MECHANISTIC KINETIC MODELING OF THE

HYDROCRACKING OF COMPLEX FEEDSTOCKS

A Dissertation
by
HANS KUMAR

Submitted to the Office of Graduate Studies of

Texas A&M University
in partial fulfillment of the requirements for the degree of
DOCTOR OF PHILOSOPHY

December 2006

Major Subject: Chemical Engineering

MECHANISTIC KINETIC MODELING OF THE

HYDROCRACKING OF COMPLEX FEEDSTOCKS

A Dissertation
by
HANS KUMAR

Submitted to the Office of Graduate Studies of

Texas A&M University
in partial fulfillment of the requirements for the degree of
DOCTOR OF PHILOSOPHY

Approved by:
Co-Chairs of Committee,
Committee Members,
Head of Department,

Gilbert F. Froment
Rayford G. Anthony
Perla B. Balbuena
Abraham Clearfield
N. K. Anand

December 2006
Major Subject: Chemical Engineering

iii

ABSTRACT
Mechanistic Kinetic Modeling of the Hydrocracking of Complex Feedstocks.
(December 2006)
Hans Kumar, B.E., Indian Institute of Technology, Roorkee;
M.S., Texas A&M University, College Station.
Co-Chairs of Advisory Committee: Dr. Gilbert F. Froment
Dr. Rayford G. Anthony

Two separate mechanistic kinetic models have been developed for the hydrocracking of
complex feedstocks. The first model is targeted for the hydrocracking of vacuum gas oil.
The second one addresses specifically the hydrocracking of long-chain paraffins, but at a
more fundamental level as compared to the first one. Both models are based on an
exhaustive computer generated reaction network of elementary steps.
In the first model, the dehydrogenation/hydrogenation steps occurring on the metal sites
to generate/consume the reactive olefinic intermediates are assumed to be very fast so
that the acid site steps are considered as the rate determining steps. The frequency
factors for acid site steps are modeled using the single-event concept and the activation
energies based on the nature of the reactant and product carbenium ions.
This model utilizes a detailed composition of the vacuum gas oil characterized by 16
different molecular classes up to carbon number 40. These classes are divided into 45
subclasses by distinguishing the isomers of a class according to the number of methyl
branches. The kinetic model is plugged into an adiabatic multi-bed trickle flow reactor
model. The model contains 33 feedstock and temperature independent parameters which
have been estimated from the experimental data.
The model has been used to study the effect of the operating conditions on the yield and
composition of various products. A sensitivity analysis of the distribution of isomers of a

iv

class among its different subclasses has been performed showing that the total
conversion increases when the content of isomers with a higher degree of branching is
increased in the feed.
In the second model, the dehydrogenation/hydrogenation steps on the metal sites are also
assumed to be rate determining. The rate coefficients for the dehydrogenation steps are
modeled depending on the nature of the carbon atoms forming the double bond. The
frequency factors for the acid site steps are modeled using the single-event concept. A
more rigorous approach has been selected to model the activation energies of the acid
site steps by implementing the Evans-Polanyi relationship. The 14 model parameters,
which are independent of the temperature and feedstock composition, have been
estimated from the experimental data. The model elucidates the effect of the relative
metal/acid activity of the catalyst on the isomerization/cracking selectivities and on the
carbon number distribution of the products.

DEDICATION

To my parents

vi

ACKNOWLEDGMENTS
I would like to express my sincere gratitude to Dr. Gilbert F. Froment and Dr. Rayford
G. Anthony for serving as the co-chairs of my committee and for their continued
guidance and support in this research. The scientific discussions with Dr. Froment have
always been very insightful and I will always be indebted to him for all the knowledge
he shared with me. His prompt responses to all my email queries are truly appreciated.
Dr. Anthony always assisted me with all the technical and non-technical issues during
this research. His encouragement and interest led this project to be completed
successfully in a timely manner. This research would not have been completed without
the generosity of Dr. Anthony providing financial support from the C. D. Holland
Professorship for this project. I am truly thankful to my advisors for the trust they
showed in my capabilities to complete this research.
I would also like to thank Dr. Perla Balbuena and Dr. Abraham Clearfield for being on
my committee and for their time, efforts and ideas they contributed during our meetings.
I am also grateful to Dr. R. E. Galiasso for various stimulating discussions and
knowledge.
My special thanks go to my friend and mentor, Dr. J. Govindhakannan. I had a
wonderful time with him discussing different aspects of research and life. His previous
contribution in this field is highly appreciated.
I am also grateful to the faculty of the Department of Chemical Engineering for being
extremely cooperative and friendly at all times during my stay at Texas A&M
University. I would especially like to thank Dr. K. R. Hall, Dr. D. M. Ford, Dr. D. F.
Shantz and Dr. M. El- Halwagi. The help that I received from our wonderful staff,
especially Towanna Hubacek, Missy Newton, Valerie Green, Ninette Portales, Barbara
Prout and Jeff Polasek, is simply unforgettable.
Many thanks to my friends and my group members for making my research at Texas
A&M University a pleasant and exciting experience. I am fortunate to have good friends

vii

like Arnab, Faisal, Amit, Vipin, Srini, Ashwini, Manish, Sanjay, Jyoti, Nishant,
Hemendra, Greg and Nitin who were always there for me in good and bad times.
Finally, I would like to express my warmest regards to my parents and family members
for their unconditional love and support without which I could have never completed this
dissertation.

viii

TABLE OF CONTENTS
Page
ABSTRACT ..................................................................................................................... iii
DEDICATION ...................................................................................................................v
ACKNOWLEDGMENTS.................................................................................................vi
TABLE OF CONTENTS ............................................................................................... viii
LIST OF TABLES ............................................................................................................xi
LIST OF FIGURES......................................................................................................... xii
CHAPTER
I

INTRODUCTION ............................................................................................1
1.1. Hydrocracking Process Description ...........................................................3
1.2. Literature Survey........................................................................................7

II HYDROCRACKING CHEMISTRY AND REACTION

aaaaaaaaaaNETWORK GENERATION .........................................................................10
2.1. Hydrocracking Catalyst............................................................................10
2.2. Reaction Mechanism ................................................................................10
2.3. Reaction Network Generation ..................................................................17
2.4. Composition of VGO ...............................................................................20
III KINETIC MODEL DEVELOPMENT FOR THE HYDROCRACKING
aaaaaaaaaiOF VGO ..........................................................................................................30
3.1. Modeling of the Frequency Factors .........................................................31
3.2. Modeling of the Activation Energies .......................................................34
3.3. Development of Rate Equations...............................................................36
3.3.1. Group (a)- PCP, Acyclic/Exo/Endo -Scission .............................38
3.3.2. Group (b)-Cyclization ....................................................................41
3.3.3. Group (c)- Dealkylation of Aromatics ...........................................42
3.3.4. Saturation of Aromatics .................................................................43
3.3.4.1. Saturation of Monoaromatics ............................................43
3.3.4.2. Saturation of Polyaromatics ..............................................48
3.4. Competitive Chemisorption on the Metal Sites .......................................51
3.5. Development of Global Rate Expressions for Acid Site Steps ................52
3.6. Elimination of the Need of Reference Olefins .........................................59
3.6.1. Lumping Coefficients with Reference Olefins (Case 1) ................60
3.6.2. Lumping Coefficients without Reference Olefins (Case 2)...........61
3.6.3. True Lumping Coefficients (Case 3)..............................................61
3.7. Competitive Chemisorption on the Acidic Sites ......................................63

ix

CHAPTER

Page

3.7.1. Dehydrogenation Equilibrium Coefficient.....................................64

3.7.2. Heats of Protonation.......................................................................65
3.7.3. Heats of Stabilization .....................................................................68
3.7.4. Entropy of Protonation...................................................................69
3.8. Parameters for Sorption............................................................................69
IV REACTOR MODEL AND PARAMETER ESTIMATION ..........................71
4.1. Reactor Model ..........................................................................................71
4.2. Estimation of Properties ...........................................................................74
4.3. Parameter Estimation ...............................................................................74
V

REACTOR SIMULATIONS ..........................................................................80

5.1. Effect of Inlet Temperature ......................................................................83

5.1.1. Effect of Temperature on the Yields of Various Commercial
IIIIIIIIIIIIIIIIIII Fractions.........................................................................................85
5.1.2. Effect of Temperature on the Yields of Various Hydrocarbon
IIIIIIIIIIIIIIIIIIIi Classes ...........................................................................................92
5.1.3. Effect of Temperature on the Yields of Hydrocarbons of
IIIIIIIIIIIIIIIIIIIIDifferent Chain Length...................................................................94
5.1.4. Effect of Temperature on the Yields of Individual Classes of
IIIIIIIIIIIIIIIIIIIIVarious Commercial Fractions.......................................................94
5.1.5. Effect of Temperature on the Iso/Normal Paraffin Ratio...............96
5.1.6. Effect of Temperature on the Carbon Number Distribution of
IIIIIIIIIIIIIIIIIII Hydrocarbon Classes .....................................................................97
5.2. Effect of Pressure .....................................................................................97
5.2.1. Effect of Pressure on the Yields of Various Commercial
IIIIIIIIIIIIIIIIIII Fractions.........................................................................................99
5.2.2. Effect of Pressure on the Yields of Various Hydrocarbon
IIIIIIIIIIIIIIIIIII Classes ...........................................................................................99
5.2.3. Effect of Pressure on the Yields of Hydrocarbons of Different
IIIIIIIIIIIIIIIIIII Chain Length................................................................................105
5.2.4. Effect of Pressure on the Yields of Individual Classes of
IIIIIIIIIIIIIIIIIII Various Commercial Fractions ....................................................105
5.2.5. Effect of Pressure on the Iso/Normal Paraffin Ratio ...................105
5.3. Sensitivity of the Distribution of Subclasses in Various Classes...........106
5.3.1. Analysis of the Distribution of Isoparaffins.................................108
5.3.2. Analysis of the Distribution of Isomers in Ring-Containing
IIIIIIIIIIIII
Hydrocarbon Classes ....................................................................110
VI KINETIC MODEL DEVELOPMENT FOR THE HYDROCRACKING
IIIIIIIIIIIIIOF PURE PARAFFINS................................................................................115
6.1. Introduction ............................................................................................115

CHAPTER

Page

6.2. Kinetics of Hydrogenation/Dehydrogenation Reactions........................117

liq
6.3. Expressing K DH
,ij in Terms of the True Thermodynamic
IIIIIIIIIIIIIIIII Equilibrium Coefficient ..........................................................................119
6.4. Modeling of the Metal Site Rate Parameters .........................................120
6.5. Modeling of the Acid Site Rate Parameters ...........................................122
6.5.1. Evans-Polanyi Relationship .........................................................124
6.5.2. Rate Equations for the Acid Site Steps ........................................124
6.6. Late Thermodynamic Lumping of Components ....................................129
6.7. Calculation of the Global Rate of Conversion of Lumps.......................129
6.8. Lumping Coefficients for the Steps on the Metal Sites .........................130
6.9. Lumping Coefficients for the Steps on the Acid Sites ...........................131
6.10. Net Rate of Formation of Paraffinic and Olefinic Lumps....................133
6.11. Reactor Model for Three-Phase Hydrocracking ..................................134
6.12. Model Degeneration into Equilibrated (de)Hydrogenation Case.........135
6.13. Estimation of the Model Parameters ....................................................136
6.13.1. Rate Parameters for the Metal Site Steps...................................136
6.13.2. Rate Parameters for Acid Site Elementary Steps.......................137
6.13.3. Parameters for Sorption in Zeolite Pores ...................................138
6.13.4. Parameters for Chemisorption on Metal Sites ...........................139
6.14. Approach to (De)Hydrogenation Equilibrium .....................................145
6.15. Application of the Model to Heavy Paraffins and Different Catalysts 146
6.15.1. Effect of Relative Metal/Acid Activity of the Catalyst..............147
6.15.2. Effect of Temperature ................................................................152
6.15.3. Effect of Total Pressure and H2/HC Molar Ratio.......................153
6.15.4. Effect of Feed Chain Length ......................................................156
VII CONCLUSIONS ..........................................................................................157
7.1. Recommendations for the Future Work....162
NOMENCLATURE.......................................................................................................164
REFERENCES...............................................................................................................170
VITA ..............................................................................................................................180

xi

LIST OF TABLES
TABLE

Page

2-1.

(Z-CN) matrix representation of VGO in terms of 462 components/lumps. ........23

2-2.

(Z-CN) matrix representation of VGO in terms of 1266 components/lumps .......27

2-3.

The number of lumps/pure components in different subclasses. ..........................28

3-1. Effect of hydrogen pressure on the aromatics content in the

ddddddhydrotreatment of Arabian Light VGO. ...............................................................46
3-2. Classification of aromatics based on the number of moles of
ddddddhydrogen required for hydrogenation...................................................................50
3-3.

Gas phase proton affinities for the formation of most stable carbenium ion........66

3-4.

Proton affinities for different classes used in the model.......................................67

4-1.

Model parameters employed in the VGO hydrocracking model. .........................76

5-1.

Reactor geometry and catalyst properties for the reactor simulations. .................81

5-2.

Composition of the VGO ......................................................................................82

5-3.

Definition of different commercial fractions based on carbon number range. .....83

5-4.

The distribution of isoparaffins selected for the simulations of Set I. ................107

5-5. The distribution of isomers selected for the ring species for the
ddddddsimulations of Set II. ..........................................................................................107

xii

LIST OF FIGURES
FIGURE

Page

1-1.

Simplified process flow diagram of a two stage hydrocracker...............................5

2-1.

Steps involved in the hydrocracking of paraffins. ................................................13

2-2.

Elementary steps of mono-ring naphthenes ..........................................................14

2-3.

Reaction network of multi-ring naphthenes..........................................................15

2-4.

Elementary steps and reactions of monoring aromatics. ......................................16

2-5.

Reaction network of multi-ring aromatics. ...........................................................17

2-6. The number of different types of elementary steps in the

iiiiiiiiiiihydrocracking of paraffins. ..................................................................................18
2-7. The total number of elementary steps in the network generated
iiiiiiiiiiifor paraffins, mono-naphthenes, di-naphthenes and mono-aromatics. ................19
2-8. The number of carbenium ions in the network generated
iiiiiiiiiifor paraffins, mono-naphthenes, di-naphthenes and mono-aromatics. .................19
3-1. Favorable configuration of the cyclization of olefinic carbenium ion
iiiiiiiiiiformed by endocyclic -scission. ..........................................................................35
3-2.

Physical and chemical phenomena inside the catalyst pellet................................37

3-3.

Mechanism for the cyclization steps.....................................................................41

3-4. Illustration of the lumping of components for the estimation of the

iiiiiiiiiiglobal rate of conversion between lumps..............................................................54
3-5. Lumping coefficients for the consumption of different classes
iiiiiiiiiiithrough pcp(s;s) steps.. ........................................................................................58
3-6. Comparison of the true lumping coefficients with the lumping
iiiiiiiiiiicoefficients calculated with and without reference olefins. .................................62
3-7. The representative molecule considered for the dehydrogenation
iiiiiiiiiiiequilibrium coefficients of naphthenes. ...............................................................65
4-1. Comparison of the experimental and predicted product composition
iiiiiiiiii in the hydrocracking of VGO. (a) Carbon number distribution (C3 to C40)
iiiiiiiiii of the total reactor effluents. (b) Carbon number distribution (C3 to C40)
iiiiiiiiii of the total paraffins. (c) Weight percent composition of normal paraffins,
iiiiiiiiii isoparaffins, one to four ring naphthenes and one to four ring aromatics.
iiiiiiiiii (d) Weight percent of different commercial fractions, i.e., LPG, light naphtha,
iiiiiiiiii heavy naphtha, kerosene, diesel and the unconverted VGO. ...............................78

xiii

FIGURE

Page

4-2. Comparison of the experimental and predicted product composition

iiiiiiiiii in the hydrocracking of VGO. (a) Carbon number distribution (C3 to C40)
iiiiiiiiii of the normal paraffins. (b) Carbon number distribution (C4 to C40) of the
iiiiiiiiii isoparaffins. (c) Carbon number distribution (C5 to C40) of the mononaphthenes.
iiiiiiiiii (d) Carbon number distribution (C10 to C40) of the dinaphthenes. .......................79
5-1. Effect of reactor inlet temperature on the (a) Temperature profile of
iiiiiiiiii solid and liquid phases along the reactor beds. (b) Weight percent conversion
iiiiiiiiii of VGO. (c) Hydrogen flux in the gas and liquid phases along the reactor beds.
iiiiiiiiii (d) Weight percent vaporization of hydrocarbons along the reactor beds. ..........86
5-2. Effect of reactor inlet temperature on the (a) Evolution of various
iiiiiiiiii commercial fractions. (b) Evolution of paraffins and one to four ring
iiiiiiiiii naphthenes. (c) Evolution of one to four ring aromatics. (d) Evolution of
iiiiiiiiii naphtheno-aromatics.. ..........................................................................................87
5-3. Effect of reactor inlet temperature on the (a) Evolution of naphthenoiiiiiiiiii aromatics. (b) Evolution of total paraffins, naphthenes, aromatics and
iiiiiiiiii naphtheno-aromatics. (c, d) Evolution of C4 to C32 hydrocarbons. .....................88
5-4. Effect of reactor inlet temperature on the (a) Evolution of different
iiiiiiiiii components of LPG. (b) Evolution of different classes of light naphtha.
iiiiiiiiii (c) Evolution of different classes of heavy naphtha. (d) Evolution of
iiiiiiiiii different classes of kerosene. (e) Evolution of different classes of diesel.
iiiiiiiiii (f) Evolution of different classes of unconverted fraction. ..................................89
5-5. Effect of reactor inlet temperature on the iso to normal paraffins
iiiiiiiiii ratio of various commercial fractions.. ................................................................90
5-6. Effect of reactor inlet temperature on the carbon number
iiiiiiiiii distribution of hydrocarbon classes......................................................................91
5-7. Effect of reactor pressure on the (a) Temperature profile of solid
iiiiiiiiii and liquid phases along the reactor beds. (b) Weight percent conversion
iiiiiiiiii of VGO. (c) Hydrogen flux in the gas and liquid phases along the reactor
iiiiiiiiii beds. (d) Weight percent vaporization of hydrocarbons along the reactor
iiiiiiiiii beds. ...................................................................................................................100
5-8. Effect of reactor pressure on the (a) Evolution of different commercial
iiiiiiiiii fractions. (b) Evolution of paraffins and one to four ring naphthenes.
iiiiiiiiii (c) Evolution of one to four ring aromatics. (d) Evolution of naphthenoiiiiiiiiii aromatics. ...........................................................................................................101
5-9. Effect of reactor pressure on the (a) Evolution of the naphthenoiiiiiiiiii aromatics. (b) Evolution of total paraffins, naphthenes, aromatics and
iiiiiiiiii naphtheno-aromatics. (c, d) Evolution of C4 to C32 hydrocarbons. ...................102

xiv

FIGURE

Page

5-10. Effect of reactor pressure on the (a) Evolution of different components

iiiiiiiiii of LPG. (b) Evolution of different classes of light naphtha. (c) Evolution of
iiiiiiiiii different classes of heavy naphtha. (d) Evolution of different classes of
iiiiiiiiii kerosene. (e) Evolution of different classes of diesel. (f) Evolution of different
iiiiiiiiii classes of unconverted fraction.. ........................................................................103
5-11. Effect of reactor pressure on the iso to normal paraffins ratio
iiiiiiiiii of different commercial fractions.......................................................................104
5-12. Molar composition of the normal, monobranched, dibranched and
iiiiiiiiii tribranched isomers among the total paraffins along the reactor length ............109
5-13. Molar composition of the unbranched, monobranched, dibranched and
iiiiiiiiii tribranched isomers for different hydrocarbon classes along the reactor
iiiiiiiiii length for case-3. ................................................................................................112
5-14. Molar composition of the unbranched, monobranched, dibranched and
iiiiiiiiii tribranched isomers for different hydrocarbon classes along the reactor
iiiiiiiiii length for case-4. ................................................................................................113
5-15. Molar composition of the unbranched, monobranched, dibranched and
iiiiiiiiii tribranched isomers for different hydrocarbon classes along the reactor
iiiiiiiiii length for case-5. ................................................................................................114
6-1. Importance of different dehydrogenation modes with respect to
iiiiiiiiii carbon number....................................................................................................122
6-2. Energy changes associated with phase transition of (a) carbenium ion
iiiiiiiiii (b) olefin.............................................................................................................125
6-3. Total and isomerization conversion of nC16 as a function of space time
iiiiiiiiii at T = 300 oC, P = 35 Bar and H2/HC ratio = 11.0.............................................142
6-4. Carbon number distribution of cracked products in the hydrocracking of nC16.
iiiiiiiiii (a) T = 300 oC, P = 35 Bar, H2/HC ratio = 11.0 and total conversion = 57 %
iiiiiiiiii (b) T = 320 oC, P = 35 Bar, H2/HC ratio = 11.0 and total conversion = 78 %...143
6-5. Percentage of isomers per carbon number in the hydrocracking of nC16
iiiiiiiiii (a) T = 300 oC, P = 35 Bar, H2/HC ratio = 11.0 and total conversion = 57 %
iiiiiiiiii (b) T = 320 oC, P = 35 Bar, H2/HC ratio = 11.0 and total conversion = 78 %...144
6-6. Evolution of the isomers of different degrees of branching for the
iiiiiiiiii hydrocracking of nC32 corresponding to equilibrated (de)hydrogenation
iiiiiiiiii condition (Ac-rds case). .....................................................................................149
6-7. Evolution of the isomers of different degrees of branching for the
iiiiiiiiii hydrocracking of nC32 corresponding to non-equilibrated (de)hydrogenation
iiiiiiiiii condition (Me-Ac-rds case)................................................................................149

xv

FIGURE

Page

6-8. Effect of relative metal/acid activity on initial isomerization selectivity

iiiiiiiiii in nC32 hydrocracking. .......................................................................................150
6-9. Simulated distribution of cracked products for Ac-rds case at various
iiiiiiiiii total conversions for the hydrocracking of nC32. ...............................................150
6-10. Simulated distribution of cracked products for Me-Ac-rds case at various
iiiiiiiiii total conversions for the hydrocracking of nC32.................................................151
6-11. Influence of the relative metal/acid activity on the maximum isomerization
iiiiiiiiii conversion for the hydrocracking of nC32. .........................................................151
6-12. Influence of temperature on the maximum isomerization conversion
iiiiiiiiii for the hydrocracking of nC32 over Cat-II. .........................................................153
6-13. Effect of total pressure on the value of o , nC 32 at T=375 oC and R=11
iiiiiiiiii for the hydrocracking of nC32 over different catalysts. ......................................155
6-14. Effect of total pressure on the total conversion, isomerization conversion
iiiiiiiiii and secondary cracking at T=375 oC and R=11 for the hydrocracking of nC32
iiiiiiiiii over Cat-I............................................................................................................155
6-15. Effect of chain length on the total conversion and secondary
iiiiiiiiii cracking expressed as the moles of cracked products formed
iiiiiiiiii per 100 moles of feed cracked at 375 oC, 150 bar and H2/HC molar ratio
iiiiiiiiii of 11.0 for a given space time over Cat-I...........................................................156

CHAPTER I
1.INTRODUCTION
Hydrocracking at high hydrogen pressures is commonly used to upgrade the heavier
fractions like vacuum gas oils (VGO) obtained from crude oil distillation.
Hydrocracking of residues and several products obtained from other refining processes,
i.e., coker gas oil, deasphalted oil, fluid catalytic cracking (FCC) cycle oils and decant
oils etc. is also carried out for getting high value products. Hydrocracking is also applied
in the production of superior quality middle distillates from Fischer-Tropsch waxes and
in the production of lube oil bases from middle distillates by removing long chain
paraffins (dewaxing). Very heavy hydrocarbons such as those extracted from tar sands
and shale can also be upgraded by hydrocracking.
A typical VGO feedstock consists of paraffinic, naphthenic, aromatic and naphthenoaromatic species, along with heteroatom impurities like sulfur, nitrogen, oxygen
compounds. Significant amount of metals may also be present. Based on the atmospheric
equivalent boiling points1 (AEBP), vacuum gas oils (VGO) can be classified as light
vacuum gas oils (LVGO) and heavy vacuum gas oils (HVGO). LVGOs have a boiling
range of about 345-430 oC. Typically, saturates in this range account for 40-60 wt% of
the fraction, consisting mainly of paraffins, and alkylnaphthenes with 1-4 rings. They
have carbon numbers between 15 and 30. The aromatic content of LVGO may be as
high as 50 wt%. Most compound types in this fraction are alkylbenzenes, naphthalenes
and phenanthrenes, with or without naphthenic rings. Sulfur compounds like benzo-,
dibenzo-, naphthobenzo-thiophenes and other derivatives are also present. The
elemental sulfur, oxygen and nitrogen content is of the order 3 wt%, 0.2 wt% and 1500
ppm respectively1.

This dissertation follows the style of Industrial & Engineering Chemistry Research.

The polar compounds in LVGOs may range from 1-10 wt%, consisting of compounds
containing nitrogen, oxygen, or both. Snyder2 identified the polar compounds in the
LVGO range of a California crude oil and reported 3.4 wt% carbazoles, 0.6% indoles,
0.5% benzocarbazoles, 0.66% pyridines, 1.74% quonolines, 0.26% benzoquonolines,
and 1.2% of pyridones etc. Furthermore, they found 1.74% aliphatic carboxylic acids,
0.97% phenols, 1.2% benzoaromatic and higher aromatic furanes, and 0.7% aliphatic
esters, ketones, and other carbonyl derivatives. The atmospheric equivalent boiling range
of HVGO is about 430540 oC with a carbon number ranging from 20 to 50. Compared
to LVGO there is a dramatic decrease in the amount of saturates and a corresponding
increase in aromatics and polar compounds. Most aromatic and naphthenic molecules in
this range primarily contain 1-4 rings. Compounds with more than 4 rings are present at
much lower concentrations. Elemental sulfur, oxygen and nitrogen content can be about
4 wt%, 0.3 wt% and 2000 ppm respectively. Snyder2 identified polar compounds in the
HVGO fraction of a California crude oil with 0.75 wt% indoles, 4.1% carbazoles, 1.3%
benzocarbazoles, 1.3% pyridines, 2.0% quonolines, 1.6% benzoquonolines, 2.0%
pyridones, 0.4% azaindoles, 0.85% dibenzofuranes, 0.8% benzonaphthofuranes, 0.15%
dinaphthofuranes, 0.08% dihydrobenzofuranes, 1.35% phenols, 1.35% aliphatic
carboxylic acids, and 2.3% aliphatic esters, ketones, and other carbonyl derivatives.
Significant differences can be found in the compositional analysis of different VGOs
depending upon the source or the origin of the feedstock.
Typically the products obtained from fractionation of the hydrocracker effluent are
defined based on their boiling point range as follows: light ends (C4- ), light naphtha (C5
80 oC), heavy naphtha (80 oC 150 oC), jet fuel/kerosene (150 oC 290 oC), diesel
fuel (290 oC 370 oC) and the unconverted fractionator bottoms (370 oC +). The
products obtained from hydrocracking are generally of high quality. The light naphtha
with a research octane number between 78 and 85 can be used for the blending stock in
the gasoline pool. The heavy naphtha is a good quality reformer feedstock for the
production of high octane gasoline. The jet fuel obtained is low in aromatics and has a
high smoke point. The diesel fraction has a relatively high cetane number and very low

sulfur content. High quality lube oils and FCC and thermal cracking feedstocks are also
obtained from the hydrocracking process3.
Environmental concern in recent years has set the agenda for the development of
hydrocracking and hydrotreatment processes to get high quality fuels with low
aromatics, sulfur and nitrogen content. Global hydrocracking capacity has increased
consistently over the last two decades3 and is expected to increase further, given the
ability of this process to upgrade heavier feedstocks. Significant developments have
been made in the hydrocracking process technologies like single-stage once through,
single-stage with recycle, two-stage process, two-stage with recycle etc. in an effort to
gain higher flexibility in the units operation and maximize the financial returns. At the
same time, different catalysts have been commercialized for processing the feedstocks of
vastly different compositions under less severe operating conditions. The development
of the fundamental kinetic models, on the other hand, for this kind of complex
conversion processes had been rather limited, with most of the kinetic models still based
on a very simplified lumped scheme for describing the product distribution. Therefore,
development of reliable mechanistic kinetic models for the hydrocracking process is an
important activity from a commercial as well as a research viewpoint.5 The design and
optimization of the hydrocracking units require a detailed kinetic model that can take
into account the complexity of the feedstock while following the rules of the underlying
carbenium ion chemistry.6 The use of comprehensive process models with an accurate
representation of hydrocracking kinetics at the elementary step level can be used to
reduce expensive experimentation in pilot plants. This kind of mathematical models can
be used for process design to predict the detailed product distribution and optimum
operating conditions for a range of feedstocks as well as for simulation and optimization
of the existing units.
1.1. Hydrocracking Process Description
Many different flow schemes have been developed for the hydrocracking process so that
various feedstocks can be processed to produce a full range of products. All of the

processes are vendor specific with respect to the reactor design and catalyst selection.
The three major schemes for hydrocracking processes can be classified as follows:
(a) Single-stage recycle hydrocracking
(b) Two-stage recycle hydrocracking
(c) Once through hydrocracking
In general, the commercial hydrocracking plants are operated at the following
conditions:3
Catalyst bed temperature

300-450o C

Pressure

85-200 bars

Liquid hourly space velocity 0.5-2.5 hr-1

H2/HC ratio

10-30 mole/mole (~3,000-10,000 Std. ft3/barrel)

Due to high hydrogen partial pressures and the use of dual function catalysts, the rate of
catalyst coking and deactivation is very low, resulting in on-stream cycle lengths of
several years.
The typical feedstocks used in hydrocracking process contain sulfur, nitrogen, and in the
case of resid feedstock, metals such as nickel and vanadium. Because such compounds
have a deleterious effect on hydrocracking catalysts, the feedstock typically requires
hydrotreatment prior to contact with the hydrocracking catalyst. For this reason, most of
the hydrocracking processes consist of two stages involving both hydrotreatment and
hydrocracking.
Figure 1-1 shows the simplified flow diagram for a two stage hydrocracking process
with recycle. The vacuum gas oil is sent to the first stage of the hydrocracker and is
severely hydrotreated. Most of the sulfur and nitrogen compounds are removed from the
oil and many of the aromatics are saturated. In addition, significant conversion to light
products occurs in the first stage. The liquid products from the first stage are sent to a
common fractionation section. To prevent overcracking, lighter products are removed by

distillation. The unconverted oil from the bottom of the fractionator is routed to the
second stage reactor section. The second reaction stage saturates almost all the aromatics
and cracks the oil feed to light products. Due to the saturation of aromatics, the second
stage produces excellent quality products. The liquid product from the second stage is
sent to the common fractionator where light products are distilled. The second stage
operates in a recycle to extinction mode with per-pass conversions ranging from 50 to
80%. The following products are obtained from fractionation: light ends, light naphtha,
heavy naphtha, jet fuel/kerosene, and diesel fuel. The fractionator bottoms containing the
unconverted feed is recycled to the second stage reactor so that it can be converted into
commercial products.

Figure 1-1. Simplified process flow diagram of a two stage hydrocracker. (source: www.abb.com).

The overhead liquid and vapor from the hydrocracker fractionator is further processed in
a light ends recovery unit where fuel gas, liquefied petroleum gas (LPG) and, naphtha
are separated. The hydrogen supplied to the reactor sections of the hydrocracker comes
from steam reformers. The hydrogen is compressed in stages until it reaches system
pressure of the reactor sections.
The catalyst in the first reactor is designed to eliminate the hetero compounds in the
feedstock and to convert the organic sulfur and nitrogen to hydrogen sulfide and
ammonia, respectively. Such catalysts typically comprise sulfided molybdenum or
tungsten and nickel or cobalt on an alumina support. The deleterious effect of H2S and
NH3 on hydrocracking catalyst is considerably lower than those of the corresponding
organic hetero compounds. The hydrotreating catalyst also facilitates the hydrogenation
of aromatics.
The hydrocracking catalyst in the second stage is designed to optimize the yields and
quality of the desired products. Various reactions such as hydrogenation,
dehydrogenation, isomerization, cracking, alkylation, dealkylation, etc. predominately
take place in the second stage reactor. Hydrogenation reactions are highly exothermic,
whereas the cracking reactions are endothermic. The amount of heat liberated in the
hydrogenation reactions is greater than the heat required for the endothermic cracking
reactions. The surplus heat released causes the reactor temperature to increase, thereby
accelerating the reaction rate. Cold hydrogen is injected between the reactor beds as a
quench to control the reactor temperature profile.
The severity of the hydrocracking operation is measured by the degree of conversion of
the feed to the lighter products. Conversion is defined as the volume percent of the feed
that disappears to form the products boiling below the desired product end point. A
given percent conversion at a low product endpoint represents a more severe operation
than does the same percent conversion at a higher product endpoint.

1.2. Literature Survey

To study the conversion of complex feedstocks through processes like hydrocracking
and catalytic cracking, most efforts have focused on the development of lumped kinetic
models in which the feedstock is divided into several lumps based on the boiling point
range. A simplified reaction network between these lumps is set up and the rate
coefficients for the global conversion of lumps are estimated from the experimental data.
For example, the kinetic model of Weekman and Nace4 for fluidized catalytic cracking
assumes that the feedstock charge is converted into the gasoline boiling fraction and the
remaining fraction by the following two equations,
k1
C1
a1C2 + a2C3
k2
C2
C3

(1.1)
(1.2)

In the above equations, C1 represents the gas oil charged, C2 represents the C5-410 oF
gasoline fraction and, C3 represents the butanes, dry gas and, coke. The coefficients a1
and a2 represent the mass of C2 and C3 produced per mass of C1 converted, respectively.
A more detailed lumped model for FCC comprising 10 lumps was developed by Jacob et
al5. Stangeland6 considered the feedstock as a series of 50 oF boiling range cuts assuming
that each heavier cut hydrocracks via a first order reaction to form a series of lighter
cuts. To achieve higher accuracy in the product yields predicted by the models, more and
more lumps were introduced by various researchers. Increasing the number of lumps
also leads to the introduction of more parameters in the kinetic model. The major
fundamental limitation of the lumped kinetic models is that the kinetic parameters
depend on the composition of the feedstock. Therefore, with every different feedstock
the kinetic model needs to be refitted and new sets of parameters have to be estimated.
Another approach for modeling the kinetics of the conversion of such complex
feedstocks is based on the notion of continuum lumping in which the reaction mixture is
considered to be a continuous mixture with respect to the feed properties like boiling
point and molecular weight. A hydrocracking model based on this approach has been
developed by Laxminarasimhan et al.7 in which the true boiling point of the mixture is

used as the characterization parameter. The rate constant of hydrocracking is assumed to

be a monotonic function of the true boiling point. A yield distribution function was
formulated from the hydrocracking data of model compounds. The resulting integrodifferential equations were solved numerically to obtain the yields of various fractions as
a function of reactor residence time. Similar to the discrete lumping, this approach is
also unable to capture the fundamental chemistry of the process providing thrust to the
development of mechanistic models.
Mechanistic models are based more closely on the chemistry of the process. Several
approaches have been developed to build the mechanistic models for catalytic and
thermal conversion of complex feedstocks. For example, Liguras et al.8, 9 represented the
feedstock using hundreds of pseudo-components whose composition was calculated
from MS and C-NMR spectroscopy and developed the kinetic model for catalytic
cracking based on the nature of carbon centers in the pseudo-components.
Quann et al.10 constructed the molecules by incrementing the structural units present in
the hydrocarbon molecules and expressed them using a vector notation which was used
to generate the reaction network. The reaction network is based on various reaction rules
selected according to the chemistry of the process. This approach is called the structural
oriented lumping (SOL).
Froment and co-workers11,

12

developed a mechanistic kinetic modeling approach

starting from the elementary steps of carbenium ion chemistry. This approach was
named as single event approach. Baltanas et al.11 generated a network of elementary
steps involving carbenium ions using a computer algorithm based on the approach
devised by Clymans et al.13 and Hillewaert et al.14. Vynckier et al.12 applied the single
event approach to complex feedstocks by introducing the concept of lumping
coefficients. Feng et al.15 used single event to model the catalytic cracking of paraffins
on a RE-Y zeolite catalyst. Svoboda et al.16 determined the single event rate parameters
for the hydrocracking of n-octane. Martens et al.17 applied single event kinetics for the
hydrocracking of C8-C12 paraffins on Pt/USY zeolites. Park and Froment18 applied the
single event kinetics along with Evans-Polanyi relationship for modeling the methanol to

olefin process over HZSM-5 catalyst, and Martinis and Froment19 to the alkylation of
isobutane with butenes.

