catalyticDehydrogenationLightAlkanesonMetals PDF

Review
pubs.acs.org/CR
Catalytic Dehydrogenation of Light Alkanes on Metals and Metal

Oxides
Jesper J. H. B. Sattler, Javier Ruiz-Martinez, Eduardo Santillan-Jimenez, and Bert M. Weckhuysen*
Inorganic Chemistry and Catalysis, Debye Institute of Nanomaterials, Utrecht University, Universiteitsweg 99, 3584 CG Utrecht, The
Netherlands
1. INTRODUCTION
1.1. Setting the Scene
Because of their extensive use as chemical building blocks, light

olens, such as propylene (propene) and ethylene (ethene), are
among the most important types of compounds in the chemical
industry. Light olens are the feedstock employed in the
production of a vast array of chemicals, including polymers
(e.g., polyethylene and polypropylene), oxygenates (e.g.,
ethylene glycol, acetaldehyde, acetone, and propylene oxide),
and important chemical intermediates (e.g., ethylbenzene and
propionaldehyde).14 The demand of these building blocks has
increased steadily over the last years, all major markets
registering a positive growth in demand and production in
recent years with the exception of the Eurozone, due to the
eects of the recent economic downturn.5 In Figure 1, the uses
as well as the recent and predicted future growth of the olen
supply and demand are shown.6 Steam cracking and uid
catalytic cracking (FCC) of naphtha, light diesel, and other oil
byproducts are the most common methods for obtaining these
light olens. For example, in 2007 less than 3% of propylene
was produced by on-purpose techniques, while FCC and steam
cracking accounted for the bulk of the propylene produced.7
A number of factors including the high-energy demands of
these processes, their low selectivity toward the production of
particular olens, dwindling petroleum reserves, and rising oil
prices are driving the petrochemical industry to search for a
more economical feedstock and more ecient conversion
technologies. In recent years, hydraulic fracturing or fracking
technologies have improved to the point where large volumes
of shale gas can be extracted in a cost-eective manner. In fact,
the United States already obtains one-quarter of their natural
gas production from shale gas deposits, and this is expected to
rise even further in the coming years due to the emphasis that is
being given to the attainment of energy independence.8,9 This
increased supply of natural gas has resulted in a drop in gas
costs of about 75% relative to 2005 prices, making this shale gas
very attractive both as an energy source and as feedstock for the
production of transportation fuels or chemicals. This includes
light olens, which can be obtained by rst converting natural
gas into synthesis gas and by subsequently converting the latter
either directly through the FischerTropsch-to-olens (FTO)
process or indirectly through the methanol-to-olens (MTO)
route with an intermediate methanol synthesis step.10,11
As shown in Table 1, in addition to methane, shale gas
deposits contain considerable amounts of natural gas liquids
CONTENTS
1. Introduction
1.1. Setting the Scene
1.2. Scope of This Review
2. Commercial Dehydrogenation Processes
2.1. Caton Process
2.2. Oleex Process
2.3. Other Patented Alkane Dehydrogenation
Processes
3. Thermodynamics of Nonoxidative Alkane Dehydrogenation
4. Overview of Nonoxidative Dehydrogenation
Catalyst Materials
4.1. Platinum-Based Catalysts
4.1.1. Nature of the Active Sites
4.1.2. Catalyst Deactivation
4.1.3. Role of the Support
4.1.4. Role of the Promoters
4.2. Chromium Oxide-Based Catalysts
4.2.1. Nature of the Active Sites
4.2.2. Catalyst Deactivation
4.2.3. Role of the Support
4.2.4. Role of the Promoters
4.3. Vanadium Oxide-Based Catalysts
4.4. Molybdenum Oxide-Based Catalysts
4.5. Gallium Oxide-Based Catalysts
4.6. Carbon-Based Catalysts
4.7. Other Formulations
5. Comparing Dierent Catalysts
5.1. In Terms of Reaction Mechanism and
Deactivation Pathway
5.2. In Terms of Propane and Isobutane Dehydrogenation Performance
6. Conclusions and Emerging Areas of Research
Author Information
Corresponding Author
Notes
Biographies
Acknowledgments
References
10613
10613
10614
10615
10615
10615
10617
10618
10620
10620
10620
10621
10621
10622
10625
10626
10627
10629
10630
10630
10633
10634
10637
10637
10640
10640
10644
10646
10647
10647
10647
10647
10648
10648
Received: May 4, 2014

Published: August 27, 2014
2014 American Chemical Society
10613
dx.doi.org/10.1021/cr5002436 | Chem. Rev. 2014, 114, 1061310653
Chemical Reviews
Review
Figure 1. Consumption of ethylene, propylene, and butylenes in 2004, 2010, and 2016 (predicted). In addition, the products made by these olens
are shown in the graph to the right. These (predicted) numbers originate from ref. 6.
worldwide, and many of these installations are already under

construction.1821
Notably, there is a similarly advantageous market situation
for the dehydrogenation of butanes, because the same
conditions apply. Increased demand, especially from the
automotive industry, combined with a decrease in supply due
to the high cost of naphtha cracking, has resulted in sharp
increases in butadiene prices. This in turn has sparked
investment in new butane dehydrogenation (BDH) installations in the United States, China, and Japan.2225
Table 1. Gas Composition of Six Dierent Shale Gas

Deposits in the United Statesa
shale deposit
Barnett Shale (avg of
four wells)
Marcellus Shale (avg
of four wells)
Fayetteville Shale
(one well)
New Albany Shale
(avg of four wells,
N2 not reported)
Antrim Shale (avg of
four wells)
Haynesville Shale
(one well)
methane
(%)
ethane
(%)
propane
(%)
carbon
oxides
(%)
nitrogen
(%)
86.8
6.7
2.0
1.7
2.9
85.2
11.3
2.9
0.4
0.3
97.3
1.0
0.0
1.0
0.7
89.9
1.1
1.1
7.9
62.0
4.2
1.1
3.8
29.0
95.0
0.1
4.8
0.1
1.2. Scope of This Review
Because dehydrogenation for the production of light olens has

become extremely relevant in recent times, we aim to provide
the reader with a complete overview of the materials used to
catalyze this reaction. First, we will introduce the subject by
considering the industrial processes that have been patented
and applied and the thermodynamics of the dehydrogenation
reaction. Next, the Pt- and CrOx-based catalysts that are
currently used in commercial processes will be examined in
detail. Additionally, other metal oxides, such as GaOx, VOx,
MoOx, and InOx, which have also proven to be promising
dehydrogenation catalysts will be surveyed. For each catalytic
material, relevant factors, such as the specic nature of the
active sites, as well as the eect of support, promoters, and
reaction feed on catalyst performance and lifetime, will be
discussed. Finally, we will compare dierent catalysts both in
terms of the reaction mechanism and deactivation pathways
and in terms of catalytic performance. To the best of our
knowledge, no attempt has been made to compare such a wide
range of dehydrogenation catalysts in a single review article,
although other reviews have been published in recent years
dealing with the dehydrogenation of light parans. For
example, Bhasin et al. wrote an extensive review on the
industrial application of dehydrogenation technologies in 2001,
which includes a comparative discussion of the state of the art
of oxidative dehydrogenation technologies.26 Two more review
articles by Vora and Bricker (UOP) were published in 2012. In
the rst review article, Vora gives a general overview of the
dehydrogenation of light olens and ethylbenzene on Pt-based
catalysts, discussing important principles for catalyst and
reactor design.27 In the second review paper, Bricker addresses
catalyst design in more detail, emphasizing mass transfer and
the use of selective hydrogen combustion to increase olen
yield.28 Other reviews focusing on a specic family of
dehydrogenation catalysts have also been published. For
example, Weckhuysen and Schoonheydt reviewed the dehydrogenation of alkanes on supported chromium oxide catalysts
in 1999.29 Dehydrogenation of alkanes on vanadium and
chromium oxides is also discussed by Jackson et al. in a book on
metal oxide catalysis published in 2009.30 The chemistry of
Larger hydrocarbons, hydrogen, and oxygen are present in trace

amounts and are not given.12
(NGL), such as ethane and propane, which are easily separated

from the natural gas.13 Consequently, the recent boom in shale
gas production has turned the United States into the lowest
cost chemical producer outside the Middle East, as illustrated
by the 126 investment projects totaling 84 billion dollars made
up to September 2013 that have been tallied by the American
Chemistry Council.14
The high availability of these relatively cheap natural gas
liquids, specically ethane, has resulted in a shift from oil-based
naphtha to shale-based ethane for steam cracking installations
in the United States, with ethane crackers being constructed
and naphtha crackers being either dismantled or converted.
Relative to naphtha cracking, steam cracking of ethane
produces a negligible amount of olens other than ethylene,
which is why the supply of propylene has dropped and its price
has risen sharply, creating opportunities for on-purpose
catalytic technologies, such as the catalytic dehydrogenation
of light parans into the corresponding olens.15 Indeed, the
protability of propane dehydrogenation is determined to a
large extent by the price dierence between propane and
propylene, which makes the current market conditions very
favorable. An additional advantage of dehydrogenation
technologies is that dehydrogenation is an on-purpose
technique, which yields exclusively a particular olen instead
of a mixture of products. In fact, industrial dehydrogenation
processes are currently optimized in such a way that they can
produce olens of polymer-quality purity. As of this writing, ca.
5 million tons of propylene is produced annually by propane
dehydrogenation (PDH).16,17 However, this number is
expected to increase signicantly in the upcoming years as
dozens of new PDH installations have been announced
10614
Chemical Reviews
Review
vanadium oxide for oxidative dehydrogenation has been further

discussed in a review by Albonetti et al.31 Another review by
Cavani et al. discusses the oxidative dehydrogenation of ethane
and propane on supported vanadium and molybdenum oxides.
The authors conclude that, although the oxidative dehydrogenation of ethane has potential, the oxidative dehydrogenation of
higher parans is still far from being commercially applied.32
The nonoxidative dehydrogenation of light parans on both
supported and bulk transition metal oxides has also been
discussed by Coperet in a review focused on CH bond
activation.33 Also of note are three reviews by Krylov, Wang,
and Ansari, where the use of CO2 as a mild oxidant for the
dehydrogenation of light alkanes is discussed.3436 Finally, in a
chapter by Caspary et al. in the most recent edition of the
Handbook of Heterogeneous Catalysis, the dehydrogenation of
alkanes is discussed broadly, albeit emphasis is placed on CrOx
and Pt catalysts, on the kinetics and mechanism of the
dehydrogenation reaction, and on reactor technology.37
2. COMMERCIAL DEHYDROGENATION PROCESSES

To date, ve main industrial processes for the dehydrogenation
of light alkanes have been patented, two of which are currently
being used to produce light olens.3841 As of April 2012, 5
million tons of propylene was being produced by the 14
propane dehydrogenation plants around the globe.16,17 The
global demand of propylene was 103 million tons that year and
is expected to increase by 45% on a yearly basis.42,43 Literally
dozens of new installations have been planned or are already
under construction, amounting to an additional 14 million tons
of propylene to be produced annually by 2018.16,17,44
Additionally, six isobutane dehydrogenation installations are
currently in operation, with three other plants planned.43 The
majority of these new facilities will be built in China or the
United States and will use either the Caton (Lummus) or the
Oleex (UOP) technologies. A compilation of all existing and
planned propane dehydrogenation installations is shown in
Figure 2. Furthermore, in Table 2, a comparison is made
between the two most important processes, Caton and
Oleex. The STAR (UHDE) process will be applied for the
rst time in two installations in the Middle East. The other two
processes, uidized bed dehydrogenation (FBD) process
(Snamprogetti and Yarsintez) and Linde-BASF PDH, have
not yet been commercially applied.
Figure 2. World map showing an overview of propane dehydrogenation installations currently in operation (lled shapes) and
installations that have been announced and are expected to start up
before 2018 (unlled shapes). The numbers represent the maximum
production of each plant (1000 tons per year). The locations of the
facilities are approximations.16,17,44,45 For some of the newly
announced PDH installations (), the industrial process to be used
remains to be announced.
reaction products.46 A schematic of a Caton dehydrogenation

installation is shown in Figure 3. The duration of the
dehydrogenation step depends on the heat content of the
catalyst bed, which decreases rapidly due to the endothermic
nature of the reaction. Part of the heat required for the reaction
is introduced to the reactors by preheating the reaction feed,
additional heat being provided by adjacent reactors that are
regenerating the coked catalysts. Moreover, fuel gas is added to
the reactor during the regeneration step to generate
supplementary heat,47 and inert material is added to the
catalyst bed to increase its heat storage capacity.48 The catalyst
remains in use for 23 years, and the progressive loss in activity
with increasing time-on-stream is counteracted by gradual
increases in temperature to aord a constant dehydrogenation
activity throughout the entire catalyst life span.29,49
2.1. Caton Process
2.2. Oleex Process
The Caton process, by CB&I Lummus, is based on the

Houdry Catadiene process, which originally was exclusively
used for the dehydrogenation of isobutane to isobutene. In
turn, isobutene was employed for the production of methyl
tertiary butyl ether (MTBE), a fuel additive used to raise the
octane number of gasoline. For environmental reasons, the use
of MTBE has declined in recent years, causing a shift in the use
of Caton installations to alternative purposes, such as propane
dehydrogenation. A Caton installation generally consists of 5
8 parallel adiabatic xed bed reactors containing a chromia
alumina catalyst. The reaction is run at temperatures of
approximately 575 C and pressures between 0.2 and 0.5 bar.26
Each reactor alternates between dehydrogenation, regeneration,
and purge steps, each lasting a few minutes (1530 min being
needed for one full cycle). Each individual reactor is made to
run continuously so there are always some units performing
dehydrogenation reactions, while other reactors are being
regenerated or purged, which results in a constant ow of
The Oleex process by UOP uses a very dierent reactor

design comprising uidized bed reactors, a catalyst regeneration
unit, and a product recovery section. A schematic representation is shown in Figure 4.50 The reaction is run using a PtSnbased catalyst at pressures between 1 and 3 bar and
temperatures ranging from 525 to 705 C. Three or four
adiabatic radial ow reactors containing the catalyst are
connected in series with preheaters in between. These gas
ow preheaters represent the main source of heat of the reactor
system. The catalyst ows through the system, the last reactor
being connected to a continuous catalyst regeneration (CCR)
unit that regenerates the catalyst by burning of any carbon
deposits and redisperses the Pt on the support material by
means of treating the catalyst with a chlorineair mixture. The
entire system is designed to operate continuously to obtain an
uninterrupted stream of reaction products. The regenerated
catalyst is then reintroduced to the rst reactor, completing an
10615
Chemical Reviews
Review
Table 2. Specics of the Caton and Oleex Commercial Dehydrogenation Processes43,44,46,5055

Caton
Oleex
catalyst
formulation
process
license
holder
used for
operating
conditions
reactor type
1820 wt % CrOx on an alumina support, promoted by 12 wt %

Na or K
CB&I Lummus
<1 wt % Pt and 12 wt % Sn on an alumina support, promoted by 01 wt %

Na or K
UOP (Honeywell)
dehydrogenation of propane, isobutane and isopentanes

575 C at 0.20.5 bar
dehydrogenation of propane and isobutane

525705 C at 13 bar
typically multiple parallel xed bed reactors
main source
of heat
supply
reaction
program
deactivation
heating of the catalyst material during regeneration
a uidized bed reactor consisting of three separate parts: several reactors in

series, a product recovery section, and catalyst regeneration section
preheating of the starting and intermediate gases before each reactor
regeneration
catalyst
lifetime
typically 12 min of dehydrogenation, 3 min of purge, 12 min of

regeneration, and 3 min of purge
deposition of coke and thermal sintering of the alumina support
leading to a loss of surface area
combustion of the coke deposits
23 years before the reactor is shut down and the catalyst is
replaced; catalyst lifetime is prolonged by increasing the
temperature to maintain high propylene yield at the cost of lower
selectivity
the separate parts of the reactor run continuously as the uidized catalyst bed
moves from one section to the next
deposition of coke and sintering of the Pt nanoparticles; in addition, catalyst
particle attrition is an issue due to the uidized bed
combustion of the coke deposits and chlorine addition to assist with Pt
redispersion
13 years, new catalyst being continuously added to the reactor system and
dusts formed due to catalyst attrition being continually removed
Figure 3. Schematic representation of a Caton dehydrogenation unit. Eight parallel reactors alternate between dehydrogenation (DH), purge, and
regeneration (regen). Olens produced are separated from side products that are subsequently mixed with oxygen and combusted to heat the
reactors during the regeneration step.
Figure 4. Schematic representation of an Oleex dehydrogenation unit. A simplied version of the gas separation unit used to separate parans,
cracking products, and hydrogen from olens is shown.
diolens and acetylenes. Polymer-quality propylene is obtained

after owing the hydrocarbons through a de-ethanizer and a
propanepropylene splitter. More detailed information on the
Caton and Oleex processes is presented in Table 2.
entire cycle every 510 days. Product recovery is done by

cooling, compressing, and drying the reactor euent, hydrogen
being cryogenically separated from the hydrocarbons. The
latter are led through a selective hydrogenation unit to remove
10616
Chemical Reviews
Review
Figure 5. Schematic representation of a Steam Active Reforming (STAR) dehydrogenation unit. The system comprises two xed bed reactors, one
located in an oven and an oxyreactor where hydrogen is combusted. A simplied version of the gas separation unit used to separate the olens from
parans, cracking products, and residual hydrogen is shown.
Figure 6. Schematic representation of the Fluidized Bed Dehydrogenation (FBD) reactor system. A simplied version of the gas separation section is
shown.
2.3. Other Patented Alkane Dehydrogenation Processes
Most of the heat is provided at the top of the reactor, where the
dehydrogenation activity is highest. The gas mixture exiting the
rst reactor is cooled prior to being introduced into the second
so-called oxyreactor, where an oxygensteam mixture is used to
selectively combust part of the hydrogen formed, shifting the
equilibrium toward higher olen yields. In addition, the
combustion of hydrogen provides the heat required for the
additional conversion of propane. The olens in the resulting
gas mixture are separated from any residual hydrogen and other
side products, which are combusted to heat the dehydrogenation reactor oven. A schematic representation of this reactor
setup is shown in Figure 5. Currently, two dehydrogenation
installations using the Uhde STAR process are being
constructed in Egypt and Iran.
The Steam Active Reforming (STAR) process developed by

Uhde utilizes a very dierent reactor setup. The reaction is run
at 69 bar and at temperatures between 500 and 600 C, steam
being added to the feed to reduce paran partial pressure and
coke formation. Because the catalyst has to be stable in the
presence of steam, PtSn supported on a (basic) zinc
aluminate is used with calcium/magnesiumaluminate as a
binder. Very little coke is deposited on the catalyst surface, and
thus dehydrogenation cycles may be as long as 7 h before
regeneration becomes necessary.
A xed bed reactor system is used, consisting of two separate
reactors placed in series.37,56 The rst is a top-red tubular
reformer-type reactor, which is externally heated by an oven.
10617
Chemical Reviews
Review
Figure 7. Schematic representation of the Linde-BASF PDH process. A simplied version of the gas separation section is shown.
Figure 8. Equilibrium conversion of C2C4 parans to olens as a function of temperature at 1 bar (left) and pressure dependence of the
dehydrogenation of propane as a function of temperature (right).
various similarities with the STAR process, as the reaction feed

is diluted with steam and an externally heated xed bed reactor
is used. Three of these reactors operate in parallel, with two in
dehydrogenation and one in regeneration mode, ensuring a
continuous ow of reaction products. A mixture of air and
steam is used to regenerate the catalyst, and the reactor is
purged before and after the regeneration step.37
Figure 6 shows a schematic representation of the Fluidized

Bed Dehydrogenation (FBD) process, licensed by Snamprogetti and Yarsintez, in which a uid catalytic cracking-like
reactor system is used for the dehydrogenation of propane. An
alkane is owed through a staged uidized bed at pressures
between 1.1 and 1.5 bar and temperatures between 550 and
600 C. A CrOx/Al2O3 catalyst, promoted with an alkali metal,
is used. The catalyst deactivates over time and is therefore
continuously transported to a regenerator connected to the
reactor to combust the carbon deposits. The heat required for
the dehydrogenation reaction is provided by heating the
catalyst material to temperatures above 700 C in the
regenerator. A fuel gas is routinely added during regeneration
as the combustion of the carbon deposits alone does not
provide enough heat to achieve these temperatures. Upon its
return to the reactor, the catalyst cools again to temperatures
below 560 C. Finally, a system comprising a jet scrubber and
cyclones in the head of the reactor is used to remove any dust
formed by catalyst attrition.48
The Linde-BASF PDH process, of which a schematic
representation is shown in Figure 7, uses a PtSn catalyst
supported on ZrO2 for dehydrogenating light parans
isothermally at a temperature of 590 C. The process shows
3. THERMODYNAMICS OF NONOXIDATIVE ALKANE

DEHYDROGENATION
From the chemical perspective, dehydrogenation is a one-step
reaction through which light parans can be converted into the
corresponding olens and hydrogen, as illustrated below for the
dehydrogenation of propane:
0
C3H8 C3H6 + H 2 (H298
= 124.3 kJ mol1)
However, the reaction is thermodynamically limited and

highly endothermic, which, according to Le Chateliers
principle, implies that higher reaction temperature and/or
lower paran partial pressure are needed to achieve high
conversions. Indeed, temperatures of 550750 C are typically
required in the dehydrogenation of C2C4 parans to obtain
alkane conversions 50% at 1 bar (see Figure 8). Moreover,
10618
Chemical Reviews
Review
Scheme 1. Examples of the Side Reactions That May Occur When 1-Butene Is Exposed to a Pt/Al2O3 Catalyst: (1)
Hydrogenolysis, (2) Thermal Cracking, (3) Catalytic Cracking, and (4) Isomerization
during the dehydrogenation of isobutane to isobutene, the

former may transform to 1- or 2-butene, which may
dehydrogenate further to 13-butadiene. The rearrangement
can take place via a carbocation intermediate formed on a
Brnsted acid site, or through an adsorbed species on the active
phase of the dehydrogenation catalyst.6264 By increasing the
temperature, the rates of both CH and CC cleavage
reactions are increased, and therefore temperature and pressure
must be carefully controlled to achieve an optimal olen yield.
In short, the balance between dehydrogenation and these side
reactions is very complex.
Fortunately, these side reactions are catalyzed by specic
sites that are not required for dehydrogenation, which means
that high olen yields can be obtained by careful catalyst design.
For instance, cracking reactions are eciently catalyzed by
Brnsted acid sites, while large Pt ensembles also display
activity for hydrogenolysis, cracking, and isomerization
reactions. Scheme 1 shows an overview of the side reactions
that may be catalyzed by Brnsted acid and Pt sites on a typical
Pt-based dehydrogenation catalyst.26
An additional problem is that the high temperatures required
to obtain high olen yields are also optimal conditions for the
formation of coke. Consequently, the performance of
dehydrogenation catalysts progressively deteriorates with
time-on-stream, making it necessary to regenerate the catalysts
frequently to preserve sucient activity. This is particularly
important when the process operates at high alkane
conversions, due to the fact that polymerized olens are
believed to be coke precursors, as will be discussed in more
detail in section 5.1. Hence, during the selection of the
operational conditions in any industrial dehydrogenation
given that the enthalpy required to dehydrogenate alkanes

decreases as the chain becomes longer (H0298 = 137 kJ mol1
for ethane, 124.3 kJ mol1 for propane, and 117.6 kJ mol1 for
isobutane dehydrogenation), a considerably larger amount of
energy is required to dehydrogenate light parans on a mass
basis. The pressure dependence is also shown in Figure 8, lower
pressures clearly resulting in higher olen yields at a given
temperature.
The CH bonds in parans and olens are considerably
more reactive than CC bonds, meaning that catalysts that
favor CH over CC bond cleavage are required to avoid side
reactions. An additional complicating factor is that olens are
considerably more reactive than their paranic counterparts,
which can further lead to unwanted side and secondary
reactions. Three types of side reactions can occur, hydrogenolysis, cracking, and isomerization. In hydrogenolysis, the
addition of hydrogen into a CC bond within a paran results
in the formation of two smaller alkanes. Because hydrogen
adatoms and/or CxHy adspecies are required for alkane
hydrogenolysis to occur,57,58 this reaction is believed to be
catalyzed by Pt sites. Cracking also results in the cleavage of a
paran to form two smaller hydrocarbons, although in this
instance no hydrogen is required. Thermal cracking hardly
occurs at dehydrogenation reaction conditions, as it requires
temperatures of 450750 C and pressures of 570 bar. This
reaction proceeds via the formation of radical intermediates,
which rearrange into an alkane and an alkene. Catalytic cracking
is dierent in the sense that it requires a catalyst with Brnsted
and Lewis acidity and proceeds via the formation of a
carbocation intermediate, albeit it also results in the formation
of an alkene and an alkane.5961 Isomerization is the
rearrangement of atoms within a molecule. For example,
10619
Chemical Reviews
Review
activation of paranic CH bonds and low activity to CC

cleavage. The rst industrial alkane dehydrogenation technology, the Pacol process (UOP), was rst commercialized in
1968. On the basis of an alumina-supported platinum catalyst,
the process greatly contributed to the widespread use of linear
olens in the detergent industry. In the early 1970s, UOP
developed the Oleex process, which was initially applied for
the production of butadiene and later extended to propylene.
4.1.1. Nature of the Active Sites. Noble metals are active
for dehydrogenation in the metallic state, and in some cases a
reduction step is necessary prior to reaction. The dehydrogenation reaction of light alkanes is insensitive to the structure of
the platinum particles; that is, such reactions are independent of
the platinum particle size or crystallographic plane exposed. As
only the amount of active sites is relevant, small particles are
preferred. However, undesired side reactions that occur during
alkane dehydrogenation, such as hydrogenolysis, isomerization,
and coke formation, are sensitive to the structure of the
platinum particles, although there is no general consensus
regarding the eect of dispersion on the rate of these reactions.
Some studies stress the markedly inverse relationship between
particle size and the rate of hydrogenolysis (see Figure 9), the
process, a compromise must be struck between olen selectivity

and paran conversion.
Notably, the reaction equilibrium can be shifted to higher
paran conversions by reducing the partial pressure of the
paran or by withdrawing the hydrogen produced. As was
shown in the previous section, two strategies have been
adopted in industry to lower paran partial pressure, reducing
the pressure below atmospheric (Caton) or diluting the
paran feed with steam (STAR process).26,30 Hydrogen can be
removed from the reaction medium by selective hydrogen
combustion (SHC), which involves the addition of an oxidant
to the reaction feed to burn o the hydrogen formed during
dehydrogenation. Relatively mild oxidants, for example, CO2,
are preferred, as the use of strong oxidants, like O2, risks also
combusting the hydrocarbon reactants and products. Through
this approach, a considerable amount of heat is generated,
which promotes the endothermic dehydrogenation reaction.
Another method deals with the addition of a solid oxygen
carrier (such as ceria) that serves as a source of oxygen for
oxidizing the hydrogen formed. The reduced ceria is reoxidized
during the regeneration step.65 Another strategy is to use a
membrane permeable to hydrogen to withdraw the latter from
the system, thereby shifting the equilibrium toward higher
olen yields. Nevertheless, as the thermal stability of these
membranes is limited, this approach is more suited for the
dehydrogenation of isobutene.6668
The oxidative dehydrogenation (ODH) of parans is related
to SHC by the fact that an oxidant is also added to the feed,
although in ODH the latter is done to directly oxidize the
paran to the respective olen. ODH has some advantages that
make it a very attractive alternative to nonoxidative
dehydrogenation, including the facts that the reaction is
exothermic (propane ODH: H0298 = 117 kJ mol1) and
not equilibrium limited. However, a good control of side
reactions, such as cracking and the excessive oxidation of
reaction products, is very challenging.
4. OVERVIEW OF NONOXIDATIVE
DEHYDROGENATION CATALYST MATERIALS
As mentioned above, two main types of formulations are
typically used for the nonoxidative dehydrogenation of light
olens: noble metal-based and metal oxide-based catalysts.
Although Pt represents the only precious metal that has been
extensively studied, various metal oxides have been successfully
tested, CrOx being the most prominent example due to its
industrial use. Furthermore, promising results have been
obtained using gallium, indium, vanadium, zinc, and molybdenum oxides. These materials are discussed in detail hereinafter.
Figure 9. Structure sensitivity of ethane hydrogenolysis on 1% Pt/

