Applied Catalysis A: General 221 (2001) 397–419

Dehydrogenation and oxydehydrogenation of paraffins to olefins

M.M. Bhasin a,∗ , J.H. McCain a , B.V. Vora b , T. Imai b , P.R. Pujadó b
a Union Carbide & UOP LLC, Subsidiary of Dow Chemical Co., P.O. Box 8361, 25303 South Charleston, WV, USA
b UOP LLC, Subsidiary of Dow Chemical Co., P.O. Box 8361, 25303 South Charleston, WV, USA

Abstract
Catalytic paraffin dehydrogenation for the production of olefins has been in commercial use since the late 1930s, while
catalytic paraffin oxydehydrogenation for olefin production has not yet been commercialized. However, there are some
interesting recent developments worthy of further research and development.
During World War II, catalytic dehydrogenation of butanes over a chromia-alumina catalyst was practiced for the production
of butenes that were then dimerized to octenes and hydrogenated to octanes to yield high-octane aviation fuel. Dehydrogenation
employs chromia-alumina catalysts and, more recently, platinum or modified platinum catalysts. Important aspects in dehydro-
genation entail approaching equilibrium or near-equilibrium conversions while minimizing side reactions and coke formation.
Commercial processes for the catalytic dehydrogenation of propane and butanes attain per-pass conversions in the range of
30–60%, while the catalytic dehydrogenation of C10 –C14 paraffins typically operates at conversion levels of 10–20%. In the
year 2000, nearly 7 million metric tons of C3 –C4 olefins and 2 million metric tons of C10 –C14 range olefins were produced
via catalytic dehydrogenation.
Oxydehydrogenation employs catalysts containing vanadium and, more recently, platinum. Oxydehydrogenation at
∼1000 ◦ C and very short residence time over Pt and Pt-Sn catalysts can produce ethylene in higher yields than in steam
cracking. However, there are a number of issues related to safety and process upsets that need to be addressed. Important
objectives in oxydehydrogenation are attaining high selectivity to olefins with high conversion of paraffin and minimizing
potentially dangerous mixtures of paraffin and oxidant. More recently, the use of carbon dioxide as an oxidant for ethane
conversion to ethylene has been investigated as a potential way to reduce the negative impact of dangerous oxidant–paraffin
mixtures and to achieve higher selectivity.
While catalytic dehydrogenation reflects a relatively mature and well-established technology, oxydehydrogenation can in
many respects be characterized as still being in its infancy. Oxydehydrogenation, however, offers substantial thermodynamic
advantages and is an area of active research in many fronts. © 2001 Elsevier Science B.V. All rights reserved.
Keywords: Paraffin dehydrogenation; Olefin; Chromia-alumina catalyst; Paraffin oxydehydrogenation; Noble metal catalysts

1. Historical overview and chromia-alumina
catalysts
Paraffin dehydrogenation for the production of
olefins has been in use since the late 1930s. During
∗ Corresponding author. Tel.: +1-304-747-4910;
fax: +1-304-747-5430.
E-mail address: bhasinmm@dow.com (M.M. Bhasin).
World War II, catalytic dehydrogenation of butanes
over a chromia-alumina catalyst was practiced for
the production of butenes, which were then dimer-
ized to octenes and hydrogenated to octanes to yield
high-octane aviation fuel.
Dehydrogenation of butanes over a chromia-alumina
catalyst was first developed and commercialized at
Leuna in Germany and was also independently de-
veloped by UOP (then Universal Oil Products) in the

0926-860X/01/$ – see front matter © 2001 Elsevier Science B.V. All rights reserved.
PII: S 0 9 2 6 - 8 6 0 X ( 0 1 ) 0 0 8 1 6 - X
398 M.M. Bhasin et al. / Applied Catalysis A: General 221 (2001) 397–419
United States, together with ICI in England. The first
UOP-designed plant came on stream in Billingham,
England, in 1940 and was soon followed by two other
units in Heysham, England, in 1941 [1]. The primary
purpose of this butane dehydrogenation was to pro-
duce butenes, which were then dimerized to octenes
using solid phosphoric acid catalysts discovered by
Schaad and Ipatieff [2].
Other companies soon followed these pioneering
efforts. For example, Phillips Petroleum built a multi-
tubular dehydrogenation reactor near Borger, TX, in
1943 [1]. However, the most significant development
was made by Houdry using dehydrogenation at less
than atmospheric pressure for higher per-pass conver-
sions. This process, which came on stream toward the
end of World War II, was also used for the production
of butenes. After the war, Houdry further developed
and commercialized the chromia-alumina dehydro-
genation system and extended it to the production of
butadiene in what became known as the CatadieneTM
process [3]. Other companies, including Shell, Gulf,
and Dow, also practiced similar dehydrogenation
technologies.
In the dehydrogenation process using chromia-
alumina catalysts, the catalyst is contained in a fixed
shallow bed located inside a reactor that may be either
a sphere, a squat vertical cylinder, or a horizontal cylin-
der. The actual design reflects a compromise between
gas flow distribution across a large cross-sectional
area and the need to maintain a low pressure drop. A
significant amount of coke is deposited on the catalyst
during the dehydrogenation step, therefore, a number
of reactors are used in parallel—some for dehydro-
genation while the rest are being purged or regen-
erated. The dehydrogenation reactions are strongly
endothermic, and the heat is provided, at least in
part, by the sensible heat stored in the catalyst bed
during regeneration (carbon burn); additional heat is
provided by direct fuel combustion and also by heat
released in the chromium redox cycle. The length of
the total reactor cycle is limited by the amount of heat
available, and can be as short as 10–20 min.
The Houdry Catadiene process was used exten-
sively for the production of butadiene, either by
itself (n-butane to butadiene) or in conjunction with
catalytic oxydehydrogenation of n-butene to butadi-
ene. The latter was commercialized by the Petro-Tex
Chemical Corp. [3] and was called the Oxo-DTM
process. A similar oxydehydrogenation approach for
the production of butadiene was also practiced by
Phillips Petroleum [3].
Large quantities of butadiene have become available
over the past 30 years, mostly as a by-product from the
thermal cracking of naphtha and other heavy hydrocar-
bons. This market shift has resulted in the shutdown
of all on-purpose catalytic dehydrogenation units for
butadiene production in North America, western Eu-
rope, and the far East.
In the late 1980s, the application of chromia-alumina
catalysts was extended by Houdry to the dehydrogena-
tion of propane to propylene and isobutane to isobuty-
lene. The new process application called CatofinTM
[4,5] operates on the same cyclic principle as in the
former Catadiene process. As of late 2000, a total of
eight Catofin units exist for the production of isobuty-
lene (including two converted older Catadiene units)
with an aggregate capacity of about 2.8 million metric
tons per annum (MTA) isobutylene. In addition, two
Catofin units were built for the production of propy-
lene, but it is understood that only one is operational
with a nameplate dehydrogenation capacity of about
250,000 MTA propylene, but usually operating only
on a seasonal basis. Plans for another 450,000 MTA
Catofin propane dehydrogenation unit in Saudi Ara-
bia have also been announced. The Catofin process
technology is currently owned by Süd-Chemie and is
offered for license by ABB Lummus.
In 1959, an alternative chromia-alumina catalytic
dehydrogenation process was developed in the for-
mer Soviet Union. This process avoided the use of the
cyclic operation by using a fluidized bed reactor con-
figuration similar to the fluidized catalytic cracking
(FCC) process used in refineries [6]. However, back-
mixing common to dense fluidized bed operations re-
sults in poor selectivity and increases the formation
of heavies, sometimes called “green oils”. Circulating
regenerated catalyst is used to provide the heat of re-
action in the riser and spent catalyst is reheated by
carbon burn in the regenerator. During the 1990s, a
large scale fluid bed isobutane dehydrogenation unit
for about 450,000 MTA isobutylene was commercial-
ized by Snamprogetti in Saudi Arabia based on tech-
nology from Yarsintez in Russia [6], but it is under-
stood that this unit has only operated at lower than de-
sign capacity. Recent literature articles report further
improvements by Snamprogetti [7,8].
 M.M. Bhasin et al. / Applied Catalysis A: General 221 (2001) 397–419 399

2. Noble metal dehydrogenation catalysts

A different approach to catalytic dehydrogenation

was first introduced in the mid-1960s for the sup-
ply of long-chain linear olefins for the production of
biodegradable detergents.
Synthetic detergents, based on the use of branched
alkylbenzene sulfonates derived from propylene
tetramer and benzene, had been introduced in the
1940s. By the early 1960s, however, it became appar-
ent that branched dodecylbenzene-based detergents,
though very active and offering excellent perfor-
mance characteristics, did not biodegrade readily and
were accumulating in the environment. The need for
biodegradable detergents prompted the development
of catalytic dehydrogenation of long-chain linear
paraffins to linear olefins.
The work on catalytic reforming with noble metal Fig. 1. Propane dehydrogenation equilibrium at 1.00 atm abs. pres-
sure.
(Pt) catalysts done in the 1940s by Haensel clearly
demonstrated that Pt-based catalysts had high activity
for the dehydrogenation of paraffins to the correspond- where xe is the equilibrium conversion, P the total ab-
ing olefins [9]. In the 1960s, Bloch [10] further ex- solute pressure and Kp is the equilibrium constant for
tended this thinking by developing Pt-based catalysts the dehydrogenation reaction. The equilibrium con-
that could selectively dehydrogenate long-chain linear stant can be easily calculated from Gibbs free energies
paraffins to the corresponding internal mono-olefins as tabulated in the API 44 report or in similar sources
with high activity and stability and with minimum of thermodynamic data. Figs. 1 and 2 illustrate the
cracking. This was the basis for the UOP PacolTM pro- equilibrium conversion levels that can be obtained for
cess for the production of linear olefins for the man- propane at 1 and 0.23 atm abs. (175 Torr), respectively.
ufacture of biodegradable detergents [11]. In 1999,
there were more than 30 commercial Pt-catalyzed de-
hydrogenation units in operation for the manufacture
of detergent alkylate.
Long-chain paraffins are both valuable and highly
prone to cracking. Therefore, in order to maintain high
selectivity and yield, it is necessary to operate at rela-
tively mild conditions, typically below 500 ◦ C, and at
relatively low per-pass conversions. While this is eco-
nomical for the production of heavy linear olefins, it
is not for the production of light olefins.
Paraffin dehydrogenation is an endothermic reaction
that is limited by chemical equilibrium and, accord-
ing to Le Chatelier’s principle, higher conversion will
require either higher temperatures or lower pressures.
In a somewhat abbreviated form for the production of
mono-olefins, this can be expressed as follows:

