Catalytic methylation of phenol on MgO - Surface chemistry and mechanism

Article abstract of DOI:10.1016/j.jcat.2009.11.019

Reactions of phenol and methanol catalyzed by MgO have been explored by kinetic measurements and in situ IR spectroscopy combined with computational studies of sorbed molecules. On MgO, methanol partly transforms to formaldehyde above 250 °C. Adsorbed phenol forms phenolate species, and the energetically preferred mode of adsorption leads to an almost orthogonal orientation of the aromatic ring with respect to the catalyst surface. All molecules involved adsorb preferably at the corner sites of MgO (three-co-ordinated Mg atoms). The main reaction products are anisole and o-cresol, the latter dominating above 300 °C. At very low conversions, salicylic aldehyde is observed as primary reaction product, being rapidly transformed to o-cresol. It is only observed during the initial accumulation of adsorbed species on the catalyst surface, but not under steady-state conditions on a fully covered catalyst surface. Therefore, o-cresol formation starts with the reaction between phenol and formaldehyde to salicylic alcohol, which in turn is rapidly transformed to salicylic aldehyde and subsequently to o-cresol. Salicylic aldehyde may also form via the bimolecular disproportionation of salicylic alcohol to o-cresol and aldehyde. The parallel reaction to o-cresol, not involving the formation of salicylic aldehyde as intermediate, proceeds via the reduction of salicylic alcohol to o-cresol by formaldehyde. The identified mechanism may open new synthetic approaches for the production of functionalized phenol derivatives and, even more importantly, for the defunctionalization of substituted phenols potentially available at large scale from deconstructed lignin.

Full text of DOI:10.1016/j.jcat.2009.11.019

Journal of Catalysis 269 (2010) 340–350
Contents lists available at ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
Catalytic methylation of phenol on MgO – Surface chemistry and mechanism
Fabrizio Cavani a,b,*, Luca Maselli a,b, Sauro Passeri a,b, Johannes A. Lercher c,
*
a
Dipartimento di Chimica Industriale e dei Materiali, ALMA MATER STUDIORUM Università di Bologna, Viale Risorgimento 4, 40136 Bologna, Italy
b
INSTM, Research Unit of Bologna, Italy
c
Technische Universität München, Department Chemie, Lichtenbergstr. 4, D-85748 Garching, Germany
a r t i c l e i n f o
a b s t r a c t
Article history:
Reactions of phenol and methanol catalyzed by MgO have been explored by kinetic measurements and
in situ IR spectroscopy combined with computational studies of sorbed molecules. On MgO, methanol
partly transforms to formaldehyde above 250 °C. Adsorbed phenol forms phenolate species, and the ener-
getically preferred mode of adsorption leads to an almost orthogonal orientation of the aromatic ring
with respect to the catalyst surface. All molecules involved adsorb preferably at the corner sites of
MgO (three-co-ordinated Mg atoms). The main reaction products are anisole and o-cresol, the latter dom-
inating above 300 °C. At very low conversions, salicylic aldehyde is observed as primary reaction product,
being rapidly transformed to o-cresol. It is only observed during the initial accumulation of adsorbed spe-
cies on the catalyst surface, but not under steady-state conditions on a fully covered catalyst surface.
Therefore, o-cresol formation starts with the reaction between phenol and formaldehyde to salicylic alco-
hol, which in turn is rapidly transformed to salicylic aldehyde and subsequently to o-cresol. Salicylic alde-
hyde may also form via the bimolecular disproportionation of salicylic alcohol to o-cresol and aldehyde.
The parallel reaction to o-cresol, not involving the formation of salicylic aldehyde as intermediate, pro-
ceeds via the reduction of salicylic alcohol to o-cresol by formaldehyde. The identified mechanism may
open new synthetic approaches for the production of functionalized phenol derivatives and, even more
importantly, for the defunctionalization of substituted phenols potentially available at large scale from
deconstructed lignin.
Received 9 September 2009
Revised 16 November 2009
Accepted 18 November 2009
Available online 29 December 2009
Keywords:
Basic catalysis
Phenol methylation
Reaction mechanism
Electrophilic substitution
Magnesium oxide
Ó 2009 Elsevier Inc. All rights reserved.
1
. Introduction
The methylation of phenol and phenol derivatives has a high
chemo-selectivity with only minor amounts reacting via
oxofunctionalization.
The major problem of the industrial process of phenol methyl-
ation is the decomposition of methanol and, consequently, a large
excess of methanol has to be fed usually in order to reach an
acceptable per-pass conversion of phenol [34]. In this context it
is surprising that the potential role of methanol surface chemistry
and eventual decomposition during the alkylation of phenols is
hardly addressed [12,24,35]. It should be noted here that usually
the alkylation of the aromatic ring with methanol is catalyzed by
acids [36,37] and that bases catalyze the alkylation of side chains
substituted to the aromatic ring (e.g., side-chain alkylation of tolu-
ene). For the latter reaction, the dehydrogenation of methanol to
formaldehyde was shown to be a prerequisite [38–40].
Acid catalysis transforms methanol into an electrophilic species
(alkoxy group or a methyl carbenium ion) able to alkylate both the
methanol oxygen and the carbon of the aromatic ring [41,42].
Bases, in contrast, catalyze dehydrogenation of methanol to form-
aldehyde that in turn primarily alkylates the carbon atoms of the
aromatic ring [41,42]. The regioselective preference of the alkyl-
ation is related to the negative charge at the carbon atom of the
aromatic ring. In the case of phenol it is the ortho and para posi-
tions relative to the O–H group. In competing reaction pathways,
industrial relevance. 2,6-Xylenol is, for example, 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 paints, 2,5-dimethylphenol is the intermediate for
the synthesis of dyes, antiseptics, and antioxidants, and 2,3,6-tri-
methylphenol is the starting compound for the synthesis of vita-
min E. The products of ortho-methylation of phenol or anisole
and of diphenols, such as guaiacol, are intermediates in the produc-
tion of skin-protecting agents and food additives.
The ring methylation of phenol to o-cresol is industrially carried
out with methanol as alkylating agent over basic catalysts [1–6].
Typical catalysts are (supported) alkali and alkaline-earth metal
oxides [7–10], (mixed) transition metal oxides [11–23], and
combination of both [24–33]. These catalysts show very high reg-
ioselectivity in the methylation of the aromatic ring as well as high
*
Corresponding authors. Fax: +39 051 209 3680.
E-mail addresses: fabrizio.cavani@unibo.it (F. Cavani), johannes.lercher@ch.
tum.de (J.A. Lercher).
0
021-9517/$ - see front matter Ó 2009 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2009.11.019

F. Cavani et al. / Journal of Catalysis 269 (2010) 340–350
341
also methyl formate decomposition products such as CO, CO
2
, CH
4
,
acetone, and by injecting the sample in a GC4300 Carlo Erba gas
chromatograph equipped with a TCD and a Carbosieve SII column.
The GC oven temperature was programed from 55 to 220 °C with a
heating rate of 10 °C/min.
and H are formed from formaldehyde. These first experiments
2
indicate that the main role of the catalyst is the formation of form-
aldehyde, which preferentially occurs with strongly basic catalysts.
The further catalytic chemistry seems to depend then on the avail-
ability of formaldehyde under reaction conditions. If this surface
chemistry is indeed prevailing, the base-catalyzed route to alkylate
aromatic rings would be a highly robust chemo-selective alkyl-
ation in the presence of strongly nucleophilic groups.
