Alkylation of toluene over basic catalysts - Key requirements for side chain alkylation

Article abstract of DOI:10.1006/jcat.1998.2253

In situ infrared spectroscopy was used to study sorption and reaction of toluene and methanol over various basic catalysts (MgO, hydrotalcites, and basic zeolites). The size of the metal cations controls the preference of sorbing methanol or toluene; i.e., the larger the metal cation, the higher the preference for toluene. The key requirements for a good catalyst for side-chain alkylation are (i) sufficient base strength to dehydrogenate methanol to formaldehyde, (ii) stabilization of sorbed toluene and polarization of its methyl group, and (iii) balanced sorption stoichiometry of the two reactants. The formation of the carbon-carbon bond mechanistically resembles an aldol condensation.

Full text of DOI:10.1006/jcat.1998.2253

JOURNAL OF CATALYSIS 180, 56–65 (1998)
ARTICLE NO. CA982253
Alkylation of Toluene over Basic Catalysts—Key Requirements
for Side Chain Alkylation
1
A. E. Palomares, G. Eder-Mirth, M. Rep, and J. A. Lercher
Department of Chemical Technology, University of Twente, P.O. Box 217, 7500 AE-Enschede, The Netherlands
Received May 5, 1998; revised July 21, 1998; accepted July 21, 1998
to formaldehyde and stabilizing the formaldehyde on the
In situ infrared spectroscopy was used to study sorption and re- catalyst (8, 16, 17); (ii) spatial constraints as found within
action of toluene and methanol over various basic catalysts (MgO, zeolite poresand inhibition oftoluene rotation in the zeolite
hydrotalcites, and basic zeolites). The size of the metal cations con-
trols the preference of sorbing methanol or toluene; i.e., the larger
the metal cation, the higher the preference for toluene. The key re-
quirements for a good catalyst for side-chain alkylation are (i) suf-
cavity (13, 17–19). It has also been suggested that specific
steric configurations of acidic and basic sites are crucial for
the formation of the required reaction complexes on the ze-
olite surface (17, 20–27). The base sites activate the carbon
ficient base strength to dehydrogenate methanol to formaldehyde,
atom of the side chain of toluene and the acid sites adsorb
(ii) stabilization of sorbed toluene and polarization of its methyl
and stabilize the toluene (21). Thus, the active center is an
assembly of acid and base sites in a cooperative action. In
contrast to these quite complex requirements, Giordano
group, and (iii) balanced sorption stoichiometry of the two reac-
tants. The formation of the carbon–carbon bond mechanistically
resembles an aldol condensation.
�c 1998 Academic Press
(
6) suggests that the primary factor governing zeolite ac-
tivity and selectivity is the overall acid–base strength, ex-
pressed as the averaged Sanderson electronegativity of the
oxide.
1
. INTRODUCTION
Because of its theoretical potential as a novel route to
In a previous communication (28), we have reported an
styrene, alkylation of toluene with methanol over basic ma- in situ IR spectroscopic study of the surface chemistry on
terials has been studied frequently over the past years (1–6). different alkali-exchanged X zeolites during toluene alky-
Despite these efforts, a consensus has not been reached so lation. We concluded that in addition to requirements listed
far concerning the mechanism of the reaction and the role above, the balanced sorption stoichiometry of the two re-
of the catalyst in directing the selectivity toward side chain actants and a strong polarization of the methyl group of
or ring alkylated products. It is known that basic materials toluene are indispensable to catalyze side chain alkylation
+
+
such as Cs - and Rb -exchanged X zeolites catalyze the of toluene.
side chain alkylation of toluene with methanol to styrene,
The purpose of the present contribution is to compare
which subsequently can be hydrogenated to ethylbenzene the surface chemistry of other basic materials with vary-
+
(
7–9). Acid materials, i.e., H–Y, H–ZSM5, etc., selectively ing activity (Cs -solid-state-exchanged Y zeolite, MgO, and
catalyze ring alkylation to xylenes (10–12). While the rates hydrotalcites with various Al/(Al + Mg) ratios) with that
for ring alkylation are quite high, the rates and yields for of alkali-exchanged X zeolites and to show that high base
side chain alkylation are still unsatisfactory, especially from strength alone is insufficient for good catalytic properties.
the point of methanol utilization. A recent study indicated, It is our goal to outline the requirements of the working
however, that alkali metal (hydr)oxides supported in zeo- catalysts during toluene alkylation. In situ IR spectroscopy
lite strongly improve the yield of the side chain alkylation and kineticmeasurementsare used to showhowchemisorp-
products (13). Also, the addition of boron (14) or zinc ox- tion and stabilization of the reactants at the surface of the
ides (15) has a promoting effect on the side chain alkylation. different basic materials relate to their catalytic properties.
Two factors have been suggested to be crucial to selec-
tive catalysis of side chain alkylation: (i) the presence of
active (basic) sites capable of dehydrogenating methanol
2. EXPERIMENTAL
2
.1. Materials
+
+
+
+
1
Ion- (Cs , Rb , K and Na -) exchanged X zeolites were
Current address: Department of Chemical and Nuclear Engineer-
prepared from Na–X (obtained from Fluka) by conven-
tional cation-exchange procedures using 0.025 M aqueous
ing, Universidad Politecnica Valencia, Avda. de los Naranjos s/n, 46022
Valencia, Spain.
