Theoretical and experimental studies on the Baeyer-Villiger oxidation of ketones and the effect of α-halo substituents

Article abstract of DOI:10.1021/jo0513966

The Baeyer-Villiger reactions of acetone and 3-pentanone, including their fluorinated and chlorinated derivatives, with performic acid have been studied by ab initio and DFT calculations. Results are compared with experimental findings for the Baeyer-Villiger oxidation of aliphatic fluoro and chloroketones. According to theoretical results, the first transition state is rate-determining for all substrates even in the presence of acid catalyst. Although the introduction of acid into the reaction pathway leads to a dramatic decrease in the activation energy for the first transition state (TS), once entropy is included in the calculations, the enthalpic gain is lost. Of all substrates examined, pentanone reacts with performic acid via the lowest energy transition state. The second transition state is also lowest for pentanone, illustrating the accelerating effect of the additional alkyl group. Interestingly, there is only a small energetic difference in the transition states leading to migration of the fluorinated substituent versus the alkyl substituent in fluoropentanone and fluoroacetone. These differences match remarkably well with the experimentally obtained ratios of oxidation at the fluorinated and nonfluorinated carbons in a series of aliphatic ketones (calculated, 0.3 kcal/mol, observed, 0.5 kcal/mol), which are reported herein. The migration of the chlorinated substituent is significantly more difficult than that of the alkyl, with a difference in the second transition state of approximately 2.6 kcal/mol.

Full text of DOI:10.1021/jo0513966

Theoretical and Experimental Studies on the Baeyer-Villiger
Oxidation of Ketones and the Effect of r-Halo Substituents
Friedrich Grein,*,† Austin C. Chen,‡ David Edwards,§ and Cathleen M. Crudden*,§
Department of Chemistry, UniVersity of New Brunswick, Fredericton, New Brunswick, E3B 6E2, Canada,
and Department of Chemistry, Queen’s UniVersity, Kingston, Ontario K7L 3N6, Canada
fritz@unb.ca; cruddenc@chem.queensu.ca
ReceiVed July 6, 2005
The Baeyer-Villiger reactions of acetone and 3-pentanone, including their fluorinated and chlorinated
derivatives, with performic acid have been studied by ab initio and DFT calculations. Results are compared
with experimental findings for the Baeyer-Villiger oxidation of aliphatic fluoro and chloroketones.
According to theoretical results, the first transition state is rate-determining for all substrates even in the
presence of acid catalyst. Although the introduction of acid into the reaction pathway leads to a dramatic
decrease in the activation energy for the first transition state (TS), once entropy is included in the
calculations, the enthalpic gain is lost. Of all substrates examined, pentanone reacts with performic acid
via the lowest energy transition state. The second transition state is also lowest for pentanone, illustrating
the accelerating effect of the additional alkyl group. Interestingly, there is only a small energetic difference
in the transition states leading to migration of the fluorinated substituent versus the alkyl substituent in
fluoropentanone and fluoroacetone. These differences match remarkably well with the experimentally
obtained ratios of oxidation at the fluorinated and nonfluorinated carbons in a series of aliphatic ketones
(calculated, 0.3 kcal/mol, observed, 0.5 kcal/mol), which are reported herein. The migration of the
chlorinated substituent is significantly more difficult than that of the alkyl, with a difference in the second
transition state of approximately 2.6 kcal/mol.
1. Introduction
First the peracid attacks the carbonyl carbon, leading to a
hemiperacetal (also known as the Criegee intermediate), and
then one of the adjacent carbon-carbon bonds migrates to the
perester oxygen, reforming the carbonyl with loss of a proton
and cleavage of the O-O bond (eq 1).
The Baeyer-Villiger reaction,1 in which a ketone is converted
into an ester or lactone upon treatment with a peracid, is a
valuable reaction because of the importance of the products,
the uniqueness of the transformation, and the difficulty of
accomplishing it by other means.2,3 The basic mechanism of
the reaction was described by Criegee more than 50 years ago.4
† University of New Brunswick.
‡ Current address: Merck Frosst Center for Therapeutic Research, Montreal,
PQ, Canada.
§ Queen’s University.
(1) Baeyer, A.; Villiger, V. Chem. Ber. 1899, 32, 3625.
10.1021/jo0513966 CCC: $33.50 © 2006 American Chemical Society
Published on Web 01/11/2006
J. Org. Chem. 2006, 71, 861-872
861

Grein et al.
In unsymmetrical ketones, migration of the two substituents
leads to different products. The relative ease of migration of a
given substituent is called its migratory aptitude. It has been
empirically determined that substituents that are best able to
stabilize a positive charge migrate preferentially and are said
to have higher migratory aptitudes.2 Despite the utility of this
rule for predicting regioselectivity, relatively subtle differences
in structure can have significant effects on the migratory
aptitudes of carbonyl substituents.5-7 Since being able to predict
the position of oxidation is crucial for the application of this
reaction to the synthesis of complex organic compounds,
understanding migratory aptitudes is critical.
reaction, autocatalysis may be observed, and the rate-determin-
ing step may change as acid is likely to affect the two steps
differently.
In an effort to understand several of the key aspects of the
Baeyer-Villiger reaction, including which step is rate-determin-
ing, what the effect of acid is on both steps, and what some of
the controlling factors with respect to migratory aptitudes are,
we embarked upon a combined theoretical and experimental
study of the energetics of the Baeyer-Villiger oxidation of
ketones. We were particularly interested in the effect of halogen
substitutents on the relative activity and migratory aptitudes
since two of us recently communicated a study in which
surprising results were obtained in a series of fluorinated
ketones.7 Specifically, we showed that the spatial orientation
of a fluorine substituent has a dramatic effect on the migratory
aptitudes of the ketone substituents.
Thus, we examined the oxidation of acetone, 3-pentanone,
fluoroacetone, chloroacetone, 2-fluoropentanone, and 2-chlo-
ropentanone via theoretical calculations using perfomic acid as
the oxidant. A variety of fluorinated and chlorinated aliphatic
ketones such as 3-fluoro and 3-chloroheptanone, among others,
were studied experimentally using mCPBA20 as the oxidant.21,22
Considering the importance of the Baeyer-Villiger reaction,
it is surprising how few theoretical studies of this reaction have
appeared. In 1997, Cardenas et al. reported ab initio studies on
the oxidation of acetone with performic acid.23 Only the second
step of the reaction was examined. At the MP2/6-31G** level,
the TS was found to lie 30 kcal/mol above the Criegee
intermediate. In agreement with previous mechanistic studies,24
it was found that cleavage of the oxygen-oxygen bond occurs
concurrently with migration of the methyl group, whereas loss
of the proton is delayed relative to that of the methyl group.
In a subsequent article, Cardenas et al.25 used semiempirical
methods (AM1 and PM3) to calculate the structure of the
Criegee intermediate and the transition state for migration in
the reaction of acetone with various peracetic and perbenzoic
acids. Relative rate constants, as evaluated from the activation
energies, correlated quite well with the experimentally deter-
mined Hammett electronic constants, supporting the notion that
the second step of the reaction is rate-determining.
In addition to the uncertainty involved in predicting relative
migratory aptitudes, the question of whether the first or second
step of the reaction is rate-determining is also under debate. In
the 1950s, Hawthorne and Emmons9 proposed that the reaction
occurs by a rate-limiting, acid-catalyzed migration. Simamura
and co-workers came to a similar conclusion after their kinetic
studies.10 Palmer and Fry, however,11 cited the observation of
kinetic isotope effects (14C) and a negative F value for the
Hammett plot of a series of p-substituted acetophenone sub-
strates as conclusive evidence for a rate-determining migration.
However, others have interpreted the 14C labeling studies as
evidence of a switch from rate-determining migration to rate-
determining addition to the carbonyl depending on the electron-
withdrawing or -donating nature of the ketone substituents.2,3,12-14
Ogata also interpreted extensive kinetic data for the oxidation
of substituted benzaldehydes in terms of a switch in the rate-
determining step from migration to addition.14 Recent compu-
tational studies by Reyes et al. support the switch in mechanism
from rate-determining migration for electron-withdrawing ke-
tones to rate-determining addition for electron-rich ketones.12
Intra and intermolecular kinetic isotope effects in the oxidation
of cyclohexanone provide strong evidence that the first step is
rate-determining in the systems examined in this study.15
The effect of acid on the reaction must also be considered.16,17
Doering and Speers18 showed that in the presence of sulfuric
acid, only 30 min are required for complete conversion,
compared to 8 days in the absence of acid.19 Kinetic studies by
Friess and Soloway also showed that the reaction was acid-
catalyzed.8 Since acid is generated as a byproduct during the
In the most comprehensive study to date, Okuno26 examined
the Baeyer-Villiger reaction of p-anisaldehyde and benzalde-
hyde with performic acid using the MP2/6-31G**//HF/3-21G
method. The transition state for the formation of the Criegee
intermediate, TS1, was also included in this study, as was the
effect of acid catalyst.27 It should be noted, however, that the
Baeyer-Villiger reaction of benzaldehdyes generally proceeds
by migration of hydride rather than the aryl group. Benzaldehyde
itself, p-Cl, p-, and m-NO2 benzaldehyde give exclusively
hydride migration regardless of the pH, and p-anisaldehyde gives
hydride migration at high pH and aryl migration at low pH.12
In the case of aliphatic aldehydes, the ratio of hydride to alkyl
(2) Krow, G. R. Org. React. 1993, 43, 251.
(3) Renz, M.; Meunier, B. Eur J. Org. Chem. 1999, 737.
(4) Criegee, R. Justus Liebigs Ann. Chem. 1948, 560, 128.
(5) Harmata, M.; Rashatasakhon, P. Tetrahedron Lett. 2002, 43, 3641.
(6) Harmata, M.; Rashatasakhon, P. Org. Lett. 2000, 2, 2913.
(7) Crudden, C. M.; Chen, A. C.; Calhoun, L. A. Angew. Chem., Int.
Ed. 2000, 39, 2851.
(8) Friess, S. L.; Soloway, A. H. J. Am. Chem. Soc. 1951, 73, 3968.
(9) Hawthorne, M. F.; Emmons, W. D. J. Am. Chem. Soc. 1958, 80,
6398.
(10) Mitsuhashi, T. M.; Miyadera, H.; Simamura, O. J. Chem. Soc.,
Chem. Commun. 1970, 1301.
(11) Palmer, B. W.; Fry, A. J. Am. Chem. Soc. 1970, 92, 2580.
(12) Reyes, L.; Castro, M.; Cruz, J.; Rubio, M. J. Phys. Chem. A 2005,
109, 3383.
(13) (a) Ogata, Y.; Sawaki, Y. J. Org. Chem. 1969, 34, 3985. (b) Ogata,
Y.; Sawaki, Y. J. Org. Chem. 1972, 37, 2953.
(14) Ogata, Y.; Sawaki, Y. J. Am. Chem. Soc. 1972, 94, 4189.
(15) Singleton, D. A.; Szymanski, M. J. J. Am. Chem. Soc. 1999, 121,
9455.
(16) Friess, S. L. J. Am. Chem. Soc. 1949, 71, 14215.
(17) Friess, S. L.; Farnham, N. J. Am. Chem. Soc. 1950, 72, 5518.
(18) Doering, W. v. E.; Speers, L. J. Am. Chem. Soc. 1950, 72, 5515.
(19) The extent of catalysis was substrate-dependent, and the results were
complicated by instability of the products in the presence of acid.
(20) mCPBA ) m-chloroperbenzoic acid.
(21) Kitazume, T.; Kataoka, J. J. Fluorine Chem. 1996, 80, 157.
(22) Itoh, Y.; Yamanaka, M.; Mikami, K. Org. Lett. 2003, 5, 4803.
(23) Cardenas, R.; Cetina, R.; Lagunez-Otero, J.; Reyes, L. J. Phys.
Chem. A 1997, 101, 192.
(24) Doering, W. v. E.; Speers, L. J. Am. Chem. Soc. 1950, 72, 5515.
(25) Cardenas, R.; Reyes, L.; Lagunez-Otero, J.; Cetina, R. J. Mol. Struct.
Theochem 2000, 497, 211.
(26) Okuno, Y. Chem.-Eur. J. 1997, 3, 212.
(27) All of these values are ∆Hliq values.
(28) Anoune, N.; Hannachi, H.; Lanteri, P.; Longeray, R.; Arnaud, C. J.
Chem. Educ. 1998, 75, 1290.
862 J. Org. Chem., Vol. 71, No. 3, 2006

