samples can be stored and analyzed afterwards, which is of particular importance in periods of rapid coke deposition.
Figure 3.5 Distribution of the six-way valve and ten-way valve in the TEOM reactor setup for storing and injecting samples of the reactor effluent to the GC (Quintana-Solorzano, 2007).
Chapter 3: Procedures 32
3.2. Recycle electrobalance reactor Beirnaert et al. (1994) presented a bench scale reactor concept: the combination of an electrobalance and recycle reactor. The use of a recycle reactor allows to work at gradientless conditions and at high conversion with an on-line gas chromatographic analysis of the reaction in this way effluent. To avoid limitations to low conversions, a recycle is added to the electrobalance reactor to ensure complete mixing. Under the conditions prevailing the rate of coke formation can be calculated directly from the increase in the catalyst mass, because then the rate of coke formation is uniform over the whole catalyst mass. The flow diagram of the recycle electrobalance setup is represented in figure 3.6.
3.2.1. Description of the recycle electrobalance reactor configuration The unit operates at temperatures up to 773 K at atmospheric pressure. Liquid and gaseous reactants can be fed. Amount of catalyst and space time can vary from 60-350 mg and 10 - 200 kg.s/mol respectively.
The reactor is made of stainless steel (AISI 316) and consists of two parts: a reactor tube and a recycle tube. The catalyst is placed in a basket which is connected with one arm of the electrobalance by kanthal wire. A thermocouple measures the temperature of the reactor gases just below the basket. The reactor pressure is measured at the top of the reactor (Beirnaert et al., 1994).
In the first minutes of the experiments, adsorption of hydrocarbons, pressure stabilization and gradual displacement of the pretreatment gas by the feed, disturb the measurement of coke content. The total amount of non-desorbable products can be calculated from the weight difference between the uncoked catalyst and the coked catalyst. Chapter 3: Procedures 33
Figure 3.6 Flow diagram of the recycle electrobalace reactor setup installed at the LPT (Diaz, 2007).
As the coking rate of the catalyst can be coupled with the catalyst activity and product selectivities, this type of reactor can be used for studying the kinetics of the ‘main’ process along with the deactivation by coke. For more information about recycle electrobalance reactor the interested reader is referred to Diaz (2007) and Beirnaert et al. (1994).
3.3. Experimental data treatment 3.3.1. Hydrocarbons outlet flow rates and carbon balances The pump controls the inlet flow rate of hydrocarbons, and the flows of diluent (helium) and internal standard are measured via mass flow controllers.
Chapter 3: Procedures 34
The outlet flow rates can be calculated from the flow rate of the internal standard and the composition of the outlet effluent determined via chromatographic analysis.
The molar flow rate of a hydrocarbon i contained in the reactor effluent can be calculated:
DME MwF DME ( )CFtA iI (3.2) i tF )( = DMECFA DME Mwi
Where:
FDME = Inlet molar flow rate of the internal standard.
Fi (t) = Outlet molar flow rate of the hydrocarbon i in the reactor effluent at a given time.
Ai (t) = Chromatographic surface area of the hydrocarbon i in the reactor effluent at a given time.
ADME = Chromatographic surface area of the internal standard.
MwDME = Molecular mass of the internal standard.
Mwi = Molecular mass of the hydrocarbon i.
CFDME = Calibration factor of the internal standard.
CFi = Calibration factor of the hydrocarbon i. To asses the quality of the experiment in terms of recovery mass, a carbon balance was performed:
nc (3.3) ∑ nF ii i %tC )( = i=1 n fc 0 ∑ Fn jj i=1
Fi = Outlet molar flow rate of the hydrocarbon i in the reactor effluent.
0 Fj = Inlet molar flow rate of the feed hydrocarbons. n j = number of carbon atoms of the hydrocarbon j contained in the feed. ni = number of carbon atoms of the hydrocarbon i contained in the reactor effluent. nc = number of carbon of hydrocarbons in the reactor effluent. Chapter 3: Procedures 35
3.3.2. Conversion, yield and selectivity The molar conversion of a (cyclo)alkane is defined as: 0 (3.4) )( − alkcalkc )( tFF )( aklc )( tX )( = 0 100 F alkc )(
The molar yield is calculated from:
i tF )( (3.5) i tY )( = 0 100 F alkc )(
The selectivity is calculated with the expression:
i tF )( (3.6) i tS )( = 0 100 )( − FF alkcalkc )(
The coking rate is obtained via derivation of a proper coke on catalyst (Cc) function with respect to time:
coke )( mtr cat (3.7) c tS )( = 0 100 [])( − alkcalkc )( )( MWtFF alkc )(
3.3.3. Coke content determination When the hydrocarbons mixture contacts the catalyst bed there is a mass change originating from density changes and hydrocarbons adsorption on the catalyst. When switching back the four-way valve from the mixture inert/hydrocarbons to inert, hydrocarbons that were reversible adsorbed will desorb from the catalyst, while others remain on it. The latter is assumed to be coke.
In the TEOM reactor the mass signal increases when the reaction starts due to a decrease in the
0 oscillating frequency of the reaction. Density changes (md) and hydrocarbon adsorption ( ma ) are also present. The mass signal gradually increases with time mainly as a result of coke deposition. Chapter 3: Procedures 36
At the end of the reaction the signal decreases due to density changes and hydrocarbon desorption
(mf). The equation to calculate the mass change per mass of catalyst can be written as: 0 (3.8) c ()( cda +−++= mmmmmtC fd )()
Figure 3.7 shows an example of how the relative mass changes of the catalyst bed typically evolve during a cracking/coking experiment in the TEOM reactor for the whole reaction period.
Figure 3.7 Diagram showing typical mass signal variations due to density changes, hydrocarbons adsorption and coke formation, during the reaction period for the cracking of n-decane/1-octene.
In the recycle electrobalance reactor density changes and hydrocarbon adsorption/desorption also occur when the reaction starts/stops. However, the mass signal decreases with time due to coke deposition. Another difference between the recycle electrobalance and the TEOM reactor is that the mass signal of the TEOM reactor is digital while that of the recycle electrobalance reactor is analogous. Chapter 4: Single-Event Microkinetic Modeling (SEMK) of Fluid Catalytic Cracking 37
Chapter 4 : Single-Event Microkinetic modeling (SEMK) of fluid catalytic cracking
4.1. Introduction
The feedstocks processed in petroleum refining and in many petrochemical processes generally consist of a homologous series of hydrocarbon families like paraffins, olefins, naphthenes, and aromatics. These series each contain a vast number of components, extending in a typical vacuum gas oil, e.g., from C15 to C40 and each of these components leads to complicated reaction pathways, contributing in their own way to the product distribution.
Conventional kinetic modeling using drastic lumping was widely used with the drawback of having rate coefficients that are feedstock dependent. Because of the complexity, but also because of incomplete chemical analysis, the kinetic modeling of these processes was based upon reduced networks consisting of a small number of reactions between pseudocomponents or lumped species, defined more by physical than by chemical properties. In recent years more detailed approaches have been developed. Liguras and Allen, 1989, described the conversion of vacuum gas oil (VGO) in terms of a relatively large number of pseudo-components, most of which are lumps in their own way. Klein et al., 1991, generated these pseudo-components from analytical characteristics using Monte-Carlo simulation. Instead, Quann and Jaffe, 1996 and Christensen et al., 1999, in their “Structure Oriented Lumping” expressed the chemical transformations in terms of typical structures of the molecules, without completely eliminating lumps and rate parameters that depend upon the feedstock composition.
SEMK modeling developed at the ‘Laboratorium voor Petrochemische Techniek’ at Ghent University at Ghent University has demonstrated to be effective in the kinetic modeling of acid catalyzed processes in general, and catalytic cracking in particular. It accounts for the detailed reaction network in catalytic cracking based on the classical carbenium ion chemistry occurring Chapter 4: Single-Event Microkinetic Modeling (SEMK) of Fluid Catalytic Cracking 38 on the Brønsted acid sites of zeolitic catalysts. (Feng et al., 1993 and Dewachtere et al., 1999). The values of these fundamental rate coefficients are invariant, regardless of the feedstock.
The formation of coke and its deactivation effect are very important issues to be addressed. Since both cracking and coking processes occur in a concerted manner, they must be simultaneously studied so that the effect of coking on the product distribution is inherently taken into account in the kinetics (Reyniers et al., 2000). A combination of SEMK rate equations for cracking of hydrocarbons and global rate equations to account for the effect of coke on the catalyst activity and selectivity was utilized by Beirnaert et al., 2001.
4.2. Carbenium ion chemistry of catalytic cracking Acid-catalyzed hydrocarbon reactions such as the zeolite-catalyzed processes encountered in petroleum refining proceed via carbenium ion intermediates (Feng et al., 1993).
Two special types have been suggested: carbenium ions are trivalent and carbonium ions are pentavalent or hexavalent. Both are represented in figure 4.1.
Figure 4.1 Differences in structure of carbenium and carbonium ions (Quintana-Solorzano R., 2007).
Carbenium ions have a tri-coordinated positively charged carbon atom containing three substituens that can be alkyl groups or hydrogen atoms. Due to the charge delocalization, there is an energy order for the carbenium ions. Depending on the number of alkyl groups bonded to the carbon charge bearing atom carbenium ions can be classified as primary, secondary or tertiary. Primary carbocations have one or zero carbons attached to the ionized carbon, secondary carbocations have two carbons attached to the ionized carbon, and tertiary carbocations have three carbons attached to the ionized carbon.
Chapter 4: Single-Event Microkinetic Modeling (SEMK) of Fluid Catalytic Cracking 39
Figure 4.2 Order of stability of examples of tertiary ( III ), secondary ( II ), and primary ( I ) carbenium ions.
Stability of the carbocation increases with the number of alkyl groups bonded to the charge- bearing carbon, that is shown in figure 4.2. Tertiary carbocations are more stable (and form more readily) than secondary carbocations; primary carbocations are highly unstable because, while ionized higher-order carbons are stabilized by hyperconjugation, unsubstituted (primary) carbons are not. Therefore, reactions such as the SN1 reaction and the E1 elimination reaction normally do not occur if a primary carbocation would be formed. An exception to this occurs when there is a carbon-carbon double bond next to the ionized carbon. Such cations as allyl cation CH2=CH- + + CH2 and benzyl cation C6H5-CH2 are more stable than most other carbocations. Molecules which can form allyl or benzyl carbocations are especially reactive.
On the other hand, carbonium ions are formed when a proton is donated to an alkane by a very strong acid. They are unstable penta-coordinated carbonations having a three center two-electron bond. Several structures have been proposed for representing carbonium ions: 1C2H (one carbon and two hydrogen atoms), 2C1H (two carbons and one hydrogen) and 3C (three carbon atoms). These are displayed in figure 4.3.
Figure 4.3 Different carbonium ions structures (a) 1C2H, (b) 2C1H (c) and 3C. (Quintana-Solorzano, 2007). Chapter 4: Single-Event Microkinetic Modeling (SEMK) of Fluid Catalytic Cracking 40
The difference in the average energy between C-H and C-C bond protonation increases with increasing the straight chain length of the hydrocarbon in the gas phase and also the energy differences are smaller than those for the corresponding carbenium ions. (Quintana-Solorzano, 2007).
The reaction pathways followed by hydrocarbons (and carbenium ions) in the so-called catalytic cracking of hydrocarbons are particularly complex due to the large number of hydrocarbons and the large number of reactions taking place. All these reactions, however, belong to a limited number of reaction families: (1) intramolecular reactions: (de)protonations, β-scissions, cyclisations and isomerizations, and (2) intermolecular reactions: hydride transfers, protolysis and alkilations.
Typical reaction families for acyclics, cyclics and aromatics are shown in tables 4.1.4.2 and 4.3 respectively.
Table 4.1 Reaction families considered in the reaction network departing from acyclic species. (Quintana- Solorzano, 2007). Chapter 4: Single-Event Microkinetic Modeling (SEMK) of Fluid Catalytic Cracking 41
Table 4.2 Reaction families considered in the reaction network departing from cyclic species. (Quintana- Solorzano, 2007).
