Al-ITQ-7, a Shape-Selective Zeolite for Acylation of 2-Methoxynaphthalene

Article abstract of DOI:10.1006/jcat.2000.3057

Acylation of 2-methoxynaphthalene with acetic anhydride has been carried out with a tridirectional 12-member ring pore zeolite named ITQ-7, which has a slightly smaller pore diameter than zeolite Beta. ITQ-7 is as active as Beta but gives better selectivity to 2-acetyl-6-methoxynaphthalene (2-AMN). This is due to the differences in the relative diffusion coefficients of 2-AMN and 1-acetyl-2-methoxynaphthalene (1-AMN), whose ratios are 2.7 and 15.5 in Beta and ITQ-7, respectively. In general, it can be said that when decreasing the zeolite crystallite size, the reaction rate increases, although the selectivity to 2-AMN decreases.

Full text of DOI:10.1006/jcat.2000.3057

Journal of Catalysis 197, 81–90 (2001)
doi:10.1006/jcat.2000.3057, available online at http://www.idealibrary.com on
Al-ITQ-7, a Shape-Selective Zeolite for Acylation
of 2-Methoxynaphthalene
P. Botella, A. Corma,1 and G. Sastre
Instituto de Tecnolog´ıa Qu´ımica, UPV-CSIC, Universidad Polite´cnica de Valencia, Avda. de los Naranjos s/n, 46022 Valencia, Spain
E-mail: acorma@itq.upv.es
Received May 16, 2000; revised August 29, 2000; accepted August 29, 2000
(2–7). Indeed, acylation of anisole by Y and Beta zeolites
was already known to occur with good yields and selectiv-
ity (5–7), with Beta giving the best results. Further efforts
on the catalyst and especially on the process design should
result in the successful commercial acylation of anisole and
veratrol using, respectively, Beta and Y zeolites (8, 9).
It would be of special interest to develop a highly
active catalyst for the selective acylation of 2-methoxy-
naphthalene (2-MN) with acetic anhydride or acid at
the 6-position, in this way producing 2-acetyl-6-methoxy-
naphthalene (2-AMN), which is an intermediate for the
preparation of Naproxen (10). An important work on the
preparation of 2-AMN using zeolites as catalysts was car-
ried out by HOECHST (11). In this work the authors stud-
ied the acylation of 2-MN by anhydrides and acid chlorides
on a series of zeolites such as HEU-1, Beta, and ZSM-12
using a ratio of 2-MN to acetic anhydride of 1 : 3, at tem-
peratures up to 573 K. They found that with acetyl chlo-
ride and acetic anhydride the ratio of 2-AMN to 1-AMN
(1-acetyl-2-methoxynaphthalene) was always lower than 1,
and only when propionic anhydride was used was the ra-
tio of 2-AMN to 1-AMN higher than 1, especially when
using ZSM-12 as the catalyst. Other zeolites such as fauja-
sites (10, 12), Beta (10, 12–14), mordenite (12), and meso-
porous molecular sieves (15) have also been studied as cata-
lysts for the acylation of 2-MN. Even if the acylation occurs
preferentially at the kinetically controlled 1-position of the
2-MN molecule, with the formation of 1-AMN the pore
dimensions of the zeolite can introduce shape selectivity
effects which can change the 2-AMN to 1-AMN yield ratio
in the products. In this sense, the Beta zeolite seems to have
good pore dimensions favoring the diffusion of the 2-AMN
molecule with respect to the 1-AMN isomer (10, 14, 16).
However, it should be considered that, while there are some
authors which claim that 1-AMN can only be formed at the
external surface of the Beta zeolite (13, 14), others claim
that the bulkier isomer can also be formed within the pores
and that a transacylation mechanism between 1-AMN and
2-MN molecules is the essential step for 2-AMN formation
(17).
Acylation of 2-methoxynaphthalene with acetic anhydride has
been carried out with a tridirectional 12-member ring pore zeolite
named ITQ-7, which has a slightly smaller pore diameter than ze-
olite Beta. ITQ-7 is as active as Beta but gives better selectivity to
2-acetyl-6-methoxynaphthalene (2-AMN). This is due to the differ-
ences in the relative diffusion coefficients of 2-AMN and 1-acetyl-
2-methoxynaphthalene (1-AMN), whose ratios are 2.7 and 15.5 in
Beta and ITQ-7, respectively. In general, it can be said that when
decreasing the zeolite crystallite size, the reaction rate increases,
c
although the selectivity to 2-AMN decreases.
2001 Academic Press
Key Words: acylation; 2-methoxynaphthalene; acetic anhydride;
Beta; ITQ-7; molecular dynamics.
INTRODUCTION
Aromatic ketones are used in the preparation of fine
chemicals, and in thissense some fragrancesand the produc-
tion of anti-inflammatories of the Naproxen type include
the acylation of aromatics in one of the reaction steps. Acy-
lations are generally carried out using Friedel–Crafts cata-
lysts, for instance, AlCl3. Acetonaphthones were already
being prepared in this way almost 70 years ago by react-
ing the corresponding aromatics with methoxy or ethoxy
groups with the corresponding acyl chlorides, using AlCl3
as a catalyst (1). The use of AlCl3 as a catalyst for acylation
reactions continues at present, despite, the fact that AlCl3 is
used in excess of the stoichiometric requirements, and HCl
and polyoxo–aluminum complexes, which are difficult to
separate, are generated as subproducts. Nevertheless, en-
vironmental pressure is pushing this process toward the
development of environmentally friendly solid acid cata-
lysts that can be regenerated and recycled and can carry
out the acylation reactions using more friendly anhydrides
or even organic acids as acylation reagents. Among the dif-
ferent solid acids, zeolites have been proven to be good
acylation catalysts with either acyl chlorides or anhydrides
1 To whom correspondence should be addressed.
81
0021-9517/01 $35.00
c
Copyright
2001 by Academic Press
All rights of reproduction in any form reserved.

