Mechanistic studies of the role of formaldehyde in the gas-phase methylation of phenol

Article abstract of DOI:10.1016/j.jcat.2008.03.012

The gas-phase reaction between phenol and formaldehyde was studied using Bronsted acidic and basic catalysts to gain insight into the components of phenol methylation with methanol. Formaldehyde and methylformate form from methanol during phenol methylati

Full text of DOI:10.1016/j.jcat.2008.03.012

Journal of Catalysis 256 (2008) 215–225
Contents lists available at ScienceDirect
Journal of Catalysis
www.elsevier.com/locate/jcat
Mechanistic studies of the role of formaldehyde in the gas-phase methylation of
phenol
N. Ballarini a,b, F. Cavani
a,b,∗
, L. Maselli a,b, S. Passeri a,b, S. Rovinetti a,b
a
Dipartimento di Chimica Industriale e dei Materiali, Alma Mater Studiorum Università di Bologna, Viale Risorgimento 4, 40136 Bologna, Italy
INSTM, Research Unit of Bologna, A Partner of Idecat NoE (6th FP of the EU), Bologna, Italy
b
a r t i c l e i n f o
a b s t r a c t
Article history:
The gas-phase reaction between phenol and formaldehyde was studied using Brønsted acidic and ba-
sic catalysts to gain insight into the components of phenol methylation with methanol. Formaldehyde
and methylformate form from methanol during phenol methylation over basic catalysts. In the basic-
catalyzed methylation of phenol with formaldehyde, the prevailing phenolic products were o-cresol and
Received 14 January 2008
Revised 10 March 2008
Accepted 15 March 2008
Available online 18 April 2008
2,6-xylenol; the same products were obtained in the methylation with methanol. When formaldehyde
was used with the acidic catalyst, the distribution of products was similar to that obtained over ba-
sic catalysts. Comparing these results with those of methylation with methylformate, methyliodide, and
methanol shows that with basic catalysts, formaldehyde was the true methylating agent in the reaction
between phenol and methanol. This indicates that the transformation of methanol differentiates the re-
action pattern in phenol methylation over basic or acidic catalysts. The role of the catalyst is to generate
the methylating species; the nature of the latter then determines the type of products obtained in phenol
methylation, whereas the catalyst type (either acidic or basic) has little influence on it.
Keywords:
Phenol methylation
o-Cresol
Formaldehyde
Basic catalysis
Acid catalysis
Magnesium–iron mixed oxide
Reaction mechanism
©
2008 Elsevier Inc. All rights reserved.
1
. Introduction
reactant, anisole is the prevailing product at low temperature. Gen-
erally, it is accepted that C-alkylation requires stronger acidic sites
than O-alkylation [24,32,37–39], and indeed less acidic zeolites are
more selective to anisole than to cresols [23,34,40–42].
The methylation of phenol and of phenol derivatives is very
significant from an industrial standpoint [1]. The products of phe-
nol and diphenol alkylation are intermediates for the synthesis of
resins and novolacs, dyes, antiseptics, antioxidants, vitamins, skin
protection agents, and food additives. Methylation can be carried
out in either the liquid phase (homogeneous or heterogeneous) or
the gas phase. The catalysts used for this reaction can be sub-
divided into two groups: (i) catalysts with acidic characteristics
The characteristics of catalysts with basic features when used
for phenol methylation are very high regioselectivity in C-methyl-
ation, because the ortho/para-methylation ratio is always much
greater than 2, and high chemoselectivity, because the O/C-
methylation ratio (a function of the basic strength of the catalysts)
generally is very low. The very high regioselectivity is explained
through the widely accepted model proposed by Tanabe [43,44],
which describes the adsorption of the phenolate anion orthogonal
to the oxide surface due to the repulsion between the highly nu-
(often modified with alkali or alkaline earth metal ions and oxides
to infer basic characteristics) and (ii) catalysts with basic features.
Among catalysts with Brønsted-type acidic characteristics, the
most studied are (i) metal phosphates (e.g., AlPO4, RE phosphates,
BPO4) [2–14]; (ii) γ -Al2O3, either as such or doped with alkali or
alkaline earth metal ions [15–22]; and (iii) zeolites in acidic form
or sometimes also exchanged with metal ions [23–36]. Catalysts
with Brønsted-type acidic characteristics are very active, and high
conversion of the aromatic substrate can be reached at moderate
cleophilic O2 anions and the aromatic ring. This makes the para
position less accessible by adsorbed methanol, whereas ortho posi-
tions, being closer to the surface, are readily accessible. All models
described in the literature later refer to the Tanabe model and thus
are focused mainly on the mode of phenol adsorption. Methanol
is assumed to adsorb on the surface, develop a methoxy species,
or be activated by interaction with the acid–base pairs [45,46].
In most cases, however, no hypothesis regarding the role of the
reactions of methanol with regard to the mechanism of phenol
methylation is reported.
−
◦
temperatures (300–350 C). The main characteristic of these ma-
terials is the preferred formation of the product of O-methylation,
especially for less-acidic catalysts; for example, when phenol is the
In previous work [47], we reported that the transformation of
methanol is the main factor determining the performance in phe-
nol methylation when catalyzed by either basic or acidic catalysts.
Specifically, catalysts that activate methanol through an acid-type
Corresponding author at: Dipartimento di Chimica Industriale e dei Materiali,
Alma Mater Studiorum Università di Bologna, Viale Risorgimento 4, 40136 Bologna,
Italy.
*
E-mail address: fabrizio.cavani@unibo.it (F. Cavani).
0
021-9517/$ – see front matter © 2008 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2008.03.012

216
N. Ballarini et al. / Journal of Catalysis 256 (2008) 215–225
◦
mechanism led to the formation of anisole, which reacted fur-
ther to yield cresols through an intramolecular rearrangement.
At high temperature, the formation of polymethylated compounds
was preferred. In contrast, the catalysts with basic features dehy-
drogenated methanol to formaldehyde, which in turn formed other
intermediates (formate and methylformate) and then decomposed
to yield CO, CO2, H2, and CH4. Evidence was found indicating that
with these catalysts, dehydrogenation of methanol is essential to
obtain alkylation of phenol; in this case, the main product was o-
cresol [47].
In the present work, we report a study of the reaction of phe-
nol with compounds formed by methanol dehydrogenation (i.e.,
formaldehyde and methylformate), highlighting their role as pos-
sible methylating agents in the reaction of methanol with phenol.
We investigated the following catalysts: MgO, the main component
in alkaline earth metal oxide-based catalysts used in the General
Electric process for the synthesis of o-cresol and 2,6-xylenol [48],
magnesium–iron mixed oxide (sample Mg/Fe/O), and a commer-
cial H-mordenite. Fe oxide is claimed to be the main component
in the optimal catalyst for the selective ring methylation of phe-
nol and 1-naphthol [49,50], whereas Mg/Fe/O catalyzes the liquid-
phase and the gas-phase methylation of both m-cresol [51] and
phenol [52,53].
Blank reactivity tests performed at 390 C with the empty steel
reactor and with the phenol/formaldehyde mixture gave a phenol
conversion of 2% (with formation of o-cresol and 2,6-xylenol) and
an overall yield of light compounds (from formaldehyde decom-
◦
position) of 10–15% at 390 C. This indicates that the reactor wall
provided a small, but nonnegligible, contribution to the reactivity.
Indeed, it also was found that the catalytic effect of the wall was
greater when a steel reactor was used in which the protective pas-
sive coating had been removed due to the prolonged exposure to a
methanol-containing stream under dehydrogenating conditions. In
◦
this case, the blank test at 390 C led to a phenol conversion of 7%
and to a relevant decomposition of formaldehyde. When conduct-
ing the same tests in a glass reactor yielded a phenol conversion
of 1%, and a formaldehyde decomposition of 5%.
