Structural and kinetic effects of chloride ions in the palladium-catalyzed allylic substitutions

Article abstract of DOI:10.1016/S0022-328X(03)00791-5

Addition of ligands to [Pd(η³-RCH-CH-CH₂) (μ-Cl)]₂ or chloride ions to cationic [(η³ -RCH-CH-CH₂)PdL₂] ⁺BF₄ ^- induces the formation of neutral complexes η

Full text of DOI:10.1016/S0022-328X(03)00791-5

Journal of Organometallic Chemistry 687 (2003) 365ꢃ376
/
www.elsevier.com/locate/jorganchem
Structural and kinetic effects of chloride ions in the palladium-
catalyzed allylic substitutions
Thibault Cantat, Emilie Ge´nin, Claire Giroud, Gilbert Meyer, Anny Jutand *
Ecole Normale Supe´rieure, De´partement de Chimie, UMR CNRS-ENS-UPMC 8640, 24, Rue Lhomond, F-75231 Paris Cedex 5, France
Received 27 May 2003; received in revised form 27 May 2003; accepted 29 July 2003
Dedicated to Professor Jean-Pierre Geneˆt on the occasion of his 60th birthday
Abstract
Addition of ligands to [Pd(h3-RCHÃ
formation of neutral complexes h1-RCHÄ
or L2ꢂ1,2-bis(diphenylphosphino)butane (dppb), 1,1?-bis(diphenylphosphino)ferrocene (dppf); Rꢂ
instead of the expected cationic complexes [(h3-RCHÃ CH2)PdL2]ꢀClꢁ. In the presence of chloride ions, the reaction of
CHÃ
morpholine with the cationic complexes [(h3-allyl)Pd(PAr3)2]ꢀBF4ꢁ (Arꢂ
4-ClÃC6H4, 4-CH3ÃC6H4) goes slower and involves both
/
CHÃ
/
CH2)(m-Cl)]2 or chloride ions to cationic [(h3-RCHÃCHÃCH2)PdL2]ꢀBF4ꢁ induces the
CH2ÃPdClL2 (RꢂH with Lꢂ(4-ClÃC6H4)3P, (4-CH3ÃC6H4)3P, (4-CF3ÃC6H4)3P
Ph with Lꢂ(4-ClÃC6H4)3P),
/
CHÃ
/
/
/
/
/
/
/
/
/
/
/
/
/
/
/
/
cationic [(h3-allyl)Pd(PAr3)2]ꢀ and neutral h1-allyl-PdCl(PAr3)2 complexes as reactive species in equilibrium with Clꢁ. The cationic
complex is more reactive than the neutral one. However, their relative contribution in the reaction strongly depends on the chloride
concentration, which controls their relative concentration. The neutral h1-allyl-PdCl(PAr3)2 may become the major reactive species
at high chloride concentration. Consequently, [Pd(h3-allyl)(m-Cl)]2 associated with ligands or cationic [(h3-allyl)PdL2]ꢀBFꢁ4 ; used
indifferently as precursors in palladium-catalyzed allylic substitutions, are not equivalent. In both situations, the mechanism of the
Pd-catalyzed allylic substitution depends on the concentration of the chloride ions, delivered by the precursor or purposely added,
that determines which species, [(h3-allyl)PdL2]ꢀ or/and h1-allyl-PdClL2 are involved in the nucleophilic attack with consequences
on the rate of the reaction and probably on its regioselectivity. Consequently, the chloride ions of the catalytic precursors [Pd(h3-
allyl)(m-Cl)]2 must not be considered as ‘innocent’ ligands.
# 2003 Elsevier B.V. All rights reserved.
Keywords: Allylic substitution; Allyl complex; Chloride ions; Kinetics; Mechanism; Palladium
1. Introduction
The cationic complex [(h3-allyl)PdIIL2]ꢀ is supposed
to be the key intermediate that reacts with the nucleo-
phile to give the substitution product and the active Pd0
catalyst [1]. It is why (h3-allyl)PdII complexes are often
used as catalytic precursors of Pd0 complexes after
reaction with the nucleophile. Two kinds of (h3-allyl)P-
dII are generally used: (i) dimeric [PdII(h3-allyl)(m-Cl)]2
(1) associated with ligands; or (ii) cationic complexes
[(h3-allyl)PdIIL2]ꢀ (2ꢀ) associated with an innocent
Nucleophilic substitutions on allylic acetates are
catalyzed by palladium complexes (Tsujiꢃ
tions (Eq. 1) [1].
The catalytic cycle is initiated by a Pd0L2 complex
(LꢂPAr3 or L2ꢂP,P; P,N ligands) involved in a
/Trost reac-
/
/
reversible oxidative addition which generates a cationic
[(h3-allyl)PdIIL2]ꢀ complexes with AcOꢁ as the counter
anion (Scheme 1) [2].
counter anion Yꢁ (Yꢁ ꢂ
[1].
/
BFꢁ4 ; PFꢁ6 ; TfOꢁ or TsOꢁ)
However, in 2001, we established that precursors 1
and 2ꢀ, Yꢁ were not equivalent [3]. Indeed, complexes
1 (RꢂH, Ph, Scheme 2) associated with 4 equivalent L
/
(PPh3), do not generate the cationic complexes 2ꢀ when
the chloride ions have not been quenched by a cation,
but give neutral complexes h1-allyl-PdIIClL2 (3) in
* Corresponding author. Tel.:
44323325.
E-mail address: anny.jutand@ens.fr (A. Jutand).
ꢀ
/
33-1-44323872; fax:
ꢀ/33-1-
0022-328X/03/$ - see front matter # 2003 Elsevier B.V. All rights reserved.
doi:10.1016/S0022-328X(03)00791-5

366
T. Cantat et al. / Journal of Organometallic Chemistry 687 (2003) 365ꢃ376
/
dimer [Pd(h3-allyl)(m-Cl)]2 with 2 equivalent of the
corresponding P,N ligand, in the absence of any chloride
scavenger. On the other hand, the decomposition of (h3-
allyl)PdII complexes may be induced by the addition of
chloride ions [7].
In 1981, Akermark et al. reported a difference of
reactivity of dimethylamine with an isolated cationic
complex
when compared to the neutral complex h1-CH3Ã
CHÃCH2ÃPdCl(PPh3)2, supposedly generated in situ by
reaction of [PdII(h3-CH3Ã
CHÃCHÃCH2)(m-Cl)]2 with 4
[(h3-CH3ÃCHÃCHÃCH2)Pd(PPh3)2]ꢀBFꢁ4 ;
/ /
CHÄ
/
/
/
/
/
equivalent PPh3. The regioselectivity of the reaction was
also different [8].
When considering palladium-catalyzed allylic substi-
tutions, the efficiency, regioselectivity and enantioselec-
tivity may be affected by the catalytic precursor or by
the presence of chloride ions purposely added, as
evidenced by Backvall [9], Hayashi [10], Lloyd-Jones
¨
[11], and Trost’s groups [12].
Scheme 1.
dynamic equilibrium (route B in Scheme 2). This
reaction with RꢂH and LꢂPPh3 was already reported
by Powell and Shaw in 1967 [4].
/
/
Moreover, we have established that complexes 3 are
always formed in the presence of chloride ions, i.e.: (i)
when chloride ions are added to 2ꢀ; BF4ꢁ (route C in
Scheme 2); (ii) when the oxidative addition of allylic
This incited us to investigate the role of chloride ions
in the possible formation of neutral h1-allyl-PdIIClL2
complexes for different mono- (Lꢂ
/PAr3) or bidentate
bisphosphines (L2ꢂP,P) and their contribution in the
/
acetates to Pd0(dba)2ꢀ
2PPh3 is performed in the
/
mechanism and kinetics of the nucleophilic attack of
morpholine.
presence of Clꢁ (route D in Scheme 2); or (iii) in the
oxidative addition of allylic chlorides to Pd0(PPh3)4
(route E in Scheme 2) [3].
Consequently, when the ligand is PPh3, the postulated
2. Results and discussion
cationic complex [(h3-allyl)Pd(PPh3)2]ꢀXꢁ (Xꢂ
/OAc
or Cl) which is supposed to react with the nucleophile is
not formed as a major complex in the presence of Clꢁ
ions which are either delivered by the catalytic precursor
or purposely added. The major complexes are neutral
h1-allyl-PdCl(PPh3)2.
2.1. Effect of chloride ions on the structure of
(allylic)PdII complexes: h3-allyl versus h1-allyl
coordination
In 2001, Braunstein et al. reported the characteriza-
2.1.1. Formation of neutral h1-allyl-PdClL2 (Lꢂ
monodentate ligand: (4-ZÃC6H4)3P, ZꢂCl, CH3, CF3)
/
tion of a h1-allyl-PdIICl(P,N) complex (P,Nꢂ
zoline)phenylphosphonite) [5]. In 2002, Kollmar and
/bis(oxa-
/
/
The two routes B and C in Scheme 2 were explored
for the synthesis of (allyl)PdII complexes in the presence
of chloride ions. Even if the monophosphine ligands (4-
Helmchen reported the synthesis of a h1-allyl-PdCl(P,N)
(P,Nꢂphosphinooxazoline) complex [6]. In both cases,
/
the h1-allyl-PdCl complexes were obtained by treating a
ZÃ
/
C6H4)3P (ZꢂCl, CH3, CF3) are not widely used as
/
Scheme 2.

