Comparative Study of the Catalytic Properties of ZSM-22 and ZSM-35/ Ferrierite Zeolites in the Skeletal Isomerization of 1-Butene

Article abstract of DOI:10.1006/jcat.1998.2174

The catalytic activities of the two catalysts ZSM-22 and ZSM-35 were compared in the skeletal isomerization of 1-butene. ZSM-22 demonstrated higher activity in 1-butene transformation, compared to that of ZSM-35. ZSM-35 was not selective to isobutene until nonselective acid sites had been poisoned by coke deposits, while ZSM-22 was selective already from the beginning. Information about the extent of coke formation was obtained by FTIR experiments and surface area measurements. ZSM-22 was more resistant towards coke formation, compared to that of ZSM-35. In order to obtain information about the reaction mechanism, 2,4,4-trimethyl-2-pentene and 1-octene were cracked over the catalysts. The selective mechanism for isobutene formation in the skeletal isomerization of 1-butene was most likely monomolecular. The bimolecular mechanism is not selective to isobutene, although it can contribute to the overall isobutene production.

Full text of DOI:10.1006/jcat.1998.2174

JOURNAL OF CATALYSIS 178, 611–620 (1998)
ARTICLE NO. CA982174
Comparative Study of the Catalytic Properties of ZSM-22 and ZSM-35/
Ferrierite Zeolites in the Skeletal Isomerization of 1-Butene
1
R. Byggningsbacka, N. Kumar, and L.-E. Lindfors
Laboratory of Industrial Chemistry, Faculty of Chemical Engineering, A˚ bo Akademi University, Biskopsgatan 8, FIN-20500 A˚ bo, Finland
Received December 15, 1997; revised May 26, 1998; accepted June 1, 1998
Production of isobutene is thermodynamically favored at
The catalytic activities of the two catalysts ZSM-22 and ZSM-35 low temperatures, as can be concluded from the thermody-
were compared in the skeletal isomerization of 1-butene. ZSM-22 namic equilibrium calculations presented in Fig. 1. These
demonstrated higher activity in 1-butene transformation, compared
to that of ZSM-35. ZSM-35 was not selective to isobutene until non-
selective acid sites had been poisoned by coke deposits, while ZSM-
calculations were made using thermodynamic data from
Ref. (15). In order to be able to use lower reaction tempera-
tures, improvements in the shape selectivity of the catalysts
are needed since dimerization of n-butene is also favored
at lower temperatures.
2
2 was selective already from the beginning. Information about
the extent of coke formation was obtained by FTIR experiments
and surface area measurements. ZSM-22 was more resistant to-
wards coke formation, compared to that of ZSM-35. In order to ob-
tain information about the reaction mechanism, 2,4,4-trimethyl-2-
pentene and 1-octene were cracked over the catalysts. The selective
In this paper we compare ZSM-22 and ZSM-35, two of
the most promising catalysts for skeletal isomerization of
1
-butene according to patents (16–18) and articles (1–2,
mechanism for isobutene formation in the skeletal isomerization of 5–7, 19–26). Although both ZSM-22 and ZSM-35 have been
-butene was most likely monomolecular. The bimolecular mecha- thoroughly explored for skeletal isomerization it is difficult
nism is not selective to isobutene, although it can contribute to the to directly compare the catalytic properties of ZSM-22 and
1
overall isobutene production.
�c 1998 Academic Press
ZSM-35 from previous publications, since different reac-
tion conditions have been used.
ZSM-35 has a one-dimensional channel system of 10-
membered-rings (4.2 � 5.4 A˚ ) and a one-dimensional chan-
nel system of eight-membered-rings (3.5 � 4.8 A˚ ). The two
1
. INTRODUCTION
There has been an ever increasing interest in the skeletal
isomerization of n-butenes over zeolites: from the indus-
trial point of view, since isobutene can be used in the re-
action with methanol to produce MTBE (methyl tert-butyl
ether), and from the academic point of view as an inten-
sive discussion has been going on whether the reaction pro-
ceeds through a monomolecular or a bimolecular reaction
mechanism (1–5). The main problem with the monomolec-
ular reaction mechanism is that this mechanism would re-
quire the formation of a highly energetically unfavourable
primary carbenium ion intermediate. Despite this, there
have recently been several papers regarding the reaction
mechanism where the authors have been in favor of the
monomolecular mechanism (3–5).
kinds of channel systems are perpendicularly intersected
and therefore ZSM-35 contains spherical cavities with a size
of about 6–7 A˚ (27). ZSM-22 consists of a one-dimensional,
1
4
2
0-membered-ring pore system with channel diameters of
.5 � 5.5 A˚ (28). Although the zeolite structures of ZSM-
2 and ZSM-35 are rather different, the difference in the
shape-selective properties might not be very large since it
is difficult for butene molecules to enter the eight-member-
ring channel system in ZSM-35 and its spherical cavities can
be modified by the coke deposits. In fact, similar selectivity
to isobutene was obtained in the isomerization experiments
over deactivated ZSM-22 and ZSM-35.
