X-ray crystal and NMR structure of a highly reactive bidentate 1,2-diamine-OsO4 complex, formally a 20-electron outer valence shell species. Mechanistic implications for the 1,2-diamine-accelerated dihydroxylation of olefins by OsO4

Full text of DOI:10.1021/ja960536d

J. Am. Chem. Soc. 1996, 118, 7851-7852
7851
X-ray Crystal and NMR Structure of a Highly
Complex,
Reactive Bidentate 1,2-Diamine-OsO
4
Formally a 20-Electron Outer Valence Shell Species.
Mechanistic Implications for the 1,2-Diamine-
Accelerated Dihydroxylation of Olefins by OsO
4
E. J. Corey,* Sepehr Sarshar, Mihai D. Azimioara,
Ronald C. Newbold, and Mark C. Noe
Department of Chemistry, HarVard UniVersity
Figure 1. X-ray crystal structure of 1‚OsO
4
. Bond lengths (Å): Os-
Cambridge, Massachusetts, 02138
(
1)-O(22), 1.740(9); Os(1)-O(21), 1.740(12); Os(1)-O(11), 1.756-
10); Os(1)-N(1), 2.329(11). Bond angles (deg): O(22)-Os-O(11),
(
9
ReceiVed February 20, 1996
8.9(7); O(11)-Os-O(12), 150.6(5); O(21)-Os-N(1), 88.2(5); O(12)-
Os-N(1), 80.1(5); N(1)-Os-N(2), 79.2(4).
The dihydroxylation of olefins by OsO4 is dramatically accel-
erated by 1,2-diamines, a phenomenon which underlies the
production of chiral 1,2-diols from olefins using various chiral
geometry. This situation is reminiscent of the observed ligand
arrangement about osmium in the 1:1 complex of OsO4 with
,2
1
,2-diamines.1 Although the high levels of enantioselectivity
4
which have been observed indicate that the chiral 1,2-diamine
may act as a bidentate ligand in the activation of OsO4, there
has been no direct evidence of this point, and even a certain
degree of skepticism, since a 1:1 bidentate diamine-OsO4 com-
plex is, in a formal sense, a 20 outer valence electron structure.
We describe herein definitive evidence for just such a structure
and provide an analysis of the mechanistic implications.
quinuclidine. The pentacoordinate quinuclidine-OsO4 com-
plex is a distorted trigonal bipyramid in which the OsO4 subunit
geometry lies between the tetrahedral structure of OsO4 and a
regular trigonal bipyramid. The observed structures of 1‚OsO4
and quinuclidine‚OsO4 can be considered as having a degree
of bonding between N and Os which is not strong enough to
convert the OsO4 tetrahedron to a regular octahedron (for 1‚-
OsO4) or trigonal bipyramid (for quinuclidine‚OsO4). Griffith,
(
R,R)-trans-1,2-Bis(N-pyrrolidino)cyclohexane (1) was pre-
4
pared from (R,R)-trans-1,2-diaminocyclohexane by reaction in
CH2Cl2 with 2.2 equiv of 1,4-dibromobutane and 4.4 equiv of
triethylamine initially at 0 °C and then at 23 °C for 20 h.
Admixture of precooled 0.1 M solutions of OsO4 and 1 in a
ratio of 1:1 at -78 °C in rigorously dry ether and slow cooling
from -50 to -80 °C gave deep red crystals, unstable above
Skapski, et al. have already pointed out that the Os-N bond
distance of 2.37 Å observed for quinuclidine‚OsO4 is longer
than expected for a single bond and that the structures in the
4
solid and solute are very similar. It is clear that the two N-Os
bonds of 1‚OsO4 (length 2.33 Å) are also partial bonds, and to
this extent it can be argued that the 1‚OsO4 is something less
than a 20 outer valence electron complex. (For example, if the
bond order of each of the Os-N bonds were 0.5, then 1‚OsO4
could be considered as an 18-electron complex).
-
40 °C and maintained at -80 °C during isolation, mounting,
and X-ray diffraction analysis.