The reactor outflow was condensed in HPLC-grade acetone.
Gases (CO, CO2, H2, CH4) were analyzed by sampling the gaseous
stream with a syringe at the reactor outlet before bubbling into
acetone, and then injecting the sample into a GC 4300 Carlo Erba
gas chromatograph, equipped with a TCD and a Carbosieve SII
column. The GC oven temperature was programmed from 55 to
220 C at a heating rate of 10 C/min. Products condensed in ace-
tone were analyzed by gas chromatography, using a GC 6000 Carlo
Erba instrument equipped with a FID and a HP-5 column. The GC
oven temperature was programmed from 50 to 250 C at a heating
rate of 10 C/min. The same GC was also equipped with a CP-
◦
◦
◦
◦
2. Experimental
Poraplot Q column (FID detector) for the separation of methanol,
dimethylether, formaldehyde, methylformate, and formic acid; in
Three catalysts were investigated; MgO, magnesium–iron mixed
◦
oxide (sample Mg/Fe/O), and a commercial H-mordenite with a
Si/Al atomic ratio of 20, supplied by Süd-Chemie AG. The MgO and
Mg/Fe/O samples were prepared by precipitation from an aqueous
solution containing corresponding metal nitrates. Details regarding
the preparation and characterization of catalysts both before and
after reaction are available elsewhere [47]. Before activity testing,
this case, the GC oven was programmed from 50 to 250 C at a
◦
heating rate of 10 C/min.
Molar yields were calculated as follows (n˙ is the molar flow):
out
n˙
product
, where product represents
in
•
•
Yield of phenolic products =
n˙
phenol
◦
the catalysts were equilibrated in the reactant mixture at 390 C
o-cresol, p-cresol, 2,6-xylenol, anisole, and polyalkylated phe-
for 10 h, to obtain a stable catalytic performance.
nols.
out
n˙
product
Catalytic tests were carried out by vaporization of an aque-
ous phenol/formaldehyde solution containing 20.6 wt% formalde-
hyde, 0.4 wt% methanol, 17.5 wt% phenol (supplied by Sigma–
Aldrich, 99 + % purity), and the balance water in a N2 stream
Yield of light products from formaldehyde:
, where
in
n˙
formaldehyde
product represents CH3OH, CO, CO2, and CH4.
out
• Yield of H2 from formaldehyde: n˙in n˙
.
H
2
formaldehyde
(
N2 gas flow 20 Nml/min). The aqueous formalin solution was
prepared to contain the minimal amount of methanol, which is
usually added in commercial formalin solutions to limit formalde-
hyde polymerization. The reactor feed composition was as follows
Conversions were expressed as follows:
conversion of phenol or formaldehyde:
(in molar fractions): nitrogen, 89.3%; formaldehyde, 1.7% (consid-
in
out
n˙
− n˙
ering all the formaldehyde oligomers as being in the monomeric
form); methanol, 0.03%; phenol, 0.46%; and water, 8.5%.
phenol or formaldehyde
phenol or formaldehyde
.
in
n˙
Overall gas residence time was 2.7 s, at a GHSVphenol of 6.2 h 1
−
.
phenol or formaldehyde
Total pressure was 1 bar. These conditions resulted in the same
residence time as in tests on phenol/methanol reactivity [47], but
this implies a phenol/formaldehyde molar ratio (1/3.7) different
from that applied previously (1/10). The gas/vapor stream was fed
Selectivity to a compound was expressed as the ratio between the
corresponding yield and the reactant conversion.
IR measurements under static conditions were performed with
a Perkin–Elmer 1750 FT-IR spectrometer equipped with a DlaTGS
detector. For recording spectra, 25 scans were co-added at a res-
3
into a stainless steel reactor containing 1 cm of catalyst shaped
−
1
either in 30- to 60-mesh particles (catalyst weight: MgO 0.85 g,
Mg/Fe/O 0.95 g), or in 1/16-inch extrudates (H-mordenite 0.64 g).
Tests with the H-mordenite were carried out in a glass reactor to
minimize dehydrogenation due to the reactor wall.
Additional tests were carried out by vaporization of a phe-
nol/methylformate (molar ratio 1/1) solution in a N2 stream (N2
gas flow 20 Nml/min). The feed composition was as follows: N2,
olution of 2 cm . MgO samples were pressed into thin, self-
◦
supported wafers and activated in situ into the IR cell at 450 C
under vacuum (10
−
5
mbar) for 2 h, until the adsorbed water
disappeared. The adsorption procedure involved contacting the
vacuum-cleaned sample wafer with vapors of either the single
reactants (i.e., formaldehyde [obtained by heating in vacuum of
paraformaldehyde, supplied by Sigma–Aldrich] and phenol) or both
reactants in a different order of adsorption at room temperature,
followed by evacuation at increasing temperatures (i.e., 100, 200,
9
0.5%; methylformate (supplied by Carlo Erba Reagenti, 97% pu-
rity), 4.7%; and phenol, 4.7%. The overall gas residence time was
.7 s.
Finally, tests were carried out while feeding a phenol/methyl-
◦
2
300, 350, 400, and 450 C). IR spectra of o-cresol (supplied by
Fluka, >99,5 purity) and salicylaldehyde (supplied by Aldrich, 98%
purity) adsorbed on MgO also were collected for comparison. The
IR spectra reported here are difference absorbance spectra of the
MgO sample with and without the adsorbed molecules.
iodide solution (molar ratio 1/10) composed of N2, 89.0%; methyl-
iodide (supplied by Sigma–Aldrich), 10.0%; and phenol, 1.0%. The
overall gas residence time was 2.7 s.

N. Ballarini et al. / Journal of Catalysis 256 (2008) 215–225
217
3
. Results
and then formic acid completely decomposed to CO2 + H2, for
an overall stoichiometry (reaction (2)) (reported for mass balance
purposes only),
3
.1. Alkylation of phenol with formaldehyde
2H2CO + H2O → CH3OH + CO2 + H2.
(2)
In fact, the experimental yield of H2 was similar to that
Figs. 1 and 2 show the results of tests carried out with the MgO
of CO2. Minor amounts of CO formed either by dehydrogenation
of formaldehyde or by decomposition of HCOOH to CO + H2O; the
latter is more likely, because the former reaction would have led
and Mg/Fe/O catalysts, respectively. Here the conversion of phenol
and the molar selectivity to the phenolic products are plotted as
a function of the reaction temperature. The yields of CH3OH, CH4,
CO, CO2, and H2, obtained by the transformation of formaldehyde,
also are reported.
to an overall yield of H2 exceeding that of CO . For the same rea-
2
son, it can be inferred that the dehydrogenation of methanol to
formaldehyde also was not extensive.
Fig. 1 shows that with MgO, the principal product of the reac-
tion between formaldehyde and phenol was o-cresol, with minor
formation of p-cresol and xylenols, mainly the 2,6-dimethyl iso-
mer. The selectivity to o-cresol was very high at low temperature
and decreased slightly with increasing reaction temperature.
The main C-containing products of formaldehyde transforma-
tion were CO2 and methanol; the two compounds formed with
similar yields. The yields of CH4 and methylformate were <0.5%;
the yield of formic acid was nil. These results indicate that due to
the presence of a large amount of water in the feed (in the forma-
lin solution), formaldehyde gave formic acid and methanol through
the Cannizzaro reaction (reaction (1)),
The overall yield of light C-containing compounds derived from
formaldehyde at 390 C was 92%, because the remaining fraction
of formaldehyde was consumed in phenol alkylation. In fact, under
these conditions, the conversion of formaldehyde was total.
