Isotope effects and the mechanism of epoxidation of cyclohexenone with tert-butyl hydroperoxide

Article abstract of DOI:10.1021/jo070777b

(Chemical Equation Presented) The mechanism of the epoxidation of 2-cyclohexen-1-one with tert-butyl hydroperoxide mediated by DBU was studied by a combination of experimental kinetic isotope effects (KIEs) and theoretical calculations. A large ¹²C/¹³C (k_12C/k _13C) isotope effect of ≈1.032 was observed at the C₃ (β) position of cyclohexenone, while a much smaller ¹²C/ ¹³C isotope effect of 1.010 was observed at the C₂ (α) position. Qualitatively, these results are indicative of nucleophilic addition to the enone being the rate-limiting step. Theoretical calculations support this interpretation. Transition structures for the addition step lead to predicted isotope effects that approximate the experimental values, while the predicted isotope effects for the ring-closure step are not consistent with the experimental values. The calculations correctly favor a rate-limiting addition step but suggest that the barriers for the addition and ring-closure steps are crudely similar in energy. The stereochemistry of these epoxidations is predicted to be governed by a preference for an initial axial addition, and the role of this preference in experimental diastereoselectivity observations is discussed.

Full text of DOI:10.1021/jo070777b

Isotope Effects and the Mechanism of Epoxidation of Cyclohexenone
with tert-Butyl Hydroperoxide
Chad F. Christian, Tetsuya Takeya, Michael J. Szymanski, and Daniel A. Singleton*
Department of Chemistry, Texas A&M UniVersity, P.O. Box 30012, College Station, Texas 77842
singleton@mail.chem.tamu.edu
ReceiVed April 13, 2007
The mechanism of the epoxidation of 2-cyclohexen-1-one with tert-butyl hydroperoxide mediated by
DBU was studied by a combination of experimental kinetic isotope effects (KIEs) and theoretical
calculations. A large 12C/ C (k12C/k13C) isotope effect of ≈1.032 was observed at the C
13
(â) position of
(R) position.
3
12
13
cyclohexenone, while a much smaller C/ C isotope effect of 1.010 was observed at the C
2
Qualitatively, these results are indicative of nucleophilic addition to the enone being the rate-limiting
step. Theoretical calculations support this interpretation. Transition structures for the addition step lead
to predicted isotope effects that approximate the experimental values, while the predicted isotope effects
for the ring-closure step are not consistent with the experimental values. The calculations correctly favor
a rate-limiting addition step but suggest that the barriers for the addition and ring-closure steps are crudely
similar in energy. The stereochemistry of these epoxidations is predicted to be governed by a preference
for an initial axial addition, and the role of this preference in experimental diastereoselectivity observations
is discussed.
Introduction
tages in many circumstances.,4,5 Diverse approaches to enan-
6
tioselective epoxidations have been developed, including
The epoxidation of R,â-unsaturated carbonyls with hydrop-
eroxides under basic conditions, the Weitz-Scheffer reaction,
is a valuable transformation in organic synthesis.,
mildness of the reaction conditions and the high selectivity for
the oxidation of electron-poor alkenes has allowed the applica-
tion of these reactions to diverse complex structures. The oldest
7
catalysis by polyamino acids (the Juli a´ -Colonna epoxidation)
8
and chiral phase-transfer catalysts, reactions employing chiral
alkyl hydroperoxides, and reactions using zinc reagents and
1
,2
The
9
10
O2 in the presence of a chiral alcohol.
The accepted two-step mechanism of these reactions (eq 1),
3
2a
first proposed by Bunton and Minkoff, involves the conjugate
and simplest reaction conditions utilize alkaline hydrogen
peroxide, but reactions using alkyl hydroperoxides have advan-
(
4) (a) Yang, N. C.; Finnegan, R. A. J. Am. Chem. Soc. 1958, 80, 5845-
5
848. (b) Payne, G. B. J. Org. Chem. 1960, 25, 275-276.
(
1) Weitz, E.; Scheffer, A. Chem. Ber. 1921, 54, 2327-2344.
(5) Yadav, V. K.; Kapoor, K. K. Tetrahedron 1995, 51, 8573-8584.
(
2) (a) Bunton, C. A.; Minkoff, G. J. J. Chem. Soc. 1949, 665-670. (b)
(6) For a review, see: Porter, M. J.; Skidmore, J. Chem. Commun. 2000,
Emmons, W. D.; Pagano, A. S. J. Am. Chem. Soc. 1955, 77, 89-92. (c)
Payne, G. B.; Williams, P. H. J. Org. Chem. 1961, 26, 651-659. (d) Yang,
N. C.; Finnegan, R. A. J. Am. Chem. Soc. 1958, 80, 5845-5848.
1215-1225.
(7) Juli a´ , S.; Masana, J.; Vega, J. C. Angew. Chem., Int. Ed. 1980, 19,
929-931. Colonna, S.; Molinari, H.; Banfi, S.; Juli a´ , S.; Masana, J.; Alvarez,
A. Tetrahedron 1983, 39, 1635-1641. Banfi, S.; Colonna, S.; Molinari,
H.; Juli a´ , S.; Guixer, J. Tetrahedron 1984, 40, 5207-5211. Geller, T.;
Roberts, S. M. J. Chem. Soc., Perkin Trans. 1 1999, 1397-1398. Kelly, D.