10

CHAPTER II
2.HYDROCRACKING CHEMISTRY AND REACTION NETWORK
GENERATION
2.1. Hydrocracking Catalyst
Hydrocracking is carried out on dual-function catalysts having a cracking function and a
hydrogenation-dehydrogenation function. The cracking function is provided by an acidic
support, whereas the hydrogenation-dehydrogenation function is provided by metals.
The different kinds of acidic support can be (i) amorphous oxides e.g., silica-alumina,
(ii) crystalline zeolites, mostly ultra stable Y-zeolites and a binder e.g., alumina, or (iii)
hybrid supports having a mixture of zeolite and amorphous oxides. Cracking and
isomerization steps take place on the acidic support. The metal can be noble metals
(palladium, platinum) or nonnoble metal sulfides from group VIA (molybdenum,
tungsten) and group VIIIA (cobalt, nickel). These metals produce the reactive olefins for
the cracking and isomerization steps and catalyze the hydrogenation of the aromatics in
the feedstock as well as heteroatom removal. The ratio between the catalysts acidic and
metal activities is adjusted in order to optimize the activity and selectivity.3
2.2. Reaction Mechanism
The well known bifunctional mechanism of hydrocracking reactions proceeding through
olefinic intermediates is widely accepted for describing the product distribution. The
model developed in this work is primarily applicable to the hydrocracking of VGO on
zeolite catalysts ignoring the activity of the binder. The feed molecules are sorbed into
the pores of zeolites from the surrounding fluid phase, which can be gas or liquid
depending upon the reactor operating conditions. Different reaction pathways are
available for the paraffinic, naphthenic and aromatic species. The sorbed paraffinic and
naphthenic species are chemisorbed on the metal sites and are dehydrogenated into the
corresponding olefinic species which migrate from the metal sites to the nearby Brnsted
acidic sites and are protonated into paraffinic and naphthenic carbenium ions. The

11

naphthenic species can dehydrogenate either in the side chain or inside the ring to
produce the corresponding olefins. The olefinic species sorbed in the zeolite pores are
assumed to be in pseudo-equilibrium with the chemisorbed olefins on the metal sites as
well as with the surface carbenium ions on the acid sites. The pseudo-equilibrium of
sorbed olefins between the metal sites as well as with the carbenium ions is modeled
using the Langmuir adsorption isotherm.
The paraffinic carbenium ions thus produced are isomerized by hydride shift (HS),
methyl shift (MS) and protonatedcyclopropane (PCP) steps. It can be noticed from
Figure 2-1 that the HS and MS steps do not change the degree of branching in the
carbenium ions whereas the PCP steps do. The isomerized carbenium ions with a higher
degree of branching crack at the carbon-carbon bond in -position with respect to the
carbon atom carrying the positive charge, yielding a smaller carbenium ion and an
olefin. The resulting carbenium ion can further crack or deprotonate on the acid sites to
produce an olefin. Similarly, the olefin can be protonated to yield another carbenium ion,
or alternatively can be hydrogenated on the metal sites to produce a paraffin. The
probability of undergoing either protonation or hydrogenation depends on the relative
strength of the acid/metal functions of the catalyst. Cyclization of paraffinic species into
ring species is also encountered. The important elementary steps that occur in the
paraffinic species are shown in Figure 2-1.
The naphthenic carbenium ions can undergo all the elementary steps mentioned above in
their alkyl side chain. Apart from that, the hydride shift and methyl shift steps can also
occur in the naphthenic rings. The PCP steps involving the ring carbon atoms result in
the expansion or contraction of the ring size. It has been shown that the hydrocracking of
cyclohexane proceeds through ring contraction forming methylcyclopentane which
undergoes the ring opening reaction. Depending on the location of the positive charge,
the naphthenic carbenium ions can undergo exocyclic -scission producing a naphthenic
olefin and severing the long alkyl chain as a paraffinic carbenium ion, or can transform
through endocyclic -scission scission producing an unsaturated carbenium ion by ring
opening. One of the main differences between the -scission in the paraffinic chain and

12

-scission in the ring is that the latter is considerably more difficult than the former.
Several explanations have been presented for this behavior. According to Eagan and
Langlois20, carbenium ions formed by endocyclic -scission of a naphthenic carbenium
ions have a high tendency to return to the cyclic form because the closeness of the
double bond and positively charged carbon atom provides a favorable configuration for
this cyclization. Studies on the hydrocracking of multi-ring naphthenes are mostly
limited up to two rings and have shown that the naphthenic rings open in a sequential
manner. The important steps in the hydrocracking of mononaphthenes are shown in
Figure 2-2.
Figure 2-3 shows the different possibilities that can occur in the hydrocracking of a tetranaphthene during hydrocracking. For simplicity, the positive charge has been omitted
from the reaction network.
Hydroconversion of alkyl aromatics proceeds either through the typical acid-site
elementary steps in their side chain or through the saturation of the aromatic ring and
subsequent isomerization and cracking in the resulting naphthenic ring. Dealkylation of
the benzyl carbenium ion formed by the direct protonation of the aromatic ring on the
acid sites is an important step in the hydrocracking of aromatics. Figure 2-4 shows the
important steps in the hydroconversion of mono-ring aromatics. Hydrocracking of
polynuclear aromatics is quite complex and involves sequential hydrogenation of the
aromatic rings along with parallel or subsequent isomerization, cracking and ring
opening reactions as shown in Figure 2-5.
To summarize, the following types of elementary steps and reactions are considered in
the current VGO hydrocracking model:
1. Dehydrogenation into olefinic species (preceding the protonation step)
2. Protonation/deprotonation
3. Hydride shift (HS) in the side chain and in the rings
4. Methyl shift (MS) in the side chain and in the rings

13

5. Protonated cyclopropane (PCP) in the side chain

6. Ring contraction/expansion through PCP in the ring
7. Acyclic -scission (scission in the side chain)
8. Exocyclic -scission (scission between ring and side chain)
9. Endocyclic -scission or ring opening (scission in the ring)
10. Cyclization (reverse of endocyclic -scission)
11. Dealkylation of aromatics
12. Hydrogenation of aromatics to naphthenes (and the reverse dehydrogenation)

-H2
H+

HS
+

+
+

MS
+

PCP
+

scission

Figure 2-1. Steps involved in the hydrocracking of paraffins.

14

-H2

Dehydrogenation

+H+

Protonation

Hydride Shift

Methyl Shift

PCP (in side chain)

PCP (in ring)

+

+
Acylic beta scission (in side chain)

+
+

Exo-cylic beta scission

Endo-cyclic beta scission

Figure 2-2. Elementary steps of mono-ring naphthenes.

15

+
+

+
+
+

Figure 2-3. Reaction network of multi-ring naphthenes. (Shown in the form of molecular reactions instead
of elementary steps for simplification).

16

-H2

Dehydrogenation

+H+
+

Protonation (in side chain)

+H+

H
Protonation (on the ring)

Hydride Shift

+
+

Methyl Shift

PCP (in side chain)

+
Acylic beta scission (in side chain)
CH4
+

+
+

Dealkylation

Figure 2-4. Elementary steps and reactions of monoring aromatics.

17

Acid Sites

+
Metal Sites

Figure 2-5. Reaction network of multi-ring aromatics. (Shown in the form of molecular reactions instead
of elementary steps for simplification).

2.3. Reaction Network Generation

Considering the large number of reaction pathways in the hydrocracking of heavy
feedstocks when expressed in terms of elementary steps, a reaction network for the
paraffins, naphthenes and aromatics has been generated up to carbon number 40 using
Boolean relation matrices and characterization vectors. The methodology and procedure
for generating the reaction network has been described in the literature13, 21, 22. Several
improvements have been made in the existing reaction network generation algorithm to

18

speed up the program execution. The use of dynamic memory allocation via linked lists
for storing and searching the intermediate olefinic and ionic species significantly
reduced the run time as compared to searching the species stored on the hard disk. The
rules for deriving the number of single-events of the PCP and cracking steps have also
been improved.
The total number of carbenium ions involved in VGO hydrocracking up to C40 is of the
order of 18 million. The total number of steps and the number of rate determining steps
are of the order of 126 and 15 million respectively. For illustrative purpose, some of the
information extracted from the computer generated reaction network has been presented
in Figure 2-6 to Figure 2-8 showing the increase in the number of species and the

Millions

number of steps with carbon number.

0.35
Npr/Depr
0.3

Nhs
Nms

Number of Steps

0.25

Npcp
Ncrk

0.2
0.15
0.1
0.05
0
0

10

15

20

25

30

35

40

45

Carbon Num ber

Figure 2-6. The number of different types of elementary steps in the hydrocracking of paraffins.

Millions

19

4
Paraffins
3.5

MNAP
DNAP

Total Number of Steps

MARO

2.5
2
1.5
1
0.5
0
0

10

15

20

25

30

35

40

45

Carbon Num ber

Figure 2-7. The total number of elementary steps in the network generated for paraffins, mono-

Millions

naphthenes, di-naphthenes and mono-aromatics.

0.6
Paraffins
MNAP

0.5

Number of Ions

DNAP
MARO

0.4

0.3

0.2

0.1

0
0

10

15

20

25

30

35

40

45

Carbon Num ber

Figure 2-8. The number of carbenium ions in the network generated for paraffins, mono-naphthenes, dinaphthenes and mono-aromatics.

20

2.4. Composition of VGO

To develop a mechanistic kinetic model requires the knowledge of the composition of
the feedstock at the molecular level. Just for the paraffins in the C5-C12 range (a typical
carbon number range for naphthas) the number of possible isomers amounts to more
than 600. The number of isomers which have actually been experimentally observed
(though not all identified) in this range is of the order of 200-40023-26. This kind of
detailed molecular analysis of light petroleum fractions like gasoline and light naphtha
can be carried out using coupled gas chromatography/mass spectroscopy. For any higher
boiling range cuts, this number becomes quickly unmanageable. For higher boiling
fractions, adding the isomers of paraffins, naphthenes, and aromatics as well as the
heteroatom compounds, the number becomes larger by several orders of magnitudes
making the complete compositional analysis of heavy fractions utterly impossible1 even
by the latest analytical techniques. However, significant developments have been made
in recent years for a quite detailed analysis of these heavy mixtures with the application
of modern analytical techniques including gas chromatography (GC), supercritical fluid
chromatography (SFC), high performance liquid chromatography (HPLC), field
ionization mass spectroscopy (FIMS), low voltage high resolution mass spectroscopy
(LV-HRMS), nuclear magnetic resonance spectroscopy (NMR) etc. Boduszynski27,

28

for example, has coupled high performance liquid chromatography (HPLC) for
separations and FIMS in the analysis of several petroleum crude oils. Similarly, Allen29
has coupled HPLC/GC compound class separations along with mass spectral,

13

C, 1H,

nuclear magnetic resonance analysis in detailed description of gas oil feedstocks. In the
work of Boduszynski, the petroleum feed was fractionated into ten boiling point cuts,
each of which was subjected to three HPLC extractive separations to determine nine subfractions: the saturates, monoaromatics, diaromatics, triaromatics, tetraaromatics,
pentaaromatics, and the polar compounds (the basic, pyrolic nitrogen and acid
components). Each of these sub-fractions was analyzed in detail by FIMS which gave
the molecular weight distribution or carbon number distribution of different homologous
series. The results available using this technique are valuable quantitative information

21

about the distribution of various components in the complex feedstocks. This kind of
analysis allows to represent the composition of a heavy petroleum fraction in a matrix
form in which the rows represent different molecular classes and the columns contain the
carbon number distribution for each class. Compounds in each class have a distinct
hydrogen deficiency number Z (as in CnH2n+Z), where Z = -2(R + DB -1), R is the
number of rings and DB is the number of double bonds in the given class. One such (ZCN) representation to characterize a VGO feedstock is presented in Table 2-1. The entire
VGO is divided into 16 molecular classes and in each class the carbon number
distribution is given up to C40 resulting in 462 pure components/lumps in total. In this
representation, all the isomers belonging to a particular class with a given carbon
number are placed in one lump with no distinction based on the degree of branching or
ring substitution in the isomeric species.
Because of the difficulties associated with this type of detailed characterization of
heavier petroleum feedstocks and at the same time its increasing importance in the
development of fundamental kinetic models, a large amount of research is directed
towards the determination of detailed composition from the bulk analysis of heavy
hydrocarbons such as ASTM distillation curves, API gravity, Reid vapor pressure
(RVP), paraffin, naphthene and aromatic (PNA) content, Conradson carbon residue
(CCR) etc. using different approaches.

22

Neurock et al.30, 31 devised a stochastic method to transform efficient sets of analytical

characterizations into molecular representations of complex feedstocks. They assembled
the important structural attributes of petroleum molecules like the number of aromatic
rings, number of naphthenic rings, number and length of paraffinic side chain etc. into
molecules according to quantitative probability density functions for each attribute
resulting in a large ensemble of representative key components from which both
molecular and global product properties were deduced. These density functions were
derived from a general Monte Carlo algorithm. In a similar work Hudebine et al.32
created the molecular composition from partial analytical data using a two-step
reconstruction method. The first step called the stochastic reconstruction assumes that oil
mixtures can be described by distributions of structural blocks. The choice of blocks and
distributions is based on expert knowledge. The transformation from a set of
distributions into a mixture of molecules is obtained by Monte-Carlo sampling while a
simulated annealing procedure adjusts the parametric distributions. The second step
termed as reconstruction by entropy maximization improves the representativeness of
the set of constructed molecules by adjusting their molar fractions. The estimation of
these mole fractions is carried out by maximizing an information entropy criterion under
linear constraints. Aye et al.33 developed a methodology for the generation of detailed
composition for the gasoline range fractions from bulk properties. They used a database
of the properties and molecular information in the form of Z-CN matrices of several
refining streams and through the linear combination of these matrices predicted the
composition of new feedstocks.

NDA

NTA

DNMA

DNDA

TNMA

CN

No of Lumps

38

37

36

31

27

23

35

31

27

23

31

27

23

27

23

23

462

LNAP LPG
HNAP
KERO
DIESEL
UNCONVERTED

CN

NMA

TNMA

DNDA

DNMA

NTA

NDA

NMA

QAR

TAR

DAR

MAR

440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462

QAR

417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439

TAR

367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389

390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416

DAR

286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308

340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366

MAR

259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285

309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339

QNA

170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192

228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258

TNA

143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169

QNA

TNA

DNA
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142

193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227

3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40

DNA

MNA
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
MNA

IPAR
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
IPAR

NPAR
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
NPAR

CN
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
CN

UNCONVERTED

DIESEL

KERO

HNAP

LNAP LPG

Table 2-1. (Z-CN) matrix representation of VGO in terms of 462 components/lumps.

Abbreviations of different classes: nPar: Normal Paraffins, iPar: Iso Paraffins, MNA: Mono-Naphthenes, DNA: Di-Naphthenes, TNA: Tri-Naphthenes, TETNA:
Tetra-Naphthenes, MAR: Mono Aromatics, DAR: Di-Aromatics, TAR: Tri-Aromatics, TETAR: Tetra-Aromatics, NMA: Naphtheno Mono Aromatics, NMA:
Naphtheno Di-Aromatics, NTA: Naphtheno Tri-Aromatics, DNMA: Di Naphtheno Mono Aromatics, DNDA: Di Naphtheno Di Aromatics, TNMA: Tri Naphtheno
Mono Aromatics

23

24

Unfortunately, the analysis of VGO even at such a detailed level (Z-CN matrix form)
does not allow development of a mechanistic kinetic model in which the model
parameters could be claimed to be independent of the feedstock composition. To explain
the reason, refer to the strict thermodynamic lumping scheme developed in the Chapter
III. This lumping scheme is based on the assumption that the isomers of a particular
carbon number belonging to a class and with a given number of methyl branches reach
thermodynamic equilibrium due to fast hydride shift and methyl shift elementary steps
on the acid sites of the catalyst. Therefore, the model requires three lumps per carbon
number for isoparaffins, i.e., monobranched, dibranched and tribranched lumps. In the
VGO composition shown in Table 2-1, however, all the isomers of a given class are
placed in one lump. Therefore, such a characterization of VGO would not allow the
calculation of the rate of conversion of less branched isomers to more branched isomers
(and vice versa) occurring through protonatedcyclopropane (PCP) mechanism, which is
an important rate determining step in the hydrocracking process. In other words,
according to the lumping scheme given in Table 2-1, all the isomers in a lump with
different number of methyl branches would have to be assumed at pseudo-equilibrium
along the entire length of the reactor which is far from reality. Vansina et al.34 compared
the experimental product distribution with the thermodynamically calculated values in
the hydrocracking of n-octane and showed that the fraction of multi-branched isomers is
always smaller than the thermodynamic equilibrium. Especially, the fraction of tribranched isomers was much smaller than their thermodynamic equilibrium values.
Similarly, Schulz et al.35 reported that in the hydrocracking of n-dodecane on different
catalysts, the ratios of the monomethyl isomers to the corresponding n-paraffins are
greater than those of thermodynamic equilibrium, whereas the amount of dimethyl
isomers are much smaller. This idea has been extended to the ring containing structures
also given the presence of long paraffinic side chains in these species and the same
reaction mechanism (i.e., PCP) for the change in the number of branches. Since the
global rate of cracking of a lump through -scission increases with the degree of
branching, placing all the isomers per carbon number of a given class in one lump would

25

not give the accurate values of the global rate of conversion through -scission steps,
which are the main elementary steps responsible for the transformation of heavy
hydrocarbons into the lighter ones. Consequently, in the current model to represent the
carbenium ion chemistry at the fundamental level, all the classes except normal paraffins
and those with four rings (naphthenic or aromatic) structures, have been further divided
into subclasses based on the number of methyl branches. Normal paraffins are obviously
pure components and the four ring structures are not divided because of their relatively
low concentrations. According to the rules set for the reaction network generation22,
paraffinic species can have a maximum of three methyl branches; ethyl and longer
branches are not allowed in the reaction network owing to relatively smaller
concentrations of this kind of isomers. Therefore, the isoparaffin class has been divided
into three subclasses, i.e., monobranched, dibranched and tribranched. The species
containing naphthenic or aromatic rings are allowed to have only one long side chain at a
given substitution, and up to three methyl branches anywhere on the ring structure or the
long side chain. Therefore, for a given carbon number of any ring containing class, the
first subclass contains the species having just the bare ring structure (i.e., the first
member of the class, e.g. benzene in monoaromatic class), or the species having only the
long side chain without any methyl branch. This subclass is referred to as unbranched in
the rest of this dissertation. The other three subclasses contain species with one, two, and
three methyl branches denoted by monobranched, dibranched and tribranched
subclasses, respectively. According to this scheme, the 16 molecular classes are divided
into 45 subclasses leading to characterization of VGO in terms of 1266 pure
components/lumps. This scheme for VGO characterization is shown in Table 2-2. The
summary of Table 2-2 has been presented in Table 2-3 showing the number of lumps for
each subclass. Naturally, it is very difficult to characterize a heavy feedstock like VGO
at such a detailed level. Mass spectrometers/high resolution mass spectrometers would
not be able to distinguish among the isomers with the same chemical formula. But
techniques like 1H and

13

C NMR can give important information about the average

population of carbon and hydrogen atoms of different nature, i.e., primary, secondary,

26

tertiary, aromatic, -position to aromatic ring etc. of a fraction separated from VGO by
GC/HPLC/SGC. This information can be utilized to get a good estimate of the
distribution of isomers with different degrees of branching to convert the VGO
composition from 16 classes to 45 subclasses. Development of more advanced analytical
methods would certainly be useful for obtaining this kind of information.
As the information on the distribution of isomers with different degrees of branching in a
class for heavy petroleum mixtures is not available, some insight has been taken from
the detailed molecular level analysis of lighter petroleum fractions. It has been generally
observed that in the straight run petroleum cuts, the isoparaffins contain primarily the
monobranched isomers with relatively smaller amounts of multibranched paraffins.
Accordingly, in the current model the distribution of monobranched, dibranched and
tribranched isomers is assumed to be 65%, 25% and 10%, respectively. In the classes
with aromatic and naphthenic rings, a distribution of 20%, 50%, 20% and 10% has been
assumed for the unbranched, monobranched, dibranched and tribranched isomers. The
distribution of the degree of branching can be significantly different in the VGOs
obtained from other processing units as compared to the straight run VGOs and need to
be accounted for accordingly. A sensitivity analysis has been performed to show the
effect of the distribution of isomers in the feed on the hydrocracking products. It will be
discussed in Chapter V.

Lumps
Cumulative

MNA3

DNA0

DNA1

DNA2

DNA3

TN A 0

TN A 1

TN A 2

TN A 3

TETN A

MAR0

MAR1

MAR2

MAR3

DAR0

DAR1

DAR2

DAR3

TA R 0

TA R 1

TA R 2

TA R 3

TETA R

NMA0

NMA1

NMA2

NMA3

NDA0

NDA1

NDA2

NDA3

N TA

DNM A0

DNM A1

DNM A2

DNM A3

DNDA

TN M A

CN

368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394

395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421

422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446

447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470

471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493

494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516

617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647

648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678

679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707

708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735

736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762

763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789

790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814

815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838

839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861

862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884

885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915

916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944

945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972

973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999

1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026

1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051

1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075

1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098

1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121

1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
1135
1136
1137
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148

1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
1163
1164
1165
1166
1167
1168
1169
1170
1171
1172
1173

1174
1175
1176
1177
1178
1179
1180
1181
1182
1183
1184
1185
1186
1187
1188
1189
1190
1191
1192
1193
1194
1195
1196
1197

1198
1199
1200
1201
1202
1203
1204
1205
1206
1207
1208
1209
1210
1211
1212
1213
1214
1215
1216
1217
1218
1219
1220

1221
1222
1223
1224
1225
1226
1227
1228
1229
1230
1231
1232
1233
1234
1235
1236
1237
1238
1239
1240
1241
1242
1243
DNDA

340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367

585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616

DNM A3

311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339

552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584

DNM A2

280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310

DNM A1

248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279

517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551

DNM A0

215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247

TETA R

181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214

TETN A

111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144

L N A P LPG

MNA2

76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110

HNAP

MNA1

1244
1245
1246
1247
1248
1249
1250
1251
1252
1253
1254
1255
1256
1257
1258
1259
1260
1261
1262
1263
1264
1265
1266

3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40

145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180

K ERO

45

D IESEL

44

U N CO N V ERTE D

43

CN

42

TN M A

41

N TA

40

NDA3

39

NDA2

38

NDA1

37

NDA0

36

NMA3

35

NMA2

34

NMA1

33

NMA0

32

TA R 3

31

TA R 2

30

TA R 1

29

TA R 0

28

D A R3

27

D A R2

26

D A R1

25

D A R0

24

M A R3

23

M A R2

22

M A R1

21

M A R0

20

TN A 3

19

TN A 2

18

TN A 1

17

TN A 0

16

DNA3

15

DNA2

14

DNA1

13

DNA0

12

MNA3

11

MNA2

10

MNA0

MNA1

TB P

37
75

MNA0

38
38

TBP

M BP

39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75

N PA R

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38

D BP

MBP

N PA R

CN

3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
CN

U N CO N V ERTE D

D IESEL

K ERO

HNAP

L N A P LPG

Subclass

DBP

Table 2-2. (Z-CN) matrix representation of VGO in terms of 1266 components/lumps. (Numbers 0 to 3 in the second row represent the number
of methyl branches in that subclass)

35 34 36 34 33 32 31 29 28 27 27 25 24 23 23 35 33 32 31 31 29 28 27 27 25 24 23 23 31 29 28 27 27 25 24 23 23 27 25 24 23 23 23 1266
110 144 180 214 247 279 310 339 367 394 421 446 470 493 516 551 584 616 647 678 707 735 762 789 814 838 861 884 915 944 972 999 1026 1051 1075 1098 1121 1148 1173 1197 1220 1243 1266

27

28

Table 2-3. The number of lumps/pure components in different subclasses.

S. No.

Subclass

No. of lumps up to C40

NPAR

38*

MBP

37

DBP

35

TBP

34

MNA0

36*

MNA1

34

MNA2

33

MNA3

32

DNA0

31*

10

DNA1

29

11

DNA2

28

12

DNA3

27

13

TNA0

27*

14

TNA1

25

15

TNA2

24

16

TNA3

23

17

TETNA

23

18

MAR0

35*

19

MAR1

33

20

MAR2

32

21

MAR3

31

22

DAR0

31*

23

DAR1

29

24

DAR2

28

25

DAR3

27

26

TAR0

27*

27

TAR1

25

28

TAR2

24

29

TAR3

23

30

TETAR

23

29

Table 2-3. Continued.

S. No.

Subclass

No. of lumps up to C40

31

NMA0

31*

32

NMA1

29

33

NMA2

28

34

NMA3

27

35

NDA0

27*

36

NDA1

25

37

NDA2

24

38

NDA3

23

39

NTA

23

40

DNMA0

27*

41

DNMA1

25

42

DNMA2

24

43

DNMA3

23

44

DNDA

23

45

TNMA

23

Total
(*) 310 pure components.

1266

30

CHAPTER III
3.KINETIC MODEL DEVELOPMENT FOR THE HYDROCRACKING
OF VGO
As shown in Chapter II, in the hydrocracking of a complex feedstock like VGO, the
number of elementary steps and reactions occurring on the acidic sites and metal sites of
the catalyst is extremely large. Yet every elementary step and reaction has a finite
contribution towards the final distribution of products. For such complex processes,
conventional kinetic modeling approaches would result in such a large number of
reaction rate coefficients that it would be truly impractical to estimate their values from
the experimental data. The solution of this problem lies in modeling the rate parameters
of these steps in addition to modeling the rate equations. The first step in developing a
methodology for modeling the rate parameters is the exploitation of the fact that in spite
of the excessive number of elementary steps occurring on the acid sites, the types of
elementary steps to which they belong is much smaller. Assigning a unique rate
coefficient to all the elementary steps of one type would be too much of a simplification,
indeed. The differences in the structures and the energy levels of the reactants and the
corresponding activated complexes of the elementary steps of a given type contribute
differently towards their frequency factors and activation energies. To account for the
effect of the structures on the frequency factors, Froment and coworkers11,

12, 36

introduced the concept of single event kinetics. In the hydrocracking of VGO, the
activation energies of the elementary steps are modeled based on the nature (secondary
or tertiary) of the reactant and product carbenium ions. A more rigorous treatment for the
modeling of the activation energies is presented in Chapter VI, dealing with the
hydrocracking of pure paraffinic feedstocks. This approach is based on the application of
the Evans-Polanyi relationship which explicitly accounts for the differences in the
energy levels of the reactants and the products starting from their heats of formation.
The present Chapter discusses the methodology for the modeling of the frequency
factors and the activation energies for acid site steps, the development of the rate

31

expressions for the individual elementary steps and the reaction mechanism and rate
expressions for the saturation of the aromatic species. Based on thermodynamic
principles, a strict lumping scheme is devised (as discussed partly in Chapter II) and the
resulting rate expressions for the global rate of conversion between the lumps are
developed.
3.1. Modeling of the Frequency Factors
Single-event kinetics has been applied to model the frequency factors of the elementary
steps taking place on acidic sites of the catalyst. In the single event kinetics, the effect of
molecular structure on the frequency factor of an elementary step is described with the
help of transition state theory and statistical thermodynamics. The rate coefficient of an
elementary step is given by transition state theory37 as,

k T
k = B
h

S o
H o

exp
exp

R
RT

(3.1)

According to statistical thermodynamics, the entropy of a species is the sum of

transnational, rotational, vibrational and electronic contributions,
o
o
o
o
S o = S Trans
+ SVib
+ S Rot
+ S Elec
o
o
o
= S ExtRot
+ S IntRot
where S Rot

(3.2)
(3.3)

The rotational part of the entropy is composed of an intrinsic term, S o and a contribution
from the symmetry of the molecule, R ln , i.e.,
o
o
S ExtRot
= SExtRot
R ln ( Ext )
o
o
S IntRot
= SIntRot
R ln ( Int )

(3.4)
(3.5)

For racemic mixtures of optically active species, an additional entropy contribution of

R ln(2 n ) due to the mixing of different enantiomers has to be considered, where n is

the number of chiral centers in the molecule.

32

o
o
S Rot
= SRot
R ln Ext n Int
2

o
o
o

where S Rot = S ExtRot + S IntRot

and

(3.6)
(3.7)

Ext Int

= Global Symmetry Number, gl

n
2

(3.8)

The global symmetry number gl quantifies all the symmetry contributions of a species.
Using the above equations, the standard entropy of activation for an elementary step can
be written as

= S

o
Trans

o
Vib

+ S

+ S

o
Elec

glR
o

+ S Rot + R ln

gl

(3.9)

The last term of equation (3.9) gives the difference in standard entropy due to the
symmetry changes in going from reactant to activated complex. Equation (3.9) can also
be written as,
S
where

glR
o

= S + R ln

gl

o
o
o
o
S o = STrans
+ SVib
+ SElec
+ SRot

(3.10)
(3.11)

Using eqns (3.1) and (3.10), the effect of changes in symmetry in going from reactant to
activated complex on the rate coefficient of an elementary step can be factored out. i.e.,

glR
k =
gl

k T
B
h

S o
H o

exp
exp RT

(3.12)

Hence, the frequency factor for this elementary step can be given by,

glR
A=
gl

k BT
S o

exp
R
h

(3.13)

33

Defining the ratio of the global symmetry numbers shown in the first bracket as the

glR
number of single-events, ne,, i.e., ne =

gl

, the frequency factor for a given elementary

step is expressed as

A = ne A

(3.14)

where the single event frequency factor A
is given by

k T
A
= B
h

S o

exp

(3.15)

For all the unimolecular elementary steps, the contribution of the translational motion to
entropy of activation would be zero as the masses of the reactant and the activated
complex in any unimolecular elementary step would be identical. It can be seen from
eqn (3.14) that the difference in symmetry between the reactant and the activated
complex has been factored out by introducing the number of single events, ne. Once the
effect of the symmetry has been factored out from the rotational entropy of activation,
the remaining intrinsic contribution arising due to the difference in the moments of
inertia of the reactant and the activated complex can be assumed to be the same for all
the elementary steps of one type considering the similarity in the structural
transformations occurring in the elementary steps of a given type. The same argument
can be given for the vibrational contribution of the entropy of activation, i.e., similar
changes in the frequencies of vibrational modes of the reactant and the activated
complex in the vicinity of the reaction center (i.e., the positively charged carbon atom
for the transformations on acidic sites), for all the elementary steps belonging to one

type. Similarly, the electronic contribution to the entropy of activation, which depends
on the number of unpaired electrons in the reactant and the activated complex would be
the same for all the elementary steps of one type. Therefore, the single event frequency
factor A
becomes independent of the structures of the reactant and activated complex for
all the different elementary steps of a given type. As a result, the application of the

34

single event concept requires only 1 independent parameter, namely the single event
frequency factor A
, to model the frequency factors of all the elementary steps of one
type. Consequently, for the six types of elementary steps occurring on the acidic sites,
i.e., PCP, acyclic -scission, endocyclic -scission, exocyclic -scission, dealkylation,
and cyclization, 6 single event frequency factor are used as the model parameters. The
values of the number of single events are estimated for all the elementary steps during
the reaction network generation and stored with the respective elementary steps.

3.2. Modeling of the Activation Energies

The activation energies for the acid site elementary steps are modeled based on the
nature of the reactant and the product carbenium ions. It is considered that the difference
in the energy levels of the reactant and product carbenium ion depends only on their
nature, i.e., secondary or tertiary. In other words, energy levels are assumed to depend
only on the number of -carbon atoms relative to the positively charged carbon atom, as
the secondary and tertiary carbenium ions have two and three -carbon atoms
respectively. Therefore, the effect of the charge delocalization on the energy levels of
carbenium ions because of the differences in the number of -carbon atoms has been
ignored in the hydrocracking model for VGO. With this assumption, four distinct
activation energies, namely, E( s ;s ) , E( s ;t ) , E(t ;s ) and E( t ;t ) are required for the four subtypes
of a given type of elementary steps, leading to 24 parameters for 6 different types of
steps.
A reduction in the total number of parameters required for modeling of the activation
energies is carried out by observing the similarity in structural transformation in
different types of elementary steps, viz., acyclic -scission, exocyclic -scission and
endocyclic -scission. It can be seen from Figure 2-2 that in all the acyclic -scission
steps, i.e., when -scission takes place in the paraffin or in the side chain of the ring
species, the double bond would be formed in the paraffinic fragment or in the side chain
of the naphthenic fragment. However, in all the exocyclic -scission steps, the double
bond would be formed in the naphthenic ring. The presence of the double bond in the

35

naphthenic ring causes an additional ring strain leading to average heat of reaction for
exocyclic steps of any given subtype, say (s;s), to be higher than the average heat of
reaction of the corresponding subtype of acyclic -scission steps, provided that other
differences are small. Considering the application of the Evans-Polanyi relationship [eqn
(6.15)], the activation energies for the exocyclic steps would be higher than those of
acyclic steps by an amount of H exo H acyc assuming the intrinsic activation
energy Eo for these two types of steps is same. Therefore, the activation energies for the
four subtypes of exocyclic steps have been obtained from activation energies of
corresponding acyclic steps by introducing only one additional parameter, Eexo ( m;n )
which accounts for the increase in the activation energies, i.e., Eexo ( m;n ) Eacyc ( m ;n ) ; m
and n represent secondary or tertiary.
It has been observed that the rate of the naphthenic ring opening via endocyclic scission is quite low compared to the rate of -scission in the paraffinic species. One
explanation for this behavior, provided by Eagan et al.20, claims that the noncyclic
carbenium ion formed by -scission of a naphthenic ion has a high tendency to return to
the cyclic form because of the favorable configuration of the olefinic carbenium ions, as
shown in Figure 3-1.

Figure 3-1. Favorable configuration of the cyclization of olefinic carbenium ion formed by endocyclic scission.

36

An approach similar to exocyclic -scission has been followed for modeling the
activation energies of the endocyclic -scission steps. Only one parameter Eendo ( m ;n )
has been introduced to account for the difference in the activation energies of a given
subtype of endocyclic steps from the corresponding subtype of acyclic steps, i.e.,

Eendo ( m ;n ) Eacyc ( m;n ) .

Cyclization steps are the reverse of endocyclic -scission. However, in the absence of
the equilibrium coefficients of the endocyclic steps at the surface of the catalyst, four
different activation energies Ecyc ( s ;s ) , Ecyc ( s ;t ) , Ecyc (t ;s ) and Ecyc ( t ;t ) are used for the four
subtypes of cyclization steps.
The reacting carbenium ions (or the benzenium ions) for the dealkylation of aromatics
are the ring protonated aromatic species. Because of the much larger delocalization of
positive charge on the aromatic ring, the energy levels of the ring protonated aromatic
species can not be distinguished based on their secondary or tertiary nature. Therefore,
depending on the nature of the paraffinic carbenium ions formed in the dealkylation
steps, only two activation energies Edealk ( s ) and Edealk (t ) are used for all the dealkylation
steps.

3.3. Development of Rate Equations

Figure 3-2 depicts physical and chemical phenomena occurring inside the catalyst pellet
during the hydrocracking of VGO. Paraffins, naphthenes and aromatic species (PNA) are
sorbed from the liquid phase to the zeolite pores. Depending upon the difference in the
reactive intermediate involved, all the elementary steps occurring on the acid sites of the
catalyst are classified into three groups:
(a) PCP, acyclic -scission, exocyclic -scission, endocyclic -scission: these steps
proceed through olefinic intermediates which are protonated into the carbenium ions for
the elementary step to occur.

37

(b) Cyclization: proceeds through diolefinic intermediates, which are protonated into
olefinic carbenium ions before the cyclization step.
(c) Dealkylation: for these elementary steps to occur, the aromatic species sorbed inside
the zeolite pores are directly protonated on the ring.

Figure 3-2. Physical and chemical phenomena inside the catalyst pellet.

38

Different rate expressions are developed for the three groups of acid site steps as
explained below. Modeling of the rate coefficients, reaction mechanism and the rate
expressions for the aromatic ring saturation on the metal sites will be dealt in the next
section.

