study of the properties of molybdenum … work is focused on the study of the properties of...
TRANSCRIPT
STUDY OF THE PROPERTIES OF MOLYBDENUM COMPOUNDS
FOR THE CATALYTIC OXIDATION OF LOGISTIC FUELS
By
OSCAR GERARDO MARIN-FLORES
A dissertation submitted in partial fulfillment of the requirements for the degree of
DOCTOR OF PHILOSOPHY
WASHINGTON STATE UNIVERSITY School of Chemical Engineering and Bioengineering
DECEMBER 2009
ii
To the Faculty of Washington State University:
The members of the Committee appointed to examine the dissertation of
OSCAR GERARDO MARIN-FLORES find it satisfactory and recommend that
it be accepted.
___________________________________
Su Ha, Ph. D., Chair ___________________________________
Grant Norton, Ph.D. ___________________________________
Richard Zollars, Ph.D.
iii
ACKNOWLEDGMENTS
My deepest thanks to my advisor, professor Su Ha, whose friendship, encouragement and
support enabled me to define my research goals and build the necessary self-confidence
to finalize this work.
To professors Grant Norton and Jeongminh Ahn, members of the Boeing team, for their
support and contributions to achieve the goal of this work.
To professor Louis Scudiero, for his support and advice.
To my fellow graduate student Sean Lin, a good friend and colleague with whom I spent
quite some time discussing not only engineering but also daily life things.
To Tim Turba and Kang Wang, for their hard work and assistance in order to complete
this project.
To the staff members, Jo Ann McCabe, Diana Thornton and Senja Estes, for the logistic
support.
To my parents, for their guidance and example, and for their lifelong generosity in
providing for my education.
And, last but not least, to my dear wife, Lisset, and my daughter, Elizabeth, for whom
adversity and unfairness have never been an excuse not to achieve noble goals, and with
whom the magnificent future awaits.
iv
STUDY OF THE PROPERTIES OF MOLYBDENUM COMPOUNDS
FOR THE CATALYTIC OXIDATION OF LOGISTIC FUELS
Abstract
by Oscar Gerardo Marin-Flores, Ph.D.
Washington State University December 2009
Chair: Su Ha
This work is focused on the study of the properties of molybdenum compounds as
catalytic materials for the reforming of logistic fuels such as gasoline and jet fuel. The
starting point for this investigation was the previous work performed in our lab by other
fellow researchers about the study of molybdenum carbide, Mo2C, as catalytic material
for the reforming of methane. Thus, encouraged by the promising results with methane,
Mo2C was evaluated as potential catalyst for the steam reforming of liquid hydrocarbons.
Our data indicates that Mo2C displays high activity for the reforming of isooctane −a
surrogate gasoline− even at temperatures as low as 700°C. However, as the temperature
becomes lower than 700°C, the catalytic activity of Mo2C experiences a decline to
eventually disappear. The particular electron configuration of the Mo atoms appears to be
responsible for the stability issues, given that it allows phase transitions that can
significantly affect the catalytic activity. Thus, Mo2C is thought to loss activity due to an
oxidation process that leads to the formation of Mo dioxide. The presence of the species
Mo4+ (assigned to the Mo dioxide phase) on the surface of active spent samples of Mo
v
carbide led us to hypothesize that this oxide phase may play a role in the catalytic activity
of Mo2C. To test this hypothesis, MoO2 was tested as reforming catalyst and our
experimental results showed that MoO2 can be used as steam reforming catalyst with a
higher activity than that of Mo2C itself. Nevertheless, the structure and properties of this
transition metal oxide led us to believe that Mo dioxide may work in a more efficient
manner under partial oxidation conditions. Our findings were in agreement with this
hypothesis and demonstrated that Mo exhibits high catalytic activity for the reforming of
both gasoline and jet fuel. In addition to the catalytic activity, Mo dioxide also displays a
significant coking resistance as compared with Ni catalysts. Finally, MoO2 was tested for
sulfur poisoning and found to tolerate concentrations as high as 1000 ppm of either
thiophene or benzothiophene without affecting in high extent the catalytic performance.
vi
TABLE OF CONTENTS Page ACKNOWLEDGEMENTS................................................................................................ iii ABSTRACT.........................................................................................................................iv LIST OF TABLES................................................................................................................x LIST OF FIGURES ............................................................................................................xii DEDICATION..................................................................................................................xvii ATRIBUTION AND OBJECTIVE ................................................................................ xviii
CHAPTER I: INTRODUCTION ........................................................................................1
1. FUEL CELLS FOR MOBILE APPLICATIONS..................................................1
2. HYDROGEN PRODUCTION BY REFORMING SYSTEMS ............................5
3. CATALYSTS FOR REFORMING OF LOGISTIC FUELS.................................8
4. MOLYBDENUM-BASED CATALYSTS FOR REFORMING
APPLICATIONS ..................................................................................................9
CHAPTER II: STUDY OF THE PERFORMANCE OF MO2C FOR ISO-OCTANE
STEAM REFORMING ......................................................................................................18
1. ABSTRACT .........................................................................................................18
2. INTRODUCTION ...............................................................................................19
3. THERMODYNAMIC ANALYSIS.....................................................................22
4. EXPERIMENTAL...............................................................................................23
vii
5. RESULTS AND DISCUSSION..........................................................................25
6. CONCLUSIONS ................................................................................................34
7. REFERENCES ....................................................................................................51
CHAPTER III: ACTIVITY AND STABILITY OF MoO2 CATALYST FOR THE
PARTIAL OXIDATION OF GASOLINE .........................................................................52
1. ABSTRACT .........................................................................................................52
2. INTRODUCTION ...............................................................................................53
3. EXPERIMENTAL...............................................................................................55
4. RESULTS AND DISCUSSION..........................................................................57
5. CONCLUSIONS ................................................................................................72
6. REFERENCES ....................................................................................................93
CHAPTER IV: X-RAY DIFFRACTION AND PHOTOELECTRON
SPECTROSCOPY STUDIES OF MoO2 AS CATALYST FOR THE PARTIAL
OXIDATION OF ISOOCTANE ........................................................................................94
5. ABSTRACT .........................................................................................................94
6. INTRODUCTION ...............................................................................................95
7. EXPERIMENTAL...............................................................................................97
8. RESULTS AND DISCUSSION..........................................................................99
9. CONCLUSIONS ..............................................................................................108
10. REFERENCES ..................................................................................................116
viii
CHAPTER V: THERMODYNAMIC AND EXPERIMENTAL STUDY OF THE
PARTIAL OXIDATION OF A JET A FUEL SURROGATE OVER MOLYBDENUM
DIOXIDE..........................................................................................................................118
1. ABSTRACT.......................................................................................................118
2. INTRODUCTION .............................................................................................119
3. THE Mo-O-C SYSTEM ....................................................................................120
4. THE PARTIAL OXIDATION OF N-DODECANE .........................................123
5. METHODOLOGY ............................................................................................124
6. EXPERIMENTAL.............................................................................................129
7. RESULTS AND DISCUSSION........................................................................130
8. CONCLUSIONS................................................................................................136
9. REFERENCES ..................................................................................................148
CHAPTER VI: NANOPARTICLE MOLYBDENUM DIOXIDE: A HIGHLY ACTIVE
CATALYST FOR PARTIAL OXIDATION OF AVIATION FUELS............................150
1. ABSTRACT .......................................................................................................150
2. INTRODUCTION .............................................................................................151
3. EXPERIMENTAL.............................................................................................154
4. RESULTS AND DISCUSSION........................................................................156
5. CONCLUSIONS ..............................................................................................162
6. REFERENCES .................................................................................................172
ix
CHAPTER VII: STUDY OF THE POISONING EFFECT OF BENZOTHIOPHENE
OVER MOLYBDENUM DIOXIDE DURING THE PARTIAL OXIDATION OF A JET
FUEL SURROGATE........................................................................................................173
1. ABSTRACT .......................................................................................................173
2. INTRODUCTION .............................................................................................174
3. EXPERIMENTAL.............................................................................................176
4. THERMODYNAMIC ANALYSIS...................................................................177
5. RESULTS AND DISCUSSION........................................................................179
6. CONCLUSIONS ..............................................................................................181
7. REFERENCES ..................................................................................................190
FINAL REMARKS ..........................................................................................................192
x
LIST OF TABLES
CHAPTER I
1. Molybdenum compounds in catalysis........................................................................16
CHAPTER III
2. XPS analysis of pretreated samples ...........................................................................89
3. Mo3d spectrum of spent sample obtained at O/C=0.72.............................................90
4. Long-Term stability test.............................................................................................91
5. XPS analysis of the spent sample obtained using 1000 ppm of thiophene................92
CHAPTER IV
1. MoO2 catalytic performance for the partial oxidation of isooctane
(700°C, 1 atm, O2/C = 0.72) ....................................................................................114
2. Peak position and atomic concentration for Mo oxidation states
as a function of the time on stream ..........................................................................115
CHAPTER V
1. MoO2 catalytic performance ....................................................................................148
xi
CHAPTER VII
1. Surface composition of fresh and spent samples .....................................................190
xii
LIST OF FIGURES
CHAPTER I
1. Technological issues for a FCHV..............................................................................12
2. Direct hydrogen fuel cell system ...............................................................................13
3. Power Electronics in an MEA....................................................................................14
4. 250 kW Fuel Cell System ..........................................................................................15
CHAPTER II
1. ISR equilibrium compositions as a function of temperature .....................................36
2. ISR hydrogen yield as a function of temperature and S/C ratio ................................37
3. Steam reforming reactor schematic ...........................................................................38
4. XRD pattern of commercial Mo2C ............................................................................39
5. Mo 3d XPS spectrum of commercial Mo2C ..............................................................40
6. Mo 3d XPS spectrum of pretreated commercial Mo2C .............................................41
7. ISR H2 yield as function of the WHSV and the S/C ratio
(Mo2C, He = 10 sccm, T = 850°C ) ...........................................................................42
8. Conversion as a function of WHSV at S/C=1
(pretreated commercial Mo2C, He=10 sccm, S/C=1, T=850°C ) ..............................43
9. Selectivity of Mo2C at S/C=1
(pretreated commercial Mo2C, He=10 sccm, S/C=1, T=850°C ) ..............................44
10. Effect of temperature on Mo2C catalytic activity ......................................................45
xiii
11. XPS Mo3d spectra of (a) sample 1 (inactive spent catalyst at 650°C)
and (b) sample 2 (active spent catalyst at 1000°C)....................................................46
12. Equilibrium constants .............................................................................................47
13. Carburization of MoO2 with isooctane at 700°C ....................................................48
14. XRD pattern of MoO2 carburized with isooctane...................................................49
15. XPS Mo3d pattern of MoO2 carburized with isooctane .........................................50
CHAPTER III
1. XPS spectra of fresh MoO2........................................................................................73
2. Equilibrium constants for the anaerobic oxidation of isooctane................................74
3. Equilibrium constants for formation of carbide phase...............................................75
4. Concentration profiles for the non-catalytic oxidation of isooctane..........................76
5. Diffraction patterns of spent samples used in the non-catalytic oxidation tests. .......77
6. Equilibrium constants for the reduction of MoO2 and MoO3 with hydrogen............78
7. Diffraction patterns of pretreated samples .................................................................79
8. Mars-van Krevelen mechanism .................................................................................80
9. Isooctane partial oxidation: thermodynamic analysis................................................81
10. Effect of O/C ratio on catalytic performance of MoO2
for the partial oxidation of isooctane at 700oC and 1 atm..........................................82
11. XRD patterns of spent samples at different O/C ratios..............................................83
12. XRD patterns of spent samples obtained at O/C=0.72 ..............................................84
13. Catalytic performance of MoO2 at different thiophene concentrations .....................85
14. XRD patterns of spent samples obtained at different thiophene concentrations .......86
xiv
15. Comparison between the performances of MoO2 and a nickel catalyst
for the partial oxidation of premium gasoline............................................................87
16. C1s XPS spectrum of spent MoO2 sample obtained after
the coking-resistance test ...........................................................................................88
CHAPTER IV
1. a) XRD pattern of the spent samples as a function of time;
b) Mo carbide to Mo dioxide ratio as a function of time .........................................109
2. (a) XPS spectra of Mo3d for MoO2 as a function of the time on stream
(0, 0.5, 9, 20 and 43 h). (b) XPS spectra of C1s for 9, 20 and 40 h on stream .......110
3. Binding energy (BE) versus oxidation number of Mo atom based on
the 43-hour sample...................................................................................................111
4. Deconvolution of XPS spectra of Mo3d from Figure 2 at 0.5 and 43 hours ...........112
5. (Left) XPS valence band as function of time on stream. (Right) UPS He I
spectra for the four samples and the clean Mo foil reference (inset).......................113
CHAPTER V
1. Ternary Phase diagram Mo-O-C at 850°C and 1 atm..............................................138
2. Equilibrium constants ..............................................................................................139
3. Diffractograms of spent samples obtained at different O2/C ratios .........................140
4. Effect of O2/C ratio on MoO2 stability (initial moles of MoO2 = 0.10)..................141
5. Diffractograms of spent samples at different space velocities (O2/C=0.7)..............142
6. Effect of space velocity on MoO2 stability (initial moles of MoO2 = 0.10) ............143
xv
7. H2 and CO yields at equilibrium conditions ............................................................144
8. Carbon formation (Inset: carbon yield profile at 850°C).........................................145
9. Hydrogen and CO yields as functions of the O2/C ratio..........................................146
10. SEM-EDX analysis of spent samples obtained from the coking-resistance test .....147
CHAPTER VI
1. Experimental arrangement for catalytic activity measurements..............................164
2. SEM and TEM images of commercially available MoO2 (a and c)
and nanoparticle MoO2 (b and d).............................................................................165
3. XRD patterns of MoO2 samples ..............................................................................166
4. Catalytic activity of nanoparticle MoO2 for partial oxidation of n-dodecane .........167
5. Effect of fuel flow rate in terms of WHSV on the catalytic activity
of nanoparticle MoO2............................................................................................................................................168
6. Hydrogen yield as a function of WHSV..................................................................169
7. SEM and TEM images of spent samples of commercial MoO2 (a and c)
and nanoparticle MoO2 (b and d).............................................................................170
8. The catalytic redox cycle based on the Mars-van Krevelen reaction mechanism ...171
CHAPTER VII
1. Equilibrium concentrations of sulfur compounds at 850°C.....................................183
2. Equilibrium constants for the reactions between Mo dioxide and sulfur oxides .....184
3. Mo dioxide catalytic activity tests at 850°C ............................................................185
4. Concentrations of CH4 and CO2 as function of time ...............................................186
xvi
5. XRD pattern of spent sample...................................................................................187
6. SEM micrographs of fresh and spent samples.........................................................188
7. TEM scan and electron diffraction pattern of spent sample ....................................189
xvii
Dedication
This dissertation is dedicated to my parents, my wife and daughter
who provided the emotional support that I required to achieve this goal
xviii
ATTRIBUTION AND OBJECTIVE
Chapters II-VII form the body of this dissertation, and either have been published, are in
review, or have been submitted to peer reviewed journals. A short summary of each
chapter is given below, including author attributes.
Chapter II, entitled “Study of the performance of Mo2C for iso-octane steam reforming”
was carried out to investigate the catalytic properties of Mo2C and the effect of operating
parameters such as temperature, steam-to-carbon ratio and space velocity, on the catalyst
performance. Our findings indicate that Mo2C displays high activity for steam reforming
of a surrogate gasoline even at temperatures as los as 700ºC, although the highest
stability, within the temperature range used in this work, was found at 850ºC. The
catalyst stability was explained in terms of two concurrent reactions: the oxidation of the
carbide phase to Mo dioxide, and the regeneration of the carbide phase via a
carburization process using the hydrocarbon as carburizing agent. This study has been
published in Catalysis Today 136 (2008) 235.
Chapter III, entitled “Activity and stability of MoO2 catalyst for the partial oxidation of
gasoline”, summarizes the results obtained when using MoO2 as catalytic material for the
reforming of both iso-octane and actual gasoline. The justification for this study was in
the presence of the species Mo4+ (commonly assigned to the Mo dioxide phase) on the
surface of the active spent samples of Mo carbide as concluded from the XPS data
xix
showed in the previous chapter. This finding led us to hypothesize that this oxide phase
may play a role in the catalytic activity of Mo2C. The experimental results reported in this
work supported this hypothesis and demonstrated that Mo dioxide exhibits not only high
catalytic activity but also significant resistance to deactivation due to coking and
tolerance to high concentration of sulfur compounds in the fuel. Based on the particular
properties displayed by this oxide and the analysis of the activity measurements, a
probable mechanism of reaction based on the Mar-van Krevelen approach was proposed.
This work has been published in Journal of Applied Catalysis A 352 (2009) 124.
Chapter IV, entitled “X-ray diffraction and photoelectron spectroscopy studies of MoO2
as catalyst for the partial oxidation of isooctane”, is the continuation of the previous
chapter. The goal of this study was to perform a deeper analysis of the catalyst structure,
at both bulk and surface levels, in order to determine the identity of the active sites on the
surface of commercial MoO2 responsible for the catalytic activity. The resulting data was
interpreted on the basis of the Mars-van Krevelen mechanism of reaction and allowed
gaining a better insight into the relationship between the catalytic activity and the
concentration of Mo4+ sites on the catalyst surface. This study has been published in
Surface Science 603 (2009) 2327. Louis Scudiero contributed with his expertise on
surface analysis techniques and valuable suggestions.
Chapter V, entitled “Thermodynamic and experimental study of the partial oxidation of a
jet a fuel surrogate over molybdenum dioxide”, is our first study focused on the
reforming of jet fuel via partial oxidation. The objective of this work was to perform a
complete analysis of the reforming process combining thermodynamic calculations and
xx
experimental results to investigate not only the theoretical efficiency of the process but
also the stability of commercial Mo dioxide during the partial oxidation of n-dodecane, a
jet fuel surrogate. The results from this investigation led us to establish a stability
window for MoO2 in terms of the O2/C ratio, which must be within the range 0.5-1.0 to
ensure the presence of enough amount of Mo dioxide in the catalyst structure. This study
is in preparation for submission to the Journal of Applied Catalysis A.
Chapter VI, entitled “Nanoparticle molybdenum dioxide: a highly active catalyst for
partial oxidation of aviation fuels”, is the result of our efforts for improving the catalytic
properties of Mo dioxide catalysts. A procedure was developed to prepare nanoparticle
MoO2 with a surface area larger than that of commercial Mo dioxide. Our measurements
indicated that the home-made nanoparticles exhibited a surface area of 45 m2/g against
4.8 for commercial MoO2. The average particle size for nanoparticle MoO2 was
approximately 20 nm and our catalytic activity measurements demonstrated that
nanoparticle MoO2 is able to prevent coking formation in a more efficient manner than
commercial Mo dioxide due to their higher concentration of active sites. This study is in
preparation for submission to the Journal of Applied Catalysis B. Grant Norton,
Jeongminh Ahn and Joe Breit provided with valuable suggestions and technical expertise
whereas Tim Turba and Wang Kang contributed with their hardwork to the
accomplishment of our objective.
Chapter VII, entitled “Study of the poisoning effect of benzothiophene over molybdenum
dioxide during the partial oxidation of a jet fuel surrogate”, deals with the analysis of the
deactivation process due to high concentrations of benzothiophene -a model sulfur
xxi
compound for aviation fuels- in the feed stream. Thermodynamic calculations as well as
experimental measurements were used to investigate the effect of sulfur compounds on
the catalyst structure, which is a serious issue for catalytic materials intended for the
reforming of logistic fuels. This work is in preparation for submission to the Journal of
Applied Catalysis B. Grant Norton, Jeongminh Ahn and Joe Breit provided with technical
expertise. Tim Turba and Wang Kang contributed with their hardwork to the completion
of this study.
1
CHAPTER I
INTRODUCTION
FUEL CELLS FOR MOBILE APPLICATIONS
Fuel cells convert chemical energy directly to electricity. It is intrinsically much more
efficient than conventional energy conversion systems and is an important new path for
efficient, clean and sustainable energy development. A major reason for the greater
interest in hydrogen energy now worldwide is that fuel cell technologies using H2 as a
fuel have advanced to the extent where many people begin to see its major commercial
application potentials. There are five types of fuel cells including polymer electrolyte
membrane fuel cell (PEMFC), alkali fuel cell, phosphoric acid fuel cell, molten carbonate
fuel cell, and solid oxide fuel cell (SOFC). Among the five types, SOFC and PEMFC are
the two most promising fuel cells.
Fuel Cell-Powered Vehicles
The most advanced use of hydrogen in the transport sector is for motor vehicles,
especially cars. This is why automotive manufacturers all over the world are investing
large amounts of money into Hydrogen Research & Development. Fuel cell vehicles are
widely regarded to be the most likely altemative fuel technology to replace internal
combustion engine vehicles in coming decades.
2
Toyota analyzed both technologies in terms of efficiency by comparing the Toyota
Highlander FCHV-adv hydrogen fuel cell vehicle and the Toyota Highlander Hybrid
gasoline-powered vehicle. The Toyota Highlander FCHV-adv hydrogen fuel cell vehicle
achieved an incredible average fuel economy of 68.3 miles per kilogram of hydrogen in a
real-world road test with the Department of Energy. On the other hand, the Toyota
Highlander Hybrid that is powered by gasoline attains an EPA-estimated 26 miles per
gallon, which indicates that Toyota hydrogen fuel cells are more than twice as efficient as
the company’s internal combustion engines.
The low emissions and high energy efficiency associated with fuel cells suggest that fuel
cells can have a strong impact on the transportation industry. In the case of automobiles,
the operation with fuel cells constrains the establishment of a hydrogen infrastructure,
while buses for public transportation require only one central hydrogen production
facility, where the refueling of a whole fleet may be scheduled to operate the facility
efficiently. Moreover, the weight of the storage tank represents a higher portion in the
total weight of a smaller vehicle, such as an automobile. Therefore, for automobiles, fuel
processing to hydrogen is also considered, if those are intended to be operated with fuel
cells.
In the early stages of fuel cell vehicle development, there was uncertainty as to how, or
whether, to hybridize a fuel cell power train. A popular notion was that fuel cell power
plants should be used exactly like internal combustion engines. That is, as stand alone
power producers that responded almost instantaneously to changes in load.
Unfortunately, experience showed that fuel cell systems did not respond to load changes
3
as quickly as their ICE counterparts. As a result, a parallel hybrid strategy was
implemented to mitigate the load response issue as well as improve fuel economy
(through regenerative braking). The first of its kind at the time, this hybridization strategy
is common throughout the fuel cell vehicle development arena today [1].
Before popularization of the fuel cell hybrid vehicle (FCHV) is possible, the following
three conditions must be met [2].
(1) Improved vehicle marketability: resolution of technological issues (see Fig. 1), cost
reduction, and enhanced appeal as a product;
(2) Establishment of societal infrastructure for hydrogen energy: technologies to produce,
transport, and store hydrogen that address concerns over CO2 emissions reduction, as well
as the establishment of codes and standards, and the establishment of infrastructure;
(3) Increased demand from society: demand stemming from strong awareness of global
warming, depletion of fossil fuels, energy security, and the like.
The fuel cell stack is the heart of a fuel cell system. However, without auxiliary
components such as the air compressor, humidifier and pressure and flow regulators, the
stack itself would not work. Fuel cell system configurations vary considerably in
different applications. A direct hydrogen fuel cell system, as shown in Fig. 2, typically
involves air supply and control, fuel supply and control, water management, and thermal
management subsystems [3].
4
Fuel Cell APU for aircrafts
Growing concerns over global climate change and national energy security are greatly
impacting future aircraft design for companies like Boeing to construct cleaner, quieter
and more fuel efficient airplanes [4]. A concept called the More Electric Aircraft (MEA)
can allow us to build more efficient aircraft by substituting hydraulically and
pneumatically driven systems by electrical ones (see Fig. 3). This increased electrical
power demand in MEA can be met by decentralizing the power producing units using
small individual devices like fuel cells [5]. Furthermore, existing aircrafts use a low
efficiency gas turbine auxiliary power unit (APU) to provide electrical power for
operating navigation systems and various other electronic devices. On average, gas
turbine APUs are 15 percent efficient at converting jet fuel into electricity. In contrast,
fuel cell APUs will be 60 percent efficient, which will reduce by three quarters the
amount of fuel an airplane uses to generate ground power. In addition, because a fuel cell
APU produces electricity electrochemically, not through combustion, its oxides of
nitrogen and sulfur emissions will be negligible and its carbon dioxide emissions will be
very low.
By replacing the conventional APU with a fuel cell APU, improvement can be made in
providing a means to obtain this power without burning as much fuel in the APU when
the airplane is on the ground and increasing the load on the main engines during the
flight. Thus, fuel cells will potentially become the primary power source and then engine
driven generators will become backup sources for future aircrafts.
5
The replacement of existing systems with electric equivalents has, and will continue to
significantly increase the electrical power requirement. It is because of the
implementation of the MEA into production airplanes, that it is even more important that
new electrical power technologies such as fuel cells continue to be advanced, not just for
stationary and ground based application, but also for aircraft implementation. This in
conjunction with the growing need for the world to decrease detrimental emissions,
including noise, is why fuel cells hold so much promise. The use of MEA on aircraft
offers significant cost benefits with lower recurring costs due to fewer parts, integration
of key sub-systems and multi-use of components. It also reduces the overall cost of
operation and ownership because MEA helps reduce fuel consumption, increasing overall
aircraft performance and energy usage. Reduced maintenance and ground support also
help to lower the cost of ownership and operation.
