Probing the reactivity of oxomanganese-salen complexes: An electrospray tandem mass spectrometric study of highly reactive intermediates

Article abstract of DOI:10.1002/1521-3765(20010202)7:3<591::AID-CHEM591>3.0.CO;2-I

Electrospray ionization in combination with tandem mass spectrometric techniques has been employed to study the formation of oxomanganese-salen complexes upon oxidation of [Mn^III(salen)]⁺ cations as well as the properties and reactions of the oxidized species in the gas phase. Two species could be characterized as the principal oxidation products: the oxomanganese(V) complex, [Mn=O(salen)]⁺, which is the actual oxygen-transfer agent in epoxidation reactions, and the dinuclear, μ-oxo bridged [L(salen)Mn-Q-Mn-(salen)L]²⁺ with two terminal ligands L; the latter acts as a reservoir species. The effects of various substituents in the 5-and 5′-positions, respectively, of the salen ligand on the reactivity of the epoxidation catalyst were determined quantitatively from CID (collision-induced dissociation) experiments and B3LYP density functional calculations. Accordingly, the effect of axial donor ligands on the reactivity of the epoxidation catalyst was studied. Electron-withdrawing substitutents on the salen ligand and additional axial ligands decrease the stability and thus enhance the reactivity of the Mn=O moiety, while electron-donating salen substituents have a strong stabilizing effect. WILEY-VCH Verlag GmbH, 2001.

Full text of DOI:10.1002/1521-3765(20010202)7:3<591::AID-CHEM591>3.0.CO;2-I

FULL PAPER
Probing the Reactivity of Oxomanganese Salen Complexes: An Electrospray
Tandem Mass Spectrometric Study of Highly Reactive Intermediates
Derek Feichtinger and Dietmar A. Plattner*[a]
Abstract: Electrospray ionization in
combination with tandem mass spectro-
metric techniques has been employed to
study the formation of oxomanga-
nese salen complexes upon oxidation
m-oxo bridged [L(salen)Mn O Mn-
(salen)L]2‡ with two terminal ligands L;
the latter acts as a reservoir species. The
effects of various substituents in the 5-
and 5'-positions, respectively, of the
salen ligand on the reactivity of the
epoxidation catalyst were determined
quantitatively from CID (collision-in-
duced dissociation) experiments and
B3LYP density functional calculations.
Accordingly, the effect of axial donor
ligands on the reactivity of the epoxida-
tion catalyst was studied. Electron-with-
drawing substitutents on the salen ligand
and additional axial ligands decrease the
stability and thus enhance the reactivity
‡
of [MnIII(salen)] cations as well as the
properties and reactions of the oxidized
species in the gas phase. Two species
could be characterized as the principal
oxidation products: the oxomanga-
Keywords: density functional calcu-
lations ´ epoxidations ´ gas-phase
ˆ
of the Mn O moiety, while electron-
‡
ˆ
nese(v) complex, [Mn O(salen)] , which
donating salen substituents have
strong stabilizing effect.
a
chemistry
´
manganese
´
is the actual oxygen-transfer agent in
epoxidation reactions, and the dinuclear,
mass spectrometry
The mechanistic scheme adopted for oxygen transfer to
organic substrates by salen complexes is based on the isolation
and characterization of an oxochromium(v) species by Kochi
and co-workers (Scheme 1).[13±15] This mechanism was in
Introduction
The oxidation of organic substrates in biological systems is
accomplished by oxygenase enzymes with high-valent oxo-
metalloporphyrin moieties in the active site.[1±9] The study of
cytochrome P-450 and porphyrin model complexes designed
to mimic its reactivity has yielded a number of synthetically
useful catalysts for the epoxidation and hydroxylation of
organic substrates.[10±12] The insight gained by the study of
metal porphyrin-based systems was readily transferred to
metal salen complexes (salen ˆ N,N'-bis(salicylidene)ethyl-
enediamine), introduced by Kochi and co-workers as versatile
epoxidation catalysts in the 1980s.[13±17] A breakthrough was
achieved in the field of enantioselective epoxidation through
the introduction of chiral manganese salen catalysts by
Jacobsen and co-workers,[18] with a similar system developed
by Katsuki and co-workers at about the same time.[19] The
Jacobsen ± Katsuki reaction is universally recognized as one of
the most useful and widely applicable methods for the
epoxidation of unfunctionalized olefins.[20, 21]
[MIII(salen)]X
O
+
PhIO
+
PhI
+
+
O
M
L
N
O
N
O
N
O
N
O
PhIO
– PhI
M
L
O
[O=MV(salen)L]+
M = Cr, Mn
+
[MIV(salen)L]+
+
Scheme 1. The mechanism of alkene epoxidation catalyzed by metal salen
complexes.
accordance with the properties and reactivity of an analogous
oxoporphinatochromium(v) complex studied earlier by
Groves and Kruper, who coined the term ªoxygen-rebound
mechanismº.[22] Switching from chromium to manganese,
Kochi and co-workers discovered a much more versatile
salen-based oxidation catalyst.[16, 17] However, mechanistic
studies on these systems have so far been hampered by the
[a] Dr. D. A. Plattner, Dr. D. Feichtinger
Laboratorium für Organische Chemie der Eidgenössischen
Technischen Hochschule Zürich, ETH-Zentrum
Universitätstrasse 16, 8092 Zürich (Switzerland)
Fax : (‡41)1-632-1280
E-mail: plattner@org.chem.ethz.ch
Supporting information for this article is available on the WWW under
http://www.wiley-vch.de/home/chemistry/ or from the author.
Chem. Eur. J. 2001, 7, No. 3
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0703-0591 $ 17.50+.50/0
591

D. A. Plattner and D. Feichtinger
FULL PAPER
fact that the catalytically active species appear only as fleeting
putative intermediates.
different samples of varying properties caused us to look for
alternatives. Amine N-oxides have been widely used as
ligands for manganese porphyrin and salen complexes.[16, 33]
Since amine N-oxides are much poorer oxidants as compared
with PhIO, the best method to prepare relatively stable m-oxo
We have found that the reactivity of short-lived intermedi-
ates in the oxidation process can readily be addressed by
transferring the metal oxo complexes to the gas phase. By
mass selection, the species of interest can be singled out and
studied separately without interference by additional com-
plexes present in solution. The reactivity of the crucial
intermediates can be monitored by directed collision with
an appropriate substrate; due to high-vacuum conditions, the
intermediatesꢁ short solution-phase lifetimes do not pose a
problem. Electrospray ionization provides a powerful tool for
the transfer of medium-to-large molecular ions to the gas
phase with minimum fragmentation.[23±25] Recently, electro-
spray mass spectrometry (ESMS) has become increasingly
popular as an analytical tool in inorganic/organometallic
chemistry.[26, 27] We have demonstrated the usefulness of
electrospray tandem mass spectrometry for mechanistic and
thermochemical analysis in organometallic chemistry in
‡
complexes was found in the mixing of [MnIII(salen)] and the
amine N-oxide (ꢀ1:10) in a slurry of iodosobenzene in
acetonitrile. A representative spectrum of the electrosprayed
supernatant solution with p-CN C6H4NMe2O as the ligand is
shown in Figure 1. The species that appear most prominently
‡
accounts on the C H activation by [CpIr(PMe3)(CH3)] ,[28]
gas-phase olefin oligomerization by ªnakedº alkylzircono-
cene cations,[29] and the oxygen-transfer reactivity as well as
V
‡
[30±32]
ˆ
coordination chemistry of [O Mn (salen)] .
