Are MnIV species involved in Mn(salen)-satalyzed Jacobsen-Katsuki epoxidations? A mechanistic elucidation of their formation and reaction modes by EPR spectroscopy, mass-spectral analysis, and product studies: Chlorination versus oxygen transfer

Article abstract of DOI:10.1021/ja0005082

EPR and ESI-MS/MS evidence is presented that in the absence of an olefinic substrate the reaction between the Mn^III(salen) complexes A1 (X = Cl) and A2 (X = PF₆) and PhIO or NaOCl as oxygen sources leads to paramagnetic Mn^IV(salen) complexes. Depending on the solvent and the counterion, two distinct Mn^IV-(salen) complexes intervene. In CH₂Cl₂, regardless of the counterion, a ClOMn^IV(salen) complex (B1) and a HOMn^IV(salen) complex (B1′) are formed by Cl and H atom abstraction from CH₂Cl₂, and the latter deprotonates to the neutral OMn^IV(salen) complex (B2). In EtOAc as solvent, only the complex B2 is obtained from A1 (X = Cl), presumably by inner-sphere electron transfer from the chloride ion. The Mn^IV(salen) complexes display the following reaction modes toward 1,2-dihydronaphthalene (1), styrene (2), and the radical probe 3 as substrates: Complex B1 chlorinates the olefins 1/2 through an electrophilic pathway to yield the 1,2-dichloro adducts 1a/2a and the chlorohydrins 1b/2b (nucleophilic trapping of the initially formed benzylic cation), while with olefin 3 the ring-opened dichloro product 3a results. Complex B2, however, epoxidizes these olefins through a radical pathway, as evidenced by the formation of isomerized stilbene oxide 4c (cis/trans ratio 36: 64) from cis-stilbene (4). The relevance of these paramagnetic Mn^IV(salen) species in Jacobsen-Katsuki catalytic epoxidations is scrutinized.

Full text of DOI:10.1021/ja0005082

J. Am. Chem. Soc. 2000, 122, 9685-9691
9685
Are MnIV Species Involved in Mn(Salen)-Catalyzed
Jacobsen-Katsuki Epoxidations? A Mechanistic Elucidation of Their
Formation and Reaction Modes by EPR Spectroscopy, Mass-Spectral
Analysis, and Product Studies: Chlorination versus Oxygen Transfer
Waldemar Adam,‡ Cordula Mock-Knoblauch,*,‡ Chantu R. Saha-Mo1ller,‡ and
M. Herderich#
Contribution from the Institut fu¨r Organische Chemie, UniVersita¨t Wu¨rzburg, Am Hubland,
D-97074 Wu¨rzburg, Germany, and Institut fu¨r Pharmazie und Lebensmittelchemie, UniVersita¨t Wu¨rzburg,
Am Hubland, D-97074 Wu¨rzburg, Germany
ReceiVed February 10, 2000
Abstract: EPR and ESI-MS/MS evidence is presented that in the absence of an olefinic substrate the reaction
between the MnIII(salen) complexes A1 (X ) Cl) and A2 (X ) PF6) and PhIO or NaOCl as oxygen sources
leads to paramagnetic MnIV(salen) complexes. Depending on the solvent and the counterion, two distinct MnIV-
(salen) complexes intervene. In CH2Cl2, regardless of the counterion, a ClOMnIV(salen) complex (B1) and a
HOMnIV(salen) complex (B1′) are formed by Cl and H atom abstraction from CH2Cl2, and the latter deprotonates
to the neutral OMnIV(salen) complex (B2). In EtOAc as solvent, only the complex B2 is obtained from A1 (X
) Cl), presumably by inner-sphere electron transfer from the chloride ion. The MnIV(salen) complexes display
the following reaction modes toward 1,2-dihydronaphthalene (1), styrene (2), and the radical probe 3 as
substrates: Complex B1 chlorinates the olefins 1/2 through an electrophilic pathway to yield the 1,2-dichloro
adducts 1a/2a and the chlorohydrins 1b/2b (nucleophilic trapping of the initially formed benzylic cation),
while with olefin 3 the ring-opened dichloro product 3a results. Complex B2, however, epoxidizes these olefins
through a radical pathway, as evidenced by the formation of isomerized stilbene oxide 4c (cis/trans ratio 36:
64) from cis-stilbene (4). The relevance of these paramagnetic MnIV(salen) species in Jacobsen-Katsuki catalytic
epoxidations is scrutinized.
Introduction
analogy to OMnIV(porph),5 was ascribed radical-type oxidations
to account for the formation of ring-opened products in the MnIII-
(salen)-catalyzed epoxidation of a radical probe.6
The first optically active MnIII(salen) complexes have been
used independently by Jacobsen1 and Katsuki2 as versatile
catalysts for the enantioselective epoxidation of unfunctionalized
olefins. The active species in these reactions is considered to
be an OMnV(salen) complex, as proposed by Kochi in his
original work on achiral derivatives.3 Indeed, Plattner4 showed
by means of an ESI-MS study that the addition of the achiral
MnIII(salen) complex to a dispersion of PhIO in CH3CN afforded
the OMn(salen) species, together with various adducts and
dimerization products thereof. Recent theoretical work con-
cluded a triplet ground state for the OMnV(salen) complex.7 In
addition to OMnV(salen), Norrby and Akermark proposed also
the intervention of an OMnIV(salen) complex to which, in
The mechanistic importance of the participation of an OMnIV-
(salen) complex in OMnV(salen)-type oxidations, in particular
the Jacobsen-Katsuki epoxidation, was pointed out by Linde
et al.7 It was stressed that small amounts of byproduct may be
the reason for nonlinear Eyring plots, which in the current
mechanistic dispute are taken as evidence for the involvement
of a metallaoxetane intermediate,8 as they generally implicate
that a second species intervenes.9
We present herein experimental evidence that in the absence
of an olefinic substrate, the reaction between MnIII(salen)
complexes A and an oxygen source (PhIO, NaOCl) leads to
‡ Institut fu¨r Organische Chemie.