1H and 13C NMR studies of 1:1 mixtures of trans-1,2-bis-
(N-pyrrolidino)cyclohexane (1) or trans-1,2-bis(N-dimethylami-
no)cyclohexane (2) with OsO4 indicate a bidentate coordination
structure in solution as well as in the solid state. For example,
a 1:1 solution of 2 and OsO4 in CD2Cl2-CFCl3 reveals
diastereotopic N-Me peaks and other peaks consistent with a
The structure of the 1:1 complex of diamine 1 with OsO4 is
shown in Figure 1.3 Clearly, the diamine 1 functions as a
bidentate ligand in the complex with OsO4. However, the ligand
attachments to osmium in the hexacoordinate complex are in
an unusual arrangement which is definitely not octahedral. For
instance, the angle O(22)-Os-N(1) is 166° and not 180°, and
the angle O(22)-Os-O(21) is 106° and not 90°. A simple way
to look at the distortion from octahedral geometry becomes
evident if one considers the OsO4 subunit in the complex. That
OsO4 subunit is in between the tetrahedral geometry of
uncoordinated OsO4 and the structure for perfect octahedral
1
C2 symmetric bidentate structure: H (500 MHz, 200 K) δ 1.02
(
2H, m), 1.56 (2H, m), 1.70 (2H, m), 1.93 (2H, m), 2.31 (6H,
1
3
s), 2.67 (6H, s), 2.77 (2H, m); C (125 MHz, 200 K) δ 22.84
(
CH2), 24.17 (CH2), 40.90 (CH3), 47.74 (CH3), 64.55 (CH). In
addition, these spectra are sharp and invariant (including
chemical shifts) over the temperature range 230-183 K, arguing
strongly against a rapid equilibrium between two equivalent
5
1
13
monodentate complexes. The H and C NMR spectra (CD2-
Cl2, 195K) of a 1:1:1 mixture of trans-1,2-bis(N-pyrrolidino)-
cyclohexane (1), N-pyrrolidinocyclohexane, and OsO4 are an
exact superposition of the spectra of a 1:1 mixture of uncom-
plexed N-pyrrolidinocyclohexane and the 1:1 complex of 1 with
OsO4; no free 1 can be detected. Therefore, it follows that the
bidentate diamine-OsO4 complex is much more stable than the
monoamine-OsO4 complex in solution.
(1) (a) Yamada, T.; Narasaka, K. Chem. Lett. 1986, 131. (b) Tokles,
M.; Snyder, J. K. Tetrahedron Lett. 1986, 27, 3951. (c) Tomioka, K.;
Nakajima, M.; Koga, K. J. Am. Chem. Soc. 1987, 109, 6213. (d) Hirama,
H.; Oishi, T.; It oˆ , S. J. Chem. Soc., Chem. Commun. 1989, 665. (e) Corey,
E. J.; Jardine, P. D.; Virgil, S.; Yuen, P.-W.; Connell, R. D. J. Am. Chem.
Soc. 1989, 111, 9243. (f) Tomioka, K.; Nakajima, M.; Koga, K. Tetrahedron
Lett. 1990, 31, 6421. (g) Hanessian, S.; Meffre, P.; Girard, M.; Beaudoin,
S.; Sanc e´ au, J.-Y.; Bennani, Y. J. Org. Chem. 1993, 58, 1991.
trans-1,2-Bis(N-pyrrolidino)cyclohexane (1) is a much better
catalyst for the reaction of osmium tetraoxide with olefins than
the related monoamine N-pyrrolidinocyclohexane. Thus, in
CD2Cl2 solution at -90 °C the reaction of osmium tetraoxide
(
2) For a review of catalytic asymmetric dihydroxylation of olefins,
see: Kolb, H. C.; VanNieuwenhze, M. S.; Sharpless, K. B. Chem. ReV.
994, 94, 2483.
1
1
(
3) The deep red orthorhombic crystal of 1‚OsO4 (0.6 × 0.4 × 0.3 mm)
(9 mM) with tetramethylethylene (9 mM) was found by H
contained four molecules in the unit cell; empirical formula, C14H26N2O4-
2
NMR measurement to be at least 10 faster with trans-1,2-bis-
Os (fw 476.57); space group P212121; a ) 10.987(3), b ) 11.250(2), c )
3
3
(N-pyrrolidino)cyclohexane than with N-pyrrolidinocyclohex-
1
2.389(2) Å; V ) 1531.3(6) Å ; d(calcd) ) 2.067 g/cm , R ) â ) γ )
0°; MoKR radiation (0.71073 Å) at 193 K; 2469 reflections collected,
9
with 2001 independent reflections; refinement method full-matrix least
(4) Griffith, W. P.; Skapski, A. C.; Woode, K. A.; Wright, M. J. Inorg.
Chem. Acta 1978, 31, L413.
