Influence of the Br?nsted and Lewis acid sites on the catalytic activity and selectivity of Fe/MCM-41 system

Article abstract of DOI:10.1016/j.apcata.2012.06.003

The system Fe₂O₃/MCM-41, with high iron dispersion, was synthesized in order to introduce Lewis acid sites into the MCM-41 structure. Two calcination atmospheres (inert and oxidant) were used to produce iron nanoclusters with different structural properties. Besides, a silylation treatment was realized on both solids with the aim of neutralizing the Br?nsted acidity associated with the MCM-41 silanol groups. The samples were characterized by atomic absorption spectroscopy, X-ray diffraction at low angles, N₂ adsorption, temperature-programmed reduction, Fourier transform infrared spectroscopy, nuclear magnetic resonance of ²⁹Si, and M?ssbauer spectroscopy. The influence of the structural changes of the nanoclusters and the effect of the simultaneous presence of Lewis and Br?nsted acid sites were evaluated studying the product distribution from the α-pinene oxide isomerization reaction. It was determined that the Br?nsted acidity of the Fe/MCM-41 system has not the sufficient acid strength to modify the product distribution which is mainly governed by the Lewis sites.

Full text of DOI:10.1016/j.apcata.2012.06.003

Applied Catalysis A: General 435–436 (2012) 187–196
Contents lists available at SciVerse ScienceDirect
Applied Catalysis A: General
journal homepage: www.elsevier.com/locate/apcata
Influence of the Brönsted and Lewis acid sites on the catalytic activity
and selectivity of Fe/MCM-41 system
N.A. Fellenza, J.F. Bengoab,1, S.G. Marchettib,∗, A. Gervasinic,∗∗
a Universidad Nacional de Rio Negro, Sede Atlántica, Belgrano 526, 8500, Viedma, Rio Negro, Argentina
b CINDECA, CONICET, CICPBA, UNLP, Fac. Ciencias Exactas, Calle 47, N◦. 257, 1900, La Plata, Argentina
c Dipartimento di Chimica Fisica ed Elettrochimica (DCFE) and Centro Interdipartimentale Materiali e Interfacce Nanostrutturati (CIMaINa), Università degli Studi di Milano, via
Camillo Golgi, 19, 20133 Milano, Italy
a r t i c l e i n f o
a b s t r a c t
Article history:
The system Fe2O3/MCM-41, with high iron dispersion, was synthesized in order to introduce Lewis acid
sites into the MCM-41 structure. Two calcination atmospheres (inert and oxidant) were used to produce
iron nanoclusters with different structural properties. Besides, a silylation treatment was realized on
both solids with the aim of neutralizing the Brönsted acidity associated with the MCM-41 silanol groups.
The samples were characterized by atomic absorption spectroscopy, X-ray diffraction at low angles,
N2 adsorption, temperature-programmed reduction, Fourier transform infrared spectroscopy, nuclear
magnetic resonance of 29Si, and Mössbauer spectroscopy. The influence of the structural changes of the
nanoclusters and the effect of the simultaneous presence of Lewis and Brönsted acid sites were evaluated
studying the product distribution from the ␣-pinene oxide isomerization reaction. It was determined that
the Brönsted acidity of the Fe/MCM-41 system has not the sufficient acid strength to modify the product
distribution which is mainly governed by the Lewis sites.
Received 26 March 2012
Accepted 2 June 2012
Available online 9 June 2012
Keywords:
POX isomerization
Lewis acid sites
MCM-41
Mössbauer spectroscopy
© 2012 Elsevier B.V. All rights reserved.
1. Introduction
silicas with tuneable pore size could be used as supports for highly
dispersed iron species. In this perspective, the Fe/MCM-41 system
In order to replace the conventional homogeneous acid catalysts
(such as metal halides, HF and H2SO4) in many fine chemical reac-
tions, numerous researches with solid acids had been made in the
last decades. Solid acid catalysts with protonic acidity (Brönsted
acid sites) have been better developed than Lewis acid containing
catalysts [1–3], in particular zeolite materials. Many oxides con-
taining dispersed metal centers on different supports appear good
candidates as Lewis acid catalysts and they have been displayed
interesting catalytic performances in reactions where Lewis acid
centers are concerned [4,5]. Among them, the Fe2O3/SiO2 system,
with the high degree of coordinative unsaturation of the nanosized
Fe2O3 clusters gives excellent catalytic properties in acid reac-
tions, such as Friedel–Crafts alkylations [6,7]. A method to obtain
such samples with the required Lewis acid sites is supporting
“nanoclusters” of iron oxides on silica materials. In order to avoid
the diffusional hindrance of large reactive molecules, mesoporous
was selected in order to produce active and selective catalysts
with Lewis acid sites associated with the Fe presence. However, it
must be considered that the bare surface of MCM-41 has a great
number of silanol groups, which have Brönsted acidity that adds
up to the introduced Lewis acid sites of the Fe species [8].
To investigate on the acid properties and catalytic performances
of the Fe/MCM-41 system, the ␣-pinene oxide (POX) isomerization
reaction was taken into account; information on the Brönsted or
Lewis catalytic action can be inferred from the obtained product
distribution [9]. Moreover, the POX isomerization is an interesting
reaction of fine chemistry, because one of the main products, cam-
pholenic aldehyde (CPA), is a very important intermediate used for
the manufacture of sandalwood fragrances, currently being inves-
tigated together with macrocyclic musks, as possible substitutes
for nitro and polycyclic musks, which are widespread used as fra-
grances in laundry detergents, fabric softeners, cleaning agents
and cosmetic products, and are recognized as damaging chemical
species to the environment.
Taking into account the co-presence of both the two kinds of
acid sites (Brönsted and Lewis type) on the Fe/MCM-41 system,
one aim of this work is to study the product distribution from the
POX isomerization obtained on different Fe/MCM-41 samples with
presence or suppression of the Brönsted acid sites of MCM-41, by
evaluating the acid-strength of the different site-types, too. In order
∗
Corresponding author. Tel.: +54 221 4210711; fax: +54 221 4211353.
Corresponding author. Tel.: +39 02 50314254; fax: +39 02 50314300.
