High-temperature and short-contact-time oxidative dehydrogenation of ethane in the presence of Pt/Al2O3 and BaMnAl11O19 catalysts

Article abstract of DOI:10.1006/jcat.2001.3192

The oxidative dehydrogenation of ethane over a Pt/_γ-Al₂O₃ catalyst in adiabatic conditions (that is at T = 800-1000°C and few milliseconds contact time) resulted in the production of ethylene with 50% yield. In order to better elucidate the single roles that gas-phase reactions and heterogeneous phase reactions might play in the high-temperature activation of ethane, a detailed experimental and theoretical investigation was addressed. The results suggested that the performance of the adiabatic reactor relied on a cooperation between catalytic reactions (responsible for deep and partial oxidation of ethane to CO_x, H₂O, and H₂) and homogeneous reactions (responsible for the formation of olefins). In autothermal conditions, thus, the main function of the Pt-catalyst appeared to be that of accelerating ignition of the radical process, through the initial combustion of ethane which can occur at temperatures as low as 200°C. It was demonstrated that the same function can be equally well accomplished by a Pt-free oxidation catalyst like the BaMnAl₁₁O₁₉ which is active in the deep oxidation of ethane. Even in the presence of the hexa-aluminate material, which has no activity in the selective oxidation of ethane to ethylene, yields to ethylene higher than 50% were observed in the autothermal reactor by preheating the ethane/air feed stream at a proper temperature.

Full text of DOI:10.1006/jcat.2001.3192

Journal of Catalysis 200, 45–58 (2001)
doi:10.1006/jcat.2001.3192, available online at http://www.idealibrary.com on
High-Temperature and Short-Contact-Time Oxidative Dehydrogenation
of Ethane in the Presence of Pt/Al O and BaMnAl O Catalysts
2
3
11 19
Alessandra Beretta and Pio Forzatti
Dipartimento di Chimica Industriale e Ingegneria Chimica “G. Natta,” Politecnico di Milano, piazza Leonardo da Vinci 32, 20133 Milano, Italy
Received July 17, 2000; revised January 18, 2001; accepted February 2, 2001; published online April 18, 2001
consequence, the one that might offer a potential for the
The oxidative dehydrogenation of ethane over a Pt/� -Al2O3 cata- industrial application. While the published results are stim-
�
lyst in adiabatic conditions (that is at T = 800� 1000 C and few ulating the discussion on the feasibility of novel processes
milliseconds contact time) resulted in the production of ethylene
with 50% yield. In order to better elucidate the single roles that gas-
phase reactions and heterogeneous phase reactions might play in
the high-temperature activation of ethane, a detailed experimental
and theoretical investigation was addressed. The results suggested
temperature and short-contact-time ODH of ethane and
that the performance of the adiabatic reactor relied on a coopera-
tion between catalytic reactions (responsible for deep and partial
oxidation of ethane to COx, H2O, and H2) and homogeneous re-
actions (responsible for the formation of olefins). In autothermal
for olefin productions (6, 7), a debate is still open on the
roles that homogeneous and heterogeneous reactions have
in the short-contact-time catalytic processes.
A few catalysts have been proposed so far for the high-
different interpretations on the mechanism of formation of
ethylene were given.
Schmidt and coworkers studied several noble-metal-
conditions, thus, the main function of the Pt-catalyst appeared to coated ceramic foam monoliths, including bimetallic cata-
be that of accelerating ignition of the radical process, through the lysts (8, 9). Up to 65% ethylene selectivity at 70% ethane
initial combustion of ethane which can occur at temperatures as conversion and practically complete O conversion were
2
�
low as 200 C. It was demonstrated that the same function can be
observed over a Pt-coated �-Al2O3 foam monolith (2).
These results were confirmed also by Flick and Huff (10). In
spite of the extremely high reaction temperatures, a hetero-
equally well accomplished by a Pt-free oxidation catalyst like the
BaMnAl11O19 which is active in the deep oxidation of ethane. Even
in the presence of the hexa-aluminate material, which has no activ-
geneous reaction mechanism was proposed to explain the
ity in the selective oxidation of ethane to ethylene, yields to ethylene
formation of olefins on the Pt-surface; the mechanism is
higher than 50% were observed in the autothermal reactor by pre-
based on �-hydrogen or �-alkyl elimination from adsorbed
heating the ethane/air feed stream at a proper temperature.
�c 2001
alkyl species (5, 10, 11). Only in the case of higher alka-
nes, a nonnegligible contribution of homogeneous steam-
Academic Press
Key Words: ethane ODH; high temperature and short contact
time; Pt/� -Al2O3 catalyst; BaMnAl11O19 catalyst; homogeneous vs cracking reactions was supposed (5, 12).
heterogeneous reactions.
Flick and Huff(13) have recentlyproposed Cr-containing
catalysts for the high-temperature and short-contact-time
selective oxidation of ethane to ethylene; they found that a
1
0% Cr2O3–ZrO2 catalyst outperformed the Pt/Al2O3 sys-
INTRODUCTION
tems. Also in this case a heterogeneous nature of ethylene
High-temperature and short-contact-time oxidative de- formation was assumed and the observed high reactivity
hydrogenation (ODH) of light alkanes has been proposed of the Cr-containing catalysts was associated to a peculiar
as an efficient route for the selective production of short- dehydrogenation activity of the Cr-oxide.
chain olefins under autothermal conditions (1). Schmidt
Bodke et al. (14) have recently shown that the selective
and coworkers have first shown that total yields to olefins conversion of ethane to ethylene at high temperature and
as high as 40% could be obtained via selective oxidation of short contact time can be greatly promoted by adding H to
2
C2–C6 paraffins by running the reaction at few milliseconds the ethane/oxygen feed stream in the presence of a Pt–Sn
contact times over Pt-coated foam monoliths; the reactor catalyst. Ethylene could be produced at 90–95% selectivity
�
operated adiabatically at temperatures of 800–1100 C, de- with over 60% ethane conversion. Apparently, the whole
pending on feed composition (2–5).
process was split in two stages: the exothermic oxidation of
The bulk of data reported in the literature on high- H and the endothermic dehydrogenation of ethane. The
2
temperature ODH of light paraffins show that, among bimetallic catalyst selectively oxidized hydrogen, while pre-
1
2
the series of (Cn + O2) reactions, the oxidative dehydro- sumably homogeneous reactions were responsible for the
genation of ethane is the most selective process and, as a dehydrogenation of ethane to ethylene.
45
0021-9517/01 $35.00
Copyright �c 2001 by Academic Press
All rights of reproduction in any form reserved.

