Influence of layer structure preservation on the catalytic properties of the pillared zeolite MCM-36

Article abstract of DOI:10.1016/j.jcat.2010.04.011

MCM-36 is a catalytically active material made by swelling and subsequent pillaring of a layered zeolite precursor, MCM-22(P). However, the swelling procedures lead to significant destruction of layer structure and alteration of framework silicon/aluminum (Si/Al) ratio, which are likely to influence the acidity and catalytic activity of the final materials. We report a milder swelling and pillaring process to prevent such structural destruction. The resulting pillared materials show higher crystallinity and improved retention of the zeolitic layer structure when compared to the ones prepared by the previously reported methods involving more aggressive treatments. Implications of such structural preservation on the physicochemical, acidic and catalytic properties of the materials have been investigated using a variety of characterization methods. The milder swelling results in catalytic materials with stronger acid sites and, at similar Si/Al ratios, higher acid site density and higher catalytic activity when compared to the MCM-36 materials produced using the previously reported more aggressive swelling process.

Full text of DOI:10.1016/j.jcat.2010.04.011

Journal of Catalysis 272 (2010) 298–308
Contents lists available at ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
Influence of layer structure preservation on the catalytic properties
of the pillared zeolite MCM-36
Sudeep Maheshwari a,1, Cristina Martínez b,1, M. Teresa Portilla , Francisco J. Llopis , Avelino Corma b,*,
Michael Tsapatsis
a
b
c
a,
**
Department of Chemical Engineering and Materials Science, University of Minnesota, Minneapolis, Minnesota, USA
Instituto de Tecnología Química (UPV-CSIC), Universidad Politécnica de Valencia, Consejo Superior de Investigaciones Científicas, Valencia, Spain
Departamento Ingeniería Química, U.V.E.G. Universidad de Valencia, Burjassot, Valencia, Spain
b
c
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 3 December 2009
Revised 4 April 2010
Accepted 14 April 2010
Available online 14 May 2010
MCM-36 is a catalytically active material made by swelling and subsequent pillaring of a layered zeolite
precursor, MCM-22(P). However, the swelling procedures lead to significant destruction of layer structure
and alteration of framework silicon/aluminum (Si/Al) ratio, which are likely to influence the acidity and
catalytic activity of the final materials. We report a milder swelling and pillaring process to prevent such
structural destruction. The resulting pillared materials show higher crystallinity and improved retention
of the zeolitic layer structure when compared to the ones prepared by the previously reported methods
involving more aggressive treatments. Implications of such structural preservation on the physicochem-
ical, acidic and catalytic properties of the materials have been investigated using a variety of character-
ization methods. The milder swelling results in catalytic materials with stronger acid sites and, at similar
Si/Al ratios, higher acid site density and higher catalytic activity when compared to the MCM-36 mate-
rials produced using the previously reported more aggressive swelling process.
Keywords:
Swelling
Pillaring
MWW zeolite
MCM-36
Layered zeolites
Xylene isomerization
Alkylation
Ó 2010 Elsevier Inc. All rights reserved.
Cracking
1
. Introduction
Earlier [8–10] and more recent [11] reports have shown that
the swelling process results in significant destruction of zeolite
MCM-36 [1] is the first pillared material with zeolitic layers
2,3]. It is derived from a layered precursor, MCM-22(P). The layers
structure with dissolution of framework silica, dealumination of
the material and reduction in microporosity within the zeolite
layers. This is likely to affect the density and nature (both
strength and accessibility) of protonic sites, which in turn, will
influence the catalytic activity of the MCM-36 materials. The
change in pore structure can also influence the selectivity [5,
12–14].
Recently, we reported an alternative procedure to swell MCM-
22(P), which results in preservation of the layer and pore structure
and of the zeolite composition [11]. Use of lower temperature for
swelling was the key to prevent the structural destruction and dis-
solution of the precursor. The resultant swollen material could be
pillared to make an analog of MCM-36, albeit with much improved
crystallinity.
[
of MCM-22(P) are intercalated by silica species, which convert into
inorganic pillars upon polymerization and hold the layers apart
creating interlayer mesoporous spaces [4]. These mesopores allow
increased access to protonic sites present on the zeolite, making
this material attractive for catalytic and sorption applications [2].
The beneficial effect of pillaring can be seen from the high catalytic
activity of MCM-36 when compared to its non-pillared counterpart
MCM-22 for a variety of reactions such as vacuum gasoil cracking
5], alkylation of isoparaffins or aromatics [6], etc.
Synthesis of MCM-36 from MCM-22(P) is a two-step process
1,4,6,7]. In the first step, MCM-22(P) is swollen by cetyltrimethyl-
ammonium cations in the presence of a base (pH > 13.5) at ele-
vated temperature (353–373 K). The swollen MCM-22(P) is then
pillared with tetraethoxysilane (TEOS) derived silica.
[
[
In this work, a detailed study describing the effect of swelling
conditions on the physiochemical properties and catalytic behavior
of MCM-36 materials is presented. X-ray diffraction (XRD), trans-
mission electron microscopy (TEM), nuclear magnetic resonance
*
Corresponding author. Fax: +34 96 3879444.
(
NMR), nitrogen adsorption and IR spectroscopy of adsorbed pyri-
*
* Corresponding author. Fax: +1 (612) 626 7246.
dine were used to characterize the structural and acid site differ-
ences in pillared materials prepared using two different swelling
conditions. The catalytic activity of these materials for a particular
E-mail addresses: acorma@itq.upv.es (A. Corma), tsapatsi@cems.umn.edu (M.
Tsapatsis).
1
Equally contributed authors.
0
021-9517/$ - see front matter Ó 2010 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2010.04.011

S. Maheshwari et al. / Journal of Catalysis 272 (2010) 298–308
299
set of reactions was tested. It is shown that for given silicon
2.3. Characterization methods
to aluminum ratio (Si/Al), MCM-36 materials prepared by
milder swelling exhibit stronger acidity and higher catalytic
activity.
The silicon and aluminum contents of the pillared materials
were determined using a 715-ES inductive coupled plasma (ICP)
optical emission spectrometer.
Nitrogen adsorption–desorption measurements were carried
out at 77 K with an Autosorb-1 analyzer (Quantachrome Instru-
ments). Prior to the measurement, samples were evacuated over-
night at 623 K and 1 mm Hg. The specific surface area and pore
size distribution were calculated by applying the BET and BJH
models, respectively, to the nitrogen adsorption data.
