Ethylbenzene isomerization on bifunctional platinum alumina-mordenite catalysts. 1. Influence of the mordenite Si/Al ratio

Article abstract of DOI:10.1006/jcat.2001.3294

The C₈ aromatic fraction is obtained from steam cracking and especially from naphtha reforming. The transformation of ethylbenzene was performed on mixtures of 0.5 wt % Pt/Al₂O₃ (75 wt %) with a platinum dispersion of 100% and of HMOR zeolites with Si/Al ratios between 6.6:1 and 180:1. The following products were found: C₂-C₆; C₆-C₈ naphthenes, cyclohexane, and methylcyclopentane; and C₆-C₁₀₊ aromatic hydrocarbons (benzene, toluene, xylenes, ethyltoluenes, trimetylbenzenes). These products resulted from six main transformations of ethylbenzene: (1) isomerization of ethylbenzene into o-, m- and p-xylenes, (2) disproportionation of ethylbenzene into benzene and diethylbenzenes, (3) dealkylation of ethylbenzene into benzene and ethylene totally transformed into ethane or into longer alkanes, (4) transalkylation of ethylbenzene-xylenes, (5) hydrogenation of C₈ aromatics, and (6) cracking of C₈ naphthenes into C₃-C₆ alkanes. With all the catalysts, hydrogenation and disproportionation were the fastest reactions and dealkylation was very slow. The rate of isomerization depended very much on the zeolite component. Reactions 2-4 occurred through acid catalysis and hydrogenation through metal catalysis, and reactions 1 and 6 through bifunctional catalysis.

Full text of DOI:10.1006/jcat.2001.3294

Journal of Catalysis 202, 402–412 (2001)
doi:10.1006/jcat.2001.3294, available online at http://www.idealibrary.com on
Ethylbenzene Isomerization on Bifunctional Platinum
Alumina–Mordenite Catalysts
1. Influence of the Mordenite Si/Al Ratio
,1
F. Moreau, S. Bernard, N. S. Gnep, S. Lacombe,† E. Merlen,† and M. Guisnet
Catalyse en Chimie Organique, Faculte´ des Sciences, UMR 6503, 40, avenue du Recteur Pineau, 86022 Poitiers Cedex, France;
and †Institut Franc¸ais du Pe´trole, 1 et 4, avenue de Bois-Pre´au, 92852 Rueil-Malmaison Cedex, France
Received March 27, 2001; revised May 21, 2001; accepted May 21, 2001
INTRODUCTION
The transformation of ethylbenzene was carried out on intimate
mixtures of 0.5 wt% Pt/Al2O3 (75 wt%) with a platinum dispersion
The C8 aromatic fraction which is obtained from steam
of 100% and of HMOR zeolites (25 wt%) with Si/Al ratios between cracking and especially from naphtha reforming is consti-
6.6 and 180. Products result from six main transformations of ethyl-
benzeneor/and ofthereaction products:thedesired isomerization of
ethylbenzene into xylenes (reaction 1), disproportionation of ethyl-
benzene (2), dealkylation followed by hydrogenation of ethylene (3),
secondary ethylbenzene–xylene transalkylation (4), hydrogenation
of ethylbenzene followed by ethylcyclohexane isomerization (5), and
secondary cracking of C8 naphthenes (6). From the comparison
of the product distributions over Pt/Al2O3, over HMOR and over
the bifunctional catalysts, reactions 2–4 were confirmed to occur
through acid catalysis and hydrogenation through metal catalysis,
and reactions 1 and 6 through bifunctional catalysis. Whatever the
tuted by the mixture of xylenes at thermodynamic equilib-
rium, i.e., approximately 25% of para, 25% of ortho, and
50% of meta, together with approximately 20% and 50%
ethylbenzene in the C8 fractions originating from reforming
and from steam cracking, respectively (1–11). The demand
for paraxylene, already very significant, is continuously in-
creasing: approximately 8% per year during the past 10
years (9); paraxylene is oxidized to terephthalic acid which
is used in the production of polyester films and fibers. Sep-
aration of ethylbenzene that is used for the production of
catalyst, the selectivity to isomers increases with ethylbenzene con- styrene requires high efficiency, and hence expansive dis-
version, which is due to thermodynamic limitations in reactions
2 and 5. A significant increase in selectivity to isomers is also ob-
served when the Si/Al ratio of the mordenite component increases
from 20 to 180. With these bifunctional catalysts, the initial rate of
disproportionation is proportional to the square of the concentra-
tion of protonic sites, which suggests that this bimolecular reaction
requires two protonic sites for its catalysis; furthermore, the rate of
dealkylation (per protonic site) depends only on the strength of the
protonic sites and the change of the isomerization rate with the bal-
ance between the metallic and acidic functions is the one expected
from a classical bifunctional mechanism. For the bifunctional cata-
tillation. It is why, in the commercial units, ethylbenzene
separation is generally omitted.
To meet the demand for paraxylene, the much less used
meta- and orthoxylenes have to be converted to additional
paraxylene. This isomerization being limited by thermody-
namic equilibrium, the separation process plays a signifi-
cant role in paraxylene production. Separation of paraxy-
lene was first carried out by crystallization. However, the
high cost of equipment, high energy consumption, and lim-
ited yield of crystallization have led to the substitution of
lysts with HMOR6.6 and 10 as acidic components, the rates of selective adsorption on molecular sieves for crystallization
disproportionation, dealkylation, and isomerization are lower than
expected and the selectivity to isomers is higher. This can be ex-
plained by strong diffusion limitations in these zeolites which, in
contrast to the other samples, do not present mesopores in addition
c
(5, 12). The raffinate going out of the separation unit enters
the isomerization unit in which the thermodynamic equi-
librium between xylene isomers is reestablished. In the iso-
merization unit, ethylbenzene must be transformed in order
to avoid its build-up in the recycle stream, with a minimum
loss and even if possible a gain in xylenes.
to micropores.
2001 Academic Press
Key Words: ethylbenzene isomerization; disproportionation; Pt;
mordenite; Si/Al ratio.
The isomerization processes differ in the way ethylben-
zene is converted, either by isomerization into xylenes or
by dealkylation or transalkylation into products which can
be easily separated (1, 3, 5–9). In the first type of process
which has the advantage of increasing xylene production,
bifunctional catalysts containing an acidic mordenite in
1 To whom correspondence should be addressed. Fax: +33-(0)5-49-45-
37-79. E-mail: michel.guisnet@univ-poitiers.fr.
402
0021-9517/01 $35.00
c
Copyright
2001 by Academic Press
All rights of reproduction in any form reserved.