It is noteworthy that when tests of phenol methylation were
carried out with MgO and feeding a methanol/phenol mixture
◦
(
<
molar ratio 10/1) [47], the overall conversion of methanol was
◦
10% at 390 C; the conversion of methanol did not increase
when water was added to the feed. Therefore, the lower reactiv-
ity of methanol compared with formaldehyde explains why the
methanol generated by formaldehyde transformation (Fig. 1, bot-
tom) was not extensively converted to light decomposition com-
pounds.
The negligible formation of methylformate indicates that be-
cause of the large excess of water, the reaction between methanol
and formic acid to yield the ester was not favored, and that the
Tischenko dimerization of formaldehyde to yield the ester was ki-
netically unfavored with respect to the Cannizzaro disproportion-
2H2CO + H2O → CH3OH + HCOOH,
(1)
Fig. 1. Top: Conversion of phenol (2), molar selectivity to o-cresol ("), p-cresol (Q)
and xylenols (F) as a function of temperature. Bottom: Molar yield of CH3OH (!),
CO (2), CO2 (Q), CH4 (×) and H2 (F) as a function of temperature. Feed compo-
sition: N2 89.3%, formaldehyde 1.7%, phenol 0.46%, methanol 0.03% and water 8.5%.
Overall gas residence time 2.7 s, GHSVphenol 6.2 h ; total pressure 1 bar. Catalyst:
MgO.
Fig. 2. Top: Conversion of phenol (2), molar selectivity to o-cresol ("), p-cresol
(Q), xylenols (F) and salicylaldehyde (1) as a function of temperature. Bottom:
Molar yield of CH3OH (!), CO (2), CO2 (Q), CH4 (×) and H2 (F) as a function of
temperature. Reaction conditions as in Fig. 1. Catalyst: Mg/Fe/O.
−1

218
N. Ballarini et al. / Journal of Catalysis 256 (2008) 215–225
Fig. 3. Left: Conversion of phenol (2), molar selectivity to anisole (×), o-cresol ("), p-cresol (Q), xylenols (F), salicylaldehyde (1) and polyalkylated phenols (∗) as a function
of temperature. Right: Molar yield of CH3OH (!), CO (2) and CO2 (Q) as a function of temperature. Reaction conditions as in Fig. 1. Catalyst: H-mordenite.
ation. Indeed, methylformate might have quickly decomposed to
methanol + CO; in this case, however, the yield of CO should have
been equimolar to that of methanol.
on the amount of formaldehyde fed; therefore, additional H2 for-
mation deriving from a co-reactant (water in this case) may cause
it to exceed the theoretical limit for the stoichiometry of formalde-
hyde dehydrogenation.
Fig. 2 reports the catalytic performance of Mg/Fe/O. This cat-
alyst was more active than MgO, as it was in the reaction be-
tween phenol and methanol reported previously [47]. The phenolic
products were o-cresol and 2,6-xylenol; no anisole was formed.
The selectivity to p-cresol was very low, whereas this compound
was obtained with selectivity >2–3% with MgO; the same differ-
ence between the two catalysts was observed in the reaction be-
The overall conversion of formaldehyde was lower than that ob-
◦
tained with MgO: 70% at 390 C, considering also the fraction of
formaldehyde reacted with phenol. Mg/Fe/O was much more ac-
tive than MgO in methanol dehydrogenation to formaldehyde in
the absence of phenol, but the two catalysts had similar activity
when phenol also was present in the stream [47]. This finding was
interpreted by assuming competition between methanol and phe-
◦
tween phenol and methanol [47]. Below 350 C, salicyladehyde (2-
nol for adsorption over the Fe3 ions, which led to inhibition of
methanol dehydrogenation. In the present case, the lower conver-
sion of formaldehyde observed with Mg/Fe/O may be attributed
either to the inhibition effect due to the presence of phenol or to
the lower activity of this catalyst in the Cannizzaro reaction. With
both the MgO and Mg/Fe/O catalysts, no deactivation phenomena
+
hydroxybenzaldehyde) also formed; its selectivity was close to 30%
◦
◦
at 250 C and then decreased down to zero at 350 C, with a cor-
responding increase in the selectivity to o-cresol. This is the main
difference with respect to tests carried out with methanol [47];
in the latter case, salicylaldehyde was not obtained. In the case of
MgO, salicyladehyde did not form even in the reaction between
phenol and formaldehyde (Fig. 1); this may be due to the fact that
this catalyst was less active than Mg/Fe/O and gave nonnegligi-
◦
were observed at reaction temperatures of 390 C and lower.
Fig. 3 reports the catalytic performance of the H-mordenite
in phenol methylation with formaldehyde. The results differ from
those obtained with methanol [47]; in the latter case, at low tem-
perature, the main reaction product was anisole, whereas over
◦
ble conversion of phenol only at temperatures above 300 C. Under
these conditions, salicylaldehyde is likely rapidly transformed to o-
cresol.
◦
Concerning the side reactions occurring on formaldehyde, main
differences with respect to MgO were as follows: (i) an higher
yield of CO, which at low temperature formed in an amount com-
300 C, the products were o-cresol, p-cresol, and polyalkylated
phenols. With formaldehyde, the nature and amount of the prod-
ucts were similar to those obtained with MgO and Mg/Fe/O. At
low temperature, the prevailing products were o-cresol and salicy-
laldehyde, whereas anisole formed with selectivity of 10%. At high
temperature, the prevailing products were o-cresol and 2,6-xylenol,
with lesser amounts of p-cresol and polyalkylated phenols.
◦
parable to CO2; (ii) a much higher yield of H2, which at 390 C
approached the theoretical limit value of 100%, and (iii) a lower
◦
yield of methanol, which had a maximum at 300 C and became
◦
nil at 390 C. The decreased yield of methanol was accompanied by
increased yields of CO2 and H2. Also with this catalyst, the yields
of formic acid and methylformate were negligible.
The formation of CO can be explained by considering the con-
tribution of formaldehyde dehydrogenation (reaction (3)),
In terms of the formation of light compounds (Fig. 3, right),
the prevailing compounds were CO and methanol, which formed
in equimolar amounts at 250 C; no formation of methane or hy-
◦
drogen was seen. This finding suggests that formaldehyde yields
methylformate by dimerization, which is then further decomposed
to yield the two light compounds. The possible formation of formic
acid, further decomposed to CO and water, cannot be excluded.
H2CO → CO + H2.
(3)
However, the relevant yield of H2 can be explained only by consid-
ering a role of H2O in the WGS reaction (reaction (4)), catalyzed by
the Fe cation,
◦
At temperatures above 250 C, the yield of methanol decreased
whereas that of CO increased; in fact, methanol reacted to yield
anisole, alkylphenols, and alkylaromatics [47].
CO + H2O → CO2 + H2.
Another contribution to H2 formation may derive from meth-
anol reforming (reaction (5)),
(4)
Two other findings are worth mentioning. First, polyalkyl-
benzenes (mainly pentamethylbenzene and hexamethylbenzene),
which were obtained in large amounts from methanol [47], formed
with an overall yield <1% from formaldehyde. Furthermore, no
short-term deactivation phenomena were observed, whereas they
were evident from methanol. In the latter case, it can be suggested
that the deactivation was due to the formation of alkylaromatics,
precursors of coke formation. This indicates that with formalde-
CH3OH + H2O → CO2 + 3H2.
Clearly, a significant contribution of these reactions might cause
the yield of H2 to exceed 100%. In fact, the yield of H2 is dependent
(5)

N. Ballarini et al. / Journal of Catalysis 256 (2008) 215–225
219
Fig. 6. Conversion of phenol (2), molar selectivity to anisole (×), o-cresol (") and
2,6-xylenol (F) as a function of temperature. Reaction conditions as in Fig. 5. Cat-
alyst: Mg/Fe/O.