T. Cantat et al. / Journal of Organometallic Chemistry 687 (2003) 365ꢃ
/376
367
palladium ligands in catalytic Tsujiꢃ
/
Trost reactions [1],
equilibrium between two h1-coordination modes 3a and
3g (Scheme 2) [3,4].
it was of academic interest to investigate their electronic
effect in the formation of neutral h1-allyl-PdIIClL2 (3).
Indeed, the positive charge on the cationic complexes
[(h3-allyl)PdIIL2]ꢀ (2ꢀ) is affected by the electronic
properties of the phosphine, with some expected con-
sequences on their reactivity with nucleophiles, includ-
ing chloride ions.
The protons of the allyl ligand of h1-CH2Ä
/
CHÃ
C6H4)3P}2 are globally located at upper
field than those of the corresponding cationic complexes
[(h3-CH2Ã C6H4)3P}2]ꢀ (compare
CHÃCH2)Pd{(4-ZÃ
Fig. 1a and b). Indeed, the 1H-NMR signals of the
central H of h1-CH2Ä
CH ÃCH2ÃPdCl{(4-ZÃC6H4)3P}2
(ZꢂCl, CH3, CF3, H) are all located at upper field than
the central H of the cationic [(h3-CH2Ã
CH ÃCH2)Pd{(4-
ZÃ Cl, CH3, CF3, H). Similarly, the
C6H4)3P}2]ꢀ (Zꢂ
signal of the four magnetically equivalent terminal H of
h1-CH2Ä
CHÃCH2ÃPdCl{(4-ZÃC6H4)3P}2 are located
at upper field than the anti-H of the cationic [(h3-CH2Ã
CHÃCH2)Pd{(4-ZÃ Cl, CF3) (itself
C6H4)3P}2]ꢀ (Zꢂ
always located at higher field than the syn-H) or very
close to the anti-H of the cationic [(h3-CH2Ã
CHÃ
CH2)Pd{(4-ZÃ CH3, H). This shift
C6H4)3P}2]ꢀ (Zꢂ
/
CH2Ã
/
PdCl{(4-ZÃ
/
/
/
/
/
/
/
/
/
As for PPh3, addition of 4 equivalent (4-ZÃ
/
C6H4)3P
CHÃ
Cl, CH3, CF3) to the dimer [Pd(h3-CH2Ã
/
/
/
/
(Zꢂ
/
/
/
CH2)(m-Cl)]2 (1) in acetone or chloroform (route B of
Scheme 2) resulted in the complete disappearance of the
dimer with formation of the neutral complexes h1-CH2Ä
/
/
/
/
/
/
CHÃ
/
CH2Ã
/
PdCl{(4-ZÃ
/
C6H4)3P}2 (3a, Zꢂ
/
Cl; 3b, Zꢂ
/
/
/
/
CH3; and 3c, Zꢂ
/
CF3, respectively). This is evidenced
by the presence in their 1H-NMR spectrum
performed in chloroform or acetone of only two
sets of signals for the protons of the allyl group
(Fig. 1a) (instead of the three sets of signals
observed for the corresponding cationic complexes
/
/
/
/
towards upper field is indicative of the coordination of
the chloride to the PdII center. This implicates a h1-
CH2Ä
/
CHÃ
/
CH2Ã
/
coordination to a PdIIL2 moieties
[(h3-CH2ÃCHÃCH2)Pdf(4-ZÃC6H4)3Pg2]ꢀBFꢁ4
(2aꢀ,
because no free phosphine was detected in the 31P-
NMR spectra [13]. Traces of the phosphonium salt [(4-
Zꢂ
/
Cl; 2bꢀ, Zꢂ
/
CH3; and 2cꢀ, Zꢂ
CF3) in Fig. 1b):
/
ZÃ
the 1H-NMR spectrum of 3aꢃ
considered [3,14].
/
C6H4)3PÃ
/
CH2Ã
/
CHÄ
/
CH2]ꢀ were also observed in
c, as when PPh3 was
(i) one well defined quintet at low field integrating for
one H; and (ii) one doublet that may be broad, located
at upper field and integrating for 4H (Fig. 1a). The
former characterizes the central H of the ligand h1-
/
1
It was also observed by H-NMR spectroscopy that
addition of 1 equivalent Clꢁ (introduced as nBu4NCl)
CH2Ä
terminal H of the ligand h1-CH2Ä
become magnetically equivalent due to a fast dynamic
/
CH Ã/CH2 whereas the latter characterizes the four
to a solution of the cationic complex [(h3-CH2Ã
/
CHÃ
BFꢁ4 ); 17 mM in
chloroform, led to the neutral complex h1-CH2Ä
CHÃ
CH2ÃPdCl{(4-ClÃC6H4)3P}2 (3a) (route C in Scheme
2).
To confirm the formation of the neutral complexes
h1-CH2Ä
CHÃCH2ÃPdCl(PAr3)2 instead of the cationic
ones [(h3-CH2Ã Pd(PAr3)2]ꢀClꢁ, conductiv-
CHÃCH2Ã
ity measurements were performed in DMF in
which free ions are usually generated [15].
solution of the isolated complex h1-CH2Ä
CHÃCH2Ã
/
/
CHÃCH2 which
/
CH2)Pd{(4-ClÃ
/
C6H4)3P}2]ꢀ (2aꢀ,
/
/
/
/
/
/
/
/
/
/
/
A
/
/
/
PdCl{(4-ClÃ
/C6H4)3P}2 (3a), 2 mM in DMF, exhibited
a low conductivity of 7 ms cmꢁ1 at 25 8C. By
comparison the conductivity of the cationic complex
[(h3-CH2ÃCHÃCH2)Pdf(4-ZÃC6H4)3Pg2]ꢀBFꢁ4
(2aꢀ,
/
BFꢁ4 ); under the same conditions, was 69 ms cmꢁ1
This allows the calculation of the theoretical conductiv-
.
ity of the putative [(h3-CH2Ã
/
CHÃ
/
CH2)Pd{(4-ClÃ
/
C6H4)3P}2]ꢀClꢁ to be close to 64 ms cmꢁ1 [16].
Consequently, one can assume that a neutral complex
h1-CH2Ä
/
CHÃ
/
CH2Ã
/
PdCl{(4-ClÃ/C6H4)3P}2 (3a) is in-
deed generated in the presence of chloride ions (routes B
and C of Scheme 2).
Similarly, when
were added to the dimer [Pd(h3-PhÃ
Cl)]2 (1?) in the absence of a chloride scavenger (route
in Scheme 2), the neutral complex h1-
CHÄCHÃCH2ÃPdCl{(4-Cl-ÃC6H4)3P}2 (3a?) was
4
equivalent (4-ClÃ
/
C6H4)3P
/
CHÃCHÃ
/
/
CH2)(m-
Fig.
(250
1. 1H-NMR
MHz,
spectroscopy
Protons
performed
of the
in
allyl
acetone-d6
ligand
TMS).
of: (a) h1-CH2Ä
/
CHÃ
/
CH2Ã
/
PdCl{(4-ClÃ
/
C6H4)3P}2 (3a); (b)
2aꢀ; BFꢁ4 ):
B
PhÃ
[(h3-CH2ÃCHÃCH2)Pdf(4-ClÃC6H4)3Pg2]ꢀBFꢁ4
(/
/
/
/
/
/
/