SBA-15 with Pt particle size ranging from 1.7 to 7.1 nm. Reprinted
with permission from ref. 69, copyright 2005, American Chemical
Society.
4.1. Platinum-Based Catalysts
latter reaction being one of the main causes of coke

formation.6972 Contrarily, others believe that for hydrogenolysis to take place, carbon deposits need to be in the
vicinity of adsorbed hydrogen, which is more likely to occur on
large Pt ensembles than on small particles.37,73,74
DFT calculations have been utilized to calculate the energy
barrier of propane dehydrogenation on at (111) and stepped
(211) Pt crystal planes by Yang et al.72 Indeed, it was found
that step sites are more reactive, as the energy barrier to form
propylene is much lower on step sites (2434 kJ mol1) than
on the at surface (6372 kJ mol1). Nevertheless, propane
adsorbs more strongly on the latter, forming propylidyne
intermediates believed to serve as precursors for both
hydrogenolysis and coke formation.72,75
In the 1960s, a new technology was developed for the synthesis

of the long-chain linear olens used in the production of
biodegradable detergents.26 This new approach was inspired by
the well-known bifunctional supported metal catalysts utilized
in catalytic reforming, where a noble metal catalyzes hydrogenation and dehydrogenation reactions and the acidic support
provides the active sites needed for isomerization, cyclization,
and hydrocracking reactions. In contrast, dehydrogenation
catalysts are monofunctional, as the acidic function must be
minimized to avoid side reactions.
Albeit all of the noble metals of group VIIIB are active in the
nonoxidative hydrogenation of alkanes, Pt is the only noble
metal utilized in commercial applications due to its superior
10620
Chemical Reviews
Review
Figure 10. Selectivity to butenes as a function of the number of butane pulses on a Pt/Al2O3 catalyst doped with dierent amounts of Na (left) and
TPO proles of the spent catalysts (right) showing that Na has a positive eect on the selectivity and that the presence of Na decreases the coke
deposition, respectively. Reprinted from ref. 84, copyright 2010, with permission from Elsevier.
4.1.2. Catalyst Deactivation. Two main processes cause

the deactivation of Pt-based dehydrogenation catalysts. First
and foremost, the side reactions that form carbon deposits on
the catalyst surface lead to the coverage of the active sites with
coke, which results in a drop in activity. As it is not possible to
completely prevent coke deposition, the catalyst needs to be
regenerated frequently by combusting the coke deposits.
However, the high temperature of both the regeneration
process and the dehydrogenation reaction triggers the
agglomeration or sintering of the platinum nanoparticles, the
concomitant loss in active sites also resulting in catalyst
deactivation. In addition to temperature, the composition of the
reaction mixture has a profound inuence on the sintering
behavior of supported Pt catalysts. As early as 1977, Fiedorow
et al. showed that Pt/Al2O3 catalysts more readily sinter in an
oxygen-containing atmosphere than in a hydrogen-containing
atmosphere at temperatures above 600 C.76,77 A third
important factor determining catalyst sintering is the
interaction between Pt and the catalyst support. Nagai et al.
reported that Pt on Al2O3 readily sinters after treatment in air at
800 C, while Pt on CeO2 remains stable due to the strong
interaction between ceria and Pt. The authors proposed that
the strength of the Ptsupport interaction depends on the
strength of the PtOM bond (where M represents the cation
in the support), which in turn shows an excellent correlation
with the electron density of the oxygen in the support. As a
result, the authors conclude the electron density of oxygen in
the support in eect controls the sintering of supported Pt
particles.78
However, Pt can be redispersed on the catalyst by the
addition of low amounts of oxygen and chlorine in the feed at
temperatures around 500 C, although this process can only
occur when the interaction between the support and the PtOx
and PtOxCly species formed is such that mobile surface
complexes are produced.76,7983
4.1.3. Role of the Support. Good supports for
dehydrogenation of alkanes have to be thermally stable,
especially under the harsh hydrothermal conditions that the
material experiences during the regeneration step. In addition,
limited support acidity is needed to avoid undesirable side
reactions, such as coke formation and alkane isomerization.
Finally, for an optimal distribution of the metallic particles, a
relatively large surface area and uniform pore size distribution

are preferred.85
High surface area alumina is the classical support employed
in platinum-based dehydrogenation catalysts due to its high
thermal stability, mechanical strength, and its exceptional ability
of maintaining the platinum nanoparticles dispersed, which is
crucial to attain stable catalyst performance. However, most
alumina supports are acidic, and therefore promoters are used
to curb support acidity. The addition of alkaline metals, such as
Li, Na, and K, poisons these acid sites, suppressing the
formation of coke on the support as shown in Figure 10.84,8689
Moreover, these promoters can modify the properties of the
platinum and suppress side reactions, such as hydrogenolysis
and isomerization, reportedly due to a reduction in the eective
size of platinum clusters.6972 Zn and Mg have also been used
to dope the support, because the resulting spinel phase is less
acidic and thermally more stable. These spinels also curtail the
sintering of platinum due to a strong metalsupport
interaction.9093 Yet another strategy is to use more stable
polymorphs of alumina, such as - or -alumina, because the
latter have considerably less Brnsted acid sites relative to alumina, albeit they also tend to have considerably lower
surface areas. It is also important to mention that to reach a
high dispersion of Pt on the support, Lewis acid sites and
amphoteric OH groups are required (which limits the use of
SiO2 and basic supports, such as MgO or ZnO). Indeed, by
hydrothermally treating -Al2O3 prior to impregnation, the
formation of additional hydroxyl groups results in a better Pt
dispersion after impregnation and calcination.94
Zeolites are a commonly used alternative support for
dehydrogenation catalysts. Dumesic and co-workers reported
PtSn supported on zeolite K-L to be a highly active, selective,
and stable catalyst in the selective dehydrogenation of
isobutane.9598 The authors suggest that in the presence of
K, zeolite L can stabilize small PtSn particles in the
micropores of the support. In addition, this catalyst shows a
high tolerance to coke deposition. This is particularly important
as the strong acid sites of zeolites make them very susceptible
to coking.85 Recently, Azzam et al. studied how the structure of
zeolite L aected the stability of a Pt/K-L catalyst during the
dehydrogenation of propane.99 Using kinetic isotope experiments, they found no evidence of isotopic scrambling,
suggesting that the dehydrogenation reaction was irreversible
10621
Chemical Reviews
Review
Figure 11. In the gures to the left, HERFD-XANES intensity contour maps (Pt-L3 edge) of Pt/Al2O3 (top) and PtSn/Al2O3 (bottom) catalysts
are shown at dierent stages of a dehydrogenation experiment. The gures to the right depict the interpretation of Iglesia-Juez et al. for these spectra
in terms of changes to the Pt nanoparticles. Reprinted from ref 122, copyright 2010, with permission from Elsevier.
properties of platinum, is necessary to obtain an optimal

catalyst. Tin is by far the most studied promoter, and all of the
platinum-based catalyst formulations that are industrially
applied include this post-transition metal, the addition of
which modies both the platinum active phase and the support.
Essentially, Sn suppresses hydrogenolysis and isomerization
reactions, minimizes metal sintering, neutralizes the acidity of
the support, and facilitates the diusion of the coke species
from the metal surface to the support. Although these benecial
eects of tin promotion are well-known and have been amply
described in the literature, the working principles of the PtSn
system are still under debate. Indeed, both geometric and
electronic eects have been invoked to explain the mechanism
through which tin inuences the catalytic properties of
platinum. In any case, it is clear that SnO2 species present on
the catalyst are reduced by the hydrocarbon feed during the
dehydrogenation reaction and/or during a prereduction step,
yielding a PtSn alloy. The reducibility of SnO2 depends
largely on the interaction with the support, strong interaction
making the reduction more dicult. However, Pt in close
proximity assists with the reduction of Sn.85
From the geometric point of view, it has been suggested that
isomerization, hydrogenolysis, and the formation of coke
precursors can all be suppressed by reducing the size of
in this system. Furthermore, the steric restraints of the zeolite

structure do not prevent coke formation via bimolecular
reactions, in contrast to what was observed for hexane
aromatization.100,101 Iglesia et al. found Pt supported on Na
[Fe]ZSM-5 zeolites to be highly selective and stable catalysts in
the dehydrogenation of light alkanes.102 These authors claim
that the replacement of the framework Al3+ for Fe3+ ions is
essential for the Pt precursor to anchor inside the 10-member
ring zeolite channels, which in turn is necessary to obtain
subnanometric Pt clusters after reduction. The resulting catalyst
displays high stability and selectivity toward light olens, which
is explained by the low acidity of the zeolitic support. Further
modications of the catalyst involving the neutralization of the
residual acid sites with Cs after Pt reduction improved the
selectivity of the catalyst in the dehydrogenation of n-pentane
and n-heptane.103 On the other hand, SBA-15 (a mesoporous
silica) was found to be an unsuitable support, as the PtSn
nanoparticles readily sinter due to the weak metalsupport
interaction.85
4.1.4. Role of the Promoters. Despite the fact that
platinum-based catalysts display very high activity in dehydrogenation reactions, their intrinsic selectivity to alkenes and
catalyst stability are not entirely satisfactory. Consequently, the
addition of promoters, which are able to modify the catalytic
10622
Chemical Reviews
Review
Figure 12. An artists impression of a PtSn nanoparticle within a PtSn/Al2O3 catalyst is shown as was discussed in the preceding section. The
surface of the alumina is covered with SnO2, which gets reduced when in close proximity to Pt and incorporated into a bimetallic PtSn
nanoparticle. Surface Pt atoms catalyze the dehydrogenation reaction according to the mechanism shown in the inset.
platinum nanoparticles.97,98,104,105 Notably, the formation of a

platinumtin alloy and/or the partial covering of the platinum
particles by tin species results in the creation of smaller
platinum ensembles. Olsbye et al. have suggested that tin
selectively covers low coordination platinum sites (such as
steps, corners, edges, and defects), which they believe are
responsible for the hydrogenolysis reaction.106,107 Furthermore,
the so-called ensemble eect can be employed to explain the
geometric advantages of Sn addition to Pt-based catalysts.
Dehydrogenation is believed to be a structure-insensitive
reaction, meaning that very small groups or even single Pt
atoms can catalyze the reaction. On the other hand, side
reactions, such as coking, are structure-sensitive, and require
relatively large ensembles of Pt. PtSn alloy formation
eectively reduces the amount of neighboring Pt atoms within
the ensembles, curbing these structure-sensitive reactions.
Moreover, it has been suggested that the sintering of Pt
nanoparticles is retarded with the addition of Sn.107 Additionally, coke precursors prefer to adsorb on large Pt ensembles,
and thus the addition of Sn facilitates the migration of these
precursors from the PtSn surface to the support, eectively
reducing the detrimental eects of coke deposition.104,108
A secondary eect of Sn addition is the modication of the

electronic properties of platinum. Alloyed metallic Sn or Sn2+
species in close contact with platinum are able to transfer
electrons to the 5d band of platinum atoms, which alters the
adsorptive and catalytic properties of the metal. The interaction
of promoters with platinum can be assessed by measuring the
heat of adsorption of distinct probe molecules through
microcalorimetric methods.109114 Dumesic et al. performed a
thorough adsorption microcalorimetry study and observed that
the heat of ethylene and isobutene adsorption decreases with
the addition of tin.115117 Infrared spectroscopy revealed that
during the adsorption of ethylene, tin suppresses the formation
of ethylidyne species and weakens the molecular adsorption of
ethylene on the platinum surface. Additional DFT calculations
explained these experimental results by revealing an increase in
the electron density of platinum, which weakens the adsorption
of ethylidyne on platinum atoms in close contact with tin. As
discussed in section 5.1, ethylidyne is suspected to be an
important precursor of coke and hydrogenolysis. Furthermore,
DFT calculations described in Figure 28 revealed that Sn
promotion signicantly reduces the dehydrogenation reaction
rate of platinum catalysts, as the energy barrier of the
dissociative adsorption of propane is increased. On the other
10623
C3H8 = 73, H2 = 27
C3H8 = 95, C2H6 = 4.5, iC4H8 = 0.5
iC4 = 33, H2 = 67
iC4 = 50, H2 = 50
C3H8 = 100
C3H8 = 75, H2 = 25
C3H8 = 100
C3H8 = 75, H2 = 25
C3H8 = 20, H2 = 25, He = 55
C3H8 = 20, H2 = 25, He = 55
C3H8 = 25, He = 75; iC4 = 33, H2 =

7, He = 60
iC4 = 33, H2 = 7, He = 60
C3H8 = 30 kPa, N2 = 70
C3H8 = 50, H2 = 50
feed composition
3130
2725
6146
1915
4029
41.739.1
24.815.7
27.512.6
1611.4
20.416.3
3720
3120
1211
3313 2010
conversion (%)b
9798
99100
9299.8
9596
5590
95.398
91.690.6
94
99.2
98
90100
9598
9295
98.399.8 96.598.6
selectivity (%)b
0.024
0.0010
0.0048
0.14
0.04
0.012
0.048
0.024
0.20
0.14
101
101
101
101
101
101
101
102
102
0.0051
0.29
0.028
1.27 101
1.09 101
4.97 101
1.76
1.75
3.94
2.090
1.41
2.20
1.45
1.43
1.13
1.83
0.0080; 0.0041
kd (h1)
1.51; 1.08
specic activity
(s1)c
2
100
126
2
12
9
12
40
2
2
140
5
3.5
166
catalyst
lifed (h)
138
139
98
140
126
141
126
142
132
133
135
136
137
102
ref
Values of conversion and selectivity of the end and start of each catalytic run are shown. From the articles considered, only the best performing catalyst is included. bFirst value is obtained at the start of
the cycle, second at the end. cSpecic activity is dened as (mol olen formed)/(mol Pt*t(s)). dCatalyst life = total time single cycle/experiment.
605
555
600
500
555
590
555
590
600
600
3.9
2.8
13.2
248
2.6
3
2.6
3.0
2.6
2.6
WHSV
WHSV
WHSV
WHSV
WHSV
WHSV
WHSV
WHSV
WHSV
WHSV
520
519
550
(2) 2 wt % Pt1 wt % Sn/CeO2/C

(3) 0.35 wt % Pt1.26 wt % Sn/Al2O3
(4) 0.5 wt % Pt0.6 wt % Sn/
MgAl2O4
(5) 0.6 wt % Pt5 wt % Ga/MgAl2O4
(6) 0.5 wt % Pt/Zn/Silicalite
(7) 0.58 wt % PtSn/K-L
(8) 1.0 wt % PtSn/SiO2
(9) 0.5 wt % Pt/Zn-Beta
(10) 0.5 wt % PtNa/SnZSM-5
(11) 0.5 wt % PtZn/NaY
(12) 0.5 wt % PtSnNa/Al-SBA-15
(13) 0.9 wt % Pt/Mg(Ga)(Al)O
(14) 0.7 wt % Pt/Mg(In)(Al)O
=
=
=
=
=
=
=
=
=
=
WHSV = 15.1; WHSV

= 18.1
WHSV = 24.8
WHSV = 3.5
WHSV = 36.6
space velocity (h1)
520
reaction
temp (C)
(1) 0.1 wt % PtNa[Fe]/ZSM-5
catalyst
Table 3. Summary of the Catalytic Data of the Pt-Based Dehydrogenation Catalysts Discusseda
Chemical Reviews
Review
10624
Chemical Reviews
Review
Figure 13. Graphs represent Raman spectra of x wt % CrOx/SBA-15 and CrOx/Al2O3 catalysts (x: a = 0.5, b = 1, c = 5, d = 10). Bands attributed to
CrO42 and CrOCr species are observed at dierent loadings and correspond with isolated Cr atoms and crystalline Cr2O3. Reprinted from ref
149, copyright 2009, with permission from Elsevier.
analogous to that of PtSn, although the electronic eect is

believed to be weaker.134
An overview of the Pt-based dehydrogenation catalysts
discussed above, including relevant reaction parameters, such as
conversion, selectivity, temperature, space velocity, and feed
composition, is given in Table 3. The activity of the catalysts is
compared using their calculated specic activity, which is
dened as the moles of olen formed per mol of Pt per second
at the start of the dehydrogenation reaction. As mentioned
before, the dehydrogenation reaction is insensitive to the
structure of the Pt, so all Pt atoms present at the surface of the
catalyst nanoparticles can be assumed to participate equally in
the dehydrogenation reaction. However, it was not possible to
calculate turnover frequency (TOF) due to the fact that Pt
dispersion values (the ratio between the Pt at the surface versus
the interior of the catalyst nanoparticles) were not forthcoming
in most of the references included in Table 3. Finally, as a proxy
for catalyst stability, we have determined the deactivation rate
(r d ), for which we assume a rst-order deactivation
mechanism:135
hand, the selectivity is increased as the energy barrier for the

desorption of propylene drops, making deep dehydrogenation
and cracking reactions less likely.114,118,119 The balance
between these two eects results in an optimum ratio of Pt
and Sn, which these calculations suggest is Pt3Sn. On the basis
of X-ray photoelectron spectroscopy (XPS) observations, Siri et
al. reported an increase of the electron density of platinum
when tin was incorporated to silica and -alumina-supported
catalysts.120 DFT calculations have shown that the addition of
Sn also reduces the heat of adsorption of hydrogen on the Pt
Sn (111) surface, which suggests that the diusion of hydrogen
atoms is more facile on the latter than on a Pt (111)
surface.73,121 Figure 11 illustrates the important role that the
electronic modication of Pt by Sn is believed to play in
inhibiting coke formation. The left part of the gure shows high
energy resolution uorescence detection X-ray absorption near
edge structure (HERFD-XANES) intensity contour maps of
the Pt L-edge of catalysts comprising alumina-supported Pt
(top) and PtSn (bottom). The right part of the gure shows a
graphical interpretation of the XANES spectra, which illustrates
the size of the metallic nanoparticles in both catalysts increasing
during the regeneration step, an eect that is more severe for
Pt/Al2O3. A shift to higher energies in the XANES spectra
during the reduction step reveals that Pt becomes more
electronegative, which is attributed to the formation of the Pt
Sn alloy. This phenomenon, which is triggered by the partial
reduction of the SnO2 present on the support, is reversed
during the oxidation step.122 Sn enrichment of the Pt
nanoparticles may be a problem, as the Pt3Sn alloy is believed
to be signicantly more active for the dehydrogenation reaction
than the PtSn alloy.114,123 Summarizing, Figure 12 shows an
artist impression of a PtSn nanoparticle on alumina.
Other metals have been studied as promoters of the activity
of platinum in the dehydrogenation of alkanes. Most notably,
zinc prevents undesired side reaction such as coke formation
and isomerization in a fashion similar to that of tin.124129
GaOx and InOx dopants are used to hinder side reactions by
the poisoning of surface Brnsted acid sites, while the
formation of PtGa and PtIn alloys has both a geometric
and an electronic eect on the Pt nanoparticles, decreasing
coke deposition further.130133 Another alloy that has been
investigated is PtGe, which is believed to function in a way
kd (h ) =
ln
1 Xend
Xend
) ln(
t
1 Xstart
Xstart
where Xstart and Xend represent the conversion at the start and
the end of an experiment, and t is the duration of the
experiment in hours, higher kd values being indicative of rapid
deactivation, that is, low stability.
4.2. Chromium Oxide-Based Catalysts
Chromium oxide-based catalysts have been one of the most

actively investigated formulations since Frey and Huppke rst
reported the dehydrogenation activity of Cr2O3 in 1933.143
During the 1940s, catalysts comprising chromia supported on
alumina were used by UOP to industrially dehydrogenate
butane to produce butene, which were dimerized and
hydrogenated to form high-octane aviation fuels. These
catalysts have been applied to several processes for the
dehydrogenation of light olens henceforth.26 Given the
industrial relevance of formulations comprising chromium
oxide, their structure under dierent reaction conditions, the
nature of the active sites, the role of the support, the eect of
alkali metals promoters, and the mechanism of the dehydrogen10625
Chemical Reviews
Review
Figure 14. In situ XANES spectra of the Cr K-edge of the 5 wt % catalysts after oxidative pretreatment (), after 90 s of dehydrogenation (black
- - -), after 15 min of PDH (gray ), and after regeneration (gray ). Notably, the Cr6+ species present on the catalyst are reduced to Cr3+
species in the rst 90 s of the reaction. Reprinted from ref 149, copyright 2009, with permission from Elsevier.
Figure 15. Graph to the left shows in situ diuse reectance spectra obtained during the dehydrogenation of propane on a CrOx/SBA-15 catalyst. In
the right set of graphs, the relative change of bands (BD) corresponding to Cr3+ and Cr6+ is shown. Both graphs illustrate that Cr3+ (band at 700
nm) is formed at the expense of Cr6+ (bands at 260 and 370 nm) during PDH. Reprinted from ref 150, copyright 2011, with permission from
Elsevier.
formation of Cr2O3 crystallites. These crystallites are more

readily formed on SBA-15 than on alumina, as the interaction
between chromium oxide and SBA-15 is relatively weak.
Prior to the 1990s, the study of the structure of chromium
oxide-based dehydrogenation catalysts had been restricted to
conditions far removed from those found under real operation.
For this catalytic system, this represented a great limitation, as
the reaction mixtures are believed to reduce chromium oxide
species. The development of in situ spectroscopic techniques
made it possible to probe these catalysts under reaction
conditions. As is shown in Figure 14 with XANES and in Figure
15 with UVvis spectroscopy, it was observed that surface Cr3+
species are produced at the expense of Cr6+ species during the
initial stages of the dehydrogenation reaction and that
polychromate is more easily reduced than monochromate.151153 Consequently, Cr3+, coordinatively unsaturated
Cr2+, as well as a mixture of both Cr3+ and Cr2+ have been all
proposed as the active sites for dehydrogenation reactions.154157 Cr6+ or Cr5+ have not been observed to be active
for the dehydrogenation reaction, albeit Cr6+/Cr5+ species do
ation reaction over these catalysts have all been intensively