R.; Roberts, S. M. Biopolymers 2006, 84, 74-89.
(3) (a) Still, W. C. J. Am. Chem. Soc. 1979, 101, 2493-2495. (b)
Schlessinger, G. R.; Bebernitz, G. R.; Lin, P.; Poss, A. J. J. Am. Chem.
Soc. 1985, 107, 1777-1778. (c) Danishefsky, S.; Hirama, M.; Gombatz,
K.; Harayama, T.; Berman, E.; Schuda, P. F. J. Am. Chem. Soc. 1979, 101,
7
5
2
020-7031. (d) Mehta, G.; Pan, S. C. Tetrahedron Lett. 2005, 46, 5219-
223. (e) Wada, S.-i.; Tanaka, R. Bioorg. Med. Chem. Lett. 2005, 15,
966-2969. (f) Rodeschini, V.; Boiteau, J.-G.; Van de Weghe, P.; Tarnus,
(8) Helder, R.; Hummelen, J. C.; Laane, R. W. P. M.; Wiering, J. S.;
Wynberg, H. Tetrahedron Lett. 1976, 1831-1834. Wynberg, H.; Greijdanus,
B. Chem. Commun. 1978, 427-428. Wynberg, H.; Marsman, B. J. Org.
Chem. 1980, 45, 158-161.
C.; Eustache, J. J. Org. Chem. 2004, 69, 357-373.
1
0.1021/jo070777b CCC: $37.00 © 2007 American Chemical Society
Published on Web 07/03/2007
J. Org. Chem. 2007, 72, 6183-6189
6183

Christian et al.
addition of a hydroperoxy or alkylperoxy anion 2 to afford a
peroxyenolate 3. The peroxyenolate then undergoes a ring-
selectivity or enantioselectivity in these reactions. This is
particularly true for epoxidations with alkyl hydroperoxides
under typical synthetic conditions and for epoxidations of cyclic
enones, where little mechanistic information is available. We
describe here a study of the epoxidation of cyclohexenone with
tert-butyl hydroperoxide (t-BuOOH)/DBU (eq 3) using a
closing intramolecular nucleophilic substitution of C2 on the
O-O bond to afford the epoxide 4. Strong evidence for an
intermediate includes the nonstereospecificity of these reactions;
unlike epoxidations with peracids,11 the stereochemistry of the
starting alkene is not necessarily retained in the epoxide. For
example, the epoxidations of both E- and Z-3-methyl-3-penten-
combination of experimental kinetic isotope effects (KIEs) and
theoretical calculations. The results provide insight into the
stereoselectivity of these reactions and should aid the further
development of stereocontrolled epoxidations.
2
-one with basic H2O2 in methanol afford predominantly the E
1
2
epoxide product.
While the overall mechanism is well-established, a consistent
picture of the rate-limiting step has been elusive. In the reaction
of Z-3-methyl-3-penten-2-one (5) with alkaline hydrogen per-
oxide, House found that Z-E isomerization occurred at rates
Results
12
Experimental Isotope Effects. The epoxidation of an enone
using a combination of an alkyl hydroperoxide and DBU in an
aprotic solvent was first reported by Schlessinger and Poss.3b
Yadav demonstrated the utility of these conditions with a range
comparable to epoxidation (eq 2). This suggests significant
5
of R,â-unsaturated carbonyls, and they have been commonly
applied in enantioselective reactions using chiral alkyl hydro-
9
peroxides. Under the prototypical epoxidation conditions
employed here (stoichiometric DBU and t-BuOOH in dichlo-
roethane at 22 °C), cyclohexenone is converted to 8 cleanly
and essentially quantitatively.
reversibility of the initial addition under these conditions. On
the other hand, substituent effect studies of the epoxidation of
4
-aryl-3-buten-2-ones have been interpreted as favoring a rate-
13
The C KIEs (k12C/k13C) for the epoxidation of cyclohex-
13
limiting addition step. In a careful study of the epoxidation
of â-deuterium labeled phenyl vinyl ketone with hydrogen
peroxide, Kelly and Roberts found that Z-E isomerization was
much faster than epoxidation under Juli a´ -Colonna conditions
using polyleucine as catalyst.14 This strongly supports rate-
limiting ring-closure. In contrast, isomerization and epoxidation
were comparable in the absence of the polyleucine. This suggests
that the rate-limiting step may vary depending on the detailed
reaction conditions. In addition, the rates of additions of
nucleophiles to enones vary substantially with substitution, and
the rate of the ring-closure step should vary with the confor-
mational freedom in the intermediate, so enone structure may
affect the rate-limiting step.
enone were determined combinatorially by NMR methodology
15
at natural abundance. Two reactions of cyclohexenone were
taken to 83 ( 2% and 89 ( 2% conversion, and the unreacted
cyclohexenone was recovered by an extractive workup followed
by flash chromatography and fractional distillation. The samples
13
of recovered cyclohexenone were analyzed by C NMR, along
with standard samples that had not been subjected to the reaction
conditions. The change in isotopic composition in each position
was determined relative to the R-methylene carbon in cyclo-
16
hexenone, with the assumption that isotopic fractionation in
this position was negligible. From the percentage conversions
and the changes in isotopic composition, the KIEs were
15
calculated as previously described.
