Well-controlled radical-based epoxidation catalyzed by copper complex immobilized on bipyridine-periodic mesoporous organosilica

Article abstract of DOI:10.1016/j.apcata.2019.02.007

The development of synthetic methods and reaction systems for safe radical reactions is of extremely industrial importance. Here we proposed new concept for a safe radical reaction system based on combined use of a mesoporous catalyst and an insoluble solid scavenger. We selected Mukaiyama epoxidation of olefin as a model radical reaction and investigated the catalysis of Cu-bipyridine complexes immobilized on trimethylsilylated bipyridine-periodic mesoporous organosilica as a solid support. The immobilized Cu complex exhibited high catalytic activity and reusability for Mukaiyama epoxidation at low substrate concentration (1 mmol) but free-radical auto-oxidation also occurred at high substrate concentration (7 mmol). Although both epoxidation reactions outside and inside the mesochannels were almost completely quenched by addition of molecular scavenger, addition of solid scavenger allowed quenching the reaction outside the mesopores but not inside the mesopores because the solid scavenger could not access the interior of the mesochannels. Thus, the combined use of a mesoporous catalyst and a solid radical scavenger would offer new reaction system for safe radical reactions.

Full text of DOI:10.1016/j.apcata.2019.02.007

Accepted Manuscript
Title: Well-Controlled Radical-Based Epoxidation Catalyzed
by Copper Complex Immobilized on Bipyridine-Periodic
Mesoporous Organosilica
Authors: Satoshi Ishikawa, Yoshifumi Maegawa, Minoru
Waki, Shinji Inagaki
PII:
S0926-860X(19)30062-6
DOI:
Reference:
https://doi.org/10.1016/j.apcata.2019.02.007
APCATA 16977
To appear in:
Applied Catalysis A: General
Received date:
Revised date:
Accepted date:
29 December 2018
1 February 2019
5 February 2019
Please cite this article as: Ishikawa S, Maegawa Y, Waki M, Inagaki S, Well-
Controlled Radical-Based Epoxidation Catalyzed by Copper Complex Immobilized on
Bipyridine-Periodic Mesoporous Organosilica, Applied Catalysis A, General (2019),
https://doi.org/10.1016/j.apcata.2019.02.007
This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that
apply to the journal pertain.

Well-Controlled Radical-Based Epoxidation Catalyzed by Copper Complex
Immobilized on Bipyridine-Periodic Mesoporous Organosilica
Satoshi Ishikawa,a,‡ Yoshifumi Maegawa,a Minoru Waki,a and Shinji Inagaki*a
a Toyota Central R&D Labs., Inc., Nagakute, Aichi 480-1192, Japan.
E-mail: inagaki@mosk.tytlabs.co.jp

Graphical Abstract
Quenching of radicals by
solid radical scavenger
Mukaiyama-epoxidation
inside pore space
Auto-oxidation
outside pore space
Cu@BPy-PMO-TMS
Solid radical scavenger
Highlights
A new concept for a safe radical reaction system
Cu-bipyridine catalyst was directly formed on the pore surfaces of BPy-PMO
The catalyst exhibited good catalytic activity and reusability for epoxidation
Addition of solid scavenger allowed the reaction solely inside the mesopore space
Abstract
The development of synthetic methods and reaction systems for safe radical reactions
is of extremely industrial importance. Here we proposed new concept for a safe radical reaction
system based on combined use of a mesoporous catalyst and an insoluble solid scavenger. We
selected Mukaiyama epoxidation of olefin as a model radical reaction and investigated the
catalysis of Cu-bipyridine complexes immobilized on trimethylsilylated bipyridine-periodic
mesoporous organosilica as a solid support. The immobilized Cu complex exhibited high

catalytic activity and reusability for Mukaiyama epoxidation at low substrate concentration (1
mmol) but free-radical auto-oxidation also occurred at high substrate concentration (7 mmol).
Although both epoxidation reactions outside and inside the mesochannels were almost
completely quenched by addition of molecular scavenger, addition of solid scavenger allowed
quenching the reaction outside the mesopores but not inside the mesopores because the solid
scavenger could not access the interior of the mesochannels. Thus, the combined use of a
mesoporous catalyst and a solid radical scavenger would offer new reaction system for safe
radical reactions.
Keywords: mesoporous materials; metal complex catalysts; radical reactions; Mukaiyama
epoxidation
1. Introduction
Radical reactions have attracted much attention because radicals can react with most
organic molecules and afford reaction products such as sterically hindered molecules and unique
polymers that cannot be synthesized by typical catalytic reactions [1]. The utilization of radicals
for organic reactions is expected to expand the scope of organic synthesis. However, radicals are
sometimes too active in nature and can cause thermal runaway, fire, and explosions [2,3].