E-mail addresses: nfellenz@quimica.unlp.edu.ar (N.A. Fellenz),
∗∗
bengoajf@quimica.unlp.edu.ar (J.F. Bengoa), march@quimica.unlp.edu.ar
(S.G. Marchetti), antonella.gervasini@unimi.it (A. Gervasini).
1
Tel.: +54 221 4210711; fax +54 221 4211353.
0926-860X/$ – see front matter © 2012 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.apcata.2012.06.003

188
N.A. Fellenz et al. / Applied Catalysis A: General 435–436 (2012) 187–196
to get this purpose, the surface silanol groups were replaced by
trimethylsilyl groups using a silylation treatment. On the other
hand, the possibility of generate different iron Lewis acid sites
in the Fe/MCM-41 system, changing the calcination atmospheres
(inert and oxidant), was studied too. The consequences produced
by these treatments on the sample properties were studied by dif-
ferent characterization techniques and evaluated analyzing their
influence on the selectivity and the activity on POX reaction.
All XRD patterns at low angles were measured using a stan-
dard automated powder X-ray diffraction system Philips PW 1710
with diffracted-beam graphite monochromator using Cu K␣ radia-
◦
tion (ꢀ = 1.5406 A) in the range 2ꢁ = 0.5–9 with steps of 0.02◦ and
˚
counting time of 2 s/step.
The textural properties, specific surface area (Sg), specific pore
volume (Vp) and pore diameter (Dp), were measured in Micromerit-
ics equipment ASAP 2020 V1.02 E.
The TPR profiles were obtained in a Quantachrome equipment,
model Quantasorb Jr. The samples (≈50 mg) were reduced between
300 and 1273 K, at a heating rate of 10 K/min, using a 5% H2/N2
mixture. Taking into account the reducible amount of iron oxide
in the samples, “K” (the sensitivity parameter expressed in s) and
“P” (the shape and resolution parameter, expressed in K) [11] were
around 75 s and 15 K, respectively.
2. Experimental
2.1. Catalyst preparation
MCM-41 support was prepared according to the methodology
proposed by Ryoo and Kim [10] using sodium silicate as silica
source, cetyl-trimethyl-ammonium chloride as surfactant under
controlled pH during the hydrothermal synthesis. For the prepa-
ration of 16 g of sample, 40 g of sodium silicate (26.1% SiO2) was
dissolved into 74 g of water. The solution was then slowly added to
38 ml of cetyl-trimethyl-ammonium chloride and 0.65 ml of NH3
under vigorous stirring at room temperature. This mixture was
heated at 373 K for 24 h in a polypropylene bottle without stir-
ring. Afterwards, it was cooled at room temperature. Then, the pH
value was regulated to approximately 11 by drop wise addition
of acetic acid under vigorous stirring. The reaction mixture was
heated again at 373 K for 24 h. This procedure for pH control and
subsequent heating was repeated twice. The resulting solid was fil-
tered, washed and dried in air at room temperature. The surfactant
was removed by calcination in N2 flow (150 cm3/min) heating the
solid from room temperature to 783 K (8 K/min) maintaining the
final temperature for 1 h. Finally, the N2 flow was replaced by air,
and the solid remained at 783 K for 6 h under air flowing.
The synthesized MCM-41 was impregnated by the incipient
wetness impregnation method with Fe(NO3)3·9H2O aqueous solu-
tion to produce a nominal Fe loading of 8% (w/w) in a single step.
The solid was dried in air at room temperature, and then was
split in two fractions; the first fraction was calcined in a flow of
N2 (60 cm3/min) from room temperature to 603 K (0.2 K/min), and
kept at the final temperature for 1 h, so obtaining the Fe/MCM-
41(N2) sample. The second fraction was calcined under the same
conditions above described, replacing the nitrogen flow by air, so
obtaining the Fe/MCM-41(air) sample.
The FT-IR absorption spectra were acquired with a Bruker IFS66
spectrometer with 1 cm−1 resolution by co-addition of 32 scans.
The samples were prepared mixing with potassium bromide, in a
1:100 proportion, in order to obtain the corresponding pellets.
Solid state 29Si nuclear magnetic resonance spectra by direct
polarization (DP) and cross-polarization (CP) 1
H
29Si with proton
decoupling and magic angle spinning (MAS) were recorded on a
Bruker Avance II-300 spectrometer equipped with a 4-mm MAS
probe operating at 300.13 MHz for protons and 59.6 MHz for 29Si.
Magic angle spinning rate was 5 kHz. Other experimental condi-
tions were: cw proton decoupling at 62.5 kHz, contact time for CP
5 ms. The chemical shifts are given in ppm from H3PO4 (0 ppm) as
standard reference.
The Mössbauer spectra were obtained in transmission geometry
with a 512-channel constant acceleration spectrometer. A source
of 57Co in Rh matrix of nominally 50 mCi was used. Velocity cali-
bration was performed against a 12 ␮m-thick ␣-Fe foil. All isomer
shifts (ı) mentioned in this paper are referred to this standard. The
temperature was varied between 25 and 298 K using a Displex DE-
202 Closed Cycle Cryogenic System. The Mössbauer spectra were
evaluated using a commercial program with constraints named
Recoil [12]. Although some spectra display magnetic relaxation, for
simplicity, Lorentzian lines with equal widths were considered for
each spectrum component. The spectra were folded to minimize
geometric effects.
2.3. Activity and selectivity measurements
A fraction of both solids was silylated with hexamethyldisi-
lazane (HMDS). In order to get this treatment the solids were
outgassed (pv < 10−3 Torr) during 3 h at 573 K. Then, a solution of 1%
(v/v) of HMDS in toluene was prepared in a glove box, in Ar atmo-
sphere, and it was added to the dehydrated solids. The mixture was
heated at 393 K, during 90 min under stirring. Finally, the treated
solids were filtered, washed with 80 cm3 of toluene, and dried in an
oven at 333 K during 16 h. The obtained solids were called Fe/MCM-
41(N2)-sil and Fe/MCM-41(air)-sil taking into account the nature
of the calcination atmospheres.
The epoxide isomerization tests using POX as substrate were
carried out at room temperature (RT) in batch conditions as detailed
in Refs. [9,13]. The catalyst sample (0.1 g) was activated into
the glass reactor at 623 K for 30 min in air and then for 30 min
under reduced pressure at the same temperature. After catalyst
activation, ␣-pinene oxide (0.66 mmol) and toluene (8 ml) were
introduced into the reactor under N2 atmosphere. The progress
of the reaction was followed by gas-chromatographic techniques
(GC from Agilent 6890 with FID detector, with a 5% phenyl-
methylpolysiloxane column, and GC–MS from Agilent 5971 series),
analyzing samples withdrawn from the reaction mixture at differ-
ent times.
