Epoxidation versus Baeyer-Villiger oxidation: The possible role of Lewis acidity in the control of selectivity in catalysis by transition metal complexes

Article abstract of DOI:10.1002/ejic.200300584

The synthesis of a series of cationic complexes of Pt^II containing triphos or triphosPO and either a solvent, a hydroxo or a hydroperoxo as the fourth ligand is reported. The complexes are characterized by IR and ³¹P{¹H} NMR spectroscopy and by molar conductivity data, and their Lewis acid characteristics determined by NMR techniques. All complexes are good catalysts for the epoxidation of olefins and the Baeyer-Villiger oxidation of ketones under mild conditions using hydrogen peroxide as oxidant. The latter reaction is more Lewis-acid demanding in terms of the electronic characteristics of the catalysts. The use of Lewis-acidic metal centers allows the selective oxidation of unsaturated ketones to the corresponding unsaturated esters without oxidation of the carbon-carbon double bond, indicating that a strong Lewis-acid character of the catalyst can be an important feature in the design of catalysts capable of separating the two processes. Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim, Germany, 2004.

Full text of DOI:10.1002/ejic.200300584

FULL PAPER
Epoxidation versus Baeyer؊Villiger Oxidation: The Possible Role of Lewis
Acidity in the Control of Selectivity in Catalysis by Transition Metal
Complexes
Agostino Brunetta[a] and Giorgio Strukul*[a]
Keywords: Lewis acids / Homogeneous catalysis / Oxidation / Peroxides
The synthesis of a series of cationic complexes of PtII con-
taining triphos or triphosPO and either a solvent, a hydroxo
characteristics of the catalysts. The use of Lewis-acidic metal
centers allows the selective oxidation of unsaturated ketones
or a hydroperoxo as the fourth ligand is reported. The com- to the corresponding unsaturated esters without oxidation of
plexes are characterized by IR and 31P{1H} NMR spectro-
the carbon-carbon double bond, indicating that a strong
scopy and by molar conductivity data, and their Lewis acid Lewis-acid character of the catalyst can be an important fea-
characteristics determined by NMR techniques. All com-
plexes are good catalysts for the epoxidation of olefins and
the Baeyer−Villiger oxidation of ketones under mild condi-
tions using hydrogen peroxide as oxidant. The latter reaction
is more Lewis-acid demanding in terms of the electronic
ture in the design of catalysts capable of separating the two
processes.
( Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim,
Germany, 2004)
Introduction
More recently, typical Lewis acids such as AlIII [4] MgII,[5]
,
and SnIV [6] have also been reported in the BV oxidation of
ketones, emphasizing the importance of having a Lewis-
acid center to promote the reaction by activation of the ke-
tone.
In transition metal chemistry, the idea of Lewis acidity is
generally associated with electron-deficient metal centers
such as the early metals in their highest oxidation state or
the first row late metals such as CuII and ZnII. This simple
notion is supported by a huge number of applications in
organic synthesis.[1] Within this framework, the idea of as-
sociating Lewis acidity with electron-rich metal centers like
PdII and PtII for catalytic applications is a relatively recent
one, the necessary condition to exploit the Lewis acidity of
the latter systems being the presence of at least one positive
charge on the complex. In the past few years, the number
of applications of these systems in catalysis has been in-
creasing, although, among the examples reported,[2] none
refers to oxidation reactions.
Several years ago we reported the first example of tran-
sition metal catalysis in the BaeyerϪVilliger (BV) oxidation
of ketones, with H2O2 as the oxidant, and using a cationic
complex of PtII as the catalyst.[3] The mechanism of the Figure 1. Reaction mechanism for the BaeyerϪVilliger oxidation of
ketones with hydrogen peroxide catalyzed by [Pt(CF3)(P-P)(solv)]ϩ-
reaction[3b] (Figure 1) clearly indicates that, having no role
type complexes
in the activation of the oxidant, the major function of Pt in
the catalytic cycle is to behave as a Lewis acid for the acti-
vation of the ketone. This view is strengthened by the anal-
ogies (Figure 1) between the quasi-peroxymetallacyclic in-
termediate 1 and the Criegee intermediate involved in the
The same complex reported in Figure 1 was also recog-
nized to be an efficient catalyst for the epoxidation of olef-
ins with H2O2 under mild conditions,[7] although in this
case the mechanism of the reaction (Figure 2) indicates that
the function of the catalyst is more complex[7d] as it acti-
vates both the olefin towards nucleophilic attack and, at the
purely organic reaction using peracids (Pt replaces Hϩ).
[a]
Department of Chemistry, University of Venice
Dorsoduro 2137, 30123 Venezia, Italy
Fax: (internat.) ϩ 39-041-257-8517
E-mail: strukul@unive.it
1030
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
DOI: 10.1002/ejic.200300584
Eur. J. Inorg. Chem. 2004, 1030Ϫ1038

Epoxidation versus BaeyerϪVilliger Oxidation
FULL PAPER
same time, hydrogen peroxide, increasing its nucleophilicity electron density of the metal. Since the presence of a coor-
by coordination.
dinated diphosphane seems to be a crucial requirement to
promote the activity of these complexes,[3,7,8] to achieve this
goal we chose tridentate ligands of the type P-P-P and P-
P-PϭO, bearing, respectively, three P donors and two P and
one O donors. This would ensure the formation of bis-cat-
ionic complexes possessing, in principle, a significant
Lewis acidity.
