Acetylation of aromatic compounds over H-BEA zeolite: The influence of the substituents on the reactivity and on the catalyst stability

Article abstract of DOI:10.1016/j.jcat.2004.12.021

The acylation with acetic anhydride of six aromatic substrates with different features (degree of activation of the aromatic ring towards electrophilic substitution, number of rings, i.e., 1 or 2) was carried out in a batch reactor at 373 K over a H-BEA zeolite (Si/Al = 15) with nitrobenzene as a solvent. The acetylation rate was found to be very dependent on the degree of ring activation, with diffusion limitations playing only a limited role. The decrease of the rate with reaction time, which was very pronounced with poorly activated and deactivated substrates, is mainly due to the inhibiting effect of acetic acid and of the products of acetic anhydride condensation.

Full text of DOI:10.1016/j.jcat.2004.12.021

Journal of Catalysis 230 (2005) 375–383
www.elsevier.com/locate/jcat
Acetylation of aromatic compounds over H-BEA zeolite: the influence
of the substituents on the reactivity and on the catalyst stability
Matteo Guidotti a,c,∗,1, Christine Canaff a, Jean-Marie Coustard b, Patrick Magnoux a,
Michel Guisnet a
a
Faculté des Sciences Fondamentales et Appliquées, Université de Poitiers, UMR CNRS 6503, 40 av. du Recteur Pineau, 86022 Poitiers cedex, France
Faculté des Sciences Fondamentales et Appliquées, Université de Poitiers, UMR CNRS 6514, 40 av. du Recteur Pineau, 86022 Poitiers cedex, France
b
c
CNR – Istituto di Scienze e Tecnologie Molecolari, via Venezian 21, 20133 Milano, Italy
Received 25 October 2004; revised 16 December 2004; accepted 17 December 2004
Available online 25 January 2005
Abstract
The acylation with acetic anhydride of six aromatic substrates with different features (degree of activation of the aromatic ring towards
electrophilic substitution, number of rings, i.e., 1 or 2) was carried out in a batch reactor at 373 K over a H-BEA zeolite (Si/Al = 15) with
nitrobenzene as a solvent. The acetylation rate was found to be very dependent on the degree of ring activation, with diffusion limitations
playing only a limited role. The decrease of the rate with reaction time, which was very pronounced with poorly activated and deactivated
substrates, is mainly due to the inhibiting effect of acetic acid and of the products of acetic anhydride condensation.
2004 Elsevier Inc. All rights reserved.
Keywords: Acetylation; Acetic anhydride; Aromatic ketones; BEA zeolite; Reaction rate; Catalyst poisoning
1. Introduction
phoric or hydrofluoric acid, involves similar problems, the
latter being much more volatile than the conventional metal
chlorides.
Aromatic ketones are useful intermediates for the syn-
thesis of a wide range of compounds used in the fine and
speciality chemical industry, such as pharmaceuticals, pesti-
cides, flavours, perfumes, UV adsorbers, and dyes [1,2]. The
acylation reactions are generally carried out in batch reactors
over corrosive conventional Friedel–Crafts catalysts, such
as AlCl3, FeCl3, or other Lewis acid metal chlorides. Be-
cause of the formation of very stable complexes between the
metal chloride and the arylketone [3], more than stoichio-
metric amounts of catalysts and a final hydrolysis step of the
1:1 molar adduct are required. Therefore, such a process
leads to the production of huge quantities of acidic by-
products to be neutralised, treated, and properly discharged.
Furthermore, the use of Brønsted acids, such as polyphos-
Acid zeolites, with their shape-selective properties and
the easy tailoring of their acidity and porosity, are par-
ticularly adapted to the selective synthesis of arylketones,
their main limitation being in the size of the synthesised
molecules. In the late 1960s, Venuto and Landis reported
m-xylene acetylation over X zeolite [4]. Since that pioneer-
ing work, several recent studies have dealt with the acylation
of aromatics catalysed by acidic zeolites [5], and two in-
dustrial processes using zeolites (H-BEA or H-FAU) were
developed for the selective synthesis of acetoanisole and ace-
toveratrole, which are, respectively, precursors of a sun pro-
tector (Parsol) and of a component in an insecticide formu-
lation (Verbutin). In both cases, noteworthy improvements
were achieved by replacement of conventional technology
(Lewis acid catalysts, acetylchloride as acylating agent, and
batch reactor) with innovative technology (zeolite catalysts,
acetic anhydride, and fixed-bed reactor) [6].
*
Corresponding author. Fax: +39-02-50314405.
E-mail address: m.guidotti@istm.cnr.it (M. Guidotti).
Temporarily in Poitiers (France) as a visiting researcher.