Product selectivity is sometimes low due to the undesired side reactions. Therefore, the
development of synthetic methods and reaction systems for safe radical reactions has been
highly desirable.
Ramos-Fernandez et al. reported the titania-catalyzed oxidative dehydrogenation of
ethyl lactate based on a free-radical reaction mechanism [4]. To improve the product yield and
minimize by-product formation, the reaction was performed in the presence of activated carbon
as a solid radical scavenger. The solid radical scavenger efficiently hampered the formation of
undesired products in the solution without a loss of catalysis on TiO2 for the desired reaction.
This unique strategy encouraged us to examine the radical reaction using a heterogeneous metal
catalyst and a solid radical scavenger.
We recently reported the preparation of a periodic mesoporous organosilica including
2,2’-bipyridine (BPy) ligands within the framework (BPy-PMO, Fig. 1) [5]. BPy-PMO has a
highly ordered mesoporous structure with a unique pore wall structure, in which BPy ligands
are densely and regularly packed, and exposed on the pore surfaces. BPy ligands have high
coordination ability; therefore, various metal-bipyridine complexes (e.g. Ir, Ru, Rh, and Mo,
etc) can be formed directly on the pore surfaces of BPy-PMO [6-10]. The metal
complex-immobilized BPy-PMOs exhibit efficient heterogeneous catalysis for various reactions
including direct C-H borylation of arenes [5,6], selective oxidation of alkanes [7], water

oxidation [8], transfer hydrogenation of nitrogen heterocycles [9], and epoxidation of olefins
[10]. However, there have been no reports on radical-based reactions using BPy-PMO as a
catalyst support.
Here, we report the immobilization of a Cu complex on BPy-PMO and application to a
radical-based epoxidation reaction (Mukaiyama epoxidation [11-13]). A Cu-bipyridine complex
was formed on the pore surfaces of BPy-PMO by reaction with CuCl2. The Cu complex
immobilized on BPy-PMO exhibits efficient catalysis for the Mukaiyama epoxidation of
cyclohexene with molecular oxygen in the presence of isobutyraldehyde. End-capping of the
surface silanol groups with trimethylsilyl groups was very effective to improve the activity and
recyclability for Mukaiyama epoxidation. Although the Mukaiyama epoxidation proceeded
efficiently at low substrate concentration (1 mmol), the radical auto-oxidation reaction occurred
at high substrate concentration (7 mmol), even in the absence of the catalyst.
Polystyrene-supported bipyridine (PS-BPy) beads were added as a solid radical scavenger to
control the undesired side reaction. The addition of the solid radical scavenger quenched the
reaction in the solution but did not quench the reaction inside the mesopore spaces of BPy-PMO
(Fig. 1). The combined use of a solid scavenger and a mesoporous catalyst may lead to the
possibility of controllable radical reactions.

Quenching of radicals by
solid radical scavenger
Mukaiyama-epoxidation
inside pore space
Auto-oxidation
outside pore space
Cu@BPy-PMO-TMS
Solid radical scavenger
Fig. 1. Schematic image of the copper-immobilized BPy-PMO nanoreactor for Mukaiyama
epoxidation in combination with a solid radical scavenger.
2. Experimental methods
2.1. Materials and methods

BPy-PMO and end-capped BPy-PMO with trimethylsilyl groups (BPy-PMO-TMS)
were prepared according to the literature, respectively [5]. PS-BPy beads (100–200 mesh, Fig.
S1, see Supplementary Material) were purchased from Sigma-Aldrich.
All other chemicals
were commercially available and were used without purification. Gas chromatography-mass
spectrometry (GC-MS; Agilent 6890GC/5973MSD) analyses were performed using a flame
ionization detector (FID) and a capillary column (HP-5MS, 0.25 mm × 30 m). Nitrogen
adsorption experiments were performed on a Nova3000e (Quantachrome) at liquid nitrogen
temperature. Pore-size distribution curves were calculated using the density functional theory
(DFT) (DFT kernel: silica, cylindrical pores, nonlinear DFT (NLDFT) equilibrium model).
Specific surface areas were calculated from the linear regions of Brunauer-Emmett-Teller (BET)
plots (P/P0 = 0.1–0.25). The t-plot method was applied to estimate the pore volumes. Powder
X-ray diffraction (XRD) measurements were performed on a Rigaku RINT-TTR with Cu-Kα
radiation (50 kV, 300 mA). Scanning electron microscopy (SEM) measurements were
performed on a Hitachi S-3600N equipped with energy dispersive X-ray spectroscopy (EDX)
detector. Ultraviolet-visible (UV-vis) spectroscopy measurements were performed on a Jasco
V-670 spectrophotometer. X-ray absorption fine structure (XAFS) measurements at the Cu
K-edge were performed at the BL14B2 beamline of SPring-8 and the data were collected in
transmission mode. Synchrotron radiation X-rays were monochromatized by a Si(311)

