The reaction of permanganyl chloride with olefins: Intermediates and mechanism as derived from matrix-isolation studies and density functional theory calculations

Article abstract of DOI:10.1002/1521-3765(20011105)7:21<4674::AID-CHEM4674>3.0.CO;2-L

Density functional theory (DFT) calculations predict that the [2+3] addition of tetramethylethylene (TME) to the MnO₂ moiety of MnO₃Cl is thermodynamically favoured over [2+1] addition (epoxidation), while the kinetic barriers for bo

Full text of DOI:10.1002/1521-3765(20011105)7:21<4674::AID-CHEM4674>3.0.CO;2-L

FULL PAPER
The Reaction of Permanganyl Chloride with Olefins: Intermediates and
Mechanism as Derived from Matrix-Isolation Studies and Density Functional
Theory Calculations
Tobias Wistuba and Christian Limberg*[a]
Dedicated to Professor Hansgeorg Schnöckel on the occasion of his 60th birthday
Abstract: Density functional theory
DFT) calculations predict that the
2‡3] addition of tetramethylethylene
TME) to the MnO moiety of MnO Cl
of the epoxidation product [ClO Mn-
at first sight surprising, is supported by
studies in solution, and, even with the
numerically equal energy barriers sug-
gested by the calculations, it can be
rationalised in terms of the much broad-
er reaction channel leading to epoxida-
tion as opposed to the much more
narrow approach path for formation of
the glycolate.
2
(
[
(
{O[C(CH ) ] }] (1) was observed. Com-
3
2 2
pound 1 was characterised by IR spec-
troscopy with the aid of isotopic-enrich-
ment experiments in combination with
DFT frequency calculations. This result,
2
3
is thermodynamically favoured over
[
2‡1] addition (epoxidation), while the
kinetic barriers for both reactions are of
comparable height. However, in an ex-
perimental investigation of the TME/
Keywords: density functional calcu-
lations ´ epoxidation ´ manganese ´
matrix isolation ´ oxidation
MnO Cl system by means of matrix-
3
isolation techniques, selective formation
Introduction
elude characterisation. In the last five or six years investiga-
tions into the reaction mechanisms of simple metal ± oxo
Transition metal oxides and their complexes play pivotal roles
in numerous important biological and nonbiological, stoichio-
metric and catalytic processes in which oxo groups are
transferred to organic substrates.[1±5] For instance, simple d -
metal ± oxo compounds and oxometal halides such as
compounds and oxometal halides have been intensified
again,[
17±30]
as it was recognised that the combination of
methods available today (elaborate quantum-mechanical
calculations, matrix-isolation and gas-phase techniques, as
well as sophisticated kinetic studies) allows deeper insights
0
[
6, 7]
[8, 9]
� [10±12]
[31]
CrO Cl ,
OsO4
MnO4
are applied extensively in
into these systems. For example, the reinvestigation of the
2
2
�
reactions in which oxygen is inserted into a C�H bond or
undergoes addition to an olefinic double bond. The wide-
spread practical use of such reagents has spurred considerable
system MnO /alkylaromatic compounds considerably im-
4
proved the understanding of the reactivity principles charac-
teristic of permanganate oxidations and led to a new way of
[
1±4, 8, 13±16]
0
[19, 20]
interest in the underlying activation mechanisms,
thinking about d -metal ± oxo compounds in general.
and kinetic studies date back to the 1940s. However, although
intense research in this area has provided a broad synthetic,
structural and mechanistic basis, many of the key steps remain
unknown, and some of the reasons for this are the high and
complex reactivities, diverse reaction pathways and the
heterogeneity of the inorganic reduction products, which thus
The reaction of permanganate with olefins recently re-
ceived renewed interest, too. The results of DFT calcula-
tions[ strongly support the original idea that permanga-
nate dihydroxylations follow the [3‡2] cycloaddition mech-
anism established earlier for OsO4[ and exclude a stepwise
process involving a [2‡2] cycloaddition (Scheme 1). The
concerted [2‡3] pathway was found to be favoured by not less
30]
[14]
26]
�
1
than 173 kJmol . The fact that the reaction rates are
[
a] Priv.-Doz. Dr. C. Limberg, T. Wistuba
Anorganisch-Chemisches Institut
Universität Heidelberg
influenced differently by donor and acceptor olefins on going
�
from OsO to MnO4 was attributed to charge differences.
4
Im Neuenheimer Feld 270
It would be of interest to isolate and study the primary
intermediates of such reactions directly, for instance, with the
aid of matrix-isolation techniques, which only recently served
69120 Heidelberg
Fax : (‡49)6221-545707
E-mail: limberg@sun0.urz.uni-heidelberg.de
to identify elusive intermediates of the system CrO Cl /
Supporting information for this article is available on the WWW under
http://wiley-vch.de/home/chemistry/ or from the author.
2
2
olefin.[
22, 23, 31]
However, the reactions of permanganate cannot
4674
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0721-4674 $ 17.50+.50/0
Chem. Eur. J. 2001, 7, No. 21