3.3.1. Group (a)- PCP, Acyclic/Exo/Endo -Scission

A paraffinic, naphthenic or aromatic (and naphtheno aromatic) species Si is sorbed from
the liquid phase into the zeolite pores.

Siliq  Sisorbed

(3.16)

The concentration of species Si inside the pore is obtained using Langmuir type sorption
isotherm

[ Si ] =

H Si CSiliq
DL

(3.17)

in which Henrys coefficient for the sorption of species Si is given by

H Si = K L, Si csat , Si and DL = 1 + K L , Si CSiliq

(3.18)

For group (a) elementary steps, the species Sisorbed is chemisorbed on the metal sites of
m
the catalyst as Sim and dehydrogenated into the corresponding olefinic species, SOij
.

Naphthenic species can dehydrogenate either in the side chain or in the ring. Aromatic
species can dehydrogenate in the side chain only.
m
Si  SOij
+ H 2liq

(3.19)

It is assumed that the catalyst has sufficient metal/acid activity to bring these
hydrogenation/dehydrogenation steps to quasi-equilibrium. The resulting olefinic species
are desorbed from the metal sites into the zeolite pores. Assuming that the Henrys

39

coefficients for the species Si and the corresponding olefinic species SOij are the same,
the concentration of the olefinic species in the zeolite pores is given by

SOij =

liq
H Si CSiliq K DH
,ij

(3.20)

DLCHliq2

these species are protonated on the acidic sites into carbenium ions, i.e.,
sorbed
+
SOij
+ H +  S Rik

(3.21)

Protonation/deprotonation steps have been shown to be potentially much faster than the
other steps on the acidic sites and therefore are assumed to be in quasi-equilibrium. The
+
is therefore obtained as,
concentration of the surface carbenium ions S Rik
+
S Rik
= K pr ( S  S + ) SOij H +
Oij
Rik

(3.22)

It should be noted that one particular olefin can produce a maximum of two carbenium
ions depending upon the location of the double bond. If the double bond is in the
terminal position, one of the produced carbenium ion will be primary and will not be
considered in the reaction network. The index k in equation (3.21) is used to describe all
the possible carbenium ions that can be produced by protonation of all the olefins
obtained from Si on dehydrogenation. The concentration of the vacant acidic sites

H + in eqn (3.22) can be eliminated by the acid site balance as,

+
H + + S Rik
= Ct

(3.23)

ik

Using eqn. (3.22), the concentration of the vacant acid site is expressed as,

H + =
1+

K
ij

Ct
+
pr ( SOij  S Rik

S
) Oij

(3.24)

In the hydrocracking of pure paraffins, it has been observed that the surface
concentration of the carbenium ions is negligible as compared to the total acid site

40

H + Ct , because the concentrations of the olefinic

concentration, so that

intermediates is several orders of magnitude smaller then the paraffinic species. In the
case of VGO hydrocracking this would not be true. The presence of significant amount
of aromatics in typical VGO feedstocks and their preferential protonation on the acid
sites would require the use of equation (3.24) without neglecting the second term in the
denominator. In previous work22 on the modeling of VGO hydrocracking, the influence
of the competitive chemisorption of aromatics on the acid sites was not considered.
In the current work, the values of the protonation equilibrium coefficients have been
estimated from the gas phase proton affinities of different species and statistical
thermodynamics to estimate the value of the denominator in eqn (3.24). This has been
described in section 3.7
Combining eqns. (3.22) and (3.24),

K pr ( S  S + ) SOij Ct
Oij
Rik
+
S Rik
=
1 + K pr ( S  S + ) SOij
Oij
Rik

(3.25)

ij

+
As can be seen from eqn (3.21), the carbenium ion S Rik
can be formed from more than

one olefins. Therefore, in the kinetic model, the concentration of the carbenium ion has
been calculated from an average over all the possible olefins [not shown in eqn (3.25)].
+
Based on the concentration of the carbenium ion S Rik
, the rate of an elementary step
+
+
S Rik
S Ruv
of group (a) is given by,

r ( m;n ) = ne ,ikuv k
( m;n )

In which DA = 1 + K pr ( S
ij

+
Oij  S Rik

K pr ( S

+
Oij  S Rik

DA

SOij Ct

(3.26)

SOij and represents the type of the elementary step

belonging to group (a). m and n represent the nature of the reactant and product

41

carbenium ions. The concentration of olefin in eqn (3.26) is eliminated in terms of the
observable concentration of species Si by using eqn (3.20),
liq
r ( m;n ) = ne ,ikuv k
( m;n ) K DH
,ij K pr ( S

+
Oij  S Rik

C
) t

H Si CSiliq
DL DACHliq2

(3.27)

3.3.2. Group (b)-Cyclization

As mentioned earlier, cyclization steps are the reverse of the endocyclic -scission. To
model the rate of cyclization steps, a reaction mechanism is proposed which proceeds
through the sequential formation of monoolefin and diolefin species followed by the
protonation on the acidic sites to produce olefinic carbenium ions. Cyclization of these
carbenium ions result in the formation of an extra naphthenic ring as shown in Figure
3-3.

-H2
SOij

Si
-H2

S DOij

SOij

+H+

+
+
S Rik

S DOij
+
+
S Rik

+
S Ruv

Figure 3-3. Mechanism for the cyclization steps.

42

Sorbed concentrations of olefin SOij and di-olefin S DOij can be obtained as

SOij =

liq
liq
K DH
( Si  SOij ) H Si CSi

S DOij =

(3.28)

DL CHliq2
liq
liq
liq
K DH
( Si  SOij ) K DH ( SOij  S DOij ) H Si CSi

(3.29)

DLCHliq2 2

With the protonation/deprotonation equilibrium coefficient of S DOij , the concentration of

+
surface carbenium ion S Rik
is given by,

H Si CSiliq
+
liq
liq
S Rik
= K DH
K
K
C
+
( Si  SOij ) DH ( SOij  S DOij )
pr ( S DOij  S Rik
) t
DL DACHliq2 2

(3.30)

+
is now given by
The rate of cyclization of carbenium ion S Rik

liq
liq
rcyc ( m ;n ) = ne,ikuv k
cyc ( m;n ) K DH
( Si  SOij ) K DH ( SOij  S DOij ) K pr ( S

+
DOij  S Rik

C
) t

H Si CSiliq
DL DACHliq2 2

(3.31)

3.3.3. Group (c)- Dealkylation of Aromatics

For the dealkylation of aromatics the reactive intermediate is formed by the protonation
of the aromatic ring itself, and therefore the dehydrogenation step to yield the olefinic
species, as is the case in paraffins and naphthenes, is not involved. The surface
concentration of the ring protonated ions is given in terms of the liquid phase
concentration of the aromatic specie S Ai as,

+
Aik

H S Ai CSliqAi

= K pr ( S  S + )Ct
Ai
Aik
DL DA

(3.32)

Similar to the rate expressions of the other groups, the rate of dealkylation steps is given
by,

rdealk ( m ) = ne,ikuv k
dealk ( m ) K pr ( S

+
Ai  S Aik

C
) t

H S Ai CSliqAi
DL DA

(3.33)

43

where m is the nature of the product paraffinic carbenium ion.

3.3.4. Saturation of Aromatics
Several experimental studies and review articles have been published on the
hydrogenation of aromatics involving model compounds like benzene, toluene, alkyl
benzenes, tetraline, naphthalene, biphenyl, phenanthrene, pyrene, chrysene, flourene and
fluoranthene etc.38-58. In most of the studies, especially those involving polyaromatic
hydrocarbons, large reaction networks have been proposed to represent the ring
saturation kinetics. In the current work, aromatic species from one to four rings are
considered in the feedstock. The following sections discuss the reaction mechanism
considered for the saturation of mono and multi-ring aromatic hydrocarbons.
3.3.4.1. Saturation of Monoaromatics
Kinetics of the hydrogenation of benzene and substituted benzene has been studied
extensively both on supported group noble metals (Pt, Pd) and on metal sulfides (MoS2,
WS2, CoMoS/Al2O3, NiMoS/Al2O3, and NiWS/Al2O3) catalysts59. Modeling for the
hydrogenation of benzene has been studied considering stepwise addition of hydrogen,
either molecularly or atomically. Aben et al.60 proposed the Langmuir type rate
expression for the hydrogenation of benzene on supported Pt, Pd, and Ni catalysts in a
flow system as,

= ko

bPH 2

1 + bPH 2

(3.34)

where b is the adsorption coefficient for molecular hydrogen and ko is the rate
coefficient. The form of the rate expression implies that hydrogen is molecularly
adsorbed and the reaction order in benzene is zero. An activation energy of
approximately 14 kcal/mol was found for all the three catalysts. In other studies61, in gas
phase hydrogenation of benzene on Ni/SiO2 catalyst the reaction order for benzene has
been shown to vary from 0.5 at 25 oC to 2-3 at 200 oC. The reaction order of hydrogen
on supported Pd catalyst has been shown to increase from 0.5 to about 4 as the

44

temperature is increased from 80 oC to 300 oC, whereas the reaction order for benzene
increased from 0 to 0.8 in the same temperature range.62,

63

Lin and Vannice39,

40, 42

proposed a mechanism in which hydrogen is added atomically and the addition of the
first hydrogen is considered as rate determining because this step breaks the resonance
structure of the benzene molecule. They also assumed that hydrogen and benzene are
adsorbed on different catalytic sites and used this mechanism to fit the benzene
hydrogenation data taken on a supported Pt catalyst. Subsequent use of this mechanism
for the liquid phase hydrogenation of benzene by Singh and Vannice57, 58, 64 also led to a
good fit to the experimental data. As the hydrocracking reactors operate in the threephase regime, the mechanism proposed by Lin et al. was first attempted in the current
model with minor simplifications by ignoring the presence of dehydrogenated benzene
species64. This mechanism considers the following sequence of steps,

H 2 + 2*  2 H *
S Ai + m  S Ai m
S Ai m + H * S A1i m + * ( RDS )
S A1i m + H *  S A 2i m + *
S A 2 i m + H *  S A 3i m + *

(3.35)

S A 3i m + H *  S A 4 i m + *
S A 4 i m + H *  S A5 i m + *
S A5i m + H *  S Ni m + *
S Ni m  S Ni + m
where S Ai and S Ni represent the aromatic and naphthenic species and S A1i to S A5i are
the reaction intermediates. Hydrogen is dissociatively chemisorbed at the * sites and the
aromatic molecule is chemisorbed at the m sites. The net rate of hydrogenation of
aromatics using this mechanism is given by
khyd K C , S Ai cm c* K C , H 2
rhyd =
C

liq 5 2
H2

liq liq3 CSliqNi

CS Ai CH 2 liq
K eqm

liq
liq
liq
1 + K C , S Ai CS Ai + K C , S Ni CS Ni + K C , SPi CSPi
i

(3.36)

45

In the above equation, adsorption terms in the denominator corresponding to H2 and

intermediate species S A1i through S A5i have been neglected. The last term in the
denominator accounts for the occupancy of the m sites by the paraffinic species present
in the reaction mixture. The values of the gas phase hydrogenation equilibrium
gas
coefficient K eqm
have been estimated using the Bensons group contribution method65 to
liq
by using the molar
get the liquid phase hydrogenation equilibrium coefficient K eqm

concentration of the liquid phase and the vapor-liquid equilibrium partition coefficients
K VLE of hydrogen, aromatic and the naphthenic species as,

liq
eqm

liq
Ctotal
gas
= K eqm
VLE
KH
2

K SVLE
Ni
VLE
KS
Ai

(3.37)

The VLE partition coefficients have been estimated using the Peng-Robinson equation
of state66.
Reactor simulations for a given VGO feedstock were performed at different reactor
pressures using eqn (3.36) and the effect of pressure on the aromatics content in the
product was compared with the experimental data of Dufresne et al.67 These authors
studied the effect of hydrogen pressure on the aromatics content in the hydrotreatment of
Arabian Light VGO containing 47 wt % aromatics. When the hydrogen pressure is
increased from 70 bar to 140 bar, the total aromatics in the products decreased from 31
wt% to 9.6 wt% at a given space time. Table 3-1 shows the effect of pressure on the
amount of individual aromatic classes.

46

Table 3-1. Effect of hydrogen pressure on the aromatics content in the hydrotreatment of Arabian
Light VGO67.

Feed VGO

Products at 70 Bar

Products at 140 Bar

Total aromatics (wt%)

47

31

9.6

Monoaromatics (wt%)

24

21

8.2

Diaromatics (wt%)

11

6.1

1.0

Triaromatics (wt%)

12

4.4

0.4

It was found from the reactor simulations that the decrease in the yield of aromatics with
the increase in reactor pressure was much smaller compared to the above experimental
data when using eqn (3.36) for modeling the hydrogenation of aromatics. This was
attributed to a low effective order of hydrogen of 0.5 based on this rate expression. A
different rate expression was tried based on the more recent work of Melis et al.68 They
proposed a model for the interpretation of hydrogenation of aromatic compounds during
hydroprocessing of gas oil considering the adsorption of hydrogen and aromatic species
on different sites,

rj =

k j x j PHn2 k j 1 xk

(1 + K j x j ) 1 + K H2 PH2

(3.38)

where x j is the weight fraction of the ith lump, k j and k j 1 are the forward and reverse
rate coefficients constrained by the equilibrium coefficient, K j is the adsorption
coefficient and n is the number of moles of H2 molecules involved in the reaction. This
model was fitted with experimental data at various pressures using a commercial
NiMo/Al2O3 catalyst. The model was also tested for different feedstocks, temperature
and LHSV and showed good agreement. However, in the absence of the kinetic
parameters in their published work, the importance of the different terms in the
denominator could not be established. A similar rate expression has been used by Korre

47

et al.48,

50

for the hydrogenation of polyaromatic hydrocarbons except that the

chemisorption of hydrogen is not considered in the rate expression,

rj =

k j K j (C j PHn2 Ck / K eqm
jk )

(3.39)

(1 + K C )
j

in which C j and Ck are the molar concentrations of the aromatic and naphthenic species
respectively. They established a relationship between molecular structures and
hydrogenation reactivities in heavy oil hydroprocessing via elucidation of the controlling
reaction pathways and kinetics of one-, two- three- and four fused ring aromatics. They
studied the hydrogenation reaction of o-xylene, tetralin, naphthalene, phenanthrene,
anthracene, pyrene, and chrysene and their multicomponent mixtures in a cyclohexane
solvent using presulfided CoMo/Al2O3 catalyst and found good agreement with the
experimental data. The same rate expression has been used in the current model with the
hydrogen partial pressure replaced by the liquid phase molar concentration of hydrogen
and liquid phase concentration of hydrocarbons by their sorbed concentrations to
account for the sorption of hydrocarbons in the zeolite pores. The sorbed concentrations
are then eliminated in terms for the observable liquid phase concentrations using eqn
(3.17) so that,

rhyd =

liq liq3 CSliqNi

comp
khyd
K
H
C C liq
,3 H 2 C , S Ai
S Ai
S Ai H 2 K eqm

liq
liq
liq
DL + K C , S Ai H S Ai CS Ai + K C , SNi H SNi CS Ni + K C , SPi H SPi CSPi
i

(3.40)

comp
In the above equation khyd
,3 H 2 is the composite rate coefficient for the hydrogenation of

monoaromatics given by the product of the rate coefficient of the surface controlling
reaction and the total concentration of the active metal sites, i.e., khyd ,3 H 2 cm . The
subscript 3H2 signifies the consumption of three moles of hydrogen consumed per mole
of aromatics.

48

3.3.4.2. Saturation of Polyaromatics

Korre et al.48,

50

estimated 45 hydrogenation rate parameters and the same number of

equilibrium ratios for 36 aromatic compounds from experimental data using the rate
expression in eqn (3.39). These values were used in the evaluation of five adsorption
parameters for ring species as a function of the number of aromatic rings and naphthenic
rings. Based on close scrutiny of the parameters and experimental observations, they
found the following trends in the hydrogenation of polyaromatics:
a) Hydrogenation of polynuclear aromatics proceeds sequentially in a ring by ring
manner. No partially hydrogenated compounds such as di- or hexahydro naphthalenes
were detected.
b) Hydrogenation reactivity increased with the number of aromatic rings. The
hydrogenation of single-aromatic ring moieties was the slowest, that of isolated two-ring
aromatics intermediate, and the hydrogenation of the middle of the three fused aromatic
ring was fastest.
c) In groups with the same number of fused aromatic rings, hydrogenation reactivity
increased with the presence of alkyl substituents and/or naphthenic rings.
Based on the observations of Korre et al. and the structures of the polyaromatic species
considered during the reaction network generation in the current work, only two
hydrogenation rate coefficients have been used for all the possible aromatic
hydrogenation reactions. One rate coefficient is used for all the reactions in which three
moles of hydrogen are consumed and another rate coefficient is used for all the reactions
in which two moles of hydrogen are consumed.

49

Table 3-2 shows all the possible hydrogenation modes for one to four ring aromatic
species that are considered in the reaction network of aromatics. The H in the ring
represents the ring to be hydrogenated. All the molecular classes of subgroup 1 have
been assumed to have the same hydrogenation rate coefficient irrespective of the length
and number of alkyl substituents on these species. The same is true for the classes of
subgroup 2. For the non-isothermal reactor model, two activation energies and two preexponential factors are required in the kinetic model. The difference in the reactivities of
the molecules of each subgroup is accounted for by the differences in their sorption
equilibrium coefficients in the zeolite pores and their chemisorption equilibrium
coefficients on the metal sites. As all the classes of subgroup 1 require 3 moles of
hydrogen per mole of hydrocarbon, the rate expression for the hydrogenation of all these
aromatics is given by eqn (3.40). For the molecular classes of subgroup 2 where only
two moles of hydrogen are consumed in hydrogenation, the rate equation has been
modified as,
k
rhyd =

comp
hyd ,2 H 2

liq liq2 CSliqNi

K C , S Ai H S Ai CS Ai CH 2 liq

K eqm

liq
liq
liq
DL + K C , S Ai H S Ai CS Ai + K C , SNi H SNi CS Ni + K C , SPi H SPi CSPi
i

(3.41)

comp
comp
The composite frequency factors Ahyd
,3 H 2 and Ahyd ,2 H 2 and the activation energies

Ehyd ,3 H 2 and Ehyd ,2 H 2 for the two subgroups have been estimated from the experimental
data.

50

Table 3-2. Classification of aromatics based on the number of moles of hydrogen required for
hydrogenation. (The symbol H in a ring represents the ring to be hydrogenated).
Subgroup-1: Each molecule requires three moles of hydrogen
H
H

Subgroup-2: Each molecule requires two moles of hydrogen

51

3.4. Competitive Chemisorption on the Metal Sites

It is well established that the chemisorption coefficient for aromatics increases with the
increase in the number of aromatic rings. This could be attributed to the increased acidbase interaction of the catalyst and aromatics. It has also been observed that the
chemisorption coefficient increases with the increase in the number of hydrogenated
rings as well, but to a lesser extent as compared to aromatic rings69.

Korre et al.46, 50

regressed the chemisorption coefficients from their extensive experimental data and
established temperature independent correlations for chemisorption coefficients on a
CoMo/Alumina catalyst as a function of the number of aromatic rings and the number of
saturated rings in the ring hydrocarbons. They found the following correlation in which
the number of aromatic rings and number of saturated rings are used as the indices,
chem
const

ln K C = p

chem
chem
paro
_ rings N AR + psat _ rings N SR
+

RT

(3.42)

chem
chem
chem
The values of pconst
, paro
_ rings and psat _ rings obtained by them were 1.324, 0.887 and

0.123 respectively for their operating conditions and catalyst. The same correlation has
chem
chem
chem
been used in the current model considering pconst
, paro
_ rings and psat _ rings to be adjustable

parameters because of the differences in the catalyst and operating conditions. As seen
from the above correlation, the effect of the length and number of side chains has not
been considered, as the aromatics studied in this group were mostly bare rings, except
for the xylenes. In a typical VGO feedstock, however, the ring species contain long alkyl
side chains. No information is currently available on the effect of long side chains on the
chemisorption coefficient of ring species. It has been considered negligible in the current
model. The chemisorption equilibrium coefficient for paraffins is given using a single
parameter by the following equation

ln K C = p chem
paraf

(3.43)

52

chem
chem
chem
chem
The dimensionless parameters pconst
, paro
_ rings , psat _ rings and p paraf would give the

chemisorption equilibrium coefficient in kgcat/kmol when calculated using eqns (3.42)

and (3.43).
3.5. Development of Global Rate Expressions for Acid Site Steps
The rate expression for the conversion of a particular carbenium ion on the acidic sites
has been developed in Sections 3.3.1 to 3.3.3 and for the hydrogenation of a given
aromatic molecule, the form of the rate expression employed in the kinetic model has
been discussed in Section 3.3.4. These rate expressions involve the concentration of a
single molecular species present in the VGO feedstock. Since the molecular level
composition of complex feedstocks like VGO is not currently accessible using the
present day analytical techniques, a certain amount of relumping of individual
components is required in the model. Unlike the conventional lumping approaches, this
lumping strategy is based on strict thermodynamic principles, grouping only the
isomeric species of a homologous series which attain a thermodynamic equilibrium
through fast interconversion*. A short review on the state-of-the-art characterization
techniques for complex petroleum mixtures, a detailed explanation of the lumping
strategy employed, and the analysis/composition of the VGO feedstock required as input
in the current model has been presented in Section 2.4. It was shown that a total of 1266
pure components/lumps were required for a VGO containing the species up to carbon
number 40. All the isomers in a particular lump belong to one molecular class and have
the same number of methyl branches. Grouping of the components based on this
approach allows the determination of the rates of individual elementary steps using the
rate expressions developed in Section 3.3. By using the exhaustive computer generated
reaction network of elementary steps, the contribution of each individual elementary step
is summed up to estimate the global rate of conversion of a lump g into another lump

To avoid the confusion in the definition of lump with the less rigorous models, the word lump will be
replaced by group of isomers or GOI in the future work and publications based on this dissertation.

53

h which are required in the development of continuity equations. This is shown

schematically in Figure 3-4 for the case of paraffinic components.
Therefore, the global rate of conversion of a lump g into another lump h through

(m; n) type of elementary steps of group (a) on the acidic sites is given as follows by
using eqn (3.27),
( g ;h )
( m;n )

H Si CSiliq
Ct
+
Oij  S Rik )
DL DACHliq2

liq
= ne ,ikuv k
( m;n ) K DH
,ij K pr ( S

(3.44)

where the summation is carried out over all the (m; n) type of elementary steps in
which carbenium ions of nature m belonging to lump g are converted to carbenium ions
of nature n belonging to lump h. In eqn (3.44), the concentration of species Si is
obtained from its equilibrium distribution in lump g as,
liq
CSiliq = ySieqmCSg

(3.45)

The same value of Henrys coefficient is used for all the isomers of a lump, so that the
denominator term for sorption in the zeolite pores is given by

Nlumps

liq
DL = 1 + K L , g CSg

g =1

(3.46)

54

Figure 3-4. Illustration of the lumping of components for the estimation of the global rate of conversion
between lumps.

Similarly, for the purpose of estimating the value of DA , the same value of equilibrium
coefficient is considered for the protonation of all olefinic species originating from the
isomers of a lump. For the aromatic lumps, the protonation is assumed to occur directly
on the aromatic ring,
N sat _ lumps

DA = 1 +

g =1

K pr ,Og SOg +

N sat _ lumps + N aro _ lumps

g = N sat _ lumps +1

K pr , Ag S Ag

(3.47)

55

where SOg is the sorbed concentration of the olefins formed from the isomers of lump

g in the case of paraffinic and naphthenic lumps and S Ag is the sorbed concentration of
aromatic lump g.
liq
The liquid phase dehydrogenation equilibrium coefficients K DH
,ij are obtained from the
gas
gas phase dehydrogenation equilibrium coefficients K DH
,ij by,

liq
Ctotal
liq
gas
K DH
=
K

,ij
DH ,ij
K HVLE
2

(3.48)

which in turn are estimated using the Bensons group contribution method. The
protonation equilibrium coefficient K pr ( S

+
Oij  S Rik

can be expressed in terms of the single

event protonation equilibrium coefficient by taking out the symmetry contributions as,

K pr ( S

+
Oij  S Rik

In previous work11,

22, 70

glSOij
= S+
glRik

K
pr ( S  S + )
Oij
Rik

(3.49)

on single event kinetic modeling of hydrocracking, the

protonation equilibrium coefficient was estimated by involving a reference olefin SOr so

+
that the protonation step SOij  S Rik
was written as,

+
SOij  SOr  S Rik

(3.50)

The reference olefin was chosen from the same class as the parent species Si in such a
way so that the double bond is located between a secondary and tertiary carbon atom. It
was further assumed that the protonation equilibrium coefficient of the reference olefin
can have only two values depending on the nature of the protonated carbenium ion.
Therefore, from eqn (3.50)

K
pr ( S

+
Oij  S Rik

= K
isom ( SOij  SOr ) K
pr ( SOr  m )

(3.51)

56

where m can be s or t. The isomerization equilibrium coefficient was calculated from the
Bensons group contribution method. It has been shown in the current work that the
involvement of the reference olefin is not necessary for obtaining the protonation
equilibrium coefficients. Therefore, irrespective of the olefinic species SOij , only two
single event protonation equilibrium coefficients, K
pr ( SOij  s ) and K
pr ( SOij t ) are required
in the model. This will be discussed in more details in section 3.6.
The final rate equation is obtained by combining eqns (3.44), (3.45), (3.48) and (3.49):
( g ;h )
( m;n )

glSOij
= ne ,ikuv S +
glRik

liq

H g CSg
gas
eqm



K DH
,ij ySi k ( m ; n ) K pr ( SOij  m )Ct

DL DACHliq2

liq
Ctotal
VLE
K H2

(3.52)

The above equation can be expressed as,

R( g( m;h;)n ) = LC( g( m;h;)n ) k
comp

( m; n )

liq
liq
Ctotal
H g CSg

DL DACHliq2 K HVLE
2

(3.53)

where LC( g( m;h;)n ) is called the lumping coefficient of formation (LCF) of lump h from
lump g and is given by,
( g ;h )
( m; n )

LC

glSOij
= ne ,ikuv S +
glRik

gas
eqm
K DH
,ij ySi

(3.54)

The lumping coefficients are functions of temperature only and do not depend on the
feedstock composition. The values of lumping coefficients are calculated at a
temperature interval of 5 oC and are utilized during the integration of the ODEs in the
non-isothermal reactor simulations. Using the central difference approach, the lumping
coefficients evaluated at temperature t are used for reactor temperature from t

t+

t
to
2

t
, where t is 5 oC. As the total number of pure components/lumps is quite large,
2

viz. 1266, and each one can be consumed through a maximum of 26 subtypes of

57

elementary steps/reactions (obtained by summing over

(m; n) , cyclization,

dealkylation and ring hydrogenation/dehydrogenation), the LCF is a three dimensional

matrix of size 1266 x 1266 x 26, and therefore requires a large amount of hard disk
space at each temperature. To save hard disk space and the CPU time in reading the
large files during the non-isothermal reactor simulations, only the non-zero elements of
LCF are stored along with their indices resulting in more than 99% saving in disk space.
To find the cumulative lumping coefficient for the consumption of a lump g to all
possible lumps through a given type of elementary step, the elements of the LCF matrix
are summed up as follows to generate a two dimensional matrix called the lumping
coefficient of consumption (LCC) of size 1266 x 26,
LCCg( m;n ) = LC( g( m;h;)n )

(3.55)

This matrix is stored along with LCF for each temperature interval to avoid the repetitive
summation during the integration of ODEs for computational efficiency. The lumping
coefficient for the consumption of different classes through pcp(s;s) elementary steps at
350 oC are plotted in Figure 3-5, showing the increase in the lumping coefficients with
the carbon number in a homologous series. As mentioned earlier (Section 2.4), the four
ring species are not divided based on the number of branches and thus have only one
curve, shown in blue.
k
comp
( m ; n ) in eqn (3.54) is the composite single event rate coefficient given by,




k
comp
( m ; n ) = k ( m ; n ) K pr ( SOij  m ) Ct

(3.56)

58

paraffins

mna

dna

tna

0.4

0.5

0.5

0.5

0.2

20

40

tetna
1

0.1

0.5

20

40

40

tetar

dnma

20

40

20

20

20

40

0.2

0.1

20

40

nta
0.2

40

20

40

tnma

0.2

0.1

0.1

20

0.4

0.2

40

nda

dnda

40

40

0.4

0.2

20

20

0.2

0.4

tar

0.5

0.2

40

40

dar

0.5

20

20

nma

0.4

mar

0.2

20

40

unbranched
mono-branched
di-branched
tri-branched
20

40

Figure 3-5. Lumping coefficients for the consumption of different classes through pcp(s;s) steps. (The xaxis is the carbon number).

Using the thermodynamic constraints on the rate coefficients of the forward and reverse
steps11, it can be shown that the composite rate coefficients for the (s;t) and (t;s)
comp

comp
isomerization steps are identical, i.e., k
isom
( s ;t ) = k isom ( t ; s ) .

From eqn (3.56), assuming that the intrinsic entropy of protonation does not depend on
the nature of the product carbenium ion, the composite single event frequency factor and
the composite activation energy are expressed as

S pr
A
comp = A
Ct exp
R

(3.57)

59

Ecomp
( m ; n ) = E ( m ; n ) + H pr ( SOij  m )

(3.58)

These parameters are estimated from the experimental data. Rate expressions similar to
eqn (3.53) are derived for the elementary steps of group (b) and (c) also.

3.6. Elimination of the Need of Reference Olefins

The concept of reference olefins was introduced by Baltanas et al.11 to model the rate
coefficients for the deprotonation steps as the protonation/deprotonation steps were not
considered to be at pseudo-equilibrium. Svoboda et al.16 later showed that the
protonation/deprotonation steps are potentially much faster than other acid sites
elementary steps and reach a pseudo-equilibrium. Based on the assumed structure of the
activated complex of the protonation step, it was hypothesized that the single event
protonation equilibrium coefficient K
pr ( S

+
Oij  S Rik

depends on the structure of the olefin

and the nature of the product carbenium ion but not on its structure. The dependency of
the structure of the reacting olefin was factored out using a reference olefin so that the
+
was given by eqn (3.51)
single event protonation equilibrium coefficient for SOij  S Rik

repeated here for reference,

K
pr ( S

+
Oij  S Rik

= K
isom ( SOij  SOr ) K
pr ( SOr  m )

(3.51)

In the proposed approach it is hypothesized based on to the similarities in the structures

that the differences in the energies and intrinsic entropies of the reacting olefins and
corresponding carbenium ions in different protonation steps are constant, and therefore,
the single event protonation equilibrium coefficient depends only on the nature of the
product carbenium ion, leading to only two values. In other words, this hypothesis
accounts for the dependence of the energies of carbenium ions on their nature as well as
their structures, however, the latter effect is cancelled out because of the similar olefin
structure. In this work a group contribution method has been developed to estimate the
heats of formation of the paraffinic carbenium ions in gas phase. By calculating the heats
of formation of the reacting olefins using the Bensons group contribution method, the

60

lumping coefficients are calculated using the true values of the protonation equilibrium
coefficients and compared with the lumping coefficients calculated based on the old and
new hypothesis, proving the superiority of the latter.
To make this comparison, the effect of the structure of the carbenium ion on the
protonation equilibrium coefficient is factored out by introducing a reference carbenium
+
ion S Rr
corresponding to the given reference olefin SOr so that,

K
pr ( S

+
Oij  S Rik

= K
isom ( SOij  SOr ) K
pr ( S

+
Or  S Rr

K
isom ( S +  S +
Rr

Rik

(3.59)

The reference carbenium ion can be secondary or tertiary, depending on the nature of the
original carbenium ion Rik+ . By comparing eqns (3.51) and (3.59) it can be observed that
in the previous approach, the isomerization equilibrium coefficients from the reference
carbenium ion to the actual product carbenium ion, K
isom ( S +  S + ) was assumed to be
Rr

Rik

unity. On the other hand, in the proposed approach the product K
isom ( SOij  SOr ) K
isom ( S +  S +
Rr

Rik

is assumed to be unity, as only two single event protonation equilibrium coefficients are
required. It can be inferred from the comparisons shown in Figure 3-6 that these two
equilibrium coefficients are correlated inversely so that their product has less deviation
from unity compared to the deviation from unity of K
isom ( S

+
Rr  S Rik

alone.

Figure 3-6 shows the lumping coefficients for the consumption of paraffins through scission steps for three different cases:

3.6.1. Lumping Coefficients with Reference Olefins (Case 1)

In this case, it is assumed that K
isom ( S +  S +
Rr

Rik

is unity. Therefore, these are the LCs based

on the old hypothesis. These plots are named as wro i.e., with reference olefins.

61

3.6.2. Lumping Coefficients without Reference Olefins (Case 2)

In this case, it is assumed that the product K
isom ( SOij  SOr ) K
isom ( S +  S +
Rr

Rik

is unity. These plots

are names as nro i.e., no reference olefin as they have been estimated without using
the reference olefin as per the new hypothesis.

3.6.3. True Lumping Coefficients (Case 3)

These lumping coefficients are calculated using the true gas phase protonation
equilibrium coefficient K
pr ( S

+
Oij  S Rik

which is estimated from on the precise values of the

heats of formation of the olefin and the protonated carbenium ion. Obviously, this case
does not need any reference olefin or reference carbenium ion and the lumping
coefficients are obtained from the complete eqn (3.59). This approach has been used in
the hydrocracking of pure paraffins [Chapter VI] as the method to estimate the heats for
formation of paraffinic carbenium ions is available. It should be noticed that this
approach can not be used currently for naphthenic and aromatic species due to the
unavailability of a group contribution type of estimation method for the heats of
formation of ring containing ions. Therefore, the lumping coefficients based on old and
new approaches are compared with the true LCs for the paraffinic species and the new
approach, which is found to be superior, has been followed for the naphthenic and
aromatic species in the VGO hydrocracking model.

62

(b) Cracking through (s;t) mode

(a) Cracking through (s;s) mode

1.0E+01

1.0E+01

DBP-true
DBP-nro
DBP-wro
TBP-true
TBP-nro
TBP-wro

1.0E+00
LCC

1.0E+00
LCC

MBP-true
MBP-nro
MBP-wro
DBP-true
DBP-nro
DBP-wro
TBP-true
TBP-nro
TBP-wro

1.0E-01

1.0E-01

1.0E-02

1.0E-02
2

10

14

18

22

26

30

34

10

14

18

22

26

30

34

Carbon Number

Carbon Number

(c) Cracking through (t;s) mode

(d) Cracking through (t;t) mode

1.0E+06

1.0E+04
DBP-true
DBP-nro
DBP-wro
TBP-true
TBP-nro
TBP-wro

1.0E+05

TBP-true
TBP-nro
TBP-wro

1.0E+03

LCC

LCC

1.0E+04
1.0E+02

1.0E+03
1.0E+01

1.0E+02
1.0E+01

1.0E+00
2

10

14

18

22

26

30

Carbon Number

34

10

14

18

22

26

30

34

Carbon Number

Figure 3-6. Comparison of the true lumping coefficients with the lumping coefficients calculated with and
without reference olefins.

It can be seen from Figure 3-6 that, except for the cracking of tribranched paraffins
through (s,t) and (t,t) modes [plots (b) and (d)], the proposed method (Case 2) gives
lumping coefficients much closer to the true values (Case 3), as compared to the old
approach (Case 1). Even in these two exceptions, the lumping coefficients in Case 1 are
only slightly better than Case 2. Based on the plots it can also be inferred
that K
isom ( SOij  SOr ) and K
isom ( S +  S +
Rr

Rik

are correlated inversely, as speculated. This

phenomenon can be explained based on the similarities of the skeleton structures of a

63

given olefin Oij with the corresponding carbenium ion Rik+ and the differences in their
structures from their respective reference species. An example can be given based on the
differences in the degree of branching. For a dibranched olefin Oij , the carbenium ion
Rik+ would be dibranched, whereas both the reference olefin and the reference carbenium
ion would be monobranched. Therefore, K
isom ( SOij  SOr ) corresponds to a decrease in the
degree of branching, whereas K
isom ( S +  S +
Rr

Rik

corresponds to an analogous increase in the

degree of branching, tending to offset their deviations from unity in a way so that their
product is more close to unity than either one separately.

