Skeletal isomerization of n-butenes: I. Mechanism of n-butene transformation on a nondeactivated h-ferrierite catalyst

Article abstract of DOI:10.1006/jcat.1996.0053

The transformation of n-butenes and of isobutene was carried out at 350°C on a fresh H-FER zeolite (Si/A1 = 13.8) with various contact times so as to obtain a large range of conversion (from 5 to 65%). With both reactants the skeletal isomerization was accompanied by the formation of various products. The skeletal isomer, propene, pentenes, octenes (traces), and n-butane (from n-butenes) or isobutane (traces from isobutene) appeared as primary products. The simultaneous formation of the skeletal isomer and of propene and pentenes proved that we were in the presence of a dimerization-cracking process. Moreover, the low amount of octenes in the products showed that dimerization was the limiting step of the process. The dimerization step involved two secondary carbenium ions, which explained its slow rate. The dimer was rapidly isomerized through Type A (alkyl shift) or Type B (via protonated cyclopropanes) isomerization into the other octenes. Lastly octenes with a trimethylpentane or a dimethylhexane skeleton were rapidly cracked into isobutene, propene, and pentenes through Type A (involving two tertiary carbenium ions) and Type B (involving one tertiary and one secondary carbenium ions) mechanisms. The transformation of n-pentenes and of n-octenes confirmed the high rate of the Type A and B isomerization and cracking steps. The slow rate of octene formation from butenes was also due to an inhibition of the diffusion of branched octenes in the narrow pores of H-FER. The shape selectivity of H-FER was also responsible for the very slow formation of isobutane by hydrogen transfer from coke precursors to isobutene and for the unexpected slow formation of propene and of pentenes from isobutene.

Full text of DOI:10.1006/jcat.1996.0053

JOURNAL OF CATALYSIS 158, 551–560 (1996)
ARTICLE NO. 0053
Skeletal Isomerization of n-Butenes
I. Mechanism of n-Butene Transformation on a Nondeactivated H-Ferrierite Catalyst
1,
M. Guisnet, * P. Andy,* N. S. Gnep,* E. Benazzi,† and C. Travers†
*
Laboratoire de Catalyse en Chimie Organique, URA CNRS 350, UFR Sciences, 40 avenue du Recteur Pineau, 86022 Poitiers Cedex, France; and
´
†
Institut Fran c¸ ais du P e´ trole, 1 et 4 Avenue de Bois-Pr e´ au, BP 311, 92506 Rueil-Malmaison Cedex, France
Received March 22, 1995; revised September 12, 1995; accepted September 18, 1995
of acidic catalysts have been reported as being active in
The transformation of n-butenes and of isobutene was carried the isomerization of n-butenes into isobutene (3):
out at 350؇C on a fresh H-FER zeolite (Si/Al ؍ 13.8) with
various contact times so as to obtain a large range of conversion
halogenated aluminas and particularly fluorinated alu-
minas (4–13),
nonhalogenated doped aluminas (14–16), e.g., silicated
(from 5 to 65%). With both reactants the skeletal isomerization
was accompanied by the formation of various products. The
skeletal isomer, propene, pentenes, octenes (traces), and n- aluminas, boron, or tin phosphate (17, 18),
butane (from n-butenes) or isobutane (traces from isobutene)
appeared as primary products. The simultaneous formation of
the skeletal isomer and of propene and pentenes proved that
we were in the presence of a dimerization-cracking process.
Moreover, the low amount of octenes in the products showed
10-membered ring zeolites such as MFl, MEL, FER,
Theta-1 (19–25), boroaluminosilicate molecular sieves
(
26), MCM-22 (27), and silicoaluminophosphate (28–30).
The skeletal isomerization of C hydrocarbons (butanes
4
that dimerization was the limiting step of the process. The or butenes) through acid catalysis compared to that of
dimerization step involved two secondary carbenium ions,
which explained its slow rate. The dimer was rapidly isomerized
through Type A (alkyl shift) or Type B (via protonated cyclo-
propanes) isomerization into the other octenes. Lastly octenes
with a trimethylpentane or a dimethylhexane skeleton were
rapidly cracked into isobutene, propene, and pentenes through
Type A (involving two tertiary carbenium ions) and Type B
larger hydrocarbons presents important particularities
31–33). Indeed the monomolecular mechanism of C4
isomerization involves necessarily a very unstable pri-
mary carbenium ion intermediate even if, as generally ad-
mitted, a protonated cyclopropane intermediate is also
involved,
(
(involving one tertiary and one secondary carbenium ions)
mechanisms. The transformation of n-pentenes and of n-oc-
tenes confirmed the high rate of the Type A and B isomerization
and cracking steps. The slow rate of octene formation from
butenes was also due to an inhibition of the diffusion of
branched octenes in the narrow pores of H-FER. The shape
selectivity of H-FER was also responsible for the very slow
formation of isobutane by hydrogen transfer from coke precur-
sors to isobutene and for the unexpected slow formation of
C
C
C
+
+
C
H
+
C
C
C
C
C
C
C
C
[a]
which it is not the case for the isomerization of larger
hydrocarbons,
propene and of pentenes from isobutene.
