Transient intermediates from Mn(salen) with sterically hindered mesityl groups: Interconversion between MnIV-phenolate and Mn III-phenoxyl radicals as an origin for unique reactivity

Article abstract of DOI:10.1021/ic702061y

In order to reveal structure-reactivity relationships for the high catalytic activity of the epoxidation catalyst Mn(salen), transient intermediates are investigated. Steric hindrance incorporated to the salen ligand enables highly selective generation of three related intermediates, O=Mn^IV(salen), HO-Mn^IV(salen), and H₂O-Mn ^III(salen^+?), each of which is thoroughly characterized using various spectroscopic techniques including UV-vis, electron paramagnetic resonance, resonance Raman, electrospray ionization mass spectrometry, ²H NMR, and X-ray absorption spectroscopy. These intermediates are all one-electron oxidized from the starting Mn ^III(salen) precursor but differ only in the degree of protonation. However, structural and electronic features are strikingly different: The Mn-O bond length of HO-Mn^IV(salen) (1.83 A) is considerably longer than that of O=Mn^IV(salen) (1.58 A); the electronic configuration of H₂O-Mn^III(salen^+?) is Mn^III-phenoxyl radical, while those of O=Mn^IV(salen) and HO-Mn^IV(salen) are Mn^IV-phenolate. Among O=Mn ^IV(salen), HO-Mn^IV(salen), and H₂O-Mn ^III(salen^+?), only the O=Mn^IV(salen) can transfer oxygen to phosphine and sulfide substrates, as well as abstract hydrogen from weak C-H bonds, although the oxidizing power is not enough to epoxidize olefins. The high activity of Mn(salen) is a direct consequence of the favored formation of the reactive O=Mn^IV(salen) state.

Full text of DOI:10.1021/ic702061y

Inorg. Chem. 2008, 47, 1674-1686
Transient Intermediates from Mn(salen) with Sterically Hindered Mesityl
Groups: Interconversion between MnIV-Phenolate and MnIII-Phenoxyl
Radicals as an Origin for Unique Reactivity
Takuya Kurahashi,† Akihiro Kikuchi,‡ Takehiko Tosha,† Yoshitsugu Shiro,‡
Teizo Kitagawa,† and Hiroshi Fujii*,†
Institute for Molecular Science & Okazaki Institute for IntegratiVe Bioscience, National Institutes
of Natural Sciences, Myodaiji, Okazaki, Aichi 444-8787, Japan, and RIKEN SPring-8 Center,
Harima Institute, 1-1-1, Kouto, Sayo, Hyogo 679-5148, Japan
Received October 18, 2007
In order to reveal structure–reactivity relationships for the high catalytic activity of the epoxidation catalyst Mn(salen),
transient intermediates are investigated. Steric hindrance incorporated to the salen ligand enables highly selective
generation of three related intermediates, OdMnIV(salen), HO-MnIV(salen), and H2O-MnIII(salen+•), each of which
is thoroughly characterized using various spectroscopic techniques including UV–vis, electron paramagnetic resonance,
resonance Raman, electrospray ionization mass spectrometry, 2H NMR, and X-ray absorption spectroscopy. These
intermediates are all one-electron oxidized from the starting MnIII(salen) precursor but differ only in the degree of
protonation. However, structural and electronic features are strikingly different: The Mn-O bond length of
HO-MnIV(salen) (1.83 Å) is considerably longer than that of OdMnIV(salen) (1.58 Å); the electronic configuration
of H2O-MnIII(salen+•) is MnIII-phenoxyl radical, while those of OdMnIV(salen) and HO-MnIV(salen) are MnIV-
phenolate. Among OdMnIV(salen), HO-MnIV(salen), and H2O-MnIII(salen+•), only the OdMnIV(salen) can transfer
oxygen to phosphine and sulfide substrates, as well as abstract hydrogen from weak C-H bonds, although the
oxidizing power is not enough to epoxidize olefins. The high activity of Mn(salen) is a direct consequence of the
favored formation of the reactive OdMnIV(salen) state.
Introduction
are known to be poor catalysts. The striking difference
between these popular ligands is worth investigating in detail
to establish structure–reactivity relationships.
In order to consider what gives rise to the reactivity
difference, electronic structures of transient intermediates are
of prime importance. It has been well established that, upon
reaction with oxidants at low temperature, Fe(porphyrin)
complexes generate OdFeIV(porphyrin) and OdFeIV(por-
phyrin+•), which are capable of oxidizing organic sub-
strates.4,5 Mn(porphyrin) complexes also generate
The reactivity of transition metal ions is considerably
dependent on the coordination environment, and thus the
design of a ligand system is critically important to achieve
desirable reactivity, even by use of the same active metal
ion. The most remarkable examples are catalytic oxygenation
reactions by porphyrin and salen complexes. In the case of
porphyrins, both Mn(porphyrin) and Fe(porphyrin) com-
plexes show high catalytic activity.1 In contrast, in the case
of salens, Mn(salen) complexes are widely utilized as highly
reactive epoxidation catalysts,2,3 while Fe(salen) complexes
(4) (a) Groves, J. T.; Haushalter, R. C.; Nakamura, M.; Nemo, T. E.;
Evans, B. J. J. Am. Chem. Soc. 1981, 103, 2884–2886. (b) Balch,
A. L.; Latos-Grazynski, L.; Renner, M. W. J. Am. Chem. Soc. 1985,
107, 2983–2985. (c) Sugimoto, H.; Tung, H.-C.; Sawyer, D. T. J. Am.
Chem. Soc. 1988, 110, 2465–2470. (d) Fujii, H. J. Am. Chem. Soc.
1993, 115, 4641–4648.
* To whom correspondence should be addressed. E-mail: hiro@ims.ac.jp.
†
National Institutes of Natural Sciences.
Harima Institute.
‡
(1) Meunier, B. Chem. ReV. 1992, 92, 1411–1456.
(2) Srinivasan, K.; Michaud, P.; Kochi, J. K. J. Am. Chem. Soc. 1986,
108, 2309–2320.
(5) (a) Chin, D.-H.; Balch, A. L.; La Mar, G. N. J. Am. Chem. Soc. 1980,
102, 1446–1448. (b) Lee, W. A.; Calderwood, T. S.; Bruice, T. C.
Proc. Natl. Acad. Sci. U.S.A. 1985, 82, 4301–4305. (c) Swistak, C.;
Mu, X. H.; Kadish, K. M. Inorg. Chem. 1987, 26, 4360–4366. (d)
Groves, J. T.; Gross, Z.; Stern, M. K. Inorg. Chem. 1994, 33, 5065–
5072.
(3) (a) Zhang, W.; Loebach, J. L.; Wilson, S. R.; Jacobsen, E. N. J. Am.
Chem. Soc. 1990, 112, 2801–2803. (b) Irie, R.; Noda, K.; Ito, Y.;
Katsuki, T. Tetrahedron Lett. 1991, 32, 1055–1058. (c) McGarrigle,
E. M.; Gilheany, D. G. Chem. ReV. 2005, 105, 1563–1602.
1674 Inorganic Chemistry, Vol. 47, No. 5, 2008
10.1021/ic702061y CCC: $40.75
2008 American Chemical Society
Published on Web 02/01/2008