3.7. Competitive Chemisorption on the Acidic Sites

To account for the strong competitive chemisorption of aromatics on the acidic sites eqn
(3.53) (as well as the rate expressions for group (b) and (c)) contains the denominator
term DA , which is given by eqn (3.47). The evaluation of DA requires the values of the
protonation equilibrium coefficients. In all the rate expressions for acid site
transformations except those for the group (c), i.e., dealkylation of aromatics, the
numerator contains the equilibrium coefficients for the protonation of the reactive
olefinic species. The concentration of these reactive olefinic species is very low because
of the small values of the dehydrogenation equilibrium coefficients and high hydrogen
pressure. Therefore, the contribution towards the total coverage of acidic sites due to the
protonation of these species would be relatively small. Because of this, the estimate of
the equilibrium coefficient for the protonation of the olefinic species to be used in the
denominator has been obtained from the literature (as discussed below) and is
considered constant for a given catalyst during the parameter estimation. On the other
hand, the (single event) protonation equilibrium coefficient in the numerator is coupled
with the single event rate coefficients of the elementary steps as seen from eqn (3.56)
and this composite rate coefficient is estimated from the experimental data on VGO
hydrocracking. Similarly for the dealkylation of the aromatics, the rate coefficient
coupled with the corresponding ring protonation equilibrium coefficients is estimated

64

from the experimental data and the ring protonation equilibrium coefficients for different
aromatic classes to be used in the denominator have been estimated from the literature
information. The approach for the estimation of various terms required to calculate the
acid site denominator is discussed below.

3.7.1. Dehydrogenation Equilibrium Coefficient

A single average value has been used for the dehydrogenation equilibrium coefficient of
paraffins to estimate the concentration of corresponding olefins. This is justified given
the small contribution of olefins towards the acid site coverage when large amount of
aromatics are present in the feed. The value of the dehydrogenation equilibrium
coefficient depends primarily on the nature of the carbon atoms where the double bond
is formed if the effect of the olefin structure in the vicinity of the double bond is ignored.
Therefore, five different dehydrogenation equilibrium coefficients can be considered for
the (p-s), (p-t), (s-s), (s-t) and (t-t) modes. The heats of dehydrogenation for these 5
modes range from approximately 105 to 126 kJ/mol. A value of 109 kJ/mol has been
retained. It corresponds to the (s-t) mode which is most important for the production of
tertiary carbenium ions having the highest occupancy on the acid sites. The olefins
produced through (p-s) and (s-s) modes can only produce secondary ions and (p-t) and
(t-t) modes have a much smaller contribution towards the formation of total olefins than
the (s-t) mode. Similarly, the entropy of dehydrogenation has been taken to be 120
J/mol/K for the (s-t) mode. For naphthenes with 1-ring to 4-rings, it has been assumed
that dehydrogenation is taking place primarily in the ring because of the higher affinity
of the ring towards the chemisorption on the metal sites. The heat of dehydrogenation
and entropy of dehydrogenation of 112 kJ/mol and 136 J/mol/K respectively for the (s-t)
mode in the molecule shown in Figure 3-7 are taken as the representative values for all
the naphthenic components/lumps. These values are used to estimate the concentration
of the olefinic species in eqn (3.47).

65

Figure 3-7. The representative molecule considered for the dehydrogenation equilibrium coefficients of
naphthenes.

3.7.2. Heats of Protonation

The values of the gas phase standard state heats of protonation are obtained from the
corresponding proton affinities (PA) compiled by Hunter and Lias71. The available
values of proton affinities of some species relevant to this work are given in Table 3-3.
A single value of PA has been considered for all the lumps of a particular class assuming
that it does not change with the carbon number. Since the proton affinities of all the
classes considered in the model are not available, certain simplifying assumptions were
made based on the trends of the above data.
For linear olefins formed from normal paraffins, PA corresponds to the formation of
secondary carbenium ions. Because of much higher stability of tertiary carbenium ions
compared to secondary ions, it has been assumed that all the isoparaffins and naphthenes
form the tertiary ions. Consequently, for the branched olefins formed from isoparaffins,
a PA of 821.4 kJ/mol is used. In the absence of any published information for multiring
naphthenic olefins, the PA of 1-methylcyclohexene has been used as the representative
value for all the naphthenes. It can be seen from the above data that increasing the
substitution on the benzene ring increases the PA, but the increase in the side chain
length does not affect it very much. An Increase in the number of aromatic rings also
increases PA.

66

Table 3-3. Gas phase proton affinities for the formation of most stable carbenium ion. (taken from
Hunter and Lias71).

Species

Gas phase proton

affinity (kJ/mol)

Linear Olefins (protonated to -sec ions, an 773.3

average value)
Branched Olefins (protonated to -ter ions, 821.4
an average value)
Cyclopentene

766.3

1-methylcyclopentene

816.5

Cyclohexene

784.5

1-methylcyclohexene

825.1

Benzene

750.4

Toluene

784.0

Ethylbenzene

788.0

Propylbenzene

790.1

Isopropyl benzene

791.6

n-butyl benzene

791.9

p-xylene

794.4

o-xylene

796.0

m-xylene

812.1

1,3,5 trimethylbenzene

836.2

Tetraline

809.7

Naphthalene

802.9

2-methylnaphthalene

831.9

Phenanthrene

825.7

Chrysene

840.9

1,2,3,4,5,6,7,8 octahydrophenthranene

846.2

67

The possible effect of the degree of substitution has been applied for different aromatic
classes based on the data of the above table. On an average, the PA of the mono
substituted aromatic species are taken in the model and an average of 10 kJ/mol have
been added to the values of PA to account for the effect of substitution on bare ring
aromatics. When the PA for naphtheno-aromatics are not available, their values are
obtained from the corresponding aromatics with the same number of aromatic rings. i.e.,
the value of PA for dinaphtheno-diaromatics (DNDA) is considered identical to that of
diaromatics (DAR). Based on these assumptions, the values shown in Table 3-4 are used
for the gas phase proton affinities of different classes in the model. The gas phase heats
of protonation are obtained as the negative of the corresponding proton affinities.

Table 3-4. Proton affinities for different classes used in the model.

Sr.

CLASS

No.

PA ( kJ/mol)

Reasoning

1.

NPA

773.3

For secondary carbenium ions

2.

IPA

821.4

For tertiary carbenium ions

3.

Naphthenes

825.1

Value of 1-methylcyclohexene

4.

MAR

791.9

The value of butyl benzene

5.

DAR, NDA, DNDA

831.9

The value of 2-methyl-naphthalene

6.

TAR, NTA

835.7

The value of phenanthrene + 10.0

kJ/mol for 1 substitution

7.

QAR

850.9

The value of chrysene + 10.0 kJ/mol for

1 substitution

8.

NMA

819.7

The value of tetraline + 10.0 kJ/mol for

1 substitution

9.

DNMA, TNMA

846.2

The value of octahydrophenanthrene

68

3.7.3. Heats of Stabilization

To get the heats of protonation at the surface of the catalyst, the heats of stabilization of
carbenium ions from gas phase to the catalyst surface are required. Refer Chapter VI for
derivation of the equation to relate the gas phase heats of protonation to the surface heats
of protonation as,

H prsur = H prgas + q( z, m)

(3.60)

where q is the negative of the relative heat of stabilization of the carbenium ions
defined as the difference in the heat of stabilization of proton and carbenium ion.
Considering that the extent of stabilization of a carbenium ion would depend primarily
on the local structure in the vicinity of the positive charge, it is assumed that q does
not depend on the carbon number and depends only on its nature and the molecular class
to which it belongs. The estimates of the values of q for paraffins have been reported
by several authors72-75. As the heats of stabilization depend on the acidity of the catalyst,
the values reported by different authors vary significantly. The difference in the heats of
stabilization of the secondary and tertiary paraffinic carbenium ions has been determined
from the experimental data on the hydrocracking of pure n-hexadecane in the current
work (Chapter VI) and is used in the VGO hydrocracking model. In the absence of any
published information, the heats of stabilization of naphthenic carbenium ions have been
given the same values as those for the tertiary paraffinic carbenium ions. The relatively
lower importance of this parameter due to the low contribution of paraffinic and
naphthenic species on the total acid site coverage justifies this assumption. For aromatics
and naphtheno-aromatics, q has been assumed to be independent of the nature of the
carbenium ions due the charge delocalization at the aromatic rings. In the absence of
information on the heats of stabilization of different aromatic ring containing classes, a
single value, q ( S A+ ) , has been considered as a parameter in the model for all these
classes. This parameter was then adjusted from the experimental data on the
hydrocracking of VGO available to us.

69

3.7.4. Entropy of Protonation

The entropy of protonation has been estimated using statistical thermodynamics based
on the loss of various degrees of freedom (DOF) in the protonation step. As the
protonation takes place, the olefinic and aromatic species from the sorbed phase are
attached to the acidic sites and thus the change in the entropy from the sorbed state to the
protonated state is of interest. As a molecule loses some of its rotational or translational
degrees of freedom, it gains some extra vibrational degree of freedom. It has been
assumed that the difference in the sum of the rotational and vibrational degrees of
freedom of the molecule from sorbed state and protonated state is negligible. Further
assuming that the sorption step results in the loss of one translation degree of freedom
and the protonated species has no translational motion, two translational DOF are taken
as the entropy of the protonation step in the model. The entropy corresponding to three
translational DOF is calculated from eqn (3.61) in which the translational partition
function is given by eqn (3.62).

St / R = ln(Qt ) + 5 / 2

(3.61)

2 mk BT 2
Qt =
V
2
h

(3.62)

3.8. Parameters for Sorption

To estimate the concentration of the sorbed species in the zeolite pores, the Henry
coefficients for sorption is required, as shown in eqns (3.17) and (3.18). Very few
experimental data is available in the literature on the sorption of complex compounds
found in the heavy petroleum feedstocks. Denayer et al.76 carried out experiments for the
sorption of alkanes, aromatics and some organic molecules in the liquid phase and in the
dense vapor phase on FAU zeolites, normally employed in the hydrocracking process.
They reported the values of the partition coefficients for normal paraffins from C5 to C16,
some light isoparaffins, benzene, toluene and mesitylene in Na-USY zeolite at room
temperature with n-octane and methanol as the mobile phases. With the non polar noctane as the mobile liquid phase, the partition coefficients of paraffins were found to be

70

very week functions of the carbon number of the adsorbate. On the other hand, with
methanol as the mobile liquid phase, the partition coefficients of paraffins increased with
the carbon number because of the polarity of the mobile phase. No data is available for
the effect of carbon number on the partition coefficients of naphthenes and aromatics. In
the hydrocracking of pure paraffins (Chapter VI), the Henry coefficients are assumed to
be constant with respect to the carbon number of adsorbent, considering the absence of
any polar compounds. In the hydrocracking of VGO on the other hand, considering the
presence of some polar compounds, it is assumed that the partition coefficients increase
with the carbon number. Three different base values of partition coefficients are
selected: one for paraffins, naphthenes and aromatics each, i.e., K paraf ,base , K Nap ,base and

K Aro,base . An increase in the partition coefficient by 8% per carbon number has been
considered to account for the effect of chain length. The increase in the value of the
partition coefficients in the ring compounds from its base value is related to the number
of saturated acyclic carbon atoms in the molecule. For example, a diaromatic C15
molecule contains 5 saturated acyclic carbon atoms, so that its partition coefficient is
calculated from the base value as,

K DAR ,C15 = K Aro ,base x (1.08 )

(3.63)

Initial estimates of the base values of partition coefficients were taken from the
literature76. The percent increase per carbon number and the base values were later
adjusted during the parameter estimation by using the VGO hydrocracking experimental
data. The Henry coefficients were calculated from the partition coefficients by dividing
them by the catalyst bulk density. The Langmuir adsorption coefficient K L, Si for lump

Si is finally obtained from (3.18) by using the saturated density csat , Si of lump Si .

71

CHAPTER IV
4.REACTOR MODEL AND PARAMETER ESTIMATION
4.1. Reactor Model
The global rate expressions developed in Chapter III are plugged into a three-phase
multibed tubular reactor model operated adiabatically with inter-stage hydrogen
quenching. Under typical industrial operating conditions, hydrocracking reactors operate
in trickle flow regime with a fixed bed of porous catalyst particles and a vapor phase and
a liquid phase flowing co-currently. It is assumed in what follows that the particles are
completely wetted and the gas and liquid phases are in plug flow. As the aromatic
hydrogenation and hydrocracking reactions proceed, H2 is transferred from gas to the
liquid phase and the lighter products formed by cracking steps are transferred from the
liquid phase into the gas phase. This inter-phase mass transfer flux is described using the
two film theory.77 For a particular lump S g , the molar flux from gas to liquid phase is
given by the following equation

N Sg

CSggas

liq
= kO ,Sg C ,VLE CSg

Sg

(4.1)

In the above equation, the overall mass transfer coefficient is given by

1
kO , Sg

1
1
+
C ,VLE
kG K Sg
kL

(4.2)

For the trickle flow regime, the value of the liquid phase mass transfer coefficient k L av
has been calculated from a correlation given by Sato78 and the gas side mass transfer
coefficient kG av is calculated from Reiss79 correlation. The gas-liquid interfacial area
C ,VLE
av is calculated by the correlation given by Charpentier.80 K Sg
is the concentration

equilibrium partition coefficient relating the equilibrium concentrations of lump S g in

the gas and liquid phases as

72

C ,VLE
K Sg
=

CSggas

(4.3)

liq
CSg

C ,VLE
VLE
K Sg
is expressed in terms of the true equilibrium partition coefficient K Sg
as

ySg
C ,VLE
K Sg
=
x
Sg

C gas
total
liq
Ctotal

gas

VLE Ctotal
K
=

Sg
liq

Ctotal

. The values of the equilibrium partition

VLE
coefficient K Sg
are calculated using the Peng-Robinson equation of state. The

continuity equations for the gas phase components/lumps and hydrogen are written as
gas
CSggas

1 dFSg
liq
g = 1,2,.........N lumps
= kO , Sg av C ,VLE CSg

dz
Sg

(4.4)

The continuity equations for components/lumps in the liquid phase also take into
account the net rate of formation of component/lump g along with the gas-liquid mass
transfer flux term,
liq
CSggas

1 dFSg
liq
= kO , Sg av C ,VLE CSg
+ RSgForm

, net g = 1,2,.........N lumps

dz
K
Sg

(4.5)

The mass transfer resistance between the liquid and solid phases and the intra-particle
diffusional resistance has been considered to be negligible. As discussed earlier, the
composition of VGO in this model is defined in terms of 1266 lumps/pure components.
Therefore in the three-phase hydrocracking of VGO, 1267 (viz., 1266 + 1 for hydrogen)
continuity equations for the liquid phase and the same number of continuity equations
for the gas-phase components/lumps are required.
A

significant

amount

of

heat

is

released

during

the

hydrogenation

of

aromatics/naphtheno-aromatics and the hydrocracking reactions. The liquid phase

temperature profile along the reactor bed is obtained from eqn (4.6)

73

dT liq
dz

N lumps
Nlumps

dFSggas
liq
Form

H
R

Sg
(

f , g ) Sg , net
dz
g =1
g =1

=
Nlumps

(F

gas
Sg

gas
P , Sg

liq
Lg

+F C

liq
P , Sg

(4.6)

g =1

In the numerator of eqn (4.6), the first term represents the total heat of reaction and is
calculated using the heats of formation of individual lumps/pure components obtained
from the Bensons group contribution method. The second term accounts for the heat
consumed during the vaporization of lighter components from the liquid to the gas
phase. The heat transfer resistance between the gas and liquid phase has been assumed to
be negligible. The temperature of the solid phase is obtained from eqn (4.7)
Nlumps

( H )R
liq
f ,g

T solid = T liq +

g =1

aLS hLS

Form
Sg , net

(4.7)

in which the solid-liquid heat transfer coefficient hLS has been estimated using the
correlation of Whitaker81.
The set of equations given by (4.4), (4.5) and (4.6) consists of 2,535 (= 1267 x 2 +1) stiff
ODEs which is solved using the backward differentiation method (BDF). The subroutine
DLSODA of the FORTRAN library ODEPACK has been used for the integration of
ODEs. The integration of this set of ODE was found to be extremely slow with
numerically calculated Jacobian matrices because of the large number of function
evaluations for each Jacobian. A subroutine has been written to get an approximate
analytical Jacobian for this system reducing the total integration time by 2-3 orders of
magnitude.
The initial conditions for the gas and liquid molar flow rates are obtained by a VLE flash
calculation at the reactor inlet temperature and pressure. A maximum allowable
temperature rise Tmax is specified for each reactor bed. The integration is carried out till
the reactor temperature reaches Tin + Tmax after which a cold stream of hydrogen at

74

temperature TH 2 is injected to quench the reaction mixture. The amount of hydrogen

required to bring the temperature of the reaction mixture back to the temperature of the
inlet to the previous bed, Tin is calculated from:
Nlumps

(F

gas
Sg

FH 2 , quench =

liq liq
CPgas
, Sg + FLg C P , Sg ) ( Tout Tin )

g =1

CPgas
, H 2 TH 2 Tin

(4.8)

4.2. Estimation of Properties

The various physical and chemical properties of the lumps/pure components involved in
the model are estimated using the following methods:
Critical temperature and critical pressure are calculated using Jobaks group contribution
method82 and API recommended methods.
Saturated and compressed liquid phase densities are estimated using Hankinson and
Thomson (HBT) method82.
The viscosities of gas and liquid phases are calculated by the method of Twu et al.83
Vapor pressures and latent heat of vaporization are calculated using the method of
Prausnitz et al.84
Heats of formation, entropies, equilibrium coefficients, and gas phase heat capacities are
calculated using Benson's group contribution method65.
Liquid phase heat capacities are estimated from Garvin et al.85

4.3. Parameter Estimation

The parameters involved in the VGO hydrocracking model are summarized in Table 4-1
with their definition and units. These parameters have been estimated from a relatively
small amount of experimental data on the three phase hydrocracking of a VGO feedstock
obtained from a lab scale tubular reactor. The product composition was available in
terms of the weight percent of normal and isoparaffins, one to four ring naphthenes, and

75

one to four ring aromatics classes. The composition of naphtheno-aromatic classes was
not available separately and was combined with the corresponding aromatics classes,
e.g., one ring aromatics include the monoaromatics, naphtheno-monoaromatics,
dinaphtheno-monoaromatics and trinaphtheno-monoaromatics. An estimate of the
carbon number distribution was available for normal paraffins, isoparaffins,
mononaphthenes, dinaphthenes and trinaphthenes classes. The initial guess of the
parameters were obtained from the literature46, 70, 72, 73, 76, 86 when available or have been
selected judiciously otherwise. This information was utilized to estimate the model
parameters first by manual tuning followed by the application of Levenberg-Marquardts
optimization algorithm to minimize the sum of square of the residuals between the
experimental and model predicted responses. The final values of the parameters have not
been reported in the dissertation because of proprietary reasons as the model has been
licensed to an industrial organization. The optimized parameters related to steps on the
acid sites have been found to satisfy the rules of carbenium ion chemistry. The activation
energies corresponding to the saturation of aromatics are in the range of 25 to 50 kJ/mol
as reported in the literature46, 59. The relative magnitude of the parameters related to the
sorption in the zeolite pores and the chemisorption on the metal sites are in accordance
with their physico-chemical nature. The fit of the model to the experimental data has
been shown as normalized parity plots in Figure 4-1 and Figure 4-2. It can be seen from
Figure 4-2 (a) and (b) that the normal paraffins are relatively overpredicted whereas the
isoparaffins are slightly under predicted. The carbon number distribution of the total
paraffins [Figure 4-1 (b)] shows a very good fit to the experimental data, however. The
agreement of the carbon number distribution of the entire reactor effluents [Figure 4-1
(a)], mononaphthenes [Figure 4-2 (c)], dinaphthenes [Figure 4-2 (d)], the composition of
different classes [Figure 4-1 (c)] and of different commercial fractions [Figure 4-1 (d)] is
quite good, considering the complexity of the reaction network and the relatively small
number of temperature and feedstock independent parameters employed in the model.

76

Table 4-1. Model parameters employed in the VGO hydrocracking model.

S.
No.

Parameter

Definition

Units

1.

comp
A
pcp

S pr
A
pcp Ct exp
R

2.

comp
E pcp
( s;s )

E pcp ( s ;s ) + H pr ( SOij  s )

[kJ/mol]

3.

comp
E pcp
( s ;t )

E pcp ( s ;t ) + H pr ( SOij  s )

[kJ/mol]

4.

comp
E pcp
( t ;t )

E pcp (t ;t ) + H pr ( SOij t )

[kJ/mol]

5.

comp
A
acyc

S pr
A
acyc Ct exp
R

[kmol/kgcat/hr]

6.

comp
Eacyc
( s;s )

Eacyc ( s ;s ) + H pr ( SOij  s )

[kJ/mol]

7.

comp
Eacyc
( s ;t )

Eacyc ( s ;t ) + H pr ( SOij  s )

[kJ/mol]

8.

comp
Eacyc
(t ;s )

Eacyc (t ;s ) + H pr ( SOij t )

[kJ/mol]

9.

comp
Eacyc
( t ;t )

Eacyc (t ;t ) + H pr ( SOij t )

[kJ/mol]

10.

comp
A
exo

S pr
A
exo Ct exp
R

[kmol/kgcat/hr]

11.

Eexo ( m;n )

Eexo ( m , n ) Eacyc ( m ,n )

[kJ/mol]

12.

comp
A
endo

S pr
A
endo Ct exp
R

[kmol/kgcat/hr]

13.

Eendo ( m;n )

Eendo ( m ,n ) Eacyc ( m , n )

14.

comp
A
dealk

A C exp S pr ring
dealk t

15.

comp
Edealk
(s)

comp
Edealk
( s ) + H pr ring

[kmol/kgcat/hr]

[kJ/mol]

[kmol/kgcat/hr]
[kJ/mol]

77

Table 4-1. Continued.

S.
No.

Parameter

Definition

Units

16.

comp
Edealk
(t )

comp
Edealk
( t ) + H pr ring

17.

comp
A
cyc

S pr diole
A
cyc Ct exp

18.

comp
Ecyc
( s;s )

Ecyc ( s ;s ) + H pr ( SDOij  s )

[kJ/mol]

19.

comp
Ecyc
( s ;t )

Ecyc ( s ;t ) + H pr ( SDOij  s )

[kJ/mol]

20.

comp
Ecyc
(t ;s )

Ecyc (t ;s ) + H pr ( SDOij t )

[kJ/mol]

21.

comp
Ecyc
( t ;t )

Ecyc (t ;t ) + H pr ( SDOij t )

[kJ/mol]

22.

comp
Ahyd
,3 H 2

Ahyd ,3 H 2 cm

[(kmol/kgcat/hr)(m3/kmol)3]

23.

comp
Ahyd
,2 H 2

Ahyd ,2 H 2 cm

[(kmol/kgcat/hr)(m3/kmol)2]

24.

Ehyd ,3 H 2

[kJ/mol]

25.

Ehyd ,2 H 2

[kJ/mol]

26.

chem
pconst

[Dimensionless]

27.

chem
psat
_ rings

[Dimensionless]

28.

chem
paro
_ rings

[Dimensionless]

29.

p chem
paraf

[Dimensionless]

30.

q ( S A+ )

q ( H + ) q ( S A+ )

31.

K paraf ,base

[Dimensionless]

32.

K Nap ,base

[Dimensionless]

33.

K Aro,base

[Dimensionless]

[kJ/mol]

[kmol/kgcat/hr]

[kJ/mol]

78

(a)

(b)

15

Predicted, wt%

Predicted, wt%

5
10

4
3
2
1

10

15

Experimental, wt%

(c)

50
40

30

Predicted, wt%

Predicted, wt%

(d)

40

20

10

Experimental, wt%

10

20

30

Experimental, wt%

40

30
20
10
0

10

20

30

40

50

Experimental, wt%

Figure 4-1. Comparison of the experimental and predicted product composition in the hydrocracking of
VGO. (a) Carbon number distribution (C3 to C40) of the total reactor effluents. (b)

Carbon number

distribution (C3 to C40) of the total paraffins. (c) Weight percent composition of normal paraffins,
isoparaffins, one to four ring naphthenes and one to four ring aromatics. (d) Weight percent of different
commercial fractions, i.e., LPG, light naphtha, heavy naphtha, kerosene, diesel and the unconverted VGO.

79

(b)
4

1.5

Predicted, wt%

Predicted, wt%

(a)
2

0.5

0.5

1.5

Experimental, wt%

(c)
2.5
2

Predicted, wt%

Predicted, wt%

(d)

Experimental, wt%

1.5
1
0.5
0

Experimental, wt%

0.5

1.5

2.5

Experimental, wt%

Figure 4-2. Comparison of the experimental and predicted product composition in the hydrocracking of
VGO. (a) Carbon number distribution (C3 to C40) of the normal paraffins. (b) Carbon number distribution
(C4 to C40) of the isoparaffins. (c) Carbon number distribution (C5 to C40) of the mononaphthenes. (d)
Carbon number distribution (C10 to C40) of the dinaphthenes.

80

CHAPTER V
5.REACTOR SIMULATIONS
The parameters estimated in Chapter IV have been used to carry out a number of nonisothermal reactor simulations in an industrial scale hydrocracking reactor for a wide
range of operating conditions. The reactor geometry taken from the literature87 and the
catalyst properties selected for these simulations are shown in Table 5-1. The
composition of the feedstock vacuum gas oil has been taken from Moustafa et al.88 and
is given in Table 5-2. The actual feedstock composition required as the input for the
reactor simulations is derived from this information as discussed in section 5.2.3. The
products obtained from the hydrocracking of this feedstock have been classified into
different commercial fractions based on their carbon number range as shown in Table
5-3. The conversion of VGO is defined based on the cracking of C20+ feed into C20 and
lighter products and is calculated by
prod
WCfeed

20 + WC 20 +
X =
100

WCfeed
20 +

(5.1)

In the above eqn, Ws have the units of weight. This criterion for calculating the
conversion is analogous to the one used in industry which is based on the end point of
diesel.
A maximum temperature increase of 15 oC per bed has been adopted for all the
simulations carried out in this study and is used as the criterion to switch to the next
catalyst bed, i.e., when the temperature of the reaction mixture reaches Tin + 15 oC ,
quenching hydrogen stream at 40 oC is injected to reduce the temperature to Tin .
Consequently, the number and length of different beds depends on the total amount of
heat released during the hydrocracking reactions and the heat release profile
corresponding to the selected operating conditions. The model can also be used to
simulate an existing hydrocracking unit with a given number of beds and pre-specified

81

bed lengths to obtain the temperature increase and conversion in each bed and the
composition of final products.

Table 5-1. Reactor geometry and catalyst properties for the reactor simulations.

Reactor Geometry
Total Bed Length

11.6 m

Reactor Diameter

3.7 m

Catalyst Properties
Catalyst Bulk Density

830 kgcat/m3r

Particle Equivalent Diameter

3.0 mm

Bed Porosity

0.4

Catalyst micropore volume

2.8x10-4 m3/kgcat

82

Table 5-2. Composition of the VGO. (weight percent).

CN

nPar

iPar

MNA DNA

TNA TETNA MAR DAR TAR NMA NDA NTA Total

14

0.827 0.047 0

0.413 0

15

0.002 0

2.330 0.359 0

16

0.016 0.002 0.004 0

2.710 1.010 0

17

0.077 0.002 0.032 0.021 0

1.287

0.180 0

2.871

0.500 0

4.242

2.320 0.950 0.323 0

0.470 0

4.195

18

0.167 0.039 0.128 0.131 0.022 0

1.610 0.915 0.459 0

0.230 3.701

19

0.365 0.152 0.339 0.312 0.109 0.018

0.718 0.335 0.126 0

0.063 2.537

20

0.729 0.333 0.614 0.613 0.247 0.196

0.246 0.168 0.078 0

0.039 3.263

21

1.010 0.570 1.030 1.050 0.541 0.447

0.086 0.113 0.047 0

0.023 4.917

22

1.430 0.910 1.470 1.380 0.650 0.723

0.012 0.020 0.010 0

6.605

23

1.560 1.270 1.920 1.770 0.873 0.793

8.186

24

1.860 1.600 2.380 2.220 1.240 0.873

10.173

25

2.070 1.720 2.790 2.460 1.370 0.693

11.103

26

1.650 1.450 2.590 2.270 1.090 0.533

9.583

27

1.270 1.360 2.380 2.110 0.897 0.439

8.456

28

0.900 1.110 2.070 1.930 0.615 0.395

7.020

29

0.538 0.723 1.550 1.520 0.403 0.235

4.969

30

0.343 0.426 1.150 1.140 0.163 0.118

3.340

31

0.183 0.268 0.712 0.608 0.090 0.082

1.943

32

0.115 0.163 0.405 0.248 0.092 0.060

1.083

33

0.040 0.070 0.236 0.015 0.125 0.047

0.533

Total 14.325 12.168 21.80 19.798 8.527 5.652

10.859 3.917 1.043 0.413 1.150 0.355 100.000

83

Table 5-3. Definition of different commercial fractions based on carbon number range. (The last
column has the atmospheric equivalent boiling points1 of normal paraffins for this carbon number
range).

Commercial

Carbon

Atmospheric Equivalent

Fraction

No.

Boiling Points (oC)

LPG

C3 C4

Light Naphtha (LNAP)

C5 C7

40 110

Heavy Naphtha (HNAP)

C8 C12

110 220

Kerosene (KERO)

C13 C15

220 260

Diesel (DIES)

C16 C20

260 350

> C20

> 350

Unconverted (UNCONV)

The simulations performed to study the effect of the reactor inlet temperature and total
pressure on the hydrocracking products are discussed in Section 5.1 and 5.1.2
respectively. The effect of the change in the distribution of isomers of a class with
different degrees of branching on the hydrocracking conversion and evolution of various
subclasses is discussed in Section 5.2.3.

5.1. Effect of Inlet Temperature

To study the effect of temperature on the VGO conversion and product distribution, the
simulations have been carried out at two different reactor inlet temperatures ( Tin ): 350
o

C and 370 oC; both at a total pressure of 150 bar and a H2 to Hydrocarbon molar ratio of

15.0 (~ 5000 Std. ft3/barrel). The VGO flow rate has been taken as 620 kmol/hr. The

84

various simulation results are shown in Figure 5-1 to Figure 5-6 with solid lines
corresponding to the simulations at 350 oC and the dotted lines at 370 oC. Figure 5-1 (a)
shows the temperature profiles of the solid and liquid phases along the reactor. At Tin =
350 oC, the total temperature increase in the reactor is 40.8 oC in three beds. It can be
seen that the length of the first bed is smaller than the second one, showing a higher rate
of heat release in first bed due to higher rates of exothermic aromatics hydrogenation
and cracking reactions. At Tin = 370 oC on the other hand, the total temperature increase
in the reactor is 58.8 oC with a total of four beds. A higher inlet temperature results in
higher rates of aromatics hydrogenation and hydrocracking reactions giving a higher rate
of temperature increase in each bed as compared to 350 oC. The heat is generated in the
solid catalytic phase from where it is transferred to the liquid phase. Because of the
higher rate of heat release, the average T between solid and liquid phases at Tin = 370
o

C is slightly higher (~ 0.4 oC) as compared to Tin = 350 oC (~0.2 oC).

The profile of the VGO conversion along the beds can be seen from Figure 5-1 (b)
indicating 88.6 % total conversion at Tin = 370 oC as compared to only 56.7 %
conversion at Tin = 350 oC. A conversion increase of approximately 26 % has been
reported in the literature89 with a temperature increase of 22 oC. Although the increase in
conversion with temperature depends on the composition of the feedstock as well as the
catalyst, the predicted increase in the total conversion is comparable to the reported
value. Figure 5-1 (c) shows the flux of hydrogen in the gas and liquid phases along the
reactor. The gas phase hydrogen flux (blue lines) drops along a bed due to the transfer of
hydrogen from gas to the liquid phase with a faster drop at high temperature, indicating
faster consumption of hydrogen. A step increase in gas phase hydrogen flux occurs at the
bed exits because of the injection of quench hydrogen. The amount of each quench
hydrogen shot is approximately 10 % of the inlet hydrogen at both temperatures. The
hydrogen flux in the liquid phase depends on the solubility of hydrogen in the liquid
phase, the rate of hydrogen transfer from gas to liquid phase and the total amount of
hydrocarbons present in the liquid phase. It can be seen from Figure 5-1 (d) that the

85

weight % vaporization of hydrocarbons increases very rapidly at higher temperature

because of more amounts of lighter hydrocarbons formed due to faster cracking,
resulting in a faster decrease in the total amount of the liquid phase hydrocarbons. This
ultimately contributes to the significantly high drop in the hydrogen flux in liquid phase
at 370 oC. Figure 5-1 (d) also indicates that at the inlet of the reactor, the percentage of
vaporization is higher at high temperature as the initial amounts of the vapor and liquid
phases are obtained by a VLE flash calculation at the reactor inlet conditions. A sudden
increase in the percentage of vaporization at the bed exits is also observed because of the
addition of quench hydrogen.

5.1.1. Effect of Temperature on the Yields of Various Commercial Fractions

Figure 5-2 (a) shows the effect of temperature on the evolution of various commercial
fractions along the reactor. It can be seen that at Tin = 350 oC, the diesel fraction (DIES)
initially increases in the first bed and remains almost steady in the last two beds,
ultimately resulting in a maximum diesel yield among all commercial fractions. At Tin =
370 oC, the yield of the diesel fraction increases quite fast in the first bed, maintains
almost a steady value in the second bed and then sharply decreases in the last two beds.
This faster increase in the diesel yield is caused by the faster cracking of VGO at higher
temperature. As approximately 64 % of VGO conversion takes place in the first two
beds itself, the cracking of diesel range hydrocarbons starts dominating in the third bed,
resulting in a decrease in the diesel yield accompanied with a sharp increase in the heavy
naphtha (HNAP) yield. It can also be observed that unlike 350 oC, the yield of heavy
naphtha is highest among all the fractions at 370 oC indicating a shift in the
hydrocracking from middle distillate to the naphtha mode. The yield of kerosene
(KERO) at 370 oC increases faster in the first bed as compared to 350 oC but reaches
almost the same value at the exit of the reactor. Both the light naphtha (LNAP) and LPG
yield increase monotonically along the reactor and show a marked increase with
temperature.

86

(a)

(b)

400

100
Solid
Liquid

80
Conversion (Wt %)

Temperature, oC

390
380
370
360

60

40

20
350
0

0.2
0.4
0.6
0.8
Fractional Reactor Length

(c)

(d)
60

800
30

700

25

20

0.2
0.4
0.6
0.8
Fractional Reactor Length

15
1

50
Vaporization (Wt%)

900

H2 Flux liq phase (kmol/hr/m2)

1000
H2 Flux gas phase (kmol/hr/m2)

0.2
0.4
0.6
0.8
Fractional Reactor Length

40
30
20
10
0

0.2
0.4
0.6
0.8
Fractional Reactor Length

Figure 5-1. Effect of reactor inlet temperature on the (a) Temperature profile of solid and liquid phases
along the reactor beds. (b) Weight percent conversion of VGO. (c) Hydrogen flux in the gas and liquid
phases along the reactor beds. (d) Weight percent vaporization of hydrocarbons along the reactor beds.
(solid lines: 350 oC, dotted lines: 370 oC).

87

(a)

(b)

50

50
NPA

LPG (C3-C4)

IPA

LNAP (C5-C7)

DIES (C16-C20)

30

20

10

MNA
DNA

40

HNAP (C8-C12)
KERO (C13-C15)

Yield (Wt %)

Yield (Wt %)

40

TNA

30

TETNA

20

10

0.2
0.4
0.6
0.8
Fractional Reactor Length

0.2
0.4
0.6
0.8
Fractional Reactor Length

(c)

(d)

12

NMA

MAR

NDA

DAR

10

DNMA

1.5

TETAR

Yield (Wt %)

Yield (Wt %)

TAR

6
4

0.5
2
0

0.2
0.4
0.6
0.8
Fractional Reactor Length

0.2
0.4
0.6
0.8
Fractional Reactor Length

Figure 5-2. Effect of reactor inlet temperature on the (a) Evolution of various commercial fractions. (b)
Evolution of paraffins and one to four ring naphthenes. (c) Evolution of one to four ring aromatics. (d)
Evolution of naphtheno-aromatics. (solid lines: 350 oC, dotted lines: 370 oC).

