The influence of the external acidity of H-ZSM-5 on its shape selective properties in the disproportionation of ethylbenzene

Article abstract of DOI:10.1006/jcat.1997.1715

The shape selectivity of H-ZSM-5 in the disproportionation of ethylbenzene was investigated, using a set of samples with the same Si/Al ratio, but different particle sizes in the range from 0.1 to 80 μm. The number of external acid sites of each sample was measured gravimetrically by the adsorption of 2,6-dimethylpyridine. The data were correlated with the results of catalytic experiments. Conversion and product distribution are strongly dependent on the external acidity which in turn correlates well with the particle size. An estimate for the diffusion coefficient could be obtained by fitting the effectiveness factor for the different particle sizes.

Full text of DOI:10.1006/jcat.1997.1715

JOURNAL OF CATALYSIS 170, 46–53 (1997)
ARTICLE NO. CA971715
The Influence of the External Acidity of H-ZSM-5 on Its Shape Selective
Properties in the Disproportionation of Ethylbenzene
Sabine Melson and Ferdi Schu¨th
Institut fu¨r Anorganische Chemie, Johann Wolfgang Goethe-Universita¨t Frankfurt, Marie Curie-Strasse 11, 60439 Frankfurt, Germany
Received February 11, 1994; revised March 12, 1997; accepted March 28, 1997
In xylene isomerization over ZSM-5 it has been observed
that the crystal size has a more prominent effect on the
shape selectivity than the site density (7). The enhanced
formation of p-xylene is attributed to the occurrence of
diffusion limitations in the Al,B-MFI sample. Kaeding (8)
states that for the disproportionation of ethylbenzene over
ZSM-5 samples with large crystal sizes a higher concen-
tration of para-diethylbenzene is formed than over smaller
crystals. He attributes this effect to the lower amount of ex-
ternal surface sites which are responsible for nonselective
reactions on the lager crystals as compared to the small crys-
tal size sample. Similar results were obtained by Shiralkar
et al. (9) and Beschmann et al.(10) in their studies of the
methylation of toluene and Bhat et al. (11) for the ethyla-
tion of toluene. Also Csicsery (12) explains such an effect
The shape selectivity of H-ZSM-5 in the disproportionation of
ethylbenzene was investigated, using a set of samples with the same
Si/Al ratio, but different particle sizes in the range from 0.1 to
80 ꢀm. The number of external acid sites of each sample was mea-
sured gravimetrically by the adsorption of 2,6-dimethylpyridine.
The data were correlated with the results of catalytic experiments.
Conversion and product distribution are strongly dependent on the
external acidity which in turn correlates well with the particle size.
An estimate for the diffusion coefficient could be obtained by fit-
c
ting the effectiveness factor for the different particle sizes.
ꢀ 1997
Academic Press
INTRODUCTION
Para-substituted dialkylbenzenes serve as important by the influence of the external surface: on small crystals
raw materials in a number of industrial processes. Such with a high external surface area, the diffusion-controlled
compounds can be produced using the shape selective shape-selective transformations proceed at a lower rate
properties of zeolites such as ZSM-5. There are several than the unrestricted surface reactions. Therefore, the para-
investigations of the influence of the ZSM-5 crystal size selectivity is increased for alkylation when the external
on conversion and para-selectivity in reactions of alkylated crystal surface is coated with residues (13). The problem
benzenes. Also the effect of the framework aluminum con- of the nonselective external surface can be reduced by in-
tent has been studied. However, there are no reports about activation of the external surface which results in a high
the measurement of the external acidity and its correlation para-selectivity, as for instance Hibino et al. (14) and Bhat
with activity and selectivity in the above-mentioned reac- etal.(11) observed for the alkylation oftoluene over ZSM-5.
tions. Paparatto et al. (1) concluded that isomerization of
One problem of previous studies (1–14) with the excep-
para-alkylated toluene over ZSM-5 and ZSM-11 proceeds tion of the recent work of Beschmann et al. (10), Arsenova
on external acid sites, and that decreasing the concentra- et al. (15), and Bhat et al. (11) was the fact that no well-
tion of external acid sites suppresses the activity, resulting defined sample set of more than two samples with different
in enhancement of para-alkylated product selectivity. The crystal sizes and the same Si/Al ratio was available. In addi-
authors proposed a two-step mechanism: in the first step, tion, there is a lack of data on the amount of acid sites on the
the para-isomer is formed by toluene alkylation with high external surface which was never measured in these studies.