82
BOTELLA, CORMA, AND SASTRE
It appeared to us that a large-pore tridirectional zeolite nanocrystalline Beta zeolite as seed. Beta 3 (large crystal)
with slightly smaller pores than those of Beta zeolites could was made without seeding. ITQ-7 samples were prepared in
be even more selective toward the formation of 2-AMN. a fluoride medium as B-ITQ-7 (19). These zeolites were ex-
Very recently, a new tridirectional large-pore zeolite named changed with a solution of Al(NO3)3 in order to obtain the
ITQ-7 (Instituto de Tecnolog´ıa Qu´ımica-7) has been syn- corresponding Al-ITQ-7, following the method indicated
thesized in the pure silica form (18) and in its acid form by Lobo and Davis (23).
˚
(19). This zeolite has two channels with 6.2 6.1 A and a
The Al content in the samples was determined by atomic
˚
third of 6.3 6.1 A which are smaller than the two major absorption spectrophotometry (Varian spectrAA-10 Plus).
˚
channels of the Beta zeolite, 7.2 6.2 A, since the third Crystallinity was measured by powder X-ray diffraction,
˚
one in Beta (5.5 5.5 A) is not of use for the acylation re- using a Phillips PW1710 diffractometer with Cu K radi-
action. In this work we have studied the acylation of 2-MN ation, and comparison with a highly crystalline standard
with acetic anhydride using Al-ITQ-7 as a catalyst. The in- sample. The acidity of the samples was established by the
fluence of zeolite crystallite size on activity and selectivity standard pyridine adsorption–desorption method (24). The
toward the 2-AMN isomer has been investigated and the surface area of the catalysts was calculated by the BET/BJH
results have been compared with those of Beta zeolites with method with N2 adsorption/desorption performed at 77 K
different compositions and crystal sizes. A higher selectiv- in a Micromeritics ASAP 2000 instrument. The crystal size
ity to 2-AMN has been observed with the Al-ITQ-7 zeo- was determined from the SEM images obtained in a JEOL
lite, and the results are explained on the basis of different 6300 scanning electron microscope.
diffusion of the two isomers in Beta and Al-ITQ-7. This as-
The most relevant physicochemical properties of these
sumption is supported by Molecular Dynamics calculations, zeolites are summarized in Table 1.
which have been performed to simulate the diffusion of 2-
methoxynaphthalene and the acylated 1-AMN and 2-AMN
products in Beta and Al-ITQ-7 zeolites.
Reaction Procedure
Experiments were made in a 25-ml three-neck round-
bottom flask, connected to a reflux cooler system, at 405 K
under an argon atmosphere and with magnetic stirring. All
reagents were supplied by Aldrich. The standard condi-
tions were the following: 100–200 mg of catalyst were acti-
vated “in situ” before the reaction by heating for 2 h at
573 K in a vacuum. Then, a mixture of 4.0 mmol of 2-
methoxynaphthalene, 2.0 mmol of acid anhydride (AA),
and 3 ml of chlorobenzene was added (2-MN/AA mo-
lar ratio = 2; catalyst/AA ratio = 1 g : 10–20 mmol acid
EXPERIMENTAL
Materials
Two commercial zeolite samples were provided by P.Q.
Industries: a Beta zeolite (CP811) and a USY zeolite
(CBV760). SSZ-33(20), UTD-1(21), and Beta samples(22)
were synthesized in our laboratory. Beta 1, 2, and 4 samples
(small crystal) were prepared by a seeding process, using
TABLE 1
Physicochemical Characteristics of Catalysts Tested in This Work
Acidity ( mol py)c
Bro¨nsted
623 K
Lewis
623 K
Area BET
Crystal size
m)b
1
Sample
Si/Ala
(m2 g
)
(
523 K
673 K
523 K
673 K
CP811
Beta 1
Beta 2
Beta 3
12.5
25
50
50
93
27.5
50
80
50
75
570
534
488
498
463
551
521
305
478
317
0.1–0.2
0.3–0.5
0.3–0.5
1–2
0.3–0.5
0.4–0.6
0.1 0.1
0.5
45
37
21
25
11
14
92
7
27
23
18
24
10
3
78
3
26
3
16
7
5
12
2
1
54
1
16
2
48
16
10
10
5
10
11
7
40
15
8
4
4
7
8
5
22
6
40
15
4
3
2
4
7
5
19
5
Beta 4
CBV760
SSZ-33
UTD-1
Al-ITQ-7
Al-ITQ-7n
1
1.2
0.02–0.05
27
6
24
13
a As-made molar ratio.
b Average size determined from the SEM images.
c Determined from the infrared spectra of adsorbed pyridine after evacuation at 523, 623, and 673 K.

