Effect of extra-framework Al formed by successive steaming and acid leaching of zeolite MCM-22 on its structure and catalytic performance

Article abstract of DOI:10.1016/j.apcata.2011.12.029

Dealuminated MCM-22 samples have been prepared by a two-step dealumination procedure. Detailed assessment of the properties of the materials obtained at each one of the successive stages, i.e. steaming (at 500 °C, 700 °C and 800 °C) and acid reflux (HCl and oxalic), has been made by XRD, N ₂ adsorption-desorption, m-xylene adsorption, ²⁷Al MAS NMR and FT-IR of pyridine adsorption. It was found that steaming generates extra-framework aluminum (EFAl) species and the majority of them cannot be extracted by the consecutive acid leaching. These extra-lattice entities block the zeolite micropores which makes the remaining Broensted acid sites isolated and inefficient. It is shown that the presence of such species vastly affects the catalytic performance of zeolite MCM-22 in the reaction of m-xylene conversion. The consequences are reduced adsorption capacity and catalytic activity, modified reaction products distribution, enhanced p-xylene selectivity, as well as altered mode of coke formation and composition of the coke precursors.

Full text of DOI:10.1016/j.apcata.2011.12.029

Applied Catalysis A: General 417–418 (2012) 76–86
Contents lists available at SciVerse ScienceDirect
Applied Catalysis A: General
journal homepage: www.elsevier.com/locate/apcata
Effect of extra-framework Al formed by successive steaming and acid leaching of
zeolite MCM-22 on its structure and catalytic performance
a
a
b
c
d,∗
R.M. Mihályi , M. Kollár , P. Király , Z. Karoly , V. Mavrodinova
a
Institute of Nanochemistry and Catalysis, Chemical Research Center, Hungarian Academy of Sciences, Pusztaszeri út 59-67, Budapest 1025, Hungary
Institute of Structural Chemistry, Chemical Research Center, Hungarian Academy of Sciences, Pusztaszeri út 59-67, Budapest 1025, Hungary
Institute of Materials and Environmental Chemistry, Chemical Research Center, Hungarian Academy of Sciences, Pusztaszeri út 59-67, Budapest 1025, Hungary
Institute of Organic Chemistry with Centre of Phytochemistry, Bulgarian Academy of Sciences, Acad. G. Bonchev str., bl. 9, 1113 Sofia, Bulgaria
b
c
d
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 21 October 2011
Received in revised form 1 December 2011
Accepted 16 December 2011
Available online 27 December 2011
Dealuminated MCM-22 samples have been prepared by a two-step dealumination procedure. Detailed
assessment of the properties of the materials obtained at each one of the successive stages, i.e.
◦
◦
◦
steaming (at 500 C, 700 C and 800 C) and acid reflux (HCl and oxalic), has been made by XRD, N2
27
adsorption–desorption, m-xylene adsorption, Al MAS NMR and FT-IR of pyridine adsorption. It was
found that steaming generates extra-framework aluminum (EFAl) species and the majority of them can-
not be extracted by the consecutive acid leaching. These extra-lattice entities block the zeolite micropores
which makes the remaining Broensted acid sites isolated and inefficient. It is shown that the presence
of such species vastly affects the catalytic performance of zeolite MCM-22 in the reaction of m-xylene
conversion. The consequences are reduced adsorption capacity and catalytic activity, modified reaction
products distribution, enhanced p-xylene selectivity, as well as altered mode of coke formation and
composition of the coke precursors.
Keywords:
MCM-22 zeolite
Steam-acid dealumination
Extra-framework Al
m-Xylene conversion
Coke formation
©
2011 Elsevier B.V. All rights reserved.
1
. Introduction
such as ZSM-5 or other structures induces rather limited meso-
porosity [6–8].
Dealumination of zeolite has been studied for several decades
The literature survey reveals that oxalic acid extracts the alu-
minum expelled upon previous hydrothermal treatment only from
large-pore but not from medium-pore zeolites because of diffusion
limitations [9,10]. Mordenite is suggested to imply diffusion prob-
lems too in removing, by mineral acid treatment, the EFAl from
its one-dimensional pore structure [11]. It is observed by Fernan-
des et al. [12] that not only in mordenite but also for zeolite beta,
steam dealumination generates highly polymerized EFAl which can
be partially removed by the subsequent acid leaching.
The particular structure of the zeolite MCM-22 holds out inter-
esting opportunities for studying the process of dealumination and
the consequences of this procedure for its catalytic performance
[13]. This zeolite is composed of two channel systems, both of them
accessible through narrow 10-MR windows [14]. The material con-
tains at least five crystallographically non-equivalent T-sites and
they are affected to different extents by the mineral acid dealu-
mination [13]. Partial dealumination takes place still during the
calcination of this lamellar structure [15]. Dealumination is more
but recently the interest towards this modification method has
been renewed as one of the several top down approaches for meso-
pore formation and creation of enhanced Broensted acid sites, both
crucial factors in determining the activity and selectivity in the
transformations of aromatic hydrocarbons on zeolite catalysts [1].
Among the typical examples for such procedure, giving rise to a
very significant catalyst improvement, are the well-known ultra-
stable Y (USY) family materials featuring mesopores [2] which are
obtained by dealumination via steam treatments and are being
widely used in the petrochemical industry. The effectiveness of
dealumination by steaming and/or acid leaching depends on both
the severity of the procedure and the zeolite structure. During this
treatment extra-framework aluminum species (EFAl) are generally
formed, which could be removed from the lattice, and secondary
mesoporosity might be created [3]. The non-framework aluminum
generated by hydrothermal treatment can be easily extracted by
the subsequent acid leaching but only in case of three-dimensional,
large-pore zeolites [4,5]. Dealumination of more siliceous zeolites,
severe on heating in air than in vacuum or N , and may occur even
2
during grinding in a mortar [16].
Studies on hydrothermally dealuminated MCM-22 zeolites
show that aluminum released from the lattice mainly stays in
the internal pore systems and then condenses into polymeric alu-
minum species, which causes a decrease in the accessibility of the
∗ Corresponding author. Tel.: +359 2 979 3961; fax: +359 2 8700225.
E-mail addresses: vmavrodinova@orgchm.bas.bg,
vesselinamavrodinova@yahoo.com (V. Mavrodinova).
0
926-860X/$ – see front matter © 2011 Elsevier B.V. All rights reserved.
doi:10.1016/j.apcata.2011.12.029

R.M. Mihályi et al. / Applied Catalysis A: General 417–418 (2012) 76–86
77
acid sites to reactant molecules [17–20] and makes its channel sys-
2. Experimental
tem tighter by deposition of octahedral aluminum debris in the
pores [21,22].
2.1. Catalysts preparation
Dealumination of MCM-22 only with mineral acids seems to be
a complex process. According to Matias et al. [13], first a rapid dis-
solution of the EF aluminum hydroxylated species formed upon
calcination takes place, followed by slower extraction of the lat-
tice Al. Higher diffusion limitations induced by the remaining
EFAl species, not removed from the inner pore system of MCM-
The MCM-22 precursor was prepared by hydrothermal syn-
3
thesis applying rotating industrial autoclave (∼1 m ) according
to the procedure described in Ref. [30]. The composition of the
synthesis gel was SiO :0.036Al O :0.024Na O:0.265Na SO :0.517
2
2
3
2
2
4
hexamethyleneimine:42.6H O. The crystallization temperature
2
2
2 zeolite, are proposed by these authors. It appears that, for the
and time were 418 K and 10 days respectively, and sample with
Si/Al molar ratio of 16 was obtained. After washing with distilled
water and drying at 343 K, the preparation was calcined in air (pro-
grammed heating, 1.5 K/min up to 823 K for 3 h) in order to remove
confined channel system of MCM-22, the ability for dealumina-
tion strongly depends on the size of the applied acid molecule.
