A highly efficient method of epoxidation of olefins with hydrogen peroxide catalyzed by changeable hexadentate 8-quinolinolato manganese(III) complexes

Article abstract of DOI:10.1016/j.jcat.2008.02.023

Novel hexadentate binding 8-quinolinolato manganese(III) complexes were proposed and conveniently synthesized for the epoxidation of olefins with aqueous hydrogen peroxide in water-acetone media with ammonium acetate and acetic acid as additives. The catalytic efficiencies of the suggested catalysts were found to be obviously superior to the traditional tetradentate salen-Mn^IIICl, due to their special hexadentate binding structures that could be easily converted to the corresponding pentadentate with pendant hydroxyl groups by opening an axial Mn{single bond}O bond in the reaction media, as supported by UV-vis spectra, in situ cyclic voltammetry, and quartz crystal microbalance characterizations.

Full text of DOI:10.1016/j.jcat.2008.02.023

Journal of Catalysis 256 (2008) 154–158
www.elsevier.com/locate/jcat
Research Note
A highly efficient method of epoxidation of olefins with hydrogen peroxide
catalyzed by changeable hexadentate 8-quinolinolato manganese(III)
complexes
Sheng Zhong, Yueming Tan, Zaihui Fu ∗, Qingji Xie ∗, Fang Xie, Xiaoping Zhou, Zhengpei Ye,
Guosheng Peng, Dulin Yin
Key Laboratory of Chemical Biology and Traditional Chinese Medicine Research (Ministry of Education of China), College of Chemistry and
Chemical Engineering, Hunan Normal University, Changsha 410081, China
Received 19 January 2008; revised 24 February 2008; accepted 25 February 2008
Abstract
Novel hexadentate binding 8-quinolinolato manganese(III) complexes were proposed and conveniently synthesized for the epoxidation of
olefins with aqueous hydrogen peroxide in water–acetone media with ammonium acetate and acetic acid as additives. The catalytic efficiencies of
III
the suggested catalysts were found to be obviously superior to the traditional tetradentate salen–Mn Cl, due to their special hexadentate binding
structures that could be easily converted to the corresponding pentadentate with pendant hydroxyl groups by opening an axial Mn–O bond in the
reaction media, as supported by UV–vis spectra, in situ cyclic voltammetry, and quartz crystal microbalance characterizations.
© 2008 Elsevier Inc. All rights reserved.
Keywords: Catalysis; Epoxidation; Hexadentate structure; Hydrogen peroxide; Olefins; 8-Quinolinolato Mn(III) complexes
1. Introduction
We have been intrigued with the idea that 8-quinolinolato
derivatives may be very promising candidates for develop-
ing novel Mn-based catalytic epoxidation with aqueous H2O2,
based on the following considerations: (i) They could sup-
press the catalase feature of Mn ions, (ii) could chelate Mn3+
ions to form hexadentate binding complexes different from
planar tetradentate binding salen Mn ones, and (iii) could
improve the epoxidative selectivity by lowering the Lewis
acidity of Mn ions. Herein we report for the first time the
valuable results obtained from using the hexadentate bind-
ing 8-quinolinolato manganese(III) complexes (Q3MnIII in
Scheme 1) to catalyze the epoxidation of olefins with aqueous
H2O2 in water–acetone media containing ammonium acetate
(NH4OAc) and acetic acid (HOAc), and in situ cyclic voltam-
metry (CV), quartz crystal microbalance (QCM), and UV–vis
spectroscopy techniques are used to identify the structural tran-
sition of Q3MnIII from hexadentate to pentadentate mode in the
catalytic process.
Epoxidation of olefins is one of the most important re-
actions in organic synthesis, because the reaction products—
epoxides—can be easily converted to wide variety of com-
pounds [1]. Consequently, there has been considerable inter-
est in the development of highly efficient epoxidation catalysts
[2–5]. In particular, nontoxic transition metals, such as Mn-
mediated epoxidation with aqueous hydrogen peroxide (H2O2),
have been appreciated as a green process [6]. But the efficient
Mn-mediated epoxidation systems reported to date have used
mainly tetradentate binding manganese complexes as catalysts
[7] and have some disadvantages, including the use of toxic sol-
vent, oxidative decomposition of organic ligands, opening by
reaction of the desired epoxide product, unsatisfactory H2O2
efficiency, and high cost.