Al-ITQ-7 ZEOLITE FOR 2-METHOXYNAPHTHALENE ACYLATION
83
anhydride). Small samples were taken periodically during responding void space, thus allowing a better initial con-
4 h. Upon recycling the catalyst, it was observed that some figuration for the MD run. The result is used as input for
deactivation occurs during the reaction. Then, in all exper- a 25-ps equilibration stage of the zeolite + sorbate system.
iments, a large number of samples were taken during the After this period, runs of 400 ps, with a timestep of 1 fs,
first hour of reaction. Products were analyzed by GC in a were carried out within a NVE microcanonical ensemble
Varian 3350 Series instrument equipped with a 30-m capil- at 450 K. The temperature was equilibrated according to
lary cross-linked 5% phenylmethylsilicone column (HP-5) the Berendsen algorithm (28). The simulations proceed
and a FID detector, using nitrobenzene as internal stan- by first assigning initial velocities to all atoms according
dard. Products were also identified by mass spectrometry to a Maxwell–Boltzmann distribution which depends on
in a Varian Saturn II GC-MS model working with a Varian the temperature of the system. From this starting point,
Star 3400 gas chromatograph and using reference samples. Newton’s equations of motion are solved using a finite time
Conversion of 2-MN has been referred to the maximum step by means of the standard Verlet algorithm (29).
conversion possible, while the selectivity to the different
In allsimulations, everyatom wasallowed to move explic-
products has been indicated according to the methoxyace- itly. Although thisincreasessubstantiallythe computational
tonaphthones distribution. Moreover, coke-like products expense, the influence of the framework flexibility has been
were formed on the catalyst during the reaction, and this be- made clear in a number of studies (30–32), especially when
came dark-brown. Catalysts were extracted in a microsoxh- the sorbate matches the pore dimensions, as is the case here.
let with CHCl3 during 12 h. The solvent was evaporated and The MD simulations have been carried out using the gen-
the remaining organic material was weighted, diluted with eral purpose DL POLY 2.11 parallel code (33), which was
chlorobenzene, and analyzed by GC-MS. Among the ad- installed on a CRAY-T3E computer at the CIEMAT. The
sorbed products we found acetic acid, 2-MN, and methoxy- simulations were run using 16 processors.
acetonaphthones. Although their percentages were usually
During the simulation, history files were saved every
less than 1 wt% , they were considered for the calculation 100 steps, and subsequent analysis used the MSD (mean
of the total mass balance. Carbon content in the used cat- square displacements) facility included in DL POLY to ob-
alyst was determined to quantify the nonextractable or- tain mean square displacements. The expression used to
ganic deposit by thermogravimetric and C-elemental anal- calculate the MSD plots was (34)
ysis performed, respectively, in a NEST instrument using
calcined kaolin as reference material and in a FISONS EA
X X
hX2(t)i = 1/(Nm Nto)
[Xi (t + t0) (Xi (t0))]2, [1]
m
to
1108 CHSN-O apparatus. Ultraviolet-visible spectroscopy
(Varian Cary 5G UV-VIS-NIR spectrophotometer) was
used to study the nature of functional groups present in
the coke. Taking into account all of these fractions it was
possible to obtain mass balances 98% .
where Nm is the number of diffusing molecules, Nto is the
number of time origins used in calculating the average, and
Xi is the coordinate of the center of mass of molecule i.
The diffusion coefficients, D, were then calculated using
the Einstein relation (34),
Molecular Dynamics Methodology
hX2(t)i = 6 D t + B,
[2]
Periodic atomistic Molecular Dynamics (MD) calcula-
tions have been performed to simulate the diffusion of
2-MN, and the acetilated products in positions 1 and 6,
1-AMN, and 2-AMN, in purely siliceous Beta and ITQ-
7. The purely siliceous unit cells of Beta and ITQ-7 (IZA
codes BEA and ISV, respectively, see Ref. 25) contain 192
atoms (64 SiO2), and in order to have a larger number of
where t is the simulation time, and B is the thermal factor
arising from atomic vibrations.
The trajectories followed by the molecules in their dif-
fusion path through the zeolite structures are visualized by
means of the “xz” and “yz” projections, which highlight
motion in the different 12-MR (member rings) channels
present in Beta and ITQ-7. The trajectory graphs contain all
the visual information needed to understand, at first sight,
the basic mechanism of the diffusion process, and they allow
us to see how the molecules diffuse through the channels.
atoms for the MD simulations, 3
3
2 macrocells con-
taining 3456 atoms were used. Thirteen molecules of ad-
sorbate were located in the void space of the macrocell in
order to complete the system. A total of six simulations
was carried out, corresponding to 2-methoxynaphthalene,
1-AMN, and 2-AMN in each of the two zeolitic structures.
The zeolite framework with the corresponding sorbate
was first optimized using the BFGS (26) technique imple-
mented in the GULP (27) code. This calculation gives a
more reliable cell size which experiences a certain change
after interaction with the corresponding sorbate and also
Interatomic potentials. Four types of interatomic poten-
tials are needed to model this system:
Vtotal =Vzeolite +Vsorbate +Vsorbate sorbate +Vzeolite sorbate. [3]
The potential for the framework, Vzeolite, was originally
allows the sorbates to relax their geometry inside the cor- derived by Catlow et al. (35) and is essentially a Born model

84
BOTELLA, CORMA, AND SASTRE
potential comprising three-body and short-range terms and
long-range Coulomb interactions. The potential for the sor-
bates, Vsorbate, was taken from Oie et al. (36) and com-
prises two- (bond), three- (angle), and four-body (dihe-
dral) interactions together with Coulomb terms. Finally,
12-6 Lennard–Jones potentials, taken from Catlow et al.