In 2008, Boeing announced that, for the first time in aviation history, it has flown a
manned airplane powered by hydrogen fuel cells. A two-seat Dimona motor-glider with
a 16.3 meter (53.5 foot) wingspan was used as the airframe. Built by Diamond Aircraft
Industries of Austria, it was modified to include a Proton Exchange Membrane (PEM)
fuel cell/lithium-ion battery hybrid system to power an electric motor coupled to a
conventional propeller. Three test flights took place in February and March at the
airfield in Ocaña, south of Madrid, operated by the Spanish company SENASA. During
the flights, the pilot of the experimental airplane climbed to an altitude of 1,000 meters
(3,300 feet) above sea level using a combination of battery power and power generated
by hydrogen fuel cells. Then, after reaching the cruise altitude and disconnecting the
6
batteries, the pilot flew straight and level at a cruising speed of 100 kilometers per hour
(62 miles per hour) for approximately 20 minutes on power solely generated by the fuel
cells. According to Boeing researchers, PEM fuel cell technology potentially could
power small manned and unmanned air vehicles. Over the longer term, solid oxide fuel
cells could be applied to secondary power-generating systems, such as auxiliary power
units for large commercial airplanes.
HYDROGEN PRODUCTION BY REFORMING SYSTEMS
There has been a large effort to produce hydrogen for fuel cell applications from various
hydrocarbon sources. For example, reforming of liquid logistic fuels has become
important for a variety of applications such as remote power production via PEMFC and
SOFC as well as ICE combustion enhancement with synthesis gas. The high-energy
density and existing refueling infrastructure of petroleum-derived heavy hydrocarbon
fuels, such as gasoline and jet fuel, have made them popular in transportation, military,
and industrial applications.
The Reforming Process
Within the context of this work, reforming is defined as the catalytic process intended to
produce hydrogen or syngas (H2+CO) using different compounds as feedstock. A
reforming process can be performed by steam reforming (SR), partial oxidation (POX)
and autothermal reforming, which is basically a combination of SR and POX. The first
one is more suitable to a stationary system rather than a mobile system because of its
7
slow start-up. The steam reforming reaction is an endothermic reaction and requires
heating (more fuel consumption). Furthermore, it is able to produce a high hydrogen
concentration in the reformed gas, about 70%. Therefore the efficiency of hydrogen
production is high. However, the steam supply of the reactor is a problem for mobile
systems. The second and the latter offer a faster start-up time and a better transient
response. These processes, although they do not require external heating, produce a
lower hydrogen concentration in the reformed gas. In fact, both POX and autothermal
reforming involve a dilution of the hydrogen with nitrogen (air is used as an oxidant)
and the hydrogen concentration is about 35% in volume. There are some additional
problems when the reformed gas is used in an internal combustion engine as additional
fuel to gasoline since it involves a reduction of volumetric efficiency [6, 7, 8].
Currently there is considerable interest in fuel reforming for pollution abatement in
automotive applications with internal combustion engines. Reforming of gasoline into H2
and other small molecules creates a fuel that burns very efficiently, thus reducing or
eliminating exhaust emissions of hydrocarbons, CO, and particulate matter. To become
commercially viable, the reforming process requires a better understanding of the
complex interactions between gas phase and surface reactions, particularly, the formation
of coking precursors. A combined approach of both experimental studies and modeling
should lead to gain a better insight into the interaction of catalytic surface reactions and
homogeneous gas-phase reactions as well as mass and heat transport.
8
Challenges in the reforming of logistic fuels
The catalytic reforming of logistic fuels has become a challenging research area mainly
due to: (i) the complexity and variation of the fuel composition, which demands a rapid
analysis of the product composition to study many components and mixtures at varying
parameters. (ii) Logistic transportation fuels are tailored to have optimum properties for
their oxidation in internal combustion engines. In order to achieve adequate ignition and
combustion properties such as volatility, heat of combustion, freezing point, ignition
delay time and others, logistic transportation fuels are blended by mixing of distilled
fractions of crude oil with processed streams of a refinery. As a result, the chemical
composition of logistic fuels changes within broad ranges.
Thus, to simplify the analysis, the studies of the catalytic partial oxidation of logistic
fuels are commonly performed using surrogate fuels such as iso-octane, for gasoline, or
n-dodecane, for jet fuel. The use of surrogates allows a reliable standardization and easy
reproduction of reforming experiments. Moreover, the influence and interaction of
dominant constituents can be explored, allowing the development of detailed models for
the partial oxidation of logistic fuels.
CATALYSTS FOR REFORMING OF LOGISTIC FUELS
The deactivation of the catalytic material during reforming of logistic fuels is one of the
major challenges in producing hydrogen for fuel cell applications. One major challenge is
associated with carbon deposition, since the high temperatures involved in reforming
9
processes favor the formation of carbon. Though carbon deposition can be somewhat
controlled using excess steam, the economics of the process does not allow operating at
high steam to carbon ratios. There are several mechanisms of carbon deposition,
including (i) chemisorption as a monolayer or physisorption in multilayers that block
access of reactants to metal surface sites, (ii) encapsulation of a metal particle and
complete deactivation of that particle, (iii) deposition in pores such that access of
reactants is denied, and (iv) formation of carbon whiskers. It has been said that
graphitization of coke deposited on the metal was one of the main reasons for
deactivation. Graphitic carbon formation on Ni surface was first reported by Dent et al.
[9]. Irrespective of the type of coke, carbon deposition always involves an initial step of
dehydrogenation and formation of unsaturated species which are able to migrate, either in
the gaseous phase or in the adsorbed phase. There is a tendency for coking to increase as
unsaturation, molecular weight, and aromaticity increase, and thus olefins and aromatics
are major coke precursors.
While coke lay down from heavy hydrocarbon and aromatic compounds at reaction
temperature can still be a concern, the poisoning of catalyst due to adsorption of sulfur
present in logistic fuels is by far a more serious problem. Extensive work has been
recently focused on the development of sulfur-tolerant catalysts. Studies have been
conducted using catalysts designed for reforming of low molecular weight hydrocarbons
and reveal that sulfur is adsorbed on the same sites as those involved in carbon formation.
Studies with catalyst surfaces, including single-crystal metal surfaces, have established
that sulfur inhibits the chemisorption of small molecules (H2, CO, NO, C2H4, etc.), and
10
thus leads to deactivation in the catalyst effectiveness for CO methanation, alkane
hydrogenolysis, olefin hydrogenation, and the water–gas shift reaction.
Conventional catalytic formulations are based on nickel as active component, which is
well-dispersed on a catalyst support such as alumina or silica [10]. Nickel-based catalysts
are prone to deactivation due to coking. In addition, while nickel alone is active in the
absence of sulfur in the feed, its performance experiences a serious decline when sulfur is
present in the feed. Addition of noble metals such as Rh can help in attaining a better
performance by reducing coking significantly and preventing sintering; however, the use
of noble metals in the formulation significantly increases the cost, precluding it thus from
extended use.
MOLYBDENUM-BASED CATALYSTS FOR REFORMING APPLICATIONS
Over a period of years, molybdenum-containing materials have found uses as catalysts in
several industrial applications [11]. For example, iron molybdate is one of the standard
catalysts for the oxidation of methanol to formaldehyde, cobalt and nickel molybdates are
catalysts in the hydrotreating of petroleum, and bismuth molybdates reportedly are active
in the ammoxidation of propylene to acrylonitrile. Table 1 shows some Mo-containing
compounds and their application as catalysts in different chemical processes.
Mo2C catalysts have drawn much attention for oxidative fuel reforming due to a number
of publications emanating from a group at Oxford, who claimed that bulk molybdenum
carbide catalysts were capable of methane reforming under stoichiometric feed conditions
11
without coking problems at high pressures [12]. Later, it was showed that these catalysts
could also be stabilized, even at atmospheric pressure, by operating a reforming reactor
with very high recycle ratios [13]. Mo2C was also studied for its performance in water-
gas shift reactions. Although considerable work was done on the performance of bulk
Mo2C catalysts for dry methane reforming, partial oxidation of methane, and steam
methane reforming reactions, few studies have been published about the reforming of
higher hydrocarbons, such as gasoline or aviation fuels. Cheekatamarla et al. [14] used
Mo2C for the reforming of iso-octane and demonstrated that bulk molybdenum carbide
(Mo2C) catalysts are essentially coke resistant at low steam/carbon (S/C) ratios and, in a
net carburizing environment, are resistant to oxidation by steam and oxygen during iso-
octane reforming [15]. Earlier studies have also shown that bulk Mo carbide has a high
resistance to sulfur in methane and iso-octane reforming.
Mo dioxide has been reported as catalyst for the dehydrogenation of alcohols [16] and the
isomerization of hydrocarbons [17]. However, to our best knowledge, no study has been
devoted to investigate the catalytic activity of Mo dioxide for the reforming of
hydrocarbons for hydrogen production. In that sense, our work will extend the knowledge
about this topic and pioneer the use of this catalytic material in a wider variety of
reforming applications.
12
Figure 1. Technological issues for a FCHV (Source: Toyota)
13
Figure 2. Direct hydrogen fuel cell system
14
Figure 3. Power Electronics in an MEA
15
Figure 4. 250 kW Fuel Cell System
16
Catalyst Application Reaction Importance
Sulfided Co-Mo or Ni-Mo on alumina
Hydrotreating, hydrodesulfurisation
Remove sulfur from crude petroleum
Oil and petroleum refining
Bi-Mo oxides Propene selective oxidation, ammoxiation
Synthesis acrolein, acrylonitrile
Making polymers and plastics
Mo-V oxides Acrolein oxidation Synthesis acrylic acid Making polymers and plastics
Fe-Mo oxides Methanol oxidation Synthesis formaldehyde
Making formalin, polymers, resins
Mo oxide on alumina Olefin metathesis Propene to ethene and butene
Olefin synthesis
Mo complexes Epoxidation Olefin to epoxide Polyether synthesis
Heteropolyacids-phosphomolybdate
Propene hydration Propene to alcohol Alcohols synthesis
Table 1: Molybdenum compounds in catalysis (Source: IMOA)
17
References
[1] Frenette G., Forthoffer D., Int. J. Hyd. Energy, 34 (2009) 3578.
[2] Morita T., Hojima K., ECS Transactions, 16 (2008) 185.
[3] Zhao H., Burke A., J. Power Sources 186 (2009) 408.
[4] Breit J., Szydlo-Moore J., 45th AIAA Aerospace Sciences Meeting and Exhibit, 8 -
11 January 2007, Reno, Nevada
[5] Adams, C., A380: `More Electric' Aircraft. Avionics October 1, 2001.
[6] Sinan O, Becerik I, Energy & Fuels, 23 (2009) 1858.
[7] Minutillo M., Int. J. Hyd. Energy, 30 (2005) 1483.
[8] Liu D. Kaun T.m Liao H., Ahmed S., J. Hyd. Energy, 29 (2004) 1035.
[9] Dent F., Moignard L., Eastwood A., Blackburn W., Trans Inst. Gas Eng. (1945-
1946) 602.
[10] Lakhapatri S., Abraham M., Appl. Catal. A: Gen, 364 (2009) 113.
[11] Tsigdinos G., Swanson W. Ind. Eng. Chem. Prod. Res. Dev. 17 (1978) 208.
[12] York A., Chem. Commun. 1 (1997) 39.
[13] Sehested J., Jacobsen C., Rokini S., Rostrup-Nielsen J., J. Catal. 201 (2001) 206.
[14] Cheekatamarla P., Thomson W., J. Power Sources, 156 (2006) 520.
[15] Cheekatamarla P., Thomson W., Appl. Catal. A: Gen, 287 (2005) 176.
[16] Balandin A., Rozhdestvenskaya I., Russ. Chem. Bull. 8 (1959) 1573.
[17] Katrib A., Leflaive P.,Hilaire L., Maire G., Catal. Lett. 38 (1996) 1.
18
CHAPTER II
STUDY OF THE PERFORMANCE OF MO2C
FOR ISO-OCTANE STEAM REFORMING
Abstract
In the present work, the performance of commercial molybdenum carbide (Mo2C) for
isooctane steam reforming has been investigated in order to determine the effects of
major operating parameters (temperature, space velocity, and steam to carbon ratio) on
the catalytic activity. While the results obtained indicate an onset reforming temperature
of 850°C, high concentrations of H2 in the reforming environment were found to reduce
the onset temperature to 750 8C. The catalytic activity at 850°C was sufficient to produce
hydrogen yields greater than 90% and carbon conversions close to 100%, with a low
selectivity to CH4 and CO2. In addition, and consistent with thermodynamic predictions,
a steam to carbon ratio of 1 appeared to optimize the reforming rates. Finally, based on
experimental observations, a reaction mechanism was formulated and used to interpret
the results obtained during catalytic activity measurements. This mechanism involves
continuous oxidation and reduction of Mo metal, which can provide activity and stability
to the catalyst when occurring at similar rates.
19
Introduction
Polymer electrolyte membrane (PEM) fuel cells have emerged as a promising energy
alternative for satisfying the growing demands of the transportation industry. No other
energy conversion technology offers the combination of benefits that fuel cells do.
Specifically, high energy efficiency and reduced greenhouse gas emissions are some of
the features that make this technology a highly attractive power source for future
generation vehicles [1].
Hydrogen is considered the ideal fuel for PEM fuel cells because it produces simple
power-generating systems with high energy efficiency and a fast response time [2] and
[3]. To achieve this high performance PEM fuel cells require high purity hydrogen and,
consequently, efforts have been primarily focused on the development of a readily
available source of pure hydrogen. Two different pathways are currently under
development to address this issue. The first one consists of off-board production of pure
hydrogen, followed by a convenient storage method which allows this stored hydrogen to
be used by the fuel cell when needed. Alternatively, the second one entails on-demand
generation of hydrogen, which occurs via an on-board fuel processor capable of
transforming a variety of feedstocks into high-purity hydrogen.
At present, technologies employed for the storage of hydrogen produced off-board
continue to be cost-prohibitive and also present further drawbacks. For instance, storage
of practical amounts of compressed hydrogen gas requires very large high-pressure
vessels; liquid hydrogen storage suffers significant evaporative loss; metal hydride
20
systems have low hydrogen storage density; and carbon nanotubes do not yet have
practical applications because of high cost and low hydrogen storage density [4]. In
addition, no hydrogen fuel supply infrastructure currently exists and much development
is needed before a practical infrastructure can be created [4] and [5]. Conversely, high-
energy-density fossil fuels already have an existing low-cost infrastructure for fuel
storage and delivery. Thus, the on-board production of hydrogen using fossil fuels
represents an interesting alternative to fulfill the requirements of hydrogen for fuel cell
applications.
On-board hydrogen production is typically performed in a fuel processor, which consists
of a reformer unit followed by several reactors designed to minimize sulfur and CO
contents of the product stream. Methanol, gasoline, and diesel are some of the liquid fuels
considered to be attractive options as potential hydrogen sources. However, gasoline
appears to make the most economic and practical sense. Well-known advantages of
gasoline are its plentiful supply, its relatively low price, the ease and safety of handling it,
and its high energy density. Moreover, a highly efficient production and distribution
infrastructure is already in operation globally.
Steam reforming, autothermal reforming, and partial oxidation are the primary methods
used in hydrocarbon processing for hydrogen production in fuel cell applications. Of
these reforming options, endothermic steam reforming is the most common technique, as
it produces higher hydrogen concentrations in the crude reformate gas compared to
partial oxidation and autothermal reforming [6], [7] and [8]. Steam reforming of liquid
hydrocarbons is largely carried out over Ni-based catalysts, and results describing their
21
performance have been reported elsewhere [9] and[10]. It has been found that these
catalysts exhibit good activity and selectivity at high temperatures, but that their stability
appears to diminish due to sintering. Addition of palladium to Ni-catalyst composition
has been determined to cause a significant improvement in the performance. Steam
reforming catalysts based on costly precious metals (platinum, ruthenium, rhodium,
rhenium) have been reported to be more effective for hydrocarbon steam reforming as
they preclude carbon deposition and inhibit methane formation [11]. Nonetheless, the
high expenses required even for small loadings of noble metal prevent these catalysts
from extensive usage. More recently, early transition metal carbides have attracted much
attention for displaying catalytic properties similar to those of noble metals [12]. Previous
studies indicate that bulk molybdenum carbide (Mo2C) catalysts show high coke
resistance at low steam to carbon ratios as well as good sulfur tolerance for both methane
and isooctane steam reforming [14] and [15]. These properties together with high thermal
stability and low cost make Mo2C an interesting alternative catalyst for gasoline steam
reforming.
The present investigation is intended to study the performance of Mo2C as a catalyst for
gasoline steam reforming under different operating conditions. Since gasoline is a
complex mixture of hydrocarbons, the analysis was simplified by using isooctane as
gasoline surrogate. The experimental results reported in this work will be analyzed in
order to determine the effect of the operating parameters on the catalytic activity. In turn,
the resulting analysis will allow us to formulate a reaction mechanism, which facilitates
22
better understanding of the catalytic behavior of Mo2C when used for isooctane steam
reforming.
Thermodynamic Analysis
Hydrogen production via steam reforming of isooctane (C8H18) is an endothermic
process, which can be expressed as follows:
C8H18 + 8H2O → 8CO + 17H2 ∆H° = 1274.9 kJ mol−1 (1)
It is generally accepted that steam reforming is the net result of a series of elementary
reactions, for some of which the catalyst surface plays a significant role. The equilibrium
composition of the reformate can be estimated by minimizing the Gibbs free energy. This
method does not demand knowledge of the exact chemical kinetics leading to the
equilibrium; instead, it only requires the identity of the chemical compounds existing in
the equilibrium state.
The equilibrium compositions of Fig. 1 show the effect of the temperature on the
reformer performance. High temperatures increase the efficiency of the reformer, as
indicated by the rise in the hydrogen yield, which in turn is due to the decline in the
concentrations of methane and unreacted water. As the temperature increases, the
equilibrium compositions approximate the values dictated by the stoichiometry of
reaction (1), i.e., 32% for CO and 68% for H2, which indicates that reaction (1) is
approaching completion.
23
Fig. 2 displays the effect of the steam to carbon molar ratio (S/C) on the hydrogen yield,
which suggests that the production of hydrogen can be optimized by using S/C ratios near
1. This result is in agreement with the stoichiometry of reaction (1) and seems to indicate
that excess steam is not favorable for the process since it will increase the amount of
unreacted water.
Experimental
Experiments were performed in a 12 mm fixed-bed tubular (quartz) reactor. A schematic
of the system used is shown in Fig. 3. The liquid feed, consisting of water and isooctane,
was vaporized at 200 °C and mixed along with the carrier gas in a vaporizer containing a
silicon carbide bed to enhance the mixing. Calibrated syringe pumps and mass flow
controllers were employed to control the flow rates. The operating temperature range of
the reactor was 650 °C −1000 °C, and the high-temperature exit stream was cooled down
to 5 °C to separate water, non-reacted isooctane, and other possible condensable
compounds from the off-gas. The dry gas product was analyzed using an SRI
chromatograph to monitor H2, CO, CO2, and CH4 concentrations.
The catalysts employed were Mo2C (Stock # 12192, Lot # L09Q054) and MoO2 (Stock #
48117, Lot # B11P16), which were purchased from Alfa Aesar. Typically, 0.5 g of
catalyst was used in each reforming experiment. The catalyst sample was supported by a
quartz wool plug placed inside the reactor as shown in Fig. 3. The spent samples were
analyzed by powder X-ray diffraction (XRD) on two different instruments: a Philips
diffractometer using Co Kα radiation with an iron filter, and a Siemens D-500 X-ray
24
powder diffractometer with Cu Kα radiation. XPS spectra were obtained with an AXIS-
165 manufactured by Kratos Analytical Inc., using an achromatic Mg Kα (1254 eV) X-
ray radiation with a power of 210 W. The binding energy was calibrated against the 4f7/2
line of clean Au to be at 84 eV. A pass energy (PE) of 80 eV was used to acquire all
survey scans. At this PE the energy resolution was about 1.2 eV. The high resolution
spectra of Mo 3d were acquired at PE of 40 eV with an energy resolution of about 0.8 eV.
The base pressure of the XPS analyzing chamber was 1.0 × 10−9 torr. As a preliminary
preparation for XPS analysis, the powdered samples were pressed into pure Indium
(99.99 pure) and analyzed in order to minimize the effects of charging. The curve fitting
of high-resolution spectra was performed using a least-squares fitting program permitting
linear or Shirley fitted backgrounds, a mixture of Gaussian and Lorentzian peak shapes,
constraints on spin-orbit splitting intervals and spin-orbit pair area ratios, as well as
constraints on peak position (binding energy), and peak width. Mo3d spin-orbit pair
intervals were set at 3.13 eV, and an area ratio of 0.666 was used. To prevent further
oxidation between the end of the experiment and the XPS or XRD analysis, the samples
were cooled down to room temperature inside the reactor using helium gas flow.
BET surface area measurements were performed using a Coulter SA-3100 automated
characterization machine.
The data was analyzed in terms of hydrogen yield, carbon conversion, water conversion
and selectivities, which were calculated as follows:
25
Results and Discussion
Characterization
Fig. 4 shows the X-ray diffraction pattern of commercial Mo2C obtained using Co Kα
radiation. As observed, the only phase detected in the bulk of the sample was β–Mo2C.
However, XPS analysis indicated that the catalyst surface not only consisted of the
carbide phase (Mo2+, 228.4 eV) but also the dioxide phase (Mo4+, 229.5 eV), the
oxycarbide phase (Mo5+, 231.2 eV), and the trioxide phase (Mo6+, 232.7 eV), as seen in
Fig. 5. The surface area of the catalyst was determined using BET analysis and found to
be 0.3 m2/g.
26
H2 pretreatment
The significant concentration of Mo atoms with oxidation degrees 5+ and 6+ on the
catalyst surface can be substantially reduced by means of a pretreatment with H2, which
is intended to increase the concentration of the carbide phase and minimize those of the
other phases. H2 pretreatments are beneficial due to their ability to reduce excess carbon
and eliminate the oxycarbide phase (Mo5+) [17]. To investigate the effect of the H2
pretreatment on the catalyst surface, 0.5 g of commercial Mo2C was subjected to a 20-
min H2 pretreatment, using a flow rate of 50 ml/min and a temperature of 850 °C. The
pretreated sample was analyzed using XPS analysis and the XPS Mo3d spectrum
obtained is displayed in Fig. 6. This analysis revealed that H2 pretreatment was highly
effective at reducing the concentration of the highest oxidation degrees existing on the
catalyst surface, namely oxycarbide Mo5+ and trioxide Mo6+. On the other hand, the
pretreatment was able to increase the concentration of carbide Mo2+ from 15% to 58%,
although the amount of the dioxide Mo4+ also experienced a raise from 14% to 39%. In
light of these results, we conclude that a short H2 pretreatment is able to eliminate Mo
with high oxidation degrees (5+, 6+), however, in doing so, the pretreatment appears to
cause an increase in the concentration of the dioxide phase Mo4+.
Catalytic Activity
The catalytic activity of commercial Mo2C for isooctane steam reforming was measured
at 850 °C, under different operating conditions. The variables investigated in the present
work were the weight hourly space velocity (WHSV), defined as the ratio of mass flow
27
rate to the mass of catalyst, and the steam to carbon molar ratio (S/C). The performance
achieved under different values of these parameters and their effect on the catalytic
activity, expressed in terms of hydrogen yield, is displayed in Fig. 7. As observed, high
performances were obtained at spaces velocities smaller than 1.8 h−1 and S/C ratios of
about 1, which agrees with the thermodynamic analysis performed in Section 2. The ratio
S/C = 1 corresponds to the stoichiometric value indicated by reaction (1) and denotes that
high concentrations of steam simply lead to unreacted water, rather than an improvement
of the catalytic activity.
Fig. 8 shows the effect of the space velocity on the conversion at S/C = 1. The decline in
the performance at higher space velocities can be related to the low surface area
measured for the catalyst, which limits the contact area between the reactants and the
catalyst surface. As observed, the conversions of isooctane and water appeared to be
slightly reduced from 92% and 81% at WHSV = 1.5 h−1, to 80% and 66%, at
WHSV = 1.8 h−1, respectively. This decrease may be attributed to the excess of reactants
fed to the reactor, which were not completely converted to products.
The selectivity of Mo2C to the reforming products at S/C = 1 is reported in Fig. 9. As
seen, the selectivity to H2 and CO was held to values around 65% and 25% within the
space–velocity range investigated in the present work. The average ratio H2/CO was 2.57,
which is higher than the value indicated by the stoichiometry of reaction (1), i.e., 2.13.
This suggests that CO was consumed by one or more side reactions taking place in the
reactor. The negligible evolution of CO2 indicates that reactions like Boudart's or water–
gas shift did not likely to happen to a significant extent, along with the reforming process.
28
CO methanation would lead to H2/CO ratios smaller than 2.13 and, therefore, this
reaction did not likely occur either. The small concentration of other species produced
during the reforming process (<10%) suggests that CO was probably consumed by
unknown reactions to form compounds that were not considered in our analysis.
Effect of the Temperature on the Catalytic Activity
The effect of operating temperature on the catalytic activity of commercial Mo2C was
also investigated. Accordingly, an experiment consisting of two parts was carried out. In
the first part, a fresh sample of commercial Mo2C was pretreated with hydrogen and then
employed to perform isooctane steam reforming at 1000 °C, with WHSV = 0.5 h−1,
S/C = 1.3, and He as carrier gas. Once the catalytic activity was measured at 1000 °C, the
temperature was decreased in 50 °C-steps and the performance was measured again at
steady state conditions, which typically took 30 min. This procedure was repeated until
the catalyst was no longer active, which happened at 650 °C. This inactive spent sample
at 650 °C (denoted as sample 1) was cooled down to room temperature in flowing He,
and then removed from the reactor for XPS analysis.
In the second part of the experiment, another fresh sample of commercial Mo2C was
subjected to the same H2 pretreatment and then used for isooctane steam reforming,
which was started at 700 °C. The procedure was basically the same as in the first part,
except that this time the temperature was systematically increased in 50 °C-steps up to
1000 °C. After the last measurement at 1000 °C, the sample was cooled down to room
temperature in flowing He, and then removed from the reactor in order to acquire its XPS
29
spectrum (denoted as sample 2). Fig. 10 shows the catalytic activity obtained in both
parts of the experiment.
