Diphosphine-palladium and -platinum complexes as catalysts for the Baeyer-Villiger oxidation of ketones: Effect of the diphosphine, oxidation of acyclic ketones, and mechanistic studies

Article abstract of DOI:10.1021/om970756f

A variety of Pd and Pt complexes of the type [(P-P)M(μ-OH)]₂²⁺ (M = Pd, Pt; P-P = a series of tetraphenyldiphosphines) were tested in the Baeyer-Villiger oxidation of ketones with hydrogen peroxide. The effect of the diphosphine-metal ring size on the catalytic activity indicates that the larger the ring, the better the catalyst and that, in general, Pt complexes are superior. The complex modified with P-P = dppb is the most active catalyst and allows for the first time the oxidation of a series of acyclic ketones. The corresponding migratory aptitude series is in full agreement with the one known for the stoichiometric organic reaction employing peracids as oxidants. A test of the reactivity of different peroxidic oxidants (H₂O₂, t-BuOOH, KHSO₅, carbamide peroxide) shows that hydrogen peroxide is the most effective. A kinetic study of the oxidation of 2-methylcyclohexanone with [(dppb)-Pt(μ-OH)]₂²⁺ as the catalyst shows typical half-order dependence on the catalyst concentration, suggesting that the hydroxy dimer opens up to form the catalytically active species. The reaction is first order in ketone and hydrogen peroxide and is independent of the acidity of the system. The reaction is suggested to proceed via a quasi-peroxymetallacyclic intermediate and bears strong similarities to the stoichiometric organic reaction.

Full text of DOI:10.1021/om970756f

Organometallics 1998, 17, 661-667
661
Dip h osp h in e-P a lla d iu m a n d -P la tin u m Com p lexes a s
Ca ta lysts for th e Ba eyer -Villiger Oxid a tion of Keton es:
Effect of th e Dip h osp h in e, Oxid a tion of Acyclic Keton es,
a n d Mech a n istic Stu d ies
Roberta Gavagnin, Maurizio Cataldo, Francesco Pinna, and Giorgio Strukul*
Department of Chemistry, University of Venice, Dorsoduro 2137, 30123 Venice, Italy
Received August 25, 1997
2+
A variety of Pd and Pt complexes of the type [(P-P)M(µ-OH)]2 (M ) Pd, Pt; P-P ) a
series of tetraphenyldiphosphines) were tested in the Baeyer-Villiger oxidation of ketones
with hydrogen peroxide. The effect of the diphosphine-metal ring size on the catalytic
activity indicates that the larger the ring, the better the catalyst and that, in general, Pt
complexes are superior. The complex modified with P-P ) dppb is the most active catalyst
and allows for the first time the oxidation of a series of acyclic ketones. The corresponding
migratory aptitude series is in full agreement with the one known for the stoichiometric
organic reaction employing peracids as oxidants. A test of the reactivity of different peroxidic
oxidants (H2O2, t-BuOOH, KHSO5, carbamide peroxide) shows that hydrogen peroxide is
the most effective. A kinetic study of the oxidation of 2-methylcyclohexanone with [(dppb)-
Pt(µ-OH)]22+ as the catalyst shows typical half-order dependence on the catalyst concentra-
tion, suggesting that the hydroxy dimer opens up to form the catalytically active species.
The reaction is first order in ketone and hydrogen peroxide and is independent of the acidity
of the system. The reaction is suggested to proceed via a quasi-peroxymetallacyclic
intermediate and bears strong similarities to the stoichiometric organic reaction.
In tr od u ction
(IV)5 have appeared, in addition to the so-called Mu-
kaiyama system,6 the latter making use of dioxygen as
the primary oxidant in the presence of a sacrificial
aldehyde. It was also possible to perform the reaction
enantioselectively7 with moderate to good ee’s using
complexes modified with chiral ligands. From a syn-
thetic point of view, the use of chiral transition-metal
catalysts appears to be the only promising alternative
to catalysis by microorganisms for the achievement of
esters with (moderately) high optical purity and there-
fore is of foremost importance.
The Baeyer-Villiger oxidation of ketones with peroxy
acids is a reaction of significant synthetic interest that
finds a variety of possible applications, e.g., the syn-
thesis of antibiotics and steroids, pheromones, mono-
mers for polymerization, etc. The whole subject has
been reviewed several times.1 The versatility of this
reaction in synthetic organic chemistry is evident from
the most recent of these reviews,1g which includes more
than 1000 references.
Attempts to apply transition-metal catalysts to this
important organic transformation have met with only
limited success.2 Catalytic methods, compared to their
stoichiometric counterparts, may offer advantages such
as the simplification of the operation conditions, the use
of cheaper reactants, or less waste product formation.
We recently reported the first unambiguous example
of transition-metal catalysis applied to the Baeyer-
Villiger oxidation. It was suggested that, to be effective,
the catalyst must be able to increase the nucleophilicity
of the oxidant (H2O2) for attack at the electrophilic
ketone center. More recently, other examples using the
same primary oxidant and based on Re(VII)4 and Ti-
However, the scope of these promising catalytic
systems is too limited to constitute a valid practical
alternative to the use of stoichiometric amounts of
organic peroxy acids. For example, they have proved
so far unable to promote the oxidation of acyclic ketones.
