Encapsulation of SALEN- and SALHD-Mn(III) complexes in an Al-pillared clay for bicarbonate-assisted catalytic epoxidation of cyclohexene

Article abstract of DOI:10.1016/j.molcata.2016.01.034

[Mn(3,5-dtSALEN)Cl] (I) and [Mn(3,5-dtSALHD)Cl] (II) complexes (3,5-dtSALEN = N,N′-bis(3,5-di-tert-butylsalicylaldehyde)ethylenediamine; 3,5-dtSALHD = N′N-bis-(3,5-di-tert-butylsalicylaldehyde)-1,2-cyclohexanediamine) were successfully encapsulated within

Full text of DOI:10.1016/j.molcata.2016.01.034

Journal of Molecular Catalysis A: Chemical 416 (2016) 10–19
Contents lists available at ScienceDirect
Journal of Molecular Catalysis A: Chemical
journal homepage: www.elsevier.com/locate/molcata
Encapsulation of SALEN- and SALHD-Mn(III) complexes in an
Al-pillared clay for bicarbonate-assisted catalytic epoxidation of
cyclohexene
Ana M. Garciaa, Viviana Morenob, Sonia X. Delgadoa, Alfonso E. Ramírezb, Luis A. Vargasb
, Miguel Á. Vicentec, Antonio Gild, Luis A. Galeanoa,∗
a Research Group on Functional Materials and Catalysis (GIMFC), Department of Chemistry, Nari˜no University, Calle 18 Cra. 50, Campus Torobajo, Pasto,
Nari˜no, Colombia
b Research Group Catalysis, Department of Chemistry, Cauca University, Sector Tulcán-Edificio Antiguo Liceo, Carrera 3 No. 3N-100, Popayán, Cauca,
Colombia
c Department of Inorganic Chemistry, Salamanca University, Plaza la Merced s/n, E-37008 Salamanca, Spain
d Department of Applied Chemistry, Los Acebos Building, Public University of Navarre, Campus Arrosadía, E-31006 Pamplona, Spain
a r t i c l e i n f o
a b s t r a c t
[Mn(3,5-dtSALEN)Cl] (I) and [Mn(3,5-dtSALHD)Cl] (II) complexes (3,5-dtSALEN = N,Nꢀ-bis(3,5-di-
tert-butylsalicylaldehyde)ethylenediamine; 3,5-dtSALHD = NꢀN-bis-(3,5-di-tert-butylsalicylaldehyde)-
1,2-cyclohexanediamine) were successfully encapsulated within a natural bentonite by using three
preparative approaches: (A) direct adsorption of every metal complex on the previously Al-pillared
bentonite, Al-PILC; (B) two-step liquid phase methodology: (i) cationic adsorption of Mn2+ in Al-PILC by
substituting its residual cationic exchange capacity (CEC), followed by (ii) diffusion of either 3,5-dtSALEN
or 3,5-dtSALHD ligands, for in-situ generation of the corresponding interlayered metal complexes;
and (C) simultaneous pillaring/encapsulation of the complexes on the raw starting clay. The materials
were characterized by cationic exchange capacity, X-ray diffraction, atomic absorption, FT-Infrared and
UV–vis spectroscopies, and N2 adsorption at 77 K. The physical encapsulation of the complexes into final
materials was proven by spectroscopic analyses. Method C yielded both highest metal incorporation and
enhanced basal space on the modified clay. All materials showed to be active catalysts in cyclohexene
epoxidation with hydrogen peroxide using acetonitrile as solvent (0.79 atm, 293 K). Addition of sodium
bicarbonate as co-catalyst led to enhanced conversion (100%) and selectivity (70%) towards the epoxide
in the presence of such a kind of heterogeneized metal-complex catalysts. The catalysts were stable and
reusable along at least two catalytic cycles.
Article history:
Received 17 September 2015
Received in revised form 30 January 2016
Accepted 30 January 2016
Available online 11 February 2016
Keywords:
Catalytic epoxidation
Encapsulated catalysts
Immobilized SALEN–Mn complex
Pillared clays
© 2016 Elsevier B.V. All rights reserved.
1. Introduction
Schiff-base ligands displaying coordination environment N2O2,
also so-called SALEN complexes, have been widely reported as sol-
Epoxidations are a class of reactions deserving of great interest
since they lead to valuable and versatile intermediates for devel-
opment of pharmaceutical active compounds [1], fine chemicals,
polymers and paints [2–5], among others. Hydrogen peroxide con-
stitutes an environmental-friendly and clean oxidizing agent for
this kind of reactions, since it only forms water as by-product,
and by its low price and wide availability [1]; however, its use
in this reaction has been recognized to be complicated because
of its inmiscibility in organic solvents [6]. Mn(III) complexes with
uble and efficient catalysts in olefin epoxidations [7–10].
By this reason, effective immobilization of this type of active
agents on solid surfaces shows several advantages, including sim-
pler handling and separation of the products, easy recovery of
the solid catalyst from the reaction mixture and lower amount
of residues [11,12]. A higher catalytic performance has also been
evidenced for such solid catalysts in comparison to the same
complexes dissolved, operating under homogeneous conditions;
appropriate immobilization allows the isolation of the active cen-
tres, preventing inactive dimers of the metal complex of the type
␮-oxo-Mn(IV) to be formed in the reaction mixture [13]. Thus, het-
erogeneization of such a type of metal complexes has received
much attention in the past few years, the attachment of several
∗
Corresponding author. Tel.: +57 3184079325; Fax: +57 2 7313106.
E-mail address: alejandrogaleano@udenar.edu.co (L.A. Galeano).
http://dx.doi.org/10.1016/j.molcata.2016.01.034
1381-1169/© 2016 Elsevier B.V. All rights reserved.

A.M. Garcia et al. / Journal of Molecular Catalysis A: Chemical 416 (2016) 10–19
11
(a)
N
N
O
N
N
Mn
Cl
Mn(CH3COO)2. 4H2O
t-Bu
HO
t-Bu
O
Bu-t
Bu-t
OH
LiCl
t-Bu
H2(3,5-dtSALEN)
Bu-t
t-Bu
[Mn(3,5-dtSALEN)Cl],
Bu-t
I
(b)
N
O
N
O
N
N
Mn
Cl
Mn(CH3COO)2. 4H2O
LiCl
t-Bu
Bu-t
HO
t-Bu
Bu-t
OH
t-Bu
H2(3,5-dtSALHD)
Bu-t
t-Bu
Bu-t
[Mn(3,5-dtSALHD)Cl],
II
Fig. 1. General sketch employed for preparation of the SALEN–Mn(III) complexes: (a) [NꢀN-bis-(3,5-di-tert-butylsalicylidene)-ethylendiamine]manganese(III) chloride
[Mn(3,5-dtSALEN)Cl)] (I) and (b) [NꢀN-bis-(3,5-di-tert-butylsalicylidene)-1,2-cyclohexanediamine]manganese(III) chloride [Mn(3,5-dtSALHD)Cl] (II).
Mn complexes on several inorganic surfaces, like silica, meso-
porous materials [14–16], activated carbon [17,18] and pillared
clays [19–22] has been then reported. It is remarkable that cata-
lysts based on pillared clays have displayed outstanding efficiency
in the epoxidation of a variety of olefins [23].
Accordingly and taking into account that several methods
have been already tested for encapsulation of SALEN complexes
in porous matrices like pillared clays, this work was focused
to study the direct immobilization of two similar complexes,
([Mn(3,5dtSALEN)Cl)] and [Mn(3,5dtSALHD)Cl], having different
substituents on the diimine bridge (Fig. 1) within the free chan-
nels in the interlayer space of an Al-PILC. Results may be compared
to previously reported methods [19]. The effect of the structure and
size of the complexes and of the insertion method on the stability
and catalytic behaviour of the final functionalized materials have
been evaluated in the cyclohexene epoxidation by H2O2 assisted
by NaHCO3 as co-catalyst.
