Selective and shape-selective Baeyer-Villiger oxidations of aromatic aldehydes and cyclic ketones with Sn-beta zeolites and H2O2

Article abstract of DOI:10.1002/1521-3765(20021018)8:20<4708::AID-CHEM4708>3.0.CO;2-U

Sn-Beta is used as a heterogeneous catalyst for the Baeyer-Villiger reaction with hydrogen peroxide. Cyclic ketones are transformed into the corresponding lactones, while unsaturated ketones are oxidized to the corresponding unsaturated lactones with very high chemoselectivity. The catalyst is also selective for the oxidation of aromatic aldehydes with H₂O₂, producing the formate ester or the corresponding hydrolyzed product, that is the alcohol. Shape-selective oxidations are observed for isomeric reactants with different molecular shapes. The catalytic Sn sites have been characterized by ¹¹⁹Sn MAS-NMR spectroscopy, and tetrahedral incorporation into the zeolite framework has been demonstrated. In situ IR spectroscopy and ¹⁸O labeling experiments have shown that the oxidation mechanism involves an intermediate of the Criegee type.

Full text of DOI:10.1002/1521-3765(20021018)8:20<4708::AID-CHEM4708>3.0.CO;2-U

FULL PAPER
Selective and Shape-Selective Baeyer± Villiger Oxidations of Aromatic
Aldehydes and Cyclic Ketones with Sn-Beta Zeolites and H O
2
2
[
a]
[a]
[a]
[a]
[b]
Michael Renz, Teresa Blasco, Avelino Corma,* Vicente Forn e¬ s, Robert Jensen,
[
b]
and Laszlo Nemeth
Dedicated to Professor Waldemar Adam on the occasion of his 65th birthday
Abstract: Sn-Beta is used as a hetero-
geneous catalyst for the Baeyer± Villig-
er reaction with hydrogen peroxide.
Cyclic ketones are transformed into the
corresponding lactones, while unsaturat-
ed ketones are oxidized to the corre-
sponding unsaturated lactones with very
high chemoselectivity. The catalyst is
also selective for the oxidation of aro-
matic aldehydes with H O , producing
the formate ester or the corresponding
hydrolyzed product, that is the alcohol.
Shape-selective oxidations are observed
for isomeric reactants with different
molecular shapes. The catalytic Sn sites
1
19
have been characterized by Sn MAS-
NMR spectroscopy, and tetrahedral in-
corporation into the zeolite framework
has been demonstrated. In situ IR spec-
1
8
troscopy and O labeling experiments
have shown that the oxidation mecha-
nism involves an intermediate of the
Criegee type.
Keywords: Baeyer± Villiger
reaction ¥ heterogeneous catalysis ¥
hydrogen peroxide
¥ oxidation ¥
tin ¥ zeolites
2
2
Introduction
ficial organic reactant. The issues encountered with peracids
make it desirable to develop catalysts that allow replacement
of the peracids by hydrogen peroxide. Some success has been
achieved with homogeneous and heterogeneous catalysts
The Baeyer± Villiger oxidation, which involves the trans-
formation of ketones into esters or lactones [Eq. (1)], is one of
the most important reactions in organic chemistry.
[1±±]
containing active centers based on As, Se, Zr, Mo, Re, or Pt,
or with soluble and solid acids.[
7, 12±14]
All of these catalysts
have at least one drawback, such as low activity or selectivity,
tedious recovery and regeneration, or high costs for starting
materials and/or for safety precautions. Exceptions in this
regard among the metal catalysts are soluble Pt complexes, for
which turnover numbers (TONs) of 4± and 60 have been
reported for cyclohexanone and 2-methylcyclohexanone,
O
O
R2
R1
R2
R1
O
The most common and suitable oxidants for this reaction
have been peracids, which entail by-products, have a low
[4]
[5]
[7]
active oxygen content, are potentially explosive, may
require strong acidity because of low reactivity.[ An addi-
tional disadvantage of peracids is their lack of selectivity for
oxidation reactions when other functional groups, such as
double bonds, are present in the reactant molecule.[ To avoid
isolation and storage of dangerous peracids, the Mukaiyama
respectively, with excellent selectivities to the lactones.
1]
Recently, advances have been reported in the application of
peroxyseleninic acids.[
1±]
With catalytic amounts of the
catalyst precursor, cyclopentanone has been converted com-
pletely into d-valerolactone, with a selectivity of 92% in
1,1,1,±,±,±-hexafluoro-2-propanol solvent.[ Among the solid
acids, sulfonic acid groups fixed to a polymer have been used
2]
1±]
method can be employed to generate peracids in situ from
corresponding aldehydes,[
6±11]
but this process needs a sacri-
in the presence of ±0% hydrogen peroxide for the Baeyer±
Villiger reaction.[
12, 15]
The best selectivities achieved with
[
a] Prof. Dr. A. Corma, Dr. M. Renz, Dr. T. Blasco, Prof. Dr. V. Forn e¬ s
such catalysts were ꢀ80% with cyclopentanone as sub-
Instituto de TecnologÌa QuÌmica, UPV-CSIC
Universidad Polit e¬ cnica de Valencia
Avenida de los Naranjos s/n, 46022 Valencia (Spain)
Fax : (‡±4)96-±877809
strate.[ Slightly lower selectivities (ꢀ70%) were achieved
12]
[12, 16]
with acid zeolites and H O .
2
2
Taking into account the limited success obtained for the
Baeyer± Villiger reaction with catalysts that activate H O ,
E-mail: acorma@itq.upv.es
2
2
[
b] Dr. R. Jensen, Dr. L. Nemeth
UOP LLC, 25 East Algonquin Road
Des Plaines, IL 60017-5017 (USA)
we have very recently reported a different catalytic strategy
that involves activation of the carbonyl group instead of the
4708
¹ 2002 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
0947-65±9/02/0820-4708 $ 20.00+.50/0
Chem. Eur. J. 2002, 8, No. 20

4
708±4717
hydrogen peroxide.[17] Among several Lewis acids, Sn within
the framework of a zeolite (Beta) is able to polarize the
carbonyl group, making it more reactive towards H O attack.
interesting possibility that by modifying this system the
selectivity to the uncommon isomer obtained when using
peracids could be increased.
2
2
If Sn peroxide is formed, it is not active for oxidation
reactions[ and cannot affect the selectivity negatively when
other functional groups are present in the molecule. In the
detailed study of the Baeyer± Villiger oxidation with Sn-Beta
and H O reported here, we have shown that with this catalyst
To study shape selectivity effects due to the geometrical
structure and site accessibility in the zeolite, a tert-butyl group
was introduced into the cyclohexanone ring instead of the
methyl group described above. Whereas 4-tert-butylcyclohex-
anone (1e) reacts readily (in dioxane 91% was converted
after ± h; Table 1, entry 9), no conversion was observed with
2-tert-butylcyclohexanone (1d, entry 7). The very bulky
ketone adamantanone (1h) has also been oxidized to lactone
with excellent conversion and selectivity [Table 1, entries 14
and 15; Eq. (±)].
18]
2
2
it is also possible to carry out shape-selective Baeyer± Villiger
oxidations; we have included the formation of lactones by
oxidation of ketones as well as the oxidation of aldehydes to
yield the corresponding ester or its hydrolysis products.