Transient Intermediates from Mn(salen)
OdMnIV(porphyrin) and O≡MnV(porphyrin), which show
oxidation activity.6 On the other hand, we recently reported
that an Fe(salen) complex does not generate high-valent
Fe-oxo species but gives salen phenoxyl radical species,
which show no monooxygenation activity.7
Transient intermediates from Mn(salen) have also been
actively investigated,8 especially in the context of reactive
intermediates for Jacobsen-Katsuki enantioselective epoxida-
tion.3 By analogy to well-established nonheme high-valent
Mn-oxo complexes,9 OdMnIV(salen) and O≡MnV(salen) have
been frequently proposed, but the exact identities of transient
intermediates still remain elusive. In previous studies, reactions
of Mn(salen) with NaOCl, PhIO, or m-CPBA (m-chloroper-
benzoic acid) at low temperatures typically produced mixtures
of transient intermediates, which were tentatively assigned as
O≡MnV(salen), MnIV-O-MnIV salen dimer, and MnIV(salen)
with Od, HO-, Cl-, m-CB-(m-chlorobenzoate), and ClO-
ligands mainly on the basis of electrospray ionization mass
spectrometry (ESI-MS) and electron paramagnetic resonance
(EPR) studies. Because more than two species were present in
the solution, these assignments were not necessarily reliable,
and thus structure–reactivity relationships could not be estab-
lished. More importantly, our previous study on Fe(salen)7
suggests that reactions of Mn(salen) with oxidants might also
generate salen phenoxyl radical species, as well as high-valent
MnV(salen) or MnIV(salen), due to comparable oxidation
potentials of the central Mn versus the salen ligand. Indeed,
many examples have been reported for metal-phenoxyl radical
complexes,10 including Mn-phenoxyl radical complexes.11
Transient intermediates from Mn(salen) are also quite
interesting as a model for the oxygen-evolving center (OEC)
in photosystem II. According to the most recent X-ray crystal
–
Figure 1. Sterically hindered Mn(salen) 1 and 2. 1 has ClO4 as a
counteranion. High-valent MnIV-phenolate intermediates OdMnIV(salen),
HO-MnIV(salen), Cl-MnIV(salen), and MnIII-phenoxyl radical intermediate
H2O-MnIII(salen+•) were generated from 1 and/or 2.
structure,12 photosynthetic oxidation of water to oxygen
proceeds at the mononuclear Mn center in the OEC. Highly
effective water oxidation is believed to come from the
efficient electron transfer from the OEC to the light-induced
P680+, which is mediated by a redox-active tyrosyl radical
(TyrZ ). Mn(salen) has a monomeric Mn center with a
vacant coordination site for the H2O binding, and a coordi-
nated phenolate as a tyrosine model. Thus, one-electron
oxidation of Mn(salen) provides an opportunity to examine
the role of electron transfer in the water oxidation in the
OEC.
Herein, we prepared sterically hindered Mn(salen) 1 and
2 bearing a protected Mn center (Figure 1), in order to
prevent the formation of dimeric species. One-electron
oxidation of 1 and 2 generates a MnIII-phenoxyl radical,
H2O-MnIII(salen+•), and MnIV-phenolate, Cl-MnIV(salen),
respectively, which provide a solid spectroscopic basis to
distinguish these two electronic configurations. Low-tem-
perature oxidation of sterically hindered 1 enables highly
selective generation of possible reactive intermediates,
OdMnIV(salen) and HO-MnIV(salen), each of which was
characterized unambiguously with various spectroscopic
methods, including UV–vis, EPR, resonance Raman, ESI-
13
•
(6) (a) Czernuszewicz, R. S.; Su, Y. O.; Stern, M. K.; Macor, K. A.; Kim,
D.; Groves, J. T.; Spiro, T. G. J. Am. Chem. Soc. 1988, 110, 4158–
4165. (b) Groves, J. T.; Stern, M. K. J. Am. Chem. Soc. 1988, 110,
8628–8638. (c) Ayougou, K.; Bill, E.; Charnock, J. M.; Garner, C. D.;
Mandon, D.; Trautwein, A. X.; Weiss, R.; Winkler, H. Angew. Chem.,
Int. Ed. Engl. 1995, 34, 343–346. (d) Groves, J. T.; Lee, J.; Marla,
S. S. J. Am. Chem. Soc. 1997, 119, 6269–6273. (e) Song, W. J.; Seo,
M. S.; George, S. D.; Ohta, T.; Song, R.; Kang, M.-J.; Tosha, T.;
Kitagawa, T.; Solomon, E. I.; Nam, W. J. Am. Chem. Soc. 2007, 129,
1268–1277. (f) Jin, N.; Ibrahim, M.; Spiro, T. G.; Groves, J. T. J. Am.
Chem. Soc. 2007, 129, 12416–12417.
(7) Kurahashi, T.; Kobayashi, Y.; Nagatomo, S.; Tosha, T.; Kitagawa,
T.; Fujii, H. Inorg. Chem. 2005, 44, 8156–8166.
(8) (a) Bryliakov, K. P.; Babushkin, D. E.; Talsi, E. P. J. Mol. Catal. A:
Chem. 2000, 158, 19–35. (b) Feichtinger, D.; Plattner, D. A.
Chem.—Eur. J. 2001, 7, 591–599. (c) Adam, W.; Mock-Knoblauch,
C.; Saha-Möller, C. R.; Herderich, M. J. Am. Chem. Soc. 2000, 122,
9685–9691. (d) Campbell, K. A.; Lashley, M. R.; Wyatt, J. K.; Nantz,
M. H.; Britt, R. D. J. Am. Chem. Soc. 2001, 123, 5710–5719. (e) Feth,
M. P.; Bolm, C.; Hildebrand, J. P.; Köhler, M.; Beckmann, O.; Bauer,
M.; Ramamonjisoa, R.; Bertagnolli, H. Chem.—Eur. J. 2003, 9, 1348–
1359.
2
MS, H NMR, and X-ray absorption spectroscopy (XAS).
Interconversion and energetics among OdMnIV(salen),
HO-MnIV(salen), and H2O-MnIII(salen+•) explain why
Mn(salen) is a far better catalyst than Fe(salen).
(9) (a) Collins, T. J.; Gordon-Wylie, S. W. J. Am. Chem. Soc. 1989, 111,
4511–4513. (b) Collins, T. J.; Powell, R. D.; Slebodnick, C.; Uffelman,
E. S. J. Am. Chem. Soc. 1990, 112, 899–901. (c) MacDonnel, F. M.;
Fackler, N. L. P.; Stern, C.; O’Halloran, T. V. J. Am. Chem. Soc.
1994, 116, 7431–7432. (d) Miller, C. G.; Gordon-Wylie, S. W.;
Horwitz, C. P.; Strazisar, S. A.; Peraino, D. K.; Clark, G. R.;
Weintraub, S. T.; Collins, T. J. J. Am. Chem. Soc. 1998, 120, 11540–
11541. (e) Mandimutsira, B. S.; Ramdhanie, B.; Todd, R. C.; Wang,
H.; Zareba, A. A.; Czernuszewicz, R. S.; Goldberg, D. P. J. Am. Chem.
Soc. 2002, 124, 15170–15171. (f) Parsell, T. H.; Behan, R. K.; Green,
M. T.; Hendrich, M. P.; Borovik, A. S. J. Am. Chem. Soc. 2006, 128,
8728–8729.
(11) (a) Müller, J.; Kikuchi, A.; Bill, E.; Weyhermüller, T.; Hildebrandt,
P.; Ould-Moussa, L.; Wieghardt, K. Inorg. Chim. Acta 2000, 297, 265–
277. (b) Mukherjee, S.; Weyhermüller, T.; Bothe, E.; Wieghardt, K.;
Chaudhuri, P. Dalton Trans. 2004, 3842–3853.
(12) Loll, B.; Kern, J.; Saenger, W.; Zouni, A.; Biesiadka, J. Nature
(London) 2005, 438, 1040–1044.
(10) Chaudhuri, P.; Wieghardt, K. Prog. Inorg. Chem. 2001, 50, 151–216.
(13) McEvoy, J. P.; Brudvig, G. W. Chem. ReV. 2006, 106, 4455–4483.
Inorganic Chemistry, Vol. 47, No. 5, 2008 1675

Kurahashi et al.
Table 1. Electrochemical Parameters (V) of 1 and 2a
1
2
1
2
1
2
compd
Ea
Ea
Ec
Ec
E1/2
E1/2
1
2
0.91
0.90
1.12
1.19
0.81
0.62
0.99
1.07
0.86
0.76
1.06
1.13
a
See the caption in Figure 2 for the measurement conditions.
UV–vis spectral changes with clear isosbestic points, as
shown in Figure 3a. The UV–vis spectrum of [1]+ features
a broad absorption centered at 905 nm, along with other
absorption maxima at 338, 420, and 550 nm. Electrochemi-
cal reduction of [1]+ at -0.40 V at 203 K regenerates the
starting MnIII complex 1. [1]+ is also generated by chemi-
cal oxidation of 1 using 1.0 equiv of a one-electron
chemical oxidant, tris(2,4-dibromophenyl)aminium hexachlo-
roantimonate (3). The broad absorption around 905 nm in
[1]+ is closely similar to that for the phenoxyl radicals
from Fe(salen),7 and absorptions at 338 and 420 could be
assigned as π-π* transitions of the coordinated phenoxyl
radical,10 suggesting that [1]+ is also a phenoxyl radical
intermediate. Further controlled-potential oxidation of [1]+
at a voltage (1.3 V) higher than Ea2 at 203 K generates
another intermediate, which shows a characteristic absorp-
tion of doubled intensity at the wavelength shifted to a
lower energy (980 nm) compared with that of [1]+. This
intermediate may correspond to [1]2+, but due to a high
oxidation potential, [1]2+ is too unstable to be investigated
by other spectroscopic methods.
Resonance Raman spectra of 1 and [1]+ were measured
at the excitation wavelength of 351.4 nm, which may come
from the π-π* transition of the phenoxyl radical. Consis-
tently, new resonance Raman bands appear at 1485 and 1604
cm-1 for [1]+, while one of two C-O stretching bands at
1273 or 1299 cm-1 in 1 disappears and the other shifts to
1283 cm-1 (Figure 4). The bands at 1485 and 1604 cm-1
for [1]+ are assigned as Y7a′ and Y8a′ modes, which consist
predominantly of C-O stretching and Cortho-Cmeta stretching,
respectively.10,16 The greater intensity of Y8a′ compared to
Y7a′ is consistent with formation of the coordinated phenoxyl
radical.10 1-d4, which is deuterated at positions 4, 6, 4′, and
6′ of the phenolate rings (for numbering, see Figure 1), shows
2H NMR signals at -1.4 and -58.1 ppm for 6/6′D and 4/4′D
of the phenolate moieties, respectively (Figure S2, Supporting
Information).8a,17 In spite of distortion from square-
pyramidal, the 6D and 6′D signals as well as the 4D and
Figure 2. Cyclic voltammograms of (a) 1 and (b) 2 in CH2Cl2 at 193 K.
Conditions: 1 × 10-3 M 1 or 2, 0.10 M Bu4NClO4 supporting electrolyte,
a saturated calomel electrode as a reference electrode, a glassy carbon
working electrode, a platinum-wire counter electrode, a scan rate of 50 mV
s-1. The potentials are referenced versus the ferrocenium/ferrocene (Fc′/
Fc) couple.
Results
MnIII-Phenoxyl Radical from 1. An X-ray crystal-
lographic analysis shows that 1 is a five-coordinate mono-
meric MnIII complex bearing H2O as an external ligand and
ClO4- as a counteranion (Figure S1, Supporting Information).
The coordination geometry of 1, which is close to square-
pyramidal, is similar to that of other Mn(salen) com-
plexes.14,15 Cyclic voltammetry of 1 does not give any clear
redox wave at room temperature. But, as shown in Figure
2a, a cyclic voltammogram of 1 at 193 K exhibits two quasi-
reversible redox waves, indicating that oxidized derivatives
of 1 are stable only at low temperatures. The redox potentials
2
4′D signals in 1 show a single H NMR signal due to fast
pseudorotation of the salen ligand, which is in contrast to 2,
giving separate signals for both 6/6′D and 4/4′D (Vide infra).
Upon addition of 0.5 and 1.1 equiv of 3 to 1, these signals
gradually disappear, but no apparent signal appears for [1]+.
This is probably because an exchange between phenolate
and the phenoxyl radical in [1]+ via pseudorotation of the
salen ligand is comparable to the NMR timescale (∼mil-
lisecond), resulting in an extreme line broadening. An ESI-
MS spectrum of [1]+ gives a singly charged signal at m/z
1128.39, which corresponds to [Mn(salen)ClO4]+. The ion
1
2
E1/2 and E1/2 are 0.86 and 1.06 V versus the ferrocenium/
ferrocene couple (Fc′/Fc), respectively. The other parameters
are summarized in Table 1. A controlled-potential oxida-
tion of 1 at a voltage (1.0 V) higher than Ea1 but lower
than Ea2 at 203 K generates an intermediate [1]+, after
(14) Pospisil, P. J.; Carsten, D. H.; Jacobsen, E. N. Chem.—Eur. J. 1996,
2, 974–980.
(16) Schnepf, R.; Sokolowski, A.; Müller, J.; Bachler, V.; Wieghardt, K.;
Hildebrandt, P. J. Am. Chem. Soc. 1998, 120, 2352–2364.
(17) Bonadies, J. A.; Maroney, M. J.; Pecoraro, V. L. Inorg. Chem. 1989,
28, 2044–2051.
(15) Kurahashi, T.; Oda, K.; Sugimoto, M.; Ogura, T.; Fujii, H. Inorg.
Chem. 2006, 45, 7709–7721.
1676 Inorganic Chemistry, Vol. 47, No. 5, 2008