As observed, at temperatures of 850 °C and higher, the response of the catalytic activity
in terms of hydrogen yield was practically the same, regardless of the direction from
which the temperature was reached. For instance, the production of hydrogen measured at
850 °C was 58% from both directions. However, at reforming temperatures between 700
and 850 °C, their catalytic activities displayed significant differences. When the
temperature was systematically decreased from 1000 to 650 °C, the catalyst showed high
activity even at temperatures as low as 700 °C, where the hydrogen yield was 70%. Yet,
when the temperature was increased from 700 to 1000 °C, the catalyst showed no activity
until it reached 850 °C. The XPS spectra of sample 1 at 650 °C and sample 2 at 1000 °C
are displayed in Fig. 11.
According to Fig. 11, the Mo3d spectrum of sample 1 (the inactive spent catalyst at
650 °C) exhibited a higher concentration of the carbide phase compared to that of sample
2 (the active spent catalyst at 1000 °C), which could indicate that the activity may not be
directly related to the concentration of carbide on the catalyst surface. The dioxide peak
displayed similar concentrations in the spectra of both samples. In addition, the
oxycarbide peak was found in both samples. However, the intensity of this peak was
higher in the spectrum of the inactive sample 1, which led us to consider a possible
relationship between the appearance of the oxycarbide phase and the catalytic activity. In
fact, when the catalyst exhibits poor activity, the concentration of unreacted isooctane in
the environment is higher, and this increases the formation rate of the oxycarbide phase.
30
The concentration of the trioxide phase was more elevated in the active sample 2 as
compared to that found in the spectrum of the inactive sample 1. This decline in the
trioxide concentration is due to its consumption in the formation the oxycarbide phase
through combination with unreacted isooctane [13].
Analysis of the Catalytic Activity
This section is intended to provide an explanation for the phenomena reported earlier in
this work. To start, a hypothesis intended to explain how the reforming process takes
place will be suggested and, thereupon, all the experimental findings that support the
hypothesis will be provided.
We hypothesize that isooctane steam reforming takes place via the following global
reactions:
Mo2C + 5H2O 2MoO2 + CO + 5H2 (2)
2MoO2 + 0.51C8H18 2Mo2C + 4.59H2 + 2.16CO + 0.92CO2 (3)
Thus, the reforming process becomes stable as long as these reactions are continuously
taking place. Reaction (2) has been considered by Darujati et al. [17] as the one causing
the oxidation of Mo2C catalysts during steam reforming processes. Reaction (2) can be
expressed as the result of the following 2-step mechanism:
• Step 1: Thermal decomposition of the carbide phase Mo2C
31
Mo2C 2Mo + C ∆H° = +53.1 kJ mol−1 (4)
• Step 2: Oxidation of the products of step 1 with steam
Mo + 2H2O(g) MoO2 + 2H2(g) ∆H° = −105.3 kJ mol−1 (5)
C + H2O CO + H2 ∆H° = +131.3 kJ mol−1 (6)
Reaction (4) appears to be thermodynamically favored by high temperatures, as observed
in Fig. 12. This is in agreement with a previous experimental study conducted by LaMont
and Thomson, who found that the rates of decomposition of Mo2C become high at
750 °C and can cause the total disappearance of the carbide phase at 850 °C [16].
Darujati et al. found that the onset temperature of reaction (2) was 600 °C [17]. In the
same studies, it was also noticed that reaction (2) appears to happen in two stages: at
temperatures below 750 °C, the oxidation rates are independent of the H2O concentration
and at temperatures of 750 °C and higher, the rate of reaction (2) increases as the
concentration of steam becomes greater. As indicated by Fig. 12, reaction (2) is a process
favored by high temperatures, which is in concordance with Darujati et al.'s experimental
study.
Based on this information, the effect of the reforming temperature on the activity of
Mo2C that is seen in Fig. 10 can be understood in terms of the thermal decomposition of
the carbide phase. At low reforming temperatures, the thermal decomposition of the
carbide phase, as described by reaction (4), becomes kinetically limited as well as
32
thermodynamically unfavorable and leads to low concentrations of MoO2 on the catalyst
surface. In turn, low concentrations of MoO2 reduce the rates of reaction (3), therefore
preventing the continuous oxidation and reduction of Mo metal according to reactions (2)
and (3). Consequently, the initiation of the reforming process would never occur and the
catalyst would become inactive for steam reforming of isooctane at low reforming
temperatures. This explanation agrees with the findings reported in Fig. 10, where the
pretreated commercial Mo2C showed no activity as the reforming temperature was raised
from 700 to 800 °C. In addition, by further increasing the reforming temperature to
850 °C, the pretreated commercial Mo2C became active, which led us to think that, at that
temperature, reactions (2) and (3) started to take over. The reforming rates at
temperatures above 850 °C were limited by the available amount of Mo metal produced
by the thermal decomposition. Thus, regardless of the direction at which the temperature
was changed, the pretreated commercial Mo2C displayed the same catalytic activity at
temperatures above 850 °C. According to Fig. 10, when the temperature decreased from
1000 to 650 °C, the pretreated commercial Mo2C was able to maintain its high catalytic
performance even at reforming temperatures as low as 700 °C. This dependency of the
catalytic performance on the temperature change direction within the range between 800
and 700 °C can be explained by considering the effect of the reforming environment. As
the reforming temperature becomes lower than 850 °C, the Mo2C thermal decomposition
rates become smaller. However, the hydrogen produced via the reforming process is able
to enhance the decomposition of the carbide phase to form Mo metal through reaction
(7), which activates the mechanism of reforming process described by reactions (2) and
(3), even though the thermal decomposition of the carbide phase occurs at low rates:
33
Mo2C + 2H2 2Mo + CH4 ∆H° = −21.6 kJ mol−1 (7)
This reaction has been previously reported by other investigators [16] and found to take
place between 650 and 800 °C. Based on this information, we believe that the high
reforming performances obtained in Fig. 10, when reducing the temperature from 850 to
700 °C, are the direct result of reaction (7).
As indicated by Fig. 10, when the temperature was raised from 700 to 800 °C, no activity
was detected using only He as the carrier gas. However, increasing the H2 concentration
in the feed stream of isooctane and water should improve the reforming performance by
promoting the formation of Mo metal via reaction (7). To confirm this, isooctane steam
reforming over commercial Mo2C was performed using H2 along with He as the carrier
gas at temperatures ranging from 700 to 800 °C. For each test, the WHSV and S/C ratio
were set to 1 h−1 and 1, respectively. At 800 °C, a flow of 10 sccm of H2 was used as the
carrier gas during the first 10 min of reforming, after which the carrier gas was switched
to He at the same flow rate. The H2 fed at the start of the experiment was sufficient to
activate the catalyst and provoke a reforming performance which was stable for at least
1 h, with a hydrogen yield of 55%. At 750 °C, the same flow rate of H2 was used as a
carrier gas intended to enhance the reforming performance. However, unlike the
experiment at 800 °C, the flow of H2 had to be continuous in order to maintain a stable
catalytic activity for at least 1 h, with a hydrogen yield of 63%. At this temperature,
switching the carrier gas to He only caused a significant drop in the reforming
performance. Finally, at 700 °C, all the efforts to enhance the catalytic activity with H2
were unsuccessful. Neither higher H2 flow rates nor longer exposure times were able to
34
activate the catalyst and increase the reforming rates. The results just described above
may be interpreted by taking into consideration a kinetic approach. As the temperature
decreases from 800 to 700 °C, reaction (7) becomes thermodynamically favored, as
indicated by Fig. 12. However, reduction of the temperature results in a decline in the
reaction rates, which appears to have a more significant impact on the reforming
performance than that which is expected from thermodynamics.
The carburization of the oxide phase with isooctane, as described by reaction (3), has
been investigated in the present work. To accomplish this, 0.5 g of commercial MoO2
was placed into the reactor and combined with 1 ml/h of previously vaporized isooctane.
The carburization was performed at 700 °C. Displayed in Fig. 13 is the evolution of the
gas products as a function of time. The high reaction rates observed during the first
60 min of reaction indicate that this process was kinetically favored. Secondly, the
declines in the concentrations of H2 and CO indicate that the process had reached
completion. From that point on, the solid formed through this reaction was catalytically
converting isooctane into other compounds, which could not been identified by GC. To
gain greater insight into the actual mechanism of the carburization process, further
investigation is required. The stoichiometric coefficients appearing in reaction (3) are
only approximations based on the concentrations of the products in the off-gas.
Conclusions
The onset temperature to perform isooctane steam reforming, using commercial Mo2C as
the catalyst and He as the carrier gas, is 850 °C.
35
The catalytic performance obtained at 850 °C indicates that isooctane steam reforming is
highly efficient at S/C ratios close to 1, in agreement with thermodynamic predictions.
The space velocity (WHSV) appears to have an insignificant effect on the activity even at
values as high as 1.8 h−1.
The catalytic activity of Mo2C is initiated by the thermal decomposition of the carbide
phase and propagated by the carburization of the dioxide phase with isooctane. The onset
temperature of the reforming process is directly related to the temperature at which Mo2C
is able to produce Mo metal. Under reducing environments with H2 as the carrier gas, the
production of Mo metal is enhanced, causing a reduction in the onset temperature to
750 °C.
36
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
600 650 700 750 800 850
Temperature (°C)
Co
mp
os
itio
n,
H 2 y
ield
H2 yield
H2
CO
CH4CO2
Figure 1. ISR equilibrium compositions as a function of temperature
37
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00
S/C ratio
H2 Y
ield
600°C
650°C
700°C
750°C
800°C
850°C
Figure 2. ISR hydrogen yield as a function of temperature and S/C ratio
38
Figure 3. Steam reforming reactor schematic
39
0
30 35 40 45 50 55 60 65
2 theta
Inte
ns
ity [
arb
itra
ry s
ca
le]
46.7°
62.0°
40.8°45.1°
Figure 4. XRD pattern of commercial Mo2C
40
3000
5000
7000
9000
11000
13000
15000
17000
19000
226228230232234236238240
Binding Energy (eV)
Inte
ns
ity
Mo2+
Mo4+
Mo6+
Mo5+
Figure 5. Mo 3d XPS spectrum of commercial Mo2
Mo2+
15%
Mo4+
14%
Mo5+
6%
Mo6+
65%
41
3000
13000
23000
33000
43000
53000
63000
73000
83000
226228230232234236
Binding Energy (eV)In
ten
sity
Mo2+
Mo4+Mo5+
Figure 6. Mo 3d XPS spectrum of pretreated commercial Mo2C
Mo2+
58%
Mo4+
39%
Mo5+
2%
Mo6+
1%
42
0.5 0.8 1.1 1.5 1.8
S/C = 0.8
S/C = 1.0
S/C = 1.3
WHSV (h-1)
75%-100%
50%-75%
25%-50%
Figure 7. ISR H2 yield as function of the WHSV and the S/C ratio
(Mo2C, He = 10 sccm, T = 850°C )
43
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
0.5 0.8 1.1 1.5 1.8
WHSV [h-1]
C conversion
H2O conversion
Figure 8. Conversion as a function of WHSV at S/C=1
(pretreated commercial Mo2C, He=10 sccm, S/C=1, T=850°C )
44
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
0.0 0.5 1.0 1.5 2.0
WHSV [h-1]
Se
lec
tivity
H2 CO
CH4 CO2
Others
Figure 9. Selectivity of Mo2C at S/C=1
(pretreated commercial Mo2C, He=10 sccm, S/C=1, T=850°C )
45
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
600 650 700 750 800 850 900 950 1000
Temperature [°C]
H2 y
ield
sample 1
sample 2
reducing T
increasing T
Figure 10. Effect of temperature on Mo2C catalytic activity
46
0
1000
2000
3000
4000
5000
6000
7000
8000
9000
225227229231233235237239
Binding Energy (eV)
Inte
ns
ity
Mo2+
Mo4+
Mo5+Mo6+
0
1000
2000
3000
4000
5000
6000
7000
8000
225227229231233235237239
Binding Energy (eV)
Inte
ns
ity
Mo2+
Mo4+Mo5+
Mo6+
Figure 11. XPS Mo3d spectra of (a) sample 1 (inactive spent catalyst at 650°C)
and (b) sample 2 (active spent catalyst at 1000°C)
Mo2+
39%
Mo4+
25%
Mo5+
21%
Mo6+
15%
Mo2+
25%
Mo4+
27%
Mo5+
12%
Mo6+
36%
47
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
700 750 800 850
Temperature [°C]
Eq
uili
bri
um
co
ns
tan
t
reaction (2)
reaction (4)
reaction (7)
Figure 12. Equilibrium constants
48
0%
5%
10%
15%
20%
25%
30%
35%
40%
0 20 40 60 80 100 120
Time [min]
Off
-ga
s c
om
po
siti
on
H2 CO
CH4 CO2
Others
Figure 13. Carburization of MoO2 with isooctane at 700°C
49
0
50
100
150
200
250
300
20 25 30 35 40 45 50 55 602 theta
Inte
ns
ity [
arb
itra
ry s
ca
le]
37.7°
34.7°
39.5°
52.5°
Figure 14. XRD pattern of MoO2 carburized with isooctane.
50
3000
13000
23000
33000
43000
53000
63000
73000
83000
226228230232234236238
Binding Energy (eV)
Inte
ns
ity
Mo2+
Mo4+Mo5+
Figure 15. XPS Mo3d pattern of MoO2 carburized with isooctane
Mo2+
55%
Mo4+
37%
Mo5+
8%
Mo6+
0%
51
References
[1] M. Wang, J. Power Sources 112 (2002) 307.
[2] D.J. Moon, J.W. Ryu, S.D. Lee, B.G. Lee, B.S. Ahn, Appl. Catal. A 272 (2004) 53.
[3] D.P. Papadias, Sh.H.D. Lee, D.J. Chmielewski, Ind. Eng. Chem. Res. 45 (2006)
5841.
[4] L.F. Brown, Int. J. Hydrogen Energy 26 (2001) 381.
[5] Q. Ming, T. Healey, L. Allen, P. Irving, Catal. Today 77 (2002) 51.
[6] Docter, A. Lamm, J. Power Sources 84 (1999) 194.
[7] S. Specchia, A. Cutillo, G. Saracco, V. Specchia, Ind. Eng. Chem. Res. 45 (2006)
5298.
[8] Esroz, H. Olgun, S. Ozdogan, J Power Sources 154 (2006) 67.
[9] J. Zhang, Y. Wang, R. Ma, D. Wu, Appl. Catal. A 243 (2003) 251.
[10] A.A. Praharso, D.L. Adesma, N.W. Trimm, Cant. Chem. Eng. J. 99 (2004) 131.
[11] L. Wang, K. Murata, Y. Matsumura, M. Inaba, Energy Fuels 20 (2006) 1377.
[12] A.P.E. York, A.J. Brungs, S.C. Tsang, M.L.H. Green, Chem. Commun. 40 (1997)
39.
[13] C. Pham-Huu, M.J. Ledoux, J. Guille, J. Catal. 143 (1993) 249.
[14] P. Cheekatamarla, W.J. Thomson, J. Power Sources 156 (2006) 520.
[15] P. Cheekatamarla, W.J. Thomson, Appl. Catal. A 287 (2005) 176.
[16] D.C. LaMont, W.J. Thomson, Appl. Catal. A 274 (2004) 173.
[17] A.R.S. Darujati, D.C. Lamont, W.J. Thomson, Appl. Catal. A 253 (2003) 397.
[18] T. Xiao, A.P.E. York, K.S. Coleman, B. John, J. Claridge, J. Sloan, M.L.H.
Charnok, Green, J. Mater. Chem. 11 (2001) 3094.
52
CHAPTER III
ACTIVITY AND STABILITY OF MoO2 CATALYST
FOR THE PARTIAL OXIDATION OF GASOLINE
Abstract
The present investigation is focused on the performance of molybdenum dioxide (MoO2)
as a catalyst for the partial oxidation of isooctane. Metallic character and high oxygen
mobility exhibited by this oxide appear to enhance its catalytic activity, which can be
explained in terms of the Mars–van Krevelen mechanism. An oxygen-to-carbon ratio
(O/C) of 0.72 seems to stabilize the catalytic performance, which could reach H2 yields of
78% and carbon conversions of 100%, at 700 °C and 1 atm, after 20 h on stream. In
addition, the catalyst was tested for sulfur tolerance using thiophene as model sulfur
compound. Our findings indicate that the catalytic activity is barely affected even at
sulfur concentrations as large as 500 ppm, after 7 h on stream. Finally, the catalyst
performance was compared to that of a nickel catalyst using premium gasoline as fuel.
MoO2 displayed a stable performance whereas the Ni catalyst deactivated due to coke
formation.
53
Introduction
Nickel (Ni) and noble metals are traditionally used as catalysts for hydrocarbon
reforming processes, which suggest that the metallic characteristic may play an important
role in the mechanism involved in such processes. Molybdenum dioxide (MoO2) exhibits
an unusual metallic electrical conductivity, which is not a common characteristic of metal
oxides. This is attributed to its relatively high density of states observed in the valence
band energy region. The existence of these free electrons is considered to enhance the
catalytic activity of Mo4+ in MoO2, unlike Mo6+ in MoO3, where all the valence electrons
of the metal are bonded to neighboring oxygen atoms [1]. The catalytic activity of MoO2
for the isomerization of hydrocarbons has been interpreted in terms of a bifunctional
(metallic–acidic) mechanism. According to this mechanism, the metallic sites existing on
the dioxide are able to dissociate hydrogen and produce active hydrogen atoms, which
bond with surface oxygen atoms to form Bronsted Mo–OH acidic functional groups [2].
Previous investigations [3], [4] and [5] have reported that the formation of MoO2
significantly affects the stability of molybdenum carbide (Mo2C) for reforming processes.
Thus, under strong oxidizing environments, such as a large steam to carbon ratio, the
carbide phase transforms into dioxide according to the following expression:
Mo2C + H2O MoO2 + CO + H2 (1)
Studies suggest that this phase transformation deactivates the carbide system. However,
the experimental evidence obtained from these works was based only on the analysis of
54
the bulk catalytic properties without investigating their surface compositions. For this
reason, we strongly believe that the conclusions about the negative influence of the
dioxide phase on the reforming reactions must be carefully examined prior to wide
acceptance. Conversely, we have recently reported the presence of MoO2 on the surface
of active Mo2C catalysts during the steam reforming of isooctane [6], which led us to
conclude that the dioxide phase may play an important role in the steam reforming of
liquid hydrocarbons.
An additional reforming process, partial oxidation, has been extensively studied
employing transition metal oxides as catalysts [7], [8] and [9]. Partial oxidation is a more
cost effective and simpler reforming process, compared to steam reforming, as it requires
no additional water. Even with these incentives, few investigations have been devoted to
the partial oxidation of liquid hydrocarbons. Furthermore, no investigation has reported
the use of MoO2 for such systems. The lack of information regarding these topics has led
us to focus our interest in the study of the MoO2 performance for the partial oxidation of
liquid hydrocarbons.
The objective of the present investigation is to understand how the various operating
conditions affect the partial oxidation of gasoline over bulk MoO2 catalyst. To simplify
this study, isooctane (2,2,4-trimethylpentane) was used to model gasoline for our
catalytic activity studies. The section 1 of this paper characterizes the physical properties
of bulk MoO2. The body of the present work combines thermodynamic analysis and
experimental results to investigate the mechanistic process of MoO2 toward the isooctane
reforming reaction. Finally, the catalyst was tested to understand how the presence of
55
sulfur compounds and coking precursors would influence their reforming activity and
long-term stability.
Experimental
Experiments were performed in a 12 mm fixed-bed tubular (quartz) reactor. The liquid
feed, consisting of either isooctane or premium gasoline, was vaporized at 200 °C and
350 °C, respectively. The vapor obtained was mixed along with air, employed as oxygen
source for the partial oxidation, using a silicon carbide bed to enhance the mixing.
Calibrated syringe pumps and mass flow controllers were employed to control the flow
rates. The exit stream was cooled down to 5 °C to separate water, non-reacted fuel, and
other possible condensable compounds from the off-gas. The dry gas product was
analyzed using an SRI chromatograph to monitor H2, CO, CO2, and CH4 concentrations.
The GC columns used for this purpose were Molecular Sieve 13X and HyesepD. The
carrier gas was a mixture of 10% hydrogen and 90% helium.
The MoO2 catalyst was purchased from Alfa Aesar. The catalyst sample was supported
by a quartz wool plug placed inside the reactor. Spent samples were analyzed by powder
X-ray diffraction (XRD) on a Philips diffractometer that employs Co Kα radiation with
an iron filter. XPS spectra were obtained with an AXIS-165 manufactured by Kratos
Analytical Inc. using an achromatic MgKα (1254 eV) X-ray radiation with a power of
210 W. The binding energy was calibrated against the 4f7/2 line of clean Au to be at
84 eV. A pass energy (PE) of 80 eV was used to acquire all survey scans. At this PE the
56
energy resolution was about 1.2 eV. The high-resolution spectra of Mo3d were acquired
at PE of 40 eV with an energy resolution of about 0.8 eV. The base pressure of the XPS
analyzing chamber was 1.0 × 10−9 Torr. Before performing any XPS analysis, the
powdered samples were pressed into pure indium (99.99% pure) in order to minimize the
effects of charging. The curve fitting of high-resolution spectra was performed using a
least-squares fitting program. Mo3d spin–orbit pair intervals were set at 3.13 eV, and an
area ratio of 0.666 was used. To prevent further oxidation between the end of the
experiment and the XPS or XRD analysis, the samples were cooled down to room
temperature inside the reactor under helium. BET surface area measurements were
performed using a Coulter SA-3100 automated characterization machine. The data was
analyzed in terms of hydrogen yield and carbon conversion, which were calculated as
follows:
For isooctane as fuel:
For premium gasoline (C7H13) as fuel:
57
Results and Discussion
Characterization
XRD analysis was employed to examine the bulk structure of the dioxide. MoO2 displays
a distorted rutile structure with monoclinic symmetry, whose unit cell parameters were
found to be: a = 5.59 Å, b = 4.84 Å, c = 5.62 Å, and β = 121.23°. These values agree with
those found in previous studies [10].
The surface analysis involved two different techniques: BET and XPS measurements.
BET surface area measurements yielded a surface area of barely 4.5 m2/g. XPS analysis
was used to determine the composition of the oxide surface. The results indicated that the
most abundant elements on the oxide surface were Mo (23%), C (29%), and O (48%).
Further analysis was performed to obtain the spectrum of each element. These results can
be seen in Fig. 1.
The catalyst surface appears to be a mixture of three different molybdenum oxide phases.
The Mo3d spectrum (Fig. 1(A)) indicates that Mo hexavalent (232.73 eV, 53%) shows
the highest concentration, followed by Mo pentavalent (231.16 eV, 29%), and finally Mo
tetravalent (229.46 eV, 18%). The de-convolution of the O1s spectrum (Fig. 1(B))
produced three peaks, which can be assigned to ionic oxygen divalent (530.28 eV, 58%),
O–H bonding (532.63 eV, 20%), and O–C bonding (531.50 eV, 22%). The C1s spectrum
(Fig. 1(C)) indicates the presence of C–O bonds on the catalyst surface (288.94 eV,
17%), as well as the existence of C–H bonds (285.89 eV, 58%) and elemental carbon
58
(284.37 eV, 25%). The information extracted from the XPS analysis allowed the
estimation of the global composition of the catalyst surface, which turned out to be
MoO1.2. This suggests a deficit of oxygen ions on the catalyst surface and the existence of
oxygen vacancies.
Non-catalytic Oxidation of isooctane Over Molybdenum Dioxide
The “non-catalytic oxidation of isooctane” occurs when the fuel is oxidized by reacting
with the oxygen existing in the MoO2 lattice structure, in absence of gaseous oxygen.
Under such operating conditions, MoO2 becomes a reactant rather than a catalytic
material. This non-catalytic oxidation of isooctane over MoO2 can be formulated as the
result of two different processes:
Full oxidation: (25/2)MoO2 + C8H18 (25/2)Mo + 8CO2 + 9H2O (2)
Partial oxidation: 4MoO2 + C8H18 4Mo + 8CO + 9H2 (3)
The equilibrium constants for these reactions are reported in Fig. 2. As observed, both
oxidation processes are strongly favored by high temperatures. However, the partial
oxidation appears to be more favorable than the full oxidation over the temperature range
considered in this analysis. The large values obtained for the equilibrium constants
suggest that partial and full oxidation are dominated by kinetics rather than
thermodynamic limitations.
59
Under reforming conditions, the Mo metal phase is fairly unstable and rapidly becomes
carbide. This carburization process can be described by the following reaction:
2Mo + CH4 Mo2C + 2H2 (4)
Nonetheless, the carbide phase can be also formed as a result of the reactions shown
below:
2MoO2 + 6CO Mo2C + 5CO2 (5)
2Mo + C Mo2C (6)
Thermodynamics allowed the estimation of the equilibrium constants for these reactions,
which are reported in Fig. 3. As seen in the figure, at temperatures above 500 °C, reaction
(4) becomes the dominant process and, consequently, the carbide phase is formed via the
carburization of the metal phase by the methane resulting from the thermal
decomposition of the fuel. On the other hand, at temperatures below 500 °C, the
formation of the carbide phase appears to be the consequence of the reduction of the Mo
dioxide phase by carbon monoxide via reaction (5), which becomes the dominant
process.
The non-catalytic oxidation of isooctane over molybdenum dioxide was experimentally
studied by oxidizing different flow rates of isooctane (0.2 ml/h and 1.0 ml/h before
vaporization) with similar amounts of Mo dioxide (0.5 g), under the same pressure and
60
temperature (1 atm and 700 °C). Helium was used as a carrier gas and fed into the system
at a flow rate of 10 sccm. The rates of formation for each species are displayed in Fig. 4.