# Institut fu¨r Pharmazie und Lebensmittelchemie.
(1) Jacobsen, E. N. In Catalytic Asymmetric Synthesis; Ojima, I., Ed.;
VCH Publishers: New York, 1993; pp 159-202.
(2) Katsuki, T. J. Mol. Catal. A: Chem. 1996, 113, 87-107.
(3) Srinivasan, K.; Michaud, P.; Kochi, J. K. J. Am. Chem. Soc. 1986,
108, 2309-2320.
(4) Feichtinger, D.; Plattner, D. A. Angew. Chem., Int. Ed. Engl. 1997,
36, 1718-1719.
(5) (a) Groves, J. T.; Stern, M. K. J. Am. Chem. Soc. 1987, 109, 3812-
3814. (b) Groves, J. T.; Stern, M. K. J. Am. Chem. Soc. 1988, 110, 8628-
8638. (c) Lee, R. W.; Nakagaki, P. C.; Bruice, T. C. J. Am. Chem. Soc.
1989, 111, 1368-1372. (d) Arasasingham, R. D.; He, G.-X.; Bruice, T. C.
J. Am. Chem. Soc. 1993, 115, 7985-7991. (e) Groves, J. T.; Lee, J.; Marla,
S. S. J. Am. Chem. Soc. 1997, 119, 6269-6273.
MnIV(salen) oxidants. Depending on the solvent (CH2Cl2 versus
(6) Linde, C.; Arnold, M.; Norrby, P.-O.; A° kermark, B. Angew. Chem.,
Int. Ed. Engl. 1997, 36, 1723-1725.
(8) Hamada, T.; Fukuda, T.; Imanishi, H.; Katsuki, T. Tetrahedron 1996,
52, 515-530.
(9) Buschmann, H.; Scharf, H.-D.; Hoffmann, N.; Esser, P. Angew.
Chem., Int. Ed. Engl. 1991, 30, 477-515.
(7) Linde, C.; A° kermark, B.; Norrby, P.-O.; Svensson, M. Angew. Chem.,
Int. Ed. Engl. 1999, 121, 5083-5084.
10.1021/ja0005082 CCC: $19.00 © 2000 American Chemical Society
Published on Web 09/26/2000

9686 J. Am. Chem. Soc., Vol. 122, No. 40, 2000
Adam et al.
Figure 1. Time profiles for the EPR spectra of the MnIV(salen)
complexes B in CH2Cl2 at 77 K (frequency 9.43 GHz, power 0.5 mW,
time constant 40.96 ms, receiver gain 1.6 × 105, modulation amplitude
10.4 G) generated from the MnIII(salen)Cl complex A1 + PhIO (5
equiv): from 30 min to 20 h and after 35 h reaction time (lowest
spectrum).
Figure 2. EPR spectrum of MnIV(salen) complexes B in CH2Cl2 at 77
K (frequency 9.43 GHz, power 0.5 mW, time constant 40.96 ms,
receiver gain 1.6 × 105, modulation amplitude 10.4 G) generated from
MnIII(salen)Cl complex A1 + PPNO (1.1 equiv) + NaOCl solution (2
M, pH 11.3, 10 equiv) after 60 min; experimental (bold spectrum) and
computer-simulated (light spectrum) spectra by assuming 80% of MnIV-
(salen) complex with D ) 0.950 and E/D ) 0.296 and 20% with D )
4.000 and E/D ) 0.160. We thank Prof. K. Wieghart (Max-Planck-
Institut, Mu¨lheim) for allowing us to measure this EPR spectrum in
his laboratory.
EtOAc) and the counterion (Cl- versus PF6-), two different
MnIV(salen) complexes are involved, one chlorinates [ClOMnIV-
(salen) as electrophilic chlorinating agent] and the other epoxi-
dizes [OMnIV(salen) as radical-type oxidizing agent] the olefinic
substrates 1-4.
Results and Discussion
signal at g ca. 8 was observed, which may be assigned to an
even-integer-spin signal of two coupled Mn atoms, e.g. (salen)-
MnOMn(salen).
Formation and Characterization of MnIV(salen) Com-
plexes. We conducted our studies with two different MnIII(salen)
complexes A, of which A1 (X ) Cl) is the complex usually
referred to as “Jacobsen’s catalyst” and is commercially
available. The complex A2 (X ) PF6) was necessary, as will
become apparent later on, to assess the origin of the chlorinated
products.
Computer simulation12 of the spectra (Figure 2) revealed two
MnIV species B with different D and E zero-field-splitting (zfs)
parameters (the first with D ) 4.000 and E/D ) 0.160, the
second with D ) 0.950 and E/D ) 0.296). Their relative
amounts change with time, i.e., after 15 min; the latter makes
up 62% of the MnIV concentration, which increases to 80% after
60 min.