Table 4.3 Reaction families considered in the reaction network departing from aromatic species. (Quintana-Solorzano, 2007). Chapter 4: Single-Event Microkinetic Modeling (SEMK) of Fluid Catalytic Cracking 42
4.3. Single-event microkinetics (SEMK) The single–event concept is based on the fundamental framework of TST (transition state theory). This concept factors the structural contributions associated with a single elementary step out of the TST rate coefficient. The number of distinct configurations taken by a reactant and its transition state is related to changes in their symmetry numbers to account for all possible occurrences of identical single events making up the elementary step (Martinis et al., 2006). The entropy changes from reactant to transition state have contributions related to the rotational (internal andexternal), translational and vibrational entropy:
= transl + ,extrot + rot int, + SSSSS vib (4.1)
It is assumed that the symmetry of the molecule is influenced by both external and internal rotational entropy:
~ (4.2) rot int, rot int, −= RSS lnσ int ~ (4.3) ,extrot ,extrot −= RSS lnσ ext
The TST equation for the elementary rate coefficient contains the enthalpy and entropy changes:
,0 ≠≠ ,0 ≠≠ ,0 ≠ Tk ΔG Tk S ΔΔ H (4.4) k = B e RT = B R ee RT h h
When optically active species are involved, the symmetry contribution has to be adjusted with the contribution owing to the mixing of the different enantiomers. For racemic mixtures, this nch contribution is equal to Rln2 , nch being the number of chiral carbons, therefore (Quintana- Solorzano, 2007):
(4.5) ~ ⎛ extσσ int ⎞ rot rot int, ,extrot rot −++= RSSSS ln⎜ ⎟ ⎝ 2nch ⎠
Chapter 4: Single-Event Microkinetic Modeling (SEMK) of Fluid Catalytic Cracking 43
(4.6) ⎛ extσσ int ⎞ σ gl = ⎜ ⎟ ⎝ 2nch ⎠
Accounting for the entropy difference between the reactant and the activated complex, eq. 4.1 results in:
~ ⎛σ r ⎞ (4.7) ,0 ≠ ,0 ≠ ,0 ≠ ,0 ≠ −Δ+Δ+Δ=Δ RSSSS ln⎜ gl ⎟ trans vib rot ⎜ ≠ ⎟ ⎝σ gl ⎠
Substituting eq. 4.7 in eq. 4.1:
~ ,0 ≠≠ ,0 ≠ ⎛σ r ⎞ Tk S ΔΔ H (4.8) k = ⎜ gl ⎟ B R ee RT ⎜ ≠ ⎟ ⎝σ gl ⎠ h
Where,
~ ,0 ≠ ,0 ≠ ,0 ≠ ~ ,0 ≠ (4.9) trans vib Δ+Δ+Δ=Δ SSSS rot
r The ratio of the global symmetry number of the reactant molecule (σ gl ) to the global symmetry
≠ number of transition state (σgl ) is defined as the number of single-events, ne:
⎛σ r ⎞ (4.10) n = ⎜ gl ⎟ e ⎜ ≠ ⎟ ⎝σ gl ⎠
Then, the rate coefficient of an elementary reaction k, can be expressed in terms of the single- ~ event rate coefficient, k :
~ (4.11) = e knk Chapter 4: Single-Event Microkinetic Modeling (SEMK) of Fluid Catalytic Cracking 44
In a network of reactions described at the molecular level the number of rate coefficients would equal the number of reactions. This is not so with the approach based upon elementary steps which is taken here. The large number of elementary steps belong to a limited and well defined number of reaction families: (de)protonation, β-scissions, isomerization, alkylation and hydride transfer. Moreover, these reactions families are decomposed in reaction types depending on the nature, primary, secondary or tertiary, of the participating carbenium ions.
Rate coefficients only depend on the symmetry of the reactant and of the transition state. This symmetry effect, denoted as the numbers of single events, ne, is defined as the ratio of the global symmetry number of the reactant to the transition state species. Explicitly accounting for this number of single events allows defining a unique single-event rate coefficient per reaction type (Baltanas et al., 1989). In table 4.4 the single-event rate coefficients per reaction family involved in the SEMK model are summarized. Those in italics can be obtained via reversibility.
Table 4.4 Single-event rate coefficients per reaction family involved in the SEMK model (Quintana- Solorzano, 2007). Chapter 4: Single-Event Microkinetic Modeling (SEMK) of Fluid Catalytic Cracking 45
4.4. Rate equations at single-event level
The reaction rate for the disappearance of a carbenium ion of type m, secondary or tertiary,
+ denoted as R1 , to produce a smaller carbenium ion of type n, secondary or tertiary, referred to as
+ R2 , and an alkene O through β-scission (Quintana-solorzano, 2007):
+ + R1 (m) Æ R2 (n) + O is written as follows:
++ ~ (4.12) β ( 21 =+→ β ),() θ + = e β nmknnmkORRr ),( θ + 1 mR )( 1 mR )(
Where:
kβ = rate coefficient for the elementary reaction.
~ β nmk ),( = single-event rate coefficient.
+ + θ + = fraction of acid sites of the catalyst covered by the carbenium ion R1 . R1
The concentration of carbenium ions is calculated by assuming pseudo-steady state approximation. Namely, the rate of formation of carbenium ions (via hydride transfer, protonation and protolysis) is equal to the rate of consumption (via deprotonation and hydride transfer).
4.5. Relumping
The single-event modeling requires a molecular analysis of the feedstock, but a detailed analysis of complex mixtures along with the resulting products after cracking is not feasible yet. Moreover, the number of species and the reaction possibilities grow exponentially with the carbon number, and hence, working with the reaction network can become impractical, so it is Chapter 4: Single-Event Microkinetic Modeling (SEMK) of Fluid Catalytic Cracking 46 necessary to work with a certain degree of lumping. An example of rate equation, in lumped form, of the transformation via β-scission of alkanes contained in lump L1 to alkanes in Lump L2 and alkanes in lump L3 is represented in equation 4.13
~ ~ ~ ( ) =+→ θ + )(),( + θ + )(),( + θ + sDstkntDtsknsDssknLLLr )(),( + (4.13) β 321 ∑ e β R1 ∑ e β R1 ∑ e β R1 ,ss ,ss ,ss ~ θ + tDttkn )(),( ∑ e β R1 ,ss
+ + The fractional coverage of the Brønsted acid sites with the carbenium ion R1 , θ + , has to be R1 calculated by assuming pseudo steady-state approximation. D(m) is the desorption term that accounts for the fraction of product carbenium ions that are desorbed either via hydride transfer to produce alkanes in L3 or via deprotonation yielding alkenes in L4. The interested reader is referred to Quintana-Solorzano (2007).
4.6. Coke formation description via elementary reactions
Coke is defined as a carbenium ion with a size and structure preventing its desorption, thus permanently covering the active site(s) it was formed upon. The literature agrees that coke mainly consists of polyaromatic moieties (Guisnet, 1997)
In the studies of several authors it has been stated that a limited number of families of elementary reactions are commonly present in the process of coke formation i.e. hydride transfers, alkylations, cyclisations, deprotonations, etc.
Coke formation affects the final product distribution and produces deactivation of the catalyst. Therefore, the calculation of the coke content on the catalyst by means of kinetic expressions is required.
The extension of the SEMK methodology to account for coke formation starts with the selection of appropriate species which can be considered as gas phase coke precursors (Quintana- Solorzano et al., 2004). The latter are defined as species, which lead via an irreversible step to Chapter 4: Single-Event Microkinetic Modeling (SEMK) of Fluid Catalytic Cracking 47 coke. These species are finally assumed to contribute directly to coke formation via relatively simple reaction pathways. The formation of the selected coke precursors is to be described in terms of the same elementary reactions as those encountered during the main cracking. This is innovative in the sense that the competition between the coking and cracking and, hence, the effect of coking on the product distribution is, at least partially, inherently accounted for in the kinetics. That complements the traditionally used deactivation functions which, in this methodology, would exclusively account for the decrease in available acid sites due to acid site coverage and to pore blockage by coke (Froment et al. 1997).
The order of reactivity towards coke of the differents families of hydrocarbons depends on their structure: polyaromatics > aromatics > alkenes > branched alkanes > normal alkanes.
Figure 4.4 Reaction scheme for coke formation in the catalytic cracking of VGO. (Moustafa et al., 2002).
Figure 4.4 shows a simplified representation of the coking scheme. The molecules considered in the model developed for the catalytic cracking of VGO are n-paraffins (individually up to C40) and further per C number i-paraffins; n-olefins; i-olefins; mono-, di-, tri-, and tetra-ring naphthenes, aromatics, and aromatic olefins; and mono-, di-, and tri-ring naphthenoaromatics. It Chapter 4: Single-Event Microkinetic Modeling (SEMK) of Fluid Catalytic Cracking 48 is clear from the preceding that coke, or the intermediate given that name, is not a dead species and may interact with other species of the reaction mixture.
4.7. Relumped rate equations for coke formation A certain degree of relumping is inevitable for practical applications. For this, formulas were developed to calculate lumped reaction rates through reconstruction of the single event kinetics. Lumping coefficients depend on the choice of the lumps and on the elementary step network, not on the fundamental rate coefficients. Global reactions rates converting lump L1 into lump L2 are obtained by summing over all elementary reaction rates converting species from lump L1 into species from lump L2.
Figure 4.5 illustrates the relumping methodology applied to the alkylation reactions for the coke formation from the coke precursors. The coke precursor from lump L1 undergoes hydride transfer to produce the corresponding carbenium ions. These are alkylated on the side chain with the olefins contained in lump L2. Evidently, each carbenium ion formed from coke precursors can react with each olefins in lump L2. As a result, the rate of consumption of the coke precursors contained in lump L1 necessarily corresponds to the summation of the rates of consumption of the species contained in that lump (Feng, 1993 and Dewachtere, 1999).
+ Figure 4.5 Relumping scheme of the side chain alkylation reactions of the group of CP ij carbenium ions formed out of a lump of gas phase coke precursors (L1) via hydride transfer with a lump of alkenes (L2) (Quintana-Solorzano, 2007).
The rate equation for the initial coke formation via side chain alkylation of the components belonging to the lump containing the coke precursors (triaromatics) denoted as L1, with the alkenes of lump L2, can be expressed by: Chapter 4: Single-Event Microkinetic Modeling (SEMK) of Fluid Catalytic Cracking 49
~ ~ ⎡ ),(),(),( + ),(),(),( PtsFtsLCtskPssFssLCssk ⎤ (4.14) 0 ⎢ β _ sidealk L2 β _ sidealk L2 ⎥ c )_( = Mwsidealkr coke ~ ~ ⎢+ ),(),(),( + ),(),(),( PtsFttLCttkPstFstLCstk ⎥ ⎣ β _ sidealk L2 β _ sidealk L2 ⎦
Where: P = Partial pressure of the lumps containing alkenes. l2 kβ (n, m) = Single-event rate coefficients of β-scission (they are used instead of the alkylation ones by applying the reversibility concept).
Mwcoke = Average molecular mass of coke (it is estimated by regression of experimental data). F Accounts for the fraction of acid sites that are covered by carbenium ions by applying the pseudo-steady state approximation. All the steps to calculate F are in Quintana-Solorzano, 2007.
As shown in figure 4.5, coke precursors have either anthracenic or phenantrenic structures. According to the reaction network, coke precursors can have 14 to 17 carbon atoms and, hence, only methyl, ethyl, or propyl side chains can be presents. Because many coke precursors are
0 lumped according to the carbon number the total rate of coke formation ( rc ) is given by (Quintana-Solorzano, 2007):
4 4 0 0 0 (4.15) c = ∑ rr CPsidealkc )__( + ∑ r CPnuclalkc )__( j=1 j=1
Chapter 5: Experimental results 50
Chapter 5 : Experimental results
Single-event MicroKinetic (SEMK) modeling is based on the fundamental chemistry involved in the catalytic cracking of hydrocarbons. Consequently, the corresponding rate coefficients are generally applicable for every type of feedstock and, hence, they can be obtained via experimental cracking of well-selected model hydrocarbons. Alkanes, cycloalkanes and aromatics are among the main families of hydrocarbons constituting a conventional catalytic cracking feed.
