The transformations involving methanol in the acid- and base-catalyzed gas-phase methylation of phenol

Article abstract of DOI:10.1016/j.jcat.2007.07.033

The alkylation of phenol with methanol was studied using a Bronsted-type acid catalyst (a H-mordenite) and basic/dehydrogenating catalysts (MgO, Fe₂O₃ and Mg/Fe/O), with the aim of investigating the reaction mechanism. The main difference between the two classes of catalysts concerned the transformations occurring on methanol. Specifically, in the former case the acid-type activation of methanol led to the development of an electrophylic species that gave rise to the formation of anisole and of C-alkylated compounds. With basic catalysts, methanol dehydrogenated to formaldehyde, which then underwent transformation to methylformate and to decomposition products, i.e., CO, CO₂, CH₄ and H₂. In this case, the prevailing compounds obtained by reaction with phenol were o-cresol and 2,6-xylenol. The dehydrogenation of methanol was found to be the key-step in the generation of the active methylating species with basic catalysts.

Full text of DOI:10.1016/j.jcat.2007.07.033

Journal of Catalysis 251 (2007) 423–436
www.elsevier.com/locate/jcat
The transformations involving methanol in the acid- and base-catalyzed
gas-phase methylation of phenol
a,1
a,∗,1
a,1
a,1
a,1
a,1
N. Ballarini , F. Cavani
, L. Maselli , A. Montaletti , S. Passeri , D. Scagliarini ,
b
b
C. Flego , C. Perego
a
Dipartimento di Chimica Industriale e dei Materiali, Alma Mater Studiorum Università di Bologna, Viale Risorgimento 4, 40136 Bologna, Italy
b
ENI SpA, Div. R&M, Via Maritano 26, 20097 S. Donato Milanese (MI), Italy
Received 1 March 2007; revised 17 July 2007; accepted 29 July 2007
Available online 14 September 2007
Abstract
The alkylation of phenol with methanol was studied using a Brønsted-type acid catalyst (a H-mordenite) and basic/dehydrogenating catalysts
MgO, Fe O and Mg/Fe/O), with the aim of investigating the reaction mechanism. The main difference between the two classes of catalysts
(
2
3
concerned the transformations occurring on methanol. Specifically, in the former case the acid-type activation of methanol led to the development
of an electrophylic species that gave rise to the formation of anisole and of C-alkylated compounds. With basic catalysts, methanol dehydrogenated
to formaldehyde, which then underwent transformation to methylformate and to decomposition products, i.e., CO, CO , CH and H . In this case,
2
4
2
the prevailing compounds obtained by reaction with phenol were o-cresol and 2,6-xylenol. The dehydrogenation of methanol was found to be the
key-step in the generation of the active methylating species with basic catalysts.
©
2007 Elsevier Inc. All rights reserved.
Keywords: Basic catalysis; Acid catalysis; Phenol methylation; Methanol dehydrogenation
1
. Introduction
Preferred alkylating agents are methanol and dimethylcar-
bonate, while more conventional reactants, such as methylchlo-
ride and dimethylsulfate, although still employed industri-
ally, are nowadays less attractive due to environmental con-
cerns [2–4].
The methylation of phenol and phenol derivatives has been
widely investigated, and several papers report about the effect
of the catalyst characteristics on the nature and distribution of
products. The reaction has a great relevance from the industrial
point of view [1]; for instance, 2,6-xylenol is the monomer for
the production of poly-(2,6-dimethyl)phenylene oxide resin; 2-
methylphenol (o-cresol) is the monomer for the synthesis of
epoxycresol novolacks; 2,5-dimethylphenol is the intermedi-
ate for the synthesis of dyes, antiseptics and antioxidants, and
The C-methylation of phenol aimed at the production of o-
cresol or of 2,6-xylenol is industrially carried out with methanol
as the alkylating agent and with catalysts possessing basic char-
acteristics [5–13]. Catalysts are made of either (i) supported
and unsupported alkali and alkaline-earth metal oxides [14–18],
or (ii) transition or post-transition mixed metal oxides [19–33],
or (iii) mixed oxides containing both alkaline-earth metals and
transition metal ions [34–44]. They are used in prevalence for
gas-phase methylation, since under liquid-phase conditions a
low conversion is generally achieved. This is attributed to the
lower reaction temperature, and to the stronger interaction that
develops between the catalyst and phenol in the condensed
phase.
2
,3,6-trimethylphenol is the starting compound for the prepa-
ration of vitamin E. The products of O-methylation of phenol
anisole) and of diphenols (e.g., guaiacol) are intermediates in
(
the production of skin protection agents and food additives.
*
Corresponding author.
E-mail address: fabrizio.cavani@unibo.it (F. Cavani).
1
The strongest basic catalysts are alkali and alkaline-earth
metal oxides, which however may deactivate by interaction
Also at: INSTM, Research Unit of Bologna, a Partner of Idecat NoE (FP6
of the EU).
0021-9517/$ – see front matter © 2007 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2007.07.033

424
N. Ballarini et al. / Journal of Catalysis 251 (2007) 423–436
with weak acid molecules, included carbon dioxide and water.
Therefore, preferred basic materials for industrial applications
are transition metal oxides (e.g., supported V/Fe mixed ox-
ides [45], or Cr oxide [46]) that are claimed to exhibit medium-
strength basic properties [1,45–47]. Similar reactivity is exhib-
wise added to another solution containing 39.75 g of Na2CO3
(Carlo Erba Reagenti) dissolved in 375 ml of distilled water.
While adding the first solution to the second one, the pH was
continuously adjusted, in order to keep it close to 10.0. Under
these conditions the precipitation of Mg(OH)2 occurred. The so
obtained slurry was left under stirring for 40 min; then the pre-
cipitate was separated from the liquid by filtration, and washed
ited by those mixed oxides that couple the O2 nucleophilicity,
typical of alkali or alkaline-earth metal oxides, with the coor-
dinating properties of transition metal oxides [1,47–49]. One
−
◦
with 6 L of distilled water at 40 C. The solid was then dried
+
◦
◦
example is MgO, in which the H -abstracting properties are
at 110 C overnight, and calcined at 450 C for 8 h in air. The
preparation of Fe2O3 was carried out with the same procedure,
using Fe(NO3)3·9H2O (Carlo Erba Reagenti) as starting mate-
rial. The same protocol was also used for the preparation of the
Mg/Fe mixed oxide, using the corresponding amount of the two
salts to obtain the desired atomic ratio between components.
The acid catalyst was a commercial H-mordenite, having
modulated through the introduction of host cations [35,44].
The addition of increasing amounts of Fe3 to MgO in sam-
ples prepared by thermal decomposition of hydrotalcite-like
precursors, with Mg/Fe atomic ratio between 2 and 6, lead
to the formation of solid solutions having general composi-
tion Mg1−xFexO1+0.5x and medium-strength basic sites [50].
+
3+
ꢀꢀ
The latter derived from the higher electronegativity of Fe as
an atomic Si/Al ratio equal to 20, shaped in 1/16 extru-
2+
compared to Mg , that decreased the charge density on the
dates (binder alumina). This sample was supplied by Süd-
Chemie AG.
neighboring O2 and made the latter less nucleophilic than O
−
atoms in MgO [50].
The XRD powder patterns of the catalysts were taken with
Ni-filtered CuKα radiation (λ = 1.54178 Å) on a Philips
X’Pert vertical diffractometer equipped with a pulse height an-
alyzer and a secondary curved graphite-crystal monochromator.
Surface area was measured by means of the BET single-point
method (N2 adsorption at the temperature of liquid N2), using
a Sorpty 1750 Fisons Instrument.
Characteristics of catalysts possessing basic features when
used as catalysts for phenol methylation are: (i) the very high
regio-selectivity in C-methylation, since the ortho/para-meth-
ylation ratio is in all cases largely higher than 2, and (ii) the high
chemo-selectivity, since the O/C-selectivity ratio, a function of
the basic strength of catalysts, is, in general, very low. A selec-
tivity to o-cresol + 2,6-xylenol as high as 98% is reported in
many patents [1].
One major problem of the industrial process of phenol
methylation is the low yield with respect to methanol, due to its
decomposition; consequently, a large excess of methanol is usu-
ally fed in order to reach an acceptable per-pass conversion of
phenol. Various solutions have been proposed to minimize this
side reaction (see, for example, [48]), amongst which the co-
feeding of water seems to be the most effective [45,49]. This as-
pect, however, is often forgotten in scientific literature, and only
few papers take into consideration the methanol decomposition
Temperature-programmed-reaction tests were performed
with a TPDRO 1100 ThermoQuest Instruments, by saturation
◦
of a He stream with methanol at 6 C, and feeding the gaseous
stream continuously to the reactor, while heating the catalyst
◦
◦
from 200 to 450 C (heating rate: 10 C/min). The effluents
were analyzed by means of a VG quadrupole. The intensity of
the following ion current signals were recorded: H2 (m/z = 2),
H2O (18), CO (28), CO2 (44), CH3OH (31 > 32 > 29 ꢁ 30),
CH4 (16), HCOOH (46), CH3OCOH (60) and H2CO (29 >
30 > 28). The most intense m/z signals of formaldehyde (29
and 30) are also typical of methanol; therefore, the pressure
of H2CO was extrapolated from the comparison of the inten-
sity profiles for signals at m/z = 29 and 30, with that one of
m/z = 31.
[
19,34,51–53] and the transformations that occur on the alkylat-
ing alcohol. On the other hand, a wide literature demonstrates
that methanol undergoes different transformations on metal ox-
ides, depending on the catalyst surface properties [13,54,55].
In the present work we correlate the catalytic performance
in the gas-phase methylation of phenol with the transforma-
tions that occur on methanol, for one acid (H-mordenite) and
three basic catalysts (MgO, Fe2O3 and Mg/Fe mixed oxide),
also possessing dehydrogenating properties. More specifically,
the aim was to determine if the nature of the methylating species
can be different when either acid or basic systems are used, and
which implications this may have on the catalytic performance
in phenol methylation.
Catalytic tests were carried out by vaporization of a metha-
nol/phenol liquid mixture (methanol/phenol molar ratio: 10/1;
liquid flow: 0.0061 ml/min; phenol supplied by Sigma Aldrich,
99+% purity; methanol supplied by Carlo Erba Reagenti) in a
N2 stream (gas flow: 20 N ml/min). The composition of the
feed gas was the following (molar fractions): methanol 0.108,
phenol 0.011, nitrogen 0.881. Overall gas residence time was
2.68 s. Total pressure was atmospheric. A high methanol/phenol
feed ratio (10/1) was used in order to better evidence the par-
allel reactions occurring on methanol, (i) dehydrogenation and
decomposition with basic catalysts, and (ii) formation of alky-
laromatics with zeolites. Some tests were also made with a
methanol/phenol feed ratio equal to 5/1. For tests made in ab-
sence of phenol, the inlet stream of methanol was maintained
the same as that one used in phenol methylation tests.
2
. Experimental
Fe/O, Mg/Fe/O and Mg/O catalysts were prepared by precip-
itation from an aqueous solution containing the corresponding
metal nitrates. For instance, to obtain 15 g of MgO, 96.15 g
of Mg(NO3)2·6H2O (Carlo Erba Reagenti, 99% purity) were
dissolved in 375 ml of distilled water. The solution was drop-
The gas/vapors stream was fed to a stainless steel reac-
ꢀ
ꢀ
3
tor (length: 30 cm, internal diameter: 3/4 ), containing 1 cm
of catalyst (catalyst weight: Mg/O 0.85 g, Mg/Fe/O 0.95 g,