The method of choice for the detection of the paramagnetic
MnIV(salen) complexes B is EPR spectroscopy, especially since
MnIV complexes give well-characterized EPR signals,10 whereas
MnV and MnIII complexes are EPR-inactive at the usual X-band
frequencies.11 In a typical experiment, a 30 mM solution of
MnIII(salen) complex A1 in CH2Cl2 was prepared and 5 equiv
of PhIO were added. The EPR spectra of this solution at 70 K,
taken after 30, 105, 265, 440, and 755 min and as late as 35 h,
showed the characteristic features of a high-spin (d3 electronic
configuration), monomeric MnIV complex, which changed
slightly with time (Figure 1). Two distinct signals were observed,
one at g ca. 2 and the other at ca. 5, of which the latter showed
a six-line hyperfine splitting with a coupling constant of A )
To quantify the amount of MnIII(salen) catalyst A1 that had
been oxidized to the MnIV(salen) complexes B, the areas under
the EPR signals were determined relative to a CuSO4-standard-
ized solution. With PhIO as oxidant, about 70% of the MnIII-
(salen) complex A1 was oxidized to the MnIV(salen) complexes
B and its EPR signal persisted for at least 35 h (Figure 3). For
NaOCl, however, at first as much as 93% of the Mn(salen)
catalyst was present as the MnIV(salen) complex B, but within
120 min it decreased to only 33% (Figure 3). Thus, all further
experiments were conducted only with PhIO as oxidant.
Treatment of complex A1 with PhIO in other solvents
common for epoxidation, such as CH2Br2, EtOAc, CH3CN, and
MeOH, also gave rise to MnIV(salen) complexes B (Figure 4).
The same was true for the addition of PhIO to a solution of A2
(PF6- as counterion instead of Cl-) in CH2Cl2. For the complex
A2 in EtOAc, however, no EPR signal could be detected upon
addition of PhIO (Figure 5).
70 G, as expected for a I ) /2 spin system.10
5
The same experiment was conducted with 10 equiv of a 2 M
NaOCl solution (pH 11.3) (Figure 2). Again, the relative
intensity and shape of the EPR signals at g ca. 2 and ca. 5 were
the same as for the PhIO oxidant and also changed slightly with
time. In this case, however, the addition of 1.1 equiv of
p-phenylpyridine N-oxide (PPNO) was necessary to form a
monomeric MnIV(salen) complex; without PPNO, only a strong
(12) The solution spectra were simulated on the basis of a spin-
Hamiltonian description of the electronic ground state: He ) D[Sz2 - S(S
+ 1)/3 + (E/D)(Sx2 - Sy2)] + µBBgjS, where S ) 1 is the spin of the coupled
system and D and E/D are the axial and rhombic zero-field parameters.
These simulations were performed with a program that was developed by
Dr. Eckhard Bill (Max-Planck-Institut fu¨r Strahlenchemie, Mu¨lheim) from
the S ) 5/2 routines of Gaffney and Silverstone (Gaffney, B. J.; Silverstone,
H. J. In Biological Magnetic Resonance; Berliner, L. J., Reuben, J., Eds.;
Plenum Press: New York, London, 1993; Vol. 13); the calculation of the
transition fields based on a Newton-Rapshon iterative method was used
as described therein.
(10) (a) Camenzind, M. J.; Hollander, F. J.; Hill, C. L. Inorg. Chem.
1983, 22, 3776-3784. (b) Burriel, R.; O’Connor, C. J.; Carlin, R. L. Inorg.
Chem. 1985, 24, 3706-3708. (c) Kessissoglou, D. P.; Li, X.; Butler, W.
M.; Pecoraro, V. L. Inorg. Chem. 1987, 26, 2487-2492. (d) Chandra, S.
K.; Basu, P.; Ray, D.; Pal, S.; Chakravorty, A. Inorg. Chem. 1990, 29,
2423-2428.
(11) McGarvey, B. R. In Transition Metal Chemistry; Carlin, R. L., Ed.;
Marcel Dekker: New York, 1966; Vol. 3, pp 89-201.

Mn(Salen)-Catalyzed Jacobsen-Katsuki Epoxidation
J. Am. Chem. Soc., Vol. 122, No. 40, 2000 9687
Figure 5. EPR spectra of MnIV(salen) complexes B in CH2Cl2 or
EtOAc at 77 K (frequency 9.43 GHz, power 2 mW, time constant 40.96
ms, receiver gain 1.6 × 105, modulation amplitude 10.4 G) generated
from MnIII(salen) complexes A1 or A2 + PPNO (2 equiv) + PhIO (1
equiv).
Figure 3. Time dependence of the concentration of MnIV(salen)
complexes B (in percent of the total Mn concentration) obtained from
the areas of the EPR spectra relative to CuSO4 as external standard:
(b) generated from MnIII(salen)Cl complex A1 + PhIO (5 equiv) and
(O) from MnIII(salen)Cl complex A1 + PPNO (1.1 equiv) + NaOCl
solution (2 M, pH 11.3, 10 equiv).
Figure 6. Electrospray mass spectrum of the MnIV(salen) complexes
B [obtained directly from the reaction mixture after 5 min of MnIII-
(salen) complex A1 addition to a slurry of PhIO in CH2Cl2].