Additional shape-selectivity effects can occur due to the
characteristics of the zeolite support when reacting substrates
have different shapes. The study also involved the character-
ization of the coordination sphere of the active Sn sites by
O
O
H2O2, Sn-beta
MTBE, 56 °C
O
119
Sn MAS-NMR spectroscopy. Mechanistic studies were
carried out using in situ IR spectroscopy to elucidate the
interaction of the reactant and of the product with the Sn-
active sites, and DR-UV spectroscopy to investigate a
potential interaction of hydrogen peroxide with the Sn-active
sites and with O-labeled cyclohexanone to establish the
reaction intermediate.
1h
2h
The lack of conversion in the case of 1d may have several
causes. It has been shown previously that, in general, ketone
1d can be converted by peracids and the tert-butyl group in
the 2-position does not prevent the Baeyer± Villiger oxida-
tion.[ Consequently, the cause must be related to the
interference of the geometries and the sizes of both the
substrate molecules and the zeolite pores. The molecular sizes
of 1d, 1e and 1h are similar. They have diameters between 6.±
and 7.1 ä (taking the longest distance as the molecular axis)
and would fit into the Beta pore of 6.6 ä Â 7.7 ä. However,
for adamantanone and 4-tert-butylcyclohexanone (1e) the
carbonyl oxygen is part of the molecular axis and therefore far
from the apolar zeolite wall. In contrast, for 2-tert-butylcy-
clohexanone (1d) the carbonyl oxygen points away from the
axis and an unfavorable interaction with the zeolite walls may
render diffusion of the molecule within the zeolite pores more
difficult. A second explanation for the lack of conversion of
18
19]
Results and Discussion
Catalytic activity
Oxidation of cyclic ketones and shape selectivity effects: T he
Baeyer± Villiger reaction of cyclohexanone (1a) to e-capro-
lactone (2a) was carried out with H O as the oxidant, Sn-
2
2
Beta-2 as the catalyst, and methyl tert-butyl ether (MTBE) as
the solvent [Eq. (2)].
1
d may be that the substituent prevents the approach of the
carbonyl group to the metal center, which is embedded in the
zeolite wall and shielded by the relatively large oxygen atoms
of the zeolite framework.[
20]
Cyclopentanone (1 f) and 2-methylcyclopentanone (1g)
could also be oxidized to the corresponding lactones with
H O using Sn-Beta as catalyst (TONs are nearly 50 and 92,
respectively; Table 1, entries 10 ± 1±). The oxidant efficiencies
were always very high and in no case was decomposition of
the hydrogen peroxide observed. Ketones that are less
reactive with respect to the Baeyer ± Villiger reaction, such
as aromatic (acetophenone) and acyclic ketones (±-octanone),
are not converted.
ATON of 68 was achieved, the lactone being the only product
observed (Table 1, entry 1). Increasing the reaction temper-
ature from 56 to 808C and use of dioxane as solvent improved
the conversion to 52% and the TON to 88 (Table 1, entry 2).
We have applied Sn-Beta to another Baeyer± Villiger reac-
tion in which a methyl substituent was introduced in position 2
or 4 of the cyclohexanone ring. A similar activity was found
for the two isomers (Table 1, entries ± ± 6), which indicate that
both reactants can diffuse readily into the pores of the zeolite
and approach to within reaction distances of the Sn frame-
work. With 2-methylcyclohexanone two regioisomeric lac-
tones were obtained in an 85:15 ratio favoring the insertion of
the oxygen between the carbonyl group and the more highly
substituted carbon atom neighbor. When conventional per-
acid oxidants were employed, the minor isomer was observed
only in traces (<5%).[ Although this behavior of the
Sn-Beta system can be regarded as less desirable, it is an
2
2
Significant advantages of the new oxidation system report-
ed in this work are the heterogeneous nature of the catalyst
and the very high selectivity to lactones obtained using H O2
2
as oxidant. However, from a practical point of view the
heterogeneous catalyst should be recyclable. In the case of Sn-
Beta, no deactivation of the catalyst was observed after
recycling four times without intermediate regeneration steps.
This allowed us to obtain a TON higher than ±00 without
regeneration. Furthermore, Sn-Beta-2 was calcined at 5008C
1]
Chem. Eur. J. 2002, 8, No. 20
¹ 2002 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
0947-65±9/02/0820-4709 $ 20.00+.50/0
4709

A. Corma et al.
FULL PAPER
Table 1. Conversions, TONs, and selectivities in the Baeyer± Villiger reaction of various substrates, catalyzed by
Sn-Beta-2.
Eq. (4)]. With 4a a mixture of
ester and phenol was observed
(Table 2, entries 1 and 2), indi-
cating that the aromatic ring is
migrating instead of the hydro-
gen atom. Moreover, the kinet-
ic results (Figure 1) show that
the ester is the primary product
formed, and it readily reacts to
give the phenolic hydrolysis
product.
The relative yields of ester
and phenol can be modified not
only by changing the level of
conversion but also by using
different solvents. The conver-
sion and selectivity to phenol
are higher with toluene than
with dioxane as solvent (Ta-
ble 2).
Entry
Substrate
R
Reaction conditions
Conversion
[%]
TON[a]
Selectivity [%]
1
R2
H
n
to lactone
2 2
for H O
1
2
±
4
5
6
7
8
9
1a
1b
1c
1
H
MTBE, 568C, 7 h
dioxane, 808C, ± h
MTBE, 568C, 7 h
dioxane, 808C, 4 h
MTBE, 568C, 7 h
dioxane, 808C, 4 h
MTBE, 568C, 7 h
MTBE, 568C, 7 h
dioxane, 808C, ± h
MTBE, 568C, 7 h
dioxane, 808C, ± h
MTBE, 568C, 7 h
dioxane, 808C, 8 h
MTBE, 568C, 7 h
MTBE, 568C, 7 h
42
52
47
80
46
70
0
68
88
86
105
81
104
0
87
1±7
48
47
47
> 98
> 98
> 98[
96[
> 98
97
> 99
> 99
> 99
> 99
> 99
> 99
±
b]
b]
1
1
Me
H
H
Me
1d
1e
1
1
tBu
H
H
tBu
±
54
91
±2
±1
±1
62
85
95[
> 98
> 98
> 98
> 98
> 98
97
> 99
> 99
> 99
> 99
> 99
> 99
> 99
> 99
10
11
12
1±
14
15
1 f
1g
0
0
H
H
H
Me
92
147
119
1h
1h
adamantanone
adamantanone
> 98
> 98
c]
[
a] TON, turnover number ˆ catalytic cycles per metal center. [b] An 85:15 mixture of the regioisomeric lactones
e-methyl-e-caprolactone and a-methyl-e-caprolactone was observed. [c] 1.1 equiv hydrogen peroxide and 60 mg
of Sn-Beta were employed; reaction time 8 h.
With the second aldehyde 4b,
in air after the fourth reaction cycle and re-used. The decrease
in TON of only 5% can be attributed to the handling of the
sample. These results support our conclusion that the Sn-Beta
catalyst has very good stability and regenerability which make
it suitable for use in a fixed bed or in a stirred continuous
reactor.
migration of both the aromatic ring and the hydrogen atom
occurs. As observed with 4a, conversion is also higher in
toluene than in dioxane. However, the solvent has only a
minor influence on selectivity to phenol for 4b.
In summary, phenols with electron-donating substituents
can be obtained from the corresponding aldehydes with
hydrogen peroxide as oxidant and Sn-Beta as catalyst.