selectivity inside the zeolite channels. In the second step In the literature, conversion and shape selectivity have been
the isomerization of the three isomers toward the thermo- correlated with the crystal size instead of the external acid-
dynamic equilibrium distribution proceeds on the external ity because the latter data were not available. Since the alu-
surface only. In addition to these suggestions, Lonyi et al. minum distribution over a crystal might be inhomogeneous
(2) proposed that the primary para-product isomerizes first (16), it is important to have also independent information
to meta- and then to ortho-diethyltoluene. Increasing con- on the number ofexternalacid sites. The aluminum distribu-
version is often correlated with decreasing para-selectivity tion within ZSM-5 crystals resulting from different synthe-
(3–6), since at high conversion consecutive isomerization sis systems has been investigated by Althoff et al. (17) with
reactions can occur easily.
electron microprobing. It has been shown that in samples
46
0021-9517/97 $25.00
c
Copyright ꢀ 1997 by Academic Press
All rights of reproduction in any form reserved.

EXTERNAL ACIDITY OF H-ZSM-5
47
synthesized from tetrapropylammonium (TPA)-containing NH4NO3 solution for 24 h, washed, and dried at 394 K.
systems there is always a pronounced enrichment of alu- By calcination at 774 K for 10 h the ammonium form was
minum in the crystal rim, the extent of which, however, can converted to H-ZSM-5. XRD of the samples showed 100%
vary. Information on the amount ofacid siteson the external crystalline MFI-structure. No amorphous material was ob-
surface is thus necessary to fully understand catalytic data. served.
We chose the adsorption of 2,6-dimethylpyridine as a probe
molecule to measure this property. This amine has been
used before as an indicator of acidity in FTIR-spectroscopic
experiments (18–20), and should be too bulky to enter the
ZSM-5 pore system. A similar approach had been used al-
ready several years ago by Karge and co-workers who used
2,6-di-tert-butylpyridine to analyze the external acidity of
other zeolite types (21). They found that catalytic activity
in the dehydration of cyclohexanol correlated almost lin-
early with the absorbance of the IR-band characteristic for
the di-tert-butylpyridinium ion for mordenites and clinop-
tilolites. Recently another bulky probe, 4-methylquinoline,
was used to analyze the external acidity of ZSM-5 (22).
However, one should bear in mind in this connection
that the term “external surface” or “external acidity” is not
easy to define on a molecular level. The approach chosen
here is clearly an operational definition, i.e., it is assumed
that 2,6-dimethylpyridine is too bulky to enter the pores
which is based on estimates using van der Waals radii, and
then external surface is defined as that part of the sample
which can react with the probe under the conditions of the
sorption experiment. The probe was chosen so that it re-
sembles the molecules under investigation in the catalytic
experiments. We chose dimethylbenzene instead of diethyl-
benzene because this molecule proved to be more suitable
for IR-spectroscopic analysis, the results of which will be
reported in a forthcoming communication.
In order to clarify the influence of the external acidity in
a reaction where shape selectivity has been observed, we
used a well-defined set of four H-ZSM-5 samples with a
Si/Al ratio of 35 in a crystal size range from 0.1 to 80 ꢀm,
resulting from an alkali-based and an alkali-free synthesis
system. The crystals have a good morphology with little
intergrowth and a narrow crystal size distribution and are
thus very well suited for this investigation. For all samples
the amount of the external acid sites has been measured and
correlated with the catalytic performance of the materials
in the disproportionation of ethylbenzene.
Characterization. The Si/Al ratios were measured by
X-ray fluorescence (XRF). The crystal size was determined
from scanning electron micrographs shown in Fig. 1. In or-
der to obtain a direct measure of the external surface of the
crystals, a method following the procedure of Grillet et al.
(26) was employed: The pores of the ZSM-5-samples were
first filled bypreadsorption ofn-nonane. For thispreadsorp-
tion the zeolites had been evacuated to 1 ꢁ 10ꢂ5 Torr and
activated at 623 K for 8 h. At room temperature n-nonane
was dosed for 24 h. These samples were transfered to a
Micromeritics ASAP 2000 volumetric sorption system and
activated at 363 K for 3 h. Variation of the activation condi-
tions for the preloaded samples did not change the results
of the analysis. By N2 adsorption at 77 K the nonmicropore
surface of the samples could be measured. This procedure,
however, was only possible for the samples with small crys-
tal size (see below).
