Formation of δ-Lactones with anti-Baeyer–Villiger Regiochemistry: Investigations into the Mechanism of the Cerium-Catalyzed Aerobic Coupling of β-Oxoesters with Enol Acetates

Article abstract of DOI:10.1002/chem.201605468

The cerium-catalyzed, aerobic coupling of β-oxoesters with enol acetates and dioxygen yields δ-lactones with a 1,4-diketone moiety. In contrast to the Baeyer–Villiger oxidation (BVO), where the higher substituted residue migrates; in the case of this oxid

Full text of DOI:10.1002/chem.201605468

A Journal of
Accepted Article
Title: Formation of δ-Lactones with anti-Baeyer-Villiger Regiochemistry
- Investigations into the Mechanism of the Cerium-Catalyzed
Aerobic Coupling of β-Oxoesters with Enol Acetates
Authors: Jens Christoffers, Irina Geibel, Anna Dierks, and Thomas
Müller
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: Chem. Eur. J. 10.1002/chem.201605468
Link to VoR: http://dx.doi.org/10.1002/chem.201605468
Supported by

10.1002/chem.201605468
Chemistry - A European Journal
Formation of δ-Lactones with anti-Baeyer-Villiger Regiochemistry –
Investigations into the Mechanism of the Cerium-Catalyzed Aerobic
Coupling of β-Oxoesters with Enol Acetates
Irina Geibel,[a] Anna Dierks,[a] Thomas Müller,[a] Jens Christoffers*[a]
[a]I. Geibel, A. Dierks, Prof. Dr. T. Müller, Prof. Dr. J. Christoffers
Institut für Chemie, Carl von Ossietzky Universität Oldenburg,
26111 Oldenburg (Germany)
E-Mail: jens.christoffers@uni-oldenburg.de
Homepage: http://www.uni-oldenburg.de/oc-christoffers/
Abstract: The cerium-catalyzed, aerobic coupling of β-oxoesters with enol acetates
and dioxygen yields δ-lactones with a 1,4-diketone moiety. In contrast to the Baeyer-
Villiger oxidation (BVO), where the higher substituted residue migrates, in this case of
an oxidative C-C coupling reaction the less substituted alkyl residue undergoes a 1,2-
shift. An endoperoxidic oxycarbenium ion comparable to the Criegee intermediate in
the BVO is proposed as a reaction intermediate and submitted to conformational
analysis by computational methods. As a result, the inverse regiochemistry is
explained by a primary stereoelectronic effect. A Hammett analysis using different
donor and acceptor substituted enol esters provides support for the oxycarbenium ion
being the crucial intermediate in the rate determining step of the conversion. An
overall mechanism is suggested with a radical chain reaction for the formation of
endoperoxides from β-oxoesters, enol acetates and dioxygen with a cerium(IV)
species as initiating reagent.
1
This article is protected by copyright. All rights reserved.

10.1002/chem.201605468
Chemistry - A European Journal
Key Topic: Reaction Mechanism
Graphical Abstract: Reaction kinetics (Hammett analysis) supported by
computational methods indicate an endoperoxidic oxycarbenium ion to be the
reaction intermediate for the cerium-catalyzed coupling of β-oxoesters, enol acetates
and dioxygen.
OAc
O
Ph
O
Ph
O
Ph
O
O
O
+ O2 (air)
HO
Ph
cat. CeCl3 7 H O
O
O
CO2Et
O
via
+
CO2Et
CO2Et
CO2Et
·
2
F3CCH2OH, 30°C, 1 d
0%
Intermediate
59%
Keywords: Lactones, Baeyer-Villiger oxidation, Oxidation, Peroxides, Cerium,
Oxoesters, Enol esters, Hammett correlation, DFT calculations, Stereoelectronic
effect
Introduction
In the year 1899 Baeyer and Villiger reported on the formation of lactones from the
reaction of cyclic ketones with monopersulfuric acid.[1] Since then, this method was
successfully established as an essential and extraordinary versatile method for the
preparation of esters and lactones from ketones and was consequently named
Baeyer-Villiger oxidation (BVO) in honor its discoverers.[2] The standard organic
reagent for this reaction is nowadays mCPBA,[3] although several other peracids or
even hydrogen peroxide or dioxygen[4] with[5] or without metal catalyst can be applied
for this transformation. A particularly interesting issue is the use of monooxygenases
2
This article is protected by copyright. All rights reserved.

10.1002/chem.201605468
Chemistry - A European Journal
in enzymatic BVO.[6] If unsymmetrically substituted ketones are submitted to the
BVO, it is generally observed that the higher substituted alkyl residue or – in cases of
aryl ketones – the phenyl ring is migrating. The origin of this regionselectivity was
understood to be a stereoelectronic effect within the so called Criegee intermediate of
the BVO (see Scheme 5 below).[7] There are however very few exceptions from this
regiochemistry. One of the prominent examples for this opposite regiochemistry is the
lactone formed by BVO of camphor,[8] which was already reported by Baeyer and
Villiger themselves.[1]
Very recently, we have discovered a new aerobic, cerium-catalyzed coupling reaction
of enol acetates and β-oxoesters leading to δ-lactones with an 1,4-dicarbonyl moiety
in the side chain.[9] With this work, we would like to uncover the origin of the unusual
regiochemistry of the lactone formation, which is contrary to expectations known from
the BVO. This recent discovery is based on our earlier work, where we found a new
route for the preparation of 1,4-diketones 5 by oxidative coupling of β-oxoesters like
compound 1 with styrene.[10] Lately, we were able to improve these reactions by
using enol esters 2 instead of olefins.[11] The cerium-catalyzed reaction utilizes
molecular oxygen (from air) and can be regarded as an oxidative Umpolung reaction.
During the course of our investigations, we had observed the formation of γ-oxo-δ-
lactones 3 as byproducts. Most recently, we had optimized this transformation to-
wards the formation of δ-lactones 3 as the main products of this cerium-catalyzed
coupling of β-oxoesters 1, enol acetates 2 and dioxygen (Scheme 1).[9] From a me-
chanistic point of view, we propose the initial formation of an endoperoxidic
1,2-dioxane derivative 6 by cerium-catalyzed coupling reaction of the two starting
materials 1 and 2 and dioxygen. Although the exact mechanism of this radical chain
reaction leading to intermediate product 6 remains unclear, we have experimental
precedence for its existence. We have isolated and fully characterized several
3
This article is protected by copyright. All rights reserved.

10.1002/chem.201605468
Chemistry - A European Journal
analogues earlier including X-ray diffraction data in some cases.[10a, 10c] Anyhow, with
the assumption of intermediate product 6, the 1,4-diketone 5 is formed by hydrolysis
of the two acetal moieties of endoperoxide 6 (actually, we were able to detect
hydrogen peroxide in the reaction mixture in these cases). When driving the
selectivity towards lactone formation by the proper choice of reaction conditions,
namely the use of CF3CH2OH (TFE) as the optimal solvent at 30°C, we propose the
loss of an acetate ion under formation of the cationic intermediate 7, which is
facilitated by hydrogen bonding with the solvent TFE. Similar to the BVO, then
heterolytic O-O bond cleavage with concerted 1,2-alkyl shift may occur. Interestingly,
the primary alkyl group migrates (as depicted in structure 7) instead of the quaternary
carbon atom (structure 7'), since compound 4 is not observed within the reaction
mixture. This regiochemistry is just contrary to the one generally observed for the
BVO. Therefore, we propose the term anti-Baeyer-Villiger (anti-BV) regiochemistry. In
this manuscript we aim to support the mechanistic picture drawn in Scheme 1 by
introduction of electron withdrawing or donating substituents at the phenyl ring in
order to destabilize or stabilize a cationic oxycarbenium intermediate 7. Furthermore,
we wish to figure out whether the anti-BV regiochemistry is the result of an electronic
or steric effect or both.
4
This article is protected by copyright. All rights reserved.