2.2. Catalyst characterization
The samples were characterized by atomic absorption spec-
troscopy (AAS), X-ray diffraction (XRD) at low angles, N2 adsorption
at 77 K (BET), temperature-programmed reduction (TPR), Fourier
transform infrared spectroscopy (FT-IR), nuclear magnetic reso-
nance (NMR) of 29Si, and Mössbauer spectroscopy (MS) at 298 and
25 K.
The Fe content of the solids was determined by atomic
absorption on an AA/AE Spectrophotometer 457 of Laboratory
Instrumentation Inc. The sample was treated in a mixture of HCl
and HF up to complete dissolution before measurement.
3. Results and discussion
3.1. Catalyst characterization and silylation effect
The ordered hexagonal structure of mesoporous MCM-41, used
as support, was verified by XRD (Fig. 1) and its textural prop-
erties were measured by N2 adsorption at 77 K (Table 1). The
impregnation, calcination in different atmospheres, and silylation
treatments, which lead to Fe/MCM-41(N2), Fe/MCM-41(N2)-sil,

N.A. Fellenz et al. / Applied Catalysis A: General 435–436 (2012) 187–196
189
Table 1
Textural properties and iron loading of the studied catalyst samples.
Samples
Sg (m2/g)
Dp(nm)
Vp (cm3/g)
CBET
%Fe (w/w)
MCM-41
912
691
494
608
476
2.7
2.9
2.4
2.6
2.6
0.88
0.64
0.59
0.50
0.33
107
108
45
92
62
–
Fe/MCM-41(N2)
Fe/MCM-41(N2)-sil
Fe/MCM-41(air)
Fe/MCM-41(air)-sil
8.9
8.6
8.4
8.4
Sg, specific surface area; Dp, average pore diameter; Vp, specific pore volume; CBET, BET model constant.
distinguished: the 758 cm−1 band is assignable to the stretching of
Si C bond and the 849 cm−1 band to rocking of CH3 groups [15].
According to Lin et al. [19] these results indicate that trimethylsi-
lyl groups of the silylant agent were anchored on the mesoporous
solid surface. Therefore, these results point to the same direction
that the analysis on the CBET value, indicating that the surface sily-
lation process, used in this work, was successful. Similar results
were obtained with Fe/MCM-41(air) and Fe/MCM-41(air)-sil (not
shown).
Fe/MCM-41(air), and Fe/MCM-41(air)-sil, did not change the struc-
tural properties of the mesoporous support, as was verified by XRD
(Fig. 1).
By analyzing the textural properties of Fe/MCM-41(N2), and
Fe/MCM-41(air), a decrease of Sg and Vp in comparison with
the support, without substantial changes in Dp, can be observed
(Table 1). This would imply a partial pore filling of the support with
the Fe oxide species. After the silylation treatment these results
were more pronounced and they were attributed to the covering of
the internal pore walls with trimethylsilyl groups. Similar results
were reported by other authors [14,15].
The Fe loading, obtained by AAS, can be considered similar over
all the catalysts within acceptable experimental errors (Table 1), so
strengthening the possibility of comparison among all samples.
The MCM-41 mesoporous solid has a great number of silanol
groups (Si OH) with pronounced hydrophilic character which are
distributed on the inner surface of the channels. The deposition
of Fe on MCM-41 should change the nature of the acid surface
turning towards surfaces with prevalent Lewis acidity [9,13,16].
However, Brönsted acidity remains on the Fe-samples due to some
silanol group presence. A silylation treatment was realized to
decrease the number of Brönsted acid sites, and so changing the
Lewis/Brönsted acid sites ratio. Besides, this treatment would mod-
ify the hydrophilic/hydrophobic nature of these catalysts. In order
to study the influence of both effects on the catalytic performance,
a reaction test, in which acidic sites are concerned, was selected:
isomerization of ␣-pinene oxide. This reaction allows studying the
prevalent presence of Lewis or Brönsted acid sites, through the
influence of the acid-site nature on the activity and particularly
on selectivity of the formed products.
More information about the efficiency of the silylation treat-
ment was obtained from the 29Si NMR spectra. The measurement of
the NMR spectra of Fe/MCM-41(N2), Fe/MCM-41(N2)-sil, Fe/MCM-
41(air) and Fe/MCM-41(air)-sil was prevented due to the magnetic
behavior of the samples; they were all weak-ferromagnetic. There-
fore, the 29Si NMR spectra were obtained with a sample of MCM-41
without Fe, before and after silylation treatment (MCM-41 and
MCM-41-sil). The presence of Fe3+ centers, acting as Lewis acid
sites, may become more acidic the neighbor silanol groups [13].
Therefore we consider that, if the silylation treatment is effective
without iron, a greater efficiency would occur when it is present.
The NMR spectra of MCM-41 and MCM-41-sil obtained by DP are
shown in Fig. 3. For both samples, the most intense peak appears
at −110 ppm, and it is associated with silicon atoms bonded to four
siloxane groups, such as (SiO4)Si. A shoulder at about −101 ppm can
be also observed, which is assigned to the presence of silanol groups
bonded to a surface silicon atom (SiO3)Si OH [17]. Besides, a peak
at about 14 ppm is only observed for MCM-41-sil. This signal can be
assigned to (CH3)3 Si species [15,20]. In Fig. 4 the NMR spectra by
CP 1H 29Si with proton decoupling and MAS are reported. In these
experiments, the signals of 29Si atoms surrounded by protons are
reinforced with respect to other Si atoms. This effect can be clearly
observed when Figs. 3 and 4 are compared. Therefore, the assign-
ments for the peaks at −101 and 14 ppm, made by DP, are right,
confirming that MCM-41-sil was silylated.
In order to get the surface hydrophobization and nullify the
silanol group action, Fe/MCM-41(N2) and Fe/MCM-41(air) were
contacted with HMDS to produce trimethylsilyl groups according
to the following scheme [17]:
Therefore, N2 adsorption results, FT-IR and NMR spectra allow
us to conclude that the silylation treatment was successfully and
hydrophobic surfaces were attained. However, it is important to
remark that this process is not totally effective since some silanols
groups remain on the surface as it is evidenced in 29Si RMN spec-
trum by the shoulder at −101 ppm.