Fig. 4. Molar selectivity to salicylaldehyde (1), o-cresol (") and xylenols (F) as
a function of phenol conversion. Temperature 250 C; residence time was varied.
Other conditions as in Fig. 1. Catalyst: Mg/Fe/O.
◦
Fig. 5. Conversion of phenol (2), molar selectivity to anisole (×), o-cresol ("),
p-cresol (Q) and 2,6-xylenol (F) as a function of temperature. Feed composition:
N2 90.5%, methylformate 4.7% and phenol 4.7%. Overall gas residence time 2.7 s;
total pressure 1 bar. Catalyst: MgO.
Fig. 7. Conversion of phenol (2), molar selectivity to o-cresol ("), 2,6-xylenol (F),
p-cresol (Q), anisole (×), 2,4-xylenol (!) and polyalkylated phenols (∗) as a func-
tion of temperature. Reaction conditions as in Fig. 5. Catalyst: H-mordenite.
hyde, the formation of coke precursors was hindered, due either
to the low amount of methanol generated or to the presence of
water, which inhibits the formation of dimethylether.
lic products were o-cresol and 2,6-xylenol. With both catalysts, the
conversion of phenol was low; in fact, most methylformate decom-
posed, and only a minor fraction of it reacted with phenol.
Tests were carried out with varying residence times to obtain
Fig. 7 shows the catalytic performance obtained with the H-
mordenite catalyst. At low temperatures, the products of reaction
with phenol were anisole and cresols (mainly o-cresol). The same
distribution of products was obtained starting from methanol [47],
whereas that obtained from formaldehyde (Fig. 3) was different. An
increase in reaction temperature led to the formation of xylenols.
In terms of the formation of light compounds (results not re-
ported), with MgO, methylformate decomposed to CO (21% yield at
information on the reaction scheme. A very low residence time
◦
(
e.g., <0.1 s) and low temperature (250 C) were used to isolate
the intermediates, which were considered highly reactive. Fig. 4
plots the selectivity to the products as a function of phenol conver-
sion for tests carried out with the Mg/Fe/O catalyst. Salicylaldehyde
clearly was a primary product, because its selectivity, if extrapo-
lated to nil conversion, exceeded zero; however, it rapidly declined
when the conversion was increased. The decreased selectivity to
aldehyde led to an increase in selectivity to o-cresol; this evidently
supports the hypothesis that the latter compound forms by con-
secutive transformation of salicyladehyde. The further reaction of
o-cresol gave the formation of 2,6-xylenol, the selectivity of which,
however, was <10% under these experimental conditions.
◦
390 C) and methanol; the yield of methane was nil. The amount
◦
of H2 generated was relatively low (<1% yield at 390 C), suggest-
ing a low fraction of methanol dehydrogenated to formaldehyde.
◦
With the Mg/Fe/O catalyst, at 390 C, the yields of CO and CO
2
were similar (19 and 14%, respectively), the yield of methane was
3%, and that of H2 was 8%. With the H-mordenite, methylformate
◦
decomposed mainly to CH3OH and CO (30% yield at 390 C); the
3.2. Alkylation of phenol with methylformate
yields of CH4 and CO2 were 3%.
The results of catalytic tests performed by feeding phenol and
3.3. Alkylation of phenol with methyliodide
methylformate are reported in Figs. 5–7. These figures show the
conversion of phenol and the molar selectivity to the phenolic
products (i.e., o-cresol, 2,6-xylenol, and anisole). With MgO (Fig. 5),
the prevailing product was anisole, whereas in tests carried out
with either methanol [47] or formaldehyde (Fig. 1), this compound
was obtained with selectivity <10%. With Mg/Fe/O (Fig. 6), the
formation of anisole was low, as it was in tests carried out with
either methanol [47] or formaldehyde (Fig. 2); the main pheno-
To demonstrate the role of the methylating agent, tests were
carried out feeding methyliodide and phenol over the Mg/Fe/O
catalyst. Results are reported in Fig. 8. The conversion of phenol
was low, because of the inefficient activation of methyliodide due
to the absence of acidic sites. The nature and relative amounts
of the products obtained in phenol methylation were the same
as those obtained in tests involving feeding methanol and phenol

220
N. Ballarini et al. / Journal of Catalysis 256 (2008) 215–225
Fig. 8. Conversion of phenol (2), molar selectivity to anisole (×), o-cresol ("),
2
,6-xylenol (F), p-cresol (Q) and polyalkylated phenols (∗) as a function of tem-
perature. Feed composition: N2 89.0%, methyliodide 10.0% and phenol 1.0%. Overall
gas residence time 2.7 s; total pressure 1 bar. Catalyst: Mg/Fe/O.
Fig. 10. FT-IR spectra recorded after adsorption of formaldehyde at room tempera-
ture on MgO and evacuation, and after heating under vacuum at increasing tem-
peratures. Room temperature (a), 100 (b), 200 (c), 300 (d), 350 (e), 400 (f) and
◦
4
50 C (g).
−
1
MgO show the presence of carbonates (1200–1700 cm
region)
−1
◦
and –OH group (3753 cm ) also after pretreatment at 450 C.
The spectrum recorded after adsorption of phenol is similar to
that reported in the literature [54,55]. The stretching vibrations of
C–H in the aromatic ring can be seen in the spectral zone between
−
1
−1
−1
are at-
3
010 and 3070 cm . Bands at 1597 cm
and 1487 cm
−1
tributed to C–C ring vibrations; bands at 1170 (and 1074) cm , to
−1
C–H bending. The band at 1283 cm
is related to C–O stretching
for the phenolate species, evidence of dissociation of the hydroxy
group. For the spectrum of salicylaldehyde, bands in the spec-
−1
tral zone between 3000 and 3090 cm
can be attributed to the
stretching of the C–H of the aromatic ring. The bands at 2876 and
−1
2780 cm
are related to the stretching of the C–H in the alde-
hydic moiety, with the latter deriving from the Fermi resonance.
−
1
The band at 1647 cm
ing. Bands at 1606 and 1532 cm
ring-stretching vibrations. At 1412 and 1458 cm
due to C–H bending, the former attributed to aldehydic C–H. The
is characteristic of aldehydic C–O stretch-
−1
are related to the aromatic
−1
lie the bands
Fig. 9. FT-IR spectra recorded after adsorption at room temperature and evacuation
on MgO (a) of: formaldehyde (b), phenol (c), o-cresol (d) and salicylaldehyde (e).
−1
band at 1332 cm
is due to the stretching of C–O, whereas that
corresponds to the stretching of C–CHO. Additional
−1
with the H-mordenite [47]. The main product at low temperature
was anisole; an increase in the reaction temperature led to the for-
mation of cresols (with the ortho-isomer the prevailing one) and
polyalkylated phenols.
These findings indicate that when the formation of the alde-
hyde was not possible, the methylating species generated with the
basic catalyst yielded the same phenolic products obtained with
methanol and with acid catalysts. These tests demonstrate that the
nature of the products obtained in phenol methylation is a func-
tion of the type of methylating species generated.
at 1155 cm
bands in the spectrum of o-cresol, when compared with the corre-
sponding spectrum of phenol, are attributed to the –C–H vibrations
−
1
of the methyl group (between 2800 and 3000 cm ) and to the
−1
−1
bending of –CH3 at 1444 cm . The bands at 3400–3800 cm are
attributed to the interaction between the adsorbed molecules and
the –OH groups of MgO.
Fig. 10 shows the spectra recorded after adsorption of formalde-
hyde at room temperature and heating at increasing temperatures.
At low temperature, the typical frequencies of formaldehyde (i.e.,
−
1
−1
2841, 2748 cm
[ν CH] and 1383, 1363 cm
[δ CH]) are present;
◦
however, these peaks disappear at temperatures above 300 C, ev-
idence of complete desorption or decomposition of formaldehyde.
