Green and selective oxidation reactions catalyzed by kaolinite covalently grafted with Fe(III) pyridine-carboxylate complexes

Article abstract of DOI:10.1016/j.cattod.2011.11.029

The immobilization of Fe(III) picolinate and Fe(III) dipicolinate complexes on kaolinite furnished heterogeneous catalysts, whose catalytic activity was evaluated. The precursor materials were kaolinite grafted with picolinic (Ka-pa) and dipicolinic (Ka-dpa) acids obtained by melting of the pyridine carboxylic acids. To obtain the catalysts Fe(Ka-pa)-n and Fe(Ka-dpa)-n (n = 1, 2, or 3 is the ligand/Fe ratio), the precursors were suspended in Fe³⁺ solutions with cation/ligand ratios of 1:1, 1:2, or 1:3. The resulting materials were characterized by thermal analyses (simultaneous TG/DTA), X-ray diffraction, UV/vis and infrared spectroscopies, and transmission electron microscopy. The grafted complexes were employed as heterogeneous catalysts in the epoxidation of cis-cyclooctene to cis-cyclooctenoxide and in the oxidation of cyclohexane to cyclohexanol and cyclohexanone at ambient temperature and pressure. Hydrogen peroxide was used as oxygen donor at a catalyst/oxidant/substrate molar ratio of 1:300:100. Fe(Ka-pa)-n catalysts were very efficient for cis-cyclooctene epoxidation (38% conversion). For cyclohexane oxidation, Fe(Ka-dpa)-n was 100% selective for cyclohexanone formation, with substrate conversion of 14%. This last series of catalysts was also very effective in the Baeyer-Villiger reaction, with 60% substrate conversion and 100% selectivity for ξ-caprolactone. After reuse (5 times), the catalysts still led to high substrate conversion.

Full text of DOI:10.1016/j.cattod.2011.11.029

Catalysis Today 187 (2012) 135–149
Contents lists available at SciVerse ScienceDirect
Catalysis Today
journal homepage: www.elsevier.com/locate/cattod
Green and selective oxidation reactions catalyzed by kaolinite covalently grafted
with Fe(III) pyridine-carboxylate complexes
Emerson H. de Fariaa,∗, Gustavo P. Riccia, Liziane Marc¸ ala, Eduardo J. Nassara, Miguel A. Vicenteb,
Raquel Trujillanob, Antonio Gilc, Sophia A. Korilic, Katia J. Ciuffia,∗, Paulo S. Calefia
a Universidade de Franca, Av. Dr. Armando Salles Oliveira, Parque Universitário, 201, 14404-600 Franca, SP, Brazil
b Departamento de Química Inorgánica, Facultad de Ciencias Químicas, Universidad de Salamanca, Plaza de la Merced, S/N, 37008 Salamanca, Spain
c Department of Applied Chemistry, Public University of Navarre, Campus of Arrosadia, Pamplona, Spain
a r t i c l e i n f o
a b s t r a c t
Article history:
The immobilization of Fe(III) picolinate and Fe(III) dipicolinate complexes on kaolinite furnished hetero-
geneous catalysts, whose catalytic activity was evaluated. The precursor materials were kaolinite grafted
with picolinic (Ka-pa) and dipicolinic (Ka-dpa) acids obtained by melting of the pyridine carboxylic acids.
To obtain the catalysts Fe(Ka-pa)-n and Fe(Ka-dpa)-n (n = 1, 2, or 3 is the ligand/Fe ratio), the precursors
were suspended in Fe3+ solutions with cation/ligand ratios of 1:1, 1:2, or 1:3. The resulting materials were
characterized by thermal analyses (simultaneous TG/DTA), X-ray diffraction, UV/vis and infrared spectro-
scopies, and transmission electron microscopy. The grafted complexes were employed as heterogeneous
catalysts in the epoxidation of cis-cyclooctene to cis-cyclooctenoxide and in the oxidation of cyclohexane
to cyclohexanol and cyclohexanone at ambient temperature and pressure. Hydrogen peroxide was used
as oxygen donor at a catalyst/oxidant/substrate molar ratio of 1:300:100. Fe(Ka-pa)-n catalysts were
very efficient for cis-cyclooctene epoxidation (38% conversion). For cyclohexane oxidation, Fe(Ka-dpa)-n
was 100% selective for cyclohexanone formation, with substrate conversion of 14%. This last series of
catalysts was also very effective in the Baeyer–Villiger reaction, with 60% substrate conversion and 100%
selectivity for ␰-caprolactone. After reuse (5 times), the catalysts still led to high substrate conversion.
© 2011 Elsevier B.V. All rights reserved.
Received 2 June 2011
Received in revised form
21 November 2011
Accepted 23 November 2011
Available online 3 January 2012
Keywords:
Kaolinite
Heterogeneous catalysis
Fe(III)–picolinates and dipicolinates
Oxidation reactions
Baeyer–Villiger
1. Introduction
context, layered materials have been successfully employed as cat-
alyst supports in hydrocarbon oxidation, enhancing the catalytic
Catalysis is the base of countless chemical processes that trans-
form crude and cheap raw material into products of high economic
value and extreme importance for humankind. Moreover, catalytic
technologies are of utmost importance for environmental sustain-
ability [1]. The most commonly employed catalytic routes make use
of metallic catalysts to improve reaction rate and selectivity [2–4].
Traditional hydrocarbon oxidation uses vast amounts of highly
toxic inorganic compounds such as potassium permanganate and
dichromate. According to Hill [5], the ideal “green” oxidizing sys-
tems are molecular oxygen and hydrogen peroxide in combination
with recyclable catalysts in non-toxic solvents. Immobilization of
metal complexes on inorganic supports such as clay minerals has
led to the successful preparation of efficient and selective hetero-
geneous catalysts for the oxidation of organic substrates. In this
activity of the metal complex and promoting product selectivity
[6]. Recently, catalysts containing iron or manganese porphyrins as
active phases have been reported [7–12].
Environmentally friendly catalytic reactions have increasingly
drawn the attention of scientists, environmentalists, and politicians
worldwide [13], as these reactions lead to low waste generation,
not to mention that they enable the use of mild conditions and
the design of highly selective and active heterogeneous catalysts,
which is desirable in an age when the control of environmental
pollution and the reduction of energy consumption are manda-
tory (sustainable chemistry). Some zeolites and natural clays
(montmorillonite, saponite, kaolinite) display interesting catalytic
properties, and the use of modified clays as catalysts has been
encouraged because they matches the sustainable chemistry con-
ditions. Layered materials were already employed as catalysts in
the beginning of the petrochemical industry, and the first hydro-
cracking process was based on acid-treated clays [2]. However, it
is believed that synthetic systems consisting of functionalized or
intercalated clays can overcome the performance of such materi-
als. Indeed, processes that promote the structural modification of
clays can tailor the characteristics of the catalysts in such a way that
∗
Corresponding authors.
E-mail addresses: eh.defaria@unifran.br (E.H. de Faria), ciuffi@unifran.br
(K.J. Ciuffi).
0920-5861/$ – see front matter © 2011 Elsevier B.V. All rights reserved.
doi:10.1016/j.cattod.2011.11.029