Powder X-ray diffraction (XRD) patterns were collected on a
Bruker AXS D5005 diffractometer using Cu Ka radiation. Data were
collected in the 2h range from 1° to 30° with a step size of 0.04° and
a step time of 3 s.
2
. Experimental
Three different MCM-36 samples were prepared in this study.
Two samples were prepared starting from the same MCM-22(P)
precursor but with different swelling conditions to assess the effect
of swelling on the Si/Al ratio, morphology and structure of the pil-
lared materials. Different swelling conditions were found to result
in significant difference of Si/Al ratio in the final materials. There-
fore, a third sample was made starting with a different precursor to
match the final Si/Al ratio of the pillared material exhibiting the
lower ratio. The details of precursor synthesis, swelling and pillar-
ing procedures are described below.
All NMR spectra were recorded at a field of 11.7 T (BRUKER
29
Avance 500). Si MAS NMR spectra were recorded at 99.37 MHz
using 4 mm rotors at a spinning speed of 8 kHz, a dwell time of
1
6.65 ls, a p/2 pulse of 4.0 ls and a recycle delay of 60 s. All spec-
27
tra were referenced with respect to tetramethylsilane (0 ppm). Al
2.1. Synthesis of MCM-22(P)
MAS NMR spectra were recorded at 130.34 MHz using 4 mm rotors
at 14 kHz spinning speed, a dwell time of 0.5
ls, a selective p/18
Two MCM-22(P) batches with different Si/Al ratios were syn-
pulse of 0.3 s and a recycle delay of 0.1 s. An aqueous solution
l
thesized using the method of Corma et al. [5,15]. For the MCM-
2 precursor with Si/Al ꢀ 45 (gel composition), 0.54 g of sodium
aluminate (MP biomedicals, USA) and 1.86 g of sodium hydroxide
97+%, Fisher) were dissolved in 233.06 g of distilled water. Subse-
quently, 14.34 g of hexamethyleneimine (HMI) (Aldrich) and
7.68 g of fumed silica (Cab-o-sil M5) were added to the mixture.
The mixture was allowed to stir for 5 h at room temperature, fol-
lowed by 11 days in rotating Teflon-lined steel autoclaves at
08 K. The crystalline product obtained after 11 days was collected
by centrifugation and repeatedly washed with distilled water to re-
duce the pH to 9.
For the precursor with Si/Al ꢀ 20, the amounts of sodium alumi-
nate and sodium hydroxide were changed to 1.37 g and 1.32 g,
respectively, and the mixture was stirred for 20 h at room temper-
ature before heating it to 408 K.
of aluminum sulfate (0.1 M) was used as the reference (0 ppm).
A FEI Tecnai G2 F30 transmission electron microscope (TEM)
equipped with a charge couple device (CCD) and operated at
2
(
3
00 kV was used for direct imaging of pillared materials. Samples
were prepared by dispersing powders in water and placing a few
drops of the dispersion on a carbon-coated copper grid followed
by air drying to evaporate off the water.
1
The concentration of acidic sites in various samples was ob-
tained by FT–IR spectroscopy using pyridine as the probe molecule.
Pyridine adsorption–desorption experiments were carried out on
4
ꢁ
1
self-supported wafers (10 mg cm ) of original samples previously
ꢁ2
activated at 673 K and 10 Pa for 2 h. After wafer activation, the
2
base spectrum was recorded, and pyridine vapors (6.5 ꢂ 10 Pa)
were admitted into the vacuum IR cell and adsorbed onto the zeo-
lite at 423 K. Desorption of pyridine was performed in vacuum over
three consecutive 1h periods of heating at 423, 523 and 623 K, each
followed by an IR measurement at room temperature. All the spec-
tra were scaled according to the sample weight. The number of
acid sites was calculated from the intensity of the band at
2
.2. Swelling and pillaring of MCM-22(P)
Both precursors were swollen with cetyltrimethylammonium
ꢁ1
ꢁ1
bromide (CTAB) at room temperature under high pH conditions.
The composition of the swelling mixture was the same as the
one reported by Kresge [1] and Corma [5]. Typically, 9 g of aqueous
slurry of MCM-22(P) (20 wt.% solids) was mixed with 35 g of an
aqueous solution of 29 wt.% CTAB (Aldrich) and 11 g of an aqueous
solution of 40 wt.% TPAOH (Alfa Aesar). The pH of the resulting
mixture was ꢀ13.8. The mixture was allowed to stir for 16 h at
room temperature, after which, the particles were recovered by
repeated cycles of centrifugation and washing with DI water (10-
min centrifugation at 10,000 rpm and redispersion in fresh water).
The process of centrifugation and washing was repeated until the
02 and 003 reflections were detected in the XRD patterns [11].
A portion of MCM-22(P) with Si/Al ꢀ 45 was also swollen at
53 K following the procedure reported by Corma et al. [5,15] for
comparison.
Pillaring of the swollen materials was performed according to
1450 cm (Lewis acid sites) and 1550 cm (Brønsted acid sites)
by applying the extinction coefficients from Emeis [16]:
2
C
B
¼ 1:88 ꢃ IA
B
ꢃ R =W
and
2
C ¼ 1:42 ꢃ IA ꢃ R =W
L
L
where B represents Brønsted acid center, L represents Lewis acid
center, C is the concentration (mmol pyridine/g catalyst), IA is the
integrated absorbance, R is the radius of the catalyst disk (cm),
and W is the weight of the disk (mg).
0
3
2.4. Catalytic testing
Catalytic cracking of n-decane and vacuum gasoil (VGO, Table 1)
was performed in an automated microactivity test (MAT) unit,
which can be operated in a continuous way up to eight cycles,
i.e., stripping–reaction–regeneration, without operator interven-
tion. The reaction zone and product recovery system have been de-
signed in accordance with ASTM D-3907. In order to vary the
the procedure reported by Barth et al. [7]. Typically, 1 g of swollen
MCM-22(P) powder was mixed with 5 g of TEOS (tetraethoxysi-
lane, Fluka), stirred for 25 h at 351 K under argon atmosphere, then
filtered and dried at room temperature. Of the resulting solid, 0.5 g
was hydrolyzed in 5 g of water (pH ꢀ 8, controlled with NaOH) for
6
h at 313 K, then filtered, dried at 300 K and calcined at 723 K un-
conversion, the catalyst to oil ratio was varied in the range of
ꢁ1
ꢁ1
der N
2
flow (140 ml min ) for 6 h and finally at 823 K in air for
0.25–0.83 g g
by keeping the amount of catalyst constant
1
2 h (temperature ramp rate of 2 K/min).
