The effect of chloride ions on the mechanism of the oxidative addition of cyclic allylic carbonates to Pd0 complexes by formation of neutral [(η1-allyl)PdClL2] complexes

Article abstract of DOI:10.1002/ejoc.200500345

The reversible oxidative addition of a cyclic allylic carbonate cis-1 to palladium(0) complexes ligated by PPh₃ is affected by chloride ions in DMF or chloroform. In the presence of chloride ions, the oxidative addition may become irreversible

Full text of DOI:10.1002/ejoc.200500345

FULL PAPER
The Effect of Chloride Ions on the Mechanism of the Oxidative Addition of
Cyclic Allylic Carbonates to Pd0 Complexes by Formation of Neutral
[(η1-allyl)PdClL2] Complexes
Thibault Cantat,[a] Nicolas Agenet,[a] Anny Jutand,*[a] Roser Pleixats,[b] and
Marcial Moreno-Mañas[b]
Keywords: Allyl ligands / Carbonates / Halides / Palladium / Reaction mechanisms
The reversible oxidative addition of a cyclic allylic carbonate
cis-1 to palladium(0) complexes ligated by PPh3 is affected
by chloride ions in DMF or chloroform. In the presence of
chloride ions, the oxidative addition may become irreversible
and the rate of isomerization of cis-1 to trans-1 is slowed
down. This is a consequence of a slower isomerization of the
cationic complex [(η3-allyl)Pd(PPh3)2]+ (cis-2+) to trans-2+ due
to the formation of neutral [(η1-allyl)PdCl(PPh3)2] (cis-3/
trans-3) complexes. Such complexes are formed instead of
the cationic cis-2+ and trans-2+ whatever the source of chlo-
ride ions, i.e., whether voluntarily added to the cationic com-
plexes cis-2+ and trans-2+, voluntarily added in the oxidative
addition of cis-1 with Pd0 complexes ligated to PPh3, or intro-
duced in the dimeric complexes [(η3-allyl)Pd(μ-Cl)]2 (cis-5/
trans-5).
(© Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim,
Germany, 2005)
The origin of the loss of stereospecificity was determined
by investigating the mechanism of the oxidative addition of
the cyclic allylic carbonate cis-1 to [Pd0(PPh3)3] or
[Pd0(dba)(PPh3)2], generated from the precursors
[Pd0(PPh3)4] or {Pd0(dba)2 and 2 PPh3} respectively.[3] The
oxidative addition was found to be reversible in DMF
(Scheme 3), leading to cationic [(η3-allyl)Pd(PPh3)2]+ (trans-
2+/cis-2+) via the common reactive [Pd0(PPh3)2] complex
(Scheme 4). The overall reaction proceeds with isomeriza-
tion at the allylic position by isomerization of the interme-
diate complexes [(η3-allyl)Pd(PPh3)2]+ (trans-2+/cis-2+) ac-
cording to an SN2 mechanism (Scheme 4),[3,4] which is re-
sponsible for the loss of stereospecificity. This study also
established that the decarboxylation of the ethyl carbonate
anion is not as fast as initially postulated.[3,5]
Introduction
Palladium(0) complexes catalyze the substitution of al-
lylic carbonates by neutral or anionic nucleophiles
(Scheme 1).[1]
Scheme 1.
It has been established by some of us that the nucleo-
philic substitution on the cyclic allylic carbonate cis-1 pro-
ceeds with a loss of stereospecificity when [Pd0(PPh3)4] is
the precatalyst (Scheme 2), whereas a nice overall retention
of configuration is achieved with [Pd2(dba)3]/2dppe.[2]
Scheme 3.
Scheme 2.
[a] Ecole Normale Supérieure, Département de Chimie, UMR
CNRS-ENS-UPMC 8640
24, Rue Lhomond, 75231 Paris, Cedex 5, France
Fax: +33-1-44323325
E-mail: Anny.Jutand@ens.fr
[b] Departament de Química, Universitat Autonoma de Barcelona,
Cerdanyola, 08193 Barcelona, Spain
Supporting information for this article is available on the
WWW under http://www.eurjoc.org or from the author.
Scheme 4.
Eur. J. Org. Chem. 2005, 4277–4286
DOI: 10.1002/ejoc.200500345
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4277

T. Cantat, N. Agenet, A. Jutand, R. Pleixats, M. Moreno-Mañas
FULL PAPER
In the case of related cis or trans cyclic allylic acetates absorbance decreases slightly with time and is stabilized af-
(CO2Me instead of Me on the cycle), Bäckvall et al. have ter addition of 5 equiv. of the cyclic allylic carbonate cis-1
observed that the stereospecificity of a nucleophilic substi- (Figure 1, a), thus attesting that an oxidative addition had
tution by Et2NH is improved when performed in the pres- taken place (Scheme 3, a).[3] However, the equilibrium in
ence of chloride ions.[4e] The effect of chloride ions on the Scheme 3 (a) is not totally shifted to its right-hand side at
efficiency, regioselectivity, and enantioselectivity of palla- this concentration of cis-1. Chloride ions (1, 2, 5, and
dium-catalyzed allylic substitutions has been also report- 10 equiv.) were then added successively using nBu4NCl as
ed.[6a–6f]
the anion source (Figure 1, a). Successive decreases of the
In 1981, Åkermark et al. reported a difference of reactiv- absorbance of [Pd0(dba)(PPh3)2] were observed, thereby
ity of dimethylamine with an isolated cationic complex [(η3- suggesting that the chloride ions affect the equilibrium of
allyl)PdL2]+BF4– (L = PPh3) when compared to the neutral Scheme 3 (a) by shifting it towards the right-hand side. It is
complex [(η1-allyl)PdClL2], hypothetically generated in situ worthwhile to note that [Pd0(dba)(PPh3)2] does not react
upon treatment of [(η3-allyl)Pd(μ-Cl)]2 with 4 equiv. of L. with chloride ions at the concentrations used above (vide
The regioselectivity of the reaction was also affected.[6g] It infra).
has been established by some of us that chloride ions, vol-
In order to discriminate between a specific effect of chlo-
untarily added as a salt or introduced by the palladium pre- ride ions due to their coordinating properties and the effect
cursor, play an important role in palladium-catalyzed sub- of increasing ionic strength due to the introduction of the
stitutions of allylic acetates.[7] Indeed, chloride ions modify ionic species nBu4N+ and Cl– (completely dissociated in
the kinetics of the oxidative addition of allylic acetates to DMF), the same UV experiments were done with
Pd0 complexes,[7a] the structure of the allylpalladium(ii) nBu4NBF4, whose anion is not coordinating. A decrease of
complexes generated in the oxidative addition by formation the absorbance of [Pd0(dba)(PPh3)2] was observed (Fig-
of neutral complexes of the type [(η1-allyl)PdClL2] (L = ure 1, b), thus indicating that the equilibrium in Scheme 3
PPh3) instead of cationic [(η3-allyl)PdL2]+ complexes,[7,8] (a) is indeed slightly affected by the ionic strength of the
and the mechanism of the second step of the catalytic cycle, medium, which favors the formation of the cationic com-
i.e., the nucleophilic attack on allylpalladium(ii) complexes, plexes 2+. Indeed, the rate of the backward reaction, i.e.,
since neutral [(η1-allyl)PdClL2] complexes were found to re- the attack of the ethyl carbonate anion on the cationic com-
act in parallel with cationic [(η3-allyl)PdL2]+ complexes.[7b] plexes 2+, must decrease in an ionic medium due to charge
We report herein that chloride ions also affect the mecha- separation.