Studies of Baeyer-Villiger Oxidation of Ketones
SCHEME 1. Uncatalyzed Baeyer-Villiger Reaction of Ketones 1a-f with Performic Acid
migration was found to depend critically on the solvent.30 These
facts unfortunately complicate correlation of the theoretical
results with experimental observations since only aryl migration
was modeled in the Okuno study. Theoretical studies of the
oxidation of aldehydes with peroxycarboxylic acids,28 of acetone
with hydrogen peroxide as oxidant,29 and of the effect of solvent
on the oxidation of aldehydes30 have also appeared.
Our study constitutes the first example of a full theoretical
treatment of both steps of the Baeyer-Villiger oxidation of a
ketone with a peracid, which is the method most commonly
employed in synthesis. In addition, we report experimental
studies on the Baeyer-Villiger oxidation of several representa-
tive fluoro and chloroketones, providing significant experimental
support for the conclusions drawn in the theoretical study.
been chosen as solvent. The polarizable continuum model
(PCM) method,33,34 as implemented in Gaussian 98, has been
applied to the gas-phase-optimized geometries, with the same
DFT functional and basis set (B3LYP/6-31++G**) employed
before.
General Reaction Profile. The reaction of acetone (1a) with
performic acid (2) produces methyl acetate (7a) and formic acid
(9) via the Criegee intermediate (4) (Scheme 1). Similarly,
oxidation of 3-pentanone (1d) produces ethylpropionate (7d).
For the halogenated substrates (1b, c, e, and f), migration of
each substituent gives rise to a different product via different
transition states. Inclusion of formic acid in the reaction scheme
has a significant effect on the geometry of the transition states
and energetics of the reaction (Scheme 2).35
As shown in Scheme 2, formic acid acts as a hydrogen bond
donor and acceptor during the addition of performic acid to the
ketone. This changes the geometry of the addition from a
strained four-membered ring to a more favorable six-membered
ring. The involvement of formic acid in the second step is
similar, but in this case, the ring size changes from a seven-
membered ring (uncatalyzed) to a nine-membered ring (cata-
lyzed).
Results for Acetone, Fluoroacetone, and Chloroacetone.
In Table 1, the energy, enthalpy, and free energy differences,
relative to the reactants, of the Criegee intermediates, the two
transition states, and the products are given for acetone,
fluoroacetone, and chloroacetone in the presence and absence
of acid.
For the substrates examined, ∆E and ∆H differ by no more
than 2 kcal/mol. However, significant differences exist between
∆E and ∆G. For the Criegee intermediate, ∆G is generally about
15 kcal/mol higher than ∆E, and for the uncatalyzed transition
states these values are 10-13 kcal/mol higher. Such systematic
differences are mainly accounted for by differences in transla-
tional and rotational entropy. In the acid-catalyzed reactions,
another 10-15 kcal/mol is added to ∆E because of the inclusion
2. Theoretical Studies
Computational Methods. Geometry optimizations were
performed with the Gaussian 98 programs,31 using Hartree-
Fock (HF), second-order Moeller-Plesset perturbation (MP2),
and density functional theory (DFT) methods with the B3LYP32
hybrid functional, at various levels of basis set. Results will be
reported for the 6-31++G** basis set having diffuse functions
and polarization functions both on the hydrogens and the first-
row atoms.
In the following sections, energy differences ∆E, enthalpy
differences ∆H, and Gibbs free energy differences ∆G, all
relative to the reactants, will be presented. The ∆E’s are
differences in the calculated total (electronic) energies, without
adjustment for zero-point energies (ZPE) or thermal effects. The
enthalpy changes, ∆H, and Gibbs free-energy changes, ∆G,
were calculated for standard temperature and pressure. ZPE
corrections are included in ∆H and ∆G, but not in ∆E.
Results will also be given for energies and Gibbs free energies
of solvation. For this purpose, dichloromethane (ꢀ ) 8.93) has
(29) Carlqvist, P.; Eklund, R.; Brinck, T. J. Org. Chem. 2001, 66, 1193.
(30) Lehtinen, C.; Nevalainen, V.; Brunow, G. Tetrahedron 2001, 57,
4741.
(31) Frisch, M. J. et al. Gaussian 98, revision A.9; Gaussian, Inc.:
Pittsburgh, PA, 1998.
(32) (a) Becke, A. D. Phys. ReV. 1988, A37, 785. (b) Lee, C.; Yang, W.;
Parr, R. G. Phys. ReV. 1988, B41, 785. (c) Becke, A. D. J. Chem. Phys.
1993, 98, 5648. (d) Stevens, P. J.; Devlin, F. J.; Chablowski, C. F.; Frisch,
M. J. J. Phys. Chem. 1994, 80, 11623.
(33) (a) Miertus, S.; Scrocco, E.; Tomasi, J. Chem. Phys. 1981, 55, 117.
(b) Tomasi, J.; Miertus, S. Chem. Phys. 1982, 65, 239.
(34) Cossi, M.; Barone, V.; Cammi, R.; Tomasi, J. Chem. Phys. Lett.
1996, 255, 327.
(35) As described by Okuno, formic acid is assumed to catalyze the
reaction without dissociation to H+.
J. Org. Chem, Vol. 71, No. 3, 2006 863