Kp
xe2 = Fig. 2. Propane dehydrogenation equilibrium at 0.23 atm abs. pres-
Kp + P sure.
400 M.M. Bhasin et al. / Applied Catalysis A: General 221 (2001) 397–419
Fig. 3. Equilibrium constants for n-paraffin dehydrogenation at
500 ◦ C.
The equilibrium constant for paraffin dehydrogena-
tion increases significantly as the carbon number in-
creases. Fig. 3 shows the equilibrium constant for the
dehydrogenation of n-paraffins ranging from ethane
to pentadecane [12]. Fig. 4 shows the temperatures
required to achieve 10–40% equilibrium conversion
based on these equilibrium constants. Fig. 4 indicates
that the temperature required for the dehydrogenation
of light paraffins is much higher than for heavy paraf-
fins. For 40% conversion, for example, the dehydro-
genation of propane requires a temperature of at least
about 580 ◦ C, while dodecane can be theoretically
dehydrogenated to the same extent at only 450 ◦ C.
The equilibrium conversion increases at higher
Fig. 4. Temperatures required to achieve 10 and 40% conversion
of C2 –C15 n-paraffins at 1 atm.
temperatures, but side reactions, coke formation, and
catalyst deactivation are also accelerated. Thus, ex-
trapolation directly from heavy olefins to light olefins
cannot be done without taking other factors into
consideration.
Production of light olefins by the catalytic dehydro-
genation of light paraffins must be able to maintain
reasonable per-pass conversion levels and high olefin
selectivity. Very importantly, it must be able to pro-
duce olefins in high yields over long periods of time
without shutdowns.
In the early 1970s, UOP introduced continuous cata-
lyst regeneration (CCR) technology that enabled noble
metal catalysts to remain at their most desirable stable
activity for several years without having to shut down
the reactor for catalyst regeneration. The combination
of noble metal catalysts operating at high severity in
conjunction with CCR technology made it possible to
design, build, and economically operate large catalytic
dehydrogenation units that can produce light olefins,
in particular, propylene and isobutylene, at high selec-
tivities while still operating at superatmospheric pres-
sures. This technology is known as the UOP OleflexTM
process. As of late 2000, there were four propane de-
hydrogenation units, five isobutane dehydrogenation
units, and one combined propane/isobutane dehydro-
genation unit of this type in commercial operation,
with an aggregate operating capacity of 900,000 MTA
polymer grade propylene and 2.3 million MTA
isobutylene. In addition, another propane dehydro-
genation unit for 350,000 MTA polymer grade propy-
lene was under design and construction.
The world propylene production capacity, based on
the use of catalytic dehydrogenation of propane has
increased steadily over the past 10 years [13] and is ex-
pected to grow even further under the right economic
conditions relative to the availability of propane; on
the other hand, environmental concerns on the use of
MTBE are expected to adversely impact the future ex-
pansion of isobutane dehydrogenation applications.
Although production of ethylene via catalytic dehy-
drogenation over Pt catalysts is very selective (about
95%), extension of this dehydrogenation technology
to ethane has not taken place due to the need for even
more severe operating conditions; higher temperatures
and lower pressures. Such conditions cause excessive
coking of the catalyst or require costlier operation
under vacuum.
 M.M. Bhasin et al. / Applied Catalysis A: General 221 (2001) 397–419 401

Fig. 5. Reactions by platinum and acid sites in light paraffin dehydrogenation with unmodified catalyst.

Practically, all existing catalytic dehydrogenation 3. Process chemistry

capacity based on Pt catalysts is based on the Oleflex
process with CCR; however, there are also two smaller
units for isobutane dehydrogenation for 118,000 and
13,000 MTA isobutylene, respectively, both based on
the STAR technology developed by Phillips Petroleum
and derived from their earlier multitubular reactor de-
sign experience. This reactor design resembles a typ-
ical steam reformer that is operated until the catalyst
deactivates as a result of coke deposition; banks of
tubes are sequentially taken out of service for cata-
lyst regeneration. The STAR technology is currently
owned and licensed by Krupp–Uhde.
The main reaction in catalytic dehydrogenation is
the formation of mono-olefins from the correspond-
ing feed paraffin. Other reactions include consecutive
and side reactions. The reaction pathways involved in
heavy paraffin dehydrogenation (e.g. detergent-range
C10 –C14 n-paraffins) are more complicated than those
in light paraffin dehydrogenation (e.g. propane and
isobutane). The main difference in reaction pathways
is that a significant amount of cyclic compounds can
form via dehydrocyclization from heavy paraffins;
this is not the case for light paraffins. Figs. 5 and 6

Fig. 6. Reactions by platinum and acid sites in heavy paraffin dehydrogenation with unmodified catalyst.
402 M.M. Bhasin et al. / Applied Catalysis A: General 221 (2001) 397–419
illustrate possible reactions that take place on Pt and
acid (A) sites, respectively, in the dehydrogenation of
light and heavy paraffins when the catalyst is not se-
lective, e.g. unmodified platinum catalysts supported
on alumina.
The consecutive reactions, the dehydrogenation of
mono-olefins to diolefins and triolefins, are catalyzed
on the same active sites as the dehydrogenation of
paraffins to mono-olefins. The consecutive reactions
that form triolefins, aromatics, dimers, and polymers
must be suppressed kinetically or by catalyst modifi-
cations.
4. Role of catalysts and supports
The discussion in this section pertains to alumina-
supported platinum catalysts. The work by Poole and
coworkers [14,15] provides an extensive review of
chromia-alumina catalysts.
The key role of dehydrogenation catalysts is to
accelerate the main reaction while controlling other
reactions. Unmodified alumina-supported platinum
catalysts are highly active but are not selective to de-
hydrogenation. Various by-products, as indicated in
Figs. 5 and 6, can also form. In addition, the catalyst
rapidly deactivates due to fouling by heavy carbona-
ceous materials. Therefore, the properties of platinum
and the alumina support need to be modified to sup-
press the formation of by-products and to increase
catalytic stability.
Fig. 7. Paraffin dehydrogenation on modified Pt catalyst.
The reaction of olefins on platinum is faster than
that of paraffins, because olefins interact with platinum
more strongly than do paraffins. The role of platinum
modifiers is to weaken the platinum–olefin interaction
selectively without affecting the platinum–paraffin in-
teraction. Arsenic, tin, germanium, lead, bismuth are
among metals reported as platinum activity modifiers.
The consecutive dehydrogenation rate of mono-olefins
and diolefins is decreased by this modification without
lowering the rate of paraffin dehydrogenation signifi-
cantly. The modifier also improves the stability against
fouling by heavy carbonaceous materials.
Platinum is a highly active catalytic element and is
not required in large quantities to catalyze the reaction
when it is dispersed on a high surface-area support.
The high dispersion is also necessary to achieve high
selectivity to dehydrogenation relative to undesirable
side reactions, such as cracking.
The typical high surface area alumina supports em-
ployed have acidic sites that accelerate skeletal isomer-
ization, cracking, oligomerization, and polymerization
of olefinic materials, and enhance “coke” formation.
Alkali or alkaline earth metals assist in the control of
the acidity. Also, ␣-alumina supports that have essen-
tially no acidity can be utilized; however, the challenge
is to obtain high dispersion of platinum on such very
low surface area supports. Therefore, acidity must be
eliminated by using suitable modifiers.
Modified catalysts possess high activity and high
selectivity to mono-olefins. The major by-products
are diolefins that can be controlled kinetically. Coke
 M.M. Bhasin et al. / Applied Catalysis A: General 221 (2001) 397–419
formation is also suppressed and, therefore, stabil-
ity is greatly improved. Over modified catalysts,
the major reaction pathways for both light and
heavy paraffin dehydrogenation systems are simpler
(Fig. 7).
Alumina has excellent thermal stability and me-
chanical strength under processing, transport, and
catalyst regeneration conditions. However, the most
important reason alumina is used as support ma-
terial is its superior capability to maintain a high
degree of platinum dispersion, which is essential
for achieving high dehydrogenation activity and
selectivity.
The catalytic reaction rate is limited by the intra-
particle mass transfer rate. If the mass transfer rate is
relatively slow, both activity and selectivity are low-
ered. As a result, the support must have a low pore
diffusional resistance (high effectiveness factor). For
a given pore volume, the surface area and the strength
of the support increase as the pore diameter decreases,
and the pore diffusional resistance decreases as the
pore diameter increases. Thus, an appropriate pore
structure must be determined for the support to achieve
optimal catalytic performance.
5. Dehydrogenation catalyst evaluation
In paraffin dehydrogenation, the rate of paraffin con-
version (x) and mono-olefin production (sx) are given
by Eqs. (1) and (2) respectively:
dx
= f1 (ki , Ki pj ) (1)
dt
d(sx)
= f2 (ki , Ki , pj ) (2)
dt
where s is the selectivity to n-mono-olefins, t the con-
tact time, f the rate function, ki the rate constant for
reaction step i, Ki the equilibrium constant for reaction
step i, and pj is the partial pressure of the j compound.
The following relationship between selectivity and
conversion can be derived from Eqs. (1) and (2):
d(sx) f2 (ki , Ki , pj )
= (3)
dx f1 (ki , Ki , pj )
The ki and Ki are a function of temperature and pj
is a function of conversion, total pressure (P), and feed
403
ratio (R). Therefore, Eq. (3) can be rewritten as
d(sx) f2 [ki (T ), Ki (T ), x, P , R]
= (4)
dx f1 [ki (T ), Ki (T ), x, P , R]
As Eq. (4) is the ratio of two functions, the rate
constants become relative values and can be expressed
as ki (T)/k0 (T), where k0 (T) is the rate constant for
the forward reaction of paraffin dehydrogenation to
mono-olefins.
Eq. (4) can be written in a functional form in F as
follows:


d(sx) ki (T )
=F , Ki (T ), x, P , R (5)
dx k0 (T )
Eq. (5) indicates that selectivity is a function of con-
version for the catalyst used (relative rate constants)
and the given reaction conditions (temperature, pres-
sure, feed ratio).
Selectivity decreases as the conversion increases
because n-mono-olefins are consecutively converted
into by-products. Selectivity decreases sharply as
conversion approaches equilibrium because the main
dehydrogenation process is limited by equilibrium,
but other reactions continue to occur. Therefore,
if side reactions are controlled, the selectivity is
improved as the equilibrium conversion becomes
higher by increasing the temperature and by decreas-
ing the pressure and the feed ratio of hydrogen to
paraffin.
The relationship between selectivity and conver-
sion can be simulated according to Eq. (5), if rate
functions, relative rate constants, and equilibrium
constants are known. Fig. 8 shows simulated selectiv-
ities to n-heptene and n-heptadiene for the dehydro-
genation of n-heptane. In this simulation, the relative
rate constants used are unity, which represents that
the catalyst possesses perfect selectivity regarding
consecutive dehydrogenation; the dehydrogenation
rate of paraffin is equal to that of mono-olefin and
diolefin. Experimental selectivities obtained over a
UOP dehydrogenation catalyst show good agreement
with the predicted values.
The rate of light paraffin conversion (Eq. (1)) over
a Pt catalyst (Oleflex type process) can be expressed
as a modified first order equation according to a
Langmuir–Hinshelwood mechanism. The rest of the
equations may be derived accordingly.
404 M.M. Bhasin et al. / Applied Catalysis A: General 221 (2001) 397–419

Snamprogetti’s dehydrogenation process consists

of a fluidized bed reactor and regeneration system.
Here too the coke build-up is very low and the
“regeneration” loop is actually a means of supplying
heat to the reactor.

7. Heat of reaction

The heat of reaction for paraffin dehydrogenation

is about 30 kcal/mol (125 kJ/mol). In a cyclic adi-
abatic operation (e.g. Houdry), heat is provided by
reheating the catalyst to a high temperature during
the regeneration step, so that the catalyst cools down
and conversion decreases during the reaction step; be-
cause several reactors are used in parallel, an average
Fig. 8. Simulation of selectivity for dehydrogenation of n-heptane. conversion is obtained. In an isothermal process (e.g.
STAR), the catalyst is loaded inside vertical tubes
inside a furnace and the heat is introduced through
6. Catalyst stability and regeneration
the tube walls. In a fluidized reactor, the temperature
profile can be maintained uniformly in the backmixed
The dehydrogenation of long-chain paraffins is per-
zone of the bed, while heat is provided by intro-
formed under relatively mild temperature conditions
ducing hot regenerated catalyst. In Oleflex adiabatic
of 400–500 ◦ C. Thus, the catalyst can maintain a long
reactors, a significant temperature drop occurs across
life even at high space velocity, and high catalyst pro-
the catalyst bed which lowers the equilibrium conver-
ductivity. Therefore, it is not economical to build fa-
sion level; a multistage reactor system with interstage
cilities for catalyst regeneration.
reheating is used for higher paraffin conversions.
Because of equilibrium limitations, the dehydro-
Fig. 9 illustrates conversion, equilibrium conver-
genation of light paraffins requires significantly higher
sion, and temperature along the catalyst bed in a
temperatures above 600 ◦ C to achieve economically
attractive conversions. The catalyst deactivation is
accelerated under high-temperature conditions, and
frequent catalyst regeneration is necessary for light
paraffin dehydrogenation. For the dehydrogenation of
light paraffins, a number of different types of reactor-
regeneration systems are commercially utilized.
Houdry’s Catofin and similar processes employ a
cyclic sequence of steps—process, purge, air regener-
ation, purge, hydrogen reduction, and back to process.
The Phillips STAR process also regenerates the cat-
alyst on a cyclic basis, but while the Houdry regener-
ation is actually a mechanism to provide the heat for
the reaction even when coke build-up is still very low,
the catalyst in the isothermal STAR process is only
regenerated after coke has accumulated to appreciable
levels that result in low catalyst activity.
UOP’s Oleflex process uses multi-stage adiabatic Fig. 9. Temperature profile and conversions of three-stage isobutane
reactors with CCR. dehydrogenation process.
 M.M. Bhasin et al. / Applied Catalysis A: General 221 (2001) 397–419
Fig. 10. Isobutane dehydrogenation.
three-stage adiabatic reactor system for the dehydro-
genation of isobutane. For propane dehydrogenation,
a four-stage reactor system becomes more economi-
cal because higher average temperatures are needed.
A multi-stage reactor system also affords lower inlet
temperatures, relative to a single stage reactor system
Fig. 11. Catofin process flow diagram.
405
(Fig. 10). Thus, thermal cracking and catalyst deacti-
vation, which are accelerated at higher temperatures,
can be controlled to low levels.
8. Process flow and reactor characteristics
8.1. Cyclical processes
As described earlier, the Houdry Catadiene process,
the Houdry Catofin process, and other similar cyclical
processes make use of parallel reactors that contain a
shallow bed of chromia-alumina catalyst. Fig. 11 il-
lustrates a schematic of such a process. This technol-
ogy has been used extensively for the production of
butadiene and, in more recent years, for the produc-
tion of isobutylene and propylene [16,17]. The feed is
preheated through a fired heater before being passed
over the catalyst in the reactors. The hot reactor ef-
fluent is cooled, compressed, and sent to the product
fractionation and recovery section. The dehydrogena-
tion reactors are refractory-lined carbon steel vessels
406 M.M. Bhasin et al. / Applied Catalysis A: General 221 (2001) 397–419
Fig. 12. Typical timing cycle for a five reactor system.
(i.e. cold wall design). In order to accommodate con-
tinuous flow of the main streams (hydrocarbons and
regeneration air), the reactors are operated on a timing
cycle that satisfies the following requirement [1]:
on-stream time + regeneration time + purge time
= total cycle time.
The number of reactors in each cycle is the prorated
time fraction of the total cycle time. Thus, with five
reactors, two reactors can be on stream simultaneously,
two on regeneration, and one on purge, evacuation, and
valve changes. Fig. 12 provides a typical timing cycle
for a five-reactor unit [1], but as many as eight reactors
in parallel have been provided in some units. The total
cycle time is usually in the range of 15–30 min.
The on-stream period at sub-atmospheric pressure
is followed by a purge. Next comes regeneration at es-
sentially atmospheric pressure, followed by purge, hy-
drogen reduction, and evacuation to reaction pressure,
after which the reactor is ready for another on-stream
period. Process streams enter and leave the reactors
through fast-acting gate valves. The gate valves can
range up to 40 in. (1 m) in diameter, are designed
for high-temperature service, and are equipped with a
pressurized inert seal in the bonnet to prevent leakage
of air into the process gas when the valve is closed.
Overall, this mechanical design has proven to be very
reliable over many years of operation.
The regeneration is done with air that has been pre-
heated through a direct fired burner or, alternatively,
with the exhaust of a gas turbine. The regeneration
step is intended to preheat the catalyst to the on-stream
temperature necessary to initiate the next process cycle
and to remove coke deposits on the catalyst. Flue gas
sensible heat may be recovered in a waste heat boiler.
The hydrogenation step prepares the catalyst for the
dehydrogenation phase and also contributes additional
heat from the reduction of Cr6+ –Cr3+ .
Another cyclical process is the Phillips STAR
(steam active reforming) process [18]. This process
uses a fixed-bed fired-tube reactor operating at a pos-
itive superatmospheric pressure. In many respects, it
is similar in design to a steam reforming furnace with
the heat of reaction provided by firing outside the
tubes, thus operating at near isothermal conditions.
M.M. Bhasin et al. / Applied Catalysis A: General 221 (2001) 397–419 407
408 M.M. Bhasin et al. / Applied Catalysis A: General 221 (2001) 397–419
Steam is used as a diluent to lower the partial pres-
sure of the reactants and, thus, to achieve reasonable
conversion levels of about 30–40% for propane and
45–55% for butanes. Steam also helps slow down the
deposition of carbon (coke) on the catalyst, thereby,
extending cycle time from minutes to hours.
Periodic catalyst regeneration or carbon burnoff is
required to maintain the activity of the catalyst. Typ-
ical cycle time is reported to be at least 8 h, with 7 h
of process time and 1 h of regeneration time. For con-
tinuous operation, various furnace modules can be op-
erated such that, for example, seven operate in the
process mode while one is in the regeneration mode.
Fig. 13 shows a schematic diagram of a STAR process
unit [18].
9. Continuous processes
Snamprogetti, an Italian company, has commer-
cialized fluidized bed dehydrogenation (FBD) for the
catalytic dehydrogenation of light paraffins using a
chromia-alumina catalyst with an alkaline promoter
[6–8], which is used primarily for the dehydrogena-
tion of isobutane to isobutylene during the manufac-
ture of MTBE. The catalyst is microspheroidal with
Fig. 14. Snamprogetti’s FBD process scheme.
an average diameter <100 ␮m and an apparent bulk
density <2000 kg/m3 [19]. The heat of reaction is
provided by circulating hot regenerated catalyst back
to the reactor. In all concepts, the FBD process is
very similar to the FCC process units commonly used
in petroleum refineries. However, because backmix-
ing has a negative effect on the yields, horizontal
baffles with suitable openings are inserted within the
fluidized bed to limit the back-flow of solids, such
that the fluidized bed is split into a series of stages,
each comparable to a CSTR [19]. A typical process
scheme is shown in Fig. 14.
Fresh feed is vaporized, mixed with the recycle of
unconverted paraffins, and fed to the fluidized reactor
through a distributor for optimal even distribution. En-
trained catalyst is removed from the product off-gas by
means of cyclones. Catalyst circulates continuously
from the reactor to the regenerator and vice-versa by
means of transfer lines. Coke deposited on the catalyst
is burnt off in the regenerator; however, because the
amount of coke is relatively small, additional fuel must
be burnt in the regenerator in order to satisfy the ther-
mal requirements of the endothermic dehydrogenation
reaction. However, while this approach is similar to
that in the Houdry process, FBD does not have a cat-
alyst reduction step with hydrogen before proceeding
 M.M. Bhasin et al. / Applied Catalysis A: General 221 (2001) 397–419 409

