Template-controlled acidity and catalytic activity of ferrierite crystals

Article abstract of DOI:10.1016/j.jcat.2009.02.017

A synthesis strategy to tailor the acid sites location in ferrierite crystals has been developed. The zeolite catalysts were synthesised in fluoride medium using different combinations of organic structure directing agents (SDAs) in the absence of inorgan

Full text of DOI:10.1016/j.jcat.2009.02.017

Journal of Catalysis 263 (2009) 258–265
Contents lists available at ScienceDirect
Journal of Catalysis
www.elsevier.com/locate/jcat
Template-controlled acidity and catalytic activity of ferrierite crystals
∗
A.B. Pinar, C. Márquez-Álvarez, M. Grande-Casas, J. Pérez-Pariente
Instituto de Catálisis y Petroleoquímica (CSIC), c/ Marie Curie 2, Cantoblanco, 28049-Madrid, Spain
a r t i c l e i n f o
a b s t r a c t
Article history:
A synthesis strategy to tailor the acid sites location in ferrierite crystals has been developed. The zeolite
catalysts were synthesised in fluoride medium using different combinations of organic structure directing
agents (SDAs) in the absence of inorganic cations. Therefore, the negative charge associated to the
incorporation of aluminium to the framework was compensated exclusively by the positive charge of
the organic SDAs. In this way, Al sitting in the zeolite framework was driven by the specific location of
the different SDA molecules within the zeolite void volume. Following this synthesis strategy, it has been
found that the distribution of strongly acidic hydroxyl groups in the proton form of the zeolites obtained
after removal of the organic templates was dependent on the combination of organic molecules used as
SDAs. Moreover, the catalytic activity of the zeolites in m-xylene and 1-butene isomerisation increased as
the relative population of strong Brönsted acid groups in sterically constrained sites inside the ferrierite
cavity decreased.
Received 3 December 2008
Revised 26 January 2009
Accepted 15 February 2009
Available online 6 March 2009
Keywords:
Ferrierite
Zeolite synthesis
Fluoride medium
Al sitting
Acid sites distribution
FTIR
© 2009 Elsevier Inc. All rights reserved.
Pyridine
Isomerisation
m-Xylene
1-Butene
1. Introduction
channels run along the [001] direction. The intersection of these
6-MR channels with the above mentioned 8-MR channels leads
to the formation of the “ferrierite cavity”, a [82626458] cage ac-
cessible through 8-MR windows [5]. The ferrierite framework was
first determined in the space group Immm [6] although reductions
of symmetry have been reported [7–9]. In Fig. 1 a representa-
tion of the structure can be seen, showing the 4 non-equivalent
T atoms (in space group Immm) and their accessibility from the
10-MR channel and the cage. It has been shown before [10–12]
that this zeolite can be conveniently synthesised in the absence
of alkaline cations by using an appropriate combination of small
and large structure directing agents (SDA). In particular, the hy-
drothermal treatment of synthesis gels containing both 1-benzyl-
1-methylpyrrolidinium (bmp) and tetramethylammonium (TMA)
cations in fluoride medium allows the crystallisation of zeolite fer-
rierite. Furthermore, computational studies reveal that TMA cations
are located inside the small cages, while the large and elongated
bmp cations are placed in the 10-MR straight channel. Hence, both
cations cooperate to build up the ferrierite structure. Moreover, in
the absence of either bmp or TMA, ferrierite was not obtained,
thus showing the key role of both cations in the crystallisation
of this zeolite in these synthesis conditions. As far as no alka-
line cations are present in the structure, the charge compensation
of the aluminium atoms located in the framework is due to the
cationic SDA only. Using CP-MAS NMR and heteronuclear correla-
tion NMR, Lobo et al. [13] found that the positively charged frag-
ment of the SDA is preferentially ordered near the negative frame-
Activity and selectivity patterns of a particular acid zeolite in
different catalytic reactions are recognised to depend upon the
combination of several factors, the number and distribution of acid
sites in the framework among them. Controlling the distribution
of framework aluminium atoms, and hence that of the acid sites,
in zeolites is a much sought objective in zeolite research, as it
would have a strong impact on catalytic activity and selectivity.
Indeed, it has been shown for ZSM-5 [1–3] and MCM-22 [4] that
Al distribution is dependent on the Si/Al ratio and on the synthesis
procedure.
The presence of different types of cavities for a given zeolite
topology introduces a high degree of complexity in the distribution
of acid sites. Many interesting zeolite catalysts possess precisely
the complex void architecture that makes them suitable candidates
for the development of specific synthesis strategies envisaging to
control the aluminium sitting and hence the acidity of the struc-
ture. Ferrierite is one of such materials. Its structure consists of
channels of 10-membered rings (10-MR, a ring constituted by 10 T
atoms linked through T–O–T bonds, where “T” represents a Si or an
Al atom tetrahedrally coordinated to 4 oxygen atoms) running par-
allel to the crystallographic [001] direction, which are intersected
by 8-MR channels parallel to the [010] direction. In addition, 6-MR
Corresponding author. Fax: +34 91 5854760.
E-mail address: jperez@icp.csic.es (J. Pérez-Pariente).
*
0021-9517/$ – see front matter © 2009 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2009.02.017

A.B. Pinar et al. / Journal of Catalysis 263 (2009) 258–265
259
Benzylpyrrolidine was methylated with methyl iodide (50% mo-
lar excess, Fluka) in ethanol. The stirring was maintained for 5
days at room temperature and ethanol was then removed under
◦
vacuum at 60 C. The resulting solid benzylmethylpyrrolidinium
iodide was exhaustively washed with diethyl ether until the result-
ing liquid was colourless. The composition of the collected yellow
solid was determined by chemical analysis (calculated values, in
wt%: C 47.5; N 4.6; H 5.9; found: C 47.3; N 4.7; H 5.8), while 13C
CP MAS-NMR confirmed the integrity of the benzylmethylpyrroli-
dinium cation (bmp). The iodide salt was then converted into the
corresponding hydroxide salt (bmpOH) by ion exchange through
an Amberlyst IRN78 resin (exch. cap.: 4 meq/g, Supelco). The ob-
tained solution of bmpOH was filtered to remove the resin and
titrated with 0.05 M HCl (Panreac) using phenolphthalein (Aldrich)
as an indicator.