2
Si OH + (CH3)3 Si NH Si (CH3)3
→ 2 Si Si (CH3)3 + NH3
In order to verify the silylation treatment efficiency, the CBET
O
value obtained from N2 adsorption measurements at 77 K, was
analyzed. Taking into account that the magnitude of CBET is an indi-
cation of the adsorption enthalpy between the adsorbate (N2) and
the solid surface, a smaller CBET value would indicate a weaker
interaction between N2 molecules and the solid surface [18]. By
comparing Fe/MCM-41(N2) and Fe/MCM-41(N2)-sil catalysts, a
decreasing of 2.4 times of the CBET value can be observed as a con-
sequence of the silylation treatment (Table 1) which is indicative of
the presence of a hydrophobic surface. In the same way, Fe/MCM-
41(air)-sil showed a decreasing of 1.5 times of its CBET value with
respect to Fe/MCM-41(air).
In order to characterize the iron species present at the cata-
lysts, Mössbauer spectra at different temperatures were collected.
The Mössbauer spectrum of Fe/MCM-41(N2) at 298 K shows a
very intense central doublet and a small sextuplet (Fig. 5). The
hyperfine parameters of the sextuplet correspond to ␣-Fe2O3 [21]
(Table 2). The hyperfine magnetic field (H) is lower than that cor-
responding to the “bulk”, probably due to the small size of the
␣-Fe2O3 crystals. This decrease in the H value could be attributed
to the magnetic collective excitations phenomenon [22]. Applying
this model and using an effective magnetic anisotropy constant
KCME = 3.5 × 103 J/m3 [23], a rough estimation of the average diam-
eter value of crystal oxide of about 35 nm for the magnetic fraction
was obtained. Taking into account that the average pore diameter
of the support is about 3 nm (Table 1) is evident that this fraction of
The H+ substitution of the silanol groups by trimethylsilyl
groups can be monitored by FT-IR [15,19]. With this purpose, the IR
spectra of Fe/MCM-41(N2) and Fe/MCM-41(N2)-sil were collected
(Fig. 2). In Fe/MCM-41(N2)-sil spectrum, two new bands can be

190
N.A. Fellenz et al. / Applied Catalysis A: General 435–436 (2012) 187–196
2θ= 2.1º
-1
758 cm
-1
849 cm
MCM-41
Fe/MCM-41(N2)
Fe/MCM-41(N2)-sil
2θ=2.2º
400 600 800 1000 1200 1400 1600 1800 2000
Wavenumber (cm-1)
Fig. 2. FT-IR spectra of Fe/MCM-41(N2) before and after silylation treatment.
Fe/MCM-41(N2)
Direct polarization
-110 ppm
MCM-41(N2)
2θ=2.2º
-101 ppm
Fe/MCM-41(N2)-sil
-110 ppm
2θ=2.2º
-101 ppm
MCM-41(N2)-sil
Fe/MCM-41(air)
14 ppm
2
4
6
8
2θ ( )
º
Fig. 1. XRD patterns at low angles of the support and the catalysts.
50
0
-50
-100
-150
-200
-250
δ (ppm)
Fig. 3. NMR spectra of MCM-41 and MCM-41-sil obtained by direct polarization.

N.A. Fellenz et al. / Applied Catalysis A: General 435–436 (2012) 187–196
191
Cross polarization
-101 ppm
MCM-41(N2)
Fe/MCM-41(N2)
T=298 K
Fe/MCM-41(N2)-sil
T=298 K
-110 ppm
-101 ppm
Fe/MCM-41(N2)
T=25K
Fe/MCM-41(N2)-sil
T=25 K
MCM-41(N2)-sil
-10
-5
0
5
10
-10
-5
0
5
10
Velocity (mm/s)
-110 ppm
Fig. 5. Mössbauer spectra of Fe/MCM-41(N2) and Fe/MCM-41(N2)-sil at 298 and
25 K.
14 ppm
relaxing ␣-Fe2O3 in Table 2). The evolution of these two broad sex-
tets when the temperature decreased would indicate that at lower
temperature the magnetic blocking process would be completed;
and the hyperfine parameters would corresponds to ␣-Fe2O3 with
extremely small sizes. Anyway, the presence of amorphous Fe2O3
“nanoclusters” cannot be excluded since the hyperfine magnetic
fields are much decreased. This complex magnetic behavior pre-
vents to determine the right value of the blocking temperature (TB)
of the sample (defined as the temperature at which 50% of the sex-
tet components have collapsed to doublets). However, a value of
about 25 K for TB can be estimated.
By applying the Neel–Brown relaxation model [26] and using
a basal magnetic anisotropy constant KBU = 2.4 × 104 J/m3, deter-
mined by Bødker and Mørup [23] for hematite particles of 5.9 nm
size, a rough estimation of the average diameter value of crys-
tal oxide of about 3 nm for the fraction of ␣-Fe2O3 (sp) at 298 K,
was obtained. By considering this result jointly with the tex-
tural measurements, we can conclude that the iron species are
predominantly located inside the MCM-41 channels. Finally, the
small doublet (5 2%) that remains at this low temperature can
be assigned to paramagnetic Fe3+ ions diffused into the MCM-41
walls. Therefore, in Fe/MCM-41(N2) there should be two fractions
of ␣-Fe2O3 with very different sizes: a small percentage, 8 1% of
crystals of about 35 nm located outside of the MCM-41 channels
structure and the great majority as extremely small crystals that
even at 25 K do not get blocked magnetically, located inside the
MCM-41 channels.
50
0
-50
-100
δ (ppm)
-150
-200
-250
Fig. 4. NMR spectra of MCM-41 and MCM-41-sil obtained by cross polarization.
the iron species is located outside the channels of the MCM-41. The
doublet could be assigned to superparamagnetic ␣-Fe2O3 particles
(␣-Fe2O3 (sp)), smaller than those that produce the sextuplet, or to
paramagnetic Fe3+ ions diffused into the support walls.