1
0021-9517/$ – see front matter
doi:10.1016/j.jcat.2004.12.021
2004 Elsevier Inc. All rights reserved.

376
M. Guidotti et al. / Journal of Catalysis 230 (2005) 375–383
Six mono- and bicyclic aromatic compounds were used as
substrates: anisole (Acros Organics; 99%), toluene (Aldrich;
99.5%), m-xylene (Fluka; 99%), fluorobenzene (Aldrich;
99%), 2-methoxynaphthalene (Lancaster; 98%), and 2-
methylnaphthalene (Aldrich; 98%). The catalytic tests were
carried out in liquid phase, at 373 K, under a dry nitrogen
atmosphere, in a three-necked round-bottomed flask con-
nected to a reflux condenser and equipped with magnetic
stirring (750 rpm). Before each catalytic test, the zeolite
was pretreated for 8 h under dry air flow (100 mL min−1) at
773 K. The standard conditions were as follows: 500 mg of
activated zeolite was added to a mixture containing 35 mmol
of substrate, 7 mmol of acetic anhydride (AA) (Fluka; 99%),
and 15 mL of nitrobenzene (Lancaster; 99%) as solvent. Sol-
vent and reagents had previously been dried on 3A molecular
sieves. Time zero of the reaction was taken at the time when
the catalyst was introduced in the mixture heated at the reac-
tion temperature. Small samples (0.1 mL) were periodically
taken and analysed by GC-FID on a capillary Varian CP-Sil-
8 CB column. Products were identified by reference samples
or by mass spectrometry.
Scheme 1.
As in the last two decades many zeolitic catalysts have
been tested under various conditions and with different acy-
lating agents, it is now possible to evaluate critically their
performance in the synthesis of aromatic ketones [5]. How-
ever, only a limited number of substrates have been used.
Most of the works were carried out on alkylarenes or alkox-
yarenes such as benzene, toluene, xylene, anisole, veratrole,
and methoxynaphthalene [7–18]. Furthermore, since the au-
thors focused their attention mainly on finding the best reac-
tion conditions for each substrate, a systematic and compar-
ative study of the performance of a catalyst in the acylation
of different aromatics under the same conditions is still lack-
ing.
Likewise, one of the main drawbacks to be solved in the
use of zeolites as acylation catalysts is the decrease with time
of the reaction rate, which is due either to the deposition
of heavy by-products within the pore system and on their
external surface or to inhibition of acylation by products of
the main and secondary reactions. Nevertheless, since the
conditions under which the reaction is carried out may affect
largely the extent of the rate decrease, the reasons for the
gradual loss of zeolite activity during the acylation reaction
have not been unequivocally determined so far.
Since the acetic anhydride was present in a substoi-
chiometric amount, the conversion of substrate (XSUB) has
been referred to the maximum possible conversion into
monoacetylated products (here 20%):
X(converted substrate)
XSUB
=
,
X(maximum substrate conversion)
where
%(SUB at time = t)
%(SUB at time = 0)
X(converted substrate) = 1 −
.
The conversion of acetic anhydride (XAA) has been com-
puted from the amount of consumed acetic anhydride:
In this work, the reactivity in acetylation with acetic an-
hydride of six aromatic molecules with different features
(monocyclic or bicyclic arenes, activated or nonactivated
substrates towards electrophilic substitution) is systemati-
cally compared over an acidic BEA zeolite (Scheme 1). The
effect of the substrate on the decrease with time of the acy-
lation rate is also evaluated, and various experiments are
carried out to specify the origin of this loss in activity.
%(AA at time = t)
XAA = 1 −
.
%(AA at time = 0)
The organic compounds retained on the external zeolite sur-
face (including the intercrystallite mesopores) and in the
micropores were recovered at the end of the experiment.
For that, the used catalyst underwent a double-extraction
methodology, as previously described [19]. First, the deac-
tivated catalyst was treated in a Soxhlet extraction appa-
ratus with dichloromethane, and the obtained solution was
analysed by GC-FID. Then, the resulting solid was dissolved
in hydrofluoric acid (40% aq.; Labosi), and the organic
species was extracted with dichloromethane and analysed by
GC-MS. With such a methodology, the largest part of the
adsorbed compounds was recovered during Soxhlet extrac-
tion. In the following zeolite dissolution step, a much less
significant part (<3% of the retained compounds after re-
action), which was neither extracted by Soxhlet treatment
nor dissolved in aqueous HF, was recovered by the final
dichloromethane extraction. Elemental analysis of the de-
activated catalyst was carried out before and after Soxhlet
treatment by dichloromethane.
2. Experimental
H-BEA zeolite (CP811B25; total Si/Al ratio = 15) was
provided by PQ Zeolites. The total pore volume of the zeo-
lite was 0.787 cm3 g−1 (micropore volume = 0.282 cm3 g−1
;
mesopore volume = 0.505 cm3 g−1), and the average crys-
tallite size (estimated from TEM) was smaller than 0.02 µm.
The concentrations of Brønsted and Lewis acid sites were
determined by pyridine adsorption at 423 K followed by IR
spectroscopy and were found to be 345 and 372 µmol g−1
,
respectively. Nitrogen adsorption measurements on the used
catalyst (after reaction with fluorobenzene) were performed
after a 1-h pretreatment period at 363 K in vacuo.

