Kinetic investigation of aerobic epoxidation of limonene over cobalt substituted mesoporous SBA-16

Article abstract of DOI:10.1039/d0cy01700k

Incorporation of low coordination Co²⁺within the structure of mesoporous silica SBA-16 has been accomplished through a facile and green “pH adjusting” method. The resulting materials were used as heterogeneous catalysts for aerobic Mukaiyama epoxidation of limonene in the presence of isobutyraldehyde, under very mild conditions. The structural integrity during the pH adjustment procedure at various loadings and states of cobalt ions within the mesoporous structure were determined using characterization techniques including nitrogen physisorption, X-ray fluorescence, diffuse reflectance UV-vis, scanning electron microscopy, temperature-programmed reduction, X-ray photoelectron spectroscopy and powder X-ray diffraction. These catalysts showed quite high reactivity for the epoxidation of limonene with high epoxide yields under optimized oxygen pressure. In this work, a thorough kinetic analysis of aerobic epoxidation of limonene was investigated to allow proposing a reaction scheme. A new mechanism, in which a surface reaction between a Co³⁺OO^?peroxo intermediate and limonene was found to be involved in the formation of the epoxidized limonene. The kinetics developed from the proposed mechanism was accurately fitted with extensive experimental initial reaction rate data. The activation energy for limonene mono epoxide formation was determined to be 22 kJ mol^?1

Full text of DOI:10.1039/d0cy01700k

View Article Online
View Journal
Catalysis
Science &
Technology
Accepted Manuscript
This article can be cited before page numbers have been issued, to do this please use: S. Madadi, J.
Bergeron and S. Kaliaguine, Catal. Sci. Technol., 2020, DOI: 10.1039/D0CY01700K.
Volume
Number
21 January 2017
Pages 313-534
7
2
This is an Accepted Manuscript, which has been through the
Royal Society of Chemistry peer review process and has been
accepted for publication.
Catalysis
Science &
Technology
rsc.li/catalysis
Accepted Manuscripts are published online shortly after acceptance,
before technical editing, formatting and proof reading. Using this free
service, authors can make their results available to the community, in
citable form, before we publish the edited article. We will replace this
Accepted Manuscript with the edited and formatted Advance Article as
soon as it is available.
You can find more information about Accepted Manuscripts in the
Information for Authors.
Please note that technical editing may introduce minor changes to the
text and/or graphics, which may alter content. The journal’s standard
Terms & Conditions and the Ethical guidelines still apply. In no event
shall the Royal Society of Chemistry be held responsible for any errors
or omissions in this Accepted Manuscript or any consequences arising
from the use of any information it contains.
ISSN 2044-4761
PAPER
Chaoqiu Chen, Yong Qin et al.
Highly dispersed Pt nanoparticles supported on carbon
nanotubes produced by atomic layer deposition for hydrogen
generation from hydrolysis of ammonia borane
rsc.li/catalysis

Page 1 of 64
Catalysis Science & Technology
View Article Online
DOI: 10.1039/D0CY01700K
Kinetic investigation of aerobic epoxidation of limonene over
cobalt substituted mesoporous SBA-16
Sara Madadi1, Jean-Yves Bergeron2, Serge Kaliaguine*1
1 Chemical Engineering Department, Laval University, Québec, G1V 0A6, Canada
2 Exploratory research leader, Soprema Inc., Drummondville, Canada
*Corresponding author (S. Kaliaguine)
E-mail : Serge.Kaliaguine@gch.ulaval.ca
Tel: 418 656-2708 poste 402708

Catalysis Science & Technology
Page 2 of 64
View Article Online
DOI: 10.1039/D0CY01700K
Abstract
Incorporation of low coordination Co2+ within the structure of mesoporous silica SBA-16
has been accomplished through a facile and green "pH adjusting" method. The resulting
materials were used as heterogeneous catalysts for aerobic Mukaiyama epoxidation of
limonene in the presence of isobutyraldehyde, under very mild conditions. The structural
integrity during the pH adjustment procedure at various loadings and state of cobalt ions
within the mesoporous structure were determined using characterization techniques
including nitrogen physisorption, X-ray fluorescence, diffuse reflectance UV–vis,
scanning electron microscopy, temperature-programmed reduction, X-ray photoelectron
spectroscopy and powder X-ray diffraction. These catalysts showed quite high reactivity
for the epoxidation of limonene with high epoxide yields under optimized oxygen
pressure. In this work, a thorough kinetic analysis of aerobic epoxidation of limonene was
investigated to allow proposing a reaction scheme. A new mechanism, in which a surface
reaction between a Co3+OO- peroxo intermediate and limonene was found to be involved
in the formation of the epoxidized limonene. The kinetics developed from the proposed
mechanism was accurately fitted with extensive experimental initial reaction rate data. The
activation energy for limonene mono epoxide formation was determined to be 22 kJ mol-
1
.
Keywords: Aerobic epoxidation; pH adjustment method; low coordination Co2+;
Limonene; kinetic.

Page 3 of 64
Catalysis Science & Technology
View Article Online
DOI: 10.1039/D0CY01700K
1. Introduction
In current polymer product development, the focus of research is increasingly directed
towards substitution of petroleum-based polymers with sustainable and renewable bio-
based polymers using environmentally benign synthetic pathways [1]. Vegetable biomass
is the major potential sustainable source of organic carbon. Monomers derived from
biomass can be divided into four categories based on hydrogen, carbon and oxygen
contents [2]. Terpenes, as hydrocarbon-rich monomers have a great potential to substitute
currently used petrochemical monomers. Limonene as a bio-based, non-food resource is a
very promising member of the terpene family, which is mainly obtained from the citrus
peel wastes. Its annual production is over 90000 tons per year and it is considered as an
economically feasible renewable feedstock [3, 4]. Its versatility as a monomer is revealed
by the great variety of polymers that can be derived from limonene [1, 5]. The primary and
secondary epoxidation products of limonene namely, limonene oxide (LMO) and limonene
dioxide (LDO), have been recently employed as biosourced monomers in the synthesis of
new green polycarbonates [6]. The polymerization of LMO and LDO with carbon dioxide
can produce limonene polycarbonate whereas limonene dicarbonate may be used as
monomer to produce the green polymer nonisocyanate polyurethane [7, 8]. These
monomers can also be utilized for the development of epoxy resins or as a cross-linker.
Considering the requirements from green chemistry and atom efficiency, there is high
incentive to exploit molecular oxygen as the greenest oxidant for the epoxidation of olefins
[9, 10]. Therefore, developing an efficient catalytic system able to activate molecular
oxygen in its ground state (spin-forbidden process) is a major challenge in this field [11].