(35), and Coulomb interactions were used to describe the
sorbate–sorbate and framework–sorbate interactions. The
charges for the sorbates were taken from a Hartree–Fock
calculation using the 6-31G basis set (37). The charges of
the sorbate atoms were then scaled by a factor of 0.6 to
decrease the effect of the fully ionic zeolite framework and
make the coulombic interaction of zeolite–sorbate more
reliable. This procedure has been successfully used ear-
lier (38). More details regarding the potential parameters
(39) and the techniques employed can be found in previous
studies (40–44).
FIG. 1. Reaction scheme of the acylation of 2-MN with acetic anhy-
dride over zeolites and possible secondary reactions over the acylation
products. ● ●, Most activated position for the electrophillic substitu-
tion on 2-MN; ●, activated position; ■, low activated position; (1) pro-
todeacylation of 1-AMN; (2) protodeacylation of 2-AMN. The dashed
line indicates the lower probability of this secondary reaction due to the
high thermodynamic stability of 2-AMN. (3) Transacylation of 1-AMN to
2-AMN.
RESULTS AND DISCUSSION
During the acylation of 2-MN with acetic anhydride it has
been found (45) that both isomers, 1-AMN and 2-AMN,
are formed as primary products. However, acylation occurs
preferentially at position 1, which is the kinetically con-
trolled position. Taking this into account, at lower levels of
conversion, and if there are not geometrical constrains, one
should expect a higher selectivity to the 1-AMN isomer.
However, when contact time increases, not only does con-
version increase, but also selectivity to the 2-AMN isomer
increases, owing to secondary reactions (17) which involve
the protodeacylation of 1-AMN as well as the migration of
the acetyl group from position 1 to the 6-position of 2-MN
(see Fig. 1) by a transacetylation mechanism. This reaction
scheme and the influence of the level of conversion on prod-
uct selectivity imply that the behavior of catalysts for the
acylation of 2-MN should be compared at a similar level of
conversion.
or 10-MR windows, gives a slightly lower activity than Beta,
but with a good selectivity to 2-AMN. The diameter of
˚
the 12-MR pores in SSZ-33 is 7.0 6.4 and 6.8 6.8 A,
˚
while that of the 10-MR pores is 5.1 5.1 A. Thus, it ap-
pears that diffusion of both acylated isomers can occur
through the 12-MR pores of this zeolite. However, when the
products are formed inside those empty volumes, 2-AMN
can diffuse faster than the bulkier 1-AMN. Consequently,
1-AMN will spend more time before diffusing out, and in
this time the consecutive protodeacylation and transacyla-
tion may occur, increasing the observed selectivity to the
In the first part of our work, we have tested a series of
zeolites with different pore dimensions, some of them with
12-MR pores, which were already used in previous works
(USY and Beta), and others with 12 10 MR pores and
14-MR pores that have not been tested before (SSZ-33,
UTD-1). The characteristics of these zeolites are given in
Table 1. The results (Fig. 2) clearly show that while all of
them were active, USY and UTD-1 zeolites, which have
the largest pores, show very low selectivity to the 2-AMN
isomer, since the most kinetically favored 1-AMN is almost
exclusively formed, even at very high levels of conversion.
On the other hand, Beta, as well as SSZ-33, was more se-
lective toward the formation of the desired isomer, even if
theywere lessactive (lower levelofconversion). Notice that
SSZ-33, which can be considered to be formed by large and
empty volumes (generated at the crossing points of 12- and
10-MR pores) connected to one another by either 12-MR
FIG. 2. Acylation of 2-MN with acetic anhydride over different zeo-
lites. Conversion of 2-MN ( ) and selectivities to 2-AMN (ᮀ) and 1-AMN
(
). Experimental conditions: T = 405 K; TOS = 4 h; 2-MN/AA = 2 :1
(M); Cat/AA = 1 g : 20 mmol.

Al-ITQ-7 ZEOLITE FOR 2-METHOXYNAPHTHALENE ACYLATION
85
FIG. 3. Acylation of 2-MN with acetic anhydride over H-BEA zeo-
lites. Conversion of 2-MN ( ) and selectivities to 2-AMN (ᮀ) and 1-AMN
FIG. 4. Acylation of 2-MN with acid anhydrides of variable carbon
chain lengths (two to four carbons) over H-BEA CP811. Conversion of
2-MN ( ) and selectivities to 2-AMN (ᮀ) and 1-AMN ( ). Experimental
conditions: T = 405 K; TOS = 4 h; 2-MN/AA = 2 : 1 (M); Cat/AA = 1 g :
20 mmol.
(
). Experimental conditions: T = 405 K; TOS = 4 h; 2-MN/AA = 2 :1
(M); Cat/AA = 1 g : 10 mmol.
2-AMN product. Nevertheless, since Beta gave the best re-
sults, it was selected as the reference zeolite against which
to compare the behavior of Al-ITQ-7. However, before this
we decided to optimize the Beta zeolite from the point of
view of the framework composition and crystallite size. To
do this, Beta samples were prepared in our laboratory with
Si/Al framework ratios of 25, 50, and 93, all with a crystal-
lite size of 0.3–0.5 m (Beta 1, 2, and 4 samples in Table 1).
The acylation results given in Fig. 3 show that after 4 h of re-
action, conversion was maximum for the samples with a 25
and 50 Si/Al ratio, while selectivity was similar in all cases.