56
0021-9517/98 $25.00
Copyright �c 1998 by Academic Press
All rights of reproduction in any form reserved.

SELECTIVE ALKYLATION OF TOLUENE OVER BASIC MATERIALS
57
solutions of potassium oxalate, rubidium acetate, or cesium 122A). The adsorption isotherms were determined gravi-
acetate at 353 K. The exchange procedure and the charac- metrically simultaneously with the differential enthalpies
terization of the samples are described in a previous paper of adsorption. The accuracy of the calorimetric measure-
(
28). Cesium-solid-state-exchanged zeolite Y was obtained ments was � 2 kJ/mol.
from Prof. Dr. H. G. Karge of the Fritz Haber Institut der
Max Planck Gesellschaft (Berlin, Germany). The Si/Al
ratio of the zeolite was 2.6 and 100% Cs exchange was re-
+
2.3. Reaction Studies
ached. Hydrotalcites with Al/(Al + Mg) ratios of (Ht 0.33)
For the in situ reaction studies, an IR cell with the char-
and (Ht 0.2) and MgO with a specific surface area of acteristics of a well-stirred continuously operating tank re-
2
00
1
3
2
00 m /g were obtained from Prof. A. Corma of the In-
actor (volume = 1.5 cm ) equipped with
gas inlet and
16
stituto de Tecnologia Quimica of Valencia (Spain).
outlet tubings and CaF2 windows was used (30). To char-
acterize the species sorbed in the material during the cata-
lytic reaction, time-resolved IR spectra were recorded at
various reaction temperatures from 373 to 723 K as the ac-
2
.2. Sorption and Coadsorption Experiments
IR spectra were obtained on a Bruker IFS88 FTIR tivated material was contacted with the reactant stream.
�
1
spectrometer (4 cm spectral resolution). For the sorption The feed gas composition had a partial pressure ratio of
experiments, a stainless steel cell with CaF2 windows that toluene/methanol = 45/15 (expressed in mbar) and He was
�
6
could be evacuated to pressures below 10 mbar was added to 1 bar. The overall flow was 3.5 ml/min. This cor-
4
used (for details see Ref. (29)). Before the adsorption was responds to a W/F ratio of 1.8 � 10 g of catalyst � s/mol of
started, the samples were activated in situ under vacuum toluene + methanol. Simultaneously, samples of the efflu-
at 773 K for 6 h. The activated sample wafer was then ent gas stream were collected in sample loops of a multiport
contacted with the appropriate pressure of methanol or valve and subsequently analyzed by gas chromatography
�
4
toluene (between 10 and 1 mbar) at constant temperature (HP5890 II, capillary column DBWAX 30m, FID, He as
T = 308 K) until adsorption–desorption equilibrium was carrier gas).
achieved (monitored by time-resolved IR spectroscopy). Because of the low reactivity of the catalysts, the small
(
For the coadsorption experiments, the catalyst was first weight permissible for an in situ infrared experiment, and
equilibrated with one reactant at a partial pressure of the minimum flow rates for a good backmixing, only low
�
3
5
� 10 mbar. Subsequently, the equilibrated material conversions and yields could be achieved in the IR reactor.
� 3
was exposed to 5 � 10 mbar of the second reactant until Therefore, parallel experiments were performed replacing
adsorption–desorption equilibrium was reached (while the IR reactor with a tubular quartz reactor (4 mm inner
keeping the partial pressure of the first reactant constant). diameter) in order to be able to directly compare the data
Then, the sample was evacuated for one hour at 308 K. The to those reported in the literature (7, 9, 13, 31). In these
sorbed molecules remaining on the catalyst were removed experiments, catalyst powders were loaded into the fixed
by temperature-programmed desorption (rate 10 K/min, bed quartz reactor and heated in situ at 723 K in flow-
up to 773 K) while the desorbing species were analyzed by ing He for 2 h before cooling to the reaction temperature
mass spectroscopy.
(698 K). A liquid mixture of toluene and methanol at a
The gravimetric and calorimetric measurements were 3 : 1 molar ratio was then pumped into the reactor system.
performed in a modified SETARAM TG-DSC 111 instru- The reactants were vaporized (partial pressure ratio 45/15
ment. The adsorption capacity for toluene and methanol (expressed in mbar)) and then mixed with flowing He to
�
3
5
was determined at 5 � 10 mbar under the same condi- obtain a constant reactor space time. A W/F of 1.2 � 10 g
tions as used for the infrared measurements.
