Gas-phase synthesis of hydroxyacetophenones by acylation of phenol with acetic acid

Article abstract of DOI:10.1016/j.jcat.2004.05.030

The gas-phase synthesis of aromatic ketones via acylation of phenol with acetic acid was studied on SiO₂-Al₂O₃, Al-MCM-41, zeolites HY and ZSM5, and tungstophosphoric acid (HPA) supported on MCM-41 and carbon. At contact times of 146 g hr/mole, the initial conversion of phenol varied between 12.5 and 19.1% (HPA/MCM-41), and the initial selectivity to o-hydroxyacetophenone (o-HAP) between 37.1 (HPA/C) and 69.1% (HY). Zeolites HY and ZSM5, which contained strong Bronsted and Lewis acid sites, produced efficiently o-HAP via both the direct C-acylation of phenol and the acylation of phenyl acetate intermediate formed from O-acylation of phenol. On strong Bronsted acid sites of HPA-based catalysts the phenol molecule essentially interacted via the benzene ring, adopting a position parallel to the surface and favoring the attack of acylium ion mainly to the oxygen of phenol.

Full text of DOI:10.1016/j.jcat.2004.05.030

Journal of Catalysis 226 (2004) 308–320
www.elsevier.com/locate/jcat
Gas-phase synthesis of hydroxyacetophenones by acylation of phenol
with acetic acid
C.L. Padró, C.R. Apesteguía ∗
Catalysis Science and Engineering Research Group (GICIC), Instituto de Investigaciones en Catálisis y Petroquímica–INCAPE (UNL-CONICET),
Santiago del Estero 2654, 3000 Santa Fe, Argentina
Received 16 March 2004; revised 26 May 2004; accepted 27 May 2004
Available online 2 July 2004
Abstract
The gas-phase synthesis of aromatic ketones via acylation of phenol with acetic acid was studied on SiO –Al O , Al-MCM-41, zeolites
2
2 3
HY and ZSM5, and tungstophosphoric acid (HPA) supported on MCM-41 and carbon. At contact times of 146 g h/mol, the initial conver-
sion of phenol varied between 12.5% (HY) and 19.1% (HPA/MCM-41), and the initial selectivity to ortho-hydroxyacetophenone (o-HAP)
between 37.1% (HPA/C) and 69.1% (HY). In all the cases, the formation of o-HAP was clearly favored in comparison to that of para-isomer.
Zeolites HY and ZSM5, which contained strong Brønsted and Lewis acid sites, produced efficiently o-HAP via both the direct C-acylation
of phenol and the acylation of phenyl acetate intermediate formed from O-acylation of phenol. HPA-based catalysts contained only Brøn-
sted acid sites and formed o-HAP exclusively as a secondary product from phenyl acetate. The o-HAP yield remained constant with time
on stream on ZSM5 but drastically decreased on the other samples because of coke formation. The superior stability of zeolite ZSM5 was
interpreted by considering that coke precursor formation is avoided into the microporous structure of this zeolite.
2004 Elsevier Inc. All rights reserved.
Keywords: Aromatic ketones; Acylation reactions; Solid acids; o-Hydroxyacetophenone synthesis
1. Introduction
ment of phenyl acetate to p-HAP is efficiently catalyzed by
methane sulfonic acid, a strong and biodegradable acid [6].
In general, the liquid-phase Fries rearrangement of phenyl
acetate forms mainly p-HAP, giving p-HAP/o-HAP ratios
between 2 and 8, but also produces significant amounts of
phenolic by-products. Considerable research work has been
performed recently to find suitable, recyclable, and environ-
mentally benign solid acid catalysts for efficiently promoting
the liquid-phase synthesis of aromatic ketones from phenyl
acetate. Strong solid acids such as ion-exchange resins, zeo-
lites, Nafion, and heteropoly acids show moderate phenyl ac-
etate conversion activity and preferentially produce p-HAP,
but also give significant amounts of phenol and are in general
rapidly deactivated because of the formation of ketene by-
products that are highly reactive coke intermediates [7–11].
Hydroxyacetophenonesare also obtained by the acylation
of phenol using acid halides [12] or acid anhydrides [13] as
acylating agents. In the liquid phase, the reaction is catalyzed
by Friedel–Crafts catalysts and produces mainly p-HAP, al-
though the isomer selectivity may be changed by changing
the solvent polarity. In the gas phase, the acylation of phe-
Aromatic ketones are valuable intermediate compounds
for the synthesis of important pharmaceuticals and fra-
grances. In particular, ortho- (o-HAP) and para-hydroxy-
acetophenone (p-HAP) are widely used for the synthesis
of aspirin and paracetamol (4-acetoaminophenol), respec-
tively [1]. o-Hydroxyacetophenone is a key intermediate for
producing 4-hydroxycoumarin and warfarin which are both
used as anticoagulant drugs in the therapy of thrombotic
disease [2], and it has been also employed for obtaining
flavonones [3,4]. Hydroxyacetophenones are commercially
produced via the Fries rearrangement of phenyl acetate in a
liquid-phase process involving the use of Lewis and Brøn-
sted acids, such as AlCl3, TiCl4, FeCl3, and HF, which pose
problems of high toxicity, corrosion, and spent acid dis-
posal [5]. Recently, it was reported that the Fries rearrange-
*
Corresponding author. Fax: 54 342 4531068.
E-mail address: capesteg@fiqus.unl.edu.ar (C.R. Apesteguía).
URL: http://www.ceride.gov.ar/gicic.
0021-9517/$ – see front matter
doi:10.1016/j.jcat.2004.05.030
2004 Elsevier Inc. All rights reserved.