Oxidation of aromatic aldehydes: The Baeyer± Villiger
oxidation of aromatic aldehydes can yield two different
products. When the migrating group is the hydrogen atom the
corresponding carboxylic acid is formed, but when the
aromatic ring migrates the aldehyde is transformed into a
formate ester which often becomes hydrolyzed to yield the
corresponding phenol (Scheme 1).
Figure 1. Yield of ester 5a and alcohol 6a versus total conversion of
aldehyde 4a in dioxane and toluene as solvents. ^ yield 5a in dioxane;
^
yield 5a in toluene;
&
6a in dioxane;
&
6a in tolene.
Table 2. Conversions, TONs, and product distributions for the Baeyer±
Villiger oxidation of the aromatic aldehydes 4a and 4b, catalyzed by Sn-
Beta-2.
Scheme 1. Formation of the ester product (migration of the aromatic ring,
left-hand side) and the acid product (migration of the hydrogen atom, right-
hand side) in the Baeyer± Villiger oxidation of aromatic aldehydes.
O
H
O
HO
O
H
O
OH
+
+
The selectivity to the different products depends on the
substituents on the aromatic ring. Electron-donating groups,
such as alkoxy, direct the reaction towards ester/phenol
formation whereas electron-withdrawing groups favor the
formation of the acid. In organic synthesis a phenol is often
the desired product, because substituted phenols are impor-
tant substrates and active compounds in the pharmaceutical
industry.[
a: R = OMe
b: R = Me
R
R
R
R
4a,b
5a,b
6a,b
7a,b
Conversion [%] TON[a] Product distribution [%]
Entry Solvent R
5
6
7
other
1
2
±
4
dioxane OMe
toluene OMe
dioxane Me
toluene Me
49
58
26
47
270
±29
166
±00
6±
1±
29
2±
±5
86
44
56
< 2
< 2
2±
2
< 2
4
21]
20
< 2
p-Anisaldehyde (4a) and p-tolualdehyde (4b) could be
oxidized by the new oxidation system Sn-Beta/H O [Table 2;
[a] TON, turnover number ˆ catalytic cycles per metal center.
2
2
4710
¹ 2002 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim 0947-65±9/02/0820-4710 $ 20.00+.50/0 Chem. Eur. J. 2002, 8, No. 20

Baeyer± Villiger Oxidation
4708±4717
Selection of the appropriate solvent can provide directly a
major yield of the desired hydrolysis product, so that an
additional saponification step such as the one required when
using organic peracids as oxidizing reagents[ may not be
necessary.
lactone, epoxide, and epoxylactone, and Ti-Beta produced the
epoxide and diols. Under the same reaction conditions MTO
lacked reactivity and selectivity, and the lactone was observed
only in minor yield.
22]
Characterization and properties of the catalytic sites: T o
1
19
Chemoselectivity of the catalyst for unsaturated ketones:
Chemoselectivity towards the lactone product with unsatu-
rated ketones has only been observed with enzymes as
catalysts,[ or with particular substrates in which the double
bond is unreactive and the Baeyer± Villiger reaction is
strongly favored. This is because Baeyer ± Villiger catalysts
such as peracids or methyltrioxorhenium (MTO) are also
good epoxidation agents. We have observed chemoselectivity
in the Baeyer± Villiger reaction with Sn-Beta catalysts and
H O [Table ±, Eq. (5)]. When reaction of bicyclohept-±-en-1-
ascertain the coordination of Sn, a Sn MAS NMR study was
undertaken with a Sn-Beta (1.6 wt% Sn) sample enriched
1
19
with Sn (Sn-Beta-2*). Sn-silicalite (MEL and MFI type) and
Sn-MCM-41 have been reported to give Sn chemical shifts d
these spectra were usually
recorded on hydrated samples, and the signals were attributed
either in octahedral
coordination. The 119Sn NMR spectrum of
dehydrated zeolite Sn-MFI gives a very broad band with the
2±]
119
[26±±0]
between À670 and À740;
IV
[26±±0]
[26, 28, ±0]
to framework Sn sites
or in
tetrahedral[
27, 29]
IV
maximum at d ˆ À688 attributed to octahedral Sn , and a
2
2
one was carried out with m-chloroperbenzoic acid (mCPBA),
a common practice, epoxides were always produced as by-
products. In contrast, Sn-Beta yielded exclusively the two
shoulder at d ˆ À4±9 assigned to framework tetrahedral sites
which shifts to d ˆ 579 when the coordination increases to five
upon hydration.[ However, the spectra were recorded under
static conditions, and therefore the signals were extremely
broad (greater than 500 ppm at half height) and the shoulder
was ascribed to ill-defined tetrahedral sites.[ For the zeolite
28]
regioisomeric lactones. Very similar results have been re-
ported with enzymes as catalysts.[
2±]
This substrate also
28]
furnishes a particular example of enhanced Baeyer± Villiger
reactivity of cyclobutanones in the presence of MTO.[
Although it is an excellent epoxidation agent, the MTO
catalyst provides only the lactone product.[
24]
119
Sn-Beta-2* after calcination and exposure to air, the Sn
MAS NMR spectrum of the hydrated sample (Figure 2b)
25]
shows two bands at d ꢀÀ 690 and À740, which are within the
IV
However, when another reactant, dihydrocarvone, in which
both double-bond oxidation and the Baeyer ± Villiger oxida-
tion can occur readily, is used to illustrate catalyst chemo-
selectivity, only Sn-Beta zeolite is fully selective for the
lactone [Table 4, Eq. (6)]. The peracid gives a mixture of
chemical shift range typical of Sn in hydrated Sn silicalites
and Sn-MCM-41 materials.[
26±±0]
It is important that the
spectrum of Sn-Beta zeolite is quite different from that of
bulk SnO or SnO impregnated on pure silica-Beta (Fig-
2
2
ure 2a) or on amorphous SiO . For these two materials, only a
2
sharp peak at d ˆ À605 has been observed, and it is believed
Table ±. Chemoselectivity of the Baeyer± Villiger reaction with various
catalysts and H O .
2 2
[±1]
to be characteristic of octahedrally coordinated Sn in SnO .
2
Since this peak was not observed in Sn-Beta-2* we can rule
out the presence of SnO2 in this sample, at least at the
detection level of NMR. Dramatic changes are observed when
zeolite Sn-Beta-2* is dehydrated (Figure 2c). The spectrum
consists of a sharp peak at d ˆ À445, which must originate
from tin in a lower coordination state. Under ambient
conditions, the original spectrum of the zeolite in the hydrated
state is recovered (Figure 2d).
O
O
O
O
O
O
+
+
O
O
Catalyst
Product distribution [%]
Sn-Beta
mCPBA
enzymes
100[a]
29
:
:
:
:
1
±4
1
:
:
:
:
1
±7
1
100[
100
a]
[
b]
[c]
MTO
1
1
[
a] As a 67:±± mixture of regioisomeric lactones; the major isomer is
shown. [b] From ref. [25]. [c] Isolated as an 87:1± regioisomeric mixture.
Table 4. Catalyst chemoselectivity in concurrent double-bond oxidation
and Baeyer± Villiger oxidation of dihydrocarvone.