Fig. 15. UOP Pacol dehydrogenation process.

to the dehydrogenation cycle; lack of this step is be- performances of the Pacol and Oleflex processes are
lieved to be detrimental to the overall performance of summarized in Table 1.
the process. Use of the Oleflex process for the dehydrogenation
UOP’s catalytic dehydrogenation processes typi- of ethane to ethylene has also been investigated but, to
cally make use of radial flow adiabatic fixed-bed (or date, the economics do not appear to be favorable be-
slowly moving bed) reactors with modified Pt-alumina cause of the low equilibrium conversion and the need
catalysts. to operate at a pressure lower than atmospheric if a rea-
The UOP Pacol process for selective long-chain sonable ethane conversion is to be expected. The cost
paraffin dehydrogenation to produce linear mono- of fractionating ethylene in an ethane–ethylene split-
olefins is shown in Fig. 15 in combination with the ter is otherwise too high. Dow Chemical has recently
UOP detergent alkylation process. The Pacol process been awarded a patent [21] for the dehydrogenation
consists of a radial-flow reactor and a product recov- of ethane over a metal-mordenite catalyst complex
ery section. Worldwide, more than 2 million MTA of at relatively low conversions in which the product
linear alkyl benzene (LAB) is produced employing
this process [20].
Table 1
The flow diagram of the UOP Oleflex process Performance of Pacol and Oleflex processes
is shown in Fig. 16. The process consists of a re-
Process Feed Conversion (%) Selectivity (%)
actor section and a product recovery section. The
reactor section consists of three or four stages of Oleflex Propane 40 90
radial-flow reactors, charge and interstage heaters, n-Butane 50 85
Isobutane 50 92
reactor feed-effluent exchangers, and the CCR sec-
tion (Fig. 17). As noted earlier, today more than 1 Pacol n-Heptane 20 90
million metric tons propylene and 2 million metric n-C10 –C13 13 90
n-C11 –C14 13 90
tons isobutylene are produced via this route [13]. The
410 M.M. Bhasin et al. / Applied Catalysis A: General 221 (2001) 397–419

Fig. 16. UOP Oleflex process.

Fig. 17. Oleflex regeneration section.