136
E.H. de Faria et al. / Catalysis Today 187 (2012) 135–149
their activity and selectivity in oxidation reactions can be enhanced
[14–17]. Some advantages of using functionalized clays are:
oxidation of menthone and tetrahydrocarvone to their correspond-
ing lactones using monopersulfuric acid (H2SO5) as oxidant, this
versatile reaction has become one of the most applied industrial
processes [34,35], being useful for the preparation of chemical,
agrochemical and pharmaceutical compounds [36]. Catalysts com-
monly employed for this oxidation are zeolites and aluminas [37],
while clays as montmorillonite [35] and kaolinite [25] have also
been utilized.
Traditionally, peracids are employed as oxidants in the
Baeyer–Villiger reaction. However, it is evident that these oxi-
dants are troublesome for large-scale utilization. One of the factors
that make a catalyst efficient for the Baeyer–Villiger reaction is its
ability to increase the electrophilicity of the peroxide. This acti-
vation enhances the oxidant’s activity toward electron deficient
centers, such as carbonyls. Metals such as Ti, V, Mo, and W may
form peroxo complexes in the presence of H2O2, but these have an
electrophilic nature, which justifies their activity in hydrocarbon
oxidation. However, the nucleophilic attack of the peroxo group
to carbonyl groups is improbable. Jacobson et al. [38] were the
first to report on the nucleophilic reactivity of peroxo complexes
containing transition metals. These authors employed the Mo(VI)-
dipicolinate complex in the oxidation of cyclopentanone at 60 ◦C
and achieved 93% product yield after 24 h. Recently, many research
groups have investigated this reaction in the presence of various
catalysts, like natural and/or synthetic clays, alumina, and transi-
tion metals supported on different matrices [25,35–38].
•
The possibility of achieving uniform intercalation or grafting of
the active species during the synthesis (without impurities).
•
The large-scale production of catalysts with homogeneous chem-
ical composition and catalytic properties.
•
Complexation of highly catalytically active transition metal ions
such as Fe3+, Co3+, and Mn3+ species with functional groups
located in the basal space of clays, giving rise to efficient, selective
heterogeneous catalysts [18–20].
As already commented, layered materials represent an inter-
esting opportunity for the development of new catalysts [2]. In this
context, kaolinite, with theoretical formula Al2Si2O5(OH)4, is a typ-
ical example of layered structure, since it is formed by combining
sheets of SiO4 tetrahedra (T) and Al(OH)6 octahedra (O) in a 1:1
proportion [21,22]. The use of layered materials as heterogeneous
catalysts is based on their surface activity, which can promote a
large variety of organic reactions such as Diels–Alders cyclization of
organic compounds, and oxidation reactions of organic compounds
such as cis-cyclooctene, cyclohexane and the Baeyer–Villiger oxi-
dation [23].
The combination of organic and inorganic constituents on the
atomic/molecular scale gives rise to materials with singular prop-
erties, thereby allowing for adjustable applications. Such materials
include those obtained by clay mineral intercalation or function-
alization, whose morphological characteristics enable superficial
and interplanar chemical modifications that lead to the effective
immobilization of complexes, such as metalloporphyrins, organic
molecules, polymers, and alkoxides [24,25]. The materials result-
ing from immobilization processes have followed a fast growing
interest in functional properties and applications in many areas
of science and technology such as catalysis, adsorption, conduc-
tors, and optical and photoactive materials [26–29]. Recently our
research group has reported the functionalization of natural kaoli-
nite with pyridine-carboxylic acids, alkoxides, and ironporphyrins
by the soft-guest displacement methodology.
This paper reports the immobilization of Fe(III)–picolinate and
Fe(III)–dipicolinate on kaolinite. The grafted complexes were used
as heterogeneous environmental-friendly catalysts in the epoxi-
dation of cis-cyclooctene to cis-cyclooctenoxide, in the oxidation
of cyclohexane to cyclohexanol and cyclohexanone, and in the
Baeyer–Villiger oxidation of cyclohexanone to ␰-caprolactone,
using hydrogen peroxide as oxidant. The nature of the catalysts
and the catalytic routes studied are summarized in Fig. 1.
2. Experimental
Catalytic oxidation reactions generally involve the use of solu-
ble salts or complexes in the presence of oxidants like O2, H2O2,
iodosylbenzene (PhIO), or KMnO4. The catalytic oxidation of the
2.1. Catalyst preparation
2.1.1. Functionalization of natural kaolinite with
C
H bonds of alkanes at ambient temperature and pressure has
Fe(III)–picolinate and Fe(III)–dipicolinate
attracted extensive and ongoing interest worldwide, by using
systems that can mimic the dioxygen activation induced by iron-
containing biological systems such as cytochrome P-450 [30].
Over the last decade the search for efficient and environmen-
tally friendly oxidation procedures that could be used to develop
green processes for many kinds of oxidation reactions has been
intensified. This can be achieved by employing clean oxidants like
hydrogen peroxide, which is considered as an ideal oxidant because
of its high active oxygen content, availability, and non-toxicity.
Moreover, it is considered to be non-polluting, since it produces
only water as product [31]. Oxidations with hydrogen peroxide are
highly atom-economic, and it is a safe, readily available and cheap
reagent.
Traditionally ironporphyrins have been reported as efficient and
selective catalysts for hydrocarbons oxidation and they are good
biomimics of cytochrome P-450 [18,25,30–33]. However, their use
is not always feasible on an industrial scale, because their synthesis
demands arduous purification procedures and chlorinated solvents
are employed in their syntheses. This encouraged us to investigate
the potential of the catalyst prepared herein as biomimetic cata-
lysts for hydrocarbons oxidation, using the clean oxidant hydrogen
peroxide.
The kaolinite employed in this work came from the munic-
ipality of São Simão in the State of São Paulo, Brazil, and was
kindly supplied by the mining company Darcy R.O. Silva & Cia.
It belongs to the ball-clay type, known for its fine granulom-
etry and for being rich in hexagonal kaolinite. Kaolinite was
purified by dispersion in water, followed by sedimentation, accord-
ing to Stock’s law and using the procedure previously described
by us [25,27–29]. The structural formula of purified kaolinite is
Si2.0Al1.96Fe0.03Mg0.01K0.02Ti0.03O7.06 [27].
Grafted kaolinite-pyridine carboxylic acids compounds were
prepared according to the methodology described by de Faria et al.
[27]. Briefly, kaolinite was first expanded with dimethylsulfoxide,
the resulting complex, designated Ka-DMSO, has the stoichiome-
try Ka(DMSO)0.45 (Ka denotes the cell-formula of kaolinite reported
above). This precursor was heated for 40 h in the presence of melted
acid (picolinic or dipicolinic) in a 1:5 molar ratio, and then the
resulting materials were washed with ethanol and oven-dried at
80 ◦C. They were designated Ka-pa and Ka-dpa when derived from
picolinic and dipicolinic acids, respectively. To obtain the derivate
Fe containing catalysts, Ka-pa and Ka-dpa solids were suspended
in a 0.1 mol/L Fe3+ solution, using the volumes needed to reach
ligand/Fe ratios, denoted as n, of 1.0, 2.0, or 3.0, and the mixture
was stirred at 55–60 ◦C for 3 h. The suspensions were centrifuged,
and the resulting solids were washed with ethanol five times.
Other reaction considered in this work is Baeyer–Villiger oxi-
dation. From the historical report by Baeyer and Villiger of the

E.H. de Faria et al. / Catalysis Today 187 (2012) 135–149
137
Fig. 1. Schematic representation of catalytic oxidations by Fe(III)–pyridine carboxylic complexes grafted on kaolinite.
The catalysts obtained were designated by a general nomenclature
Fe(Ka-pa)-n and Fe(Ka-dpa)-n when reference is made to a com-
plete series, and specifying the value of n when reference is made
to a given sample.
absorption after dissolving the samples in nitric acid, in a Mark-II
ELL-240 instrument.
Transmission electron microscopy (TEM) was performed using
a Zeiss-902 microscope over samples ground and dispersed in
ethanol by using an ultrasonic apparatus; then, a drop of the sus-
pension was placed in a Cu grid and air dried before the study.
Products from the catalytic oxidation reactions were identi-
fied using a HP 6890 Series GC System gas chromatograph (with
a flame ionization detector) equipped with a capillary column HP-
INNOWax-19091N-133, polyethylene glycol length 30 m, internal
diameter 0.25 ␮m). The products were quantified using a calibra-
tion curve obtained with a standard solution, and the conversion
was based on substrate. Technical error: about 1%.
2.1.2. Synthesis of isolated Fe(III)–picolinate and
Fe(III)–dipicolinate complexes – homogenous catalysts
The Fe(III) picolinate and dipicolinate complexes were pre-
pared according to the methodology described by Amani et al.
[39]. Picolinic (500 mg, 4.06 × 10−3 mol) or dipicolinic (500 mg,
2.99 × 10−3 mol) acid in 0.1 mol/L HCl solution (20 mL) was added
to a solution of FeCl3·6H2O (439 mg, 1.62 × 10−3 mol or 324 mg,
1.19 × 10−3 mol, respectively) in distilled water, and the resulting
solution was stirred at 55–60 ◦C for 3 h. The complexes were pre-
cipitated, and then recrystallized from acetonitrile, yielding 97%
and 95% from the reactions with picolinic and dipicolinic acid,
respectively. The resulting complexes were designated Fe(pa)3 and
Fe(dpa)3, respectively.
All solvents and reagents were purchased from Mallinckrodt,
Aldrich, or Acros Organics and were of commercial grade, unless
otherwise stated. Dichloromethane (DCM) was suspended on anhy-
˚
drous CaCl2 for 2.5 h, filtered, distilled over P2O5, and kept over 4 A
molecular sieves. cis-Cyclooctene, cyclohexane, cyclohexanol, and
cyclohexanone were purified on a basic alumina column imme-
diately before use. Hydrogen peroxide (solution in water) was
donated by Peróxidos do Brasil SA and was iodometrically titrated
before use (indicating a H2O2 content of 60 wt.%) [40].
2.2. Characterization techniques
The X-ray diffractograms of the solids were acquired on a
Siemens D-500 diffractometer operating at 40 kV and 30 mA
(1200 W), using filtered Cu K␣ radiation, varying the angle 2ꢀ from
2◦ to 65◦. All the analyses were processed at a scan speed of 2◦/min.
Infrared absorption spectra were obtained on a Perkin-Elmer
1739 spectrophotometer with Fourier transform, using the KBr pel-
let technique.
UV–vis spectra were recorded in the 200–800 nm range on a HP
8453 Diode Array Spectrophotometer. Spectra of the solid samples
were recorded in a quartz cell with 0.1 cm path length, in ethanol
suspension.
Thermal analyses (TG/DSC) were carried out in a TA Instruments
SDT Q600 Simultaneous DTA-TGA thermal analyzer, in the temper-
ature range between 25 and 1100 ◦C, at a heating rate of 20 ◦C/min
and air flow of 100 mL/min.
Element chemical analysis was carried out at Servicio General
de Análisis Químico Aplicado (University of Salamanca) by atomic
2.3. Catalytic activity
2.3.1. Oxidation of cis-cyclooctene and cyclohexane by hydrogen
peroxide
Catalytic oxidation reactions were carried out in a 2 mL glass
reactor equipped with a magnetic stirrer. In a typical hetero-
geneous catalysis reaction, 10 mg of the catalyst Fe(Ka-pa)-n or
Fe(Ka-dpa)-n were suspended in 1 mL of the solvent mixture
(dichloroethane/acetonitrile, 1:1, v/v), and the substrate (cis-
cyclooctene (125 ␮L) or cyclohexane (100 ␮L)) was added. Molar
ratio of catalyst/H2O2/substrate was 1:300:100 (the molar mass of
each catalyst was calculated from its structural formula). The oxi-
dation reactions were carried out during a controlled time interval
of 2, 4, 24, or 48 h. The reaction products were analyzed by gas
chromatography using the internal standard method.

