Tailoring mesoscopically structured H-ZSM5 zeolites for toluene methylation

Article abstract of DOI:10.1016/j.jcat.2013.12.003

Mesoscopically structured zeolites based on H-ZSM5 were designed and synthesized as highly active and shape selective catalysts for methylation of toluene by tuning diffusion and acid site concentration of the catalysts. This was achieved by combining desilication, subsequent dealumination and chemical deposition of a mesoporous SiO₂ overlayer of several nanometer thickness. The decreasing effective diffusion length in zeolite crystals achieved by desilication and dealumination increased the turnover rate of toluene by favoring activation of methanol and facilitating desorption of the produced xylenes, albeit with some loss in p-xylene selectivity. The presence of the SiO₂ overlayer increased the p-xylene selectivity by enhancing the tortuosity of the zeolite, randomly blocking pore openings at the surface, and increasing the effective diffusion path length. The final material combines the higher catalyst utilization with enhanced selectivity leading to rates comparable to the parent zeolite, but at significantly higher selectivity.

Full text of DOI:10.1016/j.jcat.2013.12.003

Journal of Catalysis 311 (2014) 271–280
Contents lists available at ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
Tailoring mesoscopically structured H-ZSM5 zeolites for toluene
methylation
a
a
c
b
a
John H. Ahn , Robin Kolvenbach , Carolina Neudeck , Sulaiman S. Al-Khattaf , Andreas Jentys ,
a,
⇑
Johannes A. Lercher
a
Department of Chemistry, Technische Universität München, Lichtenbergstraße 4, 85747 Garching, Germany
Center of Research Excellence in Petroleum Refining and Petrochemicals, King Fahd University of Petroleum & Minerals, Dhahran 31261, Saudi Arabia
Max Planck-Institut für Kohlenforschung, Kaiser-Wilhelm-Platz 1, 45470 Mülheim an der Ruhr, Germany
b
c
a r t i c l e i n f o
a b s t r a c t
Article history:
Mesoscopically structured zeolites based on H-ZSM5 were designed and synthesized as highly active and
shape selective catalysts for methylation of toluene by tuning diffusion and acid site concentration of the
catalysts. This was achieved by combining desilication, subsequent dealumination and chemical deposi-
Received 5 October 2013
Revised 2 December 2013
Accepted 5 December 2013
2
tion of a mesoporous SiO overlayer of several nanometer thickness. The decreasing effective diffusion
length in zeolite crystals achieved by desilication and dealumination increased the turnover rate of tol-
uene by favoring activation of methanol and facilitating desorption of the produced xylenes, albeit with
some loss in p-xylene selectivity. The presence of the SiO overlayer increased the p-xylene selectivity by
2
enhancing the tortuosity of the zeolite, randomly blocking pore openings at the surface, and increasing
the effective diffusion path length. The final material combines the higher catalyst utilization with
enhanced selectivity leading to rates comparable to the parent zeolite, but at significantly higher
selectivity.
Keywords:
Toluene methylation
Mesoscopically structured H-ZSM5
Methanol conversion
p-Xylene selectivity
Acid chemistry in zeolites
Ó 2013 Elsevier Inc. All rights reserved.
1
. Introduction
of o- and m-xylene, while it increases for p-xylene [15]. On the
other hand, the desilication and subsequent dealumination of H-
The reaction of toluene with methanol to xylenes has great po-
tential to become an important process in the chemical industry
1]. Among the three xylene isomers, p-xylene has the highest de-
ZSM5 shorten the diffusion path length and increase the transport
rates of all aromatic molecules [16,17], leading to enhanced activ-
ity [18–20], albeit with lower shape selectivity compared to the
parent material [20].
In this work, the strategy to synthesize small crystal zeolites with
2
a mesoporous SiO overlayer for efficient catalyst utilization and
high shape selectivity is reported. The synthesis of materials struc-
turedona meso-scale levelis described, togetherwiththecharacter-
ization of intermediate and final materials, and the investigation of
their catalytic impact on the toluene methylation kinetics.
[
mand as a key intermediate for the production of terephthalate,
which itself is an intermediate for polyester synthesis [2]. Methyl-
ation of toluene on commercially available medium pore zeolites,
however, often yields thermodynamic mixtures of the xylene iso-
mers [3–5] (i.e., ortho:meta:para xylene ratio of ꢀ22:53:25 at
6
50 K [6]). As the xylene isomers have similar boiling points, en-
ergy intensive processes, such as adsorption or fractional crystalli-
zation, are required for separation [1,2].
The para-selectivity with H-ZSM5 has been reported to be en-
hanced by increasing the zeolite crystal size [7–9], by impregnat-
ing with phosphorous or boron compounds [4,8,10,11], and by
chemical vapor (CVD) [12] or liquid deposition (CLD) [13,14] of tet-
raethyl orthosilicate (TEOS). In particular, TEOS deposition on the
zeolite particle surface is an attractive method to increase the
shape selectivity by partially blocking the pore openings as well
as by reducing the concentration of Brønsted acid sites in pore
mouth region [3,12,14]. The modification decreases the diffusivity
2. Experimental
2.1. Materials
Zeolite H-ZSM5 (Si/Al = 36; Süd-Chemie) was used as parent
material, and five different hierarchical samples were prepared
by desilication (DS), subsequent dealumination (DS–DA) and by
surface modification (SM) by CLD of TEOS, as shown in Scheme 1.
The desilicated DS sample was prepared by heating the parent H-
3
ZSM5 in a 0.2 M NaOH (>98%, Sigma–Aldrich) solution (30 cm per
⇑
Corresponding author.
E-mail address: Johannes.Lercher@ch.tum.de (J.A. Lercher).
gram of zeolite) at 340 K under stirring for 0.75 h [21]. The solution
was transferred into vials, and the solid phase was centrifuged at
0
021-9517/$ - see front matter Ó 2013 Elsevier Inc. All rights reserved.
http://dx.doi.org/10.1016/j.jcat.2013.12.003

2
72
J.H. Ahn et al. / Journal of Catalysis 311 (2014) 271–280
H-ZSM5
DS
DS-DA
*
*
*
*
*
*
*
*
*
*
*
*
*
Desilication
*
Dealumination
*
*
*
*
*
*
*
0
.2 M NaOH at
*
1 M tartaric acid
at 333K, 4 h
*
*
*
*
*
*
*
3
40K, 0.75 h
*
*
*
*
*
*
Surface Modification
TEOS (3 x 4 wt% of SiO ) in n-hexane at 353K, 1 h
2
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
2-5 nm
SM
DS-SM
DS-DA-SM
Scheme 1. Schematic representation and TEM images of the hierarchical samples prepared from the parent H-ZSM5 zeolite. DS = desilicated H-ZSM5, DS–DA = dealuminated
ꢂ
DS sample, SM = surface modified sample by chemical liquid deposition of tetraethyl orthosilicate (TEOS). Indicates the location of extra-framework Al species (EFAl).
4
000 rpm for ꢀ0.5 h and washed with deionized water. The washing
procedure was repeated three times. The sodium form of the DS zeo-
lite was exchanged into the ammonium form in 0.2 M NH Cl
>99.5%, Sigma–Aldrich) solution (30 cm3 per gram of zeolite) at
2.2. Catalyst characterization
4
The elemental composition and the crystal size distribution of
the materials were determined by atomic absorption spectroscopy
(AAS) using Unicam M Series Flame-AAS equipped with an FS 95
auto-sampler and a GF 95 graphite furnace and by dynamic light
scattering (DLS) using a Malvern Zetasizer Nano ZS system, respec-
tively. The powder X-ray diffraction (XRD; Philips X’Pert Pro sys-
(
3
53 K under stirring for 6 h. The solid was separated and the ion ex-
Cl solution was repeated three times.
change with fresh 0.2 M NH
After the third ion exchange, the zeolites were separated by centri-
fugation, washed with deionized water and dried (at 353 K) before
4
3
ꢁ1
the sample was treated in a synthetic air (flowing at 1.7 cm s
;
a
tem, kCuK = 0.154056 nm, 40 kV/40 mA) patterns were recorded
, Westfalen) at 823 K (heating rate of 0.05 K s 1) for
ꢁ
between 2h angles of 5–70° (step size of 0.017° and a scan speed
of 0.3 s per step). The signals between ꢀ5–10° and 22–25° after
background correction were integrated and compared to the par-
ent H-ZSM5 to estimate the relative crystallinity of the hierarchical
samples (see Table S1). The TEM images were taken by a Hitachi H-
7100 with a maximum acceleration of 120 kV after cutting through
the cross section of the sample.