Xia et al. [23] found differences in the effectiveness in dealumi-
nation using citric or oxalic acid, in favor of the ability of the
smaller oxalic anion to enter easier into the zeolite channels.
According to Unverricht et al. [24] the treatment of MCM-22 with
the template. Twofold ion exchange with 1 M NH NO3 solution
4
was applied for the preparation of the ammonium form. The H-
form of the sample was obtained by in situ decomposition of the
NH4+-forms in N2 at 823 K and is indicated as MCM-22 (Table 1).
The dealumination procedure was carried out by steam treatment
1
4.5 nHNO3 even for 30 h does not result in a significant dealu-
mination. The possible reason is the high stability of this zeolite
against acid attack. Oxalic acid reflux of H-MCM-22 causes par-
tial removal of the external acid sites according to Ren et al.
◦
◦
◦
at 773 K (500 C), 973 K (700 C) or 1073 K (800 C) (samples des-
ignated as D500, D700 and D800) followed by either 2 M HCl
(D700/HCl and D800/HCl) or 0.5 M oxalic (D500/Ox and D700/Ox)
acid reflux at 373 K (Table 1).
[
25].
The peculiarities of the successive steam-acid dealumination
of zeolite MCM-22 are not investigated in details. Neither the
change in the concentration and nature of the EFAl species formed
and remained at each stage of this procedure, nor their role
in the catalytic performance, are clarified sufficiently. There are
only few studies on subsequent steam-acid dealuminated MCM-
2.2. Catalysts characterization
The structure and crystallinity of the samples were verified by
a Philips PW 1810 powder diffractometer. Adsorption isotherms
were determined by N2 sorption at 77 K using a Quantachrome
NOVA Automated Gas Sorption instrument. The specific surface
2
2 zeolite [26–29]. With the exception of the contribution of Park
et al. [26] who do not find any promotion effect of the steam-
acid dealuminated MCM-22 for the reaction of methylation of
area was calculated from the N adsorption capacity by the multi-
2
2
-methylnaphtalene, other researchers [27–29] ascertain that the
ple BET method. The micropore volumes were estimated from the
alpha-s plots. The acidity of the materials was determined by FT-
IR (Nicolet Impact Type 400) spectroscopy as well as from the NH4
exchange capacity measured by temperature-programmed ammo-
nia evolution in the interval 453–923 K (Table 1).
activity of the as-modified catalysts in the reactions of alkylaro-
matic transformations (toluene and ethylbenzene) is reduced but
the selectivity towards the p-substituted aromatics is enhanced.
Presence of EFAl species is presumed only by Wu et al. [29] but
the consequences of this presence are neglected. Analysis of the
content and composition of these entities has not been made and
their effect on the catalytic performance in the examined reactions
has not been mentioned in any of the above contributions [27,28]
but the one of Wu et al. [29]. The conclusion of the only authors
that study steam-acid dealuminated MCM-22 in the reactions of
toluene and ethylbenzene conversion is that aluminum is hardly
removable from this zeolite type, but the formed EFAl species
are highly dispersed within the crystallites, without blocking the
micropores and without affecting at all the catalytic performance
The Gas Chromatographic Pulse Injection (GCPI) technique for
m-xylene adsorption consisted of introduction of m-xylene vapor
(0.16 kPa) in N2 (60 ml/min) passing through a saturator equili-
◦
brated at 8 C and introduced in a micro-flow adsorption bed loaded
with 90 mg catalyst [31]. The breakthrough time and the relative
◦
m-xylene uptake at 40 C are determined by continuously moni-
toring the exit m-xylene stream by means of FID detector of the GC
unit.
Solid-state 27Al MAS NMR experiments were carried out on Var-
ian NMR System spectrometers at Larmor frequency of 104.4 MHz
(B = 9.4 T) and 156.4 MHz (B = 14.1 T). Samples were packed into
[
27,29].
The last statement remains questionable considering the very
0
0
either 4.0 mm or 3.2 mm zirconia rotors, respectively. The magic
recent studies on acid leached [13,23] or only hydrothermally
treated [21–23] MCM-22 zeolites. The thorough 27Al MAS NMR
investigations implemented there evidence the effect of the pres-
ence, nature and variations of the extra-framework aluminum
detected after dealumination [21–23] on the reaction selectivity
in different catalytic transformations [13,23].
In the present work two-step dealumination procedure has
been performed. At the first step the samples were hydrothermally
treated, and then oxalic or hydrochloric acid reflux was applied
in order to extract the EFAl species and to generate, eventually, a
secondary mesopore channel system. The structural changes and
the alteration in the catalytic performance occurred at each dea-
lumination step have been thoroughly investigated. Our objective
was to prove whether any mesoporosity can be generated at all
by successive steam-acid dealumination of MCM-22 and to assess
the effect of the EFAl generated within the pore structure of these
materials on their catalytic activity and selectivity to isomerization
and disproportionation of m-xylene as well as on the mode of coke
formation.
angle spinning frequency (ꢀMAS: 13 kHz or 16 kHz) was set to avoid
the spectral overlap of spinning sidebands. Al chemical shifts are
27
referred to Al(NO ) (aq) using NaAlO (ı27Al = 79.3 ppm) as sec-
3
3
2
ondary reference [32]. Fine calibration of the single pulse excitation
experiment was carried out in case of the sample MCM-22 (opti-
mal parameters: 2.5 ␮s excitation pulse and 5 s relaxation delay).
The total experimental time was approximately 16 h for each sam-
ple. The spectral width was set to 625 kHz and 81920 complex
points were acquired. The spectra were zero filled to 256,000 points
and multiplied by an exponentially decaying function (lb = 20 Hz).
The backward linear prediction processing method [33] was used
to get rid of the baseline distortions. Quantitative analysis of the
aluminum species was carried out by using the relative integrals
of the relevant spectral regions. The precision of the method was
increased by averaging the data determined from three spectra for
each sample (number of scans: 5120, 10,240 and 15,360, respec-
tively).
Quantitative analysis of the 27Al MAS NMR spectra of the
parent and steam-acid dealuminated materials were significantly

7
8
R.M. Mihályi et al. / Applied Catalysis A: General 417–418 (2012) 76–86
Table 1
Aluminum content, ion-exchange capacity and textural properties of the samples.
IECb,d (mmol/g)
Surface areab,e (m /g)
2
Micropore volume
b,f
Sample
Dealumination method
Aluminum content
3
(
cm /g)
Steaming
–
Acid refluxa
ꢁAlb (mmol/g)
Proportion of the
Al speciesc (%)
MCM-22
D500
D500/Ox
–
–
1.08
1.09
0.69
0.66
–
1.13
0.63
0.66
1.14
0.65
95/–/5
0.96
0.43
0.36
–
578
534
538
–
–
515
–
521
–
481
0.18
0.15
0.15
–
–
0.15
–
0.15
–
0.14
◦
500 C, 3 h
45/14/41
51/8/41
74/4/22
53/7/40
38/21/41
48/14/38
59/8/33
35/21/44
47/12/41
◦
500 C, 3 h
Oxalic, 24 h
–
–
–
D500/Ox NH4+-form
D500/Ox H-form
D700
–
–
–
◦
◦
◦
◦
700 C, 3 h
700 C, 3 h
700 C, 3 h
800 C, 3 h
0.26
0.19
0.23
0.19
0.19
D700/Ox
D700/HCl
D800
Oxalic, 24 h
HCl, 3 h
−
◦
D800/HCl
800 C, 3 h
HCl, 3h-
a
The acid reflux was carried out 373 K using either 0.5 M H2C2O4 or 2 M HCl.