 M.M. Bhasin et al. / Applied Catalysis A: General 221 (2001) 397–419
Table 2
Characteristics of five reactor systemsa
Downflow
Low pressure drop
Plug flow – –
Catalyst addition or removal
High heat transfer, near isothermal
Variable heat transfer coefficient
a The choice of the right reactor depends on the catalyst and the selection of operating conditions. The ‘dash’ represents beneficial
characteristics of each reactor type.
ethylene is selectively recovered from the dilute
ethylene–ethane stream by alkylating it with benzene.
10. Reactor design options
The choice of reactor design plays a very important
role in the success of catalytic processes. The follow-
ing types of reactor design are commercial today for
endothermic catalytic dehydrogenation processes:
• downflow adiabatic fixed-bed;
• radial flow fixed-bed or moving bed adiabatic;
• tubular isothermal and
• fluidized bed
Table 2 summarizes the main characteristics of the
four reactor systems.
11. Other dehydrogenation technologies
The processes discussed above are for the direct cat-
alytic dehydrogenation of paraffins to the correspond-
ing olefins or of olefins to diolefins. Other approaches
have also been considered, although none has reached
the level of commercialization. Some of the most no-
table are
• halogen-assisted dehydrogenation and
• oxydehydrogenation.
Use of halogens for the dehydrogenation of paraf-
fins has been proposed in different ways. For example,
heavy paraffins were first chlorinated and then dehy-
drochlorinated to heavy olefins commercially in the
past both by Shell and by Hüls, among others. Pyrol-
ysis of methane in the presence of chlorine has been
411
Radial flow Tubular Fluidized bed
– –
–
– –
– –
proposed by Benson [22] for the production of
acetylene and ethylene. Other chlorination/dehydro-
chlorination cycles have been proposed for the
production of ethylene from ethane. Propane dehydro-
genation in the presence of iodine via a propyl iodide
intermediate has also been proposed [4,23]. Apart
from the apparent corrosion problems associated with
the use of halogens, other difficulties readily come
to mind owing to the relatively high cost of chlorine,
and even more so of iodine, and the need to either
dispose of or recycle vast quantities of halogens.
Oxydehydrogenation or oxidative dehydrogenation
can be considered in at least two different ways.
Use of oxygen to oxidize the hydrogen coproduct
from dehydrogenation, and thus to displace the dehy-
drogenation equilibrium to higher conversions. This
approach has been used commercially in the catalytic
dehydrogenation of ethylbenzene to styrene as in
the UOP Styro-PlusTM process or in the ABB Lum-
mus/UOP SMARTTM process, but to date has not
succeeded in the dehydrogenation of light or heavy
paraffins. This technology has been used in a styrene
unit at Mitsubishi Chemicals, Kashima, Japan. Al-
though a similar approach has been proposed for the
dehydrogenation of paraffins [24–27], it has not been
commercialised.
Direct use of oxygen as a means of dehydrogenat-
ing, for example, ethane to ethylene. Oxydehydro-
genation has successful commercial applications in the
conversion of n-butenes to butadiene (e.g. as in the
Oxo-D process referred to earlier), but not yet for the
production of ethylene or propylene. This subject is
analyzed in more detail in the following section.
Use of oxydehydrogenation relative to straight cat-
alytic dehydrogenation must be viewed both in terms
of safety issues and in an economic context. On the
412 M.M. Bhasin et al. / Applied Catalysis A: General 221 (2001) 397–419
latter, even though oxydehydrogenation offers advan-
tages as a means of overcoming thermodynamic equi-
librium limitations, it also leads to the total or partial
loss of byproduct hydrogen, which in some instances
can have a very significant economic impact.
Although less apparent, oxydehydrogenation also
plays a role in the work done by BP Amoco, Asahi,
and others to extend ammoxidation to the direct con-
version of propane to acrylonitrile. It is believed that
the ammoxidation of propane proceeds through a tran-
sient propylene intermediate from which acrylonitrile
is derived through a conventional ammoxidation path-
way [28].
11.1. Oxydehydrogenation of ethane and propane
Oxydehydrogenation of ethane and propane as a
route to ethylene and propylene, respectively, has
the attractive feature of removing the equilibrium
conversion restriction of dehydrogenation. Table 3
contrasts the calculated equilibrium conversions of
oxydehydrogenation of ethane and propane with the
calculated equilibrium conversions for dehydrogena-
tion. The formation of water rather than hydrogen
in oxydehydrogenation effectively removes the equi-
librium constraint on conversion at all temperatures
of interest. At the same time, oxydehydrogenation
is not without its own set of challenges that, in the
case of ethane and propane, have kept oxydehydro-
genation from being practiced on a commercial scale.
Challenges in oxydehydrogenation include handling
mixtures of paraffins and oxygen, which can be ex-
plosive at certain compositions, suitable conversion of
paraffins, which often is limited by maintaining a safe
paraffin–oxygen composition, and suitable selectivity
to olefins. The carbon oxides—carbon monoxide and
carbon dioxide—are thermodynamically more stable
Table 3
Percentage conversions at equilibrium for dehydrogenation and oxydehydrogenation of ethane and propane at atmospheric pressurea
Temperature (K) Ethane
Dehydrogenation
400 <1
600 <1
800 7 100
1000 51
a Stoichiometric mixture of alkane and oxygen for oxydehydrogenation.
than the olefins, and so catalysts must be found that
can stop the reaction at the olefin rather than allowing
it to proceed on to the oxides. Also, to compete with
steam cracking, the selectivity to olefins must be quite
high. Selectivity to ethylene from ethane in steam
cracking is reported to be about 84% at 54% ethane
conversion (800 ◦ C, 0.3 kg steam per kg feed, 0.79 s
residence time, and 154 kPa hydrocarbon partial pres-
sure), and 78% at 69% ethane conversion (833 ◦ C,
0.3 kg steam per kg feed, 0.75 s residence time, and
154 kPa hydrocarbon partial pressure) [29].
11.1.1. Ethane oxydehydrogenation
Ethane oxydehydrogenation as a route to ethylene
has been examined at temperatures in the range of
300–500 ◦ C with reducible metal oxide catalysts, and
at higher temperatures, above about 600 ◦ C, with
largely non-reducible and reducible metal oxide cata-
lysts. The latter have evolved primarily out of investi-
gations of methane oxidative coupling [80–82]. In the
oxidative coupling of methane, ethane and ethylene
are major products, and ethylene has been shown to
be derived from ethane [80–82]. The dehydrogena-
tion or oxydehydrogenation of ethane to ethylene,
must occur over these catalysts, and since the overall
selectivity is high, the selectivity to ethylene must
be good. In addition to these two lines of research,
recent work at high temperatures using Pt and Pt-Sn
on a monolith support has been reported.
12. Lower temperature ethane
oxydehydrogenation
Ethane oxydehydrogenation at temperatures in the
range of 300–400 ◦ C is conducted with reducible
metal oxide catalysts usually containing vanadium.
Propane
Oxydehadrogenation Dehydrogenation Oxydehydrogenation
100 <1 100
100 1 100
25 100
100 87 100
 M.M. Bhasin et al. / Applied Catalysis A: General 221 (2001) 397–419
Early work in this area was conducted by Thorstein-
son et al. using a vanadium molybdenum niobium
oxide catalyst [30,31], and a number of papers and
patents based on that work have been issued by Union
Carbide Corp. and other laboratories [32–39]. Ethane
is thought to react with molybdenum or vanadium
in the catalyst to form surface ethoxide, which can
then undergo a beta-elimination process to form ethy-
lene. The surface ethoxide can be oxidized further
to make surface acetate, which leads to acetic acid
on hydrolysis. Acetic acid in varying amounts is co-
produced with ethylene at pressures greater than one
atmosphere. Later work has shown that catalysts can
be made selective for either ethylene or acetic acid
by modifications in the elemental makeup of the cat-
alyst and suitable adjustments of reaction conditions
[38,39]. Selectivity to ethylene with these catalysts is
in the range of approximately 70% at approximately
70% conversion of ethane.
Ethane oxydehydrogenation at temperatures in
the range of ≥500 ◦ C may be conducted with phos-
phorous/molybdenum/antimony oxide catalysts [40]
or iron-containing solid solution catalysts stabilized
with one or more metal oxides [41]. Phosphorous/
molybdenum/antimony oxide catalysts, suitably modi-
fied with other elements give a selectivity to ethylene
in the range of 78% at 20% conversion of ethane. Iron-
containing solid solution catalysts stabilized with
metal oxide are optimally used with a flow of hydro-
gen chloride and water in addition to ethane and wa-
ter. At a contact time of 12 s at 550 ◦ C over an iron in
␣-alumina solid solution catalyst stabilized with lant-
hanum oxide, ethane conversion is in the range of 90%.
Products are ethylene (80–93% selectivity) and vinyl
chloride (1–4% selectivity). As a route to ethylene-
only, this route seems unattractive because of the pro-
duction of vinyl chloride, which must be separated
from ethylene. For vinyl chloride producers, however,
where the stream may be able to be used without
separation, such a high selectivity/conversion process
to ethylene may possibly be economically attractive.