M. Guidotti et al. / Journal of Catalysis 230 (2005) 375–383
377
Scheme 2.
In the inhibition tests, 7 mmol of acetic acid (Aldrich;
99.8%, dried on molecular sieves) or 3.5 mmol of dehy-
droacetic acid 1 (DHA) (Scheme 2; Acros Organics; 98%)
or 7 mmol of p-methoxyacetophenone (Aldrich; 99%) or
2,4-dimethylacetophenone (Lancaster; 97%) were added at
time 0 to the reaction mixture described above. In two exper-
iments, the activated zeolite (500 mg), the acetic anhydride
(7 mmol), and the solvent (15 mL) were mixed and left at
373 K for 1 h, and only after that time was the substrate
(35 mmol of either anisole or m-xylene) added.
Fig. 1. Conversion (X
m-xylene ("), toluene (1), 2-methylnaphthalene (!) and fluorobenzene
(Q) vs time.
) of anisole (F), 2-methoxynaphthalene (×),
SUB
3. Results and discussion
3.1. Acylation rate and product distribution
Nitrobenzene, which has an intermediate polarity, was
chosen as a solvent. Indeed, in previous studies [7,20] it was
shown that in the presence of nonpolar solvents there was
a strong inhibition of acetylation by the very polar prod-
ucts (acetic acid, acetylated products), whereas highly polar
solvents competed with the reactant molecules to enter the
micropores and adsorb on the acidic sites, with, therefore, a
significant limitation in the acetylation rate.
Fig. 2. Conversion of acetic anhydride (X ) during the acetylation
AA
Regardless of the substrate and the reaction time, the
products of the substrate transformation resulted essentially
from its monoacetylation. The other main product observed
in the reaction mixture was acetic acid, which resulted from
acetic anhydride consumption.
Figs. 1 and 2 show, respectively, the conversion profile of
the substrates and of acetic anhydride (AA) vs time. At the
beginning and after a reaction time of 60 min, the order of
of anisole (F), 2-methoxynaphthalene (×), m-xylene ("), toluene (1),
2-methylnaphthalene (!) and fluorobenzene (Q) vs time.
2-methylnaphthalene and fluorobenzene. Moreover, there is
no further conversion of these latter substrates after a reac-
tion time of a few minutes.
The order of initial reactivities of the benzenic substrates
is the one expected from the activating/deactivating effect
of the substituents. The polarity properties, which affect the
approach and the entrance of the aromatic molecule into the
zeolite pores, play a more limited role. Actually, even though
anisole and fluorobenzene have practically the same polarity
parameter, as proposed by Reichardt (0.198 and 0.194, re-
spectively) [21], their reactivities in heterogeneous acylation
with AA are dramatically different.
It should be noted that, even though m-xylene acetyla-
tion should be much faster than toluene acetylation (7 times
with AlCl3 catalyst [2]), in this case it occurs practically
at the same rate. This behaviour suggests the presence
of limitations in the desorption of the relatively bulky
2,4-dimethylacetophenone molecules from the zeolite mi-
cropores. Furthermore, as could be expected from the most
activating character of the methoxy group, acetylation of
2-methoxynaphthalene is much faster (30 times) than acety-
lation of 2-methylnaphthalene. However, the acetylation of
substrate reactivity was anisole > 2-methoxynaphthalene >
∼
m-xylene
zene.
toluene > 2-methylnaphthalene > fluoroben-
=
The initial rates of acetylation of the six substrates were
estimated from the slope of the tangent at time zero to the
curves in Fig. 1: anisole was found to be 2.5, 12, 13, 75,
and 340 times more reactive than 2-methoxynaphthalene,
m-xylene, toluene, 2-methylnaphthalene, and fluoroben-
zene, respectively. We also calculated turnover frequency
(TOF) values by admitting, in agreement with the literature
[2,8,9], that the acetylation reaction with acetic anhydride
occurs on the protonic sites without direct participation of
the Lewis acid sites. Whereas the values of TOF found for
anisole and 2-methoxynaphthalene are relatively high, in-
dicating the possibility of using H-BEA as a catalyst for
the acetylation of the substrates with methoxy groups, this
is not the case with the other substrates, especially with

378
M. Guidotti et al. / Journal of Catalysis 230 (2005) 375–383
Table 1
Rate and selectivity of the acetylation of substituted arenes over H-BEA
a
b
c
d
e
f
Substrates
X
r
0
TOF
(h
Product
Isomer
Distribution
(%)
SUB
−1 −1
−1
(%)
95
(mmol h
130
g
)
)
Anisole
370
150
Methoxyacetophenone
Methoxyacetonaphthone
para
ortho
99.2
0.8
2-Methoxynaphthalene
56
50
1,2-
2,6-
2,8-
74.1
22.0
2.1
others
1.8
m-Xylene
16
11
32
Dimethylacetophenone
2,4-
3,5-
2,6-
98.0
1.1
0.9
Toluene
11
5
10
29
5
Methylacetophenone
Methylacetonaphthone
para
ortho
98.7
1.3
2-Methylnaphthalene
1.7
2,6-
2,8-
1,2-
70.5
15.7
3.1
others
10.7
g
Fluorobenzene
a
1
0.35
1.1
Fluoroacetophenone
para
>99
Reaction conditions: 500 mg H-BEA-15 zeolite; substrate:AA ratio = 5; 15 mL PhNO ; 373 K; 1 bar; 60 min.