2
. EXPERIMENTAL
Although different types of zeolites have been inves-
tigated in the skeletal isomerization of 1-butene (6–13),
there is still room for improvement, both in the activity of
the catalysts, as well as in the selectivity to isobutene. The
2
.1. Preparation of the Catalysts
The ZSM-22 zeolite was synthesized using 1,6-diamino-
most promisingcatalystsfor selective skeletalisomerization hexane (Fluka) as the organic template, Ludox AS-40
of 1-butene are 10-membered-ring molecular sieves (14). (Aldrich) as the silica source, Al2(SO4)3 � 18H2O (Merck)
as the aluminum source, and KOH (Merck) as the mineral-
1
izing agent according to the method used in Ref. (29). The
ZSM-35 zeolite was synthesized using Pyrrolidine (Fluka)
To whom correspondence should be addressed. E-mail: Rune.
Byggningsbacka@abo.fi.
611
0021-9517/98 $25.00
Copyright �c 1998 by Academic Press
All rights of reproduction in any form reserved.

6
12
BYGGNINGSBACKA, KUMAR, AND LINDFORS
Acid sites of the zeolites were characterized by a
FTIR spectrometer (ATI Mattson infinity spectrometer)
equipped with a Mercury Cadmium Telluride (MCT) detec-
tor. The in situ diffusion reflectance infrared Fourier trans-
formed (DRIFT) measurements of adsorbed pyridine on a
1
/1 (weight ratio) mixture of zeolite and KBr powder were
performed in a diffuse reflectance accessory equipped with
a standard controlled environmental chamber (Spectra-
Tech, model 0030-103). The zeolites were evacuated in vac-
�
5
�
uum (10 atm) at 500 C for 2 h before pyridine adsorption
�
�
at 100 C for 30 min. The zeolite samples were kept at 100 C
in a flow of helium for 1 h to allow the pyridine to penetrate
the samples. The zeolites were evacuated at 200 C for 2 h
�
before the FTIR spectra were recorded at ambient temper-
ature. Acid sites in the fresh and deactivated as well as the
ion-exchanged (proton form) zeolites were characterized
by means of FTIR spectroscopy.
FIG. 1. The calculated thermodynamic distribution of butene isomers
as a function of temperature at equilibrium.
Surface areas of the zeolites were determined by nitrogen
adsorption (Carlo Erba Instruments). Zeolite samples of
as the organic template, Ludox AS-30 (Aldrich) as the silica
source, NaAlO2 (Riedel-de Ha e¨ n) as the aluminum source,
and NaOH (Merck) as the mineralizing agent using the
method described in Ref. (30) with some modification.
Before calcination the zeolite powder was washed, dried,
pressed, crushed, and finally sieved to 0.125–0.250 mm par-
ticle size. Calcination in order to remove the organic tem-
�
0
.2 g were evacuated at 300 C for 3 h before the surface area
measurements. The Dubinin method was used in the calcu-
lation of the surface areas. Surface areas of fresh and deac-
tivated catalysts were measured in order to obtain informa-
tion about the extent of pore blocking by the coke deposits.
�
2.3. Catalyst Testing
plate was performed at 550 C in a flow of nitrogen for 6 h,
followed by air for an additional 10 h. ZSM-22 and ZSM-35
The catalytic activities of the synthesized zeolites in
zeolites were ion-exchanged for 48 h in a 1 M NH4Cl solu- the isomerization of 1-butene (99.0% purity, AGA) to
tion in order to remove any potassium or sodium remaining isobutene were studied in a micro fixed-bed reactor sys-
from the synthesis. The ion-exchanged zeoliteswere washed tem at near atmospheric pressure (1.1 atm). The amount of
�
free of chloride ions and dried for 12 h at 80 C. The proton catalyst was varied between 0.05 and 0.5 g and the flow
+
forms of ZSM-22 and ZSM-35 were obtained after NH4 of 1-butene between 21 and 55 ml/min in order to ob-
ions had been decomposed in a calcination step performed tain the different weight hourly space velocities (WHSV)
in the same conditions as the removal of the organic tem- used in the experiments. The reactant was diluted with a
plate. The calcination steps were performed in high flows 0–55 ml/min flow of nitrogen (99.999% purity, AGA) to ob-
(
200 ml/min per gram of catalyst) in order to prevent dea- tain different partial pressures of 1-butene. The products
lumination of the zeolite structures, since ZSM-22 zeolites from the reactor were analyzed by a gas chromatograph
calcined under limited heat and mass transfer conditions (Varian 3700) equipped with a flame-ionization detector
have been found to exhibit low selectivity to isobutene (26). (FID). A capillary column (50 m � 0.32 mm ID fused-silica
PLOT Al2O3-KCL) was used for product separation.