Pillared clays (PILCs) are nanostructured materials very attrac-
tive to host metal complexes with catalytic activity. In general
terms, these materials are prepared by exchange of the original
cations present into the interlayer space of some natural clays by
voluminous, highly-charged metal-polycations, mainly Al-based,
that upon calcination give rise to pillars of their corresponding
oxides. These interlayered species provide final materials with
increased thermal stability, specific surface area and pore vol-
ume in comparison to the starting mineral [24,25]. PILCs have
been successfully prepared from polycations of several metals,
as Ti, Zr, Fe, and Cr, among others [26], but the clays pil-
lared with Al (Al-PILCs) have been the most extensively studied
[27]; these are prepared by employing the Keggin-like polyca-
2. Experimental
2.1. Materials
The preparation of the Schiff bases and their cor-
responding Mn(III) complexes was performed by using
1,2-cyclohexanediamine (98%), manganese chloride tetrahydrate
(≥98%, ACS reagent), ethylendiamine (≥99.99%, SigmaUltra), man-
ganese acetate tetrahydrate (≥99%), absolute ethanol (≥99.5%, ACS
reagent) and N,Nꢀ-dimethylformamide (≥99%), all purchased from
Sigma–Aldrich®, as well as 3,5-di-tert-butylsalicyladehyde (99%)
from Alfa–Aesar and dichloromethane (99.9%) from Mallinckrodt.
The Al-oligomeric solution used to treat the clay was prepared
from AlCl3·6H2O (≥99%, Sigma–Aldrich) and NaOH (≥99.8%
Mallinckrodt). In the catalytic experiments and the chromato-
graphic analyses, cyclohexene (99.0%, Alfa–Aesar), cyclohexene
epoxide (99.0%, Merck), hydrogen peroxide (30% w/w AR, Panreac),
acetonitrile (99.0%, Aldrich Chemical), sodium bicarbonate (99.0%,
Carlo Erba) and n-decane (≥ 99.0%, Alfa–Aesar) were used. All
reagents were employed as received.
tion [AlO4Al12(OH)24(H2O)12] ]
7+, usually abbreviated as [Al13 7+, as
interlayering species [28,29]. Therefore, the incorporation and sta-
bilization of SALEN Mn-complexes within the interlayer of Al-PILCs
may then efficiently combine in the same catalyst the high cat-
alytic activity displayed by the metal complex with the low cost,
excellent textural properties and high resistance to organic solvents
featured by the Al-PILCs [13,19,20,30,31]. Various strategies have
been assessed in order to confine such a family of metal complexes
in clay materials, finding that the final characteristics of the result-
ing catalysts strongly depend on the experimental procedure used
for the incorporation and stabilization of the metal complex into
the clay host [19].
Using of H2O2 as oxidizing agent in catalytic systems where
Mn is the active metal leads to competition of their decomposi-
tion products for the active centres, driving to deactivation [32].
Besides, it has been also shown that hydrogen peroxide behaves as
a more effective oxidizing species in the presence of activators or
co-catalysts of the reaction, like NaOH, KOH or NaHCO3 [33–35].
Thus, the use of NaHCO3 as co-catalyst has been reported to be cru-
cial for achieving good conversions in the presence of Mn-based
catalysts [34,36].
The preparation of the catalysts was made using as start-
ing material a natural bentonite from Valle del Cauca, Colombia,
(BVC). Its physicochemical and mineralogical features have been
extensively reported elsewhere [37]: cationic exchange capacity,
CEC = 89 meq/100 g clay dry basis; elemental composition (w/w%):
SiO2, 60.5; Al2O3, 24.7; Fe2O3, 10.2; MnO, 0.05; MgO, 3.07; CaO,

12
A.M. Garcia et al. / Journal of Molecular Catalysis A: Chemical 416 (2016) 10–19
Fig. 2. Methods used to incorporate/stabilize complexes (I) or (II) between the layers of the Al-PILC: method A: direct adsorption of the metal complex on the pillared clay;
method B: in-situ growing of the interlayered complex upon adsorption of SALEN/SALHD ligand on the previously Mn2+-exchanged pillared clay; and method C: simultaneous
intercalation of Al13 and encapsulation of the complex on the starting clay.
0.64; Na2O, 0.20 and K2O, 0.58. The pillaring process was performed
on the fraction of the mineral with particle diameter (dp) ≤ 2 ␮m,
obtained by differential sedimentation from a water suspension of
the raw clay mineral.
and then dried at room temperature in a desiccator. The same
procedure was used to prepare both complexes.
2.3. Preparation of the Al-pillared clay
A standard procedure previously reported was used [42]. The
2.2. Synthesis of Schiff bases and Mn(III)-complexes
7+
pillaring precursor [Al13
]
was prepared by very slow dropping
of aqueous NaOH 0.2 mol/dm3 on AlCl3·6H2O 0.2 mol/dm3, to get
a final ratio OH−/Al3+ = 2.4 (final pH around 4.3). Such a resulting
oligomeric solution was then added on a water suspension of the
clay until final nominal loading with 20 mmol Al3+/g clay. The alu-
minosilicate thereby interlayered was then washed in a dialysis
tubing cellulose membrane (Sigma) along four cycles with distilled
water, dried at 333 K and then heated in air, at 673 K for 2 h, to give
rise to the Al-pillared clay (Al-PILC).
Schiff base ligands and their corresponding complexes [NꢀN-bis-
(3,5-di-tert-butylsalicylidene)-ethylendiamine]manganese(III)
chloride [Mn(3,5-dtSALEN)Cl)] (I) and [NꢀN-bis-(3,5-di-tert-
butylsalicylidene)-1,2-cyclohexanediamine]manganese(III)
chloride [Mn(3,5-dtSALHD)Cl] (II) were prepared adopting
a
methodology widely reported in the literature [13,19,38–40].
The ligands were prepared by condensation of 3,5-di-tert-
butylsalicylaldehyde with the corresponding diamine (either,
ethylenediamine or 1,2-cyclohexanediamine) in a molar ratio 1:2
in ethanol. Complexes (I) and (II) were obtained then by adding
a saturated solution prepared with 3.66 mmol of manganese
acetate in absolute ethanol to 1.83 mmol of the corresponding
ligand dissolved in 25 cm3 of the same solvent pre-heated at
353 K under reflux; air was slowly and shortly (10–20 s) bubbled
through the reacting mixture since long oxidation might lead
to form oxo-Mn(V) species easily detected by darkening of the
brownish solution [41], which inactivate the catalytic response
of the complexes in the targeted reaction. Then, the air stream
was stopped and the heating was kept along one extra hour with
darkening of the yellow–orange solution to brown. Afterwards,
5.29 mmol of lithium chloride in ethanol were added to achieve the
access of Cl into the coordination sphere of the transition metal,
and the mixture was heated further 30 min. The resulting solid
was then dissolved in dichloromethane and refined employing
the method from Cubillos et al. [40] by means of liquid extraction
with n-heptane. The final solid was recovered by vacuum filtering
2.4. Incorporation and stabilization of the Mn(III) complexes on
Al-PILC
Every SALEN–Mn(III) complex was encapsulated in Al-PILC by
means of three approaches reported elsewhere [19], summarized
in Fig. 2. The ratio used was always 0.5 mmol complex/g clay.
2.4.1. Method A—direct adsorption of the complex on the
Al-pillared clay
A
solution of the complex in N,Nꢀ-dimethylformamide
(1.0 × 10−3 mol/dm3) was poured on Al-PILC clay and the mixture
heated to reflux (343 K) throughout 12 h under gentle stirring. The
solid was then separated by filtration, washed by centrifugation
using ethanol and then distilled water, pre-dried at 333 K and
finally heated at 393 K for 2 h until final dry state. The so pre-
pared materials were denoted as (IA)Al-PILC and (IIA)Al-PILC for
complexes (I) and (II), respectively.

