Toward understanding the unusual reactivity of mesoporous niobium silicates in epoxidation of C[dbnd]C bonds with hydrogen peroxide

Article abstract of DOI:10.1016/j.jcat.2017.09.011

Niobium-containing mesoporous silicates reveal superior activity and selectivity in epoxidation of alkenes using hydrogen peroxide as a green oxidant and, in contrast to mesoporous titanium silicates, catalyze epoxidation of both electron-rich and electron-deficient C[dbnd]C bonds. This report describes a kinetic and mechanistic investigation of epoxidation of two representative substrates, cyclooctene (CyO) and 2-methyl-1,4-naphthoquinone (MNQ), over two mesoporous niobium silicates with predominantly di(oligo)meric or isolated Nb(V) sites. The observed kinetic regularities did not depend on the state of Nb but were strictly determined by the nature of the organic substrate. The rate law established for CyO is consistent with a mechanism that involves interaction of H₂O₂ with Nb(V) sites to give a hydroperoxo species NbOOH and water, followed by oxygen transfer to a nucleophilic C[dbnd]C bond, producing an epoxide and regenerating the initial state of the catalyst. This mechanism is strongly supported by stereospecificity in epoxidation of cis-alkenes and high heterolytic pathway selectivity in the oxidation of cyclohexene. The NbOOH species is manifested by an absorption feature at 307 nm in diffuse reflectance UV–vis spectra. The addition of a base (NaOAc) causes a shift of the absorption band to 293 nm and produces a rate-retarding effect on the epoxidation reaction. Several lines of evidence, including zero reaction order in MNQ, rate-accelerating effect of base, detection of acetamide in the reaction mixture, negligible reaction rate in ethyl acetate, and recognition of weak basic sites in the niobium silicates using infrared spectroscopy of adsorbed CDCl₃, all indicate that MNQ epoxidation proceeds by another mechanism that involves rate-limiting oxidation of the solvent molecule (MeCN) to generate peroxycarboximidic acid, which reacts with electron-deficient C[dbnd]C bonds, producing an epoxy derivative and acetamide.

Full text of DOI:10.1016/j.jcat.2017.09.011

Journal of Catalysis 356 (2017) 85–99
Contents lists available at ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
Toward understanding the unusual reactivity of mesoporous niobium
silicates in epoxidation of C@C bonds with hydrogen peroxide
Irina D. Ivanchikova a, Igor Y. Skobelev a,b, Nataliya V. Maksimchuk a,b, Eugenii A. Paukshtis a,b
,
Mikhail V. Shashkov a,b, Oxana A. Kholdeeva a,b,
⇑
a Boreskov Institute of Catalysis, Lavrentieva Avenue 5, Novosibirsk 630090, Russia
b Novosibirsk State University, Pirogova Street 2, Novosibirsk 630090, Russia
a r t i c l e i n f o
a b s t r a c t
Article history:
Niobium-containing mesoporous silicates reveal superior activity and selectivity in epoxidation of alke-
nes using hydrogen peroxide as a green oxidant and, in contrast to mesoporous titanium silicates, cat-
alyze epoxidation of both electron-rich and electron-deficient C@C bonds. This report describes a
kinetic and mechanistic investigation of epoxidation of two representative substrates, cyclooctene
(CyO) and 2-methyl-1,4-naphthoquinone (MNQ), over two mesoporous niobium silicates with predom-
inantly di(oligo)meric or isolated Nb(V) sites. The observed kinetic regularities did not depend on the
state of Nb but were strictly determined by the nature of the organic substrate. The rate law established
for CyO is consistent with a mechanism that involves interaction of H2O2 with Nb(V) sites to give a
hydroperoxo species NbOOH and water, followed by oxygen transfer to a nucleophilic C@C bond, produc-
ing an epoxide and regenerating the initial state of the catalyst. This mechanism is strongly supported by
stereospecificity in epoxidation of cis-alkenes and high heterolytic pathway selectivity in the oxidation of
cyclohexene. The NbOOH species is manifested by an absorption feature at 307 nm in diffuse reflectance
UV–vis spectra. The addition of a base (NaOAc) causes a shift of the absorption band to 293 nm and pro-
duces a rate-retarding effect on the epoxidation reaction. Several lines of evidence, including zero reac-
tion order in MNQ, rate-accelerating effect of base, detection of acetamide in the reaction mixture,
negligible reaction rate in ethyl acetate, and recognition of weak basic sites in the niobium silicates using
infrared spectroscopy of adsorbed CDCl3, all indicate that MNQ epoxidation proceeds by another mech-
anism that involves rate-limiting oxidation of the solvent molecule (MeCN) to generate peroxycarbox-
imidic acid, which reacts with electron-deficient C@C bonds, producing an epoxy derivative and
acetamide.
Received 29 June 2017
Revised 15 September 2017
Accepted 15 September 2017
Keywords:
Epoxidation
Heterogeneous catalysis
Hydrogen peroxide
Mesoporous niobium silicate
Kinetics
Mechanism
Ó 2017 Elsevier Inc. All rights reserved.
1. Introduction
throughs in the field of heterogeneous liquid-phase oxidation with
H2O2 [8–10]. However, the scope of application of TS-1 is limited to
Alkene epoxidation is one of the most important transforma-
tions in the manufacture of bulk and fine chemicals/intermediates
[1,2]. Traditionally, epoxidation in organic synthesis is accom-
plished through oxidation with organic peracids or by the chloro-
hydrin route. Since these approaches are not environmentally
benign, great scientific effort has been directed to the development
of effective catalytic methods for epoxidation using aqueous
hydrogen peroxide as a green oxidant, which produces water as
the sole byproduct [3–7]. The discovery of titanium silicalite-1
(TS-1) by Eni researchers has become one of the major break-
oxidation of relatively small organic substrates, which are able to
penetrate into the micropores of this catalyst. A range of meso-
porous titanium silicates have been developed, but most of them
suffer from the gradual aggregation of Ti sites while using aqueous
H2O2 as oxidant and are selective only with electron-rich olefins
containing no allylic hydrogen atoms [11,12]. Therefore, the quest
for catalysts that would be both effective in activation of H2O2 and
broad in scope remains a challenging goal of oxidation catalysis.
In recent years, niobium-containing catalysts have attracted
growing interest in the selective oxidation community. Meso-
porous niobium silicates prepared by different methodologies
turned out to be active and recyclable catalysts for epoxidation
of various olefins using aqueous hydrogen peroxide as oxidant
[13–30]. In general, these catalysts reveal higher epoxidation selec-
⇑
Corresponding author at: Boreskov Institute of Catalysis, Lavrentieva Avenue 5,
Novosibirsk 630090, Russia.
E-mail address: khold@catalysis.ru (O.A. Kholdeeva).
https://doi.org/10.1016/j.jcat.2017.09.011
0021-9517/Ó 2017 Elsevier Inc. All rights reserved.

86
I.D. Ivanchikova et al. / Journal of Catalysis 356 (2017) 85–99
tivity than mesoporous titanium catalysts [21,30], and in addition,
they are able to epoxidize electron-deficient C@C bonds in ,b-
and solvents were the best reagent grade and were used without
further purification. The concentration of H2O2 (30% in water)
was determined iodometrically prior to use. Deionized water
(EASY pure, RF, Barnsted) was used for the preparation of catalysts.
Mesoporous niobium silicates containing mainly di(oligo)meric
(catalyst A) or isolated (catalyst B) Nb(V) sites were prepared by
the EISA technique using niobium(V) ethoxide and ammonium
niobate(V) oxalate hydrate, respectively [29]. For comparison, a
mesoporous titanium silicate with oligomeric Ti(IV) sites (catalyst
A-Ti) was prepared by the same EISA approach, following a proto-
col reported previously [41]. The catalysts were characterized by
elemental analysis, low-temperature N2 adsorption, FTIR, and DR
UV–vis techniques.
a
unsaturated carbonyl compounds [24,29]. Another interesting fea-
ture of the Nb catalysts is unusual regioselectivity toward epoxida-
tion of the less electron-rich exocyclic C@C bond in terpenes
[21,24]. Recently, Ivanchikova et al. have demonstrated that the
crucial factor that governs the rate and regioselectivity of limonene
epoxidation over mesoporous niobium silicates is the nature of the
solvent [29]; the highest ratio of exo/endo epoxides was obtained
in acetonitrile. Hence, the catalytic performance of mesoporous
niobium catalysts differs markedly from that of their Ti counter-
parts. Isolated and undercoordinated Nb(V) species were assumed
to be responsible for the high catalytic activity of niobium silicates
in
selective
oxidation
with
hydrogen
peroxide
[13,15,21,23,27,28,31,32]. The presence of Nb2O5-like nanoaggre-
gates was detrimental to the catalytic performance [23,24].
Recently, a direct comparison of the behavior of catalysts with
site-isolated and evenly dispersed di(oligo)meric Nb centers at
similar Nb loadings has shown that both types of catalysts are able
to accomplish epoxidation of both electron-rich and electron-
deficient C@C bonds [29]. Catalysts with predominately isolated
Nb centers were preferable for the selective formation of epoxides
sensitive to ring-opening and overoxidation, whereas single-site
and (di)oligomeric Nb sites were equally effective for the produc-
tion of relatively stable epoxides [29]. While high levels of under-
standing have been achieved in the areas of epoxidation over Ti, V,
Mo, W, and Re catalysts [10,11,33–35], as well as biomimetic cat-
alysts based on Fe and Mn [33,36,37], epoxidation over Nb is still
poorly understood. So far, few research groups have attempted to
rationalize the catalytic activity of niobium in H2O2-based oxida-
tion [30–32,38–40]. Detailed mechanistic information, in particu-
lar provided by kinetic and spectroscopic studies, is still limited,
although it is crucial for understanding the factors that control cat-
alytic performance.
In this work, we have studied the kinetics of H2O2-based Nb-
catalyzed epoxidation using two model substrates, cyclooctene
(CyO, electron-rich C@C bond) and 2-methyl-1,4-naphthoquinone
(MNQ, electron-deficient C@C bond), and two mesoporous nio-
bium silicate catalysts that contained predominantly site-isolated
or di(oligo)meric Nb(V) and were prepared by the same
evaporation-induced self-assembly (EISA) methodology using dif-
ferent Nb precursors. The results of the kinetics study, coupled
with additive and solvent effects, implied different mechanisms
for epoxidation of electron-rich and electron-deficient C@C bonds
over Nb(V). Diffuse reflectance (DR) UV–vis spectroscopic study
was employed to probe peroxo niobium species formed upon
interaction with H2O2. To understand better the difference in the
catalytic performance of mesoporous Nb and Ti silicates, we also
2.2. Infrared with adsorption of probe molecules
FTIR spectra were recorded on a Shimadzu FTIR 8300 spectrom-
eter with a resolution of 4 cmꢁ1, accumulating 200 scans. The cat-
alysts were pressed into thin wafers (12–25 mg/cm2) and calcined
directly in the IR cell at 450 °C in vacuum for 1 h. The spectra were
normalized to the wafer weight and presented in arbitrary units.
Pyridine was adsorbed at 180 °C for 15 min and then evacuated
at the same temperature for 30 min. The IR spectra were measured
at room temperature. CO was adsorbed at 77 K by several portions
at equilibrium pressures of 0.1, 0.4, 0.9, and 1.5 Torr. The IR spectra
were recorded at 77 K. CDCl3 was adsorbed at room temperature
by injecting a portion of 2000 lmol/g into the cell, and the spectra
were measured immediately after injection. A reference spectrum
of the corresponding catalyst run at the measurement temperature
was subtracted. Concentrations of acidic and basic sites were esti-
mated from integral intensities of the characteristic IR bands. The
number of Brønsted acid sites (BAS) was evaluated from the band
at 1530–1550 cmꢁ1 (PyH+ ions) using a molar absorption coeffi-
cient
e
equal to 3 cm/
l
mol (in the literature, e1545 usually varies
mol was measured for
from 1.25 to 3 cm/
lmol [42–47]; 3 cm/l
solid PyHCl salt [47]). The number of Lewis acid sites (LAS) was
estimated from the band at 1450 cmꢁ1 (coordinatively bonded
Py) after deconvolution of the band at 1445–1455 cmꢁ1, using
e
1450 = 3.5 cm/
in the range 1.73–3.5 cm/
of BAS and LAS were also assessed from the intensity of CO bands
at 2162–2170 and 2183–2195 cmꢁ1 using
of 2.6 and 0.9 cm/
mol, respectively [46]. The error in measuring the amounts of
l
mol (typical values reported in the literature are
l
mol [42,43,45,46]). The concentrations
e
l
BAS and LAS was 20–30%. The strength of LAS was assessed on a
scale of CO adsorption heat according to the equation Q [kJ/mol]
= 10.5 + 0.5DmCO, where DmCO = mCO(ads)–2143 [46]. The strength
of BAS in terms of proton affinity (PA) was evaluated from the shift
of the absorbance bands in the OAH stretching region due to their
interaction with CO using the equation PA [kJ/mol] = 1390–log
performed
a comparative study of H2O2 decomposition and
explored the acid–base surface properties of these catalysts by
means of FTIR spectroscopy of adsorbed probe molecules (pyridine
(Py), CO, and CDCl3).
(
D
m
/
OH D
m
OH(SiOH))/0.00226, where
of OAH vibration frequencies in the solid studied and SiO2, respec-
tively (
mOH(SiOH) was taken as 90 cmꢁ1) [46]. The uncertainty in
DmOH and DmOH(SiOH) are shifts
D
estimating the strength of BAS and LAS did not exceed 2–3%. Con-
centrations of basic sites were estimated using a molar absorption
2. Experimental
coefficient (0.16 cm/lmol) reported in the literature for CDCl3
complexes with different bases [46,47]. The error in measuring
the number of basic sites was within 20%. The strength of basic
sites was evaluated from the shift of the C–D stretching band
2.1. Materials and catalysts
Cetyltrimethylammonium bromide (CTAB, 99%+), tetraethyl
orthosilicate (TEOS, 98%+), and ammonium niobate(V) oxalate
hydrate NH4[NbO(C2O4)2(H2O)2]ꢀ3H2O (99.99%) were purchased
from Aldrich. Niobium(V) ethoxide (99.95%) was used as received
from Acros. Acetonitrile (HPLC grade, Panreac) was dried and
stored over activated 4 Å molecular sieves. Cyclohexene and cis-
cyclooctene were purchased from Aldrich and passed through a
column filled with neutral alumina before use. All other reagents
according to the equation log
D
mCD = 0.0066 ꢂ PA–4.36, where
D
mCD = 2268–mCD(ads) [47,48]. The error in estimating the strength
of basic sites was 2%.
2.3. Kinetic experiments
Kinetic experiments were performed in temperature-controlled
glass vessels under vigorous stirring (600 rpm). Reactions were ini-