Square planar PtII complexes such as 2 and 3 containing
tridentate ligands (Scheme 2) were synthesized following
the procedure outlined in Scheme 3. This was already pub-
lished for solvento complexes by Bergamini et al.[9] al-
though the species were characterized in situ (NMR evi-
dence) but never isolated. The isolation of the hydroxo and
hydroperoxo complexes of 2 could be easily accomplished
by addition of, respectively, OHϪ to the corresponding sol-
vento complex or H2O2 to the hydroxo complex. Con-
versely, in the case of 3, given the lability of the oxygen
donor, the same procedure led to the formation of a differ-
ent species that was characterized in situ by NMR spec-
troscopy. Attempts to isolate this species synthetically using
the procedure adopted for 2b (i.e. addition of a stoichio-
metric amount of OHϪ in acetone/H2O) led to the forma-
tion of a complex mixture showing a different NMR spec-
trum. Isolation of a species (albeit in moderate yield) with
the same NMR spectrum as observed in the in situ experi-
ment could be accomplished in CH2Cl2 by adding an excess
of OHϪ in water. This species is most likely a neutral bis-
hydroxo complex of the type [Pt(OH)2(P-P-PϭO)] (4) as
seems to be indicated by the full spectroscopic characteriz-
ation (see below).
Figure 2. Reaction mechanism for the epoxidation of olefins with
hydrogen peroxide catalyzed by [Pt(CF3)(P-P)(solv)]ϩ-type com-
plexes
Platinum complexes are the only catalysts known in the
literature that catalyze both the BV and the epoxidation
reactions with high activity. Moreover, they do so using hy-
drogen peroxide as the oxidant, i.e. a cheap, environmen-
tally friendly oxidant, but also an elusive one for selective,
non-radical-type oxidations. The versatility of these systems
offers a unique opportunity to investigate in more detail the
electronic properties of the metal center that allow us to
separate the BV reaction from the epoxidation reaction.
This aspect is crucial for selectivity whenever a bifunctional
molecule is oxidized, and may have an important impact
in organic synthesis, as both the BV and the epoxidation
reactions are widely used as intermediate steps in complex
synthetic procedures, leading to the preparation of pharma,
agro and fine chemicals. As an example, Scheme 1 shows
the theoretical case of dihydrocarvone where both parallel
and consecutive reactions are possible leading to the un-
selective formation of the epoxy lactone. The aim of the
present work is to establish whether the Lewis acidity of the
metal center is of importance for controlling the relative
rates of the BV and epoxidation reactions.
Scheme 2
Characterization of Complexes
All complexes were characterized by IR and 31P{1H}
NMR spectroscopy, and molar conductivity measurements.
A summary of the characterization data is reported in
Table 1. For the purpose of simplicity, and with reference
to Scheme 2, in reporting the NMR spectroscopic data Pa
refers to the phosphorus trans to P (or to O in complexes
3), Pb to the phosphorus trans to X and (only in complexes
3) Pc to the oxidized phosphorus. As shown in Table 1, in
the IR spectrum all complexes display the broad band
Scheme 1
Results and Discussion
Synthesis of Complexes
Ϫ
characteristic of non-coordinated BF4 in the 1100Ϫ960
To address the problem exposed above, one has to design cmϪ1 range. Additionally, complexes 2b and 2c show typical
a class of cationic homologous complexes of platinum() weak OϪH stretching frequencies at 3620 and 3483 cmϪ1
with very similar geometric constraints, but differing in the respectively, in agreement with previously reported hydroxo
,
Eur. J. Inorg. Chem. 2004, 1030Ϫ1038
www.eurjic.org
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1031

A. Brunetta, G. Strukul
FULL PAPER
Scheme 3
Table 1. Characterization data for complexes 2 and 3[a]
Complex
IR
31P{1H} NMR
(δ in ppm, J in Hz)
Molar conductivity
(cmϪ1
)
(ΩϪ1molϪ1cm2)
1
2a
1100Ϫ960 (b)
Pa 50.6 (s), JP,Pt ϭ 2452
115Ϫ120
1
Pb 80.2 (s), JP,Pt ϭ 3475
2
1
2b
3620 (w)
1100Ϫ960 (b)
Pa 77.7 (d), JP,P ϭ 3.7, JP,Pt ϭ 2575
79.6
2
1
Pb 39.6 (t), JP,P ϭ 3.7, JP,Pt ϭ 2625
2
1
2b[b]
2c
Pa 75.2 (d), JP,P ϭ 4.1, JP,Pt ϭ 2592
2
1
Pb 39.2 (t), JP,P ϭ 4.1, JP,Pt ϭ 2620
2
1
3483 (w)
Pa 70.9 (d), JP,P ϭ 4.3, JP,Pt ϭ 2385
90.3
172
2
1
1100Ϫ960 (b)
1100Ϫ960 (b)
1124 (s)
Pb 42.4 (t), JP,P ϭ 4.3, JP,Pt ϭ 2811
1
3a
Pa 38.2 (s), JP,Pt ϭ 3753
1
Pb 42.9 (s), JP,Pt ϭ 3890
3
Pc 55.1 (d), JPb,Pc ϭ 9.0
2
1
4[b]
1178 (s)
Pa 33.4 (d), JPa,Pb ϭ 11.0, JP,Pt ϭ 3477
8Ϫ10
3599 (w)[c]
Pb 39.4 (dd), JPa,Pb ϭ 11.0, JPb,Pc ϭ 52.5, JP,Pt ϭ 3480
2
3
1
3
Pc 31.2 (d), JPc,Pb ϭ 52.5
1
2d[d]
3d[d]
Pa 48.4 (s), JP,Pt ϭ 2459
1
Pb 82.2 (s), JP,Pt ϭ 3434
2
1
Pa 37.9 (d), JPa,Pb ϭ 6.8, JP,Pt ϭ 3900
1
Pb 43.2 (s), JP,Pt ϭ 3714
3
Pc 55.1 (d), JPb,Pc ϭ 11.1
[a]
[b]
[c]
IR in nujol; NMR in [D6]acetone at 25 °C; molar conductivity: 10Ϫ3  solution in EtOH. NMR spectrum recorded in CD2Cl2.