For a typical experiment, approximately 15 mg of the reaction was tested at 698 K. The yield was calculated re-
of catalyst � s/mol toluene + methanol was chosen and the
samples was charged into the quartz sample holder of the ferred to the methanol feed (mol styrene + mol ethylben-
�
6
balance and activated byheatingin vacuum (p < 10 mbar) zene in the outlet per mol of methanol in the feed) and
with a temperature increment of 10 K/min to 873 K and to the aromatic feed (mol styrene + mol ethylbenzene in
holding that temperature for 1 h. After cooling to 303 K, the outlet per mol of toluene in the feed). The selectiv-
toluene was discontinuously dosed into the closed system ity was defined as moles of styrene + ethylbenzene formed
and equilibrated with the surface. The equilibration was per mole of toluene reacted (selectivity to side chain alky-
followed by means of heat flow and weight changes. Re- lation) or moles of xylenes formed per mole of toluene re-
versibility was checked through desorption of the alkanes acted (selectivity to ring alkylation). Other products de-
induced by evacuation of the system at several points in tected were formaldehyde, dimethyl ether, CO, and H2O.
one experiment. The experiments were carried out in the Note that rates and yields obtained in the latter reactor are
�
3
pressure range from 10 to 13 mbar. The pressure was in excellent agreement with the values reported in the lit-
recorded with a BARATRON pressure transducer (type erature (Table 1), while the selectivities between the two

5
8
PALOMARES ET AL.
TABLE 1
groups in interaction with the sorbate. The C–H stretch-
�
1
ing vibration bands appeared between 2917 and 2955 cm
Yields Reported in the Literature for the Side Chain Alkylation
�
1
and 2771 and 2839 cm (asymmetric and symmetric vibra-
tion), the C–H deformation vibration bands between 1451
of Toluene with Methanol on Cs- and Rb–X (T = 698 K)
�
1
%
Yield
% Yield
and 1480 cm , and the C–O–H deformation vibration be-
�
1
(
mol styrene +
(mol styrene +
ethylbenzene/
mol of toluene)
tween 1400 and 1420 cm . With MgO and hydrotalcite
two peaks assigned to C–O stretching vibrations were ob-
ethylbenzene/
mol methanol)
Reference
Our work
W/F (g �s/mol)
�
1
served at 1098–1105 and 1045–1065 cm . Note that these
bands could not be seen with the zeolite samples due to the
strong absorption of their Si–O vibrations. All these ob-
served bands for methanol adsorbed on zeolites and metal
oxides agree well with results obtained by other research
groups (25, 31–33).
5
1.2 � 10
7.6
10
2.5
5
[
[
7]
9]
1.8 � 10
5
6 � 10
10–20
13
5
[
[
14]
13]
1.1 � 10
5
1 � 10
0.4–2.75
The temperature of the desorption maximum of
reactors used in the present study (i.e., the IR reactor and methanol during t.p.d. of methanol decreased in the order
the tubular reactor) were almost identical (Table 2).
MgO, hydrotalcites > Na–X > Rb–X,Cs–X > Cs–Y. With
MgO and hydrotalcites the interaction between methanol
and the surface was so strong that even at temperatures
higher than 773K some methanol-derived compounds re-
mained adsorbed.
3
. RESULTS
3
.1. Adsorption of the Individual Reactants
The bands of the IR spectra of adsorbed toluene are
The IR spectra of methanol sorbed on different samples compiled in Table 4. For toluene sorbed on the zeolitic
are shown in Fig. 1. Table 3 compiles the wavenumbers of materials only minor differences in the wavenumbers of
the bands of methanol sorbed on the various samples. All the C–C and the ring deformation vibrations (1598, 1494,
�
1
the spectra shown in this paper are difference spectra. In and 1465 cm ) were observed. The C–H stretching vi-
this mode of presentation, the bands increasing in intensity bration bands of the aromatic ring (3058–3048 and 3024–
�
�
1
after exposing the activated material to the sorbate point 3019 cm ) and the methyl group (2921–2912 and 2874–
1
upward; the bands decreasing in intensity point downward. 2857 cm ) slightly shifted to lower wavenumbers as the
�
3
After equilibration at 5 � 10 mbar, the stretching vi- size of the cation increased (25). The characteristic bands
bration of the hydrogen-bonded OH group of methanol of the out of plane vibrations were observed between 1700
�
1
� 1
and 2000 cm (34), and were especially well resolved for
appeared as a broad band (between 3500 and 3200 cm
)
with the alkali-exchanged zeolites, and as a combination of toluene sorbed on Cs–Y. With the hydrotalcites and MgO
�
1
several broad bands (between 3600 and 2900 cm ) with hy- the concentrations of toluene sorbed were small compared
drotalcitesand MgO. With MgO, additionally, two relatively to the other samples studied. A negative band at 3725–