Based on a simple oxygen balance, we were able to demonstrate that the total amount of
oxygen contained in the carbon oxide products proceeds from the lattice oxygen existing
in both the surface and the bulk of the Mo dioxide structure. The data shown in Fig. 4
was combined with the stoichiometry of reaction (2) to estimate the amount of oxygen
produced in the off-gas during both experiments. The calculated amounts of oxygen
existing in the off-gas were approximately 0.11 g at 0.2 ml/h, and 0.10 g at 1.0 ml/h of
isooctane. These values agree with the 0.13 g of oxygen existing in the 0.5 g of Mo
dioxide used in each test. The differences can be attributed to the error introduced when
using GC to measure reaction rates.
Fig. 4 indicates that the oxidation of the hydrocarbon appears to initially proceed at a
faster rate when compared to the methane formation caused by the thermal
decomposition of the fuel. However, as soon as the fuel oxidation process ceases, due the
lack of available lattice oxygen in the dioxide structure, the concentrations of both carbon
monoxide and carbon dioxide drop to almost zero. At the same time, the concentration of
methane increases, as its production rate is constant and it is no longer consumed by the
Mo metal via reaction (4).
The XRD patterns of the spent samples used in the non-catalytic oxidation tests are
shown in Fig. 5. As observed, an isooctane flow rate of 0.2 ml/h appears to enhance the
formation of crystalline molybdenum carbide. However, the carbide peak found at
61
2θ = 46.60° appears to be much smaller than that displayed by the XRD pattern of
commercial molybdenum carbide. This suggests that when using an isooctane flow rate
of 0.2 ml/h, the produced carbide phase exhibits a cluster size smaller than that of
commercial carbide. The absence of dioxide peaks is due to the total consumption of the
lattice oxygen via reactions (2) and (3).
In the XRD pattern of the sample obtained using an isooctane flow rate of 1.0 ml/h no
peaks were detected and, instead, a slight shoulder can be observed within the region
assigned to the carbide peaks (2θ = 40–50°). This finding suggests the formation of an
amorphous carbide phase as a result of the large concentration of elemental carbon
formed during the fuel decomposition.
The oxygen atoms remain divalent before and after the oxidation process takes place. In
other words, the oxidation degree of the oxygen atoms is the same in both the oxide
structure (MoO2) and the carbon oxides formed during the reaction. Consequently, the
oxygen atoms do not appear to gain electrons and are only transferred as the fuels are
oxidized. This result led us to believe that the actual oxidizing agent is the Mo4+ cation,
which ends up with an oxidation degree equal to 0 in the metallic phase.
The ability of metal oxides to transfer their lattice oxygen atoms to reacting species
increases as the strength of metal-oxygen covalent bond decreases [11]. This explains the
high oxidative capability displayed by MoO2, given that this oxide exhibits a strong ionic
characteristic [12].
62
It is important to notice that the non-catalytic oxidation of isooctane by MoO2 takes place
at significant rates despite the low concentration of Mo4+ on the surface (18%), as
indicated in Fig. 1. Thus, when the reaction occurs at 700 oC, the reduction of Mo5+ and
Mo6+ by the hydrogen molecules produced from reaction (3) is kinetically favored and
yields Mo4+. This in turn accelerates the formation of hydrogen. As a result, the process
becomes self-sustaining and the fuel oxidation process continues until the Mo dioxide
phase is completely reduced and no longer detectable using XRD.
It has been reported that the presence of Mo trioxide (MoO3) layers on the MoO2 surface
may prevent the dioxide from exhibiting catalytic activity. A pretreatment with hydrogen
has been found to be effective in minimizing the concentration of trioxide on the dioxide
surface. The next section of this paper is devoted to the study of this procedure.
H2 Pretreatment
The reduction of MoO3 to MoO2 using H2 as reducing agent is formulated as follows:
MoO3 + H2 MoO2 + H2O (7)
However, an excessive pretreatment leads to the formation of Mo metal which we have
found to be inactive for the partial oxidation of isooctane:
MoO2 + 2H2 Mo + 2H2O (8)
63
The experimental values of the equilibrium constants for reactions (7) and (8) have been
previously reported [13]. Their values are shown in Fig. 6.
As shown, the reduction of the trioxide phase to dioxide phase becomes the dominant
process over the reduction of the dioxide phase to metal phase at temperatures above
430 °C.
It has been indicated that fresh commercial MoO2 contains a small amount of an
unidentified Mo oxide on its bulk structure (see Fig. 7). The hydrogen pretreatment is
thus intended to eliminate this unknown phase whose presence we have found to
negatively affect the catalytic activity of Mo dioxide for the partial oxidation of
isooctane. A set of experiments was designed to measure the role played by the
temperature on the effectiveness of the pretreatment. To do so, two fresh samples of
commercial Mo dioxide were subjected to H2 pretreatments performed under similar
operating conditions except for the temperature, which was set to 350 °C and 450 °C. The
mass of catalyst employed in each test was 0.2 g and the duration of the pretreatment was
30 min. The pressure and the flow rate of hydrogen were measured at 1 atm and 10 sccm,
respectively.
The XRD patterns displayed in Fig. 7 indicate that when the hydrogen pretreatment is
performed at 350 °C the peak assigned to the unknown oxide (2θ = 33.23°) shifts to the
position assigned to the trioxide phase (2θ = 32.35°). This means that the temperature
was not high enough to attain the complete reduction of the unknown oxide. However, a
pretreatment performed at 450 °C was able to directly reduce the unknown oxide to
64
molybdenum dioxide, as deduced from the total disappearance of the unknown peak at
2θ = 33.23°. A side effect of the pretreatment performed at 450 °C is the slight change in
the crystal structure indicated by the shift of the peak at 2θ = 30.78° to 2θ = 30.74°,
which leads to a change in the lattice parameters.
Table 1 shows the results of the XPS analysis performed over the samples pretreated at
350 °C and 450 °C. The Mo3d spectrum of a fresh sample was added for comparison
purposes. As observed, the pretreatment at 350 °C was able to reduce the trioxide phase
to pentoxide by 10%, although it was not high enough to increase the concentration of
Mo4+. On the other hand, the pretreatment at 450 °C appears to be able to not only reduce
the concentration of trioxide by 24%, but also raise the concentration of dioxide from
18% to 29%. No metal formation was detected on the surface even at temperatures as
high as 450 °C.
Based on these results, a temperature of 450 °C appears to be high enough to remove the
unknown oxide phase from the bulk structure of the Mo dioxide and increase the
concentration of Mo divalent on the surface by 11%.
Partial oxidation of isooctane over MoO2
When gaseous oxygen is supplied to the system, the selective oxidation of hydrocarbons
using transition metal oxides becomes a catalytic process and is called partial oxidation.
Gaseous oxygen, O2, is adsorbed as O−, O2−, or incorporated into the oxide structure as
lattice oxygen via a reduction process. The electrons gained by the adsorbed oxygen are
65
those previously released during the oxidation of the hydrocarbon molecules. This
reaction mechanism formed by consecutive cycles of oxidation and reduction steps was
first studied by Mars and van Krevelen in 1954 [14] and is represented in Fig. 8.
The mechanism starts with the decomposition of the hydrocarbon on the oxide surface by
the dual sites existing on the catalyst surface. In this step, Mo tetravalent ions, Mo4+, gain
electrons from the hydrocarbon molecules. Next, lattice oxygen is transferred from the
surface into the smaller hydrocarbon molecules formed in the previous step, producing
carbon oxides, hydrogen and water. The vacancies created during this transfer are
replenished by oxygen ions coming from the metal oxide bulk structure. Finally, the
gaseous oxygen fed into the system is to replenish the catalyst bulk structure and
consume the electrons released during the first step of the process. The global reaction to
express this catalytic process is formulated as follows:
C8H18 + 4O2 8CO + 9H2 (9)
Thermodynamic analysis was performed to study the limitations of this process under
different temperatures and fuel concentrations. The results obtained were expressed in
terms of H2 yield and are shown in Fig. 9.
The calculations were modeled as a feed stream containing isooctane and air. Isooctane
partial oxidation appears to be an effective process to achieve high reforming
performances (H2 yields of 80% and higher) even at temperatures as low as 675 °C and
isooctane molar concentrations above 4%. However, this thermodynamic analysis was
66
not intended to account for stability issues, which may considerably affect the catalyst
performance. For instance, under high concentrations of air (calculated as mole of air per
gram of catalyst), molybdenum dioxide becomes unstable and susceptible to be oxidized
and eventually form Mo trioxide.
Effect of the O2/C ratio
The effect of the air concentration over the catalyst stability was investigated by
performing isooctane partial oxidation at different oxygen-to-carbon (O/C) ratios under
the same conditions of pressure and temperature (700 °C, 1 atm). Each fresh catalyst
sample was first pretreated with H2 at 450 °C for 30 min. The results of the performance
for the partial oxidation of isooctane are shown in Fig. 10.
As exhibited above, O/C ratios below 0.72 appear to be insufficient to stabilize the
dioxide phase and prevent the formation of metal and the carburization of the dioxide
phase. Thus, at O/C = 0.26 and O/C = 0.41, the performance significantly decreases after
approximately 50–60 min on stream. On the other hand, an O/C ratio of 0.72 was able to
supply enough gaseous oxygen to stabilize the bulk structure of the Mo dioxide phase for
upwards of 9 h. These results are confirmed by the XRD patterns of the spent samples,
which are displayed in Fig. 11.
The diffraction patterns of the samples obtained at O/C = 0.26 and O/C = 0.41 exhibit
only peaks assigned to the carbide phase. This suggests that the concentrations of oxygen
in the feed stream were not large enough to prevent the total consumption of the lattice
67
oxygen and, hence, the formation of the carbide phase. The shape of the carbide peaks
detected for these two samples indicates that the particle size of the resulting carbide
phase becomes smaller as the concentration of oxygen in the gas phase is reduced. The
diffraction pattern of the sample at S/C = 0.72 exhibits peaks for both the dioxide and the
carbide phases. The peak of the carbide phase (2θ = 46.60°) appears to be sharper when
compared to those obtained at smaller O/C ratios, indicating the formation of larger
cluster sizes. The peak detected for the dioxide phase at 2θ = 30.80° suggests that the
concentration of oxygen in the gas phase was large enough to delay the total reduction of
the Mo dioxide phase to carbide, for at least 9 h.
Table 2 shows the XPS Mo3d spectrum of the spent sample obtained at O/C = 0.72. The
concentrations of the Mo species on the surface of the spent sample display significant
differences compared to those of the pretreated sample before the reaction. Thus, the
concentration of Mo metal was 0%, after H2 pretreatment at 450 °C; however, after 9 h
on stream, the Mo metal concentration increased to 13%. The formation of Mo metal is
believed to take place via reactions (2) and (3). On the other hand, the concentration of
Mo4+ experienced an increase of 10%, which can be explained in terms of the reduction
process of the trioxide species (Mo6+) from 29% to 16%. The reduction of Mo6+ to Mo4+
may be the result of its reaction with either H2 or CO, which are produced during the
reforming process.
68
Long-Term Stability
After a continuous period of 20 h on stream, the catalytic activity appears to be barely
affected as inferred from the slight decline in the H2 yield (from 81% to 78%) and the
stable conversion (100%). However, the selectivity seems to be affected by the time on
stream. Thus, after a period of 20 h, the concentrations of H2 and CO were reduced by
2%, whereas the concentration of CO2 was increased by 3%. These results suggest a
change in the selectivity as a result of the large concentration of oxygen in the feed
stream, which enhances the full oxidation of the fuel (that produces CO2 and H2O) at the
expense of the partial oxidation (that produces CO and H2). The concentration of
methane, CH4, remains constant because it is formed through the thermal decomposition
of the fuel and not as a consequence of the oxidation processes.
Preliminary experiments carried out to measure the thermal stability of MoO2 led us to
conclude that sintering is not an issue for MoO2 catalysts when exposed to reforming
environments. In light of this finding, the drop in the performance cannot be attributed to
sintering effects.
The XRD pattern of the spent sample after 20 h on stream is shown in Fig. 12. The
diffraction pattern obtained after 9 h on stream is also displayed for comparison purposes.
As seen, the peaks assigned to the oxide and the carbide phases were detected in both
spectra showing similar widths, although with some differences in the intensity. The
peaks at 2θ = 30.80° and 43.80° corresponding to the dioxide phase exhibit lower
intensity after 20 h on stream, compared to those detected at the same angles but after 9 h
69
on stream. The peaks at 2θ = 40.60°, 44.80° and 46.60°, attributed to the carbide phase,
display higher intensities after 20 h on stream, compared to those detected after 9 h on
stream.
In light of these findings, the bulk structure of the molybdenum dioxide catalyst appears
to experience slight changes with increasing time on stream, which may be related to the
changes observed in the performance. Thus, as time passes, the amount of Mo dioxide in
the sample decreases and, consequently, the catalytic activity toward the partial oxidation
also decreases. At the same time, the concentration of gaseous oxygen in the feed stream
is large enough to enhance the complete oxidation of the fuel to CO2 and H2O. This
change in the reaction selectivity can be explained in terms of a shift in the reaction
mechanism from the one proposed by Mars–van Krevelen to the one formulated by Eley–
Rideal, which is based on the direct reaction between the adsorbed oxygen and the
hydrocarbon existing in the gas phase [15].
Sulfur Tolerance
The Mo dioxide catalyst was also tested for sulfur tolerance using thiophene as a sulfur
model compound. Two different concentrations of thiophene (500 ppm and 1000 ppm)
were employed in this test, performed under the same conditions at which the catalyst
exhibited the longest stability (O/C = 0.72, WHSV = 0.6 h−1 and T = 700 °C). Fig. 13
shows the H2 yield and the conversion as a function of the time on stream at the
concentrations of thiophene mentioned above.
70
As observed, a concentration of 500 ppm thiophene appears to partially affect the H2
yield, which declined by 20%, while the conversion remained practically constant after
7 h on stream. Instead, using a concentration of 1000 ppm of thiophene, the H2 yield
decreased from 86% to 43%, as the conversion also diminished from 100% to 88% after
only 4 h on stream. The drop in the catalyst performance can be attributed to the
progressive reduction in the number of sites available for the adsorption of the fuel as a
result of the increased sulfur concentration on the catalyst surface. At 500 ppm, the
adsorption of sulfur compounds appears to be relatively slow and thus no significant
effect was detected on the catalytic performance of MoO2 after 7 h on stream. However,
at 1000 ppm, the deactivation caused by sulfur adsorption significantly affects the
catalyst performance. Fig. 14 displays the diffraction patterns for the spent samples
obtained from this experiment.
As seen, a concentration of 500 ppm thiophene appears to stabilize the dioxide phase,
although, not enough to prevent the formation of some carbide, as deduced from the
small carbide peak detected at 2θ = 46.65°. On the other hand, when the concentration of
thiophene was 1000 ppm, the amount of sulfur on the oxide surface appears to be large
enough to avoid the adsorption of methane and thus, no carbide resulting from the
carburization of Mo metal was observed.
The XPS analysis of the sample obtained at a concentration of 1000 ppm is shown in
Table 4. The concentration of sulfur on the catalyst surface appears to be negligible,
which seems disagree with the interpretation of the experimental results discussed above.
However, this apparent contradiction can be explained by considering a weak adsorption
71
of the sulfur compound onto the catalyst surface [16]. Thus, the sulfur compound could
be easily removed from the catalyst surface during the cooling process, where helium was
used to flush the reactor. This finding led us to believe that the sulfur poisoning of
thiophene over MoO2 may be a reversible process; however, further investigation is
required to support this hypothesis.
Coking Resistance
To measure the resilience of MoO2 to deactivation caused by coking, the catalyst was
tested using premium gasoline as fuel, with a content of aromatics of approximately 37%.
Aromatic compounds are considered to be coking-precursors and therefore, large rates of
coking-formation can be expected. The resulting performance was compared to that
obtained using a nickel catalyst that has been found to be active for gasoline reforming.
The operating conditions for this test were the same as ones used in the previous section
(O/C = 0.72, WHSV = 0.6 h−1 and T = 700 °C). The results of this comparison are
displayed in Fig. 15.
As observed, after 7 h on stream, the catalytic performance exhibited by MoO2 was fairly
stable, obtaining a H2 yield of 94% and a conversion of 100%. The nickel catalyst started
with a H2 yield of 100% and a conversion of 68%. However, these values diminished to
59% and 37% after 4 h on stream, respectively. The low conversion obtained with the Ni
catalyst is the direct result of coke formation, which eventually plugged the reactor and
significantly reduced the catalytic performance.
72
The XPS analysis of the spent MoO2 sample (Fig. 16) shows the surface concentration of
elemental carbon (74%) increased significantly compared to that of the fresh sample
(25%), although this was not enough to cause the deactivation of the catalyst after 7 h on
stream. The high conversion obtained with the dioxide catalyst is an indication of its high
coke resistance. Nevertheless, the amount of coke accumulated on the surface may
continuously increase as we operate the reactor beyond 7 h on stream, which may
eventually give rise to stability issues.
Conclusions
The catalytic activity of MoO2 for the partial oxidation of isooctane appears to be the
result of the dual metallic–acidic character of the catalyst surface combined with its
remarkable oxygen mobility. The mechanism of reaction proposed by Mars–van
Krevelen seems to be responsible for the stability of the catalyst.
At 700 °C and after 20 h on stream, MoO2 displayed a stable performance for the partial
oxidation of isooctane with a H2 yield of 78% and a conversion of 100%.
MoO2 exhibits a significant sulfur tolerance. The adsorption of sulfur compounds seems
to affect the selectivity by enhancing the total oxidation of the fuel at the expense of its
partial oxidation.
MoO2 is able to reform actual gasoline displaying a higher coking resistance when
compared to that exhibited by an active Ni catalyst.
73
226228230232234236238240
Binding Energy (eV)
Inte
ns
ity
Mo4+ (18%)
Mo5+ (29%)
Mo6+ (53%)
Mo3d
524526528530532534536538540
Binding Energy (eV)
Inte
ns
ity
O ions 2- (58%)
O-H (20%)
O=C (22%)
O1s
276278280282284286288290292294296
Binding Energy (eV)
Inte
ns
ity
C (25%)
C-H (58%)
C=O (17%)
C1s
Figure 1. XPS spectra of fresh MoO2
74
1E-40
1E-30
1E-20
1E-10
1E+00
1E+10
1E+20
1E+30
1E+40
400 450 500 550 600 650 700 750 800
Temperature [°C]
Eq
uili
bri
um
co
ns
tan
t
Partial Oxidation
Full Oxidation
Figure 2. Equilibrium constants for the anaerobic oxidation of isooctane
75
1.E+01
1.E+02
1.E+03
1.E+04
1.E+05
1.E+06
400 450 500 550 600 650 700 750 800
Temperature [°C]
Eq
uili
bri
um
co
ns
tan
t
2 MoO2 + 6 CO = Mo2C + 5 CO2
2 Mo + C = Mo2C
2 Mo + CH4 = Mo2C + 2 H2
Figure 3. Equilibrium constants for formation of carbide phase
76
0
1
2
3
4
0 20 40 60 80 100 120Time [min]
rate
of
form
ati
on
[s
cc
m]
H2
CO
CH4
CO2
Isooctane: 0.2 ml/h
0
2
4
6
8
10
0 20 40 60 80 100
Time [min]
rate
of
form
ati
on
[s
cc
m]
H2
CO
CH4
CO2
Isooctane: 1.0 ml/h
Figure 4. Concentration profiles for the non-catalytic oxidation of isooctane
77
Figure 5. Diffraction patterns of spent samples used in the non-catalytic oxidation tests.
The spectra of both fresh carbide and fresh dioxide were added as reference.
78
0.0
0.1
0.2
0.3
0.4
0.5
300 350 400 450 500
Temperature [°C]
Eq
uil
ibri
um
co
ns
tan
t
MoO3 + H2 = MoO2 + H2O
MoO2 + H2 = Mo + H2O
MoO3 + H2 = MoO2 + H2O
MoO2 + 2 H2 = Mo + 2 H2O
Figure 6. Equilibrium constants for the reduction of MoO2 and MoO3 with hydrogen.
79
0
10
20
30
40
50
27 28 29 30 31 32 33 34 35
2 theta
Inte
ns
ity
(a
rbit
rary
sc
ale
)
Fresh
350°C
450°C
MoO3
MoOx
Figure 7. Diffraction patterns of pretreated samples
80
Decomposition
of hydrocarbon
molecule and
oxidation by
Mo4+
C8H18 � CxHy + ne− CxHy + O2−(surf) � CO, CO2, H2O, H2
Transfer of
surface oxygen
to hydrocarbon
molecules
O2−(bulk) � O2−(surf)
Replenishing
of surface with
bulk oxygen
Replenishing
of bulk oxygen
with gaseous
oxygen
O2
O2−Mo4+
C8H18CxHy + O2−
O2(gas) + ne− � O2−(bulk)
Decomposition
of hydrocarbon
molecule and
oxidation by
Mo4+
C8H18 � CxHy + ne− CxHy + O2−(surf) � CO, CO2, H2O, H2
Transfer of
surface oxygen
to hydrocarbon
molecules
O2−(bulk) � O2−(surf)
Replenishing
of surface with
bulk oxygen
Replenishing
of bulk oxygen
with gaseous
oxygen
O2
O2−Mo4+
C8H18CxHy + O2−
O2(gas) + ne− � O2−(bulk)
Figure 8. Mars-van Krevelen mechanism
81
600°C
700°C
800°C
10%8% 6%
4%2%
0%
20%
40%
60%
80%
100%
H2 y
ield
TemperatureIsooctane
concentration
80%-100%
60%-80%
40%-60%
20%-40%
Figure 9. Isooctane partial oxidation: thermodynamic analysis
82
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
0 50 100 150 200 250 300 350 400 450 500 550
Time (min)
H2 y
ield
O/C = 0.26
O/C = 0.41
O/C = 0.72
Figure 10. Effect of O/C ratio on catalytic performance of MoO2
for the partial oxidation of isooctane at 700oC and 1 atm.
83
0
10
20
30
40
50
60
25 30 35 40 45 50 55
2 theta
Inte
ns
ity
(a
.s.)
MoO2 Mo2C
O/C=0.72
O/C=0.41
O/C=0.26
Figure 11. XRD patterns of spent samples at different O/C ratios
84
30
40
50
60
70
80
25 30 35 40 45 50 55
2 theta
Inte
ns
ity
(a
.s.)
MoO2 Mo2C
20 h
9 h
Figure 12. XRD patterns of spent samples obtained at O/C=0.72
85
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
0 50 100 150 200 250 300 350 400 450
Time on-stream [min]
H2 y
ield
, C
on
ve
rsio
n
500 ppm (H2 yield)
500 ppm (conversion)
1000 ppm (H2 yield)
1000 ppm (conversion)
Figure 13. Catalytic performance of MoO2 at different thiophene concentrations
86
0
25 30 35 40 45 50
2 theta
Inte
ns
ity
[a
rbit
rary
sc
ale
] MoO2 Mo2C
1000 ppm
500 ppm
Figure 14. XRD patterns of spent samples obtained at different thiophene concentrations
87
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
0 50 100 150 200 250 300 350 400
Time on stream [min]
H2 y
ield
, C
co
nv
ers
ion
H2 yield MoO2
C conversion MoO2
H2 yield Ni catalyst
C conversion Ni catalyst
Figure 15. Comparison between the performances of MoO2 and a nickel catalyst
for the partial oxidation of premium gasoline
88
280282284286288290292
Binding Energy (eV)
Inte
ns
ity
C (74%)
C-H (19%)
C=O (7%)
C1s
Figure 16. C1s XPS spectrum of spent MoO2 sample obtained
after the coking-resistance test
89
Pretreatment
None 350°C 450°C
Mo4+ 18% 18% 29%
Mo5+ 29% 39% 41%
Mo6+ 53% 43% 29%
Table 1: XPS analysis of pretreated samples
90
Pret. 450°C O/C=0.72
Mo 0% 13%
Mo4+ 29% 39%
Mo5+ 41% 37%
Mo6+ 29% 16%
Table 2: Mo3d spectrum of spent sample obtained at O/C=0.72. Spectrum of Mo dioxide
after pretreatment at 450°C was added as reference
91
Molar concentrations TOS
(h)
H2
yield
Conver-
sion H2 CO CH4 CO2
2 81% 99% 22% 14% 4% 6%
5 83% 100% 22% 14% 4% 8%
9 82% 100% 21% 12% 3% 8%
20 78% 100% 20% 12% 4% 9%
Table 3: Long-Term stability test
92
Composition (%At)
Mo 23%
C 42%
O 35%
S 0%
Mo3d spectrum
Mo 11%
Mo4+ 35%
Mo5+ 36%
Mo6+ 18%
Table 4: XPS analysis of the spent sample obtained using 1000 ppm of thiophene
93
References
[1] Katrib, P. Leflaive, L. Hilarie, G. Maire, Catal. Lett. 38 (1996) 95.
[2] H. Al-Kandari, F. Al-Khorafi, H. Belatel, A. Katrib, Catal. Commun. 5 (2004) 225.
[3] Darujati, D. LaMont, W. Thomson, Appl. Catal. A Gen. 253 (2003) 397.
[4] P. Cheekatamarla, W. Thomson, J. Power Sources 158 (2006) 477.
[5] Darujati, W. Thomson, Appl. Catal. A Gen. 296 (2005) 139.
[6] O. Marin Flores, S. Ha, Catal. Today 136 (2008) 235.
[7] H. Nair, J. Liszka, J. Gatt, C. Baertsch, J. Phys. Chem. C 112 (2008) 1612.
[8] E. Boikov, M. Vishnetskaya, A. Emel’yanov, Y. Rufov, N. Shcherbakov,
Khimicheskaya Fizika 26 (8) (2007) 38.
[9] D. Shekhawat, T. Gardner, D. Berry, M. Salazar, D. Haynes, J. Spivey, Appl. Catal.
A Gen. 311 (2006) 8.
[10] Magneli, G. Andersson, Acta Chem. Scand. 9 (1955) 1378.