studied.
4.2.1. Nature of the Active Sites. Because the adsorption
of paran molecules takes place on CrO sites, understanding
the exact nature of these species under dehydrogenation
reaction conditions is critically important.144 On freshly
prepared catalysts a number of surface species, including
Cr6+, Cr5+, Cr3+ and Cr2+, have been identied. Furthermore,
chromates, polychromates, crystalline -chromia and amorphous chromia have been identied through a variety of
techniques.145148 The relative concentration of these species
seems to be dependent on a number of factors, such as
chromium loading (relative to the surface area), the calcination
treatment used during catalyst preparation, and the support
employed. This is illustrated in Figure 13, where chromium
species deposited on SBA-15 (a mesoporous silica) and
alumina are compared by Raman spectroscopy. Dierent
species are observed depending on the Cr loading: at low
loadings isolated CrO42 species predominate, while at higher
concentrations CrOCr species become prevalent due to the
10626
Chemical Reviews
Review
appear to be the precursor of active Cr3+ sites. This has been

assigned to the fact that reduction of Cr5+ species leads to the
formation of monomeric Cr3+ active sites. Indeed, it has been
observed that Cr5+ species that are not reduced in a hydrogen
atmosphere can be reduced by alkanes.150,158161
With respect to coordination, it appears that Cr3+ sites need
to be coordinatively unsaturated to show catalytic activity.162
Indeed, a semiquantitative relationship has been found between
the amount of pseudo-octahedral Cr3+ formed during reduction
and the dehydrogenation activity shown by the catalyst.152,153
Furthermore, it has been observed that isolated Cr3+ species
with two coordinative vacancies are formed on CrOx/Al2O3
catalysts in the presence of an alkane. Thus, it appears that
these catalysts contain coordinatively unsaturated Cr3+ under
reaction conditions. Experiments performed with Cr loadings
above monolayer coverage, which have been shown to include
crystalline chromia clusters, suggest that bulk Cr3+ ions in a
perfect octahedral environment are inactive toward alkane
dehydrogenation. Admittedly, these fully coordinated Cr3+
species are highly stable, whereas coordinatively unsaturated
Cr3+ species would try to attain the stability of bulk Cr3+ ions by
adsorbing alkane molecules from the gas phase.
With respect to nuclearity, the need for a highly dispersed
active phase is highlighted by the fact that, while activity rst
rises with increasing chromium loading, a drop in activity is
observed when the loading is such that crystalline Cr2O3 starts
forming, as illustrated in Figure 16.163165 The fact that activity
dispersed Cr3+ phase, the role of the other phases being the
replenishment of the depleted Cr3+ surface phase during long
runs.152,153,166 More detailed studies involving the calculation of
turnover frequencies (TOF) have led some workers to propose
that multinuclear Cr3+ sites are more active than isolated Cr3+
ions, a proposition that nds both support and dissent in the
literature.147,160,163,167169 Kumar et al. observed that when alumina was employed as the catalyst support, oligomeric
chromium oxide species would display the highest activity and
selectivity. However, when SBA-15 was used as the carrier,
isolated chromia species with coordination number larger than
4 showed the highest activity.149 It is clear that the optimal
nuclearity of active Cr3+ sites remains a topic still under debate.
Albeit selectivity has been found to be insensitive to
chromium loading, it does change over time.170 At rst, the
selectivity toward alkenes is fairly low and the catalyst produces
mainly carbon oxides, as is shown in Figure 17. This is
attributed to the fact that during this induction period, highly
oxidized, and thus catalytically inactive, chromium species are
being reduced to active chromium species with lower oxidation
states. Eventually, a high selectivity of 93.6% can be reached at a
conversion of 55%.29
To summarize, in recent years dehydrogenation activity has
been mostly assigned to Cr3+ species that are both dispersed
and coordinatively unsaturated. Thus, it seems that active sites
need to fulll oxidation state, coordination, and nuclearity
requirements. In fact, these and other factors, such as the origin
of Cr3+ species, have been used by some workers to classify the
catalytically active sites.171173
It should also be mentioned that many other factors could
also inuence the activity of Cr-based dehydrogenation
catalysts. For instance, the addition of CO2 as a mild oxidant
is known to improve the activity of chromium dehydrogenation
catalysts by removing the evolved hydrogen through the reverse
water gas shift reaction, which eectively shifts the equilibrium
toward product formation.174176 At the same time, CO2 is
ineective at oxidizing catalytically active Cr3+ species to
catalytically inactive Cr5+ or Cr6+ species.150,177,178
To obtain high activity, the operating conditions also need to
be considered, because activity is directly proportional to both
the reaction temperature and the partial pressure of the alkane.
Moreover, alkane conversion has been found to be inversely
proportional to space velocity, and selectivity has been found to
be directly proportional to alkene selectivity. In studies where
Cr/Al2O3 catalysts were used in the nonoxidative dehydrogenation of ethane, it was found that when the space velocity
(GHSV) decreased from 3600 to 600 h1, ethane conversion
increased from 27.6% to 31.4%, while selectivity to ethylene
decreased from 91.7% to 64.9%.179 This eect has been
attributed to the fact that lower space velocities favor ethane
activation as well as methane and coke formation.
In conclusion, it is possible to achieve high activity and
selectivity over Cr-based hydrogenation catalysts by adjusting
the composition of the catalysts and/or the operating
conditions.
4.2.2. Catalyst Deactivation. Over the course of a
catalytic cycle, CrOx dehydrogenation catalysts lose their
activity due to the deposition of coke, and as is the case with
Pt-based catalysts, they need to be periodically regenerated.
Experiments performed using a tapered element oscillating
microbalance (TEOM), shown in Figure 17, revealed that the
weight of the catalyst increases during the initial stages (1040
min) of the reaction, no additional carbon being deposited after
Figure 16. Yield of butenes at 540 C over a chromiaalumina catalyst

as a function of Cr loading. Activity increases linearly with Cr loading,
even above monolayer coverage. However, the yield drops once Cr2O3
crystallites start being detected via XRD. Reprinted from ref 163,
copyright 2000, with permission from Elsevier.
and Cr loading hold an almost linear relationship below

monolayer coverage suggests that both isolated and clustered
Cr3+ centers are catalytically active. Actually, three dierent
types of Cr3+ ions have been spectroscopically detected at these
loadings: (1) Cr3+ species formed by reduction of surface Cr6+/
Cr5+ ions; (2) isolated centers of Cr3+ ions stabilized by the
alumina surface; and (3) small amorphous (and thus XRD
invisible) Cr3+ clusters.29,49 Determining the individual
contributions of these species to catalytic performance has
proven to be extremely challenging. Some studies correlating
dehydrogenation activity with Cr-speciation suggested that a
vast majority of catalytic activity can be ascribed to the
10627
Chemical Reviews
Review
Figure 17. To the left, the conversion and selectivity observed in a 86 min period of propane dehydrogenation on a 13 wt % CrOx/Al2O3 catalyst.
Conversion drops over time, while the selectivity remains stable after increasing during the initial stages of the reaction. To the right, the mass uptake
of the catalyst as measured by a tapered element oscillating microbalance (TEOM) is shown as a function of time-on-stream. Over the rst 40 min,
the catalyst mass increases due to coke deposition, after which it remains stable. Reproduced from ref 180, with permission from the PCCP Owner
Societies.
Figure 18. Top graph shows the ex situ diuse reectance UVvis spectra of both fresh and deactivated CrOx/Al2O3. The bottom left graph shows
the extent of the shift for the bands at 600 and 450 nm, which correspond to the rst and second dd transitions, respectively. The bottom right
graph shows that the fraction of Cr3+ incorporated into the alumina framework is strongly dependent on the calcination temperature. Reprinted from
ref 49 (top graph), copyright 2004, and ref 172 (bottom two graphs), copyright 2002, both with permission from Elsevier.
into the alumina framework, which causes the amount of

catalytically active Cr species to decrease with each
regeneration step.29 This phenomenon has been measured by
UVvis spectroscopy, as shown in Figure 18. In the gure, a
CrOx/Al2O3 catalyst is treated at 1100 C for 12 h, which
results in the bands corresponding to Cr3+ species to shift,
signaling the incorporation of Cr3+ species into the alumina
this time. However, conversion continues to drop beyond this

point, which indicates that the carbon originally deposited
becomes more detrimental for the catalytic activity over time.
Although the catalysts recover most of their activity after each
reactionregeneration cycle, the loss of activity is not entirely
reversible. Indeed, the heat released through the combustion of
coke during regeneration leads to the incorporation of chromia
10628
Chemical Reviews
Review
chromiaalumina catalysts have been investigated. These

include the incipient wetness impregnation of alumina with
CrO3 or Cr(NO3)3 solutions (if an alkali ion is added, an
aqueous solution of its chromate salt is used), the
coprecipitation of Al(OH)3 and Cr(OH)3 from Al3+ and Cr3+
solutions, a type of chemical vapor deposition (CVD) involving
volatile chromium compounds being deposited on alumina
called atomic layer deposition, and the decomposition of
complexes including both alumina and chromium. All of these
methods are followed by a calcination step at approximately
600 C. Increasing the calcination temperature above 600 C
causes a loss of dehydrogenation activity, probably due to the
incorporation of Cr into the Al2O3 framework, as discussed in
the preceding section.182 Very recently, an interesting synthesis
strategy for preparing a mesoporous CrOx/Al2O3 catalyst by
using a Cr-containing metalorganic-framework lled with an
alumina precursor and subsequently combusting the organic
material was reported. The resulting catalyst shows excellent
thermal stability and is very active for isobutene dehydrogenation.189 Interestingly, increasing the pretreatment temperature
of the alumina support has a positive eect on the TOF shown
by the resulting catalyst, which has been attributed to an
increase in the number of multinuclear Cr3+ sites.172 In
addition, this pretreatment will result in a more stable Al2O3
polymorph, which is less likely to sinter. However, the use of
alumina as support is not free of drawbacks. For instance, Al2O3
can catalyze cracking and coking due to its surface acidity,
which leads to a drop in both activity and selectivity. A solution
for this is the doping of alumina-based dehydrogenation
catalysts with alkali ions, which can eectively poison these
undesirable acid sites.
The use of zirconia as an alternative support has also been
investigated. Indeed, the low surface acidity of this oxide and
the fact that chromia imparts an enhanced sintering resistance
to ZrO2 make the latter a viable substitute for alumina.159,160
De Rossi et al. have reported that catalysts with zirconia as the
support are more active in the dehydrogenation of propane and
isobutane than the corresponding catalysts supported on silica
or alumina.167,168,181,190 This has been attributed to an increase
in the amount of surface Cr5+ that is produced during the
regeneration of Cr2O3/ZrO2. The latter improves activity
because Cr5+ forms catalytically active mononuclear Cr3+ under
reaction conditions, as has already been mentioned. In turn, the
high amount of Cr5+ on this catalyst has been attributed to the
strong interaction between chromia and the basic support.170
The choice of zirconia over alumina as the catalyst support
oers the additional advantage of preventing the formation of
CrAl mixed oxides. Nevertheless, Cr2O3/ZrO2 has been
observed to deactivate faster than other less active catalysts
including Cr2O3/Al2O3, which has been partially imputed to the
higher coking rate displayed by Cr2O3/ZrO2. Additionally, XPS
studies suggest that chromiazirconia catalysts also favor the
formation of chromium carbides, which are inactive in
dehydrogenation and consequently cause catalyst deactivation.170
De Rossi and co-workers have also concluded that albeit Cr/
ZrO2 catalysts can outperform their alumina- and silicasupported counterparts, zirconia-supported formulations could
be further improved by increasing the surface area and the
thermal stability of Cr2O3/ZrO2. This has inspired other
workers to produce chromium-based dehydrogenation catalysts
with high surface areas and highly dispersed active sites. For
example, Fujdala and Tilley have reported the use of a
support. Consequently, the dehydrogenation activity shown by

the catalyst drops, and an almost linear relationship has been
found between activity and the amount of Cr6+ species on the
catalyst surface (as measured after the regeneration
step).150,166,181
Two processes are believed to lead to the dissolution of Cr3+
into the Al2O3 carrier and result in catalytically inactive Cr3+
species. First, the sintering of the alumina may cause
entrapment of the Cr species inside the bulk of the support.
As an additional eect, the loss of alumina surface area reduces
the ability of the support to stabilize Cr6+ species formed during
the regeneration step.154,182 Consequently, Cr3+ ions migrate
into the support. Indeed, it is believed that the similar ionic
radii and charge of Al3+ and Cr3+ facilitate the diusion and
subsequent incorporation of Cr3+ into the octahedral sites of
alumina, which produces a type of chromiaalumina spinel that
is inactive for the dehydrogenation reaction.145,183 Of note is
the fact that in the presence of -alumina, entrapment is
believed to be the dominant process, while in its absence
migration is favored.145,183
At elevated temperatures, -alumina undergoes a thermodynamically driven transformation to - or even -alumina, which
coincides with a considerable loss in surface area. It is unknown
whether the presence of chromia facilitates or hinders the
sintering of the alumina support, as some authors have found
the presence of chromia to enhance sintering, while other
research groups have observed the opposite.172,184186 The
transformation of -alumina to -alumina can be restrained by
doping the support with silica, zirconia, magnesia, or lanthana.
For example, Wachowski et al. observed that when a small
amount of La2O3 is added to Al2O3, the transition of -alumina
to -alumina occurs at a temperature 136 C higher relative to
when no lanthana is added.187,188 The authors found that the
use of cations having ionic radius larger than Al3+, which tend
not to dissolve in Al2O3, in combination with a preparation
method capable of aording high dopant dispersion, is most
eective in preventing the formation of -alumina.
It is important to note that the advantages oered by the
combustion of coke during the regeneration step far outweigh
these unwanted side eects. First and foremost, the heat
released in this step is harnessed to fuel the dehydrogenation
reaction, making coke deposition not completely undesirable.
Furthermore, oxidation of Cr3+ to Cr6+ during regeneration also
accomplishes the redispersion of chromia crystallites. The latter
is particularly important, given that the sintering of chromia
under dehydrogenation conditions has also been reported to be
a contributing factor to the reversible deactivation of
chromium-based catalysts.171,172
Another interesting example of a deactivation phenomenon
caused by a change in surface composition related to the nature
of the support has been observed on catalysts comprising
mesoporous silica as carrier. Indeed, by analyzing this material
via XPS before and after 6 h of time-on-stream, an important
amount of Cr3+ species was found to be reduced to Cr2+ by the
hydrogen evolved during the anoxic dehydrogenation reaction.
In turn, the resulting Cr2+ species were found to be inactive in
dehydrogenation.
4.2.3. Role of the Support. The nature of the support is
important not only because it governs the mechanical
properties of the catalyst, but also because the interaction of
the carrier with the active phase can have a profound impact on
catalyst activity and selectivity. Alumina is by far the most
common support, and a variety of methods to produce
10629
Chemical Reviews
Review
the nonoxidative dehydrogenation of propane, catalysts based

on these materials display an initial conversion of 18.1% with a
selectivity toward propylene of 87.1% at 550 C. 162
Interestingly, it has been observed that while the selectivity of
this material toward cracking decreases with time-on-stream, its
selectivity toward the alkene tends to remain constant. The
latter has been explained by assigning the occurrence of
cracking and dehydrogenation reactions to strong and medium
strength acid sites, respectively. The authors argue that strong
acid sites are rapidly deactivated, inhibiting cracking reactions
and favoring the production of the alkene.
4.2.4. Role of the Promoters. Alkali metals, such as K, Rb,
and Cs, are known to promote both the activity and the
selectivity of chromium-based dehydrogenation catalysts. This
phenomenon has been attributed to the ability of large alkali
cations to stabilize the structure of the support, to reduce
surface acidity, and/or to increase the number of active
chromium sites.154,189 Additionally, the presence of alkali metal
promoters increases the dispersion of the active phase by
assisting the formation under oxidizing conditions of Cr6+
complexes that subsequently give rise to Cr3+ under a reducing
atmosphere.
A notable exception to this promotion eect has been found
in the doping with K+ of several PDH catalysts, such as
chromia-pillared -zirconium phosphate, galliumchromium
mixed oxides pillared -zirconium phosphate, and chromia/
zirconia.162,196 For these materials, doping with K+ results in a
dramatic decrease in catalytic activity. According to some
authors, this indicates that oxygen ions play an active role in the
dissociative adsorption of propane and/or in the desorption of
hydrogen. Indeed, these workers contend that in these catalysts,
potassium forms bonds with oxygen anions, which causes some
properties, such as the Lewis acidity and the coordinative
unsaturation of the latter, to be altered in a way that hinders
catalysis.181,190 A similar argument has been used to explain the
drop in activity observed when doping chromia/zirconia
catalysts with sodium, although this eect has not been
observed for alumina-based formulations.
An artists impression of a CrOx/Al2O3 catalyst is shown in
Figure 20, while Table 4 oers an overview of the dierent
CrOx catalysts used for the dehydrogenation of light parans,
as discussed in the previous section. As was the case in Table 3,
in Tables 49 the activity of the catalysts is compared using
their specic activity, as calculated according to the denition
accompanying each table. The use of TOF was precluded by
the fact that in oxide-based catalysts there exists a wide array of
active species (e.g., monomers, oligomers, polymers, and
crystalline species) exhibiting dierent degrees of catalytic
activity and accessibility (some of these species are in the bulk
as opposed to on the surface), which makes the determination
of the actual number of active sites impossible.
cothermolysis method involving several thermolytic molecular

precursors to produce a Cr/Si/Zr/O catalyst capable of
showing propane conversions greater than 35% and propylene
selectivities exceeding 90% at 450 C.191 According to the
authors, the aforementioned values are among the best
reported in the literature, particularly at such a low temperature. Korhonen et al. attempted to combine the high surface
area of alumina with the basic nature of zirconia by covering an
alumina support with a zirconia layer.192 However, the resulting
CrOx/Al2O3ZrO2 catalyst displayed inferior activity as
compared to the benchmark CrOx/ZrO2 catalyst.
The use of silica as the support in chromia-based
dehydrogenation catalysts requires a relative high surface area
to stabilize Cr6+ sites.193 Surface area considerations also
explain the fact that titania does not constitute a good support,
as CrOx easily sinters due to its weak interaction with TiO2.
However, an active and stable propane dehydrogenation
catalyst has been obtained by preparing a titania mesoporous
material with Cr dopants.177
Temperature-programmed reduction (TPR) results suggest
that there is a direct relationship between catalytic activity and
the reducibility of the catalysts (see Figure 19).179 Indeed,
Figure 19. TPR proles of 6 wt % CrOx on dierent supports. The

catalysts with a strong reduction peak at low temperatures show the
highest catalytic activity. The gure is reproduced from ref. 179 with
kind permission from Springer Science and Business Media.
studies on the nonoxidative dehydrogenation of ethane have

revealed that catalytic activity based on ethane conversion and
ethylene yields follows the order CrOx/-Al2O3 > CrOx/SiO2 >
CrOx/MCM-41 CrOx/MgO CrOx/Si-2 (Si-2 is the
abbreviation of silicalite-2 zeolite, which has been described as a
silica analogue of the aluminosilicate zeolite ZSM-11194), a
reverse order being found for the reduction temperature of the
chromium species on the aforementioned catalyst supports.
CrMgAl and CrMg mixed oxides prepared using
layered double hydroxide (LDH) precursors have been studied
as catalysts for the nonoxidative dehydrogenation of ethane.195
Some of these catalysts were found to achieve ethane
conversions of 2730% and selectivity toward ethylene of
7175%. The method used for introducing Cr into the LDH
precursors was found to determine several properties of the
resulting mixed oxides, such as surface area, catalytic performance, coking resistance, and thermal stability against sintering.
Finally, the study of porous chromia-pillared -zirconium
phosphate materials has been particularly informative with
regards to the inuence that acidity has on catalytic behavior. In
4.3. Vanadium Oxide-Based Catalysts
Despite the excellent performance of Pt and CrOx catalysts in

paran dehydrogenation, a number of issues including catalyst
poisoning, the high cost of Pt, and environmental concerns
associated with the use of Cr have spurred the search for
alternatives. Vanadium oxides are known to be active for many
hydrocarbon oxidation reactions, including dehydrogenation.
Research on vanadium-based dehydrogenation catalysts started
in the 1980s, although at that time the focus was on oxidative
dehydrogenation using vanadiummagnesium mixed oxides.199,200
10630
Chemical Reviews
Review
Figure 20. An artists impression of the surface of a CrOx/Al2O3 catalyst in which coordinatively unsaturated Cr3+ species catalyze the
dehydrogenation reaction. Note that partially reduced CrOx species are formed on the surface of the catalyst, and that the Cr is partially dissolved in
the alumina support. The (simplied) reaction mechanism is shown in the inset.
catalyst, as OH groups serve as anchoring sites for the

vanadium precursors used to prepare these catalysts. On the
other hand, acid sites are detrimental for catalyst performance,
because they facilitate cracking and induce coke formation. A
schematic display of these vanadium species is shown in Figure
21.202
Several authors claim that two-dimensional polymeric V3+ or
4+
V species are the most active for dehydrogenation,203206 the
vanadium species bonded directly with the /-alumina support
being the most active sites in PDH. Isolated vanadium species
are also active, albeit they are susceptible to deactivation due to
coke formation. On the contrary, large V2O5 crystallites are
practically inactive in dehydrogenation.30,206
Prior to reaction, the vanadium in these catalysts is generally
present as V5+. However, the reduction of the catalyst by the
hydrocarbon feed results in the formation of V4+ and
V3+.207209 The reducibility of the vanadium species depends
on their molecular structure, VOV and VO bonds being
The nature of the vanadium present on the catalyst depends

to a large extent on the type of carrier, the support surface area,
the metal loading, and the oxidation state of vanadium.201 Far
below the theoretical monolayer coverage isolated vanadium
species are present, at intermediate surface coverage polymeric
species or sheets of vanadium oxide are favored, and above
monolayer coverage V2O5 crystallites become dominant.
Therefore, a combination of VO-support, VOV, or V
O moieties may be obtained, each having dierent catalytic
activity. In addition, depending on the support, mixed oxides
with Al, Zr, Ti, or Mg may be formed.
The surface area-to-metal loading ratio determines which
species will form. Indeed, Wu et al. reported that monomeric
species are predominantly present below 1.2 V/nm2, polymeric
vanadium species prevail between 1.2 and 4.4 V/nm2, and
crystalline V2O5 species preponderate at even higher loadings.
On the one hand, the presence of hydroxyl groups on the
surface of the support is important to obtain a well-dispersed
10631
Chemical Reviews
Review
Table 4. Summary of the Catalytic Data of the Cr-Based Dehydrogenation Catalysts Discussed
catalyst
(1) mesoporous 9 wt % CrOx/
Al2O3
(2) mixed oxide 40% Cr2O3/60%
Al2O3
(3) 73 wt % Cr2O3-pillared -ZrP
reaction
temp
(C)
space
velocity
(h1)
580
588 2
550
(4) GaCr mixed oxide on -ZrP

(20.1 wt % Cr, 9.3 wt % Ga)
(5) 1416 wt % CrOx/Al2O3
540
(6) 5 wt % CrOx/SBA-1
550
(7) 4 wt % CrOx/ZrO2
550
(8) (1 wt % Cr) CrSiZr

Xerogel
(9) Cr(2)Mg(12)Al(4)Ox (mixed
metal oxide)
(10) 5 wt % Cr, 10 wt % Ce/
SBA-15
(11) 20 wt % Cr, 1 wt % Na/
Al2O3
450
550
700
700
550
feed composition
conversion (%)a
selectivity
(%)a
C3H8 = 5, He = 95
15.710.0
98
iC4H10 = 100
2314
99.4
C3H8 = 7, He = 93
18.16.3
C3H8 = 7, He = 93
iC4H10 = 10, N2 =
90
WHSV =
1.2
WHSV =
0.3
WHSV =
0.86
WHSV =
3.2
GHSV =
3600
WHSV =
0.12
C3H8 = 6.67, CO2

= 33.3, He = 60
C3H8 = 2.5, N2 =
97.5
C3H8 = 1.9, N2 =
0.2, He = 97.1
C2H6 = 28.6, N2 =
71.4
C2H6 = 25, CO2 =
75
C3H8 = 10, N2 =
90
WHSV
3.3
WHSV
1.7
WHSV
2
WHSV
46.6
specic
activity (s1)b
kd
(h1)
catalyst
lifec (h)
ref
0.52
197
6.87 104
0.48
1.25
154
87.1
1.23 105
0.24
162
27.64.4
80
5.05 105
0.42
196
16 (start rst
cycle), 8 (30th
cycle)
3726
98 1
1.21 103
0.026
30
163
8592
1.66 104
0.14
3.75
165
60.924.3
76.287.0
2.86 105
0.26
170
37.822.7 (35 and