N. Ballarini et al. / Journal of Catalysis 251 (2007) 423–436
425
Table 1
Main characteristics of catalysts employed in the present work
2
Catalyst (atomic ratio)
Surface area (BET), m /g,
Crystalline phases (XRD)
before reaction; after reaction
before reaction; after reaction
Mg/Fe/O (Mg/Fe = 0.19)
112; 66
36; 32
68; 60
Fe O and MgFe O ; Fe O and/or MgFe O
2 3 2 4 3 4 2 4
Fe/O
Fe O hematite; Fe O magnetite
2 3 3 4
Mg/O
Mg(OH) brucite + hydroxymagnesite
2
Mg (CO ) (OH) H O; MgO periclase
5
3 4
24 2
H-mordenite (Si/Al = 20)
372; 56
H-mordenite
Fe/O 1.10 g and H-mordenite 0.64 g) shaped either in 30–60
mesh particles (Mg/O, Fe/O and Mg/Fe/O) or in extrudates (H-
mordenite). Catalyst particles were prepared by pressing the
calcined powder to obtain pellets that were then broken into
smaller granules. During catalytic measurements, the reactor
• Conversion of phenol or of methanol:
in
out
n˙
−n˙
phenol or methanol
phenol or methanol
.
in
n˙
phenol or methanol
• Selectivity to a compound is expressed as the ratio between
exit stream was condensed in 25 ml of HPLC-grade acetone for
the corresponding yield and the reactant conversion.
◦
1
h, maintained at 6 C. Products condensed in acetone were
analyzed by gas chromatography, using a GC6000 Carlo Erba
instrument equipped with a FID and a HP-5 column. The GC
3. Results
◦
oven temperature was programmed from 50 to 250 C, with a
3
.1. The characterization of catalysts
◦
heating rate of 10 C/min. Non-condensable gases (CO, CO2,
H2, CH4) were analyzed by sampling the gaseous stream with a
syringe at the reactor exit, before condensation in acetone, and
by injecting the sample in a GC 4300 Carlo Erba gas chromato-
graph, equipped with a TCD and a Carbosieve SII column. The
Table 1 summarizes the main characteristics of catalysts
employed in the present work: (i) Magnesium oxide (sample
Mg/O) is the main component in alkaline-earth oxides-based
catalysts employed in the General Electric process for the syn-
thesis of 2,6-xylenol [47–49]. (ii) Iron oxide (sample Fe/O) is
claimed to be the main component in the optimal catalyst for
the selective ring methylation of phenol and 1-naphthol [22,
27,56]. (iii) Magnesium-iron mixed oxide (sample Mg/Fe/O)
catalyzes the liquid-phase and the gas-phase methylation of m-
cresol [42,44] and of phenol [35,57]. In the case of the Mg/Fe/O
catalyst used in the present work, the choice was for a sample
having an excess of Fe, in which the chemical–physical proper-
ties of iron oxide are modified by the presence of a low amount
◦
GC oven temperature was programmed from 55 to 220 C, with
◦
a heating rate of 10 C/min.
Before carrying out the reactivity tests, Mg/O, Fe/O and
Mg/Fe/O catalysts were “equilibrated” in the reaction feed at
◦
3
90 C for 10 h, in order to obtain a stable catalytic perfor-
mance. With the H-mordenite, tests were run without any pre-
liminary equilibration, because this catalyst exhibited a slow,
continuous deactivation and it was not possible to reach a sta-
ble catalytic performance.
of Mg2 [58]; the Mg/Fe ratio, as determined in the calcined
sample by SEM/EDX, was 0.19. Finally, the acid catalyst inves-
tigated was a commercial H-mordenite, having a Si/Al atomic
ratio of 20.
+
Yields are expressed as follows:
out
n˙
product
in
•
Yield to products of phenol methylation:
, where
n˙
phenol
“
product” stands for o-cresol, p-cresol, 2,6-xylenol, 2,4-
Fig. 1 compares the X-ray diffraction patterns of the three
basic catalysts, both before (fresh samples) and after (spent
samples) reaction. In the case of Mg/O, before reaction the
oxide was strongly carbonated and hydrated (hydroxymagne-
site, Mg5(CO3)4(OH)24H2O, JCPDS file 25-0513, and brucite
Mg(OH)2, 44-1482), whereas after reaction the pattern did cor-
respond to that of MgO (periclase, 04-0829). The XRD pattern
of the fresh Mg/Fe/O sample was not well resolved, indicating
the presence of a large fraction of amorphous compound; re-
flections corresponding to both MgFe2O4 (36-0398) and Fe2O3
hematite (13-0534) were present. After reaction, reflections at-
tributed to Fe O had disappeared, while those correspond-
xylenol, anisole and polyalkylated phenols.
out
n˙
product
in
•
•
Yield to products of methanol decomposition:
where “product” stands for CO, CO2 and CH4.
,
n˙
methanol
Yield to products of methylformate decomposition:
out
n˙
product
,
where “product” stands for CO, CO2 and CH4.
in
2n˙
metylformate
•
•
Yield to H2 from methanol or from methylformate decom-
out
n˙
H
2
position: 2
n˙
.
in
methanol or methylformate
out
ꢀ
n˙
n˙
producti
Yields to alkylaromatics from methanol:
αi
,
i
in
2
3
methanol
where “producti” stands for: toluene, dimethylbenzenes,
trimethylbenzenes, tetramethylbenzenes, pentamethylben-
zene and hexamethylbenzene; αi is the number of C atoms
for each compound.
3 4
ing to either magnetite Fe O (19-0629) or the spinel phase
MgFe O (the two patterns are almost coincident) were present.
2
4
The same occurred in the case of the Fe/O sample; the XRD pat-
tern of the sample before reaction was that of Fe O , while after
2
3
reaction it corresponded to that of Fe3O4. The partial reduction
of hematite in the reaction environment, when used as the cat-
Conversions are expressed as follows:

426
N. Ballarini et al. / Journal of Catalysis 251 (2007) 423–436
in which the incorporation of a guest cation in the oxide of
a differently charged metal causes the development of struc-
tural defects and hence of less crystalline materials. The spent
Mg/Fe/O catalyst had a lower amount of amorphous component
than the fresh one; the increase of crystallinity led to a decrease
of the surface area, that however remained higher than that of
Fe/O.
3
.2. The transformation of methanol with basic catalysts:
steady-state performance
The reactivity of catalysts towards methanol was investi-
gated. Tests were carried out by feeding methanol vapors di-
luted in a N2 stream. Fig. 2 reports the yields of CO, CO2, CH4
and H2 for Mg/O, Fe/O and Mg/Fe/O samples as functions of
temperature. There was no formation of either dimethylether or
hydrocarbons (with the exception of methane).
Mg/O gave an appreciable conversion of methanol only at
◦
◦
4
50 C, while Fe/O was even active at 390 C; Mg/Fe/O yielded
◦
large amounts of decomposition products at 350 C. From the
comparison of the methanol conversion at 390 C, the following
◦
ranking was obtained: Mg/Fe/O > Fe/O ꢁ Mg/O. The higher
activity of Mg/Fe/O as compared to Fe/O can be attributed to
2
its higher surface area (66 versus 32 m /g, respectively, for
the spent catalysts), but a modification of the dehydrogenating
properties of Fe due to the presence of Mg cannot be excluded
[
27,33,35].
H2 is a co-product in the formation of the C-containing
compounds. The following stoichiometries can be assumed, for
calculation purposes only:
CH3OH → CO + 2H2,
CH3OH → 1/2CO2 + 1/2CH4 + H2,
where the first reaction corresponds to methanol decomposition
and lumps together the dehydrogenation to formaldehyde and
then to CO. The second one is the overall stoichiometry for
the transformation of methanol to formaldehyde, followed by
dimerization to yield methylformate and decomposition of the
latter into CH4 and CO2. Therefore, from a mass-balance point
of view, all reactions that lead to the formation of H2 starting
from CH3OH (in the absence of any co-reactant) are included
in these two reactions, when the assumption is made that there
is no formation of coke on catalysts (the accumulation of which
would lead to the additional formation of H2). In other words,
the amount of H2 produced should correspond to the sum of
2
CO + CH4 + CO2, and the yield of H2 to the sum of yields
of CO + 1/2CH4 + 1/2CO2. An excess of H2 with respect
to this sum might be due to the presence of non-decomposed
formaldehyde, formic acid or methylformate.
Formic acid may form either through the Cannizzaro reac-
tion
Fig. 1. X-ray diffraction patterns of fresh (calcined) samples and of spent Mg/O,
Fe/O and Mg/Fe/O.
alyst for aryl methylation, was also reported by other authors
[
59,60].
The addition of Mg2 to the Fe oxide led to a considerable
+
2CH O + H O → HCOOH + CH OH
2
2
3
increase of the surface area, as compared to Fe/O (Table 1);
this difference was present also in the spent catalysts. This may
derive from the lower crystallinity of Mg/Fe/O as compared
to Fe/O. This aspect is typical of Mg/Fe mixed oxides [42],
or through oxidation of formaldehyde by a redox-type cation,
e.g., Fe3 ; in the latter case the progressive reduction of the
cation would cause a decrease of the contribution of this re-
action. In our case, water was not fed to the reactor, neither it
+

N. Ballarini et al. / Journal of Catalysis 251 (2007) 423–436
427
mation of methylformate as the main intermediate, which then
decomposed into the two light products.
In the case of Mg/Fe/O and Fe/O, the distributions of prod-
ucts obtained with the two catalysts were similar. Also in this
case, the yield of H was not much different from the sum:
2
yield CO + 1/2 yield CH4 + 1/2 yield CO2. Indeed, very small
amounts of formaldehyde and methylformate were detected by
GC–MS. Moreover, the yield of CO was higher than that of
2
CH4; this points out for an additional mechanism for carbon
dioxide formation, for instance the decomposition of formate.
A solution of CH OH (30 wt%) in H O was vaporized in a
3
2
N2 stream and fed over the Mg/Fe/O catalyst. A lower selectiv-
ity to CH4 and an increased formation of CO2 were observed
with respect to results obtained in the absence of water. This in-
dicates that the presence of steam favored the Cannizzaro-type
disproportionation of formaldehyde and hence the formation of
formic acid; the latter in part yielded methylformate, while in
part decomposed to CO + H . On the contrary, in the absence
2
2
of H2O the preferred pathway for the formation of formate is
likely an oxidation by Fe3 , occurring on the fresh, oxidized
+
3+
2+
catalyst; the progressive reduction of Fe to Fe makes this
contribution less and less important.
3
.3. The transformation of methanol with basic catalysts:
non-steady-state performance
With the aim of determining the reaction intermediates in
methanol decomposition, and of confirming that the light com-
pounds identified (CO, CH , CO and H ) were due to the
4
2
2
dehydrogenation of methanol to formaldehyde and to the con-
secutive reactions occurring upon the latter, we carried out non-
steady-state tests by feeding methanol over the catalyst, while
raising the reaction temperature (temperature-programmed-
reaction tests). Figs. 3 and 4 show the variation of the pressure
for selected products (as evaluated from the intensity of the ion
current signal at corresponding m/z values in mass spectra),
obtained with Mg/O and Mg/Fe/O, respectively.
In the case of Mg/O, the formation of H became relevant
2
◦
above 300–350 C, in concomitance with the increase of CO.
The pressure of CO2 was very low and decreased above 350 C.
◦
No methylformate formed; the amount of HCOOH was very
low, since its pressure was 3 orders of magnitude lower than
that of CO and CO2. The amount of formaldehyde, as inferred
from the comparison of the signals at m/z = 29, 30 and 31, was
negligible.
Fig. 2. Yields of CO (2), CO (Q), CH (×) and H2 (F) as functions of the
2
4
temperature. Feed composition (molar fractions): methanol 0.12, nitrogen 0.88.
Catalysts: Mg/O (top), Fe/O (middle) and Mg/Fe/O (bottom).
was produced; in fact, there was no formation of dimethylether.
Methylformate may form either by esterification of formic acid
with methanol or by direct dimerization of formaldehyde (Tis-
chenko reaction).
Therefore, taking into consideration the information
achieved with steady-state and non-steady-state tests, it can
be inferred that Mg/O dehydrogenates methanol to formalde-
◦
hyde above 300 C, even though with a very low yield, and
In the case of Mg/O (Fig. 2, top), at high temperature the
that methylformate is rapidly formed through the dimerization
mechanism. At the high temperature required for its formation,
methylformate immediately decomposes yielding CO + CH .
yield of H approximately corresponded to the sum of yields
2
of CO + 1/2CH4 + 1/2CO2. Therefore, formaldehyde, formic
acid and methylformate were not produced. This indicates that
these intermediates are very reactive under the applied condi-
tions and quickly evolve towards the final products. The yields
of CH4 and of CO2 were similar, in agreement with the for-
2
4
The Cannizzaro reaction, followed by esterification, is another
possible mechanism for methylformate formation; however,
this reaction is unlikely, due to the fact that water was not fed
to the reactor.