ETHYLBENZENE ISOMERIZATION ON BIFUNCTIONAL CATALYSTS
403
combination with platinum supported on alumina are gen-
Sorption isothermsfor nitrogen at 77K over HMOR sam-
erally used (5–9). Indeed, whereas xylene isomerization ples were recorded using a gas adsorption system ASAP
can occur through acid catalysis, ethylbenzene isomeriza- 2010 (Micromeritics), the samples being previously pre-
tion can proceed only through bifunctional catalysis. In treated at 653 K for 2 h under secondary vacuum.
the literature, there is a good agreement with the follow-
The acidity of HMOR samples was characterized by pyri-
ing mechanism for ethylbenzene isomerization: hydrogena- dine adsorption followed by IR spectroscopy using a Nico-
tion of ethylbenzene (EB) into traces of ethylcyclohexenes let MAGNA IR 550 spectrometer. Thin wafers of about
2
(ECHE), acid isomerization of the very reactive ethylcyclo- 8 mg cm were pretreated in situ in the IR cell under air
1
hexenes into dimethylcyclohexenes (DMCHE), and then flow: 60 ml min at 773 K for 12 h and then in a vacuum
3
dehydrogenation of dimethylcyclohexenes into xylenes (X) (10 Pa) at 673 K for 1 h. Pyridine was adsorbed on the
(3, 13–16):
samples at 423 K. The IR spectra were recorded at room
temperature after the activation period and after pyridine
thermodesorption in vacuum (10 Pa) at various temper-
3
*
)
*
)
*
)
EB
ECHE
DMCHE
X.
atures: 423, 523, 623, and 723 K.
Under the conditions used for isomerization, ethylben-
zene (and xylenes) can undergo various secondary re-
actions (3, 4, 6–9, 14, 17): hydrogenation into saturated
naphthenes over the metallic sites, disproportionation,
transalkylation and dealkylation over the acidic sites, and
isomerization and hydrocracking of naphthenes through
bifunctional catalysis. Mordenite was generally chosen as
the acid component because in the narrow channels of this
zeolite disproportionation and transalkylation are limited
by steric constraints (3). Furthermore, modifications of the
acidity and porosity of the mordenite component (e.g., by
Pt/Al2O3 (PtA) was prepared by ion exchange of
-Al2O3 with a hexachloroplatinic acid solution followed
by calcination at 773 K under dry air for 4 h. The Pt con-
tent was equal to 0.53 wt% . The Pt dispersion determined
by H2–O2 titration after reduction under hydrogen flow
(30 ml min 1) at 693 K for 12 h was close to 100% .
The bifunctional Pt/Al2O3 + H-mordenite catalysts were
prepared by milling the mixture of the two components
(75 wt% Pt/Al2O3 + 25 wt% H-mordenite) and then pel-
letizing and sieving at 0.2–0.4 mm.
dealumination) are expected to lead to a significant im- 2. Ethylbenzene Transformation
provement in the selectivity to isomers.
The transformation of ethylbenzene (Fluka puriss prod-
This paper compares the results obtained in a large range
of contact times for ethylbenzene transformation over a se-
ries of intimate mixtures of 0.5 wt% Pt/Al2O3 (75 wt% ) and
of mordenites (25 wt% ) with Si/Al ratios between 6.6 and
180, hence with large differencesin acidityand porosity. The
aim is to understand the effect of these acidity and porosity
characteristics on the rates of ethylbenzene isomerization
and of the secondary reactions as well as on the selectivity
to xylene products.
uct percolated on silica gel before use) was carried out
in a fixed-bed stainless steel reactor under the following
conditions: temperature = 683 K, total pressure = 10 bar,
H2/ethylbenzene molar ratio = 4, WHSV (weight hourly
1
1
space velocity) = 5–200 g of ethylbenzene h
g
of
catalyst. Before use, the catalysts were treated in situ under
1
H2 (10 bar, 600 ml min
g
1) at 753 K for 4 h. Reaction
products were analyzed on-line by FID gas chromatogra-
phy (GC) using two fused silica J&W capillary columns: a
30-m DB-Wax and a 60-m DB-1, the first one for the sepa-
ration of ethylbenzene and xylene isomers and the second
one to obtain the distribution of all the products. Prelimi-
nary experiments have shown that a constant surface area
of the GC peaks was obtained for a time on stream (TOS) of
30–40 min. Therefore, with all the catalysts, the first analysis
of the reaction products was carried out at TOS = 45 min.
EXPERIMENTAL PART
1. Catalysts
Mordenite samples will be called HMOR followed by
the approximate value of the total Si/Al: HMOR6.6 was
supplied by the “Institut Re´gional des Mate´riaux Avance´s”
(IRMA) in Ploemeur, France. HMOR10 was a commer-
cial product from TOSOH, Amsterdam, The Netherlands.
HMOR20 was provided by PQ Zeolites B. V., Amsterdam,
The Netherlands. The other mordenite samples with Si/Al
equal to 30, 45, 90, 120, and 180 were prepared by dealu-
RESULTS AND DISCUSSION
1. Characteristics of the Zeolite Samples
The unit cell formula of all the samples (Table 1) was
mination of HMOR20 by HNO3 solutions at reflux for 4 h determined from elemental analysis and from the values of
in one or two steps with HNO3 normalities ranging from 6 framework Si/Al ratio estimated from 29Si NMR or from
to 14.5. The volume of HNO3 solution/catalyst weight ratio the position of the IR structure bands. With most of the
was equal to 10. Chemical compositions of all the mordenite samples (except HMOR10 and HMOR20), the number of
samples are grouped in Table 1.
extraframework aluminum species (EFAL) per unit cell

404
MOREAU ET AL.
TABLE 1
Characteristics of the Mordenite Samplesa
1
Pore volume (cm3
g
)
Acidity
Sample
Unit cell formula
EFAL
Micro
Meso
NH
+
NL
NH+ /Nth
HMOR6.6
HMOR10
HMOR20
HMOR30
HMOR45
HMOR90
HMOR120
HMOR180
Na0.14H6.16Al6.30Si41.70O96
Na0.01H3.68Al3.69 Si44.31O96
Na0.05H1.95Al2Si46O96
Na0.006H1.344Al1.35Si46.65O96
Na0.004H0.996Al1Si47O96
Na0.003H0.507Al0.51Si47.49O96
Na0H0.35Al0.35Si47.65O96
Na0H0.21Al0.21Si47.79O96
0
0.14
0.19
0.18
0.18
0.18
0.14
0.15
0.14
0.01
0.04
0.11
0.13
0.13
0.13
0.13
0.14
3.20
2.54
1.19
0.63
0.44
0.28
0.23
0.15
0
0.52
0.69
0.61
0.47
0.44
0.56
0.66
0.71
0.35
0.20
0.05
0
0
0
0.18
0.31
0.15
0.04
0.06
0.07
0.075
0
a EFAL: extraframework aluminum species. Nth: number of protonic sites per unit cell estimated from the unit cell formula.