 1996 Academic Press, Inc.
C
+
+
H
C
C
C
C
C
C
C
C
C
INTRODUCTION
C
C
Due to its use in the synthesis of methyltertiobutylether
MTBE), an oxygenated gasoline additive having a high
+
C
C
C [b]
(
octane number (1, 2), there is an increasing interest in the
production of isobutene from n-butenes. A large variety
The consequence, demonstrated by Chevalier et al. (33)
in the case of a very low pressure of butenes is that reaction
[a] is much slower than reaction [b] or than the scrambling
1
To whom correspondence should be addressed.
5
51
0021-9517/96 $12.00
Copyright  1996 by Academic Press, Inc.
All rights of reproduction in any form reserved.

552
GUISNET ET AL.
of 13C in C molecules which involves the same intermedi- case, the weight hourly space velocity (WHSV) was varied
4
Ϫ
1
ates as reaction [b]
(from 2 to 70 h ) to obtain a large range of reactant
conversion (from 5 to 65%). Reaction products from all
these reactants were analyzed on-line by gas chromatogra-
phy using a 50-mChrompack PLOT Al O /Na SO capil-
1
3
C
+
+
C
H
+
C
13
13
2
3
2
4
C
C
C
C
C
C
C
C
C [c]
lary column.
In spite of this, the monomolecular mechanism has been
proposed by various authors to explain the skeletal isomer-
ization of butanes (34) and of butenes (4, 12, 21).
RESULTS
1
. Preliminary Study of 1-Butene and
According to other authors (24, 35–38) a bimolecular
mechanism is responsible for the relatively rapid skeletal
isomerization of n-butanes or of n-butenes when a rela-
tively high reactant pressure is used. With n-butenes as a
reactant this mechanism involves three successive steps:
dimerization of butenes, skeletal isomerization of the di-
mers, and cracking of the octene isomers. When the various
steps are considered, propene ϩ pentenes and isobutene
are formed simultaneously. Therefore the simultaneous
formation of these alkenes (or in the case of n-butane
transformation, that of propane, pentanes, and isobutane)
constitutes a strong argument in favor of this ‘‘bimolecu-
lar’’ mechanism as shown in this paper. However, it has
sometimes been considered that the formation of propene
and pentenes was a process completely independent of the
skeletal isomerization of n-butenes (12, 26). The participa-
tion of the bimolecular process has been demonstrated for
the isomerization of n-butane into isobutane over morde-
Isobutene Transformations
A preliminary study of 1-butene and isobutene transfor-
mations was carried out under the following conditions:
T ϭ 623 K, palkene ϭ 0.1 bar, pN2 ϭ 0.9 bar, 0.05 g of
catalyst, alkene flow rate ϭ 18.4 mmol h
WHSV ϭ 20.6 g of alkene introduced per gram of catalyst
and per hour).
From n-butenes the main reaction products are isobu-
tene, propene, pentenes (iso- and n-), and n-butane. Ethyl-
ene, ethane, propane, hexenes, heptenes, and octenes are
observed in low amounts. From isobutene the main prod-
ucts are n-butenes, propene, pentenes, and octenes; ethyl-
ene, propane, n-butane, isobutane, and hexenes are also
observed. Table 1 gives the product distribution (wt%)
obtained after a 5-min reaction.
Whatever the reactant, the distribution of linear butenes
is close to that of thermodynamic equilibrium, which means
that the double bond shift is very fast. Therefore the con-
version is calculated by grouping n-butene isomers to-
gether. Figure 1 shows the change with time-on-stream
Ϫ
1
(i.e.,
nites (37) or over Fe, Mn-promoted sulfated ZrO catalyst
2
by using 13C-labeled molecules (38).
We show here that this bimolecular mechanism explains
the transformation of n-butenes and isobutene over a non-
deactivated H-FER catalyst. A process for the skeletal
isomerization of n-butenes, using this catalyst, has been
developed by Lyondel Petrochemicals. This new process
would ‘‘produce isobutene at potentially half the capital
cost of alternative technologies’’ (25, 39).