Adsorption measurementsofthe amineswere performed
employing a magnetic suspension balance (27). The sam-
ple was activated by evacuation to 5 ꢁ 10ꢂ6 Torr and then
heated at 3 K/min to 623 K and kept at that temperature
for 8 h. After cooling to room temperature the sample was
exposed to pyridine at 1 ꢁ 10ꢂ3 Torr for 4 h. After this expo-
sure time no further uptake of pyridine could be observed.
Physisorbed pyridine was removed by heating under vac-
uum at 400 K for 2 h. Readsorption after this heating did
not result in additional irreversible uptake. The results ob-
tained agree very well with the Si/Al ratios determined via
XRF so that complete saturation of the samples can be as-
sumed.
The number of the external acid sites was measured
gravimetrically by using instead of pyridine the probe 2,6-
dimethylpyridine. Adsorption and removal of physisorbed
2,6-dimethylpyridine was carried out under the same con-
ditions as for pyridine. In parallel FTIR spectroscopic mea-
surements it was shown that evacuation for 2 h at 400 K
was sufficient to remove physisorbed pyridine and 2,6-
dimethylpyridine. Both Brønsted and Lewis acid sites were
present in all samples on the internal and on the external
surface. Quantification of the fractions, however, is difficult.
EXPERIMENTAL
Materials. A collection of well-defined H-ZSM-5 sam- Investigations are under way to obtain reliable estimates on
ples with identical Si/Al ratio, but different crystal sizes was the relative amounts of the different species.
synthesized with organic templates following well-known
Ethylbenzene disproportionation was carried out on
procedures in alkali-containing (samples A, B, C) (23, 24) the powdered samples in a fixed-bed quartz reactor at
and alkali-free (sample D) (25) systems. The crystals exhibit 523 K with on-line GC analysis. No pellets were pressed
a regular morphology and show little intergrowth (Fig. 1). in order to avoid crushing of the crystals. The samples
The zeolites were calcined in air by heating at 3 K/min to were activated for 12 h in N2 atmosphere at 723 K prior
774 K for 20 h and then exchanged four times with 1 M to experiments. Selectivities were compared for identical

¨
48
MELSON AND SCHUTH
FIG. 1. Scanning electron micrographs of samples A–C and optical micrograph of sample D.
conversions by adjusting the space velocity, and the conver-
Adsorption of 2,6-dimethylpyridine was carried out to
sionsfor different sampleswere compared at identicalspace analyze the number ofacid sitesaccessible to stericallymore
velocities.
demanding di-substituted benzene derivatives which are as-
sumed to be formed mainly on the external surface in alky-
lation and isomerization reactions. The results of these ex-
periments are also summarized in Table 1. As for pyridine,
one acid site was assumed for every 2,6-dimethylpyridine
molecule adsorbed. A clear dependence between crystal
size and the amount of 2,6-dimethylpyridine adsorbed is ob-
vious from the data. It should be pointed out here that this
dependence is not a kinetic effect due to higher transport
RESULTS AND DISCUSSION
The Si/Al ratio of all samples was determined by X-ray
fluorescence to be 35. Thisisin excellent agreement with the
gravimetrically measured uptake of pyridine of the zeolites,
if a stoichiometry of one pyridine molecule per aluminum
atom is assumed (see Table 1).

EXTERNAL ACIDITY OF H-ZSM-5
49
FIG. 1—Continued
limitations on the larger crystals. Saturation of the sam- sidered. Calculated external surface areas assuming the ex-
ple with 2,6-dimethylpyridine was usually reached within ternal surface to be well approximated by the geometric
15min. Long-term experimentsover severaldaysdid not re- surface area of rectangular crystals are listed in Table 1.