[11] D. Martin, D. Duprez, J. Phys. Chem. 100 (1996) 949.
[12] F. Werfel, E. Minni, J. Phys. C Solid State Phys. 16 (1983) 6091.
[13] K. Vasilev, T. Nikolov, M. Chimbulev, Bulg. Godishnik na Visshiya
Khimikotekhnologicheski Institut, Sofiya, Volume Date 1967, 14 (1) (1971) 321–
330.
[14] M. Vannice, Catal. Today 123 (2007) 18–22.
[15] D. Liu, B. Liptak, P. Bouis, Environmental Engineers’ Handbook, Second edition,
Lewis Publishers, Boca Raton, FL, 1997.
[16] T. Jirsak, J. Rodriguez, J. Hrbek, Surf. Sci. 426 (1999) 319–335.
94
CHAPTER IV
X-RAY DIFFRACTION AND PHOTOELECTRON SPECTROSCOPY STUDIES OF
MoO2 AS CATALYST FOR THE PARTIAL OXIDATION OF ISOOCTANE
Abstract
X-ray diffraction (XRD), X-ray photoemission (XPS) as well as ultraviolet
photoemission (UPS) spectroscopy experiments on MoO2 powders were carried out to
examine the bulk, the core level energies, and the electronic structure of MoO2 samples
that were employed as catalysts for the partial oxidation of isooctane. Five fresh 0.5-g
MoO2 samples were exposed for 0, 0.5, 9, 20, and 43 h to identical reforming
environments and their spent samples were analyzed using the techniques mentioned
above. Our results indicate the rapid appearance of an intermediate Mo phase with a
binding energy of 228.5 eV and whose concentration progressively increases with time.
The oxidation state for this new phase was graphically estimated to approximately +2.6
and assigned to the compound Mo2O3, which forms on the catalyst surface as a result of
its exposure to the reforming environment. The electronic structure probed by UPS
reveals two bands, one at 1.62 eV and another at 0.55 eV below the Fermi level, that
decrease with the increasing time on stream. These results correlate very well with the
drop in the catalytic performance of MoO2 for the partial oxidation of isooctane and with
the decline in the concentration of dioxide (Mo4+) detected not only on the catalyst
surface, but also in the bulk structure, as confirmed by our XRD analysis.
95
Introduction
The production of hydrogen for fuel cell applications using hydrocarbon sources has
become challenging due to a number of issues that need to be addressed. The search for a
highly active catalytic material with elevated coking resistance and sulfur tolerance, even
at high temperatures, is still an ongoing quest. References about this topic can be
extensively found in the literature [1], [2] and [3]. For instance, Biniwale and co-workers
[4] carried out a study of the autothermal reforming of isooctane employing Ni–Mn, Ni–
W, and Rh–Ce as catalytic materials. Commercial catalytic materials for this purpose
involve the use of Ni-based catalysts which are inexpensive but at the same time well-
known for being prone to coking [5].
Molybdenum dioxide, MoO2, has been investigated as a potential catalyst for the
isomerization of light alkenes and alkanes by Al-Kandari et al. [6], and also for the
cracking of n-heptane by Song and co-workers [7]. Earlier studies of the reaction of
methane with molybdenum metal and oxides revealed the transformation of MoO3 and
MoO2 to Mo2C Koós et al. [8] investigated the carburization of MoO3 by XPS while Ma
et al. [9] studied methane aromatization over Mo-based catalysts by temperature-
programmed surface reaction (TPSR) and XPS. Wang et al. [10] characterized Mo/ZSM-
5 catalyst for the conversion of methane and benzene by XRD, XPS, ISS (ion-scattering
spectroscopy) and Infrared techniques. In all these studies Mo-oxide transformed into
Mo2C when reacting with methane. However, Marin-Flores and Ha have recently
reported that molybdenum dioxide (MoO2) is also highly active for the partial oxidation
of isooctane [11] which occurs as follows:
96
C8H18+4O2=8CO+9H2 (1)
This process is exothermic (∆H° = 659.9 kJ/mol) and has been found to proceed to full
conversion at 700 °C and 1 atm, in the presence of molybdenum dioxide. The high
activity shown by the dioxide can be explained in terms of the Mars and van Krevelen
mechanism [12], which involves oxygen provided by the oxide bulk structure to sustain
the redox cycles taking place on the catalyst surface. The unusual metallic character
displayed by this oxide appears to enhance the formation of hydrogen atoms, which, in
turn, accelerate the hydrocarbon decomposition. The high electrical conductivity
displayed by Mo dioxide is attributed to its relatively high density of states observed in
the valence band energy region. The existence of free electrons in this region enhances
the catalytic activity of Mo4+ in MoO2, unlike Mo6+ in MoO3, where all the valence
electrons of the metal are covalently bonded to neighboring oxygen atoms [7]. The
oxygen mobility can be interpreted as a result of the strong ionic character displayed by
the dioxide and the rapid formation of oxygen vacancies on the oxide structure.
Studies have been conducted to elucidate the behavior of the surface properties after
exposing the dioxide catalyst to H2 reduction [13] and [14]. Others have tried to establish
a relationship between the surface structure of the dioxide and its catalytic activity in
isomerization processes [15] and [16]. The use of molybdenum dioxide as a catalyst for
the partial oxidation of liquid hydrocarbons is a relatively new field. For this reason, no
study has previously reported the surface structure effect on the catalytic activity of the
dioxide for reforming processes. Thus, the present work is intended to investigate the role
played by the chemical states of Mo on the catalytic activity of MoO2 for the partial
97
oxidation of isooctane. Activity tests, XPS, UPS, and XRD are the methods used to
achieve the goal of this investigation.
Experimental
Activity Tests
MoO2 catalyst was purchased from Alfa Aesar and pretreated with H2 at 450 °C for 1 h to
increase the concentration of dioxide and reduce those of higher oxides (Mo5+, Mo6+),
which have been found to negatively affect the catalytic reforming activity of MoO2 for
H2 production [11]. The catalytic activity was measured using a 12 mm fixed-bed tubular
(quartz) reactor. The catalyst sample (typically 0.5 g) was supported on a quartz wool
plug and placed inside the reactor. Isooctane was vaporized at 200 °C and then mixed
with air, which was employed as the oxygen source for the partial oxidation reaction. The
operating conditions employed in the activity tests were 700 °C, 1 atm, and an oxygen to
carbon ratio (O2/C) of 0.72. The weight hourly space velocity (WHSV) was 0.6 h [1] and
the flow rate of liquid hydrocarbon fed into the system was 0.4 ml/h. These particular
operating conditions were chosen on the basis of our previous study which showed that
they lead to a stable performance for over 500 min on stream [11]. A silicon carbide bed
was employed to enhance the mixing between the reactants. The liquid fuel was injected
into the reactor using a calibrated syringe pump whereas the amount of air was supplied
using a mass flow controller. The off-gas was cooled down to 5 °C to separate water,
non-reacted isooctane, and other possible condensable compounds. The dry gas product
was then analyzed using an SRI chromatograph to monitor H2, CO, CO2, and CH4
98
concentrations. The catalytic activity was analyzed in terms of hydrogen yield and carbon
conversion, which were calculated as follows:
where the symbols stands for the molar flow rates of the different components.
Powder X-ray Diffraction
The bulk structure of the spent samples was analyzed by powder XRD on a Philips
diffractometer using Co Kα radiation with an iron filter. Because of the samples used in
this work was powder MoO2 the Bragg–Brentano optical configuration was used for the
present study.
X-ray Photoemission Spectra
The XPS spectra of the spent samples were obtained with an AXIS-165 manufactured by
Kratos Analytical Inc. using an achromatic Mg Kα (1254 eV) X-ray radiation with a
power of 210 W. The powder samples were pressed against 99.99% pure indium
flattened shots (purchased from Alfa Aesar). To minimize photoemission from the
supporting material (indium, In), a large amount of powder sample was pressed onto the
substrate. This was confirmed by either the absence or the presence of a very small In 3d
XPS signal. The spectrometer was calibrated against both the Au 4f7/2 peak at 84.2 eV
and the Ag 3d5/2 peak at 368.5 eV. Possible charging effect was precluded by switching
99
on and off the neutralizer (flood gun). Undetected shift in the Mo 3d peak due to charging
was an indication that the XPS data could be recorded without any correction. Curve
fitting of the XPS peaks was performed with the commercial CasaXPS software using
Gaussian/Lorentzian line shape. To minimize the effects of air exposure, the samples
were cooled down to room temperature inside the reactor under helium. Preliminary
experiments performed in our lab indicated that the dioxide phase is stable in air, even at
temperatures as high as 200 °C. Thus, we can safely conclude that the air exposure at
room temperature, prior to the surface analysis by photoelectron spectroscopy, was
unable to significantly affect the surface composition of the spent samples.
Ultraviolet Photoemission Spectra
The UPS spectra of the spent samples were obtained with a homemade cathode discharge
He lamp connected to the AXIS-165 by a UHV valve. The He I line at 21.21 eV was used
as the exciting source with an energy resolution of less than 150 meV as determined at
the Fermi edge of a clean polycrystalline Mo foil.
Results and Discussion
Activity Tests
The catalytic performance obtained at different times on stream (TOS) is reported in
Table 1. As seen, the conversion appeared not to be heavily affected and only diminished
from 100%, at the start of the test, to 99%, after 43 h of reaction. The hydrogen yield
reached its maximum value (91%) after 30 min on stream. Higher yields were not
100
achieved due to the formation of other H2-containing compounds, such as methane and
water. As the time of reaction becomes longer, the concentrations of H2 and CO
progressively decrease while that of CO2 experienced a slight increase. This finding can
be attributed to a change in the selectivity of MoO2, as the fuel is converted to CO2 and
H2O through a full oxidation process rather than to CO and H2 via partial oxidation. As
pointed out earlier, the catalytic activity of MoO2 for the partial oxidation of isooctane
may be explained in terms of a Mars–van Krevelen-type mechanism. Such mechanisms
imply the consumption of gas phase oxygen provided by the bulk structure to re-oxidize
the active sites previously reduced during the interaction between the hydrocarbon
molecules and the catalyst surface. The spent samples obtained from the activity tests
were analyzed and the results are reported in the next section. The following five samples
were used in this study: the one at t = 0 (after hydrogen treatment and prior to exposure to
the reforming environment) and the ones obtained after 0.5, 9.0, 20, and 43 h on stream.
Powder X-ray Diffraction
Fig. 1a reports the evolution of the catalyst bulk structure as a function of the time on
stream. Only two crystalline structures were detected in all the samples: the dioxide phase
(2θ = 30.7°, 43.6°, and 44.0°) and the carbide phase (2θ = 40.5°, 44.7°, and 46.5°). Fig.
1b shows the ratio between the XRD peak intensities of carbide (2θ = 46.5°) and dioxide
(2θ = 30.7°) as a function of the time on stream. This plot eliminates the effect of
possible changes in the peak intensities due to the different XRD measuring modes,
which may affect the reproducibility of the spectra on an absolute scale. As seen, Fig. 1b
shows a continuous increase of the ratio with time. As observed from Fig. 1, the peaks
101
assigned to the dioxide phase experienced a systematic decrease as a result of its
reduction to form the carbide phase. The decline in the amount of dioxide with time
agrees with the drop observed in the performance. Conversely, the amount of carbide in
the bulk phase appears to increase during the first 20 h. The pattern of the sample
obtained after 43 h on stream from Fig. 1a seems to disagree with this trend as it exhibits
smaller carbide peaks as compared to those detected at 20 h. The continuous decrease in
the amount of dioxide, along with the absence of additional peaks (other than those for
Mo dioxide and Mo carbide) in the 43-h-X-ray powder diffractogram from Fig. 1a led us
to believe that this reduction in the amount of carbide may be related to a degradation in
the crystalline structure of the carbide phase. Initially, the formation of the carbide phase
can be attributed to the fast diffusion of the carbon formed from the decomposition of the
fuel into the bulk structure, where oxygen vacancies are created as a result of the Mars–
van Krevelen mechanism. Thus, under the reducing environment, the dioxide phase
becomes unstable due to the higher rates of diffusion of carbon as compared to that of
oxygen, which is required to stabilize the dioxide phase. However; as the amount of
carbide increases, the number of oxygen vacancies in the bulk reduces. Consequently, the
degradation of the carbide phase may be the result of its interaction with the oxygen fed
into the system, to produce an oxygen-modified carbide compound, also known as
oxycarbide [17], which is able to exhibit amorphous structure [18].
102
X-ray Photoemission spectra
The Mo 3d XPS spectra of the catalyst samples at time 0, 0.5 9, 20 and 43 h and for the
clean polycrystalline Mo foil are shown in Fig. 2. The deconvolution of Mo 3d of the
starting material, commercial MoO2 powder, reveals the existence of different oxidation
states; Mo4+ to Mo6+ (spectrum not shown). Peak at BE = 229.5 eV and 232.7 eV were
assigned to Mo4+ and Mo 6+, respectively. In this study a peak measured at BE
231.5 eV was predominantly assigned to Mo5+. However its assignment remains
somewhat controversial in the literature. In their work on polycrystalline Mo foil Brox
and Olefjord [19] assigned a peak at 231.1 eV to a satellite of Mo4+ after oxidation in
water vapor. On the other hand, in his study on termination of the passivation of
elemental metals Barr [20] assigned the peak at about 231.4 eV to Mo5+ after exposing
polycrystalline foil to air. We based our assignment of the peak at BE = 231.5 eV to Mo5+
on the fact that the intensity ratio Mo4+/Mo5+ keeps on changing with reforming time
which is an indication that the amount of surface oxygen in our system is continuously
changing (reduced and replenished) during reforming. In Fig. 2, the XPS data indicates
the formation of a new Mo phase on the catalyst surface in between the binding energy
(BE) regions, commonly assigned to the Mo metal and Mo4+ phases. The deconvolution
of the XPS spectra allowed us to determine the position of the peaks attributed to the
different Mo oxidation states as a function of the time on stream. The information
obtained is summarized in Table 2. The oxidation state δ for the new phase was estimated
by interpolation using our experimental values combined with data from previous
investigations [21] and [22]. The results of these calculations are shown in Fig. 3. Thus,
103
after 43 h on stream the BE for the new phase was 228.5 eV which corresponds to an
oxidation state of approximately δ = 2.6+. Choi and Thompson [14] assigned a similar
Mo 3d peak (228.7 eV) to the oxidation number 3+, based on the binding energy
assignments of Mo0, Mo4+, Mo5+, and Mo6+. The oxidation state 3+ for Mo was also
suggested by Roxlo et al. [23] for their textured Mo–S system. To identify the new Mo
phase, additional information from the XPS analysis was required. For instance,
Zafeiratos et al. [24] assigns the oxidation state 3+ to a Mo oxyhydroxide phase
(MoO(OH)). However, the concentration of OH groups, associated with a BE = 531.4 eV
in the O1s spectrum (not shown), increases only by 6% between 0.5 and 43 h on stream.
This small change is not able to account for the large rise detected for the concentration
of the new phase Moδ+, which increases from 4% to 45% during the same period of time.
On the other hand, Ovari et al. [17] suggests a state of oxidation between 2+ and 4+ as a
result of the interaction between oxygen and Mo carbide to form an oxycarbide phase.
However, the appearance of an oxycarbide phase seems to be in contradiction with our
XPS results since no peak was detected in the XPS C1s signal at 283.3 eV as seen in Fig.
2b, which has been attributed to the presence of carbidic carbon involved in Mo–C bonds
[8], [9], [10] and [25]. Lu et al. [26] attribute the peak at 228.4 eV in the Mo3d spectrum
to the Mo oxide phase Mo2O3, which appears to be in good agreement with the XPS data
reported above. Thus, the new oxidation state of 2.6+ (or approximately 3+) may be the
result of the reduction of the dioxide phase to form Mo2O3 due to its interaction with the
reforming environment. This agrees with the trend observed for the bulk structure, which
progressively becomes reduced as the time on stream increases.
104
Fig. 2. (a) XPS spectra of Mo3d for MoO2 as a function of the time on stream (0, 0.5, 9,
20 and 43 h). The spectrum obtained on clean polycrystalline Mo foil is included as
reference (Mo0 = 228.08 eV). (b) XPS spectra of C 1s for 9, 20 and 40 h on stream. Note
that there is no peak corresponding to carbidic carbon (BE = 283.3 eV).
Note that the atomic concentration (At%) values listed here for the various Mo species
are based on the deconvolution of the Mo 3d peak.
The increase in the concentration of Moδ+ with time seems to be related to the decline
observed in the H2 yield, as reported in Table 1. Thus, at TOS = 0.5 h, the hydrogen yield
reached its maximum value (91%) and reduced to 77% after 43 h of reaction, whereas the
concentration of Moδ+ increased from 4% at t = 0.5 h to 45% at t = 43 h (see Fig. 4). The
concentrations of the different Mo oxidation states were obtained by curve fitting the Mo
3d spectrum. The area ratio and the energy split between Mo 3d5/2 and Mo 3d3/2 were fix
to 0.66 and 3.15 eV, respectively, to compensate the peak overlapping observed in the
Mo 3d region. In addition, the FWHM for Mo4+, Mo5+ and Mo6+ were set to 1.4 eV and
to 1.1 eV for Moδ+. These parameters allow for the best fit of our XPS Mo 3d peak.
Moreover, powder samples exhibit generally linewidth slightly broader than single
crystalline samples. For detailed information on linewidth measured for the various single
crystal Mo species, the reader is referred to the excellent work of Werfel and Minni [27].
Table 2 indicates no variations in the BE within the experimental error of ± 0.1 eV of the
species Moδ+ and Mo4+ between 0.5 and 43 h. On the contrary, slight changes were
observed for the species Mo5+ and Mo6+, whose BE remains relatively constant during the
105
first 20 h on stream to eventually decline as the time of reaction approaches the 43 h.
Kaya et al. [28] suggest that the Mo5+ and Mo6+ (higher BE states) may be more directly
affected by the environment given their stronger surface nature, while Moδ+ and Mo4+
(lower BE species) are more associated with either the sub-surface or the bulk structure.
This explains the trends of the binding energy changes as a function of the time on stream
for the various oxidation states reported in Table 2. Thus, the states Moδ+ and Mo4+
display more stable binding energies since they are located in the sub-surface of the
sample and are not fully exposed to the reforming environment. Also shown in Table 2 is
the surface atomic relative concentration of the various oxidation states of Mo with time
on stream. Because Mo6+ is an outer oxidation state [28], it is expected to be highly
affected by the reforming environment, which seems to agree with the data reported in
Table 2. Thus, the concentration of Mo6+ appears to initially decrease slightly from 19%
to 17% as a result of the strongly reducing environment. After 9 h on stream, the amount
of carbide in the bulk becomes significant (see Fig. 1a), and therefore, the migration of
oxygen into the bulk required by the Mars–van Krevelen mechanism is delayed. As a
result, the concentration of oxygen on the surface increases in concordance with the rise
in the concentration of the Mo6+ species from 17% to 23% observed after 9 h on stream.
The decline in the concentration of Mo6+ observed at 43 h can be explained by
considering the increase in the rates of oxygen diffusion from the surface into the bulk
structure to form probably an amorphous oxycarbide phase.
Fig. 5 shows the XPS valence band for the samples at 0.5, 9, 20 and 43 h on stream, and
for a clean Mo foil used as a reference material. The band emissions (see lines in Fig. 5)
106
detected for the Mo foil at about 0.3, 2.0, 4.1 (weak) and 6.4 eV are in good agreement
with the results of Werfel and Minni [27], obtained on single crystalline Mo(0 0 1). In all
the dioxide samples, the valence band region between the Fermi edge (EF) and 12 eV
displayed two distinct bands with maximum intensities at about 2 and 7 eV, which are
assigned to the Mo 4d and O 2p derived band, respectively. This is in good agreement
with the results reported by Werfel and Minni [27], who also attributed the overlapping
center region at about 6 eV to a larger contribution of the Mo cationic orbital, which
interacts with the O 2p orbital to form the covalent bonding part. A shoulder labeled (s)
in Fig. 5 at a lower BE of 2 eV was detected in all the MoO2 spectra. This peak is
characteristic of the MoO2 powder, as it was not observed for metallic Mo. The
appearance of this shoulder is related to the existence of two different BEs associated
with the oxidation state 4+, as suggested by Haber et al. [29]. One of these BEs can be
assigned to the isolated Mo4+ ions and the other one to paired double-bonded Mo4+,
where the core levels are additionally shielded by a common electron pair. The intensity
of the shoulder becomes lower as the time of reaction increases. After 43 h on stream, the
overall shape of the valence band in the region between 4 eV and the Fermi edge
becomes very similar to that of metallic Mo foil. This suggests that the metallic character
of the sample may not be heavily affected by the appearance of the new phase Moδ+,
which is confirmed by the trend observed for the catalytic performance during the late
stages of the activity tests.
107
Ultraviolet Photoemission Spectra
He I spectrum of polycrystalline Mo foil cleaned by Ar+ sputtering (full spectrum not
shown) displays a clear Fermi edge. The work function was determined experimentally to
be 4.52 eV, which agrees very well with the literature value of 4.5 eV [30]. He I spectra
of MoO2, after reforming times of 0.5, 9, 20, and 43 h, are shown in Fig. 5 (right). The
insert displays the Fermi edge of clean polycrystalline Mo foil. The valence band region
between 0 and 2.5 eV below the Fermi level for MoO2 (t = 0.5 h) shows two intense and
distinct bands. The position of these two bands are at BE = 1.62 and 0.55 eV. Tokarz-
Sobieraj et al. [31] measured a first band at 1.6 eV and a second band at 0.4 eV below EF
for MoO2 single crystal using He II (40.8 eV). They assigned these two bands to the
known two different Mo4+ sites existing in the MoO2 lattice. UPS data, seen in Fig. 5,
shows that the intensities of these two bands are significantly reduced as the time
increases from 0.5 to 9 h. This may be attributed to the gradual transformation of Mo4+
into the new state Moδ+, which eventually led to the abatement of the d-band character of
MoO2. In their theoretical work, Santos and Schmickler [32] point out that catalytic
process involving metals are strongly affected by d-band structures with energies near the
Fermi level. The metallic character of MoO2 is seen in both XPS and UPS valence band
region in the finite density of state (DOS) at the Fermi edge. Based on this information,
we can establish a relationship between catalyst performance and the UPS spectrum as
follows. At the early stages of the process (t = 0.5 h), the catalyst surface displays a high
concentration of the species Mo4+ (46%), whose d-band character is strongly related to its
high catalytic activity. The decline in the hydrogen yield observed after 9 h of reaction
108
agrees with the abatement in the Mo4+ d-band character detected by UPS. For longer
times on stream, the catalyst performance experienced a gradual decrease to end with a
hydrogen yield of 77%, after 43 h of reaction. This agrees with the reduction in the
concentration of the species Mo4+ to 27%, as observed in Fig. 4. The UPS spectra
obtained at t = 20 h and t = 43 h suggest the presence of an arising metallic character
different from that observed for Mo4+. We attribute this effect to the new phase Moδ+,
which appears to be related to the catalytic activity observed at the late stages of the
process.
Conclusions
X-Ray diffraction measurements indicate that, under the reforming conditions employed
in this work, the bulk structure MoO2 is gradually transformed into molybdenum carbide,
Mo2C. The photoemission results demonstrate that the drop in the catalytic activity of
molybdenum dioxide for the partial oxidation of isooctane can be correlated with the
decline in the concentration of Mo4+ due to the appearance of a new phase Moδ+ (δ = 2.6)
which we believe is Mo2O3. The He I UPS spectrum obtained for the 0.5-h spent sample
exhibits two bands near the Fermi edge that decrease in intensity and disappear after 9 h
on stream. XPS and UPS findings reinforce the fact that the reduction of the Mo
oxidation numbers from 4+ to δ+ (δ = 2.6) and the change in its electronic structure, in
particular the density of the d-band; significantly affect the catalytic performance of
MoO2 for the partial oxidation of isooctane.
109
28 30 32 34 36 38 40 42 44 46 48 502 theta
Inte
ns
ity
(a
.s.)
MoO2 Mo2C
t =0.5 h
t =9 h
t =20 h
t =43 h
t =0 h
a)
0
1
2
3
4
5
0 5 10 15 20 25 30 35 40 45
Time-on-stream (h)
Mo
2C
/ M
oO
2 r
ati
o
b)
Figure 1: a) XRD pattern of the spent samples as a function of time;
b) Mo carbide to Mo dioxide ratio as a function of time
110
0.0 h
0.5 h
9.0 h
20 h
43 h
240 238 236 234 232 230 228 226
Mo foil
Mo 3d
Binding Energy (eV)
Figure 2: (a) XPS spectra of Mo3d for MoO2 as a function of the time on stream (0, 0.5,
9, 20 and 43 h). The spectrum obtained on clean polycrystalline Mo foil is included as
reference (Mo0 = 228.08 eV). (b) XPS spectra of C1s for 9, 20 and 40 h on stream. Note
that there is no peak corresponding to carbidic carbon (BE=283.3 eV).
(a)
111
0
1
2
3
4
5
6
7
227 228 229 230 231 232 233
Binding Energy (eV)
Yamada et al. [16]
This work
δ = 2.6+
Ox
ida
tio
n S
tate
Figure 3: Binding energy (BE) versus oxidation number of Mo atom
based on the 43-hour sample.
112
240 238 236 234 232 230 228 226
240 238 236 234 232 230 228 226 Binding Energy (eV)
Mo δ+
(45%)
Mo 4+
(27%) Mo 5+
(15%)
Mo 6+
(13%)
43 h
0.5 h
Mo δ+
(4%)
Mo 4+
(46%)
Mo 5+
(33%) Mo 6+
(17%)
Figure 4: Deconvolution of XPS spectra of Mo3d from Figure 2 at 0.5 and 43 hours.
113
Figure 5: (Left) XPS valence band as function of time on stream. XPS valence band for
Mo foil is included as reference. (Right) UPS He I spectra for the four samples and the
clean Mo foil reference (inset).