10.1002/chem.201605468
Chemistry - A European Journal
OAc
Ph
O
Ph
O
Ph
O
Ph
O
O
O
2a (1.5 equiv.)
O
O
+
+
CO2Et
O
CO2Et
CO2Et
O2 (air)
CO2Et
cat. CeCl3 · 7 H2O
(59%)
- H
(0%)
(21%)
5a
1
3a
4
- H
OAc
O
O
Ph
O
Ph
O
Ph
O
O
- AcO
HO
HO
HO
or
CO2Et
CO2Et
CO2Et
6
7
7'
Scheme 1. Proposed mechanism for the title reaction of this investigation.
Results
BVO of model substrates. In order to find the reason for the unexpected
regiochemistry of the lactone formation we have prepared some model substrates
and submitted it to BVO (Scheme 2, Table 2); suitable standard conditions were first
of all established for the dimethyl substituted ketone 8a. Using an excess of mCPBA
in the presence of KHCO3 we could isolate the expected regioisomer 9a[12] in
quantitative yield (Table 1, entry 1),[13] however, the anti-BV regioisomer was not
formed. We then adopted these conditions for the sterically more hindered benzyl-
substituted ketone 8b.[14] The expected product 9b[15] was formed in significantly
lower yield (entry 2), but importantly, an anti-BV product was neither isolated nor
detected by GC-MS. Actually, a steric influence on the regiochemistry is not
expected, since an 1,2-alkyl shift proceeds in a suprafacial manner, thus substituents
R and R' should not influence the migration. Anyhow, there could be an electronic
influence. Therefore, we converted the β-oxoesters 8c[16] and 8d,[17] but again, only
5
This article is protected by copyright. All rights reserved.

10.1002/chem.201605468
Chemistry - A European Journal
the expected regioisomers 9c and 9d were formed, in case of the methyl-substituted
product 9c[16b] in quantitative yield (entry 3). With more steric hindrance the yield was
again lower (product 9d, entry 4).
O
O
xs. CPBA
xs. KHCO3
m
O
R
R
R'
R'
CH2Cl2, 40°C, 1 d
8a-8d
9a-9d
Scheme 2. BVO of model substrates.
Table 1. BVO of model substrates.
Entry
Starting Material
R
R'
Product
Yield[a]
1
2
3
4
8a
Me
Me
Me
Bn
Me
9a[12]
9b[15]
9c[16b]
9d
100%
48%
8b[14]
8c[16]
8d[17]
Bn
CO2Et
CO2Et
100%
67%
[a] Yield of isolated product.
We then turned to the 1,4-diketone 5a and applied the same reaction conditions as
above (Scheme 3). Again, the expected BV regiochemistry dominates the product
distribution (the combined yield of compounds 4 and 10 is 29%). But interestingly, the
anti-BV product 3a is also obtained with 13%. The low overall yield could be
interpreted either in terms of steric hindrance, but also by hydrolysis of the phenyl
ester moiety formed by side chain oxidation as present in compound 10, which could
have been occurred during the reaction or workup. Furthermore, we have converted
oxoester 1 with mCPBA under the same reaction conditions. Though the starting
material 1 was fully consumed after approx. 6 h, no unique reaction products were
isolable after column chromatography (apart from excess of mCPBA and
3-ClC6H4CO2H). The lactone expected from compound 1 by BVO was actually
6
This article is protected by copyright. All rights reserved.

10.1002/chem.201605468
Chemistry - A European Journal
reported only once before, but accessed by a completely different route.[18]
O
Ph
O
Ph
O
Ph
O
OPh
O
10 equiv. CPBA
2 equiv. KHCO3
O
m
O
O
O
O
+
+
O
CO2Et
CO2Et
CO2Et
CO2Et
CH2Cl2, 40°C, 1 d
5a
3a (13%)
4 (23%)
10 (6%)
Scheme 3. BVO of 1,4-diketone 5a.
Conversion of substituted enol esters. In order to gain evidence for a cationic
intermediate of the title reaction we investigated several enol acetates 2a–2h with
electron donating or accepting substituents at the para-position of the phenyl ring
(Scheme 4). For comparison of the yields of lactones 3a–3h identical reaction
conditions were chosen. Apart from the main products 3a–3h also the 1,4-diketones
5a–5h and the α-hydroxylated compound 11[19] were formed as byproducts, which
were isolated and quantified after purification (in the case of ester 2h, the alcohol 11
was even the main product of the conversion; Table 2, entry 8). For yields and
substituents X see Table 2. Actually, the reactions of enol esters 2a, 2b, 2f and 2h
were reported by us before.[9]
OAc
X
X
O
O
X
O
O
O
O
2a-2h (1.5 equiv.)
OH
CO2Et
+
+
O
CO2Et
CO2Et
5a-5h
CO2Et
TFE, 30°C, 18 h, air
1 mol% CeCl3 · 7 H2O
-
1
3a 3h
11
Scheme 4. Cerium-catalyzed aerobic conversion of oxoester 1 with several enol
esters 2 under comparable conditions.
7
This article is protected by copyright. All rights reserved.

10.1002/chem.201605468
Chemistry - A European Journal
Table 2. Results for different enol acetates 2a–2h.
[b]
[c]
Entry Enol Ester
X
Yields[a]
σX
kX
1
2
3
4
5
6
7
8
2b
2c
2a
2d
2e
2f
OMe 66% (3b)
15% (5b)
28% (5c)
21% (5a)
19% (5d)
27% (5e)
7% (11) –0.28
2% (11) –0.14
0.342
0.244
0.276
0.248
0.212
0.213
0.148
0.116
Me
H
59% (3c)
59% (3a)
54% (3d)
47% (3e)
44% (3f)
35% (3g)
7% (3h)
0% (11)
2% (11)
2% (11)
0
F
0.15
0.24
0.26
0.70
0.81
Cl
Br
19% (5f) 35% (11)
15% (5g) 26% (11)
2% (5h) 36% (11)
2g
2h
CN
NO2
[a] Yields of isolated products after chromatographic purification. [b] Hammett
constant for the para-substituent X; values were taken from the literature.[20] [c] Rate
constants (slopes in Figure 2).
The results summarized in Table 2 point out, that there is an obvious correlation of
the yields of lactones 3a–3h with the electron donor or acceptor properties of the
para-substituents at the phenyl rings, which are quantified by the substituent constant
σX. This correlation already qualitatively supports an oxycarbenium ion 7 as the
intermediate of the reaction. To obtain a quantitative evidence for this correlation, the
kinetics of the reaction were followed by GLC. As an essential precondition for this
attempt, all compounds collected in Table 2 were available as GC-standards. The
eight enol esters 2a–2h were submitted to the reaction with β-oxoester 1 under the
same conditions (concentrations, temperature, and stoichiometry) with mesitylene as
internal integration standard. Samples were taken in defined time intervals and
quenched by dilution with dichloromethane (DCM). In Figure 1, relative peak areas of
lactones [3x] (which were obtained by division of the peak areas of product 3x by the
peak area of the internal standard mesitylene) are plotted vs. time. At about 12 h, the
concentration of compounds 3x reached a maximum, thereafter it decreased with a
8
This article is protected by copyright. All rights reserved.