double-crystal monochromator. Ion chambers were placed front and behind the samples to
detect the intensity of incident and transmitted X-rays (I0 and I, respectively). Background
reduction was conducted using the Autobk and Spline smoothing algorithm. The κ3-weighted
extended XAFS (EXAFS) oscillations (2.0-12 Å) were Fourier transformed into R-space. The
EXAFS data was fitted in R-space (1.5-2.2 Å). The fitting parameters were the coordination
number (CN), the interatomic distance (R), the correction-of-edge energy (ΔE0), and the
Debye-Waller factor (σ2). Backscattering amplitudes and Phase-shift were calculated using the
FEFF6 code.
2.2. Formation of Cu complexes on BPy-PMOs
To a suspension of BPy-PMO (47.8 mg, 0.15 mmol) or BPy-PMO-TMS (68.5 mg,
0.20 mmol) in acetonitrile (8 mL) was added a solution of CuCl2 in acetonitrile (10 mM, 2 mL,
0.02 mmol). The reaction mixture was then stirred (500 rpm) at room temperature for 24 h. The
obtained powder was filtered, washed with acetonitrile, and dried under reduced pressure to
give Cu-BPy-PMO or Cu-BPy-PMO-TMS.
2.3. Catalytic reaction
To a solution of cyclohexene (0.10 mL, 1.0 mmol), isobutyraldehyde (0.27 mL, 3.0

mmol) and chloroform (9.62 mL) was added Cu-BPy-PMO (20 mg) or Cu-BPy-PMO-TMS (20
mg). The reaction flask as then set in an oil bath kept at 40 °C. After attaching a balloon filled
with 1 L of oxygen, the reaction was started. Aliquots of the reaction solution were collected
periodically and centrifuged to separate the solid materials. The supernatant was then mixed
with a known amount of toluene as an external standard for analysis by GC (GC-8A, Shimadzu)
with a flame ionization detector (FID) and a capillary column (HP-5, 0.53 mm × 30 m). After
the reaction was completed, the solid catalyst was filtered, washed with acetonitrile, and dried
under evacuation. The recovered catalyst was then reused in the next reaction under the same
conditions.
3. Results and discussion
3.1. Immobilization of Cu complex on BPy-PMO

To immobilize Cu complexes on the pore surfaces, BPy-PMO powder was dispersed
in acetonitrile, methanol, or deionized water containing CuCl2 and then stirred at room
temperature for 24 h. XRD patterns of the recovered Cu-immobilized BPy-PMOs were
measured to confirm whether the ordered mesoporous structure of BPy-PMO was preserved or
not. The ordered mesoporous structure was preserved when acetonitrile was used as a solvent.
The XRD pattern showed a peak at 2θ = 1.78° due to the ordered mesostructure and also peaks
at 2θ = 7-35° due to the periodic arrangement of the bipyridine groups in the pore walls, similar
to those of the original BPy-PMO (Fig. 2). However, the mesoporous structure was collapsed
when methanol or deionized water were used as solvents (Fig. S2, see Supplementary Material).
BPy-PMO powder was stirred in methanol or deionized water without the addition of CuCl2 to
check the stability of BPy-PMO in these solvents. As a result, the ordered mesoporous and pore
wall structures were completely preserved, which indicates that the presence of CuCl2 in protic
solvents such as methanol and deionized water would damage the mesoporous structure.
BPy-PMO-TMS was also treated with CuCl2 in acetonitrile. The XRD pattern of Cu-BPy-PMO-
TMS was almost the same as that of the pristine BPy-PMO-TMS, which indicates the
preservation of both the ordered mesoporous structure and the molecular-scale periodicity in the
pore walls (Fig. 2).