2.2. Synthesis of the zeolites
Fig. 1. Framework structure of FER showing the 4 non-equivalent T-atoms (in space
group Immm). T1, T2 and T3 (black atoms) are accessible from the 10-MR channel
and from the FER cavity. T4 (white atom) is accessible only from the cavity. Bridging
oxygen atoms have been omitted for clarity.
Zeolite products synthesised in fluoride medium were obtained
following the synthesis procedure reported earlier [10] from syn-
thesis gels in which the molar ratio organic/T atoms was kept
constant. The molar gel composition for the samples with bmp
and TMA as SDAs was: 0.969 SiO2:0.031 Al2O3:0.06 TMAOH:0.48
bmpOH:0.48 HF:4.6 H2O. For the samples prepared with TMA and
pyrrolidine as SDAs, the same gel composition was employed, sub-
stituting bmp by pyrrolidine. For the samples prepared in the pres-
ence of pyrrolidine as the only SDA the same molar composition
was used but with molar ratio pyrr:(Si + Al) = 0.54, thus maintain-
ing the ratio organic compounds/T atoms constant. The Si/Al ratio
in the gel was 15.7 in all the cases. In the synthesis gels in which
TMA is one of the SDAs, the ratio TMA/Al was 1.
work charge in ZSM-12, and suggested that the location of acid
sites in the zeolite might be controlled by the charge distribution
of the SDA. It could be then conceived that the framework sites
occupied by aluminium in ferrierite would be eventually depen-
dent upon the specific combination of SDAs used in the synthesis.
Moreover, it has been evidenced that the extra-framework charge-
compensating cation distribution in ferrierite zeolites changes as a
function of their Si/Al ratio [14], and from this feature it has been
concluded that the Al distribution among the different T sites is
affected as well.
All these samples underwent a post-synthesis treatment in or-
der to get their catalytically active form. They were heated in
nitrogen flow up to 200 C and kept at this temperature for 2 h,
Based upon these previous studies, we report in this work a
new approach aiming to influence the aluminium location in fer-
rierite. Our hypothesis is that the use of a combination of organic
molecules as SDAs can influence the distribution of acid sites of
the zeolite obtained, i.e., the particular combination of SDAs em-
ployed in each case could have an effect on the population of
aluminium in the different T sites of the zeolite. Therefore, our
synthesis strategy consists on the use of a combination of different
organic molecules acting as co-structure directing agents, in the
absence of sodium cations. This is an important aspect, as the neg-
ative charge associated to the incorporation of aluminium to the
tetrahedral sites by isomorphic substitution of silicon will be com-
pensated exclusively by the positive charge of the organic SDAs.
Following the previous synthesis approach using bmp and TMA as
SDAs, ferrierite samples have been obtained from gels in which
the bulky bmp cation has been substituted by pyrrolidine (pyrr)
and then both bmp and TMA have been replaced by pyrrolidine,
i.e., only pyrrolidine has been used as SDA, always maintaining a
constant total concentration of organic compounds. For comparison
purposes, a ferrierite sample has been prepared by using pyrroli-
dine at high pH in presence of sodium cations [15]. The acidity of
the calcined materials has been investigated by FTIR spectroscopy
and pyridine adsorption, and their catalytic activity in m-xylene
and 1-butene isomerisation has also been determined.
◦
followed by a treatment with a flow of oxygen enriched in ozone
◦
(60 mL/min, ca. 2 vol% ozone) at 200 C to remove, in mild con-
ditions, the organic molecules employed as SDAs. An ECO-5 ozone
generator (Salveco Proyectos, S.L.) was employed to produce the
ozone-enriched oxygen stream. In the final step prior to their use
as catalysts the samples were heated in an air flow for 2 h up to
◦
550 C and kept at this temperature for 2 additional hours.
The sample synthesised in alkaline medium following the con-
ventional procedure, FER1, crystallised from a gel with the fol-
lowing molar composition: 0.969 SiO2:0.031 Al2O3:0.323 pyrro-
◦
lidine:0.049 Na2O:11.95 H2O. The gel was heated at 175 C and
autogeneous pressure for selected periods of time, and then it was
filtrated, washed and dried overnight. The same treatment in an
ozone-enriched stream was applied to remove the organic mate-
rial occluded within the zeolite. The calcined product was subse-
quently ion-exchanged with a 1 M ammonium chloride solution
+
(Probus, 99.5 wt%) to eliminate Na cations, which were substi-
tuted by ammonium cations in this process. For this purpose the
zeolite was placed in a round-bottomed flask with the ammonium
chloride solution (1 g zeolite/100 mL solution) and it was stirred
◦
at 80 C for 8 h under reflux. Then the solid was recovered by
filtration and washed exhaustively with deionised water to elimi-
nate chloride and sodium ions. This procedure was repeated twice,
maintaining the stirring for 12 and 24 h, respectively. Complete
removal of chlorine ions was confirmed by the absence of AgCl
(white solid precipitated) when adding silver acetate solution to
the filtrate. The sample was dried overnight and calcined in an air
atmosphere for 3 h at 350 C and 3 h at 550 C to decompose am-
monium ions, thus obtaining the zeolite in its proton form.
Hereafter, the ferrierite samples will be denoted making ref-
erence to the organic cations employed as SDAs in their syn-
thesis, separated by a dash if two different molecules are em-
2. Experimental and methods
2.1. Synthesis of the SDAs
◦
◦
The synthesis of bmp was carried out by methylation of ben-
zylpyrrolidine as reported earlier [10]. The purity of the tertiary
amine was assessed by thin layer chromatography (hexane/ethyl
acetate as solvent) and chemical analysis (calculated values, in
wt%: C 82.0; N 8.7; H 9.3; found: C 81.5; N 8.7; H 9.9).

260
A.B. Pinar et al. / Journal of Catalysis 263 (2009) 258–265
Table 1
Summary of the syntheses carried out with the different combinations of SDAs. The
◦
◦
heating temperature was 175 C for FER1 samples and 150 C for the rest of the
samples.
Sample
SDA
t (days)
pH gel
TMA + bmp (fluoride medium)
10
20
30
83
FER-bmp-TMA
10.2
Pyrrolidine + TMA (fluoride medium)
7
10
FER-pyrr-TMA
8.6
Pyrrolidine (fluoride medium)
Pyrrolidine (alkaline medium)
7
10
Fig. 2. XRD patterns of samples FER-pyrr (a), FER-pyrr-TMA (b) both obtained after
10 days of hydrothermal treatment, FER1 (c) obtained after 14 days and FER-bmp-
TMA (d) obtained after 20 days. The peak highlighted with an asterisk corresponds
to an unidentified dense phase. In all cases the pattern corresponds to the calcined
sample.