2
0.5% O
0 h to obtain the Brønsted acidic form of the zeolite.
The desilicated and dealuminated (DS–DA) sample was pre-
pared from the DS sample in a 1 M 2,3-dihydroxybutanedioic acid
2
in N
2
1
3
(L-tartaric acid; >99.5%, Sigma–Aldrich) solution (20 cm per gram
of zeolite) at 333 K under stirring for 4 h [22]. This solution was
separated, washed, and dried as described above, before it was cal-
cined in synthetic air (flowing at 1.7 cm s ) at 823 K (0.05 K s
for 10 h.
3
ꢁ1
ꢁ1
)
2
The N physisorption was carried out at 77 K on a PMI auto-
mated sorptometer after outgassing the samples under vacuum
at 523 K for 2 h. The BET isotherm [23] was used to evaluate the
apparent specific surface area over a relative pressure range from
The surface modified SM samples were prepared by heating the
3
parent H-ZSM5, the DS or the DS–DA sample in hexane (25 cm per
gram of zeolite; 97%, Sigma–Aldrich) with TEOS (4 wt% SiO
2
per
0.01 to 0.1 p/p
ated by using the
hydroxylated silica [25] as the reference adsorbent. The macro-
pore volume was calculated by subtracting micro- and meso-pore
volumes from the total pore volume determined at p/p = 0.95. The
o
pore size distribution of the zeolites was evaluated by the DFT
method (cylindrical pore, NLDFT equilibrium model).
o
. The micro- and meso-pore volumes were evalu-
comparative plot [24] with non-porous
gram of zeolite; >99.0%, Sigma–Aldrich) at 353 K under stirring
for 1 h [13]. Hexane was removed with a rotary evaporator under
vacuum, and the materials were dried at 353 K, before the treat-
a
s
3
ꢁ1
ment in
(
(
a
synthetic air (flowing at 1.7 cm s
)
at 353 K
ꢁ
1
ꢁ1
0.083 K s ) for 2 h, 453 K (0.033 K s ) for 3 h, and finally 823 K
0.033 K s ) for 5 h. This procedure was repeated three times for
ꢁ
1
all samples (i.e., H-ZSM5, DS, and DS–DA) to obtain the final mate-
rial (total deposition amount of 12 wt% of SiO ).
Infrared (IR) spectroscopy (Thermo Nicolet 5700 FT-IR spec-
trometer, resolution 4 cm ) with pyridine (99.8%, Sigma–Aldrich)
ꢁ1
2

J.H. Ahn et al. / Journal of Catalysis 311 (2014) 271–280
273
and 2,6-di-tert-butyl-pyridine (2,6-DTBPy; >97%, Sigma–Aldrich)
as probe molecules was used to determine the total concentration
of Brønsted and Lewis acid sites as well as of Brønsted acid sites lo-
cated in the pore mouth regions [26], respectively. 2,6-DTBPy was
used as a probe molecule to determine the concentration of
Brønsted acid sites in the pore mouth regions, because the kinetic
diameter of 2,6-DTBPy (1.05 nm) is much larger than the size of the
H-ZSM5 micropores (0.51 ꢃ 0.55 and 0.53 ꢃ 0.56 nm [27]). All
samples were pressed into self-supporting wafers (density
removed Si, while dealumination primarily removed extra-frame-
work Al (vide infra) from the zeolite. The surface modification by
TEOS led to the deposition of a mesoporous SiO
macro- and large mesoporous volume of the zeolite domains. Note
that the SiO deposition nominally increased the Si/Al ratio.
2
overlayer in the
2
The BET and external surface areas, as well as the meso- and
macropore volumes, increased considerably after desilication, but
only marginally after subsequent dealumination. The increase in
the pore volume with diameters of approximately 4–6 nm for the
DS sample in Fig. 1 confirmed the generation of mesopores by desi-
lication (see also Scheme 1), while the pore volumes did not
change significantly after the subsequent dealumination. Deposi-
tion of the SiO overlayer decreased the micropore volume, but in-
2
creased the pore volume of pores with diameters <2 nm (Table 1
and Figs. S1–S3). This suggests that some of the micropores were
blocked by our SiO deposition, while the overall increase in the
2
pore volume of pores with a diameter between 1 and 50 nm is in
good agreement with a typical mesoporous oxide [28].
The XRD patterns and the relative crystallinity shown in Fig. 2
confirmed that the parent and mesoscopically structured H-ZSM5
samples maintained good crystallinity throughout the modifica-
tion procedures. Note that the relative crystallinity of the SM sam-
ple decreased to 90% relative to the parent H-ZSM5, because the
ꢁ2
ꢁ7
ꢀ
0.01 g cm ) and activated in vacuum (p < 10 kPa) for 1 h at
23 K (heating rate of 0.17 K s ) before the spectra of the acti-
ꢁ1
7
vated samples were measured. The samples were exposed to pyr-
idine or 2,6-DTBPy at 0.01 kPa and 423 K for 0.5 h and evacuated
for 1 h to desorb weakly bound molecules. The bands at
ꢁ
1
ꢁ1
ꢀ
1545 cm and ꢀ1450 cm in the IR spectra of pyridine ad-
sorbed were integrated to determine the total concentration of
Brønsted and Lewis acid sites, respectively. Pyridine adsorbed on
ꢁ1
Lewis acid sites resulted in two bands at 1447 cm
and at
ꢁ
1
1
455 cm , which were deconvoluted with two Gaussian functions
2
(R
values were above 0.98 in all cases). The sample was subse-
ꢁ1
quently heated to 723 K (0.17 K s ) for 0.5 h in vacuum to deter-
mine the concentration of strong Brønsted and Lewis acid sites.
In order to calculate the concentration of Brønsted acid sites inter-
acting with 2,6-DTBPy, the change of 3610 cm area of the parent
H-ZSM5 sample was correlated with the integrated area of the
ꢁ1
SiO
2
overlayer deposited is amorphous (ꢀ12 wt%). The slight in-
crease in the average crystal size from 120 to 140 nm (Fig. 3) after
desilication, dealumination, and surface modification resulted
from the removal of a small fraction of the smallest crystals after
centrifugation (slightly cloudy solution even after 0.5 h of centrifu-
gation at 4000 rpm).
+
ꢁ1
NAH stretching band of protonated 2,6-DTBPy at 3367 cm
[
3
26]. This ratio was used to relate the area of the band at
367 cm to the concentration of Brønsted acid sites interacting
ꢁ1
with 2,6-DTBPy for the mesoscopically structured samples. All
spectra were collected at 423 K and normalized to the overtone lat-
ꢁ1
tice vibration bands at 1990 and 1870 cm for comparison of the
IR spectra of different samples. Note that the acid site concentra-
tions reported are normalized to the weight of zeolite, and the
3.2. Acid site characterization of hierarchical materials
The IR spectra of the activated samples are shown in Fig. 4 and
Figs. S4–S5 (see also Fig. S6). Two distinct bands were observed at
2
materials with an SiO overlayer were further normalized to ac-
ꢁ1
ꢁ1
count for the weight of the TEOS deposited (12 wt%).