Following a preliminary account,8 we report herein the
use of a class of diphosphine hydroxy complexes of Pd-
(3) (a) Del Todesco Frisone, M.; Pinna, F.; Strukul, G. Stud. Surf.
Sci. Catal. 1991, 66, 405. (b) Del Todesco Frisone, M.; Pinna, F.;
Strukul, G. Organometallics 1993, 12, 148.
(4) Herrmann, W. A.; Fischer, R. W.; Correia, J . D. G. J . Mol. Catal.
1994, 94, 213.
(5) Bhaumik, A.; Kumar, P.; Kumar, R. Catal. Lett. 1996, 40, 47.
(6) (a) Murahashi, S.-I., Oda, Y.; Naota, T. Tetrahedron Lett. 1992,
33, 7557. (b) Murahashi, S.-I. Angew. Chem., Int. Ed. Engl. 1995, 34,
2443. (c) Bolm, C.; Schlingloff, G.; Weickardt, K. Tetrahedron Lett.
1993, 34, 3408. (d) Hamamoto, M.; Nakayama, K.; Nishiyama, Y.; Ishii,
Y. J . Org. Chem. 1993, 58, 6421. (e) Giannandrea, R.; Mastrorilli, P.;
Nobile, C. F.; Suranna, G. P. J . Mol. Catal. 1994, 94, 27.
(7) (a) Gusso, A.; Baccin, C.; Pinna, F.; Strukul, G. Organometallics
1994, 13, 3442. (b) Bolm, C.; Schlingloff, G.; Weickhardt, K. Angew.
Chem., Int. Ed. Engl. 1994, 33, 1848. (c) Bolm, C.; Schlingloff, G. J .
Chem. Soc., Chem. Commun. 1995, 1247. (d) Bolm, C.; Schlingloff, G.;
Bienewald, F. J . Mol. Catal. 1997, 117, 347. (e) Lopp, M.; Paju, A.;
Kanger, T.; Pehk, T. Tetrahedron Lett. 1996, 37, 7583.
(1) (a) Hassall, C. H. Org. React. 1957, 9, 73. (b) Leffler, J . E. Chem.
Rev. 1949, 45, 385. (c) Lee, J . B.; Uff, B. C. Q. Rev. (London) 1967, 21,
429. (d) House, H. O. Modern Synthetic Reactions; Benjamin: New
York, 1972; p 327. (e) Plesnicar, B. In Oxidation in Organic Chemistry;
Trahanovsky, W. S., Ed.; Academic Press: New York, 1978; Part C, p
254. (f) Hudlicky, M. Oxidations in Organic Chemistry; American
Chemical Society: Washington, DC, 1990; p 186. (g) Krow, G. C. Org.
React. 1993, 43, 251.
(2) Strukul, G. Angew. Chem., Int. Ed. Engl., in press.
S0276-7333(97)00756-5 CCC: $15.00 © 1998 American Chemical Society
Publication on Web 01/28/1998

662 Organometallics, Vol. 17, No. 4, 1998
Gavagnin et al.
Ch a r t 1a
a
Legend: dppm ) bis(diphenylphosphino)methane; dppe )
1,2-bis(diphenylphosphino)ethane; dppp ) 1,3-bis(diphenylphos-
phino)propane; dppb ) 1,4-bis(diphenylphosphino)butane.
Sch em e 1
2+
F igu r e 1. Effect of the diphosphine in [(P-P)Pd(µ-OH)]2
complexes as catalysts for the Baeyer-Villiger oxidation
of 2-methylcyclohexanone. Reaction conditions: Pd, 0.017
mmol; ketone, 1.7 mmol; H2O2, 1.7 mmol; DCE, 3.19 mL;
T, 25 °C.
(II) and Pt(II) (Chart 1) as catalysts for the Baeyer-
Villiger oxidation of ketones using hydrogen peroxide
as the primary oxidant. As is indicated in Chart 1, the
size of the diphosphine-metal ring was systematically
changed in order to optimize the catalytic properties of
the complexes. This has led to the discovery of a
catalyst capable of promoting the oxidation of acyclic
ketones. The essential mechanistic features of the
reaction using this class of catalysts have been inves-
tigated through a kinetic analysis.