10.1002/chem.201605468
Chemistry - A European Journal
minimal slope, probably due to decomposition reactions, e.g. ester hydrolysis or
retro-Claisen reactions. Nevertheless, since this would hamper the successive
analysis, only data up to 12 h were used for the numerical analysis. A precondition of
the linear free energy relationship according to Hammett is first order kinetics.[20]
However, data for the starting materials 1 or 2x could not be used for this kinetic
analysis for several reasons: First of all, two byproducts 5a–5h and 11 are always
formed with different rate constants or even different reaction order. Secondly, the
integration of compound 1 was not very accurate due interconversion of the keto and
enol tautomer on the GC column. Furthermore, the enol esters 2x were used in
excess. For this reason, a replacement for the concentrations of 1 or 2x was
constructed by the following term: [3x](12 h) - [3x](t), herein [3x](12 h) represents the
"end concentration" of product 3x at t = 12 h. The difference [3x](12 h) - [3x](t)
therefore shall be equal to the portion of starting material 1, which was converted with
enol ester 2x along the reaction pathway leading to lactone 3x. The complementary
fraction of compound 1 converted to byproducts 5x and 11 is faded out by this
operation. Consequently, the term ln{[3x](12 h) - [3x](t)} was plotted vs. time. In
Figure 2, the graphs clearly show first order kinetics due to their linearity. Moreover,
rate constants kX can be obtained from the regression lines. These data are listed in
the last column of Table 2. Finally, the Hammett correlation was established by
plotting log(kX/kH) vs. the para-substituent constant σX. However, the data point for X
= Me was left out, since it was not fitting at all into the linear regression line. The
negative slope of this linear correlation indicates the existence of a carbenium ion in
benzylic position as intermediate in the rate determining step of the lactone
formation. We propose, that this species is represented by the oxycarbenium ion 7.
9
This article is protected by copyright. All rights reserved.

10.1002/chem.201605468
Chemistry - A European Journal
Figure 1. Emergence of products 3a–3h; the relative peak areas [3x] obtained by
peak areas of the lactone 3x divided by the internal standard are plotted at the
ordinate.
Figure 2. Plot of ln{[3x](12 h) – [3x](t)} vs. time; for the color code see Figure 1; the
10
This article is protected by copyright. All rights reserved.

10.1002/chem.201605468
Chemistry - A European Journal
slopes of the linear equations are the rate constants kX.
Figure 3. Hammett plot with log(kX/kH) vs. substituent (para) constant σX.
Computational Results. With two stereocenters, the reaction intermediate 7 exists
as two diastereoisomers (Figure 4); they can be regarded as the trans- and cis-
cyclotautomers (α- and β-anomers within the carbohydrate terminology) of a
monocyclic open chain form. Whereas there has only one conformer to be
considered for the trans-diastereoisomer 7a (with ester E and 9-OH both axial), the
cis-diastereoisomer could adopt two conformations 7b (with E axial and 9-OH
equatorial) and 7c (with E equatorial and 9-OH axial). Although oxycarbenium ions
7b and 7c depicted in Figure 4 are enantiomeric compounds, this does not play a
role for the investigation and for the following discussion, since all compounds in this
study are racemates. Structure optimizations for all three isomers 7a–7c were carried
out using density functional theory (DFT) calculations applying the M06-2X functional
and 6-311+G(d,p) basis set[21a] (all computations were done with the Gaussian09, B1
11
This article is protected by copyright. All rights reserved.

10.1002/chem.201605468
Chemistry - A European Journal
program;[21b] see the Supporting Information for details). For these computations, E =
CO2Me was chosen for simplicity. A self-consistent reaction field (SCRF) calculation
using the PCM model with appropriate parameter modeled the solvent TFE. This
model chemistry predict compound 7c to be the thermodynamically most stable
isomer (Table 3). Both diastereoisomers 7a (trans) and 7c (cis) are comparable in
energy (E = 9 kJ mol–1). The second cis-conformer 7b is however E = 19 kJ mol–1
higher than isomer 7c, therefore, this conformation is not significantly populated at
30°C. The dihedral angles C8-C9-O1-O2 (two blue bonds) and C5-C9-O1-O2 (a blue
and a red bond) were read out of the optimized geometries. For all three isomers, the
C5-C9-O1-O2 is about 60° (syn-gauche). The formation of a product 4 with BV
constitution would require migration of the red bond C5-C9 to oxygen atom O1, while
the blue bond O1-O2 is cleaved. For both more stable isomers 7a and 7c, the angle
C8-C9-O1-O2 is about 180° (antiperiplanar). Migration of the blue C8-C9 bond to O1
forms the product 3 with anti-BV constitution. We were able to identify the transition
state for this rearrangement. The barrier is Erel = 62.5 kJ mol–1 (Grel = 54.2 kJ mol–1).
The energy profile and a movie visualizing the transformation are included in the
supporting information.
12
This article is protected by copyright. All rights reserved.

10.1002/chem.201605468
Chemistry - A European Journal
2
1O
O
Ph
HO
5
9
8
E
7
E
E
Ph
Ph
Ph
E
O
O
O
HO
O
O
O
HO
HO
7b (
,
)
7c (
,
)
eq
(
ci.s. E
ci.s.E
7a trans)
ax
..
C5
O2
C8
C5
O2
OH
C5
O2
C8
.
.
.
.
.
.
OH
C8
OH
Figure 4. Oxycarbeniumion 7, atom numbering, relative configuration of the trans-
diastereoisomer 7a, and the two cis-conformers 7b and 7c; Newman projections of
the ecliptic conformations of the C9-O1 bond show, that the blue bonds are in
antiperiplanar conformation; E = CO2Me.
Table 3. Calculated relative energies, Erel, relative Gibbs energies at 298 K, Grel, and
selected structural parameters for oxycarbenium ions 7a–7c [M06-2x/6-311+G(d,p)].
Intermediate
7a
+9
7b
7c
0
Erel / kJ mol–1
Grel / kJ mol–1
+19
+20
+13
0
dihedral angle
C8-C9-O1-O2
174°
58°
69°
51°
179°
66°
C5-C9-O1-O2
13
This article is protected by copyright. All rights reserved.

10.1002/chem.201605468
Chemistry - A European Journal
Discussion
Earlier studies on the mechanism of the BVO revealed the following picture (Scheme
5):[22] Upon a BVO of the unsymmetric ketone 11 to the ester 13, the migrating group
Rm should be positioned in antiperiplanar conformation to the O1-O2 bond in the
Criegee intermediate 12. This conformation is, for example, favourable, if Rm is larger
than R. This requirement is called the primary stereoelectronic effect. There is
actually another, so-called secondary stereoelectronic effect, discussed in the
literature, which requires the residue Rm being in antiperiplanar conformation of one
of the lone pairs of the hydroxy group at C9. However, this secondary
stereoelectronic effect does not play a role for the regioselectivity in this investigation.
In order to support the primary stereoelectronic effect, Chandrasekhar and Roy[23]
have performed the following experiment for an intramolecular BVO with a bicyclic
Criegee intermediate 14: The carboxylic acid side chain of ketone 11a was
transformed into the peracid which is in equilibrium with its cyclotautomer 14. The
analogy to the constitution of our proposed intermediate 7 is obvious. The two blue
bonds C8-C9 and O1-O2 are in antiperiplanar conformation, thus, the primary carbon
atom C9 migrates rather than the tertiary C5 leading to the product 13a with formal
anti-BV constitution.
O
Ar
O 2
..
1 O
O
O
ArCO3H
ArCO3H
R
Rm
HO
Rm
.
RmO
O
R
Rm
11
CO2H
R
9
R
.
O2
12
13
OH
O
CO2H
O
5
O
O
O 2
1
9
HO
11a
14
13a
Scheme 5. Conformation of the Criegee intermediate 12 of the BVO and
14
This article is protected by copyright. All rights reserved.