138
E.H. de Faria et al. / Catalysis Today 187 (2012) 135–149
At the end of the reaction, the Fe(Ka-pa)-n and Fe(Ka-dpa)-n cat-
alysts were recovered by centrifugation, washed in dichloroethane
and dried for 3 h at 60 ◦C before being used again in a further cis-
cyclooctene or cyclohexane oxidation reaction.
The concentration of hydrogen peroxide after the oxidation
reactions was measured by iodometric titration to determine the
real efficiency of hydrogen peroxide [40].
Fe-(Ka-pa)-3
Fe-(Ka-pa)-2
2.3.2. Oxidation of cis-cyclooctene and cyclohexane by
iodosylbenzene
Iodosylbenzene (PhIO) was obtained through hydroly-
sis of iodosylbenzene diacetate [40], and the purity of the
obtained compound was evaluated by iodometric titration
[41]. PhIO (0.023 mmol) was added to
a 2.0 mL vial sealed
with Teflon-coated silicone septum containing the cata-
a
lyst (10 mg) and 1,2-dichloroethane (1 mL). cis-Cyclooctene,
cyclohexane, cyclohexanol, all of them previously purified
on alumina column, or cyclohexane were added as sub-
strate (1.15 mmol) and di-n-butyl ether was employed as
internal standard (10 ␮L). The reaction products were ana-
lyzed by gas chromatography (GC). The products were quantified
using a calibration curve obtained with a standard solution, and
the yield was based on oxidant.
Fe-(Ka-pa)-1
2.3.3. Baeyer–Villiger (BV) oxidation of cyclohexanone
The BV oxidation reaction was carried out in a 4.0 mL glass
reactor. A 50 ␮L portion of cyclohexanone (0.48 mmol), 850 ␮L ben-
zonitrile (7.95 mmol), 5.0 ␮L di-n-butyl ether (2.94 ␮mol), 176 ␮L
of 60 wt.% (4.53 mmol) or 352 ␮L (4.53 mmol) of 30 wt.% hydrogen
peroxide and 20 mg of the catalysts were added to the reactor; the
mixture was then heated to 80 ◦C and stirred for 48 h. At the end
of the reaction, the catalysts were recovered, dried for 3 h at 60 ◦C
and used again in a further Baeyer–Villiger oxidation reaction.
In both cases, control reactions were carried out for all the oxi-
dation reactions in the absence of catalysts, in absence of oxidants
or using kaolinite grafted with picolinic and dipicolinic acids but
without containing Fe(III), while the control test to Baeyer–Villiger
oxidation reactions was conducted without benzonitrile using as
solvent a mixture of dichloroethane/acetonitrile, 1:1 (v/v), the
molar ratio of catalyst/oxidant/substrate was maintained.
Ka-pa
10
20
30
40
50
60
2 (Degrees)
Fig. 2. X-ray powder diffraction patterns (2ꢀ = 2–65◦) of the precursor Ka-pa (a), and
the catalysts Fe(Ka-pa)-1 (b), Fe(Ka-pa)-2 (c) and Fe(Ka-pa)-3 (d).
The supernatant was separated from the catalysts by centrifu-
gation and maintaining the supernatant of this reaction on a glass
reactor until 24 h, aftermost the possible products formed were
quantified by GC. Indeed, this simple test can give evidence of the
leaching of active Fe(III)-species from solid to supernatant.
peak of non-expanded kaolinite. In the case of Ka-dpa series, the
˚
basal spacing decreased from 11.90 A in the dpa-expanded kaolinite
˚
to 11.40 A in the Fe-containing solids, which suggests a more pla-
3. Results and discussion
nar arrangement of dpa molecules when complexating Fe(III) ions.
At the same time, new reflections are observed, for all the solids in
˚
this series, at 8.39 A, assigned to the formation of a tubular phase of
3.1. Characterization of kaolinite functionalized with Fe(III)
complexes
˚
˚
kaolinite due to a curling process in the layers and at 6.30 A, 5.90 A,
˚
and 5.50 A, assigned to the 0 0 2 reflection present in solids with
The X-ray diffractograms of the grafted solids derived from
kaolinite are presented in Figs. 2 and S1 (Supplementary material).
The solids intercalated with picolinic and dipicolinic acids dis-
high crystallinity. Recently, curling of single layers or even small
stacks of layers of kaolinite after intercalation and functionaliza-
tion has been reported, resulting in the so-called “halloysite-like”,
tubular, or scroll-like kaolinite. As for kaolinite functionalized with
dipicolinic acid and complexated with Fe(III), peaks at 8.39, 6.28,
˚
played basal spacings of 13.50 and 11.90 A, respectively, in both
˚
cases also presenting a small effect at 7.20 A, characteristic spac-
˚
ing of non-intercalated kaolinite [27]. The incorporation of Fe(III)
did not affect the basal distance of Ka-pa solids, although the peak
characteristic of expanded kaolinite decreased in intensity, and that
from non-expanded kaolinite increased, suggesting that part of the
ligand may leave the interlayer region to form complexes out of
it, and even that complexation process in water medium promotes
the disorder of kaolinite layers. This mainly happened in the Fe-
(Ka-pa)-3 solid, which has the higher Fe/pa ratio, in which two new
peaks characteristic of Fe–pa complexes are observed at 9.30 and
5.95 and 5.48 A were detected, ascribed to the presence of the
kaolinite tubular phase [18,32,33,42]. This phase is obtained due to
interaction between the Fe(III)–dipicolinate complex and the inor-
ganic matrix, thereby reducing the interactions between the layers
and promoting their curling [18,32,33,42]. XRD patterns revealed
˚
that the peak at 7.14 A, characteristic of pure kaolinite remain-
˚
ing from the synthesis, is displaced to 7.20 A in the case of the
kaolinite derivatives Ka-DMSO, Ka-pa, and Ka-dpa, and this peak
becomes broad after iron coordination because of the presence of
water molecules coordinated to the Fe(III) complexes located in the
˚
5.60 A. This logically provokes an increase in the intensity of the

E.H. de Faria et al. / Catalysis Today 187 (2012) 135–149
139
Table 1
justified by the strong affinity of the ligands for Fe(III), as the
entrance of Fe(III) to the interlayer region is relatively difficult,
some picolinate anions go out of this region to form the very stable
complexes. On the other hand, it became evident that lixiviation
depends on the number of carboxylate groups present on the func-
tionalized kaolinite. Indeed, the hybrids Ka-dpa were more stable
against lixiviation, probably because the dipicolinate anions, larger
than the picolinate ones, have more difficulties to leave the inter-
layer region, while Ka-pa series showed higher lixiviation of the
ligands out of the layers.
The formation of kaolinite nanotubes (or scroll-like structure)
was confirmed by TEM (Fig. 3). The micrographs showed that
the tubes have different diameters, probably because the curl-
ing process of kaolinite was incomplete. Nakagaki et al. have
assigned the formation of nanotubes in silanized kaolinite inter-
calated with a porphyrin to the strong interaction between the
silanized kaolinite layer and the porphyrin [18,32,33]. The inter-
calation of organic moieties hinders or reduces the formation
of hydrogen bonds between silica (Si O) and alumina (Al OH)
in kaolinite layers, thereby promoting curling of the layers. It is
known that curling process is favored when the hydrogen bonds
of the kaolinite structure are weakened by the intercalation or
grafting of strongly polar molecules such as amines, alcohols, etc.
Gardolinski and Lagaly have described the curling of kaolinite lay-
ers when intercalated with large amine molecules (n-alkylamine,
octadecylamine) [43]. The preparation of halloysite-like kaolin-
ite is more efficient in materials with a high degree of disorder
between the layers, thus being favored for hybrid compounds.
Interlayer basal space, d0 0 1, variation in basal space (ꢁd), and intercalation ratio (˛)
of the materials prepared in this work.
*
Sample
d0 0 1 (Å)
ꢁd (Å)
˛
(%)
Ka
Ka-DMSO
Ka-pa
Fe(Ka-pa)-1
Fe(Ka-pa)-2
Fe(Ka-pa)-3
Ka-dpa
Fe(Ka-dpa)-1
Fe(Ka-dpa)-2
Fe(Ka-dpa)-3
7.14
11.20
13.50
14.20
13.90
13.90
11.90
11.50
11.50
11.35
–
–
4.06
6.36
7.06
6.76
6.76
4.76
4.36
4.36
4.21
98
92
44
51
51
86
80
68
66
*
The intercalation ratio (˛) was determined according to the relative intensities
of the peaks relative to the basal space d0 0 1 in relation to the characteristic reflection
of purified kaolinite. ˛ = I/(I + I0), I = intensity d0 0 1 of the products, I0 = intensity d0 0 1
of the characteristic reflection of non-reacted kaolinite.
interlayers of kaolinite. Table 1 summarizes the results from XRD
patterns of grafted derivatives after complexation with Fe(III).
The intercalation ratio, that is, the percentage of layers inter-
calated, decreased from 92% in the Ka-pa solid to 44–51% in the
Fe-containing solids in this series. Although the intercalation ratio
is slightly lower in the parent solid of Ka-dpa series (86%), the
amount of layers intercalated in the Fe-containing solids was
clearly higher, 66–80% (Table 1). This suggests that the presence
of Fe3+ ions induces a partial lixiviation of the intercalated lig-
ands, forming complexes located out of the interlayer region. This is
Fig. 3. Transmission electron micrographs of the Fe(Ka-dpa)-3 materials showing the halloysite-like structure with magnifications of 25,000×.

140
E.H. de Faria et al. / Catalysis Today 187 (2012) 135–149
Fig. 4. Transmission electron micrographs with 25,000× magnification of the precursors Ka-pa (A.1) and Ka-dpa (A.2), of the catalysts obtained after complexation in
ethanolic medium: Fe(Ka-pa)-3[eth] (B.1) and Fe(Ka-dpa)-3[eth] (B.2), and of the catalysts obtained after complexation in aqueous medium: Fe(Ka-pa)-3[H2O] (C.1) and
Fe(Ka-dpa)-3[H2O] (C.2).
TEM micrographs also evidenced the presence of some crystals of
pure kaolinite with unmodified hexagonal structure, which is in
agreement with the maintenance of the 7.2 A – reflection in the
As several authors have attributed kaolinite curling to the pres-
ence of water molecules, this influence was studied by synthesizing
one of the solids in each series in the absence of water. The sam-
ples were designated as Fe-(Ka-pa)-3[eth] and Fe-(Ka-dpa)-3[eth],
according to the ethanolic reaction medium. Fig. 4 depicts the TEM
micrographs of these samples, which clearly reveals that there was
no formation of the tubular phase with halloysite-like structure in
ethanolic medium (in this figure, for avoiding confusion, [H2O] is
added to the name of the samples prepared in aqueous medium).
This confirmed that the curling process was really promoted by
the presence of water molecules and not by formation of pa– or
˚
XRD patterns.
Thus, the curling process in our solids may be induced
by the presence of water and by the coordination of Fe(III)
to the pa and dpa anions grafted to the kaolinite layer.
With the aid of the software Orion 6.53 (free available at
http://www.orionmicroscopy.com/demo.php), the internal and
external diameters of the nanotubes were determined to be 30 nm
and 80 nm, respectively, and their thickness was found to be 25 nm.