Figure 4. EPR spectra of MnIV(salen) complexes B in different solvents
at 77 K (frequency 9.43 GHz, power 2 mW, time constant 40.96 ms,
receiver gain 1.6 × 105, modulation amplitude 10.4 G) generated from
MnIII(salen) complex A1 + PPNO (2 equiv) + PhIO (1 equiv).
upon collisional activation with 35 eV. These ions are, thus,
interpreted as the PhI adduct of complex B1. The molecular
ions m/z 634.3 (35%) and m/z 636.3 were identified as the
[ClMnIV(salen)]+ species on the basis of the Cl isotopic pattern
and product ions formed by neutral loss of HCl (36 and 38 amu).
As additional proof for the assignment of the molecular ions
m/z 650.3 and m/z 634.3, the neutral loss of H35Cl (36 amu)
and H37Cl (38 amu) was monitored. Indeed, the molecular ions
m/z 650.3, 652.3 and m/z 634.3, 636.3 yielded the productions
m/z 614.3 and m/z 598.3 with a relative intensity of 3:1, as
expected for the Cl isotope distribution, which confirms the
assignment of the peaks.
In the case of complex A2, except for the molecular ion m/z
634.3, the same ions were observed as with complex A1. In
addition, PhIO adducts of these ions were also detected and
the initially formed [OMnV(salen)]+ ion (Scheme 1) was evident
at m/z 615.3.
To detect neutral complexes, a few drops of a trifluoroacetic
acid (TFA) solution in CH2Cl2 were added to the mixture to
protonate the neutral complexes. A new molecular ion at m/z
712.3 appeared, which was assigned to the [F3CCOOMnIV-
(salen)]+ cation, formed from the neutral complex B2.
A dark green solid precipitated from the CH2Cl2 solution of
the complexes B upon additon of hexane, which on dissolution
in CH2Cl2 showed the characteristic MnIV EPR signal observed
previously (Figure 1). A chlorine analysis of the solid showed
a content of 9.38% Cl for the material derived from complex
Further characterization of the MnIV(salen) complexes B was
achieved by electrospray-ionization tandem mass spectrometry
(ESI-MS/MS), which is the method of choice for the analysis
of fragile ionic paramagnetic compounds.13 Upon addition of a
CH2Cl2 solution of complex A1 to a suspension of PhIO in CH2-
Cl2, the molecular ion m/z 599.3 [MnIII(salen)]+ decreased
considerably (to 30% from 100% in the spectrum of the starting
material A1 acquired for comparison); instead, the molecular
ions m/z 616.3 (85%), 634.3 (35%), 650.3 (40%), and 854.3
(100%) dominanted (Figure 6). The assignment of these rests
on the product ions, obtained upon low-energy (25 or 35 eV),
collision-induced dissociation (CID). CID of molecular ion m/z
616.3 (complex B1′, Scheme 1) yields product ion m/z 598.3,
due to the neutral loss of H2O (18 amu), and is, thus, identified
as the [HOMnIV(salen)]+ ion. The molecular ions m/z 650.3 and
m/z 652.3 (complex B1) show the characteristic Cl isotopic
pattern. The product-ion spectrum of m/z 650.3 displays a
fragmentation pattern dominated by the neutral loss of H35Cl
(36 amu) and is assigned to the [ClOMnIV(salen)]+ ion.
Likewise, precursor ions m/z 854.3 and m/z 856.3 represent a
Cl isotopic pattern and eliminate 240 or 242 amu (PhI + HCl)
(13) Electrospray Ionization Mass Spectrometry; Cole, R. B., Ed.;
Wiley: New York, 1997.

9688 J. Am. Chem. Soc., Vol. 122, No. 40, 2000
Adam et al.
Scheme 1
A1 and 6.53% Cl from complex A2. Unfortunately, it was not
possible to grow adequate crystals of the MnIV(salen) complexes
B for X-ray analysis.
contained in the material. The complex B1 derived from A2
contains only one Cl atom per complex, which corresponds to
4.54% Cl content. The measured value of 6.53% implies that
during the reaction the Cl is transferred from the solvent to the
complex. Thus, the mechanistic interpretations in Scheme 1 are
also consistent with the chlorine analysis of the isolated MnIV
complexes.
To rationalize these EPR and MS results, we suggest the
mechanism in Scheme 1 for the formation of the various MnIV
complexes B. Reaction of the [MnIII(salen)]+X- complexes A
with PhIO or NaOCl as oxygen sources leads first to the oxo
complex [OMnV(salen)]+X-. In the absence of a substrate, this
very reactive triplet species7 abstracts a Cl or a H atom from
the CH2Cl2 solvent to afford the complexes [ClOMnIV(salen)]+X-
(B1) or [HOMnIV(salen)]+X- (B1′). That hydrogen abstraction
by the [OMnV(salen)]+X- species from organic substrates is
possible has been documented for benzylic oxidation catalyzed
by MnIII(salen) complexes, which supposedly proceeds through
the oxygen-rebound mechanism.14
With time, complex B1′ eliminates HX and forms the neutral
complex B2. Such a loss of a ligand is expected to be slow for
high-spin octahedral complexes. As the MnIV complexes B1
and B1′ bear similar oxygen functionalities, i.e., OCl versus
OH, the EPR zero-field-splitting parameters should be similar.
Indeed, in the computer simulations of the EPR spectra, these
two paramagnetic species appear as one signal with D ) 4.000
and E/D ) 0.160. The oxo complex B2, however, is distinct
and, thus, has different zero-field-splitting parameters (D )
0.950 and E/D ) 0.296 from the computer simulation). The
dehydrochlorination of the complex B1′ (X ) Cl) to B2 is
responsible for the changes in the EPR spectrum with time
(Figure 1) and, consequently, the relative amounts of the
complexes B1 + B1′ and B2 are altered.