Transient Intermediates from Mn(salen)
Figure 4. Resonance Raman spectra of a 2 mM solution of (a) 1, (b)
H2O-MnIII(salen+•) generated from 1 + 1.0 equiv of 3, and (c) reduced
3 generated from 3 + 1.0 equiv of ferrocene in CD2Cl2 at 193 K. The
excitation wavelength is 351.4 nm. The band from CD2Cl2 is denoted
with “s”. The band designated with an asterisk comes from a re-
duced 3.
Information). According to the previous perpendicular- and
parallel-mode EPR study of Jacobsen’s Mn(salen),8d this
signal was assigned as the transition between the Ms ) (2
level of an S ) 2 electron spin system, and the simulation
and temperature dependence of the signals indicated D )
-2.5 cm-1 for Jacobsen’s MnIII(salen). Variable-temperature
EPR in the range of 2.3–50 K shows that the signal at g )
7.2 for 1 obeys the Curie law just as in the case of Jacobsen’s
MnIII(salen), indicative of a negative D value also for 1. Then,
the sign of the zero-field splitting value D remains negative
before and after one-electron oxidation of 1. The accumulated
spectroscopic evidence clearly shows that [1]+ is a MnIII-
phenoxyl radical, H2O-MnIII(salen+•).
Figure 3. UV–vis spectral changes upon electrochemical oxidation of (a)
1 to H2O-MnIII(salen+•) and (b) 2 to Cl-MnIV(salen) in CH2Cl2 at 203 K.
Conditions: 0.5 × 10-3 M 1 or 2; 0.10 M Bu4NClO4 supporting electrolyte;
controlled-potential oxidation at 1.00 V vs Fc′/Fc. 1, H2O-MnIII(salen+•),
2, and Cl-MnIV(salen) are depicted as a solid line. UV–vis spectra depicted
as a dashed-dotted line are obtained after controlled-potential oxidation at
1.30 V vs Fc′/Fc.
peak corresponding to [Mn(salen)(H2O)ClO4]+ is not ob-
served, because the H2O ligand is easily lost upon ionization.
Figure 5 shows EPR spectra of [1]+ measured at various
temperatures, which clearly shows two sets of signals at g
) 4.70, 3.00, and 1.85 and g ) 5.82, 1.23, and 1.16.18 Their
temperature dependence indicates that the signals at g ) 5.82,
1.23, and 1.16 arise from the ground Kramers doublet, while
the signals at g ) 4.70, 3.00, and 1.85 originate from the
excited-state doublet. Linear least-squares fits for signal
intensities versus temperature show a negative zero-field
splitting value D ) -3 cm-1 (Figure S3, Supporting
Information). These signals are derived from an St ) 3/2 spin
system (E/D ≈ 0.18). The St ) 3/2 spin system observed for
the MnIII-phenoxyl radical complex could be explained by
an antiferromagnetic coupling between the MnIII center (S
MnIV-Phenolate from 2. Cyclic voltammetry of 2 does
not give any clear redox wave at room temperature but gives
two quasi-reversible redox waves at 193 K (Figure 2b),
indicating that oxidized derivatives of 2 are also stable at
low temperatures. The first redox potential (E1/21) is 0.76 V
versus Fc′/Fc, while the second redox potential (E1/22) is 1.13
V versus Fc′/Fc. As shown in Table 1, the ∆E value for the
first redox cycle of 2 is 0.28 V, which is much larger than
the value of 0.10 V for the phenolate-phenoxyl radical redox
couple in 1. A controlled-potential oxidation of 2 at 1.0 V
leads to UV–vis spectral changes with two isosbestic points,
as shown in Figure 3b. The UV–vis spectrum of the transient
intermediate [2]+ shows characteristic absorption maxima
at 440 and 730 nm. These spectral changes are completely
reversible, and the starting spectrum of 2 is regenerated, when
a potential of -0.30 V is applied at 203 K. The addition of
1.0 equiv of 3 to the CH2Cl2 solution of 2 also generates
exactly the same species. The UV–vis spectrum of [2]+ is
considerably different from that of H2O-MnIII(salen+•),
suggesting that [2]+ is a MnIV species. Further controlled-
potential oxidation of [2]+ at a voltage (1.3 V) higher than
4
1
) /2) and the phenoxyl radical (S ) /2) on the salen
ligand.11a Signals at g ) 4.70 and 5.82 exhibit a six-line
hyperfine splitting with A ) 82.1 and 54.8 G, respectively,
5
as expected for the I ) /2 55Mn nucleus. The starting
complex 1, which is a d4 MnIII system, shows a weak, but
well-detectable, signal at g ) 7.2 (Figure S4, Supporting
(18) Two additional signals, which were observed at a higher magnetic
field (g ) 0.84 and 0.56), could not be assigned.
Inorganic Chemistry, Vol. 47, No. 5, 2008 1677

Kurahashi et al.
Figure 6. 2H NMR spectra of a 10 mM solution of (a) 2-d4 and (b)
Cl-MnIV(salen)-d4 generated from 2-d4 + 1.1 equiv of 3 in CH2Cl2 at 203
2
K. H NMR shifts are referenced to CD2Cl2 (5.32 ppm).
Figure 7. X-band EPR spectra of Cl-MnIV(salen) generated from 2 +
1.0 equiv of 3 at 4 K in frozen CH2Cl2-toluene (7:3). Conditions:
microwave frequency, 9.56 GHz; microwave power, 2.012 mW; modulation
amplitude, 7 G; time constant, 163.84 ms; conversion time, 163.84 ms.
The sharp signal designated with an asterisk comes from remaining 3.
These spectroscopic features are consistent with formation
of a MnIV complex, rather than a MnIII-phenoxyl radical
species. An ESI-MS spectrum of [2]+ gives a single signal
at m/z 1064.34 as a singly charged ion, which corresponds
to [Mn(salen)(Cl)]+. The EPR spectrum of [2]+ shown in
Figure 7 exhibits one set of signals at g ) 5.10, 3.00, and
Figure 5. X-band EPR spectra of H2O-MnIII(salen+•) generated from 1
+ 1.0 equiv of 3 in frozen CH2Cl2-toluene (7:3). Conditions: temperature,
(a) 50 K, (b) 30 K, (c) 10 K, (d) 6 K, (e) 4 K, and (f) 2.3 K; microwave
frequency, 9.56 GHz; microwave power, 2.012 mW; modulation amplitude,
7 G; time constant, 163.84 ms; conversion time, 163.84 ms. The sharp signal
designated with an asterisk comes from remaining 3.
3
1.85, which is typical of an St ) /2 spin system (E/D ≈
Ea2 at 203 K shows UV–vis spectral changes without clear
isosbestic points, suggesting that some reactions other than
one-electron oxidation occur in this process. The resulting
intermediate, which shows a broad absorption at 980 nm
similar to [1]2+, is also too unstable to be investigated by
other spectroscopic methods.
0.20).8c,d The signal at g ) 5.10 displays a six-line
hyperfine splitting with A ) 68.5 G, as expected for the
5
I ) /2 55Mn nucleus. The signals at g ) 5.10, 3.00, and
1.85, which arise from the transition between the ms )
3
(1/2 levels of an S ) /2 spin system, show Curie-law
behavior in the range of 2.3–50 K, thus indicative of the
d3 MnIV formation with a positive zero-field splitting value
D. Variable-temperature EPR shows that the starting MnIII
precursor 2 is supposed to be an S ) 2 spin system with
a negative D,8d and thus, one-electron oxidation of 2
changes the sign of the D value from negative to positive.
The combined spectroscopic data clearly indicate that [2]+
is MnIV-phenolate, Cl-MnIV(salen).
A resonance Raman spectrum of [2]+ does not show the
band near 1485 cm-1, which is assigned to Y7a′ of the
phenoxyl radical in H2O-MnIII(salen+•) (Figure S5, Sup-
porting Information). 2-d4 shows separate signals for 4/4′D
and 6/6′D at -42.0/-47.9 and 11.5/4.5 ppm, respectively
(Figure 6).8a,17 Because of distortion from square-pyramidal
2
and slow pseudorotation of a salen ligand of 2, H NMR
signals for two phenolate rings in 2 are observed at different
Interestingly, the addition of 1.0 equiv of AgSbF6 to the
CH2Cl2 solution of Cl-MnIV(salen) at 203 K generates
another distinct species, which is identified as the MnIII-
2
positions. While H NMR signals for H2O-MnIII(salen+•)
are not detected, [2-d4]+ shows well-resolved signals at
-32.5/-60.3 ppm for 4/4′D and 25.2/-3.8 ppm for 6/6′D.
2
phenoxyl radical by UV–vis, EPR, and H NMR spectra.
1678 Inorganic Chemistry, Vol. 47, No. 5, 2008

