Isotopic tracer studies of propane reactions on H-ZSM5 zeolite

Article abstract of DOI:10.1021/jp9824860

Reactions of ¹³C-labeled alkanes show that chain growth and cyclization reactions on H-ZSM5 require the initial formation of the corresponding alkene and its extensive participation in rapid oligomerization/β-scission reactions before cyclization occurs. The role of alkene intermediates was established by the initial formation of predominantly unlabeled products from mixtures of propene and propane-2-¹³C reactants. Aromatic products of propane-2-¹³C reactions on H-ZSM5 contain similar fractions of ¹³C-atoms and binomial isotopomer distributions. Sequential formation, rapid intramolecular isomerization, and β-scission reactions of long surface chains must occur during each aromatization turnover in order to form such binomial ¹³C isotopomer distributions.

Full text of DOI:10.1021/jp9824860

9284
J. Phys. Chem. B 1998, 102, 9284-9289
Isotopic Tracer Studies of Propane Reactions on H-ZSM5 Zeolite
Joseph A. Biscardi and Enrique Iglesia*
Department of Chemical Engineering, UniVersity of California at Berkeley, Berkeley, California 94720
ReceiVed: June 4, 1998; In Final Form: August 27, 1998
Reactions of 13C-labeled alkanes show that chain growth and cyclization reactions on H-ZSM5 require the
initial formation of the corresponding alkene and its extensive participation in rapid oligomerization/â-scission
reactions before cyclization occurs. The role of alkene intermediates was established by the initial formation
13
of predominantly unlabeled products from mixtures of propene and propane-2- C reactants. Aromatic products
1
3
13
of propane-2- C reactions on H-ZSM5 contain similar fractions of C-atoms and binomial isotopomer
distributions. Sequential formation, rapid intramolecular isomerization, and â-scission reactions of long surface
chains must occur during each aromatization turnover in order to form such binomial C isotopomer
distributions.
1
3
1
. Introduction
reactant conversion. The reaction mixture was circulated using
3
-1
a graphite gear micropump (250 cm min , Cole-Parmer) that
allowed differential reactor operation (<2% reactant conversion
per pass). All propane reactions were carried out at 773 K.
The chemical composition of reaction mixtures was obtained
using capillary column gas chromatography with flame ioniza-
tion detection (Hewlett-Packard 5890, HP-1 methyl silicone
column, 50 m, 0.32 mm diameter, 1.05 µm thickness) and
electron-impact mass spectrometry (Hewlett-Packard 5972 mass
In contrast with gas-phase pyrolysis pathways, which lead
to tars and coke, non-oxidative catalytic reactions of light
alkanes on H-ZSM5 form C6-C9 aromatics. The restricted
channel environment in ZSM51 and the ability of these catalysts
to maintain a high hydrogen surface pressure during alkane
reactions3 inhibit carbon formation and lead to their unique
stability at the high temperatures and low H2 concentration
required for aromatization of alkanes. Light alkane chain growth
pathways on zeolitic acids involve dehydrogenation, dimeriza-
tion, cyclization, and aromatization steps, which proceed in
,2
,4
1
3
selective detector).
C isotopomers were calculated from
fragmentation patterns using mathematical models that correct
for the mass spectral fragmentation and natural isotopic impuri-
5
-7
parallel with acid-catalyzed and thermal cracking steps.
1
3
ties.
Numerous studies have addressed the mechanism of chain
growth and cyclization during light alkane reactions with
primary emphasis on measuring the effect of catalyst composi-
tion on reaction rate and selectivity.5-12 Here, we report direct
evidence for the mechanistic sequence and for reactive inter-
mediates using 13C isotopic tracer and exchange methods.
Propane (Matheson, chemical purity > 99.5%), propene
Matheson, chemical purity > 99.0%), ethene (Matheson,
chemical purity > 99.0%), and propane-2- C (Cambridge
Isotopes, chemical purity > 98.0%, isotopic purity > 99.0%)
were used as reactants without further purification. Helium
(
1
3
(Linde, chemical purity > 99.995%) and hydrogen (Matheson,
chemical purity > 99.5%) were purified by passage through
oxygen and molecular sieve traps (13X and 5A) at room
temperature. Helium was used as an inert diluent in most of
the catalytic studies.