Grein et al.
SCHEME 2. Effect of Acid Catalyst on the Baeyer-Villiger Oxidation of 1a-f
TABLE 1. Energies, Enthalpies, and Gibbs Free Energies for the Baeyer-Villiger Oxidation of Acetone, Fluoroacetone, and Chloroacetonea
acetone (1a)
∆H
fluoroacetone (1b)
∆H
chloroacetone (1c)
∆H
system
∆E
∆G
∆E
∆G
∆E
∆G
TS1 (3a-c)
31.5
-7.9
15.5
-
30.6
-5.2
15.3
-
12.2
10.6
-
41.7
7.5
27.8
-
40.7
38.6
-
-67.7
-67.7
32.9
-7.9
16.4
15.3
15.4
12.1
9.3
-65.7
-75.0
31.0
-5.8
15.1
14.5
14.3
12.2
9.9
-64.0
-73.3
43.3
7.8
28.3
28.0
38.1
35.0
32.8
-63.9
-73.1
31.8
-7.9
16.7
16.3
14.7
11.7
9.9
-66.3
-71.7
30.9
-5.8
15.3
15.4
13.8
11.9
10.4
-64.5
-70.2
43.5
8.3
29.3
29.3
37.8
35.4
33.8
-64.3
-69.8
Criegee (4a-c)
TS2 (5a-c)
TS2x (6b,c)
TS1-cat (10a-c)
TS2-cat (11a-c)
TS2x-cat (12b,c)
prod (7a-c)
12.4
9.5
-
-68.7
-68.7
-66.4
-66.4
prodx (8b,c)
a All values are given in kcal/mol relative to reactants, using performic acid as the oxidant and formic acid as the catalyst. Only results for B3LYP/6-
31++G** are provided. TS2x-cat refers to the acid-catalyzed second transition state in which the halogenated carbon is migrating, and all other transition
states and products are numbered accordingly.
of a molecule of acid in the transition state, leading to an
increase of about 28 kcal/mol in ∆G for acetone, and about
23-24 kcal/mol for the other systems.
are very close to the ∆E’s and ∆H’s, since the number of
molecules remains the same as one goes from reactants to
products.
Interestingly, the calculations predict that the first step will
be rate-determining for all ketones examined. Activation ener-
gies of more than 30 kcal/mol are obtained for TS1 in the
absence of acid. Presumably, the poor trajectory enforced by
the necessity of transferring the peracid proton (Scheme 1) and
forming the C-O bond at the same time is responsible for the
high energy of the uncatalyzed TS. Activation energies (∆E)
for the second TS are much lower, ranging from about 10 to
16 kcal/mol.
The addition of acid dramatically decreases the ∆E of the
first transition state, presumably because of the more favorable
angle of attack. In the case of acetone, the ∆E of the first
transition state drops from 31.5 to 12.4 kcal/mol. Significant
decreases in energy, on the order of 16 kcal/mol, are also
obtained upon inclusion of acid in TS1 for the reaction of 1b
and 1c. Formic acid has less of an effect on the energies of the
second transition states, however, only decreasing them by 4-6
kcal/mol.
Results for 3-Pentanone, Fluoropentanone, and Chloro-
pentanone. It is well-known that the methyl group has one of
the lowest migratory aptitudes of simple aliphatic substituents
and that the ease of migration (migratory aptitude) increases
with increasing substitution. To examine these effects and move
beyond the extreme case of acetone, we performed calculations
on 3-pentanone and its halogenated derivatives (Table 2). For
the Criegee intermediates, the values are very similar to those
of the acetone systems given in Table 1. The uncatalyzed first
transition states have ∆E’s in the 30 kcal/mol range and ∆G’s
in the 40 kcal/mol range. Marked differences, however, are
observed in the second transition states. For TS2-cat, ∆E and
∆H are again about 5 kcal/mol lower than that for acetone,
whereas ∆G is 10 kcal/mol lower.
Interestingly, the catalyzed first transition state is also
decreased by almost 8 kcal/mol (free energy) for 3-pentanone
compared to acetone, indicating that the larger alkyl substituent
decreases the TS for addition as well as migration. Although
the effect of additional substituents on the Baeyer-Villiger
reaction has routinely been interpreted in terms of facilitating
the migration, it is interesting to note that there is a substantial
effect on the first transition state, which remains rate-determin-
ing. As in the case of acetone, once the ∆G of the reaction is
However, as expected, the enthalpic gain resulting from
inclusion of acid in the transition structures is mostly lost once
free energy is calculated. For acetone, ∆G (TS1-cat) is ca. 28.3
kcal/mol higher than ∆E. Similar effects are seen in the
haloacetone series, although the difference between ∆G and ∆E
is ca. 23 kcal/mol instead of 28. For the products, the ∆G values
864 J. Org. Chem., Vol. 71, No. 3, 2006

Studies of Baeyer-Villiger Oxidation of Ketones
TABLE 2. Energies, Enthalpies, and Gibbs Free Energies for the Baeyer-Villiger Reaction of 3-Pentanone, 2-Fluoro-3-pentanone, and
2-Chloro-3-pentanonea
pentanone
fluoropentanone
chloropentanone
system
∆E
∆H
∆G
∆E
∆H
∆G
∆E
∆H
∆G
TS1 (3d-f)
30.6
-7.6
10.9
-
28.8
-5.8
10.0
-
41.3
8.3
23.9
-
33.1
28.5
-
-71.2
-
33.9
-6.3
12.4
12.2
14.7
6.1
6.7
-69.9
-77.6
32.9
-4.4
11.4
11.6
13.6
6.3
7.5
-68.7
-76.4
45.0
9.2
25.2
25.3
37.7
29.3
30.0
-68.7
-76.2
31.8
-5.3
13.5
16.3
15.7
7.4
9.0
-69.9
-73.5
31.2
-3.4
12.7
15.8
14.8
7.4
9.8
-68.4
-72.2
43.3
10.4
26.6
29.2
38.3
31.5
32.3
-68.4
-72.4
Criegee (4d-f)
TS2 (5d-f)
TS2x (6e,f)
cat-TS1 (10d-f)
cat-TS2 (11d-f)
cat-TS2x (12e,f)
prod (7d-f)
10.9
4.4
10.0
5.0
-
-72.7
-
-
-71.4
-
prodx (8e,f)
a All values are given in kcal/mol relative to reactants, using performic acid as the oxidant and formic acid as the catalyst. Only results for B3LYP/6-
31++G** are provided. cat-TS2x refers to the acid-catalyzed second transition state in which the halogenated carbon is migrating, and all other transition
states and products are numbered accordingly.
TABLE 3. Solvation-Adjusted Gibbs Free Energies (kcal/mol),
with Dichloromethane as Solvent, for the Baeyer-Villiger Reaction
of Acetone, Fluoroacetone, and Chloroacetonea
TABLE 4. Solvation-Adjusted Gibbs Free Energies (kcal/mol), in
Dichloromethane as Solvent, for the Baeyer-Villiger Reaction of
Pentanone, Fluoropentanone, and Chloropentanonea
acetone (1a)
fluoroacetone (1b)
chloroacetone (1c)
3-pentanone
fluoropentanone
chloropentanone
(1d)
(1e)
(1f)
TS1
43.7 (2.0)
10.4 (2.9)
29.5 (1.7)
-
45.6 (4.9)
43.4 (4.8)
-
46.0 (2.7)
10.6 (2.8)
30.6 (2.3)
30.4 (2.4)
43.6 (5.5)
40.2 (5.2)
38.5 (5.7)
-65.5 (-1.6)
-74.0 (-0.9)
46.6 (3.1)
11.3 (3.0)
31.7 (2.4)
31.4 (2.1)
43.6 (6.3)
40.8 (5.4)
39.3 (5.5)
-65.0 (-0.7)
-70.5 (-0.7)
Criegee
TS2
TS1
44.7 (3.4)
12.6 (4.3)
26.7 (2.8)
-
38.9 (5.8)
35.1 (6.6)
-
48.6 (3.6)
12.9 (3.7)
27.9 (2.7)
28.2 (2.9)
43.9 (6.2)
35.9 (6.6)
36.9 (6.9)
-70.5 (-1.8)
-77.1 (-0.9)
46.8 (3.5)
14.5 (4.1)
29.5 (2.9)
31.9 (2.7)
45.5 (7.2)
38.4 (6.9)
38.9 (6.6)
-69.2 (-0.8)
-73.0 (-0.6)
Criegee
TS2
TS2x
TS1-cat
TS2-cat
TS2x-cat
prod
TS2x
TS1-cat
TS2-cat
TS2x-cat
prod
-68.1 (-0.4)
-
prodx
-71.7 (-0.5)
-
prodx
a Energies are relative to reactants, and results are given for calculations
with B3LYP/6-31++G**, using the PCM method. ∆G of solvation is given
in parentheses.
a Energies are relative to reactants, and results are given for calculations
with B3LYP/6-31++G**, using the PCM method. ∆G of solvation is given
in parentheses.
considered, the optimal reaction path for 3-pentanone is via acid
catalysis for TS1, but not for TS2.
the new values for the overall solvation-adjusted ∆G’s relative
to the reactants. Table 4 describes the pentanone series.
As shown in Table 3, the Criegee intermediate as well as
both transition states have positive ∆Gsolv values, implying that
the reactants benefit more from solvation than the higher energy
species, whereas the products, having negative values of ∆Esolv
and ∆Gsolv, are more stabilized than the reactants. For the
Criegee intermediate and the noncatalyzed TSs, the increases
Fluorine and Chlorine Substitution. In the absence of acid,
TS1 and TS2 of fluoroacetone are energetically similar to
acetone. A very slight preference (less than 1 kcal) is noted for
migration of CH2F versus migration of CH3 in 1b. The ∆G’s
of both the first and second transition states are higher for
chloroacetone than acetone by about 2 kcal/mol, while the
catalyzed transition states are slightly lower for chloroacetone
(by 3-5 kcal/mol). The uncatalyzed second transition states for
migration of CH3 and CH2Cl are identical in energy. If the
catalyzed second transition states for fluoroacetone and chlo-
roacetone are examined, the migration of the halogenated carbon
is preferred by 1.6-2.3 kcal/mol. Although significant at the
transition state, these differences are not experimentally mean-
ingful since the reaction presumably proceeds by the lower
energy, uncatalyzed second transition state.
Interestingly, for the fluorine- and chlorine-substituted pen-
tanones, TS1 and TS2 lie energetically well above those of
unsubstituted pentanone itself. Migration of the two substituents
is equi-energetic (∆G) in the case of fluoropentanone (0.1 kcal/
mol difference), while there is a significant advantage for
migration of the nonhalogenated carbon in the case of chloro-
pentanone (2.8 kcal/mol). Again, these numbers are for the ∆G’s
of the noncatalyzed second transition states.
are relatively small, between 1.7 and 3.0 kcal/mol for ∆Gsolv
.
The trends remain the same, with little difference between the
migratory aptitudes of the halogenated and nonhalogenated
substituents and the lowest transition state energies correspond-
ing to acetone, followed by fluoroacetone and chloroacetone.
In the acid-catalyzed systems, the ∆Gsolv numbers typically
increase by 5-6 kcal/mol. As seen earlier for the ∆G’s of the
reaction, the entropy of solvation of one cat-TS molecule is
much lower than that of three molecules, from which these
transition structures are built. Thus, the catalyzed second
transition state is still ca. 10 kcal/mol higher than the uncatalyzed
TS in all three substrates. The results are similar for the
pentanone series, except slightly higher increases are observed.
The effect of including solvation in the calculations is to raise
the free energy of all of the transition states, especially the ones
that include acid catalysis. In terms of relative ordering of the
various transition states, two main effects result. In the case of
acetone, the uncatalyzed first transition state becomes lower in
energy than the catalyzed version. For all other systems, the
addition of acid still results in a lower transition state, despite
the entropic penalties. The other effect of solvation is to slightly
separate the energy of the transition states for migration of
Effect of Solvation. To assess the effect of solvation on the
Criegee intermediate and on the activation energies, calculations
were performed with dichloromethane (ꢀ ) 8.93) as solvent,
using the PCM method and B3LYP/6-31++G**. In Table 3,
the calculated values of Gibbs free energy of solvation relative
to the gas phase (∆Gsolv) are given in parentheses, along with
J. Org. Chem, Vol. 71, No. 3, 2006 865