NH+ , NL: number of protonic sites (H+) and of Lewis (L) sites per unit cell able to retain pyridine adsorbed at 423 K.
is negligible in comparison to the number of framework (Figs. 1C, 1D). Another large band also appears at 3420–
aluminum species. Furthermore, with all the samples the 3480 cm 1, which could be related to hydrogen bonding
number of sodium atoms per unit cell is very low (Table 1). between pyridine molecules and bridged hydroxyls located
With HMOR6.6 and 10, a type I isotherm was obtained in the side pockets with a shift to lower frequency values of
from nitrogen adsorption, which indicates that these zeo- the corresponding band.
lites contain essentially micropores. With all the other ze-
The concentrations of Brønsted and Lewis sites were
1
olites, the shape of the isotherms was different, showing estimated from the intensities of the bands at 1545 cm
the presence of mesopores in addition to micropores. In (PyH+) and 1450 cm 1 (PyL) by using the values of extinc-
the series of samples resulting from HMOR20, dealumina- tion coefficients measured in a previous study (19). Table 1
+
tion causes a decrease in the micropore volume and a small shows that NH , the number of protonic sites per unit cell
increase in the mesopore volume.
able to retain pyridine at 423 K as pyridinium ions is always
The following OH bands can be observed in the IR lower (1.4 to 2 times) than the theoretical number given in
1
spectra of mordenite samples: at 3744 cm correspond- the unit cell formula Nth (e.g., for HMOR6.6, Nth = 6.16
1
+
ing to silanol groups in silicon-rich debris, at 3736 cm
and NH = 3.2, Table 1). This could be explained by the
1
to silanols due to framework defects, at 3660 cm to OH
linked to extraframework Al species, and at 3610 cm 1 cor-
responding to bridged OH groups (18). Dealumination of
HMOR20 causes an increase of the silanol bands, the de-
crease of the 3610 cm 1 band, and the disappearance of the
3660 cm 1 band (Figs. 1B, 1C, 1D). With HMOR6.6 and 10
(Fig. 1A), pyridine adsorption results only in a decrease of
1
the 3610 cm band, whereas with the other samples, this
band disappears completely (Figs. 1B, 1C, 1D). This indi-
cates that all the hydroxyls located in the side pockets of
the latter samples interact with pyridine molecules, which
is not the case for those of HMOR6.6 and 10. As was
previously proposed with another series of mordenite sam-
ples (19), this change in the effect of pyridine would be
mainly due to an increase with dealumination in the ac-
cessibility of the protonic sites located in the side pockets.
Figure 1 shows also that part of the silanol groups and of
the extraframework aluminum species interact with pyri-
dine. The interaction of the silanol groups results in the
shift of the corresponding band to lower frequency values
as clearly shown with the more dealuminated samples, e.g.,
1
appearance of a new band at 3680–3700 cm after pyri-
dine adsorption over HMOR120 (curve b), and hence of a
FIG. 1. IR spectra of OH groups in HMOR10 (A), HMOR20 (B),
HMOR 90 (C), and HMOR120 (D). Spectra: (a) activated mordenites,
negative band in curve c corresponding to acidic hydroxyls (b) after pyridine adsorption at 423 K, (c) difference of spectra a–b.

ETHYLBENZENE ISOMERIZATION ON BIFUNCTIONAL CATALYSTS
TABLE 2
405
impossibility for pyridine molecules to interact with
protonic sites of the narrow channels (HMOR6.6 and
HMOR10) or to be adsorbed as pyridinium ions (other
HMOR samples). Whatever the sample, the number of
Lewis sites per unit cell able to retain pyridine adsorbed
Activity (10 3 mol h 1 g 1 Catalyst) of the Acid (A-HMOR) and
of the Bifunctional (PtA-HMOR) Catalysts in Ethylbenzene Trans-
formation (Total), and in Isomerization (Isom), Disproportionation
(Disp), Dealkylation (Dealk), and Production of C8 Naphthenes
(Hydrog) and of Ethylcyclohexane (ECH, in Parentheses)
+
at 423 K (NL) is much lower than NH (Table 1).
It should be noticed that part of the protonic sites of
HMOR6.6, 10, and 20 are able to retain pyridine ad-
sorbed at 723 K, suggesting the existence of very strong
sites. Furthermore, the ratio between the number of pro-
tonic sites able to retain pyridine adsorbed at 623 K and
at 423 K is equal to 0.3–0.35 with HMOR6.6, 10, 20, and
30, and then decreases when the Si/Al ratio increases: 0.16,
0.12, and 0.08 for Si/Al 90, 120, 180 respectively, which sug-
gests a decrease in acid strength.
Activity
Catalysta
Total
Isom
Disp
Dealk
Hydrog (ECH)
PtA-HMOR6.6
PtA-HMOR10
PtA-HMOR20
PtA-HMOR30
PtA-HMOR45
PtA-HMOR90
PtA-HMOR120
PtA-HMOR180
A-HMOR10
A-HMOR90
A-PtA
47.3
111.5
151
60.9
61.0
50.2
39.7
25
8.1
14.5
7.3
11.1
11.4
13.6
9.9
4.0
0
19.6
69.5
115
28.6
24.2
10.4
7.1
1.8
5.7
13.3
3.1
2.6
1.2
1.0
0.6
11.2
1.2
0
17.8 (9)
21.8 (7.5)
15.4 (6.2)
18.1 (7.7)
22.8 (7.6)
25.0 (17.6)
21.7 (12.5)
18.8 (15.6)
0
1.6
2. Activity, Stability of the Catalysts,
and Product Distribution
114.2
13.6
14.4
103
12.4
0
0
0
0
14.4 (14.4)
With all the catalysts, ethylbenzene transformation was
first carried out for approximately 12–15 h at a weight
hourly space velocity (WHSV) of 15 h 1, and then at dif-
ferent WHSV values (hence different contact times) for 1
to 6 h, and then finally at WHSV of 15 h 1. Figure 2 shows
that there is during the first period a decrease in conver-
sion due to coke formation (20) followed by a plateau: the
values of conversion before and after experiments at dif-
ferent contact times (i.e., at 15 and 40 h of reaction) are
very close. Dealumination leads to a slight increase in sta-
bility. Thus, the ratios of activity between 12 h and 45 min
pass from 0.70 with HMOR20 to 0.83 with HMOR180. It
should be remarked that the stability of purely acidic cata-
lysts, e.g., mixture 75 wt% Al2O3 and 25 wt% HMOR10
or HMOR90 (A-HMOR10 and A-HMOR90), is similar
to the stability of the corresponding bifunctional catalysts.
a PtA: 0.53 wt% Pt/ -Al2O3. A: -Al2O3.