4674±4685
tions: a 5% TME/Ar mixture was cocondensed with neat
O
O
Mn
O
HO
HO
H2O
MnO Cl, whose deposition rate was regulated by maintaining
3
Mn
+
O
- [H MnO ]
3
4
O
the sample at � 658C. To obtain information on isotopic shifts,
O
O
O
which are very important for band assignments, the experi-
1
6
ments were carried out not only with O MnCl but also with
3
O
Mn
O
18
O MnCl. The IR spectrum shown in Figure 1 was obtained
with O MnCl. In comparison to the spectrum recorded after
3
18
3
O
O
employment of 16
O MnCl it has the advantage of minimal
3
Scheme 1. Reaction mechanism postulated for the oxidation of olefins
with MnO Cl.
band overlap, which substantially facilitated the computa-
tional subtraction of the absorption bands of the unconsumed
starting materials. Bands of new products can thus be
immediately recognized, and consequently the spectrum of
3
be investigated by this method, as permanganate salts are not
volatile enough to allow their embedment in low-temperature
noble gas matrices. However, a neutral derivative of the
permanganate ion with sufficient volatility is available:
18
the O-enriched material is more instructive than that of its
16
16
18
O counterpart. The data for O and O are summarised in
Table 1. The relative intensities of all product bands remained
constant on variation of the experimental conditions (vide
supra), and this suggests that they all belong to a single
product. The most important region is that between 1000 and
permanganyl chloride MnO Cl. Its existence was first postu-
3
lated in 1827, but it was not until 1968 that its identity was
[
32]
confirmed unequivocally.
Subsequently, its spectroscopic
�
1
7
00 cm , where the MnˆO vibrations and certain C�O modes
absorb, and three major product bands can recognized there:
properties (IR, UV/Vis and photoelectron-spectroscopy, as
well as magnetic circular dichroism) were investigated in
�
1
16
18
[
33]
The most prominent band is at 898 cm with a D( O/ O) of
detail, but, probably because it is difficult to handle owing
to its extreme oxidative power, hydrolytic sensitivity, and
instability above � 308C, no reports on its reactivity with
respect to organic or inorganic compounds are available.
Bearing the above-mentioned mechanistic discussions con-
cerning the reactivity of permanganate in mind, we were
interested in studying the reactions of permanganyl chloride
with olefins by means of matrix-isolation techniques in order
to identify the primary products formed. The attention
focused on thermally (and not photochemically) induced
reactions, as these could then be analysed in comparison with
the permanganate system, whereby the effect of replacing an
�
1
� 1
16
18
� 1
3
9 cm , a weaker band at 818 cm (D( O/ O) ˆ 52 cm )
� 1 16 18
and a further intense band at 782 cm
(D( O/ O) ˆ
�
1
2
7 cm ). To identify the corresponding product, the struc-
tures of all reasonable species were calculated (DFT), and
their vibrational frequencies were compared with the exper-
imental data.
Calculations: Density functional calculations are more suit-
able than ab initio methods for frequency calculations on
metal ± oxo species.[ Calculations on the products resulting
from the CrO Cl /olefin system further established that at the
35]
2
2
B3LYP/LanL2DZ level of theory[
22, 23]
deviations of 100 cm
� 1
2
�
�
O
by a Cl ligand is of interest.
are not unusual and have their origins in very small differ-
ences between the theoretical and real structures.[
22, 23, 36]
However, in each case considered, the general band pattern
was in accordance with the experimental one.[
22, 23]
Our
Results and Discussion
theoretical investigation therefore started with the function-
al/basis set combination B3LYP/LanL2DZ, but as problems
were encountered in the calculations on transition states (vide
infra), we switched to the computationally more demanding
basis set 6-311G(d), which is considered to be slightly superior
for metal ± oxo compounds.[ The entire study was performed
at the B3LYP/6-311G(d) level, which has been shown in many
Reaction of MnO Cl with (CH ) CˆC(CH ) : The reactions of
3
3
2
3 2
permanganyl chloride with olefins cannot be induced at
temperatures below 50 K, which is the maximum temperature
to which noble gas matrices (even with Xe coating) can be
warmed without considerable loss of matrix quality. There-
fore, the matrix assembly was designed such that it would
37]
allow reaction between MnO Cl and the olefin during the
cases to represent a reliable combination of functional and
3
basis set for organometallic compounds.[
30b]
Its known ten-
short period of the cocondensation process: MnO Cl and the
3
olefin diluted by argon were sprayed onto a CsI window
cooled to 7 K from two different nozzles, so that, after
focussing, the beams followed a common path of 2 cm in the
gas phase before they condensed. We thus hoped to quench
any primary product formed by immediate isolation in an
argon matrix at 7 K, and to identify it by IR spectroscopy.[
Ethylene and propylene failed to react in this procedure: they
dency to overestimate the fundamental vibrations of metal ±
oxo cores, however, must be borne in mind.[
37]
The system MnO Cl/TME: product identification: Attempts
3
were made to optimise the structures of the potential products
depicted in Scheme 2 (B3LYP/6-311G(d)) in form of their
34]
singlet and triplet states with C symmetry (see Table 2). Only
1
could be matrix-isolated in unchanged form besides MnO Cl.
7, which could conceivably be formed by proton-coupled
3
electron transfer,[
18, 19]
was treated differently. It is a radical
pair, and both radicals were investigated separately in form of
their doublet states.
To identify the species responsible for the experimental IR
absorptions, the calculated vibrational frequencies were
However, the electron-rich tetramethylethylene (TME)
showed a significant reaction. Systematic variation of the
cocondensation rates and concentrations of the reagents to
achieve the best possible matrix quality and an acceptable
conversion lead to the following ideal experimental condi-
Chem. Eur. J. 2001, 7, No. 21
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0721-4675 $ 17.50+.50/0
4675

C. Limberg and T. Wistuba
FULL PAPER
Figure 1. Upper trace: the spectrum of TME isolated in an argon matrix (5 mol%) at 7 K in the region between 430 and 1500 cm 1. Middle trace: the
�
1
8
� 1
spectrum of Mn O
3
Cl isolated in an argon matrix at 7 K in the region between 430 and 1000 cm . Lower trace: the spectrum obtained after cocondensation
Cl and TME/Ar at 7 K after weighted subtraction of the spectra of the starting materials (the asterisks denote bands (positive and negative) due to
imperfect subtraction of the bands of the starting materials; they arise from slight band shifts due to different temperatures). Bottom: The line diagram
18
of Mn O
3
3
3
calculated for the IR spectra of 1 and 2, as shown in Figure 2, by means of DFT calculations at the B3LYP/6 ± 311G(d) level (a scaling factor of 0.97 was
applied).
considered. A scaling factor of 0.97 is accepted for B3LYP/6-
reasonably strong Mn�C bond. On the singlet surface we
[
37]
311G(d) calculations and was applied in the present study,
were able to optimise a structure with a manganaoxetane-like
1
too.
geometry, namely, 3* (Scheme 2), but with a very long Mn�C
�
[30a]
In contrast to the reactions with MnO , for which
distance (2.334 Š, as compared to 2.013 Š
in the manga-
4
�
calculations were performed exclusively on the singlet
naoxetane derived from MnO4 and ethylene shown in
Scheme 1; that is, the Mn�C bond is basically no longer
intact). Accordingly, the Mn�C distance in the structure of
[
30]
potential energy surfaces (PESs), crossovers to the triplet
surfaces have to be considered in the case of MnO Cl. The
3
1
MnO Cl/olefin system itself has a singlet ground state, and the
6*, as optimised from a starting geometry corresponding to
3
0
1
addition of an olefin to the metal ± oxo bonds can lead to a d
6, is also very long (2.660 Š). The frequencies calculated for
3* are not in agreement with those of the experimental bands,
2
0
1
or a d system. While d systems (such as 3 and 6) usually also
have singlet ground states, the d products (1, 2, 4, 5) proved to
2
as, for instance, an intense absorption was predicted at
�
1
be generally much more stable in their triplet states (vide infra
630 cm corresponding to a n(Mn-O-C) vibration, while the
�
1
and Table 2). The situation with MnO Cl is therefore more
region around 630 cm is free of bands in the experimental
spectrum.
3
comparable to that of CrO Cl , for which the olefin oxidation
2
2
[
22±25]
3
products also have triplet ground states,
than to
The structure of 4 could be optimised, and a subsequent
that of OsO , which reacts with olefins solely on the singlet
frequency calculation predicted the most intense band (due to
4
[
26]
surface.
We failed to find a minimum corresponding to the geometry
a n(MnOC)/n (MnO ) mode) to appear in the region around
s
2
�
1
� 1
980 cm with an isotopic shift of 90 cm , and the n(MnO)
0
1
3
� 1
of the d molecules 3 or 3 (the superscripts denote the singlet
vibration was calculated to absorb at around 650 cm with
and triplet spin states) on the PES, probably because the
medium intensity. However, no such bands are observed in
VII
3
Mn centre of MnO Cl is very electron poor and hard (as
the experimental spectrum, so that 4 can be excluded as a
3
�
3
1
opposed to that of the MnO4 ion) which has an additional
product, too. According to the calculations, 5, 6* and the
electron, and therefore it is not capable of forming a
MnO (OH)Cl part of 7 should be characterised by intense
2
4676
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0721-4676 $ 17.50+.50/0
Chem. Eur. J. 2001, 7, No. 21