FER-pyrr
FER1
7.4
14
12.1
Scharlau (99%). These experiments were carried out in a differen-
tial fixed-bed reactor operating with continuous feed and nitrogen
as carrier gas at atmospheric pressure. All the samples were previ-
ployed together as a combination of SDAs: FER-bmp-TMA (syn-
thesised with 1-benzyl-1-methylpyrrolidinium and TMA), FER-pyrr-
TMA (with pyrrolidine and TMA), FER-pyrr (with pyrrolidine) and
FER1 (with pyrrolidine in aqueous alkaline medium in the presence
of sodium). Table 1 summarises the syntheses carried out with the
different combinations of SDAs.
◦
ously activated in air flow (100 ml/min) at 450 C for 1 h. Then the
◦
flow was changed to N2 and the temperature decreased to 350 C,
keeping these conditions for 1 more hour, after which the reaction
is performed at the same temperature. In order to obtain conver-
sions of about 5%, the contact time was varied, while the molar
ratio was fixed at N2/m-xylene = 5.3. The reaction products were
analysed by gas chromatography using a 2 m × 3 mm OD packed
column filled with 16% DC-200 methylsilicone and 3% Bentone 34
on 80–100 mesh ChromosorbW. A Varian 3800 gas chromatograph
provided with a flame ionisation detector (FID) was used in the
analysis. In order to compare the activity of the catalyst in the ab-
sence of deactivation, the initial activity (V0) of all products were
calculated by extrapolating the conversion at time zero of reaction.
The transformation of 1-butene was carried out in an automatic
2.3. Characterisation of the solids
Solid products were characterised by XRD (PANalytical X´Pert
PRO-MPD diffractometer, CuKα radiation), thermal analysis (Per-
◦
kin–Elmer TGA7 instrument, heating rate 20 C/min, air flow
◦
30 mL/min, temperature range 20–900 C), chemical CHN analy-
sis (Perkin–Elmer 2400 CHN analyser), ICP-AES (Optima 3300 DV
Perkin–Elmer) and SEM/EDX (Jeol JSM 6400 Philips XL30, operat-
ing at 20 kV). 27Al MAS NMR spectra were recorded with a Bruker
AV 400 spectrometer using a BL4 probe. Pulses of 1 μs were used
to flip the magnetisation, with π/12 rad, and delays of 1 s be-
tween two consecutive pulses. The spectra were recorded while
spinning the samples at ca. 11 kHz.
◦
fixed-bed reactor under the following conditions: 400 C, atmo-
spheric pressure, N2/butene molar ratio 4, weight hourly space
velocity (WHSV) of 25 h−1. The catalysts samples ≈100 mg (pel-
letised and sieved to 0.84–0.42 mm) were previously pre-treated in
◦
situ under nitrogen flow for 1 hour at 400 C. The products were
−1
analysed on-line by gas chromatography using a capillary column
(PLOT Al2O3 50 m × 530 μm). A Varian Cp-3880 gas chromato-
graph provided with a flame ionisation detector (F.I.D.) was used in
the analysis. The conversion into the different products was calcu-
lated by considering n-butene isomers as the reactant. The activity
at 30 min time-on-stream was used to compare catalytic perfor-
mance.
Transmission FTIR spectra were recorded in the 400–4000 cm
range, at 4 cm
−1
resolution, using a Nicolet 5ZDX FTIR spectrome-
ter provided with an MCT detector. The samples were pressed into
self-supporting wafers (ca. 6 mg/cm2 thickness), placed in a glass
◦
cell with CaF2 windows, and activated in vacuum at 400 C for 8 h.
Pyridine adsorption was carried out on activated samples kept at
◦
150 C, and followed in dependence of time. The sample wafer was
◦
put in contact with pyridine vapour (8 Torr) at 150 C for a given
3. Results and discussion
period of time, subsequently degassed at the same temperature for
30 min, and then the FTIR spectrum was recorded. Pyridine was
again dosed into the cell to allow further adsorption of pyridine on
the wafer, and the cell was evacuated for 30 min prior to recording
the FTIR spectrum. This adsorption procedure was repeated several
times in order to obtain a series of FTIR spectra at increasing to-
tal contact time of the sample wafer with pyridine. The spectra
recorded after pre-treatment and after every pyridine adsorption
step were normalised to a sample thickness of 6 mg/cm2.
3.1. Synthesis of the materials
Pure zeolite ferrierite crystallised from the three different syn-
thesis gels prepared in fluoride medium. Fig. 2 collects the XRD
patterns of the corresponding products. It can be observed that
the solids obtained in the absence of bmp present sharper diffrac-
tion peaks, which suggest a higher crystallinity for these samples.
Crystal morphology of the materials was studied by scanning elec-
tron microscopy (Fig. 3). The three samples synthesised in the
absence of bmp cation are constituted by agglomerates of crystals
that exhibit the conventional platelike morphology characteristic
of ferrierite [16,17]. However, there are differences in crystal size.
2.4. Catalytic activity
The catalytic activity of the samples was tested in the iso-
merisation of m-xylene reaction; m-xylene was purchased from

A.B. Pinar et al. / Journal of Catalysis 263 (2009) 258–265
261
corresponding to these molecules. These results suggest that the
organic molecules resisted the hydrothermal treatment. Within the
void volume of the ferrierite structure, there are two possible loca-
tions for the organic guest molecules. They can be accommodated
either in the main 10-MR channel or in the ferrierite cavity. In
principle, the small size of both pyrrolidine and TMA allows them
to be occluded in any of these two locations. XRD powder diffrac-
tion studies have been recently carried out in an attempt to find
out the location of each organic molecule within the framework,
which will provide a better understanding of their role in the crys-
tallisation process. The organic content corresponds in both FER-
pyrr and FER-pyrr-TMA samples to approximately 4 molecules per
unit cell. The ferrierite framework contains two cavities per unit
cell, which can host only one molecule of either TMA or pyrroli-
dine. Therefore, these data suggest that, of the 4 organic molecules
per unit cell, 2 of them are occluded in the cavity and the other
two molecules in the 10-MR channel. Preliminary XRD Rietveld re-
finement results of both samples support this assignment [20].