2
.2. Characterization of the Zeolites
In the calculations the three n-butene isomers 1-butene,
The structure and phase purity of the synthesized zeo- cis-2-butene, and trans-2-butene were considered reac-
lites were determined by X-ray powder diffraction (XRD) tants, thus conversion, yield, and selectivity were defined
using Cu K� radiation. The XRD patterns of the synthe- as follows:
sized zeolites were similar to those previously reported in
the literature for ZSM-22 (6) and ZSM-35 (11). The bulk
(
1-butene)feed � (n-butene)effluent
(1-butene)feed
Si/Al ratio of the zeolites was measured by X-ray fluores-
cency (XRF) and found to be 53 for the ZSM-22 zeolite
and 12 for the ZSM-35 zeolite. The crystal size of the ze-
olites was determined using scanning electron microscopy
Conversion =
Yield =
(
isobutene)effluent
(1-butene)feed
(
SEM). The average width of the needle shaped ZSM-22
crystals was found to be 0.4 �m and the length was 2.9 �m.
The corresponding size of the rod shaped ZSM-35 crystals
was approximately 1.6 �m and 5.2 �m.
(
isobutene)effluent
Selectivity =
.
(1-butene)feed � (n-butene)effluent

SKELETAL ISOMERIZATION OF 1-BUTENE
. RESULTS AND DISCUSSION
613
3
TABLE 1
Differences between Proton and As-Synthesized ZSM-22
and ZSM-35 Zeolites
3
.1. Effect of Ion-Exchange on the Catalytic
Properties of the Zeolites
Yield of
Selectivity to
isobutene
(mol% )
Acid sites were characterized using FTIR of adsorbed
pyridine. Spectra recorded over the as-synthesized and pro-
ton ion-exchanged zeolites are compared in Fig. 2. The band
Ion-
WHSV Conversion isobutene
� 1
Catalyst exchange
(h
)
(mol% )
(mol% )
�
1
ZSM-22
ZSM-22
ZSM-35
ZSM-35
No
Yes
No
150
150
10
41.8
40.4
55.8
53.5
33.0
31.7
28.5
25.6
79.0
78.3
51.0
49.7
at 1548 cm which is characteristic for pyridine adsorbed
�
1
on Brønsted acid sites and the band at 1453 cm associated
with pyridine adsorbed on Lewis acid sites were present in
the ion-exchanged zeolites as well as in the as-synthesized
zeolites. The band at 1491 cm is usually associated with
pyridine adsorbed on both Lewis and Brønsted acid sites.
These experiments demonstrated that Brønsted acid sites
were present in the as-synthesized zeolites even before any
ion-exchange procedures had been performed.
Yes
10
�
�
1
Note. Reaction conditions: temperature = 400 C; partial pressure of
-butene = 0.5 atm; TOS = 10 min.
1
iments. An indication that the external diffusion was neg-
ligible was obtained from the double-bond isomerization,
which isfaster than the skeletalisomerization, dimerization,
or cracking reactions. The ratios between 1-butene, cis-2-
butene, and trans-2-butene were approximately the same in
all the experiments and fairly close to the calculated equilib-
rium values, although the ratio between trans-2-butene and
cis-2-butene exceeded the value expected from the thermo-
dynamic equilibrium calculations.
The reason for the lower activity of ZSM-35, compared
to that of ZSM-22 in 1-butene isomerization could also be
because of internal pore diffusion limitations. According to
the Si/Al ratio ZSM-35 could be expected to have higher
activity. If the internal pore diffusion is slow the crystal
size might have a profound effect on the catalytic activ-
ity since most of the 1-butene transformation would occur
close to the external surface. If internal pore diffusion is
the rate-limiting step the activity would improve with de-
creasing crystal size. This might explain why in our inves-
tigation ZSM-22 was more active in skeletal isomerization
of 1-butene, since the average crystal size of ZSM-22 was
smaller, compared to that of ZSM-35. For comparison we
also made some experiments with the less reactive molecule
n-butane. In these experiments there was in fact an im-
provement in the catalytic activity after ion-exchanging the
zeolites, and ZSM-35 was more active, compared to that of
ZSM-22.
The as-synthesized and proton ion-exchanged zeolites
were compared in the transformation of 1-butene. The re-
sultsfrom these experimentsare presented in Table 1. These
experiments indicate that there was no improvement in the
activity of the catalysts in 1-butene isomerization after ion-
+
exchanging the zeolites with NH4 and subsequent calcina-
tion. The reason for this might be that the reaction was dif-
fusion limited since 1-butene is a very reactive molecule and
the pores in the zeolites are of molecular size. Even though
the number of acid sites increased after ion-exchanging the
zeolites, the activity was unchanged since the rate limit-
ing step was the transportation of the molecules to the ac-
tive sites. Internal pore diffusion was most likely the reason
for the diffusion resistance since the external film diffusion
was negligible because of the high flows used in the exper-
Since there was no improvement in the activity in skeletal
isomerization of 1-butene after ion-exchanging the zeolites,
the as-synthesized zeolites were used in all of the following
experiments.