3745 cm and 3610 cm , characteristic for the OAH vibration of
terminal silanol groups and Brønsted acid sites, respectively
[
29,30]. Desilication of the parent H-ZSM5 significantly increased
2
.3. Kinetic experiments
the external surface area (Table 1), and consequently, the concen-
tration of terminal silanol groups (3745 cm ) increased [30]. A
band at 3670 cm appeared after desilication, which is attributed
to the OH groups of extra-framework aluminum (EFAl) species
[31]. The broad band at 3500 cm [32] decreased as a result of
the silanol nest removal (see Fig. S4, left) [33]. The decrease in
intensity of the band at 3670 cm showed that subsequent dealu-
mination of the desilicated sample removed a part of the EFAl spe-
cies. This procedure also leached some tetrahedral Al3+ from the
lattice and led to the formation of silanol nests (Fig. S4, left) [34],
as indicated by the simultaneous re-appearance of a broad band
2
at 3500 cm . Deposition of the mesoporous SiO overlayer de-
creased the bands at 3745 cm and 3610 cm (terminal SiOH
groups and Brønsted acid sites), while the intensity of the broad
ꢁ1
The catalyst samples (4–25 mg, 180–250 lm) diluted with sili-
ꢁ1
con carbide (7 times the weight of the catalyst; F46, ESK-SiC
GmbH) were held in place by quartz wool inside a quartz plug flow
reactor (0.4 cm ID). All catalysts were treated at 823 K (0.17 K s
under flowing He (1.7 cm s ; 99.996%, Westfalen) for 0.5 h prior
to the reaction. The temperature was measured by a type K ther-
mocouple in external contact to the reactor. It was maintained con-
stant by a stainless steel furnace controlled by an Eurotherm
controller (Series 2416). The toluene methylation was carried out
between 548 and 723 K at atmospheric pressure by flowing pre-
mixed toluene (>99.9%, Sigma–Aldrich) and methanol (MeOH;
ꢁ1
ꢁ
1
)
3
ꢁ1
ꢁ
1
ꢁ1
ꢁ1
ꢁ1
>99.8%, Sigma–Aldrich) feed (ptoluene = 6 kPa, pmethanol = 1.5 kPa)
into a vaporizer filled with silicon carbide. The total flow rate
was varied between 1.2 and 2.3 cm s . The reactor effluent was
analyzed by online gas chromatography (Agilent 7820A) equipped
ꢁ1
3
ꢁ1
band at 3660 cm increased because of the presence of hydrogen
bonded SiOH groups in the amorphous SiO layer [35].
The IR spectra of adsorbed pyridine (after subtracting the spec-
tra of activated sample) are shown in Fig. 5. The band at 1545 cm
results from pyridinium ions formed by Brønsted acid sites [36,37]
2
with a DB-WAX column (30 m ꢃ 0.32 mm ꢃ 0.5
lm) and a flame
ꢁ1
ionization detector. All rates were normalized by the total concen-
tration of Brønsted acid sites.
ꢁ1
and the bands at 1447 and 1455 cm characterize coordinately
bound pyridine molecules on OH groups and on extra-framework
3
+
3
. Results and discussion
Al cations [38], respectively. The two bands characterizing coor-
dinative bonding were quantitatively evaluated assuming similar
molar extinction coefficients (Table 2). The deconvoluted IR spectra
of adsorbed pyridine on H-ZSM5 and DS are shown as an example
in Fig. S7.
3.1. Chemical composition and structural characterization of
mesoscopically structured materials
The chemical composition and the textural properties of the
parent and mesoscopically structured materials are compiled in
Table 1. The Si/Al ratios indicate that desilication selectively
The concentration of Brønsted and Lewis acid sites increased
upon desilication (Table 2). Most of the newly formed Lewis acid
sites are assigned to EFAl species (1455 cm ), formed via OAAl
ꢁ1

2
74
J.H. Ahn et al. / Journal of Catalysis 311 (2014) 271–280
Table 1
Chemical composition and textural properties of the parent and mesoscopically structured materials derived from H-ZSM5.
BETa(m2
g
ꢁ1
)
S
extb(m2
g
ꢁ1
)
V
pore
d
pore < 1 nm (cm3
g
ꢁ1
)
V
pore
d
pore = 1–50 nm (cm3
g
ꢁ1
)
pore
V d g )
pore > 50 nm (cm3
ꢁ1
Catalyst
Si/Alratio
S
H-ZSM5
DS
DS–DA
SM
DS–SM
DS–DA–SM
36
27
39
42
36
47
435
482
484
434
348
435
57
96
99
24
41
84
0.12
0.12
0.12
0.09
0.06
0.08
0.03
0.05
0.05
0.08
0.08
0.06
0.17
0.23
0.24
0.13
0.14
0.20
a
S
S
BET = specific surface area analyzed according to Brunauer–Emmett–Teller.
ext = external surface area.
b
0
0
.12
.09
Terminal
silanol
Brønsted
acid
0.005
DS-DA-SM
0
0
.06
.03
0
DS-SM
SM
2
4
6
8
10
12
DS-DA
Pore Size (nm)
Fig. 1. Pore size distribution of the parent (H-ZSM5; h), desilicated (DS;
subsequently dealuminated (DS–DA; s) samples, analyzed by DFT method (cylin-
D), and
DS
drical pore, NLDFT equilibrium model).
H-ZSM5
Relative
3
800
3700
3600
3500
3400
crystallinity (%)
-1
DS-DS-SM
Wavenumber (cm )
9
8
1
7
DS-SM
SM
Fig. 4. Infrared (IR) spectra of activated samples (heated to 723 K for 1 h) measured
at 423 K under vacuum (<10 kPa). The bands at 3745 cm
represents OAH vibration of the terminal silanol groups and the Brønsted acid sites,
ꢁ
7
ꢁ1
ꢁ1
and 3610 cm
90
99
DS-DA
respectively. The spectra were normalized to the lattice vibrations.
ꢁ
1
increased significantly ðfrom447to573
with the decrease in Si/Al ratio (Table 1). Please note here, that
the concentration remained nearly constant
from447to 430 , see Fig. S10), if the weight loss after
l
mol g
Þ, consistent
9
4
DS
zeolite
100
H-ZSM5
ꢁ
1
(
lmol g
zeolite
5
15
25
35
45
55
65
3+
desilication (ꢀ25%) was accounted for, indicating that all Al , con-
2
(degrees)
tributing to Lewis or Brønsted acidity, remained in the material,
Fig. 2. Powder X-ray diffraction (XRD) patterns and relative crystallinity of the
parent H-ZSM5 and the mesoscopically structured materials.
4
while the Si was removed selectively as Si(OH) . Subsequent dea-
lumination of the DS sample by tartaric acid decreased the concen-
tration of Lewis and Brønsted acid sites (13%), because not only
most of the EFAl species, but also some tetrahedrally coordinated
bond cleavage during the removal of tetrahedrally coordinated Si
atoms from the zeolite [31]. The total concentration of acid sites
3
+
lattice Al was removed in this step (Table 2).
50
50
4
0
40
30
20
10
0
3
2
1
0
0
0
0
50
100
150
200
250
50
100
150
200
250
Crystal Size (nm)
Crystal Size (nm)
Fig. 3. Crystal size distribution measured by dynamic light scattering (DLS) method. left: H-ZSM5 (h), DS (D) and DS–DA (s), right: SM (h), DS–SM (D) and DS–DA (s).

J.H. Ahn et al. / Journal of Catalysis 311 (2014) 271–280
275
see also Fig. S11) and, thus, was present at crystal boundaries. On
the other hand, the EFAl concentration was affected inversely in
the H-ZSM5 and DS–DA–SM sample, because the formation of EFAl
species by breaking OAAl bonds was greater than their passivation
0.005
Brønsted
acid
Lewis
acid
2
by SiO deposition. This suggests that the Lewis acid sites in these
samples were most likely not accessible to react with the relatively
large TEOS molecules.