[d]
In 1,2-dichloroethane solution. NMR spectra recorded in acetone at Ϫ70 °C.
and hydroperoxo complexes of PtII.[7e,10] Similarly, in com-
plex 3a the PϭO stretching frequency of the oxygen-coordi- addition of an excess of OHϪ to 3a results in the formation
nated ligand is observed at 1124 cmϪ1 [11]
of a more static species (4) in which PϪP coupling con-
At variance with the NMR spectra mentioned above, the
.
The NMR spectroscopic data were taken in [D6]acetone. stants can be observed. The signal assigned to Pc is strongly
Complexes 2a and 3a give very broad signals in CD2Cl2 so shifted (about 20 ppm) to higher fields with respect to the
a more coordinating solvent was necessary in order to ob- parent compound, indicating that the O donor is probably
serve a relatively static spectrum at room temperature. The no longer in the coordination sphere of the metal. This view
signals are moderately broad so that PϪP coupling con- seems to be confirmed by the observation that in the IR
stants cannot be determined. In general, Pb when it is trans spectrum of the isolated species the PϭO stretching fre-
to a solvent (as in 2a and 3a), displays higher PϪPt coup- quency shifts from 1124 cmϪ1 in 3a to 1178 cmϪ1 in 4.[11]
ling constants. The same happens to Pa in 3a where it is Additionally, the PaϪPt and PbϪPt coupling constants are
trans to the labile oxygen ligand. It should be noted that practically identical (3477 and 3480 Hz respectively). The
the PbϪPt coupling constants for 2b and 2c are significantly IR spectrum of 4 in dichloroethane solution also shows a
smaller than in the case of mononuclear diphosphane com- weak band at 3599 cmϪ1, typical of hydroxo complexes of
plexes of platinum[7e,10a,12] where they fall in the PtII.[7d,7e,8a,10a] On this basis two possibilities are possible,
3000Ϫ3400 Hz range, reflecting the moderate trans effect of both in agreement with the NMR spectroscopic data but
the -OH (-OOH) ligand, as was suggested by Appleton and neither corresponding to 3b: the complex might be either a
Bennett.[12] The data reported for 2a, 2b and 3a are ident- mononuclear bis-hydroxo complex or a dinuclear µ-hydroxo
ical to those observed in situ by Bergamini et al.[9]
complex. The latter class of complexes is very stable and is
Conductivity data for 2a, 3a, and 2b, 2c taken in 10Ϫ3

relatively common in the literature in the case of PtII.[14]
A
MeOH solutions are in the range generally accepted for 1:2 possible formulation of 4 as a dinuclear µ-hydroxo complex
and 1:1 electrolytes, respectively.[13] might be supported by the values of the 1JP,Pt coupling con-
1032
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
Eur. J. Inorg. Chem. 2004, 1030Ϫ1038

Epoxidation versus BaeyerϪVilliger Oxidation
FULL PAPER
stants that fall in the 3200Ϫ3600 Hz range typical for this of a shoulder at about 2225 cmϪ1, indicative of the presence
class of complexes.[14] Addition of an excess (4:1) of HBF4 of a second ligand in the coordination sphere of the metal
in Et2O (Scheme 4) allows us to observe the corresponding causing the likely displacement of the labile O donor. Since
solvento complex 3a, while the further addition of an excess the above IR data did not provide a clear answer as to
of water leads to the formation of the aquo complex (see which of the complexes was more Lewis acidic (common
below and 3d in Table 1). These observations would be con- chemical sense would indicate 3a) it was decided to turn to
sistent with 4 being both a bis-hydroxo complex (as formu- a different technique.
Table 2. IR data of 2a and 3a after reaction with CO, acetonitrile
or 2,6-dimethylphenyl isocyanide[a]
Complex
ν(CO)
ν(CN)
ν(NC)
(cmϪ1
)
(cmϪ1
)
(cmϪ1
)
2a
2b
2150
2152
2295, 2324
2296, 2326
2204
2209
[a]
In DCE solution; resolution 0.5 cmϪ1
.
Scheme 4
Recently, Zhong et al.[16] have faced a similar problem in
the study of a homologous series of cationic bis(aryl)diim-
ine-ligated methyl complexes of PtII where the substituents
on the aromatic rings were varied systematically. The aim
of these authors was to determine the electronic effects of
the ligands on the metal center and possibly find a corre-
lation with respect to their properties in the CϪH activation
of alkanes. The problem was addressed by studying the
lated in Scheme 4) and also a dinuclear µ-hydroxo complex.
Conclusive evidence for the nature of 4 as a mononuclear
bis-hydroxo complex comes from IR spectroscopy, where
the broad strong BF4Ϫ band is not observed, and from mo-
lar conductivity data. As shown in Table 1, the conductivity
of a 10Ϫ3  solution of 4 in MeOH is practically negligible,
indicating that the complex is neutral.
1
equilibrium shown in Scheme 5 by H NMR spectroscopy,
where water is used as a Lewis base. The extent of this equi-
librium can be taken as a measure of the Lewis acidity of
the metal center. Although this method was developed for
a different class of homologous complexes of PtII, it seems
reasonable to apply it, at least semi-quantitatively, also to
the present case. The ratio between Pt-solv and PtϪOH2
complexes can be easily determined by 31P{1H} NMR spec-
troscopy at Ϫ70 °C {NMR spectroscopic data for
Acidity Measurements
A critical point for the purpose of the present work is
the determination of the Lewis acid characteristics of the
homologous complexes 2a and 3a. This was initially investi-
gated by IR spectroscopy, following addition to solutions
of 2a and 3a of CO, acetonitrile and 2,6-dimethylphenyl
isocyanide, all of which can easily displace the CH2Cl2 from
the coordination sphere of the metal. It is well-known that
both the CO, CN and NC stretching frequencies of the
above coordinated molecules are sensitive to the electron
density of the metal[14] and hence might be a valuable tool
for understanding the Lewis acid nature of the metal center.