88

(a)

(b)

0.4

60
NTA

0.35

DNDA

0.25

Yield (Wt %)

Yield (Wt %)

50

TNMA

0.3

0.2
0.15

40

PAR
NAP

30

ARO
NAP-ARO

20

0.1
10

0.05
0

0.2
0.4
0.6
0.8
Fractional Reactor Length

(c)

0.2
0.4
0.6
0.8
Fractional Reactor Length

(d)

12

12
C4

C20

C8

10

C24

10

C12

C28

Yield (Wt %)

Yield (Wt %)

C16

6
4
2
0

C32

8
6
4
2

0.2
0.4
0.6
0.8
Fractional Reactor Length

0.2
0.4
0.6
0.8
Fractional Reactor Length

Figure 5-3. Effect of reactor inlet temperature on the (a) Evolution of naphtheno-aromatics. (b) Evolution
of total paraffins, naphthenes, aromatics and naphtheno-aromatics. (c, d) Evolution of C4 to C32
hydrocarbons. (solid lines: 350 oC, dotted lines: 370 oC).

89

(a)

(b)

1.4

10
C3

NPAR

NC4

1.2

IPAR

IC4

MNAP
MARO

Yield (Wt %)

Yield (Wt %)

1
0.8
0.6

0.4
2
0.2
0

0.2
0.4
0.6
0.8
Fractional Reactor Length

0.2
0.4
0.6
0.8
Fractional Reactor Length

(c)

(d)

25

10
NPAR
IPAR

20

PAR
NAP

MNAP

ARO
NAP-ARO

MARO

15

Yield (Wt %)

Yield (Wt %)

DNAP
DARO
NMA

10

0.2
0.4
0.6
0.8
Fractional Reactor Length

(e)

0.2
0.4
0.6
0.8
Fractional Reactor Length

(f)

14

60

12

50

PAR

NAP-ARO

PAR
NAP

Yield (Wt %)

Yield (Wt %)

10

ARO

NAP-ARO

6
4

40
30
20
10

2
0

NAP
ARO

0.2
0.4
0.6
0.8
Fractional Reactor Length

0.2
0.4
0.6
0.8
Fractional Reactor Length

Figure 5-4. Effect of reactor inlet temperature on the (a) Evolution of different components of LPG. (b)
Evolution of different classes of light naphtha. (c) Evolution of different classes of heavy naphtha. (d)
Evolution of different classes of kerosene. (e) Evolution of different classes of diesel. (f) Evolution of
different classes of unconverted fraction. (solid lines: 350 oC, dotted lines: 370 oC).

90

4.5
LPG (C3-C4)
LNAP (C5-C7)

HNAP (C8-C12)
KERO (C13-C15)

iP/nP wt. ratio

3.5

DIES (C16-C20)

3
2.5
2
1.5
1
0.5

0.1

0.2

0.3

0.4
0.5
0.6
0.7
Fractional Reactor Length

0.8

0.9

Figure 5-5. Effect of reactor inlet temperature on the iso to normal paraffins ratio of various commercial
fractions. (solid lines: 350 oC, dotted lines: 370 oC).

91

(a) Total Products

(b) NPA

15

(c) IPA

2.5

Yield (wt %)

Yield (wt %)

Yield (wt %)

2
10

1.5
1
0.5

10

20
30
C-number
(d) MNA

40

20
30
C-number
(e) DNA

2
1
0

40

10

20
30
C-number
(f) TNA

40

10

20
30
C-number
(i) TAR

40

10

20
30
C-number

40

1.5

6
4
2
0

10

2.5

Yield (wt %)

Yield (wt %)

Yield (wt %)

1.5
1

0.5

0.5
0

10

20
30
C-number
(g) MAR

40

10

20
30
C-number
(h) DAR

40

1.5

0.5

Yield (wt %)

Yield (wt %)

Yield (wt %)

0.4
2

0.5

0.3
0.2
0.1

10

20
30
C-number

40

10

20
30
C-number

40

Figure 5-6. Effect of reactor inlet temperature on the carbon number distribution of hydrocarbon classes.
(diamonds: feed. circles: products at 350 oC. triangles: products at 370 oC).

92

5.1.2. Effect of Temperature on the Yields of Various Hydrocarbon Classes

Figure 5-2 (b) shows the effect of temperature on the evolution of various paraffinic and
naphthenic classes along the reactor. At 350 oC, the yields of dinaphthenes (particularly),
trinaphthenes and tetranaphthenes show a sharp decrease in the first bed with a slower
decrease in the last two beds. This occurs primarily because of the ring opening
reactions, as can be seen from the increase in the yield of mononaphthenes, and, to a
lesser extent, because of the scission in the alkyl side chains of these species. The yield
of mononaphthenes at a given temperature also increases along the reactor because of
the hydrogenation of monoaromatics and cyclization of paraffins. As the inlet
temperature is increased to 370 oC, the rate of decrease of the dinaphthenes yield
increases significantly accompanied by a marked increase in the yield of
mononaphthenes. The increase in the yield of mononaphthenes at higher temperature
also indicates higher rates of cyclization of paraffins. The yield of normal and
isoparaffins increases along the reactor because of the cracking in the alkyl side chain of
a number of ring species. The normal paraffins are converted into isoparaffins only
through PCP(s;s) elementary steps. These were found to have the highest composite
activation energy as compared to other acid site PCP and cracking steps. Because of this,
the yield of isoparaffins slightly increases with a decrease in the yield of normal
paraffins as the temperature is increased from 350 oC to 370 oC.
Figure 5-2 (c) shows the effect of temperature on the evolution of monoaromatics,
diaromatics and triaromatics along the reactor. No tetraaromatics are present in the VGO
feedstock. As the temperature is increased, all the aromatics are hydrogenated at a faster
rate, thus decreasing their yields. Figure 5-2 (d) shows the effect of temperature on the
evolution of different naphtheno-aromatics along the reactor. At 350 oC, the yield of
naphtheno-monoaromatics (NMA) first increases as they are formed from the
hydrogenation of diaromatics (DAR). Subsequent hydrogenation of NMA to
dinaphthenes occurs after 80% reactor length causing a slight decrease in their yield.
Naphtheno-diaromatics (NDA) are formed from the hydrogenation of triaromatics

93

(TAR) causing a slight increase in their yield in the first 20 % of the reactor. At the same
time, hydrogenation of NDA produces dinaphtheno-monoaromatics (DNMA) with a
decrease in their yield. When the temperature is increased to 370 oC, NMA and DNMA
are produced faster in the beginning, reach a maximum and are subsequently
hydrogenated to more saturated products in the later part of the reactor. Figure 5-3 (a)
shows the evolution of a number of four ring naphtheno-aromatics along the reactor and
the effect of temperature on their profiles. Only 0.35 wt% of naphtheno-triaromatics
(NTA) are present in the feedstock. They are hydrogenated into dinaphtheno-diaromatics
(DNDA). These are primary products as can be seen from their non-zero initial slope.
Subsequent hydrogenation of DNDA produces trinaphtheno-monoaromatics (TNMA)
with a zero slope, characterizing them as secondary products. The effect of temperature
on the evolution of these classes is analogous to NMA, NDA and DNMA classes.
Figure 5-3 (b) shows the evolution of the total amount of paraffins, naphthenes,
aromatics and naphtheno-aromatics and the effect of temperature on their profiles. As
can be concluded from the previous discussion, the total aromatics yield decreases along
the reactor and the rate of decrease increases with temperature. Because of the sequential
nature of ring saturation, the hydrogenation of naphtheno-aromatics produces another
naphtheno-aromatics except when the reacting species has only one aromatic ring.
Therefore, at 350 oC, naphtheno-aromatics keep of increasing along the reactor,
however, at 370 oC, the complete saturation of these species results in the decrease in
their yield causing an increase in the yield of naphthenes in the later part of the reactor.
The total paraffins yield increases along the reactor at both the temperatures as they are
formed from the dealkylation of aromatics and cracking in the side chain of naphthenic
and aromatic species. At higher temperature, the initial rate of increase of the paraffins
yield is increased because of the higher rate of cracking. The experimental data used in
this study for the parameter estimation showed a significant amount of cyclization of
paraffins to mononaphthenes. As the cyclization elementary steps proceeds through
diolefinic species, the effective activation energy for these steps include twice the heat of
dehydrogenation of paraffins and the composite activation energies of cyclization. A

94

value of 25-30 kcal/mol of the heat of dehydrogenation of paraffins leads to large

effective activation energies for the cyclization steps. This can be seen from the increase

in the yield of total naphthenes at high temperature at the exit of the reactor. A
significant increase in the value of rate coefficient for the cyclization of paraffins to
mononaphthenes has been reported by Filimonov et al.90 as the temperature is increased.
Part of this increase can be attributed to the complete saturation of naphtheno-aromatics,
as mentioned above; the remaining increase is due to the cyclization of the paraffins.

5.1.3. Effect of Temperature on the Yields of Hydrocarbons of Different Chain

Length
Figure 5-3 (c) and (d) show the evolution of hydrocarbons of different chain lengths
from C4 to C32 and the effect of temperature on their yields. The yields of C4, C8 and C12
increase along the reactor as they are always formed from the heavier hydrocarbons. The
rate of increase is faster at higher temperature because of higher rates of cracking. The
yield of C16 at 350 oC increases in the first 70% of the reactor and maintains almost a
constant value afterwards, indicating its rate of formation from heavier hydrocarbons to
be almost equal to its rate of cracking to lighter ones. At 370 oC, on the other hand, C16
is formed at a faster rate from the heavier hydrocarbons, reaches a maxima, and
subsequently the rate of cracking of C16 becomes more than the rate at which it is formed
from heavier hydrocarbons, resulting in a significant drop in its yield. C20 shows a
similar behavior. The yields of hydrocarbons heavier than C20 always decrease along the
reactor, the more so as the temperature is increased.

5.1.4. Effect of Temperature on the Yields of Individual Classes of Various

Commercial Fractions
Figure 5-4 (a) shows the evolution of different components of LPG along the reactor.
The yield of isobutane is always higher than that of n-butane, as can be seen from a
typical composition of LPG obtained from the hydrocracking process. The yields of all
the components increase along the reactor and the rate of increase of these yields more
than doubles with a temperature increase of 20 oC.

95

The evolution of different classes of light naphtha (LNAP: C5-C7) is shown in Figure 5-4
(b). In light naphtha also isoparaffins have the maximum yield followed by n-paraffins
and mononaphthenes. The yields of these classes increase monotonically along the
reactor. The increase is faster at the elevated temperature. Only negligible amounts of
monoaromatics are present in this fraction.
The evolution of different classes belonging to the heavy naphtha fraction (HNAP: C8C12) is shown in Figure 5-4 (c). Components belonging to the mononaphthenes class
constitute a significant amount in the heavy naphtha fraction. The yield of
mononaphthenes increases drastically from 8% to 22% as the temperature is increased
from 350 oC to 370 oC. The increase in the yield of normal and isoparaffins with
temperature is much less pronounced than that of mononaphthenes. This can be
attributed to the increased rate of cyclization of paraffins to form mononaphthenes as the
temperature is increased, already discussed above. The yield of dinaphthenes, although
low, also increases with temperature because of higher rate of cracking of the side chains
of the heavier dinaphthenes as well as the increased rate of ring opening of trinaphthenes
belonging to the HNAP fraction. The yield of monoaromatics is not affected very much
with a temperature increase. The latter increases the rate at which monoaromatics are
produced from the dealkylation and side chain scission of heavier monoaromatics, which
in turn is compensated by an increase in the rate of hydrogenation of monoaromatics
components in the HNAP fraction. Diaromatics and naphtheno-monoaromatics are
essentially absent in heavy naphtha because of the presence of very small amounts of
these species in the VGO feedstock and their faster hydrogenation.
The evolution of paraffins, naphthenes, aromatics and naphtheno-aromatics in the
kerosene (KERO: C13-C15), diesel (DIES: C16-C20) and unconverted VGO (UNCOV:
C21-C33) fraction is shown in Figure 5-4 (d), (e) and (f) respectively. At 350 oC, the
yields of paraffins and naphthenes increase along the reactor in the kerosene and diesel
fractions. On the Contrary, at 370 oC, their yields increases at a faster rate in early zones
of the bed, reaches a maximum and then drops in the last part of the reactor. The
decrease in the yield of paraffins begins faster than that of naphthenes, indicating either a

96

relatively higher rate of cracking of the paraffins of this fraction than naphthenes at the
elevated temperature or the increase in the rate of cyclization of paraffins to naphthenes.
For both the kerosene and diesel fractions, the yield of aromatics decreases along the
reactor because of ring saturation. Given the larger amount of aromatics in the diesel
fraction, its rate of decrease is significantly higher as compared to that of kerosene
fraction. The rate of aromatics hydrogenation is accelerated as the temperature is
increased. The temperature increase causes the same effect on the profiles of naphthenoaromatics in the kerosene and diesel fractions as on the total naphtheno-aromatics,
discussed above [Figure 5-3 (b)]. Paraffins and naphthenes in the unconverted VGO
keep on decreasing along the reactor at a rate that increases with temperature, as
expected. No aromatics and naphtheno-aromatics are present in the carbon number range
of unconverted fraction as can be seen from the composition of the feedstock.

5.1.5. Effect of Temperature on the Iso/Normal Paraffin Ratio

The iso/normal paraffin ratios of various commercial fractions are shown in Figure 5-5.
For LPG, light naphtha, heavy naphtha and kerosene fractions the ip/np ratio decreases
along the reactor. The decrease in the ip/np ratio of all the fractions is quite high in the
first 40% of the reactor with a relatively much slower decrease in the latter part. At a
given temperature, the ip/np ratio is always greater than unity at the exit, is highest for
the lightest fraction and decreases as the average carbon number of the fraction
increases. A drop in this ratio is observed when the reaction temperature is increased.
The decrease in the ip/np ratio with temperature has been reported by Sullivan et al.91 for
C5 and C6 fractions. Contrary to lighter fractions, the ip/np ratio of diesel increases along
the reactor. Bourne et al.92 have reported the increase in the ip/np ratio along the reactor
in their gas oil fraction (437700 F) obtained from the hydrocracking of highly aromatic
Kuwait gas oil. The current simulations showed a slight increase with temperature of this
ratio which can be attributed to a higher rate of isomerization of the long chain normal
paraffins into isoparaffins by the PCP(s;s) steps at the elevated temperature.

97

5.1.6. Effect of Temperature on the Carbon Number Distribution of Hydrocarbon

Classes
As the temperature in increased, the general trend in all the hydrocarbon classes is the
decrease in the yield of heavier hydrocarbons together with an increase in the lighter
ones [Figure 5-6 (a) to (i)]. At both 350 and 370 oC, the normal and isoparaffin yields
show a peak at carbon number 5 [Figure 5-6 (b) and (c)] which is also reflected in the
carbon number distribution of the total reactor effluents [Figure 5-6 (a)]. The peak
around carbon number 25 in different hydrocarbon classes appears because of the
corresponding peak in the feedstock. The heights of these peaks decrease (or the peaks
completely vanish) due to a higher amount of cracking as the temperature is increased.
The peak at carbon number 8-9 in the Figure 5-6 (a) comes from the corresponding
peaks in the carbon number distribution of mononaphthenes [Figure 5-6 (d)], indicating
a preferential cracking in the side chain of the mononaphthenes leaving 2-3 carbon
atoms on the naphthenic ring. Similarly, in dinaphthenes [Figure 5-6 (e)], the peak is
obtained at carbon number 13 leaving 3 carbon atoms on the dinaphthenic structure. The
ring opening of these C13 dinaphthenes creates a slight peak at C13 in the carbon number
distribution of mononaphthenes. Trinaphthenes [Figure 5-6 (f)] in the same way show a
peak at carbon number 17 (=14+3), the ring opening of which creates a small peak at C17
in the carbon number distribution of dinaphthenes. The same reasoning is applicable to
the C21 peak in the carbon number distribution of trinaphthenes. Monoaromatics,
diaromatics and triaromatics [Figure 5-6 (g) to (i)] follow similar trends with peaks at
C9, C13 and C17 respectively, the height of these peaks does not increase with the
increase in temperature because of the faster rate of hydrogenation of these aromatics at
the increased temperature.

5.2. Effect of Pressure

To study the effect of pressure on the VGO conversion and product distribution, the
simulations have been carried out at two different reactor total pressures: 150 bar and
170 bar; both at a reactor inlet temperature of 370 oC and a H2 to hydrocarbon molar

98

ratio of 15.0. The VGO flow rate has been taken as 620 kmol/hr as in the previous
simulations. The various simulation results are shown in Figure 5-7 to Figure 5-11 with
solid lines corresponding to the simulations at 150 bar and the dotted lines at 170 bar.
Figure 5-7 (a) shows the temperature profiles of the solid and liquid phases along the
reactor. As the pressure is increased from 150 to 170 bar, the concentration of hydrogen
in the liquid phase increases which in turn increases the rate of hydrogenation of the
aromatics and naphtheno-aromatics [eqn (3.41)]. The exothermic nature of these
hydrogenation reactions leads to a faster temperature rise in each bed. The total
temperature increase in the reactor at 170 bar is also increased to 63.8 oC against 58.8 oC
at 150 bar.
The effect of reactor pressure on the conversion of VGO is the combined result of
several competing phenomena: (i) As explained above, a higher pressure increases the
rate of hydrogenation of aromatics, thus favoring their cracking and, therefore, the total
VGO conversion. Higher aromatic saturation also increases the reactor temperature,
further enhancing the VGO conversion. (ii) It can be seen from eqn (3.53) that the liquid
phase concentration of hydrogen appears in the denominator of the rate expressions of
the elementary steps belonging to group (a). Therefore, a higher liquid phase hydrogen
concentration reduces the rate of PCP isomerization, acyclic -scission, exo and
endocyclic -scission elementary steps, thereby reducing the VGO conversion with total
pressure. (iii) A high reactor pressure increases the concentration of relatively lighter
hydrocarbons in the liquid phase making them susceptible to secondary cracking.
Increased amount of lighter species in the liquid phase causes a reduction in the liquid
phase mole fraction of the heavier unconverted hydrocarbons leading to a reduction in
their conversion. (iv) An increased liquid phase hydrogen concentration has a favorable
effect on VGO conversion due to enhanced hydrogenation of coke resulting in improved
catalytic actovity67. A high aromatic feedstock has a higher tendency for coke formation,
therefore, an increased pressure will have more positive effect on conversion for
aromatic feedstocks compared to the paraffinic ones.

99

As the model currently does not account for the deactivation of the catalyst due to coke
formation, the simulations results of the effect of pressure on the conversion and overall
product distribution is based on the first three phenomena only. It can be seen from
Figure 5-7 (b) that with the increase in pressure, the total conversion has dropped by
approximately 6% indicating the dominance of the (ii) and (iii) phenomena. High
amounts of paraffins and naphthenes (~ 82 wt%) in the feed supports this behavior.
The increased hydrogen flux in the liquid phase [Figure 5-7 (c)] and reduced
vaporization of hydrocarbons [Figure 5-7 (d)] are self explanatory in light of the
increased reactor pressure.

5.2.1. Effect of Pressure on the Yields of Various Commercial Fractions

Reduced conversion resulted in roughly 2% increase in the yield of diesel with 2% and
4% decrease in the yield of light naphtha and heavy naphtha, respectively. The yield of
kerosene and LPG remains almost constant [Figure 5-8(a)].

5.2.2. Effect of Pressure on the Yields of Various Hydrocarbon Classes

The main effect of pressure occurs on the aromatic and naphtheno-aromatic classes
[Figure 5-8 (c), (d) and Figure 5-9 (a)] due to their increased rate of hydrogenation. The
effect of pressure increase on these classes is similar to that of an increase in
temperature, as discussed previously. Cyclization of the paraffins to form
mononaphthenes depends on the inverse of the square of liquid phase hydrogen
concentration [eqn (3.31)] causing a decrease in cyclization rates with increase in
pressure. It can be seen from the decrease in the yield of mononaphthenes [Figure 5-8
(b)] in spite of faster hydrogenation of monoaromatics to mononaphthenes. The
decreased rate of the ring opening and side chain cracking reactions causes slight
increase in the yields of dinaphthenes, trinaphthenes and tetranaphthenes classes.

100

(a)

(b)

400

100
Solid
Liquid

80
Conversion (Wt %)

Temperature, oC

390
380
370
360

60

40

20
350
0

0.2
0.4
0.6
0.8
Fractional Reactor Length

(c)

(d)
60

800
30

700

25

20

0.2
0.4
0.6
0.8
Fractional Reactor Length

15
1

50
Vaporization (Wt%)

900

H2 Flux liq phase (kmol/hr/m2)

1000
H2 Flux gas phase (kmol/hr/m2)

0.2
0.4
0.6
0.8
Fractional Reactor Length

40
30
20
10
0

0.2
0.4
0.6
0.8
Fractional Reactor Length

Figure 5-7. Effect of reactor pressure on the (a) Temperature profile of solid and liquid phases along the
reactor beds. (b) Weight percent conversion of VGO. (c) Hydrogen flux in the gas and liquid phases along
the reactor beds. (d) Weight percent vaporization of hydrocarbons along the reactor beds. (solid lines: 150
bar, dotted lines: 170 bar).

101

(a)

(b)

50

50
LPG (C3-C4)

NPA

LNAP (C5-C7)

DIES (C16-C20)

30

20

10

MNA
DNA

40
Yield (Wt %)

Yield (Wt %)

IPA

HNAP (C8-C12)
KERO (C13-C15)

40

TNA

30

TETNA

20

10

0.2
0.4
0.6
0.8
Fractional Reactor Length

0.2
0.4
0.6
0.8
Fractional Reactor Length

(c)

(d)

12

10

MAR

NMA

DAR

NDA
DNMA

TAR

1.5

TETAR

Yield (Wt %)

Yield (Wt %)

6
4

0.5
2
0

0.2
0.4
0.6
0.8
Fractional Reactor Length

0.2
0.4
0.6
0.8
Fractional Reactor Length

Figure 5-8. Effect of reactor pressure on the (a) Evolution of different commercial fractions. (b) Evolution
of paraffins and one to four ring naphthenes. (c) Evolution of one to four ring aromatics. (d) Evolution of
naphtheno-aromatics. (solid lines: 150 bar, dotted lines: 170 bar).

102

(a)

(b)

0.4

60
NTA

0.35

DNDA

0.25

Yield (Wt %)

Yield (Wt %)

50

TNMA

0.3

0.2
0.15

PAR

40

NAP
ARO

30

NAP-ARO

20

0.1
10

0.05
0

0.2
0.4
0.6
0.8
Fractional Reactor Length

(c)

0.2
0.4
0.6
0.8
Fractional Reactor Length

(d)

12

12
C4

C20

C8

10

C24

10

C12

C28

Yield (Wt %)

Yield (Wt %)

C16

6
4
2
0

C32

8
6
4
2

0.2
0.4
0.6
0.8
Fractional Reactor Length

0.2
0.4
0.6
0.8
Fractional Reactor Length

Figure 5-9. Effect of reactor pressure on the (a) Evolution of the naphtheno-aromatics. (b) Evolution of
total paraffins, naphthenes, aromatics and naphtheno-aromatics. (c, d) Evolution of C4 to C32
hydrocarbons. (solid lines: 150 bar, dotted lines: 170 bar).

103

(a)

(b)

1.4

10
C3

NPAR

NC4

1.2

IPAR

IC4

MNAP
MARO

Yield (Wt %)

Yield (Wt %)

1
0.8
0.6

0.4
2
0.2
0

0.2
0.4
0.6
0.8
Fractional Reactor Length

0.2
0.4
0.6
0.8
Fractional Reactor Length

(c)

(d)

25

10
NPAR
IPAR

PAR
NAP

MNAP

20

ARO

NAP-ARO

MARO

15

Yield (Wt %)

Yield (Wt %)

DNAP
DARO
NMA

10

0.2
0.4
0.6
0.8
Fractional Reactor Length

0.2
0.4
0.6
0.8
Fractional Reactor Length

(e)

(f)

14

60

12

50

PAR

NAP-ARO

NAP

Yield (Wt %)

Yield (Wt %)

10

ARO

NAP-ARO

6
4

40
30
20
10

2
0

NAP
ARO

PAR

0.2
0.4
0.6
0.8
Fractional Reactor Length

0.2
0.4
0.6
0.8
Fractional Reactor Length

Figure 5-10. Effect of reactor pressure on the (a) Evolution of different components of LPG. (b) Evolution
of different classes of light naphtha. (c) Evolution of different classes of heavy naphtha. (d) Evolution of
different classes of kerosene. (e) Evolution of different classes of diesel. (f) Evolution of different classes
of unconverted fraction. (solid lines: 150 bar, dotted lines: 170 bar).

104

3
LPG (C3-C4)
LNAP (C5-C7)
HNAP (C8-C12)
KERO (C13-C15)

2.5

iP/nP wt. ratio

DIES (C16-C20)

1.5

0.5

0.1

0.2

0.3

0.4
0.5
0.6
0.7
Fractional Reactor Length

0.8

0.9

Figure 5-11. Effect of reactor pressure on the iso to normal paraffins ratio of different commercial
fractions. (solid lines: 150 bar, dotted lines: 170 bar).

105

5.2.3. Effect of Pressure on the Yields of Hydrocarbons of Different Chain Length

Higher pressures lower the total conversion of VGO. The effect of the increase of total
pressure on the hydrocarbons of different chain lengths is similar to that of a decrease in
temperature. This similarity can be observed by comparing Figure 5-9 (c), (d) with
Figure 5-3 (c), (d).

5.2.4. Effect of Pressure on the Yields of Individual Classes of Various Commercial

Fractions
The various components of LPG and all the classes of light and heavy naphtha [Figure
5-10] are reduced due to the reduced conversion with the increase in pressure as all the
components in this carbon number range are only produced by the hydrocracking
reactions. Aromatics and naphtheno-aromatics in the kerosene and diesel fractions are
decreased primarily because of more ring hydrogenation as reflected by the increase in
the amount of naphthenes in these fractions. It can be seen in Figure 5-10 (e) that the
increase in yield of naphthenes in diesel fraction is more than the total decrease in the
yields of aromatics and naphtheno-aromatics. The difference is caused by a lower rate of
cracking reactions of the diesel range naphthenic species which dominates after 70%
length of the reactor at higher pressure as against 60% at the lower pressure. Slower
cracking is also responsible for the increase of the paraffin yield in the kerosene, diesel
and unconverted fractions.

5.2.5. Effect of Pressure on the Iso/Normal Paraffin Ratio

The effect of pressure on the final values of the iso/normal ratio for heavy naphtha and
kerosene fractions is negligible [Figure 5-11]. The ratio has slightly increased for the
LPG and light naphtha fraction and shows a decrease in the diesel fraction as the
pressure is increased. As the rate coefficients of all the reactions in the model are
independent of the total pressure and the rates of the all the elementary steps which can
affect this ratio (except cyclization) vary in the same way (inversely proportional) with
pressure, this effect can not be attributed to the change in relative rates of different

106

elementary steps with pressure. An explanation for this effect can possibly be found in
the evolution of the ip/np ratio along the reactor itself. For all the fractions except diesel,
this ratio decreases with reactor length, indicating the decrease in the ratio with increase
in conversion. As the conversion is decreased with the increase in total pressure, the
ip/np ratio shows some increase with pressure. The opposite is true in the case of diesel.
For heavy naphtha, the ratio is independent of conversion in the last 25% of the reactor
length justifying a negligible effect on its value with pressure.

5.3. Sensitivity of the Distribution of Subclasses in Various Classes

The typical composition of VGO is available at the detail given in Table 2-1 in which
VGO composition is expressed in terms of the carbon number distribution of 16
molecular classes. This kind of VGO composition is converted into a more detailed
composition in which each molecular class is divided into subclasses based on the
number of methyl branches. All the simulations reported here are based on the following
distribution:
Isoparaffins for each carbon number (for which minimum three methyl branches can
exist) are considered to have 65 mole % monobranched, 25 mole % dibranched and 10
mole % tribranched isomers. In the classes with aromatic and naphthenic rings, a
distribution of 20 mole %, 50 mole %, 20 mole % and 10 mole % has been assumed for
the unbranched, monobranched, dibranched and tribranched isomers.
This section is dedicated to the study of the sensitivity of the isomer distribution on the
conversion of VGO and the evolution of the isomers of a given class along the length of
the reactor. The distribution mentioned above is referred as the base case. Two sets of
simulations have been performed. In Set-I, only the distribution of isoparaffins has been
altered while keeping the distribution of all other classes at the base values. In Set-II, the
distribution of isoparaffins has been kept at the base value and that of all other classes is
varied. The distribution of the isomers selected for different cases in Set-I and Set-II and
the corresponding conversions of VGO obtained from the reactor simulations are shown

107

in Table 5-4 and Table 5-5. All the simulations reported in this section are performed at
350 oC, 150 bar and a H2 to hydrocarbon molar ratio of 15.

Table 5-4. The distribution of isoparaffins selected for the simulations of Set I. (mole %).

(Wt %) VGO

Monobranched Dibranched Tribranched

Conversion

Base Case

65

25

10

56.7

Case 1

100

53.2

Case 2

100

62.4

Table 5-5. The distribution of isomers selected for the ring species for the simulations of Set II.
(mole %).

Unbranched Monobranched Dibranched Tribranched

(Wt %) VGO
Conversion

Base Case

20

50

20

10

59.8

Case 3

100

23.7

Case 4

100

54.5

Case 5

100

74.2

108

5.3.1. Analysis of the Distribution of Isoparaffins

The total amount of isoparaffins in the VGO feedstock is only 12.2 weight percent.
Therefore, the effect of the change in the distribution of isoparaffins on VGO conversion
while keeping the distribution of all other classes at the base values is relatively small, as
seen from Table 5-4. It can be observed however that increasing the percentage of
isomers with a higher degree of branching increases the total conversion as the number
of pathways available for cracking steps increases with the increase in the number of
methyl branches in the feedstock. Figure 5-12 shows the change in the mole percent
composition of normal, monobranched, dibranched and tribranched isomers among the
total paraffin fraction along the length of the reactor. It can be seen that the tribranched
paraffins in the base case decrease very fast in the first 20% length of the reactor without
any counter increase in the amount of dibranched isomers. This indicates the fast
cracking of the tribranched paraffins. In both Case 1 and Case 2, the final composition of
isomers approaches the base case values within 5 % range. The amount of isomers with
different degrees of branching are approaching some steady state values in each case,
however, it should be noted that these values are not the thermodynamic equilibrium
values. If all the paraffinic isomers (NPA, MBP, DBP and TBP) of a given carbon
number were able to reach thermodynamic equilibrium among them, the total amount of
tribranched isomer fraction would have been the highest with decreasing amounts of
dibranched, monobranched and normal paraffins, in that order. This can be explained
qualitatively based on the increase in the number of isomers with increased number of
methyl branches and a relatively less Gibbs free energy of formation of isomers with
more methyl branches, particularly if the crowding of methyl branches is absent.
Although the curves shown in Figure 5-12 represent the distribution of isomers for the
entire carbon number range (C3 to C33) and not just for one particular carbon number, it
can be said conclusively that the final distribution is very far from the thermodynamic
equilibrium distribution as seen from maximum amount of monobranched isomers with
very small amounts of di- and tribranched isomers. The deviation of the hydrocracking
products from the thermodynamic equilibrium showing large amounts of monobranched

109

and smaller amounts of multibranched paraffins is well documented34, 35 and is supported

by these simulations. The above explanation also clarifies the need to divide the
isoparaffins for a given carbon number into mono-, di-, and tribranched lumps, given the
lack of thermodynamic equilibrium among them.

60

50

Mole %

40

30

20

10

0.1

0.2

0.3

0.4
0.5
0.6
0.7
Fractional Reactor Length

0.8

0.9

Figure 5-12. Molar composition of the normal, monobranched, dibranched and tribranched isomers
among the total paraffins along the reactor length. (Blue: NPA. Green: MBP. Red: DBP. Cyan: TBP. Solid
lines: Base Case. Dotted lines: Case-1. Dashed lines: Case-2.).

110

5.3.2. Analysis of the Distribution of Isomers in Ring-Containing Hydrocarbon

Classes
In Set II simulations, the distribution of isoparaffins in the feedstock is maintained at
their base values. Three different cases are compared with the base case to analyze the
effect of the distribution of isomers with different number of methyl branches in the
hydrocarbon classes containing one or more of naphthenic or aromatic rings (i.e., class 3
to 16 in Table 2-1. Case 3 simulation corresponds to a VGO feed composition in which
100 % of the isomers in each of classes 3 to 16 are assumed to be unbranched, i.e., not
having any methyl branch anywhere on the ring or along the side chain. Case 4 and Case
4 assumed similarly the isomers to be 100% monobranched and 100 % dibranched,
respectively. The comparison of the three cases with the base case is shown in Figure
5-13 to Figure 5-15.
Similar to the isoparaffins, increasing the isomers with more degree of branching
increases the VGO conversion, however, the increase in conversion is substantial, unlike
with isoparaffins, because of the large amount of ring containing species (73.5 wt %) in
the VGO.
It can be seen from Figure 5-13 (Case 3) that for mononaphthenes and dinaphthenes
classes [Figure (b) and (c)] unbranched fraction decreases quite rapidly to approach the
base case composition with a similar increase in the monobranched isomer fraction. For
the aromatics and naphtheno-monoaromatics, the rate at which the unbranched fraction
decreases and monobranched fraction increases is very low and stays very far from the
base case, even at the end of the reactor. This is probably because of the lower carbon
number range (only up to C22 as against C33 for naphthenes) of these species in the VGO
feedstock, [Table 5-2] as the global rate coefficient for PCP isomerization increases with
the carbon number as can be deciphered from the lumping coefficient plots [Figure 3-5]
for PCP(s;s) mode. Analogous behavior can be seen in Case 4 [Figure 5-14].
In Case 5, when the feedstock contains 100% dibranched isomers, 47% conversion takes
place in only first 13% of the reactor with a sharp decrease in the amount of dibranched

111

isomers in each class [Figure 5-15]. This decrease is caused by the high rate of cracking
of the dibranched isomers as well as to their sequential isomerization into the
monobranched and unbranched isomers through PCP, seen by a rapid increase in the
amount of monobranched and unbranched isomers approaching their base case values.
The rapid increase in the monobranched isomers associated with the cracking of
dibranched isomers can be explained by the fact that the total number of methyl
branches in the products of -scission is always one less than the number of methyl
branches in the reactant, i.e., the two fragments obtained from the cracking of
dibranched mononaphthenes (in the side chain) would have only one methyl branch in
total.
Similar to the Set I simulation results for isoparaffins, it can be observed for ring
containing species also that the final distribution of the hydrocracking products is very
far from their thermodynamic equilibrium with respect to the isomers with different
number of methyl branches, illustrating the requirement to have separate lumps for the
four subclasses for each class per carbon number.
As a summary of this section, the VGO conversion is quite sensitive to the distribution
of the isomers inside the classes. When the distribution is changed for the isomers of
only one class, the change in the conversion is not very large, but the cumulative effect
of the change in the isomer distribution of all the classes becomes significant. The
conversion increases when the amount of isomers with more methyl branches is
increased. The isomer distribution in the products converges to the similar values (within
roughly 10% range) irrespective of the distribution of the isomers in the feedstock,
especially at the high conversions. The final distribution of isomers in a class at the
reactor exit is quite different from the corresponding thermodynamic equilibrium.

112

(a) Paraffins

(b) Mono-Naphthenes

60

100
80
Mole %

Mole %

40

20

60
40
20

0.5

Fractional Reactor Length

80

80

60

60

40
20

40
20

0.5

Fractional Reactor Length

(e) Di-Aromatics

(f) Nap-Mono-Aromatics

100

80

80

60

60

Mole %

Mole %

0.5
Fractional Reactor Length

100

40
20
0

(d) Mono-Aromatics

100

Mole %

Mole %

(c) Di-Naphthenes

100

0.5
Fractional Reactor Length

40
20

0.5

Fractional Reactor Length

0.5

Fractional Reactor Length

Figure 5-13. Molar composition of the unbranched, monobranched, dibranched and tribranched isomers
for different hydrocarbon classes along the reactor length for case-3. (Blue: NPA. Green: MBP. Red: DBP.
Cyan: TBP. Solid lines: Base Case. Dotted lines: Case 3.).