For a reaction such as the 2-MN acylation in which one
of the isomers is bulkier than the other and, consequently,
its diffusion out of the channels is much slower, the sizes
of the crystallites can play an important role on the prod-
uct selectivity observed for two reasons. First, by decreas-
ing the crystallite size of the zeolite the probability for
the bulkiest product to be formed on the external surface
strongly increases. Second, the pore length decreases and
consequently the diffusion coefficient for the bulkier prod-
uct (1-AMN) increases proportionally more. Both concepts
lead to the same conclusion, i.e., when decreasing zeolite
crystallite size the selectivity for the bulkier product should
increase. Table 2 shows indeed that when working at low
levels of conversion (11% ), in order to avoid consecutive
reactions, the reaction rate increases when decreasing the
crystallite size from 1–2 m (Beta 3 sample in Table 1) to
0.3–0.5 m (Betas 1, 2, and 4 in Table 1 and 2), and the se-
lectivity to 2-AMN decreases. These results are consistent,
but do not say whether 1-AMN is formed exclusively on
the zeolite surface or also inside the pores of Beta. At this
point, we decided to carry out the acylation of 2-MN using
anhydrides with different chain lengths. Results in Fig. 4
show that the selectivity to the 1-AMN product decreases
TABLE 2
Acylation of 2-Methoxynaphthalene with Acetic Anhydride over Al-ITQ-7 and Comparison with H-Beta Zeolites: Initial
Rates for the Different Isomers Detecteda
XT 2-MN
wt% b
XT 2-MN
wt% c
d
d
d
e
e
e
Catalyst
r2-AMN
r1-AMN
rOther AMN
S2-AMN
S1-AMN
SOther AMN
CP811
Beta 1
Beta 2
Beta 3
Beta 4
Al-ITQ7
Al-ITQ7n
14
10
11
11
17
15
10
69
80
84
73
77
40
28
16.83
20.57
26.45
11.70
47.02
11.64
4.11
15.39
26.45
32.33
9.76
49.96
6.4
0
0
0
0.98
0
1.09
0.78
0.82
1.09
0.94
1.82
0.54
0.91
1.29
1.22
0.77
1.06
0.55
1.86
0
0
0
0.04
0
0
0
0
0
7.64
a Experimental data: solvent, chlorobenzene; T = 405 K; catalyst/Ac2O = 1 g : 10 mmol; 2-MN/AA = 2 : 1 (M).
b Conversion data considered for the calculation of the reaction rates.
c Maximum conversion reached at 4 h of reaction.
1
d Initial reaction rate in mmol g 1 h
.
e Kinetic selectivity indicated as the ratio between the reaction rate for a determinated product and the rates for the rest.

86
BOTELLA, CORMA, AND SASTRE
as the length of the anhydride chain increases. Then, if the
bulkier 1-AMN was formed solely on the external surface
we should not observe the selectivity to 1-AMN to decrease
but, if anything, to increase in the case that the isomer acy-
lated in position 6 starts to have problems forming inside
the pores and the rate of reaction decreases. Therefore, we
conclude that the formation of the 1-AMN product in Beta
takes place not only on the external surface of the zeolite
˚
but also within the 7.2 6.2 A pores.
From the above study of the acylation reaction with zeo-
lite Beta, we have selected as a reference catalyst the Beta
3 sample with a Si/Al framework molar ratio of 50 and crys-
tallite size of 1–2 m. Then, a sample of Al-ITQ-7 with a
Si/Al ratio of 50 and a crystallite size between 1 and 2
m
was prepared and the results obtained at different times
are presented in Fig. 5a. The initial rate of the reaction was
calculated from the slope of the curve of conversion ver-
sus time, when reaction time goes to zero. An analogous
methodology was followed for the Beta 3 sample (Fig. 5b).
The initial reaction rate (Table 2) indicates that Al-ITQ-7
is as active as Beta. Moreover, from the evolution of con-
version with reaction time (Fig. 5) it is also evident that Al-
ITQ-7 decays much faster. This is not surprising if one takes
into account that the pores in Al-ITQ-7, by being smaller,
will be more easily plugged by residual hydrocarbons. In
order to see this, both catalysts, Beta and Al-ITQ-7, were
taken after being used for 4 h and then washed and finally
soxhlet-extracted. The residual amount of coke within the
pores which could not be extracted was determined by
thermogravimetric and elemental analysis to be 6.1 and
FIG. 6. UV-VIS spectra of Beta 3 (solid line) and Al-ITQ-7 (dashed
line) samples (Si/Al = 50; catalyst average crystal size, 1–2 m) after their
use in the acylation of 2-methoxynaphthalene with acetic anhydride. Ex-
perimental conditions: temperature = 405 K; TOS = 4 h; Cat/AA = 1 g :
10 mmol; 2-MN/AA = 2 : 1 (M).
5.5, respectively, for Beta and Al-ITQ-7. The results show
that a similar amount of hydrocarbons remain adsorbed
in Al-ITQ-7 even at a clearly lower maximum conversion.
The UV-visible analysis of the deactivated catalyst showed
(Fig. 6) the typical bands of acetophenones and polyaro-
matics at 245–260 and 290–300 nm, respectively.
The selectivity of the two zeolites has been compared
from the initial reaction rates for the formation of the differ-
ent isomers, and the results show that the ratio for 2-AMN
to the other products formed is 1.09 for the best Beta and
1.82 for Al-ITQ-7. This is clear proof that the pores of Al-
ITQ-7, by being smaller than those of Beta, give a bet-
ter shape selectivity effect for the formation of 2-AMN.