Transient Intermediates from Mn(salen)
Figure 8. UV–vis spectral changes of 1 to OdMnIV(salen) upon reaction
with O3 in CH2Cl2 at 203 K. Inset: A plot of absorbance at 600 nm versus
the equivalents of added ferrocene.
Removal of the Cl ligand on the fifth coordination site as
AgCl and the coordination of an SbF6- counteranion or H2O
in solvent induce intramolecular electron transfer from
phenolate to the MnIV center, resulting in conversion to the
MnIII-phenoxyl radical. This observation clearly indicates that
the external ligand plays a primary role in determining the
electronic structure of Mn(salen) in higher oxidation states.
OdMnIV(salen) and HO-MnIV(salen). OdMnIV(salen)
was prepared by reaction of 1 with O3, which is an oxidant
frequently employed to generate high-valent metal-oxo
species.4c,19 Upon reaction with O3 at 203 K, 1 shows
absorption spectral changes with clear isosbestic points to
generate a transient intermediate with absorptions at 355,
430, and 600 nm (Figure 8). The absorption spectrum of the
transient intermediate is similar to that of MnIV-phenolate,
Cl-MnIV(salen), rather than that of the MnIII-phenoxyl
radical, H2O-MnIII(salen+•). A redox titration of the transient
intermediate with ferrocene indicates that this intermediate
2
is in the one-electron oxidized state from 1. A H NMR
Figure 9. (a) XANES spectra of 1 (solid), HO-MnIV(salen), and
OdMnIV(salen) (20 mM frozen propionitrile solution) at 10 K. Inset: The
pre-edge region around 6540 eV. (b) The k2-weighted Fourier transform
magnitude of Mn K-edge EXAFS spectra of 1 (top), HO-MnIV(salen)
(middle), and OdMnIV(salen) (bottom), calculated in the k range of
measurement of this transient intermediate at 203 K shows
broad NMR signals at 23.4 and -33.2 ppm (Figure S6,
Supporting Information). The EPR spectrum shows signals
at g ) 4.20, 3.50, and 1.95, which arise from the transition
between the ms ) (1/2 levels of an St ) 3/2 spin system (E/D
≈ 0.05),8c,d and the signals show Curie-law behavior in the
range of 2.3–50 K, indicative of a positive D. These 2H NMR
and EPR data for the transient intermediate are also consistent
with formation of MnIV species, in comparison with those
for MnIV-phenolate, Cl-MnIV(salen) versus the MnIII-phe-
noxyl radical, H2O-MnIII(salen+•). ESI-MS spectrometry at
low temperatures does not give any molecular ion peak,
suggesting that this intermediate is a neutral complex.
Attempts to detect MndO vibrations by resonance Raman
spectroscopy with an excitation wavelength of 351.4, 413.1,
or 600 nm or by Fourier transform infrared experiments were
not successful.
2–12 Å-1
.
absorption near-edge structure (XANES) regions (Figure 9a),
the pre-edge height is not increased at all upon formation of
OdMnIV(salen) from 1, which is in contrast to significantly
increased pre-edge heights for MnV complexes.8e,20 Both the
pre-edge and the absorption edge for OdMnIV(salen) is
shifted to higher energy by 0.1 and 0.6 eV, respectively,
compared with the MnIII precursor 1 (Table 2). The extended
X-ray absorption fine structure (EXAFS) spectrum of
OdMnIV(salen) shows a clearly resolved peak at a phase-
shifted distance R′ ≈ 1 Å, as shown in Figure 9b. Curve-
fitting suggests that this peak corresponds to a bond distance
The coordination structure of OdMnIV(salen) was inves-
tigated by XAS. As seen from the Mn K-edge X-ray
(20) (a) Weng, T.-C.; Hsieh, W.-Y.; Uffelman, E. S.; Gordon-Wylie, S. W.;
Collins, T. J.; Pecoraro, V. L.; Penner-Hahn, J. E. J. Am. Chem. Soc.
2004, 126, 8070–8071. (b) Lansky, D. E.; Mandimutsira, B.; Ram-
dhanie, B.; Clausén, M.; Penner-Hahn, J.; Zvyagin, S. A.; Telser, J.;
Krzystek, J.; Zhan, R.; Ou, Z.; Kadish, K. M.; Zakharov, L.; Rheingold,
A. L.; Goldberg, D. P. Inorg. Chem. 2005, 44, 4485–4498.
(19) (a) Gross, Z.; Nimri, S. Inorg. Chem. 1994, 33, 1731–1732. (b) Fujii,
H.; Yoshimura, T.; Kamada, H. Inorg. Chem. 1997, 36, 6142–6143.
Inorganic Chemistry, Vol. 47, No. 5, 2008 1679

Kurahashi et al.
arise from the transition between the ms ) (1/2 levels of an
Table 2. XAS Edge Energies, Pre-Edge Energies, and Heights for 1,
HO-MnIV(salen), and OdMnIV(salen)
8c,d
3
St ) /2 spin system (E/D ≈ 0.20),
and the signals obey
edge
pre-edge
energy (eV)
pre-edge
peak height
the Curie law in the range of 2.3–50 K, indicative of a
positive D. ESI-MS of this intermediate21 gives a singly
charged ion at m/z 1046.51, which corresponds to [Mn(sale-
n)(OH)]+, indicating formation of a MnIV complex bearing
OH as an external ligand, HO-MnIV(salen). Comparison of
the XANES regions (Figure 9a, Table 2) shows almost
identical spectra for OdMnIV(salen) and HO-MnIV(salen),
indicating closely similar electronic and coordination struc-
tures. However, the EXAFS spectrum of HO-MnIV(salen)
is strikingly different from that of OdMnIV(salen) and shows
no clear resolved peak at a phase-shifted distance R′ ≈ 1 Å
(Figure 9b). In curve fitting of the EXAFS spectrum of
HO-MnIV(salen), at most, two shells can be used due to a
resolution. Upon formation of HO-MnIV(salen), the Mn-N/
Osalen bond distances may not differ considerably, which is
indeed the case for OdMnIV(salen). Then, the MnIV-OH
bond may be grouped into the shorter shell (∼1.81 Å, N )
2) or the longer shell (∼1.99 Å, N ) 2) from the Mn-N/
compd
energy (eV)a
1
6549.3
6549.9
6549.9
6538.8
6539.0
6538.9
0.045
0.044
0.046
HO-MnIV(salen)
OdMnIV(salen)
a The edge positions were reported as those at a normalized absorption
of 0.5.
Osalen bonds. We thus carried out two possible two-shell curve
fittings as shown in Table 3, and the best fit gives a
MnIV-OH bond of 1.83 Å, which is in agreement with the
crystallographically determined MnIV-OH distance (1.81 Å)
in a mononuclear MnIV complex with two OH ligands.29b
Interconversions among all the intermediates described herein
are summarized in Scheme 1.
Figure 10. UV–vis spectral changes of H2O-MnIII(salen+•) (---) to
HO-MnIV(salen) (—) and OdMnIV(salen) (···) upon addition of 1.0 equiv
and 2.0 equiv of Bu4NOH in CH2Cl2 at 203 K.
of 1.58 Å (Table 3), which is almost comparable to the value
of 1.55 Å for the fully characterized MnVdO complexes.9a–d
This peak, which is shorter than the other Mn-N/Osalen
bonds, could be assigned as the MnIVdO bond. It is thus
clearly indicated that the intermediate generated by reaction
of 1 with O3 is OdMnIV(salen). O≡MnV(salen), the most
likely candidate for the reactive species for epoxidation, is
not detected at all, probably because O≡MnV(salen) is so
reactive as to be readily reduced even at 203 K.
Reactivity of OdMnIV(salen). Reactions of OdMnIV
(salen) with various substrates were investigated at 223 K.
At this temperature, OdMnIV(salen) in CH2Cl2 gradually
decomposes (t1/2 ≈ 9 h). OdMnIV(salen) is more stable in
CD2Cl2 than in CH2Cl2 (kH/kD ) 3.5), suggesting that
hydrogen atom abstraction from CH2Cl2 molecules
is one of the decomposition pathways. Reactions of
OdMnIV(salen) with triphenylphosphine and methyl p-tolyl
sulfide (100 equiv) in CH2Cl2 under an Ar atmosphere give
triphenylphosphine oxide and methyl p-tolyl sulfoxide in 70
( 15 and 8.8 ( 1% yields, respectively. The 18O incorpora-
tion from the 18O-labeled OdMnIV(salen) was found to be
24 and 39% for triphenylphosphine oxide and methyl p-tolyl
sulfoxide, respectively. When the reaction is carried out under
O2 atmosphere, the yields of triphenylphosphine oxide and
methyl p-tolyl sulfoxide are increased to 180 ( 15 and 39
( 2%, suggesting that reactions of OdMnIV(salen) with
phosphine and sulfide partially proceed by a radical chain
mechanism. Reaction of OdMnIV(salen) with norbornylene
was also examined. But the addition of norbornylene (100
equiv) to the CH2Cl2 solution of OdMnIV(salen) does not
change the rate of disappearance of OdMnIV(salen). Indeed,
the yield of exo-2,3-epoxynorbornane after 10 h is found to
be less than 1%. Although OdMnIV(salen) does possess the
ability of oxygen atom transfer toward phosphine and sulfide,
the oxidizing power is not enough for epoxidation of olefins.
This suggests that OdMnIV(salen) might not be a reactive
intermediate for Jacobsen-Katsuki enantioselective epoxi-
HO-MnIV(salen) was prepared from H2O-MnIII(salen+•)
by the addition of a base at low temperatures. Upon addition
of 1.0 equiv of Bu4NOH at 203 K, H2O-MnIII(salen+•) shows
absorption spectral changes with isosbestic points at 420 and
715 nm, and the characteristic absorption at 905 nm
completely diminishes along with a concomitant increase in
absorption at 720 nm, as shown in Figure 10. Further addition
of 1.0 equiv of Bu4NOH induces slight spectral changes with
an isosbestic point at 740 nm, resulting in a UV–vis spectrum
with absorption maxima at 355, 430, and 600 nm, which is
2
exactly the same as that for OdMnIV(salen). H NMR and
EPR spectroscopy also shows that the intermediate gen-
erated by the addition of 2.0 equiv of Bu4NOH to
H2O-MnIII(salen+•) is identical with OdMnIV(salen). On the
other hand, the addition of 1.0 equiv of CF3SO3H to
OdMnIV(salen) regenerates the intermediate with absorption
at 720 nm, but further addition of CF3SO3H does not
regenerate H2O-MnIII(salen+•).
The intermediate with absorption at 720 nm is MnIV-
phenolate, not the MnIII-phenoxyl radical, as clearly indicated
from the absorption features and the well-resolved 2H NMR
signals (Figure S7, Supporting Information). This intermedi-
ate shows EPR signals at g ) 5.10, 3.10, and 1.85, which
(21) 2,6-Di-tert-butylpyridine is utilized instead of Bu4NOH, because
Bu4NOH disturbs the ionization of analytes.
1680 Inorganic Chemistry, Vol. 47, No. 5, 2008