nism (kinetics and isomerization) of the oxidative addition
However, one notices that for the same concentration of
of cyclic allylic carbonates to Pd0 complexes by forming added nBu4NBF4 or nBu4NCl, the effect of the latter is
neutral [(η1-allyl)PdII] chloride complexes.
much higher (compare Figures 1, parts a and b), which sug-
Results and Discussion
Evidence of a Specific Effect of Chloride Ions on the
Thermodynamics of the Oxidative Addition of the Cyclic
Allylic Carbonate cis-1 to Pd0 Complexes
The complex [Pd0(dba)(PPh3)2], quantitatively generated
from a solution of Pd0(dba)2 (1 mm) and 2 equiv. of PPh3
in DMF, was characterized by UV spectroscopy by an ab-
sorption band at λmax = 396 nm (Figure 1, part a).[9a,9b] Its Scheme 5.
Figure 1. UV spectroscopy at 396 nm in DMF at 20 °C. a) Initial absorbance of [Pd0(dba)(PPh3)2] (1 mm) generated from Pd0(dba)2
(1 mm) and PPh3 (2 mm). The arrows indicate the addition of nЈ = 5 equiv. of cis-1 at t = 100 s and then successive additions of n equiv.
of nBu4NCl. b) The same experiment as in part a of this figure but with addition of nBu4NBF4. In both cases, n is the total number of
added equivalents.
4278
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
Eur. J. Org. Chem. 2005, 4277–4286

The Mechanism of the Oxidative Addition of Cyclic Allylic Carbonates to Pd0
FULL PAPER
Figure 2. Oxidative addition of cis-1 (0.1 m) to the palladium(0) species generated in situ from Pd(dba)2 (1 mm) and PPh3 (2 mm) in DMF
at 15 °C. a) As monitored by UV spectroscopy: variation of 2lnx + 1 – x vs. time [x = (D – Dlim)/(D0 – Dlim)] {D: absorbance of
[Pd0(dba)(PPh3)2] at time t; D0: initial absorbance of [Pd0(dba)(PPh3)2]; Dlim: absorbance at infinite time}. b) Kinetics of the formation
of the cationic complex 2+·EtOCO2– generated in the oxidative addition under the same conditions as in part a of this figure, as monitored
by conductivity measurements. c) Variation of 2lnx + 1 – x vs. time [x = (κlim – κ)/(κlim – κ0)] (κ: conductivity of 2+·EtOCO2 at time t;
–
–
κ0: residual initial conductivity; κlim: final conductivity of 2+·EtOCO2 ).
gests a specific effect of chloride ions on the equilibrium of ments[10] under the same experimental conditions ([cis-1] =
Scheme 3 (a). At high chloride concentrations (n = 0.1 m). The kinetic curve is shown in Figure 2, b. The curve
10 equiv.), the residual absorbance is almost reached (Fig- is not S-shaped, which indicates that the neutral intermedi-
ure 1, a). This suggests a reaction of the chloride ions with ate complex [(η2-allyl-OCO2Et)Pd0(PPh3)2], which is sup-
the cationic complexes 2+, which generates the new com- posedly formed in a first complexation step (as observed
plexes trans-3/cis-3 (Scheme 5). Such a reaction shifts the with acyclic allylic acetates),[10] does not accumulate before
equilibrium in Scheme 3 (a) totally towards its right-hand its ionization to the cationic complex 2+. In other words,
side. The structure of complexes trans-3/cis-3 will be con- the rate of the ionization step is faster than the rate of the
firmed later on.
complexation step. Consequently, the rate of formation of
the cationic complex 2+, as monitored by conductivity mea-
surements, should be similar to the rate of disappearance
of the Pd0 complex, as monitored by UV spectroscopy (Fig-
Specific Effect of Chloride Ions on the Kinetics of the
Oxidative Addition of the Cyclic Allylic Carbonate cis-1 to
Pd0 Complexes in DMF
ure 2, a). Indeed, the plot of 2lnx + 1 – x vs. time [x =
–
(κlim – κ)/(κ
– κ0); (κ: conductivity of 2+·EtOCO2 at
lim
The reaction of the allylic carbonate cis-1 with
[Pd0(dba)(PPh3)2] (1 mm) in DMF was performed with a
high concentration of cis-1 (0.1 m) such that the equilibrium
in Scheme 3 (a) was totally shifted towards its right-hand
side (Scheme 6). The kinetics of the reaction was monitored
by UV spectroscopy by recording the decrease of the ab-
sorbance of [Pd0(dba)(PPh3)2] at λ = 396 nm with time.
time t; κ0: residual initial conductivity; κlim: final conductiv-
–
ity of 2+·EtOCO2 )] is linear (Figure 2, c). The observed
rate constant (kobs = 6.2×10–2 s–1 in DMF at 15 °C) was
calculated from the slope and was found to be close to the
rate of disappearance of the Pd0 complex (kobs
=
7.2×10–2 s–1; vide supra) within the accuracy of both tech-
niques.
In order to anticipate any effect of the ionic strength that
might occur when the kinetics of the oxidative addition of
cis-1 is investigated in the presence of nBu4NCl, the kinetics
was first investigated in the presence of nBu4NBF4 with a
concentration in the range 0.3–0.9 m, i.e., at high ionic
strength. nBu4NBF4 was added to [Pd0(dba)(PPh3)2] (1 mm)
before the addition of cis-1 (0.1 m) and the kinetics was
monitored by UV spectroscopy, as in the absence of
nBu4NBF4 (vide supra). The rate of the reaction increased
slightly upon increasing the nBu4NBF4 concentration. Fig-
Scheme 6.
The kinetic law of the overall reaction in Scheme 6 is
given in Equation (1), where x = (D – Dlim)/(D0 – Dlim) {D
= absorbance of [Pd0(dba)(PPh3)2] at time t, D0 is the initial
absorbance of [Pd0(dba)(PPh3)2], and Dlim the absorbance
at infinite time}.[9c]
BF4
ure 3 (a) shows the linear variation of log(kobs /kobs0) with
BF4
the square root of nBu4NBF4 concentration (kobs : ob-
0
2lnx + 1 – x = –kK[cis-1]t/C0 = –kobs
t
(1)
served rate constant in the presence of nBu4NBF4; kobs
:
observed rate constant in the absence of nBu4NBF4, as de-
termined in Figure 2, a), which is in agreement with the
Debye–Hückel law, thus attesting that the accelerating effect
is indeed due to the ionic strength.[11]
The plot of 2lnx + 1 – x vs. time is linear (Figure 2, a).