�
1
� 1
� 1
narrow bands were observed at 3475 and 3618 cm . With 3747 cm and a positive band between 3670 and 3700 cm
the nonzeoliticmaterialsa negative peak appeared between were observed in the difference spectra, which varied in par-
3
�
1
734–3760 cm characteristic of oxide surface hydroxyl allel in intensity. During t.p.d. of toluene the temperature
TABLE 2
Comparison between the Selectivity (mol %) Obtained in the Side Chain Alkylation of Toluene with Methanol
4
in the IR Reactor (W/F = 1.8 � 10 g � s/mol Toluene + Methanol) and in the Tubular Quartz Reactor (W/F =
5
1
.2 � 10 g � s/mol Toluene + Methanol, Toluene/Methanol = 3) at 698 K over Different Materials
Selectivity to side
chain alkylation
Selectivity to ring
alkylation
Selectivity to side-
chain alkylation
Selectivity to ring
alkylation
5
5
4
4
(
W/F = 1.2 � 10 g �
(W/F = 1.2 � 10 g �
(W/F = 1.8 � 10 g �
(W/F = 1.8 � 10 g �
Sample
s/mol)
s/mol)
s/mol)
s/mol)
Hydrotalcites
MgO
Cs–X
Rb–X
K–X
traces
traces
78
65
46
traces
traces
82
70
41
14
22
43
97
—
10
20
39
95
83
Na–X
Cs–Y
3
—
5
17

SELECTIVE ALKYLATION OF TOLUENE OVER BASIC MATERIALS
TABLE 4
59
IR Bands of Toluene Sorbed on Na–X, Cs–X, Cs–Y, Hydrotalcite,
and MgO at 5 � 10� mbar and 308 K
3
�
� 1
Sample
Na–X
�C-H aromatic [cm 1]
�C-H aliphatic [cm ]
3055, 3024
3048, 3021
2921, 2859
2916, 2857
Cs–X
Cs–Y
Ht 0.2, Ht 0.33
MgO
3048, 3021
3083, 3053, 3024
3082, 3052, 3025
2912, 2874, 2858
2950, 2926, 2871
2958, 2926, 2873
and 2869–2859 cm 1, and C–C ring vibrations at 1495 and
1
�
�
1
+
598 cm ) were the main spectral features with the Cs
exchanged zeolites. On the other hand, the bands charac-
teristic of sorbed methanol (the O-H stretching vibration
�
1
at 2900–3600 cm and the C-H stretching vibrations of the
�
1
methyl group at 2815–2771 and 2945–2917 cm ) were the
most important ones after coadsorption of methanol and
toluene on hydrotalcite and MgO. Indeed, the quantita-
tive evaluation of the IR spectra showed that the ratio of
FIG. 1. Difference IR spectra of methanol sorbed on MgO, hydrotal- toluene : methanol in the sorbed phase was 1 : 30 with MgO,
cite, Cs–X, and Cs–Y ( p = 5 � 10� mbar, T = 308 K).
3
1 : 19 with hydrotalcite, 2 : 1 with Cs–X, and 3 : 1 with Cs–Y.
T.p.d. of coadsorbed toluene and methanol from the vari-
of the desorption rate maximum increased in the order ous materials (Fig. 3) also indicated that methanol was the
MgO < hydrotalcite < Na–X < Rb,Cs–X < Cs–Y.
main desorbing species at every temperature with hydro-
The adsorption isotherms of toluene on Na–X and Cs–X talcite and MgO, while it was toluene with Cs–X and Cs–Y.
show that the amount of toluene sorbed on Cs–X is larger
than on Na–X. The heats of adsorption for Na–X and Cs–X 3.3. Reaction Studies
(
extrapolated to zero loading) are 76 kJ/mol and 92 kJ/mol
With zeolitic materials reaction products were not de-
tected in the gas phase below 523 K. Above 523 K, the
respectively, for MgO it is much lower (50 kJ/mol).
3
.2. Coadsorption of the Reactants
The IR spectra recorded after coadsorption of the two
reactants are compiled in Fig. 2. Only bands characteris-
tic for sorbed toluene and methanol were observed. The
bands typical for sorbed toluene (C-H stretching vibrations
�
1
of the aromatic ring at 3055–3045 and 3025–3020 cm , C-H
stretching vibrations of the methyl group at 2918–2916
TABLE 3
IR Bands of Methanol Sorbed on Na–X, Cs–X, Cs–Y, Hydro-
�
3
talcite, and MgO at 5 � 10 mbar and 308 K
�O-H [cm 1]
�
�C-H methyl [cm ]
� 1
Sample
Na–X
Cs–X
Cs–Y
3336
3244
3375
2956, 2839
2941, 2820
2940, 2829
2940, 2817
Ht 0.33
3600–2900
3
3
739 (negative band)
3600–2900
740 (negative band)
Ht 0.2
MgO
2941, 2815
2939, 2917,
FIG. 2. Difference IR spectra and quantitative evaluation of toluene
3600–2900, 3617, 3481
758 (negative band)
methanol occupation after coadsorption of toluene and methanol on MgO,
3
2824, 2795, 2772
�
3
hydrotalcite, Cs–X, and Cs–Y ( p = 5 � 10 mbar, T = 308 K).

6
0
PALOMARES ET AL.

SELECTIVE ALKYLATION OF TOLUENE OVER BASIC MATERIALS
61
FIG. 6. Difference IR spectra of a hydrotalcite with an Al/Al + Mg =
0
.2 during the reaction of toluene and methanol as a function of the reac-
tion temperature (T = 423–743 K).
FIG. 4. Products formed during temperature programmed reaction
of toluene and methanol over hydrotalcite with an Al/Al + Mg ratio of
Cs–X, at low temperatures, the main bands were charac-
teristic for sorbed toluene, while a new band was observed
0
.2. (47 mg of sample, 10 K/min, 6 mbar toluene, 3 mbar methanol).
�
1
around 1660–1680 cm . Thisband reached itsmaximum in-
tensity around 433 K and disappeared completely at 523 K.