(
TOS) in the total conversion of n-butenes and in their
conversion into the main products. The conversion into
isobutene increases slightly during the first 3 h then de-
creases. A quasi-identical curve is obtained for the conver-
sion into octenes. On the other hand there is a rapid initial
decrease in the conversion into the other products: ethyl-
ene, propene, propane, n-butane, pentenes, hexenes, and
heptenes. Therefore the selectivity to isobutene increases
from 57 wt% at TOS ϭ 5 min to about 92 wt% at TOS Ͼ
EXPERIMENTAL
2
00 min. This selectivity increase is probably due to the
Ferrierite (FER) was prepared in fluoride medium at
formation of carbonaceous deposits (coke) (40). The same
phenomenon is observed for isobutene transformation, the
catalyst deactivation, however, being less pronounced.
The selectivity to n-butenes increases from 91.5 wt% at
TOS ϭ 5 min to 97.5 wt% at TOS ϭ 450 min. It can be
remarked that this selectivity to the skeletal isomers is
much higher than that observed in n-butene transfor-
mation.
the Laboratoire des Mat e´ riaux Min e´ raux in Mulhouse
ϩ
4
(
France). Its H-form was prepared by NH ion exchange
which was performed three times with an excess of 10 N
ammonium nitrate solution during 4 h at 373 K; its Na, K
content was less than 50 ppm and its Si/Al ratio was equal
to 13.8. Prior to use, it was calcined in situ under dry air
flow at 773 K.
The transformations of 1-butene (99.4% pure), isobutene
(
99.5% pure), propene (99.4% pure), 1-pentene (97%
2
. Kinetic Study of 1-Butene and of
Isobutene Transformations
pure), and 1-octene (98% pure) were carried out in a flow
reactor in the presence of nitrogen at 623 K under atmo-
spheric pressure; the reactant and nitrogen partial pres-
To limit the effect of deactivation on the activity and
sures were equal to 0.1 and 0.9 bar, respectively. In every selectivity, the kinetic study of 1-butene and of isobutene

SKELETAL ISOMERIZATION OF n-BUTENE, I
553
TABLE 1
Distribution (wt%) of the Products of n-Butene and of Isobutene Transformations after 5-min Reaction
؊
1
for a WHSV Value of 20.6 h
Xa
C
1
2
Ϫ
C
2
Ϫ
C
n-C2
Ϫ
i-C2
Ϫ
n-C
i-C
0
C
2
Ϫ
C
2
Ϫ
C
2
Ϫ
C
8
2
Ϫ
⌺
2
3
3
4
4
4
4
5
6
7
n-Butenes
Isobutene
14.6
17.3
19.4
3.6
0.5
0.1
0
61.2
0
4.5
0.65
10.7
2.25
1.35
0.15
0.85
0
0.5
1.4
100
100
0.25
91.5
0.1
a
Conversion.
transformations was carried out at low TOS values. A time) on the conversion of n-butenes into isobutene, pro-
rectangular pulse technique was used for determining pene, pentenes, and n-butane. From this type of figure the
the effect of the operating parameters: injecting the kinetic nature of the products (primary, secondary, etc.,
reactant for 3 min (a constant value of the reactant according to the classification of Abbot and Wojciechowski
concentration was obtained after about 2 min), analyzing (41)) can be deduced. Moreover, the rates of transforma-
the products by on-line gas chromatography (50 min), tion of the reactant into the primary products can be deter-
and the catalyst being kept during this time under a mined from the initial slope of the curves. The curves
Ϫ
1
nitrogen flow of 4000 ml h
giving the yields as a function of the conversion (Fig. 3)
confirm the kinetic nature of the products.
2.1. Effect of contact time: Reaction scheme. Figure 2
In n-butene transformation, isobutene, propene, pen-
shows as an example the effect of contact time (actually tenes, n-butane, ethylene, and octenes appear as primary
the reverse of WHSV, i.e., a value proportional to contact products while ethane, propane, isobutane, hexenes, and
FIG. 1. Influence of time-on-stream (TOS) on the conversion of n-butenes into various products. Operating conditions: Pn-butene ϭ 0.1 bar;
Ϫ
1
WHSV ϭ 20.6 h
.

554
GUISNET ET AL.
FIG. 2. Transformation of n-butenes on a nondeactivated H-FER sample. Effect of contact time on the yields of the main reaction products.
heptenes are secondary. It must be emphasized that the pene, pentenes, isobutane (only traces), and octenes
molar ratios between propene and pentenes and between appear as primary products while ethylene, propane, n-
ethylene and hexenes are always greater than 1. Thus butane, and hexenes are secondary products. Moreover,
the molar ratio between propene and pentenes passes there is also a secondary mode of formation of propene
from about 2 at zero conversion to about 3 at 60% and pentenes. Like for n-butene transformation the
conversion. In isobutene transformation, n-butenes, pro- propene/pentenes molar ratio is between 2 and 3 and
FIG. 3. Transformation of n-butenes on a nondeactivated H-FER sample. Yields of products versus conversion of n-butenes.