400 m)
29.6
91.194.7
1.85 104
0.12
6.08
191
71
6.54 104
195
5546
9697
0.059
175
4737
8089
0.069
198
7.41 106
a
First value is obtained at the start of the cycle, second at the end. bSpecic activity is dened as (mol olen formed)/(mol Cr*t(s)). cCatalyst life =
total time single cycle/experiment.
Figure 21. Various vanadium oxide species present on a catalyst support (S): isolated vanadium oxide species (a), dimers (b), two-dimensional
chains of vanadium oxide species (c), and V2O5 crystallites (d). Reprinted from ref 201, copyright 2003, with permission from Elsevier.
Ogonowski et al. reported that the eect of CO2 depends

strongly on the catalyst support. Relatively weak basic sites on
carriers, such as those present on an activated carbon support,
facilitate the reverse water gas shift reaction, which lifts the
thermodynamic limitation of the reaction and increases alkane
conversion. When these basic sites are too strong, as is the case
on Al2O3 and ZnO2, CO2 irreversibly adsorbs and poisons these
sites. This increases the acidity of the support, which in turn
increases coke formation and lowers selectivity. If these basic
sites are absent, as in the case of SiO2, no benecial eects of
CO2 are observed.
Alumina is commonly used as a support for vanadium oxide
catalysts, although Sokolov et al. concluded through testing a
wide range of Si- and Al-based supports that using silica to dope
the alumina support provides benecial eects on catalyst
stability and activity.214 Furthermore, as was shown by Harlin et
al., the performance of supported vanadium catalysts can be
improved through the addition of MgO, either as a promoter or
as a component of a mixed oxide. According to these workers,
the increase in dehydrogenation selectivity stems from the basic
nature of MgO, resulting in reduced coke formation.208
An overview of the dierent VOx catalysts used for the
dehydrogenation of light alkanes is shown in Table 5.
more easily reduced than VO-support bonds. Harlin et al.

studied the eects of vanadium oxide reduction with H2, CO,
and CH4 prior to the dehydrogenation reaction. In all cases, the
reduction was incomplete, and yielded a mix of vanadium oxide
species, including V3+, V4+, and V5+ (see Figure 22). Higher
initial catalytic activities were obtained in all cases, although the
eect was strongest after reduction with CO. The authors
suggested that when the catalyst was reduced with H2 and CH4,
detrimental OH groups were formed. V4+ was determined to
be the most active species in butane dehydrogenation.
Initially, deactivation is caused by the strong adsorption of
reactants on the active sites and to a lesser extent to coke
deposition. When the coke is removed through the calcination
of the catalyst, the latter may deactivate further as vanadium
oxide species sinter and form larger V2O5 crystallites or mixed
oxides such as AlVO4, both of which are less active.206 V2O5
crystallites can revert to active polymeric vanadium species by
treating the catalyst under an oxygen atmosphere at 600 C.210
In several studies, CO2 is added to the feed as a mild oxidant
to increase alkene yields.211213 Dierent reasons for this
benecial eect of CO2 have been suggested, such as the
removal of acid sites that catalyze unwanted side reactions, coke
gasication, oxidation of overly reduced vanadium centers, and
hydrogen removal by the reverse watergas shift reaction.
10632
Chemical Reviews
10633
As some of these catalysts require an activation period, the maximum (as opposed by the initial) conversion is considered to determine the deactivation rate. bFirst value is obtained at the start of the
cycle, second at the end. cSpecic activity is dened as (mol olen formed)/(mol Cr*t(s)). dCatalyst life = total time single cycle/experiment. eCatalyst is prereduced by CO prior to the dehydrogenation
reaction.
600
WHSV = 1.9
600
215
3
0.86
1230 to 1- and 13
butene
7015
212
0.72
WHSV = 5
580
(4) 4.9 wt % VOx4.7 wt %

Mg/Al2O3e
(5) 5.4 wt % VOx4.5 wt %
Zr/Al2O3e
(6) 4 wt % V36 wt % MgOx/
Cact
(7) 3.5 wt % VOx/Al2O3
i-C4H10 = 14.3, CO2

= 85.7
GHSV = 72 (at a pressure of C4H10 = 3, N2 = 97
0.5 bar)
45.526.5
76.485.0
5.77 104
1.17
208
1.34; 3.12
9.32 104; 9.40 104
8590 7788 (to ibutene)
4436; 5435
6238
WHSV = 5
580
(3) 5.2 wt % VOx/Al2O3c
i-C4H10 = 10, N2 =
90
i-C4H10 = 10, N2 =
90
32.98.5 (1560 m)
nC4H10 = 100
600
(2) 3.5 wt % VOx/-Al2O3
GHSV = 14 400
0.25
207
0.085
11.5
9.46 104
205
1
2.22
202
1.33
1.04
1.75 10
5576 (to all butene

isomers)
56.1 (all butene
isomers)
6585 (to i-butene)
113
nC4H10 = 10
WHSV = 2.6, N2 = 90
520
conversion (%)b
feed composition
(1) 2.2 wt % VOx/-Al2O3
Molybdenum oxides are also frequently used as catalysts in

hydrocarbon conversion reactions, including dehydrogenation.
To our knowledge, the rst report of the dehydrogenation
activity of a molybdenum oxide-based catalyst was of MoO3/
Al2O3 in the dehydrocyclization of n-heptane, which was
published in 1946.216 Since that rst report, the nonoxidative
dehydrogenation of alkanes on MoOx has been studied
scarcely, as emphasis has been given to the study of oxidative
dehydrogenation on these materials.
The chemistry of molybdenum oxide is comparable to that of
vanadium oxide, as Mo can be present as MoOx monomers,
polymers, or MoO3 crystallites, depending on the molybdenum
loading, the type of support used, and catalyst preparation
conditions.217
Harlin et al. noted that optimal activity was achieved with
monolayer coverage of molybdenum oxide on an alumina
support. In the fresh catalyst, molybdenum is present as Mo6+,
but analogously to catalysts comprising vanadium and
chromium oxide, MoO3 is reduced by the hydrocarbon feed
during reaction to create the active species.218 These reduced
sites are believed to be Mo4+ and Mo5+. Further reduction
results in sites that are more active, but the latter catalyze
cracking and deactivate fast due to coke deposition. In fact, at
space velocity (h1)
4.4. Molybdenum Oxide-Based Catalysts
reaction temp
(C)
Figure 22. Vanadium 2p3/2 XPS spectra of a VOx/Al2O3 catalyst

calcined (top) and reduced with CO (bottom). Reduction yields a
mixture of V5+, V4+, and V3+. Reprinted from ref 207, copyright 2000,
with permission from Elsevier.
catalyst
Table 5. Summary of the Catalytic Data of the V-Based Dehydrogenation Catalysts Discusseda
selectivity (%)b
specic activity (s1)c
kd (h1)
catalyst
lifed (h)
ref
Review
Chemical Reviews
Review
loadings above monolayer coverage, the catalyst is more easily

reduced, which results in a drop in catalyst activity.218,219 The
same authors also point out that such a catalyst deactivates due
to the formation of catalytically inactive Al 2 (MoO 4 ) 3
crystallites. To prevent the formation of this mixed oxide,
they added magnesium oxide to obtain a molybdenum
magnesium mixed oxide that also shows dehydrogenation
activity. It was found that an additional benecial eect of
magnesia addition is that the resulting mixed oxide has lower
acidity, which increases alkene selectivity and reduces coke
deposition. The reduction of the catalyst with hydrogen prior to
reaction increases conversion, although selectivity decreases
due to the formation of acid sites.220 Overall, MoO3 catalysts
deactivate rapidly due to coking and generally aord a relatively
low conversion and selectivity.
The formation of molybdenum oxycarbides (MoOxCy) has
been reported to occur during butane dehydrogenation on
unsupported molybdenum oxide. These materials are also
active in the dehydrogenation of parans, and a high activity
close to the thermodynamic equilibrium is initially obtained,
although over a period of a few hours the selectivity drops to
zero and the reaction yields only methane, as shown in Figure
23. During this period, the oxycarbide becomes depleted from
oxygen and a wide range of carbides active in hydrogenolysis
are formed.222 To improve the stability of the oxide, steam can
be added to the feed (which hampers the growth of these
carbides) or the catalyst can be doped with vanadium (which
increases the oxophilicity of the catalyst). X-ray diractograms
showing this transition are shown in Figure 24. However, the
introduction of steam is not free of drawbacks, as it induces
sintering. Hydrogen addition prevents coke formation and
increases conversion, albeit it is believed that H2 competitively
adsorbs on the active sites of the catalyst.221223
Consistent with the fact that MoOxCy is a catalyst active in
butane dehydrogenation, Harlin et al. showed that MoO3
deposited on SiC displays signicant catalytic properties
when hydrogen and steam are present in the feed.224 As
compared to bulk MoO3 and MoO3 supported on SiO2 and
Al2O3, higher selectivities and stability are obtained over
MoO3/SiC, albeit at the cost of a lower conversion. This was
explained by the increased reducibility of MoO3/SiC as
compared to the MoO3/Al2O3 catalyst. A summary of the
Mo-based catalysts used for the dehydrogenation of light
parans is shown in Table 6.
4.5. Gallium Oxide-Based Catalysts
Figure 23. Yield of dierent carbon molecules during butane

dehydrogenation on a MoOxCy catalyst at dierent steam partial
pressures. When no steam is added, the selectivity to C4 alkenes drops
rapidly and the methane yield reaches 100% after 5 h. The
hydrogenolysis reaction is suppressed when steam is added to the
feed. Reprinted from ref 221, copyright 1999, with permission from
Elsevier.
The dehydrogenation activity of gallium oxide supported on

ZSM-5, which was rst reported in the late 1980s in studies
dealing with conversion of propane to aromatics, has been the
target of renewed attention in recent years.225228 In the
interim, both bulk and supported Ga2O3 have been used as
dehydrogenation catalysts.
The exact nature of the active sites in Ga-based dehydrogenation catalysts is still under debate. Akin to Al2O3, Ga2O3 has
dierent polymorphs of which -Ga2O3 is not only the most
stable, but also the most active Ga species in PDH according to
some authors.229 -Ga2O3 has a monoclinic structure, in which
gallium atoms are equally distributed between tetrahedral and
octahedral congurations. This crystal structure has a high
concentration of weak Lewis acid sites (coordinatively
unsaturated tetrahedral ones on the surface of the material),
which are deemed to be centers of dehydrogenation activity.230
Indeed, a heterolytic dissociation reaction mechanism has been
proposed where H adsorbs on a Ga+ Lewis acid site and

C3H7+ on a neighboring oxygen, although these species can also
be adsorbed reversed.33,231 The assignment of Lewis acid sites
as the active species was substantiated by Chen et al.,232 who,
by relating the results of NH3-TPD measurements with activity
data for dierent GaxAl10xO15 mixed oxides, showed that a
high concentration of Lewis acid groups in the form of
coordinatively unsaturated tetrahedral Ga3+ cations is a
prerequisite for dehydrogenation activity (see Figure 25). Xu
et al. reported dehydrogenation activity on both Lewis and
Brnsted acid sites on GaO3/H-ZSM-5. A zeolitic support
displaying a low concentration of medium and strong acid sites
and a relatively high concentration of weak acid sites was found
to lead to the most active and stable catalyst.233,234 In addition,
Michorczyk et al. concluded that acidic groups catalyze the
10634
Chemical Reviews
Review
Figure 24. XRD diractograms of the fresh Mo oxycarbide catalyst before dehydrogenation, but after pretreatment (top left), as well as after
dehydrogenation reaction with no steam (top right), with 5 Torr of steam (bottom left), and with 18 Torr of steam (bottom right). In the absence of
steam, the well-dened MoO2 and MoOxCy species disappear during reaction and are replaced by more amorphous MoOxCy species, which are
enriched with carbon and presumably responsible for hydrogenolysis. The addition of steam prevents this transformation and with 18 Torr of steam,
MoO2 and ordered MoOxCy species are still observed on the spent catalysts. Reprinted from ref 221, copyright 1999, with permission from Elsevier.
Table 6. Summary of the Catalytic Data of the Molybdenum Oxide-Based Dehydrogenation Catalysts Discusseda
catalyst
reaction
temp (C)
space
velocity
(h1)
feed composition
(1) 13.4 wt % MoO3/Al2O3
560
WHSV = 2
n-C4H10 = 10, N2 = 90
(2) 12.8 wt % MoO38.4

wt % Mg/Al2O3
(3) MoO3
550
WHSV = 2
(4) 14.2 wt % MoOx/SiC
560
WHSV = 2
n-C4H10 = 9.3, N2 =
83.9, H2O = 6.8
n-C4H10 = 10, N2 = 90,
H2O = 3.1
conversion
(%)b
specic
activity (s1)c
kd
(h1)
catalyst
lifed (h)
ref
50
7.21 105
2.78
0.17
218
82
7.66 105
1.12
428
5584
2.96 10
0.28
7.5
221
2217
3871
5.41 105
0.32
224
2118 (second
min 25)
1311
selectivity
(%)b
As some of these catalysts require an activation period, the maximum (as opposed to the initial) conversion is considered to determine the
deactivation rate. bFirst value is obtained at the start of the cycle, second at the end. cSpecic activity is dened as (mol olen formed)/(mol
Mo*t(s)). dCatalyst life = total time single cycle/experiment.
high energy barrier. Nevertheless, theoretical calculations by

Pidko et al. indicate that these relatively stable [HGaOH]+
sites are preferentially regenerated by water desorption
resulting in a reduced active site.236 Murata and co-workers
have observed an increase in activity after a prereduction step
with hydrogen at 550 C, while a simultaneous increase of
Ga+H and GaOH sites was observed via NH3 TPD. These
authors proposed that these sites also form during reaction and
may be responsible for the dehydrogenation activity of the silica
supported Ga2O3 catalyst.237,238
Contradictory information has been reported regarding the
reducibility of Ga2O3. Rodriguez et al. reported that Ga2O3/
Al2O3 is not reduced by hydrogen at 700 C, but the presence
of Pd on the catalyst can assist the reduction of Ga3+ to Ga+ and
reaction after observing that promotion with a base, such as

potassium, resulted in a decrease in activity.235
Several authors have performed DFT calculations in an eort
to understand the dehydrogenation reaction mechanism on
Ga2O3. Studies by Liu et al. regarding propane dehydrogenation
on a perfect Ga2O3 (100) surface revealed that the CH bond
is activated by a surface oxygen that abstracts an hydrogen from
the propane molecule, thus forming a propyl species, which
coordinates with a Ga surface site and a hydroxyl group. A
second hydroxyl group is formed via -hydrogen elimination,
but for these hydroxyl groups to recombine, either as H2O or as
H2, a high energy barrier needs to be overcome. This barrier is
much lower over a gallium hydridehydroxide species,
although the formation of the gallium hydride also involves a
10635
Chemical Reviews
Review
with CrOx (vide supra), where a weak metalsupport

interaction results in an increased reducibility. Indeed, Meitzner
et al. showed that Ga2O3/H-ZSM-5 could be reduced by
hydrogen or hydrocarbons at dehydrogenation reaction
conditions. Finally, regarding the reduction of Ga2O3, it is
important to note that Ga2O is gaseous and metallic gallium is
liquid at dehydrogenation reaction conditions.240
A number of studies have shown that the addition of CO2 to
the reactant feed improves the olen yield.241245 Nakagawa et
al. found that the addition of steam and CO2 decreased coke
deposition on a Ga2O3/TiO2 catalyst by the Boudouart
reaction, while also assisting ethylene desorption.241,242
However, other authors claimed that the increase in activity
was caused by the reverse watergas shift reaction
(RWGS).243,244 Additionally, Michorczyk et al. noted that
this eect could only be seen at relative low levels of coke
deposition.235 It was suggested that CO2 could also poison the
basic sites required for the dissociative adsorption of the alkane.
Additionally, Hensen et al. studied the promoting eect of
water with experiments and theory and observed an increase in
the conversion of propane by cofeeding water.246 The authors
attributed this promoting eect to the formation of partially
hydrolyzed gallium species, more specically, to the formation
of binuclear hydroxyl-bridged Ga3+ cations.
Dierent carriers have been used in gallium oxide-based
dehydrogenation catalysts, the inclusion of acid and basic sites
on the support being important to achieve a high conversion.
Xu et al. reported that in the absence of an oxidizing agent, the
conversion obtained using Ga2O3 supported on ZrO2, Al2O3,
and TiO2 was high, medium, and low, respectively, while SiO2
and MgO-supported Ga2O3 were inactive in alkane dehydrogenation.233 The authors proposed the activity to be dependent
on the presence of acid sites of medium strength on the catalyst
surface. However, a dierent trend was observed when CO2
was present, as conversion was observed to be high on Ga2O3/
TiO2 and low on Ga2O3/ZrO2 and Ga2O3/Al2O3. This was
attributed to the two contrary roles of CO2, which on the one
Figure 25. Concentration of Lewis acid sites versus propane

conversion measured 15 min after the start of the PDH reaction on
GaxAl10xO15. A clear correlation between the conversion and the
surface density of weak Lewis acid sites is observed. For these
experiments, CO2 is used to improve conversion. Reprinted from ref
232, copyright 2008, with permission from Elsevier.
Ga0 at temperatures between 300 and 550 C and form an alloy

(in which Ga0 is immobilized). Indeed, it has been reported
that the reduction of Ga2O3 can be promoted by the proximity
of other, more easily reduced metals, such as Pt and
Pd.131133,238 For instance, on a PtGa/H-beta catalyst,
reduced gallium hydrides were observed after hydrogen
reduction at 550 C.239 It has also been proposed that the
reducibility of these species depends on the interaction of
gallium oxide with the catalyst support, which is also the case
Table 7. Summary of the Catalytic Data of Gallium Oxide-Based Dehydrogenation Catalysts Discusseda
reaction
temp
(C)
space
velocity
(h1)
(1) 1.7 wt % Ga2O3/SiO2
550
(2) Ga2O3/MTS (mesoporous

silica) ratio Ga/Si = 0.05
(3) -Ga2O3
550
(4) Ga2O3
600
(5) 5 wt % Ga2O3/ZrO2
600
WHSV =
0.97 h1
WHSV =
0.33 h1
WHSV =
0.15 h1
WHSV =
1.2 h1
WHSV =
0.3 h1
catalyst
500
(6) 5 wt % Ga2O3/TiO2
(7) Ga8Al2O15
(8) 5 wt % Ga2O3/H-ZSM-5
(w Si/Al = 97)
(10) 0.001 wt % Pt, 1.2 wt %
Ga/Al2O3
500
650
590
WHSV =
0.3 h1
WHSV =
0.36 h1
GHSV =
400 h1
feed composition
selectivity
(%)b
conversion
(%)b
specic
activity (s1)c
kd (h1)
catalyst
lifed (h)
ref
C3H8 = 10, Ar = 90
2320
8680
4.97 104
0.036
249
i-C4H10 = 10, He =
90
C3H8 = 2.5, N2 =
97.5
C3H8 = 17, CO2 =
83
C3H8 = 2.5, N2 =
97.5
C3H8 = 2.5, CO2 =
5, N2 = 92.5
(a) C3H8 = 2.5, N2 =
97.5
(b) C3H8 = 2.5, CO2
= 5, N2 = 92.5
C2H6 = 3, CO2 = 15,
Ar = 82
C3H8 = 100
4640
58.3 to i-C4H8
5.52 105
0.092
2.67
230
3312
9595
6.95 107
0.21
229
313
9582
3.55 105
0.67
235
396; 328
74
1.90 105
0.77
233
73
1.54 105
0.56
51.722.5
91.698.2
2.39 106
0.168
232
49.733.1e
91.798.0e
2.30 106
0.089
23.714.5
85.993.7
2.84 105
0.0086
70
244
39
86
0.25
247
As some of these catalysts require an activation period, the maximum (as opposed to the initial) conversion is considered to determine the
deactivation rate. bFirst value is obtained at the start of the cycle, second at the end. cSpecic activity is dened as (mol olen formed)/(mol
Ga*t(s)). dCatalyst life = total time single cycle/experiment. eValues obtained at 15 min and 8 h.
10636
Chemical Reviews
Review
activity. The authors hypothesized that the carbon species

formed on the catalyst surface were in fact catalyzing the
reaction. The highest activity was observed at 700 C, a
temperature at which the carbon deposits display a very
ordered structure. In view of this, the authors decided to test
their hypothesis by using carbon nanobers (CNFs) as a
catalyst under similar conditions. Tellingly, CNFs proved to be
catalytically active. It was concluded that radicals formed on the
surface of coke facilitate the dehydrogenation reaction, albeit at
elevated temperatures the noncatalytic conversion of propane is
also relatively high. However, elevated temperatures also lead to
a drop in selectivity, as cracking becomes the dominant
reaction. Interestingly, the latter was not the case in the work of
McGregor et al., who reported a selectivity toward butenes of
80%. Furthermore, hybrid nanocarbon, comprising a diamond
core and a graphitic shell, displayed activity for the
dehydrogenation of propane, which was attributed to the
presence of either ketone groups or structural defects in the
graphitic material.251
In previous work, CNFs had been shown to be active in the
oxidative dehydrogenation of n-butane, although in this case
surface oxygen groups were believed to be the active sites.252
Furthermore, nitrogen doped carbon nanotubes (CNTs)
showed some activity in the oxidative dehydrogenation of
propane, although conversion and selectivity were relatively
low.253
Liu et al. prepared carbon-based ordered mesoporous
systems that showed activity in the dehydrogenation of propane
in the absence of steam or oxidants.254 The active sites were
proposed to be oxygen-containing groups (CO), which are
formed during a pretreatment step involving nitric acid. The
catalyst shows remarkable long-term stability, displaying
conversion and selectivity values of 45% and 85%, respectively,
after 100 h of time on stream at 600 C. Experiments in which
carbon nanotubes were tested under similar conditions resulted
in low activity.
Neylon et al. studied the activity of supported early transition
metal (groups 5 and 6) nitrides and carbides in butane
dehydrogenation,255,256 as the catalytic properties of these
materials are believed to be similar to those of platinum.
Interestingly, Mo, V, and W carbides and nitrides showed
activity in both dehydrogenation and hydrogenolysis, displaying
specic activities similar to those of a commercial PtSn/Al2O3
catalyst. Niobium carbide and nitride did not show any activity.
The group 6 compounds tested were found to be more active,
albeit they primarily catalyzed hydrogenolysis. In Table 8 are
shown the catalytic data of the carbon-based dehydrogenation
catalysts as reported by McGregor et al. and Liu et al.
In addition, silica supported metal suldes, such as ZnS, FeS,
CuS, CoS, MnS, NiS, and MoS, were tested for the
dehydrogenation of isobutane, and were found to perform
signicantly better than their respective oxides. The materials
show higher initial yields than commercial CrOx and PtSn
catalyst, but they deactivate over a few hours due to sulfur
leaching.257
hand removes H2 from the surface of the catalyst by the reverse

watergas shift reaction, while on the other hand absorbs
strongly on the basic sites of the support. Thus, the low
conversion observed when zirconia or alumina were employed
as the support can be ascribed to the relatively high amount of
basic sites on these oxides. Notably, the structure of the support
is also important, as a mesoporous GaAl mixed oxide
exhibited an increased resistance to catalyst coking.245
Work on the promotion of GaOx-based dehydrogenation
catalysts appears to be limited, although a patent by Iezzi et al.
claims that a catalyst comprising 1.2 wt % Ga2O3, <99 ppm of
Pt, and 1.6 wt % of SiO2 on an Al2O3 support represents a
highly active formulation in the dehydrogenation reaction.247
Indeed, when this catalyst is used for PDH, a conversion of
39% and a selectivity of 85% can be obtained after 150 h of
time-on-stream at 580 C, with frequent regenerations at 650
C. These values are the highest among those reported for Gabased dehydrogenation catalysts. However, which metal, Pt or
Ga, is predominately responsible for the high activity of this
catalyst, is not discussed.
In recent work by Sattler et al., a 3 wt % Ga/Al2O3 catalyst
was promoted with 1000 ppm of Pt and 0.25 wt % of K to
obtain a highly active, selective, and stable PDH catalyst.248
Evidently, the function of K was to reduce coke formation by
poisoning Brnsted acid sites. However, Pt is believed to act as
a promoter for the GaOx. Indeed, albeit 1000 ppm Pt/Al2O3
was found to be virtually inactive, 1000 ppm of Pt in
combination with 3 wt % Ga aorded conversions close to
the equilibrium and an propylene selectivity <95%. In the
absence of Pt, propane conversion is more than halved. It was
suggested that the role of Pt is to assist hydrogen desorption,
which facilitates the regeneration of the gallium hydride
hydroxide species. The catalytic data on the Ga-based
formulations discussed above are summarized in Table 7.
4.6. Carbon-Based Catalysts
Carbon materials have been tested as the support in catalysts

employed for various reactions, mostly because of their high
surface area and the fact that these materials are both
inexpensive and environmentally friendly. Moreover, discrete
functionalities of activated carbon can catalyze an array of
reactions, including alkane dehydrogenation.250 McGregor et al.
have reported that a VOx/Al2O3 catalyst completely encapsulated by coke deposits (see Figure 26) is active in butane
dehydrogenation.215 The fact that catalytic activity remained
the same despite complete coke encapsulation is remarkable, as
this extent of coking usually results in a drastic decrease in
4.7. Other Formulations

Figure 26. TEM images of VOx/Al2O3 after butane dehydrogenation
at 550 C (left) and 700 C (center and right). The high magnication
image of the catalyst deactivated at 700 C clearly shows that the
catalyst is completely enveloped in a carbon layer a few nanometers
thick. Nevertheless, this catalyst still displays a relatively high activity at
700 C. Reprinted from ref 215, copyright 2010, with permission from
Elsevier.
Indium is known to share many of the catalytic properties of

gallium, a prominent example being the activity that both
metals display in NOx reduction.258 These similarities are
relatively unsurprising, particularly taking into account the
adjacency of Ga and In within the same group of the periodic
table. Interestingly, Halasz et al. have tested the reactivity of
10637
Chemical Reviews
Review
Table 8. Summary of the Catalytic Data of Carbon-Based Dehydrogenation Catalysts

catalyst
(1) coked 3.5 wt %
VOx/Al2O3; CNF
(2) mesoporous carbon
reaction
temp (C)
700
600
space velocity (h1)