Transient Intermediates from Mn(salen)
Table 3. EXAFS Curve Fitting Results for 1, HO-MnIV(salen), and OdMnIV(salen)
compd
Na
R (Å)b
σ2 (Å2)c
∆E0 (eV)d
R (%)e
1
Mn-Oexternal_ligand
Mn-N/Osalen
1
1
1
1
1
1
2
2
1
2
2
1
2
2
2.166
1.847
1.861
1.983
1.970
1.83
1.83
1.98
1.96
1.81
0.0024
0.0022
0.0021
3.96
3.96
3.96
15.1
15.1
6.71
6.71
8.11
7.81
6.61
7.81
-0.56
6.65
5.78
11.8
0.00002
0.00002
-0.0025
-0.0025
-0.0063
-0.0058
-0.0049
-0.0058
-0.0020
-0.0033
-0.0059
HO-MnIV(salen)
OdMnIV(salen)
Mn-Oexternal_ligand
Mn-N/Osalen
3.84
4.41
14.8
Mn-Oexternal_ligand
Mn-N/Osalen
1.96
1.58
1.81
1.99
Mn-Oexternal_ligand
Mn-N/Osalen
a Coordination number, fixed at the value shown. b Interatomic distances estimated from EXAFS curve fitting for HO-MnIV(salen) and OdMnIV(salen),
c
d
and those determined by X-ray crystal structural analysis for 1 (fixed parameter in the fitting). Debye–Waller factor. Shift in photoelectron energy
e
zero. R factors, defined as ∑|yexp(i) - ytheo(i)|/∑|yexp(i)|, where yexp and ytheo are experimental and theoretical data points, respectively.
Scheme 1
Table 4. Hydrogen Atom Abstraction from Selected Substrates by OdMnIV(salen)
substrate
BDE (kcal/mol)
product
yield (%)
k (s-1
)
1,4-cyclohexadienea
9,10-dihydroanthracenea
cyclohexeneb
76c
benzene
anthracene
2-cyclohexen-1-ol
2-cyclohexen-1-one
b
28 ( 3
2.09 × 10-3
1.11 × 10-4
6.46 × 10-5
78c
3.0 ( 1.0
5.0 ( 1.1
7.7 ( 0.7
81.6d
a
100 equiv of substrate was added to the CH2Cl2 solution of OdMnIV(salen). 500 equiv of cyclohexene was added to the CH2Cl2 solution of
c
d
OdMnIV(salen). Ref 32. Ref 43.
dation, although reactivity of the MnIVdO unit in less bulky
Mn(salen) catalysts utilized for enantioselective epoxidation
still remains to be investigated.
contaminated during a long reaction. It is interesting to note
that neither HO-MnIV(salen) nor H2O-MnIII(salen+•) gives
those oxidation products at all.
OdMnIV(salen) was reacted with 9,10-dihydroanthracene
and 1,4-cyclohexadiene. Anthracene and benzene are gener-
ated as major products, respectively. Table 4 summarizes
yields of oxidation products and pseudo-first-order rate
constants k for the decay of OdMnIV(salen) in the presence
of the substrate. As the bond dissociation energy (BDE)
of the C-H bond is increased, the reaction rate decreases.
This is consistent with the hydrogen atom abstraction
mechanism.
On the other hand, the addition of cyclohexene (500 equiv)
to the CH2Cl2 solution of OdMnIV(salen) slightly increases the
rate of disappearance of OdMnIV(salen), indicating that
OdMnIV(salen) does react with cyclohexene. Analysis of the
reaction mixture after 10 h revealed that major products are
2-cyclohexen-1-ol (5.0 ( 1.1%) and 2-cyclohexen-1-one (7.7
( 0.7%). Cyclohexene oxide is generated as a minor product
(1.2 ( 0.1%).22 Other possible oxidation products such as
2-cyclohexen-1-one oxide and bicyclohexene are not detected.
When the reaction is carried out under an O2 atmosphere, the
yields of 2-cyclohexen-1-ol and 2-cyclohexen-1-one are in-
creased to 10 ( 1.5 and 15 ( 2.1%, respectively, but the yield
of cyclohexene oxide remains at 1.2 ( 0.1%. When the 18O-
labeled OdMnIV(salen) is reacted with cyclohexene, no 18O
incorporation into 2-cyclohexen-1-ol and 2-cyclohexen-1-one
is observed. Thus, the oxygen atom in 2-cyclohexen-1-ol and
2-cyclohexen-1-one is not derived from OdMnIV(salen) but
most probably comes from oxygen gas, which might be
(22) Formation of 2-cyclohexen-1-ol and 2-cyclohexen-1-one exactly
coincides with the decay of OdMnIV. (salen), suggesting that both are
indeed derived from OdMnIV(salen) (Figure S9). In contrast, the
formation of cyclohexene oxide is saturated in the very beginning of
the reaction, albeit in the presence of OdMnIV(salen). When
OdMnIV(salen) is prepared via a ligand exchange from H2O-
MnIII(salen) and 2 equiv of Bu4NOH, reaction with cyclohexene
produced 2-cyclohexen-1-ol and 2-cyclohexen-1-one exclusively.
These observations suggest that cyclohexene oxide is not derived from
OdMnIV(salen).
Inorganic Chemistry, Vol. 47, No. 5, 2008 1681

Kurahashi et al.
Discussion
level as the excited-state doublet. Thus, one-electron oxida-
tion of the Mn center in MnIII(salen) changes the sign of the
D value from negative to positive, while the D value remains
negative upon one-electron oxidation of the phenolate in
MnIII(salen). But such correlations might be limited to the
present system, because signs and magnitudes of D are
dependent on coordination environments.8d,11a,24d,e EPR is
also a sensitive probe for the fifth external ligand, and E/D
values vary depending on Cl (∼0.20), OH (∼0.20), OMe
(∼0.07), and dO (∼0.05) as an external fifth ligand.
Interestingly, this trend is completely parallel to the UV–vis
absorption shifts in the phenolate-to-MnIV charge-transfer
band. The previous resonance Raman study showed that the
C-O stretching of the phenoxyl radical (Y7a′) appeared at
different positions depending on metal ions even in the same
ligand system, and this was attributed to different extents of
double-bond character of the phenoxyl radical C-O bond,
which are influenced by metal ions.26 Interestingly, in the
case of phenoxyl radicals from Mn(salen) and Fe(salen), Y7a′
bands are observed at exactly the same position (1485 cm-1),
which may indicate that the natures of phenoxyl radicals from
MnIII(salen) and FeIII(salen) are exactly the same, irrespective
of different redox potentials and Lewis acidities of Mn
and Fe.
In light of the present spectroscopic data, some previous
observations might be worth re-examining. One is a report
that describes two crystallographically characterized mono-
nuclear Mn complexes with tridentate ONO coordination,
which were assigned as MnIV-phenolate.27 Quite interest-
ingly, the CH2Cl2 solutions of these show absorptions at 640
nm (ε ) 1945 M-1 cm-1)/890 nm (ε ) 1090 M-1 cm-1)
and 640 nm (ε ) 3850 M-1 cm-1)/970 nm (ε ) 900
M-1cm-1), respectively. The absorptions at 640 nm are
consistent with MnIV-phenolate, but the absorptions at 890
and 970 nm rather suggest the MnIII-phenoxyl radical. This
might indicate that the reported system contains an equilib-
rium between MnIV-phenolate and the MnIII-phenoxyl radical
in solution. Another example is a paper that reports a high-
valent Mn(salen) species from Jacobsen’s catalyst.8e Reaction
of Jacobsen’s catalyst with m-CPBA generates a transient
intermediate with absorption at 684 nm (ε ) 4000 M-1
cm-1). This transient intermediate was assumed to be
MnIV(salen+•) with the 684 nm absorption assigned as a salen
ligand radical. However, the present study casts doubt on their
Spectroscopic Features of MnIV-Phenolate versus
MnIII-Phenoxyl Radical. Absorption features in the visible
region could be a fingerprint to distinguish between MnIV-
phenolate and MnIII-phenoxyl radical formulations. MnIV-
phenolate, Cl-MnIV(salen), shows an intense absorption at
730 nm (ε ) 6022 M-1 cm-1), which could be assigned to
the phenolate-to-MnIV charge-transfer band. The phenolate-
to-MnIV charge-transfer band is shifted to higher energy upon
replacement of the fifth ligand from Cl (730 nm) to OH (720
nm), OMe (600 nm),23 and dO (<600 nm). Consistent with
our observation, MnIV complexes containing phenolate
groups are reported to show intense absorption (ε )
2000–6000 M-1 cm-1) in the range of 500–650 nm.24 On
the other hand, the MnIII-phenoxyl radical, H2O-
MnIII(salen+•), exhibits a broad absorption at 905 nm (ε )
7000 M-1 cm-1). By analogy to our previous result on
Fe(salen),7 further one-electron oxidation of H2O-MnIII-
(salen+•) may most probably generate a MnIII-diphenoxyl
radical, H2O-MnIII(salen2+••), which shows an absorption
of doubled intensity at 980 nm. According to the recent
experimental and theoretical studies on metal-organic radical
complexes by Wieghardt and Neese,25 these absorptions at
905 and 980 nm are most likely assigned to intervalence
charge transfer between the phenolate and the phenoxyl
radical and ligand-to-ligand charge transfer between the
phenoxyl radicals for H2O-MnIII(salen+•) and H2O-
MnIII(salen2+••), respectively. But the only example of the
MnIII-phenoxyl radical complex from the monophenolate
complex also shows an absorption maximum at 1015 nm (ε
) 1800 M-1 cm-1),11a and thus a metal-to-ligand charge
transfer assignment may not be completely ruled out for the
905 nm band in the present case. Variable-temperature EPR
could differentiate between MnIV-phenolate and the MnIII-
phenoxyl radical from the present MnIII(salen). Temperature
dependence of EPR signals from all of the MnIV(salen)’s
investigated herein shows that transitions between the ms )
(1/2 levels are the lowest energy levels, thus indicative of
positive D values. In contrast, H2O-MnIII(salen+•) exhibits
EPR signals, which arise from transitions from both the ms
) (1/2 and Ms ) (3/2 levels, and their temperature
dependence reveals a negative D value because of the Ms )
(3/2 level as the ground Kramers doublet and the ms ) (1/2
(23) MeO-MnIV(salen) was prepared by addition of 1 equiv of NaOMe
in MeOH to H2O-MnII.I (salen+•) in CH2Cl2 at 203 K. The electronic
structure was confirmed by UV-vis, EPR, 2H NMR, and ESI-MS
spectroscopy (Figure S10).
(25) (a) Herebian, D.; Bothe, E.; Neese, F.; Weyhermüller, T.; Wieghardt,
K. J. Am. Chem. Soc. 2003, 125, 9116–9128. (b) Herebian, D.;
Wieghardt, K. E.; Neese, F. J. Am. Chem. Soc. 2003, 125, 10997–
11005. (c) Blanchard, S.; Neese, F.; Bothe, E.; Bill, E.; Weyhermüller,
T.; Wieghardt, K. Inorg. Chem. 2005, 44, 3636–3656. (d) Ray, K.;
Begum, A.; Weyhermüller, T.; Piligkos, S.; van Slageren, J.; Neese,
F.; Wieghardt, K. J. Am. Chem. Soc. 2005, 127, 4403–4415. (e) Ray,
K.; Weyhermüller, T.; Neese, F.; Wieghardt, K. Inorg. Chem. 2005,
44, 5345–5360. (f) Chlopek, K.; Bothe, E.; Neese, F.; Weyhermüller,
T.; Wieghardt, K. Inorg. Chem. 2006, 45, 6298–6307. (g) Kokatam,
S.; Ray, K.; Pap, J.; Bill, E.; Geiger, W. E.; LeSuer, R. J.; Rieger,
P. H.; Weyhermüller, T.; Neese, F.; Wieghardt, K. Inorg. Chem. 2007,
46, 1100–1111.
(24) (a) Matsushita, T.; Hirata, Y.; Shono, T. Bull. Chem. Soc. Jpn. 1982,
55, 108–112. (b) Okawa, H.; Nakamura, M.; Kida, S. Bull. Chem.
Soc. Jpn. 1982, 55, 466–470. (c) Chin, D.-H.; Saywer, D. T.; Schaefer,
W. P.; Simmons, C. J. Inorg. Chem. 1983, 22, 752–758. (d) Hartman,
J. R.; Foxman, B. M.; Cooper, S. R. Inorg. Chem. 1984, 23, 1381–
1387. (e) Lynch, M. W.; Hendrickson, D. N.; Fitzgerald, B. J.;
Pierpont, C. G. J. Am. Chem. Soc. 1984, 106, 2041–2049. (f) Tirant,
M.; Smith, T. D. Inorg. Chim. Acta 1986, 121, 5–11. (g) Kessissoglou,
D. P.; Li, X.; Butler, W. M.; Pecoraro, V. L. Inorg. Chem. 1987, 26,
2487–2492. (h) Chandra, S. K.; Basu, P.; Ray, D.; Pal, S.; Chakravorty,
A. Inorg. Chem. 1990, 29, 2423–2428. (i) Mikuriya, M.; Jie, D.;
Kakuta, Y.; Tokii, T. Bull. Chem. Soc. Jpn. 1993, 66, 1132–1139. (j)
Adam, B.; Bill, E.; Bothe, E.; Goerdt, B.; Haselhorst, G.; Hildenbrand,
K.; Sokolowski, A.; Steenken, S.; Weyhermüller, T.; Wieghardt, K.
Chem.—Eur. J. 1997, 3, 308–319.
(26) Itoh, S.; Kumei, H.; Nagatomo, S.; Kitagawa, T.; Fukuzumi, S. J. Am.
Chem. Soc. 2001, 123, 2165–2175.
(27) Dutta, S.; Basu, P.; Chakravorty, A. Inorg. Chem. 1991, 30, 4031–
4037.
(28) Green, M. T.; Dawson, J. H.; Gray, H. B. Science 2004, 304, 1653–
1656.
1682 Inorganic Chemistry, Vol. 47, No. 5, 2008