DS-DA-SM
DS-SM
Quantitative evaluation of the acid sites at the external surface
and in the pore mouth of the zeolite was performed via adsorption
of 2,6-DTBPy. The corresponding IR spectra are shown in Fig. 6. The
concentrations of Brønsted and Lewis acid sites determined by 2,6-
DTBPy adsorption are summarized in Table 2 and Fig. 7. A fraction
SM
ꢁ1
of terminal SiOH groups (3745 cm ) and Brønsted acid sites
ꢁ
1
(
3610 cm ) interacted with 2,6-DTBPy, as shown by the decrease
ꢁ1
DS-DA
DS
in the intensity of both bands. New bands appeared at 3367 cm
and 1616 cm , which are assigned to the NAH and C@C stretch-
ing vibrations of 2,6-DTBPyH , respectively. The correlation be-
ꢁ1
+
+
tween these bands (shown in Fig. S9) indicates that both bands
can be used for quantification of the acid sites probed.
ꢁ1
ꢁ1
H-ZSM5
1400
After desilication, the bands at 3367 cm and 1616 cm in
Fig. 6 increased compared to the parent H-ZSM5. We attribute this
to the fact that a substantial fraction of Brønsted acid sites became
accessible to 2,6-DTBPy through the generation of mesopores,
increasing in this way the concentration of Brønsted acid sites in
the pore mouth region. These Brønsted acid sites were primarily
removed during subsequent dealumination with tartaric acid, the
kinetic diameter (0.68 nm) of which is slightly larger than the size
of the H-ZSM5 micropores (ꢀ0.55 nm). The decrease in intensities
1
700
1600
1500
-1
Wavenumber (cm )
Fig. 5. IR spectra after adsorption of pyridine at 423 K, 0.01 kPa and outgassing for
h under vacuum (spectra of the activated sample subtracted). The bands at 1447
1
ꢁ
1
and 1455 cm represent the coordinately bonded pyridine molecules on SiOH
ꢁ
1
groups and EFAl species, respectively and the band at 1545 cm pyridinium ions
formed on Brønsted acid sites.
ꢁ1
ꢁ1
of the bands at 3367 cm and 1616 cm compared to DS sample
confirmed the loss of Brønsted acid sites in the pore mouth region.
Surface modification with TEOS decreased the Brønsted acid
site concentration to a similar extent, whether the sample was
treated by desilication or subsequent dealumination prior to depo-
2
The surface modification by the SiO overlayer deposition de-
creased only the fraction of Brønsted acid sites accessible by 2,6-
DTBPy, because the kinetic diameter of TEOS (0.96 nm) is larger
than the size of the H-ZSM5 micropores, and consequently, only
Brønsted acid sites at the pore mouth region were affected
(Fig. 7). Thus, the total loss Brønsted acid site concentration is
equal to the decrease in sites accessible to 2,6-DTBPy (ꢀ30–
2
sition of SiO overlayer (DSM in Table 2; loss of 30–40 lmol of
Brønsted acid sites per g zeolite). Note that most of the Brønsted
acid sites were strong and more than 90% of pyridine was still ad-
sorbed on all sites even after the samples were outgassed to 723 K.
2
Deposition of the mesoporous SiO overlayer onto the DS sam-
ple significantly decreased the Lewis acid site concentration,
whereas it slightly increased for the H-ZSM5 and DS–DA. Hence,
the surface modification procedure affects the concentration of Le-
wis acid sites via the removal of EFAl by reacting with the depos-
40 lmol per g zeolite).
3.2.1. Toluene methylation on parent and modified H-ZSM5
The turnover rates of toluene at different reaction temperatures
2
ited SiO (site passivation) and via generation of EFAl by mild
and the apparent energy of activation are shown in Table 3 (the
conversions of reactants, turnover, and formation rates of metha-
nol and the aromatic products are shown in Tables S3–S5). The
reaction orders of toluene methylation were 0.6–0.7 with respect
to methanol and 0.4 with respect to toluene under the reaction
conditions used [39]. This indicates that the slow diffusing
dealumination (breaking OAAl bond). The Lewis acid site concen-
tration for the DS–SM sample decreased, because the significant
fraction of EFAl formed during desilication (DS sample) was re-
moved during the condensation reaction of TEOS. Most of the EFAl
formed during desilication was accessible to TEOS (DSM DS, Table 2;
Table 2
Concentration of Brønsted and Lewis acid sites determined by adsorption of pyridine and 2,6-di-tert-butyl-pyridine (2,6-DTBPy).
mol g zeoliteꢁ1)
Lewis acid sites (lmol g zeolite )
ꢁ1
Brønsted acid sites (
l
Totala
Strongb
2,6-DTBPy accessiblec
% 2,6-DTBPy interactiond
Total
EFAl
SiOH
H-ZSM5
DS
DS–DA
SM
DS–SM
DS–DA–SM
380
398
345
351
366
312
371
370
332
327
331
293
90
204
131
62
170
92
24
51
38
18
47
29
67
175
55
76
129
74
3
102
16
16
72
64
73
39
60
57
46
28
SM H-ZSM5e
29
32
33
44
39
39
28
34
39
–
–
–
ꢁ9
46
ꢁ19
ꢁ13
30
ꢁ12
4
16
ꢁ7
D
D
D
e
SM DS
SM DS–DAe
a
b
c
After adsorption of pyridine at 423 K and outgassing for 1 h under vacuum.
After subsequently heating the samples to 723 K for 0.5 h.
Total concentration of Brønsted acid sites (both strong and weak) located in the pore mouth region. After adsorption of 2,6-DTBPy at 423 K and outgassing for 1 h under
vacuum.
d
e
Defined as% of 2,6-DTBPy interacting with total Brønsted acid sites.
Difference between before and after the deposition of SiO overlayer.
2

2
76
J.H. Ahn et al. / Journal of Catalysis 311 (2014) 271–280
3
367 cm-1
1616 cm-1
the pores, because the turnover rates were affected by changes in
0.0015
the mesoporosity and the transport path length in the catalyst (Ta-
ble 3). The Weisz–Prater criterion [40] (Table S6) suggests that the
reaction is influenced by the transport of large aromatic products,
while the transport of the main product (p-xylene) and of the reac-
tants (toluene, methanol) is not likely influencing the reaction rate.
Thus, a simplified toluene turnover rate can be formulated by using
a competitive Langmuir formalism for the adsorption of methanol
and toluene [47] neglecting the disproportionation of toluene
(<1%) in Eq. (1):
DS-DA-SM
DS-SM
SM
DS-DA
DS
K
Tol
p
þ
K
MeOH
p
Tol
MeOH
r
TM / kTM
P
P
n
i¼1
n
i¼1
1 þ KTolpTol
with p ¼ fðp
K
n
pn 1 þ KMeOHpMeOH
þ
n
K pn
; pTolÞ
ð1Þ
n
MeOH
where KTM is the reaction rate constant, KMeOH and KTol are the
adsorption equilibrium constant of toluene and methanol, pMeOH
and pTol are the partial pressures of methanol and toluene, while
H-ZSM5
n n
K and p are the adsorption constant and partial pressure of the
products. Eq. (1) indicates that a higher concentration of o- and
m-xylene as well as of multiple alkylated products inside the zeolite
pores reduces both the local concentration of toluene and the sur-
face coverage of methanol. The measured concentration of these ad-
sorbed bulky aromatic products was much higher than the
predicted coverage at equilibrium [41], emphasizing their impor-
tance in reducing the steady-state concentrations of toluene and
methanol.
3
800
3200
2600
2000
1400
-1
Wavenumber (cm )
Fig. 6. Changes in IR spectra after adsorption of 2,6-di-tert-butyl-pyridine (2,6-
DTBPy) at 423 K, 0.01 kPa and outgassing for 1 h in vacuum (spectra of activated
samples are subtracted). The characteristic bands at 3367 cm (NAH vibration)
and 1616 cm (C@C vibration) appear from 2,6-DTBPy interaction with the zeolite.