4
6
BERETTA AND FORZATTI
Pt-Rh gauzes have been applied by Goetsch and Schmidt for rapid ignition) and homogeneous reactions (responsi-
(
15) and by Holmen and coworkers (16, 17) to the se- ble for the partial oxidation of propane to propylene and
lective oxidation of alkanes to olefins. They showed that cracking to ethylene and methane).
large amounts of ethylene and ethylene/propylene mixtures
In this work, the same approach was extended to the
could be produced from ethane/air and propane/air feeds, study of the single roles of homogeneous and heteroge-
respectively, in the presence of one single gauze at contact neous reactions in the activation of ethane/air mixtures. The
times in the order of microseconds. However, Holmen and nonadiabatic annular reactor was used to isolate catalytic
coworkers (16, 17) proposed that the production of olefins reactions and homogeneous reactions in the presence of a
occurred mainly in the gas-phase, since even higher olefin Pt/� -Al2O3 catalyst. New pieces of evidence were searched
yieldscould be obtained in the emptyreactor when an exter- for supporting the correct interpretation of the perfor-
nal heat source was applied; it was also proposed that the mance of a short contact time adiabatic reactor, wherein
main role of the gauze was primarily represented by the both presumably occur. Also, in order to verify the effective
rapid ignition of the homogeneous process through initial role of Pt as active component, a Pt-free hexa-aluminate
oxidation reactions.
catalyst was herein studied for comparison.
Recently, theoretical approaches have also been devel-
oped to investigate the contributions of homogeneous and
heterogeneous reactions in the short contact time partial
oxidation of ethane. These are based on the combination
of detailed kinetic schemes for the gas-phase and the solid
phase, respectively. Huff et al. (18) concluded that the role
of the catalyst was to initiate by heat release the gas-phase
process, while Zerkle et al. (19) proposed that catalytic ox-
idation reactions drove both homogeneous and heteroge-
neous cracking of ethane to ethylene.
EXPERIMENTAL
Catalysts. A commercial Pt/� -Al2O3 catalyst (Engel-
hard ESCAT 24) was studied in the present work. It con-
sists of eggshell Pt-coated � -Al2O3 particles with a BET
2
surface area of nearly 95 m /g and a pore volume of
3
8
4 mm /g (estimated on the basis of Hg-intrusion porosime-
try). The powders were ground and sieved with a 325 mesh
sieve down to an average particle size of about 15 �m. Pre-
vious studies suggest that the nature of support affects the
performance of the Pt-catalyst; Flick and Huff (10) found
that � -Al2O3 pellets were less selective to ethylene than
The authors have previously tested a Pt/� -Al2O3 catalyst
in the oxidative dehydrogenation of propane and ethane
in adiabatic conditions. At few milliseconds contact time
�
and temperatures of 800–1000 C, the partial oxidation of
�
-Al2O3 monoliths. In this work, the choice of Pt/� -Al2O3
propane produced 60% total selectivity to olefins (propy-
lene and ethylene) with alkane conversion higher than 50%
catalyst depends on the continuity with previous work, as
this is the same catalyst already studied by the authors in
the oxidative dehydrogenation of propane.
(
20). The oxidative dehydrogenation of ethane resulted in
the selective production of ethylene with yields as high as
0% (21). The results of the adiabatic tests were compared
In order to verify the possible application of a Pt-free ox-
idation catalyst to the ignition of homogeneous reactions, a
Ba-MnAl11O19 catalyst was also tested. The hexa-aluminate
catalyst has a proved activity for total oxidation of methane
and bio-gas and is characterized by a good thermal stabil-
ity; it was developed in our laboratory for application to gas
turbine catalytic combustors. The preparation and charac-
terization of the catalyst has been reported in detail else-
where (23). The catalyst powders were ground and sieved
with 325 mesh sieve.
5
with theoretical evaluations on the expected role of gas-
phase oxidative pyrolysis of the hydrocarbon; in the cases
of both ethane and propane ODH, the measured product
distributions were found very close to those characteristic
of a purely homogeneous process. This suggested that in
the adiabatic reactor gas-phase reactions might control the
conversion of ethane and propane.
In order to better clarify the single roles of homogeneous
and heterogeneous reactions, the oxidative dehydrogena-
tion of propane was studied by means of an annular re-
Catalyst deposition. The catalyst fine particles were de-
actor wherein very small amounts of catalyst were loaded posited via aqueous slurry onto mullite tubes, thus forming
and tested at high flow rates; temperature was externally thin (about 50 �m) and short (1–5 cm) catalyst layers in
controlled and varied from low to high values (22). Exper- the central portion of the ceramic supports. Prior to de-
iments with increasing amounts of the Pt/� -Al2O3 catalyst position, the mullite tubes were treated in acid bath to in-
proved that the catalyst was uniquely active in the produc- crease the surface roughness and enhance the catalyst ad-
tion of COx, H2O, and H2. Conversely, by testing the re- hesion. Few milligrams of catalysts are deposited via slurry
activity of the propane/air mixture in the empty reactor, it (10–100 mg) and this allows the realization of experiments
was shown that homogeneous oxidative pyrolysis occurred at extremely high gas hourly space velocity. The same cata-
�
above 600 C and resulted in the selective production of lyst slurry was used to deposit the catalyst onto the FeCrAl-
propylene and ethylene. The overall performance of the loy tissue (20); the fibrous metallic support was previously
�
adiabatic reactor for propane ODH could be thus explained calcined at 1000 C to promote the formation of a superficial
as a combination of heterogeneous reactions (responsible layer of inert �-Al2O3.