2
. Experimental Conditions
.1. Catalyst Preparation. Na ions in Na-ZSM5
+
2
+
(
Zeochem) were replaced with NH4 by contacting the zeolite
with a solution of 0.67 M NH4NO3 (Fisher, Certified ACS,
98.0%) at 353 K for 10 h. Solids were separated from the
Propane conversion turnover rates are reported as the molar
rate of propane conversion per g-atom Al in the catalyst charge.
Batch reactor data are shown as product site yields (moles of
product g-atom Al) or reactant turnovers (moles of C3H8
converted per g-atom Al) as a function of contact time. The
slopes of these plots give propane conversion turnover rates or
product site-time yields. Hydrocarbon selectivities are reported
on a carbon basis as the percentage of the converted propane
that appears as each reaction product. Hydrogen selectivities
are reported as the percentage of the hydrogen atoms in the
converted propane that appear within a given hydrocarbon
product or as H2.
>
solution by filtering. The exchange procedure was repeated
three times using fresh NH4NO3 solutions in order to ensure
complete exchange of Na cations (residual Na content < 0.07%
by elemental analysis). The catalyst samples were then dried
in flowing air at 383 K for 20 h and calcined in flowing air at
+
+
7
73 K for 20 h in order to decompose NH4 to H , which
resulted in the acidic form of the zeolite (H-ZSM5). The
elemental composition of all samples was obtained by atomic
emission spectroscopy (Galbraith Laboratories). The atomic Si/
Al ratio was 14.5 ( 0.9.
2.2. Catalytic Reaction Studies. Kinetic studies on H-ZSM5
catalysts were performed in a gradientless recirculating reactor
operated in batch mode. Chemical compositions of reaction
products were measured as a function of reaction time and
3
. Results and Discussion
.1. Catalytic Reactions of Propane on H-ZSM5. On
3
H-ZSM5, the initial products of propane conversion at 773 K
are primarily propene, H2, and equimolar amounts of ethene
and methane. Small amounts of ethane and aromatics are also
*
Corresponding author. Fax: 510-642-4778. E-mail: iglesia@
cchem.berkeley.edu.
1
0.1021/jp9824860 CCC: $15.00 © 1998 American Chemical Society
Published on Web 10/27/1998

Propane Reactions on H-ZSM5 Zeolite
J. Phys. Chem. B, Vol. 102, No. 46, 1998 9285
Figure 1. Ethene and propene selectivities during propane conversion
Figure 2. Light alkane selectivities during propane conversion on
on H-ZSM5 (773 K, 26.6 kPa C
3
H
8
, 74.7 kPa He).
3 8
H-ZSM5 (773 K, 26.6 kPa C H , 74.7 kPa He).
TABLE 1: Propane Product Distribution on H-ZSM5 (773
SCHEME 1: Bimolecular Mechanism Involving Hydride
Transfer to Adsorbed Carbenium Ions
3 8
K, 26.6 KPa C H , 74.7 KPa He, 24.9% Propane
Conversion, 5.4 ks Contact Time)
carbon
hydrogen
selectivity (%)
selectivity (%)
methane
ethane
ethene
22.4
33.6
13.3
12.6
11.8
19.2
5.5
11.8
16.8
15.8
20.4
12.8
3.4
SCHEME 2: Monomolecular Mechanism Involving
Alkane Dehydrogenation and Then Adsorption of
Alkenes on Bronsted Acid Sites
propene
C
C
4
-C
6
+
6
aromatics
benzene
1.3
toluene
6.5
2.9
2.8
1.4
+
C
8
hydrogen (H
2
)
4.0
formed initially (<5%). The initial rate of methane formation
-
4
-1
(
7.7 × 10 s ) is faster than the initial rate of propene
value (1.44).15 This indicates that double bond, cis/trans, and
skeletal isomerizations occur readily and become quasi-
equilibrated steps on H-ZSM5 even at low conversions.