sult in any additional uptake of 2,6-dimethylpyridine which For sample A a spherical shape was assumed as the SEM
proves that equilibrium had been reached. Crystals larger photographs suggest. As can be derived from Table 1, the
than 20 ꢀm with a correspondingly small external surface concentration of acid sites on the external surface is larger
did not show any weight gain beyond the sensitivity limit of by a factor of around 10 than in the bulk, assuming a pene-
the balance and thus could not be analyzed gravimetrically. tration depth of 2,6-dimethylpyridine of one unit cell. There
The amount of 2,6-dimethylpyridine adsorbed is surpris- are three possible explanations for this effect: (i) the cal-
ingly high, if just the geometrical external surface is con- culated geometrical surface area could be too small due to

¨
MELSON AND SCHUTH
50
However, even taking into account that the measured
TABLE 1a
Summary of the Results of Pyridine Adsorption on all Samples
external surface area is larger than the calculated surface
area based on a smooth crystal surface, the amount of alu-
minum analyzed by 2,6-dimethylpyridine adsorption is still
higher by a factor of about 4 than that expected from the
bulk Si/Al ratio. Due to the fact that no additional uptake of
2,6-dimethylpyridine is observed even for long times, pene-
tration of 2,6-dimethylpyridine into the pore system seems
to be rather improbable. There is a report in the literature
that 2,6-dimethylpyridine can be formed over ZSM-5 (28)
and the formation is assumed to be taking place in the pore
system. However, there is no information on the crystal
size, and the reaction was carried out at 723 K where there
might indeed be some penetration of the molecules into
the pore system. Under the conditions of the adsorption
experiments, however, we exclude this possibility. The high
concentration of acid sites on the external surface can be
better explained byan enrichment ofaluminum towardsthe
rims of the crystals. Since the crystals used were all synthe-
sized using TPA as the template, enrichment of Al is to be
expected. It can amount to a factor of 5 or more (16, 29, 30)
which, together with the presence of surface roughness and
cracks, explains the high concentration of acid sites mea-
sured by 2,6-dimethylpyridine.
If the density of acid sites on the external surface is cal-
culated from the amount adsorbed and the measured ex-
ternal surface areas, one arrives at about three sites per
square nanometer which in turn means that there is only
a space of about 0.3 nm2 per adsorbed molecule of 2,6-
dimethylpyridine. This is only about twice the space re-
quirement assumed for a nitrogen molecule adsorbed on
surfaces. Space requirements for the 2,6-dimethylpyridine
can most probably not be based on the kinetic diameter. For
pyridine, on-top adsorption on acid sites could be demon-
strated (31), so that the space requirement is probably best
described by the projection of the shape of the molecule
along the C2 axis. This is difficult to estimate, but 0.3 nm2
should be sufficient to accommodate a 2,6-dimethylpyridine
molecule, if the estimate is simply based on van der Waals
radii (approx. 0.14 nm2). However, it should be pointed out
that substantially higher surface concentrations of acid sites
than measured here could probably not be detected by this
probe molecule due to packing constraints.
Pyridine
adsorbed
(mg/gcat)
Acid site
concentration
(ꢀmol/g)
Sample
Crystal size (ꢀm)
Si/Al ratio
A
B
C
D
0.1
5–10
15
35
35
35
35
36.2
36.5
36.0
36.1
458
462
456
457
80
Note. Si/Al ratio measured with X-ray fluorescence. One pyridine
molecule assumed per acid site.
the presence of surface roughness and cracks, (ii) an enrich-
ment of aluminum toward the edges of the crystals could
be present, and (iii) 2,6-dimethlypyridine might penetrate
deeper into the crystal than one unit cell.
In order to obtain a better estimate of the external sur-
face area, N2 adsorption measurements were performed on
samples for which the micropore system had been blocked
by preadsorption of nonane. Only samples A and B could
be analyzed using this method, since the external surface
area of the samples with larger crystal size (C and D) was
too small to be reliably determined using this method. The
external surface area obtained using this method is a factor
of 3 larger than the calculated surface areas (52 m2/g for
sample A and 2.7 m2/g for sample B). The N2 isotherms of
the samples completely filled with nonane showed no in-
dication of the presence of micropores which proves that
only the external surface has been measured in these exper-
iments. The reason for the measured external surface being
higher by a factor of 3 than the calculated external surface is
most probably due to cracks formed during the calcination
of the materials (clearly visible in the optical microscope
on the large crystals) and to surface roughness.