ODH OF ETHANE AT HIGH TEMPERATURE AND SHORT CONTACT TIME
47
Reactors. Two kinds of reactors were used to realize ac-
tivity tests at short contact times (20–22): the annular reac-
tor, which is a nonadiabatic reactor wherein heterogeneous
and homogeneous reactions could be studied separately,
and the adiabatic packed-bed reactor, which is an insulated
reactor wherein homogeneous and heterogeneous phases
are instead intimately interconnected.
The annular reactor consists of a quartz tube (ID = 0.7
or 0.9 cm), wherein the catalyst-coated mullite tube is in-
serted. The gas flows in between the tubes and contacts
longitudinally the catalyst layer. This reactor configuration
is especially suitable to study high-temperature and short-
contact-time processes as extremely high space velocities
6
(
>10 L/h/kgcat) can be easily realized with no pressure
drop. The catalyst temperature is well characterized, as it
can be measured from inside the coated mullite tube by
a sliding thermocouple. The reactor is heated by an ex-
ternal furnace; thus experiments can be run at any reac-
tion temperature; the authors were especially interested in
performing ethane ODH tests in the temperature range
FIG. 1. Homogeneousoxidative dehydrogenation ofethane. Effect of
residence time and reactor temperature on ethane and oxygen conversion.
Feed composition, C2H6/O2/N2 = 11/1/4.
�
of 200–700 C (that is below the usual operating window
of authothermal reactors) in order to isolate the contribu-
tion of the single heterogeneous phase. Still, given the high
volume/surface ratio of this single-channel reactor, tests in
the absence of catalyst were performed for investigating the 1/1 = ethane/oxygen ratio at varying flow rate and oven
role of homogeneous reactions. temperature. Data have been plotted as functions of the
Adiabatic high-temperature tests were instead per- reactor residence time (the ratio between the empty re-
formed in an insulated reactor, wherein catalyst-coated actor volume and the volumetric flow rate, evaluated at
FeCrAlloy elements were packed in a quartz tube (0.7 cm the reactor temperature). At the intermediate tempera-
�
i.d.) in between two guard beds of ceramic particles. The ture of 690 C, appreciable conversions of reactants were
external furnace was used to preheat the feed stream and observed at 0.18 s and they increased rapidly with in-
the inlet portion of the reactor. The outlet portion of the creasing residence time. At slightly higher temperature
�
reactor, instead, was cooled by an external Cu coil for water (oven T = 725 C) reaction was so fast that conversion of
circulation. The axial temperature profile was measured by ethane and oxygen kept over 90% over the whole range
0
0
1
6
�
a thin thermocouple sliding inside a 1 capillary located of residence times. At the reaction temperature of 670 C,
along the reactor axis.
instead, conversion of ethane and oxygen moderately in-
Analysis of the product mixture was realized by combin- creased with time and amounted to 22 and 11% , respec-
ing the results of three separate gas-chromatographic mea- tively, at the contact time of 0.85 s. In all the experiments
surements, obtained respectively from a Molecular Sieve ethylene was produced in large amounts. For instance, at
�
column using He as carrier gas (separating O2, N2, CO and the intermediate temperature of 690 C, the distribution of
CH4), a Popapak Q column (separating CO2, H2O, and hy- the C-containing products (as reported in Fig. 2a) was char-
drocarbons), and a Molecular Sieve column using Ar as acterized by extremely high selectivity to ethylene; this de-
carrier gas (for separation of H2 and N2). All the columns creased from 95% to nearly 60% at increasing contact time.
were connected to thermoconductivity detectors.
The nonzero selectivity of ethylene at zero ethane conver-
sion clearly shows that ethylene was a primary product of
ethane decomposition. CO was produced in much smaller
amount and its selectivity increased with increasing resi-
dence time in line with the behavior of a secondary product.
Methane and CO2 were also produced and were favored
by contact time; they were only minor products as their se-
lectivity maintained below 7 and 4% , respectively. Among
the H-containing species (Fig. 2b), water was formed in
large quantity while the selectivity of H2 was below 10% .
ROLE OF HOMOGENEOUS REACTIONS IN ETHANE ODH:
TESTS IN THE ANNULAR REACTOR
The gas-phase oxidative dehydrogenation of ethane was
studied in the annular reactor in the absence of any catalyst
by varying reactor temperature, feed flow rate, and feed
composition.
Effects of residence time and temperature. Figure 1 re- Concerning the effect of temperature, the process became
ports the results observed with an ethane/air mixture with increasingly selective towards ethylene with decreasing T,