In the present work, we explore the catalysis with MgO as typ-
ical (albeit not most active) catalyst with the aim of establishing
the detailed mechanism of catalysis alkylation of aromatic com-
pounds with nucleophilic functionalization. Phenol is used as sub-
strate for conversion, as its simplicity offers a better route for
probing the surface chemical species by a combination of in situ
vibrational spectroscopy and theoretical calculations. The mea-
surements and calculations allow us to postulate and test the ki-
netic sequence of elementary steps and to lead to clear design
parameters for the selective alkylation of functionalized aromatic
compounds.
Yields are expressed as follows:
out
product
n_
Yield to products of phenol methylation ¼
;
in
phenol
n_
where ‘‘product” stands for o-cresol, 2,6-xylenol, anisole, and sali-
cylic aldehyde.
Conversion is expressed as follows:
in
out
n_
ꢀ n_
phenol
phenol
Conversion of phenol :
:
in
n_
phenol
Selectivity to a compound is expressed as the ratio between the
corresponding yield and the reactant conversion.
The computational study was carried out on a model Mg12O
12
cluster. The cluster geometry was taken from Ref. [43] where the
Mg–O distance is 2.10850 Å. Density functional theory (DFT) calcu-
lation was performed using the B3LYP hybrid density functional,
and the 6-31G(d,p) basis set was used. The geometry of all ad-
sorbed molecules was fully optimized, but the substrate cluster
was not relaxed in the optimization process in agreement with
an experimental study that demonstrated a poor relaxation in this
oxide [44]. In order to improve the visual comparison between
experimental and calculated IR spectra, wavenumbers have been
scaled by a factor of 0.9613 [45]. The adsorption energies were
computed as the difference between the energy of MgO/molecule
system and the sum of the energies of the separated fragments,
2
. Experimental
The employed catalysts were a commercial MgO, supplied by
Acros Organics, and a self-synthesized MgO, having specific surface
2
areas of 12 and 68 m /g, respectively [42].
Catalytic tests were carried out by vaporization of a methanol/
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, purity 99%)
E
ads = E(MgO–MOL) ꢀ EMgO ꢀ EMOL. Correction for the zero point con-
2
in a N stream (gas flow 20 N ml/min). The liquid mixture was
tribution and, when needed, basis set superimposition error (BSSE),
by the counterpoise method, were applied. All calculations were
carried out using Gaussian 03 program package [46]. Optimized
structures are visualized using GaussView program [47]. Calcula-
tions have been carried out on IBM SP5/512 supercomputer at
CINECA.
fed by means of a syringe pump (Precidor Type 5003, Infors
HT). The composition of the feed gas was the following (molar
fractions): methanol 0.108, phenol 0.011, and nitrogen 0.881.
The residence time was 2.7 s (in experiments varying the tem-
perature), or was varied in order to obtain very low phenol con-
versions (<0.5%), at 300 °C. In the latter case, experiments were
carried out at contact time lower than 0.2 s, achieved by varying
the liquid stream from 0.037 to 0.003 ml/min, and correspond-
The IR spectra after adsorption of reactant and potential prod-
uct molecules (phenol, methanol, o-cresol, salicylic alcohol, and
salicylic aldehyde) were recorded using a Perkin Elmer 1750 FT-
IR Spectrometer equipped with a DTGS detector at a resolution of
2
ingly the N stream in order to keep constant the reagent/nitro-
gen molar ratio; each test was carried out by loading 0.2 g of
fresh catalyst. Total pressure was one bar. In tests made by feed-
ing salicylic alcohol (Sigma Aldrich, purity 99.9%) and water, the
composition of the feed gas was (molar fractions): salicylic alco-
hol 0.004, water 0.437, and balance nitrogen. In tests made by
feeding salicylic alcohol or salicylic aldehyde in formalin
ꢀ
1
2
cm . MgO samples were pressed into self-supported wafer
and activated in situ in the IR cell at 450 °C under vacuum
ꢀ7
(
<10 mbar) for 2 h. Before measuring the IR spectra, the activated
sample was contacted with 2–4 mbar of the adsorbent either at
50 °C, or at increasing temperatures from ambient to 550 °C. For
2
recording spectra, 25 scans were co-added. The IR spectra reported
are displayed as difference absorbance spectra of the sample with
and without the absorbed molecule.
(
20.6 wt.% formaldehyde in water, containing 0.4% methanol),
the composition of the feed gas was (molar fractions): salicylic
alcohol 0.005 or salicylic aldehyde 0.0005, formaldehyde 0.054,
water 0.354, methanol 0.002, and balance nitrogen.
The in situ IR spectra were recorded using a Bruker IFS 88 FT-IR
spectrometer equipped with an in situ flow cell, which was de-
scribed previously [48,49]. For recording spectra, 60 scans
The gas/vapor stream was fed into a tubular glass reactor
3
(
length 300 mm, inner diameter 19 mm) containing either 1 cm
ꢀ1
(
1 scan/s) were co-added at a resolution of 4 cm . Either methanol
of catalyst (catalyst weight 0.85 g, experiments with varying tem-
perature) or variable amounts of catalyst (experiments with vary-
ing residence time) shaped in 250–500 lm particles. Catalyst
particles were prepared by pressing the calcined powder to obtain
pellets that were then broken into smaller granules. During cata-
lytic experiments, the reactor exit stream was condensed in
or a mixture of methanol and phenol (molar ratio 10/1) was vapor-
ized in a He stream (20 mol% of organic components in the He
stream) and fed into the cell. Spectra are reported after subtraction
of the spectrum of MgO.
2
5 ml of HPLC-grade acetone for 1 h maintained at 6 °C. Products
3. Results
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 oven temperature was programed from 50 to
3.1. Catalytic phenol methylation
2
(
50 °C, with a heating rate of 10 °C/min. Non-condensable gases
CO, CO , H , and CH ) were analyzed by sampling the gaseous
stream with a syringe at the reactor exit, before condensation in
The catalytic gas-phase methylation of phenol with methanol
over low-surface-area MgO yields o-cresol and 2,6-xylenol as the
principal reaction products (see Fig. S1 of Supplementary material)
2
2
4

3
42
F. Cavani et al. / Journal of Catalysis 269 (2010) 340–350
[
41,42]. At low temperatures, i.e., below 300 °C, also anisole
different sites of a (MgO)12 cluster representing MgO surface
sites. In order to account for the differences in the co-ordination
of Mg and O, the sites for adsorption considered are those corre-
sponding to the corner, step (edge), and terrace positions. Table 1
compiles the main molecular properties in the adsorbed state,
i.e., adsorption energy, interatomic distances, and the COH angle
at equilibrium. For comparison, the reference values for the cor-
responding isolated molecules are also given.
formed (selectivity less than 25%), but its selectivity decreased
and disappeared above 350 °C. The variation in yield at 300 °C as
a function of the conversion of phenol is shown in Fig. 1. These
tests were carried out by varying the gas residence time, under
conditions aimed at obtaining very low phenol conversion (less
than 0.5%), in order to detect the formation of the truly primary
reaction products.
The primary products were salicylic aldehyde and anisole; how-
ever, also a direct contribution for o-cresol cannot be excluded
indicating that three parallel reaction pathways exist. With
increasing conversion, salicylic aldehyde was reduced to o-cresol.