113

(a) Paraffins

(b) Mono-Naphthenes

60

100
80
Mole %

Mole %

40

20

60
40
20

0.5

Fractional Reactor Length

80

80

60

60

40
20

40
20

0.5

Fractional Reactor Length

(e) Di-Aromatics

(f) Nap-Mono-Aromatics

100

80

80

60

60

Mole %

Mole %

0.5
Fractional Reactor Length

100

40
20
0

(d) Mono-Aromatics

100

Mole %

Mole %

(c) Di-Naphthenes

100

0.5
Fractional Reactor Length

40
20

0.5

Fractional Reactor Length

0.5

Fractional Reactor Length

Figure 5-14. Molar composition of the unbranched, monobranched, dibranched and tribranched isomers
for different hydrocarbon classes along the reactor length for case-4. (Blue: NPA. Green: MBP. Red: DBP.
Cyan: TBP. Solid lines: Base Case. Dotted lines: Case 4.).

114

(a) Paraffins

(b) Mono-Naphthenes

60

100
80
Mole %

Mole %

40

20

60
40
20

0.5

Fractional Reactor Length

80

80

60

60

40
20

40
20

0.5

Fractional Reactor Length

(e) Di-Aromatics

(f) Nap-Mono-Aromatics

100

80

80

60

60

Mole %

Mole %

0.5
Fractional Reactor Length

100

40
20
0

(d) Mono-Aromatics

100

Mole %

Mole %

(c) Di-Naphthenes

100

0.5
Fractional Reactor Length

40
20

0.5

Fractional Reactor Length

0.5

Fractional Reactor Length

Figure 5-15. Molar composition of the unbranched, monobranched, dibranched and tribranched isomers
for different hydrocarbon classes along the reactor length for case-5. (Blue: NPA. Green: MBP. Red: DBP.
Cyan: TBP. Solid lines: Base Case. Dotted lines: Case 5.).

115

CHAPTER VI
6.KINETIC MODEL DEVELOPMENT FOR THE HYDROCRACKING
OF PURE PARAFFINS
6.1. Introduction
Hydrocracking is a process of considerable flexibility and can be used for the conversion
of a wide range of feedstocks into a variety of desired products.3 In addition to the
hydrocracking of vacuum gas oil (VGO) and other refinery residues, hydrocracking is
also applied to crack the Fischer-Tropsch products for the

production of middle

distillate of very high quality.93, 94 The synthetic diesel produced by this process has a
cetane number of more than 74 with zero sulfur content.95,

96

This has motivated the

development of a detailed model for the hydrocracking of pure paraffins with a more
fundamental approach than considered for the VGO hydrocracking in Chapters III to V.
Hydrocracking of paraffins is carried out on bifunctional catalysts with a metal function,
generally provided by Pt to catalyze the dehydrogenation into olefins and an acid
function catalyzing isomerization and cracking. The acid function is often provided by
Y-zeolites because of their wider pore structure that minimizes the diffusion resistance
for the bulky molecules of heavy feedstocks.
It has been shown in several studies11, 17, 97 on hydroisomerization and hydrocracking of
model compounds ranging from C8 to C12 paraffins that on a well-balanced catalyst (i.e.
having sufficient metal activity), hydrogenation/dehydrogenation steps are intrinsically
much faster than the elementary steps taking place on the acid sites. As a result, the
hydrogenation/dehydrogenation reactions attain quasi-equilibrium and the elementary
steps on the acid sites are the rate determining steps, represented by Ac-rds. Under
these conditions, the concentration of a particular olefin intermediate depends only on
the concentration of the corresponding saturated paraffin and monobranched feed
isomers are the only primary products, whereas multibranched and cracked molecules
are secondary products. The concentration of feed isomers is high and secondary

116

cracking is relatively limited. On the other hand, if the catalyst does not have sufficient
metal activity to equilibrate the hydrogenation/dehydrogenation steps, the situation
becomes considerably more complex. Under these so called non-ideal hydrocracking
conditions, the concentration of a particular olefin intermediate depends on the
concentration of all the paraffins present in the reaction mixture.98 When the rate
determining steps shift from acidic sites to both the acid and metal sites of the catalyst
(represented as Me-Ac-rds) multi-branched feed isomers and cracked products become
primary products of hydrocracking. It also results in increased cracking selectivities and
more secondary cracking. Degnan and Kennedy98 studied this phenomenon by
conducting experiments on n-heptane hydroisomerization on physical mixtures of beta
zeolite and Pt/Al2O3 particles by varying the amount of metal containing component.
They reported that the cracking vs. isomerization selectivity varies significantly with the
metal loading up to a point beyond which the degree of isomerization is independent of
the metal loading. They interpreted their results using a simplified dual-site model.
Debrabandere and Froment99 studied the gas-phase hydrocracking of C8 to C12 nparaffins on a Pt/USY zeolite for a variety of operating conditions. For a catalyst with
given metal to acid activity, high temperatures, low pressures and high hydrogen to
hydrocarbon ratios shift the rate determining step from acidic sites to the metal sites of
the catalyst. These effects are more pronounced when the reacting hydrocarbons have
longer side chains. Thybaut et al.100 explained this behavior in gas-phase hydrocracking
of paraffins using a lumped reaction scheme and simplified analytical expressions.
In the work reported here, a detailed mechanistic kinetic model has been developed for
the hydrocracking of heavy paraffins in which the rate determining step is assumed to
occur on both acid and metal sites of the catalyst. The generalized nature of the model
also makes it applicable to the specific case in which the hydrogenation/dehydrogenation
steps are at equilibrium, which has been conventionally referred to as ideal
hydrocracking or Ac-rds. Because of the fundamental nature of the rate parameters,
this model can be used for the hydrocracking of mixtures of paraffins.

117

6.2. Kinetics of Hydrogenation/Dehydrogenation Reactions

Hydrogenation/dehydrogenation

reactions

are

modeled

using

molecular

dehydrogenation mechanism and are represented by the following sequence of steps

occurring on the metal sites:
sorption

Pi liq
Pi sorbed

(K )

(i)

(K )

(ii)

L , Pi

( zeolite )

Pm
Pi sorbed + m

C , Pi

kdh

(K

ij

Oij m + H 2 m
Pm
+ m

kdh ij

(K

Oijsorbed + m
Oij m

sorbed

Oij

+H

C ,Oij

pr / depr

R+

ik

ads
DH ,ij

(iii)
(iv)
(v)

HS / MS / PCP
Rik+
Ril+

(vi)

Cracking
Rik+
R +jm + Ouvsorbed

(vii)

H 2liq + m
H 2 m

(K )

desorption

Oijliq
Oijsorbed

( zeolite )

(viii)

C ,H2

(K )
L ,Oij

(ix)

Scheme 1. Reaction mechanism for metal and acid sites steps considered in the model

The elementary steps taking place on the acid sites have also been listed in Scheme 1.
When the rate determining step of the hydrogenation/dehydrogenation sequence on the
metal site is the surface dehydrogenation step, i.e., step (iii), the rate of formation of
olefin Oij is given by

118

Form
rOij
, metal =

K C ,Oij COij K C , H 2 CHliq2

kdhij K C , Pi [CPi ]
ads

K DHij

c2
m

liq
1 + K C , Pi [CPi ] + K C ,Oij COij + K C , H 2 CH 2
i
ij

(6.1)

The square brackets represent concentrations of the sorbed species in the zeolite pores.
In a three-phase reactor operating in the trickle flow regime, it is assumed that the entire
external catalyst surface is covered with a liquid film of hydrocarbons and the sorption
of hydrocarbons into the pores of the zeolite proceeds from the liquid phase. It is further
assumed that hydrogen is directly chemisorbed from the liquid phase onto the metal sites
of the catalyst. The concentration of the sorbed species can be eliminated in terms of the
observable liquid phase concentrations using the following expression:

[CPi ] =

H Pi CPiliq
DL

(6.2)

in which Henrys coefficient for the sorption of paraffin Pi in the pores of the zeolite is
given by

liq
H Pi = K L , Pi csat , Pi and DL = 1 + K L , PiCPiliq + K L ,Oij COij

i
ij

(6.3)

By using equation (6.2) and a similar eqn for olefins, eqn (6.1) can be written as

Form
rOij
, metal

liq

K C ,Oij H Oij C Oij

K C , H 2CHliq2 2
liq
kdhij K C , Pi H PiCPi
cm
ads
K DH

,ij

=
2
DC DL

K H C liq
K H C liq
where DC = 1 + C , Pi Pi Pi + C ,Oij Oij Oij + K C , H 2 CHliq2
DL
DL
i
ij

(6.4)

(6.5)

119

The equilibrium constant for the overall dehydrogenation reaction in liquid phase, i.e.,
liq
K DH
can be expressed in terms of the equilibrium constant for the surface
,ij

dehydrogenation step as
K

liq
DHij

ads
K L , Pi K C , Pi K DHij

(6.6)

K L ,Oij K C ,Oij K C , H 2

Considering that the difference in the saturated liquid molar volume of a paraffin and the
corresponding olefin is negligible, the saturation concentration of paraffins in the pores
of zeolite csat , Pi can be assumed to be the same as that of the corresponding olefin csat ,Oij .
Eqns (6.4) and (6.6) can be combined to give the rate of formation of olefin Oij on the
metal sites as,

Form
rOij
, metal

liq liq
liq COij
CH
k dhij K C , Pi H Pi CPi liq 2 cm2
K DH ,ij

=
DC2 DL

(6.7)

liq
6.3. Expressing K DH
,ij in Terms of the True Thermodynamic Equilibrium

Coefficient
The true thermodynamic equilibrium coefficient K DH for the dehydrogenation reaction
Pi = Oij + H 2 can be expressed in terms of the activities of the reactants and products as
follows:

K DH

f
(
=

Oij

o
f Oij

( f

)( f

Pi

H2

f Pio

f Ho 2

(6.8)

Considering ideal gas at 1 atm (=Po) to be the standard state for all the components, the
liquid-phase fugacities of reacting species can be written as:

fPi = xPi Pi f PiL (T , P)

L
fOij = xOij Oij f Oij
(T , P )

(6.9)
(6.10)

120

fH 2 = xH 2 H H 2 (T , P )
yielding

K DH

xOij xH 2 Oij
=

xPi Pi

(6.11)
VLE
L
f Oij (T , P ) K H 2

L
f Pi (T , P ) Po

(6.12)

Therefore, the equilibrium coefficient relating the liquid-phase concentrations of

reacting species can be expressed in terms of the true gas-phase thermodynamic
equilibrium coefficient by

liq
DH

liq liq
C Oij
CH 2
Pi

=
=
K

DH
CPiliq
Oij

f PiL (T , P ) Po
L
VLE
f Oij (T , P) K H 2

liq
Ctotal

(6.13)

It has been assumed that the ratio of the liquid phase fugacities of pure components and
the ratio of activity coefficients of a paraffin and corresponding olefin are close to one,
resulting in

Po
liq
K DH
= K DH VLE
KH
2

liq
Ctotal

(6.14)

The gas-phase equilibrium coefficient is calculated using Bensons group contribution

method65 and the Henry coefficient of hydrogen in the liquid phase is calculated using
the Peng-Robinson equation of state66.

6.4. Modeling of the Metal Site Rate Parameters

Eqn (6.7) contains the rate coefficients for the dehydrogenation reactions, k dhij . For a
particular paraffin molecule Pi, several dehydrogenation rate coefficients k dhij may be
required, depending upon the number and structure of the olefinic isomers Oij produced
from Pi. This leads to a large number of dehydrogenation rate coefficients, even for the
hydrocracking of a simple paraffinic feedstock. For example, for n-decane, 411
hydrogenations / dehydrogenations are required to describe its hydrocracking at the
fundamental level. The number of possible hydrogenations/dehydrogenations increases
exponentially with the number of carbon atoms in the feedstock. The total number of

121

hydrogenations/dehydrogenations taking place in the hydrocracking of a feedstock with

n carbon atoms in the heaviest molecule is obtained by the cumulative sum of all the
olefinic isomers up to carbon number n. Given the large number of olefinic isomers and
associated hydrogenation/dehydrogenations, it is obvious that some simplification is
required to reduce the number of hydrogenation/dehydrogenation rate parameters.
Baltanas et al.101 considered one rate coefficient per carbon number in their model,
leading to 9 independent rate coefficients for the hydrocracking of C16 and resulting in
18 temperature independent parameters, A and E. The number of these parameters keeps
on increasing linearly with the carbon number of the feedstock. Given the highly nonlinear nature of the kinetic model, it is obvious that the meaningful estimation of such a
large number of parameters may be subject to considerable errors and would be
practically impossible for heavier feedstocks. A more fundamental approach is required
to model the rate coefficients for dehydrogenation reactions with a minimum number of
truly independent parameters. The number of these independent parameters and their
values should not change with the carbon number and composition of the feedstock. In
the methodology proposed here, it is assumed that the activation energy for a particular
dehydrogenation step depends on the ease with which the hydrogen atoms can be
removed from two neighboring carbon atoms of a paraffin molecule. The energy
required to remove a hydrogen atom depends on the nature of the parent carbon atom:
primary, secondary or tertiary102. The activation energies for the dehydrogenation of
paraffins would then depend on the location of the double bond in the product olefin, or
in other words, on the nature of the carbon atoms forming the double bond in the olefin.
Consequently, all the dehydrogenation rate coefficients can be modeled using five
different activation energies depending upon the nature of the resulting double bonded
carbon atoms, i.e., (p-s), (p-t), (s-s), (s-t) and (t-t). A single value of the pre-exponential
factor is used for all the dehydrogenation modes. Although the corresponding five
fundamental rate coefficients are independent of the number of carbon atoms in the
reactant, the effective rate coefficient for the dehydrogenation of a particular paraffinic
molecule increases with the number of carbon atoms. Figure 6-1 shows the relative

122

importance of the different modes of hydrogenation/dehydrogenation steps as a function

of the carbon number. The percentage of (s-s) mode is highest and increases with the
carbon number. The percentage of (t-t) mode is the lowest and decreases with the carbon
number.

100
p-s
p-t
s-s

80

s-t
t-t

Percent

60

40

20

0
0

10

15

20

25

30

35

Carbon Number

Figure 6-1. Importance of different dehydrogenation modes with respect to carbon number.

6.5. Modeling of the Acid Site Rate Parameters

The single-event concept11, 36 has been applied to model the frequency factors for the
acid site steps. Refer to Chapter III for the theory of single event kinetics. The activation
energies of all these steps are modeled using the linear free-energy relationship of
Evans-Polanyi103. The five different kinds of elementary steps in paraffins hydrocracking

123

namely, protonation/deprotonation, hydride shift, methyl shift, PCP and - scission are
named as five different types. In the VGO hydrocracking model it was assumed that
the energy level of a particular carbenium ion is determined only by its nature, i.e.,
secondary or tertiary. Therefore, it was generally considered that all the elementary steps
of a particular type, in which a reactant carbenium ion of nature m is converted into a
product carbenium ion of nature n, have the same value of activation energy (m and n
can be -s or t) *. As a result of this, for one type of elementary step, a total of four
1

subtypes and consequently four different activation energies were used.

Considering that the energy of a carbenium ion depends only on its nature is equivalent
to saying that the energy of carbenium ions depends only on the number of carbons in
alpha position with respect to the positive charge, because all the secondary carbenium
ions have two alpha carbons and the tertiary carbenium ions have three alpha carbons. It
can be seen, however, that there is a significant difference in the energies of carbenium
ions with the same carbon number and same nature, depending upon the possible amount
of charge delocalization in the vicinity of the charge carrying carbon atom. Hence, the
energy of carbenium ions of a particular carbon number not only depends on the number
of alpha carbon atoms but also on their nature, although the latter effects are much
weaker.
To eliminate the effects of the above assumption, the Evans-Polanyi relationship has
been introduced into the single-event kinetic model for the hydrocracking of paraffins, as
done by Park & Froment74 for the simpler case of methanol to olefin (MTO) process and
Martinis and Froment73 for the alkylation of isobutane with butenes. The Evans-Polanyi
approach requires the calculation of the heat of formation of every individual carbenium
ion to get the heat of reactions of elementary steps. The activation energy of a particular
step is then calculated by using the linear free energy relationship of Evans and Polanyi.

* From here on, all these steps will be referred to as belonging to one subtype.

124

6.5.1. Evans-Polanyi Relationship

According to the Evans-Polanyi relationship, the activation energy varies linearly with
the energy change associated with the elementary step as,

E = Eo + (H rs )

if H rs > 0

(6.15)

E = Eo + (1 )(H rs ) if H rs < 0

where is the transfer coefficient and Eo is the intrinsic activation energy for the
elementary steps of the given type. In the above equation, H rs is the heat of reaction at
the catalyst surface for the particular elementary step. Consequently, to model the
activation energies of all the elementary steps of one type, only two independent
parameters, Eo and are required, provided the heat of individual elementary steps at
the surface of the catalyst are available.

6.5.2. Rate Equations for the Acid Site Steps

Using the single-event approach and the Evans-Polanyi relationship, the rate of a PCP
elementary step, e.g., Rik+ (m) R +jl (n) is given by

rpcp ,( m;n )

'
Eo , pcp + pcp
H rs


= ne, pcp Apcp exp

RT

'
where pcp
= pcp

if H rs > 0

'
= (1 pcp ) if H rs < 0
and pcp

CRik+

(6.16)
(6.17)
(6.18)

The concentration of surface carbenium ion, CR+ is determined using the protonation
ik
equilibrium coefficient between the sorbed olefins and the surface carbenium ions. i.e.,

C + = K pr COij H +

Rik

(6.19)

To get the value of the protonation equilibrium coefficient, the entropy of protonation

S prs and the heat of protonation H prs at the surface are required. If the effect of the

125

structure of the reacting olefin and the product carbenium ion is accounted for using the
single-event concept, a single value for the intrinsic entropy of protonation, S pr would
be sufficient to obtain the entropies of all the protonation steps. The heat of protonation
is calculated from the heats of formation of the involved species according to

H prs = H sf ( R + ) H sf ( H + ) H sorbed
(O)
f

(6.20)

In the above equation , the superscript s refers to the surface species. In three phase
2

hydrocracking, consideration of the effective latent heat of the carbenium ion is

required to get the heats of formation in the liquid phase from the gas phase heats of
formation. The energy changes associated with the phase transition for a carbenium ion
and olefin are shown in Figure 6-2. The heats of formation of surface carbenium ions are
obtained by subtracting the heat of stabilization of these species from their liquid phase
heats of formation. i.e.,

H sf ( R + ) = H gf ( R + ) H vap ( R + ) H stab ( R + )

Figure 6-2. Energy changes associated with phase transition of (a) carbenium ion (b) olefin.

For the sake of simplicity, the subscripts of the olefins and carbenium ions have been dropped.

(6.21)

126

The heat of stabilization of carbenium ions H stab ( R + ) can be written as the sum of the
heat of sorption of carbenium ions from liquid phase into the zeolites pores

H sorp ( R + ) and the heat of chemisorption from sorbed phase to the surface q ( R + ) i.e.,
H stab ( R + ) = H sorp ( R + ) + q( R + )

(6.22)

Dumesic et al.72 considered a constant value for the energy of stabilization of carbenium
ions, q ( R + ) regardless of the nature of the carbenium ion. Martinis and Froment73
distinguished between the energy of stabilization of secondary and tertiary ions They
also accounted for the effect of chain length on the stabilization energy. The latter effect,
however, was found to be very small and has been neglected in the current model.
The stabilization of an olefin from the liquid phase to the zeolite pores involves only the
heat of sorption H sorp (O ) so that,

H sorbed
(O) = H gf (O) H vap (O) H sorb (O)
f

(6.23)

The heat of stabilization of a proton from the gas phase to the surface is represented by
q ( H + ) so that,

H sf ( H + ) = H gf ( H + ) q( H + )

(6.24)

Substituting eqns (6.21), (6.22), (6.23), and (6.24) into eqn (6.20) and assuming that the
latent heats of vaporization of an olefin and the corresponding carbenium ion are
identical , the heat of protonation becomes:

H prs = H gf ( R + ) H gf ( H + ) H fg (O ) + [q ( H + ) q ( R + )]
H sorp ( R + ) H sorp (O )

(6.25)

The last bracket in eqn (6.25) is the difference in the heat of sorption of a carbenium ion
and the corresponding olefin. For the same number of carbon atoms and the same
skeleton structures of these two species, the difference in their heats of sorption is
assumed to be negligible. The middle bracket is the difference in the heats of

127

chemisorption of a proton and the heat of chemisorption of a carbenium ion and is

represented by q ( R + ) 74. Depending upon the nature of the carbenium ion, it can have
two values for a catalyst with given acid strength and can be written as q (m) , where
(m) can be (s) for secondary or (t) for the tertiary ions. Eqn (6.25) is now written as
H prs = H fg ( R + ) H fg ( H + ) H fg (O ) + q (m)

(6.26)

The gas phase heat of formation of a proton is known to be 367.2 kcal/mol72 at 298 K.
The gas phase heats of formation of olefins at 298 K are obtained by using Bensons
group contribution method. To obtain the gas phase heats of formation of carbenium
ions at 298 K, a Benson like group contribution method has been developed.
Experimental data on the heats of formation of a number of paraffinic carbenium ions
have been taken from Lias et al.71 Quantum chemical calculations have been performed
using Gaussian 03 to obtain the heats of formation of the ions not available in the
literature. G3 theory104 has been used for ions having up to 5 carbon atoms and
G3MP2105 theory for ions with 6 carbon atoms. Secondary and tertiary ions are regressed
separately because of the different amount of charge delocalization in the vicinity of the
carbon atom carrying the positive charge.
To derive the heat of protonation at the reaction temperature T from the standard state
temperature of 298 K, the difference in the heat capacities of the reactant and product
species, CPrxn is required. Given the similar skeleton structures of the reactant and
product species, it is assumed that CPrxn is negligible. Consequently, K pr at the reaction
temperature T can be given by applying the single-event approach as,
Oij

K pr =

gl

Rik+

gl

S pr
exp
R

H prs
exp
RT

(6.27)

Eliminating H prs in terms of the gas phase heat of protonation H prg the following
equation is obtained for the surface concentration of carbenium ions,

128

O
S
ij
C + = gl+ exp pr
Rik Rik
R
gl

H prg + q (m)
+
exp
COij H

RT

(6.28)

Svoboda et al.16 showed that in the hydrocracking of paraffins the total concentration of
the carbenium ions is negligible with respect to the total concentration of the acidic sites.
More recently, by taking typical values for the protonation equilibrium coefficients and
concentration of sorbed olefins, Thybaut et al.106 showed that at 506 K, the concentration
of the carbenium ions is approximately 7.10x10-4 mol/kgcat, to be compared with a total
concentration of acid sites, which is typically of the order of 0.1-0.8 mol/kgcat.
Substituting the surface concentration of carbenium ions from eqn (6.28) into (6.16) and
replacing H + by Ct , the rate of the PCP step can be written as,
S pr
gl
= ne , pcp R A
pcp Ct exp
R
gl

Oij

rpcp ( m;n )

+
ik

'

H prg + q (m) + Eo , pcp + pcp

H rs
exp

RT

COij

(6.29)

In the above equation, the heat of reaction of a PCP step is given by

H rs = H sf ( R +jl ) H sf ( Rik+ )
= H gf ( R +jl ) H vap ( R +jl ) H sorp ( R +jl ) q ( R +jl )

(6.30)

H gf ( Rik+ ) H vap ( Rik+ ) H sorp ( Rik+ ) q ( Rik+ )

Assuming the difference in the latent heat of vaporization and heat of sorption of the
reactant

and

product

carbenium

ions

to

be

negligible

and

adding

and

subtracting q ( H + ) in the right hand side of equation (6.30), the heat of reaction of the
PCP step can be written as,

H rs = H rg + [ q(n) q(m)]

(6.31)

where n and m represent the nature of the product and reactant carbenium ions,
respectively. Substituting equation (6.31) into eqn (6.29) leads to,

129

S pr
gl
= ne , pcp R A
pcp Ct exp
R
gl

Oij

rpcp ( m;n )

+
ik

comp

E pcp

( m;n )
C
exp

Oij

RT

comp
g
'
g
where E pcp
( m ; n ) = H pr + q ( m) + Eo , pcp + pcp ( H r + [ q ( n ) q ( m ) ] )

(6.32)

(6.33)

is the composite activation energy for the PCP step. In eqn (6.32) the sorbed
concentration of olefin is expressed by a Langmuir isotherm and is given by an equation
analogous to eqn (6.2).

6.6. Late Thermodynamic Lumping of Components

Studies97, 107, 108 on hydroisomerization and hydrocracking of model compounds from C8
to C12 reveal, that the isomers of any carbon number fraction with a particular degree of
branching reach thermodynamic equilibrium because the HS and MS steps are much
faster as compared to the PCP steps. Therefore, all the isomers for a particular carbon
number with a given degree of branching are placed in one lump with the distribution of
isomers in the lump governed by thermodynamic equilibrium. With this late strict or
thermodynamic lumping, the model requires four lumps/pure components per carbon
number for paraffins and four lumps per carbon number for olefins. For paraffins, these
lumps/pure components are designated as n-paraffins, mono-branched paraffins, dibranched paraffins and tri-branched paraffins and, similarly for olefins, n-olefins, monobranched olefins, di-branched olefins, and tri-branched olefins. A n-paraffin, is
obviously not a lump but a pure component. Based on the above late lumping scheme,
the kinetic model for the hydrocracking of a paraffin mixture of up to C16 requires a total
of 48 lumps/pure components for paraffins and 48 lumps (only C3 will be a pure
component in the case of olefins) for olefins.

6.7. Calculation of the Global Rate of Conversion of Lumps

The global rate of consumption of a particular paraffinic lump is the sum of the rates of
consumption of all the isomers of that lump through dehydrogenation. Similarly, the rate
of consumption of a particular olefinic lump is calculated by summing up the rate of

130

consumption of all the olefinic isomers of that lump through all the possible
hydrogenation steps. To carry out these sums requires going through the entire set of
computer generated hydrogenation/dehydrogenation steps, to calculate the equilibrium
distribution of the isomers of every lump, and the thermodynamic equilibrium
coefficients for all the hydrogenation /dehydrogenation steps. The global rate of
conversion of the olefinic lumps through acid site elementary steps is obtained in the
same way, but it also requires the heats of reaction for all the PCP and -scission steps as
well as the heats of protonation. This is computationally very demanding for heavy
feedstocks, so that a systematic method was developed to calculate the so called
lumping coefficients for the steps on the metal and the acidic sites of the catalyst. Their
values are pre-calculated at a number of temperature intervals and stored in the computer
hard disk, to be used later in the parameter estimation and reactor simulations.

6.8. Lumping Coefficients for the Steps on the Metal Sites

Eqn (6.7) expresses the rate of formation of a particular olefin on the metal sites of the
catalyst. This equation is used to calculate the net rate of formation of a particular
paraffinic lump Pg, by summation over the rate of formation of every isomer Pi of this
lump as,
Form
Pg , net

KC , Pg H Pg cm2 5
liq 5
liq liq
=
k
LCP
C

k
LCO

dh , q
g ,q
Pg
dh , q
g , q COg C H 2
DC2 DL q =1

q =1

(6.34)

In the above equation, q represents the index for the mode of dehydrogenation and g for
lump number. LCPg , q is the Lumping Coefficient for the consumption of Paraffins of
lump g through the qth mode of dehydrogenation and LCOg ,q is the Lumping Coefficient
for the consumption of Olefins of lump g through the qth mode of hydrogenation. The
value of LCPg , q is calculated as
paraffin
isomers

LCPg ,q =

i ,q

i =1

yPieqm

(6.35)

131

where the index i runs over all the paraffin isomers of lump g. Further, ni ,q is the number
of dehydrogenation reactions for the particular paraffin Pi through qth mode and yPieqm is
the equilibrium mole fraction of the isomer Pi in the lump g. Similarly, the value of
LCOg ,q is calculated as
paraffin
isomers

LCOg , q =

ni ,q

eqm
yOij

K
i =1

j =1

liq
DH ,ij

(6.36)

eqm
as the equilibrium mole fraction of the olefin isomer Oij in the lump Og.
with yOij

6.9. Lumping Coefficients for the Steps on the Acid Sites

Equation (6.32) gives the rate of a PCP elementary step in which an m type of carbenium
ion is converted into an n type of carbenium ion. Since the protonation/deprotonation
steps are assumed to be in pseudo-equilibrium, the rate of consumption/formation of
carbenium ions would also be equal to the rate of consumption/formation of the
corresponding olefinic species. Therefore, to get the global rate of conversion of an
olefinic lump, g, into another olefinic lump, h, through PCP steps of subtype (m;n), a
summation has to be carried out over the rate of all such steps, yielding
R

( g ;h )
pcp ( m ;n )

S
= A
pcp Ct exp pr
R

E comp

glij
pcp ( m ;n )

C
ne, pcp R+ exp
Oij
ik
steps
RT

gl

(6.37)

It can be seen from eqn (6.33) that the composite activation energy involves the heat of
protonation of the reacting olefin and heat of reaction of the corresponding elementary
step so that the composite activation energies for different PCP (m;n) subtype of steps
would be different.
To formulate the lumping coefficients for the acid site steps, the composite activation
energy as given in eqn (6.33) is split into two parts. i.e.,
comp

E pcp
( m ;n ) = E pcp ( m ;n ) 1 + E pcp ( m ; n ) 2

(6.38)

132

where E pcp ( m;n ) = H prg + 'pcp H rg

1

(6.39)

and E pcp ( m;n ) = Eo , pcp + q (m) + 'pcp ( q (n) q (m) )

2

(6.40)

As stated above, the value of the terms in E pcp ( m;n ) is different for each elementary
1
step and their evaluation requires the estimation of the heats of formation of individual
carbenium ions and olefins. The terms in E pcp ( m;n ) , on the other hand, depend only on
2

the nature of the reactant and product carbenium ions and are identical for all the PCP
(m;n) subtype of steps. By splitting the composite activation energy in this way and

taking E pcp ( m;n ) out of the summation, eqn (6.37) can be written as,
2

( g ;h )
pcp ( m ;n )

S
= A
pcp Ct exp pr
R

E pcp ( m;n )
E pcp ( m ;n )

glij
2
1
ne , pcp + exp
COij
exp

steps

RT
RT

glRik

(6.41)

Expressing the concentration of the olefin isomer Oij in terms of the concentration of the
eqm
olefinic lump g by using its equilibrium mole fraction yOij
, the global rate of conversion

of the olefinic lump g into lump h by PCP (m;n) subtype of steps can be written as:

( g ;h )
( g ;h )

C exp S pr
R pcp
=
LC
A
( m;n )
pcp ( m ;n ) pcp t
R

E pcp ( m;n )

2
C
exp

g
RT

(6.42)

( g ;h )
where LC pcp
( m ; n ) is the corresponding lumping coefficient given by,

LC

( g ;h )
pcp ( m ; n )

E pcp ( m;n )
glij

1 eqm
= ne, pcp R+ exp
y Oij

RT

glik
steps

(6.43)

A similar approach is applied to formulate the rate of consumption/formation of a lump

through cracking steps. The lumping coefficients for PCP/cracking are functions of pcp ,

cr , and temperature and do not depend on any other model parameter.

133

In eqn (6.42), the concentration of the olefinic lump g in the sorbed phase is eliminated
in favor of its concentration in the liquid phase so that,

( g ;h )
( g ;h )

C exp S pr
R pcp
=
LC
A
pcp ( m ;n ) pcp t
( m;n )
R

E pcp ( m;n ) H C liq

2 Og Og
exp

DL
RT

(6.44)

Given the form of the rate expressions in eqn (6.44), the separate estimation of the
single-event frequency factor for PCP/cracking and the entropy of protonation is not
possible. Therefore, composite single-event frequency factors for PCP and cracking
steps are defined as,
S
comp
A
pcp
= A
pcp Ct exp pr
R

(6.45)

S pr
A
crcomp = A
cr Ct exp
R

(6.46)

Introducing these composite frequency factors, eqn (6.44) becomes after rearranging:
E

( g ;h )
( g ;h )

comp H Og exp pcp ( m ;n ) 2 C liq
R pcp
=
LC
A
( m;n )
pcp ( m ;n ) pcp

Og
RT
DL

(6.47)

6.10. Net Rate of Formation of Paraffinic and Olefinic Lumps

A paraffinic lump/pure component can only be consumed or produced on the metal sites
of the catalyst. Therefore, the net rate of formation of a paraffinic lump Pg is directly
given by eqn (6.34). Olefinic lumps, on the other hand are produced/consumed on both
the metal and acid sites. The rate of formation of a particular olefinic lump Og on the
metal sites is given by the negative of the rate of formation of the corresponding
paraffinic lump Pg, i.e.,
Form
Form
ROg
, metal = RPg , metal

(6.48)

134

The rate of formation of an olefinic lump Og on the acid sites is obtained by summing up
the rate of formation of this lump from all the other olefinic lumps by PCP/cracking
steps. The net rate of formation of lump Og is then given by
Form
Form
Form
ROg
, net = ROg , metal + ROg , acid

(6.49)

Nlumps
Form
where ROg
, acid =

R
i =1

(i ; g )
pcp ( m ;n ) / cr ( m ; n )

(6.50)

( m;n )

The summation over (m;n) implies summation over all the four subtypes namely,
(s;s),(s;t),(t;s) and (t;t) of PCP and cracking steps.

The rate of formation of hydrogen is written as the sum of the rate of formation of
olefins from the hydrogenation/dehydrogenation reactions on metal sites, i.e.,
Nlumps

Form
H 2 , net

= RPiForm
, net

(6.51)

i =1

6.11. Reactor Model for Three-Phase Hydrocracking

Similar to the continuity equations developed in Chapter IV for the hydrocracking of
VGO, continuity equations are written for the gas phase paraffinic components/lumps
and hydrogen as
gas
gas
CPg

1 dFPg
liq
g = 1,2,.........N lumps
= kO , Pg av C ,eqm CPg
K

dz
Pg

(6.52)

The continuity equations for paraffinic components in the liquid phase also take into
account the net rate of formation of component/lump g along with the gas-liquid flux
term,
liq
gas
CPg

1 dFPg
liq
Form
+ RPg
= kO , Pg av C ,eqm CPg
, net g = 1,2,.........N lumps
K

dz
Pg

(6.53)

As the equilibrium coefficient for the dehydrogenation of paraffins is of the order of

10-4-10-5 and the reactor is operated under high hydrogen partial pressures, the

135

concentration of olefins are several orders of magnitude smaller than the corresponding
paraffins. Therefore it has been assumed that the olefinic components exist only in the
liquid phase. Consequently, the continuity equations for olefinic lumps/components do
not contain the mass-transfer term, i.e.,
liq
1 dFOg
Form
= ROg
g = 1,2,.........N lumps
, net
dz

(6.54)

For three-phase hydrocracking of n-hexadecane, a total of 48 lumps/pure components for

paraffins and the same number of lumps/components for olefins is required. As a result,
the model consists of 97 (= 48+48+1) continuity equations for the liquid phase
components and 49 (= 48+1) continuity equations for the gas-phase components,
including hydrogen. The set of ordinary differential equations defined above is solved
for the initial conditions given by the feed composition:
FPggas

z =0

FPgliq

z =0

FOgliq

z =0

= FPggas,o
= FPgliq,o
=0

g = 1,2,......N lumps

Approximate analytical solutions have been obtained to see that the difference in the
eigen values of ODEs for a paraffin and corresponding olefin are of the order of
kdh K dheqm . Such a large difference in the eigen values leads to a stiff set of ODEs which

has to be integrated using backward differentiation formula (BDF), e. g. Gears method.