Resu lts a n d Discu ssion
Syn th esis a n d Ch a r a cter iza tion . The complexes
shown in Chart 1 have been prepared by following the
general procedure outlined some 25 years ago by Bush-
nell et al.9 for [(PPh3)2Pt(µ-OH)]2 (BF4)2. The procedure
involves (Scheme 1) the introduction of the diphosphine
into the coordination sphere of the metal using a
suitable starting material (trans-(MeCN)2PdCl2 or (COD)-
PtCl2). Chloride abstraction from (P-P)MCl2 (M ) Pd,
Pt) is then performed using 2 equiv of AgBF4 in CH2-
Cl2/acetone solution. The chlorides are initially replaced
by water, present as an impurity in reagent grade
acetone, and upon addition of diethyl ether, the bridging
hydroxy dimers precipitate out of solution.10
2+
F igu r e 2. Effect of the diphosphine in [(P-P)Pt(µ-OH)]2
complexes as catalysts for the Baeyer-Villiger oxidation
of 2-methylcyclohexanone. Reaction conditions: Pt, 0.017
mmol; ketone, 1.7 mmol; H2O2, 1.7 mmol; DCE, 3.19 mL;
T, 25 °C.
a singlet at δ 37.1 ppm. Additionally, the molar
conductivity determined for a 10-3 M solution in MeOH
gave 207 Ω-1 mol-1 cm2, typical for a 2/1 electrolyte.11
Effect of th e Dip h osp h in e. The effect of the
diphosphine-metal ring size was tested by varying the
length of the hydrocarbon chain linking together the
phosphorus donors, as is indicated in Chart 1. The
purpose of this was to examine the steric factors that
may influence the reactivity of the complex, in analogy
with previous results obtained with [(P-P)Pt(CF3)(CH2-
Cl2)]+ (P-P ) a variety of diphosphines) in the epoxi-
dation of 1-octene.12 In this latter case, it was found
that an increase in the bite angle of the diphosphine
caused a decrease in the catalytic activity of up to 2
orders of magnitude, which correlated with the space
available for coordination of the olefin.12
All Pt derivatives and most of the Pd derivatives have
been reported previously (see Experimental Section).
Only (dppb)PdCl2 and [(dppb)Pd(µ-OH)]2 (BF4)2 (1d ) are
new compounds, and they have been characterized by
IR, NMR, and (only for 1d ) molar conductivity data. The
IR spectrum of the complex (dppb)PdCl2 shows typical
Pd-Cl stretchings at 289 and 306 cm-1, while the 31P-
{1H} NMR spectrum shows a singlet at δ 29.8 ppm (free
dppb at δ -16.2 ppm). The IR spectrum of 1d shows a
medium-intensity O-H stretching band at 3593 cm-1
,
The test substrate chosen was 2-methylcyclohex-
anone, and the reaction conditions are reported in
Figures 1 and 2, which summarize the results obtained.
In all cases the selectivity of the reaction was >99%, as
ꢀ-heptalactone was the only reaction product observed.
As shown, in the case of both Pd and Pt complexes there
and a strong broad band between 995 and 1100 cm-1
typical of BF4-, while the 31P{1H} NMR spectrum shows
(8) Strukul, G.; Varagnolo, A.; Pinna, F. J . Mol. Catal. 1997, 117,
413.
(9) Bushnell, G. W.; Dixon, K. R.; Hunter, R. G.; McFarland, J . J .
Can. J . Chem. 1972, 50, 3694.
(10) (a) Pisano, C.; Consiglio, G.; Sironi, A.; Moret, M. J . Chem. Soc.,
Chem. Commun. 1991, 421. (b) Sperrle, M.; Gramlich, V.; Consiglio,
G. Organometallics 1996, 15, 5196.
(11) Geary, W. J . Coord. Chem. Rev. 1971, 7, 81.
(12) Zanardo, A.; Michelin, R. A.; Pinna, F.; Strukul, G. Inorg. Chem.
1989, 28, 1648.

Diphosphine-Palladium and -Platinum Complexes
Organometallics, Vol. 17, No. 4, 1998 663
is a dependence on the nature of the diphosphine and
the reactivity follows the order dppm < dppe < dppp <
dppb. The catalytic activity is also dependent on the
metal (Pd or Pt).
The Pd complexes are less active, as the amount of
ꢀ-heptalactone obtained is much lower, and the catalyst
is able to perform just a few turnovers. The reaction
solution, initially pale yellow, changed to orange within
about 1 h and eventually darkened with concomitant
precipitation of a brown solid, most likely Pd metal.
Attempts to extend the catalyst lifetime using more
coordinating solvents such as THF and EtOH were
unsuccessful, in part because of solvent oxidation. In
contrast, the Pt complexes were stable under the condi-
tions used and turnovers as high as 55 were obtained.
The reactivity sequence found for the different diphos-
phines is opposite to that previously observed for the
epoxidation reaction12 and hence must be related to a
different steric effect. It is possible that breakdown of
the hydroxy bridge in the coordinatively saturated
dimeric starting species to yield coordinatively unsatur-
ated (P-P)Pt(OH) monomeric fragments is the prelimi-
nary step for catalytic activity. We tried to find some
possible structural indications of this by analyzing
crystallographic data found in the literature for some
of the complexes used in this study. X-ray structural
analyses of (µ-hydroxy)platinum complexes containing
dppm,13 dppp,14 and dppb13 have been published re-
cently. Except for the expected increase in the P-Pt-P
bite angle, which was observed also for the monomeric
palladium diphosphine dichlorides,15 no clear steric
factor involving the hydroxy bridge could be found along
the series. For example, while in the case of the dppm
and dppb derivatives the two platinum coordination
planes are coplanar,13 in the case of the dppp derivative
the two coordination planes form a dihedral angle of
about 40°.14
Effect of th e Oxid a n t. The reactivity of other
peroxidic oxidants in the oxidation of 2-methylcyclohex-
anone was checked using the most active catalyst (2d ).
