cient alkenes investigated appeared unreactive in our condi-
tions (Table 1, entries 4, 5, and 9). In contrast, 1-octene ap-
peared rather reactive. Indeed, when using 0.004 equiv of cata-
lyst and 0.25 equiv of CAN, 1-octene was efficiently epoxidized
and converted into the corresponding diol in situ upon ring-
opening with water. The CAN-catalyzed nucleophilic attack of
water on the epoxide was further demonstrated as stirring 1-
octene with CAN (0.25 equiv) at RT in a 1:1 water/acetonitrile
mixture afforded the corresponding diol in 61% isolated yield.
In addition, a control reaction using H218O afforded the bis 18O-
labeled compound as the only diol, again pertaining the fact
that the inserted oxygen atoms come from water molecules
activated in the coordination sphere of the manganese ion.
Other substrates, including benzyl alcohol and thioanisole,
were also tested. In these cases, control reactions revealed that
benzyl alcohol was equally oxidized to the corresponding alde-
hyde, regardless of whether the catalyst was present or not.
Indeed, after 24 h at room temperature, benzaldehyde was
present in 85% GC yield in the absence of catalyst and 76%
GC yield in the presence of catalyst (Table 1, entry 7). A similar
observation was made upon catalytic oxidation of thioanisole
in the presence of 0.004 equiv of catalyst: the sulfoxide was
evidenced in 68% GC yield, and 91% GC yield if 0.002 equiv of
epoxide was used. In the absence of catalyst, the expected
sulfoxide was isolated in 87% yield, definitely proving that the
catalyst-free CAN-mediated oxidation proceeds efficiently.
We have described herein the chemical activation of a mono-
nuclear manganese(III)-hydroxido complex with CeIV as oxidant.
The resulting highly oxidized manganese species is capable to
perform the oxygen atom transfer reaction to electron-rich al-
kenes. Experiments with isotopically labeled water clearly sup-
port that the oxygen atom originates from the water molecule.
These results stand as a proof-of-concept for the use of water
as oxygen source to perform the oxidation of organic sub-
strates. Work on fine-tuning the electrochemical properties of
manganese mononuclear complexes for the photochemical ac-
tivation of a bound water molecule with the defining aim to
exclude chemical oxidants is in progress.
Table 1. Selected [tBuLMnIIIOH]+-catalyzed oxidation of various sub-
strates.
Entry Substrate
Product[a]
Yield [%]
TON
1
2
3
4
5
6
7
8
9
cyclooctene
cyclooctene
cyclooctene
styrene
a-pinene
1-octene
benzyl alcohol
thioanisole
but-3-enoic acid 2-(oxiran-2-yl)acetic acid
9-oxabicyclo[6.1.0]nonane 56
9-oxabicyclo[6.1.0]nonane 56[b]
9-oxabicyclo[6.1.0]nonane 56 (48)[c]
15
137
15
0
1
3
n.d.
n.d.
0
2-phenyloxirane
a-pinene epoxide
octane-1,2-diol
0
4
10
benzaldehyde
76[b] (85)[d]
68 (87)[d]
0
(methylsulfinyl)benzene
[a] Conditions: substrate 0.4 mmol/1 mL acetonitrile; catalyst 0.004 equiv,
CAN (0.25 equiv in 1 mL H2O, RT, 0.5 h). [b] Conditions: substrate
0.4 mmol/1 mL acetonitrile; CAN (2 equiv in 1 mL H2O, RT, 24 h). [c] Isolat-
ed yield, 1.5 h. [d] Catalyst-free reaction. n.d.: not determined.
DFT calculations were undertaken to gain insight into the elec-
tronic description of the doubly oxidized species. In this case,
deprotonation of the axial hydroxyl group is more likely and
our results point to the formation of a metalloradical species
(see Supporting Information). We investigated the catalytic
performance of [tBuLMnIIIOH]+ by using conditions similar to
those used by Nam et al. previously (catalyst: 0.004 equiv; CAN
0.25 equiv). Interestingly, the catalytic run with cyclooctene
showed after 0.5 h at room temperature the presence of the
corresponding epoxide in 56% GC yield as the sole product
(Table 1, entry 1). However, when the reaction was left for
a longer time, the yield decreased (50% after 1 h and 30%
after 2 h), probably due to the ring opening of the epoxide in
the presence of excess water and a Lewis acid. Upon increas-
ing the amount of oxidant up to 2 equiv, the analytical yield
remained stable, but the TON increased by an order of magni-
tude, as the oxidant was not the limiting reagent anymore
(entry 2). Unfortunately, using these conditions, the selectivity
of the reaction decreased and the amount of by-products in-
creased correlatively. Interestingly, increasing the catalytic
charge did not appear beneficial. Indeed, when using 8 times
more catalyst (0.02 equiv) the amount of epoxide remained in
the 40% range, while the amount of byproducts increased
drastically. Notably, when [tBuLMnIIICl]+ was used a milder ep-
oxidation proceeded and the epoxide was detected in 46%
yield after 1 h and 56% (48% isolated yield) after 1.5 h (see
Table S4). We also noted that the ability of the [tBuLMnIIICl]+
complex for the catalytic oxidation is similar to that of the
[tBuLMnIIIOH]+ complex. Such a chemical reactivity can be ex-
plained by the rapid exchange of the axial chloro ligand with
a hydroxo group. This is supported by both the electrochemi-
cal behavior and the mass spectrum of the [tBuLMnIIICl]+ com-
plex in the presence of water (see Figures S1B, S2).
Experimental Section
Syntheses: The ligand tBuLH and complex [tBuLMnOH](ClO4)
were synthesized according to
a
published procedure.[9]
2[tBuLMnC](MnCl4): 83.1 mg (0.420 mmol) of MnCl2·4H2O dissolved
in 2 mL of ethanol was added under argon to one equivalent of
tBuLH dissolved in 8 mL of ethanol in the presence of one equiva-
lent of NEt3. The brown–red solution was stirred for 6 h at room
temperature. The solution was concentrated by evaporation. A
brown powder was precipitated by addition of diethyl ether, col-
lected by filtration, and dried in vacuum (yield 75%). Elemental
analysis for 2[tBuLMnCl](MnCl4): calcd (%): C 53.88, H 5.77, N 8.67;
Remarkably, when 18O-labeled water (H218O) was used in
combination with acetonitrile as solvent the only product of
the reaction was the isotopically labeled cyclooctene epoxide
(see Figures S9–S10), bringing conclusive evidence that the
oxygen atom originates from a water molecule being activated
through a two-electron, two-proton process. More challenging
substrates were also tested. Unfortunately, the electron-defi-
found (%):
[tBuLMnCl]+.
C 53.73, H 5.85, N 8.80. ESI-MS: m/z=574.22
CAN-mediated oxidation: In a 10 mL round-bottom flask, 0.001 g
(0.0016 mmol) of catalyst, 0.4 mmol of substrate (alkene alcohol or
thiol), and 1 mL of acetonitrile were introduced. An aqueous solu-
tion of CAN [0.055 g (0.1 mmol) in 1 mL] was added dropwise and
the reaction was allowed to stir for the appropriate amount of
ChemSusChem 2012, 5, 2147 – 2150
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2149