This chapter starts with a discussion of the obtained experimental data of the catalytic cracking of a mixture consisting of n-decane, methylcyclohexane, butylcyclohexane and 1-octene over LZ- Y20 zeolite in the TEOM reactor. The composition of the mixture can be found in table 5.1. The effect of the operating conditions on conversion, product distribution and coke formation is assessed. A quantitative analysis of the experimental data is very useful for the construction of the kinetic model. In table 5.2 a summary of the experimental conditions is shown, this information can be found more detailed in appendix B. Composition (mol/mol mixture) 1-octene 0.158 n-decane 0.274 methylcyclohexane 0.279 butylcyclohexane 0.289 Table5.1 Composition of the feed Temperature [K] 693 693 693 693 693 693 693 723 723 723 723 723 723 723 753 753 753 753 753 753 753 Total P 400 400 400 400 400 300 200 [103 Pa] W/F 5.0 10.0 15.0 20.0 25.0 10.0 10.0
[kgcat s/ mol mixture] Table5.2 Experimental conditions investigated for the catalytic cracking in the TEOM reactor. Chapter 5: Experimental results 51
In the second part of this chapter, the experimental data obtained by R. Van Borm with the recycle electrobalance reactor will be discussed. The influence of process conditions on the kinetics of the cracking and coking process were investigated too. Different zeolites were investigated using an iso-octane or methylcyclohexane as feed. A summary of the experimental conditions is shown in table 5.3.
3 ZEOLITE FEED T (K) PT (10 Pa) W/F (kgcat s/mol) CBV 500 Iso-octane 723 ~100 32.21 - 84.01 CBV 500 Iso-octane 748 ~100 10.30 - 143.77 CBV 720 Iso-octane 723 ~100 21.74 - 131.55 CBV 720 Iso-octane 748 ~100 14.06 - 86.46 CBV 760 Iso-octane 723 ~100 84.01 - 214.83 CBV 760 Iso-octane 748 ~100 12.96 - 174.57 LZ-Y20 Iso-octane 748 ~100 8.33 - 79.68 Y 62 Iso-octane 748 ~100 16.97 - 161.16 Y 62 Iso-octane 723 ~100 74.49 - 83.01 Y 62 Iso-octane 763 ~100 46.88 - 84.11 CBV 500 Methylcyclohexane 748 ~100 51.18 - 67.60 CBV 720 Methylcyclohexane 748 ~100 42.49 - 62.23 CBV 760 Methylcyclohexane 748 ~100 30.61 - 43.18 LZ-Y20 Methylcyclohexane 748 ~100 33.06 - 53.57 Table 5.3 Experimental conditions investigated for the catalytic cracking in the recycle electrobalance reactor.
5.1 Experimental results of the TEOM® Reactor 5.1.1. Effect of operating conditions on activity
5.1.1.1 Influence of the space time In figures 5.1, 5.2, 5.3, and 5.4, the conversion profiles as a function of time for different space times, are shown for every component of the mixture. In all the cases, except in case of 1-octene, Chapter 5: Experimental results 52 it is clear that the conversion dramatically drops during the first 20 minutes of reaction. This is due to the deactivating effect of coke formation. It is generally known that the initial catalyst cracking activity is mainly associated to very strong Brønsted acid sites. The presence of strong acid sites accelerates the protolysis reactions, whose activation energy is typically high. Moreover, due to a higher carbenium ion lifetime adsorbed on strong acid sites than on weaker acid sites, propagation of the catalytic cracking reactions via hydride transfer is favored (Quintana-Solorzano, 2007).
However, the catalyst activity decay becomes very slow at higher times on stream indicating that, after a period of strong deactivation, the remaining fraction of acids sites still active for the cracking of hydrocarbons is very slowly deactivated by coke. Meloni et al., 2001 and Hopkins et al., 1996, have also reported this in case of cracking of pure hydrocarbons on pure zeolites. These acid sites do not deactivate by coke and allow the cracking of hydrocarbons after some time.
In figures 5.1-5.3 the effect of space time on the conversion can also be observed. Conversion increases with increasing space time. Higher space time indicates a higher amount of catalyst active sites available for the reaction, and therefore a higher conversion. This is observed at every temperature level.
Conversion vs time (nC10) 100 W/F = 5 kgs/mol 80 W/F = 10 kgs/mol
60 W/F = 15 kgs/mol W/F = 20 kgs/mol 40 W/F = 25 kgs/mol
Conversion % 20
0 0 2000 4000 6000 8000 10000 Time (s)
Figure 5.1 n-Decane conversion as a function of time for different space times, at 753 K and 400 103 Pa. Chapter 5: Experimental results 53
Conversion vs time (buchex) 100 80 W/F = 5 kgs/mol 60 W/F = 10 kgs/mol 40 W/F = 15 kgs/mol 20 Conversion % W/F = 20 kgs/mol 0 W/F = 25 kgs/mol 0 2000 4000 6000 8000 10000 Time (s)
Figure 5.2 Butylcyclohexane conversion as a function of time for different space times, at 753 K and 400 103 Pa.
Conversion vs time (mchex) 100
80 W/F = 5 kgs/mol 60 W/F = 10 kgs/mol 40 W/F = 15 kgs/mol W/F = 20 kgs/mol
Conversion % 20 W/F = 25 kgs/mol 0 0 2000 4000 6000 8000 10000 Time (s)
Figure 5.3 Methylcyclohexane conversion as a function of time for different space times, at 753 K and 400 103 Pa.
1-Octene conversion was always higher than 98%, regardless of temperature, pressure, space time or time on stream. In case of 1-octene, it is observed that there is no dependency between conversion/time. The high conversion of this component indicates that it is responsible for the catalyst deactivation, also due to its low concentration in the mixture the secondary reactions of 1-octene are less important at the deactivation of the catalyst. Chapter 5: Experimental results 54
Conversion vs time (1-octene)
100.0 W/F = 5 kgs/mol 99.8 W/F = 10 kgs/mol 99.6 W/F = 15 kgs/mol 99.4 W/F = 20 kgs/mol 99.2 W/F = 25 kgs/mol
Conversion % 99.0 98.8 0 2000 4000 6000 8000 10000 Time (s)
Figure 5.4 1-Octene conversion as a function of time for different space times, at 753 K and 400 103 Pa.
5.1.1.2 Influence of the pressure The influence of the total pressure was also studied and is represented in figures 5.5, 5.6 and 5.7. The effect of the pressure is not very clear. The largest differences were found during the first minutes of the reaction, namely when the catalyst is being deactivated fast. An increase in the pressure results in an increased conversion for all the components, as a result of the higher adsorption of hydrocarbons on the catalyst surface that promote the hydride transfer reaction.
Conversion vs time (nC10)
50 40 Ptotal = 400 kPa 30 Ptotal = 300 kPa 20 Ptotal = 200 kPa 10 Conversion % 0 0 5000 10000 Time (s)
Figure 5.5 Conversion of n-decane as a function of time for different pressures (723 K, W/F = 10 kgs/mol)
Chapter 5: Experimental results 55
Conversion vs time (mchex)
80
60 Ptotal = 400 kPa 40 Ptotal = 300 kPa 20 Ptotal = 200 kPa
Conversion % 0 0 5000 10000 Time (s)
Figure 5.6 Conversion of methylcyclohexane as a function of time for different pressures (723 K, W/F = 10 kgs/mol)
Conversion vs time (buchex)
100 80 Ptotal = 400 kPa 60 Ptotal = 300 kPa 40 20 Ptotal = 200 kPa Conversion % 0 0 5000 10000 Time (s)
Figure 5.7 Conversion of butylcyclohexane versus time for different pressure (723 K, W/F = 10kgs/mol)
5.1.1.3. Influence of the temperature Depending on the operating conditions, the deposition of coke on the zeolite varies between 0.006-0.03 kg coke/kg catalyst.
In figures 5.8, 5.9 and 5.10 is observed that during the first minutes of reaction the conversion decays fast independently of the component and of the temperature. Conversion increases with increasing temperature. It is well known that cracking reactions are favored at higher reaction temperatures.
Chapter 5: Experimental results 56
Conversion vs time (nC10)
60
40 T = 693 K T = 723 K 20 T = 753 K Conversion % Conversion 0 0 3000 6000 9000 12000 15000 Time (s)
Figure 5.8 Conversion of n-decane as a function of time for different temperatures (W/F = 5 kg·s/mol, PT = 400 kPa).
Conversion vs time (mchex)
60 T = 693 K 40 T = 723 K
20 T = 753 K Conversion % 0 0 3000 6000 9000 12000 15000 Time (s)
Figure 5.9 Conversion of methylcyclohexane as a function of time for different temperatures (W/F = 5 kg·s/mol, PT = 400 kPa).
Conversion vs time (buchex) 80
60 T = 693 K 40 T = 723 K T = 753 K 20 Conversion % 0 0 3000 6000 9000 12000 15000 Time (s)
Figure 5.10 Conversion of butylcyclohexane as a function of time for different temperatures (W/F = 5 kg·s/mol, PT = 400 kPa).
Chapter 5: Experimental results 57
5.1.2. Product distribution In figures 5.11-5.13, the most important pathways are represented in the cracking of n-decane, methylcyclohexane and butylcyclohexane respectively.
Figure 5.11 Reaction pathways at the carbenium ion level for the cracking of n-decane in the presence of 1-octene. P, O and C+ are for alkanes, alkenes and alkylcarbenium ions, respectively. Mo, di and tr are for methyl, dimethyl and trimethyl and substituted alkanes alkenes and alkylcarbenium ions. All the steps are reversible (except for protolysis) but in order not to overload the figure the reserve steps were not drawn. Reaction families are identified by different colors, viz., black: protolysis, red: PCP-isomerizations, blue: β-scissions, purple: hydride transfer, green: deprotonation and brown: set of reactions occurring on carbenium ions formed via protolysis (Quintana-Solorzano, 2007). Chapter 5: Experimental results 58
Figure 5.12 Reaction pathways at the carbenium ion level for methylcyclohexane. N, N+, cO, A, P, O, O+ and C+ are for cycloalkanes, cycloalkylcarbenium ions, cycloalkenes aromatics, alkanes, alkanes, alkenylcarbenium ions and alkylcarbenium ions respectively. Mo, di and tr are for methyl, dimethyl and trimethyl and substituted species and et for and ethyl substituent of cycloalka(e)nes (or cycloalkylcarbenium ion). All the steps are reversible (except for protolysis) but in order not to overload the figure the reserve steps were not drawn. Reaction families are identified by different colors, viz., black: protolysis, red: PCP-isomerizations, blue: acyclic β-scissions, brown: endocyclic β-scissions, purple: hydride transfer, green: deprotonation and grey: steps leading to aromatics (Quintana-Solorzano, 2007). Chapter 5: Experimental results 59
Figure 5.13 Reaction pathways at the carbenium ion level for butylcyclohexane. Symbols description as in figure 5.12 plus ms, ds, trs and tes for mono, di, tri and tetracycloalkane substituted hydrocarbons and DA for diaromatics. All step are reversible except for the reaction to coke and prololysis but have been omitted. Reaction families are identified by different colors, viz., black: protolysis, red: PCP-isomerizations, blue: acyclic β-scissions, brown: endocyclic β-scissions, purple: hydride transfer, green: deprotonation and gray: steps leading to aromatics (Quintana- Solorzano, 2007). Chapter 5: Experimental results 60
The evolution of the selectivity per carbon number and per hydrocarbon family with time can be seen in figure 5.14. The main reaction products of cracking the mixture n-decane + methylcyclohexane + butylcyclohexane/1-octene are propylene, 1-butene, 2-butene, and alkanes such as propane, isobutene and isopentane.