For the samples crystallised from gels containing bmp and TMA,
chemical analysis and 13C MAS NMR spectrum (not shown) con-
firmed that these organic cations kept their integrity within the
zeolite. Furthermore, molecular mechanics calculations revealed
the templating role of both cations. The most stable configuration
of the system corresponds to the location of bmp in the 10-MR
channels of the ferrierite structure, being too large to be occluded
inside the small cages, while TMA is occluded within the FER cages
[10]. This result was confirmed by 13C CP MAS NMR, as the ob-
served chemical shift of the 13C resonance corresponding to TMA
ions (57.4 ppm) was in the range of the signals reported for TMA
occluded in other zeolitic cavities with a size similar to that of
the FER cage [21,22]. According to the chemical analysis results,
the nitrogen content of the samples is approximately 0.12 mol per
100 g of sample, which is equivalent to 3 N atoms per unit cell.
Both TGA and chemical analysis (Table 2) indicate that the organic
content of FER-bmp-TMA samples slightly decreases with the crys-
tallisation time. This observation might be connected with the fact
that ferrierite-related layered materials obtained by different pro-
cedures are formed by the stacking of ferrierite layers [23–27]. If
the layers are not fully condensed in the as-made material, ad-
ditional organic molecules could eventually be located in the in-
terlayer space, leading to an organic content slightly higher than
that expected for fully connected tridimensional crystals, as could
be the case of the FER-bmp-TMA samples crystallised at shorter
times.
Fig. 3. SEM images of the samples FER-bmp-TMA obtained after 20 d of hydrother-
mal treatment (up, left), FER1 obtained after 14 d (up, right), FER-pyrr-TMA obtained
at 10 d (down, left) and FER-pyrr obtained after 10 days (down, right).
Table 2
Chemical analysis and organic content of the samples.
Sample
t
Chemical analysis (wt%) C/Nc TGAa
SEM/EDX
organic (wt%) Si/Al
(d)b
C
H
N
FER-bmp-TMA 10
8.85
9.3
8.79
8.4
1.84
1.83
1.79
1.66
1.68 6.1
1.73 6.3
1.65 6.2
1.62 6.0
13.5
13.7
12.9
11.7
15.1
15.5
16.0
15.6
20
30
83
FER-pyrr-TMA
FER-pyrr
7
10
7.81
7.58
1.85
1.86
2.03 4.5
1.97 4.5
11.8
11.6
15.4
15.9
7
10
7.63
7.35
1.75
1.74
2.03 4.4
2.00 4.3
11.5
11.2
15.2
15.0
FER1
14
5.48
1.38
1.61 4.0
9.6
15.8
a
◦
Weight loss in the range 200–900 C as measured by thermogravimetric analy-
sis.
b
Heating time (days).
Molar ratio.
c
For the samples synthesised with pyrrolidine, the amount of
organic material occluded in the framework significantly varies de-
pending on the synthesis procedure. The FER1 samples, prepared
in the presence of alkaline cations, show an organic content lower
than that found for the FER-pyrr samples, prepared in fluoride
medium with no alkaline cations. Bearing in mind that both sets of
samples present a similar Al content, this result provides an indi-
rect evidence for the incorporation of sodium cations to the inner
volume of the zeolite in the FER1 samples for negative charge com-
pensation, which was confirmed by the ICP-AES results.
For all the samples prepared in fluoride medium, the number
of organic molecules occluded within the zeolite is higher than the
amount of aluminium atoms incorporated to the zeolitic network.
This fact evidences that a fraction of the positive charge associated
to the organic molecules employed as SDAs would be most proba-
bly counter balanced by the incorporation of fluoride anions within
the void volume of the zeolite during the crystallisation process.
+
The sample synthesised in presence of Na at high pH possesses
smaller crystals, with dimensions of 3 × 2.5 μm, while those of
samples FER-pyrr and FER-pyrr-TMA are about 22×13 μm on aver-
age. This increase in the crystal size due to the presence of fluoride
anions in the synthesis gel is commonly reported for zeolite ma-
terials [18,19]. In contrast, ferrierite obtained from gels with bmp
and TMA as SDAs presents needle-like crystals ca. 7 μm long. This
crystal habit has been scarcely observed for ferrierite and remark-
ably differs from the morphology observed for the rest of samples.
The Si/Al ratio of the crystals, measured by SEM/EDX, is within
the range 15–16 for the four sets of samples (Table 2), close to the
one of the synthesis gels. This result corresponds to an aluminium
content of ca. 2.2 atoms per unit cell of ferrierite. ICP-AES results
confirmed the presence of sodium in the FER1 samples (Na/Al mo-
lar ratio of 0.13).
The organic content of the as-made samples for different crys-
tallisation times is collected in Table 2. The C/N atomic ratio, cal-
culated from the elemental analysis, for the samples containing
either TMA and pyrrolidine (FER-pyrr-TMA) or pyrrolidine alone
(FER-pyrr and FER1) as SDAs is close to 4, the theoretical value
3.2. FTIR characterisation of hydroxyl groups
FTIR spectra in the O–H stretching region of the ferrierite sam-
◦
ples outgassed at 400 C are disclosed in Fig. 4. All the samples

262
A.B. Pinar et al. / Journal of Catalysis 263 (2009) 258–265
◦
Fig. 4. Left: FTIR spectra in the νOH region of ferrierite samples after outgassing at 400 C (full lines) and subsequent adsorption of pyridine by contact with the amine vapour
◦
at 150 C for 4 h (dotted lines): (a) FER-pyrr-TMA obtained after hydrothermal treatment for 7 d, (b) FER-pyrr at 10 d, (c) FER-bmp-TMA at 20 d and (d) FER1. Right: Intensity
−1
of the bridging OH band at 3601 cm
(stars).
as a function of pyridine adsorption time, for samples FER-pyrr-TMA (triangles), FER-pyrr (squares), FER-bmp-TMA (circles) and FER1
of ferrierite zeolite in acid form, and tentatively assigned to Al–OH
species [29–31]. However, the exact nature of the hydroxyl species
giving rise to the latter band has not yet been clarified. Recent re-
sults indicate that the acid strength of these hydroxyls is higher
than that expected for extra-framework Al species [32].