Experiments were carried out in 1,2-dichloroethane (DCE).
In the case of CO, solutions of the complexes were satu-
rated with the gas prior to recording the IR spectrum, while
with acetonitrile or 2,6-dimethylphenyl isocyanide variable
amounts of reactant could be added. This allowed us to
distinguish between the free and coordinated ligand.
Table 2 summarizes the results obtained. As can be seen, in
all cases the coordinated ligand vibrates at higher frequenc-
ies with respect to the free ligand (CO 2143 cmϪ1; MeCN
2255 cmϪ1; 2,6-dimethylphenyl isocyanide 2121 cmϪ1),
which is a good indication that the complexes are indeed
Lewis acids. However, the addition of CO or CH3CN re-
sults in negligible differences between the complexes, only
with 2,6-dimethylphenyl isocyanide could a 5 cmϪ1 shift to
higher frequency be observed in the case of 3b. It has to be
pointed out that with the latter complex the addition of an
excess of isocyanide determines a further shift of the band
[Pt(H2O)(P-P-P)](BF4)2
(2d)
and
[Pt(H2O)(P-P-Pϭ
O)](BF4)2 (3d) are reported in Table 1}. By applying mass
balance to Pt and H2O one can easily solve a four-equation
system, which allows us to calculate the equilibrium con-
stant (Scheme 5). This was found to be 110 for 2a and 1400
for 3a indicating that the latter is considerably more Lewis
acidic than the former. Indeed, as the NMR experiments
were carried out in acetone the solvento species are not ex-
actly 2a and 3a, as in Scheme 2, but more correctly the cor-
responding acetone derivatives. The limitations of these
maximum to higher frequencies as a result of the formation Scheme 5
Eur. J. Inorg. Chem. 2004, 1030Ϫ1038
www.eurjic.org
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1033

A. Brunetta, G. Strukul
FULL PAPER
NMR experiments when used for the interpretation of the
catalytic reactions that will be considered below are numer-
ous: (i) the technique was developed for a different class of
complexes, although the principle is quite general; (ii) the
equilibrium constants are determined at Ϫ70 °C, while the
catalysis is carried out at higher temperatures; (iii) the sol-
vento complexes considered are actually Pt(acetone) instead
of Pt(dichloromethane). Even with these limitations we be-
lieve that the overall meaning of the experiment is quite
clear, i.e. 3a is more acidic than 2a.
Catalytic Activity
Having ascertained that 2a and 3a differ significantly in
their Lewis acidity, their activity in the BV oxidation and
in the epoxidation of olefins with hydrogen peroxide was
compared. 2-Methylcyclohexanone and cyclooctene were Figure 3. Reaction profiles for the epoxidation of cyclooctene with
hydrogen peroxide catalyzed by homologous PtII complexes:
chosen as prototype substrates and a substrate/H2O2/cata-
lyst ratio equal to 100:100:1 was chosen. Reactions were
carried out in DCE.
In both processes 2a was found to be inactive at 20 °C,
whereas at 70 °C a moderate activity was observed. In Fig-
ure 3 a comparison between 2a, 2b and 2c in the epoxid-
ation of cyclooctene is reported. As can be seen the hom-
[(triphos)Pt(solv)]2ϩ (2a), diamonds; [(triphos)Pt(OH)]ϩ (2b),
circles; [(triphos)Pt(OOH)]ϩ (2c), squares; numbers on the curves
indicate the rate expressed as a zero-order rate constant; reaction
conditions: catalyst 0.017 mmol; substrate 1.7 mmol; 35% H2O2
1.7 mmol; DCE 3 ml; N2 1 atm.
The activity of 3a is significantly different, although the
ologous complexes display a very similar catalytic behavior: reaction profile has the same shape as that shown for 2a in
their maximum activity (expressed as zero-order rate con- Figure 3. As can be seen from Table 3, this catalyst is al-
stant) is essentially the same: a very short induction time is ready active at 20 °C especially in the oxidation of 2-methyl-
observed with 2b and 2c, indicating a relatively quick con- cyclohexanone, while the activity is very poor in the epoxid-
version into the catalytically active species, and only the ation of cyclooctene. At 70 °C it is approximately as active
amount of epoxide formed seems to differ markedly. The as 2a in the epoxidation reaction and more active in the BV
reactions are in any case very selective as the epoxide was oxidation (Table 3), but seems to deactivate more quickly.
the only product observed, with no evidence for glycol or
The reactivity shown by 2a with respect to cyclooctene is
further oxidation products. It has to be pointed out that observed also in the epoxidation of other olefins (Table 4).
under these experimental conditions a certain degree of ir- Even in this case a similar activity between 2a and 2b is
reproducibility was observed (approximately Ϯ10% of the observed. However, by the end of the epoxidation reactions,
maximum activity and Ϯ25% of the maximum conversion) some epoxide hydrolysis occurs with formation of the corre-
as the result of a deactivation process that will be con- sponding glycol, probably promoted by the acid catalyst
sidered more thoroughly below. Similar results (Table 3) and due to the lower stability of these epoxides with respect
were observed for the oxidation of 2-methylcyclohexanone to cyclooctene epoxide. Internal olefins are epoxidized more
(2-heptalactone was the only oxidation product) both in efficiently than terminal olefins (1-octene) and this con-
terms of activity and conversion. Like in epoxidation, the trasts with the behavior observed for other cationic PtII
reaction must be carried out at 70 °C, and deactivation of complexes {e.g. [Pt(CF3)(P-P)(solv)]ϩ}, which are specific
the catalyst is evident.
for the epoxidation of terminal olefins.[7] The α,β-unsatu-
Table 3. Catalytic activity of complexes 2 and 3 in the epoxidation of cyclooctene and the BaeyerϪVilliger oxidation of methylcyclohexa-
none[a]
Catalyst
Substrate
Temp.