428
N. Ballarini et al. / Journal of Catalysis 251 (2007) 423–436
Fig. 4. Pressure of selected molecular ions as functions of temperature in tem-
perature-programmed-reaction tests. Top: selected substances. Bottom: mole-
cular ion and main fragments for methanol and formaldehyde. Catalyst: Mg/
Fe/O. Feed methanol, carrier He.
Fig. 3. Pressure of selected molecular ions as functions of temperature in tem-
perature-programmed-reaction tests. Top: selected substances. Bottom: molec-
ular ion and main fragments for methanol and formaldehyde. Catalyst: Mg/O.
Feed methanol, carrier He.
3
.4. The transformation of methanol with the acid catalyst
As far as tests with Mg/Fe/O concern (Fig. 4), the amount
of each product was higher than that obtained with Mg/O, con-
The tests of methanol transformation were also carried out
with the H-mordenite. In this case, two different classes of
compounds formed: (i) alkylbenzenes, ranging from toluene
to hexamethylbenzene, and (ii) light decomposition products.
Fig. 5 (top) plots the conversion of methanol, the overall yield
of alkylaromatics and that of light compounds. Fig. 5 (middle)
reports the detail of the yields of the light compounds.
firming the results obtained with steady-state tests. H O and CO
2
formed in relevant amount even at low temperature, when the
formation of HCOOH occurred. Therefore, under these condi-
tions the decomposition of formic acid into CO + H O is prob-
2
able. The detection of formic acid confirms the high activity of
Mg/Fe/O in the dehydrogenation or oxidative dehydrogenation
of methanol to formaldehyde; the latter was then oxidized by
Alkylaromatics were the prevailing compounds in the tem-
◦
3
+
perature range comprised between 250 and 350 C; in this range
Fe to yield the formate species.
the formation of light decomposition products was nil (only
methane formed). In a recent work [61] we reported that in
the gas-phase methylation of phenol catalyzed by H-BEA ze-
olites, the main products of methanol transformation were pen-
tamethyl and hexamethylbenzene; the latter, and polynuclear
aromatics as well, were responsible for the progressive catalyst
deactivation.
The partial pressures of CO2 and CH4 had the same trend
above 300–350 C, confirming the presence of one common
◦
intermediate, i.e., methylformate. It is worth mentioning that
methylformate was not detected even under non-steady con-
ditions, probably because of its high reactivity and rapid de-
composition to light compounds. The decrease of intensity of
◦
the m/z = 31 signal above 350 C was due to a relevant in-
◦
crease of methanol conversion. The relative decrease of in-
tensity for signals at m/z = 29 and 30 was lower than that
of m/z = 31; this indicates the presence of low, but non-
negligible, amounts of formaldehyde.
With the H-mordenite, at above 350 C, the formation of
alkylaromatics decreased considerably and the main product
was methane, with minor amounts of CO, CO2 and H2. The
plot of methanol conversion as a function of temperature also

N. Ballarini et al. / Journal of Catalysis 251 (2007) 423–436
429
Fig. 6. Top: conversion of phenol (2), selectivity to o-cresol ("), p-cresol (Q),
,6-xylenol (F) and anisole (×) as functions of temperature. Bottom: yields
2
of CO (2), CO (Q), CH (×) and H (F) as functions of temperature. Feed
2
4
2
composition (molar fractions): methanol 0.108, phenol 0.011, nitrogen 0.881.
Catalyst: Mg/O.
formaldehyde and decomposition; however, methanol yielded
small amounts of alkylbenzenes. At higher temperature, methyl-
formate decomposed to CH4 + CO2. The formation of H2 may
derive either from the decomposition of formic acid or from
the dehydrogenation of methanol, both formed by hydrolysis
of methylformate (H2O is the co-product in the formation of
alkylbenzenes).
3
.5. The methylation of phenol with the basic catalysts
Fig. 5. Top: conversion of methanol (F), yield of alkylaromatics (Q) and yield
of light compounds (") as functions of temperature. Middle: yields of CO (2),
CO (Q), CH (×) and H2 (F) as functions of temperature. Feed compo-
The performance of Mg/O, Fe/O and Mf/Fe/O catalysts in
2
4
sition (molar fractions): methanol 0.12, nitrogen 0.88. Catalyst: H-mordenite.
phenol methylation is shown in Figs. 6–8. For each catalyst the
following values are reported as a function of the reaction tem-
perature: (i) the conversion of phenol and the selectivity to the
aromatic products, calculated with respect to the phenol con-
verted, and (ii) the yields of the light products (CO2, CH4, CO
and H2), calculated with respect to the inlet methanol.
Bottom: yields of CO (2), CO (Q), CH (×) and H2 (F) as functions of
2
4
temperature. Feed composition (molar fractions): methylformate 0.12, nitrogen
.88. Catalyst: H-mordenite.
0
suggests a change of the prevailing mechanism of methanol
◦
transformation at 350–400 C, with respect to that one occur-
ring at lower temperatures.
3
.5.1. Mg/O (Fig. 6)
The conversion of phenol was low; the main product was
When tests were carried out by feeding methylformate
◦
◦
(Fig. 5, bottom), CO was the main product below 350 C, to-
o-cresol, with small amounts of 2,6-xylenol (at 450 C) and
anisole. p-Cresol formed in very low quantity. The conversion
of methanol was comparable to that achieved under the same
conditions without phenol (Fig. 2, top). However, there were
gether with methanol, both deriving from the acid-catalyzed
decomposition of the reactant. The absence of H confirms
that the methanol formed did not undergo dehydrogenation to
2

430
N. Ballarini et al. / Journal of Catalysis 251 (2007) 423–436
Fig. 7. Top: conversion of phenol (2), selectivity to o-cresol ("), 2,6-xylenol
F), p-cresol (Q) and anisole (×) as functions of temperature. Bottom: yields
(
Fig. 8. Top: conversion of phenol (2), selectivity to o-cresol ("), 2,6-xylenol
(F), p-cresol (Q) and anisole (×) as functions of temperature. Bottom: yields
of CO (2), CO2 (Q), CH4 (×) and H2 (F) as functions of temperature. Feed
composition (molar fractions): methanol 0.108, phenol 0.011, nitrogen 0.881.
Catalyst: Mg/Fe/O.
of CO (2), CO (Q), CH (×) and H (F) as functions of temperature. Feed
2
4
2
composition (molar fractions): methanol 0.108, phenol 0.011, nitrogen 0.881.
Catalyst: Fe/O.
differences in the distribution of the light products, since the
formation of CH4 was lower than that of CO2 in the presence of
phenol, whereas the two compounds formed in almost equimo-
lar amounts in the absence of phenol.
3.5.3. Mg/Fe/O (Fig. 8)
The catalyst was more active than Mg/O and Fe/O in phe-
nol and methanol conversion. The surface area of Mg/Fe/O was
almost twice that of Fe/O (Table 1); this may explain the dif-
ference in reactivity between these two catalysts. By contrast,
the surface area of Mg/O was higher than that of Fe/O, but the
activity of the two samples was comparable. This points out
for an important role of the Fe sites. It is worth noting that
also with Mg/Fe/O a competition effect between phenol and
methanol caused a decrease of methanol conversion to light
products as compared to that achieved in the absence of phe-
nol (Fig. 2, bottom). The distribution of the phenolic products
was also similar to that one obtained with Fe/O. The main prod-
uct at low temperature was o-cresol, but its selectivity decreased
in favor of 2,6-xylenol when the reaction temperature was in-
creased.
3
.5.2. Fe/O (Fig. 7)
The performance of Fe/O in phenol methylation was simi-
lar to that of Mg/O. The prevailing product was o-cresol, with
low formation of anisole and traces of p-cresol. One difference
with respect to Mg/O concerns the transformation of methanol;
in fact, with Fe/O the conversion of methanol was lower, as
demonstrated by the lesser formation of light compounds, with
respect to tests done without phenol (Fig. 2, middle). This indi-
cates that methanol and phenol competed for adsorption on the
same sites; the presence of the aromatic reactant (even though
with a partial pressure which was 1/10th that of methanol) in-
hibited the dehydrogenation of methanol. The presence of a
competitive effect between phenol and methanol indicates a rel-
evant role of the Fe–O Lewis acid–base pair in the interaction
with O and H atoms, respectively, of the hydroxy group of the
two molecules. The relative distribution of the light compounds
also changed with respect to that one obtained in the absence
of phenol; the main effect was the decrease of the selectivity to
CO, with also a slight increase of the selectivity to CH4.
Tests were carried out with the Mg/Fe/O catalyst and with a
methanol/phenol feed ratio equal to 5/1, by keeping the partial
pressure of methanol the same as for tests reported in Fig. 8,
while the phenol partial pressure was increased. At 390 C,
the conversion of phenol decreased from 28% (obtained with
a feed ratio of 10/1), to 15% (feed ratio 5/1); the rate of phe-
◦