368
T. Cantat et al. / Journal of Organometallic Chemistry 687 (2003) 365ꢃ376
/
generated, as well as when 1 equivalent Clꢁ was
CH2Cl) but higher than the JHH coupling between an
allylic and a vinylic proton (e.g. 7 Hz in PhCHÄCH Ã
added to
a
solution of the cationic complex
/
/
[ (h3-PhÃCHÃCHÃCH2) Pd f(4-ClÃC6H4)3Pg2]ꢀ BFꢁ4
CH2Cl). Moreover, the allylic CH2 protons of the
neutral complex (3.15 ppm in Fig. 2a) became magne-
tically equivalent when compared to the two differen-
tiated anti-H and syn-H of the cationic complex (Fig.
2b). Their JHH coupling with the central H was 9 Hz,
which is higher than the JHH coupling between an allylic
(2a?ꢀ,BFꢁ4 ) in chloroform (route C in Scheme 2). In
/
both cases, 1? and 2a?ꢀ,
in the allyl series, the 1H-NMR signals of the protons of
the allyl ligand of h1-PhÃ
CHÄCHÃCH2ÃPdCl{(4-ClÃ
/
BFꢁ4 were totally converted. As
/
/
/
/
/
C6H4)3P}2 (3a?) were located at higher field
(Fig. 2a) than those of the cationic complex
[ (h3-PhÃCHÃCHÃCH2) Pd f(4-ClÃC6H4)3Pg2]ꢀ BFꢁ4
and a vinylic proton (e.g. 7 Hz in PhCHÄ
/
CH Ã
/
CH2Cl)
due to the contribution of the 3g structure. Conse-
quently, the dynamic equilibrium between 3g and 3a
observed in the allyl series also operates in the cinnamyl
(2a?ꢀ,BFꢁ4 ) (Fig. 2b).
/
Whereas in the cinnamyl chloride (E)-PhCHÄ
CH2Cl, the signal of the proton PhCH was located at
lower field (6.63 ppm) than that of the central PhCHÄ
CH (6.30 ppm), the situation was reversed
/
CHÃ
/
1
series (Scheme 2, Rꢂ
/
Ph) and the observed H-NMR
spectrum assigned to (3a?) is in fact an average spectrum
due to a fast dynamic equilibrium between 3g and 3a
/
in the complex h1-PhÃ
/
CHÄ
/
CHÃ
/
CH2Ã
/
PdCl{(4-ClÃ
/
(Scheme 2, Rꢂ
This was ultimately confirmed by the fact that the
cationic complex (2a?ꢀ,BF4ꢁ) was characterized in 31P-
/H).
C6H4)3P}2 (Fig. 2a). This suggests that the PhCH
proton in the latter neutral complex was shifted to
/
upper field because of a PdII coordination, as in h1-PhÃ
/
NMR spectroscopy by two doublets due to non
equivalent phosphine ligands whereas the complex
generated in the presence of chloride ions exhibited
only one singlet attesting for 2 equivalent ligands sitting
in a trans position in the neutral complexes 3g and 3a
CH(PdClL2)Ã
dynamic equilibrium between two h1 coordination of
the cinnamyl ligand (3g and 3a in Scheme 2, RꢂH).
/
CHÄ/CH2, and the involvement of a fast
/
This was confirmed by the fact that, in the neutral
complex, the JHH coupling between PhCH and the
central H was 13 Hz, an average value that is lower than
(Scheme 2, RꢂPh). Although JPH couplings were
/
observed for the protons of the allyl ligand of the
cationic (h3-allyl)PdII complexes (Fig. 2b), no JPH
coupling was detectable in the h1-allyl-PdII chloride
complexes (Fig. 2a). This suggests that either the
dynamic equilibrium between 3g and 3a occurs via the
dissociation of the phosphine ligand or that the JPH
coupling is too small to be detected at 250 MHz because
of the cis coordination of the phosphine and the
cinnamyl group. As observed in Fig. 2a, the phospho-
nium salt [cinnamyl-P(4-Cl-C6H4)3]ꢀ was not formed in
appreciable amount in contrast with [cinnamyl-PPh3]ꢀ
[3,14].
a trans JHH coupling (e.g. 15.5 Hz in (E)-PhCH Ä
/
CH Ã
/
2.1.2. Formation of neutral h1-allyl-PdCl(P,P) (P,Pꢂ
bidentate ligand: dppb, dppf)
/
The route B in Scheme 3 was explored for the
synthesis of (allyl)PdII complexes ligated by bispho-
sphines, in the presence of chloride ions.
The complexes h1-CH2Ä
and h1-CH2Ä
CHÃCH2Ã
/
CH-CH2-PdCl(dppb) (5d)
/
/
/
PdCl(dppf) (5e) were synthe-
sized by addition of respectively 2 equivalent 1,2-
bis(diphenylphosphino)butane (dppb) or 1,1?-bis(diphe-
nylphosphino)ferrocene (dppf) to the dimer [Pd(h3-
CH2Ã
/
CHÃCH2)(m-Cl)]2 (1). As in the monodentate
/
1
series, the H-NMR spectrum performed in CDCl3 or
acetone-d6 exhibited only two sets of signals for the
allylic protons: (i) one well defined quintet at low field
integrating for one H, the central H of the ligand h1-
Fig. 2. 1H-NMR spectroscopy performed in CDCl3 (250 MHz, TMS).
CH2Ä
higher field and integrating for 4H which characterizes
the four terminal H of h1-CH2Ä
CHÃCH2 that became
/
CH Ã
/
CH2; and (ii) one broad multiplet, located at
Protons of the allyl ligand of: (a) h1-PhCHÄ
/
CHÃ
/
CH2Ã
/
PdCl{(4-ClÃ
/
C6H4)3P}2 (3a?); (b) [(h3-PhCHÃCHÃCH2)Pdf(4-ClÃC6H4)3Pg2]ꢀBFꢁ4
(/
2a?ꢀ; BFꢁ4 ):
/
/
/

T. Cantat et al. / Journal of Organometallic Chemistry 687 (2003) 365ꢃ
/376
369
Scheme 3.
Scheme 4.
(
2aꢀ; BFꢁ4 ); performed in DMF at room temperature,
/
magnetically equivalent due to a fast dynamic equili-
brium (5g and 5a in Scheme 3).
produced the substitution product 10 (Scheme 4),
identified by 1H-NMR spectroscopy [21]. In a first
approach, the simple nonsubstituted allyl complex was
investigated to by-pass regioselectivity problems.
A solution of h1-CH2Ä
/
CHÃ
/
CH2Ã/PdCl(dppb) (5d), 2
mM in DMF, exhibited a low conductivity of 12 ms
cmꢁ1 [18]. The theoretical conductivity of the putative
[(h3-CH2Ã
/
CHÃ
that could have been formed (route G in Scheme 3), was
estimated from the known conductivity of [(h3-CH2Ã
CHÃ
CH2)Pd(dppb)]ꢀAcOꢁ [2a] and found to be close
to 124 ms cmꢁ1 [19]. Consequently, one can assume that
a neutral complex h1-CH2Ä
CHÃCH2ÃPdCl(dppb) (3d),
/
CH2)Pd(dppb)]ꢀClꢁ, 2 mM in DMF,
The reaction was first monitored by cyclic voltam-
metry. The complex 2aꢀ; BFꢁ4 (2 mM) in DMF (con-
taining nBu4NBF4, 0.3 M) exhibited a reduction peak
/
/
R1 at Ep ꢂꢁ1:39 V versus SCE and no oxidation peak.
R1
After addition of 20 equivalent morpholine, a new
reduction peak R2 appeared at less negative potential:
/
/
/
is indeed generated in the presence of chloride ions
(route B in Scheme 3) instead of the cationic complex,
even in a dissociative solvent as DMF [15]. This was
EpR ꢂꢁ1:02 V; while the reduction peak R1 of
2
2aꢀ; BFꢁ4 was no longer observed. When the potential
was scanned to oxidation first, an oxidation peak O3
1
ultimately checked by running a H-NMR spectrum in
DMF-d7 which exhibited the quintet and the doublet as
unique signals characteristic of the neutral complex h1-
was detected at EpO ꢂꢀ0:25 V: From these voltammo-
3
grams, one concludes that the morpholine reacts with
2aꢀ; BFꢁ4 (its reduction peak R1, whose reduction peak
current is proportional to its concentration, was no
longer detected) to generate a Pd0 complex characterized
by its oxidation peak O3. This reaction also generates a
species, which is reduced at R2. The latter was identified
as the protonated morpholine 9 (Scheme 4). The
presence of 9 more easily reduced than 2aꢀ; BF4ꢁ did
not allow the investigation of the kinetics of the reaction
of 2aꢀ; BFꢁ4 with morpholine, by monitoring the decay
of its reduction peak current versus time. Moreover, the
reaction was too fast at 25 8C to be monitored by cyclic
voltammetry [22].
CH2Ä
/
CHÃ
/
CH2Ã
/
PdCl(dppb), as observed in chloro-
form and acetone (vide supra).
2.2. Effect of chloride ions on the kinetics of the
nucleophilic attack on allylic-PdII complexes
Cationic complexes [(h3-allyl)PdL2]ꢀ [20] are postu-
lated to be the reactive species in the nucleophilic attack
(Scheme 1) [1]. As evidenced above, cationic complexes
[(h3-allyl)PdL2]ꢀBFꢁ4 (Lꢂ
PAr3) are easily transformed
/
to the neutral complexes h1-allyl-PdClL2 in the presence
of chloride ions. Consequently, it was of interest to
investigate the effect of chloride ions on the kinetics of
the nucleophilic reaction.
The
kinetics
of
the
nucleophilic
attack
was thus investigated by UV spectroscopy as
reported by Kuhn and Mayr for the cationic complex
[(h3-PhÃCHÃCHÃCH2)Pd(PPh3)2]ꢀBFꢁ4 in dichloro-
methane [23]. In that case, the decay of the absorbance
of [(h3-PhÃCHÃCHÃCH2)Pd(PPh3)2]ꢀBFꢁ4 in the
presence of morpholine was monitored versus
2.2.1. Kinetics of the reaction of morpholine with the
cationic complexes [(h3-CH2ÃCHÃCH2)PdL2]ꢀBF4ꢁ
(Lꢂ
/
(4-ClÃ
/
C6H4)3P, (4-CH3Ã
/
C6H4)3P) in the absence
time.
However,
since
the
complex
of chloride ions
Morpholine was chosen as nucleophile. Its reaction
with
[ (h3-PhÃCHÃCHÃCH2) Pd f(4-ClÃC6H4)3Pg2]ꢀ BFꢁ4
[(h3-CH2ÃCHÃCH2)Pdf(4-ClÃC6H4)3Pg2]ꢀBFꢁ4
(
2aꢀ; BFꢁ4 ) did not exhibit any absorbance in the range
/