Reaction of Permanganyl Chloride with Olefins
4674±4685
�
1
18
Table 1. Frequencies [cm ] obtained experimentally by the cocondensation of Mn O
6
3
Cl with a TME/Ar mixture (5%) and calculated frequencies (B3LYP,
3
3
-311G(d); a scaling factor of 0.97 was applied) for 1 and 2 (intensities are given in parentheses). The assignments are based on the computational results.
18O
MnCl/TME
Exptl:
D( O/ O)
Calcd: [ O]- 1
18
3
Calcd:
D( O/ O) for 1
Calcd: [ O]- 2
18
3
Calcd:
D( O/ O) for 2
Assignment
3
1
6
18
16
18
3
16
18
3
1
483 (0.37)
464 (0.40)
0
1503 (0.08)
6
0
0
CH
3
bend
1
489 (0.14)
1493 (0.03)
1481 (0.04)
0
0
0
0
0
0
1
0
1481 (0.04)
1
1
474 (0.02)
472 (0.02)
CH
CH
CH
3
3
3
bend
bend
bend
1
1
1
1
1
1
1
1
1
1
457 (0.37)
448 (0.37)
437 (0.30)
393 (0.30)
388 (0.97)
372 (0.50)
214 (0.37)
175 (0.90)
115 (0.60)
022 (0.22)
0
0
1455 (0.05)
1448 (0.05)
1406 (0.06)
1394 (0.09)
1389 (0.13)
1366 (0.11)
1189 (0.11)
1167 (0.18)
1117 (0.27)
1013 (0.04)
0
0
0
0
0
1
1
0
0
2
1456 (0.07)
1449 (0.02)
0
0
0
4
0
5
2
1394 (0.04)
1384 (0.10)
1372 (0.10)
1217 (0.02)
1184 (0.11)
1149 (0.11)
1128 (0.38)
0
1
0
1
0
CH
3
bend
n(CC)
d(CCC)
1(CH )
3
1
1
9
98 (1)
45
n(MnˆO)
898 (1)
39
948 (1)
43
n
as(MnO
2
)
9
9
8
25 (0.13)
22 (0.06)
29 (0.81)
4
1
26
n
n
n
n
n
n
as(MnO
2
, ring)
8
18 (0.47)
82 (0.62)
52
27
925 (0.32)
746 (0.82)
46
28
s
s
s
(MnO
(MnO
(COC)
2
2
)
7
65 (0.07)
25
, ring)
7
7
6
16 (0.27)
67 (0.35)
20
24
as(MnO
(MnO
2
, ring)/nas(C
, ring)/n (C O )
2 2
O )
s
2
s
2 2
5
54 (0.16)
10
20
4
489 (0.22)
15
512 (0.09)
459 (0.18)
447 (0.16)
394 (0.09)
9
1
6
4
3
21 (0.41)
97 (0.08)
n(MnCl)
13
Cl
Cl
Mn
this is further evidence against the C�H
3 1
Me2
C
Cl Me2
Cl
Mn
O
Mn
O
O
C
activation products 5, 6* and 7. This leaves
O
CMe2
O
Mn
O
CMe2
CMe2
3
3
O
only 1 and 2 as potential products, and
their optimized structures are depicted in
Figure 2.
O
CMe2
O
O
O
O
C
C
1
2
Me2
Me2
3
3*
The theoretical spectrum of 32, which
would appear to be the most likely product
by analogy with the permanganate ion
(Scheme 1), and that of the epoxidation
product 31 are displayed in form of line
diagrams in Figure 1. There are several
arguments which point to an assignment
CMe2
CMe
CMe2
CMe
Cl
CMe2
OH
Mn
OH
Mn
CMe2
CMe2
Cl MeC
Mn
Cl
C
Cl
C
Mn
CH2
H
2
H
2
O
O
O
O
O
O
O
O
O
HO
4
5
6
6*
Cl
Me
C
3
Mn.
O
H C
2
.
CMe2
to 1:
HO
3
O
1) In agreement with the experiment, 1 is
7
predicted to show three bands with large
3
Scheme 2. Potential products of the reaction between MnO Cl and TME.
16
18
O/ O isotopic shifts (the numbers on
the lines in Figure 1): 948 (43), 925 (46)
Table 2. Free reaction energies DG(293.15 K/0.001 atm) [kJmol� 1] of the
compounds 1 ± 6 and 1a ± 3a relative to the corresponding starting
materials, as calculated by DFT [B3LYP/6 ± 311G(d)]. DG(293.15 K/
� 1
� 1
3
and 746 cm (28 cm ), as opposed to 2, which should
display five such bands.
) As can be seen in Figure 1, the frequencies calculated for
2
0
.001 atm) for 7, treated as two doublet radicals, is � 215 kJmol� 1
.
3
1
match the experimental values far better, and the
1
2
3*
4
5
6*
1a
2a
3a
intensity pattern is better reproduced, too. Comparatively
�
1
singlet
� 79 � 161 ‡ 74 � 47
� 95 ‡ 21 � 31 � 152
±
large deviations (50 and 107 cm after scaling) are only
triplet � 102 � 217
±
� 74 � 154 � 66 � 203 ‡ 19
±
found for the n (MnO ) and n (MnO ) vibrations, as
s
2
as
2
expected given the known tendency of B3LYP/6-311G(d)
calculations to overestimate fundamental vibrations of
�
1
absorptions of the MnOH unit in the region around 600 cm ,
�
1
but there is neither a band at this position (Æ150 cm ), nor is
the fingerprint region of the experimental spectrum in
agreement with an organic fragment CH CMeˆCMe , and
[37]
metal ± oxo compounds (similar deviations are conse-
[38]
quently also found for MnO Cl itself ). Here, the agree-
3
2
2
ment between experiment and theory is better with the
Chem. Eur. J. 2001, 7, No. 21
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0721-4677 $ 17.50+.50/0
4677

C. Limberg and T. Wistuba
FULL PAPER
band at 782 cm 1 if it were assigned to one of themÐa
�
triplet would be expected on isotopic scrambling, that is, a
�
1
band intermediate between 782 and 809 cm should
appear and have the highest intensity. The right-hand side
of Figure 3 illustrates this hypothetical case for one of
�
1
these vibrations. Since, however, the band at 782 cm
corresponds to the ring-stretching vibration of the epoxide
3
ligand in 1, the above-mentioned isotopically scrambled
mixture of MnO Cl species leads to only the two known
3
�
1
bands at 809 and 782 cm (as only one O atom is involved
in the vibration) with an intensity ratio of 2:1 (Figure 3,
left), and this lends further strong support to the assign-
3
ment of the experimental spectrum to 1.
1
8
5
) Finally, [ O]TME complexed with OˆCrCl had already
2
[23]
� 1
been found previously to absorb at 788 cm with an
isotopic shift of 28 cm , which is in excellent agreement
with the present results.
�
1
3
3
Figure 2. Structures of 1 and 2 as optimised by DFT methods (B3LYP/6 ±
311G(d)) with bond lengths [Š] and angles [8].
�
1
LanL2DZ basis set (53 and 84 cm , unscaled), but as this
could not be later used for calculating the transition states,
these results are not included in Figure 1 for reasons of
consistency.
3
) Even more importantly, the deviations of the isotopic shifts
calculated for 1 from those obtained experimentally are
3
well within the acceptable range, while the rather large
3
deviations in the case of 2 cannot be explained. Further-
more, two additional medium-intensity bands were calcu-
3
� 1
lated for 2 to appear at 716 and 667 cm , while there are
no bands at all present in the region between 780 and
�
1
5
50 cm in the experimental spectrum.
3
3
4
) To definitely distinguish between 2 and 1, a decisive
experiment with O/ O isotopically scrambled MnO Cl
was performed. If the experimental band at 782 cm ,
which appears at 809 cm on using O MnCl, belonged to
Figure 3. Left: the bands appearing at 809 cm on using 16
� 1
3
O MnCl and at
1
6
18
� 1
18
782 cm on using
O
3
MnCl and their behaviour on employing a mixture of
3
16
16
18
16 18
�
1
3 2
O MnCl, O2 OMnCl and O O MnCl with a ratio of 1:1.8:1. Right: the
band pattern calculated (B3LYP/6-311G(d)) for the situations on the
�
1
16
3
(C-O-C) of 31 and n(MnO
assumption that the band corresponds to n
of 2.
s
2
)
3
3
2
and not to 1, it would have to be assigned to one of the
3
four vibrations calculated to absorb at 829, 765, 716 or
�
1
6
(
67 cm with isotopic shifts of a comparable magnitude
The intense bands of the experimental spectrum must
�
1
3
20 ± 30 cm ). According to the calculations all of these
vibrations involve both O atoms of the glycolate ring; the
employed mixture of
therefore be assigned to 1. This does not necessarily mean
3
that no 2 is formed, since at low concentrations, most of its
1
6
16
18
3
O MnCl,
3
O2 OMnCl and
bands would be masked by those of 1. To determine an upper
1
6
18
3
3
O O MnCl with a ratio of 1:1.8:1 should therefore lead
limit for the formation of 2 relative to 1, we reconsidered the
spectrum with regard to this aspect. A 1:1 mixture of the two
compounds would give an overlap of the two theoretical
2
to three IR spectroscopically distinguishable isotopic
combinations for this glycolate ring ( O/ O, O/ O and
O/ O with a ratio of 4.8:5.6:1). This means that for any of
the four above-mentioned vibrationsÐand hence for the
1
6
16
16
18
1
8
18
3
spectra shown in Figure 1. If 2 were present in the matrices,
its most intense band should certainly be visible independ-
4678
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0721-4678 $ 17.50+.50/0
Chem. Eur. J. 2001, 7, No. 21