At higher temperatures another new band (as a shoulder)
formation of formaldehyde was observed. Products of the
alkylation of toluene with methanol appeared between 573
and 593 K. Depending on the zeolite, side-chain-alkylated
� 1
was observed with Cs–X at 1610 cm (as was previously re-
ported (28) for Na–X). The presence of this band is difficult
to unequivocally establish due to the overlap with the band
(
(
Cs–X) or ring-alkylated products (Cs–Y) were formed
Table 2). The situation was different with the hydrotalcites
� 1
of sorbed toluene at 1600 cm . For clarification, Cs–X was
and MgO. In this case, side-chain-alkylated products were
only observed in very low concentrations, the main reaction
products being formaldehyde, dimethylether, CO, H2 and
water up to 673 K, H2 and CO between 673 and 823 K, and
CO2 and H2O at higher temperatures (see Fig. 4).
The in situ IR spectra obtained in the temperature pro-
grammed reaction of toluene and methanol over Cs–X and
hydrotalcite are shown in Figs. 5 and 6, respectively. With
exposed only to methanol. It was seen (Fig. 7, line c) that
� 1
in this case also a band appeared around 1610 cm . Note
� 1
that the shoulder at 1610 cm was observed in the spectra
during reaction over Cs–X zeolite but not during reaction
over Cs–Y (Fig. 7, line b).
With hydrotalcite (Fig. 6) and MgO, methanol was the
main adsorbent and, thus, nearly all bands were related
FIG. 5. Difference IR spectra of Cs–X during the reaction of toluene
FIG. 7. Difference IR spectra at 643 K of (a) methanol and toluene
and methanol as a function of the reaction temperature (T = 423–743 K). on Cs–X, (b) methanol and toluene on Cs–Y, (c) methanol on Cs–X.

6
2
PALOMARES ET AL.
to methanol and its reaction products. The most intense vibrations at 2939 and 2824 cm 1, appearing as shoulders
�
�
1
band appeared at 1602 cm , together with two other in Fig. 1, are attributed to a methyl group vibrating rather
bands at 1380 and 1369 cm (35). These bands (especially freely, probably related to the narrow O–H bands observed
�
1
�
1
that at 1602 cm ) appeared already at very low tempera- at higher wavenumber.
tures (333 K) and increased in intensity until 673 K. Then,
The results from the t.p.d. of methanol show that it is
their intensity decreased, but the bands were still visible at bound strongly when the radius of the accessible metal
�
1
7
1
73 K. A new band was also observed at 1089 cm and cation is small (MgO, hydrotalcite, and Na–X) and the inter-
�
1
118 cm (starting from 473 K), reaching its maximum in- action becomes very weak when the radius is large (Cs–X,
�
1
tensity at 673 K, when the bands at 1109 and 1060 cm dis- Cs–Y). Thus, we conclude that methanol is coordinately
appeared.
bound to the metal cation via the lone pair of its oxy-
gen and the strength of the interaction between methanol
and the catalyst is mainly determined by this bonding. This
is clearly seen when the decreasing strength of methanol
sorption with increasing size of the cation in a series of
alkali-exchanged X-zeolites is considered (28). The increas-
4
. DISCUSSION
4
.1. Adsorption and Coadsorption of the Reactants
When methanol is sorbed on the alkali-cation-exchanged ing base strength of the lattice oxygen in this series leads to
zeolites (28), all bands typical for hydroxyl and methyl more pronounced hydrogen bondingbetween the methanol
group stretching vibrations shift to lower wavenumbers in- hydroxyl group and the lattice oxygen. However, the contri-
dicating a weakening of the OH and CH bands due to the bution of that interaction to the overall strength of sorption
interactions with the catalyst. The wavenumbers of the OH is minor. Note that these interactions are most pronounced
stretching vibration band decreased in the order Cs–Y > for the most basic catalysts of this study, i.e., MgO and the
Na–X > Cs–X, Rb–X. This shift results from two effects, hydrotalcites.
i.e., the increasing interaction between the OH groups and
The infrared spectra of sorbed toluene show a slightly
the oxygen of the zeolite lattice and the increasing inter- downward shift of the C–H stretching vibration bands of the
molecular interactionsbetween sorbed methanolmolecules aromatic ring and the methyl group as the size of the cation
(
36). In the case of methanol sorbed on MgO and hydro- increases. These shifts to lower wavenumbers are attributed
talcites several overlapping broad bands appear indicating to a combination of the strong interactions of the aromatic
strong hydrogen-bonded hydroxyl groups of the surface ring with the cation (acting as electron withdrawing, i.e.,
and of sorbed methanol. The most important suggestion electron pair acceptor (40) sites) and to direct interactions
from these results is that the hydrogen bonds of a molecule with the lattice oxygen. These observations are in line with
like methanol become stronger the more polar (in case of reports by Barthomeuf et al. (41) and Jobic et al. (42) on the
the oxygen, the more basic) the surface is.
adsorption of benzene on Na–Y zeolite.