Furthermore, the Pt-Al2O3 component of the bifunc-
tional catalysts, actually the Al2O3 (25 wt% )–Pt-Al2O3
(75 wt% ) mixture (A-PtA), is only active for ethylbenzene
hydrogenation (last line, Table 2) and its stability is quasi
perfect.
For each catalyst, the conversion of ethylbenzene at the
plateau was plotted as a function of contact time (taken
here as the reverse of WHSV) and the activity estimated
from the slope of the initial tangents to the curves (Fig. 3).
PtA-HMOR20and PtA-HMOR10are the most active cata-
lysts, the more dealuminated sample (PtA-HMOR180) be-
ing the least active (Table 2).
The distribution of reaction products obtained (on the
stabilized catalysts) for a conversion of 35% over the
PtA-HMOR samples and over A-HMOR10 is reported in
FIG. 2. Conversion of ethylbenzene versus time on stream (TOS) at
various contact times. (A) Conversion on fresh catalysts (PtA-HMOR10
(+), PtA-HMOR20 (᭛), PtA-HMOR90 (1), and PtA-HMOR120 (᭺))
at a weight hourly space velocity (WHSV) of 15 h 1. (B) Experiments at
FIG. 3. Ethylbenzene transformation on PtA-HMOR10 (+), PtA-
HMOR20 (᭛), PtA-HMOR90 (1), and PtA-HMOR120 (᭺). Conversion
(X, % ) versus contact time (h) ( is taken as the reverse of the weight
hourly space velocity, WHSV).
different space velocities (hence different contact times); WHSV equal to
1
5, 30, 60, and 150 h 1. (C) Conversion at WHSV of 15 h
.

406
MOREAU ET AL.
TABLE 3
Distribution (Wt%) of All the Products of Ethylbenzene Transformationa
PtA-
PtA-
PtA-
PtA-
PtA-
A-
HMOR6.6
HMOR10
HMOR20
HMOR90
HMOR120
HMOR10
Conversion
C2
34.2
1.9
34.2
3.2
34.9
6
37.6
1.5
35
1.3
37.9
8.5
C3
1.3
1.7
1.8
1.1
1.1
1.4
C4 (i/n)
C5 (i/n)
C6
2.2 (1.6)
0.9 (2.57)
0.6
3.0 (1.5)
1.3 (1.84)
0.8
2.5 (1.39)
1.0 (2.03)
0.8
2.0 (1.27)
1.0 (1.78)
0.5
2.0 (1.21)
1.1 (1.72)
0.5
1.2 (1.61)
0.4 (4.33)
0.3
N6
3.4
3.9
2.7
2.5
2.1
0.7
N7
0.4
0.5
0.2
0.2
0.2
0.2
N8
14.8
14.3
2.9
37.3
2.8
10.8
22.2
3.1
24.1
3.1
5.6
36.3
2.2
6.6
2.6
10.9
10.1
2.8
47.8
2.7
13.2
8.7
2.5
53.9
2.1
0.8
50.3
1.9
0.4
2.4
Benzene
Toluene
Xylenes
ET
TMB
DEB
DMEB
0.3
14.4
1.0
0.2
19.6
0.8
0
28.8
0.4
0.4
12.5
1.8
0.3
9
1.2
0
29.2
0.3
C10
+
1.5
1.7
2.5
2.2
0.8
2.0
100
100
100
100
100
100
a N6, N7, N8: C6, C7, C8 naphthene. ET: ethyltoluenes. TMB: trimethylbenzenes. DEB: diethylbenzenes.
DMEB: dimethylethylbenzenes.
Table 3. With the bifunctional catalysts, the following (probably hydrogenation of benzene and toluene and then
products were observed:
isomerization), of trimethylbenzenes (disproportionation
of xylenes), and of heavy aromatic products. Furthermore,
part of the C3–C6 alkanes could result from transforma-
tion of ethylene through oligomerization–cracking steps
followed by hydrogenation. However, with the bifunctional
catalysts, there is a good agreement between the exper-
imental values of the yields in benzene and the corre-
sponding values calculated from the yields in DEB, DMEB,
and ethane (produced simultaneously with benzene), which
indicates that the oligomerization–cracking process can be
neglected.
C2–C6 alkanes.
C6–C8 naphthenes, cyclohexane and methylcyclopen-
tane (N6); methylcyclohexane and dimethylcyclopentanes
(N7); ethylcyclohexane and isomers, dimethylcyclohex-
anes, ethylmethylcyclopentanes, and trimethylcyclopen-
tanes (N8).
+
C6–C10 aromatic hydrocarbons: benzene, toluene,
xylenes (X), ethyltoluenes (ET), trimethylbenzenes
(TMB), diethylbenzenes (DEB), dimethylethylbenzenes
+
(DMEB), and C11 (formed only at very high conversions).
With the purely acid catalysts, reactions 1, 5, and 6
do not occur at low conversions and are practically neg-
ligible at high conversions. C3–C6 hydrocarbons which
are observed do not result from naphthene cracking but
from oligomerization–cracking of ethylene resulting from
dealkylation: indeed, the experimental values of the ben-
zene yield are equal to the calculated values, only if it is
considered that C3–C6 hydrocarbons result from ethylene
transformation.
With Pt-Al2O3 (actually A-PtA), ethylcyclohexane is
the only product at low conversions. At high conversions,
trimethyl- and ethylmethylcyclopentanes appear in low
amounts resulting from ethylcyclohexane transformation.
From the comparison of A-PtA, PtA-HMOR, and A-
HMOR, it can be concluded that reactions 2–4 occur
through acid catalysis, hydrogenation through metal cata-
lysis, the second part of reaction 5 (isomerization of N8), re-
action 1 (isomerization), and reaction 6 (cracking) through
bifunctional catalysis (21).
These products result from six main transformations of
ethylbenzene or/and of the primary products:
1. Isomerization of ethylbenzene into ortho-, meta- and
paraxylenes.
2. Disproportionation of ethylbenzene into benzene and
diethylbenzenes.
3. Dealkylation of ethylbenzene into benzene and ethy-
lene totally transformed into ethane (or into longer alka-
nes).
4. Transalkylation of ethylbenzene–xylenes which leads
to ethyltoluene (ET) and toluene or to dimethylethylben-
zene (DMEB) and benzene.
5. Hydrogenation of C8 aromatics (with isomerization of
the products).