SKELETAL ISOMERIZATION OF n-BUTENE, I
555
TABLE 2
Rates of Transformation of Alkenes into Apparent Primary Products on the
Nondeactivated Ferrierite
Ϫ
1
g
Ϫ1)
Rate (mmol h
Product:
Reactant
Propene
Isobutene
68
n-Butenes
Pentenes
Butane
Hexenes
Octenes
n-Butenes
Isobutene
Propene
n-Pentenes
-Octenes
9
1.2
7 (n ϩ i)
1 (n ϩ i)
4 (n)
0.2 (i)
1-2
0.7 (n ϩ i)
3.4 (n ϩ i)
141
40
100
1300 (i)
50 (n ϩ i)
1
33
6
4500 (i)
the distribution of pentenes is slightly different from that conversion, the pressure of reactant being equal to 0.1 bar.
at thermodynamic equilibrium.
The following main products were obtained:
The rates of n-butene and of isobutene transformations
into the apparent primary reaction products are reported
in Table 2. Curiously the ratio between the skeletal isomer-
ization rates of isobutene and of n-butenes is greater than
the ratio expected from the thermodynamic equilibrium
from propene: hexenes, butenes, pentenes, ethene,
propane
from 1-pentene: the thermodynamic equilibrium mix-
ture of n-pentenes, isopentenes, pro-
pene, ethene, butenes, hexenes
(
2 instead of 1.3). This greater value cannot be due to
from 1-octene:
n-octenes and most of the isooctenes
more than 20 GC peaks), butenes,
pentenes, propene.
deactivation phenomena because the initial deactivation
of n-butene isomerization (Fig. 1) or of isobutene isomer-
ization is slow. Furthermore, n-butenes are transformed
more rapidly into propene and pentenes than isobutene
(
From the effect of the contact time (carried out with
(
about 7 times) and more slowly into octenes (about 5 the rectangular pulse technique) it is possible to classify
times). Moreover the rate of transformation of n-butenes these products into primary and secondary ones. From
into n-butane is at least 20 times greater than that of isobu- propene, only hexenes appear as primary products, from
tene into isobutane.
n-pentenes, only isopentenes, and from n-octenes, isooc-
tenes, butenes, pentenes, and propene. Table 2 compares
the rates of transformation of the reactants into these pri-
mary products with those of n-butenes and isobutene trans-
formations.
2.2. Influence of the operating parameters. The effect
of the operating parameters is as follows: partial pressure
of the reactant, temperature, on the initial rates of forma-
tion of the main primary products was determined either
by using various samples of catalysts under different condi-
tions or by varying the operating conditions on one sample.
In the range of pressure under consideration (0.05 to 0.3
bar) the reaction order with respect to n-butenes is close
to 1 for the formation of propene and pentenes while it is
smaller than 1 (about 0.5) for the skeletal isomerization
of n-butenes. On the other hand, the order with respect
to isobutene for its isomerization into n-butenes is close
to 1. The activation energy for n-butene isomerization is
DISCUSSION
The preliminary study of n-butene and isobutene trans-
formations on H-FER confirms the significant positive ef-
fect deactivation by coke deposit has on the selectivity of
the skeletal isomerization (29, 40). This curious positive
effect of coke was recently attributed to the development
of a new reaction mechanism involving as active sites car-
bocations blocked in the zeolite pores (40). The present
article is limited to the establishment of reaction schemes
and mechanisms on nondeactivated catalyst samples.
Ϫ
1
equal to 14 Ϯ 1 kcal mol and that for isobutene isomeriza-
Ϫ
1
tion is equal to 10 Ϯ 1 kcal mol . The values for the
formation of propene and of pentenes from n-butenes are
lower, but rapid deactivation makes it impossible to deter-
mine them with precision.