All data were related to 1 g of sample, calcined at 973 K.
b
c
The proportion of the different aluminum species were determined by integrating the 27Al MAS NMR spectra (see Fig. 2). The sequence is: tetrahedral/5-coordinated or
distorted tetrahedral/octahedral environments, respectively.
d
The ion-exchange capacity (IEC) of the samples, obtained as the amount of ammonia, evolved from the NH4+ form sample in the 453–923 K temperature range during
the TPAE run.
e
Calculated by the multiple BET method from the N2 adsorption isotherms.
Estimated from the alpha-s plots.
f
influenced by the relaxation delay, if shorter than 5 s. The relative
intensities of the resonances corresponding to the different coordi-
nation environments (Al species) were determined at 14 T in order
to minimize the major source of inaccuracy because of the inability
to detect all the aluminum species at low magnetic field [34].
from lattice positions. The authors found out that the XRD pattern
of ITQ-1, the pure silica analog of MCM-22, was better resolved
than those of Al-MCM-22 and indicated an improved crystallinity.
We suggest the same effect for our dealuminated materials with the
exception of the most severely heated D800 and D800/HCl samples
(Fig. 1). Some amorphization may have occurred as a result of the
2
.3. Catalytic studies
higher steaming temperature in this case. No diffraction peaks due
to another crystalline phase were observed for our steamed and
steamed/acid-leached materials.
The test reaction has been carried out in a fixed-bed flow reac-
tor at atmospheric pressure, reaction temperatures of 523 K and
23 K and various contact times. N carrier gas was passed through
a saturator filled with m-xylene and equilibrated at 293.2 K so that
partial pressure of 0.9 kPa to be attained. On-line analysis of the
reaction products has been performed using HP–GC with 25 m FFAP
capillary column.
Wu et al. [29] also reported comparable or even higher intensity
of the XRD reflections of their steam-acid dealuminated materials
compared to the corresponding parents.
6
2
The chemical composition, ion-exchange capacity and the tex-
tural properties of the parent and dealuminated samples are
presented in Table 1. The difference between the total aluminum
content (1.08 mmol/g) and the IEC (0.96 mmol/g) of the initial
material indicates that about 11% of Al is in extra-framework posi-
tions. The amount of aluminum in the steamed samples D500, D700
and D800 remains similar to that of the parent material but their
ion-exchange capacity decreases significantly by 60%, 77% and 83%,
respectively. The results indicate that substantial portion of the Al
2.4. Coke determination
Thermogravimetric microbalance (Setaram TG 92) experiments
described in Ref. [35] were used for determination of the adsorbed
carbonaceous species on the spent catalysts. The preparation
procedure of the used catalysts included the following. The cat-
alytic experiments were aborted after chosen time period (usually
1
20 min), the volatiles were eliminated by heating in N for 30 min
500
2
at the reaction temperature. Then the samples were cooled down
D800/HCl
still in N , taken out from the reactor, placed in the microbalance
2
and heated first in Ar by raising the temperature (10 K/min) up to
D800
8
23 K. Then the flow of Ar was exchanged to air flow and the aged
catalyst was left for another 1 h at 823 K. The registered weight loss
corresponds to the amount of the intermediates and products des-
orbed up to 823 K in Ar, the so called “light coke”, and to that of the
strongly held “hard coke” removed in air at 823 K.
D700/HCl
D700/Ox
D700
3
. Results and discussion
D500/Ox
3.1. Physicochemical characteristics of the samples
D500
MCM-22
The X-ray diffraction patterns (Fig. 1) show that between 5 and
1
5
5 2ꢂ the intensity of the reflections for the samples steamed at
00 C and 700 C (D500 and D700) and their acid-treated varieties
5
10
15
20
25
30
35
40
45
50
◦
◦
Bragg angle (2Θ)
(
D500/Ox, D700/Ox and D700/HCl) is higher than that of the par-
ent MCM-22 material. The enhanced intensity of these reflections,
according to Camblor et al. [36], can be due to the removal of Al
Fig. 1. XRD patterns of the parent, steamed and steamed/acid-leached MCM-22
samples.

R.M. Mihályi et al. / Applied Catalysis A: General 417–418 (2012) 76–86
79
intensity decrease of the resonances at the 70–50 ppm region. The
above reversible change of octahedrally coordinated aluminum to
tetrahedral species suggest that three-fold coordinated framework
−
◦
Al atoms with SiO defect sites could be formed upon hydrothermal
treatment at 500 C (see below in Scheme 1, route II).
Calculations on the relative intensity of the signal of the peaks
corresponding to the different Al species present in the steamed
only and in the acid-refluxed materials were made. The results
reveal (Table 1) that the various types of EFAl species formed dur-
ing steaming and present in differing proportions of both essential
kinds (in accordance to Fig. 2 and Ref. [40]) are only partially
removed by acid treatment. From all dealuminated materials, the
sample steamed at the lowest temperature and then refluxed with
oxalic acid (D500/Ox) possesses the highest portion of tetrahedrally
coordinated Al and the lowest percentage of extra-lattice alu-
minum. The most severe steaming generates the highest amount
of EFAl on the expense of lattice aluminum which cannot be
removed out of the zeolite structure by the subsequent acid leach-
ing (D800/HCl). As a whole, about 10–20% of the aluminum atoms
in the steamed materials that have been in tetrahedral positions
restore their tetrahedral coordination after the acid reflux. This
occurs on the expense of the 5-coordinated or distorted tetrahe-
dral Al atoms, since the portion of the octahedrally coordinated
Al remains practically unchanged (Table 1). The nonframework
species, not able to be extracted by the acid (whatever they are),
may: (i) compensate the framework charge of the zeolite resulting
in a decrease in Broensted acid site concentration [41], (ii) interact
with part of the structural Al and consequently reduce their acid
strength [42] or, finally, (iii) hinder the access to the effective acid
sites by blocking the pores, as Xia et al. [23] have proposed. Each one
of these possibilities seems probable as well for our steam-acid dea-
luminated materials, considering the consensus of all investigators
on the presence of extractable EFAl in the lattice of the dealumi-
nated MCM-22 zeolites, regardless of the applied procedure for
their preparation.
Fig. 2. Central region of the 27Al MAS NMR spectra of the parent, steamed and acid-
treated MCM-22 samples. The value of magnification is indicated at the right side
of each spectrum. The relative intensities of the resonances are given in Table 1.
atoms has been removed from tetrahedral positions, the higher the
more rigorous the hydrothermal treatment has been.
The subsequent acid leaching results in a similar decrease, by
about 40%, of the total Al content for all acid-treated materials.
The results suggest that both extra-lattice and tetrahedrally coor-
dinated aluminum atoms have been expelled from the zeolite
structure upon the acid reflux (Table 1) and the total dealumination
degree does not depend so much on the acid used.
3
.1.2. FT-IR studies
In order to obtain information about the type of the acid sites in
3
.1.1. 27Al MAS NMR
Solid state 27Al MAS NMR spectroscopy has become an effi-
the parent, the steamed and the subsequently acid-treated materi-
als, FT-IR spectra before and after Py adsorption were made (Fig. 3).
The spectra in the hydroxyl stretching region show (Fig. 3A), that
steaming results in a strong reduction (D500) or even in the disap-
pearance (D700 and D800) of the band attributed to bridging OH
groups (3620 cm ). Simultaneously, an increase in the intensity
of the external Si OH groups (3745 cm ) is observed. The follow-
ing acid treatment step leads to an increased concentration of the
internal silanols (3727 cm ) on the expense of the external Si OH
cient method to determine the coordination state of the aluminum
species in zeolites. The non-equivalent positions of identical
aluminum coordination states may strongly overlap, but the reso-
nances of tetrahedral and octahedral environments are sufficiently
resolved to determine the Al(IV)/Al(VI) ratio, respectively [36].