370
T. Cantat et al. / Journal of Organometallic Chemistry 687 (2003) 365ꢃ376
/
300ꢃ800 nm, another strategy was chosen. After addi-
/
tion of 10 equivalent morpholine to 2aꢀ; BF4ꢁ (0.5 mM)
in DMF at 25 8C, an absorption band appeared at 323
nm (absorbance Dꢂ
complex 8 (Scheme 4), as attested by the UV spectro-
scopy performed on a solution of Pd0{(4-ClÃ
C6H4)3P}4
and the substitution product 10.
/
0.34) which characterizes the Pd0
/
When less than 2 equivalent morpholine was added to
2aꢀ; BFꢁ4 (0.5 mM) in DMF, the absorbance of the Pd0
complex 8 was less than 0.34. One additional equivalent
morpholine was required to observe the absorbance of
0.34 obtained when a large excess morpholine was used.
Addition of more morpholine did not affect the final
absorbance. This shows that the complete substitution
on 2aꢀ (Eq. 4 in Scheme 4) requires 2 equivalent
morpholine. The deprotonation involved in a second
step (Eq. 3 in Scheme 4) is thus faster than the
nucleophilic reaction (Eq. 2 in Scheme 4). Consequently,
the rate of formation of the Pd0 complex 8 (Eq. 4) which
could easily be monitored by UV spectroscopy, is
indicative of the rate of the nucleophilic attack (Eq. 2).
In order to anticipate any effect of the ionic strength
that might occur when the nucleophilic attack of
morpholine will be investigated in the presence of Clꢁ
introduced as nBu4NCl (free ions in DMF, vide supra),
the kinetics of reaction (4) was investigated in the
absence and presence of nBu4NBF4 (0.3 M), i.e. at
low and high ionic strength. To observe reasonable
reaction times, the kinetics of the reaction in Eq. 4 (rate
Fig. 3. Kinetics of the reaction of morpholine with the cationic
complex [(h3-CH2ÃCHÃCH2)Pd(PAr3)2]ꢀBFꢁ4 (Arꢂ
4-ClÃC6H4)
2aꢀ; BFꢁ4 ) (0.5 mM) in DMF containing nBu4NBF4 (0.3 M) at
25 8C, monitored by UV spectroscopy. (a) Absorbance at 323 mm
of the Pd0 complex 8 versus time, after addition of 10 equivalent
morpholine; (b) variation of ln x versus time (xꢂ([8]lim [8])/[8]lim
(Dlim D)/Dlim with D: absorbance of at and Dlim final
absorbance of 8 determined in Fig. 1a). ln xꢂ kexpt; (c) reaction
order in morpholine: plot of kexp versus the morpholine concentration.
kexp kh3[morpholine].
/
/
(
/
ꢁ
/
/
ꢁ
/
ꢂ/
constant: kh3) was performed at ꢁ25 8C. Fig. 3a
/
ꢁ
/
8
t
:
exhibits the absorbance of the Pd0 complex 8, generated
after addition of 10 equivalent morpholine to 2aꢀ; BF4ꢁ
(0.5 mM) in DMF, versus time.
/
ꢁ
/
ꢂ/
According to the mechanism postulated in Scheme 4,
the kinetic law of reaction (4) is given in given in Eq. (5)
(M: morpholine).
determined in the absence of nBu4NBF4: kh3
ꢂ
4.2 Mꢁ1
/
d[2aꢀ]=dtꢂꢁd[8]=dtꢂꢁkh3[2aꢀ][M]
(5)
sꢁ1 (DMF, ꢁ
/
25 8C). Comparison of the two values
In the presence of excess morpholine (pseudo first
order conditions), the integration of Eq. (5) gives Eq. (6)
indicates that the effect of the increasing ionic strength
on the rate of the nucleophilic attack of morpholine on
the cationic complex 2aꢀ was relatively low (accelerat-
ing effect of 30%) despite the fact that a cationic PdII
complex was involved in the reaction.
(xꢂ/([8]lim
ꢁ/[8])/[8]lim
ꢂ/(Dlim
ꢁD)/Dlim with D: absor-
/
bance of 8 at t and Dlim: final absorbance of 8).
lnxꢂꢁkh3[M]tꢂꢁkexp
t
(6)
The reaction of morpholine (2 mM) with the cationic
complex
The plot of ln x versus time was linear (Fig. 3b):
ln xꢂ kexpt. Plotting kexp versus morpholine concen-
/
ꢁ
/
[(h3-CH2ÃCHÃCH2)Pdf(4-MeÃC6H4)3Pg2]ꢀBFꢁ4
tration afforded a straight line (Fig. 3c), which
confirms that the reaction order in morpholine is
one in agreement with the mechanism of Scheme 4:
(
2bꢀ; BFꢁ4 ) (1 mM) was investigated at 25 8C by UV
spectroscopy in the absence of nBu4NBF4 [24]. The
/
reaction was found to be slower than that with
2aꢀ; BFꢁ4 ; with a rate constant of kh3
kexp
ꢂkh3[M]. This reaction order was also found by
/
ꢂ
1 Mꢁ1 sꢁ1
/
Kuhn and Mayr for the related complex
[(h3-PhÃCHÃCHÃCH2)Pd(PPh3)2]ꢀBFꢁ4 in dichloro-
methane [23].
The value of kh3 was then calculated from the slope of
the straight line in Fig. 3c: kh3
(DMF, 25 8C) (Fig. 4, Table 1) [25]. This is consistent
with the fact that the ligand (4-Me-C6H4)3P of the
cationic PdII center is more electron rich than (4-ClÃ
C6H4)3P, which disfavors the external nucleophilic
/
ꢂ
/
5.6 Mꢁ1 sꢁ1 (DMF,
ꢁ
/
25 8C, nBu4NBF4, 0.3 M). The rate constant was also
attack on the allyl ligand [1].