A.M. Garcia et al. / Journal of Molecular Catalysis A: Chemical 416 (2016) 10–19
13
2.4.2. Method B—in-situ growing of the complex by adsorption of
2.6. Catalytic experiments
the Schiff-base on the Al-pillared clay previously exchanged with
Mn2+
The experiments of the cyclohexene epoxidation were per-
formed using hydrogen peroxide as oxidizing agent in a glass,
jacketed-batch-reactor open to the atmosphere, at room conditions
(0.79 atm and 293 K in average). In a general experiment, the reactor
was charged with 3.0 cm3 of acetonitrile, 1 mmol of cyclohexene,
1 mmol of sodium bicarbonate and 0.02 g catalyst. Dropping of
4–12 mmol of H2O2 was performed along 5 h of reaction under
constant flow rate and stirring, while samples of 0.1 cm3 were
taken every hour until 8 h as final time of reaction. Each sample
was microfiltered (Hydrophobic PTFE, 0.45 ␮m) and then analysed
by gas chromatography in a Shimadzu GC-14A device equipped
with a FID detector and a Varian VF-1 (15 m × 0.25 × 0.25 ␮m) cap-
illary column, using the following program: the temperature of
the column started and was kept at 60 ◦C for 4 min, then raised to
80 ◦C (5 ◦C/min), 4 min, then until 120 ◦C (5 ◦C/min), 1 min, 150 ◦C
(20 ◦C/min), 1 min, and finally to 250 ◦C (20 ◦C/min) during 1 min.
Calibration curves were built in advance for cyclohexene and cyclo-
hexene oxide. The response factors were determined by using the
areas below every peak. Therefore, the concentrations of reactant
and the targeted product were calculated by extrapolation of the
signal of FID detector within the corresponding calibration curve.
Every run was made using a gradient of temperature between 333
and 523 K throughout 29 min, where the temperature of both the
injector and the detector was maintained at 523 K. For assessment
of the catalyst’s reusability, the solid was recovered from the reac-
tion mixture, washed twice with acetonitrile and then five times
with methanol in order to remove any remaining reactant or prod-
uct; every catalyst was then reused in the cyclohexene epoxidation
under identical experimental set-up.
The Al-PILC was treated with a 2.0 mol/dm3 aqueous solution of
MnCl2·4H2O under stirring for 8 h, in order to obtain the Mn2+
-
exchanged form of the Al-pillared clay by cation exchange; the
incorporated Mn content is denoted as Mnincorporated. The super-
natant was retired and the procedure was entirely repeated once
again. Then, the solid was washed by centrifugation with distilled
water along four cycles and dried at 333 K (Mn,Al-PILC). Afterwards,
a 0.01 mol/dm3 solution of the Schiff-base in methanol was added
on a suspension of the solid (1 g Mn,Al-PILC/100 cm3 methanol) in
proper amount for a final ratio Mnincorporated:ligand = 1:1 (equiva-
lent to a ratio 0.6125 mmol ligand/g Mn,Al-PILC), and the mixture
was then refluxed at 323 K for 8 h with stirring. The materials
were washed with dichloromethane by centrifugation to retire the
excess of either ligand or complex deposited on the external sur-
face of the clay, and finally dried at 333 K, giving rise to (IB)Al-PILC
and (IIB)Al-PILC materials for the immobilized complexes (I) and
(II), respectively.
2.4.3. Method C—simultaneous
[Al13]
7+-intercalation/complex-encapsulation on the starting clay
An ethanolic solution of the complex (2 mmol of complex/dm3)
7+
was slowly poured on the [Al13
]
oligomeric solution, reaching
the ratios employed above: 20 mmol Al3+/g clay and 0.5 mmol com-
plex/g clay. The resulting mixed solution was slowly dropped under
strong shaking on a suspension of the clay previously swollen for
24 h in water (2.0% w/v) at room temperature. The resulting suspen-
sion was stirred for 1 h and then allowed to stand for 2 h. The solid
was recovered by centrifugation, washed with distilled water until
final conductivity close to 10 ␮S/cm, and the excess of the complex
removed by two sequential cycles of Soxhlet extraction, the first
cycle with ethanol and the second one with dichloromethane, until
no detection of the complex in the liquid phase. The final solid was
dried at 333 K and heated at 473 K for 2 h from room temperature,
under a heating rate of 1 K/min. Solids were denoted as (IC)Al-PILC
and (IIC)Al-PILC for complexes (I) and (II), respectively.
3. Results and discussion
3.1. Characterization of Schiff-base ligands and their
Mn(III)-complexes
The SALEN-like ligands and derived Mn(III)-complexes were
characterized by UV–vis, FTIR and NMR spectroscopies. The UV–vis
spectra (CH2Cl2) showed characteristic signals in ꢁmax 280 nm and
around 330 nm corresponding to M–L charge transfer transitions
2.5. Physicochemical characterization of catalysts
∗
n → ␲ and ␲ → ␲∗, respectively, the last one due to the pres-
The 200–700 nm UV–vis spectra of the Schiff-bases and their
corresponding Mn(III) complexes were recorded in a Merck UV–vis
Pharo 300 spectrophotometer on dichlorometane 1.0 mmol/dm3
solutions. The 400–4000 cm−1 FT-IR spectra were obtained from
pellets of every sample in KBr (spectral grade, Sigma–Aldrich) in a
Thermo Scientific Nicolet FTIR-6700 spectrophotometer with res-
olution of 4 cm−1, from average of 32 scans. 1H NMR spectra of
the ligands in CDCl3 were made in a 400 MHz Bruker UltrashieldTM
apparatus.
ence of the imino group (
C N) [18]. Besides, both complexes
(CH2Cl2) showed a signal with ꢁmax close to 430 nm, that could
be also attributed to M–L charge transfer, as well as other very
weak signal around 500 nm due to a d–d transition characteristic
in d4 manganese ions [38]. The FTIR spectra of ligands and com-
plexes displayed the expected characteristic signals for this kind
of compounds. Both ligands H2(3,5dtSALEN) and H2(3,5-dtSALHD)
showed, respectively, the following (ꢂmax/cm−1): 3438–3432 (OH,
The content of Mn in the solids was determined by atomic
absorption spectroscopy (AAS) in a PerkinElmer 2380 spectrom-
eter on the samples previously digested in a HF/HNO3 mixture. The
cationic exchange capacity, CEC, of the clay materials was measured
by extensive exchange with 2.0 mol/dm3 ammonium acetate fol-
lowed by several cycles of distilled water washings, drying (333 K)
and micro-Kjeldahl determination of the NH3(g) released from the
(NH4)+-exchanged solids. The XRD patterns were recorded on pow-
dered samples in the range of 2–15◦ 2ꢀ at 2◦ 2ꢀ/min in a Siemens
D-500 Diffractometer, using Ni-filtered Cu K␣ (ꢁ = 1.5405 Å) inci-
dent radiation. Specific surface area of the solid materials was
determined at 77 K in a Quantachrome CHEMBET 3000 equip-
ment; the samples were degassed at 463 K for 2.5 h under a He/N2
70/30 stream. Specific surface area were calculated using the BET
equation (SBET).
ꢂ); 2995–2990 (
ꢂas); 2869–2863 ( CH3, ꢂs), and 1629–1630 (
C
H, ꢂ); 1595–1594 (C C, ꢂ); 2962–2961 ( CH3,
C, ꢂ), the last
N
signal proving the formation of the imine bridge. Both complexes
showed the same signals exhibited by the ligands; however, the
signal of the imine is particularly useful since it became shifted to
lower wavenumbers, 1611 and 1608 cm−1 for complexes (I) and (II)
respectively, in comparison with their corresponding ligands, evi-
dencing the formation of the organometallic adducts. The signals
at 567–546 (Mn–O, ꢂ) and 489–485 (Mn–N, ꢂ) are featured by the
linkage between the metal and the N2O2 coordination sphere. The
characterization of the ligands by NMR revealed the following; (i)
H2(3,5-dtSALEN) 1H NMR (ı ppm): 13.66 (s, 2 H, OH), 8.42 (s, 2 H,
CH N), 7.40 (d, 2 H, ArH), 7.10 (d, 2 H, ArH), 3.95 (m, 2 H, C NCH),
1.47 (s, 18 H, t-butyl). (ii) H2(3,5-dtSALHD) 1H NMR (ı ppm): 13.72
(s, 2 H, OH), 8.32 (s, 2 H, CH N), 7.33 (d, 2 H, ArH), 7.01 (d, 2 H,

14
A.M. Garcia et al. / Journal of Molecular Catalysis A: Chemical 416 (2016) 10–19
(IC)Al-PILC solid at 1616 cm−1 (the other solids showed a broad
peak between 1680 and 1650 cm−1, in which this signal may prob-
ably be overlapped with the bending of adsorbed water, observed in
the starting clay at 1638 cm−1). This suggests that A and B methods
promoted covalent interactions of the complexes with the host-
ing inorganic matrix, due in part to the higher temperature of the
treatment. In the same sense, the antisymmetric stretching from
tetrahedral SiO2 in smectites, observed at 1032 cm−1 for BVC solid,
shifted to higher frequencies, in the range 1060–1100 cm−1, in all
the modified solids, also suggesting host-guest interactions of the
complexes with the clay. In addition, a set of bands can be seen
in the 1500–1200 cm−1 region, similar to those displayed by the
free complexes in this region. Finally, Mn O and Mn N stretchings,
expected at wavenumbers near to 575 cm−1 and 492 cm−1, respec-
tively, unfortunately overlapped with the strong Mg O stretching
of the clay, avoiding the use of these key signals from SALEN for
evaluating the successful immobilization of the complexes inside
the inorganic matrix.