GHSV = 72 (at a
pressure of 0.5 bar)
WHSV = 0.59
feed
composition
conversion
(%)a
C4H10 = 3,
N2 = 97
C3H8 = 5,
N2 = 95
50; 41
69.344.5
selectivity (%)a
80; 80 to 1- and
13-butene
62.285.1
specic
activity (s1)b
8.21 105
kd
(h1)
0.010
catalyst
lifec (h)
ref
215
100
254
a
c
First value is obtained at the start of the cycle, second at the end. bSpecic activity is dened as (mol olen formed)/(mol CO group*t(s)).
Catalyst life = total time single cycle/experiment.
Figure 27. Propane conversion on a series of 20% In80% metal mixed oxides showing InAlOx to be the most active formulation (left). Albeit other
In:Al ratios were tested, 20% In80% Al aorded the highest propylene yield. The TPR proles of dierent InAl mixed oxides after H2 reduction at
500 C and subsequent treatment with CO2 at 600 C are also shown (right). The peak at around 700 C (labeled ) represents the reduction of
In2O3 to In0; the total hydrogen consumption showing the In2O3 is nearly completely reduced to In0. The peak at 300 C (labeled ) corresponds to
the reduction of small In2O3 domains in the mixed oxide and suggests the formation of small metallic In particles. The absence of this peak in the top
trace suggests that CO2 cannot reoxidize these particles at 600 C. The graph to the left is reprinted from ref 260, copyright 2010, with permission
from Elsevier. The graph to the right is reprinted from ref 262, copyright 2010, with permission from Elsevier.
than SiO2, as the interaction of the indium oxide was stronger

with the former carriers, resulting in a better distribution of
small nanoparticles.
CO2 was reported by the same authors to have a benecial
eect, which was ascribed to the occurrence of the reverse
watergas shift reaction (RWGS). Furthermore, suciently
high CO2:hydrocarbon ratios proved capable of retarding
catalyst deactivation.
Albeit supported iron oxides have been known to be active in
ethylbenzene dehydrogenation since 1946,264 it was not until
the 1990s that Geus and co-workers investigated the activity of
supported iron oxide in the dehydrogenation of 1-butene to 1
3 butadiene.265268 The authors determined that Fe3+ is the
catalytically active species. The catalysts deactivate rapidly due
to coke deposition and reduction of Fe2O3 to Fe3O4. However,
catalyst stability is signicantly enhanced by the addition of
potassium. On the one hand, K stabilizes the active phase by
forming KFeO2, which donates electrons to the active iron
species and prevents their reduction. On the other hand, the
acidic groups responsible for coke formation are poisoned by
potassium ions. Optimal stability was obtained with a FeOx
monolayer, regardless of the type of support employed. An
propane on Ga- and In-loaded ZSM-5 and found both catalysts

to be active in PDH and display similar activity and selectivity
values.259 However, both Ga/ZSM-5 and In/ZSM-5 did not
exclusively yield propylene, but also larger aromatic molecules.
Whereas dehydrogenation and aromatization activity were
found to be associated with Lewis acidity, cracking and coke
formation were related to Brnsted acidity.
Chen et al. mixed In2O3 with other metal oxides in an eort
to obtain an active PDH catalyst.260 A mixed oxide consisting of
20% In2O3 and 80% Al2O3 showed the highest propylene yield,
as shown in Figure 27. Small In2O3 clusters on the surface of
the catalyst were reduced under reaction conditions to form
small metallic In particles. Because no correlation between
surface acidity and reactivity was observed, these metallic
particles were believed to be the species active in dehydrogenation. The reduction of these particles is a slow process during
reaction, and thus a considerable amount of time is required to
attain the optimal yields. By prereducing the catalyst with
hydrogen, high conversions were immediately obtained. Similar
conclusions were reached in a study involving a series of oxidesupported In2O3-based catalysts prepared by incipient wetness
impregnation.261 Al2O3 and ZrO2 were more suitable supports
10638
Chemical Reviews
Review
Table 9. Summary of the Catalytic Data of the Indium and Iron-Based Dehydrogenation Catalysts Discusseda
catalyst
reaction
temp (C)
(1) In2Al8O15
600
(2) 10 wt % In2O3/
Al2O3
(3) 6 wt % K3.1 wt %
FeO2/MgO
(4) 3.4 wt % Fe3O4/
Cact
600
600
650
space
velocity
(h1)
WHSV
0.15
WHSV
0.15
WHSV
0.35
WHSV
=
=
=
=7
feed composition
conversion (%)b
C3H8 = 2.5, CO2 =

10, N2 = 87.5
C3H8 = 2.5, CO2 =
10, N2 = 87.5
1-C4H8 = 5, H2O =
30, N2 = 65
i-C4H10 = 30, CO2 =
70
35.729.3
(3 h, 8 h)
212421
(30 m, 3 h, 8 h)
3925
3220
(0.5 h, 4.5 h)
specic
activity (s1)c
kd
(h1)
catalyst
lifed (h)
ref
76.577.5
2.22 106
0.058
260
298385
5.41 106
0.034
261
0.67
20
266
selectivity (%)b
8082 to 13butadiene
7165
(to i-C4H8)
0.16
4.5
250
Because catalyst reduction by the hydrocarbon stream tends to be a prerequisite for dehydrogenation activity, the table shows both optimal and
nal conversion values along with the corresponding selectivity. The times at which these conversions are obtained are shown in parentheses. bFirst
value is obtained at the start of the cycle, second at the end. cSpecic activity is dened as (mol olen formed)/(mol ln*t(s)). dCatalyst life = total
time single cycle/experiment.
Scheme 2. Dehydrogenation Mechanism As Proposed by Horiuti and Polanyi on a Pt Surfacea
The intermediates depicted have been observed by IR spectroscopy during the hydrogenation reaction at low temperatures and are supported by
theoretical calculations. The dissociative adsorption of propane may also take place by the abstraction of the -hydrogen, forming a 2-propyl
intermediate, but for brevity this pathway is not shown. High paran pressures favor the formation of propylidyne species (step h-j), which are coke
precursors. These species are relatively stable and can only be hydrogenated when high concentrations of hydrogen are present on the Pt
surface.292297
The conversion of light alkanes to aromatics on Zn

supported on H-ZSM-5 zeolites has received considerable
attention.100,271275 Isolated Zn2+ species with a tetrahedral
orientation were deemed to be the active sites, with a reaction
mechanism similar to that reported on GaOx-based dehydrogenation catalysts.276280 These species resided in the cation
exchange sites of the zeolite, eectively providing the countercharge for two Al species. The proximity of Al species was
found to be important, as calculations show basic Al-bound
oxygen is required to perform the CH cleavage and form an
AlOH group. This is important, as, similar to Ga, ZnH and
ZnOH groups are dicult to regenerate. Furthermore, these
species were found to be highly stable, which is important as
volatile Zn species may be formed at dehydrogenation reaction
conditions.
The activity of zinc titanate in the dehydrogenation of light
olens has also been studied, the catalysts being prepared either
as hydrogels or as mixed oxides. Chen et al. investigated mixed
oxides with dierent zinc to titanium ratios and found that a
Zn:Ti ratio of 2 resulted in the most active catalyst. Cubic
Zn2TiO3 was deemed to be the active species in dehydrogenation, although the overall activity of these catalysts was found
to be limited.281283 More recently, isolated Zn2+ species
deposited on a silica support were found to be active in
interesting sulfate-promoted FeOx/Al2O3 catalytic system was

reported by Sun et al., which showed excellent activity in
propane dehydrogenation, accompanied by relatively high
stability.269 The sulfate groups were found to withdraw
electrons from Fe, enhancing the acidity of Fe sites and
improving activity. Unfortunately, the sulfate groups were not
completely stable, and the loss of sulfate groups as SO2 resulted
in the irreversible deactivation of the catalyst by carburization
and sintering.
In a recent publication by Yun et al., an active propane
dehydrogenation catalyst was obtained by incorporating Fe ions
in the framework of ZSM-5 zeolites. The catalytic activity was
explained by the redox properties of these sites (Fe2+ and Fe3+
redox cycle) that resulted in the formation of radical
intermediates.270 Shimada et al. investigated the use of activated
carbon, promoted by Fe, as a catalyst for the dehydrogenation
of isobutane.250 In the absence of iron, activated carbon was
only able to aord an isobutene yield of 10%, albeit this value
increased to 40% upon the addition of 3.5 wt % Fe3O4 in the
presence of CO2. In addition to dehydrogenation, the iron
oxide also catalyzes the RWGS reaction. Indeed, XRD results
show that the iron oxide is reduced by the hydrocarbon feed,
and subsequently reoxidized by the CO2 in the feed.
10639
Chemical Reviews
Review
Figure 28. Energy proles for propane dehydrogenation to propylene on Pt (111) (black), Pt3Sn (111) (blue), and Pt2Sn (111) (red) surfaces, as
obtained by DFT calculations. The intermediates are shown along with their corresponding energies. The recombination of the hydrogen adatoms is
not considered.114 The values of potential energy for all intermediates are given relative to the potential energy of a propane molecule in the gas
phase and a clean catalyst surface.
propane dehydrogenation.284 The active sites were determined

to be coordinatively unsaturated Zn2+ Lewis acids sites located
on a tri-membered siloxane ring on the silica support. During
the catalytic reaction, the Zn is transformed to tricoordinated
species due to the temperature increase. DFT calculations
suggested a reaction mechanism akin to that observed over
CrOx and GaOx catalysts.
In Table 9 are shown the catalytic activity data of the indiumand iron-based dehydrogenation catalysts discussed.
However, because the dehydrogenation reaction is an

equilibrium-controlled reaction, compounds can react back
and forth leading to isotopic scrambling, which complicates the
interpretation of experimental results.
Infrared spectroscopy is another valuable tool to detect
intermediates formed during dehydrogenation reactions on Ptbased catalysts. Because of experimental limitations, this is done
through the study of the reverse hydrogenation reaction, which
takes places at considerably lower temperatures. This approach
assumes that the same reaction intermediates are encountered,
which needs to be veried due to the signicantly dierent
reaction conditions. During the adsorption of ethylene on both
Pt and PtSn surfaces at temperatures ranging from 100 to
27 C, several species are detected, including ethylidyne as well
as -bonded and di--bonded ethylene (the propylene
analogues of which are shown in Scheme 2).116,117,298300
Cremer et al. used sum frequency generation IR to elucidate
the role of these ethylene adspecies on a Pt (111) surface.301,302
During ethylene hydrogenation, ethylidyne was found to be a
spectator species, although one competing for active sites on
the catalyst surface. On the other hand, -bonded ethylene was
determined to be the main intermediate of ethylene hydrogenation, although di--bonded ethylene was also converted at
a lower rate. Ethyl groups are only detected at very high
hydrogen pressures, as these species readily convert to di-ethylene otherwise. Nevertheless, Zaera et al. have reported that
the reaction mechanism of ethylene hydrogenation involves a
-bonded ethylene and an ethyl intermediate.303 Furthermore,
when propylene adsorption on Pt (111) was studied by
reectionabsorption IR spectroscopy, most of the propylene
was found to be coordinated as di--bonded, albeit -bonded
propylene was found to predominate at high surface coverages.292 On the contrary, DFT calculations showed that
propylene is preferentially adsorbed in the disigma
mode.75,114 DFT calculations performed on Pt, Pt3Sn, and
Pt2Sn (111) surfaces during the PDH reaction have revealed
that the dissociative adsorption of the alkane is the rate-limiting
step. In the presence of Sn, this energy barrier is increased, as is
5. COMPARING DIFFERENT CATALYSTS

5.1. In Terms of Reaction Mechanism and Deactivation
Pathway
Even though the dehydrogenation reaction has been studied for

almost a century, the exact mechanism is still a topic under
debate, particularly because this reaction is catalyzed by a vast
array of materials as dierent as noble metals and transition
metal oxides. The so-called HoriutiPolanyi mechanism, which
was proposed in 1934, is most commonly used to describe
catalytic dehydrogenation reactions.285 This mechanism follows
LangmuirHinshelwood kinetics, where all of the surface sites
of the catalyst are considered to be identical. The Horiuti
Polanyi mechanism consists of four main steps: dissociative
adsorption of the paran, CH cleavage of a second hydrogen
atom, formation of a hydrogen molecule, and subsequent
desorption of both the hydrogen and the olen, as is shown in
Scheme 2ag.286 Notably, both CH cleavage
steps229,263,287289 and the dissociative adsorption of the
paran have been suggested as the rate-limiting step of the
dehydrogenation reaction.96,98,144,290
Several studies have been performed on Pt and PtSn
catalysts to gain a detailed understanding of the mechanism of
the dehydrogenation reaction. First, isotopic labeling experiments have been used to identify the reaction intermediates.
For example, cofeeding D2 during ethane dehydrogenation
resulted in a single deuterium being attached to the ethane
molecules, suggesting an ethyl reaction intermediate.291
10640
Chemical Reviews
Review
Scheme 3. Proposed Reaction Mechanisms for the Dehydrogenation of Ethane on CrOx Catalysts144,306,307
Figure 29. Energy prole for propane dehydrogenation to propylene on CrOx catalyst obtained by DFT calculations. The intermediates from
Scheme 3 are shown along with their relative energies. Note that for the second section of the scheme, a second propane molecule is
dehydrogenated.314,315 The values of potential energy for all intermediates are given relative to the potential energy of a propane molecule in the gas
reduces the amount of highly active Pt step sites, on which

reaction intermediates adsorb strongly.304 The latter observation is rendered particularly noteworthy by the fact that isotopic
labeling experiments, performed by Galvita et al. during ethane
dehydrogenation, have revealed that ethylene can form
methane and coke through cracking, the conversion proceeding
via an ethylidyne intermediate.132,290
The dehydrogenation mechanism on chromia-based catalysts
is somewhat dierent, as both the chromium and the oxygen
atoms of the oxide are believed to participate in the reaction.305
Quantum chemical methods (DFT) have been used by
Lillehaug et al. to study the mechanism of ethane dehydrogenation on a series of clusters containing a single Cr3+ ion bonded
to OH or OSi groups species, which served as a model for
the dierent active sites suggested in literature.306,307 The rst
step they proposed is CH activation via -bond metathesis
resulting in the formation of CrC and OH bonds, followed
by a -H transfer to a Cr site.306,308 The nal step proposed is
the formation of hydrogen and the desorption of the latter and
ethylene. An alternative mechanism, in which a CrH species
activates ethane by forming molecular hydrogen and an
shown in Figure 28, therefore decreasing the activity of the

catalyst.114,118 On Pt (111) the dehydrogenation may proceed
by a 1- or a 2-propyl intermediate, while on the PtSn alloys
the 1-propyl intermediate is favored. The positive eect of Sn
on catalytic activity is explained by the facile desorption of
propylene, which hinders further dehydrogenation of the
adsorbed propylene.
Microcalorimetry has aorded complementary information
on the dehydrogenation mechanism over Pt-based catalysts, a
lower heat of adsorption of ethylene being detected on the Pt
Sn relative to the Pt surface (94135 and 157 kJ mol1,
respectively).117,296 These relatively high heat of adsorption
values are generally indicative of the dissociative adsorption of a
hydrocarbon. Density functional theory (DFT) calculations on
the adsorption of ethane on dierent Pt and PtSn alloy
surfaces, (111) and (211), have revealed that electron transfer
takes place from Sn to Pt. Keeping in mind the fact that Sn
addition reduces the heat of adsorption, the formation of
ethylidyne becomes energetically less favorable because of this
electron transfer.117,118,289,304 In addition, by studying these
dierent PtSn alloys, it was shown that the presence of Sn
10641
Chemical Reviews
Review
Figure 30. Energy prole for propane dehydrogenation to propylene on GaOx catalyst, as determined by DFT calculations on a Ga (100) surface.
Intermediates analogous to those displayed in Scheme 3 are shown along with their relative energies. Note that two pathways for the
dehydrogenation of propane are considered, one involving a radical intermediate and another an ethyl intermediate, the former having a lower
energy barrier.314,315 The values of potential energy for all intermediates are given relative to the potential energy of a propane molecule in the gas
Scheme 4. Proposed Pathway Linking the Formation of Strongly Adsorbed Propyl Species to That of an Aromatic Ring through
a Series of Cracking (ae), Deep Dehydrogenation (fi), and/or Dimerization (f,g,j,k) Reactions293,294,321a
Note that the adsorbed hydrocarbons shown in the scheme are examples of species formed on the Pt surface, and that they actually represent a
family of similar species.
unsaturated, and aromatic carbon species, along with acetates

and carboxylates, have been detected by diuse reectance
infrared Fourier transform spectroscopy (DRIFTS).192,197,311
This is consistent with a model in which the rst step of the
reaction is the abstraction of hydrogen from the alkane to form
an alkoxide species (ethoxide and isopropoxide being formed
from ethane and propane, respectively). However, the
reduction of the catalyst prior to the dehydrogenation reaction
led to a dramatic decrease in the amounts of acetates and
carboxylates detected, albeit the reaction rate remained
unaected. Evidently, the latter throws into question the role
of acetic and carboxylic species as reaction intermediates. An
alternative interpretation is that these species are formed during
the reduction of the Cr6+ species by hydrocarbons present in
the feed. Parenthetically, the involvement of hydrocarbons in
the formation of reduced active chromium sites, the
participation of both chromium and oxygen atoms in the
reaction, and the proposition that CH activation represents
adsorbed ethyl species, has been proposed on the basis of the

results of isotopic labeling experiments performed during
ethane dehydrogenation on a CrOx/Al2O3 catalyst.144 The two
main mechanisms proposed for the dehydrogenation of light
alkanes on chromia/alumina catalysts are shown in Scheme 3.
In these mechanisms, either the dissociative adsorption144,305 or
-H elimination is believed to be the rate-limiting step.306,309,310
Figure 29 shows two reaction cycles in an energy diagram, as
calculated by Lillehaug et al. Indeed, the energy barrier is the
highest for the initial CH activation, although this energy is
signicantly lower when the bond is activated by the CrH
intermediate (i.e., the right part of Scheme 3). Nevertheless,
hydrogen desorption is preferred, as the energy barrier is lower.
In situ FTIR spectroscopy has provided evidence of the
formation of acetaldehyde and acetone intermediates as ethane
and propane are adsorbed on a mesoporous chromiaalumina
catalyst.197 Furthermore, during the dehydrogenation of
propane on a CrOx catalyst, surface species including aliphatic,
10642
Chemical Reviews
Review
Scheme 5. Proposed Mechanisms for the Growth of Polyaromatic Species by Either DielsAlder-type Addition, Propylene
Addition via Radical Chemistry, or Fusion of Small Domains into Larger Ones322,323a
The species will continue to grow in a similar fashion, and at a certain point are considered to be graphene-like sheets. The molecules in the scheme
are only simplied hydrogen-decient coke precursors that may be present on the surface of the catalyst. Such species are directly bonded to the Pt
surface and in reality will contain signicant amounts of hydrogen, either in hydrogen-terminated aromatic rings or in aliphatic side chains.
eect of coke is largely dependent on catalyst design and the

location of the coke deposits.320
The process of coke formation is very complex and not very
well understood. Brnsted acid and others active sites of the
catalysts serve as seeding points to which olens and parans
are added until large graphitic structures are formed. Indeed,
olens in the stream can easily react with the carbocation
intermediates formed on these acid sites. Additionally, strong
adsorption of reaction intermediates, such as the ethylidyne or
ethyne species discussed above,75,132,290 and/or products on
active sites can also induce polymerization and lead to the
formation of graphitic sheets. In Scheme 4, several pathways for
the formation of the rst aromatic ring structures from
propylidyne and similar molecules are proposed, which include
cracking, deep dehydrogenation, and polymerization reactions.
Coke species may grow from the resulting aromatic rings by
either DielsAlder-type addition, propylene addition to a
radical, or fusion of smaller domains, as shown in Scheme 5.
When these polyaromatic species are still relatively small, they
can also migrate from the metal to the support (where they can
undergo further oligomerization and condensation). In fact, the
addition of Sn to the Pt-based catalyst facilitates this migration
by weakening the binding of the hydrocarbon to the
metal.304,324,325 In addition, acid sites present on the support
may protonate these small polyaromatics, making them more
reactive toward gas-phase olens.322,326 Nevertheless, studies
on these processes are scant, which is why the mechanism
explaining the progression from adsorbed atomic carbon to
aromatic rings is still largely unknown. Therefore, more detailed
studies are needed to better understand this process. For
example, it would be benecial to gain a fundamental
understanding on the steps proposed in Schemes 4 and 5 to
develop strategies to prevent coke formation through these
routes.
The location of coke is very important, as coke on the
support has a milder negative eect on catalyst activity.320,324
However, it is important to note that the presence of carbon
deposits on the surface does not always negatively inuence
catalyst performance. Indeed, a low amount of coke deposition
the rst reaction step are somewhat reminiscent of the Phillips

catalyst and of the mechanism of ethylene polymerization
recently proposed by Coperet and co-workers over this
formulation.312,313 However, additional work will be necessary
to investigate the relationship between the mechanisms of
ethylene polymerization and ethane dehydrogenation, particularly in view of the dierent nature of the reactions and the
reaction conditions.
The reaction mechanism on GaOx has been studied by Liu et
al. by DFT calculations on a Ga (100) surface (Figure 30).231 A
mechanism similar to that shown in Scheme 3 was proposed,
except for the fact that the initial CH activation takes place
involving a radical, which can adsorb on the Ga site as an ethyl
species, or directly undergo the second CH cleavage step. The
desorption of both hydrogen and propylene was found to be
energetically unfavorable, which was used to explain the rapid
coking and the relative low activity of GaOx-based dehydrogenation catalysts. Indeed, while heating Ga2O3 in an ethane
atmosphere, the evolution of Ga-ethyl and hydroxyl surface
species was observed by DRIFTS at 150 C.316
To summarize, most of the experimental results suggest that
the dehydrogenation of light alkanes proceeds via the Horiuti
Polanyi mechanism, where either the dissociative adsorption of
the alkane or the second CH bond cleavage is the ratedetermining step. Evidence of fundamental dierences in the
reaction mechanism on metal oxide and noble metal catalysts is
lacking, although reports of mechanistic studies involving metal
oxide catalysts are admittedly scant.
All dehydrogenation catalysts deactivate with time-on-stream,
several processes being responsible for the deactivation of the
dierent types of catalysts discussed in this Review. However,
there has been a clear consensus in the literature since the mid1970s identifying coke deposition as the main cause of
deactivation.77,317319 Coke is a collective name given to a
large family of highly graphitized hydrocarbon molecules that
tend to form on the catalyst surface during reaction. Coke
accumulation eectively hinders the diusion of reactants to the
catalyst surface by blocking both pores and active sites, which
results in a drop in catalytic activity. Therefore, the detrimental
10643
Chemical Reviews
Review
Figure 31. Propylene yield during propane dehydrogenation (left) and isobutene yield during isobutane dehydrogenation (right) versus the
deactivation rate of catalysts included in Tables 39. Each data point is labeled with the number included in the rst column of the relevant table, to
distinguish between dierent catalysts. Catalyst systems displaying high yields combined with low deactivation rates are preferred.
Figure 32. Yield of propylene and isobutene versus WHSV (left and right, respectively) of the catalysts included in Tables 39. High yields obtained
at high space velocities are preferred. Therefore, because Pt-based catalysts can aord comparable yields to catalysts based on metal oxides while
operating at higher space velocities, they are deemed superior in this respect.
formation,324,327,328 whereas high partial pressures of olens

and parans exacerbate coke deposition. Even the nature of the
alkane reactant has an eect, as coke formed during propane
dehydrogenation on a PtSn/Al2O3 catalyst is reported to be
more detrimental than coke formed during ethane dehydrogenation (although such a comparison is dicult because these
reactions run at dierent temperatures).132 Alternatively, coke
deposition can be reduced by adding a promoter element, the
addition of Sn to Pt-based dehydrogenation catalysts being the
best-known example.
has been observed to increase the catalytic activity by

facilitating the adsorption of the alkane near the active
dehydrogenation site.320 Furthermore, highly active sites
responsible for side reactions, such as hydrogenolysis, are the
rst to be poisoned by coke, resulting in higher selectivity to
the olen product. Coke becomes detrimental when large
graphitic sheets are formed that cover the catalytically active
sites.173,180 To address this coke-induced deactivation, a
regeneration step under an oxidative atmosphere has been
incorporated to industrial dehydrogenation processes. This
process is relatively straightforward, as coke is easily combusted
in air at dehydrogenation reaction temperatures. The
ambivalent nature of coke deposits is further highlighted by
the fact that their presence on deactivated catalysts is essential
in the industrial processes that use coke combustion as the
main source of the heat required to run the dehydrogenation
reaction.
Several strategies have been applied to control coke
formation. For example, coke deposition can be dramatically
reduced by using a support with little to no Brnsted acidity, or
by poisoning the acid sites. Changes in the reaction conditions,
such as cofeeding steam, hydrogen, or CO2, can reduce coke
5.2. In Terms of Propane and Isobutane Dehydrogenation

