Dehydrogenation of n-butane on the industrial microspherical chromia-alumina catalyst AOK-73-24 in a membrane reactor

Article abstract of DOI:10.1134/S0965544113060054

The basic features of the catalytic membrane dehydrogenation of n-butane have been studied in a reactor with membrane modules based on Pd/Ag foil of 9.3 and 30 μm thickness and in the absence of the membrane (temperature, 500-550°C; feed space velocity, 150-1200 h^-1). It has been shown that in the absence of the membrane, the dehydrogenation of n-butane occurs in the kinetic region. The membrane thickness has a significant effect on the performance of the catalytic membrane reaction, presumably, because of the difference in the rate of H₂ withdrawal from the reaction mixture. In the reactor with the 9.3-μm thick Pd/Ag-foil, the reaction is controlled by the H₂ formation rate. As the thickness of the palladium foil is increased to 30 μm, the reaction proceeds to the diffusion region, thereby resulting in both enhancement of selectivity for butenes and a reduction of the yield of hydrocarbon deposits. Pleiades Publishing, Ltd., 2013.

Full text of DOI:10.1134/S0965544113060054

ISSN 0965ꢀ5441, Petroleum Chemistry, 2013, Vol. 53, No. 6, pp. 401–406. © Pleiades Publishing, Ltd., 2013.
Original Russian Text © L.P. Didenko, V.I. Savchenko, L.A. Sementsova, O.G. Sheverdenkina, L.A. Bykov, 2013, published in Neftekhimiya, 2013, Vol. 53, No. 6, pp. 449–455.
Dehydrogenation of nꢀButane on the Industrial Microspherical
Chromia–Alumina Catalyst AOKꢀ73ꢀ24 in a Membrane Reactor
L. P. Didenko, V. I. Savchenko, L. A. Sementsova, O. G. Sheverdenkina, and L. A. Bykov
Institute of Problems of Chemical Physics, Russian Academy of Sciences, Chernogolovka, Moscow oblast, Russia
eꢀmail: ludi@icp.ac.ru
Received April 3, 2013
Abstract—The basic features of the catalytic membrane dehydrogenation of
reactor with membrane modules based on Pd/Ag foil of 9.3 and 30 m thickness and in the absence of the
membrane (temperature, 500–550°C; feed space velocity, 150–1200 h–1). It has been shown that in the
absence of the membrane, the dehydrogenation of ꢀbutane occurs in the kinetic region. The membrane
thickness has a significant effect on the performance of the catalytic membrane reaction, presumably, because
of the difference in the rate of H2 withdrawal from the reaction mixture. In the reactor with the 9.3ꢀ m thick
Pd/Agꢀfoil, the reaction is controlled by the H2 formation rate. As the thickness of the palladium foil is
increased to 30 m, the reaction proceeds to the diffusion region, thereby resulting in both enhancement of
selectivity for butenes and a reduction of the yield of hydrocarbon deposits.
nꢀbutane have been studied in a
µ
n
µ
µ
Keywords: palladium membrane module, catalytic dehydrogenation,
DOI: 10.1134/S0965544113060054
nꢀbutane, chromia–alumina catalyst
Catalytic dehydrogenation of lower alkanes is an ity, thus opening the possibility of obtaining in princiꢀ
alternative petrochemical process for manufacturing ple the two desired products olefin and hydrogen in the
the valuable chemicals olefins from inexpensive and catalytic membrane process of dehydrogenation of
available natural gas and petroleum gas feedstock, light alkanes.
whose reserves are large and their use is highly ineffecꢀ
tive.
Previously, we investigated the reaction of catalytic
dehydrogenation of propane in a combination reactor
with a membrane module based on Pd/Ag foil of a
30 m thickness and showed that as a result of Н2
removal from the reaction mixture, the yield of propyꢀ
lene increased by a factor of 1.6–2 [7]. In this study,
the palladium membrane module was used for the
removal of Н2 from the products of catalytic dehydroꢀ
genation of nꢀbutane, another component of natural
gas or petroleum gas.