13. Higher temperature ethane
oxydehydrogenation
Ethane oxydehydrogenation at temperatures in the
range of 650–800 ◦ C is conducted with a variety
413
of materials, most not containing the elements
molybdenum and vanadium, which are common
in lower-temperature catalysts. As opposed to the
lower-temperature catalysts which tend to produce
coproducts, the higher temperature processes tend to
produce only ethylene as C2 product, though some do
produce methane from cracking reactions. Oxygenated
products other than carbon oxides largely are absent,
likely because of their instability at high temperature,
especially in contact with the catalyst and oxygen.
Li and coworkers at the Lanzhou Institute
of Chemical Physics, China, have shown that
a Na2 WO4 -Mn/SiO2 catalyst (a high-selectivity
methane coupling catalyst) at 700 ◦ C is capable of
giving greater than 70% selectivity to ethylene (and
10% selectivity to methane) at >70% conversion of
ethane [42]. Wang and coworkers at the National In-
stitute of Materials and Chemical Research, Japan,
found that lithium chloride on sulfated zirconia at
650 ◦ C yielded 70% selectivity to ethylene (and 2%
to methane) at 98% conversion of ethane [43,44]. Al-
though the catalyst did show some deactivation with
time, it can still maintain an ethylene yield as high
as 50% after 24 h. Lin et al. at the Tokyo Institute
of Technology, Japan, and Chonnam National Uni-
versity, Korea, found that a SrBi3 O4 Cl3 catalyst at
640 ◦ C is capable of giving 90% selectivity to ethy-
lene at a 25% conversion of ethane [45]. Co-feeding
HCl did not change the conversion or the selectivity,
but did slow the activity decrease along with the loss
of chlorine observed in its absence.
Dang et al. at the Lanzhou Institute of Chemical
Physics, China, have examined a vanadium oxide
on barium carbonate catalyst at 650 ◦ C that gave
good activity [46]. At 34% conversion of ethane,
they obtained 76% selectivity for ethylene. Au and
coworkers at Hong Kong Baptist University examined
SrFeO3−δ Clσ and Ho2 O3 with BaCl2 [47,48]. Of the
two, SrFeO3−δ Clσ gave the higher yield of ethylene.
At 680 ◦ C, SrFeO3−δ Clσ yielded 90% ethane con-
version and 70% selectivity. Over 40 h, there was no
drop-off in yield or in chlorine content of the cata-
lyst. Longer-duration tests were not conducted. The
Ho2 O3 with BaCl2 catalyst at 640 ◦ C yielded 57%
conversion and 68% selectivity to ethylene.
Choudhary et al. at the National Chemical Labo-
ratory, India, have examined strontium and other rare
earth oxides deposited on sintered low surface area
414 M.M. Bhasin et al. / Applied Catalysis A: General 221 (2001) 397–419
supports precoated with lanthanum and other rare earth
oxides [49,50]. A Sr-Nd2 O3 catalyst showed the high-
est activity and selectivity for ethane. At 800 ◦ C, the
Sr-Nd2 O3 catalyst was capable of giving 60% con-
version of ethane and greater than 80% selectivity to
ethylene.
Workers at Phillips Petroleum Company have ex-
amined catalysts comprising lithium, titanium, and
manganese, and catalysts comprising cobalt, phos-
phorous, and at least one promoter selected from a list
of elements [51,52]. At 650 ◦ C, the lithium/titanium
catalyst with manganese gave an ethane conversion of
47% and a selectivity to ethylene of 75%. At 670 ◦ C,
a catalyst containing cobalt, phosphorous potassium,
and zirconium where the cobalt was introduced as
cobalt sulfide, and calcining of the catalyst was ac-
complished in the absence of oxygen at 670 ◦ C, gave
an 85% conversion of ethane and a 86% selectivity to
ethylene (25 vol.% ethane/75 vol.% air at 3 psig and
2400 GHSV). Addition of a halogen-containing com-
pound, such as methyl chloride, increases ethylene
yield. There is an indication that the catalyst loses
activity with use over time for oxydehydrogenation of
hydrocarbons. If catalyst activity and selectivity can
be maintained over long periods of time, this catalyst
seems like a good potential candidate for use in an
economical process for ethane oxydehydrogenation
to ethylene.
14. Ethane oxydehydrogenation over
monolith catalysts
Ethane oxydehydrogenation over Pt and Pt-Sn
coated monolith catalysts at high temperatures and
extremely short contact times has been the subject of
work by Schmidt and coworkers at the University of
Minnesota for the past several years [53–58]. Work
along similar lines has also been reported to be un-
der way in Russia. Similar results are reported over
platinum gauze and platinum coated pellets [59–61].
Conversions of ethane on the order of 70% and se-
lectivities on the order of 65% are obtained over a
Pt-coated ceramic foam monolith at approximately
1000 ◦ C at contact times on the order of 1 ms. With
the addition of tin to the catalyst and with the addi-
tion of hydrogen to the feed, selectivity to ethylene
could be increased to above 85% at a conversion of
ethane of 67% [56,57]. The authors’ consideration
on mechanism follows: “Qualitatively, this process
must involve first the oxidation of H2 to H2 O, which
generates heat and removes O2 , followed by dehy-
drogenation of C2 H6 to produce C2 H4 and H2 , and
all of these reactions occur within ∼10−3 s. Possi-
ble mechanisms to explain the results are (i) purely
catalytic reactions on the Pt–Sn surface, (ii) purely
homogeneous reaction and (iii) catalytic H2 oxidation
followed by homogeneous ethane decomposition. We
will consider each of these mechanisms and show
that, while each of these gives partial interpretation of
results, none appears to be totally satisfactory” [56].
Hydrogen is produced in the process in greater
amount than fed, so that hydrogen required for the
process can be recycled from downstream equipment.
Such high selectivity to ethylene coupled with low
contact time and high conversion of ethane makes this
process an attractive possibility as an alternative to
steam cracking of hydrocarbons as a commercial route
to ethylene. While the possibility is bright, a number
of questions need to be resolved. Among these are the
questions of long-term catalyst activity, long-term se-
lectivity, and whether such mixtures of ethane, hydro-
gen, and oxygen in the ratios needed, reported to be
2:2:1 at 950 ◦ C, can be handled safely on a commercial
scale. In addition, issues related to process start-ups,
shutdowns, and re-starts have to be addressed.
15. Ethane oxydehydrogenation with
carbon dioxide
Another approach that is being pursued in sev-
eral laboratories is to use CO2 as a mild oxidant for
oxydehydrogenation of hydrocarbons. [42,62–68]. In
addition to ethane, oxidative dehydrogenation of ethyl-
benzene [67] and of propane has also been demon-
strated [65,66,68]. Extensive work has been done on
oxidative dehydrogenation of ethane with CO2 by
Xu et al. [62] of the Dalian Institute of Chemical
Physics. Thus, reaction temperature was significantly
lowered from that used in steam cracking, though it
is still quite high. High selectivity was obtained at
∼800 ◦ C, 1000 h−1 GHSV and 0.1 MPa (∼1 atm) us-
ing several catalysts containing oxides of Cr, Cr–Mn,
Cr–Mn–Ni, Cr–Mn–La onto a Silicalite-2 (Si-2) ze-
olite. Highest conversion/selectivity observed were
 M.M. Bhasin et al. / Applied Catalysis A: General 221 (2001) 397–419
69.8%/90.8% at 800 ◦ C and 800 h−1 and 72.2%/90.4%
at 820 ◦ C/1000 h−1 over a Cr/Si-2 catalyst. These con-
versions approached equilibrium for these two condi-
tions; which were calculated to be 72.8 and 77.6%.
Oxydehydrogenation using CO2 according to the
authors [62] has the following special features.
• CO2 acts as a mild oxidant for the oxidative dehy-
drogenation of C2 H6 to yield C2 H4 .
• A significant lowering of the reaction temperature
as compared to the steam cracking process, resulting
in lower coke formation on the catalyst.
• C2 H6 conversion and C2 H4 selectivity are higher
in this process than in all other processes of C2 H4
production from C2 H6 . Furthermore, no C3 + and
C2 H2 products are formed.
The dehydrogenation of ethane is facilitated by the
reaction of CO2 with H2 (reverse water-gas shift) to
make CO + H2 O and also with ethane and methane as
shown as follows:
C2 H6 = C2 H4 + H2
CO2 + H2 = CO + H2 O
C2 H6 + 2CO2 = 4CO + 3H2
CH4 + CO2 = 2CO + 2H2
In addition, ethane hydrogenolysis to methane also
provides favorable free energy of reaction.
C2 H6 + H2 = 2CH4
The authors suggest that ethane oxydehydrogena-
tion with carbon dioxide takes place according to the
following overall stoichiometry:
16C2 H6 + 9CO2
= 14C2 H4 + 12CO + 6H2 O + 12H2 + CH4
It can be seen from this reaction scheme that in or-
der to increase the selectivity to ethylene, the forma-
tion of methane must be suppressed and that the key
problem is to develop a catalyst that can suppress the
thermodynamically favorable side reactions.
The work of Xu et al. [62] also demonstrated
that the inclusion of steam in the feed leads to a
three-fold reduction in coke formation at 800 ◦ C.
These authors also studied the C2 H6 + CO2 reac-
tion in FCC tail-gas to increase the ethylene content.
415
Thus, using a tail-gas containing 18.8% C2 H6 and
19.2% C2 H4 , the ethylene content was increased to
25.4–27.2% depending on the composition of the
catalyst. This reaction was carried out at 1073 K,
0.1 MPa, 1000 h−1 GHSV and CO2 :C2 H6 mole ratio
of 1:1. The Cr/Si-2 catalyst gave the lowest ethylene
enhancement (25.4%) while Cr-Mn-La/Si-2 catalyst
gave the highest ethylene enhancement to 27.2%.
The corresponding ethane conversions/selectivity are
60.6%/79.6% and 63.6%/85.8%, respectively.
Some of the other early work in this area was