E.H. de Faria et al. / Catalysis Today 187 (2012) 135–149
141
a - Ka-pa
FeCl3
pa
(Ka-pa)
Fe-(Ka-pa)-3
196 LMCT
b - Fe-(Ka-pa)-1
c - Fe-(Ka-pa)-2
d - Fe-(Ka-pa)-3
d
c
245
217
220
b
a
350
265
200
300
400
Wavelength (nm)
500
600
4000
3500
3000
2500
2000
1500
1000
500
Wavenumber (cm-1)
Fig. 6. UV/vis spectra of the compounds Fe(III) chloride, picolinic acid (pa), Ka-pa
and Fe(Ka-pa)-3.
Fig. 5. Infrared spectra of the compound Ka-pa (a), and the catalysts Fe(Ka-pa)-1
(b), Fe(Ka-pa)-2 (c) and Fe(Ka-pa)-3 (d).
free electron pair of the nitrogen atoms in pa or dpa anions grafted
on kaolinite to the t2g, Eg empty Fe(III) orbitals.
dpa–Fe(III) complexes. In aqueous medium, two factors may have
contributed to the formation of the tubular phases: the presence of
water molecules and the reduced cohesion force between kaolinite
layers after insertion of pa and dpa molecules.
The thermal behavior of the catalysts Fe(Ka-pa)-3 and Fe(Ka-
dpa)-3 was investigated by thermogravimetry and differential
thermal analysis under oxygen atmosphere (Figs. 7 and S4). Both
catalysts showed high thermal stability, above 300 ◦C. A first mass
loss step, with an associated endothermic peak, was observed at
60–65 ◦C, being attributed to the elimination of small amounts of
solvent adsorbed in the clay. Before the intense mass loss ascribed
to the thermal decomposition of the complexes, both solids showed
a small endothermal effect close to 270 ◦C, best observed in the DTG
curves, probably due to the removal of small amounts of water. The
decomposition of the organic ligands appears as a strongly exother-
mic process, more intense in the Fe(Ka-dpa)-3 solid, centered at
404 ◦C. This process is actually composed of two steps, centered at
345 and 480 ◦C in Fe(Ka-pa)-3 solid and at 400 and 498 ◦C in the case
of Fe(Ka-dpa)-3. This may be reasonably ascribed to the removal in
the first process of the ligands forming complexes out of the inter-
layer space and in the second process of the ligands located inside
the interlayer space. The fact that the temperature of the second
process is considerably higher for the Ka-dpa series may be due to
the two carboxylate groups in dpa anions, which makes them to be
more strongly bonded to Fe(III) cations. In both cases, the presence
Figs. 5 and S2 show the FTIR spectra of Ka-pa, and Ka-dpa
precursors, and of the catalysts obtained after complexation with
Fe(III), the assignment of the bands is summarized in Table S1 (Sup-
plementary material). The infrared spectra of parent kaolinite
displayed four bands in the 3500–3800 cm−1 region (3694, 3668,
3650, 3618 cm−1), which can be assigned to the stretching of the
inter- and intralamellar hydroxyls [27]. The displacement of these
bands evidenced the intercalation of molecules in this region. In
Ka-pa and Ka-dpa solids, the bands typical of the asymmetric
and symmetric stretching of the carboxylic acid groups were also
observed at 1689, 1566, and 1478 cm−1 [27–29].
After complexation with Fe(III), the bands attributed to the
stretching of C N, C O and C H bonds shifted from 1649, 1615,
1570, 1480–1454, 1345, and 1297 cm−1 to 1678, 1642, 1604,
1499–1480, and 1290 cm−1, thereby confirming the coordination
of Fe(III) to the N atom and the carboxylate group of the picoli-
nate anions grafted on kaolinite. The band observed at 1705 cm−1
is ascribed to the C O stretching vibration of the pa carboxylic acid
group coordinated to the Fe(III) ion. The appearance of a band at
1480 cm−1 can also be ascribed to the formation of the Fe(Ka-pa)-n
and Fe(Ka-dpa)-n complexes in the basal space of kaolinite. Álvaro
et al. have reported the incorporation of Fe(III) picolinate complexes
into zeolite Y and Na-mordenite, observing that the presence of the
band at 1480 cm−1 can be taken as a spectroscopic evidence for the
formation of Fe(pa)3 complex [4].
100
TG
DTG
DTA
95
65
90
Figs. 6 and S3 show the Diffuse-Reflectance UV–vis spectra of
the solids Fe(Ka-pa)-3 and Fe(Ka-dpa)-3, compared to the solu-
tion spectra of Fe(III) chloride, the ligands pa and dpa, and the
complexes Ka-pa and Ka-dpa. The UV–vis spectra of Ka-pa and Ka-
dpa display bands at 217 and 265 nm; and 204, 220, and 260 nm,
respectively, characteristic of the absorption of picolinic and dipi-
colinic acids grafted in the inorganic matrix. For the Fe-containing
solids, the bands at 196, 220, and 265 nm are associated to ligand
to metal charge-transfer (LMCT). These bands can reasonably be
attributed to the complexation of Fe(III) species with the ligands
grafted on kaolinite [44]. However, other authors [45] consider that
these bands are due to the d–d transfer to 2Eg state. In our solids,
this LMCT band seems to be due to the electron transfer from the
270
85
498
80
75
70
998
400
400
200
600
800
1000
Temperature (ºC)
Fig. 7. TG/DTA measurements of Fe(Ka-dpa)-3 solid performed in O2 atmosphere.

142
E.H. de Faria et al. / Catalysis Today 187 (2012) 135–149
Table 2
Thermal analysis results for Ka-pa and Ka-dpa solids.
Sample
Decomposition temperature, ◦C (Mass loss, %)
H2O/ethanol
Grafted ligands
Complexed ligands
Kaolinite dehydroxylation
Ka-pa
25–190 (3)
25–190 (3)
25–190 (4)
25–190 (3)
25–200 (3.5)
25–200 (1)
25–200 (1)
25–200 (2)
150–380 (24)
200–380 (5)
200–380 (7)
200–380 (15)
200–380 (14)
200–380 (6)
200–380 (7)
200–380 (7)
–
400–460 (13)
400–600 (13)
400–600 (12)
–
380–420 (7)
280–420 (8)
280–420 (8)
450–1100 (11)
400–1100 (13)
400–1100 (18)
400–1100 (15)
400–1100 (16)
400–1100 (12)
400–1100 (11)
400–1100 (11)
Fe(Ka-pa)-1
Fe(Ka-pa)-2
Fe(Ka-pa)-3
Ka-dpa
Fe(Ka-dpa)-1
Fe(Ka-dpa)-2
Fe(Ka-dpa)-3
of Fe(III) coordinated to the picolinates increases the thermal stabil-
ity with respect to the corresponding Ka-pa and Ka-dpa precursors,
the picolinate anions are eliminated at higher temperatures; in the
first case, the pa units decompose at 370 ◦C and in the second one at
375 and 473 ◦C [27]. At higher temperatures, the dehydroxylation
of kaolinite at 470 ◦C and the formation of a phase close to mullite
at ca. 1000 ◦C are observed [27,46–49]. The temperature and mass
loss in the different processes are summarized in Table 2.
oxidant, and high temperatures (∼160 ◦C) and pressures (15 bar).
Under these conditions, conversion levels are as low as 4%, with
80% selectivity to K-A oil and a cyclohexanol/cyclohexanone molar
ratio of 2:1 [54]. Although the oxidant used in this system is cheap,
it is only efficient at temperatures around 70 ◦C. Therefore, the
search for catalysts that are efficient at mild temperatures is a great
challenge for both researchers and industries, as emphasized by
Schuchardt et al. [55,56].
The ligands in the precursors are divided in two types in the final
solids, those coordinated or not coordinated to Fe(III) cations. The
mass loss corresponding to each type is shown in Table 2. A good
agreement is observed between the initial ligand amount and the
final values, the small differences may be due to a small leaching
of ligands during the incorporation of Fe, and to relative changes
in composition due to the variation of the amount of water and
of Fe. The amount of ligands is very similar for the solids in each
series, but rather different between both series, ca. 13% in Ka-pa
series and 8% in dpa series, which is probably related to the higher
ligand ability of dpa by the presence of a second carboxylic group.
It is noticeable that the amount of coordinated ligands is almost
constant in the three solids of a series; this perfectly matches with
the fact that the amount of Fe is also almost constant in the three
solids, in spite of the different amount added in the synthesis. Thus,
the solids in Ka-pa series have Fe contents close to 6.7% and those
in Ka-dpa series close to 5.30% (expressed as Fe2O3, on dry basis
and once subtracted the Fe present in the original kaolinite). It is
remarkable that the Fe contents are almost constant in each series,
as observed in the contents of the ligands. The ratio dpa/Fe results to
be close to 2.0, while the ratio pa/Fe is close to 1.0, which seems to be
related to the accessibility of Fe(III) to the ligands in the interlayer
region.
The results of the oxidation of cis-cyclooctene catalyzed by
Fe(Ka-pa)-n and Fe(Ka-dpa)-n heterogeneous catalysts and the
corresponding homogeneous catalysts Fe(pa)3 and Fe(dpa)3 using
hydrogen peroxide as oxidant are summarized in Table 3. The
conversion was analyzed after 2, 4, 24 and 48 h of reaction. It is
remarkable that, in all cases, the selectivity to cyclooctene oxide
was 100%.
The heterogeneous catalysts led to high conversions of cis-
cyclooctene to cis-cyclooctene epoxide, mainly in the Fe(Ka-pa)-n
series, in which reached values up to 37%, as indicated with total
selectivity to the epoxide. These conversions were similar to those
obtained for the corresponding homogeneous system. This clearly
suggests that the access to the active Fe(III) sites is easy, lowering
the amount of catalyst needed for reaching good conversion values
at ambient temperature and pressure. This is an important advan-
tage with respect to some reported biomimetic systems consisting
of metalloporphyrins heterogeneized in solid matrices, in which
the difficult access of the substrate and the oxidant to the active
catalytic sites makes that high oxidant/catalyst ratios have to be
used in order to obtain good substrate conversions, also making
difficult the reuse of the catalysts.
The lower conversions in the Fe(Ka-dpa)-n solids is probably
because their lower interlayer heights difficult the access of the
substrate to the interlayer region, mainly in the Fe(Ka-dpa)-3 cat-
alyst, which contains the higher amount of organic ligand. The
catalysts Fe(Ka-pa)-n, in turn, exhibit larger basal distance, allow-
ing a more facile access of the substrate to the catalytic sites in the
matrix and explaining the higher yields obtained. Taken together,
these data prove that Fe(III) plays a fundamental role in the oxi-
dation reaction. This explanation is reinforced by the amount of
complexes present in the basal space of kaolinite, the higher ligand
amount in Ka-(pa)1.59 in relation to Ka-(dpa)0.38 contributes to the
efficiency of the catalyst [26]. The catalysts based on Ka-pa precur-
sor present higher layer disorder that those based on Ka-dpa, which
can be assigned to the high amount of picolinic moieties (grafted,
intercalated or adsorbed) in the interlayer space of the kaolinite.
This higher disorder promotes the accessibility of the substrate to
the Fe(III)–picolinate active sites, explaining the higher conversion
of Ka-pa solids to cis-cyclooctene (37%) when compared to series
Ka-dpa (15%). On the other hand, the longer reaction time nec-
essary for substrate conversion in the case of the heterogeneized
catalysts is expectable considering the morphology of the kaoli-
nite matrix, which may hinder the access of the reagents to the
active catalytic sites. In any case, it can be said that the performance
3.2. Oxidation of cis-cyclooctene and cyclohexane by hydrogen
peroxide
For various reasons, cis-cyclooctene is often used as a diagnos-
tic substrate whenever a new catalytic system is synthesized. First,
the high stability of cis-cyclooctenoxide, its main oxidation prod-
uct, enables easy evaluation of the catalyst efficiency [50]. Besides,
epoxides are very useful intermediates in the chemical industry,
since they are the starting point for the preparation of a wide variety
of products.
On the other hand, cyclohexane is relatively inert and its oxida-
tion is well reported in the literature, thereby enabling comparison
to bibliographic data. This substrate also makes possible the eval-
uation of the catalyst selectivity because cyclohexanol and/or
cyclohexanone can be obtained as oxidation products [50–53]. The
industrial importance of the oxidation of cyclohexane is great; as it
leads into a mixture of cyclohexanol and cyclohexanone – called K-
A oil – employed in the synthesis of textile fibers, such as nylon-6
and nylon-6.6 [51]. This reaction is carried out at industrial level
employing soluble Co(II) salts as catalysts, molecular oxygen as