The objective of this study was to reveal the basic
features of the catalytic membrane dehydrogenation
of nꢀbutane on the industrial microspherical chroꢀ
The development of industrial processes for cataꢀ
lytic dehydrogenation of light hydrocarbon feedstock
is hindered by their high energy and capital intensity
due to the endothermicity of the reaction and thermoꢀ
dynamic constraints on the product yield. Thus, there
is a need for new technological solutions aimed at
enhancing the reaction performance, the feedstock
conversion, and the selectivity for the desired prodꢀ
ucts.
μ
One of the solutions is the membrane technology.
As a result of combining the reaction process with
hydrogen removal from the reaction zone, the feedꢀ
stock to olefin conversion may increase and the selecꢀ
tivity for cracking products can decrease. The concept
of the catalytic membrane process was first formulated
and implemented in studies by Academician
V.M. Gryaznov and coworkers [1–6], which gave rise
to a wealth of research in this field over the world.
mia–alumina catalyst AOKꢀ73ꢀ24.
EXPERIMENTAL
The reaction was run in the combined membrane
reactor including a hydrogenꢀpermeable module
based on Pd/Ag foil (15.0 wt % Ag); the foil was
The severe conditions of the dehydrogenation obtained by the process designed at the Moscow Speꢀ
reaction impose high requirements on the membrane cialty Alloy Processing Plant. The membrane modules
material, which should possess heat resistance, MM 1 and MM 2 used had palladium foils of 9.3 and
mechanical strength, and resistance to deactivation by 30 μm thickness, respectively. The designs of the reacꢀ
hydrocarbon deposits, along with a high Н2 selectivity. tor and the membrane module are detailed in [7, 8]. A
From this standpoint, palladium is of the greatest 2 cm3 (2.3 g) portion of the industrial microspherical
interest since palladium foil has the 100% Н2 selectivꢀ chromia–alumina catalyst AOKꢀ73ꢀ24, prepared by
401

402
DIDENKO et al.
impregnating the hydrargillite thermochemical actiꢀ selectivity. The selectivity for the formation of hydroꢀ
vation product with aqueous solutions of carbon deposits (SС, %) was calculated as
chromium and potassium salts (TU (Technical Speciꢀ
,
SС = 100 − S
+ S
+ S
+ S
+ S
C3H6 ΣC4H8
(
)
fications) 6ꢀ68ꢀ208ꢀ04), was placed in the retentate
compartment of the reactor. In this study, we used the
same batch of the catalyst and its regeneration was not
CH4
C2H4
C2H6
where
SC H , SC H , SC H , SΣC H are the selectivities
2 4 2 6 3 6 2 8
practiced. nꢀButane (100%) at a preset rate was fed on
for methane, ethylene, ethane, propylene, and
butenes, respectively, in %.
the catalyst through the holes located on the periphery
of the retentate compartment, and the resulting prodꢀ
ucts were withdrawn through the central hole and sent
to a chromatograph. As a driving force for the removal
of Н2 through the membrane, the stripping gas nitroꢀ
gen was used, which was fed counterꢀcurrently to
butane to the permeate compartment. The flow rates
of the stripping gas, which corresponded to the maxiꢀ
mum possible withdrawal of Н2 under the chosen conꢀ
ditions and were determined on the basis of a prelimiꢀ
nary study of the hydrogen permeability of MM 1 and
MM 2, were 40 and 150 cm3/min, respectively.
The gas flow rates were controlled with RRGꢀ12
gas flow regulators (Electropribor, Zelenograd). The
reactor was heated in an electric oven. The temperaꢀ
ture on the membrane and in the oven was monitored
with chromel–alumel thermocouples. The product
composition was analyzed in the onꢀline mode using a
Kristall 5000 chromatograph with flameꢀionization
and thermal conductivity detectors. The Н2 content in
the products was determined on a column packed with
RESULTS AND DISCUSSION
The basic features of the reaction were studied by
varying the temperature in the range of 500–550°С
and the feed space velocity within 150–1200 h–1. The
influence of H2 removal through the membrane on the
yield and composition of butenes was determined on
the basis of the results of comparative experiments
under the same conditions without H2 withdrawal
(“nonmembrane” reaction); for this purpose, a
gastight stainless steel plug was installed instead of the
membrane module.