ꢁ
1
+
ꢁ
1
After desilication and subsequent dealumination, the increase
in the toluene turnover rates is attributed to the decrease in the
effective diffusion length ꢀ45% and 63% (based on o-xylene, see
Table S7), thereby reducing the steady-state concentrations of o-
xylene, m-xylene, and multiple alkylated reaction products [41].
The increase in turnover rate and the decrease in effective path
length after dealumination (DS–DA) indicate that EFAl species con-
tribute to the retention of bulky aromatic molecules inside the
pores. Consequently, their removal (Fig. 7) also decreases the con-
centration of these molecules inside the pores. EFAl (Table 2) spe-
cies formed upon desilication decrease the transport rate by either
physically blocking pore openings or by chemically interacting
with the aromatic molecules.
6
5
4
3
2
1
00
00
00
00
00
00
0
1
75
1
1
29
70
6
7
0
7
6
2
5
5
7
9
4
2
9
6
204
131
290
289
2
The partial blocking of the pore openings by the SiO overlayer
2
14
220
194
196
significantly increased the diffusion path length, especially for the
large aromatic molecules (>70%, see Table S7) [15]. In situ IR spec-
troscopy (Fig. S1 of Ref. [41]) showed that the relative concentra-
tion of larger aromatic products with respect to toluene
increased after surface modification by 30%, which results in a low-
er concentration of the reactants inside the zeolite pores because of
competitive adsorption (Eq. (1)). Thus, we conclude that o-xylene,
m-xylene and multiple aromatic products accumulate toward the
steady-state concentration, while the reaction to p-xylene is (nei-
ther externally nor internally) diffusion limited on any of the inves-
tigated catalysts. Hence, the differences in the activity of the
catalysts are attributed to differences in the (surface) chemistry in-
side the pores rather than to varying effectiveness factors.
H-ZSM5
SM
DS
DS-SM DS-DA DS-DA-SM
Fig. 7. Concentration of internally and externally accessible Brønsted as well as
Lewis acid sites per gram of a zeolite. White represents Lewis acid sites, gray and
black the Brønsted acid site accessible and not accessible by bulky 2,6-DTBPy,
respectively.
molecules formed by multiple alkylations accumulate in the zeo-
lite crystals. As this degree of accumulation varies with the partial
pressure of methanol and toluene, the competitive adsorption of
higher substituted products reduces the increase in the coverage
of both reactants (Eq. (1)) causing a rate dependence lower than
one with respect to the pressure of toluene and methanol.
In general, desilication and subsequent dealumination in-
creased and surface modification decreased the toluene turnover
rates. This trend was especially pronounced at lower temperatures.
The apparent activation energy for toluene methylation decreased
slightly with desilication and dealumination, but increased with
The alkylation reaction proceeds via the activation of methanol
to either a methoxonium ion or a methyl carbenium ion and the
subsequent reaction with an aromatic molecule. The concentra-
tions of adsorbed xylene and trimethylbenzene after leaching de-
creased and after surface modification with SiO increased, which
2
leads to the discussed variation in the concentration of toluene
and methanol inside the pores (Eq. (1)).
surface modification by an external SiO
e.g., from 81 to 101 kJ mol with H-ZSM5 and after surface mod-
ification, respectively (Arrhenius plot shown in Fig. S12).
The reaction of toluene methylation is concluded to be con-
trolled by the accumulation of larger reaction products within
2
overlayer deposition,
In addition, the significant presence of aromatic products from
toluene methylation in the zeolite pores [42] increased, conse-
quently, the probability of competitive methylation reactions, i.e.,
methylation of xylene to TriMB and of TriMB or tetramethylben-
zene (TetraMB). The latter (xylene and TriMB) methylation
ꢁ1

J.H. Ahn et al. / Journal of Catalysis 311 (2014) 271–280
277
Table 3
Turnover rates of toluene at different reaction temperatures and the apparent energy of activation for toluene methylation.
573 K
623 K
673 K
723 K
a
b
Toluene turnover rate
Toluene turnover rate
Toluene turnover rate
Toluene turnover rate
E
app
ꢁ
2
ꢁ1
ꢁ2
ꢁ1
ꢁ2
ꢁ1
ꢁ2
ꢁ1
ꢁ1
(10 mol [s mol H]
)
(10 mol [s mol H]
)
(10 mol [s mol H]
)
(10 mol [s mol H]
)
(kJ mol )
HZSM5
DS
DS–DA
SM
DS–SM
DS–DA–SM
2.1
2.7
3.7
0.84
0.92
1.5
7.5
8.9
11
4.5
4.2
6.5
14
16
19
12
12
15
20
20
25
19
20
24
81
75
68
101
90
87
a
Measured based on the toluene consumption (ptoluene = 6.0 kPa, pmethanol = 1.5 kPa, 10 mg of catalyst and total flow rate = 2.3 cm3
Apparent energy of activation based on the toluene turnover rates measured between 548 and 623 K.
ꢁ1
s
).
b
reaction was more significant for zeolites with higher diffusion
constraints, e.g., SM compared to H-ZSM5, leading to lower toluene
turnover rates. Note that the methylation of larger aromatic prod-
ucts is also favored compared to toluene, because the activation
energy of methylation reactions of the benzene ring decreases with
number of methyl substitutions (ꢀ10 kJ mol difference between
xylene and toluene methylation [43,44]), and therefore, the differ-
ences in the rates are more pronounced at lower reaction temper-
atures [45] (Table 3, see also Section 3.2.2).
120
105
90
ꢁ1
In the reactions studied, one would assume the apparent energy
of activation to increase with a decreasing intra-particle diffusion
limitation; however, the opposite has been observed. The energy
of activation increases linearly with the diffusion path length as
shown in Fig. 8 and Table 3, which reflects the increasing coverage
of the large aromatic products competing for the active sites with
the reactants (methanol and toluene). Therefore, the activation en-
ergy is a measure for the temperature dependence of the coverage
of the heavy alkylated products compared to the reactant mole-
cules. As the concentration of adsorbed reactants (toluene and
methanol) decreases, the increasing concentration of heavier prod-
ucts inside the pores increases the apparent activation energy. As
the concentration of the latter increase with pore length, the
apparent activation energy increases simultaneously. Note that in
the absence of competitive product adsorption, the activation en-
ergy for the reaction would be identical for all catalysts, because
both the local steric environment at the reactive sites and the
chemical composition are not affected by the modification.
In summary, the differences in the catalytic activity are attrib-
uted to the change in the relative concentration of adsorbed spe-
cies in the pores caused by the competitive adsorption of large
aromatic products leading to a lower concentration of the reactant
molecules at the active sites. Consequently, the toluene turnover
rates increased with DS and DS–DA samples (vice versa for the
75
60
0
50
100
150
200
250
Estimated pore length (nm)
Fig. 8. Apparent activation energy of toluene methylation (from Table 3) vs. the
estimated pore length, calculated from the o-xylene uptake rates with IR Spectros-
copy at 403 K [41]. These values were estimated by assuming that the intrinsic
diffusion coefficients were the same between the samples and that the changes
were caused by the differences in the effective diffusion lengths. The filled symbols
represent surface modified samples (SM (j), DS–SM (N) and DS–DA–SM (d)) and
unfilled the non-modified samples (parent H-ZSM5 (h), DS (D) and DS–DA (s)).
The fraction of MeOH for LH formation on the catalysts studied
here is summarized in Table 4. Surprisingly, by decreasing the
reaction temperature, the fraction of methanol used for light
hydrocarbons formation increased from 0.21 at 723 K to 0.48 at
573 K for the parent H-ZSM5 (Table 4). Aromatic products, such
as xylenes and TriMBs, are methylated further and light hydrocar-
bons are dealkylated from these highly methylated aromatics
[48,49], as these molecules cannot exit the pores. Under these con-
ditions, dealkylation becomes favorable over methylation.