I.D. Ivanchikova et al. / Journal of Catalysis 356 (2017) 85–99
87
tiated by addition of H2O2 into a MeCN solution containing organic
2.6. Stereospecificity tests
substrate (CyO or MNQ), a catalyst, and an internal standard for GC.
The total volume of the reaction mixture was 1 mL. The reaction
temperature was 50 °C for CyO and 70 °C for MNQ. Typical reaction
Oxidation of cis-/trans-stilbenes and cis-b-methylstyrene was
carried out at 50 °C in MeCN (1 mL) using the following concentra-
tions of the reagents: alkene 0.1 M, H2O2 0.1 M, and catalyst 0.003
mmol Nb. The ratio of cis- and trans-epoxides was determined by
1H NMR and GC–MS techniques.
time was 1–2 h for CyO and 2–6 h for MNQ. Samples (1 lL) of the
reaction mixture were withdrawn periodically during the reaction
with a syringe and analyzed. Substrate conversions and epoxide
product yields were quantified using biphenyl as internal standard.
Each experiment was reproduced at least 2–3 times. To rule out the
possibility of evaporative losses of the substrate, blank experi-
ments without catalyst and oxidant were carried out at the reac-
tion temperature using the internal standard.
2.7. Oxidation of cyclohexene
Catalytic oxidation of cyclohexene was carried out at 30 and 50
°C in MeCN (1 mL) using the following concentrations of the
reagents: CyH 0.1 M, H2O2 0.1 M (taken as a 50% solution in water),
and 0.003 mmol Nb or Ti catalyst. The product yields and substrate
conversions were quantified by GC using biphenyl as an internal
standard. For GC analysis, we used the method described by Shul’-
pin [50], which is based on treatment of the reaction mixture with
triphenylphosphine to reduce unreacted H2O2 and cyclohexenyl
hydroperoxide formed during the reaction.
2.3.1. Reaction order in catalyst
The catalyst amount was varied in the range 7.5–45 mg, which
corresponds to 0.0015–0.009 mmol of Nb. Concentrations of other
reactants were held constant: substrate 0.1 M, H2O2 0.1 (CyO) or
0.4 M (MNQ).
2.3.2. Reaction order in substrate
The initial substrate concentration was varied between 0.01 and
0.2 M while maintaining a constant concentration of H2O2 (0.1 and
0.4 M for CyO and MNQ, respectively) and catalyst amount (0.015
g; 0.003 mmol Nb).
2.8. Hydrogen peroxide decomposition
Decomposition of H2O2 (0.4 M) was studied in the absence of
organic substrate at 80 °C in MeCN (5 mL) in the presence of either
Nb or Ti catalyst (0.015 mmol Nb or Ti). Aliquots of 0.2 mL were
taken during the reaction course, and H2O2 concentration was
determined by iodometric titration. Four parallel experiments
were carried out. The temperature dependence of the decomposi-
tion rate was studied in the range 40–80 °C.
2.3.3. Reaction order in H2O2
The initial oxidant concentration was varied in the range 0.05–
0.4 M (for CyO) or 0.1–0.8 M (MNQ). The concentration of water in
these experiments was kept constant by addition of corresponding
amounts of H2O. The concentrations of other reactants were as fol-
lows: CyO or MNQ 0.1 M, catalyst 0.015 g (0.003 mmol Nb).
2.9. DR UV–vis study of H2O2 interaction with Nb catalyst
2.3.4. Reaction order in H2O
Catalyst B (0.57 g, 0.11 mmol Nb) was placed in a mixture of
MeCN (6 mL) and 30% H2O2 (2 mL), and the slurry was stirred at
room temperature for 15 min. Then the catalyst was separated
by filtration and divided into two portions. The first portion was
treated with H2O (5 mL) to remove physisorbed H2O2, while the
second was subjected to treatment with an aqueous solution of
NaOAc (5 mL, 0.16 mmol) for 15 min at room temperature. The cat-
alyst was separated by filtration under vacuum, dried in, air and
used for DR UV–vis measurements.
The initial concentration of water was varied from 0.2 to 1.7 M
(CyO) and from 0.8 to 3.1 M (MNQ). Other parameters were held
constant: substrate 0.1 M, catalyst 0.015 g (0.003 mmol Nb), H2O2
0.1 M (CyO) or 0.4 M (MNQ).
2.3.5. Determination of activation energies
Temperature dependence of the reaction rate was studied in the
range 30–80 °C in MeCN using the following reaction conditions:
CyO or MNQ 0.1 M, H2O2 0.1 or 0.4 M for CyO or MNQ, respectively,
and 0.015 g catalyst (0.003 mmol Nb).
2.10. Instrumentation
2.4. Initial rate determination and evaluation of rate laws
GC analyses were performed using a Chromos GH-1000 gas
chromatograph equipped with a flame ionization detector and a
quartz capillary column (30 m ꢂ 0.25 mm) filled with BP-5. Sub-
strate conversions and product yields in all cases were determined
by the internal standard technique. GC–MS analyses were carried
out using an Agilent 7000 GC/MS system equipped with a CM–
WAX fused silica capillary column (10 m ꢂ 0.25 mm ꢂ 0.25 mm).
The mass spectrometer worked in EI mode at 70 eV, scan range
40–550 m/z. The identification of components was based on com-
puter matching of their mass spectra with those of the NIST’11
mass spectral library (NIST MS Search 2.0, 2005). Nitrogen adsorp-
tion measurements were performed at 77 K using a NOVA 1200
instrument (Quantachrome) within the partial pressure range
10ꢁ4–1.0. The catalysts were degassed at 150 °C for 24 h before
the measurements. Surface areas were determined by the BET
analysis. Pore size distributions were calculated from the adsorp-
tion branches of the nitrogen isotherms by the regularization pro-
cedure, using reference local isotherms calculated for a cylindrical
silica pore model in the framework of the density functional theory
(DFT) approach. Special software provided by Quantachrome Corp.
was used. Mean pore diameters were calculated as mathematical
expectation values from the pore size distributions. Niobium con-
The initial rate method was employed to determine the reaction
orders. Initial rates were calculated as d[CyO(or MNQ)]/dt at t = 0.
Since epoxides were the sole products at low conversion, the initial
reaction rates of substrate consumption and epoxide accumulation
were close to each other. The rate law was derived by applying a
steady-state approximation to concentrations of all active species.
Procedures for calculating the initial rates and fitting the rates with
the derived law were similar to those described by some of us ear-
lier [49]. For the detailed description of the kinetic modeling pro-
cedure and derivation of the rate law, see the Supporting
Information (SI).
2.5. Effects of additives
Acid (HClO4) or base (NaOAc) was added into the reaction mix-
tures before addition of the oxidant in an amount close to the
amount of Nb in the catalyst loaded (0.003 mmol). Turnover fre-
quencies (TOF) were determined from the initial rates and calcu-
lated as TOF = (moles of substrate consumed)/(moles of Nb ꢂ
time).

88
I.D. Ivanchikova et al. / Journal of Catalysis 356 (2017) 85–99
tent in the solids was determined by ICP–AES using a Thermo Sci-
entific iCAP–6500 instrument. The state of niobium in the catalysts
was characterized by DR UV–vis spectroscopy under ambient con-
ditions using a Shimadzu UV–VIS 2501PC spectrometer in the 190–
900 nm range with a resolution of 2 nm; BaSO4 was used as stan-
dard reference. The infrared (IR) spectra were recorded on a Shi-
madzu FTIR 8300 spectrometer.
[54]), mixed oxides [32], and supported niobium oxides [57], but
their number remained lower relative to the number of LAS. So
far, very few experimental studies have dealt with evaluation of
basic sites in Nb-containing silica materials. On the basis of ESR,
NO/FTIR, and catalytic studies, Ziolek and co-workers suggested
that Lewis acidic and basic sites appear simultaneously in the
Nb-containing molecular sieves upon dehydration [55,56]. The
total acidity and basicity, evaluated by temperature-programmed
desorption (TPD) of Py and CO2, respectively, were found to be
close in Nb-KIT-6 and increased with enlarging Nb loading [58].
In this work, we employed FTIR spectroscopy of weakly (CO) and
strongly (Py) interacting bases to probe semiquantitatively the
number and strength of LAS and BAS in mesoporous niobium sili-
cates A and B in comparison with titanium silicate A-Ti prepared
by the same method. In addition, adsorption of CDCl3 was used
to estimate the number and strength of basic sites. Fig. S2 shows
FTIR spectra of Py adsorbed at 423 K on the three catalysts studied.
The band at 1530–1550 cmꢁ1 corresponding to PyH+ ions [43–46]
is present in the spectra of both Nb catalysts but is missing in the
spectrum of the Ti catalyst. The number of BAS is larger for catalyst
A with oligomeric Nb sites, which agrees with the literature
3. Results and discussion
3.1. Catalyst preparation and characterization
In our previous studies, we demonstrated that EISA is an afford-
able and versatile method for the preparation of hydrothermally
stable and leaching-tolerant mesoporous titanium [41] and nio-
bium silicates [29] with a weakly ordered structure. This method
enables easy control of the state of Nb active centers and, depend-
ing on the choice of the niobium precursor, favors the formation of
either small oligomers (type A) or site-isolated Nb(V) species (type
B) [29]. Using the EISA approach, we synthesized niobium silicates
of these two types (hereinafter, catalysts A and B) and used them
for exploring kinetic regularities of CyO and MNQ epoxidation with
H2O2. For comparison, we also prepared a mesoporous titanium sil-
icate (designated as catalyst A-Ti) by the EISA method. All three
catalysts had similar textural properties and metal content
(Table 1). A complete characterization of these catalysts by XRD,
SEM, N2 adsorption, and spectroscopic techniques (DR UV–vis,
Raman, and XPS) has been reported elsewhere [29,41]. DR UV–
vis spectra of catalysts A, B, and A-Ti are shown in Fig. S1.
[32,52,54,57]. The bands at 1607–1610 cmꢁ1
to complexes of Py with LAS (PyL), while the band at 1600 cmꢁ1
8a) corresponds to H-bonded Py [43,44,46]. The absorption fea-
ture at 1448 cmꢁ1
19b) is observed as a result of overlapping
bands of coordinatively bonded (1450 cmꢁ1
and H-bonded
(m8a) are attributed
(m
(m
)
(1440 cmꢁ1) Py. The concentrations of BAS and LAS estimated from
the intensity of the characteristic bands at 1545 and 1450 cmꢁ1
respectively, are given in Table 2.
,
One of the most reliable and informative approaches to explor-
ing BAS and LAS in heterogeneous catalysts is adsorption of weakly
interacting probe molecules, e.g., CO at 77 K, monitored by IR spec-
troscopy [46,51,59]. The spectra of CO adsorbed onto catalyst A at
different pressures are shown in Fig. 1. A similar picture was
observed for catalyst B (Fig. S3). At low CO pressure, two bands
at 2163 and 2193 cmꢁ1 are present. The first can be attributed to
CO interacting with surface BAS through H-bonding, while the sec-
ond corresponds to CO coordinatively bonded to LAS [46,59]. With
increasing CO pressure, the band at 2193 cmꢁ1 shifts to 2191 cmꢁ1
and its intensity increases. Simultaneously, a new absorption fea-
ture at 2157 cmꢁ1 arises on the left side of the 2163 cmꢁ1 band
and increases in intensity. This band is due to CO molecules H-
bonded to silanol groups [46,59]. Therefore, two types of BAS can
be distinguished in the niobium silicates. The IR spectra of CO
adsorbed onto catalyst A-Ti (Fig. S4) are similar to the correspond-
ing spectra reported earlier for titanium silicates prepared by dif-
ferent methods [59] and reveal only two characteristic bands, at
2157 and 2186 cmꢁ1 (the latter shifts to 2183 cmꢁ1 at high CO
pressure), which indicates that BAS different in strengths from
SiOH groups are absent.
3.2. FTIR study of surface acid–base properties
Infrared spectroscopy permits discrimination between various
types of surface sites based on the effects on the spectral features
of the adsorbed probe molecules induced by acid–base interaction
[46,51]. Several IR studies employing probe molecules (Py, CD3CN,
NO) revealed Lewis acid sites (LAS) in dehydrated Nb-containing
molecular sieves [40,52–56]. Only one type of LAS was identified
by adsorption of CD3CN onto Nb-b [40]. Brønsted acid sites (BAS)
are not present in this zeolite, as shown by Py [53] and CD3CN
[40] adsorption. On the other hand, BAS appeared in niobium-
rich mesostructured materials (Nb-SBA-15 [52] and Nb-KIT-6
Table 1
Metal content and textural characteristics of Nb and Ti silicates.
Catalyst
M, wt.%a
SBET, m2/g
Vp, cm3/gb
Dp, nmc
A
B
A-Ti
1.96
1.80
2.35
1190
1280
1270
0.56
0.57
0.54
2.8
2.8
2.7
The concentrations of BAS and LAS determined from the inten-
sity of the CO bands at 2162–2170 and 2183–2195 cmꢁ1, respec-
tively, are provided in Table 2. The values acquired using Py and
CO probe molecules are in good agreement, which implies fairly
a
b
c
Based on elemental analysis data for calcined samples.
Mesopore volume.
Mean pore diameter.
Table 2
Acidity and basicity of Nb- and Ti-containing silicates.a
Catalyst
Brønsted acid sites
Py
Lewis acid sites
Basic sites
CDCl3
CO
Py
CO
N,
lmol/g
PA, kJ/mol
N,
l
mol/g
N,
lmol/g
Q, kJ/mol
N,
lmol/g
PA, kJ/mol
N, lmol/g
A
B
A-Ti
12
5
<2
1200
1200
–
12
10
<1
33
21
12
36
35
32
32
23
15
780
780
740
760
520
640
a
See Section 2.2 for calculation details and measurement errors.