3
.2. Effect of Time on Stream (TOS) on the Activity
and Selectivity of ZSM-22 and ZSM-35
In the 1-butene skeletal isomerization experiments the
catalysts are modified by coke deposits. The surface ar-
�
eas of fresh catalysts and catalysts deactivated at 400 C for
2
FIG. 2. FTIR spectra of adsorbed pyridine on (A) as synthesized
ZSM-22, (B) ion-exchanged ZSM-22, (C) as-synthesized ZSM-35, and
0 h were measured in order to obtain information about
(
D) ion-exchanged ZSM-35.
the extent of pore blocking by the coke deposits. These

6
14
BYGGNINGSBACKA, KUMAR, AND LINDFORS
TABLE 2
Differences between Fresh and Deactivated Catalysts
Surface
area
(m /g)
Yield of
isobutene
(mol% )
Selectivity to
isobutene
(mol% )
TOS
(min)
Conversion
(mol% )
2
Catalyst
ZSM-22
ZSM-22
ZSM-35
ZSM-35
1
1200
1
255.7
92.2
400.6
39.6
49.4
29.4
16.7
0.0
36.9
25.7
8.8
74.7
87.3
52.6
—
1200
0.0
�
� 1
Note. Reaction conditions: temperature = 400 C; WHSV = 150 h
partial pressure of 1-butene = 0.5 atm.
;
surface area measurements demonstrated that there was
a substantial decrease in the surface area between fresh
and deactivated catalysts. The results from the surface area
FIG. 4. Conversion (ᮀ), yield of isobutene (o), and selectivity to
experiments are presented in Table 2, together with con- isobutene (x) over ZSM-22 as a function of TOS. Reaction conditions:
�
temperature = 400 C; partial pressure of 1-butene = 0.5 atm; WHSV =
version, yield of isobutene, and selectivity to isobutene for
both fresh and deactivated catalysts. Deactivated ZSM-35
catalysts did not show any activity at all when the WHSV of
�
1
1
50 h .
�
1
1
-butene was 150 h , but it was active at lower WHSV. The
The main difference between ZSM-22 and ZSM-35 in the
decrease in the surface area was higher for ZSM-35, com- 1-butene isomerization experiments was observed in the se-
pared to ZSM-22. The reason for this might be that coke lectivity to isobutene as a function of time on stream (TOS).
deposits are formed more easily inside the spherical cavi- Results from deactivation experiments over ZSM-22 and
ties of ZSM-35 than in the one-dimensional channel system ZSM-35 zeolites are presented in Figs. 4 and 5, respectively.
of ZSM-22.
ZSM-35 demonstrated high selectivity to isobutene after
FTIR spectra of adsorbed pyridine over the fresh and de- the nonselective acid sites had been poisoned by coke de-
activated are compared in Fig. 3. As can be seen the number posits, while ZSM-22 showed high selectivity already from
of acid sites accessible for pyridine decreased substantially the beginning. It can be concluded from the figures that
for the deactivated catalysts, compared to the fresh cata- the selectivity to isobutene over deactivated ZSM-22 and
lysts. The activity of the deactivated catalysts was found to ZSM-35 obtained the same value, but ZSM-22 was more
be restored after burning off the coke deposits.
active, compared to ZSM-35 since the contact time needed
in order to acquire the same level of conversion and yield of
FIG. 5. Conversion (ᮀ), yield of isobutene (o), and selectivity to
isobutene (x) over ZSM-35 as a function of TOS. Reaction conditions:
�
FIG. 3. FTIR spectra of adsorbed pyridine on (A) fresh ZSM-22, temperature = 400 C; partial pressure of 1-butene = 0.5 atm; WHSV =
�
1
(
B) deactivated ZSM-22, (C) fresh ZSM-35, and (D) deactivated ZSM-35.
10 h .

SKELETAL ISOMERIZATION OF 1-BUTENE
615
TABLE 3
coke deposits. It was found that the selectivity to isobutene
increased mainly during the first hours of TOS. The selec-
tivity increased from 53 to 73% over ZSM-35 during the
first 2 h when the conversion was kept constant by changing
the WHSV, but only by 7% during the next 14 h. There
was also an improvement in the selectivity to isobutene
over ZSM-22 because of coke deposits, but the selectivity
increased only by 2% during 20 h of TOS. Active sites on
the outer surface of the zeolite might be poisoned, making
the zeolite more shape-selective. Carbon deposits might
poison preferentially stronger Brønsted acid sites known
Effect of TOS on the Product Selectivity over ZSM-22
and ZSM-35 Zeolites
Catalyst
ZSM-22
ZSM-35
TOS (min)
10
150
660
60
10
20
970
3
WHSV (h� 1)
Selectivity (mol% )
Ethene
Propane
Propene
Isobutane
n-Butane
Isobutene
Pentenes
Hexenes
Heptenes
Octenes
0.5
0.2
9.0
0.3
2.6
79.0
5.2
0.4
0.3
2.5
0.2
0.1
6.9
0.2
2.6
80.1
4.5
0.5
0.5
4.3
3.3
1.2
22.2
0.6
0.3
0.1
7.6
0.5 to induce dimerization reactions. Carbonaceous deposits
3.3
0.8
79.1
6.7
0.4
0.4
can fill irregularities in the zeolite lattice, modifying the
53.2
10.2
1.3
1.3
3.0
space inside the zeolite pores. In fact, coke deposits inside
the cavities of ZSM-35 were most likely needed in order
to receive high selectivity to isobutene. On the other hand,
4.1 since ZSM-22 does not contain any cavities because of the
Yield of isobutene
Conversion
33.0
41.8
34.4
43.0
21.6
40.6
33.4
42.2
one-dimensional channel system, the modifications by the
coke deposits were not as dramatic as for ZSM-35. The
difference in selectivity between ZSM-22 and ZSM-35
has previously been compared by Mooiweer et al. (19)
and, according to their results, ZSM-35 demonstrated a
�
Note. Reaction conditions: temperature = 400 C; partial pressure of
-butene = 0.5 atm.