Catalysis Science & Technology
Page 4 of 64
View Article Online
DOI: 10.1039/D0CY01700K
Cobalt-catalyzed epoxidation of alkenes with molecular oxygen is widely investigated.
Tang et al. [12] and Beier et al. [13] respectively reported the catalytic performances of
the Co2+-exchanged faujasite zeolites and cobalt-based metal–organic framework STA-
12(Co) for the epoxidation of olefins with O2 in the presence of DMF. In another study,
Quek and co-workers prepared a series of framework substituted Co-TUD-1 catalysts, by
direct hydrothermal synthesis, cobalt substituted MCM-41 and cobalt cations substituted
NaX faujasite zeolite via an ion exchange method. The catalytic performance of these
catalysts was tested in liquid phase trans-stilbene epoxidation with molecular oxygen as
oxidant in the presence of DMF [14]. The solvent used, DMF, was however likely acting
as an oxygen-transfer agent or “sacrificial” solvent forming considerable amounts of by-
products such as N-formyl-N-methylformamide [13, 15]. Therefore, it was found that the
atom-efficiency of epoxidation with molecular oxygen in DMF and other amide solvents
is poor. These were classified as hazardous solvents, hence the method is far from being a
“green” technology [16, 17]. Molecular oxygen is the most desirable oxidant, however
being a triplet causes requirement of a sacrificial solvent or reductant to perform the
epoxidation of the target compound in liquid phase. This results in the production of an
oxidized form of this reductant (scheme1). Catalytic epoxidation of ethylene in the gas
phase utilizing silver catalysts is reported as a rare exception [18].
Mesoporous molecular sieves attracted much attention in the field of catalysis due to their
high surface area, uniform pore size distribution and large pore size for easy diffusion of
bulky reactant molecules [19]. These characteristics of mesoporous materials make them
suitable hosts for grafting metal ions in order to enhance the density of active sites [20].
Nevertheless, effective introduction of heteroatoms in the framework of mesoporous

Page 5 of 64
Catalysis Science & Technology
View Article Online
DOI: 10.1039/D0CY01700K
materials prepared under strongly acidic conditions is rather difficult and therefore proper
selection of a preparation procedure is important to achieve a homogeneous metal
dispersion on the support. Recently, much effort has been made to incorporate
heteroelements into mesoporous molecular sieves by two methods of “postsynthesis
grafting” [21-24] and “direct-synthesis” [25]. In a first part of this study, Co/SBA-16
prepared by post-grafting impregnation of the support with a solution of cobalt
acetylacetonate (Co(acac)2) was found active in epoxidation of limonene with oxygen in
the presence of isobutyraldehyde and ethylacetate as a green solvent under very mild
conditions [23] . Although the postsynthesis procedure can allow grafting high-content of
heteroatoms into mesoporous silica, it tends to form metal oxides on the external surface
or in the pores at relatively high heteroatoms content [26, 27]. Wu et al. [28] introduced an
efficient procedure designated as “pH-adjusting” for the grafting of Al and Ti into
mesostructured of SBA-15, during which large amounts of heteroatoms could be
substituted into the mesophase. This pH-adjusting method was implemented in another
study by Cui et al. to substitute Co2+ into the structure of SBA-15 which then exhibited
high activity in aerobic epoxidation of styrene [26]. In the present study, we also
incorporated a certain amount of cobalt in suitable oxidation state into mesoporous silica
SBA-16 by the “pH-adjusting” method. The results were compared with our previous
results obtained with catalysts produced by postsynthesis grafting of SBA-16 with cobalt
acetylacetonate. This comparison allowed investigating the effect of the catalyst
preparation procedure and use of a chelated precursor, for improving aerobic Mukaiyama
epoxidation of limonene. A systematic study of the catalyzed epoxidation of limonene

Catalysis Science & Technology
Page 6 of 64
View Article Online
DOI: 10.1039/D0CY01700K
under pressurized oxygen was carried out. The initial reaction rates of aerobic epoxidation
of limonene were analyzed and the reaction mechanism discussed.
2. Materials and methods
2.1 Catalyst preparation
The general procedure for SBA-16 synthesis and pH-adjusting method was according to
references [28-30]. The synthesis of Co/SBA-16 using the pH-adjusting method was
carried out as follows: 5.35 g of Pluronic F127(BASF) was dissolved at 318 K in a solution
containing 257.2 g distilled water and 10.6 g concentrated hydrochloric acid employing a
magnetic stirrer. After complete dissolution, n-butanol was added all at once to the
solution. Finally, after 1 h stirring the silica source, tetraethylorthosilicate (TEOS, Aldrich,
98%), was added to the solution and agitated for 4 h at 318 K. A desired amount of
Co(NO3)2·6H2O (Alfa Aesar) was then added to the mixture, followed by additional
stirring at 318 K for 20 h. The mixture was then transferred into an autoclave for
hydrothermal aging at 373 K for two days. The pH value of the synthesis mixture was then
adjusted to 7.5 by adding ammonia at room temperature dropwise followed by another 2
days of aging at 373 K. The obtained solid was filtered , washed thoroughly with deionized
water and dried at room temperature in vacuum oven. Finally, the surfactant was removed
by ethanol Soxhlet extraction. A sample of SBA-16 was prepared by this pH adjusting
method without adding the cobalt source.

Page 7 of 64
Catalysis Science & Technology
View Article Online
DOI: 10.1039/D0CY01700K
2.2 Catalyst characterization
Nitrogen adsorption-desorption isotherms were recorded using a Quantachrome Nova
2000 instrument (Autosorb-1 software) at 77 K. Before adsorption measurements, the
sample was outgassed under vacuum at 353 K for 10 h. Specific surface area was
determined from Brunauer–Emmett–Teller (BET) equation. The pore volume was obtained
as the adsorbed nitrogen volume at the maximum relative pressure (P/P0 = 0.99). The
nonlocal density functional theory (NLDFT) method was employed to calculate the pore
size distribution (PSD). The cobalt content in each sample was determined by X-ray
fluorescence spectroscopy (XRF; Rigaku ZSX Primus II). The X-ray diffraction (XRD)
analyses were performed using a Rigaku D-Max-Ultima III diffractometer with nickel-
filtered Cu Kα radiation of wavelength 1.5406 A° (40 Ma, 40 Kv). XPS analyses were
carried out with a KRATOS Axis-Ultra electron spectrometer (UK), using an Al
monochromatic radiation operating at 300 watts. High energy resolution Co2p lines were
used for chemical analysis. These were recorded at 20 eV or 40 eV pass energy and a step
size of 0.05 eV or 0.1 eV respectively, depending on the element peak intensity. Binding
energy scale is calibrated against standard samples: Au4f7/2: 84.0eV, Ag3d5/2: 368.2eV,
Cu2p3/2: 932.6eV [31]. In addition to the absolute Co2p binding energies, the 2p1/2, 2p3/2
spin–orbital splittings and intensities of satellite structures were also used for the analysis
of XPS data. Cobalt coordination was determined using a Varian Cary 300 Bio UV-visible
spectrophotometer. The powdered samples were pelletized and loaded into a quartz cell.
Pure siliceous SBA-16 was selected as reference and the spectra were collected in the range
of 200–800 nm. Temperature-programmed reduction with H2 (H2-TPR) was carried out
using an RXM-100 analyzer (ASDI). Typically, 50 mg sample was firstly pretreated in a