6. Cracking of C8 naphthenes (N8) into C3–C6 alkanes.
It should be remarked that additional reactions are nece-
ssary to explain the formation of N6, N7 naphthenes

ETHYLBENZENE ISOMERIZATION ON BIFUNCTIONAL CATALYSTS
407
For each catalyst, the conversion of ethylbenzene
For the formation of these C8 naphthenes (Fig. 5), a quasi
through the six main reactions presented above was plot- plateau in conversion is obtained from short contact times:
ted as a function of contact time (1/WHSV from to 0.005 to 1/WHSV equal to 0.03 h. The plateau is obtained even at
0.2 h). and as a function of total conversion. With the bi- shorter contact times when only ethylcyclohexane is consid-
functional catalysts, the slope of the tangent to the curves at ered (Fig. 6). This suggests that thermodynamic equilibrium
zero conversion was positive for reactions 1, 2, 3, and 5 (the between ethylbenzene and ethylcyclohexane and even the
corresponding products are primary) and equal to zero for isomers of ethylcyclohexane is rapidly established. The con-
the products of reactions 4 and 6 which are therefore sec- version at thermodynamic equilibrium of ethylbenzene into
ondary products (Fig. 4). The secondary nature of these ethylcyclohexane was estimated to be equal to 2% from the
latter products was quite expected since ethylbenzene– values of the standard Gibbs energies of formation (22).
xylenes transalkylation is necessarily consecutive to iso- Values very close to this thermodynamic value were ob-
merization, and cracking is consecutive to naphthene for- tained on pure PtA for contact times greater than 0.03 h.
mation. Another important remark is that whereas the With the bifunctional catalysts, the maximum percentage of
conversion of ethylbenzene through isomerization, dealky- ethylcyclohexane was lower (Fig. 6). This lower value seems
lation, transalkylation, and cracking increases continuously to be due to a rapid isomerization of ethylcyclohexane. In-
with contact time, this is not the case for the conversion into deed, the more significant this isomerization, the lower the
disproportionation products which passes through a maxi- maximum percentage of ethylcyclohexane.
mum with the most active catalysts(i.e., at high conversions)
and especially for the conversion into C8 naphthenes.
The distribution of xylenes depends on contact time, and
hence on ethylbenzene conversion, as shown in Fig. 7. At
FIG. 4. Ethylbenzene transformation on PtA-HMOR10 (+), PtA-HMOR20 (᭛), PtA-HMOR90 (1), and PtA-HMOR120 (᭺). (A) Yield in
isomers (Isom), (B) yield in disproportionation products (Disp), (C) yield in dealkylation products (Dealk), (D) yield in transalkylation products
(Transalk), and (E) yield in cracking products (crack) versus conversion of ethylbenzene (X, % ).

408
MOREAU ET AL.
FIG. 7. Change in the percentage of p-xylene (% para) in the xylene
FIG. 5. Ethylbenzene transformation on PtA-HMOR10 (+), PtA-
mixture as a function of ethylbenzene conversion (X, % ): PtA-HMOR10
(+), PtA-HMOR20 (᭛), PtA-HMOR90 (1), PtA-HMOR120 (᭺), and
PtA-HMOR180 (ᮀ).
HMOR20 (᭛), PtA-HMOR90 (1), and PtA-HMOR120 (᭺). Yield in
C8 naphthenes (N8) versus contact time (h) ( is taken as the reverse of
the weight hourly space velocity, WHSV).
low conversion, the proportion of paraxylene is greater
than that at thermodynamic equilibrium (25.1% (22)),
whereas the proportion of metaxylene is lower. The dif-
ference between experimental and equilibrium values in-
creases with the degree of dealumination (Fig. 7). For con-
versions greater than 10–20% , the xylene distribution is
very close to the thermodynamic equilibrium distribution
(Fig. 7). Disproportionation of ethylbenzene leads also at
low conversion to a mixture of diethylbenzenes (DEB)
richer in para isomer than at equilibrium: e.g., 36.2% para,
58.2% meta, 5.6% ortho with PtA-HMOR6.6 and 68%
para, 31% meta, and 1% ortho with PtA-HMOR180 at zero
conversion instead of 28% para, 52% meta, and 20% ortho
at thermodynamic equilibrium (22, 23). The preferential
formation of para isomers suggests the existence of shape
selectivity effects. However, this proposal is in contradic-
tion with the increase in selectivity to para isomers with the
degree of dealumination, and hence with the accessibility
to the acid sites. A tentative suggestion is that this increase
is due to the significant decrease in disproportionation
(Fig. 4); indeed disproportionation of para isomers on a
HFAU zeolite was shown to be faster than disproportiona-
tion of meta and ortho isomers (24).
3. Influence of the Catalyst Characteristics
on the Initial Rates
It should be remarked that all the catalytic results dis-
cussed here were obtained at the plateau of conversion,
whereas the physicochemical characteristics were estab-
lished on the fresh catalysts. However, the degree of de-
sactivation depends only slightly on the catalyst: the ratio
between ethylbenzene conversion at 12 h and 45 min is be-
tween 0.70and 0.83with allthe bifunctionalcatalysts. More-
over, there is practically no difference between the prod-
uct distributions measured at isoconversion on the fresh
and the stabilized catalysts. Therefore, the correlations be-
tween rates and physicochemical properties which are pre-
sented hereafter can be considered as semiquantitatively
valid.
The initial rates of reactions 1, 2, 3, and 5 were estimated
from the values of the slopes at zero conversion of the tan-
gents to the curves of yields versus contact time. It should
be remarked that there is a large imprecision in the val-
ues of hydrogenation rates because of the rapid establish-
ment of a plateau in conversion corresponding to thermo-
dynamic equilibrium. However, as can be expected, similar
values are found for the rates of formation of C8 naphthenes
with all the bifunctional catalysts and with A-PtA (Table 2).
Moreover, the rates of ethylcyclohexane formation are sim-
ilar on A-PtA and on the bifunctional catalysts with highly
dealuminated HMOR as the acid component for which the
isomerization of ethylcyclohexane is relatively slow.
Table 2 shows that with all the bifunctional catalysts, hy-
drogenation is the fastest reaction. Disproportionation is
also very fast whereas dealkylation is very slow. Finally, the
initial rate of isomerization depends very much on the cata-
lyst. To discuss the effect of the protonic sites on the rates
of reactions 1, 2, and 3 which occur through acid or bifunc-
tional catalysis, the turnover frequency values (TOF) were
FIG. 6. Ethylbenzene transformation on A-PtA ( ), PtA-HMOR10
(+), PtA-HMOR20 (᭛), PtA-HMOR90 (1), and PtA-HMOR120(᭺).
Yield in ethylcyclohexane (ECH) versus contact time (h) ( is taken
as the reverse of the weight hourly space velocity, WHSV).
+
estimated by dividing the initial rates by nH the concen-
tration of protonic sites per gram of catalyst drawn from
pyridine adsorption experiments.