1. Generalities
On the nondeactivated H-FER catalysts the transforma-
tion of n-butenes and of isobutene is very complex. Indeed
the skeletal isomerization is accompanied by the formation
3
. Transformation of Side Products of Butene
Conversion
The transformation of propene, 1-pentene, and 1-octene of various products. This formation involves one or several
was carried out under the standard conditions of butene of the following reactions: dimerization of alkenes (or al-

556
GUISNET ET AL.
value deduced from the thermodynamic equilibrium data
(
2 instead of 1.3).
ii. The activation energy for isobutene isomerization is
lower than that found for n-butene isomerization although
the reverse is expected from the exothermicity of n-butene
isomerization.
iii. The reaction order with respect to n-butenes is equal
to 0.5 while the one with respect to isobutene is equal to
1. If it is supposed that the reversible n-butene–isobutene
isomerization occurs through a monomolecular mechanism
this would mean that isobutene is weakly adsorbed, n-
butene adsorption being relatively strong although the con-
FIG. 4. Scheme of n-butene transformation on a nondeactivated H- trary is expected from the stability of the corresponding
FER sample.
carbocations.
Therefore it can be concluded that n-butene and isobu-
tene isomerization reactions do not involve the same steps
and hence do not occur through the same mechanisms.
Since isobutene isomerization occurs with only the slow
formation of propene and pentenes (which requires neces-
sarily a dimerization-cracking mechanism with an interme-
kylation of one alkene by another), cracking, and hydride
transfer from coke precursors to alkenes. Moreover, if
the same reversible steps are involved in n-butene and
isobutene isomerizations, certain results cannot be easily
explained, namely the following.
i. The ratio between the initial rates of skeletal isomer- diate isomerization of the C dimer), it could be supposed
8
ization of isobutene and of n-butenes is greater than the that this reaction occurs mainly through a monomolecular
FIG. 5. Transformation of n-butenes through the dimerization-cracking mechanism: (a) Dimerization steps, (b) isomerization of the dimers,
and (c) cracking steps. MHept, methylheptyl; DMH, dimethylhexyl; TMP, trimethylpentyl.

SKELETAL ISOMERIZATION OF n-BUTENE, I
557
FIG. 5—Continued

558
GUISNET ET AL.
mechanism. On the other hand the skeletal isomerization (Type A isomerization) into all the carbenium ions with
of n-butenes which is accompanied by the formation of a methylheptane or dimethylhexane skeletons (steps 1b and
large amount of propene and pentenes would occur prefer- 2b). Isomerization with supplementary branching (Type B
entially through the dimerization-cracking mechanism. isomerization, i.e., through protonated cyclopropanes) can
However, the monomolecular mechanism which requires also occur with formation of carbenium ions with a trimeth-
the formation of a very unstable primary carbenium ion ylpentane skeleton (step 3b).
(
reaction [a]) is most unlikely. Also, in disfavor of this
The cracking of carbenium ions with a monobranched
monomolecular mechanism the activation energy for iso- skeleton involves two secondary carbenium ions as inter-
butene isomerization is lower than that for n-butene isom- mediates (Type C cracking), while the cracking of di-
erization whose intermediates are more stable. A more branched carbenium ions can occur through Types B1 and
realistic proposal is that n-butene and isobutene skeletal B2 (one secondary ϩ one tertiary carbenium ions involved)
isomerizations occur through the dimerization-cracking or through Type C steps. Lastly scission of carbenium ions
mechanism but that these isomerization reactions do not with a tribranched skeleton can occur through Type A
involve the same C intermediates.
(two tertiary carbeniums ions involved) or through Type
B steps (see Fig. 5).
8
According to Brouwer (32), the rates of the various
cracking and isomerization steps decrease as follows:
2
. n-Butene Transformation: Reaction Scheme
and Mechanisms
The scheme in Fig. 4 could be proposed to explain the
primary or secondary apparent formations of the products.
Hydride shift Ͼ A isomerization Ͼ A cracking
Ͼ B isomerization Ͼ B1, B2 cracking
Ͼ C cracking.
It must be underlined that most likely the C carbenium
8
ions (rather than the octenes desorbed from the zeolite)
2
Ϫ
2
Ϫ
are the real intermediates in the production of C , C , However, Martens et al. (43) report that Type A cracking
2
3
2
Ϫ
2
Ϫ
2
Ϫ
i-C , C , C . Because these products appear as primary is faster than Type A isomerization. Therefore the slowest
4
5
6
ones, steps 2, 3, and 4 are necessarily faster than step 1. step of the dimerization-cracking mechanism would be the
Furthermore line 1 in Table 2 shows that step 2 is faster dimerization steps (steps 1a and 2a), which are the reverse
than step 3 (about 4 times) and much faster than step 4 steps of C cracking reactions and even more that at the
2
Ϫ
(
20–50 times). Hydride transfer to n-C4 (step 5) is about reaction temperature, dimerization compared to cracking
1
1
5 times slower than butene isomerization (through steps is highly unfavored.