Reaction of Permanganyl Chloride with Olefins
ently, but in close proximity to the most intense band of 1
4674±4685
3
� 1
to ca. 70 cm ). Consequently, the deviations of the band
�
1
3
(
898 cm ) on the high-energy side. Only one absorption can
positions in the spectrum calculated for 1 from the exper-
�
1
1
be observed there: a band at 926 cm , whose integral ratio
with respect to the band at 898 cm is 1:13. If this band indeed
belonged to 32, the corresponding additional absorptions
calculated to appear at 925 and 829 cm might be assigned to
a weak band at 845 cm , perturbed by subtraction of a band
imental absorptions are much smaller than those for 1, and
�
1
hence the product must have been formed in its thermody-
namically more stable triplet state.
�
1
�
1
The system MnO Cl/ethylene: thermodynamics and kinetics:
3
3
belonging to unconsumed MnO Cl, as well as a band which is
Having identified 1 as the main product of the MnO Cl/TME
3
3
3
� 1
almost completely covered by the band of 1 at 809 cm ,
gas-phase reaction, the question arises why a [2‡1] and not a
3
�
respectively. Two medium-intensity absorptions remain for 2,
with calculated frequencies of 716 and 667 cm , and these
could tentatively be correlated with two very weak bands at
[2‡3] cycloaddition (as suggested for MnO ) is the main
4
�
1
reaction pathway. We hoped to gain more insights from DFT
calculations of the thermodynamics and kinetics governing
the reaction of MnO Cl with ethylene to give 1a and 2a.
�
1
[39]
around 650 cm which can be barely distinguished from the
noise. On the basis of such assignments and the above-
mentioned integration, a maximum of 7% of the converted
3
3
3
Cl
Mn
TME can be estimated to yield 2, and 93% 1. This relative
yield of 7% represents an upper limit, as there is no residual
band in the experimental spectrum whose intensity exceeds
the intensities of the bands discussed in this context.
Cl
Mn
O
CH2
O
O
O
O
O
CH2
C
C
1a
H2
2a
H2
It can therefore safely be concluded that, rather surpris-
3
3
Scheme 4. The hypothetical ethylene-derived products 1a and 2a.
3
3
ingly, 1 and not 2 is the main reaction product (Scheme 3). In
3
1
fact it is even quite certain that 1 and not 1 is formed: In the
singlet state the two d electrons occupy a weakly antibonding
MnˆO p* orbital, while in the triplet state one of them
occupies a more strongly antibonding MnˆO orbital; that is,
the MnˆO bonds are stronger in the singlet state and absorb
� 1
The triplet state of MnO Cl lies 86 kJmol higher in energy
than the singlet state (in accordance with the formal d
configuration of the metal centre), and as the reaction was
induced below room temperature, it is reasonable to assume
that it starts from the singlet state. The reaction path of lowest
energy leading to 1a involves a crossover from the singlet
surface to the triplet surface immediately before the singlet
transition state (TS) is reached from the reactant side
3
0
at much higher frequencies (D n(MnˆO) is calculated to be
s/t
3
Cl
Cl
Cl
Mn
O
Mn
+
Mn
O
X
O
O
CMe2
O
O
O
O
CMe2
O
(
Figure 4).
C
C
2
Me2
1
Me2
First, the internal reaction coordinate (IRC) was traced
starting from the TS of the singlet surface in both directions
towards the starting materials in their singlet states and
Scheme 3. Experimental outcome of the reaction between MnO
TME.
3
Cl and
Figure 4. Energy profile for the reaction involving epoxidation of ethylene along the IRC path in the region where crossing of the two PESs occurs.
Chem. Eur. J. 2001, 7, No. 21
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0721-4679 $ 17.50+.50/0
4679

C. Limberg and T. Wistuba
FULL PAPER
product 1a in steps of 0.1 amu1/2 bohr (Figure 4). Subsequent-
1
point, where the singlet ± triplet splitting is less than
0.1 kJmol . It is located just before the singlet TS, and thus
�
1
ly, the reaction between MnO Cl in its triplet state and
3
ethylene was investigated: both compounds approach each
other in an unsymmetric fashion to give the radical compound
the rate-determining activation barrier for the reaction of
3
MnO Cl with ethylene to give 1a corresponds closely to the
3
.
2
3
(
(
O) ClMnOCH CH ( 3a) without passing through a TS
amount of energy required to surmount this singlet TS
2
2
3
=
� 1
=
Figure 4). Compound 3a is a local minimum on the triplet
(DG (293.15 K, 0.001 atm) ˆ 102 kJmol ; DE is calculated
3
� 1
surface that is separated from the product 1a by a barrier
with a free enthalpy of only 19 kJmol (both the minimum
Figure 5) and the corresponding TS were optimised, and
to be 43 kJmol ; see Figures 7 and 4). The oxo-transfer
�
1
process is not concerted: the two emerging C�O bonds are
(
quite dissimilar at the point of intersection of the two PESs
(
Figure 8). Once this maximum has been reached on the
singlet PES, the reaction can proceed along the triplet surface,
passing the radical complex as a local minimum and finally
3
reaching the triplet product 1a (Figure 7). Such crossing from
singlet to triplet PESs can be facilitated under the adiabatic
assumption by spin ± orbit coupling.[
24, 25]
Let us now turn to the [2‡3] cycloaddition pathway.
3
� 1
1
Compound 2a is 51 kJmol more stable than 2a and
therefore represents the end point of the IRC. Furthermore
3
� 1
3
2
a lies 137 kJmol lower in energy than 1a, so that its
Figure 5. Structure of 33a as optimised by DFT methods (B3LYP/6 ±
11G(d)) with bond lengths [Š] and angles [8].
formation should indeed be thermodynamically favoured.
However, thermodynamic control is excluded, since the
interconversion of 1a and 2a is prohibited by substantial
barriers (105/242 kJmol , see Figure 7). The TS between the
starting materials and 2a (see Figure 9) is situated on the
singlet surface, only 58 kJmol [DG (293.15 K, 0.001 atm)]
above the starting materials (spin crossover proceeds on the
product side). In consequence, 2a appears to be not only the
thermodynamically but also the kinetically favoured product.
3
3
3
3
� 1
tracing the IRC showed that the TS is connected to both 3a
3
3
and 1a, as shown in Figure 6, which plots the electronic
�
1
=
energy. Intramolecular radical trapping by the terminal oxo
.
2
3
ligands of (O) ClMnOCH CH to give 2a, which is also
2
2
3
possible in principle, is prohibited by the rather large distance
between these oxo ligands and the carbon-centred radical, as
compared to the distance between the C-centred radical and
the alkoxidic oxygen atom, which, in addition, has a higher
The system MnO Cl/TME: thermodynamics and kinetics:
3
Naturally, the enthalpies for the formation of the glycolate
and epoxide complexes from reactions of MnO Cl with the
olefinic bonds change to a certain extent on going from
ethylene to TME: DG [ 1a/ 2a](293.15 K, 0.001 atm) is
22 kJmol larger than DG [ 1/ 2](293.15 K, 0.001 atm), and
the reactions become more exothermic. According to the
3
spin density than the oxo O atoms. Compound 1a is
3
�
1
1
3
6 kJmol more stable than the corresponding product 1a
3
3
on the singlet surface, and it therefore clearly represents the
ground state of the product. Given this information on the
singlet and triplet PESs, it was possible to locate their crossing
f
�
1
3
3
f
2.
3
3
Figure 6. Energy profile for the conversion of the alkoxide radical (O)
2
ClMnCH
2
CH ( 3a) to the end product 1a along the IRC path.
4680
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0721-4680 $ 17.50+.50/0
Chem. Eur. J. 2001, 7, No. 21