Grein et al.
the former carbonyl (C2-O1) at 1.39-1.40 Å, the O3-O4 bond
at 1.44-1.45 Å, and the newly formed O1-H5 bond at 0.97 Å
for all six ketones.
In all of the second transition states examined (Table 6), the
O-O bond has stretched from 1.45 Å in the Criegee intermedi-
ates to 1.95-1.99 Å at the transition states. The incipient bond
between the migrating CH3 or CH2CH3 group (Rmig ) migrating
group) and the peroxide oxygen is remarkably constant, being
2.00-2.01 Å for the acetone series and 2.06-2.11 Å for the
pentanone structures (uncatalyzed TS2). The transition state
structures for migration of the halogenated substituents (TS2x)
all feature longer incipient XCH2-O bonds than observed for
the corresponding methyl migrations, implying that these
transition states are less advanced. In the case of fluoroacetone,
the difference is the smallest at 0.06 Å, but in all other examples,
this bond is longer by about 0.1 Å. The O-O bond is also
shorter in these cases, consistent with a less advanced transition
state.
In the catalyzed second transition state there is little difference
in the O-O bond for migration of CH3 or CH2X, with the
exception of TS2x-cat for chloropentanone (12f), where the
O-O bond is relatively short at 1.92 Å. Elongation of C-Rmig
is observed in all cases for TS2 and TS2x upon addition of acid.
The transfer of the hydroxyl proton from the Criegee intermedi-
ate to the acid catalyst (H5-O7) is significantly more advanced
in the acetone series than the pentanone series.
3. Experimental Results
FIGURE 1. Relative free energies including solvation (dichlo-
romethane) for the Baeyer-Villiger oxidation of 1a, d, e, and f.
The relatively few published studies of oxidations of R-ha-
loketones36 indicate a preference for migration of the nonha-
logenated carbon of varying magnitudes. However, these studies
can be complicated by the fact that migration of the halogen-
bearing carbon leads to a hydrolytically unstable species that
can easily decompose under the reaction conditions (eq 2).
Hydrolysis of the product from oxidation at the halogen-bearing
carbon leads eventually to production of the corresponding
aldehyde and HX, which then further increases the rate of
hydrolysis of the ester, and the aldehyde itself is likely unstable
under oxidizing conditions. This series of reactions complicates
the analysis of the relative migratory aptitudes of these
substrates, since low-yielding reactions with poor mass balances
may actually indicate a propensity for migration of the
halogenated substituent.
CHFCH3 and CH2CH3 in fluoropentanone such that migration
of the ethyl group is more favorable by 0.3 kcal/mol compared
to migration of the fluorinated substituent (Figure 1). Overall,
pentanone (green) is the most reactive substrate for both the
first and the second transition states and chloroacetone (black)
is the least.
Structures of Criegee Intermediates and Transition States.
The calculated structures for the oxidation of acetone and
pentanone are shown in Figure 2. In the halogenated series, three
conformers were examined in each case, and in some cases,
the difference between the different conformers was found to
be dramatic. The most stable conformers are shown in Figure
2. A comparison of the halogen conformers is given in the
Supporting Information.
Key bond distances for TS1, the Criegee intermediate, and
TS1-cat are shown in Table 5. On the basis of the length of the
incipient carbon-oxygen bond, the uncatalyzed first transition
state is the most advanced for fluoro- and chloroacetone (3b/c,
C2-O3 bond ) 2.10 and 2.12 Å) and least for chloro- and
fluoropentanone 3e/f (C2-O3 ) 2.23 Å). Interestingly, the length
of the incipient O-H bond does not correlate with the C-O
bond, being longest in 3b (1.11Å) and shortest in 3e/f (1.04
and 1.03 Å).
Despite the significant rearrangement required in TS1 for the
inclusion of acid, the distance for the incipient C-O bond
changes little for most of the compounds examined. A dramatic
exception is fluoropentanone (10e), in which a 0.15 Å contrac-
tion is observed upon addition of acid. With this change, the
most advanced transition states are observed for fluoroacetone
and fluoropentanone, with C2-O3 bond lengths of 2.08 Å in
both cases.
To assess the selectivity of the oxidation in halogenated
systems, several fluoro- and chloroketones were prepared.
(36) (a) Itoh, Y.; Yamanaka, M.; Mikami, K. Org. Lett. 2003, 5, 4803.
(b) Marigo, M.; Bachmann, S.; Halland, N.; Braunton, A.; Jorgensen, K.
A. Angew. Chem., Int. Ed. 2004, 43, 5507. (c) Kobayashi, S.; Tanaka, H.;
Amii, H.; Uneyama, K. Tetrahedron 2003, 59, 1547. (d) Shiozaki, M.; Arai,
M. J. Org. Chem. 1989, 54, 3754. (e) Overberger, C. G.; Kaye, H. J. Am.
Chem. Soc. 1967, 89, 5640. (f) Ahmad, M. S.; Farooq, Z. Ind. J. Chem.
1977, 233. (g) Smissman, E. E.; Bergen, J. V. J. Org. Chem. 1962, 27,
2316. (h) Ali, S. M.; Roberts, S. M. J. Chem. Soc., Perkin Trans. 1 1976,
1934. (i) Bollinger, J. E.; Courtney, J. L. Aust. J. Chem. 1964, 17, 440.
The bond distances in the Criegee intermediates are remark-
ably constant across the series, with the C-O bond length of
866 J. Org. Chem., Vol. 71, No. 3, 2006