The peak at 1342 cm
3
.4. IR spectra recorded after adsorption and co-adsorption of reactants
with the MgO catalyst
−1
◦
in the spectra recorded at 100 and 200 C
is due to the formate species. The main result of this experiment
is that formaldehyde was weakly bound to the surface sites of
MgO and did not remain adsorbed at those temperatures at which
the reaction between phenol and formaldehyde occurred. This sug-
Fig. 9 plots the IR spectrum of MgO along with spectra of the
sample recorded after the sorption of vapors of formalin, phenol,
salicylaldehyde, and o-cresol at room temperature. The spectra of

N. Ballarini et al. / Journal of Catalysis 256 (2008) 215–225
221
Fig. 11. FT-IR spectra recorded after adsorption at room temperature on MgO of
phenol and then of formaldehyde and evacuation, and after heating under vacuum
at increasing temperatures. Room temperature (a), phenol and then formaldehyde,
Fig. 12. FT-IR spectra recorded after adsorption at room temperature on MgO of
formaldehyde and then of phenol and evacuation, and after heating under vacuum
at increasing temperatures. Room temperature (a), formaldehyde and then phenol,
◦
◦
room temperature (b), 100 (c), 200 (d), 300 (e), 350 (f), 400 (g), 450 C (h).
room temperature (b), 100 (c), 200 (d), 300 (e), 350 (f), 400 (g), 450 C (h).
gests that the latter reaction also may occur through a Rideal-type
mechanism involving adsorbed phenolate and gas-phase formalde-
hyde.
the aldehyde was not isolated at higher temperatures, probably be-
cause of its rapid transformation into o-cresol.
Fig. 11 plots the spectra recorded after preadsorption of phenol
at room temperature and subsequent adsorption of formaldehyde
and heating up to 450 C. Fig. 12 plots the spectra recorded by
4
. Discussion
◦
4
.1. Formaldehyde as the alkylating agent: A comparison with methanol
inverting the order of reactants adsorption on MgO. In the former
case, formaldehyde peaks were not evident, due to the stronger ad-
sorption of phenol over surface sites. In contrast, when phenol was
adsorbed after formaldehyde, the aromatic adsorbed on MgO, but
there was no substantial variation in the concentration of adsorbed
formaldehyde. These data can be interpreted based on previous
reports of the adsorption of these molecules over MgO (see also
the Section 4). In fact, phenol interacts strongly with Mg–O pairs
Our findings indicate that analogies exist between the gas-
phase methylation of phenol with methanol [47] and that with
formaldehyde. With basic catalysts, the performance in phenol
methylation was similar when either of the two reactants was fed.
On the other hand, the catalyst type (either basic or acidic) had
little influence on the nature of the products when formaldehyde
was used as the methylating agent, whereas the nature of products
obtained over the two catalysts was quite different when methanol
was the reagent [47]. This finding indicates that the methylat-
ing species in phenol methylation with methanol were not the
same when basic or acid catalysts were used but were the same
when the reaction was carried out with formaldehyde. Further-
more, it implies that not methanol itself, but one of the products
of methanol transformation (i.e., formaldehyde or methylformate),
plays the major role in the basic-catalyzed methylation of phenol.
These hypotheses are discussed more in detail below.
[54,55]. In contrast, a weaker interaction between the aldehydic
O atom and Mg develops with formaldehyde. Therefore, the ad-
sorption of formaldehyde left several Mg–O pairs on the surface
available for the adsorption of phenol; the latter did not displace
formaldehyde, because a prerequisite for phenol adsorption is the
accessibility of free Mg–O pairs. On the contrary, the preadsorption
of phenol led to saturation of the surface; because the interaction
of formaldehyde with the catalyst was not strong enough to com-
pete efficiently with the aromatic, only a very small amount of
formaldehyde was adsorbed.
When the temperature was increased, new bands appeared. In
The results obtained in phenol methylation with formaldehyde
over MgO and Mg/Fe/O can be explained by considering formalde-
hyde to be the true alkylating agent in the reaction of phenol
methylation with methanol and also with formaldehyde. According
to this hypothesis, the dehydrogenation of methanol to formalde-
hyde is the necessary requisite for methylation with basic catalysts.
A large proportion of formaldehyde is then decomposed to light
compounds, but a part of it reacts with phenol to yield o-cresol.
Therefore, the true molar ratio between formaldehyde and phenol
is much lower than the feed ratio (3.7/1) and also much lower
than the methanol/phenol feed ratio used in tests with methanol
(10/1) [47]. Indeed, the extent of methanol decomposition ob-
served in methanol/phenol reactivity tests was much lower than
that of formaldehyde; in fact, the yield of light compounds from
−1
the first case (Fig. 11), the bands at 1448 and 1250 cm
were due
to o-cresol formation, attributed to –CH3 bending and C–O stretch-
ing, respectively. This means that the low amount of formaldehyde
adsorbed reacted rapidly with the phenolate species. In the second
case (Fig. 12), the bands attributed to formaldehyde disappeared at
◦
temperatures above 300 C. Relevant features are the concomitant
−1
development of a shoulder at 1642 cm
448 (δ CH) cm , both corresponding to the most intense bands
(ν C=O) and a peak at
−1
1
of adsorbed salicylaldehyde (Fig. 9). These features disappeared at
◦
00 C, while bands attributed to o-cresol developed. These results
3
are in agreement with those of the catalytic tests and confirm that
◦
in the temperature range of 200–300 C, salicyladehyde, the inter-
mediate in the formation of o-cresol, can be isolated. In contrast,

222
N. Ballarini et al. / Journal of Catalysis 256 (2008) 215–225
methanol never exceeded 15% [47], whereas with formaldehyde,
the extent of formaldehyde transformation was almost total. This
difference is due to the higher reactivity of the aldehyde compared
with methanol, as well as the presence of a large amount of water,
which favors the disproportionation of formaldehyde. Despite this,
however, the conversion of phenol with formaldehyde was similar
to that with methanol. This indicates that formaldehyde is much
more reactive than methanol to the aromatic ring, due to the elec-
trophilic properties of the C atom in the carbonyl moiety.
to form a partial negative charge on the C atom. An aldol-type
condensation occurs, with C–C bond formation. The formation of
formaldehyde and an intermediate formate species in the side-
chain alkylation of alkylaromatics with methanol also has been
proposed by others [63–65].
In terms of the alkylation of activated aromatics with methanol,
to the best of our knowledge, the hypothesis that formaldehyde
may play a direct role was first cited in a patent [66], which
specified alkaline earth oxides as catalysts for the gas-phase re-
action between methanol and phenol. In the open literature, some
authors have suggested a possible role of aldehydes in the alkyla-
tion with alcohols, but no experimental proof has been provided
[67–74]. For instance, Gopinath et al. [67] identified formalde-
hyde, dioxymethylene, and formate by means of IR study when
feeding methanol over Cu/Co/Fe/O catalysts; however, they argued
that methanol dehydrogenation did not occur in the presence of
phenol, due to the competition of the two reactants for adsorp-
tion on the same sites. Radhe Shyam et al. [71] suggested that
in the vapor-phase alkylation of pyridine with methanol over a
Zn/Mn/Fe/O ferrospinel system, formaldehyde attacks intermedi-
ate dihydropyridine to yield 3-picoline, but they provided no ex-
perimental evidence supporting formaldehyde as the true alkylat-
ing agent. Klemm et al. [72] proposed that in the methylation of
1-naphthol catalyzed by alumina, the main methylating agents are
either the adsorbed methanol (coordinated to the acidic site) or
the methyl carbonium ion, but did not exclude the possibility that
either formaldehyde (formed by dehydrogenation of methanol) or
The nature of the phenolic products and their relative quanti-
ties also were quite similar when either of two reactants was used
with the basic catalysts. One minor difference concerns the for-
mation of anisole; in the reaction between phenol and methanol
over MgO, anisole was obtained with selectivity <10%, whereas
in the reaction with formaldehyde it did not form at all. Another
difference concerns the formation of salicylaldehyde as the reac-
tion intermediate; this was detected with selectivity close to 30%
at low temperature in the phenol/formaldehyde reaction with the
Mg/Fe/O catalyst and with the H-mordenite but was not observed
at all with these catalysts in the phenol/methanol reaction. This is
likely due to the fact that with methanol, the rate-determining step
of the reaction is the dehydrogenation of methanol to formalde-
hyde [47], which then rapidly attacks the aromatic ring; this im-
plies a very low concentration of the adsorbed phenolic interme-
diates under these conditions. When instead formaldehyde is fed
directly, the concentration of the phenolic intermediates is rele-
vant, and the latter may partially desorb into the gas phase.