(0.5 g) and changing the amount of feed. Before each experiment,

3
00
S. Maheshwari et al. / Journal of Catalysis 272 (2010) 298–308
Table 1
Vacuum gasoil properties.
(HP5890II) equipped with a Supelco-WAX10 capillary column
60 m length, 0.2 mm inner diameter) and a flame ionization
detector. Initial rates and selectivities were calculated as detailed
elsewhere [18].
(
ꢁ
1
Density (15 °C) g cc
Aniline point (°C)
K (UOP)
Sulfur (wt.%)
Nitrogen (ppm)
Na (ppm)
0.916
79.2
11.82
1.65
1261
0.18
<0.1
0.3
3
. Results and discussion
Cu (ppm)
Fe (ppm)
3.1. Si/Al ratio and surface area
Ni (ppm)
0.2
V (ppm)
Average molecular weight
0.4
407
The physicochemical properties of the three samples are listed
in Table 2. The sample name indicates the preparation condition
and final Si/Al ratio of the pillared material. For example,
MCM36-RT-23 indicates that the sample was prepared by swelling
at room temperature (RT) and the final Si/Al ratio in the pillared
material is 23. HT stands for high temperature (353 K).
Aromatic carbon (ndM) (%)
Naphthenic carbon (ndM) (%)
Paraffinic carbon (ndM) (%)
Aromatic rings/molecule (ndM)
Naphthenic rings/molecule (ndM)
22.96
15.16
61.88
1.17
1.01
Distillation curve ASTM D-1160 (°C)
Samples MCM36-RT-54 and MCM36-HT-24 were prepared
starting from the same MCM-22(P), but swelling at room temper-
ature and 353 K, respectively. It can be seen that swelling at high
temperature (which is the usual procedure reported in the litera-
ture) results in considerable dissolution of framework silica due
to the reactive basic conditions. The resulting material (MCM36-
HT-24) has a lower Si/Al ratio when compared to the precursor de-
spite the silica incorporation during the pillaring process. Swelling
at room temperature prevents framework silica dissolution to a
great extent. The final pillared materials (MCM36-RT) have higher
Si/Al ratio than the corresponding precursor materials, as expected,
due to incorporation of silica during the pillaring process.
Nitrogen adsorption/desorption isotherms of the three pillared
samples are shown in Fig. 1. An increase in the adsorption volume
up to relative pressures of ꢀ0.3 indicates the presence of mesop-
ores created by pillaring in all three samples. However, the in-
crease in adsorption volume continues to higher pressure values
in sample MCM36-HT-24, indicating the presence of larger mesop-
ores as well as a wider pore size distribution. As already observed
from decrease in the Si/Al ratio, MCM36-HT-24 is produced follow-
ing significant dissolution of framework silica. Such dissolution is
likely to create secondary pore structure with larger mesopores
in addition to the ones produced by pillaring. The BJH pore size dis-
5
%
10%
30%
50%
436
70%
459
90%
512
3
19
352
414
the system was purged with a 30 cc/min N
reaction temperature. After reaction, stripping of the catalyst was
carried out for 15 min using a N flow of 30 cc/min. During the
2
flow for 30 min at the
2
reaction and stripping steps, the liquid products were collected
in the corresponding glass receivers located at the exit of the reac-
tor, kept at 280 K by means of a computer-controlled bath. Mean-
while, the gaseous products were collected in a gas burette by
water displacement. After stripping, the catalyst was regenerated
at 813 K for 3 h, in a 100 cc/min stream of air. Catalytic cracking
reaction was performed at 773 K and 60 s time on stream (TOS)
for both feedstocks. The gases were analyzed using a Varian 3600
GC equipped with two detectors, a thermal conductivity detector
(
TCD) for analysis of H
molecular sieve column, and a flame ionization detector (FID) for
C1–C6 hydrocarbons separated in a 30 m Plot/Al column. In
the case of n-decane cracking, the liquids were analyzed in a Varian
900 GC with a 100 m, 0.25 mm ID Petrocol DH column, and in the
2 2
and N , which were separated in a 15 m
2 3
O
3
case of gasoil by simulated distillation (SIMDIS) in a Varian 3800
GC equipped with a 10 m, 0.53 mm ID MXT-2887 metal column
supplied by Restek Corp. and using the STARD SD software from
Table 2
Varian following the ASTM D-2887 standard. The CO
2
formed dur-
Physicochemical properties of MCM36 samples.
ing the regeneration step is monitored and quantified by means of
an IR cell (Rosemund Binos-100). More details can be found in [17].
The meta-xylene isomerization and benzene and toluene alkyl-
ation reactions were conducted in a fixed-bed continuous glass
down-flow reactor (11 mm internal diameter). The catalyst was di-
luted with glass and placed in the reactor bed.
2
Sample
Si/Al in precursor gel
Si/Al (ICP)
S
BET (m /g)
MCM36-RT-54
MCM36-HT-24
MCM36-RT-23
45
45
20
54
24
23
789
862
742
The meta-xylene isomerization was carried out with a meta-xy-
lene/N
toluene was fed in excess to alcohol in a molar ratio of 4:1, while
the N flow was fixed to achieve a 10:1 M ratio of N to alcohol.
Prior to addition of the reactants, the catalyst was heated to
2
molar ratio of 1:4. In the alkylation reaction, benzene or
2
2
6
3
23 K at a heating rate of 5 K/min, under a flow of nitrogen. After
0 min, the temperature was raised to 723 K and kept at that tem-
perature for 1 h. The reactor was then cooled to reaction tempera-
ture: 623 K for meta-xylene isomerization and 553 K for aromatics
alkylation. When the temperature was stabilized, the reactants in
liquid-phase were fed at the top of the reactor. The reactor exit
was connected to
a multisampling controlled heated valve
equipped with a hydrocarbon detector at the outlet, in order to de-
tect the moment when the products fill the first loop. This moment
is considered the zero reaction time, and the gas sample is kept in
the first loop. The rest of the loops were automatically filled at pre-
programmed times on stream (10, 30, 60, 180 and 360 s). The prod-
ucts of the reactions were analyzed in a gas chromatograph
Fig. 1. Nitrogen adsorption isotherms for (a) MCM36-RT-23, (b) MCM36-RT-54 and
(c) MCM36-HT-24.