1
isobutene was only 1/15 of that needed for ZSM-35. Several much higher selectivity to isobutene than ZSM-22. The
papers have been published about the positive effect of discrepancies between this paper and Ref. (19) might be in
coke on the selectivity to isobutene over ZSM-35 (2, 20– the quality of the ZSM-22 catalyst sample and the reaction
2
2). When the effect of deactivation is evaluated, similar conditions used. Since shape-selectivity is required in
conversion should be used since the selectivity to isobutene order to obtain high selectivity to isobutene in skeletal
also improved when the WHSV of 1-butene was increased isomerization of 1-butene the quality of the zeolite sample
and conversion decreased. Even if fresh and deactivated is of great importance. The quality of the zeolite sample
ZSM-35 catalysts are evaluated at similar conversions, as depends upon synthesis parameters, type and purity of
in Table 3, it is evident that there is a substantial increase reactants, and postsynthesis treatments.
in the selectivity to isobutene because of changes in the
The reaction scheme in Fig. 6 is proposed in order to ex-
shape-selective properties of the zeolite caused by the plain the product distribution in Table 3. Since the ratio
FIG. 6. Proposed reaction scheme to explain the obtained products.

6
16
BYGGNINGSBACKA, KUMAR, AND LINDFORS
between propene and pentene was not one, as expected high selectivity to isobutene received after the very short
from the direct cracking of octenes, the formation of TOS over zeolites with one-dimensional channels systems
=
nonenes (C ) are included in Fig. 6 in order to explain the such as ZSM-22.
9
excess formation of propene. Another explanation could
The other proposed reaction mechanisms are bimolecu-
be that propene was less reactive in the coke formation. As lar. The dimerization, isomerization, and cracking scheme
the catalysts became deactivated the ratio between propene for the butene isomers is given in Fig. 8. The rate of A
and pentene became closer to one because of the improve- isomerization (methyl shift) is faster than B isomeriza-
ments in the shape-selective properties of the zeolites. Sat- tion (chain branching), while the rate of the dimerization/
urated hydrocarbons were produced mainly because of the cracking steps is decreasing as follows: A cracking (tertiary
hydrogen transfer reactions in coke formation. Because the carbenium ions) > B cracking (secondary and tertiary car-
rate of coke formation over ZSM-35 was lower for the deac- benium ions) > C cracking (secondary carbenium ions). If
tivated catalyst, compared to the fresh catalyst, the selectiv- isobutene was only produced through a bimolecular re-
ity to propane, isobutane, and n-butane reduced with TOS. action mechanism, the problem would be to explain the
Since the rate of coke formation was higher on fresh cata- high isobutene selectivity observed in the skeletal isomer-
lysts, deactivated catalysts were used in all of the following ization of 1-butene since the main by-products propene
experiments.
and pentene are also produced through bimolecular reac-
tions.
In order to discriminate between the monomolecular and
bimolecular mechanism, an attempt was made to obtain in-
formation about the selectivity of the bimolecular mecha-
3
.3. Cracking of 2,4,4-Trimethyl-2-Pentene (TMP)
and 1-Octene over ZSM-22 and ZSM-35
Two different mechanisms, a monomolecular and a bi- nisms through cracking of 2,4,4-trimethyl-2-pentene (TMP)
molecular, have been proposed in the literature for the and 1-octene. Resultsfrom these experimentsare presented
formation of isobutene (1–5). The different steps in the in Table 4.