I.D. Ivanchikova et al. / Journal of Catalysis 356 (2017) 85–99
89
in Ti catalysts, for which typical values of
D
m
OH lie in the range 90–
100 cmꢁ1 and are close to the
DmOH established for silica [59]. Also,
the strength of LAS assessed on the scale of CO adsorption heat [46]
(Table 2) turned out to be higher for the Nb catalysts than for the Ti
one. On the other hand, the strength of LAS for the Nb catalysts A
and B is approximately the same within experimental error (2–3%).
The number and strength of basic sites were assessed using
CDCl3 as a probe molecule [46]. This method was first suggested
by Paukshtis and co-workers [60] and then employed by other
researchers in exploring basic sites in heterogeneous catalysts
[48,51]. The IR spectra of adsorbed CDCl3 are presented in Fig. S4.
The broad band centered at 2262 (catalysts A and B) cmꢁ1 corre-
sponds to C–D stretching vibration of CDCl3 adsorbed onto week
basic sites [46,60]. Previously, weak basic sites were identified in
Nb2O5 oxide by IR spectroscopy using CHCl3 as a probe molecule
(D
mCH 13 versus 38 cmꢁ1 for Al2O3 [45]). For catalyst A-Ti, the
value of mCD is very close to that measured for CDCl3 adsorbed on
silica (2265 cmꢁ1) [48,60]. Estimates of the number of basic sites
and their strength in terms of PA are provided in Table 2. Three
main conclusions can be drawn from these data: (1) the difference
in PA between A and A-Ti (780 versus 740 kJ/mol) is outside the
uncertainty of the measurements and indicates that weak basic
sites are present in the Nb silicates and absent in the Ti silicate
(PA 733 kJ/mol was reported for SiO2 [48]); (2) among the Nb sili-
cates, the number of weak basic sites is larger for catalyst A, which
contains di(oligo)meric Nb sites; and (3) the estimated concentra-
tions of basic sites are of the same order of magnitude as the gen-
eral number of NbAOANb/NbAOASi bonds in the Nb silicates
calculated on the basic of the elemental analysis data under the
assumption that one Nb(V) is surrounded by four bridging oxygen
atoms along with one terminal oxygen (@O or AOH).
Fig. 1. FTIR spectra of CO adsorbed at 77 K on catalyst A (carbonyl stretching
region): (1) CO pressure 0.1 Torr, (2) 0.4 Torr, (3) 0.9 Torr, and (4) 1.5 Torr. In the
insert: (5) spectrum of OAH stretching region measured after CO adsorption at 0.4
Torr. The spectrum of catalyst A before interaction with adsorbate molecules was
used for background subtraction.
good accessibility of the Nb sites in both catalysts. Similarly to
other Nb-containing silica materials [52,54,57,58], the number of
LAS is significantly larger than the number of BAS. One can notice
that the concentration of LAS seems underestimated relative to the
number of Nb sites calculated on the basis of the elemental analy-
Subramaniam and co-workers demonstrated that the total acid-
ity and basicity increase with increasing niobium loading in Nb-
KIT-6 [58]. This agrees with our finding that the number of both
BAS and basic sites is larger in catalyst A than in catalyst B.
sis data (ca. 200
in the molar extinction coefficient
ular adsorbate and adsorbent [51,59]. Since the true value of
l
mol/g). The reason for that could be uncertainty
, which depends on the partic-
for
e
e
the mesoporous Nb and Ti silicates is unknown, the values deter-
mined for the bulk oxides were used in the calculations (see Sec-
tion 2.2), which could lead to underestimation of the number of
LAS. Nevertheless, a rough estimate and comparison of the number
of sites within a series of catalysts can be used to assess their rel-
ative accessibility [59]. Since the concentration of LAS seems to be
close for catalysts A and B within experimental error, which is 20–
30%, we may infer that Nb sites possess similar accessibility in both
Nb catalysts. This can be envisioned if catalyst A contains evenly
dispersed coordinatively unsaturated dimers or small oligomers
of Nb(V); otherwise the number of LAS would be expected to
decrease relative to the number in catalyst B, which has mostly
site-isolated Nb.
3.3. General features of epoxidation kinetics
The epoxides derived from CyO and MNQ possess fairly good
stability toward epoxide ring opening and rearrangement, which
results in excellent epoxidation selectivity in H2O2-based oxidation
over the Nb catalysts (both A and B) [29]. For this reason, these two
compounds were chosen as model substrates for kinetic studies.
Previously, we demonstrated by a hot filtration test and elemental
analysis data [61] that oxidation catalysis over Nb silicates pre-
pared by EISA has a truly heterogeneous nature and there is no nio-
bium leaching [29]. For both substrates, the reaction rate was
negligible in the absence of catalyst or oxidant. Kinetic plots for
CyO and MNQ consumption over catalysts A and B revealed no
induction period over the entire range of concentrations used.
The initial epoxidation rates and epoxide yields were not affected
by light, which ruled out any photochemical process over the nio-
bium silicates. Kinetic curves acquired under Ar and in air coin-
cided completely, indicating that molecular oxygen is not
involved, at least in the rate-limiting step of the oxidation reac-
tions. Addition of the conventional radical scavenger ionol pro-
duced no rate-retarding effect. All these facts collectively imply a
nonradical oxidation mechanism or a mechanism with short radi-
cal chains.
Although quantitative determination of acid or base strength
appears to be extremely difficult, relative sequences of these quan-
tities for comparable groups of solid catalysts can be established
[51]. The strength of BAS can be evaluated from the shift of the
absorbance band in the OAH stretching region caused by the inter-
action with probe molecules, such as CO or MeCN [32,46,51,59].
The insert in Fig. 1 shows a difference spectrum of this region mea-
sured for catalyst A. While the band at 3640 cmꢁ1 is due to SiOH
groups H-bonded with CO, the broad band at 3460–3470 cmꢁ1
can be assigned to BAS associated with Nb [46]. The values of pro-
ton affinity (PA) estimated for BAS on the basis of the band shift
relative to the frequency of the OAH stretching before CO adsorp-
tion are presented in Table 2. The band shift
D
m
OH of ca. 270 cmꢁ1
The reaction rates were not affected by the rate of stirring of the
reaction mixture in the range of 500–1000 rpm, showing that
external diffusion of reactants to catalyst particles does not limit
the overall oxidation rate. In line with this, epoxidation rates were
proportional to the catalyst loading. Surprisingly, kinetic regulari-
ties did not depend on the state of Nb centers (isolated or
and the corresponding value of PA (1200 kJ/mol) indicate that BAS
in the niobium silicates are of moderate strength. This is consistent
with the conclusions made for other Nb–silica catalysts on the
basis of studies by NH3 TPD [54] and FTIR with adsorbed MeCN
[32]. The strength of BAS in Nb catalysts is significantly higher than

90
I.D. Ivanchikova et al. / Journal of Catalysis 356 (2017) 85–99
oligomerized) but were solely determined by the nature of the
organic substrate (vide infra).
tion rate exhibited saturation with increasing concentrations of
alkene (Fig. 2a) and oxidant (Fig. 2b) and increased proportionally
to the catalyst loading (Fig. 2c). The reaction was mildly inhibited
by H2O (Fig. 2d). In contrast to epoxidation over the microporous
zeolite Nb-b, where the rate of epoxidation depended inversely
on the C6H10O concentration [40], addition of CyO epoxide (ca.
40 mol.% of the amount of alkene) did not affect the reaction rate,
3.4. Kinetics of CyO epoxidation
Fig. 2 shows the dependence of initial rates of CyO consumption
on concentrations of alkene, hydrogen peroxide, niobium, and
water. Notwithstanding the type of catalyst (A or B), the epoxida-
A
Catalyst
Catalyst
B
3.00
2.80
2.40
2.00
1.60
1.20
0.80
2.50
2.00
1.50
1.00
0.50
0.00
(a)
(a)
0.00
0.05
0.10
0.15
0.20
0.00
0.05
0.10
[CyO]
0.15
0.20
[CyO], M
3.50
3.00
2.50
2.00
1.50
1.00
0.50
2.40
2.00
1.60
1.20
0.80
0.40
(b)
(b)
0.00
0.10
0.20
0.30
0.40
0.00
0.10
0.20
0.30
0.40
[H2O2], M
[H2O ], M
2
6.00
5.00
4.00
3.00
2.00
1.00
0.00
6.00
5.00
4.00
3.00
2.00
1.00
0.00
(c)
(c)
1.50 3.00 4.50 6.00 7.50 9.00
n(Nb) x 103, mmol
1.50 3.00 4.50 6.00 7.50 9.00
n(Nb) x 103, mmol
2.50
2.25
2.00
1.75
1.50
1.25
1.00
0.75
0.50
2.50
2.25
2.00
1.75
1.50
1.25
1.00
0.75
0.50
(d)
(d)
0.25 0.50 0.75 1.00 1.25 1.50 1.75
[H2O], M
0.25 0.50 0.75 1.00 1.25 1.50 1.75
[H2O], M
Fig. 2. Plots of the initial rate of CyO oxidation (MeCN, 50 °C) versus concentration of alkene (a), oxidant (b), catalyst (c), and water (d): experimental points (j) and fitting
) to an equation given in Section 3.9. For other reaction conditions, see Section 2.3.
(