T. Cantat et al. / Journal of Organometallic Chemistry 687 (2003) 365ꢃ
/376
371
Fig. 4. Kinetics of the reaction of morpholine (2 mM) with the
cationic complex [(h3-CH2ÃCHÃCH2)Pd(PAr3)2]ꢀBFꢁ4 (Arꢂ
4-Me-
C6H4) ( 1 mM) in DMF at 25 8C, monitored by UV
2bꢀ; BFꢁ4 ) (C0ꢂ
spectroscopy at 296 mm. Variation of 1/x versus time (xꢂ([8]lim
[8])/[8]lim (Dlim D)/Dlim with D: absorbance of 8 at t and Dlim: final
absorbance of 8. 1/xꢂ2kh3C0tꢀ1 [25].
/
/
/
/
ꢁ/
ꢂ/
ꢁ/
/
/
Scheme 5.
Table 1
Comparative intrinsic reactivity of cationic [(h3-allyl)PdL2]ꢀ and
neutral h1-allyl-PdClL2 complexes with morpholine in DMF at
25 8C (Scheme 5)
(C0ꢂ0.5 mM). It was of interest to determine whether
/
the cationic complex 2aꢀ remained the reactive complex
even in the presence of a large excess of Clꢁ or whether
both complexes 2aꢀ and 3a react in parallel with
morpholine to give the same substitution product 10
(Scheme 5). The reaction of morpholine with 2aꢀ is
known to be an external attack onto the h3-allyl ligand
[1] whereas reaction of morpholine with 3a would be an
attack of the h1-allyl ligand by a SN2? substitution, as
proposed by Akermark et al. [8,26].
Lꢂ
/
(4-ZÃ
/
C6H4)3P kh1 (Mꢁ1 sꢁ1
)
kh3 (Mꢁ1 sꢁ1
)
KCl (Mꢁ1
)
Zꢂ
/
Cl
0.91
0.29
94
1
58ꢄ
103
/
Zꢂ
/
Me
2.4ꢄ
103
/
2.2.2. Kinetics of the reaction of morpholine with the
cationic complexes [(h3-CH2ÃCHÃCH2)PdL2]ꢀBF4ꢁ
According to Scheme 5, the kinetic law is given in Eq.
(7) (M: morpholine).
(Lꢂ
of chloride ions
The effect of chloride ions on the reactivity of
2aꢀ; BFꢁ4 (C0ꢂ
0.5 mM) with morpholine was investi-
/
(4-ClÃ
/
C6H4)3P); (4-CH3Ã/C6H4)3P)) in presence
d[8]=dtꢂkh3[2aꢀ][M]ꢀkh1[3a][M]
(7)
/
gated in DMF by UV spectroscopy, by monitoring the
formation of the Pd0 complex 8 versus time. The
chloride ions were introduced as nBu4NCl, in the
Eq. (8) is obtained upon considering steady state
approximation for complex 2aꢀ.
ꢀ
ꢁ
d[8]
dt
kh3kꢁCl
kꢀCl[Clꢁ] ꢀ kh3[M]
concentration range of 4.9ꢃ8.8 mM to minimize any
/
ꢂ[3a][M] kh1 ꢀ
(8)
effect of the variation of the ionic strength (vide supra).
The nucleophilic attack went slower and slower when
the Clꢁ concentration was increased. This cannot be
interpreted as a consequence of the increasing ionic
strength, since the effect of the ionic strength was to
slightly accelerate the nucleophilic attack (vide supra). A
specific effect of nBu4NCl must thus be considered.
As established in Section 2.1.1, the addition of
Clꢁ as nBu4NCl to the cationic complex
[ (h3 -CH2 ÃCHÃCH2) Pd f(4-ClÃC6 H4)3 Pg2 ]ꢀ BFꢁ4
It simplifies into Eq. (9) when considering that
kꢀCl[Clꢁ]ꢀ
/
kh3[M].
ꢀ
ꢁ
d[8]
dt
kh3kꢁCl
kꢀCl[Clꢁ]
ꢂ[3a][M] kh1 ꢀ
ꢀ
ꢁ
kh3
KCl[Clꢁ]
ꢂ[3a][M] kh1 ꢀ
(9)
The integration of Eq. (9) gives Eq. (10) (when [Clꢁ]ꢁ
10C0) with an apparent rate constant kapp expressed in
Eq. (11).
/
(
2aꢀ; BFꢁ4 ) induced the formation of the neutral com-
/
plex h1-CH2Ä
/
CHÃ
/
CH2Ã
/
PdCl{(4-ClÃ
/
C6H4)3P)}2 (3a).
It was checked by 1H-NMR spectroscopy in chloroform
that reaction of the isolated neutral complex 3a with
morpholine gave the substitution product 10. However,
the fact that the rate of the nucleophilic attack on the
cationic complex 2aꢀ; BFꢁ4 depends on the Clꢁ con-
centration suggests that the two complexes 2aꢀ and 3a
are in equilibrium with the Clꢁ (Scheme 5) with
ꢀ
ꢁ
kh3
KCl[Clꢁ]
ln xꢂ[M] kh1 ꢀ
t
(10)
(11)
kh3
KCl[Clꢁ]
kapp ꢂkh1 ꢀ
The rate constant kapp was determined by plotting
ln x versus time (xꢂ([8]lim [8])/[8]lim (Dlim D)/
Dlim with D: absorbance of 8 at t and Dlim: final
KCl[Clꢁ]ꢀ
/
1, because the cationic complex 2aꢀ was
/
ꢁ
/
ꢂ
/
ꢁ
/
no longer detected in the presence of 1 equivalent Clꢁ

372
T. Cantat et al. / Journal of Organometallic Chemistry 687 (2003) 365ꢃ376
/
absorbance of 8) for different concentrations of Clꢁ,
added to the cationic complex 2aꢀ (C0ꢂ
0.5 mM) in
value of kh1). But if [Clꢁ]ꢂ
/
5ꢄ
10ꢁ3 M, the contribu-
/
/
tion of the cationic complex 2aꢀ will be only 0.35 of
that of the neutral complex 3a which then plays the
essential role in the reaction with morpholine.
DMF just before the addition of morpholine (15
equivalent). Since the nucleophilic attack was consider-
ably slower in the presence of Clꢁ, the reactions were
performed at 25 8C.
Consequently, depending on the chloride concentra-
tion, one may then switch from a nucleophilic attack on
the h3-allyl ligand of the cationic complex to a SN2?
substitution on the h1-allyl ligand of the neutral
complex, which may affect the rate of the reaction and
probably the regioselectivity of the reaction. This double
effect was observed by Akermark et al. when reacting
dimethylamine independently with an isolated cationic
complex [(h3-CH3ÃCHÃCHÃCH2)Pd(PPh3)2]ꢀBFꢁ and
The plot of kapp versus the reciprocal of the Clꢁ
concentration was linear (Fig. 5a). This confirms the
mechanism proposed in Scheme 5. kh1 was determined
from the intercept and kh3/KCl from the slope.
kh1 ꢂ0:91 Mꢁ1 sꢁ1 kh3=KCl ꢂ1:6ꢄ10ꢁ3 sꢁ1
with
a
neutral complex h1-CH3Ã
/
CHÄ
/
CHÃ4
/
CH2Ã
/
The value of kh3 was determined at 25 8C under
stoichiometric conditions (2 equivalent morpholine) and
ꢂ
94 Mꢁ1 sꢁ1
/
PdCl(PPh3)2, supposedly generated in situ by reaction
of [PdII(h3-CH3Ã
CHÃCHÃCH2)(m-Cl)]2 with 4 equiva-
in the absence of chloride ions: kh3
(25 8C). The value of the equilibrium constant KCl was
then calculated: KClꢂ5.8ꢄ
104 Mꢁ1 (DMF, 25 8C).
/
/
/
lent PPh3 [8]. However, in this work, neither the
equilibrium between the cationic and the neutral com-
plex was considered, nor the effect of the chloride
concentration.
/
/
From the values of the equilibrium and rate con-
stants, it comes out that the equilibrium between the
cationic complex 2aꢀ and the neutral complex 3a lies in
favor of 3a (for [Clꢁ]ꢂ
/
0.5 mM, then [3a]/[2aꢀ]ꢂ
30).
/
Similarly, the rate of the reaction of morpholine (2
equivalent)
with
the
cationic
complex
Since the intercept of the straight line of Fig. 5a differs
from zero, it ensues that the neutral complex 3a reacts in
parallel with the cationic complex 2aꢀ (Scheme 5). The
cationic complex 2aꢀ is intrinsically considerably more
reactive than the neutral complex 3a (by a factor 100).
However, in the presence of Clꢁ, its concentration may
be very low and the neutral complex 3a may become the
[ (h3 -CH2 ÃCHÃCH2) Pd f(4-MeÃC6 H4)3 Pg2]ꢀ BFꢁ4
(/
2bꢀ; BFꢁ4 ) (C0ꢂ
1 mM) in DMF decreased upon
/
addition of chloride ions at 25 8C. The equilibrium
constant KCl and the rate constant kh1 were determined
as for complex 2aꢀ; BF4ꢁ (vide supra) in DMF at 25 8C
(Fig. 5b, Table 1).
main reactive complex. Indeed, if [Clꢁ]ꢂ
/
5ꢄ
10ꢁ4 M,
/
Comparison of that values to those found for the
C6H4)3P (Table 1) shows that the forma-
the contribution of the cationic complex 2aꢀ (expressed
by the value of kh3/KCl[Clꢁ]) will be 3.5 times higher
than that of the neutral complex 3a (expressed by the
ligand (4-ClÃ
/
tion of the neutral complex h1-allyl-PdClL2, expressed
by the value of KCl, is less favored than for (4-ClÃ
/
C6H4)3P. This is consistent with the fact that the positive
charge on the PdII is lower when the more electron rich
phosphine (4-MeÃ
/
C6H4)3P is considered.
Comparison of the kh3 values shows that, as expected,
the cationic [(h3-allyl)Pd(PAr3)2]ꢀ is less reactive when
the electron rich (4-MeÃ
/
C6H4)3P is considered com-
pared to (4-ClÃC6H4)3P because the allyl ligand is less
/
electron deficient. Comparison of the kh1 values shows
that the neutral h1-allyl-PdCl(PAr3)2 complex is only
slightly less reactive when the electron rich (4-MeÃ
/
C6H4)3P is considered compared to (4-ClÃC6H4)3P.
/
This is in favor of a SN2? mechanism with Pd(PAr3)2
as the leaving group located far from the impact of the
nucleophilic attack.
3. Conclusion
Fig. 5. Kinetics of the reaction of morpholine with the cationic
complexes [(h3-CH2ÃCHÃCH2)Pd(PAr3)2]ꢀBFꢁ4 in DMF in the pre-
sence of various amount of Clꢁ ions added as nBu4NCl at 25 8C,
monitored by UV spectroscopy. Variation of kapp versus the reciprocal
The complexes [Pd(h3-allyl)(m-Cl)]2 or cationic
[(h3-allyl)PdL2]ꢀBFꢁ4 used as precursors in Pd-catalyzed
allylic substitutions are not equivalent. Indeed, addition
of ligand to the former induces the formation of neutral
of Clꢁ concentration (see Eq. (11)). (a) Arꢂ
C0ꢂ0.5 mM), morpholine (15 equivalent); (b) Arꢂ
2bꢀ; BFꢁ4 ; C0ꢂ
1 mM), morpholine (2 equivalent).
/
4-ClÃ
/
C6H4
(
/
2aꢀ; BFꢁ4
C6H4
;
complexes h1-allyl-PdClL2 (Lꢂ
CH3ÃC6H4)3P, (4-CF3ÃC6H4)3P; L2ꢂ
/
(4-ClÃ
/
C6H4)3P, (4-
dppb, dppf) in-
/
/
4-MeÃ
/
(/
/
/
/
/