4
8
BERETTA AND FORZATTI
FIG. 3. Homogeneousoxidative dehydrogenation ofethane. Effect of
residence time and temperature on ethylene selectivity. Feed composition,
C2H6/O2/N2 = 1/1/4.
FIG. 2. Homogeneous oxidative dehydrogenation of ethane. Effect
�
of residence time on product distribution at 690 C reactor temperature.
Feed composition, C2H6/O2/N2 = 1/1/4.
while the selectivity of CO became larger and larger in the
opposite direction. Figure 3 compares the C-selectivity of
ethylene at three levels of temperature investigated; it was
�
practically the only reaction product (with water) at 670 C,
while its selectivity amounted to nearly 40% (an almost
�
equal selectivity was observed for CO) at 725 C.
�
Effect of feed composition. At 725 C and 0.4 s con-
tact time, the influence of feed composition was investi-
gated by varying the ethane/oxygen ratio between 1 and
2
.6. Figure 4a shows that ethane conversion decreased pro-
gressively at increasing feed ratio. However, the converted
molar flow of ethane had a net increase. The process be-
came progressively more selective toward the production
of ethylene, since the olefin C-selectivity increased from
4
0 to 65% (Fig. 4b); conversely, CO selectivity was a de-
creasing function of the feed ratio and amounted to less
than 20% at ethane/oxygen = 2.6. Methane selectivity was
almost constant around 10% , while CO2 was detected in
small amount. With increasing ethane/oxygen feed ratio,
and presumably with increasing ethylene concentration,
FIG. 4. Homogeneous oxidative dehydrogenation of ethane. Effect
of ethane/oxygen feed ratio on ethane and oxygen conversion and on
�
the distribution of C-containing products. Feed, ethane/air; T = 725 C;
C2+ species were increasingly formed; their contribution residence time, 0.4 s.

ODH OF ETHANE AT HIGH TEMPERATURE AND SHORT CONTACT TIME
49
ing conditions of the short contact time adiabatic reactors).
The gas-phase process is extremely selective toward the
formation of ethylene since the olefin is the primary prod-
uct of ethyl radical decomposition, while COx are terminal
products of the radical process. For a given value of ethane
conversion, ethylene selectivity is then a slightly increasing
function of ethane/oxygen feed ratio (21).
ROLE OF THE Pt/� -Al2O3 CATALYST: TESTS
IN THE ANNULAR REACTOR
Once the important role of gas-phase reactions was ver-
ified, attention was then focused on the effective role of
the heterogeneous reactions. These were studied through
a perturbative analysis, wherein operating conditions were
chosen in order to minimize the extent of gas-phase ethane
decomposition (high flow rate and low reactant concentra-
tions) and the effect of the addition of increasing amounts
of catalysts on the product distribution was analyzed. Ex-
periments were performed at varying ethane/oxygen feed
ratio. The results reported in Fig. 6 were obtained in the
annular reactor with a total feed stream of 1 L/min (eval-
uated under Normal conditions, T = 273 K and P = 1 atm)
and molar feed composition ethane : oxygen : nitrogen was
FIG. 5. Measured difference between reactor axial temperature pro-
file and oven wall temperature profile at varying feed composition. Con-
ditions as in Fig. 4.
has been summed up in Fig. 4b and included the production
of propylene, propane, and butenes. With increasing ethane
concentration in the feed stream, the reaction was less and
less exothermic. Figure 5 plots for this set of experiments
the measured difference between the reactor axial temper-
ature profile and the oven wall temperature profile. At the
entrance of the reactor a hot spot was measured wherein
most ofthe reaction presumablyoccurred;the hot-spot tem-
perature reduced with increasing ethane/oxygen ratio.
2
:2 :96. Three cases were compared: (1) absence of cata-
lyst on the mullite tube (that is the residual contribution of
homogeneous reactions in the empty volume of the reac-
tor, as a reference), (2) use of a mullite tube coated with a
1
.5-cm-long catalyst layer (catalyst load � 10 mg), (3) use of
a mullite tube coated with a 5-cm-long catalyst layer (cata-
lyst load � 35 mg). In the absence of any catalyst, flow rate
and feed composition were such that conversion of ethane
�
and oxygen were appreciable over 650–700 C, only, but gas-
Theory of ethane oxidative pyrolysis. Elsewhere, the au- phase reactions were very fast and conversion of ethane and
�
thors have simulated the expected behavior of a purely O2 amounted to 80 and 60% , respectively, at 770 C oven
homogeneous plug flow reactor operating under the same temperature. In the presence of the Pt/Al2O3 catalyst, re-
�
conditions as those explored experimentally (21). Simula- actants conversion started already at 250 C. Up to about
�
tions were performed by means of a detailed kinetic scheme 600 C oven temperature, reactants conversion did not in-
of homogeneous oxidative pyrolysis of hydrocarbons (24, crease significantly with temperature, but increased with
�
2
5). The data herein reported are fully in line with the increasing amount of catalyst. Above 600 C, both on the
theoretical predictions. Experiments and theory concur to shorter and on the longer catalyst layers ethane conversion
evidence that for a given pressure and feed composition, had a sharp increase, which could be partly explained by
the product distribution is practically a unique function the onset of the homogeneous reactions. Concerning the
of the degree of ethane conversion, and does not depend on single reaction products (Fig. 7), CO2 was only produced
the single values of temperature and contact time. In gen- in traces in the empty annular reactor; in the presence
eral, ethylene selectivity is a decreasing function of ethane of the short layer of catalyst CO2 productivity increased
conversion; for instance, at C2/O2 = 1, 95% selectivity is signifi- cantly and could completely explain ethane con-
�
measured and calculated at ethane conversions below 10% , version up to 500–550 C. Further increase was caused by
and selectivity decreases down to 60% at increasing con- the use of a longer catalyst layer. In both cases, the out-
version up to 80% . According to model predictions, such let CO2 flow decreased with increasing temperature above
�
�
a trend is valid either at intermediate temperature (600 C) 500 C.
and contact times in the order of 0.5–1 s (the operating con- Water was produced in appreciable amounts in the ab-
�
ditions for the experiments in the annular reactor), either at sence of catalyst above 700 C. With addition of a small
1
�
000 C at contact times in the order of 5–10 ms (the operat- catalyst load, however, its production initiated at low

5
0
BERETTA AND FORZATTI
FIG. 6. Ethane ODH in the presence of Pt/� -Al2O3 catalyst at varying catalyst layer length. Conversion of ethane and oxygen. Feed composition,
C2H6/O2/N2 = 2/2/96; flow rate, 1 L/min at 273 K, 1 atm.
�
Above 500 C, CO was formed in the presence of the
temperature and increased with increase of the catalyst
layer length. The production of water and CO2 completely short catalyst layer and the production of this species in-
accounted for the conversion of ethane and oxygen ob- creased significantly with temperature; such a production
served in the presence of the catalyst in the temperature was higher than that associated to the single contribution
�
window between 250 and 500 C; thus deep oxidation of of gas-phase reactions. Also, CO molar flow increased with
ethane was firstly catalyzed by Pt/Al2O3 at low T.
further increase of the catalyst load. A similar trend was
FIG. 7. Ethane ODH in the presence of Pt/� -Al2O3 catalyst at varying catalyst layer length. Outlet flow rate of CO2, H2O, CO, and H2. Conditions
as in Fig. 6.