An alternative hypothesis is that methanol is the true alkylat-
ing agent in the reaction of phenol methylation with methanol
and also with formaldehyde. In fact, a fraction of formaldehyde
is transformed into methanol and formic acid by the Cannizzaro
reaction. However, it is noteworthy that in the reaction between
phenol and formaldehyde, the concentration of methanol in the
reaction environment was low, in the best case equimolar to the
amount of CO2 formed. Therefore, in formaldehyde/phenol tests
the effective molar ratio between methanol and phenol is around
+
the hydroxymethyl carbonium ion ( CH2OH) might be the true
electrophilic agent [73].
Our catalytic tests of phenol methylation with formaldehyde
have demonstrated that the nature of the methylating species de-
termines the type of phenolic products. On the other hand, the
catalyst affects the type of methylating agent when methanol is
the reactant, whereas it has little effect on it with formaldehyde.
This finding also supports the hypothesis that in the presence
of a basic catalyst, the true active species in phenol methyla-
tion with methanol is not the alcohol. This would explain why at
low temperature, the main product of the reaction between phe-
nol and methanol with the H-mordenite catalyst is anisole [47],
which, conversely, is formed in low yield starting from formalde-
hyde. In fact, under these conditions, methanol is protonated by
the acidic catalyst and reacts either with another molecule of
methanol to yield dimethylether or with phenol to yield anisole.
When formaldehyde is the reactant, the electrophilic carbonyl
bond rapidly attacks the aromatic ring.
1, much lower than that in tests carried out by directly feeding
the methanol/phenol mixture [47]. If methanol were the true alky-
lating agent even in the formaldehyde/phenol tests, then a much
lower conversion of phenol than that obtained in methanol/phenol
tests would be expected, because the conversion of phenol is
greatly affected by the methanol/phenol feed ratio [47]. But at the
same temperature and contact time, the conversion of phenol was
greater with formaldehyde than with methanol. Finally, it is note-
worthy that much of the formaldehyde transformation to methanol
(
Figs. 1 and 2) was due to the excess of water fed. (Water is a core-
Therefore, our findings support the hypothesis that the main
role of the catalyst (either acidic or basic) is in generation of the
methylating species, whereas it has little influence on the reac-
tion pathway between the adsorbed phenol and the methylating
species or on the type of phenolic products finally obtained. If this
hypothesis is correct, then the use of a basic catalyst and a methy-
lating agent that does not dehydrogenate should yield the same
phenolic products obtained with the acidic catalyst and methanol,
that is, anisole at low temperature and cresols/polylakylphenols
at high temperature. In fact, with a basic catalyst, methyliodide
actant of the Cannizzaro reaction.) In the tests carried out with
methanol, no water was fed, and the amount of water generated
in the reaction was comparatively low [47].
All of the data obtained support the hypothesis that in the
gas-phase phenol alkylation with methanol catalyzed by basic cat-
alysts, the true alkylating agent is not methanol, but formaldehyde.
We next discuss this hypothesis in more detail.
4
.2. Formaldehyde as the alkylating agent in the basic-catalyzed
δ+
reaction between phenol and methanol
(the activation of which can occur only by generation of a CH3
species) gave the same distribution of products (Fig. 8) as obtained
with methanol/phenol and H-mordenite [47].
The IR spectra recorded after co-adsorption of phenol and
formaldehyde demonstrated the following:
In the literature, the direct alkylation of aromatics with alde-
hydes and ketones has been less studied than the corresponding
alkylation with alcohols, alkyl halides, or alkenes, because of the
tendency to yield oligomeric byproducts [56–60]. Lercher et al.
[
61,62] reported that in the basic-catalyzed methylation of toluene
1. Phenol developed a much stronger interaction with the MgO
surface than formaldehyde.
2. The reaction between phenol and formaldehyde required the
preadsorption and activation of phenol and the development
of the phenolate species; however, formaldehyde does not
to styrene with methanol, the necessary conditions for obtain-
ing side-chain alkylation are the dehydrogenation of methanol to
formaldehyde and the development of a strong interaction be-
tween the methyl group of toluene and the basic oxygen atoms

N. Ballarini et al. / Journal of Catalysis 256 (2008) 215–225
223
Fig. 13. Schematic representation of the main reactions occurring in the gas-phase methylation of phenol with methanol in basic catalysis. Dotted arrows represent reactions
kinetically less favored. a: favored in the presence of water (tests with formalin solution). b: favored in the presence of water and of Fe.
have to be in the adsorbed state to allow an electrophilic at-
tack on the activated aromatic ring.
hydrogenation of formaldehyde). In this case, a parallel primary
formation of o-cresol should also contribute to phenol conversion;
indeed, the data given in Fig. 4 do not exclude this possibility.
The transformation of salicylaldehyde to o-cresol also may occur
through reaction of the former with either methanol or formalde-
hyde, leading to reduction of the carbonyl to the methyl group and
to oxidation of the reductant to CO2.
3
. When the reaction was between adsorbed phenol and ad-
sorbed formaldehyde, salicylaldehyde could be isolated as the
reaction intermediate at moderate reaction temperatures (i.e.,
◦
below 300 C).
4
. Salicylaldehyde was the intermediate in o-cresol formation.
Therefore, the reaction between phenol and methanol in ba-
sic catalysis likely includes a reductive step by either methanol or
formaldehyde, for an overall stoichiometry,
Reactivity tests conducted with varying residence times (Fig. 4)
confirmed that salicylaldehyde was one primary product in the re-
action between phenol and formaldehyde, and that salicylaldehyde
was consecutively transformed to o-cresol. The possible presence
of other reactions leading to o-cresol cannot be excluded, however.
Salicyladehyde may form by rapid dehydrogenation of salicy-
lalcohol, which, by analogy with the hydroxymethylation reaction,
can be assumed to be the initial product of the reaction between
phenol and formaldehyde. If salicylalcohol is consecutively trans-
formed before it can desorb into the gas phase, then it cannot be
identified as a primary reaction product. Other reactions that may
occur on salicylalcohol include condensation with another phe-
nol molecule to yield (PhOH)–CH2–(PhOH), a possible precursor of
cresol [75], and the reduction of the hydroxymethyl to a methyl
group by formaldehyde or by hydrogen (the latter formed by de-
4.3. Methylformate as the alkylating agent: Comparison with methanol
Methylformate is a weak alkylating agent used in the Zerbe–
Jage process [76] for the synthesis of anisole from alkali metal
phenolate through the classical base-catalyzed Williamson ether
synthesis. Methylformate is generated in-situ by reaction between

224
N. Ballarini et al. / Journal of Catalysis 256 (2008) 215–225
◦
methanol and CO at 180 C and high pressure. The reaction be-
tween phenol and methylacetate catalyzed by Mg-zeolites also
mation of a more electrophilic species; formaldehyde then either
reacts rapidly with phenol to yield o-cresol or gives rise to the
parallel formation of methylformate. At low temperature, methyl-
formate may yield anisole by reaction with phenol, but at high
temperature, it decomposes to lighter compounds. Salicylaldehyde
is the intermediate product in the reaction between phenol and
formaldehyde and is the precursor for the formation of o-cresol.