Thus, o-cresol is produced via a direct and an indirect pathway
via formation of salicylic aldehyde. At phenol conversion above
The adsorption energies on the terrace site are in general weak,
but are higher for step sites and the highest on corner sites, i.e., the
values increase with decreasing co-ordination of the atom. With
increasing strength of interaction the phenolic O–H bond length
in the molecule increases causing a more pronounced distortion
in the molecular geometry. For adsorption on terrace sites, the
O–H group interacts associatively. Upon adsorption on co-ordin-
atively unsaturated sites the O–H group dissociates. It should be
noted that adsorption of phenol on MgO is reported to lead to phe-
nolate species and surface O–H groups [50]. The energetically most
favored adsorption configuration is that one in which the adsorbed
molecule adopts an orthogonal orientation with respect to the
MgO surface. In agreement with the literature [7,9], we conclude
that this orientation is caused by the repulsion of the aromatic ring
by the electron-rich surface oxygen.
0
.05%, anisole and o-cresol were the only reaction products. The
two products do not interconvert. It is known from literature that
in the presence of acid catalysts anisole may be an alkylating agent
[
41]. With the present MgO catalysts, however, anisole did not con-
vert. At 450 °C, a small amount of phenol was formed from anisole
decomposition because methylanisole was not observed. Thus,
over MgO anisole is a stable compound.
Salicylic aldehyde is likely formed by the reaction between phe-
nol and formaldehyde, the latter having formed by methanol dehy-
drogenation [41,42]. Salicylic aldehyde is
a
very reactive
The IR spectra for sorptive molecules adsorbed over the differ-
ent defective sites of MgO and for the corresponding gas-phase
molecule were calculated. In the case of phenol (Fig. S2 of Supple-
mentary material) the spectra of the adsorbed molecules are sim-
ilar, the only relevant difference being the frequency of C–O
intermediate that can be recovered only under very unusual reac-
tion conditions, i.e., low temperatures as well as extremely low con-
tact times and phenol conversions (Fig. 1). At low temperatures,
however, the dehydrogenation of methanol to formaldehyde is less
favored [41]. The question arises, therefore, whether salicylic alde-
hyde is the reaction intermediate in the formation of o-cresol under
reaction conditions leading to high phenol and methanol conver-
sions. It should be noted that the reaction between phenol and
formaldehyde should lead to the formation of salicylic alcohol as
a first intermediate and not to the formation of salicylic aldehyde.
In order to explore the proposed reaction pathway theoretical cal-
culations and IR spectroscopic measurements were combined to
understand surface structures and potential intermediates.
ꢀ1
stretching band at 1300 cm when phenol is adsorbed in the cor-
ner position, but at considerably lower values for the step
ꢀ1
ꢀ1
(1237 cm ) and the terrace (1241 cm ) positions, pointing out
for differences in the C–O bond strength. The same is observed
for o-cresol.
Experimental IR spectra were recorded after adsorption of the
molecules on synthetic high-surface-area MgO; Fig. 3 compiles
the difference IR spectra of phenol, o-cresol, salicylic alcohol, and
salicylic aldehyde, at 250 °C. Attribution of bands to specific vibra-
tion modes was done by comparing experimental with calculated
IR spectra, for adsorption at the corner site, which according to
the computational study is the energetically preferred one. For
example, Table S1 summarizes the experimental and calculated
frequencies and the corresponding vibration modes for phenol
adsorption.
3.2. Sorption of phenol, o-cresol, salicylic alcohol, and salicylic
aldehyde on MgO
Fig. 2 compiles the optimized models for the adsorption of
phenol, o-cresol, salicylic alcohol, and salicylic aldehyde on three
In the case of spectrum recorded after phenol adsorption, a small
ꢀ1
negative band at 3740 cm indicates that some phenol molecules
interact with residual O–H groups on the MgO surface. Two small
ꢀ
1
bands of perturbed O–H groups at 3705 and 3581 cm are also ob-
ꢀ1
served. Between 3010 and 3070 cm bands of stretching vibra-
ꢀ1
tions of C–H in the aromatic ring are observed, at 1602 cm and
ꢀ
1
at 1491 cm
the C–C ring vibrations, and at 1170 and at
ꢀ
1
1
1
1
074 cm
the band of C–H bending. The band at 1300 and
264 cm are attributed to C–O stretching mode. As with phenol,
ꢀ
in the case of o-cresol a small negative band is observed at
ꢀ
1
3
3
740 cm , together with two bands of perturbed O–H groups at
ꢀ1
713 and 3585 cm . As for phenol, the adsorption at 250 °C leads
to an alteration of vibration modes of phenolic OH in the 1200 cm
region, due to the dissociation of the hydroxyl group.
ꢀ
1
The difference IR spectrum for salicylic alcohol indicates
ꢀ1
adsorption on O–H groups (negative band at 3740 cm ) and
ꢀ1
shows perturbed O–H groups at 3701 and 3584 cm . The bands
ꢀ1
between 2800 and 2950 cm are attributed to C–H stretching of
the hydroxymethyl group, while bands between 3000 and
ꢀ
1
3
100 cm
are attributed to the C–H stretching vibrations of
Fig. 1. Selectivity to products as function of phenol conversion in the reaction
between phenol and methanol over low-surface-area MgO. Reaction conditions: T
ꢀ1
the aromatic ring. Band at 1449 cm is attributed to the bending
of the C–H in the hydroxymethyl group. At 1597 and 1538 cm
the aromatic C–C stretching vibrations and at 1281 cm
ꢀ1
3
00 °C; residence time range: 0.04–0.4 s; methanol-to-phenol molar feed
ꢀ1
ratio = 10/1. Symbols: anisole (ꢀ), salicylic aldehyde (N), and o-cresol (j).
the

F. Cavani et al. / Journal of Catalysis 269 (2010) 340–350
343
Fig. 2. Models of adsorption for (from top to bottom) phenol, o-cresol, salicylic alcohol, and salicylic aldehyde at the different sites of the (MgO)12 cluster (from left to right):
corner site, step site, and terrace site.
stretching of the bond between the O atom and the aromatic ring
the adsorbed alcohol is tentatively attributed to a fraction of sal-
icylic aldehyde formed upon adsorption (see below). The main
difference between the experimental and the gas-phase spectrum
ꢀ
1
are observed. The band at 1153 cm is assigned to the C–CH
2
OH
ꢀ1
stretching vibration. The band at 1629 cm in the spectrum of

3
44
F. Cavani et al. / Journal of Catalysis 269 (2010) 340–350
Table 1
Main molecular properties of phenolic reactants, intermediates, and products adsorbed on MgO surface: adsorption energy (Eads), interatomic distances (d), and angle (Å) at
equilibrium. In parenthesis are given the BSSE-corrected adsorption energies.