14 12 10 8 6 4 2 0
Binding Energy (eV)
MoO 2
Foil
Mo 4d met O 2p derived band
Mo 4d met
4d 2p 2
20
43
MoO
0.5
= EF
Binding Energy (eV)
3 2.5 2 1.5 1 0.5 0
0.5
20
9
43 EF
0.00.51.0
Mo Foil
XPS UPS
s
1 2
3
4
114
TOS (h) H2 yield Conversion H2% CH4% CO% CO2%
0.5 91% 100% 23% 5% 16% 5%
1.0 90% 100% 23% 4% 15% 5%
1.5 83% 100% 22% 4% 14% 6%
3.0 83% 100% 21% 4% 13% 7%
5.0 83% 100% 22% 4% 14% 8%
9.0 82% 100% 21% 3% 12% 8%
16.0 79% 100% 20% 3% 11% 9%
20.0 78% 100% 20% 4% 12% 9%
27.0 76% 99% 20% 3% 12% 8%
43.0 77% 99% 20% 3% 12% 8%
Table 1: MoO2 catalytic performance for the partial oxidation of isooctane
(700°C, 1 atm, O2/C = 0.72).
115
TOS (h) 0.0 At% 0.5 At%
9.0 At% 20.0 At% 43.0 At%
Moδδδδ+ - - 228.5 4 228.5 14 228.5 23 228.5 45
Mo4+ 229.5 47 229.5 46 229.5 35 229.5 28 229.5 27
Mo5+ 231.3 34 231.4 33 231.4 28 231.4 26 231.0 15
Mo6+ 232.8 19 232.8 17 232.8 23 232.8 23 232.5 13
Note that the atomic concentration (At %) values listed here for the various Mo species
are based on the deconvolution of the Mo 3d peak.
Table 2: Peak position and atomic concentration for Mo oxidation states as a function of the time on stream
116
References
[1] X. Chen, A. Tadd, J. Schwank, J. Catal. 251 (2007) 374.
[2] H. Ibrahim, P. Kumar, R. Idem, Energy Fuels 21 (2) (2007) 570.
[3] Y. Lu, J. Chen, Y. Liu, Q. Xue, M. He, J. Catal. 254 (2008) 39.
[4] R. Biniwale, A. Mizuno, M. Ichikawa, Appl. Catal. A: Gen. 276 (2004) 169.
[5] Y. Rezgui, M. Guemini, Ind. Eng. Chem. Res. 47 (2008) 4056.
[6] S. Al-Kandari, H. Al-Kandari, F. Al Kharafi, A. Katrib, J. Phys.: Conf. Ser. 100
(2008) 012005.
[7] J. Song, P. Chen, S. Kim, G. Somorjai, R. Gartside, F. Dautzenergb, J. Mol. Catal.
A: Chem. 184 (2002) 197.
[8] Ákos Koós, Albert Oszko, Appl. Surf. Sci. 253 (2007) 3022.
[9] Ding Ma, Yuying Shu, Mojie Cheng, Yide Xu, Xinhe Bao, J. Catal. 194 (2000)
105.
[10] Dingjun Wang, Jack H. Lunsford, Michael P. Rosyneck, J. Catal. 169 (1997) 347.
[11] O. Marin-Flores, S. Ha, Appl. Catal. A: Gen. 352 (2009) 124.
[12] M. Vannice, Catal. Today 123 (2007) 18.
[13] Katrib, P. Leflaive, L. Hilarie, G. Maire, Catal. Lett. 38 (1996) 95.
[14] G. Choi, L. Thompson, Appl. Surf. Sci. 93 (1996) 143.
[15] F. Al-Kharafi, H. Al-Kandari, A. Katrib, Catal. Lett. 123 (2008) 269.
[16] Katrib, J.W. Sobczak, M. Krawczyk, L. Zommer, A. Benadda, A. Jablonski, G.
Maire, Surf. Interf. Anal. 34 (2002) 225.
[17] L. Ovari, J. Kiss, A. Farkas, F. Solymosi, J. Phys. Chem. B 109 (2005) 4638.
[18] P. Delporte, F. Meunier, C. Pham-Huu, P. Vennegues, M. Ledoux, J. Guille, Catal.
Today 23 (1995) 251.
117
[19] B. Brox, I. Olefjord, Surf. Interf. Anal. 13 (1988) 3.
[20] L. Barr Terry, J. Phys. Chem. 82 (1978) 1801.
[21] M. Yamada, J. Yasumaru, M. Houalla, D. Hercules, J. Phys. Chem. 95 (1991) 7037.
[22] E. Broclawik, A. Foti, V. Smith, J. Catal. 62 (1980) 185.
[23] C. Roxlo, H. Deckman, J. Gland, S. Cameron, R. Chianelli, Science 235 (1987)
1629.
[24] S. Zafeiratos, G. Papakonstantinou, M.M. Jacksic, S.G. Neophytides, J. Catal. 232
(2005) 127.
[25] P. Perez-Romo, C. Potvin, J. Manoli, M. Chehimi, G. Djega-Mariadassou, J. Catal.
208 (2002) 187.
[26] Y. Lu, C. Clayton, Corros. Sci. 29 (1989) 927.
[27] F. Werfel, E. Minni, J. Phys. C: Solid State Phys. 16 (1983) 6091.
[28] S. Kaya, J. Weissenrieder, D. Stacchiola, T. Todorova, M. Sierka, J. Sauer, S.
Shaikhutdinov, H. Freund, Surf. Sci. 602 (2008) 3338.
[29] J. Haber, W. Marczewski, J. Stoch, L. Ungier, Berichte der Bunsen-Gesellschaft 79
(1975) 970.
[30] U. Petermann, I.D. Baikie, B. Lägel, K.M. Dirscherl, Mater. Res. Soc. Symp. 615
(2000) G6.6.1.
[31] R. Tokarz-Sobieraj, K. Hermann, M. Witko, A. Blume, G. Mestl, R. Schlogl, Surf.
Sci. 489 (2001) 107.
[32] E. Santos, W. Schmickler, Chem. Phys. Chem. 7 (2006) 2282.
118
CHAPTER V
THERMODYNAMIC AND EXPERIMENTAL STUDY OF
THE PARTIAL OXIDATION OF A JET A FUEL SURROGATE OVER
MOLYBDENUM DIOXIDE
Abstract
In the present work, a combination of thermodynamic calculations and experimental
results was used to investigate the activity and stability of commercial Mo dioxide
(MoO2) as catalyst for the partial oxidation of aviation jet fuels. N-dodecane was used as
surrogate jet fuel. Our results indicate that the stability window for MoO2 is strongly
affected by the O2/C molar ratio. Thus, the formation of elemental carbon on the catalyst
structure can be prevented using O2/C ratios higher than 0.5. However, O2/C ratios higher
than 1.0 enhance the formation of Mo trioxide which is volatile and leads to the
irreversible loss of catalytic material. The activity was measured at 850°C and 1 atm and
our findings indicate that, within the stability window determined earlier, the production
rates of H2 and CO can reach values as high as 78% and 92%, respectively. The coking
resistance of MoO2 was compared with that of a commercial nickel catalyst by
performing activity tests under coke-promoting conditions. EDX analysis of the spent
samples shows that MoO2 is much more resistant to deactivation by coking than
commercial nickel catalyst. Based on these results, Mo dioxide appears as a promising
catalyst for the partial oxidation of jet fuels.
119
Introduction
Many investigators have focused their efforts on the development of high-performance
catalysts for the external reforming of transportation fuels [1, 2, 3]. This process involves
the use of an on-board fuel processor to transform the hydrocarbon fuel into a gas
mixture that is directly used by fuel cell devices. The U.S. standard fuel for commercial
aviation is denominated Jet A and is formed by a mixture of hundreds of hydrocarbons of
different structure/properties. The high level of complexity of this mixture increases
significantly the challenges to develop new catalytic materials for reforming purposes.
The use of surrogate fuels significantly simplifies the analysis. Hence, many attempts
have been made to develop surrogates with well-known features and that emulate specific
properties of the fuel [4, 5]. Among these substitutes, n-dodecane is the most widely used
surrogate for Jet A fuel as it is the paraffinic hydrocarbon with the largest concentration
in real jet A fuels [6].
Ni-based compounds are the catalytic materials currently utilized for the reforming of
hydrocarbon mixtures such as jet fuels. However, the high content of aromatic
compounds in the fuel negatively affects the performance of these catalysts due to the
formation of carbonaceous deposits, which rapidly lead to the deactivation of the
catalytic material [7]. Thus, the search for more stable catalysts for the reforming of
complex hydrocarbon mixtures is still an on-going task.
Molybdenum dioxide (MoO2) has been recently reported to be highly active for the
reforming of both isooctane and gasoline via partial oxidation processes [8]. The activity
120
of this transition metal oxide has been explained in terms of a Mars-van-Krevelen-type
mechanism of reaction, which involves the consumption of nucleophilic oxygen ions
provided by the lattice structure of the oxide to sustain the redox cycles taking place on
the catalyst surface [8]. The major drawback in the use of Mo dioxide as catalytic
material for reforming processes is the loss of activity due to phase transitions occurring
as a result of the exposure of the catalyst to the reforming environment. The stability
problem showed by Mo dioxide has not been sufficiently studied by previous
investigators, and thus, the information about the system Mo-C-O is scarce. A theoretical
study based on thermodynamic principles can reduce the intricacy of the problem and
help us better understand the behavior of such a complex system under various reforming
environments. This approach has been extensively used in different systems and the
results are reported everywhere [9], [10], [11], [12].
Thus, the present work will aid to overcome the lack of information with regard to the
performance and stability of Mo dioxide as catalyst for the partial oxidation of a jet fuel
surrogate. To achieve this goal, a thermodynamic study of both the Mo-O-C system and
the reforming process was carried out to investigate the stability of the MoO2 phase under
reforming environments as well as the catalytic activity of this transition metal oxide for
the production of syngas from a jet A fuel surrogate.
The Mo-O-C system
A thermodynamic analysis to study the stability of the catalytic material was developed
on the basis of some assumptions formulated to simplify the high level of complexity
121
exhibited by the system. As found in our previous research performed with gasoline [8],
the only Mo phases detected under reforming environments were exclusively those
formed with either C or O. The amounts of these elements in the system are related to
each other by the oxygen to carbon molar ratio (O2/C), which is defined as the ratio of the
number of moles of molecular oxygen to the number of moles of carbon contained in the
hydrocarbon fuel fed into the system. Hence, our analysis was focused on the effect that
carbon and oxygen may have on the Mo dioxide phase as the hydrogen contained in the
fuel was assumed to have low impact on the catalyst stability.
Several intermediate oxides with Mo oxidation states between 4+ and 6+ have been
reported in the literature. Thus, Kihlborg [13, 14] was able to identify the crystal structure
of intermediate Mo oxides existing between 500°C and 800°C. The Mo oxides found by
this investigator, in addition to MoO2 and MoO3, were monoclinic Mo4O11 (η-oxide),
orthorrombic Mo4O11 (γ-oxide), tetragonal Mo5O14 (θ-oxide), Mo17O47 (χ-oxide), Mo8O23
(β-oxide), monoclinic Mo9O26 (β'-oxide), and triclinic Mo9O26 (ζ-oxide). Kihlborg’s
findings indicate that MoO2 and MoO3 are the only phases stable at temperatures above
800°C. The melting point of MoO2 has been found to be above 2200°C [15]. However, at
lower temperatures MoO2 can vaporize as a result of its disproportionation to Mo trioxide
and Mo metal, forming MoO2 vapor as a side product. Nevertheless, Burns et al. [16] and
Blackburn et al. [17] found that these reactions take place at a very small extent, even at
temperatures above 1200°C. Mo trioxide is a binary Mo oxide with relatively low
melting point (795°C, [18]) and elevated vapor pressures. The molecules of MoO3 in the
vapor phase appear to be unstable as monomers and rapidly produce polymeric forms
122
([19, 20, 21]), which, below 1000°C, are mainly trimer Mo3O9, tetramer Mo4O12 and
pentamer Mo5O15 [22].
The binary phases Mo-C exhibit a similar complexity to that shown by Mo-O phases.
Several stable and metastable phases have been reported in the literature. Thus, Lu et al.
[23] reported seven molybdenum carbide phases: hexagonal Mo2C (β-carbide),
orthorrombic Mo2C (β’or α-carbide), hexagonal MoC1-x (η-carbide), cubic MoC1-x (δ-
carbide), hexagonal MoC (γ-carbide), and hexagonal MoC1-x (γ’-carbide). Only two
carbide phases are thermodynamically stable at room temperature, the β’-Mo2C phase
and the γ-MoC phase. The β’-phase transforms into the hexagonal β-Mo2C at
approximately 1230°C [24]. The η and δ carbide phases are stable only at higher
temperatures. The hexagonal γ’-carbide phase is a metastable phase that appears to be
stabilized by oxygen and, therefore, it should be considered as an oxycarbide. Kuo et al.
found that this phase transforms into the more stable γ-carbide phase at about 800°C [25].
However, some other investigators believe γ-carbide phase is also a metastable phase
stabilized by oxygen or by other impurities [26].
The formation of ternary phases increases the intricacy of the system Mo-C-O in itself
already complicated due to the formation of different binary phases. Mo oxycarbides
(MoCxOy) are ternary phases which can form as a result of either the reduction of Mo
oxides [27] or the oxidation of carbon-containing Mo phases [28]. Oxycarbides have
been reported as active catalytic materials for different processes [29, 30]; however, the
123
stability of such oxycarbide phases decreases at high temperatures [31]. Thus, at 800°C
oxycarbides can thermally decompose to produce Mo carbide, Mo dioxide and Mo metal.
The partial oxidation of n-dodecane
The partial oxidation of n-dodecane can be represented as follows:
C12H26 + 6 O2 � 12 CO + 13 H2 (1)
The transformation of the hydrocarbon into CO and H2 has been explained using different
reaction schemes like the one proposed by Pacheco et al. [32], who assume the
hydrocarbon to undergo combustion followed by both steam and carbon dioxide
reforming. This reaction scheme can be described as
C12H26 + 37/2 O2 � 12 CO2 + 13 H2O (2)
C12H26 + 12 H2O � 12 CO + 25 H2 (3)
C12H26 + 12 CO2 � 24 CO + 13 H2 (4)
Reaction (2) describes the catalytic combustion of the hydrocarbon fuel, while reactions
(3) and (4) represent the steam and dry reforming processes, respectively. In addition, the
steam reforming of dodecane can also lead to the formation of CO2 and H2, according to:
C12H26 + 24 H2O � 12 CO2 + 37 H2 (5)
124
The presence of CO and water in the system suggests that the water-gas shift reaction
may take place as shown below:
CO + H2O � CO2 + H2 (6)
Methane evolution can be explained in terms of the methanation process, which is
formulated as follows:
CO + 3 H2 � CH4 + H2O (7)
Finally, the formation of coke is associated to the Boudouard reaction, which is
formulated as shown below:
2CO � CO2 + C (8)
Methodology
The use of equilibrium constants of known reactions to determine the composition of an
equilibrium mixture has become the traditional and most simple approach to perform a
thermodynamic analysis for multicomponent systems. However, when the equilibrium
compositions must be determined by a large number of simultaneous reactions, the
computational work required for this purpose becomes largely tedious.
A more general method for solving this complicated problem is through the minimization
of the Gibbs function. According to the second law of thermodynamics, at the
125
equilibrium state, the system has the lowest possible Gibbs free energy from all possible
states.
Let us consider a closed system allowed to perform only pressure-volume work. For a
system with c components and π phases the total Gibbs free energy is given by:
∑∑= =
=c
i
ji
j
jinG1 1
µπ
(9)
where j refers to the phase (j=1,…, π) and i to the components (i = 1, ..., c). n stands for
the number of moles and µ for the chemical potential.
The problem now is to find the set of values for n that minimizes G at constant T and P,
subject to the constraints imposed by the elemental balances.
k
c
i j
jiijk bna =∑∑= =1 1
π
k = 1,…, m (10)
where aijk is the number of atoms of the kth element in a mole of the ith species in the jth
phase. bk is defined as the total number of atoms of the kth element in the reaction
mixture. Thus, using vector notation, the problem is formulated as follows:
),,(min nPTGn
(11)
ban =
126
To solve this minimization problem a Lagrange function is defined as shown below:
( )ban −−=
−−= ∑ ∑∑
= = =
m
k
Tc
i j
kjiijkk GbnaGL1 1 1
λλλλπ
λ (12)
The chemical potential can be expressed in terms of the fugacities if̂ as follows:
0
0,
ˆln
i
ipureii
f
fRT+= µµ (13)
where iµ represents the chemical potential for the component i in the mixture, and
0, pureiµ stands for the chemical potential of pure i. The standard states 0
if for the
fugacities are chosen based on the phase to which the component i belongs. At low
pressures and high temperatures the ideal gas model describes accurately the behavior of
a gas phase; therefore, the fugacity of a gas component G
if̂ equals the partial pressure of
such a component. For condensed phases, the fugacity equals the vapor pressure in the
case of liquids, or the sublimation pressure, in the case of solid phases. The standard state
for both gas components and condensed phases was assumed to be 1 bar.
Based on the information provided in section II.1, the following phases were considered
for our analysis:
Solid phases: Mo, γ-MoC, α-Mo2C, MoO2
Liquid phases: MoO3
127
Gas-phase compounds: CO, CO2, O2, (MoO3)x
The vapor pressure data provided by Ghosh [15] led us to consider MoO3 as the only
condensed phase able to produce gas-phase compounds.
The expressions used in our calculations are shown below:
Catalyst stability analysis (T in K and DG in J/mol):
• Solid γ-MoC (Mo monocarbide):
29430046.110368.8 240
)(−−×=∆ −
−TTG
sMoCγ (J/mol)
• Solid α-Mo2C (Mo carbide):
49170469.310301.2 230
)(2−−×−=∆ −
−TTG
sCMoα (J/mol)
• Solid MoO2:
58880010912.110891.8 2230
)(2−×+×−=∆ − TTG
sMoO (J/mol)
• Liquid MoO3:
60.2435200
log576.4)(3
+−=T
Pvap
lMoO
128
70260010306.210850.1 2220
)(3−×+×−=∆ − TTG
lMoO
• MoO3 in gas-phase (MoO3)x:
1326 10377.210986.410416.1 −−− ×+×+×−= TTx
522150 10674.310054.310776.1)3(
×+×−×−=∆ − TTGxMoO
(J/mol)
• Carbon oxides:
88035019.880 −−=∆ TGCO
(J/mol)
419192461.310
2−=∆ TG
CO (J/mol)
Partial oxidation process (T in °C and ∆G in J/mol):
• N-dodecane:
4321340 10913.110174.110368.710922.12612
×+×+×+×−=∆ −− TTTGHC
• Carbon dioxide
523380 10943.3091.310315.110085.92
×−−×+×−=∆ −− TTTGCO
• Carbon monoxide
129
5124360 10349.110075.910356.210220.1 ×−×−×+×=∆ −− TTTGCO
• Water
5122360 10297.210479.410117.110536.32
×−×+×+×−=∆ −− TTTGOH
• Methane
4122350 10292.510295.810114.310152.14
×−×+×+×−=∆ −− TTTGCH
Experimental
Commercial MoO2 catalyst (99%, metal basis) was purchased from Alfa Aesar. The
catalytic activity tests were performed using a 12 mm fixed-bed tubular (quartz) reactor.
n-dodecane (99+ %) purchased from Alfa Aesar was used as jet A fuel surrogate. A
syringe pump was used to inject the liquid fuel into a pre-heater at 500°C to be vaporized.
The amount of air was regulated using a mass flow controller. The mixing of vaporized
fuel and air was enhanced using a silicon carbide bed in the pre-heater. The operating
conditions in the reactor were 850°C and 1 atm. For each test, 0.5 g of fresh catalyst (no
pretreated) was supported on a quartz wool plug and placed inside the reactor. The off-
gas was cooled down to 5°C to separate water, non-reacted isooctane, and other possible
condensable compounds. The dry gas product was then analyzed using an SRI
chromatograph to monitor H2, CO, CO2, and CH4 concentrations.
130
The bulk structure of the spent samples was analyzed by powder x-ray diffraction (XRD)
on a Philips diffractometer using Co Kα radiation with an iron filter.
The catalytic performance was analyzed in terms of three parameters:
• H2 yield: the ratio between the hydrogen atoms produced in the form of molecular
hydrogen from the off-gas and the total amount of hydrogen atoms in the fuel stream
• CO yield: the ratio between the carbon atoms produced in the form of the carbon
monoxide and the total amount of carbon atoms in the fuel stream
• Conversion: the ratio between the total amount of carbon atoms produced in the form
of carbon monoxide, carbon dioxide and methane and the total amount of carbon
atoms in the fuel stream
Results and Discussion
Analysis of the catalyst stability
Using the methodology explained earlier, the ternary phase diagram Mo-C-O was
calculated at 850°C and 1 atm. The temperature chosen for our analysis lies within the
typical range used for the reforming of liquid hydrocarbons such as jet fuel or gasoline
due to the fact that it favors the kinetics of the reforming process. In addition, at 850°C
ternary phases such as oxycarbides are not stable and, therefore, the analysis of the
system becomes simpler. As a result, the diagram obtained considers only those phases
131
formed by Mo, C and O that are stable under the conditions of pressure and temperature
selected for this study. Figure 2 shows the calculated ternary phase diagram showing the
minimized Gibb free energy in a grayscale on the background. The results of our
calculations indicate that the maximum stability is attained by those regions where solid
Mo dioxide, solid Mo carbide and liquid Mo trioxide are present and coexisting with a
gas phase consisting basically of carbon dioxide.
Two important conclusions can be drawn from this diagram. The first one is related to the
formation of liquid Mo trioxide in equilibrium with a mixture of polymeric trioxide
species in the gas phase. Since our reactor behaves as an open system, the continuous
removal of the gas products will lead to a progressive reduction of Mo in the system
which, in turn, will produce a continuous decline in the catalytic activity. Therefore, the
formation of Mo trioxide must be avoided. The second one is the formation of carbon as
a result of using low O2/C ratios in the reaction. The ternary phase diagram indicates that
elemental carbon only coexists with Mo monocarbide and a gas phase formed by carbon
monoxide. Thus, as long as the formation of MoC is prevented, the risk of deactivation
due to coking will be minimized.
Effect of O2/C ratio
The O2/C ratio plays a significant role on the catalytic activity since it defines the type of
environment to which the catalytic material is being exposed. As deduced from Figure 1,
when using Mo dioxide as starting material, O2/C ratios between 0.5 and 1.0 lead to the
132
appearance of Mo carbide. This can be explained considering the following reactions as
responsible for the equilibrium between these phases [33, 34]:
2 MoO2 + 3 C = Mo2C + 2 CO2 (14)
Mo2C + 3 O2 = 2 MoO2 + CO2 (15)
The equilibrium constants for these reactions (see Figure 2) are significantly different and
indicate that the oxidation of the carbide phase by gas oxygen (reaction 15) is
thermodynamically favored in a much larger extent than the carbothermic reduction of
the Mo dioxide phase (reaction 14). However, a slow kinetics of reaction 15 may reverse
this trend and make it become the process responsible for the Mo dioxide stability.
Hence, increasing the concentration of gas oxygen in the system should enhance the rapid
oxidation of the carbide phase to Mo dioxide and favor thus the stability of this phase, as
predicted by Thermodynamics.
This hypothesis was experimentally validated by performing activity tests at O2/C ratios
ranging from 0.5 to 1. A fixed flow rate of liquid fuel was fed into the pre-heater along
with a proper flow rate of air to produce a mixture with a specific O2/C ratio, which was
then fed into the reactor for 3 hours. The spent catalyst samples were analyzed using X-
ray powder diffraction and the diffractograms obtained are shown in Figure 4. As
noticed, only peaks associated to the Mo carbide phase were detected in the patterns of
the spent samples obtained at O2/C ratios of 0.5 and 0.6. However, at a ratio of 0.7 the
Mo dioxide peaks started to show up and became larger as the ratio was systematically
133
increased from 0.7 to 1.0. At a ratio of 1.0 a small amount of MoO3 crystals were
observed at the bottom of the reactor, which indicates the formation and sublimation of a
small amount of this oxide.
In light of these results, the concentration of oxygen appears to be of key significance to
stabilize the Mo dioxide phase with an onset O2/C ratio of about 0.7. This is in good
agreement with additional thermodynamic calculations whose results are shown in Figure
5, and which were obtained considering a small and constant initial concentration of
MoO2 in the system (10% mol). The equilibrium compositions were calculated at
different initial O2/C ratios within the range 0.5-1.0. These calculations show that onset
ratio for the appearance of the Mo dioxide in the bulk of the sample is approximately
0.73 and that larger ratios increase the amount of dioxide in the system at expenses of the
carbide phase. At an O2/C ratio of 1.0 the only solid phase appears to be Mo dioxide;
however, larger values promote the appearance of the Mo trioxide phase, which, as
explained before, must be avoided by any means.
Effect of the space velocity
The increase in the space velocity directly affects the concentration of reactants and, thus,
is expected to produce a raise in the rates of reactions 14 and 15. The intensity of this
effect on each reaction is closely related to the way the reaction rates depend on the
concentration of the reactants. Thus, the kinetics of reaction 14 is function of the
concentration of C whereas that of reaction 15 is dependant on the gas O2 concentration.
If the concentration of both reactants were systematically increased in similar ratios, then
134
the reaction rate more affected would be the one responsible for the stability of the
catalytic material.