In this
report, we present a systematic study of substituent and
ligand effects on the intrinsic reactivity of high-valent
oxomanganese salen complexes that are the catalytically
active species in the Kochi ± Jacobsen ± Katsuki epoxidation.
Figure 1. Electrospray mass spectrum of a solution of [Mn(salen)]ClO4, p-
CN-N,N-dimethylaniline N-oxide, and iodosobenzene in acetonitrile; this
shows the formation of N-oxide ligated manganese(iii) and oxomanganese
species.
‡
in the spectrum are [MnIII(salen)p-CN C6H4NMe2O] (m/z:
Results and Discussion
483), [MnV O(salen)p-CN C6H4NMe2O] (m/z: 499), and
[p-CN C6H4NMe2O(salen)MnIV O MnIV(salen)OMe2N-p-
CN C6H4]2‡ (m/z: 491). The signal at m/z: 325 is due to the
H -bridged N-oxide adduct [p-CN C6H4NMe2O H
ONMe2-p-CN C6H4] . Amine N-oxides apparently are much
better ligands than iodosobenzene or acetonitrile, which are
both displaced, and they are very effective in stabilizing the m-
oxo bridged manganese(iv) complexes.
‡
ˆ
m-Oxomanganese complexes: Previously, we reported the
detection of oxomanganese(v) salen complexes by electro-
spray of in situ mixtures of MnIII(salen) and suitable oxygen
transferring agents, for example, iodosobenzene.[30, 32] The
oxidized species most prominent in the spectrum are the
‡
‡
V
‡
ˆ
parent oxo-complex [O Mn (salen)] and the dinuclear, m-
oxo bridged complex with two terminal PhIO ligands
[PhIO(salen)Mn O Mn(salen)OIPh]2‡; the latter acts as a
reservoir species for parking the catalytically active complex
in a more persistent form [Eq. (1)]. In addition to the
Effects of electronic tuning of the salen ligand: The possibility
to generate and stabilize a species too reactive to be
detectable in condensed phases tempted us to extend our
experiments to study electronic effects of salen substituents
on the reactivity of the catalytically active oxomanganese(v)
species. Evidently, the way to probe the influence of electronic
tuning, namely CID (collision-induced dissociation) threshold
measurements[34] of differently substituted [MnIII(salen) ´
PhIO] adducts, was precluded by the sheer size of the
molecules. In the case of the cyclometallation reaction
V
‡
ˆ
analytical detection of the [O Mn (salen)] cation, we were
‡
ˆ
able to demonstrate that [O Mn(salen)] displays oxygen-
transfer reactivity with respect to suitable substrate molecules
(olefins, sulfides).[30, 31]
V
‡
‡
[O Mn (salen)] ‡ [MnIII(salen)] > [(salen)MnIV O MnIV(salen)]2‡ (1)
ˆ
m-Oxomanganese(iv) complexes without terminal ligands
or with acetonitrile instead of iodosobenzene were conspic-
uously absent in all the mass spectra recorded. Since
acetonitrile binds readily to the .[MnIII(salen)] complex and
can only be removed by applying moderate tube lens
potentials (>40 V), we came to the conclusion that m-oxo
bridged species have to be coordinated by much better ligands
in order to increase their lifetimes in solution to a level, at
which they become detectable by ESMS. Iodosobenzene is
clearly efficient in stabilizing a m-oxo complex,[32] but the
lability of the I O bond and the problems experienced with
‡
‡
[CpIr(PMe3)(CH3)] ! [CpIr(h2-CH2PMe2)] ‡ CH4,
we
were able to demonstrate that reliable thermochemical
data can be obtained by CID threshold measurements
with our experimental setup,[28b] but it also became clear from
this study that the number of degrees of freedom and
the quality of the frequencies obtained from quantum
chemical calculations needed for the RRKM (Rice ± Ram-
sperger± Kassel ± Marcus) correction pose a serious problem
when solution-phase species with a full ligand sphere are
studied.
592
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0703-0592 $ 17.50+.50/0
Chem. Eur. J. 2001, 7, No. 3

Reactivity of Oxomanganese Salen Complexes
591±599
The presence of the m-oxomanganese(iv) complexes led us
to devise an alternative experiment to assess substituent
following order: NO2 > Cl  H > F > CH3 > OMe. The net
effect of the 5,5'-dichloro-substituted ligand as compared with
the unsubstituted salen is zero within the limits of exper-
imental error. With the electron-donating substituents, the
5,5'-dimethyl-oxomanganese(v) complex is twice as stable as
the unsubstituted one. The maximum stability is achieved by
the 5,5'-dimethoxy substitution, which yields an oxomanga-
nese(v) species fifteen times more stable than the unsubsti-
tuted salen complex (Figure 2). A small stabilizing effect was
ˆ
effects on the reactivity of the Mn O species. When a
dinuclear m-oxo complex with different salen ligands on each
side is fragmented, the ratio of the resulting two oxomanga-
nese(v) complexes (and, accordingly, the corresponding
MnIII(salen) fragments) will reflect their kinetic stability
(Scheme 2). On the assumption that the reverse barrier for
Scheme 2. Possible fragmentation products upon CID of m-oxomanga-
nese(iv) salen complexes with salen ligands with different substituents in
the 5- and 5'-positions.
both reaction channels is roughly equal, entropic factors
cancel out, and the energy distribution of the reactant ions can
be approximated by a Boltzmann distribution; the observed
difference in kinetic stability will reflect the difference in
thermodynamic stability as well (the assumptions used for the
evaluation of the branching ratios and the physical back-
ground on which they are based are specified in the Support-
ing information).[35] Thus, it is possible to establish an ordering
of the relative stabilities of oxomanganese(v) salen com-
plexes; this stability depends on the electronic influence of the
salen substituents.