*
Corresponding authors.
E-mail address: fzhhnu@tom.com (Z. Fu).
0021-9517/$ – see front matter © 2008 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2008.02.023

S. Zhong et al. / Journal of Catalysis 256 (2008) 154–158
155
2. Experimental
2.1. Preparation and characterizations of Q3MnIII complexes
The general preparation procedure for the Q3MnIII com-
plexes is as follows. First, 5 mL of Mn(OAc)2 aqueous solution
(1 mol L−1) was added dropwise to 50 mL of stirred tetrahy-
drofuran (THF) containing 15 mmol of 8-hydroxyquinoline (8-
HQ) or its derivative, followed by adding 30 wt% aq. H2O2
(0.57 g, 5 mmol) into the solution and adjusting the solution
pH to 6–7 with ammonia (measured using PHS-3C pH meter,
China). Then, the reaction mixture was refluxed for 2 h, and
the resulting precipitate was filtrated and washed with ethanol
for at least three times. After air-drying, a deep-yellow solid
(80–92% yield) was obtained and denoted as 1a to 1e (see
Scheme 1). For the sake of comparison, a hexadentate binding
5-chloro-7-iodo-8-quinolinolato FeIII (Q3FeIII) and traditional
salen–MnIIICl complexes also was prepared.
CHN and metal elemental analyses of these complexes indi-
cate that 1a–1e and Q3Fe complexes are hexadentate, whereas
5-nitryl-8-quinolinolato MnIII (denoted as 1*f) is tetradentate
(see Supporting Information). The diffusion reflection UV–vis
spectra of the solid Q3MnIII complexes were recorded from 200
to 800 nm on an UV-3310 spectrophotometer (BaSO4 as a stan-
dard, Hitachi). The CV and QCM techniques used to charac-
terize these complexes are described in detail in the Supporting
Information.
III
Fig. 1. The diffusion reflection UV–vis spectra of some typical solid Q Mn
complexes.
3
the metal-to-ligand charge transfer (MLCT) reported for other
Mn(III) complexes [8]. In addition, the two very weak bands in
the 530–583 nm region may be attributed to the spin-allowed
d–d transitions of the centered Mn ions, supporting the suppo-
sition that the Q3MnIII complexes have a distorted octahedral
geometry [9].
2.2. Epoxidation procedure
3.2. Catalytic epoxidations
The general procedure for epoxidation is as follows. To
a cooled (10 ◦C) and stirred water–acetone (v/v 1:3, 3 mL)
mixture of olefin (1 mmol), catalyst Q3MnIII (0.02 mmol),
NH4OAc (0.2 mmol), and HOAc (0.1 mmol), 10% H2O2
(1.5 mmol) was added dropwise over a specified period (gen-
erally within a half of the total reaction time length) to improve
the utilization efficiency of H2O2. After H2O2 was consumed,
the catalyst was separated from the reaction mixture by fil-
tration, and the filtrate was analyzed on an Agilent 6890N
gas chromatography (GC) with a HP-5 quartz capillary col-
umn (30 m × 0.32 mm × 0.25 µm) and flame ionization de-
tector (FID) using ultra-pure nitrogen as a carrier gas (rate
1.0 mL/min). Both the injector and detector temperature were
250 ◦C, and the column temperature was between 90 and
The epoxidation of styrene with aqueous H2O2 was used to
check the catalytic performances of Q3MnIII, Q3FeIII, 5-nitryl-
8-quinolinolato MnIIICl (1*f), and salen–MnIIICl complexes,
and the obtained results are listed in Table 1. To our delight,
the hexadentate Q3MnIII complexes (1a–1e), with the help of
NH4OAc–HOAc, exhibited modest to high activity and excel-
lent epoxidative selectivity in water–acetone media, whereas
the tetradentate 1*f and salen–MnIIICl, as well as the hexaden-
tate Q3FeIII, all gave very poor epoxide yields under the same
reaction conditions (entries 6–8). These findings suggest that
the ligand, metal ion, and chelate mode had a significant affect
on the catalytic efficiency of these catalysts. Entries 2–5 fur-
ther show that the substituents of 8-hydroxyquinoline (8-HQ)
had a large influence on the catalytic performances of Q3MnIII
complexes. Compared with 8-HQ-based Q3MnIII (1a), the 5-
and 7-halogenated Q3MnIII catalysts (1b–1e) significantly im-
proved the epoxide yield (80–98%). The activity followed an
increasing sequence of 1e > 1d > 1c > 1b, most likely due to
the increase of electron-donating effect of halogen substituents
from Cl, Br to I on the ligand’s aryl ring and the high resistance
of halogenated 8-HQ ligands against oxidative degradation of
the Q3MnIII, as reported for chlorinated porphyrin–MnIII [10].