C.L. Padró, C.R. Apesteguía / Journal of Catalysis 226 (2004) 308–320
309
Scheme 1. para- and ortho-HAP formation from the acylation of phenol with acetic acid.
nol with acetic acid or acetic anhydride has been carried out
on Al-MCM-41 and zeolite ZSM5, at temperatures between
520 and 573 K [14–19]. Scheme 1 shows the synthesis of
p- and o-HAP from the acylation of phenol with acetic acid.
It was found that the gas-phase phenol acylation on solid
acids forms predominantly o-HAP, giving o-HAP/p-HAP
molar ratios higher than 80 [14,16]. However, the experi-
mental o-HAP yields were moderate, particularly because of
the significant formation of phenyl acetate [18,19]. The reac-
tion network of the gas-phase acylation of phenol with acetic
acid was studied by Neves et al. [17], but more detailed stud-
ies are needed to ascertain the exact requirements of acid
site density and strength to efficiently promote the selec-
tive o-HAP formation. Our understanding of the deactivation
mechanism under reaction conditions is also lacking, despite
the fact that the potential use of solid acids in gas-phase acy-
lation reactions is often limited because of the rapid activity
decay caused by coke formation.
In this paper, we perform a detailed study of the gas-
phase acylation of phenol with acetic acid over differ-
ent solid acids. Specifically, we prepared and character-
ized HPA/MCM-41, Al-MCM-41, zeolites HY and ZSM5,
HPA/carbon, and SiO2–Al2O3 catalysts. Our goal was to re-
late the structural properties and the surface acid site density
and strength of the solids with their ability for the efficient
catalysis of the phenol acylation reaction to yield o-HAP.
Results will show that zeolite ZSM5 does not deactivate on
stream and produces selectively o-HAP at high rates. The
unique performance of zeolite ZSM5 is based on its high
density of strong Lewis and Brønsted acid sites that promote
the formation of o-HAP both by direct C-acylation of phenol
and by phenyl acetate obtained from O-acylation of phenol,
and on its microporous structure that hinders the formation
coke precursor intermediates.
Sodium silicate solution (14% NaOH and 27% SiO2, Al-
drich), cetyltrimethylammonium bromide (Aldrich), alu-
minum isopropoxide (Aldrich), and deionizated water were
used as the reagents. The composition of the synthesis gel
was 7SiO2–xAl2O3–2.7Na2O–3.7CTMABr–1000H2O. The
pH was adjusted to 10 using a 0.1 M H2SO4 solution, and
then the gel was transferred to a Teflon-lined stainless-steel
autoclave and heated to 373 K in an oven for 96 h. After crys-
tallization, the solid was washed with deionized water, dried
at 373 K, and finally calcined at 773 K for 4 h. HPA/MCM-
41 was prepared following the procedure of Kozhenikov et
al. [21] by stirring a suspension of 1 g of MCM-41 in 30 cm3
of an aqueous solution of HPA (H3PW12O40 ·6H2O, Merck)
at room temperature for 24 h. After evaporation of the sol-
vent at 333 K under vacuum, the solid was dried at 353 K
and calcined in air at 573 K. HPA/C was prepared by stirring
a suspension of carbon (from Westvaco) in a hydrochloric
HPA solution at pH 1.6 for 24 h. The HY zeolite was pre-
pared by triple ion exchange of a commercial NaY zeolite
(UOP-Y 54) with ammonium acetate (Sigma, 99%) at 298 K
and subsequent calcination in air. The H form of ZSM5 ze-
olite (Zeocat Pentasil PZ-2/54) was obtained following the
same procedure used for HY. SiO2–Al2O3 (Ketjen LA-LPV)
was calcined at 773 K.
2.2. Catalyst characterization
The crystalline structures of MCM-41, Al-MCM-41,
HPA/C, and HPA/MCM-41 were determined by X-ray dif-
fraction (XRD) using a Shimadzu XD-D1 diffractometer and
Ni-filtered CuKα radiation. BET surface areas (Sg), mean
¯
pore diameter (dp), and pore-size distribution were measured
by N2 physisorption at its boiling point in a Quantochrome
Corporation NOVA-1000 sorptometer. Elemental compo-
sitions were measured by atomic absorption spectroscopy
(AAS).
2. Experimental
Acid site densities were determined by temperature-
programmed desorption (TPD) of NH3 preadsorbed at
373 K. Samples (200 mg) were treated in He (60 cm3/min)
at 773 K for 1.5 h and then exposed to a 1% NH3/He stream
for 40 min at 373 K. Weakly adsorbed NH3 was removed by
2.1. Catalyst preparation
Pure silica MCM-41 and Al-MCM-41 mesoporous ma-
terials were synthesized according to Edler and White [20].

310
C.L. Padró, C.R. Apesteguía / Journal of Catalysis 226 (2004) 308–320
flowing He at 373 K for 2 h. The temperature was then in-
creased at 10 K/min and the NH3 concentration in the efflu-
ent was measured by mass spectrometry (MS) in a Baltzers
Omnistar unit.
Table 1
Chemical composition and physical properties of the catalysts used in this
work
Samples
Surface area
2
Pore diameter
Si/Al
HPA
(%)
¯
7.4
7.4
5.5
d
(Å)
S
(m /g)
p
g
The nature of surface acid sites was determined by
Fourier transform infrared spectroscopy (FT-IR) by using
pyridine as probe molecule and a Shimadzu FTIR-8101M
spectrophotometer. The spectral resolution was 4 cm−1 and
50 scans were coadded. Sample wafers were formed by
pressing 20–40 mg of the catalyst at 5 tons/cm2 and trans-
ferred to a sample holder made of quartz. An inverted T-
shaped Pyrex cell containing the sample pellet was used.
The two ends of the short arm of the T were fitted with CaF2
windows. All the samples were initially outgassed at 723 K
for 4 h and then a background spectrum was recorded after
cooling the sample at room temperature. Data were obtained
after admission of pyridine, adsorption at room tempera-
ture, and sequential evacuation at 298, 423, 573, and 723 K.
Spectra were always recorded at room temperature. Differ-
ence spectra were obtained by substracting the background
spectrum recorded previously.
HY
NaY
ZSM5
660
700
350
560
925
1010
1700
6
2.4
2.4
20
11.3
18
∞
–
–
–
–
–
–
–
–
–
–
–
100
28
30
SiO –Al O
45
30
34
22
–
2
2 3
Al-MCM-41
MCM-41
Carbon
HPA
HPA/C
390
505
30
29
HPA/MCM-41
ꢀ
Si (%) = [Yi/ Yi]100. Product yields (ηi, mol of product
i/mol of phenol fed) were calculated as ηi = SiXP.
3. Results and discussion
3.1. Catalyst characterization
Coke formed on the catalysts during reaction was mea-
sured by temperature-programmed oxidation (TPO). Sam-
ples (50 mg) were heated in a 3% O2/N2 stream at 10 K/min
from room temperature to 1073 K. The evolved CO2 was
converted to methane by means of a methanation catalyst
(Ni/kieselghur) operating at 673 K and monitored using a
flame ionization detector.
The physicochemical characteristics (surface area, pore
diameter, chemical composition) of the samples used in this
work are summarized in Table 1. BET surface areas of
MCM-41 and carbon samples are higher than 1000 m2/g.
However, after impregnation with HPA the Sg values drop-
ped to 505 (HPA/MCM-41) and 390 m2/g (HPA/C). In the
case of carbon, the HPA molecules blocked the carbon mi-
cropore structure and, as a consequence, the average pore
diameter increased from 22 to 30 Å. In contrast, the HPA
addition to MCM-41 did not change the support pore diam-
eter size, probably because MCM-41 exhibits a uniformly
sized mesoporous structure.
XRD patterns of Al-MCM-41 and MCM-41 samples
showed that both samples are well crystallized and exhibit
a strong diffraction peak at 2.2◦ corresponding to 100 reflec-
tion; no other crystalline phases were observed. On the other
hand, no crystalline HPA structure was detected on HPA-
supported samples, thereby indicating that the HPA phase is
well dispersed on MCM-41 and carbon supports.
Sample acid properties were probed by TPD of NH3
preadsorbed at 373 K. The obtained TPD curves are shown
in Fig. 1. The NH3 surface densities for acid sites were ob-
tained by deconvolution and integration of TPD traces and
are presented in Table 2. The NH3 evolved from MCM-41
and carbon supports was negligible. The acid site strength
of HPA-based samples was clearly higher compared to the
other samples. Pure HPA showed a sharp NH3 desorp-
tion peak at about 910 K, which accounts for the strong
Brønsted acid sites present on this material. This high-
temperature desorption peak was also detected on HPA/C
and HPA/MCM-41 samples, which presented an additional
low-temperature peak at ca. 493 K. The evolved NH3
from HY, ZSM5, and SiO2–Al2O3 samples gave rise to
2.3. Catalytic testing
The gas-phase acylation of phenol (Merck, > 99%) with
acetic acid (Merck, 99.5%) was carried out in a fixed-bed,
continuous-flow reactor at 553 K and 101.3 kPa. Samples
were sieved to retain particles with 0.35–0.42 mm diam-
eter for catalytic measurements and pretreated in air at
773 K for 2 h before reaction, except samples containing
HPA that were pretreated in air at 573 K for 2 h. Phe-
nol (P) and acetic acid (AA) were introduced (P/AA = 1)
via a syringe pump and vaporized into flowing N2 to give
a N2/(P + AA) ratio of 45. Standard catalytic tests were
conducted at a contact time (W/FP0) of 146 g h/mol and
gas-hour space velocity (GHSV) of 235 cm3 STP/(g min).
The exit gases were analyzed on-line using a Hewlett-
Packard 5890 chromatograph equipped with a Supelcowax
10 column and a flame ionization detector. Data were
collected every 25 min for about 6 h. The main prod-
ucts of phenol acylation with acetic acid were phenyl ac-
etate (PA), ortho-hydroxyacetophenone (o-HAP) and para-
hydroxyacetophenone (p-HAP); para-acetoxyacetophenone
(p-AXAP) was detected in trace amounts. Phenol conversion
(XP, mol of phenol reacted/mol of phenol fed) was calcu-
ꢀ
ꢀ
ꢀ
lated as: XP = Yi/( Yi + YP), where Yi is the molar
fraction of products formed from phenol, and YP is the out-
let molar fraction of phenol. The selectivity to product i (Si ,
mol of product i/mol of phenol reacted) was determined as:

C.L. Padró, C.R. Apesteguía / Journal of Catalysis 226 (2004) 308–320
311
Fig. 1. TPD profiles of NH on: (A) HPA-based samples; (B) SiO –Al O ,
3
2
2 3
Al-MCM-41, and acid zeolites. Heating rate: 10 K/min.
Fig. 2. FT-IR spectra in the hydroxyls stretching region of: (A) SiO –Al O
2 3
2
Table 2
and Al-MCM-41; (B) NaY, ZSM5, and HY. (a) Degassed at 723 K for 4 h;
(b) after pyridine adsorption at room temperature and desorption at 423 K
for 0.5 h; (c) difference spectra (a)–(b).
Characterization of sample acidity: TPD of NH and FTIR of pyridine
3
Catalyst
TPD of NH
FTIR of pyridine
3
2
(µmol/g) (µmol/m ) Brønsted sites Lewis sites (L) L/B
(B) (area/g)
(area/g)
HY samples obtained after evacuation at 723 K for 4 h, and
after pyridine adsorption at room temperature and desorp-
tion at 423 K for 0.5 h. Difference spectra are also included.
The position of the IR band for the O–H stretching modes
(νOH) over degassed Al-MCM-41 and SiO2–Al2O3 is al-
most the same, about 3745 cm−1 (Fig. 2A), but the νOH
wavenumber on oxides is not directly related to the acid
strength of the hydroxyl group [22]. More useful informa-
tion on Brønsted acid site strength is obtained by analyzing
the difference spectra in Fig. 2A. It is observed, in fact, that
after desorption at 423 K of the pyridine adsorbed at room
temperature, the base is almost completely eliminated on
Al-MCM-41 but remains adsorbed to a significant extent
on SiO2–Al2O3. This result reflects the stronger acidity of
surface OH groups on SiO2–Al2O3 compared to Al-MCM-
41. On the other hand, no absorption bands were detected
in the hydroxyl stretching on the NaY sample (Fig. 2B),
thereby indicating that the hydroxyl group concentration on
this zeolite is negligible. The spectrum of degassed zeolite
HY
NaY
ZSM5
SiO –Al O
2 3
Al-MCM-41
HPA
1380
280
770
1005
340
539
315
2.1
0.4
2.2
1.8
0.4
90.0
0.8
0.7
310
n.d.
465
525
341
204
135
–
1.5
–
1.0
3.0
4.2
–
a
337
68
32
–
–
–
2
HPA/C
HPA/MCM-41 352
–
–
–
–
a
n.d., not detected.
a peak at 483–493 K and a broad band between 573 and
773 K. In contrast, zeolite NaY desorbed NH3 in a sin-
gle TPD peak centered at about 480 K. On a weight basis,
the HY zeolite exhibited the highest surface acid density
(1380 µmol/g), probably reflecting the higher aluminum
content, followed by SiO2–Al2O3 (1005 µmol/g) and ZSM5
(770 µmol/g) samples. For the other samples, the amount
of evolved NH3 was between 550 and 280 µmol/g. Pure
HPA presented the highest acid site density on an areal basis
(90.0 µmol/m2), followed by zeolites ZSM5 (2.2 µmol/m2)
and HY (2.1 µmol/m2).
ZSM5 shows two absorption bands at 3610 and 3745 cm−1
,
respectively. The band at 3610 cm−1 corresponds to Si–
OH–Al bridging hydroxyl groups [23] whereas the asym-
metric band at 3745 cm−1 is attributed to the stretching
Fig. 2 shows the FT-IR spectra in the hydroxyl stretch-
ing region of SiO2–Al2O3, Al-MCM-41, NaY, ZSM5, and