The desired products of the catalytic dehydrogenaꢀ
tion of nꢀbutane were 1ꢀbutene, transꢀ2ꢀbutene, cisꢀ2ꢀ
butene, and a small amount (2–4%) of 1,3ꢀbutadiene.
The products also included Н2; the cracking products
СН4, С2Н4, С2Н6, and С3Н6 gases; and hydrocarbon
deposit (HCD).
Since both the Pd/Ag foil and stainless steel meshes
can exhibit catalytic activity in the nꢀbutane dehydroꢀ
genation reaction, blank runs in the absence of a cataꢀ
lyst were conducted prior to the study. The measureꢀ
molecular sieves 13X (2 mm
argon). The hydrocarbon composition of the products
was determined on an HPꢀAl/KCl column (0.5 mm
30 m; 80 ; carrier gas, helium). The amount of the
× 2 m; 50°С; carrier gas,
×
°С
ments were made in the membrane reactor at
Т =
products was calculated using the method of absolute
calibration. The relative error of the analysis did not
exceed 2%.
550 with the closed exit from the permeate comꢀ
°
С
partment. It was found that at a space velocity of
600 h–1, the feedstock conversion was 1.2%. Thus, the
The conversion of butane (
α
, %) and selectivity for
dehydrogenation of nꢀbutane on the foil and the mesh
proceeds to a small extent, and barely affects the charꢀ
acteristics of the reaction
products (S, %) were calculated by the following forꢀ
mulas:
in
C4H10
out
X
Vin – X
Vout
C4H10
Figures 1a, 1b, and 1c present the results of the
experiments without the withdrawal of Н2 from the
reaction and with its withdrawal through MM 1 at a
feed space velocity varied from 300 to 1200 h–1. It can
be seen that the yield of butenes in the nonmembrane
reaction (Fig. 1a) does not depend on the feed space
velocity, thereby suggesting the absence of diffusion
control of the reaction under these conditions. At the
same time, the yield of butenes in the catalytic memꢀ
brane reaction increases with the feedstock space
velocity to be ~33% at 900 h–1 and remains actually
unchanged with the further increase in the space
velocity (Fig. 1b). The yield of HCD decreases from
~45% at a feed space velocity of 300 h–1 to ~18% at
900 h–1 or higher (Fig. 1c). Note that the HCD accuꢀ
mulates not only in the catalyst bed, but also partly on
the surface of stainless steel mesh in the membrane
module without affecting the Pd/Ag foil. To remove
α = ꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀ × 1 0 0 ,
in
X
Vin
C4H10
nXprodVout
S = ꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀ × 100,
in
C4H10
out
Vout)
C4H10
4(X
Vin – X
where Vin is the volumetric flow rate of butane at the
in
C4H10
reactor inlet, cm3/min;
X
concentration of
nꢀ
butane in the gas stream at the reactor inlet, volume %;
вых
C4H10
X
is the concentration of nꢀbutane at the reactor
outlet, volume %; Vout is the volumetric flow rate of the
products at the reactor outlet, cm3/min; Xprod is the
concentration of the product in the mixture outflowꢀ
ing from the reactor, volume %; and
n is the number of
carbon atoms in the product molecule.
The yield of total butenes was calculated as the HCD after each run, the meshes were held in a muffle
product of ꢀbutane conversion and total butene furnace at = 600 for four hours.
n
Т
°С
PETROLEUM CHEMISTRY Vol. 53
No. 6
2013

DEHYDROGENATION OF
n
ꢀBUTANE ON THE INDUSTRIAL MICROSPHERICAL
403
100
90
80
70
60
50
40
30
20
10
(а)
60
50
40
30
20
10
(b)
60
50
40
30
20
10
(c)
1
2
4
3
3
1
2
4
3
1
0
0
0
40
80
120
40
80 120
40
80
120
Time, min
Time, min
Time, min
Fig. 1. Effect of the feed space velocity on the yields of total C butenes and hydrocarbon deposit at
T = 550°C and a Pd/Ag foil
µm (a) without removal of H and (b, c) with H removed at a stripping gas flow rate of 40 cm /min. nꢀButane
4
3
thickness of 9.3
space velocity: (
2 2
–1
1
) 300, (2) 600, (3) 900, and (4) 1200 h .