Phenol-isolated
Phenol-corner
Phenol-step
Phenol-terrace
ads (kcal mol 1
ꢀ
)
ꢀ43
ꢀ27 (ꢀ20)
2.021
1.055
1.421
1.346
ꢀ17 (ꢀ10)
2.691
1.647
1.010
1.363
E
d (Mgcry–Omol) (Å)
d (Ocry–Hmol) (Å)
d (O–H) (Å)
1.909
1.001
1.709
1.342
128.77
0.966
1.389
108.83
d (C–O) (Å)
A (CÔH) (°)
128.71
111.29
o-Cresol-isolated
o-Cresol-corner
o-Cresol-step
o-Cresol-terrace
ads (kcal mol 1
ꢀ
)
ꢀ40
ꢀ24 (ꢀ17)
2.037
1.070
1.393
1.349
ꢀ14 (ꢀ8)
2.661
1.646
1.010
1.366
E
d (Mgcry–Omol) (Å)
d (Ocry–Hmol) (Å)
d (O–H) (Å)
1.912
1.000
1.704
1.340
141.00
0.965
1.370
109.41
d (C–O) (Å)
A (CÔH) (°)
122.65
111.08
Salicylic alcohol-isolated
Salicylic alcohol-corner
Salicylic alcohol-step
Salicylic alcohol-terrace
ꢀ
E
ads (kcal mol 1
)
ꢀ54
ꢀ35 (ꢀ22)
2.057
1.037
1.519
1.346
107.82
2.258
1.415
1.051
1.431
ꢀ30 (ꢀ16)
2.492
1.053
1.449
1.316
104.65
2.317
1.948
0.981
1.441
d (Mgcry–Omol) (Å)
d (Ocry–Hmol) (Å)
d (O–H) (Å)
1.997
0.972
2.426
1.315
52.245
2.384
1.749
0.996
1.439
0.977
1.361
106.94
d (C–O) (Å)
A (CÔH) (°)
d (Mgcry–OAlc) (Å)
d (Ocry–HAlc) (Å)
d (O–HAlc) (Å)
0.966
1.444
d (CH
2
–OAlc) (Å)
Salicylic aldehyde-isolated
Salicylic aldehyde-corner
Salicylic aldehyde-step
Salicylic aldehyde-terrace
ads (kcal mol 1
ꢀ
)
ꢀ61
ꢀ25 (ꢀ17)
2.152
1.007
1.586
1.306
128.91
2.198
1.241
ꢀ4 (6)
2.244
1.031
1.496
1.308
142.70
2.273
1.225
E
d (Mgcry–Omol) (Å)
d (Ocry–Hmol) (Å)
d (O–H) (Å)
2.049
0.974
1.758
1.295
115.98
2.101
1.253
0.989
1.339
107.03
d (C–O) (Å)
A (CÔH) (°)
d (Mgcry–OAld) (Å)
d (C@OAld) (Å)
1.235
(
1
NIST library) concerns the C–O stretching vibration shifting from
3.3. IR spectra of methanol under reaction conditions
ꢀ1
ꢀ1
237 cm
(in the gas-phase spectrum) to 1291 cm
(in the
experimental one).
For adsorbed salicylic aldehyde, bands between 3000 and
Fig. 4 compares the difference spectra registered in the presence
of flowing methanol at 250 °C over the high-surface-area MgO with
the gas-phase spectrum of methanol. The shape and intensity of
the bands indicate that some contribution of gas-phase methanol
is present in the difference spectra. On the other side, the differ-
ence between the spectrum of methanol in the gas-phase and that
of flowing methanol over MgO also show that bands due to surface
structures can be discriminated from the overlaying gas-phase
spectra.
ꢀ1
3
090 cm
(attributed to the C–H stretching bands of the aro-
ꢀ1
matic ring) as well as bands at 2876 and 2779 cm (attributed
the stretching of the C–H band of the aldehyde group) are ob-
served; the latter one derives from the Fermi resonance [51].
The characteristic band of the aldehyde C@O stretching vibration
ꢀ1
appears at 1638 cm . The strong shift to lower wavenumbers
ꢀ1
compared to the gas-phase spectrum (1680 cm , NIST library)
is due to the interaction of the carbonyl group with Mg2 cations
+
An increase in the feed time led to an increase in the intensity
ꢀ1
ꢀ1
acting as Lewis acid sites. The bands at 1606 and 1535 cm are
attributed to the aromatic ring stretching vibrations, while the
bands between 1409 and 1467 cm to C–H bending vibrations.
for bands at 1009 (shoulder) and 1031 cm ; the same occurred
ꢀ1
for the most intense peak at 1061 cm . Bands at 1116 and
ꢀ
1
ꢀ1
1160 cm are assigned to adsorbed methanol. The latter band is
ꢀ
1
The band at 1333 cm is attributed to C–O, being remarkably
attributed to the rocking vibration of methyl group [36]. In this re-
gion, the bands due to vibrations of gas-phase methanol are easily
shifted with respect to the corresponding bands in o-cresol and
ꢀ
1
ꢀ1
salicylic alcohol spectra. The band at 1149 cm corresponds to
the vibration of C–CHO. According to the computational model,
the salicylic aldehyde is deprotonated leading to an enhanced
resonance effect and the consequent alteration of the vibrations
of the carbonyl group. The stabilization of the aldehyde group
recognized. Above 1300 cm , bands at 1359, 1394, 1455, and
ꢀ
1
1584 cm
were observed. In the wavenumber region of C–H
stretching bands characteristic of gas-phase methanol overlapped
with bands attributed to adsorbed methanol. A broad band was ob-
ꢀ1
served between 3550 and 3730 cm , together with a negative
2ꢀ
2+
ꢀ1
by the basic O and the Mg cation is in agreement with the
shift experimentally found for the carbonyl bond stretching. It
is important to note that the aldehyde interacts with MgO at
peak at 3747 cm
groups.
attributed to hydrogen bonding on MgOH
One relevant aspect of this experiment is the asynchronous var-
iation of the intensities during the reaction of methanol, especially
2
50 °C but does not undergo any transformation such as the dis-
ꢀ1
mutation via the Tishchenko reaction [52]. This implies that sal-
icylic aldehyde can be isolated as an intermediate in the reaction
between phenol and methanol.
in the region between 1200 and 1800 cm . The relative intensities
ꢀ1
of bands at 1322, 1332, 1359, 1394, 1455, 1477, and 1584 cm re-
mained approximately constant. The bands with maxima at 1610

F. Cavani et al. / Journal of Catalysis 269 (2010) 340–350
345
Fig. 4. Details of spectral regions from the spectra recorded by flowing methanol at
50 °C (dotted spectrum: gas-phase methanol) over the high-surface-area MgO.
Spectra plotted at increasing times from bottom to top.
2
COO– at 1322, 1359, and 1394 cmꢀ1, and
m
as COO– at 1639, 1610,
ꢀ1
ꢀ1
and 1584 cm , while the
m
CH
3
falls at 1332 cm . The shoulder
ꢀ1
at 2790 cm
s
is likely due to the combined mode m (COO–) +
ꢀ1
m(CH). The band with maxima at 1610 and 1639 cm (and the cor-
responding symmetric stretching vibrations) can be attributed to
the formation of bidentate carbonates, while that one at
ꢀ
1
1
(
584 cm
is related to the development of formate species
, 2790 C–H) [54]. Therefore, the data indicate
1584 as, 1390
m
m
s
m
the initial formation of formates followed by the slower formation
of carbonates.
Fig. 5 displays the spectra recorded at 450 °C. Under these con-
ditions, the contribution of gas-phase methanol was lower than
that at 250 °C. The band in the region between 1500 and
ꢀ
1
ꢀ1
1
700 cm has maxima at 1612 and 1628 cm , with shoulders
ꢀ1
at 1590 and 1560 cm . The band is similar to that obtained by
dosing CO over MgO [55]. In contrast with what observed at
50 °C, there was no asynchronous variation of bands; this is likely
2
2
due to the fact that the reaction having the higher activation en-
ergy was kinetically favored at the higher temperature. An evident
phenomenon is the development of negative peaks in the hydroxyl
ꢀ1
ꢀ1
zone at 3747 cm and at 1540 cm
(mCOO–) in the carbonates
zone; it is an evidence of the presence of a methanol decomposi-
ꢀ1
tion phenomenon. At 2780 and 2750 cm are the bands due to
the stretching vibration of the aldehydic C–H bond.
Fig. 3. Difference IR spectra of phenol (top spectrum), o-cresol (middle top
spectrum), salicylic alcohol (middle bottom spectrum), and salicylic aldehyde
(
bottom spectrum) adsorbed over high-surface-area MgO.