S. Maheshwari et al. / Journal of Catalysis 272 (2010) 298–308
301
and long-range ordering of layers in the two MCM36-RT samples
and are in agreement with the physicochemical properties dis-
cussed above. Again, no major qualitative differences are observed
for the materials prepared by RT swelling starting from the two dif-
ferent precursors.
3.3. Transmission electron microscopy
Transmission electron microscopy was employed for the direct
visualization of crystal morphology and layer structure. The start-
ing MCM-22(P) material consists of thin circular disks that are
about 500–1000 nm in diameter and 50–100 nm in thickness
(
images not shown) [11]. TEM images of MCM36-RT-54 and
MCM36-HT-24 are shown in Fig. 4 for comparison. MCM36-RT-
4 shows a remarkable preservation of crystal morphology during
5
the swelling and pillaring process. The crystals retain their circular
disk morphology and sharp facets as seen in Fig. 4a (the crystals are
oriented along their thinner edge in this image). Fig. 4b shows a
higher resolution image of the same sample. Stacks of MWW layers
can be seen in each crystal (appearing as dark bands separated by
white bands corresponding to gallery space). The long-range order-
ing of layers, suggested by the presence of higher order reflections
in the XRD patterns, can be directly visualized in this micrograph.
The remarkable ordering of the layers suggests a nearly perfect pil-
laring of material, without major changes to the crystal morphol-
ogy. The interlayer spacing calculated from these TEM images is
in agreement with those obtained from XRD. Images of MCM36-
RT-23 sample show similar structural properties (not shown).
MCM36-HT-24, on the other hand, shows a highly disordered
structure (Fig. 4c). The disk-like morphology and sharp crystal fac-
ets are no longer visible. Unlike MCM36-RT-54 that showed or-
dered stacks of layers throughout the sample, this material
shows small domains of stacked layers surrounded by disordered
amorphous domains (Fig. 4d). It appears that swelling and pillaring
in this sample proceeds with considerable fragmentation and par-
tial dissolution of crystals, destroying the overall morphology and
creating disordered domains. Note that the only difference in the
preparation of this sample when compared to MCM36-RT-54 is
the temperature of swelling and subsequent water washings. Thus,
room temperature swelling results in a well-ordered pillared mate-
rial with remarkable structural preservation.
Fig. 2. BJH pore size distribution of MCM36-RT-23 (d), MCM36-RT-54 (j) and
MCM36-HT-24 ( ).
D
tribution shown in Fig. 2 clearly shows that MCM36-HT-24 has
bigger mesopores (ꢀ2.7 nm) when compared to the two MCM36-
RT samples (ꢀ1.8 nm). It is important to remark that the initial
Si/Al ratio of the precursor is not affecting the room temperature
swelling procedure and that both samples, MCM36-RT-54 and
MCM36-RT-23, have similar textural properties and a narrow mes-
opore size distribution.
Surface areas of all three samples were estimated by applying
the BET method to the nitrogen adsorption isotherms shown
above, and values are reported in Table 2. The two MCM-36 sam-
ples prepared via room temperature swelling (MCM36-RT-54 and
MCM36-RT-23) show very similar BET surface areas. The values re-
ported are typical of pillared MCM-36 materials [6]. MCM36-HT-
2
4 has the highest surface area among the three samples, presum-
ably due to creation of secondary pore structure by dissolution of
framework silica as well as severe fragmentation during the
high-temperature swelling process [11].
3.2. X-ray diffraction
X-ray diffractograms of the three samples are shown in Fig. 3.
The samples show a 001 peak around 2h ꢀ 2° corresponding to
an interlayer distance of about 4.5 nm, which indicates successful
pillaring [1,3,6]. However, the materials prepared by room temper-
ature swelling (traces a and b) show much sharper peaks and high-
er order reflections (002 and 003) when compared to the one
prepared by high-temperature swelling (trace c). These observa-
tions indicate a higher crystallinity, better structural preservation,
3.4. Multinuclear solid-state NMR investigation
Solid-state NMR was employed to understand the effect of
swelling conditions on the local bonding environment in the pil-
2
9
lared materials. Fig. 5, trace a, shows the Si NMR spectra of
Fig. 3. XRD patterns from the three pillared materials: (a) MCM36-RT-23, (b) MCM36-RT-54 and (c) MCM36-HT-24. The figure has been divided into two parts for better
visualization of 0 0 1 peak and traces have been shifted vertically for clarity.

3
02
S. Maheshwari et al. / Journal of Catalysis 272 (2010) 298–308
Fig. 4. TEM images of MCM36-RT-54 (a and b) and MCM36-HT-24 (c and d) at various magnifications.
Fig. 5. 29Si MAS NMR spectra of (a) MCM36-RT-23 and (b) MCM36-HT-24.
Fig. 6. 29Si CP MAS NMR spectra of (a) MCM36-RT-23 and (b) MCM36-HT-24.
MCM36-RT-23. Three well-resolved peaks (at ꢁ104, ꢁ110,
centration of Q3 species. This is most likely due to the structural
destruction during the swelling process (as already suggested by
the decrease in Si/Al ratio and partial dissolution of zeolite crystals
ꢁ
119 ppm) along with two less resolved shoulders (around ꢁ113
and ꢁ116 ppm) can be seen in the ꢁ104 to ꢁ120 ppm range, which
4
4
correspond to crystallographically non-equivalent Q tetrahedral
noted in TEM images) that results in bond breaking around Q sites
n
3
sites (Q stands for X4ꢁnSi[OSi]
ꢁ
n
) [19–22]. A small peak around
creating Q sites.
3
29Si CP MAS NMR spectra of the two pillared materials are
shown in Fig. 6. Both materials show an intense peak around
98 ppm is attributed to Q sites [23,24] arising from the silanol
groups on the MWW layer surface as well as those present on
the silica pillars. The overall spectrum is very similar to the one ob-
tained from an MWW structure (e.g., MCM-22) [19–22], suggesting
that the local bonding environment stays well preserved in this
material.