Only the B1 and B1′ complexes are detected by ESI-MS
analysis, because these are cationic. The additionally observed
[ClMnIV(salen)]+ ion (m/z 634.3 in Scheme 1) in the case of
complex A1 is explained by OCl/Cl ligand exchange from
complex B1. Upon addition of TFA to the solution of the
complexes B1 and B2, the neutral complex B2 is protonated to
form B1′ again; however, the OH- ligand is quickly replaced
by the CF3CO2- anion and the resulting [F3CCOOMnIV(salen)]+
complex is detected by ESI-MS analysis (m/z 712.3 in Scheme
1).
Addition of the nonpolar solvent hexane to the solution of
complexes B1 and B2 mainly precipitated the ionic complexes
B1 and B1′. For the complex B1 derived from A1, which
contains also a Cl- as counterion, i.e., two Cl atoms per
complex, a Cl content of 10.33% was calculated. The measured
value of 9.38% suggests that about 25% of complex B1′ is also
In nonhalogenated solvents such as EtOAc, however, there
can be no halogen abstraction from the solvent. Since hydroxyl-
ation of hydrocarbons through H abstraction is documented for
OMnV(salen) species,14 it should in principle be possible to
abstract a H atom from EtOAc and, in fact, more likely than
from CH2Cl2 (the bond dissociation energies are 397.5 kJ mol-1
for EtOAc15 and 411.7 kJ mol-1 for CH2Cl216). However,
chloride ions must be available to generate the OMnIV(salen)
species in EtOAc. Indeed, by starting from A1 (X ) Cl), MnIV
complexes were detected by EPR spectroscopy also in a variety
of nonhalogenated solvents (Figure 4), whereas from A2 (X )
PF6) in EtOAc no paramagnetic signals were seen (Figure 5).
These findings are consistent with the following rationale: The
initially formed OMnV(salen) complex (from A1) suffers inner-
sphere electron transfer with the ligated Cl- counterion to be
reduced to the OMnIV(salen) complex B2. With the noncoor-
dinating PF6- counterion in complex A2 and no chlorine source
(CH2Cl2, Cl-), such a redox process is unlikely and the OMnV-
(salen) complex is destroyed without formation of a MnIV(salen)
species. The failure to detect paramagnetic intermediates by EPR
spectroscopy is taken as evidence for this.
Our in-depth EPR and MS studies reveal that the reaction of
MnIII(salen) complexes A with an oxygen-atom source in the
absence of an olefin leads to paramagnetic MnIV complexes B
(Scheme 1) if either a Cl- serves as counterion or CH2Cl2 is
employed as solvent. Moreover, the present data provide
circumstantial experimental evidence for the very reactive
OMnV(salen) complex as the initial reactive species in the MnIII-
(salen)-catalyzed epoxidation. It is difficult to envisage what
highly reactive manganese species other than MnV(salen) may
serve for the formation of the MnIV(salen) complexes B. To
assess the role of the herein identified, novel paramagnetic MnIV-
(salen) complexes B as oxygen-transfer agents, we explored their
propensity to epoxidize olefins.
Oxidation of Olefins by the MnIV(salen) Complexes B. As
model substrates, two conjugated olefins were chosen, namely
1,2-dihydronaphthalene (1) and styrene (2), which have been
epoxidized catalytically by MnIII(salen) complexes (Table 1).
(14) (a) Hamada, T.; Irie, R.; Mihara, J.; Hamachi, K.; Katsuki, T.
Tetrahedron 1998, 54, 10017-10028. (b) Komiya, N.; Noji, S.; Murahashi,
S.-I. Tetrahedron Lett. 1998, 39, 7921-7924. (c) Lee, N. H.; Lee, C.-S.;
Jung, D.-S. Tetrahedron Lett. 1998, 39, 1385-1388. (d) Larrow, J. F.;
Jacobsen, E. N. J. Am. Chem. Soc. 1994, 116, 12129-12130.
(15) Bordwell, F. G.; Harrelson, J. A., Jr.; Zhang, X. J. Org. Chem. 1991,
56, 4448-4450.
(16) CRC Handbook of Chemistry and Physics, 76th ed.; Lide, D. R.,
Ed.; CRC Press: Boca Raton, 1995.

Mn(Salen)-Catalyzed Jacobsen-Katsuki Epoxidation
J. Am. Chem. Soc., Vol. 122, No. 40, 2000 9689
Table 1. Oxidation of 1,2-Dihydronaphthalene (1), Styrene (2),
and Olefin 3 by the in-situ-Generated MnIV(salen) Complexes B1
and B2
To gain mechanistic insight into these oxidations, the olefin
3 was chosen as radical probe.6,17 The MnIV(salen) complexes
generated from A1 and A2 in CH2Cl2 gave with substrate 3 the
chlorinated, ring-opened products 3a and 3a′ (entries 10 and
13); in fact, no epoxide (3c) could be detected. Again, also
addition of the isolated complex to a CH2Cl2 solution of olefin
3 resulted in the same products (entry 11). In contrast, the MnIV-
(salen) complex formed from A1 in EtOAc oxidized the olefin
3 in low conversion cleanly to the epoxide 3c with no ring-
opening products (entry 12). Complex A2 in EtOAc was again
unreactive toward the olefin (entry 14).