2
b
c
d
Conversion of substrate after 60 min.
Initial rate of monoacetylation.
Turn-over frequency ([mol
−1 −1 +
) after 2.5 min. The amount of H was computed according to the amount
H+ s
] [mol
]
SUB
of Brønsted acid sites.
e
In all cases selectivity to acetylated products was > 98% after 60 min.
Obtained from GC-analysis of the reaction mixture after 60 min.
Only p-fluoroacetophenone isomer was detected.
f
g
naphthalene derivatives is slower than that of benzenic deriv-
atives with the same substituting group, whereas the reverse
could be expected [22]. This is most likely due to limitations
in the desorption from the BEA micropores of the acetylated
naphthalenic products, which are bulkier than the acetylated
benzenic products.
With regard to the chemoselectivity of the reaction, in
all cases no other products but acetylated derivatives were
detected in the GC analysis of the reaction mixture. There-
fore the selectivity for acetylated species can be considered
higher than 98%.
As far as regioselectivity is concerned, para-disubstituted
acetophenones were almost exclusively obtained in the case
of anisole, toluene, and fluorobenzene (Table 1). As the for-
mation of para-disubstituted products is strongly favoured
also with homogeneous and nonzeolitic solid acid cata-
lysts (>97% in almost all cases [2,8,23]), such very high
para-selectivity is likely due to the intrinsic features of the
substrate (electronic effects and intramolecular steric con-
straints) rather than to the shape selectivity of the BEA
zeolite. In the case of m-xylene, the synergic ortho,para-
orienting effect of the two methyl groups led mainly to 2,4-
dimethylacetophenone. In any event, very small amounts of
the other two less favoured isomers were detected, namely
3,5-dimethylacetophenone and 2,6-dimethylacetophenone.
With the bicyclic aromatic substrates, 2-methoxynaph-
thalene and 2-methylnaphthalene, a wider number of acety-
lated derivatives was possible, and, in this case, the presence
of the oxygen atom affected markedly the isomer distribu-
tion of the products:
– The acylation of 2-methoxynaphthalene led predomi-
nantly to 2-methoxy-1-acetonaphthone (Table 1). The
preferential acetylation of this molecule in position 1
has been shown previously [7,11] and is consistent with
the lower charge density on carbon 1 with respect to
that on the other carbon atoms obtained from semi-
empirical MO calculations [24]. However, at long reac-
tion times, 2-methoxy-1-acetonaphthone was found to
undergo isomerisation through a transacylation process
(from 2-methoxy-1-acetophenone to the residual 2-
methoxynaphthalene) to the more thermodynamically
favoured species, that is, to 6-methoxy-2-acetonaphtho-
ne [11]. In fact, a more detailed examination of the
evolution of the isomer distribution versus the reaction
time showed that the fraction of the 2,6-isomer increased
gradually from 14% after 5 min to 22% after 60 min. It is
worth highlighting that there was still an increase in the
2,6-isomer after the complete consumption of AA (after
30 min), in agreement with the proposed transacylation
mechanism [11].
– The acylation of 2-methylnaphthalene led predomi-
nantly to 6-methyl-2-acetonaphthone, and only a small
amount of the product resulting from acetylation in
position 1 (2-methyl-1-acetonaphthone) could be ob-
served (Table 1). Such behaviour was also observed

M. Guidotti et al. / Journal of Catalysis 230 (2005) 375–383
379
with AlCl3 in the presence of nitromethane as a sol-
vent [25], whereas 2-methyl-1-acetonaphthone was the
main product in the presence of non-nitro-containing
solvents, such as CHCl3 or CS2 [26]. The large differ-
ence found on this substrate with respect to the selec-
tivity of 2-methoxynaphthalene acetylation in the same
solvent (nitrobenzene) seems surprising. Indeed, on both
reactants carbon 1 is the most activated position for elec-
trophilic substitution [22]. Therefore, the most likely
explanation is that the large difference in the distribu-
tion of acetylated products comes from differences in
the relative rates of acetylation (in position 1) and of
the following isomerisation into the 2,6-isomer. That
is to say, the ratio between the rates of isomerisation
and acetylation would be greater in 2-methyl than in
2-methoxynaphthalene transformations. However, since
no papers in the open literature dealt with the heteroge-
neous acetylation of 2-methylnaphthalene, further stud-
ies are needed to confirm this assertion.
other substrates, it is smaller than 1, that is, 0.7–0.8 with
2-methoxynaphthalene, 0.3 with m-xylene and toluene, 0.1–
0.15 with 2-methylnaphthalene, and lower than 0.1 with
fluorobenzene. This means that a part of AA (most of it with
the poorly activated aromatics) was not used for the forma-
tion of monoacetylated products, but for the production of
other compounds. Moreover, the catalyst deactivation seems
to be related to the formation of these compounds, which,
therefore, could poison the active sites or block the access
to the reactant molecules. Deactivation of H-BEA zeolite
during acetylation reactions was attributed to highly polar
species strongly retained in the zeolite pores [12–16,18].