3
ꢁ98 to ꢁ100 ppm range, corresponding to the Q silanol species
3
2
[23,24]. The Q and Q species are more rapidly polarized than
4
the Q species in CP mode due to their closer proximity to the sil-
anols and give intense peaks [25]. However, this peak is much
broader in MCM36-HT-24 sample. This could be indicative of dif-
ferent types of silanol groups present in this sample (e.g., silanol
groups on cups of MWW layers, silanol groups created by breaking
The spectrum of MCM36-HT-24 (Fig. 5, trace b) appears very
different. The peaks are much broader and less resolved, indicating
a large dispersion in the local bonding environment. A broad shoul-
der can be seen around the ꢁ100 ppm range, indicating a high con-
4
of Si–O–Si bonds around other Q species). The same peak in

S. Maheshwari et al. / Journal of Catalysis 272 (2010) 298–308
303
much higher concentration of Brønsted acid sites and with a higher
proportion of sites with medium-to-strong acidity. This result is
important from the catalytic point of view, since activity and selec-
tivity of the catalyst will be related to the total number of Brønsted
acid sites and their acid strength distribution. MCM36-RT-54 has a
lower number of total Brønsted acid sites, in agreement with its
lower Al content, but the proportion of stronger acid sites, i.e.,
those able to retain pyridine at 623 K, is high and comparable to
that measured for MCM36-RT-23 (see Table 3).
On the other hand, MCM36-HT-24 has significantly more Lewis
acid sites, suggesting larger amount of extraframework Al species
created by some dealumination of the material during the high-
2
7
temperature swelling. This is in contrast with the Al MAS NMR
spectra (Fig. 7) that show similar amounts of extraframework Al
species (at 0 ppm) for both samples. Miltenburg et. al. [27] re-
ported a similar observation upon alkali treatment of MCM-22 zeo-
lite, which resulted in increase in the Lewis acid sites without any
Fig. 7. 27Al MAS NMR spectra of (a) MCM36-RT-23 and (b) MCM36-HT-24.
MCM36-RT-23 sample is very sharp, indicative of a single type of
silanol species (most likely the ones present on the cups of
MWW layers). A peak around ꢁ90 ppm is also visible in both sam-
2
7
changes in the Al NMR spectra when compared to the parent zeo-
2
7
lite. Previous reports indicate that Al NMR can sometimes under-
estimate the amount of extraframework Al species if they are in a
highly distorted coordination [29]. These Al species are not de-
tected in NMR and are often referred to as ‘NMR silent’ or ‘invisible
aluminum’ in the zeolite literature [30]. Presence of such, ‘NMR si-
lent,’ extraframework aluminum in MCM36-HT-24 can be the rea-
son for the observed discrepancy in acidity and NMR results.
2
ples, which arises due to the Q species [19,20] created by Si–O–Si
bond breaking during swelling or pillaring step. This peak is more
intense in MCM36-HT-24 indicative of greater bond breaking.
2
7
Al NMR results for the pillared materials are shown in Fig. 7.
The peak at 55 ppm in both samples is due to the framework alu-
minum [22,23], while a small peak around 0 ppm is due to the
extraframework aluminum, known to be created by dealumination
during calcination of the samples [23].
3.6. Catalytic testing
3.5. Acidic properties of pillared materials
3.6.1. Cracking of n-decane
Cracking of n-decane was used as a test reaction to evaluate the
intrinsic activity of the pillared materials. Since diffusion limita-
tions are not expected for n-decane, the cracking activity of the cat-
alyst should be related to the number and strength of the Brønsted
acid sites, with no effect of pore structure or crystal size [5].
Fig. 8 shows the n-decane conversion obtained with the three
samples at different contact times. The first-order kinetic rate con-
stants are given in Table 4. It can be seen that the activity of the
catalysts follows the order MCM36-RT-23 > MCM36-RT-
54 > MCM36-HT-24, which correlates very well with the total acid-
ity of the samples as measured by pyridine adsorption–desorption.
Indeed, the turnover frequency values (TOF) calculated from the
first-order kinetic rate and the concentration of Brønsted acid sites
retaining pyridine at 423 K and 523 K give similar values for the
three samples (see Table 4). The results indicate that the higher
activity of pillared samples obtained by room temperature swell-
So far, we discussed the effect of swelling conditions on the
physiochemical and structural properties of resulting pillared
materials. It was found that swelling at high temperature results
in significant dissolution of framework silica and greater destruc-
tion of crystal structure. These changes are likely to affect the acid-
ity and catalytic properties of the samples.
The acidic properties of all three samples were evaluated using
infrared spectroscopy following pyridine adsorption. The results
are summarized in Table 3. MCM36-RT-54 and MCM36-HT-24,
although prepared from the same MWW precursor, present very
different acidity due to significant desilication of the latter sample.
The sample prepared by high-temperature swelling (MCM36-HT-
2
4
(
4) shows slightly lesser number of total Brønsted acid sites (at
23 K) as well as lower amount of strongest Brønsted acid sites
at 623 K), when compared with the pillared material prepared
via room temperature swelling (MCM36-RT-54). Such a reduction
in Brønsted acidity due to desilication has been reported for several
zeolites [26–28].
Since Si/Al ratio plays a deciding role in the final concentration
of acidic sites in zeolites, it is interesting to compare the two pil-
lared materials with similar Si/Al ratio obtained using different
swelling conditions, MCM36-RT-23 and MCM36-HT-24. It can be
easily seen that at a given Si/Al ratio in the final pillared material,
the room temperature swelling procedure gives a sample with
Table 3
Acidity as determined by FT–IR combined with pyridine adsorption at various
temperatures.
Sample
Acidity (
Brønsted
23 K 523 K 623 K
lmol pyridine/g)
B623/ Lewis
B423
4
423 K 523 K 623 K
MCM36-RT-54
MCM36-HT-24
MCM36-RT-23
56
62
108
50
47
98
33
21
65
0.60
0.34
0.60
19
61
40
17
44
31
17
42
30
Fig. 8. n-Decane cracking. Conversion obtained using MCM36-RT-23 (d), MCM36-
RT-54 (j) and MCM36-HT-24 (D).

3
04
S. Maheshwari et al. / Journal of Catalysis 272 (2010) 298–308
Table 4
First-order kinetic rate constants and turnover frequencies in the n-decane cracking at
73 K and 60 s TOS. Selectivities and representative ratios obtained at 35% n-decane
fragmented crystals with lower secondary porosity, possibly
increasing recracking of molecules before they can diffuse out from
the catalyst, and thus producing more gases and less liquid fuels.