(Scheme 2) [39]. The methylation of xylenes or TriMBs was favored
at lower reaction temperatures relative to toluene methylation, be-
cause the activation energy of toluene methylation was higher
than that of xylenes or TriMBs. This in turn led to higher probabil-
ity for the formation of light hydrocarbon via dealkylation from
highly methylated aromatic molecules.
2
samples with SiO overlayer).
3.2.2. Influence of the mesoscopically structured pores and reaction
temperature on methanol usage
The methanol used for alkylating toluene, the aromatic prod-
ucts and for the formation of light hydrocarbons (LH), was dis-
cussed previously as a function of the zeolite framework type
The methanol usage toward formation of light hydrocarbons
also increased as the effective diffusion length became longer,
i.e., the MeOH for LH formation fraction increased from 0.30 to
[
39]. The fraction of MeOH for LH formation was defined as follows:
Fraction of MeOH for LH formation
ꢃ C þ 3 ꢃ C þ 4 ꢃ C þ . . .
Methanolin ꢁ ðMethanol þ 2 ꢃ DMEÞ
0
.48 to 0.75, at 573 K with DS–DA, H-ZSM5, and SM samples,
respectively (Table 4). We attribute this to a longer residence time
of the highly substituted aromatic molecules in the zeolite pores,
which resulted in a higher probability for product methylation
and eventual dealkylation to form light hydrocarbon (Scheme 2).
2
2
3
4
¼
ð2Þ
out
Methanol and dimethyl ether (DME) were treated as one reac-
tant (accounting for the stoichiometry of 1/2 for DME), because
methanol is reversibly dehydrated to DME and both are able to
methylate aromatic molecules via a similar mechanism [46,47]. A
fraction of unity would indicate that methanol was used only for
the formation of light hydrocarbons and none for the methylation
or the formation of aromatics.
3.2.3. Influence of the pores structured on a mesoscale level and the
reaction temperature on selectivities
The selectivities of xylene isomers and TriMBs in the aromatic
products are shown in Table 5. The selectivity to p-xylene decreased
with desilication and dealumination, but increased significantly

2
78
J.H. Ahn et al. / Journal of Catalysis 311 (2014) 271–280
2
with the deposition of an external SiO overlayer and in general on
all catalysts with the reaction temperature. Note that the xylenes
and TriMBs selectivities in aromatics did not change significantly
with the zeolite modifications or with reaction temperatures.
The overall impact of the structural changes on a mesoscale le-
vel with respect to the p-xylene selectivity as a function of the tol-
uene turnover rates at 723 K is shown in Fig. 9, left and as a
function per weight of zeolite in Fig. 9, right. Desilication and sub-
sequent dealumination increased the toluene turnover rate by ꢀ5%
and 30%, respectively, but decreased the p-xylene selectivity with-
in xylenes from ꢀ65% to 55% and 50%. Deposition of the SiO
layer on these materials decreased the toluene turnover rate by
5%, but increased p-xylene selectivity significantly (from 50–
2
over-
ꢀ
6
5% to 75–85%).
The variation in p-xylene selectivity with catalyst modification
is the result of the combination of several pathways. During tolu-
ene methylation, the p-xylene isomer can be generated by toluene
methylation, xylene isomerization, and dealkylation of highly
methylated aromatic molecules [39]. Thus, the p-xylene formation
rate can be formulated assuming that neither the reactants nor p-
xylene are diffusion limited and a certain concentration of o-, m-
xylene, and higher alkylated products is present according to Eq.
Scheme 2. Methylation and dealkylation of light hydrocarbons during the reaction
of toluene with methanol inside H-ZSM5 zeolites. Toluene and p-xylene are not
diffusion limited, while larger aromatic molecules, such as m- and o-xylenes, Tri-
and TetraMBs are accumulated inside the pores. Most of the light hydrocarbons are
formed from highly methylated aromatic products before leaving the zeolite pores
as less-methylated aromatic molecules.
(
3). Disproportionation reaction can be neglected, as it is slow un-
der these reaction conditions, i.e., <1% of the toluene converted
underwent disproportionation.
r
pX / rTM þ rmXiso þ rDealk
ð3Þ
1
selectivity vs. C conversion shown in Fig. S14). Note that the p-xy-
K
Tol
p
þ
K
MeOH
p
Tol
MeOH
r
pX / kTM
P
P
S
TM
n
i¼1
n
i¼1
lene selectivity decreased more pronounced at toluene conversion
levels above ꢀ15% (compared to the lower toluene conversion lev-
els) for the H-ZSM5 sample (Fig. 10, empty square), because the
methanol was nearly depleted at this point (methanol conversion
1
þ KTolpTol
K
n
pn 1 þ KMeOHpMeOH
þ
n
K pn
þ kiso
h
mX þ kDe
h
HAP
ð4Þ
Here, KTM, Kiso, and kDe are the reaction rate constants of the tol-
>90%). Under these conditions, the isomerization became the dom-
uene methylation, the m-xylene isomerization, and the dealkyla-
tion, respectively and STM is the intrinsic selectivity for the
formation of p-xylene via methylation of toluene. hmx and hHAP
are the surface coverages of m-xylene and higher alkylated prod-
ucts. The toluene conversion for the surface modified samples with
inant reaction and the p-xylene selectivity started to approach the
thermodynamic equilibrium.
The relative contribution of dealkylation compared to the other
two pathways can be evaluated by comparing the fraction of meth-
anol used for LH formation with the p-xylene selectivity (Fig. 11) as
both were produced simultaneously in the dealkylation of multiple
alkylated aromatics. The slope of the linear correlation observed
increased with temperature and confirmed the decrease of the
dealkylation rate relative to the methylation and isomerization
rates. It was discussed in the previous section that the rate of tol-
uene methylation decreased with surface modification, while it is
shown here that the p-xylene formation rate increased as the tem-
perature increased from 573 to 723 K (Table 5). Hence, the isomer-
ization is concluded to be the most dominant pathway in
controlling the selectivity of the xylenes formed [41,50]. Likewise,
the p-xylene selectivity increased significantly after the surface
SiO
2
overlayer was generally lower by ꢀ1–4% compared to the
non-modified catalysts (see Table S4) at the same reaction condi-
tions (space velocity, temperature, partial pressures, and similar
1
C conversions). This is attributed to the higher concentration of
large aromatic products in the pores, which results in a lower local
concentration of toluene and adsorbed methanol at the active sites.
Moreover, the higher rate for the methylation of the products com-
pared to that of toluene further lowers the toluene turnover rate
(
Section 3.2.1).
The overall selectivity of p-xylene decreased as the toluene con-
version increased. For the surface modified materials (SM, DS–SM,
and DS–DA–SM), the selectivity to p-xylene was clearly higher at
the same toluene conversion level (Fig. 10, at 723 K; p-xylene
2
modification with the mesoporous SiO layer, because the effective
diffusion path length increased resulting in a strong increase in the
concentration of o-, m-xylene, and higher alkylated products inside
the pores. This situation led to an increase in the rate of both isom-
erization and dealkylation, which over-compensated the decrease
in the toluene methylation rate leading to the increased p-xylene
formation rate observed.
It was previously reported that the concentration of Brønsted
acid sites located in the pore mouth region does not affect the p-
xylene selectivity significantly [3,12,14], which was confirmed by
comparing the Brønsted acid site concentrations (Table 2) with
the selectivities (Table 5). A correlation between the p-xylene
selectivity and the concentration of Brønsted acid sites located in
the pore mouth region was not found. The DS-SM sample with
the highest p-xylene selectivity observed within the series of
catalysts studied has also the highest concentration of acid sites
Table 4
Fraction of methanol (MeOH) used for the formation of light hydrocarbons (LH)a at
different reaction temperatures during toluene methylation for the parent H-ZSM5
and its mesoscopically structured samples.