The oxidants chosen were hydrogen peroxide, potassium
caroate (KHSO5), tert-butyl hydroperoxide, and carba-
mide peroxide (CP), i.e., the adduct between urea and
hydrogen peroxide. The reaction was tested at room
temperature, under the usual reaction conditions, and
the results are summarized in Figure 3. As can be seen,
H2O2 is the most effective oxidant, while t-BuOOH is a
very poor one. Even after several days the maximum
conversion observed was less than 1%. Moderate oxida-
tion properties are displayed by KHSO5, while in the
case of CP, despite a good initial activity, the maximum
conversion never exceeded 10% (CP is sparingly soluble
in DCE, the solvent chosen for the comparison test).
Oxid a tion of Acyclic Keton es. Given the good
catalytic properties of 2d in combination with H2O2 as
the oxidant, we have tested the reactivity for the
oxidation of a series of simple acyclic ketones (Chart 2).
These are all methyl ketones in which the second group
was chosen in order to determine a possible order of
migratory aptitude, in analogy to what is known for the
organic reaction.1
F igu r e 3. Effect of the oxidant in the Baeyer-Villiger
oxidation of 2-methylcyclohexanone using [(dppb)Pt(µ-
2+
OH)]2 as catalyst: (circles) t-BuOOH; (diamonds) CP;
(triangles) KHSO5; (squares) H2O2. Reaction conditions: Pt,
0.017 mmol; ketone, 1.7 mmol; oxidant, 1.7 mmol; DCE as
solvent; T, 25 °C.
F igu r e 4. Oxidation of substituted methyl ketones (see
Chart 2) catalyzed by [(dppb)Pt(µ-OH)]22+. Reaction con-
ditions: Pt, 0.017 mmol; ketone, 1.7 mmol; H2O2, 1.7 mmol;
DCE as solvent; T, 25 °C.
Ch a r t 2
Reactions were performed at room temperature with
the exception of the oxidation of acetophenone, which
was observed only at 58 °C. A summary of the results
is reported in Figure 4. No activity was observed in the
case of 2-hexanone even after 24 h at 70 °C. In general,
the order of migratory aptitude is
tertiary alkyl > vinyl > secondary alkyl > phenyl >
primary alkyl
This is in agreement with the series found for the
organic reaction.1 The results here reported represent
the first observation of the Baeyer-Villiger oxidation
of an acyclic ketone under catalytic conditions using a
transition-metal complex.
(13) Li, J . J .; Sharp, P. R. Inorg. Chem. 1996, 35, 604.
(14) Bandini, A. L.; Banditelli, G.; Demartin, F.; Manassero, M.;
Minghetti, G. Gazz. Chim. Ital. 1993, 123, 417.
The oxidation of methyl vinyl ketone represents also
(15) Steffen, W. L.; Palenik, G. J . Inorg. Chem. 1976, 15, 2432.
an interesting example of chemoselectivity. It is known1

664 Organometallics, Vol. 17, No. 4, 1998
Gavagnin et al.
Sch em e 2
Ta ble 1. Su m m a r y of th e Kin etic Da ta for th e
Oxid a tion of 2-Meth ylcycloh exa n on e w ith
Hyd r ogen P er oxid e Ca ta lyzed by
a
[(d p p b)P t(µ-OH)]2(BF 4)2
103[Pt2]
(M)
[ketone]
(M)
103[H2O2(org)]
(M)
105 × init rate
(M s-1
)
0.10
0.50
1.0
2.5
5.0
0.50
0.50
0.50
0.50
0.50
0.50
0.50
0.50
0.50
0.50
0.50
0.50
0.50
0.50
0.10
0.20
1.0
53
53
53
53
53
53
53
53
53
3.31
8.98
20.2
82
0.48
1.01
1.30
2.13
2.49
0.20
0.39
2.23
3.57
0.06
0.13
0.28
1.96
2.82
F igu r e 5. Kinetic analysis of the oxidation of 2-methyl-
cyclohexanone with hydrogen peroxide using [(dppb)Pt(µ-
2+
OH)]2 as catalyst: effect of catalyst concentration.
2.0
0.50
0.50
0.50
0.50
0.50
140
a
Reactions performed in DCE at 10 °C.
that with unsaturated ketones the Baeyer-Villiger
oxidation is accompanied by some epoxidation of the
carbon-carbon double bond. A stoichiometric reaction
using m-chloroperbenzoic acid (MCPBA) as the oxidant,
under conditions similar to those reported for the
catalytic reaction in Figure 4, indicates that the unsat-
urated ester forms initially, followed by the epoxy ketone
and finally by the epoxy ester (Scheme 2). In contrast,
the catalytic reaction with hydrogen peroxide and 2d
yields only the unsaturated ester. This result is at
variance with some previous findings obtained with
other Pt(II) complexes of the type [(P-P)Pt(CF3)(CH2-
Cl2)]+ or (P-P)Pt(2-van) + HClO4 (P-P ) various
diphosphines; 2-van ) the dianion of 2-vanillin), with
which a variety of R,â-unsaturated ketones could be
conveniently epoxidized.16
F igu r e 6. Kinetic analysis of the oxidation of 2-methyl-
cyclohexanone with hydrogen peroxide using [(dppb)Pt(µ-
OH)]22+ as catalyst: effect of ketone (circles) and hydrogen
peroxide (squares) concentrations.