N. Ballarini et al. / Journal of Catalysis 251 (2007) 423–436
431
Fig. 9. Conversion of phenol (2), selectivity to o-cresol ("), 2,6-xylenol (F),
p-cresol (Q), anisole (×), 2,4-xylenol (!) and polyalkylated phenols ( ) as
functions of temperature. Feed composition (molar fractions): methanol 0.108,
phenol 0.011, nitrogen 0.881. Catalyst: H-mordenite.
Fig. 10. Conversion of anisole (2), selectivity to o-cresol ("), 2,6-xylenol
(F), p-cresol (Q), phenol (×) and 2,4-xylenol (!) as functions of tempera-
ture. Feed composition (molar fractions): anisole 0.01, nitrogen 0.99. Catalyst:
H-mordenite.
nol transformation remained the same in the two series of tests.
Therefore, the reaction rate order for the partial pressure of phe-
nol was close to zero, possibly indicating a saturation of surface
sites devoted to phenol adsorption and activation. On the other
hand, the conversion of methanol to yield methylated phenols
and light compounds was not affected by the variation of phe-
nol partial pressure.
been investigated by feeding vapors of anisole in the ab-
sence of methanol. The results plotted in Fig. 10 show that
anisole is very reactive at temperatures higher than 300 C.
◦
Therefore, the anisole formed in phenol methylation may ac-
tually be an intermediate in the formation of C-alkylated
compounds, especially for high phenol conversion. In fact,
data in Fig. 9 show that in phenol methylation the selectiv-
◦
ity to anisole was the lowest (3%) at 350 C, in correspon-
3
.6. The methylation of phenol with the acid catalyst
dence of the highest phenol conversion, while at 400 and
◦
4
50 C it was higher than 10%, due to the decrease of phenol
The catalytic performance in phenol methylation of the H-
conversion.
In tests made by feeding anisole (Fig. 10), comparable
mordenite is illustrated in Fig. 9. The catalyst was very active;
◦
at 350 C the conversion of phenol was 80%. A further in-
amounts of phenol and o-cresol were obtained, and minor
amounts of p-cresol (which formation increased with temper-
ature), of 2,4- and 2,6-xylenols and of polyalkylated phenols.
Since methylanisoles were not formed at all, it is possible to
exclude the intermolecular methylation:
crease of temperature led to a decrease of conversion. This
can be attributed to two concomitant phenomena, (i) a pro-
gressive deactivation of the catalyst, due to the formation of
condensed aromatics (coke precursors) by transformation of
methanol [61–66], and (ii) a considerable increase in the par-
allel reactions of methanol decomposition to light compounds.
Indeed, the spent catalyst was coked, and its surface area was
considerably lower than that one of the fresh catalyst (Table 1).
The in situ regeneration of the catalyst by treatment in nitrogen-
diluted air increased the activity. For instance, the conversion of
2 anisole → phenol + methylanisole.
Therefore the formation of o-cresol mainly occurred by in-
tramolecular rearrangement of anisole. The formation of phenol
is not to be attributed to the hydrolysis of anisole, since wa-
ter was not present. However, it is known that alkylarylethers
may decompose at high temperature, to form alkylphenols (or-
tho > para) and phenols [67]. Noteworthy, we detected the
formation of ethylene and of minor amounts of light alkanes
(methane, ethane and propane); this suggests that the transfor-
mation of anisole to phenol indeed occurred through a high-
temperature, radical-like bimolecular mechanism with genera-
tion of two molecules of phenol and one of ethylene.
One further aspect of tests made by feeding anisole is the
absence of any evident deactivation phenomena at high tem-
perature, in contrast with what observed in the case of phenol
methylation. This was due to the absence of methanol in the
reaction feed, reactant for the formation of coke precursors; in
fact, polyalkylated aromatics were not produced at all.
◦
phenol at 390 C after the regeneration treatment was 87% (be-
fore regeneration, 64%); in correspondence, the selectivity to
anisole decreased from 12% to less than 1%. The structural in-
tegrity of the spent catalyst was checked by means of X-ray
diffraction.
The main products at low temperature were o-cresol and
anisole, but also 2,6-xylenol and p-cresol formed in relevant
amounts. The increase of temperature led to a lower selectiv-
ity to all mono-alkylated compounds, in favor of the forma-
tion of xylenols and polyalkylated compounds. However, above
◦
3
50 C the decrease of phenol conversion caused an increase of
the selectivity to o-cresol and anisole.
Anisole itself may act as an alkylating agent, favoring the
formation of cresols and polyalkylates. This possibility has

432
N. Ballarini et al. / Journal of Catalysis 251 (2007) 423–436
mates, with incorporation of the O2 species. The nucleophilic
−
4
. Discussion
attack on the carbonyl moiety occurs with the concomitant ab-
+
2−
4
.1. The transformation of methanol over basic
straction of an H by a second O species. The attack at the
carbonyl produces a dioxymethylene complex, further evolved
to the formate species. The desorption of HCOOH, or its de-
composition, leads to the reduction of the metal ion.
(
dehydrogenating) catalysts
It is worth making a short overview of the properties of metal
oxides in methanol transformation, as inferred from the wide
literature available in this field. Higher alcohols dehydrate to
olefins on acidic catalysts, while they dehydrogenate to alde-
hydes or ketones on catalysts having basic features; these re-
actions have often been used as test reactions to estimate the
surface properties of catalysts [68]. The alcohol typically used
is 2-propanol, due to the stability of the cation that develops by
protonation and elimination of water, and to the easy detection
of the corresponding ketone, i.e. acetone.
Formic acid on alumina may decompose either through a
dehydrative route to H O + CO, or through a dehydrogenative
2
route to H2 + CO2 [78–80], via intermediate surface formate
species [81]. The formation of methylformate from methanol
may occur through different mechanisms, depending on the
nature of the oxide (either easily reducible or non reducible),
the conditions (temperature, pressure, presence of H O), and
2
the surface concentration of formaldehyde, the latter being pri-
marily a function of the dehydrogenative properties of the ox-
ide. Methylformate decomposes to yield either CH + CO or
For what specifically concerns methanol, even in the 60’s
infrared studies have demonstrated that the chemisorption of
methanol on alumina leads to the formation of methoxide in
4
2
CH3OH + CO, or gives back formaldehyde through a reverse
Tischenko reaction [82]. In the presence of water, methylfor-
mate decomposes to formic acid and methanol.
◦
the 35–170 C range; its transformation at higher temperatures
into formate-like surface compound [69,70] is accompanied by
evolution of H2, with the possible intermediate formation of
formaldehyde [71].
In conclusion, several examples exist in literature showing
that methanol undergoes different transformations on metal ox-
ides, depending on redox properties of the cation and on the
Several reviews have examined the nature of the species
that develop by interaction between methanol and redox-type
solid oxides [54,55]. The interaction with cations having Lewis-
type acid properties yields an undissociated CH3OHads, bonded
species [72]. Dissociated methoxy species are preferentially
formed over basic (ionic) oxides (Bi2O3, Fe2O3, NiO, ZnO,
ZrO2). The covalent and Lewis-type acid oxides (WO3, SiO2,
V2O5, Nb2O5, MoO3) and the amphoteric ones (CeO2, TiO2)
produce both undissociated and dissociated terminal methoxy.
Bridging methoxy, the intermediate for the formate species, can
also form, e.g., on ZnO [73].
“
basicity” of the oxygen anion. It is worth underlying that in our
tests the formation of light products (CO, CH , CO and H ) is
4
2
2
indicative of the dehydrogenation of methanol to formaldehyde
and of the further transformation of the latter. Other mecha-
nisms that might lead to the formation of the same compounds,
e.g., a Fischer–Tropsch or methanation leading to the formation
of CH and H O from CO + H , possibly catalyzed by Fe in
4
2
2
Mg/Fe/O and Fe/O catalysts, can be excluded due to the follow-
ing reasons: (i) the same light products also formed with Mg/O,
although in lower amount, and (ii) no metallic Fe was found in
spent samples.
The interaction of methanol with non-reducible metal ox-
ides has been the object of several investigations as well.
MgO is known to catalyze the dehydrogenation of methanol
to formaldehyde [74]. Methanol gives rise to physisorption,
chemisorption or heterolytic dissociation [75] via acid–base
The catalysts investigated possessed different activity in the
dehydrogenation of methanol to formaldehyde and in the con-
secutive transformation of formaldehyde. The preferred path-
way for the transformation of formaldehyde was a function of
the catalyst type. With Mg/O, formaldehyde dimerized to yield
methylformate; however, due to the relatively high temperature
at which methanol dehydrogenated, the ester readily decom-
posed to CO + CH . The overall conversion of methanol over
−
mechanism, with formation of the CH3O species [76]. The
development of the adsorbed formate species occurs through
a Cannizzaro-type reaction with intermolecular disproportion-
ation; a nucleophilic attack of the O2− surface species to the
2
4
Mg/O was not remarkably affected by the presence of phenol;
therefore, the adsorption of the aromatic compound was not fa-
vored over that of methanol.
carbonyl (with development of the formate), is accompanied by
a hydride transfer to a second molecule of adsorbed formalde-
hyde, which is converted to a methoxy species [73,77]. The for-
mate may finally decompose to CO and H2, without any supply
In the case of Fe-containing samples, more active than
Mg/O, the presence of Fe3 in the fresh catalyst may favor
oxidative mechanisms (the oxidehydrogenation of methanol to
H2CO + H2O and the oxidation of formaldehyde to formate).
Formic acid decomposed to either CO + H2O or CO2 + H2 (the
former reaction being preferred at low temperature), or yielded
methylformate. Due to the high reactivity of these samples
(especially of Mg/Fe/O), methylformate was obtained even at
+
2−
of O from the solid. The disproportionation mechanism may
also occur on metal oxides such as ZnO, through intermedi-
ate dioxymethylene species; finally, carbonate and bicarbonate
species develop, with evolution of CO2. Alternatively, the ad-
sorbed methoxy and formate species may yield methylformate;
however, the latter also forms by the direct dimerization of ad-
sorbed formaldehyde (Tischenko reaction).
◦
An overview of the interactions that formaldehyde may de-
velop with metal oxides has been reported in the review of
Barteau [13, and references therein]. Over redox-type oxides
temperature lower than 400 C. The decomposition of methyl-
formate led to the formation of either CH3OH + CO or CH4 +
CO2, the former reaction being preferred at low temperature,
the latter at high temperature.
(
ZnO, CuO, etc.), formaldehyde gives rise to adsorbed for-