60.0 100.0 O2 55.0 P1 P2 O3 50.0 P3 80.0 O4 45.0 P4 O5 40.0 P5 O6 35.0 P6 60.0 O7 30.0 P7 O8 P8 25.0 selectiv, mol% selectiv, mol% 40.0 P9 20.0 15.0 20.0 10.0 5.0 0.0 0.0 012345678 012345678 time, ks time, ks
20.0 16.0
A6 18.0 N5 14.0 16.0 A7 N6 12.0 A8 14.0 N7 10.0 A9 12.0 A10 N8 10.0 8.0
selectiv, mol% N9 8.0 selectiv, mol% 6.0 6.0 4.0 4.0 2.0 2.0
0.0 0.0 012345678 02468 time, ks time, ks
70 48.0 nO4 44.0 nP4 60 iO4 40.0 nO5 iP4 36.0 50 iO5
nP5 % 32.0 nO6 iP5 40 iO6 28.0 nP6 nO7 24.0 iP6 30 iO7
20.0 mol selectivity, selectivity, mol% nP7 nO8 16.0 20 iO8 iP7 12.0 iP8 8.0 10 4.0 0 0.0 012345678 012345678 time, ksec time, ksec Figure 5.14 Product selectivities as a function of the per carbon number and per hydrocarbon family (T = 723 K, P = 400 kPa and W/F = 5kg s/mol). Chapter 5: Experimental results 61
Figures 5.15 and 5.16 represent the initial selectivities as a function of the conversion. This is very useful to evaluate the nature of the formed reaction products. At higher conversion it is observed at every temperature level that the selectivity of propane and isobutane increases with conversion in contrast to isobutene and propylene that decreases. These results agree with the literature on cracking of n-decane/1-octene and butylcyclohexane/1-octene (Quintana-Solorzano, 2007). The decrease or increase in the selectivity to certain products with conversion is associated to secondary reactions, in particular hydride transfer and β-scission of primary products. Isoalkanes are primary products but they are also produced via secondary reactions as their selectivity increases with conversion. Normal alkanes are only produced via primary cracking. Propane selectivity increases because of secondary cracking contribution. The ratio of isoalkanes to normal alkanes is clearly higher than one and increases with the conversion. (iso)Alkenes are unstable primary products, which are highly susceptible to secondary reactions as they systematically decreased with increasing conversion. It is clear that the selectivity to alkanes increases at the expense of alkenes via hydride transfer reaction (Corma et al., 1991).
Selectivity vs Conversion
60 50 Propane 40 Isobutane 30 isobutene 20 isopentane propylene 10 Selectivity, mol % 0 0 20406080100 Conversion, mol%
Figure 5.15 Initial product distribution as a function of nC10 + mchex + buchex initial conversion, at 723
K, P = 400 kPa and W/F = 5-25 kgcats/mol.
Chapter 5: Experimental results 62
Selectivity vs conversion
80 70 Propane 60 50 Isobutane 40 Isobutene 30 Isopentane 20 propylene Selectivity, mol% 10 0 0 20406080100 Conversion, mol%
Figure 5.16 Initial product distribution as a function of nC10 + mchex + buchex initial conversion, at 753
K, P = 400 kPa and W/F = 5-25 kgcats/mol.
5.1.3. Effect of operating conditions on coke formation Figure 5.17 illustrates the coke content as a function of time on stream at 753 K. Coke formation was found to be very rapid, especially in the first 20 minutes. A sharp increase in coke content is observed. This indicates the existence of an autocatalytic effect, which explains the high conversion during that period of time. It is observed also, that the coke formation is higher when the space time increases as was expected based on literature references (Quintana-Solorzano, 2007).
Deposited coke vs time
3.5E-02 W/F=5 Kgs/mol
3.0E-02 W/F=15 kgs/mol 2.5E-02 2.0E-02 W/F=20 kgs/mol 1.5E-02 W/F=25 kgs/mol 1.0E-02 coke/kg cat) 5.0E-03
Deposited coke (kg 0.0E+00 0246810 Time (103s)
Figure 5.17 Deposited coke as a function of time for different space times and 753 K and 400 103Pa of total pressure.
Chapter 5: Experimental results 63
Figures 5.18 and 5.19, show the amount of deposited coke versus time for different temperatures. In contrast to what is mentioned in literature (Reyniers et al., 2000), the experimental data show that an increase in the temperature results in an increase in coke formation. When analyzing coke formation in terms of the operating variables studied, normally, two contrary effects have to be accounted for: at higher reaction temperatures more coke precursors are available due to higher conversion, nevertheless, coke precursors formed can be more easily desorbed from the catalyst. In this case this behaviour can be explained by the lower surface coverage at higher temperature, although the rate coefficient is increased.
Deposited coke vs time
0.035 0.030 0.025 T = 693 K 0.020 T = 723 K 0.015 T = 753 K 0.010 coke/kg cat) 0.005 Deposited coke (kg 0.000 0246810 Time (103s)
3 Figure 5.18 Deposited coke as a function of time (PT = 400 10 Pa, W/F = 10 kg·s/mol).
Deposited coke vs time
0.030
0.020 T = 693 K 0.010 T = 723 K
0.000 T = 753 K coke/kg cat) 0.00 5.00 10.00
Deposited coke (kg -0.010 Time (103s)
3 Figure 5.19 Deposited coke as a function of time (PT = 400 10 Pa, W/F = 20 kg·s/mol).
5.1.4. Conclusions
Chapter 5: Experimental results 64
Conversion dramatically drops during the first 20 minutes of reaction. This is due to the deactivating effect of coke formation. The presence of strong acid sites accelerates the protolysis reactions, whose activation energy is typically high. However, the catalyst activity decay becomes very slow at higher times on stream indicating that, after a period of strong deactivation, the remaining fraction of acids sites still active for the cracking of hydrocarbons is very slowly deactivated by coke.
Higher space time indicates a higher amount of catalyst active sites available for the reaction, and therefore a higher conversion. This is observed at every temperature level. An increase in the pressure results in an increased conversion for all the components, as a result of the higher adsorption of hydrocarbons on the catalyst surface that promote the hydride transfer reaction.
Conversion increases with increasing temperature. It is well known that cracking reactions are favored at higher reaction temperatures.
The main reaction products of cracking the mixture n-decane + methylcyclohexane + butylcyclohexane/1-octene are propylene, 1-butene, 2-butene, and alkanes such as propane, isobutene and isopentane. The decrease or increase in the selectivity to certain products with conversion is associated to secondary reactions, in particular hydride transfer and β-scission of primary products.
In contrast to what is mentioned in literature the experimental data show that an increase in the temperature results in an increase in coke formation. In this case this behaviour can be explained by the lower surface coverage at higher temperature, although the rate coefficient is increased.
Chapter 5: Experimental results 65
5.2 Experimental results of the recycle electrobalance reactor. Catalyc cracking experiments using iso-octane or methylcyclohexane as feed have been performed over a series of Y zeolites in order to study the effect of the catalyst acid properties of the cocking behaviour
Experimental data have been obtained for every Y-zeolite available, at a constant hydrocarbon partial pressure, varying space time and different temperature, as it is represented in table 5.3.
5.2.1. Influence of the zeolite The available experimental data allow to study the effect of acidity on methylcyclohexane and iso-octane cracking. The physical properties of the four Y zeolites investigated are represented in table 5.3. Catalyst Si/Al bulk Si/Al frame EFAL Active sites Structure (mmol
NH3/g) LZ-Y20 H-USY 2.6 30.0 Yes 0.99 FAU CBV 500 NH4-Y 2.6 3.9 Yes 1.50 FAU CBV 720 USY 15 16 Yes 0.60 FAU CBV 760 H-USY 30 100 Yes 0.23 FAU Table 5.3 Physical properties of the zeolites investigated.
Figures 5.20, 5.21 and 5.22 show the coke deposition during methylcyclohexane cracking as a function of space time. At all conditions investigated the coke deposition on the zeolites decreases in the order: CBV 500>CBV 720>LZY20>CBV 760. This indicates that coke deposition increases with decreasing Si/Al framework ratio. At equal times on stream (180 s, 1140 s and 2040 s) it is observed that the coke deposition increases with increasing space time and thus with conversion. It is observed too, that the deposition of coke on CBV 500 is higher than on other zeolites, which can be due to a higher number of active sites (1.5 mmol NH3/g). Chapter 5: Experimental results 66
Deposited coke vs W/F (180 s)
0.01
0.008 CBV500 0.006 CBV720 0.004 CBV760
coke/kg cat) 0.002 LZY20
Deposited coke (kg 0 0 50 100 150 W/F (kgs/m ol)
0 3 Figure 5.20 Deposited coke as a function of space time, t = 3 min, T = 748 K, p mchex = 7 10 Pa, Feed = methylcyclohexane.
Deposited coke vs W/F (1140 s)
0.05
0.04 ) CBV500 0.03 CBV720
0.02 CBV760 LZY20 coke/kg cat 0.01 Deposited coke(kg 0 0 50 100 150 W/F (kgs/mol)
0 3 3 Figure 5.21 Deposited coke as a function of space time, t = 19 min, T = 748 K, p mchex = 7 10 Pa, Feed = methylcyclohexane.
Deposited coke vs W/F (2040 s)
0.06 0.05 CBV500 0.04 CBV720 0.03 CBV760 0.02
coke/kg cat) 0.01 LZY20
Deposited coke (kg (kg coke Deposited 0 0 50 100 150 W/F (kgs/mol)
0 3 Figure 5.22 Deposited coke as a function of space time, t = 34 min, T = 748 K, p mchex = 7 10 , Feed = methylcyclohexane. Chapter 5: Experimental results 67
The same trend is observed in case of iso-octane cracking. The increased coke deposition with decreasing Si/Al framework ratio is represented in figure 5.23.
Deposited coke vs W/F (2040 s)
0.14 0.12 CBV500 0.1 CBV720 0.08 0.06 CBV760 0.04 LZY20 coke/kg cat) 0.02
Deposited coke (kg 0 0 50 100 150 200 W/F (kgs/mol)
0 3 Figure 5.23 Deposited coke as a function of space time, t = 34 min, T = 748 K, p mchex = 7 10 , Feed = iso- octane.
As shown in figure 5.24 and 5.25, conversion decreases with time and with the amount of deposited coke. It is also observed that in case of CBV 500 this decrease occurs faster and more pronounced than the other zeolites.
Conversion vs time
80 70 60 CBV500 50 CBV720 40 30 CBV760 20 LZY20 Conversion % Conversion 10 0 0 2000 4000 6000 Time (s)
3 Figure 5.24 Conversion as a function of time, T = 748 K, piC8 ~7 10 Pa, W/F ~70 Kg·s/mol, Feed = iso- octane
Chapter 5: Experimental results 68
Conversion vs deposited coke
80 70 60 CBV500 50 CBV720 40 CBV760 30 LZY20 20 Conversion % 10 0 0.00 0.03 0.06 0.09 0.12 0.15 Deposited coke (kg coke/kg cat)
3 Figure 5.25 Conversion as a function of deposited coke, T = 748 K, piC8 ~7 10 Pa, W/F ~70 Kg·s/mol, Feed = iso-octane.
5.2.2. Influence of the feed Figures 5.26-5.28 show the influence of the feed on the cocking behaviour of the zeolite. The same trend is observed for all the zeolites tested. The results obtained indicate that when the reactor is fed with iso-octane the conversion, the coking rate and the deposited coke are higher. This is due to higher reactivity of iso-octane compared to methylcyclohexane.
Conversion vs deposited coke
50.00 40.00 30.00 M_LZY20 20.00 O_LZY20 10.00 conversion % 0.00 0.00 0.01 0.01 0.02 Deposited coke (kg coke/kg cat)
3 Figure 5.26 Conversion as a function of deposited coke, T = 748 K, pp ~7 10 Pa, W/F ~63 Kg·s/mol.
Chapter 5: Experimental results 69
Deposited coke vs time
0.014 0.012 0.010 0.008 M_LZY20 0.006 O_LZY20 0.004 coke/kg cat) 0.002
Deposited coke (kg 0.000 0 2000 4000 6000 8000 Time (s)
3 Figure 5.27 Deposited coke as a function of time, T = 748 K, pp ~7 10 Pa, W/F ~63 Kg·s/mol.
Coking rate vs time 0.14 0.12 0.10 0.08 M_LZY20 0.06 O_LZY20 0.04
coking rate (kg 0.02 coke/min kg cat ) 0.00 0 2000 4000 6000 8000 time (s)
3 Figure 5.28 Coking rate as a function of time, T = 748 K, pp ~7 10 Pa, W/F ~63 Kg·s/mol.
5.2.3. Conclusions At all conditions investigated the coke deposition on the zeolites decreases in the order: CBV 500>CBV 720>LZY20>CBV 760. This indicates that coke deposition increases with decreasing Si/Al framework ratio. The deposition of coke on CBV 500 is higher than on other zeolites, which can be due to a higher number of active sites.
It is observed that the coke deposition increases with increasing space time and thus with conversion.
Conversion decreases with time and with the amount of deposited coke. It is also observed that in case of CBV 500 this decrease occurs faster and more pronounced than on the other zeolites. Chapter 5: Experimental results 70
The results obtained indicate that when the reactor is fed with iso-octane the conversion, the coking rate and the deposited coke are higher. This is due to higher reactivity of iso-octane compared to methylcyclohexane.