ETHYLBENZENE ISOMERIZATION ON BIFUNCTIONAL CATALYSTS
409
For the acidicreactions2and 3, the TOF valueswere plot-
+
ted versus nH (Figs. 8A, 8B). Figure 8A shows that with the
series of dealuminated samples (Si/Al from 20 to 180), TOF
+
for disproportionation is proportional to nH , and hence
2
+
that the catalyst activity is proportional to (nH ) . This rela-
tion is the one expected for reactions requiring two protonic
sites for their catalysis (25, 26), which would be the case for
disproportionation occurring through diphenylmethane
intermediates (21, 27) (Fig. 9): two different protonic
sites would be involved in steps 1–3 and steps 4–6. With
large-pore zeolites, this mechanism is more likely than the
deethylation–ethylation mechanism which was shown to
occur with average pore size zeolites (28, 29). Therefore, the
density of acid sites seems to be the parameter determining
the disproportionation rate, the acidic strength (higher with
HMOR20 than with the other zeolites) having apparently
+
no effect. However, the relationship between TOF and nH
is not valid for PtA-HMOR6.6 and 10 and A-HMOR10.
The very low values of the TOF found with these catalysts
are probably related to the quasi absence of mesopores
in these zeolites (Table 1). Indeed, it is well demonstrated
that mesopores play a significant role in reactions catalyzed
by monodimensional zeolites allowing a quasi tridimen-
sional diffusion of organic molecules (30, 31). In the
absence of mesopores, reactant transformation would be
FIG. 9. Mechanism of ethylbenzene disproportionation over large-
pore zeolites.
significantly limited by diffusion, and hence would occur
only in the outer part of the crystallites. Only a small part of
the protonic sites would therefore participate in reactions
with therefore very low reaction rates and TOF values.
This part (actually the effectiveness factor of the cata-
lyst) can be roughly estimated from the linear correlation
+
+
between TOF and nH . Thus, with PtA-HMOR10, the nH
value corresponding to the experimental value of TOF
(310h 1) isequalto 30 molg 1 ofcatalyst (Fig. 8A) instead
of the measured value of 220, which means that only 14%
of the protonic sites of HMOR10 would participate in dis-
proportionation ( = 0.14). With PtA-HMOR6.6, the per-
centage of the active protonic sites estimated by the same
way is equal to less than 3% ( = 0.03).
The TOF values obtained for dealkylation with PtA-
HMOR6.6, PtA-HMOR10, and A-HMOR10 are also
lower than for the other samples (Fig. 8B). Further-
more, quasi identical values of TOF are found with the
other samples, except with PtA-HMOR20 which is 2.5
times more active. This could be ascribed to the pres-
ence of very strong acidic sites on HMOR20, able to re-
tain pyridine adsorbed at 723 K. For this monomolecu-
lar reaction which therefore requires only one acid site
for its catalysis, there is as expected no effect of the
density of the acidic sites. It should be remarked that
TOF in dealkylation is lower (3–10 times depending
FIG. 8. Turnover frequency values TOF (h 1) for disproportionation
(A) and for dealkylation (B) versus the concentration of protonic sites per
gram of catalyst, nH+ ( mol g 1 catalyst) able to retain pyridine adsorbed
at 423 K.

410
MOREAU ET AL.
on the sample) than TOF in disproportionation. There- genating/acid function balance. The shape of the curve is
fore, it is most likely that dealkylation requires stronger the one expected from a bifunctional catalysis: at low value
+
acid sites than disproportionation. This could be foreseen of nPt/nH , TOF increases with this ratio, which means that
from the reaction mechanism, disproportionation requiring the limiting step is catalyzed by the metallic sites whereas
the following carbocations which present a high stability, at high values, TOF is constant, i.e., the limiting step is cata-
lyzed by the acidic sites. Curiously, the values obtained with
PtA-HMOR6.6 and 10 seem to be located on the curve.
This can be explained by a compensation effect: indeed,
in the crystallites of these mordenite samples which do
not present mesopores, the transformation of the olefinic
intermediates is, like ethylbenzene disproportionation or
whereas dealkylation requires the formation of a very
unstable ethyl carbenium ion CH3–CH+2 . The high acid
strength required for dealkylation could explain why the
presence of very strong acid sites (e.g., on HMOR20) has a
very positive effect on the TOF in dealkylation. For dispro-
portionation which does not require a high acidic strength,
the effect of this parameter on TOF should be more limited,
and hence would be masked by the large effect of the acid
site density.
The part of the protonic sites of HMOR6.6 and HMOR10
which participate in dealkylation (actually the effectiveness
factor ) can be estimated from the comparison between
their TOF values and those of the other catalysts. As the
strength of the protonic sites of HMOR6.6 and HMOR10
is comparable to that of HMOR20, the TOF value of this
latter catalyst was chosen to estimate . Only 5% of the
protonic sites of HMOR6.6 ( = 0.05) would participate
in ethylbenzene dealkylation, which would be the case for
19% of the sites of HMOR10.
dealkylation, significantly limited by diffusion, and hence
would occur only in the outer part of the crystallites. Thus,
the ratio between the concentrations of accessible Pt atoms
and of accessible protonic sites is greater than the calcu-
+
lated nPt/nH value. On the other hand, the actual TOF
value (i.e., per accessible protonic site) is greater than the
value reported in Fig. 10.
Similar curves are also obtained for the formation
of dimethylcyclohexanes, methylethylcyclopentanes, and
trimethylcyclopentanes which result from ethylcyclohex-
ane isomerization through bifunctional catalysis.
4. Influence of the Zeolite on the Product Distribution
The distributions of the products obtained over the bi-
functional catalysts were compared at various conversions.
Practically no difference was observed between the distri-
butionson the fresh and on the stabilized catalysts, provided
however that these distributionsare compared at isoconver-
sion. The distributions (per reaction) determined for zero
(extrapolated values) and for 35% conversion are given in
Table 4. A large effect of conversion can be observed: a sig-
nificant increase in the percentages of isomerization and
dealkylation, the appearance of cracking and transalky-
lation products at high conversions, and a significant
decrease in disproportionation and hydrogenation. All of
these changes could be expected from the shape of the
curves giving conversions versus contact time. The de-
crease in disproportionation is due to the establishment
of thermodynamic equilibrium and to the secondary trans-
formation of diethylbenzenes (shown by a maximum in
Fig. 4B) by dealkylation into ethylene which is rapidly
hydrogenated and into ethylbenzene which can undergo
other transformations. The plateau in hydrogenation prod-
ucts (Fig. 5) is due to the establishment of thermody-
namic equilibrium for low values of contact time and
hence a low value of conversion. Cracking products re-
sult from secondary transformation of C8 naphthenes, and
transalkylation products from secondary transformation
of xylenes with ethylbenzene. This secondary transforma-
tion of xylenes is slow compared to their rate of for-
mation and the percentage of xylenes increases signifi-
cantly at the expense of those of hydrogenation and of
disproportionation.