◦
After pyridine adsorption at 150 C, the spectra showed char-
acteristic bands of pyridinium ions at 1545 and 1638 cm−1, and
weaker bands due to pyridine interacting with Lewis sites, at 1455
and 1620 cm−1. In addition, the presence of weak bands at 1445
and 1596 cm−1, as well as the disappearance of the terminal
silanols band at 3747 cm−1, pointed out to the presence of pyri-
dine H-bonded to silanols. A decrease in the intensity of the band
−1
located at 3601 cm was also observed, due to the proton transfer
from strong Brönsted sites to pyridine. However, this band was not
completely removed when the samples remained in contact with
◦
pyridine at 150 C for as long as 90 min. Therefore, in order to ver-
Fig. 5. 27Al MAS NMR spectrum of the calcined FER-pyrr-TMA sample obtained after
ify that pyridine adsorption reached equilibrium, FTIR spectra were
recorded at increasing contact times. Fig. 4 (right) shows that the
bridging hydroxyls band intensity levels off with increasing pyri-
dine adsorption time, demonstrating that after pyridine adsorption
for 90 min, equilibrium was reached for all samples except FER-
bmp-TMA. For the later sample, adsorption for as long as ca. 4 h
was required to reach equilibrium. The results evidence that not
10 days of heating.
−1
exhibit an intense asymmetric band at ca. 3601 cm
with a weak
shoulder at 3550 cm−1, characteristic of bridging hydroxyl groups
(Si–OH–Al), and a weaker band at 3747 cm−1, corresponding to
−1
terminal silanols (cf. Ref. [28]). The band at 3601 cm
shows sim-
ilar intensity for all the samples synthesised in fluoride medium,
which indicates that all these samples possess comparable concen-
tration of strong Brönsted sites, in agreement with their similar
Si/Al ratio. The spectra do not show additional hydroxyl bands that
could be attributed to extra-framework species. Indeed, in calcined
sample FER-pyrr-TMA, only tetrahedrally coordinated Al species are
detected by 27Al MAS NMR (Fig. 5), as the spectrum shows a single
resonance at around 52 ppm. The absence of a resonance corre-
sponding to octahedral species suggests that extra-framework Al
species would be present in low concentration and in a highly
distorted environment. In the FTIR spectrum of sample FER1, the
bridging hydroxyls band shows lower intensity, revealing that the
concentration of strong Brönsted sites in this sample is around
25% lower than the amount in the samples synthesised in fluo-
◦
all the acid sites are accessible to pyridine at 150 C. It is notewor-
thy that the samples exhibit apparent differences on accessibility
of the bridging hydroxyls to pyridine. Besides, the intensity of the
−1
band at 3650 cm
observed in the spectra of sample FER1 re-
mained constant, indicating that these hydroxyl species were not
accessible to pyridine. Nevertheless, complete removal of this band
was observed when less bulky basic probe molecules, such as am-
monia and CO, were adsorbed [32].
The location and accessibility of Si(OH)Al groups in ferrierite
has been extensively studied, as this is a key factor which deter-
mines the catalytic properties of this zeolite. Martucci et al. [33],
in a neutron powder diffraction study of deuterated ferrierite car-
ried out at 2 K, identified two Brönsted acid sites, one of them
heading towards the centre of the ferrierite cage, and the second
one heading towards the 10-membered ring channel. The first one
corresponds to a hydroxyl group bridging Si and Al atoms both
ride medium. In addition, the spectrum shows a broad shoulder at
−1
ca. 3650 cm
that has been previously observed in FTIR spectra

A.B. Pinar et al. / Journal of Catalysis 263 (2009) 258–265
263
Table 3
hand, the Al content of these samples corresponds to ca. 2.2 Al
atoms per unit cell. Therefore, up to 91% of the Al could be located
in T4 sites and, hence, have an associated bridging hydroxyl inside
the ferrierite cage. In this case, pyridine would be accessible to at
most 9% of the total acid sites, a situation that would correspond
to the sample synthesised in fluoride medium from pyrrolidine as
the only SDA (FER-pyrr), for which 10% of acid sites interact with
pyridine. This is the sample with the less accessible acid sites.
Replacement of pyrrolidine by other SDA molecules led to the sys-
tematic changes in acid sites accessibility described above, which
can be rationalised as follows. It has been observed that partial
replacement of pyrrolidine by TMA increases accessibility (sample
FER-pyrr-TMA compared to sample FER-pyrr). Preliminary XRD re-
finement results of sample FER-pyrr-TMA together with molecular
dynamics calculations [20] show the preferential location of TMA
in the ferrierite cage. Some of the cages would be filled with this
cation, while others still contain pyrrolidine, with no TMA present
in the main channel. Hence, the partial replacement of pyrrolidine
by TMA seems to “pump” protons from inaccessible cage locations
to those in more open environment accessible from the main chan-
nel.
Accessibility of the Brönsted acid sites to pyridine found in the IR adsorption exper-
iments for the samples synthesised with the different combinations of SDAs.
Sample
% of acid sites accessible to pyridine
FER-pyrr
10
18
36
52
FER-pyrr-TMA
FER-bmp-TMA
FER1
located in T4 positions, and the second one bridging T1 and T3
cations (Fig. 1). Nonetheless, it is important to note that due to
the low temperature at which this study was conducted, only the
most stable location for the acidic hydroxyl can be found. In con-
trast, in an FTIR study of ferrierite samples conducted at room
temperature, Zholobenko et al. [34] proposed up to four frame-
work hydroxyls in pores of different size. These hydroxyls were
assigned to Si(OH)Al groups in the 10-MR channel, in the extended
ring at the intersection of 8- and 6-MR channels, and in 8- and 6-
membered rings. Wichterlová et al. [28] reported on the inability
for pyridine to probe all the Brönsted sites in ferrierite when ad-
◦
sorbed at relatively low temperatures (130 C). This was attributed
When all the pyrrolidine molecules in the cages are replaced
by TMA and those in 10-MR channels by bmp (sample FER-bmp-
TMA compared to sample FER-pyrr-TMA), the average accessibility
of acid sites further increases. In this way, more Al atoms are
driven to positions where the associated bridging OH groups are
more readily available to interact with pyridine.
to the constrained access of pyridine to Brönsted sites located in-
side the ferrierite cavity. Nevertheless, it has been proposed that
pyridine can pass the 8-MR window and that a combination of
high temperature and pyridine partial pressure allows to quantita-
tively probe Brönsted acidity in ferrierite [35].