It is known from the literature that substituents in the 5-
and 5'-positions of the salen ligand strongly perturb the redox
properties of [metal(salen)] complexes.[36] We chose the
following salen derivatives for our studies: 5,5'-dinitro-, 5,5'-
difluoro-, 5,5'-dichloro-, 5,5'-dimethyl-, 5,5'-dimethoxy-, and
the unsubstituted salen itself. The results of the fragmentation
of the ªmixedº m-oxo complexes are shown in Table 1. The
destabilizing effect on the oxomanganese(v) complex is in the
Figure 2. Top: electrospray mass spectrum of
a
solution of
[Mn((MeO)2 salen)]ClO4, [Mn(salen)]ClO4, p-CN-N,N-dimethylaniline
N-oxide, and iodosobenzene in acetonitrile, and this shows the formation
of homogeneous and mixed m-oxo complexes with terminal N-oxide
ligands; bottom: daughter-ion spectrum (0.09 PaAr, collision energy
34 eV) of the mixed dinuclear complex [L((MeO)2 salen)Mn O Mn-
(salen)L]2‡ (m/z: 521, L ˆ p-CN-N,N-dimethylaniline N-oxide), and this
‡
shows both the fragmentation to [MnL((MeO)2 salen)] (m/z: 543) and to
the corresponding oxo complex (m/z: 559).
found with the 5,5'-difluorosubstituted salen. Conversely,
substitution with electron-withdrawing groups decreases the
stability of the oxygen transferring species markedly. The
dinuclear m-oxo complex with the 5,5'-dinitrosubstituted salen
can only be seen right after mixing with amine N-oxide/PhIO
and disappears within seconds from the spectrum. Accord-
ingly, the formation of mixed dinuclear complexes with 5,5'-
dinitrosalen is very difficult to detect, and we found them to
be too short-lived for conducting fragmentation experiments.
The product yields in the fragmentation experiments reflect
the trend in stability of the oxomanganese(v) ions that one
would conceive intuitively when one considers the electronic
Table 1. Branching ratios (IR/IH) and differences in activation energies for
the two principal pathways of the dissociation of the ªmixedº m-oxo
complex. A value >1 indicates a stabilizing effect of the substituent on the
oxomanganese(v) species.
[a]
Substituent
OMe
Me
F
H
Cl
NO2
±
IR/IH
14.6
2.16
1.39
(1)
1.03
Æ 2.6
Æ 0.16
0.32
1.56
Æ 0.10
Æ 0.11
‡
sp
0.78
0.07
0
0
0.11
0.79
3.98)
DE
3.93
0.35
0.55
(
[a] The branching ratio could not be determined due to the extremely fast
decay of the m-oxo complex.
Chem. Eur. J. 2001, 7, No. 3
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0703-0593 $ 17.50+.50/0
593

D. A. Plattner and D. Feichtinger
FULL PAPER
ˆ
properties of the substituted salen ligands. The Mn O moiety
in these high-valent complexes is stabilized by electron-
donating and destabilized by electron-withdrawing substitu-
tents. This result can easily be rationalized by the electron
deficiency imposed on the manganese center upon oxidation
Two different approaches have been taken in order to relate
the experimental branching ratio of one dissociation reaction
to absolute differences in activation energy (see Supporting
information). Based on the linear free energy relationship
displayed in Figure 3, all the remaining dissociation energies
can be interpolated. The thermochemistry of the homodes-
motic reaction shown in Equation (2) reflects the energy
to MnV O. The differences in stability derived from gas-phase
ˆ
experiments are quite pronounced. The ordering of the
substituent effects suggests an underlying linear free energy
relationship. When we plotted ln(k1/k2) versus Hammett
parameters found in the literature, the best correlation was
obtained with the s values given by H. C. Brown et al.
(Figure 3).[37] These electrophilic substituent constants are
‡
‡
ˆ
[Mn(salen)] ‡ [O Mn(5,5'-(CH3)2 salen)]
!
(2)
‡
‡
ˆ
[O Mn(salen)] ‡ [Mn(5,5'-(CH3)2 salen)]
‡
difference of the transition states of the two possible
fragmentation pathways. For an accurate determination of
the thermochemistry, the hybrid Hartree ± Fock/density func-
tional method (B3LYP) together with the 6 ± 311G* basis set
was employed. The fully optimized structures of reactants and
products are shown in Figure 4.[38]
In an alternative approach, the energetics of the dissocia-
tion reaction (Scheme 2) have been calculated using a
statistical model of the collision process and of the unim-
olecular reaction kinetics (see Supporting information for
details). The DFT (density functional theory) calculation gave
1
a difference between the activation energies of 5.0 kJmol ,
whereas the statistical kinetic model gave an energy differ-
1
ence of 2.3 kJmol . The DFT calculations should give a very
accurate description of the thermochemistry as a result of the
high level of theory employed and the homodesmotic nature
of the reaction. The kinetic model, on the other hand,
provides an approximate lower boundary for the energy
differences by treating the dication as an ideally ergodic
system. However, the major substructures (Mn(salen), termi-
nal ligands) are only linked by Mn O bonds and thus
probably weakly coupled. Because of the approximations
used in the statistical model, we are inclined to favor the
energetics derived from the DFT calculation.
Figure 3. Correlation of the branching ratios of ªmixedº m-oxomanga-
nese(iv) salen complexes and s values of the 5,5'-substituents.
‡
How do our results (displaying the ªintrinsicº reactivity of
the active catalyst) compare with those for the reactivity in
solution? It was recognized at an early stage by Kochi that
[Mn(salen)] complexes with electron-donating substituents,
such as the 5,5'-dimethoxy derivative, give only poor yields of
epoxide, whereas the catalyst with 5,5'-dinitro substituents
gave the best product yields.[16] While the reactivity differ-
ences seen with differently substituted achiral salens suggest
based on the solvolysis of tert-cumyl chlorides, that is, for
reactions, in which an electron-donating group interacts with a
developing positive charge in the transition state. The good
‡
correlation with the s values in our experiments can be
rationalized by the analogy between a developing carbocation
and the increase in oxidation state of the manganese center in
the course of the fragmentation.
Figure 4. B3LYP/6 ± 311G* structures of the manganese(iii) and oxomanganese(v) salen complexes of the homodesmotic reaction given in Equation (2).