3
.4. IR Spectra during reaction of methanol and phenol
ꢀ
1
and 1639 cm appeared only after some minutes after methanol
was admitted and finally became more intense than the band cen-
Fig. 6 reports the spectra recorded, while feeding methanol and
ꢀ1
tered at 1584 cm . These bands are attributable to stretching
vibrations of carbonates having different co-ordination [53]:
phenol at 250 °C over the high-surface-area MgO, after subtraction
of the MgO spectrum, and with the further subtraction of the
m
s

3
46
F. Cavani et al. / Journal of Catalysis 269 (2010) 340–350
Fig. 5. Details of spectral regions from the spectra recorded by flowing methanol at
50 °C (dotted spectrum: gas-phase methanol) over the high-surface-area MgO.
4
Spectra plotted at increasing times from bottom to top.
Fig. 6. Difference IR spectra of surface species while passing methanol and phenol
over high-surface-area MgO and the IR spectra in the presence of methanol only
corresponding spectra obtained when only methanol was fed at
the same time on stream (Fig. 4). This allows to minimize the inter-
ference of methanol (in the gas-phase and adsorbed state) and of
its products of transformation. Clearly, this is valid only provided
the reaction network of methanol transformation is the same both
in the absence and in the presence of phenol. Indeed, the spectra
obtained have negative bands, likely due to the fact that the corre-
sponding products formed from methanol either did not form in
the presence of phenol or readily reacted with the latter.
(Fig. 4) at 250 °C. Spectra plotted at increasing times from bottom to top.
ꢀ
1
3
027, and 3058 cm , and the decrease in the intensity for bands
ꢀ1
at 2918, 2805, and 2771 cm . This indicates that at short times
on stream the relative concentration of the adsorbed species
changes remarkably. At first, the band at 1448 cm is in line with
ꢀ1
It is evident that the spectrum recorded after reaction includes
not only the bands relative to adsorbed reactants, but also addi-
tional bands due to the reaction products. One important feature
the rapid formation and adsorption of o-cresol. With time on stream,
ꢀ1
an aromatic compound with a C@O group (1625 cm , i.e., salicylic
aldehyde) increases in concentration. Both products seem to be sta-
ble under the reaction conditions probed, and evidences for the
interconversion are not observable in the in situ IR spectra.
ꢀ1
was the band at 1597 cm that was present both in the spectrum
recorded after reaction and in the spectra of reaction intermediates
ꢀ1
as well (Fig. 3). Also the band at 1488 cm was present in all spec-
tra, but not in that of adsorbed salicylic alcohol. The band at
ꢀ
1
1
080 cm is likely to be attributed to methanol. An intense peak
3.5. The reactivity of salicylic alcohol
ꢀ1
at 1625 cm was also present in the spectrum of adsorbed sali-
cylic aldehyde (but it cannot be excluded to be due to a formate
species). In the spectral region between 3200 and 2600 cm , the
band at 2805 cm can be attributed to methanol or to a product
of its transformation, while bands between 3058 and 3027 cm
are due to aromatic C–H vibrations.
Spectroscopic measurements and catalytic tests suggest that
one key intermediate of the reaction is salicylic alcohol. In order
to understand the possible elementary steps for the transformation
of salicylic alcohol into o-cresol, we performed reactivity tests by
feeding either a salicylic alcohol/water solution or a salicylic
ꢀ
1
ꢀ1
ꢀ1
The main time-dependent features of the spectra include (i) the
alcohol/formalin aqueous solution, both vaporized in a N stream.
Table 2 summarizes the results obtained at 300 °C.
2
ꢀ1
progressive increase in the intensity of band at 1448 cm attrib-
uted to the CH bending vibration in o-cresol, (ii) the concomitant
decrease in the intensity of the aldehydic C–H stretching band be-
3
In both cases, there was a total conversion of salicylic alcohol;
this explains why this compound was not isolated as an intermedi-
ate in phenol methylation even at a very low contact time (Fig. 1).
In the absence of formaldehyde, salicylic alcohol produced phenol
and o-cresol in an almost equimolar amount. The C balance, how-
ever, was close to 80–85%, highlighting the formation of heavy
ꢀ1
tween 2700 and 2800 cm , and (iii) the increase in the band at
ꢀ1
1
625 cm
The spectra reported in Fig. 6 show the progressive increase in
the intensity for bands at 1297, 1340, 1388, 1448, 1488, 1625,
.

F. Cavani et al. / Journal of Catalysis 269 (2010) 340–350
347
Table 2
OH
OH
OH
O
Reactivity of salicylic alcohol over MgO at 300 °C.
OH
Product
Selectivity (%)
2
+
H O
2
+
Feed: salicylic
alcohol in water
Feed: salicylic alcohol
in aqueous formalin
ð4Þ
Phenol
o-Cresol
45
55
0
0
0
10
35
24
3
1
2
,6-Xylenol
Salicylic aldehyde
-Hydroxy-
-methylbenzaldehyde
Toluene
Tests made with salicylic alcohol and formalin yielded several
compounds (Table 2); the prevailing ones were o-cresol, 2,6-xyle-
nol, and 2,4,6-trimethylphenol, also with the presence of non-neg-
ligible amounts of phenol and 4H-benzo[1,3]dioxine. The mono-,
di-, and tri-methylated phenols form by a reaction of phenol and
o-cresol (the main compounds formed in the absence of formalin)
with formaldehyde; the absence of the corresponding hydroxyme-
thylated intermediates confirms that the latter are highly reactive.
The dioxine forms by O-hydroxymethylation on the two hydroxy
groups of salicylic alcohol, with the formation of a cyclic and rela-
tively stable intermediate. The non-cyclic, less stable compound is
the 2-hydroxy-3-methylbenzaldehyde, which formed in much
smaller amount.
2
3
0
0
0
1
13
13
2
4
,4,6-Trimethylphenol
H-benzo[1,3]dioxine
by-products; the latter were identified as di-aryl compounds,
formed either by condensation of the hydroxymethyl groups of
two salicylic alcohol molecules or the ring-alkylation of one sali-
cylic alcohol molecule onto a second one. It is worthy of note that
in the absence of catalyst still there was a total conversion of sali-
cylic alcohol, and the same products obtained in the presence of
MgO were also obtained, but with the predominance of o-cresol,
while salicylic aldehyde formed in an amount equivalent to phenol.
At 200 °C, both in the presence and in the absence of MgO, the total
salicylic alcohol fed was converted to these heavy by-products. It is
apparent that in the presence of a relatively high partial pressure of
salicylic alcohol, the alcohol reacts very quickly with another mol-
ecule of alcohol, both in the presence and in the absence of cata-
lysts, yielding di-aryl compounds, which possibly further
decompose according to the stoichiometry:
Among the light compounds, we noted the very significant for-
mation of CO
methane or hydrogen. The absence of H
out reactions involving H O as the reactant (e.g., WGS of CO), and a
2
, with only traces of CO, and no formation of either
2
makes it possible to rule
2
Cannizzaro-like disproportion of formaldehyde leading to metha-
nol (that was also not formed in the reaction; the small amount
of methanol detected was that originally present in the formalin
solution) and formic acid; in fact, the decomposition of the acid
would lead either to CO and H , or to H O and CO. Also, the ab-
2 2 2
sence of methane makes it possible to rule out a Tishchenko-like
dimerization of formaldehyde to methylformate as well as its
OH
OH
OH
decomposition into CO
was most likely due to the Oppenauer-type oxidation:
2
and CH
4
. Therefore, the formation of CO
2
OH
OH
2
H O
OH
O
OH
+
2
ð5Þ
ð1Þ
+
H CO
+
CO2
2
OH
OH
OH
OH
When the same reaction was carried out without the MgO cat-
alyst, the predominant product was 4H-benzo[1,3]dioxine, with
around 60% selectivity. o-Cresol also formed, but there were no
di- and tri-methylated compounds. Therefore, at the conditions
used (T 300 °C) the catalyst is essential for the C-hydroxymethyla-
tion reaction, but it is not necessary for the O-hydroxymethylation
of salicylic alcohol, and the redox reaction between salicylic alde-
hyde and formaldehyde.