In regard to the enantioselectivity of the chlorination and
epoxidation reactions, the enantiomeric excess (ee) was deter-
mined by multidimensional gas chromatography (MDGC) and
HPLC on chiral phases. Thus, the reaction between the MnIV-
(salen) complexes B formed from A1 and PhIO in CH2Cl2 with
olefin 1 (entry 1) afforded the products 1a (5% ee), 1b (23%
ee), and 1c (25% ee) in much lower enantioselectivity compared
to the catalytic mode of operation (53% ee for 1c in the control
reaction described below).
m.b.b convnc
entry complexa solvent [%]
[%]
product distributiond
1a
47
60
5
1b
22
26
1c
31f
14f
95f
16f
To ascertain whether the identified products persist under the
reaction conditions and, thus, are not transformed into each
other, several control reactions were conducted. Thus, the
independently synthesized epoxides 1c and 3c were submitted
to a CH2Cl2 solution of the in situ-formed MnIV complexes B
(from complex A1). They were, however, reisolated unchanged
after a reaction time of 2 h. The same was true for chlorohydrin
1b, which was not transformed into the epoxide 1c. The
formation of product 3a′, however, was shown to originate from
the oxidative cleavage of the dichloroproduct 3a, since submis-
sion of the latter to a CH2Cl2 solution of the MnIV complexes
B gave within 1 h the cleavage product 3a′ in 6% yield.
To assess the relevance of the herein described MnIV(salen)
complexes B in the catalytic epoxidation reactions with MnIII-
(salen) complexes, the oxidation of the olefins 1, 2, and 3 was
conducted under the usual catalytic conditions (1 equiv of the
olefin, 5-7 mol % of the catalyst A1, 0.3 equiv of PPNO, and
1.5-2.0 equiv of the oxygen source PhIO in CH2Cl2 at 20 °C)
1
2
3
4
5
A1e
A1g
A1
A2
A2
CH2Cl2
73
78
34
20
40
0
CH2Cl2 > 95
EtOAc
CH2Cl2
EtOAc
64
56
72
68
16
2a
62
2b
h
2c
38
6
7
8
9
A1
A1
A2
A2
CH2Cl2
EtOAc
CH2Cl2
EtOAc
29
27
35
19
64
18
34
0
100
14
86
h
3a
53
63
3a′
47
37
3c
10
11
12
13
14
A1
A1g
A1
A2
A2
CH2Cl2
55
55
12
46
0
CH2Cl2 >95
EtOAc
CH2Cl2
EtOAc
>95
93
100
43
57
67
a The MnIII(salen) complex A was first oxidized at 20 °C for 1-2 h
and then the resulting MnIV(salen) complexes B were added to the
olefin. b The mass balances are defined as the sum of all products and
unreacted olefin versus initially used olefin, and were determined by
1H NMR spectroscopy (see Supporting Information). c Determined from
the 1H NMR spectra by assuming that 1 equiv of the MnIV(salen)
complex was consumed per 1 equiv of product. d Normalized to 100%
conversion; very minor amounts of oxidative decomposition products
of the salen ligand were also detected. e No PPNO was added. f Minor
amounts of naphthalene were also formed as byproducts. g The isolated
MnIV(salen) complex was used. h Product 2b was formed in such minor
1
and the products submitted to H NMR and GC-MS analyses.
Only the usual epoxides 1c and 2c were observed; in the case
of 1,2-dihydronaphthalene (1), also 30% naphthalene was found.
More significantly, the chlorinated products 1a and 2a were
not detected. With olefin 3, the catalytic oxidation with MnIII-
(salen)/PhIO gave mainly epoxide 3c; however, several ring-
opened products were also found as byproducts, in particular
the dichloro product 3a. Thus, these findings corroborate the
preliminary results reported by A° kermark and Norrby6 with
MnIII(salen)/PhIO.
These oxidation results may be rationalized in terms of the
different reactivity of the two MnIV complexes B1 and B2
(Scheme 2). The MnIV(salen) complex B1 chlorinates the olefins
in a stepwise pathway, as evidenced by the chlorine-containing
ring-opening product 3a. In this context, it is important to
emphasize that the chlorinated products were also observed
when the isolated MnIV(salen) complexes B were added to a
CH2Cl2 solution of the olefin 1 or 3. Thereby it has been
unequivocally demonstrated that the MnIV(salen) complex B1
is the chlorinating agent and not a reactive Cl species (Cl2,
HOCl), which might have been formed from CH2Cl2 during
the generation of the MnIV(salen) complexes B. Further evidence
for this fact is the enantiomeric excess of the reaction products
(23% for 1b and 5% for 1a) as these engage the chiral salen
ligand in the oxygen-transfer step.
1
amounts that it was only detected by GC-MS and not by H NMR
analysis.
With complexes A1 and A2, mainly dichlorination of the olefin
1 (entries 1, 2, and 4) was observed in CH2Cl2, to afford trans-
1,2-dichloro-3,4-dihydronaphthalene (1a). The oxyfunctionalized
products trans-1-hydroxy-2-chloro-3,4-dihydronaphthalene (1b)
and 1,2-epoxy-3,4-dihydronaphthalene (1c) were formed in
lower amounts. These products were obtained not only with
the in situ-generated MnIV complexes B but also with the
isolated complex that was redissolved in fresh CH2Cl2 (entry
2). In EtOAc, however, complex A1 led in low conversion
almost exclusively to the epoxide 1c (entry 3), while complex
A2 in EtOAc was unreactive toward olefin 1. The results with
styrene (2) as substrate (entries 6-9) parallel those obtained
with 1,2-dihydronaphthalene (1), except that only traces of the
styrene chlorohydrin (2b) were detected. This lack of reactivity
is in agreement with the fact that under these conditions no
EPR signal was detected (Figure 5) and, thus, no MnIV(salen)
species was generated to conduct oxidation.