Catalysis Science & Technology
Page 8 of 64
View Article Online
DOI: 10.1039/D0CY01700K
quartz reactor under flowing argon at 823 K for 2 h. After cooling to about 298 K, a H2–
Ar (5% H2) mixture flow of 30 ml/min was introduced into the reactor and the temperature
was raised to 1273 K at a rate of 10 K/min. The H2 consumption was recorded using a TCD
detector. Transmission electron microscopy (TEM) was performed on a JEOL JEM-1230
operating at 60–120 kV. The samples were prepared according to the method described
elsewhere [32].
2.3 Catalytic epoxidation experiments
2.3.1 Atmospheric pressure tests
a 100 ml three-necked round bottom flask equipped with a liquid condenser was used
for aerobic epoxidation of limonene under atmospheric pressure. In a typical reaction,
3.6 mmol of limonene, 170 mg of catalyst, 12 mmol isobutyraldehyde and 10 ml
ethylacetate as a solvent, were added to the reactor. The mixture was stirred
vigorously during the reaction for a specified period of time using a magnetic stirrer
and heated to the desired temperature on a Fisher Scientific Thermoplate equipped
with a thermocouple. The reaction was started by bubbling O2 at atmospheric pressure
into the reaction mixture at the rate of 10 ml/min. After the reaction, the catalyst was
separated by centrifugation and the liquid products were analyzed using a CP-3800
gas chromatograph (Varian Inc.) equipped with a flame ionization detector (FID).
Products were quantified based on calibration curves of identified compounds of
known retention time and confirmed by mass spectroscopy. The selectivity, yield of
epoxides, and the conversion of limonene were calculated on the basis of gas
chromatography results using Eqs. (1a) and (1b):

Page 9 of 64
Catalysis Science & Technology
View Article Online
DOI: 10.1039/D0CY01700K
Limonene Conversion:
푛푖 ― 푛푓
퐶표푛푣푒푟푠푖표푛(%) =
Product selectivity:
푆푒푙푒푐푡푖푣푖푡푦(%) =
∗ 100
(1a)
(1b)
푛푖
푛푝푟표푑푢푐푡
푛푖 ― 푛푓
∗ 100
2.3.2 High pressure tests
a high-pressure stainless steel autoclave (201C-4873 Parker Autoclave Engineers) with a
capacity of 100 mL equipped with a mechanical stirrer, an automatic temperature control
system and a sampling microvalve, was used for the pressurized liquid phase epoxidation
of limonene. In a typical run, specified amounts of limonene and isobutyraldehyde and
required amount of catalyst were charged into the autoclave. The mixture was stirred
vigorously under constant stirring at 500 rpm during the reaction for a specified period of
time and heated to the desired temperature. The reactor was then fed with oxygen at up to
145 psi pressure. During the reaction, the pressure and temperature were kept constant.
Samples were withdrawn at fixed time intervals and analyzed by GC for kinetic analysis.
3. Results and discussion
The preparation of the Co/SBA-16 catalyst by the pH adjusting method is described in
Scheme 2. Initially, mixing water, HCl, F127, butanol and TEOS is followed by stirring at
318 K for 4 h. Secondly, Co(NO3)2.6H2O is added followed by additional stirring at 318 K
for 20 h and hydrothermal aging for 48 h at 373 K. According to Zhao et al. [33], the

Catalysis Science & Technology
Page 10 of 64
View Article Online
DOI: 10.1039/D0CY01700K
absence of sufficiently strong electrostatic or hydrogen-bonding interactions at pH 2.0–7.0
leads to the formation of disordered porous silica. Therefore, the mesostructure is formed,
at low pH value (pH<2) which must be adjusted to neutral (pH=7.5) by adding ammonium
hydroxide for grafting of Co2+ species. Then, hydrothermal aging is performed for another
48 h at 373 K during which a large amount of cobalt species can be introduced into the
mesophase. Finally, Soxhlet extraction allows removal of the surfactant.
According to [26], the optimum adjusted pH value of 7.5 is high enough for Co
incorporation but low enough to avoid SBA-16 hydrolysis [28]. Lou and co-workers [34]
used sodium hydroxide and Cui and Wu used ammonia for adjusting the pH of their
synthesis system. They observed that the resultant salts (NaCl and NH4Cl) helped to
improve ordering the mesostructure of SBA-15 [26, 28].
In a previous study , several SBA-16 catalysts with different Co/Si ratios ranging from 1.1
to 9.7% were synthesized and used for epoxidation of limonene using molecular oxygen.
Catalytic activity was shown to increase as the ratio was raised up to 4.5% which is the
upper limit at which isolated tetrahedrally coordinated Co2+ species can be produced by
this method. Above this value (4.5 to 9.7%), the activity was found to decrease. Over this
Co/Si range some octahedrally coordinated Co3+ in the form of Co3O4 was also formed
[23].
Therefore, in the present study in order to examine the effect of the catalyst preparation
procedure for improving catalytic activity, the pH adjusting method was used to prepare
Co/SBA-16 using cobalt nitrate (Co(NO3)2·6H2O).