N. Ballarini et al. / Journal of Catalysis 251 (2007) 423–436
433
perature is increased the equilibrium is shifted towards the
methoxy species [83]; the latter acts as an electrophylic alky-
lating agent on alkylaromatics [84,85]. In the methanol-to-
olefins process, the first step is the dehydration of methanol
to dimethylether. Two mechanisms have been proposed: (a)
an indirect pathway, in which the adsorbed methanol first re-
acts with methoxy species and then converts into dimethylether
in the presence of another methanol molecule [86]; (b) a di-
rect pathway, in which two methanol molecules react on an
acid site with formation of dimethylether and H2O in one step
[
87]. Surface methoxy species SiO(CH3)Al play a role in the
Fig. 11. Reaction network of methanol decomposition to light products.
formation of dimethylether [88]. The conversion of the equi-
librium mixture of methanol and dimethylether (and water as
well) is dominated by a “hydrocarbon pool” route [89,90],
in which methanol is directly added onto reactive organic
compounds to form aliphatic and aromatic hydrocarbons. The
methoxy species also plays a role in the kinetic “induction pe-
riod,” leading to the reactive hydrocarbon pool. Alternative
“direct” mechanisms have been proposed, in which a carbe-
The interaction of phenol with Mg/Fe/O was much stronger
than with Mg/O, and the presence of phenol decreased the con-
version of methanol, as inferred from the low yields of light
compounds obtained during the reaction of phenol methylation.
Therefore, an important contribution of the Lewis-acid charac-
teristics of Fe3 [35,42] can be hypothesized. With both Fe/O
and Mg/Fe/O, in the presence of phenol there was a decrease of
the selectivity to CO and an increase of that to CH4.
+
+
nium ion CH3 reacts with dimethylether to generate either a
carbonium ion CH –CH –OCH , or a oxonium ylide species.
+
3
3
3
Furthermore, tests made with Mg/Fe/O and variation of the
phenol partial pressure evidenced that an increase of the latter
led to a decrease of phenol conversion, and that the reaction rate
of phenol methylation was substantially independent from this
reaction parameter. These data are consistent with a reaction
mechanism in which the rate-limiting step is the activation of
methanol, and with a Langmuir–Hinshelwood-type adsorption
model in which, despite the large excess of methanol in the feed
stream, phenol interacts strongly with Fe–O pairs and there-
fore competes with methanol for adsorption. This also confirms
that the higher activity of the Fe-containing catalysts (Fe/O and
Mg/Fe/O) as compared to Mg/O was due to the intrinsic dehy-
Other mechanisms include a carbene species: CH , as the re-
action intermediate (see the review [91], for an analysis of the
2
several mechanisms proposed in the literature). The surface-
stabilized ylide or the carbene species [92] are responsible for
the methylation of aliphatic compounds and the formation of
hydrocarbons, both aliphatics and aromatics (polymethylben-
◦
zenes) [93]. At T > 420 C, relevant amounts of methane and
formaldehyde were found, whose formation was attributed to a
reaction between methanol and methoxy groups [94].
In the case of the H-BEA zeolite, the main products in
methanol transformation at high temperature were light aliphat-
ics, both saturated and unsaturated, hexamethylbenzene and
pentamethylbenzene [61], while ZSM-5 gave mostly dimethyl
and trimethylbenzenes [62]. These compounds were further
converted to naphthalene derivatives that were finally respon-
sible for the formation of coke precursors and the deactivation
of the zeolite [63].
3+
drogenating properties of the Fe cation.
Fig. 11 summarizes the different routes of methanol transfor-
mation over metal oxides having basic/dehydrogenating prop-
erties, as inferred from the analysis of the scientific literature.
With our systems and under the conditions employed, the main
reactions involved were different in function of the catalyst
characteristics. Mg/O was poorly active in methanol dehydro-
genation to formaldehyde; the latter then dimerized to methyl-
formate, which finally decomposed to light products. With Fe-
containing catalysts, an additional contribution to methylfor-
mate formation derived from the oxidation of formaldehyde
to formate, at least in the fresh, oxidized catalyst. The inter-
mediate compounds formed by methanol transformation (i.e.,
formaldehyde and formic acid) were identified in trace amounts
only under non-steady-state conditions. This means that under
the conditions used for catalytic tests these intermediates were
very reactive, and were quickly transformed into the light com-
pounds.
It is worth mentioning that hexamethylbenzene also formed
by reaction between phenol and methanol over γ -Al O at
2
3
◦
T > 400 C [64,65], while in benzene and toluene methylation
the same compounds wee not formed [66]. Lower temperatures
led to the formation of anisole and cresols. Therefore, the phe-
nolic group was proposed to play an important role in the for-
mation of hexamethylbenzene through oxygen-containing in-
termediates.
With our H-mordenite catalyst, the formation of polyalkyl-
benzenes occurred preferentially in the temperature range be-
◦
tween 250 and 350 C, while at higher temperature the de-
composition to light compounds became the main reaction of
methanol transformation (Fig. 5). A possible role of coke, de-
posited on zeolite surface, in the dehydrogenation of methanol
and decomposition to light compounds can not be excluded.
A change of the main reaction occurring on methanol also
corresponds to a variation in the distribution of products ob-
tained in phenol methylation. In the low-temperature range, the
prevailing compounds were anisole and cresols. The increase
4
.2. The transformation of methanol over zeolites
Over zeolites, the adsorption of methanol can generate
+
framework-bonded methoxonium CH3OH2 and methoxy spe-
cies that can co-exist at low temperature, but when the tem-