I.D. Ivanchikova et al. / Journal of Catalysis 356 (2017) 85–99
91
thereby indicating that adsorption of the main oxidation product
onto the mesoporous Nb catalyst is insignificant.
OAO bond breaking in the hydroperoxo intermediate, similarly
to that in homogeneous Ti-containing polyoxometalates [35].
Activation energies of H2O2 decomposition over catalysts A and
B were determined from the Arrhenius plots shown in Fig. 5.
It is noteworthy that the apparent activation energies for CyO
epoxidation are significantly lower than the Ea established for
decomposition of H2O2: 11.9 vs. 20.5 kcal/mol for catalyst A and
10.9 vs 16.6 kcal/mol for catalyst B. Therefore, one might expect
that epoxidation selectivity (or, in case of epoxides sensitive to ring
opening, total selectivity to epoxide, diol, and their overoxidation
products) might be improved by decreasing the reaction tempera-
ture. To verify this suggestion, we have studied the effect of tem-
perature on the selectivity of cyclohexene oxidation.
The rate of CyO oxidation exhibited a typical Arrhenius depen-
dence, which implies that there is no change in the rate-limiting
step in the temperature interval used. The values of the apparent
activation energy (Ea) for catalysts A and B estimated from the
Arrhenius plots (Fig. 3) are close, 11.9 and 10.9 kcal/mol, respec-
tively. Earlier, an activation energy of 10.3 kcal/mol was reported
for epoxidation of styrene over a mesoporous catalyst Nb-KIT-5
[27]. Note that such values of Ea are typical of reactions controlled
by chemical interaction, because for reactions controlled by diffu-
sion processes, Ea is expected to be <4 kcal/mol [62].
As one can judge from the plots shown in Fig. 3, the rate of CyO
epoxidation was always a little higher for catalyst B than catalyst
A, which seems reasonable taking into account greater isolation
of Nb sites in the former.
3.6. Effect of temperature on epoxidation of cyclohexene
The data presented in Table 3 show that the total heterolytic
pathway selectivity (we define it as the sum of epoxide, trans-
cyclohexane-1,2-diol, and 2-hydroxycyclohexane-1-one) increased
from 70–71% to 85–88% at similar alkene conversions when the
reaction temperature was reduced from 50 to 30 °C. This effect
was less pronounced for the titanium catalyst A-Ti.
3.5. Hydrogen peroxide decomposition
Unproductive decomposition of hydrogen peroxide is the main
reaction that competes for H2O2 with the target selective oxidation
reaction [26,29,63,64]. Taking this into account, we evaluated the
activity of catalysts A and B in H2O2 decomposition in the absence
of any organic substrate. Fig. 4 shows kinetic plots of H2O2 decay
over two niobium silicates and titanium silicate of type A (the
amount of Nb or Ti was similar in all the experiments). No H2O2
conversion occurred in the absence of any catalyst.
Among the Nb silicates, catalyst B was a bit more active than
catalyst A (TOF 1.05 versus 0.85 minꢁ1). More important is that
both Nb catalysts are significantly less active in this reaction than
the Ti catalyst prepared by the same EISA technique (TOF 3.3
minꢁ1). This agrees with the smaller amounts of homolytic oxida-
tion products formed from substrates with reactive allylic H atoms
(e.g., cyclohexene and limonene) over niobium silicates than over
their Ti analogues [21,30] (see also Section 3.6). It is generally
believed that acid sites of moderate strength are involved in the
mechanism of epoxidation of alkenes with H2O2, while strong acid-
ity activates the decomposition of H2O2 [32]. The results of the FTIR
spectroscopic study presented in Table 2 allow one to conclude
that the difference in the H2O2 decomposition rates observed for
the Nb and Ti catalysts is certainly not related to the difference
in their Brønsted acidity, which decreases in the order A > B > A-
Ti. We may assume that higher activity of the Ti silicate in H2O2
decomposition can be related to lower energy cost of homolytic
0.4
0.3
0.2
0.1
0.0
0
20
40
60
80
100
Time, min
Fig. 4. H2O2 decomposition in the absence of organic substrate over catalysts A ( ),
B (j), and A-Ti ( ). Reaction conditions: H2O2 0.4 M, catalyst 0.015 mmol Nb or Ti,
80 °C, MeCN 5 mL. Points acquired from four parallel experiments are shown for
each curve.
-5
-5
Ea = 16.6 0.8 kcal/mol
-6
Ea = 10.9 0.4 kcal/mol
-6
-7
-8
-9
-7
-8
Ea = 11.9 0.6 kcal/mol
Ea = 20.5 0.3 kcal/mol
-10
-9
0.0028
0.0029
0.0030
1/T, K-1
0.0031
0.0032
0.0028
0.0030
0.0032
0.0034
1/T, K-1
Fig. 3. Arrhenius plots for CyO oxidation over catalysts A ( ) and B (j). Reaction
conditions: CyO 0.1 M, H2O2 0.1 M, catalyst 0.003 mmol Nb, MeCN 1 mL, 30–80 °C.
Fig. 5. Arrhenius plots for H2O2 decomposition over catalysts A ( ) and B (j).
Reaction conditions: H2O2 0.4 M, Nb 0.02 mmol, MeCN 7 mL, 40–80 °C.

92
I.D. Ivanchikova et al. / Journal of Catalysis 356 (2017) 85–99
Table 3
Effect of temperature on selectivity of CyH oxidation over Nb and Ti catalysts.
Catalyst
T, °C
Time, h
CyH conv., %
Heterolytic pathway selectivity, %a
Homolytic pathway selectivity, %b
A
B
A-Ti
A
B
30
30
30
50
50
50
1
1
2
0.5
0.3
2.5
23
30
24
30
34
35
85
88
70
70
71
65
11
9
21
18
21
27
A-Ti
Note: Reaction conditions: CyH 0.1 mmol, 50% H2O2 0.1 mmol, catalyst 0.003 mmol Nb or Ti, MeCN 1 mL.
a
Total selectivity of heterolytic oxidation products (cyclohexene oxide + trans-cyclohexane-1,2-diol + 2-hydroxycyclohexanone).
Total selectivity of homolytic oxidation products (cyclohexenyl hydroperoxide + 2-cyclohexene-1-ol + 2-cyclohexene-1-one).
b
Interestingly, an opposite trend was predicted for zeolite Nb-b,
for which the activation enthalpy appeared to be significantly
higher for cyclohexene epoxidation than for H2O2 decomposition
(17.2 vs 10.8 kcal/mol) and therefore improvement of selectivity
was expected with rising reaction temperature [40]. Unfortunately,
no data on product selectivity at different temperatures were pro-
vided to verify this prediction.
scopic studies, it has been inferred to be the active epoxidizing spe-
cies. Although the ability of mesoporous niobium silicates to
decompose H2O2 is lower than that of their titanium counterparts
(Fig. 4), it is not negligible, and one might expect a contribution of
radical species to the oxidation process. On the other hand, one of
the main arguments against a radical mechanism of epoxidation
over Nb is the high heterolytic pathway selectivity in the oxidation
of cyclohexene. This olefin is a well-known test substrate for dis-
crimination between homolytic and heterolytic oxidation pro-
cesses, since it possesses highly reactive CAH bonds in the allylic
position, which react easily with radical species, resulting in allylic
oxidation products [68]. For niobium silicates, total heterolytic
pathway selectivity in cyclohexene oxidation typically varies in
the range 70–99%, depending on the degree of Nb site isolation
and the reaction conditions [13,15,23,24,28–30]. In general, it is
superior to that of mesoporous titanium silicates, for which allylic
oxidation products are present in pronounced amounts when H2O2
is employed as oxidant [11,12]. For example, SiO2-grafted Nb(V)
catalysts (Nb–SiO2) demonstrated 88–90% selectivity toward
cyclohexene oxide and trans-diol at ca. 100% total yield, while
Ti–SiO2 prepared by the same methods gave only 65–67% of these
products at 69–93% total yield [30]. The data presented in Table 3
also demonstrate that Nb silicates prepared by EISA (both A and B)
reveal higher heterolytic pathway selectivity than titanium silicate
A-Ti (85–88% vs 70% at 30 °C). A similar trend was established for
oxidation of limonene, another substrate with reactive allylic
hydrogen atoms [21].
Finally, stereospecificity in epoxidation of cis-alkenes is another
strong argument in favor of a nonradical epoxidizing species and
heterolytic concerted epoxidation mechanism that involves no
radical or ionic intermediate capable of rotation over CAC bonds
(such a species would lead to at least partial loss of the initial con-
figuration and formation of trans-epoxide). Indeed, epoxidation of
methyl oleate (methyl cis-9-octadecenoate) in the presence of
mesoporous Nb catalysts produced solely methyl cis-9,10-
epoxyoctadecanoate [17,22]. In contrast, some amounts of the cor-
responding trans-isomer formed in the presence of mesoporous
titanium-silicates under similar reaction conditions [69]. The oxi-
dation of cis-b-methylstyrene over Nb silicates (both catalysts A
and B) also proceeded with retention of the cis-configuration, pro-
ducing cis-epoxide as the main oxidation product (Scheme 1). The
amount of the corresponding trans-epoxide was less than 1%.
3.7. Presumably active species: Heterolytic vs. homolytic mechanism of
alkene epoxidation
Various oxidizing species were proposed as active forms
responsible for liquid-phase selective oxidation over heteroge-
neous niobium catalysts. Among them were both radical forms,
including HOÅ [39], superoxo NbVO2ꢁÅ [38] and NbIV–(O2)ꢁ [40],
and a species designated as Nb–Oꢁ [55,56], and nonradical forms,
such as side-on peroxo Nb(g
2–O2) [65,66] and end-on hydroper-
oxo [14,31,32] species. Various structures proposed for the nonrad-
ical peroxo niobium species are shown in Fig. 6.
Generation of HOÅ and HO2Å in Nb-catalyzed oxidation is evident,
since unproductive decomposition of hydrogen peroxide always
occurs along with organic substrate oxidation (H2O2 utilization
efficiency in alkene epoxidation is typically in the range of 50–
65%) [26,29]. The formation of NbOÅ and NbO2Å species upon
decomposition of H2O2 over Nb silicates may occur via mecha-
nisms similar to those proposed for Ti silicates [10,67]. Generation
of the superoxo niobium species NbVO2ꢁÅ on the surface of Nb cat-
alysts was previously confirmed by ESR [38,39]. A shoulder at
320 nm in DR UV–vis spectra of H2O2-treated amorphous Nb2O5
has been assigned to such species [39]. More recently, a similar
absorption band (310 nm) was found for H2O2-treated Nb-Beta
and assigned to an unusual superoxide designated as NbIV–(O2)
or NbIV–(O2)ꢁ [40]. On the basis of results of in situ UV–vis spectro-
III
I
IV
II
H2O2
H3O+
V
Nb,Si-catalyst
Conv. 50%
10%
Select. 60%
20%
6%
Scheme 1. Oxidation of cis-b-methylstyrene over mesoporous niobium silicate.
Reaction conditions: alkene 0.1 M, H2O2 0.1 M, catalyst 0.003 mmol Nb, MeCN 1 mL,
50 °C, 6 h.
Fig. 6. Structures of peroxo and hydroperoxo niobium species proposed in the
literature: I [65], II [66], III [32], IV [14,32], and V [31].