10.1002/chem.201605468
Chemistry - A European Journal
Chandrasekhar's model system 11a for the primary stereoelectronic effect; the
numbers of C9, O1 and O2 and the color codes red and blue are adopted from the
1,2-dioxabicyclo[4.3.0]nonane skeleton of species 7 in Figure 4.
The question for an overall mechanistic picture is, whether an endoperoxidic 3-aryl-3-
acetoxy-1,2-dioxabicyclo[4.3.0]nonane derivative 6, an intermediate compound,
which could so far neither be isolated from nor detected in the reaction mixture,
exists. However, in our preceding investigations on the cerium-catalyzed, aerobic
conversion of β-oxoesters or α-acetyllactones with styrene or α-methylstyrene we
were able to isolate and characterize several congeners 15 of the structural type
represented by proposed intermediate 6. Five of these compounds 15a–15e were
even suitable for single crystal X-ray crystallography (Figure 5). One compound is
monocyclic (15c), two are annulated (15a and 15b), and two others are spirocyclic
(15d and 15e). The structures were already published earlier[10a, 10c] and crystal data
have been deposited.[24] Structural similarities are apart from the 1,2-dioxane ring the
tertiary alcohol at C9, the quaternary carbon center C5 carrying an ester moiety and
a phenyl substituent at C3. The most significant difference is of course the acetoxy
group at C3, which is supposed to be actually the reason for the intrinsic instability of
compound 6 compared to isolable compounds 15.
15
This article is protected by copyright. All rights reserved.

10.1002/chem.201605468
Chemistry - A European Journal
2
OAc
Ph
Me
Ph
1O
O
O
O
O
O
O
Ph
3
HO
HO
HO
5
9
CO2Et
Ph
CO2Et
CO2Me
6
15a
15b
Me
Ph
O
O
Ph
O
O
O
O
HO
Me
HO
Me
HO
Me
O
Me CO2Et
O
O
15c
15d
15e
Figure 5. Comparision of postulated intermediate 6 with structurally characterized
analogues 15a–15e.
With the data collected above the following general view of the overall reaction
mechanism can be developed; it consists of three stages:
(1) Start of the radical chain reaction (Scheme 6),
(2) propagation with coupling of enol ester 2a and dioxygen (Scheme 7), and
(3) decomposition of the 1,2-dioxane derivative 7 (Scheme 8).
Stage (1): The radical chain reaction starts with the α-oxidation of the β-oxoester 1 by
a Ce(IV) species, a process, which is actually well known for the stoichiometric
oxidation of carbonyl compounds with CAN. Actually, the precatalyst in our case is a
Ce(III) salt. It is proposed, that Ce(III) coordinates to the β-oxoester 1 under
formation of a diketonato complex 16 (Scheme 6) in similarity of the well-known
chemistry of Fe(III) and other trivalent metal ions.[25] The anionic ligand then
facilitates the oxidation of Ce(III) to Ce(IV) with dioxygen and the Ce(IV)-diketonato
complex 17a is formed. Formation of a radical species 17b with spin density at the α-
position can be proposed as ligand-to-metal charge transfer within Ce(IV)-diketonato
complex 17a.[26] In the catalytic cycle, dissociation of complex 17b liberates the ligand
16
This article is protected by copyright. All rights reserved.

10.1002/chem.201605468
Chemistry - A European Journal
18 and Ce(III) is finally reoxidized to Ce(IV) by air (upper cycle in Scheme 7).
Obviously, coordination of Ce(III) to β-oxoesters shifts the redox potential and makes
this air-oxidation possible. Therefore, the β-oxoester 1 is not only the substrate of the
reaction, but also plays the role of a ligand for Ce(III) centers necessary for the
reoxidation their Ce(IV).
2+
3+
+ ¼ O2
+ H+
O
O
O
O
O
O
+III
CeLn
+III
CeLn
+IV
CeLn
- ½ H2O
EtO
EtO
EtO
16
17a
17b
Scheme 6. Start of the radical chain reaction by air-oxidation of Ce(III)-diketonato
complex 16.
The radical chain reaction of stage (2) is summarized in Scheme 7. After initial
formation of the α-radical 18 by Ce(IV), addition to the enol acetale 2a furnishes the
benzyl radical 19, which gains further stabilization by the +M-effect of the acetoxy
group. Reaction of radical 19 with dioxygen leads to peroxy-species 20, which closes
the radical chain reaction by H atom abstraction from oxoester 1 with regeneration of
α-radical 18 under formation of hydroperoxide 21. The latter undergoes a
cyclotautomerization to the 1,2-dioxane derivative 6. Using the carbohydrate
terminology, compound 21 defines the open-chain form, and compound 6 exists as
two equilibrating α- and β-anomers (epimers at C9). In the case of an unspecified
interruption of the radical chain mechanism, the α-radical can continuously be
regenerated by Ce(IV) oxidation of the oxoester 1. Furthermore and as pointed out
above in Scheme 6, the formation of Ce(III)-diketonato complexes facilitates the
reoxidation of Ce(III) complexes with dioxygen to Ce(IV) species.
17
This article is protected by copyright. All rights reserved.

10.1002/chem.201605468
Chemistry - A European Journal
+ O2
LnCe3+
LnCe4+
Ph
OAc
O
O
+
CO2Et
CO2Et
2a
18
1
Ph OAc
O
CO2Et
O O
OAc
O OH
OAc
19
OAc
Ph
Ph
Ph
O
O
O
O
HO
+ O2
9
CO2Et
CO2Et
CO2Et
20
21
6
Scheme 7. Chain propagation and formation of 1,2-dioxane derivative 6.
Finally, in stage (3), compound 6 could find two pathways for decomposition to
product 3a and byproduct 5a (Scheme 8): First of all, acid catalyzed hydrolysis could
furnish the 1,4-diketone 5a. Byproduct of this process is hydrogen peroxide. The
obviously preferred reaction pathway is, however, the loss of an acetate ion from
compound 6. We propose that this process is facilitated by H bonding of acetate to
the solvent TFE. The oxycarbenium ion 7, which could be regarded as a species
analogous to the Criegee intermediate of the BVO, is particularly stabilized by
conjugation with the phenyl ring as well as the oxygen atom O2 of the dioxane
moiety. Actually, the Hammett correlation depicted in Figure 3 strongly supports the
formation of this intermediate cation in the rate determining step. Electron donating
substituents at the para-position of the phenyl ring result in an acceleration of the
reaction, while electron withdrawing decrease the reaction rates. Furthermore, a
18
This article is protected by copyright. All rights reserved.