Page 11 of 64
Catalysis Science & Technology
View Article Online
DOI: 10.1039/D0CY01700K
3.1 BET analysis
As shown in Fig.1A all samples exhibit type-IV N2 adsorption/desorption isotherms
demonstrating the presence of good quality cage-like pores [35]. Pore size distribution was
calculated using the adsorption branch and the non-linear DFT method assuming a
cylindrical pore geometry. This may slightly underestimate the pore size for the spherical
cage-like geometry [29] (Fig.1B). The specific surface area and the total pore volume were
obtained from the standard BET method applied between 0.03–0.29 P/P0 and the volume
of N2 adsorption at P/P0 = 0.99. Table 1, displays the primary mesopores average diameter
(11.9 nm) of SBA-16 prepared by pH adjusting, with volume of mesopores and specific
surface area estimated to be 0.93 cm3 g−1 and 380 m2 g−1, respectively. An obvious
reduction in surface area, pore volume and pore diameter can be observed upon
introduction of cobalt. For Co/SBA-16 samples the surface area, pore volume and pore
diameter decrease could possibly be attributed to the filling of the pores with cobalt species
during the pH adjustment procedure [28]. From the results of XRF (Table 1), the actual
concentration of Co in the tested Co/SBA-16 catalysts was almost identical to the nominal
Co loading calculated from the precursor addition for Co/Si ratio lower than 9.1%.
From the results shown in Fig.1B, much larger pore diameters were obtained by the pH
adjusting preparation method compared to those by postgrafting impregnation with toluene
Co(acac)2 described in [23]. It seems that the high temperature calcination used for
template removal in postgrafting impregnation caused meaningful shrinkage of the
structure compared to template removal by soxhlet extraction used in the pH adjusting
method [35, 36]. However, the lower surface area of materials obtained by the pH adjusting
method suggests the possibility of insufficient template removal from micropores using

Catalysis Science & Technology
Page 12 of 64
View Article Online
DOI: 10.1039/D0CY01700K
only extraction. Another reason for the lower surface area might be related to the high pH
value, since a high OH- concentration speeds up the hydrolysis during SBA-16
mesostructure synthesis. [37].
The catalytic activity of the epoxidation catalyst is especially affected by its accessibility
of the reactants to the active sites. Higher cyclohexene conversion was found by Berube et
al. [38] in epoxidation of cyclohexene using calcined Ti-SBA-15 catalyst prepared at
higher hydrothermal temperature (140ºC) which enhances the formation of larger pores
and favors the diffusion of the reactants. An increased accessibility to the active sites for
the reactants was suggested to explain this result. In fact, pore enlargement occurs together
with a pronounced increase in pore volume upon increasing hydrothermal treatment
temperature and time [39]. On the other hand, low temperature hydrothermal aging could
have resulted in higher silanol density. An increased density of accessible silanol groups
on the surface of both aged and calcined at lower temperatures materials could lead to
more Co-O-Si bonds, thus yielding higher dispersion and substitution of the Co2+ sites in
tetrahedral coordination.
3.2 X-ray diffraction analysis
Fig.2A displays the X-ray diffraction patterns for the Co/SBA-16 samples (Co/Si=1.1-
8.8%). The low angle region, exhibits a single intense peak at low 2θ indexed to the
reflection of ordered mesoporous materials with Im3m cubic symmetry. As illustrated on
Fig.2A, with the increase of the cobalt loading, the peak intensity decreased. This is likely
associated to the formed dispersed cobalt oxide occupying some of the pores in the SBA-

Page 13 of 64
Catalysis Science & Technology
View Article Online
DOI: 10.1039/D0CY01700K
16 and the reduced orderliness of the catalyst structure. The symmetry of mesoporous
structure was however essentially maintained even at higher cobalt content. These results
show that a certain amount of Co can be substituted into SBA-16 by the pH-adjusting
method, without essentially destroying the structural ordering.
Only a broad feature centered at 22.2º and ascribed to the amorphous structure of SBA-16
silica framework can be observed in wide angle XRD patterns of the different Co/SBA-16
samples prepared by pH adjusting method (Fig.2B). No peaks of the crystalline cobalt
oxide were observed for Co/SBA-16 (8.8%). Co/SBA-16 with almost the same cobalt
content prepared by the method of postgrafting with Co(acac)2 in [23], showed weak peaks
at about 37◦, 45◦, suggesting the presence of some highly dispersed Co3O4 nanoclusters
which may have formed during the high temperature calcination.
3.3 Transmission Electron Microscopy (TEM)
As shown in Fig.3, TEM images illustrate well-ordered body-centered cubic (bcc) arrays
of mesopores with three dimensional channels, indicating a 3-D spherical (Im¯3m
symmetry) mesostructure. The TEM images of (A, B) Co/SBA-16 (Co/Si=1.1%) and (C,
D) Co/SBA-16 (Co/Si=8.8%) catalysts at different magnifications indicate that the typical
morphology of SBA-16 was well retained during pH adjustment and cobalt incorporation.
The result coincides with the results obtained from the XRD studies and demonstrates that
the pH-adjusting method allows maintaining the highly ordered mesostructure with high
dispersion of cobalt species even when a rather large amount of Co is grafted.

Catalysis Science & Technology
Page 14 of 64
View Article Online
DOI: 10.1039/D0CY01700K
3.4 Diffuse Reflectance UV-vis spectroscopy (DR-UV)
The UV/vis spectra of SBA-16 and Co-incorporated SBA-16 catalysts are depicted in
Fig.4. All samples showed three overlapping absorption peaks at around 530, 599 and 648
nm which are especially distinct in the spectrum of the low Co (1.1%) catalyst. Increasing
cobalt content led to gradual decrease in relative intensity of peak at 648 nm, followed by
appearance of another peak around 480 nm.
Typically, three kinds of Co species may form during the decomposition of the cobalt
precursor. Tetrahedral and octahedral Co2+ in CoO and octahedral Co3+ in Co3O4 are the
main species present in supported Co catalysts [40]. Since only the isolated tetrahedral Co2+
species contribute to the epoxidation of alkenes [12], it is desired to obtain materials with
high content of 4-coordinated Co species. As can be seen from this figure, SBA-16 spetrum
exhibits no absorption peak in the wavelength region of 400–800 nm. The three line pattern
in the region of 550–700 nm was repeatedly ascribed to 4 A 2 ®4 T 1 transition of Co 2+ ions
in tetrahedral configuration (520, 580 and 655 nm in [13]; 540, 570 and 615 in [12, 41]).
Therefore the presence of isolated tetrahedrally coordinated Co2+ in the framework of
SBA-16 is ascertained in these supported catalysts [42]. The 480 nm band is typical of
Co2+ in octahedral symmetry and can be attributed either to [Co(H2O)6]2+ or to
[Co(H2O)5OH]+ l located in the lattice framework of SBA-16 [43, 44]. Unlike the results in
our previous paper using Co(acac)2 impregnated Co/SBA-16, the absence of absorption
peaks at 300-400 nm and 710 nm, which could be assigned to the absorbance of Co3O4, the
absence of octahedral Co3+ even at the highest cobalt content (Co/Si=8.8%) is confirmed
[23]. This result indicates that not only a large amount of Co species could be introduced