312
C.L. Padró, C.R. Apesteguía / Journal of Catalysis 226 (2004) 308–320
vibration of terminal SiOH groups located either at the
boundaries of the zeolite crystal or at the surface of non-
crystalline material [24]. Fig. 2B shows that evacuation
at 423 K does not remove to any extent the pyridine ad-
of pyridine, adsorption at room temperature, and sequen-
tial evacuation at 298, 423, 573, and 723 K. The pyridine
absorption bands at around 1540 cm−1 and between 1440
and 1460 cm−1 arise from pyridine adsorbed on Brønsted
and Lewis acid sites, respectively, on zeolites [27–30], Al-
MCM-41 [30,31], and SiO2–Al2O3 [32]. Fig. 3A shows that
the spectrum collected following evacuation at 298 K con-
tains broad absorption bands, characteristic of the presence
of physisorbed pyridine. Pyridine molecules interacting via
H bonding with weakly acidic surface OH groups (bands at
1446 and 1597 cm−1) are removed after evacuation at 423 K
and the resulting FT-IR spectrum shows well-defined ab-
sorption peaks. We decided then to characterize the sample
acidity by using spectra collected after evacuation at 423 K.
Fig. 3B compares the FT-IR spectra obtained on samples HY,
NaY, ZSM5, Al-MCM-41, and SiO2–Al2O3 after pyridine
evacuation at 423 K. The relative contributions of Lewis and
Brønsted acid sites were obtained by deconvolution and inte-
gration of pyridine absorption bands appearing in Fig. 3B at
around 1450 and 1540 cm−1, respectively. Results are given
in Table 2. In general, the band at 1540 cm−1 characteristic
for pyridinium ions does not change in wavenumber upon
varying the sample acidity but the frequency of the band ac-
counting for coordinately bound pyridine increases with the
strength of interaction [22]. FT-IR spectra of Fig. 3B show
that on Al-MCM-41 and SiO2–Al2O3 samples the coordi-
nately bound pyridine band appears at 1455 cm−1, which
reflects the adsorption of pyridine on Lewis acid sites as-
sociated with tricoordinate Al atoms. In agreement with
the results obtained by TPD of NH3, the amount of pyri-
dine adsorbed on Al-MCM-41 after evacuation at 423 K
is clearly lower compared to acid zeolites or SiO2–Al2O3,
reflecting the moderate acidic character of mesoporous Al-
MCM-41 sample. On the other hand, the areal peak rela-
tionship between Lewis and Brønsted sites, L/B, was higher
on Al-MCM-41 than on SiO2–Al2O3 (Table 2). The rela-
tive density of Brønsted and Lewis acid sites on Al-MCM-
41 depends, of course, on the Si/Al ratio, but the value of
L/B = 4.2 determined for our Si/Al = 18 sample is close
to that reported for Al-MCM-41 samples containing similar
Si/Al ratios [31]. Consistent with the IR characterization of
hydroxyl groups in Fig. 2, the pyridine absorption spectrum
on zeolite NaY does not reveal the presence of surface Brøn-
sted sites. The band representing Lewis acid centers appears
on NaY at 1443 cm−1, a frequency considerable lower com-
pared to those determined for similar bands on Al-MCM-41
and SiO2–Al2O3 samples. This frequency shift has been in-
terpreted by considering that the pyridine is adsorbed on
NaY by a polarization of the molecule in the field due to
cation [28]; i.e., the interaction is associated with Na rather
than to Al. Fig. 3A shows that on zeolite HY the pyridine ab-
sorption bands associated with Brønsted and Lewis acid sites
sorbed at room temperature on bridging OH at 3610 cm−1
,
reflecting the strong adsorption of pyridine on Si–OH–Al
groups of zeolite ZSM5. In contrast, the interaction of pyri-
dine with terminal silanol OH groups at 3745 cm−1 is
very weak. Finally, the IR spectrum of zeolite HY matrix
shows three absorption bands which correspond to high-
frequency Si–OH–Al groups (3620 cm−1), low-frequency
Si–OH–Al groups (3550 cm−1), and terminal silanol groups
(3743 cm−1) [25]. Difference spectra in Fig. 2B show that
pyridine strongly adsorbs on high-frequency Si–OH–Al
groups associated to OH groups located in large cavities.
The low-frequency band at 3550 cm−1 attributed to OH
groups in the sodalite cages is less affected by pyridine ad-
sorption. Very weakly acidic SiOH groups at 3743 cm−1 are
also affected by pyridine adsorption, probably because of a
perturbation of these groups by the pyridine molecules ad-
sorbed on neighboring Lewis sites [26].
The density and nature of surface acid sites were deter-
mined from the FT-IR spectra of adsorbed pyridine. Fig. 3A
shows the spectra obtained on zeolite HY after admission
appear after evacuation at 423 K at 1542 and 1454 cm−1
,
Fig. 3. FT-IR spectra of pyridine adsorbed on: (A) zeolite HY at 298 K and
respectively. These two bands are present even after evacu-
ation of zeolite HY at 723 K, thereby indicating that zeolite
HY contains strong Brønsted and Lewis acid sites. But it
evacuated at increasing temperatures; (B) SiO –Al O , Al-MCM-41, and
2
2 3
acid zeolites at 298 K and evacuated 423 K for 0.5 h. Dotted lines indicate
−1
−1
the presence of Lewis (1450 cm ) and Brønsted (1540 cm ) sites.

C.L. Padró, C.R. Apesteguía / Journal of Catalysis 226 (2004) 308–320
313
should be noted that the L/B ratio on HY increases with the
evacuation temperature from 1.5 (423 K) to 2.2 (723 K). The
increase of the L/B ratio with the desorption temperature
on zeolite HY has already been reported [26] and probably
reflects the fact that a part of the Brønsted sites correspond-
ing to the bridged hydroxyl groups of the structure are of
medium strength. Nevertheless, some of the Brønsted sites
display strong acidity, as evidenced by the FTIR data. These
stronger Brønsted sites are associated either to bridged OH
interacting with the Lewis sites or to extraframework OH
groups [33]. The very strong Lewis sites of zeolite HY have
been associated [26] with aluminum atoms that are either
extraframework (tetracoordinated) or associated with frame-
work defects (tricoordinated). On zeolite ZSM5, the band of
adsorbed pyridine on Lewis acid sites is split in two overlap-
ping peaks at 1445 and 1455 cm−1, respectively. A similar
two-peak band for the pyridine adsorption on Lewis acid
sites on ZSM5 was reported from researchers of the Mo-
bil Company [34]. These authors suggested [34] that the
Lewis acidity on ZSM5 is due to Al located in the zeolite
framework, probably generated during calcination, resulting
in partial hydrolysis of Al–O bonds. Zeolite ZSM5 contained
a similar concentration of Brønsted and Lewis acid sites (Ta-
ble 2), but in contrast to zeolite HY the ratio L/B did not
vary significantly by increasing the desorption temperature
of adsorbed pyridine. It was determined, in fact, that after
degassing at 723 K the L/B ratio was 0.92 that was similar
to that measured following evacuation at 423 K (L/B = 1.0).
Fig. 4. Phenol conversion and product selectivities as a function of time
on stream on: (A) zeolite HY; (B) zeolite ZSM5 [553 K, 101.3 kPa total
0
pressure, W/F = 146 g h/mol, P/AA = 1, N /(P + AA) = 45].
2
P
3.2. Catalytic testing
Fig. 4A shows phenol conversion (XP) and product selec-
tivities (Si) obtained at 553 K on zeolite HY and typically
illustrates the time-on-stream behavior of the catalysts dur-
ing the reaction. It is observed that the phenol conversion
practically remained unmodified but the product selectivity
significantly changed during the catalytic test. The PA se-
lectivity increased with time on stream at the expense of
the formation of ortho and para-HAP isomers. On the other
hand, the formation of o-HAP compared to that of p-HAP
was highly favored. Qualitatively, a similar catalytic behav-
ior on stream was observed for the other samples, excepting
ZSM5. In fact, as it is shown in Fig. 4B, on ZSM5 not only
XP but also the product selectivities did not change with time
on stream.
Table 3
Phenol conversion (X ), selectivities to o-HAP (S
), p-HAP
P
o-HAP
(S
), and PA (S ), at times zero and 4 h
PA
p-HAP
Catalysts
t = 0 h
P
t = 4 h
P
0
0
0
0
S
X
S
S
X
S
S
S
o-HAP p-HAP PA
o-HAP p-HAP PA
HY
NaY
ZSM5
15.0 69.1
15.9 62.0
18.2 67.1
8.9
–
22.0 15.4 10.4
38.0 8.1 8.1
32.7 18.2 65.3
46.5 13.1 30.9
58.3 15.9 21.9
77.5 24.3 2.4
61.0 15.5 15.8
48.0 21.7 2.6
1.4
–
88.2
91.9
34.7
68.1
76.4
96.3
83.3
97.4
0.2
1.2
1.6
1.6
1.5
–
–
Al-MCM-41 12.9 52.3
SiO –Al O
HPA
HPA/C
1.0
1.1
0.5
0.9
–
16.0 39.1
22.3 21.0
15.3 37.0
2
2
3
HPA/MCM-41 19.1 52.0
0
P
T = 553 K, W/F = 146 g h/mol; P/AA = 1, N /(P + AA) = 45.
2
The conversion and selectivity values obtained for all the
samples at times zero and 4 h are given in Table 3. Values
at time zero were obtained by extrapolating to zero the time-
on-stream curves using semilogarithmic plots. Zeolite HY
showed the highest initial selectivity values to o-HAP and
p-HAP (69 and 8.9%, respectively), but after 4 h of reaction
the o-HAP selectivity decreased to only 10.4%. Zeolite NaY
also produced ab initio predominantly o-HAP, but did not
form p-HAP and the phenol conversion diminished signifi-
cantly on stream. Pure HPA and HPA/MCM-41 were among
the most active catalysts and efficiently converted phenol to
PA, but they rapidly lost their ability for producing the o-
HAP isomer. The HPA/C sample was more stable for o-HAP
formation and after 4 h on stream still exhibited a nonneg-
ligible So-HAP value (15.8%). SiO2–Al2O3 and Al-MCM-41
were more stable than HPA-based catalysts for o-HAP for-
mation; in particular, it is noted that the So-HAP value on
Al-MCM-41 was 30.9% after 4 h on stream. But the most
stable catalyst was zeolite ZSM5; actually, we did not ob-
serve on ZSM5 sample any significant activity decay for the
o-HAP formation rate during the catalytic run.