The improved structural preservation can also be partially respon-
sible for the high gases production from MCM36-RT-23.
As observed for n-decane cracking, the olefin/paraffin ratios in
the gaseous products are lowest for MCM36-RT-23 due to a high
extent of hydrogen transfer reactions in this sample, owing to its
higher density of Brønsted acid sites. This higher hydrogen transfer
capacity is in agreement with the higher coke selectivity obtained
with this sample.
7
conversion.
MCM36-
RT-23
MCM36-
RT-54
MCM36-
HT-24
K (s 1) ꢂ 10
ꢁ
2
2.60
2.4
2.6
1.07
1.9
2.1
0.82
1.3
1.7
ꢁ
1
ꢁ1
4
4
K/B423 (s /mol Py g ) ꢂ 10
ꢁ
1
ꢁ1
K/B523 (s /mol Py g ) ꢂ 10
C1–C4 (wt.%)
C5+ (wt.%)
Coke (wt.%)
22.1
11.4
1.5
21.2
10.8
3.0
19.5
10.5
5.1
Gases composition (wt.%)
C1 + C2
Propane
Propylene
n-Butane
Isobutane
Total butenes
Isobutene
0.9
4.2
4.3
3.2
5.9
3.6
1.6
1.1
3.4
4.9
2.6
5.2
4.1
1.8
1.0
2.9
5.2
2.1
4.1
4.2
1.8
3.6.3. Xylene isomerization–disproportionation
The isomerization and disproportionation of meta-xylene has
been proposed as a test reaction, which gives important informa-
tion concerning the structure of zeolitic catalysts [18,35–40]. In
this reaction, the para-/ortho-xylene ratio depends on the size of
the micropores, and on the possible presence of lobes, cages or
crossing channels. Moreover, since meta-xylene isomerization is
monomolecular, whereas disproportionation is a bimolecular reac-
tion involving a bulkier transition state, the relative extent of these
two reactions (isomerization/disproportionation ratio) will indi-
cate the possible presence of void spaces, large enough to allow
the bimolecular process to occur.
Ratios
Propylene/propane
Butenes/butanes
Isobutene/isobutane
1.02
0.39
0.28
0.15
0.66
1.44
0.52
0.34
0.21
0.71
1.78
0.67
0.44
0.23
0.79
(
C1 + C2)/isobutane
C3/C4
The initial conversion of meta-xylene over the three pillared
samples is shown in Fig. 10. It can be clearly seen that MCM36-
RT-23 is again the most active sample due to its high Brønsted acid
site density. The other two samples have very similar activity,
although a closer look indicates a slightly higher value for
MCM36-HT-24. This may be due to easier reactant accessibility
as a result of more fragmented structure created by partial dissolu-
tion of the crystal during the high-temperature swelling.
ing is related to the preservation of a larger amount of zeolite-
bridging hydroxyl groups in these materials. Table 4 shows the
selectivities to different cracking products obtained with all three
pillared materials, as well as some selectivity parameters, such as
the olefin to paraffin ratio in the C3 and C4 fractions and the isobu-
tene/isobutane ratio, which give information on the extent of
hydrogen transfer that occurs during the cracking process. The
higher the above ratios, the lower the extent of hydrogen transfer
reactions [31–33]. By taking that into account, it can be seen that
MCM36-RT-23 gives the highest hydrogen transfer. This sample
is also the one with the highest Brønsted acid site density (Table 3),
and therefore it will favor bimolecular reactions like hydrogen
transfer to a larger degree [32,34]. Selectivity to gases and liquids
is very similar for the three catalysts. The amount of coke is lower
with the well-pillared samples due to the presence of 10 MR chan-
nels in each layer where coke formation is hindered.
It is well known [18,35–40] that the shape and dimensions of
the zeolite pores have great influence over the para/ortho selectiv-
ity observed during meta-xylene isomerization. The selectivity ob-
tained over medium pore (10MR) zeolites is higher (>2.5) due to
differences in the diffusion of both isomers through the narrow
channels of these zeolites [41]. On the other hand, the large pore
zeolites (12MR) show para/ortho selectivities which are only
slightly above the equilibrium value (para/ortho ꢀ 1), often in the
range of 1.0–1.4 [38]. Zeolites containing both 10MR and 12MR
pores show intermediate behavior [12,37]. The three pillared sam-
ples analyzed here give para/ortho ratios between 1.4 and 2.0,
which are typical of materials containing both medium (10MR)
and large pores (12MR) [12,37]. However, MCM36-RT-54 has
slightly higher para/ortho ratio (ꢀ2 when compared to ꢀ1.5 for
the other samples), which is very close to that obtained for a
MCM-22 material [12]. Minimal structural degradation during
the synthesis of this sample results in preservation of 10MR pore
channel within the layers, which favors para-xylene production
[12]. On the other hand, the para/ortho ratio for MCM36-RT-23 is
lowest despite its structural similarity to MCM36-RT-54. This is
likely due to the larger number of acid sites in this sample, together
with a larger proportion of the sites with strongest acidity, which
increases the average lifetime of adsorbed species favoring second-
ary rearrangement reactions and bringing down the selectivity to
para-xylene.
3.6.2. Vacuum gasoil cracking
While n-decane cracking measures the intrinsic activity of the
material, vacuum gasoil (VGO) cracking activity depends on the
accessibility of the bulky hydrocarbon molecules to the acidic sites
[
5]. Moreover, for the three pillared layered zeolites (for which
accessibility to the acidic sites present in the 12MR cups is
expected to be similar), one may expect the overall VGO cracking
activity to be a function of total concentration of Brønsted acid
sites. The results obtained for VGO cracking (Fig. 9) support this
hypothesis, since MCM36-RT-23, which has the highest amount
of Brønsted acid sites, is the most active sample. The other two
samples have very similar acid site density and hence similar
activity.