This uncertainty in the rate-limiting step and a lack of
knowledge of the structural characteristics of the rate-limiting
transition state thwarts a detailed understanding of diastereo-
Table 1 shows the results of two separate KIE determinations
(
each based on six sets of spectra) for each of the two reactions.
13
The independent sets of C KIEs agree within the standard
deviation of the measurements, though the KIEs here are more
variable than in previous KIE determinations on cyclohex-
enone. Despite the variability, the qualitative features of the
KIEs are apparent. Only the C2 and C3 KIEs differ significantly
from unity, with a relatively large C3 isotope effect and a smaller
KIE at C2. The qualitative interpretation of the large C3 KIE is
that the rate-limiting step involves a substantial bonding change
at C3, as would be expected for a rate-limiting addition of
(
9) (a) Adam, W.; Bheema Rao, P.; Degen, H.-G.; Saha-M o¨ ller, C. R.
J. Am. Chem. Soc. 2000, 122, 5654-5655. (b) Adam, W.; Bheema Rao,
1
7
P.; Degen, H.-G.; Saha-M o¨ ller, C. R. Eur. J. Org. Chem. 2002, 630-639.
(
c) Aoki, M.; Seebach, D. HelV. Chim. Acta 2001, 84, 187-207. (d) Lattanzi,
A.; Cocilova, M.; Iannece, P.; Scettri, A. Tetrahedron: Asymmetry 2004,
1
5, 3751-3755.
10) Enders, D.; Zhu, J.; Raabe, G. Angew. Chem., Int. Ed. 1996, 35,
725. Yu, H.-B.; Zheng, X.-F.; Lin, Z.-M.; Hu, Q.-S.; Huang, W.-S.;
(
1
Pu, L. J. Org. Chem. 1999, 64, 8149-8155.
11) For an example of stereospecificity with an electron-poor alkene,
see: Bartlett, P. A.; Chouinard, P. M. J. Org. Chem. 1983, 48, 3854-
855.
(
-
t-BuOO to cyclohexenone. The medium-sized C2 KIE is less
3
(
(
(
12) House, H. O.; Ro, R. S. J. Am. Chem. Soc. 1958, 80, 2428-2433.
(15) Singleton, D. A.; Thomas, A. A. J. Am. Chem. Soc. 1995, 117,
9357-9358.
13) Temple, R. D. J. Org. Chem. 1970, 35, 1275-1280.
14) Kelly, D. R.; Caroff, E.; Flood, R. W.; Heal, W.; Roberts, S. M.
(16) The R-carbon of cyclohexenone is more separate from other peaks
in the 13C NMR rather than the â- or γ-methylene carbons, and its use as
a standard gives a lower variance in the isotope effects.
(17) Frantz, D. E.; Singleton, D. A.; Snyder, J. P. J. Am. Chem. Soc.
1997, 119, 3383-4. Frantz, D. E.; Singleton, D. A. J. Am. Chem. Soc.
2000, 122, 3288-3295.
Chem. Commun. 2004, 2016-2017. See also: Carrea, G.; Colonna, S.;
Meek, A. D.; Ottolina, G.; Roberts, S. M. Tetrahedron: Asymmetry 2004,
5, 2945-2949. Kelly, D. R.; Roberts, S. M. Chem. Commun. 2004, 2018-
020. Mathew, S. P.; Gunathilagan, S.; Roberts, S. M.; Blackmond, D. G.
1
2
Org. Lett. 2005, 7, 4847-4850.
6184 J. Org. Chem., Vol. 72, No. 16, 2007

Epoxidation of Cyclohexenone with tert-Butyl Hydroperoxide
TABLE 1. Experimental and Predicted 13C Kinetic Isotope Effects
ert
transfer to the enolate oxygen to afford a 3-t -butylperoxycy-
(
k12C/k13C) for the Epoxidation of Cyclohexenone with
clohexenol can also be envisioned but are subject to the same
limitation.
a
t-BuOOH/DBU
C1
C2
C3
C4
C5
A low barrier is predicted for 11 to complete the epoxidation
by ring-closure via transition structure 12. The alternative
equatorial process via 13, 14, and 15 is similar but 2-3 kcal/
mol higher in energy. For the axial pathway, the addition step
is predicted to be rate-limiting, since 10 is 0.7 kcal/mol higher
in energy than 12. In contrast, the equatorial pathway would
have the ring-closure step as mainly rate-limiting. In either case,
these relative energies should not be considered as highly
reliable due to the limitations of the calculation model, the
solvent model, and the calculation method itself. A more reliable
conclusion from the calculated results is that the addition and
ring-closure steps should be roughly comparable in barrier. This
is consistent with the idea that the rate-limiting step may depend
on the detailed reaction conditions.