314
C.L. Padró, C.R. Apesteguía / Journal of Catalysis 226 (2004) 308–320
0
i
0
P
Fig. 5. Product distribution for phenol/acetic acid reactions on: (A) HPA/C; (B) ZSM5; (C) HY. Product yields (η ) and phenol conversion (X ) at t = 0 as a
function of contact time [553 K, 101.3 kPa total pressure, P/AA = 1].
3.3. Reaction network
and ηo0-HAP with space time than zeolite ZSM5 (Fig. 5C).
Initial formation rates of o-HAP and PA were similar on HY
0
0
−3
1.24 × 10 mmol/(hm2),
∼
=
∼
=
sample, i.e., ro-HAP,0
r
The effect of contact time on the product distribution was
determined in order to identify primary and secondary reac-
tion pathways. The observed deactivation, however, required
that each data point be obtained on a fresh catalyst and that
initial product yields be obtained by extrapolating to ini-
tial time on stream using semilogarithmic plots. The product
yields at t = 0 (ηi0) and the corresponding phenol conver-
sion values on HPA/C are shown as a function of contact
time in Fig. 5A. The local slopes of the curves in Fig. 5A
give the rate of formation of each product at a specific phe-
nol conversion and residence time. The nonzero initial slope
for PA shows that it forms directly from the acylation of
phenol with acetic acid. PA yield increases with increasing
W/FP0 but reaches a maximum value as it converts to larger
acylation products in secondary reactions with increasing
space time. The initial zero slope of the o-HAP yield curve
is consistent with o-HAP formation from conversion of pri-
mary PA products. Formation of p-HAP on HPA/C was not
significant in the W/FP0 range studied. Fig. 5B shows the
ηi0 vs W/FP0 plot obtained on ZSM5. In contrast with re-
sults obtained on HPA/C, formation of o-HAP on ZSM5
displays a nonzero initial slope, thereby suggesting the di-
rect C-acylation of phenol to o-HAP. The PA yield curve
shows an initial nonzero slope and then reaches a maximum,
indicating that PA is formed by direct O-acylation of phe-
nol and then is converted to secondary products. p-HAP is
formed in very small amounts on ZSM5. From the initial
slopes of the yield curves in Fig. 5B we determined the
initial formation rates of o-HAP and PA, respectively, and
but lower than on ZSM5P.Ap,0-HAP was a secondary product
formed on nonnegligible amounts on HY. In fact, ηp0-HAP in-
creased with contact time and the selectivity to p-HAP for
W/FP0 = 500 g h/mol was about 18%.
In order to gain more details on the reaction mecha-
nism of phenol acylation with acetic acid, we studied the
conversion of PA on ZSM5 (T = 553 K, P = 101.3 kPa,
PPA = 0.638 kPa). PA is the key intermediate to yield o-HAP
from phenol by a consecutive two-step mechanism. The aim
of this study was to investigate the direct transformation of
PA to p- and o-HAP via a Fries rearrangement mechanism on
zeolite ZSM5. Conversion of PA to p- and o-HAP isomers
by Fries rearrangement involves the intramolecular migra-
tion of the acyl group to the para- and ortho-positions of
the aromatic ring, respectively (Scheme 2A). The product
yields at t = 0 and the corresponding PA conversion values
on ZSM5 are shown as a function of contact time in Fig. 6.
PA was essentially converted to phenol. The total selectivity
0
∼
to o- and p-HAP isomers reached about 28% forXPA 40%,
=
which is consistent with previous results reported on acid
zeolites [35,36], but rapidly decreased for higher XP0A val-
ues and was negligible for XP0A = 80%, probably because of
consecutive transformation of both isomers. Besides, a rapid
activity decay for o-HAP formation was observed in con-
trast with the high stability that ZSM5 showed for the or-
tho-isomer synthesis from phenol and AA in Fig. 4B. We
conclude then that the gas-phase formation of o-HAP from
PA via a Fries rearrangement is not a significant reaction
pathway on ZSM5.
obtained ro0-HAP,0 = 2.43 × 10−3 mmol/(h m2) and rP0A,0
=
Finally, to obtain further insight on the P/AA acylation
mechanism an additional catalytic test was carried out on
ZSM5 by cofeeding phenol with PA (553 K, P = 101.3 kPa,
PPA = PP = 0.638 kPa, W/FP0 = 43.1 g h/mol). In Fig. 7
we plotted the evolution of PA conversion and product se-
lectivities as a function of time. Initially, PA is readily and
selectively converted to o-HAP showing that ZSM5 effi-
ciently catalyzes the acylation of phenol with PA. PA decom-
poses to phenol and forms simultaneously an acylium ion
which in turns attacks the phenol molecule to yield o-HAP
1.80×10−3 mmol/(hm2) (ri0,0, is the formation rate of prod-
uct i at XP → 0 and t = 0). Thus, on ZSM5 the direct
C-acylation rate of phenol to o-HAP is higher than the O-
acylation rate of phenol to PA. In a previous work using
zeolite ZSM5 of Si/Al = 40 and under experimental con-
ditions similar to those in the present work, Neves et al. [17]
found that O-acylation of phenol to the PA pathway was fa-
vored compared to that of C-acylation of phenol to o-HAP.
Qualitatively, zeolite HY showed similar evolutions of ηP0A