When comparing catalyst selectivity plots (Fig. 9), it can be seen
that the most active sample, MCM36-RT-23, is less selective to li-
quid fuels (gasoline and diesel) and more selective to gases and
coke. This is most likely due to higher acid site concentration that
results in the secondary cracking of molecules from diesel and gas-
oline to give gases. Interestingly, when comparing MCM36-RT-54
and MCM36-HT-24, which give very similar activity and have sim-
ilar acid site concentration, the former is slightly more selective to-
ward gases and coke and less selective toward liquids. A higher
degree of structural preservation in MCM36-RT-54 results in less
The disproportionation of xylenes to give trimethylbenzene and
toluene is a bimolecular secondary reaction, which could be fa-
vored by the presence of large cages and high acid site density
[42]. Fig. 10 shows that the overall selectivity of all three samples
toward isomerization is high (ratio > 10) indicating the predomi-
nant effect of the 10MR pores. In the case of MCM36-RT-23, the
higher density of acid sites enhances the bimolecular reaction
resulting in a lower isomerization to disproportionation ratio [43].

S. Maheshwari et al. / Journal of Catalysis 272 (2010) 298–308
305
Fig. 9. Vacuum gasoil cracking. Conversion, selectivity and olefin/paraffin ratio in the gas products obtained using MCM36-RT-23 (d), MCM36-RT-54 (j) and MCM36-HT-24
(D
).
3
.6.4. Benzene and toluene alkylation
Aromatics alkylation serves as a useful test reaction to study
the olefin to produce C6 and C9 olefins that can be further trans-
formed through cracking, isomerization and alkylation giving ole-
fins and other alkylbenzenes (toluene, cumene, butylbenzenes,
etc.).
zeolite topology and is of significant commercial interest in the
petrochemical industry [18,37,43,44].
The alkylation of aromatic molecules with propanol occurs
through an electrophilic substitution and is considered to proceed
via a carbenium ion-type mechanism [45,46]. The alcohol is con-
verted into activated alkene by dehydration on the acid sites of
the zeolite. The resulting olefin can follow two major reaction
pathways: (1) alkylation with the aromatic, which can further re-
act to give di- and tri-alkyl-aromatics; and (2) oligomerization of
The activity and selectivity of the three pillared samples for
alkylation of benzene with isopropanol has been obtained at short
time on stream (TOS) to minimize the effect of catalyst deactiva-
tion. The results presented in Fig. 11 show that MCM36-RT-23
gives the highest benzene conversion in agreement with its high
Brønsted acid site density and with the catalytic results reported
above. Selectivity to cumene (CUM) is high (between 80% and

3
06
S. Maheshwari et al. / Journal of Catalysis 272 (2010) 298–308
Fig. 10. Meta-xylene isomerization. Conversion (A), selectivity (B) and isomerization/disproportionation ratio (C) obtained using MCM36-RT-23 (d), MCM36-RT-54 (j) and
MCM36-HT-24 ( ).
D
Fig. 11. Alkylation of benzene with isopropanol. Conversion (A), selectivity to cumene (B) and to cumene + DIPB (C) obtained using MCM36-RT-23 (d), MCM36-RT-54 (j)
and MCM36-HT-24 ( ).
D

S. Maheshwari et al. / Journal of Catalysis 272 (2010) 298–308
307
Fig. 12. Alkylation of toluene with isopropanol. Conversion (A), selectivity to cymene (B) and to cymene + DIPT (C) obtained using MCM36-RT-23 (d), MCM36-RT-54 (j) and
MCM36-HT-24 ( ).
D
9
0%) for all three samples, and no significant differences are ob-
Brønsted acid site density and a larger proportion of stronger acidic
sites for a given Si/Al ratio, when compared to the one prepared
using the method previously reported in the literature. The higher
acidity makes these materials very active for a host of catalytic
reactions such as cracking, isomerization and alkylation, and better
structural preservation affects the product distribution obtained
from these reactions.
served. As diisopropylbenzene (DIPB) can be converted to cumene
in a second stage of the process by transalkylation with unreacted
benzene, it can also be considered as a product of interest. Close to
1
00% (CUM + DIPB) selectivity values are obtained with all three
samples.
Although cumene is the major propylbenzene obtained, unde-
sirable n-propylbenzene (NPB) can also be formed by secondary
bimolecular transalkylation reactions between cumene and ben-
zene [47,48]. Only traces of NPB are produced under our reaction
conditions, as expected from these types of materials with non-
crossing 10MR channels.
Results from alkylation of toluene with isopropanol to produce
cymene and other products are shown in Fig. 12. The product dis-
tribution and trends are very similar to those obtained for benzene
alkylation. All three materials show high selectivity to cymene
Acknowledgments
We gratefully acknowledge Sandeep Kumar for the TEM images
and Dr. Sonjong Hwang for the NMR spectra of the MCM-36 mate-
rials. We thank Professor Frank Bates for his input since the early
stages of this work and Professor Aditya Bhan for suggestions dur-
ing the writing of this manuscript. This work was made possible
with the financial support from the NSF (CBET-0403574 (NIRT)
and 0327811), and PROMETEO and Consolider from Generalitat
Valenciana and Ministerio de Ciencia e Innovación. Parts of this
work were carried out in the University of Minnesota I.T. Charac-
terization facility, which receives partial support from the NSF
through the NNIN program. M.T. Portilla gratefully acknowledges
CSIC for JAE fellowship.
(
>90%) and MCM36-RT-23 presents the highest activity due to its
higher acid density. Only traces of NPT are formed in our reaction
conditions.
4
. Conclusions
A modified method to produce MCM-36 materials with higher
References
crystallinity and well-preserved layer structure has been demon-
strated. Swelling at room temperature (instead of high tempera-
tures) and extensive washing with water are keys to achieve
such high structural integrity in the pillared materials. Elemental
analysis, TEM, XRD and nitrogen adsorption revealed that the
method preserves overall particle morphology and microporosity
within the layers and prevents dissolution of the framework silica.
The improved structural preservation is reflected in the acidity and
catalytic activity of the resulting pillared materials. The MCM-36
materials prepared by the method presented here exhibit higher
[
1] C.T. Kresge, W.J. Roth, K.G. Simmons, J.C. Vartuli, WO Patent 92/011935, Mobil
Oil Corporation, 1992.
[2] S. Laforge, P. Ayrault, D. Martin, M. Guisnet, Appl. Catal. A 279 (2005) 79.
[
3] W.J. Roth, C.T. Kresge, J.C. Vartuli, M.E. Leonowicz, A.S. Fung, S.B. McCullen,
Stud. Surf. Sci. Catal. 94 (1995) 301.
[
4] J. Kornatowski, J.O. Barth, K. Erdmann, M. Rozwadowski, Micropor. Mesopor.
Mater. 90 (2006) 251.