monomolecular mechanism are given in Fig. 7. This mech-
1-Octene is not directly produced in dimerization of n-
anism requires a highly energetically and ther-modyna- butene since this would require a primary carbenium ion
mically unfavourable ring opening to form a primary car- and therefore be substantially slower than the A, B, and
benium ion as an intermediate and has therefore been C dimerization steps in Fig. 8. Since direct cracking of 1-
considered to be too slow in order to explain the high octene also would require a primary carbenium ion, most
rate of isobutene formation in skeletal isomerization of n- of the cracking occurred after the skeletal isomerization
butene. On the other hand, the monomolecular reaction of 1-octene. Consequently only small amounts of ethene
mechanism could explain the high selectivity to isobutene, and hexene were observed in the 1-octene cracking ex-
if bimolecular reactions were inhibited by the shape-selec- periments. Because of the skeletal isomerization, most of
tive properties of the zeolites. A “pseudo-monomolecular” the different octene isomers were detected in the product
mechanism has also been proposed in the literature (2). stream, in addition to butene, propene, and pentene from
According to this mechanism the active sites would be the the cracking reactions. Cracking of 1-octene was not very
coke, instead of the Brønsted acid sites and would not selective to isobutene. Less than half the products from the
therefore require a primary carbenium ion. The problem cracking reactions were butene isomers. The received prod-
with “pseudo-monomolecular” mechanism is to explain the uct distribution is close to that expected from the cracking
FIG. 7. The monomolecular reaction scheme in skeletal isomerization of 1-butene.

SKELETAL ISOMERIZATION OF 1-BUTENE
617
FIG. 8. The dimerization, isomerization, and cracking scheme for the butene isomers.

6
18
BYGGNINGSBACKA, KUMAR, AND LINDFORS
TABLE 4
Selectivity to Different Products when 1-Octene and 2,4,4-
Trimethyl-2-Pentene Were Cracked over Deactivated ZSM-22 and
ZSM-35
tal isomerization of 1-butene. In dimerization of n-butene,
TMP would most likely be produced in a reaction between
isobutene and trans-2-butene or cis-2-butene as described
in Fig. 9.
Because the double-bond isomerization is fast, the equi-
Reactant
Catalyst
TMP
1-Octene
librium between 1-butene, trans-2-butene, and cis-2-butene
�
ZSM-22
ZSM-35
ZSM-22
ZSM-35 is established rapidly and at 400 C only 30% of n-butene is
left as 1-butene. Consequently, the majority of the dimer-
ization reactions between isobutene (as a tertiary carbe-
nium ion) and n-butene would proceed according to the
Selectivity (mol% )
Propene
trans-2-Butene
2.9
9.4
5.2
72.5
8.0
2.0
1.3
4.3
2.4
87.2
3.7
1.1
27.3
11.1
6.6
18.2
8.1
27.6
10.2
6.1
21.4
7.0
mechanism in Fig. 9. In order to receive high selectivity to
TMP, the B1 dimerization in Fig. 8 would also have to be
faster than B2 dimerization. If this were the case, isobutene
would be produced mainly through cracking of TMP while
the by-products, propene and pentenes, would be produced
in dimerization between isobutene and 1-butene and sub-
sequent cracking of the dimers. TMP molecules have been
considered to be too large to be formed inside channels
of 10-membered-ring zeolites (4). On the other hand, the
cracking rate of TMP was much higher, compared to 1-
octene. This indicates than TMP can reach the active sites
in the zeolites.
1
-Butene
Isobutene
cis-2-Butene
Pentenes
27.2
27.0
�
Note. Temperature = 400 C.
scheme in Fig. 8, if cracking takes place before isomer-
ization to TMP molecules. The number of cracking re-
actions producing propene and pentene would then be
higher than the cracking reactions producing butene iso-
mers. From the cracking of 1-octene it is possible to exclude
that the selective mechanism in skeletal isomerization of
Although it might be difficult to explain a selectivity of
over 90% with a bimolecular mechanism, this mechanism
could still contribute to the overall isobutene production.
1
-butene to isobutene would be: n-butene + n-butene →
octenes → isobutene + n-butene. According to the 1-oct-
ene cracking experiments, the highest expected selectiv-
ity to isobutene for this bimolecular reaction mechanism
would be approximately 27% , while on the other hand, over
3
.4. Effect of Temperature, WHSV of 1-Butene,
and Partial Pressure of 1-Butene
9
0% selectivity was observed in the skeletal isomerization
of 1-butene. The cracking rate of 1-octene was also too slow
The relative reaction rate between a monomolecular and
to explain the high production rate of isobutene in skeletal a bimolecular mechanism should also depend on the re-
isomerization of 1-butene.
action conditions. The changes in the product selectivity
Cracking of TMP produced mostly isobutene, since A over ZSM-35 as a function of WHSV, partial pressures of
cracking was faster than both B2 cracking and B isomeri- 1-butene, and temperature are presented in Table 5. The
zation. Isobutene exceeds the value predicted by the ther- changes in the product distribution over ZSM-22 as a result
modynamic calculations because isobutene (A cracking) of WHSV, partial pressure of 1-butene, and temperature
was kinetically favored over the other butene isomers were similar to those of ZSM-35.
(
B and C cracking) and the equilibrium was approached
Results from 1-butene isomerization experiments over
from the opposite side, compared to monomolecular skele- ZSM-35 using different WHSV are presented in Fig. 10.
FIG. 9. Reaction mechanism for TMP formation and subsequent cracking to isobutene.