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001 0947-6539/01/0703-0594 $ 17.50+.50/0 Chem. Eur. J. 2001, 7, No. 3
594

Reactivity of Oxomanganese Salen Complexes
591±599
that the more electron-deficient ligand will give the more
effective catalyst, the interplay between epoxidation efficien-
cy and selectivity is much more subtle and less predictable for
asymmetric epoxidation. In 1991, Jacobsen reported the
dramatic effects of electronic tuning, that is, using different
substituents in the para-position to the ligating oxygens (5,5'-
disubstitution), on the enantioselectivity of the [Mn(salen)]
catalyzed epoxidation of cis-disubstituted olefins.[39] In a more
recent account, Jacobsen and co-workers systematically
investigated the correlation between enantioselectivity and
the electronic character of the Jacobsen catalyst by varying
the substituents in the 5,5'-positions (NO2, Cl, H, Me, and
OMe).[40] In all cases, electron-donating groups on the catalyst
were found to give higher enantioselectivities in epoxidation,
whereas electron-withdrawing substitutents led to decreased
enantioselectivity. The possibility that the substitutents might
induce conformational distortions in the oxygen transferring
complexes or provoke changes in the Mn O bond lengths and
thus in the substrate ± ligand interactions was dismissed as
highly improbable by these authors. Instead, the electronic
effect on the enantioselectivity was attributed to changes in
for possible stabilizing/destabilizing effects of donor ligands
ˆ
on the Mn O moiety. The following ligands were studied: p-
CN-N,N-dimethylaniline N-oxide, p-Me-N,N-dimethylaniline
N-oxide, p-Br-N,N-dimethylaniline N-oxide, N,N-dimethyl-
aniline N-oxide, acetonitrile, pyridine N-oxide, and triethyl-
phosphine oxide.
the reactivity of the MnV O moiety.
ˆ
Our gas-phase experiments provide the first direct probe
for the intrinsic properties of the reactive intermediates in the
Jacobsen ± Katsuki epoxidation. The intermolecular experi-
ment, in which the two metal centers in the ªmixedº m-oxo
complex compete for the bridging oxygen atom, provides us
with a direct measure of the oxygen-transfer capabilities of
the oxomanganese(v) species, without the need for an olefinic
substrate. We have shown that electron-donating substituents
indeed stabilize the oxomanganese(v) moiety, which should
attenuate its reactivity and give a relatively milder oxidant.
Electron-withdrawing substituents, on the other hand, desta-
bilize the oxomanganese(v) moiety and make it a more
reactive oxidant. Accordingly, the milder oxidant, which
delivers the oxygen to the substrate in a more product-like
transition state, will achieve a higher degree of stereochemical
communication between substrate and catalyst, whereas the
more reactive oxidantꢁs differentiation of the diastereomeric
transition structures will be much poorer. The analogous
trends in the reactivities seen in solution-phase and in our gas-
phase stability measurements indicate that the mechanism of
oxygen transfer to the substrate in solution is primarily
governed by the intrinsic reactivity of the oxomanganese(v)
complexes.
Scheme 3. Possible fragmentation products upon CID of m-oxomanga-
nese(iv) salen complexes with two different terminal ligands on each side.
The first axial ligand studied was pyridine N-oxide. Kochi
had already reported the use of pyridine N-oxide in man-
ganese and chromium salen catalyzed epoxidations, which
resulted in a significant enhancement of epoxide yields.[14, 16]
By carefully adjusting the stoichiometry of the in situ mix-
ture of [MnIII(salen)]ClO4, iodosobenzene, and pyridine N-
oxide as well as the spraying conditions, we succeeded in
generating the asymmetrical dinuclear m-oxo complex
[pyO(salen)Mn O Mn(salen)]2‡ (m/z: 376.5) in the gas
phase. The primary fragmentation products observed are
‡
‡
[Mn O(salen)] (m/z: 337) and [MnIIIpyO(salen)] (m/z:
416), with no traces of alternative fragmentation detectable
(Scheme 4). Fragmentation of the bridging m-oxo bond takes
place exclusively on the side where the terminal ligand is
bound.
ˆ
The relative destabilizing effects of different amine N-
oxides on the Mn oxo moiety can be seen from the
fragmentation of the m-oxomanganese(iv) complex with p-
CN-N,N-dimethylaniline N-oxide and pyridine N-oxide, re-
Effects of axial ligands: The reactivity of salen catalysts in
epoxidation reactions cannot only be tuned by substitution of
the salen, but also by adding donor ligands to the reaction
mixture.[14, 17] The question arises whether or not donor
ligands directly modulate the reactivity of the oxygen trans-
ferring species by ligation.[33, 41] The methodology employed
by us for the investigation of electronic influences on the
stability of the oxomanganese(v) complexes can be extended
to study the effects of additional axial ligands. For this
purpose, a m-oxomanganese(iv) complex with two different
terminal ligands has to be prepared, which upon fragmenta-
tion will give rise to two differently ligated MnV O species
(Scheme 3). This experiment provides us with a direct probe
ˆ
Scheme 4. Fragmentation of the asymmetrical m-oxo complex
[(salen)Mn O Mn(salen)Opy]2‡ upon CID.
Chem. Eur. J. 2001, 7, No. 3
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0703-0595 $ 17.50+.50/0
595

D. A. Plattner and D. Feichtinger
FULL PAPER
spectively, as terminal ligands (Figure 5). Fragmentation of
the Mn O bond occurs predominantly on the side where the
aniline N-oxide is bound; this means that pyridine N-oxide is
much less destabilizing and should thus be the less effective
Figure 6. Top: daughter-ion spectrum (0.13 PaAr, collision energy 20 eV)
‡
of [p-CN C6H4NMe2O MnIII(salen)] , and this shows fragmentation of
the Mn O as well as of the N O bond; bottom: daughter-ion spectrum
‡
(0.13 PaAr, collision energy 20 eV) of [MnIII(pyO)(salen)] , and this shows
exclusive loss of pyridine N-oxide.
Figure 5. Top: electrospray mass spectrum of a solution of [Mn(salen)]-
ClO4 and iodosobenzene with pyridine N-oxide and p-CN-N,N-dimethyl-
aniline N-oxide in acetonitrile; bottom: daughter-ion spectrum
(0.08 PaAr, collision energy 14 eV) of the dinuclear m-oxo complex
[L1(salen)Mn O Mn(salen)L2]2‡ (m/z: 457.5, L1 ˆ pyridine N-oxide, L2 ˆ
p-CN-N,N-dimethylaniline N-oxide), and this shows exclusive fragmen-
Table 2. Branching ratios (IL1/IL2) for the two principal pathways of the dissociation
of the ªmixedº m-oxo complex with two different terminal ligands (L2 ˆ p-
CN C6H4NMe2O). A value >1 indicates a stabilizing effect of the terminal ligand
2
ˆ
on the Mn O moiety with respect to the reference ligand L .
Ligand C6H5NMe2O p-Br C6H4NMe2O p-Me C6H4NMe2O[a] Et3PO pyO
‡
tation to [MnL2(salen)] (m/z: 483) and to the oxo complex
IL1/IL2
1.12
Æ 0.17
1.30
Æ 0.13
1.30
Æ 0.15
1.33 > 10
Æ 0.12
1
‡
ˆ
[L (salen)Mn O] (m/z: 432).