The value of kobs = 7.2×10–2 s–1 was determined from the
slope of the straight line and the value of kK = 7.2×10–4 s–1
(DMF, 15 °C) as well.
The kinetics of formation of the ionic complexes
The kinetics of the reaction of allylic carbonate cis-1
–
2+·EtOCO2 was monitored by conductivity measure- (0.1 m) with [Pd0(dba)(PPh3)2] (1 mm) in DMF was then in-
Eur. J. Org. Chem. 2005, 4277–4286
www.eurjoc.org
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4279

T. Cantat, N. Agenet, A. Jutand, R. Pleixats, M. Moreno-Mañas
FULL PAPER
Figure 3. a) Effect of the ionic strength in the reaction of cis-1 (0.1 m) with the Pd0 complex generated from Pd(dba)2 (1 mm) and PPh3
BF4
(2 mm) in DMF at 15 °C. Variation of log(kobs /kexp0) vs. the square root of nBu4NBF4 concentration. b) Specific effect of chloride
ions in the reaction of cis-1 (0.1 m) with the Pd0 complex generated from Pd(dba)2 (1 mm) and PPh3 (2 mm) in DMF at 15 °C. Variation
of log(kexpCl/kexp0) vs. the square root of nBu4NCl concentration. aЈ) Theoretical effect of the ionic strength due to the presence of
nBu4NCl. The straight line in part aЈ of this figure has been rebuilt from the straight line of part a.
vestigated in the presence of various amounts of nBu4NCl ence of various amounts of chloride ions, added to the Pd0
in the range 1–10 mm, i.e., at lower concentrations than that complex as nBu4NCl before the cyclic allylic carbonate cis-
of nBu4NBF4 (0.3–0.9 m) in the previous experiments. An 1 (Scheme 7).
accelerating effect was observed that was more pronounced
than the one due to the ionic strength. In a first approach,
log(kobsCl/kobs0) (kobsCl: observed rate constant in the pres-
ence of nBu4NCl; kobs0: observed rate constant in the ab-
sence of nBu4NCl, determined as in Figure 2, a) was plotted
against the square root of nBu4NCl concentration (Fig-
ure 3, b). A straight line was not obtained. Each data point
Scheme 7.
In CDCl3, and starting from a stoichiometric amount of
could be compared to the expected one due to the effect of
the ionic strength. The latter is represented by the straight
line in Figure 3 (aЈ), which was rebuilt from the straight line
in Figure 3, a. A comparison of Figures 3 (parts b and aЈ)
cis-1 and [Pd0(PPh3)4] (20 mm each), the isomerization of
cis-1 to trans-1 afforded the thermodynamic ratio 38:62,
which was reached within 10 min (curve a in Figure 4), as
1
observed in the first H NMR spectrum recorded. When
shows that, for a similar salt concentration of 0.01 m, the
the same reaction was performed in the presence of a stoi-
rate of the oxidative addition is 3.4 times faster in the pres-
chiometric amount of nBu4NCl (20 mm), the same ratio was
ence of nBu4NCl than in the presence of nBu4NBF4. Conse-
1
also obtained in the first H NMR spectrum (curve b in
quently, a specific effect of chloride ions is observed which
might be due to the formation of anionic [Pd0(PPh3)2Cl]
Figure 4). At these concentrations, the kinetics of the isom-
erization is too fast to be monitored accurately by 1H NMR
spectroscopy and the effect of chloride ions could not be
determined.
–
species[7a] that lead to the neutral [(η1-allyl)PdCl(PPh3)2]
complexes trans-3/cis-3. The accelerating effect is not large
because the competition between dba and Cl– for the coor-
dination of Pd0(PPh3)2 is strongly in favor of dba. Indeed,
it is only upon addition of a large amount of nBu4NCl
(0.75 m) that 30% of [Pd0(dba)(PPh3)2] (C0 = 1 mm) disap-
peared due to the formation of [Pd0(PPh3)2Cl]–, as attested
by UV spectroscopy.[7a] At low chloride concentrations,
[Pd0(PPh3)2Cl]– is present at only trace levels.
Specific Effect of Chloride Ions on the Kinetics of the
Isomerization of the Cyclic Allylic Carbonate cis-1 to trans-
1 Induced by Pd0 Complexes
As indicated in Schemes 3 and 4, isomerization of the
cyclic allylic carbonate cis-1 to trans-1 is observed in the
presence of [Pd0(PPh3)4] or [Pd0(dba)(PPh3)2].[3] The rate of
isomerization was monitored by H NMR spectroscopy in
CDCl3 by the integration of the signal of the allylic protons
Figure 4. Effect of chloride ions on the kinetics of the isomerization
of cis-1 to trans-1 induced by [Pd0(PPh3)4] in CDCl3. a) cis-1 +
[Pd0L4] (20 mm/20 mm); b) cis-1 + [Pd0L4] + Cl– (20 mm/20 mm/
20 mm); c) cis-1 + [Pd0L4] + Cl– (200 mm/20 mm/20 mm); d) cis-1 +
[Pd0L4] + Cl– (200 mm/20 mm/200 mm).
1
of cis-1 and trans-1 located at δ = 5.24 and 5.10 ppm respec-
When the concentration of cis-1 was increased to
tively. The same experiment was also performed in the pres- 200 mm, the isomerization was slower and the effect of chlo-
4280
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
Eur. J. Org. Chem. 2005, 4277–4286

The Mechanism of the Oxidative Addition of Cyclic Allylic Carbonates to Pd0
FULL PAPER
rides could be observed, as seen by comparing curves c and with Cl– as the counteranion. This strongly suggests the for-
d in Figure 4. The isomerization proceeded slower upon in- mation of the neutral [(η1-allyl)PdCl(PPh3)2] complexes cis-
creasing the chloride concentration. In another experiment 3 and trans-3 (Scheme 9). A 31P NMR spectrum recorded
involving [Pd0(PPh3)4] (37 mm) and cis-1 (74 mm), the isom- on the same solution did not exhibit the two singlets char-
erization to trans-1 was complete within 35 min. When the acteristic of the cationic complexes cis-2+ and trans-2+ (En-
same reaction was performed in the presence of nBu4NCl try 1 in Table 2) but two new sharp singlets were detected
(37 mm), an induction period without isomerization was at δ = 23.92 and 23.39 ppm (Entry 2 in Table 2), which were
observed. The isomerization started to be significant after assigned to the neutral complexes cis-3 and trans-3.
60 min, indicating that the isomerization cis-1 to trans-1 is
slowed down in the presence of chlorides.