and 2). Hydride transfer to isobutene (step 6) is very slow, The participation of Type B isomerization steps in n-
which could be due to the bulk of the reaction intermediate butene transformation (step 3b) is confirmed by experi-
while hydride transfer to propene (step 7) and to ethylene ments with n-pentenes and n-octenes. Indeed the skeletal
(
5
step 8) appears to occur at rates similar to those in step isomerization of n-pentenes which occurs through proton-
. Hydrogen transfer reactions can be considered as irre- ated cyclopropane intermediates (Type B isomerization)
versible while dimerization, alkylation, and cracking are is much faster than the transformation of n-butenes into
reversible (Fig. 4). isobutene (20 times), into propene (150 times), and into
The main steps involved in the dimerization-cracking pentenes (200 times). Moreover, this is demonstrated by
mechanism of n-butene transformation are described in the apparent direct formation of isobutene and isopentenes
Fig. 5. For the sake of simplification the steps involving from 1-octene: the formation of these products requires
primary carbenium ions as intermediates which are very two or three successive Type B isomerization steps (for
slow are not included. Three types of steps can be distin- the formation of carbenium ions with a di- or tribranched
guished: (a) the dimerization steps, (b) the isomerization of skeletons) followed by Type A or B cracking steps.
the dimers, and (c) the cracking of the C carbenium ions.
Figure 5 shows that isobutene, propene, and pentenes
8
Dimerization (steps 1a and 2a) leads to carbenium ions can be formed from n-butenes through various pathways.
with a 3-methylheptane and with a 3,4-dimethylhexane Isobutene can result from A and B1 cracking of carbenium
ϩ
skeleton, respectively, the second one being certainly fa- ions with trimethylpentane skeleton (TMP ) and from B1
vored because of the larger amount of 2-butenes. Indeed, and B2 cracking of carbenium ions with a dimethylhexane
ϩ
ϩ
even at very low contact times, a thermodynamic equilib- skeleton (DMH ). The formation of TMP requires at
rium mixture of n-butenes (22.5% 1-butene, 77.5% 2-bu- least one Type B isomerization of the carbenium ions re-
ϩ
tenes) is formed. It must be emphasized that steps 1a and sulting from n-butene dimerization while that of DMH
2
a are the reverse steps of Type C cracking reactions (i.e., can occur through a Type A isomerization. Propene and
ϩ
involving two secondary carbenium ions (42)).
The carbenium ions formed through steps 1a and 2a can B1, B2, and C cracking (two possibilities) of DMH ; and
be transformed through hydride transfer and alkyl shift from C cracking of carbenium ions with a methylheptane
pentenes can result from Type B2 cracking of TMP ; from
ϩ

SKELETAL ISOMERIZATION OF n-BUTENE, I
559
skeleton. C cracking steps, being very slow compared to energy is lower for isobutene than for n-butene isomer-
B cracking, can probably be neglected. From this figure it ization.
can be concluded that the dimerization-cracking mecha-
The propene ϩ pentenes formation is much slower than
nism must lead not only from n-butenes to propene and isobutene isomerization, which was not expected from the
pentenes but also to isobutene. However, as observed, dimerization-cracking mechanism. Indeed the formation
isobutene formation is faster than propene ϩ pentenes of these products involves practically the same steps as
formation for it involves a very rapid A cracking step.
isobutene isomerization:
3. Isobutene Transformation: Reaction Scheme
C
and Mechanisms
+
C
+
C
~
H, ~CH
3
C
C
C
C
C
C
C
C
C
C
C
[g]
[h]
The formation of n-butenes, propene, and pentenes as
C
C
primary products shows that the cracking steps of the
dimerization-cracking mechanism are, as is the case for n-
butene transformation, faster than the dimerization steps.
However, a larger amount of octenes is found. This shows
that the difference between the rates of dimerization and
cracking steps is less pronounced than in the case of n-
butene transformation. In agreement with this proposal,
ethylene appears as a secondary product, whereas it is
an apparent primary product of n-butene transformation.
Moreover, to the primary mode of propene ϩ pentenes
formation is added another mode of formation (probably
via secondary transformation of n-butenes). Furthermore
hydrogen transfer from coke precursors to isobutene is 20
times slower than that to n-butenes. This is probably due
to steric contraints in the formation of the transition state
of hydrogen transfer to isobutene.
+
C
+
C
C
C
C
C
C
C
C
C
C + C
C
C
C
C
This means most likely that either the formation of the
,3,4-TMP carbenium ion (reaction [g]) or its cracking
reaction [h]) are strongly inhibited in the H-FER pores.