These species could be:
– acetylated products: polyacetylated products and even
the desired monoacetylated products, which are bulkier
and more polar than the reactant molecules and hence
could be more strongly retained by the hydrophilic zeo-
lite;
3.2. Origin of the decrease in the acetylation rate with time
or
Fig. 1 shows that, with all of the substrates, the forma-
tion of monoacetylated derivatives stopped after a reaction
time of 20–30 min. However, whereas with anisole this can
be related to the complete consumption of acetic anhydride,
it is not the case with the other substrates (Fig. 2). Thermo-
dynamic limitations can also be excluded, as was shown in
the case of 2-methylnaphthalene acetylation. Indeed, no re-
action of 6-methyl-2-acetonaphthone with acetic acid (the
reverse reaction of the 2-methylnaphthalene acetylation) oc-
curred over H-BEA at 373 K.
Another important difference between the transforma-
tion of anisole and that of the other substrates is shown in
Fig. 3, in which the ratio of the substrate conversion to the
AA conversion (XSUB/XAA) is plotted versus time. With
anisole, this ratio is close to 1, which indicates a selective
transformation of AA into methoxyacetophenones. With the
– products resulting from the transformation of AA mole-
cules: acetic acid, condensation products.
These two possibilities are critically examined hereafter on
the basis of the composition of the products retained on the
zeolite and the effect of some of these products on the rate
of acetylation.
3.2.1. Analysis of the products retained on the zeolite
The products retained on the H-BEA zeolite after acety-
lation of each substrate for 60 min were extracted and
analysed. They can be classified into two groups: the organic
materials that can be recovered by simple Soxhlet extrac-
tion with dichloromethane of the used catalyst and those
that could be extracted only after destruction of the zeolite
by means of concentrated hydrofluoric acid. The first group
corresponds roughly to the products adsorbed to the external
surface of the zeolite and in the intercrystallite mesopores,
whereas the second group is formed by the organic species
strongly retained in the microporous channels of the zeolite.
The composition of the mixture obtained by Soxhlet ex-
traction (first group) was approximately the same as that of
the reaction mixture. In fact, the largest part of the com-
pounds recovered by this extraction procedure was due to
a layer of reaction mixture physisorbed to the catalyst parti-
cles.
The composition of the second group (i.e., the com-
pounds trapped within the zeolite micropores) is reported
in Table 2. First of all, it is worth noting that the carbon
content on the catalyst after Soxhlet extraction, which is a
rough measure of the global amount of these organic de-
posits, was approximately of the same order of magnitude
in all cases. Actually, the carbon content was slightly higher
after the reaction on bicyclic aromatics. In general, these
Fig. 3. Substrate conversion (X
) to acetic anhydride conversion (X
)
AA
SUB
ratio vs time in the acetylation of anisole (F), 2-methoxynaphthalene
(×), m-xylene ("), toluene (1), 2-methylnaphthalene (!) and fluoroben-
zene (Q).

380
M. Guidotti et al. / Journal of Catalysis 230 (2005) 375–383
Table 2
Composition of organic deposits strongly retained inside the zeolite pores after 1 h reaction
a
Substrates
Anisole
Carbon content
(wt%)
Organic compound extracted after HF
mineralisation of used catalyst
b
(wt%)
1.45
Monoacetylated anisole
Diacetylated anisole
93
6
1
c
Others
2-Methoxynaphthalene
m-Xylene
2.01
Monoacetylated methoxynaphthalene
Diacetylated methoxynaphthalene
Others
43
32
19
6
Unreacted methoxynaphthalene
1.90
1.80
Monoacetylated xylene
Diacetylated xylene
37
35
14
14
d
AA derivatives
Others
Toluene
Diacetylated toluene
Triacetylated toluene
AA derivatives
81
7
6
Monoacetylated toluene
Others
4
2
2-Methylnaphthalene
2.64
1.74
Monoacetylated methylnaphthalene
AA derivatives
Others
43
21
19
17
Diacetylated methylnaphthalene
Fluorobenzene
a
AA derivatives
Others
80
20
From CHNS elemental analysis on the used catalyst after Soxhlet extraction and before HF mineralisa-
tion.
b
c
See Section 2 for details.