−
1
The 2 Al MAS NMR spectra of the parent and the steamed and
steamed/acid-leached materials are shown in Fig. 2. Intense res-
onances at the high frequency region (ca. 70 → 50 ppm) were
observed in case of all samples. These overlapping resonances are
assigned to charge-balanced, tetrahedrally coordinated framework
aluminum atoms. The gradual decrease and broadening of these
signals is also accompanied by the appearance of another slight res-
onance at ca. 30 ppm (in particular for D700 and D800) attributed
to either pentahedral or tetrahedrally-distorted Al species (frame-
7
−1
−
1
−
1
−1
(3745 cm ) and to the appearance of new bands at 3700 cm and
−
1
at 3527 cm in case of the acid-leached samples (Fig. 3A).
The band at 3727 cm 1 is also registered by Meloni et al. [43]
in MCM-22 materials with higher Si/Al ratio and corresponds to
the “hydroxyl nests” detected on steamed [21], acid-leached [13]
or steam and acid-treated [29] zeolites of the MCM-22 type. In our
−
n+
−1
work Al interacting with Al ) [38–40], all they generated as a
case the intensity of the band at 3727 cm is the highest for the
result of steaming and slightly reduced after the acid reflux. A
very weak resonance at ca. 0 ppm can be detected for the cal-
cined parent zeolite (Fig. 2, MCM-22). Upon steaming at 500 C a
broad and a sharp resonance appear in the spectrum (Fig. 2, D500).
The broad resonance is assigned to polymeric aluminum species
most severely steam-treated D800/HCl.
−
1
−1
The new bands appeared at 3700 cm and 3527 cm after acid
reflux of the steamed samples are ascribed by Corma at al. [15] and
Meloni et al. [43], to the particular, also internal silanols (defect
sites), located preferentially in the supercages and 10 MR sinusoidal
channels, respectively. Their concentration is again the highest for
◦
[
(
21,23], which cannot be removed by the following acid treatment
Fig. 2, D500/ox).
−
1
D800/HCl. The broad band around 3500 cm detectable in the FT-
IR spectra of D700/Ox, D700/HCl and D800/HCl (Fig. 3A) is also
observed in ZSM-5 zeolite and is ascribed to silanols interacting
through hydrogen bonding [44,45]. The strong intensity reduc-
The sharp line at 0 ppm disappeared completely upon liquid
phase ammonium ion exchange, whereas the resonances due to
+
tetrahedral Al increased (Fig. 2, D500/ox NH4 -form). Transfor-
+
−1
mation of this NH4 -form to H-form by thermal deammoniation
tion of the 3620 cm band attributed to the bridging OH groups
results in the restoration of the sharp line at 0 ppm and the
or its complete disappearance for the more severe steam-treated

8
0
R.M. Mihályi et al. / Applied Catalysis A: General 417–418 (2012) 76–86
Scheme 1.
materials could be ascribed to the removal of practically all tetra-
hedral Al atoms. This observation is in contrast to our 27Al MAS
NMR results (Fig. 2) that proved the presence of tetrahedral Al in
all dealuminated materials. Our explanations for this contradiction
are that, as a result of the dealumination, (i) part of the bridg-
ing OH groups was neutralized by cationic extra-lattice Al species
zeolite. The intensity of these Lewis-connected Py bands is a bit
higher for D500 and decreases with the steaming temperature. Acid
treatment leads to partial extraction of this type of adsorption sites
(D500/Ox) or to a change in their distribution (D700/HCl/Ox and
D800/HCl) (Fig. 3B).
FT-IR spectra of Py adsorbed on only acid-leached MCM-22
zeolites [13] and on steamed mordenites [12] reveal much lower
density of acid sites than the concentration of lattice Al atoms in
their structure. The authors ascribe this phenomenon to a con-
finement effect within the micropores caused by the EFAl entities
deposited during the dealumination procedure which restrict the
access of the Py molecules to the zeolite acid sites. The same effect
cannot be ruled out in our case, too.
(
route I in Scheme 1) and (ii) the remained tetrahedral framework
Al was transformed to tri-coordinated lattice aluminum (route II in
Scheme 1), in accordance to the mechanism suggested by Altwasser
et al. [46].
The authors also assume that NH3 can coordinate to
these threefold-coordinated Al atoms and heals the framework
(
−)
Si O Al bridges (Scheme 1). The latter suggestion explains why
our steamed as well as acid-treated materials still possess some
ion-exchange capacity (Table 1), but in the same time the bridging
hydroxyls are practically missing in their FT-IR spectra.
The difficulty in removing these EFAl species out of the zeolite
pores can be explained by the 27Al MAS NMR results of Meriaudeau
et al. [21] and Ma et al. [22] on hydrothermally treated MCM-22
materials. These authors claim that deeper steam dealumination
occurs at higher temperatures and the aggregation and condensa-
tion of the expelled octahedral extra-framework aluminum atoms
into polymeric aluminum species takes place at vigorous steaming.
In case of the initial sample pyridine adsorption causes the
−1
disappearance of the band at 3620 cm
appearance of the characteristic pyridinium band at 1543 cm
Fig. 3B). For the steamed materials the intensity of the pyridinium
(not shown) and the
−
1
(
◦
◦
band is significantly lower than that of the initial MCM-22 sam-
ples, and further reduces for the samples undergone acid leaching.
Py adsorption also leads to a small intensity decrease of the bands
If such species have been formed at 700 C and 800 C compared to
◦
500 C in our case, then they are supposed to be eliminated with
more difficulty by the next acid reflux, as our experiments indeed
have shown. Milder steaming generates not aggregated [21] and
monomeric Al species [22] that should be more easily extractable
by the subsequent acid treatment.
−
1
−1
.
at 3700 cm and at 3727 cm
Strong bands attributed to Py coordinated to Lewis acid sites
−1
−1
and 1446 cm in case of the parent
also appear at 1453 cm
3
620
0.1
1543
A 3
745
B
0.1
1
453
1446
1437
MCM-22
D500
D500/Ox
MCM-22
3
727
D500
D500/Ox
D700
3
700
3527
D700/Ox
D700
D700/HCl
D800
D700/Ox
D700/HCl
D800
D800/HCl
D800/HCl
3
800
3700
3600
3500
3400
1550
1500
1450
-
1
-1
Wavenumber (cm )
Wavenumber (cm )
Fig. 3. FT-IR spectra of the parent and dealuminated samples (A) after evacuation at 673 K for 1 h and (B) after Py adsorption at 473 K and evacuation at 473 K for 0.5 h.

R.M. Mihályi et al. / Applied Catalysis A: General 417–418 (2012) 76–86
81
2
2
2
1
1
1
1
1
1
1
1
1
1
20
10
00
90
80
70
60
50
40
30
20
10
00
7000
MCM-22
D700/HCl
D800/HCl
D500/ox
D500
6000
5
4
000
000
D700
(
p/p )
o
3000
0.0 0.2 0.4 0.6 0.8 1.0
800
MCM-22(I)
D800/HCl
D700/HCl
D500/Ox
2
1
000
000
0
600
400
200
D500/ox
D700/HCl
D700/Ox
0.0
0.2
0.4
0.6
0.8
1.0
0
10
20
30
40
50
Relative pressure (p/po)
Time (min)
Fig. 4. N2 adsorption and desorption isotherms of the parent MCM-22 zeolite, the
Fig. 5. Uptake curves of m-xylene adsorption at 313 K over the parent MCM-22 and
steamed and the consecutively steam-acid dealuminated zeolites.
the steam-acid dealuminated materials.
Results suggest that steam treatment of MCM-22 leads to reduc-
tion of the framework Al atoms and formation of octahedrally and
pentahedrally coordinated EFAl species. The successive acid treat-
ment results in additional but only partial (40%) extraction of both
framework and extra-framework aluminum.
3.1.3.2. m-Xylene uptake. On Fig. 5 the m-xylene uptake over the
initial and the steam-acid dealuminated materials in dependence
on the adsorption time measured by GCPI technique was presented.