Chapter 6: Parameter estimation in presence of coke 71
Chapter 6 : Parameter estimation in presence of coke
6.1. Objective function The objective function for parameter estimation used is the minimization of the weighted sum of the squares of the residuals between calculated and experimental molar yields:
nrespn resp nexp (6.1) ˆ ˆ 2 ,...,, bbb p bRSS )( = ∑∑ jk ∑( ij )( yyyyw ikikij ) ⎯−− ⎯→⎯⎯ min j k i=1
Where: RSS (b) = The residual sum of squares to be minimized and b is the optimal parameters vector. nexp = Number of independent experiments. nresp = Number of responses. p = Number of parameters to be estimated.
yˆij = Calculated yield of response j in experiment i.
yij = Experimental yield of response j in experiment i. wjk = The weight factor of responses k and j. The weight factor is calculated with the empirical equation:
n −a (6.2) ⎡ resp ⎤ ⎢∑ yij ⎥ ⎣⎢ j =1 ⎦⎥ ww iii == −a nexp ⎡nresp ⎤ ∑∑⎢ yij ⎥ i=11⎣⎢ j = ⎦⎥
Where ‘a’ can take values between 0 and 1. When zero is used all the responses are equally weighted, and 1 is utilized when the weight factors express the real relative importance in the components.
Chapter 6: Parameter estimation in presence of coke 72
ODRPACK 2.01 solver (Boggs et al., 1992) was used to estimate the parameters that minimize the objective function.
FORTRAN 77 code was used for performing the calculations. For the estimation of the kinetic parameters involved in the model for coke formation based on initial coking rates, the following objective function was defined: n exp (6.3) ˆ 0 0 b bRSS )( ∑( , ,kckc )⎯ℜ−ℜ= ⎯→ min k =1 Where: ˆ 0 ℜ ,kc = Calculated average initial coking rate.
0 ℜ ,kc = Experimental average initial coking rate for a given experiment k. nexp = Number of experiments accounted for the regression. b = Vector of parameters to be determined.
ODRPACK 2.01 solver was also used for finding the parameters that minimize the coke function.
The average initial experimental coking rate can be easily obtained because the coke content on the bed is available at different times on stream. The calculated average initial coking rate is obtained via integration of the local coking rates along the catalyst bed: 1 W (6.4) ˆ 0 =ℜ 0dWr c ∫ c W 0
The equation 6.4 can be approximated using the trapezoid formula:
ˆ 0 1 0 0 0 0 (6.5) c =ℜ [ 1, cc 2, 2...2 nc −1, ++++ rrrr ,nc ] etep etep 2nstep
0 r ,ic is the calculated initial coking rate for the catalyst mass in a given element i, which depends on the partial pressure of the corresponding gas phase coke precursors. The gas phase composition along the catalyst bed is calculated via integration of the reactor continuity equations Chapter 6: Parameter estimation in presence of coke 73
(eq. 6.14) The composition in the gas phase of a given species is a function of the inlet conditions and its net production rate. The latter can be obtained also via integration of the corresponding reactor model equations (eq. 6.14).
6.2. Statistics The F-test is used to determine the global significance of the regression:
⎡ nexp nresp ⎤ (6.6) ˆ 2 ⎢∑∑ ()yw ijj ⎥ TSS ⎣ i==11j ⎦ p p Freg = = RSS nexp nresp yyw ˆ 2 − pnn ∑∑ ()− ijijj exp resp i==11j
( exp resp − pnn )
Where TSS = Regression sum of squares. RSS = Residual sum of squares.
The calculated F-value ( Freg ) is compared to the tabulated one for a probability level of 1-α, i.e.
Ftab(p, nexp nresp -p, 1-α). When the former is higher than the latter, the regression is considered as globally significant.
A t-test is used to asses the individual significance of the estimated model parameters. Each parameter is tested against a reference t value, bi , which is typically assumed zero: (6.7) − bb ii tcalc = σˆ()bi
Then tcalc is usually the ratio between the values of bi to its standard deviation σˆ(bi ). The latter correspond to the square root of the diagonal elements of the covariance matrix of the parameters ˆ denoted as bV )( ii : Chapter 6: Parameter estimation in presence of coke 74
−1 ⎡ ⎤ (6.8) ˆ T )( = ⎢∑ JJwbV iii ⎥ ⎣ ⎦
Where J is the Jacobian matrix:
⎡∂ i bF )( ⎤ (6.9) Ji = ⎢ ⎥ ⎣ ∂b ⎦
When tcalc is calculated, it is compared with a tabulated value obtained for n-p degrees of freedom and a probability level of 1-α, i.e., tab − pnt −α 2/1,( ) . This way a significant parameter always satisfies:
calc > tab − pntt −α )2/1,( (6.10) Where: α = 0.05 (chosen) p = number of responses in the objective function.
The confidence interval of the estimated parameters delimits the range in which the optimal value of βi is located within a selected probability level of 1-α:
tabi pntb −−− α )2/1,( ⋅σˆ )( ≤ β ≤ + tabiii − pntbb −α )2/1,( ⋅σˆ bi )( (6.11)
6.3. Experimental reactor model equations The continuity equation for the gas phase component j in an experiment i can be written as follows: 0 1 ⎛ ∂yF ⎞ ∂yˆˆ (6.12) alkc )( ⎜ ij ⎟ ij − ⎜ ⎟ + = Rij Ω c νρ ⎝ ∂t ⎠ ⎛W ⎞ ∂⎜ 0 ⎟ ⎝ FA ⎠
at t=0 yˆij = 0
at W yˆ = 0 0 = 0 ij F alkc )( Chapter 6: Parameter estimation in presence of coke 75
Where: W 0 = Space-time for the (cyclo)alkane. F alkc )(
Rij = Net production rate of product j in experiment i. For coke deposition the continuity equation is: ∂C (6.13) c = R ∂t c
at t=0 C c = 0
Where;
Rc = Net production rate of coke. t = Time on stream.
The continuity equations account for the deformation of different lumps in a continuous, pseudo- homogeneous, one-dimensional reactor model. This equation was derived in the absence of concentration and thermal gradients, while taking into account that plug flow and isobaric behavior are guaranteed.
In absence of coke formation the accumulation term of eq. 6.12 is removed and the set of partial differential equations is changed to a set of to ordinary differential equation (ODE’s).
ydˆij (6.14) = Rij ⎛W ⎞ d⎜ o ⎟ ⎝ F alkc )( ⎠ at W yˆ =0 ∀ t o = 0 ij F alkc )(
The LSODA integration routine is used for the integration of the continuity equations. Chapter 6: Parameter estimation in presence of coke 76
6.4. Estimation of activation energies For the estimation of the activation energies, experimental data obtained by cracking the mixture n-decane + methylcyclohexane + butylcyclohexane / 1-octene in a TEOM reactor were used. The estimation of activation energies of the various elementary reactions involved in the kinetics was performed via non-isothermal regression of these data.
During the non isothermal-regression, a reparameterized form of the Arrhenius equation was required:
~ ⎡ E ⎛ 11 ⎞⎤ (6.15) ⎜ ⎟ Ak rep exp⎢ ⎜ −−= ⎟⎥ ⎣⎢ ⎝ TTR m ⎠⎦⎥
~ ⎡ E ⎤ (6.16) = + ACA exp − rep ,Ht ⎢ ⎥ ⎣ RTm ⎦
Where: ~ k = single-event rate coefficient. E = Activation energy.
Arep = Preexponential factor.
Tm = Average temperature over the experiments. ~ A = Single-event preexponential factor.
C ,Ht + = Total concentration of acid sites on the catalyst.
6.5. Rate coefficients A full reaction network starting from individual compounds for all hydrocarbon families involves a huge number of reactions species and reactions. However, the reactions can be grouped in a limited number of reaction families (protonation, β-scission, protolytic scission, etc.). Within a reaction family the rate coefficients depend, as discussed earlier, on the symmetry of the reactant Chapter 6: Parameter estimation in presence of coke 77 and of the transition state. An overview of the reaction rate coefficients per hydrocarbon family is given in table 6.1 (Quintana-Solorzano, 2007). Elementary reaction per hydrocarbon family Number of parameters Alkanes Hydride transfer 2 Protonation 2 Deprotonation 3 Pcp isomerization 4 β-scission 4 Protolytic scission 3 Cycloalkanes Hydryde transfer 2 Hydryde transfer alkenes 1 Protonation 2 Deprotonation 3 Endocyclic β-scission 4 Endocyclic protolytic scission 3 Hydride transfer cyclic (di)alkenes and (poly)aromatic cyclic alkenes 1 Deprotonation cyclic olefinic carbenium ions 2 Deprotonation cyclic diolefinic carbenium ions and (poly)aromatic carbenium ions 1 Aromatics Hydride transfer 2 Protonation 2 Deprotonation 3 Dealkylation 2 Disproportionation 1 Cyclisation 1 Table 6.1 Overview of elementary representative reactions per hydrocarbon family (Quintana-Solorzano, 2007). Chapter 6: Parameter estimation in presence of coke 78
Ten different reaction types are involved in coke formation (table 6.2 and 6.3), the ten activation energies should be calculated.
Table 6.2 Summary of the rate-determining steps in coke formation at the molar level corresponding to the alkylation of alkylcarbenium ions and aromatic carbenium ions according to the extended network including coking (Quintana-Solorzano, 2007).
Table 6.3 Summary of the rate-determining steps in coke formation at the molar level corresponding to the nucleus alkylation of aromatics according to the extended network including coking (Quintana-Solorzano, 2007).
A first reduction in the number of parameters is possible considering that the side chain alkylation on aromatic carbenium ions is analogous to the alkylation of alkylcarbenium ions. As a result, only four rate coefficients are sufficient to describe the (side chain) alkylation, decreasing the number of kinetic parameters from ten to six. For the (side chain) alkylation the number of kinetic parameters can be further reduced. If it is assumed that the structure (and nature) of the formed alkylated species do not have any effect in the coking reactions as they are being considered as coke. Hence, two rate coefficients that only depend on the nature of the departing carbenium ion, i.e. secondary or tertiary, are sufficient (Quintana-Solorzano, 2007). These rate coefficients are denoted as:
~ ~ (side_chain) alkylation Æ k _ sidealk (s; m), k _ sidealk (t; m)
Chapter 6: Parameter estimation in presence of coke 79
For the nucleus alkylation of aromatics that occurs via electrophylic attack by an alkylcarbenium ion, two reaction types (rate coefficients) are involved depending on the nature the latter carbenium ion:
~ ~ Nucleus alkylation Æ k _ nuclalk (s), k _ nuclalk (t)
The final number of activation energies to be estimated via regression for the coking model amounts to four.
6.6. Results The parameter estimation was performed by adjusting the four parameters related with the formation of coke specified in section 6.5 and using the objective function eq. 6.3. Activation energies for the main cracking reactions were taken from Quintana-Solorzano, 2007. It is shown in tables 6.4 and 6.5. In all the estimations performed, the obtained parameters were not significantly different from zero, i.e., the estimated t-values were much lower than the tabulated ones (2.161). Moreover, the calculated F-value for the significance of the regression was 1.62 for the 95% confidence interval, being also lower than the tabulated value of 2.95.
Table 6.4 Activation energies in kJ(mol)-1 and 95% individual confidence limits for the various reaction types involved in the cracking of alk(e)nes. They have been estimated via non-isothermal regression of the n-decane/1-octene cracking data. (Quintana-Solorzano, 2007)
Chapter 6: Parameter estimation in presence of coke 80
Table 6.5 Activation energies in kJ(mol)-1 and 95% individual confidence limits for the various reaction types involved in the cracking of cyclic hydrocarbons. They have been estimated via non-isothermal regression of the methylcyclohexane/1-octene and n-butylcyclohexane/1-octene cracking data. (Quintana- Solorzano, 2007)
Figure 6.1 shows the parity diagrams of calculated molar yield as a function of experimental molar yield for the responses accounted for in the objective function classified by hydrocarbon type, i.e. alkanes, alkenes, cycloalka(e)nes, and aromatics. The agreement between calculated and measured values is not good.
Figure 6.2 shows the parity diagram of the calculated initial coking rate as a function of experimental initial coking rate. It is clear that the values calculated are not similar to the experimental ones. The values calculated by the program are very low.