[a] Value extrapolated from
Br C6H4NMe2O because of overlap between parent and daughter peaks in the
a
competition experiment against L2 ˆ p-
promoter for oxidation catalysis. N,N-dimethylaniline N-
oxide and its derivatives are not only capable of oxygen
transfer to manganese porphyrin,[42] but also to manga-
nese salen complexes (see Figure 6 top), albeit less efficiently
than iodosobenzene.[43] Pyridine N-oxide, on the other hand,
does not transfer oxygen to Mn salen complexes, as can be
seen from the fragmentation depicted in Figure 6 (bottom).
The only loss observed is that of pyridine N-oxide (95 mass
units). The alternative fragmentation of the N O bond cannot
be detected even at collision energies of up to 50 eV.
The surprisingly big difference in destabilization between
p-CN-N,N-dimethylaniline N-oxide and pyridine N-oxide
prompted us to test N,N-dimethylaniline N-oxides with
different substituents in the para-position. As can be seen
from Table 2, the only ratio significantly deviating from 1
comes from the competition experiment between the p-CN
and the p-Br derivatives. As the site of substitution is far
removed from the site of coordination, the ratios of ꢀ1
daughter-ion spectrum of [p-Me C6H4NMe2O(salen)Mn O Mn(salen)OMe2N-p-
CN C6H4]2‡
.
between the Br-, Me-, and H-substituted aniline N-oxides are
not unexpected. A small but significant difference is observed
between p-CN-N,N-dimethylaniline N-oxides and Et3PO; the
latter also acts as a promoter in metal salen catalyzed
epoxidations.[14] Triethylphosphine oxide binds very strongly
‡
to [MnIII(salen)] and has only a slightly less destabilizing
ˆ
effect on the Mn O moiety than p-CN-N,N-dimethylaniline
ˆ
N-oxide. As a result of the very strong P O bond, however,
the phosphine oxide does not transfer oxygen to the metal
center.
The following picture emerges from our electrospray MS
data. Since we have never detected m-oxo complexes without
terminal ligands in the in situ mixtures, we conclude that the
additional axial ligands strongly promote formation of
596
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0703-0596 $ 17.50+.50/0
Chem. Eur. J. 2001, 7, No. 3

Reactivity of Oxomanganese Salen Complexes
591±599
of the chlorides in acetonitrile were treated with an equimolar amount of
silver perchlorate and stirred for fiveteen minutes. After removal of AgCl
by filtration, the solution was diluted to 10 5 m.
dinuclear complexes. This can easily be rationalized by
regarding the formation of the m-oxomanganese(iv) species
‡
ˆ
as a partial oxygen transfer from [L(salen)Mn O] to
Synthesis of para-substituted dimethylaniline N-oxides: Dimethylaniline N-
oxide, p-CN-dimethylaniline N-oxide, p-methyl-dimethylaniline N-oxide,
and p-Br-dimethylaniline N-oxide were prepared according to the proce-
dure of Cymerman Craig and Purushothaman[45] by MCPBA (m-chloro-
peroxybenzoic acid) oxidation of the respective N,N-dimethylanilines (all
bought and used as received from Fluka). All N-oxide spectra matched the
analytical data given in the literature. For the experiments with terminal
ligands, the amine N-oxide or Et3PO was used in a tenfold excess with
respect to the manganese complex. Solutions that contained m-oxomanga-
nese salen complexes were prepared by adding stock solutions (10 3 m in
CH3CN) of the terminal ligand to a suspension of iodosobenzene in
acetonitrile, and then adding the manganese(iii) salen salt just prior to
electrospray.
‡
[MnL(salen)] . As evidenced in our stability measurements,
axial ligands enhance the oxygen-transfer capability and thus
the formation of dinuclear m-oxo complexes. The conpropor-
tionation can be viewed as an escape route for the extremely
reactive oxomanganese(v) species to ªhideº in a more
persistent form. After disproportionation of the dinuclear
‡
ˆ
complex, that is, after the release of [L(salen)Mn O] , the
coordinating ligand stays at the metal center and will now
promote the reactivity of the catalyst in the epoxidation (or in
the back reaction to the m-oxo complex).
Instruments: For the mass spectrometric measurements, the slightly
modified FinniganMATTSQ7000 electrospray tandem mass spectrometer
described previously[28b] (octopole, quadrupole, octopole, quadrupole
setup) was used. The first octopole was fitted with an open cylindrical
sheath around the rods into which a collision gas could be bled for
thermalization or reaction up to 2.5 Pa.
Conclusion
Electrospray ionization tandem mass spectrometry is a
convenient approach to the study of transition metal cata-
lyzed reactions and reactive intermediates that are too elusive
for solution-phase characterization, as has been demonstrated
General ESMS setup for the experiments: All quantitative measurements
were carried out in daughter-ion mode, that is, the first quadrupole was
used to mass select ions of a single mass, which were then collided with a
target gas in the second octopole. The second quadrupole was operated in
scanning mode in order to detect the ionic fragments. The collision energy
could be varied by applying different potentials (to a lens in front of the
second octopole), which altered the velocity of the ions on their way into
the collision region (the collision energies are given in eV, lab frame). The
incoming ions were thermalized in the first octopole with argon at a
pressure of ꢀ1.3 Pa and at a temperature of 708C. The tube lens was
typically operated at 70 V (referenced to m/z: 500).
‡
ˆ
for the oxomanganese salen complex [O Mn(salen)] . Once
in the gas phase, reactivity studies of species that have so far
escaped any solution-phase characterization can be conducted
both qualitatively as well as quantitatively. As shown for
electronic and ligand effects on oxomanganese(v) salen
complexes, insight into the intrinsic reactivity of transient
species can be gained far beyond the limits of solution-phase
techniques. The similarities and the differences between the
observed gas-phase and solution-phase chemistry will then
yield the information that is necessary to dissect the complex
mechanisms of transition metal catalysis.
Determination of branching ratios
CID of complexes of the type [p-CN-C5H4NMe2O(salen)Mn O-Mn(5,5'-
R2 salen)OMe2N-p-CN-C5H4]: A set of 122 separate measurements of
[p-CN-C5H4NMe2O(salen)Mn O Mn(5,5'-(CH3)2 salen)OMe2N-p-CN-
C5H4]2‡ was taken at collision energies ranging from 4 ± 44 eV (lab frame,
collision gas Ar, that is, 0.15 ± 1.68 eV in c.o.m. frame) and from 4 ± 24 eV
(lab frame, collision gas Xe, that is, 0.46 ± 2.76 eV in c.o.m. frame),
respectively, in order to obtain information about the collision-energy
dependence and the statistical errors of the experiments. The collision gas
was at a pressure of 0.027 Pa as measured by a cold cathode gauge.