Studies of Baeyer-Villiger Oxidation of Ketones
FIGURE 2. Calculated structures for key intermediates and transition states in the presence and absence of acid.
Unfortunately, fluoroacetone and chloroacetone were both highly
resistant to oxidation, which is not surprising due to the very
low migratory aptitude of the methyl substituent (or the CH2X
group). Rather than employ unusual catalysts or conditions to
force this reaction to go to completion, we decided to employ
representative ketones in our experimental study. Although it
would be ideal to employ exactly the same ketones in both cases,
we felt it was more desirable to have experimental conditions
that were as close as possible to the simulated conditions.
Considering the difficulties previously described with mass
balance in the Baeyer-Villiger reactions of halogenated ketones,
we were also concerned about the volatility of the pentanone
derivatives. Thus, we prepared ketones 16-22, which represent
a relatively wide range of substrates, including straight-chain,
cyclic, and benzylic aliphatic ketones. In particular, hexanone
derivatives 19 and 22 are very close analogies of the pentanone
derivatives examined in the theoretical studies. The synthetic
routes employed for the preparation of the requisite ketones are
shown in Scheme 3.
R-Haloketones 16-22 were then subjected to Baeyer-
Villiger oxidation with mCPBA (Table 7). In all of the ketones
examined, oxidation occurred preferentially at the nonhaloge-
nated substituent. In the case of the R-fluoroketones, the
selectivity was remarkably consistent throughout a range of
substrates. In all cases, mass balances were between 83 and
93%. Both the 2-fluorocyclohexanone substrate 16 and the
dibenzylketone 17 reacted with 70:30 preference for oxidation
at the nonhalogenated carbon. Ketones 18 and 19, which are
better mimics of 2-fluoropentanone (1e), reacted with 75:25 and
67:33 selectivities, respectively. Again, the major isomer
observed was that of oxidation at the nonhalogenated carbon.
This ratio of products translates into a ∆∆Gq for their formation
from the Criegee intermediate of 0.4-0.63 kcal/mol. This
corresponds very well to the differences in free energy predicted
for 2-fluoro-3-pentanone (1e) (0.3 kcal/mol, including entropy
and using the solvation model, Table 4). In the case of 19, which
is the closest analogy to 1e, we observe the best match with
J. Org. Chem, Vol. 71, No. 3, 2006 867

Grein et al.
TABLE 5. Key Bond Distances for TS1, Criegee, and TS1-cat
the range of 10-15 kcal/mol, but ∆G is 33-40 kcal/mol (Tables
1 and 2). Despite this fact, addition to the ketone via the acid-
catalyzed manifold is still more favorable for all ketones
examined, with the exception of acetone (when solvation is
included in the calculations). In this case, the addition via the
uncatalyzed pathway is more favorable by 1.9 kcal/mol.
Inclusion of entropy in the calculations also has an effect on
the Criegee intermediate. Examination solely of ∆E or ∆H leads
to the untenable conclusion that this intermediate, which has
never been isolated, is more stable than the starting material by
5-6 kcal/mol. However, when entropy is included in the
calculations, the Criegee intermediate becomes 7-8 kcal/mol
higher in energy than the starting materials.
structure
TS1 (3a)
TS1 (3b)
TS1 (3c)
TS1 (3d)
TS1 (3e)
TS1 (3f)
Criegee (4a)
Criegee (4b)
Criegee (4c)
Criegee (4d)
Criegee (4e)
Criegee (4f)
TS1-cat (10a) 1.29
TS1-cat (10b) 1.29
TS1-cat (10c) 1.29
TS1-cat (10d) 1.30
TS1-cat (10e) 1.29
C2-O1 C2-O3 O1-H5 O3-H5 O7-H6 O7-H5 O3-H6
1.29
1.30
1.29
1.29
1.29
1.29
1.40
1.40
1.40
1.40
1.39
1.40
2.17
2.10
2.12
2.16
2.23
2.23
1.45
1.44
1.44
1.45
1.45
1.44
2.18
2.08
2.15
2.18
2.08
2.20
1.07
1.11
1. 05
1.07
1.04
1.03
0.97
0.97
0.97
0.97
0.97
0.97
1.02
1.05
1.04
1.01
1.04
1.03
1.46
1.39
1.52
1
1.60
1.60
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
For the second transition state, the inclusion of acid leads to
a small decrease in ∆E and ∆H, but the entropic penalty is too
great to counteract this. Thus, it is more energetically favorable
for the reaction to proceed in the absence of acid catalyst for
the second step.
The comparison between acetone and pentanone is also
instructive. Experimentally, it has been shown that migration
of methyl groups occurs extremely slowly, and therefore we
chose to include the Baeyer-Villiger oxidation of 3-pentanone
and its derivatives to permit more meaningful comparisons with
experimental results. As expected, the reaction of pentanone
was significantly enhanced relative to that of acetone. Relative
to that of acetone, the calculated first and second transition states
were both lowered by about 7 kcal/mol (solvated ∆G values).
In fact, pentanone was the most reactive substrate of all those
examined. The addition of a fluorine substituent on either
acetone or pentanone had a slightly decelerating effect on both
steps of the reaction. Surprisingly, the transition state for
migration of the nonfluorinated carbon in fluoropentanone was
only 0.3 kcal/mol lower than for migration of CHFCH3. A larger
preference of 2.6 kcal/mol was calculated for migration of CH2-
CH3 versus CHClCH3 in chloropentanone. The first transition
state was elevated in all the chlorinated systems, implying that
the overall rate will be slower, a fact borne out in experimental
studies.
1.23
1.57
1.20
1.30
1.26
1.22
1.29
1.23
1.47
1.49
1.59
1.50
1.53
1.14
1.17
1.20
1.15
1.20
TS1-cat (10f)
1.29
the theoretical studies (∆∆Gq calculated ) 0.3 kcal/mol and
observed ) 0.42 kcal/mol).
The oxidation of R-chloroketones 20-22 occurred signifi-
cantly more slowly. For example, 3-chloroheptan-4-one (22)
took 18 days to go to 22% conversion compared to the fluoro
derivative 19, which was 55% complete after just 4 days. This
is in agreement with our calculations, which show that TS1,
the rate-determining step, is highest in the chlorinated ketones.
The calculations also predict a greater selectivity for the
oxidation of the aliphatic substituent relative to the chlorinated
substituent. For chloropentanone, ∆∆Gq is calculated to be on
the order of 2.4 kcal/mol. For ketones 20 and 22, we only
observe the isomer resulting from oxidation at the nonhaloge-
nated carbon. This is reasonable since a ∆∆Gq of 2.4 kcal/mol
translates into a product ratio of ca. 97:3. In one case,
chlorocyclohexanone 21, we did observe the minor product in
ca. 10% yield. This corresponds to a ∆∆Gq of 1.1 kcal/mol.
Again, we note that the mass balances in all cases were high
(usually greater than 80%) and therefore a high regioselectivity
can be inferred.
The few experimental studies in which halogenated ketones
are subjected to Baeyer-Villiger oxidation are generally
synthetic in focus, and thus the minor product (oxidation of the
halogenated carbon) is rarely a consideration. Thus, to put our
theoretical work in context, we carried out the Baeyer-Villiger
oxidation of halogenated ketones, specifically to determine the
regioselectivity of the reaction experimentally. In all of the
fluoroketones examined, a ratio of approximately 70:30 in favor
of migration of the nonfluorinated substituent was determined.
This ratio translates to a ∆∆Gq of between 0.42 and 0.63 kcal/
mol, which is remarkably similar to the predicted value of 0.3
kcal/mol for 2-fluoropentanone. Interestingly, the observed
regioselectivity for oxidation of 3-fluorohexanone (0.42 kcal/
mol) was actually closest to the calculated value of 0.3 kcal/
mol.
4. Discussion
To the best of our knowledge, this study is the only theoretical
examination of the complete Baeyer-Villiger reaction of a
ketone, including both the addition and rearrangement steps.
Six ketones were examined theoretically, and seven experimen-
tally. Remarkable agreement was found between theory and
experiment, even for ketones that had significant structural
differences.
In the theoretical studies, the first step of the reaction was
found to be rate-determining in the presence or absence of acid.
The presence of acid has a significant effect on the enthalpy of
the addition, improving the geometry of the attack and resulting
in a decrease in ∆H of 16-19 kcal/mol. However, including
an extra molecule in the transition state results in a large entropic
penalty and a higher free energy. Thus, ∆H for TS1-cat is in
In the chlorinated series, the preference for migration of the
nonchlorinated substituent has been difficult to measure since
in many cases only one product was observed. This is also
consistent with the larger ∆∆Gq obtained in our calculations,
which predict ratios on the order of 97:3 (nonhalogenated/
halogenated). Our experiments with the chlorinated systems also
revealed a significantly decreased rate of reaction, again
consistent with theory, which showed that the transition state
868 J. Org. Chem., Vol. 71, No. 3, 2006