Transient Intermediates from Mn(salen)
assignment, because the 684 nm absorption most likely
originates from a phenolate-to-MnIV charge transfer, not a
phenoxyl radical (a salen ligand radical). The present spectro-
scopic consideration could also provide a foundation for future
investigations on the long-sought but still unknown intermediate
in Jacobsen-Katsuki enantioselective epoxidation.
OdMnIV(salen) during synchrotron radiation, which might
cause a rather small change in the pre-edge feature, because
no appreciable difference was observed for the first and the
final data in the present XAS data, which are composed of
10 different sequential measurements for the same sample.
In the absence of an authentic MnIVdO complex, which is
crystallographically characterized and is subjected to a
thorough XAS study, the XAS data for the crystallographi-
cally characterized O≡MnV(HMPA-B) complex is best for
comparison (H4(HMPA-B) ) 1,2-bis(2-hydroxy-2-methyl-
propanamido)benzene).20a Upon formation of MnV≡O from
MnIII, the pre-edge transition is shifted to higher energy by
1.21 eV. Because the pre-edge energy shift on forming
MnIVdO from MnIII is then expected to be less than 1.21
eV, the pre-edge energy shift of 0.1 eV for the present case
is somewhat small but does not fall far away from the
expected range. It is also noted that different measurement
conditions for 1 as a solid and OdMnIV(salen) in solution
might be a possible cause for a small pre-edge energy change.
According to the quantitative consideration of the pre-
edge intensity of metal-oxo complexes by Pecoraro and
Penner-Hahn,20a formation of d2 O≡MnV(HMPA-B) from
MnIII accompanies a ca. 9-fold increase in pre-edge intensity,
which is also observed for the crystallographically character-
ized d2 O≡CrIV(porphyrin).30a The crystallographically char-
acterized d4 OdFeIV(TMC) complex also shows an increase
in pre-edge intensity,30c which is 6 times smaller than the
increase in the case of the d2 O≡MnV(HMPA-B) and
O≡CrIV(porphyrin) complex (TMC ) tetra(N-methyl)cy-
clam). The recent study on the O≡FeV(B) complex shows a
3-fold increase in the pre-edge intensity upon forming d3
FeV≡O from FeIII (H4B ) 3,3,6,6,9,9-hexamethyl-3,4,8,9-
tetrahydro-1H-1,4,8,11-benzotetraazacyclotridecine-2,5,7,10
(6H,11H)-tetraone).31 In contrast, it was reported very
recently by Nam et al. that the pre-edge intensity is not
increased at all upon formation of the six-coordinate
O≡MnV(porphyrin) from its MnIII precursor.6e Thus, the
extent of increase in pre-edge intensity varies depending on
an electronic configuration and symmetry around the metal
center in metal-oxo complexes. A precise explanation for
the low pre-edge intensity in the present d3 OdMnIV(salen)
may be difficult at present, but several possibilities could be
pointed out. One is symmetry around the metal center, which
does affect pre-edge intensity.6e,30b Although the present
OdMnIV(salen) does not possess a trans axial ligand, the
coordination geometry might be distorted from square-
pyramidal upon forming MnIVdO due to steric hindrance,
resulting in a more centrosymmetric metal center than other
metal-oxo species. Another is a spin state of a metal center,
Mononuclear MnIVdO and MnIV-OH from Mn(salen).
In spite of their importance as key intermediates postulated
in both artificial and enzymatic reactions, very few MnIV
species with terminal oxo ligands are thoroughly
characterized,6a–c,9f thus limiting our knowledge of their
structures and reactivity. The most important finding in this
context is that OdMnIV(salen) is readily interconverted to
HO-MnIV(salen) via an acid–base equilibrium. Although
their spectroscopic properties including UV–vis and EPR are
similar, EXAFS analysis clearly shows a shorter Mn-O bond
for OdMnIV(salen) by 0.25 Å compared with HO-
MnIV(salen) (1.58 versus 1.83 Å). The previous XAS study
reveals that chloroperoxidase compound II (CPO-II) bears
an FeIV-OH bond (1.82 Å), instead of an FeIVdO one (ca.
1.65 Å).28 Then, the increase in the MIVdO bond length upon
protonation is similar between Mn and Fe complexes. In
FeIVdO chemistry,5 isolation and characterization of the
FeIV-OH unit instead of FeIVdO in CPO-II is a rather special
case, and this was attributed to strong axial electron donation
by thiolate ligation.28 On the other hand, protonation of the
present OdMnIV(salen) is a consequence of basicity of the
oxo ligand in the MnIVdO unit, which was already discussed
in detail by Groves and Spiro.6a Indeed, there have been other
examples, in which MnIV-OH species were isolated and
characterized instead of MnIVdO.29 The Mn-O bond length
in the present five-coordinate OdMnIV(salen) (1.58 Å) is
shorter than that in the five-coordinate OdMnIV(porphyrin)
(1.69 Å), which was previously estimated by the EXAFS
analysis.6c This might be possibly ascribed to a difference
in coordination environments between salens and porphyrins,
or to the slightly distorted square-pyramidal geometry in the
present sterically hindered Mn(salen).
Another point to be noted is the pre-edge features for the
present OdMnIV(salen). The present OdMnIV(salen) shows
a pre-edge and an absorption edge shifted to higher energy
by 0.1 and 0.6 eV, respectively, and the intensity of the pre-
edge of OdMnIV(salen) is not increased significantly,
compared with the MnIII precursor 1. The shift of the
absorption edge to higher energy by 0.6 eV is consistent with
the change in the oxidation state from MnIII to MnIV, but a
small change in the pre-edge feature may require particular
attention in comparison with previous results. However, we
can rule out X-ray damage such as photoreduction of
d3 MnIVdO of S ) /2 versus d3 FeV≡O of S ) /2. The
high-spin configuration in the present OdMnIV(salen) might
disturb Mn(3d) + O(2p) orbital mixing, a major factor in
facilitating a normally forbidden 1s f 3d pre-edge transition
in metal-oxo complexes, thus resulting in a lower pre-edge
intensity.
3
1
(29) (a) Makino, R.; Uno, T.; Nishimura, Y.; Iizuka, T.; Tsuboi, M.;
Ishimura, Y. J. Biol. Chem. 1986, 261, 8376–8382. (b) Yin, G.;
McCormick, J. M.; Buchalova, M.; Danby, A. M.; Rodgers, K.; Day,
V. W.; Smith, K.; Perkins, C. M.; Kitko, D.; Carter, J. D.; Scheper,
W. M.; Busch, D. H. Inorg. Chem. 2006, 45, 8052–8061.
(30) (a) Penner-Hahn, J. E.; Benfatto, M.; Hedman, B.; Takahashi, T.;
Doniach, S.; Groves, J. T.; Hodgson, K. O. Inorg. Chem. 1986, 25,
2255–2259. (b) Westre, T. E.; Kennepohl, P.; DeWitt, J. G.; Hedman,
B.; Hodgson, K. O.; Solomon, E. I. J. Am. Chem. Soc. 1997, 119,
6297–6314. (c) Lim, M. H.; Rohde, J.-U.; Stubna, A.; Bukowski,
M. R.; Costas, M.; Ho, R. Y. N.; Münck, E.; Nam, W.; Que, L., Jr
Proc. Natl. Acad. Sci. U.S.A. 2003, 100, 3665–3670.
(31) Tiago de Oliveira, F.; Chanda, A.; Banerjee, D.; Shan, X.; Mondal,
S.; Que, L. Jr.; Bominaar, E. L.; Münck, E.; Collins, T. J Science
2007, 315, 835–838.
Inorganic Chemistry, Vol. 47, No. 5, 2008 1683