This selectivity effect clearly disappears when the size of
crystallites of Al-ITQ-7 is decreased from 1–2 to 0.02–
0.05 m (Table 2). Thus, it appears that the improved se-
lectivity to 2-AMN with Al-ITQ-7 can be due to product
diffusion shape selectivity.
In order to study the changes in the diffusion coefficients
for the reactant 2-MN and the products when going from
Beta to Al-ITQ-7, a Molecular Dynamics study was under-
taken with these two zeolites.
In the case of Beta zeolite, Fig. 7 presents the trajectories
of 1-AMN which indicate that the diffusion occurs through
the x-direction (Fig. 7a) and through the y-direction
(Fig. 7b). The distance traveled by the 1-AMN molecule
is not very large, and in any case the molecules diffuse in
either direction but none of the molecules is observed to
change from one channel to another by passing through
the channel intersection. This is probably due to the large
size of the naphthalene ring which precludes the dihedral
folding needed to follow the center of the void space in the
rotation of 90 when changing from one channel along [010]
to a channel along [100] or vice versa.
FIG. 5. Acylation of 2-MN with acetic anhydride over Al-ITQ-7
(a) and Beta 3 (b) samples (Si/Al = 50, catalyst average crystal size,
1–2 m). Conversion of 2-MN (ᮀ) and selectivities to 2-AMN (᭺) and
1-AMN (4). Experimental conditions: T = 405 K; 2-MN/AA = 2 : 1 (M);
Cat/AA = 1 g : 10 mmol.

Al-ITQ-7 ZEOLITE FOR 2-METHOXYNAPHTHALENE ACYLATION
87
and will be discussed later. Same considerations apply for
the smaller 2-MN reactant molecule.
In the case of Al-ITQ-7, the diffusion of 1-AMN through
the channels is presented in Fig. 8, which shows that the
molecules remain very close to the initial starting point
through the 400-ps simulation. It appears then that the
diffusion is very restricted in all channels and an “exten-
sive local motion” appears. This behavior is in marked con-
trast with the diffusion of the same sorbate in the Beta
zeolite (Fig. 7). It is clear that the slightly smaller size of
the channels of ITQ-7 with respect to those of Beta is the
reason for the observed difference. The size of the chan-
˚
nels of ITQ-7 are (2 ) 6.2 6.1 and (1 ) 6.3 6.1 A,
while the channels of Beta are (2 ) 7.2 6.2 and (1 )
FIG. 7. Trajectories of the center of mass of the 13 1-AMN molecules
diffusing through the Beta structure at 450 K. Projections in “xz” (a) and
˚
“yz” (b) are shown. The 12-MR channels of dimensions 7.2 6.2 A which
run parallel to [100] and [010] are shown.
The diffusion path of 2-AMN in Beta is similar, in the
sense that no diffusion from one channel to another is ob-
served, but nevertheless the trajectories in this case are
longer due to the smaller size of this isomer. This fact is re-
flected in the corresponding diffusion coefficients (Table 3)
TABLE 3
Diffusion Coefficients Obtained from the
Plots in Fig. 10a
Sorbate
Beta
ITQ-7
FIG. 8. Trajectories of the center of mass of the 13 1-AMN molecules
diffusing through the ITQ-7 structure at 450 K. Projections in “xz”
(a) and “yz” (b) are shown. The 12-MR channels of dimensions 6.2
2-MN
2-AMN
1-AMN
21.71
10.44
3.90
5.88
3.87
0.25
˚
6.1 A which run parallel to [100] and [010] are shown. A third system of
a Diffusion coefficient (10 6 cm2 s 1).
12-MR channels of dimensions 6.3 6.1 A runs parallel to [001].
˚

88
BOTELLA, CORMA, AND SASTRE
˚
5.5 5.5 A. It follows that, although the third channel is
wider in ITQ-7 than in Beta, the other two channels (run-
ning through [100] and [010]) are larger in Beta and this
makes it easier for 1-AMN to diffuse. It was recognized
early on that when the sizes of the sorbate and the ze-
olite void system are similar, small variations in the sor-
bate size can change the diffusivity by several orders of
magnitude (46). Although this is not the case here, as will
be seen from the diffusion coefficients below, it is quite
clear that a slightly smaller pore size is, for the 1-AMN
molecule, a strong impediment to diffuse. Results from
Fig. 8 also indicate that diffusion is only observed along
the [001] direction in ITQ-7. This is better seen (Fig. 8b)
in the molecule located around (y, z) ( 15, 7). This mo-
˚
tion corresponds to the channel of 6.3 6.1 A, which is
the largest of the three channels of ITQ-7 if one considers
the larger dimension of the opening the effective channel
size. This consideration also explains the better diffusion of
1-AMN in Beta (Fig. 7), where the larger dimension of the
˚
opening through which the molecule diffuses is 7.2 A.
Looking at the diffusion of 2-AMN in ITQ-7 we see much
wider trajectories as correspond to the smaller size of this
isomer (Fig. 9). The diffusion here is wider in two senses:
first because of the larger distance traveled through the
[001] channels and second because the channels through
[100] and [010] are also used for diffusion. This makes the
ITQ-7 a structure in which the three channels can be used
for the diffusion of 2-AMN. This was not possible in the case
of Beta because of the smaller size of one of the channels
˚
(5.5 5.5 A). The case of 2-AMN diffusing in ITQ-7 shows
another unique feature in the sense that some molecules
are observed to change from one channel to another (see
Figs. 9a and 9b). This is facilitated by the wide space avail-
able at the intersection between the channels in ITQ-7,
which makes it possible for the bulky and rigid naphthalene
unit to change its direction 90 without excessive repulsive
proximity to the channel walls.