Moreover, the solid (IC)Al-PILC shows not only the more intense
bands in comparison to the other two immobilization methods,
but also the imine group signal (ca. 1616 cm−1) closer to that of the
free [Mn(3,5-dtSALEN)Cl)] complex; indeed, it is noteworthy for the
same material the presence of strong signals at 2900–2800 cm−1
corresponding to C H stretching of the methyl groups from the t-
butyl substituents. All these observations, together with its higher
efficiency in the Mn uptake (see Mn(IE) in Table 1), strongly suggest
that method C was the best procedure to successfully attach the
complex to the aluminosilicate without significant modification of
its molecular structure.
Mn contents in the modified materials showed insertion effi-
ciencies between 60 and 90% (Table 1). Besides, the content of Al
increased in comparison to the starting clay because of the incor-
poration of the alumina pillars, causing a parallel decrease in Mg,
Si and Ca contents.
It is worth noting that methods B and C displayed higher
insertion percentages than method A. The lower incorporation
efficiencies for method A could be due to the easier leaching of
the complexes from the solids during the washing steps, as this
method leads to the complexes mainly incorporated on the exter-
nal surface of the solids. The high insertion of Mn in (IB)Al-PILC
and (IIB)Al-PILC solids indicates that Mn homoionization of the Al-
pillared clay before the treatment with the ligands favoured the
metal uptake and stabilization in the clay matrix; as confirmed
by the higher compensation of CEC observed for the solids pre-
pared using method B (CCEC in Table 1). Finally, solids obtained from
method C displayed the highest Mn contents, probably because
under simultaneous pillaring and encapsulation the complexes
underwent easier access into the porous hosts.
The CEC values demonstrated that a compensation of the CEC
of the clay by Al13 pillars was achieved in all the modified mate-
rials. Moreover, it must be also accounted that solids obtained by
method A showed very similar CCEC values that Al-PILC, whereas
materials modified by either methods B or C displayed significantly
higher values. It may be inferred that B and C methods allowed the
complexes to get immobilized mostly into the interlayer space of
the clay mineral, thus contributing to further compensation of the
starting CEC. Regarding the nature of the complex, materials mod-
ified with complex II showed higher efficiency in the Mn uptake
and higher CEC compensation than those modified with complex
I, except for method C. This was probably due to the presence of a
voluminous substituent near the imine bridge of II. In spite of the
bigger size of complex II respect to complex I, the adsorption of
the complex (method A) or of the ligand (method B) of system II
was apparently more efficient in terms of Mn uptake, under very
similar values of compensation of the initial CEC. This suggested
Fig. 3. FTIR spectra of the Mn-complexes and their clay-derived solids: (a) [Mn(3,5-
dtSALEN)Cl)] (I) and (b) [Mn(3,5-dtSALHD)Cl] (II) complexes by every method of
preparation.
ArH), 3.35 (m, 2 H, C NCH), 1.91–1.56 (m, 8 H, cyclohexyl-H), 1.44
(s, 18 H, t-butyl). Thus, it was confirmed that the targeted ligands
and complexes were successfully prepared.
3.2. Characterization of the solid, immobilized Mn(III)–SALEN
catalysts
The FTIR spectra of the materials are compared in Fig. 3. The
typical signals of the aluminosilicate together with those from the
Mn-complexes are observed. First of all, the intense and broad band
in the region 3700–3300 cm−1 that can be assigned to the OH
,
stretching in Al OH and Si OH groups of the starting clay [13],
remained for the immobilized complexes obtained by the three
methods, suggesting that no covalent interaction was established
with the metal centre of the SALEN-complexes, i.e. that mainly
physical encapsulation took place. The solids derived from method
C for both metal-complexes clearly displayed stronger signals; that
is somewhat expectable since the final solids were calcined at lower
temperature than their counterparts of the methods A and B (473 K
vs. 673 K), and a higher fraction of surface hydroxyls remains in
the solids. This band exhibits a couple of shoulders in the edge
of higher frequency (ca. 3632 cm−1), that may be related to the
hydroxyl groups from the Al-pillars. On the other hand, the immo-
bilization of both Mn(III)-complexes was evidenced by the shifting
of the imine group signal (
both ligands and appearing at lower frequency (1620–1600 cm−1
for the corresponding complexes, and being observed for
C
N), observed at 1690–1640 cm−1 for
)

A.M. Garcia et al. / Journal of Molecular Catalysis A: Chemical 416 (2016) 10–19
15
Table 1
Elemental composition (% w/w)a, compensation of the initial cation exchange capacity (CCEC), BET specific surface area (SBET), and basal spacing (d001) for the starting clay,
the Al-pillared clay and the Mn-complex-functionalized solids.
b
c
d
Sample
Fe2O3
CaO
MgO
SiO2
K2O
Al2O3
Mn
Mn(IE) (%)
CCEC (%)
SBET (m2/g)
d001 (nm)
BVC
Al-PILC
4.19
1.79
3.36
3.50
4.19
3.95
4.53
4.07
0.21
0.04
0.04
0.03
0.03
0.04
0.03
0.03
0.58
0.58
0.45
0.35
0.18
0.75
0.35
0.58
57.3
49.0
44.8
53.3
51.1
49.8
47.3
41.9
0.14
0.05
0.10
0.07
0.14
0.24
0.11
0.20
7.16
9.18
10.2
11.3
12.8
13.1
13.3
13.8
0.03
0.04
1.89
2.40
2.52
2.11
2.53
2.22
–
–
–
80
208
50
65
73
38
16
55
1.42
1.70
1.59
1.65
2.76
1.70
1.71
1.90
34
36
56
56
37
59
55
(IA)Al-PILC
(IB)Al-PILC
(IC)Al-PILC
(IIA)Al-PILC
(IIB)Al-PILC
(IIC)Al-PILC
68.8
87.4
91.7
76.8
92.1
80.8
a
The metal contents are expressed in terms of their corresponding most stable oxides; except Mn, which is expressed as element content since it is retained on the clay
in its complex form. All values are dry-basis.
b
Mn(IE): insertion efficiency of Mn.
c
CCEC = percentage of the CEC of the starting clay that was compensated by the sequential Al-pillaring and Mn-complex immobilization treatments.
Determined from the corresponding powder XRD patterns.
d
2.76 nm
that adverse adsorption of complex II into the micropores of the
pillared clay was overcome by an extra stability provided to the
complex by the cyclohexane substituent on the diimine bridge, but
this bigger size is apparently not determinant on the stabilization
within the interlayer space of bentonite, since the complexes are
featured by a planar structure that fits well into the quasi-infinite
two dimensional interlayer space available in Al-PILC, only lim-
ited by the average distance between the alumina pillars. However,
somewhat unexpectedly, the method C, in which the Al13 interca-
lation and complexes encapsulation were simultaneous, showed
lower Mn uptake, although similar CEC compensation for complex
II respect to complex I.