Performance
Given that a large variety of catalysts are active in the

nonoxidative dehydrogenation of C2C4 parans, as documented in Tables 39, a comprehensive comparison of the
activity and stability of these catalysts is desirable, particularly
because said comparison stands to reveal qualitative dierences
between various formulations. We have attempted this
comparison focusing on three performance criteria, olen
yield (obtained at the start of the experiment), stability
(expressed as the deactivation rate), and the space velocity at
which the rst two criteria were measured (expressed as
10644
Chemical Reviews
Review
WHSV). In this approach, the eects of various factors, such as

pressure, temperature, preparation method, promoter, and
catalyst support, are not considered. The reasons to ignore
these eects include the fact that this information is missing in
many articles (pressure being a notable example), the absence
of a clear trend (as tends to be the case with temperature), and
an excessively large variance within a single variable to allow a
meaningful comparison (catalyst preparation being an instance
of such a variable). It is important to note that catalysts were
only included in this comparison if the space velocity,
deactivation rate, and initial yield could be determined from
the information provided in the original source. Also of note is
the fact that the performance criteria compared were measured
during propane or isobutane dehydrogenation, which are the
reactions that have been researched most intensively.
In Figure 31, the yields of propylene and isobutene are
plotted versus the deactivation rate of Pt-, CrOx-, GaOx-, and
In2O3-based catalysts. Moreover, in the plot showing PDH
results, the data points representing the catalytic performance
of dierent catalyst types are grouped around similar values, the
best initial yields being similar for Pt and CrOx and slightly
higher for GaOx. Additionally, PDH results show that the
dierent catalyst types deactivate at dierent speeds, with Pt
being particularly stable, CrOx deactivating rapidly, and other
systems deactivating even faster. Another interesting observation stemming from the PDH data is that within each type of
catalyst high yields often coincide with relatively low
deactivation rates, which is counterintuitive. Although considerably fewer data points are available for the isobutane
dehydrogenation, the Pt-based catalysts also display a superior
stability as compared to the metal oxides at similar yields.
Notably, the VOx-based catalysts show excellent yields (40%),
albeit they deactivate very fast.
Space velocity is another important factor that has been
considered to compare the performance of dierent dehydrogenation catalysts. In Figure 32 are plotted propylene and
isobutene yield aagainst the weight hourly space velocity
(WHSV).
To be commercially attractive, dehydrogenation catalysts
must provide high olen yields at high space velocities. In other
words, catalysts should display high specic activity, which is to
say good conversion at short contact times. Figure 32 illustrates
that while the yields obtained over GaOx-, CrOx-, and Pt-based
propane dehydrogenation catalysts are comparable, GaOx-,
CrOx-, and Pt-based catalysts operate at relatively low,
intermediate, and high space velocities, respectively.
Taking into account that Pt-based catalysts typically have low
loadings (<1 wt %), whereas GaOx- and CrOx-based catalysts
generally have loadings between 5 and 20 wt %, it is clear that
on a weight basis Pt represents the most active formulation in
PDH. This is conrmed when the average specic activities
from Tables 39 are compared to one another. Indeed, for the
dehydrogenation of propane on Pt-based catalysts, the average
specic activity is 2.86 101 s1, which is considerably higher
than the average specic activities over CrOx and GaOx, which
are 7.49 105 and 8.17 105, respectively. In Scheme 6, the
proposed active sites for propane dehydrogenation of these
three catalysts are shown. In the case of platinum-based
catalysts, any Pt0 atom located at the surface is believed to
catalyze the reaction, while in CrOx- and GaOx-based catalysts
Cr3+ and Ga3+ must be coordinatively unsaturated to display
catalytic activity. This similarity between the active sites of Gaand Cr-containing catalyst, and their dierence to Pt, may
Scheme 6. Proposed Active Sites for the Pt-, CrOx-, and

GaOx-Based Propane Dehydrogenation Catalysts, the Order
of Magnitude of Their Calculated Specic Activity, and the
Reaction Rates (k/A) Calculated by DFT
therefore explain the relative specic activity values of these

formulations.
However, it should be noted that specic activity values are a
very rough approximation of the real TOF of these catalysts. In
case of Pt catalysts, all Pt (not only the surface Pt) is taken into
account in our specic activity calculation, and additional errors
are introduced for the metal oxides. In the case of CrOx
catalysts, only coordinatively unsaturated Cr3+ species are active
in the dehydrogenation reaction, yet our specic activity
calculations take into account a considerable amount of species
that are either coordinatively saturated or in a dierent
oxidation state. Although Ga3+ is likely the only stable species
during the dehydrogenation reaction, our specic activity
calculation involving GaOx catalysts considers a signicant
amount of the Ga that is either present in the framework of the
support or incorporated in larger Ga2O3 crystallites. Nevertheless, the dierence between the specic activity calculated
for Pt and the metal oxides is in the range of 4 orders of
magnitude, which cannot be explained solely by the errors
introduced in our calculations. For the dehydrogenation of
isobutane, the same dierences of specic activities between the
Pt and metal oxide catalysts are observed.
It is important to mention that catalysts tend to deactivate
faster when higher space velocities are used. Notably, metal
oxide-based catalysts deactivate faster than the Pt-based
formulations, despite the fact that the latter are commonly
used at higher space velocities, which highlights the remarkable
stability of Pt catalysts.
The dierences in the specic activity values described
previously could be explained by comparing the activation
energies calculated by DFT methods previously discussed in
section 4. In these calculations, the energy barrier for the rate
limiting step on PtSn system is lower (109 kJ mol1) than in
the CrOx and GaOx catalysts (161 and 163 kJ mol1,
respectively). Ea aects the reaction rate, as described by the
Arrhenius equation:
k = A eEa / RT
where k is the reaction rate constant, A is the pre-exponential
factor, R is the gas constant, and T is the temperature. By
calculating k/A at a temperature of 600 C and assuming A is
comparable for the three systems considered, the rate constant
is 103 times higher for the PtSn catalysts as compared to the
metal oxide catalysts. The dierence in the calculated specic
10645
Chemical Reviews
Review
Figure 33. To the left, a schematic representation of the pilot-scale reactor used for propane dehydrogenation is shown. Two UVvis probes are
inserted in the reactor to visualize the deactivation of the CrOx/Al2O3 catalyst as a function of time and catalyst bed height (the spectra acquired are
shown in the center of the gure). This technique is excellent to visualize the coke-induced darkening of the catalyst extrudates from green to black,
as is shown in the photographs to the right. Reproduced from ref 198, copyright 2013, by permission of The Royal Society of Chemistry.
Al2O3 (Oleex) catalysts. These catalysts and technologies have

been advanced to a point where the reaction can be run
continuously for several years. Nevertheless, current dehydrogenation technologies still oer room for improvement, and
thus further advances are needed to lower noble metal loading,
increase catalyst stability, and/or improve the energy eciency
of the dehydrogenation process. Moreover, steps will need to
be taken to overcome the reactions biggest drawbacks, that is,
the limited equilibrium conversion and the prevalence of side
reactions at high temperatures.
Although alternative catalysts, most notably metal oxides
such as GaOx, MoOx, and VOx, have received signicant
attention over the last decades, a practicable alternative to the
PtSn and CrOx systems has not yet been developed. Indeed,
these alternative formulations deactivate too fast, mainly due to
coke deposition and to the fact that their activity cannot be fully
recovered by a regeneration step. Thus, further research will
also be necessary to address these issues.
Nevertheless, advancements can also be expected in other
areas. Oxidative dehydrogenation does not have the thermodynamic limitations shown by nonoxidative dehydrogenation.
Because an oxidant combusts the hydrogen formed during
reaction, full conversion of the paran can be achieved. O2 and
CO2 have been the most investigated oxidizing agents, each
having distinct advantages and disadvantages. In the case of O2,
coke deposition is nonexistent, the reaction temperature can be
lowered to about 400 C, and the dehydrogenation reaction
becomes exothermic (H2 + 1/2O2 H2O (H0298 = 242 kJ
mol1)). Nevertheless, new challenges arise as overoxidation of
the hydrocarbons in the feed must be avoided to maintain high
olen selectivity.252,329,330 CO2, being a much milder oxidant,
can eectively consume H2 under dehydrogenation conditions,
activity values for propane dehydrogenation on the Pt and

metal oxide catalysts (Tables 3, 4, and 7) is approximately 104,
as is shown in Scheme 6. The higher activity observed for the
Pt-based catalysts may therefore be explained by the
signicantly lower energy barrier of the rate-limiting step.
However, care must be taken when comparing these Ea values,
as the methods used for obtaining these are not identical.
Furthermore, these simulations assume model surfaces that do
not resemble real catalyst systems, and therefore the results are
not necessarily representative of the experimentally observed
reaction rates. Nevertheless, these combined theoretical and
experimental data are very encouraging and highlight the need
for more extensive studies on the reaction rate on Pt- and metal
oxide-based dehydrogenation catalysts.
To summarize, although metal oxide-based catalysts can
aord yields comparable to those obtained using Pt-based
formulations, the latter can operate at higher space velocities
and exhibit lower deactivation rates than metal oxide catalysts.
6. CONCLUSIONS AND EMERGING AREAS OF

RESEARCH
Because of their widespread use as feedstock in the chemical
industry, the demand for olens is expected to remain high for
the foreseeable future. Additionally, the recent surge in shale
gas production has rendered the cost of natural gas very
economical as compared to that of crude oil. Therefore,
alternative methods for obtaining olens, dehydrogenation and
MTO, have become very competitive relative to more
traditional methods, such as uid catalytic cracking and steam
cracking. Indeed, dozens of new PDH projects have been
announced in the past few years. The installations proposed or
being built will employ either CrOx/Al2O3 (Caton) or PtSn/
10646
Chemical Reviews
Review
These measurements show that the duration of the

dehydrogenation and regeneration phases could be dynamically
adjusted to keep the catalyst in its most active state, as
moderate amounts of coke are believed to assist the
dehydrogenation reaction while excessive amounts lead to
catalyst deactivation (vide supra). The latter would undoubtedly represent an improvement to the way in which reversible
deactivation is currently addressed. As an additional advantage,
UVvis spectroscopy can be used to monitor the long-term
irreversible deactivation of the catalyst. Indeed, because the
relative amounts of active chromia and of chromium that has
become irreversibly deactivated due to its migration into the
support can be measured by UVvis spectroscopy, this
technique provides a means to determine when the catalyst
needs replacing.
This analysis has been taken a step further by designing,
constructing, and testing probes for operando Raman spectroscopy, which were used to obtain information on the
chemical nature of the coke deposits formed during reaction.
Furthermore, by using the operando UVvis spectra as a
measure for catalyst darkening, a quantitative analysis of the
coke deposits could be performed during the initial stages of
the dehydrogenation reaction by operando Raman spectroscopy.49,180,346
Therefore, the monitoring of industrial chromiaalumina
catalysts by combined time-resolved UVvis and Raman
spectroscopies can be employed to improve the overall
eciency of future alkane dehydrogenation processes.
although it is not reactive enough to oxidize the hydrocarbons

in the reaction mixture. However, much less heat is formed
when CO2 is used, so the dehydrogenation reaction remains
endothermic (H2 + CO2 H2O + CO (H0298 = 41 kJ
mol1)), as the RWGS does not provide enough heat to
compensate for the highly endothermic dehydrogenation
reaction. The cofeeding of CO2 may result in additional coke
deposition, due to the increased CO and decreased H2 partial
pressures brought about by the reverse watergas shift
reaction. On the other hand, coke may be gasied by CO2.
Finally, CO2 competitively adsorbs on the surface, specically
on the basic sites of the catalyst materials, which can negatively
aect their performance.162,175,331 In short, nding a suitable
combination of catalyst material and oxidizing agent will be
pivotal in the development of an eective reaction system for
the oxidative dehydrogenation of parans.
An alternative solution lies in the use of solid oxygen carriers,
such as CeO2, SbO4, In2O3, WO3, PbOx, Tl2O3, and BiO3,
capable of selectively oxidizing the hydrogen evolved by the
SHC reaction.332334 Noble metals, such as Pt and Au, can also
catalyze the SHC process, but in this case oxygen is cofed to
facilitate the hydrogen combustion.335,336 Because the SHC
reaction consumes the lattice oxygen of metal oxide catalysts,
the latter require frequent regeneration with oxygen. Moreover,
the reaction rate is relatively low and some of these metal
oxides are highly toxic. These issues need to be resolved for this
approach to become viable for industrial application.
Another method of removing the hydrogen from the reaction
mixture is the introduction of membranes to the reactor system.
This approach allows hydrogen to diuse out of the reactor,
shifting the equilibrium toward the desired olen product. 337344 A large variety of membranes have been
investigated, such as Pd-coated mesoporous alumina, PtAg
composites, and membranes involving either molybdenum- or
carbon-based molecular sieves. The inverse relationship
between permeability and selectivity of the membranes is a
serious problem, as both qualities are desired. Catalyst
deactivation represents an additional challenge, as coke
deposition increases when the hydrogen partial pressure
drops. Nevertheless, nearly all studies report increased olen
yields when a membrane system is used.
It is very important to heed the interplay between catalyst
deactivation and regeneration in industrial settings. Indeed, for
both the Pt and the CrOx formulations used industrially, the
elimination of coke from the surface of the spent catalysts may
result in irreversible catalyst deactivation if the regeneration
conditions are not carefully controlled. Currently, this is done
in a very crude manner, by monitoring the gas composition at
the reactor outlet. However, recent reports suggest that in the
near future, operando spectroscopy will provide the means for
more sophisticated process control.198 Indeed, Raman/UVvis
spectroscopy can be used to monitor noninvasively the amount
and type of coke as the latter forms and accumulates on the
catalyst surface.173,180,345,346 As is shown in Figure 33,
spatiotemporal dierences in the coke deposition on a catalyst
bed consisting of CrOx/Al2O3 extrudates were visualized in a
pilot scale reactor using two UVvis probes, which collected
spectra in a continuous fashion during propane dehydrogenation.198 Coke deposition was observed to occur faster at the
top of the catalyst bed, which was attributed to the temperature
gradient measured in the reactor. Furthermore, coke
combustion proceeded in an inhomogeneous fashion, coke at
the top of the catalyst bed being combusted rst.
AUTHOR INFORMATION
Corresponding Author
*E-mail: b.m.weckhuysen@uu.nl.
Notes
The authors declare no competing nancial interest.

Biographies
Jesper J. H. B. Sattler was born in Veenendaal, The Netherlands, in

1986. He obtained his masters degree in chemistry at Utrecht
University. His thesis was on the subject of the intergrowth of zeolite
materials studied by in situ microspectroscopy under the supervision
of Dr. Marianne H. F. Kox and Prof. Bert M. Weckhuysen. He
continued working under the guidance of Prof. Bert M. Weckhuysen
for his Ph.D. research, which focusses on the deactivation of propane
dehydrogenation catalysts.
10647
Chemical Reviews
Review
Bert M. Weckhuysen was born in Aarschot, Belgium, in 1968. After

receiving his Ph.D. in 1995 with Prof. Robert Schoonheydt from
Leuven University, he worked as a postdoctoral fellow with Prof. Israel
Wachs at Lehigh University and Prof. Jack Lunsford at Texas A&M
University. Since 2000, he is full professor of inorganic chemistry and
catalysis at Utrecht University. He has authored over 340 publications
in peer-reviewed scientic journals. Weckhuysen has received several
research awards for his work on in situ catalyst characterization, the
most recent being the 2011 Paul H. Emmett Award in Fundamental
Catalysis from the North American Catalysis Society, the 2012
International Catalysis Award from the International Association of
Catalysis Societies, the 2013 Royal Society of Chemistry Bourke
Award, and the 2013 Spinoza Award from The Netherlands
Organization for Scientic Research.
Javier Ruiz-Martinez was born in Granada, Spain, in 1979. He

completed his masters degree in Chemical Engineering at the
University of Granada and the Norwegian University of Science and
Technology in 2003. After graduation, he started his Ph.D. under the
supervision of Prof. Antonio Sepulveda-Escribano and Prof. Francisco
Rodriguez-Reinoso at the University of Alicante, on the synthesis and
characterization of PtSn catalysts for selective hydrogenation and
steam reforming reactions. During his Ph.D., he worked as a visiting
researcher with Prof. James Anderson at the University of Aberdeen
and with Prof. Burtron Davis and Dr. Gary Jacobs at the University of
Kentucky Center for Applied Energy Research. After obtaining his
Ph.D. in 2009, he moved to Utrecht University as a postdoctoral
research fellow in the group of Prof. Bert Weckhuysen, where he
obtained a personal NWO Veni grant in 2012. Since October 2013, he
is an assistant professor within the same group.
ACKNOWLEDGMENTS
This work was funded by The Netherlands Research School
Combination-Catalysis (NRSC-C). J.R.-M. also acknowledges
NWO for his VENI grant. Jelle M. Boereboom is kindly
acknowledged for designing Figures 12 and 20 and the table of
content gure.
REFERENCES
(1) Budavari, S.; ONeil, M.; Smith, A.; Heckelman, P.; Obenchain, J.
In The Merck Index, 12th ed.; Budavari, S., Ed.; Merck & Co.: NJ,
1996; pp 13481349.
(2) McCoy, M.; Reisch, M.; Tullo, A. H.; Short, P. L.; Tremblay, J.-F.
Chem. Eng. News 2006, 84, 59.
(3) Ethylene Uses and Market Data, http://www.icis.com/Articles/
2007/11/05/9075777/ethylene-uses-and-market-data.html (accessed
Oct 25, 2013).
(4) Market Study: Propylene (UC-1705), http://www.ceresana.com/
en/market-studies/chemicals/propylene/ (accessed Oct 25, 2013).
(5) The Global Olens and Polyolens Markets in 2011 - Slow
Growth in Demand Amid Political and Economical Crisis, http://
www.researchandmarkets.com/research/bdd2c0/the_global_olens
(accessed Oct 25, 2013).
(6) Nexant. Industry Report - Independent Market Report on the
Global and Indonesian Petrochemicals Industry, http://www.chandraasri.com/UserFiles/201105151926340.NexantIndustryReport2011.pdf
(7) Radclie, C. The FCC unit as a propylene source, http://www.
digitalrening.com/article/1000312#.UmpROtK8BWQ (accessed Oct
25, 2013).
(8) Tullo, A. H. Chem. Eng. News 2012, 90, 10.
(9) McFarland, E. Science 2012, 338, 340.
(10) Torres Galvis, H. M.; de Jong, K. P. ACS Catal. 2013, 3, 2130.
(11) Stocker, M. Microporous Mesoporous Mater. 1999, 29, 3.
(12) Bullin, K.; Krouskop, P. Composition Variety Complicates
Processing Plans for US Shale Gas, http://www.bre.com/portals/0/
Eduardo Santillan-Jimenez was born in San Luis Potosi, Mexico, in

1980. While studying chemistry at the Autonomous University of San
Luis Potosi, he interned in the laboratory of Prof. David Atwood at the
University of Kentucky (Lexington, KY), working on the removal of
mercury from water. After graduation, he joined the University of
Kentucky (UK) chemistry graduate program, performing his doctoral
research on the selective catalytic reduction of nitrogen oxides at the
UK Center for Applied Energy Research (CAER) and at the
University of Alicante (Spain), under the guidance of Prof. Mark
Crocker and Prof. Agustin Bueno-Lopez. After obtaining his Ph.D. in
2008, he worked as a postdoctoral fellow in the group of Prof. Bert
Weckhuysen at Utrecht University (The Netherlands), where part of
his work centered on the study of propane dehydrogenation catalysts
via operando spectroscopy. In 2010, he returned to the UK-CAER,
where he now holds the position of Principal Research Scientist.
10648
Chemical Reviews
Review
technicalarticles/KeithBullin-CompositionVariety_USShaleGas.pdf
(13) Chemical Composition of Natural Gas, http://www.uniongas.
com/about-us/about-natural-gas/Chemical-Composition-of-NaturalGas#L (accessed Oct 23, 2013).
(14) Americas New Energy Future: The Unconventional Oil & Gas
Revolution and the US Economy, http://www.ihs.com/info/ecc/a/
americas-new-energy-future-report-vol-3.aspx (accessed Oct 23, 2013).
(15) Schut, J. H. How Shale Gas Is Changing Propylene, http://
plasticsengineeringblog.com/2013/02/20/how-shale-gas-is-changingpropylene/ (accessed Oct 23, 2013).
(16) Propylene from Propane via Dehydrogenation (similar to UOP
Oleex), http://base.intratec.us/home/chemical-processes/
propylene/propylene-from-propane-via-dehydrogenation (accessed
Oct 23, 2013).
(17) Propylene from Propane via Dehydrogenation (similar to
Lummus CATOFIN), http://base.intratec.us/home/chemicalprocesses/propylene/propylene-from-propane-via-dehydrogenation-2
(18) Buonomo, F.; Sanlippo, D.; Triro, F. In Handbook of
Heterogeneous Catalysis; Ertl, G., Knozinger, H., Weitkamp, J., Eds.;
Wiley-VHC Verlag GmbH: Weinheim, Germany, 1997; pp 2140
2151.
(19) Alperowicz, N. Engineering & construction: Shale boom
reinvigorates the industry, could lead to shortage of labor recources,
http://www.chemweek.com/lab/Engineering-and-construction-Shaleboom-reinvigorates-the-industry-could-lead-to-shortage-of-laborresources_52648.html (accessed Oct 23, 2013).
(20) Hua, L. China propane imports to surge on new PDH projects,
http://www.icis.com/Articles/2013/05/21/9670487/china-propaneimports-to-surge-on-new-pdh-projects.html (accessed Oct 23, 2013).
(21) Shale Gas, Competitiveness, and New US Chemical Industry
Investment: An Analysis Based on Announced Projects, http://
chemistrytoenergy.com/sites/chemistrytoenergy.com/les/shale-gasfull-study.pdf (accessed Oct 23, 2013).
(22) Richardson, J. Market intelligence: BD capacity on the way,
http://www.icis.com/Articles/2012/03/12/9539978/Marketintelligence-BD-capacity-on-the-way.html (accessed Jun 26, 2014).
(23) Richardson, J. Butadiene Oversupply Threat, http://www.icis.
com/blogs/asian-chemical-connections/2012/05/butadieneoversupply-threat/ (accessed Jun 26, 2014).
(24) Naylor, L. Lower Europe LLDPE imports support Europe
pricing picture, http://www.icis.com/Articles/2012/03/12/9539978/
Market-intelligence-BD-capacity-on-the-way.html (accessed Jun 26,
2014).
(25) Lopez, J. BASF, Linde plan to jointly develop butadiene
technology, http://www.icis.com/resources/news/2014/06/03/
9786384/basf-linde-plan-to-jointly-develop-butadiene-technology/
(accessed Jun 26, 2014).
(26) Bhasin, M. M.; McCain, J. H.; Vora, B. V.; Imai, T.; Pujado, P. R.
Appl. Catal., A 2001, 221, 397.
(27) Vora, B. V. Top. Catal. 2012, 55, 1297.
(28) Bricker, J. C. Top. Catal. 2012, 55, 1309.
(29) Weckhuysen, B. M.; Schoonheydt, R. A. Catal. Today 1999, 51,
223.
(30) Jackson, S. D.; Stair, P. C.; Gladden, L. F.; McGregor, J. In Metal
Oxide Catalysis; Jackson, S. D., Hargreaves, J. S. J., Eds.; Wiley-VHC
Verlag GmbH: Weinheim, Germany, 2009; pp 595612.
(31) Albonetti, S.; Cavani, F.; Trifiro, F. Catal. Rev.: Sci. Eng. 1996,
38, 413.
(32) Cavani, F.; Ballarini, N.; Cericola, A. Catal. Today 2007, 127,
113.
(33) Coperet, C. Chem. Rev. 2010, 110, 656.
(34) Krylov, O. V.; Mamedov, A. K.; Mirzabekova, S. R. Ind. Eng.
Chem. Prod. Res. Dev. 1995, 34, 474.
(35) Wang, S.; Zhu, Z. H. Energy Fuels 2004, 18, 1126.
(36) Ansari, M. B.; Park, S.-E. Energy Environ. Sci. 2012, 5, 9419.
(37) Caspary, K. J.; Gehrke, H.; Heinritz-Adrian, M.; Schwefer, M. In
Handbook of Heterogeneous Catalysis; Ertl, G., Knozinger, H.,
Weitkamp, J., Eds.; Wiley-VHC Verlag GmbH: Weinheim, Germany,