I.D. Ivanchikova et al. / Journal of Catalysis 356 (2017) 85–99
93
We also attempted to check stereospecificity in the oxidation of
cis-stilbene, an olefin substrate highly prone to the formation of
trans-epoxide in case of radical epoxidizing species [70]. Cis-
epoxide was found in the reaction mixture in a low yield, while
trans-epoxide was not present at all. However, the main oxidation
product was benzaldehyde; therefore, this test was less unambigu-
ous than that with cis-b-methylstyrene, where cis-epoxide was the
principal reaction product. Trans-stilbene was ca. fivefold less reac-
tive than cis-stilbene, indicating that steric factors have a strong
impact on the oxidation process over the Nb catalysts.
NbOH species with H2O2 via Eqs. (1)–(5). Peroxo and hydroperoxo
species may be present at an equilibrium (counteranion is not
shown) that would shift right or left, depending on the surface
and medium acidity:
+
+ H+
a
ð6Þ
The distinctive features of epoxidation over mesoporous nio-
bium -silicates, such as minor or negligible amount of allylic oxida-
tion products and the retention of configuration of cis-olefins, are
only consistent with a heterolytic mechanism. Therefore, we can
rule out any radical species as the species responsible for the selec-
tive formation of epoxides and suggest a concerted transfer of oxy-
gen atoms from a peroxo niobium species to a nucleophilic C@C
bond.
In 1991, Clerici first suggested that hydroperoxo TiAOOH rather
than peroxo Ti(
2–O2) is responsible for the catalytic activity of TS-
g
1 [75]. One of the main indications in favor of TiAOOH was a rate-
inhibiting effect of basic compounds on the catalytic activity and,
in opposition, a rate-accelerating effect of acids [75,76]. This sug-
gestion was further supported by chemical, spectroscopic, and
computational studies, and TiAOOH species became a widely
accepted concept over the years, extended later to other Ti-
containing molecular sieves (for recent reviews, see Refs.
[10,11]). This prompted us to investigate effects of acids and bases
on the catalytic performance of mesoporous niobium silicates. The
results are provided in Table 4, where we also include data
acquired for the mesoporous catalyst A-Ti. Addition of sodium
acetate in an amount close to the amount of Nb sites retarded
the rate of CyO oxidation substantially (TOF 0.18 vs. 0.39 minꢁ1),
which strongly supports the hypothesis about hydroperoxo nio-
bium(V) as the real epoxidizing species in the system. Indeed,
according to Eq. (6), bases are expected to transform NbOOH to
3.8. Active niobium species: Hydroperoxo vs. peroxo
The next question that arises is whether the active epoxidizing
species is peroxo Nb(g
2–O2) or hydroperoxo NbOOH (the latter
may also include hydroxyl and water molecules, as shown in
Fig. 6). By analogy with titanium, the formation of peroxo com-
plexes is associated with the reaction of the Nb@O group with
H2O2, while the formation of hydroperoxo species can be envisaged
in two steps [32]:
Nb(g
2–O2). A similar effect was found for the mesoporous Ti cata-
Nb@O þ H2O2 ¢ NbðOÞ2 þ H2O
NbAOH þ H2O2 ¢ NbOOH þ H2O
Nb@O þ H2O2 ¢ NbðOHÞOOH
ð1Þ
ð2Þ
ð3Þ
lyst (TOF 0.22 vs. 0.35 minꢁ1). In contrast to alkene epoxidation
over TS-1 [76], the influence of acid on the CyO oxidation rate
turned out to be minor for mesoporous Nb and Ti silicates (see
Table 4). This may be rationalized if we remember that TS-1 is
hydrophobic and contains mostly tetrahedrally coordinated Ti
(OSi)4 sites, which are partially hydrolyzed in the presence of water
to produce highly reactive TiAOH species, whereas mesoporous Nb
and Ti silicates already have a significant number of M–OH (M = Nb
or Ti) groups on their surface.
Strukul and co-workers suggested that one Nb(V) center can
bear two OH groups, one of which is replaced with a hydroperoxo
group upon interaction with H2O2, leading to Nb(OH)OOH (struc-
ture IV in Fig. 6) [14]:
The peroxo species formed upon interaction of the Nb catalysts
with H2O2 was probed by a DR UV–vis spectroscopic technique.
Fig. 7 shows changes in the DR UV–vis spectrum of catalyst B after
treatment with aqueous H2O2. The appearance of the new absorp-
tion band with a maximum at 307 nm (curve B) responsible for the
pale yellow color of the H2O2-treated sample is associated with
reaction of the Nb(V) center with hydrogen peroxide.
NbðOHÞ2 þ H2O2 ¢ NbðOHÞOOH þ H2O
ð4Þ
More recently, Notestein and co-workers also assumed that
individual Nb sites may hydrolyze under reaction conditions to
form two NbOH bonds with both NbOH on the same Nb atom
[31]. Then they suggested that one of the hydroxyls takes the role
of assisting (similarly to the role of alcohol in Ti-catalyzed oxida-
tion [10]) in the rate-limiting oxygen transfer step, while the other
forms the bound hydroperoxo species (species V in Fig. 6). In this
case, the water molecule remains coordinated on the Nb(OH)
OOH species:
Table 4
Effects of acid and base additives on CyO and MNQ oxidation with H2O2 over Nb and
Ti silicates.
NbðOHÞ2 þ H2O2 ¢ NbðOHÞðH2OÞOOH
ð5Þ
Catalyst
Substrate
CyO
Additive
TOF, minꢁ1a
As implied by vibrational spectroscopic [25,71,72] and compu-
tational [53,73,74] techniques, both Nb@O and NbAOH groups can
be present on the surfaces of niobium silicates. The ratio between
them, as well as the degree of isolation/polymerization, may
depend on the preparation methodology as well as on the hydra-
tion or dehydration state of the surface. Tielens et al. revealed that
the most favorable Nb(V) structure in zeolite Nb-b is one having a
Nb(V)OH group linked to the zeolitic walls by four NbAOSi bonds
[53]. The Nb(V)@O site is less stable but becomes similar in stabil-
ity when hydrated. Later on, it was shown using periodic density
functional theory for a systematic investigation of a series of model
zeolites that the M@O group is more stabilized for M = vanadium,
while M–OH is more favorable for M = niobium or tantalum [73].
Therefore, generation of both peroxo and hydroperoxo species
might be expected through interaction of the surface Nb@O and
A
–
0.39
NaOAc
HClO4
–
NaOAc
HClO4
–
NaOAc
HClO4
–
0.18b
0.38
A-Ti
A
CyO
0.35
0.22
0.37
0.04
0.15
0.04
No reaction
MNQ
MNQ
A-Ti
Notes: Reaction conditions: substrate 0.1 mmol, catalyst 0.003 mmol Nb or Ti,
additive 0.003 mmol (if any), H2O2 0.1 mmol and 50 °C (CyO) or H2O2 0.4 mmol and
70 °C (MNQ), MeCN 1 mL.
a
TOF = (moles of substrate consumed)/(moles of metal ꢂ time), determined by
GC from initial rates of substrate consumption.
b
Similar result with 2 equiv. of NaOAc.

94
I.D. Ivanchikova et al. / Journal of Catalysis 356 (2017) 85–99
(for recent reviews, see Refs. [10,11]). Interestingly, DFT calcula-
tions recently performed using model compounds, early-transi
6
4
2
0
tion-metal-substituted Keggin polyoxometalates [PTM(O2)(H)
A
B
C
nꢁ
W11O39
]
(TM @ Ti(IV), V(V), Zr(IV), Nb(V), Mo(VI), W(VI), and
Re(VII)), showed that, similarly to Ti(IV), the oxygen transfer
energy barrier is ca. 9 kcalꢀmolꢁ1 lower for NbOOH than for Nb
(O2) [35].
3.9. Rate law and mechanism of CyO epoxidation over niobium
silicates
The kinetic regularities established for CyO oxidation over the
niobium silicates (saturation kinetics and first-order rate depen-
dence on catalyst were shown in Fig. 2) are similar to those
observed earlier for alkene epoxidation with H2O2 over TS-1 [10]
and grafted Ta-SBA-15 [91], heterogeneous catalysts known for
their high heterolytic pathway selectivity. Hence, we may suppose
that an Eley–Rideal-type mechanism can be employed for descrip-
tion of Nb-catalyzed CyO epoxidation with H2O2. Such mechanism
may involve the following steps: adsorption of water and H2O2 on
undercoordinated Nb centers (the presence of LAS has been con-
firmed by FTIR spectroscopy with adsorption of CO and Py), H2O2
activation through reversible interaction with NbOH to give
NbOOH followed by electrophilic oxygen transfer from NbOOH to
C@C bond, producing epoxide and regenerating the initial state
of the catalyst. The equilibrium of adsorption/desorption of water
on/from NbOOH site should also be taken into account. We can
neglect the step of adsorption of epoxide product on Nb centers
because no effect of the epoxide on the reaction rate was estab-
lished experimentally:
200
250
300
350
400
450
, nm
Fig. 7. DR UV–vis spectra of catalyst B before (A) and after treatment with H2O2/
H2O (B) and H2O2/H2O/NaOAc (C). Treatment conditions: catalyst 0.57 g (0.11 mmol
Nb), MeCN 6 mL, 30% H2O2 2 mL, H2O 5 mL, NaOAc (if any) 0.16 mmol, 15 min, room
temperature.
Recently, DR UV–vis absorption features at 320 [39] and 335
[40] nm were observed for amorphous niobia and zeolite Nb-
Beta, respectively, after interaction with H2O2 and, as was already
mentioned above, these bands were assigned to superoxo niobium
species, referring to TS-1. However, it is widely accepted that the
DR UV–vis absorption at 350–400 nm (specifically, 385 nm for
TS-1), which appears after interaction of titanium silicates with
H2O2 in addition to the parent Ti4+O2ꢁ ? Ti3+Oꢁ ligand-to-metal
charge transfer (LMCT) band at 208–250 nm, is ascribed to LMCT
from a OAO moiety to the Ti center and therefore belongs to
(hydro)peroxo rather than superoxo titanium species [10,77–81].
A yellow or light yellow color is characteristic of several Nb(V) per-
K1
NbAOH þ H2O ¢ NbAOHðH2OÞ
ð7Þ
ð8Þ
ð9Þ
K2
NbAOH þ H2O2 ¢ NbAOHðH2O2Þ
oxo complexes, for which bidentate (g
2) coordination of the per-
k3
oxo ligand has been proved by single crystal X-ray analysis [82–
84]. Generally, niobium peroxo complexes with a peroxo-to-Nb
molar ratio of 1:1 reveal UV–vis absorption features at 255–280
nm [85–88], while complexes with Nb bearing more than one per-
oxo ligand may absorb also at 325–380 nm [84–86,89]. Impor-
tantly, after treatment of the Nb peroxo derivative with aqueous
NaOAc, the DR UV–vis absorption feature shifted from 307 to
293 nm (curve C in Fig. 7). A blue shift was previously observed
in the DR UV–vis spectrum of H2O2-treated TS-1 after subsequent
dosage with an NH3 or NaOH aqueous solution, and it was attribu-
ted to the transformation of an open, end-on hydroperoxo
(„SiO)3(H2O)2TiOOH species into a more stable side-on peroxo
complex [90]. Concerning niobium peroxo species, it was men-
tioned in the literature that their color changed from yellow at
pH ꢃ1 to pale yellow at pH ꢃ3 [87]. On the other hand, a pro-
nounced bathochromic shift was observed in the UV–vis spectra
of the H2O2–Nb2O5–H2SO4 system as the sulfuric acid concentra-
tion was increased [85]. All these facts collectively have led us to
the conclusion that the absorption feature at 307 nm arising after
interaction of catalyst B with aqueous H2O2 can be assigned to a
protonated peroxo Nb(V) species (i.e., the hydroperoxo species
NbOOH), which undergoes transformation to the peroxo species
NbAOHðH2O2Þ ¢ NbAOOHðH2OÞ
k
ꢁ3
K4
NbAOOH þ H2O ¢ NbAOOHðH2OÞ
ð10Þ
ð11Þ
k5
NbAOOH þ CyO ¢ NbAOH þ epoxide
By applying a steady-state approximation to the concentration
of the active species NbOOH, the rate law represented by Eq. S1
(see the SI for details) was derived. After fitting of the experimental
data with this rate law, parameters k and K4 were found to be
ꢁ3
close to zero (Table S1), indicating that the formation of NbOOH
from NbOH and H2O2 is almost irreversible and is accompanied
by instant release of water molecules. Under this assumption, a
simplified rate law can be derived (see SI for details):
K2k3½H2O2ꢄnðNbÞ0
V 1 þ K1½H2Oꢄ þ
ꢀ
ꢁ
W0 ¼
ð12Þ
K2k3½H2O2
k5½CyOꢄ
ꢄ
where n(Nb)0 is the total number of accessible Nb sites and V is the
volume of the reaction mixture.
Nb(g
2–O2) after deprotonation under the action of base. The nega-
Eq. (12) describes the observed kinetic regularities well, includ-
ing the dependence of the initial reaction rate on the concentration
of water (fitting of the experimental points is shown in Fig. 2).
However, we should also mention here that a rate law similar to
that depicted by Eq. (12) might be obtained under the assumption
that Nb(OH)2 reacted with H2O2 via Eq. (4). Therefore, our kinetic
results are, in principle, consistent with both NbOOH and Nb(OH)
OOH active species.
tive effect of base on the catalytic activity of niobium silicates in
epoxidation of electron-rich alkenes (Table 4) allowed us to sug-
gest that NbOOH rather than Nb(g
2–O2) is the true epoxidizing
species that operates with H2O2. Whether the hydroperoxo moiety
has g1 or g2 coordination is still an open question. It should be
noted, however, that the debate about the coordination mode of
the hydroperoxo group in titanium silicate catalysts is still ongoing