(a)
(IC)Al-PILC
(IB)Al-PILC
(IA)Al-PILC
Al-PILC
I (a.u)
1.42 nm
10
BVC
2
4
6
8
12
14
The BET specific surface areas of the materials are also summa-
rized in Table 1; SBET decreased for the heterogeneized complexes
compared to the Al-pillared clay Al-PILC. This can be attributed
to the insertion of the complexes within the porous network of
the pillared aluminosilicate provoking a partial blockage of the
pores. Besides, the catalysts displaying lower specific surface area
are those prepared with complex II, which has the voluminous
cyclohexyl substituent on the diimine bridge, in according with the
effect reported elsewhere [43]. In general, the materials prepared
by method C undergone lower decreases in the specific surface area
upon complex immobilization. In addition, as somehow expectable,
the larger size of the SALEN ligand decreased in higher amount the
final specific surface area of the catalysts prepared using method
B, in which the in-situ formation of the complexes by diffusion of
the ligand through the clay pores up to the previously exchanged
Mn2+ cations is obviously more sterically hindered in the case of
complex II.
2 theta (°)
(b)
1.90 nm
1.71 nm
(IIC)Al-PILC
(IIB)Al-PILC
(IIA)Al-PILC
Al-PILC
I (a.u)
1.42 nm
BVC
2
4
6
8
10
12
14
2 theta (°)
The XRD powder patterns for the starting clay and the materials
modified with complex I (a) and complex II (b) are compared in
Fig. 4, and their corresponding basal spacings are given in Table 1.
All the modified solids displayed a shifted d001 signal respect to the
starting mineral (6.20◦ 2ꢀ, 1.42 nm), showing that the three meth-
ods allowed to incorporate the SALEN complexes without loss of
the expanded structure of the pillared clay. The solids prepared by
methods A y B, in which the immobilization was performed on the
clay previously pillared, basal spacings (1.59–1.71 nm) were pretty
similar to that of the pillared bentonite Al-PILC (1.70 nm); hence,
the incorporation of Mn did not significantly decrease the basal
spacing of the precursor solids, in agreement with the behaviour
reported by Cardoso et al. [19]. However, the broadening of the
d001 peaks in the complex-functionalized solids evidenced a loss of
c-axis ordering, suggesting a heterogeneous distribution of the Mn-
complexes and the alumina pillars, although without drastically
affecting the thermal stability of the solids.
Fig. 4. Small-angle powder XRD patterns of the starting and modified materials: (a)
complex (I)—derived catalysts and (b) complex (II)—derived catalysts.
distribution of the pillars within the interlayer space of the clay,
taking into account that method C involved simultaneous interca-
lation/encapsulation of the SALEN complexes. At the same time,
the basal spacing of (IC)Al-PILC with the smallest Schiff-base lig-
and, was clearly higher than that of (IIC)Al-PILC; suggesting a very
interesting role of the SALEN-like complexes in the intercalation of
the clay with the Al-polycations. Apparently, complex I induced a
swelling of the clay layers in the presence of the pillaring polyca-
tions significantly higher than the typical values reported for the
Keggin adduct, around 0.7–0.9 nm, as measured before or after the
final thermal treatment, to more than twice that value (∼1.8 nm).
Furthermore, although solid (IIC)Al-PILC displayed the higher basal
spacing in between the set of materials prepared with complex II
(1.9 nm), this effect was not so important as in the case of (IC)Al-
PILC. While complex II promoted a more homogeneous distribution
of the pillaring species throughout the interlayer of the clay by
The solids prepared using method C displayed basal spacings
(1.90 nm–2.76 nm) clearly higher than Al-PILC. The presence of the
complex in the system apparently led to enhanced, widespread

16
A.M. Garcia et al. / Journal of Molecular Catalysis A: Chemical 416 (2016) 10–19
100
100
80
60
40
20
0
NaHCO3 + (IC)Al-PILC
molar ratio = 1.0
molar ratio = 0.1
80
60
40
20
0
NaHCO3
molar ratio = 2.0
(IC)Al-PILC
0
2
4
6
8
0
2
4
6
8
time (h)
time (h)
Fig. 6. Catalytic performance of the bicarbonate-assisted cyclohexene epoxidation
in the presence or absence of (IC)Al-PILC catalyst.
Fig. 5. Effect of the NaHCO3/cyclohexene molar ratio on the conversion of cyclo-
hexene by H2O2 in the presence of (IC)Al-PILC catalyst.
Table 2
serving as a kind of organometallic spacer, complex I seemed to
intervene in the aggregation of the Keggin-like species probably
leading them to form dimers by self-assembly around the Mn-
complex or inducing larger polynuclear species of the metal before
and even during the intercalation stage [44].
Cyclohexene conversion as a function of the H2O2/cyclohexene molar ratio in the
presence of the Mn–SALEN clay catalyst (IC)Al-PILC.
Molar ratio H2O2/cyclohexene a Cyclohexene conversion (%) H2O2 efficiency b (%)
4
10
12
0
99
99
0.0
9.9
8.2
3.3. Bicarbonate-assisted cyclohexene epoxidation by H2O2
a
Catalyst loading = 0.02 g; NaHCO3 loading = 1.0 mmol.
Calculated as follows: H2O2 efficiency (%) = (mmol of H2O2 reacted for cyclohex-
b
The effect of the amount of bicarbonate used as co-catalysts
was first investigated by carrying out a series of experiments using
(IC)Al-PILC as catalyst and adding various amounts of NaHCO3 (see
Fig. 5). The increase from 0.1 to 1.0 in the bicarbonate/cyclohexene
molar ratio led to enhance the final cyclohexene conversion from
92 to 99% after 8 h of reaction, with similar trends at lower times.
Moreover, further increase in the NaHCO3 amount, up to twice the
substrate, caused the opposite effect, decreasing the conversion at
8 h of reaction as well as making slower the rate of reaction for
times higher than 1 h.
ene conversion/mmol of H2O2 added) × 100.
demonstrates the key role played by NaHCO3 for the activation
of H2O2 as oxidizing agent in epoxidation, reported mainly to be
related with the enhanced rate of the epoxidation against the para-
site disproportionation of the hydrogen peroxide [34]. It is obvious
that the combination of various properties as the formation of the
active peroxymonocarbonate species, the catalytic activity of the
Mn(III)–SALEN complex and the stability and encapsulation effi-
ciency of the hosting pillared clay allowed to obtain improved
catalytic performance, avoiding at the same time the full leach-
ing of the complex from the solid and preventing the formation
of inactive species. Other series of catalytic experiments showed
the effect of the molar ratio H2O2/cyclohexene in the conversion of
the olefin, using (IC)Al-PILC keeping constant the amounts of cat-
alyst and bicarbonate (0.02 g and 1 mmol, respectively) (Table 2).
Using a molar ratio of 4.0, no conversion at all was found, but when
the molar ratio was 10 conversion achieved was 99%. A further
increase in the molar ratio to 12 did not conduce to any further
improvement of the cyclohexene conversion, so it can inferred
that a ratio H2O2/cyclohexene between 4 and 10 allows to dra-
matically enhance the olefin conversion in the presence of the
Mn–SALEN immobilized clay-catalysts; although a higher concen-
tration of peroxide may lead to the formation of other products
of the cyclohexene oxidation contributing to increase the olefin
conversion [36], the disproportionation of hydrogen peroxide also
forms higher amounts of water, decreasing the overall efficiency
of the oxidizing agent. The global efficiency of conventional cat-
alytic epoxidations is often affected by the presence of water. It
happens mainly by the ring opening of the epoxides by water catal-
ysed by Lewis acid-sites [34]. However, in our case it must be
realized that two phases with very different polarity are into the
same reacting mixture, where the epoxide of course is favoured
in the organic phase (acetonitrile) [46]. Therefore, one of the main
advantages of the bicarbonate-assisted epoxidations by H2O2 is in
fact the lower susceptibility of the system to the presence of water.