done using highly selective catalysts for the oxidative
coupling of methane by Liu et al. of the Lanzhou
Institute of Chemical Physics [42]. Their work,
employing a selective methane coupling catalyst
(Na2 WO4 -Mn/SiO2 ), showed that >70% conversion
and selectivity could be achieved at 700–750 ◦ C and
space velocities of >30,000 h−1 , employing O2 as the
oxidant. This is in contrast with 53% conversion/97%
selectivity at 800 ◦ C and 69.5% conversion/90.5%
selectivity at 850 ◦ C employing C2 H6 :CO2 = 1:1 and
3600 h−1 space velocity. These catalysts were sta-
ble for 100 h of operation. The authors propose that
surface lattice oxygen is responsible for selective oxy-
dehydrogenation while the bulk lattice oxygen is re-
sponsible for deep (non-selective) oxidation of ethane.
Nakagawa et al. [63] studied a series of metal ox-
ides for ethane dehydrogenation in the presence of
CO2 at 650 ◦ C, C2 H6 :CO2 = 5:25 ml/min and SV =
900 h−1 ml (g-cat)−1 . The order of activity was as
follows:
Ga2 O3 > Cr 2 O3 > V2 O5 > TiO2
> Mn3 O4 > In2 O3 > ZnO > La2 O3 .
The ethylene selectivity was generally >85% in the
presence of CO2 . Ethylene yields for Ga2 O3 in the
presence of CO2 was approximately twice than in its
absence. The ethylene yield enhancement with Cr2 O3
was only a modest 1.2%, while in the case of V2 O5 ,
there was a 2.7% detrimental effect. The authors pro-
pose involvement of acid sites in the dehydrogenation
reaction.
Wang et al. [64] of Japan studied the effect of sul-
fate and Na-treatment silica on the dehydrogenation
of ethane with CO2 at 650 ◦ C and 1 atm. Sulfating
silica (348 m2 /g) with (NH4 )2 SO4 (and calcining
at 700 ◦ C/3 h to give ∼2 wt.% SO4 2− ), and adding
416 M.M. Bhasin et al. / Applied Catalysis A: General 221 (2001) 397–419
5 wt.% Cr gave the best catalytic performance.
A conversion/selectivity/yield of 53.9/87.6/55.2%
was observed at 650 ◦ C/1 atm. With a 10:50:40 =
ethane:CO2 :N2 mixture. The corresponding 5% Cr
catalyst (3.34 m2 /g on silica gave 56.1/92.9/52.1%
conversion/selectivity/yield. However, Na-Cr/SiO2
gave very poor performance of 4.9/98.2/4.8%. Further-
more, the surface area dropped to 3.1 m2 /g. In com-
parison, silica alone of 348 m2 /g gave 2.2/97.2/0.2%
conversion/selectivity/yield. Dehydrogenation is pro-
posed to occur by abstraction of H by oxygen species
which in turn are formed from surface carbonates
decomposition. At low temperatures, decomposition
of carbonates absorbed on strong basic sites restricts
formation and mobility of oxygen resulting in lower
activity of Cr/Na-SiO2 catalyst. However, sulfation
of silica favors reaction with hydrocarbon and hence,
higher activity (along with formation of CO + H2 ).
The Cr/SiO2 and Cr/SO4 2− -SiO2 were shown to be
stable for 5–6 h of operation.
Macho and coworkers [65,66] studied oxy-
dehydrogenation of propane and butane over
Mn-Cr-K/Y-Al2 O3 to give propene and butenes, re-
spectively, according to the following reaction:
C3 H8 + CO2 ⇔ C3 H6 + CO + H2 O
C4 H10 + CO2 ⇔ C4 H8 + CO + H2 O
Splitting of the C–C bond proceeds simultaneously
leading to the formation of both higher and lower hy-
drocarbons as shown in the following equation. These
reactions are preferred at relatively higher tempera-
tures while the relatively lower temperatures prefer
straight dehydrogenation with CO2 . In addition, coke
formation is more prevalent at relatively higher tem-
peratures.
2C3 H6 ⇔ C4 H8 + C2 H4
4C3 H8 + 4CO2 ⇔ 3C4 H8 + 4CO + 4H2 O
C3 H8 + 3CO2 ⇔ C2 H4 + 4CO + 2H2 O
C3 H8 ⇔ C3 H6 + H2
C2 H4 + CO2 ⇔ CH4 + 2CO
The authors propose that:
• at middle temperatures (850–1000 K) dehydrogena-
tion of alkanes proceeds first and hydrogen gener-
ated reacts with CO2 giving CO + H2 O;
• at higher temperatures (>950 K), alkanes may react
directly with oxygen left from CO2 .
The authors also explain the reactivity of catalysts
in terms of acidic and basic nature of supports and
the reactivity of CO2 with such supports. Typically,
good active catalysts appear to approach thermody-
namic equilibrium. However, temperatures required
to achieve higher conversions and yields also lead to
higher levels of coking and by-product formation. No
economic assessment of the CO2 -based dehydrogena-
tion processes is available. The authors are not aware
of any commercialization efforts using this reaction.
16. Propane oxydehydrogenation
Propane oxydehydrogenation combining high
propane conversion and high propylene selectivity has
proved an elusive goal, and propane oxydehydrogena-
tion as a route to propylene appears to be far from
realizing its commercialization potential. One diffi-
culty is that propylene is more easily oxidized than
is propane, so that selectivity tends to decline rapidly
with conversion. Another is that at temperatures above
about 700 ◦ C, propane cracking becomes significant,
and a variety of products other than propylene are
produced. Efforts to increase propylene selectivity
have centered in the areas of more selective catalysts
at temperatures below 700 ◦ C, more selective cata-
lysts and conditions at high temperatures, membrane
reactors, and cyclic-operation reactors.
A variety of catalysts for propane oxydehydrogena-
tion have been examined recently at temperatures be-
low 700 ◦ C [69–75]. Particularly, notable for its high
selectivity to propylene is a vanadia-silica-zirconia
catalyst of Rulkens and Tilley at the University of
California, Berkley, CA [75]. These workers using
a molecular precursor route, produced an 18/36/46
V2 O5 -SiO2 -ZrO2 catalyst which gave 81.5% selec-
tivity to propylene at 8% conversion of propane at
550 ◦ C. The presence of zirconium is important in re-
taining high selectivity, likely by stabilizing the dis-
persion of vanadia. These catalysts also are among the
most active for propane oxydehydrogenation.
Ranzi and coworkers at the Polytechnic University
of Milan [76] and Choudhary et al. at the National
Chemical Laboratory of India [77] have examined
 M.M. Bhasin et al. / Applied Catalysis A: General 221 (2001) 397–419
oxydehydrogenation of propane at high temperatures
with and without catalysts. Selectivities to olefins,
comprising primarily ethylene and propylene, can be
obtained in selectivities of 50–80%. Conversion of
propane is on the order of 90% under these conditions.
Alfonso, Julbe, Farrusseng, Menéndez, and Santa-
marı́a at the University of Zaragoza, Spain, and the
Laboratoire des Matériaux et Procédés Membranaires,
France, examined the oxidative dehydrogenation of
propane over V-Al2 O3 catalytic membranes which al-
low separation of propane and oxygen feeds [78]. Se-
lectivities of 51% to propylene were obtained at 8%
conversion of propane at 550 ◦ C, which is higher than
the 44% selectivity obtained at the same temperature
and conversion with premixed feeds.
Creaser et al. at Lulea University of Technol-
ogy, Sweden, University of Waterloo, Canada, and
Chalmers University of Technology, Sweden, exam-
ined cyclic operation of the oxidative dehydrogena-
tion of propane over a V-Mg-O catalyst [79]. In a
cyclic reactor where oxygen and propane were alter-
nately passed over the catalyst, propylene selectivity
was considerably increased at 510 ◦ C compared to
steady state operation with mixed feeds. At a 1:1
propane:oxygen ratio, selectivity to propylene was
78% at 4.3% conversion of propane with cyclic op-
eration compared to 55% selectivity to propylene at
5.5% conversion of propane with mixed feeds. These
conversions/selectivities are still far from being com-
mercially attractive.
17. Summary
It is interesting to note that though oxydehydro-
genation is thermodynamically favored at even am-
bient conditions, no catalyst has been discovered that
can provide such a true oxidative conversion at tem-
peratures below ∼300 ◦ C. Catalysts discovered by
Union Carbide that operate at 300–400 ◦ C [80–82],
are the best known catalysts for producing ethylene
(with co-production of some acetic acid in varying
amounts). Thus, a true oxydehydrogenation catalyst
is yet to be discovered. Another explanation may
very well be that all of the currently known cata-
lysts proceed through a dehydrogenation step (C–H
activation and bond breakage) followed by oxida-
tion of the evolved H2 or a surface H-containing
417
intermediate. This certainly appears to be the case with
the majority, if not all, of the catalysts that are active
for oxydehydrogenation that operate at 500–800 ◦ C.
However, under these conditions, it is very difficult
to avoid undesirable partial/total oxidation to CO and
CO2 via gas phase and surface reactions. Therefore,
the challenge for oxydehydrogenation catalysis is
to develop highly active and selective catalysts for
totally selective oxidation of alkanes to alkenes.
References
[1] G.F. Hornaday, F.M. Ferrell, G.A. Mills, Manufacture
of mono- and diolefins from paraffins by catalytic
dehydrogenation, in: Advances in Petroleum Chemistry and
Refining, Vol. 4, Interscience, Pans, 1961.
[2] V.N. Ipatieff, R.E. Schaad, Mixed polymerization of butenes
by phosphoric acid catalyst, Ind. Eng. Chem. 30 (1938) 596–
599.
[3] A.L. Waddams, Chemicals from Petroleum, 4th Edition, Gulf
Publishing Company, Houston, 1978, 1980.
[4] A.H. Weiss, The manufacture of propylene (158th Meeting of
the American Chemical Society, 10–12 September 1969), in:
Refining Petroleum for Chemicals, Advances in Chemistry
Series, Vol. 97, American Chemical Society, Washington, DC,
1970.
[5] R.G. Craig, D.C. Spence, Catalytic dehydrogenation of
liquefied petroleum gas by the Houdry Catofin and Catadiene