2
2
squares of F ; GOF on F , 1.030. Final R indices [I > 2σ(I)], R1 ) 0.0444,
wR2 ) 0.1042; R indices (all data), R1 ) 0.0534, wR2 ) 0.1104; absolute
structure parameter 0.04(3). Detailed X-ray crystallographic data are
available from the Cambridge Crystallographic Data Center, 12 Union Road,
Cambridge, CB2 1EZ, U.K.
q
(5) ∆G ) <3.5 kcal mol for an equilibrium between two hypothetical
interconverting monodentate complexes is calculated by the method of:
Bovey, F. A.; Jelinski, L.; Mirau, P. A. Nuclear Magnetic Resonance
Spectroscopy, 2nd ed.; Academic Press: New York, 1987; p 300.
S0002-7863(96)00536-7 CCC: $12.00 © 1996 American Chemical Society

7852 J. Am. Chem. Soc., Vol. 118, No. 33, 1996
Communications to the Editor
ane.6 Similarly, a faster rate of reaction (by >10 ) was observed
for trans-1,2-bis(dimethylamino)cyclohexane (2) than for N,N-
dimethylaminocyclohexane. Given that the bidentate complex
of 1 or 2 with OsO4 is the only one observed in solution and
the fact that diamines 1 or 2 are much more potent as
accelerators of the reaction of OsO4 with olefins than are
monoamines, it follows that the transition state is bidentate with
regard to the diamine.7
2
complexes via attachment of the olefinic carbons to the two
equatorial oxygens of the 1,2-diamine-OsO4 complex. Initial
olefin π-coordination to Os occurs in the geometry shown in 6
(for the reaction of 2‚OsO4-Me2CdCMe2) followed by a rapid
90° rotation about the Os-olefin axis to form the trans-dioxo
osmate adduct 4. This process is analogous to that proposed
for the enantioselective dihydroxylation of olefins by bis-
cinchona alkaloid-OsO4 complexes, except that for the bis-
cinchona case the π-complexation step is fast and reversible
Diamine 3 is an excellent ligand for the enantioselective
dihydroxylation of olefins by OsO4 at -78 °C, with enantio-
meric excess in the region of 95% for a range of olefins from
styrene to trans-3-hexene.1 These reactions are strongly
accelerated by the ligand 3. For example, trans-3-hexene
undergoes complete reaction with 3‚OsO4 at -78 °C in CH2-
Cl2 within 30 min, as compared to no reaction at all with OsO4
alone or with the analogous monoamine mesitylmethyl-R-
9
and the conversion to osmate ester is rate limiting. This change
in rate-limiting step is consistent with the higher electron density
on diamine-complexed OsO4 relative to monoamine-complexed
OsO4. In the diamine-OsO4 complex the olefin π-complex
should be destabilized and the rate of rearrangement to the
osmate ester enhanced relative to the monoamine-OsO4 system.
e
1
e
phenylethylamine.
1H NMR studies at low temperature were undertaken to
determine the nature of the primary product of the reaction of
a 1,2-diamine-OsO4 complex with an olefin. When a frozen
solution of 2‚OsO4 in CFCl3-CD2Cl2 in an NMR tube at 180
K was overlayered with tetramethylethylene in CFCl3-CD2-
Cl2, two frozen layers resulted. The temperature in the NMR
instrument was increased gradually until the layers started to
melt and mix at 210 K; the spectrum observed was that of 2‚-
OsO4 and tetramethylethylene. Within 30 s at 215 K, the
formation of the trans-dioxo osmate(VI) ester (4) resulted, as
1
judged by the new H NMR peaks which coincided with those
The dihydroxylation pathway via a rate-limiting π-complex-
ation (as for 6) and subsequent fast rearrangement to a trans-
dioxo osmate ester readily explains the observed absolute
stereochemistry of dihydroxylation with the most effective
in the spectrum of authentic 4.8 The thermodynamically less
stable cis-dioxo osmate ester 5 could not be detected, indicating
either that it is not a primary [3 + 2] adduct or that, if it is
kinetically preferred, it must isomerize to the more stable trans-
dioxo osmate ester 4 very rapidly at 215 K. In the event that
the cis-dioxo osmate ester 5 is a transitory intermediate in the
formation of the trans-dioxo osmate ester 4, the observed
absolute stereochemical course of the dihydroxylation of olefins
1c
1e
1g,10
diamines such as Tomioka’s, Jardine’s (3), and Hanessian’s.