[
[
5] A. Corma, V. Fornes, J. Martinez-Triguero, S.B. Pergher, J. Catal. 186 (1999) 57.
6] Y.J. He, G.S. Nivarthy, F. Eder, K. Seshan, J.A. Lercher, Micropor. Mesopor. Mater.
25 (1998) 207.
[7] J.-O. Barth, J. Kornatowski, J.A. Lercher, J. Mater. Chem. 12 (2002) 369.

3
08
S. Maheshwari et al. / Journal of Catalysis 272 (2010) 298–308
[
8] R. Ravishankar, D. Bhattacharya, N.E. Jacob, S. Sivasanker, Micropor. Mater. 4
1995) 83.
9] R. Schenkel, J.O. Barth, J. Kornatowski, J.A. Lercher, Stud. Surf. Sci. Catal. 142A
2002) 69.
[26] J.C. Groen, S. Abello, L.A. Villaescusa, J. Perez-Ramirez, Micropor. Mesopor.
Mater. 114 (2008) 93.
[27] A. Miltenburg, J. Pawlesa, A.M. Bouzga, N. Zilkova, J. Cejka, M. Stoecker, Top.
Catal. 52 (2009) 1190.
(
[
(
[
[
[
10] P. Wu, D. Nuntasri, J. Ruan, Y. Liu, M. He, W. Fan, O. Terasaki, T. Tatsumi, J. Phys.
Chem. B 108 (2004) 19126.
11] S. Maheshwari, E. Jordan, S. Kumar, F.S. Bates, R.L. Penn, D.F. Shantz, M.
Tsapatsis, J. Am. Chem. Soc. 130 (2008) 1507.
[28] J.S. Jung, J.W. Park, G. Seo, Appl. Catal. A 288 (2005) 149.
[29] P.P. Man, J. Klinowski, J. Chem. Soc., Chem. Commun. (1988) 1291.
[30] R. Rachwalik, Z. Olejniczak, J. Jiao, J. Huang, M. Hunger, B. Sulikowski, J. Catal.
252 (2007) 161.
[31] J.O. Petunchi, G.A. Sill, W.K. Hall, J. Catal. 141 (1993) 323.
[32] A. Corma, M. Faraldos, A. Martinez, A. Mifsud, J. Catal. 122 (1990) 230.
[33] C. Franco Parra, M.R. Goldwasser, F. Fajula, F. Figueras, Appl. Catal. 17 (1985)
217.
[34] L.A. Pine, P.J. Maher, W.A. Wachter, J. Catal. 85 (1984) 466.
[35] M. Guisnet, N.S. Gnep, S. Morin, Micropor. Mesopor. Mater. 35–36 (2000) 47.
[36] P.A. Jacobs, J.A. Martens, Stud. Surf. Sci. Catal. 28 (1986) 23.
[37] F.J. Llopis, G. Sastre, A. Corma, J. Catal. 227 (2004) 227.
[38] J.A. Martens, J. Perez-Pariente, E. Sastre, A. Corma, P.A. Jacobs, Appl. Catal. 45
(1988) 85.
[39] D.H. Olson, W.O. Haag, ACS Symp. Ser. 248 (1984) 275.
[40] L.B. Young, S.A. Butter, W.W. Kaeding, J. Catal. 76 (1982) 418.
[41] P.B. Weisz, Pure Appl. Chem. 52 (1980) 2091.
12] A. Corma, C. Corell, F. Llopis, A. Martinez, J. Perez-Pariente, Appl. Catal. A 115
(
1994) 121.
[
[
[
13] A. Corma, J.B. Monton, A.V. Orchilles, Appl. Catal. 16 (1985) 59.
14] R. Kumar, P. Ratnasamy, J. Catal. 116 (1989) 440.
15] A. Corma, V. Fornes, S.B. Pergher, T.L.M. Maesen, J.G. Buglass, Nature 396
(
1998) 353.
[
[
16] C.A. Emeis, J. Catal. 141 (1993) 347.
17] A. Corma, C. Martinez, F. Ortega, O. Bermudez, Appl. Catal. A 230 (2002)
1
11.
[
[
18] A. Corma, V.I. Costa-Vaya, M.J. Diaz-Cabanas, F.J. Llopis, J. Catal. 207 (2002) 46.
19] R. Aiello, F. Crea, F. Testa, G. Demortier, P. Lentz, M. Wiame, J.B. Nagy, Micropor.
Mesopor. Mater. 35–36 (2000) 585.
[
20] M.A. Camblor, A. Corma, M.-J. Diaz-Cabanas, C. Baerlocher, J. Phys. Chem. B 102
(
1998) 44.
[42] J. Perez-Pariente, E. Sastre, V. Fornes, J.A. Martens, P.A. Jacobs, A. Corma, Appl.
Catal. 69 (1991) 125.
[
[
21] G.J. Kennedy, S.L. Lawton, M.K. Rubin, J. Am. Chem. Soc. 116 (1994) 11000.
22] S.L. Lawton, A.S. Fung, G.J. Kennedy, L.B. Alemany, C.D. Chang, G.H. Hatzikos,
D.N. Lissy, M.K. Rubin, H.-K.C. Timken, J. Phys. Chem. 100 (1996) 3788.
23] W. Kolodziejski, C. Zicovich-Wilson, C. Corell, J. Perez-Pariente, A. Corma, J.
Phys. Chem. 99 (1995) 7002.
[43] A. Corma, V. Martinez-Soria, .E. Schnoeveld, J. Catal. 192 (2000) 163.
[44] A. Corma, F.J. Llopis, C. Martinez, G. Sastre, S. Valencia, J. Catal. 268 (2009) 9.
[45] J.R. Anderson, T. Mole, V. Christov, J. Catal. 61 (1980) 477.
[46] Y. Du, H. Wang, S. Chen, J. Mol. Catal. A: Chem. 179 (2002) 253.
[47] J. Cejka, B. Wichterlova, Catal. Rev. – Sci. Eng. 44 (2002) 375.
[48] E.G. Derouane, H. He, S.B.D.-A. Hamid, I.I. Ivanova, Catal. Lett. 58 (1999) 1.
[
[
[
24] D. Ma, F. Deng, R. Fu, X. Han, X. Bao, J. Phys. Chem. B 105 (2001) 1770.
25] S.L. Burkett, M.E. Davis, J. Phys. Chem. B 98 (1994) 4647.