An experiment done by feeding salicylic aldehyde with formalin
at 350 °C confirmed that the reaction (5) might be the one possible
pathway for o-cresol formation. In fact, salicylic aldehyde gave al-
most total (98%) conversion to o-cresol (selectivity 30%) and 2,6-
xylenol (selectivity 70%).
OH
+
+
CO
ð2Þ
Heavy compounds may also form by the MgO-catalyzed con-
densation reaction of two adsorbed molecules of salicylic aldehyde,
the latter having been formed by alcohol dehydrogenation. More-
over, phenol may also form by direct decomposition of salicylic
alcohol:
OH
OH
ð3Þ
OH
4
. Discussion
+
H CO
2
4.1. The interaction of reactants and products with MgO
However, the formation of CO was very low, thus suggesting
that a direct disproportion is likely the dominant mechanism for
o-cresol formation from salicylic alcohol (reaction (4)), even
though in the presence of MgO, the salicylic aldehyde formed re-
mains strongly bound to the catalyst surface. Indeed, the spent cat-
alyst was brown, thus confirming the accumulation of heavy
compounds on its surface during reaction.
The interaction of methanol with MgO (as model oxide for basic
oxides) has been studied by several authors [56–61] and it is well
established that MgO catalyzes the dehydrogenation of methanol
[62]. Chemisorption occurs via heterolytic dissociation [58] involv-
ꢀ
ing the formation of hydroxyl groups and CH
3
O
anions [63].
ꢀ
2+
Further abstraction of H by an accessible Mg cation or a carbo-
naceous residue completes the step. Formation of surface formate

3
48
F. Cavani et al. / Journal of Catalysis 269 (2010) 340–350
species may occur through nucleophilic interaction of an adjacent
oxygen atom with the carbon of the methylene group. Alterna-
tively two sorbed formaldehyde molecules may disproportionate
lower the temperature is, i.e., the lower the degree of dehydroge-
nation of methanol, the more significant the formation of anisole
becomes. We also found that reacting phenol with formalin yielded
the same product distribution as that obtained by reacting phenol
with methanol.
(via a nucleophilic attack) to formate and a methoxy species
60,64]. Formate may decompose to CO and H . Alternatively, ad-
[
2
sorbed methoxy and formate species may yield methylformate
Tishchenko reaction) which is able to decompose at high temper-
atures to methane and CO [65].
The present data show that already at 250 °C the interaction of
methanol with the clean catalytic surface generates formate and
carbonate species, giving unequivocal evidence for the dehydroge-
nation of methanol and of the strong interaction of the resulting
products with the basic surface. At high temperature the extent
of methanol decomposition was relevant indicating that the form-
aldehyde generated is very reactive and is rapidly transformed to
The catalytic tests demonstrate that at moderate temperature
(i.e., 300 °C) salicylic aldehyde is a primary reaction product (next
to o-cresol and anisole) which can be isolated only at very low
phenol conversions (low contact time). As it is seen as product
accumulating also with higher time on stream (see Fig. 6) we con-
clude that when more products and reactants are sorbed at the sur-
face, salicylic aldehyde reacts preferentially with surface species to
o-cresol. Under conditions of low surface coverage, salicylic alde-
hyde is not further converted to consecutive products because
the low concentration of adsorbed methanol and formaldehyde
makes the secondary reaction less likely. When the concentration
of molecules at the adsorbed state increases, salicylic aldehyde is
rapidly transformed to o-cresol. We speculate at present that the
salicylic aldehyde transformation may occur in bimolecular dispro-
portionation-like reaction (intermolecular H-transfer) from ad-
sorbed methanol, formaldehyde, or another functionalized
aromatic compound.
(
2
4 2
light compounds, i.e., CO, CH , and CO .
In line with the earlier reports [23,50,66,67] the present data
show that the adsorption of phenol and phenol analogs (o-cresol,
salicylic alcohol, and salicylic aldehyde) on highly co-ordinatively
unsaturated sites (corner sites) is energetically favorable. In this
case, the interaction between MgO and the adsorbate is very strong
and leads to the dissociation of the sorbate at room temperature.
The reaction between an activated aromatic molecule and form-
aldehyde is also known as hydroxymethylation, and leads to the
formation of salicylic alcohol in the case of phenol. This reaction
occurs in the liquid phase at moderate temperatures, and is cata-
lyzed by acidic and basic catalysts [68,69]. The IR spectra of ad-
sorbed salicylic alcohol (see characteristic bands at 1653 and
4
.2. The reaction between adsorbed reactants
In the literature, the mechanism of reaction between phenol
and methanol catalyzed by basic oxides is believed to occur by a
direct electrophilic substitution on the aromatic ring with only
few papers considering a possible role of formaldehyde. Palomares
et al. [39,40] demonstrated that in the base-catalyzed side-chain
methylation of toluene to styrene, it is important that (i) methanol
needs to be dehydrogenated to formaldehyde forming a strongly
positively polarized C atom (through the interaction between the
carbonyl and the metal cation) and (ii) the methyl group of toluene
needs to interact with the basic oxygen atoms to form a partial
negative charge on the C atom. In consequence, an aldol-type C–
C bond formation may occur followed by dehydration and the for-
mation of the C@C moiety. The limiting factor in this chemistry is
the low stability of the intermediately formed formaldehyde which
would lead to partial decomposition at the reaction temperatures
required for side-chain alkylation.
ꢀ
1
1629 cm in Fig. 3) indicate that at least a significant fraction is
rapidly transformed to salicylic aldehyde.
4.3. The mechanism of transformation of salicylic alcohol into o-cresol
Reactivity tests made by feeding salicylic alcohol showed that
this compound, the very first intermediate in the C-hydroxymethy-
lation of phenol by formaldehyde, is very reactive and is rapidly
transformed into either heavy di-aryl compounds or mono-aryl
phenolic compounds. The formation of heavy products (a reaction
occurring even in the absence of catalyst) is significant when sali-
cylic alcohol is directly fed into the reactor, but is likely less impor-
tant during the reaction of phenol methylation.
The transformation of alcohol into salicylic aldehyde, a primary
product of the reaction (Fig. 1), may occur via either direct dehy-
drogenation, or disproportionation (reaction (4)), which would ex-
plain the experimental evidence of both phenolic compounds
being primary products (Fig. 1), as well as the substantial absence
The question arises if intermediately formed formaldehyde
would also play an important role in the C-alkylation of phenol
over basic catalysts. Due to the higher intrinsic reactivity of the
aromatic ring, methylation may occur under milder reaction condi-
tions and with catalysts having lower basic strength than for the
side-chain alkylation of toluene.
of light compounds (CO, CO
alcohol.