Note that the wavenumber of the symmetric methyl
With MgO and the hydrotalcites, only traces of toluene
stretching band decreased in the order Na–X > Cs–Y > were adsorbed on the hydroxyl groups, giving rise to a
�
�
1
Rb–X > Cs–X > hydrotalcite > MgO. In line with an ear- band at 3743 cm . Upon contact with toluene a band ap-
1
lier paper of Takezawa and Kobayashi (37) this indicates pears at 3670 cm (in parallel to the decrease of the band
�
1
increasing strength of interaction between the lattice oxy- at 3743 cm ), attributed to the perturbed OH groups
gens of the catalyst and the hydrogen atoms of the methyl hydrogen-bonding toluene. Such hydrogen-bonding inter-
�
1
group of methanol. It suggests that MgO (�s C–H 2795 cm
)
actionswith aromaticmoleculeshave been shown to be very
is the most basic and Na–X (�s C–H 2839 cm ) the least basic weak (43), even when adsorbed on very acidic hydroxyl
material.
groups. For the current sample of MgO the low uptake and
�
1
In the case of the hydrotalcite, the presence of two broad the dominance of this type of interaction suggests that only
OH stretching vibrational bands reflects well the contribu- few hydroxyl groups and partially unsaturated Mg2 exist
+
tion of the perturbed surface OH groups and the OH group on the surface of the sample studied.
of methanol. An exact contribution is not possible at this
point, but previous experience with sorption on acidic zeo- increased with increasing size of the metal cation, as seen
lites indicates the existence of cyclic sorbed species (38).
from the increasing thermal stability during t.p.d. of toluene
In contrast to methanol, the sorption strength of toluene
For MgO the case is more complex and the relatively when going from MgO, hydrotalcite, and Na–X to Cs–X and
narrow bands that appear with the broad bands, show dra- Cs–Y. The stronger bonding results from the specific inter-
matically lower hydrogen-bonding strength and/or dissoci- action of the aromatic ring with the metal cation. The better
ation of methanol and the formation of new hydroxyl bands the match of the delocalized �-orbitals of toluene (Lewis
(
33, 39). Moreover a significant fraction of the molecules is base) and the orbitals of the electron-pair-accepting alkali
concluded to be sorbed on Mg cations which also should cation (Lewis acid), the higher the stability of the bond.
lead to a relatively sharp OH band. The C–H stretching Note that this is clearly observed also in more complex
2+

SELECTIVE ALKYLATION OF TOLUENE OVER BASIC MATERIALS
63
+
systems, such as the K -channels used by Dougherty et al. earlier reports of Yashima et al. (48) and Wieland et al. (49),
44, 45). In consequence, this leads to a low mobility of the who reported formaldehyde formation and subsequent
(
aromatic molecules sorbed in alkali-exchanged X zeolites side chain alkylation of toluene. We attribute this different
with large cations, as suggested also from earlier NMR behavior to the existence of CsO moieties in the zeolites
studies (19, 46). This low mobility, however, is not the result used by the above-mentioned authors (48, 49), in accor-
of the steric constraints in the zeolite pores induced by dance with Hathaway et al. (50). The absence of such very
the larger size of the metal cation, as shown by the similar basic oxide clusters leaves the parent Cs–Y zeolite inactive
uptake rates on all samples studied. Note in this context for methanol dehydrogenation. Note in this context that
that Bellat et al. (47) reported even higher uptake rates Lavalley et al. (31) did not mention any formaldehyde
for the K–Y/p-xylene system than for the Na–Y/p-xylene formation on Cs–Y at elevated temperatures (623 K).
system. The stronger interaction of the larger cations is
reflected clearly in the heats of toluene adsorption (76 nificant fraction of the sites were interacting with toluene
kJ/mol with Na–X and 92 kJ/mol with Cs–X). (see Figs. 2 and 3) and sorbed formate species were ob-
It should be emphasized that with Cs–X zeolites a sig-
The coadsorption experiments fully confirm the results served. Both materials showed an appreciable activity for
and discussion. MgO and hydrotalcite preferentially adsorb side-chain alkylation. Thus, we conclude that from the ba-
methanol, while Cs–X and Cs–Y adsorb mainly toluene. For sic materials reported in this work, only Cs–X possess the
the hydrotalcites and MgO, 95–97% of sorption capacity is adequate balance of both catalytic functions to catalyze the
taken up by methanol, while with Cs–Y 75% of the sorbed side-chain alkylation of methanol. MgO, hydrotalcites, and
molecules are toluene.
Cs–Y were either unable to bind toluene sufficiently or did
not transfer methanol into its dehydrogenated state.
4
.2. Reaction Studies
Previous reaction studies demonstrated that depend-
5
. CONCLUSIONS
ing on the alkali ion exchanged into the X-zeolite, side
chain or ring alkylation is preferred. X zeolites with larger
The present study, demonstrates that several require-
cations (preferentially adsorbing toluene over methanol) ments have to be met to successfully catalyze the side chain
showhigh selectivityto styrene and ethylbenzene. The more alkylation of methanol. The catalyst needs to (i) have suffi-
acidic X zeolites (i.e., those with smaller metal cations and cient base strength to dehydrogenate methanol to formal-
preferentially sorbing methanol over toluene) show high dehyde, (ii) stabilize toluene on the surface or in the pores
selectivity to xylenes (see, e.g., Palomares et al. (28)).
and activate its methyl group, and (iii) establish adequate
Hydrotalcites and MgO readily catalyze methanol de- sorption stoichiometry between toluene and methanol.