434
N. Ballarini et al. / Journal of Catalysis 251 (2007) 423–436
of temperature and of phenol conversion caused an increase
of the selectivity to xylenols and polyalkylated phenols. At
temperature higher than 350 C the selectivity to polyalkylated
phenol decreased and that to anisole increased, because of the
lower phenol conversion. Moreover, a relevant increase in the
formation of o-cresol was observed, while the selectivity to
i.e., methylformate. Due to the presence of the carbonyl moiety,
these compounds are more electrophylic than methanol, and
hence are able to attack the aromatic ring even in the absence
of an acid-catalyzed activation of the molecule. Moreover, the
deprotonation of phenol and the generation of the phenolate
species further activate the aromatic substrate.
◦
2
,6-xylenol remained constant. The change in the type of mech-
This hypothesis is also supported by the comparison of the
reactants conversion obtained with the three basic catalysts. In
fact, Mg/O was clearly the least active catalyst in both methanol
and phenol conversion. In Fe/O and Mg/Fe/O, the high activ-
ity has to be attributed to the dehydrogenating properties of
Fe ; also, an oxidative mechanism may contribute to methanol
activation, at least in the fresh catalyst Therefore, over Fe-
containing catalysts the dehydrogenation of methanol repre-
sents the key-step for the generation of the active compound
for the electrophylic attack to the aromatic ring.
anism involved in methanol activation and the generation of a
different methylating species may contribute to a modification
of the products obtained in phenol methylation.
3+
4
.3. The role of methanol transformation in the acid and
basic-catalyzed methylation of phenol
The implications that methanol transformations may have
on catalytic performance in phenol methylation are matter of
discussion. Results indicate that differences between acid and
basic catalysts concern both the nature of aromatic products ob-
tained and the type of transformations occurring on methanol.
Specifically, we can distinguish three different behaviors:
The implications of methanol dehydrogenation on the mech-
anism of phenol methylation, in terms of regio- and chemo-
selectivity, is currently the object of our investigation. Several
questions are still open, and specifically:
(
1) Mg/O dehydrogenates methanol to formaldehyde, but pos-
sesses low activity in this reaction; in phenol methylation,
it gives products of both O-methylation and C-alkylation,
the latter being largely preferred, especially at high tem-
perature. Amongst C-alkylated products, o-cresol is greatly
favored over p-cresol (yielded in very low amount) and 2,6-
xylenols.
(1) If the dehydrogenation of methanol is the pre-requisite for
the generation of the active compound in phenol methyla-
tion over metal oxides having basic/dehydrogenating prop-
erties, which is the true methylating species amongst those
formed by methanol transformation, i.e., formaldehyde and
methylformate? Are these compounds more reactive than
methanol in phenol methylation?
(
2) Fe/O and Mg/Fe/O are active in methanol dehydrogenation,
even at relatively low temperature. The distribution of prod-
ucts in phenol methylation is similar to that obtained with
Mg/O, except for the lower formation of anisole. These
catalysts, especially Mg/Fe/O, are also more active than
Mg/O in phenol methylation. In this case a competition
between the reactions of methanol dehydrogenation and
phenol methylation occurs, likely due to the strong adsorp-
(2) Which is the mechanism that leads to the formation of the
methylated phenol, when the first step of the reaction is not
a direct methylation by methanol?
(3) Is the generation of a reactive species different from
methanol the reason for the high chemo- and regio-
selectivity experimentally observed in phenol methylation
over metal oxides having basic/dehydrogenating proper-
ties? May the presence of metal cations having pronounced
Lewis-acid characteristics also affect the selectivity?
tion of phenol on Fe3 sites having Lewis-type properties.
3) H-mordenite gives an acid-catalyzed activation of methanol
and a wide distribution of products of phenol methylation,
with relevant formation of both anisole and C-alkylated
compounds. The prevailing reaction of methanol transfor-
+
(
5. Conclusions
This work compares the performance in phenol methyla-
tion of three catalysts having basic/dehydrogenation proper-
ties (Mg/O, Mg/Fe/O and Fe/O) with that of an acid catalyst,
a H-mordenite. The main difference between the two classes of
catalysts concerned the transformations occurring on methanol.
With basic catalysts the key-step was the dehydrogenation
of methanol to formaldehyde, with generation of the active
species for the methylation of phenol. In this case, the prevail-
ing compounds obtained by reaction with phenol were o-cresol
and 2,6-xylenol. Furthermore, formaldehyde underwent paral-
lel transformations to methylformate and light products, i.e.,
CO, CO2, CH4 and H2. The basic/dehydrogenating property of
these catalysts played its main role in the activation of methanol
and in the generation of an active species more electrophylic
than methanol itself.
◦
mation at temperature lower than 350 C is the formation of
polymethylated benzenes, precursors of coke formation. In
the high-temperature range, however, the prevailing mech-
anism of methanol activation is dehydrogenation; corre-
spondingly, the distribution of products obtained in phenol
methylation is slightly modified, with the prevailing forma-
tion of o-cresol over anisole.
With the three basic/dehydrogenating catalysts, the conver-
sion of phenol began exactly when the conversion of methanol
to light compounds also started. Therefore, these data support
the hypothesis that the dehydrogenation of methanol is the nec-
essary step for the generation of the active compound for phenol
methylation. In other terms, it is possible that the true “methy-
lating” agent is not methanol, but either formaldehyde or some
compound formed by transformation of formaldehyde itself,
The acid-type activation of methanol led to the develop-
ment of an electrophylic species and to the formation of anisole