As the figure reveals, the adsorption capacity of the dealuminated
samples amounts to about 62–70% of that of the parent MCM-22
depending on the severity of the dealumination procedure. Same
sequence in the capacity decrease for N and m-xylene adsorption
3.1.3. Adsorption studies
2
is found on the steamed and acid treated samples but the uptake
decrease for the bigger m-xylene molecule is larger than that of
nitrogen. The most rigorously steam-acid dealuminated D800/HCl
The proposal about pore blockage by EFAl species in zeolite
MCM-22 treated by nitric acid only have been presumed before by
Matias et al. [13] but declined back by the authors themselves with
the argument that “. . .such proposal could seem surprising when
the nitrogen adsorption data are considered”. This conclusion is
based on the absence of change in the micropore volume after the
acid treatment of the parent zeolite. The same argument against
the possibility of micropore blocking by EFAl species is given by
Wu et al. [29] considering the slight changes in the micropore vol-
ume of the steam-acid dealuminated materials compared to the
parent one. In both contributions, however, the micropore volume
is measured not with an adsorbate with a similar size of the reactant
displays the lowest adsorption capacity for both N and m-xylene.
2
Diffusional barriers caused by the EFAl species remained deposited
after both stages of dealumination should be responsible for the
observed reduction in the adsorption capacity.
Analogously, reduced capacity towards n-hexane and more
markedly for the bigger cyclohexane molecule has been observed
by Ding et al. [47] on only steamed ZSM-5 compared to the starting
material. The authors also found lower total pore volume as well as
lower micropore area and volume for the hydrothermally treated
samples but they attribute this effect to the narrowing of the zeolite
channels as a result of the replacement of the longer Al O bonds
by the shorter Si O bonds.
molecule but with small N molecule [13]. Not to neglect this very
2
important consideration, adsorption experiments not only with N2
but also with the m-xylene reactant have been performed in the
present investigation.
3.2. m-Xylene transformation
3.1.3.1. N2 adsorption. In Fig. 4 the N2 adsorption and desorp-
tion isotherms of the parent, steamed and steamed/acid-leached
samples are presented. As the data show, the decrease in the micro-
pore volume for the mildly steam-acid treated samples (D500,
D500/Ox) determined by N2 adsorption is about 17% (Table 1).
The catalytic properties of the parent material were compared
to that of its dealuminated modifications obtained at each dealu-
mination steps, steaming and subsequent acid reflux, respectively.
Whatever the reaction temperature and the contact time applied,
the products expected from m-xylene isomerization (p- and o-
isomer) and disproportionation (toluene and trimethylbenzenes)
are observed. The conversion on the steamed at 773 K and 973 K
and their acid-treated modifications (Fig. 6A and B, respectively)
together with the parent, are presented as a function of the time on
stream (TOS). In accordance with the loss of framework Al atoms
and respectively of Broensted acid sites (IEC, FT-IR and 27Al MAS
NMR data) the samples reveal decreasing catalytic activity after
the consecutive steps of steaming and acid leaching, compared to
the parent material. The activity of the more severely steamed cat-
alysts as well as those of their acid-leached modifications is much
lower.
The micropore volume of the sample hydrothermally treated at
◦
8
00 C is the lowest (Fig. 4). Comparing the isotherm of the steamed
(
D500, D700, D800) and subsequently acid-treated samples, it
can be concluded that upon acid reflux the micropore volume
of the samples practically does not change. Results suggested
that the lower N2 adsorption capacity of the dealuminated zeo-
lites is not due to the loss of crystallinity considering the even
higher intensity of the XRD reflections of the dealuminated sam-
ples (with the exception of D800/HCl, Fig. 1), but rather to the
presence of EFAl entities inside the pores, causing some adsorp-
tion limitations even for the small N2 molecule. BET surface area
decreases also, not substantially, as a result of dealumination
(
Table 1). The absence of hysteresis of the isotherms (inset in
The vastly reduced catalytic activity of the dealuminated mate-
rials compared to their parent corresponds to the loss of accessible
acid sites detected by the FT-IR measurements of adsorbed Py
Fig. 4) proved that mesopores were not formed even after acid
leaching.

8
2
R.M. Mihályi et al. / Applied Catalysis A: General 417–418 (2012) 76–86
A
B
3
2
2
1
1
0
5
0
5
0
5
0
30
25
20
15
10
5
Cont. time=0.04h
MCM-22
Cont. time=0.82h
D700
Cont. time=0.13h
MCM-22
Cont. time=0.82h
D500
D700/Ox
D700/HCl
D800/HCl
D500/Ox
0
0
40
80
120
160
0
40
80
120
160
TOS (min)
TOS (min)
Fig. 6. Total m-xylene conversion as a function of the TOS for the dealuminated catalysts steamed at 773 K (A), and at 973 K (B), and subsequently acid refluxed. Reaction
temperature 623 K.
(
Fig. 3B) and is in accordance with the reduction in their m-xylene
of both lattice and extra-lattice Al present in the steamed samples.
The proportion between the rest of them, left after the acid reflux,
do not change significantly and does not depend on the type of the
acid used. This result points to the contribution of each individual
dealumination step, the steaming one being more essential.
In order to assess the selectivity of the catalysts, it was indis-
pensable to apply quite differing reaction conditions for to attain
close degree of m-xylene conversion on the parent and the dealu-
minated materials. In Fig. 7 the conversion and the p- to o-xylene
ratio in dependence on the TOS are presented for diverse contact
times and reaction temperatures. As the data show, whatever the
reaction conditions and the dealumination procedure applied, at
the adjusted close or equal conversion, the selectivity towards the
p-xylene isomer is higher in case of the dealuminated catalysts
and increases in the course of the reaction compared to that of the
parent material. It is obvious that the observed effect is not a con-
sequence neither from the difference in the reaction temperatures
nor from the prolonged contact times, since longer contact times
(used for the dealuminated catalysts) must facilitate the equilibra-
tion between the three isomer products which was not the case in
our experiments.
Wu et al. [29] and Park et al. [27,28] also observed enhanced
p-isomer selectivity in the reactions of toluene and ethylbenzene
transformations respectively, for consecutively steam-acid dealu-
minated MCM-22 zeolites. According to these authors, the selective
replacement of framework Al from the sinusoidal channels and
the surface cups [29] or their preferential removal only from the
external surface [27,28] suppresses the secondary isomerization
of the primarily formed p-methyl/ethyl substituted benzenes and
this is the reason for their enhanced fraction in the final prod-
uct. The suggestion of Wu et al. [29] about the capability for a
specific and more intense replacement of aluminum atoms from
the sinusoidal channels of MCM-22 upon steam-acid dealumina-
tion appears somewhat doubtful, however, considering the last
results of Matias et al. [13]. Following the change in the activity
of the dealuminated by HNO3 zeolite MCM-22 in the test reac-
tion of methylcyclohexane transformation these authors found out
that the Al atoms are extracted first from the outer cups, and then
adsorption capacity (Fig. 5). The sequence in the diminish of the
activity follows the severity of steaming both for the steamed and
acid leached catalysts (Fig. 6) and corresponds to the one ascer-
tained by the Py and m-xylene adsorption measurements. These
observations indicate that part of the EFAl species generated upon
steaming and remained not extracted by the subsequent acid treat-
ment are most probably responsible for the substantial loss of acid
sites by both neutralizing part of them and giving rise to limitations
in the access of the reactant molecule to the rest. 29Si MAS NMR
studies of highly siliceous MCM-22 prepared by hydrothermal dea-
lumination confirm the presence of at least one buried T-site per
unit cell that is not accessible to a channel wall [37].