Experimental KIEs
expt 1
expt 2
expt 1
expt 2
average
std dev
0.998 1.010 1.033 1.005 0.996
1.003 1.009 1.027 1.002 0.999
b
b
1.010 1.037 1.002 0.995
1.010 1.029 1.008 1.003
1.001 1.010 1.032 1.004 0.998
0.000 0.004 0.003 0.004
Predicted KIEs
axial addition (10)
equatorial addition (13)
axial ring-closure (12)
1.004 1.008 1.027 0.997 0.999
1.004 1.007 1.031 1.001 1.001
1.006 1.012 0.997 0.996 0.999
equatorial ring-closure (15) 1.005 1.012 1.000 0.999 1.001
a
The experimental isotope effects were measured at 22 °C. The predicted
b
isotope effects were calculated for 25 °C. Due to practical limitations,
the NMR data was collected with a time between pulses inadequate for
relaxation of the carbonyl carbon precluding an accurate integration.
The formation of termolecular transition structures such as
10 from separate reactants should involve a substantial entropic
penalty. Including an entropy estimate based on the harmonic
frequencies at 25 °C, the free-energy barrier (at standard state)
for reaction via 10 would be 37.1 kcal/mol. This is too high by
about 12-15 kcal/mol for a reaction that proceeds in less than
a day at 25 °C. However, consideration of the following two
issues may ameliorate this problem. The first issue is that the
gas-phase entropy of compounds is decreased on dissolution in
organic solvents, typically by 15 eu. From this, the entropy
loss for forming a termolecular transition state might be expected
to be decreased compared to a gas-phase calculation by roughly
readily interpreted. A more detailed interpretation of these KIEs
will be possible with the aid of theoretical calculations.
Theoretical Mechanisms. The epoxidation of cyclohexenone
with t-BuOOH mediated by the simplified DBU model 9 was
studied in B3LYP calculations using 6-31G* and 6-31+G**
basis sets and using a PCM solvent model for dichloroethane
with full geometry optimization in all cases. The involvement
of charged species limits the reliability of a calculational
mechanistic study of this reaction by itself, but consideration
of the experimental isotope effects allows assessment of the
calculational predictions. In turn, the prediction of isotope effects
from the theoretical models facilitates a detailed interpretation
of the experimental isotope effects.
18
3
0 eu, decreasing the free-energy barrier by 9 kcal/mol. The
second issue is that t-BuOOH may form a complex with DBU.
The formation of a complex between t-BuOOH and 9 is
predicted to be downhill by 4.2 kcal/mol in dichloroethane
A series of conformationally varying pathways for the
epoxidation were explored. The diastereotopic faces of a half-
chair cyclohexenone may undergo addition of the tert-butylp-
eroxy anion in either an “axial” or an “equatorial” orientation.
(
B3LYP/6-31+G**/PCM + zpe) but is predicted to be ender-
gonic by 5.6 kcal/mol after inclusion of the gas-phase entropy
estimate. However, to roughly evaluate the energetics of
complex formation under the solution reaction conditions, 1
equiv of DBU was added to a 1 M solution of t-BuOOH in
dichloroethane at 22 °C. This leads to a temperature rise of 4
(The discussion here will consistently use the axial/equatorial
notation to denote epoxidation pathways and products, always
basing the description on the orientation of the initial addition.)
The pathways explored allowed for (1) axial versus equatorial
attack on the cyclohexenone, (2) two possible orientations of
DBU model 9, and (3) an anti (O-O-CdC dihedral angle of
approximately 180°) versus gauche (O-O-CdC approximately
°
C. If the heat capacity of the solution is estimated to be the
-1
-1 19
same as that of dichloroethane (129 J mol K ), this
temperature rise translates into a heat of mixing at 1 M of 1.5
kcal/mol, suggesting that a substantial portion of the t-BuOOH/
DBU mixture forms a complex. (Complete proton transfer is
7
5°) orientation of the attacking oxygen atoms versus the
+
unlikely because t-BuOOH and DBUH have similar pKa’s in
carbon-carbon double bond. Only the lowest-energy axial and
equatorial pathways from the B3LYP/6-31+G**/PCM calcula-
tions are presented here (Figure 1); structures and energies
associated with the full set of pathways as well as gas-phase
and 6-31G* calculations are given in the Supporting Information.
The lowest-energy calculated pathway involves an axial
addition of the peroxide with oxygens oriented anti versus the
carbon-carbon double bond, as in transition structure 10. This
transition structure leads, by a minimum-energy path, to the
zwitterionic enolate/iminium intermediate 11. As a side process,
water, but proton transfer will be disfavored in less polar
solvents.) The predicted free-energy barrier for reaction of a
t-BuOOH/9 complex with cyclohexenone is only 31.5 kcal/mol.
Considering these issues along with the limitations of the
calculated and solvent models, the calculational barrier is
reasonably consistent with the facility of the experimental
reaction.
Predicted Isotope Effects. The 13C KIEs associated with the
transition structures in Figure 1 were predicted from the scaled
20
theoretical vibrational frequencies using conventional transition
1
1 could readily undergo proton transfer to afford neutrals 9
(
18) Wilhelm, E.; Battino, R. Chem. ReV. 1973, 73, 1-9.
and 3-tert-butylperoxycyclohexanone. However, 3-tert-butylp-
eroxycyclohexanone is predicted to be 9.6 kcal/mol uphill in
free-energy (B3LYP/6-31+G**/PCM + zpe + harmonic en-
tropy and enthalpy estimates at 25 °C) from separate starting
materials. From this, 3-tert-butylperoxycyclohexanone should
not build up under the reaction conditions and its formation
must be reversible. Alternative side processes such as proton
(
19) Rastorguev, Y. L.; Ganiev, Y. A. IzV. Vyssh. Uchebn. ZaVed., Neft
Gaz. 1967, 10, 79-82.