hydrogenation; i.e., formaldehyde and its decomposition Only materials which provide all these characteristics show
products were observed in adsorbed state and in the gas appreciable catalytic activity.
phase. After contact with the reactant mixture chemisorbed
These requirements lead us to suggest a pathway for the
�
1
formaldehyde (formate bands at 1610–1602 cm (39)) was reaction (Scheme 1) similar to an aldol-type condensation,
the most abundant surface species. The formation of CO which was discussed in detail in Ref. (28), as also postulated
and H2 was observed as the formate decomposed (673 K). by Itoh et al. (21). One elementary step is the abstraction of
Other CO containing compounds adsorbed at different po- the hydrogen from methanol adsorbed on the basic materi-
�
1
sitions (as evidenced by bands at 1109 and 1060 cm ) were als leading to the formation of formaldehyde and formates
also seen. With increasing temperature these compounds with a C-atom positively polarized. Note that this does not
transform to surface species characterized by bands at 1089 require redox properties in the oxide catalysts (see, e.g.,
�
1
and 1118 cm (mixture of carbonate, methoxy, etc.) (33) Ref. (51)). In a parallel reaction step, toluene is strongly
and finally decompose into CO2 and H2O at high tempera- retained on the large cations due to the interactions of the
tures (773 K). In the temperature interval monitored the toluene �-electrons (electron pair donor) and the cation
surface concentration of toluene was minimal. Thus, we (electron pair acceptor). This interaction withdraws elec-
conclude that the low toluene concentration prevents an tron from the aromatic ring and facilitates the polarization
appreciable production of ethylbenzene.
of the methyl C–H, much as an electron withdrawing sub-
In contrast, with Cs–Y strong sorption of toluene side stituent (e.g., a nitro group) on the aromatic ring would
was observed (see Figs. 2 and 3). However, while formate- (40, 51). Additionally the methyl group of toluene inter-
type species are present after sorption of methanol on acts with the oxygen atoms of the basic framework further
Cs–X (see, e.g., Ref. (32)), they were not observed with polarizing the methyl C–H bond. Note that this leads in
Cs–Y (see Fig. 7). This implies that Cs–Y was unable to the transition state (when the hydrogen is being split off as
+
dehydrogenate/activate methanol. Thus, we attribute the H ) to a negatively charged methyl carbon. Under these
absence of catalytic activity for side chain alkylation to the conditions, the positively charged C atom of formaldehyde
absence of chemisorbed formaldehyde. This contrasts with would react with the activated carbon atom of toluene to

6
4
PALOMARES ET AL.
ACKNOWLEDGMENTS
The financial support of the European Community, Human Capital
and Mobility Project, under Grant CIPA-CT94-0184 is acknowledged. We
thank Dr. Florian Eder and Ir. G. Nivarthy for the gravimetric measure-
ments and Prof. Dr. H. G. Karge and Prof. A. Corma for providing the
samples used in this study.
REFERENCES
1
2
3
4
5
. Hattori, H., Chem. Rev. 95, 537 (1995).
. Dartt, C. B., and Davis, M. E., Catalysis Today 19, 151 (1994).
. Kollar, J. (Ed.), Adv. Chem. Technol. 1, 1 (1979).
. Pines, H., and Arrigo, J. T., J. Amer. Chem. Soc. 79, 4958 (1957).
. Garces, J. M., Vrieland, E., Bates, S. I., and Scheidt, F. M., Stud. Surf.
Sci. Catal. 20, 67 (1985).
6
7
8
9
. Giordano, N., Pino, L., Cavallaro, S., Vitarelli, P., and Rao, B. S., Zeo-
lites 7, 131 (1987).
. Yashima, T., Sato, K., Hayasaka, T., and Hara, N., J. Catal. 26, 303
(
1972).
. Sidorenko, Y. N., Galich, P. N., Gutyrya, V. S., Ilin, V. G., and Neimark,
I. E., Dokl. Akad. Nauk. SSSR 173, 132 (1967).
. Engelhardt, J., Szanyi, J., and Valyon, J., J. Catal. 107, 296 (1987).
1
1
0. Chen, N. Y., Kaeding, W. W., and Dwyer, T., J. Am. Chem. Soc. 101,
783 (1979).
1. Kaeding, W. W., Chu, C., Young, L. B., and Butter, S. A., J. Catal. 67,
59 (1981).
6
1
1
1
1
2. Young, L. B., Butter, S. A., and Kaeding, W. W., J. Catal. 76, 418 (1982).
3. Hathaway, P. E., and Davis, M. E., J. Catal. 119, 497 (1989).
4. Wieland, W. S., Davis, R. J., and Garces, J. M., Catal. Today 28, 443
(
1996).
5. Archier, D., Coudurier, G., and Nacchache, C., in “Proceedings of the
th Zeolite Conference, II, Montreal” (R. von Ballmoos et al., Eds.),
1
9
p. 525. Butterworth–Heinemann, London, 1993.
1
1
6. King, S. T., and Garces, J. M., J. Catal. 104, 59 (1987).
7. Philippou, A., and Anderson, M. W., J. Am. Chem. Soc. 116, 5774
SCHEME 1
(
1994).