These results indicate that a further study is required. Chapter 6: Parameter estimation in presence of coke 81
Alkanes Alkenes
0.8 0.7 0.7 0.6 0.6 0.5 0.5 0.4 0.4 0.3 0.3 0.2 0.2
y calc, mol/mol 0.1 y cal,mol/mol 0.1 0 0 0 0.2 0.4 0.6 0.8 0 0.2 0.4 0.6 0.8 y exper, mol/mol y exper, mol/mol
Aromatics Cycloalka(e)nes
0.25 0.6
0.2 0.5
0.4 0.15 0.3 0.1 0.2 0.05
y calc,mol/mol y calc, mol/mol 0.1
0 0 0 0.05 0.1 0.15 0.2 0.25 0 0.1 0.2 0.3 0.4 0.5 0.6 y exper, mol/mol y exper, mol/mol
Figure 6.1 Parity diagrams comparing calculated with experimental molar yields of the responses accounted for in the objective function for the estimation of activation energies.
Mixture
0.0001
0.00008
0.00006
0.00004
0.00002 (kgcoke/100kgcat/ks) Initial coking calc rate, 0 0 0.02 0.04 0.06 0.08 0.1 Initial coking rate, exp (kgcoke/100kgcat/ks)
Figure 6.2 Parity diagram comparing calculated and experimental initial coking rate.
Chapter 6: Parameter estimation in presence of coke 82
6.7. Conclusions A parameter estimation accounting for the formation of coke has been performed. In this case, only the kinetic parameters related to the formation of coke were adjusted. However, the estimated parameters resulted non-significant.
For a further study, a higher number of kinetic parameters should be adjusted instead of only the four parameters considered in this work. Chapter 7: Conclusions 83
Chapter 7 : Conclusions
SEMK modeling of the catalytic cracking of hydrocarbons plays an important role in the development of this work. This approach allows to deal with coke in terms of elementary reactions. The formation of coke and its precursors is entirely described in terms of elementary reactions, i.e., alkylations, cyclisations, hydride transfers and deprotonations, which constitute a subset of those occurring during the main cracking.
During the cracking of the mixture n-decane, methylcyclohexane, butylcyclohexane and 1-octene over LZY20 in the TEOM reactor the deactivating effect of coke on the conversion occurs in the first 20 minutes of the reaction. This dramatic conversion drop indicates that coke deactivation is selective to the strongest acid sites. The presence of strong acid sites accelerates the protolysis reactions, whose activation energy is typically high. Moreover, due to a higher carbenium ion lifetime adsorbed on strong acid sites than on weaker acid sites, propagation of the catalytic cracking reactions via hydride transfer is favored.
The conversion increases with temperature and space time. Higher space times indicate a higher amount of catalyst active sites available for reactions, and therefore, a higher conversion.
Butylcyclohexane is the component in the mixture with higher conversion. Alkenes and alkanes are the type of hydrocarbons formed in the highest proportions. Main cracking products are propylene, isobutane, isopentane, isobutene, trans-2-butene, cis-2-butene and 1-butene. The cracking of the mixture also yields the formation of aromatics and cycloalka(e)s, the latter being higher. Naphthalene and methylindanes are also observed but not in high proportions.
Alkenes are formed in highest proportion, this indicates that β-scission and deprotonation of carbenium ions are favored compared to protolytic scission and hydride transfer of carbenium ions.
Chapter 7: Conclusions 84
Coke formation increases with time on stream, which emphasizes the importance of secondary transformations of primary cracking products in coke formation. Total pressure has less influence on conversion but anyway it has a positive effect in coke formation.
The catalytic cracking of iso-octane and methylcyclohexane, has been performed in a recycle electrobalance reactor at different temperatures, partial pressures and space times, using different catalysts. A series of Y faujasites has been tested to study the influence of acid properties on iso- octane and methylcyclohexane conversion. The coke deposition on the zeolite increases with increasing framework Si/Al ratio of the zeolite, so it decreases in the order: CBV 500>CBV 720>LZY20>CBV 760 zeolite. It is also observed that when the reactor is fed with iso-octane the conversion, the coking rate and the deposited coke are higher compared to methylcyclohexane. This indicates that iso-octane is more reactive than methycyclohexane.
A parameter estimation accounting for the formation of coke has been performed. In this case, only the kinetic parameters related to the formation of coke were adjusted. However, the estimated parameters resulted non-significant. The kinetic parameters determined previously (Quintana-Solorzano, 2007 ) for the cracking of n-decane/1-octene, methylcyclohexane/1-octene and butylcyclohexane/1-octene are not suitable to describe the cracking of the mixture n-decane, methylcyclohexane, butylcyclohexane, 1-octene.
In the future, in the parameter estimation a higher number of kinetic parameters should be adjusted instead of only the four parameters considered in this work. It would be interesting to work with different zeolites and broader range of temperatures in the TEOM reactor to have a clear knowledge of the influence of the catalyst properties and the temperature on coke deposition.
For the recycle electrobalance reactor it is suggested to work with different partial pressures to know how this variable affects the coke deposition. Appendix A: Operation of the set up 85 APPENDIX A: Operation of the set up 1. Operational specifications of the TEOM 1500 PMA sensor unit
Mass sensitivity = 1 µg Time resolution = 0,1 s Pressure ranges: • Normal mode = vacuum to 106 Pa • High pressure mode = up to 6 106 Pa Maximum gas flow rate = 3,3 10-5 m3/s Temperature ranges: • Preheating zone = 323 – 773 K • Main zone = 323 – 873 K
Figure 1: Gas inlet and outlet flow directions + optics position + catalyst bed in the TEOM reactor 2. Procedure when loading and unloading the TEOM reactor (see Figure 2 + also see TEOM manual)
1. Unscrew two inlet lines (hydrocarbon feed + He) 2. Open box 3. Remove insulation 4. Unloose black clamp screw and glide clamp around sensor unit 5. Tighten black clamp screw 6. Open main nut with a nut spanner
Appendix A: Operation of the set up 86
Figure 2: Sensor unit containing the tapered element, heat parts, optics, inlet and outlet gas positions and thermocouples 7. Unloose black clamp screw and glide clamp into original position 8. Pull the red lever downwards 9. Lift sensor unit, turn it 90° and remove it carefully 10. Remove cap (check indication mark) 11. Remove quartz fibre and catalyst from the catalyst basket 12. Load new quartz fibre and catalyst powder Attach the cap 13. Renew sealing ring at the main nut + use lubricant 14. Place TEOM sensor unit and plug the electronic connections on the right side of the box 15. Pull the red lever upwards 16. Hand tighten main nut
Appendix A: Operation of the set up 87 17. Unloose black clamp screw and glide clamp around sensor unit 18. Tighten black clamp screw 19. Tighten main nut with a nut spanner 20. Unloose black clamp screw and glide clamp into original position 21. Connect inlet lines. Use new swagelok sealing rings 22. Before attaching the insulation again and increasing the temperature in the reactor, check whether the set-up is hermetic (leakages mainly arise at the main nut and at both back pressure regulators) 23. Attach insulation 24. Close the box
3. Checklist before performing an experiment
Check all gas cylinders. The minimal required pressures are: • Air to GC ………………………………….. 30 bar • He to GC ………………………………….. 30 bar • He to TEOM ………………………………. 30 bar • H2 ………………………………………….. 20 bar • Air to TEOM + BP2 ………………………. 20 bar • N2 (make-up + pressurize liquid N2 vessel) ..20 bar
Pressure regulators relaxate the cylinder pressure respectively into: • Air to GC ………………………………………... 4 bar • He to GC …………………………………………10 bar • He to TEOM …………………………………….. 5 bar • H2 …………………………………………...…… 5 bar • Air to TEOM + BP2 …………………………….. 6 bar • N2 (make-up + pressurize liquid N2 vessel) ……... 5 - 7 bar
Settings for the GC-FID: • Air ………………….. 40 psi (1 psi = 6,8948 103 Pa = 6,8948 10−2 bar) • H2 …………………... 15 psi • Carrier (He) ………… 60 psi • Column pressure …… 27 psi
Ignition of the flame of the GC-FID: • Completely open air regulator • Open H2 regulator until you hear a click • Light the air + H2 mixture at the detector with a lighter (electrical ignition of the GC-FID does not work properly) • When a stable flame has established (check signal A): ¾ Completely open the H2 regulator ¾ Completely open the make-up gas regulator
Appendix A: Operation of the set up 88 4. Procedure when performing an experiment 4.1 Stop the regeneration program of the catalyst
• Stop the decoking4 program in Labview • Close red valve (air for regeneration) on control panel 4.2 Set process conditions
In Labview: • Load Ugent.standby Ö exit Ö start in manual mode • Uncheck option to save file at gatetime intervals, otherwise too large files will be created • Check that the 4 WV is in off position • Input temperature Preheat • Input temperature Zone Control • Input setpoints of mass flow controllers MFC 2 and MFC 4
Input setpoint MFC 3 based on calibration curve.
Regulate pressures BP 1 and BP 2 (manometric pressure).
When pressures have stabilized: • In Labview: input setpoints heaters A, B and C • Push buttons A, B and C on the control panel to activate the respective heaters • Activate heater Reactor Outlet (typically 220 °C, depending on feed and pressure) • Activate heater Analysis Line (typically 220 °C) • Activate heater He Purge Line (typically 170 – 180 °C) • Activate heater Reactor Inlet (sample) (typically 220 °C)
Continuously control pressures in BP 1 and BP 2, and check whether these are equal.
DME heater: 1/3 between 0 and 2 A Ö ± 0,7 A.
Start pump when temperature and pressure have stabilized. The temperatures for heaters A, B and C may still be 30 °C lower than the setpoint when the pump is started.
Open water tap for Liebig cooler.
Start the chronometer and measure volumetric flow of the feed (use balance + feed density) and compare with setpoint.
Start or check the storage of the coke signal data in Labview.
Stabilization for 1 h 20 min (starting from the moment when the liquid feed has been pumped).
Prepare ATLAS files for the GC analysis of the reactor effluent samples (see ATLAS manual).
Appendix A: Operation of the set up 89
Remember to control the pressures in BP 1 and BP 2 frequently. From the moment that the hydrocarbon feed is sent to the reactor, BP 1 and BP 2 may not be adjusted anymore, since this would influence/falsify product distribution and coke signal.
5 minutes before the start of the catalytic cracking reactions (= before switching the 4 WV from off to on position), open the DME line: • Pressure regulator of the DME cylinder Ö 3,5 bar • Input setpoint mass flow controller (see calibration curve, typical value is 10,0)
Frequently measure/control volumetric flow of the pump.
4.3 Perform the GC-FID analyses
4.3.1 Online analysis: • Switch 6 WV to ON position in Labview • Wait 15 s • Simultaneously push start button on GC-FID and start button on integrator • Wait 5 s • Switch 6 WV to OFF position in Labview when this online analysis is not the last online analysis. When performing the last online analysis leave the 6 WV in ON position. The following offline analyses will all be performed while leaving the 6 WV in online (ON) position.
Appendix A: Operation of the set up 90 Figure 3: Examples of the 6 WV and 10 WV positions for respectively storing samples (6 WV OFF position) and injecting samples of the reactor effluent to the GC-FID (6 WV ON position)
4.3.2 Offline analysis: • The 6 WV is in ON position and the 10 WV should be in the loop previous to the one which contains the sample you would like to inject during the offline analysis • Switch the 10 WV to the loop you want to inject (normally the next loop) • Wait 10 s • Simultaneously push start button on GC-FID and start button on integrator
4.3.3 Storing a sample (for a later offline analysis): • The 6 WV is in offline position • The 10 WV “activates” the loop which will contain the sample • At the reaction time you want a sample stored, move the 10 WV to the next position and the sample will be stored in the previous loop
4.4 Stopping procedure
Close 4 WV.
Stop pumping the feed and turn off the balance.
Close DME line.
Close water tap for Liebig cooler.
Wait 45 min Ö purging of the lines.
Stop storing data in files in Labview, approximately 20-40 minutes after the 4 WV was closed and the feeding of hydrocarbons to the reactor was stopped.
Switch off all temperature regulators or decrease their setpoints, except for heater C (6 WV and 10 WV box) and the heater of the GC-FID inlet line.
Decrease the setpoints of the He flows.