Therefore, the nature of the products obtained by reaction be-
tween phenol and methanol is governed by the transformations
occurring on methanol, which are a function of catalyst type. This
also is demonstrated by the results of the reaction between phenol
and formaldehyde over the acid catalyst, with the formation of the
same products typically obtained in the basic-catalyzed methyla-
tion of phenol, and the reaction between phenol and methyliodide
over the basic catalyst, with the formation of the same products
obtained in the acid-catalyzed methylation of phenol.
◦
yields anisole as the main reaction product at T < 350 C, whereas
o-cresol and xylenols form at higher temperatures [77].
Results obtained in phenol methylation with methanol over
basic catalysts [47] demonstrated that with these catalysts, methyl-
formate is a product of methanol transformation that may rep-
resent the true alkylating species. Moreover, the yield of the
main products obtained by the high-temperature decomposition
of methylformate (i.e., CH4 and CO2) was significantly affected by
the presence of phenol. In fact, whereas in the absence of phe-
nol, the two compounds formed in almost equimolar amounts, the
presence of the aromatic led to a substantial decrease in methane
formation compared with carbon dioxide. This may be interpreted
as being due to the insertion of the methyl group on the aromatic
ring from intermediately formed methylformate, with the remain-
ing part of the molecule released in the form of carbon dioxide
and hydrogen.
Acknowledgments
Our findings for the reaction between phenol and methylfor-
mate indicate that a very low amount of anisole was formed with
the catalyst characterized by the stronger dehydrogenation activity
INSTM is acknowledged for providing doctoral grants to S.P. and
L.M. The authors thank Süd-Chemie for providing the H-mordenite
sample and Polynt SpA for providing the methanol-lean formalin
solution.
(Mg/Fe/O), whereas anisole was a predominant product of the re-
action with MgO. This suggests that methylformate can act directly
as the methylating agent, with the introduction of the methyl
group to yield anisole and the concomitant release of CO2. In fact,
almost no CH4 was generated from methylformate decomposition.
The C atom of the methyl group in the ester was not sufficiently
electrophilic to attack the aromatic ring, whereas it did react with
the O atom in the phenolate. Conversely, when methylformate was
significantly decomposed to CH3OH + CO and methanol was de-
hydrogenated to formaldehyde (i.e., with Mg/Fe/O), the prevailing
product was that of C-methylation, o-cresol.
Consequently, our findings demonstrate that with basic cata-
lysts and under mild reaction conditions (at which methylformate
was not significantly decomposed), the ester may act as a methy-
lating agent on phenol to yield anisole, but when methanol was
the reactant [47], methylformate formed only under those condi-
tions in which the alcohol was dehydrogenated to formaldehyde.
The latter then either formed methylformate (which, however, is
rapidly decomposed due to the high reaction temperature) or re-
acted with phenol to yield C-alkylated compounds. In tests involv-
ing feeding the formalin solution, methylformate formed only in
small amounts, because the Cannizzaro reaction was the preferred
path for formaldehyde transformation.
References
[1] H. Fiege, in: Ullmann’s Encyclopedia of Industrial Chemistry, vol. A8, 2002,
p. 25.
[
[
2] G. Sarala Devi, D. Giridhar, B.M. Reddy, J. Mol. Catal. 181 (2002) 173.
3] B.M. Reddy, G. Sarala Devi, P.M. Sreekanth, Res. Chem. Int. 28 (6) (2002) 595.
[4] F.M. Bautista, J.M. Campelo, A. Garcia, D. Luna, J.M. Marinas, A. Romero, J.A.
Navio, M. Macias, Appl. Catal. A 99 (1993) 161.
[5] V. Durgakumari, S. Narayanan, L. Guczi, Catal. Lett. 5 (1990) 377.
[6] R. Pierantozzi, A.F. Nordquist, Appl. Catal. 21 (1986) 263.
[7] F. Nozaki, I. Kimura, Bull. Chem. Soc. Jpn. 50 (3) (1977) 614.
[8] J.M. Campelo, A. Garcia, D. Luna, J.M. Marinas, M.S. Moreno, Stud. Surf. Sci.
Catal. 41 (1988) 249.
[9] L. Calzolari, F. Cavani, T. Monti, C. R. Acad. Sci. IIC 3 (2000) 533.
[10] F. Cavani, T. Monti, in: M.E. Ford (Ed.), Catalysis of Organic Reactions, Marcel
Dekker, New York, 2000, p. 123.
[11] F. Cavani, T. Monti, D. Paoli, Stud. Surf. Sci. Catal. 130 (2000) 2633.
[12] F. Cavani, T. Monti, P. Panseri, B. Castelli, V. Messori, WO Patent 74,485, as-
signed to Borregaard Italia, 2001.
[
13] L. Gilbert, M. Janin, A.M. Le Govic, P. Pommier, A. Aubry, in: J.R. Desmurs,
S. Ratton (Eds.), The Roots of Organic Development, Elsevier, Amsterdam, 1996,
p. 48.
[14] Y. Shiomi, Y. Nakamura, T. Manabe, S. Furusaki, M. Matsuda, M. Saito, US Patent
5,248,835, assigned to UBE Ind., 1993.
[
[
15] E. Santacesaria, D. Grasso, D. Gelosa, S. Carrà, Appl. Catal. 64 83, 101, 1990.
16] R. Tleimat-Manzalji, D. Bianchi, G.M. Pajonk, Appl. Catal. A 101 (1993) 339.
The distribution of products obtained with H-mordenite and
methylformate (Fig. 7) was quite similar to that obtained with
methanol [47] but different than that obtained with formaldehyde.
Therefore, at low temperature, either the methylformate activated
by protonation of the carbonylic O atom or the methanol ob-
tained by decomposition of methylformate, reacted with phenol to
[
17] Y. Fu, T. Baba, Y. Ono, Appl. Catal. A 166 (1998) 419 and 425.
18] Y. Fu, T. Baba, Y. Ono, Appl. Catal. A 178 (1998) 219.
[19] L. Kiwi-Minsker, S. Porchet, R. Doepper, A. Renken, Stud. Surf. Sci. Catal. 108
1997) 149.
20] S. Porchet, L. Kiwi-Minsker, R. Doepper, A. Renken, Chem. Eng. Sci. 51 (11)
[
(
[
[
(
1996) 2933.
◦
produce anisole. At high temperature (i.e., T > 350 C), methanol
21] L. Kiwi-Minsker, S. Porchet, P. Moeckli, R. Doepper, A. Renken, Stud. Surf. Sci.
Catal. 101 (1996) 171.
demonstrated ring-substitution to cresols and polyalkylated phe-
nols.
[22] M.C. Samolada, E. Grigoriadou, Z. Kiparissides, I.A. Vasalos, J. Catal. 152 (1995)
2.
5
[
[
[
23] P.D. Chantal, S. Kaliaguine, J.L. Grandmaison, Appl. Catal. 18 (1985) 133.
24] R. Pierantozzi, A.F. Nordquist, Appl. Catal. 21 (1986) 263.
25] S.C. Lee, S.W. Lee, K.S. Kim, T.J. Lee, D.H. Kim, J. Chang Kim, Catal. Today 44
(1998) 253.