O
O
O
O
+
+
O
O
O
O
Oxidant
Conversion [%]
Product distribution [%]
Sn-Beta/H
mCPBA
Ti-Beta/H O
2 2
2
O
2
68
85
46
9
100 :
0
:
:
:
:
0
18
0
11 : 71
0 : 79[
±0 : ±±
a]
[
b]
MTO/H
2
O
2
20
[
a] The ™missing∫ 21% comprised products from ring-opening of the
Figure 2. 119Sn MAS NMR spectra of: a) SnO
2
-Beta; b) hydrated Sn-Beta;
epoxide. [b] Reaction conditions as in ref. [25].
c) dehydrated Sn-Beta; d) rehydrated Sn-Beta.
Chem. Eur. J. 2002, 8, No. 20
¹ 2002 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
0947-65±9/02/0820-4711 $ 20.00+.50/0
4711

A. Corma et al.
FULL PAPER
Accordingly, we attribute the signal at d ˆ À445 of the
Reaction mechanism
IV
dehydrated Sn-Beta-2* zeolite to tetrahedral Sn at frame-
Spectroscopic study: To study the role of Sn in the Baeyer±
Villiger reaction, we began with two hypotheses: Sn interacts
with H O to form SnÀOOH, which attacks the carbonyl
work sites, which increases its coordination under ambient
conditions by bonding to extra water molecules. The presence
of two 1 Sn NMR peaks upon hydration could be explained
tentatively by the presence of two different coordination
2
2
19
group (although organoperoxo complexes of tin(iv) did not
exhibit any oxygen transfer reactions);[ alternatively, Sn
interacts with the carbonyl group, which becomes polarized
and consequently can be attacked by the nucleophilic H O .
18]
IV
states, penta- and hexacoordinated Sn . However, consider-
ing the position of the low-field peak, we think that hydration
induces changes in the local symmetry of the Sn environments
leading to two different octahedral sites.
2
2
To analyze the first hypothesis, we measured the DR-UV/
Vis spectrum of a sample of Sn-Beta treated with H O .
2
2
Activity tests with cyclohexanone and Sn-Beta with differ-
ent tin contents showed that the amount of tin that can be
incorporated into network positions is limited. When the Sn
content was increased from 0.8 to 1.6 wt% (Table 5; compare
entries 1 and 2), conversion increased. When a larger amount
of Sn was present (2.±6 wt%; entry ±), both conversion and
TON decreased. A similar relationship between activity and
the amount of metal incorporated into the zeolite framework
has been observed previously in the case of Ti-silicalite (TS-
Contrary to the case of a Ti-Beta sample, where the
interaction of Ti with H O results in the formation of a band
2
2
at 425 nm associated with a TiÀOOH species (Figure ±), for
Sn-Beta no band in the DR-UV/Vis region was observed. The
spectrum was exactly the same for silica-Beta as for pure
treated with H O (Figure ±).
2
2
[
±2]
1),
where a maximum in activity for alkane or alkene
oxidation was found for samples containing ꢀ1.5% Ti. For
higher Ti contents, some of the Ti forms Ti ± O ± Ti pairs and/
or is in extra-framework positions, with a deleterious effect on
conversion. By analogy, it seems that most of the Sn should be
isolated in framework positions and it should be catalytically
active for Sn-Beta samples with ꢁ2 wt% Sn. For higher Sn
contents, some of Sn must be in extra-framework positions,
that is, in SnO crystals. Indeed, we have observed that SnO
2
2
highly dispersed on silica gel or on pure silica-Beta zeolite
gives very little, if any, catalytic activity (Table 5, entries 6 and
7
), indicating that the presence of framework Sn is necessary
Figure ±. DR-UV/Vis spectra of zeolite samples treated with hydrogen
for catalyzing the Baeyer± Villiger reaction with Sn-Beta
zeolite.
Furthermore, we have confirmed experimentally that the
Sn-zeolite active site remains intact during the reaction and
that oxidation does not occur by homogeneous catalysis
through leached Sn. Therefore, the reaction was conducted to
peroxide. (––) Sn-Beta‡H
2
O;
(
* * * *
)
Ti-Beta‡H
2 2
O ; (- - - -) Si-Be-
ta‡H O .
2
2
Thus, we conclude that even if H O can be coordinated to
2
2
tin in zeolite framework positions, such coordination proceeds
without deprotonation and therefore without activation of the
H O . This conclusion is consistent with the null activity of Sn-
±0% conversion, the catalyst was filtered out at reaction
temperature, and then the reaction was allowed to proceed in
the absence of solid catalyst. Under these experimental
conditions the reaction did not proceed, and further con-
version was not observed after four hours of reaction time.
Moreover, in another experiment in which SnCl was used as
catalyst the TON was 5, instead of the value of 68 obtained
with the Sn-Beta-2 zeolite.
2
2
Beta for epoxidation of olefins.
The second hypothesis (the interaction of a tin Lewis acid
with a carbonyl compound) is well known and documented in
the literature. Lewis acid ± carbonyl adducts have been
observed and investigated by X-ray analysis. Average bond
lengths (dM ± O), valence angles (aC-O-M), and torsion
angles (qX-C-M-O) for boron, aluminum, and tin complexes
were found by a systematic survey of the Cambridge
4
Table 5. Conversions, TONs, and selectivities in the Baeyer± Villiger
reaction of cyclohexanone (1a), catalyzed by various metal species.
Structural Database on crystal structures of Lewis acid ± car-
[±±]
bonyl compound complexes.
In the late 1950s it was
Entry
Catalyst
Conver-
sion [%]
TON[a]
Lactone
selectivity [%]
discovered that NMR signals and IR bands of carbonyl
compounds were shifted upon complexation with a Lewis
acid.[ In order to explore our second hypothesis–that Sn
polarizes the carbonyl group, favoring the H O attack–we
±4]
1
2
±
4
5
6
7
Sn-Beta-1 (0.8 wt% Sn)
Sn-Beta-2 (1.6 wt% Sn)
Sn-Beta-± (2.± wt% Sn)
Ti-Beta (1.2 wt% Ti)
Al-Beta (0.5 wt% Al)
26
42
27
4
19
0
87
68
±0
4
10
0
> 98
> 98
> 98
> 98
> 98
2
2
have made use of these two spectroscopic techniques.
Unfortunately, the introduction of molecules into the zeolite
structure broadens the NMR signals significantly, and small
shifts are difficult to identify. Another limitation was the high
dilution of the active sites combined with the corresponding
low sensitivity of the analytical method.
[
b]
SiO
2 2
¥ SnO
[
b]
Beta ¥ SnO
2
0
0
[
a] TON, turnover numberˆ catalytic cycles per metal center. [b] The same
amount of Sn was employed as for the Sn-Beta-2 sample.
4712
¹ 2002 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
0947-65±9/02/0820-4712 $ 20.00+.50/0
Chem. Eur. J. 2002, 8, No. 20

Baeyer± Villiger Oxidation
4708±4717
However, we determined that IR spectroscopy is an
appropriate technique to study the interaction of a ketone
detected. An IR spectrum identical to the one for the pure
silica-Beta zeolite was observed when a SnO -impregnated
2
(
cyclohexanone) with a tin Lewis acid site, on the basis that
silica-Beta sample (1.6 wt% Sn) was subjected to cyclohex-
anone adsorption/desorption. There is no shift of the IR
the carbonyl band in the IR region is very strong and
therefore allows the detection of small amounts of adsorbed
cyclohexanone.[ Therefore, cyclohexanone was adsorbed
and then desorbed successively at 50, 100 and 2008C. Fig-
ure 4a shows the IR spectra for Sn-Beta-2. After adsorption
at room temperature, an excess of cyclohexanone was physi-
carbonyl band of cyclohexanone with SnO but it does occur
2
±5]
with Sn tetrahedrally coordinated in the framework positions
of the zeolite.