FIG. 9. Trajectories of the center of mass of the 13 2-AMN molecules
diffusing through the ITQ-7 structure at 450 K. Projections in “xz”
(a) and “yz” (b) are shown. The 12-MR channels of dimensions 6.2
˚
6.1 A which run parallel to [100] and [010] are shown. A third system of
˚
The diffusion coefficients for 2-MN, 1-AMN, and 2-AMN 12 MR channels of dimensions 6.3 6.1 A runs parallel to [001].
in Beta and ITQ-7 from the MD simulations are obtained
from the MSD with Eq. [2]. The MSD plots are shown in
Fig. 10, and the corresponding diffusion coefficients are and this justifies the larger diffusivity in Beta than in ITQ-7.
shown in Table 3. The linearity of all the plots in Fig. 10 A different behavior could be expected at higher loadings,
is a confirmation of the sufficient sampling reached in the where the effect of the larger void space of ITQ-7 may over-
MD simulations.
A comparison of the structural effect shows that all the
come the diffusion limiting factor of smaller channel sizes.
A comparison of the sorbate size gives decreasing dif-
diffusion coefficients for a given sorbate are larger in Beta fusion coefficients as the sorbate size increases (2-MN >
than in ITQ-7. As mentioned earlier, the sorbates find two 2-AMN > 1-AMN). The trend is not the same over the two
˚
accessible channels in Beta of 7.2 6.2 A, and a channel structures and, for example, similar factors (2.1 and 1.5) are
˚
of 6.3 6.1 A for the case of 1-AMN, and two additional found between the relative diffusion coefficients of 2-MN
˚
channels of 6.2 6.1 A for the cases of 2-AMN and 2-MN, and 2-AMN in Beta and ITQ-7, respectively, whereas very
in ITQ-7. From this, a possibility of a larger diffusion of different factors (2.7 and 15.5) are obtained for the relative
2-AMN and 2-MN in ITQ-7 with respect to Beta could be diffusion coefficients of 2-AMN and 1-AMN in Beta and
envisaged. However, the larger size of the available chan- ITQ-7, respectively. The reason for the large difference be-
nels in Beta is more important than the number of channels tween the diffusion coefficients of 2-AMN and 1-AMN in

Al-ITQ-7 ZEOLITE FOR 2-METHOXYNAPHTHALENE ACYLATION
REFERENCES
89
1. Kuroda, C., Sci. Papers Phys. Chem. Res. 18, 51 (1932).
2. Corma, A., Chem. Rev. 95, 559 (1995).
3. Spagnol, M., Gilbert, L., and Alby, D., Ind. Chem. Libr. 8, 29 (1996).
4. Fang, R., Harvey, G., Kouwenhoven, H. W., and Prins, R., Appl. Catal.
A General 130, 67 (1995).
5. Corma, A., Climent, M. J., Garc´ıa, H., and Primo, J., Appl. Catal. 49,
109 (1989).
6. Harvey, G., Vogt, A., Kouwenhoven, H. W., and Prins, R., Proc. Int.
Zeol. Conf. 9th 2, 363 (1993).
7. Rohan, D., Canaff, C., Fromentin, E., and Guisnet, M., J. Catal. 177,
296 (1998).
8. Spagnol, M., Gilbert, L., Guillot, H., and Tirel, Ph-J., Patent PCT, Int.
Appl. WO 97 48,665 (1997).
9. Gilbert, L., and Spagnol, M., Patent PCT, Int. Appl. WO 97 17,324
(1997).
10. Harvey, G., and Mader, G., Collect. Czech. Chem. Commun. 57, 862
(1992).
11. Neuber, M., and Leupold, E. I., Eur. Pat. Appl. EP 459495 B1 (1991).
12. Yadav, G. D., and Krishnan, M. S., Chem. Eng. Sci. 54, 4189 (1999).
13. Harvey, G., Binder, G., and Prins, R., Stud. Surf. Sci. Catal. 94, 397
(1995).
FIG. 10. Mean square displacements for 2-MN, 2-AMN, and 1-AMN
in Beta and ITQ-7 zeolites. The diffusion coefficients obtained according
to Eq. [2] (see text) are shown in Table 3.
14. Heinichen, H. K., and Ho¨lderich, W. F., J. Catal. 185, 408 (1999).
15. Gunnewegh, E. A., Gopie, S. S., and van Bekkum, H., J. Mol. Catal.
A Chem. 106, 151 (1996).
16. Bharathi, P., Waghmode, S. B., Sivasanker, S., and Vetrivel, R., Bull.
Chem. Soc. Jpn. 72, 2161 (1999).
17. Fromentin, E., Coustard, J. M., and Guisnet, M., J. Catal. 190, 433
(2000).
18. Villaescusa, L. A., Barret, P. A., and Camblor, M., Angew. Chem. Int.
Ed. 38, 1997 (1999).
ITQ-7 can be found in the fact that 2-AMN uses three chan-
nels whereas the 1-AMN uses only one channel. This fact
confirms that a higher selectivity to 2-AMN can be expected
in ITQ-7 than in Beta during the acylation of 2-MN with
acetic anhydride.
19. Corma, A., D´ıaz-Caban˜as, M. J., and Forne´s, V., Angew. Chem. Int.
Ed. 39, 2346 (2000).