T. Cantat et al. / Journal of Organometallic Chemistry 687 (2003) 365ꢃ
/376
373
stead of the expected cationic complexes [(h3-
allyl)PdL2]ꢀClꢁ. Those neutral complexes h1-allyl-
PdClL2 are also formed when chloride ions are added
to [(h3-allyl)PdL2]ꢀBFꢁ4 : The neutral complexes h1-
allyl-PdClL2 are in equilibrium with the cationic com-
CH3, CF3), dppb, dppf (Aldrich and Strem) and the
anhydrous nBu4NCl (Aldrich) were commercial. The
allyl chloride, cinnamyl chloride, morpholine were
commercial (Aldrich) and used after filtration on
alumina. Pd0{(4-ClÃ
/
C6H4)3P}4 [28], [Pd(h3-CH2Ã
CHÃCH2)(m-Cl)]2
[4,29] were synthesized according to published proce-
/
CHÃ
/
plexes [(h3-allyl)PdL2]ꢀ and Clꢁ (Lꢂ
(4-CH3ÃC6H4)3P).
The reaction of morpholine with the cationic com-
plexes [(h3-allyl)Pd(PAr3)2]ꢀBFꢁ4 (Arꢂ
4-ClÃC6H4, 4-
CH3ÃC6H4) goes slower when performed in the presence
/
(4-ClÃ/C6H4)3P,
CH2)(m-Cl)]2 and [Pd(h3-PhCHÃ
/
/
/
dures.
The 31P-NMR spectra were recorded on a Bruker
spectrometer (103 MHz) using H3PO4 as an external
1
reference, the H-NMR spectra on a Bruker spectro-
/
/
/
of chloride ions. The reaction involves both cationic
[(h3-allyl)Pd(PAr3)2]ꢀ and neutral h1-allyl-PdCl(PAr3)2
complexes, as reactive species. The cationic complex is
more reactive than the neutral one. Since the two
complexes are involved in equilibrium with the chloride
ions, their relative contribution in the reaction strongly
depends on the chloride concentration, which controls
their relative concentration. At high chloride concentra-
tion the neutral h1-allyl-PdCl(PAr3)2 becomes the major
reactive species.
Consequently, if the catalytic precursor of a Pd-
catalyzed allylic substitution is a dimeric complex
[Pd(h3-allyl)(m-Cl)]2 associated with monodentate phos-
phine ligands, in the absence of any chloride ions
scavenger, the mechanism of the catalytic reaction may
involve the neutral complex h1-allyl-PdClL2 as the main
reactive one and the role of the cationic complex [(h3-
allyl)PdL2]ꢀ may be minimized. This will depend on the
initial concentration of [Pd(h3-allyl)(m-Cl)]2 which con-
trols the chloride ions concentration.
Whatever the catalytic precursor, when chloride ions
are purposely added in a catalytic allylic substitution,
their concentration is high compared to the previous
situation. The competition between the cationic and the
neutral complexes in their reaction with the nucleophile
may then be in favor of the neutral complex h1-allyl-
PdClL2.
In both situations, the rate as well as probably
the regioselectivity of the reaction may depend on
the chloride concentration, which determines which
species, h1-allyl-PdClL2 or/and [(h3-allyl)PdL2]ꢀ, are
involved in the nucleophilic attack. This illustrates
the fact that the chloride ions of the catalytic
precursors [Pd(h3-allyl)(m-Cl)]2 must not be considered
as ‘innocent’ ligands [27]. Work is in progress to test
the effect of chloride ions on the regioselectivity of
the reaction in the case of mono- and bidentate
ligands.
meter (250 or 200 MHz). FAB mass spectrometry was
performed on a JEOL MS 400 spectrometer. The UV
spectra were recorded on a spectrophotometer DU 7400
Beckman. A 1 mm length immersion probe equipped
with fiber optics (Hellma) was used for the reactions
performed at ꢁ25 8C. The conductivity measurements
/
were performed with a conductivity meter (CDM210
Radiometer Analytical) equipped with a cell whose
constant k was close to 1 cmꢁ1. Cyclic voltammetry
was performed with a homemade potentiostat and a
wave form generator GSTP4 (Radiometer Analytical).
The current was recorded on an oscilloscope Nicolet
301. The cyclic voltammetry was performed at 0.2 V sꢁ1
at a steady gold disk electrode (id 0.5 mm). The counter
electrode was a platinum wire (apparent area 0.2 cm2)
and the reference a calomel electrode SCE separated
from the cell by a bridge containing 3 ml DMF with
nBu4NBF4, 0.3 M. The cell contained 12 ml of DMF
with nBu4NBF4, 0.3 M.
General procedure for the conductivity measurements.
They were performed in
25 8C containing 15 ml of DMF. 0.03 mmol of
the allylꢃ
PdII complexes 3a or 5d were introduced
a cell thermostated at
/
and the conductivity measured just after introduction
and then versus time to monitor the complexes
stability.
General procedure for the kinetics of the reaction of
morpholine
[(h3-CH2ÃCHÃCH2)PdL2]ꢀBFꢁ4 (2aꢀ
C6H4)3P and 2bꢀ : Lꢂ
(4-CH3 ÃC6H4)3P) monitored
with
cationic
complexes
:
Lꢂ(4-ClÃ
/
/
/
/
by UV spectroscopy in the absence and presence of
chloride ions. The reactions were performed in a
thermostated cell equipped with a 1 mm length immer-
sion probe. A solution of 25 ml of DMF containing
2aꢀ; BFꢁ4 (12 mg, 0.0125 mmol) was transferred under
argon to the cell thermostated at ꢁ25 8C. The wave-
/
length was selected at 323 nm. After addition of the
appropriate amount of morpholine, the increase of the
absorbance was recorded versus time until a constant
value was reached (Fig. 3a). In other experiments,
various amount of nBu4NCl were introduced from a
mother solution in DMF, before the addition of
morpholine.
4. Experimental
General methods. All experiments were performed
using standard Schlenk techniques under argon atmo-
sphere. DMF was distilled on CaH2 and kept under
argon. The phosphines PPh3, (4-ZÃ
/
C6H4)3P (ZꢂCl,
/