Studies of Baeyer-Villiger Oxidation of Ketones
TABLE 6. Key Bond Distances for TS2, TS2x, TS2-cat, and TS2x-cat
structure
TS2 (5a)
TS2 (5b)
TS2 (5c)
TS2 (5d)
TS2 (5e)
C2-Rmig
Rmig-O3
O3-O4
O1-H5
H5-O8
H5-O7
O7-H6
H6-O8
1.83
1.82
1.82
1.83
1.84
1.84
1.83
1.87
1.90
1.83
1.88
1.96
1.89
1.90
1.89
1.88
1.90
1.89
1.88
1.93
1.97
1.88
1.97
2.00
2.00
2.01
2.00
2.06
2.11
2.10
2.00
2.06
2.09
2.06
2.17
2.18
1.98
1.98
1.98
2.09
2.10
2.09
1.98
2.04
2.10
2.09
2.19
2.19
1.98
1.99
1.99
1.98
1.99
1.95
1.98
1.98
1.96
1.98
1.94
1.94
1.99
2.01
2.02
1.96
1.99
1.96
1.99
1.98
1.97
1.96
1.94
1.92
1.05
1.08
1.08
1.04
1.04
1.03
1.05
1.05
1.04
1.04
1.02
1.02
1.00
1.01
1.00
0.99
0.99
.99
1.46
1.39
1.38
1.53
1.50
1.50
1.46
1.47
1.49
1.53
1.56
1.57
-
-
-
-
-
-
TS2 (5f)
-
TS2x (6a)
-
TS2x (6b)
-
TS2x (6c)
-
TS2x (6d)
-
TS2x (6e)
-
TS2x (6f)
-
TS2-cat (11a)
TS2-cat (11b)
TS2-cat (11c)
TS2-cat (11d)
TS2-cat (11e)
TS2-cat (11f)
TS2x-cat (12a)
TS2x-cat (12b)
TS2x-cat (12c)
TS2x-cat (12d)
TS2x-cat (12e)
TS2x-cat (12f)
1.66
1.60
1.62
1.73
1.70
1.70
1.66
1.67
1.68
1.73
1.76
1.77
1.04
1.06
1.05
1.03
1.03
1.03
1.04
1.04
1.03
1.03
1.02
1.02
1.47
1.43
1.45
1.52
1.51
1.52
1.47
1.48
1.50
1.52
1.55
1.56
-
-
-
-
-
1.00
0.99
0.99
0.99
0.98
0.98
-
-
-
-
SCHEME 3. Preparation of Ketones 16-22
the predicted energy of the Criegee intermediates and both
transition states. Acid catalysts generally decrease the ∆G of
the first transition state, but increase the ∆G of the second one.
Solvation stabilizes the products more than the starting materials
and the starting materials more than the transition states or
Criegee intermediate. The optimal reaction path for acetone
proceeds over the uncatalyzed step for both TS1 and TS2,
whereas the reaction with pentanone proceeds best via TS1-cat
and TS2.
Relative to that of the other ketones examined in this study,
migration of the ethyl group in pentanone seems to be the most
energetically favorable. The introduction of a fluorine substituent
decreases the migratory aptitude by a surprisingly moderate
amount, both experimentally and theoretically. Chlorination
decreases the experimentally observed rate of reaction, consistent
with theoretical predictions. Both theory and experiment show
the relative propensity for migration of a chlorinated substituent
to be significantly attenuated relative to a nonhalogenated
substituent. Work is ongoing to extend both the theoretical and
experimental work to other classes of substrates including ethers,
amines, and aromatic ketones.
for attack at the carbonyl was highest in the case of the
chlorinated substituents.
5. Summary and Conclusions
6. Experimental Section
Several interesting features of the Baeyer-Villiger reaction
have been revealed in our combined theoretical and experimental
study. The reaction of acetone and 3-pentanone with performic
acid is highly exothermic, with the products lying on average
60-80 kcal/mol lower in energy than the starting materials. The
first step is rate-determining for all substrates examined, even
in the presence of acid. The inclusion of entropy in the
calculations is vital since it makes a significant difference in
Commercial grade mCPBA was purified using an adaptation of
the protocol described by Bortolini et al.38 Diisopropylamine was
distilled from CaH2 and stored cold over potassium hydroxide
pellets. N-Chlorosuccinimide was recrystallized from dichlo-
(37) Sever, R. R.; Root, T. W. J. Phys. Chem. B 2003, 107, 10848.
(38) Bortolini, O.; Campestrini, S.; Difuria, F.; Modena, G.; Valle, G.
J. Org. Chem. 1987, 52, 5467.
J. Org. Chem, Vol. 71, No. 3, 2006 869

Grein et al.
TABLE 7. Baeyer-Villiger Oxidation of r-Haloketones
then at room temperature overnight. At 0 °C, the reaction was
quenched with 40 mL of distilled water and extracted with diethyl
ether (4 × 100 mL). The combined organic layers were washed
with brine (200 mL), dried over Na2SO4, filtered, and concentrated
in vacuo. The product mixture was purified by column chroma-
tography (silica gel, 10:1 hexanes/ethyl acetate) to afford fluoro-
ketone 17 as a white powder. Subsequent recrystallization from
boiling hexanes gave 0.73 g (91%) of 17 as crystalline white plates.
Spectra are consistent with those reported elsewhere.42
Synthesis of 2,9-Diphenyl-3-fluorononan-5-one (18). Following
the procedure for the synthesis of 17, diisopropylamine (0.54 mL,
4.1 mmol), n-BuLi (1.7 mL of 2.5 M hexane solution, 4.3 mmol),
chlorotriethylsilane (0.94 mL, 5.6 mmol), and the parent ketone
were reacted to give 0.76 g (71%) of the corresponding enol silyl
ether. A portion of this material (0.70 g, 2.1 mmol) was then
fluorinated with Selectfluor (1.34 g, 3.78 mmol). The crude product
mixture thus obtained was then purified by column chromatography
(silica gel, 30:1 hexanes/ethyl acetate) to give 0.67 g (73%) of 18.
1H NMR (400 MHz, CDCl3) δ 7.3-7.6 (m, 10H), 4.73 (ddd, J )
49.0, 8.8, 4.0 Hz, 1H), 2.61-2.82 (m, 6H), 1.91-2.10 (m, 4H).
19F NMR (376 MHz, CDCl3) δ -196.0 (m).
Synthesis of 2-Fluoro-3-heptanone (19). LDA was generated
by the dropwise addition of 1.37 M n-BuLi (25.5 mL, 35 mmol)
to diisopropylamine (35.7 mmol, 5.0 mL) in 40 mL of dry THF.
The LDA solution was cooled to -78 °C, and to this 4-heptanone
(20.0 mmol, 2.8 mL) and trimethylsilyl chloride (39.4 mmol, 5.0
mL) were slowly added as a solution in 20 mL of dry THF. The
reaction flask was left to warm to room temperature over 16 h.
THF was removed in vacuo, and the reaction mixture was diluted
with 20 mL of anhydrous DMF. Selectfluor (40.0 mmol, 14.2 g)
was next added as a suspension in 20 mL of DMF at 0 °C and left
to stir at ambient temperature for 16 h. The reaction was quenched
with distilled water (100 mL) and extracted with diethyl ether (3
× 30 mL). Partial pressure distillation at 250 mmHg followed.
Ketone 18 was isolated in the fraction boiling at 104-107 °C. Yield
44%. 1H and 13C NMR spectra are consistent with published data.43
Preparation of (()-2-Chlorodibenzylacetone (20). In a 100-
mL RBF equipped with a magnetic stir bar, 2-hydroxydibenzyl-
acetone (0.091 g, 0.40 mmol) was suspended in 8 mL of CCl4.
This mixture was cooled to 0 °C, and then a solution of
triphenylphosphine (0.13 g, 0.48 mmol) in 20 mL of CCl4 was
added. The reaction vessel was then fitted with an oven-dried reflux
condenser, and the reaction suspension was refluxed in darkness
for 10 h. After cooling to room temperature, 100 mL of pentane
was added. The resulting suspension was then thoroughly triturated
and filtered. The filtrate was concentrated to reveal a cloudy residue.
The crude product mixture thus obtained was purified via column
chromatography (silica gel, gradient elution, 30:1 to 20:1 to 15:1
hexane/ethyl acetate) to afford 93 mg (76% yield) of chloroketone
20. Spectral data were consistent with literature data.41 1H NMR
(400 MHz, CDCl3) δ 7.16-7.31 (m, 8H), 6.97 (m, 2H), 5.37 (s,
1H), 3.80 (d, J ) 16.0 Hz, 1H), 3.64 (d, J ) 16 Hz, 1H). 13C {1H}
NMR (100 MHz, CDCl3) δ 206.1, 136.3, 132.5, 129.5, 129.2, 128.8,
128.6, 127.1, 126.9, 78.1, 44.4.
a Ratio of oxidation at the two substituents determined by 1H NMR
analysis of the crude reaction mixture. b Overall yield of the two products
along with any recovered starting material. c Reaction only proceeded to
20% conversion. d Reaction only proceeded to 22% conversion.
romethane. The concentration of n-butyllithium used was deter-
mined via titration against BHT and fluorene or against N-benzyl-
benzamide. Commercially available 4-tert-butylcyclohexanone was
recrystallized from hexanes. 4-Heptanone was distilled prior to use.
All other chemicals were used without further purifications.
Fluoroketone 16,39 the parent ketone of 18,40 and chloroketone 2041
were prepared according to literature procedures.
Synthesis of 2-Fluorodibenzylacetone (17). A 100-mL RBF
was charged with a solution of diisopropylamine (0.62 mL, 4.7
mmol) in 20 mL of THF. This mixture was cooled to -78 °C and
stirred at this temperature for 10 min. n-BuLi (2.2 mL of 2.16 M
hexane solution, 4.8 mmol) was then added dropwise over 5 min,
and the resulting pale yellow solution was stirred at -78 °C for a
further 45 min. A solution of 1,3-diphenylacetone (0.98 g, 4.7
mmol) and chlorotriethylsilane (0.94 mL, 5.6 mmol) in 30 mL of
THF was then added dropwise via a syringe. The reaction mixture
was stirred cold for 1 h and then warmed to room temperature and
stirred for a further 2 h. The reaction mixture was concentrated in
vacuo, and the resulting residue was triturated with pentane (100
mL). The insoluable salts were filtered off to reveal, after
concentration in vacuo, 1.23 g (83%) of the corresponding enol
silyl ether. A portion of this material (1.14 g, 3.5 mmol) was then
diluted with 30 mL of DMF. To this solution was added finely
powdered Selectfluor (1.50 g, 4.23 mmol) as a suspension in 30
mL of DMF. The resulting mixture was stirred at 0 °C for 2 h and
Synthesis of cis-4-tert-Butylchlorocyclohexanone (21). A 250-
mL RBF equipped with a magnetic stir bar and addition funnel
was charged with a solution of enolsilyl ether 1444 (7.60 g, 33.6
mmol) in 50 mL of dried dichloromethane at 0 °C. A solution of
freshly recrystallized N-chlorosucccinimide (5.40 g, 40.4 mmol)
in 100 mL of dichloromethane was added dropwise via the addition
funnel. The reaction mixture was stirred at 0 °C for 8 h and then
was warmed to room temperature and allowed to stir for another
16 h. Upon recooling to 0 °C, the reaction was quenched with 100
mL of distilled water. The organic layer was separated and washed
with brine (100 mL) and dried over Na2SO4, filtered, and
concentrated in vacuo. The resulting oily mixture of products was
isolated by repeated column chromatography (silica gel, 20:1
(39) Denmark, S. E.; Wu, Z. C.; Crudden, C. M.; Matsuhashi, H. J. Org.
Chem. 1997, 62, 8288.
(40) Seebach, D.; Corey, E. J. J. Org. Chem. 1975, 40, 231.
(41) Fort, A. W. J. Am. Chem. Soc. 1962, 84, 2620.
(42) Rozen, S.; Menahem, Y. J. Fluorine Chem. 1980, 16, 19.
870 J. Org. Chem., Vol. 71, No. 3, 2006