It should be noted that the [2 + 2] cycloaddition (osmaox-
etane) pathway involving 1‚OsO4 is inconsistent with the
observed absolute stereocourse of the C2 symmetric 1,2-
11
diamine-OsO4 reaction and with quantum mechanical cal-
1
by OsO4 complexed with various C2 symmetric 1,2-diamines
12
culations.
can be explained by the [3 + 2] cycloaddition pathway which
was hypothesized earlier.1 However, it is not obvious that such
a facile isomerization of 5 to 4 is a reasonable possibility. The
most favorable pathway for the conversion 5 f 4 would appear
to be one involving dissociation of one of the Os-N bonds,
pseudorotation, and Os-N recombination. At least for the case
of the trans-dioxo osmate ester 4, Os-N dissociation appears
to be a very slow process even at 23 °C, as indicated by the
following observations: (1) no ligand exchange of 4 with
N,N,N,N-tetramethylethylenediamine over 24 h and (2) no
decomposition of 4 upon silica gel chromatography. These facts
argue that 4 may be the kinetically (as well as thermodynami-
cally) favored product, in which case a direct [3 + 2]
cycloaddition pathway predicts the wrong absolute configuration
for the osmate ester or 1,2-diol product.1 However, an indirect
pathway, which correctly predicts absolute configuration, ap-
pears deserving of serious consideration.
Structure 6, a distorted cubic octahedral ligand arrangement
about Os, is geometrically and sterically feasible for a highly
reactive, metastable intermediate or transition state. The
symmetry-allowed interaction of electrons in oxygen p-orbitals
of the equatorial OsO2 unit with π* of the olefin drives the 90°
rotation which converts 6 to 4. Thus, weak π-coordination of
the olefin to Os causes sufficient π-electron transfer to the metal
so that nucleophilic binding of oxygen to carbon is induced.
e
Acknowledgment. This research was supported by grants from the
National Institutes of Health and the National Science Foundation. We
are grateful to Dr. Axel K. Fischer for helpful advice.
Supporting Information Available: Stereopairs of the osmaox-
etane, the π-complex, and the trans-dioxo osmate ester from trans-
stilbene, OsO , and the Tomioka diamine (1 page). See any current
4
c,e
masthead page for ordering and Internet access instructions.
JA960536D
(
8) NMR spectrum of 4: 1H (500 MHz) δ 1.31 (2H, dddd, J ) 2.0, 3.0,
9
1
.4, 11.6 Hz), 1.51 (2H, dddd, J ) 2.9, 3.0, 3.3, 12.2 Hz), 1.86 (6H, s),
.88 (6H, s), 1.97 (2H, dddd, J ) 2.9, 8.2, 9.4, 11.6 Hz), 2.21 (2H, dddd,
J ) 2.0, 8.2, 8.3, 12.2 Hz), 2.83 (6H, s), 3.29 (6H, s), 3.32 (2H, dd, J )
1
3
.3, 8.3 Hz); H (500 MHz, 200 K) δ 1.02 (2H, m), 1.21 (2H, m), 1.53
(
2H, m), 1.68 (8H, m), 1.86 (2H, m), 2,77 (4H, m), 3.10 (2H, m), 3.19
(
2H, m), 3.51 (2H, m); 13C (125 MHz, 200 K) δ 22.93, 25.28, 23.92, 24.20,
4
9.52, 56.42, 65.65.
The following hypothesis is advanced for the enantioselective
(
9) Corey, E. J.; Noe, M. C. J. Am. Chem. Soc. 1996, 118, 319.
dihydroxylation of olefins by C2 symmetric 1,2-diamine-OsO4
(10) A referee has suggested that a transition state involving olefin, OsO4,
and bidentate 1,2-diamine might also be accessed from an intermediate
monodentate 1,2-diamine-OsO4 complex, without specifying structural data
for the transition state or the basis for enantioselectivity.
(11) A [2 + 2] cycloaddition pathway for the chiral 1,2-diamine-OsO4
enantioselective dihydroxylation which would have to involve one of the
equatorial oxygens clearly predicts formation of the wrong enantiomer.
(12) A. Veldkamp and G. Frenking (J. Am. Chem. Soc. 1994, 116, 4937)
predict that for diamine-coordinated OsO4 the osmaoxetane structure “is
not a minimum on the potential energy hypersurface.”
(
6) The reaction was followed by measuring the decrease in the
concentration of tetramethylethylene as a function of time.
7) Although the faster observed rates of reaction of olefins and OsO4
(
with 1,2-diamines as compared with structurally analogous monoamines
must be due at least in part to a more favorable equilibrium for 1,2-diamine
complexation, our data do not allow an assessment of the degree to which
1
,2-diamine rate enhancement may be due to a greater rate constant for the
subsequent reaction with olefin.