10.1002/chem.201605468
Chemistry - A European Journal
conformational analysis of the both diastereoisomers of compound 7, namely trans-
isomer 7a and cis-isomer 7c (Figure 4), has pointed out that only the C8-C9 bond is
in antiperiplanar position to the O1-O2 bond. Thus, the formal anti-BV regiochemistry
of the product is the result of the so-called primary stereoelectronic effect, which is
operative in the rearrangement of the intermediate 7. The second cis-conformer 7b
cannot find a pathway for decomposition to product 3a due to the lack of appropriate
antiperiplanar orientation of the migrating bonds. In addition, it is too high in energy to
be significantly populated at 30°C.
OAc
Ph
O
O
Ph
O
O
HO
+ H2O, cat. H
CO2Et
CO2Et
- AcOH, - H2O2
6
5a
- AcO
Ph
O
Ph
O
O
O
HO
O
CO2Et
CO2Et
- H
7
3a
Scheme 8. Decomposition of intermediate product 6 either by acid catalyzed
hydrolysis under formation of 1,4-diketone 5a or via oxycarbenium ion 7 with
formation of lactone 3a with anti-BV constitution.
Conclusion
δ-Valerolactone derivatives 3 with an 1,4-diketone moiety are formed by aerobic,
cerium-catalyzed coupling of β-oxoester 1, enol acetates 2 and dioxygen. This
transformation can be regarded as an Umpolung reaction, since the normally
nucleophilic α-carbon atom of the β-oxoester becomes an electrophilic radical center
reacting with the electron rich olefin 2. Furthermore, the catalyst metal cerium is non-
19
This article is protected by copyright. All rights reserved.

10.1002/chem.201605468
Chemistry - A European Journal
toxic and environmentally benign, atmospheric oxygen is abundant and the
conversion is highly atom-economic. Therefore, this reaction could be considered as
a contribution to the field of sustainable chemistry. In this article, an overall
mechanistic picture of the reaction was drawn consisting of three parts: (1) Initiation
by a Ce(IV) species, (2) radical chain propagation and (3) decomposition of
endoperoxide 6. First of all, a radical chain reaction is started by α-oxidation of the
oxoester 1 by a Ce(IV) species, which was formed by air-oxidation of Ce(III). This
step is actually making the overall transformation an Umpolung. A precondition for
the oxidation of Ce(III) to Ce(IV) is the coordination of the metal center to the β-
oxoester 1 under formation of a diketonato-complex 16. A respective Ce(IV)-
diketonate species 17a is then undergoing intramolecular ligand-to-metal charge
transfer (complex 17b), thus, generating a Ce(III) center and the radical species 18.
Secondly, the radical chain propagates by addition of radical 18 to the enol ester 2
before addition to dioxygen occurs. The radical chain is closed by α-hydrogen
abstraction from the oxoester 1 by the peroxy radical 20 under formation of the
hydroperoxide 21, which is in equilibrium with its cyclotautomer 6. The latter
intermediate 6 could so far neither be isolated nor spectroscopically detected,
although its existence is supported by structural precedence of similar endoperoxides
15, which were formed in reactions of oxoesters with styrene or α-methyl styrene in
the presence of dioxygen. The third part exhibits similarities to the BVO. Herein
compound 6 eliminates an acetate ion and forms an oxycarbenium ion 7. The
existence of this ion 7 is strongly supported by the kinetics of the reaction of several
donor and acceptor substituted enol acetates 2a–2h. Donor substituents accelerate
the reaction, whereas acceptor substituents slow it down. An analysis of the first
order kinetics resulted in rate constants k, which were linearly correlated with the
para-substituent constants σ in a Hammett plot. The negative slope provides
20
This article is protected by copyright. All rights reserved.

10.1002/chem.201605468
Chemistry - A European Journal
evidence for the existence of the benzyl cation 7, which could be considered as an
analog to the Criegee intermediate of the BVO. Furthermore, the relative energies
and optimized conformations of three possible stereoisomers 7a–7c were obtained
by DFT calculations. A conformational analysis of the two relevant isomers 7a and 7c
explains the regiochemistry of the rearrangement reaction by a primary stereoelec-
tronic effect. In contrast to the BVO, where the higher substituted residue is migra-
ting, the primary alkyl residue of intermediate 7 undergoes an 1,2-alkyl shift and
forms the δ-valerolactone 3.
Experimental Section
General: Preparative column chromatography was carried out using Merck SiO2 (35–
70 µm, type 60 A) with hexanes, tert-butyl methyl ether (MTBE), and ethyl acetate
1
(EA) as eluents. TLC was performed on aluminum plates coated with SiO2 F254. H-,
19F- and 13C-NMR spectra were recorded on a Bruker Avance DRX 500 and 300
instruments. Multiplicities of carbon signals were determined with DEPT experiments.
HRMS spectra of products were obtained with a Waters Q-TOF Premier (ESI) spec-
trometer. IR spectra were recorded on a Bruker Tensor 27 spectrometer equipped
with a "GoldenGate" diamond ATR unit. GLC analyses were performed on a Focus
GC (Thermo-Fisher) on a SP-Sil 19CB capillary column (Macherey-Nagel, 30 m, 0.25
mm) with H2 as carrier gas. The starting materials 1 and 8a were commercially
available. The following compounds were literature known and prepared according to
14b]
published procedures: 2-methylcyclopentanone,[14a,
2c,[27] 2d,[27] 2e,[27] 2g,[27]
8c,[16] and 8d.[17] The synthesis of enol acetates 2a, 2b, 2f and 2h and their reactions
of oxoester 1 yielding 3a, 3b, 3f, 3h, 5a, 5b, 5f, 5h and 11 were reported by us
earlier.[9]
21
This article is protected by copyright. All rights reserved.

10.1002/chem.201605468
Chemistry - A European Journal
2-Benzyl-2-methylcyclopentanone (8b): Under exclusion of air and moisture
(nitrogen atmosphere), nBuLi (1.2 mL of a 2.5 mol/L solution in hexane, 3.0 mmol)
was added dropwise within 10 min to a solution of HMDS (0.6 mL, 3.0 mmol) in abs.
hexane (1.2 mL). After heating the mixture for 1 h to reflux, abs. THF (5 mL) was
added and the mixture was cooled to –78°C. A solution of 2-methylcyclopentanone
(269 mg, 2.74 mmol) in abs. THF (1 mL) was added dropwise within 15 min and the
mixture was stirred for further 3 h at 0°C. Benzyl bromide (515 mg, 3.01 mmol) was
then added dropwise at this temperature. The resulting mixture was stirred at
ambient temperature overnight, subsequently diluted with hydrochloric acid (5 mL, 1
mol/L) and extracted with Et2O (3 x 5 mL). The combined organic layers were
washed with brine (10 mL), dried (MgSO4), filtered and the solvent was evaporated in
vacuo. The residue was purified by column chromatography (SiO2, hexanes/MTBE
10:1, Rf = 0.23) to yield the title compound 8b (115 mg, 0.61 mmol, 22%) as a
1
colorless liquid. H NMR (300 MHz, CDCl3): δ = 1.02 (s, 3H), 1.59–1.86 (m, 3H),
1.91–2.11 (m, 2H), 2.24–2.35 (m, 1H), 2.59 (d, J = 13.3 Hz, 1H), 2.86 (d, J = 13.3 Hz,
1H), 7.10–7.18 (m, 2H), 7.18–7.29 (m, 3H) ppm. The data are in accordance with
literature values.[14c]
General procedure A for BVO (GPA):[13] Cyclopentanone derivative 8 (1.0 equiv.)
was added to a mixture of mCPBA (6.0–10 equiv., 70% in H2O) and KHCO3 (2.0
equiv.) in DCM (6 L/mol). The mixture was heated to reflux for 5–24 h, subsequently
cooled to ambient temperature and washed with an aqueous NaHSO3 solution (10
L/mol, 10% w/w). The aqueous layer was extracted with DCM (2 x 10 L/mol). The
combined organic layers were first washed with a saturated aqueous K2CO3 solution
(10 L/mol) and then with brine (10 L/mol), dried (MgSO4), filtered and the solvent was
22
This article is protected by copyright. All rights reserved.