As is expected for reactions occurring through bifunc-
tionalcatalysis(32–36), the balance between hydrogenating
and acid functions determines for a large part the catalyst
activity and the TOF of the protonic sites for ethylbenzene
isomerization. Figure 10 shows the change of TOF versus
+
nPt/nH , the ratio between the concentrations of accessible
platinum atoms and of protonic sites able to retain pyridine
adsorbed at 423 K, taken as representative of the hydro-
FIG. 10. Turnover frequency values TOF (h 1) for ethylbenzene iso-
merization versus the ratio between the number of accessible platinum
atoms (nPt) and the concentration of protonic sites (nH+ ) able to retain
pyridine adsorbed at 423 K.

ETHYLBENZENE ISOMERIZATION ON BIFUNCTIONAL CATALYSTS
411
TABLE 4
Distribution (Wt%) of the Reaction Products Extrapolated at Zero Conversion (A) and at 35%
Conversion of Ethylbenzene (B) in Isomerization (Isom), Disproportionation (Disp), Dealkyla-
tion (Dealk), Cracking Products (Crack), and Production of C8 Naphthenes (Hydrog)a
Catalyst
Isom
Disp
Dealk
Crack
Hydrog
Transalk
A
PtA-HMOR6.6
PtA-HMOR10
PtA-HMOR20
PtA-HMOR30
PtA-HMOR45
PtA-HMOR90
PtA-HMOR120
17.1 (27.4)
13.0 (16.2)
4.8 (5.3)
18.2 (25.9)
18.7 (29.9)
27.1 (54.0)
24.9 (55.0)
41.4
62.3
76.2
47.0
39.6
20.7
17.9
3.8
5.1
8.8
5.1
4.3
2.4
2.5
0
0
0
0
0
0
0
37.6
19.6
10.2
29.7
37.4
49.8
54.7
0
0
0
0
0
0
0
B
PtA-HMOR6.6
PtA-HMOR10
PtA-HMOR20
PtA-HMOR30
PtA-HMOR45
PtA-HMOR90
PtA-HMOR120
41.4 (46.3)
28.4 (31.6)
6.9 (7.3)
31.0 (34.1)
34.3 (37.6)
53.1 (60.1)
60.1 (67.9)
29.7
37.1
47.1
32.9
34.4
21.2
14.8
8.3
12.5
28.1
15.5
11.7
5.3
5.8
7.1
8.2
8.2
6.6
4.8
5.1
10.6
10.2
5.4
9.0
8.8
4.2
4.4
4.4
3.4
4.2
4.0
3.2
11.6
11.5
5.3
a The percentage of isomers indicated in parentheses was calculated without considering C8 naphthenes
in the products.
+
Whatever the conversion, a large increase in the selec- and nH and from the TOF in dealkylation. Whatever the
tivity to xylenes can be observed when the Si/Al ratio of case, it can be concluded that the particular selectivity of
the zeolite component passes from 20 to 120. Thus, at 35% PtA-HMOR6.6 and 10 catalysts (Table 4B) is due to the
conversion, the selectivity in isomerization passes from 7% participation in ethylbenzene transformation of only the
to 60% if all the products are considered, or from 7.5% protonic sites of the outer part of the zeolite crystallites as
to 68% if C8 naphthenes, which in the isomerization unit was previously proposed by Fernandes et al. (7).
are recycled with the C8 aromatic cut, are not considered.
This increase is due to a significant decrease in the selectiv-
ity to disproportionation products, from 47 to 15% and to
CONCLUSIONS
dealkylation products, and from 28 to 5% (Table 4). A slight
From this comparison of the activity and selectivity of this
decrease in the selectivity to cracking products is also ob-
series of PtAl2O3 (75 wt% )–mordenite (25 wt% ) catalysts,
served whereas the selectivity to transalkylation products
the following conclusions can be drawn:
remains practically constant.
Curiously, the selectivityto xylenesofPtA-HMOR10and
especially of PtA-HMOR6.6 is higher than the selectivity of
PtA-HMOR20. This can be related to the participation in
the various reactions of only the protonic sites of the shell
of the crystallites owing to diffusion limitations in these
mordenite samples which have practically no mesopores
(7). At 35% conversion, the selectivity of PtA-HMOR6.6
1. With all the catalysts, ethylbenzene undergoes, in
addition to the desired isomerization reaction, three di-
rect transformations: hydrogenation into C8 naphthenes
which is very limited by thermodynamic equilibrium, dis-
proportionation, and dealkylation. Secondary reactions of
transalkylation ethylbenzene–xylenes and of naphthene
cracking are also observed.
+
(nPt/nH = 0.07) is comparable to that of PtA-HMOR45,
2. With all the catalysts, hydrogenation and dispropor-
tionation are the fastest reactions and dealkylation is very
slow;the rate ofisomerization dependsverymuch on the ze-
olite component. For hydrogenation, thermodynamic equi-
librium is established at relatively short contact times.
3. Except for the catalysts having HMOR6.6 and 10
as acid components, the activity in disproportionation is
proportional to the square of the concentration in pro-
tonic sites, which suggests the participation of two pro-
tonic sites in the bimolecular transformation; the activity in
+
i.e., of a sample with nPt/nH equal to 0.42. Therefore, if all
the protonic sites of this latter catalyst participate in isomer-
ization, it would be the case for only 17% (0.07 100/0.42)
of those of PtA-HMOR6.6. PtA-HMOR10 has an isomer-
ization selectivity close to that of a sample having a Si/Al
ratio between 20 and 30. Therefore, 30% of the acid sites of
PtA-HMOR10 would participate in ethylbenzene isomer-
ization. These values are greater than those calculated from
the linear correlation between TOF in disproportionation

412
MOREAU ET AL.
dealkylation depends on the concentration in protonic sites 12. Franck, H.-G., and Stadelhofer, J. W., “Industrial Aromatic Chemistry,
Raw Materials, Process, Products.” Springer-Verlag, Berlin, 1988.
13. Gnep, N. S., and Guisnet, M., Bull. Soc. Chim. Fr. 5–6, 429 (1977).
14. Gnep, N. S., and Guisnet, M., Bull. Soc. Chim. Fr. 5–6, 435 (1977).
15. Ro¨bschla¨ger, K. H., and Christoffel, E. G., Ind. Eng. Chem. Prod. Res.
and also on their acid strength; the change of the activ-
ity in ethylbenzene isomerization of the protonic sites (the
turnover frequency) with the balance between the hydro-
genating and acid activities is the one expected from a bi-
functional mechanism.