(20) The calculations used the program QUIVER (Saunders, M.; Laidig,
K. E.; Wolfsberg, M. J. Am. Chem. Soc. 1989, 111, 8989-8994). B3LYP
frequencies were scaled by a factor of 0.9614 (Scott, A. P.; Radom, L. J.
Phys. Chem. 1996, 100, 16502-16513). The exact choice of scaling factor
makes little difference in the calculated KIEsvarying the scaling factor
from 0.93 to 0.97 changes the 13C KIEs by less than 0.001.
J. Org. Chem, Vol. 72, No. 16, 2007 6185

Christian et al.
FIGURE 1. Lowest-energy axial (top) and equatorial (bottom) pathways for the epoxidation of cyclohexenone with t-BuOOH mediated
by DBU model 9 in B3LYP/6-31+G**/PCM calculations. Energies (B3LYP/6-31+G**/PCM + zpe) are in kcal/mol relative to separate starting
materials.
state theory by the method of Bigeleisen and Mayer.21 Tunneling
corrections were applied using the one-dimensional infinite
parabolic barrier model.22 Such KIE predictions have proven
highly accurate in reactions not involving hydrogen transfer as
long as the calculation accurately depicts the mechanism and
transition state geometry.23
The results are summarized in Table 1. The addition transition
structures 10 and 13 lead to predicted KIEs of about 1.030 at
C3 and 1.008 at C2, with small isotope effects predicted for the
remaining carbons. In contrast, ring-closure transition structures
Stereoselectivity with 4-Methylcyclohexenone. The pre-
dicted preference for the axial addition pathway should have
stereochemical consequences. With cyclohexenones containing
a substituent at C4, C5, or C6, the preferred epoxide product
should arise from adding the substituent to 10 in its most
favorable position, and this would normally be assumed to have
the group in a pseudoequatorial orientation. However, an
equatorial substituent at C4 in 10 could sterically interact with
the incoming tert-butylperoxy anion, so that the preferred
transition state geometry is not obvious.
1
2 and 15 lead to predicted KIEs near unity at C3 and 1.012 at
To explore this issue, the epoxidation of 4-methylcyclohex-
enone was investigated. The three lowest-energy transition
structures for the addition step, 16-18, are shown in Figure 2.
The axial addition transition structures 16 and 17, analogues of
10, differ by having pseudoequatorial and pseudoaxial orienta-
tions of the methyl group, respectively, while 18, an analogue
of 13, is the most favored transition structure for the equatorial
addition. Notably, the lowest-energy transition structure 17
places the methyl group in a pseudoaxial position. When the
methyl group is pseudoequatorial, as in 16, the preferred
antiorientation of the attacking oxygen atoms versus the
carbon-carbon double bond is distorted to avoid interaction
of the distal oxygen with the methyl group, so that the
O-O-CdC dihedral angle is decreased from 175° in 10
to 157° in 16. With a pseudoaxial methyl group, as in 17,
the antiorientation of the peroxide is unhindered (the O-O-
CdC angle in 17 is 169°) and 17 is favored despite the
concomitant strain of the pseudoaxial methyl group. The
energy cost of proceeding through an equatorial addition,
as in 18, remains higher, in part due to a C3-C4 eclipsing
interaction present in 18 (and 13) that is not present in 16 and
17 (or 10).
C2. These predictions fit well with conventional expectations.
In the addition step, C3 is undergoing a major σ-bonding change
and is expected to exhibit a substantial normal isotope effect,
while the bonding change at C2 is more minor and the predicted
KIE, while still significantly greater than unity, is low. In the
ring-closure step, little change is occurring at C3, so the predicted
isotope effect is near unity. The predicted isotope effect for C2
at 1.012 is near the low end of primary carbon KIEs (KIEs
associated with a σ-bonding change in the rate-limiting step)
but is consistent with the relatively low KIEs observed in other
epoxidation reactions.2
3d,24
(
21) (a) Bigeleisen, J.; Mayer, M. G. J. Chem. Phys. 1947, 15, 261-
67. (b) Wolfsberg, M. Acc. Chem. Res. 1972, 5, 225-233. (c) Bigeleisen,
J. J. Chem. Phys. 1949, 17, 675-678.
22) Bell, R. P. The Tunnel Effect in Chemistry; Chapman & Hall:
London, 1980; pp 60-63.
23) (a) Beno, B. R.; Houk, K. N.; Singleton, D. A. J. Am. Chem. Soc.
996, 118, 9984-9985. (b) Meyer, M. P.; DelMonte, A. J.; Singleton, D.