SKELETAL ISOMERIZATION OF 1-BUTENE
619
TABLE 5
Effect of WHSV, Partial Pressure of 1-Butene, and Temperature
on the Product Distribution over Deactivated ZSM-35
WHSV
30
0.5
400
Partial pressure
Temperature
WHSV (h 1)
�
3
5
10
1.0
400
3
10
0.5
500
Pressure (atm)
Temperature ( C)
0.5
400
0.5
0.5
�
400
350
Selectivity (mol% )
Ethene
Propane
0.3
0.1
7.6
0.0
0.1
2.3
0.2
0.1
6.0
0.3
0.1
8.8
0.1
0.1
5.8
0.4
0.1
2.8
Propene
Isobutane
n-Butane
0.5
0.8
0.2
0.4
0.4
0.7
0.6
1.3
0.5
1.2
0.4
0.3
Isobutene
Pentenes
79.1
6.7
94.5
1.6
83.7
5.0
72.3
7.5
75.5
5.3
93.7
2.0
Hexenes
0.4
0.0
0.2
0.8
0.6
0.3
Heptenes
Octenes
Yield of isobutene
Conversion
0.4
4.1
33.4
42.2
0.0
0.9
11.1
11.7
0.1
3.6
31.6
37.7
1.0
7.3
28.1
38.8
0.5
0.0
0.0
26.1
28.0
FIG. 11. Conversion (ᮀ), Yield of isobutene (o), and selectivity to
10.4
24.1
32.0
isobutene (x) over the deactivated ZSM-35 catalyst as a function of
partial pressure of 1-butene. Reaction conditions: temperature = 400�C;
�
1
WHSV = 10 h
.
of 1-butene because of dimerization reactions. Since dimer-
Isobutene is a very reactive molecule since it is reacting ac- ization reactions were favored over cracking reactions at
cording to the A or B mechanisms in the dimerization reac- high partial pressures of 1-butene, an increasing selectivity
tions. Therefore it easily reacts in consecutive reactions, re- to hexenes, heptenes, and especially octenes was observed
ducing the selectivity to isobutene, especially at low WHSV. with increasing partial pressures of 1-butene. The selectiv-
This was evident when the thermodynamic equilibrium be- ity to propene and pentenes increased at the expense of
tween the butene isomers was established, since then the isobutene with increasing partial pressure of 1-butene be-
main reactions were dimerization and subsequent cracking cause of increasing dimerization between butene isomers
of the dimers, increasing the selectivity to the by-products and subsequent cracking to propene and pentenes. An in-
propene and pentenes.
dication that isobutene might be partly produced through
Figure 11 presents results from experiments using differ- dimerization and cracking reactions can be found in Fig. 11,
ent partial pressures of 1-butene over deactivated ZSM-35. since the yield ofisobutene increased with increasingpartial
The conversion increased with increasing partial pressures pressure of 1-butene.
FIG. 10. Conversion (ᮀ), yield of isobutene (o), and selectivity to
FIG. 12. Conversion (ᮀ), yield of isobutene (o), and selectivity to
isobutene (x) over the deactivated ZSM-35 catalyst as a function of
isobutene (x) over the deactivated ZSM-35 catalyst as a function of tem-
�
WHSV. Reaction conditions: temperature = 400 C; partial pressure of 1- perature. Reaction conditions: partial pressure of 1-butene = 0.5 atm;
�
1
butene = 0.5 atm.
WHSV = 10 h
.

6
20
BYGGNINGSBACKA, KUMAR, AND LINDFORS
6. Simon, M. W., Suib, S. L., and O’Young, C.-L., J. Catal. 147, 484 (1994).
Results from experiments using different temperatures
7. Xu, W.-Q., Yin, Y.-G., Suib, S. L., Edwards, J. C., and O’Young, C.-L.,
over ZSM-35 are presented in Fig. 12. The selectivity
to isobutene increased with increasing temperature, since
dimerization reactions became less favored, compared to
cracking reactions. Especially, the selectivity to octene iso-
mers was greatly reduced at high temperatures.
J. Phys. Chem. 99, 9443 (1995).
8
9
. Asensi, M. A., Corma, A., and Mart ´ı nez, A., J. Catal. 158, 561 (1996).
. Bianchi, D., Simon, M. W., Nam, S. S., Xu, W.-Q., Suib, S. L., and
O’Young, C.-L., J. Catal. 145, 551 (1994).
10. Woo, H. C., Lee, K. H., and Lee, J. S., Appl. Catal. A 134, 147
1996).
(
1
1
1
1. Xu, W.-Q., Yin, Y.-G., Suib, S. L., Edwards, J. C., and O’Young, C.-L.,
J. Phys. Chem. 99, 9443 (1995).
2. Xu, W.-Q., Yin, Y.-G., Suib, S. L., and O’Young, C.-L., J. Catal. 150, 34
4
. CONCLUSIONS
ZSM-22 and ZSM-35 are both promising catalysts for
skeletal isomerization of 1-butene, as can be concluded
from the results presented in this paper. The selective re-
action mechanism producing isobutene was most likely
monomolecular over both the catalysts, since cracking of
(
1994).