Do not decrease the pressures in the back pressure regulators yet. This can only be done when the temperature in the 6 WV and 10 WV box has sufficiently decreased.
After the last GC-FID analysis: • Switch off heater C (6 WV and 10 WV box) and the heater of the GC-FID inlet line • Switch 6 WV in offline position
Appendix A: Operation of the set up 91 • Decrease pressures in BP 1 and BP 2 to 0 Pa from the moment the temperature in the 6 WV and 10 WV box has decreased to 35 °C
4.5 Regeneration of the (deactivated) catalyst
This procedure, which usually takes place at night, can only start after the last GC-FID analysis has been performed.
In Labview: • Close Ugent.standby (manual mode) Ö ON/OFF button • Run in program mode Ö load decoking4 • Exit • Open red valve (air for regeneration) on control panel
Remark: in the visual interface (VI) in Labview one cannot see the real-time process conditions of the decoking4 program.
5. Replacing the feed
Collect organic residues in the appropriate recipients.
Clean beakers and calibrated flasks with acetone + dry them with compressed air.
Fill beaker with the “new” feed.
Purge pump with this feed: • Open switch “open/close” • Push “purge” button (there is a default purge time, but one can also stop the purge process by simply clicking the purge button a second time). • Close switch “open/close”
Start purging the feed lines with the feed component by pumping it through the feed line for 5 minutes.
When starting a first experiment with this “new” feedstock, the feed line must be purged during 2 hours, instead of the usual 1 h 20 min.
6. Additional information
• Heater A (HT A) = regulates temperature in the evaporator • HT B = regulates temperature in the 4 way valve (4 WV) box
Appendix A: Operation of the set up 92 • HT C = regulates temperature in the 6 and 10 way valve (6 WV and 10 WV) box • Based on the flowsheet in Figure 4 one can deduce how the regulation of the pressures in the different lines by the two back pressure regulators depends on the position of the 4 WV
Appendix A: Operation of the set up 93
Figure 4: Flow diagram of the TEOM setup at the Laboratorium voor Petrochemische Techniek of the Ghent University
Appendix B: Experimental conditions of TEOM reactor 94
APPENDIX B: Experimental conditions of TEOM reactor
T, K 693 693 693 693 693 693 693
T, K 723 723 723 723 723 723 723
T, K 753 753 753 753 753 753 753
P operation, kPa 300 300 300 300 300 200 100
P atm, kPa 100 100 100 100 100 100 100
Total pressure, KPa 400 400 400 400 400 300 200
MW 1-octene, kg/mol 0.1122
Density 1-octene, kg/m3 715
\ MW n-decane, kg/mol 0.1423
Density n-decane, kg/m3 726.53
MW methylcyclohexane, kg/mol 0.0982
Density 1-methylcyclohexane, kg/m3 770.00
MW butylcyclohexane, kg/mol 0.1403
Density buthylcyclohexane, kg/m3 818.00
Appendix B: Experimental conditions of TEOM reactor 95
Is 1-octene going to be used? Yes or not YES YES YES YES YES YES YES
Molar ratio 1-octene/(n-decane+mchex=buchex) 0.1765 0.1765 0.1765 0.1765 0.1765 0.1765 0.1765
Molar ratio n-decane/(n-decane+mchex=buchex) 0.3333 0.3333 0.3333 0.3333 0.3333 0.3333 0.3333
Molar ratio mchex/(n-decane+mchex=buchex) 0.3333 0.3333 0.3333 0.3333 0.3333 0.3333 0.3333
Molar ratio buchex/(n-decane+mchex=buchex) 0.3333 0.3333 0.3333 0.3333 0.3333 0.3333 0.3333
mol 1-octene/mol mixture 0.1500 0.1500 0.1500 0.1500 0.1500 0.1500 0.1500
mol n-decane/mol mixture 0.2833 0.2833 0.2833 0.2833 0.2833 0.2833 0.2833
mol mchex/mol mixture 0.2833 0.2833 0.2833 0.2833 0.2833 0.2833 0.2833
mol 1-buchex/mol mixture 0.2833 0.2833 0.2833 0.2833 0.2833 0.2833 0.2833
Necessary for proper vaporizat. dilution(mol mixture/mol 0.050 0.050 0.050 0.050 0.050 0.050 0.050 He) 0.1247 0.1247 0.1247 0.1247 0.1247 0.1247 0.1247 Mixture MW, kg/mol 761.60 761.60 761.60 761.60 761.60 761.60 761.60 Mixture density, kg/m3 5.0 10.0 15.0 20.0 25.0 10.0 10.0 W/F, kgcat s/mol mixture 5.9 11.8 17.6 23.5 23.5 11.8 11.8 W/F, kgcat s/mol (nC10+mchex+buchex) 3.00E-05 3.00E-05 3.00E-05 3.00E-05 3.00E-05 3.00E-05 3.00E-05 W, kgcat
Mol flow of He, mol/s 1.20E-04 6.00E-05 4.00E-05 3.00E-05 2.40E-05 6.00E-05 6.00E-05
MFC-3 and MFC-4 ……… . flow of Helium, ml/min 161.3 80.6 53.8 40.3 32.3 80.6 80.6
Appendix B: Experimental conditions of TEOM reactor 96
MFC-2 ……………………………… Helium as a purge, 70 70 70 70.0 70.0 70.0 70.0 ml/min
Molar flow of reactant mixture, mol/s 6.00E-06 3.00E-06 2.00E-06 1.50E-06 1.20E-06 3.00E-06 3.00E-06
Molar flow of reactant mixture, mol/h 0.0216 0.0108 0.0072 0.0054 0.00432 0.0108 0.0108
Mass flow of mixture, kg/s 7.48E-07 3.74E-07 2.49E-07 1.87E-07 1.50E-07 3.74E-07 3.74E-07
Mass flow of mixture, kg/h 2.69E-03 1.35E-03 8.98E-04 6.73E-04 5.39E-04 1.35E-03 1.35E-03
Vol flow of mixture, m3/s 9.82E-10 4.91E-10 3.27E-10 2.46e-10 1.96E-10 4.91E-10 4.91E-10
Value to be used in the liquid HC pump . Vol flow of mixture, ml/min 0.059 0.029 0.020 0.015 0.012 0.029 0.029
Appendix C: Experimental conditions of the recycle electrobalance reactor 97 APPENDIX C: Experimental conditions of the recycle electrobalance reactor
1. Experimental conditions for methylcyclohexane In table I.1, an overview is given of the conditions applied in the experiments with methylcyclohexane.
pºmch pt F°mch W/Fºmch Conversion Experiment T (ºC) (103 Pa) (103 Pa) (10-6 mol/s) (kg.s/mol) (mol%)
M_LZY20_03 475 6,85 101,88 0,3 63,756 53,571
M_LZY20_04 475 6,88 100,99 0,3 63,756 51,098
M_LZY20_05 475 7,18 102,5 0,36 71,64 52,701
M_LZY20_06 475 7,130 102,37 0,31 81,54 51,903
M_LZY20_07 475 6,530 102,72 0,28 56,412 46,105
M_LZY20_08 475 6,81 102,33 0,3 42,984 47,608
M_LZY20_09 475 7,36 102,89 0,35 27,396 45,133
M_LZY20_10 475 7,18 103,2 0,31 30,744 49,249
M_LZY20_11 475 7,11 101,75 0,34 14,976 33,057
M_CBV720_01 475 7,16 99,92 0,32 30,204 42,489
M_CBV720_02 475 7,34 100,9 0,33 38,844 48,786
M_CBV720_03 475 6,84 100,96 0,3 53,460 51,825
M_CBV720_04 475 7,19 100,59 0,32 60,552 52,685
M_CBV720_05 475 6,86 100,85 0,3 63,864 47,109
M_CBV720_06 475 7,02 101,51 0,31 83,700 62,235
M_CBV760_03 475 7,04 101,77 0,32 79,452 41,952
M_CBV760_04 475 6,97 101,65 0,3 56,608 32,180
M_CBV760_06 475 6,85 101 0,304 37,870 30,612
Appendix C: Experimental conditions of the recycle electrobalance reactor 98 M_CBV760_07 475 6,99 102,34 0,31 42,696 39,891
M_CBV760_08 475 6,99 102,61 0,28 114,340 43,178
M_CBV500_01 475 6,96 102,26 0,28 91,476 56,527
M_CBV500_02 475 6,91 102,14 0,26 98,350 67,463
M_CBV500_03 475 7,05 101,94 0,3 73,692 67,469
M_CBV500_04 475 6,91 101,48 0,3 64,188 67,603
M_CBV500_05 475 6,92 101,73 0,3 49,248 72,316
M_CBV500_06 475 6,7 102,07 0,289 33,370 51,179
2. Experimental conditions for iso-octane
Appendix C: Experimental conditions of the recycle electrobalance reactor 99
Appendix C: Experimental conditions of the recycle electrobalance reactor 100
Appendix C: Experimental conditions of the recycle electrobalance reactor 101
Appendix D: Calibration factors for the TEOM reactor 102 APPENDIX D: Calibration factors for the TEOM reactor
numb C Compound name Calib Factor Mol Weight atoms ID CF MW nc metane 1.070 16.04 1 etylene 1.180 28.05 2 ethane 0.950 30.07 2 propylene 1.030 42.08 3 propane 0.960 44.09 3 DME 2.550 46.03 2 isobutane 1.030 58.12 4 isobutene 1.070 56.10 4 1-butene 0.940 56.10 4 1,3-butadiene 1.000 54.00 4 n-butane 1.000 58.12 4 trans-2-butene 1.020 56.11 4 cis-2-butene 0.980 56.11 4 3-methyl-1-butene 1.094 70.13 5 isopentane 1.007 72.15 5 1-pentene 1.136 70.13 5 2-methyl-1-butene 1.107 70.13 5 n-pentane 1.048 72.15 5 trans-2-pentene 1.049 70.13 5 cis-2-pentene 1.049 70.13 5 2-methyl-2-butene 1.035 70.13 5 2,2-dimethylbutane 1.094 86.17 6 cyclopentene 1.161 68.00 5 4-methyl-1-pentene/3-methyl-1-pentene 1.096 84.16 6 cyclopentane 1.156 70.13 5 2,3-dimethyl-butane 0.989 86.17 6 cis-4-methyl-2-pentene 1.036 84.16 6 2-methylpentane 1.023 86.17 6 trans-4-methyl-2-pentene 1.036 84.16 6 3-methylpentane 1.023 86.17 6 1-hexene+2-methyl-1-pentene 1.131 84.16 6 n-hexane 1.058 86.17 6 t-&-c-hexene-3 1.059 84.16 6 t-2-hexene 1.059 84.16 6 2-methyl-pentene-2 1.047 84.16 6 cis-3-methyl-2-pentene 1.047 84.16 6 3-methyl-cyclopentene 1.116 80.00 6 cis-2-hexene 1.059 84.16 6 trans-3-methyl-pentene-2 1.047 84.16 6 methylcyclopentane 1.112 84.16 6
Appendix D: Calibration factors for the TEOM reactor 103 2,4-dimethylpentane 1.006 100.00 7 3,4-dimethyl-pentene-1/4,4-dime-c- penten 1.070 98.19 7 1-methylcyclopentene 1.114 82.16 6 benzene 1.082 78.10 6 3-methyl-1-hexene 1.098 98.19 7 2-methyl-3-cis-hexene 1.056 98.19 7 3,3-dimethylpentane 1.096 100.23 7 cyclohexane 0.875 84.18 6 2-methyl-t-3-hexene 1.035 98.19 7 4-methyl-1-hexene 1.098 98.19 7 4-methyl-t/c-2-hexene 1.088 98.19 7 2-methylhexane 1.035 98.00 7 2,3-dimethylpentane 1.006 98.19 7 cyclohexene 1.045 82.14 6 3-methylhexane 1.035 98.19 7 3,4-dimethyl-c-2-pentene 1.025 98.19 7 1,3-cis-dimethylcyclopentane 1.081 98.19 7 1,3-trans-dimetylcyclopentane 1.081 98.19 7 1,2-trans-dimethylcyclopentane 1.081 98.19 7 2,2,4-trimethylpentane 1.065 98.19 7 trans-2-heptene 1.065 98.19 7 n-heptane 1.083 96.19 7 cis-2-heptene 1.065 98.19 7 1,2-dimethylcyclopentane 1.083 96.19 7 methylcyclohexane 1.021 98.19 7 ethyl-cyclopentane 1.111 98.19 7 2,4-dimethylhexane 1.083 96.19 7 1,2,4-trimethylcyclopentane 1.060 112.21 8 3,3-dimethylhexane 1.097 114.23 8 2,3,4-trimethylpentene 0.992 114.23 8 toluene 1.048 92.14 7 2,3,3-trimethylpentane 1.103 96.19 7 1-methylcyclohexene 1.021 96.19 7 3,4-dimethylhexane 1.021 96.19 7 1c,2t,3-trimethylcyclopentane 1.058 112.24 8 1t,4-dimethylcyclohexane 1.006 112.24 8 3t-ethylmethylcyclopentane 1.084 112.24 8 2t-ethylmethylcyclopentane 1.084 112.24 8 1,1-methylethylcyclopentane 1.165 112.24 8 1c,4-dimethylcyclohexane 1.006 112.24 8 1-octene 1.125 112.24 8 i-propylcyclopentane 1.084 112.24 8 1c,2-dimethylcyclohexane 1.006 112.24 8 1c,3c,5-trimethylcyclohexane 0.995 126.27 9 ethylbenzene 1.055 106.18 8 1c,2t,2t-trimethylcyclohexane 1.067 126.27 9 p-xylene/m-xylene 1.022 106.18 8 3,4-dimethylheptane 1.027 128.29 9 o-xylene 1.022 106.18 8 i-butylcyclopentane 1.089 126.27 9 1-methyl-1-ethylcyclohexane 0.964 126.27 9
Appendix D: Calibration factors for the TEOM reactor 104 1-methyl-3-ethylbenzene 1.032 120.21 9 1-methyl-4-ethylbenzene 1.032 120.21 9 1,3,5-trimethylbenzene 1.003 104.15 9 1-methyl-2-ethylbenzene 1.032 120.21 9 1,2,4-trimethylbenzene 1.003 104.15 9 i-butylbenzene 1.044 134.24 10 n-decane 1.010 142.15 10 1,2,3-trimethylbenzene 1.003 104.15 9 n-butylcyclohexane 1.049 140.30 10 1-methyl-3-n-propylbenzene 1.040 134.24 10 1-methyl-4-n-propylbenzene 1.040 134.24 10 1,4-dimethyl-2-ethylbenzene 1.014 134.24 10 1,3-dimethyl-4-ethylbenzene 1.014 134.24 10 1,2-dimethyl-4-ethylbenzene 1.014 134.24 10 1,3-dimethyl-2-ethylbenzene 1.014 134.24 10 1,2-dimethyl-3-ethylbenzene 1.014 134.24 10 1,2,3,4-tetra-methylbenzene 0.988 134.24 10 1,2,3,5-tetra-methylbenzene 0.988 134.24 10 5-methyl-indane 1.089 132.24 10 4-methyl-indane 1.089 132.24 10 2-methyl-indane 1.089 132.24 10 1-methyl-indane 1.089 132.24 10 naphthalene 1.061 128.2 10
References 105
References
Baltanas M. A., Van Raemconck K. K., Froment G. F., Mohedas S. R. (1989), Ind. Eng. Chem. Res., 28 (1989) 899. Bartholomew CH (2001) Mechanisms of catalyst deactivation. Applied Catalysis A: General 212: 17-60.