2008; pp 32063229.
(38) Pujado, P. R.; Vora, B. V. Hydrocarbon Process. 1990, 69, 65.
(39) Buyanov, R. A.; Pakhomov, N. A. Kinet. Catal. 2001, 42, 64.
(40) Altani, A. M. Oil Gas Eur. Mag. 2004, 30, 36.
(41) Tinnemans, S. J. Combined operando Raman/UV-Vis-NIR
spectroscopy as a tool to study supported metal oxide catalysts at
work. Ph.D. Thesis, Utrecht University, 2006.
(42) Research and Markets: China Propylene Market, 20132015,
http://www.businesswire.com/news/home/20130913005218/en/
Research-Markets-China-Propylene-Market-2013-2015 (accessed Oct
23, 2013).
(43) Ondrey, G. UOPs C3 Oleex process selected for propylene
plant in the U.S. Gulf Coast, http://www.che.com/only_on_che/
latest_news/UOPs-C3-Oleex-process-selected-for-propylene-plantin-the-U-S-Gulf-Coast_10539.html (accessed Oct 23, 2013).
(44) Boswell, C. On-purpose technologies ready to ll propylene gap,
http://www.icis.com/Articles/2012/04/16/9549968/on-purposetechnologies-ready-to-ll-propylene-gap.html (accessed Oct 23, 2013).
(45) Yu, L.; Chen, K. China Insight: Propane dehydrogenation
projects boom, but risks loom, http://www.icis.com/Articles/2012/
04/02/9546232/China-Insight-Propane-dehydrogenation-projectsboom-but-risks.html (accessed Oct 23, 2013).
(46) CATOFIN Dehydrogenation, http://www.cbi.com/images/
uploads/tech_sheets/CatonDehyrogenation-12.pdf (accessed Oct
23, 2013).
(47) Ercan, C.; Gartside, R. J. Can. J. Chem. Eng. 1996, 74, 626.
(48) Sanlippo, D.; Miracca, I. In Proceedings of the DGMK Conference
Oxidation and Functionalization: Classical and Alternative Routes and
Sources; Ernst, S., Gallei, E., Lercher, J. A., Rossini, S., Santacesaria, E.,
Eds.; German Society for Petroleum and Coal Science and Technology
(DGMK): Hamburg, 2005; p 773.
(49) Nijhuis, T. A.; Tinnemans, S. J.; Visser, T.; Weckhuysen, B. M.
Chem. Eng. Sci. 2004, 59, 5487.
(50) UOP Oleex process for light olen production, http://pet-oil.
blogspot.nl/2012/10/uop-oleex-process-for-light-olen.html (accessed Oct 23, 2013).
(51) Honeywells UOP Technology Selected For Petrochemical
Production In China, http://www.dailynance.com/rtn/pr/
honeywell-s-uop-technology-selected-for-petrochemical-production-inchina/rd637412763/ (accessed Oct 23, 2013).
(52) CCR Platforming & Oleex, http://www.uop.com/products/
equipment/ccr-regeneration/ (accessed Oct 23, 2013).
(53) New Clariant CATOFIN Propane Dehydrogenation Catalysts
Delivers Signicant Savings, http://newsroom.clariant.com/newclariant-caton-propane-dehydrogenation-catalyst-delivers-signicantsavings/ (accessed Oct 23, 2013).
(54) Sanlippo, D.; Miracca, I.; Triro, F. In Encyclopedia of
Hydrocarbons: Rening and Petrochemicals; Amadei, C., Ed.; ENI:
Rome, 2006; Vol. II, pp 687700.
(55) Jensen, S. F.; Roennekleiv, M.; Rytter, E.; Souraker, P.
International Patent WO 2000072967 A1, 2000.
(56) The Uhde STAR process: Oxydehydrogenation of light parans
to olens, http://www.thyssenkrupp-uhde.de/leadmin/documents/
brochures/uhde_brochures_pdf_en_12.pdf (accessed Oct 23, 2013).
(57) Sinfelt, J. H. J. Phys. Chem. 1964, 68, 344.
(58) Cortright, R. D.; Watwe, R. M.; Dumesic, J. A. J. Mol. Catal. A:
Chem. 2000, 163, 91.
(59) Willems, P. A.; Froment, G. F. Ind. Eng. Chem. Res. 1988, 27,
1966.
(60) Olah, G. A.; Molnar, A. Hydrocarbon Chemistry, 2nd ed.; Wiley
Inc.: Hoboken, NJ, 2003; pp 3037.
(61) Rahimi, N.; Karimzadeh, R. Appl. Catal., A 2011, 398, 1.
(62) Barron, Y.; Maire, G.; Muller, J. M.; Gault, F. G. J. Catal. 1966,
5, 428.
(63) Garin, F.; Aeiyach, S.; Legare, P.; Maire, G. J. Catal. 1982, 77,
323.
(64) Olah, G. A.; Molnar, A. Hydrocarbon Chemistry, 2nd ed.; Wiley
Inc.: Hoboken, NJ, 2003; pp 161185.
10649
Chemical Reviews
Review
(65) Grasselli, R. K. Catal. Today 1999, 24, 141.

(66) Dittmeyer, R.; Hollein, V.; Daub, K. J. Mol. Catal. A: Chem.
2001, 173, 135.
(67) Ravanchi, M. T.; Kaghazchi, T.; Kargari, A. Desalination 2009,
235, 199.
(68) Sheintuch, M.; Simakov, D. S. A. In Membrane Reactors for
Hydrogen Production Processes; De Falco, M., Marrelli, L., Iaquaniello,
G., Eds.; Springer-Verlag Ltd.: London, 2011; pp 183200.
(69) Rioux, R. M.; Song, H.; Hoefelmeyer, J. D.; Yang, P.; Somorjai,
G. A. J. Phys. Chem. B 2005, 109, 2192.
(70) Song, H.; Rioux, R. M.; Hoefelmeyer, J. D.; Komor, R.; Niesz,
K.; Grass, M.; Yang, P.; Somorjai, G. A. J. Am. Chem. Soc. 2006, 128,
3027.
(71) Santhosh Kumar, M.; Chen, D.; Walmsley, J. C.; Holmen, A.
Catal. Commun. 2008, 20, 747.
(72) Yang, M.-L.; Zhu, Y.-A.; Fan, C.; Sui, Z.-J.; Chen, D.; Zhou, X.G. Phys. Chem. Chem. Phys. 2011, 13, 3257.
(73) Cortright, R. D.; Dumesic, J. A. J. Catal. 1994, 148, 771.
(74) Spivey, J. J.; Roberts, G. W.; Goodwin, J. G.; Kim, S.; Rhodes,
W. D. In Catalysis; Spivey, J. J., Roberts, G. W., Eds.; The Royal
Society of Chemistry: London, 2004; pp 320348.
(75) Yang, M.-L.; Zhu, Y.-A.; Fan, C.; Sui, Z.-J.; Chen, D.; Zhou, X.G. J. Mol. Catal. A: Chem. 2010, 321, 42.
(76) Fiedorow, R. M. J.; Chahar, B. S.; Wanke, S. E. J. Catal. 1978,
51, 193.
(77) Moulijn, J. A.; van Diepen, A. E.; Kapteijn, F. Appl. Catal., A
2001, 212, 3.
(78) Nagai, Y.; Hirabayashi, T.; Dohmae, K.; Takagi, N.; Minami, T.;
Shinjoh, H.; Matsumoto, S. J. Catal. 2006, 242, 103.
(79) Adler, S. F.; Keavney, J. J. J. Phys. Chem. 1960, 64, 208.
(80) Lieske, H.; Lietz, G.; Spindler, H.; Volter, J. J. Catal. 1983, 81, 8.
(81) Lee, T. J.; Kim, Y. G. J. Catal. 1984, 90, 279.
(82) Le Normand, F.; Borgna, A.; Garetto, T. F.; Apesteguia, C. R.;
Moraweck, B. J. Phys. Chem. 1996, 100, 9068.
(83) Monzon, A.; Garetto, T. F.; Borgna, A. Appl. Catal., A 2003,
248, 279.
(84) Bocanegra, S. A.; Castro, A. A.; Guerrero-Ruz, A.; Scelza, O. A.;
de Miguel, S. R. Chem. Eng. J. 2006, 118, 161.
(85) Zhang, Y.; Zhou, Y.; Shi, J.; Zhou, S.; Sheng, X.; Zhang, Z.;
Xiang, S. J. Mol. Catal. A: Chem. 2014, 381, 138.
(86) Casella, M. L.; Siri, G. J.; Santori, G. F.; Ferretti, O. A.; RamirezCorredores, M. M. Langmuir 2000, 16, 5639.
(87) Siri, G. J.; Bertolini, G. R.; Casella, M. L.; Ferretti, O. A. Mater.
Lett. 2005, 59, 2319.
(88) De Miguel, S. R.; Bocanegra, S. A.; Vilella, I. M. J.; GuerreroRuz, A.; Scelza, O. A. Catal. Lett. 2007, 119, 5.
(89) Tasbihi, M.; Feyzi, F.; Amlashi, M. A.; Abdullah, A. Z.;
Mohamed, A. R. Fuel Process. Technol. 2007, 88, 883.
(90) Rennard, R. J.; Freel, J. J. Catal. 1986, 98, 235.
(91) Aguilar-Rios, G.; Valenzuela, M. A.; Armendariz, H.; Salas, P.;
Domnguez, J. M.; Acosta, D. R.; Schifter, I. Appl. Catal., A 1992, 90,
25.
(92) Bocanegra, S. A.; Guerrero-Ruz, A.; de Miguel, S. R.; Scelza, O.
A. Appl. Catal., A 2004, 277, 11.
(93) Lai, Y.; He, S.; Li, X.; Sun, C.; Seshan, K. Appl. Catal., A 2014,
469, 74.
(94) Mironenko, R. M.; Belskaya, O. B.; Talsi, V. P.; Gulyaeva, T. I.;
Kazakov, M. O.; Nizovskii, A. I.; Kalinkin, A. V.; Bukhtiyarov, V. I.;
Lavrenov, A. V.; Likholobov, V. A. Appl. Catal., A 2014, 469, 472.
(95) Cortright, R. D.; Dumesic, J. A. Appl. Catal., A 1995, 129, 101.
(96) Cortright, R. D.; Levin, P. E.; Dumesic, J. A. Ind. Eng. Chem. Res.
1998, 37, 1717.
(97) Hill, J. M.; Cortright, R. D.; Dumesic, J. A. Appl. Catal., A 1998,
168, 9.
(98) Cortright, R. D.; Hill, J. M.; Dumesic, J. A. Catal. Today 2000,
55, 213.
(99) Azzam, K. G.; Jacobs, G.; Shafer, W. D.; Davis, B. H. Appl.
Catal., A 2010, 390, 264.
(100) Biscardi, J. A.; Iglesia, E. J. Catal. 1999, 182, 117.
(101) Azzam, K. G.; Jacobs, G.; Shafer, W. D.; Davis, B. H. J. Catal.

2010, 270, 242.
(102) Waku, T.; Biscardi, J. A.; Iglesia, E. Chem. Commun. 2003, 9,
1764.
(103) Li, X.; Iglesia, E. Chem. Commun. 2008, 8, 594.
(104) Lieske, H.; Sarkany, A.; Volter, J. Appl. Catal. 1987, 30, 69.
(105) Wu, J.; Peng, Z.; Bell, A. T. J. Catal. 2014, 311, 161.
(106) Virnovskaia, A.; Morandi, S.; Rytter, E.; Ghiotti, G.; Olsbye, U.
J. Phys. Chem. C 2007, 111, 14732.
(107) Nagaraja, B. M.; Shin, C.-H.; Jung, K.-D. Appl. Catal., A 2013,
467, 211.
(108) Boudart, M.; Aldag, A.; Benson, J. E.; Dougharty, N. A.; Girvin
Harkins, C. J. Catal. 1966, 6, 92.
(109) Cardona-Martinez, N.; Dumesic, J. A. Adv. Catal. 1992, 38,
149.
(110) Ruiz-Martnez, J.; Sepulveda-Escribano, A.; Anderson, J. A.;
Rodrguez-Reinoso, F. Catal. Today 2007, 123, 235.
(111) Ruiz-Martnez, J.; Coloma, F.; Sepulveda-Escribano, A.;
Anderson, J. A.; Rodrguez-Reinoso, F. Catal. Today 2008, 133135,
35.
(112) Ruiz-Martnez, J.; Sepulveda-Escribano, A.; Anderson, J. A.;
Rodrguez-Reinoso, F. Phys. Chem. Chem. Phys. 2009, 11, 917.
(113) Soares, O. S. G. P.; O rfao, J. J. M.; Ruiz-Martnez, J.; SilvestreAlbero, J.; Sepulveda-Escribano, A.; Pereira, M. F. R. Chem. Eng. J.
2010, 165, 78.
(114) Yang, M.; Zhu, Y.; Zhou, X.; Sui, Z.; Chen, D. ACS Catal.
2012, 2, 1247.
(115) Cortright, R. D.; Dumesic, J. A. J. Catal. 1995, 157, 576.
(116) Natal-Santiago, M. A.; Podkolzin, S. G.; Cortright, R. D.;
Dumesic, J. A. Catal. Lett. 1997, 45, 155.
(117) Shen, J.; Hill, J. M.; Watwe, R. M.; Spiewak, B. E.; Dumesic, J.
A. J. Phys. Chem. B 1999, 103, 3923.
(118) Nykanen, L.; Honkala, K. J. Phys. Chem. C 2011, 115, 9578.
(119) Gao, J.; Zhao, H.; Yang, X.; Koel, B. E.; Podkolzin, S. G. Angew.
Chem., Int. Ed. 2014, 53, 3641.
(120) Siri, G. J.; Ramallo-Lopez, J. M.; Casella, M. L.; Fierro, J. L. G.;
Requejo, F. G.; Ferretti, O. A. Appl. Catal., A 2005, 278, 239.
(121) Fearon, J.; Watson, G. W. J. Mater. Chem. 2006, 16, 1989.
(122) Iglesias-Juez, A.; Beale, A. M.; Maaijen, K.; Weng, T. C.;
Glatzel, P.; Weckhuysen, B. M. J. Catal. 2010, 276, 268.
(123) Deng, L.; Shishido, T.; Teramura, K.; Tanaka, T. Catal. Today
2013, 232, 33.
(124) Silvestre-Albero, J.; Sanchez-Castillo, M. A.; He, R.; SepulvedaEscribano, A.; Rodriguez-Reinoso, F.; Dumesic, J. A. Catal. Lett. 2001,
74, 17.
(125) Silvestre-Albero, J.; Serrano-Ruiz, J. C.; Sepulveda-Escribano,
A.; Rodrguez-Reinoso, F. Appl. Catal., A 2005, 292, 244.
(126) De Cola, P. L.; Glaser, R.; Weitkamp, J. Appl. Catal., A 2006,
306, 85.
(127) Yu, C.; Xu, H.; Ge, Q.; Li, W. J. Mol. Catal. A: Chem. 2007,
266, 80.
(128) Silvestre-Albero, J.; Serrano-Ruiz, J. C.; Sepulveda-Escribano,
A.; Rodrguez-Reinoso, F. Appl. Catal., A 2008, 351, 16.
(129) Vu, B. K.; Song, M. B.; Ahn, I. Y.; Suh, Y.-W.; Suh, D. J.; Kim,
W.-I.; Koh, H.-L.; Choi, Y. G.; Shin, E. W. Appl. Catal., A 2011, 400,
25.
(130) Jablonski, E. L.; Castro, A. A.; Scelza, O. A.; de Miguel, S. R.
Appl. Catal., A 1999, 183, 189.
(131) Sun, P.; Siddiqi, G.; Chi, M.; Bell, A. T. J. Catal. 2010, 274,
192.
(132) Siddiqi, G.; Sun, P.; Galvita, V.; Bell, A. T. J. Catal. 2010, 274,
200.
(133) Sun, P.; Siddiqi, G.; Vining, W. C.; Chi, M.; Bell, A. T. J. Catal.
2011, 282, 165.
(134) Ballarini, A. D.; de Miguel, S. R.; Castro, A. A.; Scelza, O. A.
Appl. Catal., A 2013, 467, 235.
(135) Serrano-Ruiz, J. C.; Sepulveda-Escribano, A.; RodrguezReinoso, F. J. Catal. 2007, 246, 158.
(136) Barias, O. A.; Holmen, A.; Blekkan, E. A. J. Catal. 1996, 158, 1.
10650
Chemical Reviews
Review
(137) Salmones, J.; Wang, J.-A.; Galicia, J. A.; Aguilar-Rios, G. J. Mol.

Catal. A: Chem. 2002, 184, 203.
(138) Zhou, Y.; Davis, M. D. U.S. Patent 5214227, 1993.
(139) Barri, S. A. I.; Tahir, R. U.S. Patent 5126502, 1992.
(140) Stagg, S. M.; Querini, C. A.; Alvarez, W. E.; Resasco, D. E. J.
Catal. 1997, 168, 75.
(141) Zhang, Y.; Zhou, Y.; Huang, L.; Xue, M.; Zhang, S. Ind. Eng.
Chem. Res. 2011, 50, 7896.
(142) Duan, Y.; Zhou, Y.; Zhang, Y.; Sheng, X.; Xue, M. Catal. Lett.
2011, 141, 120.
(143) Frey, F. E.; Huppke, W. F. Ind. Eng. Chem. Res. 1933, 25, 54.
(144) Olsbye, U.; Virnovskaia, A.; Prytz, .; Tinnemans, S. J.;
Weckhuysen, B. M. Catal. Lett. 2005, 103, 143.
(145) Weckhuysen, B. M.; de Ridder, L. M.; Schoonheydt, R. A. J.
Phys. Chem. 1993, 97, 4756.
(146) Weckhuysen, B. M.; Schoonheydt, R. A.; Mabbs, F. E.;
Collison, D. J. Chem. Soc., Faraday Trans. 1996, 92, 2431.
(147) Cavani, F.; Koutyrev, M.; Trifiro, F.; Bartolini, A.; Ghisletti, D.;
Iezzi, R.; Santucci, A.; Del Piero, G. J. Catal. 1996, 158, 236.
(148) Weckhuysen, B. M.; Wachs, I. E. J. Chem. Soc., Faraday Trans.
1996, 92, 1969.
(149) Santhosh Kumar, M.; Hammer, N.; Ronning, M.; Holmen, A.;
Chen, D.; Walmsley, J. C.; Oye, G. J. Catal. 2009, 261, 116.
(150) Michorczyk, P.; Ogonowski, J.; Zenczak, K. J. Mol. Catal. A:
Chem. 2011, 349, 1.
(151) Weckhuysen, B. M.; Wachs, I. E. J. Phys. Chem. 1996, 100,
14437.
(152) Weckhuysen, B. M.; Bensalem, A.; Schoonheydt, R. A. J. Chem.
Soc., Faraday Trans. 1998, 94, 2011.
(153) Weckhuysen, B. M.; Verberckmoes, A. A.; Debaere, J.; Ooms,
K.; Langhans, I.; Schoonheydt, R. A. J. Mol. Catal. A: Chem. 2000, 151,
115.
(154) Marcilly, C.; Delmon, B. J. Catal. 1972, 24, 336.
(155) Ashmawy, F. M. J. Chem. Soc., Faraday Trans. 1980, 76, 2096.
(156) Lugo, H. J.; Lunsford, J. H. J. Catal. 1985, 91, 155.
(157) Grunert, W.; Saffert, W.; Feldhaus, R.; Anders, K. J. Catal.
1986, 99, 149.
(158) Selwood, P. W. J. Am. Chem. Soc. 1970, 92, 39.
(159) Cimino, A.; Cordischi, D.; De Rossi, S.; Ferraris, G.; Gazzoli,
D.; Indovina, V.; Minelli, G.; Occhiuzzi, M.; Valigi, M. J. Catal. 1991,
127, 744.
(160) Bruckner, A.; Radnik, J.; Hoang, D. L.; Lieske, H. Catal. Lett.
1999, 60, 183.
(161) Mimura, N.; Okamoto, M.; Yamashita, H.; Oyama, S. T.;
Murata, K. J. Phys. Chem. B 2006, 110, 21764.
(162) Perez-Reina, F. J.; Rodrguez-Castellon, E.; Jimenez-Lopez, A.
Langmuir 1999, 15, 8421.
(163) Hakuli, A.; Kytokivi, A.; Krause, A. O. I. Appl. Catal., A 2000,
190, 219.
(164) Kyto kivi, A.; Jacobs, J. P.; Hakuli, A.; Merilainen, J.;
Brongersma, H. H. J. Catal. 1996, 162, 190.
(165) Michorczyk, P.; Pietrzyk, P.; Ogonowski, J. Microporous
Mesoporous Mater. 2012, 161, 56.
(166) Hakuli, A.; Kytokivi, A.; Krause, A. O. I.; Suntola, T. J. Catal.
1996, 161, 393.
(167) De Rossi, S.; Ferraris, G.; Fremoiotti, S.; Garrone, E.; Ghiotti,
G.; Campa, M. C.; Indovina, V. J. Catal. 1994, 148, 36.
(168) De Rossi, S.; Casaletto, M. P.; Ferraris, G.; Cimino, A.; Minelli,
G. Appl. Catal., A 1998, 167, 257.
(169) Gaspar, A. B.; Brito, J. L. F.; Dieguez, L. C. J. Mol. Catal. A:
Chem. 2003, 203, 251.
(170) Zhang, X.; Yue, Y.; Gao, Z. Catal. Lett. 2002, 83, 19.
(171) Airaksinen, S. M. K.; Kanervo, J. M.; Krause, A. O. I. Stud. Surf.
Sci. Catal. 2001, 136, 153.
(172) Puurunen, R. L.; Weckhuysen, B. M. J. Catal. 2002, 210, 418.
(173) Nijhuis, T. A.; Tinnemans, S. J.; Visser, T.; Weckhuysen, B. M.
Phys. Chem. Chem. Phys. 2003, 5, 4361.
(174) Takehira, K.; Ohishi, Y.; Shishido, T.; Kawabata, T.; Takaki, K.;
Zhang, Q.; Wang, Y. J. Catal. 2004, 224, 404.
(175) Shi, X.; Ji, S.; Wang, K. Catal. Lett. 2008, 125, 331.
(176) Michorczyk, P.; Ogonowski, J.; Niemczyk, M. Appl. Catal., A
2010, 374, 142.
(177) Ma, F.; Chen, S.; Wang, Y.; Chen, F.; Lu, W. Appl. Catal., A
2012, 427428, 145.
(178) Shishido, T.; Shimamura, K.; Teramura, K.; Tanaka, T. Catal.
Today 2012, 185, 151.
(179) Yang, H.; Xu, L.; Ji, D.; Wang, Q.; Lin, L. React. Kinet. Catal.
Lett. 2002, 76, 151.
(180) Tinnemans, S. J.; Kox, M. H. F.; Nijhuis, T. A.; Visser, T.;
Weckhuysen, B. M. Phys. Chem. Chem. Phys. 2005, 7, 211.
(181) De Rossi, S.; Ferraris, G.; Fremiotti, S.; Cimino, A.; Indovina,
V. Appl. Catal., A 1992, 81, 113.
(182) Puurunen, R. L.; Beheydt, B. G.; Weckhuysen, B. M. J. Catal.
2001, 204, 253.
(183) Weckhuysen, B. M.; De Ridder, L. M.; Grobet, P. J.;
Schoonheydt, R. A. J. Phys. Chem. 1995, 99, 320.
(184) Vuurman, M. A.; Hardcastle, F. D.; Wachs, I. E. J. Mol. Catal.
1993, 84, 193.
(185) Simon, S.; van der Pol, A.; Reijerse, E. J.; Kentgens, A. P. M. J.
Chem. Soc., Faraday Trans. 1995, 91, 1519.
(186) El-Shobaky, H. G.; Ghozza, A. M.; El-Shobaky, G. A.;
Mohamed, G. M. Colloids Surf., A 1999, 152, 315.
(187) Wachowski, L.; Kirszensztejn, P.; Lopatka, R.; Czajka, B. Mater.
Chem. Phys. 1994, 37, 29.
(188) Rossignol, S.; Kappenstein, C. J. Inorg. Mater. 2001, 3, 52.
(189) Zhao, H.; Song, H.; Xu, L.; Chou, L. Appl. Catal., A 2013, 456,
188.
(190) De Rossi, S.; Ferraris, G.; Fremiotti, S.; Indovina, V.; Cimino,
A. Appl. Catal., A 1993, 106, 125.
(191) Fujdala, K.; Tilly, T. D. J. Catal. 2003, 218, 123.
(192) Korhonen, S. T.; Airaksinen, S. M. K.; Banares, M. A.; Krause,
A. O. I. Appl. Catal., A 2007, 333, 30.
(193) Martyanov, I.; Sayari, A. Catal. Lett. 2008, 126, 164.
(194) Bibby, D. M.; Milestone, N. B.; Aldridge, L. Nature 1979, 280,
664.
(195) Tsyganok, A.; Green, R. G.; Giorgi, J. B.; Sayari, A. Catal.
Commun. 2007, 8, 2186.
(196) Alcantara-Rodrguez, M.; Rodrguez-Castellon, E.; JimenezLopez, A. Langmuir 1999, 15, 1115.
(197) Shee, D.; Sayari, A. Appl. Catal., A 2010, 389, 155.
(198) Sattler, J. J. H. B.; Gonzalez-Jimenez, I. D.; Mens, A. M.; Arias,
M.; Visser, T.; Weckhuysen, B. M. Chem. Commun. 2013, 49, 1518.
(199) Chaar, M. A.; Patel, D.; Kung, M. C.; Kung, H. H. J. Catal.
1987, 105, 483.
(200) Wachs, I. E.; Weckhuysen, B. M. Appl. Catal., A 1997, 157, 67.
(201) Weckhuysen, B. M.; Keller, D. E. Catal. Today 2003, 78, 25.
(202) Volpe, M.; Tonetto, G.; de Lasa, H. Appl. Catal., A 2004, 272,
69.
(203) Wu, Z.; Kim, H.-S.; Stair, P. C.; Rugmini, S.; Jackson, S. D. J.
Phys. Chem. B 2005, 109, 2793.
(204) Wu, Z.; Stair, P. J. Catal. 2006, 237, 220.
(205) Jackson, S. D.; Rugmini, S. J. Catal. 2007, 251, 59.
(206) McGregor, J.; Huang, Z.; Shiko, G.; Gladden, L. F.; Stein, R. S.;
Duer, M. J.; Wu, Z.; Stair, P. C.; Rugmini, S.; Jackson, S. D. Catal.
Today 2009, 142, 143.
(207) Harlin, M. E.; Niemi, V. M.; Krause, A. O. I. J. Catal. 2000,
195, 67.
(208) Harlin, M. E.; Niemi, V. M.; Krause, A. O. I.; Weckhuysen, B.
M. J. Catal. 2001, 203, 242.
(209) Takahara, I.; Saito, M.; Inaba, M.; Murata, K. Catal. Lett. 2005,
102, 201.
(210) Steinfeldt, N.; Muller, D.; Berndt, H. Appl. Catal., A 2004, 272,
201.
(211) Ogonowski, J.; Skrzynska, E. React. Kinet. Catal. Lett. 2006, 88,
293.
(212) Ogonowski, J.; Skrzynska, E. Catal. Lett. 2006, 111, 79.
(213) Ogonowski, J.; Skrzynska, E. Catal. Lett. 2008, 121, 234.
10651
Chemical Reviews
Review
(214) Sokolov, S.; Stoyanova, M.; Rodemerck, U.; Linke, D.;