2
(
(
1
A. J. Am. Chem. Soc. 1999, 121, 10865-10874. (c) DelMonte, A. J.; Haller,
J.; Houk, K. N.; Sharpless, K. B.; Singleton, D. A.; Strassner, T.; Thomas,
A. A. J. Am. Chem. Soc. 1997, 119, 9907-9908. (d) Singleton, D. A.;
Merrigan, S. R.; Liu, J.; Houk, K. N. J. Am. Chem. Soc. 1997, 119, 3385-
3
386.
(24) Singleton, D. A.; Merrigan, S. R. J. Am. Chem. Soc. 2000, 122,
1
1035-11036.
6186 J. Org. Chem., Vol. 72, No. 16, 2007

Epoxidation of Cyclohexenone with tert-Butyl Hydroperoxide
FIGURE 3. 13C KIEs for other additions to R,â-unsaturated carbonyl
compounds. (a) Base-catalyzed â-addition of dimethyl malonate in
methanol at 64 °C. See ref 25. (b) Free-radical polymerization at
6
0 °C. See ref 26.
observed C2 KIE of 1.010 is of ambiguous diagnostic value.
As noted above, epoxidation reactions tend to exhibit small
13
primary C KIEs, in part because the equilibrium isotope effect
2
3d
for an epoxidation is significantly inverse. Because of this,
the 1.010 KIE is qualitatively consistent with rate-limiting ring-
closure. However, the R-carbon in other additions to R,â-
unsaturated carbonyl compounds can exhibit a significantly
1
3
normal C KIE (Figure 3), such as 1.007 in free-radical
polymerization.
The theoretically predicted isotope effects serve to pin down
the rate-limiting step. Despite the concerns expressed above,
1
3
the predicted C3 C KIEs based on 12 and 15 for rate-limiting
13
ring-closure are near unity. The large observed C KIE of
approximately 1.030 is thus inconsistent with rate-limiting
closure of the epoxide ring. In comparing the experimental KIEs
with the predicted KIEs based on 10 and 13 for rate-limiting
addition, the agreement is not as good as typically observed in
other examples. However, considering the difficulty of ac-
curately modeling this reaction due to the involvement of
charged intermediates, along with the spread of the experimental
KIEs, the correspondence of experimental and predicted KIEs
is quite reasonable, particularly for the key C3 and C2 positions.
Overall, the KIEs very strongly support rate-limiting addition
to the enone.
FIGURE 2. Lowest-energy transition structures for the addition step
of the epoxidation of 4-methylcyclohexenone with t-BuOOH mediated
by DBU model 9 in B3LYP/6-31+G**/PCM (for dichloroethane)
calculations. Energies (B3LYP/6-31+G**/PCM + zpe) are in kcal/
mol relative to separate starting materials. Many hydrogens have been
removed for clarity.
2
3
Discussion
The natural-abundance 13C KIE determination did not work
The choice of rate-limiting step and the factors affecting this
choice may be best understood by considering the relative facile
nature of forward versus backward reactions for the enolate
intermediates such as 11 or 14. The ring-closure of 11 faces a
very low barrier because of the weakness of the O-O bond
and because its geometry is well positioned for the SN2
as well as normal for the current reaction. Aside from the
relatively high run-to-run variability in the 13C KIE, particularly
at C3 where the standard deviation is 0.004, it is disturbing to
note that the C4 and C5 KIEs were not consistently close to
their expected values of near unity. Nonetheless, the overall
pattern in the KIEs was sufficiently reproducible for the purpose
at hand of distinguishing the rate-limiting step in the epoxidation
reaction.
+
displacement by C2. A slight repositioning of the 9‚H coun-
terion to hydrogen bond with the tert-butoxy oxygen appears
to aid in catalyzing the displacement. On the other hand, the
cleavage of 11 back to starting materials is also understandably
facile. From the isotope effect analysis above, we know that
the barrier for the backward cleavage of the enolate in reality
is higher than that for ring-closure. If the ring-closure step were
Some care must be taken in the qualitative interpretation of
these isotope effects. The observed C3 KIE fits well with
conventional expectations and with the â 13C KIEs observed in
other additions to R,â-unsaturated carbonyl compounds (Figure
2
5,26
3
).
However, as we have previously reported, the presence
2
0% rate-limiting, as would be expected from the calculated
of weak bonds in reactive intermediates can result in secondary
13
13
relative energies of 10 and 12, the predicted KIE for C would
C KIEs ( C KIEs that are not the result of a σ-bonding change
3
be 1.022, which is at the limit of consistency with experiment.
in the rate-limiting step) that are so large as to mimic primary
1
3
27
From this, it appears that the difference in energy between
addition and ring-closure transition states in reality is at least
as high as the calculated difference between 10 and 12.
The higher energy of 14 along the equatorial pathway versus
C KIEs. Because the C3-O bond in intermediates resembling
3
should be weak, the KIE to be expected at C3 if ring-closure
were the rate-limiting step was not clear. In addition, the
1
1 on the axial pathway may be ascribed to a near eclipsing
(
25) Singleton, D. A.; Schulmeier, B. E. J. Am. Chem. Soc. 1999, 121,
313-9317.
26) Singleton, D. A.; Nowlan, D. T., III; Jahed, N.; Matyjaszewski, K.
Macromolecules 2003, 36, 8609-8616.