ODH OF ETHANE AT HIGH TEMPERATURE AND SHORT CONTACT TIME
51
FIG. 8. Ethane ODH in the presence of Pt/� -Al2O3 catalyst at varying catalyst layer length. Outlet flow rate of ethylene and methane. Conditions
as in Fig. 6.
found for the formation of hydrogen: though appreciable
Additional experiments were then realized in the ab-
at high temperature in the empty reactor, its production sence of oxygen. In the three cases of empty reactor, pres-
�
was anticipated at 500 C and was progressively enhanced ence of a short catalyst layer and presence of a longer cata-
by the addition of increasing amounts of catalyst. In the lyst layer, conversion of ethane was the same in the whole
range of intermediate and high temperature, the catalyst temperature range investigated. The products uniquely ob-
was thus active in the production of synthesis gas, presum- served in the outlet stream were ethylene and hydrogen; the
ably via partial oxidation of ethane. Ongoing research is addition of catalyst did not affect the production of these
presently addressing the mechanism and kinetics of CO /H2 species due to single gas-phase reactions (Fig. 10).
production over the Pt-catalyst.
These results were in complete agreement with those
Concerning the production of ethylene, Fig. 8 shows that obtained from previous experiments with propane/
the olefin molar flow was maximum, though small, in the oxygen/nitrogen mixtures (22). Also in the case of propane,
empty reactor, while it decreased with increasing cata- in fact, it was observed that the addition of catalyst pro-
lyst load. The same trend was shown by methane outlet moted the formation of the deep oxidation and partial ox-
flow. These results suggest that hydrocarbons were formed idation products, while decreasing the production of ethy-
uniquely in the gas-phase while the catalyst did not con- lene and propylene, which was maximum in the absence
tribute to their production. On the opposite the addition of of any catalyst layer. Also in that case, then, the oxidation
the Pt/Al2O3 catalyst contributed to partly decompose the activity of the catalyst decreased with decreasing O2 feed
amount of ethylene and methane produced by the homo- concentration, and in the absence of oxygen only gas-phase
geneous reactions.
reactions occurred.
Analogous experiments were carried out by halving the
oxygen feed content. Results are shown in Fig. 9. The de-
crease of oxygen concentration caused a significant de-
crease of ethane conversion, but it still grew with the ad-
ETHANE OXIDATION TESTS OVER BaMnAl11O19
IN THE ANNULAR REACTOR
dition of catalyst and the comparison with the tests in the
The Ba-, Mn-substituted hexa-aluminate catalyst was
�
empty reactor showed that ethane conversion up to 600 C first tested in the annular reactor, in order to estimate its ac-
could be explained only by the contribution of the catalyst tivity in the oxidation of ethane. A lower activity than that
phase. The production of all the species was of course lower of the Pt/Al O catalyst was expected; thus the experiments
2
3
than in the previous case. Still, the comparison of the three were carried out at relatively low flow rate and high reac-
cases clearly showed the promoting effect of the catalyst tant concentration. The results reported in Fig. 11 were ob-
addition on the production of CO2 and H2O (which were tained with a 1.7-cm-long catalyst layer, a total flow rate of
3
formed only at high temperature in the empty reactor) and 120 cm (NTP)/min and an ethane/air feed with ethane/
on the formation of CO and H2, which became important oxygen ratio = 1. They are compared with the results of ex-
�
products above 500 C. Concerning the hydrocarbons, also periments which were carried out with the Pt-catalyst under
in this case the production of ethylene was maximum in the the same operating conditions. The results of blank experi-
absence of any catalyst, while it decreased with increasing ments in the empty reactor are also reported as a reference.
length of the catalyst layer. A similar trend was shown by
methane production.
The activity of the noble metal catalyst was so high that
the reactant conversions likely underwent mass transfer

5
2
BERETTA AND FORZATTI
FIG. 9. Ethane ODH in the presence of Pt/� -Al2O3 catalyst at varying catalyst layer length. Outlet flow rate of the single reaction products. Feed
composition, C2H6/O2/N2 = 2/1/97; flow rate, 1 L/min at 273 K, 1 atm.
limitation as soon as reaction started. Instead, the activ- indicates that the onset of homogeneous reactions could
ity of the BaMnAl11O19 catalyst was significantly lower; thoroughly explain the data. The flow rate of the single re-
ethane and oxygen conversionswere first observed at nearly action products are reported in Fig. 12; they refer uniquely
�
4
50 C oven temperature and they increased progressively to the BaMnAl11O19 which was tested with a 1.7-cm-long
�
with increasing T. Over 700 C, the conversion of reactants and a 5-cm-long catalyst layer. CO2 and H2O were the
was complete also in the case of the BaMnAl11O19 catalyst, only reaction products observed in the presence of the
but comparison with the activity of the gas-phase clearly catalyst and their production increased with increase of