I.D. Ivanchikova et al. / Journal of Catalysis 356 (2017) 85–99
95
selectivity in H2O2-based oxidation of electron-rich alkenes. This
agrees with the results recently reported by Notestein and co-
workers, who revealed that direct cyclohexene epoxidation rates
correlate well with an ionic bond character of the supported oxide
[30]. However, higher heterolytic oxidation selectivity could also
be due to a higher energy barrier to homolytic OAO bond breaking
in the hydroperoxo intermediate, similarly to homogeneous Ti-
containing polyoxometalates [35].
H2O2
3.10. Kinetics of MNQ epoxidation
The reaction rates of MNQ oxidation over catalysts A and B
turned out to be at least two orders of magnitude lower than the
rates of CyO epoxidation. Fig. 8 shows dependences of the initial
rates of MNQ consumption over catalyst A on concentrations of
all reagents. Similar dependences were found for catalyst B. The
reaction was found to be first-order in catalyst (Fig. 8a) and H2O2
(Fig. 8c), zero-order in MNQ (Fig. 8b), and nearly zero-order in
H2O (Fig. 8d). Therefore, the kinetic regularities found for MNQ dif-
fer significantly from those established for CyO (Fig. 2).
H2O
Scheme 2. Proposed catalytic cycle for alkene epoxidation with H2O2 over Nb
Similar epoxidation rates were found for MNQ and another qui-
none, 2,3,5-trimethyl-1,4-benzoquinone. Coupled with reaction
order zero in MNQ, this fact indicates that the quinone substrate
does not participate in the rate-limiting step of the epoxidation
process. Apparent activation energies determined from Arrhenius
dependences (Fig. 9) are 13.9 and 12.3 kcal/mol for catalysts A
and B, respectively, which is somewhat high compared with Ea of
CyO epoxidation (see Fig. 3). However, their values still remain
lower than the activation energies of H2O2 decomposition, which
were 20.5 and 16.6 kcal/mol for catalysts A and B, respectively
(see Fig. 5).
silicates.
The catalytic cycle proposed for epoxidation of electron-rich
alkenes is shown in Scheme 2. It is fully consistent with all the
results acquired in this work, including the kinetic and spectro-
scopic data, the rate retarding effect of bases, high heterolytic path-
way selectivity in the epoxidation of cyclohexene, and
stereospecificity of the epoxidation of cis-b-methylstyrene.
A comparative FTIR spectroscopic study using probe molecules
(Py and CO) performed on niobium and titanium silicates prepared
by the same EISA technique showed that the strength of LAS is
somewhat higher for Nb catalysts (see Table 2 and discussion in
Section 3.2), which may account for their superior heterolytic
As one can see from the plots in Fig. 9, the rate of MNQ epoxi-
dation was a little higher for catalyst A than for catalyst B in the
range of temperatures used, while an opposite trend was observed
5,0
4,0
3,0
2,0
1,0
5
4
3
2
1
(a)
9
(b)
0
0,0
1
2
3
4
5
6
7
8
0,00 0,04 0,08 0,12 0,16 0,20
n(Nb) x 103, mmol
[MNQ], M
4,0
3,0
2,0
1,0
0,0
7,0
6,0
5,0
4,0
3,0
2,0
1,0
(c)
(d)
0,0
1,0
1,5
2,0
2,5
3,0
3,5
0,0
0,2
0,4
[H2O2], M
0,6
0,8
[H2O], M
Fig. 8. Plots of the initial rate of MNQ oxidation (MeCN, 70 °C) over catalyst A versus concentration of catalyst (a), substrate (b), oxidant (c), and water (d). For other reaction
conditions, see Section 2.3.

96
I.D. Ivanchikova et al. / Journal of Catalysis 356 (2017) 85–99
times upon addition of NaOAc. This fact has prompted us to the
thought that probably we deal with the so-called Payne oxidation,
which normally requires basic conditions [93]. In this case, the
epoxidation rate is essentially independent of olefin structure
because the slow step in the process is oxidation of the solvent,
MeCN, to give a peroxycarboximidic acid intermediate, H3CC
(@NH)OOH, which is the real epoxidizing species that reacts with
electron-deficient C@C bonds. Indeed, conversion of MNQ became
negligible (<5%) when acetonitrile was replaced with another sol-
vent, ethyl acetate. Finally, acetamide (the product derived from
peroxyimidic acid) was found in the reaction mixture in an amount
comparable to that of epoxide formed from MNQ. No acetamide
was present when the oxidizing substrate was electron-rich alkene
(CyO or limonene) or titanium silicate A-Ti was used as the cata-
lyst. Earlier, on the basis of specific tests and spectroscopic analy-
ses, Guidotti et al. excluded the formation of peroxyimidic acid
during alkene epoxidation with H2O2 over mesoporous titanium
silicates [94]. On the other hand, the same team suggested that
Payne-type oxidation may contribute to the epoxidation of methyl
oleate due to the presence of residual amounts of basic transester-
ification catalysts in the commercial samples of this substrate [69].
Therefore, we may conclude that the ability to epoxidize
-7
-8
Ea = 13.9 1.4 kcal/mol
-9
Ea = 12.3 0.3 kcal/mol
-10
0.0028
0.0029
0.0030
1/T, K-1
0.0031
0.0032
Fig. 9. Arrhenius plots for MNQ oxidation over catalysts A ( ) and B (j). Reaction
conditions: MNQ 0.1 M, H2O2 0.4 M, catalyst 0.003 mmol Nb, MeCN 1 mL, 40–80 °C.
in the epoxidation of CyO (Fig. 3). Possible reasons for this will be
addressed in the following sections.
electron-deficient double bonds in a, b-unsaturated carbonyl com-
pounds, characteristic of mesoporous Nb silicates, is in fact related
to their ability to oxidize MeCN and accomplish Payne-type
oxidation.
3.11. Active species responsible for MNQ epoxidation
As already mentioned, the rate of MNQ epoxidation over Nb sil-
icates was not affected by light, molecular oxygen, or addition of
radical chain scavengers. We should also recall here that meso-
porous titanium silicates are practically inactive in epoxidation of
3.12. Proposed mechanism for MNQ epoxidation
Two alternative mechanisms, both involving peroxycarbox-
imidic acid and consistent with the observed kinetic regularities,
can be envisaged for MNQ epoxidation over niobium silicate cata-
lysts. The first one may involve steps leading to the formation of
hydroperoxo species NbOOH (Eqs. (7)–(10)), followed by rate-
limiting oxidation of MeCN with this species to produce peroxy-
imidic acid and oxygen transfer from the latter to C@C bonds, giv-
ing epoxy derivatives (MNQO) and acetamide:
a,b-unsaturated carbonyl compounds. Accordingly, epoxyquinones
were minor byproducts in the oxidation of alkylphenols to benzo-
quinones over Ti catalysts [92], but their yield greatly increased
under the same conditions if Nb -catalysts were employed [29].
Indeed, no reaction occurred when MNQ was subjected to oxida-
tion using H2O2 and catalyst A-Ti (Table 3). Therefore, we may con-
clude that there is no correlation between the catalyst’s ability to
decompose hydrogen peroxide (Fig. 4) and epoxidize MNQ. All
these facts collectively imply that a nonradical oxidation mecha-
O
NbAOOH þ H3CCN H!2 NbAOH þ H3CCð@NHÞOOH
ð13Þ
ð14Þ
nism operates in the Nb-catalyzed oxidation of
carbonyl compounds.
a, b-unsaturated
H3CCð@NHÞOOH þ MNQ ! H3CCð@OÞNH2 þ MNQO
It is remarkable that the effect of base on the epoxidation of
MNQ is opposite to the effect found for CyO epoxidation (see
Table 4). The rate of MNQ oxidation increased approximately four
An alternative mechanism may involve hydrogen peroxide acti-
vation through adsorption and dissociation on a basic site, which
is, most likely, NbOSi or NbONb (the second option is shown
H2O2
δ-
δ-
δ+
MeCN
Scheme 3. Tentative mechanism for epoxidation of a,b-unsaturated carbonyl compounds with H2O2 over Nb catalysts.

I.D. Ivanchikova et al. / Journal of Catalysis 356 (2017) 85–99
97
below), slow oxidation of MeCN with HOOꢁ to afford an anionic
indicate that epoxidation of
a,b-unsaturated carbonyl compounds
intermediate that takes protons from the catalyst, thereby regener-
ating its initial form and producing peroxyimidic acid, followed by
oxygen transfer from the peroxy acid to C@C bond (Eq. (14)):
proceeds through the so-called Payne oxidation, which involves
rate-limiting oxidation of the solvent molecule (MeCN) by a peroxo
niobium species or, more likely, HOOꢁ formed upon interaction of
H2O2 with a basic site (NbAOANb or NbAOASi), leading to the for-
mation of peroxycarboximidic acid, H3CC(@NH)OOH, which reacts
with electron-deficient C@C bonds, producing epoxy derivatives
and acetamide. A comparative FTIR spectroscopic study using var-
ious probe molecules (Py, CO, and CDCl3) performed on the nio-
bium and titanium silicates prepared by the same EISA technique
revealed two main differences: (1) the amount and strength of
LAS and BAS is somewhat higher in Nb silicates than in their Ti
counterparts and (2) Nb silicates possess weak basic sites while
Ti silicates do not; the number of basic sites is larger for catalysts
with di(oligo)meric Nb sites. The higher strength of LAS may
account for the superior heterolytic pathway selectivity in H2O2-
based oxidation of alkenes over mesoporous niobium silicates,
while the higher strength of BAS may be the reason for various
rearrangements observed in the oxidation of acid-sensitive sub-
strates. The presence of weak basic sites may explain why meso-
H2O2 þ NbONb ¢ NbðH2O2ÞONb
NbðH2O2ÞONb ¢ NbðOHÞNb þ HOOꢁ
HOOꢁ þ H3CCN ! H3CCðOOHÞ@Nꢁ
ð15Þ
ð16Þ
ð17Þ
H3CCðOOHÞ@Nꢁ þ NbðOHÞNb ! H3CCð@NHÞOOH
þ NbONb
ð18Þ
Previously, participation of peroxycarboximidic acid in alkene
epoxidation with H2O2 in the presence of nitriles was implied for
the well-known hydrotalcite basic catalysts [95–97], and a similar
mechanism engaging hydrogen peroxide activation on basic sites
was proposed [97]. Although basic sites in niobium silicates are
weaker than in layered double hydroxides (PA 780 vs 850–920 kJ/-
mol [98]), we may tentatively assume their participation in nucle-
ophilic activation of hydrogen peroxide. In turn, the inertness of
porous niobium silicate catalysts are able to epoxidize
a,b-
unsaturated carbonyl compounds while their Ti analogs are not.
The results obtained in this work allowed us to suggest that the
presence of both Lewis acidic and basic sites ensures the ability
of mesoporous niobium silicate to activate hydrogen peroxide by
two different modes, leading to electrophilic (NbOOH) and nucle-
ophilic (HOOꢁ) active species, which, in turn, accomplish epoxida-
tion of electron-rich and electron-deficient C@C bonds,
respectively. Further spectroscopic, computational, and kinetic
studies on model Nb-containing systems are in progress in our
group to gain further insight into the mechanisms of H2O2 activa-
tion and Nb-catalyzed oxidations.
titanium silicates in epoxidation of a,b-unsaturated carbonyl com-
pounds can be explained by the lack of basic sites in these cata-
lysts. Recent DFT calculations revealed that the most basic site in
Nb-b is the bridging oxygen of the NbAOASi site [53]. Our esti-
mates by FTIR using adsorbed CDCl3 showed a similar strength of
basic sites in catalysts A and B (see Table 2). On the other hand,
somewhat higher activity of catalyst A in MNQ epoxidation
(Fig. 9) correlates with a higher concentration of basic sites in this
catalyst, which in turn, correlates with the number of NbAOANb
connectivities.
Scheme 3 presents a tentative mechanism for the selective
epoxidation of
a,b-unsaturated carbonyl compounds (e.g., MNQ)
Acknowledgments
catalyzed by niobium silicates. This mechanism is compatible with
all the data obtained in this work (kinetics, products, additives, and
solvent effects) and is also supported by the detection of basic sites
in the niobium catalysts.
The authors thank M.G. Clerici for fruitful discussions. The help
of Dr. S.A. Yashnik in DRS UV–vis measurements is greatly appre-
ciated. This work was carried out in the framework of budget pro-
ject 0303-2016-0005 for the Boreskov Institute of Catalysis and
partially supported by the Russian Foundation for Basic Research
(Grant 16-03-00827).
4. Conclusions
The results acquired in this work using kinetic, product, and
spectroscopic tools indicated two different mechanisms for H2O2-
based epoxidation of electron-rich and electron-deficient C@C
bonds over mesoporous niobium silicates. Epoxidation of
electron-rich alkenes involves adsorption of H2O2 on coordina-
tively unsaturated Nb(V) sites, interaction with H2O2 to give a
hydroperoxo species NbOOH and water, followed by a concerted
electrophilic oxygen transfer from NbOOH to C@C bonds, produc-
ing epoxides and regenerating the initial state of the catalyst. This
mechanism is compatible with all the data gained in this study and
reported in the literature for mesoporous Nb silicates, specifically
the rate law established for CyO epoxidation (first order in catalyst
and fractional, <1, orders in alkene and oxidant), rate-retarding
effect of base, stereospecificity in the epoxidation of cis-alkenes,
and high heterolytic pathway selectivity in oxidation of alkenes
with highly reactive allylic hydrogen atoms, such as cyclohexene
and limonene. The hydroperoxo niobium species is manifested in
DRS UV–vis by an absorption band at 307 nm, which reveals a blue
shift after treatment with an aqueous solution of NaOAc.
Appendix A. Supplementary material
Supplementary data associated with this article can be found, in
the online version, at https://doi.org/10.1016/j.jcat.2017.09.011.
References
[1] R.A. Sheldon, M.C.A. van Vliet, in: Fine Chemicals through Heterogeneous
Catalysis, Wiley-VCH, Weinheim, 2001, p. 473.
[2] F. Cavani, J.H. Teles, Sustainability in catalytic oxidation: an alternative
approach or a structural evolution?, ChemSusChem 2 (2009) 508–534
[3] G. Strukul (Ed.), Catalytic Oxidations with Hydrogen Peroxide as Oxidant,
Kluwer Academic, Dordrecht, 1992.
[4] B.S. Lane, K. Burgess, Metal-catalyzed epoxidations of alkenes with hydrogen
peroxide, Chem. Rev. 103 (2003) 2457–2474.
[5] G. Grigoropoulou, J.H. Clark, J.A. Elings, Recent developments on the
epoxidation of alkenes using hydrogen peroxide as an oxidant, Green Chem.
5 (2003) 1–7.
[6] R. Sheldon, I.W.C.E. Arends, U. Hanefeld, Green Chemistry and Catalysis, Wiley-
VCH, Weinheim, 2007.
[7] M.G. Clerici, O.A. Kholdeeva (Eds.), Liquid Phase Oxidation via Heterogeneous
Catalysis: Organic Synthesis and Industrial Applications, Wiley, Hoboken, New
Jersey, 2013.
[8] B. Notari, in: D.D. Eley, W.O. Haag, B.C. Gates (Eds.), Adv. Catal. 41 (1996) 253–
334.
[9] C. Perego, A. Carati, P. Ingallina, M.A. Mantegazza, G. Bellussi, Production of
titanium containing molecular sieves and their application in catalysis, Appl.
Catal. A: Gen. 221 (2001) 63–72.
Several lines of evidence, including the established rate law
(zero order in substrate and first order in catalyst and H2O2), the
rate-accelerating effect of base, the formation of acetamide in the
amount close to that of epoxide, the lack of reaction in ethyl acet-
ate, and the detection of weak basic sites on the surface of niobium
silicates by IR spectroscopy using CDCl3 as a probe molecule, all