C.L. Padró, C.R. Apesteguía / Journal of Catalysis 226 (2004) 308–320
315
Scheme 2. Phenyl acetate conversion reactions. (A) Fries rearrangement; (B) P/PA acylation.
Fig. 7. Product distribution for phenol/phenyl acetate reactions on ZSM5.
Selectivities (S ) and phenyl acetate conversion (X ) as a function of
Fig. 6. Product distribution for phenyl acetate conversion reactions on
0
0
i
PA
ZSM5. Product yields (η ) and phenyl acetate conversion (X ) at
i
PA
time on stream [553 K, 101.3 kPa total pressure, P = P = 0.638 kPa,
P
PA
t = 0 as a function of contact time [553 K, 101.3 kPa total pressure,
0
P
W/F = 43.1 g h/mol].
P
= 0.638 kPa].
PA
(Scheme 2B). However, Fig. 7 also shows that the o-HAP
formation rate from P/PA acylation rapidly diminishes on
stream despite that ZSM5 is highly stable for forming o-
HAP from P/AA acylation (Fig. 4B).
cuss below the surface acidity requirements for the efficient
catalysis of reaction steps leading from phenol to o-HAP.
3.4. Surface acid property and catalytic performance
Results of Figs. 5–7 allow us to propose the reaction net-
work for the synthesis of o-HAP from the acylation of phe-
nol with AA described in Scheme 3. Phenol initially reacts
with AA via two parallel acylation reactions: by O-acylation
phenol is transformed to PA and by C-acylation yields di-
rectly o-HAP. Formation of o-HAP also takes place via sec-
ondary reactions from PA, either by Fries rearrangement or
by P/PA acylation. Regarding the production of p-HAP, only
zeolite HY formed p-HAP, but as a minor product (Table 3)
and via secondary reaction pathways (Fig. 5C). Neves et
al. [17] have proposed that p-HAP on acid zeolites is formed
through hydrolysis of p-acetoxyacetophenoneresulting from
the selective autoacylation of phenyl acetate. The relative
rate of the different pathways involved in Scheme 3 greatly
depends on the solid acid employed. In this regard, we dis-
The acylating agent or acylium ion CH3CO+ is formed
on our solid acids from AA either on Brønsted or Lewis
acid sites, as depicted in Scheme 4. Generated electrophilic
CH3CO+ may then attack the phenol molecule either by
electrophilic substitution of the ortho-hydrogen in the aro-
matic ring forming o-HAP or, alternatively, by O-acylation
of the OH group producing PA. Results in Fig. 5A showed
that on the HPA/C sample, which contains only surface
Brønsted sites, phenol is initially converted exclusively to
PA. This is explained by considering that on strong Brøn-
sted acid sites of HPA-based catalysts the phenol molecule
essentially interacts via the benzene ring [37], adopting a
position parallel to the surface and favoring the attack of

316
C.L. Padró, C.R. Apesteguía / Journal of Catalysis 226 (2004) 308–320
Scheme 3. Reaction network for the synthesis of o-HAP from phenol and acetic acid on solid acids.
acid sites, probably as shown in Scheme 5B. As proposed
by other authors [37,38], phenol is adsorbed on Lewis acid
centers predominantly in vertical orientation and allows the
electrophilic attack of benzene rings in the ortho-position,
because stabilization of the ortho-isomer intermediate is fa-
vored as compared to the intermediate formation in the para-
position [17]. This is consistent with the fact that we never
observed the direct formation of p-HAP from phenol on the
samples used in this work.
PA is consecutively converted to o-HAP on HPA/C, and
zeolites HY and ZSM5 samples (Fig. 5). After formation
and activation on Brønsted acid sites [35], PA may be trans-
formed in o-HAP either by intramolecular Fries rearrange-
ment or by its decomposition and consecutive phenol acyla-
tion (Scheme 6, intermolecular acylation). Our results show
that on zeolite ZSM5 the conversion of PA to o-HAP via
a Fries rearrangement is not a significant reaction pathway
(Fig. 6). We must conclude then that on ZSM5, and prob-
ably on zeolite HY, PA decomposes to phenol and acylium
ions which consecutively react and yield o-HAP by inter-
molecular C-acylation on Lewis acid sites, as described in
Scheme 6. In contrast, HPA-based catalysts would convert
PA to o-HAP essentially via an intramolecular Fries re-
arrangement on strong Brønsted acid sites (Scheme 6) be-
cause, as noted above, HPA catalysts do not promote the
acylation intermolecular pathway to obtain o-HAP by C-
acylation of phenol.
Scheme 4. Formation of the acylating agent on: (A) Brønsted acid sites, (B)
Lewis acid sites.
acylium ion mainly to the oxygen of phenol, as shown in
Scheme 5A.
On zeolites ZSM5 and HY both o-HAP and PA are pri-
mary products (Fig. 5), but phenol is directly converted to
o-HAP at higher (ZSM5) or similar (HY) rates than to PA.
Zeolite NaY that contains only Lewis acid sites produces
o-HAP with more than 60% selectivity for a phenol con-
version of about 16%, similarly to the values found on ze-
olites HY and ZSM5, which contain similar Lewis acid site
concentrations than NaY but L/B ratios between 1 and 1.5
(Table 2). These results suggest that C-acylation of phe-
nol to yield o-HAP takes place mainly on surface Lewis
In summary, catalysts containing Lewis and Brønsted
acid sites such as zeolites ZSM5 and HY exhibit a superior
performance for producing o-HAP because they efficiently
promote the two main reaction pathways leading from phe-
nol to o-HAP, i.e., the direct C-acylation of phenol, and
the O-acylation of phenol forming PA intermediates which

C.L. Padró, C.R. Apesteguía / Journal of Catalysis 226 (2004) 308–320
317
Scheme 5. O- and C-acylation of phenol on Brønsted (A) and Lewis (B) acid sites, respectively.
Scheme 6. Formation of o-HAP from PA via different reaction pathways.
are consecutively transformed via intermolecular P/PA C-
acylation.
3.5. Catalyst deactivation
Data in Table 3 show that although phenol conversion re-
mains practically constant on stream during the acylation of
phenol with AA, the selectivity to PA rapidly increases at
the expense of o-HAP over all the catalysts, except zeolite
ZSM5. In Fig. 8 we have plotted the evolution of the activity
a for the formation of o-HAP as a function of time on stream.
The activity a is defined as a = ro-HAP/ro0-HAP, where ro-HAP
and ro0-HAP are the formation rates of o-HAP at times t and
zero, respectively. Fig. 8 shows that a does not change with
time on ZSM5, but rapidly decreases on the other samples,
particularly on samples HY and HPA/MCM-41.
Coke formed on ZSM5, Al-MCM-41, HY, SiO2–Al2O3,
and HPA/MCM-41 was determined by analyzing the sam-
ples after the catalytic tests by temperature-programmed ox-
idation. The resulting TPO profiles are shown in Fig. 9. The
evolved CO2 from HPA/MCM-41 gave rise to a well-defined
Fig. 8. Acylation of phenol with acetic acid: Activity (a) for o-HAP
formation as a function of time [553 K, 101.3 kPa total pressure,
0
P
W/F = 146 g h/mol, P/AA = 1, N /(P + AA) = 45].
2
TPO peak with a maximum at about 795 K. The TPO profile
for zeolite HY showed that CO2 evolves in a broad band be-
tween 600 and 1000 K, containing three overlapping peaks.