The effects of additives were further studied with catalyst
1e as an example (entries 9–12). In the absence of additives,
1e demonstrated relatively low catalytic efficiency (entry 9).
Among the additives examined, the NH4OAc–HOAc mixture
1
170 ◦C. The isolated epoxides were determined by H NMR
and also satisfactorily identified by comparing their MS spectra
with those of the authentic samples.
3. Results and discussion
3.1. Characterization of Q3MnIII complexes
The diffusion reflection UV–vis spectra of the solid Q3MnIII
complexes (1b, 1d, and 1e) were very similar (Fig. 1), consist-
ing of two ban∗ds in the 269–346 nm region that can be assigned
to the π → π transitions of ligands. The complexes also ex-
hibited a strong band at ca. 400–428 nm, probably as a result of

156
S. Zhong et al. / Journal of Catalysis 256 (2008) 154–158
Table 1
III
a
The epoxidation of olefins with aqueous H O catalyzed by Q Mn complexes
2
2
3
Entry
Catalyst
Substrate
Additive
(mmol)
Time
(h)
Conversion
(%)
Selectivity for
epoxide (%)
1
2
3
4
5
1a
1b
1c
1d
1e
2
2
4
2
2
2
2
48
5
2
60
82
85
90
100
30
8
21
27
84
88
25
98
98
98
98
98
95
90
25
96
98
95
96
Styrene
NH OAc (0.2)
4
+ HOAc (0.1)
b
6
7
1*f
c
III
Salen–Mn
III
8
9
10
11
12
Q Fe
3
1e
1e
1e
1e
No additive
NH OAc (0.2)
4
Styrene
HOAc (0.1)
14
2
NH OAc (0.2)
4
+ NH ·H O (0.1)
3
2
d,e
13
14
1e
1e
1e
1e
1e
1e
1e
1e
1-Hexene
1-Octene
β-Pinene
α-Methyl styrene
20
24
2
3
4
4
4
4
62
50
100
98
95
95
83
95
99
98
85
85
85
80
d,e
15
16
f
HOAc (0.1)
d,g
17
+ NH OAc (0.2)
4
d,g
18
Cyclohexene
d,g
19
93
73
20
a
trans-2-hexene-1-ol
III
Unless otherwise specified, the reactions were performed using substrate (1 mmol), catalyst Q Mn (2% molar percentage), 10% H O (1.5 mmol) and
water/acetone (1/3, 3 mL) at 10 C; H O was added dropwise to the solution. Conversion and selectivity for olefins were determined by GC using an internal
3
2 2
◦
2
2
standard technique.
b
Using MnCl as a Mn source.
2
c
d
e
f
III
ꢀ
Salen–Mn Cl was prepared from N,N -bis(salicylidene) ethylethylenediamine and MnCl following reported procedures [2].
2
0.05 mmol NH OAc and 0.05 mmol HOAc were added.
4
1.5 mmol H O (30%) and water/acetone (v/v 1/4, 3 mL) were added.
2
2
1e of 3.5% molar percentage was used.
Entry 17 used the fresh 1e, 18 and 19 employed the recovered 1e once and twice, respectively.
g
(pH 7.13 in the water–acetone medium) gave the highest epox-
ide yield (entry 5), although NH4OAc (pH 8.09 in its water–
acetone solution) was very effective in increasing the epoxida-
tive efficiency (entry 10), and HOAc (pH 5.23 in its water–
acetone solution) also exhibited good results, provided that re-
action time was prolonged (entry 11). However, the weak basic
additives NH4OAc–NH3·H2O (pH 9.24 here) gave very disap-
pointing results (entry 12). These findings indicate that the pH
value of reaction medium plays a key role in the catalytic per-
formance of the Q3MnIII complexes.