Additional measurements at 0.013 and 0.007 Pa yielded the same branching
ratios. The measurements of the remaining mixed dinuclear m-oxo
complexes were conducted at a pressure of 0.027 PaAr and 34 eV collision
energy (lab frame). Under these conditions, the total fragment yield was
about 10%, and secondary fragmentation products were only present in
negligible amounts. Spectra were added in the course of 2 ± 3 minutes at a
rate of one scan per second. Please note that the CID spectra shown in
Figures 2 and 5 were recorded at significantly higher gas pressures (0.07 ±
0.13 PaAr) in order to represent all different parent and fragment peaks in
an adequate size.
Experimental Section
Materials: Diacetoxyiodobenzene, pyridine N-oxide, 5-chloro-2-hydroxy-
benzaldehyde, 5-methyl-2-hydroxybenzaldehyde, 5-methoxy-2-hydroxy-
benzaldehyde, and 5-nitro-2-hydroxybenzaldehyde were purchased from
Aldrich and Fluka, respectively, while 5-fluoro-2-hydroxybenzaldehyde
was obtained from Melford Laboratories (UK). Triethylphosphine oxide
was obtained from Strem. Acetonitrile (HPLC grade quality) for the
preparation of the electrospray solutions was purchased from Fluka.
Synthesis of salen ligands: Salen (N,N'-bis(salicylidene)ethylene diamine)
was bought from Fluka and used as received. The 5,5'-substituted salen
ligands were synthesized from the respective aldehydes and ethylene
diamine according to the procedure given by Jacobsen et al.[44] All spectra
matched the analytical data given in the literature. The 5,5'-difluorosalen
ligand has so far not been described in the literature.
For the calculation of the branching ratio, the ratio between the inten-
sities of the manganese(iii) fragment masses corresponding to
‡
[p-CN C5H4NMe2O Mn(5,5'-R2 salen)] and [p-CN C5H4NMe2O Mn-
‡
(salen)] was chosen. To account for the secondary fragmentation products
N,N'-Bis(5,5'-difluorosalicylidene)ethylene diamine (5,5'-Difluorosalen):
Yield 50%. M.p. 205 ± 206.58C; 1H NMR (200 MHz, CDCl3, 258C,
‡
‡
[Mn(5,5'-R2 salen)] and [Mn(salen)] derived from the loss of the
terminal ligand, the total ion yield for each channel was calculated by
adding the intensity of the secondary fragment to its respective parent. The
branching ratio of every experiment was assigned a statistical weighting
factor proportional to the total fragment intensities in order to calculate
mean values.
ˆ
TMS): d ˆ 8.32 (s, 2H; ArCH N), 6.89 ± 7.05 (m, 6H; Ar), 3.98 (4H;
NCH2CH2N); IR (KBr pellet): nÄ ˆ 3447(w), 2939(w), 2909(w), 2853(w),
1634(s), 1498(s), 1364(m), 1140(m), 831(m), 702(s) cm 1; elemental analysis
calcd (%) for C16H14N2O2F2 (304.3): C 63.15, H 4.64, N 9.21; found C 63.09,
H 4.77, N 9.01.
Synthesis of manganese(iii) salen complexes: The manganese(iii) salen
complexes were prepared either following the procedure of Kochi et al.[16]
([Mn(5,5'-dinitrosalen)]PF6 and [Mn(5,5'-difluorosalen)]PF6) or the pro-
cedure of Jacobsen et al.[44] [(Mn(salen)]Cl, [Mn(5,5'-dimethoxysalen)]Cl,
[Mn(5,5'-dimethylsalen)]Cl, and [Mn(5,5'-dichlorosalen)]Cl). In the case of
manganese(iii)PF6 salts, the electrospray solutions were prepared by
dissolving the salt in CH3CN and diluting to 10 5 m concentration. Solutions
CID of complexes of the type [L1(salen)Mn O Mn(salen)L2]2‡
: The
‡
branching ratio was calculated from the intensities of [Mn(salen)L1] and
‡
[Mn(salen)L2] . In contrast to the experiments with different salen ligands
on each side of the m-oxo complex, the ratio could not be corrected for
secondary fragmentation products in this case, as loss of the terminal ligand
‡
led to the same product [Mn(salen)] for both channels. The experimental
‡
conditions were chosen such that the [Mn(salen)] peak intensity never
Chem. Eur. J. 2001, 7, No. 3
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0703-0597 $ 17.50+.50/0
597

D. A. Plattner and D. Feichtinger
FULL PAPER
exceeded 5% of the primary fragment intensities. In the case of acetonitrile
or pyridine N-oxide, there was an additional reaction channel, and this
yielded an asymmetrical [(salen)Mn O Mn(salen)L]2‡ fragment by dis-
sociation of one of the terminal ligands. Those species could also be
obtained in high enough intensities to conduct CID experiments by
applying higher tube lens potentials (ꢀ90 V). The CID experiment on
[(salen)Mn O Mn(salen)NCCH3]2‡ turned out to be inconclusive, since
the only fragment mass which can be unambiguously assigned to a specific
[21] C. T. Dalton, K. M. Ryan, V. M. Wall, C. Bousquet, D. G. Gilheany,
Top. Catal. 1998, 5, 75 ± 91.
[22] J. T. Groves, W. J. Kruper, J. Am. Chem. Soc. 1979, 101, 7613 ± 7615.
[23] M. Yamashita, J. B. Fenn, J. Phys. Chem. 1984, 88, 4451 ± 4459.
[24] The basic electrospray source for mass spectrometry is described by:
C. M. Whitehouse, R. N. Dreyer, M. Yamashita, J. B. Fenn, Anal.
Chem. 1985, 57, 675 ± 679.
[25] A complete monograph on the technique is: Electrospray Ionization
Mass Spectrometry (Ed.: R. D. Cole), Wiley, New York, 1997.
[26] a) R. Colton, J. C. Traeger, Inorg. Chim. Acta 1992, 201, 153 ± 155;
b) R. Colton, A. DꢁAgostino, J. C. Traeger, Mass Spectrom. Rev. 1995,
14, 79 ± 106, and references cited therein.
[27] a) V. Katta, S. K. Chowdhury, B. T. Chait, J. Am. Chem. Soc. 1990, 112,
5348 ± 5349; b) S. R. Wilson, Y. H. Wu, Organometallics 1993, 12,
1478 ± 1480; c) A. O. Aliprantis, J. W. Canary, J. Am. Chem. Soc. 1994,
116, 6985 ± 6986; d) L. A. P. Kane-Maguire, R. Kanitz, M. M. Sheil, J.
Organomet. Chem. 1995, 486, 243 ± 248; e) C. Q. Jiao, B. S. Freiser,
S. R. Carr, C. J. Cassidy, J. Am. Chem. Soc. Mass Spectrom. 1995, 6,
521 ± 524; f) K. S. Wang, X. L. Han, R. W. Gross, G. W. Gokel, J. Am.