318
C.L. Padró, C.R. Apesteguía / Journal of Catalysis 226 (2004) 308–320
Fig. 11. Initial deactivation of o-HAP formation rate, d = −(da/dt) , as
0
t=0
a function of conversion of phenol at t = 0 on HY, ZSM5, and HPA/C sam-
ples. Acylation of phenol with acetic acid [553 K, 101.3 kPa total pressure,
P/AA = 1, N /(P + AA) = 45].
2
deactivation, d0(h−1) = −(da/dt)t=0, as the initial slope of
the activity versus time curves obtained for different W/FP0
values on HY, HPA/C, and ZSM5 samples. The resulting d0
values are represented as a function of XP0 in Fig. 11. Zeo-
lite ZSM5 does not deactivate by increasing the conversion
of phenol up to 25%, but on HPA/C and HY samples d0 in-
creases with XP0, thereby suggesting that coke is formed via
consecutive side reactions from the products of the reaction.
As noted previously, zeolite ZSM5 does not deactivate
during the acylation of phenol with AA (Fig. 4B), despite
that it rapidly deactivates when PA or a P + PA mixture is
used as reactants (Fig. 7). This is explained by taking into
account that in the case that PA is added alone or in a P + PA
mixture, the conversion of PA to P occurs with simultane-
ous formation of ketenes (CH2=C=O) which are extremely
reactive and unstable compounds that dimerize to diketenes
and polymerize very quickly [39]. As a matter of fact, the
formation of ketenes is the main reason of the significant cat-
alyst activity decay observed during the liquid-phase Fries
rearrangement of PA on solid acids [10]. In contrast, the
acylation of phenol with AA produces concomitantly water
(Schemes 3 and 4) and any consecutive formation of ketenes
from PA is neutralized by the presence of moisture. In fact,
water rapidly reacts with ketenes to produce acetic acid [40]
and thereby suppresses the reaction pathway forming coke
from ketene precursors. Thus, we conclude that because of
the presence of water in the reaction products, ketenes are
not responsible for the catalyst activity decay shown in Fig. 8
for zeolite HY and HPA/C.
Fig. 9. Acylation of phenol with acetic acid. Temperature-programmed ox-
idation of the catalysts after the 6 h runs. 10 K/min heating rate.
Fig. 10. Activity for o-HAP formation at 5 h on stream (a ) as a function
5 h
of the amount of carbon formed during acylation of phenol with acetic acid.
Coke burnt on zeolite ZSM5 was qualitatively similar to that
of zeolite HY, but the amount of CO2 evolved from ZSM5
was significantly lower. Finally, the curves of CO2 evolved
from the oxidation of coke formed on samples Al-MCM-41
and SiO2–Al2O3 were qualitatively similar and both pre-
sented a maximum at about 870 K.
The amount of carbon on the samples was measured from
the area under the curves of Fig. 9 and the values ranged
from 29.4 mgC/g catalyst on ZSM5 to 153 mg C/g on HY.
In Fig. 10 we plotted the activity for the formation of o-HAP
at 5 h on stream (a5 h = ro5-HhAP/ro0-HAP) during the acylation
of phenol with AA as a function of the amount of carbon
on the sample determined from Fig. 9. It is observed that
a5 h decreases with the amount of carbon on the sample,
thereby suggesting that the activity decay for the formation
of o-HAP is caused by coke formation. Fig. 10 also shows
that lna5 h vs %C is a linear plot, suggesting that the o-HAP
formation rate diminishes exponentially with the amount of
carbon formed on the samples.
Finally, we studied the effect that feeding o-HAP alone or
in a o-HAP + AA mixture has on catalyst deactivation using
SiO2–Al2O3 and ZSM5 samples. o-HAP was fed on SiO2–
Al2O3 for 4 h (T = 553 K, Po-HAP = 0.45 kPa, W/Fo0-HAP
=
580 g h/mol). We did not detect the formation of any prod-
uct from o-HAP and the amount of coke formed after 4 h
on stream was not significant (%C = 1.2). It seems that o-
HAP is not strongly adsorbed on SiO2–Al2O3 and does not
form heavier products either. Thus, o-HAP would not be re-
sponsible by itself for the catalyst activity decay measured
on SiO2–Al2O3 in Fig. 8. The coinjection of o-HAP with
In an attempt to obtain insight on the species responsible
for coke formation we quantitatively evaluated the effect that
varying space time W/FP0 (and consequently XP0) has on o-
HAP formation rate decay by determining the initial catalyst

C.L. Padró, C.R. Apesteguía / Journal of Catalysis 226 (2004) 308–320
319
Scheme 7. Formation of coke precursors from o-HAP/AA conversion reactions.
AA (T = 553 K, PAA = 1.8 kPa, Po-HAP = 0.45 kPa) on
SiO2–Al2O3 formed phenol and two heavy unknown prod-
ucts. A rapid activity decay was observed and after 4 h on
stream the amount of carbon deposited was 13.5%. This ob-
served SiO2–Al2O3 deactivation is consistent with results
reported by Jayat and co-workers [17]. These authors, in
fact, found that coke formed on MFI zeolites during the acy-
lation of phenol with AA is mainly constituted by methyl-
naphthols, 2-methylchromone, and 4-methylcoumarin. For-
mation of 2m-cromone and 4m-coumarin may take place
from o-HAP and AA as depicted in Scheme 7, via the initial
formation of o-acetoxyacetophenone. In contrast, on ZSM5
we did not observe any catalyst deactivation by cofeeding o-
HAP with AA and the amount of coke formed after 4 h of
stream was only 1.5%. Besides, on ZSM5 we did not detect
the formation of heavy compounds among the reaction prod-
ucts (the only reaction products were PA and phenol). These
later results strongly suggest that on ZSM5 the formation of
coke precursors via the reaction shown in Scheme 7 does not
take place, probably because the narrow pore-size structure
of ZSM5 hinders the formation of condensed intermediate
compounds, such as o-AXAP.
Lewis acid sites promote also the direct C-acylation of phe-
nol in the ortho-position to yield o-HAP. Zeolites ZSM5 and
HY, which contain strong Lewis and Brønsted acid sites,
produce o-HAP at high rates because they efficiently cat-
alyze the two main reaction pathways leading from phenol
to o-HAP, i.e., the direct C-acylation of phenol and the O-
acylation of phenol forming the PA intermediate which is
consecutively transformed via intermolecular phenol/PA C-
acylation. Solid acids of moderate acid strength such as Al-
MCM-41 or SiO2–Al2O3 are less active and selective for
o-HAP formation than zeolites HY or ZSM5.
All the samples used in this work, except zeolite ZSM5,
rapidly deactivate on stream because of coke formation.
Coke precursors are formed from secondary condensation
reactions, probably between o-HAP and acetic acid, and not
from any ketene formation by phenyl acetate decomposi-
tion. The superior stability of zeolite ZSM5 is explained by
considering that its narrow pore-size structure hampers the
formation of coke precursor compounds.
Acknowledgments
We thank the Universidad Nacional del Litoral (UNL),
Consejo Nacional de Investigaciones Científicas y Técnicas
(CONICET), and Agencia Nacional de Promoción Cientí-
fica y Tecnológica (ANPCyT), Argentina, for the financial
support of this work. We are grateful to H. Cabral for IR
spectroscopy measurements.
4. Conclusions
Gas-phase acylation of phenol with acetic acid on solid
acids produces essentially o-hydroxyacetophenone and phe-
nyl acetate, together with minor amounts of p-hydroxyaceto-
phenone, but the product formation rates, the reaction mech-
anism, and the catalyst activity decay greatly depend on the
surface acid site density and strength of the solids. HPA-
based catalysts, such as HPA/C or HPA/MCM-41, that con-
tain only strong Brønsted acid sites, promote exclusively
the O-acylation of phenol forming phenyl acetate which
is then consecutively transformed to o-HAP, probably via
an intramolecular Fries rearrangement. Samples containing
References
[1] J. Fritch, O. Fruchey, T. Horlenko, US patent 4954652, Hoechst
Celanese Corporation (1990).
[2] I. Uwaydah, M. Aslam, C. Brown, S. Fitzhenry, J. McDonough, US
patent 5696274 (1997).
[3] M.J. Climent, A. Corma, S. Iborra, J. Primo, J. Catal. 151 (1995) 60.