When tin is replaced by titanium, a smaller shift of the
À1
carbonyl band can be observed. The signal is shifted ±2 cm
À1
À1
cally adsorbed, presenting a typical wavenumber of 171± cm
with Ti, compared with 48 cm with Sn (Figure 4c). After
for the carbonyl band. After desorption at 1008C, three bands
remained: two were in the region of the uncoordinated
desorption at 1008C only the shifted carbonyl band is
detected, and at 2008C all the cyclohexanone has been
desorbed. Both observations, the smaller shift and the easier
desorption, suggest that the interaction of the carbonyl
oxygen with framework titanium is weaker than the inter-
action between the carbonyl group and the tin center. This is
probably why the activity of Ti-Beta is lower than that of Sn-
Beta (see Table 5).
carbonyl group of cyclohexanone, and one was shifted to
À1
1
665 cm . The latter was the only band still detected after
desorption at 2008C. This band can be attributed to a tin ±
oxygen interaction since it does not appear when cyclo-
hexanone is adsorbed/desorbed on a pure silica-Beta zeolite
(
Figure 4b). With this zeolite, after adsorption of cyclohex-
anone and after desorption at 508C, the band of the normal
carbonyl group was obtained. After further desorption at
As with the ketone substrate, the lactone produced during
reaction can also be coordinated to the Sn center. This is
demonstrated when the corresponding series of IR spectra are
recorded with e-caprolactone as the adsorbed organic com-
1
008C, practically all the cyclohexanone was eliminated, and
À1
in any case no bands at wavenumbers <1700 cm were
Figure 4. IR spectra of Beta samples before adsorption of cyclohexanone or e-caprolactone (zeolite), after adsorption (258C) and after desorption at the
corresponding temperature (508C, 1008C, 2008C): a) cyclohexanone and Sn-Beta; b) cyclohexanone and Si-Beta; c) cyclohexanone and Ti-Beta; d) e-
caprolactone and Sn-Beta; e) e-caprolactone and Si-Beta.
Chem. Eur. J. 2002, 8, No. 20
¹ 2002 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
0947-65±9/02/0820-471± $ 20.00+.50/0
4713

A. Corma et al.
FULL PAPER
pound and with pure silica-Beta and a Sn-Beta as zeolite
samples. With both zeolites the same results are obtained for
e-caprolactone and cyclohexanone. In Sn-Beta, the lactone is
adsorbed by the metal center, and a shift of the carbonyl band
is still present after desorption at 2008C (Figure 4d). In pure
silica-Beta, no shift is observed for the carbonyl band, and the
lactone is almost completely desorbed at 1008C and com-
pletely eliminated after desorption at 2008C (Figure 4e). The
coordination of the lactone product indicates that, at high
conversions, the substrate competes for coordination to the
active center (necessary for activation) with the increasing
amount of product.
In summary, it has been shown by IR measurements that
the cyclohexanone substrate can coordinate to framework Sn
centers, polarizing the carbonyl group and increasing the
density of positive charge on the carbon atom of the carbonyl
group. This makes the positively charged carbon atom of the
carbonyl group more prone to a nucleophilic attack by
hydrogen peroxide.
in the latter both oxygen atoms are chemically equivalent, and
consequently a statistical 50:50 distribution of carbonyl and
alcoholic oxygen would be expected for the labeled products.
Also, a carbonyl oxide pathway could be excluded because
according to this mechanism all labeled oxygen would be
detected as alcoholic, and none as carbonyl.
For the present H O /Sn-Beta oxidation system, IR meas-
2
2
urements showed that the carbonyl group could be activated
by coordination to the Lewis acid tin center. It is possible that
the oxygen coordinated to the Lewis acid is transformed into a
good leaving group after the attack by hydrogen peroxide, and
in this way a carbonyl oxide or a dioxirane may be formed. To
find out which of these two mechanisms operates in our case,
1
8
we chose O-labeled 2-methylcyclohexanone as reactant.
During reaction, it is transformed into the corresponding e-
caprolactone, which disintegrates in a mass spectrometer
predominantly into two different fragments, each containing
one oxygen atom [Eq. (7)].
18O labeling experiments: Although the Baeyer± Villiger
oxidation was discovered as early as 1899, the mechanism of
the reaction was not clarified until 195±.[ One potential
mechanism, involving 1,2,4,5-tetraoxanes as intermediates,
could be excluded since these
1]
peroxides were isolated and it
R1
R2
O
O
R1
R2
was observed that they do not
transform into the rearrange-
ment products.[
This outcome makes it feasible to analyze the crude
reaction mixture directly by GC-MS without any additional
treatment. If the substrate is not 100% labeled, the ratio
calculated from mass peaks 128 and 1±0 gives the labeling
content of the whole molecule, which can be compared with
the labeling content at the carbonyl position. The latter is
obtained from the ratio of mass peaks 84 and 86.
To verify the reliability of the test reaction, the Baeyer±
Villiger reaction was first carried out with mCPBA. For this
oxidant the mechanism is well established and the labeled
oxygen should only be detected in the carbonyl position.
Indeed, the ratios of the mass peaks 84/86 and 128/1±0 were
identical at 0.11:0.89 [Table 6, Eq. (8)].
O
O
±6]
1
,2,4,5-tetraoxanes
In the 1950s, a dioxirane, a
carbonyl oxide, and an adduct
of the peracid and a carbonyl group (which was called a
Criegee intermediate) were discussed as potential intermedi-
ates (Scheme 2).
with 18O proved that the rearrangement could proceed only
via the Criegee intermediate.[ All labeled oxygen was
detected as carbonyl oxygen in benzoic acid phenyl ester,
the Baeyer± Villiger product. This result was incompatible
with a mechanism involving a dioxirane intermediate because
[±7]
A benzophenone labeling experiment
±8]
Taking the method of preparation of the labeled substrate
into account, it was clear that any water molecule (unlabeled)
present in the reaction solution would lower the labeling
content in the substrate molecule. Consequently, not only was
the reaction carried out under anhydrous conditions by using
±
ä molecular sieves to absorb water, but the solvent and
mCPBA were dried well and stored over ± ä molecular sieves
before use. With these precautions, it was demonstrated that
Table 6. Labeled product distribution from Baeyer± Villiger reaction of
1
8
2
-methyl[ O]cyclohexanone with mCPBA.
1
8
Reaction conditions
Labeled ( O) content [%]
1
2
±
4
5
6
Scheme 2. Potential mechanism for the Baeyer± Villiger rearrangement
with peracids via the carbonyl oxide, the dioxirane, or the ™Criegee∫
intermediate and their consequences for the position of a labeled oxygen
atom in the product.
±[a]
0
±[a]
70
mCPBA, MTBE, ± ä molecular sieves 89 81 89
Sn-Beta, H , MTBE, MgSO , ±0 min 89 57 71
0
0
2
O
2
4
[a] Product not formed with mCPBA.
0947-65±9/02/0820-4714 $ 20.00+.50/0
4714
¹ 2002 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2002, 8, No. 20

Baeyer± Villiger Oxidation
4708±4717
the Baeyer± Villiger reaction with peracids and 18O-labeled
Conclusion
2-methylcyclohexanone was a good test to establish the
reaction mechanism.