(°C)
105 ϫ max. rate
Max. conv.
(%)
(MsϪ1
)
2a
2b
2c
2a
2b
3a
3a
3a
3a
cyclooctene
70
70
70
70
70
20
70
20
70
1.5
1.4
1.4
1.5
1.6
0.05
1.2
2.0
7.0
20
40
25
11
13
2
11
40
25
2-methylcyclohexanone
cyclooctene
2-methylcyclohexanone
[a]
Experimental conditions: catalyst 0.017 mmol; substrate 1.7 mmol; 35% H2O2 1.7 mmol; DCE 3 ml; N2 1 atm.
www.eurjic.org Eur. J. Inorg. Chem. 2004, 1030Ϫ1038
1034
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Epoxidation versus BaeyerϪVilliger Oxidation
FULL PAPER
Table 4. Catalytic activity of complexes 2 and 3 in the oxidation of other substrates[a]
Catalyst
Substrate
Temp.
(°C)
105 ϫ max. rate
Max. conv.
(%)
(MsϪ1
)
2a
2b
2a
2b
2b
2a
2b
3a
3a
3a
3a
3a
3a
3a
cyclohexene
1-octene
70
70
70
70
70
70
70
20
70
20
70
20
70
20
1.0
1.3
0.3
0.4
1.0
2.0
1.5
0.5
4.5
0.5
6.2
0.2
5
20
20
10
15
10
15
12
5
15
5
11
5
4-methyl-2-pentene
2-cyclohexen-1-one
cyclopentanone
cyclohexanone
3,3-dimethyl-2-butanone
2-allylcyclohexanone
6
11
0.8
[a]
Experimental conditions: catalyst 0.017 mmol; substrate 1.7 mmol; 35% H2O2 1.7 mmol; DCE 3 ml; N2 1 atm.
rated ketone 2-cyclohexen-1-one is selectively epoxidized in lyst to control the selectivity of the reaction?’’ — some com-
agrement with previous findings with PtII complexes.[17]
petition experiments and reactions on bifunctional sub-
Complex 3a is less effective than 2a in the epoxidation of strates were carried out.
olefins, but far more efficient in the BV oxidation of ketones
Initially, 2a and 3a were tested in the oxidation of a 1:1
(Table 4). Notably, the oxidation of open-chain ketones is mixture of cyclooctene and 2-methylcyclohexanone; the re-
also possible, a unique ability among the BV oxidation cata- sults are collected in Table 5. As can be seen, in the case of
lysts known in the literature, that is shared only with [Pt(µ- 3a the reactivity observed seems to be simply the addition
2ϩ
OH)(P-P)]2 complexes.[8a]
of the results observed in the reactions with the individual
substrates. Conversely, the behavior of 2a is somehow sur-
prising and in contrast to the results observed with the indi-
vidual substrates. In fact, some preference for the BV reac-
tion seems to be evident at 70 °C. This trend becomes quite
clear in the oxidation of unsaturated ketones. Scheme 6
summarizes the results observed with dihydrocarvone and
2-allylcyclohexanone. In both cases the exclusive product is
the lactone, independently of the catalyst used. This might
Competition Reactions and the Possible Role of Lewis
Acidity
The overall picture that emerges from an analysis of the
catalytic activity of 2a and 3a is that the latter is far more
active in the BV oxidation, whereas 2a is probably a better
epoxidation catalyst, although in this case the differences
are less evident. This seems to confirm the considerations
made in the Introduction, where the reactions were com-
pared mechanistically and indicate that the BV oxidation is
more Lewis acid demanding as far as the catalyst properties
are concerned.
To clarify this point and possibly answer the main ques-
tion raised — ‘‘can we exploit the Lewis acidity of the cata-
Table 5. Competition experiments in the oxidation of a mixture of
cyclooctene and 2-methylcyclohexanone[a]
Product
Temp. 105 ϫ max. rate Max. conv.
Catalyst
(°C) (MsϪ1
)
(%)
2a
2a
3a
3a
cyclooctene epoxide 20
2-heptalactone
cyclooctene epoxide 70
2-heptalactone
cyclooctene epoxide 20
2-heptalactone
cyclooctene epoxide 70
2-heptalactone
Ϫ
Ϫ
0.8
1.5
0.05
1.5
1.0
7.0
20
20
2
30
10
25
[a]
Experimental conditions: catalyst 0.017 mmol; substrates 0.85
ϩ0.85 mmol; 35% H2O2 1.7 mmol; DCE 3 ml; N2 1 atm.