10.1002/chem.201605468
Chemistry - A European Journal
removed in vacuo. The crude product was purified by column chromatography to give
the lactone 9.
6,6-Dimethyltetrahydropyran-2-one (9a): According to the GPA, 2,2-
dimethylcyclopentanone (8a, 56 mg, 0.50 mmol), mCPBA (744 mg, 70% in H2O, 3.02
mmol) and KHCO3 (106 mg, 1.06 mmol) in DCM (3 mL) were heated to reflux for 5 h
to yield the title compound 9a (64 mg, 0.50 mmol, 100%) after chromatography (SiO2,
hexanes/EA 1:1, Rf = 0.28) as a colorless oil. 1H NMR (300 MHz, CDCl3): δ = 1.39 (s,
6H), 1.72–1.74 (m, 2H), 1.85–1.90 (m, 2H), 2.46 (t, J = 6.9 Hz, 2H) ppm. The data
were in accordance with literature values.[12]
6-Benzyl-6-methyltetrahydropyran-2-one (9b): According to the GPA, 2-benzyl-2-
methylcyclopentanone (8b, 91 mg, 0.48 mmol), mCPBA (1.18 g, 70% in H2O, 4.77
mmol) and KHCO3 (95 mg, 1.00 mmol) in DCM (3 mL) were heated to reflux for 24 h
to yield the title compound 9b (48 mg, 0.23 mmol, 48%) after chromatography (SiO2,
hexanes/MTBE 1:1, Rf = 0.18) as a colorless oil. 1H NMR (300 MHz, CDCl3): δ = 1.36
(s, 3H), 1.58–1.88 (m, 4H), 2.23–2.34 (m, 1H), 2.42–2.52 (m, 1H), 2.88 (d, J = 13.7
Hz, 1H), 3.00 (d, J = 13.7 Hz, 1H), 7.19–7.38 (m, 5H) ppm. The data were in
accordance with literature values.[15]
Ethyl 2-methyl-6-oxotetrahydropyrane-2-carboxylate (9c): According to the GPA,
β-oxoester 8c (92 mg, 0.54 mmol), mCPBA (758 mg, 70% in H2O, 3.07 mmol) and
KHCO3 (104 mg, 1.04 mmol) in DCM (3 mL) were heated to reflux for 24 h to yield
the title compound 9c (101 mg, 0.54 mmol, 100%) after chromatography (SiO2,
hexanes/MTBE 1:2, Rf = 0.23) as a yellow oil. 1H NMR (300 MHz, CDCl3): δ = 1.26 (t,
23
This article is protected by copyright. All rights reserved.

10.1002/chem.201605468
Chemistry - A European Journal
J = 7.1 Hz, 3H), 1.58 (s, 3H), 1.63–1.88 (m, 3H), 2.17–2.21 (m, 1H), 2.36–2.62 (m,
2H), 4.15–4.26 (m, 2H) ppm. The data were in accordance with literature values.[16]
Ethyl 2-benzyl-6-oxotetrahydropyrane-2-carboxylate (9d): According to the GPA,
β-oxoester 8d (126 mg, 0.51 mmol), mCPBA (1.24 g, 70% in H2O, 5.01 mmol) and
KHCO3 (106 mg, 1.06 mmol) in DCM (3 mL) were heated to reflux for 24 h to yield
the title compound 9d (90 mg, 0.34 mmol, 67%) after chromatography (SiO2,
hexanes/MTBE 1:1, Rf = 0.23) as a colorless oil. 1H NMR (300 MHz, CDCl3): δ = 1.21
(t, J = 7.1 Hz, 3H), 1.61–1.83 (m, 3H), 2.10–2.37 (m, 2H), 2.50–2.58 (m, 1H), 3.13 (d,
J = 14.1 Hz, 1H), 3.24 (d, J = 14.1 Hz, 1H), 4.18 (q, J = 7.1 Hz, 2H), 7.19–7.31 (m,
5H) ppm. 13C{1H} NMR (75 MHz, CDCl3): δ = 14.07 (CH3), 16.99 (CH2), 28.61 (CH2),
29.46 (CH2), 43.82 (CH2), 62.17 (CH2), 85.93 (C), 127.17 (CH), 128.21 (2 CH),
130.66 (2 CH), 134.21 (C), 169.94 (C), 171.61 (C) ppm. IR (ATR): nu(tilde) = 3031
(w), 2964 (w), 2928 (w), 1739 (vs), 1497 (w), 1455 (w), 1444 (w), 1371 (w), 1355 (w),
1329 (w), 1228 (s), 1188 (s), 1118 (m), 1083 (vs), 1052 (s), 1017 (m), 988 (w), 935
(m), 859 (w), 757 (m), 701 (s), 667 (w), 647 (w), 611 (w) cm–1. HRMS (ESI): calcd.
+
285.1097 (for C15H18NaO4 ); found 285.1102 [M + Na+]. C15H18O4 (262.31).
BVO of Ethyl 2-oxo-1-(2-oxo-2-phenylethyl)cyclopentane-1-carboxylate (5a):
According to the GPA, 1,4-diketone 5a (278 mg, 1.01 mmol), mCPBA (2.45 mg, 70%
in H2O, 10.0 mmol) and KHCO3 (198 mg, 1.98 mmol) in DCM (6 mL) were heated to
reflux for 24 h. The crude mixture was purified by column chromatography (SiO2,
hexanes/MTBE 1:1 → 1:2) to yield lactone 3a [37 mg, 0.13 mmol, 13%, Rf = 0.28
(hexanes/MTBE 1:2)] in the first fraction as a colorless oil. Secondly, lactone 10 [17
mg, 60 µmol, 6%, Rf = 0.16 (hexanes/MTBE 1:2)] was eluted as a colorless oil. As
the third fraction lactone 4 [66 mg, 0.23 mmol, 23%, Rf = 0.13 (hexanes/MTBE 1:2)]
24
This article is protected by copyright. All rights reserved.