Kondratenko, E. V. Catal. Sci. Technol. 2014, 4, 1323.
(215) McGregor, J.; Huang, Z.; Parrott, E. P. J.; Zeitler, J. A.;
Nguyen, K. L.; Rawson, J. M.; Carley, A.; Hansen, T. W.; Tessonnier,
J.-P.; Su, D. S.; Teschner, D.; Vass, E. M.; Knop-Gericke, A.; Schlogl,
R.; Gladden, L. F. J. Catal. 2010, 269, 329.
(216) Russell, A. S.; Stokes, J. J. Ind. Eng. Chem. 1946, 38, 1071.
(217) Xie, S.; Chen, K.; Bell, A. T.; Iglesia, E. J. Phys. Chem. B 2000,
104, 10059.
(218) Harlin, M. E.; Backman, L. B.; Krause, A. O. I.; Jylha, O. J. T. J.
Catal. 1999, 183, 300.
(219) Lopez Cordero, R.; Gil Llambias, F. J.; Lopez Agudo, A. Appl.
Catal. 1991, 74, 125.
(220) Mitchell, P. C. H.; Wass, S. A. Appl. Catal., A 2002, 225, 153.
(221) Ledoux, M. J.; Meunier, F.; Heinrich, B.; Pham-huu, C.; Harlin,
M. E.; Krause, A. O. I. Appl. Catal., A 1999, 181, 157.
(222) Frank, B.; Cotter, T. P.; Schuster, M. E.; Schlogl, R.;
Trunschke, A. Chem.Eur. J. 2013, 19, 16938.
(223) Lee, J. S.; Oyama, S. T.; Boudart, M. J. Catal. 1987, 106, 125.
(224) Harlin, M. E.; Krause, A. O. I.; Heinrich, B.; Pham-huu, C.;
Ledoux, M. J. Appl. Catal., A 1999, 185, 311.
(225) Gnep, N. S.; Doyemet, J. Y.; Seco, A. M.; Ribeiro, F. R.;
Guisnet, M. Appl. Catal. 1988, 43, 155.
(226) Meriaudeau, P.; Naccache, C. J. Mol. Catal. 1989, 50, L7.
(227) Price, G. L.; Kanazirev, V. J. Catal. 1990, 126, 267.
(228) Meitzner, G. D.; Iglesia, E.; Baumgartner, J. E.; Huang, E. S. J.
Catal. 1993, 140, 209.
(229) Zheng, B.; Hua, W.; Yue, Y.; Gao, Z. J. Catal. 2005, 232, 143.
(230) Nesterenko, N. S.; Ponomoreva, O. A.; Yuschenko, V. V.;
Ivanova, I. I.; Testa, F.; Di Renzo, F.; Fajula, F. Appl. Catal., A 2003,
254, 261.
(231) Liu, Y.; Li, Z. H.; Lu, J.; Fan, K. J. Phys. Chem. C 2008, 112,
20382.
(232) Chen, M.; Xu, J.; Su, F.; Liu, Y.; Cao, Y.; He, H.; Fan, K. J.
Catal. 2008, 256, 293.
(233) Xu, B.; Zheng, B.; Hua, W.; Yue, Y.; Gao, Z. J. Catal. 2006,
239, 470.
(234) Xu, B.; Li, T.; Zheng, B.; Hua, W.; Yue, Y.; Gao, Z. Catal. Lett.
2007, 119, 283.
(235) Michorczyk, P.; Ogonowski, J. Appl. Catal., A 2003, 251, 425.
(236) Pidko, E. A.; Hensen, E. J. M.; van Santen, R. A. J. Phys. Chem.
C 2007, 111, 13068.
(237) Takahara, I.; Saito, M.; Inaba, M.; Murata, K. Catal. Lett. 2004,
96, 29.
(238) Rodrguez, L.; Romero, D.; Rodrguez, D.; Sanchez, J.;
Domnguez, F.; Arteaga, G. Appl. Catal., A 2010, 373, 66.
(239) Chao, K. J.; Wei, A. C.; Wu, H. C.; Lee, J. F. Microporous
Mesoporous Mater. 2000, 3536, 413.
(240) Butt, D. P.; Park, Y.; Taylor, T. N. J. Nucl. Mater. 1999, 264,
71.
(241) Nakagawa, K.; Okamura, M.; Ikenaga, N.; Kobayashi, T. Chem.
Commun. 1998, 3, 1025.
(242) Nakagawa, K.; Kajita, C.; Okumura, K.; Ikenaga, N.; NishitaniGamo, M.; Ando, T.; Kobayashi, T.; Suzuki, T. J. Catal. 2001, 203, 87.
(243) Li, H.; Yue, Y.; Miao, C.; Xie, Z.; Hua, W.; Gao, Z. Catal.
Commun. 2007, 8, 1317.
(244) Shen, Z.; Liu, J.; Xu, H.; Yue, Y.; Hua, W.; Shen, W. Appl.
Catal., A 2009, 356, 148.
(245) Michorczyk, P.; Kustrowski, P.; Kolak, A.; Zimowska, M. Catal.
Commun. 2013, 35, 95.
(246) Hensen, E. J. M.; Pidko, E. A.; Rane, N.; van Santen, R. A.
Angew. Chem., Int. Ed. 2007, 38, 7273.
(247) Iezzi, R.; Bartolini, A.; Buonomo, F. U.S. Patent 0198428,
2002.
(248) Sattler, J. J. H. B.; Gonzalez-Jimenez, I. D.; Luo, L.; Stears, B.
A.; Malek, A.; Barton, D. G.; Kilos, B. A.; Kaminsky, M. P.; Verhoeven,
M. W. G. M.; Koers, E. J.; Baldus, M.; Weckhuysen, B. M. Angew.
Chem., Int. Ed. 2014, 53, 9251.
(249) Saito, M.; Watanabe, S.; Takahara, I.; Inaba, M.; Murata, K.
Catal. Lett. 2003, 89, 213.
(250) Shimada, H.; Akazawa, T.; Ikenaga, N.; Suzuki, T. Appl. Catal.,
A 1998, 168, 243.
(251) Wang, R.; Sun, X.; Zhang, B.; Sun, X.; Su, D. Chem.Eur. J.
2014, 20, 1.
(252) Zhang, J.; Liu, X.; Blume, R.; Zhang, A.; Schlogl, R.; Su, D. S.
Science 2008, 322, 73.
(253) Chen, C.; Zhang, J.; Zhang, B.; Yu, C.; Peng, F.; Su, D. Chem.
Commun. 2013, 49, 8151.
(254) Liu, L.; Deng, Q.-F.; Agula, B.; Zhao, X.; Ren, T.-Z.; Yuan, Z.Y. Chem. Commun. 2011, 47, 8334.
(255) Levi, R. B.; Boudart, M. Science 1973, 181, 547.
(256) Neylon, M. K.; Choi, S.; Kwon, H.; Curry, K. E.; Thompson, L.
T. Appl. Catal., A 1999, 183, 253.
(257) Wang, G.; Li, C.; Shan, H. ACS Catal. 2014, 4, 1139.
(258) Perdigon-Melon, J.; Gervasini, A.; Auroux, A. J. Catal. 2005,
234, 421.
(259) Halasz, J.; Konya, Z.; Fudala, A .; Beres, A.; Kiricsi, I. Catal.
Today 1996, 5861, 293.
(260) Chen, M.; Xu, J.; Liu, Y.-M.; Cao, Y.; He, H.-Y.; Zhuang, J.-H.
Appl. Catal., A 2010, 377, 35.
(261) Chen, M.; Wu, J.-L.; Liu, Y.-M.; Cao, Y.; Guo, L.; He, H.-Y.;
Fan, K.-N. Appl. Catal., A 2011, 407, 20.
(262) Chen, M.; Xu, J.; Cao, Y.; He, H.-Y.; Fan, K.-N.; Zhuang, J.-H.
J. Catal. 2010, 272, 101.
(263) Li, Q.; Sui, Z.; Zhou, X.; Chen, D. Appl. Catal., A 2011, 398,
18.
(264) Wenner, R. R.; Dybal, E. C. Catal. Eng. Prog. 1948, 45, 275.
(265) Stobbe, D. E.; van Buren, F. R.; Hoogenraad, M. S.; van Dillen,
A. J.; Geus, J. W. J. Chem. Soc., Faraday Trans. 1991, 87, 1639.
(266) Stobbe, D. E.; van Buren, F. R.; van Dillen, A. J.; Geus, J. W. J.
Catal. 1992, 562, 548.
(267) Boot, L. A.; van Dillen, A. J.; Geus, J. W.; van Buren, F. R. J.
Catal. 1996, 163, 186.
(268) Boot, L. A.; van Dillen, A. J.; Geus, J. W.; van Buren, F. R. J.
Catal. 1996, 163, 195.
(269) Sun, Y.; Tao, L.; You, T.; Li, C.; Shan, H. Chem. Eng. J. 2014,
244, 145.
(270) Yun, J. H.; Lobo, R. F. J. Catal. 2014, 312, 263.
(271) Guisnet, M.; Gnep, N. S.; Alario, F. Appl. Catal., A 1992, 89, 1.
(272) Almutairi, S. M. T.; Mezari, B.; Magusin, P. C. M. M.; Pidko, E.
A.; Hensen, E. J. M. ACS Catal. 2012, 2, 71.
(273) Biscardi, A.; Iglesia, E. Catal. Today. 1996, 31, 207.
(274) Mole, T.; Anderson, J. R.; Creer, G. Appl. Catal. 1985, 17, 141.
(275) Wang, L.; Tao, L.; Xie, M.; Xu, G.; Huang, J.; Xu, Y. Catal. Lett.
1993, 21, 35.
(276) Biscardi, J. A.; Meitzner, G. D.; Iglesia, E. J. Catal. 1998, 202,
192.
(277) Aleksandrov, H. A.; Vayssilov, G. N. Catal. Today 2010, 152,
78.
(278) Kazansky, V. B.; Subbotina, I. R.; Rane, N.; van Santen, R. A.;
Hensen, E. J. M. Phys. Chem. Chem. Phys. 2005, 7, 3088.
(279) Kazansky, V. B.; Pidko, E. A. J. Phys. Chem. B 2005, 109, 2103.
(280) Pidko, E. A.; Van Santen, R. A. J. Phys. Chem. C 2007, 111,
2643.
(281) Eastman, A. D. U.S. Patent 4327238, 1980.
(282) Eastman, A. D. U.S. Patent 4371730, 1983.
(283) Chen, Z. X.; Derking, A.; Koot, W.; van Dijk, M. P. J. Catal.
1996, 161, 730.
(284) Schweitzer, N. M.; Hu, B.; Das, U.; Kim, H.; Curtiss, L. A.;
Stair, P. C.; Miller, T.; Hock, A. S. ACS Catal. 2014, 4, 1091.
(285) Horiuti, I.; Polanyi, M. Trans. Faraday Soc. 1934, 30, 1164.
(286) Biloen, P.; Dautzenberg, F. M.; Sachtler, W. M. H. J. Catal.
1977, 50, 77.
(287) Chen, K.; Iglesia, E.; Bell, A. T. J. Catal. 2000, 192, 197.
(288) Sokolov, S.; Stoyanova, M.; Rodemerck, U.; Linke, D.;
Kondratenko, E. V. J. Catal. 2012, 293, 67.
10652
Chemical Reviews
Review
(289) Hauser, A. W.; Gomes, J.; Bajdich, M.; Head-Gordon, M.; Bell,
A. T. Phys. Chem. Chem. Phys. 2013, 15, 20727.
(290) Galvita, V.; Siddiqi, G.; Sun, P.; Bell, A. T. J. Catal. 2010, 271,
209.
(291) Virnovskaia, A.; Rytter, E.; Olsbye, U. Ind. Eng. Chem. Res.
2008, 47, 7167.
(292) Zaera, F.; Chrysostomou, D. Surf. Sci. 2000, 457, 71.
(293) Zaera, F.; Chrysostomou, D. Surf. Sci. 2000, 457, 89.
(294) Chrysostomou, D.; Zaera, F. J. Phys. Chem. B 2001, 105, 1003.
(295) Valcarcel, A.; Ricart, J.; Clotet, A.; Illas, F.; Markovits, A.;
Minot, C. J. Catal. 2006, 241, 115.
(296) Watwe, R. M.; Spiewak, B. E.; Cortright, R. D.; Dumesic, J. A.
J. Catal. 1998, 180, 184.
(297) Chrysostomou, D.; Chou, A.; Zaera, F. J. Phys. Chem. B 2001,
105, 5968.
(298) Ohtani, T.; Kubota, J.; Kondo, J. N.; Hirose, C.; Domen, K. J.
Phys. Chem. B 1999, 103, 4562.
(299) De La Cruz, C.; Sheppard, N. J. Chem. Soc., Faraday Trans.
1997, 93, 3569.
(300) Essen, J. M.; Haubrich, J.; Becker, C.; Wandelt, K. Surf. Sci.
2007, 601, 3472.
(301) Cremer, P. S.; Su, X.; Shen, Y. R.; Somorjai, G. A. J. Am. Chem.
Soc. 1996, 118, 2942.
(302) Cremer, P. S.; Stanners, C.; Niemantsverdriet, J. W.; Shen, Y.
R.; Somorjai, G. A. Surf. Sci. 1995, 328, 111.
(303) Zaera, F.; Janssens, T. V. W.; Ofner, H. Surf. Sci. 1996, 368,
371.
(304) Nykanen, L.; Honkala, K. ACS Catal. 2013, 3, 3026.
(305) Airaksinen, S. M. K.; Harlin, M. E.; Krause, A. O. I. Ind. Eng.
Chem. Res. 2002, 41, 5619.
(306) Lillehaug, S.; Brve, K. J.; Sierka, M.; Sauer, J. J. Phys. Org.
Chem. 2004, 17, 990.
(307) Lillehaug, S.; Jensen, V. R.; Brve, K. J. J. Phys. Org. Chem.
2006, 19, 25.
(308) Carra, S.; Forni, L.; Vintani, C. J. Catal. 1967, 9, 154.
(309) Kao, J. Y.; Piet-Lahanier, H.; Walter, E.; Happel, J. J. Catal.
1992, 133, 383.
(310) Suzuki, I.; Kaneko, Y. J. Catal. 1977, 47, 239.
(311) Airaksinen, S. M. K.; Banares, M. A.; Krause, A. O. I. J. Catal.
2005, 230, 507.
(312) Conley, M. P.; Delley, M. F.; Siddiqi, G.; Lapadula, G.; Norsic,
S.; Monteil, V.; Safonova, O. V.; Coperet, C. Angew. Chem., Int. Ed.
2014, 53, 1872.
(313) Delley, M. F.; Conley, M. P.; Coperet, C. Catal. Lett. 2014,
144, 805.
(314) Lillehaug, S.; Brve, K. J.; Sierka, M.; Sauer, J. J. Phys. Org.
Chem. 2004, 17, 990.
(315) Lillehaug, S.; Jensen, V. R.; Brve, K. J. J. Phys. Org. Chem.
2006, 19, 25.
(316) Kazansky, V. B.; Subbotina, I. R.; Pronin, A. A.; Schlogl, R.;
Jentoft, F. C. J. Phys. Chem. B 2006, 110, 7975.
(317) Noda, H.; Tone, S.; Otake, T. J. Chem. Eng. Jpn. 1973, 7, 1974.
(318) Toei, R.; Nakanishi, K.; Yamada, K.; Okazaki, M. J. Chem. Eng.
Jpn. 1975, 8, 131.
(319) Swift, H. E.; Beuther, H.; Rennard, R. J. Ind. Eng. Chem. Prod.
Res. Dev. 1976, 15, 131.
(320) Menon, P. G. J. Mol. Catal. 1990, 59, 207.
(321) Chen, Y.; Vlachos, D. G. J. Phys. Chem. C 2010, 114, 4973.
(322) Wang, B.; Ma, X.; Caffio, M.; Schaub, R.; Li, W.-X. Nano Lett.
2011, 11, 424.
(323) Otero, G.; Gonzalez, C.; Pinardi, A. L.; Merino, P.; Gardonio,
S.; Lizzit, S.; Blanco-Rey, M.; Van de Ruit, K.; Flipse, C. F. J.; Mendez,
J.; de Andres, P. L.; Martn-Gago, J. A. Phys. Rev. Lett. 2010, 105,
216102.
(324) Larsson, M.; Hulten, M.; Blekkan, E. A.; Andersson, B. J. Catal.
1996, 164, 44.
(325) Li, Q.; Sui, Z.; Zhou, X.; Zhu, Y.; Zhou, J.; Chen, D. Top. Catal.
2011, 54, 888.
(326) Land, T. A.; Michely, T.; Behm, R. J.; Hemminger, J. C. Surf.

Sci. 1992, 264, 261.
(327) Li, Q.; Sui, Z.; Zhou, X.; Chen, D. Appl. Catal., A 2011, 398,
18.
(328) Sattler, J. J. H. B.; Beale, A. M.; Weckhuysen, B. M. Phys. Chem.
Chem. Phys. 2013, 15, 12085.
(329) Malleswara Rao, T. V.; Deo, G.; Jehng, J.-M.; Wachs, I. E.
Langmuir 2004, 20, 7159.
(330) Ruettinger, W.; Benderly, A.; Han, S.; Shen, X.; Ding, Y.; Suib,
S. L. Catal. Lett. 2011, 141, 15.
(331) Zhu, Q.; Takiguchi, M.; Setoyama, T.; Yokoi, T.; Kondo, J. N.;
Tatsumi, T. Catal. Lett. 2011, 141, 670.
(332) Grasselli, G. R.; Stern, D. L.; Tsikoyiannis, J. G. Appl. Catal., A
1999, 189, 1.
(333) Van der Zande, L. M.; de Graaf, E. A.; Rothenberg, G. Adv.
Synth. Catal. 2002, 344, 884.
(334) Late, L.; Thelin, W.; Blekkan, E. A. Appl. Catal., A 2004, 262,
63.
(335) Waku, T.; Biscardi, J. A.; Iglesia, E. J. Catal. 2004, 222, 481.
(336) Late, L.; Rundereim, J. I.; Blekkan, E. A. Appl. Catal., A 2004,
262, 53.
(337) Kikuchi, E.; Uemiya, S.; Sato, N.; Inoue, H.; Ando, H.;
Matsuda, T. Chem. Lett. 1989, 18, 489.
(338) Matsuda, T.; Koike, I.; Kubo, N.; Kikuchi, E. J. Membr. Sci.
1993, 77, 283.
(339) Collins, J. P.; Schwartz, R. W.; Sehgal, R.; Ward, T. L.; Brinker,
C. J.; Hagen, G. P.; Udovich, C. A. Ind. Eng. Chem. Res. 1996, 35, 4398.
(340) Kiwi-Minsker, L.; Wolfrath, O.; Renken, A. Chem. Eng. Sci.
2002, 57, 4947.
(341) Sznejer, G.; Sheintuch, M. Chem. Eng. Sci. 2004, 59, 2013.
(342) Bobrov, V. S.; Digurov, N. G.; Skudin, V. V. J. Membr. Sci.
2005, 253, 233.
(343) Liang, W.; Hughes, R. Catal. Today 2005, 104, 238.
(344) Medrano, J. A.; Julia, I.; Garc, F. R.; Li, K.; Herguido, J.; Mene,
M. Ind. Eng. Chem. Res. 2013, 52, 3723.
(345) Chua, Y. T.; Stair, P. C. J. Catal. 2003, 213, 39.
(346) Bennici, S. M.; Vogelaar, B. M.; Nijhuis, T. A.; Weckhuysen, B.
M. Angew. Chem., Int. Ed. 2007, 46, 5412.
10653

catalyticDehydrogenationLightAlkanesonMetals PDF

Hochgeladen von

Dokumentinformationen

Originaltitel

Copyright

Verfügbare Formate

Dieses Dokument teilen

Dokument teilen oder einbetten

Freigabeoptionen

Stufen Sie dieses Dokument als nützlich ein?

Sind diese Inhalte unangemessen?

Copyright:

Verfügbare Formate

catalyticDehydrogenationLightAlkanesonMetals PDF

Hochgeladen von

Copyright:

Verfügbare Formate

Review

Catalytic Dehydrogenation of Light Alkanes on Metals and Metal

Because of their extensive use as chemical building blocks, light

Received: May 4, 2014

dx.doi.org/10.1021/cr5002436 | Chem. Rev. 2014, 114, 1061310653

worldwide, and many of these installations are already under

Table 1. Gas Composition of Six Dierent Shale Gas

1.2. Scope of This Review

Because dehydrogenation for the production of light olens has

Larger hydrocarbons, hydrogen, and oxygen are present in trace

(NGL), such as ethane and propane, which are easily separated

dx.doi.org/10.1021/cr5002436 | Chem. Rev. 2014, 114, 1061310653

vanadium oxide for oxidative dehydrogenation has been further

2. COMMERCIAL DEHYDROGENATION PROCESSES

reaction products.46 A schematic of a Caton dehydrogenation

2.1. Caton Process

2.2. Oleex Process

The Caton process, by CB&I Lummus, is based on the

The Oleex process by UOP uses a very dierent reactor

dx.doi.org/10.1021/cr5002436 | Chem. Rev. 2014, 114, 1061310653

Table 2. Specics of the Caton and Oleex Commercial Dehydrogenation Processes43,44,46,5055

1820 wt % CrOx on an alumina support, promoted by 12 wt %

<1 wt % Pt and 12 wt % Sn on an alumina support, promoted by 01 wt %

dehydrogenation of propane, isobutane and isopentanes

dehydrogenation of propane and isobutane

typically multiple parallel xed bed reactors

heating of the catalyst material during regeneration

a uidized bed reactor consisting of three separate parts: several reactors in

typically 12 min of dehydrogenation, 3 min of purge, 12 min of

diolens and acetylenes. Polymer-quality propylene is obtained

entire cycle every 510 days. Product recovery is done by

dx.doi.org/10.1021/cr5002436 | Chem. Rev. 2014, 114, 1061310653

2.3. Other Patented Alkane Dehydrogenation Processes

The Steam Active Reforming (STAR) process developed by

dx.doi.org/10.1021/cr5002436 | Chem. Rev. 2014, 114, 1061310653

various similarities with the STAR process, as the reaction feed

Figure 6 shows a schematic representation of the Fluidized

3. THERMODYNAMICS OF NONOXIDATIVE ALKANE

However, the reaction is thermodynamically limited and

dx.doi.org/10.1021/cr5002436 | Chem. Rev. 2014, 114, 1061310653

during the dehydrogenation of isobutane to isobutene, the

given that the enthalpy required to dehydrogenate alkanes

dx.doi.org/10.1021/cr5002436 | Chem. Rev. 2014, 114, 1061310653

activation of paranic CH bonds and low activity to CC

process, a compromise must be struck between olen selectivity

Figure 9. Structure sensitivity of ethane hydrogenolysis on 1% Pt/

4.1. Platinum-Based Catalysts

latter reaction being one of the main causes of coke

In the 1960s, a new technology was developed for the synthesis

dx.doi.org/10.1021/cr5002436 | Chem. Rev. 2014, 114, 1061310653

4.1.2. Catalyst Deactivation. Two main processes cause

relatively large surface area and uniform pore size distribution

dx.doi.org/10.1021/cr5002436 | Chem. Rev. 2014, 114, 1061310653

properties of platinum, is necessary to obtain an optimal

in this system. Furthermore, the steric restraints of the zeolite

dx.doi.org/10.1021/cr5002436 | Chem. Rev. 2014, 114, 1061310653

platinum nanoparticles.97,98,104,105 Notably, the formation of a

A secondary eect of Sn addition is the modication of the

dx.doi.org/10.1021/cr5002436 | Chem. Rev. 2014, 114, 1061310653

C3H8 = 25, He = 75; iC4 = 33, H2 =

(2) 2 wt % Pt1 wt % Sn/CeO2/C

WHSV = 15.1; WHSV

space velocity (h1)