“Others” corresponds to a large number of minor heavy compounds (e.g., polycyclic aromatic com-
pounds linked by carbonyl bridges).
d
Compounds derived from AA condensation and/or oligomerisation only.
data mean that the large differences between the decreases
in acetylation rates of the substrates are not due simply to
large differences in the total quantity of organic deposits. As
an example, even though the carbon contents on the BEA
samples used in anisole and fluorobenzene acetylation was
similar (Table 2; 1.45 and 1.74%, respectively), the loss in
activity appeared to be negligible in the first case and com-
plete in the second (Fig. 1). Furthermore, nitrogen adsorp-
tion measurements showed that the significant decrease in
rate observed with poorly reactive substrates was not due to
a blockage of the access to the active sites of the zeolite.
Indeed, neither the total pore volume (VTOT) nor the micro-
pore volume (Vµ) of the H-BEA-15 zeolite changed notice-
ably before and after reaction with fluorobenzene (before
reaction: VTOT = 0.787 cm3 g−1, Vµ = 0.282 cm3 g−1; af-
tive abundances. It is worth noting that a large part of or-
ganic compounds, which are either hydrolysable or soluble
in acidic aqueous solutions, could not be recovered because
of the use of a concentrated hydrofluoric acid aqueous so-
lution. However, the list in Table 2 displays a representative
although partial picture of the carbonaceous compounds re-
tained in the zeolite channels.
These compounds can be classified into two main cate-
gories:
1. The species resulting from mono- or poly-(di- and some-
times tri-)acetylation of the substrate. As was previ-
ously shown in anisole [13] and veratrole acetylation
[10], most of these di- and tri-acetylations do not oc-
cur on the deactivated aromatic ring but on the side
chain. The relative amount of mono- and polyacetylated
compounds depends very much on the substrate: practi-
cally only monoacetylated products from anisole; com-
parable amounts of monoacetylated and diacetylated
products from 2-methoxynaphthalene, m-xylene, and
2-methylnaphthalene; essentially polyacetylated (mainly
diacetylated) products from toluene; last, no acetylated
products from fluorobenzene (Table 2). This large dif-
ference in the amount of acetylated products retained
ter reaction and Soxhlet extraction: VTOT = 0.723 cm3 g−1
,
Vµ = 0.222 cm3 g−1; i.e., a decrease of only 8% of the total
pore volume and 20% of the micropore volume). Therefore,
the decrease in acetylation rate is likely due to inhibition of
acetylation by strongly adsorbed species on the active sites
rather than to a plugging of the zeolite channels by organic
deposits.
The main compounds recognised after the zeolite min-
eralisation are listed in Table 2, together with their rela-

M. Guidotti et al. / Journal of Catalysis 230 (2005) 375–383
381
within the zeolite micropore results most likely from dif-
ferent causes, such as the difference in polarity between
the substrate and the acetylated products, the difference
between the size of substrate molecules and pore open-
ings, and the reactivity of the substrate.
2. The species resulting from AA condensation (so-called
AA derivatives in Table 2). Three compounds were
recognised as major components, namely, 3-acetyl-
4-hydroxy-6-methyl-2-pyrone 1 (dehydroacetic acid;
DHA), 3-acetyl-2,6-dimethyl-4-pyrone 2, and 2,7-di-
methylpyrano[4,3-b]pyran-4,6-dione 3 (Scheme 2), and
they all can be obtained by multiple condensation of AA
mediated by an acidic zeolite. In particular, DHA can be
obtained from the condensation of four molecules of AA
over a zeolitic acid site [27]. An important observation
is that these AA derivatives do not appear in the acety-
lation of anisole and 2-methoxynaphthalene substrates,
which occur quickly and with only a limited decrease
in the rate with the reaction time. On the other hand,
these derivatives are the major products retained on the
zeolite during the acetylation of fluorobenzene, which
occurs very slowly and with a significant decrease in re-
action rate. This suggests that the AA derivatives play a
more significant role than acetylated compounds in the
inhibition of the acetylation reaction.
Fig. 4. Effect of inhibiting agents on the conversion of m-xylene. Simulta-
neous addition of zeolite, AA and m-xylene ("); addition of zeolite and AA
and, after 1 h aging at 373 K, addition of m-xylene (E); simultaneous ad-
dition of zeolite, AA, m-xylene and 3.5 mmol of DHA (2); simultaneous
addition of zeolite, AA, m-xylene and 7 mmol of acetic acid (×).
3.2.2. Effect of acetylated products and AA derivatives on
the acetylation rate
First, the role of acetylated products as inhibiting species
in the acylation reaction was tested, by the addition, from
the beginning, of 7 mmol of acetylated products (4-methoxy-
acetophenone and 2,4-dimethylacetophenone for anisole and
m-xylene acetylation, respectively) in the reaction mixture.
A decrease in the initial reaction rate and TOF (ca. 1.4–
1.8 times less) was observed in both cases. However, after
60 min of reaction, the conversions of anisole and m-xylene
were close to those observed in the absence of acetylated
products (96% for anisole and 14% for m-xylene, respec-
tively). Therefore, the autoinhibition of acetylation cannot
explain the complete poisoning of the catalyst that is ob-
served during the acylation of poorly activated or deactivated
substrates.