5
. Conclusion
Fig. 13 summarizes the results obtained in the present study
[26] M. Marczewski, G. Perot, M. Guisnet, Stud. Surf. Sci. Catal. 41 (1988) 273.
and in our previous work on phenol methylation with methanol
over basic catalysts [47]. In the gas-phase methylation of phe-
nol with methanol, the main role of the catalyst (either basic or
acidic) is in generating the methylating agent. With the acidic cat-
alyst, the activation of methanol generates an electrophilic species
that reacts with phenol to produce anisole at low temperature,
whereas at high temperature, the preferred products of the reac-
tion are C-alkylated compounds. Conversely, with basic catalysts,
the dehydrogenation of methanol is a required step in the for-
[
[
27] M. Marczewski, J.-P. Bodibo, G. Perot, M. Guisnet, J. Mol. Catal. 50 (1989) 211.
28] L. Garcia, G. Giannetto, M.R. Goldwasser, M. Guisnet, P. Magnoux, Catal. Lett. 37
1996) 121.
(
[29] S. Balsamà, P. Beltrame, P.L. Beltrame, P. Carniti, L. Forni, G. Zuretti, Appl.
Catal. 13 (1984) 161.
[
30] P. Beltrame, P.L. Beltrame, P. Carniti, A. Castelli, L. Forni, Appl. Catal. 29 (1987)
3
27.
31] R.F. Parton, J.M. Jacobs, H. van Ooteghem, P.A. Jacobs, Stud. Surf. Sci. Catal. 46
1989) 211.
[32] S. Namba, T. Yashima, Y. Itaba, N. Hara, Stud. Surf. Sci. Catal. 5 (1980) 105.
[
(

N. Ballarini et al. / Journal of Catalysis 256 (2008) 215–225
225
[
[
[
[
33] Z.-H. Fu, Y. Ono, Catal. Lett. 21 (1993) 43.
[56] Z. Hou, T. Okuhara, Chem. Comm. (2001) 1686.
[57] A. Corma, H. Garcia, J. Miralles, Microporous Mesoporous Mater. 43 (2001)
161.
[58] J. Tateiwa, E. Hayama, T. Nishimura, S. Uemura, Chem. Lett. (1996) 59.
[59] J. Tateiwa, E. Hayama, T. Nishimura, S. Uemura, J. Chem. Soc. Perkin Trans. 1
(1997) 1923.
34] M. Renaud, P.D. Chantal, S. Kaliaguine, Can. J. Chem. Eng. 64 (1986) 787.
35] G. Moon, W. Böhringer, C.T. O’Connor, Catal. Today 97 (2004) 291.
36] M.D. Romero, G. Ovejero, A. Rodriguez, J.M. Gomez, I. Agueda, Ind. Eng. Chem.
Res. 43 (2004) 8194.
[
37] R.F. Parton, J.M. Jacobs, D.R. Huybrechts, P.A. Jacobs, Stud. Surf. Sci. Catal. 46
(
1989) 163.
[60] T. Nishimura, S. Ohtaka, A. Kimura, E. Hayama, Y. Haseba, H. Takeuchi, S. Ue-
mura, Appl. Catal. A 194 (2000) 415.
[
[
[
38] J.B. Moffat, Catal. Rev. Sci. Eng. 18 (1978) 199.
39] A. Corma, H. Garcia, J. Primo, J. Chem. Res. Synop. (1988) 40.
40] P.D. Chantal, S. Kaliaguine, J.L. Grandmaison, Stud. Surf. Sci. Catal. 19 (1984) 93.
41] L.B. Young, N.J. Skillman, US Patent 4,371,714, assigned to Mobil Oil Co, 1983.
[61] A.E. Palomares, G. Eder-Mirth, J.A. Lercher, J. Catal. 168 (1997) 442.
[62] A.E. Palomares, G. Eder-Mirth, M. Rep, J.A. Lercher, J. Catal. 180 (1998) 56.
[63] S.T. King, J.M. Garces, J. Catal. 104 (1987) 59.
[
[
42] P. Espeel, R. Parton, H. Toufar, J. Martens, W. Hölderich, P. Jacobs, in:
J. Weitkamp, L. Puppe (Eds.), Catalysis and Zeolites: Fundamentals and Appli-
cations, Springer-Verlag, Berlin, 1999, p. 377, chap. 6.
[64] H. Itoh, A. Miyamoto, Y. Murakami, J. Catal. 64 (1980) 284.
[65] G. Madhavi, S.J. Kulkarni, K.V.V.S.B.S.R. Murthy, V. Viswanathan, K.V. Raghavan,
Appl. Catal. A 246 (2003) 265.
[
43] K. Tanabe, T. Nishizaki, in: G.C. Bondet al. (Ed.), Proceed. 6th ICC, The Chemical
Society, London, 1977, p. 863.
[66] S.B. Hamilton, US Patent 3,446,856, assigned to Schenectady, 1969.
[67] T. Mathew, M. Vijayaraj, S. Pai, B.B. Tope, S.G. Hegde, B.S. Rao, C.S. Gopinath,
J. Catal. 227 (2004) 175.
[68] S. Sato, K. Koizumi, F. Nozaki, J. Catal. 178 (1998) 264.
[69] S. Sato, R. Takahashi, T. Sodesawa, K. Matsumoto, Y. Kamimura, J. Catal. 184
(1999) 180.
[70] A.H. Padmasri, A. Venugopal, V. Durgakumari, K.S. Rama Rao, P. Kanta Rao,
J. Mol. Catal. A 188 (2002) 255.
[71] A. Radhe Shyam, R. Dwivedi, V.S. Reddy, K.V.R. Chary, R. Prasad, Green Chem. 4
(2002) 558.
[72] L.H. Klemm, J. Shabtai, D.R. Taylor, J. Org. Chem. 33 (1968) 1480.
[73] E.I. Heiba, P.S. Landis, J. Catal. 3 (1964) 471.
[74] A.R. Gandhe, J.B. Fernandes, J. Mol. Catal. A 226 (2005) 171.
[75] I.B. Borodina, E.B. Pomakhina, Ts.M. Ramishvili, O.A. Ponomareva, A.I. Rebrov,
I.I. Ivanova, Russ. J. Phys. Chem. 80 (2006) 892.
[76] C. Zerbe, F. Jage, Brennstoff-Chem. 5 (16) (1935) 88.
[77] K. Shanmugapriya, S. Saravanamurugan, M. Palanichamy, B. Arabindoo, V. Mu-
rugesan, J. Mol. Catal. A 223 (2004) 177.
[
[
[
44] K. Tanabe, Stud. Surf. Sci. Catal. 20 (1985) 1.
45] K. Sreekumar, S. Sugunan, J. Mol. Catal. A 185 (2002) 259.
46] T. Tsai, F. Wang, Catal. Lett. 73 (2001) 167.
[
47] N. Ballarini, F. Cavani, L. Maselli, A. Montaletti, S. Passeri, D. Scagliarini, C. Flego,
C. Perego, J. Catal. 251 (2007) 423.
[48] US Patent 4,933,509, assigned to General Electric, 1989.
[49] H. Grabowska, W. Kaczmarczyk, J. Wrzyszcz, Appl. Catal. 47 (1989) 351.
[50] H. Grabowska, W. Mista, J. Trawczynski, J. Wrzyszcz, M. Zawadzki, Appl. Catal.
A 220 (2001) 207.
[
51] F. Cavani, L. Maselli, D. Scagliarini, C. Flego, C. Perego, Stud. Surf. Sci. Catal. 155
2005) 167.
(
[52] S. Velu, C.S. Swamy, Appl. Catal. A 162 (1997) 81.
[53] US Patent 6,897,175, assigned to General Electric, 2005.
[54] R.N. Spitz, J.E. Barton, M.A. Barteau, R.H. Staley, A.W. Sleight, J. Phys. Chem.
B 90 (1986) 4067.
[55] D.R. Taylor, K.H. Ludlum, J. Phys. Chem. 76 (1972) 2882.