6
176 J . Org. Chem., Vol. 64, No. 17, 1999
Limberg and Wistuba
coordination sphere has been structurally characterized
so far, while pentacoordinated compounds are known.12,32
Consequently, a complex 2* may not be as favorable as
the calculation of its energy suggests. First of all, entropy
Exp er im en ta l a n d Com p u ta tion a l Deta ils
Com p u ta tion a l Meth od . Density functional calculations
were carried out with the Gaussian/DFT34 series of programs.
The structures were optimized with the LanL2DZ basis set
using in each case the B3LYP formulation of density functional
theory, i.e., Becke’s three-parameter exchange functional and
the Lee-Yang-Parr correlation functional. The simulations
of the NMR spectra of 3a and 3b were performed using Win
Daisy.
VI
changes have to be considered: Cr is very small, and
while a fifth ligand still has a lot of freedom (the
structures of three different conformational isomers were
optimized for 2, and their energy differences were not
higher than 2 kJ /mol), a sixth one certainly does not.
Frequency calculations allowed the determination of the
entropies, leading to a ∆G208 value of only -16 kJ /mol
Exp er im en ta l P r oced u r e. NMR spectra of solutions were
recorded using a Bruker AC 200 instrument operating at 200
MHz. CrO Cl (99.99% pure), propylene oxide (99% pure), and
2 2
2 2
for the reaction of two molecules of 1 to 2* and CrO Cl .
cis- and trans-2-butylene (96% pure) were supplied by Aldrich,
and ethylene oxide (99.8%) was supplied by Fluka. The
reagents were further purified by fractional distillation in
vacuo by passage through traps maintained at successively
lower temperatures prior to use. The traps were cooled by
immersion in liquid nitrogen, acetone/dry ice, or “slush” baths
each composed of a suitable organic material in solid-liquid
equilibrium.
This would still lead to considerable disproportion, but
the error of these calculations has to be considered at
that stage: the ∆G value is the result of three structure
optimizations and three frequency calculations. Bearing
in mind the results obtained with the same method for
the conformers of 3a , it is easy to imagine that the error
for the ∆G value mentioned above can fall into a region
which is decisive for the positioning of the corresponding
equilibrium. We therefore still consider 2 as the major
species in solution before the actual reaction, in agree-
ment with the experimental findings. However, as al-
ready suggested by those, the calculations support the
relevance of 2* for the consideration of the equilibria in
solution. The same is true for an oxirane complex of 1
The NMR samples were prepared by means of a vacuum
-3
line (at a background pressure of <10 mbar). Chromyl
chloride dissolves in, and reacts with, hydrocarbon-based
greases, and so recourse was made to a combination of Young’s
greaseless valves and ground-glass joints lubricated typically
with Voltalef fluorocarbon grease. Since all reactants employed
were sufficiently volatile, their stoichiometries could be ad-
justed with use of standard manometric techniques. Typically
the epoxide was allowed to enter the line at a certain pressure
and was then condensed at the bottom of a Pyrex NMR tube.
The appropriate amount of CD Cl was distilled on top of it.
(
6), the structure of which was optimized, too. A com-
3
3
plexation energy of -29.2 kJ /mol was deduced, which
is naturally less negative than the one found for chromyl
chloride as the alkoxide ligand lowers the Lewis acidity
of the chromium center. The energy of reaction i was
calculated to amount to -147.9 kJ /mol (the most stable
conformation of 1 calculated was considered), which is a
reasonable driving force for it to proceed even at very
low temperatures. Surprisingly, reaction v (Scheme 9)
was found to be more exothermic (-161.8 kJ /mol) than
reaction iv (-138.8 kJ /mol), so that for purely thermo-
dynamic reasons the formation of the chromatdiester (Od
2
2
Subsequently the line was filled with chromyl chloride at the
pressure required, and this quantity was added to the NMR
tube via condensation, too. The latter was flame sealed and
placed in a -90 °C slush where it was allowed to anneal to
that temperature. After its contents had liquefied, they were
mixed by shaking of the tube at that temperature, and the
tube was quickly transferred into the NMR spectrometer which
had been precooled to -95 °C. Subsequently the temperature
was adjusted to -80 °C and the first spectrum recorded.
For clarification of the stereochemistry of the cleavages,
typically 1 mmol of epoxide and 1 mmol of chromyl chloride
were co-condensed together with dichloromethane (20 cm ),
3
)
2
Cr(OCH
2
CH
2
Cl)
2
, 5, should have been expected. The
and the mixture was warmed to -50 °C and allowed to stir
for 2 h. Subsequently 10 cm3 of water was added at that
temperature by means of a syringe and the whole mixture
annealed to room temperature under constant shaking of the
reaction vessel. The layers were separated, and the aqueous
reason for reaction v not being observed must therefore
have its origin in the kinetics. There are some transition
states which dependent on Scheme 4 would lead to 5, and
it is easy to find reasons why these should not be formed.
However, as such a discussion would be purely specula-
tive, we will refrain from going into detail here.
3
one was extracted three times with 5 cm of diethyl ether. All
organic phases were combined and investigated by gas chro-
matography on a Chrompack CP 9002. Evaporation of the
solvents yielded the chlorohydrins which could be investigated
spectroscopically. Authentic samples were prepared by co-
condensing stoichiometric amounts of the epoxides in diethyl
In conclusion the mechanism depicted in Scheme 4 is
in good agreement with all experimental findings and
therefore provides a picture explaining the observations
made and elucidating the critical features of the mech-
anism for ring-opening of epoxides by chromyl chloride
at low temperatures. Current research concerns the
synthesis of stable derivatives of 1, making use of the
information gained in the present study, as such com-
pounds are expected to show a higher selectivity in
oxidation reactions than chromyl chloride.
2 4
ether with gaseous HCl, which had been dried over H SO .
After 1 h of reaction time at -30 °C the mixtures were
annealed to room temperature and investigated as described
above.
Ack n ow led gm en t. Financial support by the Deut-
sche Forschungsgemeinschaft (scholarship (C.L.) and
funding) and the Fonds der Chemischen Industrie is
gratefully acknowledged. Further, we thank A. Frick for
the NMR simulation as well as Professor G. Huttner
for his generous support.
(
33) The formation of 6 will also contribute to the deviations found
for the later periods of the kinetic investigations.
34) Gaussian 94, Revision E.2: Frisch, M. J .; Trucks, G. W.;
(
Schlegel, H. B.; Gill, P. M. W.; J ohnson, B. G.; Robb, M. A.; Cheeseman,
J . R.; Keith, T.; Petersson, G. A.; Montgomery, J . A.; Raghavachari,
K.; Al-Laham, M. A.; Zakrzewski, V. G.; Ortiz, J . V.; Foresman, J . B.;
Peng, C. Y.; Ayala, P. Y.; Chen, Wong, W. M. W.; Andres, J . L.;
Replogle, E. S.; Gomperts, R.; Martin, R. L.; Fox, D. J .; Binkley, J . S.;
Defrees, D. J .; Baker, J .; Stewart, J . P.; Head- Gordon, M.; Gonzalez,
C.; Pople, J . A. Gaussian, Inc., Pittsburgh, PA, 1995.
Su p p or tin g In for m a tion Ava ila ble: Structural and spec-
troscopic data concerning the optimized structures in Figure
4. This material is available free of charge via the Internet at
http://pubs.acs.org.
J O982357F