TABLE 1b
Summary of the Characterization of the External Acidity
Concentration
of external External/
Calcd
external
Measured
external
2,6-DMP
adsorbed
acid sites
(ꢀmol/g)
internal surface area surface area
acidity
Sample (mg/gcat
)
(m2/g)
(m2/g)
The difference between the number of external acid sites
in the four samples is strongly reflected in their catalytic
behavior. The conversion of ethylbenzene to benzene and
the isomeric diethylbenzenes strongly increases with an in-
crease of the number of external acid sites (Fig. 2). This is in
A
B
C
D
29.7
2.0
0.8
ꢃ
277
18
8
ꢃ
1.53
0.04
0.02
ꢃ
16.5
1.1
0.4
52.0
2.7
ꢃ
0.1
ꢃ
Note. The number of external acid sites has been calculated from the agreement with published data mentioned in the Introduc-
adsorption of 2,6-dimethyl-pyridine (2,6-DMP) assuming one molecule
per acid site. For the calculation of the external surface area rectangular
crystals were assumed (for the 0.1 ꢀm crystal spheres). The measured
external surface area was determined by BET analysis with nitrogen after
tion in which the particle size instead of the amount of ex-
ternal acid sites had been correlated with the activity of the
samples in reactions with different disubstituted benzenes.
On the other hand, the selectivity decreases with increas-
ing external acidity (Fig. 3). This is in agreement with Chen
pore filling with nonane.
ꢃ
Nonmeasurable values.

EXTERNAL ACIDITY OF H-ZSM-5
51
FIG. 2. Dependence of ethylbenzene conversion on the concentra-
tion of external acid sites at constant Treactor = 523 K, WHSV = 0.06 hꢂ1
.
FIG. 4. Selectivity to the isomers at different conversions over sample
A. ᭺, molar ratio p/m; ᭿, molar ratio p/o.
Concentration of external acid sites for sample D was estimated assuming
an identical density as for sample C.
(13, 34). The product distribution is strongly dependent on
the space velocity. Figure 4 shows the concentration of the
products formed at different conversions achieved by vary-
ing the space velocity over sample A. The thermodynamic
equilibrium between para- and meta-diethylbenzene is al-
ready reached at high space velocities while the conversion
is still low. Decreasing the space velocity leads to the for-
mation of increasing amounts of the ortho-isomer.
et al. (32, 33) who observed that the para-selectivity for re-
actions of toluene depends on the crystal size. Large crys-
tals which have a correspondingly small external surface
showed a high para-selectivity. The conversions and selec-
tivities observed here are not linearly correlated with the
amount of external acid sites, since both are governed by
the interplay between reaction and diffusion, resulting in
nonlinear dependences.
Two alternative explanations have been suggested to ra-
tionalize the selectivity in the isomerization of disubstituted
benzenes, primarily for the xylene isomerization. The first
was put forward by Wei (35) and Weisz et al. (36, 37). These
authors assume that all isomers can enter the pore system.
Shape selectivity then depends on the length of the intraze-
olitic diffusion path and is therefore governed by the rela-
tion between intrinsic reactivities and diffusion coefficients
of the different isomers. This in turn is dependent on the
structure type, the modification, and the crystal size. In or-
der to explain the lack ofpara-selectivityat high conversion,
Wei (35) proposed product reentry into the zeolitic channel
system for further reaction.
Para- and meta-diethylbenzene are the only products
observed in our study for low values of the external
acidity. The formation of m-diethylbenzene increases and
that of p-diethylbenzene decreases with increasing exter-
nal acidity. o-Diethylbenzene can only be detected by
GC when the thermodynamic equilibrium between p- and
m-diethylbenzene has already been reached. (Equilibrium
values calculated by Kaeding (8) for diethylbenzene at
523 K are p/m = 0.57; p/o = 1.95). Similar results were ob-
served during the alkylation of ethylbenzene with ethanol
which results in the primary formation of p-diethylbenzene
The other approach, put forward by Fraenkel (38),
Derouane (39), and Nayak and Riekert (40), assumes that
there are para-selective events in the pore system due to
diffusional constraints while external acid sites are respon-
sible for non-selective transformations.
Our results suggest that the meta- and ortho-isomers are
produced mainly on the external surface of the crystals,
which is in agreement with data of Karge et al. (41) who re-
ported that o-diethylbenzene cannot enter the ZSM-5 pore
system. If a major fraction of the acid sites is accessible
for the reactants as well as for the product, isomerization
until equilibrium is reached is easily understandable. The
increase of ortho-diethylbenzene formation with decreas-
ing space velocities can be rationalized by the fact that
the reactants are in contact with the catalyst for a longer
time, thus allowing the system to approach the equilibrium
FIG. 3. Selectivity to para-, meta-, and ortho-diethylbenzene in ethyl-
benzene disproportionation over samples A–D in dependence on the con-
centration of external acid sites at a conversion of 4% and a reactor tem-
perature of 523 K. ᭺, molar ratio p/m; ᭿, molar ratio p/o. The letters
indicate the sample.