320
C.L. Padró, C.R. Apesteguía / Journal of Catalysis 226 (2004) 308–320
[4] M.T. Drexler, M.D. Amiridis, J. Catal. 214 (2003) 136.
[5] K.G. Davenport, Ch.B. Hilton, US patent 4524217, Celanese Corp.
[22] J.A. Lercher, Ch. Gründlig, G. Eder-Mirth, Catal. Today 27 (1996)
353.
[23] P.A. Jacobs, R. von Ballmoos, J. Phys. Chem. 86 (1982) 3050.
[24] M.T. Aronson, R.J. Gorte, W.E. Farneth, J. Catal. 105 (1987) 455.
[25] A. Corma, V. Fornés, F. Rey, Zeolites 13 (1993) 56.
[26] A. Boréave, A. Aroux, C. Guimon, Micropor. Mater. 11 (1997) 275.
[27] E.P. Parry, J. Catal. 2 (1963) 371.
[28] J.W. Ward, J. Catal. 10 (1968) 34.
[29] H. Knözinger, Adv. Catal. 25 (1976) 184.
[30] J. Wang, L. Huang, H. Chen, Q. Li, Catal. Lett. 55 (1998) 157.
[31] A. Sakthivel, S.E. Dapurkar, N.M. Gupta, S.K. Kulshreshtha, P. Sel-
vam, Micropor. Mesopor. Mater. 65 (2003) 177.
[32] G. Busca, Catal. Today 41 (1998) 191.
[33] R.A. Beyerlin, G.B. Mc Vicker, L.N. Yacullo, J.J. Ziemak, J. Phys.
Chem. 92 (1988) 1967.
[34] G.I. Woolery, G.H. Kuehl, H.C. Timken, A.W. Chester, J.C. Vartuli,
Zeolites 19 (1997) 288.
[35] V. Pouilloux, J.-P. Bodibo, I. Neves, M. Gubelnann, G. Perot, M. Guis-
net, in: M. Guisnet, et al. (Eds.), Heterogeneous Catalysis and Fine
Chemicals II, Elsevier, Amsterdam, 1991, p. 513.
[36] V. Borzatta, G. Busca, E. Poluzzi, V. Rossetti, M. Trombetta, A. Vac-
cari, Appl. Catal. A 257 (2004) 85.
(1985).
[6] A. Commarieu, W. Hölderich, J.A. Laffitee, M. Dupont, J. Mol. Catal.
A: Chem. 182 (2002) 137.
[7] C.S. Cundy, R. Higgins, S.A. Kibby, B.M. Lowe, R.M. Paton, Tetra-
hedron Lett. 30 (1989) 2281.
[8] A. Vogt, H. Kouwenhoven, R. Prins, Appl. Catal. 123 (1995) 37.
[9] F. Jayat, M.J. Sabater Picot, M. Guisnet, Catal. Lett. 41 (1996) 181.
[10] A. Heidekum, M.A. Harmer, W.F. Hölderich, J. Catal. 176 (1998) 260.
[11] E.F. Kozhevnikova, J. Quartararo, I.V. Kozhevnikov, Appl. Catal.
A 245 (2003) 69.
[12] J. Mueller, W. Wiersdorff, W. Kirschenlohr, G. Schwantje, US patent
4508924, BASF (1985).
[13] G.N. Mott, US patent 4607125, Celanese Corporation, 1986.
[14] M. Guisnet, D.B. Lukyanov, F. Jayat, P. Magnoux, I. Neves, Ind. Eng.
Chem. Res. 34 (1995) 1624.
[15] B.B. Gupta, US patent 4668826, Celanese Corporation, 1987.
[16] K.G. Bhattacharyya, A.K. Talukdar, P. Das, S. Sivasanker, Catal. Com-
mun. 2 (2001) 105.
[17] I. Neves, F. Jayat, P. Magnoux, G. Perot, F.R. Ribeiro, M. Gubelman,
M. Guisnet, J. Mol. Catal. 93 (1994) 169.
[37] K.V.R. Chary, K. Ramesh, G. Vidyasagar, V.V. Rao, J. Mol. Catal. A:
Chem. 198 (2003) 195.
[18] F. Jayat, M. Guisnet, M. Goldwasser, G. Giannetto, Stud. Surf. Sci.
Catal. 105 (1997) 1149.
[19] Y.V. Subba Rao, S.J. Kulkarni, M. Subrahmanyam, A.V. Rama Rao,
Appl. Catal. A 133 (1995) L1.
[20] K.J. Edler, J.W. White, Chem. Mater. 9 (1997) 1226.
[21] I.V. Kozhenikov, K.R. Kloestra, A. Sinnema, H.W. Zandbergen, H. van
Bekkum, J. Mol. Catal. A: Chem. 114 (1996) 287.
[38] K. Tanabe, T. Nishizaki, in: G.C. Bond, P.B. Wells, F.C. Tompkins
(Eds.), Proceedings of the Sixth International Conference on Catalysis,
vol. 2, The Chemical Society, London, 1977, p. 863.
[39] T.T. Tidwell, Acc. Chem. Res. 23 (1990) 273.
[40] E. Bothe, A.M. Dessouki, D. Schulte-Frohlinde, J. Phys. Chem. 84
(1980) 3270.