ODH OF ETHANE AT HIGH TEMPERATURE AND SHORT CONTACT TIME
53
FIG. 10. Ethane ODH in the presence of Pt/� -Al2O3 catalyst at varying catalyst layer length. Outlet flow rate of ethylene and hydrogen. Feed
composition, C2H6/N2 = 2/98; flow rate, 1 L/min at 273 K, 1 atm.
catalyst load; other species, including CO, H2, ethylene, dynamics of temperature. In few seconds the reactor tem-
�
�
and methane were formed above 650 C but their produc- perature increased up to 900 C, and then the following T
tion was almost entirely accounted for by the contribution increase slowed down and temperature was approximately
of homogeneous reactions. Thus, opposite to the Pt-Al2O3 constant after 30 min from the feed injection.
catalyst, the mixed oxide based catalyst was proven to be
At varying ethane/oxygen feed ratio, reactor tem-
uniquely active in the deep oxidation of ethane, while no perature, ethane conversion, and product distribution
activity was shown in the partial oxidation to CO and H2.
changed. The maximum temperature of the catalytic
portion established close to the theoretical adiabatic tem-
perature and decreased with increasing C2/O2 ratio in line
with the decreasing exothemic content of the reaction.
Conversion of ethane decreased from over 95% to nearly
ODH OF ETHANE IN AUTOTHERMAL CONDITIONS
OVER Pt/� -Al2O3
Tests of ethane ODH in the presence of the Pt-Al2O3 40% . CO and H2 were the major products at equimolar
catalyst were realized in the autothermal reactor at varying feed composition but the production of ethylene prevailed
inlet temperature and feed composition. Results have been at ethane/O2 ratios >1.4. The olefin selectivity grew to
reported in detail elsewhere (17) and are herein summa- 55% .
rized in Table 1.
At higher preheat temperature, a similar trend was ob-
In all the experiments with the Pt-catalyst, reactor light served. At increasing ethane/oxygen feed ratios, ethane
off occurred instantaneously as soon as the ethane/air conversion and temperature decreased but ethylene se-
stream was let in. As an example, Fig. 13 reports the typical lectivity increased. A larger production of ethylene was
FIG. 11. Comparison of ethane and oxygen conversions in the three cases of absence of catalyst, presence of a 1.3-cm-long Pt/Al2O3 catalyst and
3
presence of a 1.7-cm-long BaMnAl11O19 catalyst. Feed composition, C2H6/O2/N2 = 1/1/4; flow rate, 120 cm /min at 273 K, and 1 atm.

5
4
BERETTA AND FORZATTI
FIG. 12. Ethane oxidation tests over the BaMnAl11O19 catalyst. Outlet flow rate of CO2, H2O, CO, H2, ethylene, and methane over 1.7-cm-long
and 5-cm-long catalyst layers and in the absence of any catalyst. Operating conditions as in Fig. 11.

ODH OF ETHANE AT HIGH TEMPERATURE AND SHORT CONTACT TIME
55
TABLE 1
Preliminary experiments were realized over the Pt/� -
Al2O3 catalyst with the purpose of studying the effect of
H2 cofeed. No significant promotion of ethylene formation
was found and this is in line with the same findings of Bodke
et al. (14); they reported in fact a moderate H2 effect for the
single Pt-coated foam monoliths, while the selective oxida-
tion of H2 resulted a peculiar characteristic of the bimetallic
Sn–Pt system.
Performances of the Pt-Containing Autothermal Reactor
%
C-mol Selectivity
�
C2/O2 C2 conv. (% ) T ( C)
CO2
CO
C2H4 CH4 C3 + C4
�
Tin = 200 C
0
1
1
1
2
.98
.36
.43
.74
.12
95.3
75.5
68.2
55.2
41.0
998
960
947
924
890
9.6
47.5 37.4
28.1 56.4
30.4 54.2
27.8 56.6
30.0 53.4
5.8
4.2
3.6
2.8
2.0
0.3
2.6
1.8
2.2
1.2
8.7
10.0
10.6
13.4
ETHANE ODH IN AUTOTHERMAL CONDITIONS
OVER BaMnAl11O19
�
Tin = 400 C
The hexa-aluminate catalyst was deposited onto the
FeCrAlloy fibrous support and tested in the adiabatic
reactor (catalyst load, 90 mg). The purpose of the present
experiments was to verify the possible use of a Pt-free
oxidation catalyst for realizing the ignition of homoge-
neous reactions. Preheat temperature for the inlet gas
stream was set to the minimum temperature at which
1
1
1
2
.27
.45
.72
.0
92.1
84.6
70.7
59.9
1002
985
953
936
8.3
8.1
7.9
7.9
27.1 53.9
23.0 59.8
20.5 63.9
20.8 64.7
6.7
5.7
4.4
3.6
4.0
3.4
3.3
3.0
Note. Feed, ethane/air; total flow rate, 1 L(NTP)/min; contact time,
ms.
5
the BaMn-catalyst resulted active in the oxidation of
�
observed than at the lower preheat temperature. An ethy- ethane, that is 450–500 C. Ethane–air mixtures were fed
lene yield as high as 50% was observed at ethane/O2 = 1.45. at varying ethane/oxygen ratios with a total flow rate of 1
Elsewhere (21) these data were compared with the simula- L(NTP)/min, corresponding to approximately 5 ms contact
tions of a purely homogeneous process and it was suggested time in the catalyzed portion of the reactor.
that, at increasing reactor temperature, the contribution of
Figure 14 reports an example of light-off of the reactor;
the catalyst phase to the production of COx became less im- while in the presence of Pt/Al O catalyst light-off was in-
2
3
portant;thiswaslikelydue to the lower sensitivityofhetero- stantaneous, in the case of the Pt-free catalyst two distinct
geneous kinetics to the increase of temperature, compared phases appeared during reactor ignition. Initially, temper-
with homogeneous kinetics.
ature increased slowly; within several minutes, the reactor
�
The Pt-containing reactor was stable as temperature and temperature grew from the preheat value to nearly 700 C
product distribution kept constant for several hours, until but subsequently a sharp increase of temperature occurred
operating conditions were changed. Light-off of the reactor up to the final temperature value which kept constant with
always occurred very rapidly even after weeks of operation time and depended on feed composition. During the initial
with no need of regeneration. Apparently, the catalyst did slow T increase only oxidation products were detected in
not suffer from coking or deactivation.
the outlet product mixture, while after reach of the final
FIG. 13. ODH of ethane in autothermal reactor in the presence of
FIG. 14. ODH of ethane in autothermal reactor in the presence of
Pt/Al2O3 catalyst. Light-off of the reactor. Flow rate, 1 L/min at 273 K and
BaMnAl11O19 catalyst. Light-off of the reactor. Flow rate, 1 L/min at 273 K
and 1 atm; feed, ethane/air; preheat temperature, 500 C.
�
�
1
atm; feed, ethane/air; preheat temperature, 200 C.