Studies of Baeyer-Villiger Oxidation of Ketones
hexanes/ethyl acetate) to afford 3.74 g (59%) of ketone 21 along
with about 8-10% of the trans isomer, which was separated by
chromatography. The spectra obtained were identical to those
reported elsewhere.29
Synthesis of 3-Chloro-4-heptanone (22). 4-Heptanone (5.03
mL, 36.00 mmol) was dissolved in 40 mL of anhydrous CCl4 and
maintained at 0 °C during the dropwise addition of sulfuryl chloride
(2.90 mL, 36.10 mmol). The reaction flask was then warmed to 50
°C for 3.5 h. CCl4 was removed by distillation at atmospheric
pressure. Isolation of pure 22 was accomplished by column
chromatography (silica gel, 8:1 hexanes/methylene chloride, Rf
0.41), 3.153 g (59% yield). 22 displayed spectra consistent with
previously published data.45
(102 mg, 0.59 mmol), and sodium bicarbonate (41 mg, 0.50 mmol)
were reacted to give, after purification by column chromatography
(silica gel, gradient elution, 14:1 to 12:1 to 10:1 hexane/ethyl
acetate), 72 mg (60% yield) of 17(ox) and 27 mg (24% yield) of
17(F-ox).
1
Fluorolactone 17(ox): IR (NaCl) ν 1760 (s). H NMR (400
MHz, CDCl3) δ 7.36 (m, 10H), 5.81 (d, J ) 47.6 Hz, 1H), 5.20
(dd, J ) 33.2, 12.0 Hz, 2H). 13C {1H} NMR (100 MHz, CDCl3) δ
168.3, 134.9 (d, J ) 18.3 Hz), 133.9, 129.6 (d, J ) 2.4 Hz), 128.8,
128.6, 128.5, 128.1, 126.7, 89.3 (d, J ) 184.9 Hz), 67.3. 19F NMR
(376 MHz, CDCl3) δ -180.6 (d, J ) 48.9 Hz). MS (EI, 70 eV),
m/e: 91 (C7H7, 100), 244 (m+, 17). HRMS (EI), m/e: calcd for
C15H13O2F (m+): 244.0900, found: 244.0908. TLC mobility, Rf
0.46 (in 8:1 hexane/ethyl acetate, visualized with UV and p-
anisaldehyde).
Baeyer-Villiger Oxidation of trans-2-Fluoro-4-t-butylcyclo-
hexanone (16).
1
Fluorolactone 17(F-ox): IR (NaCl) ν 1764 (s). H NMR (400
MHz, CDCl3) δ 7.36 (m, 10H), 7.21 (d, J ) 55.2 Hz, 1H), 3.77 (d,
J ) 1.2 Hz, 2H). 13C {1H} NMR (100 MHz, CDCl3) δ 169.6, 134.5
(d, J ) 22.3 Hz), 132.7, 130.2 (d, J ) 1.6 Hz), 129.3, 128.7, 128.6,
127.4, 126.1 (d, J ) 5.5 Hz), 101.7 (d, J ) 220.0 Hz), 41.0. 19F
NMR (376 MHz, CDCl3) δ -123.0 (d, J ) 54.9 Hz). MS (EI, 70
eV), m/e: 91 (C7H7, 83), 225 (m+ - F, 23), 244 (m+, 7). HRMS
(EI), m/e: calcd for C15H13O2F (m+): 244.0900, found: 244.0889.
TLC mobility, Rf 0.54 (in 8:1 hexane/ethyl acetate, visualized with
UV and p-anisaldehyde).
In a 25-mL RBF were combined fluoroketone 16 (446 mg, 2.59
mmol), mCPBA (530 mg, 3.07 mmol), and sodium bicarbonate (218
mg, 2.59 mmol) in 13 mL of HPLC grade chloroform. The resulting
suspension was then stirred under a static nitrogen atmosphere in
darkness at ambient temperature and monitored periodically via
thin-layer chromatography. When the reaction was deemed com-
plete, chloroform was removed in vacuo to reveal a white solid
residue. The residue was then taken up in hexane (50 mL), and the
resulting suspension was filtered. After washing the precipitate with
copious amounts of hexane, we concentrated the filtrate in vacuo.
The resulting yellow oil, after 1H NMR analysis, was then purified
via repeated column chromatography (silica gel, gradient elution,
14:1 to 12:1 hexane/ethyl acetate) to afford 291 mg (66% yield) of
16(ox) and 120 mg (27% yield) of 16(F-ox).
Baeyer-Villiger Oxidation of (()-2,9-Diphenyl-3-fluornonan-
5-one (18).
1
Fluorolactone 16(ox): IR (NaCl) ν 1746 (s). H NMR (400
Following the optimized protocol outlined for the oxidation of
fluoroketone 16, fluoroketone 18 (665 mg, 2.34 mmol), mCPBA
(466 mg, 2.7 mmol), and sodium bicarbonate (227 mg, 2.7 mmol)
were reacted to give, after purification by way of column chroma-
tography (silica gel, 30:1 hexane/ethyl acetate), 111 mg (16% yield)
of 18(ox) and 38 mg (6% yield) of 18(F-ox).
MHz, CDCl3) δ 5.39 (ddd, J ) 49.2, 6.8, 3.2 Hz, 1H), 4.59 (ddd,
J ) 12.8, 8.8, 2.4 Hz, 1H), 4.25 (m, 1H), 2.36 (m, 1H), 2.11-2.18
(m, 1H), 1.60 (m, 3H), 0.91 (s, 9H). 13C {1H} NMR (100 MHz,
CDCl3) δ 170.1 (d, J ) 22.0 Hz), 90.5 (d, J ) 185.0 Hz), 67.7 (d,
J ) 6.0 Hz), 42.9 (d, J ) 2 Hz), 32.6, 29.5 (d, J ) 21.0 Hz), 29.4,
27.3. 19F NMR (376 MHz, CDCl3) δ -188.5 (m). MS (EI, 70 eV)
m/e: 57 (C4H9, 100), 132 (m+ - C4H9, 31), 173 (m+ - CH3, 6).
HRMS (EI) m/e: calcd for C9H14O2F (m+ - CH3): 173.0979,
found: 173.0970. TLC mobility, Rf 0.45 (in 4:1 hexane/ethyl
acetate, visualized with PMA).
1
Fluorolactone 18(ox): H NMR (400 MHz, CDCl3) δ 7.14-
7.31 (m, 10H), 4.86 (dt, J ) 49.2, 6.0 Hz, 1H), 4.17 (t, J ) 6.0
Hz, 2H), 2.79 (t, J ) 6.8 Hz, 2H), 2.67 (t, J ) 8.0 Hz, 2H), 2.13-
2.24 (m, 2H), 1.98 (m, 2H). 19F NMR (376 MHz, CDCl3) δ -192.8
(m). TLC mobility, Rf 0.33 (in 20:1 hexane/ethyl acetate, visualized
with UV and PMA).
1
Fluorolactone 16(F-ox): IR (NaCl) ν 1756 (s). H NMR (400
Fluorolactone 18(F-ox): 1H NMR (400 MHz, CDCl3) δ 7.15-
7.33 (m, 10H), 7.43 (dt, J ) 56.2, 4.0 Hz, 1H), 3.38 (t, J ) 8.0
Hz, 2H), 2.69 (t, J ) 7.6 Hz, 2H), 2.52 (t, J ) 7.6 Hz, 2H), 2.22
(m, 2H), 2.01 (m, 2H). 19F NMR (376 MHz, CDCl3) δ -129.2 (dt,
J ) 55.9, 18.0 Hz). TLC mobility, Rf 0.49 (in 20:1 hexane/ethyl
acetate, visualized with UV and PMA).
MHz, CDCl3) δ 5.94 (dd, J ) 50.0, 5.2 Hz, 1H), 2.80 (ddt, J )
14.2, 7.4, 2.0 Hz, 1H), 2.65 (td, J ) 13.6, 2.0 Hz, 1H), 2.44 (ddt,
J ) 14.8, 5.2, 2.4 Hz, 1H), 2.03 (m, 1H), 1.76 (tt, J ) 12.0, 2.4
Hz, 1H), 1.60 (m, 1H), 1.35 (m, 1H), 0.92 (s, 9H). 13C {1H} NMR
(100 MHz, CDCl3) δ 172.6, 106.6 (d, J ) 231.0 Hz), 42.9 (d, J )
3.0 Hz), 35.3 (d, J ) 4.0 Hz), 33.9 (d, J ) 22.0 Hz), 32.7, 27.4,
23.6. 19F NMR (376 MHz, CDCl3) δ -131.7 (dd, J ) 50.0, 32.0
Hz). MS (EI, 70 eV) m/e: 57 (C4H9, 100), 153 (m+ - CH3 - HF,
9), 173 (m+ - CH3, 1). HRMS (EI), m/e: calcd for C9H14O2F (m+
- CH3): 173.0979, found: 173.0969. TLC mobility, Rf 0.61 (in
4:1 hexane/ethyl acetate, visualized with p-anisaldehyde).
Baeyer-Villiger Oxidation of 3-Fluoro-4-heptanone (19).
Baeyer-Villiger Oxidation of 2-Fluorodibenzylacetone (17).
Following the optimized protocol outlined for the oxidation of
fluoroketone 16, fluoroketone 17 (112 mg, 0.49 mmol), mCPBA
Following the optimized protocol outlined for the oxidation of
fluoroketone 16, fluoroketone 19 (857 mg, 6.48 mmol), mCPBA
(1.350 g, 7.82 mmol), and sodium bicarbonate (546 mg, 6.50 mmol)
were reacted to 55% conversion to afford a crude mixture containing
19(ox) (64%) and 19(F-ox) (36%).
(43) Enders, D.; Faure, S.; Potthoff, M.; Runsink, J. Synthesis 2001, 2307.
(44) Denmark, S. E.; Dappen, M. S. J. Org. Chem. 1984, 49, 798.
(45) Marigo, M.; Bachmann, S.; Halland, N.; Braunton, A.; Jorgensen,
K. A. Angew. Chem., Int. Ed. 2004, 43, 5507.
J. Org. Chem, Vol. 71, No. 3, 2006 871