Beirnaert H., Alleman J., Marin G.B. (2001), Ind. Eng. Chem. Res., 40 (2001) 1337.
Beirnaert H.C., Vermeulen R., Froment G.F. A recycle electrobalance reactor for the study of catalyst deactivation by coke. Studies in Surface Science and catalysis, Vol. 88, 1994 Elsevier Science B.V.
Boggs P.T., Byrd R.H., Rogers J.E., Schnabel R.B., ODRPACK v.2.01 User’s reference guide (1992). Http://www.netlib.com Bayraktar O., Kugler E. L. Characterization of coke on equilibrium fluid catalytic cracking catalyst by temperatured-programmed oxidation. Appl. Catal., A 2002, 233, 197.
Bollas G. M., Vasalos I.A., Lappas A.A., Iatridis D.K., Tsioni G.K. (2004). Bulk molecular characterization approach for the simulation of FCC feedstocks. Ind. Eng. Chem. 2004, 43, 3270-3281. Breck D.W., U.S. Patent 3,130,007 (1964)
Butt JB & Petersen EE (1988) Activation, deactivation and poisoning of catalysts. Academic Press, Inc., New York.
Chen, J., Heck, R.M., Farrauto, R.J., (1992). Deactivation regeneration and poison – resistant catalysts: commercial experience in stationary pollution abatement, Catal. Today11: 517-545.
Chen D., Grønvold A., Rebo H. P., Moljord K. and Holmen A. Catalyst deactivaton stuied by conventional and oscillating microbalance reactors. Appl. Cat. A, 137, L1-8 (1996).
Corma A., Mocholi M., Orchilles V. Appl. Catal., 7 (1991) 307.
Dewachtere N. V., Santaella F., Froment G. F. Chem. Eng.Sci., 54 (1999) 3653.
Diaz P. (2007). Single-event microkinetis o (cyclo)alkane cracking on acid catalysts. Master thesis, Universiteit Gent.
References 106 Espinat D., Freund E., Dexpert H., Martino G. Localization and structure of carbonaceous deposits on reforming catalysts. J. Catal. 1990, 126, 496.
Espinat D., Freund E., Dexpert H., Martino G., Couzi M., Lespade P. Cruege F. Characterization of the coke formed on reforming catalysts by laser Raman spectroscopy. Appl. Catal. 1985, 13, 343.
Feng W., Vynckier E., Froment G. F., Ind. Eng. Chem. Res., 32 (1993) 3997
Forzatti P & Lietti L (1999) Catalyst deactivation. Catalysis Today 52: 165-181.
Froment G. F. (1997). Stud. Surf. Sci. Catal. 111 (1997) 53.
Froment G. F., in catalyst deactivation, Bartholomew C.H. and Fuentes G.A. (Eds.), vol.53, Elsevier science, Amsterdam, 1997, p.53.
Gambicki S., Future Directions in Oil Refining, Stud. Surf. Sci. Catal. 130 (2000) 147.
Guisnet, M.; Magnoux, P. Appl. Catal. 212 (2001) 83.
Guisnet M., Magnoux P., Martin D., in Catalyst deactivation, C.H. Bartholomew and G. A. Fuentes (Eds.), Vol. 53, Elsevier Science, Amsterdam, 1997, p.1.
Guisnet, M.; Magnoux, P. Influence of the pore structure. Appl. Catal. 1989, 54, 1.
Guisnet M., Magnoux P., Martin D. (1997). Roles of acidity and pore structure in the deactivation of zeolites by carbonaceous deposits. In catalyst deactivation 1997 , Stud. Surf. Sci. Catal. 111, New York 1997 p 1.
Harding R.H., Peters A.W., Nee J.R.D. (2001). New developments in FCC catalyst technology. Appl. Catal. A: Gen. 221 (2001) 389-396
Hershkowitz F., Madiara P. D. Simultaneous measurement of adsorption, reaction, and coke using a pulsed microbalance reactor. Ind. Chem. Res., 32, p2969-2974 (1993).
In-Su Han, Riggs J.B., Chung C. B., Chem. Eng. Proc., 43 (2004) 1063.
Ito T., Bonardet J. L., Fraissard J. Nagy J. B. Andre C. Gabelica Z. Derouane E. G. About coke deposition on zeolite HY. A 129Xe-NMR study. Appl. Catal. 1988, 43, L5.
Karge H. G. Coke formation on zeolites. Stud. Surf. Sci. Catal. 1991, 58, 531.
References 107 Karge H. G. , Lange J. P., Gutsze A., Laniecki M. Coke formation through the reaction of olefins over hydrogen mordenite. II. In-situ EPR measurements under on-stream contitions. J. Catal. 1988, 114, 114.
Koon, C. L.; Akbar, F.; Hughes, R.; Tyagi, Y. R.; Castro Diaz, M.; Martin, S. C.; Hall, P. J.; Snape, C. E. Development of an experimental protocol to evaluate FCC stripper performance in terms of coke yield and composition. Trans. Inst. Chem. Eng. 2000, 78, 738.
Lerner, B. A., Himpsl, F. L. (1997). Feed molecular properties explain cracking results. Oil Gas J. 1997, 95 (48), 61.
Li C., Stair P. C. Ultraviolet Raman spectroscopy characterization of coke formation in zeolites. Catal. Today 1997, 33, 353.
Liu DR & Park J-S (1993) Electron microprobe characterization of phosphorus containing deposits on used automotive catalyst surfaces. Applied Catalysis B: Environmental 2: 49-70.
Marcilly C. (2002). Present status and future trends in Catalysis for refining and petrochemicals. J. Catal., 216 (2003) 47-62
Martinis J.M., Foment G. F. (2006). Alkylation on soilid acids. Part2. Single-event kinetic modeling. Ing. Chem. Res. 2006, 45, 954-967.
Maya-Yescas R., Leon-Becerril E., Salazar-Sotelo D. (2004). Translation of MAT Kinetic data to model industrial catalytic cracking units. Chem. Eng. Technol. 2004, 27, No.7
Milhaccea E., Pop G., Bozga G., Muntean O., unpublished results, 1994. Moulijn, J. A., van Diepen, A. E., Kapteijn, F., (2001). Catalyst deactivation: is it predictable?What to do?, Appl. Cat. A Gen. 212; 3-16
Moustafa T. M., Froment G. F. Kinetic modeling of coke formation and deactivation in the catalytic cracking of vacuum gas oil. Det. Chem. Eng. Texas 77843-3122. (2002).
Plank C. J., Nace D. M. (1955). Coke formation and its relation ship to cumene cracking. Ind. Eng. Chem. 1955, 47 (11), 2374.
Paweewan B., Barrie P. J., Gladden L. F. (1997). Coking during ethane conversion on ultrastable zeolite Y. Applied catalysis A: General (1998) 353-362.
Petersen, E. E., Bell, A. T., (1987). Catalyst Deactivation, New York, Marcel Dekker, Inc.
Prahan A. R., Wu J. F. Jong S. J. Chen W. H. Tsai T. C. Liu S. B. Influences of zeolite structure on formation and location of coke: A 129Xe and 13C CP-MAS NMR study. Appl. Catal., A 1997, 159, 187.
References 108
Querini C.A., Fung S. C. Catal. Today, 37 (1997) 277.
Quintana-Solorzano R. (2006-2007). Single-event of microKinetics for coking in catalytic cracking: development and application. Phd.Thesis at LPT, UGent (2007).
Quintana-Solorzano R., Thybaut G. B., Marin G. B., Lødeng R., Holmen A., Catal. Today, 107-108 (2005) 619.
Quintana-Solorzano R., Thybaut G. B., Marin G. B., Lødeng R., Holmen A., Proc. 13th Int. Cong. Catal., Paris France, 11-16 July 2004, paper P2-214.
Reyniers M-F, Beirnaert H., Marin G.B. (2000). Influence of coke formation on the conversion of hydrocarbons I. Alkanes on a USY-zeolite. Applied catalysis A: General 202 (2000) 49-63.
Snape C. E., McGhee B. J., Martin S. C., Andresen J. M. Structural characterization of catalytic coke by solid-state 13C-NMR spectroscopy. Catal. Today 1997, 37, 285.
Scherzer J., Octane Enchancing Zeolitic FCC Catalysts, Marcel Dekker, New York, 1990.
Service W. J. (1960). Catalytic cracking feedstocks. Pet Chem. Eng. 1960, 32, C31.
Wang G., Gao J., Wu C. Evolutionary desing on FCC reactors driven by the high temperature and short contact time demands. Petroleum science and technology. Vol. 22, Nos. 11 & 12, pp. 1581-1594, 2004.
Wilson, J. W. Fluid Catalytic Cracking Technology andOperation; PennWell Publishing Co.: Tulsa, Oklahoma, U.S.A., 1997. www.britannica.com www.engineering.purdue.edu www.iza-structure.org www.rpco.com
Zhu W., van de Graaf J. M.,.van den Broeke L. J. P, F. Kapteijn F., and Moulijn J. A. TEOM: A Unique technique for measuring adsorption properties. Light Alkanes in Silicalite-1. Ind. Eng. Chem. Res. 1998, 37, 1934-1942