6.12. Model Degeneration into Equilibrated (de)Hydrogenation Case

The above approach automatically allows the model to also predict the product
distribution for a catalyst which has sufficient metal activity, i.e., when the
hydrogenation/dehydrogenation reactions are equilibrated. When the ratio of rate
coefficients of acid site steps (PCP/cracking) to the rate coefficients of dehydrogenation
steps tends to zero, the set of ODE representing the continuity equations for the liquid

136

phase paraffins and olefins degenerates into the set of ODEs given by eqn (6.55) and
algebraic equations given by eqn (6.56),
liq
gas
CPg

1 dFPg
liq
Form
+ ROg
= kO , Pg av C ,eqm CPg
g = 1,2,.........N lumps
, acid
K

dz
Pg

app
DH ( Pg Og )

liq liq
COg
CH 2
=
liq
CPg

g = 1,2,.........Nlumps

(6.55)

(6.56)

The above set of algebraic equations relates the liquid phase concentration of any
paraffinic lump Pg and the corresponding olefinic lump Og through the apparent
app
.
dehydrogenation equilibrium coefficient K DH

app
K DH
gives the average value of the

dehydrogenation equilibrium coefficient for the dehydrogenation of lump Pg into lump

Og through all possible isomers and is

related to the individual dehydrogenation

equilibrium coefficients by,

5 isomers

eqm
ni ,q yPi
q =1 i =1

app
K DH
( Pg Og ) =
paraffin
eqm
isomers
yOij
liq
i =1 j =1 K DH ,ij
paraffin

(6.57)

6.13. Estimation of the Model Parameters

The model parameters are estimated from the isothermal experimental data on threephase hydrocracking of n-hexadecane in a tubular reactor. The parameters involved in
the model are divided into four categories:

6.13.1. Rate Parameters for the Metal Site Steps

As discussed earlier, to model the hydrogenation/dehydrogenation reactions, five
activation energies for five different modes of dehydrogenation and one frequency factor
is required.

137

6.13.2. Rate Parameters for Acid Site Elementary Steps

The lumping coefficients for the acid site steps are functions of the Evan-Polanyi
parameters pcp and cr and therefore, simultaneous estimation of these two parameters
along with other model parameters requires the recalculation of the lumping coefficients
at each iteration during the parameters estimation. As the value of alpha should lie
between 0 and 1, the recalculation of lumping coefficients at every iteration has been
avoided by calculating them once at several pairs of alpha at an interval of 0.1 and
estimating the remaining model parameters at these different pairs of alpha, maintaining
them constant. The pair of alpha corresponding to the best set of parameters has been
retained as final.
To find out the actual number of independent parameters that can be determined from
experimental data, eqn (6.40) is written for all the four subtypes of PCP steps. i.e.,

E pcp ( s ;s ) = Eo , pcp + q ( s )
2
E pcp ( s ;t ) = Eo , pcp + q ( s ) + 'pcp ( q (t ) q ( s ) )
2

(6.58)

E pcp (t ;s ) = Eo , pcp + q (t ) + 'pcp ( q ( s ) q (t ) )

2
E pcp (t ;t ) = Eo , pcp + q (t )
2
Defining Eo, pcp = Eo , pcp + q ( s ) and q (t ; s ) = q (t ) q ( s ) the above equations can be
written as,

E pcp ( s ;s ) = Eo, pcp

2
E pcp ( s ;t ) = Eo, pcp + 'pcp q (t ; s )
2
E pcp (t ;s ) = Eo, pcp + (1 'pcp )q (t ; s )
2
E pcp (t ;t ) = Eo, pcp + q (t ; s )
2

(6.59)

138

In the same way, Eo,cr = Eo ,cr + q ( s ) is defined for the cracking steps and equations
analogous to (6.59) are written for the cracking steps. It can be seen from the above
transformation of parameters that for constant values of alpha, the four original
parameters used in eqn (6.58), i.e., Eo , pcp , Eo ,cr , q (t ) and q ( s ) are not mutually
independent. Only three re-parameterized entities defined as Eo, pcp , Eo,cr and q (t ; s ) are
mutually independent parameters regarding the acid site activation energies that can be
determined from the experimental data. Apart from these three parameters two
comp
composite single-event frequency factors A
pcp
and A
crcomp are also independent parameters

for the acid site steps.

6.13.3. Parameters for Sorption in Zeolite Pores

From the experiments76 on the adsorption of alkane molecules up to carbon number 16,
it was found that the Henry coefficients of alkanes of different chain length in non-polar
n-octane mobile phase are almost identical at high liquid densities and increases slightly
with carbon number at relatively lower liquid densities. Moreover, Henry coefficients in
the liquid phase have been found to be very weak functions of temperature meaning a
relatively small heat of sorption, in contrast to the high heat of sorption in the gas
phase109. Therefore, in the hydrocracking of pure paraffins, only one temperature
independent Henry coefficient has been used irrespective of the chain length and degree
of branching of the molecules. The Langmuir coefficient for sorption K L, Pg for a
paraffinic lump g is obtained as the ratio of the Henry coefficient and the saturation
concentration of lump g, csat , g . Because of the existence of liquid like densities of
hydrocarbons inside the zeolite pores, the saturation concentration of a lump is obtained
by dividing the pore volume of the zeolite by the saturated molar volume of lump g in
the liquid phase70, i.e.,
csat , g =

v pore
liq
vsat
,g

(6.60)

139

Because of very high correlation in the sorption equilibrium coefficients and the kinetic
rate parameters, the Henry coefficients for the olefins have been assumed to be the same
as those of the paraffins.

6.13.4. Parameters for Chemisorption on Metal Sites

The independent estimation of the chemisorption parameters was found to be difficult
because the form of the rate expression leads to a high correlation between the
chemisorption equilibrium coefficient, Henry coefficient and rate coefficient for
dehydrogenation. The chemisorption equilibrium coefficient for a particular paraffinic
and olefinic lump at the metal sites were initially assumed to be equal and a function of
carbon number represented by a relationship analogous to that given by Denayer109 for
the gas phase sorption of paraffins in zeolite pores. However, from the initial parameter
estimation, the carbon number dependency came out to be very weak. Moreover, the
denominator term Dc for chemisorption was close to unity. Consequently, it was
assumed that the chemisorption equilibrium coefficients are small and independent of
the carbon number. This allowed coupling the entropy and heat of chemisorption for
paraffins and olefins along with the pre-exponential factor and activation energies of
dehydrogenation, respectively.
Summarizing, the model contains a total of 14 temperature and feedstock composition
independent parameters whose values were estimated from the experimental data by
minimizing the sum of squares of the residuals between the experimental and model
predicted responses using the Levenberg-Marquardt optimization algorithm.
nresp nobs

S ( ) =

i =1

y exp yijcal

ij

j =1

(6.61)

In the absence of replicate experimental data no weighting factors are used. 34 responses
representing the molar flow rates of normal paraffins and iso-paraffins, the total
conversion and isomerization yield of n-hexadecane, and the C4/C12 molar ratio are used
per set of experimental data. The distribution of mono-branched and multi-branched

140

isomers for a given carbon number was available only from C6 to C9 and this
information has also been used in the parameter estimation. For a particular Pt/USY
catalyst (Cat-I) experimental data was available at a total pressure of 35 atm., at two
temperatures, 300 oC and 320 oC, and molar H2/HC ratio of 6.0 to 11.0. The parameters
related to the acid sites have been found to satisfy the rules of carbenium ions chemistry.
The ratio of the composite single-event frequency factor for -scission to that of PCP is
in agreement with the values reported earlier17,

72

. The estimated parameters lead to

effective activation energies of PCP and -scission which are close, resulting in the
selectivity vs. conversion behavior with respect to temperature observed by Steijns et
al.97 for the Ac-rds case. The five effective activation energies for dehydrogenation on
the metal sites are of the order of 80 kJ/mol.
The product distribution obtained by using the parameters for Cat-I is shown in Figure
6-3 to Figure 6-5. The agreement between the experimental and calculated conversions,
the distribution of cracked products as well as the percentage of isomers for the entire
product range of C3 to C13 is very good. The percentage of isomers in the cracked
products increases with the carbon number. The skewed product distribution of Figure
6-4 reveals a significant amount of secondary cracking. For this catalyst, a single-event

141

kinetic model with rate determining step only on the acidic sites predicted a much
smaller amount of secondary cracking and a poorer fit of the experimental data.
The single-event kinetic models previously developed for the hydrocracking of paraffins
are based on four rate coefficients per type of elementary step, namely k(s;s), k(s;t),
k(t;s) and k(t;t). It was found in previous work that such a model consistently over-

predicted the yields of C3 and C4 products. One of the reasons for this behavior is the
inability of these models to distinguish among the rate coefficient of cracking steps in
which a C3/C4 product is formed from those in which both the cracked products are
larger than C4. Because of the lesser degree of charge distribution, the heats of formation
of 2-propyl and s-butyl carbenium ions are relatively higher than those of the larger
secondary ions whose stability is slightly increased due to the presence of a larger
number of -carbons with respect to the positive charge. The implementation of the
Evans-Polanyi relationship in the current model accounts for such subtle differences in
the stability of ions and accordingly assigns a higher activation energy to the cracking
steps in which a C3 or C4 ions is formed, resulting in an improved fit of the experimental
data.

142

80

Total conversion

% Conversion

60

Isom conversion

40

20

0
100

150

200

250

300

350

400

450

500

Space T ime (gcat/mol/hr)

Figure 6-3. Total and isomerization conversion of nC16 as a function of space time at T = 300 oC, P = 35
Bar and H2/HC ratio = 11.0. (Points: experimental, lines: predicted).

143

(a)
Moles per 100 moles of products

8
7

exp

cal

5
4
3
2
1
0
0

Moles per 100 moles of products

(b)

4
6
8
Carbon Number

10

12

14

14
12

exp

10

cal

8
6
4
2
0
0

6
8
Carbon Number

10

12

14

Figure 6-4. Carbon number distribution of cracked products in the hydrocracking of nC16. (a) T = 300 oC,
P = 35 Bar, H2/HC ratio = 11.0 and total conversion = 57 % (b) T = 320 oC, P = 35 Bar, H2/HC ratio =
11.0 and total conversion = 78 %.

144

(a)
100

% Iso/(Iso+Normal)

80
60

40
20

exp
cal

0
0

6
8
10
Carbon Number

12

14

16

(b)
100

% Iso/(Iso+Normal)

80
60

40
exp

20

cal
0
0

6
8
10
Carbon Number

12

14

16

Figure 6-5. Percentage of isomers per carbon number in the hydrocracking of nC16 (a) T = 300 oC, P = 35
Bar, H2/HC ratio = 11.0 and total conversion = 57 % (b) T = 320 oC, P = 35 Bar, H2/HC ratio = 11.0 and
total conversion = 78 %.

145

6.14. Approach to (De)Hydrogenation Equilibrium

Depending

upon

the

relative

metal/acid

activity

of

the

catalyst,

the

hydrogenation/dehydrogenation steps do not necessarily attain equilibrium. The initial

cracking selectivity, So can be used as a measure to quantify the approach to
(de)hydrogenation

equilibrium

because

So

would

be

zero

when

the

hydrogenation/dehydrogenation steps are at equilibrium and it would increase as the

deviation of these steps from the equilibrium increases. It was observed, however, that
the deviation in the value of So is not very sensitive with respect to the relative metal/acid
activity of the catalyst. To quantify the approach to equilibrium, a new variable
g

(C
=

liq
Og

liq
CHliq2 CPg

app
DH ( Pg Og )

has been defined for a particular lump g. To compare the approach

to (de)hydrogenation equilibrium for different catalysts, the value of at zero

conversion for the reacting normal paraffin (i.e., o , feed ) has been found to be useful.
When

the

catalyst

has

sufficient

metal

activity

to

bring

the

hydrogenation/dehydrogenation steps to equilibrium, the value of for all the lumps

equals unity. With a catalyst having a relatively weaker metal/acid activity, the value of
for different lumps deviates from unity. For example, in the hydrocracking of nC16 the

value of for the feed molecule decreases from unity as the relative metal/acid activity
decreases, whereas for all other lumps, becomes greater than unity. A weak metal
function is not able to dehydrogenate the paraffinic feed nC16 fast enough (i.e., compared
to the rate at which the corresponding olefinic nC16 lump is consumed on the acid sites)
to bring the concentration of olefinic nC16 lump at its equilibrium value, resulting in a
value of smaller than 1. At the same time, a weak metal function is not able to
hydrogenate the branched olefinic isomers of C16 and the olefinic products formed on the
acid sites fast enough (compared to the rate they are produced on the acid sites) to
reduce their concentration to the corresponding (de)hydrogenation equilibrium
concentration. Consequently, these olefins attain a steady state at concentrations higher
than their equilibrium concentrations, resulting in larger than unity. The deviation

146

from unity increases as the relative metal/acid activity decreases. Here it should be
noticed that in the hydrocracking of paraffin mixtures, the value of for the heaviest
paraffin should be used for quantifying the approach to equilibrium, as the other reacting
paraffins would also be formed by the cracking of the heaviest paraffin, and therefore the
value of for them may actually be larger than unity.
The value of o , feed obtained from the reactor simulations depends on the values of the
model parameters, which in turn depend on the relative metal/acid activity of a particular
catalyst, the reactor operating temperature, total pressure and H2/HC ratio. In three-phase
hydrocracking, the effect of total pressure and H2/HC ratio on o , feed is much smaller
than that of temperature.
The value of o ,nC16 obtained from the simulations for Cat-I is 0.774 at 300 oC and
0.571 at 320 oC indicating that higher temperatures lead to larger deviation from the
equilibrated hydrogenation/dehydrogenation steps. Similar conclusions have been
arrived at in previous experimental studies35, 99. These results point to higher effective
activation energies of acid site steps compared to the effective activation energies of
the metal site steps.

6.15. Application of the Model to Heavy Paraffins and Different Catalysts

The model has been used to simulate the three-phase hydrocracking of heavy paraffins
ranging from C16 to C32 over a number of operating conditions. To study the effect of the
relative metal/acid activity of the catalyst on the product distribution, two new set of
parameters (Set-II and Set-III) have been derived from Set-I, which was obtained from
the experimental data on Cat-I. Set-II has been obtained from Set-I by decreasing the
composite frequency factor for dehydrogenation by 50% and increasing the composite
frequency factors for PCP and -scission by 35%, while keeping all other parameters
unchanged. Set-II, therefore, corresponds to a hypothetical catalyst Cat-II, having a
relatively lower metal/acid activity than Cat-I. The composite frequency factors for PCP
and -scission in Set-III (corresponding to a hypothetical catalyst Cat-III) were chosen at

147

30% of the corresponding values in Set-I. The Ac-rds case can be considered as the
asymptotic behavior of the Me-Ac-rds with respect to the relative metal/acid activity of
the catalyst, and therefore a large number has been taken for the composite frequency
factor for dehydrogenation for Cat-III so that it corresponds to Ac-rds case. It should be
mentioned that the possibility of direct isomerization and hydrogenolysis of paraffins on
the metal sites of the catalyst has not been considered in the current study.

6.15.1. Effect of Relative Metal/Acid Activity of the Catalyst

The effect of the relative metal/acid activities on product distribution is illustrated in
Figure 6-6 to Figure 6-11 for the hydrocracking of nC32 at 375 oC, 150 bar and a H2/HC
molar ratio of 11.0. Figure 6-6 and Figure 6-7 show the evolution of C32 isomers with
different degrees of branching vs. the total conversion. Figure 6-6 corresponds to Cat-III
resulting in o , nC 32 = 0.99. The well-known Ac-rds behavior can be seen from the initial
slopes of the curves indicating mono-branched paraffins to be the primary products and
di-branched and tri-branched products formed sequentially from the mono-branched
paraffins. Figure 6-7 corresponds to Cat-II resulting in o , nC 32 = 0.40, representing a
significant deviation from the equilibrated hydrogenation/dehydrogenation steps.

In

contrast to Figure 6-6, di-branched and tri-branched isomers are not really secondary
products but are formed directly from nC32. The appearance of the multi-branched
isomers of the feed as the primary products can be used as a possible criterion to identify
the lack of metal activity of the catalyst in an experimental study. Alvarez et. al110 using
catalysts having different metal/acid activities observed this phenomenon in the
hydrocracking of n-decane. When the intrinsic hydrogenation/dehydrogenation rates are
comparable or smaller than the intrinsic rates of acid sites PCP steps, di-branched and
tri-branched olefinic isomers of the feed are produced from the mono-branched olefinic
isomers at a rate faster than that at which mono-branched olefinic isomers could be
hydrogenated into mono-branched paraffins. Consequently, all the mono-, di- and tribranched olefinic isomers are hydrogenated almost simultaneously, exhibiting a nonzero initial selectivity for multi-branched isomers. An analogous behavior for cracking

148

can be observed in Figure 6-8 showing the isomerization selectivity as a function of total
conversion for the above mentioned three catalysts. The initial isomerization selectivity
for Cat-III (i.e., Ac-rds) is unity, while that for Cat-I and Cat-II (i.e., Me-Ac-rds), 0.95
and 0.82 respectively, indicating that normal paraffins are directly cracked into smaller
products, even before their isomerization into the branched paraffins could take place.
The explanation for this behavior is the same as described for the direct appearance of
multi-branched feed isomers and has been observed experimentally.110, 111
Figure 6-9 and Figure 6-10 show the moles of cracked products formed per 100 moles of
C32 cracked. Figure 6-9 corresponds to the hydrocracking on Cat-III, resulting in
o ,nC 32 = 0.99 and Figure 6-10 corresponds to Cat-I for which o , nC 32 is 0.67. It can be

observed that secondary cracking takes place even in Ac-rds hydrocracking (Cat-III),
however, in Me-Ac-rds (Cat-I) the amount of secondary cracking is significantly higher
for comparable values of the total conversion. This is because at the same operating
conditions, the concentrations of cracked olefins for Cat-I are higher than those on CatIII due to their lower rates of hydrogenation. This leads to higher concentrations of the
corresponding carbenium ions as the protonation/deprotonation steps maintain pseudoequilibrium with the olefins. The existence of higher concentrations of the carbenium
ions of cracked products increases their rate of cracking compared to Ac-rds
hydrocracking, resulting in increased secondary cracking.
At the same time, the lower metal/acid activity of Cat-I leads to higher concentrations of
the branched C32 carbenium ions as compared to Cat-III. This increases the rate of
cracking of the branched isomers, resulting in a lower isomerization conversion. This is
shown in Figure 6-11 in which the isomerization conversion is plotted against the total
conversion for the three catalysts. For Cat-III, the maximum isomerization conversion is
46.7 % as compared to 24.3 % for Cat-II for which the value of o , nC 32 is 0.4. This
behavior has been observed in several experimental studies.99, 112

149

100.0
o = 0.99

NP
MBP

80.0

DBP

Mole %

T BP
60.0

40.0

20.0

0.0
0.0

20.0

40.0

60.0

80.0

100.0

T otal Conversion %

Figure 6-6. Evolution of the isomers of different degrees of branching for the hydrocracking of nC32
corresponding to equilibrated (de)hydrogenation condition (Ac-rds case).

100.0
o = 0.4

NP
MBP

80.0

DBP

Mole %

T BP
60.0

40.0

20.0

0.0
0.0

20.0

40.0

60.0

80.0

100.0

T otal Conversion %

Figure 6-7. Evolution of the isomers of different degrees of branching for the hydrocracking of nC32
corresponding to non-equilibrated (de)hydrogenation condition (Me-Ac-rds case).

150

1.0
Cat-III

Isom. selectivity

0.8

Cat-I

0.6
Cat-II
o = 0.99
0.4
o = 0.67
0.2
o = 0.40
0.0
0.0

20.0

40.0

60.0

80.0

100.0

T otal conversion, %

Figure 6-8. Effect of relative metal/acid activity on initial isomerization selectivity in nC32 hydrocracking.

25.0
o = 0.99

XT = 54 %
XT = 76 %

Moles/100 moles of feed cracked

20.0

XT = 87 %
XT = 90 %

15.0

10.0

5.0

0.0
2

10

14

18

22

26

30

Carbon Number

Figure 6-9. Simulated distribution of cracked products for Ac-rds case at various total conversions for the
hydrocracking of nC32.

151

30.0
o = 0.67

XT = 54 %
XT = 76 %

Moles/100 moles of feed cracked

25.0

XT = 88 %
XT = 91 %

20.0

15.0
10.0

5.0
0.0
2

10

14

18

22

26

30

Carbon Number

Figure 6-10. Simulated distribution of cracked products for Me-Ac-rds case at various total conversions
for the hydrocracking of nC32.

Isom. conversion, %

60.0

50.0

Cat-III

40.0

Cat-I

o = 0.99

30.0
Cat-II

o = 0.67

20.0
o = 0.40
10.0

0.0
0.0

20.0

40.0

60.0

80.0

100.0

T otal conversion, %

Figure 6-11. Influence of the relative metal/acid activity on the maximum isomerization conversion for
the hydrocracking of nC32.

152

6.15.2. Effect of Temperature

Figure 6-12 shows the isomerization conversion plotted against total conversion for the
hydrocracking of nC32 on Cat-II at a total pressure of 150 bar, H2/HC molar ratio of 11.0
and three different temperatures. It can be seen that the maximum isomerization
conversion (or yield) decreases when the temperature is increased. The decrease in the
maximum isomerization yields as well as in the values of o , nC 32 show that the
deviation from the equilibrated hydrogenation/dehydrogenation steps increases with
temperature. This phenomenon has been observed experimentally99 and can be attributed
to higher effective activation energies of acid site steps than the effective activation
energies of metal site steps. The effective activation energies of acid site steps include
the heat of sorption in the zeolite pores, heat of chemisorption on the metal sites, heat of
protonation of olefins and the intrinsic activation energies of the acid site steps.
Similarly, the effective activation energies for metal site steps include the heat of
sorption in the zeolite pores, heat of chemisorption on the metal sites, and the intrinsic
activation energies of dehydrogenation steps.
When

the

catalyst

has

hydrogenation/dehydrogenation

sufficient
equilibrium

metal
(i.e.,

activity

o , nC 32 =1.0)

to
at

the

achieve
highest

temperature, the value of o , nC 32 remains 1.0 at all lower temperatures. In that case the
decrease in the maximum isomerization conversion with an increase in temperature
would be significantly lower or negligible, attributed entirely to the differences in the
effective activation energies of PCP and cracking steps, in contrast to a catalyst with
weak metal/acid activity. No experimental data could be found documenting this
behavior for the three-phase hydrocracking of long chain paraffin like C32 on a so-called
well-balanced catalyst. Experiments on gas-phase hydrocracking of C8 to C12 paraffins
on well-balanced catalysts show a unique curve i.e., independent of temperature.97

153

60.0

50.0
Isom. conversion, %

300 o C
40.0
350 o C
30.0
375 o C

o = 0.82
o = 0.54

20.0
o = 0.40
10.0

0.0
0.0

20.0

40.0

60.0

80.0

100.0

T otal conversion, %

Figure 6-12. Influence of temperature on the maximum isomerization conversion for the hydrocracking of
nC32 over Cat-II.

6.15.3. Effect of Total Pressure and H2/HC Molar Ratio

The effect of total pressure at a constant temperature of 375 oC and H2/HC molar ratio of
11.0 has been simulated for three-phase hydrocracking of nC32. Figure 6-13 shows that
the increase in the total pressure does not change the value of o , nC 32 for Ac-rds case
(i.e., Cat-III). For Cat-I and II, the values of o , nC 32 increase with total pressure,
indicating that over a catalyst with insufficient metal activity, an increase in the total
pressure brings the hydrogenation/dehydrogenations steps closer to equilibrium. The
shift is more pronounced over a catalyst with a lower metal/acid activity.
Figure 6-14 shows the effect of total pressure on the total conversion, isomerization
conversion, and the secondary cracking for Cat-I. Secondary cracking is represented by
the moles of cracked products formed per 100 moles of feed cracked. The behavior of
Figure 6-14 can be explained as follows. In three-phase hydrocracking the reaction rates

154

are governed by the liquid phase concentrations (or fugacities) of hydrocarbons and
hydrogen. With the increase in total pressure, liquid phase fugacity of hydrogen
increases significantly with only negligible increase in the fugacity of hydrocarbons.
Increased hydrogen fugacity decreases the intrinsic rate of dehydrogenation of the feed
paraffin and therefore, the total conversion. A higher liquid phase fugacity of hydrogen
increases the rate of hydrogenation of the olefinic isomers of the feed and of the olefins
produced by cracking, so that an increased total pressure favors the isomerization
conversion but decreases the secondary cracking. These trends correspond to an increase
of o , nC 32 with total pressure.
In three-phase hydrocracking at constant temperature and total pressure, change in the
H2/HC ratio has negligible effect on the liquid phase fugacities of hydrogen and
hydrocarbons and alters primarily the rate of mass-transfer of these species between the
gas and liquid phases. Therefore, it has small effect on the total conversion of paraffinic
feedstocks provided there are no mass transfer limitations. A higher H2/HC ratio, on the
other hand, provides more driving force to the lighter cracked products to escape from
the liquid phase into the gas phase leading to a decrease in secondary cracking.

155

1.2
Cat-III
1

0.8

o, nC32

Cat-I
0.6
Cat-II
0.4

0.2

0
100

110

120

130

140

150

160

Pressure, Bar

Figure 6-13. Effect of total pressure on the value of

o , nC 32 at T=375 oC and R=11 for the

hydrocracking of nC32 over different catalysts.

400

100
90
80

350

Conversion, %

70
60
300
50
40
Isom. Conv.

30

250

20
10
100

110

120

130

140

150

Moles of cracked products formed per

100 moles of feed cracked

T otal Conv.

200
160

Pressure, Bar

Figure 6-14. Effect of total pressure on the total conversion, isomerization conversion and secondary
cracking at T=375 oC and R=11 for the hydrocracking of nC32 over Cat-I.

156

6.15.4. Effect of Feed Chain Length

Figure 6-15 shows the effect of the chain length on the total conversion and secondary
cracking. With chain length, reactivity of paraffins increases because of the increase in
the available reaction pathways resulting in higher conversions. The chain length of
primary cracking products formed from a long chain hydrocarbon is likely to be longer
than those formed from short chain hydrocarbons, making them more susceptible to
further cracking and thus leading to increased secondary cracking.

250

50
240

Total conversion, %

45
40

230
35
30

220

25
20

Moles of cracked products formed per

100 moles of feed cracked

55

210
12

16

20

24

28

32

36

Feed chain length

Figure 6-15. Effect of chain length on the total conversion and secondary cracking expressed as the moles
of cracked products formed per 100 moles of feed cracked at 375 oC, 150 bar and H2/HC molar ratio of
11.0 for a given space time over Cat-I.

157

CHAPTER VII
7.CONCLUSIONS
This research is focused on the development of mechanistic kinetic models for the
hydrocracking of complex feedstocks of industrial relevance. Two separate models are
developed, one for the hydrocracking of vacuum gas oils and the other for the
hydrocracking of long chain paraffins. Because of the complex structures of the
molecules present in a typical vacuum gas oil and the gigantic reaction network, some
simplifying assumptions are made in the hydrocracking model of vacuum gas oil. In the
hydrocracking model of paraffin feeds, on the other hand, the modeling approach
adopted for VGO has been extended to an even more detailed and fundamental level
which allowed to investigate the effect of the relative strength of the metal/acid activity
of the hydrocracking catalyst on the hydrocracking product distribution.
Both models are developed using an exhaustive network of elementary steps, generated
by a computer algorithm. The elementary steps occurring on the acid sites of the catalyst
are based on the underlying carbenium ion chemistry of the hydrocracking process. The
reaction network for the saturation of aromatic species is based on the sequential
hydrogenation of the aromatic rings.
In the VGO hydrocracking model, it is assumed that the dehydrogenation/hydrogenation
reactions preceding/following the acid site transformations attain quasi-equilibrium, so
that the elementary steps on the acid sites are considered to be the rate determining steps.
The frequency factors for these steps are modeled using the single event concept and the
activation energies are based on the nature of the reactant and the product carbenium
ions. This leads to a tractable number of model parameters in spite of a very large
number of steps actually considered in the model. The number of activation energies
required in the model has been further reduced by identifying the similarity in the
structural transformation in the acyclic, exocyclic and endocyclic -scission steps.

158

An investigation of the available experimental data on the hydrocracking of VGO

revealed that the amount of mono-ring species found in the hydrocracking products was
significantly higher than what could be formed by the opening of the multi-ring species.
This suggested the importance of the cyclization of paraffins in the formation of monoring naphthenes. A reaction mechanism proceeding through diolefinic species has been
proposed in this work to model the rate of the cyclization steps. A significant increase in
the rate of cyclization with temperature has been reported in the literature. It is supported
by this mechanism, as shown from the reactor simulations.
To model the equilibrium coefficients of the protonation steps, a new hypothesis has
been proposed in this work which eliminates the involvement of the reference olefins,
practiced previously. This hypothesis has been corroborated by calculating the true
values of the lumping coefficients using the gas phase heats of formation of carbenium
ions and comparing them with the lumping coefficients calculated with the existing and
proposed methods.
As the reactions on zeolite catalysts occur through the reactants sorbed in the zeolite
pores, the kinetic model have been developed to account for the sorption of
hydrocarbons belonging to different classes from the liquid phase to the zeolite pores. As
reported in the literature, the reactivity of bulky molecules decreases with molecular size
on the zeolite catalysts due to their reduced probability to enter the pores. The current
modeling approach allows to account for the reduced reactivity of the bulky molecules
by using the smaller values of their sorption equilibrium coefficients.
The model for the hydrocracking of VGO explicitly considers the strong competitive
chemisorption of aromatic/polyaromatic species on the acid sites of the catalyst to
account for the influence of the aromatic content of the feedstock on the rates of acid site
transformations. The values of the protonation equilibrium coefficients required to
calculate the acid site coverage have been estimated from the gas-phase proton affinities
of the aromatics and their heats of stabilization from gas phase to acid sites. The entropic
part of the protonation equilibrium coefficients has been estimated using statistical
thermodynamics.

159

The increase in the hydrogenation reactivity of aromatics with the number of aromatic
and saturated rings is well known from the experimental literature. This increase in the
reactivity is caused both by the higher chemisorption coefficients of polyaromatics on
the metal sites as well as the higher values of their intrinsic hydrogenation rate
coefficients. The current model accounts for both these important aspects to model the
rate of aromatic hydrogenation reactions at a fundamental level.
The total number of molecular species considered in this model is of the order of 1
million. It was not possible to write a continuity equation for each individual species
because of the mathematical problems that would arise during the integration of such a
large number of ordinary differential equations as well as the limitations of the current
analytical techniques for the identification of individual molecules in the heavy
petroleum mixtures. Therefore, a strict lumping strategy has been devised which is
based on the attainment of thermodynamic equilibrium among the isomers placed in
each lump. This approach leads to 1266 pure components/lumps to be used in this
model. Although a lumping of components is employed in the model, the global rates of
conversion among the lumps are calculated by accounting for the contribution of all the
reactions of every component present in a lump. The concentration profiles of all these
lumps and hydrogen are obtained in the gas and liquid phases along the reactor by
integrating 2534 ordinary differential equations. An ordinary differential equation for the
temperature profile of the gas/liquid phases and an algebraic equation for the calculation
of the temperature of the catalyst at each axial position in the reactor are solved
simultaneously with the 2534 mass balance equations. The model has been used to
estimate the parameters from the experimental data on VGO hydrocracking.
A variety of simulations have been performed in this work using the estimated
parameters to analyze the effect of operating conditions on the product distribution. The
model is capable of providing the product distribution at any level of detail desired by
the user. For example, the evolution of various commercial fractions, the evolution of
various hydrocarbon classes, the evolution of paraffins, naphthenes and aromatics in
each commercial fraction, the evolution of each hydrocarbon class in a commercial

160

fraction, the evolution of the carbon number distribution of each class as well as each
commercial fraction and so on. The effect of temperature and pressure on the product
distribution predicted by the model is supported by the information available in the open
literature. The effect of the distribution of the isomers with different degree of branching
in a class on the reactivity of vacuum gas oil has also been studied, showing the increase
in the VGO conversion with an increase in the content of isomers with higher number of
methyl branches. The simulations carried out in this study reveal the deviation of the
isomer distribution from the corresponding thermodynamic distribution, confirming the
published experimental data.
The potential of the model for commercial applications is clear considering the wealth of
information it produces compared to the available models based on lumps mainly
defined by their physical properties. For example, the prediction of the yield of
aromatics in the hydrocracking products is very important to the refining industry
considering the growing environmental regulations imposed by federal agencies. Apart
from that, the properties of the products obtained from the hydrocracking process are
significantly affected by their aromatic content, e.g., aromatics improve the quality of
the gasoline fraction by enhancing its octane number, while they deteriorate the quality
of kerosene and diesel in terms of their smoke point and cetane number. Therefore, the
accurate prediction of the amount and composition of aromatics in various commercial
fractions and the effect of operating conditions on their yields as predicted by the present
model can be extremely valuable to the refiners.
The hydrogen consumption for different feedstocks and the knowledge of the
temperature increase during hydrocracking reactions are also very important parameters
for the hydrocracker operators. The conventional lumped models obtain the amount of
hydrogen consumption using the average C/H ratios of the feed and products, whereas
the total heat generated is calculated by taking a constant value for the heat evolved per
mole of hydrogen consumed. The current model considers a detailed reaction network
for the hydrogenation of mono- and multiring aromatics which allows the prediction of
precise values for of hydrogen consumption, depending on the aromatic content and

161

composition of the feed, as well as on the conversion level. The detailed energy balance
based on the heats of formation of all the reacting species along with the heat effects
associated with the vaporization of lighter hydrocarbons formed during the cracking
reactions permits the model to accurately predict the temperature profile along the
reactor. This kind of information can be efficiently used for the optimization of
commercial plants with respect to the operating conditions as well as the efficient
selection of the feedstock for a desired product slate.
The other contribution of this research is the development of a generalized mechanistic
kinetic model for the hydroisomerization and hydrocracking of long chain paraffins.
Unlike the assumption made in the VGO model, this model considers the
dehydrogenation/hydrogenation

reactions

preceding/following

the

acid

site

transformations also to be rate determining like some of the elementary steps occurring
on the acid sites of the catalyst. This enables the model to simulate hydrocracking on
catalysts with low metal activity. The activation energies of acid site steps in this model
are obtained by the use of the Evans-Polanyi relationship. Quantum chemical
calculations have been performed to obtain the energies of the gas phase carbenium ions.
These are used to develop a group contribution method for the faster estimation of heats
of formation of all the carbenium ions. This model contains 14 independent parameters
which are invariant with respect to the carbon number of the feedstock and the reactor
operating conditions. These parameters have been estimated from experimental data on
the hydrocracking of n-hexadecane. A new parameter o , feed is introduced in this model
which quantifies the approach of the hydrogenation/dehydrogenation steps to the
corresponding equilibrium. The effects of the metal activity, of temperature, pressure
and H2/HC ratio on the product distribution and on the shift in the rate determining step
have been investigated. The model has potential application for selecting the optimum
operating conditions in the hydrocracking of mixtures of paraffins as encountered in
Fisher-Tropsch waxes, now produced starting from natural gas, and for a more efficient
selection of the catalyst.

162

7.1. Recommendations for the Future Work

Similar to the approach taken for the hydrocracking of pure paraffins, the VGO model
can also be built considering the rate determining steps to be both on metal and acid sites
of the catalysts. A firm ground work has been laid down in this research to model the
rate coefficients of the metal site dehydrogenation/hydrogenation reactions of paraffinic
species which can be extended to the ring species present in VGO.
Another important development in the VGO model is the implementation of the EvansPolanyi relationship for the modeling of activation energies. This requires the energies of
the carbenium ions with complex naphthenic and aromatic structures. Calculation of the
energies of these species, although computationally expensive, can be carried out
considering the ever increasing processor speeds. Uncertainties in the energies of such
species using currently available ab initio theories, however, need to be addressed in this
effort. Similar to the acid site steps, the modeling of the activation energies of the metal
site rate coefficients using the Evans-Polanyi relationship in the two-site rate
determining VGO model would also be a significant theoretical development.
The current model does not account for the deactivation of the catalyst due to coke
deposition. Although the rate of coking in hydrocracking is very low because of the high
hydrogen pressures, the prediction of the decrease of the catalyst activity with run time
and the knowledge of the rate of temperature increase required to maintain the
conversion levels are of great importance to the refining industry. Modeling of the rate
of coke formation using a fundamental approach as advanced in the current model would
be highly recommended. At the same time, addition of the hydrodesulfurization and
hydrodenitrification reactions to the current reaction network would be of great practical
importance. The deleterious effect of the organic nitrogen on the acidic sites of the
catalyst also needs to be considered.
The need for reliable kinetic models with high predictive capabilities for such kind of
complex processes is ubiquitous in the oil industry. Development of such models is
possible only through fundamental mechanistic approaches which require a detailed

163

composition of the feedstock. As discussed in this dissertation, to obtain this type of

detailed feedstock composition is very difficult, practically impossible, in a typical
refinery. This opens up the avenues for the development of theoretical approaches that
can predict the detailed feedstock composition from a partial analysis, which is generally
available in the oil industry. Integration of such techniques with the detailed mechanistic
models as developed in this work would consummate the current research.