¨
MELSON AND SCHUTH
52
concentrations by isomerization on the non-shape-selective
external surface. Lonyi et al. (2) discuss the formation of
meta- and ortho-ethyltoluene in the ethylation of toluene in
terms of a subsequent isomerization of the p-isomer formed
first. Our data indicate that the isomerization of the primary
product p-diethylbenzene proceeds on the external surface
in a sequential reaction first to the meta-isomer, and finally
from the meta- to the ortho-isomer. With increasing exter-
nal acidity sufficient meta-diethylbenzene is produced from
the p-diethylbenzene formed shape selectively in the pore
system, which then reacts further to the ortho-isomer via a
1,2-alkyl-shift reaction. Allen and Yats (42) described that
in the liquid phase o-xylene does not isomerize directly to
p-xylene, or vice versa. Similar results were obtained for va-
por phase reactions by Hanson and Engel (43). Although
there seems to be some mobility of the ortho- and meta-
isomers of xylene in the pore system of ZSM-5, as Mirth
et al. (44) determined (the ratio of the diffusion coeffi-
cients for the three xylene isomers by FTIR spectroscopy at
100–200ꢄC is around para : ortho : meta = 1000 : 10 : 1), the
contributions of meta- and even more so of ortho-xylenes
should be negligible for the conversion in the pore system.
This is even more important for the diethylbenzenes, where
the molecules are bulkier.
From the dependence of the conversion on the particle
size it is possible to obtain a rough estimate for the dif-
fusion coefficient of ethylbenzene in ZSM-5. If a catalyst
efficiency of 100% is assumed for sample A with very small
particle sizes, the catalyst efficiency of the larger particle
size materials can be fitted using the diffusion coefficient
as the free parameter under the assumption of a constant
turnover frequency (TOF) for all samples. The assumption
of an efficiency of 100% seemed reasonable for a material,
which has most of the active sites located at or near the
external surface. A reasonable fit of the data could other-
wise only be obtained, if very low catalyst efficiencies were
assumed for sample A (below 10% ) which did not seem
plausible. The TOF has been calculated from the conver-
sion over sample A to be 3.4 ꢁ 10ꢂ5 sꢂ1, assuming that the
concentration in the crystal did not substantially differ from
the gas-phase concentration (catalyst efficiency assumed as
100% ).
FIG. 5. Dependence of the effectiveness factor on the crystal length
along (100). Solid lines are fitted curves. Curve B: D = 1.2 ꢁ 10ꢂ12 cm2/s.
For estimating the sensitivity of the analysis, curves are shown which
are obtained using different values of the diffusion coefficient. (A) D =
1.2 ꢁ 10ꢂ13 cm2/s; (B) D = 1.2 ꢁ 10ꢂ12 cm2/s; (C) D = 1.2 ꢁ 10ꢂ11 cm2/s.
1.2 ꢁ 10ꢂ12 cm2/s at 523 K (Fig. 5). This is somewhat lower
than diffusion coefficients estimated for para-xylene by
Mirth et al. (44) (6 ꢁ 10ꢂ12 cm2/s at 373 K) and apprecia-
bly lower than those reported by other authors (45–47) for
ethylbenzene (around 4 ꢁ 10ꢂ9 cm2/s at 373 K). These data
were obtained at low loadings from uptake measurements.
However, even at higher loadings high mobilities have been
reported (of the order of 10ꢂ10 cm2/s (48)). One should bear
in mind, however, that the diffusion coefficient obtained in
this study is only an effective one, valid under reaction con-
ditions in the ethylbenzene disproportionation, which can
not easily be related to results from sorption experiments.
Diffusion might be strongly hindered by the bulky isomers
which can be present in the channel intersection, but not
diffuse fast. Reduction of the diffusion rates by coking ap-
pears to be rather improbable, since no sign of coking of
the catalysts was detected.