5
6
BERETTA AND FORZATTI
FIG. 15. ODH of ethane in autothermal reactor in the presence of BaMnAl11O19 catalyst. Effect of ethane/O2 feed ratio on reactants conversion,
�
reactor temperature, distribution of products. Flow rate, 1 L/min at 273 K and 1 atm; feed, ethane/air; preheat temperature, 500 C. Dashed curves,
calculated selectivity of ethylene and CO for a purely homogeneous adiabatic reactor, operating at the same inlet temperature and degree of reactant
conversion.
temperature level ethylene was also present among the re- in low amounts at all composition tested with selectivi-
action products. The temporal profile of temperature seems ties in the order of 10% each. Hydrogen was produced in
to reflect the relative rate and the stepwise onset of hetero- large amount at low ethane/oxygen feed ratios but its H-
geneous reactions and homogeneous reactions; the low ox- selectivity decreased with increasing C-feed content. Ethy-
idation activity of the BaMnAl11O19 catalyst was presum- lene yield was over 50% for ethane/oxygen >1.3.
ably responsible for the initial slow temperature increase
Experimental results were compared with the simula-
of the reactor, while once the threshold for ignition of ho- tions of a detailed kinetic scheme of homogeneous reac-
mogeneous reactions was reached, then sudden light-off of tions and Fig. 15 also shows in dashed lines the calcu-
oxidative pyrolysis occurred.
lated C-selectivity of ethylene and CO. Calculations refer
Once ignited, the Pt-free autothermal reactor behaved to an isothermal adiabatic reactor with inlet temperature
�
similarly to the one with the Pt-catalyst. Temperatures of 500 C, which operated at the same conversion level of
were in general higher likely due to the higher preheat reactants as measured in the experiments, at few millisec-
temperature level. As reported in Fig. 15, at equimolar onds contact time. A very close agreement was found be-
�
ethane/oxygen ratio, temperature was about 1150 C and it tween experimental and predicted data.
�
decreased to 980 C at ethane/oxygen ratio = 2. Ethane con-
Indeed the hexa-aluminate catalyst behaved as ignitor of
version was also higher than in the presence of the Pt/Al2O3 the high-temperature homogeneous process; presumably
catalyst. As expected, ethane conversion decreased with because of the high temperature level reached and of the
increasing C2/O2 feed ratio; complete conversion was ob- low intrinsic activity of the catalyst, homogeneous reactions
served at C2/O2 = 1 and nearly 75% conversion was ob- governed the final performance of the reactor, with conse-
tained at feed ratio close to 2. The product distribution quent minimum production of COx and maximum selectiv-
is also shown in Fig. 15. Ethylene selectivity increased ity to ethylene.
significantly with increasing ethane/oxygen feed ratio and
Concerning the catalyst stability, after few consecutive
amounted to 68% at C2/O2 = 2. Accordingly, CO selectiv- cycles consisting of light-off, high-temperature operation,
ity decreased down to 15% . CO2 and CH4 were produced and shut down, the BaMnAl11O19 was unable to activate the