In the membrane reactor with MM 2, the effect of seen that an increase in the yield of the desired prodꢀ
the feedstock space velocity on the yield of butenes is ucts relative to the “nonmembrane” reference value
substantially different. As can be seen from Fig. 2, the is achieved only at low feedstock space velocity of
yield of butenes increases with a decrease in the space 150 h–1. The yield of butenes in the membrane process
velocity and is ~ 54% at 150 h–1 of butane. The speꢀ
cific feature of the reaction is a lower yield of HCD as
compared with the nonmembrane process (Fig. 3). A
reduction in the yield of HCD is not characteristic of
100
the catalytic membrane reaction. Typically, the yield
of HCD is higher in the case of H2 withdrawal from the
90
reaction mixture, a difference that is believed to be due
to the weakening of the stabilizing effect of H2 [9, 10].
80
70
60
50
40
30
20
10
The influence of temperature on the yields of
butenes and HCD in the presence of the membrane
module with the Pd/Agꢀfoil of varying thickness is
illustrated in Table 1.
It can be seen that in the range of 500–550°С, the
1
yield of butenes increases by approximately a factor of
5 when MM 1 and MM 2 are used. The yield of HCD
weakly depends on temperature in the reactor with
MM 2, whereas it increases with the temperature in
the case of MM 1. As a result, the selectivity for
butenes is higher in the catalyst–MM 2 membrane
system. From the data in Table 2 it is seen that the
selectivity for 1ꢀbutene, transꢀ2ꢀbutene and cisꢀ2ꢀ
butene is higher in the reactor with MM 2, rather than
MM 1; as the temperature increases, this difference
increases.
2
0
40
80
120
Time, min
The results of comparative runs of the catalytic
membrane reaction with the MM 2 membrane and the
nonmembrane reaction at various feedstock space
velocities are presented in Figs. 4a and 4b. It can be
Fig. 2. Effect of the feed space velocity on the yield of total
C butenes at = 550°C. Foil thickness, 30 m; stripping
ꢀButane space velocity:
T
µ
4
3
gas flow rate, 150 cm /min.
n
–1
(1) 150 and (2) 900 h .
PETROLEUM CHEMISTRY Vol. 53
No. 6
2013

404
DIDENKO et al.
50
(а)
50
(b)
40
30
40
30
2
2
1
20
10
20
10
1
0
0
40
80
120
40
80
120
Time, min
Time, min
Fig. 3. Yield of hydrocarbon deposit in experiments (
1
) with and (2) without H withdrawal at a Pd/Ag foil thickness of 30 µm,
2
–1
T = 550°C, and feed space velocities of (a) 150 and (b) 900 h
.
is ~55% under these conditions; however, it decreases from 30 to 9.3
μm. Consequently, the aboveꢀdescribed
with time and becomes equal to the “nonmembrane” differences in the characteristics of the reaction may
value after 120 min.
be due to different rates of Н2 withdrawal through
MM 1 and MM 2.
In similar comparative experiments with the MM 1
membrane, the yield of butenes did not exceed
the nonmembrane value at any feed space velocity.
Figures 5a–5c show that the greatest difference
between the catalytic membrane and nonmembrane
In the reactor with the MM 1 membrane having a
higher Н2 permeability and, hence, a higher Н2 withꢀ
drawal rate, the yield of HCD is significantly greater,
being ~45% at a butane space velocity of 300 h–1 and a
temperature of 550°C; this leads to catalyst deactivaꢀ
tion and, as a consequence, lowering the yield of
butenes. An increase in the feed space velocity diminꢀ
reactions at 550
of 300 h–1; as the space velocity increases, the differꢀ
ence is reduce; at 500 , it is negligible.