Dihydrogen (H2) selectivity remains low and independent of
conversion on H-ZSM5 (4.0-6.0%), while light alkane (C2
and C4) selectivities increase with increasing propane conversion
-4 -1
formation (3.0 × 10 s ). Ethene and propene are formed
with high selectivity at low conversions and decrease with
increasing conversion as they are converted to more stable
products (e.g. aromatics, alkanes) (Figure 1). Ethene selectivity
decreases faster than propene selectivity with increasing conver-
sion. This may suggest that ethene converts to stable products
faster than propene on H-ZSM5. This is not supported,
however, by the individual turnover rates using strictly ethene
or propene feeds. On H-ZSM5, propene conversion turnover
(Figure 2). Most of the adsorbed hydrogen species formed in
dehydrogenation steps appear to transfer to carbonaceous surface
species and form C2 or C4 alkanes, rather than recombining and
desorbing into the gas phase as H . The C and C alkane
-4
-1
rates (15.0 × 10 s ) are faster than ethene conversion
2
2
4
-4
-1
selectivities increase with increasing conversion because of the
higher abundance of hydrogen atoms on the surface as the rate
of formation of aromatics increases with increasing contact time.
turnover rates (2.4 × 10 s ). The ethene selectivity decreases
faster because, at longer contact times, acid-catalyzed alkene
oligomerization/â-scission reactions control propane conversion
selectivity. These oligomerization/â-scission reactions are
constrained by the zeolite pore structure and by the thermody-
namics governing these alkene reactions. On H-ZSM5, Tabak
3
.2. Mechanism of Propane Activation Reactions on
Bronsted Acid Sites. It has been suggested that carbocations
are the reactive surface intermediates in alkane reactions
1
6,17
1
4
catalyzed by Bronsted acids.
However, these surface
et al. have shown that alkenes first polymerize on Bronsted
acid sites to form dimers and trimers, followed by isomerization,
â-scission, and further oligomerization. Thus, these alkene
oligomerization/cracking reactions on Bronsted acid sites lead
to a more rapid decrease in ethene selectivity during propane
conversion because the formation of ethene, unlike propene,
requires unstable primary carbocations.
intermediates may not be true carbocation species, but instead
alkoxide species that retain cationic character only in activated
1
8-22
complexes during chemical reactions.
Mechanistic proposals for light alkane reactions on acid
catalysts include hydride transfer from the alkane to an adsorbed
2
3-26
carbenium ion (Scheme 1),
or protonation of alkenes,
At low conversions, aromatic products, consisting of a
mixture of benzene, toluene, xylenes, and a low concentration
of larger aromatics, form slowly. The predominant aromatic
product of propane is not benzene, the expected product of c
simple dimerization steps, but toluene. Aromatic concentrations
increase with conversion because aromatics are secondary
products of propane. At higher conversions, reactions of
propane on H-ZSM5 lead to aromatics, alkenes, and alkanes
with a broad molecular weight distribution (Table 1).
formed in preceding dehydrogenation steps to form a carbenium
2
7-28
ion (Scheme 2),
or direct interaction of the alkane with a
proton to form a carbonium ion, followed by H2 desorption to
2
6
form a carbenium ion (Scheme 3).
The relative contributions of these mechanistic routes in the
initial activation of propane on H-ZSM5 can be measured using
1
3
reaction mixtures of propane and propene. Propane-2- C
1
2
13
12
( CH3- CH2- CH3) was used as a reactant to distinguish
between the propane fed initially and the propane formed from
propene hydrogenation. The propane conversion rates were
The ratio of 1-butene and isobutene to cis-2-butene and trans-
13
2
-butene on H-ZSM5 instantaneously reaches its equilibrium
calculated from those products containing C, which form only

9286 J. Phys. Chem. B, Vol. 102, No. 46, 1998
Biscardi and Iglesia
1
3
Figure 4. C content in propane, propene, benzene, toluene, and
1
3
Figure 3. Propane-2- C turnovers versus contact time during propane
conversion on H-ZSM5 with and without unlabeled propene (773 K,
13
xylenes on H-ZSM5 using propene/propane-2- C mixtures (773 K,
13
5
.6 kPa C
3
H
8
-2- C, 1.1 kPa C
H
3 6
, 94.6 kPa He).