98
I.D. Ivanchikova et al. / Journal of Catalysis 356 (2017) 85–99
[10] For recent review see: M.G. Clerici, M.E. Domine, in: Liquid Phase Oxidation via
Heterogeneous Catalysis: Organic Synthesis and Industrial Applications,
Wiley, Hoboken, New Jersey, 2013, pp. 21–93.
[37] K.P. Bryliakov, E.P. Talsi, Active sites and mechanisms of bioinspired oxidation
with H2O2, catalyzed by non-heme Fe and related Mn complexes, Coord. Chem.
Rev. 276 (2014) 73–96.
[11] O.A. Kholdeeva, in: Liquid Phase Oxidation via Heterogeneous Catalysis:
Organic Synthesis and Industrial Applications, Wiley, Hoboken, New Jersey,
2013, pp. 127–219.
[12] O.A. Kholdeeva, Recent developments in liquid-phase selective oxidation using
environmentally benign oxidants and mesoporous metal silicates, Catal. Sci.
Technol. 4 (2014) 1869–1889.
[13] I. Nowak, B. Kilos, M. Ziolek, A. Lewandowska, Epoxidation of cyclohexene on
Nb-containing meso- and macroporous materials, Catal. Today 78 (2003) 487–
498.
[14] F. Somma, P. Canton, G. Strukul, Effect of the matrix in niobium-based aerogel
catalysts for the selective oxidation of olefins with hydrogen peroxide, J. Catal.
229 (2005) 490–498.
[38] M. Ziolek, P. Decyk, I. Sobczak, M. Trejda, J. Florek, H. Golinska, W. Klimas, A.
Wojtaszek, Catalytic performance of niobium species in crystalline and
amorphous solids—gas and liquid phase oxidation, Appl. Catal. A Gen. 391
(2011) 194–204.
[39] M. Ziolek, I. Sobczak, P. Decyk, K. Sobanska, P. Pietrzyk, Z. Sojka, Search for
reactive intermediates in catalytic oxidation with hydrogen peroxide over
amorphous niobium(V) and tantalum(V) oxides, Appl. Catal. B Environ. 164
(2015) 288–296.
[40] D.T. Bregante, P. Priyadarshini, D.W. Flaherty, Kinetic and spectroscopic
evidence for reaction pathways and intermediates for olefin epoxidation on
Nb in *BEA, J. Catal. 348 (2017) 75–89.
[41] I.D. Ivanchikova, M.K. Kovalev, M.S. Mel’gunov, A.N. Shmakov, O.A. Kholdeeva,
User-friendly synthesis of highly selective and recyclable mesoporous
titanium-silicate catalysts for the clean production of substituted p-
benzoquinones, Catal Sci. Technol. 4 (2014) 200–207.
[15] I. Nowak, M. Ziolek, Effect of texture and structure on the catalytic activity of
mesoporous niobosilicates for the oxidation of cyclohexene, Micropor.
Mesopor. Mater. 78 (2005) 281–288.
[16] F. Somma, G. Strukul, Niobium containing micro-, meso- and macroporous
silica materials as catalysts for the epoxidation of olefins with hydrogen
peroxide, Catal. Lett. 107 (2006) 73–81.
[42] T.R. Hughes, H.M. White, A study of the surface structure of decationized Y
zeolite by quantitative infrared spectroscopy, J. Phys. Chem. 71 (1967) 2192–
2201.
[17] A. Feliczak, K. Walczak, A. Wawrynczak, I. Nowak, The use of mesoporous
molecular sieves containing niobium for the synthesis of vegetable oil-based
products, Catal. Today 140 (2009) 23–29.
[43] C.A. Emeis, Determination of integrated molar extinction coefficients for
infrared absorption bands of pyridine adsorbed on solid acid catalysts, J. Catal.
141 (1993) 347–354.
[18] M. Selvaraj, S. Kawi, D.W. Park, C.S. Ha, A merit synthesis of well-ordered two-
dimensional mesoporous niobium silicate materials with enhanced
hydrothermal stability and catalytic activity, J. Phys. Chem. C 113 (2009)
7743–7749.
[44] F. Thibault-Starzyk, B. Gil, S. Aiello, T. Chevreau, J.-P. Gilson, In situ
thermogravimetry in an infrared spectrometer: an answer to quantitative
spectroscopy of adsorbed species on heterogeneous catalysts, Micropor.
Mesopor. Mater. 67 (2004) 107–112.
[19] A. Feliczak-Guzik, I. Nowak, Mesoporous niobosilicates serving as catalysts for
synthesis of fragrances, Catal. Today 142 (2009) 288–292.
[45] M. Tamura, K. Shimizu, A. Satsuma, Comprehensive IR study on acid/base
properties of metal oxides, Appl. Catal. A Gen. 433–434 (2012) 135–145.
[46] E.A. Paukshtis, E.N. Yurchenko, Study of the acid–base properties of
heterogeneous catalysts by infrared spectroscopy, Russ. Chem. Rev. 52
(1983) 242–258.
[20] I. Nowak, Frontiers in mesoporous molecular sieves containing niobium: from
model materials to catalysts, Catal. Today 192 (2012) 80–88.
[21] C. Gallo, R. Tiozzo, F. Psaro, M. Carniato, Guidotti, Niobium metallocenes
deposited onto mesoporous silica via dry impregnation as catalysts for
selective epoxidation of alkenes, J. Catal. 298 (2013) 77–83.
[22] C. Tiozzo, C. Bisio, F. Carniato, L. Marchese, A. Gallo, N. Ravasio, R. Psaro, M.
Guidotti, Epoxidation with hydrogen peroxide of unsaturated fatty acid
methyl esters over Nb(V)-silica catalysts, Eur. J. Lipid Sci. Technol. 115
(2013) 86–93.
[47] E.A. Paukshtis, Infrakrasnaya Spektroskopiya
v Geterogennom Kislotno-
Osnovnom Katalize (Infrared Spectroscopy in Heterogeneous Acid-Base
Catalysis), Nauka, Novosibirsk, 1992 (in Russian).
[48] V.N. Panchenko, M.M. Matrosova, J. Jeon, J.W. Jun, M.N. Timofeeva, S.H. Jhung,
Catalytic behavior of metal–organic frameworks in the Knoevenagel
condensation reaction, J. Catal. 316 (2014) 251–259.
[23] C. Tiozzo, C. Bisio, F. Carnato, A. Gallo, S.L. Scott, R. Psaro, M. Guidotti,
Niobium–silica catalysts for the selective epoxidation of cyclic alkenes: the
generation of the active site by grafting niobocene dichloride, Phys. Chem.
Chem. Phys. 15 (2013) 13354–13362.
[49] I.Y. Skobelev, O.V. Zalomaeva, O.A. Kholdeeva, J.M. Poblet, J.J. Carbó,
Mechanism of thioether oxidation over di- and tetrameric Ti centres: kinetic
and DFT studies based on model Ti-containing polyoxometalates, Chem. Eur. J.
21 (2015) 14496–14506.
[24] C. Tiozzo, C. Bisio, F. Carniato, M. Guidotti, Grafted non-ordered niobium-silica
materials: versatile catalysts for the selective epoxidation of various
unsaturated fine chemicals, Catal. Today 235 (2014) 49–57.
[25] R. Turco, A. Aronne, P. Carniti, A. Gervasini, L. Minieri, P. Pernice, R. Tesser, R.
Vitiello, M. Di Serio, Influence of preparation methods and structure of
niobium oxide-based catalysts in the epoxidation reaction, Catal. Today 254
(2015) 99–103.
[50] G.B. Shul’pin, Metal-catalyzed hydrocarbon oxygenations in solutions: the
dramatic role of additives: a review, J. Mol. Catal. A Chem. 189 (2002) 39–66.
[51] H. Knözinger, Infrared spectroscopy for the characterization of surface acidity
and basicity, in: G. Ertl, H. Knözinger, F. Schüth, J. Weitkamp (Eds.), Handbook
of Heterogeneous Catalysis, second ed., Wiley-VCH, Weinheim, 2008, p. 1135.
[52] M. Trejda, A. Tuel, J. Kujawa, B. Kilos, M. Ziolek, Niobium rich SBA-15 materials
–
preparation, characterisation and catalytic activity, Micropor. Mesopor.
[26] C. Tiozzo, C. Palumbo, R. Psaro, C. Bisio, F. Carniato, A. Gervasini, P. Carniti, M.
Guidotti, The stability of niobium-silica catalysts in repeated liquid-phase
epoxidation tests: a comparative evaluation of in-framework and grafted
mixed oxides, Inorg. Chim. Acta 431 (2015) 190–196.
Mater. 110 (2008) 271–278.
[53] F. Tielens, T. Shishido, S. Dzwigaj, What do the niobium framework sites look
like in redox zeolites? A combined theoretical and experimental investigation,
J. Phys. Chem. C 114 (2010) 3140–3147.
[27] A. Ramanathan, R. Masheswari, B. Subramanian, Facile styrene Epoxidation
with H2O2 over novel niobium containing cage type mesoporous silicate, Nb-
KIT-5, Top. Catal. 58 (2015) 314–324.
[54] A. Ramanathan, R. Macheswari, D.H. Barich, B. Subramaniam, Niobium
incorporated mesoporous silicate, Nb-KIT-6: synthesis and characterization,
Micropor. Mesopor. Mater. 190 (2014) 240–247.
[28] A. Ramanathan, H. Zhu, R. Maheswari, P.S. Thapa, B. Subramaniam,
Comparative study of Nb-incorporated cubic mesoporous silicates as
epoxidation catalysts, Ind. Eng. Chem. Res. 54 (2015) 4236–4242.
[29] I.D. Ivanchikova, N.V. Maksimchuk, I.Y. Skobelev, V.V. Kaichev, O.A. Kholdeeva,
Mesoporous niobium-silicates prepared by evaporation-induced self-
assembly as catalysts for selective oxidations with aqueous H2O2, J. Catal.
332 (2015) 138–148.
[55] M. Ziolek, I. Sobczak, I. Nowak, P. Decyk, A. Lewandowska, J. Kujawa, Nb-
containing mesoporous molecular sieves—a possible application in the
catalytic processes, Micropor. Mesopor. Mater. 35 (36) (2000) 195–207.
[56] M. Ziolek, I. Sobczak, A. Lewandowska, I. Nowak, P. Decyk, M. Renn, B.
Jankowska, Oxidative properties of niobium-containing mesoporous silica
catalysts, Catal. Today 70 (2001) 169–181.
[57] J. Datka, A.M. Turek, J.M. Jehng, I.E. Wachs, Acidic properties of supported
niobium oxide catalysts: an infrared spectroscopy investigation, J. Catal. 135
(1992) 186–199.
[58] H. Zhu, A. Ramanathan, J.F. Wu, R.V. Chaudhari, B. Subramaniam, Effects of
tunable acidity and basicity of Nb-KIT-6 catalysts on ethanol conversion:
experiments and kinetic modeling, AIChE J. 63 (2017) 2888–2899.
[59] N.N. Trukhan, A.A. Panchenko, E. Roduner, M.S. Melgunov, O.A. Kholdeeva, J.
Mrowiec-Białon´ , A.B. Jarze˛bski, FTIR spectroscopic study of titanium-
containing mesoporous silicate materials, Langmuir 21 (2005) 10545–10554.
[60] E.A. Paukshtis, N.S. Kotsarenko, L.G. Karakchiev, Investigation of proton-
acceptor properties of oxide surfaces by IR spectroscopy of hydrogen-bonded
complexes, React. Kinet. Catal. Lett. 12 (1979) 315–319.
[61] R.A. Sheldon, M. Wallau, I.W.C.E. Arends, U. Schuchardt, Heterogeneous
catalysts for liquid-phase oxidations: philosophers’ stones or Trojan horses?,
Acc Chem. Res. 31 (1998) 485–493.
[62] P. Atkins, J. De Paula, Elements of Physical Chemistry, sixth ed., Oxford
University Press, 2013.
[63] W. Yan, A. Ramanathan, P.D. Patel, S.K. Maiti, B.B. Laird, W.H. Thompson, B.
Subramaniam, Mechanistic insights for enhancing activity and stability of Nb-
incorporated silicates for selective ethylene epoxidation, J. Catal. 336 (2016)
75–84.
[64] S.K. Maiti, A. Ramanathan, W.H. Thompson, B. Subramaniam, Strategies to
passivate Brønsted acidity in Nb-TUD-1 enhance hydrogen peroxide
[30] N.E. Thornburg, A.B. Thompson, J.M. Notestein, Periodic trends in highly
dispersed groups IV and
epoxidation with H2O2, ACS Catal. 5 (2015) 5077–5088.
[31] N.E. Thornburg, S.L. Nauert, A.B. Thompson,
V supported metal oxide catalysts for alkene
J.M.
Notestein,
Synthesisꢁstructure–function relationships of silica-supported niobium(V)
catalysts for alkene epoxidation with H2O2, ACS Catal. 6 (2016) 6124–6134.
[32] A. Aronne, M. Turco, G. Bagnasco, G. Ramis, E. Santacesaria, M. Di Serio, E.
Marenna, M. Bevilacqua, C. Cammarano, E. Fanelli, Gel derived niobium–
silicon mixed oxides: characterization and catalytic activity for cyclooctene
epoxidation, Appl. Catal. A Gen. 347 (2008) 179–185.
[33] S.T. Oyama (Ed.), Mechanisms in Homogeneous and Heterogeneous
Epoxidation Catalysis, Elsevier, Amsterdam, 2008.
[34] N. Rösch, C. Di Valentin, I.V. Yudanov, in: F. Maseras, A. Lledós (Eds.),
Computational Modelling of Homogeneous Catalysis, Kluwer Academic
Publishers, Dordrecht, 2002, Ch. 12, pp. 289–324.
[35] N. Antonova, J.J. Carbó, U. Kortz, O.A. Kholdeeva, J.M. Poblet, Mechanistic
insights into alkene epoxidation with H2O2 by Ti- and other TM-containing
polyoxometalates: role of the metal nature and coordination environment, J.
Am. Chem. Soc. 132 (2010) 7488–7497.
[36] L. Que Jr., W.B. Tolman, Biologically inspired oxidation catalysis, Nature 455
(2008) 333–340.