From Fig. 5 it can be deduced that the onset of diffu-
sion limitations is expected for crystals in the size range of
several micrometers. This agrees well with the findings of
Arsenova et al. (15), who did not observe diffusion limita-
tions for crystals up to 1 ꢀm in size, but at the slightly higher
reaction temperature of 563 K.
Assuming plate geometry which is reasonable for the
pore system of ZSM-5, the catalyst efficiency ꢁ is express-
ed as
ꢁ = tanh ϕ/ϕ;
with ϕ being the Thiele modulus where
CONCLUSIONS
By measuring the external acidity of a series of well-
defined H-ZSM-5 samples using 2,6-dimethylpyridine as a
probe molecule we could demonstrate that the activity as
well as the shape-selective properties in the disproportiona-
tion of ethylbenzene are clearly correlated with the external
acidity. Over large crystals (80 ꢀm) with a very low exter-
ϕ = L(r/D)1/2
and L is the thickness of the crystal along [100], r is the nal acidity only p-diethylbenzene is formed in concentra-
reaction constant, and D is the diffusion coefficient. tions above the detection limit of the GC. With decreasing
Nonlinear regression of the catalyst efficiency yielded crystal size and thus an increasing number of external acid
an effective diffusion coefficient for ethylbenzene of D = sites, first the meta-isomer, and only for very small crystals

EXTERNAL ACIDITY OF H-ZSM-5
53
(100 nm) also the ortho-isomer, are observed as reaction 15. Arsenova, N., Haag, W. O., and Karge, H. G., Stud. Surf. Sci. Catal.
105, 1293 (1997).
16. von Ballmoos, R., and Meier, W. M., Nature 289, 782 (1981).
17. Althoff, R., Schulz-Dobrick, B., Schu¨th, F., and Unger, K., Micro-
products. The catalyst efficiency decreases from around 1
for the smallest crystals down to about 0.1 for the 80 ꢀm ma-
terial. The measured external acidity is much higher than
porous Mater. 1, 207 (1993).
expected from the Si/Al ratio and the crystal size. The ex-
planation for this fact is surface roughness and the presence
of cracks on the one hand, and an enrichment of aluminum
in the crystal rim on the other hand. The disproportionation
of ethylbenzene most probably proceeds shape-selectively
to p-diethylbenzene in the pore system of the ZSM-5. The
18. Matulewicz, E. R., Kerkhof, F., Moulijn, J., and Reitsma, H., J. Colloid
Interface Sci. 77, 110 (1980).
19. Corma, A., Rodellas, C., and Fornes, V., J. Catal. 88, 374 (1984).
20. Jacobs, P. A., and Heylen, C. F., J. Catal. 34, 267 (1974).
21. Karge, H. G., Ko¨sters, H., and Wada, Y., in “Proceedings of the Sixth
International Zeolite Conference” (D. Olson and A. Bisio, Eds.),
p. 308. Butterworths, London, 1984.
p-diethylbenzene is then isomerized on the external sur- 22. Weber, R. W., Fletcher, J. C. Q., Mo¨ller, K. P., and O’Connor, C. T.,
Microporous Mater. 7, 1 (1996).
23. U.S. Patent 3,926,782 (1975).
24. U.S. Patent 3,702,886 (1975).
25. Kumar, D., Ph.D. Thesis, Brunel University, UK (1993).
face of the crystals via 1,2-alkyl-shift reactions first to the
meta-isomer and subsequently to o-diethylbenzene.
These results suggest that if a catalyst is desired which
has a high para-selectivity as well as a high conversion, it is
best to use crystals which are as small as possible, and block
or remove the external acidic sites, following procedures
which have been described in the literature (11, 49, 50).
26. Grillet, Y., Llewellyn, P. L., Kenny, M. B., Rouquerol, F., and
Rouquerol, J., Pure Appl. Chem. 56, 2157 (1993).
27. Luce, G., and Gast, T., Wa¨gen Dosieren 20 (1989).
28. Van der Gaag, F., Adriaansens, R., van Bekkum, H., and van Geem,
P. , Stud. Surf. Sci. Catal. 49, 283 (1990).
29. Tissler, A., Polanek, P., Girrbach, U., Mu¨ller, U., and Unger, K., Stud.
Surf. Sci. Catal. 46, 399 (1989).
30. Debras, G., Gourgue, A., B. Nagy, J., and Declippeleir, G., Zeolites 5,
369 (1985).