When the oxidative addition of the allylic carbonate cis-
1 (20 mm) to [Pd0(dba)(PPh3)2] (20 mm), generated from
This inhibiting effect of chloride ions on the rate of isom- Pd0(dba)2 + 2 PPh3, was performed in CDCl3 in the pres-
erization of cis-1 to trans-1 can be explained by the forma- ence of nBu4NCl (20 mm), the same 1H NMR spectrum was
tion of complexes [(η1-allyl)PdCl(PPh3)2] (cis-3 and trans-3; obtained, thus attesting to the formation of the same com-
Scheme 8). The formation of such complexes would de- plexes cis-3 and trans-3 (Scheme 9).
crease the rate of the SN2 isomerization of the cationic com-
The following part of this work is devoted to the vali-
plexes cis-2+ and trans-2+ by decreasing their concentration. dation of the mechanism proposed in Scheme 8 by the inde-
In other words, part of the palladium catalyst is converted pendent synthesis of the neutral complexes [(η1-allyl)
into the neutral complexes cis-3 and trans-3, which cannot PdCl(PPh3)2] (cis-3 and trans-3).
isomerize directly due to the cyclic structure of the η1-allylic
ligand.
Synthesis of the Complexes [(η1-allyl)PdCl(PPh3)2] (trans-3
and cis-3) by Treating the Dimer [(η3-allyl)Pd(μ-Cl)]2
(trans,trans-5 and cis,cis-5) with Triphenylphosphane
The synthesis of the dimeric complexes 5 was adapted
from the Bosnich procedure[12] using a mixture of the cyclic
allylic chlorides cis-4 and trans-4 (65:35) (Scheme 10).
The oxidative addition of the Pd0 generated by reduction
2–
of PdCl4 by CO was slow, and a large excess of the allylic
chlorides was required. The dimeric complex was isolated
as a yellow powder and was not very stable, as currently
1
observed in the cyclic series.[4d,13] The H NMR spectrum
in CDCl3 (Figure S4 in the Supporting Information) re-
vealed the formation of two complexes trans,trans-5 and
cis,cis-5, as evidenced by two distinct doublets for the Me
group as well as two different sets of signals for the allylic
Scheme 8.
protons Hc and Hs, which were discriminated by a COSY
A specific reaction of chloride ions was indeed observed. experiment (Entry 2 in Table 1). The cis and trans structures
The reaction was monitored by 1H NMR spectroscopy per- were assigned by comparison with related complexes re-
formed on a solution of cis-1, [Pd0(PPh3)4], and nNBu4Cl ported by Kurosawa et al. (CO2Me instead of Me)[14] in
at the same concentration of 20 mm in CDCl3. A fast cis-1/ which the signals of the allylic protons (Hc and Hs) in the
trans-1 isomerization was observed first (vide supra Fig- trans dimer are located at higher field than those of the cis
ure 4, b). After 4 h, the signals of the allylic carbonates cis- complex. The ratio cis,cis-5/trans,trans-5 (67:33) was de-
1 and trans-1 were no longer detected (Figure S3 in the Sup- duced from the 1H NMR spectrum. The stereochemistry of
porting Information) and the free carbonate anion the oxidative addition of such cyclic allylic chlorides 4 to
EtOCO2– had been released into solution, as attested by the palladium(0) is not clear and seems to depend strongly on
presence of the quadruplet for the CH2 protons of its Et the solvent.[14]
group at δ = 3.72 ppm (Figure S3). Two new sets of allylic
The dimeric complexes cis,cis-5/trans,trans-5 (67:33) were
protons were detected whose total integration was one pro- treated with PPh3 (successively PPh3/Pd = 1, 2, 3) in the
ton, as determined by comparison with the integration of absence of any chloride scavenger and the reactions were
the CH2 protons of EtOCO2– at δ = 3.72 ppm. Close vinylic monitored by H and 31P NMR spectroscopy in CDCl3.
1
protons were also observed whose total integration corre- Two kinds of complexes were generated depending on the
sponded to two protons (Entry 1 in Table 1, and Figure S3). ratio PPh3/Pd, as attested by their different 1H and 31P
These two sets of vinylic and allylic protons do not be- NMR spectra. When the PPh3/Pd ratio was 1, the mixed
long to the known cyclic allylic chlorides cis-4 and trans-4 complexes cis-6 and trans-6 (70:30) were generated
(see Exp. Sect. and Figure S2), and neither do they belong (Scheme 11, Entry 5 in Table 2).
to the cationic complexes cis-2+ and trans-2+ (Entry 4 in
Table 1 and Figure S6), which could have been generated oxidative addition of cis-4 and trans-4 (65:35) to the Pd0
These complexes are the same as those isolated after the
Eur. J. Org. Chem. 2005, 4277–4286
www.eurjoc.org
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4281

T. Cantat, N. Agenet, A. Jutand, R. Pleixats, M. Moreno-Mañas
FULL PAPER
Table 1. 1H NMR spectra (250 MHz vs. TMS) in CDCl3 of the Me protons of the cycle and allylic and/or vinylic protons of allylpalladium
complexes.
1
[a] The cis/trans ratio is that obtained in the reaction of Scheme 9 with [Pd0(PPh3)4]. [b] H NMR (400 MHz). [c] The cis/trans ratio is
that obtained in the reaction of Scheme 12.
Table 2. 31P NMR shifts (101 MHz vs. H3PO4) in CDCl3.
n
Origin
δ [ppm]
–
1
cis-2+·BF4
Scheme 14
Scheme 14
Scheme 9
Scheme 9
Scheme 15
Scheme 15
Scheme 13
Scheme 13
Scheme 11
Scheme 11
Scheme 12
Scheme 12
–
23.73 (34%)
22.85 (66%)
23.92 (42%)
23.39 (58%)
23.92 (44%)
23.40 (56%)
23.90 (43%)
23.38 (57%)
23.91 (70%)
22.97 (30%)
23.91 (55%)
22.97 (45%)
25.9/24.3
–
trans-2+·BF4
cis-3
2
3
4
5
6
7
Scheme 9.
trans-3
cis-3
trans-3
cis-3
trans-3
cis-6
trans-6
cis-6
trans-6
generated from Pd0(dba)2 + 2 PPh3 (Scheme 12, Entry 3 in
Table 1, Entry 6 in Table 2, Figure S5 in the Supporting In-
formation). Surprisingly, only one phosphane ligand is in-
corporated into complexes 6. This is due to the formation
of [PdCl2(PPh3)2] (δ =23.4 ppm) as a by-product, which
consumes part of the PPh3 ligand.
cis-7+·Cl–/trans-
7+·Cl–[a]
The same complexes cis-6 and trans-6 were also obtained
in the oxidative addition of cis-4 and trans-4 (65:35) to
[Pd0(PPh3)4] in CDCl3, along with the formation of the al-
lylic phosphonium salts cis- and trans-7+·Cl–.[16] The forma-
[a] No attempt was made to discriminate the cis and trans struc-
tures.
tion of the latter from the allylic chlorides cis-4 and trans- deduced by comparison with related complexes (CO2Me in-
4 is catalyzed by Pd0 complexes.[16]
stead of Me), which were synthesized by an oxidative ad-
The protons of the two complexes 6 were assigned from dition of the corresponding cyclic allylic chlorides to
a COSY spectrum and the cis and trans structures were [Pd0(PPh3)4], as reported by Kurosawa et al.[15]
4282
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
Eur. J. Org. Chem. 2005, 4277–4286

The Mechanism of the Oxidative Addition of Cyclic Allylic Carbonates to Pd0
FULL PAPER
establishes the complexation of the Pd atom by two phos-
phane ligands in the complexes cis-3 and trans-3.