We have shown that Sn-Beta with Sn tetrahedrally coordi-
nated in zeolite framework positions is an excellent hetero-
geneous catalyst for Baeyer± Villiger oxidations of cyclic
ketones and aromatic aldehydes, with H O as oxidant. The
The oxidation of labeled 2-methylcyclohexanone with
hydrogen peroxide catalyzed by Sn-Beta [Eq. (8)] could not
be carried out under anhydrous conditions, primarily because
water (unlabeled) is always a by-product. Moreover, molec-
ular sieves could not be employed for drying because hydro-
gen peroxide would be adsorbed as well as water. Therefore,
only the solvent was pre-dried with molecular sieves and the
2
2
system provides excellent catalytic activity and shape selec-
tivity, and also chemoselectivity with unsaturated ketones.
Unlike previous catalysts, Sn-Beta activates the carbonyl
group but does not activate H O . The carbonyl group is then
2
2
reaction was carried out in the presence of MgSO within a
attacked selectively by the H O . Labeling experiments with
2 2
O have demonstrated unambiguously that the mechanism of
4
1
8
short reaction time (±0 min). The total amount of labeled
oxygen in the lactone product decreased from 89 to 71%, but
the determining observation from a mechanistic viewpoint
was that 100% of the labeled oxygen was found at the
carbonyl position. This demonstrates unambiguously that the
mechanism of the Baeyer± Villiger oxidation with hydrogen
peroxide catalyzed by Sn-Beta proceeds through a Criegee
adduct of the hydrogen peroxide with the activated ketone
substrate and not through a dioxirane intermediate. Therefore
we conclude that the Baeyer ± Villiger oxidation of cyclic
ketones using H O as oxidant and Sn-Beta as catalyst follows
the reaction for the Baeyer± Villiger oxidation with Sn-Beta
and H O occurs through a Criegee-type intermediate.
2
2
Experimental Section
2
2
the mechanism in Scheme ±.
Synthesis of the catalyst: Sn-Beta zeolites with different Sn contents were
synthesized according to the following procedure.[ Tetraethyl orthosili-
cate (TEOS) was hydrolyzed in an aqueous solution of tetraethylammo-
±9]
4
nium hydroxide (TEAOH) with stirring. An aqueous solution of SnCl ¥
1
19
5
2
H O was then added. In the case of a Sn-enriched sample (Sn-Beta-2*),
1
19
metallic tin enriched with the Sn isotope (82.9%) was dissolved in
hydrochloric acid and this solution was used as the source of Sn. The
synthesis mixture was stirred until the ethanol formed upon hydrolysis of
TEOS had been evaporated. HF was added to the resulting clear solution,
and a thick gel was formed with the composition SiO
TEAOH:7.5H
O:0.54HF. Three gels were prepared with x ˆ 0.0042,
.008±, and 0.0125, respectively. They were seeded with nanocrystalline
20 nm crystals) pure silica-Beta at a 1 wt% level. Well crystallized Sn-Beta
zeolites with Sn contents of 0.8 (Sn-Beta-1), 1.6 (Sn-Beta-2), and 2.± wt%
Sn-Beta-±), as determined by chemical analysis, were obtained by
2 2
:xSnO :0.54-
2
0
(
(
crystallization in a rotating Teflon-lined stainless steel autoclave at 1408C
for 20 d.
XRD showed the Sn-Beta zeolites calcined at 5008C to be highly crystalline
zeolites, and no peaks of SnO
Figure 5). Nitrogen adsorption experiments on the calcined Sn-Beta
samples gave an isotherm very similar to that of pure silica-Beta with
À1 and BETsurface areas of 450 ±
micropore volumes of 0.20 ± 0.21 cm
À1. The IR spectra in the framework vibration regions of Sn-Beta-1
475 m
2
were observed in the diffractogram
(
Scheme ±. Catalytic cycle for the Baeyer± Villiger reaction with hydrogen
peroxide, catalyzed by Sn-Beta. First the ketone is coordinated to the Lewis
acid center, and thereby the carbonyl group is activated (in situ IR
spectroscopy). Then the hydrogen peroxide attacks the more electrophilic
carbonyl carbon atom (in situ IR spectroscopy). After the rearrangement
step the lactone product is replaced by a new substrate molecule. This
mechanism is analogous to that for the peracids and has been proved by an
±
g
2
g
and pure silica-Beta are shown in Figure 6.
1
8
O labeling experiment.
The formation of the second product, the regioisomeric a-
methyl-e-caprolactone, confirms the conclusions regarding
the reaction mechanism, although fragmentation into a
species with only a carbonyl oxygen (ÀH CO in this case;
2
m/z 98 with 16O and m/z 100 with O) was not predominant.
The ratio of the 128/1±0 mass peaks was identical to the ratio
for the 98/100 mass peaks, namely 0.±0:0.70 and 0.±0:0.70,
respectively [Eq. (8)]. The reaction mechanism proceeds via
the same Criegee intermediate as the main product, but
oxygen is inserted into the C ± C bond between the carbonyl
carbon and its less substituted neighbor.
18
Figure 5. X-ray diffraction patterns of calcined Sn-Beta and pure silica-
Beta zeolites.
Chem. Eur. J. 2002, 8, No. 20
¹ 2002 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
0947-65±9/02/0820-4715 $ 20.00+.50/0
4715

A. Corma et al.
FULL PAPER
by GC-MS; after 2 h the theoretical maximum 18O content of 89.1% was
reached. The resultant ketone was used without any further treatment or
‡
purification. MS: m/z (%): 114 (8±.4) [M] , 112 (9.29), 68 (100).
Preparation of a solution of H
±5% H ; 5 mL; 4.52 g) was stirred with MTBE (20 mL; 18.4 g). MgSO
was added to remove the water. After decantation, the drying process was
repeated, and the MgSO was removed by filtration. The H content of
the resultant sample was 7.1 Æ 0.6%.
2 2 2 2
O in MTBE: An aqueous H O solution
(
2
O
2
4
4
2 2
O
Preparation of a solution of mCPBA in MTBE: mCPBA (500 mg) was
dissolved in MTBE (5 mL; ±.70 g). The solution was dried with MgSO and
4
the final content of mCPBA was 9.7%. For the reaction under anhydrous
conditions, this solution was dried with molecular sieves.
1
6
Oxidation of 2-methyl[ O]cyclohexanone (unlabeled) with H
lyzed by Sn-Beta-2: 2-Methylcyclohexanone (48 mg; 0.428 mmol) was
added to a mixture of an H solution (7.1% H ; 1±9 mg; 0.291 mmol)
2 2
O cata-
2
O
2
2 2
O
in MTBE, and MTBE (1.50 g) and the reaction was started with the
addition of Sn-Beta-2 (25 mg). The reaction mixture was heated to 568C for
±
h. The solids were removed and the solution was submitted directly to GC
Figure 6. IR spectra in the framework vibration region of calcined Sn-Beta
and pure silica-Beta zeolites.
and GC-MS analysis. Conversion 44%; products e-methyl-e-caprolactone
and a-methyl-e-caprolactone (8±:17 mixture). Unconverted 2-methyl-
‡
cyclohexanone: MS m/z (%): 114 (0.472), 112 (91.7) [M] , 110 (1.08 Â
The spectra of the two samples are very similar except for a shoulder at
˱
À1
10 ), 68 (100). a-Methyl-e-caprolactone: MS m/z (%): 1±0 (0.02±2), 128
ꢀ
960 cm exhibited by the Sn-containing Beta. This band has also been
‡
˱
(
1.±7) [M] , 126 (2.67 Â 10 ), 86 (2.0±), 84 (78.5), 82 (0.280), 55 (100). a-
observed in other zeolites containing framework metals and has been
attributed to Si ± O ± M vibrations.[26] Although its origin is still under
‡
Methyl-e-caprolactone: MS m/z (%): 1±0 (0.0206), 128 (28.5) [M] , 126
˱
(
9.69 Â 10 ), 100 (0.644), 98 (15.6), 96 (0.0±98), 56 (95.9), 55 (100).
discussion, it is generally taken as an indication of metal incorporation into
the framework.