Chem. Soc. 1995, 117, 7680 ± 7686; g) L. Ripa, A. Hallberg, J. Org.
Chem. 1996, 61, 7147 ± 7155; h) Z. M. Dzhabieva, V. P. Kozlovskii,
Yu. M. Shulꢁga, A. F. Dodonov, G. P. Belov, Russ. Chem. Bull. 1996,
45, 474 ± 476; i) J. M. Brown, K. K. Hii, Angew. Chem. 1996, 108, 679 ±
682; Angew. Chem. Int. Ed. Engl. 1996, 35, 657 ± 659; j) K. K. Hii,
T. D. W. Claridge, J. M. Brown, Angew. Chem. 1997, 109, 1033 ± 1036;
Angew. Chem. Int. Ed. Engl. 1997, 36, 984 ± 987; k) B. H. Lipshutz,
K. L. Stevens, B. Lames, J. G. Pavlovich, J. P. Snyder, J. Am. Chem.
Soc. 1996, 118, 6796 ± 6797; l) R. Saf, R. Schitter, C. Mirtl, F. Stelzer, K.
Hummel, Macromolecules 1996, 29, 7651 ± 7656; m) K. A. Hirsch,
S. R. Wilson, J. S. Moore, J. Am. Chem. Soc. 1997, 119, 10401 ± 10412;
n) M. S. Stephan, A. J. J. M. Teunissen, G. K. M. Verzijl, J. G. de Vries,
Angew. Chem. 1998, 110, 688 ± 690; Angew. Chem. Int. Ed. 1998, 37,
662 ± 664.
‡
reaction channel corresponded to [Mn(salen)(NCCH3)] , which was only a
minor peak in the spectrum.
Computational methods: Quantum chemical calculations for the com-
pounds shown in Figure 4 have been performed using the Gaussian94 series
of programs[46] on DECAlpha 8400 computers at ETH Zürich. The quintet
and triplet ground states of the manganese(iii) and manganese(v) com-
plexes, respectively, were fully optimized without symmetry constraints
using the hybrid Becke3Lee ± Yang ± Parr (B3LYP) exchange correlation
functional.[47] A 6 ± 311G* valence triple zeta ‡ polarization basis set was
‡
ˆ
used. The structure and energy of [O Mn(salen)] have been reported in a
previous account.[31]
Acknowledgements
The authors thank Prof. Peter Chen for financial support and his
continuous interest in our work and Martin Jufer for his assistance in
some of the experiments. We would also like to thank the Competence
Center for Computational Chemistry at ETH Zürich for generous
allocation of computing resources.
[1] W. J. Mijs, C. R. H. I. De Jonge, Organic Syntheses by Oxidation with
Metal Compounds, Plenum, New York, 1986.
[2] C. C. Reddy, G. A. Hamilton, K. M. Madyashta, Biological Oxidation
Systems, Academic Press, San Diego, 1990.
[3] R. Sato, T. Omura, Cytochrome P-450, Kodansha, Tokyo, 1978.
[4] R. E. White, M. J. Coon, Annu. Rev. Biochem. 1980, 49, 315 ± 356.
[5] T. G. Traylor, P. S. Traylor in Active Oxygen in Biochemistry (Eds.: J. S.
Valentine, C. S. Foote, J. F. Liebman, A. Greenberg), Blackie Aca-
demic, London, 1995, pp. 84 ± 187.
[28] a) C. Hinderling, D. A. Plattner, P. Chen, Angew. Chem. 1997, 109,
272 ± 274; Angew. Chem. Int. Ed. Engl. 1997, 36, 243 ± 244; b) C.
Hinderling, D. Feichtinger, D. A. Plattner, P. Chen, J. Am. Chem. Soc.
1997, 119, 10793 ± 10804.
[29] D. Feichtinger, D. A. Plattner, P. Chen, J. Am. Chem. Soc. 1998, 120,
7125 ± 7126.
[30] D. Feichtinger, D. A. Plattner, Angew. Chem. 1997, 109, 1796 ± 1798;
Angew. Chem. Int. Ed. Engl. 1997, 36, 1718 ± 1719.
[31] D. A. Plattner, D. Feichtinger, J. El-Bahroui, O. Wiest, Int. J. Mass
Spectrom. 2000, 195/196, 351 ± 362.
[6] Cytochrome P-450, Structure, Mechanism and Biochemistry (Ed.: P. R.
Ortiz de Montellano), Plenum, New York, 1995.
[32] D. Feichtinger, D. A. Plattner, J. Chem. Soc. Perkin Trans. 2 2000,
1023 ± 1028.
[33] The use of p-cyanodimethylaniline N-oxide as an oxidant for
manganese(iii) porphyrin complexes has been pioneered by Bruice
et al. M. F. Powell, E. F. Pai, T. C. Bruice, J. Am. Chem. Soc. 1984, 106,
3277 ± 3285.
[34] a) K. M. Ervin, P. B. Armentrout, J. Chem. Phys. 1985, 83, 166 ± 189;
b) S. K. Loh, D. A. Hales, L. Lian, P. B. Armentrout, J. Chem. Phys.
1989, 90, 5466 ± 5485; c) R. H. Schultz, P. B. Armentrout, Int. J. Mass
Spectrom. Ion Processes 1991, 107, 29 ± 48; d) R. H. Schultz, K. C.
Crellin, P. B. Armentrout, J. Am. Chem. Soc. 1991, 113, 8590 ± 8601;
e) N. F. Dalleska, K. Honma, L. S. Sunderlin, P. B. Armentrout, J. Am.
Chem. Soc. 1994, 116, 3519 ± 3528; f) M. T. Rodgers, K. M. Ervin, P. B.
Armentrout, J. Chem. Phys. 1997, 106, 4499 ± 4508.
[7] T. Mann in Oxidases and Related Systems (Eds.: T. E. King, H. S.
Mason, M. Morrison), Liss, New York, 1988, pp. 29 ± 49.
[8] Y.-C. Fann, N. C. Gerber, P. A. Osmulski, L. P. Hager, S. G. Sligar,
B. M. Hoffmann, J. Am. Chem. Soc. 1994, 116, 5989 ± 5990.
[9] a) K. McMillan, D. S. Bredt, D. J. Hirsch, S. H. Snyder, J. E. Clark,
B. S. S. Masters, Proc. Natl. Acad. Sci. USA 1992, 89, 1141 ± 1145;
b) K. A. White, M. A. Marletta, Biochemistry 1992, 31, 6627 ± 6631;
c) M. A. Marletta, J. Biol. Chem. 1993, 268, 12231 ± 12234.
[10] J. T. Groves, W. J. Kruper, T. E. Nemo, R. S. Myers, J. Mol. Catal. 1980,
7, 169 ± 177.