To select the right option, the Mössbauer spectrum at 25 K was
obtained (Fig. 5). The spectrum displayed a broad central line and
three peaks at both sides, two of them are broad and the other
narrow. Comparing this spectrum with that obtained at 298 K is
obvious that the superparamagnetic relaxation phenomenon is
present; the background is curved, there are very broad peaks
and the central signal had decreased but it had not disappeared.
Therefore, this complex behavior was fitted with three sextets and
one doublet. Two sextuplets are relaxing and they were simu-
lated using the relaxation model of two states of Blume and Tjon
[24] and the other one (the narrower) is non-relaxing. This non-
relaxing sextet has parameters of ␣-Fe2O3 which has undergone
Morin transition, therefore, their diameters are larger than ≈20 nm
[25] (called ␣-Fe2O3 in Table 2). The second sextet has not com-
pleted the magnetic blocked state, yet (called partially magnetically
blocked ␣-Fe2O3 in Table 2). The third sextet is in relaxation pro-
cess with a fluctuation speed higher than the previous one (called
In order to verify that the silylation treatment did no affect the
nature and the size of the iron species present in the catalyst, the
Mössbauer spectra of Fe/MCM-41(N2)-sil at 298 and 25 K were
obtained (Fig. 5). It can be seen that the fitting and the hyperfine

192
N.A. Fellenz et al. / Applied Catalysis A: General 435–436 (2012) 187–196
Fe/MCM-41(air)
Fe/MCM-41(air)
T=298 K
Fe/MCM-41(N2)
200
400
600
800
1000
1200
1400
T (K)
Fig. 7. TPR profiles of Fe/MCM-41(N2) and Fe/MCM-41(air).
fitting process. Both sextets were assigned, in the same way that
in Fe/MCM-41(N2), to ␣-Fe2O3 with different magnetic relaxation
speeds. For the same reasons that in this solid the presence of
amorphous Fe2O3 “nanoclusters” cannot be discarded. Taking
into account that at 25 K, the magnetic blocking is lower than in
Fe/MCM-41(N2), we can conclude that the iron oxide particles are
lower than 3 nm. The doublet was assigned to paramagnetic Fe3+
ions diffused into the MCM-41 walls. Its percentage was nearly
equal to that obtained with Fe/MCM-41(N2): 5 2% for Fe/MCM-
41(N2) vs. 7 3% for Fe/MCM-41(air). Therefore, in Fe/MCM-41(air)
were detected the same species that in Fe/MCM-41(N2), but all iron
oxide crystals are located inside of the MCM-41 pores and its aver-
age diameter is smaller. Taking into account that the impregnation
step was common in both systems, we can conclude that during the
calcination step in N2 stream, a fraction of the iron crystals migrate
outside of the MCM-41 pores and growth up by sintering process. In
a previous work [27] a similar result was obtained by studying the
effect of the calcination atmospheres on the structural properties
of Fe/amorphous SiO2 system. These results were explained taking
into account that the solid–solid wetting phenomenon can be dras-
tically modified when the composition of the gaseous environment
is changed. Thus, it has been established that the condition by
spreading of the iron oxide on the support is more easily satisfied
in oxidant atmosphere [28]. Therefore, in Fe/MCM-41(air) the
spreading conditions could be satisfied more easily, leading to the
production of islands like as “raft” of iron oxide. These “rafts” would
strongly interact with the inner walls of the MCM-41 channels,
and their migration at the outside surface would be hindered.
In order to confirm these different interaction strengths
between the iron oxides species and the MCM-41 support –
depending on the used calcination atmosphere- the TPR assays with
Fe/MCM-41(N2) and Fe/MCM-41(air) were obtained (Fig. 7). The
Fe/MCM-41(N2) TPR profile was fitted with three Gaussian peaks
Fe/MCM-41(air)
T=25K
-10
-5
0
5
10
Velocity (mm/s)
Fig. 6. Mössbauer spectra of Fe/MCM-41(air) at 298 and 25 K.
parameters obtained at both temperatures (Fig. 5 and Table 2) are
identical – within the experimental errors – to that obtained before
the silylation treatment.
The Mössbauer spectra of Fe/MCM-41(air) at 298 and 25 K
are shown in Fig. 6. At 298 K only a doublet had been detected
assignable to ␣-Fe2O3 (sp) or to paramagnetic Fe3+ ions diffused
into the support walls. It is important to remark that the magnetic
signal, corresponding to ␣-Fe2O3 located outside the MCM-41
channels, was not detected. Again, the spectrum obtained at 25 K
showed a superparamagnetic relaxation with a partial magnetic
blocking. Two relaxing sextuplets and a doublet were used in the

N.A. Fellenz et al. / Applied Catalysis A: General 435–436 (2012) 187–196
193
Table 2
Mössbauer parameters of the catalyst samples measured at 298 K and 25 K.
Temperature
298 K
Species
Parameters
Fe/MCM-41 (N2)
Fe/MCM-41 (N2)-syl
Fe/MCM-41 (air)
␣-Fe2O3
H (T)
51.0 0.2
0.37 0.03
−0.24 0.06
50.5 0.2
0.37 0.03
−0.26 0.06
–
–
–
–
ı (mm/s)
2ε (mm/s)
%
15
1
11
1
ꢂ (mm/s)
ı (mm/s)
%
0.73 0.01
0.35 0.01
0.75 0.01
0.35 0.01
0.70 0.01
0.35 0.01
100
␣-Fe2O3 (sp) and/or
paramagnetic Fe3+
85
1
89
1
25 K
␣-Fe2O3
H (T)
54.1 0.2
0.47 0.02
0.37 0.05
54.0 0.2
0.47 0.04
0.33 0.07
–
–
–
–
ı (mm/s)
2ε (mm/s)
%
8
1
5
1
H (T)
45.8 0.4
0.38 0.04
0*
46.1 0.3
0.50 0.03
0*
45.2 0.8
Partially magnetically
blocked ␣-Fe2O3
ı (mm/s)
2ε (mm/s)
%
0.45 0.07
0*
66
3
66
3
57
6
H (T)
41.5*
41.5*
41.5*
Relaxing ␣-Fe2O3
ı (mm/s)
2ε (mm/s)
%
ꢂ (mm/s)
ı (mm/s)
%
0.48 0.06
0.48 0.04
0.47 0.03
0*
0*
0*
21
3
22
2
37
7
5
Paramagnetic Fe3+
diffused into MCM-41
walls
0.59 0.07
0.37 0.04
0.59 0.06
0.37 0.04
0.7 0.1
0.45 0.04
3
5
2
7
1
H: hyperfine magnetic field in Tesla; ı: isomer shift (all the isomer shifts are referred to ␣-Fe at 298 K); 2ε: quadrupole shift; ꢂ: quadrupole splitting.