Reaction of Permanganyl Chloride with Olefins
4674±4685
3
3
3
Figure 7. Reaction profile of the suggested mechanisms for the formation of 1a and 2a from MnO Cl and ethylene.
results of Rösch et al.,[40] this means that the activation barrier
for the [2‡3] cycloaddition of TME is lower than that for
ethylene, and a similar trend can also be expected for the
epoxidation.
To analyse the influence of permethylation on these
barriers more accurately, we performed further calculations,
which were based on the results described above for ethylene.
The rate-determining activation barrier for the epoxidation of
3
ethylene by MnO Cl (resulting in 1a) corresponds closely to
3
the energy of the corresponding TS on the singlet surface.
Assuming that the same is true for TME,[ we determined
the structure and energy of the TS between the starting
39]
1
� 1
materials and 1. It is located only 45 kJmol above the
�
1
starting materials, that is, the barrier is lowered by 57 kJmol
on going from ethylene to TME, which represents an
enormous substituent effect. This effect is less pronounced
for the [2‡3] cycloaddition, whose activation barrier was
approximated[ by monitoring the optimum energies for
Figure 8. The structure of the rate-determining transition state for the
epxidation of ethylene by MnO Cl, as optimised by DFT methods (B3LYP/
-311G(d)) with bond lengths [Š] and angles [8].
3
6
39]
geometries in which TME approached MnO Cl symmetrically
3
in five steps with fixed equally long O ´´´ C distances in a [2‡3]
�
1
� 1
fashion. It amounts to 45 kJmol (cf. 58 kJmol for ethyl-
ene), so that according to our calculations the barriers on the
3
3
way to 1 and 2 are in fact numerically equal. Changing the
basis sets or functionals used for the calculations is likely to
change the calculated absolute (and to a smaller extent maybe
even the relative) energies somewhat.[
22±25, 30, 35, 36]
Moreover,
it has to be considered that we have made some approxima-
tions and assumptions[ and that (as for any method) there
are certain error ranges valid for results obtained by DFT
methods. In an extreme case the epoxidation barrier could be
39]
�
1
overestimated by a few kJmol , while the cycloadditon
�
1
barrier could be underestimated by a few kJmol , and this
would lead to a kinetic control on the basis of these energy
barriers. However, the fact that DFT calculations at a
relatively high level of theory predicted the barriers to be of
Figure 9. The structure of the rate-determining transition state for the
Cl, as optimised by DFT methods
B3LYP/6 ± 311G(d)) with bond lengths [Š] and angles [8].
[
(
2‡3] cycloaddition of ethylene to MnO
3
Chem. Eur. J. 2001, 7, No. 21
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0721-4681 $ 17.50+.50/0
4681

C. Limberg and T. Wistuba
FULL PAPER
an equal height must also be dealt with. This is the basis on
which the question why MnO Cl reacts with TME selectively
3
3
by direct epoxidation to give 1 must be answered, even
though this pathway is not favoured thermodynamically and
there is no clear-cut kinetic reason (as far as the barriers are
3
3
concerned) for the selective formation of 1 rather than 2. Is a
more constructive explanation available for the discrepancy
between experimental and theoretical results than errors
made by theory?
Reaction channels: In seeking an answer it is first necessary to
consider the reaction conditions. The cocondensation of the
�
5
two gases takes place at very low pressures (ca. 10 mbar), so
that the mean free pathway of the molecules amounts to 1 ±
2
cm. Since the common path followed the MnO Cl and TME
3
gases after leaving their separate nozzles is only 2 cm, the
molecules can collide only once or twice before condensation
occurs, and these collisions will additionally not be very
effective, because the beams flow almost parallel to one
another and the MnO Cl molecules still have the thermal
3
energy characteristic for a temperature of � 658C. It is
therefore unlikely that the molecules already react with each
other in the gas phase. More collisions occur during the
process of condensation on the CsI window, so that the
reaction can proceed there, even though the molecules have
already lost some of their thermal energy at that stage. This
means that the activation barriers must be very low, even
lower than those calculated above. Apparently the calcula-
tions overestimate these barriers, a finding which is not
unusual, especially as the barriers were calculated for a
temperature of 293.15 K, while the actual reaction probably
takes place far below that temperature. However, for the
present problem only relative trends, such as the fact that the
Figure 10. Reaction channels for the epoxidation and [2‡3] cycloaddition
with MnO Cl.
3
spond to a four-electron/four-obital interaction[41] and the
transition state structure shown in Figure 8 is of course just
one possibility, in which the olefin approaches from the
bottom side of the plane defined by the three oxo ligands. It
can, however, also approach from above, and while entering it
is also free to rotate around the C�O axis, which must
3
3
barriers leading to 1 and 2 are of comparable magnitude, are
of interest, and these should be reliable. Note, however, that
calculated free activation enthalpies are not necessarily
identical with those determined experimentally in kinetic
=
investigations. The DE values can be reliably calculated for a
subsequently be established as a bond in the TS. All this
makes these reaction channels much broader than those of the
[2‡3] cycloaddition, and a corresponding projection on a
hemisphere is shown in Figure 10. Epoxidation is thus
statistically favoured over [2‡3] cycloaddition. The latter
practically requires frontal collision of the two molecules,
while epoxidation could conceivably also occur in a less direct
specific two-dimensional reaction profile, and an evaluation
of the calculated frequencies for the structures involved
=
results in DG values like those discussed so far here.
=
However, the real, measurable DG is determined by the
three-dimensional PES, in which additional entropy contri-
butions (which are disregarded in the corresponding partition
function used in the calculation) can become important, for
instance, the number of entrances a given system has to a
particular path and their accessibility. Hence, the reaction
channels should also be taken into account.
3
collision, that is, formation of 1 is additionally supported by
the higher efficiency of the collisions that lead to it. Never-
3
theless, some 2 should be formed, too, and on the basis of our
IR spectra 7% of the converted TME may indeed have
3
3
3
Formation of 2 can only occur if the olefin approaches one
reacted to give 2 (vide supra). Compound 1, however, is
clearly the main product. The fact that ethylene shows no
reactivity can be explained in terms of the somewhat higher
activation barriers (vide supra), which may be only just too
high to be crossed during the process of condensation.
MnO unit rather symmetrically within its plane, so that the
2
approach path for a successful reaction is quite narrow. There
are three such planes, and therefore three reaction channels
exist. A projection of these on a hemisphere is shown in
Figure 10. In contrast, as outlined above, the reaction
coordinate calculated for the epoxidation of ethylene re-
vealed that the olefin is attacked unsymmetrically by one oxo
group, so that there are many possible ways in which the olefin
Behaviour of MnO Cl in solution: The above conclusions
3
should still hold if the rather exotic ªsolventº argon is
replaced by a more common, but still relatively inert solvent.
To test this hypothesis, we performed additional experiments
can approach the oxo ligands of MnO Cl and be epoxidised.
3
The orbital interaction initialising the reaction will corre-
on the reactivity of MnO Cl towards olefins in the liquid
3
4682
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0721-4682 $ 17.50+.50/0
Chem. Eur. J. 2001, 7, No. 21