between the initial rate and the square root of the Pt
dimer concentration. This observation is an indication
that the coordinatively saturated bridging hydroxy
dimer opens up to yield a catalytically active monomeric
species. A first-order dependence is observed for both
the ketone and hydrogen peroxide concentrations (Fig-
ure 6).
Kin etic Stu d ies. A kinetic study was performed by
analyzing the initial rate of reaction for the oxidation
of 2-methylcyclohexanone with hydrogen peroxide using
2d as catalyst. The reaction was performed in a two-
phase DCE/H2O solvent system at 10 °C, a temperature
that seems to constitute a good compromise between
activity and the accuracy of rate data.
A summary of the kinetic data is reported in Table 1.
Since the reaction occurs in the organic phase, the
concentration of H2O2 in DCE was calculated by deter-
mining the partition curve of H2O2 between H2O and
DCE using aqueous solutions ranging from 1.0 to 20 M.
The effect of the individual reactants on the reaction
rate is shown in Figures 5 and 6.
Mech a n ism . A simple mechanism that accounts for
the above experimental observations is shown in Scheme
3. A fast dissociation equilibrium in which the ketone
and hydrogen peroxide are actively involved is sug-
gested as the initial step. This is followed by the rate-
determining intramolecular nucleophilic attack in 3d
between the coordinated hydroperoxy ligand and the
carbonyl carbon, the electrophilic character of which is
increased by coordination. This step is described as a
two-stage process leading to the formation of a quasi-
peroxymetallacyclic intermediate (4d ), in which a simple
rearrangement yields the lactone and a monomeric,
coordinatively unsaturated Pt+-OH species that quickly
dimerizes to form the starting complex.
The effect of Pt concentration (Figure 5) is of frac-
tional order, and a linear relationship is observed
A scheme of this type yields the following mathemati-
cal expression for the reaction rate:
(16) Baccin, C.; Gusso, A.; Pinna, P.; Strukul, G. Organometallics
1995, 14, 1161.
d[lactone]/dt ) kK1/2[ketone][H2O2][Pt2(OH)2]1/2

Diphosphine-Palladium and -Platinum Complexes
Organometallics, Vol. 17, No. 4, 1998 665
Sch em e 3
Effect of Acid ity. Since commercial 35% hydrogen
peroxide is stabilized with acid (apparent pH 2.0), the
possible active role of H+ in the catalysis has been
considered. It is known10,14,19 that bridging hydroxy
complexes of Pt(II) can be protonated to give the
corresponding aquo complexes, and this would modify
the catalytic cycle according to Scheme 4.
The effect on the initial rate of formation of ꢀ-hepta-
lactone in the catalytic system, following the addition
of different amounts of either perchloric acid or potas-
sium hydroxide to the aqueous phase, was studied. The
addition of OH- or H+ in excess (up to 10 times) with
respect to the catalyst showed no effect on the reaction
rate. This seems to be in agreement with Scheme 4,
where the equilibrium characterized by K1 (the only one
considered in Scheme 3) is mathematically equivalent
to the sum of the two equilibria characterized by K2 and
K3. From a kinetic point of view, the rate law will hold,
again provided that the mass balance with respect to
the different platinum species does not change.
Sch em e 4
The step by step 31P{1H} NMR experiment reported
above was repeated by dissolving 2d in CD2Cl2 (step 1)
and then adding a 10-fold excess of HClO4 (step 2) and
finally a large excess of 2-methylcyclohexanone and
H2O2 (step 3). The results are summarized in Figure
7, where spectra A-C correspond to steps 1-3, respec-
tively. As can be seen, the addition of an excess of acid
to 2d (spectrum A) produces a new species (spectrum
B) that was identified as [(dppb)Pt(OH2)2]2+ by com-
parison with an authentic sample prepared by following
the method reported by Bandini et al.14 for [(dppe)Pt-
(OH2)2]2+ (see Experimental Section). The addition of
the ketone and hydrogen peroxide (spectrum C) restores
the initial compound 2d and produces a small amount
of a new species showing two nonequivalent phosphorus
We have tried to determine the mass balance for Pt
through 31P{1H} NMR experiments. The catalyst 2d
(40 mg) was dissolved in CD2Cl2, and the spectrum was
run, showing a singlet at δ 5.1 ppm flanked by Pt
1
satellites exhibiting a J P-Pt coupling constant of 3546
Hz. Addition of 200 µL of 2-methylcyclohexanone did
not change the spectrum nor did the further addition
of 100 µL of hydrogen peroxide. A second experiment
in which the order of addition of the two reactants was
reversed gave the same results, indicating that the only
platinum species macroscopically present in the system
is the starting compound, the concentration of which is
just the initial 2d concentration. Under these circum-
stances, the above expression also becomes the rate law
and is in full agreement with the experimental observa-
tions reported in Figures 5 and 6.