N. Ballarini et al. / Journal of Catalysis 251 (2007) 423–436
435
and cresols. Methanol also underwent the parallel formation of
[34] S. Sato, K. Koizumi, F. Nozaki, J. Catal. 178 (1998) 264.
35] S. Velu, C.S. Swamy, Appl. Catal. A 162 (1997) 81.
[36] S. Velu, C.S. Swamy, Appl. Catal. A 145 (1996) 225.
[
alkylbenzenes, precursors of coke formation. At temperature
◦
higher than 350 C methanol mainly decomposed to light prod-
[37] A.H. Padmasri, A. Venugopal, V. Durgakumari, K.S. Rama Rao, P. Kanta
Rao, J. Mol. Catal. A 188 (2002) 255.
ucts; under these conditions, the preferred products of phenol
methylation were o-cresol and 2,6-xylenol.
It is proposed that the different nature of the methylating
species generated by methanol activation may be the reason for
the different type of products obtained in phenol methylation
with acid and basic catalysts.
[38] S. Sato, R. Takahashi, T. Sodesawa, K. Matsumoto, Y. Kamimura, J.
Catal. 184 (1999) 180.
[39] S. Velu, S. Sivasanker, Res. Chem. Int. 24 (6) (1998) 657.
[
[
[
40] M. Bolognini, F. Cavani, D. Scagliarini, C. Flego, C. Perego, M. Saba,
Catal. Today 75 (2002) 103.
41] F. Cavani, C. Felloni, D. Scagliarini, A. Tubertini, C. Flego, C. Perego,
Stud. Surf. Sci. Catal. 143 (2002) 953.
42] M. Bolognini, F. Cavani, C. Felloni, D. Scagliarini, C. Flego, C. Perego,
in: D.G. Morrell (Ed.), Catalysis of Organic Reactions, Dekker, New York,
2002, p. 115.
Acknowledgments
INSTM is acknowledged for the Ph.D. grant of Luca Maselli
and Sauro Passeri. Eni SpA is acknowledged for the Ph.D. grant
of Diana Scagliarini. Süd-Chemie is acknowledged for provid-
ing the sample of H-mordenite.
[
[
[
43] M. Bolognini, F. Cavani, D. Scagliarini, C. Flego, C. Perego, M. Saba,
Microporous Mesoporous Mater. 66 (2003) 77.
44] F. Cavani, L. Maselli, D. Scagliarini, C. Flego, C. Perego, Stud. Surf. Sci.
Catal. 155 (2005) 167.
45] US Patent 3,855,318 (1974), assigned to Asahi Chem.
References
[46] US Patent 4,208,857 (1977), assigned to Mitsui Toatsu Chem.
47] US Patent 3,446,856 (1964), assigned to General Electric.
[
[
1] H. Fiege, in: Ullmann’s Encyclopedia of Industrial Chemistry, vol. A8,
Wiley–VCH, Weinheim, 2002, p. 25.
[48] US Patent 4,933,509 (1989), assigned to General Electric.
[49] US Patent 4,041,085 (1975), assigned to General Electric.
[50] J.S. Valente, F. Figueras, M. Gravelle, P. Kumbhar, J. Lopez, J.-P. Besse,
J. Catal. 189 (2000) 370.
[2] M. Eissen, J.O. Metzger, E. Schmidt, U. Schneidewind, Angew. Chem.
Int. Ed. 41 (2002) 414.
[
[
3] R.A. Sheldon, R.S. Downing, Appl. Catal. A 189 (1999) 163.
4] P.T. Anastas, M.M. Kirchoff, T.C. Williamson, Appl. Catal. A 221 (2001)
[51] K.-T. Li, I. Wang, K.-R. Chang, Ind. Eng. Chem. Res. 32 (1993) 1007.
[52] J. Kaspi, G.A. Olah, J. Org. Chem. 43 (16) (1978) 3142.
[53] P.D. Chantal, S. Kaliaguine, J.L. Grandmaison, Appl. Catal. 18 (1985)
133.
[54] G. Busca, Catal. Today 27 (1996) 457.
[55] J.C. Lavalley, Catal. Today 27 (1996) 377.
3.
[
[
[
[
5] H. Hattori, Chem. Rev. 95 (1995) 537.
6] Y. Ono, T. Baba, Catal. Today 38 (1997) 321.
7] K. Tanabe, W.F. Hölderich, Appl. Catal. A 181 (1999) 399.
8] R.F. Parton, J.M. Jacobs, D.R. Huybrechts, P.A. Jacobs, Stud. Surf. Sci.
Catal. 46 (1989) 163.
[56] H. Grabowska, R. Klimkiewicz, W. Tylus, P.J. Godowski, Appl. Catal.
A 240 (2003) 111.
[9] R.J. Davis, J. Catal. 216 (2003) 396.
[57] US Patent 6,897,175 (2005), assigned to General Electric.
[58] WO Patent 2004/014832A2, assigned to Vanetta SpA.
[59] H. Grabowska, W. Mista, L. Syper, J. Wrzyszcz, M. Zawadzki, Appl.
Catal. A 144 (1996) L1.
[
[
[
[
[
10] Y. Ono, J. Catal. 216 (2003) 406.
11] G. Busca, Phys. Chem. Chem. Phys. 1 (1999) 723.
12] A. Cimino, F.S. Stone, Adv. Catal. 47 (2002) 141.
13] M. Barteau, Chem. Rev. 96 (1996) 1413.
[
60] H. Grabowska, W. Mista, L. Syper, J. Wrzyszcz, M. Zawadzki, J.
Catal. 160 (1996) 134.
14] K. Tanabe, T. Nishizaki, in: G.C. Bond, et al. (Eds.), Proceedings of the
6th ICC, The Chemical Society, London, 1977, p. 863.
[61] M. Bregolato, V. Bolis, C. Busco, P. Ugliengo, S. Bordiga, F. Cavani, N.
Ballarini, L. Maselli, S. Passeri, I. Rossetti, L. Forni, J. Catal. 245 (2007)
[
[
[
[
[
[
[
[
[
[
15] K. Tanabe, Stud. Surf. Sci. Catal. 20 (1985) 1.
16] S. Sato, K. Koizumi, F. Nozaki, Appl. Catal. A 133 (1995) L7.
17] R. Bal, S. Mayadevi, S. Sivasanker, Org. Process Res. Dev. 7 (2003) 17.
18] R. Bal, B.B. Tope, S. Sivasanker, J. Mol. Catal. A 181 (2002) 161.
19] V. Durgakumari, S. Narayanan, J. Mol. Catal. 65 (1991) 385.
20] K. Tanabe, K. Shimazu, H. Hattori, K. Shimazu, J. Catal. 57 (1979) 35.
21] A.S. Reddy, C.S. Gopinath, S. Chilukuri, J. Catal. 243 (2006) 278.
22] H. Grabowska, W. Kaczmarczyk, J. Wrzyszcz, Appl. Catal. 47 (1989) 351.
23] S. Velu, C.S. Swamy, Appl. Catal. A 145 (1996) 141.
2
85.
[
[
[
62] Ø. Mikkelsen, S. Kolboe, Microporous Mesoporous Mater. 29 (1999) 173.
63] M. Bjørgen, U. Olsbye, S. Kolboe, J. Catal. 215 (2003) 30.
64] N.M. Cullinane, S.J. Chard, C.W.C. Dawkins, in: N. Rabjohn (Ed.), Or-
ganic Syntheses, coll. vol. IV, Wiley and Sons, New York, 1963, p. 520.
65] P.S. Landis, W.O. Haag, J. Org. Chem. 28 (1963) 585.
66] L.H. Klemm, J. Shabtai, D.R. Taylor, J. Org. Chem. 33 (1968) 1480.
67] H. Hard, H.S. Eleuterio, J. Am. Chem. Soc. (1954) 519.
68] Y. Iwasawa, Acc. Chem. Res. 30 (1997) 103.
69] R.C. Greenler, J. Chem. Phys. 37 (1962) 2094.
70] R.O. Kagel, J. Phys. Chem. 71 (1967) 844.
71] J. Shabtai, L.H. Klemm, D.R. Taylor, J. Org. Chem. 35 (1970) 1075.
72] L.J. Burcham, L.E. Briand, I.E. Wachs, Langmuir 17 (2001) 6164 and
[
[
[
[
[
[
[
[
24] V.V. Rao, K.V.R. Chary, V. Durgakumari, S. Narayanan, Appl. Catal. 61
(1990) 89.
[25] A.R. Gandhe, J.B. Fernandes, J. Mol. Catal. A 226 (2005) 171.
[26] A.R. Gandhe, J.B. Fernandes, S. Varma, N.M. Gupta, J. Mol. Catal. A 238
(2005) 63.
[
[
[
[
27] H. Grabowska, W. Mista, J. Trawczynski, J. Wrzyszcz, M. Zawadzki,
Appl. Catal. A 220 (2001) 207.
6
175.
73] N.D. Lazo, D.K. Murray, M.L. Kieke, J.H. Haw, J. Am. Chem. Soc. 114
1992) 8552.
74] P. Mars, J.J. Scholten, P. Zwietering, Adv. Catal. 14 (1963) 35.
[
28] R. Klimkiewicz, H. Grabowska, H. Teterycz, Appl. Catal. A 246 (2003)
(
125.
[
29] T. Mathew, N.R. Shiju, K. Sreekumar, B.S. Rao, C.S. Gopinath, J.
Catal. 210 (2002) 405.
30] T. Mathew, S. Shylesh, B.M. Devassy, C.V.V. Satyanarayana, B.S. Rao,
C.S. Gopinath, Appl. Catal. A 261 (2004) 292.
[75] C. Di Valentin, A. Del Vitto, G. Pacchioni, S. Abbet, A.S. Wörz, K. Judai,
U. Heiz, J. Phys. Chem. B 106 (2002) 11961.
[76] M. Ziolek, J. Kujawa, O. Saur, J.C. Lavalley, J. Phys. Chem. 97 (1992)
9761.
[77] X.D. Peng, M.A. Barteau, Langmuir 5 (1989) 1051.
[78] G. Patermarakis, Appl. Catal. A 252 (2003) 231.
[79] T. Yumura, T. Amenomori, Y. Kagawa, K. Yoshizawa, J. Phys. Chem.
A 106 (2002) 621.
[
[
31] T. Mathew, B.S. Rao, C.S. Gopinath, J. Catal. 222 (2004) 107.
32] T. Mathew, N.R. Shiju, B.B. Tope, S.G. Hegde, B.S. Rao, C.S. Gopinath,
Phys. Chem. Chem. Phys. 4 (2002) 4260.
[33] T. Mathew, M. Vijayaraj, S. Pai, B.B. Tope, S.G. Hegde, B.S. Rao, C.S.
Gopinath, J. Catal. 227 (2004) 175.

436
N. Ballarini et al. / Journal of Catalysis 251 (2007) 423–436
[
[
[
[
[
[
[
[
[
80] K.S. Kim, M.A. Barteau, Langmuir 4 (1988) 945.
81] D.E. Fein, I.E. Wachs, J. Catal. 210 (2002) 241.
[89] I.M. Dahl, S. Kolboe, J. Catal. 149 (1994) 458.
[90] B. Arstad, S. Kolboe, J. Am. Chem. Soc. 123 (2001) 8137.
[91] J.F. Haw, W. Song, D.M. Marcus, J.B. Nicholas, Acc. Chem. Res. 26
(2003) 317.
[92] P.E. Sinclair, C.R.A. Catlow, J. Phys. Chem. B 101 (1997) 295.
[93] W. Wang, A. Buchholz, M. Seiler, M. Hunger, J. Am. Chem. Soc. 125
(2003) 15260.
82] J.S. Francisco, J. Am. Chem. Soc. 125 (2003) 10475.
83] J. Rakoczy, T. Romotowski, Zeolites 13 (1993) 256.
84] G. Mirth, J. Lercher, J. Catal. 132 (1991) 244.
85] A. Corma, G. Sastre, P.M. Viruela, J. Mol. Catal. A 100 (1995) 75.
86] T.R. Forester, R.F. Howe, J. Am. Chem. Soc. 109 (1987) 5076.
87] J. Bandiera, C. Naccache, Appl. Catal. 69 (1991) 139.
88] W. Wang, M. Seiler, M. Hunger, J. Phys. Chem. B 105 (2001) 12553.
[94] O. Dewaele, V.L. Geers, G.F. Froment, G.B. Marin, Chem. Eng. Sci. 54
(1999) 4385.