Scheme 6
Eur. J. Inorg. Chem. 2004, 1030Ϫ1038
www.eurjic.org
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1035

A. Brunetta, G. Strukul
FULL PAPER
be partly due to the moderate activity of 2a observed with other contains two phosphorus and one oxygen donors.
terminal olefins. However, we believe that other factors Spectroscopic characterization indicates that the solvento
might cooperate to explain the results of this competitive derivative of the latter class of complexes (3a) is a stronger
experiment. For example, if the complex is sufficiently Lewis acid than its counterpart 2a, although both are rather
Lewis acidic, even a metal center like PtII becomes more strong. Their use as catalysts in the BV oxidation of ketones
oxophilic and, in the presence of a molecule containing a and in the epoxidation of olefins has shown that the former
CϭC double bond and a CϭO double bond, it becomes reaction is quite Lewis-acid demanding with respect to the
more prone to interact with a simple donor like oxygen and properties of the catalyst. This has allowed the complete
more reluctant to undergo back-bonding to the olefin. In control of the parallel oxidation processes in the oxidation
other words, the discrimination may arise both from the of bifunctional substrates such as dihydrocarvone or 2-allyl-
intrinsic catalytic activity and from the coordinating ability cyclohexanone, where the exclusive formation of the BV
of the metal, both properties being determined by the Lewis oxidation product was observed. The views presented in
acidity of the central metal atom.
this work may probably be generalized and could be of help
in the design of a new generation of selective, more active
catalysts.
Deactivation of the Catalyst
As was mentioned above, a deactivation process is evi-
dent when reactions are carried out at 70 °C. This limits
significantly the reproducibility of the catalytic experiments
and seems to be independent both of the catalyst and of
the oxidation reaction studied. When the metal species are
isolated at the end of a catalytic reaction they can be recog-
nized either as [PtCl(P-P-P)](BF4) (from 2a) or, at least
partly, as [PtCl2(P-P-PϭO)] (from 3a). In the case of 3a
the situation is more complex as some free, fully oxidized
phosphane is also observed in the 31P{1H} NMR spectrum
of the isolated material and therefore some phosphorus-free
Pt species must also be formed. The most likely source of
chloride is the solvent and it was demonstrated that the
same chloro derivatives could be obtained by simply boiling
the starting complexes in DCE, CHCl3 or CH2Cl2. In the
case of 3a the observation that only the chloro complex is
formed is a good indication that the other decomposed
species are most likely generated in the catalytic reactions
by hydrogen peroxide, which is responsible for the full oxi-
dation of the phosphane. These observations imply that
during the catalytic processes there is a parallel reaction
between the catalyst and the solvent that leads to the com-
plete conversion into catalytically inactive chloro deriva-
tives. The inactivity of the chloro complexes was demon-
strated by blank catalytic tests. The moderate stability of
boiling DCE or chloroform and their tendency to form
traces of HCl, which would explain our results, are quite
Experimental Section
Apparatus: IR spectra were recorded on a Nicolet 750 spectropho-
tometer either in nujol mulls using KBr plates or in CH2Cl2 solu-
tion using CaF2 windows. 31P{1H} and H NMR spectra were re-
1
corded on a Bruker AC200 spectrometer, operating in FT mode,
using as external reference 85% H3PO4 and TMS, respectively.
Negative chemical shifts are upfield from the reference. GLC meas-
urements were taken on a HewlettϪPackard 5890A gas chromato-
graph equipped with a FID detector (gas carrier He). Identification
of products was made by GLC by comparison with authentic
samples. Molar conductivity measurements were made on 10Ϫ3
solutions in MeOH with a CDM 83 Radiometer Copenhagen in-
strument, equipped with a CDC 334 immersion cell.

Materials: Solvents were dried and purified according to standard
methods. Substrates were purified by passing through neutral alum-
ina and stored in the dark at low temperature. Hydrogen peroxide
(35% Fluka), triphos (Aldrich), HBF4 (54% in diethyl ether; Ald-
rich) are commercial products and were used without purification.
The following compounds were prepared according to literature
procedures:
[PtCl2(COD)],[18]
[PtMe2(COD)],[19]
[PtMe2-
(triphos)],[9] [PtMe2(triphosPO)].[9] The preparation of new com-
plexes was performed under dry N2 by using Schlenk techniques,
although all of them were found to be air stable once isolated.
Preparation of Complexes: Some of the complexes reported here
were previously observed by in situ NMR spectroscopy, but never
well-known, although we have used DCE as solvent in PtII- isolated and fully characterised.
catalyzed oxidations with hydrogen peroxide for the past
[(triphos)Pt(CH2Cl2)](BF4)2 (2a): Triphos (0.3 g, 0.56 mmol) was
20 years and never observed such problems. The case of
dichloromethane is less straightforward as this solvent boils
at 42 °C and is recognized to be quite stable. We may specu-
late that a possible explanation for the above behavior may
dissolved in dry N2-saturated CH2Cl2 (5 mL) and a solution of
[PtMe2(COD)] (0.19 g, 0.57 mmol) in dry N2-saturated CH2Cl2
(5 mL) was added to this solution. The mixture was stirred for 2
hours under N2 and then 200 µL of 54% HBF4 was slowly added
rely on a direct attack on the CϪCl bond by the strongly to the solution. This was stirred for 20 minutes, during which time
Lewis acidic Pt2ϩ metal centers.
rapid CH4 evolution was observed. Water was added to the solution
to eliminate traces of acid, the two phases were separated and the
organic one was dried with anhydrous Na2CO3. The solvents were
slowly evaporated from the solution in vacuo to give a white-yellow
solid. This was recrystallized from CH2Cl2/Et2O (0.46 g, yield
Conclusions
82%). C35H35B2Cl2F8P3Pt: calcd. C 42.54, H 3.57; found C 42.27,
H 3.39.