(17) Donnelly, J. A.; Hoey, J. G.; O’Brien, S.; O’Grady, J. J. Chem.
Soc., Perkin Trans. 1 1973, 2030-2033.

9690 J. Am. Chem. Soc., Vol. 122, No. 40, 2000
Adam et al.
Scheme 2
glycals.24 Although the enzyme itself has been a topic of
considerable interest during the past decades, model systems
are still scarce.23,25 In addition to the Fe(porph) complexes
mentioned above, there is the example of a polymer-supported
Mn(porph) complex that catalyzes the chlorination of dimedon
by H2O2 and Cl-.25 Consequently, we have observed for the
first time that MnIII(salen) complexes also display CPO-type
activity, besides their usual role as effective catalysts in
asymmetric epoxidation.
In contrast to B1, complex B2 does epoxidize the olefins 1-3.
When CH2Cl2 is used as solvent, both complexes B1 and B2
are formed (cf. Scheme 1) and, thus, chlorination as well as
epoxidation were observed (Table 1, entries 1, 2, 4 and 6, 8).
However, in EtOAc as solvent, from A1 the B2 MnIVcomplex
is formed (Scheme 1) and accordingly only epoxide is found
(Table 1, entries 3, 7, and 12); moreover, from A2 in EtOAc
no oxidation was detected (Table 1, entries 5, 9, and 14).
As for the mechanism of the epoxidation by complex B2,
the exclusive formation of the epoxide 3c (no ring-opened
products) suggests a nonradical/noncation oxygen transfer;
however, the phenyl substitution at the carbinyl-radical site
renders this radical clock rather slow (k ) 3.6 × 108 s-1).26
A
more appropriate radical probe is the cis/trans isomerization of
cis-stilbene, for which the rate constant for rotation around the
C-C single bond is ca. 1011 s-1 27
. Therefore, cis-stilbene was
As far as the mechanism of the chlorination by complex B1
is concerned, it is important to realize that the radical probe 3
does not differentiate between a radical and an electrophile
addition to the double bond.18 Indeed, electrophilic addition
would lead to a cyclopropylcarbinyl cation, whose nonclassical
behavior is well established.19 Recent solvolysis experiments
of 3-arylcyclobutyl tosylates and ab initio computations have
shown that with a cation-stabilizing group such as phenyl on
the cyclopropyl ring, solvolysis only affords homoallylic
products.20 Thus, the formation of the ring-opened product 3a
may in principle be envisioned through a radical as well as a
cation pathway. The radical option seems, however, to be
unlikely, because the abstraction of a Cl atom from the MnIV-
(salen) complex B1 would regenerate the very reactive OMnV-
(salen) complex. By contrast, in the cation alternative, the
positively polarized Cl atom in complex B1 is electrophilically
added to the double bond to afford a benzylic cation, which is
nucleophilically trapped by chloride ions or water21 to afford
the dichloro products and the chlorohydrins (Scheme 3). In the
case of olefin 3, the ring-opened dichloro product is obtained
due to the rearrangement of the benzylic cation before nucleo-
philic trapping.
treated with a solution of the complex B2 formed from A1 and
PhIO in EtOAc to afford the stilbene oxide as the only product
in a 36:64 cis/trans ratio at 14% conversion (Scheme 2). This
fact confirms that a stepwise radical mechanism is operating in
the oxygen-transfer process of the MnIV(salen) complex B2,
which is consistent with the findings by Groves and Bruice5
for the epoxidation with analogous OMnIV(porph) complexes.
We conclude that the MnIV(salen) complexes described herein
react with olefins either by stepwise electrophilic chlorination
(complex B1) or by stepwise radical epoxidation (complex B2).
Are the herein described MnIV(salen) complexes B of any
relevance in the usual Jacobsen-Katsuki catalytic epoxidation
reaction with MnIII(salen) complexes? Under the conditions
typical for such MnIII(salen)-catalyzed epoxidations (olefin,
oxygen source, catalyst, and additive all together present from
the start), none of the chlorinated products were detected for
the olefins 1 and 2 as substrates. Therefore, under theses
conditions, no significant amount of the ClOMnIV(salen)
(24) Liu, K. K.-C.; Wong, C.-H. J. Org. Chem. 1992, 57, 3748-3750.
(25) Labat, G.; Meunier, B. J. Chem. Soc., Chem. Commun. 1990, 1414-
1416.
(26) The rate constant for the opening of the secondary benzyl radical
in trans-1-benzyl-2-phenylcyclopropane has been estimated to have this
value (Hollis, R.; Hughes, L.; Bowry, V. W.; Ingold, K. U. J. Org. Chem.
1992, 57, 4284-4287). Since such secondary and tertiary radicals usually
ring-open at similar rates (Newcomb, M.; Tanaka, N.; Bouvier, A.; Tronche,
C.; Horner, J. H.; Musa, O. M.; Martinez, F. N. J. Am. Chem. Soc. 1996,
118, 8505-8506), this value was assumed for the radical clock 3.