Figure 6 displays the patterns of the spent samples obtained after activity tests performed
under the same operating conditions except for the space velocity, which was increased
from 0.8 to 6.2 h−1. As seen, at a WHSV=0.8 h−1, the peaks of both phases MoO2 and
Mo2C were detected with XRD, which suggests reactions 14 and 15 took place at similar
rates. However, the concentration of the Mo2C phase appears to decrease at a WHSV=3.6
h−1, to eventually disappear at a WHSV=6.2 h−1. These results suggest the Mo carbide
oxidation process is not only thermodynamically more favorable but also exhibits a faster
kinetics compared to that of the carbothermic reduction of the Mo dioxide phase. This is
in concordance with the fact that the oxidation of the carbide phase requires the diffusion
of a gas phase such as oxygen into a solid phase, Mo2C. Instead, the reduction of Mo
dioxide implies the diffusion of a solid phase, C, into another solid phase, MoO2, which
is a slower process.
The results of our thermodynamic calculations indicate that a decrease in the WSHV (as a
result of increasing the initial amount of MoO2) should lead to an increase in the
equilibrium concentration of Mo dioxide. However, the experimental results of Figure 6
appear to be in disagreement with this trend, which can be attributed to kinetic effects, as
explained earlier.
135
Analysis of the catalytic process
Catalyst performance
The effect of temperature and O2/C ratio on the yields of H2 and CO at equilibrium
conditions is shown in Figure 7. At a fixed O2/C ratio, the production of H2 appears to
reach a maximum value at a certain temperature between 500 to 1000ºC. In addition, the
optimum H2 yield becomes larger as the temperature increases within the temperature
range used in our calculations. On the other hand, at a fixed temperature within the range
500-800ºC , the H2 produced at equilibrium conditions reaches a maximum value at a
O2/C ratio between 0.5 and 1. At higher temperatures the maximum H2 yield is always
attained at an O2/C ratio of 0.5.
The production of CO at equilibrium conditions displays a somewhat different behavior.
Thus, at a fixed O2/C ratio, the CO yield always increases as the temperature becomes
higher. At fixed temperatures, the CO yield appears to exhibit a pattern similar to that
found for the H2 yield. Thus, a fixed temperatures between 500-800ºC, the CO yield
reaches a maximum value at an O2/C ratio within the range 0.5-1.0 whereas at higher
temperatures the maximum CO production is achieved at an O2/C ratio of 0.5.
Hence, the results of our thermodynamic calculations indicate that, at temperatures below
800ºC, the production of syngas can be optimized by tuning the O2/C ratio. However, at
higher temperatures the maximum rate of syngas formation is achievable only at an O2/C
ratio of 0.5.
136
Coking resistance
The production of elementary carbon was investigated using thermodynamic calculations
and the results are shown in Figure 8. As observed, at temperatures higher than 800ºC,
the carbon formation takes place at O2/C ratios lower than 0.5. Instead, at lower
temperatures, O2/C ratios larger than 0.5 are required to prevent carbon formation. The
inset in Figure 8 displays the amount of carbon formed at equilibrium conditions at a
temperature of 850ºC. Our calculations were used to test the resistance of commercial
MoO2 to deactivation by coking. To do so, an activity test was performed under coke
promoting conditions, this is, 850°C, 1 atm, and an O2/C ratio of 0.4. N-dodecane was
used as fuel. For comparison purposes, another activity test was carried out under similar
conditions using a commercial Ni catalyst. After 4 hours of operation the samples were
removed and analyzed using SEM-EDX to determine the amount of coke formed in each
spent sample. The results are shown in Figure 9. As noticed, the commercial Ni catalyst
exhibits a significant number of carbon deposits in the structure which indicates its poor
resistance to coking deactivation and agrees with the continuous decay observed in the
performance. Instead, Mo dioxide shows no significant amount of coke in the structure,
which explains the stable performance obtained during the reforming process. EDX data
support these observations as they indicate a much larger amount of carbon in the nickel
catalyst as compared to that obtained in the Mo dioxide catalyst. The high coking
resistance displayed by Mo dioxide can be explained on the basis of the reverse
Boudouard reaction, which appears to be enhanced by Mo dioxide active sites. This
explanation appears to be in agreement with the experimental data reported by Carrasco-
137
Marin et al. [35], who claim that Mo dioxide promotes the CO2 gasification of activated
carbons.
Conclusions
In light of the results reported in this work, Mo dioxide appears as a promising catalyst
for the partial oxidation of aviation jet fuels, not only in terms of activity but also
stability. Our thermodynamic calculations led us to conclude that, in order to maximize
the production of syngas, the reformer must operate using a O2/C ratio of 0.5, since at this
value the Mo dioxide phase is stable and the formation of H2O and CO2 is minimized.
138
3
4
5 7
2
1
6
-2 x 104
-12x 104
1: MoO3L + gas
2: MoO2+MoO3L+gas
3: Mo+MO2C+MoO2
4: Mo2C+MoO2+gas
5: MoC+Mo2C+gas
6: MoC+gas
7: MoC+C+gas
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9CMo
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
O
0.1
0.2
0.3
0.4
0.5 (O2/C=0.50)
0.6 (O2/C=0.75)
0.7 (O2/C=1.17)
0.8
0.9
Standard
Gibbs Free Energy (J/mol)
Pure MoO2
Figure 1. Ternary Phase diagram Mo-O-C at 850°C and 1 atm
139
-50
0
50
100
150
200
500 550 600 650 700 750 800 850 900 950 1000
Temperature (°C)
ln(K
eq)
2 MoO2 + 3 C = Mo2C + 2 CO2
Mo2C + 3 O2 = 2 MoO2 + CO2
Figure 2. Equilibrium constants
140
20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 50 52 54 56 58 60
2 theta
Inte
ns
ity
(a
.s.)
Mo dioxide
Mo carbide
O2/C=1.0
O2/C=0.9
O2/C=0.6
O2/C=0.5
O2/C=0.7
O2/C=0.8
Figure 3. Diffractograms of spent samples obtained at different O2/C ratios
141
0.00
0.05
0.10
0.15
0.20
0.25
0.30
0.35
0.40
0.45
0.50
0.5 0.6 0.7 0.8 0.9 1.0
O2/C ratio
mo
les
at
eq
uili
bri
um
MoO2
Mo2C
MoC
CO
CO2
Figure 4. Effect of O2/C ratio on MoO2 stability (initial moles of MoO2 = 0.10)
142
20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 50 52 54 56 58 60
2 theta
Inte
ns
ity
Mo dioxide
Mo carbide
WHSV=6.2 h-1
WHSV=3.6 h-1
WHSV=0.8 h-1
Figure 5. Diffractograms of spent samples at different space velocities (O2/C=0.7)
143
0.00
0.05
0.10
0.15
0.20
0.25
0.30
0.35
0.40
0.45
0.50
0.0 0.1 0.2 0.3 0.4
MoO2 initial moles
mo
les
at
eq
uili
bri
um
MoC
MoO2
Mo2C
O2/C=0.7
CO
CO2
Figure 6. Effect of space velocity on MoO2 stability (initial moles of MoO2 = 0.10)
144
500
600
700
800
900
1000
0.5
0.6
0.7
0.8
0.9
1
0.2
0.4
0.6
0.8
1
O2/C ratio
Temperature (°C)
H2
yie
ld
500600
700800
9001000
0.5
0.6
0.7
0.8
0.9
1
0
0.2
0.4
0.6
0.8
1
O2/C ratio
Temperature (°C)
CO
2y
ield
Figure 7. H2 and CO yields at equilibrium conditions
145
500
600
700
800
900
1000 0.2
0.4
0.6
0.8
10
0.2
0.4
0.6
0.8
O2/C ratioTemperature (°C)
C y
ield 0%
10%
20%
30%
40%
50%
60%
70%
0.2 0.4 0.6 0.8 1.0
O2/C ratio
Ca
rbo
n y
ield
Figure 8. Carbon formation (Inset: carbon yield profile at 850°C)
146
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
0.5 0.6 0.7 0.8 0.9 1.0O2/C ratio
H2 Y
ield
(%
)
Mo dioxide
Equilibrium
gas phase
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
0.5 0.6 0.7 0.8 0.9 1.0O2/C ratio
CO
Yie
ld (
%)
Mo dioxide
Equilibrium
gas phase
Figure 9. Hydrogen and CO yields as functions of the O2/C ratio
147
Figure 10. SEM-EDX analysis of spent samples obtained from the coking-resistance test
148
Concentrations (mol %)
O2/C Conversion H2 CO CH4 CO2
0.5 99% 22% 24% 0% 1%
0.6 100% 16% 18% 2% 5%
0.7 100% 12% 15% 2% 8%
0.8 100% 9% 12% 2% 9%
0.9 100% 8% 10% 1% 11%
1.0 100% 6% 7% 1% 12%
Table 1: MoO2 catalytic performance
149
References
[1] Villoria J., Alvarez-Galvan M., Navarro R., Briceño Y., Gordillo Alvarez F., Rosa
F., Fierro J., Catal.Today, 138(2008) 135-140
[2] Moon D., Ryu J., Catal. Letters, 89(2003) 207-212
[3] Lee I., Catal. Today, 136(2008) 258-265
[4] Natelson R., Kurman M., Cernansky P., Miller D., Fuel 87 (2008) 2339-2342
[5] Kahandawala M., DeWitt M., Corporan E., Sidhu S., Energy & Fuels 22 (2008)
3673-3679
[6] Shi L., Bayless D., International Journal of Hydrogen Energy 33 (2008) 1067-1075
[7] Gould B., Chen X., Schwank J., Appl. Catal. A: General 334 (2008) 277-290
[8] O. Marin-Flores, Su Ha, Appl. Catal. A: General, 352 (2008) 124.132
[9] Luo N., Zhao X., Cao F., Xiao T, Fang D., Energy & Fuels 2007, 21, 3505–3512
[10] Dalle Nogare D., Baggio P., Tomasi C., Mutri L., Canu P., Chemical Engineering
Science 62 (2007) 5418 – 5424
[11] Shuo L., Zhang, K, Fang L., Li Y., Energy & Fuels 2008, 22, 1365–1370
[12] Derevich I., Ermolaev V., Krylova A., Pershukov V., Theoretical Foundations of
Chemical Engineering, 2006, Vol. 40, No. 2, pp. 183–189
[13] Kihlborg L., Acta Chem. Scand. 13 (1959) 954-962
[14] Kihlborg L., Acta Chem. Scand. 23 (1969) No 5
[15] Ghosh A., J. Less-Common Metals, 17 (1969) 329-333
[16] Burns R., DeMaria G., Drowart J., Grimley R., J. Chem. Phys., 32 (1960) 1363-
1366
[17] Blackburn P., Hoch M., Johnston H., J. Phys. Chem., 62 (1958) 769-773
150
[18] Perry R., Green D., Perry’s Chemical Engineering’s Handbook, Seventh Edition,
(1997) McGraw-Hill
[19] Berkowitz J., Inghram M., Chupka W., J. Chem. Phys., 26 (1957) 842-846
[20] Norman J., Staley H., J. Chem. Phys., 43 (1965) 3804-3806
[21] Ackermann R., Thorn R., Alexander C., Tetenbaum M., J. Phys. Chem., 64 (1960)
350-355
[22] Kazenas E., Tsvetkov Y., Zhurnal Neorganicheskoi Khimii, 14 (1969) 11-13
[23] Lu J., Hugosson H., Eriksson O, Nordstrom L., Jansson U., Thin Solid Films, 370
(2000) 203-212
[24] Prankatz B., Properties of Elements and Oxides, Bulletin 672 (1982)
[25] Kuo K., Hagg G., Nature, 170 (1952) 245-246
[26] Toth L., Transition Metal Carbides and Nitrides, (1991) Academic Press
[27] Delporte P., Meunier F., Pham-Huu C., Vennegues P., Ledoux M., Guille J., Catal.
Today 23 (1995) 251-267.
[28] Ovari L., Kiss J., Farkas A., Solymosi F., J. Phys. Chem. B 109 (2005) 4638-46445
[29] Hwu H., Zellnet M., Chen J., J. Catal. 229 (2005) 30-44
[30] Liu P., Rodriguez J., J. Phys. Chem. B, 110 (2006) 19418-19425
[31] Li X., Li Y., Chem. Eur. J., 10 (2004) 433-439
[32] Pacheco M., Sira J., Kopasz J., Applied Catalysis A: General 250 (2003) 161–175
[33] Chaudhury S., Mukerjee S., Vaidya V., Venugopal V., J. Alloys Comp. 261 (1997)
105-113
151
CHAPTER VI
NANOPARTICLE MOLYBDENUM DIOXIDE: A HIGHLY ACTIVE CATALYST
FOR PARTIAL OXIDATION OF AVIATION FUELS
Abstract
The More Electric Airplane (MEA) concept may be the most innovative development in
aviation since the Wright “Flyer” and certainly represents the most transformative change
in commercial aviation since the first use of jet engines in 1952. One of the key
requirements for enabling the MEA is a fuel-flexible fuel reformer that operates directly
on logistics fuels such as Jet-A. In this paper, we show that nanoparticle molybdenum
dioxide (MoO2) synthesized directly from a reducing ethylene glycol/water solution can
catalyze the partial oxidation of dodecane (a C-12 hydrocarbon surrogate for Jet-A fuel)
at weight-hourly-space-velocities up to 10 h-1. Even at these very high flow rates the
MoO2 nanoparticle catalyst shows a remarkably high fuel conversion of > 90% with a
hydrogen yield of > 70% and an exceptional coking resistance. Under similar
environments, conventional Ni-based catalysts and commercial low surface-area MoO2
quickly deactivate due to coking. Our results demonstrate that in its nanoparticle form
MoO2 represents a very promising alternative to expensive noble metals for the
reformation of various logistics fuels and is an important step towards realization of the
MEA.
152
Introduction
Growing concerns over global climate change and national energy security are requiring
a reduction in the consumption of fossil fuels 1. These concerns are greatly impacting
future aircraft design and commercial manufacturers such as Boeing Commercial
Airplanes are working on cleaner, quieter, and more fuel-efficient airplanes. A concept
called the More Electric Airplane (MEA) will allow greater fuel efficiency by
substituting hydraulically and pneumatically driven systems by those based on electrical
energy 2. The increased electrical power demand in a MEA can be met by decentralizing
the power-producing units using small individual devices such as fuel cells. Furthermore,
existing commercial aircraft use a low efficiency gas turbine auxiliary power unit (APU)
to provide electrical power for operating navigation systems and various other electronic
devices. By replacing the conventional APU with a solid oxide fuel cell (SOFC) APU,
improvements can be made in providing a means to obtain auxiliary power without
consuming excessive amounts of fuel when the airplane is on the ground or when the
load is increased on the main engines during flight 3. Thus, fuel cells may become the
primary electrical power source with engine-driven generators serving a backup role on
future airplanes.
An important practical requirement for the use of fuel cells on commercial and military
airplanes is that they must operate using kerosene-based aviation fuels (such as Jet-A and
JP-8), which are already on board. The existing approach for fuel cell systems operating
on Jet-A fuel (the standard kerosene-based commercial aviation fuel) requires a fuel
reformation process in which the Jet-A is mostly converted to hydrogen and carbon
153
monoxide 4-6. This syngas mixture is fed into SOFCs where it is electrochemically
converted to H2O and CO2, and produces electrical power. To develop a high
performance fuel reforming system that can operate with Jet-A fuel, a catalyst with the
following attributes is required 7:
• High oxidation activity toward Jet-A fuel
• High resistance to coking
• Stability at high operating temperatures (i.e., higher than 700oC 8)
• High sulfur tolerance (e.g., aviation fuels typically contain 500 ppm of sulfur)
Conventional Ni-based catalysts quickly deactivate under Jet-A fuel reforming
environments due to coke formation and sulfur poisoning 9, 10. Thus, it is of extreme
importance to develop new catalytic materials that can directly process Jet-A fuel.
Molybdenum dioxide (MoO2) is a transition metal oxide that has long been known to be
active for hydrocarbon decomposition 11, 12 and has more recently shown to display high
reforming activity for various long-chain hydrocarbons 13, 14. Marin-Flores and Ha
reported that MoO2 is highly active for reforming isooctane (a gasoline surrogate) via
partial oxidation 15. This process is exothermic (∆H°= −659.9 kJ/mol) and in the presence
of MoO2 proceeds to full conversion at 700°C and 1 atm. The catalytic activity shown by
MoO2 can be explained in terms of the Mars-van Krevelen mechanism 16, 17, which
154
involves the consumption of nucleophilic oxygen ions provided by the oxygen sub-lattice
with the purpose of sustaining the redox cycles taking place on the catalyst surface.
The capability of MoO2 to selectively transfer lattice oxygen into hydrocarbon molecules
can be attributed to the rapid formation of oxygen vacancies in its oxygen sub-lattice 18.
Consequently, MoO2 can catalyze reactions like the carbon dioxide assisted gasification
of activated carbons and chars, which would typically lead to rapid catalyst deactivation
through coking 19. Another interesting, and very relevant, feature of MoO2 is its high
sulfur tolerance. Molybdenum dioxide has been shown to maintain a high reforming
activity for gasoline even in the presence of 500 ppm of model sulfur compounds 17,
which is a typical concentration in most aviation fuels.
Molybdenum dioxide is also an unusual transition metal oxide because of its high
metallic-like electrical conductivity (1.1 x 104 Scm-1 at 300K in bulk samples) 20, which is
associated with its mixed interatomic bonding and a relatively high density of states at the
Fermi level. The existence of free electrons in this region enhances the catalytic activity
of Mo4+ in MoO2, unlike that of Mo6+ in MoO3, where all the valence electrons of the
metal are covalently bonded to neighboring oxygen atoms 13. The metallic conductivity of
MoO2 makes it a material of interest for many applications, including catalysis. Most
recently it has been studied as a possible anode for lithium-ion batteries 20.
Despite its interesting catalytic properties, a very limited number of studies have been
conducted examining the potential of MoO2 as a catalyst for reforming processes 19. Such
studies were carried out using commercial MoO2, with particle sizes in the range of a few
155
micrometers and Brunauer, Emmett, and Teller (BET) surface areas <10 m2/g. In this
paper, we report the catalytic performance of nanoparticle MoO2 for the partial oxidation
of a Jet-A fuel surrogate, n-dodecane (C12H26). We chose n-dodecane for this study
because paraffinic hydrocarbons (with C between 8 and 16) constitute the most abundant
components present in Jet-A fuels. By utilizing nanoparticles we have shown that it is
possible to significantly increase the total reactive surface area and thus achieve
reforming processes with much higher efficiency levels than those of commercial MoO2.
Experimental
Catalyst preparation and characterization
Nanoparticle MoO2 was synthesized by reduction of molybdenum trioxide (MoO3)
powder in a 1:3 volume ratio of ethylene glycol to distilled water 21. The mixture was
combined in a 45 ml Teflon-lined general-purpose vessel (Parr Instrument Company),
which was subsequently sealed and heated to 180°C for 12h. The liquid ratio of 1 to 3
was chosen because it yields pure single phase MoO2, without the need for any post-
synthesis reduction. After cooling, the dark colored MoO2 was filtered and air dried at
100°C. It is relevant to note that the process used to produce the MoO2 nanoparticles is
scalable and large quantities of catalyst could easily be prepared.
BET surface area measurements were carried out using a Coulter SA-3100 automated
characterization machine and previous degassing of the sample under vacuum at 513 K
during 30 minutes. The morphology of the powder samples was examined using both
156
SEM (FEI Sirion operated at 15 kV) and TEM (Philips CM-200 operated at 200 kV). X-
ray diffraction patterns were obtained using the Bragg-Brentano optical configuration in a
Philips diffractometer with Co Kα radiation and Fe filter.
Activity tests
Figure 1 shows a schematic diagram of the experimental setup employed to measure the
activity of the catalysts under partial oxidation conditions. The catalyst samples, in
powder form, supported by a quartz wool plug were placed inside a 12 mm fixed-bed
tubular (quartz) reactor. No catalyst pretreatment was applied prior to the activity tests.
The liquid feed, consisting only of n-dodecane, was fed into the furnace at 500°C, where
it was vaporized and mixed with air, employed as oxygen source. A silicon carbide bed
inside the furnace was used to enhance the mixing of the reactants. A calibrated syringe
pump and a mass flow controller allowed the control of the flow rates of liquid fuel and
air, respectively. The exit stream was cooled down to 5°C to separate water, non-reacted
fuel, and other possible condensable products from the gas product. The composition of
the dry off-gas was monitored using an SRI chromatograph with a thermal conductivity
detector (TCD) and a Shincarbon packed column to determine the concentrations of H2,
CO, CO2, and CH4.
The catalytic performance was analyzed in terms of three parameters:
• H2 yield: the ratio between the hydrogen atoms produced in the form of molecular
hydrogen from the off-gas and the total amount of hydrogen atoms in the fuel stream.
157
• CO yield: the ratio between the carbon atoms produced in the form of the carbon
monoxide and the total amount of carbon atoms in the fuel stream.
• Conversion: the ratio between the total amount of carbon atoms produced in the form
of carbon monoxide, carbon dioxide and methane and the total amount of carbon
atoms in the fuel stream
Results and discussion
Characterization of catalytic materials
Figure 2 shows a montage of scanning electron microscope and transmission electron
microscope images comparing commercially available MoO2 (Alfa Aesar) with
nanoparticle MoO2. The commercial powder consists of irregular shaped particles with a
wide range of sizes. Most of the particles have diameters >200nm, although occasionally
much smaller particles were found. In contrast, the nanoparticle MoO2 has a much more
uniform particle size and shape, consisting of equiaxed particles with average size of
about 20nm. The selected area diffraction patterns (SADPs) confirm the presence of
MoO2 in both sets of powders.
Figure 3 shows X-ray diffraction patterns obtained from both the commercially available
MoO2 and nanoparticle MoO2. In both cases, only the peaks for the distorted rutile
structure of the dioxide phase are present. The two most intense peaks at 2θ = 30.5° and
44.0° are due to overlapping of reflections from the (011) and (110), and (020), (-211),
158
(111) and (200) planes, respectively. Using the Scherrer method the average particle size
was determined from peak broadening to be 23 nm for the synthesized nanoparticle
MoO2. This value is consistent with the TEM observations.
The BET surface area of the nanoparticle MoO2 was determined to be 48 m2/g, which is
about an order of magnitude greater than that of the commercially available material.
Catalytic activity measurement
To investigate the catalytic activity of nanoparticle MoO2 for the partial oxidation of n-
dodecane, tests were conducted at 850°C and 1 atm, with a weight-hourly-space-velocity
(WHSV) of 1.1 h-1 and an O2/C ratio of 0.5. The process can be described by equation 1:
C12H26 + 6 O2 � 12 CO + 13 H2 (1)
The operating temperature was chosen to prevent carbon formation since the reverse
Boudouard reaction is thermodynamically favored if the operating temperature is <700°C
8. The O2/C ratio was set to 0.5 on the basis of our thermodynamic calculations, which
show that ratios <0.5 promote carbon formation in the catalytic material whereas ratios
>0.5 lead to an increase in the concentrations of CO2 and H2O as a result of complete
oxidation of the fuel.
Figure 4 shows the time-evolution of the catalytic performance of nanoparticle MoO2
under the specified operating conditions. The process takes about 1h to reach steady state
conditions at which point the H2 yield, CO yield and conversion become approximately
159
85%, 85% and 80%, respectively. As expected from the stoichiometry of the reaction, an
O2/C ratio of 0.5 leads to a small concentration of CO2 whereas the concentrations of H2
and CO attain values of about 25% and 20%, respectively. The reduced concentration of
CH4 suggests that thermal cracking of the fuel was taking place at low rates.
Figure 5 shows the effect of the WHSV on the catalytic performance of nanoparticle
MoO2. The operating conditions chosen for this test were the same as described above:
850°C, 1 atm, and an O2/C ratio of 0.5. The WHSV range considered was 1−10 h−1. As
noticed, the conversion decreased very slightly from 97% to 91% as the WHSV increased
from 1 to 10 h−1. The H2 yield and the CO yield gradually decreased from 90% to 73%
and from 88% to 72%, respectively whereas the concentrations of H2, CO, CO2 and CH4
appeared not to be significantly affected by operation at very high flow rates, which
suggests that no changes occurred in the reaction mechanism.
Figure 6 is particularly significant because it compares the effect of WHSV, i.e., fuel
flow rate, on the catalytic performance of both nanoparticle MoO2 and commercially
available MoO2. The H2 production rates at equilibrium conditions determined from
thermodynamic calculations have been included as a reference (i.e., the equilibrium value
would represent a theoretical maximum production rate). As seen, the H2 production rates
for both catalysts are initially similar, displaying a linear increase as the WHSV was
increased from 1 to 7 h−1; the nanoparticles give rates slightly closer to the equilibrium
values. The inability of both catalytic materials to produce reaction rates closer to those at
equilibrium conditions can be attributed to two factors. The first one is related to the
160
kinetics of the reaction, this is, the time of contact between the reactants and the catalyst
active sites seems to be insufficient to attain equilibrium. The second one has to do with
the inherent selectivity of the catalytic material, which promotes the formation of other
hydrogen-containing side products at expenses of molecular H2. However, the activity of
the commercial material exhibits a significant drop in the H2 production rate as the
WHSV becomes >7 h-1, while the nanoparticle catalyst shows a continuous increase in H2
production rates.
After reaction at large WHSVs the catalyst samples were reexamined using TEM and
SEM (Figure 7). The SEM micrograph of the spent commercial catalyst (Figure 7a)
indicates significant differences in terms of morphology comparing to its original starting
material (Figure 1a). The TEM images show that the spent particles have become
encapsulated by graphitic carbon (i.e. coking), which has caused deactivation. On the
other hand, the shape and particle size of the nanoparticle MoO2 (Figures 7b and d)
remain practically unaltered by the reforming environment even at the very high flow
rates. TEM images of the MoO2 nanoparticles show the complete absence of the graphitic
carbon, i.e., there is no coking.
Selected area diffraction patterns from the reacted samples are shown in Figures 7(e) and
7(f) for the commercial MoO2 and nanoparticle MoO2 samples, respectively. The SADP
for the spent commercial catalyst shows reflections from both Mo2C and MoO2.