In our experimental conditions, the ferrierite samples exhibited
marked differences on accessibility of pyridine to acidic hydrox-
yls (Table 3). In sample FER1, 52% of the bridging hydroxyls were
accessible to pyridine. This value is in good agreement with previ-
ous pyridine adsorption measurements carried out on a commer-
cial ferrierite (Si/Al ratio of 8.4), for which 60% of the Brönsted
An additional increase in accessibility is obtained for samples
+
synthesised in the presence of both Na cations and pyrrolidine.
The inorganic cations could be located either in the channel or
in the cage, and therefore it would not be straightforward to ac-
count for the observed effect in terms of cation location. It is
+
◦
worth to note that it has been shown that Na ions are pref-
sites were found accessible to pyridine at 130 C [28]. In compari-
erentially located in the ferrierite cage after partial ion-exchange
son with the reference sample, the zeolites synthesised in fluoride
medium exhibited lower accessibility, being this value strongly
dependent on the combination of SDAs used. Among the later
samples, the one synthesised with bmp and TMA cations (FER-
bmp-TMA) showed the highest accessibility (36%). The number of
Brönsted sites accessible to pyridine decreased to 18% when bmp
cations were replaced by pyrrolidine in the synthesis gel (sample
FER-pyrr-TMA), and it decreased further to 10% when also TMA
cations were replaced by pyrrolidine (sample FER-pyrr). These re-
sults evidence that the distribution of strong Brönsted acid sites
between the ferrierite cages and the more accessible sites exposed
to the 10-MR channels is determined by the actual combination of
SDAs used in the synthesis of ferrierite. It is worth noting that the
accessibility of a given Brönsted acid site is not only determined
by which one is the T site that is occupied by the Al atom, but
also by which one of the four oxygen atoms in its first coordina-
tion shell the proton is bound to. Hydroxyl groups associated to Al
atoms in T4 sites would be non accessible to pyridine, as the four
possible locations for a hydroxyl next to a T4 site are inside the
ferrierite cage (Fig. 1). Furthermore, the zeolites might also con-
tain non accessible Brönsted sites associated to Al atoms located
in T1, T2 and T3 sites, even though they are in the 10-membered
rings, for only 3 of the 4 oxygen atoms surrounding each one of
these sites are also located in the 10-ring channel while the fourth
oxygen atom is inside the ferrierite cavity. Therefore, the presence
of strong Brönsted sites non accessible to pyridine could eventually
be related not only to Al atoms located in T4 sites but also to some
specific locations of the bridging OH species in the framework next
to a given Al atom.
+
+
of NH4 -ferrierite [36]. However, inclusion of Na ions together
with pyrrolidine species into the pores during the zeolite synthesis
+
could lead to a different preferential location of Na ions.
In order to rationalise all these results, it would be possible to
establish a relative trend for Al location on either cage or chan-
nel sites among the several templates, as follows. For cage sites,
pyrrolidine would have higher tendency than TMA to attract Al,
while for channel locations the order would be pyrrolidine > bmp.
In this way, the fraction of Al located at accessible 10-MR channel
sites could eventually be tailored as a function of the particular se-
lected combination of SDAs, for each one seems to possess some
specific influence toward Al sitting in its vicinity.
3.3. Activity of the ferrierite samples in m-xylene isomerisation and
1-butene skeletal isomerisation
The observed differences in acid sites accessibility should have
consequences on the catalytic activity of these ferrierite materi-
als in acid-catalysed reactions. To check this influence, the cat-
alytic transformation of two molecules with different size have
been studied, the isomerisation/disproportionation of the relatively
bulky m-xylene, and the skeletal isomerisation of the compara-
tively smaller 1-butene. The results are collected in Table 4. In
both cases, large differences in the overall reaction rates are ob-
served among the different samples, being the most active the one
+
synthesised at high pH in the presence of Na with pyrrolidine as
SDA (FER1 sample). Interestingly, among the catalysts prepared in
presence of fluoride anions, the one obtained from gels containing
pyrrolidine as the only SDA presents the lowest activity.
In the case of m-xylene reaction, the synthesis method em-
ployed to obtain the ferrierite catalysts has also a significant in-
fluence on the p-xylene/o-xylene ratio. The most active catalyst
The organic content determined for the as-synthesised samples
indicates that every ferrierite cage is occupied by one SDA cation.
It is therefore expected that, at most, one T4 site per cavity (i.e.
two T4 sites per unit cell) can be occupied by Al. On the other

264
A.B. Pinar et al. / Journal of Catalysis 263 (2009) 258–265
Table 4
Activity and selectivity of the different samples in the isomerisation–disproportiona-
tion of m-xylene and in the 1-butene isomerisation.
Sample
m-Xylene isomerisation
n-Butene isomerisation
d
Reaction rateb
(mol/(g h))
P/O
ratio
I/D
ratio
Xc
(%)
S
(%)
SBET
=
(i-C4
)
(m2/g)
FER-pyrra
FER-pyrr-TMA
FER-bmp-TMA
FER1
5.60E−04
6.40E−03
2.20E−02
3.90E−02
n.d.
1.4
1.4
4.6
n.d.
4.3
10.9
46.2
9.6
79.3
79.1
69.2
24.6
350
341
305
244
14.8
24.0
40.2
a
Conversion lower than 0.1%.
Calculated for conversions lower than 5%.
b
c
Conversion at time-on-stream = 30 min.
d
Selectivity to isobutene (time-on-stream = 30 min).
Fig. 7. Activity of the different samples in 1-butene isomerisation (time-on-stream =
30 min) as a function of the accessibility of pyridine to their acid sites in the IR
experiments.
hydroxyls that protonate pyridine after prolonged adsorption. For
the samples synthesised in fluoride medium, this parameter would
be proportional to the number of acid sites per unit mass of cata-
lyst that can be reached by pyridine and protonate this molecule,
as all these zeolites possess similar density of bridging hydroxyls.
For sample FER1, however, it has to be taken into account that the
concentration of bridging hydroxyls is ca. 25% lower. Nevertheless,
due to the significantly higher acid sites accessibility determined
for that sample, it would possess the largest content of acid sites
able to protonate pyridine. It can be clearly observed in Fig. 6
that the activity in the m-xylene isomerisation increases with the
fraction of acid sites that interact with pyridine. However, it is in-
teresting to note that the relationship between catalytic activity
and number of acid sites accessible to pyridine is not linear. The
increase of activity is more prominent as the number of accessible
OH groups increases. This result suggests that a fraction of the hy-
Fig. 6. Activity of the different samples in m-xylene isomerisation as a function of
the accessibility of pyridine to their acid sites in the IR experiments.