It turns out that the second stage of dealumination, the acid
reflux, instead of increasing the Broensted acidity as it happens
with the steam-acid dealuminated mordenites, provokes like in
beta zeolite (Ref. [12]), only partial removal of the EFAl species gen-
erated upon the previous steaming. The remaining, not effectively
dislodged and most probably polymerized EFAl species may block
the acid sites, whereas the cationic-type aluminum may endure re-
dispersion and may neutralize part of these sites in accordance to
Fernandes et al. [12]. If such process of aluminum redistribution
occurs in the large pore zeolites, it should be more reinforced by
the acid attack in the narrow 10 MR sinusoidal channels and the 10
MR openings of the supercages of zeolite MCM-22. Thus, it should
result in additional reduction of the concentration and accessibility
of the remaining Broensted acid sites.
It is worth mentioning here, that the type of the acid used for
reflux does not affect the catalytic action considering the behav-
ior of both materials steamed at 973 K and subsequently treated
with HCl or oxalic acid (Fig. 6B). Contrary to this, substantial dif-
ference is observed by Xia et al. [23] in the acidity as well as in the
selectivity to dimethyl ether produced from syngas when direct
treatment by individual acids (oxalic or citric) is applied on MCM-
2
2. Our characterization results above indicate however, that when
preliminary steaming step is applied, the type of the acid used upon
the next reflux procedure appears to be insignificant. The severity of
steaming determines preferentially the content and likely the type

R.M. Mihályi et al. / Applied Catalysis A: General 417–418 (2012) 76–86
83
T =523 K
react.
1
1
5
0
5
4
3
2
1
A
Cont. time =1.85 h
D500/Ox
Cont. time=0.1 h
MCM-22
Cont. time=1.85 h
D700/HCl
equilibrium
Cont. time=0.07 h
MCM-22
0
50
100
150 0
50
100
150
TOS (min)
TOS (min)
1
5
0
5
0
4
B
T
=523 K
react.
3
2
1
Cont. time=0.1h
1
MCM-22
T
=623 K
react.
Cont. time=0.82h
D700/HCl
D700/Ox
equilibrium
D700
0
50
100
150
0
50
100
150
TOS (min)
TOS (min)
Fig. 7. m-Xylene conversion and p- to o-xylene ratio in dependence on the TOS at different contact times and reaction temperature of 523 K (A), 523 K and 623 K (B) for the
parent and dealuminated catalysts.
preferentially from the larger supercages. The narrow sinusoidal
channels are hardly affected by the acid treatment because of dif-
fusion limitations for the transport of the HNO3 molecules alone.
Considering the succession of the steam-acid treatment in our case,
even more embarrassed transport of the acid molecules together
with the dissolved aluminum species, generated during the pre-
ceding hydrothermal treatment, might be expected, compared to
the solely acid leaching. The acid molecules will suffer in our case of
even more impeded diffusion through the micropores obstructed
by the EFAl debris, especially when the most confined sinusoidal
channels are considered. The lattice Al atoms located there should
be the most unlikely to be extracted upon the next acid reflux.
Considering the significant decrease in the catalytic activity of
our dealuminated catalysts presented above and their lower m-
xyleneadsorptioncapacity(Fig. 5), werathersuggestinaccessibility
but not a selective extraction of framework Al from the sinusoidal
channels. Matias et al. [13] also ascribe the diminish in the activ-
ity in methylcyclohexane transformation not only to the proton
concentration decrease by partial extraction of the framework alu-
minum (by acid only in their case) but also to the limitation in
the access of the reactant molecules to these acid sites caused
by the presence of the EFAl barriers. Thus the Wu’s suggestion
about the increased p-xylene selectivity of the steam-acid dealu-
minated MCM-22 should be a little bit modified in accordance to
our experimental results. The higher p-selectivity might be indeed
determined by the reduced number of acid sites in the sinusoidal
channels, which, however, are not extracted by the acid leaching
but are inaccessible because of the pore/site blockage by EFAl gen-
erated upon steaming. Dilution of the surface acid sites should also
have occurred since the most accessible centers must be the most
affected by the acid attack [13].
Since the acid treatment is not able to withdraw out of the nar-
row MCM-22 channel system all the EFAl entities generated by
steaming, then these entities would remain captured within the
zeolite micropores. This may cause, as well, diffusional limitations
for the bulkier o-isomer and will reinforce shape selectivity effect
in m-xylene transformation. This assumption is supported by the
reduced adsorption capacity towards m-xylene determined above
(Fig. 5) which should be due to blocking of the sinusoidal chan-
nels and/or narrowing of the 10 MR openings of the supercages

8
4
R.M. Mihályi et al. / Applied Catalysis A: General 417–418 (2012) 76–86
A 10
100
80
60
40
20
100
a
b
c
9
0
D500/Ox
8
6
4
2
80
Cont. time=1.85 h
7
0
MCM-22
60
Cont. time=0.1 h
5
0
40
3
0
20
1
0
0
40 80 120 160
0
40 80 120 160
TOS, min
0
40 80 120 160
B 10
100
80
60
40
20
100
80
a
b
c
8
6
4
2
D700/HCl
Cont. time=1.85 h
MCM-22
60
Cont. time=0.07 h
40
20
0
40 80 120 160
0
40 80 120 160
TOS (min)
0
40 80 120 160
Fig. 8. Total m-xylene conversion (a), selectivity to isomerization (b), and to side reactions of m-xylene transformation (c), in dependence on the time on stream on D500/Ox
A) and D700/HCl (B) (contact time = 1.85 h) compared with MCM-22 (contact time = 0.1 and 0.07 h) at Treact = 523 K.
(
2.0
1.5
1.0
0.5
0.0
1.0
20
a
c
b
18
D500/Ox
Cont. time=1.85 h
0.8
0.6
0.4
0.2
16
14
MCM-22
Cont. time=0.1 h
12
10
8
6
4
2
0
40
80
120
10 60
40
80
120
160
0
40
80
120
160
TOS (min)
Fig. 9. Distribution of toluene and TMB reaction products over D500/Ox and MCM-22 at close total m-xylene conversion (depicted in Fig. 8) in dependence on the TOS.
Treact = 523 K, cont. time = 1.85 h for D500/Ox and 0.1 h for MCM-22).
(

R.M. Mihályi et al. / Applied Catalysis A: General 417–418 (2012) 76–86
85
Table 2
where m-xylene isomerization preferentially occurs [48]. It is in
accordance, as well, to the selectivity effect revealed by hydrother-
mally dealuminated MCM-22 zeolites in the reaction of n-octane
hydroconversion [21].
Weight loss determined by TGA upon heating in Ar from room temperature to 873 K
and kept in air for 1 h at 873 K.
Samples
Heating in Ar
03–498 K
Heating in air
At 873 K
Total loss
It should be noticed here that the p- to o-isomer ratio of the
dealuminated catalysts increases with TOS (Fig. 7), an effect not
mentioned and discussed by others studied the p-selectivity in
the conversions of alkyl substituted aromatics [27–29]. This effect
appears to be related to the deactivation of the dealuminated cata-
lysts and may evolve from the gradual elimination of the remained
acid sites by carbonaceous compounds. Coke precursors may also
build up additional diffusion barriers, reduce the proton concen-
tration and suppress the secondary isomerization of the primarily
formed and most rapidly diffusing p-xylene isomer.
A different mode of deactivation of the parent material and its
dealuminated modifications can be suggested in accordance to the
observed reaction products distribution discussed below. In Fig. 8
the selectivity towards the isomerization (Siso.) and the side reac-
tions, i.e. dealkylation and disproportionation (Sdealk + dispropr.), at
almost equal degree of m-xylene conversion for two steam-acid
dealuminated catalysts is compared to their parent. Two main dis-
tinctions in the performance of both types of catalysts can be seen:
3
498–873 K
Over 498 K
(Ar + air), %
Tmax
% Loss
Tmax
% Loss
% Loss
MCM-22
D500/Ox
368 K
363 K
7.1
3.7
709 K
831 K
0.5
1.31
1.89
0.11
2.35
1.42
The disproportionation reaction of m-xylene should occur not
inside the narrowed pores but preferentially and more facilitated
on the surface hemycages. Then gradual elimination by progressive
coke formation on the remaining there acid surface sites starts,
most probably by the site poisoning mechanism of deactivation
proposed by Guisnet et al. [51]. In the same time, the deactivation
rate of both catalysts is analogous (Fig. 8a).