Reaction of Permanganyl Chloride with Olefins
4674±4685
phase. Epoxide complexes should again be the main primary
products, although at somewhat higher temperatures inter-
conversion to the more stable diolates could be expected if the
corresponding barriers are low enough. According to the
calculations, in the case of ethylene it is quite substantial
TME is treated with MnO Cl by the procedure described
3
above, the epoxide is again obtained as the main product
(77.2%) together with the product of its Lewis acid catalysed
rearrangement (pinacolone, 12.3%) and some pinacol
(10.4%), probably resulting from epoxide hydrolysis. Con-
�
1
3
(105 kJmol ). Any of these primary products should undergo
sequently, 1 is formed in solution as the main product, too,
subsequent aggregation reactions leading to a solid, whose
workup with polar, water-containing solvents provides the
free organic oxidation products (epoxides or diols).
and this means that the epoxidation is again the preferred
reaction pathway; this is additional support for the above
interpretation.
�
Treating trans-5-decene at � 808C with a solution of
A question remains: Why does MnO4 not epoxidise
MnO Cl in CFCl immediately leads to a suspension of a
olefins? The present study showed that the seemingly
3
3
�
brown precipitate in a colourless solution. Workup with wet
acetonitrile (at � 808C or after warming to room temper-
ature) gave mainly trans-1,2-dibutyloxirane (37.5%) in addi-
analogous reaction of an olefin with MnO4 now must be
considered as a completely different case, as the replacement
�
2�
of Cl by O not only influences the singlet ± triplet splitting,
but also introduces an electronic charge. This leads to a
[
42]
tion to the chlorinated products threo-6-chlorodecan-5-ol
17.6%; essentially free of erythro isomer), the correspond-
ing 6-chlorodecan-5-one (36.1%) and 5,6-dichlorodecane
8.8%), as shown in GC/MS measurements by comparison
with authentic samples (Scheme 5). Neither threo- or erythro-
(
situation in which, according to additional calculations, the
�
epoxidation of ethylene by MnO is endothermic (DE ˆ
4
�
1
[30]
(
‡53 kJmol ), while the [2‡3] cycloaddition
is highly
exothermic, as in the case of MnO Cl. Moreover, the negative
3
5
6
,6-dihydoxodecane nor the products of their overoxidation,
-hydroxydecan-5-one, 5,6-decanedione or valeric acid, could
charge leads to higher activation barriers, that is, to kinetic
control, which, however, again favours the [2‡3] cycloaddi-
tion. It is therefore understandable, that diols rather than
epoxides are isolated from permanganate oxidations.
be detected. As MnO Cl already starts to decompose above
3
�
308C, it is impossible to weigh it without sacrifying part of it.
This of course prohibits an accurate mass balance. The above
products account for 60% of the MnO Cl employed, which is
3
satisfactory considering that a significant part of the MnO Cl
3
Conclusion
would have decomposed already before it came into contact
with the olefins and that the formation of some of the
products can be assumed to require two equivalents of
3
In summary, the high selectivity for the product 1 has its
origin, at least partly, in the greater number of approach paths
for the olefin leading to 31, which are in addition much
MnO Cl.
3
3
Our interest now focused on the oxidation of TME in
solution, which can be compared directly with the reaction
performed with the aid of the matrix isolation technique.
Tetramethyethylene was shown previously to react quantita-
tively with the oxo groups of metal oxide chlorides,[
probably because the individual steps relevant to oxygenation
proceed faster due to the electron richness of the olefin, so
that side reactions (e.g., chlorinations) are suppressed. When
broader than those which yield 2. This situation leads to a
statistical preference for the epoxidation path, which might be
additionally favoured by lower requirements for the effec-
tiveness of the corresponding collisions. It may also be that the
epoxidation barrier is slightly lower than that of [2‡3]
cycloaddition, but we found no indication of this in our
calculations. The present study thus provides precedent for
the importance of considering all factors that guide a reaction
so as be able to correctly predict an
23]
experimental result.
C4H9
H
C4H9
Product discrimination due to
different approach paths, one of
which is less accessible, might also
provide an answer to the long-
standing question why CrO Cl ep-
Cl
O
+
C4H9
CFCl3, -80°C
CH3CN, H2O, RT
C H
HO
Cl
4
9
C4H9
H
+
Mn
O
O
2
2
C4H9
O
H
C4H9
H
H
oxidizes olefins, while OsO and
MnO4 dihydroxylate them, al-
4
Cl
Cl
Cl
C4H9
�
+
though all these metal ± oxo com-
pounds should favour a [2‡3] cy-
cloaddition in the primary step
according to calculations per-
formed for ethylene.[
C4H9
O
C4H9
17, 23±26, 30, 31]
O
Cl
Mn
OH
O
CFCl3, -80°C
CH3CN, H2O, RT
Currently we are investigating
whether it is possible to design the
matrix set-up in a way that allows
even ethylene to react or that gives
compound 31 sufficient time and
HO
+
+
+
O
O
O
3
[
Mn]
H2O
energy to rearrange to 2 before the
Scheme 5. Products obtained from the oxidation of trans-5-decene with MnO
3
Cl after aqueous workup.
reaction mixture is quenched.
Chem. Eur. J. 2001, 7, No. 21
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0721-4683 $ 17.50+.50/0
4683