119Sn MAS NMR spectra: These were recorded with a Varian VXR-S400
WB spectrometer at a frequency of 149.0 MHz using a 5 mm high-speed
dotty probe. Samples were packed into silicon nitride rotors and spun at
1
8
Oxidation of 2-methyl[ O]cyclohexanone with mCPBA in the presence of
3
was added to mCPBA solution (9.67% mCPBA; 506 mg; 0.28± mmol) in
MTBE (dried over ± ä molecular sieves), anhydrous MTBE (1.00 g), and
18
ä molecular sieves: 2-Methyl[ O]cyclohexanone (5± mg; 0.464 mmol)
8
kHz. The spectra were recorded using pulses of ±.5 ms to flip the
magnetization at 608 with recycle delays of 50 s. Chemical shifts are
referred to tetramethyltin using SnO
±
ä molecular sieves, and the reaction mixture was heated to 568C for 1 h.
The solids were removed by filtration and the solution was submitted
directly to GC and GC-MS analysis. Conversion 14%; e-methyl-e-
caprolactone was the exclusive product. Unconverted 2-methyl[ O]cyclo-
hexanone: MS m/z (%): 114 (7±.8) [M] , 112 (17.8), 68 (100); O content
8
8
2
as a secondary reference (d ˆ À604).
For measurements with dehydrated samples, the zeolites were treated
under dynamic vacuum at 4008C overnight and transferred to the rotor
within a glovebox under an atmosphere of nitrogen. The spectrum of
rehydrated samples were acquired after exposing the ™dehydrated sam-
ples∫ to ambient conditions for several hours.
18
‡
18
‡
0.5%; e-methyl-e-caprolactone: MS m/z (%): 1±0 (1.26) [M] , 128 (0.155),
18
18
6 (90.6), 84 (11.4), 57 (100). O content in the molecule 89.2%;
O
content in carbonyl position 89.1%.
In situ IR measurements: For the IR study of the interaction of cyclo-
1
8
Oxidation of 2-methyl[ O]cyclohexanone with H
2
O
2
catalyzed by Sn-
Beta: 59 mg (0.517 mmol) of 2-methyl[ O]cyclohexanone was added to the
mixture of an H solution (7.1%; 11± mg; 0.2±6 mmol) in MTBE,
anhydrous MTBE (1.6± g) and MgSO (±9 mg). The reaction was started
with the addition of Sn-Beta-2 (25 mg). The reaction mixture was heated to
68C for ±0 min. To stop the reaction, by removing the water and the
unconverted H , molecular sieves (± ä) were added. After 1 h, the solids
hexanone with Sn-Beta and other zeolites, the solid was first treated
1
8
À4
overnight at 4008C under dynamic vacuum (10 Torr) to remove adsorbed
2
O
2
2
H O, then cyclohexanone was adsorbed and the IR spectrum was recorded.
Afterwards, the ketone was desorbed successively at 50, 100, and 2008C.
An IR spectrum was recorded after each desorption.
4
5
Oxidations: Ketones and solvents of the highest purity available (ꢂ98%)
were purchased from Aldrich and used without further purification. The
only exceptions were bicyclo[±.2.0]hept-2-en-6-one (Merck; purity >95%)
and hydrogen peroxide (Fluka; ±5 wt% in water). GC analyses were
carried out on an HP 5890 gas chromatograph equipped with a 25 m HP-5
column and an adequate temperature programming capability. A Fisons 80±5
gas chromatograph coupled with a Fisons MD800 mass spectrometer was
used for GC-MS analyses to identify products. The labeling experiments
were analyzed by GC-MS on an Agilent HP6890 apparatus.
2
O
2
were removed by filtration and the solution was analyzed directly by GC
and GC-MS. Conversion 20%; the product was an 85:15 mixture of e-
methyl-e-caprolactone and a-methyl-e-caprolactone. Unconverted 2-meth-
yl[ O]cyclohexanone: MS m/z (%): 114 (51.4) [M] , 112 (±8.6), 68 (100);
O content 57.0%. e-Methyl-e-caprolactone: MS m/z (%): 1±0 (1.1±) [M] ,
28 (0.485), 86 (86.8), 84 (±4.4), 57 (100). O content in the molecule
0.0%; O content in carbonyl position 71.6%. a-Methyl-e-caprolactone:
MS m/z (%): 1±0 (19.7) [M] , 128 (8.5±), 100 (11.5), 98 (4.80), 56 (100), 55
(99.6). O content in the molecule 69.8%; O content in carbonyl position
70.2%.
18
‡
18
‡
18
1
7
18
‡
1
8
18
General procedure for the Baeyer± Villiger oxidation with H
Beta: The ketone (ꢀ1 mmol) and hydrogen peroxide (±5%
1.5 mmol) were dissolved in MTBE (±.00 g), or the ketone (ꢀ1 mmol)
and hydrogen peroxide (±5% H
; ꢀ± mmol) were dissolved in dioxane
±.00 g). A sample (50 mg) of the catalyst was added, and the reaction
2
O
2
and Sn-
2 2
H O ;
ꢀ
Determination of the configuration of the regioisomers in the oxidation of
2
O
2
1
bicyclohept-3-en-1-one by
H
NMR spectroscopy: Major isomer, 1-
, 258C): d ˆ 2.48
(
oxabicyclooct-4-en-2-one:[
40]
1
H NMR (±00 MHz, CDCl
±
mixture was stirred and heated to 568C (MTBE) or 808C (dioxane) for 7 h.
The reaction was followed by gas chromatography, and the products were
identified by comparison with reference samples by GC-MS spectroscopy,
(
6
dd, J ˆ 0.8, 18 Hz, 1H), 2.75 (m, ±H), ±.5± (m, 1H), 5.15 (ddd, J ˆ 2.4, ±.9,
Hz, 1H, 7-H), 5.60 (m, 1H, 4-H or
H
H
1
5-H), 5.80 (m, 1H, 4-H or 5-H).
H
1
or after purification by H NMR spectroscopy. In the case of the aromatic
6
O 2
7
H 5
O
aldehydes, the aldehyde (±.7 mmol) and hydrogen peroxide (±5%;
8
Minor isomer, 2-oxabicyclooct-4-en-1-
3
H
4
5
.1 mmol) were dissolved in dioxane (±.0 g) or toluene (±.0 g). A sample
1
H
H
one:
H
NMR (±00 MHz, CDCl
±
,
H
(
50 mg) of catalyst Sn-Beta-2 was added, and the reaction mixture was
2
2
58C): d ˆ 2.75 (m, ±H), ±.15 (dt, J ˆ
H
H 1 O
stirred and heated to 808C for 7 h.