31. Jacobs, W. P. J. H., Schunk, S., and Schu¨th, F., Microporous Mater., in
press.
32. Chen, N. Y., Kaeding, W. W., and Dwyer, F. G., J. Am. Chem. Soc. 101,
6783 (1979).
ACKNOWLEDGMENTS
This research was funded by the Deutsche Forschungsgemeinschaft
under Grant Schu744/1-1. We thank Dr. G. Kokotailo and Dr. D. Kumar
who supplied the samples. We also thank the referees for several valuable
suggestions.
33. Chen, N. Y., J. Catal. 114, 17 (1988).
REFERENCES
34. Kim, J. H., Namba, S., and Yashima, T., Stud. Surf. Sci. Catal. 46, 71
(1989).
1. Paparatto, G., Moretti, E., Leofanti, G., and Gatti, F., J. Catal. 105, 227 35. Wei, J. A., J. Catal. 76, 433 (1982).
(1987).
36. Weisz, P. B., Pure Appl. Chem. 52, 2091 (1980).
2. Lonyi, F., Engelhardt, J., and Kallo, D., Zeolites 11, 169 (1991).
3. Young, L. B., Butter, S. A., and Kaeding, W. W., J. Catal. 76, 418
(1982).
37. Haag, H. W., Lago, R. M., and Weisz, P. B., Faraday Discuss. Chem.
Soc. 72, 317 (1982).
38. Fraenkel, D., Ind. Eng. Chem. Res. 29, 1814 (1990).
4. Babu, G. P., Santra, M., Shiralkar, V. P., and Ratnasamy, J. Catal. 100, 39. Derouane, E. G., Stud. Surf. Sci. Catal. 5, 5 (1980).
458 (1986). 40. Nayak, V. S., and Riekert, L., Appl. Catal. 23, 403 (1986).
5. Ku¨rschner, U., Jerschkewitz, H. G., Schreier, E., and Vo¨lter, J., Appl. 41. Karge, H., Wada, Y., Weitkamp, J., Ernst, S., Girrbach, U., and Beyer,
Catal. 57, 167 (1990).
H., Stud. Surf. Sci. Catal. 20, 101 (1984).
6. Beschmann, K., and Riekert, L., J. Catal. 141, 548 (1993).
7. Sulikowski, B., and Klinowski, J., Z. Phys. Chem. 177, 93 (1992).
8. Kaeding, W. W., J. Catal. 95, 512 (1985).
42. Allen, R. H., and Yats, L. D., J. Am. Chem. Soc. 81, 5289 (1955).
43. Hanson, K. L., and Engel, A. J., Am. Inst. Chem. Eng. 13, 260 (1967).
44. Mirth, G., Cejka, J., and Lercher, J. A., J. Catal. 139, 24 (1993).
9. Shiralkar, V. P., Joshi, P. N., Eapen, M. J., and Rao, B. S., Zeolites 11, 45. Niessen, W., and Karge, H., Micropor. Mater. 1, 1 (1993).
511 (1991).
46. Bu¨low, M., Caro, J., Ro¨hl-Kuhn, B., and Zibrowius, B., Stud. Surf. Sci.
Catal. 46, 505 (1989).
10. Beschmann, K., Riekert, L., and Mu¨ller, U., J. Catal. 145, 243
(1994).
47. Ruthven, D. M., Eic, M., and Richard, E., Zeolites 11, 647 (1991).
11. Bhat, Y. S., Das, J., Rao, K. V., and Halgeri, A. B., J. Catal. 159, 368 48. Fo¨rste, C., Ka¨rger, J., Pfeifer, H., Rickert, L., Bu¨low, M., and Zika´nova,
(1996).
A., J. Chem. Soc. Faraday Trans. 86, 881 (1990).
49. Yashima, T., Sakaguchi, Y., and Namba, S., Stud. Surf. Sci. Catal. 7, 739
(1981).
50. Kaeding, W. W., Chu, C., Young, L. B., Weinstein, B., and Butter, S. A.,
J. Catal. 67, 159 (1981).
12. Csicsery, S. M., Zeolites 4, 202 (1984).
13. Kim, J., Namba, S., and Yashima, T., Bull. Chem. Soc. Jpn. 61, 1051
(1988).
14. Hibino, T., Niwa, M., and Murakami, Y., J. Catal. 128, 551 (1991).