Another effect of the H-FER shape selectivity is the
ϩ
2
(
slow formation of octenes. Indeed step [f] (B cracking)
being slower than step [e] (A isomerization) and slower
than the reverse of step [d] (A cracking), a large amount
of octenes should be observed in the products of isobutene
transformation. This is not the case even if the formation
of octenes from isobutene is faster than from n-butenes
Table 2). This can be attributed to limitations in the diffu-
sion of branched dimers in the narrow pores of H-FER
24), hence in the desorption of these dimers from the
zeolite crystallites.
As suggested previously, isobutene isomerization does
not occur through the reverse steps of n-butene isomeriza-
tion. Indeed the most likely pathway of isobutene isomer-
ization is as follows:
(
(
C
+
+
CONCLUSION
C
C
C + C
C
C
C
C
C
C
C
C
C
[d]
[e]
[f]
The reaction schemes of n-butene and of isobutene
transformations on a nondeactivated H-FER sample were
established. The formation of the main products, skeletal
isomer(s), propene, and pentenes, involves dimerization
steps followed by isomerization of the dimers and their
cracking. The dimerization step is much slower than the
isomerization and cracking steps because the latter involve
more stable carbenium ions and because the diffusion of
branched dimers in the narrow pores of H-FER is inhib-
ited. The shape selectivity of H-FER limits also the forma-
tion of propene, pentenes, and isobutane from isobutene.
There is no need to involve a monomolecular mechanism
to explain the skeletal isomerization of butenes on the
nondeactivated H-FER catalyst.
C
C
C
C
+
+
C
~H, ~CH
3
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
+
+
C
C
C
C
C
C
C + C
C
C
C
C
C
All these steps are rapid. Indeed step [d] is the reverse
step of Type A cracking, step [e] is an A type isomerization,
and step [f] a B2 type cracking. This explains why the rate
of isobutene isomerization is twice as fast as the rate of n-
butene isomerization (instead of the 1.3 value expected
from a reversible process) and also why the activation
ACKNOWLEDGMENT
P. Andy acknowledges the R e´ gion Poitou Charentes for a scholarship.

560
GUISNET ET AL.
2
3. Grandvallet, P., Mooiweer, H. H., Kortfeek, A. G., and Kraushaar-
Czarnetzki, B., EP Patent No. 0501577 A1 (1992), assigned to Shell
International Research Maatschappij B.V..
REFERENCES
1
2
. Juguin, B., Torck, B., and Martino, G., in ‘‘Catalysis by Acids and
Bases’’ (B. Imelik et al., Eds.), Stud. Surf. Sci. Catal., Vol. 20, p. 253.
Elsevier, Amsterdam, 1985.
. Maxwell, I. E., Naber, J. E., and de Jong, K. P., Appl. Catal., A:
General 113, 153 (1994).
. Butler, A. C., and Nicolaides, C. P., Cataly. Today 18, 443 (1993).
. Choudhary, V. R., and Doraiswamy, L. K., J. Catal. 23, 54 (1971).
. Choudhary, V. R., ‘‘Chemical Industry Developments,’’ Vol. VIII,
No. 7, p. 32. Incorporating CP & E, 1974.
. Choudhary, V. R., Ind. Eng. Chem., Prod. Res. Dev. 16, 12 (1977).
. Raghavan, N. S., and Doraiswamy, L. K., J. Catal. 48, 21 (1977).
. Juguin, B., and Miguel, J., EP Patent No. 66485 (1982), assigned to
Institut Fran c¸ ais du P e´ trole.
2
4. Mooiweer, H. H., de Jong, K. P., Kraushaar-Czarnetzki, B., Storck,
W. H. J., and Krutzen, S. C. H., in ‘‘Zeolite 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.
5. Powers, D. H., Murray, B. D., Winquist, B. H. C., Callender; E. M.,
and Varner, J. H., EP Patent No. 0523823 A2 (1992), assigned to
Lyondell Petrochemical Company.
2
3
4
5
2
6. Bianchi, D., Simon, M. W., Nam, S. S., Xu, W. Q., Suib, S. L., and
O’Young, J. Catal. 145, 551 (1994).
6
7
8
2
7. Bundens, R. G., Keville, K. M., Huss, A., Chu, C. T. W., and Husain,
A., U.S. Patent No. 5146029 (1990), assigned to Mobil Oil Corpo-
ration.
2
2
8. Gaffney, A. M., and Jones, C. A., U.S. Patent No. 5107050 (1990),
assigned to Ares Chemical Technology L.P..
9. Gajda, G. J., U.S. Patent No. 5132484 (1991), assigned to UOP.
9
. Juguin, B., and Miguel, J., U.S. Patent No. 4434315 (1984), assigned
to Institut Fran c¸ ais du P e´ trole.