(a)
(b)
5
10 15 20 25 30 35
1
2
3
4
5
2θ / degree (Cu-Kα)
2θ / degree (Cu-Kα)
Fig. 2. (a) Low and (b) medium scattering-angle XRD patterns of BPy-PMO (blue line),
Cu-BPy-PMO (red line), BPy-PMO-TMS (green line), and Cu-BPy-PMO-TMS (purple line).
Fig. 3 shows nitrogen adsorption/desorption isotherms for BPy-PMO and
BPy-PMO-TMS before and after immobilization of the Cu complexes. All were typical type IV
isotherms, which indicate the presence of uniform mesoporosity, although the adsorption
amounts decreased after the formation of Cu complexes on the pore surfaces. Table 1 lists the
BET surface areas (SBET), density functional theory pore diameters (dDFT) and mesopore
volumes (Vp) of Cu-immobilized BPy-PMO and BPy-PMO-TMS. Cu-BPy-PMO and
Cu-BPy-PMO-TMS have high SBET of 573 and 597 m2 g-1, and dDFT of 4.1 and 3.8 nm,
respectively, which were almost the same or slightly smaller than those of pristine BPy-PMO
and BPy-PMO-TMS, respectively.

(a)600
(b) 400
500
300
400
300
200
100
0
200
100
0
0
0.2 0.4 0.6 0.8
1
0
0.2 0.4 0.6 0.8
1
Relative Pressure / P/P0
Relative Pressure / P/P0
Fig. 3. Nitrogen adsorption/desorption isotherms of (a) BPy-PMO (blue line) and Cu-BPy-PMO
(red line), (b) BPy-PMO-TMS (green line), and Cu-BPy-PMO-TMS (purple line). Closed
circles, adsorption; open circles, desorption.
The amounts of Cu loading were measured using EDX and are listed in Table 1. EDX
analysis showed that the Cu/Si atomic ratios were 0.07 and 0.04 for Cu-BPy-PMO and
Cu-BPy-PMO-TMS, respectively. The Si/BPy molar ratio was 2/1 for BPy-PMO and 10/3 for
BPy-PMO-TMS [5]; therefore, the molar ratios of Cu to the total BPy ligands (Cu/BPy) were
calculated to be 0.14 and 0.13, respectively (assuming that the silanol groups (Si-OH) at the
pore surfaces (2/3 of the total Si) were completely trimethylsilylated for BPy-PMO-TMS [5]).

The Cu complexes are formed only on the BPy ligands exposed on the surface because the pore
walls are composed of three layers of Si-BPy-Si units. Thus, the molar ratio of Cu to the surface
BPy ligands (Cu/BPysurf) were 0.21 and 0.20 for Cu-BPy-PMO and Cu-BPy-PMO-TMS,
respectively, which indicates that approximately 20% of the surface BPy ligands coordinate
with Cu. The Cl/Cu atomic ratios were 1.0 for both samples, which were apparently lower that
of the ideal ratio of 2 (CuCl2). This indicates the loss of Cl anions during the metalation process.
Table 1. Physical parameters of Cu-immobilized BPy-PMO and BPy-PMO-TMS
a
a
a
SBET
dDFT
nm
4.3
4.1
3.8
3.8
Vp
b
Sample
Cu/Sib
Cu/BPysurf
Cl/Cub
m2 g-1
693
cc g-1
0.50
0.38
0.41
0.36
BPy-PMO
Cu-BPy-PMO
-
-
-
573
0.07
-
0.23
-
1.0
-
BPy-PMO-TMS
Cu-BPy-PMO-TMS
654
597
0.04
0.22
1.0
aMeasured by N2 adsorption at liquid N2 temperature (-196 °C). SBET calculated by the BET
b
method. dDFT calculated by the DFT method. Vp calculated by the t-plot method. Estimated
from EDX measurement. Cu/BPysurf calculated by taking into account the following
assumptions: (1) the crystal wall of BPy is composed of three layers; (2) the Si/BPy ratio in the

crystal wall is 2 in BPy-PMO, and (3) 10/3 in BPy-PMO-TMS.
The electronic structure of the Cu-bipyridine complex in Cu-BPy-PMO and
Cu-BPy-PMO-TMS was evaluated using UV-vis spectroscopy. Fig. S3 shows UV-vis absorption
spectra for Cu-BPy-PMO and BPy-PMO-TMS together with Cu(bpy)Cl2 as a model complex
[14]. Cu(bpy)Cl2 exhibited an absorption centred at 710 nm, which is attributable to the d-d
transitions of Cu2+ in a tetragonally distorted octahedral configuration [15,16]. Cu-BPy-PMO
and Cu-BPy-PMO-TMS showed a broad absorption centred at 770 nm. The peak broadening
and red-shift of the absorption band compared with that for Cu(bpy)Cl2 may be due to the
changes in electronic state of Cu2+ by the attachment of silicon atoms at both sides of the
bipyridine ligand [17]. Similar changes in UV-vis spectra were observed in other
metal-immobilized BPy-PMO.5 Cu-BPy-PMO also had a shoulder peak at 980 nm. Taking into
account there was no absorption in Cu-BPy-PMO-TMS, the additional absorption may be
caused by the formation of an unidentified copper complex by reaction between the silanol
groups and CuCl2.
The coordination state of the copper centres in Cu-BPy-PMO and Cu-BPy-PMO-TMS
were studied using XAFS measurements. X-ray absorption near-edge spectroscopy (XANES)
measurements at the Cu K-edge of Cu-BPy-PMO and Cu-BPy-PMO-TMS showed a