C. Limberg and T. Wistuba
FULL PAPER
[
[
4] Organic Synthesis by Oxidation with Metal Compounds (Eds.: W. Mijs,
C. R. H. I. de Jonge), Plenum, New York, 1986.
5] H. K. Singh in Cytochrome P-450: Structure, Mechanism, and
Biochemistry (Ed.: R. R. Ortiz de Montellano), 2nd ed., Plenum,
New York, 1995; A. E. Shilov, A. A. Shteinman, Acc. Chem. Res. 1999,
Experimental Section
Computational method: Density functional calculations were carried out
with the Gaussian/DFT[ series of programs. The B3LYP formulation of
43]
[44]
density functional theory was used with the basis sets LanL2DZ and
-311G(d). Harmonic vibrational frequencies and infrared intensities were
3
2, 763.
6
[
[
[
6] G. Cainelli, G. Cardillo, Chromium Oxidations in Organic Chemistry,
Springer, Berlin, 1984; K. B. Wiberg in Oxidation in Organic
Chemistry (Ed.: K. B. Wiberg), Academic Press, New York, 1965, 69.
7] F. H. Westheimer, Chem. Rev. 1949, 49, 419; F. Freeman in Organic
Syntheses by Oxidation with Metal Compounds (Ed.: W. J. Mijs,
C. R. H. I. de Jonge), Plenum, New York, 1986, Chap. 2, pp. 41.
8] a) K. B. Sharpless, K. Akashi, J. Am. Chem. Soc. 1975, 97, 5927;
b) K. B. Sharpless, T. C. Flood, J. Am. Chem. Soc. 1971, 93, 2316; L. M.
Hjelmeland, G. H. Loew, J. Am. Chem. Soc. 1977, 99, 3514; review: M.
Schröder, Chem. Rev. 1980, 80, 187.
predicted at these levels by numerical second derivatives by using
analytically calculated first derivatives. All points were characterised as
minima or first-order saddle points by frequency analysis.
Matrix-isolation experiments: Cooling by liquid helium gave temperatures
�
7
of about 7 K inside a shroud maintained at a pressure below 10 mbar.
Olefin/argon mixtures were prepared in a vacuum line by using standard
manometric techniques and then deposited on the cold support. A CsI
window was used for all the IR experiments and MnO
3
Cl samples (held at
�
658C) were deposited continuously, typically over a period of 20 min; the
rate of deposition of the material to be isolated was regulated by its own
low volatility at that temperature.
[
9] C. Döbler, G. M. Mehltretter, U. Sundermeier, M. Beller, J. Am.
Chem. Soc. 2000, 122, 10289.
IR spectra were recorded in transmission mode in the range of 4000 ±
[
[
10] F. A. Carey, R. J. Sundberg, Advanced Organic Chemistry, Plenum
Press, New York, 1977.
11] G. L. Lee, T. Chen, J. Org. Chem. 1991, 56, 5341, and references
therein; G. L. Lee, T. Chen, J. Am. Chem. Soc. 1993, 115, 11231.
[12] W. A. Waters, Quart. Rev. 1958, 12, 29; H. B. Henbest; W. R. Jackson,
B. C. G. Robb, J. Chem. Soc B 1966, 803.
[13] K. B. Sharpless, A. Y. Teranishi, J.-E. Bäckvall, J. Am. Chem. Soc.
1977, 99, 3120.
�
1
� 1
300 cm
and with
a
resolution of 1 cm
on a Bruker IFS 88 v
spectrophotometer.
Starting materials: Tetramethylethylene (99% pure) was supplied by
Aldrich and further purified by fractional condensation in high vacuum
prior to use. Argon (99.999% pure) was supplied by Messer Griesheim.
1
8
O-Enriched MnO
3
Cl was prepared by
Isotopically scrambled MnO Cl was prepared by heating a
O for 60 h at 1008C. After removal of the
a
procedure published for
[45]
MnO
solution of KMnO
3
F.
3
1
6
18
4
in H
2
O/H
2
[14] G. Wagner, J. Russ. Phys. Chem. Soc. 1895, 27, 219; K. B. Wiberg, K. A.
Saegebarth, J. Am. Chem. Soc. 1957, 79, 2822.
[15] K. A. Jùrgensen, R. Hoffmann, J. Am. Chem. Soc. 1986, 108, 1867.
volatile substances, as-prepared material was employed for the synthesis of
32d]
MnO
3
Cl.[
[
16] M. Sono, M. P. Roach, E. D. Coulter, J. H. Dawson, Chem. Rev. 1996,
6, 2841.
17] LReO : D. V. Deubel, G. Frenking, J. Am. Chem. Soc. 1999, 121, 2021;
Solution experiments: In
a
typical experiment MnO
3
Cl (0.045 g,
9
0.325 mmol) was condensed into a flask equipped with a greaseless tap,
[
3
CFCl
3
(20 mL) was cocondensed and the mixture was warmed to � 808C. It
M. A. Pietsch, T. V. Russo, R. B. Murphy, R. L. Martin, A. K. Rapp e ,
Organometallics 1998, 17, 2716; S. Köstlmeier, V. Nasluzov, W. A.
Herrmann, N. Rösch, Organometallics 1997, 16, 1786.
was subsequently cocondensed onto a frozen (� 1968C) mixture of a
1.5 molar excess of the olefin and CFCl (20 mL). After warming the
3
reaction vessel to � 808C with stirring, the solution immediately turned
[
[
18] CrO
2
Cl
2
/C�H: G. K. Cook, J. M. Mayer, J. Am. Chem. Soc. 1994, 116,
colourless, and a brown solid precipitated. The suspension was stirred at
1
855; G. K. Cook, J. M. Mayer, J. Am. Chem. Soc. 1995, 117, 7139.
�
808C for 1 h and at room temperature for 1 h. All volatile substances
�
19] MnO
4
/C�H: K. A. Gardner, J. M. Mayer, Science 1995, 269, 1849;
were removed and investigated by GC/MS with a standard to give a mass
balance. They contain unconverted olefin and epoxide. The brown residue
was treated with wet acetonitrile, and again all volatile substances were
removed and investigated by GC/MS. They contained epoxide, carbonyl
compounds and chlorinated products.
K. A. Gardner, L. L. Kuehnert, J. M. Mayer, Inorg. Chem. 1997, 36,
069.
2
[
[
20] J. M. Mayer, Acc. Chem. Res. 1998, 31, 441.
21] Reactions of metal ± oxo compounds in the gas phase: S. Shaik, M.
Filatov, D. Schröder, H. Schwarz, Chem. Eur. J. 1998, 4, 193; D.
Schröder, H. Schwarz, S. Shaik, Struct. Bonding 2000, 97, 91; D.
Schröder, S. Shaik, H. Schwarz, Acc. Chem. Res. 2000, 33, 139; P.
Jackson, J. N. Harvey, D. Schröder, H. Schwarz, Int. J. Mass Spect.
Authentic samples of the isomers of 5,6-decanediol were prepared by
treatment of trans-5-decene with OsO and subsequent aqueous workup, as
4
well as by acid/base-catalysed hydrolysis of trans-1,2-dibutyloxirane.
Samples of 6-hydroxydecane-5-one, 5,6-decanedione and valeric acid were
2
001, 204, 233.
22] CrO Cl /olefin: C. Limberg, R. Köppe, H. Schnöckel, Angew. Chem.
998, 110, 512; Angew. Chem. Int. Ed. 1998, 37, 496.
4
obtained by reaction of trans-5-decene with an excess of KMnO . Epoxides
[
2
2
were generally synthesised by oxidation of the olefins with m-chloroper-
benzoic acid. erythro-6-Chlorodecan-5-ol can be synthesised by opening
the corresponding epoxide with HCl, and 5,6-dichlorodecane is the main
1
[
[
23] CrO
24] CrO
2
Cl
Cl
2
/olefin: C. Limberg, R. Köppe, Inorg. Chem. 1999, 38, 2106.
/olefin: M. Torrent, L. Deng, T. Ziegler, Inorg. Chem. 1998, 37,
2
2
2
product when the olefin is treated with Cl .
1
307.
[
[
25] CrO
2
Cl
2
/olefin: M. Torrent, L. Deng, M. Duran, M. Sol a , T. Ziegler,
Can. J. Chem. 1999, 77, 1476.
26] OsO
/olefin: a) S. Dapprich, G. Ujaque, F. Maseras, A. Lled o s, D. G.
4
Acknowledgement
Musaev, K. Morokuma, J. Am. Chem. Soc. 1996, 118, 11660; b) U.
Pidun, C. Boehme, G. Frenking, Angew. Chem. 1996, 108, 3008;
Angew. Chem. Int. Ed. Engl. 1996, 35, 2817; c) M. Torrent, L. Deng, M.
Durun, M. Sol aÁ , T. Ziegler, Organometallics 1997, 16, 13; d) A. J.
Del Monte, J. Haller, K. N. Houk, K. B. Sharpless, D. A. Singleton, T.
Strassner, A. A. Thomas, J. Am. Chem. Soc. 1997, 119, 9907.
This work was supported by the Deutsche Forschungsgemeinschaft
Habilitanden and Heisenberg scholarships to C.L. and financial support
by the SFB 247), the Fonds der Chemischen Industrie and the Dr. Otto
Röhm Gedächtnisstiftung.
(
[
27] CrO
b) L. Deng, T. Ziegler, Organometallics 1997, 16, 716.
[28] CrO Cl /MeOH: B. S. Ault, J. Am. Chem. Soc. 1998, 120, 6105.
2 2
Cl /MeOH: a) T. Ziegler, J. Li, Organometallics 1995, 14, 214;
[
1] R. H. Holm, Chem. Rev. 1987, 87, 1401; R. H. Holm, Coord. Chem.
Rev. 1990, 100, 183.
2
2
[29] K. N. Houk, T. Strassner, J. Am. Chem. Soc. 2000, 122, 800.
[30] a) K. N. Houk, T. Strassner, J. Org. Chem. 1999, 64, 800; b) T.
Strassner, M. Busold, J. Org. Chem. 2001, 66, 672.
[
2] W. A. Nugent, J. M. Mayer, Metal ± Ligand Multiple Bonds, Wiley,
New York, 1988; L. K. Woo, Chem. Rev. 1993, 93, 1125; R. H. Holm in
Bioinorganic Catalysis (Ed.: J. Reedijk), Marcel Dekker, New York,
[31] C. Limberg, Chem. Eur. J. 2000, 6, 2083.
1
993, pp. 347 ± 393; K. B. Sharpless, Tetrahedron 1994, 50, 4235, and
[32] a) J. Dumas, Ann. Chim. Phys. 1827, 36, 81/2; b) B. Aschoff, Chem.
Zentral. 1860, 31, 838; B. Aschoff, Monatsber. Kgl. Preuû. Akad. Wiss.
Berlin 1860, 474/85, 483/5; c) B. Franke, J. Prakt. Chem. 1887, 36, 31;
d) T. S. Briggs, J. Inorg. Nucl. Chem. 1968, 30, 2866.
references therein; E. N. Jacobsen, L. Deng, Y. Furukuwa, L.
Martinez, Tetrahedron 1994, 50, 4324, and references therein.
3] J. H. Enemark, C. G. Young, Adv. Inorg. Chem. 1993, 40, 1.
[
4684
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0721-4684 $ 17.50+.50/0
Chem. Eur. J. 2001, 7, No. 21