Formation of the Neutral Complexes [(η1-allyl)PdCl(PPh3)2]
(cis-3 and trans-3) by Treatment of the Cationic
Complexes [(η3-allyl)Pd(PPh3)2]+ (cis-2+ and trans-2+) with
Chloride Ions
The cationic complexes [(η3-allyl)Pd(PPh3)2]+ (cis-2+ and
trans-2+) were synthesized by treating the dimeric com-
plexes cis,cis-5/trans,trans-5 (67:33) with PPh3 (PPh3/Pd =
2) in the presence of NaBF4 as a chloride scavenger, accord-
ing to a related procedure reported by Powell and Shaw
(Scheme 14).[17] The cationic complexes cis-2+ and trans-2+
with BF4– as the counteranion were isolated and fully char-
acterized by 1H and 31P NMR spectroscopy (Entry 4 in
Table 1, Entry 1 in Table 2, Figure S6 in the Supporting In-
formation), and FAB and electrospray mass spectrometry.
Two sets of signals are clearly distinguishable for the allylic
protons (central Hc and syn Hs), corresponding to the two
isomers with a cis-2+/trans-2+ ratio of 34:66. According to
reports of related complexes,[4d] the protons of the trans-2+
complex are located at higher field than those of the cis-2+
complex.
Scheme 10.
Scheme 11.
Scheme 12.
Scheme 14.
When the dimeric complexes cis,cis-5/trans,trans-5
(67:33) were treated with 4 equiv. of PPh3 per dimer (PPh3/
When the mixture of the isolated cationic complexes cis-
2+/trans-2+ (34:66) in CDCl3 was treated with 10 equiv. of
Pd = 2), two new complexes were generated and assigned
1
Cl–, the H and 31P NMR spectra changed to those of the
1
as cis-3 and trans-3 (Scheme 13). Indeed, their H and 31P
neutral complexes [(η1-allyl)PdCl(PPh3)2] (cis-3 and trans-
3; Scheme 15) already observed (Scheme 9 and Figure S3).
This clearly establishes that the cationic complexes cis-2+
and trans-2+ generated in the oxidative addition of cis-1 re-
act with chloride anions to form the neutral complexes cis-
3 and trans-3, as postulated in Scheme 8.
NMR spectra (Entry 1 in Table 1, Entry 4 in Table 2) are
the same as those obtained for the complexes generated
from the oxidative addition of cis-1 to Pd0 complexes per-
formed in the presence of added chlorides (Scheme 9).
Scheme 13.
The FAB mass spectrum does not contain the peak m/z
= 760 for the complexes cis-3 and trans-3, but instead one
at m/z = 725 (major one in the isotopic pattern characteris-
tic of the Pd atom) corresponding to the cleavage of Cl–, as
also observed in related complexes.[8e] The detection of a
mass peak at m/z = 630, which corresponds to the fragment
Scheme 15.
Conclusions
It has been shown that whatever the source of chloride
Pd(PPh3)2 after the cleavage of Cl– and the allylic group, ions, i.e. voluntarily added to the cationic complexes cis-2+
Eur. J. Org. Chem. 2005, 4277–4286
www.eurjoc.org
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4283

T. Cantat, N. Agenet, A. Jutand, R. Pleixats, M. Moreno-Mañas
FULL PAPER
and trans-2+, voluntarily added during the oxidative ad-
dition of cis-1 with Pd0 complexes ligated to PPh3, or intro-
duced in the dimeric complexes cis- and trans-[(η3-allyl)
Pd(μ-Cl)]2 (5), the neutral complexes cis- and trans-[(η1-al-
lyl)PdCl(PPh3)2] (3) are generated instead of the cationic
cis- and trans-[(η3-allyl)Pd(PPh3)2]+ (2+) (Scheme 16).
Materials: DMF was distilled from calcium hydride under vacuum
and kept under argon. PPh3 and nBu4NCl were commercial
(Acros). Pd(dba)2,[18] [Pd(PPh3)4],[19] cis-1,[20] and 5-methyl-2-cy-
clohexen-1-ol[21] were prepared according to reported procedures.
UV Experiments: These were performed in a thermostatted 1-mm
path-length cell on mixtures of Pd0(dba)2 (1 mm) and 2 equiv. of
PPh3 in DMF with a suitable amount of the allylic carbonate cis-
1, without or with known amounts of nBu4NBF4 or nBu4NCl in-
troduced from a mother solution in DMF.
Conductivity Measurements: Pd(dba)2 (8.6 mg, 0.015 mmol) and
PPh3 (7.9 mg, 0.03 mmol) were added to 15 mL of DMF at 15 °C
and the residual conductivity, κ0, was measured. cis-1 (273 μL,
1.5 mmol) was then added. The conductivity, κ, was measured with
time until a constant value, κlim, was obtained.
Synthesis of 3-Chloro-5-methylcyclohexene (4): This compound was
synthesized from the corresponding alcohol[21] according to a pub-
lished procedure.[22] 5-Methyl-2-cyclohexen-1-ol (4.78 g, 0.043 mol)
was dissolved in 55 mL of anhydrous diethyl ether and thionyl chlo-
ride (3.21 mL, 0.044 mol) was added. After 1 h the solvent and
excess of thionyl chloride were evaporated under vacuum. After
distillation (b.p. 42 °C, 15 Torr), 3.35 g of a colorless liquid was
collected (60% yield) as a mixture of cis-4 and trans-4 (65:35),
whose characterization data were as reported previously.[22] 1H
NMR (400 MHz, CDCl3): δ = 1.03 (d, J = 6.3 Hz, 3 H, cis-Me
major), 1.04 (d, J = 6.5 Hz, 3 H, trans-Me minor), 1.49 (m, 1 H,
MeCH in cis and trans), 1.8 (m, 2 H, CH2CCl), 2.03–2.23 (m, 2 H,
CH2C= in cis and trans), 4.59–4.66 (m, 1 H, CHCl in cis and trans),
5.70–5.91 (m, 2 H, HC=CH in cis and trans) ppm. 13C NMR
(63 MHz, CDCl3): cis-4: δ = 21.71, 29.56, 33.06, 42.28, 56.63,
129.21, 129.85 ppm; trans-4: δ = 21.33, 23.30, 33.55, 40.00, 55.94,
127.23, 131.37 ppm.
Scheme 16. For clarity, the stereochemistry of the complexes is not
shown (see text and Scheme 8).