It occurs thanks to the buffer effect exerted by the bicarbonate,
where the slightly basic pH promotes formation of the peroxy-
monocarbonate species useful enhancing the rate of epoxidation,
while the ring opening of the epoxides towards the undesired diol
Some reports, like that from Yao and Richardson [45], have
demonstrated that bicarbonate is a good activator of H2O2 in
aqueous catalytic epoxidations of alkenes, in terms of allowing
higher conversions, while decreasing the parasite disproportion-
ation. Therefore, the use of bicarbonate as co-catalyst has shown
to be crucial in order to achieve maximum olefin conversion in the
presence of Mn–SALEN complexes at lower times of reaction. Lane
et al. [34] proposed that a key factor to get high conversions in
bicarbonate-assisted olefin epoxidations with H2O2 is the forma-
tion of the peroxymonocarbonate ion (HCO4)−, a nucleophile more
reactive than hydrogen peroxide, formed in the pH range between
7.0 and 9.0. Therefore, the result observed can be explained con-
sidering that peroxymonocarbonate ion is effectively stabilized at
bicarbonate/cyclohexene molar ratio of 1.0, but for ratios signif-
icantly exceeding this value (as in the case of 2.0), the pH may
−
surpass 9.0, the deprotonation and decomposition of HCO4 to
2−
CO3 become significant according to the equilibrium in Eq. (1),
strongly decreasing the concentration of HCO4 and affecting the
−
conversion of the olefin.
HCO−3 + H2O2 ꢀ HCO− + H2O
(1)
4
Afterwards, some experiments were made with or without the
addition of catalyst to point out the role truly played by this and
the bicarbonate ion in the olefin conversion (Fig. 6). NaHCO3 even
in the absence of the catalyst was capable to contribute to the
conversion of cyclohexene (though below 30%, 8 h), showing that
the peroxymonocarbonate ion is itself active, although fairly less
efficient when the (IC)Al-PILC catalyst is present (99%, 8 h). In the
presence of the same complex-immobilized clay catalyst but with-
out addition of bicarbonate co-catalyst, a more rather negligible
conversion was achieved (around 6.0%) upon 8 h of reaction. This

A.M. Garcia et al. / Journal of Molecular Catalysis A: Chemical 416 (2016) 10–19
17
Table 3
Cyclohexene epoxidation catalysed by SALEN- and SALHD- Mn(III) complexes immobilized on the Al-pillared bentonite.
Catalyst
Run
Conversion a (%)
Selectivity b(%)
Yield c (%)
H2O2 efficiency d(%)
TON e
(IA)Al-PILC
(IB)Al-PILC
(IC)Al-PILC
First
First
First
Second
First
First
First
Second
99
99
99
85
96
97
100
84
63
63
70
ND
30
26
77
ND
62
62
69
ND
29
25
77
ND
9.9
9.9
9.9
8.5
9.6
9.7
10.0
8.4
90
71
75
ND
38
27
95
ND
(IIA)Al-PILC
(IIB)Al-PILC
(IIC)Al-PILC
a
Determined by GC-FID; time of reaction = 8 h. ND—not determined.
b
c
Determined as follows: selectivity (%) = (mmol epoxide/mmol cyclohexene reacted) × 100.
Yields were calculated as the product (conversion × selectivity) divided by 100 in each case.
d
e
Calculated as follows: H2O2 efficiency (%) = (mmol of H2O2 reacted for cyclohexene conversion/mmol of H2O2 added) × 100.
TON = (mmol cyclohexene epoxide/mmol Mn in the catalyst).
becomes inhibited in absence of the Lewis acid-sites. Besides, it can
be inferred that the equilibrium of Eq. (1) contributes to decrease
the fraction of the H2O2 decomposed via disproportionation, in
other words, to increase the efficiency of the H2O2 in the reac-
tion. Moreover, the obtained hydrogen peroxide efficiencies were
low due to the theoretical relationship of 1 mol of hydrogen per-
oxide required to epoxidize 1 mol of cyclohexene (Tables 2 and 3).
However, it must be also remarked that our results displayed effi-
ciencies in close correlation with other studies carried out also in
aqueous medium and with hydrogen peroxide as oxidizing agent,
as the results reported by Lane et al. [34], who found that the
bicarbonate-assisted epoxidation by hydrogen peroxide of a num-
ber of olefins, including cyclohexene, required 10 equivalents of
this oxidizing agent for the full conversion of the olefin. As it can be
seen, all of our catalysts showed cyclohexene conversions between
96 and 100% when used with a molar ratio H2O2/cyclohexene = 10
(Table 3). Furthermore, Costa et al. [49] more recently reported con-
versions ranging between 11.8 and 25.8%, but using a significantly
lower ratio H2O2/cyclohexene = 1.2. Remarkably, no bicarbonate
co-catalyst was used at all.
The high conversions here reported have been found in
several olefin’s epoxidations, but mostly performed in the
presence of organic oxidizing agents like iodosylbenzene or meta-
chloroperoxybenzoic acid [19,47], in order to achieve proper
miscibility in the organic mixtures. Also because the presence of
water, a product of the disproportionation of H2O2, promotes the
decomposition of the epoxide decreasing the selectivity. Recently,
it was also attained a 50% styrene conversion by using hydro-
gen peroxide in acetonitrile as solvent, but a ratio H2O2/styrene
of 5.0 was required and this value was enhanced when the reac-
tion was carried out in more polar solvents [48]. Before, Costa
et al. [49] obtained increasing cyclohexene conversions in the range
of 20–26% in the epoxidation with H2O2 catalysed by Mn-, Co-
or Fe-metalloporphyrins anchored on MCM-41, but using a fairly
lower ratio H2O2/cyclohexene of 1.2. Finally, a new Mn–SALEN type
complex, [Mn(saldien)(N3)], immobilized on SBA-15 mesoporous
material was recently reported catalysing cyclohexene oxidation
by hydrogen peroxide with conversions of 11–38% as a function
of the peroxide dosage and time of reaction [50]. In our case,
H2O2/cyclohexene ratios exceeding 10 were needed probably to
overcome not only the low miscibility in the organic mixture of
reaction, but also a partial parasite degradation of the oxidizing
agent in the presence of some traces of Fe in the inorganic hosting
mineral; moreover, there must be remarked that most of the iron
content in the starting clay here used is located within the lattice
of the mineral as reported in advance [37], and then no substantial
contribution of this reaction in the peroxide consumption should
be expected.
of the immobilized complexes, both Mn(III)-complexes delivered
almost the same close to full conversions of cyclohexene. How-
ever, in terms of both selectivity and yield the catalysts based in
complex I were in general more effective (60–70%) than those of
complex II (30–77%), as well as fairly less sensitive regarding the
method of preparation. Moreover, the solid (IIC)Al-PILC displayed
the highest catalytic performance in terms of conversion, selec-
tivity and yield in a single catalytic run. This should be probably
related to the enhanced physicochemical properties of this cata-
lyst in terms of both basal spacing and specific surface area with
respect to the counterparts obtained by methods A and B (Table 1).
In addition, the enhanced selectivity demonstrated by the two
solids prepared through simultaneous Al13-intercalation/complex
encapsulation (method C), irrespective the nature and size of the
Mn–SALEN complex, suggests that this method guarantees more
selective confinement and stabilization of the complex within the
interlayer space of the pillared clay. Similar improvement of styrene
epoxidation by iodosylbenzene using Mn(III)–SALEN immobilized
on clay minerals have been reported, where the simultaneous
Al13-intercalation/complex encapsulation was the method with the
lower leaching of the complex [19]. In other words, a considerable
fraction of the complex or the ligand may remain deposited on the
external surface of the hosting clay when using methods A and B,
respectively, and although more easily accessible, their catalytic
performance could not control the nature of the products, control
more related to the shape and size of the solid porosity, and being
more susceptible of leaching along the first catalytic run.
Finally, regarding the catalytic activity displayed by the clay-
catalysts, the top turn-over number (TON) value was reached for
(IIC)Al-PILC catalyst (95), which was slightly higher than the one
reported elsewhere by Lane et al. [34], who used dissolved MnSO4
as active species. This clearly shows that our catalysts did not
undergo loss of activity by the immobilization on the solid phase
but they were a little bit more active than the metal ions. This
probably obeyed to intrinsic higher activity of the immobilized
Schiff–Mn(III) complexes, as well as the promoting effect pro-
vided by the clay-catalytic support. In a similar way, Fraile et al.
[51] reported catalytic activity for Mn–SALEN complex [NꢀN-bis-
salicylidene-ethylendiamine]manganese(III) chloride, supported
on several clays. In the case of the commercial bentonites, the
TON numbers were significantly lower than ours (6.2 and 12.8,
for direct exchange or metal exchange followed by treatment with
the ligand, respectively), probably suggesting that simultaneous
intercalation/complex-encapsulation employed in this work seri-
ously improved the catalytic response of the SALEN-like complex.