4. The behavior of the catalysts with HMOR6.6 and 10
acid components is particular. This can be related to the ab-
sence of mesopores in the mordenite crystallites, the conse-
quence being that acidic reactions occur only in the shell of
Dev. 18, 347 (1979).
16. Nitta, M., and Jacobs, P., in “Catalysis by Zeolites” (B. Imelik,
C. Naccache, Y. Ben Taarit, J. C. Vedrine, G. Coudurier, and H.
Praliaud, Eds.), Stud. Surf. Sci. Catal., Vol. 5, p. 251. Elsevier,
Amsterdam, 1980.
17. Hsu, Y. S., Lee, T. Y., and Hu, H. C., Ind. Eng. Chem. Res. 27, 942
(1988).
the crystallites. The effectiveness factor for HMOR6.6 and 18. Guisnet, M., Ayrault, P., and Datka, J., Polish J. Chem. 71, 1445
(1997).
10 could be as low as 0.03–0.15 and 0.15–0.30, respectively.
5. The bifunctional catalysts which are the more selec-
tive in ethylbenzene isomerization are those for which the
19. Guisnet, M., Ayrault, P., and Datka, J., Polish J. Chem. 71, 1455
(1997).
20. Moreau, F., unpublished results.
limiting step is the acid isomerization of ethylcyclohexene
intermediates. This can be obtained for high values of the
ratio between the concentrations of accessible Pt sites and
of protonic sites accessible by reactant molecules, i. e., for
bifunctional catalysts containing a very dealuminated mor-
denite and also for bifunctional catalysts containing mor-
denite without mesopores.
21. Silva, J. M., Ribeiro, M. F., Ribeiro, F. R., Benazzi, E., and Guisnet,
M., Appl. Catal. 125, 15 (1995).
22. Stull, D. R., Westrum, E. F., and Sinke, G. C., “The Chemical Thermo-
dynamics of Organic Compounds.” Wiley, New York, 1969.
23. Koeding, W. W., J. Catal. 95, 512 (1985).
24. Morin, S., Gnep, N. S., and Guisnet, M., J. Catal. 159, 296 (1996).
25. Guisnet, M., Avendano, F., Bearez, C., and Chevalier, F., J. Chem. Soc.,
Chem. Commun. 336 (1985).
26. Guisnet, M., Acc. Chem. Res. 23, 392 (1990).
27. Karge, H. G., Ladebeck, J., Sarbak, Z., and Hatada, K., Zeolites 2, 94
(1982).
REFERENCES
28. Amelse, J. A., in “Proceeding, 10th North American Meeting of the
Catalyst Society” (J. Ward, Ed.), Stud. Surf. Sci. Catal., Vol. 38, p. 165.
1988.
29. Silva, J. M., Ribeiro, M. F., Ribeiro, F. R., Benazzi, E., and Guisnet,
M., Appl. Catal. 125, 1 (1995).
30. Gnep, N. S., Roger, P., Cartraud, P., Guisnet, M., Juguin, B., and
Hamon, C., C. R. Acad. Sci. Paris, Se´rie II, 309, 1743 (1989).
31. Meima, G. R., van der Aalst, M. J. M., Samson, M. S. V., Garces, J. M.,
and Lee, J. G., in “Proceedings, 9th International Zeolite Conference”
(R. van Ballmoos, J. B. Higgins, and M. M. J. Treacy, Eds.), Vol. II,
p. 327. Butterworths Heinemann, Boston, 1992.
32. Weitkamp, J., Ind. Eng. Chem. Prod. Res. Dev. 21, 550 (1982).
33. Guisnet, M., and Pe´rot, G., in “Zeolites: Science and Technology”
(F. R. Ribeiro, A. E. Rodrigues, L. D. Rollmann, and C. Naccache,
Eds.), NATO Series E, Vol. 80, p. 397. Martinus Nijhoff, The Hague,
1984.
34. Weitkamp, J., and Ernst, S., in “Guidelines for Mastering Prop-
erties of Molecular Sieves. Relationship between the Physico-
chemical Properties of Zeolitic Systems and Their Low Dimen-
sionality” (D. Barthomeuf, E. G. Derouane, and W. Ho¨lderich,
Eds.), NATO ASI Series B, Vol. 221, p. 343. Plenum, New York,
1990.
35. Guisnet, M., Alvarez, F., Giannetto, G., and Pe´rot, G., Catal. Today 1,
415 (1987).
36. Alvarez, F., Ribeiro, F. R., Pe´rot, G., Thomazeau, C., and Guisnet, M.,
J. Catal. 162, 179 (1996).
1. Beck, J. S., and Haag, W. O., in “Handbook of the Heterogeneous
Catalysis” (G. Ertl, H. Kno¨zinger, and J. Weitkamp, Eds.), Vol. 5,
p. 2136. VCH, Weinheim, 1997.
2. Chen, N. Y., Garwood, W. E., and Dwyer, F. G., in “Shape
Selective Catalysis in Industrial Applications” (H. Heinemann,
Ed.), Chemical Industries Vol. 36, p. 203. Dekker, New York,
1989.
3. Olson, D. H., and Haag, W. O., ACS Symposium Series 284, Chap. 14,
p. 275. Am. Chem. Soc., Washington, DC, 1984.
4. Ribeiro, M. F., Travers, C., Raatz, F., Marcilly, C., and Ribeiro,
F. R., in “Zeolites: Facts, Figures, Future” (P. A. Jacobs and R. A. van
Santen, Eds.), Stud. Surf. Sci. Catal., Vol. 49, Part B, p. 1349. Elsevier,
Amsterdam, 1989.
5. Alario, F., Barraque´, M., Marcilly, C., in “Ge´nie des Proce´de´s,” Tome
J3, p. J 5920. Techniques de l’Inge´nieur, Paris, 1996.
6. Sie, S. T., de Vries, A. F., Mesters, C. M. A. M., Boon, A. Q. M.,
Bottenberg, K., and Trautrims, B., Erdo¨l Erdgas Kohle, November,
463 (1996).
7. Fernandes, L. D., Monteiro, J. L. F., Sousa-Aguiar, S. F., Martinez, A.,
Corma, A., J. Catal. 177, 363 (1998).
8. Guisnet, M., l’Actualite´ Chimique, April, 9 (1989).
9. John, H. H., Neubauer, H. D., and Birke, P., Catal. Today49, 211(1999).
10. Baraque´, M., l’Actualite´ Chimique, Octobre, 18 (1997).
11. Gonzalez, H., Rodriguez, A., Cedeno, L., Ramirez, J., and Aracil, J.,
Ind. Eng. Chem. Res. 35, 3964 (1996).