*
Parameters held fixed in fitting.
with maxima at 673, 830 and 986 K. They could be assigned to the
following reduction steps:
the catalyst surface [9]. Acid surfaces with presence of Brönsted acid
sites lead to some other products, like PCP, TCV, and dimerization
products, decreasing the CPA selectivity. In the present catalysts,
Brönsted acid sites could be generated from silanol groups, espe-
cially if they are located near to the Fe3+ centers [13]. Since the
silylation treatment is expected to decrease the number of silanol
groups, the silylated Fe/MCM-41 samples may turn the product
distribution towards species formed by the Lewis acid site trans-
formation of POX.
The results obtained in the ␣-pinene oxide transformation over
the four Fe/MCM-41samples are presented in Tables 3 and 4.
Figs. 9 and 10 display the most significant results in terms of POX
conversion and CPA selectivity. Fe/MCM-41(N2) catalyst showed a
fast POX conversion during the first 10 min and the reaction pro-
ceeded up to complete POX conversion in about 30 min. Selectivity
to CPA started at ca. 70% and it little decreased down to 60% at total
POX conversion. It is important to remark that the TCV produc-
tion is nearly negligible; therefore, the silanol groups did not have
Brönsted acidity strong enough to produce this by-product. The
silylation treatment (Fe/MCM-41(N2)-sil sample) did not remark-
ably show different results in terms of product distribution. Total
POX conversion was achieved at longer time (60 min of reaction) in
comparison with Fe/MCM-41(N2). We can conclude that the more
hydrophobic nature of the surface – after the silylation treatment
– makes more difficult the arrival of the polar reactive molecules
up to the Fe3+ sites.
␣-Fe2O3/amorphousFe2O3 → Fe3O4 → FeO → ␣-Fe
Beyond 1250 K, the profile showed a small increase that could
be attributed to the beginning of the Fe2+ ions reduction diffused
inside the SiO2 walls. The Fe/MCM-41(air) TPR profile showed
important differences with respect to Fe/MCM-41(N2). Thus, only
the two first peaks were detected at 712 and 826 K. The third
peak assigned to the reduction of FeO to ␣-Fe did not appear.
Instead, a very intense H2 consumption was detected at 1283 K.
Taking into account the high temperature of this peak, it could be
attributed to the reduction of Fe2+ ions diffused inside the MCM-
41 walls. A great quantity of this species would be produced after
the FeO formation due to a very important interaction with the
support. Therefore, the Fe2+ diffusion into the SiO2 lattice com-
petes with the final reduction. As consequence the FeO → ␣-Fe
step at about 900–1000 K did not occur. This different reduction
behavior would confirm the distinct interaction degree between
iron oxide species and the support, and would justify the differ-
ent crystal sizes obtained when the calcination atmospheres were
changed.
3.2. Catalytic test of ˛-pinene oxide isomerization
The catalytic test results over Fe/MCM-41(air) and Fe/MCM-
41(air)-sil were very different from the same materials calcined in
inert atmosphere. They showed very lower activity in terms of POX
conversion but higher selectivity to CPA (Table 4 and Figs. 8 and 9).
Reaction times as long as 24 h were necessary to obtain a complete
POX conversion. The low initial POX conversion (around 10%) was
accompanied by a very high initial selectivity to CPA (more than
90%) on both Fe/MCM-41(air) and Fe/MCM-41(air)-sil. For longer
times of reaction POX conversion increased and selectivity to CPA
little decreased; it was around 70% after 250 min of reaction. A cer-
tain degree of deactivation of the most active Lewis sites could
be inferred to explain the observed fall of selectivity. The effect
of the silylation treatment on Fe/MCM-41(air) was very similar to
The isomerization of POX can be viewed as a test for under-
standing the action of the Brönsted or Lewis acid sites and their
relative ratio, besides its recognized importance in specialty chem-
istry [29–32]. In this work, it has been used as model reaction able to
enlighten the presence of active catalytic Lewis sites from the prod-
uct distribution [9,13]. The acid POX rearrangement can give several
products since this molecule has a high reactivity (Fig. 8) [29]. Thus
over acid catalysts, CPA, transcarveol (TCV), trans-sobrerol (TSB),
pinocamphone (PCP), and paracymene (CIM) can be observed as
the principal reaction products. Among them, CPA is the most inter-
esting product since is a very important intermediate used for the
manufacture of fragrances. High CPA selectivities from acid POX
isomerization can be obtained only if Lewis acid sites are working at

194
N.A. Fellenz et al. / Applied Catalysis A: General 435–436 (2012) 187–196
O
HO
P
P
iinenee oxide
OOX
O
Caamphholeniic alddeehydee
CPA
4-Isopropenyl-1-methyl-2-cyclohexen-1-ol
HXN
OH
O
Is
o
-
c
m
m
p
p
h
h
o
le
yd
e
ALD
Cis-caren-4-ol
OH
OH
CRRN
O
Pinocarveolo
PCV
Pino
ccanphone
P
Cis-Carveol
Trans-carveol
cis/trans-CARV
Fig. 8. Scheme of the ␣-pinene oxide isomerization reaction over Fe/MCM-41 catalysts.
that already observed over Fe/MCM-41(N2) and Fe/MCM-41(N2)-
sil samples.
Therefore, the observed different catalytic behavior would be
assigned to more subtle structural differences. We can speculate
that calcination in oxidant conditions helps the spreading of the
iron species on the inner walls of the MCM-41 (more oxidant
atmosphere) and favor the oxide/support interaction (as it was
demonstrated by TPR); only one type of geometry of Fe3+ site is
prevalently formed. These iron sites have high coordinative unsat-
uration and can act as acid Lewis sites in the catalytic reaction. All
Taking into account the results obtained from the character-
ization study, we know that the iron species were the same in
both sample series (calcined in inert and oxidant atmospheres),
the most part of Fe-species are extremely small nanoparticles
located inside the channels of the MCM-41 support (the popula-
tion of large ␣-Fe2O3 particles in Fe/MCM-41(N2) is very small).