The results from the coke determinations for the couple of
spent catalysts MCM-22 and D500/Ox which reveal same degree
of m-xylene conversion and deactivation rate (Fig. 8a) are pre-
sented in Table 2. In accordance with the interpretations of Bibby
et al. [53] and Pradhan et al. [54], three well established peaks of
water elimination (303 K-498 K, not shown), and replacement of
the more mobile carbonaceous deposits-“soft” coke (498–873 K)
are detected in argon flow. The strongly retained, more condensed
compounds considered as “hard” coke are liberated upon air com-
bustion at 873 K. According to the data, the dealuminated catalyst
(
i) the selectivity towards the side reactions of m-xylene conver-
sion of the parent material decreases with the TOS on the expense
of its isomerization and that corresponds to the deactivation rate.
In case of the dealuminated catalysts the values for Siso. and Sdispr.
remain almost constant, and (ii) the dealuminated samples reveal
lower selectivity towards isomerization and higher for dispropor-
tionation compared to their parent.
is able to adsorb twice lower amount of H O, and its total weight
2
loss is also about half of its parent, most probably because of
the restricted void pore volume. The highly volatile coke precur-
sors released from it in argon, at the high temperature interval
According to Laforge et al. [49], the deactivation of the cata-
lysts of the MCM-22 type upon m-xylene conversion reflects the
decrease in the intensity of disproportionation which occurs on the
acid sites located in the supercages where coke preferentially accu-
mulates. This process is associated with the pore blocking effect
strongly pronounced for these so called “trap cages” (large cages
with small apertures) of MCM-22 [50], according to the classifica-
tion of Guisnet et al. [51]. Our parent material displays analogous
performance (Fig. 8A and B). If one considers the confined space
in the large supercages of our steam-acid dealuminated catalysts,
however, then m-xylene disproportionation (which involves bulky
bimolecular intermediates) should occur exceptionally at the acid
sites remaining on the external hemicages. The deactivation of
these surface sites should be due to “site poisoning” rather than
to “pore blockage” (Ref. [51]) and should lead to a proportional
and constant change in the selectivity of both isomerization and
disproportionation as our results actually show. Desorption of the
bulk disproportionation products (TMB isomers) from these large
external cups with small depth would be easier and faster, and their
retention time should be shorter enough to prevent substantial
secondary transformations [52] as it is shown below.
(498–873 K), are about three times higher, than that of the parent.
On the contrary, the parent material reveals preferential formation
of much more condensed coke deposits eliminated in air. Obviously,
the composition of the coke over the parent and the dealuminated
catalysts differ in a great extent. The more condensed deposits on
the former are probably accumulated into the supercages as was
suggested above. These compounds are strongly embarrassed to
leave the “trap cages” and endure intense condensation. Higher
temperature and air flow are needed for their removal from the
aged catalyst. In contrast, coke precursors on the dealuminated
materials comprises of more volatile compounds, since they grow
on the surface cups and can readily desorb from there, at lower
temperature and still in Ar. The process of formation of these coke
precursors is determined by the larger space available on the exter-
nal surface of the dealuminated catalysts, independently from the
much longer contact time applied for the catalytic experiments
with them.
Thus, it can be suggested that the disposition of the residual EFAl
in the three different pore systems of MCM-22 affects the reaction
products distribution in both reactions of isomerization and dispro-
portionation in m-xylene transformation, as well as it determines
the mechanism of carbonaceous compounds formation. This obser-
vation is in accord to the conclusion of Guisnet et al. [51] that both
coke formation and deactivation are shape-selective processes.
Comparison of the side products distribution on both types of
catalysts, parent and dealuminated, supports the suggestion made
above. Larger yields of toluene and TMB are detected on D500/Ox
(
Fig. 9a and b) which indicates facilitated reaction of m-xylene dis-
proportionation over the dealuminated catalyst. The Tol:TMB mole
ratio is constant (Fig. 9c) however, and amounted to 3 indicating
some permanent loss of TMB isomers, but much lower than that
of the parent material (starting from 11 and decreasing along with
the catalyst deactivation). This much higher selectivity in favor of
toluene (Fig. 9c) is accompanied by formation of ethyltoluene and
C2–C4 hydrocarbons (not shown). This is an indication for more
intense occurrence of secondary reactions of transformation of the
TMB molecules remained trapped in the “empty” micropores of
this catalyst. Such a process is strongly restricted by the pres-
ence of EFAl entities in the channel system of the dealuminated
materials.
4. Conclusion
The effect of the residual EFAl species can be generalized as
follows:
1
. The presence of EFAl entities causes blocking of the zeolite
micropores and leads to significant reduction in their adsorp-
tion capacity and catalytic activity. This effect is assumed to be
a particular feature of the steam-acid dealuminated MCM-22.

8
6
R.M. Mihályi et al. / Applied Catalysis A: General 417–418 (2012) 76–86
2
. The proposed by other authors selective replacement of pro-
[19] P. Canizares, A. Carrero, Appl. Catal. A: Gen. 248 (2003) 227–237.
[
20] L. Teyssier, M. Thomas, C. Bouchy, J. Martens, E. Guillon, in: A. Gedeon, P. Mas-
siani, F. Baboneau (Eds.), Zeolites and related materials: Trends, Targets and
Challenges, Stud. Surf. Sci. Catal., vol. 174, Elsevier, Amsterdam, 2008, p. 969.
ton sites by such steam-acid dealumination of MCM-22 zeolite
should be implied not only as definite extraction and elimina-
tion of the framework Al atoms connected to bridging protons
but also as a process of “burying” of these acid sites under the
EFAl species. This makes most of the proton sites inaccessible and
inefficient and is the reason for the observed enhanced p-xylene
selectivity.
. The presence of residual EFAl species limits the access of the reac-
tant molecules to the catalytically active sites, leads to reduced
activity in m-xylene conversion, affects the products distribution
in both reactions of m-xylene isomerization and disproportion-
ation and determines the mode of coke formation.
[21] P. Meriaudeau, Vu A. Tuan, Vu T. Nghiem, F. Lefevre, Vu T. Ha, J. Catal. 185 (1999)
78–385.
3
[
[
22] D. Ma, F. Deng, R. Fu, X. Han, X. Bao, J. Phys. Chem. B 105 (9) (2001) 1770–1779.
23] J. Xia, D. Mao, W. Tao, Q. Chen, Y. Zhang, Y. Tang, Microporous Mesoporous
Mater. 91 (2006) 33–39.
[24] S. Unverricht, M. Hunger, S. Ernst, H.G. Karge, J. Weitkamp, in: J. Weitkamp, H.G.
Karge, H. Pfeifer, W. Holderich (Eds.), Zeolites and Related Microporous Mate-
rials, State of the Art 1994, Stud. Surf. Sci. Catal., vol. 84, Elsevier, Amsterdam,
1994, p. 37.
3
[25] X. Ren, J. Liang, J. Wang, J. Porous Mater. 13 (2006) 353–357.
[
[
[
[
26] J.-N. Park, J. Wang, S.-I. Hong, C.W. Lee, Appl. Catal. A: Gen. 292 (2005) 68–75.
27] S.-H. Park, H.-K. Rhee, React. Kinet. Catal. Lett. 78 (2003) 73–80.
28] S.-H. Park, H.-K. Rhee, Appl. Catal. A: Gen. 219 (2001) 99–105.