Page 15 of 64
Catalysis Science & Technology
View Article Online
DOI: 10.1039/D0CY01700K
into the pure silica SBA-16 by pH-adjusting, but also that the Co species introduced by this
method are essentially located at 4-coordinated sites, especially at lower cobalt content.
3.5 X-Ray Photoelectron Spectroscopy (XPS)
High-resolution Co 2p photoelectron lines of the cobalt containing catalysts are reported
in Fig.5. There are no clear differences between the Co 2p spectra of Co/SBA-16 samples
with different cobalt loadings. The binding energy of Co 2p3/2 exhibits an evident peak at
781.5 eV in all spectra. The cobalt species with such a binding energy could be assigned
to Co2+ in an isolated state. Moreover, all of the prepared samples show a very high
intensity of the shake-up satellite peak centered at around 787.5 eV that could be related to
the presence of some Co2+ ions in octahedral symmetry typical of CoO [45]. The satellite
peak position of the Co2+compound is located at higher binding energy than the main peak
of Co 2p3/2 by 5–6 eV. In addition, the satellite peak position of the Co3+ compound would
be higher than the main peak by 10–11 eV and weaker [46]. Therefore, these results clearly
indicate that most of the surface cobalt species in these catalysts are under the form of
Co2+, which is consistent with the UV/vis, XRD and TEM results. This is in disagreement
with the results of Wang et al. [46] who observed the characteristic peak of Co3O4 in the
XPS lines of their catalysts prepared by incipient wetness impregnation of SBA-15 with
cobalt nitrate hexahydrate at relatively low Co loading. At pH higher than that of its PZC
(pH 3), the silica surface is negatively charged and there is a certain attraction between
cobalt ions and the surface. XPS shows uniform distribution of cobalt ions on the surface
under these preparation conditions. At pH value below that of the silica PZC, Co2+ ions are

Catalysis Science & Technology
Page 16 of 64
View Article Online
DOI: 10.1039/D0CY01700K
repelled by positively charged silica surface and contributes to the non-uniform cobalt
distribution and consequently lower cobalt dispersion [47].
In the catalysts discussed in [23], prepared by impregnation of cobalt acetylacetonate, the
Co 2p3/2 peak appeared at higher binding energy (BE = 782 eV) than that of the equivalent
catalyst (Co/Si=1.1% see below) prepared from cobalt nitrate (BE = 781.5 eV). Although
the results of both catalysts indicate that most of the surface cobalt species are in the form
of Co2+, these Co2+ ions in the acetylacetonate prepared samples have a stronger interaction
with the SBA-16 support probably in the form of a cobalt silicate phase. This hypothesis
could be supported by the higher temperatures required for decomposing the organic cobalt
salt precursors as compared to cobalt nitrate [45].
3.6 Hydrogen Temperature Programmed Reduction (H2-TPR)
Temperature programmed reduction profiles (Fig.6) show the reduction behavior of air
exposed materials at different Co/Si ratios, which reflects the electronic interaction of the
supported phase with the carrier.
For comparison, the H2-TPR line of Co3O4 is reported. It shows only one peak centered at
673 K indicating that the Co2+ and Co3+ ions in the Co3O4 lattice are essentially reduced
simultaneously. A broad feature observed between 623 and 773 K for all catalysts tested,
should be associated with some Co2+ species since both XRD and XPS indicate absence of
Co3+ in these samples. These species may likely be in a coordination environment similar
to that of the Co ions in the Co3O4 lattice namely associated oxidic cobalt clusters and in
octahedral coordination.

Page 17 of 64
Catalysis Science & Technology
View Article Online
DOI: 10.1039/D0CY01700K
At higher temperature the catalysts H2-TPR lines exhibit an intense series of peaks around
923-1123 K. A distinct line at about 973 K is dominant in the low cobalt content material
and does not seem to change in intensity at higher cobalt loading. It may be safely assumed
that this particular peak is due to the active epoxidation sites, namely the isolated Co2+ ions
in tetrahedral coordination.
The broad feature between 973 and 1123 K is obviously associated with Co2+ species
efficiently protected against reduction by a strong interaction with the silica lattice. It may
therefore be hypothesized that these Co2+ions are involved in silicate like associations and
do not act as active sites in the epoxidation process.
The above results are in contrast with those obtained by Martínez et al. [45] who observed
the formation of Co3O4 in samples prepared by impregnation of SBA-15 with cobalt nitrate.
The results reported in Fig.6 are also in contrast with the H2-TPR observed for samples
obtained by impregnation of SBA-16 with cobalt acetylacetonate in our previous study
[23]. In this previous study a large reduction peak observed at 673 K at the higher cobalt
content (Co/Si=9.7%) was ascribed to the reduction of Co3+ surface species.
3.7 Investigation of limonene aerobic epoxidation
In [23] surface response methodology was implemented in order to find the optimal
catalytic conditions of limonene epoxidation over Co/SBA-16 and provide a general
overview of the effects of the main test parameters on limonene conversion and yields
of mono- and diepoxide. The estimated reaction conditions for maximum yield of
epoxides were as follows: 0.3 mmol/ml limonene concentration, 1 mmol/ml

Catalysis Science & Technology
Page 18 of 64
View Article Online
DOI: 10.1039/D0CY01700K
isobutyraldehyde concentration (IBA/limonene ratio=3.3) and 14.3 mg/ml catalyst
amount at 301 K for 200 min. Since surface methodology is based on a curve fitting
process, it is not a substitute for a thorough kinetic analysis of the reaction. In the
present part of this study, Co/SBA-16 at different Co loadings prepared by the pH
adjusting method, were evaluated as catalysts for epoxidation of limonene. The tests
were performed under atmospheric and higher oxygen pressure, using the optimum
reaction conditions established previously, and an extensive kinetic study was
performed.
3.7.1 Limonene epoxidation at atmospheric pressure
Table 2 summarizes the catalytic performance of Co/SBA-16 at different cobalt loadings
under optimum reaction conditions. The best epoxide yield (85%) was achieved using
Co/Si=1.1%. Increasing cobalt loadings (Co/Si ratio) to 8.8% decreased the yield to 65%.
These results suggest that the higher epoxide selectivity is associated with the higher
density of isolated tetrahedral Co2+ species with some parallel limonene oxidation being
catalyzed by the higher coordination Co2+ ions identified by UV-vis, XPS and TPR
analyses. Moreover, the catalytic results showed the highest optimized epoxide yield over
Co/SBA-16(1.1%) prepared by pH adjusting method which is a lower cobalt content,
compared to the optimal Co/SBA-16(4.5%) prepared by post-grafting of Co(acac)2 reached
in [23]. The superiority of this catalyst prepared by the pH adjusting method may be related
to the higher surface density of silanol groups accessible to the Co2+ ions in the SBA-16
support. This higher silanol density results from no calcination being applied before
introduction of the cobalt precursor. At strongly acidic pH, the Co2+ions in the SBA-16
mesopores do not link to the silica surface. Adjusting the pH value to 7.5 results in the