I.D. Ivanchikova et al. / Journal of Catalysis 356 (2017) 85–99
99
utilization and reduce metal leaching during ethylene epoxidation, Ind. Eng.
Chem. Res. 56 (2017) 1999–2007.
[polyaminocarboxylato bis(N-oxido)]niobate(V) complexes, Eur. J. Inorg.
Chem. 737–743 (2003).
[65] P. Chagas, H.S. Oliveira, R. Mambrini, M. Le Hyaric, M.V. de Almeida, L.C.A.
[84] A. Maniatakou, C. Makedonas, C.A. Mitsopoulou, C. Raptopoulou, I. Rizopoulou,
A. Terzis, A. Karaliota, Synthesis, structural and DFT studies of a peroxo-
niobate complex of the biological ligand 2-quinaldic acid, Polyhedron 27
(2008) 3398–3408.
[85] N. Adler, C.F. Hiskey, Spectra of the peroxy complexes of niobium in sulfuric
acid, J. Am. Chem. Soc. 79 (1957) 1827–1830.
Oliveira,
A novel hydrofobic niobium oxyhydroxide as catalyst: Selective
cyclohexene oxidation to epoxide, Appl. Catal. A Gen. 454 (2013) 88–92.
[66] H. Shima, M. Tanaka, H. Imai, T. Yokoi, T. Tatsumi, J.N. Kondo, IR observation of
selective oxidation of cyclohexene with H2O2 over mesoporous Nb2O5, J. Phys.
Chem. C 113 (2009) 21693–21699.
[67] C.W. Yoon, K.F. Hirsekorn, M.L. Neidig, X. Yang, T.D. Tilley, Mechanism of the
decomposition of aqueous hydrogen peroxide over heterogeneous TiSBA15
and TS-1 selective oxidation catalysts: insights from spectroscopic and density
functional theory Studies, ACS Catal. 1 (2011) 1665–1678.
[68] R.A. Sheldon, J.K. Kochi, Metal-Catalyzed Oxidations of Organic Compounds,
Academic Press, New York, 1981.
[86] N. Adler, C.F. Hiskey, Composition of the peroxy complexes of niobium(V) in
sulfuric acid, J. Am. Chem. Soc. 79 (1957) 1831–1834.
[87] Y. Narendar, G.L. Messing, Synthesis, decomposition and crystallization
characteristics of peroxoꢁcitratoꢁniobium: an aqueous niobium precursor,
Chem. Mater. 9 (1997) 580–587.
[88] Q. Geng, Q. Liu, P. Ma, J. Wang, J. Niu, Synthesis, crystal structure and
[69] M. Guidotti, E. Gavrilova, A. Galarneau, B. Coq, R. Psaro, N. Ravasio, Epoxidation
of methyl oleate with hydrogen peroxide. The use of Ti-containing silica solids
as efficient heterogeneous catalysts, Green Chem. 13 (2011) 1806–1811.
[70] O.A. Kholdeeva, V.A. Grigoriev, G.M. Maksimov, M.A. Fedotov, A.V. Golovin, K.I.
Zamaraev, Polyfunctional action of transition metal substituted
heteropolytungstates in alkene epoxidation by molecular oxygen in the
presence of aldehyde, J. Mol. Catal. A Chem. 114 (1996) 123–130.
[71] L.J. Burcham, J. Datka, I.E. Wachs, In situ vibrational spectroscopy studies of
supported niobium oxide catalysts, J. Phys. Chem. B 103 (1999) 6015–6024.
[72] X.T. Gao, I.E. Wachs, M.S. Wong, J.Y. Ying, Structural and reactivity properties
of Nb-MCM-41: comparison with that of highly dispersed Nb2O5/SiO2
catalysts, J. Catal. 203 (2001) 18–24.
[73] A. Wojtaszek, M. Ziolek, F. Tielens, Probing acid–base properties in group V
aluminum containing zeolites, J. Phys. Chem. C 116 (2012) 2462–2468.
[74] D.C. Tranca, A. Wojtaszek-Gurdak, M. Ziolek, F. Tielens, Supported and inserted
monomeric niobium oxide species on/in silica: a molecular picture, Phys.
Chem. Chem. Phys. 17 (2015) 22402–22411.
photocatalytic properties of an unprecedented arsenic-disubstituted
4.5ꢁ
Lindqvist-type peroxopolyoxoniobate ion: {As2Nb4(O2)4O14H1.5
}
, Dalton
Trans. 43 (2014) 9843–9846.
[89] S.R. Gogoi, J.J. Boruah, G. Sengupta, G. Saikia, K. Ahmed, K.K. Bania, N.S. Islam,
Peroxoniobium(V)-catalyzed selective oxidation of sulfides with hydrogen
peroxide in water: a sustainable approach, Catal. Sci. Technol. 5 (2015) 595–
610.
[90] A. Zecchina, S. Bordiga, C. Lamberti, G. Ricchiardi, C. Lamberti, G. Ricchiardi, D.
Scarano, G. Petrini, G. Leofanti, M. Mantegazza, Structural characterization of
Ti centres in Ti-silicalite and reaction mechanisms in cyclohexanone
ammoximation, Catal. Today 32 (1996) 97–106.
[91] D.A. Ruddy, T.D. Tilley, Kinetics and mechanism of olefin epoxidation with
aqueous H2O2 and a highly selective surface-modified TaSBA15 heterogeneous
catalyst, J. Am. Chem. Soc. 130 (2008) 11088–11096.
[92] O.V. Zalomaeva, N.N. Trukhan, I.D. Ivanchikova, A.A. Panchenko, E. Roduner, E.
P. Talsi, A.B. Sorokin, V.A. Rogov, O.A. Kholdeeva, EPR study on the mechanism
of H2O2-based oxidation of alkylphenols over titanium single-site catalysts, J.
Mol. Catal. A Chem. 277 (2007) 185–192.
[75] M.G. Clerici, Oxidation of saturated hydrocarbons with hydrogen peroxide,
catalysed by titanium silicalite, Appl. Catal. 68 (1991) 249–261.
[76] M.G. Clerici, P. Ingallina, Epoxidation of lower olefins with hydrogen peroxide
and titanium silicalite, J. Catal. 140 (1993) 71–83.
[93] G.B. Payne, P.H. Deming, P.H. Williams, Reactions of hydrogen peroxide. VII.
Alkali-catalyzed epoxidation and oxidation using a nitrile as co-reactant, J.
Org. Chem. 26 (1961) 659–663.
[77] S. Bordiga, A. Damin, F. Bonino, G. Ricchiardi, C. Lamberti, A. Zecchina, The
structure of the peroxo species in the TS-1 catalyst as investigated by resonant
Raman spectroscopy, Angew. Chem. Int. Ed. 41 (2002) 4734–4931.
[78] W. Lin, H. Frei, Photochemical and FT-IR probing of the active site of hydrogen
peroxide in Ti silicalite sieve, J. Am. Chem. Soc. 124 (2002) 9292–9298.
[79] S. Bordiga, A. Damin, F. Bonino, C. Lamberti, Single site catalyst for partial
oxidation reaction: TS-1 case study, Top. Organomet. Chem. 16 (2005) 37–68.
[80] R.L. Brutchey, D.A. Ruddy, L.K. Andersen, T.D. Tilley, Influence of surface
modification of TiꢁSBA15 catalysts on the epoxidation mechanism for
cyclohexene with aqueous hydrogen peroxide, Langmuir 21 (2005) 9576–
9583.
[94] M. Guidotti, C. Pirovano, N. Ravasio, B. Lґazaro, J.M. Fraile, J.A. Mayoral, B. Coq,
A. Galarneau, The use of H2O2 over titanium-grafted mesoporous silica
catalysts: a step further towards sustainable epoxidation, Green Chem. 11
(2009) 1421–1427.
[95] S. Ueno, K. Yamaguchi, K. Yoshida, K. Ebitani, K. Kaneda, Hydrotalcite catalysis:
heterogeneous epoxidation of olefins using hydrogen peroxide in the presence
of nitriles, Chem. Commun. 295–296 (1998).
[96] M.A. Aramendía, V. Borau, C. Jiménez, J.M. Luque, J.M. Marinas, J.R. Ruiz, F.J.
Urbano, Epoxidation of limonene over hydrotalcite-like compounds with
hydrogen peroxide in the presence of nitriles, Appl. Catal. A Gen. 216 (2001)
257–265.
[81] E. Gianotti, C. Bisio, L. Marchese, M. Guidotti, N. Ravasio, R. Psaro, S. Coluccia, Ti
(IV) catalytic centers grafted on different siliceous materials: spectroscopic
and catalytic study, J. Phys. Chem. C 111 (2007) 5083–5089.
[82] M.K. Harrup, G.-S. Kim, H. Zeng, R.P. Johnson, D. VanDerveer, C.L. Hill,
Triniobium polytungstophosphates. Syntheses, structures, clarification of
isomerism and reactivity in the presence of H2O2, Inorg. Chem. 37 (1998)
5550–5556.
[97] M.D. Romero, J.A. Calles, M.A. Ocaña, J.M. Gómez, Epoxidation of cyclohexene
over basic mixed oxides derived from hydrotalcite materials: activating agent,
solvent and catalyst reutilization, Micropor. Mesopor. Mater. 111 (2008) 243–
253.
[98] M.N. Timofeeva, A.E. Kapustin, V.N. Panchenko, E.O. Butenko, V.V. Krupskaya,
A. Gil, M.A. Vicente, Synthetic and natural materials with the brucite-like
layers as high active catalyst for synthesis of 1-methoxy-2-propanol from
methanol and propylene oxide, J. Mol. Catal. A Chem. 423 (2016) 22–30.
[83] D. Bayot, B. Tinant, B. Mathieu, J.-P. Declercq, M. Devillers, Spectroscopic and
structural
characterizations
of
novel
water-soluble
peroxo