Table 3
Catalytic performances of the Fe/MCM-41(N2) and Fe/MCM-41(N2)-sil catalysts in the reaction of isomerization of ␣-pinene oxide in terms of POX conversion and selectivity
to all the products.
Catalyst code
Sample
mass (g)
Reaction
Time (min)
Conversion (%)
POX
Selectivity (%)
CPA
ALD
PCV
PCP
cis-CARV
trans-CARV
CRN
HXN
Other
Fe/MCM-41(N2)
0.1039
1
3
5
10
20
30
48.6
64.1
67.9
85.0
93.4
97.6
67.5
66.2
71.3
60.8
59.5
59.3
7.5
7.4
8.2
7.3
7.3
7.3
2.8
2.9
3.2
2.8
2.8
2.8
0.0
0.0
0.0
1.3
1.4
1.4
0.0
0.0
0.0
2.0
2.1
2.0
0.0
0.0
0.0
2.0
2.1
2.2
9.1
8.9
9.1
9.1
9.6
9.7
7.1
7.4
0.0
7.7
8.2
8.3
5.1
6.4
7.0
6.2
6.2
6.2
Fe/MCM-41(N2)-sil
0.1042
1
3
5
10
20
30
60
30.5
49.1
59.6
70.5
76.8
84.5
95.5
75.9
68.6
66.6
64.0
66.9
66.6
59.9
0.0
7.6
7.9
7.7
8.2
8.2
7.8
3.9
3.7
3.6
3.4
3.4
3.4
3.1
0.0
0.0
0.0
0.0
0.0
0.0
1.3
0.0
0.0
0.0
1.6
0.0
0.0
2.0
0.0
0.0
0.0
0.0
0.0
0.0
2.1
7.5
7.4
8.3
8.6
8.0
8.3
9.2
5.2
5.7
6.3
6.4
6.1
6.1
7.0
7.0
6.3
6.4
6.4
6.4
6.4
6.3
POX: ␣-pinene oxide; CPA: campholenic aldehyde; ALD: iso-campholenic aldehyde; PCV: pinocarveol; PCP: pinocamphone; cis-CARV: cis-carveol; trans-CARV: trans-carveol;
CRN: cis-caren-4-ol; HXN: 4-isopropenyl-1-methyl-2-cyclohexen-1-ol.

N.A. Fellenz et al. / Applied Catalysis A: General 435–436 (2012) 187–196
195
Table 4
Catalytic performances of the Fe/MCM-41(air) and Fe/MCM-41(air)-sil catalysts in the reaction of isomerization of ␣-pinene oxide in terms of POX conversion and selectivity
to all the products.
Catalyst code
Sample
mass (g)
Reaction
Time (min)
Conversion (%)
POX
Selectivity (%)
CPA
ALD
PCV
PCP
cis-CARV
trans-CARV
CRN
HXN
Other
Fe/MCM-41(air)
0.1031
1
3
5
10
20
30
60
120
240
1260
8.9
11.0
11.6
15.2
17.9
24.1
23.6
31.2
42.6
70.9
94.6
91.9
93.8
90.4
87.0
76.9
83.8
76.0
71.0
67.4
0.0
0.0
0.0
0.0
0.0
9.2
0.0
8.0
7.6
8.1
8.1
9.5
7.8
9.1
8.9
8.6
7.8
7.0
8.0
8.8
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
1.7
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
3.4
4.2
6.1
8.2
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
2.6
0.0
0.0
0.0
0.0
0.0
3.6
4.4
4.0
3.8
3.2
4.0
Fe/MCM-41(air)-sil
0.1022
1
3
5
10
20
30
60
120
240
1440
8.0
11.2
14.7
18.4
19.2
23.5
30.2
44.8
55.8
100.0
93.6
92.3
88.8
82.5
89.3
85.5
78.1
68.9
66.5
63.3
0.0
0.0
0.0
0.0
0.0
0.0
0.0
10.0
10.8
10.5
8.6
8.8
9.5
9.3
8.8
9.7
8.4
7.8
7.0
7.6
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
1.1
1.3
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
1.5
0.0
0.0
0.0
0.0
0.0
0.0
6.0
6.2
6.9
7.3
0.0
0.0
0.0
0.0
0.0
0.0
2.4
2.4
3.0
3.3
0.0
0.0
0.0
2.7
0.0
2.8
2.9
2.3
2.4
3.2
POX: ␣-pinene oxide; CPA: campholenic aldehyde; ALD: iso-campholenic aldehyde; PCV: pinocarveol; PCP: pinocamphone; cis-CARV: cis-carveol; trans-CARV: trans-carveol;
CRN: cis-caren-4-ol; HXN: 4-isopropenyl-1-methyl-2-cyclohexen-1-ol.
100
90
80
70
60
50
100
80
60
40
20
0
of them would have the same geometry and would favor the CPA
production.
Taking into account that, when N2 stream was used during cal-
cinations, a higher magnetic blocking temperature and a lower
specific surface area were obtained, it could be inferred that, clus-
tering of the iron species occurs inside the MCM-41 channels. As a
consequence, different crystalline planes, corners, steps, could be
exposed and Fe3+ sites with appropriate geometry to be more active
in adsorbing reactant would appear.
Fe/MCM-41(air)
100
0
50
150
200
250
100
80
60
40
20
0
Time (min)
80
70
60
50
100
90
80
70
60
50
100
80
60
40
20
0
Fe/MCM-41(N2)
Fe/MCM-41(air)-sil
0
10
20
30
Time (min)
0
50
100
150
200
250
100
80
60
40
20
0
80
70
60
50
Time (min)
Fig. 10. Selectivity to ␣-campholenic aldehyde formed in the isomerization of ␣-
pinene oxide (left axis, filled marker) and POX conversion (right axis, empty marker)
vs. reaction time over Fe/MCM-41(air) and Fe/MCM-41(air)-sil catalysts.
4. Conclusions
Fe/MCM-41(N2)-sil
Two Fe/MCM-41 systems calcined in different atmospheres
(inert and oxidant) were studied. Both have extremely small ␣-
Fe2O3 and/or amorphous Fe2O3 species. However, when air is
used a more pronounced spreading with smaller size and a large
oxide/support interaction was obtained.