characteristic weak peak attributable to the 1s→3d transition at 8977 eV, as well as that for
Cu(bpy)Cl2 (Fig. S4, see Supplementary Material). The absorption edge in the rising part was
observed at 8985.2, 8984.6, and 8984.9 eV for Cu-BPy-PMO, Cu-BPy-PMO-TMS, and
Cu(bpy)Cl2, respectively. The chemical shifts relative to the K-edge energy of Cu foil (8979 eV)
were 5.6-6.2 eV, which is characteristic value of +2 oxidation state of the copper centre.
Curve-fitting analyses of EXAFS Fourier transforms were conducted using the
available crystal structure of Cu(bpy)Cl2 as a model complex to generate the theoretical model
(Figs. S5-6, see Supplementary Material) [18]. The radial structural functions of
Cu-BPy-PMO-TMS were almost fitted to the model structure by adopting the crystal structural
parameter of Cu(bpy)Cl2 (Table S1, see Supplementary Material). However, the EXAFS Fourier
transform of Cu-BPy-PMO showed deviation from the fitting curve under the same analytical
conditions, which suggests that the local coordination structure of Cu-BPy-PMO may differ
slightly from that of Cu(bpy)Cl2, possibly due to the formation of an unidentified copper
complex as observed for the UV-vis absorption spectra.
3.2. Mukaiyama epoxidation by Cu-immobilized BPy-PMO catalysts
The catalytic properties of Cu-BPy-PMO and Cu-BPy-PMO-TMS for Mukaiyama epoxidation
with dioxygen were evaluated (Table 2). Cyclohexene and isobutyraldehyde were selected as

substrate and co-reagent, respectively. Under general epoxidation conditions, auto-oxidation
occurs, even in the absence of a catalyst and especially at high substrate concentration. We
confirmed that almost no reaction occurred within 150 min without a catalyst when 1 mmol of
cyclohexene and 3 mmol of isobutyraldehyde were used in chloroform solvent under an oxygen
atmosphere at 40 °C. Thus, the catalysis evaluation was conducted under the same conditions.
1,2-Epoxycyclohexane (1), 2-cyclohexene-1-ol (2), 2-cyclohexene-1-one (3) and
1,2-cyclohexanediol (4) were obtained as reaction products (Scheme 1). Among these products,
the selectivity toward 1 was ca. 90% in all cases; therefore, only the yield of 1 was used for
evaluation of the catalytic activity.
Scheme 1. Reaction scheme of Mukaiyama epoxidation of cyclohexene by catalyzed by
Cu-BPy-PMO and Cu-BPy-PMO-TMS in the presence of O2 and isobutyraldehyde.

Table 2. Mukaiyama epoxidation of cyclohexene catalyzed by Cu-BPy-PMO and
Cu-BPy-PMO-TMS in the presence of O2 and isobutyraldehyde.a
By-product selectivity [%]e
Conversion
[%]b
Selectivity
[%]c
Yield
[%]d
6.5
Catalyst
2
3
4
Cu-BPy-PMO
Cu-BPy-PMO-TMS
1st reaction
7.3
89
2.5
6.4
1.9
17.3
16.3
22.8
89
90
92
15.6
14.7
21.0
2.6
2.1
2.0
8.6
5.4
4.0
0
Cu-BPy-PMO-TMS
2nd reaction
2.9
2.0
Cu-BPy-PMO-TMS
3rd reaction
a
All reactions were conducted with cyclohexene (1 mmol) and isobutyraldehyde (3 mmol) in
c
chloroform (10 mL) at 40 °C under oxygen (1 atm). b Conversion of cyclohexene. Selectivity
d
of 1 was calculated as yield of 1/total yield of reaction products. Yield of 1 based on
e
cyclohexene. Selectivity of by-products was calculated as yield of 2 or 3 or 4/total yield of
reaction products, respectively.