Entries 13–19 further illustrate that the Q3MnIII catalyst 1e
also was very efficient for the epoxidation of other unfunction-
alized alkenes. 1-Hexene and 1-octene as nonactivated termi-
nal olefins could give >50% conversion and modest to excel-
lent selectivity using a relatively low terminal oxidant excess
(1.5 equivalents of H2O2; see entries 13 and 14). β-Pinene
was nearly quantitatively oxidized to its epoxide without cy-
clobutane rupture (entry 15). Entry 16, for the epoxidation of
α-methyl styrene, illustrates that even extremely acid-sensitive
epoxide can be formed with 98% selectivity, indicating that the
ligand’s basicity of the Q3MnIII can suppress the opening side
reactions of the desired epoxide by decreasing the Lewis acidity
of Mn3+ ions. The reuse of the catalyst 1e was checked by the
epoxidation of cyclohexene; the catalyst could be filtered from
the reaction media and reused twice without loss of activity
(entries 17–19). Finally, entry 20 illustrates that the trans-2-
hexene-1-ol as a functionalized alkene also could be selectively
oxidized to form the corresponding epoxide with modest yield,
but the hydroxyl of allylic alcohol was partly oxidized to form
the corresponding aldehyde as byproduct.
3.3. Catalytic mechanism
To investigate the catalytic mechanism of the Q3MnIII epox-
idation system, in situ CV and QCM methods were used to
study the Q3MnIII, Q3FeIII, 5-nitryl-8-quinolinolato MnIIICl
(1*f), and salen–MnIIICl complexes, as detailed in the Support-
ing Information. Therein, the main results and conclusions are
summarized as follows:
1. Cyclic voltammography for 1e and its ligand (e, 5-Cl-7-I-
8-HQ) in water–acetone medium showed high quantitative
similarity, even at different potential-scan rates and dif-
ferent solution-pH values. Ligand-redox peaks were also
found for 1b and 1d under identical conditions. These find-
ings suggest that these pristine hexadentate Q3MnIII cata-
lysts with Mn3+(d4) ions could cleave at the axial MnIII–O
coordination site in the water–acetone medium to give pen-
tadentate structures with pendant electroactive hydroxyl
groups, as depicted in Scheme 1.

S. Zhong et al. / Journal of Catalysis 256 (2008) 154–158
157
−5
Fig. 2. UV–vis spectra of (A) an acetone solution of catalyst 1e (6 × 10 M,
−5
4 mL); (B) a water–acetone (v/v 1/3, 4 mL) solution of 1e (6 × 10 M) and
NH OAc–HOAc; (C) adding 3 µL 30% H O to the solution B; (D) adding
4
2 2
0.2 mmol styrene to the solution C. Inset shows visible absorption spectral
changes observed upon adding 1 µL 30% H O to a acetone–water (v/v 3/1)
2
2
solution of catalyst 1e and NH OAc–HOAc (B), spectral scanning was taken at
4
average 3 min intervals (C –C ).
Scheme 1. Proposed catalytic cycle.
1
7
with λmax ∼ 412 nm is assigned to the metal to ligand charge
transfer (MLCT) [8]. When 1e was dissolved in water–acetone
containing NH4OAc–HOAc, this MLCT absorbance decreased
significantly and exhibit some blue shifting (λmax ∼ 402 nm,
see B in Fig. 1), possibly due to the cleavage of axial Mn–O
bond of 1e to form intermediate 2 (see Scheme 1). After aque-
ous H2O2 was added to this solution, repeated scanning re-
vealed a successive change in the MLCT absorbance (inset
in Fig. 2); specifically, the MLCT band immediately became
shifted to λmax ∼ 424 nm and its intensity increased signifi-
cantly after the addition of H2O2 (B vs C1 in the inset), indi-
cating that H2O2 interacted with 2 to form a new intermediate.