Scheme 3 is a simplified view of the actual mecha-
nism. The initial equilibrium might be a fast, kineti-
cally indistinguishable, stepwise process, in which the
hydroxy bridge is initially cleaved by the weak acid
H2O2, as has been observed by ourselves8 and others17
for other weak acids, followed by coordination of the
ketone to the vacant coordination site of the monomeric
hydroperoxyplatinum species. An alternative view, still
kinetically indistinguishable, might be the initial cleav-
age of the hydroxy bridge by the ketone, followed by
external nucleophilic attack of free hydrogen peroxide
on the resulting Pt(OH)(ketone) species, as suggested
for the Baeyer-Villiger oxidation of ketones with Pt-
(CF3) derivatives.3b However, this posibility seems
unlikely because of the ease with which diphosphine
hydroxyplatinum complexes undergo exchange with
hydrogen peroxide, a pathway that has been exploited
for the preparation of stable hydroperoxy complexes.18
2
1
nuclei (δP1 3.7 ppm, J P-P 28.5 Hz, J P-Pt 3181 Hz; δP2
2
1
2.4 ppm, J P-P 28.5 Hz, J P-Pt 3887 Hz). The presence
of this species does not change significantly the plati-
num mass balance and seems to support the mechanism
proposed, as it can be consistent with the existence of
either 3d or 4d on the basis of the P-Pt coupling
constants, with P1 being trans to a σ-bonded oxygen and
P2 trans to a neutral oxygen donor.
In conclusion, the interpretation given above for the
effect of H+ in the system seems correct and the possible
involvement of the additional equilibria shown in Scheme
4 appears to have no influence on the course of the
catalytic reaction.
Na tu r e of th e Oxygen Tr a n sfer Step . As is shown
in Scheme 3, the oxygen transfer is suggested to occur
via a two-stage process in which the formation of a
quasi-peroxymetallacyclic species (4d ) is involved. Al-
though species of this type have never been isolated,
since the early report of Mimoun20 they have been often
implied in the oxidation of unsaturated substrates,21
(19) Bandini, A. L.; Banditelli, G.; Cinellu, M. A.; Sanna, G.;
Minghetti, G.; Demartin, F.; Manassero, M. Inorg. Chem. 1989, 28,
404.
(20) Mimoun, H.; Charpentier, R.; Mitschler, A.; Fischer, J .; Weiss,
R. J . Am. Chem. Soc. 1980, 102, 1047.
(21) See for example: (a) Mimoun, H. In Comprehensive Coordina-
tion Chemistry; Wilkinson, G., Gillard, R. D., McCleverty, J ., Eds.;
Pergamon: Oxford, U.K., 1987; Chapter 61.3, p 317. (b) J ørgensen, K.
A. Chem. Rev. 1989, 89, 431. (c) Strukul, G. In Catalytic Oxidations
with Hydrogen Peroxide as Oxidant; Strukul, G., Ed.; Kluwer: Dor-
drecht, The Netherlands, 1992; Chapter 6, p 177.
(17) Longato, B.; Corain, B.; Bonora, G. M.; Pilloni, G. Inorg. Chim.
Acta 1987, 137, 75.
(18) Strukul, G.; Ros, R.; Michelin, R. A. Inorg. Chem. 1982, 21, 495.
(b) Zanardo, A.; Michelin, R. A.; Pinna, F.; Strukul, G. Inorg. Chem.
1989, 28, 1648.

666 Organometallics, Vol. 17, No. 4, 1998
Gavagnin et al.
F igu r e 7. 31P{1H} NMR experiment in CD2Cl2: (A) [(dppb)Pt(µ-OH)]22+; (B) excess HClO4 added; (C) excess 2-methyl-
cyclohexanone and excess H2O2 added. The inset shows the satellite region of spectrum C.
Sch em e 5
of hydrogen peroxide (a very poor oxidant in noncata-
lyzed reactions) and (ii) it provides a facile pathway for
loss of HO- from the intermediate. The latter in
particular seems crucial, as the poor reactivity of
hydrogen peroxide in this reaction is generally related
to HO- being a poor leaving group. This argument,
supported by the experimental evidence that the order
of migratory aptitude is the same as in the organic
reaction, seems to suggest that in the present Pt-
catalyzed reaction the actual rate-determining step is
most likely the rearrangement of the quasi-peroxyplati-
nacyclic intermediate species.
Con clu sion s
including ketones.3 Clearly, the simple kinetic analysis
reported above does not distinguish between a single-
stage process and a two-stage process, as the rate
equation would be the same in both cases. Although
not required by the rate law, the suggestion that the
rate-determining step is a two-stage process has the
advantage of being similar to the well-known mecha-
nism of the stoichiometric Baeyer-Villiger oxidation
using peracids.1 It is known that this reaction proceeds
(Scheme 5) through the formation of a peroxidic inter-
mediate (the so-called “Criegee intermediate”22), from
which the ester product forms via a simple rearrange-
ment. Of the two stages, the second is generally
considered to be rate determining,1e,g as this is in
agreement with both the order of reactivity of different
peracids (which parallels the properties of X-COO- as
a leaving group) and the order of migratory aptitude
for different alkyl (aryl) groups.