1
1
1
0. Cheng, Z. X., and Ponec, V., Appl. Catal. 118, 127 (1994).
1. Cheng, Z. X., and Ponec, V., J. Catal. 148, 607 (1994).
2. Szabo, J., Perrotey, J., Szabo, G., Duchet, J. C., and Cornet, D., J.
Mol. Catal. 67, 76 (1991).
3. Szabo, J., Szabo, G., van Gestel, J., and Cornet, D., Appl. Catal., A:
General 96, 319 (1993).
4. Juguin, B., and Martino, G., SA Patent No. 90/2870 (1991) and FR
Patent No. 2657605 (1991), assigned to Institut Fran c¸ ais du P e´ trole.
5. Travers, C., Dagand, J., Burzynski, J. P., and Courty, P., EP Patent
No. 588660 (1992), assigned to Institut Fran c¸ ais du P e´ trole.
6. Nilsen, B. P., Omferlo, J. H., and Gates, B. C., Ind. Eng. Chem.
Fundam. 25, 337 (1986).
30. Gajda, G. J., and Barger, P. T., U.S. Patent No. 5191146 (1991),
assigned to UOP.
31. Brouwer, D. M., Rec. Trav. Chim., Pays-Bas 87, 1435 (1968).
32. Brouwer, D. M., in ‘‘Nato Adv. Stud. Inst., Series E’’ (R. Prins and
G. C. A. Schuit, Eds.), Vol. 39, p. 137. Sijthoff & Noordhoff, Alphen
aan den Rijn, 1980.
33. Chevalier, F., Guisnet, M., and Maurel, R. in ‘‘Proceedings, 6th Inter-
national Congress on Catalysis, London, 1976’’ (G. C. Bond, P. B.
Wells, F. C. Tompkins, Eds.), Vol. 1, p. 478. The Chemical Society,
London, 1976.
1
1
1
1
1
3
4. Garin, F., Seyfried, L., Girard, P., Maire, G., Abdulsamad, A., and
Sommer, J., J. Catal. 151, 26 (1995).
3
5. Guisnet, M., Gnep, N. S., Bearez, C., and Chevalier, F., in ‘‘Catalysis
by Zeolites’’ (B. Imelik et al., Eds.), Stud. Surf. Sci. Catal., Vol. 5,
p. 77. Elsevier, Amsterdam, 1980.
7. Nilsen, B. P., Stoeker, M., and Riis, T., in ‘‘Catalysis by Acids and
Bases’’ (B. Imelik et al., Eds.), Stud. Surf. Sci. Catal., Vol. 20, p. 253.
Elsevier, Amsterdam, 1985.
3
3
3
6. Bearez, C., Chevalier, F., and Guisnet, M., React. Kinet. Catal. Lett.
1
8. La Ginestra, A., Patrono, P., Massucci, M. A., Galli, P., Ferragina,
C., and Mancini, C., in ‘‘Proceedings, 9th International Congress on
Catalysis, Calgary, 1988’’ (M. J. Phillips and M. Ternan, Eds.), Vol.
2
2, 405 (1983).
7. Bearez, C., Avendano, F., Chevalier, F., and Guisnet, M., Bull. Soc.
Chim. Fr. 3, 346 (1985).
8. Adeeva, V., Lei, G. D., and Sachtler, W. M. H., Appl. Catal., A:
General 118, L11 (1994).
1
, p. 449. Chem. Institute of Canada, Ottawa, 1988.
1
9. O’Young, C. L., Xu, W. Q., Simon, M. W., and Suib, S. L., in ‘‘Zeolite
and Related Microporous Materials : State of the Art 1994’’ (J. Weit-
kamp et al., Eds.), Stud. Surf. Sci. Catal., Vol. 84, p. 1671. Elsevier,
Amsterdam, 1994.
3
4
9. Chem. Eng. News, March 23, 7 (1992).
0. Guisnet, M., Andy, P., Travers, C. and Benazzi, E., J. Chem. Soc.,
Chem. Commun., 1685 (1995).
2
2
0. Simon, M. W., Suib, S. L., and O’Young, C. L., J. Catal. 147, 484 (1994).
1. Xu, W. Q., Yin, Y. G., Suib, S. L., and O’Young, C. L., J. Catal. 150, 42. Weitkamp, J., Jacobs, P. A., and Martens, J. A., Appl. Catal. 8,
4
1. Abbot, J., and Wojciechowski, B. W., J. Catal. 90, 270 (1984).
3
4 (1994).
123 (1983).
2
2. Simon, M. W., Xu, W. Q., Suib, S. L., and O’Young, Microporous
Mat. 2, 477 (1994).
43. Martens, J. A., Jacobs, P. A., and Weitkamp, J., Appl. Catal. 20,
239 (1986).