Fig. 4 shows reaction kinetic curves for the epoxidation of cyclohexene to 1. In the
absence of a catalyst, almost no product formation was observed. Likewise, almost no product
was formed when BPy-PMO or BPy-PMO-TMS were used. However, the use of Cu-BPy-PMO
promoted epoxidation without an induction period and the yield of 1 increased continuously
with the reaction time. The product yield after 150 min was 6.5%, which corresponds to a
turnover number (TON) of 9. Cu-BPy-PMO-TMS exhibited a higher reaction rate with a
turnover frequency (TOF) within the initial 60 min (TOF = 14.4 h-1) that was 4.3 times higher
than that of Cu-BPy-PMO. The reaction proceeded continuously with the reaction time and the
product yield reached 15.6% after 150 min, which corresponds to a TON of 30. The higher
catalytic activity of Cu-BPy-PMO-TMS than that with Cu-BPy-PMO is considered to be for
following reasons: (1) TMS treatment of the surface silanol groups contribute to smooth mass
transfer within the mesochannels because oxygenated products such as epoxides tend to interact
with the silanol groups. Such a promotion effect of diffusion by trimethylsilylation has been
reported previously for mesoporous silica-based oxidation catalysts [19-23]. (2) The TMS
treatment also excludes the formation of an undesired Cu complex by reaction between CuCl2
and the silanol groups at the BPy-PMO surfaces, which was suggested by UV-vis absorption

spectra and EXAFS analysis.
15
10
5
0
0
30
60
90
120 150
Reaction time / min
Fig. 4. Time dependent yield of 1 over Cu-BPy-PMO-TMS (closed circles), Cu-BPy-PMO
(open circles), BPy-PMO-TMS (closed triangles), BPy-PMO (open triangles), and without
catalyst (closed squares). All reactions were conducted with cyclohexene (1 mmol) and
isobutyraldehyde (3 mmol) in chloroform (10 mL) at 40 °C under oxygen (1 atm).
Cu-BPy-PMO-TMS exhibited high reusability for Mukaiyama epoxidation.
Cu-BPy-PMO-TMS was recovered by filtration under air and washed with solvent, followed by
drying at 60 °C under reduced pressure. The recycle reactions were conducted under identical
conditions. The recovered Cu-BPy-PMO-TMS exhibited almost no loss of activity and
selectivity for at least two recycle uses (Fig. 5, Table 2).

16
14
12
10
8
6
4
2
0
0
75 150 225 300 375 450
Reaction time / min
Fig. 5. Recycle experiment using Cu-BPy-PMO-TMS. All reactions were conducted with
cyclohexene (1 mmol) and isobutyraldehyde (3 mmol) in chloroform (10 mL) at 40 °C under
oxygen (1 atm) in the presence of Cu-BPy-PMO-TMS or recovered Cu-BPy-PMO-TMS.
Fig.
6
shows N2 adsorption/desorption isotherms for Cu-BPy-PMO and
Cu-BPy-PMO-TMS before and after the reactions. Cu-BPy-PMO showed a large decrease in
the amount of adsorption after the first reaction, while Cu-BPy-PMO-TMS showed a very small
decrease in the amount of adsorption, even after the third reaction. The BET surface areas of
Cu-BPy-PMO and Cu-BPy-PMO-TMS after the first reaction were largely different at 141 and
543 m2 g-1, respectively (Table S2). XRD analysis also showed that the low angle reflections
due to the ordered mesoporous structure were retained for Cu-BPy-PMO-TMS, although they
were largely decreased for Cu-BPy-PMO after the first reaction (Fig. S7, see Supplementary
Material). These results indicate that Cu-BPy-PMO-TMS has higher structural stability than

Cu-BPy-PMO for Mukaiyama epoxidation. The improved structural stability by TMS treatment
would also contribute to the high catalytic activity and high reusability.
(a)400
(b)400
300
300
200
100
0
200
100
0
0
0.2 0.4 0.6 0.8
1
0
0.2 0.4 0.6 0.8
1
Relative Pressure / P/P0
Relative Pressure / P/P0
Fig. 6. Nitrogen adsorption/desorption isotherms for (a) Cu-BPy-PMO and (b)
Cu-BPy-PMO-TMS before (blue line) and after the first (red line), second (green), and third
reaction (purple line). Closed circles, adsorption; open circles, desorption.
Hot filtration experiments were also conducted to evaluate the leaching of copper
species. Fig. 7 shows reaction kinetic curves before and after removal of the PMO catalyst when
the reaction had proceeded for 60 min. No significant change in the product yield was observed
after removal of the PMO catalyst, which indicates that the catalytic reaction occurred only over
Cu-BPy-PMO-TMS.