374
T. Cantat et al. / Journal of Organometallic Chemistry 687 (2003) 365ꢃ376
/
4.1. [h3-CH2Ã
(2aꢀ,BF4ꢁ)
/
CHÃ
/
CH2)Pd{(4-ClÃ
/
C6H4)3P}2]ꢀBF4ꢁ
4.4. [(h1-CH2Ä
(3b)
/
CHÃ
/
CH2)PdCl{(4-CH3Ã/C6H4)3P}2
It was synthesized according to route A in Scheme 2.
A solution of tri(4-chlorophenyl)phosphine (0.366 g, 1
mmol) in acetone (3 ml) was added to a solution of
The complex was synthesized from [Pd(h3-CH2Ã
/
CHÃ
/
CH2)(m-Cl)]2 according to the procedure used for 3a
(route B in Scheme 2). H-NMR (200 MHz, CDCl3): d
2.34 (s, 18H, CH3 of ligand), 3.72 (m, 4H, CH2), 5.65
1
[Pd(h3-CH2Ã
/
CHÃ
/
CH2)(m-Cl)]2 (0.10 g, 0.27 mmol) in
acetone (5 ml), immediately followed by NaBF4 (0.161
g, 1.5 mmol) in water (6 ml). A gray precipitate was
formed which was isolated and dried: 0.425 g (81.5%
(quintet, 1H, JHH
JHH 8 Hz, aromatic H), 7.62 (d, 12H, JHH
aromatic H). 31P-NMR (103 MHz, CDCl3): d 21.17 (s).
MS (FABꢀ) C45H47ClP2Pd: m/zꢂ755 [Mꢁ
Cl]ꢀ.
ꢂ/10 Hz, central H), 7.45 (d, 12H,
ꢂ
/
ꢂ
/
8 Hz,
1
yield). H-NMR (200 MHz, CDCl3): d 3.80 (ddd, 2H,
/
/
/
JHH
JHH
ꢂ
/
13, JPH
7 Hz, syn-H), 6.02 (tt, 1H, JHH
7.32 (m, 24H, aromatic H). 31P-NMR
acetone-d6): 22.85 (s). MS
) C39H29Cl6P2Pdꢀ: m/zꢂ879 [M]ꢀ, 838 [Mꢁ
C3H5]ꢀ, 365 [Mꢁ 1 ligand]ꢀ.
C3H5ꢁ
ꢂ
/
7, JHH
ꢂ
/
6 Hz, anti-H), 4.01 (br. d, 2H,
ꢂ
/
ꢂ
/
13, JHH 7 Hz,
ꢂ
/
4.5. [h3-CH2Ã
C6H4)3P}2]ꢀBF4ꢁ (2cꢀ,BF4ꢁ)
/
CHÃ
/
CH2)Pd{(4-CF3Ã
/
central H), 7.19ꢃ
(103 MHz, DMFꢀ
(FABꢀ
/
/
d
/
/
/
The complex was synthesized from [Pd(h3-CH2Ã
/
CHÃ
/
/
/
CH2)(m-Cl)]2 (0.05 g, 0.137 mmol) according to the
procedure used for 2aꢀ; BFꢁ4 (route A in Scheme 2). The
precipitate was crystallized from dichloromethane. 0.23
4.2. (h1-CH2Ä
/
CHÃ
/
CH2)PdCl{(4-ClÃ
/
C6H4)3P}2 (3a)
1
g of white crystals was collected (72% yield). H-NMR
It was synthesized according to the route B in Scheme
2. A solution of tri(4-chlorophenyl)phosphine (0.64 g,
1.75 mmol) in acetone (5 ml) was added to a solution of
(250 MHz, acetone-d6): d 3.84 (ddd, 2H, JHH
JPH 7 Hz, anti-H), 4.32 (br. d, 2H, JHH 7 Hz, syn-
H), 6.26 (tt, 1H, JHH 12, JHH 7 Hz, central H), 7.62
ꢂ12,
/
ꢂ
/
ꢂ
/
ꢂ
/
ꢂ
/
[Pd(h3-CH2Ã
/
CHÃCH2)(m-Cl)]2 (0.16 g, 0.44 mmol) in
/
(s, 24H, aromatic H). 31P-NMR (103 MHz, acetone-d6):
d 24.04 (s).
acetone (15 ml). After evaporation of the solvent, an
orange precipitate was formed which was isolated as a
pure compound and dried: 0.62 g (77% yield). 1H-NMR
4.6. [(h1-CH2Ä
(3c)
/
CHÃ
/
CH2)PdCl{(4-CF3Ã
/
C6H4)3P}2
(250 MHz, CDCl3): d 3.72 (br. d, 4H, JHH
CH2), 5.61 (quintet, 1H, JHH 10 Hz, central H), 7.35ꢃ
7.4 (m, 24H, aromatic H). 31P-NMR (103 MHz,
CDCl3): d 21.55 (s). MS (FABꢀ) C39H29Cl7P2Pd: m/
zꢂ879 [Mꢁ Clꢁ
Cl]ꢀ, 838 [Mꢁ C3H5]ꢀ.
ꢂ/10 Hz,
ꢂ
/
/
The complex was synthesized from [Pd(h3-CH2Ã
/
CHÃ
/
/
CH2)(m-Cl)]2 (0.05 g, 0.137 mmol) according to the
procedure used for 3a (route B in Scheme 2). 0.18 g of
/
/
/
/
Complex 3a was also generated in situ by addition of
1 equiv nBu4NCl (9 mmol) to 2aꢀ; BF4ꢁ (9 mmol) in
CDCl3 (route C in Scheme 2). The reaction was
1
crystals was collected (59% yield). H-NMR (250 MHz,
acetone-d6): d 3.61 (d, 4H, JHH
(quintet, 1H, JHH
ꢂ10 Hz, CH2), 5.70
/
1
ꢂ10 Hz, central H), 7.61 (s, 24H,
/
quantitative. H-NMR (250 MHz, CDCl3): d 3.72 (br.
d, 4 H, CH2), 5.64 (quintet, 1H, JHH
aromatic H). 31P-NMR (103 MHz, acetone-d6): d 23.33
(s).
ꢂ10 Hz, central
/
H), 7.35ꢃ
/
7.4 (m, 24H, aromatic H). The 1H-NMR
spectrum also exhibits the four sets of protons of the
butyl group of nBu4Nꢀ.
4.7. [h3-PhÃ
C6H4)3P}2]ꢀBF4ꢁ (2a?ꢀ,BF4ꢁ)
/
CHÃ
/
CHÃ
/
CH2)Pd{(4-ClÃ
/
4.3. [h3-CH2Ã
C6H4)3P}2]ꢀBF4ꢁ (2bꢀ,BF4ꢁ)
/
CHÃ
/
CH2)Pd{(4-CH3Ã
/
The complex was synthesized from [Pd(h3-PhÃ
/
CHÃ
/
CHÃCH2)(m-Cl)]2 according to the procedure used for
/
The complex was synthesized according to the proce-
dure used for 2aꢀ; BF4ꢁ (route A in Scheme 2). The
2aꢀ; BFꢁ4 (route A in Scheme 2). Pale yellow crystals
1
(61% yield). H-NMR (250 MHz, CDCl3): d 3.67 (dd,
precipitate was crystallized from dichloromethaneꢃ
/pet-
1H, JHH
ꢂ
/
7, JPH
11, JPH
10, JPH
11, JHH
ꢂ7 Hz, syn-H of CH2), 3.72 (dd, 1H,
/
roleum ether. 0.33 g of pale yellow needles was collected
1
(72% yield). H-NMR (250 MHz, CDCl3): d 2.32 (s,
JHH
JHH
JHH
ꢂ
/
ꢂ
ꢂ
/
11 Hz, anti-H of CH2), 5.69 (dd, 1H,
10 Hz, PhCH-anti), 6.34 (ddd, 1H,
10, JHH 7 Hz, central H), 6.8ꢃ7.0 (m,
7.4 (m, 24H, aromatic H of ligand).
31P-NMR (103 MHz, CDCl3): d 22.95 (d, 1P, JPP
45
Hz), 24.62 (d, 1P, JPP 45 Hz). MS (FABꢀ
C45H33Cl6P2Pdꢀ: m/zꢂ955 [M]ꢀ, 838 [Mꢁ
PhC3H4]ꢀ.
ꢂ
/
/
18H, CH3 ligand), 3.48 (ddd, 2H, JHH
anti-H), 3.95 (br. d, 2H, Jꢂ6 Hz, syn-H), 5.93 (tt 1H,
JHH 13, JHH 6 Hz, central H), 7.07 (s, 24H, aro-
matic H). 31P-NMR (103 MHz, acetone-d6): d 22.02 (s).
MS (FABꢀ
) C45H47P2Pdꢀ: m/zꢂ755 [M]ꢀ, 714 [Mꢁ
C3H5]ꢀ.
ꢂ/13, JPH
ꢂ/7 Hz,
ꢂ
/
ꢂ
/
ꢂ
/
/
/
5H, H of Ph), 7.1ꢃ
/
ꢂ
/
ꢂ
/
ꢂ
/
ꢂ
/
/)
/
/
/
/
/