Therefore, chloride ions play a specific role in the oxidat-
ive addition of the cyclic allylic carbonate cis-1 to Pd0 com-
plexes ligated by PPh3. Indeed, in the presence of chloride
ions: (i) the oxidative addition might become irreversible,
(ii) the rate of isomerization of the cyclic allylic carbonate
cis-1 to trans-1 is slowed down as a consequence of a slower
isomerization of trans-2+ to cis-2+, and (iii) neutral [(η1-
allyl)PdCl(PPh3)2] complexes are formed. Consequently,
catalytic reactions performed with the cyclic allylic carbon-
ate cis-1 or trans-1 should be more stereospecific in the
presence of chloride ions due to the irreversibility of the
oxidative addition and the slower isomerization of trans-2+
to cis-2+. This has indeed been observed by Bäckvall et al.
in palladium-catalyzed allylic substitutions by the amine
Et2NH with related cis or trans cyclic allylic acetates
(CO2Me instead of Me on the cycle) when performed in the
presence of LiCl, who noted that “the presence of chlorides
led to a remarkable improvement of stereospecificity” (substi-
tution with retention of configuration).[4d]
Characterization of the Phosphonium Salts trans-7+·Cl– and cis-
7+·Cl–: Triphenylphosphane (11.5 mg, 0.044 mmol) was dissolved
in CDCl3 in an NMR tube along with 3.5 μL (0.026 mmol) of the
cyclic allylic chloride trans-4/cis-4 (35:65) in the presence of 5 mg
1
(0.0087 mmol) of Pd(dba)2. H NMR (250 MHz, CDCl3, TMS): δ
= 0.92 (d, J = 15 Hz, 3 H, CH3), 1.45 (m, 1 H, CH–Me), 2.2 (m,
4 H, =CCH2 and CH2CP+), 5.62 (t, J = 9 Hz, 1 H, CHP+), 6.07
(m, 2 H, HC=CH) ppm. 31P NMR (101 MHz, CDCl3): δ = 25.9
(s) and 24.3 (s) ppm in the ratio 72:28. FAB-MS: m/z = 357 [M]+,
262 [M – C7H11 + H]+.
Synthesis of the Dimeric Complexes trans,trans-5 and cis,cis-5: Etha-
nol (4 mL) and 3-chloro-5-methylcyclohexene (4; 1.95 g, 8.5 mmol),
synthesized as above, were added to sodium tetrachloropalladate
(626 mg, 2.1 mmol) in 0.9 mL of degassed water. Carbon monoxide
was then bubbled through the solution at 45 °C for 60 min. A yel-
low precipitate appeared and the solution was kept at –20 °C for
3 d. The yellow precipitate was then filtered, washed with water,
ethanol, and diethyl ether, and dried to give 288 mg of a mixture
of trans,trans-5 and cis,cis-5 (58% yield), which was kept at –20 °C
in the dark (Figure S4 in the Supporting Information). cis,cis-5:
Experimental Section
General Methods: All experiments were performed under argon
using standard Schlenk techniques. 1H NMR spectra were recorded
with a Bruker spectrometer (400 MHz or 250 MHz) with TMS as
an internal reference. 31P NMR spectra were recorded with a
Bruker spectrometer (101 MHz) with H3PO4 as external reference.
UV spectra were recorded with a DU 7400 Beckman spectropho-
tometer. The conductivity was measured with a conductivity meter
CDM210 (Radiometer Analytical). The cell constant was 1 cm–1.
The conductivity was recorded against time using a computerized
home-made program. FAB mass spectra were recorded with a
JEOL MS 700 spectrometer with the Magic Bullet matrix. Electro-
spray mass spectrometry was performed with a T100LC spectrome-
ter (Jeol AccuTOF JMS).
1
67% of the mixture. H NMR (250 MHz, CDCl3): δ = 0.88 (d, J
= 6.4 Hz, 3 H, Me), 1.48 (m, 3 H), 2.0 (m, 2 H), 5.06 (dd, JHc,Hs
= 6.2, JHs,Hα = 6.2 Hz, 2 H, Hs), 5.38 (t, JHc,Hs = 6.2 Hz, 1 H,
1
Hc) ppm. trans,trans-5: 33% of the mixture. H NMR (250 MHz,
CDCl3): δ = 0.86 (d, J = 7 Hz, 3 H, Me), 1.48 (m, 3 H), 2.0 (m, 2
H), 4.87 (m, 2 H, Hs), 5.33 (t, JHc,Hs = 6.7 Hz, 1 H, Hc) ppm.
–
–
Synthesis of the Cationic Complex trans-2+·BF4 /cis-2+·BF4 : A
solution of triphenylphosphane (44 mg, 0.168 mmol) in 5 mL of
acetone was added to a solution of the dimer trans,trans-5/cis,cis-5
(20 mg, 0.042 mmol) in 6 mL of acetone. Water (2 mL) was then
4284
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
Eur. J. Org. Chem. 2005, 4277–4286

The Mechanism of the Oxidative Addition of Cyclic Allylic Carbonates to Pd0
FULL PAPER
added followed by a solution of sodium tetrafluoroborate (0.52 g,
4.76 mmol) in 3 mL of water. A grey precipitate of trans-2+·BF4–
and cis-2+·BF4– (66:34) was formed and collected by filtration
vent was evaporated to give 36.4 mg of the complexes trans-6 and
cis-6 as an orange solid (86% yield). trans-6: 48%. 1H NMR
(250 MHz, CDCl3): δ = 0.72 (d, J = 6 Hz, 3 H, Me), 1.0–1.16 (m,
2 H), 1.69–1.87 (m, 2 H), 2.28–2.42 (m, 1 H), 4.10 (m, 1 H, HsЈ),
5.60–5.62 (m, 2 H, Hc, HsЈЈ), 7.42 (m, 9 H, PPh3), 7.66 (m, 6 H,
–
1
(15 mg, 22% yield). trans-2+·BF4 : H NMR (250 MHz, CDCl3): δ
= 0.57 (d, J = 6.4 Hz, 3 H, Me), 0.94–1.05 (m, 3 H), 1.57 (m, 2 H),
5.08 (m, 2 H, Hs), 6.22 (t, JHc,Hs = 7.3 Hz, 1 H, Hc), 7.23–7.40 (m, PPh3) ppm. 31P NMR (101 MHz, CDCl3): δ = 22.97 (s) ppm. cis-
30 H, PPh3) ppm. 31P NMR (101 MHz, CDCl3): δ = 22.85 (s) ppm.
6: 52%. H NMR (250 MHz, CDCl3): δ = 0.76 (d, J = 6 Hz, 3 H,
1
cis-2+·BF4 : H NMR (250 MHz, CDCl3): δ = 0.59 (d, J = 6.5 Hz,
Me), 1.0–1.16 (m, 2 H), 1.69–1.87 (m, 2 H), 2.28–2.42 (m, 1 H),
–
1
3 H, Me), 0.94–1.05 (m, 3 H), 1.57 (m, 2 H), 5.18 (dd, JHc,Hs = 6.5, 4.13 (m, 1 H, HsЈ), 5.66 (t, J = 7 Hz, 1 H, Hc), 5.84 (q, J = 7 Hz,
JHs,Hα = 6.5 Hz, 2 H, Hs), 6.34 (t, JHc,Hs = 6.5 Hz, 1 H, Hc), 7.23– HsЈЈ), 7.42 (m, 9 H, PPh3), 7.66 (m, 6 H, PPh3) ppm. 31P NMR
7.40 (m, 30 H, PPh3) ppm. 31P NMR (101 MHz, CDCl3): δ = 23.73
(s) ppm. FAB-MS: m/z = 725 [M], 630 [M – allyl(C7H11)], 463 [M –
PPh3], 368 [Pd(PPh3)].