Page 19 of 64
Catalysis Science & Technology
View Article Online
DOI: 10.1039/D0CY01700K
transformation of these ions to an oxo form, condensation with adjacent silanols and
anchoring of metal atoms to the mesopore walls.
The high content of low coordination metal ions and the well ordered mesostructure are
both in line with the previous observations of Wu et al. [28] for the synthesis of Al-SBA-
15 and Ti-SBA-15 and Cui et al. [26] for Co-SBA-15 syntheses by the pH adjusting
method. In addition, the high catalytic performance observed at 1.1% Co/Si ratio which
was only reached at 4.5% in the postsynthesis prepared catalysts (see Table 5 in [23]) is
likely associated with the intrawall micro/meso pores being filled with remaining template
in the pH adjustment prepared material (see Fig.1B). This suggests that during the latter
synthesis, Co is not grafted within these intrawall pores so that similar surface density of
tetrahedral Co in the main channels are reached in these two samples at different overall
Co contents. Consequently, it is suggested that the observed high selectivity of the
Co/SBA-16 catalysts prepared by the pH adjusting method are associated with the high
dispersion of isolated tetrahedral cobalt(II) species on the surface of the main channels of
highly ordered mesoporous SBA-16.
The results in Table 2 also show a decreasing epoxide yield at increasing cobalt loading.
Since no other oxidized compound was detected among the products in liquid phase, it may
be hypothesized that some complete oxidation to CO2 was happening in spite of the low
reaction temperature. Cobalt oxides are known to be among the best total oxidation
catalysts. These results may therefore mean that the surface oxidic cobalt clusters whose
density increases with cobalt loading would catalyze this oxidation.

Catalysis Science & Technology
Page 20 of 64
View Article Online
DOI: 10.1039/D0CY01700K
3.7.2 Limonene epoxidation under oxygen pressure
The reaction was investigated at increasing oxygen pressure, the rest of the parameters
being kept at their optimum values. As seen from Fig.7, after 200 min the conversion of
limonene is almost complete. Epoxidation of limonene with oxygen produces mainly 1,2-
limonene epoxide and some 8,9-limonene epoxide was also observed after 30 min of
reaction (not shown). Because of the electrophilic nature of the oxidant, epoxidation first
occurs at more substituted double bonds therefore 1,2-limonene oxide is the dominant
product. Similar observations were reported in reference [48] for limonene epoxidation
using H2O2 as the oxidizing agent and in references [2, 22] for limonene epoxidation with
dimethyldioxirane and tert-butyl hydroperoxide respectively.
The selectivity of 1,2 limonene oxide was increased by increasing pressure from 15 to 44
psi, then decreased with further increasing up to 116 psi whereas the selectivity towards
limonene dioxide increased with enhancing oxygen pressure up to 58 psi and then
decreased which could be possibly due to the secondary oxidation or ring opening of
epoxide at higher pressure. Consequently, the highest yield of 95% limonene epoxides was
achieved under the optimized pressure of 44 psi. Almost complete limonene transformation
into its epoxides was achieved using Co-SBA/16(1.1%) under 44 psi oxygen pressure. This
dependency of epoxide conversion and selectivity on pressure was also witnessed by Sha
et al. [49] in epoxidation of cyclooctene using Co(II) coordinated metal-organic framework
and acetonitrile as the solvent. Schuster and Hölderich also mentioned that at atmospheric
pressure the system did not work properly, because due to the low pressure the oxygen
uptake of the limonene and α-pinene in fluorobenzene solution were erratic and the
catalytic results could not be reproduced [50].

Page 21 of 64
Catalysis Science & Technology
View Article Online
DOI: 10.1039/D0CY01700K
In our kinetic studies of limonene epoxidation in the presence of isobutyraldehyde and
dioxygen at 44 psi oxygen pressure to produce epoxidized limonene with Co/SBA-16
(1.1%) as catalyst, the experimental data for initial rates presented in Tables (3-7) and Figs.
(9-13) were used. An example of experimental monitoring of monoepoxides and diepoxide
concentrations is reported in Fig. 8. This shows that diepoxide is produced by secondary
epoxidation of monoepoxides. The total yield of epoxides under optimized condition is
58% selectivity of 1,2 limonene and 37% limonene dioxide for the total consumption of
limonene. After 250 min, limonene is entirely converted and the concentrations of epoxides
start deceasing, presumably upon secondary oxidation. Since no other oxidation product is
detected in the liquid phase, this oxidation may be complete yielding CO2. In Supporting
Information (Fig. S1) we provide C13 NMR spectrum of the sample prepared by aerobic
epoxidation of limonene over Co-SBA-16(1.1%) catalyst under optimum condition, which
shows that the two trans isomers of LDO as well as the two cis isomers are in similar
contents in the product.
The recycling tests of the catalyst Co/SBA-16 (1.1%) were performed under repeated runs,
while the reaction conditions kept unchanged in their optimum conditions (the same
conditions as used in Fig. 8). Between cycles the solid catalyst was recovered from the
reaction mixture and washed with ethanol then dried and used in the next run. For three
repeated cycles the conversion and selectivity remained almost unchanged. Moreover, no
reaction was observed after catalyst filtration out of the reaction medium. These
observations suggest the stability and recyclability of the catalyst Co/SBA-16 (1.1%)
prepared by the pH adjusting method.