with PMO catalyst
15
10
5
without PMO catalyst
(after filtration)
0
0
50
100
150
Reaction time / min
Fig. 7. Results of hot filtration experiments. Cu-BPy-PMO-TMS was removed at 60 min from
the start of the reaction. Closed circles, time dependent yield of 1 without removal of the
catalyst; open circles, time dependent yield of 1 after the removal of Cu-BPy-PMO-TMS. All
reactions were conducted with cyclohexene (1 mmol) and isobutyraldehyde (3 mmol) in
chloroform (10 mL) at 40 °C under oxygen (1 atm).
The molecular structure of the Cu complex immobilized on BPy-PMO-TMS after the
reaction was investigated using elemental and XAFS analyses. Fig. 8a shows the Cu/BPysurf and
Cl/Cu molar ratios for Cu-BPy-PMO-TMS before and after reactions measured by EDX
analysis. The Cu/BPysurf ratios were almost constant before and after the reactions
(approximately 0.2), which was consistent with the result that indicated almost no leaching of
Cu species by the hot-filtration experiment. The Cl/Cu ratios were also almost constant

(approximately 1), which suggests that the coordination structure of Cu was not apparently
changed after the reactions. The XAFS Fourier transforms before and after catalysis were
almost identical, which suggests that the coordination structure around Cu was retained after the
reaction (Fig. 8b).
2
(a)
0.2
1
0.1
0
0
(b)
0
2
4
6
R / 10-1 nm
Fig. 8. (a) Cu/BPysurf ratio (black) and Cl/Cu ratio (grey) of Cu-BPy-PMO-TMS before and
after the catalytic reactions. (b) Cu K-edge EXAFS Fourier transform (κ = 20-120 nm-1) of
Cu-BPy-PMO-TMS before (black line) and after the first reaction run (grey line).
3.3. Mukaiyama epoxidation in the presence of radical scavengers
The free radical auto-oxidation reaction occurs with Mukaiyama epoxidation at high
substrate concentrations. Fig. 9a shows the time-dependent yield of 1 without a catalyst at high

concentrations of cyclohexene (7 mmol) and isobutyraldehyde (7 mmol). Epoxide 1 was
produced after an induction period of 50 min and production increased steeply. After the
reaction for 6 h, 1 was obtained in 23% yield with formation of by-products 2, 3, and 4 (Table 3).
The selectivity of 1 relative to reaction products was decreased from 89% to 71% by increasing
substrate concentrations in the absence of catalyst (vs. Table 2). A radical scavenger was added
to the reaction mixture to control the free radical auto-oxidation. When
2,6-di-tert-butyl-p-cresol (BHT) was added into the reaction mixture as a molecular radical
scavenger, the reaction was almost completely stopped and no product was afforded (Fig. 9a).
This result indicates that the radical scavengers quench the radicals generated by auto-oxidation.
During the investigation, auto-oxidation was also found to be quenched when PS-BPy beads
were added into the reaction mixture. This suggests that PS-BPy also acts as a radical scavenger
for Mukaiyama epoxidation.
Fig. 9b shows the time-dependent yield of 1 in the presence of the Cu-BPy-PMO-TMS
catalyst at high concentrations of cyclohexene and isobutyraldehyde. Epoxide 1 was produced
without the induction period and production increased at an almost constant rate. After the
reaction for 6 h, 23% yield of 1 was obtained along with by-products 2, 3, and 4 (Table 3). By
using Cu-BPy-PMO-TMS, the selectivity of 1 was increased to 79% even at high substrate
concentrations. Then, BHT was added into the reaction mixture to control the free radical

auto-oxidation in the presence of the Cu-BPy-PMO-TMS catalyst. However, the reaction was
almost completely stopped, similar to the auto-oxidation reaction without a catalyst. This result
indicates that the molecular scavenger quenched the radicals generated in the solution and also
in the mesopore spaces of Cu-BPy-PMO-TMS because the molecular scavenger can easily
penetrate the mesochannels because of its molecular size (0.9 nm) is smaller than the pore
diameter (3.8 nm).
(a) 25
(b) 25
Cu-BPy-PMO-TMS
No catalyst
20
20
15
10
5
15
10
5
+ PS-BPy
+ BHT
+ PS-BPy
+ BHT
0
0
0
100 200 300 400
Reaction time / min
0
100 200 300 400
Reaction time / min
Fig. 9. Time dependent yield of 1 at high substrate concentration (7 mmol) without a catalyst (a,
closed circles) and with the Cu-BPy-PMO-TMS catalyst (b, closed circles). A polymer radical
scavenger (PS-BPy, closed triangles) or a molecular radical scavenger (BHT, closed squares)
was added. All reactions were conducted with cyclohexene (7 mmol) and isobutyraldehyde (7
mmol) in chloroform (10 mL) at 40 °C under oxygen (1 atm).