10.1002/chem.201605468
Chemistry - A European Journal
was isolated as a colorless solid (mp. 107°C). All spectroscopic data for compound
3a are in accordance with the literature.[9]
1
Ethyl 6-oxo-2-(2-oxo-2-phenylethyl)tetrahydropyran-2-carboxylate (4): H NMR
(500 MHz, CDCl3): δ = 1.29 (t, J = 7.1 Hz, 3H), 1.76–1.93 (m, 2H), 2.02 (td, J = 13.6
Hz, J = 4.5 Hz, 1H), 2.21–2.26 (m, 1H), 2.49 (ddd, J = 18.5 Hz, J = 10.6 Hz, J = 8.0
Hz, 1H), 2.66–2.72 (m, 1H), 3.61 (d, J = 17.2 Hz, 1H), 3.67 (d, J = 17.2 Hz, 1H),
4.26–4.35 (m, 2H), 7.43–7.46 (m, 2H), 7.54–7.57 (m, 1H), 7.90–7.92 (m, 2H) ppm.
13C{1H} NMR (125 MHz, CDCl3): δ = 14.03 (CH3), 16.72 (CH2), 28.55 (CH2), 30.26
(CH2), 46.76 (CH2), 62.26 (CH2), 82.64 (C), 128.07 (2 CH), 128.60 (2 CH), 133.45
(CH), 136.38 (C), 169.39 (C), 171.36 (C), 194.14 (C) ppm. IR (ATR): nu(tilde) = 3002
(w), 2975 (w), 2923 (w), 1743 (s), 1732 (s), 1688 (s), 1596 (w), 1580 (w), 1466 (w),
1451 (w), 1440 (w), 1359 (m), 1312 (m), 1232 (m), 1196 (s), 1163 (m), 1133 (m),
1081 (s), 1054 (s), 1024 (m), 1006 (m), 978 (w), 936 (m), 859 (w), 764 (m), 747 (s),
695 (s), 687 (m), 660 (w), 623 (m) cm–1. HRMS (ESI): calcd. 313.1046 (for
+
C16H18NaO5 ); found 313.1049 [M + Na+]. C16H18O5 (290.32).
1
Ethyl 6-oxo-2-[(phenoxycarbonyl)methyl]tetrahydropyran-2-carboxylate (10): H
NMR (500 MHz, CDCl3): δ = 1.30 (t, J = 7.1 Hz, 3H), 1.74–1.83 (m, 1H), 1.89–1.95
(m, 1H), 2.04 (td, J = 13.7 Hz, J = 4.4 Hz, 1H), 2.24–2.29 (m, 1H), 2.51 (ddd, J = 18.5
Hz, J = 10.8 Hz, J = 7.9 Hz, 1H), 2.61–2.72 (m, 1H), 3.08 (d, J = 16.1 Hz, 1H), 3.37
(d, J = 16.1 Hz, 1H), 4.24–4.35 (m, 2H), 7.09–7.12 (m, 2H), 7.20–7.25 (m, 1H), 7.34–
7.39 (m, 2H) ppm. 13C{1H} NMR (125 MHz, CDCl3): δ = 14.11 (CH3), 16.97 (CH2),
28.60 (CH2), 30.23 (CH2), 43.26 (CH2), 62.64 (CH2), 82.99 (C), 121.48 (2 CH),
126.08 (CH), 129.44 (2 CH), 150.26 (C), 166.84 (C), 168.92 (C), 170.74 (C) ppm. IR
(ATR): nu(tilde) = 2955 (w), 2928 (w), 2856 (w), 2364 (w), 2342 (w), 1753 (vs), 1593
25
This article is protected by copyright. All rights reserved.

10.1002/chem.201605468
Chemistry - A European Journal
(w), 1492 (w), 1459 (w), 1374 (w), 1311 (w), 1232 (m), 1196 (s), 1165 (m), 1150 (m),
1109 (w), 1085 (m), 1055 (w), 1021 (w), 938 (w), 751 (w), 691 (w) cm–1. HRMS (ESI):
+
calcd. 329.0996 (for C16H18NaO6 ); found 329.0993 [M + Na+]. C16H18O6 (306.31).
1-(4-Methylphenyl)vinyl acetate (2c): The product was prepared by the method of
Hagemeyer and Hull.[28] A mixture of 4-methylacetophenone (5.0 g, 37 mmol),
isopropenyl acetate (7.5 g, 75 mmol) and four drops of conc. H2SO4 (ca. 80 mg) was
heated to reflux for 18 h. Subsequently, the mixture was directly submitted to column
chromatography (SiO2, hexanes/MTBE 2:1, Rf = 0.47) to yield the title compound 2c
1
(2.65 g, 15.0 mmol, 40%) after chromatography as a light yellow oil. H NMR (300
MHz, CDCl3): δ = 2.28 (s, 3H), 2.35 (s, 3H), 4.97 (d, J = 1.8 Hz, 1H), 5.43 (d, J = 1.8
Hz, 1H), 7.13–7.19 (m, 2H), 7.34–7.39 (m, 2H) ppm. The data were in accordance
with literature values.[29]
General procedure B (GPB):[27] Under exclusion of air and moisture (nitrogen
atmosphere), a solution of nBuLi (1.05 equiv., 2.5 mol/L in hexane) was added
dropwise over 10 min to a solution of HMDS (1.05 equiv.) in abs. hexane (0.5 L/mol)
and the mixture was heated to reflux for 1 h. After adding abs. THF (1.5 L/mol) the
solution was cooled to –78°C. At this temperature, a solution of the acetophenone
derivative (1.0 equiv.) in abs. THF (1 L/mol) was added dropwise within 30 min and
the mixture further stirred for 30 min. Subsequently, a solution of Ac2O (1.05 equiv.)
in abs. THF (1 L/mol) was added dropwise within 15 min and the resulting mixture
was stirred for further 1.5 h at this temperature. Subsequently, the mixture was stirred
overnight at ambient temperature and then was acidified with hydrochloric acid (3
L/mol, 1 mol/L) and extracted with Et2O (3 x 1.5 L/mol). The combined organic layers
26
This article is protected by copyright. All rights reserved.

10.1002/chem.201605468
Chemistry - A European Journal
were dried (MgSO4) and the solvent was removed in vacuo. The residue was purified
by column chromatography to yield the enolester 2.
1-(4-Fluorophenyl)vinyl acetate (2d): According to the GPB, reaction of nBuLi (6.1
mL, 2.5 mol/L in hexane, 15 mmol), HMDS (3.1 mL, 15 mmol), 4-fluoroacetophenone
(2.61 g, 14.5 mmol) and Ac2O (1.55 g, 15.2 mmol) gave the title compound 2d (2.00
g, 11.1 mmol, 77%) after chromatography (SiO2, hexanes/MTBE 5:1, Rf = 0.32) as a
1
yellow liquid. H NMR (300 MHz, CDCl3): δ = 2.28 (s, 3H), 5.01 (d, J = 2.2 Hz, 1H),
5.40 (d, J = 2.2 Hz, 1H), 7.01–7.07 (m, 2H), 7.42–7.47 (m, 2H) ppm. The data were
in accordance with literature values.[27]
1-(4-Chlorophenyl)vinyl acetate (2e): According to the GPB, reaction of nBuLi (6.1
mL, 2.5 mol/L in hexane, 15 mmol), HMDS (3.1 mL, 15 mmol), 4-chloroacetophenone
(2.85 g, 14.5 mmol) and Ac2O (1.55 g, 15.2 mmol) gave the title compound 2e (2.21
g, 11.2 mmol, 78%) after chromatography (SiO2, hexanes/MTBE 5:1, Rf = 0.29) as a
1
yellow liquid. H NMR (300 MHz, CDCl3): δ = 2.28 (s, 3H), 5.05 (d, J = 2.3 Hz, 1H),
5.46 (d, J = 2.3 Hz, 1H), 7.30–7.33 (m, 2H), 7.38–7.41 (m, 2H) ppm. The data were
in accordance with literature values.[27]
1-(4-Cyanophenyl)vinyl acetate (2g): According to the GPB, reaction of nBuLi (7.0
mL, 2.5 mol/L in hexane, 15 mmol), HMDS (3.1 mL, 15 mmol), 4-cyanoacetophenone
(2.71 g, 14.5 mmol) and Ac2O (1.55 g, 15.2 mmol) gave the title compound 2g (2.12
g, 11.3 mmol, 78%) after chromatography (SiO2, hexanes/MTBE 5:1, Rf = 0.33) as a
1
yellow liquid. H NMR (300 MHz, CDCl3): δ = 2.29 (s, 3H), 5.21 (d, J = 2.5 Hz, 1H),
5.60 (d, J = 2.5 Hz, 1H), 7.54–7.57 (m, 2H), 7.63–7.66 (m, 2H) ppm. The data were
in accordance with literature values.[27]
27
This article is protected by copyright. All rights reserved.