Subsequently, the MLCT absorbance gradually decreased with
time (C1 to C7), implying the further change in such a new
intermediate. After several hours, a new absorption band cen-
tered around 500 nm was developed, the maximum of which
was obscured by the tailing of strongly absorbing species at
λmax > 500 nm (C in Fig. 2), as reported by Kochi et al. us-
ing salen–Mn mediated epoxidations with iodosylbenzene [11].
This spectral change should be assigned to the formation of
Q3MnV=O (5 in Scheme 1) through a heterolytic cleavage of
O–O bond. After the addition of the substrate (styrene), it gave
a spectrum D that illustrates that the Q3MnV=O species is in-
volved in oxygen atom transfer stage. But if the salen–MnIIICl
was treated with aqueous H2O2 under the aforementioned ex-
perimental conditions, its MLCT absorbance continuously de-
creased, and a slight blue shift of the salen’s π → π∗ band
(λmax ∼ 337 nm) occurred simultaneously with time (Fig. 3).
This spectral change likely implies that the salen–MnIIICl in-
duces homolytic cleavage of the O–O bond of H2O2 to generate
free radicals in solution, which leads to spurious reactions that
can degrade the catalyst [7], as supported by the foregoing cat-
alytic experiment (see entry 7).
2. Obvious Mn-redox peaks for the effective hexadentate
Q3MnIII catalysts were not seen in the water–acetone
medium, but their Mn-redox peaks were evident in the
N,N-dimethylformamide (DMF) medium, likely due to
the limited solubility values, the enhanced steric hindrance
of the ligands (hexadentate) to the interfacial electron
transfer of Mn, and the possible masking of the ligand-
redox peaks against Mn-redox peaks in the water–acetone
medium.
3. In contrast, the Mn-redox peaks of 1*f became obvious in
both water–acetone and DMF media, suggesting that this
complex has a different coordination structure, support-
ing the elemental analysis results demonstrating that this
complex is tetradentate. In addition, the tetradentate salen–
MnIIICl gave its obvious Mn-redox peaks in water–acetone
medium.
4. No redox waves of ligand and Fe ions were recorded for
Q3FeIII, supporting that this complex with Fe3+(d5) ions
owns a stable hexadentate structure (elemental analysis),
and it could not be converted to the pentadentate structure
with pendant electroactive hydroxyl groups.
5. The QCM experiments showed that the dissolution of the
1e coated on a QCM Au electrode became faster soon after
the HOAc was added to an aqueous NH4OAc, but such a
phenomenon was not observed for the Q3FeIII under the
same experimental conditions, also demonstrating in the
coexistence of HOAc and NH4OAc that the hexadentate
1e was able to open to give a more soluble pentadentate
structure, but the Q3FeIII with a stable hexadentate struc-
ture could not.
The 1e epoxidation system was further examined by UV–
vis spectroscopy; the results are illustrated in Fig. 2, in which
A is the spectrum of 1e in acetone where an absorption band

158
S. Zhong et al. / Journal of Catalysis 256 (2008) 154–158
dation of olefins with aqueous H2O2, and have introduced in
situ CV and QCM methods for studying the catalytic mecha-
nisms of the Q3MnIII complexes. This system has the follow-
ing advantages: (i) utilization of more cost-effective and green
water–acetone solvent, (ii) facile operation, and (iii) very high
catalytic efficiency and stability. The system also may be ex-
ploited in the future for industrial applications and asymmetric
epoxidation. Moreover, the CV and QCM tests provide much
useful information on catalytic mechanism, and they can be
expected to find wider applications in catalytic science and
technology, including electrochemistry-assisted redox cataly-
sis.
Acknowledgments
This work was supported by the National Natural Sci-
ence Foundation of China (20573035, 20675029, 90713018,
20335020) and the Natural Science Foundation of Hunan
Province (05JJ40022).
III
−4
Fig. 3. UV–vis spectra of (E) a solution of salen–Mn Cl (1 × 10 M) in
acetone–water (v/v 3/1, 4 mL) containing NH OAc–HOAc; (F) adding 2 µL
30% H O to the solution E. Inset shows the successive changes of UV–vis
4
2
2
absorption spectra observed upon adding 2 µL 30% H O to the solution E,
2
2
spectral scanning was taken at average 5 min intervals (F –F ).