In this work we have synthesized two classes of homolo-
gous cationic complexes of PtII containing tridentate li-
gands and bearing only one potentially vacant coordination
site: one class contains three phosphorus donors while the was dissolved in dry N2-saturated CH2Cl2 (10 mL) and a solution
[(triphosPO)Pt(CH2Cl2)](BF4)2 (3a): Triphos (0.80 g, 1.50 mmol)
1036
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
Eur. J. Inorg. Chem. 2004, 1030Ϫ1038

Epoxidation versus BaeyerϪVilliger Oxidation
FULL PAPER
of [PtMe2(COD)] (0.48 g, 1.44 mmol) in dry N2-saturated CH2Cl2 tration of the commercial 35% H2O2 solution was checked iodo-
(5 mL) was added to this solution. The mixture was stirred for 2 metrically prior to use.
hours under N2 and then a 3% solution of H2O2 (5 mL, 2 mmol) The following general procedure was followed: the required amount
was added; after vigorously stirring for one hour the mixture was of catalyst was placed in solid form (0.017 mmol) in the reactor,
filtered. The two phases were separated and the organic phase was
washed with water and dried with anhydrous Na2CO3, filtered and
reduced to a small volume. 500 µL of 54% HBF4 was then slowly
added to the solution. This was stirred for 20 minutes during which
rapid time CH4 evolution was observed. Water was added to the
solution to eliminate traces of acid, the two phases were separated
and the organic one was dried with anhydrous Na2CO3. The sol-
which was evacuated and filled with N2. Purified, N2-saturated ole-
fin/ketone was added under N2 flow, followed by the required
amount of 1,2-dichloroethane (3 mL). After thermostating at the
required temperature for a few minutes, the H2O2 solution in the
appropriate amount was injected through the septum and the time
was started.
All reactions were monitored by GLC by direct injection of samples
vents were slowly evaporated from the solution in vacuo to obtain taken periodically from the reaction mixtures with a microsyringe.
a white-yellow solid. This was recrystallized from CH2Cl2/Et2O Separation of the products was performed on a 25 m HP-5 capillary
(1.15 g, yield 80%). C35H35B2Cl2F8OP3Pt: calcd. C 41.86, H 3.51; column using a flame ionization detector. The maximum rate was
found C 41.57, H 3.42.
expressed as a zero-order rate constant from analysis of conversion
vs. time curves.
[(triphos)Pt(OH)]BF4 (2b): Compound 2a (0.50 g, 0.51 mmol) was
suspended in CDCl3 (15 mL). A solution of NaOH (0.1 ) was
added (15 mL) and the mixture was vigorously stirred for 30 mi-
nutes. The organic phase was separated, washed with water and
dried with anhydrous Na2CO3, filtered and slowly evaporated in
vacuo to obtained a white solid. This was recrystallized from
CH2Cl2/Et2O (0.37 g, yield 87%). C34H34BF4OP3Pt: calcd. C 49.00,
H 4.11; found C 49.17, H 4.03.
[1]
Lewis Acids in Organic Synthesis (Ed.: H. Yamamoto), Wiley-
VCH: Weinheim, 2000.
[2]
[2a] K. Hori, J. Ito, T. Ohta, I. Furukawa, Tetrahedron 1998, 54,
12737. [2b] D. Ferraris, B. Young, T. Dudding, T. Lectka, J. Am.
[2c]
Chem. Soc. 1998, 120, 4548.
A. K. Ghosh, H. Matsuda,
[2d]
Org. Lett. 1999, 1, 2157.
S. Oi, E. Terada, K. Ohuchi, T.
Kato, Y. Tachibana, I. Inoue, J. Org. Chem. 1999, 64, 8660. [2e]
K. Hori, H. Kodama, T. Ohta, I. Furukawa, J. Org. Chem.
[(triphos)Pt(OOH)]BF4 (2c): Compound 2b (0.62 g, 0.74 mmol)
was suspended in CDCl3 (5 mL). A 3.5% solution of H2O2 (1.5 mL,
1.5 mmol) was added and the mixture was vigorously stirred for 30
minutes. The white precipitate obtained was filtered off, washed
with Et2O and recrystallized from CH2Cl2/Et2O (0.35 g, yield 55%).
C34H34BF4O2P3Pt: calcd. C 48.07, H 4.03; found C 48.15, H 4.11.
[2f]
1999, 64, 5017.
K. Pignat, J. Vallotto, F. Pinna, G. Strukul,
Organometallics 2000, 19, 5160. [2g] J. Hao, M. Hatano, K. Mi-
[2h]
kami, Org. Lett. 2000, 2, 4059.
T. Hattori, Y. Suzuki, O.
[2i]
Uesugi, S. Oi, S. Miyana, Chem. Commun. 2000, 73.
A.
Fürstner, D. Voigtländer, W. Schrader, D. Giebel, M. Reetz,
[2j]
Org. Lett. 2001, 3, 417.
N. M. Brunkan, P. S. White, M. R.
Gagne, Organometallics 2002, 21, 1565. [2k] N. M. Brunkan, M.
´
[(triphosPO)Pt(OH)2] (4): Compound 3a (0.30 g, 0.30 mmol) was
dissolved in CH2Cl2 (4 mL) under N2. A 4  solution of NaOH
(3 mL) in water was added and the mixture was stirred for one
hour. The organic phase was separated, washed with water and
dried with anhydrous Na2CO3. The solvents were slowly evapo-
rated from the solution in vacuo to obtain a white solid. This was
redissolved in warm acetone and kept at Ϫ15 °C for 24 hours. The
white solid obtained was washed with diethyl ether and dried
(0.070 g, yield 30%). C34H35O3P3Pt: calcd. C 52.38, H 4.52; found
C 52.12, H 4.33.
´
R. Gagne, Organometallics 2002, 21, 1576.
[3] [3a]
M. Del Todesco Frisone, R. Giovanetti, F. Pinna, G. Stru-
kul, Stud. Surf. Sci. & Catal. 1991, 66, 405. [3b] M. Del Todesco
Frisone, F. Pinna, G. Strukul, Organometallics 1993, 12, 148.