The comparison between the Pd and Pt homologous
species shows that the latter are more active and stable
in the Baeyer-Villiger oxidation of ketones. Although
an analogous result had been previously observed in the
epoxidation of olefins,21c it is still somehow surprising
because it is contrary to the commonly observed superior
catalytic properties of Pd, particularly in oxidation
reactions.
The systematic change of the size of the diphosphine-
metal ring led to catalyst 2d which, for the first time,
is capable of oxidizing open-chain ketones. Although
the structural reasons are not understood, this repre-
sents a useful application of transition-metal catalysis
to this important organic transformation.
The order of migratory aptitude for different alkyl/
aryl groups in methyl ketones is in agreement with the
organic transformation using peracids.
The mechanistic study points to the following impor-
tant factors in the catalysis: (i) the ability of the
complex to increase both the electrophilicity of the
carbonyl carbon and the nucleophilicity of hydrogen
peroxide and (ii) the assistance given by the metal
center to dissociation of HO- from the peroxymetalla-
cyclic intermediate, thus making a typically “green”
reagent like hydrogen peroxide a valuable oxidant for
the reaction.
As can be seen, the quasi-peroxyplatinacyclic inter-
mediate bears strong similarities to the Criegee inter-
mediate (with Pt+ instead of H+). However, the role of
the platinum complex in the overall reaction is not
simply that of a Lewis acid, as it plays two other
important functions: (i) it increases the nucleophilicity
(22) (a) Criegee, R. J ustus Liebigs Ann. Chem. 1948, 560, 127. (b)
Lee, J . B.; Uff, B. C. Q. Rev. 1967, 21, 431.

Diphosphine-Palladium and -Platinum Complexes
Organometallics, Vol. 17, No. 4, 1998 667
benzene and evaporated to dryness a few times until a white
solid was obtained. This was suspended in Et2O, filtered,
washed with Et2O, and dried in vacuo (yield 75%).
Exp er im en ta l Section
Ap p a r a tu s. IR spectra were taken on a Nicolet FTIR
Magna 750 and on a Digilab FTS 40 interferometers either in
solid (KBr pellets) or in CH2Cl2 solution using CaF2 windows.
1H and 31P{1H} NMR spectra were recorded on a Bruker AC
200 spectrometer operating in FT mode, using as external
references TMS and 85% H3PO4, respectively. Negative
chemical shifts are upfield from the reference. UV-vis spectra
were recorded on a Perkin-Elmer Lambda 5 spectrophotom-
eter. Conductivity measurements were performed on a Ra-
diometer instrument using 10-3 M solutions in MeOH at 25
°C. GC measurements were taken on a Hewlett-Packard
5790A gas chromatograph equipped with a 3390 automatic
integrator. GC-MS measurements were performed on a
Hewlett-Packard 5971 mass selective detector connected to a
Hewlett-Packard 5890 II gas chromatograph. Identification
of products was made with GC or GC-MS by comparison with
authentic samples.
Ma ter ia ls. Solvents were dried and purified according to
standard methods. Ketone substrates were purified by passing
through neutral alumina, prior to use. Hydrogen peroxide
(35% from Acros, 60% from Degussa), m-chloroperbenzoic acid
(MCPBA, 75% from J anssen), carbamide peroxide (CP, from
Peroxid Chemie), t-BuOOH (80% from Fluka), Oxone
(2KHSO5‚KHSO4‚K2SO4, from J anssen), dppm, dppe, diphoe,
dppp, and dppb (all from Strem), and most of the synthetic
reagents were commercial products and were used without
purification.
The following compounds were prepared according to lit-
erature procedures: [(dppm)Pt(µ-OH)]2(BF4)2,13,14 [(dppe)Pt-
(µ-OH)]2(BF4)2,13 [(dppp)Pt(µ-OH)]2(BF4)2,13,14 [(dppb)Pt(µ-OH)]2-
(BF4)2,13,14 [(dppe)Pd(µ-OH)]2(BF4)2,23 [(dppp)Pd(µ-OH)]2(BF4)2.10a
[(d p p b)P d Cl2]. To a solution of the complex (MeCN)2PdCl2
(0.50 g, 1.93 mmol) in N2-saturated C2H4Cl2 (50 mL) was added
solid dppb (0.822 g, 1.93 mmol). The resulting solution was
stirred at room temperature for 24 h. Then it was concen-
trated to small volume under reduced pressure and the
resulting pale yellow solid was filtered, washed with Et2O, and
dried in vacuo (yield 94%).
Anal. Calcd (found) for C28H32O2P2PtB2F8: C, 40.47 (40.23);
H, 3.85 (3.57). IR (KBr pellet): 1100-990 cm-1
.
31P{1H} NMR
1
(CD2Cl2): δ 4.6 ppm (s); J P-Pt 3930 Hz.