2 2
, H ) in tests made by feeding salicylic
Indirect evidence for such chemistry has been reported in the
literature. Using IR spectroscopy Mathew et al. [23] studied the
development of formaldehyde, dioxymethylene, and formate spe-
cies from methanol on Cu/Co/Fe/O catalysts (decomposition of
the products above 300 °C), but also reported that due to the com-
petition of phenol for adsorption on the sites responsible for meth-
anol oxidation, decomposition did not occur in the presence of
both compounds. Padmasri et al. [27] also speculated that in the
reaction between phenol and isobutanol on MgO and on calcined
hydrotalcites, the alkoxy species might dehydrogenate forming
an adsorbed isobutyraldehyde intermediate. The latter then was
reported to react with phenol to yield either o-butenylphenol, after
dehydration of the intermediate alcohol, or 2-butenylphenol.
In recent contributions [41,42] it was shown by us that under
conditions favoring methanol dehydrogenation, formaldehyde is
the true reacting species with the phenolate yielding o-cresol and
Finally, the consecutive transformation of salicylic aldehyde to
yield o-cresol occurs by an Oppenauer-type oxidation which ex-
plains the formation of CO (reaction (5)).
2
A similar reaction may occur between salicylic alcohol and
formaldehyde to yield o-cresol as one primary reaction product:
OH
OH
OH
+
H CO
+
CO2
+
H2
2
ð6Þ
This reaction may prevail over that involving two molecules of
salicylic alcohol (reactions (1) and (2)) under conditions of low sur-
face concentration of the adsorbed intermediates, that is, under the
conditions normally used for phenol methylation.
2
,6-xylenol. The products obtained at low temperatures and con-
ventional contact times are anisole and o-cresol, the latter domi-
nating between 300 and 450 °C. It is important to note that the

F. Cavani et al. / Journal of Catalysis 269 (2010) 340–350
349
Fig. 7. Overall reaction scheme of phenol methylation with methanol catalyzed by MgO.
Tests made without the catalyst showed that under the condi-
tions used, the catalyst is essential for the C-hydroxymethylation
reaction (phenol to salicylic alcohol), but it may not be necessary
for the disproportion of salicylic alcohol to o-cresol and salicylic
aldehyde, and the consecutive reaction between salicylic aldehyde
and formaldehyde to yield o-cresol.
It is worth noting that in the reaction between phenol and
methanol, salicylic aldehyde can be isolated only when extremely
low contact times are used (Fig. 1), i.e., under conditions at which
the surface coverage by adsorbed species is very low. When, on the
contrary, reaction conditions are used that may favor the accumu-
lation of adsorbed intermediates on the catalyst surface, e.g., by
feeding salicylic alcohol and by using low reaction temperatures,
the formation of heavy compounds is kinetically preferred. Finally,
when reacting phenol and methanol at normal reaction conditions
the concentration of adsorbed salicylic alcohol and aldehyde is
very low, and the two compounds readily react with methanol
and formaldehyde to yield o-cresol.
The third primary product, anisole, is only observed at low tem-
peratures. It is formed by direct reaction between activated metha-
nol and phenol with water as the leaving group. Because at high
temperature, i.e., under conditions at which methanol dehydroge-
nation occurs with feasible rates, the only reaction products are o-
cresol and 2,6-xylenol, we conclude that the reaction occurs in a
concerted mechanism using the polarization of the phenol OH and
methanol C–O bond. Because we compare the selectivity at different
temperature for two different routes, if the activation energies for
both are far different, the selectivity will be different. It seems, the
anisole formation with very low activation energy and at low tem-
perature, it is the favorite product. Alternatively it could be specu-
lated that anisole is formed by a reaction between phenolate and
methylformate (formed by dimerization of HCHO). At higher tem-
The nature of the products generated by methanol interacting
with the catalyst is a function of the reaction conditions, and it af-
fects the primary products obtained in phenol methylation. Under
conditions at which the extent of methanol dehydrogenation is
low, i.e., low temperature, the main primary products of reaction
are anisole and o-cresol. Under conditions more favorable for
methanol dehydrogenation, anisole is no longer formed, and o-cre-
sol becomes the only reaction product. 2,6-Xylenol forms in signif-
icant concentrations above 350 °C and for high phenol conversion.
The adsorption of phenol on MgO generates a phenolate species,
and the energetically preferred mode of adsorption is on the corner
site of MgO, with an almost orthogonal orientation of the aromatic
ring with respect to the catalyst surface. The adsorption of o-cresol
and of the reaction intermediates, salicylic alcohol, and salicylic
aldehyde at the corner site of MgO is also energetically favored.
In all cases, the adsorbed molecules dissociate forming a pheno-
late-like species and a hydroxyl group.
The reaction between adsorbed phenolate and formaldehyde
generates salicylic alcohol via hydroxymethylation, which is rap-
idly transformed to salicylic aldehyde. At low temperature and for
very low conversions and short residence time, salicylic aldehyde
is one of the primary reaction products. However, the aldehyde is
very rapidly transformed into o-cresol, and is not detectable for
phenol conversion higher than 0.05%. Below 300 °C anisole and o-
cresol are formed, the latter being formed via either direct reduc-
tion of salicylic alcohol by formaldehyde, or the bimolecular dispro-
portion of salicylic alcohol to o-cresol and salicylic aldehyde.
In situ spectroscopy shows salicylic aldehyde to be the reaction
intermediate in o-cresol formation. The latter is observed during
the initial accumulation of adsorbed species on the catalyst sur-
face, i.e., while a low surface coverage prevails. It is not observed
under steady conditions, i.e., when the catalyst surface is fully cov-
ered, in line with the fact that salicylic aldehyde is only observed at
extremely low conversions.
perature, methylformate readily decomposes to, e.g., CH
4 2
and CO .
Overall the results demonstrate that subtle tailoring of the base
properties allows to use methanol in a dual role as alkylating agent
and as reducing agent of the alcohol/aldehyde group created.
While this is important for the selective alkylation of phenols,
the insight into the reaction may even be more important for the
defunctionalization of substituted phenols potentially available at
large scale from deconstructed lignin.
5
. Conclusions
The combination of catalytic tests, IR spectroscopy of adsorbed
reactants, products, and possible intermediates together with a
computational study as well as in situ IR spectra during reaction al-
lows for a full description of the reaction network in the methyla-
tion of phenol (see Fig. 7).

3
50
F. Cavani et al. / Journal of Catalysis 269 (2010) 340–350
[
[
30] M. Bolognini, F. Cavani, D. Scagliarini, C. Flego, C. Perego, M. Saba, Catal. Today
5 (2002) 103.
31] F. Cavani, C. Felloni, D. Scagliarini, A. Tubertini, C. Flego, C. Perego, Stud. Surf.
Sci. Catal. 143 (2002) 953.
Acknowledgments
7
This work has been conducted in the framework of Network of
Excellence IDECAT (FP6 of the EU). INSTM is acknowledged for the
Grant of SP and LM.
[32] M. Bolognini, F. Cavani, D. Scagliarini, C. Flego, C. Perego, M. Saba, Micropor.
Mesopor. Mater. 66 (2003) 77.
[
33] F. Cavani, L. Maselli, D. Scagliarini, C. Flego, C. Perego, Stud. Surf. Sci. Catal. 155
2005) 167.
34] US Patent 4,933,509, 1989, assigned to General Electric.
(
[
Appendix A. Supplementary material
[35] P.D. Chantal, S. Kaliaguine, J.L. Grandmaison, Appl. Catal. 18 (1985) 133.
[
36] 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) 283.
37] R. Pierantozzi, A.F. Nordquist, Appl. Catal. 21 (1986) 263.
38] S.T. King, J.M. Garces, J. Catal. 104 (1987) 59.
39] A.E. Palomares, G. Eder-Mirth, J.A. Lercher, J. Catal. 168 (1997) 442.
40] A.E. Palomares, G. Eder-Mirth, M. Rep, J.A. Lercher, J. Catal. 180 (1998) 56.