0
10
20
30
40
50
60
Time (min)
Fig. 9. Selectivity to ␣-campholenic aldehyde produced in the isomerization of ␣-
pinene oxide (left axis, filled marker) and POX conversion (right axis, empty marker)
vs. reaction time over Fe/MCM-41(N2) and Fe/MCM-41(N2)-sil catalysts.
In order to decrease the Brönsted acidity of the silanol groups
present on the MCM-41, a silylation treatment over both series of

196
N.A. Fellenz et al. / Applied Catalysis A: General 435–436 (2012) 187–196
catalysts was carried out. Their efficiency was verified by different
techniques.
The POX isomerization allowed determining that the Brönsted
acidity of the Fe/MCM-41 system has not the sufficient acid strength
to decrease the selectivity towards CPA, since sylilated and non-
sylilated samples have the same selectivity.
[6] Y. Sun, S. Walspurger, J.-P. Tessonier, B. Louis, J. Sommer, Appl. Catal. A 300
(2006) 1–7.
[7] N. He, S. Bao, Q. Xu, Appl. Catal. A 169 (1998) 29–36.
[8] A. Corma, V. Fornes, M.T. Navarro, J. Perez-Pariente, J. Catal. 148 (1994)
569–574.
[9] N. Ravasio, F. Zaccheria, A. Gervasini, C. Messi, Catal. Commun. 9 (2008)
1125–1127.
[10] R. Ryoo, J.M. Kim, J. Chem. Soc. Chem. Commun. (1995) 711–712.
[11] D.A.M. Monti, A. Baiker, J. Catal. 83 (1983) 323–335.
[12] K. Lagarec, D.G. Rancourt, Mossbauer spectral analysis software, Dep. of Phys.
University of Otawa, Version 1.0., 1998.
On the other hand, the different calcination atmospheres lead
to different types of acid Lewis sites of Fe3+. When air is used, the
spreading conditions could be satisfied more easily, leading to the
production of islands like as “raft” of iron oxide. Therefore, only one
type of Fe3+ sites with high coordinative unsaturation is present.
Instead, when the calcination was realized under N2 stream, clus-
tering of the iron species occurs inside the MCM-41 channels. As
a consequence, more than one type of Fe3+ sites with appropriate
geometry to POX isomerization appear. However, their geometry is
less adequate in order to obtain a high CPA selectivity in comparison
with the iron sites located on the iron “raft”.
This study shows how it is be possible to tune the nature of the
acid sites of a selected material by introducing new components;
in this case highly unsaturated Fe species confirm to be good Lewis
acid centers able to act as active sites in reactions requiring such
sites.
[13] A. Gervasini, C. Messi, P. Carniti, A. Ponti, N. Ravasio, F. Zaccheria, J. Catal. 262
(2009) 224–234.
[14] R.R. Sever, R. Alcala, J. Dumesic, T.W. Root, Microporous Mesoporous Mater. 66
(2003) 53–67.
[15] J. Bu, H.K. Rhee, Catal. Lett. 66 (2000) 245–249.
[16] A. Gervasini, C. Messi, D. Flahaut, C. Guimon, Appl. Catal.
113–121.
A 367 (2009)
[17] M. Ojeda, F.J. Pérez-Alonso, P. Terreros, S. Rojas, T. Herranz, M. López Granados,
J.L. García Fierro, Langmuir 22 (2006) 3131–3137.
[18] R. Mariscal, M. Lòpez-Granados, J.L.G. Fierro, J.L. Sotelo, C. Martos, R. Van
Grieken, Langmuir 16 (2000) 9460–9467.
[19] K. Lin, L. Wang, F. Meng, Z. Sun, Q. Yang, Y. Cui, D. Jiang, F.S. Xiao, J. Catal. 235
(2005) 423–427.
[20] S. Haukka, A. Root, J. Phys. Chem. 98 (1994) 1695–1703.
[21] E. Murad, J.H. Johnston, in: G.J. Long (Ed.), Mössbauer Spectroscopy Applied to
Inorganic Chemistry, vol. 2, Plenum Publishing Corporation, 1987.
[22] S. Mørup, H. Topsøe, Appl. Phys. 11 (1976) 63–66.
[23] F. Bødker, S. Mørup, Europhys. Lett. 52 (2000) 217–223.
[24] M. Blume, J.A. Tjon, Phys. Rev. 165 (1968) 446–456.
[25] R.E. Vandenberghe, E. De Grave, C. Landuydt, L.H. Bowen, Hyperfine Interact.
53 (1990) 175–196.
Acknowledgement
[26] W.F. Brown Jr., Phys. Rev. 130 (1963) 1677–1686.
[27] A.M. Alvarez, J.F. Bengoa, M.V. Cagnoli, N.G. Gallegos, S.G. Marchetti, R.C. Mer-
cader, Hyperfine Interact. 161 (2005) 3–9.
[28] S.H. Overbury, P.A. Bertrand, G.A. Somorjai, Chem. Rev. 75 (1975) 547–560.
[29] W.F. Holderich, J. Roseler, G. Heitmann, A.T. Liebens, Catal. Today 37 (1997)
353–366.
[30] L. Alaerts, E. Seguin, H. Poelman, F. Thibault-Starzyk, P.A. Jacobs, D.E. De Vos,
Chem. Eur. J. 12 (2006) 7353–7363.
[31] J. Kaminska, M.A. Schwegler, A.J. Hoefnagel, H. van Bekkum, Recl. Trav. Chim.
Pays-Bas 111 (1992) 432–437.
[32] G. Neri, G. Rizzo, S. Galvagno, G. Loiacono, A. Donato, M.G. Musolino, R.
Pietropaolo, E. Rombi, Appl. Catal. A 274 (2004) 243–251.
The authors acknowledge the financial support of ANPCyT (PICT
No. 00549), which allowed the development of this work.
References
[1] A. Corma, Chem. Rev. 97 (1997) 2373–2419.
[2] A. Corma, Chem. Rev. 95 (1995) 559–614.
[3] A. Corma, H. Garcia, Chem. Rev. 103 (2003) 4307–4365.
[4] G. Busca, Chem. Rev. 107 (2007) 5366–5410.
[5] G. Sartori, R. Maggi, Chem. Rev. 106 (2006) 1077–1104.