27) Singleton, D. A.; Wang, Z. J. Am. Chem. Soc. 2005, 127, 6679-
685.
9
interaction of C3 and C4 in 14 versus the nearly perfectly
staggered 11. The barrier for backward cleavage of 14 via 13
is slightly lower than ring-closure via 15, so that ring-closure
in this case would be predicted to be mainly rate-limiting. The
(
(
6
J. Org. Chem, Vol. 72, No. 16, 2007 6187

Christian et al.
difference in predicted rate-limiting steps for the axial and
equatorial pathways may be thought of as resulting from the
ease of backward cleavage of 14 via 13, since this loses the
strain of the eclipsing interaction.
With the knowledge that the addition step is rate-limiting,
consideration of the calculated transition structures provides
insight into the stereochemistry of these epoxidations with
substituted cyclohexenones. For 5-substituted cyclohexenones,
there is a substantial preference for both hydrogen peroxide-
and alkyl hydroperoxide-mediated reactions to afford axial
2
8,29
epoxidation.
(We are again defining “axial” epoxidation
based on the orientation of the initial addition, assuming that
substituents on the cyclohexenone have been placed in a
preferred pseudoequatorial position.) This is illustrated by the
5
,29
widely used epoxidation of carvone (19) to afford 20. 6-
3
0
Substituted cyclohexenones also tend to be epoxidized axially,
3
1
though there has been one report of an exception. The
observation of axial epoxidation in these cases is in agreement
with the calculated preference for axial transition structure 10
over 13 or 15 in the equatorial pathway.
The epoxidation of some other substituted cyclohexenones
has until now been stereochemically more enigmatic. With
4
-monosubstituted cyclohexenones, such as 21 and 23, there is
a tendency for epoxidation to occur on the face opposite the
-substituent(22 and 24).32 This would be considered equatorial
4
epoxidation if the 4-substituent is pseudoequatorial during the
reaction. However, when the 4-substituent is held rigidly
3
3
pseudoequatorial, as in 26, axial epoxidation predominates.
This strange reversal of stereochemistry between 23 and 26 is
difficult to explain if the methyl group in 23 remains pseu-
doequatorial. However, the complete set of observations is
understandable from the theoretical prediction that the preferred
transition state has the 4-substituent flipped to axial, as in 17.
When the 4-substituent is equatorial, there is a substantial steric
interaction with the distal oxygen of an incoming hydroperoxy
or tert-butylhydroperoxy anion. This hindered attack would lead
to the minor product 25 from 23. The major product 24 could
potentially be formed in two ways, either by placing the methyl
group pseudoaxial in an axial transition state analogous to 10,
as in 17, or by placing the methyl group pseudoequatorial in
an equatorial pathway analogous to 13/15, as in 18. The energy
ring-flip is possible, and the general preference for axial
epoxidation affording 27 is likely augmented by a steric effect
of the axial bridgehead methyl group inhibiting the rate-limiting
ring-closure step for the nonobserved equatorial epoxidation.
Conclusions
In epoxidations of electron-deficient alkenes mediated by
hydroperoxides under basic conditions, reactivity and stereo-
selectivity will be controlled by the energy of the transition state
for the rate-limiting step. Either the addition step or the ring-
closure step may in principle be rate-limiting, with no obvious
qualitative reason to favor one over the other. In the work of
Kelly and Roberts on the epoxidation of phenyl vinyl ketone
under Juli a´ -Colonna conditions, it appears clear that the ring-
34
cost of placing the methyl group pseudoaxial, while significant,
is still less than the substantial cost of the equatorial pathway,
and the axial epoxidation pathway remains preferred. In 26, no
1
4
(
28) (a) Gravel, D.; Bordeleau, J. Tetrahedron Lett. 1998, 39, 8035-
038. (b) Kreiser, W.; K o¨ rner, F. HelV. Chim. Acta 1999, 82, 1610-1629.
c) Hanazawa, T.; Wada, T.; Masuda, T.; Okamoto, S.; Sato, F. Org. Lett.
001, 3, 3975-3977. (d) Hatcher, M. A.; Peleg, S.; Dolan, P.; Kensler, T.
W.; Sarjeant, A.; Posner, G. H. Bioorg. Med. Chem. 2005, 13, 3964-3976.
29) (a) Daniewski, A. R.; Garofalo, L. M.; Hutchings, S. D.; Kabat, M.
M.; Liu, W.; Okabe, M.; Radinov, R.; Yiannikouros, G. P. J. Org. Chem.
closure step is rate-limiting. Here, in the first detailed
mechanistic study with a cyclic enone, the experimental kinetic
isotope effects supplemented by theoretically predicted isotope
effects strongly support the addition step as being rate-limiting.