However, when the solid scavenger PS-BPy was added into the reaction mixture,
epoxide 1 was produced with a decrease in production rate to approximately 55% of the rate
without a scavenger. After the reaction for 6 h, 1 was obtained in 15% yield along with
by-products 2, 3, and 4 (Table 3). Although the selectivity of 1 was almost comparable to
reaction system only using Cu-BPy-PMO-TMS, we clearly observed improved selectivity
compared with blank reaction, possibly due to the suppression of undesired reactions such as
radical-based allylic oxidation and ring opening reaction. Low product yield and low conversion
of cyclohexene suggest that PS-BPy quenches the radicals in the solution but could not quench
the reaction inside the mesopore spaces because the PS-BPy beads could not penetrate into the
mesochannels. Thus, the combined use of Cu-BPy-PMO-TMS and PS-BPy enables the
radical-based reaction to proceed solely inside the mesochannel spaces. The reaction is limited
and controlled inside the mesochannels, which could be advantageous to achieve
well-controlled radical reactions without thermal runaway, fire, or explosions.

Table 3. Product distributions of Mukaiyama epoxidation of cyclohexene without or with Cu –
BPy-PMO-TMS at high substrate concentration.a
By-product selectivity [%]e
Conversion
[%]b
Selectivity
Yield
[%]d
23
Catalyst
[%]c
70
2
3
4
Blank
32.5
4.0
1.6
10
9.0
15
Cu-BPy-PMO-TMS
Cu-BPy-PMO-TMS
+PS-BPy
29.0
79
23
10.4
19.1
78
15
1.6
11.4
9.0
a
All reactions were conducted with cyclohexene (7 mmol) and isobutyraldehyde (7 mmol) in
b
c
chloroform (10 mL) at 40 °C under oxygen (1 atm). Conversion of cyclohexene. Selectivity
d
of 1 was calculated as yield of 1/total yield of reaction products. Yield of 1 based on
e
cyclohexene. Selectivity of by-products was calculated as yield of 2 or 3 or 4/total yield of
reaction products, respectively.

4. Conclusion
Cu-bipyridine complexes were formed on the pore surfaces of BPy-PMO and
BPy-PMO-TMS, and evaluated for catalysis of Mukaiyama epoxidation. BPy-PMO-TMS
formed a Cu complex which is almost the same coordination structure as Cu(bpy)Cl2, while
BPy-PMO formed a mixture of the desired Cu(bpy)Cl2 complex and an unidentified Cu species,
possibly due to coordination with the surface silanol groups. Cu-BPy-PMO-TMS exhibited
higher catalytic activity and recyclability for the epoxidation of cyclohexene with molecular
oxygen than Cu-BPy-PMO. Auto-oxidation of cyclohexene occurred in the presence and
absence of a catalyst at high substrate concentration. The addition of a solid radical scavenger
could quench auto-oxidation in the solution but could not quench the reaction in the mesopore
spaces of BPy-PMO. This result is very important because the combined use of a mesoporous
catalyst and a solid scavenger enables radical-based reactions to proceed in a controllable
manner and thus extends the scope of organic synthesis chemistry using radicals.

Conflicts of interest
There are no conflicts to declare.
Author information
Present Address
‡S.I.: Kanagawa University, 3-27-1, Rokkakubashi, Kanagawa-ku, Yokohama 221-8686, Japan
Acknowledgements
The authors thank Ms Ayako Ohshima (Toyota Central R&D Laboratories, Inc.) for conducting
the GC-MS analyses. This work was supported by the ACT-C program of the Japan Science and
Technology Agency (JST) (grant number JPMJCR12Y1).
Appendix A. Supplementary data
Supplementary material related to this article can be found, in the online version.
References
[1] A. Studer, D. P. Curran, Angew. Chem. Int. Ed. 55 (2016) 58-102.
[2] M. L. You, M. Y. Liu, S. H. Wu, J. H. Chi, C. M. Shu, J. Therm. Anal. Calorim. 96 (2009)