E.H. de Faria et al. / Catalysis Today 187 (2012) 135–149
143
Table 3
Conversion of cis-cyclooctene to cyclooctenoxide catalyzed by Ka-pa and ka-dpa-containing systems.a
Catalyst
2 h (%)
TON
4 h (%)
TON
24 h (%)
TON
48 h (%)
TON
Ka-pa
–
–
–
–
–
Fe(pa)3
7.0
4.0
4.0
7.0
–
2349
134
134
234
7.0
5.0
4.0
10.0
–
2349
168
134
335
25.0
22.0
22.0
35.0
–
8391
741
741
36.0
29.0
34.0
37.0
–
12083
977
1141
1241
Fe(Ka-pa)-1
Fe(Ka-pa)-2
Fe(Ka-pa)-3
Ka-dpa
1174
Fe(dpa)3
4.0
4.0
4.0
4.0
914
69
69
4.0
4.0
4.0
4.0
914
69
69
7.0
9.0
9.0
9.0
1600
157
157
157
11.0
12.0
14.0
15.0
2515
209
243
342
Fe(Ka-dpa)-1
Fe(Ka-dpa)-2
Fe(Ka-dpa)-3
69
69
a
Conditions: molar ratio of catalyst/oxidant/cis-cyclooctene was 1:300:100; solvent mixture acetonitrile/dichloroethane (ACN/DCE, 1:1, v/v). Homogeneous catalysts
Fe(pa)3 and Fe(dpa)3 were obtained under the same conditions as the heterogeneous catalysts. All the reactions were conducted at 25 ◦C and ambient pressure. Turnover
Number (TON) = sum of products conversion (mol)/amount of “active component” (mol).
of the Fe(III)–picolinate grafted catalysts is good. It can again be
remarked that all the reactions involved herein proceed by a typical
heterogeneous catalysis mechanism, since Fe–picolinate and dipi-
colinate complexes did not leach from the support during the entire
process.
The consumption of hydrogen peroxide was investigated after
the catalysis oxidation reactions. For all reactions, when Fe(Ka-pa)-
n and Fe(Ka-dpa)-n were employed as heterogeneous catalysts the
hydrogen peroxide was fully consumed during the reaction, this
gives good evidences that the Fe(III) picolinates and dipicolinates
grafted on kaolinite are able to transfer the oxygen to the sub-
strate with high efficiency. The catalysts were successfully reused
in the oxidation reaction using hydrogen peroxide as oxidant. The
catalysts were separated from the reaction mixture after each
experiment by simple filtration, washed and dried before being
used in the subsequent runs. Good conversions were maintained
for the first five reuses.
As for cyclohexane oxidation, all heterogeneous catalysts based
on Ka-dpa precursor were more selective for cyclohexanone for-
mation compared to the homogeneous reaction, however all the
catalysts based on Ka-pa present low selectivity for ketone or alco-
hol. This fact can be explained by the morphology of these solids,
the easy accessibility to their active sites promote the fast oxida-
tion of cyclohexane and the diffusion of this product to the reaction
medium. But in the specific case of solids based on Ka-dpa the dif-
ficult accessibility and release of products from the active sites
promote the higher selectivity to the ketone, the cyclohexanol
obtained is reoxidized by hydrogen peroxide and only ketone is
released from Fe-(Ka-dpa)-n matrix, as seen from the results pre-
sented in Table 4.
the substrate (cyclohexane) favors its oxidation (to cyclohexanol)
and re-oxidation (to cyclohexanone), as previously reported by
Nakagaki et al. for ironporphyrins immobilized on sylanized
halloysite nanotubes [37–39]. It is generally expected that iron-
porphyrins functioning as biomimetic models of cytochrome
P-450 should lead to higher yields of cyclohexanol than of
cyclohexanone. In fact, high selectivity for the alcohol product
has been observed by several authors. However, the fact that
Fe(III)–dipicolinate immobilized on kaolinite was more selective
for cyclohexanone (100%) shows that these materials grafted
on kaolinite outperformed the catalysts currently employed by
the industry, not to mention that all the reactions carried out
here were accomplished in mild conditions (ambient tempera-
ture and pressure, environmentally friendly oxidant). Experiments
conducted under similar conditions, but in the absence of cata-
lysts and with materials that were not complexed with Fe(III),
namely Ka-pa and Ka-dpa, did not give any oxidation products,
proving that Fe(III) plays a crucial role in the catalytic oxidation
reactions [57–59].
The selectivity achieved with the catalysts Fe(Ka-dpa)-n is
higher than those reported by Esmelindro et al. [19], who used a
binuclear iron complex as catalyst in the oxidation of cyclohexane
by hydrogen peroxide. These authors observed a conversion of 19%
of cyclohexane, but the selectivity cyclohexanol/cyclohexanone
was close to 2:1 (12.65% of cyclohexanol and 6.59% of cyclohex-
anone). In this homogeneous system, the binuclear Fe(III) complex
acted as methanemonooxygenase model, with high selectivity for
the alcohol. However, no data on the reuse of the catalyst were
reported for this system [60]. Again in this case, the reaction was
entirely heterogeneous; no leaching of Fe–picolinate and dipicoli-
nate complexes was observed.
Our present results in the oxidation of cis-cyclooctene and
cyclohexane demonstrate the catalytic potential of iron complexes
supported on clays, mainly regarding the high selectivity toward
cyclohexanone. These data are even more interesting given that
selectivity for cyclohexanol is usually found, as in the case of the
industrial process for production of adipic acid and ␧-caprolactam.
Moreover, the Fe–picolinate–clay system is advantageous because
it is heterogeneous, much cheaper than other systems, the catalysts
are synthesized in mild conditions – low temperature and ambient
pressure – and they are active in catalytic oxidation reactions using
an eco-friendly oxidant as H2O2.
With the aim of investigating the influence of temperature in
the catalytic oxidation and of optimizing this process, the best
catalysts (Fe(Ka-pa)-3 and Fe(Ka-dpa)-3) were employed in both
oxidation reactions at 55 ◦C, as it has been reported that in some
cases higher temperatures facilitate the diffusion of the substrate to
the active catalytic sites, thereby enhancing conversion of the reac-
tants. The catalyst/oxidant/substrate molar ratio was maintained at
1:300:100, the same used at ambient temperature.
In particular, the catalysts Fe(Ka-dpa)-1, Fe(Ka-dpa)-2, and
Fe(Ka-dpa)-3 were the most selective for cyclohexanone forma-
tion, furnishing ketone yields of 5, 5, and 7% after 48 h, respectively
(ambient temperature and pressure). The formation of cyclohex-
anone in higher proportion can be explained by the amount of
complex present in the basal space of the clay, which must resem-
ble the surroundings of the metal ion in the biomimetic Fe model,
thereby giving rise to site-isolation in the case of the grafted com-
plexes. In the specific case of Fe(Ka-dpa)-n, the basal space of
˚
the catalysts is smaller (11.4 A), not to mention the existence of
˚
phases with basal distances of 8.4, 6.3 and 5.5 A, assigned to the
presence of “halloysite-like” structures with internal diameter of
30 nm. These tubular phases with shorter basal distance are prob-
ably the reason for the higher selectivity of these catalysts. The
difficult release of products from the active catalytic sites certainly
causes further oxidation of the initially formed cyclohexanol to
cyclohexanone. It is important to bear in mind that the empty
spaces inside the kaolinite nanotubes are occupied by hydrophobic
organic molecules, as picolinates, carrying complexes displaying
catalytic activity, and the restricted movement and diffusion of