CONCLUSIONS
20. Zones, S. I., and Nakagawa, Y., Microp. Mater. 3, 547 (1994).
21. Freyhardt, C. C., Tsapatsis, M., Lobo, R. F., Balkus, K. J., Jr., and Davis,
M. E., Nature 381, 295 (1996).
Zeolite Al-ITQ-7, a tridirectional zeolite with some
smaller channels than Beta, is active for the acylation of 2-
methoxynaphthalene with acetic anhydride, giving a higher 22. Botella, P., Corma, A., Lo´pez-Nieto, J. M., Valencia, S., Lucas, M. E.,
and Sergio, M., Appl. Catal. A General 203, 251 (2000).
23. Lobo, R. F., and Davis, M. E., Microp. Mater. 3, 61 (1994).
24. Chica, A., and Corma, A., J. Catal. 187, 167 (1999).
25. Meier, W. M., Olson, D. H., and Baerlocher, Ch., “Atlas of Zeo-
selectivity to the desired 2-acetyl-6-methoxynaphthalene
than that obtained with Beta. It has been shown by Molec-
ular Dynamics calculation that this is due to the much
lower diffusion coefficient of 1-AMN in Al-ITQ-7 than in
Beta.
By using anhydrides of different molecular sizes it has
been shown that, at least in the Beta zeolite, the bulkiest
isomer is formed not only on the external surface but also
within the pores of the zeolites. In general, a decrease in
lite Structure Types,” 4th ed. Elsevier, Amsterdam (1996). Also at
http://www.iza-structure.org.
26. Shannon, D. F., Math. Comp. 24, 647 (1970).
27. Gale, J. D., J. Chem. Soc. Faraday Trans. 93, 629 (1997).
28. Berendsen, H. J. C., Postma, J. P. M., van Gunsteren, W., DiNola, A.,
and Haak, J. R., J. Chem. Phys. 81, 3684 (1984).
29. Verlet, L., Phys. Rev. 159, 98 (1967).
crystallite size increases the reaction rate but decreases the 30. Demontis, P., Fois, E. S., Suffriti, G. B., and Quartieri, S. J., Phys. Chem.
94, 4329 (1990).
selectivityto the lessbulkyand more desired 2-AMN. An ef-
ficient way to remove the surface acid sites without blocking
the pore mouths combined with the use of Beta or Al-ITQ-
31. Deem, M. W., Newsam, J. M., and Creighton, J. A., J. Am. Chem. Soc.
114, 7198 (1992).
32. Sastre, G., Catlow, C. R. A., and Corma, A., J. Phys. Chem. B 103, 5187
7 with small crystallites should maximize both conversion
and selectivity to 2-AMN.
(1999).
33. Smith, W., and Forester, T. R., J. Mol. Graph. 14, 136 (1996).
34. Allen, M. P., and Tildesley, D., “Molecular Simulation of Liquids,”
Oxford Univ. Press, Oxford, 1980.
35. Catlow, C. R. A., Freeman, C. M., Vessal, B., Tomlinson, S. M., and
Leslie, M., J. Chem. Soc. Faraday Trans. 87, 1947 (1991).
36. Oie, T., Maggiora, T. M., Christoffersen, R. E., and Duchamp, D. J.,
Int. J. Quant. Chem. Quant. Biol. Symp. 8, 1 (1981).
37. Ditchfield, R., Hehre, W. J., and Pople, J. A., J. Chem. Phys. 54, 724
(1971).
ACKNOWLEDGMENTS
The authors thank the Comisio´n Interministerial de Ciencia y Tec-
nolog´ıa in Spain (Project MAT 97-0561 and MAT 97-1016-C02-01) for
financial support and the Centro de Investigaciones Energeticas, Medio
Ambientales y Tecnologicas for the use of their computing facilities.

90
BOTELLA, CORMA, AND SASTRE
38. Shubin, A. A., Catlow, C. R. A., Thomas, J. M., and Zamaraev, K. I.,
43. Auerbach, S. M., Henson, N. J., Cheetham, A. K., and Metiu, H. I.,
Proc. R. Soc. Lond. A 446, 411 (1994).
J. Phys. Chem. 99, 10600 (1995).
39. Sastre, G., Raj, N., Catlow, C. R. A., Roque-Malherbe, R., and Corma, 44. Sastre, G., Catlow, C. R. A., Chica, A., and Corma, A., J. Phys. Chem.
A., J. Phys. Chem. B 102, 3198 (1998).
B 104, 416 (2000).
40. Yashonat, S., Thomas, J. M., Nowak, A. K., and Cheetham, A. K.,
Nature 331, 601 (1988).
41. Schrimpf, G., Schlenkrich, M., Brickmann, J., and Bopp, P., J. Phys.
Chem. 96, 7404 (1992).
42. Nicholas, J. B., Trouw, F. R., Mertz, J. E., Iton, L. E., and Hopfinger,
A. J., J. Phys. Chem. 97, 4149 (1993).
45. Kouwenhoven, H. W., and van Bekkum, H., in “Handbook of Het-
erogeneous Catalysis” (G. Ertl, H. Kno¨zinger, and J. Weitkamp, Eds.),
Vol. 5, p. 2358. VCH, Weinheim, 1997.
46. Schro¨ter, H. J., and Ju¨ntgen, H., in “Adsorption Science and Technol-
ogy” (A. E. Rodrigues, M. D. LeVan, and D. Tondeur, Eds.), NATO
ASI E158. Kluwer, Dordrecht, 1988.