3. Gielgens, L. H., Veenstra, I. H. E., Ponec, V., Haanepen, M. J., and
van Hooff, J. H. C., Catal. Lett. 32, 195 (1995).
4. Houzvicka, J., Hansildaar, S., and Ponec, V., J. Catal. 167, 273 (1997).
5. Reid, R. C., Prausnitz, J. M., and Poling, B. E., “The Properties of
Gases and Liquids.” McGraw–Hill, New York, 1987.
6. Rahmim, I., Huss, A., Lissy, D. N., Klocke, D. J., and Johnson,
I. D., U.S. Patents 5,157,194 (1992) and 5,237,121 (1993); assigned to
Mobile Oil Corporation.
7. Grandvallet, P., de Jong, K. P., Mooiweer, H. H., Kortbeek, A. B., and
Kraushaar-Czarnetzki, B., Europ. Patent 501,577 (1992); assigned to
Shell International Research Maatschappij B.V.
8. Powers, D. H., Murray, B. D., Winquist, B. H. C., Callender, E. M.,
and Varner, J. H., Europ. Patent, 523,838 (1993); assigned to Lyondell
Petrochemical Company.
9. Mooiweer, H. H., de Jong, K. P., Kraushaar-Czarnetzki, B., Stork,
W. H. J., and Krutzen, B. C. H., in “Zeolites and Related Microporous
Materials: State of the Art 1994” (J. Weitkamp et al., Eds.), Stud. Surf.
Sci. Catal., Vol. 84, p. 2327. Elsevier, Amsterdam, 1994.
0. Xu, W.-Q., Yin, Y.-G., Suib, S. L., and O’Young, C.-L., J. Phys. Chem.
99, 758 (1995).
1
1
1
-octene was not selective to isobutene. The selectivity to
1
1
1
1
isobutene increased with increasing temperature, decreas-
ing partial pressure of 1-butene, and increasing WHSV
of 1-butene for both catalysts, which also supports the
monomolecular mechanism. ZSM-22 was more selective
to isobutene, compared to that of ZSM-35 during the first
hours of TOS, while the deactivated catalysts demonstrated
similar selectivity. Both ZSM-22 and ZSM-35 can be used as
synthesized without any ion-exchanging steps, since there
was no improvement in the activity in 1-butene transforma-
tion after the zeolites had been ion-exchanged in order to
increase the number of Brønsted acid sites. The main ad-
vantages of using ZSM-22 would be that approximately 15
times higher WHSV can be used and the total selectivity to
isobutene is higher over ZSM-22, since no deactivation is
needed in order to get high selectivity to isobutene.
2
2
2
2
2
1. Seo, G., Jeong, H. S., Jang, D.-L., Cho, D. L., and Hong, S. B., Catal.
Lett. 41, 189 (1996).
2. Meriaudeau, P., Tuan, V. A., Le, N. H., and Szabo, G., J. Catal. 169, 397
(
1997).
3. Meriaudeau, P., Naccache, C., Le, H. N., Vu, T. A., and Sazabo, G.,
J. Mol. Catal. A 123, L1 (1997).
4. Pellet, R. J., Casey, D. G., Huang, H.-M., Kessler, R. V., Kuhlman,
E. J., O’Young, C.-L., Sawicki, R. A., and Ugolini, J. R., J. Catal. 157,
ACKNOWLEDGMENTS
Financial support from the Finnish Graduate School in Chemical En-
gineering (GSCE) is gratefully acknowledged.
423 (1995).
2
2
5. Xu, W.-Q., Yin, Y.-G., Suib, S. L., Edwards, J. C., and O’Young, C.-L.,
J. Catal. 163, 232 (1996).
6. Byggningsbacka, R., Lindfors, L. E., and Kumar, N., Ind. Eng. Chem.
Res. 8, 2990 (1997).
REFERENCES
1
. Guisnet, M., Andy, P., Gnep, N. S., Benazzi, E., and Travers, C., J. Catal.
58, 551 (1996).
. Guisnet, M., Andy, P., Gnep, N. S., Travers, C., and Benazzi, E.,
J. Chem. Soc. Chem. Commun., 1685 (1995).
1
27. Vaughan, P. A., Acta Crystallogr. 21, 983 (1996).
28. Kokotailo, G. T., Schlenker, J. L., Dwyer, F. G., and Valyocsik, E. W.,
Zeolites 5, 349 (1985).
2
3
4
5
. Houzvicka, J., and Ponec, V., Appl. Catal. A 145, 95 (1996).
. Houzvicka, J., Diefenbach, O., and Ponec, V., J. Catal. 164, 288 (1996).
. Meriaudeau, P., Bacaud, R., Hung, L. N., and Vu, A. T., J. Mol. Catal. 30. Plank, C. J., Rosinski, E. J., and Rubin, M. K., U.S. Patent 4,016,245
A 110, L177 (1996). (1977).
29. Kumar, N., Lindfors, L. E., and Byggningsbacka, R., Appl. Catal. A
139, 189 (1996).