°С is observed at a feed space velocity
°С
These results indicate a significant influence of the ishes the deactivating effect of HCD by virtue of
membrane module type (MM 1 or MM 2) on the yield decreasing their amount; thus, the yield of butenes
of the desired products. Since these membranes differ increases, approaching the nonmembrane value. The
in the Pd/Ag foil thickness and, hence, hydrogen perꢀ characteristics of the catalytic membrane and nonꢀ
meability, it may be assumed that this is the main cause membrane reactions at 500
of the differences observed. Our earlier study [11] of data suggest that the process in the catalyst–MM 1
the hydrogen permeability of MM 1 and MM 2 at membrane system is controlled by the rate of formaꢀ
400 showed that the permeability increases by a facꢀ tion of Н2 in the dehydrogenation reaction of nꢀ
°С differ slightly. These
Т
=
°С
tor of 3.5 as the thickness of the Pd/Ag foil decreases butane. Owing to the high Н2 permeability of the
Table 1. Influence of temperature on the yield of butenes and hydrocarbon deposit in the reactor with membrane modules
with Pd/Ag foil of different thicknesses* at a feedstock space velocity of 1200 h–1 and stripping gas flow rates of 40 cm3/min
(MM 1) and 150 cm3/min (MM 2)
Temperature,
°
C
500
550
Yield, %
Butenes
HCD
Butenes
HCD
Foil thickness,
µm
9.3
30
Without membrane
6.2
4.3
8.0
3.1
8.1
5.2
30.8
21.1
41.2
14.2
10.1
22.5
* Data by the 80th min of the reaction.
PETROLEUM CHEMISTRY Vol. 53
No. 6
2013

DEHYDROGENATION OF
n
ꢀBUTANE ON THE INDUSTRIAL MICROSPHERICAL
405
Table 2. Selectivity for the formation of butenes in the presence of membrane modules with Pd/Ag foil of different thicknesses*
at a feedstock space velocity of 1200 h–1 and stripping gas flow rates of 40 cm3/min (MM 1) and 150 cm3/min (MM 2)
Temperature, 500 550
°C
Type of membrane module
Selectivity for butaneꢀ1, %
MM 1
30.0
MM 2
39.4
MM 1
22.2
MM 2
32.5
Selectivity for transꢀbuteneꢀ2, %
Selectivity for cisꢀbuteneꢀ2, %
* Data by the 80th min of the reaction.
27.4
13.0
36.0
17.1
30.4
13.0
42.1
18.5
membrane, the Н2 removal rate can exceed the rate of tion occurring in the kinetic region, the catalytic
its formation in the dehydrogenation reaction, and the
latter is the rateꢀlimiting step in the catalytic memꢀ
brane process as a whole. For this reason, the “memꢀ
brane effect”, i.e., enhancement of the performance
relative to the “nonmembrane” reaction, cannot be
achieved. This result suggests an important role of the
catalyst in the catalytic membrane system, as has been
noted in a number of studies [12–14]. For a productive
catalytic membrane system to be obtained, a highꢀperꢀ
meability membrane should be combined with a highꢀ
performance catalyst.
membrane reaction proceeds to the diffusion region,
as indicated by the effect of the feed space velocity on
the yield of butenes over the entire range examined.
The increase in the yield of butenes with respect to
the nonmembrane value at a feed space velocity of
150 h–1 and 550
°С cannot be called “membrane
effect” in the strict sense of the word. The increase is
associated with the different effects of the space velocꢀ
ity on the parameters of these reactions. Whereas the
yield of butenes in the nonmembrane reaction is indeꢀ
pendent of the feed space velocity over the velocity
range examined, that in the catalytic membrane proꢀ
cess increases with the decreasing space velocity of
butane, i.e., with an increase in the time of its contact
with the catalyst, and becomes higher than the nonꢀ
On passing to the MM 2 membrane based on the
Pd/Ag foil of a 30 μm thickness, the rate of Н2 withꢀ
drawal from the reaction mixture decreases and the
features of the catalytic membrane reaction in quesꢀ
tion are significantly altered. First of all, its occurrence
region changes: in contrast to the nonmembrane reacꢀ membrane value at a space velocity of 150 h–1.
100
90
80
70
60
50
40
30
20
10
(b)
100
90
(а)
80
70
1
60
50
2
1
2
40
30
20
10
0
0
40
80
120
40
80
120
Time, min
Time, min
Fig. 4. Change in the yield of total butenes in the comparative experiments (1) with H withdrawal (stripping gas flow
2
3
150 cm /min) and (2) without H withdrawal at a Pd/Ag foil thickness of 30 µ]m, T = 550°C, and feed space velocities of (a) 150
2
–1
and (b) 900 h
.