144
E.H. de Faria et al. / Catalysis Today 187 (2012) 135–149
Fig. 8. Effect of hydroquinone on conversion of cis-cyclooctene to cyclooctenoxide.
In the case of Fe(Ka-pa)-3, its activity at 55 ◦C was lower
than at RT, probably due to a partial leaching of Fe(III)–picolinate
from the interlayer space of the clay. According to Shul‘pin et al.,
simple salts have lower activity than complexes, and in the case of
homogeneous catalysis the activity of the complexes is much lower
compared to that of the supported catalysts [57]. On the other hand,
in presence of hydrogen peroxide, the higher temperature might
provoke the decomposition of the complexes immobilized in the
interlayer space of kaolinite. However, in the case of Fe(Ka-dpa)-3,
the increase of the reaction temperature to 55 ◦C promoted higher
substrate conversion. The fact that dipicolinate anion has two car-
boxylic groups gives rise to a more stable material than in the case
of the picolinate, as these groups are covalently bound to the clay
via two hydroxyl groups, what justifies a lower leaching degree.
Specifically for the case of cyclohexane oxidation, the results prove
that the higher temperature enhances the diffusion of the substrate,
thereby increasing cyclohexanone yields (14%). The higher selectiv-
ity for cyclohexanone (100%) compared to reactions conducted at
ambient temperature was attributed to the presence in this cata-
lyst of a tubular phase of kaolinite (halloysite-like), as confirmed
by XRD and TEM analyses.
To study the mechanisms involved in the formation of the
products during cis-cyclooctene and cyclohexane oxidation, all the
catalysts were studied again in similar conditions, but in the pres-
ence of the radical scavenger hydroquinone. Fig. 8 displays the
conversion observed for cis-cyclooctene oxidation reaction in the
presence of this radical scavenger. These data strongly suggest that
a nonradical pathway is responsible for the selective oxidation reac-
tion. In fact, the presence of radical active species was ruled out for
both substrates, which was corroborated by the absence of byprod-
ucts like cyclohexylhydroperoxide. The oxidation is thus probably
performed via a non-radical mechanism involving high-valent Fe-
oxo species.
The efficiency and stability of Fe–picolinate–clay system was
analyzed by evaluation of its reuse. For this purpose, the solid
catalysts were separated from the reaction mixture after each
experiment by simple filtration and dried before being used in
a subsequent run. In the case of the samples Fe-(Ka-pa)-n and
Fe(Ka-dpa)-n, the catalysts were reused in five consecutive runs
without any decrease in activity. On the other hand, to determine
if catalysis is genuinely heterogeneous the catalysts Fe(Ka-pa)-n
and Fe(Ka-dpa)-n were filtered off the reaction mixture after a cat-
alytic cycle, an extra amount of oxidant was added to the resulting
supernatant, and the oxidation reaction was allowed to proceed
under the same initial conditions for 72 h. After this period, no
cis-cyclooctene oxide yields were detected, indicating that the cat-
alytic activity of the solid catalysts is truly heterogeneous and also
demonstrating that the Fe(III)–pyridine carboxylic complexes play
an essential role in the catalysis.

E.H. de Faria et al. / Catalysis Today 187 (2012) 135–149
145
Table 5
Cyclohexane oxidation by PhIO using the indicated catalysts and after different reactions times.a
Catalysts
2 h
OL
4 h
OL
24 h
48 h
OL
ONE
ONE
OL
–
ONE
ONE
Ka-pa
Fe(Ka-pa)-3
–
–
–
16.0
–
–
–
14.5
–
16.0
–
–
–
17.0
Ka-dpa
Fe(Ka-dpa)-3
–
–
–
10.0
–
–
–
9.5
–
–
–
10.0
–
–
–
11.0
a
Conditions: catalyst/oxidant/cyclohexane = 10 mg catalyst, 5 mg PhIO, 150 ␮L cyclohexane; solvent mixture acetonitrile/dichloroethane (ACN/DCE, 1:1, v/v). All the
reactions were conducted at 25 ◦C and ambient pressure. % = product yield.
The activity of heterogeneous catalysis is kinetically controlled
by two principal factors, i.e., the number of active sites on catalyst
surface and the turnover number (TON) of reaction on each active
site. Homogeneous catalysts generally have high turnover numbers
due the easiness of diffusion of substrate and products, while solid
catalysts suffer some irreversible chemical changes in their surfaces
affecting the reaction mechanism and the diffusion and access to
active sites. Thus, picolinate and dipicolinate complexes in homo-
geneous medium shows high TON values (see Tables 3, 4 and 7),
while the lower TON in heterogeneous catalysts can be explained
due to the difficult access to active sites located in the interlayer
space of kaolinite. However, it may be emphasized that these het-
erogeneous catalysts based on kaolinite can be easily reutilized by
several times with constant catalytic activity.
functionalized part of the clay faces the interior part of the formed
tubes, thus the active sites are more protected and the access of the
substrate and the oxidant to these sites more restricted, thereby
becoming a limiting factor for their reactivity.
The results obtained for the oxidation of cyclohexane are shown
in Table 5. The precursors Ka-pa and Ka-dpa were again not active
for this reaction, and the Fe-containing solids showed relatively
high cyclohexane conversions, between 9.5 and 17.0%. Various
aspects may be remarked in these results. First of all, it is very sig-
nificant that selectivity was complete for cyclohexanone for both
catalysts, cyclohexanol was not detected. Also in both cases, the
conversion is almost constant with time; in fact, high conversions
are observed for the shorter time considered (2 h), slightly decreas-
ing after 4 h (confirming the tendency observed in the oxidation of
cyclooctene), increasing a little for higher reaction times. This again
confirms that the catalytically active sites are easily accessible to
the reactant and the oxidant. Finally, conversion is again signifi-
cantly higher when using Fe(Ka-pa)-3 catalysts than when using
Fe(Ka-dpa)-3.
Cyclohexanone can be formed in a single step, by direct
oxidation of cyclohexane, or in two steps, by the oxidation of cyclo-
hexanol previously formed by oxidation of cyclohexane. Trying to
gain information on the catalytic activity of the solids, cyclohexanol
was used as substrate, being oxidized under the same conditions.
The results obtained are shown in Table 6. Both Fe(Ka-pa)-3 and
Fe(Ka-dpa)-3 catalysts are really able to oxidize cyclohexanol to
cyclohexanone. However, in the case of the matrix based on the
Ka-pa precursor, with much larger basal spacing, cyclohexanol does
not remain in contact with the active site for long, being rapidly
released into solution, while for the Ka-dpa series solids, the kaoli-
nite tubes present in their structure make cyclohexanol diffusion
to the solution slower, allowing the oxidation of the alcohol to the
ketone. This is more clearly evident when reactions using H2O2
as oxidants are analyzed; cyclohexanol and cyclohexanone are
formed. Nevertheless, catalysts based on the Ka-dpa precursor give
selectivities of up to 87% for cyclohexanone when the reaction is
performed at 60 ◦C, and 100% selectivity is achieved in the reactions
carried out at ambient temperature and using PhIO and/or H2O2. No
conversion was observed when the solids Ka-pa and Ka-dpa were
employed in the same conditions.
3.3. Oxidation of cis-cyclooctene and cyclohexane by PhIO
Iodosylbenzene has various interesting peculiarities as oxidant:
it gives good oxidant conversion; it is relatively inert in the absence
of iron complexes, forming with them high-valent ferryl radicals
FeIV(O)+• and PhI; it is a polymeric solid that does not contain
weak O H bonds, thus eliminating the occurrence of free radi-
cal chain reactions normally initiated by oxidants such as alkyl
hydroperoxides, R
O O H [60]. By these reasons, the oxidation
of cis-cyclooctene, cyclohexane, and cyclohexanol with this oxy-
gen source was investigated, catalyzed by the precursors Ka-pa and
Ka-dpa and the catalysts Fe(Ka-pa)-3 and Fe(Ka-dpa)-3.
The oxidation of cis-cyclooctene was in all cases null when using
the precursors, even when the reaction was carried out for 48 h, and
in spite of the small amount of Fe in the parent kaolinite. In opposite,
the Fe-containing solids Fe(Ka-pa)-3 and Fe(Ka-dpa)-3 catalyzed
quantitatively the reaction in all the conditions considered, even
after only 2 h of reaction. It is noteworthy that the heterogeneous
systems furnished higher yields compared to the parent hetero-
geneous catalysts (without iron). The fact that the conversion is
very high from short reaction times is a very promising result,
since one of the usual disadvantages of heterogeneous catalysts
is the fact that they require long reaction times, because the matrix
hinders the access of the substrate to the active catalytic sites and
slows down the diffusion of the products. Active sites in our grafted
kaolinite seem to be easily available, probably because of the lay-
ered structure and relatively high interlayer height of both catalytic
systems.
It has been reported that the addition of aminoacids to cat-
alytic systems composed of metal catalysts considerably increases
It is important to highlight that catalysts based on the Ka-dpa
precursor gave lower product yields compared to the catalysts
based on Ka-pa. This may be justified by the more difficult access
of the reactants to the active sites in the former materials, and by
the smaller quantity of Fe(III) complex immobilized in their inter-
layer space. Indeed, Fe-(Ka-pa)-n and Fe(Ka-dpa)-n catalysts show
distinct morphologies. In the particular case of Ka-dpa, the mate-
rials obtained after complexation with Fe(III) in aqueous media
presented tubular morphology; the tubular nature of the Fe(Ka-
dpa)-n catalysts might contribute to their selectivity, once the
Table 6
Oxidation of cyclohexanol to cyclohexanone by PhIO catalyzed by using the indi-
cated catalysts and after different reactions times.a
Catalysts
2 h
4 h
24 h
48 h
Ka-pa
Fe(Ka-pa)-3
–
17.0
–
48.0
–
90.0
–
100
Ka-dpa
Fe(Ka-dpa)-3
–
16.0
–
18.0
–
75.0
–
75.0
a
Conditions: the same as in Table 5.