(FER1) exhibits the expected p/o ratio for zeolites with one-
directional 10-MR channels [37–39], and a negligible dispropor-
tionation activity, which indicates that the reaction occurs mainly
within the 10-MR channels, with a very small contribution of the
external surface. On the contrary, the samples synthesised in flu-
oride medium show relatively low p/o values, and comparatively
higher disproportionation activity than the one obtained at high
pH. This higher tendency to yield o-xylene and disproportionation
products is an indication of a more open environment and less dif-
fusional restrictions for the reaction in the catalysts synthesised in
fluoride medium. This implies that in the latter catalysts the ratio
between the number of accessible acid sites on the external sur-
face and the number of accessible acid sites in the bulk would be
higher than that of sample FER1. Despite this fact, the samples pre-
pared in fluoride medium present lower activity than FER1, which
suggests that a larger fraction of active centres would not be acces-
sible to the m-xylene molecule. Indeed, these differences in activity
could not be accounted for by differences in surface area. Table 4
shows an opposite trend between surface area and activity, i.e.,
the sample with the lowest activity possesses the highest surface
area. This result, consistent with the obtained p/o values, is an
additional evidence that the relative contribution of sites located
in external surface to the overall activity of the samples prepared
in fluoride medium would be higher than that of the one ob-
tained in alkaline medium, and suggests that the reaction is barely
affected by the reported differences in crystal size among the dif-
ferent catalysts. Hence, the observed differences in activity should
be attributed to differences in the accessibility of the reactant m-
xylene molecules to the acid sites of the ferrierite samples we have
prepared. Fig. 6 depicts the relationship between the activity of the
catalysts and the acid sites accessibility. Strong Brönsted sites ac-
cessibility is given as percentage of the total number of bridging
◦
droxyl groups that are able to interact with pyridine at 150 C are
not available to activate the bulkier m-xylene even at such high
◦
temperature as 350 C, the temperature at which the catalytic test
has been performed.
In contrast to results on m-xylene conversion, when the iso-
merisation of the smaller 1-butene molecule is considered, a nearly
linear relationship between acidity and activity is found for the
samples synthesised in fluoride medium (Fig. 7) in agreement with
its molecular size being smaller than that of m-xylene. These ob-
servations suggest that hydroxyl groups other than those buried
inside the cage and associated to T4 sites might be available for re-
action if appropriate substrate molecules are used. These Brönsted
sites would be associated to the Al atoms located in T1, T2 and T3
positions, which would not be equivalent in terms of their acces-
sibility from the 10-MR channel for bulky molecules. It has been
reported [40–42] that the formation of isobutene on ferrierite cat-
alysts is a function of the number of acid hydroxyl groups located
in the 10-MR channel, which is consistent with the acidity results
reported here. A complete account of the catalytic behaviour of
these samples in 1-butene conversion is reported elsewhere [43].
It is worth to note that the catalyst FER1 do not seem to follow
the linear relationship between activity and density of accessible
bridging hydroxyls found for the samples synthesised in fluoride
medium, taking into account that the estimated density of bridg-
ing hydroxyls in catalyst FER1 is ca. 25% lower. This result might
be explained assuming that at least part of the hydroxyl species
responsible for the band at 3650 cm−1, which are not accessible
to pyridine, could be accessible and able to activate 1-butene.
The analysis of the characterisation and catalytic activity results
discussed above strongly supports the conclusion that the accessi-

A.B. Pinar et al. / Journal of Catalysis 263 (2009) 258–265
265
bility of acid sites to the relatively bulky reagent molecules like
Supplementary material
pyridine, m-xylene and butene in these samples is strongly de-
pendent on the specific combination of structure directing agents
used in the synthesis. It is interesting to observe that the pres-
ence of sodium ions in the gel is sufficient to enormously increase
the acid site accessibility, as can be seen by comparing the results
corresponding to the sample synthesised in fluoride medium using
pyrrolidine as SDA (FER-pyrr) with that obtained at high pH in the
presence of the same SDA (FER1). On the other hand, significant
differences in the accessibility of the acid sites among the catalysts
prepared in fluoride medium have been evidenced. In this case
the negative charge associated to the aluminium atoms can only
be neutralised by the bulky organic SDAs, as there are no sodium
cations present in the synthesis gel. Therefore, these differences
in the accessibility of the acid centres are an indication that the
specific organic SDA employed in each case have a marked effect
on the location pattern of Al atoms associated to these acid sites
in the framework. In other words, we have found evidences that
strongly suggest that the Al location in the ferrierite framework
can be influenced in a decisive manner by the structure directing
agents present in the gel, including sodium among them.
Regarding the effect of the different SDA, the results we present
suggest that this effect would be dependent upon the particular
template molecules used in the synthesis, although we can not yet
offer a clear explanation as to why specific combination of tem-
plate molecules behave so differently. As compared with the small
sodium cation, which would be able to occupy a variety of differ-
ent positions, the bulkiness of the organic SDAs would fix them in
some specific configuration inside of either the main channel or
the cage, making the Al sitting much more specific as well.
The online version of this article contains additional supple-
mentary material.
Please visit DOI: 10.1016/j.jcat.2009.02.017.
References
ˇ
[1] B. Wichterlová, J. Deˇdecˇek, Z. Sobalík, J. Cejka, Stud. Surf. Sci. Catal. 135 (2001)
344.
[2] J. Deˇdecˇek, D. Kaucký, B. Wichterlová, O. Gonsiorová, Phys. Chem. Chem. Phys. 4
(2002) 5406.
ˇ
[3] V. Gábová, J. Deˇdecˇek, J. Cejka, Chem. Commun. (2003) 1196.
ˇ
[4] J. Deˇdecˇek, J. Cejka, M. Oberlinger, S. Ernst, Stud. Surf. Sci. Catal. 142 (2002) 23.
[5] http://iza-structure.org/databases.
[6] P.A. Vaughan, Acta Crystallogr. 21 (1966) 983.
[7] R.E. Morris, S.J. Weigel, N.J. Henson, L.M. Bull, M.T. Janicke, B.F. Chmelka, A.K.
Cheetham, J. Am. Chem. Soc. 116 (1994) 11849.
[8] M.P. Attfield, S.J. Weigel, F. Taulelle, A.K. Cheetham, J. Mater. Chem. 10 (2000)
2109.