Catalysis Science & Technology
Page 22 of 64
View Article Online
DOI: 10.1039/D0CY01700K
3.8 Kinetics of aerobic epoxidation of limonene over Co/SBA-16 in
the presence of isobutyraldehyde
The mechanistic pathway of this epoxidation depends critically on the fate of adsorbed
RCHO and its interation with oxygen. There are two visions of this fate reported in
literature. One school advocates rapid decomposition [23, 51, 52] as:
(푆푖푂)4퐶표2 + ⋯푅퐻퐶 = 푂→(푆푖푂)4퐶표2 + + 푅퐶푂• + 퐻•
(2)
푅퐶푂•
The produced
radical being released to the liquid phase where it reacts with oxygen:
푅퐶푂• + 푂2→ 푅퐶푂•3
(3)
As shown in Scheme 3 , the peroxy radical may then either react with limonene in the liquid
phase (pathway A) or react with the tetrahedral Co2+ site yielding an adsorbed peroxy
species on the oxidized Co3+. The interaction of this specie with limonene produces the
epoxide (pathway B).
Some other researchers have proposed that a different surface peroxide is formed by
oxidative insertion of the oxygen molecule in the adsorbed aldehyde complex [53, 54]:
(푆푖푂)4퐶표2 + ⋯푅퐻퐶 = 푂 + 푂2→(푆푖푂)4퐶표3 + 푂푂 ― + 푅퐶퐻푂
(4)
A wide variety of Co2+ complexes are well known to bind and activate dioxygen in the
· -
form of a superoxo-Co3+ complex (Co3+-O2 ) [55-57]. This cobalt superoxo complex was
introduced by earlier studies as the active epoxidizing agent for epoxidation of alkenes [12,
13, 53, 54].
In Scheme 4, we propose a variant of the literature proposed mechanism which uses this
peroxy intermediate (pathway C).

Page 23 of 64
Catalysis Science & Technology
View Article Online
DOI: 10.1039/D0CY01700K
In the text below we will first report extensive initial rate data for limonene consumption
(Rl), 1,2- and 8,9-monoepoxide generation (R1,2e and R8,9e) and isobutyraldehyde oxidation
(Ra). These data will then be discussed in terms of the rate equations developed for
pathways A to C and shown in supporting information.
When the rate determining step (RDS) is assumed to be the homogenous reaction in the
bulk solution (mechanism pathway A), the total rate of limonene epoxidation (Re) and
isobutyraldehyde oxidation (Ra) can be described by Eqs. (5) and (6) respectively. A
detailed derivation of these equations can be found in the Supporting Information.
1
1
2
푅푒 = ― 0.16퐾2[퐿]2 ―0.8퐾1[푆][퐿] + 0.8퐾3 [푅퐶퐻푂]2 퐶표
(5)
2 +
[
]
[퐿]
1
2
3
푅푖 = 퐾4[푆][푅퐶퐻푂] ― 0.2퐾5[퐿][푅퐶퐻푂] + 퐾6 [푅퐶퐻푂]2 [퐶표2 +
]
(6)
푅퐶푂•
Where [S] is the density of surface species responsible for
various constants are explicited in supporting information.
termination step. The
3
If the rate determining step (RDS) is assumed to be the surface reaction between surface
peroxy complex and limonene (step 3 in pathway B, (Eq. (S25) in Supporting
Information), then the rate of reaction at low limonene concentration can be obtained as:
2 +
[
]
[퐿]
푅푒 = ― 푅푖 = 퐾푐[푅퐶퐻푂] 퐶표
(7)
Where,
[
]
푘푎푘푏푘푐 푂2
퐾푐 =
[
]
푘푏푘 ―푏 푂2
(
[
]
)
푘 ―푎 + 푘푏 푂2 푘 ―푏
―
(
)
]
[
푘 ―푎 + 푘푏 푂2

Catalysis Science & Technology
Page 24 of 64
View Article Online
DOI: 10.1039/D0CY01700K
For pathway C assuming that the RDS is the oxidative insertion of oxygen in the adsorbed
aldehyde (step 2 in Scheme 4, Equation (S42) in supporting information):
훼1[푂2]
2 +
[
퐶표
]
[푅퐶퐻푂]
푅푒 =
(8)
훼2 + 훼3[푂2]
which, assuming formation of complex C (Equation (11 below)) may also be written as:
훼1[푂2]
[푅퐶퐻푂]0
푘푐
2 +
[
퐶표
]
푅푒 =
(9)
훼2 + 훼3[푂2]
1 + [퐿]
푘푐′
Figs (9-13) and Tables (3-7) indicate the variation of the initial epoxidation rate (Re = R1,2e
+ R8,9e)) and initial limonene and isobutyraldehyde consumption rates (Rl, Ra), which can
be measured from the slope of tangent of initial limonene monoepoxides, limonene and
isobutyraldehyde concentrations variations as functions of time.
3.8.1 Effect of isobutyraldehyde concentration
As shown in Table 3, these results revealed that increasing the concentration of
isobutyraldehyde from 0.33 to 2.6 M resulted in an increase of initial limonene
consumption and epoxidation of limonene as well as of initial isobutyraldehyde oxidation
rates as illustrated in Fig 9(a), (b).
The main features of these data are:
1) Whereas from Fig. 9b isobutyraldehyde oxidation rate (Ra) is consistent with
Equation 6, the rate of limonene epoxidation (Re) does not follow Equation 5 as the
curve Re ([RCHO]) is linear. Thus only the RCHO oxidation RDS can be a
푅퐶푂•
homogeneous radical process involving reaction of
3 (pathway A).

Page 25 of 64
Catalysis Science & Technology
View Article Online
DOI: 10.1039/D0CY01700K
2) Re being systematically higher than Ra implies that pathway B cannot be the process
for limonene epoxidation. Pathway B would impose Ra = Re and a simultaneous
homogeneous reaction consuming only RCHO would impose Ra ≥ Re.
3) Re < Rl also expressed as a = Re/Rl < 1 (see supporting information) indicates that
some minor reaction parallel to epoxidation is consuming limonene. Part of this
consumption does not involve the aldehyde (Rl = 0 at [RCHO] = 0) but another part
also consumes the aldehyde, not generating epoxide ( at 0.33 M ≥ [RCHO] ).
3.8.2 Effect of the catalyst concentration
Experimental data obtained at varying catalyst content, expressed as a concentration of
Co2+ ions, are reported in Table 4 and Fig. 10. It appears that :
1) The initial rate of isobutyraldehyde oxidation does not depend on catalyst
concentration confirming that this reaction is a homogeneous process. Also Re ≥ Ra
as soon as there is catalyst in the system.
2) In the presence of catalyst both Rl and Re vary linearly with its concentration and
the two lines being parallel indicates that contrary to epoxidation the rate of extra
consumption of limonene does not vary with catalyst content. Thus whereas
epoxidation appears as a heterogeneous catalytic reaction, this extra consumption
is through a homogeneous reaction.
3) Both Ra and Rl have measurable initial rates in absence of catalyst.
3.8.3 Effect of limonene concentration
Table 5 and Figure 11 show the initial rate data obtained at increasing limonene
concentration from 0.09 to 0.94 M. All three curves of Rl, Re and Ra versus [L] show the