Kurahashi et al.
Figure 11. Acid–base equilibrium between MnIV-phenolate and MnIII-
phenoxyl radical for Mn(salen), in comparison with the case of Fe(salen).
Interconversion and Energetics between MnIV-Pheno-
late and MnIII-Phenoxyl Radical as a Determinant Fac-
tor for Reactivity. Among OdMnIV(salen), HO-MnIV-
(salen), and H2O-MnIII(salen+•), only OdMnIV(salen) shows
hydrogen-abstracting ability. Protonation of OdMnIV(salen)
reversibly generates HO-MnIV(salen), which shows con-
siderably reduced hydrogen-abstracting ability. The reactivity
difference between MnIVdO and MnIV-OH was also noted
by Busch et al. very recently.32 They compared rate constants
for hydrogen-abstraction reactions with the fully character-
ized dihydroxo MnIV complex at pH 4.0 and 8.4 and found
that reactions at pH 8.4 are ∼14 times faster than those at
pH 4.0. The thermodynamic basis for the reactivity difference
remains unknown, because estimated BDE values of OH
bonds that form from MnIVdO and MnIV-OH are very
similar. The present study reveals for the first time the
structural difference between MnIVdO and MnIV-OH, which
might be somewhat correlated with the peculiar reactivity
difference. H2O-MnIII(salen+•) is formally regarded as a
diprotonated form of OdMnIV(salen), but the electronic
structure is not MnIV-phenolate any more, but the MnIII-
phenoxyl radical instead, which has no hydrogen-abstracting
ability for substrates such as cyclohexene. The most critical
feature in relation to the catalytic activity of Mn(salen) is that
the formation of inactive H2O-MnIII(salen+•) is considerably
unfavorable in energetics among OdMnIV(salen),
HO-MnIV(salen), and H2O-MnIII(salen+•) (Figure 11). Indeed,
H2O-MnIII(salen+•) is readily converted to HO-MnIV(salen)
by addition of Bu4NOH, while HO-MnIV(salen) cannot be
converted back to H2O-MnIII(salen+•) even in the presence of
strong protic acid CF3SO3H.33 Energetics that favor the generation
of high-valent metal-oxo species is exactly the evidence for the
high catalytic activity of Mn(salen).
Figure 12. Relevance of the present model system to the photosynthetic
water oxidation by the OEC complex.
phenoxyl radical. Attempts to generate OdFeIV(salen) by
reaction with m-CPBA also resulted in formation of the FeIII-
phenoxyl radical, and we did not observe any evidence for
an FeIV species with Mössbauer spectroscopy. In the case
of Fe(salen), the FeIII-phenoxyl radical, H2O-FeIII(salen+•),
would be a thermodynamic sink (Figure 11). High-valent
species OdFeIV(salen) and HO-FeIV(salen), which are
candidates for potent oxidants, are readily converted to
inactive H2O-FeIII(salen+•) via protonation and intramo-
lecular electron transfer. The absence of even trace FeIV
species might indicate that generation of inactive H2O-FeIII
(salen+•) is considerably favored in energetics among
OdFeIV(salen), HO-FeIV(salen), and H2O-FeIII(salen+•),
which results in no catalytic activity of Fe(salen). Such
energetics between high-valent and ligand-radical species
could also determine the oxidizing activity of metal ions in
porphyrin and other ligand systems.
Relevance to Photosynthetic Water Oxidation by the
OEC Complex. One of the key steps in photosynthetic water
oxidation is sequential one-electron oxidations and proton
transfers mediated by the uncoordinated tyrosyl radical
(TyrZ), which is positioned at ∼5.0 Å away from the Mn
center and is proposed to function as a hydrogen atom
abstractor.34 The present model system, which has a phenoxyl
radical coordinated to the MnIII center and which exhibits
conversion from H2O-MnIII(salen+•) to HO-MnIV(salen) via
proton and electron transfer, is thus quite relevant (Figure
12). Among other related model systems,35 the present model
is the first to evaluate the energetics between MnIV-phenolate
and the MnIII-phenoxyl radical. When water is bound on Mn
as H2O, one-electron oxidation of Mn(salen) generates the
MnIII-phenoxyl radical. But once H2O on Mn is deprotonated
to OH, the MnIII-phenoxyl radical is readily converted to
more stable MnIV-phenolate via intramolecular electron
transfer from the MnIII center to the phenoxyl radical.
In our previous study on Fe(salen), one-electron oxidation
of both FeIII(salen)(Cl) and FeIII(salen)(H2O) generates
the FeIII-phenoxyl radical.7 Upon one-electron oxidation,
FeIII(salen)(OH) is readily protonated by a small amount of
acid that might be formed by electrochemical water oxidation
to give FeIII(salen)(H2O), which is then oxidized to the FeIII-
(34) Hoganson, C. W.; Babcock, G. T. Science 1997, 277, 1953–1956.
(35) (a) Pecoraro, V. L.; Baldwin, M. J.; Caudle, M. T.; Hsieh, W.-Y.;
Law, N. A. Pure Appl. Chem. 1998, 70, 925–929. (b) Gupta, R.;
Borovik, A. S. J. Am. Chem. Soc. 2003, 125, 13234–13242. (c)
Maneiro, M.; Ruettinger, W. F.; Bourles, E.; McLendon, G. L.;
Dismukes, G. C. Proc. Natl. Acad. Sci. U.S.A. 2003, 100, 3707–3712.
(e) Mayer, J. M.; Rhile, I. J.; Larsen, F. B.; Mader, E. A.; Markle,
T. F.; DiPasquale, A. G. Photosynth. Res. 2006, 87, 3–20.
(32) Yin, G.; Danby, A. M.; Kitko, D.; Carter, J. D.; Scheper, W. M.; Busch,
D. H. J. Am. Chem. Soc. 2007, 129, 1512–1513.
(33) As demonstrated in the conversion of Cl-MnIV(salen) to H2O-
MnIII(salen+•), reversed electron transfer indeed occurs from the
phenolate to the MnIV center. In this case, the thermodynamically
unfavorable process is enforced by removal of Cl as an AgCl
precipitate from the equilibrium system.
1684 Inorganic Chemistry, Vol. 47, No. 5, 2008

Transient Intermediates from Mn(salen)
spectra were obtained with an LCT time-of-flight mass spectrometer
equipped with an electrospray ionization interface (Micromass). The
low-temperature ESI-MS method was reported previously.7 The
500 MHz NMR spectra were measured on an LA-500 spectrometer
(JEOL). Elemental analyses were conducted on a CHN corder MT-6
(Yanaco). Gas chromatography–mass spectroscopy (GC-MS) analy-
sis was performed on a QP-5000 GC-MS system (Shimadzu)
equipped with a capillary gas chromatograph (GC-17A, CBP5-
M25–025 capillary column). Quantitative analysis was performed
with GC-MS using undecane or benzophenone as an internal
standard. Experiments were repeated three times, and averaged
yields relative to 1 were reported.
XAS Measurements. Mn K-edge XAS data were obtained at
SPring-8, beam line BL01B1, under ring conditions of 8 GeV and
ca. 100 mA (Proposal No. 2007A1090). The XAS datum of 1 in
the solid state was collected at room temperature, and the XAS
data of OdMnIV(salen) and HO-MnIV(salen) in frozen propionitrile
solution were collected at 10 K. The data collections were carried
out in fluorescence mode with a Lytle detector, using a Si(111)
double-crystal monochromator. A Cu foil was utilized to calibrate
the energy. In the present XAS data, which are composed of 10
different sequential measurements for the same sample, the first
and the final data were carefully compared, but no appreciable
difference was observed, suggesting no X-ray damage such as
photoreduction during synchrotron radiation.
The XAS data were Fourier-transformed between k ) 2 and 12
Å-1 and processed in a standard manner using WinXAS software
(version 3.1).37 Theoretical EXAFS signals were calculated using FEFF
(version 8.4).38 Structural parameters obtained by curve fittings are
summarized in Table 3. The coordination number (N) was fixed as
indicated in Table 3 in curve-fitting procedures. The other parameters,
including the interatomic distance (R), the shift in photoelectron energy
zero (∆E0), and the Debye–Waller factor (σ), were allowed to vary.
Curve fitting for structurally characterized 1, in which R was fixed,
shows that parameters for OdMnIV(salen) and HO-MnIV(salen) were
reasonably well-converged. Curve fittings were carried out only for
atoms which are coordinated to Mn. Curve-fitting results are shown
in Figure S8 (Supporting Information).
Materials. CH2Cl2 was purchased from Kanto as anhydrous
solvent and was stored in the presence of 4A molecular sieves. To
remove traces of HCl, CH2Cl2 was passed through activated alumina
under an Ar atmosphere just before use. Bu4NClO4 was purchased
from Kanto and was dried over P2O5 in vacuo. CD2Cl2 was
purchased from Acros and was stored in the presence of 4A
molecular sieves. CD2Cl2 was also passed through activated alumina
under an Ar atmosphere just before use. Propionitrile (99%),
AgClO4·H2O (99.999%), AgSbF6 (98%), and phenol-d6 were
purchased from Aldrich and were used as received. sec-Butyl-
lithium and silica gel (60N, spherical neutral, particle size 40–100
µm) were purchased from Kanto. Bu4NBr3 was purchased from
TCI. Mn(OAc)2·4H2O was purchased from Wako. H218O was
purchased from Cambridge Isotope Laboratories. The oxidant, 3,
was prepared according to the reported method39 and was assayed
with titration by ferrocene. m-CPBA was purchased from Nacalai
and was purified by washing with a phosphate buffer.40 The purity
of m-CPBA was checked with iodometry. The synthesis of bis(3,5-
dimesitylsalicylidene)-1,2-dimesitylethylenediamine was already
reported elsewhere.15 The synthesis of bis(3,5-dimesitylsalicylidene)-
Therefore, the present model suggests a proton and electron
transfer mechanism, in which deprotonation of the Mn-bound
H2O triggers the electron transfer from Mn to TyrZ. It is
interesting to note that a recent time-resolved X-ray study
on the OEC complex argues against the validity of the
hydrogen atom abstraction model and leads to a proposal of
an alternative mechanism, in which TyrZ brings an electro-
static effect on the Mn-bound H2O to induce a deprotonation
reaction that is a prerequisite to the subsequent electron
transfer to TyrZ.36 Notably, the Fe complex in the same salen
framework generates the FeIII-phenoxyl radical irrespective
of an external ligand, and electron transfer from the FeIII
center to the phenoxyl radical does not occur in any case.7
Photosynthetic water oxidation is carried out exclusively by
the Mn center, not by Fe, possibly due to feasible electron
transfer from the Mn center to the tyrosyl radical.
Conclusion
We described herein the selective preparation and thorough
characterization of OdMnIV(salen), HO-MnIV(salen), and
H2O-MnIII(salen+•). OdMnIV(salen) shows oxygen atom
transfer and hydrogen atom abstraction abilities, while
HO-MnIV(salen), which only differs in the degree of
protonation, but bears a longer Mn-O bond compared to
OdMnIV(salen), shows reduced oxidizing activity. Poor
oxidizing species H2O-MnIII(salen+•), which is also in a one-
electron oxidized state with H2O as an external ligand, is
not described as MnIV-phenolate but the MnIII-phenoxyl
radical, indicating that transition from active MnIV-phenolate
to the inactive MnIII-phenoxyl radical is controlled by the
degree of protonation of the fifth oxygen ligand. Ener-
getics among OdMnIV(salen), HO-MnIV(salen), and
H2O-MnIII(salen+•) favor the generation of active
OdMnIV(salen), which is exactly the key for high catalytic
activity of Mn(salen). The present study not only reveals
the electronic basis for the high catalytic activity of Mn(salen)
far beyond Fe(salen) but also may provide valuable insight
into the development of efficient oxidation catalysts in the
future.
Experimental Section
Instrumentation. Cyclic voltammograms were measured with
an ALS612A electrochemical analyzer. UV–vis spectra were
recorded on an Agilent 8453 (Agilent Technologies) equipped with
a USP-203 low-temperature chamber (UNISOKU). UV–vis mea-
surements upon controlled-potential electrochemical oxidation were
reported in the previous paper.7 EPR spectra were recorded in a
quartz cell (d ) 5 mm) at 4 K on an E500 continuous-wave X-band
spectrometer (Bruker) with an ESR910 helium-flow cryostat
(Oxford Instruments). Resonance Raman spectra were measured
with a quartz spinning cell kept at 193 K, using an MC-100DG
100 cm single polychromator (Ritsu Ohyo Kogaku) equipped with
an LN/CCD-1100-PB liquid-nitrogen-cooled CCD detector (Roper
Scientific). A BeamLok 2080 argon ion laser (Spectra Physics) was
utilized as an excitation source. The laser power at the sample was
about 10 mW. Raman shifts were calibrated with indene. ESI-MS
(37) Ressler, T. J. Synchrotron Radiat. 1998, 5, 118–122.
(38) Ankudinov, A. L.; Bouldin, C. E.; Rehr, J. J.; Sims, J.; Hung, H. Phys.
ReV. B: Condens. Matter Mater. Phys. 2002, 65, 1041071–11.
(39) Schmidt, W.; Steckhan, E. Chem. Ber. 1980, 113, 577–585.
(40) Schwartz, N. N.; Blumbergs, J. H. J. Org. Chem. 1964, 29, 1976–
1979.
(36) Haumann, M.; Liebisch, P.; Müller, C.; Barra, M.; Grabolle, M.; Dau,
H. Science 2005, 310, 1019–1021.
Inorganic Chemistry, Vol. 47, No. 5, 2008 1685