[9] Y. Yokomori, J. Wachsmuth, K. Nishi, Microporous Mater. 50 (2001) 137.
[10] A.B. Pinar, L. Gómez-Hortigüela, J. Pérez-Pariente, Chem. Mater. 19 (2007) 5617.
[11] A.B. Pinar, R. García, J. Pérez-Pariente, Collect. Czech. Chem. Commun. 72
(2007) 666.
[12] A.B. Pinar, J. Pérez-Pariente, L. Gómez-Hortigüela, WO2008116958-A1.
[13] D.F. Shantz, R.F. Lobo, C. Fild, H. Koller, Stud. Surf. Sci. Catal. 130 (2000) 845.
[14] K. Frolich, R. Bulánek, P. Nachtigall, in: Proc. of the 40th Symp. on Catalysis,
Prague, 2008, p. 40.
[15] C.J. Plank, E.J. Rosinski, M.K. Rubin, US Patent 4 016 245 (1977).
[16] A. Kuperman, S. Nadimi, S. Oliver, G.A. Ozin, J.M. Garces, M.M. Olken, Na-
ture 365 (1993) 239.
[17] W.J. Smith, J. Dewing, J. Dwyer, J. Chem. Soc. Faraday Trans. 1 85 (1989) 3623.
[18] J.L. Guth, H. Kessler, R. Wey, Stud. Surf. Catal. Sci. 28 (1986) 121.
[19] L.A. Villaescusa, M.A. Camblor, Recent Res. Devel. Chem. 1 (2003) 93.
[20] A.B. Pinar, et al., manuscript in preparation.
[21] W. Fan, S. Shirato, F. Gao, M. Ogura, T. Okubo, Microporous Mesoporous
Mater. 89 (2006) 227.
4. Conclusions
[22] S. Hayashi, K. Suzuki, K. Hayamizu, J. Chem. Soc. Faraday Trans. 1 85 (1989)
2973.
[23] L. Schreyeck, P. Caullet, J.C. Mougenel, J.L. Guth, B. Marler, Microporous Mater. 6
(1996) 259.
[24] T. Ikeda, Y. Akiyama, Y. Oumi, A. Kawai, F. Mizukami, Angew. Chem. Int. Ed. 43
(2004) 4892.
[25] A. Burton, R.J. Accardi, R.F. Lobo, Chem. Mater. 12 (2000) 2936.
[26] D.L. Dorset, G.J. Kennedy, J. Phys. Chem. B 108 (2004) 15216.
[27] R. Millini, L.C. Carluccio, A. Carati, G. Bellussi, C. Perego, G. Cruciani, S. Zanardi,
Microporous Mesoporous Mater. 74 (2004) 59.
[28] B. Wichterlová, Z. Tvaru˚ žková, Z. Sobalík, P. Sarv, Microporous Mesoporous
Mater. 24 (1998) 223.
[29] Y.S. Jin, A. Auroux, J.C. Vedrine, Appl. Catal. 37 (1988) 1.
[30] M. Trombetta, G. Busca, S. Rossini, V. Piccoli, U. Cornaro, A. Guercio, R. Catani,
R.J. Willey, J. Catal. 179 (1998) 581.
[31] V.L. Zholobenko, E.R. House, Catal. Lett. 89 (2003) 35.
[32] B. Onida, et al., manuscript in preparation.
Ferrierite crystals having the same Al content, 2.2 Al/u.c., have
been synthesised from sodium-free gels in fluoride media by using
some specific combinations of structure directing agents, namely
tetramethylammonium (TMA), pyrrolidine and benzylmethylpyrro-
lidinium (bmp). It has been found that the distribution of bridging
hydroxyl groups between the ferrierite cage accessible through 8-
MR windows and the 10-MR channel varies according to the spe-
cific combination of templates used in the synthesis. Considering
also the reference ferrierite sample synthesised in alkaline medium
+
from a gel containing both Na and pyrrolidine, the accessibility
◦
of pyridine to the Brönsted acid sites at 150 C decreases in the
order Na/pyrrolidine > TMA/bmp > TMA/pyrrolidine > pyrrolidine.
A good correlation between acid sites accessibility and catalytic ac-
tivity in the isomerisation of m-xylene and conversion of n-butene
has been found. Hence, this synthesis strategy of ferrierite crystals
allows to gain an effective control on acid sites distribution in the
zeolite framework, and might eventually be extended to other suit-
able zeolite materials by a rational choice of appropriate structure
directing agents.
[33] A. Martucci, A. Alberti, G. Cruciani, P. Radaelli, P. Ciambelli, M. Rapacciulo, Mi-
croporous Mesoporous Mater. 30 (1999) 95.
[34] V.L. Zholobenko, D.B. Lukyanov, J. Dwyer, W.J. Smith, J. Phys. Chem.
(1998) 2715.
B 102
[35] J.A.Z. Pieterse, S. Veefkind-Reyes, K. Seshan, L. Domokos, J.A. Lercher, J. Catal.
187 (1999) 518.
ˇ
[36] P. Sarv, B. Wichterlová, J. Cejka, J. Phys. Chem. B 102 (1998) 1372.
[37] J. Martens, J. Pérez-Pariente, E. Sastre, A. Corma, P.A. Jacobs, Appl. Catal. 45
(1988) 85.
[38] A. Corma, E. Sastre, J. Catal. 129 (1991) 177.
ˇ
[39] G. Mirth, J. Cejka, J.A. Lercher, J. Catal. 139 (1993) 24.
Acknowledgments
ˇ
[40] B. Wichterlová, N. Žilkova, E. Uvanova, J. Cejka, P. Sarv, C. Paganini, J.A. Lercher,
Appl. Catal. A 182 (1999) 297.
[41] L. Domokos, L. Lefferts, K. Seshan, J.A. Lercher, J. Mol. Catal. A Chem. 162 (2000)
147.
[42] J. Cejka, B. Wichterlová, P. Sarv, Appl. Catal. A 179 (1999) 217.
[43] C. Márquez-Alvarez, A.B. Pinar, R. García, M. Grande-Casas, J. Pérez-Pariente,
Top. Catal., in press.
A.B. Pinar acknowledges the Spanish MICINN (former Ministry
of Education) for a PhD grant. The authors thank Dr. T. Blasco for
collecting the NMR spectra. This work has been financially sup-
ported by the MICINN (project CTQ2006-06282).
ˇ