1
3
6
.67 kPa C
3
H
8
3 6
-2- C, 0-1.07 kPa C H , balance He).
SCHEME 3: Monomolecular Mechanism Involving
Direct Protonation of the Reacting Alkane
13
from the propane-2- C in the feed. This propane/propene
reaction mixture allows the measurements of true propane
conversion rates during propene addition without contributions
by the products of propene reactions. Bimolecular reaction
1
3
pathways (Scheme 1) would lead to higher propane-2- C
activation turnover rates when propene is added to the propane-
Figure 5. Isotopomer distribution of benzene at different propane
conversions (773 K, 5.6 kPa C H -2- C, 1.1 kPa C H , 94.6 kPa He).
3 8 3 6
13
2
- C reactants; monomolecular propane activation pathways
Schemes 2 and 3) would be unaffected by the presence of
propene.
13
(
TABLE 2: Isotopomer Distribution of Benzene on H-ZSM5
1
3
(
773 K, 5.6 KPa C
3 8 3 6
H -2- C, 1.1 KPa C H , 94.6 KPa He, at
13
Turnover rates for propane-2- C conversion on H-ZSM5
1
.6% Propane Conversion)
did not increase when unlabeled propene was added (Figure
no. of 13
C
mole
3
), suggesting that, at propane partial pressures, propane forms
in benzene
fraction
pathway to benzene
2 unlabeled propenes
1 unlabeled propene + 1 labeled propane
2 labeled propanes
reactive intermediates via monomolecular routes. The propane-
1
3
-4 -1
0
1
2
0.739
0.237
0.024
2
- C turnover rate on H-ZSM5 is 6.1 × 10
s
with pure
-
4
-1
propane and only 5.8 × 10
s
when unlabeled propene is
present. These data rule out the possibility that trace amounts
of alkenes could significantly influence the rate of propane
The initial benzene, toluene, and xylene products on H-ZSM5
are predominantly unlabeled, consistent with their direct forma-
tion from the unlabeled propene in the feed (Figure 4). The
isotopic content in these products increases in parallel with that
+
conversion by providing C3 carbocations that activate propane
in hydrogen-transfer bimolecular reactions.
The remaining question is how monomolecular pathways
occur: by direct protonation of a propane molecule to form
13
in propene, because they are formed from increasingly C-
13
enriched propene. The C content in toluene and benzene is
+
+
carbonium ions (C3H9 ), which then decompose to C3H7 and
slightly lower than in propene because of the fast initial
formation of unlabeled larger alkenes, which are diluted slowly
with time as the products of the enriched propane reactant begin
+
26
H2 (or C2H5 and CH4) or by independent dehydrogenation
of alkanes on redox sites followed by gas-phase transfer of
alkenes to protonic sites where they undergo conventional acid
13
to contribute to the observed products. The C fraction is
catalysis.2
7,28
Our studies support the suggestion of propane
similar in all aromatic products, suggesting that they all form
from a common pool of reactive intermediates derived from
activation by a monomolecular route, but provide no direct
evidence for distinguishing among these mechanisms on
H-ZSM5.
13
similar precursors. Additionally, the fraction of C within each
of these aromatic products (e.g. benzene) follows a binomial
distribution as the reaction proceeds (Figure 5), which indicates
that C-C bonds are broken and re-formed many times in each
precursor species in the time required for the formation of an
3
.3. Isotopic Tracer Studies of the Initial Formation of
Propene Reactive Intermediates. The initial rate of propane
dehydrogenation and the reactive nature of the propene formed
were probed by measuring the chemical and isotopic composi-
tion of reaction products formed from reactions of unlabeled
1
3
aromatic molecule. The C distribution in the initial benzene
product is shown in Table 2.
13
13
At higher conversions (and contact times) the 13C fractions
in ethene and butene products are very similar to those in
propene and propane-2- C. The C content in reaction
products (extrapolated to zero conversion or contact time)
provides a direct measure of the fraction of the products that
formed directly from propane without requiring propene inter-
mediates. Isotopic contents in products below that in C3H8-2-
1
3
propene (Figure 6). However, the initial C content in ethene
is higher than in propene, suggesting that some of the ethene
forms initially via direct cracking of labeled propane. The initial
13
13
C (0.33) would require contributions from unlabeled propene.