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.jcat.2009.11.019.
[
[
[
[
References
[41] N. Ballarini, F. Cavani, L. Maselli, A. Montaletti, S. Passeri, D. Scagliarini, C.
Flego, C. Perego, J. Catal. 251 (2007) 423.
[
[
[
[
[
[
[
1] H. Hattori, Chem. Rev. 95 (1995) 537.
2] Y. Ono, T. Baba, Catal. Today 38 (1997) 321.
3] R.J. Davis, J. Catal. 216 (2003) 396.
4] Y. Ono, J. Catal. 216 (2003) 406.
5] A. Cimino, F.S. Stone, Adv. Catal. 47 (2002) 141.
6] M. Barteau, Chem. Rev. 96 (1996) 1413.
7] K. Tanabe, T. Nishizaki, in: G.C. Bond et al. (Eds.), Proceedings of Sixth ICC, The
Chemical Society, London, 1977, p. 863.
8] S. Sato, K. Koizumi, F. Nozaki, Appl. Catal. A 133 (1995) L7.
9] K. Tanabe, H. Hattori, T. Sumiyoshi, K. Tamaru, T. Kondo, J. Catal. 53 (1978) 1.
[42] N. Ballarini, F. Cavani, L. Maselli, S. Passeri, S. Rovinetti, J. Catal. 256 (2008) 215.
[43] S. Sasaki, K. Fujino, Y. Takeuchi, Proc. Jpn. Acad. 55 (1979) 43.
[44] E.A. Colbourn, J. Kendrick, C. Mackrodt, Surf. Sci. 126 (1983) 350.
[45] M.W. Wong, Chem. Phys. Lett. 256 (1996) 391.
[46] M.J. Frisch et al., Gaussian Inc., Wallingford CT, Gaussian 03, Revision C.02,
2004.
[47] R. Dennington II, T. Keith, J. Millam, K. Eppinnett, W.L. Hovell, R. Gilliland,
GaussView, Version 3.09, Semichem Inc., Shawnee Mission, KS, 2003.
[48] G. Mirth, J.A. Lercher, J. Catal. 132 (1991) 244.
[49] G. Mirth, J.A. Lercher, J. Catal. 147 (1994) 199.
[50] R.N. Spitz, J.E. Barton, M.A. Barteau, R.H. Staley, A.W. Sleight, J. Phys. Chem. B
90 (1986) 4067.
[51] D.F. Eggers Jr., W.E. Lingren, Anal. Chem. 28 (1956) 1328.
[52] K. Tanabe, K. Saito, J. Catal. 35 (1974) 247.
[53] Y. Fukuda, K. Tanabe, Bull. Chem. Soc. Jpn. 46 (1973) 1616.
[54] G. Busca, J. Lamette, J.-C. Lavalley, V. Lorenzelli, J. Am. Chem. Soc. 109 (1987)
5197.
[55] J.V. Evans, T.L. Whateley, Trans. Faraday Soc. 63 (1967) 2769.
[56] G. Busca, Catal. Today 27 (1996) 457.
[57] J.C. Lavalley, Catal. Today 27 (1996) 377.
[58] J. Günster, G. Liu, J. Stultz, S. Krischok, D.W. Goodman, J. Phys. Chem. 104
(2000) 5738.
[59] C. Di Valentin, A. Del Vitto, G. Pacchioni, S. Abbet, A.S. Wörz, U. Heiz, J. Phys.
Chem. B 106 (2002) 11961.
[
[
[
[
[
[
[
[
[
[
[
[
10] R. Bal, B.B. Tope, S. Sivasanker, J. Mol. Catal. A 181 (2002) 161.
11] K.T. Li, I. Wang, K.R. Chang, Ind. Eng. Chem. Res. 32 (1993) 1007.
12] V. Durgakumari, S. Narayanan, J. Mol. Catal. 65 (1991) 385.
13] K. Tanabe, K. Shimazu, H. Hattori, K. Shimazu, J. Catal. 57 (1979) 35.
14] T. Kotanigawa, K. Shimokawa, Bull. Chem. Soc. Jpn. 47 (1974) 950 and 1535.
15] H. Grabowska, W. Kaczmarczyk, J. Wrzyszcz, Appl. Catal. 47 (1989) 351.
16] S. Velu, C.S. Swamy, Appl. Catal. A 145 (1996) 141.
17] V.V. Rao, K.V.R. Chary, V. Durgakumari, S. Narayanan, Appl. Catal. 61 (1990) 89.
18] S. Karuppannasamy, K. Narayanan, C.N. Pillai, J. Catal. 66 (1980) 281.
19] H. Grabowska, W. Mista, J. Trawczynski, J. Wrzyszcz, M. Zawadzki, Appl. Catal.
A 220 (2001) 207.
[
[
20] K. Sreekumar, S. Sugunan, Appl. Catal. A 230 (2002) 245.
21] T. Mathew, N.R. Shiju, K. Sreekumar, B.S. Rao, C.S. Gopinath, J. Catal. 210 (2002)
4
05.
[
[
22] T. Mathew, S. Shylesh, B.M. Devassy, C.V.V. Satyanarayana, B.S. Rao, C.S.
Gopinath, Appl. Catal. A 261 (2004) 292.
23] T. Mathew, M. Vijayaraj, S. Pai, B.B. Tope, S.G. Hegde, B.S. Rao, C.S. Gopinath, J.
Catal. 227 (2004) 175.
24] S. Sato, K. Koizumi, F. Nozaki, J. Catal. 178 (1998) 264.
25] S. Velu, C.S. Swamy, Appl. Catal. A 119 (1994) 241.
26] S. Velu, C.S. Swamy, Appl. Catal. A 162 (1997) 81.
27] A.H. Padmasri, A. Venugopal, V. Durgakumari, K.S. Rama Rao, P. Kanta Rao, J.
Mol. Catal. A 188 (2002) 255.
[60] X.D. Peng, M.A. Barteau, Langmuir 5 (1989) 1051.
[61] L.J. Burcham, L.E. Briand, I.E. Wachs, Langmuir 17 (2001) 6164.
[62] P. Mars, J.J. Scholten, P. Zwietering, Adv. Catal. 14 (1963) 35.
[63] M. Ziolek, J. Kujawa, O. Saur, J.C. Lavalley, J. Phys. Chem. 97 (1992) 9761.
[64] N.D. Lazo, D.K. Murray, M.L. Kieke, J.H. Haw, J. Am. Chem. Soc. 114 (1992) 8552.
[65] T. Mori, S. Hoshino, A. Neramittagapong, J. Kubo, Y. Morikawa, Chem. Lett.
(2002) 390.
[66] D.R. Taylor, K.H. Ludlum, J. Phys. Chem. 76 (20) (1972) 2882.
[67] H. Hattori, K. Shimazu, N. Yoshii, K. Tanabe, Bull. Chem. Soc. Jpn. 49 (1976) 969.
[68] J. Lecomte, A. Finiels, P. Geneste, C. Moreau, Appl. Catal. A 168 (1998) 235.
[69] M. Ardizzi, F. Cavani, L. Dal Pozzo, L. Maselli, R. Mezzogori, Stud. Surf. Sci. Catal.
158 (2005) 1953.
[
[
[
[
[
[
28] S. Sato, R. Takahashi, T. Sodesawa, K. Matsumoto, Y. Kamimura, J. Catal. 184
(
1999) 180.
29] T.M. Jyothi, T. Raja, M.B. Talawar, B.S. Rao, Appl. Catal. A 211 (2001) 41.