C. Bolm, O. Beckmann, C. Palazzi, Can. J. Chem. 2001, 79,
1593.
C. Bolm, O. Beckmann, A. Cosp, C. Palazzi, Synlett 2001,
1461.
[4]
[5]
[6]
[7]
A. Corma, L. T. Nemeth, M. Renz, S. Valencia, Nature 2001,
412, 423.
[7a] G. Strukul, R. A. Michelin, J. Chem. Soc., Chem. Commun.
[7b]
1984, 1538.
G. Strukul, R. A. Michelin, J. Am. Chem. Soc.
[7c]
Determination of the Acidity of the Complexes. NMR Method: In
these experiments 1a and 2a were generated in an NMR tube from
a known amount of precursor. The use of the isolated complexes
2a and 3a was found to be less reproducible as they contain traces
of water. Typically 4.5 ϫ 10Ϫ5 mol of [PtMe2(triphos)] or [PtMe2-
(triphosPO)] were dissolved in dry [D6]acetone (0.7 mL). A stoi-
chiometric quantity of 54% HBF4 (12 µL, 0.9 ϫ 10Ϫ4 mol) was
added and, after evolution of methane, the 31P{1H} NMR spec-
trum was registered at Ϫ70 °C. Then, increasing volumes of H2O
(1, 1.5, 2, 3 µL) were introduced into the tube, and the 31P{1H}
NMR spectrum at Ϫ70 °C was registered each time. A new species
formed (2d or 3d). By comparing the intensity of the signals at δ ϭ
56 (3a, at Ϫ70 °C the signal shifts to lower fields) and 55 ppm (3d)
and at δ ϭ 51 (2a) and 48.5 ppm (2d) the equilibrium constant
could be calculated as shown in Scheme 5.
1985, 107, 7563.
Strukul, Organometallics 1987, 6, 728.
R. Sinigalia, R. A. Michelin, F. Pinna, G.
[7d]
A. Zanardo, F.
Pinna, R. A. Michelin, G. Strukul, Inorg. Chem. 1988, 27, 1966.
[7e]
A. Zanardo, F. Pinna, R. A. Michelin, G. Strukul, Inorg.
Chem. 1989, 28, 1648.
[8]
[9]
[8a] R. Gavagnin, M. Cataldo, F. Pinna, G. Strukul, Organomet-
[8b]
allics 1998, 17, 661.
C. Paneghetti, R. Gavagnin, F. Pinna,
G. Strukul, Organometallics 1999, 18, 5057.
[9a] G. Annibale, P. Bergamini, V. Bertolasi, E. Bosco, M. Catta-
briga, R. Rossi, Inorg. Chim. Acta 2002, 333, 116. [9b] G. Annib-
ale, P. Bergamini, M. Cattabriga, Inorg. Chim. Acta 2001,
316, 25.
[10] [10a] G. Strukul, R. Ros, R. A. Michelin, Inorg. Chem. 1982, 21,
[10b]
495.
M. T. Atlay, M. Preece, G. Strukul, B. R. James, J.
Chem. Soc., Chem. Commun. 1982, 406.
[11] [11a]
K. Nakamoto, Infrared Spectra of Inorganic and Coordi-
nation Compounds, 2nd edition, Wiley, New York, 1970, p. 115.
[11b]
R. M. Silverstein, F. X. Webster, Spectrometric Identifi-
Catalytic Reactions: These were carried out in a 10 mL round-bot-
tomed flask equipped with a stopcock for vacuum/N2 operations
and a sidearm fitted with a screw-capped silicone septum to allow
sampling. Stirring was performed by a Teflon-coated bar driven
externally by a magnetic stirrer. Reactions at 70 °C were carried
out using an oil bath connected with a thermostat. The concen-
cation of Organic Compounds, 6th ed., Wiley, New York, 1998,
p. 109.
[12]
[13]
T. G. Appleton, M. A. Bennett, Inorg. Chem. 1978, 17, 738.
W. J. Geary, Coord. Chem. Rev. 1971, 7, 81.
[14] [14a]
G. W. Bushnell, K. R. Dixon, R. G. Hunter, J. J. McFar-
land, Can. J. Chem. 1972, 50, 3694. [14b] B. Longato, B. Corain,
Eur. J. Inorg. Chem. 2004, 1030Ϫ1038
www.eurjic.org
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1037

A. Brunetta, G. Strukul
FULL PAPER
[17]
[18]
[19]
G. M. Bonora, G. Pilloni, Inorg. Chim. Acta 1987, 137, 75. [14c]
C. Baccin, A. Gusso, F. Pinna, G. Strukul, Organometallics
1995, 14, 1161.
J. X. McDermott, J. F. White, G. M. Whitesides, J. Am. Chem.
Soc. 1976, 98, 6521.
R. Bassan, K. H. Bryars, L. Judd, A. W. G. Platt, P. G. Pringle,
Inorg. Chim. Acta 1986, 121, L41.
Received August 20, 2003
A. L. Bandini, G. Banditelli, F. Demartin, M. Manassero, G.
Minghetti, Gazz. Chim. Ital. 1993, 23, 417.
[14d]
J. J. Li, P. R.
Sharp, Inorg. Chem. 1996, 35, 604.
[15]
[16]
J. P. Collman, L. S. Hegedus, J. R. Norton, R. G. Finke, Prin-
ciples and Applications of Transition Metal Chemistry, Univer-
sity Science Books, Mill Valley, 1987, p. 110.
H. A. Zhong, J. A. Labinger, J. E. Bercaw, J. Am. Chem. Soc.
2002, 124, 1378.
Early View Article
Published Online February 3, 2004
1038
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
Eur. J. Inorg. Chem. 2004, 1030Ϫ1038