Finally, Costa et al. [49] obtained much higher TON values using
a Mn–porphyrin complex immobilized on MCM-41 (1.54 × 105).
Such a significant difference could be of course explained by a
higher intrinsic activity of the Mn–porphyrin, but it must also be
accounted that these authors calculated the TON number based on
The conversion, selectivity, yield and TON values displayed
by all the solids are summarized in Table 3. Regarding the size

18
A.M. Garcia et al. / Journal of Molecular Catalysis A: Chemical 416 (2016) 10–19
the yield of all products and not only to the desired epoxide. In addi-
tion, as long as a pretty lower molar ratio H2O2/cyclohexene = 1.2
was used, a higher probability of the molecules of the oxidizing
agent to be activated by a metal centre is expectable, before direct
attack on the olefin or the parasite reaction of the peroxide dispro-
portionation to take place, improving overall catalytic activity.
This coincides with previous studies using hybrid catalysts of the
Mn(III)–SALEN type stabilized on pillared clays by a similar strat-
egy, where high olefin conversions and good selectivities toward
the epoxide were also achieved, using iodosylbenzene as oxidizing
agent [19].
also attributed to blocking of the active sites by products located
in the pores of the catalysts or by side reactions leading to form
␮-oxo-Mn(IV) dimeric units of the complexes, well known as inac-
tive to catalyse cyclohexene epoxidation [54]. The used catalysts of
method C ((IC)Al-PILC-U and (IIC)Al-PILC-IU) in corrected version)
were characterized by FTIR spectroscopy after the first catalytic
experiment (it is now shown as Fig. 1S, in Supplementary mate-
rial) and compared against the corresponding catalysts before use
((IC)Al-PILC and (IIC)Al-PILC) and the starting isolated complexes
(I) and (II). Such spectra basically showed that the signal of imine
group weakens after the first catalytic cycle, though more drasti-
cally in the case of (IIC)Al-PILC-U than (IC)Al-PILC-U; in addition,
a broadening of the signals was observed in the region 1565 cm−1
until around 1440 cm−1, attributed to the stretch of aromatic link-
age C C, probably overlapped with a signal displayed in the same
region by the vibration of deformation in water adsorbed on the
starting clay (BVC) (see Fig. 3). It suggested that partial decompo-
sition of both complexes took place after the first catalytic run, but
with significant fraction of the immobilized complexes still active
and stable within the catalyst’s framework.
Another factor probably related to the increased selectivity dis-
played by (IIC)Al-PILC catalyst with respect to (IC)Al-PILC would
be the higher steric hindrance exerted by the voluminous cyclo-
hexane ring on the diimine bridge [13]. This suggests that not only
the preparative strategy but also the size of the substituents on the
diimine ligand bridge may strongly control the overall catalytic per-
formance of the SALEN–Mn(III)/Al-pillared clay catalysts in olefin
epoxidations. It has been widely stated that the activity and selec-
tivity in asymmetric synthesis of epoxides are strongly affected
by the way the olefin approaches the catalytic centre and then,
in heterogeneized complexes, the orientation at which it becomes
stabilized within the solid matrix plays an important role on the
enantiomeric selectivity [52]. However, in our case the selectivity
was assessed in terms of the fraction of the olefin finally converted
to the desired epoxide in either optical form. Hence, selectivity is
mainly governed by the pathway either, homolytic or heterolytic,
the hydrogen peroxide decomposition follows in the presence of
the catalyst and the co-catalyst. It is well known that the heterolytic
pathway leads to increased incidence of side-products than the
homolytic one. Although the homolytic route may induce the fur-
ther decomposition of the epoxide by the free radicals, it has been
reported that the incidence of this on the selectivity is negligible in
comparison to the influence exerted by the heterolytic decompo-
sition (disproportionation) of the peroxide [53]. In the other hand,
different selectivity to the epoxide found in our results as a func-
tion of the final basal spacing displayed by the clay-catalysts (see
for instance d001 values of (IC)Al-PILC vs (IC)Al-PILC in Table 1)
can be explained in terms of bigger pore sizes and surfaces when
d001 increases instead than different orientations displayed by the
metal complexes within the framework of the functionalized pil-
lared clay. Apparently in the case of the cyclohexene epoxidation,
the selectivity to the epoxide is favoured by basal spacings as close
as possible to those values typical of pillared clays (1.70–1.90 nm),
probably because higher pore sizes miss shape-selectivity induced
by the pore channels of the modified aluminosilicate.
The highly active catalysts ((IC)Al-PILC and (IIC)Al-PILC) were
selected for reusing tests. A slight decrease in the cyclohexene con-
version was observed in the second catalytic run for both catalysts
(see Table 3). This showed that SALEN- and SALHD–Mn(III) com-
plexes displayed high stability even under the aggressive oxidizing
conditions of reaction when they were stabilized by simultaneous
intercalation/encapsulation. It has been proposed that this type of
catalysts would be employed up to four sequential runs without
significant loss in their catalytic properties [13,19,20]. The slow
deactivation may occur by decomposition of the metal complexes
within the hosting aluminosilicate involving oxidation exerted by
the excess of hydrogen peroxide; according to Cardoso et al. [19],
the deactivation of the catalytic centres can occur by decomposition
of the complex inside the matrix by the oxidant. In this partic-
ular case, since the H2O2 is a very strong oxidizing agent, attack
on the Mn–SALEN type complexes may take place by means of at
least two ways: the disaggregation of the complex in the metal ion
and the ligand predominantly by the change in the oxidation state
of the metal centre, as well as the straightforward attack on the
structure of the ligands, due to their organic nature. It could be
4. Conclusions
Two Mn(III) complexes with Schiff bases – [Mn(3,5-dtSALEN)Cl]
and [Mn(3,5-dtSALHD)Cl] – were successfully synthesized with
pretty high yields and featured the typical expected signals
by UV–vis and FTIR and spectroscopies. Immobilization of both
Mn complexes on
a natural bentonite by three experimen-
tal approaches allowed to obtain six heterogeneized catalysts;
simultaneous Al13-pillaring/complex encapsulation of the alumi-
nosilicate (method C) showed the highest Mn uptake efficiency
exceeding 90%, against below 70% when every complex was directly
adsorbed on the clay previously pillared by a conventional proce-
dure (method A). The same immobilization procedure also led to
enhanced basal spacing in the final materials and improved com-
pensation of the cationic exchange capacity of the starting clay;
indeed, it was observed that such a type of SALEN-like Mn(III) com-
plexes incorporated by this procedure may significantly improve
the intercalation of the starting clay by acting as pillar-spacers
and leading to more homogeneous pillar distribution or by giv-
ing rise to highly aggregated, larger Al-based oligocations, able to
expand the starting mineral to basal spacings beyond those typical
of Al-pillared clays. The modified materials exhibited high catalytic
performance in the bicarbonate-assisted cyclohexene epoxidation
by hydrogen peroxide, in a cleaner reaction pathway with the cat-
alytic species entrapped in a low cost inorganic hosting mineral
and using a more environmental-friendly oxidizing agent than the
usual iodosylbenzene. All materials reached conversions in the
range 84–100% of the starting cyclohexene; the solids prepared
by method C displayed the higher conversions of the olefin (up to
100%) and selectivities to the targeted epoxide (up to 77%), which
decreased up to 85% when used in a second catalytic cycle, showing
that pillars were able to keep trapped the active metal complexes
in a great extent, though the typical structure of the metal com-
plexes was partially affected by once finished the first catalytic cycle
(according to FTIR of used materials).
Acknowledgements
Authors gratefully acknowledge funding by projects
VIPRI—Narin˜o University (agreement 031/2009) and ID-3920
from Cauca University. V. Moreno thanks the grant received
from COLCIENCIAS, program “Young Researchers”. A. Gil and
M.A. Vicente thank the support from the Spanish Ministry of
Economy and Competitiveness (MINECO) and the European

A.M. Garcia et al. / Journal of Molecular Catalysis A: Chemical 416 (2016) 10–19
19
Regional Development Fund (FEDER) (project MAT2013-47811-
C2-R).