146
E.H. de Faria et al. / Catalysis Today 187 (2012) 135–149
alkane conversion in the case of cyclohexane oxidation by H2O2, the
aminoacids seem to play the role of co-catalysts. In our system, the
picolinic acids seem to play a similar role, which can be attributed
to the fact that their structure, because of the vicinity between the
amino and the acid groups, is in a certain way similar to that of the
aminoacids.
Shul‘pin et al. [57] described that the catalytic activity of a dinu-
clear iron complex in the oxidation of cyclohexane was improved by
using H2O2 as oxidant, while according to Nizova et al. [61], this oxi-
dation occurs by a radical mechanism that involves initial reduction
of a Fe(III) cation by H2O2 and consequent formation of hydroper-
oxy radicals and Fe(II); the role of the aminoacid is to facilitate the
proton transfer between the coordinated H2O2 and the iron ion,
thereby producing the Fe
O OH fragments [62]. In our catalytic
system, the use of a radical trap (hydroquinone) did not modify the
results, suggesting the occurrence of a non-radical mechanism of
oxidation. However, when triphenylphosphine (PPh3) was added
to the reaction medium, the conversion of cyclohexanol to cyclo-
hexanone was higher (Table 7). In the case of a radical mechanism,
formation of both cyclohexanol and cyclohexanone increases after
addition of PPh3. This has been reported to occur in the GC injector
through transformation of alkyl hydroperoxides, which are reduced
by PPh3, to give the alcohol [62].
On the basis of the results obtained with hydroquinone and PPh3
a mechanism for the catalytic activity of Fe-picolinate-kaolinite
catalysts can be proposed (Fig. 9). This mechanism is similar
to that reported by Gozzo [62], who detected the reduction of
the Fe(III)–pyridine carboxylate complex to the corresponding
Fe(II)–pyridine carboxylate complex in the presence of excess
H2O2. These authors reported that the Fe(II) complex subsequently
reacted with H2O2 and, in the presence of pyridine from the amino
acid residue, a non-radical mechanism then took place. In this
non-radical mechanism, the active species responsible for oxy-
gen transfer to the substrate is HO FeIV
O OH, which leads to
the formation of alkyl hydroperoxide, the latter of which produces
cyclohexanol. In the presence of PPh3, the hydroperoxide is reduced
to the alcohol, which accounts for the occurrence of the non-radical
mechanism, as previously discussed by Gozzo in the case of the
GIF-catalysts mechanism involving Fe(II)–bispicolinate complexes
[62]. Pyridine derivatives are believed to be essential for the system
as a trap for hydroxyl radical, thereby preventing Fenton chemistry.
3.4. Baeyer–Villiger oxidation reaction of cyclohexanone
Table 8 displays the results from the Baeyer–Villiger reaction
carried out for oxidation of cyclohexanone to ␰-caprolactone, using
Fe(Ka-pa)-3 and/or Fe(Ka-dpa)-3 as catalysts. The reaction was
carried out under various different conditions, and the following
conclusions are found:
•
Presence of benzonitrile: it is necessary for the reaction, in the
absence of this solvent, there was no ␰-caprolactone formation.
This effect has already been observed previously, and it is pro-
posed that benzonitrile participates in the transference of oxygen
[6,34–38,63].
•
Temperature: it is a decisive factor. When the reactions were
conducted at ambient temperature, the conversions were lower
than 2% after 24 h of reaction while, as observed in Table 8, much
higher conversions are obtained when the reaction is carried out
at 60 ◦C.
•
Conversions of 45 and 60% were obtained after 24 h of reaction.
The selectivity was complete for ␰-caprolactone, no other prod-
ucts were detected.
Presence of Fe(III) complexes: they are the true catalysts, when
the reaction was carried out in the absence of any catalyst, or in
•

E.H. de Faria et al. / Catalysis Today 187 (2012) 135–149
147
H
O
Fe3+(Ka-pa)
H+
Fe2+(Ka-pa)
O
+
O
O
+
H
H
H
O
H
O
O
O
Fe4+(Ka-pa)
O
H
O
O
H
O
H
Fe2+(Ka-pa)
Fe2+(Ka-pa)
OH
+ pa
+
O
H
H
O
H
H
O
H
PPh3
H
O
H
O
Fe2+(Ka-pa)
O
O
Fe4+(Ka-pa)
cyclohexylhydroperoxide
O
Fig. 9. Possible mechanism for the oxidation of cyclohexane (Adapted from Gozzo mechanism for GIF-catalysts).
the presence of the solids non-containing iron, Ka-pa and Ka-dpa,
very low substrate conversions were observed (less than 1%).
to those carried out in aqueous medium [24,34–38]. The layered
structure of the support resulted in good maintenance of its prop-
erties. Even under various catalytic cycles, problems related to the
diffusion of the reactants to the iron–picolinate complexes, active
sites or of the products out of the catalyst were not observed, in
contrast with what is observed for other supports, such as silica
and alumina.
It has been proposed that the Baeyer–Villiger reaction con-
sists of two steps: the first one comprises a nucleophilic attack
on the carbonyl, with formation of the Criegee adduct, and the
second one involves a concerted rearrangement of the interme-
diate, thereby generating an ester and a carboxylic acid [6,63].
Water is formed as byproduct when H2O2 is employed as oxidant,
while a carboxylic acid is formed if using a peracid as oxidant. A
detailed mechanism for the Baeyer–Villiger oxidation catalyzed by
The reaction was carried out using two different hydrogen
peroxide concentrations (30 and 70 wt.%), this implies different
amounts of water in the reaction medium. For both catalysts, the
best results were achieved with 70 wt.% H2O2. In some cases, the
presence of water can deactivate a catalyst, since water molecules
can coordinate to its active sites. In this sense, further investigations
of the hydrophilicity of Fe-picolinate-kaolinite system are neces-
sary, because this is an essential parameter when H2O2 is employed
in aqueous medium. Steffen et al. [64], studying the Baeyer–Villiger
reaction catalyzed by commercial alumina or alumina prepared by
the sol–gel process, have reported that the reactions accomplished
with H2O2 in anhydrous medium furnished higher yields compared
1st step:
H
O
H
N
HO
NH
O
O
O
-
H
Fe3+(Ka-pa)
Fe3+(Ka-pa)
Fe3+(Ka-pa)
step:
2nd
HO
NH
HO
NH
-
O
HN
O
O
O
Fe3+(Ka-pa)
Fe3+(Ka-pa)
+
O
Fig. 10. Possible mechanism for the Baeyer–Villiger oxidation in the presence of benzonitrile (Adapted from Ruiz [6]).