(101 MHz, CDCl3): δ = 23.91 (s) ppm.
Characterization of trans-6 and cis-6: These complexes were gener-
ated in situ as shown in Scheme 12. Pd(dba)2 (5 mg, 0.0088 mmol)
and PPh3 (4.6 mg, 0.0017 mmol) were introduced into an NMR
tube containing 0.5 mL of CDCl3. The complex [Pd0(dba)(PPh3)2],
generated in situ, exhibits two broad signals at δ = 27.03 and
24.92 ppm. They disappeared after addition of 1.16 μL
(0.0088 mmol) of the allylic chlorides trans-4 and cis-4 (35:65). Two
new 31P NMR singlets were observed in CDCl3 at δ = 23.91 and
22.97 ppm (in a 55:45 ratio). The 1H NMR spectrum showed a
total consumption of the allylic chlorides and exhibited the signals
of complexes trans-6 and cis-6 reported above. The same reaction
was performed with [Pd0(PPh3)4] (11.5 mg, 0.01 mmol) in CDCl3
under stoichiometric conditions; the same 1H and 31P NMR signals
assigned to the complexes trans-6 and cis-6 were observed but the
allylic phosphonium salts were also detected as major compo-
nents.[16] In addition, [PdCl2(PPh3)2] (δ = 23.4 ppm) and OPPh3 (δ
= 29.1 ppm) were also detected.
Characterization of Complexes trans-3 and cis-3: These complexes
were generated in situ, as shown in Scheme 9, from [Pd0(PPh3)4]
(11.56 mg, 0.01 mmol), nBu4NCl (2.8 mg, 0.01 mmol), and cis-1
(1.8 μL, 0.01 mmol) in 0.5 mL of CDCl3. The course of the reac-
1
tion was monitored by H NMR spectroscopy. The isomerization
cis-1 to trans-1 was observed first. After 4 h, the signals of cis-1
and trans-1 were no longer detectable and the spectrum exhibited
the signals of cis-3 and trans-3. cis-3: 42%. 1H NMR (250 MHz,
CDCl3): δ = 0.73 (d, J = 6.5 Hz, 3 H, Me), 1.95–2.1 (m, 2 H), 4.26–
4.35 (m, 1 H, H3), 5.75 (dd, J = 8.5, J = 5 Hz, 1 H, H1), 5.84–5.85
(m, H2), 7.2–7.5 (m, 18 H, PPh3), 7.6–7.7 (m, 12 H, PPh3) ppm.
The signals of the aliphatic protons of the cycle are not given due
to overlapping with the protons of the CH3CH2CH2 chain of
nBu4NCl. 31P NMR (101 MHz, CDCl3): δ = 23.92 ppm. trans-3:
1
58%. H NMR (250 MHz, CDCl3): δ = 0.77 (d, J = 6.5 Hz, 3 H,
Me), 1.95–2.1 (m, 2 H), 4.18–4.23 (m, 1 H, H3), 5.69 (br. d, J =
8.5 Hz, 1 H, H1), 5.87–5.89 (m, H2), 7.2–7.5 (m, 18 H, PPh3), 7.6–
7.7 (m, 12 H, PPh3) ppm. 31P NMR (101 MHz, CDCl3): δ = 23.39
(s) ppm. ESI-MS (methanol): m/z = 725 [M – Cl]+, 630 [M – Cl –
allyl(C7H11)].
Supporting Information: 1H NMR spectra of compounds cis-1, cis-
4/trans-4, complexes cis-3/trans-3, cis,cis-5/trans-trans-5, cis-6/trans-
–
–
6, and cis-2+BF4 /trans-2+BF4 .
Characterization of the Complexes trans-3 and cis-3: These com-
plexes were generated in situ, as shown in Scheme 9, from
Pd0(dba)2 (5.8 mg, 0.01 mmol), PPh3 (5.6 mg, 0.02 mmol),
nBu4NCl (2.8 mg, 0.01 mmol), and cis-1 (1.8 μL, 0.01 mmol) in
Acknowledgments
This work was supported by the Centre National de la Recherche
Scientifique, the Ministère de la Recherche et de la Technologie
(Ecole Normale Supérieure), and by projects BQU-04002 (MCyT
of Spain) and 2001-SGR-00181 (DURSI, Generalitat de Catalu-
nya). This collaborative work was performed under the auspices
of COST program D24-0013-02. Johnson Matthey Plc. is greatly
thanked for a loan of Na2PdCl4.
1
0.5 mL of CDCl3. The reaction was monitored by H NMR spec-
troscopy. The isomerization cis-1 to trans-1 was observed first. Af-
ter 7 h, the signals of cis-1 and trans-1 were no longer detectable
1
and the spectrum exhibited the H and 31P NMR signals of cis-3
and trans-3 reported above. A COSY experiment allowed the dis-
crimination between the H1 and H2 protons.
Characterization of the Complexes trans-3 and cis-3: This time these
complexes were generated as shown in Scheme 15. A mixture of
trans-2+·BF4 and cis-2+·BF4 (66:34; 8 mg, 0.01 mmol), was intro-
duced into an NMR tube containing 0.5 mL of CDCl3, followed
by nBu4NCl (28 mg, 0.1 mmol). The 1H and 31P NMR spectra were
the same as those reported just above.
–
–
[1] a) J. Tsuji, I. Shimizu, I. Minami, Y. Ohashi, T. Sugiura, K.
Takahashi, J. Org. Chem. 1985, 50, 1523–1529; b) J. Tsuji, Tet-
rahedron 1986, 42, 4361–4401; c) S. A. Godleski, in Comprehen-
sive Organic Synthesis, vol. 4 (Eds: B. M. Trost, I. Fleming),
Pergamon, Oxford, 1991; d) M. Moreno-Mañas, R. Pleixats, in
Handbook of Organopalladium Chemistry for Organic Synthesis
(Ed.: E.-i. Negishi), Wiley, New York, 2002, vol. II, p. 1707–
1767.
Synthesis of the Complexes trans-3 and cis-3: This time these com-
plexes were generated as shown in Scheme 13. A solution of tri-
phenylphosphane (44.3 mg, 0.169 mmol) in 5 mL of acetone was
added to a stirred solution of the dimeric complexes trans,trans-5
and cis,cis-5 (20 mg, 0.0422 mmol) in 10 mL of acetone. After
30 min, the solvent was evaporated. The complexes trans-3 and cis-
[2] M. Moreno-Mañas, L. Morral, R. Pleixats, J. Org. Chem. 1998,
63, 6160–6166.
[3] C. Amatore, S. Gamez, A. Jutand, G. Meyer, M. Moreno-
Mañas, L. Morral, R. Pleixats, Chem. Eur. J. 2000, 6, 3372–
3376.
1
3 were collected as an orange solid. The H and 31P NMR spectra
were the same as those reported above.