Reaction of Permanganyl Chloride with Olefins
4674±4685
[
33] R. M. E. Vliek, P. R. Boudewijn, P. J. Zandstra, Chem. Phys. Lett.
[41] The oxidation must involve both the HOMO and LUMO of the
oxidant, as well as the HOMO and LUMO of the olefin.[
18, 19]
The
1976, 39, 405; E. Diemann, E. L. Varetti, A. Müller, Chem. Phys. Lett.
1977, 51, 460; P. J. Aymonino, H. Schulze, A. Müller, Z. Naturforsch. B
1969, 24, 1508; J. Jasinski, S. L. Holt, J. Chem. Soc. 1972, 1046; E. L.
HOMO of MnO
[42] The source of chlorine is not entirely clear for some of the products.
Dichlorination is probably due to partial decomposition of MnO Cl to
give free Cl Abstraction of Cl from the solvent CFCl by
3
Cl is primarily an oxygen lone pair.
Varetti, A. Müller, Z. Anorg. Allg. Chem. 1978, 442, 230; J. P. Jasinski,
S. L. Holt, J. H. Wood, J. W. Moskowitz, J. Chem. Phys. 1975, 63, 1429;
S. Rettrup, S. Bo, C. J. Ballhausen, Acta Chem. Scand., 1990, 44, 853.
34] The optical properties of matrices thus obtained are not as good as
those of matrices generated by prior matrix isolation of a given
substrate followed by a subsequent photolytically induced reaction.
Hence, the bands in the IR spectra of such matrices are comparatively
broad.
3
2
.
3
intermediate alkoxide radicals (analogous to 3a, vide supra) is also
conceivable.
[
[43] Gaussian 98, Revision A.6, M. J. Frisch, G. W. Trucks, H. B. Schlegel,
G. E. Scuseria, M. A. Robb, J. R. Cheeseman, V. G. Zakrzewski, J. A.
Montgomery, Jr., R. E. Stratmann, J. C. Burant, S. Dapprich, J. M.
Millam, A. D. Daniels, K. N. Kudin, M. C. Strain, O. Farkas, J. Tomasi,
V. Barone, M. Cossi, R. Cammi, B. Mennucci, C. Pomelli, C. Adamo,
S. Clifford, J. Ochterski, G. A. Petersson, P. Y. Ayala, Q. Cui, K.
Morokuma, D. K. Malick, A. D. Rabuck, K. Raghavachari, J. B.
Foresman, J. Cioslowski, J. V. Ortiz, B. B. Stefanov, G. Liu, A.
Liashenko, P. Piskorz, I. Komaromi, R. Gomperts, R. L. Martin,
D. J. Fox, T. Keith, M. A. Al-Laham, C. Y. Peng, A. Nanayakkara, C.
Gonzalez, M. Challacombe, P. M. W. Gill, B. Johnson, W. Chen, M. W.
Wong, J. L. Andres, C. Gonzalez, M. Head-Gordon, E. S. Replogle,
J. A. Pople, Gaussian, Inc., Pittsburgh PA, 1998.
[
[
35] M. Torrent, P. Gili, M. Duran, M. Sol aÁ , J. Chem. Phys. 1996, 104, 9499.
36] A.B e rces, T. Ziegler, J. Phys. Chem. 1995, 99, 11417; A. B e rces, J.
Phys. Chem. 1996, 100, 16538.
[
[
37] I. Bytheway, M. W. Wong, Chem. Phys. Lett. 1998, 282, 219.
�
1
[33]
38] Vibrational frequencies [cm ] of MnO
calculated [B3LYP/6 ± 311G(d), scaling factor of 0.97]): nas(MnO
51.5 versus 1064.6, nas(MnO ) 889.9 versus 1012.3, n(MnCl) 458.9
versus 468.0.
3
Cl (experimental
versus
3
)
9
3
[
39] The detailed investigation of the TME system is computationally very
demanding.
[44] A. D. Becke, J. Chem. Phys. 1993, 98, 5648.
[
40] The activation energies of [2‡3] cycloadditions between olefins and
metal oxo moieties can be correlated with the reaction energies on the
basis of Marcus theory: P. Gisdakis, N. Rösch, J. Am. Chem. Soc. 2001,
[45] E. L. Varetti, R. R. Filgueira, A. Müller, Spectrochim. Acta, 1981, 37A,
369.
123, 697.
Received: January 30, 2001
Revised: June 20, 2001 [F3037]
Chem. Eur. J. 2001, 7, No. 21
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0721-4685 $ 17.50+.50/0
4685