.7, 7.5 Hz, 1H), ±.60 (m, 1H), 4.25
H
6
Preparation of 18O-labeled 2-methylcyclohexanone: 2-Methylcyclohexa-
(dd, J ˆ 1.5, 9.± Hz, 1H, ±-H), 4.44 (dd,
J ˆ 7.0, 9.± Hz, 1H, ±-H), 5.66 (m, 1H,
4-H or 5-H), 5.88 (m, 1H, 4-H or 5-H).
7
O 2
H 5
8
1
8
18
3
H
none (500 mg, 4.46 mmol) and H
2
O (ꢀ96% O; 1.00 g, 49.8 mmol) were
4
H
H
H
stirred in a biphasic system at room temperature. The reaction was followed
4716
¹ 2002 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
0947-65±9/02/0820-4716 $ 20.00+.50/0
Chem. Eur. J. 2002, 8, No. 20

Baeyer± Villiger Oxidation
4708±4717
[
[
[
21] G. La˚mann, S. Pˆtsch (Max-Delbrueck-Centrum f¸r Molekulare
Medizin), DE4±44645, 1995 [Chem. Abstr. 1995, 123, 74882].
22] I. M. Godfrey, M. V. Sargent, J. A. Elix, J. Chem. Soc. Perkin Trans. 1
Acknowledgement
The authors thank UOP LLC and CICYT (MAT2000-1±92) for financing
this work. M.R. is grateful to the Spanish Ministry of Science and
Technology for a ™Ram o¬ n y Cajal∫ Fellowship.
1
974, 1±5± ± 1±54.
2±] D. R. Kelly, P. W. H. Wan, J. Tang inBiotechnology, Vol. 8a (Eds.: H.-J.
Rehm, G. Reed, A. P¸hler, P. Stadler; Volume Ed.: D. R. Kelly),
Wiley-VCH, Weinheim, 1998.
[
24] W. A. Herrmann, R. W. Fischer, J. D. G. Correira, J. Mol. Catal. 1994,
94, 21± ± 22±.
[
[
[
1] M. Renz, B. Meunier, Eur. J. Org. Chem. 1999, 7±7 ± 750.
2] G. R. Krow, Org. React. 1993, 43, 251 ± 798.
±] C. Bolm, O. Beckmann, T. K. K. Luong, in Transition Metals for
Organic Synthesis, Vol. 2 (Eds.: M. Beller, C. Bolm), Wiley-VCH,
Weinheim, 1998, pp. 21± ± 218.
[25] A. M. F. Phillips, C. Rom aƒ o, Eur. J. Org. Chem. 1999, 1767 ± 1770.
[26] N. K. Mal, A. Bhaumik, V. Ramaswamy, A. A. Belhekar, A.
Ramaswamy, Stud. Surf. Sci. Catal. 1995, 94, ±17 ± ±24.
[27] P. Fejes, J. B. Nagy, K. Kov a¬ cs, G. Vank o¬ , Appl. Catal. A: General 1996,
145, 155 ± 184.
[28] N. K. Mal, V. Ramaswamy, S. Ganapathy, A. V. Ramaswamy, Appl.
Catal. A: General 1995, 125, 2±± ± 245.
[29] K. Chaudari, T. K. Das, P. R. Rajmohanan, K. Lazar, S. Sivasanker,
A. J. Chandwadkar, J. Catal. 1999, 183, 281 ± 291.
[±0] N. K. Mal, V. Ramaswamy, P. R. Rajamohanan, A. V. Ramaswamy,
Micropor. Mater. 1997, 12, ±±1 ± ±40.
[±1] N. J. Clayden, C. M. Dobson, A. Fern, J. Chem. Soc. Dalton Trans.
1989, 84± ± 847.
[
[
[
4] G. Strukul, Nature 2001, 412, ±88 ± ±89.
5] H. Heaney, Aldrichim. Acta 1993, 26, ±5 ± 45.
6] C. Bolm, G. Schlingloff, K. Weickhardt, Tetrahedron Lett. 1993, 34,
±
405 ± ±408.
[
7] G. Strukul, Angew. Chem. 1998, 110, 1256 ± 1267; Angew. Chem. Int.
Ed. 1998, 37, 1198 ± 1209.
[
[
8] R. Raja, J. M. Thomas, G. Sankar, Chem. Commun. 1999, 525 ± 526.
9] K. Kaneda, S. Ueno, T. Imanaka, J. Chem. Soc. Chem. Commun. 1994,
7
97 ± 798.
[
[
[
10] I. C. Chisem, J. Chisem, J. H. Clark, New J. Chem. 1998, 81 ± 82.
11] T. Yamada, T. Takai, O. Rhode, T. Mukaiyama,Chem. Lett. 1991, 1 ± 4.
12] J. Fischer, W. F. Hˆlderich, Appl. Catal. A: General 1999, 180, 4±5 ±
[±2] B. Notari, Adv. Catal. 1996, 41, 25± ± ±±4.
[±±] S. Shambayati, S. L. Schreiber, in Comprehensive Organic Synthesis,
Vol. 1 (Eds.: B. M. Trost, I. Fleming), Pergamon, Oxford, 1991,
pp. 28± ± ±24.
[±4] M. Santelli, J.-M. Pons, Lewis Acids and Selectivity in Organic
Synthesis, CRC Press, Boca Raton, 1995, pp. 1 ± 19.
[±5] B.-P. Susz, R. Weber, Helv. Chim. Acta 1970, 53, 2085 ± 2097.
[±6] W. Dilthey, M. Inckel, H. Stephan, J. Prakt. Chem. 1940, 154, 219 ± 2±7.
[±7] W. v. E. Doering, L. Speers, J. Am. Chem. Soc. 1950, 72, 5515 ± 5518.
[±8] W. v. E. Doering, E. Dorfman, J. Am. Chem. Soc. 1953, 75, 5595 ± 5598.
[±9] S. Valencia, A. Corma (UOP LCC), US596847±A, 1999 [Chem.
Abstr. 1999, 131, ±01204].
4
4±.
[
1±] G.-J. ten Brink, J.-M. Vis, I. W. C. E. Arends, R. A. Sheldon, J. Org.
Chem. 2001, 66, 2429 ± 24±±.
[
[
14] K. Neimann, R. Neumann, Org. Lett. 2000, 2, 2861 ± 286±.
15] G. A. Olah, T. Yamato, P. S. Iyer, N. J. Trivedi, B. P. Singh, G. K. S.
Prakash, Mater. Chem. Phys. 1987, 17, 21 ± ±0.
[
[
[
[
[
16] C. D. Chang, S. D. Hellring (Mobil Oil Corp.), US4870192, 1989
[
Chem. Abstr. 1990, 112, 158077].
17] A. Corma, L. T. Nemeth, M. Renz, S. Valencia, Nature 2001, 412, 42± ±
25.
18] M. T. H. Tarafder, W. M. Z. W. Yunus, S. Silong, Aust. J. Chem. 1997,
0, 229 ± 2±2.
19] P. A. Ochsner, K. F. De Polo (L. Givaudan & Cie), DE2510861, 1975
Chem. Abstr. 1976, 84, 21996].
4
[40] P. A. Grieco, J. Org. Chem. 1972, 37, 2±6± ± 2±64.
5
[
Received: January 28, 2002
20] J. C. van der Waal, E. J. Creyghton, P. J. Kunkeler, K. Tan, H.
van Bekkum, Top. Catal. 1998, 4, 261 ± 268.
Revised: June 18, 2002 [F±829]
Chem. Eur. J. 2002, 8, No. 20
¹ 2002 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
0947-65±9/02/0820-4717 $ 20.00+.50/0
4717