164

NOMENCLATURE
av = Gas-liquid interfacial area, [mi2/mr3]
aLS = Liquid-solid interfacial area, [mi2/mr3]

A
= Single event frequency factor for type of elementary steps, [1/hr]

A
comp = Composite single event frequency factor for type of steps, [kmol/kgcat/hr]
cm = Total concentration of the metal sites, [kmol/kgcat]
csat , Pi = Sat. concentration of paraffin Pi in the zeolite pore, [kmol/kgcat]

CPgas
, H 2 = Gas phase heat capacity of hydrogen, [kJ/mol/K]
CPgas
, Sg = Gas phase heat capacity of lump S g , [kJ/mol/K]
CPliq, Sg = Liquid phase heat capacity of lump S g , [kJ/mol/K]
gas
Ctotal
= Total concentration of the gas phase, [kmol/m3]
liq
Ctotal
= Total concentration of the liquid phase, [kmol/m3]

Ct = Total concentration of the acid sites, [kmol/kgcat]

CHliq2 = Concentration of hydrogen in the liquid phase, [kmol/m3]

CPiliq = Concentration of paraffin Pi in the liquid phase, [kmol/m3]
CSiliq = Concentration of species Si in the liquid phase, [kmol/m3]
liq
CSg
= Concentration of lump S g in the liquid phase, [kmol/m3]

CSggas = Concentration of lump S g in the gas phase, [kmol/m3]

[CPi ] = Concentration of sorbed paraffin, [kmol/kgcat]

165

COij = Concentration of sorbed olefins, [kmol/kgcat]

C + = Concentration of surface carbenium ions Rik+ , [kmol/kgcat]

Rik
Eo ,k = Intrinsic activation energy for elementary steps of type k, [kJ/mol]

E ( m ;n ) = Activation energy for the steps of (m; n) subtype, [kJ/mol]

Ecomp
( m ; n ) = Composite activation energy for the steps of ( m; n) subtype, [kJ/mol]
fi = Partial fugacity of component i at reaction temperature and pressure, [bar]
fi o = Fugacity of pure component i at standard state conditions, [bar]

FSggas = Gas phase flow rate of lump S g , [kmol/hr]

FSgliq = Liquid phase flow rate of lump S g , [kmol/hr]
FH 2 , quench = Molar flow rate of the hydrogen stream for quench, [kmol/hr]
h = Plancks constant, [kJ.hr]
hLS = Solid-liquid heat transfer coefficient, [kJ/m2/K/hr]
H exo = Heat of the exocyclic -scission elementary steps, [kJ/mol]
H acyc = Heat of the acyclic -scission elementary steps, [kJ/mol]

H fg (i ) = Heat of formation of specie i in the gas phase, [kJ/mol]

H sf (i ) = Heat of formation of species i at the surface, [kJ/mol]
H liq
f , g = Heat of formation of lump S g in the liquid phase, [kJ/mol]
H sorbed
(O) = Heat of formation of olefin in the sorbed phase, [kJ/mol]
f
H vap (i ) = Latent heat of vaporization of species i, [kJ/mol]

166

H stab ( R + ) = Heat of stabilization of carbenium ion, R+ from liquid phase to the surface,

[kJ/mol]
H sorp (i ) = Heat of sorption of specie i from liquid phase to the sorbed phase, [kJ/mol]
H pr ( SOij  m ) = Enthalpy of protonation of olefinic species SOij

H + = Concentration of vacant acid sites, [kmol/kgcat]

H Si = Henry coefficient for the sorption of paraffin Si, [m3/kgcat]
k B = Bolztmans constant, [kJ/molecule/oC]
k
( m;n ) = Single event rate coefficient for the elementary step of type (m; n) , [1/sec]
kO , Sg = Overall mass transfer coefficient for lump S g , [mi/hr]
kG = Gas phase mass transfer coefficient, [mi/hr]
k L = Liquid phase mass transfer coefficient, [mi/hr]
kdhij = Dehydrogenation rate coefficient, [kgcat/kmol/hr]
comp
khyd
= Composite rate coefficient for the hydrogenation of monoaromatics,
,3 H 2

[(kmol/kgcat/hr)(m3/kmol)3]

K HVLE
= Vapor liquid equilibrium coefficient for hydrogen, [dimensionless]
2
C ,VLE
K Sg
= Concentration based vapor liquid equilibrium coefficient for lump S g ,

[dimensionless]
K C , H 2 = Chemisorption equilibrium coefficient for hydrogen at the metal site, [m3/kmol]
K C , Pi = Chemisorption equilibrium coefficient of paraffin Pi at the metal site,

[kgcat/kmol]

167

ads
K DH
,ij = Dehydrogenation equilibrium coefficient at the metal site, [dimensionless]

liq
K DH
,ij = Equilibrium coefficient for the overall dehydrogenation reaction in the liquid

phase, [kmol/m3]
liq
K DH
( Si  SOij ) = Equilibrium coefficient for the given overall dehydrogenation reaction in

the liquid phase, [kmol/m3]

K L , Pi = Sorption equilibrium coefficient of paraffin Pi in the zeolite pore, [m3/kmol]

K pr = protonation/deprotonation equilibrium coefficient, [dimensionless]

ne = Number of single-events, [dimensionless]
N lumps = Total number of lumps/pure components, [dimensionless]
N sat _ lumps = Number of saturated lumps/pure components, [dimensionless]
N aro _ lumps = Number of aromatic lumps/pure components, [dimensionless]
N Sg = Molar flux of lump S g from gas phase to liquid phase, [kmol/hr/mi2]
q (m) = Heat of stabilization of m type of carbenium ion, from sorbed phase to the

surface, [kJ/mol]
q (m) = Relative heat of stabilization of a proton with respect to an m type of paraffinic

carbenium ion, [kJ/mol]

q ( z , m) = Relative heat of stabilization of a proton with respect to an m type of

carbenium ion of class Z, [kJ/mol]

R = Universal gas constant, [kJ/mol/K]

R( g( m;h;)n ) = Global rate of conversion of lump g into lump h through (m; n) type of
elementary steps, [kmol/kgcat/hr]

168

Form
RPg
, metal = Rate of formation of paraffin lump Pg on metal sites, [kmol/kgcat/hr]

Form
ROg
, metal = Rate of formation of Olefin lump Og on metal sites, [kmol/kgcat/hr]

Form
ROg
, acid = Rate of formation of Olefin lump Og on acid sites, [kmol/kgcat/hr]

3
RSgForm
, net = Net rate of formation of lump S g on acid sites, [kmol/mr /hr]

[ Si ] = Sorbed concentration of species

Si in the zeolite pores, [kmol/kgcat]

+
+
S Rik
= Concentration of surface carbenium ions S Rik
, [kmol/kgcat]

S pr = Intrinsic entropy of protonation, [kJ/mol/K]

St = Entropy corresponding to three translational degrees of freedom, [kJ/mol/K]
T liq = Liquid phase temperature, [K]
T solid = Solid phase temperature, [K]
TH 2 = Temperature of the hydrogen stream used for quench, [K]
Tin = Bed inlet temperature, [K]
Tout = Bed outlet temperature, [K]
v pore = Micro-pore volume in the zeolite, [m3/kgcat]
3
liq
vsat
, g = Saturated liquid phase volume of lump g, [m /kmol]

WCfeed
20 + = Weight of C20 + hydrocarbons in the feed, [kg]
WCprod
20 + = Weight of C20 + hydrocarbons in the products, [kg]
WCfeed
20 + = Weight of C20 + hydrocarbons in the feed, [kg]
xi = Mole fraction of component i in the liquid phase, [dimensionless]

169

X = weight percent conversion of VGO

yPieqm = Equilibrium mole fraction of paraffin i in the corresponding lump,

[dimensionless]
eqm
yOij
= Equilibrium mole fraction of olefin ij in the corresponding lump, [dimensionless]

z = Axial length along the reactor, [mr]

Greek symbols

k = Transfer function for elementary steps of type k

i = Activity coefficient of component i in the liquid phase, [dimensionless]
= Reactor cross-sectional area, [mr2]

Sg = Latent heat of vaporization of lump S g , [kJ/mol]

= Parameter for quantifying approach to equilibrium, [dimensionless]

EXT = External rotational symmetry number, [dimensionless]

INT = Internal rotational symmetry number, [dimensionless]

gli = Global symmetry number of species i, [dimensionless]

170

REFERENCES
(1) Algelt, K. H.; Boduszynski, M. M., Composition and Analysis of Heavy Petroleum
Fractions. Marcel Dekker, Inc.: New York, 1994.
(2) Snyder, L. R., Nitrogen and Oxygen Compound Types in Petroleum - Total Analysis
of a 400-700 Degrees F Distillate from a California Crude Oil. Analytical Chemistry
1969, 41, (2), 314-320.
(3) Scherzer, J.; Gruia, A. J., Hydrocracking Science and Technology. Marcel Dekker,
Inc.: New York, 1996.
(4) Weekman, V. W.; Nace, D. M., Kinetics of Catalytic Cracking Selectivity in Fixed,
Moving, and Fluid Bed Reactors. AIChE Journal 1970, 16, (3), 397-404.
(5) Jacob, S. M.; Gross, B.; Voltz, S. E.; Weekman, V. W., Lumping and Reaction
Scheme for Catalytic Cracking. AIChE Journal 1976, 22, (4), 701-713.
(6) Stangela.B. E., Kinetic-Model for Prediction of Hydrocracker Yields. Industrial &
Engineering Chemistry Process Design and Development 1974, 13, (1), 71-76.
(7) Laxminarasimhan, C. S.; Verma, R. P.; Ramachandran, P. A., Continuous lumping
model for simulation of hydrocracking. AIChE Journal 1996, 42, (9), 2645-2653.
(8) Liguras, D. K.; Allen, D. T., Structural Models for Catalytic Cracking. 1. ModelCompound Reactions. Industrial & Engineering Chemistry Research 1989, 28, (6), 665673.
(9) Liguras, D. K.; Allen, D. T., Structural Models for Catalytic Cracking. 2. Reactions
of Simulated Oil Mixtures. Industrial & Engineering Chemistry Research 1989, 28, (6),
674-683.
(10) Quann, R. J.; Jaffe, S. B., Structure-Oriented Lumping - Describing the Chemistry
of Complex Hydrocarbon Mixtures. Industrial & Engineering Chemistry Research 1992,
31, (11), 2483-2497.
(11) Baltanas, M. A.; Vanraemdonck, K. K.; Froment, G. F.; Mohedas, S. R.,
Fundamental Kinetic Modeling of Hydroisomerization and Hydrocracking on NobleMetal-Loaded Faujasites. 1. Rate Parameters for Hydroisomerization. Industrial &
Engineering Chemistry Research 1989, 28, (7), 899-910.
(12) Vynckier, E.; Froment, G. F., Modeling of the Kinetics of Complex Porcesses
Based upon Elementary Steps. Elsevier Science Publishers BV: Amsterdam, The
Netherlands, 1991; Vol. 131.

171

(13) Clymans, P. J.; Froment, G. F., Computer-Generation of Reaction Paths and RateEquations in the Thermal-Cracking of Normal and Branched Paraffins. Computers &
Chemical Engineering 1984, 8, (2), 137-142.
(14) Hillewaert, L. P.; Dierickx, J. L.; Froment, G. F., Computer-Generation of Reaction
Schemes and Rate-Equations for Thermal-Cracking. AIChE Journal 1988, 34, (1), 1724.
(15) Feng, W.; Vynckier, E.; Froment, G. F., Single-Event Kinetics of Catalytic
Cracking. Industrial & Engineering Chemistry Research 1993, 32, (12), 2997-3005.
(16) Svoboda, G. D.; Vynckier, E.; Debrabandere, B.; Froment, G. F., Single-Event
Rate Parameters for Paraffin Hydrocracking Oil a Pt/Us-Y Zeolite. Industrial &
Engineering Chemistry Research 1995, 34, (11), 3793-3800.
(17) Martens, G. G.; Marin, G. B.; Martens, J. A.; Jacobs, P. A.; Baroni, G. V., A
Fundamental Kinetic Model for Hydrocracking of C-8 to C-12 Alkanes on Pt/US-Y
Zeolites. Journal of Catalysis 2000, 195, (2), 253-267.
(18) Park, T. P.; Froment, G. F., Kinetic Modeling of the Methanol to Olefins Process.
1. Model Formulation. Ind. Eng. Chem. Res. 2001, 40, 4172-4186.
(19) Martinis, J. M.; Froment, G. F., Alkylation on Solid Acids. Part 2. Single-Event
Kinetic Modeling. Ind. Eng. Chem. Res. 2006, 45, 954-967.
(20) Egan, C. J.; White, R. J.; Langlois, G. E., Selective Hydrocracking of C9Alkylcyclohexanes to C12-Alkylcyclohexanes on Acidic Catalysts - Evidence for Paring
Reaction. Journal of the American Chemical Society 1962, 84, (7), 1204-1215.
(21) Baltanas, M. A.; Froment, G. F., Computer-Generation of Reaction Networks and
Calculation of Product Distributions in the Hydroisomerization and Hydrocracking of
Paraffins on Pt-Containing Bifunctional Catalysts. Computers & Chemical Engineering
1985, 9, (1), 71-81.
(22) Govindhakannan, J. Modeling of a Hydrogenated Vacuum Gas Oil Hydrocracker.
Ph.D. Dissertation, Texas Tech. University, Lubbock TX, 2003.
(23) Miller, R. L.; Ettre, L. S.; Johansen, N. G., Quantitative-Analysis of Hydrocarbons
by Structural Group Type in Gasolines and Distillates. 3. Combined Use of Liquid and
Gas-Chromatography. Journal of Chromatography 1983, 264, (1), 19-32.
(24) Miller, R. L.; Ettre, L. S.; Johansen, N. G., Quantitative-Analysis of Hydrocarbons
by Structural Group Type in Gasolines and Distillates. 2. Liquid-Chromatography.
Journal of Chromatography 1983, 259, (3), 393-412.

172

(25) Johansen, N. G.; Ettre, L. S.; Miller, R. L., Quantitative-Analysis of Hydrocarbons

by Structural Group Type in Gasolines and Distillates. 1. Gas-Chromatography. Journal
of Chromatography 1983, 256, (3), 393-417.
(26) Szakasits, J. J.; Robinson, R. E., Hydrocarbon Type Determination of Naphthas
and Catalytically Reformed Products by Automated Multidimensional GasChromatography. Analytical Chemistry 1991, 63, (2), 114-120.
(27) Boduszynski, M. M., Characterization of Heavy Crude Components. Abstracts of
Papers of the American Chemical Society 1985, 190, (SEP), 78-88.
(28)
Boduszynski, M. M., Composition of Heavy Petroleums. 2. Molecular
Characterization. Energy & Fuels 1988, 2, (5), 597-613.
(29) Allen, D. T., Structural Models of Catalytic Cracking Chemistry. Abstracts of
Papers of the American Chemical Society 1991, 201, 27-35.
(30) Neurock, M.; Nigam, A.; Trauth, D.; Klein, M. T., Molecular Representation of
Complex Hydrocarbon Feedstocks through Efficient Characterization and Stochastic
Algorithms. Chemical Engineering Science 1994, 49, (24A), 4153-4177.
(31) Trauth, D. M.; Stark, S. M.; Petti, T. F.; Neurock, M.; Klein, M. T., Representation
of the Molecular-Structure of Petroleum Resid through Characterization and MonteCarlo Modeling. Energy & Fuels 1994, 8, (3), 576-580.
(32) Hudebine, D.; Verstraete, J. J., Molecular Reconstruction of LCO Gasoils from
Overall Petroleum Analyses. Chemical Engineering Science 2004, 59, (22-23), 47554763.
(33) Aye, M. M. S.; Zhang, N., A Novel Methodology in Transforming Bulk Properties
of Refining Streams into Molecular Information. Chemical Engineering Science 2005,
60, (23), 6702-6717.
(34) Vansina, H.; Baltanas, M. A.; Froment, G. F., Hydroisomerization and
Hydrocracking. 4. Product Distribution from N-Octane and 2,2,4-Trimethylpentane.
Industrial & Engineering Chemistry Product Research and Development 1983, 22, (4),
526-531.
(35) Schulz, H. F.; Weitkamp, J. H., Zeolite Catalysts - Hydrocracking and
Hydroisomerization of N-Dodecane. Industrial & Engineering Chemistry Product
Research and Development 1972, 11, (1), 46-53.
(36) Froment, G. F., Single Event Kinetic Modeling of Complex Catalytic Processes.
Catalysis Reviews-Science and Engineering 2005, 47, (1), 83-124.

173

(37) Eyring, H., The Activated Complex and the Absolute Rate of Chemical Reactions.
Chemical Reviews 1935, 17, (1), 65-77.
(38) Girgis, M. J.; Gates, B. C., Reactivities, Reaction Networks, and Kinetics in HighPressure Catalytic Hydroprocessing. Industrial & Engineering Chemistry Research
1991, 30, (9), 2021-2058.
(39) Lin, S. D.; Vannice, M. A., Hydrogenation of Aromatic-Hydrocarbons over
Supported Pt Catalysts. 1. Benzene Hydrogenation. Journal of Catalysis 1993, 143, (2),
539-553.
(40) Lin, S. D.; Vannice, M. A., Hydrogenation of Aromatic-Hydrocarbons over
Supported Pt Catalysts. 2. Toluene Hydrogenation. Journal of Catalysis 1993, 143, (2),
554-562.
(41) Lin, S. D.; Vannice, M. A., Hydrogenation of Aromatic-Hydrocarbons over
Supported Pt Catalysts. 3. Reaction Models for Metal-Surfaces and Acidic Sites on
Oxide Supports. Journal of Catalysis 1993, 143, (2), 563-572.
(42) Lin, S. D.; Vannice, M. A.; Herrmann, J. M.; Wang, D.; Apestequia, C.; Duprez,
D.; Figueras, F.; Conner, W. C.; Kiperman, S. L.; Hall, W. K.; Blackmond, D. G.;
Grunert, W.; Butt, J. B.; Schulz, H., Toluene Hydrogenation over Supported Platinum
Catalysts. Studies in Surface Science and Catalysis 1993, 75, 861-874.
(43) Girgis, M. J.; Gates, B. C., Catalytic Hydroprocessing of Simulated Heavy Coal
Liquids. 2. Reaction Networks of Aromatic-Hydrocarbons and Sulfur and Oxygen
Heterocyclic-Compounds. Industrial & Engineering Chemistry Research 1994, 33, (10),
2301-2313.
(44) Girgis, M. J.; Gates, B. C., Catalytic Hydroprocessing of Simulated Heavy Coal
Liquids. 1. Reactivities of Aromatic-Hydrocarbons and Sulfur and Oxygen HeterocyclicCompounds. Industrial & Engineering Chemistry Research 1994, 33, (5), 1098-1106.
(45) Klein, M. T.; Korre, S. C.; Read, C. J.; Russell, C. L., Hydrocracking of Model
Polynuclear Aromatics - Pathways, Kinetics, and Structure Reactivity Correlations.
Abstracts of Papers of the American Chemical Society 1993, 205, 43-51.
(46) Korre, S. C.; Klein, M. T., Development of Temperature-Independent Quantitative
Structure/Reactivity Relationships for Metal- and Acid-Catalyzed Reactions. Catalysis
Today 1996, 31, (1-2), 79-91.
(47) Korre, S. C.; Klein, M. T.; Quann, R. J., Effect of Temperature on Quantitative
Structure-Reactivity Correlations for Hydrocracking Polynuclear Aromatics. Abstracts
of Papers of the American Chemical Society 1995, 210, 120-128.

174

(48) Korre, S. C.; Klein, M. T.; Quann, R. J., Polynuclear Aromatic-Hydrocarbons

Hydrogenation. 1. Experimental Reaction Pathways and Kinetics. Industrial &
Engineering Chemistry Research 1995, 34, (1), 101-117.
(49) Korre, S. C.; Klein, M. T.; Quann, R. J., Hydrocracking of Polynuclear Aromatic
Hydrocarbons. Development of Rate Laws Through Inhibition Studies. Industrial &
Engineering Chemistry Research 1997, 36, (6), 2041-2050.
(50) Korre, S. C.; Neurock, M.; Klein, M. T.; Quann, R. J., Hydrogenation of
Polynuclear Aromatic-Hydrocarbons. 2. Quantitative Structure-Reactivity Correlations.
Chemical Engineering Science 1994, 49, (24A), 4191-4210.
(51) Rahier, H.; Denayer, J. F.; Van Mele, B., Low-Temperature Synthesized
Aluminosilicate Glasses - Part IV - Modulated DSC Study on the Effect of Particle Size
of Metakaolinite on the Production of Inorganic Polymer Glasses. Journal of Materials
Science 2003, 38, (14), 3131-3136.
(52) Rautanen, P. A.; Aittamaa, J. R.; Krause, A. O. I., Liquid Phase Hydrogenation of
Tetralin on Ni/Al-2 O-3. Chemical Engineering Science 2001, 56, (4), 1247-1254.
(53) Rautanen, P. A.; Aittamaa, J. R.; Krause, K. O. I., Solvent Effect in Liquid-Phase
Hydrogenation of Toluene. Industrial & Engineering Chemistry Research 2000, 39,
(11), 4032-4039.
(54) Rautanen, P. A.; Lylykangas, M. S.; Aittamaa, J. R.; Krause, A. O. I., Liquid-Phase
Hydrogenation of Naphthalene and Tetralin on Ni/Al2O3: Kinetic Modeling. Industrial
& Engineering Chemistry Research 2002, 41, (24), 5966-5975.
(55) Lylykangas, M. S.; Rautanen, P. A.; Krause, A. O. I., Liquid-Phase Hydrogenation
Kinetics of Multicomponent Axomatic Mixtures on Ni/Al2O3. Industrial & Engineering
Chemistry Research 2002, 41, (23), 5632-5639.
(56) Lylykangas, M. S.; Rautanen, P. A.; Krause, A. O. I., Hydrogenation and
Deactivation Kinetics in the Liquid-Phase Hydrogenation of Isooctenes on Pt/Al2O3.
Industrial & Engineering Chemistry Research 2004, 43, (7), 1641-1648.
(57) Singh, U. K.; Vannice, M. A., Kinetics of Liquid-Phase Hydrogenation Reactions
over Supported Metal Catalysts - A Review. Applied Catalysis A-General 2001, 213,
(1), 1-24.
(58) Singh, U. K.; Vannice, M. A., Kinetics of Liquid-Phase Hydrogenation Reactions
over Supported Metal Catalyst - A Review. Applied Catalysis A-General 2003, 255, (2),
361-361.

175

(59) Stanislaus, A.; Cooper, B. H., Aromatic Hydrogenation Catalysis - A Review.

Catalysis Reviews-Science and Engineering 1994, 36, (1), 75-123.
(60) Aben, P. C.; Platteeu, J. C.; Stoutham, B., Hydrogenation of Benzene over
Supported Plantinum, Palladium and Nickel Catalysts. Recueil Des Travaux Chimiques
Des Pays-Bas 1970, 89, (5), 449-455.
(61) Vanmeerten, R. Z. C.; Coenen, J. W. E., Gas-Phase Benzene Hydrogenation on a
Nickel-Silica Catalyst. 1. Experimental-Data and Phenomenological Description.
Journal of Catalysis 1975, 37, (1), 37-43.
(62) Chou, P.; Vannice, M. A., Benzene Hydrogenation over Supported and
Unsupported Palladium. 1. Kinetic-Behavior. Journal of Catalysis 1987, 107, (1), 129139.
(63) Chou, P.; Vannice, M. A., Benzene Hydrogenation over Supported and
Unsupported Palladium. 2. Reaction Model. Journal of Catalysis 1987, 107, (1), 140153.
(64) Singh, U. K.; Vannice, M. A., Kinetic and Thermodynamic Analysis of LiquidPhase Benzene Hydrogenation. AIChE Journal 1999, 45, (5), 1059-1071.
(65) Cohen, N.; Benson, S. W., Estimation of Heats of Formation of OrganicCompounds by Additivity Methods. Chemical Reviews 1993, 93, (7), 2419-2438.
(66) Peng, D.; Robinson, D. B., New 2-Constant Equation of State. Industrial &
Engineering Chemistry Fundamentals 1976, 15, (1), 59-64.
(67) Dufresne, P.; Bigeard, P. H.; Billon, A., New Developments in Hydrocracking:
Low Pressure High-Conversion Hydrocracking. Catalysis Today 1987, 1, (4), 367-384.
(68) Melis, S.; Erby, L.; Sassu, L.; Baratti, R., A Model for the Hydrogenation of
Aromatic Compounds During Gasoil Hydroprocessing. Chemical Engineering Science
2004, 59, (22-23), 5671-5677.
(69) Lavopa, V.; Satterfield, C. N., Poisoning of Thiophene Hydrodesulfurization by
Nitrogen-Compounds. Journal of Catalysis 1988, 110, (2), 375-387.
(70) Martens, G. G.; Marin, G. B., Kinetics for Hydrocracking Based on Structural
Classes: Model Development and Application. AIChE Journal 2001, 47, (7), 1607-1622.
(71) Hunter, E. P. L.; Lias, S. G., Evaluated Gas Phase Basicities and Proton Affinities
of Molecules: An Update. Journal of Physical and Chemical Reference Data 1998, 27,
(3), 413-656.

176

(72) Dumesic, J. A.; Rudd, D. F.; Aparicio, L. M.; Rekoske, J. E.; Trevino, A. A., The
Microkinetics of Hetrogeneous Catalysis. American Chemical Society: Washington, DC,
1993.
(73) Martinis, J. M.; Froment, G. F., Alkylation on solid acids. Part 2. Single-Event
Kinetic Modeling. Industrial & Engineering Chemistry Research 2006, 45, (3), 954-967.
(74) Park, T. Y.; Froment, G. F., Kinetic Modeling of the Methanol to Olefins Process.
1. Model Formulation. Industrial & Engineering Chemistry Research 2001, 40, (20),
4172-4186.
(75) Park, T. Y.; Froment, G. F., Kinetic Modeling of the Methanol to Olefins Process.
2. Experimental Results, Model Discrimination, and Parameter Estimation. Industrial &
Engineering Chemistry Research 2001, 40, (20), 4187-4196.
(76) Denayer, J. F.; Bouyermaouen, A.; Baron, G. V., Adsorption of Alkanes and Other
Organic Molecules in Liquid Phase and in the Dense Vapor Phase: Influence of Polarity,
Zeolite Topology, and External Fluid Density and Pressure. Industrial & Engineering
Chemistry Research 1998, 37, (9), 3691-3698.
(77) Froment, G. F.; Depauw, G. A.; Vanrysselberghe, V., Kinetic Modeling and
Reactor Simulation in Hydrodesulfurization of Oil Fractions. Industrial & Engineering
Chemistry Research 1994, 33, (12), 2975-2988.
(78) Sato, Y.; Hirose, H.; Takahashi, F.; Toda, M., First Pacific Chemical Engineering
Congress 1972, 187-195.
(79) Reiss, L. P., Cocurrent Gas-Liquid Contacting in Packed Columns. Industrial &
Engineering Chemistry Process Design and Development 1967, 6, (4), 486-499.
(80) Charpentier, J. C., Recent Progress in Gas-Liquid Mass Transfer in Packed Beds.
Chem. Eng. Journal 1976, 11, 161-171.
(81) Whitaker, S., Forced Convection Heat Transfer Correlations or Flow in Pipes, Past
Flat Plates, Single Cylinders, Single Spheres, and for Flow in Packed Beds and Tube
Bundles. AIChE J. 1972, 18, (2), 361-371.
(82) Poling, B. E.; Prausnitz, J. M.; O'Connell, J. P., The Properties and Gases and
Liquids. Fifth ed.; McGraw-Hill: New York, 2001.
(83) Twu, C. H., Internally Consistent Correlation for Predicting Liquid Viscosities of
Petroleum Fractions. Industrial & Engineering Chemistry Process Design and
Development 1985, 24, (4), 1287-1293.

177

(84) Macknick, A. B.; Prausnitz, J. M., Vapor-Pressures of Heavy Liquid Hydrocarbons

by a Group-Contribution Method. Industrial & Engineering Chemistry Fundamentals
1979, 18, (4), 348-351.
(85) Garvin, J., Determine Liquid Specific Heat for Organic Compounds. Chemical
Engineering Progress 2002, 98, (5), 48-50.
(86) Martens, G. G.; Thybaut, J. W.; Marin, G. B., Single-Event Rate Parameters for the
Hydrocracking of Cycloalkanes on Pt/US-Y Zeolites. Industrial & Engineering
Chemistry Research 2001, 40, (8), 1832-1844.
(87) Pacheco, M. A.; Dassori, C. G., Hydrocracking: An Improved Kinetic Model and
Reactor Modeling. Chemical Engineering Communications 2002, 189, (12), 1684-1704.
(88) Moustafa, T. M.; Froment, G. F., Kinetic Modeling of Coke Formation and
Deactivation in the Catalytic Cracking of Vacuum Gas Oil. Industrial & Engineering
Chemistry Research 2003, 42, (1), 14-25.
(89) Weismantel, G. E., Petroleum Processing Handbook. Marcel Dekker: New York,
1992.
(90) Filimono.Va; Rogov, S. P.; Osipov, L. N.; Agafonov, A. V.; Perezhig.Iy; Popov, A.
A.; Khavkin, V. A., Rates of Reaction of Individual Groups of Hydrocarbons in
Hydrocracking. International Chemical Engineering 1972, 12, (1), 21-26.
(91) Sullivan, R. F.; Meyer, J. A., Catalyst Effects on Yields and Product Properties in
Hydrocracking. ACS Symposium Series 1975, 20, 28-51.
(92) Bourne, K. H.; Bennett, R. N., Hydrocracking for Middle Distillates - Study of
Process Reactions and Corresponding Product Yields and Qualities. Abstracts of Papers
of the American Chemical Society 1972, 164, G45-G62.
(93) Bhringer, W.; Kotsiopoulos, A.; Fletcher, J. C., Selective Fischer-Tropsch Wax
Hydrocracking Opportunity for Improvement of Overall Gas-to-Liquids Processing.
Paper presented at the International R&D Forum on Oil, Gas and Petrochemicals, Kuala
Lumpur, Malaysia, 2004.
(94) Schweitzer, J. M.; Galtier, P.; Schweich, D., A Single Events Kinetic Model for the
Hydrocracking of Paraffins in a Three-Phase Reactor. Chemical Engineering Science
1999, 54, (13-14), 2441-2452.
(95) Fox, J. M., The Different Catalytic Routes for Methane Valorization - An
Assessment of Processes for Liquid Fuels. Catalysis Reviews-Science and Engineering
1993, 35, (2), 169-212.

178

(96) Gregor, J. H., Fischer-Tropsch Products as Liquid Fuels or Chemicals - An

Economic-Evaluation. Catalysis Letters 1991, 7, (1-4), 317-331.
(97) Steijns, M.; Froment, G. F., Hydroisomerization and Hydrocracking. 3. KineticAnalysis of Rate Data for Normal-Decane and Normal-Dodecane. Industrial &
Engineering Chemistry Product Research and Development 1981, 20, (4), 660-668.
(98) Degnan, T. F.; Kennedy, C. R., Impact of Catalyst Acid Metal Balance in
Hydroisomerization of Normal Paraffins. AIChE Journal 1993, 39, (4), 607-614.
(99) Debrabandere, B.; Froment, G. F., Influence of the Hydrocarbon Chain Length on
the Kinetics of the Hydroisomerization and Hydrocracking of n-Paraffins. In
Hydrotreatment and Hydrocracking of Oil Fractions, Froment, G. F.; Delmon, B.;
Grange, P.` eds., Elsevier: Amsterdam, 1997.
(100) Thybaut, J. W.; Narasimhan, C. S. L.; Denayer, J. F.; Baron, G. V.; Jacobs, P. A.;
Martens, J. A.; Marin, G. B., Acid-Metal Balance of a Hydrocracking Catalyst: Ideal
Versus Nonideal Behavior. Industrial & Engineering Chemistry Research 2005, 44,
(14), 5159-5169.
(101) Chavarria, J. C.; Ramirez, J.; Gonzalez, H.; Baltanas, M. A., Modelling of nHexadecane Hydroisomerization and Hydrocracking Reactions on a Mo/H BetaAlumina Bi-functional Catalyst, Using the Single Event Concept. Catalysis Today 2004,
98, (1-2), 235-242.
(102) Pines, H., The Chemistry of Catalytic Hydrocarbon Conversions. Academic Press
Inc.: New York, 1981.
(103) Evans, M. G.; Polanyi, M., Inertia and Driving Force of Chemical Reactions.
Transactions of the Faraday Society 1938, 34, (1), 0011-0023.
(104) Curtiss, L. A.; Raghavachari, K.; Redfern, P. C.; Rassolov, V.; Pople, J. A.,
Gaussian-3 (G3) Theory for Molecules Containing First and Second-Row Atoms.
Journal of Chemical Physics 1998, 109, (18), 7764-7776.
(105) Curtiss, L. A.; Raghavachari, K.; Redfern, P. C.; Pople, J. A., Assessment of
Gaussian-3 and Density Functional Theories for a Larger Experimental Test Set. Journal
of Chemical Physics 2000, 112, (17), 7374-7383.
(106) Thybaut, J. W.; Marin, G. B.; Baron, G. V.; Jacobs, P. A.; Martens, J. A., Alkene
Protonation Enthalpy Determination from Fundamental Kinetic Modeling of Alkane
Hydroconversion on Pt/H-(US)Y-Zeolite. Journal of Catalysis 2001, 202, (2), 324-339.
(107) Martens, G.; Froment, G. F., Kinetic Modeling of Paraffins Hydrocracking Based
upon Elementary Steps and the Single Event Concept. In Reaction Kinetics and the

179

Development of Catalytic Processes, Froment, G. F.; Waugh, K. C. eds., Elsevier:

Amsterdam, 1999.

(108) Martens, J. A.; Tialen, M.; Jacobs, P. A., Attempts to Rationalize the Distribution
of Hydrocracked Products. III. Mechanistic Aspects of Isomerization and Hydrocracking
of Branched Alkanes on Ideal Bifunctional Large-Pore Zeolite Catalysts. Catalysis
Today 1987, 1, 435-453.
(109) Denayer, J. F.; Baron, G. V.; Souverijns, W.; Martens, J. A.; Jacobs, P. A.,
Hydrocracking of n-Alkane Mixtures on Pt/H-Y Zeolite: Chain Length Dependence of
the Adsorption and the Kinetic Constants. Industrial & Engineering Chemistry Research
1997, 36, (8), 3242-3247.
(110)
Alvarez, F.; Ribeiro, F. R.; Perot, G.; Thomazeau, C.; Guisnet, M.,
Hydroisomerization and Hydrocracking of Alkanes - Influence of the Balance Between
Acid and Hydrogenating Functions on the Transformation of n-Decane on PtHY
Catalysts. Journal of Catalysis 1996, 162, (2), 179-189.
(111) Guisnet, M.; Alvarez, F.; Giannetto, G.; Perot, G., Hydroisomerization and
Hydrocracking of n-Heptane on the Zeolites. Effect of the Porosity and of the
Distribution of Metallic and Acid Sites. Catalysis Today 1987, 1, 415-426.
(112) Kuznetsov, P. N., Study of n-Octane Hydrocracking and Hydroisomerization over
Pt/HY Zeolites Using the Reactors of Different Configurations. Journal of Catalysis
2003, 218, (1), 12-23.

180

VITA
Hans Kumar was born in Shamli, India on April 20, 1979. He received a Bachelor of
Engineering degree in chemical engineering from the Indian Institute of Technology,
Roorkee, India in July 2000. Then he worked as a process engineer at the Process Design
and Development Division of Engineers India Limited, New Delhi, India from July 2000
to July 2002. He joined the graduate program in the Department of Chemical
Engineering, Texas A&M University in August 2002 and obtained his Master of Science
degree in chemical engineering in August 2004 and Doctor of Philosophy degree in
chemical engineering in December 2006. He can be contacted through email
[email protected] or through his parents at his permanent address:

Hans Kumar
c/o Kuldeeep Gupta
H.No:108, Kaka Nagar
Shamli - 247776
U. P. (India)
Phone: (91)-1398-257416