processes, in: R.A. Meyers (Ed.), Handbook of Petroleum
Refining Processes, McGraw-Hill, New York, 1986 (Chapter
4.1).
[6] D. Sanfilippo, F. Buonomo, G. Fusco, I. Miracca, Paraffins
Activation through Fluidized Bed Dehydrogenation: the
Answer to Light Olefins Demand Increase, Elsevier,
Amsterdam, Stud. Surf. Sci. Catal. 119 (1998) 919–924.
[7] R. Iezzi, A. Bartolini, U.S. Patent 5,633,421 (27 May 1997),
assigned to Snamprogetti.
[8] E.C. Luckenbach, F.A. Zenz, Giovanni Papa, A. Bertolini,
U.S. Patent 5,656,243 (12 August 1997), assigned to
Snamprogetti.
[9] V. Haensel, U.S. Patent 2,602,772 (8 July 1952), and other
patents on catalytic reforming with Pt catalysts assigned to
UOP.
[10] H.S. Bloch, UOP discloses new way to make linear
alkylbenzene, Oil Gas J. 79–81 (1967); U.S. Patent
3,448,165 (3 June 1969), and other patents on the catalytic
dehydrogenation of paraffinic hydrocarbons assigned to
Herman Bloch and UOP.
[11] R.C. Berg, B.V. Vora, Detergent alkylate, in: Encyclopedia
of Chemical Processing and Design, Vol. 15, Marcel Dekker,
New York, 1982, pp. 266–284.
[12] T. Imai, B.V. Vora, J.C. Bricker, Development of
dehydrogenation catalysts and processes, Petroleum and
Petrochemical College, Chulalongkorn University, Bangkok,
December 1989.
418 M.M. Bhasin et al. / Applied Catalysis A: General 221 (2001) 397–419
[13] J.H. Gregor, R. Antonelli, T.D. Foley, E.C. Arnold, Increased
opportunities for propane dehydrogenation, DeWitt World
Petrochemical Review, Houston, TX, 23–25 March 1999.
[14] C.P. Poole, D.S. MacIver, The physical–chemical properties
of chromia-alumina catalysts, in: Advances in Catalysis, Vol.
17, Academic Press, New York, 1967, pp. 223–314.
[15] C.P. Poole, W.L. Kehl, D.S. MacIver, Physical properties of
coprecipitated chromia-alumina catalysts, J. Catal. 1 (1962)
407–415.
[16] E.L. Tucci, J.M. Dufallo, R.J. Feldman, Commercial
performance of the Houdry Catofin process for isobutylene
production for MTBE, in: Proceedings of the Workshop on
Catalysts and Catalytic Processes, Research Institute, King
Fahd University of Petroleum and Minerals, Dhaharan, Saudi
Arabia, 6 November 1991.
[17] S. Gussow, R. Whitehead, Isobutane dehydrogenation by
Catofin as feed for motor fuel ether, in: Proceedings of the
NPRA Annual Meeting, San Antonio, TX, 17–19 March
1991.
[18] R.O. Dunn, F.M. Brinkmeyer, G.F. Schuette, The Phillips
STAR process for the dehydrogenation of C3 , C4 , and C5
paraffins, in: Proceedings of the NPRA Annual Meeting, New
Orleans, LA, 22–24 March 1992.
[19] F. Buonomo, G. Fusco, I. Miracca, G. Papa, D. Sanfilippo,
Fluid bed dehydrogenation of light paraffins: the FBD
technology, in: Proceedings of Petrotech ’96, Bahrain, 10–12
June 1996.
[20] R.J. Lawson, T.R. Fritsch, Developing a clean, efficient
process, Asia-Pacific Chem. 99 (1999) 8–9.
[21] A.Q. Campbell, J.M. Garcés, T.M. May, R.F. Pogue,
Preparation of ethylbenzene and substituted derivatives by
alkylation using unpurified reaction products of ethylene
prepared by dehydrogenation of ethane, World Patent WO
96/34843 (7 November 1996), assigned to Dow Chemical Co.
[22] S. Benson, Conversion of methane using chlorine, U.S. Patent
4,199,533 (22 April 1996), assigned to the University of
Southern California.
[23] A.H. Weiss, Which propylene process is best? Hydrocarbon
Process. 47 (1968) 123–127.
[24] T. Laegreid, M. Rönnekleiv, A. Sobbaken, in: Proceedings of
the DGMK Conference, 11–12 November 1993, p. 147.
[25] R. Lodeng, P. Soaker, Reactor for catalytic dehydrogenation
of hydrocarbons with selective oxidation of hydrogen, U.S.
Patent 5,997,826 (7 December 1999), assigned to Den Norske
Stats Oljeselskap A.S. (Statoil).
[26] J.G. Tsikoyiannis, D.L. Stern, R.K. Grasselli, J. Catal. 184
(1999) 77.
[27] P.A. Agaskar, R.K. Grasselli, J.N. Michaels, P.T. Reischman,
D.L. Stern, J.G. Tsikoyiannis, Process for the catalytic
dehydrogenation of alkanes to alkenes with simultaneous
combustion of hydrogen, U.S. Patent 5,430,209 (4 July 1995),
assigned to Mobil Oil Corp.
[28] J.F. Brazdil, A.P. Cavalcanti, J.P. Padolewski, Method for
preparing vanadium antimony oxide-based oxidation and
ammoxidation catalysts, U.S. Patent 5,693,587 (2 December
1997), assigned to The Standard Oil Company of Ohio
(Sohio/BP).
[29] Ullmann’s Encyclopedia of Industrial Chemistry, 5th Edition,
Vol. A10, VCH Verlagsgesellschaft, GmbH, Germany, 1987,
p. 54.
[30] E.M. Thorsteinson, T.P. Wilson, F.G. Young, P.H. Kasai, J.
Catal. 52 (1978) 116.
[31] E.M. Thorsteinson, T.P. Wilson, F.G. Young, P.H.U.S. Kasai,
U.S. Patent 4,250,346 (1981), assigned to Union Carbide.
[32] J.H. McCain, U.S. Patent 4,524,236 (18 June 1985), assigned
to Union Carbide.
[33] R.M. Manyik, J.L. Brockwell, J.E. Kendall, U.S. Patent
4,899,003 (6 February 1990), assigned to Union Carbide.
[34] K. Ruth, R. Kieffer, R. Burch, J. Catal. 175 (1998) 16.
[35] K. Ruth, R. Burch, R. Kieffer, J. Catal. 175 (1998) 27.
[36] M. Kitson, U.S. Patent 5,260,250 (9 November 1993),
assigned to BP Chemicals.
[37] W. Ueda, N.F. Chen, K. Oshihara, Kinet. Catal. 40 (1999)
401 (English).
[38] W. Ueda, N.F. Chen, K. Oshihara, Chem. Commun. 517
(1999).
[39] C. Hailett, European Patent Application, Publication
0,480,594,A2 (15 April 1992), assigned to B.P. Chemicals.
[40] M. Kutyrev, F. Cavani, F. Trifiro, European Patent
Application, Publication 0,544,372,A1 (2 June 1993),
assigned to Eniricherche S.p.A.
[41] J. Magistro, U.S. Patent 5,087,791 (11 February 1992),
assigned to B.F. Goodrich.
[42] Y. Liu, J. Xue, X. Liu, R. Hou, S. Li, in: A. Parmaliana, et
al. (Eds.), Natural Gas Conversion V, Elsevier, Amsterdam,
Stud. Surf. Sci. Catal. 119 (1998) 593,
[43] S. Wang, K. Murata, T. Hayakawa, S. Hamakawa, K. Suzuki,
Catal. Lett. 59 (1999) 173.
[44] S. Wang, K. Murata, T. Hayakawa, S. Hamakawa, K. Suzuki,
Chem. Commun. 103 (1999).
[45] S.W. Lin, Y.C. Kim, W. Ueda, Bull. Chem. Soc. Jpn. 71
(1998) 1089.
[46] Z. Dang, J. Gu, J. Lin, D. Yang, Catal. Lett. 54 (1998) 129.
[47] H.X. Dai, C.F. Ng, C.T. Au, Catal. Lett. 57 (1999) 115.
[48] C.T. Au, K.D. Chen, H.X. Dai, Y.W. Liu, C.F. Ng, Appl.
Catal. A 177 (1999) 185.
[49] V.R. Choudhary, S.A.R. Mulla, V.H. Rane, J. Chem. Technol.
Biotechnol. 71 (1998) 167.
[50] V.R. Choudhary, B.S. Uphade, S.A.R. Mulla, U.S. Patent
5,712,217 (27 January 1998), assigned to Council of Scientific
and Industrial Research, India.
[51] A.D. Eastman, J.B. Kimble, U.S. Patent 4,450,313 (22 May
1984), assigned to Phillips Petroleum.
[52] A.D. Eastman, J.P. Guillory, C.F. Cook, J.B. Kimble, U.S.
Patent 4,497,971 (5 February 1985), assigned to Phillips
Petroleum.
[53] L.D. Schmidt, M. Huff, World Patent 96/13475 (9 May 1996),
assigned to Regents of the University of Minnesota.
[54] S.S. Bharadwaj, L.D. Schmidt, World Patent 96/33149 (24
October 1996), assigned to Regents of the University of
Minnesota.
[55] C. Yokoyama, S.S. Bhardadwaj, L.D. Schmidt, World Patent
97/26987 (31 July 1997), assigned to Regents of the
University of Minnesota.
 M.M. Bhasin et al. / Applied Catalysis A: General 221 (2001) 397–419
[56] A.S. Bodke, D.A. Olschki, L.D. Schmidt, E. Ranzi, Science
285 (1999) 712.
[57] M. Huff, L.D. Schmidt, in: M.M. Bhasin, D.W. Slocum (Eds.),
Methane and Alkane Conversion Chemistry, Vol. 227, Plenum
Press, New York, 1995.
[58] A.S. Bodke, L.D. Schmidt, Catal. Lett. 63 (1999) 113.
[59] S.S. Bharadwaj, L.D. Schmidt, J. Catal. 155 (1995) 403.
[60] D.W. Flick, M.C. Huff, J. Catal. 178 (1998) 315.
[61] D.I. Iordanoglou, A.S. Bodke, L.D. Schmidt, J. Catal. 187
(1999) 400.
[62] L. Xu, L. Lin, Q. Wang, L. Yan, D. Wang, W. Liu, in: A.
Parmaliana, et al. (Eds.), Natural Gas Conversion V, Elsevier,
Amsterdam, Stud. Surf. Sci. Catal. 119 (1998) 605.
[63] K. Nakagawa, M. Okamura, N. Ikenaga, T. Suzuki, S. Wang,
Chem. Commun. 1025 (1998).
[64] S. Wang, K. Murata, T. Hayakawa, S. Hamakawa, K. Suzuki,
Chem. Lett. 569 (1999).
[65] M. Kralik, V. Macho, Petrol. Coal 40 (1998) 115.
[66] V. Macho, M. Kralik, E. Jurecekova, L. Jurecek, Petrol. Coal
39 (1997) 46.
[67] M. Sugino, H. Shimada, T. Turuda, H. Miura, N. Ikenaga, T.
Suzuki, Appl. Catal. A 121 (1995) 125.
[68] Takahara, M. Saito, Chem. Lett. 973 (1966).
[69] H. Jachow, World Patent 99/42404 (26 August 1999), assigned
to BASF.
[70] G. Descat, World Patent 98/24742 (11 June 1998), assigned
to ELF Atochem.
419
[71] S. Wang, K. Murata, T. Hayakawa, S. Hamakawa, K. Suzuki,
Chem. Lett. 25 (1999).
[72] P. Viparelli, P. Ciambelli, L. Lisi, G. Rupooplo, G. Russo,
J.C. Volta, Appl. Catal. A 184 (1999) 291.
[73] D. Creaser, B. Andersson, R.R. Hudgins, P.L. Silverston,
Chem. Eng. Sci. 54 (1999) 4365.
[74] Z.M. Fang, Q. Hong, Z.H. Zhou, S.J. Dai, W.Z. Weng, H.L.
Wan, Catal. Lett. 61 (1999) 39.
[75] R. Rulkens, T.D. Tilley, J. Am. Chem. Soc. 120 (1998)
9959.
[76] A. Beretta, P. Forzatti, E. Ranzi, J. Catal. 184 (1999) 469.
[77] V.R. Choudhary, V.H. Rane, A.M. Rajput, AIChE J. 44 (1998)
2293.
[78] M.J. Alfonso, A. Julbe, D. Farrusseng, M. Menéndez, J.
Santamarı́a, Chem. Eng. Sci. 54 (1999) 1265.
[79] D. Creaser, B. Andersson, R.R. Hudgins, P.L. Silverston,
Chem. Eng. Sci. 54 (1999) 4437.
[80] G.E. Keller, M.M. Bhasin, Synthesis of ethylene via oxidative
coupling of methane. 1. Determination of active catalysts, J.
Catal. 73 (1999) 9.
[81] M.M. Bhasin, Feasibility of ethylene synthesis via catalytic
oxidative coupling of methane, in: paper presented at the
Methane Conversion Symposium, Auckland, New Zealand,
27 April to 1 May 1987.
[82] M.M. Bhasin, Feasibility of Ethylene Synthesis via Oxidative
Coupling of Methane, Elsevier, Amsterdam, Stud. Surf. Sci.
Catal. 36 (1988) 343.