1
8. Unland, M. L., and Barker, G. E., in “Chemical Industries Series 5,
Catalysis of Organic Reactions” (W. R. Moser, Ed.), p. 51. Dekker,
New York, 1981.
form the new carbon–carbon bond at the methyl group.
Most likely the reaction would proceed via a concerted 19. Sefcik, M. D., J. Am. Chem. Soc. 101, 2164 (1979).
2
2
2
0. Rakoczy, J., React. Kinet. Catal. Lett. 48, No. 2, 401 (1992).
1. Itoh, H., Miyamoto, A., and Murakami, Y., J. Catal. 64, 284 (1980).
2. Lacroix, C., Deluzarche, A., Kiennemann, A., and Boyer, A., J. Chim.
Phys. 81, n. 7/8, 481 (1984).
3. Itoh, H., Hattori, T., Suzuki, K., and Miyamoto, A., J. Catal. 79, 21
(1983).
4. Eder-Mirth, G., Wanzenbock, H. D., and Lercher, J. A., Stud. Surf.
Sci. Catal. 94, 449 (1995).
5. Mielczarski, E., and Davis, M. E., Ind. Eng. Chem. Res. 29, 1579 (1990).
6. Eder-Mirth, G., and Lercher, J. A., Recl. Trav. Chim. Pays-Bas 115,
mechanism avoiding the existence of long lived ionic in-
termediates. We would emphasize the strong analogy to a
classical aldol condensation by stressing the improved ease
of that reaction if an electron withdrawing group is placed
on the aromatic ring (51).
The elementary step mentioned first may proceed on
alkali-cation-exchanged X zeolites, hydrotalcite, and MgO,
but not with Cs–Y (as it has an insufficient base strength);
2
2
2
2
the other is only possible with basic materials having large
157 (1996).
+
accessible cations, such as the Cs cations in the exchanged 27. Eder-Mirth, G., Collect. Czech. Chem. Commun. 60, 421 (1995).
28. Palomares, A. E., Eder, G., and Lercher, J. A., J. Catal. 168, 442 (1997).
29. Mirth, G., and Lercher, J. A., J. Phys. Chem. 95, 3736 (1991).
30. Mirth, G., Eder, F., and Lercher, J. A., Appl. Spectrosc. 48, 194 (1994).
31. Ziolek, M., Czyzniewska, J., Lamotte, J., and Lavalley, J. C., Catal.
Lett. 37, 223 (1996).
X and Y zeolites, but not with hydrotalcite and the MgO
studied in the present report. This is not so much related
to the base strength of the lattice oxygen, but to the in-
ability to retain and polarize toluene under reaction condi-
tions. Note that Cs–X is the only material reported here 32. Unland, M. L., J. Phys. Chem. 82, 580 (1978).
3
3
3
3. Tench, A. J., Giles, D., and Kibblewhite, J. F. J., Trans. Faraday Soc.
7, 854 (1971).
4. Dzwigaj, S., De Mallmann, A., and Barthomeuf, D., J. Chem. Soc.
Faraday Trans. 86, 431 (1990).
5. Busca, G., Lamotte, J., Lavalley, J. C., and Lorenzelli, V., J. Am. Chem.
Soc. 109, 5197 (1987).
that fulfills both conditions. These results also demon-
strate that side chain alkylation does not depend solely
on basic properties of a catalytic material and is, hence,
inadequate for probing the base strength of unknown
materials.
6

SELECTIVE ALKYLATION OF TOLUENE OVER BASIC MATERIALS
65
3
3
3
3
4
6. Palomares, A. E., Rep, M., and Lercher, J. A., in preparation.
7. Takezawa, N., and Kobayashi, H., J. Catal. 28, 335 (1973).
8. Van Santen, R. A., and Kramer, G. J., Chem. Rev. 95, 637 (1995).
9. Kagel, R. O., and Greenler, R. G., J. Chem. Phys. 49, n. 4, 1638 (1965).
0. McMurry, J., “Organic Chemistry,” Brooks/Cole, Pacific Grave, CA, 47. Bellat, J.-P., and Simonot-Grange, M.-H., Zeolites 15, 124 (1995).
44. Dougherty, D. A., Science 271, 163 (1996).
45. Kumpf, R. A., and Dougherty, D. A., Science 261, 1708 (1993).
46. Borovkov, V. Y., Hall, W. K., and Kazanski, V. B., J. Catal. 51, 437
(1978).
1
988.
48. Yashima, T., Sato, K., Hayasaka, T., and Hara, N., J. Catal. 26, 303
(1972).
49. Wieland, W. S., Davis, R. J., and Garces, J. M., J. Catal. 173, 490 (1998).
50. Hathaway, P. E., and Davis, M. E., J. Catal. 116, 279 (1989).
51. March, J., “Advanced Organic Chemistry,” Wiley–Interscience, New
York, 1992.
4
4
1. De Mallmann, A., and Barthomeuf, D., J. Phys. Chem. 93, 5936 (1989).
2. Fitch, A. N., Jobic, H., and Renouprez, A., J. Phys. Chem. 90, 1311
(
1986).
4
3. Derewinski, M., Haber, J., Ptaszynski, J., Lercher, J. A., and
Rumplmayer, G., Stud. Surf. Sci. Catal. 28, 957 (1986).