148
E.H. de Faria et al. / Catalysis Today 187 (2012) 135–149
Table 8
Amount of ␰-caprolactone obtained (%) by Baeyer–Villiger oxidation of cyclohexanone by 30 and 60 wt.% H2O2 using the indicated catalysts.a
Oxidant
2 h
4 h
24 h
30% H2O2
60% H2O2
30% H2O2
60% H2O2
30% H2O2
60% H2O2
Catalyst
Ka-pa
Fe(pa)3
Fe(Ka-pa)-3
Fe(Ka-pa)-3 A.T.b
Fe(Ka-pa)-3 WBzc
–
38.0
21.0
–
–
55.0
38.0
–
–
59.0
27.0
–
–
65.0
43.0
–
–
68.0
33.0
–
–
79.0
60.0
–
–
–
–
–
–
–
Ka-dpa
Fe(dpa)3
Fe(Ka-dpa)-3
Fe(Ka-dpa)-3 A.T.b
Fe(Ka-pa)-3 WBz.c
–
22.0
12.0
–
–
33.0
31.0
–
–
28.0
14.0
–
–
53.0
39.0
–
–
40.0
24.0
–
–
68.0
45.0
–
–
–
–
–
–
–
a
Conditions: catalyst/oxidant/cyclohexane = 20 mg catalyst, 176 ␮L H2O2, 850 ␮L benzonitrile, acetonitrile:dichloroethane: ACN/DCE (1:1, v/v). All the reactions were
conducted at 60 ◦C and ambient pressure. The stoichiometric ratio (number of moles) of peroxide was maintained; only the concentration of water in the system was varied.
b
Reactions were conducted at ambient temperature and pressure.
Reactions were conducted without benzonitrile, which was substituted by a mixture ACN/DCE (1:1, v/v).
c
iron-picolinate-kaolinite catalysts is proposed in Fig. 10. In the case
of the electrophilic activation of the substrate, the acid sites in
iron-picolinate-kaolinite catalysts, as well as electron withdrawing
substituents, are able to activate the carbonyl group in cyclohex-
anone, thereby enhancing the polarizability of the C O double
bond, which in turn facilitates the nucleophilic attack of H2O2. The
peroxide thus bonded to Fe cations is able to oxidize benzonitrile
to peroxycarboximidic acid intermediate, which in turn can inter-
act with cyclohexanone forming the Criegee adduct and allowing
the oxidation to caprolactone. In this sense, one of the factors that
make a catalyst efficient for the Baeyer–Villiger reaction is its abil-
ity to increase the nucleophilicity of the peroxide, this activation
promotes higher affinity of the oxidant for electron deficient cen-
ters [24,34–38], as is the case of the carbonyl group immobilized
by its interaction with the picolinate-kaolinite system.
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.cattod.2011.11.029.
References
[1] Ullmann’s Encyclopedia of Industrial Chemistry, 7th edition, Wiley-VCH Verlag,
Weinheim, 2005.
[2] G. Centi, S. Peranthoner, Microporous Mesoporous Mater. 107 (2008) 3–15.
[3] Z.L. Lu, E. Lindner, H.A. Mayer, Chem. Rev. 102 (2002) 3543–3578.
[4] M. Álvaro, B. Ferrer, H. Garcia, A. Sanjuán, Tetrahedron 55 (1999) 11895–11902.
[5] R.D. Dewhurst, A.F. Hill, M.K. Smith, Angew. Chem. Int. Ed. 43 (2004) 476–478.
[6] J.R. Ruiz, C. Jiménez-Sanchidrián, R. Llamas, Tetrahedron 62 (2007)
11697–11703.
[7] R.A. Reziq, D. Avnir, I. Miloslavski, H. Schumann, J. Blum, J. Mol. Catal. A: Chem.
185 (2002) 179–185.
[8] A.P.M. Wight, E. Davis, Chem. Rev. 102 (2002) 3589–3614.
[9] Z.L. Lu, E. Lindner, H. Mayer, Chem. Rev. 102 (2002) 3543–3578.
[10] A. Godelitsas, D. Charistos, C. Tsipis, P. Misaelides, A. Filippidis, M. Schindler,
Microporous Mesoporous Mater. 61 (2003) 69–77.
[11] M. Hartmann, L. Kevan, Chem. Rev. 99 (1999) 635–663.
[12] R. Bechara, D. Balloy, J. Dauphin, J. Grimblot, Chem. Mater. 11 (1999)
1703–1711.
[13] J.M. Thomas, R. Raja, Catal. Today 117 (2006) 22–31.
[14] L.R.D. da Silva, L.C. Garla, Quim. Nova 2 (1998) 169–174.
[15] A.N. Pour, Y. Zamani, A. Tavasoli, S.M.K. Shahri, S.A. Taheri, Fuel 87 (2008)
2004–2012.
[16] V.H. Deshpande, D.E. Ponde, V.J. Bulbule, A. Sudalai, A.S. Gajare, J. Org. Chem.
63 (1998) 1058–1063.
[17] S. Letaïef, B. Casal, P. Aranda, M.A. Martín-Luengo, E. Ruiz-Hitzky, Appl. Clay Sci.
22 (2003) 263–277.
[18] S. Nakagaki, F.L. Benedito, F. Wypych, J. Mol. Catal. A: Chem. 217 (2004)
121–131.
[19] O.A.C. Antunes, M.C. Esmelindro, E.G. Oestreicher, H. Marquéz-Alvarez, C.
Dariva, S.M.S. Egues, C. Fernandes, A.J. Bortoluzzi, V. Drago, J. Inorg. Biochem.
99 (2005) 2054–2061.
[20] K.C. Gupta, A.K. Sutar, Coord. Chem. Rev. 252 (2008) 1420–1450.
[21] J.E. Gardolinsky, G. Lagaly, Clay Miner. 40 (2005) 547–566.
[22] J. Murakami, T. Itagaki, K. Kuroda, Solid State Ionics 172 (2004) 279–282.
[23] P.T. Anastas, M.M. Kirchhoff, T.C.A. Williamsom, Catalysis A 221 (2001) 3–13.
[24] R.L. Frost, J. Kristof, E. Horvat, J.T. Kloprogge, Spectrochim. Acta. A 56 (2000)
1191–1204.
4. Conclusions
Kaolinite covalently grafted with picolinate and dipicolinate
anions proved to be viable supports for the effective immobiliza-
tion of iron by forming complexes. The catalytic results obtained for
the complexes covalently grafted into the basal space of kaolinite
demonstrated that these catalysts perform well in cis-cyclooctene
and cyclohexane oxidation, their activities are similar to those dis-
played by the corresponding homogeneous complexes. In the case
of cyclohexane oxidation, the catalysts Fe(Ka-dpa)-n are efficient
and highly selective, which is attributed to the presence of a kaoli-
nite tubular phase. Compared to the parent homogeneous catalysts,
the heterogeneous catalysts were advantageous with respect to
cis-cyclooctene and cyclohexane oxidation reactions since they
allowed for the use of mild conditions (temperature and ambient
pressure, use of non-polluting oxidants) and promoted high activity
and product selectivity. The main advantage of the kaolinite grafted
complexes is their easy separation from the reaction mixture by
simple filtration of the solid, thus enabling catalyst reuse.
[25] N. Bizaia, E.H. de Faria, G.P. Ricci, P.S. Calefi, E.J. Nassar, K.A.D.F. Castro, S.
Nakagaki, K.J. Ciuffi, R. Trujillano, M.A. Vicente, A. Gil, S.A. Korili, Appl. Mater.
Interface 11 (2009) 2667–2678.
Acknowledgments
[26] P. Gómez-Romero, C. Sanchez, New J. Chem. 29 (2005) 57–58.
[27] E.H. de Faria, O.J. Lima, K.J. Ciuffi, E.J. Nassar, M.A. Vicente, R. Trujillano, P.S.
Calefi, J. Colloid Interface Sci. 335 (2009) 210–215.
[28] L.R. Ávila, E.H. de Faria, K.J. Ciuffi, E.J. Nassar, P.S. Calefi, M.A. Vicente, R. Trujil-
lano, J. Colloid Interface Sci. 341 (2010) 186–193.
[29] E.H. de Faria, K.J. Ciuffi, E.J. Nassar, M.A. Vicente, R. Trujillano, P.S. Calefi, Appl.
Clay Sci. 48 (2010) 516–521.
[30] S.L. Jain, P. Bhattacharyya, Inorg. Chim. Acta 359 (2006) 4398–4402.
[31] A. Goti, F. Cardona, Green Chem. React. (2008) 191–212.
[32] S. Nakagaki, F. Wypych, J. Colloid Interface Sci. 315 (2007) 142–157.
Spanish authors thank financial support from Spanish Ministry
of Science and Innovation (MICINN) and the European Regional
Development Fund (FEDER) through projects MAT2010-21177-
C02, and Junta de Castilla y León (SA009A11-2). The Brazilian
group thanks support from Brazilian Research funding agencies
FAPESP and CNPq, and Peróxidos do Brasil (Solvay) for supplying
the 70 wt.% aqueous hydrogen peroxide solution.

E.H. de Faria et al. / Catalysis Today 187 (2012) 135–149
149
[33] S. Nakagaki, G.S. Machado, M. Halma, A.A.S. Marangon, K.A.D.F. Castro, N. Mat-
[51] J.M. Thomas, R. Raja, G. Sankar, R.G. Bell, Acc. Chem. Res. 34 (2001) 191–200.
toso, F. Wypych, J. Catal. 242 (2006) 110–117.
[34] K. Weissermel, H.-J. Arpe, (C.R. Lindley, Trans.), Industrial Organic Chemistry,
vol. 237, 3rd ed., VCH Publishers, Weinheim, 1997.
[52] R.G. Pearson, J. Am. Chem. Soc. 85 (1963) 3533–3543.
[53] I. Fernandez, J.R. Pedro, A.L. Roselló, R. Ruiz, I. Castro, X. Ottenwaelder, Y. Jour-
naux, Eur. J. Org. Chem. 7 (2001) 1235–1247.
[35] G.-J. ten Brink, I.W.C.E. Arends, R.A. Sheldon, Chem. Rev. 104 (2004) 4105–4123.
[36] Z. Lei, G. Ma, C. Jia, Catal. Commun. 8 (2007) 305–309.
[54] R. Bechara, D. Balloy, J.Y. Dauphin, J. Grimblot, Chem. Mater. 11 (1999)
1703–1711.
[37] Z. Lei, L. Wei, R. Wang, G. Ma, Catal. Commun. 9 (2008) 2467–2469.
[38] S.E. Jacobson, R. Tang, F.J. Mares, Chem. Commun. 12 (1978) 888–889.
[39] V. Amani, N. Safari, H.R. Khavasi, P. Mirzei, Polyhedron 26 (2007) 4908–4914.
[40] J.G. Sharefkin, H. Saltzmann, Org. Synth. 43 (1963) 62–64.
[41] H.J. Lucas, E.R. Kennedy, M.W. Forno, Org. Synth. 43 (1963) 483.
[42] S. Brunauer, P.H. Emmett, E. Teller, J. Am. Chem. Soc. 60 (1938) 309–319.
[43] J.E.F.C. Gardolinski, G. Lagaly, Clay Miner. 40 (2005) 547–556.
[44] V. Umamaheswari, W. Böhlman, A. Pöppl, A. Vinu, M. Hartmann, Microporous
Mesoporous Mater. 89 (2006) 47–57.
[55] U. Schuchardt, W. Carvalho, E.V. Spinacé, Synlett 10 (1993) 713–718.
[56] U. Schuchardt, D. Cardoso, R. Sercheli, R. Pereira, R.S. Da Cruz, M.C. Guerreiro,
D. Mandelli, E.V. Spinacé, E.L. Pires, Appl. Catal. A 211 (2001) 1–17.
[57] G.B. Shul‘pin, C.C. Golfeto, G. Süss-Fink, L.S. Shul‘pina, D. Mandelli, Tetrahedron
Lett. 46 (2005) 4563–4567.
[58] G. Dubois, A. Murphy, T.D. Stack, Org. Lett. 14 (2003) 2469–2472.
[59] J.G. Carriazo, M.A. Centeno, J.A. Odrizola, S. Moreno, R. Molina, Appl. Catal. A:
Gen. 317 (2007) 120–128.
[60] J.T. Groves, J. Inorg. Biochem. 100 (2006) 434–447.
[61] G. Nizova, B. Krebs, G. Süss-Fink, S. Schindler, L. Westerheide, G.L. Cuervo, B.
Shulˇıpin, Tetrahedron 58 (2002) 9231–9237.
[45] A. Jitianu, M. Crisan, A. Meghea, I. Rau, M. Zaharescu, J. Mater. Chem. 12 (2002)
1401–1407.
[46] J.J. Tunney, C. Detellier, Chem. Mater. 5 (1993) 747–748.
[47] K.B. Brandt, T.A. Elbokl, C. Detellier, J. Mater. Chem. 13 (2003) 2566–2572.
[48] J.J. Tunney, C. Detellier, Can. J. Chem. 75 (1997) 1766–1772.
[49] J.E. Gardolinski, P. Peralta-Zamora, F. Wypych, J. Colloid Interface Sci. 211 (1999)
137–141.
[62] F. Gozzo, J. Mol. Catal. A: Chem. 171 (2001) 1–22.
[63] J.R. Ruiz, C. Jiménez-Sanchidrián, R. Llamas, Tetrahedron 63 (2007)
1435–1439.
[64] A. Steffen, S. Teixeira, J. Sepulveda, R. Rinaldi, U. Schuchardt, J. Mol. Catal. A:
Chem. 287 (2008) 41–44.
[50] K.J. Ciuffi, E.J. Nassar, L.A. Rocha, Z.N. da Rocha, S. Nakagaki, G. Mata, R. Trujillano,
M.A. Vicente, S.A. Korili, A. Gil, Appl. Catal. A 319 (2007) 153–162.