Grein et al.
Fluorolactone 19 (F-ox): IR (neat) ν 1740 cm-1. 1H NMR (500
MHz, CDCl3) δ 6.30 (dt, J ) 55.9, 5.1 Hz, 1H), 2.38 (t, J ) 7.4
Hz, 2H), 1.85 (m, 2H), 1.71 (m, 2H), 1.00 (m, 6H). 19F NMR (400
MHz, CDCl3) δ -146.34 (m). 13C {1H} NMR (500 MHz, CDCl3)
δ 172.30, 104.16 (d, J ) 220.9 Hz), 36.34, 26.94 (d, J ) 22.9 Hz),
22.30, 13.89, 7.55 (d, J ) 5.9 Hz).
(253 mg, 1.46 mmol), and sodium bicarbonate (121 mg, 1.44 mmol)
were reacted to give, after purification by way of column chroma-
tography (silica gel, gradient elution, 20:1 to 18:1 to 16:1 hexane/
ethyl acetate), 23 mg (8% yield) of 21(Cl-ox) and 154 mg (57%
yield) of 21(ox) as white solids.
Chlorolactone 21 (Cl-ox): 1H NMR (400 MHz, CDCl3) δ 6.29
(dd, J ) 7.8, 3.2 Hz, 1H), 2.71 (ddd, J ) 14.2, 7.0, 2.2 Hz, 1H),
2.41 (m, 2H), 2.03 (m, 2H), 1.40-1.52 (m, 2H), 0.93 (s, 9H).
Note: compound too unstable and isolated in too small an amount
to obtain 13C NMR spectra. TLC mobility, Rf 0.29 (in 10:1 hexane/
ethyl acetate, visualized with p-anisaldehyde).
1
Fluorolactone 19(ox): IR (neat) ν 1762 cm-1. H NMR (500
MHz, CDCl3) δ 4.88 (ddd, J ) 49.1, 7.0, 4.5 Hz, 1H), 4.17 (t, J )
6.7 Hz, 2H), 1.95 (m, 2H), 1.70 (m, 2H), 1.00 (m, 6H). 19F NMR
(400 MHz, CDCl3) δ -130.99 (m). 13C {1H} NMR (500 MHz,
CDCl3) δ 170.36, 90.28 (d, J ) 184.2 Hz), 67.28, 26.22 (d, J )
1.5 Hz), 10.64, 9.19 (d, J ) 4.2 Hz).
Chlorolactone 21(ox): 1H NMR (400 MHz, CDCl3) δ 4.89 (dd,
J ) 6.4, 2.0 Hz, 1H), 4.78 (ddd, J ) 12.8, 10.4, 1.2 Hz, 1H), 4.32
(ddd, J ) 12.8, 6.2, 2.0 Hz, 1H), 2.24 (m, 1H), 2.13 (m, 1H), 1.87
(tt, J ) 12.0, 2.8 Hz, 1H), 1.79 (m, 1H), 1.49 (dddd, J ) 13.6,
12.0, 10.4, 2.0 Hz, 1H), 0.92 (s, 9H). 13C {1H} NMR (100 MHz,
CDCl3) δ 169.6, 68.6, 58.3, 43.9, 32.6, 31.4, 30.5, 27.5. TLC
mobility, Rf 0.28 (in 10:1 hexane/ethyl acetate, visualized with
p-anisaldehyde).
Baeyer-Villiger Oxidation of 2-Chlorobenzylacetone (20).
Baeyer-Villiger Oxidation of 3-Chloro-4-heptanone (22).
Following the optimized protocol outlined for the oxidation of
fluoroketone 16, chloroketone 20 (0.084 g, 0.34 mmol), mCPBA
(0.070 mg, 0.41 mmol), and sodium bicarbonate (29 mg, 0.35
mmol) were reacted to give, after purification by way of column
chromatography (silica gel, 40:1 hexane/ethyl acetate), 40 mg (45%
yield) of 20(ox). A second fraction containing 20 (36 mg, 43%
recovery) was also collected.
Following the optimized protocol outlined for the oxidation of
fluoroketone 16, chloroketone 22 (1.550 g, 10.43 mmol), mCPBA
(1.985 g, 11.47 mmol), and sodium bicarbonate (876 mg, 10.43
mmol) were reacted for 18 days to afford a crude mixture containing
22 (78%) and 22(ox) (22%). Bulb to bulb distillation (100 mmHg,
120 °C) afforded a sample (1.18 g, 76% mass balance) of 22 and
22(ox).
1
Chlorolactone 20(ox): IR (neat) ν 1752 cm-1. H NMR (400
MHz, CDCl3) δ 7.02-7.29 (m, 10H), 5.46 (s, 1H), 5.23 (d, J )
12.4 Hz, 1H), 5.16 (d, J ) 12.4 Hz, 1H). 13C {1H} NMR (100
MHz, CDCl3) δ 173.0, 139.3, 131.4, 129.9, 128.8, 128.6, 127.4,
127.1, 126.9, 79.9, 71.7. HRMS (EI), m/z: calcd for C15H13O2Cl
(m+ - H): 260.0604, found: 260.0604.
1
Chlorolactone 22(ox): IR (neat) ν 1747 cm-1. H NMR (500
Baeyer-Villiger Oxidation of trans-2-Chloro-4-t-butylcyclo-
hexanone (21).
MHz, CDCl3) δ 4.25 (dd, J ) 7.6, 6.0 Hz, 1H), 4.14-4.18 (m,
2H), 2.06-2.10 (m, 2H), 1.71-1.74 (m, 2H), 1.04-1.08 (m, 3H),
0.96-0.99 (m, 3H). 13C {1H} NMR (500 MHz, CDCl3) δ 170.17,
67.85, 59.37, 28.81, 22.26, 11.04, 10.91. HRMS (CI), m/z: calcd
for C7H13O2Cl (m+ - H): 165.0682, found: 165.0724.
Supporting Information Available: Complete ref 31, spectra
for selected compounds, discussion of effect of halogen conformers
on the stability of TS2, and comparison with other theoretical work.
This material is available free of charge via the Internet at
http://pubs.acs.org.
Following the optimized protocol outlined for the oxidation of
fluoroketone 16, chloroketone 21 (251 mg, 1.32 mmol), mCPBA
JO0513966
872 J. Org. Chem., Vol. 71, No. 3, 2006