[4] For postulated and established SN2 mechanisms, see: a) T. Tak-
ahashi, Y. Jinbo, K. Kitamura, J. Tsuji, Tetrahedron Lett. 1984,
25, 5921–5924; b) P. B. Mackenzie, J. Whelan, B. Bosnich, J.
Am. Chem. Soc. 1985, 107, 2046–2054; c) H. Kurosawa, S. Ogo-
shi, N. Chatani, Y. Kawasaki, S. Murai, I. Ikeda, Chem. Lett.
1990, 1745–1748; d) J.-E. Bäckvall, K. L. Granberg, A. Heu-
Synthesis of the Complexes trans-6 and cis-6: These complexes were
generated as shown in Scheme 11. A solution of triphenylphos-
phane (22 mg, 0.084 mmol) in 5 mL of acetone was added to a
stirred solution of the dimeric complexes trans,trans-5 and cis,cis-5
(20 mg, 0.0422 mmol) in 10 mL of acetone. After 30 min, the sol-
Eur. J. Org. Chem. 2005, 4277–4286
www.eurjoc.org
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4285

T. Cantat, N. Agenet, A. Jutand, R. Pleixats, M. Moreno-Mañas
FULL PAPER
mann, Isr. J. Chem. 1991, 31, 17–24; e) K. L. Granberg, J.-E.
Bäckvall, J. Am. Chem. Soc. 1992, 114, 6858–6863 and ref. [3].
[5] F. Ozawa, T. Son, S. Ebina, K. Osakada, A. Yamamoto, Orga-
nometallics 1992, 11, 171–176.
[6] a) J. E. Bäckvall, R. E. Nordberg, J. Am. Chem. Soc. 1981, 103,
4959–4960; b) R. E. Nordberg, J. E. Bäckvall, J. Organomet.
Chem. 1985, 285, C24; c) M. Kawatsura, Y. Uozumi, T. Haya-
shi, Chem. Commun. 1998, 217–218; d) G. C. Lloyd-Jones, S. C.
Stephen, Chem. Commun. 1998, 2321–2323; e) G. C. Lloyd-
Jones, S. C. Stephen, Chem. Eur. J. 1998, 4, 2539–2549; f) B. M.
Trost, F. D. Toste, J. Am. Chem. Soc. 1999, 121, 4545–4554; g)
B. Åkermark, G. Åkermark, L. S. Hegedus, K. Zetterberg, J.
Am. Chem. Soc. 1981, 103, 3037–3040.
[13]
a) C. W. Alexander, W. R. Jackson, W. B. Jennings, J. Chem.
Soc. B 1971, 2241–2243; b) K. Dunne, F. J. McQuillin, J. Chem.
Soc. C 1970, 2196–2200.
H. Kurosawa, S. Ogoshi, Y. Kawasaki, S. Murai, M.-a. Mi-
yoshi, I. Ikeda, J. Am. Chem. Soc. 1990, 112, 2813–2814.
[14]
[15]
H. Kurosawa, H. Kajimaru, S. Ogoshi, H. Yoneda, K. Miki,
N. Kasai, S. Murai, I. Ikeda, J. Am. Chem. Soc. 1992, 114,
8417–8424.
+
[16]
Authentic samples of the phosphonium salts trans-7 and cis-
7+ with Cl– as the counteranion were purposely synthesized to
detect their possible formation in the reactions investigated in
this work. The reaction of the allylic chlorides trans-4 and cis-
4 (35:65) with PPh3 in excess (5 equiv.) was extremely slow. Af-
ter 3 d, two singlets were observed in the 31P NMR spectrum
at δ = 25.9 and 24.3 ppm in CDCl3. The formation of the phos-
phonium salts is catalyzed by Pd0(dba)2. With 33% of the cata-
lyst, the allylic chlorides trans-4/cis-4 (35:65) were totally con-
[7] a) C. Amatore, A. Jutand, M. A. M’Barki, G. Meyer, L. Mott-
ier, Eur. J. Inorg. Chem. 2001, 873–880; b) T. Cantat, E. Génin,
C. Giroud, G. Meyer, A. Jutand, J. Organomet. Chem. 2003,
687, 365–376.
[8] For η1-allylpalladium complexes see: a) J. Powell, B. L. Shaw,
J. Chem. Soc. A. 1967, 1839–1849; b) P. Fitton, M. P. Johnson,
J. E. McKeon, Chem. Commun. 1968, 6–7; c) H. Kurosawa, S.
Ogoshi, Bull. Chem. Soc. Jpn. 1998, 71, 973–984; d) P.
Braunstein, F. Naud, A. Dedieu, M.-M. Rohmer, A. DeCian,
S. J. Rettig, Organometallics 2001, 20, 2966–2981; e) M.
Kollmar, G. Helmchen, Organometallics 2002, 21, 4771–4775.
[9] a) C. Amatore, A. Jutand, G. Meyer, Inorg. Chim. Acta 1998,
273, 76–84; b) C. Amatore, A. Jutand, Coord. Chem. Rev. 1998,
1
verted into the phosphonium salts trans-7+ and cis-7+. The H
NMR spectra of the two salts are very similar, although they
could be differentiated by their 31P NMR spectra. However,
since these compounds only appear as by-products, their ste-
reochemistry was not investigated.
[17]
J. Powell, B. L. Shaw, J. Chem. Soc. A 1968, 774–777.
[18] Y. Takahashi, T. Ito, S. Sakai, Y. Ishii, Chem. Commun. 1970,
1065–1066.
178–180, 511–528; c) C. Amatore, A. Jutand, F. Khalil, M. A. [19] D. T. Rosevear, F. G. H. Stone, J. Chem. Soc. A 1968, 164–167.
M’Barki, L. Mottier, Organometallics 1993, 12, 3168–3178.
[10] A. Jutand, Eur. J. Inorg. Chem. 2003, 2017–2040.
[11] a) As the starting reagents are neutral, the effect of the ionic
strength must be in operation at the level of the activated com-
plex; b) P. W. Atkins, Physical Chemistry, Oxford University
Press, 3rd ed., 1986, p. 237–245.
[12] a) P. R. Auburn, P. B. Mackenzie, B. Bosnich, J. Am. Chem.
Soc. 1985, 107, 2033–2046; b) P. B. Mackenzie, J. Whelan, B.
Bosnich, J. Am. Chem. Soc. 1985, 107, 2046–2054.
[20] M. Moreno-Mañas, J. Ribas, A. Virgili, J. Org. Chem. 1988,
53, 5328–5335.
[21] a) R. L. Frank, H. K. Hall, J. Am. Chem. Soc. 1950, 72, 1645–
1648; b) J. P. Blanchard, H. L. Goering, J. Am. Chem. Soc.
1951, 73, 5863–5865.
[22] A. Vallribera, N. Serra, J. Marquet, M. Moreno-Mañas, Tetra-
hedron 1993, 49, 6451–6462.
Received: May 12, 2005
Published Online: August 26, 2005
4286
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
Eur. J. Org. Chem. 2005, 4277–4286