1
12
Supplementary material
Based on our findings, a mechanism for Q3MnIII system
is proposed (Scheme 1). Theoretically, because of the Jahn–
Teller effect of Mn3+ ions with d4 [12], it is anticipated that the
Q3MnIII (1) has a distorted octahedral structure with an axial
Mn–O bond longer than its equatorial Mn–O bond. The axial
Mn–O bond may be more easily cleaved to form a more sol-
uble pentadentate structure with pendant hydroxyl groups (2)
in the presence of NH4OAc–HOAc due to competitive O–H
bonding. Then the 2, which has a basic structure similar to
that of porphyrin– or salen–MnIII with imidazole as its axial
ligand [13,14], may activate H2O2 to form an intermediate 3.
The 3 eventually may be converted to form an intermediate
Q3MnV=O (5) after undergoing a pathway of forming an in-
termediate 4 with the help of OAc− ions, as reported for the
porphyrin–MnIII catalyzed epoxidation of alkenes [15]. No-
tably, the pendant O–H group near the Mn3+ ions in inter-
mediate 2 likely acts as a highly active and competitive co-
ordination group to prevent the catalyst from forming the μ-
oxomanganese dimers [7]. On the other hand, it easily forms in-
tramolecular hydrogen bond with MnIII–O–OAc group in inter-
mediate 4 (Scheme 1), which can activate the O–O bond toward
heterolytic cleavage to form MnV=O active species. These two
effects can explain why the Q3MnIII has much better stability
and higher efficiency of H2O2 utilization compared with salen–
MnIIICl in water–acetone medium containing NH4OAc–HOAc.
The online version of this article contains additional supple-
mentary material (preparation of Q3FeIII complex; elemental
analyses; CV and QCM tests of the catalysts).
Please visit DOI: 10.1016/j.jcat.2008.02.023.
References
[1] S.P. de Visser, S. Shaik, J. Am. Chem. Soc. 125 (2003) 7413.
[2] W. Zhang, J.L. Loebach, S.R. Wilson, E.N. Jacobsen, J. Am. Chem.
Soc. 112 (1990) 2801.
[3] J. Rudolph, K.L. Reddy, J.P. Chiang, K.B. Sharpless, J. Am. Chem.
Soc. 119 (1997) 6189.
[4] K. Matsumoto, Y. Sawada, B. Saito, K. Sakai, T. Katsuki, Angew. Chem.
Int. Ed. 44 (2005) 4935.
[5] M. Colladon, A. Scarso, P. Sgarbossa, R.A. Michelin, G. Strukul, J. Am.
Chem. Soc. 128 (2006) 14006.
[6] B.S. Lane, K. Burgess, J. Am. Chem. Soc. 123 (2001) 2933.
[7] B.S. Lane, K. Burgess, Chem. Rev. 103 (2003) 2457.
[8] S.M. Crawford, Spectrochim. Acta 19 (1963) 255.
[9] S. Biswas, K. Mitra, S.K. Chattopadhyay, B. Adhikary, C.R. Lucas, Tran-
sition Metal Chem. 30 (2005) 393.
[10] J.P. Renaud, P. Battioni, J.F. Bartoli, D. Mansuy, J. Chem. Soc. Chem.
Commun. (1985) 888.
[11] K. Srinivasan, P. Michaud, J.K. Kochi, J. Am. Chem. Soc. 108 (1986)
2309.
[12] T.J. Hubin, J.M. McCormick, N.W. Alcock, D.H. Busch, Inorg. Chem. 40
(2001) 435.
[13] P. Battioni, J.P. Renaud, J.F. Bartoli, M. Reina-Artiles, M. Fort, D. Man-
suy, J. Am. Chem. Soc. 110 (1988) 8462.
4. Conclusion
[14] T. Schwenkreis, A. Berkessel, Tetrahedron Lett. 34 (1993) 4785.
[15] P.L. Anelli, L. Banfi, F. Legramandi, F. Montanari, G. Pozzi, S. Quici,
J. Chem. Soc. Perkin Trans. 1 (1993) 1345.
In summary, for the first time we have developed hexaden-
tate Q3MnIII complexes as very efficient catalysts for epoxi-