[11] Metalloporphyrins in Catalytic Oxidations (Ed.: R. A. Sheldon),
Dekker, New York, 1994.
[12] Metalloporphyrin Catalyzed Oxidations (Eds.: F. Montanari, L. Case-
lla), Kluwer, Dordrecht, 1994.
[35] A similar method that translates kinetic branching ratios into
thermochemical data in cases when the difference in entropy change
between two different fragmentation pathways can be neglected was
introduced by Cooks. The assumptions inherent in the ªCooksꢁ kinetic
methodº were essentially the same as those required for linear free
energy relationships, that is, Hammett s1 correlations. The kinetic
method has been widely used for the determination of proton
affinities and gas-phase basicities of many organic compounds, and
this has yielded values in excellent agreement with independent
methods. a) R. G. Cooks, J. S. Patrick, T. Kotiaho, S. A. McLuckey,
Mass Spectrom. Rev. 1994, 13, 287 ± 339; b) R. G. Cooks, P. S. H.
Wong, Acc. Chem. Res. 1998, 31, 379 ± 386, and references cited
therein.
[13] T. L. Siddall, N. Miyaura, J. C. Huffman, J. K. Kochi, J. Chem. Soc.
Chem. Commun. 1983, 1185 ± 1186.
[14] E. G. Samsel, K. Srinivasan, J. K. Kochi, J. Am. Chem. Soc. 1985, 107,
7606 ± 7617.
[15] K. Srinivasan, J. K. Kochi, Inorg. Chem. 1985, 24, 4671 ± 4679.
[16] K. Srinivasan, P. Michaud, J. K. Kochi, J. Am. Chem. Soc. 1986, 108,
2309 ± 2320.
[17] K. Srinivasan, S. Perrier, J. K. Kochi, J. Mol. Catal. 1986, 36, 297 ± 317.
[18] W. Zhang, J. L. Loebach, S. R. Wilson, E. N. Jacobsen, J. Am. Chem.
Soc. 1990, 112, 2801 ± 2803.
[19] R. Irie, K. Noda, Y. Ito, N. Matsumoto, T. Katsuki, Tetrahedron Lett.
1990, 31, 7345 ± 7348.
[20] E. N. Jacobsen in Comprehensive Organometallic Chemistry II, Vol. 12
(Eds.: G. W. Wilkinson, F. G. A. Stone, E. W. Abel, L. S. Hegedus),
Pergamon, New York, 1995, Chapter 11.1.
[36] a) W. M. Coleman, R. R. Goehring, L. T. Taylor, J. G. Mason, R. K.
Boggess, J. Am. Chem. Soc. 1979, 101, 2311 ± 2315; b) W. M. Coleman,
598
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0703-0598 $ 17.50+.50/0
Chem. Eur. J. 2001, 7, No. 3

Reactivity of Oxomanganese Salen Complexes
591±599
R. K. Boggess, J. W. Hughes, L. T. Taylor, Inorg. Chem. 1981, 20, 700 ±
706; c) W. M. Coleman, R. K. Boggess, J. W. Hughes, L. T. Taylor,
Inorg. Chem. 1981, 20, 1253 ± 1258.
[43] In our ESMS experiments, we have always used a combination of
amine N-oxides and iodosobenzene in order to get higher yields of the
m-oxomanganese(iv) complexes and better reproducibility of the
oxidized species than would be possible with N-oxides as the only
oxygen source.
[44] W. Zhang, E. N. Jacobsen, J. Org. Chem. 1991, 56, 2296 ± 2298.
[45] J. Cymerman Craig, K. K. Purushothaman, J. Org. Chem. 1970, 35,
1721 ± 1722.
[46] M. J. Frisch, G. W. Trucks, H. B. Schlegel, P. M. W. Gill, B. G. Johnson,
M. A. Robb, J. R. Cheeseman, T. Keith, G. A. Petersson, J. A.
Montgomery, K. Raghavachari, M. A. Al-Laham, V. G. Zakrzewski,
J. V. Ortiz, J. B. Foresman, J. Cioslowski, B. B. Stefanov, A. Nanayak-
kara, M. Challacombe, C. Y. Peng, P. Y. Ayala, W. Chen, M. W. Wong,
J. L. Andres, E. S. Replogle, R. Gomperts, R. L. Martin, D. J. Fox, J. S.
Binkley, D. J. Defrees, J. Baker, J. P. Stewart, M. Head-Gordon, C.
Gonzalez, J. A. Pople, Gaussian 94, Revision C.3, Gaussian, Pitts-
burgh PA, 1995.
[37] H. C. Brown, Y. Okamoto, J. Am. Chem. Soc. 1958, 80, 4979 ± 4987.
[38] The ground state of the manganese(iii) complex is quintet. As for the
ground state of the manganese(v) oxo complex, calculations on salen
model systems point to triplet ground states: a) T. Strassner, K. N.
Houk, Org. Lett. 1999, 1, 419 ± 421; b) C. Linde, B. Škermark, P.-O.
Norrby, M. Svensson, J. Am. Chem. Soc. 1999, 121, 5083 ± 5084; c) L.
Cavallo, H. Jacobsen, Angew. Chem. 2000, 112, 602 ± 604; Angew.
Chem. Int. Ed. 2000, 39, 589 ± 592. Preliminary calculations in our
group on the singlet states of oxomanganese(v) salen show them very
close in energy to the triplets. In view of the above cited model studies,
we decided to use the triplet energies for the homodesmotic reaction.
[39] E. N. Jacobsen, W. Zhang, M. L. Güler, J. Am. Chem. Soc. 1991, 113,
6703 ± 6704.
[40] M. Palucki, N. S. Finney, P. J. Pospisil, M. L. Güler, T. Ishida, E. N.
Jacobsen, J. Am. Chem. Soc. 1998, 120, 948 ± 954.
[41] N. S. Finney, P. J. Pospisil, S. Chang, M. Palucki, R. G. Konsler, K. B.
Hansen, E. N. Jacobsen, Angew. Chem. 1997, 109, 1798 ± 1801; Angew.
Chem. Int. Ed. Engl. 1997, 36, 1720 ± 1723.
[47] a) A. D. Becke, Phys. Rev. A 1988, 38, 3098 ± 3100; b) C. Lee, W. Yang,
R. G. Parr, Phys. Rev. B 1988, 37, 785 ± 789; c) P. J. Stevens, F. J. Devlin,
C. F. Chablowski, M. J. Frisch, J. Phys. Chem. 1994, 98, 11623 ± 11627.
Received: March 20, 2000
[42] D. Feichtinger, D. A. Plattner, unpublished results.
Revised version: August 7, 2000 [F2372]
Chem. Eur. J. 2001, 7, No. 3
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0703-0599 $ 17.50+.50/0
599