PETROLEUM CHEMISTRY Vol. 53
No. 6
2013

406
DIDENKO et al.
Т
= 550°C
(а)
Т
= 550°C
(б)
Т
= 500°C
(в)
70
60
50
40
30
20
10
70
60
50
40
30
20
10
70
60
50
40
30
20
10
2
1
2
1
2
1
0
0
0
40
80
120
40
80
120
40
80
120
Time, min
Time, min
Time, min
Fig. 5. Yield of total butenes in the (
1
) catalytic membrane and (
2) “nonmembrane” reactions. Foil thickness, 9.3 µm. nꢀButane
–1
space velocity: (a) 300 or (b, c) 1200 h
.
Thus, it has been shown that the catalytic memꢀ
brane dehydrogenation of ꢀbutane on the industrial
REFERENCES
n
1. V. M. Gryaznov, Dokl. Akad. Nauk SSSR 189, 794
microspherical chromia–alumina catalyst AOKꢀ73ꢀ
24 significantly differs in features when Pd/Ag memꢀ
branes of various thicknesses are used, with the differꢀ
ences being likely due to the different rates of H2 withꢀ
drawal from the reaction mixture. It is assumed that
the reaction with the use of the membrane module
(1969).
2. V. M. Gryaznov, SU Patent No. 274,092 (1964).
3. N. N. Mikhalenko, E. V. Khrapova, and
V. M. Gryaznov, Zh. Fiz. Khim. 60, 511 (1986).
4. V. M. Gryaznov, M. M. Ermilova, L. S. Morozova,
et al., J. LessꢀComm. Met. 89, 529 (1983).
based on the Pd/Agꢀfoil of a 9.3
ited by the H2 formation rate. The catalytic membrane
reaction in the reactor with the Pd/Ag foil of a 30
μm thickness is limꢀ
5. N. V. Orekhova, M. M. Ermilova, and V. M. Gryaznov,
Dokl. Akad. Nauk SSSR 321 789 (1991).
μm
6. E. M. Chistov, A. P. Mischenko, N. R. Roshan, and
V. M. Gryaznov, in Abstracts of the Second Conference on
Catalysis in Membrane Reactors (Moscow, 1996).
thickness occurs in the diffusion region. With this foil,
the yield of the hydrocarbon deposit is lower, a case
that is untypical of the membrane dehydrogenation of
alkanes; as a result, a higher selectivity for butenes is
observed. The yield of total butenes of about 55%,
which is greater than that in the nonmembrane proꢀ
cess, has been achieved in the reactor with the memꢀ
7. L. P. Didenko, V. I. Savchenko, L. A. Sementsova,
et al., Pet. Chem. 53, 27 (2013).
8. L. P. Didenko, M. S. Voronetskii, L. A. Sementsova,
et al., Al’tern. Energ. Ekol., No. 10, 154 (2010).
9. M. Sheintuch and R. M. Dessau, Chem. Eng. Sci. 51
535 (1996).
,
brane module based on palladium foil of a 30 μm
thickness at
T = 550°C and a feed space velocity of
10. J. P. Collins, R. W. Schwartz, R. Sehgal, et al., Ind. Eng.
1
150 h⎯ .
Chem. Res. 35, 4398 (1996).
11. L. P. Didenko, V. I. Savchenko, V. V. Barelko, and
L. A. Bykov, Al’tern. Energ. Ekol. (2013) (in press).
ACKNOWLEDGMENTS
12. P. Ciavarella, D. Casanave, S. Moueddeb, et al., Catal.
Today 67, 177 (2001).
This work was performed on the equipment of the
Center for Collective Use “New petrochemical proꢀ
cesses, polymer composites, and adhesives” under the
13. S. Miachon and J. A. Dalmon, Top. Catal. 29, 59
(2004).
Russian Academy of Sciences Presidium Basic 14. D. Casanave, P. Ciavarella, and K. Fiaty, Chem. Eng.
Sci. 54, 2807 (1999).
Research Program no. 3 “Energy aspects of deep proꢀ
cessing of fossil and renewable hydrocarbon sources.”
Translated by S. Zatonsky
No. 6 2013
PETROLEUM CHEMISTRY Vol. 53