573 K
623 K
673 K
723 K
H-ZSM5
DS
DS–DA
SM
DS–SM
DS–DA–SM
0.48
0.36
0.30
0.75
0.68
0.60
0.42
0.32
0.28
0.57
0.54
0.52
0.30
0.26
0.20
0.46
0.45
0.40
0.21
0.22
0.15
0.35
0.35
0.31
a
Defined in Eq. (2). Data was measured during toluene methylation at ptoluene
6.0 kPa, pmethanol = 1.5 kPa, 10 mg of catalyst and total flow rate = 2.3 cm3
ꢁ1
=
s .

J.H. Ahn et al. / Journal of Catalysis 311 (2014) 271–280
279
Table 5
Selectivity of xylene isomers in xylenes, xylenes and trimethylbenzenes (triMBs) in aromatics during toluene methylation.
a
5
73 K (6–15%)b
623 K (31–46%)b
673 K (65–75%)b
723 K (82–89%)b
Xylene
selectivity
p:m:o [%]
Xylenes
(TriMBs)
[mol%]c
Xylene
selectivity
p:m:o [%]
Xylenes
Xylene
selectivity
p:m:o [%]
Xylenes
Xylene
selectivity
p:m:o [%]
Xylenes
(TriMBs)
[mol%]c
(TriMBs)
(TriMBs)
c
c
[mol%]
[mol%]
H-ZSM5
DS
DS–DA
SM
DS–SM
DS–DA–
SM
38:28:35
33:29:38
33:28:39
47:25:28
41:27:32
40:27:33
90 (5.8)
92 (5.9)
92 (5.1)
87 (8.0)
86 (8.8)
87 (8.0)
51:25:24
42:29:29
41:29:31
70:17:13
60:21:18
57:23:20
89 (8.3)
87 (9.1)
90 (8.0)
88 (7.3)
85 (9.9)
86 (9.8)
60:24:16
52:29:19
49:29:22
81:13:6
76:15:9
71:18:11
88 (9.2)
86 (9.7)
89 (9.2)
89 (6.5)
87 (8.6)
87 (9.2)
64:24:12
56:30:14
52:31:17
85:11:4
84:12:5
78:16:7
89 (8.8)
87 (9.1)
89 (9.0)
92 (5.3)
90 (6.4)
89 (7.6)
a
b
c
Measured at ptoluene = 6.0 kPa, pmethanol = 1.5 kPa, 10 mg of catalyst and total flow rate = 2.3 cm3
ꢁ1
s
.
C
1
(methanol and DME) conversion (Table S3).
mol% within aromatics; rest is detected as 1,2,4,5-tetraMB.
9
8
7
6
5
0
0
0
0
0
SM
90
80
SM
DS-SM
DS-SM
DS-DA-SM
DS-DA-SM
Surface
modification
70
H-ZSM5
H-ZSM5
Desilication
60
50
DS
DS
Dealumination
DS-DA
0.26
DS-DA
0.18
0.20
0.22
0.24
65
70
75
80
85
90
-
1
Toluene turnover rate (mol [sec mol H] )
Toluene consumption rate
-1
)
Fig. 9. Selectivity of p-xylene within xylenes vs. toluene turnover rate (left, per mole of Brønsted acid site) and consumption rate (right, per gram of zeolite) during toluene
methylation at 723 K (ptoluene = 6.0 kPa, pmethanol = 1.5 kPa, 10 mg of catalyst, total flow rate = 2.3 cm3
s
ꢁ1
). The filled symbols represent surface modified samples (SM (j), DS–
SM (N) and DS–DA–SM (d)) and unfilled the non-modified samples (parent HZSM5 (h), DS (
D
) and DS–DA (s)).
90
80
70
60
50
40
30
20
100
SM
DS-SM
DS-DA-SM
723 K
673 K
8
6
0
0
623 K
H-ZSM5
DS
DS-DA
5
73 K
40
2
0
equilibrium
10
0
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0
5
15
20
Fraction of MeOH for LH formation
Toluene conversion (%)
Fig. 11. p-Xylene selectivity vs. fraction of MeOH for LH formation during toluene
methylation, measured at 573 K (s), 623 K ), 673 K (e) and 723 K (h)
(ptoluene = 6.0 kPa, pmethanol = 1.5 kPa, 10 mg of catalyst, total flow rate = 2.3 cm
Fig. 10. p-Xylene selectivity vs. the conversion of toluene during toluene methyl-
(
D
ation, measured at 723 K (ptoluene = 6.0 kPa, pmethanol = 1.5 kPa, 4–25 mg of catalyst,
3
ꢁ1
total flow rate = 1.2–2.3 cm3
ꢁ1
s ).
s
1
, C (methanol and DME) conversion = 40–99%). The
filled symbols represent surface modified samples (SM (j), DS–SM (N) and DS–DA–
SM (d)) and unfilled the non-modified samples (parent H-ZSM5 (h), DS (
DS–DA (s)).
overlayer, simultaneously enhanced the p-xylene selectivity and
toluene turnover rate (Fig. 9; maintained the activity in per weight
D) and
2
of zeolite basis). Desilication and dealumination prior to SiO depo-
ꢁ
accessible for 2,6-DTBPy (152
l
mol g 1) compared to the parent H-
sition resulted in decreased transport path length of bulky aro-
matic products by more than 2ꢃ (Table S7 [41]) reducing the
amount of large reaction products and, consequently, led to a more
efficient usage of the active sites and higher toluene turnover rates.
ꢁ1
ZSM5 and DS–DA samples (90 and 131
lmol g , respectively)
with a lower p-xylene selectivity.
In conclusion, the series of these zeolite modifications, i.e., desi-
lication, dealumination, and deposition of mesoporous SiO
2
SiO deposition enhanced the isomerization and dealkylation rate
2

2
80
J.H. Ahn et al. / Journal of Catalysis 311 (2014) 271–280
by increasing the local concentration of larger reaction products
and, in turn, increased the p-xylene selectivity.
References
[
[
[
[
1] W. Vermeiren, J.P. Gilson, Top. Catal. 52 (2009) 1131.
2] T. Tsai, S. Liu, I. Wang, Appl. Catal. A 181 (1999) 355.
3] T. Hibino, M. Niwa, Y. Murakami, J. Catal. 128 (1991) 551.
4] W.W. Kaeding, C. Chu, L.B. Young, B. Weinstein, S.A. Butter, J. Catal. 67 (1981)
4
. Conclusions
159.
[
[
[
5] H. Vinek, J.A. Lercher, J. Mol. Catal. 64 (1991) 23.
6] R.D. Chirico, W.V. Steele, J. Chem. Eng. Data 42 (1997) 784.
7] Y.S. Bhat, J. Das, K.V. Rao, A.B. Halgeri, J. Catal. 159 (1996) 368.
Tailoring H-ZSM5 zeolites with a series of well-known modifi-
cation steps enabled synthesis of a new generation of catalysts that
combine high activity and high selectivity to p-xylene via methyl-
ation of toluene. The desilication (DS) and dealumination (DA) with
NaOH and tartaric acid, respectively, and deposition of a mesopor-
[8] L.B. Young, S.A. Butter, W.W. Kaeding, J. Catal. 76 (1982) 418.
9] O. Mikkelsen, P.O. Rønning, S. Kolboe, Micropor. Mesopor. Mater. 40 (2000) 95.
[
[
10] J.P. Breen, R. Burch, M. Kulkarni, D. McLaughlin, P.J. Collier, S.E. Golunski, Appl.
Catal. A 316 (2007) 53.
[11] H. Vinek, G. Rumplmayr, J.A. Lercher, J. Catal. 115 (1989) 291.
ous SiO
2
overlayer onto the zeolite surface were used to modify the
materials by reducing the size of the crystal domains and main-
taining their shape selectivity. The Lewis and Brønsted acid site
concentrations in the pore mouth region (accessible by 2,6-di-
tert-butyl-pyridine) increased substantially after desilication, but
these sites were effectively removed by subsequent dealumination
with tartaric acid, without causing major changes to the mesopo-
rosity. The surface modification by deposition of a mesoporous
SiO overlayer further decreased the concentration of Brønsted
2
acid sites at the pore entrance and increased transport path length
of the bulkier o- and m-xylene isomers.