C fraction in ethene suggests that 38% of the ethene forms

Propane Reactions on H-ZSM5 Zeolite
J. Phys. Chem. B, Vol. 102, No. 46, 1998 9287
1
3
13
Figure 6. C content in reaction products of propene/propane-2- C
13
3 8 3 6
mixtures on H-ZSM5 (773 K, 5.6 kPa C H -2- C, 1.1 kPa C H , 94.6
kPa He).
1
3
13
Figure 7. C distribution in benzene formed from propane-2- C on
13
TABLE 3: 13C Distribution in Ethene on H-ZSM5 at 1.6%
3 8
H-ZSM5 (773 K, 6.7 kPa C H -2- C, 94.6 kPa He).
13
Propane Conversion (773 K, 5.6 KPa C
3 8
H -2- C, 1.1 KPa
3 6
C H
, 94.6 KPa He)
no. of 13
C
mole
in ethene
fraction
percent from origin molecule
0
1
2
0.81
0.19
0.00
61.6% from unlabeled propene
38.4% from labeled propane
from the direct cracking of labeled propane (Table 3). The
initial higher 13C content in butene than in propene shows that
butene is formed from both ethene dimerization and from
13
propene via chain growth and subsequent cracking, hence a
C
content between both 13C contents in ethene and propene.
From the initial unlabeled products and the dilution of these
1
3
products as the reaction proceeds, reactions of propane-2- C
and propene indicate that propene is a reactive intermediate
during propane conversion on H-ZSM5. These data give
evidence for initial propane dehydrogenation and propane
cracking but provide limited information about the chain growth
of these alkene intermediates.
1
3
13
Figure 8. C distribution in toluene formed from propane-2- C on
13
3 8
H-ZSM5 (773 K, 6.7 kPa C H -2- C, 94.6 kPa He).
3
.4. Formation and Scission of Carbon-Carbon Bonds.
Reactions of positionally labeled alkanes can be used to study
the mechanism of cyclization reactions.2
9-31
Specifically, the
13
analysis of the content and location of C in the products of
13
propane-2- C reactions can establish the reaction pathways
required for the conversion of propene intermediates to stable
aromatic products. For example, a simple cyclization of two
13
13
propene-2- C molecules formed from propane-2- C would lead
13
to benzene-1,4- C in the absence of intramolecular scrambling.
In contrast, the formation and expected intramolecular isomer-
ization of large chains as adsorbed alkoxide or cationic
intermediates would lead to products with a random number
1
3
13
and location of 13C-atoms.
Figure 9. C distribution in xylenes formed from propane-2- C on
13
3 8
H-ZSM5 (773 K, 6.7 kPa C H -2- C, 94.6 kPa He).
Disappointingly, all information about the location and
1
3
13
content of C-atoms in the propane-2- C is lost during an
aromatization turnover on H-ZSM5. Even at low conversion,
the alkene reactants.32,33 Both chain length and methyl group
position appear to reach equilibrium values. The broad carbon
number distribution within aromatic products of propane reac-
tions on H-ZSM5 (Table 1) is also consistent with extensive
C-C bond formation and cleavage during aromatization reac-
tions. These fast reactions lead to similar average 13C content
and isotopomer distributions in all aromatic products, suggesting
that these aromatic products form from a common pool of gas-
phase and adsorbed alkenes.
13
all aromatic products of propane-2- C reactions on H-ZSM5
show a binomial distribution of 13C-atoms, instead of the single
isotopomer expected from propene dimerization and hexene
cyclization (Figures 7-9). This binomial isotopomer distribu-
tion suggests that propane aromatization on H-ZSM5 involves
rapid chain lengthening and shortening reactions and intra-
molecular methyl shifts, which break and re-form every C-C
bond many times in the time required for an aromatic turnover.