[23] T. Luts, W. Suprun, D. Hofmann, O. Klepel, H. Papp, J. Mol. Catal. A: Chem. 261
(2007) 16–23.
[24] G. Nagendrappa, Appl. Clay Sci. 53 (2011) 106–138.
[25] L.A. Galeano, M.A. Vicente, A. Gil, Catal. Rev. 56 (2014) 239–287.
Appendix A. Supplementary data
[26] A. Gil, S.A. Korili, R. Trujillano, M.A. Vicente, Appl. Clay Sci. 53 (2010) 97–105.
[27] S. Arellano, T. Gallardo, T. Osorio, Rev. Soc. Quim. Mex. 44 (2000) 204–208.
[28] G. Johansson, Acta Chem. Scand. 14 (1960) 771–773.
[29] F. Figueras, Catal. Rev. 30 (1988) 457–499.
[30] R. Mokaya, W. Jones, J. Catal. 172 (1997) 211–221.
[31] I. Kuz´niarska-Biernacka, A.R. Silva, R. Ferreira, A.P. Carvalho, J. Pires, C. Freire,
B. de Castro, New J. Chem. 28 (2004) 853–858.
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.molcata.2016.01.
034.
[32] G. Grigoropoulou, J.H. Clark, J.A. Elings, Green Chem. 5 (2003) 1–7.
[33] R.A. Sheldon, Top. Curr. Chem. 164 (1993) 21–34.
[34] B.S. Lane, M. Vogt, V.J. De Rose, K. Burgess, J. Am. Chem. Soc. 124 (2002)
11946–11954.
[35] M. Ghorbanloo, H.H. Monfared, C. Janiak, J. Mol. Catal. A: Chem. 345 (2011)
12–20.
[36] B. Qi, X.-H. Lu, D. Zhou, Q.-H. Xia, Z.-R. Tang, S.-Y. Fang, T. Pang, Y.-L. Dong, J.
Mol. Catal. A: Chem. 322 (2010) 73–79.
[37] L.A. Galeano, A. Gil, M.A. Vicente, Appl. Catal. B 100 (2010) 271–281.
[38] A. Silva, C. Freire, B. De Castro, New J. Chem. 28 (2004) 253–260.
[39] A. Silva, V. Budarin, J. Clark, B. De Castro, C. Freire, Carbon 43 (2005)
2096–2105.
[40] J. Cubillos, S. Vásquez, C. Montes De Correa, Appl. Catal. A 373 (2010) 57–65.
[41] A. Chellamani, P. Kulanthaipandi, S. Rajagopal, J. Org. Chem. 64 (1999)
2232–2239.
[42] L.A. Galeano, A. Gil, M.A. Vicente, App. Catal. B 104 (2011) 252–260.
[43] R. Ji, K. Yu, L.-L. Lou, Z. Gu, S. Liu, J. Inorg. Organomet. Polym. Mater. 20 (2010)
675–683.
[44] S. Bi, C. Wang, Q. Cao, C. Zhang, Coord. Chem. Rev. 248 (2004) 441–455.
[45] H. Yao, D.E. Richardson, J. Am. Chem. Soc. 122 (2000) 3220–3221.
[46] G. Salomao, M. Olsen, V. Drago, C. Fernandes, L. Cardozo, O. Antunes, Catal.
Commun. 8 (2007) 9–72.
[47] J. Huang, J. Cai, C.M. Li, X. Fu, Inorg. Chem. Commun. 44 (2014) 20–24.
[48] M.B. Ansari, E.-Y. Jeong, S.-E. Park, Green Sustain. Chem. 2 (2012) 1–7.
[49] A. Costa, G. Ghesti, J. de Macedo, V. Braga, M. Santos, J. Dias, S. Dias, J. Mol.
Catal. A: Chem. 282 (2008) 149–157.
References
[1] G. Franz, R.A. Sheldon, in: B. Elvers, S. Hawkins, G. Shulz (Eds.), UIImann’s
Encyclopedia of Industrial Chemistry, vol. A, John Wiley and Sons, Inc.,
Weinheim, 1991, pp. 261–311.
[2] M.G. Clerici, G. Bellussi, U. Romano, J. Catal. 129 (1991) 159–167.
[3] S. Bhattacharjee, J.A. Anderson, J. Mol. Catal. A: Chem. 249 (2006) 103–110.
[4] U. Biermann, W. Friedt, S. Lang, W. Luhs, G. Machmuller, J.O. Metzger, M.R.
Klaas, H.J. Schafer, M.P. Schneider, Angew. Chem. Int. Ed. 39 (2000)
2206–2224.
[5] M.A.R. Meier, J.O. Metzger, U.S. Schubert, Chem. Soc. Rev. 36 (2007)
1788–1802.
[6] A. Corma, H. García, Chem. Rev. 102 (2002) 3837–3892.
[7] K. Srinivasan, P. Michaud, J. Kochi, J. Am. Chem. Soc. 108 (1986) 2309–2320.
[8] R. Irie, K. Noda, Y. Ito, N. Matsumoto, T. Katsuki, Tetrahedron Lett. 31 (1990)
7345–7348.
[9] E. Jacobsen, W. Zhang, A. Muci, J. Am. Chem. Soc. 113 (1991) 7063–7064.
[10] T. Katsuki, Coord. Chem. Rev. 140 (1995) 189–214.
[11] I. García-Bosch, X. Ribas, M. Costas, Adv. Synth. Catal. 351 (2009) 348–352.
[12] D.E. De Vos, M. Dams, B.F. Sels, P.A. Jacobs, Chem. Rev. 102 (2002) 3615–3640.
[13] P. Das, I. Kuz´niarska-Biernacka, A.R. Silva, A. Carvalho, J. Pires, C. Freire, J. Mol.
Catal. A: Chem. 248 (2006) 135–143.
[14] D. Chatterjee, A. Mitra, J. Mol. Catal. A: Chem. 144 (1999) 363–367.
[15] B. Choudary, N. Sreenivasa, M. Lakshmi, P. Lakshmi, Catal. Lett. 76 (2001)
213–218.
[50] S. Alavi, H. Hosseini-Monfared, M. Siczek, J. Mol. Catal. A: Chem. 377 (2013)
16–18.
[51] J.M. Fraile, J.I. García, J. Massam, J.A. Mayoral, J. Mol. Catal. A: Chem. 136
(1998) 47–57.
[52] G. Oxford, D. Dubbeldam, L. Broadbelt, R. Snurr, J. Mol. Catal. A: Chem. 334
(2011) 89–97.
[53] R. Sheldon, J. Van Doorn, J. Catal. 31 (1973) 427–437.
[54] D.D. Agarwal, R.P. Bhatnagar, R. Jain, S. Srivastava, J. Chem. Soc. Perkin Trans. 2
(1990) 989–992.
[16] I. Kuz´niarska-Biernacka, A. Silva, A. Carvalho, J. Pires, C. Freire, Langmuir 21
(2005) 10825–10834.
[17] B. Cardoso, A.R. Silva, B. Castro, C. Freire, Appl. Catal. A 285 (2005) 110–118.
[18] K. Yu, Z. Gu, R. Ji, L.-L. Lou, S. Liu, Tetrahedron 65 (2009) 305–311.
[19] B. Cardoso, J. Pires, A. Carvalho, I. Kuzniarska-Biernacka, A. Silva, B. Castro, C.
Freire, Microporous Mesoporous Mater. 86 (2005) 295–302.
[20] B. Cardoso, J. Pires, A.P. Carvalho, I. Kuz´niarska-Biernacka, A.R. Silva, C. Freire,
B. de Castro, Eur. J. Inorg. Chem. 5 (2005) 837–844.
[21] L. Gandía, M.A. Vicente, A. Gil, Appl. Catal. A 196 (2000) 281–292.
[22] M.A. Vicente, R. Trujillano, K.J. Ciuffi, E.J. Nassar, S.A. Korili, A. Gil, et al.,
Pillared Clays and Related Catalysts, in: A. Gil (Ed.), Springer, Berlin, 2010, pp.
301–318.