The theoretical calculations here correctly predict that the
addition step should be rate-limiting, and it is notable that the
literature stereochemical observations are generally consistent
with a preferred rate-limiting axial addition step as stereochemi-
cally controlling. However, the similarity of barriers for the
addition and ring-closure steps should be recognized. Because
of this, it is reasonable to suppose that the rate-limiting step
may change depending on the detailed reaction conditions and
the substrate structure. Indeed, the rate-limiting step for the
equatorial epoxidation pathway is predicted to be ring-closure
instead of addition. This overall mechanistic uncertainty com-
plicates the understanding and control of diastereoselectivity
or enantioselectivity in these reactions, as it may be uncertain
what combination of transition states along major and minor
8
(
2
(
2
002, 67, 1580-1587. (b) Okamura, W. H.; Aurrecoechea, J. M.; Gibbs,
R. A.; Norman, A. W. J. Org. Chem. 1989, 54, 4072-4083. (c) Pogrebnoi,
S.; Saraber, F. C. E.; Jansen, B. J. M.; de Groot, A. Tetrahedron 2006, 62,
1
743-1748. (d) Zhao, X. Z.; Tu, Y. Q.; Peng, L.; Li, X. Q.; Jia, Y. X.
Tetrahedron Lett. 2004, 45, 3713-3716. (e) Mehta, G.; Kumaran, R. S.
Tetrahedron Lett. 2003, 44, 7055-7059.
(
30) Solladie, G.; Hutt, J. J. Org. Chem. 1987, 52, 3560-3566.
(
31) Barrett, A. G. M.; Capps, N. K. Tetrahedron Lett. 1986, 27, 5571-
5
574.
(32) (a) Rodeschini, V.; Van de Weghe, P.; Salomon, E.; Tarnus, C.;
Eustache, J. J. Org. Chem. 2005, 70, 2409-2412. (b) Hoke, M. E.; Brescia,
M.-R.; Bogaczyk, S.; DeShong, P.; King, B. W.; Crimmins, M. T. J. Org.
Chem. 2002, 67, 327-335.
(33) Moreno-Dorado, F. J.; Guerra, F. M.; Aladro, F. J.; Bustamante, J.
M.; Jorge, Z. D.; Massanet, G. M. Tetrahedron Lett. 1999, 55, 6997-7010.
(34) Rickborn, B.; Lwo, S.-Y. J. Org. Chem. 1965, 30, 2212-2216.
6188 J. Org. Chem., Vol. 72, No. 16, 2007

Epoxidation of Cyclohexenone with tert-Butyl Hydroperoxide
pathways is deciding the selectivity. However, knowledge of
the rate-limiting step provides insight into the stereochemistry
of epoxidations of interest, and this information is readily
available from kinetic isotope effects.
45:55 mixture of 2-cyclohexenone and 2,3-epoxy-2-cyclohexenone.
This fraction was chromatographed on a 4 × 30 cm flash silica gel
column using 1:1 chloroform/hexanes as eluent to afford 0.5 g of
2-cyclohexenone (>99% purity by GC) along with mixtures of
2-cyclohexenone and 2,3-epoxy-2-cyclohexenone in various ratios.
A second reaction performed by an analogous procedure proceeded
to 89.1% conversion.
Experimental Section
Epoxidation of 2-Cyclohexenone with tert-Butyl Hydroper-
oxide and DBU. A mixture of 19.2 g (0.20 mol) of 2-cyclohex-
enone, 3.4 g of dodecane (internal standard), 27.6 g (0.16 mol) of
DBU, and 50 mL (0.21 mol) of 4.1 M t-BuOOH in dichloroethane35
in 600 mL of dichloroethane was prepared at 0 °C and allowed to
warm slowly to 22 °C. Aliquots were periodically removed, and
the conversion was checked by GC. After 18 h at 22 °C, the
conversion was determined to be 83.4%. The reaction mixture was
diluted with 300 mL of chloroform and 300 mL of water and stirred
NMR Measurements. NMR samples were prepared using 200
mg of cyclohexenone in a 5 mm NMR tube filled to a 5 cm sample
height with CDCl
3
. The 13C spectra were recorded at 100.5 MHz
using inverse gated decoupling, 60 s delays, and a 5.0 s acquisition
time to collect 400 000 points. Integrations were determined
numerically using a constant equal integration region for peaks
compared. A zeroth-order baseline correction was generally applied,
but in no case was a first-order (tilt) correction applied. Six spectra
were obtained for each of two independent samples of cyclohex-
enone. The raw integration results are shown in the Supporting
Information.
for 30 min. The organic layer was separated, washed with H
dried (anhydrous MgSO ), and concentrated under vacuum below
2
O,
4
3
0 °C. The residue was then chromatographed on a 5 × 45 cm
flash silica gel column using 10% ethyl acetate/30-60% petroleum
ether as eluent to afford 9.5 g of a mixture of 2-cyclohexenone
and 2,3-epoxy-2-cyclohexenone in the approximate ratio of 17:83.
Vacuum transfer of the volatiles from this mixture on a water
aspirator followed by fractional distillation using a 10 cm Vigreaux
column afforded a 5.2 g fraction (bp 90-95 °C) of an approximately
Acknowledgment. We thank NIH grant No. GM-45617 and
the Robert A. Welch Foundation for support of this research.
Supporting Information Available: Energies and full geom-
etries of all calculated structures, and NMR integration results for
all reactions. This material is available free of charge via the Internet
at http://pubs.acs.org.
(35) Sharpless, K. B.; Verhoeven, T. R. Aldrichchimica Acta 1979, 12,
6
3-74.
JO070777B
J. Org. Chem, Vol. 72, No. 16, 2007 6189