Kurahashi et al.
1,2-dimesitylethylenediamine-d4 was carried out in the same manner
except for the formylation of 1,3-dimesityl-2-(methoxymethoxy)-
benzene-d3, in which sec-butyl-lithium was utilized instead of
n-butyl-lithium. 2,4-Dibromophenol-d3 was prepared by bromination
of phenol-d6 with Bu4NBr3.41 CAUTION! The perchlorate salts
used in this study are potentially explosiVe and should be handled
in a small amount with great care.
Synthesis of 2. Following the method reported by Jacobsen and
Zhang,42 Mn(OAc)2·4H2O (251.2 mg, 1.02 mmol) was added to a
solution of bis(3,5-dimesitylsalicylidene)-1,2-dimesitylethylenedi-
amine (501 mg, 0.51 mmol) in dry ethanol (30 mL). The mixture
was heated to reflux for 1 h. Then, ethanol was removed by rotary
evaporation, and the residue was dissolved in CH2Cl2 (30 mL).
The organic layer was washed with 0.1 M HCl aqueous solution
(20 mL × 3) and saturated NaCl aqueous solution (20 mL × 1)
and was dried over anhydrous Na2SO4. The solvent was removed
by rotary evaporation, and the residue was purified by silica flash
column chromatography (CH2Cl2/methanol 50:1). The residue
obtained after evaporation was dissolved in a minimum amount of
CH2Cl2 and was passed through a membrane filter (Cosmonice Filter
S, Nacalai). The residue was further purified by precipitating from
hot acetonitrile solution to give 2 (227.4 mg, 0.21 mmol). Anal.
calcd for C70H74N2O2ClMn·H2O: C, 77.58; H, 7.07; N, 2.58. Found:
C, 77.76; H, 6.90; N, 2.61.
Preparations of OdMnIV(salen), HO-MnIV(salen), MeO-
MnIV(salen), and H2O-MnIII(salen+•) via Ligand Exchange
Reactions. For preparations of OdMnIV(salen) and HO-MnIV(salen),
the solution of 1.0 and 2.0 equiv of Bu4NOH dissolved in a minimum
amount of CH2Cl2 was added to the CH2Cl2 solution of
H2O-MnIII(salen+•) at 193 K. For preparation of MeO-MnIV(salen),
the solution of 1.0 equiv of NaOMe dissolved in a minimum amount
of methanol was added to the CH2Cl2 solution of H2O-MnIII(salen+•)
at 193 K. In order to convert Cl-MnIV(salen) to H2O-MnIII(salen+•),
the solution of 1.0 equiv of AgSbF6 dissolved in a minimum amount
of CH2Cl2 was added to the CH2Cl2 solution of Cl-MnIV(salen) at
193 K. The concentrations are 0.5 mM for UV–vis, 0.5 mM for ESI-
2
MS, 1.0 mM for EPR, 10 mM for H NMR, and 20 mM for XAS.
Ligand exchange reactions were accomplished immediately, and
thereafter the samples were subjected to the physical measurements.
For ESI-MS measurements, the solution was diluted with an excess
of cold CH2Cl2. The solvent mixture of CH2Cl2/toluene (7:3) was
utilized for EPR measurements. The XAS sample of HO-MnIV(salen)
was prepared in propionitrile. The CH2Cl2 solution of 18OdMnIV(salen)
was prepared from the 18O-labeled 1, 3, and Bu4N18OH in exactly the
same manner as described above. The 18O-labeled 1 was prepared by
dissolving 1 (∼50 mg) in H218O (0.1 mL) and anhydrous THF (0.9
mL) and then evaporating the solvent in vacuo. This procedure was
repeated three times. Bu4N18OH was prepared similarly.
Reactions of OdMnIV(salen) with Various Substrates. The
OdMnIV(salen) solution was prepared by reaction of 1 (0.4
µmol) in CH2Cl2 (0.4 mL) with O3. The resulting solution was
frozen in liquid nitrogen and thawed at 193 K under a vacuum
to remove excess O3 gas. This freeze–thaw cycle was repeated
five times, and finally Ar or O2 gas was introduced to the reaction
vessel. Then, the solution of substrate (40 µmol) in CH2Cl2 (40
µL), which was degassed via five freeze–thaw cycles just prior
to use, was added to the OdMnIV(salen) solution at 223 K. After
being stirred for 5 min for triphenylphosphine, 30 min for 1,4-
cyclohexadiene and 9,10-dihydroanthracene, 1 h for methyl
p-tolyl sulfide, and 10 h for norbornylene and cyclohexene at
223 K under an Ar or O2 atmosphere, the reaction was quenched
by the addition of Bu4NI (4 µmol) in degassed CH2Cl2 (20 µL).
A quantitative analysis of products was carried out by GC-MS.
The averaged yields of products were determined from three
independent experiments and were reported relative to 1. A blank
experiment in the absence of 1 showed that no triphenylphos-
phine oxide was formed. The 18O-labeled OdMnIV(salen)
intermediate, which was prepared as described above and was
degassed via a freeze–thaw method, was subjected to reactions
with substrates in exactly the same manner. The 18O incorpora-
tion was determined by GC-MS.
Synthesis of 1. AgClO4·H2O (29.2 mg, 0.13 mmol) in THF (1
mL) was added to a solution of 2 (136.4 mg, 0.13 mmol) in THF
(5 mL). The mixture was stirred at room temperature for 30 min
and was then filtered to remove AgCl precipitate. The solvent was
removed by rotary evaporation, and the residue was dried in vacuo.
The residue dissolved in a minimum amount of CH2Cl2 was passed
through a membrane filter (Cosmonice Filter S, Nacalai). Crystal-
lization from CH2Cl2-hexane gave 1 (79.8 mg, 0.070 mmol). Anal.
calcd for C70H74N2O6ClMn·2H2O: C, 72.12; H, 6.74; N, 2.40.
Found: C, 72.17; H, 6.92; N, 2.56.
Preparations of H2O-MnIII(salen+•) and Cl-MnIV(salen). The
solution of 1.0 equiv of 3 dissolved in a minimum amount of CH2Cl2
was added to the CH2Cl2 solution of 1, 2, 1-d4, and 2-d4 at 193 K.
The concentrations are 0.5 mM for UV–vis, 0.5 mM for ESI-MS, 1.0
mM for EPR, 2.0 mM for resonance Raman, and 10 mM for 2H NMR.
One-electron oxidations were accomplished immediately, and the
samples were subjected to the physical measurements just after addition
of 3. For ESI-MS measurements, the solution was diluted with an
excess of cold CH2Cl2. The solvent mixture of CH2Cl2/toluene (7:3)
was utilized for EPR measurements.
Preparation of OdMnIV(salen). The stream of O3 and O2
prepared by UV-irradiation to the O2 gas was passed through the
CH2Cl2 solution of 1 at 193 K. The concentrations are 0.5 mM for
Acknowledgment. We thank Mr. Seiji Makita (IMS) for
elemental analysis and Dr. Hajime Tanida (JASRI/SPring-8)
for assistance in the measurement of X-ray absorption spectra.
This work was supported by grants from the Ministry of
Education, Culture, Sports, Science and Technology, Japan, and
from the Japan Science and Technology Agency, CREST.
2
UV–vis, 0.5 mM for ESI-MS, 1.0 mM for EPR, 10 mM for H
NMR, and 20 mM for XAS. For ESI-MS measurements, the
solution was diluted with an excess of cold CH2Cl2. The solvent
mixture of CH2Cl2/toluene (7:3) was utilized for EPR measure-
ments. The XAS sample was prepared in propionitrile.
Supporting Information Available: The ORTEP view (Figure
S1) and crystallographic data (Table S1) of 1, X-ray crystallographic
files in CIF format for 1, and Figures S2–S10. This material is
available free of charge via the Internet at http://pubs.acs.org.
(41) Kajigaeshi, S.; Kakinami, T.; Okamoto, T.; Nakamura, H.; Fujikawa,
M. Bull. Chem. Soc. Jpn. 1987, 60, 4187–4189.
(42) Zhang, W.; Jacobsen, E. N. J. Org. Chem. 1991, 56, 2296–2298.
(43) Denisov, E. T.; Denisova, T. G. Handbook of Antioxidants: Bond
Dissociation Energies, Rate Constants, ActiVation Energies and
Enthalpies of Reactions, 2nd ed.; CRC Press: Boca Raton, FL, 2000.
IC702061Y
1686 Inorganic Chemistry, Vol. 47, No. 5, 2008