[12] H.P. Röger, M. Krämer, K.P. Möller, C.T. O’Connor, Micropor. Mesopor. Mater.
21 (1998) 607.
[
13] S. Zheng, H.R. Heydenrych, H.P. Röger, A. Jentys, J.A. Lercher, Top. Catal. 22
(
2003) 101.
[14] J. C
ˇejka, N. Zˇ ilková, B. Wichterlová, G. Eder-Mirth, J.A. Lercher, Zeolites 17
(
1996) 265.
[
[
15] S. Zheng, A. Jentys, J.A. Lercher, J. Catal. 241 (2006) 304.
16] J.C. Groen, W. Zhu, S. Brouwer, S.J. Huynink, F. Kapteijn, J.A. Moulijn, J. Pérez-
Ramírez, J. Am. Chem. Soc. 129 (2007) 355.
17] F.C. Meunier, D. Verboekend, J.P. Gilson, J.C. Groen, J. Pérez-Ramírez, Micropor.
Mesopor. Mater. 148 (2012) 115.
18] C.H. Christensen, K. Johannsen, E. Törnqvist, I. Schmidt, H. Topsøe, C.H.
Christensen, Cat. Today 128 (2007) 117.
19] C.H. Christensen, K. Johannsen, I. Schmidt, C.H. Christensen, J. Am. Chem. Soc.
[
[
[
[
[
[
The turnover rate of toluene and p-xylene selectivity was con-
cluded to be strongly influenced by the presence of large aromatic
products inside the pores and increased simultaneously at rela-
tively high reaction temperatures (>673 K) with H-ZSM5 after desi-
125 (2003) 13370.
20] C. Fernandez, I. Stan, J.P. Gilson, K. Thomas, A. Vicente, A. Bonila, J. Pérez-
Ramírez, Chem. Eur. J. 16 (2010) 6224.
21] J.C. Groen, L.A.A. Peffer, J.A. Moulijn, J. Pérez-Ramírez, Colloids Surfaces A:
Physicochem. Eng. Aspects 241 (2004) 53.
22] J. Kim, M. Choi, R. Ryoo, J. Catal. 269 (2010) 219.
[23] S. Brunauer, P.H. Emmett, E. Teller, J. Am. Chem. Soc. 60 (1938) 308.
[24] M. Kruk, M. Jaroniec, J. Choma, Carbon 36 (1998) 1447.
25] S.J. Gregg, Adsorption Surface Area and Porosity, second ed., Academic Press
Inc., New York, 1982.
26] A. Corma, V. Fornes, L. Forni, F. Marquez, J. Martinez-Triguero, D. Moscotti, J.
Catal. 179 (1998) 451.
27] Database of Zeolite Structures. <http://www.iza-structure.org/databases/>.
28] S. Scholz, J.A. Lercher, Chem. Mater. 23 (2011) 2091.
29] N. Tøpsoe, K. Pedersen, E.G. Derouane, J. Catal. 70 (1981) 41.
lication, dealumination, and subsequent SiO
2
overlayer deposition.
Increasing the relative concentration of these species by the SiO
2
overlayer reduced the toluene conversion rate, because the higher
concentration of bulky aromatic products (increase by 30%) led to a
decreased concentration of toluene and adsorbed methanol at the
active site as well as a lower probability of alkylating toluene. By
shortening the effective diffusion length via desilication/dealumi-
[
[
[
[
[
2
nation (decrease by 63%) prior to SiO deposition, the steady-state
concentration of bulky aromatic molecules in the pores could be
decreased. This new material increased p-xylene selectivity and
toluene turnover rates from 64% to 78% and by ꢀ20%, respectively
[30] G. Qin, L. Zheng, Y. Xie, C. Wu, J. Catal. 95 (1985) 609.
[
31] G. Lietz, K.H. Schnabel, C. Peuker, T. Gross, W. Storek, J. Volter, J. Catal. 148
1994) 562.
32] K. Yamagishi, S. Namba, T. Yashima, J. Phys. Chem. 95 (1991) 872.
(
[
(
and maintained the catalytic activity in per weight of zeolite ba-
[33] M. Kustova, M.S. Holm, C.H. Christensen, Y. Pan, P. Beato, T.V.W. Janssens, F.
Joensen, J. Nerlov, Stud. Surf. Sci. Catal. 174 (2008) 117.
sis), however, with the drawback of a slightly lower selectivity of
methanol usage.
[
[
[
34] D. Ma, F. Deng, R. Fu, X. Han, X. Bao, J. Phys. Chem. B 105 (2001) 1770.
35] R.A. Shaikh, S.G. Hegde, A.A. Behlekar, B.S. Rao, Catal. Today 49 (1999) 201.
36] J.P. Marques, I. Gener, P. Ayrault, J.C. Bordado, J.M. Lopes, F. Ramôa Ribeiro, M.
Guisnet, Micropor. Mesopor. Mater. 60 (2003) 251.
Acknowledgments
[37] E.P. Parry, J. Catal. 2 (1963) 371.
[
38] L.H. Ong, M. Dömök, R. Olindo, A.C. van Veen, J.A. Lercher, Micropor. Mesopor.
Mater. 164 (2012) 9.
The authors acknowledge the support from the Ministry of
Higher Education, Saudi Arabia for the establishment of the Center
of Research Excellence in Petroleum Refining and Petrochemicals
at King Fahd University of Petroleum and Minerals (KFUPM). The
Deutsche Forschungsgemeinschaft (DFG Project JE 260-10/1) and
the ‘‘Fonds der Chemischen Industrie’’ are gratefully acknowledged
for financial support (R.K.). The authors would also like to thank
Gary Haller, Jeongnam Kim, Mirek Derewinski, Oliver Gutiérrez,
and Stefan Schallmoser for helpful technical discussions, Xaver
[
39] J.H. Ahn, R. Kolvenbach, S.S. Al-Khattaf, A. Jentys, J.A. Lercher, ACS Catal. 3
(2013) 817.
40] P.B. Weisz, Science 178 (1973) 433.
[
[
41] J.H. Ahn, R. Kolvenbach, S.S. Al-Khattaf, A. Jentys, J.A. Lercher, Chem. Commun.
(
2013), http://dx.doi.org/10.1039/c3cc46197a.
[42] G. Mirth, J.A. Lercher, J. Catal. 132 (1991) 244.
[
[
43] D. Lesthaeghe, B. De Sterck, V. Van Speybroeck, G.B. Marin, M. Waroquier,
Angew. Chem. Int. Ed. 46 (2007) 1311.
44] D.M. McCann, D. Lesthaeghe, P.W. Kletnieks, D.R. Guenther, M.J. Hayman, V.
Van Speybroeck, M. Waroquier, J.F. Haw, Angew. Chem. Int. Ed. 47 (2008)
5179.
[
[
45] S. Arrhenius, Z. Phys. Chem. 4 (1889) 226.
46] S. Svelle, S. Kolboe, O. Swang, U. Olsbye, J. Phys. Chem. B 109 (2005) 12874.
2
Hecht for N physisorption experiments, Martin Neukamm for
AAS measurements, and Axel Dreier for TEM images.
[47] I.I. Ivanova, A. Corma, J. Phys. Chem. B 101 (1997) 547.
[
[
[
48] U. Olsbye, M. Bjørgen, S. Svelle, K. Lillerud, S. Kolboe, Catal. Today 106 (2005)
08.
49] D. Lesthaeghe, A. Horré, M. Waroquier, G.B. Marin, V. Van Speybroeck, Chem.
Eur. J. 15 (2009) 10803.
1
Appendix A. Supplementary material
50] G. Mirth, J.A. Lercher, J. Catal. 147 (1994) 199.
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.jcat.2013.12.003.