However, the SADP from the nanoparticle catalyst shows reflections that can be indexed
exclusively to Mo2C (this was confirmed by XRD, which showed peaks only for Mo2C).
161
The partial oxidation of liquid hydrocarbons over MoO2 catalysts appears to follow the
Mars-van Krevelen reaction mechanism. According to this mechanism, the hydrocarbon
molecules are first adsorbed on the metal oxide surface. The next stage in the mechanism
involves the oxidation of adsorbed hydrocarbons following an insertion of oxygen atoms
provided by reduction of the metal oxide surface. The reduced metal oxide surface is re-
oxidized and surface oxygen vacancies are replenished by gas phase oxygen. This redox
catalytic cycle is illustrated in Figure 8.
In heterogeneous catalysis, the number of available surface active sites per weight of
catalyst increases as the particle size decreases. Because of the very large surface area to
volume ratio for nanoparticles, the high coking resistance displayed by the nanoparticle
MoO2 at high space velocities can be attributed to the existence of a large number of
active sites available on their surface to catalytically reform the high volume of fuel into
syngas. Given that the SADP and corresponding XRD analysis (not shown) of the spent
nanoparticle MoO2 sample showed only Mo2C, then the reaction mechanism must
involve the participation of only a few layers of surface oxide (i.e., there is a core-shell
structure consisting of a Mo2C core with a very thin MoO2 shell). The presence of
molybdenum carbide in the sample might be thought as responsible for the observed
catalytic activity. However, to the best of our knowledge, molybdenum carbide does not
promote partial oxidation of liquid hydrocarbons and, instead, Mo2C has been found to
catalyze the aromatization of liquid hydrocarbons such as n-octane 22. Additionally, as
part of this work, Mo2C was tested for the partial oxidation of n-dodecane under the same
operating conditions as those for Mo dioxide and only poor activity was detected.
162
During the reforming reaction, the MoO2 nanoparticles are transformed into Mo2C
through a carbothermal process that can be described as follows:
2 MoO2 + 4 C � Mo2C + 4 CO (2)
This reaction is thermodynamically favored (Keq=1.27×106, at 850°C) and the mechanism
of reaction has been previously investigated 23, 24. However, the mechanism of reaction
involves surface MoO2 active sites which still exist even after the whole bulk of the
sample has become carbide. EDX (data not shown) supports this hypothesis since it
indicates a high concentration of oxygen element which must come only from the
surface/subsurface layers given that the bulk has been completely carburized. Thus, the
nanoparticle surface contains enough active sites of the MoO2 phase to sustain the surface
redox catalytic cycle even at high WHSVs. A different behavior was observed for the
commercial low surface area MoO2 particles. At WHSVs <7 h-1, the performance was
similar to that of nanoparticles. However, at WHSVs >7 h-1, the insufficient number of
surface active sites of the MoO2 phase could not sustain the redox catalytic cycle (Figure
8). Thus, the reaction rates of thermal cracking and gas phase oxidation become larger
than that of the Mars-van Krevelen based surface reaction. The thermal cracking
enhances the formation of methane and elementary carbon whereas the gas phase
oxidation leads to the formation of carbon oxides. The carbon produced by the thermal
cracking of the fuel then forms the ring-like structures that encapsulate the catalyst
particles. This is in agreement with the particle morphology observed in the SEM and the
encapsulation of particles by graphitic carbon seen in the TEM. The particle
163
encapsulation caused by the excess carbon prevents the interaction of the hydrocarbon
molecules from the fuel with the active sites located on the catalyst surface and, as a
result, the catalyst performance is expected to decline. In addition, the carbon ring-like
structures may prevent the diffusion of carbon into the bulk of the particles, which
explains why the SADPs from these samples show reflections from MoO2.
Conclusions
In summary, the present study demonstrates that molybdenum dioxide nanoparticles
possess high catalytic activity for the production of syngas for fuel cell applications via
the partial oxidation of a Jet-A fuel surrogate. The coking resistance and high fuel
oxidation activity displayed by MoO2 make this catalyst suitable for the reforming of
various logistics fuels. In addition, the metallic-like electrical conductivity of MoO2
allows it be used directly as a reforming anode in a SOFC by replacing the conventional
nickel-based anode electrode. Successful development of SOFCs using direct fuel
conversion will eliminate the need for an external fuel reforming system, thereby
minimizing the weight and complexity of the overall system.
164
GC
Cooling
Bath
Condenser
Fuel pump
Vaporizer
Reactor
Mass Flow
Controllers
He AirThermo-
couple
SiC bed
Catalyst
TCD detectorShincarbon Column
Figure 1. Experimental arrangement for catalytic activity measurements
165
Figure 2. SEM and TEM images of commercially available MoO2 (a and
c) and nanoparticle MoO2 (b and d). The corresponding selected area
electron diffraction patterns as shown in (e) and (f), respectively. The
index of electron diffraction pattern for nanoparticle MoO2 (Figure 2 (f))
matches to that of pure burke MoO2.
166
24 28 32 36 40 44 48 52 56 60
2 theta
Inte
ns
ity
(a
.s.)
Commercial MoO2
Nanoparticle MoO2
(0 1 1)
(1 1 0)
(0 2 0)
(-2 1 1)
(1 1 1)
(2 0 0)
Figure 3. XRD patterns of MoO2 samples
167
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
0 60 120 180 240 300 360 420 480
Time (min)
Yie
lds
, c
on
ve
rsio
n,
mo
lar
co
nc
en
tra
tio
n
H2
COCO2
CH4
H2 yield
CO yield
Conversion
Figure 4. Catalytic activity of nanoparticle MoO2 for partial oxidation of n-dodecane.
168
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
0 1 2 3 4 5 6 7 8 9 10 11
WHSV (h-1
)
H2
COCO2 CH4
Yie
lds
, c
on
ve
rsio
n,
mo
lar
co
nc
en
tra
tio
n
H2 yield
CO yield
Conversion
Figure 5. Effect of fuel flow rate in terms of WHSV on the catalytic activity of
nanoparticle MoO2.
169
0
40
80
120
160
200
0 1 2 3 4 5 6 7 8 9 10 11
WHSV (h-1
)
H2 r
ate
(m
icro
mo
l/s
-g)
Equilibrium
Nanoparticle MoO2
Commercial MoO2
Figure 6. Hydrogen yield as a function of WHSV.
170
Figure 7. SEM and TEM images of spent samples of commercial MoO2 (a and c) and
nanoparticle MoO2 (b and d). The corresponding SADP patterns are shown in Figures (e)
and (f), where a symbol “o” stands for the index of burke MoO2 and a symbol “c” stands
for the index of burke Mo2C.
171
Figure 8. The catalytic redox cycle based on the Mars-van Krevelen reaction mechanism.
172
References
[1] National Academy of Sciences, What You Need To know About Energy. National
Academic Press: Washington, D.C., 2008.
[2] Adams, C., A380: `More Electric' Aircraft. Avionics October 1, 2001.
[3] Daggett, D. L.; Lowery, N.; Wittmann, J., Fuel Cell APU for Commercial Aircraft.
In The Hydrogen Expo Conference, Hamburg, 2005.
[4] Ibarreta, A. F.; Sung, C. J. International Journal of Hydrogen Energy 2006, 31, (8),
1066-1078.
[5] Roychoudhury, S.; Lyubovsky, M.; Walsh, D.; Chu, D.; Kallio, E. Journal of
Power Sources 2006, 160, (1), 510-513.
[6] Shi, L.; Bayless, D. J. International Journal of Hydrogen Energy 2008, 33, (3),
1067-1075.
[7] Steele, B. C. H. Solid State Ionics 1996, 86-8, 1223-1234.
[8] Zhan, Z. L.; Barnett, S. A. Science 2005, 308, (5723), 844-847.
[9] Matsumura, M.; Hirai, C. Journal of Chemical Engineering of Japan 1998, 31, (5),
734-740.
[10] Gould, B. D.; Chen, X. Y.; Schwank, J. W. Applied Catalysis a-General 2008, 334,
(1-2), 277-290.
[11] Griffith, R. H. Nature 1936, 137, 538.
[12] Griffith, R. H. Nature 1950, 165, 486-487.
[13] Katrib, A.; Leflaive, P.; Hilaire, L.; Maire, G. Catalysis Letters 1996, 38, (1-2), 95-
99.
[14] Katrib, A.; Sobczak, J. W.; Krawczyk, M.; Zommer, L.; Benadda, A.; Jablonski, A.;
Maire, G. Surface and Interface Analysis 2002, 34, (1), 225-229.
[15] Marin-Flores, O. G..; Ha, S. Catalysis Today 2008, 136, (3-4), 235-242.
173
[16] Vannice, M. A. Catalysis Today 2007, 123, (1-4), 18-22.
[17] Marin-Flores, O. G.; Ha, S. Applied Catalysis a-General 2009, 352, (1-2), 124-132.
[18] McCarroll, W. H., Solid state chemistry Encyclopedia of Inorganic Chemistry John
Wiley & Sons: New York, 1994; Vol. 6.
[19] Carrasco-Marin, F.; Riverautrilla, J.; Utrerahidalgo, E.; Morenocastilla, C. Fuel
1991, 70, (1), 13-16.
[20] Shi, Y.; Guo, B.; Corr, S. A.; Shi, Q.; Hu, Y.-S.; Heier, K. R.; Chen, L.; Seshadri,
R.; Stucky, G. D. Nano Letters 2009, doi: 10.1021/nl902423a.
[21] Chen, X. Y.; Zhang, Z. J.; Li, X. X.; Shi, C. W.; Li, X. L. Chemical Physics Letters
2006, 418, (1-3), 105-108.
[22] Szechenyi, A.; Solymosi, F. Applied Catalysis a-General 2006, 306, 149-158.
[23] Chaudhury, S.; Mukerjee, S. K.; Vaidya, V. N.; Venugopal, V. Journal of Alloys
and Compounds 1997, 261, (1-2), 105-113.
[24] Patel, M.; Subrahmanyam, J. Materials Research Bulletin 2008, 43, (8-9), 2036-
2041.
174
CHAPTER VII
STUDY OF THE POISONING EFFECT OF BENZOTHIOPHENE
OVER MOLYBDENUM DIOXIDE DURING THE
PARTIAL OXIDATION OF A JET FUEL SURROGATE
Abstract
The present work is intended to investigate the effect that organic sulfur compounds may
have on the catalytic performance of Mo dioxide during the reforming of aviation jet
fuels. N-dodecane and benzothiophene were used as surrogates jet fuel and sulfur model
compound, respectively. Thermodynamic calculations led us to focus the study on the
poisoning effect of sulfur oxides since the formation of H2S under partial oxidation
conditions appears not to be favorable. The activity tests were performed at 850°C and
indicated that MoO2 displays an important tolerance to deactivation by organic sulfur
compounds even at concentrations as high as 1000 ppmw. XPS analysis indicates the
presence of small concentration of sulfur compounds, which was enough to significantly
reduce the catalytic performance. SEM and TEM images of the spent samples suggest a
close relationship between sulfur poisoning and coking formation.
175
Introduction
Extensive research has been focused on finding alternative power sources that could help
overcome the strong fossil fuel dependency of the economy. Within this context, fuel cell
technology has attracted attention due to not only the high efficiency achieved but also
the possibility of minimizing greenhouse emissions. Particularly in the aircraft industry, a
hybrid system consisting of a conventional internal combustion engine (ICE) for
propulsion and a solid oxide fuel cell (SOFC) powered auxiliary power unit (APU)
appears as an attractive approach to optimize the fuel consumption through the utilization
of an environmentally friendly technology. SOFC systems operating with hydrogen attain
the highest efficiencies although this type of fuel cells can operate with syngas (H2+CO)
as well. H2 lacks a suitable distribution infrastructure; therefore, it is desirable the on-
demand production of hydrogen/syngas using the jet fuel already present on the aircraft.
One of the major challenges in doing so is the high content of sulfur compounds existing
in fossil fuels, the raw material for transportation fuels. Jet fuel is a complex hydrocarbon
mixture characterized by the presence of sulfur, which can achieve concentrations as high
as 3000 ppmw.
Previous research [1] has determined that the two major sulfur compounds in jet fuels are
2,3-dimethylbenzothiophene and 2,3,7,-dimethylbenzothiophene. These organic sulfur
compounds are known to undergo cracking at high temperatures to produce H2S, which
decomposes to release elemental sulfur. The sulfur thus generated may bind to the
catalyst structure forming stable compounds with the transition metals existing in the
catalytic material, leading eventually to sulfur poisoning.
176
Desulphurization is a process widely used to remove most of the sulfur from logistic
fuels; however, small amounts of remaining sulfur can lead to a rapid catalyst
deactivation. Thus, fuel processors for SOFCs are required to generate syngas-rich
reformate for extended periods of time in the presence of sulfur and deliver syngas with
little or no sulfur to the fuel cell stack. For that reason, the development of new reforming
catalysts with high sulfur-tolerance is of key significance for the processing of jet fuels to
produce the syngas required by SOFCs.
Many investigators have tried to obtain formulations for sulfur-tolerant catalysts. Noble
metals supported on ceria are catalytic materials that display high hydrogen yield,
improved stability and sulfur tolerance [2, 3]. These features are attributed to the high
dispersion of the noble metal and the addition of metal oxide additives. The presence of
Pd and Rh in ceria-based catalysts improves the performance in terms of initial hydrogen
yields; however, the catalyst deactivation is evident after 12 hours of operation, when the
concentration of sulfur in the fuel is 50 ppm. Another approach involves the use of
catalysts with perovskite structure, which have been found to be active for the
autothermal reforming of sulfur-containing fuels [4]. The presence of small amounts of
either K or Ru significantly increases not only the activity of these catalysts but also their
stability at sulfur concentrations of 220 ppm and below. More recently, Lakahpatri et al.
[5] synthesized Rh-Ni catalysts supported on alumina and, on the basis of his findings, he
concluded that concentrations of sulfur of 1000 ppmw are able to reduce the performance
by 50% after10 hours on stream. The mechanism of deactivation of these catalysts
177
involves the saturation of Ni sites with sulfur, which enhances the formation of graphitic
carbon.
As seen, the attempts describe above were not able to provide sulfur tolerant catalysts to
handle concentrations of sulfur in jet fuels as high as 3000 ppm. On the other hand, the
catalytic materials currently developed are mainly based on expensive noble metals and
their rates of deactivation are still far from the ones required for practical applications.
Molybdenum dioxide MoO2 has been reported as an active catalyst for the partial
oxidation of gasoline [6], and with a sulfur tolerance of up to 1000 ppm. For that reason,
MoO2 emerges as a novel catalytic material to reform sulfur-containing jet fuels via
partial oxidation.
Experimental
Commercial MoO2 catalyst (99%, metal basis) was purchased from Alfa Aesar. The
catalytic activity tests were performed using a 12 mm fixed-bed tubular (quartz) reactor.
N-dodecane (99+ %) purchased from Alfa Aesar was used as jet A fuel surrogate. A
syringe pump was used to inject the liquid fuel into a pre-heater at 500°C to be vaporized.
The amount of air was regulated using a mass flow controller. The mixing of vaporized
fuel and air was enhanced using a silicon carbide bed in the pre-heater. The operating
conditions in the reactor were 850°C and 1 atm. For each test, 0.5 g of fresh catalyst (no
pretreated) was supported on a quartz wool plug and placed inside the reactor. The off-
gas was cooled down to 5°C to separate water, non-reacted isooctane, and other possible
178
condensable compounds. The dry gas product was then analyzed using an SRI
chromatograph to monitor H2, CO, CO2, and CH4 concentrations.
The bulk structure of the spent samples was analyzed by powder x-ray diffraction (XRD)
on a Philips diffractometer using Co Kα radiation with an iron filter.
The information was analyzed in terms of H2 yield, CO yield and conversion, which are
defined as follows:
in
HC
out
H
n
nyieldH
2612
2
2
26
2•
•
×
×=
in
HC
out
CO
out
CO
out
CH
n
nnnConversion
2612
24
12•
•••
×
++=
in
HC
out
CO
n
nyieldCO
261212•
•
×
=
Thermodynamic Analysis
Figure 1 shows the results of the thermodynamic calculations based on the minimization
of Gibbs free energy for the reaction between air and the model sulfur compound
benzothiophene (BT), at 850°C. As noticed, the amount of oxygen strongly affects the
identity of the sulfur compound resulting from this process. O2/BT molar ratios below
179
10.5 enhance the formation of H2S whereas ratios greater than 10.5 promote the
formation of sulfur oxides. N-dodecane partial oxidation must be performed at O2/C
ratios higher than 0.5 to prevent coke formation. The molar ratio oxygen to
benzothiophene (O2/BT) for a concentration of the sulfur compound equal to 3000 ppmw
becomes larger than 1500 and, therefore, sulfur oxides are expected to be the promoters
of the catalyst deactivation due to sulfur.
The interaction between sulfur oxides and molybdenum dioxide was analyzed combining
thermodynamic calculations with data from the literature. The chemical reaction between
MoO2 and SO2 is not thermodynamically favored; however, the adsorption of SO2 by Mo
dioxide has been reported by many investigators. Furthermore, SO2 may dissociate in
presence of Mo dioxide to form elementary sulfur, which may react with the dioxide to
form Mo disulfide and regenerate sulfur dioxide. This process is thermodynamically
favorable and suggests a dynamic equilibrium between SO2 adsorbed and elementary
sulfur on the surface of Mo oxide.
SO3 seems to play a different role in the deactivation process. Thus, the oxidation of Mo
dioxide by SO3 is thermodynamically feasible at high temperatures to produce SO2 and
MoO3. This reaction becomes fairly inconvenient at high temperatures given that the
formation of MoO3 leads to the loss of catalytic material as the melting point of Mo
trioxide is approximately 800°C. The variation of the equilibrium constants of the
processes mentioned above with the temperature is shown in Figure 2.
180
Results and Discussion
Effect of the concentration of benzothiophene on the catalyst performance
Figure 3 summarizes the catalytic performance of MoO2 obtained during the partial
oxidation of n-dodecane containing different concentrations of sulfur compound. As seen,
concentration of up to 1000 ppmw of BT in the fuel appear to barely affect the catalytic
activity of Mo dioxide producing a drop of about 10% in the hydrogen yield, respect to
that obtained without BT in the fuel after 7 hours of operation. The CO yield displays a
similar behavior although the decline in this parameter was only of approximately 5%,
after the same time on stream. However, the total fuel conversion exhibits a different
trend, remaining unaffected after the same period of time. Higher concentrations of sulfur
compound intensify these patterns. Thus, 3000 ppmw of BT provokes a drop in the
hydrogen yield of approximately 40%, whereas CO yield and conversion reduce by 45%
and 30%, respectively.
The concentrations of CH4 and CO2 appear to increase with both time and concentration
of BT in the fuel, as seen in Figure 4. The presence of BT appears to promote the
formation of CH4 and CO2, which can be attributed to a change in the mechanism of
reaction dominating in the system. Thus, as the concentration of BT increases, the fuel
cracking process seems to prevail over the reforming, producing CH4 and C. The amount
of carbon formed seems to be responsible for the decrease in the total conversion;
however, some of the carbon combines with oxygen to produce CO2.
181
Effect of the presence of benzothiophene on the catalyst structure
An additional test was performed to elucidate the effect of the presence of BT in the fuel
over the catalyst structure. Thus, a fresh catalyst sample was exposed for 2 hours to a
partial oxidation environment using a mixture dodecane-BT with a concentration of 5%
(weight basis) of sulfur compound as fuel. The operating conditions were 850ºC, 1 atm,
and an O2/C ratio of 0.5.
The x-ray diffractogram of the spent sample (Figure 5) shows the presence of only Mo
dioxide and Mo carbide in the bulk structure, which indicates a limited interaction
between the sulfur compound and the catalyst structure. The formation of carbide is
thermodynamically favored at an O2/C ratio of 0.5 and seems not to be affected by the
presence of the sulfur compound in the fuel.
The SEM micrographs of fresh and spent samples shown in Figure 6 indicate a
significance difference in terms of particle size. The reduction in the particle size
observed in the spent sample may be related to the formation of the carbide phase. The
morphology appears to be the same in both samples.
The electron diffraction pattern shown in Figure 7 indicates the presence of two phases in
the spent sample, in agreement with the x-ray diffraction pattern shown in Figure 5. In
addition, electron diffraction results discard the formation of amorphous sulfur-
containing phases that could not be detected using XRD.
182
The XPS analysis of the spent sample (Spent) reveals significant changes in the surface
composition respect to that of a fresh sample, as deduced from data shown in Table 1.
The amount of carbon significantly increases from 14.0 to 73.5%, which suggests the
presence of elementary carbon on the outer structure of the catalyst particles, given that
such amount of carbon can not be attributed only to the formation of a carbide phase. The
total atomic concentration of sulfur on the surface is barely 1.9%, which may be related
to a poor adsorption of the sulfur compound on the catalyst surface. The composition
obtained after successive sputtering steps indicates the presence of sulfur in the
subsurface, which suggests the migration of sulfur into deeper layers. In addition, the
concentration of carbon appears to be directly related to that of sulfur, which may
indicate a close relationship between sulfur adsorption and coking formation.
Conclusions
Molybdenum dioxide shows tolerance to concentrations of benzothiophene in the fuel as
high as 1000 ppmw. XPS analysis indicates the presence of small concentrations of
sulfur-containing compounds on the surface of the spent samples, which appears to
activate the formation of carbon. Thus, the drop in the performance may be attributed to a
combination of sulfur adsorption and coking-formation as a result of the presence of
benzothiophene in the fuel, as suggested by SEM and TEM images.
183
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
5 10 15 20 25
O2/BT ratio
Su
lfu
r c
om
po
un
d y
ield
(%
)
H2S
SO2
SO3
Figure 1. Equilibrium concentrations of sulfur compounds at 850°C
184
0
10
20
30
40
50
60
70
0 200 400 600 800 1000
T (°C)
Ln
(K
eq) MoO2 + 3S = MoS2 + SO2
MoO2 + SO3 = MoO3 + SO2
Figure 2. Equilibrium constants for the reactions between Mo dioxide and sulfur oxides
185
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
0 60 120 180 240 300 360 420
Time (min)
H2 y
ield
no BT
1000 ppmw BT
2000 ppmw BT
3000 ppmw BT
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
0 60 120 180 240 300 360 420
Time (min)
CO
yie
ld
no BT
1000 ppmw BT
2000 ppmw BT
3000 ppmw BT
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
0 60 120 180 240 300 360 420
Time (min)
Co
nv
ers
ion
no BT
1000 ppmw BT
2000 ppmw BT
3000 ppmw BT
Figure 3. Mo dioxide catalytic activity tests at 850°C
186
0%
1%
2%
3%
4%
5%
6%
7%
8%
9%
10%
0 60 120 180 240 300 360 420
Time (min)
CH
4 c
on
ce
ntr
ati
on
no BT
1000 ppmw BT
2000 ppmw BT
3000 ppmw BT
0%
1%
2%
3%
4%
5%
6%
7%
8%
9%
10%
0 60 120 180 240 300 360 420
Time (min)
CO
2 c
on
ce
ntr
ati
on
no BT
1000 ppmw BT
2000 ppmw BT
3000 ppmw BT
Figure 4. Concentrations of CH4 and CO2 as function of time
187
20 25 30 35 40 45 50 55 60
2 theta
Inte
ns
ity
Mo dioxide
Mo carbide
spent
fresh
Figure 5. XRD pattern of spent sample
188
Figure 6. SEM micrographs of fresh (a) and spent (b) samples
189
Figure 7. TEM scan and electron diffraction pattern of spent sample
190
Atomic Concentration (%)
Mo C O N S
Fresh 17.75 14.04 35.45 32.76 0.00
Spent 5.57 73.48 7.83 10.88 1.89
Sputt1 7.37 67.35 9.08 14.25 1.94
Sputt2 7.50 69.72 8.42 12.38 1.99
Sputt3 9.14 63.35 8.21 17.72 1.57
Sputt4 10.23 59.98 7.79 20.95 1.06
Sputt5 13.23 51.75 8.92 25.15 0.95
Table 1: Surface composition of fresh and spent samples
191
References
[1] Lee, I. and Ubanyionwu H., Fuel 87 (2008) 312.
[2] Azad, A.; Duran, M.; McCoy, A.; and Abraham, M., Appl. Catal. A-Gen 332
(2007) 225.
[3] McCoy, A.; Duran, M.; Azad, A.; Chattopadhyay, S; and Abraham, M., Energy &
Fuels 21 (2007) 3513.
[4] Dinka, P.; Mukasyan, A., J. Power Sources 167 (2007) 472.
[5] Lakhapatri, S. and Abraham, M., Appl. Catal. A-Gen 364 (2009) 113.
[6] Marin Flores, O. and Ha, S., Appl. Catal. Gen-A 352 (2009) 124.
192
FINAL REMARKS
The present work led us to believe that molybdenum compounds display very high
activity for the reforming of logistic fuels. Mo carbide (Mo2C) for steam reforming of
gasoline, and Mo dioxide (MoO2) for the partial oxidation of jet fuel, appear to be very
promising catalytic materials with interesting properties such as coking resistance and
sulfur tolerance. These features are of major relevance when dealing with liquid
hydrocarbons derived from fossil fuels. However, the reforming activity of these
compounds was not tested under autothermal conditions, which has been found to be the
most efficient way to operate a catalytic reformer.
The Mars-van Krevelen mechanism of reaction proposed to explain the catalytic activity
of Mo dioxide was able to provide an interpretation for our experimental observations.
However, a deeper study is required to confirm the actual occurrence of such a
mechanism.
Finally, the combination of reforming activity and electrical conductivity exhibited by
these two compounds led us to believe in the possibility of using Mo compounds for the
fabrication of direct solid oxide fuel cells to produce electrical power directly from the
logistic fuel. The fact that Mo carbide has been used in PEM fuel cells in the past and the
preliminary data obtained with Mo dioxide on this matter appears to support this belief.
This possibility may generate a wide range of applications for these compounds.