(27) For the ethyl radical, a rotation barrier of Ea ) 0.06 kcal/mol has
been determined experimentally (Sears, T. J.; Johnson, P. M.; Jin, P.; Oatis,
S. J. Chem. Phys. 1996, 104, 781-792). For a value of A ) 8.7 × 1012 s-1
(determined by the torsional motion of 290 cm-1 in ethane, cf.: Horn, B.
A.; Herek, J. L.; Zewail, A. H. J. Am. Chem. Soc. 1996, 118, 8755-8756),
the Arrhenius equation gives a rate constant for rotation of 7.9 × 1012 s-1
about the C-C bond. In the benzyl radical derived from the epoxidation of
stilbene (cf. structure),
This reaction sequence bears a remarkable resemblance to
that proposed for the chlorination catalyzed by chloroperoxidase
(CPO),22 which has been supported by recent studies on Fe-
(porph) chemical model systems (Scheme 4).23 Besides the
chlorination of activated C-H bonds (e.g., in dimedon or
aromatic compounds), CPO also catalyzes the halohydration of
(18) Newcomb, M.; Le Tadic-Biadatti, M.-H.; Chestney, D. L.; Roberts,
E. S.; Hollenberg, P. F. J. Am. Chem. Soc. 1995, 117, 12085-12091.
(19) Wiberg, K. B.; Hess, B. A., Jr.; Ashe, A. J., III In Carbonium Ions;
Olah, G. A., Schleyer, P. v. R., Eds.; Wiley-Interscience: New York, 1972;
Vol. III, pp 1295-1345.
(20) Wiberg, K. B.; Shobe, D.; Nelson, G. L. J. Am. Chem. Soc. 1993,
115, 10645-10652.
(21) Both nucleophiles should be present in the solution in low
concentrations. In the case of A1 as starting material, it is also possible
that the counterion is the source of the chloride ion.
(22) Libby, R. D.; Shedd, A. L.; Phipps, A.; Beachy, T. M.; Gerstberger,
S. M. J. Biol. Chem. 1992, 267, 1769-1775.
(23) (a) Wagenknecht, H.-A.; Claude, C.; Woggon, W.-D. HelV. Chim.
Acta 1998, 81, 1506-1520. (b) Wagenknecht, H.-A.; Woggon, W.-D.
Angew. Chem., Int. Ed. Engl. 1997, 36, 390-392.
the steric hindrance should be larger in view of the two phenyl groups and
the bulky OMn(salen) moiety. A conservative estimate of 1-3 kcal/mol is
proposed for the rotation barrier, which results in rate constants of rotation
between 1.6 × 1012 and 5.5 × 1010 s-1. We are grateful to Prof. M.
Newcomb (Wayne State University, Detroit) for advice on this issue.

Mn(Salen)-Catalyzed Jacobsen-Katsuki Epoxidation
J. Am. Chem. Soc., Vol. 122, No. 40, 2000 9691
Scheme 3
Scheme 4
complex (B1) is formed. With olefin 3, however, the chlorinated
products detected with B1 confirm the preliminary studies by
Norrby and A° kermark,6 in which the observed ring-opened
products have been proposed to originate from oxidations by
MnIV(salen) complexes. Nevertheless, even if only trace amounts
of MnIV(salen) complexes B were formed and they would not
effectively compete to reveal detectable quantities of byproducts,
the nonlinearities in Eyring plots, as suggested by Akermark7
et al. and observed by Katzuki8 et al., may derive therefrom.
complexes, which for the first time were fully characterized by
means of EPR and MS analyses as well as product studies, have
been postulated in the literature to be responsible for radical-
type oxidation products.
Acknowledgment. The Deutsche Forschungsgemeinschaft
(SFB 347 “Selektive Reaktionen Metall-aktivierter Moleku¨le”)
and the Fonds der Chemischen Industrie are thanked for
generous financial support, Drs. E. Bill and H. Hummel (Max-
Planck-Institut fu¨r Strahlenchemie, Mu¨lheim) for help with the
interpretation of the EPR spectra and their simulation, and Dr.
M. Bro¨ring (Institut fu¨r Anorganische Chemie, Wu¨rzburg) for
fruitful discussions. We are grateful to Prof. P. Schreier (Institut
fu¨r Pharmazie und Lebensmittelchemie, Wu¨rzburg) for making
his facilities available to us to carry out the ESI-MS measure-
ments. This work is dedicated to Prof. Dr. R. Neidlein
(Heidelberg) on the occasion of his 70th birthday.
Conclusion
Clearly, and mechanistically and preparatively important, the
present results demonstrate that in the MnIII(salen)-catalyzed
epoxidation, the order of mixing of the components is impor-
tant: In the absence of an olefinic substrate, the initially formed
OMnV(salen) complex transforms into the MnIV(salen) com-
plexes B, i.e., the hypochlorite-ligated B1 in CH2Cl2 and the
oxo one B2 in EtOAc. B1 is responsible for electrophilic
chlorination (chloroperoxidase-type activity), while B2 performs
radical-type epoxidation. Be this as it may, fortunately, under
the usual Jacobsen-Katsuki epoxidation conditions (all com-
ponents together from the very start), the MnIV(salen) complexes
B play a minor, if at all noticeable, role. Nevertheless, these
Supporting Information Available: General experimental
procedures and the characterization of the products (PDF). This
material is available free of charge via the Internet at
http://pubs.acs.org.
JA0005082