Since acetic acid and the “AA derivatives” are the typical
products obtained from the reaction of the AA on the zeolite,
further acylation tests were carried out in which the acylat-
ing agent (AA) and the presumed inhibiting species (acetic
acid or dehydroacetic acid 1, as the most representative of
the AA derivatives recognised in the organic deposit analy-
sis) were added. In these experiments, the presence of either
acetic acid or DHA led to an almost complete inhibition of
the acetylation of m-xylene (Fig. 4), whereas it had no detri-
mental effect on the acetylation of anisole (Fig. 5). On the
other hand, previous tests in which AA was replaced by ei-
ther acetic acid or DHA confirmed that none of them is an
effective acetylating agent under these conditions.
Fig. 5. Effect of inhibiting agents on the conversion of anisole. Simultane-
ous addition of zeolite, AA and anisole ("); addition of zeolite and AA and,
after 1 h aging at 373 K, addition of anisole (E); simultaneous addition of
zeolite, AA, anisole and 3.5 mmol of DHA (2); simultaneous addition of
zeolite, AA, anisole and 7 mmol of acetic acid (×).
To confirm these observations, AA, the solvent, and the
activated zeolite were mixed without the substrate at the re-
action temperature for a period of 60 min. The substrates
were then added and the reaction was followed as a function
of time. The 1 h “aging” of AA on the zeolite led to a sig-
nificant decrease in the rate of m-xylene acetylation (Fig. 4).
On the other hand, the same treatment had only a small in-
hibiting effect on the initial rate of anisole acetylation, and
a full conversion of the substrate was obtained after 60 min
(Fig. 5). It is worth highlighting that, after the 1 h “aging”
of AA and before the substrate addition, a partial consump-
tion of AA with acetic acid formation was recorded. This
consumption could be due to the interaction of AA with
the activated zeolite with formation of the active acylating
species and of AA derivatives from these species. Several
active acylating species have been postulated to be formed:
acylium ions [16], electrophilic complexes between AA and
zeolitic Brønsted acid sites [17], or ketene molecules [28].

382
M. Guidotti et al. / Journal of Catalysis 230 (2005) 375–383
All of them may play a role not only in the arene acetyla-
tion, but also in the formation of the AA derivatives.
From a practical point of view, the acidic zeolites (espe-
cially the BEA zeolite) are a viable alternative to Friedel–
Crafts catalysts, insofar as very activated and polar aromatic
substrates are used. In contrast, whenever poorly activated
or nonactivated substrates are concerned, the rapid decrease
with time of the acetylation rate could represent a major
drawback in the development of zeolite-catalysed acetyla-
tion processes.
Another series of experiments was carried out to deter-
mine the inhibiting effect of the products formed during the
acetylation of poorly activated or nonactivated substrates on
anisole transformation. Thus m-xylene was acetylated under
typical conditions for 60 min, and, after that time, an aliquot
(7 mmol) of fresh anisole was added to the reaction mixture
without the addition of fresh AA. As soon as anisole was
added, a high rate of methoxyacetophenone formation was
detected, and, after the following 60 min, the anisole con-
version was close to the maximum value (ca. 80%) expected
from the residual amount of AA after m-xylene transfor-
mation. A similar behaviour was observed with the addi-
tion of anisole to the fluorobenzene reaction mixture: the
H-BEA zeolite was completely nonactive towards fluoroben-
zene acetylation, but the addition of fresh anisole (without
the addition of fresh AA) gave rise to rapid formation of
methoxyacetophenone. Such observations show that the in-
hibiting species, which are formed on the catalyst surface,
are able to stop the acetylation of weakly activated substrates
but unable to stop that of anisole. Thanks to their high po-
larity and hydrophilicity and to the presence of an electron-
rich coordinating oxygen atom, the molecules of anisole are
probably able to compete with the AA derivatives and the
acetic acid molecules for access to the hydrophilic zeolite
micropores and to the acidic protonic sites, which is not the
case for m-xylene and fluorobenzene molecules.
Acknowledgment
M. Guidotti gratefully acknowledges the Regional Coun-
cil of the Poitou-Charentes region for a fellowship.
References
[1] M.G. Franck, J.W. Stadelhofer, in: Industrial Aromatic Chemistry,
Springer, Berlin/Heidelberg, 1987.
[2] H.W. Kouwenhoven, H. van Bekkum, in: G. Ertl, H. Knoetzinger,
J. Weitkamp (Eds.), Handbook of Heterogeneous Catalysis, vol. 5,
Wiley–VCH, Weinheim, 1997, p. 2358.
[3] H. van Koningsveld, J.J. Scheele, J.C. Jansen, Acta Crystallogr. C 43
(1987) 294.
[4] P.B. Venuto, P.S. Landis, in: D.D. Eley, H. Pines, P.B. Weisz (Eds.),
Adv. Catal., vol. 18, Academic Press, New York, 1968, p. 259.