Such a turnover requires the formation of C6 or larger molecules,
a process that involves oligomerization reactions, but also a
significant number of â-scission and methyl shift events. These
oligomerization, â-scission and isomerization reactions quickly
establish an alkene mixture that is independent of the size of
The 2-butene product shows a binomial distribution of 13C
on H-ZSM5 (Figure 10). This distribution suggests that even
at very low propane conversion (<6%) most of the carbons in
2-butene have undergone extensive C-C bond breaking
and making during oligomerization/cracking cycles, thus

9288 J. Phys. Chem. B, Vol. 102, No. 46, 1998
Biscardi and Iglesia
1
3
13
Figure 13. 13C distribution in propene formed from propane-2- C on
13
Figure 10. C distribution in 2-butene formed from propane-2- C
1
3
13
on H-ZSM5 (773 K, 6.7 kPa C
3
H
8
-2- C, 94.6 kPa He).
H-ZSM5 (773 K, 6.7 kPa C
H
3 8
-2- C, 94.6 kPa He, at 5.9, 11.7, and
3
5.7% propane conversion).
1
3
13
Figure 11. C distribution in ethene formed from propane-2- C on
1
3
13
13
H-ZSM5 (773 K, 6.7 kPa C
5.7% propane conversion).
3
H
8
-2- C, 94.6 kPa He, at 5.9, 11.7, and
Figure 14. C distribution in unreacted propane-2- C on H-ZSM5
13
3
(773 K, 6.7 kPa C H -2- C, 94.6 kPa He, at 5.9, 11.7, and 35.7%
3
8
propane conversion).
larger alkenes (from propene) are present at similar propane
13
conversions. These large alkenes crack to dilute the C-ethene
13
derived from propane-2- C cracking, thus lowering the percent-
age of ethene formed from direct propane cracking at similar
1
3
propane conversion. The ethene C enrichment from Figure
9
, extrapolated to zero contact time, agrees with the fraction of
C in ethene derived from pure propane-2- C feeds.
1
3
13
Similar to the ethene distribution, ethane initially contains
1
3
mostly one C at low conversions (Figure 12). As conversion
increases, the ethane isotopic distribution closely resembles that
of ethene, suggesting that during propane conversion on
H-ZSM5, ethane is formed by direct hydrogenation of ethene
using H-atoms formed in C-H bond activation steps.
1
3
13
13
Figure 12. C distribution in ethane formed from propane-2- C on
The initial propene C distribution formed on H-ZSM5 is
1
3
H-ZSM5 (773 K, 6.7 kPa C
5.7% propane conversion).
3
H
8
-2- C, 94.6 kPa He, at 5.9, 11.7, and
13
also not binomially distributed and contains mostly one C from
3
13
the direct dehydrogenation of propane-2- C (Figure 13). As
producing a nearly binomial 2-butene 13C isotopomer distribu-
tion.
conversion increases, propene begins to acquire a binomial
13
C
distribution and forms increasingly in alkene cycles. The
At low conversions, ethene and ethane 13C distributions are
not binomial and contain mostly one 13C (Figures 11 and 12).
Initially, ethene is formed primarily from the direct cracking
propane C distribution does not change much during reaction,
13
up to about 35% conversion (Figure 14). It slowly begins to
1
3
13
acquire multiple C atoms, primarily via hydrogenation of
C
13
of propane-2- C and not from oligomerization/cracking cycles,
which would lead to a binomial 13C distribution in ethene.
However, as conversion increases, ethene approaches a binomial
scrambled propene, but, even at 35% propane conversion, the
propane still contains >87% of the singly-labeled isotopomer
suggesting that propane dehydrogenation/hydrogenation steps
are not equilibrated.