T. Cantat et al. / Journal of Organometallic Chemistry 687 (2003) 365ꢃ
/376
375
4.8. (h1-PhÃ
(3a?)
/
CHÄ
/
CHÃ
/
CH2)PdCl{(4-ClÃ
/
C6H4)3P}2
References
[1] (a) J. Tsuji, Palladium Reagents and Catalysts, John Wiley &
Sons, Chichester, 1996, p. 290;
It was synthesized from [Pd(h3-PhÃ
CH2)(m-Cl)]2 according to route B in Scheme 2 or by
addition of 1 equiv nBu4NCl to 2a?ꢀ; BF4ꢁ in chloro-
form (route C in Scheme 2). 1H-NMR (250 MHz,
/
CHÃ
/
CHÃ
/
(b) S.A. Godleski, in: B.M. Trost, I. Flemming (Eds.), Compre-
hensive Organic Synthesis, vol. 4, Pergamon, Oxford, 1991;
(c) C.G. Frost, J. Howard, J.M.J. Williams, Tetrahedron:
Asymmetry 3 (1992) 1089;
(d) G. Consiglio, R. Waymouth, Chem. Rev. 89 (1989) 257;
CDCl3): d 3.15 (d, 2H, JHH
JHH 13 Hz), 6.06 (dd, 1H, JHH
central H), 7.26ꢃ7.61 (m, 29H, H of Ph and aromatic H
of ligand). 31P-NMR (103 MHz, CDCl3): d 21.89 (s).
MS (FABꢀ) C45H33Cl7P2Pd: m/zꢂ955 [Mꢁ
Cl]ꢀ, 838
[Mꢁ
Cl-PhC3H4]ꢀ.
ꢂ
/
9 Hz), 5.28 (d, 1H,
(e) A. Pfaltz, Acc. Chem. Rev. 26 (1993) 339ꢃ345;
/
ꢂ
/
ꢂ
/
13, JHH 9 Hz,
ꢂ
/
(f) T. Hayashi, in: I. Ojima (Ed.), Catalytic Asymmetric Synthesis,
/
VCH, New York, 1993, p. 325;
(g) B.M. Trost, D.L. Van Vranken, Chem. Rev. 96 (1996) 395;
(h) B.M. Trost, Acc. Chem. Res. 29 (1996) 355;
/
/
/
/
(i) E.I. Negishi, Handbook of Organopalladium Chemistry for
Organic Synthesis, 2002, vol. 2, Wiley Interscience, New York,
pp. 1663ꢃ2027.
/
4.9. (h1-CH2Ä
/
CHÃ
/
CH2)PdCl(dppb) (5d)
[2] (a) C. Amatore, A. Jutand, G. Meyer, L. Mottier, Chem. Eur. J. 5
(1999) 466;
(b) C. Amatore, S. Gamez, A. Jutand, Chem. Eur. J. 7 (2001)
1273;
It was synthesized according to route B in Scheme 3:
dppb (0.114 g, 0.27 mmol) in acetone (5 ml) was added
(c) C. Amatore, S. Gamez, A. Jutand, G. Meyer, L. Mottier,
Electrochim. Acta 46 (2001) 3237.
to a solution of [Pd(h3-CH2Ã
/
CHÃCH2)(m-Cl)]2 (0.05 g,
/
0.136 mmol) in acetone (2.5 ml). After half an hour, the
solvent was concentrated. A pale yellow precipitate
appeared. After filtration, the solid was dissolved in
dichloromethane and crystallized in petroleum ether.
0.115 g was collected (70% yield). H-NMR (250 MHz,
CDCl3): d 1.91 (br. s, 4H, CH2 of dppb), 2.83 (br. s, 4H,
CH2 of dppb), 3.75 (br. m, 4H, CH2), 5.63 (quintet, 1H,
[3] C. Amatore, A. Jutand, M.A. M’Barki, G. Meyer, L. Mottier,
Eur. J. Inorg. Chem. (2001) 873.
[4] J. Powell, B.L. Shaw, J. Chem. Soc. A (1967) 1839.
[5] P. Braunstein, F. Naud, A. Dedieu, M.-M. Rohmer, A. DeCian,
S.J. Rettig, Organometallics 20 (2001) 2966.
[6] M. Kollmar, G. Helmchen, Organometallics 21 (2002) 4771.
[7] G. Malaise´, L. Barloy, J.A. Osborn, N. Kyritsakas, C. R. Chimie
5 (2002) 289.
1
[8] B. Akermark, G. Akermark, L.S. Hegedus, K. Zetterberg, J. Am.
Chem. Soc. 103 (1981) 2037.
[9] (a) J.E. Backvall, R.E. Nordberg, J. Am. Chem. Soc. 103 (1981)
JHH
ꢂ
/
10 Hz, central H of CH2ꢂ
/
CH Ã
/
CH2), 7.43 (br. s,
1
12 H, aromatic H), 7.55 (br. s, 8 H, aromatic H). H-
NMR (250 MHz, DMF-d7): d 1.18 (br. s, 4 H, CH2 of
dppb), 2.96 (br. s, 4H, CH2 of dppb), 3.75 (br. m, 4H,
CH2), 5.98 (quintet, 1H, JHH
CH2ꢂCH ÃCH2), 7.54 (br. s, 12H, aromatic H), 7.69
(br. s, 8H, aromatic H). 31P-NMR (103 MHz, CDCl3): d
18.47 (s). FABꢀ MS C31H33P2PdCl: m/zꢂ573 [Mꢁ
Clꢁ
Cl]ꢀ, 532 [Mꢁ C3H5]ꢀ.
¨
4959;
(b) R.E. Nordberg, J.E. Backvall, J. Organomet. Chem. 285
¨
ꢂ10.5 Hz, central H of
/
(1985) C24.
[10] M. Kawatsura, Y. Uozumi, T.J. Hayashi, Chem. Soc. Chem.
Commun. (1998) 217.
/
/
[11] (a) G.C. Lloyd-Jones, S.C. Stephen, J. Chem. Soc. Chem.
Commun. (1998) 2321;
/
/
/
/
/
(b) G.C. Lloyd-Jones, S.C. Stephen, Chem. Eur. J. 4 (1998) 2539.
[12] B.M. Trost, F.D. Toste, J. Am. Chem. Soc. 121 (1999) 4545.
[13] A mixed neutral complex (h3-CH2Ã
/
CHÃ
generated only by addition of two equiv. PPh3 to the dimer
[Pd(h3-CH2Ä
CHÃCH2)(m-Cl)]2 (Route F of Scheme 2) [3,4].
Attempt to synthesize such a complex with (4-CF3ÃC6H4)3P led
to a complex mixture.
[14] In a previous work [3], we established that the phosphonium salt
was formed from the neutral complex h1-RCHÄ
CHÃCH2Ã
/
CH2)PdCl(PPh3) 4 was
4.10. (h1-CH2Ä
/
CHÃ
/
CH2)PdCl(dppf) (5e)
/
/
It was synthesized according to the procedure used for
1
5d. Orange crystals (52% yield). H-NMR (250 MHz,
/
CDCl3): d 3.99 (br. m, 4H, ꢂ
Cp), 5.95 (quintet, 1H, JHH
CH2ÄCH ÃCH2), 7.48 (m, 12H, aromatic H), 7.60 (m,
8H, aromatic H). 31P-NMR (103 MHz, CDCl3): d 22.96
(s). (FABꢀ) C37H33ClFeP2PdCl: m/zꢂ701 [Mꢁ
Cl]ꢀ,
660 [MꢁClꢁ
C3H5]ꢀ.
/
CH2), 4.43 (m, 10H, H of
/
/
/
ꢂ/10.5 Hz, central H of
PdCl(PPh3)2 by a reversible reductive elimination which also
generated a Pd0 complex. This reaction became preponderant in
the presence of any chemical able to stabilize or to react with the
Pd0 complex (dba, extra PPh3, dioxygen) [3].
/
/
/
/
/
[15] For a review on the characterization of cationic PdII complexes by
conductivity measurements in DMF, see: A. Jutand, Eur. J. Inorg.
Chem. (2003) 2017.
/
/
[16] This value was calculated at 25 8C by addition of the conductivity
of 2aꢀ; BFꢁ4 ; 2 mM in DMF (69 ms cmꢁ1) determined above (see
text) to that of nBu4NCl, 2 mM in DMF (131 ms cmꢁ1) [17] and
subtraction of the conductivity of nBu4NBF4, 2 mM in DMF
(136 ms cmꢁ1) [2a].
[17] A. Jutand, A. Mosleh, Organometallics 14 (1995) 1810.
[18] The conductivity slightly increased to 20 ms cmꢁ1. This might be
due to the formation of a small amount of an allyl-phosphonium
salt [14].
Acknowledgements
This work has been supported by the Centre National
de la Recherche Scientifique (CNRS-ENS-UPMC,
UMR 8640) and the Ministe`re de la Recherche (Ecole
Normale Supe´rieure). We thank Johnson Matthey for a
generous loan of sodium tetrachloropalladate.

376
T. Cantat et al. / Journal of Organometallic Chemistry 687 (2003) 365ꢃ376
/
[19] This value has been calculated at 25 8C from the known
CH2)Pd(dppb)]ꢀAcOꢁ 2 mM in
absorption bands appeared at 296 and 356 nm. The kinetics was
monitored by recording the increase of the absorbance of the Pd0
at 296 nm.
conductivity of [(h3-CH2Ã
/
CHÃ
/
DMF (82 ms cmꢁ1) [2b] by addition of the conductivity of
nBu4NCl, 2 mM in DMF (131 ms cmꢁ1) [17] and subtraction of
the conductivity of nBu4NOAc, 2 mM in DMF (89 ms cmꢁ1) [2a].
[20] (a) J. Powell, B.L. Shaw, J. Chem. Soc. A (1968) 774;
(b) T. Hayashi, A. Yamamoto, T. Hagihara, J. Org. Chem. 51
(1986) 723.
[25] Since the reaction was performed under stoichiometric conditions,
according to the mechanism of Scheme 4, the kinetic law is then 1/
xꢂ
step (deprotonation). If k2ꢀ
2kh3C0tꢀ1.
/
kh3C0(2ꢀ
/
kh3/k2)tꢀ
/
1 if k2 is the rate constant of the second
/
kh3, the rate law is then 1/xꢂ
/
/
[21] M. Baboule`ne, J.L. Torregrosa, V. Spe´ziale, A. Lattes, Bull. Soc.
[26] For SN2? substitution of morpholine on allyl bromide see: G.
Butler, R.L. Bunch, J. Am. Chem. Soc. 71 (1949) 3120.
[27] For a review on the halide effects in transition metal catalysis, see:
K. Fagnou, M. Lautens, Angew. Chem. Int. Ed. 41 (2002) 26.
[28] Y. Takahashi, Ts. Ito, S. Sakai, Y. Ishii, J. Chem. Soc. Chem.
Commun. (1970) 1065.
Chim. Fr. 11ꢃ/12 (1980) 565.
[22] The kinetics was monitored at ꢁ/25 8C. Monitoring the kinetics
by cyclic voltammetry at such a low temperature was thus a
problem due to the high resistivity of the solution.
[23] O. Kuhn, H. Mayr, Angew. Chem. Int. Ed. 38 (1999) 343.
[24] In that case the complex exhibited an absorption band at 314 nm
which disappeared upon addition of morpholine. Two new
[29] P.M. Auburn, P.B. Mackenzie, B. Bosnish, J. Am. Chem. Soc.
107 (1985) 2033.