[5] A. Corma, H. Garcìa, Chem. Rev. 103 (2003) 4307;
A. Bezuhanova, Appl. Catal. A: Gen. 229 (2002) 127, and references
therein.
[6] P. Marion, R. Jacquot, S. Ratton, M. Guisnet, in: M. Guisnet,
J.P. Gilson (Eds.), Zeolites for Cleaner Technologies, Imperial College
Press, London, 2002, p. 281.
[7] E. Fromentin, J.M. Coustard, M. Guisnet, J. Mol. Catal. A: Chem. 159
(2000) 377.
4. Conclusions
The initial rate and stability of acetylation of aromatic
substrates with acetic anhydride over a H-BEA zeolite were
shown to be very dependent on the substrate. Thus the ini-
tial rate of anisole acetylation is 340 times greater than that
of fluorobenzene acetylation; the decrease in rate with reac-
tion time is much faster in the latter case than in the former
one. The reactivity of the aromatic compounds is primarily
controlled by electronic factors: activating substituents on
the aromatic ring have a strong positive influence on the re-
activity of the substrates, deactivating substituents, a strong
negative influence. Diffusion limitations play a more limited
role, as shown by the small decrease in the acetylation rate
from the monocyclic to the bicyclic aromatic substrates.
The decrease in acetylation rate is mainly due to an in-
hibiting effect of the products of acetic anhydride transfor-
mation: acetic acid and condensation products. Such an ef-
fect is small in the acetylation of substrates such as anisole,
which are highly polar and hydrophilic and can therefore
favourably compete with the inhibiting species to enter the
H-BEA zeolite micropores and approach the active sites. On
the other hand, it is very pronounced with hydrophobic sub-
strate molecules such as methyl- and fluorosubstituted aro-
matics, which cannot easily enter the zeolite micropores in
presence of the inhibiting species.
[8] M. Spagnol, L. Gilbert, D. Alby, in: J.R. Desmurs, S. Ratton (Eds.),
The Roots of Organic Development, Elsevier, Amsterdam, 1996, p. 29.
[9] A. Corma, M.J. Climent, H. Garcia, J. Primo, Appl. Catal. 49 (1989)
109.
[10] C. Guignard, V. Pèdron, F. Richard, R. Jacquot, M. Spagnol, J.M.
Coustard, G. Pérot, Appl. Catal. A 234 (2002) 79.
[11] E. Fromentin, J.M. Coustard, M. Guisnet, J. Catal. 190 (2000) 433.
[12] P. Moreau, A. Finiels, P. Meric, J. Mol. Catal. A: Chem. 154 (2000)
185.
[13] D. Rohan, C. Canaff, E. Fromentin, M. Guisnet, J. Catal. 177 (1998)
296.
[14] E.G. Derouane, C.J. Dillon, D. Bethell, S.B. Derouane-Abd Hamid,
J. Catal. 187 (1999) 209.
[15] K. Smith, Z. Zhenhua, P.K.G. Hodgson, J. Mol. Catal. A 134 (1998)
121.
[16] U. Freese, F. Heinrich, F. Roessner, Catal. Today 49 (1999) 237.
[17] S.D. Kim, K.H. Lee, J.S. Lee, Y.G. Kim, K.E. Yoon, J. Mol. Catal.
A 152 (2000) 33.
[18] P. Botella, A. Corma, J.M. Lopez-Nieto, S. Valencia, R. Jacquot,
J. Catal. 195 (2000) 161.
[19] D. Rohan, C. Canaff, P. Magnoux, M. Guisnet, J. Mol. Catal. A:
Chem. 129 (1998) 69.
[20] V. Moreau, Ph.D. thesis, University of Poitiers (2003) p. 109.
[21] C. Reichardt, Chem. Rev. 94 (1994) 2319.
[22] K.P.C. Vollhardt, in: Organic Chemistry, W.H. Freeman and Company,
New York, 1987.
[23] J.R. Desmurs, M. Labrouillère, C. Le Roux, H. Gaspard, A. Laporterie,
J. Dubac, Tetrahedron Lett. 38 (1997) 8871.

M. Guidotti et al. / Journal of Catalysis 230 (2005) 375–383
383
[24] S. Pivsa-Art, O. Kazumi, M. Miura, S. Murata, M. Nomura, J. Chem.
[26] J.M. Bonnier, J. Rinaudo, Bull. Soc. Chim. Fr. (1971) 2094.
[27] E. Fromentin, Ph.D. thesis, University of Poitiers (2000) p. 76.
Soc., Perkin Trans. 1 13 (1994) 1703.
[28] M.L.M. Bonati, R.W. Joyner, M. Stockenhuber, Catal. Today 81
(2003) 653.
[25] P.H. Gore, A.Y. Miri, J. Rinaudo, J.M. Bonnier, Bull. Soc. Chim. Fr.
(1978) 104.