Syn th esis of La cton es. Lactones used as standards for
gas chromatography determinations in the individual catalytic
reactions were synthesized from the starting ketone (20 mmol)
in 25 mL of CH2Cl2, to which 20 mmol of MCPBA was added
under N2 with stirring. After a few hours the solid MCBA
was filtered off and the solution containing the ester/lactone
was used for qualitative identification in the GC analysis.
Ca ta lytic Rea ction s. These were carried out in a 25 mL
round-bottomed flask equipped with a stopcock for vacuum/
N2 operations and a side arm fitted with a screw-capped
silicone septum to allow sampling. Constant temperature
((0.1 °C) was maintained by water circulation through an
external jacket connected with a thermostat. For reactions
carried out at temperatures >25 °C the reaction vessel was
equipped with a reflux condenser. Stirring was performed by
a Teflon-coated bar driven externally by a magnetic stirrer.
The absence of diffusional problems was determined by the
conversion vs time plot independence of the stirring rate in
randomly selected catalytic experiments. The concentration
of the commercial H2O2 solution was checked iodometrically
prior to use.
The following general procedure was followed.
The required amount of catalyst was placed as a solid in
the reactor, which was evacuated and filled with N2. Purified,
N2-saturated ketone was added under N2 flow, followed, if
necessary, by the required amount of solvent. After thermo-
stating at the required temperature for a few minutes, the
H2O2 solution in the appropriate amount was injected through
the septum and time was started.
All reactions were monitored with GLC by direct injection
of samples taken periodically from the reaction mixtures with
a
microsyringe. Initial rate data were determined from
conversion vs time plots. Separation of the products was
performed on a 25 m HP-5 capillary column using a flame
ionization detector.
Anal. Calcd (found) for C28H28Cl2P2Pd: C, 55.69 (55.07); H,
4.64 (4.45). IR (KBr pellet): 306, 289 cm-1 (PdCl). 31P{1H}
NMR (CD2Cl2): δ 29.8 ppm (s).
The amount of residual H2O2 at different times was deter-
mined by sampling 10 µL aliquots from the aqueous phase.
These were diluted in water and titrated iodometrically.
Deter m in a tion of H2O2 Con cen tr a tion in C2H4Cl2. This
was performed according to a modified procedure of the method
described by Wolfe.24 To a typical blank reaction mixture (5
mL of C2H4Cl2 + 0.25 mL of 2-methylcyclohexanone) was
added 1.0 mL of aqueous H2O2 of known concentration. After
the mixture was stirred for a few minutes, a 4 mL portion of
the organic phase was separated. To this was added an equal
amount of a 2.1 M solution of TiCl4 in 6 N HCl, and the
mixture was stirred for 10 min. A 3.0 mL portion of the orange
aqueous phase was separated and diluted to 25 mL with the
same 2.1 M solution of TiCl4 in 6 N HCl. Spectrophotometric
analysis was carried out at 416 nm (ꢀ 583 cm-1 M-1). The
Lambert-Beer law was obeyed in the concentration range
explored.
[(d p p b)P d (µ-OH)]2(BF 4)2 (1d ). The complex [(dppb)PdCl2]
(0.50 g, 0.828 mmol) was placed in a round-bottomed flask, to
which CH2Cl2 (70 mL) and reagent grade acetone (55 mL) were
added. After the pale yellow solution was saturated with N2,
1.69 mL (1.66 mmol) of a AgBF4 solution in acetone was added.
The mixture was stirred in the dark for 1 h, and then AgCl
was filtered off. The resulting yellow solution was concen-
trated to small volume under reduced pressure, and the
addition of Et2O resulted in the precipitation of a pale yellow
solid that was filtered, washed with Et2O, and dried in vacuo
(Yield 92%).
Anal. Calcd (found) for C56H58O2P4PdB2F8: C, 58.72 (59.07);
H, 4.55 (4.42). IR (C2H4Cl2 solution): 3353 cm-1 (OH). IR
(KBr pellet): 1100-995 cm-1 31P{1H} NMR (CD2Cl2): δ 37.1
.
ppm (s). Conductivity (10-3 M in MeOH): 207 Ω-1 mol-1 cm2.
[(d p p b)P t(OH2)2](BF 4)2. The complex 1d (0.50 g, 0.34
mmol) was dissolved in N2-saturated CH2Cl2 (5 mL). To the
solution was added a large excess of 11 M HClO4 (0.313 mL,
3.4 mmol) under N2. The mixture was stirred for 1 h and then
brought to dryness in vacuo. The residue was treated with
Ack n ow led gm en t. This work was supported jointly
by the MURST and CNR. Thanks are expressed to
Professors G. Albertin and G. Annibale of this depart-
ment for helpful discussions. We thank also Miss T.
Fantinel for skillful technical assistance.
(23) Bushnell, G. W.; Dixon, K. R.; Hunter, R. G.; McFarland, J . J .
Can. J . Chem. 1972, 50, 3694.
(24) Wolfe, W. C. Anal. Chem. 1962, 34, 1328.
OM970756F