1
3
C distribution, indicating that ethene is mostly formed from
alkene cycles that break and re-form C-C bonds. Reactions
3.5. Mechanism of Propane Aromatization on H-ZSM5
Zeolites. Propane reaction pathways have been extensively
discussed in the literature primarily on the basis of observed
1
3
of propane-2- C/propene mixtures, described in section 3.3,
suggested that only a fraction of the initial ethene product came
13
5-12
from direct cracking of propane-2- C; this discrepancy arose
effects of catalyst composition on reaction rate and selectivity.
because, with the addition of propene, a higher abundance of
We have obtained more direct evidence for the sequence of steps

Propane Reactions on H-ZSM5 Zeolite
J. Phys. Chem. B, Vol. 102, No. 46, 1998 9289
SCHEME 4. Reaction Network for Propane Conversion
on H-ZSM5
Along with aromatics, a small amount of hydrogen (H2) is
also formed as a result of the recombinative desorption of
hydrogen adsorbed on the zeolite surface. However, on
H-ZSM5, the major hydrogen disposal pathway is by hydrogen-
transfer reactions, which lead to the use of carbonaceous
intermediates as stoichiometric sinks for the removal of the
hydrogen species formed in sequential dehydrogenation steps.
Acknowledgment. This work was supported by the National
Science Foundation (Grant CTS-96-13632) and by University
of California startup funds. We acknowledge the technical
assistance of Mr. Joseph E. Baumgartner (Exxon Research and
Engineering Co.) in the design of the microreactor and in the
development of protocols for chemical and isotopic analyses
of reaction products.
using isotopic tracer and exchange methods. These methods
have established the nature of individual reaction steps in this
reaction sequence.
The reaction network shown in Scheme 4 is consistent with
our results and with previous studies of acid-catalyzed reac-
References and Notes
5
,6
tions.
Propene, initially formed via a dehydrogenation of
(
1) Chen, N. Y.; Garwood, W. E. J. Catal. 1978, 52, 453.
propane, and ethene, initially formed from propane cracking,
undergo rapid acid-catalyzed chain growth reactions within
(2) Anderson, J. R.; Foger, K.; Mole, T.; Rajadhyaksha, R. A.; Sanders,
J. V. J. Catal. 1979, 58, 114.
+
(3) Iglesia, E.; Baumgartner, J. E.; Price, G. L. J. Catal. 1992, 134,
49.
zeolite channels by oligomerization or by addition to C3
5
carbocations formed directly from propane. These surface
species undergo rapid methyl shift isomerization reactions; as
the chain lengthens, â-scission reactions become faster and
ultimately establish a pool of carbocation species and alkene
intermediates with a broad size distribution. These alkene
intermediates continue to undergo hydrogenation-dehydroge-
nation reactions, commonly described as hydrogen transfer, in
(
4) Iglesia, E.; Baumgartner, J. E.; Meitzner, G. D. In Proceedings of
the 10th International Congress on Catalysis; Guczi, L., Solymosi, F.,
Tetenyi, P., Eds.; Elsevier: Budapest, 1992.
(5) Guisnet, M.; Gnep, N. S.; Aittaleb, D.; Doyemet, J. Y. Appl. Catal.
1
992, 87, 255.
(
6) Guisnet, M.; Gnep, N. S.; Alario, F. Appl. Catal. 1992, 89, 1.
(7) Giannetto, G.; Monque, R.; Galiasso, R. Catal. ReV. 1994, 36, 271.
(
8) Kitagawa, H.; Sendoda, Y.; Ono, Y. J. Catal. 1986, 101, 12.
9) Yashima, T.; Sasaki, T.; Takahashi, K.; Watanabe, S.; Namba, S.
(
+
which light alkenes accept hydrogen to form alkanes. C6
Sekiyu Gakkaishi 1988, 31, 154.
alkenes tend to cyclize and further dehydrogenate to kinetically
stable and thermodynamically favored aromatic products.
Cracking of larger chains also occurs with the concomitant
removal of hydrogen by its addition to light alkene products of
cracking reactions. Propane aromatization on H-ZSM5 occurs
via a complex sequence of dehydrogenation, oligomerization,
â-scission, and cyclization steps (Scheme 4).
(10) Ono, Y.; Kanae, K. J. Chem. Soc.-Faraday Trans. 1991, 87, 669.
(
11) Kwak, B. S.; Sachtler, W. M.; Haag, W. O. J. Catal. 1994, 149,
65.
12) Engelen, C.; Wolthuizen, J. P.; van Hooff, J. H. Appl. Catal. 1985,
19, 153.