ODH OF ETHANE AT HIGH TEMPERATURE AND SHORT CONTACT TIME
57
�
process at the same inlet temperature of 500 C. However, in the final performance of the reactor, being responsible
�
a treatment in air at 500 C for several hours during which of the production of olefins. Indeed, the observed product
CO2 was formed in traces reestablished the catalyst activ- distributions compared well with the predictions of the ex-
ity; afterward, in fact, ignition of the reactor occurred with pected performance of a purely homogenous adiabatic re-
the same dynamics as shown in Fig. 14. The BaMnAl11O19 actor.
thus suffered from reversible deactivation, presumably
The hypothesis that the main role of the Pt-catalyst
due to coke deposition during the high temperature opera- was that to accelerate the ignition of homogeneous reac-
tion that the lower oxidation activity of the catalyst was not tions through oxidation reactions was further verified by
sufficient to prevent. A deeper understanding of the deac- testing for comparison a Pt-free oxidation catalyst. The
tivation phenomenon was beyond the scope of the present BaMnAl11O19 was proved to be uniquely active in the total
work, wherein the BaMnAl11O19 catalyst has been simply oxidation of ethane to CO2 and H2O. However, in the adia-
�
taken as a model of Pt-free oxidation catalysts.
batic reactor, at the proper preheat temperature of 500 C,
the hexa-aluminate catalyst produced light-off of the reac-
tor and yields over 50% to ethylene were observed from
ethane/air mixtures; thus, once ignited the Pt-free adiabatic
reactor performed even better than the Pt-containing reac-
tor. The observed product distribution was compared with
the predictions of a purely homogeneous reactor and a very
good match wasfound between the measured and predicted
ethylene formation. These data suggest that in principle
high ethylene yields can be obtained by exploiting the com-
bination of an oxidation catalyst with the fast homogeneous
reactions of oxidative pyrolysis.
CONCLUSIONS
Several works in the literature have shown that oxidative
dehydrogenation of ethane (and other light alkanes) over
Pt-catalysts leads to the selective production of ethylene,
when the process is carried out in autothermal reactors
�
at temperatures close to 1000 C and at millisecond con-
tact times. Similar results were obtained by the authors by
carrying out the reaction over a Pt/� -Al2O3 catalyst sup-
ported over a metallic tissue under adiabatic conditions.
However, in the author’s experience the short contact time
autothermal reactors are not versatile tools for the kinetic
investigation: temperature is not controlled by the operator
and cannot be freely varied, high reactant concentrations
are needed and the comparison with blank experiments in
ACKNOWLEDGMENTS
This work was performed under the financial support of ENI and Snam-
progetti. The authors thank Dr. Luca Basini (Snamprogetti) for his contri-
bution to the discussion of the work. Prof. Ranzi (Politecnico di Milano)
the absence of catalyst (an important “home-work” when is also gratefully thanked for sharing his expertise on the oxidative py-
high temperatures are involved) cannot be realized. Fur- rolysis of hydrocarbons and for providing the model simulations herein
reported.
thermore, the short contact times are realized in the adi-
abatic reactors by supporting the catalyst over high void
fraction materials which allow to use high flow rates while
minimizing pressure drops (foam monoliths, gauzes, and
others); thus, the contribution of homogeneous reactions
in the gas volume which is surrounded by the catalyst sup-
port should not be excluded a priori.
REFERENCES
1
2
3
4
5
. Baerns, M., and Buyevskava, O., Catal. Today 45, 13 (1998).
. Huff, M., and Schmidt, L. D., J. Phys. Chem. 97, 11815 (1993).
. Huff, M., and Schmidt, L. D., J. Catal. 149, 127 (1994).
. Huff, M., and Schmidt, L. D., J. Catal. 155, 82 (1995).
. Dietz, A. G., Carlsson, A. F., and Schmidt, L. D., J. Catal. 176, 459
(1996).
. Pereira, C., Science 285, 670 (1999).
. Schmidt, L. D., Siddall, J., and Bearden, M., AIChE J. 46, 1492 (2000).
. Huff, M., Torniainen, P. M., and Schmidt, L. D., Catal. Today 21, 113
In this work a nonadiabatic structured reactor was used
to investigate separately the single contributions of homo-
geneous and heterogeneous reactions in the short contact
time oxidative dehydrogenation of ethane over the Pt/� -
Al2O3 catalyst. In the absence of catalyst, activity tests
showed that the radical gas-phase process was highly se-
lective toward the formation ethylene; theoretical analyses
have confirmed the experiments. The studyofthe single role
6
7
8
(
1994).
9
. Yokoyama, C., Bharadwaj, S. S., and Schmidt, L. D., Catal. Lett. 38,
181 (1996)
1
0. Flick, D. W., and Huff, M. C., J. Catal. 178, 315 (1998).
11. Huff, M., and Schmidt, L. D., AIChE J. 42, 3483 (1996).
�
of heterogeneous reactions showed that from low (200 C)
1
2. Faravelli, T., Goldaniga, A., Ranzi, E., Dietz, A., Davis, M., and
Schmidt, L. D., in “Natural Gas Conversion V. Studies in Surface
Science and Catalysis” (A. Parmaliana et al., Eds.), Vol. 119, p. 575.
Elsevier, Amsterdam, 1998.
3. Flick, D. W., and Huff, M. C., Appl. Catal. A 187, 13 (1999).
4. Bodke, A. S., Olschki, D. A., Schmidt, L. D., and Ranzi, E., Science
up to high temperatures the Pt/� -Al2O3 catalyst was active
in the deep and partial oxidation of ethane to COx, while
no positive proof was found of an heterogeneous forma-
tion of ethylene. It is thus believed that in the short contact
time adiabatic reactor, a synergism between catalytic com-
bustion and gas-phase oxidative pyrolysis is established;
the former seems responsible for the rapid light-off of the
reactor, while the latter seems to have an important role
1
1
2
85, 712 (1999).
1
1
5. Goetsch, D. A., and Schmidt, L. D., Science 271, 1560 (1996).
6. Lødeng, R., Lindv a˚ g, O. A., Kvisle, S., Reiner-Nielsen, H., and
Holmen, A., in “Natural Gas Conversion V. Studies in Surface Science

5
8
BERETTA AND FORZATTI
and Catalysis” (A. Parmaliana et al., Eds.), Vol. 119, p. 641. Elsevier, 21. Beretta, A., Ranzi, E., and Forzatti, P., Chem. Eng. Sci., in press.
Amsterdam, 1998.
7. Lødeng, R., Lindv a˚ g, O. A., Kvisle, S., Reiner-Nielsen, H., and
Holmen, A., Appl. Catal. A 187, 25 (1999).
22. Beretta, A., Piovesan, L., and Forzatti, P., J. Catal. 187, 455(1999).
23. Groppi, G., Bellotto, M., Cristiani, C., Forzatti, P., and Villa, P. L.,
Appl. Catal. A 104, 101 (1993).
1
1
1
2
8. Huff, M., Androulakis, I. P., Sinfelt, J. H., and Reyes, S. C., J. Catal. 24. Dente, M., and Ranzi, E., in “Pyrolysis: Theory and Industrial Prac-
1
91, 46 (2000).
tice” (L. F. Albtight et al., Eds.), Chap. 7. Academic Press, New York,
1983.
25. Ranzi, E., Faravelli, T., Gaffuri, P., Sogaro, A., D’Anna, A., and
Ciajolo, A., Comb. Flame 108, 24 (1997).
9. Zerkle, D. K., Allendorf, M. D., Wolf, M., and Deutschmann, O.,
J. Catal. 196, 18 (2000).
0. Beretta, A., Ranzi, E., and Forzatti, P., J. Catal. 184, 469 (1999).