29] P. Wu, T. Komatsu, T. Yashima, Microporous Mesoporous Mater. 22 (1998)
3
43–356.
Acknowledgments
[
30] R. Ravishankar, D. Bhattacharya, N. Jacob, S. Sivasankar, Microporous Meso-
porous Mater. 4 (1995) 83–93.
Support of this work in the framework of the Hungarian-
Bulgarian Inter-Academic Exchange Agreement (project no. 17)
is gratefully acknowledged. The support of the Hungarian GVOP-
[31] R.M. Mihályi, I. Kolev, V. Mavrodinova, M. Kollár, T. Korányi, Ch. Minchev, in: K.
Hadjiivanov, V. Valtchev, S. Mintova, G. Vayssilov (Eds.), Advanced Micro- and
Mesoporous Materials, Heron Press, 2008, pp. 363–371.
[
32] D.L. Bryce, G.M. Bernard, M. Gee, M.D. Lumsden, K. Eichele, R.E. Wasylishen,
Can. J. Anal. Sci. Spectrosc. 46 (2001) 46–82.
3
.2.1.-2004-04-0210/3.0 grant is gratefully acknowledged.
[
[
[
33] P. Koehl, Progress NMR Spectrosc. 34 (1999) 257.
34] T.H. Chen, B.H. Wouters, P.J. Grobet, Eur. J. Inorg. Chem. (2000) 281–285.
35] V. Mavrodinova, M. Popova, Y. Neinska, Ch. Minchev, Appl. Catal. A: Gen. 210
References
(
2001) 397–408.
[
1] J. Cejka, A. Corma, S. Zones (Eds.), Zeolites and Catalysis. Synthesis, Reactions
and Applications, Wiley-VCH Verlag GmbH&Co. KGaA, Weinheim, 2010.
2] (a) C. Choifeng, J.B. Hall, B.J. Huggins, R.A. Beyerlein, J. Catal. 140 (1993)
[36] M.A. Camblor, A. Corma, M.J. Dias-Cabanas, J. Phys. Chem. B 102 (1998) 44–51.
[37] G.J. Kennedy, S.L. Lawton, A.S. Fung, M.K. Rubin, S. Steuernagel, Catal. Today 49
(1999) 385–399.
[
3
95–405;
[38] A. Corma, V. Fornes, A. Martinez, J. Sanz, ACS Symp. Ser. 375 (1988) 17–33.
[39] J.P. Gilson, G.C. Edwards, A. Petes, K. Rajagopalan, R.F. Wormsbecher, T.G.
Roberie, M.P. Shatlock, J. Chem. Soc. Chem. Commun. (1987) 91–92.
[40] N. Malicki, G. Mali, A-A. Quoineaud, P. Bourges, L. Simon, F. Thibault-Starzyk,
C. Fernandez, Microporous Mesoporous Mater. 129 (2010) 100–105.
[41] A. Corma, A. Martinez, C. Martinez, Appl. Catal. A: Gen. 134 (1996) 169–182.
[42] J. Miller, P. Hopkins, B. Meyers, G. Ray, R. Roginski, G. Zajac, N. Rosenboum, J.
Catal. 138 (1992) 115–128.
(
(
b) R.A. Beyerlein, C. Choi Feng, J.B. Hall, B.J. Huggins, G.J. Ray, Top. Catal. 4
1997) 27–42.
[
[
[
[
3] A. Corma, V. Fornes, F. Rey, Appl. Catal. 59 (1990) 267–274.
4] J. Scherzer, ACS Symp. Ser. 248 (1984) 157–200.
5] J. Linch, F. Raatz, P. Dufresne, Zeolites 7 (1987) 333–340.
6] J.L. Molz, H. Heinichen, W.F. Holderich, J. Mol. Catal. A: Chem. 136 (1998)
1
75–184.
[
[
[
7] C.S. Triantafillidis, A.G. Vlessidis, L. Nalbandian, N.P. Evmiridis, Microporous
Mesoporous Mater. 47 (2001) 369–388.
8] G.C. Groen, J.A. Moulijn, J. Perez-Ramirez, Microporous Mesoporous Mater. 87
[43] D. Meloni, S. Laforge, D. Martin, M. Guisnet, E. Rombi, V. Solinas, Appl. Catal. A:
Gen. 215 (2001) 55–66.
[44] M.S. Holm, S. Svelle, F. Joensen, P. Beato, C.H. Cristensen, S. Bordiga, M. Bjorgen,
Appl. Catal. A: Gen. 356 (2009) 23–30.
[45] C. Fernandez, I. Stan, J-P. Gilson, K. Thomas, A. Vicente, A. Bonilla, J. Perez-
Ramirez, Chem. Eur. J. 16 (2010) 6224–6233.
[46] S. Altwasser, J. Jiao, S. Steuernagel, J. Weitkamp, M. Hunger, in: E. van Santen,
L. Callanan, M. Claeys (Eds.), Proc. 14th Int. Zeol. Conf., Cape Town, 2004, pp.
1212–1213.
[47] C. Ding, X. Wang, X. Guo, S. Zhang, Catal. Commun. 9 (2007) 487–493.
[48] S. Laforge, D. Martin, M. Guisnet, Microporous Mesoporous Mater. 67 (2004)
235–244.
[49] S. Laforge, D. Martin, J.L. Paillaud, M. Guisnet, J. Catal. 220 (2003) 92–103.
[50] D. Meloni, D. Martin, M. Guisnet, Appl. Catal. A: Gen. 215 (2001) 67–79.
[51] M. Guisnet, L. Coste, F. Ramoa Ribeiro, J. Mol. Catal. A: Chem. 305 (2009)
69–83.
(
2005) 153–161.
9] M.R. Minas, I. Rahmim, A.S. Fung, J.R.A. Huss, US Patent 5321194 (1994).
[
[
[
[
[
10] M.R. Apelian, A.S Fung, J.K. Gordon, F. Thomas, Prepr. 14th North American
Meeting of the Catalysis Society, Snowbird, Utah, 1995.
11] M. Muller, G. Harvey, R. Prince, Microporous Mesoporous Mater. 34 (2000)
35–147.
12] L.D. Fernandes, J.L.F. Monteiro, E.F. Sousa-Aguilar, A. Martines, A. Corma, J. Catal.
77 (1998) 363–377.
13] P. Matias, J.M. Lopes, P. Ayrault, S. Laforge, P. Magnus, M. Guisnet, F. Ramoa
Ribeiro, Appl. Catal. A: Gen. 365 (2009) 207–213.
1
1
14] M.E. Leonowicz, J.A. Lawton, S.L. Lawton, M.K. Rubin, Science 264 (1994)
1
910–1913.
[
[
15] A. Corma, C. Corell, J. Perez-Pariente, Zeolites 15 (1995) 2–8.
16] A.Corma, C. Corell, V. Fornes, W. Kolodziejski, J. Perez-Pariente, Zeolites 15
[52] P. Matias, J.M. Lopes, S. Laforge, P. Magnoux, P.A. Russo, M.M.L. Ribeiro Carrot,
M. Guisnet, F. Ramoa Ribeiro, J. Catal. 259 (2008) 190–202.
[53] D.M. Bibby, N.B. Milestone, J.E. Petterson, L.A. Aldridge, J. Catal. 97 (1986)
493–502.
[54] A.R. Pradhan, J.F. Wu, S.J. Jong, T.C. Tsai, S.B. Liu, Appl. Catal. A: Gen. 165 (1997)
489–497.
(
1995) 576–582.
17] S.M. Campbel, D.M. Bibby, J.M. Coddington, R.F. Howe, R.H. Meinhold, J. Catal.
61 (1996) 338–349.
18] T. Masuda, Y. Fujikata, S.R. Mukai, K. Hashimoto, Appl. Catal. A: Gen. 172 (1998)
3–83.
[
[
1
7