(13) Price, G. L.; Iglesia, E. Ind. Eng. Chem. Res. 1989, 28, 839.
14) Tabak, S. A.; Krambeck, F. J.; Garwood, W. E. AIChe J. 1986,
2, 1526.
15) Stull, D. R.; Westrum, E. F.; Sinke, G. C. The Chemical
4
(
(
3
(
Thermodynamics of Organic Compounds; Robert E. Keieger Publishing
Co.: Malabar, FL, 1987.
4
. Conclusions
(16) Thomas, C. L. Ind. Eng. Chem. 1949, 41, 2564.
(
17) Greensfelder, B. S.; Voge, H.; Good, G. M. Ind. Eng. Chem. 1949,
The results from the isotopic 13C labeling studies provide
41, 2573.
(18) Kazansky, V. B.; Senchenya, I. N. J. Catal. 1989, 119, 108.
(19) Kazansky, V. B.; Senchenya, I. N. J. Mol. Catal. 1992, 74, 257.
20) Kazansky, V. B.; Senchenya, I. N.; Frash, M.; van Santen, R. A.
additional evidence for the propane reaction pathways suggested
by previous researchers.5
-7
These isotopic results suggest that
(
propane aromatization occurs on H-ZSM5 by a complex
reaction sequence, with a first step consisting of propane
dehydrogenation to propene or propane cracking to ethene and
Catal. Lett. 1994, 27, 345.
(21) Kazansky, V. B.; Frash, M. V.; van Santen, R. A. Catal. Lett. 1994,
8, 211.
2
9
4
1
(22) Lercher, J. A.; van Santen, R. A.; Vinek, H. Catal. Lett. 1994, 27,
1
3
methane (confirmed by reactions of propene/propane-2- C
1.
(23) Abbot, J.; Wojciechowski, B. W. Can. J. Chem. Eng. 1985, 63,
62.
24) Corma, A.; Planelles, J.; Sanchez-Marin, J.; Tomas, F. J. Catal.
985, 93, 30.
25) Gates, B. C.; Katzer, J. R.; Schuit, G. C. Chemistry of Catalytic
Processes; McGraw-Hill: New York, 1979.
26) Haag, W. O.; Dessau, R. In Proceedings of the 8th International
Congress on Catalysis; Dechema: Berlin, 1984; Vol. 305.
27) Mills, G.; Heinemann, H.; Miliken, T.; Oblad, A. Ind. Eng. Chem.
13
mixtures). Reactions of propane-2- C show that chain growth
on H-ZSM5 involves rapid alkene oligomerization/cracking
cycles. Stable products, such as aromatics, alkanes, and H2,
exit this cycle via cyclization reactions and hydrogen transfer
or desorption pathways (Scheme 4). These alkene reactions
are necessary to explain the distribution of aromatic products.
(
(
(
13
Products of propane-2- C reactions on H-ZSM5 contain
(
13
binomial distributions of C-atoms. Benzene products contain
1953, 45, 134.
one to five 3C-atoms, instead of the two C-atoms expected
1
13
(28) Sommer, J.; Habermacher, D.; Hachoumy, M.; Jost, R.; Reynaud,
A. Appl. Catal. 1996, 146, 193.
from direct propene cyclodimerization. Reactions involve rapid
oligomerization/cracking cycles of intermediate alkenes; such
reactions break and re-form every C-C bond many times during
one aromatization turnover. Rapid oligomerization/cracking
(29) Iglesia, E.; Baumgartner, J. E.; Price, G. L.; Robbins, J. L.; Rose,
K. D. J. Catal. 1990, 125, 95.
(30) Davis, B. H.; Venuto, P. B. J. Org. Chem. 1971, 36, 337.
(
31) Price, G. L.; Egedy, G. R. J. Catal. 1983, 84, 461.
(32) Garwood, W. E. Prepr.sAm. Chem. Soc., DiV. Pet. Chem. 1982,
13
cycles and intervening methyl shifts lead to the same C content
2
7, 563.
and to binomial 13C distributions in all aromatic products.
(33) Garwood, W. E. ACS Symp. Ser. 1983, 218, 383.