10.1002/chem.201605468
Chemistry - A European Journal
Ethyl
3-[2-oxo-2-(4-methylphenyl)ethyl]tetrahydropyran-2-one-3-carboxylate
(3c): CeCl3 · 7 H2O (5 mg, 13 µmol) was added to a mixture of β-oxoester 1 (200 mg,
1.28 mmol) and enol ester 2c (338 mg, 1.92 mmol, 1.5 equiv.) in TFE (1 mL), and the
reaction mixture was stirred at 30°C for 24 h under an atmosphere of air.
Subsequently, all volatile materials were removed in vacuo and the residue was
purified by column chromatography (SiO2, hexanes/MTBE 2:1 → 1:1) to yield 1,4-
diketone 5c [103 mg, 0.36 mmol, 28%, Rf = 0.47 (hexanes/MTBE 1:1)] in the first
fraction as a colorless oil. Secondly, the α-hydroxylated compound 11 [4 mg, 20
µmol, 2%, Rf = 0.30 (hexanes/MTBE 1:1)]) was obtained as a colorless oil. As the
third fraction the title compound 3c [229 mg, 0.75 mmol, 59%, Rf = 0.13
1
(hexanes/MTBE 1:1)] was eluted as a colorless oil. Lactone 3c: H NMR (300 MHz,
CDCl3): δ = 1.26 (t, J = 7.1 Hz, 3H), 1.78–1.86 (m, 1H), 2.03–2.19 (m, 3H), 2.35 (s,
3H), 3.56 (d, J = 18.6 Hz, 1H), 3.97 (d, J = 18.6 Hz, 1H), 4.12–4.29 (m, 2H), 4.47–
4.55 (m, 1H), 4.60–4.71 (m, 1H), 7.17–7.23 (m, 2H), 7.78–7.85 (m, 2H) ppm. 13C{1H}
NMR (75 MHz, CDCl3): δ = 13.82 (CH3), 20.87 (CH2), 21.46 (CH3), 30.13 (CH2),
45.46 (CH2), 50.84 (C), 62.04 (CH2), 69.83 (CH2), 127.96 (2 CH), 129.12 (2 CH),
133.42 (C), 144.37 (C), 169.80 (C), 171.33 (C), 196.16 (C) ppm. IR (ATR): nu(tilde) =
2978 (w), 2936 (w), 1730 (s), 1718 (vs), 1659 (s), 1607 (m), 1453 (w), 1405 (m),
1351 (m), 1281 (m), 1250 (s), 1224 (s), 1197 (vs), 1180 (s), 1167 (vs), 1117 (s), 1068
(m), 1022 (m), 986 (m), 858 (w), 809 (m), 639 (m), 572 (m) cm–1. HRMS (ESI): calcd.
+
327.1203 (for C17H20NaO5 ); found 327.1207 [M + Na+]. C17H20O5 (304.34).
1
Ethyl 2-[2-oxo-2-(4-methylphenyl)ethyl]cyclopentanone-2-carboxylate (5c): H
NMR (300 MHz, CDCl3): δ = 1.22 (t, J = 7.1 Hz, 3H), 2.01–2.23 (m, 3H), 2.39 (s, 3H),
2.44–2.68 (m, 3H), 3.45 (d, J = 18.5 Hz, 1H), 3.81 (d, J = 18.5 Hz, 1H), 4.15 (q, J =
28
This article is protected by copyright. All rights reserved.

10.1002/chem.201605468
Chemistry - A European Journal
7.1 Hz, 2H), 7.21–7.26 (m, 2H), 7.80–7.85 (m, 2H) ppm. 13C{1H} NMR (75 MHz,
CDCl3): δ = 13.94 (CH3), 19.79 (CH2), 21.60 (CH3), 33.34 (CH2), 37.71 (CH2), 43.33
(CH2), 57.40 (C), 61.56 (CH2), 128.07 (2 CH), 129.23 (2 CH), 133.79 (C), 144.25 (C),
170.70 (C), 196.26 (C), 215.12 (C) ppm. IR (ATR): nu(tilde) = 2979 (w), 2923 (w),
1750 (s), 1721 (vs), 1679 (s), 1607 (m), 1464 (w), 1404 (m), 1352 (w), 1260 (w),
1225 (s), 1209 (m), 1181 (s), 1151 (s), 1103 (m), 1019 (w), 984 (w), 924 (w), 860 (w),
+
811 (m), 561 (m) cm–1. HRMS (ESI): calcd. 311.1254 (for C17H20NaO4 ); found
311.1249 [M + Na+]. C17H20O4 (288.34).
1
Ethyl 2-hydroxycyclopentanon-2-carboxylate (11): H NMR (500 MHz, CDCl3): δ
= 1.29 (t, J = 7.1 Hz, 3H), 2.08–2.14 (m, 3H), 2.43–2.53 (m, 3H), 3.63 (s, 1H), 4.23–
4.28 (m, 2H) ppm. The data were in accordance with literature values.[9]
Ethyl
3-[2-oxo-2-(4-fluorophenyl)ethyl]tetrahydropyran-2-one-3-carboxylate
(3d): According to the procedure for 3c given above, reaction of CeCl3 · 7 H2O (5 mg,
13 µmol), β-oxoester 1 (202 mg, 1.28 mmol) and enol ester 2d (347 mg, 1.92 mmol,
1.5 equiv.) in TFE (1 mL) gave a crude mixture, which was submitted to column
chromatography (SiO2, hexanes/MTBE 1:1) to yield 1,4-diketone 5d (72 mg, 0.25
mmol, 19%, Rf = 0.43) in the first fraction as a colorless oil. Secondly, the α-
hydroxylated compound 11 (5 mg, 23 µmol, 2%, Rf = 0.29) was obtained as a
colorless oil. As the third fraction the title compound 3d (212 mg, 0.69 mmol, 54%, Rf
= 0.20) was eluted as a colorless solid (mp. 108°C). Lactone 3d: 1H NMR (300 MHz,
CDCl3): δ = 1.28 (td, J = 7.1 Hz, J = 1.2 Hz, 3H), 1.78–1.91 (m, 1H), 2.05–2.25 (m,
3H), 3.55 (d, J = 18.5 Hz, 1H), 3.98 (d, J = 18.6 Hz, 1H), 4.13–4.32 (m, 2H), 4.49–
4.59 (m, 1H), 4.62–4.72 (m, 1H), 7.06–7.15 (m, 2H), 7.91–8.01 (m, 2H) ppm. 13C{1H}
NMR (75 MHz, CDCl3): δ = 13.91 (CH3), 20.94 (CH2), 30.21 (CH2), 45.54 (CH2),
29
This article is protected by copyright. All rights reserved.