Catalysis Science & Technology
Page 26 of 64
View Article Online
DOI: 10.1039/D0CY01700K
very unusual decreasing trend as functions of the main reactant concentration. This would
correspond to a negative order of reaction. Similar trend for conversion by aerobic
epoxidation was previously reported for limonene in [23] and oleic acid in [58]. Other
significant aspects of these results include:
1) Rl ≥ Re over the whole range of [L] with the curves merging asymptotically at high
limonene concentration.
2) Ra < Re at all values of [L].
Summing up the conclusions made at this point, the mechanism must include pathway A
for the aldehyde oxidation and a pathway which, like C, would allow a primary epoxidation
rate (Re) higher than aldehyde oxidation rate (Ra). The pathway C described in Scheme 4
does not however predict values of Re and Ra decreasing at increasing [L]. In order to
resolve this contradiction it may be proposed that there is a competition between the
aldehyde adsorption (step 1 in Scheme 4).
(푆푖푂)4퐶표2 + + 푅퐶퐻푂 ⇌ (푆푖푂)4퐶표2 + 퐶푅퐻푂
(10)
and formation of a complex with limonene
푘푐
(11)
⇌
+ 푅퐶퐻푂
푘푐′
The enhanced competition of this complex with the adsorption process would explain the
decrease in both epoxidation and aldehyde oxidation rates upon increasing limonene
concentration.

Page 27 of 64
Catalysis Science & Technology
View Article Online
DOI: 10.1039/D0CY01700K
As shown in Fig. 11 this varation is correctly represented by Equation (12) which is derived
from Equation (9).
푎
푅푒 =
(12)
1 + 푏[퐿]
Therefore, by curve fitting of experimental data parameters a and b with an average error
±11.4% were estimated to be 5.3 10-2 M min-1 and 0.65 M-1, respectively.
3.8.4 Effect of temperature
The dependence of epoxidation reaction rate to temperature was studied between 293 to
331 K, as indicated in Table 6 and Fig. 12. The Arrhenius plot of limonene epoxidation
reported in Fig.12 yields an activation energy of (22 kJ/mol) which is much less than the
activation energy (48 kJ mol-1) reported by Wentzel et al. for α-pinene epoxidation using
homogeneous nickel(II)-β-diketonate complexes [59]. From Eyring equation, the
parameters ∆H‡ and ∆S‡ were calculated to be 19.4 kJ mol-1 and -169.5 J ºK-1 mol-1,
respectively. This value of DH, is quite close to our experimental value. The large negative
value of ∆S‡ indicates a rigid transition state for the rate-determining step.
Since this low activation energy is suggestive of a rate being limited by internal diffusion
of the reactant a test was performed on a catalyst sample obtained by grinding followed by
sedimentation of the coarser particles in order to isolate a population of smaller particles
of the same catalytic material. The average particle size was decreased from 20 to 5
microns. The results of that test are displaid in Table 7. They show essentially unchanged
values of the three initial reaction rates.These are clear proofs that none of these rates is
mass transfer limited under the reported conditions.

Catalysis Science & Technology
Page 28 of 64
View Article Online
DOI: 10.1039/D0CY01700K
The Arrhenius plot for isobutyraldehyde oxidation shows a change in slope around 331 K
usually associated with a change in RDS from chemical reaction control at lower
temperature to control by some mass transfer process at higher temperature. The latter may
be oxygen transfer from the gas phase which would be lowered by the decrease in oxygen
solubility at increasing temperature.
These differences between the two reactions thermal behaviour confirms the above
discussed differences in mechanisms.
3.8.5 Effect of oxygen Pressure
The epoxidation reaction was investigated at changing initial oxygen pressure over the
range of 44 to 145 psi. The dependence of epoxidation reaction against oxygen pressure is
depicted in Fig.13.
Increasing oxygen pressure from 87 to 145 psi almost did not have any influence on
epoxidation rate (Table 8). Sheldon in Ref. [60], mentioned oxygen pressure will affect
the ratio of epoxide to allylic by-product by changing the ratio of addition to abstraction of
radical to the olefin double bond. It is therefore suggested that oxygen pressure
enhancement particularly at initial reaction stage, increases the addition of peroxy radical
to the double bond of limonene (steps 2 to 4 in Scheme 4) instead of hydrogen abstraction
to produce peracid and allylic radicals. Hence, the rate of epoxide formation increases with
rising oxygen pressure. As shown in Fig. 13 this varation is correctly represented by
Equation (13) which is derived from Equation (9) assuming Henry’s law.
푎푃푂2
푅푒 =
(13)
1 + 푏푃푂2

Page 29 of 64
Catalysis Science & Technology
View Article Online
DOI: 10.1039/D0CY01700K
By curve fitting of experimental data parameters a and b with an average error ±9.6% were
estimated to be 0.12 10 -2 M min-1psi-1 and 0.013 psi-1, respectively.
On the other hand, increasing oxygen pressure may cause secondary oxidation or ring
opening of epoxide during the time course of the reaction. Consequently, as discussed
earlier, under our reaction conditions the highest yields of limonene epoxides were
achieved under the optimized pressure of 44 psi.
Conclusions
A novel series of practical and green heterogeneous catalysts has been introduced in this
paper for the aerobic epoxidation of limonene by coordination unsaturated Co(II) supported
on mesostructured SBA-16. The preparation of these materials by the pH adjusting method
was thoroughly investigated showing that this method allowed selectively producing the
active sites, namely isolated tetrahedrally coordinated Co2+ ions in the silica support. This
new method affords cage-like materials of larger pore volumes and pore sizes making them
more attractive for catalysis involving diffusion of bulky molecules such as limonene.
The mesoporous Co/SBA-16 (Co/Si=1.1%) prepared by the pH adjusting method using
cobalt(II) nitrate showed 85% yield of limonene epoxides under optimum conditions at
atmospheric pressure and ambient temperature. This catalyst exhibited almost (∼100%)
limonene epoxides yield under 44 psi oxygen pressure.
Using the optimized catalyst a thorough investigation of initial reaction rates allowed
making various conclusions about the mechanism of aerobic epoxidation of limonene. The